Development and application of glyco-analytical tools for ...

Development and application of

glyco-analytical tools

for biotechnology

A thesis submitted to the

University of Manchester

for the degree of

Doctor of Philosophy

in the

Faculty of Science and Engineering.

2017

Michel Riese

School of Chemistry

2

Table of Contents

Table of Contents .............................................................................................................................. 2

Abstract ............................................................................................................................................... 3

Declaration .......................................................................................................................................... 4

Copyright Statement .......................................................................................................................... 5

Acknowledgement ............................................................................................................................. 7

Structure of this thesis ....................................................................................................................... 8

Chapter 1 Introduction .................................................................................................................. 9 1.1 Carbohydrates ....................................................................................................... 9 1.2 Analytical tools for carbohydrates ...................................................................17 1.3 Applications for glyco-analytical tools ............................................................30 1.4 References ...........................................................................................................42

Chapter 2 Objectives of this thesis............................................................................................. 53 2.1 Simple, quantitative and non-destructive GOase assay ................................53 2.2 Carbohydrate arrays for fast and sensitive hydrolase characterisation .......54 2.3 Completing the N-acetylneuraminic acid toolkit ...........................................54 2.4 Glycolipids in Parkinson’s disease ...................................................................55

Chapter 3 Simple, quantitative and non-destructive Galactose Oxidase assay .................... 56 3.1 Summary ..............................................................................................................56 3.2 Contribution ........................................................................................................56 3.3 Introduction ........................................................................................................57 3.4 Target glucosides ................................................................................................60 3.5 Experimental Section .........................................................................................63 3.6 Results & Discussion .........................................................................................70 3.7 Conclusion ...........................................................................................................86 3.8 Appendix .............................................................................................................88 3.9 References ...........................................................................................................90

Chapter 4 Carbohydrate arrays for fast and sensitive hydrolase characterisation ............... 92 4.1 Summary ..............................................................................................................92 4.2 Contribution ........................................................................................................92 4.3 Manuscript ...........................................................................................................92 4.4 Supporting Information ................................................................................. 128

Chapter 5 Completing the N-acetylneuraminic acid toolkit ................................................. 142 5.1 Summary ........................................................................................................... 142 5.2 Contribution ..................................................................................................... 142 5.3 Manuscript ........................................................................................................ 142 5.4 Supporting Information ................................................................................. 162

Chapter 6 Endogenous modulation of neuronal dopamine transport ................................ 215 6.1 Summary ........................................................................................................... 215 6.2 Contribution ..................................................................................................... 215 6.3 Manuscript ........................................................................................................ 215 6.4 Supporting Information ................................................................................. 227

Chapter 7 Discussion and Outlook .......................................................................................... 236

Word count: 44032

3

Abstract

Carbohydrates are the most diverse family of biomolecules in nature. From a panel of

mono-saccharides organisms build a vast variety of glycans and glycoconjugates with

essential biological functions in energy metabolism, cellular communication and structural

integrity to name a few. The wide array of architectures found in glycans is orchestrated by

carbohydrate active enzymes which control glycosidic bonds between mono-saccharides and

perform additional carbohydrate modifications. In order to assess biological functions,

structural information is essential as is the ability to modify carbohydrates to use as biological

probes. This thesis addresses analytical tools to gain insights into carbohydrates and the

enzymes involved in the glycan metabolism.

An easy NMR assay is presented to monitor enzymatic oligo-saccharide oxidation with

minimal sample preparation, while the regio-selective oxidation products are valuable targets

as well as precursors for industrial biotechnology applications.

The adaptation of a glyco array platform for rapid screening for glycoside hydrolase

activities of fungal enzymes towards mixed oligo-saccharide libraries advances the analytical

possibilities and provides a tool for the identification of novel enzymatic activities.

While state-of-the-art N-glycan analysis solves the problem of isobaric linkage isomers

through the application of ion mobility, traditional methods heavily rely on exo-glycoside

hydrolases. The discovery and proven applicability of an α2,6-‘pseudosialidase’ completes

the analytical toolbox for N-acetylneuraminic acid terminated N-glycans.

The identification of glucosylsphingosine as an endogenous modulator of DAT-mediated

dopamine transport is an exciting discovery and may reveal a new dimension to the etiology

of Parkinson’s disease.

The methods presented in this thesis provide glycoscientists with tools to further analyse

glycans, CAZymes and their impact on biotechnology.

4

Declaration

No portion of the work referred to in the thesis has been submitted in support of an

application for another degree or qualification of this or any other university or other institute

of learning.

5

Copyright Statement

i. The author of this thesis (including any appendices and/or schedules to this thesis)

owns certain copyright or related rights in it (the “Copyright”) and s/he has given

The University of Manchester certain rights to use such Copyright, including for

administrative purposes.

ii. Copies of this thesis, either in full or in extracts and whether in hard or electronic

copy, may be made only in accordance with the Copyright, Designs and Patents Act

1988 (as amended) and regulations issued under it or, where appropriate, in

accordance with licensing agreements which the University has from time to time.

This page must form part of any such copies made.

iii. The ownership of certain Copyright, patents, designs, trade marks and other

intellectual property (the “Intellectual Property”) and any reproductions of copyright

works in the thesis, for example graphs and tables (“Reproductions”), which may be

described in this thesis, may not be owned by the author and may be owned by third

parties. Such Intellectual Property and Reproductions cannot and must not be made

available for use without the prior written permission of the owner(s) of the relevant

Intellectual Property and/or Reproductions.

iv. Further information on the conditions under which disclosure, publication and

commercialisation of this thesis, the Copyright and any Intellectual Property and/or

Reproductions described in it may take place is available in the University IP Policy

(see http://documents.manchester.ac.uk/DocuInfo.aspx?DocID=487), in any

relevant Thesis restriction declarations deposited in the University Library, The

University Library’s regulations (see http://www.manchester.ac.uk/library/

aboutus/regulations) and in The University’s policy on Presentation of Theses

6

Für Kai.

7

Acknowledgement

First and foremost, I would like to thank Professor Sabine L. Flitsch for taking me on as

a student and offering constant motivation and supervision throughout the project. Sabine’s

enabling and supportive guidance allowed me to pursue a variety of ideas.

I would also like to thank all members of the Turner-Flitsch lab - past and current - for

their help, scientific input and enjoyable working environment as well as their introduction

into the ‘Mancunian’ life style: Anthony, Niki, Shahed, Rachel, Matthew, Juan, Jack, Sasha,

Sarah, Will, Paula, Mark, Ian, Hannah, Peter, Isobel, Emma, Sarah, Liz, Chris, Chris, Chantel,

Susanne, Fabio, Lorna, James, Scott, Lucy, Bas, Cesar, Jolanda, Declan, Syed, Mirja, Paul,

Steph, Antonio, Jason, Kun and everybody whom I haven’t mentioned. I have made many

friends amongst you for which I am grateful. A special thanks to Matthew, Rachel and Mark

who remotely ensured I’d make it through to submission day.

In times like these, my gratitude extends to the European Commission for funding

‘TINTIN’ (Grant No. 266025) within the Marie Skłodowska-Curie Actions programme.

International networks like these enable scientific and cultural achievements and provide

excellent training opportunities. After all, I almost learned Spanish thanks to: Natalie, Aoife,

Teodora, Yasmina, Jesús, Carmen, Gerard, María, Nicolò, Moussa, Csaba, Lois, Cécile and

Gavin.

I would like to thank my family Undine & Heiko, Kai and Karin & Gerd for unconditional

support as well as fuelling my curiosity for decades.

A particularly big “Thank you!” to Antje who joined me for my ‘British adventure’ and

keeps on inspiring me with an unbelievable amount of support and love. I wouldn’t have

succeeded without you.

Thank you everyone!

8

Structure of this thesis

This thesis in presented in ‘Journal Format’. Chapters 1 and 2 provide an overall

introduction to the field of carbohydrates and carbohydrate active enzymes and describe

analytical challenges. Chapter 3 is presented in a classic format while Chapters 4, 5 and 6 are

presented in manuscript format. Chapter 4 has been published in Scientific Reports 7, Article

number: 43117 (2017) doi:10.1038/srep43117 on the 21st February 2017. Chapter 5 has been

submitted to ‘Glycobiology’ and accepted on the 20th December 2017. Chapter 6 has been

prepared in ‘Nature Letter’ format with an anticipated submission in 2018. Chapter 7

summarises the thesis’ achievements and compares them to current literature.

9

Chapter 1 Introduction

1.1 Carbohydrates

The linear paradigm of biological information that is stored in DNA, passed on to RNA

and applied in translated proteins is insufficient to explain biology on a molecular level. Along

with lipids and small molecule metabolites, carbohydrates are crucial mediators of

biochemical processes fine-tune signalling and recognition processes as well as fulfilling

structural roles and acting as energy intermediates.1

Compared to nucleic acids and peptides, carbohydrates form the most complex and varied

group of biopolymers. Whereas the former are linear hetero-polymers of which the

sequences are directly encoded, carbohydrates lack a molecular blueprint. Instead, their

synthesis is regulated through metabolic states and expression levels of carbohydrate-active

enzymes.

Figure 1.1: Scheme of the glycosidic bond (in ManNAc-β1,4-Gal) and overview of possible

isomerisation and/or substitution patterns defining carbohydrates.

As poly-hydroxylated ketones or aldehydes, carbohydrates or saccharides are structurally

diverse. The carbonyl motif is likely to exist as intramolecular hemiacetals producing five- or

OO

O

HOHO

OH

HOOH

OH

OH

4. anomericposition

3. regiochemistryof linkages

2. stereochemistry/identity

1. composition andsequence ofmonomers

Anatomy of glycans

NH

O 5. substitutions

Chapter 1 - Introduction

10

six-membered rings mutarotating in aqueous conditions. Depending on the configuration of

the asymmetric centres, various isomeric mono-saccharides are possible. Additionally, the

basic set of building blocks is subject to post-glycosylational modifications expanding the

range of mono-saccharide units (Figure 1.2).2

Figure 1.2: Panel of common mono-saccharides in human N-glycans. Hexoses glucose

(Glc), galacose (Gal), mannose (Man) and derivatives N-acetyl-glucosamine (GlcNAc) and

N-acetyl-galactosamine (GalNAc). Sialic acid (N-acetyl-neuraminic acid, Neu5Ac) is

characteristic in N-glycans in vertebrates.

1.1.1 Poly-saccharides

Mono-saccharides can polymerise to form di-, oligo- or poly-saccharides (n > 12) via

glycosidic bonds. A layer of complexity is added through a) the regiochemistry, b) the

conformation of the anomeric position and c) the branching of these linkages. The

importance of linkage variation is easily illustrated by the structural and functional difference

between the all-glucose polymers cellulose (β1,4-linked) and amylose (α1,4-linked). Whereas

long cellulose fibrils are essential for plant cell walls, plants use helices of amylose to store

energy in form of glucose units in order to relieve osmotic pressure. The addition of α1,6-

linkages to the structure of glycogen results in a heavily branched structure that is exploited

by muscle tissue to maximise the surface area for future hydrolysis.

O

OH

O

HOHO

HOOH

OH

OH

OH

OHOHO

OH

OH

OHOHO

OH OH

OH OH

NHO

OHOH

OHO

NH

OH

OHO

O

COOH

OH

HO

HN

HOOH

OH

O

D-glucose, Glc D-galactose, Gal D-mannose, Man

N-acetyl-D-glucosamine, GlcNAc N-acetyl-D-galactosamine, GalNAc N-acetyl-neuraminic acid, Neu5Ac


11

Furthermore, poly-saccharides are found in insects’ exoskeletons (chitin) or as rheology

modifiers in the extra-cellular matrix (glycosaminoglycans).

In contrast to poly-saccharides, glycoconjugates contain a non-carbohydrate aglycon.

Glycopeptides and glycolipids are naturally occurring glycoconjugates originating from the

process of glycosylation which involves the enzymatic transfer of a glycan from an activated

donor onto an acceptor substrate.

1.1.2 N-glycans

Glycans transferred onto an asparagine residue within the peptide’s consensus sequence

Asn-Xxx-Ser/Thr (where Xxx is any amino acid but proline) are described as N-linked

glycans, or N-glycans.3 Typically, Glc3Man9GlcNAc2 is pre-synthesised at the endoplasmic

reticulum membrane and co-translationally transferred from its dolichol phosphate anchor

by an oligo-saccharyltransferase followed by trimming and re-glycosylation along the

secretory pathway in eukaryotes.4

A multitude of diverse glycans coat mammalian cell surfaces and decorate glycoproteins5

to act as receptors or ligands in recognition processes like cell adhesion, immune system,

host-pathogen interaction and crucially fertilisation.6–9 Due to the essential roles of N-glycans

the effects of aberrant glycosylation are severe ranging from infertility10 and various forms

of cancer11–15 to foetal mortality16.


12

Figure 1.3: N-glycosylation pathway in mammals. Glc3Man9GlcNAc2 is transferred onto the

nascent peptide from the preformed glycan-dolichol phosphate precursor by an oligo-

saccharyl-transferase in the endoplasmic reticulum. Initial glycan trimming and folding of the

peptide is followed by further processing of the glycan antennae to form high mannose-,

complex- and hybrid-type N-glycans. From “Essentials of glycobiology”,

Varki et al.1


13

1.1.3 Glycolipids

Glycoconjugates with lipid linked to carbohydrates are described as glycolipids. The

subgroup of glycosphingolipids (GSL) is of great importance in vertebrate biology and shares

the characteristic amino alcohol sphingosine aglycon. Closely related to these are

glycocerolipids which are glycosphingolipids linked to a fatty acid via an amide bond.

Figure 1.4: Panel of glycolipids in mammals. Glucosylated ceramide is the focal point with

a multitude of glycolipids derived from it. Similarly, the aminoalcohol aglycon is found with

a free amino group, giving rise to the family of glycosphingosines. Adapted from Essentials

of glycobiology, Varki et al.1

Looking at the biosynthesis of glycosphingolipids, it is easy to understand why knock-

outs of the GlcCer synthase gene in mice are embryo-lethal: the entire family of GSLs

originates from the precursor glucoceramide.17 These experiments demonstrate the essential

roles of GSLs in developmental biology modulating intercellular coordination. Hence, GSLs

are particularly abundant in membranes of tissue that heavily depend on intercellular


14

communication, especially myelin sheets in the brain.18 As part of ‘lipid rafts’ in plasma

membranes, GSLs interact with pathogens19 and endogenous glycoproteins20 as well as

laterally with membrane-associated receptors to modulate signalling.21,22

While the GlcCer synthase knock-out genotype is embryo-lethal, the knock-out of

β-glucocerebrosidase, which catalyses the reverse reaction, is not. Because of the essential

role of glycolipids in the embryonic development of membrane-rich organs, the effects of

an anabolic enzyme deficiency emerge earlier compared to the effects of pathologic variants

of catabolic enzymes resulting in substrate accumulation causing sphingolipidoses, a subclass

of lysosomal storage disorders. The associated phenotype shows severe consequences,

especially mental effects due to the involvement of the central nervous system (see section

1.3.3 for glycolipids in Parkinson’s disease). The family of sphingolipidoses display diverse

phenotypes but are usually caused by a single-gene disorder.

1.1.4 CAZymes

The various groups of carbohydrates are structurally diverse and the resulting complex

structures fulfil crucial tasks in biology. Meticulous control over the biosynthesis of glycans

is essential to ensure correct function. Simple mono-saccharides are part of the energy

metabolism and therefore a significant proportion is consumed by catabolic processes.

Meanwhile, resulting energy equivalents are spent on the enzymatic synthesis of sugar

nucleotides, which serve as building block for highly-regulated anabolic processes.

Carbohydrate active enzymes (CAZymes) are employed by organisms to perform

individual steps in the process of glycan synthesis. The two main classes of CAZymes are

a) glycosyltransferases (GT, with over 380,000 members identified) and b) glycoside

hydrolases (GH, with over 500,000 members)a,23.

a Number of entries in “Carbohydrate Active Enzymes database”, http://www.cazy.org/, December 2017


15

Glycosyltransferases (EC 2.4.x.y) catalyse the transfer of a carbohydrate moiety from an

activated donor to a specific acceptor to form a specific de novo glycosidic bond retaining

or inverting the anomeric configuration.24,25 Mammalian Leloir-type GTs utilise sugar

nucleotides donors like UDP-Glc, -Gal, -GlcNAc, GDP-Man and CMP-Neu5Ac.26

Figure 1.5: Glycosyltransferase mechanism with inversion (top) or retention (bottom) of the

stereo-chemistry of the glycosidic bond.

Glycoside hydrolases (EC 3.2.1.x) catalyse the hydrolysis of existing glycosidic bonds at

the non-reducing end (exo) or within the glycan (endo) without the use of any energy-

intensive cofactors. Therefore, glycoside hydrolases perform mainly catabolic functions (e.g.

glycogen hydrolysis to provide rapid energy, degradation of biomass or pathogen defence as

part of the innate immune system).27,28

In biotechnological applications, glycoside hydrolases are exploited to degrade complex

mixtures of poly-saccharides to yield the desired compounds for further processing.29 The

OHOHO

OH

OH

OUDP

OHOHO

OH

OH

OUDP

RO

H

H

H

OHOHO

OH

OH

OUDP

ROH

B: BH+

OHOHO

OH

OH

H

UDP

OR

B:

OHOHO

OH

OH

OUDP

H

O

OHOHO

OH

OH

HO UDP

H

HO

OR

Inversion

Retension - ‘Koshland-type’

OHOHO

OH

OH

OUDP

H

H

OR OHO

HOOH

OH

OUDP

H

H

O ROHO

HOOH

OH

HO UDP

H

OR

OO O

HO

R

‡

Retension - SNi mechanism


16

main benefit of using enzymatic degradation over acid hydrolysis is the preservation of the

mono-saccharides generated from the hydrolysis which can be broken down in harsh acidic

conditions.30 For cellulose degradation in particular a more integrated approach to hydrolysis

is required due to the presence of lignin. Fungal genome sequencing revealed numerous

putative CAZymes which require biochemical characterisation to fully address the

biotechnological challenges.

Glycoside hydrolases also mediate biological signalling as well: Co- and post-translational

N-glycan trimming performed in the endoplasmic reticulum is GH-dependent and crucially

influences the residence time of the glycopeptide within the organelle to ensure correct

folding before migrating to the Golgi apparatus for further processing.

Another example of signalling involvement is the glycolipid homeostasis. As described

above, glycosphingolipids in particular form an essential part of the organism’s

communication. Therefore, the sequential and fine-tuned hydrolysis of glycolipids is

performed by a panel of exo-glycoside hydrolases to “switch off” signals conveyed through

the presence of these glycosylated lipids.


17

1.2 Analytical tools for carbohydrates

Given the complex nature and vast structural variety amongst glycans, analytical tools are

essential in the discovery of structure-function relationships. Routinely, glycans are subjected

to nuclear magnetic resonance spectroscopy (NMR), mass spectrometry (MS), glycan arrays

and liquid chromatography (LC) separation (following glycosidase treatment).

1.2.1 Carbohydrate nuclear magnetic resonance spectroscopy

Nuclear magnetic resonance (NMR) spectroscopy is a powerful analytical tool based on

the magnetic moment of nuclei in magnetic fields. Any nucleus with a spin quantum number

I > 0 possesses a magnetic moment (µ) proportional to the gyromagnetic ratio (γ) with µ = γI

and can therefore be detected by NMR spectroscopy. Therefore, nuclei with spin quantum

numbers I = ½ like 1H, 13C and 15N are especially important for probing carbohydrates and

other biomolecules. With a high natural abundance, 1H (99.98%) is the most sensitive

nucleus whereas 13C (1.11%) and 15N (0.37%) are far less abundant and require extended

acquisition time or isotopic labelling of the sample.

Changes in the individual chemical environment of a nucleus can affect its resonance

frequency through electronic shielding as observable through NMR experiments. Compared

to known standards the characteristic variation in chemical shifts allows for structural

elucidation of mono-, oligo-, and poly-saccharides as well as glycoconjugates.

Primary structural analysis of saccharides by NMR spectroscopy is performed in several

ways.31,32 Characteristic proton signals outside the bulk region (3-4 ppm) can act as

‘structural-reporter groups’.33 In particular, anomeric protons are found in the range of 4.5-

5.5 ppm which simplifies assignment. Integration of these resonances is a good starting point

to assess the number of mono-saccharides present. Information on the anomeric

configuration of 4C1 D-pyranoses can be extracted from the vicinal coupling constant

between H-1 and H-2 (axial-axial ~7 Hz, equatorial-axial ~4 Hz, axial-equatorial and


18

equatorial-equatorial <2 Hz)34, as well as the H-1 resonance with the α-anomer shifted

downfield compared to the β-anomer. Sialic acids which lack an anomeric proton may be

assessed via alternative signals (e.g. H-4 for Neu5Ac). The addition of non-carbohydrate

substituents (e.g. OMe, OAc, OSO3) can be identified by the characteristic downfield shifts

for protons (~0.2-0.5 ppm).35

Based on this information, databases36,37 provide access to composition, linkages,

sequence and structural motifs. In addition to 1H-NMR, 13C and 15N spectra can provide

useful information. Where the signal dispersion in the 13C channel is preferable compared to

1H, 15N is especially useful for amino- and N-acetylated saccharides. The first 2D-NMR

spectra of glucose were published in the 1980’s laying the foundation for structural studies

on complex carbohydrates to succeed.38

1.2.2 Mass spectrometry of carbohydrates

Mass spectrometry (MS) is a very versatile group of methodologies ultimately determining

the mass-to-charge (m/z) ratio of an analyte. The process can be separated into two sub-

processes: 1) ionisation and 2) mass analysis (separation and detection).

Carbohydrates, especially oligo- and poly-saccharides, are structurally complex as

described earlier. For MS to provide the desired structural information careful consideration

of the experimental setup is essential.

1.2.2.1 Ionisation

Analytes have to be ionised to determine the mass-to-charge ratio. Common glycans

ionise in positive mode with the addition of a small cation (e.g. H+, Li+, Na+, etc.), whereas

inherently charged species (e.g. phosphoglycolipids in negative mode) may ionise readily.39

In order to retain the structural properties of the analyte particular attention has to be

paid to the ionisation conditions. Oligo- or poly-saccharides are likely to fragment in source

causing glycosidic bonds to break. The loss of structural information can be detrimental to


19

the analytical value. Terminal sialic acids in particular are prone to dissociate from their

aglycon under harsh conditions.40

Amongst common ionisation techniques soft methods with a low chance of in-source

fragmentation can be distinguished from hard methods known to heavily fragment sample

materials. In carbohydrate applications, the two most commonly used techniques are

electrospray ionisation (ESI) and matrix-assisted laser desorption ionisation (MALDI).

ESI is an attractive method for the ionisation of biomolecules41,42 because it is unlikely to

cause fragmentation, can be used at atmospheric pressure and can be set up in a liquid

chromatography workflow providing on-line ESI-MS data. Additionally, sample quantities

required are low due to flow rates between nL/min to µL/min. To ensure optimal ionisation

efficiency, desalting prior to analysis is essential.

Liquid sample is pumped into a capillary which is subjected to a large electric field (106

V/m). This charge polarises the sample material and draws the continuous flow towards the

counter electrode. At the nozzle of the capillary a cone is formed and subsequently

destabilised with increasing force. The resulting jet releases charged droplets repelling each

other. These spreading droplets are electrostatically attracted and accelerated towards the

counter electrode (Figure 1.6).43,44


20

Figure 1.6: Schematic depiction of the electrospray ionisation (ESI). Analyte is pumped

through a capillary with a voltage applied, forming charged droplets. Solvent is displaced

(according to the (a) ion evaporation model or (b) charge residue model) and charged analyte

travels to the mass analyser. Adapted from Konermann et al.45

During the travel phase, the charged droplets start to shrink due to evaporation to a point

of destabilisation caused by repulsive Coulomb forces. Fission causes the droplets to shrink

even further ultimately releasing de-solvated ions according to unclear mechanisms.

According to the ion evaporation model46 (IEM, Figure 1.6 a) which is prevalent for small

molecules like glycans, analyte ions are ejected from a critically charged droplet, whereas

larger molecules (e.g. proteins) are believed to ionise according to the charge residue model47

(CRM, Figure 1.6 b). Here, the analyte is left with the droplet’s charge following solvent

evaporation producing multiply charged species.

Another soft ionisation method routinely used in carbohydrate analysis is matrix-assisted

laser desorption ionisation (MALDI). Unlike ESI, analytes are desorbed and ionised from a

solid phase through UV irradiation with a pulsed laser.43,48 The solid phase is created by co-

crystallisation of sample material and an excess of matrix on a conductive surface, usually

stainless steel or gold (Figure 1.7). The choice of matrix depends on the specific application

and can influence the quality of the results. Generally, matrices can be categorised as aromatic


21

organic acids (e.g. 2,5-dihydroxybenzoic acid as ‘all-round matrix’, α-cyano-4-hydroxy-

cinnaminic acid for peptides or 2,4,6-trihydroxyacetophenone for glycans). These

compounds fulfil two objectives providing a source or sink of protons based on the

polarisation of the instrument and secondly absorbing and distributing the UV radiation

emitted by the laser.49

Figure 1.7: Schematic depicting a MALDI source. Analyte (red) is co-crystallised with an

excess of matrix (grey).

Following in vacuo irradiation, the photo-ionised matrix desorbs and expands rapidly

transitioning into the gas phase accompanied by analyte molecules.43 For the analyte

ionisation two models have been proposed: according to the ‘Lucky Survivors’ theory,

charged analytes are preformed prior to crystallisation which ultimately survive sublimation.

Alternatively, secondary collision between charged matrix and analyte in the gas phase are

suggested to transfer ionisation.48,50 Due to the abundance of matrix (fragment) ions the

resolution of analytes below m/z of 500 is poor and often lost in background noise.


22

1.2.2.2 Mass analyser

In order to determine m/z values of the generated analyte ions, mass analysers are

connected in-line to ion sources. Frequently ESI sources producing a constant stream of

analytes are paired with quadrupole mass analysers (QMS), whereas gated MALDI sources

work well with time-of-flight (ToF) analysers.

The central element in quadrupoles are four electrodes arranged in parallel at the vertices

of a square, with opposing rods connected to the same direct current (DC) polarity. This

potential is overlaid with a potential in an alternating radio frequency (RF). This setup allows

for ions travelling through to be focussed or deflected and subsequently discharged. Based

on mass and charge state, ions are more or less likely to be focussed by a particular pair of

electrodes.51 By altering the DC and RF parameters, stable trajectories for specific m/z

species are formed allowing ranges of ions to be scanned. Especially the scanning speed and

overall compact build render quadrupole mass analysers a popular choice in LC-MS systems.

In ToF mass analysers, ions are accelerated by an electrical potential (V) and travel

through a vacuum of given length (L). The time-of-flight is inversely proportional to the

velocity, which can be expressed as

𝑇𝑂𝐹 = 𝐿 ∙ '𝑚

2𝑧𝑒𝑉

with the mass (m) and charge (z) of the analyte and the charge of an electron (e). Therefore,

ions with smaller m/z reach the detector first.43 To compensate for slight differences in

kinetic energy, reflectron mirrors can be used to re-focus species with identical m/z values

and hereby increase the resolution of the mass analyser at the expense of sensitivity.52

Following separation according to their respective mass-to-charge ratio, ions hit the mass

detector. Electron multipliers or microchannel plate detectors are commonly used to detect

analyte ions exploiting the principle of secondary emission, where an incoming particle


23

induces the emission of secondary particles from a suitable material. Stacking these elements

can lead to a cascading effect which results in electronic amplification of the initial signal.

While ESI-QMS is often very convenient as biological carbohydrate samples are dissolved

in aqueous solution, its salt tolerance is comparably low. Additional sample preparation (i.e.

desalting) is required. However, an LC-MS setup with an in-line ESI-Quad can be employed

as a form of sample preparation, but with limitations regarding the high-throughput potential

of the methodology.

MALDI-ToF is popular due to its versatility in sample requirements and low in-source

fragmentation. Very small amounts of solid or analyte solution can be analysed following co-

crystallisation with matrix which is rapid. While this ionisation technique is more salt tolerant,

MALDI is limited to analytes of approximately m/z > 500 as a result of the abundance of

matrix-derived ions.

Neither of the two methods described bears the potential to fully characterise

carbohydrates. Due to the high degree of isomerisation amongst mono-saccharides, m/z

values alone can be insufficient for the precise assignment of glycan structures with various

linkage positions and configurations (i.e. α1,6- vs. β1,4-linked) or even isobaric mono-

saccharide components (Glc vs. Man).

1.2.2.3 Ion mobility

Recently, ion mobility spectrometry has gained popularity in the glyco-analytical

community as is can be used in conjunction with mass spectrometry (IM-MS) and is able to

provide additional information on isobaric carbohydrates. The application of IM-MS adds a

new dimension to the field. Analytes are separated based on their size and shape, specifically

their rotationally averaged cross section area to charge ratio (Ω/z), with the collision cross-

section (CCS) Ω calculated from the drift time.53


24

This thesis focusses on travelling-wave IMS (TWIMS), where a stacked ring ion guide

(SRIG) consists of ring electrodes through which a symmetric potential wave propagates

continuously. While the direct current applied to the ring electrodes focusses the ions

laterally, analyte ions are propelled through the device by ‘surfing’ the potential wave in case

of higher-mobility species and therefore arriving earlier. Lower mobility ions tend to ‘roll

over’ and reside inside the drift cell for longer (Figure 1.8).54

Figure 1.8: Schematic mechanism of travelling wave ion mobility spectrometry. Analytes

interact with a travelling wave are separated based on their mobility (blue - low, red - high).55

Based on calibration experiments under stable conditions in the drift cell, the CCS can be

calculated. However, for the purpose of (e.g. isobaric) separation, TWIMS can be used

without calibration in-line as IM-MS.

With the help of ion mobility separation of known saccharide libraries, it is now possible

to determine glycan structures in unprecedented detail based on mass as well as

morphology.56,57 This key methodology elevated the discipline of structural carbohydrate

analysis to a new level. The discrimination between isobaric mono-saccharide species as well


25

as linkage conformers is accessible through the mobility separation of characteristic

fragments. However, current limitations exist with regards to the availability of glycan

standards and instrumentation, as well as the throughput of the techniques.

1.2.3 Glycan arrays

While the structural assignment of glycans is a critical element, further biochemical

characterisation is required to understand their biological function entirely. This

characterisation depends on the specific carbohydrate but might involve its biosynthetic

route, degradation pathway, or identification of interaction partners from complex mixtures

in varying conditions like temperature or pH. With numerous biological functions still to be

assigned, functional glycan analysis faces several challenges:

First of all, to analyse the number of experiments required to determine the various factors

involved a high-throughput format with little sample use is beneficial.58 Obtaining pure

carbohydrates from biological sources is challenging due to the microheterogeneity of

glycans. Therefore, miniaturisation is important due to sample quantity constraints.

With regards to glycan-binding, multivalency effects are notoriously difficult to mimic in

solution, which is why an immobilised setup can result in a more accurate testing

environment. If combined with a conductive surface, ligand immobilisation can be exploited

as a platform for MALDI-ToF MS. Traditional approaches like isothermal calorimetry,

surface plasmon resonance or enzyme-linked lectin assays depend on difficult to obtain

quantities of analyte and lack high-throughput capabilities.59 Finally, carbohydrate binding is

greatly influenced by clustering effects which are often disregarded in the aforementioned

methods.60

These challenges can be overcome with carbohydrate (micro-)arrays, alternatively known

as glycan arrays, which consist of a number of immobilised substrates on a solid support

allowing for control over spatial distribution.59 Glycoconjugates have been presented in


26

various array formats, for example on silica61, in ELISA-type well plates62 and more recently,

on functionalised self-assembled monolayers (SAM) on gold.63 In particular SAMs on gold

mimic the physicochemical properties of cell membranes64 and are easy to analyse in situ

through MALDI-ToF MS without prior derivatisation or labelling. As a tuneable platform,

glycan arrays are suited to study glycan-lectin binding58,65,66, bacterial adhesion67,68 and

enzymatic reactions69–71 in a high-throughput manner with small sample quantities (<10-6 L).

Figure 1.9: Schematic representation of glycan arrays. Depicted: hydrophobic self-

assembled monolayer on gold-coated steel targets for MALDI-ToF. Glycan arrays like these

are used to analyse the on-chip synthesis or degradation as well as the lectin-binding of

glycans.

Similarly, enzymatic carbohydrate transformations benefit from immobilised array

formats. The reduction in analyte quantity can be valuable when sample or enzyme supplies

are limited. Furthermore, product identification is aided by the compatibility of the platform

with MALDI-ToF MS. This is especially the case in multistep syntheses with successive

enzymatic steps the format allows for sequential reaction control.72 Beloqui et al. used a

similar platform to screen glycoside hydrolase activities from natural sources.73


27

1.2.4 Liquid chromatography of carbohydrates

In addition to the above mentioned analytical tools, liquid chromatography (LC) is applied

as separation methodology to further increase the analytical resolution.74–78 The underlying

principal is based on differing affinities between analytes and stationary phase. Poorly

retained analytes travel faster in the mobile phase due to higher solubility and consequently

elute faster than well-retained compounds. Two column materials are commonly used for

glycan analysis applications (Figure 1.10).

C18 reversed-phase (RP) columns are used to separate hydrophilic carbohydrates which

require derivatisation to increase hydrophobicity and improve separation. This step is

combined with the introduction of chromophores to utilise spectroscopic detection,

traditionally, by reductive amination of the aldehyde moiety with aniline derivatives.79,80

The introduction of hydrophilic interaction chromatography (HILIC) with polar

stationary and less-polar, organic mobile phases (acetonitrile) in combination with ultra-

performance liquid chromatography (UPLC/UHPLC) gives access to rapid baseline-

separation of underivatised glycans.81,82 An increasingly polar gradient allows for the elution

of hydrophilic carbohydrates. With eluents dissolved in comparably high organic solvents

which evaporate readily, subsequent in-line ESI-MS is simplified making HILIC separation

a popular choice. The recent arrival of Water’s commercial GlycoWorks system integrates

enzymatic N-glycan release, labelling and separation.12,83


28

Figure 1.10: Column materials commonly used in glyco-specific applications. Reverse phase

(RP) requires derivatisation to improve separation. Hydrophilic interaction chromatography

(HILIC) is preferred as analytes elute in high organic solvents, improving ESI compatibility.

Usually, elution is detected by spectroscopic systems (if chromophores are present) or in-

line mass spectrometers. Subsequent analyte identification is achieved by comparison to

known (external) standards aided by databases containing compound chromatograms and

spectra (e.g. GlycoBase).77,78 To ensure globally standardised experimental parameters,

retention times are referenced to dextran polymer ladders as ‘glucose units’ (GU).84

LC-MS analysis is applicable in structure confirmation experiments, where known

glycosylation patters are compared to previous chromatograms of similar samples.

Quantification of glycoforms of patterns of heterogeneously glycosylated samples is not

feasible with techniques described earlier. For de novo structure identification, more detailed

analysis is required to resolve larger glycans in particular. The impact on retention times

caused by single linkage isomers for example is minimal and therefore requires an elaborate

separation to achieve resolution. Additional glycosidase treatment can improve the

experimental elucidation dramatically and has been the gold standard in structural

identification.

SiO

OCH3

SiO

OO

O NH

OR

15O

RP column material

HILIC column material


29

1.2.5 Docking

For some applications, the tools described so far are not sufficient to answer biological

questions due to solubility limitations, especially within cell biology where lipid bilayer

membranes are more complex than just two-dimensional matrices. Amphiphilic glycolipids

are integral parts of these subcellular structures which makes their functional assessment

even more challenging as the majority of biochemical assays have been developed for

aqueous systems.

With significant advances in computational power and databases filled with three-

dimensional structures, molecular docking is currently more feasible and has become a useful

tool in life sciences. Molecular docking experiments predict orientation, location and

subsequently affinity between interacting molecules simulating molecular recognition in silico.

Typically, small molecule ligands with various degrees of freedom are docked to predicted

binding sites in rigid target proteins within a force field. Possible ligand orientations are

iterated and intermolecular complexes between the rigid target and the ligand conformers

are evaluated by a scoring function similar to this85:

∆𝐺/012013 = ∆𝐺526 + ∆𝐺898: + ∆𝐺;/<12 + ∆𝐺28=<95 + ∆𝐺><?=

Based on repulsive and electrostatic forces, hydrogen bonding, desolvation energy and

torsional entropy the binding energy is calculated. Docking compound libraries or families

of ligand poses and comparison of the scoring results is a suitable method to understand the

molecular architecture of the interaction. Traditionally, molecular docking results find wide

application in pharmaceutical drug discovery where virtual screening can support the

identification of lead compounds.86 Few examples can be found where molecular docking

approaches provide insights into the biology of amphiphilic glycolipids.87,88 However,

molecular docking cannot replace biochemical characterisation of molecular interactions and

will remain supportive tool.


30

1.3 Applications for glyco-analytical tools

1.3.1 Industrial biotechnology

One of the first successful examples of industrial biotechnology involving carbohydrates

was the process of anaerobic alcoholic fermentation turning glucose into carbon dioxide and

the valuable commodity ethanol. To this day, these whole cell catalysts are of importance in

the food and drink industry.

Lately, traditional chemical synthesis, especially in the fine chemicals and pharmaceutical

industry, has been combined with or replaced by biocatalytic processes.89 The success of

biocatalysis can be explained by two main reasons: Firstly, biocatalysis is able to accomplish

multiple challenging chemical steps at once and secondly, it is able to replace toxic heavy

metal catalysts for aqueous reactions at ambient temperatures to advance “green chemistry”.

Enzymes act as (bio-)catalysts to provide easy access to several classes of chiral

compounds like optically pure amines. Particularly challenging regio-, stereo- or chemo-

selectivity is achievable due to the nature of enzymes. Their complex three-dimensional

structure consisting of chiral L-amino acids results in precise targeting of substrates and

inherent stereo-selectivity.

Beyond the natural enzymes with limited substrate promiscuity “enzyme engineering”

broadens the possibilities. Optimising the biocatalyst’s amino acid sequence by means of

molecular biology gives rise to a wider range of tailored enzymes with desirable compound

specificities and improved activities for substrates of interest.90,91

Additionally, enzyme-catalysed reactions can be more sustainable and environmentally

friendly. Proteins are biocompatible and biodegradable, work in ambient aqueous conditions

and with the potential for improved economy. These benefits can solve common problems

like organic solvent or heavy metal catalyst usage as well as high energy costs.92


31

As mentioned before, carbohydrate chemistry is governed by stereo- and regio-selective

chemistry which is orchestrated by CAZymes in nature. The industrial exploitation of

carbohydrates as feed stocks, materials and pharmaceuticals requires glyco-analytical tools in

order to characterise properties and profile activities. The aforementioned (CAZymes,

section 1.1.4) specific hydrolysis of plant-derived poly-saccharides is a suitable example: The

generation of higher-value compounds for poly-saccharides requires characterisation of the

biocatalysts as well as the products generated.

1.3.2 Pharmaceutical quality control

Over half of the top-selling biopharmaceutical drugs and therapeutic proteins, mostly

monoclonal antibodies, are glycosylated.93 Several examples of glycosylated therapeutics are

summarised in Table 1.1.

Production of ‘biologics’ is realised in cell culture or fermentation processes which creates

challenges regarding the control of glycosylation profiles and particular glycan isoforms.94,95

As glycosylation is not directly genetically encoded but controlled through the expression

patterns of CAZymes, multiple isoforms with varying physicochemical properties are

generated. Pathway analysis and metabolic engineering resulted in production cell lines with

a tuneable spectrum of glycosylation patterns.96–98 However, absolute control is still

unachievable. Commercial solutions to separate peptide production and in vitro glycan

synthesis are available.99 This two-step process offers versatility at increased production cost.

Glycosylation as a post-translation modification performing a multitude of biological

functions in therapeutic proteins modulating in vivo clearance100,101, efficacy102 and

immunogenicity103. Therefore, precise control over glycosylation is essential for the efficacy

and safety of biopharmaceuticals.102

In human IgGs, Asn297 in the Fc region is glycosylated in a typical bi-antennary complex

core type fashion which is crucial for the biological function mediated via antibody-


32

dependent cellular cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC).

Hence, precise control over glycosylation profiles are vital for the safety and efficacy of

therapeutic IgGs.104

Intravenous immunoglobulin samples for example contain approximately 19% (mono-

and di-)sialylated glycoforms. α2,6-sialylation is predominant in humans, whereas Chinese

hamster ovary cells produce α2,3-sialylated recombinant IgGs.105 In murine myeloma cells,

terminal sialylation is found to contain non-human Neu5Gc instead of Neu5Ac, which is

potentially immunogenic in humans as 85% of the healthy population express anti-Neu5Gc

antibodies.106,107

The therapeutic IgG Cetuximab contains 30% α1,3-galactosylated glycans with led to

cases of anaphylactic reactions in the US due to anti-α1,3-Gal antibodies in patients receiving

Cetuximab treatment.103

In conclusion, the control during production as well as the analysis of glycoforms during

development and quality control are crucial for economic and clinical safety aspects. In order

to orchestrate the multitude of glycan profiles precise tools for analysis and glycan

engineering are required. This includes optimised production hosts, as well as highly-specific

(exo-)glycosidases together with LC-MS systems.


33

Table 1.1: Panel of selected glycosylated therapeutic proteins approved for clinical

application. Excerpt from Solá et al.102

Non-proprietary name Name (Company) Indication Glycans Production Alpha 1-antitrypsin Prolastin®

(Talecris Biotherapeutics)

Treatment of congenital α1AT deficiency with

emphysema

3 N-Linked Tissue fractionation

(human placenta)

Antithrombin III Atryn® (Ovation

Pharmaceutics)

Prevention of peri-operative and peri-

partum thromboelitic events

4 N-Linked Milk fractionation (transgenic

goats) C1-esterase-inhibitor Cinryze®

(CSL) Treatment of

hereditary angioedema 6 N-Linked 7 O-Linked

Plasma fractionation

(human) Darbopoetin alfa ARANESP®

(Amgen) Treatment of anemia

associated with chronic renal failure

5 N-Linked 1 O-Linked

CHO cells

Epoetin multiple Treatment of anemia associated with chronic

renal failure (CRF)


CHO cells

Eptacog alfa (CF VIIa) NovoSeven® (Novo Nordisk)

Treatment of spontaneous and

surgical bleedings in haemophilia A and B


BHK cells

Fibrinogen Haemocomplettan® (CSL)

Haemorrhagic diatheses in hypo-,

dys-, or afibrinogenaemia

5 N-Linked Plasma fractionation

(human)

Idursulfase Elaprase® (Shire)

Treatment of Mucopolysaccharidosis

I

8 N-Linked Human cells

Insulin multiple Treatment of diabetes multiple multiple Interferon beta-1a Rebif®

(Pfizer) Treatment of multiple

sclerosis 1 N-Linked CHO cells

Urokinase alfa Abbokinase® (ImaRx

Therapeutics)

Treatment of acute massive pulmonary

emboli


HK cells


34

1.3.3 Glycolipids in Parkinson’s disease

Parkinson’s disease (PD) is a neurodegenerative disease affecting the central nervous

system (CNS), in particular the motor system through the loss of dopaminergic neurons

(Figure 1.11). With an incidence of 1-2 per 1000 people and 1 in 100 people over 60 years

PD108 is the second most common neurodegenerative disease.109

Figure 1.11: Schematic of a dopaminergic synapse. Dopamine (red symbols) is

synthesised and packaged into vesicles, where it is stored. Upon arrival of a neuronal

stimulus, vesicles fuse with the pre-synaptic membrane to release the neurotransmitter into

the synaptic cleft. Dopamine receptors on the post-synaptic membrane are activated and

subsequently cause adjacent ion channels (not shown) to open to complete

neurotransmission. Dopamine is recycled via the dopamine transporter (DAT) and either

repackaged into dopaminergic vesicles or broken down via monoamine oxidase.


35

The first outline of the disease’s symptoms was published by James Parkinson’s “Essay

on the shaking palsy” in 1817 describing the characteristic bradykinesia, tremor and rigidity

as well as mental effects.110 The mental disorders like e.g. depression, psychosis, sleep

disorders and dementia (reviewed in Chaudhuri et al., 2006) advance with age and severity of

the disease.111 To this day, diagnosis is based on these ambiguous clinical criteria.112 No

reliable sensitive biomarkers are available that could aid diagnostic accuracy. However, the

dopamine transporter (DAT) and vesicular monoamine transporter 2 (VMAT-2) are

promising targets for neuroimaging (Figure 1.11).113

As a cure for PD is still unavailable, disease management focusses on pharmacological

treatment to reduce the symptoms. Therapeutic approaches are centred around replacing the

depleted dopamine levels. Levodopa (L-Dopa, Figure 1.12) is a dopamine precursor that can

penetrate the blood-brain barrier where it is converted and can diminish motor symptoms.114

Other strategies are based on the administration of dopamine agonists with a multitude of

side effects or monoamine oxidase inhibitors to reduce dopamine catabolism and thereby

extend neurotransmission.115

1.3.3.1 Etiology of Parkinson’s disease

While the cause of PD is unknown in most cases, PD is neuropathologically characterised

through the presence of proteinaceous aggregates called Lewy bodies containing α-synuclein

(α-syn) in dopaminergic neurons located in the substantia nigra pars compacta. The resulting

death of the affected neurons leads to impaired motor neuron stimulation causing a reduction

of voluntary movement facility.109 Alpha-synuclein is mainly found in the brain and is

involved in the neuronal vesicular transport in presynaptic terminals. It has been shown that

α-syn interacts with phospholipids and is partially structured.116 Furthermore, α-syn has a

tendency to oligomerise and subsequently polymerise to form fibrils which accumulate to

form plaques. Whereas the role of Lewy bodies for the death of neurons is still debated, the


36

removal of α-syn through aggregation does affect the neurons capability to traffic dopamine

efficiently causing neurodegeneration.117

Oxidative stress due to increased levels of reactive oxygen species (ROS) is attributed to

dopamine metabolism and dopamine auto-oxidation (Figure 1.12).118,119 Dopamine oxidises

to dopamine-quinone via the semi-quinone compound reducing oxygen to superoxide.

Additionally, the metabolism of dopamine via monoamine oxidase produces homovanillic

acid and equivalents of hydrogen peroxide.

PD patients’ substantia nigra have been shown to have decreased levels of antioxidants. As

a consequence, PD neurons are predisposed to suffer from ROS poisoning through DNA

damage, lipid oxidation of electron transport chain uncoupling resulting in cell death.120–122

Furthermore, it has been shown that oxidative stress can promote α-syn aggregation,

linking the cytotoxic events creating a pathogenic web (Figure 1.14).123 Alongside Lewy body

formation and oxidative stress, altered protein handling is suggested as a pathogenic

mechanism.124 Mitochondrial dysfunction has been linked to idiopathic PD explaining the

implications of (environmental) neurotoxins like rotenone and 1-methyl-4-phenylpyridinium

(MPP+) which interfere with Complex I in the electron transport chain.125


37

Figure 1.12: Dopamine metabolism. Tyrosine is hydroxylated to L-Dopa by the tyrosine

hydroxylase (TH), before the rate-limiting decarboxylation via amino acid decarboxylase

(AADC) gives dopamine. Dopamine is broken down into homovanillic acid (HVA) via a

cascade of monoamine oxidase (MAO), aldehyde dehydrogenase and catechol-O-methyl

transferase (COMT) releasing hydrogen peroxide. In the presence of oxygen, dopamine can

auto-oxidase to the quinone producing superoxide. DOPAC: 3,4-dihydroxyphenylacetic

acid, 3-MT: 3-methoxyl-tyrosine.

HO

O

OHNH2

HO

O

OHNH2

HO

HO

HO NH2

HO

HO O

HO

MeO NH2

HO

MeO O

O

O NH2

HO

O NH2

MAO

MAO

COMT

COMT

TH

AADC

O2

O2 O2

O2

O2

H2O2

O2

H2O2

tyrosine

L-Dopa

dopamine

3-MT

OH

OH

DOPAC

dopamine-semiquinone

dopamine-quinone

HVA


38

1.3.3.2 Risk factors

The most prominent risk factor for PD is age. Age impacts cellular physiology in many

ways: oxidative stress increases, decreased lysosomal capability and progressive decline of

glucocerebrosidase to mention only a few.126,127 Decreased ROS tolerance is important, as the

disturbed dopamine metabolism puts aged neurons under additional stress. Decreased

neuronal lysosomal capacity, causing a direct loss of normal proteolytic activity, leads to an

inability to clear α-syn aggregates.

Reports of a predisposition of cocaine users to PD is based on the drug’s affinity for the

dopamine transporter. Blocking DAT-mediated re-uptake of the neurotransmitter depletes

presynaptic cytosolic dopamine and extends signal duration. Additionally, remaining

dopamine, as well as cocaine are metabolised producing neurotoxic ROS in the process.128,129

Amongst genetic risk factors, pathogenic mutations in the SNAC gene have been

identified to increase α-syn’s tendency to oligomerise.130,131 SNAC-associated forms of PD

show a decreased response to levodopa therapy and generally earlier onset of the disease.109

A significant discovery is the correlation between heterozygous mutations in the GBA1

gene found in Gaucher’s disease and PD.132 GBA encodes the lysosomal

β-glucocerebrosidase (GCase) which cleaves the glycosidic bond in glucocereboside and –

sphingosine (Figure 1.13). However, the precise contribution of GCase to the pathology of

PD remains elusive.

1.3.3.3 Etiology of Gaucher’s disease

Gaucher’s disease (GD) is a lysosomal storage disorder (LSD) associated with the lack of

GCase activity, which in turn leads to the accumulation to glycolipids. The activity reduction

can be the result of missense mutations in the GBA1 gene of which approximately 200

variants are known.133 GD severity differs depending on the degree of functional depletion

caused by the respective mutation. Clinically, three types of GD have been established: GD


39

type 1 is the most common from without CNS involvement. Symptoms range from

asymptomatic to early-onset of splenomegaly and hepatomegaly in childhood. Type 2 GD

describes the acute neuronopathic form which is lethal in childhood due to severe symptoms

including neurological impairment. GD with neurological involvement but late-onset is

described as type 3. Symptoms like seizures and dementia manifest in adolescence and

patient’s life expectancy lies between 30 and 40 years.133

Figure 1.13: Reaction scheme depicting β-glucocerebrosidase (GCase) which cleaves the

glycosidic bond in glucoceremide and –sphingosine.

Recent reports suggest a reduced GCase activity results in a predisposition to PD.134,135

Heterozygous carriers of loss-of-function GBA1 mutations are known to be younger when

diagnosed with PD and show increased severity of symptoms.136 Additionally, GCase activity

in heterozygous carriers is diminished to about 50-70% compared to non-carriers.134

Interestingly, early stage nonGBA1 PD patients are reported to show a significant reduction

in GCase activity and lysosomal capacity.137 Furthermore, an increase in GlcSph levels has

OHOHO

OHO

OH

HN

OHOHO

OH OH

OH

HN

HOGCase

(β-glucosyl)sphingosine, R = H(β-glucosyl)ceramide, R = COCH2(CH2)15CH3

OHR OHR


40

been shown in nonGBA1 PD patients and even in healthy individuals above the age of

sixty.127

Until enzyme replacement therapies (ERT) became available in 1991, care of GD was

restricted to symptomatic treatment. Today, treatment options are diverse: ERT offers relief

through the application of recombinant GCase resulting in decreased GlcCer/GlcSph levels.

Substrate reduction therapy (SRT) aims at inhibiting the ceramide glucosyltransferase and

thereby reducing the rate of GlcCer anabolism. With the rise of ERT, SRT application was

limited to patients with mild conditions. In an attempt to rescue residual activity of

improperly folded GCase in vivo, small molecule chaperons have been considered. The cough

medicine ‘ambroxol’ showed significant recovery rates in a clinical study with GD patients.138

With the multitude of GCase variants occurring, the identification of suitable compounds is

challenging but may provide a valuable addition to the therapeutic spectrum.139

Lysosomal storage disorders pose a serious threat to post-mitotic neurons as the

lysosomal function is reduced. The particularly extensive metabolic activity and high

membrane turnover in synapses explains their vulnerability. As a consequence of lysosomal

deficiency and resulting defects in axonal transport, synapses fail to transmit neuronal stimuli

resulting in “physiological death”, rather than cell death.140 Despite type 1 GD’s classification

as non-neuronopathic, this form is often associated with PD.134

Mazzulli et al. discovered a pathogenic positive feedback loop in which reduced GCase

activity causes lysosomal stress resulting in a decreased α-syn turnover. The resulting

oligomers and fibrils put even more stress on the lysosomal capacity leading to impaired

autophagy which results in an increase in cytosolic α-syn and eventually α-syn oligomers. The

presence of these pathogenic oligomers inhibits endoplasmic trafficking of the

LIMP2/GCase complex to the Golgi apparatus and ultimately to the lysosome.141,142


41

A recent report showed a direct molecular interaction between GlcSph and α-syn

aggregation, suggesting the lyso-form of the accumulating glycolipid as the toxic

compound.143 This study demonstrates the importance of glycoconjugates and suitable glyco-

analytical tools to discover valuable insights.

Figure 1.14: Schematic summary of contributing factors in the pathobiochemistry of

Parkinson’s disease (PD). The death of dopaminergic neurons leads to reduced

neurotransmission resulting in the PD. Aging influences lysosomal function and GCase

activity negatively, while ROS are increased. Reduced GCase activity caused by Gaucher’s

disease leads to increased glycolipid levels which positively influence α-syn polymerisation.

Green arrows: increase, purple arrows: decrease, black arrow: transition.

To summarise (Figure 1.14) the pathobiochemical landscape surrounding the etiology of

PD, age is the predominant factor. Aging neurons struggle to maintain α-syn homeostasis

and lysosomal function and are more vulnerable to oxidative stress. Additionally, the

pathology of GD plays an important role, especially in conjunction with age. Molecular

interactions between GCase and α-syn as well as between glycolipids and α-syn have been

shown to contribute to neuronal dysfunction.


42

1.4 References

1. Varki, A., Cummings, R. D. & Esko, J. D. Essentials of Glycobiology, 3rd edition. (Cold

Spring Harbor Laboratory Press, 2017).

2. Yu, H. & Chen, X. Carbohydrate post-glycosylational modifications. Org. Biomol.

Chem. 5, 865–872 (2007).

3. Spiro, R. G. Protein glycosylation: nature, distribution, enzymatic formation, and

disease implications of glycopeptide bonds. Glycobiology 12, 43R–56R (2002).

4. Dell, A., Galadari, A., Sastre, F. & Hitchen, P. Similarities and differences in the

glycosylation mechanisms in prokaryotes and eukaryotes. International Journal of

Microbiology 2010, 1–14 (2010).

5. Apweiler, R., Hermjakob, H. & Sharon, N. On the frequency of protein glycosylation,

as deduced from analysis of the SWISS-PROT database. Biochim. Biophys. Acta - Gen.

Subj. 1473, 4–8 (1999).

6. Maverakis, E. et al. Glycans in the immune system and The Altered Glycan Theory of

Autoimmunity: A critical review. J. Autoimmun. 57, 1–13 (2015).

7. Sperandio, M., Gleissner, C. A. & Ley, K. Glycosylation in immune cell trafficking.

Immunological Reviews 230, 97–113 (2009).

8. Morris, H. R. et al. Gender-specific glycosylation of human glycodelin affects its

contraceptive activity. J. Biol. Chem. 271, 32159–32167 (1996).

9. Schröter, S., Osterhoff, C., McArdle, W. & Ivell, R. The glycocalyx of the sperm

surface. in Human Reproduction Update 5, 302–313 (1999).

10. Fukuda, M. N. & Akama, T. O. In vivo role of α-mannosidase IIx: Ineffective

spermatogenesis resulting from targeted disruption of the Man2a2 in the mouse.

Biochimica et Biophysica Acta - General Subjects 1573, 382–387 (2002).

11. Paszek, M. J. et al. The cancer glycocalyx mechanically primes integrin-mediated

growth and survival. Nature 511, 319–325 (2014).

12. Bones, J., Mittermayr, S., O’Donoghue, N., Guttman, A. & Rudd, P. M. Ultra

Performance Liquid Chromatographic Profiling of Serum N -Glycans for Fast and

Efficient Identification of Cancer Associated Alterations in Glycosylation. Anal. Chem.

82, 10208–10215 (2010).

13. Chen, Y.-T. et al. Identification of novel tumor markers for oral squamous cell

carcinoma using glycoproteomic analysis. Clin. Chim. Acta 420, 45–53 (2013).


43

14. Lauc, G. et al. Loci Associated with N-Glycosylation of Human Immunoglobulin G

Show Pleiotropy with Autoimmune Diseases and Haematological Cancers. PLoS

Genet. 9, (2013).

15. Bull, C. et al. Targeting Aberrant Sialylation in Cancer Cells Using a Fluorinated Sialic

Acid Analog Impairs Adhesion, Migration, and In Vivo Tumor Growth. Mol. Cancer

Ther. 12, 1935–1946 (2013).

16. Wurm, D., Löffler, G., Lindinger, A. & Gortner, L. Congenital disorders of

glycosylation type Ia as a cause of mirror syndrome. J. Perinatol. 27, 802–804 (2007).

17. Yamashita, T., Wada, R. & Proia, R. L. Early developmental expression of the gene

encoding glucosylceramide synthase, the enzyme controlling the first committed step

of glycosphingolipid synthesis. Biochim. Biophys. Acta - Gen. Subj. 1573, 236–240 (2002).

18. Yu, R. K., Tsai, Y. T. & Ariga, T. Functional roles of gangliosides in

Neurodevelopment: An overview of recent advances. Neurochemical Research 37, 1230–

1244 (2012).

19. Schengrund, C. L. ‘Multivalent’ saccharides: Development of new approaches for

inhibiting the effects of glycosphingolipid-binding pathogens. Biochemical Pharmacology

65, 699–707 (2003).

20. Schnaar, R. L. Brain gangliosides in axon-myelin stability and axon regeneration.

FEBS Letters 584, 1741–1747 (2010).

21. Yoon, S.-J., Nakayama, K., Hikita, T., Handa, K. & Hakomori, S. Epidermal growth

factor receptor tyrosine kinase is modulated by GM3 interaction with N-linked

GlcNAc termini of the receptor. Proc. Natl. Acad. Sci. U. S. A. 103, 18987–18991

(2006).

22. Inokuchi, J. I. GM3 and diabetes. Glycoconjugate Journal 31, 193–197 (2014).

23. Lombard, V., Golaconda Ramulu, H., Drula, E., Coutinho, P. M. & Henrissat, B. The

carbohydrate-active enzymes database (CAZy) in 2013. Nucleic Acids Res. 42, (2014).

24. Sinnott, M. L. Catalytic Mechanisms of Enzymic Glycosyl Transfer. Chem. Rev. 90,

1171–1202 (1990).

25. Zhang, R., Yip, V. L. Y., Withers, S. G. & Columbia, B. Mechanisms of enzymatic

glycosyl transfer. Compr. Nat. Prod. II Chem. Biol. 385–422 (2010).

26. Leloir, L. F. & Cardini, C. E. Biosynthesis of glycogen from uridine diphosphate

glucose. J. Am. Chem. Soc. 79, 6340–6341 (1957).

27. Ghose, T. K. Measurement of cellulase activities. Pure Appl. Chem. 59, (1987).


44

28. Johnson, L. N. The early history of lysozyme. Nature Structural Biology 5, 942–944

(1998).

29. Chundawat, S. P. S., Beckham, G. T., Himmel, M. E. & Dale, B. E. Deconstruction

of Lignocellulosic Biomass to Fuels and Chemicals. Annu. Rev. Chem. Biomol. Eng. 2,

121–45 (2011).

30. Carvalho, D. M. & Colodette, J. L. Comparative study of acid hydrolysis of lignin and

polysaccharides in biomasses. BioResources 12, 6907–6923 (2017).

31. Barb, A. W. & Prestegard, J. H. NMR analysis demonstrates immunoglobulin G N-

glycans are accessible and dynamic. Nat. Chem. Biol. 7, 147–153 (2011).

32. Lundborg, M. & Widmalm, G. Structural analysis of glycans by NMR chemical shift

prediction. Anal. Chem. 83, 1514–1517 (2011).

33. Vliegenthart, J. F. G., Dorland, L. & van Halbeek, H. High-resolution, 1H-Nuclear

magnetic resonance spectroscopy as a tool in the structural analysis of carbohydrates

related to glycoproteins. Adv. Carbohydr. Chem. Biochem. 41, 209–374 (1983).

34. Jansson, P.-E., Kenne, L. & Widmalm, G. Casper—a computerised approach to

structure determination of polysaccharides using information from n.m.r.

spectroscopy and simple chemical analyses. Carbohydr. Res. 168, 67–77 (1987).

35. van Halbeek, H. in (eds. Grant, D. M. & Harris, R. K.) 1107–1137 (John Wiley &

Sons Ltd, 1996).

36. van Kuik, J. A., Hard, K. & Vliegenthart, J. F. A 1H NMR database computer program

for the analysis of the primary structure of complex carbohydrates. Carbohydr. Res. 235,

53–68 (1992).

37. Doubet, S., Bock, K., Smith, D., Darvill, A. & Albersheim, P. The complex

carbohydrate structure database. Trends Biochem. Sci. 14, 475–477 (1989).

38. Curatolo, W., Neuringer, L. J., Ruben, D. & Haberkorn, R. Two-dimensional J-

resolved 1H-nuclear magnetic resonance spectroscopy of α,β-d-glucose at 500 MHz.

Carbohydr. Res. 112, 297–300 (1983).

39. Cancilla, M. T., Penn, S. G., Carroll, J. A. & Lebrilla, C. B. Coordination of Alkali

Metals to Oligosaccharides Dictates Fragmentation Behavior in Matrix Assisted Laser

Desorption Ionization / Fourier Transform Mass Spectrometry. J. Am. Chem. Soc. 118,

6736–6745 (1996).


45

40. Hunnam, V. et al. Ionization and fragmentation of neutral and acidic

glycosphingolipids with a Q-TOF mass spectrometer fitted with a MALDI ion source.

J. Am. Soc. Mass Spectrom. 12, 1220–1225 (2001).

41. Fenn, J. B., Mann, M., Meng, C. K., Wong, S. F. & Whitehouse, C. M. Electrospray

ionization for mass spectrometry of large biomolecules. Science (80-. ). 246, 64–71

(1989).

42. Yamashita, M. & Fenn, J. B. Electrospray ion source. Another variation on the free-

jet theme. J. Phys. Chem. 88, 4451–4459 (1984).

43. Watson, J. T. & Sparkman, O. D. Introduction to Mass Spectrometry. Introduction to Mass

Spectrometry: Instrumentation, Applications and Strategies for Data Interpretation: Fourth Edition

(John Wiley & Sons, Ltd, 2007).

44. Kebarle, P. & Verkcerk, U. H. Electrospray: From Ions in solution to Ions in the gas

phase, what we know now. Mass Spectrom. Rev. 28, 898–917 (2009).

45. Konermann, L., Ahadi, E., Rodriguez, A. D. & Vahidi, S. Unraveling the mechanism

of electrospray ionization. Anal. Chem. 85, 2–9 (2013).

46. Iribarne, J. V. On the evaporation of small ions from charged droplets. J. Chem. Phys.

64, 2287 (1976).

47. Fernandez De La Mora, J. Electrospray ionization of large multiply charged species

proceeds via Dole’s charged residue mechanism. Anal. Chim. Acta 406, 93–104 (2000).

48. Karas, M., Glückmann, M. & Schäfer, J. Ionization in matrix-assisted laser

desorption/ionization: Singly charged molecular ions are the lucky survivors. J. Mass

Spectrom. 35, 1–12 (2000).

49. El-Aneed, A., Cohen, A. & Banoub, J. Mass spectrometry, review of the basics:

Electrospray, MALDI, and commonly used mass analyzers. Applied Spectroscopy Reviews

44, 210–230 (2009).

50. Knochenmuss, R. Ion formation mechanisms in UV-MALDI. Analyst 131, 966 (2006).

51. Miller, P. E. & Denton, M. B. The quadrupole mass filter: Basic operating concepts.

J. Chem. Educ. 63, 617 (1986).

52. Wiley, W. C. & McLaren, I. H. Time-of-flight mass spectrometer with improved

resolution. Rev. Sci. Instrum. 26, 1150–1157 (1955).

53. Mason, E. A. & Schamp, H. W. Mobility of Gaseous Ions in Weak Electric Fields.

Ann. Phys. 4, 233–270 (1958).


46

54. Giles, K. et al. Applications of a travelling wave-based radio-frequency-only stacked

ring ion guide. Rapid Commun. Mass Spectrom. 18, 2401–2414 (2004).

55. Shvartsburg, A. A. & Smith, R. D. Fundamentals of traveling wave ion mobility

spectrometry. Anal. Chem. 80, 9689–9699 (2008).

56. Both, P. et al. Discrimination of epimeric glycans and glycopeptides using IM-MS and

its potential for carbohydrate sequencing. Nat. Chem. 6, 65–74 (2013).

57. Hofmann, J., Hahm, H. S., Seeberger, P. H. & Pagel, K. Identification of carbohydrate

anomers using ion mobility–mass spectrometry. Nature (2015).

58. Blixt, O. et al. Printed covalent glycan array for ligand profiling of diverse glycan

binding proteins. Proc. Natl. Acad. Sci. U. S. A. 101, 17033–17038 (2004).

59. Oyelaran, O. & Gildersleeve, J. C. Glycan arrays: recent advances and future

challenges. Current Opinion in Chemical Biology 13, 406–413 (2009).

60. Lundquist, J. J. & Toone, E. J. The cluster glycoside effect. Chem. Rev. 102, 555–578

(2002).

61. Tang, P. W., Gool, H. C., Hardy, M., Lee, Y. C. & Felzi, T. Novel approach to the

study of the antigenicities and receptor functions of carbohydrate chains of

glycoproteins. Biochem. Biophys. Res. Commun. 132, 474–480 (1985).

62. Roy, R. Syntheses and some applications of chemically defined multivalent

glycoconjugates. Current Opinion in Structural Biology 6, 692–702 (1996).

63. Gray, C. J. et al. Label-Free Discovery Array Platform for the Characterization of

Glycan Binding Proteins and Glycoproteins. Anal. Chem. 89, 4444–4451 (2017).

64. Love, J. C., Estroff, L. A., Kriebel, J. K., Nuzzo, R. G. & Whitesides, G. M. Self-

assembled monolayers of thiolates on metals as a form of nanotechnology. Chemical

Reviews 105, 1103–1169 (2005).

65. Kuno, A. et al. Evanescent-field fluorescence-assisted lectin microarray: A new

strategy for glycan profiling. Nat. Methods 2, 851–856 (2005).

66. Pilobello, K. T., Krishnamoorthy, L., Slawek, D. & Mahal, L. K. Development of a

lectin microarray for the rapid analysis of protein glycopatterns. ChemBioChem 6, 985–

989 (2005).

67. Qian, X. et al. Arrays of self-assembled monolayers for studying inhibition of bacterial

adhesion. Anal. Chem. 74, 1805–1810 (2002).

68. Bundy, J. & Fenselau, C. Lectin-based affinity capture for MALDI-MS analysis of

bacteria. Anal. Chem. 71, 1460–1463 (1999).


47

69. Sanchez-Ruiz, A., Serna, S., Ruiz, N., Martin-Lomas, M. & Reichardt, N. C. MALDI-

TOF mass spectrometric analysis of enzyme activity and lectin trapping on an array

of N-glycans. Angew. Chemie - Int. Ed. 50, 1801–1804 (2011).

70. Serna, S., Yan, S., Martin-Lomas, M., Wilson, I. B. H. & Reichardt, N. C.

Fucosyltransferases as synthetic tools: Glycan array based substrate selection and core

fucosylation of synthetic N-glycans. J. Am. Chem. Soc. 133, 16495–16502 (2011).

71. Blixt, O. et al. Glycan microarrays for screening sialyltransferase specificities. Glycoconj.

J. 25, 59–68 (2008).

72. Šardzík, R. et al. Chemoenzymatic synthesis of O-mannosylpeptides in solution and

on solid phase. J. Am. Chem. Soc. 134, 4521–4 (2012).

73. Beloqui, A., Sanchez-Ruiz, A., Martin-Lomas, M. & Reichardt, N.-C. A surface-based

mass spectrometry method for screening glycosidase specificity in environmental

samples. Chem. Commun. 48, 1701–1703 (2012).

74. Rudd, P. M. et al. Oligosaccharide sequencing technology. Nature 388, 205–207 (1997).

75. Lareau, N. M., May, J. C. & McLean, J. A. Non-derivatized glycan analysis by reverse

phase liquid chromatography and ion mobility-mass spectrometry. Analyst 140, 3335–

3338 (2015).

76. Vanderschaeghe, D., Festjens, N., Delanghe, J. & Callewaert, N. Glycome profiling

using modern glycomics technology: Technical aspects and applications. Biological

Chemistry 391, 149–161 (2010).

77. Campbell, M. P., Royle, L., Radcliffe, C. M., Dwek, R. A. & Rudd, P. M. GlycoBase

and autoGU: Tools for HPLC-based glycan analysis. Bioinformatics 24, 1214–1216

(2008).

78. Royle, L. et al. HPLC-based analysis of serum N-glycans on a 96-well plate platform

with dedicated database software. Anal. Biochem. 376, 1–12 (2008).

79. Bigge, J. C. et al. Nonselective and Efficient Fluorescent Labeling of Glycans Using 2-

Amino Benzamide and Anthranilic Acid. Anal. Biochem. 230, 229–238 (1995).

80. Hase, S. Pyridylamination as a means of analyzing complex sugar chains. Proc. Jpn.

Acad. Ser. B. Phys. Biol. Sci. 86, 378–390 (2010).

81. Buszewski, B. & Noga, S. Hydrophilic interaction liquid chromatography (HILIC)-a

powerful separation technique. Anal Bioanal Chem 402, 231–247 (2012).


48

82. Hägglund, P., Bunkenborg, J., Elortza, F., Jensen, O. N. & Roepstorff, P. A new

strategy for identification of N-glycosylated proteins and unambiguous assignment of

their glycosylation sites using HILIC enrichment and partial deglycosylation. J.

Proteome Res. 3, 556–566 (2004).

83. Wuhrer, M., De Boer, A. R. & Deelder, A. M. Structural glycomics using Hydrophilic

interaction chromatography (HILIC) with mass spectrometry. Mass Spectrom. Rev. 28,

192–206 (2009).

84. Guile, G. R., Rudd, P. M., Wing, D. R., Prime, S. B. & Dwek, R. A. A Rapid High-

Resolution High-Performance Liquid Chromatographic Method for Separating

Glycan Mixtures and Analyzing Oligosaccharide Profiles. Anal. Biochem. 240, 210–226

(1996).

85. Jain, A. N. Scoring functions for protein-ligand docking. Curr. Protein Pept. Sci. 7, 407–

20 (2006).

86. Meng, X.-Y., Zhang, H.-X., Mezei, M. & Cui, M. Molecular Docking: A Powerful

Approach for Structure-Based Drug Discovery. Curr. Comput. Aided-Drug Des. 7, 146–

157 (2011).

87. Nadas, J., Li, C. & Wang, P. G. Computational structure activity relationship studies

on the CD1d/Glycolipid/TCR complex using AMBER and AUTODOCK. J. Chem.

Inf. Model. 49, 410–423 (2009).

88. Nyholm, P. G. et al. Two distinct binding sites for globotriaosyl ceramide on

verotoxins: Identification by molecular modelling and confirmation using deoxy

analogues and a new glycolipid receptor for all verotoxins. Chem. Biol. 3, 263–275

(1996).

89. Liese, A., Seelbach, K. & Wandrey, C. Industrial Biotransformations, Second Edition.

Industrial Biotransformations, Second Edition (2006).

90. Turner, N. J. Directed evolution drives the next generation of biocatalysts. Nature

Chemical Biology 5, 567–573 (2009).

91. Bornscheuer, U. T. et al. Engineering the third wave of biocatalysis. Nature 485, 185–

194 (2012).

92. Sheldon, R. A. in Green Biocatalysis (ed. Patel, R. N.) 1–15 (John Wiley & Sons, Inc,

2016).

93. Walsh, G. Biopharmaceutical benchmarks 2014. Nat. Biotechnol. 32, 992–1000 (2014).


49

94. Hossler, P., Khattak, S. F. & Li, Z. J. Optimal and consistent protein glycosylation in

mammalian cell culture. Glycobiology 19, 936–949 (2009).

95. Higgins, E. Carbohydrate analysis throughout the development of a protein

therapeutic. Glycoconjugate Journal 27, 211–225 (2010).

96. Hossler, P., Mulukutla, B. C. & Hu, W. S. Systems analysis of N-glycan processing in

mammalian cells. PLoS One 2, (2007).

97. Baik, J. Y. et al. Metabolic engineering of Chinese hamster ovary cells: Towards a

bioengineered heparin. Metab. Eng. 14, 81–90 (2012).

98. Del Val, I. J., Polizzi, K. M. & Kontoravdi, C. A theoretical estimate for nucleotide

sugar demand towards Chinese Hamster Ovary cellular glycosylation. Sci. Rep. 6,

(2016).

99. Thomann, M. et al. In vitro glycoengineering of IgG1 and its effect on Fc receptor

binding and ADCC activity. PLoS One 10, (2015).

100. Sinclair, A. M. Erythropoiesis stimulating agents: Approaches to modulate activity.

Biologics: Targets and Therapy 7, 161–174 (2013).

101. Yang, W. H. et al. An intrinsic mechanism of secreted protein aging and turnover. Proc.

Natl. Acad. Sci. 112, 13657–13662 (2015).

102. Solá, R. J. & Griebenow, K. Glycosylation of Therapeutic Proteins: An Effective

Strategy to Optimiza Efficacy. BioDrugs 24, 9–21 (2011).

103. Chung, C. H. et al. Cetuximab-Induced Anaphylaxis and IgE Specific for Galactose-

α-1,3-Galactose. N. Engl. J. Med. 358, 1109–1117 (2008).

104. Mimura, Y. et al. Glycosylation engineering of therapeutic IgG antibodies: challenges

for the safety, functionality and efficacy. Protein Cell 1–16 (2017).

105. Takeuchi, M. et al. Comparative study of the asparagine-linked sugar chains of human

erythropoietins purified from urine and the culture medium of recombinant Chinese

hamster ovary cells. J. Biol. Chem. 263, 3657–3663 (1988).

106. Zhu, A. & Hurst, R. Anti-N-glycolylneuraminic acid antibodies identified in healthy

human serum. Xenotransplantation 9, 376–381 (2002).

107. Tangvoranuntakul, P. et al. Human uptake and incorporation of an immunogenic

nonhuman dietary sialic acid. Proc. Natl. Acad. Sci. 100, 12045–12050 (2003).

108. de Lau, L. M. & Breteler, M. M. Epidemiology of Parkinson’s disease. Lancet Neurol.

5, 525–535 (2006).


50

109. Tysnes, O. B. & Storstein, A. Epidemiology of Parkinson’s disease. Journal of Neural

Transmission 124, 901–905 (2017).

110. Parkinson, J. Neuropsychiatry Classics: An essay on the Shaking Palsy. J.

Neuropsychiatry Clin. Neurosci. 14, 223–236 (2002).

111. Chaudhuri, K. R., Healy, D. G. & Schapira, A. H. Non-motor symptoms of

Parkinson’s disease: diagnosis and management. Lancet Neurol. 5, 235–245 (2006).

112. Jankovic, J. Parkinson’s disease: clinical features and diagnosis. J. Neurol. Neurosurg.

Psychiatry 79, 368–76 (2008).

113. Wang, J., Hoekstra, J. G., Zuo, C., Cook, T. J. & Zhang, J. Biomarkers of Parkinson’s

disease: current status and future perspectives. Drug Discov. Today 18, 155–62 (2013).

114. Hunter, K. R., Shaw, K. M., Laurence, D. R. & Stern, G. M. Sustained levodopa

therapy in parkinsonism. Lancet (London, England) 2, 929–31 (1973).

115. Schapira, A. H. V. Progress in neuroprotection in Parkinson’s disease. European Journal

of Neurology 15, 5–13 (2008).

116. Uversky, V. N. Neuropathology, biochemistry, and biophysics of a-synuclein

aggregation. J. Neurochem. 103, 17–37 (2007).

117. Schulz-Schaeffer, W. J. The synaptic pathology of alpha-synuclein aggregation in

dementia with Lewy bodies, Parkinson’s disease and Parkinson’s disease dementia.

Acta Neuropathol. 120, 131–143 (2010).

118. Spina, M. B. & Cohen, G. Dopamine turnover and glutathione oxidation: implications

for Parkinson disease. Proc. Natl. Acad. Sci. U. S. A. 86, 1398–400 (1989).

119. Asanuma, M., Miyazaki, I. & Ogawa, N. Dopamine- or L-DOPA-induced

neurotoxicity: The role of dopamine quinone formation and tyrosinase in a model of

Parkinson’s disease. Neurotox. Res. 5, 165–176 (2003).

120. Zhang, J. et al. Parkinson’s Disease Is Associated with Oxidative Damage to

Cytoplasmic DNA and RNA in Substantia Nigra Neurons. Am. J. Pathol. 154, 1423–

1429 (1999).

121. Napolitano, A., Manini, P. & D’Ischia, M. Oxidation Chemistry of Catecholamines

and Neuronal Degeneration: An Update. Curr. Med. Chem. 18, 1832–1845 (2011).

122. Halliwell, B. Oxidative stress and neurodegeneration: Where are we now? Journal of

Neurochemistry 97, 1634–1658 (2006).

123. Hashimoto, M. et al. Oxidative stress induces amyloid-like aggregate formation of

NACP/α-synuclein in vitro. Neuroreport 10, 717–721 (1999).


51

124. Schapira, A. H. et al. Perspectives on recent advances in the understanding and

treatment of Parkinson’s disease. European Journal of Neurology 16, 1090–1099 (2009).

125. Schapira, A. H. V et al. Mitochondria in the etiology and pathogenesis of parkinson’s

disease. Ann. Neurol. 44, S89–S98 (1998).

126. Carmona-Gutierrez, D., Hughes, A. L., Madeo, F. & Ruckenstuhl, C. The crucial

impact of lysosomes in aging and longevity. Ageing Research Reviews 32, 2–12 (2016).

127. Rocha, E. M. et al. Progressive decline of glucocerebrosidase in aging and Parkinson’s

disease. Ann. Clin. Transl. Neurol. 2, 433–438 (2015).

128. Dackis, C. A. & Gold, M. S. New concepts in cocaine addiction: The dopamine

depletion hypothesis. Neurosci. Biobehav. Rev. 9, 469–477 (1985).

129. Lloyd, S. A., Faherty, C. J. & Smeyne, R. J. Adult and in utero exposure to cocaine

alters sensitivity to the parkinsonian toxin 1-methyl-4-phenyl-1,2,3,6-

tetrahydropyridine. Neuroscience 137, 905–913 (2006).

130. Zarranz, J. J. et al. The new mutation, E46K, of alpha-synuclein causes Parkinson and

Lewy body dementia. Ann. Neurol. 55, 164–73 (2004).

131. Appel-Cresswell, S. et al. Alpha-synuclein p.H50Q, a novel pathogenic mutation for

Parkinson’s disease. Mov. Disord. 28, 811–813 (2013).

132. Aharon-Peretz, J., Rosenbaum, H. & Gershoni-Baruch, R. Mutations in the

glucocerebrosidase gene and Parkinson’s disease in Ashkenazi Jews. N. Engl. J. Med.

351, 1972–7 (2004).

133. Jmoudiak, M. & Futerman, A. H. Gaucher disease: pathological mechanisms and

modern management. Br. J. Haematol. 129, 178–88 (2005).

134. Sidransky, E. et al. Multicenter Analysis of Glucocerebrosidase Mutations in

Parkinson’s Disease. N. Engl. J. Med. 361, 1651–1661 (2009).

135. Bras, J. et al. Complete screening for glucocerebrosidase mutations in Parkinson

disease patients from Portugal. Neurobiol. Aging 30, 1515–1517 (2009).

136. Neumann, J. et al. Glucocerebrosidase mutations in clinical and pathologically proven

Parkinson’s disease. Brain 132, 1783–1794 (2009).

137. Murphy, K. E. et al. Reduced glucocerebrosidase is associated with increased -

synuclein in sporadic Parkinson’s disease. Brain 137, 834–848 (2014).

138. Narita, A. et al. Ambroxol chaperone therapy for neuronopathic Gaucher disease: A

pilot study. Ann. Clin. Transl. Neurol. 3, 200–215 (2016).


52

139. Mistry, P. K. et al. Gaucher disease: Progress and ongoing challenges. Molecular Genetics

and Metabolism 120, 8–21 (2017).

140. Cantuti-Castelvetri, L. & Bongarzone, E. R. Synaptic failure: The achilles tendon of

sphingolipidoses. J. Neurosci. Res. 94, 1031–1036 (2016).

141. Mazzulli, J. R. et al. Gaucher disease glucocerebrosidase and α-synuclein form a

bidirectional pathogenic loop in synucleinopathies. Cell 146, 37–52 (2011).

142. Mazzulli, J. R., Zunke, F., Isacson, O., Studer, L. & Krainc, D. alpha-Synuclein-

induced lysosomal dysfunction occurs through disruptions in protein trafficking in

human midbrain synucleinopathy models. Proc Natl Acad Sci U S A (2016).

143. Taguchi, Y. V. et al. Glucosylsphingosine Promotes α-Synuclein Pathology in Mutant

GBA-Associated Parkinson’s Disease. J. Neurosci. 37, 9617–9631 (2017).

53

Chapter 2 Objectives of this thesis

This thesis aims to provide enhanced analytical tools to study carbohydrates and their

functional biology as well as their application in biotechnology. In particular, three specific

biotechnological objectives are addressed as described in Chapters 3, 4 and 5, while a novel

perspective on down-stream effects of CAZyme dysfunction within neurotoxcicity is

presented in chapter six.

2.1 Simple, quantitative and non-destructive GOase assay

Galactose Oxidase (GOase; EC 1.1.3.9) is an oxygen-dependent single copper enzyme

that oxidises the primary hydroxyl group in terminal carbohydrates. It has applications

ranging from glycan labelling to synthetic carbohydrate chemistry.

Previously, a glucoside-specific variant GOase F2 has been developed through means of

directed evolution by Rannes et al.b However, soluble expression of the enzyme in E. coli is

low-yielding and reaction monitoring requires either an enzymatic cascade reaction to detect

hydrogen peroxide production or time-consuming HPLC analysis which additionally

requires either derivatisation or non-quantitative mass spectrometry.

Chapter 3 describes the development of a simple 1H-NMR screen to provide structural

information on the regiospecific oxidation of mono- and oligo-saccharides. Signal intensity

of characteristic protons can be quantitative and the experimental set-up, unlike mass

spectrometry, allows for material recovery. A route to industrially relevant oxidation

protocols is presented utilising model compounds.

b Rannes, J. B.; Ioannou, A.; Willies, S. C.; Grogan, G.; Behrens, C.; Flitsch, S. L.; Turner, N. J.

J. Am. Chem. Soc. 2011, 133 (22), 8436–8439.

Chapter 2 - Objectives

54

2.2 Carbohydrate arrays for fast and sensitive hydrolase

characterisation

Industrial biotechnology that relies on sustainable feed stocks is limited by the availability

of well-characterised hydrolytic enzymes that degrade plant-based lignocellulose to simpler

carbohydrates. Sequencing of microbial genome sequencing has revealed numerous putative

carbohydrate active enzymes (CAZymes). However, biochemical characterisation of these

enzymes is key to their successful implementation in industrial processes. In order to rapidly

gain an understanding of the substrate scope of candidate enzymes, fast and sensitive high-

throughput methods are required.

The manuscript in Chapter 4 describes the development and application of carbohydrate

arrays coupled with MALDI-ToF MS to profile CAZyme activities from Aspergillus niger

towards specific libraries of oligo-saccharides from commercial and natural sources on mixed

glycan arrays.

2.3 Completing the N-acetylneuraminic acid toolkit

The group of complex glycoconjugates in human biology consists of not only glycolipids

(see Chapter 5) but also peptidoglycans and proteoglycans (e.g. extracellular matrix),

glycosides (e.g. glycogen or DNA) and glycoproteins (N-glycosylation). Structurally, this

group is very diverse with different monomers as well as regio- and stereo-chemical

variations of individual linkages. In order to a) analyse and b) manipulate this challenging

chemistry, dedicated CAZymes with specific activities are required to complement traditional

analytical tools. Unfortunately, the portfolio of available specific glycosidases is incomplete.

In humans the N-glycans, regioisomers α2,6- or α2,3-sialic acid (Neu5Ac) linked to

galactosides are found to convey distinct information. Hence, the sialylation state of glycans

especially in therapy is of great importance.

Chapter 2 - Objectives

55

The manuscript presented in Chapter 5 addresses the gap in the toolbox of Neu5Ac active

enzymes by adding a highly specific α2,6-selective ‘pseudosialidase’ and show-cases its

applicability in glycan-engineering setups.

2.4 Glycolipids in Parkinson’s disease

The pathology of Parkinson’s disease (PD), responsible for the rapid death of

dopaminergic neurons, is not fully understood and a cure remains elusive. Currently, L-Dopa

treatment is the best option for patients but only acts to reduce severity of the symptoms.

Furthermore, an early diagnosis of PD is essential for successful therapy.

While many biomarker studies have been completed, neurologists still rely on symptom-

based diagnosis. Sufferers of Gaucher’s disease (GD) have a predisposition towards early-

onset PD. GD affects the lysosomal enzyme β-glucosylcerebrosidase (GCase) resulting in

the accumulation of glycolipids. The empirical connection between the two pathologies has

been the subject of numerous academic studies mainly focussing on the interaction between

GCase, reduced lysosomal activity and α-synuclein fibril formation.

In Chapter 6 the correlation between glycolipids and dopamine homeostasis in PD is

analysed from a new perspective and a mechanism is proposed to further elucidate the

pathobiochemistry.

56

Chapter 3 Simple, quantitative and non-destructive

Galactose Oxidase assay

3.1 Summary

Galactose oxidase is a potent tool in synthetic carbohydrate chemistry because of its

unique ability to regio-selectively introduce versatile carbonyl motifs into carbohydrates. The

expression of the glucoside-adapted F2 variant has been improved and specific oxidations

of target compounds were monitored by NMR. The industrial applicability was demonstrated

using scale-up oxidations of lactose.

The work presented in this chapter is in progress and prepared for submission to

ChemBioChem. See Chapter 7 for planned future experiments supporting the relevance of

this work.

3.2 Contribution

Drs Peter Both & Susanne Herter performed codon optimisation and sub-cloning, SH

established expression and purification methodology, MR expressed and purified protein for

the experiments presented, MR designed and performed analytical experiments, SH helped

with data interpretation, target molecules were suggested by MR and discussed with BASF

SE, Germany as industrial partner. MR wrote the manuscript.

Chapter 3 - Galactose Oxidase

57

3.3 Introduction

Galactose Oxidase (GOase; EC 1.1.3.9) is an oxygen-dependent single copper enzyme

secreted by various fungi and was discovered in Fusarium sp. formerly known as Polyporus

circinatus and Gibberella zeae.1 It catalyses the oxidation of primary alcohols (RCH2OH),

especially hexoses, to aldehydes converting molecular oxygen to hydrogen peroxide. Its

unusual mechanism caused it to be studied intensively.2

RCH2OH+O2→RCHO+H2O2

The redox chemistry performed by Galactose Oxidase (GOase) is facilitated by a well-

described copper(II)-tyrosyl radical. This radical is based on the post- translational formation

of a tyrosinate ligand consisting of a tyrosine residue (Y272) covalently linked to a nearby

cysteine (C228).3 Tryptophan (W290) stabilises this unusual radical within the active site. The

self-processing formation of this redox cofactor requires only the apoprotein, copper and

oxygen.4,5

The copper ion in the active site (Figure 3.1) is coordinated by four amino acid side chains:

tyrosine (Y495), the tyrosyl radical (Y272/C228) and two histidines (H496 and H581) as well

as water in the inactive state or the substrate during binding respectively.6


58

Figure 3.1: Galactose oxidase active site representation. Cu(II) is coordinated by histidines

H581 and H496 and tyrosine Y495 as well as the tyrosyl radical Y272/C228 which is

stabilised by tryptophane W290. In the ‘inactive’ state (shown) water is bound to Cu(II).

Upon substrate binding H2O is replace by the alcohol substrate (not shown).6

GOase exists in three oxidative states: An oxidised “active” (Cu(II), Tyr•) and a reduced

form (Cu(I), Tyr) which are involved in the catalytic cycle (Figure 3.2). A semi-reduced form

(Cu(II), Tyr) which is catalytically inactive. It can be oxidised to become active again. Upon

substrate (RCH2OH) binding H2O is displaced. The substrate is oxidised by the transfer of

a “proton-coupled electron” to the tyrosyl radical. Following the exit of the product the

“reduced” state of the enzyme holds Cu(I). Oxygen oxidises the active site in a similar

binding mode as the substrate. The “proton-coupled electron” is transferred back and

hydrogen peroxide as well as the “active” GOase are formed.7,8


59

Figure 3.2: Galactose oxidase mechanism. Following substrate binding to the ‘active’ state

a proton-coupled electron is transferred from the substrate to the tyrosyl group. The product

leaves the active site leaving the enzyme “reduced”. Upon O2 entrance and binding the

proton- coupled electron is transferred back to form H2O2, reoxidising the active site.6

The ability of galactose oxidase to oxidise glycosides chemo- and regio-selectively is very

interesting for glyco-chemical purposes. In oligo-saccharides only terminal glycosides are

oxidised by galactose oxidase, leaving all the other residues unaltered. Furthermore, only the

6-position (primary alcohol) is oxidised to the aldehyde (in equilibrium with its conjugated

hydrate).


60

Figure 3.3: Galactose Oxidase (GOase) oxidises terminal glucoside motifs in saccharides

producing hydrogen peroxide. The resulting aldehyde can be found in equilibrium with the

conjugated hydrate.

Rannes et al. demonstrated the versatility of GOase variants ranging from the natural

substrate galactose to glucosides and mannosides and even N-acetylated glucosides.9 The

presented F2 variant contains several mutations and offers by far the broadest substrate

scope amongst galactose oxidase variants. This variant was developed from previous

generations by means of directed evolution and screening against D-glucose. Interestingly, in

the F2 variant two neighbouring amino acids (Q406E and Y405F) are altered compared to

M3 causing an increase in activity towards D-glucose, possibly due to the close proximity of

these residues to the C-4 and -2 positions, since D-glucose (and D-mannose) are C-4 and -2

epimers of D-galactose.9 Focussing on glucosides and derived oligo- and poly-saccharides,

GOase F2 offers a good biocatalytic route to the desired oxidised products.

3.4 Target glucosides

The chemo- and regio-selective modification especially in glycans is extremely challenging

due to the abundance of similar functional groups. Targeting a specific subtype without the

use of protection group chemistry is very demanding. Only few examples for regio-selective

chemistry in unprotected saccharides are known for example the TEMPO-mediated

oxidation of the primary hydroxyl group in the 6-position.10 The nature of enzymatic

biotransformations allows for an intrinsic selectivity which can simplify the process in many

ways. Galactose Oxidase for example introduces a versatile carbonyl motif at the 6-position

of terminal glycosides without the need of protection groups.

O

OH R

OH

HOHO

O

OH R

OH

HOHO

1/2 O2 H2O2

1/2 O2

GOase HOO

OH RHO

HO

O

H2O


61

From an industrial perspective, glucosides are much more interesting than galactosides

because of their natural abundance. Glucose and especially glucosides make up vast

percentages of biomass in forms of starch and cellulose. These glucose polymers mainly

consist of α1,4- (in amylose) and β1,4-linked (in cellulose) glucose units. Whereas cellulose’s

natural properties provide plants with structure, starch (amylose and α1,6-branched

amylopectin) is their energy reservoir in form of polymerised glucose to relieve osmotic

pressure.

The physicochemical properties of cellulose (especially the low solubility in water) make

it a very demanding compound to work with. Besides, cellulose is found in plant cell walls

along with lignin, which would require further separation. However, amylose is much more

accessible because of its increased water solubility and therefore an anticipated substrate to

subject to enzymatic oxidation.

BASF as an industrial partner is especially interested in the chemical functionalisation of

starch (i.e. amylose). The targeted oxidation of hydroxyl groups to carbonyl groups would

allow for a) the modification of physical properties (e.g. viscosity) and b) further

functionalisation including cross linking linear amylose fibrils with diamines for example.

To study the effects of alpha- and beta-linkages in glucosidic polymers and the influence

of oligo-/polymer length during the oxidation mediated by galactose oxidase a panel of

substrates (Figure 3.4) containing D-galactose (1) and D-glucose (4) as well as their 1-O-

methylated compounds (2, 3, 5 and 6) has been identified. Furthermore, glucosidic oligo-

saccharides in alpha (maltose 10, maltotriose 11) and beta-linkage (cellobiose 12, cellotriose

13) were included. Additionally, raffinose (7), lactose (8) and sucrose (9) round up the panel

with lactose being of special industrial interest as well.

Lactose accumulates in significant amounts in various industrial dairy processes and is

cheaply available. Consequently, it was identified as an interesting target by BASF. The


62

terminal galactoside motif is a very promising target for galactose oxidase and would result

in the formation of an easily obtainable highly water soluble dialdehyde from sustainable

sources. In this study, the oxidation of the above-mentioned substrates by different galactose

oxidase variants was investigated. Insights into activities and mechanism are very important

to define and understand the substrate scope of the enzyme.

Compound R1 R2 R3 R4

D-galactose (1) H OH H/OH H/OH

α-D-methylgalactoside (2) H OH OMe H

β-D-methylgalactoside (3) H OH H OMe

D-glucose (4) OH H H/OH H/OH

α-D-methylglucoside (5) OH H OMe H

β-D-methylglucoside (6) OH H H OMe

raffinose (7) H OH (1,4)Glc-β(1,2)Fru H

lactose (8) H OH (1,4)Glc H

sucrose (9) OH H (1,2)Fru H

maltose (10) OH H (1,4)Glc H

maltotriose (11) OH H [α(1,4Glc)]2 H

cellobiose (12) OH H H (1,4)Glc

cellotriose (13) OH H H [β(1,4Glc)]2

Figure 3.4: Substrates of interest (1-13) and corresponding oxidised targets (1a-13a).

OR1HO

R2

R3

R4OH

OH

OR1HO

R2

R3

R4OH

O

1, 2,..., 13 1a, 2a,..., 13a

GOase


63

3.5 Experimental Section

3.5.1 Materials

- Sodium phosphate buffer (NaPi), 100 mM, pH 7.4

- Sodium phosphate buffer, 50 mM, pH 8.0, NaCl 300 mM

- Sodium phosphate buffer, 50 mM, pH 8.0, NaCl 300 mM, desthiobiotin 5 mM

- Horseradish Peroxidase, stock 75 U/mL, final 33.75 U/mL, Sigma Aldrich

- Catalase from bovine liver, stock 330 U/mL, final 165 U/mL, Sigma Aldrich

- ABTS, stock 0.4 mg/mL, stock 0.18 mg/mL, Sigma Aldrich

- D2O, Sigma Aldrich

- all saccharides, Sigma Aldrich

3.5.2 Sub-cloning and expression

The GOase F2 gene was sub-cloned into a pET30a vector. For plasmid production, the

vector is transformed into E. coli XL-1 blue super-competent cells (Agilent Technologies)

using materials and methods described in the supplier’s manual. A single colony of E. coli

XL-1 blue cells containing plasmid GOase F2 is picked from overnight plates and used to

inoculate 5 mL LB medium supplemented with 1 µL of kanamycin per mL (30 mg mL-1

stock solution). The biomass of one 5 mL overnight culture is used for subsequent plasmid

extraction according to materials and methods described for the QUIAGEN plasmid mini

kit.

For enzyme production, the vector is transformed into One Shot E. coli BL21 StarTM

(DE3) cells. A single colony of E. coli BL21 StarTM (DE3) cells containing plasmid GOase

F2 is picked from overnight plates and used to inoculate 5 mL LB medium supplemented

with 1 µL of kanamycin/mL (30 mg mL-1 stock solution). 500 µL of the overnight starter

culture is used to inoculate 250 mL of an auto induction medium (8ZY-4LAC) as described


64

by Deacon & McPherson16 and supplemented with 250 µL of kanamycin (30 mg mL-1 stock

solution) in a 2 L baffled flask. Cells are grown at 26°C and 250 rpm for 60 h.

Protein purification was carried out according to Deacon & McPherson.16 Cells were

disrupted using Triton-X 100 lysis buffer followed by centrifugation of the lysate at 20.000

rpm for 30 min at 4°C and subsequent dialysis of the crude extract into 50 mM NaPi buffer

(300 mM NaCl, pH 8.0) for 12 h at 4°C. Protein purification was accomplished using Strep-

Tag-II columns (5 mL capacity) eluting with a 5 mM desthiobiotin NaPi buffer (300 mM

NaCl, pH 8.0).

Following purification, the protein samples were dialysed into 50 mM sodium phosphate

buffer (pH 7.4) for 24 h at 4°C. In this step copper-sulfate was added to the dialysis buffer

to accomplish the copper-loading of the protein. After copper-loading, the protein samples

were dialysed into copper-free 50 mM sodium phosphate buffer (pH 7.4) over night at 4°C

and subsequently concentrated to desired concentrations.

Under the described conditions the protein does not elute as a sharp peak, but in a rather

broad manner (Figure 3.5). Following purification, the yield of GOase F2 ranges between

50 mg to 60 mg per 250 mL of culture. For the quantification of protein yielded after

expression the BCA Protein Assay KIT by Thermo Fisher Scientific Inc. was used according

to manufacture’s instructions.


65

Figure 3.5: SDS-PAGE of fractions obtained during the protein purification of GOase F2

(approx. 68.5 kDa) using a Strep-Tag-II column.

3.5.3 Biotransformations

3.5.3.1 Glycoside oxidation

The reaction mixture was set up in a 2 mL reaction tube as follows: Stock solutions of

substrates (α-D-methylgalactoside (2), β-D-methylgalactoside (3), α-D-methylglucoside (5), β-

D-methyl-glucoside (6), maltose (10), maltotriose (11), cellobiose (12) and cellotriose (13))

were made up with water. The solution was then vortexed for 10 min to saturate with

atmospheric O2, and 50 µL were added to the reaction. HRP and catalase were added to the

reaction from a 10 mg/mL stock solution. Purified GOase F2 variant was added to the

reaction (see Table 3.1). The reaction was then left for 24 h at 30°C in an orbital shaker at

1000 rpm. Subsequent work up included spinning down of insoluble particles, filtration

through filter tips and freeze-drying over night before dissolving the samples in D2O prior

to NMR experiments.


66

Table 3.1: Components for the biotransformation reactions.

component stock solution volume (µL) final concentration

GOase F2 4.0 mg/mL 50 1.00 mg/mL

HRP 10 mg/mL 21.8 1.09 mg/mL

catalase 10 mg/mL 2.20 0.11 mg/mL

substrate 1040 mM 50.0 260 mM

buffer 100 mM ad 200

3.5.3.2 Lactose oxidation

The reaction mixture was set up in a 2 mL reaction tube as follows: A stock solution of

D-lactose (8) was made up in water. The solution was then vortexed for 10 min prior to

adding to the reaction to saturate with atmospheric O2. Catalase (with and without HRP) was

added to the reaction from a 10 mg/mL stock solution. Dialysed GOase wt (BASF) was

added to the reaction (see Table 3.2). The reaction was then left for 24 h at 30°C in an orbital

shaker at 1000 rpm. Subsequent work up included spinning down of insoluble particles,

filtration and freeze-drying over night before dissolving the samples in D2O prior to NMR

experiments.


67

Table 3.2: Components for the biotransformation reactions.


GOase F2 20.0 mg/mL 25.0 1.00 mg/mL

HRP 10 mg/mL 55.0 1.1 mg/mL


substrate 400 mM 31.25, 62.5, 125 and 250 25, 50, 100 and 200 mM


GOase F2 20.0 mg/mL 50.0 1.00 mg/mL

HRP - - -


substrate 400 mM 31.25, 62.5, 125 and 250 25, 50, 100 and 200 mM


3.5.3.3 Lactose oxidation – scale-up

The reaction mixture was set up in a 500 mL baffled shake flask as follows: A stock

solution of D-lactose (8) was made up in water. The solution was then vortexed for 10 min

prior to adding to the reaction to saturate with atmospheric O2. Catalase was added to the

reaction from a 10 mg/mL stock solution. Dialysed GOase wt (BASF) was added to the

reaction (see Table 3.3). The reaction was then left for 48 h at 30°C in a shaking incubator

at 250 rpm. Subsequent work up included spinning down of insoluble particles, filtration and

freeze-drying over night before dissolving the samples in D2O prior to NMR experiments.

Table 3.3: Components for the biotransformation scale-up reactions.


GOase F2 20 mg/mL 10 and 40 1 and 8 mg/mL

catalase 10 mg/mL 10 0.1 mg/mL

substrate 400 mM 25 and 50 100 and 200 mM

H2O ad 100


68

3.5.4 Analytics

3.5.4.1 Specific activity and kinetics

Specific activities are recorded based on the well-known HRP-ABTS assay, where ABTS

is reduced by HRP and H2O2, which is released during the oxidation of substrate compounds.

This assay is compatible with a 96-well format and is used for specific activities and kinetic

studies. Kinetics are recorded as substrate concentration dependant specific activities.

The oxidation of 1 mol substrate releases 2 mol e- which reduce 2 mol ABTS, therefore

z = 2.

The spectral photometer software gives initial rates ∆E in units per minute. These are

calculated to give specific activities in µmol/min mg with: vtotal = 200 µL, z = 2,

ελ=420 nm = 36 mM-1cm-1, venzyme = 10 µL, cenzyme = 0.1 mg/mL, d = 0.55 cm,

MWenzyme = 68.5kDa.

Δ𝐸 =ΔAΔt

𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐𝑎𝑐𝑡𝑖𝑣𝑖𝑡𝑦 =ΔE ∙ 𝑣><>Q9

𝑧 ∙ 𝜀 ∙ 𝑣81STU8 ∙ 𝑐81STU8 ∙ 𝑑

[𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐𝑎𝑐𝑡𝑖𝑣𝑖𝑡𝑦] =μmol

𝑚𝑖𝑛 ∙ 𝑚𝑔

Composition of the 96-well assay

1. add 10 µL of the enzyme preparation (0.05 mg/mL).

2. add 90 µL of the reaction solution.

3. start the reaction by adding 100 µL of the substrate solution.

4. measure absorption at λ = 420 nm for 45 min.


69

5. determine the slope (∆OD · min-1) of the initial linear region of the plotted graph

(absorption vs. time).

Plotting the absorption against the time and subsequent determination of the initial slopes

gives the specific activities for GOase variants and substrates according to the equation

above.

3.5.4.2 NMR spectroscopy

NMR experiments were carried out for the starting materials (2, 3, 5, 6, 8, 10, 11, 12 and

13) and the crude reaction mixtures of (2a, 3a, 5a, 6a, 8a, 10a, 11a, 12a and 13a). The starting

material (30 mg) was added to an NMR tube and dissolved in D2O. Crude reaction mixtures

were centrifuged, filtered and frozen with liquid nitrogen for 5 min. Following freeze-drying

for >5 h the samples were dissolved in 500 µL D2O and then transferred into a Wilmad

NMR tube 5 mm (Sigma-Aldrich). D2O was added to reach the desired volume. Spectra were

recorded at 400 MHz and 20°C.


70

3.6 Results & Discussion

3.6.1 GOase F2 oxidises various glucosides

3.6.1.1 Specific activities

To test the activity of galactose oxidase (GOase) wild-type enzyme versus the optimised

F2 variant specific activities were obtained using a coupled, doubly-indirect horseradish

peroxidase-ABTS (HRP-ABTS) assay.9 The in situ generated H2O2 is turned over to H2O and

ABTS is oxidised to its colourful intermediate radical (see Figure 3.7). The change in

absorption at λ = 420 nm is measured spectrophotometrically.

To re-evaluate the substrate scope of the GOase F2 variant and to benchmark the enzyme

further experiments with mono- and oligo-saccharides (listed in Figure 3.4) were performed.

As a comparison GOase wild-type enzyme provided by BASF was investigated. The

determination of specific activities is a characterising tool in case kinetic measurements are

not obtainable. Table 3.4 summarises the specific activities for GOase (wt and F2) for a panel

of mono- and oligo-saccharides.


71

Table 3.4: Specific activities of galactose oxidase variants for the tested substrates. No values

were obtained for maltose (10) and cellotriose (13) due to poor fitting parameters or

insufficient quantities of material. Values calculated from triplicates given as mean ± standard

deviation.

Specific activity / µmol/(min mg)

Substrate GOase wild-type GOase F2

D-galactose (1) 1.11 ± 0.01 5.64 ± 0.03

α-D-methylgalactoside (2) 0.89 ± 0.02 5.15 ± 0.04

β-D-methylgalactoside (3) 1.05 ± 0.01 4.74 ± 0.02

D-glucose (4) 0.07 ± 0.01 2.07 ± 0.04

α-D-methylglucoside (5) > 0.01 0.78 ± 0.01

β-D-methylglucoside (6) > 0.01 1.61 ± 0.05

raffinose (7) 0.91 ± 0.01 3.32 ± 0.03

lactose (8) 0.11 ± 0.01 2.20 ± 0.04

sucrose (9) > 0.01 0.02 ± 0.01

maltose (10) n.d. n.d.

maltotriose (11) > 0.01 0.14 ± 0.01

cellobiose (12) > 0.01 0.20 ± 0.01

cellotriose (13) n.d. n.d.

Apart from the overall increased activity of GOase F2 over GOase wt (for D-galactose

(1)) the increase in specific activity of GOase F2 for D-glucose (3) is 6 times higher than

GOase wt. This trend towards increased activity is not limited to glucosides only. Raffinose

and lactose (gal-terminated) are better substrates for the F2 variant as well.

Figure 3.6 presents the specific activities for each substrate normalised to the activity

against D-galactose (1). This relative comparison of activities shows that the activity of GOase

F2 for glucosides (4, 5, 6) is significantly improved over the wild-type, which is hardly active.

However, the β-glucoside (6) is favoured over the alpha-anomer (5).


72

Beyond the mono-saccharides the specific activity for lactose (8) is improved. Together

with the industrial partner BASF, the dialdehyde 6-oxo-lactose (8b) was identified as an

interesting target molecule which is accessible through this enzymatic route.

Looking at the overall low activities of GOase wt it is very likely that stability issues

influenced the outcome in these assays, especially since BASF determined higher values

before shipping the enzyme preparations. Those preparations were copper-loaded via dialysis

before use. However, the rest-activity proved to be useful for applications in

biotransformation reactions.

Figure 3.6: Relative comparison of specific activities (Table 3.4) normalised to D-galactose

(set to 100%). GOase wt (black bars) possess a very narrow substrate spectrum excluding

glucose-terminated saccharides. GOase F2 (grey bars) shows increased activity with

glucosides as well as lactose.


73

3.6.1.2 Kinetics

To fully understand the mechanistic details behind the change in substrate specificity

kinetic parameters give information about binding and turnover of substrate molecules.

Kinetic parameters for galactose oxidase F2 were obtained using the known HRP-ABTS

assay in a time- and substrate concentration-dependent manner. Table 3.5 summarises the

values for KM as well as for kcat. GOase F2s KM for D-galactose (1) is determined to be

(180 ± 20) mM which is comparable with a literature value for GOase wt, whereas kcat is

determined at (20 ± 12) s−1, which is 55 times lower that the literature value.11

Table 3.5: Kinetic parameters (KM and kcat) for the tested substrates. No values were obtained

for compounds 2, 10 and 11 due to unsuccessful fitting routines. Compound 13 was not

determined because of unavailability in sufficient quantities. Values calculated from

triplicates and subsequent fitting of parameters. For plots see Section 3.8.1. n.d. = not

determined due to limited substrate availability, u.f. = unsuccessful fit.

Substrate KM / mM kcat / 1/s kcat/KM / M-1s-1

D-galactose (1) 180 ± 20 20 ± 12 111

lit. GOase wt for (1)11 175 1180 6742

α-D-methylgalactoside (2) n.d. n.d. n.d.

β-D-methylgalactoside (3) 275 ± 35 14 ± 2.0 51

D-glucose (4) 450 ± 88 11 ± 1.0 24

α-D-methylglucoside (5) 500 ± 176 2.7 ± 0.6 5

β-D-methylglucoside (6) 450 ± 112 5.4 ± 0.9 12

raffinose (7) 230 ± 168 3.5 ± 1.7 15

lactose (8) 600 ± 318 8.0 ± 2.0 13

sucrose (9) 350 ± 130 0.4 ± 0.1 1

maltose (10) n.d. n.d. n.d.

maltotriose (11) u.f. u.f. u.f.

cellobiose (12) 350 ± 87 1.7 ± 0.3 5

cellotriose (13) n.d. n.d. n.d.


74

The KM values are determined to be in the high mM rage with large errors. Since the KM

gives the substrate concentration at which half vmax is reached, > KM · 2 should be the highest

concentration in the substrate range while recording kinetic data. Unfortunately, most of the

substrates are not soluble in molar concentrations. Therefore, vmax is not determined with

enough certainty because the Michaelis-Menten curve does not reach saturation.

Subsequently the quality of KM is poor as well. All values (KM and kcat) are obtained from a

software based fitting routine. In the cases of 2, 10 and 11 the fitting of the raw data resulted

in no values for KM and kcat at all.

Additionally, the obtained raw data before fitting from three independent measurements

varies significantly. The used HRP-ABTS assay seems to be highly sensitive. Generally, the

stability of the chromophore (ABTS•+) is pH dependent and can be over-oxidised to be

colourless (Figure 3.7).

However, despite the high errors of the obtained values, looking at kcat/KM as a

measurement for the quality of a biocatalyst, it is notable that the β-glucoside (6) is favoured

over the α-glucoside (5). Similarly, lactose (8) has a high KM, nevertheless the kcat/KM is

comparable to the methylglucosides.


75

Figure 3.7: Oxidation states of 2,2’-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid

(ABTS). A single electron oxidation gives rise to the mono- radical ABTS•+ which has an

absorption maximum at around 420 nm. A further oxidation would result in the colour-less

non-radical ABTS++.

S

NN N

N

S

-O3S

SO3-

S

N•+N N

N

S

-O3S

SO3-

S

N•+N N

N•+

S

-O3S

SO3-

ABTS

ABTS

ABTS++

- e-

- e-

+ e-

+ e-


76

3.6.1.3 NMR spectroscopy

The way specific activities and kinetic parameters were determined gives information

about the production of H2O2, which is linked to substrate consumption. To study the

formation of product an NMR assay was developed analysing the crude reaction mixture.

Initially, 13C-NMR spectra were recorded to monitor the formation of carbonyl carbons in

the 6-oxo-compounds in characteristic windows around 86 to 89 ppm for the hydrate and

approximately 186 ppm for the aldehyde (not shown).

To simplify the process and learn something about quantitative turnover 1H - NMR was

used in a similar approach directly analysing the crude mixture. The 1H -NMR spectra (Figure

3.8 and Figure 3.9) show the formation of products from the 1-O-methylated mono-

saccharides:12,13

• α-D-methylgalactoside-6-hydrate (2a): 1H -NMR (400 MHz, D2O): δ(ppm) = 5.08

(1H, Ha-6, d), 4.07 (1H, Ha-4, dd), 3.93 (1H, H-4, dd), 3.38 (3H, Ha- OMe, s), 3.37

(3H, H-OMe, s);

• α-D-methylglucoside-6-hydrate (5a): 1H -NMR (400MHz, D2O): δ(ppm) = 5.23 (1H,

Ha-6, d), 3.38 (3H, Ha-OMe, s), 3.37 (3H, H-OMe, s);

• β-D-methylglucoside-6-hydrate (6a): 1H -NMR (400MHz, D2O): δ(ppm) = 5.23 (1H,

Ha-6, d), 4.34 (2H, Ha/H-1, m), 3.54 (3H, Ha-OMe, s), 3.53 (3H, H-OMe, s).


77

Furthermore, the oxidation products (10a-13a) of the di- and tri-saccharides maltose,

maltotriose, cellobiose and cellotriose were identified by comparing the spectra of mixed

samples to the respective starting materials (Figure 3.10 and Figure 3.11).

• maltose-12-hydrate (10a): 1H -NMR (400MHz, D2O): δ(ppm) = 5.44 (1H, Ha-7, d),

5.38 (1H, H-7, d), 5.24 (1H, Ha-12, d), 5.18 (2H, Ha/H-1β, d), 4.62 (2H, Ha/H-1α,

d);

• maltotriose-18-hydrate (11a): 1H -NMR (400MHz, D2O): δ(ppm) = 5.41 (2H, Ha/H-

13, d), 5.37 (2H, Ha/H-7, d), 5.24 (1H, Ha-18, d), 5.19 (2H, Ha/H-1β, d), 4.62 (2H,

Ha/H-1α, d);

• cellobiose-12-hydrate (12a): 1H -NMR (400 MHz, D2O): δ(ppm) = 5.24 (1H, Ha-12,

d), 5.19 (2H, Ha/H-1β, d), 4.62 (2H, Ha/H-1α, d), 4.48 (1H, Ha-7, d), 4.47 (1H, H-

7, d);

• cellotriose-18-hydrate (13a): 1H-NMR (400 MHz, D2O): δ(ppm) = 5.24 (1H, Ha-18,

d), 5.19 (2H, Ha/H-1β, d), 4.62 (2H, Ha/H-1α, d), 4.49 (4H, Ha/H-13, Ha/H-7, m).


78

Figure 3.8: 1H-NMR spectra of crude reactions oxidising methylated galactosides using

GOase F2. (a) Spectrum of crude mix containing 2 and 2a. Estimated yield by OMe signals:

37%. (b) Spectrum of crude mix containing 3 and 3a. Estimated yield by H/H’-4 signals:

20%.

3.33.43.53.63.73.83.94.04.14.24.34.44.54.64.74.84.95.05.1f1 (ppm)

5.0

2.9

1.6

1.0

1.0

O

OH

OH

H

OH

OH

OH H

H

O

CH3

alpha-D-methylgalactoside-6 hydrate (2a)

3.33.43.53.63.73.83.94.04.14.24.34.44.54.64.74.84.95.05.1f1 (ppm)

15.0

4.3

1.0

0.6

2.0

0.6

1.9

1.3

O

OH

OH

H

OH

OH

OH H

HO

CH3

beta-D-methylgalactoside-6 hydrate (3a)


79

Figure 3.9: 1H-NMR spectra of crude reactions oxidising methylated glucosides using

GOase F2. (a) Spectrum of crude mix containing 5 and 5a. Estimated yield by OMe signals:

60%. (b) Spectrum of crude mix containing 6 and 6a. Estimated yield by OMe signals: 40%.

3.13.23.33.43.53.63.73.83.94.04.14.24.34.44.54.64.74.84.95.05.15.2f1 (ppm)

2.1

3.1

1.0

O

OH

H

H

OH

OH

OH H

OH

O

CH3

alpha-D-methylglucoside-6 hydrate (5a)

3.13.23.33.43.53.63.73.83.94.04.14.24.34.44.54.64.74.84.95.05.15.2f1 (ppm)

4.4

2.9

1.6

1.5

2.5

1.0

O

OH

H

H

OH

OH

OH H

OHO

CH3

beta-D-methylglucoside-6 hydrate (6a)


80

Figure 3.10: 1H-NMR spectra of crude reactions oxidising α-linked oligomers using GOase

F2. (a) Spectrum of crude mix containing 10 and 10a. Estimated yield by H-12 signals: 64%.

(b) Spectrum of crude mix containing 11 and 11a. Estimated yield by H-18 signals: 100%.

3.13.23.33.43.53.63.73.83.94.04.14.24.34.44.54.64.74.84.95.05.15.25.35.45.5f1 (ppm)

0.57

0.35

0.65

0.57

0.39

0.61

0.34

0.61

1 α/β

OOH

OH

O

OH

OH

O

OH

OHOH

OHH

OH

H

7

12

Maltose-12-hydrate (10a)

3.13.23.33.43.53.63.73.83.94.04.14.24.34.44.54.64.74.84.95.05.15.25.35.45.5f1 (ppm)

0.61

1.00

0.62

0.42

0.91

0.97

0.92

OOH

OH

O

OH

OH

O

OH

O

OH

OHH

O

OH

OHOH

OHH

OH

H

1 α/β

7

13

18

Maltotriose-18-hydrate (11a)


81

Figure 3.11: 1H-NMR spectra of crude reactions oxidising β-linked oligomers using GOase

F2. (a) Spectrum of crude mix containing 12 and 12a. Estimated yield by OMe signals: 26%.

(b) Spectrum of crude mix containing 13 and 13a. Estimated yield by OMe signals: 74%.

3.23.33.43.53.63.73.83.94.04.14.24.34.44.54.64.74.84.95.05.15.25.3f1 (ppm)

1.00

0.62

0.41

0.26

OOH

OH

H

OH

OH

O O

OH

OHOH

OHOH

H

1 α/β 7

12

Cellobiose-12-hydrate (12a)

3.23.33.43.53.63.73.83.94.04.14.24.34.44.54.64.74.84.95.05.15.25.3f1 (ppm)

2.00

0.64

0.41

0.78

OOH

OH

H

OH

OH

O O

OH

OH

OHOH

H

O O

OH

OHOH

OH

H

1 α/β

713

18

Cellotriose-18-hydrate (13a)


82

The application of 1H- instead of 13C-NMR makes a quantitative analysis possible. Upon

integration of signals the ratios between residual substrate and formed product can be

translated into yields. Table 3.6 summarises the turnover of the analysed compounds.

Table 3.6: Calculated yields of oxidation of α- and β-linked mono- and oligo-saccharides

after integration of 1H-NMR signals. Signal ratios in the crude spectra give conversions.

product yield / %

α-D-methylgalactoside-6-hydrate (2a) 37

α-D-methylglucoside-6-hydrate (5a) 60

maltose-12-hydrate (10a) 64

maltotriose-18-hydrate (11a) >99

β-D-methylgalactoside-6-hydrate (3a) 20

β-D-methylglucoside-6-hydrate (6a) 40

cellobiose-12-hydrate (12a) 26

cellotriose-18-hydrate (13a) 74

The analysed compounds can be grouped according to their structure into α- and β-linked

saccharides as well as mono- and oligo-saccharides. As described earlier, the glucosides show

60 to 100% more yield than their galactoside equivalent. This higher activity of the F2 variant

towards glucosides is notable.

Additionally, α-glycosides are preferred as these anomers yield 50 to 100% higher

conversions than the β-equivalents. This is the case for the investigated di- and tri-saccharides

as well with maltose (10) and maltotriose (11) being converted to 64 and 100 % respectively.

Cellobiose (12) and cellotriose (13) are not as well accepted substrates.

However, longer saccharides yield higher conversions. Looking back at a putative natural

function of GOase it is likely that it is used to oxidise natural occurring glycans rather than

freely available D-galactose. Therefore, the increased activity of GOase for longer chains may

be explained by its intrinsic preference for carbohydrate chains.


83

3.6.2 Lactose oxidation

3.6.2.1 Oxygen availability

The comparably high specific activity of galactose oxidase (wild-type and F2) for lactose

led to the identification of the same as a very interesting substrate for industrial purposes

(Figure 3.12) with applications for the resulting dialdehyde (8b).

Figure 3.12: Enzymatic oxidation of lactose (8). Galactose oxidase oxidises the 6-position

of the terminal galactoside specifically. The product forms an equilibrium between the

hydrate and the corresponding aldehyde (8a). Catalase acts as hydrogen peroxide removing

agent and partly recycles oxygen. Upon ring opening, 6-oxo lactose presents a dialdehyde

(8b) with various possible industrial applications.

BASF provided GOase wt in form of yeast culture supernatant. With the supplied enzyme

substrate concentrations between 25 mM to 200 mM were tested to get full conversion as

this would simplify the purification. Especially with higher lactose concentrations the

availability of oxygen can be a limiting factor which ultimately limits the scale and set-up of

large scale biotransformations.

Following the idea of 13C-NMR to monitor reactions switching to 1H-NMR resulted in

increased speed and resolution of the analysis. Several characteristic signals (H-6 of the

O

OH

OHOH

HO1/2 O2 H2O2

1/2 O2

GOase

8

O OHO

OH

HO

OH O

OH

OH

HO O OHO

OH

HO

OH

O

O

OH

OH

HO O OHHO

OH

HOO

catalase8a

8b

O


84

galactoside shifts downfield to 5.15 ppm) were found to prove the structure and give

information about conversions12. With lower substrate concentration higher conversions can

be observed giving full conversion at 50 mM lactose (Figure 3.13). This might be explained

with the limitation of available oxygen in solution.

The O2 concentration can be very important. Therefore, different hydrogen peroxide

removal strategies were employed finding that the application of catalase without horse

reddish peroxidase (HRP) gives higher yields (Table 3.7). Upon switching to catalase as the

only hydrogen peroxide removing agent oxygen is partly recycled, in contrast to HRP which

does not generate oxygen at all.

Table 3.7: Calculated yields of lactose oxidation based on integration of 1H-NMR signals at

4.10 and 4.45 ppm for varying substrate concentrations and hydrogen peroxide removal

techniques. Full conversion is reached at ≤50 mM substrate concentration without HRP.

Values calculated form single spectra.

[lactose] / mM + HRP - HRP

25 99% >99%

50 96% >99%

100 32% 60%

200 16% 30%


85

Figure 3.13: 1H-NMR spectra (excerpt) of lactose (starting material (SM); A) and lactose

oxidation reactions at varying substrate concentrations (B- E). Signals at 5.15, 4.45 and

4.10 ppm represent H-6, H-1 and H-4 of the galactoside in (2), respectively. The H-1 doublet

at 4.42 ppm shifts down-field to 4.45 ppm in the oxidised product (2).

3.6.2.2 Preparative Lactose oxidation

In order to demonstrate the industrial applicability of the lactose process scale-up

reactions were run. On gram-scale the reactions did not yield full conversion at 100 mM

possibly due to limited oxygen availability. However, the formation of the 6-oxo-species (8a)

was identified with > 70% yield (for [GOase wt] = 8 mg/mL and [lactose] = 100 mM) by

1H-NMR as described previously. Product isolation proved to be challenging. A

recrystallisation approach according to Bund et al. and de Souza et al. failed under the

conditions tested.14,15

Future improvements could involve a compartmentalised approach to separate the

enzymes from the substrate reservoir and push the reaction to completion.


86

3.7 Conclusion

The goal of this study was the evaluation the substrate scope of galactose oxidase (GOase)

amongst possible saccharides of interest. Specific activities and kinetic parameters showed,

that the GOase F2 variant is a potent candidate to oxidise the target glucosides, as well as

lactose.

The sensitive ABTS-assay lead to specific activities and kinetic parameters with big error

bars. Nevertheless, the extracted KM in the high millimolar range revealed the problem of

limited solubility of the substrates. Furthermore β-glucosides appeared to be the better

substrates with higher KM/kcat.

The NMR studies showed that GOase F2 does turn over glucosides to a higher degree

that galactosides. Beyond the known mono-saccharide substrates, the versatility of GOase

F2 has been demonstrated oxidising di- and even tri-saccharides with increasing efficiency.

The NMR experiments also revealed a tendency towards α-glucosides being preferred.

Looking back at the kinetic data though, it is possible, that β-glucosides are too good

substrates which leads to H2O2 poisoning the active site of the enzyme and ultimately to

inactivity. In contrast, the α-anomers might be slower substrates and therefore the enzyme

lasts longer to lead to higher conversion.

The H2O2 removal strategy is crucial. For the tested substrates catalase without HRP

seems to work fine. Catalase partly recycles oxygen which enhances the conversion.

Additionally, it has been found that for other GOase substrates little amounts of HRP in the

reaction mixture can reactivate GOase active site which is likely to work for lower substrate

concentrations or lower oxygen requirements.

Since GOase exclusively oxidises terminal glycosides the configuration of the glycosidic

bond might be of importance possibly interfering with the binding site. Consequently α- and


87

β-anomers bind differently which could influence the distance between 6-OH and the Cu-

tyrosyl radical leading to poor reaction efficiency.

Generally speaking, these findings suggest that GOase F2 is a valid candidate for the

industrial oxidation of lactose and glucosidic oligomers in α-configuration similar to amylose.


88

3.8 Appendix

3.8.1 Supporting information to Section 3.6.1.2 (Michaelis-Menten plots GOase kinetics)

Specific activities were determined for a range of substrate concentrations ensuring the

overall substrate consumption never exceeded 20 %. Three independent measurements were

performed, averages calculated, and standard deviations calculated. Specific activities were

plotted against the respective substrate concentration to yield Michaelis-Menten plots. Fitting

to a model was achieved using Origin 9.0.

𝑦 =𝑉UQ^ ∙ 𝑥𝐾a + 𝑥

Values for Vmax were used to calculate kcat . Errors values were retrieved through the

fitting routine.


89

Figure 3.14: Michaelis-Menten plots of GOase kinetic data. Data were plotted and curves

(red lines) were fitted to data points disregarding outliers (red squares) with Michaelis-

Menten-routine within Origin 9.0.


90

3.9 References

[1] G Avigad, D Amaral, C Asensio, and B L Horecker. The D-galactose oxidase of

Polyporus circinatus. J. Biol. Chem., 237:2736–43, 1962.

[2] J W Whittaker. The radical chemistry of galactose oxidase. Arch. Biochem. Biophys.,

433(1):227–39, 2005.

[3] N Ito, S E Phillips, C Stevens, Z B Ogel, M J McPherson, J N Keen, K D Yadav, and

P F Knowles. Novel thioether bond revealed by a 1.7 A crystal structure of galactose

oxidase. Nature, 350(6313):87–90, 1991.

[4] M S Rogers, A J Baron, M J McPherson, P F Knowles, and D M Dooley. Galactose

Oxidase Pro-Sequence Cleavage and Cofactor Assembly Are Self-Processing

Reactions. J. Am. Chem. Soc., 122(5):990–991, 2000.

[5] M S Rogers, R Hurtado-Guerrero, S J Firbank, M A Halcrow, D M Dooley, S E V

Phillips, P F Knowles, and M J McPherson. Cross-link formation of the cysteine 228-

tyrosine 272 catalytic cofactor of galactose oxidase does not require dioxygen.

Biochemistry, 47(39):10428–39, 2008.

[6] L Que and W B Tolman. Biologically inspired oxidation catalysis. Nature,

455(7211):333–340, 2008.

[7] C D Borman, C G Saysell, C Wright, and A G Sykes. Mechanistic studies on the single

copper tyrosyl-radical containing enzyme galactose oxidase. Pure Appl. Chem.,

70(4):897–902, 1998.

[8] J W Whittaker. Free radical catalysis by galactose oxidase. Chem. Rev., 103(6):2347–63,

2003.

[9] J B Rannes, A Ioannou, S C Willies, G Grogan, C Behrens, S L Flitsch, and N J Turner.

Glycoprotein labeling using engineered variants of galactose oxidase obtained by

directed evolution. J. Am. Chem. Soc., 133(22):8436–8439, 2011.

[10] N J Davis and S L Flitsch. Selective oxidation of monosaccharide derivatives to uronic

acids. Tetrahedron Lett., 34(7):1181–1184, 1993.

[11] L D Kwiatkowski, M Adelman, R Pennelly, and D J Kosman. Kinetic mechanism of

the Cu(II) enzyme galactose oxidase. J. Inorg. Biochem., 14(3):209–22, 1981.


91

[12] V Bonnet, R Duval, and C Rabiller. Oxidation of galactose and derivatives catalysed

by galactose oxidase: structure and complete assignments of the NMR spectra of the

main product. J. Mol. Catal. B Enzym., 24-25:9–16, 2003.

[13] S Singh, S Nambiar, R A Porter, T L Sander, K G Taylor, and R J Doyle. Dialdosides-

(1,5) of glucose and galactose: synthesis, reactivity, and conformation. J. Org. Chem.,

54(10):2300–2307, 1989.

[14] R K Bund and A B Pandit. Rapid lactose recovery from buffalo whey by use of ‘anti-

solvent, ethanol’. J. Food Eng., 82(3):333–341, 2007.

[15] R Rosa de Souza, R Bergamasco, S Claúdio da Costa, X Feng, S H Bernardo Faria, and

M Luiz Gimenes. Recovery and purification of lactose from whey. Chem. Eng. Process.

Process Intensif., 49(11):1137–1143, 2010.

[16] S E Deacon and M J McPherson. Enhanced expression and purification of fungal

galactose oxidase in Escherichia coli and use for analysis of a saturation mutagenesis

library. Chembiochem, 12:593–601, 2011.

92

Chapter 4 Carbohydrate arrays for fast and sensitive

hydrolase characterisation

4.1 Summary

Following the enzymatic oxidation of oligo-saccharides in solution monitored by NMR,

the next step was developing a screening platform to characterised fungal glycoside

hydrolases and their activity towards plant derived oligo- and poly-saccharides. A

hydrophobic SAM platform coupled with an adapted reductive amination strategy forms the

basis for a rapid MALDI-ToF assay.

4.2 Contribution

JvM designed the overall study. ALD performed NMR experiments JvM performed

biological experiments. As part of this thesis MR established the SAM formation

methodology (based on suggestions by Dr A. Ruiz-Sanchez). MR and BT performed

reductive aminations (CJG advised). MR, JvM and BT performed MALDI-ToF experiment.

JvM, BT and MR analysed data was. All authors wrote the manuscript.

4.3 Manuscript

This manuscript is published in Scientific Reports 7, Article number: 43117 (2017)

doi:10.1038/srep43117.

Chapter 4 - Carbohydrate arrays

93

Application of carbohydrate arrays coupled with mass

spectrometry to detect activity of plant-polysaccharide

degradative enzymes from the fungus Aspergillus niger

Jolanda M van Munster1*, Baptiste Thomas2, Michel Riese2, Adrienne L Davis3, Christopher

J Gray2, David B Archer1, Sabine L Flitsch2

1 Fungal Biology and Genetics, School of Life Sciences, University of Nottingham, University

Park, Nottingham NG7 2RD, UK 2 Chemical Biology, Manchester Institute for Biotechnology, University of Manchester, 131

Princess Street, Manchester M1 7DN, United Kingdom 3 School of Chemistry, University of Nottingham, University Park, Nottingham NG7 2RD,

UK

Email addresses: [email protected],

[email protected], [email protected],

[email protected], [email protected],

[email protected], [email protected]

* Corresponding author

Dr J.M. van Munster, Manchester Institute for Biotechnology, University of Manchester,

131 Princess Street, Manchester M1 7DN, United Kingdom,

[email protected]


94

Abstract

Renewables-based biotechnology depends on enzymes to degrade plant lignocellulose to

simple sugars that are converted to fuels or high-value products. Identification and

characterization of such lignocellulose degradative enzymes could be fast-tracked by

availability of an enzyme activity measurement method that is fast, label-free, uses minimal

resources and allows direct identification of generated products.

We developed such a method by applying carbohydrate arrays coupled with MALDI-ToF

mass spectrometry to identify reaction products of carbohydrate active enzymes (CAZymes)

of the filamentous fungus Aspergillus niger. We describe the production and characterization

of plant poly-saccharide-derived oligo-saccharides and their attachment to hydrophobic self-

assembling monolayers on a gold target. We verify effectiveness of this array for detecting

exo- and endo-acting glycoside hydrolase activity using commercial enzymes, and

demonstrate how this platform is suitable for detection of enzyme activity in relevant

biological samples, the culture filtrate of A. niger grown on wheat straw.

In conclusion, this versatile method is broadly applicable in screening and characterisation

of activity of CAZymes, such as fungal enzymes for plant lignocellulose degradation with

relevance to biotechnological applications as biofuel production, the food and animal feed

industry.

Introduction

The availability of well-characterised, affordable and efficient carbohydrate active

enzymes (CAZymes) that are capable of modifying or degrading plant-derived carbohydrates

underpins the food and feed industries as well as renewables-based biotechnology. Of

particular interest are enzymes capable of degrading complex lignocellulose, generating

simple sugars from which biochemicals and second generation biofuels can be produced.

This transformation would benefit from more efficient and cheaper enzyme mixtures,


95

enabled by the discovery of enzymes with improved stability, novel or improved catalytic

mechanisms or other helper proteins that contribute to synergistic substrate degradation1,2.

Genome sequencing of many microbes with high lignocellulose degradative capacity has

resulted in the discovery of many potential valuable CAZymes, (available in the CAZy

database3), but their biochemical characterization is lagging behind4. Furthermore, detailed

understanding of the regulation behind gene expression and protein secretion in fungal

enzyme production strains could optimize the industrial enzyme production process5,6.

The fungus Aspergillus niger is extensively used as an industrial producer of organic acids

and enzymes7. The genome of this fungus encodes a large set of CAZymes for the

degradation of plant poly-saccharides8, and these genes are expressed in response to

cultivation on lignocellulosic substrates that are of relevance as feedstock for biofuel

production, such as wheat straw, willow and sugar cane bagasse9-11. Despite the abundance

of research on A. niger CAZymes, the biochemical characterisation of many of these enzymes

is incomplete. This lack of knowledge on enzyme specificities, in particular substrate and

product range, prevents a complete understanding of the enzymatic machinery responsible

for the degradation of lignocellulose. Methods for rapid, sensitive detection and

characterisation of enzyme activity on plant-derived substrates are therefore an essential tool.

However, key challenges are posed by a trade-off between method throughput and derived

information content, and the limited availability of characterised complex substrates.

Carbohydrate arrays are eminently suited to resolve the key challenges associated with

characterising CAZyme activity, namely simultaneous screening of multiple reactions and

conditions on minimal amounts of substrates. Carbohydrate arrays, including those carrying

plant-based oligo-saccharides,12,13 have been successfully used to screen for enzyme activities

as well as binding specificities of antibodies and carbohydrate binding proteins (for recent

reviews see14-16). These arrays offer high analytical resolution as the complex structural

features of the poly-saccharides are separated in sub-structures. However, a limitation of


96

many arrays is their reliance on labelled substrates, binding domains or antibodies to aid

visualisation. The coupling of carbohydrate arrays with MALDI-ToF MS enables sensitive,

rapid and label-free visualisation of enzymatic products directly at the array surface and can

also allow for structural features to be ascertained through tandem MS17. Recently,

carbohydrate arrays coupled with MS have been applied to determine the substrate specificity

of carbohydrate binding modules and to detect and identify exo-glycosidase enzyme

activities, active on terminal residues of the attached carbohydrates13,14, as well as limited

endo-acting activity on short oligo-saccharides18.

Here, we expanded on this work by generating MALDI-ToF MS compatible carbohydrate

arrays with plant-derived oligo-saccharides, including those with a high degree of

polymerisation (DP), and applying them to detect and identify substrates and products of

both endo- and exo- acting fungal CAZymes, including a proof of principle application

towards detection of activity of lignocellulose-active CAZymes enzymes secreted by A. niger.

From plant poly-saccharides we produced, isolated and characterized oligo-saccharides with

a high DP. We use these, as well as commercially available oligo-saccharides, to generate a

carbohydrate array on a hydrophobic self-assembling monolayer (SAM) of alkanethiols

coating the gold surface of a MALDI-ToF target (Figure 4.1). We show that this system can

be used to detect both endo- and exo-acting enzyme activity and we apply the arrays to detect

substrate changes caused by activity of CAZymes of A. niger.


97

Figure 4.1: Schematic overview of carbohydrate array construction and use. After formation

of a hydrophobic SAM of alkanethiols on a gold-coated target, hexadecylamine-labelled

oligo-saccharides are immobilised via hydrophobic-interaction. Products of enzyme activity,

here exemplified by removal of DP2, as well as the original substrate, can be identified with

MALDI-ToF MS.


98

Methods

Fungal strains and growth conditions.

The fungus Aspergillus niger strain AB4.119 was grown on potato dextrose agar slopes at

28°C until spores were produced. All AB4.1 cultures were supplemented with 10 mM uridine.

Spores were harvested with 0.1 % (v/v) Tween 20 and liquid cultures of 100 mL aspergillus

minimal medium (AMM)9 with 1 % glucose were inoculated with 106 spores mL-1 and

incubated at 28°C, 150 RPM for 48 h. Mycelium was harvested using Miracloth, washed with

AMM without carbon source, and 1.5 g (wet weight) was transferred to cultures of 100 mL

AMM with 1 % (w/v) ball milled wheat straw and incubated for 24 h at 28°C, 150 rpm.

Preparation and composition of the straw has been described previously9. Mycelium and

culture filtrate were separated by filtration through Miracloth, the culture filtrate was

concentrated 20-fold using Vivaspin 20 columns with a 10 kDa Mw cut-off and frozen until

further analysis.


99

Gene expression analysis.

Fungal mycelium was disrupted by grinding it in liquid nitrogen using a mortar and pestle.

RNA was isolated using Trizol (Invitrogen), followed by clean-up and DNAse treatment

using the NucleoSpin RNA purification kit (Macherey-Nagel). Absence of genomic DNA

was verified by PCR. cDNA was synthesized using SuperScript 3 reverse transcriptase

(Invitrogen) using 0.5 µg total RNA and oligo(dT) primer. To measure expression of cbhA,

cbhB, xynB and actA, qRT-PCR was performed on an Applied Biosystems 7500 Fast Real-

Time PCR system in a 10 µl total volume with 2 µl 4x diluted cDNA, 0.2 µM of each primer

(Supplementary Table S1) and 5 µl FAST SYBR-Green Master Mix (Applied Biosystems).

qRT-PCRs were performed with a 95°C 20 s initial denaturation, followed by 40 cycles of

95°C for 3 s and 60°C for 30 s, with all measurements in technical duplicate. Production of

a single product was verified using a melt-curve. Gene expression levels were calculated from

a genomic DNA standard curve, and corrected for expression of the reference gene20 actA.

Values are given as mean standard error of biological triplicate experiments.

Protein analysis.

Proteins in fungal culture filtrate were separated on 4-20% Tris-glycine Novex SDS-

PAGE gels (Invitrogen) and visualized with silver staining21. For protein identification,

proteins were precipitated by adding 5 mL trichloroacetic acid to 20 mL culture filtrate,

incubating for 1 h at 4°C and spinning at 16060 g for 10 min. Pellets were washed twice with

200 µL acetone, loaded onto a 10% SDS-PAGE gel and electrophoresis performed until the

dye front had moved approximately 1 cm into the separating gel. Each gel lane was then cut

into a single slice, on which in-gel tryptic digestion was performed using a DigestPro

automated digestion unit (Intavis Ltd).

Peptides were analysed generally as described by22 with the following exceptions: an

Acclaim PepMap C18 nano-trap column (Thermo Scientific) was used for injection. Peptides

were resolved with a 7 segment gradient ( 1-6 % solvent B over 1 min, 6-15 % B over 58


100

min, 15-32 % B over 58 min, 32-40 % B over 5 min, 40-90 % B over 1 min, held at 90 % B

for 6 min and then reduced to 1 % B over 1 min), Tandem mass spectra were acquired in

the mass range m/z 300 to 2000, followed by MS/MS for the top twenty multiply charged

ions. Data were processed with Proteome Discoverer software v1.4 (Thermo Scientific),

searched against the A. niger SwissProt database (downloaded 23-03-16) using the SEQUEST

algorithm, with a 10 ppm peptide precursor tolerance and a maximum of 1 missed cleavage,

and peptide data was filtered to meet a false discovery rate of 1 % while retaining only

proteins identified by ≥ 2 unique peptides. The mass spectrometry proteomics data were

deposited in the PRIDE repository with dataset identifiers PXD005699 and

10.6019/PXD005699.

Isolation of hemicellulose and production of oligo-saccharides.

The hemicellulose fraction of wheat straw was isolated generally as described elsewhere23.

Briefly, 3 % (w/v) wheat straw was suspended in 0.5 M KOH at 40°C for 2.5 h, after which

the suspension was filtered through Miracloth and neutralized with 6 M acetic acid. Polymers

were precipitated by adding 3 volumes of ethanol and incubation at 4°C for ≥ 1 h. Precipitate

was collected by centrifugation, washed with 70% (v/v) ethanol and lyophilized, resulting in

the recovery of 2.4 g hemicellulose enriched fraction from 16 g of wheat straw.

Oligo-saccharides were generated by hydrolysis of 100 mg hemicellulose aliquots in 2 mL

0.5 M H2SO4 for 10 min at 100°C, after which samples were cooled and neutralized with

sodium carbonate. Water-soluble material was collected and lyophilized. The material (0.7 g)

was dissolved in 10 mM ammonium bicarbonate, rid of particles by centrifugation, and

separated on size on a Bio-Gel P4 fine (Biorad) column with a diameter of 0.9 cm and length

of 70 cm. The column was equilibrated and run in 10 mM ammonium bicarbonate at 12 mL

h-1 using a peristaltic pump, and elution fractions of 3 mL were collected for 5 hours.

Wheat arabinoxylan (Medium viscosity, Megazyme) was hydrolysed to generate oligo-

saccharides using Thermomyces lanuginosus endo-xylanase (X2753, Sigma Aldrich). A 1 % (w/v)


101

solution of arabinoxylan in boiling water was prepared, and cooled to 60°C. Xylanase (8 mg

mL-1 suspension in sodium citrate buffer pH 6) was added as 200 µL per g arabinoxylan and

incubated for 30 min at 60°C. The suspension was boiled for 5 min to stop enzyme activity,

cooled and lyophilized. Recovered soluble material (1 g) was dissolved in 4 mL 10 mM

ammonium bicarbonate and separated on the Bio-Gel column as described above.

Characterization of oligo-saccharides.

Fractions containing oligo-saccharides were analysed by TLC, MALDI-ToF-MS and

NMR. TLC was performed by separating samples on TLC silica 60 plates with aluminium

backing (Merck), and a liquid phase of butanol: ethanol: water in 5:5:3 ratio (v/v).

Carbohydrates were visualized by dipping the plate in orcinol solution (250 mg orcinol

(Sigma) in 95 mL ethanol, 5 mL H2SO4) and heating the plate to 100°C. Samples were

prepared for MALDI-ToF by co-crystalizing them with a 5% 2,5-dihydroxybenzoic acid

(DHB) matrix solution in methanol. MALDI-ToF MS was performed using the Bruker

Ultraflex 3 in positive mode.

Selected samples were exchanged with deuterium oxide and analysed by NMR.

Measurements were made on a Bruker AV(III)500 NMR spectrometer using a dual 1H/13C

helium-cooled cryoprobe operating with a sample temperature of 298 K. Spectra were

approximately referenced in the 1H dimension using the deuterium lock (equivalent to setting

the HDO peak = 4.75 ppm). Accurate referencing was achieved by setting the easily

identifiable reducing end α-xylose resonance equal to 5.184 ppm. This procedure gives

chemical shifts directly equivalent to those referenced relative to internal acetone at

(d=2.225 ppm) as given in24. The 13C chemical shifts were measured relative to external DSS

(4,4-dimethyl-4-silapentane-1-sulfonic acid) set to -1.6 ppm and thus are indicative only.

1H NMR spectra were recorded with 30 degree pulses, a data acquisition time of 3.17 s

and a relaxation delay of 1 s. The spectral width was 20.6 ppm. Data were zero-filled and

multiplied by an exponential window function with lb = 0.3 Hz prior to Fourier


102

transformation. Spectra were baseline corrected and integrated using the standard routine in

the Bruker TOPSPIN software (no lineshape fitting was attempted).

HSQC spectra were acquired phase sensitive in echo/anti-echo mode in a 1k x 256 data

matrix using the Bruker pulse program hsqcetgpsisp2.2 which utilises double inept transfer.

13C decoupling was applied during acquisition. Prior to Fourier transformation, data points

in the F1 dimension were doubled using linear prediction (64 coefficients), the data matrix

was zero-filled to 2k x 1k real data points. A cosine-squared window function was applied to

the data.

TOCSY spectra were acquired phase sensitive in echo/anti-echo mode in a 2k x 512 data

matrix using the Bruker pulse program dipsi2etgpsi which utilises the DIPSI 2 sequence for

mixing. The mixing time was set to 15 ms to observe direct correlations or 80 ms to observe

coupling networks. Prior to Fourier transformation, data points in the F1 dimension were

doubled using linear prediction, the data matrix was zero-filled to 4k x 1k real data points. A

cosine-squared window function was applied to the data.

Preparation of carbohydrate arrays.

Maltose (Sigma-Aldrich), cellotetraose, cellohexaose and xylohexaose (Megazyme) were

coupled to hexadecylamine via reductive amination; 1-5 mg carbohydrate with

hexadecylamine (2 : 1 molar ratio), 1 M sodium cyanoborohydride in 1 mL of 90% (v/v)

DMSO, 10% (v/v) acetic acid and incubated overnight at 70°C. Reactions were diluted with

methanol before use.

Reductive amination of non-commercial oligo-saccharides (hemicellulose fractions H5,

H6, H7, H9 and arabinoxylan fractions AX10, AX11, AX12, AX15) was performed

according to an alternative method based on the work of Gildersleeve et al.25. Briefly, oligo-

saccharides (0.4 mg) were dissolved in a mixture of sodium borate buffer (31 µL of a 400

mM solution, pH 8.5) and sodium sulfate buffer (21 µL of a 3 M solution, 50°C), then

hexadecylamine (12 µL from 6.43 mg of hexadecylamine in 3 mL of methanol) and sodium


103

cyanoborohydride (1.77 mg) were added. The reaction was allowed to warm at 56°C.

Labelling of the oligo-saccharides could be visualised by application of 1 µL of α-cyano-4-

hydroxycinnamic acid (CHCA) (20 mg mL-1 in a mixture of 50% (v/v) acetonitrile and 50 %

(v/v) water containing 0.1 % of trifluoroacetic acid (TFA)) followed by MALDI-ToF MS.

After 8 h, the reaction mixture was cooled at room temperature, diluted in water/methanol

(2 mL, typically in a 4:1 ratio) and stored at room temperature.

Gold-coated MALDI target plates (AB Sciex Ltd (AB plates)) were cleaned with a mixture

of 30% (v/v) hydrogen peroxide and 70 % (v/v) sulphuric acid for 20 min, rinsed extensively

with water, followed by methanol and dried under a stream of nitrogen. Then, the plates

were incubated overnight in 1-undecanethiol (37.5 µL in 20 mL of methanol) forming a

hydrophobic SAM. Plates were washed with methanol, dried under a stream of nitrogen, and

1 µL of hexadecylamine labelled oligo-saccharides, diluted to 0.2 mg mL-1 in water:methanol

(4:1, vol/vol), were spotted on each well. After incubation for 20 min in a sealed container,

plates were washed twice with water. MALDI-ToF MS confirmed immobilization after

application of 1 µL of DHB (15 mg mL-1 in a mixture of 50% (v/v) acetonitrile and 50 %

(v/v) water containing 0.1% of TFA) directly on the gold plate.

Enzyme reactions on carbohydrate arrays.

Reaction conditions for commercial enzymes were designed for incomplete substrate

degradation, thus allowing identification of both substrates and products in one mass

spectrum. β-glucosidase from almond (49290, Sigma-Aldrich) was prepared as 1 mg mL-1 (≥

6 U mL-1) in water, then 2 µL was spotted on the carbohydrate array and the plate was

incubated for 30 min at 37°C in a sealed container. Endo-xylanase from Thermomyces

lanuginosus (X2753, Sigma Aldrich) was prepared as 5 ng mL-1 (≥ 10 mU mL-1), mixed 1:1 with

100 mM citric acid-sodium phosphate buffer (pH 8), then 2 µL was spotted on the

carbohydrate array and the plate was incubated for 10 min at 30°C in a sealed container.

Measurements to determine the minimal enzyme amount that can reliably be detected were


104

performed with a dilution range of xylanase, incubated at pH 6 for 2 h at 30°C. Culture

filtrates from A. niger grown on wheat straw (1-fold or 20-fold concentrated) were mixed 1:1

with 100 mM citric acid-sodium phosphate buffer (pH 4), then 2 µL were spotted on the

carbohydrate array and the plate was incubated for 2 h at 30°C in a sealed container. After

incubation, the plates were dipped into distilled water and shaken gently for 1 min, and then

dried under a stream of nitrogen. Reaction products were identified with MALDI-ToF MS

and MS-MS on a Ultraflex II TOF/TOF, using 2’, 4’,6’-trihydroxyacetophenone

monohydrate (THAP) (10 mg mL-1 in acetone) or DHB (15 mg mL-1 in a mixture of 50%

(v/v) acetonitrile and 50% (v/v) water containing 0.1% of TFA) as matrix.

Mass spectrometric analysis of arrays.

Gold AB plates were loaded into the instrument using the MTB AB adapter (Bruker).

MALDI-ToF mass spectra (1500 shots/spectrum) were recorded on a Bruker Ultraflex II

instrument with a Smartbeam I laser in positive reflector ion mode. The instrument was

calibrated between m/z 700-3500 using a solution of peptide calibration mix II (Bruker

Daltonics, Bremen). Data were analysed and normalised using FlexAnalysis version 3.0

(Bruker).

Results

Hydrolytic enzymes secreted by A. niger grown on wheat straw.

The fungus A. niger produces a range of plant poly-saccharide degradative enzymes in

response to lignocellulose. The fungus was cultivated in liquid cultures containing glucose in

order to obtain biomass, and resulting mycelium was washed and transferred for 24 h to

liquid cultures containing ball milled wheat straw. These cultivation conditions have

previously been found to highly induce a large number of genes that encode (putative) plant

poly-saccharide active enzymes9. qRT-PCR showed that such genes were induced as

transcript levels of the cellobiohydrolases encoding genes cbhA, cbhB and the xylanase


105

encoding gene xynB were strongly increased compared to the repressive glucose conditions

(Figure 4.2). Culture filtrate was harvested and analysis using SDS-PAGE; a broad range of

proteins are indeed secreted in wheat straw cultures (Figure 4.2). Shotgun proteomics of

precipitated proteins from the wheat straw culture filtrate (Supplementary Table S4.2) shows

that the most abundant proteins are annotated as (putative) CAZymes active on cellulose

and hemicelluloses, in particular (arabino)xylan, and their predicted activities constitute a

mixture of exo- and endo-acting enzymes. Cellulolytic enzymes include (putative)

cellobiohydrolases CbhA (GH7) and CbhB (GH7), CbhC (GH6), endoglucanase EglB

(GH5-5), and several putative β-glucosidases. Xylanolytic enzymes include (putative)

endoxylanases XynA (GH10) and XynB (GH11), β -xylosidase xlnD (GH3), α-

arabinofuranosidases AxhA (GH62), AbfB (GH54), AfbA (GH51). Based on these results,

cellulose oligo-saccharides as well as (arabino)xylan-derived oligo-saccharides were

prioritized as substrates for the carbohydrate arrays.

Figure 4.2: Analysis of gene expression and protein secreted by A. niger. (a) Expression of

genes cbhA (encoding cellobiohydrolase CbhA), cbhB (encoding cellobiohydrolase CbhB),

and xynB (encoding xylanase XynB), as % of expression of actin encoding gene actA, in

liquid cultures containing glucose or wheat straw as carbon source, represented as mean

standard error, n = 3, (b) SDS-PAGE gel of proteins secreted by A. niger in these cultures.


106

Production of oligo-saccharides.

To obtain broad structural variety, (arabino)xylan oligo-saccharides were generated from

multiple sources. The xylan-rich hemicellulose fraction isolated from wheat straw was

digested using mild acid hydrolysis, whereas commercially obtained wheat arabinoxylan was

enzymatically hydrolysed. The obtained oligo-saccharide mixtures (referred to as fractions

Hn and AXn respectively, where n indicates the fraction number) were subjected to size

exclusion chromatography to remove monomers and low molecular weight oligo-

saccharides, as well as the remaining poly-saccharide. Separation and visualisation of the

soluble hemicellulose-derived hydrolysis products (H) on TLC, (Supplementary Figure

S4.1a), showed a range of oligo-saccharides as well as monomers. MALDI-ToF MS indicated

a series of pentose-oligo-saccharides with a degree of polymerisation (DP) of up to at least

15 (Supplementary Figure S4.1). TLC analysis of arabinoxylan derived oligo-saccharides

(AX)(Supplementary Figure S4.1b) showed generation of an oligo-saccharide series, and

MALDI-ToF MS identified masses consistent with pentose oligo-saccharides with a range

of DPs from 5 to well over 20 (Supplementary Figure S4.1). Oligo-saccharide fractions were

selected that contained mainly oligo-saccharides with a DP of 5-15 (fractions H5-H9) and 5-

20 (fractions AX10-AX15).

The identity and structure of obtained oligo-saccharides was further analysed with

NMR. Analysis of fraction H8 showed that it contained, as main component, a linear

β-(1,4)-Xylp oligo-saccharides with an average DP of ~ 7, as derived from chemical

shifts observed in the 1H-1H TOCSY NMR spectrum of fraction H8 (Supplementary

Table S4.3). The 1D 1H NMR and 2D 1H-13C HSQC NMR chemical shifts

(Supplementary Table S4.4) of fractions H6, H7, H8 also corresponded with oligo-

saccharides with a linear β-1,4-Xylp backbone. Based on chemical shifts reported for similar

structures26, the oligo-saccharides contained 4-O-methyl-glucuronic acid (MeGlcAp) and


107

glucuronic acid (GlcAp) decorations (Figure 4.3). The H-fractions varied in backbone DP

and number of decorations. Integration of the 1D 1H NMR signals (Supplementary Table

S4.4) indicated that the oligo-saccharides have an average backbone DP of 10.4, 8.3 and 7.9

for H6, H7 and H8 respectively, with an average of 1.1, 0.8 and 0.6 decorations per oligo-

saccharide.


108

Figure 4.3: NMR analysis of oligo-saccharides. (a) 1D 1H spectrum of arabinoxylan fraction

AX14, (b) 1D 1H spectrum of wheat hemicellulose fraction H8 and (c) HSQC spectrum

showing the anomeric reporter resonances of fraction H8 (red), overlaid with that of fraction

AX14 (black). The assigned carbohydrate monomers and linkage types are indicated.


109

1D 1H NMR and 2D 1H-13C HSQC NMR chemical shifts of fractions AX14 and AX15,

reported in Supplementary Table S4.4, indicated that these samples also contain oligo-

saccharides with a linear β-1,4-Xylp backbone (Figure 4.3). No (Me)GlcAp decorations were

observed, but the oligo-saccharides contained one or more xylose residues with a single or

double α-Araf substitutions, as indicated by comparison of chemical shifts with those

reported for similar structures27. Integration of the 1D 1H NMR signals indicated that the

oligo-saccharides have an average xylose backbone DP of 4.9 and 4.3 for AX14 and AX15

respectively. The ratio of oligo-saccharides with single and double α-Araf substitutions

differed slightly between fraction AX14 and AX15. On average, AX14 had 0.9 single or 0.4

double Araf substituted xylose residues respectively and AX15 had 0.7 single or 0.3 double

Araf substituted xylose residues.

Carbohydrate array production.

A hydrophobic self-assembling monolayer (SAM) on a gold plate was used as a scaffold

for the immobilization of various labelled carbohydrates via hydrophobic interaction. A small

library of oligo-saccharides activated for attachment to the hydrophobic SAM was generated,

by linking glycans covalently to a hydrophobic hexadecylamine tail using reductive amination.

Using a classical method with 90% (v/v) DMSO, 10% (v/v) acetic acid28, commercially

obtained cellotetraose, cellohexaose and xylohexaose were readily attached to

hexadecylamine. However, as this procedure failed with isolated oligo-saccharides from

fraction AX (AX10, AX11, AX12, AX15) or fraction H (H5, H6, H7, H9), we developed an

alternative method, largely inspired by the work of Gildersleeve et al.25, that successfully

labelled oligo-saccharides, as identified with MALDI-ToF MS. The unlabelled oligo-

saccharides were observed using DHB or THAP as matrix, while the labelled oligo-

saccharides can be analysed solely with CHCA. Carbohydrates labelled with hexadecylamine

were immobilized on a gold plate that was functionalized with a hydrophobic SAM of 1-

undecanethiol. After incubation, the plate was washed with water to remove non-


110

immobilised carbohydrates, as well as the NaBH3CN from the reductive amination step,

which may interfere with MS measurements and enzymatic activity.

All DP oligo-saccharide fragments were retained upon labelling of AX and H fractions

(Supplementary Figure S4.2). Some discrepancies in relative intensity are observed between

unlabelled and labelled (for example DP6 is relatively more intense in the labelled fraction

AX10). This may be the result of alterations in the ionisation efficiency of the two systems,

differences in solubility as well as preferential labelling of certain oligo-saccharides depending

on their degree of polymerization and the presence of substitutions25,29-32. The signal intensity

ratios between peaks during reductive amination reactions with an excess of hexadecylamine

(Supplementary Table S4.6, procedure a vs b), indicated that preferential labelling contributed

to the observed effect, as over time low DP oligo-saccharides were initially labelled while the

final signal intensity ratios were more in agreement with the unlabelled oligo-saccharides.

Influence of degradation of high DP oligo-saccharides was excluded as no increase in signals

corresponding to DP1-4 are observed after labelling and monitoring of the reductive

amination reaction of fraction AX10 for up to 3 days showed that the signal intensity ratios

between peaks were stable from 4 h up 3 days (Supplementary Table S4.6, procedure a).

Conveniently, after immobilization onto the SAM, both low and high DP labelled oligo-

saccharides can still be measured by MS, albeit for high DP oligo-saccharides with a much

reduced signal (Supplementary Figure S4.2). This may be a result of a reduced ionisation

efficiency for these non-covalently immobilised systems, or preferential attachment of low

DP oligo-saccharides. However, we show here that higher DP oligo-saccharides are

successfully immobilized on this array platform, with products of well over DP10 detected

in H9 and AX10 fractions with considerable signal intensity. Furthermore, comparison of

peak ratios of oligo-saccharides on arrays generated on separate days showed that interday

reproducibility was excellent, with a variation in peak ratios below 4% relative standard

deviation (Supplementary Table S4.6).


111

To estimate the surface capacity to bind labelled oligo-saccharide and the effect of oligo-

saccharide concentration, oligo-saccharides were applied on the plate in a range of

concentrations from 3.1 mg mL-1 to 0.2 mg mL-1 in water:methanol (4:1, v/v). Regardless of

the dilution, the spectra obtained were similar in all respects, suggesting that a 0.2 mg mL-1

concentration is sufficient for efficient reproducible array formation. This allows for 5000

assays per mg carbohydrate and equals application of 0.1-0.2 nmol carbohydrate.

The successful attachment of a range of commercial oligo-saccharides as well as oligo-

saccharides in AX- and H-fractions shows that not only oligo-saccharides with a low DP can

be attached, but also linear and decorated oligo-saccharides of considerably higher DP. Such

longer oligo-saccharides (mainly DP7-DP10 range) are rarely available commercially, and

often not tested for carbohydrate array platforms18,33. However, these substrates are essential

for adequate detection and characterisation of endo-acting enzymes since these enzymes

often have extensive substrate binding clefts in which multiple binding sites contribute to

substrate and product specificity34.

Carbohydrate arrays are suitable to detect enzyme activity.

In order to establish whether these carbohydrate arrays can be used as a method for the

detection and identification of products resulting from enzyme activity, carbohydrate arrays

were incubated with exo- and endo-acting enzymes and, after washing to remove sample and

non-attached degradation products, reaction products were identified by MALDI-ToF MS.

Exo-acting activity of β-glucosidase cleaves terminal glucose residues from cellulose oligo-

saccharides. Incubation of β-glucosidase on cellotetraose on carbohydrate arrays resulted, as

expected, in a reaction product with masses corresponding to hexadecylamide-labeled

cellotriose (Figure 4.4). Endoxylanase from Thermomyces lanuginosus acting on xylohexaose

containing carbohydrate arrays resulted in products with masses corresponding to

hexadecylamide labelled xylopentaose, xylotetraose and xylotriose (Figure 4.4). The

minimum xylanase enzyme concentration resulting in activity that could reliably be detected


112

on the arrays with the xylohexaose substrate was 2.5 ng µL-1, equivalent to application of ≥

5 µU of activity per reaction.

Figure 4.4: MALDI-ToF MS analysis of commercial enzyme activity on carbohydrate arrays.

Spectra showing reaction products of incubation of (A) β-glucosidase on cellotetraose-

containing arrays, with DP4 + Na 914.5 m/z, DP4 + H 892.5 m/z, DP3 + Na 752.4 m/z,

DP3 + H 730.5 m/z, (B) endoxylanase on xylohexaose-containing arrays with DP6 + H

1036.7 m/z, DP5 + H 904.6 m/z, DP4 + H 772.6 m/z, DP3 + H 640.6 m/z. Reaction

conditions were chosen such that incomplete substrate degradation allowed identification of

both substrates and products in one mass spectrum. Intensities of peak resulting from

background substrate degradation or impurities were ≤2.2% (Supplementary, Figure S4.3).


113

The identification of these reaction products shows that the application of carbohydrate

arrays that display enzyme substrates, combined with label free detection by MALDI-ToF

MS, is suitable for the detection of CAZyme reaction products and that both activity resulting

from endo- and exo-acting enzymes can be detected.

Carbohydrate arrays are suitable to detect enzyme activity in biological samples.

Culture filtrates of the A. niger grown on wheat straw are complex with regard to

composition, they contain proteins, organic acids, carbohydrates remaining from the

lignocellulose substrate as well as unknown components including fungal metabolites. To

test the suitability of carbohydrate arrays to identify products from enzyme activity in these

samples, culture filtrates were incubated with carbohydrate arrays containing commercial

xylohexaose and cellotetraose, as well as the arabinoxylan and hemicellulose derived oligo-

saccharides.

Wheat straw culture filtrate incubated on cellotetraose containing arrays resulted in the

generation of a masses corresponding to hexadecylamide-labeled cellobiose (Figure 4.5).

These results are indicative of activity of non-reducing end acting cellobiohydrolases, which

cleave cellobiose from the non-reducing substrate end, as well as possible β-glucosidase or

endo-cellulase activity. These enzymes were also identified by proteomics to be present in

the culture filtrate as highly abundant proteins. Incubation of wheat straw culture filtrate on

immobilised xylohexaose resulted mainly in products with masses corresponding to labelled

xylotetraose and xylotriose, with minor amounts of xylopentaose (Figure 4.5). MS-MS

confirmed the identity of these products. These results are indicative of activity of enzyme(s)

that sequentially cleaves xylose, preferentially cleaves either xylobiose or xylotriose from the

non-reducing end of the substrate, or by random hydrolysis of the substrate. Based on the

proteomics results, these activities are likely displayed by XynA, XynB and/or XlnD that

were identified in the culture filtrate; these enzymes have endo-xylanase and β-xylosidase

activities35-37.


114

Figure 4.5: MALDI-ToF MS analysis of A. niger enzyme activity on carbohydrate arrays.

Spectra showing reaction products of incubation of enzymes from A. niger wheat straw

culture filtrate on (A) cellotetraose-containing arrays, with DP4 + H 893.1 m/z,

DP2 + H 568.8 m/z, (B) xylohexaose-containing arrays, with DP6 + H 1036.5 m/z,

DP5 + H 904.6 m/z, DP4 + H 772.5 m/z, DP3 + H 640.5 m/z.

We studied the enzyme activity of a culture filtrates of A. niger on the high DP oligo-

saccharides of the AX and H fractions under several parameters including the reaction time

(30 min, 2 h or 4 h) and pH (range from pH 3 to pH 7). On both AX and H fractions,

enzyme activity in the wheat straw culture filtrates resulted in increased signal intensities

corresponding to DP2, DP3 and often also DP4, while peaks from DP5 up to DP11

decreased, and those higher than DP12 were not identified (Figure 4.6, Figure 4.7). Activity

could be expressed using the signal intensity ratio DP2/DP5, DP3/DP5 and, to a lesser

extent, DP2/DP3.


115

Figure 4.6: MALDI-ToF MS analysis of effect of pH on activity of A. niger enzymes on

fraction AX10. MALDI-ToF MS spectra before and after incubation with enzymes from A.

niger wheat straw culture filtrate on carbohydrate arrays containing fraction AX10, at 30 °C

for 4 h using various pH conditions (from pH 3 to pH 7).


116

Figure 4.7: MALDI-ToF MS analysis of effect of pH on activity of A. niger enzymes on

fraction H7. Maldi-ToF MS spectra before and after incubation with enzymes from A.

niger wheat straw culture filtrate on carbohydrate arrays containing fraction H7, at 30 °C for

4 h using various pH conditions (from pH 3 to pH 7).

During incubation of the culture filtrates of A. niger at pH 4 and 30°C, variation of the

time of incubation from 0 h to 2 h to 4 h indicated that the longer reaction time decreased


117

the amount of DP5 significantly, and increased the ratios DP2/DP5 and DP3/DP5 from

1.1 to 2.7 to 7.7 (± 0.0) and from 0.3 to 0.7 to 1.0 (± 0.1), respectively for AX10. These ratios

increased for H7 from 0.9 to 1.7 to 4.3 (± 0.1) and from 1 to 1.8 to 2.2 (± 0.2). Longer

incubation did not enhance the activity of the enzymes as the same ratios were obtained after

4 h and 8 h (data not shown).

The effect of the pH on wheat straw culture filtrate enzyme activity on the oligo-

saccharides of fractions AX and H fractions was tested in incubations at 30°C for 4 h. For

fraction AX10 (Figure 4.6), degradation of high DP oligo-saccharides (DP4 upward) was

observed at all pH values, and main products were DP2 and DP3. The ratios DP2/DP5 and

DP2/DP3 were found highest at pH 4 (6.2 and 5.8 respectively) (Table 4.1) and lowest at

pH 7. Interestingly, the signal corresponding to DP3 and DP4 was found to be most

abundant at pH 6 and pH 7, indicating accumulation of these oligo-saccharides under these

conditions.


118

Table 4.1: Ratio of the intensities of signal calculated for DP2/DP5, DP3/DP5, DP2/DP3

concerning fractions AX10 and H7 after incubation with A. niger enzymes at 30°C for 4 h

using various pH conditions (from pH 3 to pH 7). Values are given as mean ± standard

deviation, n=2.

DP2/DP5 DP3/DP5 DP2/DP3

AX10 H7 AX10 H7 AX10 H7 pH 3 2.2 ± 0.1 3.2 ± 0.0 0.5 ± 0.0 3.4 ± 0.3 4.3 ± 0.1 0.9 ± 0.0 pH 4 7.7 ± 0.0 4.4 ± 0.1 1.0 ± 0.1 2.4 ± 0.2 7.8 ± 0.1 1.8 ± 0.1 pH 5 2.2 ± 0.1 2.1 ± 0.0 0.5 ± 0.0 1.7 ± 0.1 4.3 ± 0.1 1.2 ± 0.1 pH 6 4.6 ± 0.1 2.7 ± 0.0 1.4 ± 0.1 4.9 ± 0.0 3.2 ± 0.1 0.5 ± 0.0 pH 7 2.1 ± 0.1 1.9 ± 0.1 2.1 ± 0.0 4.2 ± 0.1 1.1 ± 0.1 0.5 ± 0.1

No enzyme (starting material) 1.1 ± 0.0 0.7 ± 0.3 0.3 ± 0.0 0.8 ± 0.2 3.4 ± 0.1 0.8 ± 0.1

Regarding activity of the enzymes on fraction H7 (Fig. 7), the DP2 to DP3 product ratio

was >1 solely at pH 4 and pH 5 (Table 4.1). In addition, as described for the fraction AX10,

the DP2 to DP3 ratio was lower at pH 6 and pH 7 (0.5) than for the starting oligo-saccharide

(0.8), reflecting an accumulation of tri-saccharide.

A. niger seemed able to degrade the higher DP oligo-saccharides and produce low DP

oligo-saccharides both in acid and in neutral condition (from pH 3 to pH 7) in both AX and

H oligo-saccharide fractions. However, pH dependent activity was observed since the

enzyme mixture was less active at pH 6-7, resulting in an accumulation of tri-saccharide

(DP3) and tetra-saccharide (DP4). These results have clearly indicated that the overall

optimum pH for this mixture of enzymes on these substrates, using our platform, was at

pH 4.


119

Discussion

Through this study, we have confirmed that the carbohydrate microarray-based

technology combined with MALDI-ToF MS is a powerful strategy for the assessment of

enzyme activity in many simultaneous experiments using small amounts of valuable

carbohydrates. Linear as well as decorated oligo-saccharides, with both low and high DP, can

be efficiently immobilised on the arrays, after which their enzymatic degradation can be

monitored by MALDI-ToF MS.

Both covalent and non-covalent attachment of carbohydrates on arrays has been

described15-17. Some of these platforms require chemical modification via several synthetic

steps to generate carbohydrates suitable for attachment, which is readily performed with

monomers but less straight forward for oligo-saccharides. To enable the generation of arrays

with a range of oligo-saccharides, which are generally available in small quantities, we selected

an attachment strategy that requires minimal synthetic modification; the attachment of a

hydrophobic anchor via reductive amination, which can immobilise the attached

carbohydrate on a hydrophobic SAM. A relatively short hydrophobic anchor (C12)

immobilised poorly in our hands during initial test, while longer (C18, double C16) anchors,

as described18, provided effective immobilisation, resulting in the employment of the C16

hexadecylamine immobilisation anchor described here.

Reductive amination of oligo-saccharide has been widely used in the past for several

applications including conjugation of carbohydrate to proteins or other scaffolds in order to

produce molecules of pharmaceutical utility, as well as derivatization of biological sugars

affording better detection and analysis by liquid chromatography and/or mass

spectrometry38. The procedure generally employed for reductive amination of free oligo-

saccharide was described by Roy et al39 in 1984, using sugar in the presence of NaBH3CN

and aqueous sodium borate buffer for 24 h at 37-50°C. The alternative protocol developed


120

by the group of Gildersleeve25 is rather similar, except for the use of salt additives such as

sodium sulfate. The authors have suggested that high salt concentrations could make the

removal of water molecule easier, and hence, might favour formation of the imine. We

explored the role of the sodium sulfate buffer by performing reductive amination with or

without this salt additives; the presence of sodium sulfate buffer accelerates the rate of the

reaction.

Purification of labelled oligo-saccharides, which is time consuming and leads to a loss of

material, can be avoidable following our approach. As an example, with solely 0.4 mg of

oligo-saccharide starting material, we could perform up to 2000 array experiments. However,

without such purification no estimate can be obtained of the yield of the labelling reaction,

making it impossible to compare yield between different oligo-saccharide fractions.

However, by comparing the signal intensity obtained during the MS analysis for each

compound in the AX and H fractions over time, we showed that oligo-saccharides with a

low DP were labelled more effectively than those with the highest DP, which seems to be

remedied by addition of amine.

The identification of reaction products from degradation of plant-biomass derived oligo-

saccharides in complex samples as these fungal culture filtrates, as well of that of commercial

enzyme preparations, shows that the hydrophobic carbohydrate array coupled with MS was

a platform that was well suited for measurement of such enzyme activities. The enzyme

activities that were detected using the linear commercial oligo-saccharides confirmed that the

platform is suitable for detection of both exo- and endo-acting enzymes. Analysing enzyme

activity on a single substrate, as performed here, gives a good indication of the presence or

absence of degradative activities. It however has limited value in deriving a mechanism of

action for an enzyme because the activity of a single enzyme during substrate degradation

can result in multiple products: to study the substrate/product profile of a single enzyme in


121

detail, separate incubations of the enzyme with a set of defined substrates of different lengths

will be required.

For example, β-glucosidase hydrolyses the terminal residue of cellulose oligo-saccharides,

and degradation of substrate cellotetraose can thus theoretically result in cellotriose, and after

another hydrolysis event on the same molecule, also in cellobiose. We did not observe the

formation of cellobiose under our reaction conditions, which – besides steric effects - may

point towards cellotetraose being the minimum length substrate for this particular enzyme,

or, more likely may reflect a higher turnover rate for cellotetraose vs cellotriose. Similarly,

different fits of substrates in the active cleft of the endo-acting enzyme may result in a range

of different products, which may or may not be degraded further dependent on the minimum

substrate length required by the enzyme and any differences in kinetics that it displays for

these different substrates. This explains the range of products resulting from endo-xylanase

activity observed in this study.

Using mixtures of such plant derived oligo-saccharides with high DP, i.e. AX and H

fractions we demonstrated that the optimum conditions required to observe an activity of

the culture filtrates of A. niger on arabinoxylan or xylan oligo-saccharides was an incubation

at pH 4 for 4 h at 30°C. At higher pH, accumulation of DP3 and DP4 oligo-saccharides was

observed. For the oligo-saccharides from the AX fractions, this could be due to reduced

activity of α-arabinofuranosidases such as AbfB and AxhA that removes the arabinose

decorations from the xylose backbone. However, since the same phenomenon was also

observed for the H-fraction oligo-saccharides, it is more likely due to a reduction in an exo-

acting β-xylosidase that degrades the xylose backbone of DP4 and DP3 to DP2. The

proteomics results indicated that such an enzyme, XlnD, was indeed present in the culture

filtrates of A. niger grown on wheat straw.

Some divergence existed of the activity of A. niger enzymes in the culture filtrate when

degrading oligo-saccharides in the H7 and AX10 fractions; the latter seems to be a better


122

substrate for the enzyme mixture as the reduction in maximum observed DP was much

greater for AX (from DP20 to DP9) than for H (from DP14 to DP11). It is surprising that

the AX fraction seems to be degraded more efficiently, since the structure is more complex

(i.e. decorated with arabinose) compared to H (linear xylose series). However, the complex

structure also means that a higher number of enzymes that are all highly abundant present in

the culture filtrate can act on the oligo-saccharide; not only the XynA GH10 and XynB

GH11 endo-xylanases but also the arabinofuranosidases AfbB and AxhA. Furthermore,

GH10 xylanases as XynA - which is the most abundant protein identified in the culture

filtrate - are reported to more efficiently degrade arabinose-decorated xylose oligo-saccharide

compared to undecorated oligo-saccharides34, thus possibly contributing to quicker turnover

of AX than H oligo-saccharide series. Alternatively, steric interference may contribute to the

observed differences in degradation between the H and AX fractions, as close proximity of

substrate molecules on the SAM may limit their availability to enzymatic modification or

degradation. This underlines the need to immobilise a minimal amount of substrate on the

SAM to limit such interferences.

No signal corresponding to DP1 was observed in all the conditions assayed, probably

because the reductive amination yielded an unnatural ring-opened product, which could be

not recognized as a substrate by the enzyme and led to an accumulation of DP2. This also

suggested that A. niger enzymes need a closed-ring hemiacetal mono-saccharide as backbone

in order to be active.

In conclusion, we reported on the production of a carbohydrate array with plant polymer-

derived oligo-saccharides, and its application to detect activity of carbohydrate active

enzymes in commercial preparations as well as biological samples. To our knowledge, this is

the first report describing the use of a carbohydrate array with high DP oligo-saccharide

substrates in combination with MALDI-ToF MS. By using mass spectrometry instead of

labelled antibodies in combination with the arrays, we made a significant advancement in


123

detection of the substrate alterations that result from enzyme activities. This opens up the

possibility of distinguishing different enzyme activities, such as exo- or endo-active enzymes,

based on their product profiles. It also offers the opportunity to detect and distinguish

between enzyme activities with different mechanisms to hydrolyse the substrate, such as

LMPOs, lyases and endo-acting hydrolyses, as well as potential new mechanisms. This

platform could form a basis for on-chip identification of complex enzyme products in

combination with a recently developed carbohydrate sequencing method based on ion

mobility-mass spectrometry (IM-MS)40-42.39

References

[1] Klein-Marcuschamer, D., Oleskowicz-Popiel, P., Simmons, B. A. & Blanch, H. W. The

challenge of enzyme cost in the production of lignocellulosic biofuels. Biotechnol Bioeng

109, 1083-1087 (2012).

[2] Chundawat, S. P., Beckham, G. T., Himmel, M. E. & Dale, B. E. Deconstruction of

lignocellulosic biomass to fuels and chemicals. Annu Rev Chem Biomol Eng 2, 121-145

(2011).

[3] Lombard, V., Golaconda Ramulu, H., Drula, E., Coutinho, P. M. & Henrissat, B. The

carbohydrate-active enzymes database (CAZy) in 2013. Nucleic Acids Res 42, D490-495

(2014).

[4] Gilbert, H. J. The biochemistry and structural biology of plant cell wall deconstruction.

Plant Physiol 153, 444-455 (2010).

[5] Kubicek, C. P., Mikus, M., Schuster, A., Schmoll, M. & Seiboth, B. Metabolic

engineering strategies for the improvement of cellulase production by Hypocrea jecorina.

Biotechnol Biofuels 2, 19 (2009).

[6] Znameroski, E. A. et al. Induction of lignocellulose-degrading enzymes in Neurospora

crassa by cellodextrins. Proc Natl Acad Sci U S A 109, 6012-6017 (2012).

[7] Ward, O. P., Qin, W. M., Dhanjoon, J., Ye, J. & Singh, A. Physiology and

Biotechnology of Aspergillus. Adv Appl Microbiol 58C, 1-75 (2005).

[8] Andersen, M. R., Giese, M., de Vries, R. P. & Nielsen, J. Mapping the polysaccharide

degradation potential of Aspergillus niger. BMC Genomics 13, 313 (2012).


124

[9] Delmas, S. et al. Uncovering the genome-wide transcriptional responses of the

filamentous fungus Aspergillus niger to lignocellulose using RNA sequencing. PLoS

Genet. 8, e1002875 (2012).

[10] Pullan, S. et al. RNA-sequencing reveals the complexities of the transcriptional

response to lignocellulosic biofuel substrates in Aspergillus niger. Fungal Biol Biotechnol 1,

3 (2014).

[11] de Souza, W. R. et al. Transcriptome analysis of Aspergillus niger grown on sugarcane

bagasse. Biotechnol Biofuels 4, 40 (2011).

[12] Pedersen, H. L. et al. Versatile high resolution oligosaccharide microarrays for plant

glycobiology and cell wall research. J Biol Chem 287, 39429-39438 (2012).

[13] Agger, J. W. et al. Discovery of LPMO activity on hemicelluloses shows the importance

of oxidative processes in plant cell wall degradation. Proc Natl Acad Sci U S A 111, 6287-

6292 (2014).

[14] Rillahan, C. D. & Paulson, J. C. Glycan microarrays for decoding the glycome. Annu

Rev Biochem 80, 797-823 (2011).

[15] Park, S., Gildersleeve, J. C., Blixt, O. & Shin, I. Carbohydrate microarrays. Chem Soc

Rev 42, 4310-4326 (2013).

[16] Donczo, B., Kerekgyarto, J., Szurmai, Z. & Guttman, A. Glycan microarrays: new

angles and new strategies. Analyst 139, 2650-2657 (2014).

[17] Gray, C. J., Weissenborn, M. J., Eyers, C. E. & Flitsch, S. L. Enzymatic reactions on

immobilised substrates. Chem Soc Rev 42, 6378-6405 (2013).

[18] Beloqui, A., Sanchez-Ruiz, A., Martin-Lomas, M. & Reichardt, N. C. A surface-based

mass spectrometry method for screening glycosidase specificity in environmental

samples. Chem Commun (Camb) 48, 1701-1703 (2012).

[19] van Hartingsveldt, W., Mattern, I. E., van Zeijl, C. M., Pouwels, P. H. & van den

Hondel, C. A. Development of a homologous transformation system for Aspergillus

niger based on the pyrG gene. Mol Gen Genet 206, 71-75 (1987).

[20] Bohle, K. et al. Selection of reference genes for normalisation of specific gene

quantification data of Aspergillus niger. J Biotechnol 132, 353-358 (2007).

[21] Yan, J. X. et al. A modified silver staining protocol for visualization of proteins

compatible with matrix-assisted laser desorption/ionization and electrospray

ionization-mass spectrometry. Electrophoresis 21, 3666-3672 (2000).

[22] Matia-Gonzalez, A. M., Laing, E. E., Gerber, A. P. Conserved mRNA-binding

proteomes in eukaryotic organisms. Nat Struct Mol Biol 22, 1027-1033 (2015).


125

[23] Sun, R. C. & Tomkinson, J. Characterization of hemicelluloses obtained by classical

and ultrasonically assisted extractions from wheat straw. Carbohyd polym 50, 263-271

(2002).

[24] Kormelink, F. J. et al. Characterisation by 1H NMR spectroscopy of oligosaccharides

derived from alkali-extractable wheat-flour arabinoxylan by digestion with endo-(1--

>4)-beta-D-xylanase III from Aspergillus awamori. Carbohydr Res 249, 369-382 (1993).

[25] Gildersleeve, J. C., Oyelaran, O., Simpson, J. T. & Allred, B. Improved procedure for

direct coupling of carbohydrates to proteins via reductive amination. Bioconjug Chem 19,

1485-1490 (2008).

[26] Pena, M. J. et al. Arabidopsis irregular xylem8 and irregular xylem9: implications for

the complexity of glucuronoxylan biosynthesis. Plant Cell 19, 549-563 (2007).

[27] Vliegenthart, J. F. G. H., R.A.; Kamerling, J.P. A 1H-NMR spectroscopic study on

oligosaccharides obtained from wheat arabinoxylans, in Xylans and Xylanases (ed J.

Visser) 17-37 (Elsevier Science Publishers B.V., 1992).

[28] Bigge, J. C. et al. Nonselective and efficient fluorescent labeling of glycans using 2-

amino benzamide and anthranilic acid. Anal Biochem 230, 229-238 (1995).

[29] Stowell, S. R. et al. Human galectin-1 recognition of poly-N-acetyllactosamine and

chimeric polysaccharides. Glycobiology 14, 157-167 (2004).

[30] Beaudoin, M. E., Gauthier, J., Boucher, I. & Waldron, K. C. Capillary electrophoresis

separation of a mixture of chitin and chitosan oligosaccharides derivatized using a

modified fluorophore conjugation procedure. J Sep Sci 28, 1390-1398 (2005).

[31] Evangelista, R. A. C., F.A.; Guttman, A. Reductive amination of N-linked

oligosaccharides using organic acid catalysts. J. Chromatogr. A 745, 273-280 (1996).

[32] Rustighi, I. et al. Analysis of N-acetylaminosugars by CE: a comparative derivatization

study. Electrophoresis 30, 2632-2639 (2009).

[33] Ban, L. et al. Discovery of glycosyltransferases using carbohydrate arrays and mass

spectrometry. Nat Chem Biol 8, 769-773 (2012).

[34] Pollet, A., Delcour, J. A. & Courtin, C. M. Structural determinants of the substrate

specificities of xylanases from different glycoside hydrolase families. Crit Rev Biotechnol

30, 176-191 (2010).

[35] Do, T. T, Quyen, D. T.,Nguyen, T.N., Nguyen, V.T., Molecular characterization of a

glycosyl hydrolase family 10 xylanase from Aspergillus niger. Protein Expr Purif 92, 196-

202 (2013).


126

[36] Selig, M. J. et al. Heterologous expression of Aspergillus niger beta-D-xylosidase (XlnD):

characterization on lignocellulosic substrates. Appl Biochem Biotechnol 146, 57-68 (2008).

[37] Ruanglek, V. et al. Cloning, expression, characterization and high cell-density

production of recombinant endo-1,4-β-xylanase from Aspergillus niger in Pichia pastoris.

Enzyme Microb Technol. 41, 19-25 (2007).

[38] Ruhaak, L. R. et al. Glycan labeling strategies and their use in identification and

quantification. Anal Bioanal Chem 397, 3457-3481 (2010).

[39] Roy, R., Katzenellenbogen, E. & Jennings, H. J. Improved procedures for the

conjugation of oligosaccharides to protein by reductive amination. Can J Biochem Cell

Biol 62, 270-275 (1984).

[40] Both, P. et al. Discrimination of epimeric glycans and glycopeptides using IM-MS and

its potential for carbohydrate sequencing. Nat Chem 6, 65-74 (2014).

[41] Lanucara, F., Holman, S. W., Gray, C. J. & Eyers, C. E. The power of ion mobility-

mass spectrometry for structural characterization and the study of conformational

dynamics. Nat Chem 6, 281-294 (2014).

[42] Gray, C. J. et al. Applications of ion mobility mass spectrometry for high throughput,

high resolution glycan analysis. Biochim Biophys Acta 1860, 1688-1709 (2016).

[43] van Munster, J. M. et al. The role of carbon starvation in the induction of enzymes that

degrade plant-derived carbohydrates in Aspergillus niger. Fungal Genet Biol (2014).


127

Acknowledgements

We acknowledge the Proteomics Facility of the University of Bristol, for their excellent

technical support of the protein identifications via their proteomics service. We thank Kevin

Butler and Mick Cooper, School of Chemistry, University of Nottingham, for acquiring the

NMR spectra and for training on the MALDI-ToF MS equipment, respectively. This study

was supported by the Biotechnology and Biological Sciences Research Council (BBSRC)

(BB/G01616X/1 and BB/K01434X/1), the Sustainable Chemical and Biological Processing

Priority Group, University of Nottingham, the IBCatalyst Glycoenzymes for bioindustries

(BB/M02903411), IBCatalyst Chemo-enzymatic production of speciality glycans

(BB/M028836/1) and EU FP7 ITN training in neurodegeneration therapeutics, intervention

and neurorepair 'TINTIN' (608381).

Author contributions

JvM, BT, AD, MR, CJG, DA and SF designed the experiments, JvM, BT, AD and MR

performed the experiments, JvM, BT, AD, MR, DA and SF analysed the data, JvM and BT

wrote the manuscript. All authors reviewed the manuscript.

Additional information

Competing financial interests: The authors declare no competing financial interests.


128

4.4 Supporting Information

Application of carbohydrate arrays coupled with mass

spectrometry to detect activity of plant-polysaccharide

degradative enzymes from the fungus Aspergillus niger

Jolanda M van Munster1*, Baptiste Thomas2, Michel Riese2, Adrienne L Davis3, Christopher

J Gray2, David B Archer1, Sabine L Flitsch2


129

Figure S4.1: Thin layer chromatography of a oligo-saccharides generated by acid hydrolysis

of wheat hemicelluloses and b of oligo-saccharides generated by enzymatic hydrolysis of

wheat arabinoxylan, with relevant fractions labelled below the sample application site

(numbers H1-H11 and AX5-AX18) and M indicating the H and AX mixture of oligo-

saccharides before size separation, c MALDI-ToF MS spectra with indicated the DP (DPx)

and spacing of the main pentose oligo-saccharide series and spacing of a minor secondary

series consisting of 4Me-GLcA decorated xylose oligo-saccharides for oligo-saccharides

generated by acid hydrolysis of wheat hemicelluloses and d for oligo-saccharides generated

by enzymatic hydrolysis of wheat arabinoxylan. All labelled peaks are sodium adducts.

H1 H2 H3 H4 H5 H6 H7 H8 H9 H10 H11 M

b

AX5 AX6 AX7 AX8 AX9 AX10 AX11 AX12 AX13 AX14 AX15 AX16 AX17 AX18 M

a


130

Figure S4.1 continued: MALDI-ToF MS spectra with indicated the DP (DPx) and spacing

of the main pentose oligo-saccharide series and spacing of a minor secondary series

consisting of 4Me-GLcA decorated xylose oligo-saccharides for oligo-saccharides generated

by acid hydrolysis of wheat hemicelluloses and d for oligo-saccharides generated by

enzymatic hydrolysis of wheat arabinoxylan. All labelled peaks are sodium adducts.


131

Figure S4.1 continued: MALDI-ToF MS spectra with indicated the DP (DPx) and spacing

of the main pentose oligo-saccharide series and spacing of a minor secondary series

consisting of 4Me-GLcA decorated xylose oligo-saccharides for oligo-saccharides generated

by acid hydrolysis of wheat hemicelluloses and d for oligo-saccharides generated by

enzymatic hydrolysis of wheat arabinoxylan. All labelled peaks are sodium adducts.


132

Figure S4.2: a MALDI-ToF MS spectra for fraction AX10 with (a1) unlabeled starting oligo-

saccharide, (a2) labelling of free oligo-saccharide by reductive amination and (a3)

immobilization of labelled carbohydrates on the gold plate via hydrophobic interactions b

MALDI-ToF MS spectra for fraction H9 with (b1) unlabeled starting oligo-saccharide, (b2)

labelling of free oligo-saccharide via reductive amination and (b3) immobilization of labelled

carbohydrates on the gold plate via hydrophobic interactions.


133

Figure S4.3: MALDI-ToF MS analysis of signals derived from background hydrolysis

and/or impurities in carbohydrate arrays used with commercial enzyme activity. Spectra

showing substrate incubated with buffer only under reaction conditions of a β-glucosidase

on cellotetraose-containing arrays, with DP4 + Na 914.5 m/z, DP4 + H 892.5 m/z (100%),

DP3 + Na 752.4 m/z, DP3 + H 730.5 m/z (2.2%) while DP3 peak resulting from enzyme

activity (Fig. 4) was 48%. b endoxylanase on xylohexaose-containing arrays with DP6 + H

1036.9 m/z (100%), DP5 + H 904.9 m/z (1.2%), DP4 + H 772.9 m/z (0.6%), DP3 + H

640.8 m/z (2.1%) while peaks resulting from enzyme activity (Fig. 4) were 5.3%, 9.1% and

6.4% respectively.


134

Table S4.1: qRT-PCR primers.

gene gene ID primer name sequence (5'-3') reference

actA An15g00560 An15g00560Fw TCCTGGGTCT GGAGAGCGGT G

43

An15g00560Rv CTGCATACGG TCGGAGATAC CGGG

43

cbhA An07g09330 cbhAFw CCAGCAAGCC GGAACGCTCA 9

cbhARv AACGCGCCGT TTAGCCCACA 9

cbhB An01g11660 cbhBFw CCAGCGATGG CAGCTGCAC 9

cbhBRv CTGCCGGACG TGGTCACACC 9

xynB An01g00780 xynBFw CACGACTCTG TCGCCCAGCG 9

xynBRv GGGGGTCAGT GGTCCAGCCA 9


135

Table S4.2: Shotgun proteomics identification of proteins in the filtrate of A. niger cultures grown in the presence of wheat straw with Score: SEQUEST protein score, #PSMs: total number of identified peptide sequences.

Biol

ogica

l rep

licat

e A

# P

SMs

1437

863

759

601

376

430

247

149

178

101

72

69

90

60

57

48

32

53

32

32

24

23

25

25

19

38

15

16

23

12

# U

niqu

e Pe

ptid

es

38

12

4 13

11

10

10

22

10

29

6 15

9 13

14

8 5 6 6 15

5 6 6 11

8 16

7 6 4 6

Cove

rage

87.7

7

52.4

3

36

61.4

5

35.8

7

34.5

1

44.4

8

34.5

8

31.4

2

36.1

6

36.6

7

36.2

6

22.2

2

34.4

5

23.6

6

24.6

2

25.5

4

19.6

4

30.2

9

21.7

5

19.5

7

17.2

4

24.2

9

19.7

27.8

8

22.4

7

16.8

2

15.5

5

18.5

2

13.2

2

Scor

e

4386

.343

249

3181

.423

578

2421

.212

367

2057

.845

71

1273

.707

949

1214

.966

406

735.

6361

805

403.

6227

429

332.

4203

77

276.

1128

501

284.

6185

062

225.

6573

155

270.

6945

381

141.

2109

748

162.

2628

663

119.

9707

662

101.

3156

013

114.

5235

491

118.

8116

159

71.6

4827

18

64.6

1966

348

58.3

6288

941

88.5

2926

564

50.4

6855

557

49.6

6892

433

80.5

8373

761

50.5

9748

769

54.3

9299

071

45.2

2074

592

29.4

5250

034

Biol

ogica

l rep

licat

e B

# P

SMs

1293

694

715

502

356

328

207

150

169

111

65

64

71

88

54

56

34

49

32

33

25

30

20

24

23

36

17

14

27

15

# U

niqu

e Pe

ptid

es

38

12

5 13

11

10

10

22

10

28

6 14

10

16

13

8 4 9 7 15

5 6 7 10

9 15

8 5 3 6

Cove

rage

87.7

7

52.4

3

36.4

4

61.4

5

35.8

7

34.5

1

44.4

8

34.4

5

28.4

33.3

7

36.6

7

31.9

7

20.9

38.7

3

21.9

6

25.3

8

23.3

7

23.2

5

31.5

9

20.5

6

19.5

7

17.2

4

24.2

9

16.9

7

25.2

4

20.5

7

20.5

2

14.0

1

10.8

9

13.2

2

Scor

e

3836

.167

32

2510

.741

362

2237

.920

687

1629

.407

482

1160

.001

724

988.

6003

529

591.

2075

629

410.

8237

553

332.

6012

958

307.

6644

777

233.

4921

172

213.

7005

343

208.

3747

042

197.

5940

342

163.

5840

549

131.

3712

02

120.

3914

742

107.

8468

797

104.

2714

376

77.2

9581

523

74.7

5479

639

72.0

5174

184

64.5

7549

357

61.8

1282

258

58.9

5624

59

55.7

0176

053

49.5

3932

083

48.0

0241

899

45.8

5378

671

45.3

1327

009

calc.

pI

6.65

4.32

5.45

4.86

4.44

4.30

4.55

4.89

4.55

4.78

5.02

5.05

4.60

5.43

4.74

4.67

5.03

5.08

4.61

5.21

4.49

4.51

4.07

4.83

5.19

5.38

5.47

5.20

4.53

4.30

MW

[kD

a]

35.4

6

48.2

2

24.0

4

35.8

1

52.4

8

56.1

7

30.5

3

87.1

6

36.5

4

93.1

7

28.5

7

55.8

9

57.1

0

59.1

1

82.0

7

41.2

5

19.1

7

48.8

0

41.2

2

109.

64

46.3

2

42.1

0

38.6

7

71.1

8

45.5

0

93.6

7

58.8

2

57.1

7

48.1

0

67.9

1

# A

As

327

452

225

332

499

536

281

804

331

860

270

513

531

537

765

394

184

443

383

1007

460

406

350

660

416

841

541

521

459

628

Des

crip

tion

Prob

able

endo

-1,4

-bet

a -xy

lana

se C

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

xlnC

PE

=1 S

V=

1 - [

XY

NC_

ASP

NC]

Prob

able

1,4 -

beta

-D-g

luca

n ce

llobi

ohyd

rolas

e A

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

cbhA

PE

=3

SV=

1 - [

CBH

A_A

SPN

C]

Prob

able

endo

-1,4

-bet

a -xy

lana

se B

OS=

Asp

ergil

lus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

xlnB

PE

=3

SV=

1 - [

XY

NB_

ASP

NC]

Prob

able

alph

a -L -

arab

inof

uran

osid

ase

axhA

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

axhA

PE

=3 S

V=

1 - [

AX

HA_

ASP

NC]

Prob

able

alph

a-L-

arab

inof

uran

osid

ase

B O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=ab

fB P

E=3

SV

=1

- [A

BFB_

ASP

NC]

Prob

able

1,4-

beta

- D- g

luca

n ce

llobi

ohyd

rolas

e B

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

cbhB

PE

=3

SV=

1 - [

CBH

B_A

SPN

C]

Prob

able

feru

loyl

este

rase

A O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=fa

eA P

E=

3 SV

=1

- [FA

EA_

ASP

NC]

Prob

able

exo-

1,4-

beta

-xyl

osid

ase

xlnD

OS=

Asp

ergil

lus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

xlnD

PE

=3

SV=

1 - [

XY

ND

_ASP

NC]

Prob

able

endo

-bet

a-1,

4-gl

ucan

ase

B O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=eg

lB P

E=3

SV

=1

- [E

GLB

_ASP

NC]

Prob

able

beta

- glu

cosid

ase

A O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=bg

lA P

E=

3 SV

=1

- [BG

LA_A

SPN

C]

Prob

able

feru

loyl

este

rase

C O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=fa

eC P

E=

3 SV

=1

- [FA

EC_

ASP

NC]

Prob

able

man

nosy

l-olig

osac

char

ide

alpha

-1,2

-man

nosid

ase

1B O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=m

ns1B

PE

=3

SV=

1 -

[MN

S1B_

ASP

NC]

Prob

able

rham

noga

lactu

rona

te ly

ase

A O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=rg

lA P

E=

3 SV

=2

- [RG

LA_ A

SPN

C]

Ext

race

llular

exo-

inul

inas

e in

uE O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=in

uE P

E=2

SV

=1

- [IN

UE_

ASP

NC]

Prob

able

beta

-glu

cosid

ase

M O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=bg

lM P

E=3

SV

=1

- [BG

LM_A

SPN

C]

Asp

artic

pro

teas

e pe

p1 O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=pe

p1 P

E=3

SV

=1

- [PE

PA_A

SPN

C]

Cell

wal

l pro

tein

phi

A O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=ph

iA P

E=

2 SV

=1

- [PH

IA_A

SPN

C]

Prob

able

alph

a-ga

lacto

sidas

e B

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

aglB

PE

=3

SV=

1 - [

AG

ALB

_ASP

NC]

Prob

able

man

nan

endo

-1,4

-bet

a-m

anno

sidas

e A

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

man

A P

E=

1 SV

=1

- [M

AN

A_A

SPN

C]

Prob

able

beta

-gala

ctos

idas

e A

OS=

Asp

ergil

lus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

lacA

PE

=3

SV=

1 - [

BGA

LA_A

SPN

C]

Prob

able

gluc

an en

do-1

,3-b

eta-

gluc

osid

ase

eglC

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

eglC

PE

=3

SV=

2 - [

EG

LC_A

SPN

C]

Prob

able

endo

-xylo

galac

turo

nan

hydr

olas

e A

OS=

Asp

ergi

llus n

i ger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

xghA

PE

=3

SV=

1 - [

XG

HA_

ASP

NC]

Prob

able

arab

inog

alact

an e

ndo -

beta

- 1,4

-gala

ctan

ase

A O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=ga

lA P

E=

3 SV

=1

- [G

AN

A_A

SPN

C]

Prob

able

alph

a-ga

lacto

sidas

e D

OS=

Asp

ergil

lus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

aglD

PE

=3 S

V=

2 - [

AG

ALD

_ASP

NC]

Prob

able

gluc

an 1

,3-b

eta-

gluc

osid

ase

A O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=ex

gA P

E=

3 SV

=1 -

[EX

GA

_ASP

NC]

Prob

able

alph

a-gl

ucur

onid

ase

A O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=ag

uA P

E=3

SV

=1

- [A

GU

A_A

SPN

C]

Beta

-glu

curo

nida

se O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=A

n02g

1189

0 PE

=1

SV=

1 - [

GU

S79_

ASP

NC]

Prob

able

feru

loyl

este

rase

B O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=fa

eB P

E=3

SV=

1 - [

FAE

B_A

SPN

C]

Prob

able

1,4-

beta

-D- g

luca

n ce

llobi

ohyd

rolas

e C

OS=

Asp

ergil

lus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

cbhC

PE

=3

SV=

1 - [

CBH

C_A

SPN

C]

Prob

able

alph

a -L -

arab

inof

uran

osid

ase

A O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=abf

A P

E=

3 SV

=1

- [A

BFA

_ASP

NC]

Acc

essio

n

A2Q

FV7

A2Q

PG2

A2Q

7I0

A2Q

FV9

A2R

511

A2Q

AI7

A2Q

SY5

A2Q

A27

A2Q

PC3

A2R

AL4

A2Q

YU

7

A2Q

AS2

A2R

2L1

A2R

0E0

A5A

BF5

A2R

3L3

A2R

2S8

A2Q

EJ9

A2Q

KT4

A2Q

AN

3

A2Q

H21

A2Q

K83

A2R

B93

A2R

2S6

A2R

AR6

A2R

3X3

A2Q

EQ

6

A2R

0Z6

A2Q

YR9

A2Q

7E0


136

Table S4.2 (continued): Shotgun proteomics identification of proteins in the filtrate of A. niger cultures grown in the presence of wheat straw with Score: SEQUEST protein score, #PSMs: total number of identified peptide sequences.

Biol

ogica

l rep

licat

e A

# P

SMs

11

24

10

14

10

NA

6 12

8 15

9 10

8 8 9 12

NA

4 4 3 2 2 3 NA

4

# U

niqu

e Pe

ptid

es

2 5 3 7 7 NA

3 2 6 4 3 5 2 3 3 7 NA

4 2 2 2 2 2 NA

2

Cove

rage

4.67

23.2

4

13.5

6

18.9

5

11.6

NA

11.5

8

14.1

1

5.9

13.6

2

11.9

6

10.1

4

5.83

5.4

8.97

13.1

8

NA

5.91

11.9

4

9.92

6.63

0.88

13.4

8

NA

2.97

Scor

e

37.9

5039

654

59.3

9916

742

22.1

7836

38

25.8

1601

477

22.9

4560

409

NA

20.7

1233

189

21.6

1398

59

15.3

2841

229

28.6

0390

46

22.0

7549

024

22.4

2315

507

24.9

5914

817

18.0

5999

267

14.4

6537

042

22.7

8718

197

NA

6.31

0669

422

8.49

9233

723

10.0

4579

258

4.50

6391

764

4.31

3843

966

4.43

7262

297

NA

0

Biol

ogica

l rep

licat

e B

# P

SMs

14

18

13

18

15

12

9 11

9 14

11

10

7 10

11

10

6 11

4 3 4 3 4 2 4

# U

niqu

e Pe

ptid

es

2 4 5 9 12

4 4 2 6 4 4 5 4 5 5 7 3 7 2 2 3 2 3 2 3

Cove

rage

4.67

19.5

7

15.6

3

22.6

16.5

2

10.6

9

16.5

5

14.1

1

6

13.6

2

13.3

2

10.1

4

15.0

2

11.9

2

16.0

9

13.0

4

9.09

12.1

4

11.9

4

9.92

9.67

0.88

18.4

4

6.71

4.25

Scor

e

42.8

6552

119

42.1

5743

279

39.4

8248

076

39.3

1278

515

24.5

3107

798

24.4

9345

97

22.1

8053

246

21.7

6690

674

21.2

2608

018

19.7

5820

541

19.3

9040

112

18.5

9540

629

16.7

4226

558

16.6

1905

289

16.3

4141

445

14.9

2676

353

14.8

1848

919

14.7

4084

735

8.22

4465

847

7.27

5989

771

7.11

0866

427

5.92

9324

269

4.44

7160

482

4.07

4769

974

2.44

4819

45

calc.

pI

4.22

4.39

4.82

4.86

4.91

4.87

5.20

3.88

5.00

4.82

4.48

4.79

4.23

5.03

4.34

5.66

5.20

5.29

10.5

5

4.88

6.77

4.55

10.1

3

4.91

5.11

MW

[kD

a]

34.4

6

34.6

0

47.2

6

48.3

5

86.5

1

34.0

3

45.8

7

25.4

7

111.

74

34.3

0

38.0

9

46.9

4

46.9

3

59.1

8

39.7

6

82.5

4

47.8

2

103.

96

14.1

4

26.6

7

37.8

0

122.

02

14.9

9

31.9

3

78.3

7

# A

As

321

327

435

438

793

318

423

241

1017

323

368

434

446

537

379

736

440

931

134

262

362

1257

141

298

706

Des

crip

tion

Prob

able

arab

inan

endo

-1,5

-alp

ha- L

-ara

bino

sidas

e A

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=a

bnA

PE=

3 SV

=1

- [A

BNA

_ASP

NC]

Prob

able

pect

ines

tera

se A

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

pmeA

PE=

3 SV

=1

- [PM

EA

_ASP

NC]

Prob

able

exop

olyg

alact

uron

ase

X O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=pg

aX P

E=3

SV

=1

- [PG

LRX

_ASP

NC]

Prob

able

exop

olyg

alact

uron

ase

B O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=pg

xB P

E=

3 SV

=1

- [PG

XB_

ASP

NC]

Prob

able

alph

a-fu

cosid

ase

A O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=af

cA P

E=

3 SV

=1

- [A

FCA

_ASP

NC]

Prob

able

arab

inan

endo

-1,5

- alp

ha- L

-ara

bino

sidas

e C

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=a

bnC

PE=

3 SV

=1

- [A

BNC_

ASP

NC]

Puta

tive

galac

tura

n 1,

4-al

pha -

galac

turo

nida

se C

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

rgxC

PE

=2 S

V=

1 - [

RGX

C_A

SPN

C]

Prob

able

xylo

gluc

an-s

peci

fic e

ndo-

beta

-1,4

-glu

cana

se A

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=x

geA

PE

=3

SV=

1 - [

XG

EA_

ASP

NC]

Prob

able

beta

-gala

ctos

idas

e B

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

lacB

PE=

3 SV

=2

- [BG

ALB

_ASP

NC]

Prob

able

pect

ate

lyase

A O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=pl

yA P

E=3

SV

=1

- [PL

YA

_ASP

NC]

Prob

able

endo

poly

galac

turo

nase

I O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=pg

aI P

E=

3 SV

=1

- [PG

LR1_

ASP

NC]

Prob

able

exop

olyg

alact

uron

ase

A O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=pg

xA P

E=

3 SV

=1

- [PG

XA

_ASP

NC]

Prob

able

rham

noga

lactu

rona

se A

OS=

Asp

ergil

lus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

rhgA

PE

=3

SV=

1 - [

RHG

A_A

SPN

C]

Prob

able

alph

a-ga

lacto

sidas

e A

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

aglA

PE

=3

SV=

1 - [

AG

ALA

_ASP

NC]

Prob

able

pect

in ly

ase

A O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=pe

lA P

E=

3 SV

=1

- [PE

LA_A

SPN

C]

Alp

ha-x

ylos

idas

e A

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=a

xlA

PE

=1

SV=

1 - [

XY

LA_A

SPN

C]

Prob

able

exop

olyg

alact

uron

ase

C O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=pg

xC P

E=

3 SV

=2

- [PG

XC_

ASP

NC]

Beta

- man

nosid

ase

A O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=m

ndA

PE

=3

SV=1

- [M

AN

BA_A

SPN

C]

Hist

one

H2A

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

httA

PE

=3

SV=

1 - [

H2A

_ASP

NC]

Prob

able

cutin

ase

1 O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=A

n14g

0217

0 PE

=3

SV=

1 - [

CUTI

1_A

SPN

C]

Prob

able

endo

poly

galac

turo

nase

B O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=pg

aB P

E=3

SV

=1

- [PG

LRB_

ASP

NC]

End

ochi

tinas

e A

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

ctcA

PE

=2

SV=

1 - [

CHIA

1_A

SPN

C]

Hist

one

H2B

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

htb1

PE

=3 S

V=

1 - [

H2B

_ASP

NC]

40S

ribos

omal

pro

tein

S0

OS=

Asp

ergi

llus n

iger

(stra

in C

BS 5

13.8

8 /

FGSC

A15

13) G

N=

rps0

PE=

3 SV

=1

- [RS

SA_A

SPN

C]

Prob

able

rham

noga

lactu

rona

te ly

ase

B O

S=A

sper

gillu

s nig

er (s

train

CBS

513

.88

/ FG

SC A

1513

) GN

=rg

lB P

E=3

SV=

1 - [

RGLB

_ASP

NC]

Acc

essio

n

A2Q

T85

A2Q

K82

A2R

060

A2Q

HG

0

A2R

797

A5A

AG

2

A2R

AY

7

A2Q

877

A2Q

A64

A2Q

V36

A2Q

AH

3

A2Q

W66

A2Q

YE

5

A2Q

L72

A2R

3I1

A2Q

TU5

A2Q

EW

2

A2Q

WU

9

P0C9

53

A2R

2W3

A2Q

CV8

A2Q

UQ

2

A2Q

Y49

A2Q

E32

A5A

BH4


137

Table S4.3: Upfield 1H chemical shifts of xylose oligo-saccharide as determined from the

2D TOCSY spectrum of fraction H8.

Chemical shift (p.p.m.) Protons Internal β-Xylp Non-reducing terminal β-Xylp

1H 4.48 4.46 2H 3.29 3.26 3H 3.556 3.42 4H 3.79 3.62 5H 4.10, 3.38 3.97, 3.31


138

Table S4.4: Assignment of chemical shifts of 1H and 1H-13C HSQC NMR anomeric

resonances. Xylp1 is reducing end xylopyranose; β-Xylp2/3/4/5 is the total of unsubstituted Xylp

residues other than the reducing end; α-Araf-XN-O- single (3s) or double (2 and 3d)

arabinofuranosidase decoration on unidentified Xylp via α-1,3-, or α-1,2- and α-1,3- linkage

respectively; β-XylpN-Af-O-3/2,3 Xylp with single/double Araf decoration respectively; (4-O-

Me)-α-GlcA-XN-C2 signal for GlcA and 4MeGlcA attached via α-1,2 to an unidentified Xylp,

β-XylpN-(Me)GlcA unidentified Xylp decorated with GlcA or 4MeGlcA, NA not identified

in this fraction, NR not resolvable.

Chemical shift (p.p.m.) Fraction AX14 Fraction H8

Residue 1H (1D) 13C (HSQC) 1H (1D) 13C (HSQC) α-Xylp1 5.184 93.1 5.183 92.9 β-Xylp1 4.584 97.6 4.584 97.6

β-Xylp2/3/4/5 ~4.47 ~102.7 ~4.48 ~102.8 α-Araf-XN-O3s 5.395 108.7 NA NA α-Araf-XN-O-2 5.224 109.8 NA NA α-Araf-XN-O-3d 5.273 109.2 NA NA β-XylpN-Af-O-2,3 4.640 100.9 NA NA β-XylpN-Af-O-3 4.43-4.45 NR NA NA

4-O-Me-α-GlcA-XN-2 NA NA 5.285 ~98.7 α-GlcA-XN-2 NA NA 5.303 ~98.7 β-XylpN-GlcA NA NA 4.641 ~102.7 β-XylpN-MeGlcA NA NA 4.623 ~102.7


139

Table S 4.5: Signal integration ratios 1H NMR. With the sum of the signal for the α- and β-

anomer of the reducing end xylopyranose (Xylp1) set to 1, β-XylpN-Af-O-3 + β-Xylp2/3/4/5

combined ratio for β-XylpN-Af-O-3 and β-Xylp2/3/4/5, other notations as in Table S3b.

Fraction Signal integration ratio 1H (1D) H6 H7 H8 AX14 AX15

Xylp1 1 1 1 1 1 β-Xylp2/3/4/5 8.2 6.5 6.3 NR NR

(4-O-Me)-α-GlcA-XN-C2 1.1 0.7 0.6 NA NA β-XylpN-(Me)GlcA 1.2 0.8 0.6 NA NA β-XylpN-Af-O-2,3 NA NA NA 0.4 0.3

β-XylpN-Af-O-3 + β-Xylp2/3/4/5 NA NA NA 3.5 3.0 α-Araf-XN-O-3s NA NA NA 0.9 0.7 α-Araf-XN-O-2 NA NA NA 0.4 0.3 α-Araf-XN-O-3d NA NA NA 0.4 0.3


140

Table S4.6: Ratio of the intensities of MALDI-ToF MS signal, calculated for DP5/DP8,

DP6/DP8, DP5/DP6 of fraction AX10 after labelling via reductive amination using various

reaction times and procedures. Procedure a: Excess of oligo-saccharide, hexadecylamine

(limiting reagent), excess of NaBH3CN, sodium borate/sodium sulfate buffer. Procedure b:

Oligo-saccharide (limiting reagent), excess of hexadecylamine, excess of NaBH3CN, sodium

borate/sodium sulfate buffer. Procedure c: Excess of oligo-saccharide, hexadecylamine

(limiting reagent), excess of NaBH3CN, sodium borate buffer.

DP5/DP8 DP6/DP8 DP5/DP6 Procedure a b c a b c a b c

1 hour 0.32 0.40 0.35 0.70 0.79 0.74 0.46 0.52 0.47 2 hours 0.39 0.39 0.38 0.77 0.80 0.76 0.50 0.50 0.50

4 hours 0.54 ± 0.02 0.37 0.40 0.92 ±

0.02 0.81 0.85 0.59 ± 0.01 0.48 0.47

8 hours 0.55 0.36 0.47 0.92 0.77 0.89 0.60 0.47 0.53 1 day 0.56 0.35 0.52 0.93 0.75 0.99 0.60 0.46 0.52

3 days 0.56 0.33 0.55 0.92 0.74 0.95 0.60 0.45 0.53 Unlabelled oligosacc. 0.24 0.53 0.45


141

Table S4.7: Interday reproducibility of array generation, given as mean with standard

deviation of the ratio of the intensities of MALDI-ToF MS signal, calculated for DP5/DP8,

DP6/DP8, DP5/DP6 of fraction AX10 and H9 immobilised over multiple days.

DP5/DP8 DP6/DP8 DP5/DP6 AX10 day 1 1.71 1.82 0.94 AX10 day 2 1.77 1.81 0.99 AX10 day 3 1.75 1.81 0.99

AX10 variability 1.74 ± 0.03 1.81 ± 0.01 0.97 ± 0.03 H9 day 1 1.93 1.68 1.15 H9 day 2 1.93 1.62 1.19 H9 day 3 1.97 1.65 1.24

H9 variability 1.94 ± 0.02 1.65 ± 0.03 1.19 ± 0.05

142

Chapter 5 Completing the N-acetylneuraminic acid toolkit

5.1 Summary

While plant-derived saccharides have industrial biotechnological relevance, N-glycans

form an integral part of eukaryotic biology. With an increasing number of therapeutic

glycoproteins being released, glycosylation is considered a crucial parameter to fine-tune

efficacy and prevent immunogenicity. Sialic acids (Neu5Ac) serves as a capping moiety in

human N-glycans. However, the enzymatic tool box for efficient analysis and remodelling

lacks an efficient α2,6-sialidase. This chapter presents the discovery and characterisation of

an α2,6-‘pseudo-sialidase’ and its application of complex glycoproteins.

5.2 Contribution

PB designed the study. PB identified candidate enzymes and recombinant protein was

provided by Prozomix Ltd., UK. As part of this thesis the analytical challenges were

addressed. PB and MR interpreted kinetic data. PB and KH performed biotransformation

reactions. MR designed, performed and interpreted N-glycan release and -analysis

experiments. MR and CJG co-designed, -performed and -analysed the ion mobility

experiments. EGP performed the NMR studies. LPC and JV provided compounds. The

manuscript was co-authored by PB, MR and CLG and edited by all authors.

5.3 Manuscript

This manuscript has been accepted for publication in Glycobiology on

19th December 2017.

Chapter 5 - ‘Pseudosialidase’

143

Development and application of a highly α2,6-selective

pseudosialidase.

Peter Both,*† Michel Riese,† Christopher J. Gray,† Kun Huang,† Edward G. Pallister, † Iaroslav

Kosov,† Louis P. Conway,‡ Josef Voglmeir,‡ and Sabine L. Flitsch*†

† School of Chemistry & Manchester Institute of Biotechnology, The University of

Manchester, Manchester M1 7DN, U.K.

‡ Glycomics Glycan Bioengineering Research Center, College of Food Science and

Technology, Nanjing Agricultural University, Nanjing 210095, China.

Abstract

Within human biology, combinations of regioisomeric motifs of α2,6- or α2,3-sialic acids

linked to galactose are frequently observed attached to glycoconjugates. These include

glycoproteins and glycolipids, with each linkage carrying distinct biological information and

function. Microbial linkage-specific sialidases have become important tools for studying the

role of these sialosides in complex biological settings, as well as being used as biocatalysts

for glycoengineering. However, currently, there is no α2,6-specific sialidase available. This

gap has been addressed herein by exploiting the ability of a Photobacterium sp. α2,6-

sialyltransferase to catalyze trans-sialidation reversibly and in a highly linkage-specific

manner, acting as a pseudosialidase with excess of cytidine monophosphate. Selective, near

quantitative removal of α2,6-linked sialic acids was achieved from a wide range of sialosides

including small molecules conjugates, simple glycan, glycopeptide and finally complex

glycoprotein including both linkages.


144

Introduction

Structure-function studies of glycans rely on the elucidation of their structure and on the

isolation of sufficient amounts of pure well-defined material.1,2 Structural analysis can be

addressed using the enzymes responsible for glycan biosynthesis. These enzymes fall into

two broad categories: glycosyltransferases and trans-glycosidases which are responsible for

the formation of specific glycosidic linkages from certain activated glycosides,3-8 and

glycosidases that hydrolyze particular glycosidic linkages. All of these enzymes currently

characterized are highly stereo-selective for formation and cleavage of α- or β-glycosidic

bond, but the glycosyltransferases tend to show higher regioselectivities in glycan bond

formation.9-13

Over the years a number of elegant studies have highlighted the reversibility of many

reactions involving glycosidic bonds and in particular glycosidases have been used ‘in reverse’

for synthesis of glycosidic linkages.14-17 Similarly, reactions catalyzed by inverting

glycosyltransferases have been shown to be reversible (including GT80, GT52 and GT42

sialyltransferases),18-21 although there are no examples of applying them as highly selective

exo-sialidases.

Sialyltransferases are of particularly interest because their substrates (e.g. N-

acetylneuraminic acid, Neu5Ac; a sialic acid) are important building blocks of animal

glycomes. Sialic acid moieties are abundant in the gangliosides of neural membranes in the

brain and in many glycoproteins, as well as bacteria of the microbiota of higher animals.22,23

Sialylation is a terminal/capping modification of the non-reducing ends or branches of

glycans, with a particularly common motif being sialyl-galactoside either in α2,3- or α2,6-

glycosidic linkage.23

Sialic acids are involved in a wide range of biological processes, ranging from cancer to

microbial infections.24-27 Human influenza viruses preferentially bind α2,6-linked sialic acid,


145

swine viruses bind both α2,6- and α2,3-linked sialic acid, while avian and equine viruses

display higher affinity toward α2,3-linked sialic acid.28-31 The great diversity, complexity and

heterogeneity of sialosides represent a serious challenge in the identification of glycan

structures and their function,23 and consequently sialidases have become important tools for

the analysis of sialo-glycoconjugates. Sialidases are highly selective for the α-sialyl linkage,

but show reduced specificity for regioisomeric structures, and with one know exception32

preferentially hydrolyze α2,3-linkages over α2,6-linkages (Figure 1, top; Figure S2).33 The

most commonly used sialidase to assess α2,3-linkages is NanB from Streptococcus pneumoniae,

which produces 2,7-anhydro-Neu5Ac from α2,3-sialosides.34,35 Hence, an enzyme with a

broad substrate scope able to selectively desialylate α2,6-linked Neu5Ac in native

glycoproteins, is currently missing in the glycomics toolbox.

Herein we discuss the exploitation of a α2,6-sialyltransferase as a pseudosialidase that

specifically removes α2,6-linked Neu5Ac by promoting the reverse reaction through addition

of excess CMP. Loss of Neu5Ac was shown by various chromatographic techniques, NMR,

mass spectrometry and ion mobility spectrometry for a wide range of sialosides including

small molecules conjugates, simple glycans, glycopeptides and finally glycoproteins. This

pseudosialidase was highly specific for α2,6-linked Neu5Ac compared to the other

commonly occurring regiosomer α2,3-linked Neu5Ac, for which selective sialidases are

already available.


146

Experimental Section

Sialidase Activity Assays

Sialidase activity of sialidases was assessed using an equimolar mixture of sialosides 1 and

2 (0.25 mM each) and 7.7 µM of enzyme incubated at 37 °C. Salmonella typhinurium LT2

sialidase reaction mixtures contained 50 mM sodium citrate buffer pH 3.5. Reaction mixtures

for the two putative sialidases contained 50 mM tris(hydroxymethyl)-aminomethane (Tris)

buffer pH 7.5. Reactions were incubated at 37 °C and samples taken after 1, 2, 4 and 6 hours.

Sialyltransferase Activity Assay

Sialyltransferase activity was assessed using reaction mixtures containing 50 mM Tris

buffer (pH 7.5), 0.25 mM compound 3, 1 mM CMP-Neu5Ac and 20 mM MgCl2. Reactions

were incubated for six h at 37 °C.

Pseudosialidase Activity Assay.

Pseudosialidase activity of enzymes was evaluated using an equimolar mixture (0.25 mM

each) of sialosides 1 and 2 containing 50 mM Tris buffer (pH 7.5), 4 mM CMP and 20 mM

MgCl2. Reactions were incubated for six h at 37 °C. In the case of Photobacterium sp. JT-ISH-

224 α2,6-sialyltransferase samples were taken after 1, 2, 4 and 6 hours. The pseudosialidase

activity of Photobacterium sp. JT-ISH-224 α2,6-sialyltransferase towards sialoside 2 (0.25 mM)

was assessed using 50 mM Tris buffer pH 7.5, 4 mM CMP, 20 mM MgCl2 and 7.7 µM of

enzyme in the reaction mixtures which were incubated at 37 °C. Samples were taken after 1,

2, 4 and 6 hours.


147

Kinetic Measurements.

The reaction mixtures for kinetic measurements contained 50 mM Tris buffer pH 7.5, 10

mM CMP, 20 mM MgCl2 and 1.54 µM of enzyme at different concentrations (8, 6, 4, 2, 1

and 0.5 mM) of sialoside 1. Reactions were incubated for 60 min at 37 °C. In none of the

reactions did conversion of sialoside 1 to compound 3 exceed 20 %. Conversion was in the

linear range for both the highest and the lowest concentration based on measurements at 30,

60 and 90 min.

Reverse Phase High-Performance Liquid Chromatography (RP HPLC).

Samples containing compounds 1, 2 and/or 3 were analyzed using RP HPLC.

Phenomenex C18 (2)column 250 x 2 mm 5 micron with UV detection at 245 nm (Eluents:

A – 50 mM ammonium formate pH 4.5, B – acetonitrile, flow rate 0.4 mL/min, 0-10 min

isocratic 10 % B, 10-30 min linear gradient 10 to 30 % B and from 30-45 isocratic 10 % B).

Pseudosialidase Reaction Followed by 13C NMR.

To follow pseudosialidase reaction using 13C labeled sialyllactose with 13C NMR a 200

µlenzymatic solution containing 6 mM uniformLy labeled 13C-pyruvate, 6 mM N-

acetylmannosamine (ManNAc), 6 mM CTP, 100 mM Tris buffer (pH 7.5), 10 mM MgCl2

10 µL (of 17.2 mg/mL) purified Escherichia coli K12 aldolase, 2 µL (of 6.3 mg/mL) Neisseria

meningitidis CMP-NeuAc synthase, 2.5 U Saccharomyces cerevisiae pyrophosphatase, was

incubated overnight at 37 °C to synthesize [1,2,3- 13C]-CMP-NeuAc. Following incubation

the enzymes were removed via ultrafiltration. 20 mM of lactose and 2 µlof (3 mg/mL)

Enterobacteriaceae bacterium FGI 57 α2,6-sialyltranferase was added to the filtrate and incubated

overnight to yield 13C labeled α2,6-sialyllactose labeled the C1, C2, and C3 position of the

sialic acid. The sialyltransferase was then removed through ultrafiltration. Some of the excess

lactose was broken down using a β1,4-galactosidase. Following removal of the β1,4-


148

galactosidase, 25 mM CMP and 25 µl(of 14.6 mg/mL) Photobacterium sp. JT-ISH-224 (total

reaction volume increased to 300 µL), was added to the α2,6-sialyllactose mixture and

incubated overnight at 37 °C. Following incubation the enzyme was removed via

ultrafiltration, and the filtrate diluted to 650 µL using D2O and placed in an NMR tube. N.B.

NMR spectra were obtained separately for each stage of the reaction (Figure S6). The

samples were analyzed using a Varian VMS and Bruker Avance II+ 500 MHz spectrometer.

All 13C NMR spectra were obtained using 2048 scans, 1.0223616 acquisition time, 254.8 ppm

spectral width. The C3 carbon of the Neu5Ac was used as a diagnostic carbon (38-44 ppm)

in the NMR to follow the enzymatic reactions. 1 µLaliquots were also taken before and after

the pseudosialidase reaction and analyzed using an Agilent 6510 QTOF connected to an

Agilent 1200 series LC. Flow injection used was 0.3 mL/min 50 % acetonitrile 0.1 % formic

acid.

Pseudosialidase Activity Towards Egg Yolk peptido-N-glycan.

The pseudosialidase activity of Photobacterium sp. JT-ISH-224 α2,6-sialyltransferase

towards egg yolk protein derived α2,6-disialylated bi-antennary peptido-N-glycan from

Ludger was evaluated in reaction mixtures containing 50 mM Tris buffer pH 7.5, 4 mM CMP,

20 mM MgCl2, 0.25 mM peptido-N-glycan and 7.7 µM of enzyme which were incubated at

37 °C for 4 and 16 hours.

Enzymatic Treatment of Fetal Calf Fetuin.

Reactions containing either Photobacterium sp. JT-ISH-224 α2,6-sialyltransferase or

Photobacterium damsela (Pda2,6ST) α2,6-sialyltransferase were prepared using 50 mM Tris

buffer pH 7.5, 4 mM CMP, 20 mM MgCl2, 2 mg/mL fetuin and 7.7 µM of enzyme which

were incubated at 37 °C for 16 hours. Reaction containing Salmonella typhinurium LT2 sialidase

(NanH) consisted of 50 mM sodium citrate buffer pH 6, 2 mg/mL fetuin and 7.7 µM of


149

enzyme which was incubated at 37 °C for 16 hours. Reaction of 25 µLcontaining Streptococcus

pneumoniae NanB consisted of 50 mM sodium phosphate buffer pH 6, 2 mg/mL fetuin and

2 µLof commercial enzyme (following the standard protocol provided by Sigma). Samples

were incubated at 37 °C for 16 hours. Untreated fetuin (from Sigma) served as a control.

Glycan Release, Labeling and Analysis.

Glycans were released and labeled using Waters’ GlycoWorks RapiFluor-MS N-Glycan

Kit and analyzed by hydrophilic interaction liquid chromatography (HILIC) using Waters’

ultra-performance liquid chromatography (UPLC) instrument. The N-glycans were cleaved

from the peptide/protein using PNGase F following detergent-assisted thermal

denaturation. Subsequent labeling with the RapiFluor-MS tag and solid phase extraction

yielded a purified complex mixture of labeled N-glycans for analysis. 20 µLof the glycan

mixture were injected and separated on a ACQUITY UPLC Glycan BEH Amide 130Å

column (1.7 µm, 2.1 x 150 mm) using a gradient (50 mM ammonium formate, pH 4.4 and

acetonitrile; 25 % buffer, 75 % acetonitrile to 46 % buffer, 54 % acetonitrile over 35 min)

with a flow rate of 0.4 mL/min at 60°C over 55 min. Retention times were calibrated against

dextran ladders to calculate glucose units (GU) values which were compared against the

GlycoBase database. Matching species were confirmed by MS.

Traveling Wave Ion Mobility-Mass Spectrometry (TWIMS-MS) Analysis.

Samples (approximately 1-10 µM, 45 mM ammonium acetate pH 7, 25 % DMF, 53.625

% acetonitrile in water) were infused into a Synapt G2-Si HDMS (Waters, UK) by static

nanoelectrospray ionization using pulled borosilicate emitters (World Precision Instruments,

USA, thin-wall capillary, 4” length, 1.2 mm OD). The capillary, cone voltage and source

temperature were typically set to 0.8-1.5 kV, 25 V and 40 °C respectively. No cone gas was


150

used. The trap DC entrance, bias and exit were set to 0, 45 and 3 V. The IM travelling wave

speed was set to 1000 m/s and the wave height set at 40 V. Nitrogen drift gas flow was set

at 90 mL/min for all experiments. The helium and argon flow were set to 180 and 2 mL/min

respectively for the helium and trap cell. The trap voltage was set to 25 V for all collision-

induced dissociation data and was otherwise set to 4 V for MS acquisitions. The transfer

voltage was set to 2 V throughout. The mass measurements were calibrated using 2 mg/mL

NaI calibrant in 50% Isopropyl alcohol. Drift times were calibrated to a mix of dextran 1000

(0.1 mg/mL) and 5000 oligomers (0.5 mg/mL) in the presence of 1 mM NaH2PO4 in 50%

MeOH, whose CCS have been previously verified by DTIMS.

Mass spectra and ATDs were processed using MassLynx V4.1 (Waters, UK) and

OriginPro 9.1 (OriginLabs, USA) respectively. ATDs were calibrated and subsequently

normalized to their maximum intensity. Gaussian distributions were fitted to these spectra

and the center of this fitted peak was taken as the peaks CCS.

Protein Gel Electrophoresis, Western Blot and Lectin Binding.

Samples of commercially available fetal calf fetuin and asialofetuin (both from Sigma),

fetal calf fetuin treated with pseudosialidase and fetal calf fetuin treated with nonspecific

sialidase NanH were subject to sodium dodecyl sulfate polyacrylamide gel electrophoresis

(SDS-PAGE) [4 µg load of each in 30 µL wells]. Reference gel was stained with Instant Blue

(Expedeon).

Two gels were treated for 5 min with Towbin buffer and subject to semi-dry Western blot

(WB) using PVDF membranes (BioRad) activated by methanol and presoaked in Towbin

buffer. Protein transfer to the membrane took 50 min at 15 V.

Membrane blocking was achieved by 1 h incubation with ‘Protein-Free (TBS) Blocking

Buffer’ (Pierce) at room temperature (RT). Lectin binding buffer consisted of 10 mM sodium

phosphate, 150 mM NaCl, 1 % (v/v) Tween 20, pH 7.8. The final concentration of lectin


151

for fluorescein labelled SNA I (from Sambucus nigra; Vector Laboratories) binding was 7.5

µg/mL (20 mL), while the final concentration of lectin for fluorescein labelled MAL I (from

Maackia amurensis; Vector Laboratories) binding was 10 µg/mL (20 mL). After a 1 h

incubation of the membranes with the respective lectins at RT images were obtained using

Typhoon Trio imaging system (blue laser, fluorescein specific filter, 50 µ pixels).

Results and Discussion

Pseudosialidase Activity Towards Simple Sialosides.

Given that efforts to identify new α2,6-sialidases so far had not been successful, we

decided to explore alternative strategies. Previous studies had shown that GT80 family38

sialyltransferases can catalyze both directions of the classical sialyltransferase reaction from

CMP-Neu5Ac to acceptor, which can be prevented by remocing CMP from the reaction

mixture.19-21 We decided to exploit this reversibility by adding excess CMP with the aim to

drive the reaction towards CMP-Neu5Ac formation, thus effectively using the enzyme as a

pseudosialidase. To ensure practical utility, it was important to demonstrate that the

specificity for the α2,6-linkage would be retained in the ‘hydrolysis’ reaction and the enzyme

can operate on complex glycoconjugates such as glycoproteins.

The GT80 family α2,6-sialyltransferases from Photobacterium damsela20 and Photobacterium sp.

JT-ISH-22439 were heterologously expressed in E. coli and their expression and activity

compared to three putative sialyltransferases from Neisseria gonorrhoae (GT52),

Enterobacteriaceae bacterium FGI 57 (GT52) and Campylobacter insulaenigrae (GT42) (Section

S2.1.). The GT52 members exhibited α2,6-sialyltransferase activity as well as α2,6-

pseudosialidase activity, but their bacterial expression was poor, which led us to explore more

suitable candidates. The enzyme from Campylobacter insulaenigrae (a GT42 member) was

confirmed to be an α2,3-sialyltransferase and also exhibited α2,3-pseudosialidase activity.


152

These results are in agreement with Thorson and co-workers’ hypothesis of the reversible

nature of inverting glycosyltransferases’ activity.18

Figure 1: Timecourse of Photobacterium sp. JT-ISH-224 α2,6-sialyltransferase’s

pseudosialidase activity toward an equimolar mixture of α2,6- (1) and α2,3-sialosides (2).

Tilted purple squares represent Neu5Ac, yellow circles represent galactose.

The pseudosialidase activity of JT-ISH-224 α2,6-sialyltransferase provided the most

promising results. The enzyme was tested against an equimolar mixture (0.25 mM each) of

sialosides 1 and 2 (Figure 1) in the presence of excess CMP with samples being taken after 1,

2, 4 and 6 hours. Analysis by RP HPLC showed hydrolysis of the α2,6-isomer 1 to 3 at ~95

% after 6 hours. The pseudosialidase activity toward α2,3-isomer 2 in the mixture was below

the detection limit. A sample of pure α2,3-isomer 2 under the same reaction conditions

showed very small amounts of hydrolysis (1.5 %; Figure S4).


153

Measurement of kinetics parameters for hydrolysis of 1 (Figure S5; Section S3.2.) showed

Km to be 4.9 ± 0.4 mM, kcat = 10 ± 1 min-1 and kcat/Km = 2 ± 0.4 mM-1min-1.

Previous work on trans-sialidase activity of GT80 family sialyltransferases suggests

formation of CMP-Neu5Ac.21 To confirm this, 13C labeled α2,6-sialyllactose (Neu5Ac labeled

at C 1, 2 and 3) was synthesized and used as a substrate, which allowed us to track the fate

of Neu5Ac by 13C NMR spectroscopy (Section S3.3.) and mass spectrometry (MS).

Interestingly, a decrease in the α2,6-sialyllactose signal and corresponding increase in free

Neu5Ac were observed, while no CMP-Neu5Ac intermediate was detected (Figures S6 and

S7), suggesting very fast further hydrolysis of CMP-Neu5Ac by the enzyme under the

reaction conditions. However, under the tested reaction conditions no sialidase activity was

detected in the absence of CMP.

Pseudosialidase Activity Towards Egg Yolk peptido-N-glycan.

Next, the activity of the α2,6-pseudosialidase towards the α2,6-disialylated bi-antennary

peptido-N-glycan (A2-peptide isolated from egg yolk; kindly provided by Ludger; Figure 2,

top) was assessed using ultra performance liquid chromatography (UPLC) coupled with

MS.40,41 Figure 2 shows that the enzyme does not discriminate between the two antennae

(partial hydrolysis generates near equimolar peaks of both mono-sialylated isomers, bottom

trace) and desialylation can be pushed to completion (Figure S8, bottom trace). The MS

profile of each peak is represented in the Supporting Information (Figures S9 – S11; Table

S1).


154

Figure 2: Hydrophilic interaction liquid chromatography (HILIC) elution profile (using

fluorescence detection)40 of A2 peptide(top) before and (bottom) after 4 hours of treatment

with pseudosialidase. Symbols represent: tilted purple square = Neu5Ac; yellow circle =

galactose; blue square = N-acetylglucosamine; green circle = mannose; / = α2,6-linkage.

Pseudosialidase Activity Towards Fetuin.

Finally, the α2,6-pseudosialidase was tested against a complex mixture of glycans on an

intact native protein, with commercially available bovine fetuin as a model substrate. The N-

glycosylation profile of fetuin (which contains three potential N-glycosylation sites) had been

analyzed previously to contain a number of bi- and tri-antennary structures with both α2,3-

and α2,6-linked terminal Neu5Ac moieties ranging from mono- to tri-sialylated structures

(Figure 3A; Figures S13 – S15; Table S2). Untreated fetuin, fetuin treated with NanB (an

α2,3-sialidase) and fetuin treated with non-specific sialidase NanH served as controls. UPLC-

MS analysis suggested that the majority of α2,6-linked Neu5Ac had been removed from the

glycans of bovine fetuin by the pseudosialidase (Figure 3B; Figures S16 - 18; Table S3).

Compared to the previous α2,6-disialylated bi-antennary peptido-N-glycan the fetuin

sample contained bi-antennary structures with di-α2,6-Neu5Ac, di-α2,3-Neu5Ac, and mixed

α2,6-/α2,3-Neu5Ac based on retention times and corresponding m/z values. Upon α2,6-

pseudosialidase treatment of fetuin the di-α2,3-Neu5Ac glycan is unchanged, while peaks

representing di-α2,6-Neu5Ac and mixed α2,6-/α2,3-Neu5Ac glycans disappear from the

chromatogram and new peaks corresponding to their α2,6-desialylated equivalents emerge.


155

Similar patterns are observed for tri-antennary structures. Treatment of fetuin with NanB

did not lead to complete removal of α2,3-linked Neu5Ac (Figure 3C; Figures S19 – S22;

Table S4). Treatment of fetuin with unspecific sialidase also did not proceed to completion

(Figure 3D; Figures S23 and S24; Table S5). However, a new peak not seen in previous

samples arose suggesting the presence of a tri-antennary structure with 1,3-linked galactose

(20.489 min, m/z 773.65 [M+3H]3+) instead of the usual 1,4-linked galactose (20.829 min,

m/z 773.50 [M+3H]3+) on one of the antennae.

In order to confirm that desialylation was generally α2,6-specific, ion mobility (IM) MS

was performed on selected samples (Section S3.6.). IM-MS is an emerging method for the

discrimination of isomeric structures and glycan sequencing.42-44 IM-MS2 had previously been

reported to be capable of easily and rapidly discerning Neu5Ac-α2,3/6-Gal-GlcNAc

terminating glycans based on the differing mobilities of the terminal B3-product ion (m/z

657) generated by collision-induced dissociation (detailed description in the SI).45,46 Initially,

the entire sample pool was analyzed (i.e. no quadrupole isolation of a specific m/z), so the

fragments represent the global glycan composition (Figure 4). The α2,6-disialylated bi-

antennary peptido-N-glycan (A2) gave rise to a B3 fragment with a collision cross section

(CCS) of 235 2 in close agreement with the values previously recorded by Kolarich and

coworkers (Figure 4A).45 In comparison, fetuin produced isomeric fragments corresponding

to α2,6- and α2,3-Neu5Ac terminating B3-product ions whose CCS were 235 and 246 Å2

respectively (Figure 4B). Treatment with the pseudosialidase resulted in a significant

reduction of the α2,6-Neu5Ac signal or was abolished for the α2,6-disialylated bi-antennary

peptido-N-glycan (A2) (Figure 4C).

Treatment of fetuin with the α2,3-specific sialidase NanB, resulted in the opposite

spectrum, where almost no α2,3-Neu5Ac signal was observed (Figure 4D). Treatment with

Salmonella typhimurium LT2 nonspecific sialidase (NanH) also resulted in the m/z 657 peak

being below detection level (data not shown).


156

Similar trends were also observed for quadrupole selected glycoforms, where after

treatment with the pseudosialidase glycoforms were observed to be enriched for α2,3-

Neu5Ac terminating glycans or if no α2,3-Neu5Ac glycoforms were present, their signal was

abolished (Figures S30-S34).

Likewise, treatment of fetuin with NanB resulted in enrichment of α2,6-Neu5Ac

glycoforms and a much greater number of desialylated species were observed within the mass

spectrum compared to the control (Figures S30, S32, S33 and S35). See SI section 3.6.1. for

more detail.

Figure 3: Hydrophilic interaction liquid chromatography (HILIC) elution profile

(fluorescence detection)40 of pseudosialidase treated native bovine fetuin N-glycans (B).

Controls used are (A) N-glycans released from untreated bovine fetuin, (C) N-glycans from

bovine fetuin treated with α2,3-sialidase NanB and (D) N-glycans released from bovine

fetuin treated with unspecific sialidase NanH. Symbols represent: tilted purple square =

Neu5Ac; yellow circle = galactose; blue square = N-acetylglucosamine; green circle =

mannose; / = α2,6-linkage; \ = α2,3-linkage; | = α2,(3 or 6)-linkage.


157

Figure 4: Collision cross section (CCS) distribution associated with the diagnostic

[Neu5Acα2,3/6-Galβ1,4-GlcNAc+H]+ tri-saccharide B3-product ion (m/z 657)45 generated

after collision-induced dissociation of all glycans (no m/z selection within the quadrupole)

derived from α2,6-disialylated bi-antennary peptido-N-glycan (A), fetuin (B) fetuin after

treatment with the α2,6-pseudosialidase (C) and fetuin after treatment with the α2,3-sialidase

NanB (D). Shaded in yellow and blue respectively are Gaussian fits associated with the α2,6-

sialylated and α2,3-sialylated B3-product ion.

In addition to MS-based analysis protein samples were analysed by protein gel

electrophoresis under denaturing conditions and Western blots using SNA I lectin from

elderberry bark, and MAL I lectin from Maackia amurensis.(Figure S36).47 These experiments

were in agreement with data obtained from the HILIC-UPLC profiles of the samples

(Figure 3; Section S3.7. for more detail).


158

Conclusions

In conclusion, we have demonstrated that the heterologously expressed Photobacterium sp.

JT-ISH-224 α2,6-sialyl-transferase is able to specifically target α2,6-linked Neu5Ac in

complex mixture of N-glycans of a native protein leading to removal of the target moiety

over α2,3-linkage in high yields. Hence, this α2,6-pseudosialidase should be a valuable

addition to the glycan analysis and remodeling toolbox.

Supporting Information

Additional experimental procedures, protein sequences, supplementary results containing

sialidase treatment profiles of compounds and MS data. This material is available free of

charge on the ACS Publications website as PDF

Corresponding Authors

[email protected], [email protected]

Author Contributions

The manuscript was written through contributions of all authors. / All authors have given

approval to the final version of the manuscript.

Notes

The authors declare no competing financial interests.

Acknowledgment

This study was supported by the BBSRC, EPSRC, InnovateUK: IBCatalyst

[BB/M02903411] and by the European Union’s Seventh Framework Programme

[FP7/2007–2013] under grant agreement No. 266025. The authors would like to thank

Ludger Ltd. for donating the egg yolk peptido-N-glycan.


159

References

[1] Dosekova, E.; Filip, J.; Bertok, T.; Both, P.; Kasak, P.; Tkac, J. Med. Res. Rev. 2017, 37,

514-626.

[2] Wang, L. X.; Amin, M. N. Chem. Biol. 2014, 21, 51-66.

[3] Rich, J. R.; Withers, S. G. Nat. Chem. Biol. 2009, 5, 206-215.

[4] Wang, L. X. Carbohydr. Res. 2008, 343, 1509-1522.

[5] Chen, X.; Varki, A. ACS Chem. Biol. 2010, 5, 163-176.

[6] Schmaltz, R. M.; Hanson, S. R.; Wong, C. H. Chem. Rev. 2011, 111, 4259-4307.

[7] Wang, L. X.; Lomino, J. V. ACS Chem. Biol. 2012, 7, 110-122.

[8] Šardzík, R.; Green, A. P.; Laurent, N.; Both, P.; Fontana, C.; Voglmeir, J.;

Weissenborn, M. J.; Haddoub, R.; Grassi, P.; Haslam, S. M.; Widmalm, G.; Flitsch, S.

L. J. Am. Chem. Soc. 2012, 134, 4521-4524.

[9] Schedin-Weiss, S.; Winblad, B.; Tjernberg, L. O. FEBS J. 2014, 281, 46-62.

[10] Paleček, E.; Tkáč, J.; Bartošík, M.; Bertók, T.; Ostatná, V.; Paleček, J. Chem. Rev. 2015,

115, 2045-2108.

[11] Stanton, R.; Hykollari, A.; Eckmair, B.; Malzl, D.; Dragosits, M.; Palmberger, D.;

Wang, P.; Wilson, I. B.; Paschinger, K. Biochim. Biophys. Acta. 2017, 1861, 699-714.

[12] Paschinger, K.; Wilson, I. B. Glycoconj. J. 2016, 33, 273-283.

[13] Wilson, I. B.; Paschinger, K. Anal. Bioanal. Chem. 2016, 408, 461-471.

[14] Vic, G.; Hastings, J. J.; Crout, D. H. G. Tetrahedron Asymmetry. 1996, 7, 1973-1984.

[15] Hancock, S. M.; Corbett, K.; Fordham-Skelton, A. P.; Gatehouse, J. A.; Davis, B. G.

Chembiochem. 2005, 6, 866-875.

[16] Bojarova, P.; Kren, V. Chimia (Aurau). 2011, 65, 65-70.

[17] Lu, L.; Liu, Q.; Jin, L.; Yin, Z.; Xu, L.; Xiao, M. PLoS One. 2015, 10, e0140531. D.O.I.:

10.1371/journal.pone.0140531.

[18] Zhang, C.; Griffith, B. R.; Fu, Q.; Albermann, C.; Fu, X.; Lee, I. K.; Li, L.; Thorson, J.

S. Science. 2006, 313, 1291-1294.

[19] Lairson, L. L.; Wakarchuk, W. W.; Withers, S. G. Chem. Commun. (Camb.). 2007, 4, 365-

367.

[20] Cheng, J.; Huang, S.; Yu, H.; Li, Y.; Lau, K.; Chen, X. Glycobiology. 2010, 20, 260-268.

[21] Mehr, K.; Withers, S. G. Glycobiology. 2016, 26, 353-359.

[22] Freeze, H. H.; Eklund, E. A.; Ng, B. G.; Patterson, M. C. Annu. Rev. Neurosci. 2015, 38,

105-125.


160

[23] Varki, N. M.; Varki, A. Lab. Invest. 2007, 87, 851-857.

[24] Stowell, S. R.; Ju, T.; Cummings, R. D. Annu. Rev. Pathol. 2015, 10, 473-510.

[25] Borsig, L.; Wong, R.; Hynes, R. O.; Varki, N. M.; Varki, A. Proc. Natl. Acad. Sci. U.S.A.

2002, 99, 2193-2198.

[26] Kato, K.; Ishiwa, A. Trop. Med. Health. 2015, 43, 41-52.

[27] Lubin, J. B.; Lewis, W. G.; Gilbert, N. M.; Weimer, C. M.; Almagro-Moreno, S.; Boyd,

E. F.; Lewis, A. L. Infect. Immun. 2015, 83, 3126-3136.

[28] Wang, Z.; Chinoy, Z. S.; Ambre, S. G.; Peng, W.; McBride, R.; de Vries, R. P.; Glushka,

J.; Paulson, J. C.; Boons, G. J. Science. 2013, 341, 379-383.

[29] Ito, T.; Couceiro, J. N.; Kelm, S.; Baum, L. G.; Krauss, S.; Castrucci, M. R.; Donatelli,

I.; Kida, H.; Paulson, J. C.; Webster, R. G.; Kawaoka, Y. J. Virol. 1998, 72, 7367-7373.

[30] Connor, R. J.; Kawaoka, Y.; Webster, R. G.; Paulson, J. C. Virology. 1994, 205, 17-23.

[31] Wilks, S.; de Graaf, M.; Smith, D. J.; Burke, D. F. Vaccine. 2012, 30, 4369-4376.

[32] Corfield, A. P.; Higa, H.; Paulson, J. C.; Schauer, R. Biochim. Biophys. Acta. 1983, 744,

121-126.

[33] Varki, A.; Schauer, R. Sialic Acids. In Essentials of Glycobiolgy, 2nd ed.; Varki, A.;

Cummings, R. D.; Esko, J. D.; Freeze, H. H.; Stanley, P.; Bertozzi, C. R.; Hart, G. W.;

Etzler, M. E.; Eds.; Cold Spring Harbor (NY): Cold Spring Harbor Laboratory Press;

2009. Chapter 14.

[34] Xu, G.; Kiefel, M. J.; Wilson, J. C.; Andrew, P. W.; Oggioni, M. R.; Taylor, G.L. J. Am.

Chem. Soc. 2011, 133, 1718-1721.

[35] Gut, H.; King, S. J.; Walsh, M. A. FEBS Lett. 2008, 582, 3348-3352.

[36] Huang, K.; Wang, M. M.; Kulinich, A.; Yao, H. L.; Ma, H. Y.; Martínez, J. E.; Duan,

X. C.; Chen, H.; Cai, Z. P.; Flitsch, S. L.; Liu, L.; Voglmeir, J. Carbohydr. Res. 2015, 415,

60-65.

[37] Minami, A.; Ishibashi, S.; Ikeda, K.; Ishitsubo, E.; Hori, T.; Tokiwa, H.; Taguchi, R.;

Ieno, D.; Otsubo, T.; Matsuda, Y.; Sia, S.; Inada, M.; Suzuki, T. FEBS Open Bio. 2013,

3, 231-236.

[38] Lombard, V.; Golaconda Ramulu, H.; Drula, E.; Coutinho, P. M.; Henrissat, B. Nucleic.

Acids. Res. 2014, 42, D490-495.

[39] Tsukamoto, H.; Takakura, Y.; Mine, T.; Yamamoto, T. J. Biochem. 2008, 143, 187-197.

[40] Bones, J.; Mittermayer, S.; O’Donoghue, N.; Guttman, A.; Rudd, P. M. Anal. Chem.

2010, 82, 10208-10215.


161

[41] Campbell, M. P.; Royle, L.; Radcliffe, C. M.; Dwek, R. A.; Rudd, P. M. Bioinformatics

2008, 24, 1214-1216.

[42] Both, P.; Green, A. P.; Gray, C. J.; Šardzík, R.; Voglmeir, J.; Fontana, C.; Austeri, M.;

Rejzek, M.; Richardson, D.; Field, R. A.; Wildmalm, G.; Flitsch, S. L.; Eyers, C. E. Nat.

Chem. 2014, 6, 65-74.

[43] Hofmann, J.; Hahm, H. S.; Seeberger, P. H.; Pagel, K. Nature. 2015, 526, 241-244.

[44] Gray, C.J.; Schindler, B.; Migas, L. G.; Pičmanová, M.; Allouche, A. R.; Green, A. P.;

Mandal, S.; Motawia, M. S.; Sánchez-Pérez, R.; Bjarnholt, N.; Møller, B. L.; Rijs, A. M.;

Barran, P. E.; Compagnon, I.; Eyers, C. E.; Flitsch, S.L. Anal. Chem. 2017, 89, 4540-

4549.

[45] Hinnenburg, H.; Hofmann, J.; Struwe, W. B.; Thader, A.; Altmann, F.; Varón Silva,

D.; Seeberger, P. H.; Pagel, K.; Kolarich, D. Chem. Commun. (Camb.] 2016, 52, 4381-

4384.

[46] Hofmann, J.; Struwe, W. B.; Scarff, C. A.; Scrivens, J. H.; Harvey, D. J.; Pagel, K. Anal.

Chem. 2014, 86, 10789-10795.

[47] Cummings, R. D.; Etzler, M. E. Antibodies and Lectins in Glycan Analysis. In

Essentials of Glycobiolgy, 2nd ed.; Varki, A.; Cummings, R. D.; Esko, J. D.; Freeze,

H. H.; Stanley, P.; Bertozzi, C. R.; Hart, G. W.; Etzler, M. E.; Eds.; Cold Spring Harbor

(NY): Cold Spring Harbor Laboratory Press; 2009. Chapter 45.


162


Development and application of a highly α2,6-selective

pseudosialidase.

Peter Both,*† Michel Riese,† Christopher J. Gray, † Kun Huang,† Edward G. Pallister,†

Iaroslav Kosov,† Louis P. Conway,‡ Josef Voglmeir,‡ and Sabine L. Flitsch*†


163

Chemo-enzymatic synthesis of UV active sialylgalactosides from 5-bromo-4-chloro-

3-indolyl-β-D-galactopyranoside (X-Gal)

Chemo-enzymatic synthesis of sialosides 1 and 2

X-Gal-α2,6-Neu5Ac (1) and X-Gal-α2,3-Neu5Ac (2) were chemo-enzymatically

synthesized according to a previously reported method.S1 Briefly, N-acetylglucosamine (44

mg, 0.20 mmol) was dissolved in 50 mL of water and combined with sodium pyruvate (53

mg, 0.48 mmol), cytidine triphosphate (38 mg, 79 µmol), X-Gal (3) (16 mg, 39 µmol) , MgCl2

(11 mg, 1.2 mmol), N-aceylglucosamine-2-Epimerase (50 mU from Pedobacter heparinus), sialic

acid aldolase (25 mU from Escherichia coli), CMP-sialic acid synthase (20 mU from Neisseria

meningitidis), sialyltransferase (20 mU of Campylobacter jejuni α2,3-Sialyltransferase or 25 mU

Photobacterium damsalae α2,6-Sialyltransferase), in 2-(N-morpholino)ethanesulfonic acid

(MES) buffer (final concentration of 40 mM, pH 6.5 to a final volume of 50 mL). The

solutions were incubated at 37°C until ≥80 % conversions were observed by HPLC analysis.

Excess X-Gal was digested using β-galactosidase from Escherichia coli (~10 U, 37 °C, 30 min)

and the X-Gal-labeled sialic acids were isolated by solid phase extraction (RP-SPE, yielding

7 mg and 10 mg, respectively), and analyzed using nuclear magnetic resonance (NMR)

spectroscopy.

Concentration assessment of 1 and 2 was performed by HPLC using standard curve of

the commercially available compound 3.

Sialidases and sialyltransferases used in this study

Photobacterium sp. JT-ISH-224 α2,6-sialyltransferase, Salmonella typhinurium LT2 sialidase

(NanH), Paeniclostridium sordellii 8483 sialidase, Saccharothrix xinjiangensis sialidase, Neisseria

gonorrhoae α2,6-sialyltransferase, Enterobacteriaceae bacterium FGI 57 α2,6-sialyltransferase and

Campylobacter insulaenigrae α2,3-sialyltransferase were obtained from Prozomix. Photobacterium

damsela α2,6-sialyltransferase was expressed and purified as reported previously.S2 Streptococcus

pneumoniae sialidase (NanB) was purchased in the form of aqueous solution from Sigma.


164

Molecular weights and amino acid sequences of recombinant enzymes provided by

Prozomix, Ltd.

Photobacterium sp. JT-ISH-224 α2,6-sialyltransferase (MW = 58.8 kDa):

MGSSHHHHHHSSGLVPRGSHMEENTQSIIKNDINKTIIDEEYVNLEPINQSNISFTKHSWVQTC

GTQQLLTEQNKESISLSVVAPRLDDDEKYCFDFNGVSNKGEKYITKVTLNVVAPSLEVYVDHAS

LPTLQQLMDIIKSEEENPTAQRYIAWGRIVPTDEQMKELNITSFALINNHTPADLVQEIVKQAQ

TKHRLNVKLSSNTAHSFDNLVPILKELNSFNNVTVTNIDLYDDGSAEYVNLYNWRDTLNKTDNL

KIGKDYLEDVINGINEDTSNTGTSSVYNWQKLYPANYHFLRKDYLTLEPSLHELRDYIGDSLKQ

MQWDGFKKFNSKQQELFLSIVNFDKQKLQNEYNSSNLPNFVFTGTTVWAGNHEREYYAKQQINV

INNAINESSPHYLGNSYDLFFKGHPGGGIINTLIMQNYPSMVDIPSKISFEVLMMTDMLPDAVA

GIASSLYFTIPAEKIKFIVFTSTETITDRETALRSPLVQVMIKLGIVKEENVLFWADLPNCETG

VCIAV

Salmonella typhinurium LT2 sialidase (MW = 44.2 kDa):

MGSSHHHHHHSSGLVPRGSHMTVEKSVVFKAEGEHFTDQKGNTIVGSGSGGTTKYFRIPAMCTT

SKGTIVVFADARHNTASDQSFIDTAAARSTDGGKTWNKKIAIYNDRVNSKLSRVMDPTCIVANI

QGRETILVMVGKWNNNDKTWGAYRDKAPDTDWDLVLYKSTDDGVTFSKVETNIHDIVTKNGTIS

AMLGGVGSGLQLNDGKLVFPVQMVRTKNITTVLNTSFIYSTDGITWSLPSGYCEGFGSENNIIE

FNASLVNNIRNSGLRRSFETKDFGKTWTEFPPMDKKVDNRNHGVQGSTITIPSGNKLVAAHSSA

QNKNNDYTRSDISLYAHNLYSGEVKLIDDFYPKVGNASGAGYSCLSYRKNVDKETLYVVYEANG

SIEFQDLSRHLPVIKSYN


165

Paeniclostridium sordellii 8483 sialidase (MW = 44 kDa):

MGSSHHHHHHSSGLVPRGSHMSNLNTTNEPQKTTIFNKNDNMWNAQYFRIPSLQTLADGTMLAF

SDIRYNGAADHAYIDIGAAKSTDNGQTWEYKTVMENDRIDSTFSRVMDSTTVVTDTGRIILIAG

SWNKNGNWASSTTSLRSDWSVQMVYSDDNGETWSDKVDLTTNKARIKNQPSNTIGWLGGVGSGI

VMSDGTIVMPIQIALRENNANNYYSSVIYSKDNGETWTMGNKVPDPKTSENMVIELDGALIMSS

RNDGKNYRASYISYDLGSTWEVYDPLHNKISTGNGSGCQGSFIKVTAKDGHRLGFISAPKNTKG

GYVRDNITVYMVDFDDLSKGVRELCIPYPEDGNSSGGGYSCLSFNNGKLSILYEANGNIEYKDL

TDYYLSLENNKKLK

Saccharothrix xinjiangensis sialidase (MW = 45.9 kDa):

MGSSHHHHHHSSGLVPRGSHMAPQNTQATPHFSTVDDPAMDAPQQHTAYVLFQGGHRESLNGIT

YHSFRIPAIVRTNAGTLLAFAEGRVKSNQDHGNINLVYKRGVNNGSKASDWSGLKEAVGAGMGT

WGNPTPVVDRATGTIWLFLSWNAADKSLGGGTNPDTGEPTSPIRQWGERRVYAMKSTDDGLTFT

GLDGSSRPTDLTEALLPKTKADGSTWAWDAMGPGAGLYTSGGALVIPAQHRNIYSTDHGRTWKV

QKLAEGTGEATITELADGTLYRNDRPGGTTWAEIAKRRFVARGGLDSGFAPFTPDDTLLDPKNQ

ASVLQYNNDAPARTIFLNSASTEVRTKMRVRLSYDGARTWPVSRPLSDGPSAPGAGTEGGYSSM

AKTSDYRIGALVESNLDVGDRTSARSIVFRKFNLSWILHGCAC

Neisseria gonorrhoae α2,6-sialyltransferase (MW = 42.9 kDa):

MGSSHHHHHHSSGLVPRGSHMDRVNQGERNAVSLLKDKLFNEEGKPVNLIFCYTILQMKVAERI

MAQHPGERFYVVLMSENRNEKYDYYFNQIKDKAERAYFFYLPYGLNKSFNFIPTMAELKVKSML

LPKVKRIYLASLEKVSIAAFLSTYPDAEIKTFDDGTNNLIRESSYLGGEFAVNGAIKRNFARMM

VGDWSIAKTRNASDEHYTIFKGLKNIMDDGRRKMTYLPLFDASELKAGDETGGTVRILLGSPDK

EMKEISEKAAKNFNIQYVAPHPRQTYGLSGVTALNSPYVIEDYILREIKKNPHTRYEIYTFFSG

AALTMKDFPNVHVYALKPASLPEDYWLKPVYALFRQADIPILAFDDKNQSHGKSK


166

Enterobacteriaceae bacterium FGI 57 α2,6-sialyltransferase (MW = 37.7 kDa):

MGSSHHHHHHSSGLVPRGSHMKKVIKEETIEYDDIEFVYFSKTYDDKQSKYYELISVHAKKSTF

ITGVYSFKLIQELKAKFKGRSYDKVLLASLDDSINHYLLSFCDFNQLITFDDGVGNIIKTGAYF

IEDSRRSLKKRFFTLVHILLGRKYYLNLIKQRSDRHYTIYKGFENCVPNAKAISLFDFHLSPKT

FNNVTHVFLGTMFNEITIQQHEGDYLKKELLSYMLGIQGSVKYIPHPRSRDREFTPYEFESNGI

AEEIVFELLSHGYVVNLYGFASSCQFNLMNIRGVNIILLDSPRVNMAVKEAINMLLEKIPSSNY

INIQS

Campylobacter insulaenigrae α2,3-sialyltransferase (MW = 37.3 kDa):

MGSSHHHHHHSSGLVPRGSHMEEKNALICGNGPSLREIEYEKLPKNYDVFRCNQFYFEDKYYAG

KNIQYAFFNPYVFFEQYYTIKNIIDKDEYDIKNIVCSSFGLESIDSRNLLEFFYNYFPDTIFGF

DLIKQLKEFHSFIKFNEIYNEQRITSGIYMCAFAVAMGYKNIYISGIDFYSNKNQPYLFKYQTN

NVLKLIPEFKNEIKATIHSKNFDLKALEFLSKTYNVNFYSLNQNSELSKYIKLASIIRGNNDFV

IDNKPKDYIDDILLPANNTYKKFKKFPLPVIKNNLWFRLIKDLVRLPSDIKHYLKDKR


167

Sodium Dodecyl Sulphate - PolyAcrylamide Gel Electrophoresis profile of sialidases and

Photobacterium sp. JT-ISH-224 α2,6-sialyltransferase produced by Prozomix

Figure S1: SDS-PAGE profile of a) Photobacterium sp. JT-ISH-224 α2,6-sialyltransferase, b)

Salmonella typhinurium LT2 sialidase, c) Paeniclostridium sordellii 8483 sialidase and d) Saccharothrix

xinjiangensis sialidase.


168

Sialidases: Activity towards α2,3- and α2,6-sialoside

Sialidase activity of sialidases was assessed using an equimolar mixture of sialosides 1 and

2 (0.25 mM each) and 7.7 µM of enzyme incubated at 37 °C. Salmonella typhinurium LT2

sialidase reaction mixtures contained 50 mM sodium citrate buffer pH 3.5. Reaction mixtures

for the two putative sialidases contained 50 mM tris(hydroxymethyl)-aminomethane (Tris)

buffer pH 7.5. Reactions were incubated at 37 °C and samples taken after 1, 2, 4 and 6 hours.

Samples containing compounds 1, 2 and/or 3 were analyzed using RP HPLC.

Phenomenex C18 (2)column 250 x 2 mm 5 micron with UV detection at 245 nm (Eluents:

A – 50 mM ammonium formate pH 4.5, B – acetonitrile, flow rate 0.4 mL/min, 0-10 min

isocratic 10 % B, 10-30 min linear gradient 10 to 30 % B and from 30-45 isocratic 10 % B).


169

Figure S2: Timecourse of sialidase activity using equimolar mixtures of sialosides 1 and 2.

(A) Salmonella typhinurium LT2 sialidase; (B) Paeniclostridium sordellii 8483 sialidase; (C)

Saccharothrix xinjiangensis sialidase.

0102030405060708090

100

0 1 2 3 4 5 6pe

ak a

rea

/ %t / h

123

0102030405060708090

100

0 1 2 3 4 5 6

peak

are

a / %

t / h

123

0

10

20

30

40

50

60

0 1 2 3 4 5 6

peak

are

a / %

t / h

123

A

B

C


170

Photobacterium sp. JT-ISH-224 α2,6-sialyltransferase

Sialyltransferase activity was assessed using reaction mixtures containing 50 mM Tris

buffer (pH 7.5), 0.25 mM compound 3, 1 mM CMP-Neu5Ac and 20 mM MgCl2. Reactions

were incubated for six hours at 37 °C. All above mentioned reactions were stopped by

addition of 1:1 volume of methanol.

Pseudosialidase activity: timecourse

Pseudosialidase activity of enzymes was evaluated using an equimolar mixture (0.25 mM

each) of sialosides 1 and 2 containing 50 mM Tris buffer (pH 7.5), 4 mM CMP and 20 mM

MgCl2. Reactions were incubated for six h at 37 °C. In the case of Photobacterium sp. JT-ISH-

224 α2,6-sialyltransferase samples were taken after 1, 2, 4 and 6 hours.


towards sialoside 2 (0.25 mM) was assessed using 50 mM Tris buffer pH 7.5, 4 mM CMP,

20 mM MgCl2 and 7.7 µM of enzyme in the reaction mixtures which were incubated at 37

°C. Samples were taken after 1, 2, 4 and 6 hours.

All above mentioned reactions were stopped by addition of 1:1 volume of methanol.

Figure S4: Pseudosialidase activity toward sialoside 2 (α2,3-linked).

0

20

40

60

80

100

0 1 2 3 4 5 6

peak

are

a / %

time / h

2 3


171

Kinetic parameters for compound 1

The reaction mixtures for kinetic measurements contained 50 mM Tris buffer pH 7.5, 10

mM CMP, 20 mM MgCl2 and 1.54 µM of enzyme at different concentrations (8, 6, 4, 2, 1

and 0.5 mM) of sialoside 1. Reactions were incubated for 60 min at 37 °C. In none of the

reactions did conversion of sialoside 1 to compound 3 exceed 20 %. Conversion was in the

linear range for both the highest and the lowest concentration based on measurements at 30,

60 and 90 min.

All above mentioned reactions were stopped by addition of 1:1 volume of methanol.

Figure S5: Hanes-Woolf plot for compound 1 using Photobacterium sp. JT-ISH-224 α2,6-

sialyltransferase. Rates (v) are equal to the concentration of compound 3 produced in the

reactions.


172

𝑉UQ^ =1𝑏 = 909.1 𝜇𝑀 ℎj ≈ 15 ± 1𝜇𝑀 𝑚𝑖𝑛j

∆𝑉UQ^ = n𝜕𝑉UQ^𝜕𝑏 n ∙ ∆𝑏 = 67.2 𝜇𝑀 ℎj = 1.1𝜇𝑀 𝑚𝑖𝑛j

𝒌𝒄𝒂𝒕 =𝑉UQ^[𝐸] ≈ 𝟏𝟎 ± 𝟏𝒎𝒊𝒏{𝟏

𝑲𝒎 = 𝑎 ∙ 𝑉UQ^ =𝑎𝑏 ≈ 𝟒. 𝟗 ± 𝟎. 𝟒𝒎𝑴

∆𝐾a = n𝜕𝐾U𝜕𝑎 n ∙ ∆𝑎 + n

𝜕𝐾U𝜕𝑏 n ∙ ∆𝑏 = 0.4𝑚𝑀

𝒌𝒄𝒂𝒕𝑲𝒎

≈ 𝟐. 𝟎 ± 𝟎. 𝟒𝒎𝑴{𝟏𝒎𝒊𝒏{𝟏

Following pseudosialidase reaction using 13C labeled sialyllactose with 13C NMR

To follow pseudosialidase reaction using 13C labeled sialyllactose with 13C NMR a 200

µlenzymatic solution containing 6 mM uniformLy labeled 13C-pyruvate, 6 mM N-

acetylmannosamine (ManNAc), 6 mM CTP, 100 mM Tris buffer (pH 7.5), 10 mM MgCl2 10

µl(of 17.2 mg/mL) purified Escherichia coli K12 aldolase, 2 µl(of 6.3 mg/mL) Neisseria

meningitidis CMP-NeuAc synthase, 2.5 U Saccharomyces cerevisiae pyrophosphatase, was

incubated overnight at 37 °C to synthesize [1,2,3- 13C]-CMP-NeuAc. Following incubation

the enzymes were removed via ultrafiltration. 20 mM of lactose and 2 µlof (3 mg/mL)

Enterobacteriaceae bacterium FGI 57 α2,6-sialyltranferase was added to the filtrate and incubated

overnight to yield 13C labeled α2,6-sialyllactose labeled the C1, C2, and C3 position of the


173

sialic acid. The sialyltransferase was then removed through ultrafiltration. Some of the excess

lactose was broken down using a β1,4-galactosidase. Following removal of the β1,4-

galactosidase, 25 mM CMP and 25 µl(of 14.6 mg/mL) Photobacterium sp. JT-ISH-224 (total

reaction volume increased to 300 µl), was added to the α2,6-sialyllactose mixture and

incubated overnight at 37 °C. Following incubation the enzyme was removed via

ultrafiltration, and the filtrate diluted to 650 µlusing D2O and placed in an NMR tube. N.B.

NMR spectra were obtained separately for each stage of the reaction (Figure S6). The samples

were analyzed using a Varian VMS and Bruker Avance II+ 500 MHz spectrometer. All 13C

NMR spectra were obtained using 2048 scans, 1.0223616 acquisition time, 254.8 ppm

spectral width. The C3 carbon of the Neu5Ac was used as a diagnostic carbon (38-44 ppm)

in the NMR to follow the enzymatic reactions. 1 µl aliquots were also taken before and after

the pseudosialidase reaction and analyzed using an Agilent 6510 QTOF connected to an

Agilent 1200 series LC. Flow injection used was 0.3 mL/min 50 % acetonitrile 0.1 % formic

acid.


174

Figure S6: 13C NMR spectra of the C3 carbon position for (A) [1,2,3-13C]-Neu5Ac, (B)

[1,2,3-13C]-CMP-Neu5Ac, (C) sialylation of lactose with [1,2,3-13C]-CMP-Neu5Ac, (D)

sialidase reaction of labeled sialyllactose.

Three separate products were observed during the progress of the reaction: Neu5Ac (C3

δ 39.3 (d JC3-C2 = 41 Hz), CMP-Neu5Ac (C3 δ 41.2 (dd JC3-C2= 41 Hz, JC3-P1 = 10 Hz) and

sialyllactose (C3 δ 40.1 (d JC3-C2= 42 Hz). The predicted and observed m/z were as follows:

[1,2 3-13C]-Neu5Ac: predicted [M-H]- = 311.109, observed [M-H]- = 311.108; [1,2 3-13C]-

CMP-Neu5Ac predicted [M-H]- = 616.150, observed [M-H]- = Not observed; 13C labeled

sialyllactose predicted [M-H]- = 635.214, observed [M-H]- = 635.214.

A

B

C

D


175

Figure S7: LC-ESI MS spectrum before pseudosialidase reaction (bottom) and after

pseudosialidase reaction (top). A clear fall in the sialyllactose signal at m/z 635 (red box)

after pseudosialidase reaction (top) compared with before reaction (bottom). No evidence

of any CMP-Neu5Ac in either mass spectrum (m/z 616; gray box).

The 13C NMR spectrum following the pseudosialidase reaction of the labeled α2,6-

sialyllactose does not indicate the presence of CMP-Neu5Ac however, an increase of the free

Neu5Ac was observed. This agrees well with what has been previously observed when using

other techniques such as HPLC.S3

Activity toward α2,6-disialylated biantennary egg yolk N-glycopeptide


towards egg yolk protein derived α2,6-disialylated bi-antennary peptido-N-glycan from

Ludger was evaluated in reaction mixtures containing 50 mM Tris buffer pH 7.5, 4 mM CMP,

20 mM MgCl2, 0.25 mM peptido-N-glycan and 7.7 µM of enzyme which were incubated at

37 °C for 4 and 16 hours.

Before pseudosialidase reaction

After pseudosialidase reaction


176

Figure S8: Desialylation of α2,6-disialylated bi-antennary egg yolk N-glycopeptide.

Upon inspection we noticed a shoulder on the desialylated species which becomes more

prevalent as time goes on (Figure S8, bottom). This shoulder is absent within the fetuin

samples, suggesting this peak does not arise from the α2,6-pseudosialidase isomerizing the

structure. No degalactosylated species were observed in this sample and the mass spectrum

of the peak after 4 and 16 hours are near identical (Figure S10 top and S11) suggesting the

presence of an additional isomer present in the starting material but inseparable by the

conditions used in HILIC UPLC. IM-MS analysis confirmed the presence of only α2,6-

sialylated precursors (Figure 5).


177

untreated peptido-N-glycan, 23.3 min

Figure S9: α2,6-disialylated biantennary egg yolk N-glycopeptide MS profile.

S8.1


178

Pseudosialidase treated peptido-N-glycan (4 h), 17.8, 20.6, 20.8 and 23.4 min

Figure S10: Desialylation of α2,6-disialylated biantennary egg yolk N-glycopeptide after

4 hours with MS data.

S8.2

S8.3

S8.4

S8.5


179

pseudosialidase treated peptido-N-glycan (16 h), 17.4 min

Figure S11. Desialylation of α2,6-disialylated biantennary egg yolk N-glycopeptide after

16 hours with MS data.

S8.6


180

Table S1: Peptido-N-glycan control and pseudosialidase mediated desialylation with MS

data.

Sample Retention time [min]

found m/z calc. m/z Fragment

untreated peptido-N-glycan S8.1 23.3 845.91 845.66 A2G(4)2S(6,6)2

[M+3H]3+

pseudosialidase treated peptido-N-glycan (4 h)

S8.2 17.8 651.83 651.60 A2G(4)2 [M+3H]3+ 977.32 976.89 A2G(4)2 [M+2H]2+

S8.3 20.6 748.89 748.63 A2G(4)2S(?)1

[M+3H]3+

1122.90 1122.44 A2G(4)2S(?)1 [M+2H]2+

S8.4 20.8 748.68 748.63 A2G(4)2S(?)1

[M+3H]3+

1122.77 1122.44 A2G(4)2S(?)1 [M+2H]2+

S8.5 23.4 846.01 845.66 A2G(4)2S(6,6)2 [M+3H]3+

pseudosialidase treated peptido-N-glycan

(16 h) S8.6 17.7

651.79 651.60 A2G(4)2 [M+3H]3+

977.30 976.89 A2G(4)2 [M+2H]2+

Treatment of native bovine fetuin

Treatment of a glycoprotein (bovine fetuin): Reactions containing either Photobacterium sp.

JT-ISH-224 α2,6-sialyltransferase or Photobacterium damsela (Pda2,6ST) α2,6-sialyl-transferase

were prepared using 50 mM Tris buffer pH 7.5, 4 mM CMP, 20 mM MgCl2, 2 mg/mL fetuin

and 7.7 µM of enzyme which were incubated at 37 °C for 16 hours. Reaction containing

Salmonella typhinurium LT2 sialidase (NanH) consisted of 50 mM sodium citrate buffer pH 6,

2 mg/mL fetuin and 7.7 µM of enzyme which was incubated at 37 °C for 16 hours. Reaction

of 25 µLcontaining Streptococcus pneumoniae NanB consisted of 50 mM sodium phosphate

buffer pH 6, 2 mg/mL fetuin and 2 µLof commercial enzyme (following the standard

protocol provided by Sigma). Samples were incubated at 37 °C for 16 hours. Untreated fetuin

(from Sigma) served as a control.


181

Glycans were released and labeled using Waters’ GlycoWorks RapiFluor-MS N-Glycan

Kit and analyzed by hydrophilic interaction liquid chromatography (HILIC) using Waters’

ultra-performance liquid chromatography (UPLC) instrument. The N-glycans were cleaved

from the peptide/protein using PNGase F following detergent-assisted thermal

denaturation. Subsequent labeling with the RapiFluor-MS tag and solid phase extraction

yielded a purified complex mixture of labeled N-glycans for analysis. 20 µLof the glycan

mixture were injected and separated on a ACQUITY UPLC Glycan BEH Amide 130Å

column (1.7 µm, 2.1 x 150 mm) using a gradient (50 mM ammonium formate, pH 4.4 and

acetonitrile; 25 % buffer, 75 % acetonitrile to 46 % buffer, 54 % acetonitrile over 35 min)

with a flow rate of 0.4 mL/min at 60 °C over 55 min. Retention times were calibrated against

dextran ladders to calculate glucose units (GU) values which were compared against the

GlycoBase database.41 Matching species were confirmed by MS.


182

Figure S12. Native bovine fetuin desialylation. (A) untreated fetuin. (B) fetuin treated with

pseudosialidase. (C) fetuin treated with heavily α2,3-preferential NanB. (D) fetuin treated

with non-specific NanH.


183

Figure S13. Mass spectra of desialylation of native bovine fetuin control, Fig. S12A (peaks

1-4).

S12A

.1

S12A

.2

S12A

.3

S12A

.4


184


5-8).

S12A

.5

S12A

.6

S12A

.7

S12A

.8


185


9-12).

S12A

.9

S12A

.10

S12A

.11

S12A

.12


186

Table S2. Peak assignment of desialylation of native bovine fetuin control, Fig. S12A based

on retention time and m/z values.

Retention time [min] found m/z calc. m/z Fragment

S12A.1 14.9 773.96 773.81 M5 [M+2H]2+ S12A.2 19.4 1021.32 - - S12A.3 20.5 1122.56 1122.44 A2G(4)2S(?)1 [M+2H]2+ S12A.4 20.8 845.73 845.66 A2G(4)2S(??)2 [M+3H]3+ S12A.5 21.9 845.88 845.66 A2G(4)2S(??)2 [M+3H]3+ S12A.6 22.9 846.09 845.66 A2G(4)2S(??)2 [M+3H]3+ S12A.7 24.7 1064.35 1064.40 A3G(4)3S(???)3 [M+3H]3+ S12A.8 25.6 1064.75 1064.40 A3G(4)3S(???)3 [M+3H]3+ S12A.9 26.4 1064.86 1064.40 A3G(4)3S(???)3 [M+3H]3+ S12A.10 26.9 1176.54 - - S12A.11 27.4 1064.74 1064.40 A3G(4)3S(???)3 [M+3H]3+ S12A.12 27.7 1176.69 - -


187

Figure S16. Mass spectra of desialylation of native bovine fetuin treated with

pseudosialidase, Fig. S12B (peaks 1-4).

S12B

.1

S12B

.2

S12B

.3

S12B

.4


188



S12B

.5

S12B

.6

S12B

.7

S12B

.8


189



S12B

.9

S12B

.10

S12B

.11


190

Table S3. Peak assignment of desialylation of native bovine fetuin treated with

pseudosialidase, Fig. S12B based on retention time and m/z values.


S12B.1 14.9 773.42 773.81 M5 [M+2H]2+

S12B.2 17.7 651.72 651.60 A2G(4)2 [M+3H]3+ 977.13 976.89 A2G(4)2 [M+2H]2+

S12B.3 19.1 1122.73 1122.44 A2G(4)2S(3)1 [M+2H]2+ 748.79 748.63 A2G(4)2S(3)1 [M+3H]3+

S12B.4 21.0 773.56 773.31 A3G(?,?,?)3 [M+3H]3+ 1159.79 1159.46 A3G(?,?,?)3 [M+2H]2+

S12B.5 22.2 870.78 870.34 A3G(4)3S(3)1 [M+3H]3+ S12B.6 23.4 967.71 967.37 A3G(4)3S(??)2 [M+3H]3+ S12B.7 23.6 967.69 967.37 A3G(4)3S(??)2 [M+3H]3+ S12B.8 24.4 968.05 967.37 A3G(4)3S(??)2 [M+3H]3+ S12B.9 24.9 1064.86 1064.40 A3G(4)3S(???)3 [M+3H]3+ S12B.10 25.6 1064.76 1064.40 A3G(4)3S(???)3 [M+3H]3+

S12B.11 26.9 1168.47 1168.73 A3S(6)1G(4,4,3)3S(???)3 [M+3H]3+


191

Figure S19. Mass spectra of desialylation of native bovine fetuin treated with heavily α2,3-

preferential NanB, Fig. S12C (peaks 1-4).

S12C

.1

S12C

.2

S12C

.3

S12C

.4


192



S12C

.5

S12C

.6

S12C

.7

S12C

.8


193



S12C

.9

S12C

.10

S12C

.11

S12C

.12


194



S12C

.13

S12C

.14

S12C

.15

S12C

.16


195

Table S4. Peak assignment of desialylation of native bovine fetuin treated with heavily α2,3-

preferential NanB, Fig. S12C based on retention time and m/z values.

Retention time [min]

found m/z calc. m/z Fragment

S12C.1 14.6 773.93 773.81 M5 [M+2H]2+ S12C.2 20.3 748.88 - - S12C.3 20.7 773.49 773.31 A3G(?,?,?)3 [M+3H]3+

1159.60 1159.46 A3G(?,?,?)3 [M+2H]2+ S12C.4 21.5 1131.03 - - S12C.5 21.8 846.10 845.66 A2G(4)2S(??)2 [M+3H]3+ S12C.6 22.9 870.77 870.34 A3G(4)3S(?)1 [M+3H]3+ S12C.7 23.4 - - - S12C.8 24.0 968.05 967.37 A3G(4)3S(??)2 [M+3H]3+ S12C.9 24.1 967.78 967.37 A3G(4)3S(??)2 [M+3H]3+ S12C.10 24.9 967.64 967.37 A3G(4)3S(??)2 [M+3H]3+ S12C.11 25.1 967.94 967.37 A3G(4)3S(??)2 [M+3H]3+

S12C.12 25.6 1064.95 1064.38 A3G(4)3S(???)3 [M+3H]3+ S12C.13 26.1 - - - S12C.14 26.3 1064.95 1064.38 A3G(4)3S(???)3 [M+3H]3+ S12C.15 26.8 1080.03 - - S12C.16 27.3 1072.17 - -


196

Figure S23. Mass spectra of desialylation of native bovine fetuin treated with non-specific

NanH, Fig. S12D (peaks 1-4).

S12D

.1

S12D

.2

S12D

.3

S12D

.4


197

Figure S24. Mass spectra of desialylation of native bovine fetuin treated with non-specific

NanH, Fig. S12D (peaks 5-7).

S12D

.5

S12D

.6

S12D

.7


198

Table S5. Peak assignment of desialylation of native bovine fetuin treated with non-specific

NanH, Fig. S12D based on retention time and m/z values.


S12D.1 14.7 785.22 784.80 M5 [M+H+Na]2+

S12D.2 17.5 651.74 651.60 A2G(4)2 [M+3H]3+ 977.30 976.89 A2G(4)2 [M+2H]2+

S12D.3 20.5 773.65 773.31 A3G(?,?,?)3 [M+3H]3+ S12D.4 20.8 773.50 773.31 A3G(?,?,?)3 [M+3H]3+ S12D.5 21.8 870.67 870.34 A3G(4)3S(?)1 [M+3H]3+ S12D.6 23.1 870.78 870.34 A3G(4)3S(?)1 [M+3H]3+ S12D.7 24.1 967.83 967.37 A3G(4)3S(??)2 [M+3H]3+


199

Figure S25. Native bovine fetuin desialylation. (A) fetuin treated with pseudosialidase. (B)

fetuin treated with Pda2,6ST


200

Figure S26. Mass spectra of desialylation of native bovine fetuin treated with Pda2,6ST, Fig.

S25B (peaks 1-4).

S25B

.1

S25B

.2

S25B

.3

S25B

.4


201


S25B (peaks 5-8).

S25B

.5

S25B

.6

S25B

.7

S25B

.8


202


S25B (peaks 9-12).

S25B

.9

S25B

.10

S25B

.11

S25B

.12


203


S25B (peaks 13-15).

S25B

.13

S25B

.14

S25B

.15


204

Table S6: Peak assignment of desialylation of native bovine fetuin treated with Pda2,6ST,

Fig. S25B based on retention time and m/z values.


S25B.1 14.9 784.74 784.80 M5 [M+H+Na]2+ S25B.2 17.7 977.10 976.89 A2G(4)2 [M+2H]2+

S25B.3 19.1 748.82 748.63 A2G(4)2S(3)1 [M+3H]3+ 1122.69 1122.44 A2G(4)2S(3)1 [M+2H]2+

S25B.4 20.3 1122.83 1122.44 A2G(4)2S(6)1 [M+2H]2+ S25B.5 20.5 - - - S25B.6 20.8 - - - S25B.7 22.2 870.82 870.34 A3G(4)3S(3)1 [M+3H]3+ S25B.8 23.4 967.79 967.37 A3G(4)3S(??)2 [M+3H]3+ S25B.9 23.6 - 967.37 A3G(4)3S(??)2 [M+3H]3+ S25B.10 24.4 967.66 967.37 A3G(4)3S(??)2 [M+3H]3+ S25B.11 24.7 - - - S25B.12 24.8 1064.67 1064.40 A3G(4)3S(???)3 [M+3H]3+ S25B.13 25.6 1064.84 1064.40 A3G(4)3S(???)3 [M+3H]3+ S25B.14 26.4 1064.46 1064.40 A3G(4)3S(???)3 [M+3H]3+

S25B.15 26.9 1176.43 1176.06 A3S(6)1G(4,4,3)3S(???)3 [M+H+2Na]3+


205

Traveling wave ion mobility-mass spectrometry (TWIMS-MS) analysis

Sialylated glycans were analyzed according to the protocol of Kolarich et al. with slight

adjustments.S5 Samples (approximately 1-10 µM, 45 mM ammonium acetate pH 7, 25 %

DMF, 53.625 % acetonitrile in water) were infused into a Synapt G2-Si HDMS (Waters, UK)

by static nanoelectrospray ionization using pulled borosilicate emitters (World Precision

Instruments, USA, thin-wall capillary, 4” length, 1.2 mm OD). The capillary, cone voltage

and source temperature were typically set to 0.8-1.5 kV, 25 V and 40 oC respectively. No

cone gas was used. The trap DC entrance, bias and exit were set to 0, 45 and 3 V. The IM

travelling wave speed was set to 1000 m/s and the wave height set at 40 V. Nitrogen drift

gas flow was set at 90 mL/min for all experiments. The helium and argon flow were set to

180 and 2 mL/min respectively for the helium and trap cell. The trap voltage was set to 25

V for all collision-induced dissociation data and was otherwise set to 4 V for MS acquisitions.

The transfer voltage was set to 2 V throughout. The mass measurements were calibrated

using 2 mg/mL NaI calibrant in 50% Isopropyl alcohol. Drift times were calibrated to a mix

of dextran 1000 (0.1 mg/mL) and 5000 oligomers (0.5 mg/mL) in the presence of 1 mM

NaH2PO4 in 50% MeOH, whose CCS have been previously verified by DTIMS.S6

Mass spectra and ATDs were processed using MassLynx V4.1 (Waters, UK) and

OriginPro 9.1 (OriginLabs, USA) respectively. ATDs were calibrated and subsequently

normalized to their maximum intensity. Gaussian distributions were fitted to these spectra

and the center of this fitted peak was taken as the peaks CCS.


206

TWIMS-MS analysis of selected glycoforms

Selected species were also mass selected prior to tandem mass spectrometry and

measurement of the mobility of the diagnostic B3-product ion, to prove these fragments

arose from the intact glycan and not solely the presence of any of this tri-saccharide within

the sample or an alternative precursor. The α2,6-disialylated bi-antennary A2G(4)2S(6,6)2

glycan ([M+2H]2+, m/z 1268) from the bi-antennary peptido-N-glycan produced a CCS

distribution similar to the entire glycan population, namely presence of a single α2,6-Neu5Ac

terminating tri-saccharide (Figure S31). m/z 1268 was also selected within the fetuin sample,

however consisted of α2,3-Neu5Ac containing glycoforms as well as α2,6-Neu5Ac (Figure

S32). Overall, there was a greater amount of α2,6-Neu5Ac observed within these glycoforms,

similar to what was observed within the UPLC data. Interestingly, the tri-antennary tri-

sialylated glycoforms associated with m/z 1596 ([M+2H]2+) from fetuin displayed quite a

distinct CCS distribution with α2,3-Neu5Ac terminating glycoforms being as abundant as

α2,6-Neu5Ac, highlighting this methods potential ability to discern not only the α2,3/6-

Neu5Ac ratio on the global glycans released, but also specific glycoforms (Figure S33). After

treatment with the pseudosialidase, glycoforms associated with this precursor were primarily

α2,3-Neu5Ac containing further highlighting the enzyme’s activity upon native glycans and

preference towards α2,6-Neu5Ac (note m/z 1607 [M+Na+H]2+ precursor was used as it was

more abundant in this sample than m/z 1596) (Figure S34). Treatment of fetuin with α2,3-

sialidase NanB, glycoforms associated with this precursor were primarily α2,6-Neu5Ac

containing (Figure S35).


207

Figu

re S

31: F

ragm

enta

tion

of A

2-pe

ptid

e co

ntro

l and

Col

lisio

n cr

oss

sect

ion

(CC

S) d

istrib

utio

n as

soci

ated

with

the

dia

gnos

tic [

Neu

5Acα

2,3/

6-G

alβ1

,4-

Glc

NA

c+H

]+ tr

i-sac

char

ide

B 3-p

rodu

ct io

n (m

/z 6

57).


208

Figu

re S

32:

Frag

men

tatio

n of

bi-s

ialy

late

d A

2 gl

ycan

s of

fet

uin

cont

rol

and

Col

lisio

n cr

oss

sect

ion

(CC

S) d

istrib

utio

n as

soci

ated

with

the

dia

gnos

tic

[Neu

5Acα

2,3/

6-G

alβ1

,4-G

lcN

Ac+

H]+

tri-s

acch

arid

e B 3

-pro

duct

ion

( m/z

657

).


209

Figu

re S

33:

Frag

men

tatio

n of

tri -

sialy

late

d A

3 gl

ycan

s of

fet

uin

cont

rol

and

Col

lisio

n cr

oss

sect

ion

(CC

S) d

istrib

utio

n as

soci

ated

with

the

dia

gnos

tic

[Neu

5Acα

2,3/

6-G

alβ1

,4-G

lcN

Ac+

H]+

tri-s

acch

arid

e B 3

-pro

duct

ion

( m/z

657

).


210

Figu

re S

34: F

ragm

enta

tion

of tr

i- sia

lylat

ed A

3 gl

ycan

s of f

etui

n tre

ated

with

pse

udos

ialid

ase

and

Col

lisio

n cr

oss s

ectio

n (C

CS)

dist

ribut

ion

asso

ciat

ed w

ith th

e

diag

nost

ic [N

eu5A

cα2,

3/6-

Galβ

1,4-

Glc

NA

c+H

]+ tr

i -sac

char

ide

B 3-p

rodu

ct io

n (m

/z 6

57).


211

Figu

re S

35: F

ragm

enta

tion

of tr

i- sia

lyla

ted

A3

glyc

ans o

f fet

uin

treat

ed w

ith α

2,3-

spec

ific

Nan

B an

d C

ollis

ion

cros

s sec

tion

(CC

S) d

istrib

utio

n as

soci

ated

with

the

diag

nost

ic [N

eu5A

cα2,

3/6-

Galβ

1,4-

Glc

NA

c+H

]+ tr

i-sac

char

ide

B 3-p

rodu

ct io

n (m

/z 6

57).


212

Glycoprofiling of fetuin containing samples by Lectin assisted Western blots.

Samples of commercially available fetal calf fetuin and asialofetuin (both from Sigma),

fetal calf fetuin treated with pseudosialidase and fetal calf fetuin treated with nonspecific

sialidase NanH were subject to sodium dodecyl sulfate polyacrylamide gel electrophoresis

(SDS-PAGE) [4 µg load of each in 30 µL wells]. Reference gel was stained with Instant Blue

(Expedeon).

Two gels were treated for 5 min with Towbin buffer and subject to semi-dry Western blot

(WB) using PVDF membranes (BioRad) activated by methanol and presoaked in Towbin

buffer. Protein transfer to the membrane took 50 min at 15 V.

Membrane blocking was achieved by 1 h incubation at room temperature (RT) with

‘Protein-Free (TBS) Blocking Buffer’ (Pierce).

Lectin binding buffer consisted of 10 mM sodium phosphate, 150 mM NaCl, 1 % (v/v)

Tween 20, pH 7.8. The final concentration of lectin for fluorescein labelled SNA I (from

Sambucus nigraI; Vector Laboratories) binding was 7.5 µg/mL (20 mL), while the final

concentration of lectin for fluorescein labelled MAL I (from Maackia amurensis; Vector

Laboratories) binding was 10 µg/mL (20 mL). After a 1 h incubation of the membranes with

the respective lectins at RT images were prepared using Typhoon Trio imaging system (blue

laser, fluorescein specific filter, 50 µ pixels).

SNA I lectin isolated from elderberry bark, binds preferentially to sialic acid attached to

terminal galactose in α2,6- and to a lesser degree, α2,3- linkage. Binding is also inhibited to

some extent by lactose or galactose. This lectin does not appear to bind sialic acid linked to

N-acetylgalactosamine.

Maackia amurensis lectin I (MAL I) binds Gal-β1,4-GlcNAc but tolerates substitution of

N-acetyllactosamine with sialic acid at the 3 position of galactose. However, MAL I does not

appear to bind this structure when substitution with sialic acid is on the 6 position of

galactose.


213

Figure S36: Glycoprofile of fetuin containing samples by SDS-PAGE and lectin-assisted

Western blots.

The theoretical MW for non-glycosylated mature fetuin is ~36 kDa. The glycoprofile of

fetuin and asialofetuin samples varies from batch to batch. However, fetuin usually produces

a major band around 60 kDa, while asialofetuin usually produces a major band around 50

kDa with a number of lower degradation bands (rarely seen in fetuin samples).

Compared to the fetuin sample the sample treated with pseudosialidase displays a slight

overall shift towards lower MWs, while there is a clear enrichment of the lower band at the

expense of the upper band. SNA I binding is lower in the sample treated with pseudosialidase

what is in accordance with lower affinity of the lectin towards α2,3-linked Neu5Ac. There is

also an increase of MAL I binding in the sample treated with pseudosialidase due to the

increase of N-acetyllactosamine residues upon α2,6-specific desialylation.


214

The sample treated with nonspecific sialidase NanH displays an additional MW shift of

bands and generation of two additional faint low MW bands. While the two main bands with

higher MW show diminished binding with both lectins the two faint bands show a more

significant binding to both lectins. Based on HILIC-UPLC data approximately 2/5 of the

N-glycans present in the sample still contain Neu5Ac (single or two per glycan) with both

α2,3- and α2,6-linkages present. Since these faint bands are of MWs, which correspond to

protein degradation products, give stronger signal with both lectins, we believe they may be

richer in sialylated remnants.

References

[S1] Huang, K.; Wang, M. M.; Kulinich, A.; Yao, H. L.; Ma, H. Y.; Martínez, J. E.; Duan,

X. C.; Chen, H.; Cai, Z. P.; Flitsch, S. L.; Liu, L.; Voglmeir, J. Carbohydr. Res. 2015, 415,

60.

[S2] Cheng, J.; Huang, S.; Yu, H.; Li, Y.; Lau, K.; Chen, X. Glycobiology. 2010, 20, 260.

[S3] Mehr, K.; Withers, S. G. Glycobiology. 2016, 26, 353.

[S4] Campbell, M. P.; Royle, L.; Radcliffe, C. M.; Dwek, R. A.; Rudd, P. M. Bioinformatics

2008, 24, 1214.

[S5] Hinnenburg, H.; Hofmann, J.; Struwe, W. B.; Thader, A.; Altmann, F.; Varón Silva,

D.; Seeberger, P. H.; Pagel, K.; Kolarich, D. Chem. Commun. (Camb.) 2016, 52, 4381.

[S6] Hofmann, J.; Struwe, W. B.; Scarff, C. A.; Scrivens, J. H.; Harvey, D. J.; Pagel, K. Anal.

Chem. 2014, 86, 10789.

215

Chapter 6 Endogenous modulation of neuronal dopamine

transport

6.1 Summary

The final chapter discusses the effects of glycolipids and their associated catabolic

enzymes on human health. While the etiology of Parkinson’s disease is not fully understood

a correlation between Gaucher’s and Parkinson’s disease has been identified. Especially the

role of the lysosomal β-glucocerebrosidase has been of particular interest. The following

manuscript in preparation reveals a new perspective on the involvement of

glucosylsphingosine in the neurodegenerative process of dopaminergic neurons.

6.2 Contribution

As part of this thesis the general idea was development. MR and SLF designed the study.

MR designed, and TD performed computational experiments. MR provided compounds to

YMG. YMG and TL performed cell culture experiments and analysed the data. MR

interpreted the data and wrote the manuscript. All authors edited the manuscript.

6.3 Manuscript

This manuscript is prepared in the Nature Letters format. Additional experiments are in

progress. Submission is anticipated for 2018 to Nature Letters.

Chapter 6 - Dopamine transport

216

Endogenous modulation of neuronal dopamine

transport

Michel Riese,† Teodora Djikic, ‡ Yasmina Marti Gil,§ Thorsten Lau, § Partick Schloss,§ Kemal

Yelekci,‡ and Sabine L. Flitsch*†

† School of Chemistry & Manchester Institute of Biotechnology, The University of

Manchester, Manchester M1 7DN, U.K.

‡ Department for Bioinformatics and Genetics, Faculty of Science, Kadir Has University,

Istanbul, Turkey.

§ Biochemistry Laboratory, Department of Psychiatry and Psychotherapy, Central Institute

of Mental Health, Medical Faculty Mannheim, Heidelberg University, Germany.

Abstract

Parkinsons disease (PD) is the second most common neurodegenerative disorder and is

characterised by the loss of dopaminergic neurons in the substaia nigra pars compacta resulting

in altered dopamine signalling.

The most cited genetic risk factors are mutations in the gba1 gene1,2. Dysfunctional

glucocerebrosidase leads to the accumulation of glycolidips and is associated with the

lysosomal storage disorder Gauchers disease (GD).

Increased levels of glucosylsphingosine as well as reduced activity of GBA1 are found to

be linked with lower survival of neurons. However, the underlying molecular mechanism are

yet to be understood3.

In this study, we demonstrate a direct interaction between the neuronal dopamine

transporter (DAT) and glucosylsphingosine (GlcSph).

Similar to cocaine, a known inhibitor of DAT, GlcSph (IC50 = approx. 2 µM) shows a

ten-fold increased inhibition of DAT-mediated dopamine transport. With the recently


217

described model of the human DAT it was possible to gain insights into possible mechanisms

of the interaction between GlcSph and DAT. Whereas dopamine binds the open-out

conformation, GlcSph shows a much stronger binding to the open-in conformation.

The discovery of an endogenous DAT inhibitor associated with a known genetic risk

factor is the first interaction of its kind to be described and offers a molecular explanation

towards the correlated pathologies of PD and GD. The underlying mechanism needs further

investigation, but this interaction might reveal a new dimension to the field.

Introduction

Amongst neurodegenerative diseases Parkinson’s disease (PD) is the second most

common after Alzheimer’s disease. PD is age-dependent and affects 1% of people over the

age of 60 and even up to 4% in higher age-groups worldwide4. The neuropathology of PD is

characterised by the loss of dopaminergic neurons in the substaia nigra pars compacta leading to

a depletion of the neurotransmitter dopamine. Amongst the symptoms reduced facilitation

of voluntary movements (tremor, bradykinesia and rigidity) and non-motor symptoms are

prevalent4,5. Patients are treated with dopamine agonists like L-Dopa to reduce the severity

of symptoms. To the present day no cure for PD is available.

The etiology of PD is complex and still unclear. However, genetic (e.g. alpha-synuclein

and gba1) and non-genetic risk factors have been studied extensively4 suggesting oxidative

stress, inflammation, mitochondrial dysfunction and protein sorting as well as small molecule

involvement are important features of a heterogeneous pathology6.

Over the last decade, studies linking Gaucher’s disease (GD) with PD attracted a lot of

attention7. GD patients are described to have an up to thirteen-fold increased probability to

develop PD with earlier on-set2,8. The hydrolase glucocerebrosidase (GCase) is encoded by

the GBA1 gene and cleaves the glycosidic bonds of glucosylcerebrosides and –sphingosines

in lysosomes. Mutations in the gba1 gene are the main cause for the lysosomal storage


218

disorder GD where reduced glucocerebrosidase (GC) activity leads to the accumulation of

glucosylceramide (GlcCer) and -sphingosine (GlcSph). Even aging individuals with two

intact alleles of gba1 show a significant GlcSph accumulation due to reduced GCase activity9.

Some studies focussed on the formation of synucleinopathies through intracellular

interactions of GCase with its interaction partners (e.g. LIMP-2) and alpha-synuclein on the

other hand10. Others investigated the effects of increased glycolipid concentrations on

membrane dynamics and autophagy and a few targeted calcium localisation. In this context,

the roles, distribution and targets of accumulated substrates (glucosylsphingosine/-ceramide)

of GCase are yet to be understood.

Here, we present the perspective of glucosylsphingosine as a naturally occurring

metabolite interacting with crucial transporter proteins. Starting from tyrosine, dopaminergic

neurons synthesise the neurotransmitter dopamine which is stored in synaptic vesicles. The

vesicular monoamine transporter 2 (VMAT2) is responsible for the dopamine transport

across the membrane in anti-port with protons. Preloaded synaptic vesicles are located in the

axon terminal. Upon electrical stimulation, these vesicles fuse with the presynaptic

membrane to release the content into the synaptic cleft where dopamine excites postsynaptic

receptors coupled to gated ion channels and causes signal transduction (Figure 6.1A). Excess

dopamine is removed by dopamine transporter (DAT) and co-transported with Na+ into the

cytosol and either oxidised by monoamine oxidase (MAO) or redistributed into synaptic

vesicles by VMAT2.

Interference of dopamine transporters by small molecules (e.g. cocaine) is known to alter

the dopamine flux. Firstly, dopamine clearance from the synaptic cleft is effected which may

lead to overstimulation and secondly dopamine less available to be repackaged into synaptic

vesicles. To combat dopamine depletion, neurons increase the rate of neurotransmitter

synthesis. Dopamine is subsequently stored in vesicles to prevent its auto-oxidation and ROS

formation.


219

Several studies propose that DAT has the ability of the reverse transport. Following this

idea, we suggest that it is possible for endogenous compounds to bind DAT from its

cytosolic side (e.g. hDAT open-in conformation).11–13

Figure 6.1: Schematic dopamine flux in presynaptic membranes. (A) Dopamine is

synthesised from tyrosine and stored in synaptic vesicles which fuse with the pre-synaptic

membrane to release the neurotransmitter upon stimulation. The dopamine transporter

(DAT) re-uptakes dopamine into the pre-synaptic neuron for recycling. (B) Inhibition of

DAT (by cocaine and GlcSph) leads to synaptic accumulation of dopamine. (altering

dopamine homeostasis as well as inhibiting the synthesis of the precursor L-Dopa by down-

regulation of tyrosine hydroxylase).

Similar to cocaine, a known DAT inhibitor14, glucosylceramide (GlcCer) and especially its

lysofrom glucosylsphingosine (GlcSph) are small molecules with characteristic properties.

Presumably, the primary amine GlcSph is the more active compound due to its solubility.

When comparing the molecular structure of GlcSph to dopamine, the motif of a d,e-di-

hydroxy-substituted primary amine stands out (Figure 6.2). Based on this shared motif we

postulate that GlcSph is a dopamine antagonist, blocking DAT, disrupting the dopamine flux

and ultimately interfering with dopaminergic neurotransmission.


220

Results

Molecular docking

In order to assess GlcSph’s potential to act as a dopamine analogue a recent model of the

human DAT (hDAT) was used for molecular docking experiments. The homology model of

hDAT is based on the crystal structure of DAT from Drosophila spec. and is validated against

dopamine and known inhibitors.15 Similarly, a panel of amines (Figure 6.2) was docked into

hDAT’s two distinct states: open-out and open-in. The strength of each respective docking

result was calculated to give free energy ((DG) and inhibitory constant (Ki). Dopamine and

cocaine, a known inhibitor for DAT, were found to bind to hDAT with expected free

energies (DGdopamine-out = -4.57 kcal/mol and DGcocaine-out = -8.26 kcal/mol respectively, Table

S1) to the open-out state. DAT’s affinity for dopamine was reduced in the open-in

conformation and for cocaine was low (DGdopamine-in = -5.79 kcal/mol). Docking of the

glycolipids which are known to accumulate due to gba1 mutations show millimolar inhibitory

constants Ki for GlcCer in both open-in and open-out states. GlcSph shows similar levels of

association to hDAT in the open-out state compared to dopamine. Interestingly, the

calculated affinity for GlcSph (DGGlcSph = -7.58 kcal/mol, Ki(GlcSph) = 2.77 µmol, Table

S1) in the open-in conformation was higher than for dopamine and comparable to the

strength of cocaine’s binding to the open-out state (Table S1). This tight interaction between

GlcSph and hDAT, especially in its open-in state, lead to the detailed analysis of the substrate

binding site. While dopamine’s and GlcSph’s position and orientation in the open-out state

are similar with amino acids Asp79 and Tyr156 interacting with the substrate’s amino groups,

dopamine’s planar architecture results in additional hydrogen bonding between Ala77 and

Ser422 and dopamine’s hydroxyl groups. However, GlcSph’s lipid motif causes further van-

der-Waals interactions with amino acids Cys319, Phe320 and Leu485 which are unoccupied

in the case of dopamine (Fig S1).


221

In the open-in conformation, the substrates’ amine groups are coordinated by ionic

interactions between Asp79 as well as hydrogen bonding with the peptide backbone of amino

acids Ala77 and Val78. While this amine anchor locks both molecules into a similar position,

dopamine has got few additional interactions with hDAT. Leu322 is a hydrogen bond

acceptor for the hydroxyl groups and the aromatic ring shows pi-stacking with Phe76. While

these interactions are not present in case of GlcSph, its more complex architecture provides

van-der-Waals and hydrophobic interactions between the lipid tail and amino acids Trp267

and Leu418. Especially the additional hydroxyl groups of the glucosyl moiety show great

potential for extensive hydrogen bonding with amino acids Ala77, Asp421, Ser422, Gly425

and Gly426. Additionally, the alpha-hydroxyl group forms a hydrogen bond with the peptide

backbone of Leu418 (Fig. S2). Interestingly, both dopamine and GlcSph appear to be located

in a very similar position inside hDAT with close alignment of the two structures in their

docked and energy-minimised structures (Figure 6.2).


222

Figure 6.2: Substrate panel for the human dopamine transporter homology model (hDAT)

which alternates between three states: open-out, closed and open-in. The natural substrate

dopamine features a d,e-di-hydroxy-substituted primary amine motif (red) which is also

present in glucosylsphingosine. GlcCer and cocaine feature derivatives of this common

structure. The open-in docked and energy-minimised structures of GlcSph and dopamine

superimposed: the primary amine is oriented in a very similar location with polar substituents

facing in the same area.

Glucosylsphingosine and -ceramide differentially affect hDAT-dependent ASP+

uptake

These in silico findings lead to test GlcSph’s ability as an endogenous inhibitor of the

dopamine transporter. Lau and co-workers developed a comprehensive assay to probe

neurotransmitter uptake in vitro using a fluorescent monoamine transporter substrate

(ASP+)16. Cells expressing hDAT were treated with increasing concentrations of GlcSph or

cocaine as a control for uptake inhibition before hDAT-dependent ASP+ uptake was

quantified. Figure 6.3 shows the decrease in hDAT-mediated ASP+ uptake with various


223

concentrations of GlcSph. While cocaine (50 µM) exposure reduces ASP+ uptake to about

25%, 1-3 µM GlcSph is sufficient to block 50% of hDAT’s neurotransmitter uptake capacity.

Figure 6.3: GlcSph blocks the human dopamine transporter (hDAT) and thereby reduces

ASP+ uptake by transgenic cells. HEK293 cells stably expressing the human dopamine

transporter were treated with various concentrations of GlcSph or 50 µM cocaine prior to

incubation with ASP+. Subsequent quantification by confocal laser scanning microscopy

revealed a significant dose-dependent increase in ASP+ uptake by physiological

concentrations of GlcSph. As expected, cocaine significantly reduced hDAT-dependent

uptake.

Contro

l

0,1µM

GluS

ph

0,3µM

GluS

ph

1µM G

luSph

3µM G

luSph

10µM

GluS

ph

Cocain

e0

25

50

75

100

125

FASP+[%] ***

******

*** ***

***


224

Discussion

Looking at glucosylsphingosine, a naturally occurring substrate it suggests that the shared

structural motifs with the neurotransmitter dopamine could lead to biological interference.

Molecular dynamics simulations of the human dopamine transporter in complex with

substrates reveal that glucosylsphingosine could have similar binding modes compared to

dopamine. Our findings suggest two possible mechanisms of interaction:

Firstly, GlcSph could be a competitive inhibitor for dopamine in the open-out state of

DAT reducing the rate of dopamine transport across the pre-synaptic membrane.

Secondly, GlcSph demonstrates a high probability to bind to DAT in the open-in

conformation. A strong interaction could trap DAT in the final phase of the transport cycle

and prevent its return to the open-out state abolishing DAT-mediated dopamine transport

completely.

The reduction of dopamine analogue uptake in vitro underlines GlcSph’s potential to

modulate dopamine distribution with down-stream effects like dopamine accumulation

leading to altered signal clearance in the synaptic cleft and dopamine depletion in pre-synaptic

neurons. Inefficient recycling of dopamine can cause increased catecholamine metabolism

linked to cytotoxic ROS formation17.

Additionally, the vesicular monoamine transporter 2 (VMAT-2) is believed to be affected

by MPTP+ in a similar fashion18. We postulate that VMAT-2 might be targeted by GlcSph

causing further perturbation of the neuronal dopamine homeostasis and dopamine packaging

into vesicles escalating into neurotoxicity.


225

References

1. Neudorfer, O. et al. Occurrence of Parkinson’s syndrome in type 1 Gaucher disease.

QJM 89, 691–694 (1996).

2. Aharon-Peretz, J., Rosenbaum, H. & Gershoni-Baruch, R. Mutations in the

glucocerebrosidase gene and Parkinson’s disease in Ashkenazi Jews. N. Engl. J. Med.

351, 1972–7 (2004).

3. Alcalay, R. N. et al. Glucocerebrosidase activity in Parkinson’s disease with and without

GBA mutations. Brain 138, 2648–2658 (2015).

4. de Lau, L. M. & Breteler, M. M. Epidemiology of Parkinson’s disease. Lancet Neurol. 5,

525–535 (2006).

5. Chaudhuri, K. R., Healy, D. G. & Schapira, A. H. Non-motor symptoms of

Parkinson’s disease: diagnosis and management. Lancet Neurol. 5, 235–245 (2006).

6. Sampson, T. R. et al. Gut Microbiota Regulate Motor Deficits and Neuroinflammation

in a Model of Parkinson’s Disease. Cell 1–12 (2016).

7. Sidransky, E. Gaucher disease and parkinsonism. Mol. Genet. Metab. 84, 302–304

(2005).

8. Alcalay, R. N. et al. Glucocerebrosidase activity in Parkinson’s disease with and without

GBA mutations. Brain 138, 2648–2658 (2015).

9. Rocha, E. M. et al. Progressive decline of glucocerebrosidase in aging and Parkinson’s

disease. Ann. Clin. Transl. Neurol. 2, 433–438 (2015).

10. Mazzulli, J. R. et al. Gaucher disease glucocerebrosidase and α-synuclein form a

bidirectional pathogenic loop in synucleinopathies. Cell 146, 37–52 (2011).

11. Jones, S. R., Gainetdinov, R. R., Wightman, R. M. & Caron, M. G. Mechanisms of

amphetamine action revealed in mice lacking the dopamine transporter. J. Neurosci. 18,

1979–1986 (1998).

12. Leviel, V. The reverse transport of DA, what physiological significance? Neurochem. Int.

38, 83–106 (2001).

13. Fog, J. U. et al. Calmodulin Kinase II Interacts with the Dopamine Transporter C

Terminus to Regulate Amphetamine-Induced Reverse Transport. Neuron 51, 417–429

(2006).

14. Reith, M. E. A., Meisler, B. E., Sershen, H. & Lajtha, A. Structural requirements for

cocaine congeners to interact with dopamine and serotonin uptake sites in mouse brain

and to induce stereotyped behavior. Biochem. Pharmacol. 35, 1123–1129 (1986).


226

15. Djikic, T. & Yelekci, K. Human Dopamine Transporter: The first implementation of

a combined in silico/in vitro approach revealing the substrate and inhibitor specificities.

submitted (2017).

16. Lau, T., Proissl, V., Ziegler, J. & Schloss, P. Visualization of neurotransmitter uptake

and release in serotonergic neurons. J. Neurosci. Methods 241, 10–17 (2015).

17. Marchitti, S. A., Deitrich, R. A. & Vasiliou, V. Neurotoxicity and Metabolism of the

Catecholamine-Derived 3,4-Dihydroxyphenylacetaldehyde and 3,4-

Dihydroxyphenylglycolaldehyde: The Role of Aldehyde Dehydrogenase. Pharmacol.

Rev. 59, 125–150 (2007).

18. Hogan, K. A., Staal, R. G. W. & Sonsalla, P. K. Analysis of VMAT2 binding after

methamphetamine or MPTP treatment: Disparity between homogenates and vesicle

preparations. J. Neurochem. 74, 2217–2220 (2000).


227


Endogenous modulation of neuronal dopamine

transport

Michel Riese,† Teodora Djikic, ‡ Yasmina Marti Gil,§ Thorsten Lau, § Partick Schloss,§ Kemal

Yelekci,‡ and Sabine L. Flitsch*†


228

Materials and Methods

Molecular dynamics simulation

Substrates were docked in a 3D model of human dopamine transporter (hDAT) that was

previously obtained in our laboratory.144 The open-out conformation was created from

drosophila’s DAT (PDB access code: 4M48).145 The open-in conformation was modelled on

the basis of LeuT (PDB access code: 3TT3)146, using BIOVIA D.S 2016. The sequences were

aligned using the secondary structures of transmembrane proteins. Sequence similarity

between hDAT and LeuT was estimated to be 40.4%. 20 models were created using “Build

Homology Model” protocol. And they were verified using MODELLER plug-in of BIOVIA

D.S and the best model (Normalized DOPE score = -0.32) was chosen.

The energy of ligands and proteins was minimised and the structures were protonated to

pH = 7.4 in the same programme. Dopamine, cocaine, glucosylceramide and

glucosylsphingosine were docked in AutoDock (www.autodock.org).147 All the compounds

were set to be flexible, while the proteins were kept rigid. AutoDock adds a free-energy

scoring function created from a linear regression analysis, the AMBER force field, and a large

set of diverse protein-ligand complexes with known inhibition constants.148 As a centre of

binding pocket, CA atom of amino acid Phe320, was chosen. Due to the size of molecules

and number of active torsions a grid box was set to be 70 (each grid point is 0.375 Å) in all

directions for glucosylceramide and glucosylsphingosine and number of evaluation

25,000,000 was used. Whereas the grid box for dopamine and cocaine was set to 50 in all

directions and the number of evaluations was 5,000,000. For the docking studies the

Lamarckian genetic algorithm was used as search algorithm. For visualisation of non-bonded

interactions BIOVIA Discovery Studio 2016 was used.


229

Calculation of binding energies

Autodock uses a molecular mechanics model for enthalpic contributions such as vdW

and hydrogen bonding, and an empirical model for entropic changes upon binding. Each

component is multiplied by empirical weights found from the calibration against a set of

known binding constants. The scoring function of Autodock is represented in following

equation:

𝛥𝐺 = 𝛥𝐺526 ��𝐴0�𝑟0��

−𝐵0�𝑟0��

0�

+ 𝛥𝐺;{/<12�𝐸(𝑡) �𝐶0�𝑟0��

−𝐷0�𝑟0��

� +0�

𝛥𝐺898: �𝑞0𝑞�𝜀𝑟0��0�

+ 𝛥𝐺><?𝑁><? + 𝛥𝐺=<9 ��𝑆0𝑉� + 𝑆�𝑉0�𝑒{?��

�

��

0�

where ΔG stands for free energy of binding, rij is the magnitude of the distance between

i and j atoms, qi and qj are the charge at points i and j respectively, in C and ε0 is the

permittivity of a vacuum, S - solvation term for atom V –atomic fragmental volume of atom

σ – Gaussian distance constant; it is a sum of van der Walls, hydrogen bonds, electrostatics

(Coulomb's Law), torsions and desolvatation energies.

Quantification of hDAT-dependent ASP+ uptake in Glucosylsphingosine and

Glucosylceramide exposed cells

Transgenic HEK293 cells constitutively expressing hDAT were cultured as described

(HEK293-hDAT)149: cells were maintained in Dulbecco’s modified Eagle’s medium

supplemented with 10% fetal bovine serum, penicillin (100 U/mL), streptomycin

(100 µg/mL), and geneticin (200 µg/mL) at 37°C and 5 % CO2. Glucosylsphingosine

(GluSph, Sigma) or Glucosylceramide (GluCer, Avantis) were applied at 0.1, 0.3, 1, 3 or


230

10 µM for 10 min in confluent HEK293-hDAT populations. Subsequently, HEK293-hDAT

were loaded with 10 µM 4-(4-(dimethylamino)styryl)-N-methylpyridinium iodide (ASP+,

#D288, Life Technologies) for 30 s. Then, cells were washed with medium devoid of both,

dye and lipid, before being mounted for confocal microscopy to avoid background staining.

All lipids and fluorescent dye used were applied in pH- and temperature-adjusted bath

solutions to avoid temperature or pH shifts that might interfere with protein function or

transport.

ASP+ live cell imaging was performed as described before using a Leica TCS SP5

confocal imaging system attached to a DM IRE2 microscope.150,151 Images were acquired

with a HCX PL APO 63× oil planchromat lens with a NA 1.40 (Leica, Mannheim, Germany)

and a DPSS laser to excite ASP+ (561 nm). HEK-hDAT cells were kept at 37°C during

image acquisition using a microscope microheating system (Ibidi, Planegg, Germany). The

Z-stacks were acquired with sections taken every 0.5 µm and all images were exported as tiff-

files. Data analysis was performed according to Lau et al.151 Confocal z-stacks were imported

into NIH ImageJ (version 1.45s; National Institutes of Health, Bethesda, MD) to generate z-

projections and subsequent intensity quantification using the MultiMeasure plugin in defined

regions of interest (ROIs). For all ROIs, the integrated densities of the fluorescence were

determined. At least 30 cells per treatment were quantified in three independent experiments.

For data presentation and comparison of individual experiments, ASP+ fluorescence

intensity was normalized to each experiment’s control values. Statistical analysis was

performed using GraphPad Software (GraphPad Software Inc., La Jolla, USA): one-way

ANOVA and post hoc Tukey tests were calculated using experimental raw data; p < 0.05

was considered significant. The results are displayed in bar graphs as means ± SEMs.


231

Results

Molecular dynamics simulation

Table 6.1. Results of the MD simulation based on the homology model of hDAT. Substrates

were modelled into two distinct states and the free energy as well as inhibitory concentrations

were calculated.

Open-out Open-in Compound ΔG / kcal/mol Ki / µmol ΔG / kcal/mol Ki / µmol Dopamine -4.57 450.62 -5.79 56.73 Cocaine -8.26 0.88 - -

Glucosylsphingosine (GlcSph) -4.95 233.84 -7.58 2.77 Glucosylceramide (GlcCer) -2.64 11610.00 -3.96 1250.00


232

Figure S6.4. Docking of substrates into the open-out hDAT. The natural substrate

dopamine shows a binding comparable to literature values with regards to energy, orientation

and molecular interactions (top). The primary amine GlcSph is located in the same binding

pocket featuring similar interactions compared to the native one, e.g. ionic interaction

between amine and Asp79 (bottom).


233

Figure S6.5. Docking of substrates into open-in hDAT. The natural substrate dopamine

shows a binding comparable to literature values with regards to energy, orientation and

molecular interactions (top). The primary amine GlcSph is located in the same binding

pocket featuring similar interactions compared to the native one, e.g. ionic interaction

between amine and Asp79 (bottom).


234

Quantification of hDAT-dependent ASP+ uptake in glucosylsphingosine and

Glucosylceramide exposed cells

A B

C

Figure 6.6. Imaging of HEK 293 cells transfected with the human dopamine transporter

(hDAT) were treated with varying concentrations of GlcSph (and cocaine as control) before

incubation with ASP+ and subsequent image acquisition by confocal laser scanning

microscopy (B). Whereas cocaine (50 µM) reduces the hDAT-dependent ASP+ uptake to

about 25%, physiological concentrations of GlcSph cause a reduction to 75% to 60% of

control fluorescence intensities (A). Although deviations from control fluorescence

intensities were observed, over all experiments performed GlcCer treatment had no

significant effect on hDAT-dependent ASP+ uptake (C).

Contro

l

0,1µM

GluS

ph

0,3µM

GluS

ph

1µM G

luSph

3µM G

luSph

10µM

GluS

ph

Cocain

e0

25

50

75

100

125

FASP+[%] ***

******

*** ***

***

Contro

l

0,1µM

GluC

er

0,3µM

GluC

er

1µM G

luCer

3µM G

luCer

10µM

GluC

er0

25

50

75

100

125

150

FASP+[%]

Control 0.1 µM GluSph

0.3 µM GluSph 1.0 µM GluSph

3.0 µM GluSph 10 µM GluSph


235

References

1. Djikic, T. & Yelekci, K. Human Dopamine Transporter: The first implementation

of a combined in silico/in vitro approach revealing the substrate and inhibitor

specificities. submitted (2017).

2. Penmatsa, A., Wang, K. H. & Gouaux, E. X-ray structure of dopamine transporter

elucidates antidepressant mechanism. Nature 503, 85–90 (2013).

3. Krishnamurthy, H. & Gouaux, E. X-ray structures of LeuT in substrate-free

outward-open and apo inward-open states. Nature 481, 469–474 (2012).

4. Morris, G. M. et al. Software news and updates AutoDock4 and AutoDockTools4:

Automated docking with selective receptor flexibility. J. Comput. Chem. 30, 2785–

2791 (2009).

5. Cornell, W. D. et al. A Second Generation Force Field for the Simulation of

Proteins, Nucleic Acids, and Organic Molecules. J. Am. Chem. Soc. 117, 5179–

5197 (1995).

6. Hummerich, R. et al. DASB - In vitro binding characteristics on human

recombinant monoamine transporters with regard to its potential as positron

emission tomography (PET) tracer. J. Neurochem. 90, 1218–1226 (2004).

7. Martí, Y., Matthaeus, F., Lau, T. & Schloss, P. Methyl-4-phenylpyridinium

(MPP +) differentially affects monoamine release and re-uptake in murine

embryonic stem cell-derived dopaminergic and serotonergic neurons. Mol. Cell.

Neurosci. 83, 37–45 (2017).

8. Lau, T., Proissl, V., Ziegler, J. & Schloss, P. Visualization of neurotransmitter

uptake and release in serotonergic neurons. J. Neurosci. Methods 241, 10–17

(2015).

236

Chapter 7 Discussion and Outlook

Recently, the focus in glycoscience has shifted to carbohydrate active enzymes underlining

their central role in the field: CAZymes mediate the metabolism of carbohydrates and as

such are intimately involved in metabolic diseases involving glycoconjugates. Additionally,

CAZymes have become valuable tools in biotechnology, similar to restriction enzymes in

molecular biology. This thesis presents a number of projects aiming to increase the

understanding of CAZymes by developing analytical tools such as NMR and mass

spectrometry, ultimately leading to new biotechnological processes and perspectives on cell

biology.

A label-free NMR method devoid of complex sample preparations presented in Chapter 3

enables the rapid analysis of oxidation products resulting from the glycan oxidation mediated

by galactose oxidase. Oligo-saccharides up to tri-saccharides were easily assigned and

quantified. This novel activity screen complements indirect measurements of GOase activity

and reveals the identity of the product. Screening of α- and β-glucosides suggests a preference

for α-linkages. Together with the improvement in expression, GOase variants prove to be a

versatile tool for the regiospecific oxidation of glucosides and galactosides. Future efforts to

optimise the redox potential through oxygen supply and hydrogen peroxide removal should

increase yields at higher substrate concentrations. Therefore, GOase-mediated oxidation of

glucosides would be suitable for industrial application. Furthermore, the presented NMR

analysis of oxidised glycans with minimal sample preparation is a valuable tool. Compared

to Bonnet et al.’s work, the chromatographic separation of by-products can be avoided

resulting in a more streamLined analysis. Additional applications are conceivable where

regiospecific oxidation reactions of glycans are performed. The importance of 6-oxo-

saccharide derivatives is given through potential applications in biotechnology using plant-

Chapter 7 - Discussion

237

derived poly-saccharides as feedstock. With the discovery of ω-transaminases, amine

dehydrogenases and reductive aminases, a two-step enzymatic route towards amino sugars

is conceivable. These bi-functional carbohydrate derivatives serve important functions as

CAZyme inhibitors and biological probes with the addition of reporter groups. Currently

these compounds are accessed through hydrogenation of the respective azide in a multistep

synthesis with the use of protection group chemistry. Again, the NMR methodology can

provide a simplified analysis of the products formed in potential cascades.

Additionally, 6-amino-aldohexoses or possibly 5-amino-aldopentoses may find possible

application as chiral seven- or six-membered heterocyclic compounds respectively, with

hydroxyl groups configurations based on the choice of mono-saccharide. This potential

access to a variety of cyclic imines (or amines following reduction) with well-defined stereo-

centres may find applications in medicinal chemistry since a high proportion of drugs

contains cyclic chiral amines.

Figure 7.1: Retrosynthesis of chiral cyclic imines through region-specific enzymatic

oxidation and subsequent amination.

Chapter 4 discusses the adaptation of the glycan array platform on hydrophobic SAMs

provides a technology capable of screening fungal glycoside hydrolase activities on

heterogeneous carbohydrate libraries extracted from biological sources. The alternative

strategy for labelling through reductive amination of oligo-saccharide libraries devoid of

workup is simple to perform and should find broad application. Similarly, the generation of

hydrophobic SAMs on gold surfaces is accessible for most laboratories. However, the

compatibility of gold-coated MALDI targets with existing instrumentation can be a limitation

and therefore restrict the application of this glycan array.

OHOHO

OH

OH

1. GOase

2. TA, AmDH or RedAm OH

OHOHO

OH

NH2

OH

HO

HO

HO

OH

OHO

HO

HO

OH

N

NH2


238

Nevertheless, profiling hydrolytic activity can be a valuable tool to assign biochemical

functions to putative CAZymes mined from metagenome data. Initial experiments are on-

going in order to transfer the (hydrophobic) SAM glycan array technology onto an updated

gold chip format compatible with an in-house MALDI-IM-MS instrument. The addition of

ion mobility separation opens up a new dimension in glycan analysis and will vastly expand

the possibilities of the array technology presented here. In particular the analysis of partially

hydrolysed glycan libraries including branched oligo-saccharides is going to benefit

dramatically.

Conventional structural analysis of poly-saccharides is based mainly on mono-saccharide

composition, linkage analysis and chromatographic separation to determine overall size

distribution. While mono-saccharide analysis requires total hydrolysis of the poly-saccharide,

linkage information is accessible following methylation of the intact saccharide prior to

hydrolysis, subsequent acetylation and gas chromatographic separation of derivatised mono-

saccharides. The use of ion mobility separation of mono-saccharides (generated via collision

induced dissociation) and di-saccharides reveals composition and linkage information is an

elegant, in-line technology compared to time consuming chemical hydrolysis and

derivatisation.


239

Figure 7.2: Schematic of traditional saccharide analysis workflow involving multiple

derivatisation steps followed by GC-MS analysis.

In conclusion, the presented MALDI-ToF screen provides valuable information on the

hydrolytic activity towards mixed linear saccharides. The future addition of ion mobility will

expand the experimental scope and address traditional shortcomings of poly-saccharide

analysis.

While the aforementioned ion mobility separation of glycans is reserved for those with

access to state-of-the-art instrumentation, (N-)glycan analysis still relies heavily on the use of

exo-glycoside hydrolases in order to sequentially and specifically dissect analytes prior to LC-

MS analysis. The α2,6-sialyltransferase presented in Chapter 5 has been demonstrated to act

as α2,6-‘pseudo-sialidase’. The specific removal of α2,6-linked Neu5Ac from complex

glycans and even glycoproteins has potential applications in glyco-engineering settings where

remodelling therapeutic or otherwise active peptides require homogeneous glycosylation.

With the increasing importance of therapeutic glycoproteins (see Table 1.1), quality control

over the specific composition is essential for efficacy and safety of the drug.

OHOHO

OH

OH

Methylation

MeIO OHO

OH

OH

OH

OMeOMeO

OMe

OMe

O OMeO

OMe

OMe

OMe

Hydrolysis

H+OMeO

MeOOMe

OMe

HO OMeO

OMe

OMe

OMeOH

Reduction

NaBD4

CHDOH

CHOMe

MeOHC

CHOMe

CHOH

CH2OMe

CHDOH

CHOMe

MeOHC

CHOH

CHOH

CH2OMe

Acetlyation

Ac2O

CHDOH

CHOMe

MeOHC

CHOMe

CHOAc

CH2OMe

CHDOH

CHOMe

MeOHC

CHOAc

CHOAc

CH2OMe

terminal Glc 4-linked Glc


240

Furthermore, an α2,6-specific sialidase may find applications in cancer and pathogen

biology. Recent discoveries of hypersialylation in various forms of cancer suggest a

correlation between the extend of terminally sialylated N-glycans (on Fas and TNFR1) and

immune-evasive behaviour of carcinoma cells. The expression levels of the sialyltransferase

ST6Gal-I in abnormal cells is increased and may be countered by a specific α2,6-

‘pseudosialidase’.

Additionally, the human influenza virus recognises and attaches to α2,6-sialylation on host

cells, while most pathogens preferentially bind to α2,3-sialylated glycans which are common

throughout the animal kingdom. This specificity switch is another example of the crucial role

of linkage configuration in terminal sialic acids where the presented CAZyme provides a

diagnostic or potentially therapeutic value.

From an analytic perspective, the linkage discrimination performed on a mixture of

analytes is a first of its kind. Traditionally, a parent ion would have been selected, fragmented

to give the diagnostic fragment and then mobility separated. This novel approach of

fragmenting the entire ion population originating from a heterogeneous sample ensures the

sample wide absence of a Neu5Acα2,6-Galβ1,4-GlcNAc fragment. The imprecise analytical

approach has got the potential to analyse glycomes more globally and provide information

beyond the usual single ion species scope that is present in state-of-the-art ion mobility glycan

analysis.

Chapter 6 offers a novel perspective on the pathobiochemistry of Parkinson’s disease

with respect to the current literature surrounding the correlation between GCase-mediated

lysosomal deficiency and neurotoxcicity.

While the etiology of Parkinson’s disease is still not fully understood, current literature

suggests multiple aspects with aging being the central factor. A significant proportion of the

research on the correlation between Gaucher’s disease and Parkinson’s disease focusses on

the reduced activity of the lysosomal β-glucocerebrosidase, which is the most important


241

genetic risk factor and the resulting impact on the lysosomal function. Other reports describe

the critical role of the dopamine metabolism and its involvement in neuronal ROS levels.

The identification of a shared chemical motif between the glycolipid glucosylsphingosine and

the neurotransmitter dopamine lead to the discovery of an endogenous modulator of DAT-

mediated dopamine transport. The presented in silico findings are based on a three-stage

homology model of the human dopamine transporter and supported by a significant

reduction of DAT-mediated ASP+ uptake in vitro. Both models used to probe GlcSph’s

influence on DAT are validated against the known inhibitor cocaine. These conclusions form

the basis to connect the currently unrelated but simultaneously observed events of elevated

glycolipid and ROS levels, two key neurotoxicity mediators.

Following publication of the data, further experiments are anticipated targeting

intracellular ROS levels, additional in vitro experiments with neuronal cell models and in

particular VMAT-2 inhibition are planned to complement the data presented so far. The

vesicular monoamine transporter 2 is of central importance in the dopamine homeostasis

and is likely to be affected by GlcSph because of VMAT-2’s structural similarity to DAT.


242

Figure 7.3: Updated mechanistic flowchart. The conclusion from chapter 6 is a direct

linkage between elevated glucosylsphingosine (GlcSph) levels and dopamine (DA) transport

(red arrow). Possible downstream effects like altered DA metabolism or reduced

neurotransmission could explain how glycolipid metabolism and ROS-mediated

neurotoxicity are interconnected.

The diverse selection of analytical approaches presented in this thesis offers new methods

towards a better understanding of carbohydrates, underlying mechanisms and their impact

on biotechnology and human health. While the focus lies on the analysis of carbohydrates,

the results provide useful insights into CAZymes. In each chapter, the glycan subjects are

embedded in their respective enzymatic environment. Glycobiologists today have to

understand that they need to become versed in biocatalysis in order to study their chosen

subject and draw meaningful conclusions from it.

Development and application of glyco-analytical tools for ...

Documents