Page 1
doi.org/10.26434/chemrxiv.8217479.v1
Functionalized Titanium Oxide Nanowire Substrate for Surface-AssistedLaser Desorption/ionization Imaging Mass SpectrometryEwelina P. Dutkiewicz, Han-Jung Lee, Cheng-Chih Hsu, Yu-Liang Yang
Submitted date: 03/06/2019 • Posted date: 04/06/2019Licence: CC BY-NC-ND 4.0Citation information: Dutkiewicz, Ewelina P.; Lee, Han-Jung; Hsu, Cheng-Chih; Yang, Yu-Liang (2019):Functionalized Titanium Oxide Nanowire Substrate for Surface-Assisted Laser Desorption/ionization ImagingMass Spectrometry. ChemRxiv. Preprint.
Imaging mass spectrometry (IMS) is a powerful technique that enables analysis of various molecular speciesat a high spatial resolution with low detection limits. In contrast to the standard matrix-assisted laserdesorption/ionization mass spectrometry (MALDI-MS) approach, surface-assisted laser desorption/ionization(SALDI) is more effective in the detection of small molecules due to the absence of interfering backgroundsignals in low m/z ranges. We developed a functionalized TiO2 nanowire as a solid substrate for IMS oflow-molecular-weight species in biological specimens. We prepared TiO2 nanowires using the inexpensivemodified hydrothermal process and subsequently functionalized it chemically with various silane analogs toovercome the problem of superhydrophilicity of the substrate. Chemical modification changed the selectivity ofimprinting of samples deposited on the surface of the plate and thus improved the detection limits. Due to theenhanced performance, the functionalized TiO2 nanowire substrate could be successfully used for imaging ofcomplex native samples. We applied our new substrate to image distribution of the secondary metabolites in(1) petal of the medicinal plant Catharanthus roseus and (2) microbial co-culture of Burkholderia cenocepacia869T2 vs Phellinus noxius. We observed that secondary metabolites are distributed heterogeneously in apetal, which is consistent with previous results reported for the C. roseus plant leaf and stem. We verified thesemi-quantitative capabilities of the imprinting/imaging approach by comparing results using standard LC-MSanalysis of the plant extracts. Several bacteria-related metabolites produced by B. cenocepacia 869T2 inpresence of P. noxius, which were unable to be detected by MALDI-MS approach, were revealed by our newlydeveloped approach. This suggested that the functionalized TiO2 nanowire substrates-based SALDI is apowerful technique complementary to MALDI-MS.
File list (2)
download fileview on ChemRxivFunctionalized titanium oxide nanowire substrate for surfac... (2.36 MiB)
download fileview on ChemRxivSI_Functionalized titanium oxide nanowire substrate for su... (5.65 MiB)
Page 2
Functionalized titanium oxide nanowire substrate for surface-assisted laser
desorption/ionization imaging mass spectrometry
Ewelina P. Dutkiewicz1,2, Han-Jung Lee1, Cheng-Chih Hsu2*, Yu-Liang Yang1*
1Agricultural Biotechnology Research Center, Academia Sinica, Taipei, Taiwan
2 Department of Chemistry, National Taiwan University, Taipei, Taiwan
* Corresponding authors
1
Page 3
ABSTRACT
Detection and localization of low-molecular-weight species in biological specimens provide
important information about biochemical processes taking place in the organism. Imaging
mass spectrometry (IMS) is a powerful technique that enables analysis of various molecular
species at a high spatial resolution with low detection limits. In contrast to the standard
matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) approach,
surface-assisted laser desorption/ionization (SALDI) is more effective in the detection of
small molecules due to the absence of interfering background signals in low m/z ranges. We
developed a functionalized TiO2 nanowire as a solid substrate for IMS of
low-molecular-weight species in biological specimens. We prepared TiO2 nanowires using the
inexpensive modified hydrothermal process and subsequently functionalized it chemically
with various silane analogs to overcome the problem of superhydrophilicity of the substrate.
Chemical modification changed the selectivity of imprinting of samples deposited on the
surface of the plate and thus improved the detection limits. Due to the enhanced performance,
the functionalized TiO2 nanowire substrate could be successfully used for imaging of complex
native samples. We applied our new substrate to image distribution of the secondary
metabolites in (1) petal of the medicinal plant Catharanthus roseus and (2) microbial
co-culture of Burkholderia cenocepacia 869T2 vs Phellinus noxius. We observed that
secondary metabolites are distributed heterogeneously in a petal, which is consistent with
previous results reported for the C. roseus plant leaf and stem. We verified the
semi-quantitative capabilities of the imprinting/imaging approach by comparing results using
standard LC-MS analysis of the plant extracts. Several bacteria-related metabolites produced
by B. cenocepacia 869T2 in presence of P. noxius, which were unable to be detected by
MALDI-MS approach, were revealed by our newly developed approach. This suggested that
the functionalized TiO2 nanowire substrates-based SALDI is a powerful technique
2
Page 4
complementary to MALDI-MS.
1. INTRODUCTION
Imaging mass spectrometry (IMS) is a highly versatile and sensitive approach to characterize
and localize molecules in situ in biological specimens. Matrix-assisted laser
desorption/ionization imaging mass spectrometry (MALDI-IMS) is one of the imaging
techniques that enables analysis across the wide molecular mass range, at high spatial
resolution with low detection limits.1,2 MALDI employs a laser to desorb and ionize analyte
molecules mixed with the excess of the organic matrix, that facilitate the desorption/ionization
process. Typical MALDI matrices are small molecules with high UV light absorption and
good desorption/ionization abilities.3 Those traditional matrices work well for analysis of
larger molecules such as lipids, peptides, and proteins, nevertheless, their use in the analysis
of small molecules is limited due to the presence of interfering matrix-related signals in low
m/z range. To overcome this drawback, various carbon–4, semiconductor–5, and metal–
based6,7 materials have been proposed as alternatives to traditional organic MALDI matrices.8
As a matter of fact, the first approach to use the inorganic matrix for laser
desorption/ionization (LDI) was presented as early as in 1988 by Tanaka et al.9 At that time,
ultra-fine 30-nm size cobalt powder prepared in glycerol as a dispersant was used to analyze
proteins and polymers. Inspired by Tanaka, Sunner at al.4 used much larger 2-150 µm
graphite particles prepared in glycerol for analysis of low molecular weight analytes as well
as peptides and proteins at a much higher resolution. For the first time, the term – graphite
surface-assisted laser desorption/ionization (graphite SALDI) was used. Since then, a large
number of different materials have been reported as SALDI substrates with varying degrees of
performance.10,11
Back in 1988, Tanaka et al.9 have already defined characteristics of the material suitable
to assist LDI, such as strong absorption in the UV range, low heat capacity and large surface
3
Page 5
area per volume unit. Hence, metal-based materials, comprising chemically stable metal
oxides showed promising performance for SALDI-MS.12 Particularly, TiO2 powder
characterized by its strong UV absorption, mixed with a liquid medium showed the lowest
background noise and the highest intensity during the analysis of low-molecular-weight polar
and non-polar analytes.12 Furthermore, TiO2 nanomaterials were developed to increase the
surface-to-volume ratio and enhance ionization efficiency. Organic dispersive liquid medium
was eliminated to completely reduce interferences in the low mass range below m/z 1000.
TiO2 thin films13,14, nanoparticles15,16, nanotubes17,18, and nanowires19 were developed as very
effective materials for SALDI-MS.
Unlike the other TiO2 nanomaterials, the feasibility of TiO2 nanowires for SALDI-MS
has not been broadly explored due to the inconvenient manufacturing process. Conventional
nanowire synthesis is through the vapor-liquid-solid (VLS) mechanism involving chemical
vapor deposition in the presence of a catalyst under the vacuum.20,21 However, Kim et al. have
recently presented that TiO2 nanowires can be easily synthesized through the modified
hydrothermal process simply by changing aqueous solutions at ambient atmospheric
conditions.19 Nanowires can be synthesized from the planar Ti plate, which is a more
attractive approach for IMS applications. Potentially, TiO2 nanowire surface can be used as a
solid LDI-IMS substrate, although it has a notable superhydrophilic character, in which water
droplets deposited on its surface spreads completely and deteriorate the sensitivity, as well as
the lateral resolution of subsequent IMS.
In this paper, we present TiO2 nanowires prepared on a polished Ti surface through the
modified hydrothermal process as a solid substrate for LDI-IMS for the first time. Importantly,
the TiO2 nanowire surface was chemically functionalized with various silane analogs to
overcome the problem of superhydrophilicity. In addition, chemical modifications changed
the selectivity of imprinting of samples deposited on the surface of the functionalized TiO2
nanowire plate, improving the signal to noise ratios. As a result, our new LDI-IMS substrates
4
Page 6
showed a greater performance for imprinting and imaging of the distribution of metabolites in
very fragile specimens such as flower petals and agar media. Here, we demonstrate the
imaging of secondary metabolites in medicinal plant Catharanthus roseus petal and microbial
co-culture of Burkholderia cenocepacia 869T2 vs Phellinus noxius.
2. EXPERIMENTAL SECTION
Preparation of functionalized TiO2 nanowire plates
The TiO2 nanowires were synthesized on a polished Ti plate (25 × 75 mm, thickness 0.127
mm, 99.7% trace metal basis, Sigma-Aldrich, Germany) using a modified hydrothermal
process.19 The Ti plates were degreased by soaking in methanol (ACS grade, Macron Fine
Chemicals, USA) for 30 min and dried under a nitrogen stream. The etching process was
carried out by dipping the Ti plates in 10 M potassium hydroxide (BioXtra, ≥85%,
Sigma-Aldrich) prepared in deionized water at room temperature for 24 h with mild shaking.
After alkali solution treatment, the TiO2 nanowire plates were thoroughly rinsed with
deionized water and then dipped in it for 96 h at room temperature with mild shaking. The
deionized water was changed repeatedly every 24 h to efficiently remove alkali residues as
well as etched particles. Finally, the TiO2 nanowire plates were dried under a nitrogen stream
and subjected to heat treatment at 600 °C for 1 h in a muffle furnace (MF-25 P, Enshine
Scientific Corporation, Taiwan).
Following hydrothermal process, the TiO2 nanowire plate was degreased by soaking in
methanol (LC-MS grade, J.T. Baker, USA) for 30 min, dried under a nitrogen stream and
subjected to O2/Ar plasma cleaning process (Solarus model 950, Gatan, USA) for 3 min to
remove any organic contaminants and to generate surface hydroxyl groups for further
chemical modification. The TiO2 nanowire plate was then immediately immersed in silane
analogue solution [10-3 M, TiO2_1: trichloro(3,3,3-trifluoropropyl)silane (97%, Aldrich,
Steinheim, Germany); TiO2_2: trichloro(1H,1H,2H,2H-perfluorooctyl)silane (97%, Aldrich),
5
Page 7
and TiO2_3: trichlorooctadecylsilane (≥90%, Aldrich)] prepared in n-hexane (Puriss, ACS,
≥99%, Sigma-Aldrich) for 4 h at room temperature. The modified TiO2 nanowire plate was
washed in dichloromethane (Puriss, ACS, ≥99%, Honeywell Riedel-de Haën, Germany) and
consequently in 2-propanol (LC-MS grade, 99.9%, Sigma-Aldrich) for 1 h each time and
dried under a gentle stream of nitrogen. The modified TiO2 nanowire plates were prepared
weekly and stored in the drying cabinet until use.
Preparation of sandwich imprints for petal analysis
C. roseus plants were kept in the plant growth chamber (24 °C, 45% RH, 12 h of light per day).
Fresh flowers were dissected and a single petal was placed on the TiO2 nanowire plate with its
upper surface facing the plate (in one experiment, intact flower was imprinted). Then, the
petal was covered with a silica TLC plate from the top to absorb the excessive liquids leaking
from the petal during imprinting. Bottom of the TiO2 nanowire plate, as well as the TLC plate,
were affixed to the commercial stainless steel plates of the same size with conductive
double-sided tape to ensure electric contact. Petal was manually imprinted using a machine
vise press (MC power vice, VIP 100, Herbert, Taiwan) by applying a load of 1 ton over an
area of 45 × 100 mm for 2 min. Petal tissue residue was immediately removed with tweezers
from the surface of the plate after imprinting. A mixture of calibration standards was
deposited next to the imprint and left for drying in the air for a few minutes before subjecting
the sample for SALDI-IMS analysis. For comparison of different cultivars of C. roseus, a spot
of reserpine (2 µL, 5 mg/L) was deposited next to each imprint as internal standard and
imaged together with the imprint.
Preparation of vinca alkaloid standard solutions and calculation of limits of detection
Standard stock solutions of catharanthine (Cayman Chemical Company, USA), serpentine
(Toronto Research Chemicals, Canada), vindoline (≥98%, Sigma), and vinblastine (≥98%,
6
Page 8
Sigma) were prepared in amber glass vials in methanol (LC-MS grade, J.T. Baker) and
acetonitrile (LC-MS grade, J.T. Baker, USA) mixture (10:2, v/v) and further diluted with 90%
methanol (LC-MS grade, J.T. Baker). 2 µL of working sample solution was deposited on the
functionalized TiO2_2 nanowire plate. The limit of detection (LOD) was estimated to be equal
to 3-times signal-to-noise ratio (S/N) value. S/N was calculated according to the following
formulae: S/N=2.5×(S-0.5N)/N, where S is the maximum peak intensity, at the center of
specific m/z (for known concentration of a standard compound), N is noise equal to the
amplitude of the signal at the same m/z (±1.0 Da) in blank mass spectrum calculated as
3-times root-mean-square.
SALDI-MS analysis
SALDI-IMS experiments were carried out using autoflex speed MALDI-TOF-TOF mass
spectrometer equipped with the Smartbeam-IITM laser (Bruker Daltonic, Germany) operated
by flexControl (version 3.4) and flexImaging (version 3.0) software (Bruker Daltonic).
Normally, mass spectra were acquired in positive-ion mode, in m/z 160-1500 range with
deflection of the ions below m/z 160. Laser settings were as follows: 1000 Hz repetition rate,
45% intensity (with global attenuation of 20%), “large” laser size corresponding to ~ 100 µm
diameter laser footprint. Mass spectrometer settings were as follows: ion source 1: 19.00 kV,
ion source 2: 16.55 kV, lens: 5.50 kV, reflector: 20.95 kV, reflector 2: 9.50 kV, pulsed ion
extraction: 150 ns. The detector was set to 2190 V, while digitizer was set to 5 GS/s. Each
pixel in the MS image is related to a signal in a single mass spectrum acquired as a sum of
1500 laser shots. The IMS experiments of petal were performed at spatial resolution from 250
to 500 µm. The mass spectrometer was calibrated with the mixture of universal MALDI
matrix (Fluka, St. Gallen, Switzerland), reserpine (Fluka) and peptide calibration standard
(Bruker Daltonic).
7
Page 9
Additional methods are described in the Supporting Information file.
3. RESULTS AND DISCUSSION
Preparation and characterization of functionalized TiO2 nanowire plate
The TiO2 nanowires were synthesized on a polished Ti plate using the modified hydrothermal
process described before.19,22 As shown in Figure 1a, polished Ti plate was exposed to alkali
treatment at room temperature and then it was thoroughly washed with H2O. The modified
hydrothermal process is a simple procedure carried out under ambient atmospheric conditions
which makes it suitable for mass production. Afterward, TiO2 nanowire plate was subjected to
heat treatment at 600 °C to transform TiO2 lepidocrocite to anatase crystal form.19,23 It has
been already reported that properties of TiO2 nanowires manufactured according to the above
procedure are effective for ionization of various molecules. Pyun’s group prepared arrays of
TiO2 nanowires and applied them to assist analysis of amino acid and peptide standards19,24,
as well as benzylpenicillin in milk25 by LDI-MS. In our study, we adapted the TiO2 nanowires
production procedure and further introduced a chemical modification to prepare the solid
substrate for LDI-IMS (Figure 1a). To verify its efficacy, we applied the new substrate to
imprint and image distribution of the secondary metabolites in a petal of the medicinal plant C.
roseus (Figure 1b) and bacteria/fungi co-culture of B. cenocepacia 869T2 vs P. noxius
(Figure 1c).
We confirmed the successful synthesis of three-dimensional networked structure of TiO2
nanowires through investigating surface morphology by scanning electron microscopy
(Figure 2a, left). Potentially, TiO2 nanowire can be readily used as a solid substrate for IMS,
nevertheless, the wetting properties and surface termination of the substrate have a profound
impact on desorption and ionization efficiency and sensitivity. Due to the superhydrophilic
character of the TiO2 nanowire material, the analyte droplet deposited on this substrate
8
Page 10
spreads completely on the surface (Figure 2b, left). A high spreading of the droplet
containing the analyte molecules leads to a significant decrease of sensitivity of detection. To
solve this problem, Pyun’s group prepared an array of multiple TiO2 nanowire zones (300 µm
diameter) on a target plate19. Ti plate was covered with a protective parylene film, then
specific zones were masked with a layer of metal, furthermore plate was subjected to plasma
treatment to expose unmasked Ti, and finally, the metal mask was removed and plates were
etched by KOH solution. After the heat treatment process, parylene coating was removed and
as a result, an array of multiple TiO2 nanowire zones was created.19 Lo et al. stamped TiO2
nanotube layers with small SiO2 loops to restrain sample deposited inside and used them for
analysis of small molecules and peptides.17 SiO2 sol was used as a plaster on the surface of the
nanotube to create a 2-3 mm rings for restricting sample solution within them. Piret et al.
modified TiO2 nanotubes chemically with octadecyltrichlorosilane. At first, the TiO2
nanotube surface was subjected to UV/ozone treatment, then it was reacted with
octadecyltrichlorosilane and treated with UV/ozone irradiation again to adjust
hydrophobicity.18
All the mentioned approaches were developed to overcome the problem of the
superhydrophilic character of TiO2 nanomaterials with the purpose of sample spot deposition,
but not for imaging. Some of them are quite complex procedures and moreover, none of them
were tested specifically for modification of TiO2 nanowires over larger substrate area. In our
work, we simply modified TiO2 nanowire substrate chemically to adapt it for imaging
purposes. Surface of TiO2 nanowire was subjected to O2/Ar plasma to remove any organic
contaminants and to generate surface hydroxyl groups prior to chemical modification with
three different silane analogs: (1) trichloro(3,3,3-trifluoropropyl)silane, (2)
trichloro(1H,1H,2H,2H-perfluorooctyl)silane, and (3) trichlorooctadecylsilane (Figure 1a).
We noticed that the modification with trichlorosilanes was more efficient and robust than with
chlorosilanes, probably due to the higher stability of trichlorosilanes. We confirmed that
9
Page 11
chemical modification did not affect the three-dimensional networked structure of TiO2
nanowires by scanning electron microscopy (Figure 2a, right). The successful modification
was confirmed by simply depositing a water droplet on the plate and the change of surface
wettability (Figure 2b). Water droplet deposited on the non-modified plate (TiO2_0) spread
completely, while droplets deposited on the modified plates (TiO2_1, TiO2_2, and TiO2_3)
remained with high contact angles. Importantly, we observed that chemical modification
changed the selectivity of imprinting of samples deposited on the surface of the functionalized
TiO2 nanowire plate. As shown in Figure 2c, mass spectral profiles of C. roseus petal
imprinted on four different TiO2 nanowire substrates are drastically different. Although there
is only a minor difference between mass spectra corresponding to imprint on the
non-modified plate TiO2_0 and plate TiO2_1 modified with short trifluoropropyl chain, the
imprinting quality was significantly improved due to its decreased hydrophilicity (Figure S1).
While petal is pressed against a superhydrophilic surface TiO2_0, liquid released from the
plant spreads extensively on the surface that lowers the sensitivity of detection and greatly
affects the resolution of imprinting.
We noticed a high imprinting selectivity towards C. roseus secondary metabolites,
called vinca alkaloids while using plate TiO2_2, functionalized with perfluorooctyl chain
(Figure 2c). We detected high signals at m/z 335.2, 337.2, and 349.2 and lower signal at m/z
457.2 corresponding to the oxidized catharanthine, catharanthine, serpentine, and vindoline,
respectively in a petal imprint (Figure 2c). When using plate TiO2_1, we could also detect
those signals, although their S/N ratios were much lower comparing to plate TiO2_2. Notably,
choosing a nanomaterial with an appropriate affinity for a specific analyte improves detection
limits. When using plate TiO2_3, functionalized with octadecyl chain, we did not detect
signals corresponding to vinca alkaloids, although we could detect many other signals related
to plant metabolites, that could not be observed in blank mass spectra (Figure 2c).
We also verified that the amount of metabolites imprinted on TiO2 nanowire plate and
10
Page 12
their ionization efficiency is sufficient to perform successful SALDI-MS imaging. We
compared the distribution of petal metabolites imprinted on functionalized nanowire plate
TiO2_3 in assistance of nanomaterial only as well as imprint covered with a standard organic
MALDI matrix (Figure S2). We found that the distribution of vinca alkaloids in the petal
imprint in both cases was very similar, what proves that TiO2 nanowire itself is an effective
substrate for ionization of small molecules. Moreover, when using TiO2 nanowire substrate
only, we could image higher number of small molecules (molecular weight < 800 g/mol),
while in the presence of organic MALDI matrix we observed interfering matrix-related
signals in low m/z (Figure S2). On the other hand, while applying standard MALDI matrix,
signal intensities were higher and we could detect a few molecules of larger molecular
weights, which were not detected while using TiO2 nanowire only.
In another experiment, we tried to enhance ionization efficiency of molecules imprinted
on TiO2 nanowire substrate by supplying H+ ions externally. Trifluoroacetic acid (0.1%) was
sprayed over the petal imprint and we could observe the slight improvement of signal
intensities, however, we could not detect any new signals (Figure S3). Moreover, we also
investigated the possibility to rescan samples in positive and in negative ion modes. TiO2
nanowire plate can possibly act as the universal substrate for analysis in both polarity modes
since this is a matrix-free approach. During the analysis of petal extract deposited on the
surface TiO2 nanowire plate, we detected multiple signals related to plant metabolites ionized
in positive-ion mode. As expected, we could also detect several metabolite signals ionized in a
negative-ion mode such as m/z 165.0, 220.9 or 740.4 (14 signals in total), although a high
background noise derived from the substrate itself was observed (Figure S4).
Application of functionalized TiO2 nanowire substrate for IMS of C. roseus secondary
metabolites in a petal
11
Page 13
IMS has been profoundly utilized as a valuable tool for molecular spatiotemporal imaging of
natural compounds in plants.26,27 Nevertheless, investigation of the spatial distribution of
low-molecular-weight metabolites is still challenging. As we have mentioned earlier, the
application of a leading imaging technique – MALDI-IMS is limited in this case, due to the
presence of interfering matrix-related signals in low m/z range. Here, we describe a prominent
application of functionalized TiO2 nanowire substrates for IMS of C. roseus secondary
metabolites in a petal. For most of the imaging experiments, plant tissue is cryo-sectioned
(longitudinally or cross-sectioned) into thin slices of 8-50 µm. For MALDI-IMS, plant tissue
section is deposited on a conductive sample carrier (stainless steel or glass covered with
indium tin oxide), dehydrated, covered with organic matrix and dried prior to analysis.27 In
the case of petal analysis, preparation of thin planar longitudinal sections is particularly
difficult since the petal itself is too thin to section. On the other hand, depositing the intact
petal on the sample carrier may affect conductivity and lower ionization efficiency.
Alternatively, flower bud with spirally folded petals can be cross-sectioned and imaged.28
Even so, multiple sections would have to be analyzed and all the results would have to be
merged to reconstruct the distribution of molecules within the whole petal. A relatively simple
approach is to prepare petal imprint and perform the IMS directly on the imprinted substrate.
As such, the surface hydrophobicity and ionization efficient upon laser excitation of the
substrate play crucial roles in obtaining an accurate representation of molecules distribution of
the imprints.
C. roseus, also known as Madagascar periwinkle, is a flowering plant producing a vast
number of secondary metabolites that belong to the group of terpenoid-indole alkaloids, called
vinca alkaloids.29 Two of these alkaloids – vinblastine and vincristine are powerful anticancer
drugs preventing cell division, although they are biosynthesized at very low concentrations.
Vinblastine is produced mostly in plant leaves (but also in flower, stem, seeds, and root) by
coupling two precursors – catharanthine and vindoline. C. roseus is one of the most studied
12
Page 14
medicinal plants and a model organism for biotechnological studies of vinca alkaloids.29
We imprinted C. roseus petal on functionalized TiO2_2 nanowire substrate and we
detected multiple signals related to vinca alkaloids that were ionized as protonated molecules
[M+H]+ (Figure 3). We detected highly abundant ions at m/z 335.2, 337.2, and 349.2,
corresponding to the oxidized catharanthine, catharanthine, and serpentine, respectively in the
petal imprint. We also detected smaller ion signals at m/z 457.2 and 793.4 corresponding to
vindoline and precursor of vinblastine – anhydrovinblastine. Putative identification of
alkaloids was based on SALDI-MS/MS and HR LC-MS/MS experiments (Figures S5 and
S6). In the case of oxidized catharanthine, catharanthine, serpentine, and vindoline, MS/MS
spectra corresponding to molecules detected in petal extracts were compared with spectra
corresponding to commercial standards. In the case of anhydrovinblastine, due to the lack of
commercial standard, identification was based on a comparison of MS/MS spectra with data
found in the literature.30,31 Besides identified vinca alkaloids, we could detect numerous
sample-related signals in petal imprint.
We investigated the spatial distribution of several vinca alkaloids in dark pink and white
petals of C. roseus flower (Figure 3a-c). In one experiment, we also imaged the whole flower
to compare the distribution of metabolites between petals (Figure 3a). We observed that vinca
alkaloids are not distributed homogeneously in a petal (Figure 3a-c). For example,
catharanthine (m/z 337.2) accumulated in the central and end tips of a petal, while serpentine
(m/z 349.2) was distributed more homogeneously in the center of a petal. In case of vindoline
(m/z 357.2) and anhydrovinblastine (m/z 793.4), tiny amounts were distributed within the
whole petal, while slightly higher amounts were accumulated in the central tip. Distribution of
alkaloids in different petals of the same flower was quite consistent, although the relative
concentration of specific alkaloids was higher in some petals than in others (Figure 3a). Our
results were in agreement with the recent studies, reporting the cell-specific synthesis of vinca
alkaloids in C. roseus stem and leaf leading to low yield during natural biosynthesis of
13
Page 15
vinblastine.32,33 Here, we reported similar observation for petal revealed by our new method.
Although we could effectively ionize the standard compound of vinblastine (m/z 811.4)
deposited on TiO2 substrate, it was not found in the petal imprint. The possible reason may be
due to the low concentration of this metabolite in a petal or inefficiency in transferring by
imprinting. Moreover, we observed that limit of detection of vinblastine (1.43 ng) deposited
on TiO2_2 is much higher than the limit of detection of other studied metabolites –
catharanthine (0.29 ng), serpentine (0.12 ng), and vindoline (0.21 ng), suggesting that
SALDI-MS using our modified TiO2 substrate are more suitable for compounds of lower
molecular weights.
Interestingly, we observed quite severe oxidation of some vinca alkaloids in LDI-ion
source resulting in the appearance of M-1 signals in positive-ion mode. Initially, we identified
signal at m/z 355.2 as the in-source oxidized form of catharanthine, since we could detect high
signals at m/z 335.2 and 337.2 for the pure standard of catharanthine. The signal at m/z 355.2
was already reported by another group for MALDI-MS experiment of C. roseus plant sections,
but its identity was not revealed.28 Here, we proposed the structure of this species (Figure S7).
Surprisingly, we noticed that the distribution of metabolite at m/z 335.2 in a petal was
different than the distribution of the catharanthine (Figure 3a-c). To verify accuracy of the
imprinting method for semi-quantitative purposes, we prepared extracts (13.3 mg/mL) of
three sections of the petal – (1) end tip, (2) middle part and (3) central tip and subjected them
to SALDI-MS on TiO2 and standard LC-QQQ-MS analysis (Figure 3d, Table S2). Relative
amounts of vinca alkaloids obtained by the two methods were similar (for more results, see
Figure S8). Both datasets were in agreement with imaging experiment and they confirmed
our observations related to the heterogeneous spatial distribution of vinca alkaloids in a petal.
In the cases of the signal detected at m/z 337.2, catharanthine was one of the major
components, however, there was also a contribution of other isomers such as tabersonine and
vindolinine, that was revealed by LC-QQQ-MS analysis (Figure 3d). In the case of
14
Page 16
anhydrovinblastine, we observed a poor relation of relative amounts of this metabolite in three
petal sections analyzed by SALDI-MS and LC-QQQ-MS. After verification of the accuracy of
the imprinting method for semi-quantitative purposes, we hypothesized that catharanthine was
oxidized at different rates in different regions of the petal before it was harvested for IMS.
This is not surprising, as the other molecular species in the petal may possess anti-oxidative
properties, which could regulate the extent of the oxidation of catharanthine. In addition, the
in-source oxidation may also contribute to the presence of oxidized catharanthine.
Moreover, we investigated the differences in vinca alkaloid profiles between five
cultivars of C. roseus plant (Figure 3e). Four petals of each flower picked from a different
cultivar (Figure S9) were imaged and averaged mass spectra corresponding to each image
were subjected to Partial Least Squares – Discriminant Analysis (PLS-DA). We observed a
clear separation of data points corresponding to five different colors of petals (Figure 3e).
Clusters of data points related to intensively tinted petals – dark pink and dark purple were
closer to each other but far away from the cluster of data points related to white petals.
Clusters of data points related to lightly tinted petals – light purple and pink/white are
between clusters of points related to intensively tinted petals and white petals. Serpentine and
oxidized catharanthine were the most important metabolites responsible for the separation of
clusters in PLS-DA plots. Semi-quantitative comparison of all discussed vinca alkaloids in
different cultivars is presented in Figure 3f. Interestingly, the ratio of oxidized catharanthine
and catharanthine signals in different cultivars was found to be different.
As mentioned above, the distribution of vinca alkaloids in a petal is not homogenous. In
a few particular cases, we observed an interesting large metabolic feature that looked like a
“metabolic loop” in the petal, where we could clearly see the increased accumulation of
several vinca alkaloids in proximity to each other (Figure 4). Since the two crucial precursors
(catharanthine and vindoline) were present at the increased concentrations in close proximity,
we could also observe the increased concentration of the intermediates of the final product in
15
Page 17
vinblastine biosynthetic pathway (Figure 4a) in this specific region of the petal (Figure 4b).
Remarkably, we were able to detect almost all of the intermediates of the metabolic pathway
leading to the production of vinblastine. The precursors and a final product which give rise to
a “metabolic loop” were thus proposed.
Application of functionalized TiO2 nanowire substrate for IMS of microbial co-culture
Here, we describe a possible application of functionalized TiO2 nanowire substrates for IMS
of microbial co-culture. Burkholderia cenocepacia 869T2 was isolated as an endophyte and is
capable of inhibiting several phytopathogens such as Phellinus noxius, which causes the
brown root rod disease.34 In the IMS analysis of B. cenocepacia 869T2 vs P. noxius co-culture,
interestingly, we found that [M+Na]+ signals at m/z 453.2, 539.4, 625.7, and [M+K]+ signal at
m/z 727.8, corresponding to polymers of poly-(R)-3-hydroxybutyrate (Figure 5 and Figure
S10). Notably, ion signals of poly-(R)-3-hydroxybutyrate polymers were only observed by
SALDI-IMS using the functionalized TiO2_2 nanowire substrate but were not revealed by
MALDI-MS (Figure 5b). The poly-(R)-3-hydroxybutyrate polymers serve primarily as an
energy source and also enhance the resistance of bacterial cells to various stress conditions. In
addition, the unique spatial distribution of signals at m/z 969.2, 1030.6, 1269.2, and 1358.5
implies their production was interfered by the fungal cells in the co-cultural condition.
However, we were not able to detect any signals from fungal cell areas. One of the reasons is
both fungal mycelia and TiO2_2 nanowire substrate are superhydrophobic, therefore the fungal
metabolites were difficult to transfer to the TiO2_2 nanowire substrate for IMS analysis.
CONCLUSIONS
We developed a functionalized TiO2 nanowire as a solid substrate for IMS of
low-molecular-weight species. We prepared TiO2 nanowires through the inexpensive modified
16
Page 18
hydrothermal process and subsequently functionalized it with various silane analogs. New
TiO2 nanowire substrate was found to be effective to assist LDI-MS as well as IMS of small
molecules. Chemical modification improved detection limits by manipulating the
hydrophilicity of the substrate and the selectivity of imprinting of samples deposited on the
surface of the plate. The modification procedure is straightforward and cost-effective,
allowing flexibility in adjusting the properties of the substrate. Due to the enhanced selectivity,
functionalized TiO2 nanowire substrate could be successfully used for imaging of complex
native biological samples. Moreover, the same substrate can be used for analysis in both
positive and negative ion modes. Furthermore, we successfully applied our TiO2 substrate to
image distribution of the secondary metabolites in (1) petal of the medicinal plant C. roseus
and (2) microbial co-culture of B. cenocepacia 869T2 vs P. noxius. The heterogeneous
distribution of the secondary metabolites in a petal of C. roseus was detected. The
bacteria-related metabolites produced in the presence of fungi were also revealed. We verified
the semi-quantitative capabilities of the imprinting/imaging approach by comparing results
with standard LC-MS analysis of plant extracts. As for now, functionalized TiO2 nanowire is
suitable as a solid substrate for MS of small molecules, where the interferences of the organic
matrix for MALDI-MS are inevitable. We believe that our modified TiO2 substrates will have
great applicability in many other biological and clinical studies.
ASSOCIATED CONTENT
Supporting information
Additional methods (SEM, MALDI-IMS analysis for microbial sample, SALDI-IMS analysis
for microbial sample, SALDI-MS/MS, LC-MS), Table S1 (Optimized settings and retention
times for multiple reaction monitoring (MRM) LC-QQQ-MS method), Figure S1 (Imprinting
of C. roseus petal on functionalized TiO2 nanowire substrates), Figure S2 (Spatial distribution
of metabolites in C. roseus petals – comparison between petal imprint on TiO2 substrate and
17
Page 19
imprint additionally covered with MALDI matrix), Figure S3 (Spatial distribution of
metabolites in C. roseus petals – comparison between petal imprint on TiO2 substrate and
imprint additionally covered with 0.1% trifluoroacetic acid), Figure S4 (Repeated scans of the
same C. roseus petal extract sample spot in positive- and negative-ion modes), Figure S5
(SALDI-MS/MS spectra of the signal at m/z 335, 337, 349, and 457), Figure S6
(LC-HR-MS/MS spectra of white petal extract), Figure S7 (MS/MS fragmentation pattern of
catharanthine and oxidized catharanthine), Figure S8 (Comparison of relative amounts of
vinca alkaloids in extracts from three sections of white petal analyzed by SALDI-MS and
LC-QQQ-MS), Figure S9 (Flowers of five cultivars of C. roseus plant investigated in the
study), Figure S10 (LC-HR-MS/MS spectra of poly-(R)-3-hydroxybutyrate from the
microbial extract).
AUTHOR INFORMATION
Corresponding Author
*E-mail: Y.-L. Yang: [email protected]
C.-C. Hsu: [email protected]
ORCID: Y-L Yang: 0000-0002-3533-5148
C.-C. Hsu: 0000-0002-2892-5326
Author Contributions
E.P.D. prepared TiO2 nanowire substrate, optimized analysis conditions and collected most of
the data included in the manuscript, E.P.D. and H.-J.L. developed and optimized conditions of
TiO2 nanowire substrate preparation, E.P.D, C.-C.H, and Y.-L.Y designed experiments,
reviewed the data and wrote the manuscript.
Notes
The authors declare no competing financial interest.
18
Page 20
ACKNOWLEDGMENTS
This research was supported by Ministry of Science and Technology (MOST), R.O.C. (Grant
nos.: 104-2320-B-001-019-MY2, 107-2321-B-001-038- and 108-2636-M-002-008-), and
Center for Emerging Materials and Advanced Devices, National Taiwan University (NTU).
We thank W.-C. Lai, Surface and Nanoscience Laboratory, Institute of Physics, Academia
Sinica for taking SEM images, H.-J. Huang, ASCEM, Academia Sinica, for the assistance and
use of the Solarus plasma cleaner, and Metabolomics Core Facility, Agricultural
Biotechnology Research Center (ABRC), Academia Sinica, for performing LC-QQQ-MS and
LC-HR-MS analysis of petal extracts. We also thank B.-W. Wang, ABRC, Academia Sinica,
for assistance during TiO2 nanowire substrate preparation, H.-D. Cheng, ABRC, Academia
Sinica, for taking care of C. roseus plants and Y.-M. Lai, ABRC, Academia Sinica, for
collecting the IMS data of microbial co-culture samples.
REFERENCES
(1) Caprioli, R. M.; Farmer, T. B.; Gile, J. Molecular Imaging of Biological Samples:
Localization of Peptides and Proteins Using MALDI-TOF MS. Anal. Chem. 1997, 69
(23), 4751–4760.
(2) Norris, J. L.; Caprioli, R. M. Analysis of Tissue Specimens by Matrix-Assisted Laser
Desorption/Ionization Imaging Mass Spectrometry in Biological and Clinical Research.
Chem. Rev. 2013, 113 (4), 2309–2342.
(3) Baker, T. C.; Han, J.; Borchers, C. H. ScienceDirect Recent Advancements in
Matrix-Assisted Laser Desorption / Ionization Mass Spectrometry Imaging. Curr. Opin.
Biotechnol. 2017, 43, 62–69.
(4) Sunner, J.; Dratz, E.; Chen, Y.-C. Graphite Surface-Assisted Laser
Desorption/Ionization Time-of-Flight Mass Spectrometry of Peptides and Proteins from
19
Page 21
Liquid Solutions. Anal. Chem. 1995, 67 (23), 4335–4342.
(5) Wei, J.; Buriak, J. M.; Siuzdak, G. Desorption–Ionization Mass Spectrometry on
Porous Silicon. Nature 1999, 399, 243–246.
(6) Owega, S.; Lai, E. P. C.; Bawagan, A. D. O. Surface Plasmon Resonance-Laser
Desorption/Ionization-Time-of-Flight Mass Spectrometry. Anal. Chem. 1998, 70 (11),
2360–2365.
(7) Schurenberg, M.; Dreisewerd, K.; Hillenkamp, F. Laser Desorption/Ionization Mass
Spectrometry of Peptides and Proteins with Particle Suspension. Anal. Chem. 1999, 71
(1), 221–229.
(8) Shi, C. Y.; Deng, C. H. Recent Advances in Inorganic Materials for LDI-MS Analysis
of Small Molecules. Analyst 2016, 141, 2816–2826.
(9) Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.; Yoshida, Y.; Yoshida, T. Protein and Polymer
Analyses up to m/z 100 000 by Laser Ionization Time-of-Flight Mass Spectrometry.
Rapid Commun. Mass Spectrom. 1988, 2 (8), 151–153.
(10) Law, K. P.; Larkin, J. R. Recent Advances in SALDI-MS Techniques and Their
Chemical and Bioanalytical Applications. Anal. Bioanal. Chem. 2011, 399, 2597–2622.
(11) Lu, M.; Yang, X.; Yang, Y.; Qin, P.; Wu, X.; Cai, Z. Nanomaterials as Assisted Matrix
of Laser Desorption/Ionization Time-of-Flight Mass Spectrometry for the Analysis of
Small Molecules. Nanomaterials 2017, 7, 87 (4), 1–21.
(12) Kinumi, T.; Saisu, T.; Takayama, M.; Niwa, H. Matrix-Assisted Laser
Desorption/Ionization Time-of-Flight Mass Spectrometry Using an Inorganic Particle
Matrix for Small Molecule Analysis. J. Mass Spectrom. 2000, 35 (3), 417–422.
(13) Torta, F.; Fusi, M.; Casari, C. S.; Bottani, C. E.; Bachi, A. Titanium Dioxide Coated
MALDI Plate for on Target Analysis of Phosphopeptides. J. Proteome Res. 2009, 8 (4),
1932–1942.
(14) Sonderegger, H.; Rameshan, C.; Lorenz, H.; Klauser, F.; Klerks, M.; Rainer, M.; Bakry,
20
Page 22
R.; Huck, C. W.; Bonn, G. K. Surface-Assisted Laser Desorption/Ionization-Mass
Spectrometry Using TiO 2-Coated Steel Targets for the Analysis of Small Molecules.
Anal. Bioanal. Chem. 2011, 401 (6), 1963–1974.
(15) Castro, A. L.; Amorim Madeira, P. J.; Nunes, M. R.; Costa, F. M.; Florencino, H.
Titanium Dioxide Anatase as Matrix for Matrix-Assisted Laser Desorption/Ionization
Analysis of Small Molecules. Rapid Commun. Mass Spectrom. 2009, 22, 3761–3766.
(16) Lorkiewicz, P.; Yappert, M. C. Titania Microparticles and Nanoparticles as Matrixes for
in Vitro and in Situ Analysis of Small Molecules by MALDI-MS. Anal. Chem. 2009,
81 (16), 6596–6603.
(17) Lo, C.-Y.; Lin, J.-Y.; Chen, W.-Y.; Chen, C.-T.; Chen, Y.-C. Surface-Assisted Laser
Desorption/Ionization Mass Spectrometry on Titania Nanotube Arrays. J. Am. Soc.
Mass Spectrom. 2008, 19, 1014–1020.
(18) Piret, G.; Kim, D.; Drobecq, H.; Coffinier, Y.; Melnyk, O.; Schmuki, P.; Boukherroub,
R. Surface-Assisted Laser Desorption/Ionization Mass Spectrometry on Titanium
Dioxide (TiO2) Nanotube Layers. Analyst 2012, 137, 3058–3063.
(19) Kim, J.-I.; Park, J.-M.; Hwang, S.-J.; Kang, M.-J.; Pyun, J.-C. Top-down Synthesized
TiO2 Nanowires as a Solid Matrix for Surface-Assisted Laser Desorption/Ionization
Time-of-Flight (SALDI-TOF) Mass Spectrometry. Anal. Chim. Acta 2014, 836, 53–60.
(20) Huang, M. H.; Wu, Y.; Feick, H.; Tran, N.; Weber, E.; Yang, P. Catalytic Growth of
Zinc Oxide Nanowires by Vapor Transport. Adv. Mater. 2001, 13 (2), 113–116.
(21) Kang, M. J.; Pyun, J. C.; Lee, J. C.; Choi, Y. J.; Park, J. H.; Park, J. G.; Lee, J. G.; Choi,
H. J. Nanowire-Assisted Laser Desorption and Ionization Mass Spectrometry for
Quantitative Analysis of Small Molecules. Rapid Commun. Mass Spectrom. 2005, 19,
3166–3170.
(22) Kasuga, T.; Hiramatsu, M.; Hoson, A.; Sekino, T.; Niihara, K. Formation of Titanium
Oxide Nanotube. Langmuir 1998, 14, 3160–3163.
21
Page 23
(23) Castro, A. L.; Madeira, P. J. A.; Nunes, M. R.; Costa, F. M.; Florencio, M. H. Titanium
Dioxide Anatase as Matrix for Matrix-Assisted Laser Desorption/Ionization Analysis of
Small Molecules. Rapid Commun. Mass Spectrom. 2008, 22, 3761–3766.
(24) Kim, M.; Park, J.; Yun, T. G.; Noh, J.; Kang, M.; Pyun, J. TiO2 Nanowires from
Wet-Corrosion Synthesis for Peptide Sequencing Using Laser Desorption/Ionization
Time-of-Flight Mass Spectrometry. Appl. Mater. Interfaces 2018, 10, 33790–33802.
(25) Kim, J.-I.; Park, J.-M.; Noh, J.-Y.; Hwang, S.-J.; Kang, M.-J.; Pyun, J.-C. Analysis of
Benzylpenicillin in Milk Using MALDI-TOF Mass Spectrometry with Top-down
Synthesized TiO2 Nanowires as the Solid Matrix. Chemosphere 2016, 143, 64–70.
(26) Bjarnholt, N.; Li, B.; D’Alvise, J.; Janfelt, C. Mass Spectrometry Imaging of Plant
Metabolites - Principles and Possibilities. Nat. Prod. Rep. 2014, 31, 818–837.
(27) Boughton, B. A.; Thinagaran, D.; Sarabia, D.; Bacic, A.; Roessner, U. Mass
Spectrometry Imaging for Plant Biology: A Review. Phytochem. Rev. 2016, 15, 445–
488.
(28) Nakabayashi, R.; Hashimoto, K.; Toyooka, K.; Saito, K. Top-down Metabolomic
Approaches for Nitrogen-Containing Metabolites. Anal. Chem. 2017, 89, 2698–2703.
(29) Van der Heijden, R.; Jacobs, D. I.; Snoeijer, W.; Hallard, D.; Verpoorte, R. The
Catharanthus Alkaloids: Pharmacognosy and Biotechnology. Curr. Med. Chem. 2004,
11, 607–628.
(30) Hemalatha, R. G.; Pradeep, T. Understanding the Molecular Signatures in Leaves and
Flowers by Desorption Electrospray Ionization Mass Spectrometry (DESI MS)
Imaging. J. Agric. Food Chem. 2013, 61 (31), 7477–7487.
(31) Hu, B.; Wang, L.; Ye, W.-C.; Yao, Z.-P. In Vivo and Real-Time Monitoring of
Secondary Metabolites of Living Organisms by Mass Spectrometry. Sci. Rep. 2013, 3
(1), 2104.
(32) Roepke, J.; Salim, V.; Wu, M.; Thamm, A. M. K.; Murata, J.; Ploss, K.; Boland, W.; De
22
Page 24
Luca, V. Vinca Drug Components Accumulate Exclusively in Leaf Exudates of
Madagascar Periwinkle. Proc. Natl. Acad. Sci. 2010, 24, 15287–15292.
(33) Yamamoto, K.; Takahashi, K.; Mizuno, H.; Anegawa, A.; Ishizaki, K.; Fukaki, H.;
Ohnishi, M.; Yamazaki, M.; Masujima, T.; Mimura, T. Cell-Specific Localization of
Alkaloids in Catharanthus Roseus Stem Tissue Measured with Imaging MS and
Single-Cell MS. Proc. Natl. Acad. Sci. 2016, 113 (14), 3891–3896.
(34) Ho, Y.-N.; Chiang, H.-M.; Chao, C.-P.; Su, C.-C.; Hsu, H.-F.; Guo, C.; Hsieh, J.-L.;
Huang, C.-C. In Planta Biocontrol of Soilborne Fusarium Wilt of Banana through a
Plant Endophytic Bacterium, Burkholderia Cenocepacia 869T2. Plant Soil 2015, 387
(1–2), 295–306.
23
Page 25
FIGURES
Figure 1. Preparation protocol of functionalized TiO2 nanowire substrate and its applications. a) Major steps during
preparation of the TiO2 nanowire substrate. b) Application of the TiO2 nanowire substrate for IMS of C. roseus petal. c)
Application of the TiO2 nanowire substrate for IMS of the microbial co-culture of Burkholderia cenocepacia 869T2 vs
Phellinus noxius.
24
Page 26
Figure 2. Characterization of the functionalized TiO2 nanowire substrate. a) Scanning electron microscopy images of
nanowires created on the surface of the polished Ti plate before (left) and after (right) chemical modification (magn. 30,000).
b) Optical images of a water droplet deposited on the surface of TiO2 nanowire substrate before (first left) and after
modification with various silanes (top and side views). c) Average mass spectra related to images of C. roseus petal imprinted
on four different TiO2 nanowire substrates. Signal intensities were normalized with respect to total ion currents.
25
Page 27
Figure 3. Spatial distribution of vinca alkaloids in C. roseus petals. Distribution of alkaloids in a) intact small flower, b)
single dark pink petal, and c) single white petal of a bigger flower. Signals at m/z 335.2, 337.2, 349.2, 457.2, and 793.4
correspond to the oxidized form of catharanthine, catharanthine, serpentine, vindoline, and anhydrovinblastine, respectively.
Raster step was 250 µm. The color bar represents MS signal intensity. d) Comparison of relative amounts of vinca alkaloids
in extracts from three sections of white petal analyzed by SALDI-MS and LC-QQQ-MS. C – catharanthine, T – tabersonine,
V – vindolinine. Data were scaled for comparison. Error bar: SD (3 technical replicates) e) Profiling of petals from five
different cultivars of C. roseus plant. (PLS-DA scores plot, four petals of each color were imaged). f) Comparison of relative
amounts of vinca alkaloids in petals of five different cultivars of C. roseus plant. Colors: DP – dark pink, DPP – dark purple,
LPP – light purple, PW – pink/white, W – white. Error bar: 1.5 SD (four biological replicates). All the signal intensities were
normalized with respect to total ion currents (except (e), where signal intensities were normalized with respect to IS signal
intensity).
26
Page 28
Figure 4. The special features of metabolites observed in C. roseus petal. a) Biosynthetic pathway of vinblastine. b) MS
images of C. roseus petal imprint. Signals at m/z 337.2, 369.2 2, 335.2 2, 457.2 2, 791.4 2, 793.4 2, and 811.4 2 correspond to
catharanthine, intermediate of catharanthine, the oxidized form of catharanthine, vindoline, an intermediate of
anhydrovinblastine, anhydrovinblastine, and vinblastine, respectively. Notice, that vinblastine was not observed due to its low
concentration in the petal. Raster step was 250 µm. Signal intensities were normalized with respect to total ion currents. The
color bar represents MS signal intensity.
27
Page 29
Figure 5. Spatial distribution of bacterial metabolites in the co-culture of Burkholderia cenocepacia 869T2 vs Phellinus
noxius. The white colony in the center is B. cenocepacia 869T2, and the white mycelia on the right are P. noxius. a) Selected
signals related to the bacterial metabolites. Signals at m/z 453.2, 539.4, 625.7, and 727.8, correspond to
poly-(R)-3-hydroxybutyrate. Raster step was 1000 µm. The color bar represents MS signal intensity. b) Average mass spectra
(m/z 180-840) of MALDI-IMS and SALDI-IMS of the co-culture of B. cenocepacia 869T2 vs P. noxius. The signals of
poly-(R)-3-hydroxybutyrate shown in b) are highlighted. Signal intensities were normalized with respect to total ion currents.
28
Page 30
download fileview on ChemRxivFunctionalized titanium oxide nanowire substrate for surfac... (2.36 MiB)
Page 31
Supporting information
Functionalized titanium oxide nanowire substrate for surface-assisted laser
desorption/ionization imaging mass spectrometry
Ewelina P. Dutkiewicz1,2, Han-Jung Lee1, Cheng-Chih Hsu2*, Yu-Liang Yang1*
1Agricultural Biotechnology Research Center, Academia Sinica, Taipei, Taiwan 2 Department of Chemistry, National Taiwan University, Taipei, Taiwan
SUPPLEMENTARY METHODS
SEM
The morphology of the nanowires was characterized using a scanning electron microscope
instrument (Inspect F, FEI, Japan) operated at 5 kV.
MALDI-IMS analysis for microbial sample
For the microbial co-culture sample, the 10 μL of Burkholderia cenocepacia 869T2 broth was
added to the PDA agar plate and the 8 mm Phellinus noxius mycelia plug from a fresh
growing fungal plate (7 days) was placed at the opposite edge of the spot at a 1.5 cm distance.
After 4 days of incubation, areas of agar media containing the microbial colonies and fungal
mycelia were excised from the Petri dish and transferred onto the MALDI stainless steel
target plate. The universal MALDI matrix (Fluka, St. Gallen, Switzerland) was spread on top
of the agar by using HTX TM-SprayerTM (HTX Technologies LLC, USA). The parameters
used for each agars were 1200 mm/min velocity, 0.1 mL/min flow rate, 12 passes, 3 mm line
spacing and a nozzle temperature of 80 °C. Once the sample was completely covered with
matrix, it was exposed in a 37 °C incubator for overnight until it was deemed dried. MALDI-
IMS was collected using an autoflex speed MALDI-TOF-TOF mass spectrometer equipped
with the Smartbeam-IITM laser (Bruker Daltonic, Germany) operated by flexControl (version
3.4) and flexImaging (version 3.0) software (Bruker Daltonic). The mass spectra were
acquired in positive-ion mode, in m/z 100–2000 range. Laser settings were as follows: 500 Hz
repetition rate, 25% intensity (with global attenuation of 20%), “large” laser size
corresponding to ~ 100 µm diameter laser footprint. Mass spectrometer settings were as
1
Page 32
follows: ion source 1: 19.00 kV, ion source 2: 16.6 kV, lens: 8.7 kV, reflector: 21.0 kV,
reflector 2: 9.4 kV, pulsed ion extraction: 150 ns. The detector was set to 2190 V, while
digitizer was set to 5 GS/s. Each pixel in the IMS is related to a signal in a single mass
spectrum acquired as a sum of 1000 laser shots. The IMS experiments were performed at a
spatial resolution of 1000 µm. The mass spectrometer was calibrated with the mixture of
universal MALDI matrix (Fluka, St. Gallen, Switzerland) and peptide calibration standard
(Bruker Daltonic).
SALDI-IMS analysis for microbial sample
The areas of agar media containing the microbial colonies and fungal mycelia were excised
from the Petri dish and transferred onto the TiO2 nanowire plate with its upper surface
(microbial cells) facing the plate. The microbial samples were manually imprinted by
applying a load of one 500 mL beaker for 2 min. The microbial samples were immediately
removed with tweezers from the surface of the plate after imprinting. SALDI-IMS
experiments were carried out using autoflex speed MALDI-TOF-TOF mass spectrometer
equipped with the Smartbeam-IITM laser (Bruker Daltonic, Germany) operated by flexControl
(version 3.4) and flexImaging (version 3.0) software (Bruker Daltonic). The mass spectra
were acquired in positive-ion mode, in m/z 60-2500 range. Laser settings were as follows: 500
Hz repetition rate, 80% intensity (with global attenuation of 20%), “large” laser size
corresponding to ~ 100 µm diameter laser footprint. Mass spectrometer settings were as
follows: ion source 1: 19.00 kV, ion source 2: 16.55 kV, lens: 5.50 kV, reflector: 20.95 kV,
reflector 2: 9.50 kV, pulsed ion extraction: 150 ns. The detector was set to 2190 V, while
digitizer was set to 5 GS/s. Each pixel in the MS image is related to a signal in a single mass
spectrum acquired as a sum of 1000 laser shots. The IMS experiments were performed at a
spatial resolution of 1000 µm. The mass spectrometer was calibrated with the mixture of
universal MALDI matrix (Fluka, St. Gallen, Switzerland) and peptide calibration standard
(Bruker Daltonic).
SALDI-MS/MS
SALDI-MS/MS experiments were carried out in the LIFT mode using the same MALDI-
TOF-TOF instrument as for SALDI-MS and MALDI-MS experiments. Mass spectra were
acquired in positive-ion mode. Laser settings were as follows: 200 Hz repetition rate, 45%
intensity (with global attenuation of 20%), “large” laser size corresponding to ~ 100 µm
diameter laser footprint. Mass spectrometer settings were as follows: ion source 1: 6.00 kV,
2
Page 33
ion source 2: 5.30 kV, lens: 3.00 kV, reflector: 27.00 kV, reflector 2: 11.70 kV, Lift 1: 19.00
kV, Lift 2: 4.80 kV, pulsed ion extraction: 120 ns. The detector was set to 2225 V, while
digitizer was set to 5 GS/s. In fragmentation mode, detector gain boost was set to 180%, laser
power boost to 100%, and analog offset to -1.0%. PLMS was “on”. The isolation window was
set to -3+1 Da for parent ion at m/z 335.2, -1+3 Da for parent ion at m/z 337.2, and -3+3 for
remaining parent ions. MS/MS spectra presented in Figure S5 were acquired as a sum of
1500 laser shots in parent-ion mode and 1500 shots in fragment-ion mode.
LC-MS
LC-Orbitrap-MS
The ACQUITY ultra-high performance liquid chromatography system (Waters, USA), fitted
with ACQUITY UPLC BEH C18 column (100 mm × 2.1 mm, 1.7 μm; Waters), was used to
separate vinca alkaloids. Mobile phase A consisted of 0.1% formic acid in the water, while
mobile phase B consisted of 0.1% FA in acetonitrile. Gradient elution was as follows: 95% A
at 0 min, 0.5% A at 6 min, 0.5% A at 8 min, 95% A at 8.2 min, and 95% A at 10 min. The flow
rate was kept at 0.4 mL/min, column temperature was set to 40 °C, injection volume was 10
µL. The Orbitrap Elite mass spectrometer equipped with HESI ion source (Thermo Fisher
Scientific, USA) was used as a detector. MS was operated in a positive-ion mode in m/z 100-
1500 range at 15000 resolution. Source voltage was set to 3.5 kV.
LC-QQQ-MS
The ACQUITY ultra-high performance liquid chromatography system (Waters), fitted with
ACQUITY UPLC BEH C18 column (100 mm × 2.1 mm, 1.7 μm; Waters), was used to
separate vinca alkaloids. Mobile phase A consisted of 0.1% formic acid in the water, while
mobile phase B consisted of 0.1% FA in acetonitrile. Gradient elution was as follows: 95% A
at 0 min, 0.5% A at 6 min, 0.5% A at 8 min, 95% A at 8.2 min, and 95% A at 10 min. The flow
rate was kept at 0.4 mL/min, column temperature was set to 40°C, injection volume was 1 µL.
The Xevo TQ-S mass spectrometer (Waters) equipped with ESI ion source was used as a
detector. MS was operated in positive-ion mode. Source voltage was set to 3.3-3.4 kV.
Optimized settings and retention times for multiple reaction monitoring (MRM) are shown in
Table S1.
3
Page 34
SUPPLEMENTARY TABLES
Table S1. Optimized settings and retention times for multiple reaction monitoring (MRM)
LC-QQQ-MS method.
Compound Collision voltage (V) Monitored transition RT (min.)
Anhydrovinblastine 40 793 > 563 2.88
Ajmalicine 20 353 > 210 2.61
Catharanthine 15 337 > 173 2.69
Oxidized catharanthine 15 335 > 228 2.66
Reserpine (IS) 30 609 > 397 3.32
Serpentine 35 349 > 261 2.66
Tabersonine 30 337 > 180 2.60
Vinblastine 50 811 > 224 2.69
Vindoline 50 457 > 173 2.88
Vindolinine* 20 337 > 320 2.25
19S Vindolinine* 20 337 > 320 2.15
Yohimbine 20 355 > 212 2.29
*Vindolinine and 19S Vindolinine were quantified together (peak areas were summed up).
4
Page 35
SUPPLEMENTARY FIGURES
Figure S1. Imprinting of C. roseus petals on functionalized TiO2 nanowire substrates. Images were taken before removing petal residues after imprinting. Arrows indicate liquid released from the petal and spread extensively on a highly hydrophilic surface.
5
Page 36
Figure S2. Spatial distribution of metabolites in C. roseus petals – comparison between petal imprint on TiO2 substrate (TiO2 only) and imprint additionally covered with MALDI matrix (TiO2 and matrix). a) Optical images of white and dark pink petals and their imprints on TiO2_2 plate, b) distribution of vinca alkaloids discussed in the main text, c) distribution of metabolites for which ionization was enhanced in presence of matrix, d) distribution of metabolites for which images can be obtained with TiO2 only, e) distribution of other interesting metabolites related to petal color (the same result for TiO2 only and TiO2 and matrix). Imprints were covered with a layer of universal MALDI matrix (10 mg/mL) by HTX TM-SprayerTM (one pass, 0.1 mL/min, lines 1 mm away; HTX Technologies LLC, USA). Raster step was 400 µm. Signal intensities were normalized with respect to total ion currents.
6
Page 37
Figure S3. Spatial distribution of metabolites in C. roseus petals – comparison between petal imprint on TiO2 substrate (TiO2 only) and imprint additionally covered with 0.1% trifluoroacetic acid (TiO2 and TFA). a) Optical images of white and dark pink petals and their imprints on TiO2_2 plate, b) distribution of vinca alkaloids discussed in the main text, c) distribution of metabolites for which ionization was enhanced in presence of TFA. Imprints were covered with a layer of TFA (0.1% in 90% methanol) by HTX TM-SprayerTM (one pass, 0.1 mL/min, lines 1 mm away; HTX Technologies LLC, USA). Raster step was 400 µm. Signal intensities were normalized with respect to total ion currents.
7
Page 38
Figure S4. Repeated scans of the same C. roseus petal extract sample spot in positive- and negative-ion modes. Raw mass spectra for blank and white petal extract (50 mg/mL) sample spot collected in a) positive- and b) negative-ion modes deposited on TiO2_2 plate. c) Images of sample spots of white and dark pink petal extracts. Signals related to plant metabolites detected in negative-ion mode – some of the signals are specific to the color of petal. All new signals are highlighted in spectra presented in b). Raster step was 400 µm. Signal intensities were normalized with respect to total ion currents.
8
Page 39
Figure S5 a-b. SALDI-MS/MS spectra of the signal at m/z 335. a) Mass spectra scaled to a base peak, b) magnified spectra. Upper spectra correspond to catharanthine standard (10 mg/L), while lower spectra correspond to white petal extract (50 mg/mL) deposited on TiO2_2 plate. Spectra collected in positive-ion mode.
9
Page 40
Figure S5 c-d. SALDI-MS/MS spectra of the signal at m/z 337. a) Mass spectra scaled to a base peak, b) magnified spectra. Upper spectra correspond to catharanthine standard (10 mg/L), while lower spectra correspond to white petal extract (50 mg/mL) deposited on TiO2_2 plate. Spectra collected in positive-ion mode.
10
Page 41
Figure S5 e-f. SALDI-MS/MS spectra of the signal at m/z 349. a) Mass spectra scaled to a base peak, b) magnified spectra. Upper spectra correspond to serpentine standard (10 mg/L), while lower spectra correspond to white petal extract (50 mg/mL) deposited on TiO2_2 plate. Spectra collected in positive-ion mode.
11
Page 42
Figure S5 g-h. SALDI-MS/MS spectra of the signal at m/z 457. a) Mass spectra scaled to a base peak, b) magnified spectra. Upper spectra correspond to vindoline standard (10 mg/L), while lower spectra correspond to white petal extract (50 mg/mL) deposited on TiO2_2 plate. Spectra collected in positive-ion mode.
12
Page 43
Figure S6. HR LC-MS/MS spectra of white petal extract. MS/MS spectra of signals at a) m/z 335.2, b) m/z 337.2, c) m/z 349.2, d) m/z 457.2, and e) m/z 793.4. Spectra collected in positive-ion mode.
13
Page 44
Figure S7. MS/MS fragmentation pattern of catharanthine and oxidized catharanthine.
14
Page 45
Figure S8. Comparison of relative amounts of vinca alkaloids in extracts from three sections of white petal analyzed by SALDI-MS and LC-QQQ-MS. Signals at m/z 353.2, 355.2, and 811.4 correspond to ajmalicine, yohimbine, and vinblastine, respectively. Data were scaled for comparison. Error bar: SD (three technical replicates).
15
Page 46
Figure S9. Flowers of five cultivars of C. roseus plant investigated in the study.
16
Page 47
Figure S10. HR LC-MS/MS spectra of poly-(R)-3-hydroxybutyrate from the microbial extract. MS/MS spectra of signals at a) m/z 453.2, b) m/z 539.2, c) m/z 625.2, d) and m/z 727.3. Spectra collected in positive-ion mode.
17
Page 48
download fileview on ChemRxivSI_Functionalized titanium oxide nanowire substrate for su... (5.65 MiB)