Enhanced secretory production of hemolysin-mediated cyclodextrin glucanotransferase in Escherichia coli by random mutagenesis of the ABC transporter system

Page 1

Egt

KFa

b

c

d

a

ARRA

KACDEE

1

thtttepic(taptru

0d

Journal of Biotechnology 150 (2010) 453–459

Contents lists available at ScienceDirect

Journal of Biotechnology

journa l homepage: www.e lsev ier .com/ locate / jb io tec

nhanced secretory production of hemolysin-mediated cyclodextrinlucanotransferase in Escherichia coli by random mutagenesis of the ABCransporter system

heng Oon Lowa, Nor Muhammad Mahadib, Raha Abdul Rahimc, Amir Rabud,arah Diba Abu Bakard, Abdul Munir Abdul Muradd, Rosli Md. Illiasa,∗

Department of Bioprocess Engineering, Faculty of Chemical Engineering, Universiti Teknologi Malaysia, 81310 Skudai, Johor, MalaysiaMalaysia Genome Institute, Universiti Kebangsaan Malaysia, Bangi, Selangor, MalaysiaDepartment of Cell and Molecular Biology, Faculty of Biotechnology and Biomolecular, Sciences, University Putra Malaysia, MalaysiaSchool of Biosciences & Biotechnology, Faculty of Science & Technology, Universiti Kebangsaan Malaysia, Bangi, Selangor, Malaysia

r t i c l e i n f o

rticle history:eceived 20 April 2010eceived in revised form 1 October 2010ccepted 3 October 2010

a b s t r a c t

The hemolysin transport system was found to mediate the release of cyclodextrin glucanotransferase(CGTase) into the extracellular medium when it was fused to the C-terminal 61 amino acids of HlyA(HlyAs61). To produce an improved-secretion variant, the hly components (hlyAs, hlyB and hlyD) wereengineered by directed evolution using error-prone PCR. Hly mutants were screened on solid LB-starch

eywords:lpha-hemolysin transport systemyclodextrin glucanotransferaseirected evolution

plate for halo zone larger than the parent strain. Through screening of about 1 × 104 Escherichia coliBL21(DE3) transformants, we succeeded in isolating five mutants that showed a 35–217% increase in thesecretion level of CGTase-HlyAs61 relative to the wild-type strain. The mutation sites of each mutant werelocated at HlyB, primarily along the transmembrane domain, implying that the corresponding region wasimportant for the improved secretion of the target protein. In this study we describe the finding of novel

e for

xtracellular secretionscherichia coli

site(s) of HlyB responsibl

. Introduction

Understanding the mechanism of bacterial protein transloca-ion has uncovered new insights into the secretory production ofeterologous proteins. Improvements in gene expression and inhe host strain have increased overall recombinant protein produc-ion in Escherichia coli (Choi and Lee, 2004). However, the capacityo secrete significant amounts of proteins into the extracellularnvironment may be limited by insufficient capacity of the trans-ort machinery (Sung et al., 2010). Therefore, applied research has

ncreasingly focused on devising strategies to enhance the effi-iency of heterologous protein secretion. Recently, Sommer et al.2009, 2010) demonstrated an effective extracellular secretion sys-em using bacteriocin release protein (BRP) with the benefit offfinity purification using maltose binding protein (MBP). Targetrotein fused to MBP was translocated to the periplasmic space via

he N-terminal signal-dependent Sec or Tat pathway and was theneleased to the medium by permeabilisation of outer membranesing BRP.

∗ Corresponding author. Tel.: +60 7 5535564; fax: +60 7 5581463.E-mail address: [email protected] (R. Md. Illias).

168-1656/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.jbiotec.2010.10.001

enhancing secretion of CGTase in E. coli.© 2010 Elsevier B.V. All rights reserved.

The secretion of recombinant proteins into the surroundingmedium is preferable to cytoplasmic or periplasmic expressionfor several reasons. Extracellular secretion enables the synthesisof active proteins by preventing the formation of inactive proteinaggregates (inclusion bodies) and the proteolytic degradation ofproteases. Both are major problems when high levels of proteinexpression are desired, even though the target protein is directedto the periplasmic space (Sriubolmas et al., 1997; Baneyx, 1999).Additionally, the extracellular environment is relatively conduciveto oxidation, which may aid in the proper folding of proteins byfacilitating the formation of disulfide bonds. Furthermore, extra-cellular protein secretion could ease the downstream purificationprocess and eventually lead to much higher protein yields.

Gram-negative bacteria secrete a wide range of proteins throughtheir inner and outer membranes. Several pathways are dedicatedto these tasks; they are generally characterised as either Sec-dependent or Sec-independent pathways (Kostakioti et al., 2005).The type I secretion pathway is Sec-independent and allows for theexport of proteins directly from the cytoplasm into the extracellu-

lar medium, bypassing the formation of a periplasmic intermediate.One of the most extensively studied type I secretion systems isthe alpha-hemolysin (HlyA) system found in uropathogenic E. coli.Unlike in the Sec pathway, the secretion signal for hemolysin islocated at the C-terminus of HlyA at amino acid 50–60, and it

Page 2

4 iotechnology 150 (2010) 453–459

iKb1(HNstfHto2s2st

HtaPuscn(pbHrwFrism2Ep2aw

irasmwpp

2

2

abgta(

Table 1Oligonucleotides used in this study.

Name Sequence (5′–3′)a

FHlyAs GGCCGGATCCGATGCATTAGCCTATGGAAGTRHlyAs GCGCGAAGCTTTTATGCTGATGCTGTCAAAGpET AsBD R GGCCGAGCTCTTAACGCTCATGTAAACTTTCTGTFHlyAs218 GCGCGGATCCGCGGAAATTCTCTTGCAAAAAATGTATTATCCGT-F GCGGTACCGACGTAACAAACAAAGTCCGT-R GGCCGGTACCCCAATTAATCATAACCGTCUT-F ATATGGTACCTCGTCCACTCGCAACGACUT-R ATATGGTACCACCAATGCGGGCCTGGAGGAATL448F-R GATAACTTTCAGTTGGAGAGTTAAACACATCACCAAGGCGGGL448F-F CCCGCCTTGGTGATGTGTTTAACTCTCCAACTGAAAGTTATCF175L-R ACCACCTGAAGAAAAAGGGGF175L-F CCCCTTTTTCTTCAGGTGGTV621A-F GCAAGGGCGCTGGCGAACAACCCTAAAATACTCV162A-F ATTGAAACCCTTGTTGTATCTGCTTTTTTAV162A-R GTAAAAAAGCAGATACAACAAGGGTTTCAATV162A-F(S) CCCTTGTTGTATCCGCTTTTTTACAATTATTTGCV162A-R(S) GCAAATAATTGTAAAAAAGCGGATACAACAAGGGBamF CCATGGCTGATATCGGATCCEcoR CCAGATGAGCGGGTAAGAATTCEcoF CCGGTACGTGAAAAGGACGAAAATGAATTC

54 K.O. Low et al. / Journal of B

s not removed after being transported (Mackman et al., 1987;oronakis et al., 1989). The hemolysin secretion is carried outy an inner membrane (HlyB-HlyD) complex (Thanabulu et al.,998) with the trimeric outer membrane (OM)-anchored TolCWandersman and Delepelaire, 1990; Thanabulu et al., 1998).lyB is an ATP-binding cassette (ABC) protein containing a large-terminal transmembrane domain (TMD) fused to a highly con-

erved C-terminal nucleotide-binding domain (NBD). It recogniseshe C-terminal of HlyA and harnesses the energy via ATP hydrolysisor the secretion of the target protein (Benabdelhak et al., 2003).lyD is a membrane fusion protein (MFP) that has a single N-

erminal TMD linked to a large periplasmic domain. The interactionf the signal sequence with HlyD recruits TolC (Balakrishnan et al.,001) and triggers the formation of a contiguous exit channel thatpans the cell envelope (Thanabulu et al., 1998; Balakrishnan et al.,001). Importantly, recombinant proteins fused to the HlyA signalequence (HlyAs) can be recognised and secreted by the hemolysinranslocator (Fernández et al., 2000; Fraile et al., 2004).

Mutational studies have been performed on HlyAs, HlyB andlyD to identify the important elements required for successful

ranslocator complex assembly and protein secretion (Chervauxnd Holland, 1996; Hui and Ling, 2002; Holland et al., 2005;iementa et al., 2005). Hui and co-workers (2000, 2002), via sat-ration mutagenesis, demonstrated that disruption of amphiphilictructure of the first �-helix of HlyAs or reduction of non-positiveharged residue at the C-terminal final eight residues of HlyAs sig-ificantly reduces HlyA secretion. On the other hand, Schulein et al.1994) showed that substitution at the last C-terminal 33 residues,articularly L475, E477 and R478, and deletion of the region inetween L127 and L170 of HlyD, significantly reduces or abolisheslyA secretion. Recently, six residues distributed at the periplasmic

egion of HlyD, discovered using chemical random mutagenesis,ere identified to reduce HlyA secretion (Piementa et al., 2005).

or HlyB, Blight et al. (1994a) and Sheps et al. (1995), via chemicalandom mutagenesis, have identified regions in HlyB, particularlyn the ATPase domain, when substituted significantly reduces HlyAecretion. However, fewer studies have focused on the enhance-ent of protein secretion (Eom et al., 2005; Sugamata and Shiba,

005) by engineering the translocator machinery. Related to that,om et al. (2005) performed error-prone PCR on a lipase ABC trans-orter of Pseudomonas fluorescens and produced four mutants with.6–3.2 fold increase in lipase secretion level. Similarly, Sugamatand Shiba (2005) isolated two HlyB mutants, via PCR mutagenesis,ith 27–15 fold higher secretion level of subtilisin E.

Cyclodextrin glucanotransferase (CGTase) is an industriallymportant enzyme that produces cyclodextrin (CD) when CGTaseeacts with starch. CD has wide applications in the food, cosmeticnd pharmaceutical industries. In this paper, we describe the con-truction of a heterologous, secretory CGTase expression systemediated by the hemolysin transport system in E. coli. The systemas further improved by the use of directed evolution via error-rone PCR to create Hly variants exhibiting improved secretoryroduction of recombinant CGTase.

. Materials and methods

.1. Bacterial strains, DNA manipulation and growth conditions

E. coli strains used in this study were as follows: JM109 was useds the cloning host; BL21 (DE3) (Novagen) was used for recom-

inant protein production; and J96 contains the alpha-hemolysinene coding region (ATCC). Standard recombinant DNA manipula-ion techniques and PCR were performed as described (Sambrooknd Russell, 2001). E. coli strains were grown in liquid Luria–BertaniLB) medium or on solid LB agar plates at 37 ◦C. For secretion analy-

SacR CAAGCTTGTCGACGGAGCTC

a Relevant restriction sites used for cloning of PCR products are underlined.

sis, cells were grown in 2× LB broth at 200 rpm at 30 ◦C. Kanamycin(30 �g/ml) was added to the growth medium as needed.

2.2. Plasmid construction

All oligonucleotide primers used in this study are listed inTable 1. The coding region of the mature CGTase gene (GenebankAccession No. AY770576) lacking a stop codon was PCR cloned fromour laboratory clone (Ong et al., 2008) using primers CGT-F andCGT-R. The resulting 2022-bp DNA fragment was end digested withKpnI and ligated into the corresponding site of pET-29(a) (Novagen),yielding pCGT23. The coding region of the hly gene, from the 5′

end of C-terminal 61 amino acid of HlyA (HlyAs61) to the 3′ end ofhlyD, was PCR cloned from total DNA isolated from E. coli J96 usingprimers FHlyAs and pET AsBD R. The resulting 3838-bp DNA frag-ment was end digested with BamHI and SacI before ligation into thecorresponding sites of pCGT23, yielding pCGT-AsBD. The DNA frag-ment encoding Glomerella cingulata cutinase cDNA was amplifiedby PCR using primers CUT-F and CUT-R from a laboratory clone (AbuBakar et al., 2001). The resulting 579-bp DNA fragment was enddigested with KpnI and substituted into the corresponding site ofpCGT-AsBD and p079-08F, yielding pCUT-AsBD and pCUT-07908F,respectively. Overlap-extension PCR (Urban et al., 1997) was per-formed to construct an expression plasmid with the C-terminal218 amino acid of HlyA (HlyAs218). Briefly, the coding region of theHlyAs218 gene was PCR amplified from total DNA isolated from E.coli J96 using primers FHlyA218 and RHlyAs. The resulting 657-bpDNA fragment and pCGT-AsBD were used as templates with outerprimers FHlyAs218 and pET AsBD R for full length gene ampli-fication. The resulting 4306-bp DNA fragment was end digestedwith BamHI and SacI before ligation into the corresponding sites inpCGT23, yielding pCGT-AsBD218. To show the dependency of theABC transporter on recombinant CGTase or cutinase secretion, thehlyB and hlyD genes were not included in the plasmid construct.Briefly, HlyAs61 was PCR cloned from plasmid pCGT-AsBD usingprimers FHlyAs and RHlyAs. The resulting 186-bp DNA fragmentwas end digested with BamHI and HindIII before ligation into the

corresponding sites of pCGT23, yielding pCGT-As. The DNA frag-ment encoding cutinase was amplified as described above, digestedwith KpnI and substituted into the corresponding site of pCGT-As,yielding pCUT-As.

Page 3

K.O. Low et al. / Journal of Biotechnology 150 (2010) 453–459 455

F to rane

psftptVVmep

n

2

wuei72Dwt

2

c0kss

2

acstborTuf

2

((

3.3. Characterisation of mutant variants

The relative CGTase activity of the variants is shown in Fig. 4.Variant 079-08F showed the highest increase of secretion level

ig. 1. Expression plasmid construct of pCGT-AsBD and the target region subjectednzyme sites, are indicated by arrows.

Site directed mutagenesis of the hlyB gene for plasmids pL448F,L448F218, pFA, pAB, pCA, pFC, pFCB, pFAB and pFCAB were con-tructed by overlapping extension PCR (Andreas et al., 1997). Theollowing oligonucleotides were designed to insert codon substitu-ions into the template with outer primers BamF (or FHlyAs218 forL448F218 and p101-12C218) and EcoR: L448F-R and L448F-F forhe L448F mutation, F175L-R and F175L-F for the F175L mutation,162A-F and V162A-R for the V162A mutation, V162A-F(S) and162A-R(S) for V162A with silent mutation (t483c and t639a). Plas-id p101-12C218 was constructed via megaprimer PCR (Colosimo

t al., 1999) with mutagenic oligonucleotide V621A-F and EcoR, andCGT-AsBD218 as template for full length amplification.

All of the DNA sequences amplified by PCR were confirmed byucleotide sequencing.

.3. Random mutagenesis and mutant library construction

Random mutagenesis of the hly gene (HlyAs61, HlyB and HlyD)as performed using error-prone PCR (epPCR) from pCGT-AsBDsing the following four primers: BamF, EcoR, EcoF and SacR.pPCR was performed using a 100 �l reaction mixture contain-ng 160 mM (NH4)2SO4, 670 mM Tris–HCI, pH 8.0, 0.1% Tween 20,mM MgCI2, 1 mM dCTP, 1 mM dTTP, 0.2 mM dATP, 0.2 mM dGTP,5 pmol primers (each), 5 U Taq DNA polymerase, ∼5 ng templateNA and 0.5 mM MnCI2. The resulting mutagenic DNA fragmentas end digested with appropriate restriction enzymes and substi-

uted for the corresponding region of pCGT-AsBD.

.4. Screening for improved secretion activity

E. coli BL21 (DE3) harbouring mutagenised pCGT-AsBD wasultured on solid LB agar plates containing 0.5% soluble starch,.5 mM isopropyl-�-d-thiogalactopyranoside (IPTG) and 30 �g/mlanamycin at 30 ◦C. The amount of secreted CGTase-HlyAs61 wasemiquantitatively estimated by comparing the sizes of clear halosurrounding the CGTase-secreting colonies.

.5. Assay of cyclodextrin glucanotransferase (CGTase) activity

CGTase activity was determined using the phenolphthaleinssay previously described (Ho et al., 2005) with a slight modifi-ation. The reaction mixture containing 1.0 ml of 4% (w/v) solubletarch in 0.1 M phosphate buffer (pH 6.0) and 0.5 ml of enzyme solu-ion was incubated at 60 ◦C for 15 min. The reaction was stoppedy the addition of 5.0 ml of 0.03 M NaOH. Five hundred microlitersf 0.02% (w/v) phenolphthalein in 5 mM Na2CO3 were added to theeaction mixture and incubated at room temperature for 15 min.he reduction in colour intensity was measured at 550 nm. Onenit of enzyme activity was defined as the amount of enzyme thatormed 1 �mol �-CD/min.

.6. SDS-PAGE and Western blot

Sodium dodecyl sulfate-polyacrylamide gel electrophoresisSDS-PAGE) was performed using a 10% polyacrylamide gelSambrook and Russell, 2001). For Western blotting, equal volumes

dom mutagenesis. Primers for error-prone PCR, as well as the associated restriction

of culture supernatant were concentrated with trichloroacetic acid(10%, v/v), separated by SDS-PAGE and transferred onto nitrocellu-lose membrane (Pierce). Immunoblotting was performed using theS-Tag Western blot kit (Novagen) according to the manufacturer’sinstructions.

3. Results

3.1. Secretion of CGTase by the hemolysin transport system

A truncated CGTase gene lacking its N-terminal signal sequencewas cloned in frame upstream of the C-terminal 61 amino acidof HlyA (HlyAs61)-HlyB-HlyD operon (Fig. 1). Cultivation of therespective recombinant E. coli cells onto LB-starch (LBS) agar plateindicated that CGTase secretion was dependent on the hemolysinABC transporter (HlyB-HlyD) (Fig. 2).

3.2. Isolation of Hly mutants with improved secretion of CGTase

Two pairs of primers were used to generate random pointmutations in the hly genes: BamF–EcoR and EcoF–SacR (Fig. 1).Approximately 5000 transformants were generated from each ofthe primer pairs. Mutant candidates that formed large halo zones onLBS plates were selected and the level of secreted CGTase-HlyAs61was determined by a quantitative CGTase assay. Five variants (070-11B, 079-08F, 083-10E, 093-05A, and 101-12C) that showed thehighest secretion levels were obtained with the BamF–EcoR primerpair (Figs. 1 and 3). On the other hand, no significant variant wasobtained from the primer pair EcoF–SacR.

Fig. 2. Secretion of enzymatic active CGTase-HlyAs61 by the hemolysin transportsystem in E. coli. E. coli cells were incubated on LBS plate at 30 ◦C for 24 h; 1, E.coli (pET29(a) [negative control]); 2, E. coli (pCGT-As [negative control]); 3, E. coli(pCGT-AsBD).

Page 4

456 K.O. Low et al. / Journal of Biotechnology 150 (2010) 453–459

Table 2Mutation of selected Hly variants.

Round Variant Amino acid and nucleotide substitutions a

Error-prone PCR (mutant library) 070-11B F175L(T547G), (c1866t)079-08F V162A(T485C)083-10E (t12c), D37G(A110G), (t483c), V273A(T818C), I355L(A1063T)093-05A (t3007c)b, (t483c), (t639a)101-12C (t483c), V621A(T1862C)

Site-directed mutagenesis L448F L448F(C1342T)CA (t3007c),b (t483c), (t639a), V621A(T1862C)FA (t3007c),b (t483c), (t639a), V162A(T485C)AB (t3007c),b (t483c), (t639a), F175L(T547G)FC V162A(T485C), (t483c), V621A(T1862C)FAB V162A(T485C), F175L(T547G), (c1866t)FCB V162A(T485C), (t483c), V621A(T1862C), F175L(T547G), (c1866t)

V16

n by

(1sabt(rvimafi(c

3

a

Ffscc19

FCAB

a Substitutions of nucleotide are shown in parentheses. Silent mutations are showb Mutation refers to HlyA.

217%) compared to that of the wild type, while variants 101-2C, 093-05A, 083-10E and 070-11B exhibited an increased level ofecretion of 130, 56, 41 and 35%, respectively. As shown in Fig. 4B, andditional protein band was also detected below the major proteinands (CGTase-HlyAs61). This is most likely corresponding to pro-eolytic degradation product which is also observed in other studiesSugamata and Shiba, 2005; Caspers et al., 2010). DNA sequencingevealed the position of the mutation in each variant (Table 2). Eachariant contained an average of 2.6 base changes, which resultedn one to three amino acid substitutions and up to three silent

utations. Two of the most improved secretion variants, 079-08Fnd 101-12C, had a valine residue substituted with alanine at therst transmembrane segment (V162A) and at the ATPase domainV621A) of HlyB, respectively, suggesting a structurally importanthange that could enhance protein secretion.

.4. Silent mutant 093-05A

Variant 093-05A contained three silent mutations but no aminocid substitutions and to our surprise, it conferred 56% increase

ig. 3. Starch hydrolysis activity of transformants harbouring the CGTase-Hly-usion gene (pCGT-AsBD). Semi-quantitative estimation of the extracellularecretion of CGTase was performed by comparing the sizes of clear halos. E. coliells were incubated on LBS plates at 30 ◦C for 24 h; 1, E. coli (pCGT-As [negativeontrol]); 2, E. coli (pCGT-AsBD [wild-type]); 3, E. coli (p070-11B); 4, E. coli (p083-0E); 5, E. coli (p093-05A); 6, E. coli (p101-12C)i; 7, E. coli (p079-08F); 8, E. coli (pFA);, E. coli (pFCB); 10, E. coli (pFAB); 11, E. coli (pFCAB).

2A(T485C), (t483c), V621A(T1862C), (t3007c),b (t639a), F175L(T547G), (c1866t)

lowercase letters. All mutations refer to HlyB, except for b.

in secretion levels compared to the wild-type. Two of the silentmutations were located at TMD-HlyB, while another was locatedat HlyAs61 (Table 2). To the best of our knowledge, this is the firstreport of silent mutation(s) in HlyAs and/or HlyB that could result inan enhanced-secretion phenotype, which encouraged us to furthercharacterise the silent mutations. We introduced the silent muta-tions into three improved-secretion variants (070-11B, 079-08Fand 101-12C), which resulted in variants AB, FA and CA, respec-tively (Table 2). Expression studies revealed that the addition ofsilent mutations into 070-11B (i.e., variant AB) further improvedits CGTase-HlyAs61 secretion level (Fig. 5). However, the silentmutations resulted in a substantial decrease in secretion level for079-08F and 101-12C. Despite this result, the mechanism by whichsilent mutations affect protein secretion is unknown.

3.5. Effectiveness of mutation V621A and L448F on CGTasesecretion

Mutation L448F in HlyB was reported to result in a mutant thathad 27-fold-higher levels of subtilisin E secretion activity than thewild-type did at 23 ◦C (Sugamata and Shiba, 2005). To evaluate the

Fig. 4. Comparison of the amount of CGTase-HlyAs61 secreted by wild-type andmutants. (A) Relative amount of CGTase-HlyAs61 secreted by wild-type and mutantsin the culture supernatant. CGTase activity was measured 16 h after induction withIPTG (40 �M) at 30 ◦C. The enzyme activity was expressed as percentage relative towild-type. Each bar represents the mean ± standard error (n = 3). NC, negative con-trol; WT, wild-type. (B) Immunoblot analysis of the CGTase-HlyAs61 in the culturesupernatant.

Page 5

K.O. Low et al. / Journal of Biotechnology 150 (2010) 453–459 457

Fig. 5. Effect of silent mutations on CGTase-HlyAs61 secretion for improved-secretion variants. E. coli BL21(DE3) strains carrying p070-11B (070-11B), p079-08F(079-08F), p101-12C (101-12C), and the respective addition of silent mutations(sww

edsamaacittwHVbs

Fec1alab

Fig. 7. Comparison of the amount of CGTase-HlyAs61 secreted by 079-08F-basedvariants. (A) Relative amount of CGTase-HlyAs61 secreted by single and multiple pos-

093-05A) pAB (AB), pFA (FA) and pCA (CA) were cultured, and protein expres-ion was induced by IPTG (40 �M) at 30 ◦C for 16 h. The level of secreted CGTaseas measured using the CGTase assay. Secretion activity was calculated relative toild-type activity (100%). Each bar represents the mean ± standard error (n = 3).

ffectiveness of this mutation on CGTase translocation, we intro-uced mutation L448F into our construct (Fig. 1) and compared itsecretion with mutant 101-12C (mutation V621A), both of whichre located at HlyB-NBD. Taking into account the possible experi-ental differences between Sugamata and Shiba’s study (Sugamata

nd Shiba, 2005) and our study, constructs with longer (218 aminocid) and shorter (61 amino acid) C-terminal signal sequences werereated (see Section 2.2). Surprisingly, mutation L448F resultedn significantly lower secretion of CGTase-HlyAs, independent ofhe length of the C-terminal signal sequence and post-inductionemperature, compared to wild-type (Fig. 6). Only mutant L448F,ith a shorter signal sequence, showed wild-type levels of CGTase-

lyAs61 secretion when expressed at 23 ◦C. In contrast, mutation621A outperformed mutation L448F, in all parameters tested,y a factor of 19–84%. In addition, the shorter C-terminal signalequence was found to be superior to the longer signal sequence

ig. 6. Secretion activity of CGTase-HlyAs using C-terminal signal sequence of differ-nt lengths and different protein expression temperature. E. coli BL21(DE3) strainsarrying pCGT-AsBD (WT), pCGT-AsBD218 (WT218), p101-12C (101-12C), p101-2C218 (101-12C218), pL448F (L448F), and pL448F218 (L448F218) were cultured,nd protein expression was induced by IPTG for 16 h (30 ◦C) and 24 h (23 ◦C). Theevel of secreted CGTase-HlyAs was measured using the CGTase assay. Secretionctivity was calculated relative to wild-type activity (100%). WT, wild-type. Eachar represents the mean ± standard error (n = 3).

itive mutations variants in the culture supernatant. CGTase activity was measured16 h after induction with IPTG (40 �M) at 30 ◦C. The enzyme activity is expressedas percentage relative to wild-type. Each bar represents the mean ± standard error(n = 3). (B) Immunoblot analysis of CGTase-HlyAs61 in the culture supernatant.

for all constructs at all tested temperatures based on extracellularCGTase enzyme activity.

3.6. Combination of mutations for enhanced secretion activity

The combination of positive mutations can often furtherimproves the desired characteristics of a protein (Goh et al., 2009;Hikaru et al., 2009). Therefore, four variants were created by com-bining mutations V162A, V621A, F175L and silent mutations indifferent orders. All four variants showed further improvements inCGTase-HlyAs61 secretion levels, ranging from 6 to 27% improve-ment, compared to the mutant 079-08F (mutation V162A) (Fig. 7).

3.7. Secretory expression of other protein

By using the 079-08F variant, we tested the secretory expressionof another target protein. Cutinase from G. cingulata was chosen ascandidate for secretory expression by the mutated Hly system. Asshown in Fig. 8, secretory expression of cutinase was not detectedin wild-type HlyB but was clearly observed in variant 079-08F after4 h of induction. This indicates that the HlyB-V162A mutation (vari-ant 079-08F) increases the protein translocation efficiency of thetransport machinery. The results suggest that variant 079-08F hasthe potential to efficiently secrete not only CGTase but also otherproteins, as demonstrated on cutinase.

4. Discussion

The aim of this study was to construct a hemolysin-mediatedheterologous protein secretion system in E. coli and to furtherimprove the ABC transporter system for the enhanced secretoryproduction of target protein. We employed an alkalophilic CGTase,a commercially valuable enzyme from locally isolated Bacillus sp.G1 (Ong et al., 2008), as the target protein. The N-terminal truncated

CGTase fused upstream to the C-terminal 61 amino acid of HlyA wassecreted into the culture medium. For the system to be competi-tive for industrial applications, we performed directed evolutionon Hly components via epPCR. The constructed mutant library wasscreened on LBS plates at 30 ◦C, from which about 1.5%, or 150,

Page 6

458 K.O. Low et al. / Journal of Biotech

Fig. 8. Secretion of cutinase by wild-type and mutant HlyB genes. Secretion of cut-HlyAs was evaluated by using E. coli cultures carrying pCUT-AsBD (wild type [WT])otc

csl3

mtrEcTsg0wd

tasdLaibAotTdmiamab1vctp

rP

ment of protein secretion. The ability to secrete target proteins in

61

r pCUT-07908F (079-08F) or pCUT-As (negative control [N]). Culture media wereaken at 4, 8 and 12 h after IPTG induction (40 �M), as indicated, at 30 ◦C. Negativeontrol cells were cultured for 12 h after induction.

lones were subjected to a quantitative CGTase assay. Using thistrategy, five mutant clones were isolated with improved secretionevels ranging from 35 to 217% better than that of the wild-type at0 ◦C.

DNA sequencing analysis revealed that most of the positiveutations were distributed across the TMD-HlyB. It is well known

hat the TMD of the ABC transporter is responsible for substrateecognition and specificity. A similar observation was reported byom et al. (2005) in which amino acid changes in the TMD of TliDaused an increase in TliA lipase secretion. They reasoned that theliD mutants had more efficient interaction with the C-terminalignal sequence of TliA and thus facilitated the secretion of tar-et proteins. Similarly, our mutants (i.e. mutant 070-11B, 093-05A,79-08F and 083-10E) might also confer more efficient interactionsith HlyAs, and the results point to the importance of the TMDuring the translocation of target proteins.

In addition, we isolated 093-05A, a silent mutant that containshree silent mutations and no amino acid substitution but displaysn improved-secretion phenotype. Silent mutations in the hlyAequence had been shown to result in HlyA hypersecretion by intro-ucing rare codons that slow down the translation rate (Lee andee, 2005) and resulted in reduction of intracellular HlyA proteinggregates (inclusion bodies) (Gupta and Lee, 2008). These stud-es underline the importance of improving HlyA-secretion yieldy slowing the translation rate using synonymous codon changes.nother theory is that the rate of translation in a specific regionf a gene may affect the folding, specific activity and function ofhe encoded polypeptide (Komar et al., 1999; Sarfaty et al., 2006).o examine the validity of this hypothesis in our system, we intro-uced the silent mutations (093-05A) into three single, positiveutation variants. We showed that silent mutations did indeed

nfluence the secretion of target proteins despite the fact that nomino acid substitution occurred. Interestingly, the effect of silentutations was not the same in all of the positive mutants. Secretory

ctivity was improved in mutant F175L (i.e., 070-11B against AB)ut not in mutant V162A (i.e., 079-08F against FA) or V621A (i.e.,01-12C against CA). Despite these results, we cannot rule out thealidity of our hypothesis; further investigation is needed to elu-idate the exact mechanism. It would be very exciting to elucidatehe role of silent mutation(s) and codon preference in membrane

rotein biogenesis and protein transport.

101-12C contains a single amino acid substitution in the HlyBegion (V621A) near the C-loop and at the N-terminal of thero-loop. HlyB-V621 is highly conserved throughout the repeat-in-

nology 150 (2010) 453–459

toxin (RTX) family of ABC transporters (Pasteurella haemolytica LktBand Proteus vulgaris HlyB); it is also highly conserved among otherNBDs such as the Termococcus litoralis maltose importer (MalK), thePyrococcus furiosus DNA repair enzyme (Rad50) and the Mus muscu-lus bile salt export pump (ABCBB). However, the hypothetical ABCtransporter ATP-binding protein MJ0796 (from Methanococcus jan-naschii), HisP (from Salmonella typhimurium) and CysA, the putativeABC transporter (from Alicyclobacillus acidocaldarius), all have Alaat position 161, 169 and 161, respectively; this is also the case formutant 101-12C. Several studies have pointed out the importanceof the Pro loop region in the overall translocation cycle, particu-larly in its interaction with the TMD (Petronilli and Ames, 1991;Blight et al., 1994b; Liu et al., 1999; Zaitseva et al., 2005). Althoughthe mechanism by which V621A contributes to improved secre-tion activity is unclear, the mutation may facilitate the interactionbetween the NBD and the TMD or between the catalytic domainand the helical domain. The secretion level of mutant 101-12C wasfound to be higher than that of the wild-type at 23 ◦C and 30 ◦C,which indicates that the mutation has the ability to improve thesecretion activity of the transport machinery, albeit at a lower ratein lower temperature.

Mutant HlyB-L448F is shown to be able to increase the secre-tion of subtilisin E, human c-Myc and PTEN, suggesting the effectof the mutation is universal and not protein specific (Sugamata andShiba, 2005). However, secretion of CGTase-HlyAs by mutant HlyB-L448F resulted in contradicting outcome. There are three mainexperimental differences between studies by Sugamata and Shiba(2005) and ours: the reporter molecule (subtilisin E vs CGTase),the length of C-terminal signal sequence (218 amino acids vs61 amino acids) and the post-induction temperature (23 ◦C vs30 ◦C). The latter two variables were shown not contributing tothe observed phenotype of HlyB-L448F on CGTase-HlyAs secre-tion. Therefore, we envisaged that the choice of reporter molecule,at least for CGTase-HlyAs, influences the secretion efficiency ofmutant HlyB-L448F. Indeed, ABC transporter mutants that conferincreased secretion of TliA lipase, secrete wild-type level of PrtAprotease (Eom et al., 2005). Besides, Sugamata and Shiba (2005)showed that HlyB-L448F secretes a lower level of PTEN than subtil-isin E, also suggesting that the secretion efficiency may be proteindependent.

Multiple mutations showed a step-wise increase in CGTase-HlyAs61 secretion level. Interestingly, combination of closelylocated (based on primary sequence) beneficial mutations [i.e.,mutant FCB and FAB (see Table 2)] resulted in greater incrementof target protein secretion level than distantly located beneficialmutations (i.e., mutant FC). Our observation suggested that thebeneficial effect of the closely located mutations may be similarto each other and thus could be ‘amplified’ to a greater extent. Itshould be noted that since the ABC transporter, particularly theTMD, is responsible for substrate molecule recognition (Zhang et al.,1993; Thanabulu et al., 1998), we believe that our HlyB mutants (i.e.,070-11B, 079-08F, FCB, FAB and FCAB) conferred an increased inter-action with the substrate molecule. To clarify the contribution of themutation(s), further study is necessary to examine the structure-function relationships among the single mutant, the combinationmutants and the mutants substituted with all other amino acid ateach position.

In this study, we produced mutants with improved secretionlevels by performing directed evolution of the E. coli hemolysintransport system using activity screening with the starch plateassay. We found novel sites/regions important for the enhance-

large quantities extracellularly should provide enormous potentialfor industrial and clinical applications. We are currently investigat-ing the effectiveness of our mutants in the secretion of a numberof heterologous proteins. Elucidating the mechanism by which the

Page 7

iotechn

ma

A

BMN

R

A

A

B

B

B

B

B

C

C

C

C

E

F

F

G

G

H

H

K.O. Low et al. / Journal of B

utations alter the level of protein secretion will require furthernalysis.

cknowledgement

This project was supported by the Genomics and Moleculariology Initiatives Programme of the Malaysia Genome Institute,inistry of Science, Technology and Innovation Malaysia (Projecto. 07-05-MGI-GMB011).

eferences

bu Bakar, F.D., Cooper, D.M., Zamrod, Z., Mahadi, N.M., Sullivan, P.A., 2001. Cloningand characterisation of the cutinase-encoding gene and cDNA from the fun-gal phytopathogen Glomerella cingulata. Asia Pac. J. Mol. Biol. Biotechnol. 9 (2),119–130.

ndreas, U., Susanne, N., Jaeger, K.-E., 1997. A rapid and efficient method for site-directed mutagenesis using one-step overlap extension PCR. Nucleic Acids Res.25 (11), 2227–2228.

alakrishnan, L., Hughes, C., Koronakis, V., 2001. Substrate-triggered recruitment ofthe TolC channel-tunnel during type I export of hemolysin by Escherichia coli. J.Mol. Biol. 313, 501–510.

aneyx, F., 1999. Recombinant protein expression in Escherichia coli. Curr. Opin.Biotechnol. 10, 411–421.

enabdelhak, H., Kiontke, S., Horn, C., Ernst, R., Blight, M.A., Holland, I.B., Schmitt,L., 2003. A specific interaction between the NBD of the ABC-transport HlyB anda C-terminal fragment of its transport substrate haemolysin A. J. Mol. Biol. 327,1169–1179.

light, M.A., Pimenta, A.L., Lazzaroni, J.-C., Dando, C., Kotelevets, L., Seror,S.J., Holland, I.B., 1994a. Identification and preliminary characterization oftemperature-sensitive mutations affecting HlyB, the translocator required forthe secretion of haemolysin (HlyA) from Escherichia coli. Mol. Gen. Genet. 245,431–440.

light, M.A., Pimenta, A.L., Lazzaroni, J.-C., Dando, C., Kotelevets, L., Séror,S.J., Holland, I.B., 1994b. Identification and preliminary characterization oftemperature-sensitive mutations affecting HlyB, the translocator required forthe secretion of haemolysin (HlyA) from Escherichia coli. Mol. Gen. Genet. 245,431–440.

aspers, M., Brockmeier, U., Degering, C., Eggert, T., Freudl, R., 2010. Improvementof Sec-dependent secretion of a heterologous model protein in Bacillus subtilisby saturation mutagenesis of the N-domain of the AmyE signal peptide. Appl.Microbiol. Biotechnol. 86 (6), 1877–1885.

hervaux, C., Holland, I.B., 1996. Random and directed mutagenesis to elucidate thefunctional importance of helix II and F-989 in the C-terminal secretion signal ofEscherichia coli hemolysin. J. Bacteriol. 178, 1232–1236.

hoi, J.H., Lee, S.Y., 2004. Secretory and extracellular production of recombinantprotein using Escherichia coli. Appl. Microbiol. Biotechnol. 64, 625–635.

olosimo, A., Xu, Z., Novelli, G., Dallapiccola, B., Gruenert, D.C., 1999. Simple ver-sion of “megaprimer” PCR for site-directed mutagenesis. Biotechniques 26, 870–873.

om, G.T., Song, J.K., Ahn, J.H., Seo, Y.S., Rhee, J.S., 2005. Enhancement of the efficiencyof secretion of heterologous lipase in Escherichia coli by directed evolution of theABC transport system. Appl. Environ. Microbiol. 71, 3468–3474.

ernández, L.A., Sola, I., Enjuanes, L., de Lorenzo, V., 2000. Specific secretion of activesingle-chain Fv antibodies into the supernatants of Escherichia coli cultures byuse of the hemolysin system. Appl. Microbiol. Biotechnol. 66, 5024–5029.

raile, S., Munoz, A., de Lorenzo, V., Fernández, L.A., 2004. Secretion of proteins withdimerization capacity by the haemolysin type I transport system of Escherichiacoli. Mol. Microbiol. 53, 1109–1121.

oh, K.M., Mahadi, N.M., Hassan, O., Raja Abdul Rahman, R.N.Z., Md Illias, R., 2009.A predominant �-CGTase G1 engineered to elucidate the relationship betweenprotein structure and product specificity. J. Mol. Catal. B: Enzym. 57, 270–277.

upta, P., Lee, K.H., 2008. Silent mutations result in HlyA hypersecretion by reducingintracellular HlyA protein aggregates. Biotechnol. Bioeng..

ikaru, N., Okada, K., Onodera, T., Ogasawara, W., Okada, H., Morikawa, Y., 2009.

Directed evolution of endoglucanase III (Cel12A) from Trichoderma reesei. Appl.Microbiol. Biotechnol. 83, 649–657.

o, K.S., Said, M., Hassan, O., Kamaruddin, K., Ismail, A.F., Rahman, R.A., NikMahmood, N.A., Illias, R.Md., 2005. Purification and characterization of cyclodex-trin glucanotransferase from alkalophilic Bacillus sp. G1. Process Biochem. 40,1101–1111.

ology 150 (2010) 453–459 459

Holland, I.B., Schmitt, L., Young, J., 2005. Type I protein secretion in bacteria, the ABC-transporter dependent pathway. Mol. Membr. Biol. 22 (1–2), 29–39 (Review).

Hui, D., Ling, V., 2002. A combinatorial approach toward analyzing functionalelements of the Escherichia coli hemolysin signal sequence. Biochemistry 41,5333–5339.

Hui, D., Morden, C., Zhang, F., Ling, V., 2000. Combinatorial analysis of the structuralrequirements of the Escherichia coli hemolysin signal sequence. J. Biol. Chem.275, 2713–2720.

Komar, A.A., Lesnik, T., Reiss, C., 1999. Synonymous codon substitutions affect ribo-some traffic and protein folding during in vitro translation. FEBS J. 462, 387–391.

Koronakis, V., Koronakis, E., Hughes, C., 1989. Isolation and analysis of the C-terminalsignal directing export of Escherichia coli hemolysin protein across both bacterialmembranes. EMBO J. 8, 595–605.

Kostakioti, M., Newman, C.L., Thanassi, D.G., Stathopoulus, C., 2005. Mechanismsof protein export across the bacterial outer membrane. J. Bacteriol. 187,4306–4314.

Lee, P.S., Lee, K.H., 2005. Engineering HlyA hypersecretion in Escherichia coli basedon proteomic and microarray analyses. Biotechnol. Bioeng. 89, 195–205.

Liu, P.-Q., Liu, C.E., Ames, G.F.-L., 1999. Modulation of ATPase activity by physicaldisengagement of the ATP-binding domains of an ABC transporter, the histidinepermease. J. Biol. Chem. 274, 18310–18318.

Mackman, N., Baker, K., Gary, L., Haigh, R., Nicaud, J.-M., Holland, I.B., 1987. Releaseof a chimeric protein into the medium from Escherichia coli using the C-terminalsecretion signal of haemolysin. EMBO J. 6, 2835–2841.

Ong, R.M., Goh, K.M., Mahadi, N.M., Hassan, O., Raja Abdul Rahman, R.N.Z., Illias,R.Md., 2008. Cloning, extracellular expression and characterization of a pre-dominant �-CGTase from Bacillus sp. G1 in E. coli. J. Ind. Microbiol. Biotechnol.35, 1705–1714.

Petronilli, V., Ames, G.F.-L., 1991. Binding protein-independent histidine permeasemutants. J. Biol. Chem. 266, 16293–16296.

Piementa, A.L., Racher, K., Jamieson, L., Blight, M.A., Holland, I.B., 2005. Mutations inHlyD, part of the type I translocator for hemolysin secretion, affect the foldingof the secreted toxin. J. Bacteriol. 187, 7471–7480.

Sambrook, J., Russell, D.W., 2001. Molecular Cloning: A Laboratory Manual. ColdSpring Harbor, New York.

Sarfaty, C.K., Oh, J.M., Kim, I.-W., Sauna, Z.E., Calcagno, A.M., Ambudkar, S.V., Gottes-man, M.M., 2006. A “silent” polymorphism in the MDR1 gene changes substratespecificity. Sciencexpress, 1–4.

Schulein, R., Gentschev, I., Schlor, S., Gross, R., Goebel, W., 1994. Identification andcharacterization of two functional domains of the hemolysin translocator pro-tein HlyD. Mol. Gen. Genet. 245, 203–211.

Sheps, J.A., Cheung, I., Ling, V., 1995. Hemolysin transport in Escherichia coli. Pointmutants in HlyB compensate for a deletion in the predicted amphiphilic helixregion of the HlyA signal. J. Biol. Chem. 270 (24), 14829–14834.

Sommer, B., Friehs, K., Flaschel, E., 2010. Efficient production of extracellular proteinswith Escherichia coli by means of optimized coexpression of bacteriocin releaseproteins. J. Biotechnol. 145 (4), 350–358.

Sommer, B., Friehs, K., Flaschel, E., Reck, M., Stahl, F., Scheper, T., 2009. Extracellularproduction and affinity purification of recombinant proteins with Escherichiacoli using the versatility of the maltose binding protein. J. Biotechnol. 140 (3–4),194–202.

Sriubolmas, N., Panbangred, W., Sriurairatana, S., Meevootisom, V., 1997. Localiza-tion and characterization of inclusion bodies in recombinant Escherichia coli cellsoverproducing penicillin G acylase. Appl. Microbiol. Biotechnol. 47, 373–378.

Sugamata, T., Shiba, T., 2005. Improved secretory production of recombinant pro-teins by random mutagenesis of hlyB, an alpha hemolysin transporter fromEscherichia coli. Appl. Environ. Microbiol. 71, 656–662.

Sung, H.Y., Seong, K.K., Kim, J.F., 2010. Secretory production of recombinant proteinsin Escherichia coli. Recent Pat. Biotechnol. 4, 23–29.

Thanabulu, T., Koronakis, E., Hughes, C., Koronakis, V., 1998. Substrate-inducedassembly of a contiguous channel for protein export from E. coli: reversiblebridging of an inner-membrane translocase to an outer membrane exit pore.EMBO J. 17, 6487–6496.

Urban, A., Neukirchen, S., Jaeger, K.E., Journals, O., 1997. A rapid and efficient methodfor site-directed mutagenesis using one-step overlap extension PCR. NucleicAcids Res. 25, 2227–2228.

Wandersman, C., Delepelaire, P., 1990. TolC, an Escherichia coli outer membrane pro-tein required for hemolysin secretion. Proc. Natl. Acad. Sci. U.S.A. 87, 4776–4780.

Zaitseva, J., Jenewein, S., Oswald, C., Jumpertz, T., Holland, I.B., Schmitt, L., 2005. Amolecular understanding of the catalytic cycle of the nucleotide-binding domainof the ABC transporter HlyB. Biochem. Soc. Trans. 33, 990–995.

Zhang, F., Sheps, J.A., Ling, V., 1993. Complementation of transport-deficient mutantsof Escherichia coli alpha-hemolysin by second-site mutations in the transporterhemolysin B. J. Biol. Chem. 268 (26), 19889–19895.