Functional analysis of the C-terminal domain of the WbaP protein ...

Glycobiology vol. 20 no. 11 pp. 1389–1401, 2010doi:10.1093/glycob/cwq104Advance Access publication on June 29, 2010

Functional analysis of the C-terminal domain of the WbaPprotein that mediates initiation of O antigen synthesis inSalmonella enterica†

Kinnari B Patel2, Sarah E Furlong2,and Miguel A Valvano1,2,3

2Department of Microbiology and Immunology, Infectious Diseases ResearchGroup, Siebens-Drake Research Institute, and 3Department of Medicine,University of Western Ontario, London, Ontario, N6A 5C1, Canada

Received on May 21, 2010; revised on June 26, 2010; accepted onJune 27, 2010

WbaP catalyzes the transfer of galactose-1-phosphate ontoundecaprenyl phosphate (Und-P). The enzyme belongs to alarge family of bacterial membrane proteins required forinitiation of the synthesis of O antigen lipopolysaccharideand polysaccharide capsules. Previous work in our labora-tory demonstrated that the last transmembrane helix andC-terminal tail region of WbaP (WbaPCT) are sufficient forenzymatic activity. Here, we demonstrate the cytoplasmiclocation of theWbaPC-terminal tail and show thatWbaPCT

domain N-terminally fused to thioredoxin (TrxA–WbaPCT)exhibits improved protein folding and enhanced transferaseactivity. Alanine replacement of highly conserved chargedor polar amino acids identified seven critical residues for en-zyme activity in vivo and in vitro. Four of these residues arelocated in regions predicted to be α-helical. These regionsand their secondary structure predictions are conserved indistinct WbaP family members, suggesting they may con-tribute to form a conserved catalytic center.

Keywords: lipopolysaccharide/membrane protein /O antigen/sugar transferase

Introduction

The outer membrane of Gram-negative bacteria is an asymmet-ric lipid bilayer with an outer leaflet consisting predominantlyof lipopolysaccharide (LPS) molecules. LPS has a lipid A por-tion embedded in the outer membrane, a core oligosaccharideand an O-specific polysaccharide (O antigen) that is exposedto the bacterial surface. O antigen biogenesis is a multistep pro-cess commencing at the cytoplasmic face of the plasmamembrane (Raetz and Whitfield 2002; Valvano 2003). The ini-tiation reaction leads to the formation of a phosphoanhydridebond between a sugar-1-phosphate (commonly delivered as a

uridine diphosphate [UDP] sugar) and undecaprenyl phosphate(Und-P) with the concomitant release of uridine monopho-sphate. Subsequent steps in O antigen synthesis are completedvia one of two major pathways designated ABC transporter- andWzy-dependent pathways (Raetz and Whitfield 2002; Valvano2003). In the Wzy-dependent pathway, glycosyltransferases se-quentially add sugars to form a complete Und-P-P-linked Oantigen subunit facing the cytosol, which is translocated tothe periplasm by a process mediated by the Wzx protein. O an-tigen subunits are subsequently polymerized by the Wzypolymerase, and they acquire a particular length distributionby an unknown mechanism that requires Wzz, also called chainlength determinant or O antigen copolymerase (Morona et al.2000). The polymer is finally ligated to independently synthe-size lipid A-core oligosaccharide and the LPS molecule istransported to the outer membrane (Raetz and Whitfield2002; Valvano 2003) by the recently discovered Lpt pathway(Ruiz et al. 2009; Sperandeo et al. 2009).Enzymes catalyzing the initiation reaction can be grouped

into two distinct families: polyisoprenyl-phosphate N-acetyl-hexosamine-1-phosphate (PNPT) and polyisoprenyl-phosphatehexose-1-phosphate (PHPT) transferases (Valvano 2003).PNPTs include prokaryotic and eukaryotic members whilePHPTs are restricted to bacteria. WecA, a UDP-GlcNAc:Und-P GlcNAc-1-P transferase commonly found in the Enterobac-teriaceae, is a prototypic member of the PNPT family. TheSalmonella enterica WbaP has UDP-galactose (Gal):Und-PGal-1-P transferase activity and is a prototypic member of thePHPT family (Reeves 1993; Wang et al. 1996; Valvano 2003;Saldías et al. 2008). Other characterized PHPT members initiatethe synthesis of glycans for S-layer protein glycosylation, suchas Geobacillus stearothermophilus WsaP (Steiner et al. 2007),and exopolysaccharides, such as Xanthomonas campestrisGumD (Katzen et al. 1998), Streptococcus pneumonia Cps2E(Cartee et al. 2005; Xayarath and Yother 2007),Klebsiella pneu-moniae ORF14 (Arakawa et al. 1995; Drummelsmith andWhitfield 1999), and Escherichia coli K-12 WcaJ (Stevensonet al. 1996).WbaP is a basic membrane protein with five predicted trans-

membrane (TM) helices (Saldías et al. 2008). We havepreviously demonstrated three domains in WbaP: a hydropho-bic, membrane embedded N-terminal region spanning four TMhelices, a central periplasmic loop, and a large C-terminal do-main comprising the fifth TM helix with the remainder of theprotein predicted to face the cytosol (Saldías et al. 2008). Thelast 219 residues encompassing the tail region are sufficient forgalactosyl-phosphate transferase activity in vivo (Wang et al.

© The Author 2010. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected] 1389

1To whom correspondence should be addressed: Tel: +1-519-661-3427; Fax:+1-519-661-3499; e-mail: [email protected]†Dedicated to the memory of Professor Thammaiah Viswanatha, friend andmentor, deceased on November 23, 2008.

Downloaded from https://academic.oup.com/glycob/article-abstract/20/11/1389/1987815by gueston 08 April 2018

1996; Saldías et al. 2008). In this study, we characterize inmore detail the C-terminal domain of WbaP. An N-terminalthioredoxin (TrxA) fusion to this domain led to enhanced pro-tein folding and membrane expression. Together, trypsinaccessibility and green fluorescent protein (GFP) localizationexperiments demonstrated that the C-terminus is cytoplasmic,and analysis of GFP-fused proteins demonstrated that WbaPlocalizes to the poles of the cells. A targeted amino acid re-placement strategy directed at highly conserved charged andpolar residues identified three arginines (R319, R377, R401),one lysine (K331), one glutamic acid (E383), and two asparticacids (D382 and D458) that are critical for in vivo and in vitroWbaP activity. We also show that residues R377, D331, E383,and D458 are located in three putative α-helical segments,which are highly conserved in other members of the WbaPfamily, suggesting they may form part of a catalytic center.

ResultsThe C-terminus of WbaP resides in the cytoplasm

WbaP consists of five predicted TM helices, a large periplas-mic loop between TM-IV and TM-V (PL2), and a largecytoplasmic tail region (Figure 1) (Saldías et al. 2008). Previ-ous work has demonstrated that only the WbaP C-terminus(WbaPCT), from phenylalanine-258 (F258) to tyrosine-476(Y476), is required for transferase activity (Wang et al. 1996;Saldías et al. 2008). To experimentally establish which portionof the C-terminus resides in the cytoplasm, we conducted pro-tease accessibility assays on spheroplasts as described inMaterials and methods. We opted for this method as an alter-native approach to study protein topology due to the poormembrane expression of WbaP constructs fused to reporterproteins such as LacZ (β-galactosidase) and PhoA (alkalinephosphatase) (data not shown). The protease accessibility strat-

1390

Fig. 1. Graphic representation of the WbaP parental protein and the various deleted derivatives encoded by plasmids used in this study. All constructs contain eithera FLAG or 6xHis epitope tag for detection on the N- or C-terminus. Plasmid pKP42 encodes a construct that also has an N-terminal TrxA fusion followed by a TEVcleavage site. Predicted transmembrane (TM), periplasmic loops (PL), and cytoplasmic loops (CL) are indicated.

Fig. 2. Cytoplasmic localization of the C-terminus by trypsin accessibility experiments in spheroplasts and with inverted membrane vesicles. Samples were treatedwith trypsin, soybean trypsin inhibitor (STI), and/or Triton X-100 as indicated in the panels. Upon treatment, samples were analyzed by immunoblot withanti-FLAG M2 monoclonal antibodies. (A). Spheroplasts of DH5α cells expressing WbaPM1–R354–FLAG or WbaPFLAG–M1–I144/F258–Y476 (both ~43 kDa).Spheroplasts were incubated for 30 min at room temperature. Lanes: 1 and 5, no trypsin; 2, trypsin and STI; 3 and 6, trypsin alone; 4, trypsin and Triton X-100.(B) IMVs were prepared by lysing cells with the French press and were incubated at room temperature for 30 min. Lanes: 1 and 2, no trypsin; 2 and 4,trypsin alone. Samples were separated by 16% SDS–PAGE and arrows indicate trypsin cleavage products containing a FLAG tag.

KB Patel et al.


egy relies on the principle that after formation of spheroplasts,regions of an integral membrane protein that are exposed to theperiplasm would be susceptible to proteolysis, while regionsfacing the cytoplasm would be protected (Dai et al. 1996).Trypsin cleaves selectively after lysine and arginine residues(Olsen et al. 2004). Due to its basic nature, WbaP containsmultiple trypsin cleavage sites spanning the entire length ofthe protein. Since we were unable to obtain a recombinantplasmid expressing full-length WbaP with C-terminal or N-ter-minal tags (Saldías et al. 2008) (data not shown), we usedtruncated WbaP derivatives containing N-terminally (FLAG–WbaPM1–I144/F258–Y476, encoded by pKP10) and C-terminally(WbaPM1–R354–FLAG, encoded by pKP2) fused FLAG epi-topes (Figure 1), which enabled us to detect trypsin cleavagefragments from both ends of the protein. E. coli DH5α cellsexpressing the respective constructs were converted to sphero-plasts and after incubation with trypsin, soybean trypsininhibitor (STI) was added before lysis of spheroplasts byfreeze–thaw. Samples were separated by sodium dodecylsul-fate polyacrylamide gel electrophoresis (SDS–PAGE) andanalyzed by immunoblot. The samples corresponding to un-treated FLAG-tagged WbaP constructs localize to the innermembrane, as they are detected in both spheroplasts andmembrane vesicles, and each have an apparent mass of approx-imately 43 kDa (Figure 2A, lanes 1 and 5 and Figure 2B, lanes 1and 3). A high molecular mass band of approximately 98 kDawas observed also with both proteins, although this band wasmuch more intense in the case of WbaPM1–R354–FLAG(Figure 2A). We interpreted this band as due to oligomerizationof WbaPM1–R354–FLAG either as an artifact of the mild dena-turing conditions we routinely use to process membraneproteins by SDS–PAGE (Amer and Valvano 2002; Maroldaet al. 2004; Lehrer et al. 2007; Perez et al. 2008) (see Materialsand methods) or alternatively, protein–protein interactionsmediated by the periplasmic loop that is absent in FLAG–WbaPM1–I144/F258–Y476 (Figure 1). Trypsin treatment of spher-oplasts expressing WbaPM1–R354–FLAG resulted in a band ofapproximately 36 kDa (Figure 2A, lane 3), which correspondsto the mass of a product cleaved within the first predictedperiplasmic loop (PL1; Figure 1). Trypsin treatment of sphero-plasts expressing FLAG–WbaPM1–I144/F258–Y476 resulted in aband of approximately 9 kDa, which also corresponded to aproduct cleaved in PL1 (Figure 2A, lane 6). There were no de-tectable bands corresponding to the predicted cytoplasmic tailregion, suggesting that this region was not accessible to proteol-ysis. To confirm that the C-terminus is susceptible to trypsincleavage, we isolated inverted membrane vesicles (IMVs) byFrench press disruption of cells expressing WbaPM1–R354–FLAG and FLAG–WbaPM1–I144/F258–Y476 and treating theIMVs with trypsin prior to SDS–PAGE. Trypsin treatment ofIMVs expressing WbaPM1–R354–FLAG resulted in severalbands in the region of the gel corresponding approximately to7–12 kDa (Figure 2B, lane 2), and treatment of IMVs expressingFLAG–WbaPM1–I144/F258–Y476 resulted in bands of approxi-mately 30 kDa, 29 kDa, and 22 kDa (Figure 2, lane 4). Thesebands correspond to the expected products from cleavage siteslocated within the predicted cytoplasmic tail. Together, these re-sults suggest that amino acids located to at least the C-terminal20 kDa of WbaP are in the cytoplasm, in agreement with thebioinformatic predictions. In our trypsin experiments usingspheroplasts or total membranes vesicles, we observed no pro-

teolytic cleavage products from the large predicted periplasmicloop 2 (PL2). This region is predicted to have complex second-ary structure (Saldías et al. 2008), which could prevent access oftrypsin.

An N-terminal TrxA fusion improves folding of overexpressedWbaPF258–Y476 and enhances activity in vivoOverexpressed FLAG–WbaPF258–Y476, herein referred to asWbaPCT, which contains 24 residues from PL2, TM-V andthe WbaP C-terminal cytosolic tail (Figure 1), can restore Oantigen surface expression in the S. Typhimurium ΔwbaPstrain MSS2 (Figure 3). However, as compared to the parentalstrain, WbaPCT does not mediate production of full-length Oantigen since only short O antigen bands can be visible inthe gel. One possible explanation for this result was proteinmisfolding and possible formation of inclusion bodies uponoverexpression of WbaPCT. TrxA is a 12-kDa protein that sta-bilizes membrane proteins when fused to the N-terminus of atarget protein and also prevents the formation of inclusion bod-ies (LaVallie et al. 1993). Therefore, we constructed a WbaPCTderivative with TrxA fused to the N-terminus (TrxA–WbaPCT)under the expression of the arabinose-inducible pBAD promot-er. To facilitate further purification, this construct also has a6xHis N-terminal tag. We investigated the functionality ofTrxA–WbaPCT in the strain MSS2 and found that the additionof the N-terminal 6xHis–TrxA fusion partner resulted in a pro-tein with the ability to restore O antigen production to similarlevels as in the parental strain LT2 (Figure 3). We also usedGFP fusions to determine whether improved activity of over-expressed TrxA–WbaPCT was associated with proper proteinfolding. Proper folding of GFP fused to the C-terminus of a

1391

Fig. 3. Complementation of S. Typhimurium strain MSS2 (ΔwbaP) withoverexpressed WbaPCT and TrxA–WbaPCT. LPS samples were prepared andseparated by 12% SDS–PAGE. Gels were stained with silver nitrate for Oantigen detection. “Polymeric OAg” denotes the region of the gel containinglipid A-core OS bands covalently linked to O antigen polysaccharides ofvarying molecular mass. “Core + 1 OAg” indicates the band corresponding tolipid A-core OS with one covalently attached O antigen subunit.

Structure–function of WbaP C-terminal domain


target protein is highly influenced by the correct folding of thefusion partner such that only folded fusion protein will becomefluorescent (Geertsma et al. 2008). Membrane proteins fused toGFP migrate as a doublet on SDS–PAGE with the faster migrat-ing bands representing folded proteins that fluoresce in gelunder ultraviolet light (UV) and the slower migrating bands re-presenting misfolded proteins unable to fluoresce (Geertsmaet al. 2008). Since we established that the C-terminus of WbaPis cytoplasmic and GFP can only fluoresce when properlyfolded in the cytosol (Feilmeier et al. 2000), we constructedGFP fusions to the C-terminus of WbaPCT derivatives withand without the N-terminal TrxA fusion. Total membranes fromMSS2 cells expressing the GFP fusion constructs were separat-ed by SDS–PAGE and detected with anti-GFP antiserum(Figure 4A). In addition to the misfolded and folded forms ofprotein detected (Figure 4A, white and black arrows, respec-tively) there were also lower molecular weight bands whichlikely correspond to degraded GFP. When the SDS–PAGEgel was observed under UV light, a dense band representingproperly folded TrxA–WbaPCT–GFP protein and a very faintband representing the properly folded WbaPCT–GFP proteinwere detected (Figure 4B). These results suggest that the fusionof TrxA to WbaPCT improves protein folding.

Next, we used fluorescence microscopy to observe the ex-pression of WbaPCT–GFP and TrxA–WbaPCT–GFP inMSS2 cells. Although cells expressing the two constructs fluo-resced (Figure 4C and D), dense punctuate expression aroundthe membrane and at the poles were only seen in cells contain-ing TrxA–WbaPCT–GFP (Figure 4D). This pattern oflocalization of WbaPCT–GFP in bacterial cell bodies is similarto that observed in E. coli K-12 for the initiating glycosyltrans-ferase WecA (Lehrer et al. 2007) and further confirms thecytoplasmic localization of the C-terminus of WbaP. Therefore,we concluded that the TrxA fusion improves membrane ex-pression and folding of WbaPCT, resulting also in fullcomplementation of O antigen expression in MSS2.

Topological location of the TrxA fusion partner in themembraneAlthough TrxA is a cytoplasmic protein, based on the topolog-ical model previously described (Saldías et al. 2008), fusion ofTrxA to the N-terminus of WbaPCT should result in localizationof TrxA to the periplasm. However, rapid folding of TrxA aftertranslation prevents it from crossing the plasma membrane(Huber et al. 2005). To analyze how TrxA–WbaPCT–GFP issituated in the inner membrane, we took advantage of a tobac-co etch virus (TEV) protease site engineered at the fusionendpoint between the TrxA tag and WbaPCT. Protease cleavageexperiments were carried out on spheroplasts using a 6xHis-tagged TEV protease with a mass of 25 kDa, which alsoserved as an internal control. In addition to TrxA–WbaPCT,the cytosolic FLAG-tagged transcriptional regulator from Bur-kholderia cenocepacia, BCAL0381-FLAG (Aubert et al. 2008),was also expressed in the cells as a lysis control. Spheroplastswere treated with TEV protease, and the periplasmic fractionwas collected and analyzed by immunoblot. No cleavage wasdetected in the periplasmic fraction (Figure 5A, lane 3), butcleavage was detectable upon treatment of lysed cells withTEV protease, resulting in a 15 kDa-cleavage product (Figure5A, lane 4). BCAL0381-FLAG was detected only upon lysisof cells (Figure 5B, lane 4) and not in the periplasmic fractions(Figure 5B, lanes 2 and 4). This suggests that the TrxA partner inthe fusion protein is cytoplasmic and that the TM helix in thefusion protein TrxA–WbaPCT must be “pinched in” or looselyassociated with the inner membrane.

Functional characterization of TrxA–WbaPCT in vitroWe next investigated whether TrxA influences the galactosyl-phosphate transferase activity of TrxA–WbaPCT in vitro. Totalmembranes prepared from MSS2, providing both enzyme andendogenous Und-P acceptor, were incubated with 14C-labeledradioactive UDP-Gal. The lipid fraction containing the Und-P-P-Gal product was extracted with 1-butanol and the radioac-tive counts measured. The assay was performed using 12 mMMg2+ and 80 μg of total membrane protein from MSS2 bacteriaexpressing WbaPCT (encoded by pKP12), TrxA–WbaPCT

(encoded by pKP42) or carrying the pBAD24 vector control.Total membranes expressing WbaPCT or TrxA–WbaPCT

resulted in the incorporation of 1.98 ± 0.28 and 2.51 ±0.28 pmoles of Gal per mg of total membrane protein into thebutanol extractable fraction, respectively, while membranes pre-

1392

Fig. 4. Expression and electrophoretic mobility of GFP fusion proteins. Totalmembranes of MSS2 (ΔwbaP) cells expressing WbaPCT–GFP and TrxA–WbaP–GFP were analyzed by immunoblot using anti-GFP antibodies (A) andby in-gel GFP fluorescence using UV light that provides excitation at 395 nm(B). Black and white arrows indicate the position of folded (fluorescent) andmisfolded (nonfluorescent) protein, respectively. Samples were separated by14% SDS–PAGE. Fluorescence microscopy of MSS2 (ΔwbaP) cellsexpressing WbaPCT–GFP (C) or TrxA–WbaPCT–GFP (D). The micrographwas taken with a ×100 oil immersion objective, resulting in a magnification of1000-fold.

KB Patel et al.


pared from cells with vector control incorporated 0.12 ±0.01 pmoles of Gal per mg of total membrane protein. Usingthe analogous sugar-phosphate transferase WecA, we have pre-viously shown that the butanol fraction is highly enriched forundecaprenyl-P-P-sugar species (Lehrer et al. 2007). We alsotested the activity of TrxA–WbaPCT under varying concentra-tions of Mg2+ and Mn2+ and found that the fusion protein hada metal cofactor profile similar to the parental full-length proteinas described previously (Osborn et al. 1962) (data not shown).This demonstrates that the TrxA fusion does not disrupt the en-zymatic activity of WbaPCT.

Identification and in vivo functional analysis of conservedamino acids in TrxA–WbaPCT

To identify residues within the C-terminus of WbaP that arerequired for enzyme function, we first used bioinformaticscomparing several hundred WbaP homologs with the Sequenc-ing Alignment andModeling system (SAM-T02) (Karchin et al.2003). This analysis revealed that R319, K331, R333, R377,D382, E383, R401, D446, Y449, and D458 (the numbers denotethe position of these residues in the S. Typhimurium LT2WbaP)were the most conserved charged or polar residues in pro-teins annotated as putative galactosyl-phosphate transferases(Figure 6). These conserved amino acids were individually re-placed by alanine in TrxA–WbaPCT, and the ability of eachmutant protein to complement LPS O antigen surface expres-sion in the ΔwbaP mutant MSS2 was evaluated by silverstaining. The majority of the alanine replacements abolishedO antigen production (Figure 7A), suggesting the mutatedWbaP derivatives loss enzymatic activity (see below). In con-trast, R333A and Y449A resulted in proteins that mediatedpartial complementation, as noticed by the formation of lipidA-core OS with one O antigen oligosaccharide unit, whilethe D446A mutant protein could complement full-length O an-

tigen (Figure 7A). We also mutated two moderately conservedresidues, R368 and Q386. Both mutant proteins mediated theformation of lipid A-core OS with one O antigen unit(Figure 7A), suggesting that WbaP function was reduced butnot completely abolished. In addition to the replacement mu-tants constructed above, an alanine replacement of the highlyconserved glycine-393 (G393) was constructed, which resultedin a protein that mediated a full complementation of O antigensynthesis (Figure 7A). All of the mutant proteins were detectedin total membrane fractions. However, band shifts were seenupon SDS–PAGE analysis in mutants R401A, D446A, andD458A, as compared to relative migration of the parentalTrxA–WbaPCT protein (Figure 7B). This behavior could beattributed to differential binding of SDS by these proteins(Rath et al. 2009), potentially indicating changes in secondarystructure.To validate the mutagenesis strategy, six nonconserved resi-

dues along the C-terminus: isoleucine-352 (I352), valine-372(V372), phenylalanine-388 (F388), glutamic acid-409 (E409),tryptophan-453 (W453), and leucine-462 (L462) were replacedby alanine and the mutant proteins tested for the ability to com-plement O antigen production. Cells expressing these WbaPmutant proteins produced full-length O antigen (Figure 8), sup-porting the notion that nonconserved amino acids do not play arole in WbaP function.To further characterize the alanine replacement mutants with

no O antigen production, additional replacements with residuesconserving charge and/or structure were constructed. Thus, ar-ginine residues were mutated to lysine and glutamine, asparticacid residues were mutated to glutamic acid and asparagine,and glutamic acid residues were mutated to aspartic acid andasparagine. WbaP derivatives containing the conserved replace-ment mutations could not complement O antigen production(data not shown). In addition, we also further investigated

1393

Fig. 5. Cytoplasmic localization of TrxA in TrxA–WbaPCT. MSS2 (pKP42, p0381-FLAG) cells expressing both TrxA–WbaPCT and the FLAG-tagged cytoplasmictranscriptional regulator BCAL0381-FLAG were incubated for 30 min at 37°C (lane 1). Spheroplasts were incubated alone (lane 2), with TEV protease (lane 3), andwith TEV protease after lysis (lane 4). Periplasmic fractions were collected and analyzed by immunoblot. Cleavage of TrxA-WbaPCT was detected in lysedspheroplasts treated with 6xHis-tagged TEV protease resulting in a 15 kDa product. (A) Anti-His IgG2a monoclonal antibody. (B) Anti-FLAG M2 monoclonalantibody.



whether a large aromatic residue is required at the position ofY449, since the Y449A mutant mediated only a partial comple-mentation of O antigen production. The Y449S replacementmutant mediated partial complementation while the Y449F re-placement led to restoration of full-length O antigen, suggestingthat the aromatic group and not the hydroxyl group is requiredfor full enzyme activity (data not shown). Because maintainingthe same net charge with a different amino acid was not suffi-cient to restore activity, we concluded that R319, K331, R377,D382, E383, R401, and D458 are absolutely critical for enzymeactivity in vivo.

Functional characterization of the highly conserved chargedor polar residues in the C-terminal domain of WbaPThe enzymatic activity assay, as described above for the paren-tal TrxA–WbaPCT protein, was performed under excess ofexogenous Und-P substrate to characterize in more detail thefunctional defects of the alanine replacement mutants. All mu-tant proteins were initially expressed in MSS2 (ΔwbaP).

However, background galactosyl-phosphate transferase activityfluctuated in this strain over time and growth conditions makingit difficult to assess the functionality of mutant proteins withlow levels of enzymatic activity. This could not be resolvedby introducing deletions in the two core galactosyltransferasegenes waaI and waaB (Wollin et al. 1983; Kadam et al.1985), as well as in the putative glucosyltransferase genefor colanic acid synthesis wcaJ (White et al. 2003) (data notshown). Therefore, we used E. coli strain MV501 (wecA::Tn10), as this strain cannot produce Und-P-P-N-acetylglucosa-mine (GlcNAc) due to the absence of WecA (Alexander andValvano 1994). The galactosyltransferase WbbD acts on theGlcNAc substrate to catalyze the transfer of Gal (Riley et al.2005). Therefore, in the absence of WecA, Gal cannot be incor-porated to the O antigen. Accordingly, this strain demonstrateda low and reproducible background activity in the lipid-associated fraction.

To quantify the amounts of parental TrxA–WbaPCT andmutant proteins present in membrane fractions we used quan-

1394

Fig. 6. Representation of conserved amino acids in WbaP C-terminus obtained by the SAM-T02 protein structure prediction server. The server, using sequencesfrom putative WbaP homologs, identifies conserved amino acids. The size of the letter representing the amino acid is proportional to the level of conservation of thisresidue on the automatic alignment. The position of residues targeted by mutagenesis is indicated below the letter. Charged and polar residues that are highlyconserved (R319, K331, R333, R377, D382, E383, R401, D446, Y449, and D458) and two moderately conserved residues (R368, Q386) were mutated as well as ahighly conserved glycine (G393). Numbers in boxes indicate the positions of nonconserved residues (I352, V372, F388, E409, W453, and L462) chosen forreplacement mutagenesis.

KB Patel et al.


titative immunoblotting, as described previously (Pedersen etal. 1999; Lehrer et al. 2007). For the protein standard, we pu-rified a 6xHis-tagged lambda phage lytic transglycosylase(LT). An example of this analysis is shown in Figure 9A.The relative intensities of the bands were determined by den-sitometry, using the program ImageJ, and were used to derive astandard curve to calculate the amount of TrxA–WbaPCT andits derivatives in the total membrane preparations (see Materi-als and methods). The mutant proteins were all well expressed

in MV501, and total membrane fractions showed banding pro-files that were identical to those found in MSS2 (data notshown). To determine the optimal Mg2+ concentration forTrxA–WbaPCT activity in MV501, total membranes were test-ed with various concentrations of Mg2+ and 25 mM was foundto be optimal and was used in the assay (data not shown). All ofthe alanine mutants were tested for activity in vitro. From these,only G393A and D446A were active (136.7 ± 16.2 and 53.9 ±8.7 pmol/mg protein, respectively, compared to 88.7 ±7.9 pmol/mg protein for the parental construct; Figure 9B).Likewise, upon addition of 50 μM exogenous Und-P, onlythe activity of G393A and D446A increased to 950.5 ± 79.5and 318.5 ± 37.7 pmol/mg protein, respectively, compared to539.0 ± 125.0 pmol/mg for the parental construct (Figure 9B).There is no reported structural information on WbaP or any

other members of the PHPT family, and solved structures withenough sequence homology to provide templates for proteinmodeling are not available. However, secondary structure pre-dictions using protein structure prediction (PSI-PRED) indicatethat R377, D382, E383, and D458, are located on three puta-tive α-helical regions. Comparisons with other members of thePHPT family indicate similar predictions for the S. pneumo-niae Cps2E, E. amylovora AmsG, E. coli K-12 WcaJ, G.stearothermophilus WsaP, and Caulobacter crescentus PssY(Figure 10). This suggests that these regions are structurallyconserved and support the notion that they may contribute toform a putative catalytic center.

Discussion

Using a combination of protease accessibility and C-terminalGFP fusion experiments, we have demonstrated that the C-ter-

1395

Fig. 7. Complementation of O antigen synthesis in MSS2 (ΔwbaP) by plasmids encoding site-directed alanine mutants of TrxA–WbaPCT. (A) Mutant proteinsretaining the ability to partially or completely complement O antigen synthesis are marked with an asterisk (*). Only mutations G393A and D446A resulted inproteins with the ability to complement full-length O antigen. LPS samples were prepared and separated by 12% SDS–PAGE. Gels were silver stained and LPSdetected. “Polymeric OAg” denotes the region of the gel containing lipid A-core OS bands covalently linked to O antigen polysaccharides of varying molecularmass. “Core + 1 OAg” indicates the band corresponding to lipid A-core OS with one covalently attached O antigen subunit. (B) Total membranes were prepared andseparated by 14% SDS–PAGE. Proteins were detected with anti-His6 IgG2a monoclonal antibodies. Band shifting was observed with mutants R401A, D446A, andD458A (arrows).

Fig. 8. Site-directed alanine mutations in nonconserved residues along the C-terminus do not affect function. LPS was extracted and analyzed as describedabove.



minus of WbaP, a prototypic member of the PHPT family, iscytoplasmic. This agrees with computer predictions of the to-pology of WbaP (Saldías et al. 2008) and other members of thePHPT family, such as G. stearothermophilus WsaP (Steiner etal. 2007) and S. pneumoniae Cps2E (Xayarath and Yother2007). It was previously established that the C-terminal do-main of WbaP is absolutely required for enzymatic activity,while the N-terminus of WbaP is dispensable for the galacto-syl-phosphate transfer function (Wang et al. 1996; Saldías et al.2008). In this study, we observed a partial complementation ofO antigen synthesis when the C-terminus was overexpressed in

MSS2 under the control of the pBAD promoter, resulting in anO antigen LPS banding pattern consistent with a defect in Oantigen polymerization. This defect could be due to instabilityof the overexpressed protein in the inner membrane or the ac-cumulation of misfolded proteins interfering with O antigenpolymerization. Fusions to proteins such as glutathione S-transferase, maltose binding protein and TrxA can improvefolding of recombinant proteins (Smith and Johnson 1988;LaVallie et al. 1993; Sonezaki et al. 1994). Indeed, overexpres-sion of a TrxA–WbaPCT chimeric protein restored normal Oantigen production by MSS2, with the same banding patternas that observed in the parental strain. GFP fusions to WbaPCTand TrxA–WbaPCT allowed us to better understand the foldingstate of the fusion proteins. These experiments revealed thatthe addition of TrxA enhanced protein folding and led to apunctuate expression of WbaPCT in the membrane, as detectedby fluorescence microscopy. The pattern of distribution ofTrxA–WbaPCT in the bacterial membrane mimics that ob-served with the E. coli WecA protein (Lehrer et al. 2007)and suggests that these proteins could be localized in microdo-mains within the membrane.

The TrxA–WbaPCT fusion protein included TM-V, the lastpredicted TM helix of WbaP (Figure 1). ATEV protease acces-sibility experiment revealed that, contrary to our expectation,the TrxA portion of the fusion protein resided in the cytoplasmor at cytosolic face of the inner membrane. We know from pre-vious work that TM-V is required for function (Saldías et al.2008), but this helix may not be required to span the membraneto yield a functional protein as suggested by the topology ofTrxA–WbaPCT. Several possibilities may account for a func-tional TrxA–WbaPCT in association to the inner membrane,but we favor the idea that the TM-V sequences form a mem-brane hairpin and that distortion of this helix does not affectpossible interactions with the Und-P, also present in the mem-brane, making it possible to retain catalytic activity.

Highly conserved residues have been identified by sequencealignment within proteins of the PHPT family, such as G.stearothermophilus WsaP and C. crescentus HfsE, PssY, andPssZ (Steiner et al. 2007; Toh et al. 2008). To our knowledge,this is the first report investigating these residues in detail todetermine which ones play a critical role in enzyme function.Of the 10 highly conserved polar or charged residues targeted,seven were important for O antigen production and only mu-tation D446A resulted in a protein that retain full functionalityas determined in vivo by complementation of O antigen pro-duction and in vitro by a galactosyl-phosphate transfer assay.In contrast, alanine replacements in nonconserved residues re-sulted in functional proteins that were able to mediateproduction of polymeric O antigen when expressed in MSS2.While some of the identified residues resulting in nonfunction-al proteins may be directly involved in catalysis, substratebinding, and/or metal cofactor coordination, some may alsobe critical to retain the structure of a putative catalytic center.The latter is suggested by the observed band shifting of themutant proteins by immunoblotting, which could be attributedto changes in secondary structure as it has been observed forother integral membrane proteins (Rath et al. 2009).

To further explore a functional role of the critical residuesidentified in this study, we performed conservative replacementswith amino acids that would preserve or reverse polarity without

1396

Fig. 9. In vitro transferase activity of site-directed alanine mutants. (A)Quantitative immunoblotting to determine the amounts of TrxA–WbaPCT andderivative proteins in total membrane preparations. The relative amounts of6xHis molecules in the samples were determined by comparison with knownamounts (50–200 ng) of purified 6xHis-lambda phage lytic transglycosylase(LT), which were included on the same gel. The graph is a plot of the relativepixel densities of the 6xHis–LT bands, which were calculated with the programImageJ, versus the amounts of purified protein loaded in the lanes. The insetshows the blot with the different amounts of 6xHis–LT (left four lanes) andTrxA–WbaPCT (right two lanes). The amount of WbaP6xHis–TrxA–F258–Y476was deduced from a standard curve. (B) Enzyme assays were performed with25 mM Mg2+ and 14C-labeled UDP-galactose with both endogenous Und-P(white bars) and 50 μM exogenous Und-P (gray bars). The lipid phase wasextracted as described in Materials and methods. Activity is expressed asincorporation of galactose-1-P per mg of His-tagged protein at 30 min. Totalmembrane proteins of 40 μg and 50 μg were used for the endogenous andexogenous Und-P experiments, respectively. All assays were carried out at37°C for 30 min in duplicate.

KB Patel et al.


major structural compromise. However, none of these replace-ments resulted in functional proteins. We also replaced thehighly conserved G393 to alanine as an additional control,as we did not expect this residue to be important in catalysisor substrate binding. Although as expected, the mutation didnot affect the ability of the protein to complement O antigensynthesis, we reproducibly observed a higher in vitro activityof the G393A mutant relative to that of the parental construct.This suggests that this residue may be involved in increasingprotein stability or regulating the catalytic reaction. Structuralinformation would be required to better understand the func-tion of G393 in WbaP activity.

Since functionally important residues in the PHPT family oftransferases have not been identified, it is difficult to assign arole to the residues in this study. In the PNPT family, importantresidues have been investigated in the two well-characterizedE. coli members, WecA and MraY. In both proteins, a pairof highly conserved sequential aspartic acid residues, D90/D91 in WecA (Lehrer et al. 2007) and D115/D116 in MraY(Lloyd et al. 2004), are thought to be involved in coordinatingMg2+ or Mn2+. Although WecA and WbaP belong to distinctfamilies of sugar transferases, they require divalent cationsMg2+ or Mn2+ and perform the same biochemical reaction us-ing a UDP-sugar and Und-P, which results in an Und-P-P-sugar product joined by a phosphoanhydride bond formation.Like WecA, WbaP contains two sequential, highly conserved,negatively charged residues, D382/E383, which could be in-volved in Mg2+ coordination.

In summary, we have demonstrated that C-terminal tail of S.enterica WbaP is cytoplasmic and have identified critical ami-no acids in this domain, laying groundwork for furtherinvestigation of PHPT proteins and the elucidation of their cat-alytic mechanism. The ability to stably express the catalyticdomain of WbaP will also aid in studies leading to obtainstructural information, which will ultimately facilitate a betterunderstanding of the mechanism of galactosyl-phosphate trans-fer by this enzyme.

Materials and methodsBacterial strains, plasmids, media and growth conditionsBacterial strains and the plasmids used in this study are listedin Table I. Bacteria were grown aerobically at 37°C in Luria–Bertani (LB) medium (Difco, Franklin lake, New Jersey)(10 mg/mL tryptone, 5 mg/mL yeast extract, 5 mg/mL NaCl).Media were supplemented with 100 μg/mL ampicillin, 30 μg/mL chloramphenicol, or 40 μg/mL kanamycin as appropriate.

Recombinant DNA methods, PCR, and cloning strategiesPlasmid DNAwas isolated using the Qiagen miniprep kit (Qia-gen Inc., Mississauga, Ontario, Canada). Digestion withrestriction enzymes, ligation with T4-ligase, and transformationwere carried out as described by Maniatis et al. (1982). DNAsequences were determined using an automated sequencer atthe York University Core Molecular Biology and DNA Se-quencing Facility, Toronto, Ontario, Canada. Polymerasechain reaction (PCR) amplifications were carried out with PwoIDNA polymerase (Roche Diagnostics, Laval, Quebec, Canada)in a Perkin Elmer 2400 GeneAmp PCR system. The oligonu-cleotides used for these experiments are listed in Table II.Plasmid pKP2 was constructed by PCR amplification of a1062-bp fragment using primers 591 and 2159 and S. entericaserovar Typhimurium LT2 DNA as template. This fragment wasdigested with SmaI and SalI and ligated to these sites of pBAD-FLAG (Table I). pSEF4 was constructed by PCR amplificationusing LT2 DNA as template and primers 3164 and 1152. The1132-bp fragment was digested with SmaI and ligated to this siteof pBADNTF. Plasmid pKP41 was constructed by PCR ampli-fication of a 774-bp fragment, using primers 3599 and 258 andplasmid pSEF4 as a template. This fragment was digested withHindIII and ligated to this site of pEB-T7. To construct pKP42,we used pKP41 as a DNA template and the primers 3601 and3602. The 1311-bp fragment obtained by PCR amplificationwas digested by NcoI and ligated to the NcoI and SmaI sitesof pBAD24. The plasmid pKP43 was constructed in a similar

1397

Fig. 10. ClustalW alignment and PSI-PRED secondary structure prediction of regions of WbaP and homolog proteins containing conserved residues R377, D382,E383, G393, R401, and D458. Conserved residues critical for WbaP activity in vivo and in vitro are indicated with arrowheads and G393 is marked with an asterisk.The sugar transferases compared are as follows: S. enterica Typhimurium WbaP (accession number NP_461027), S. pneumoniae Cps2E (accession numberAAD10174), E. amylovora AmsG (accession number YP_003531611), E. coli K-12 WcaJ (accession number AP_002647), G. stearothermophilusWsaP (accessionnumber AAR99615), and C. crescentus PssY (accession number AAK22153). Below the alignment PSI-PRED secondary structure predictions are depicted with α-helices shown as black boxes and β-strands as white arrows. Residues R377, D382, E383, and D458 are predicted to reside in α-helices, whereas G393 and R401are predicted to be in unstructured regions.



1398

Table I. Bacterial strains and plasmids

Strain or plasmid Relevant propertiesa Source or reference

StrainsS. entericaLT2 Salmonella enterica serovar Typhimurium, wild-type S. MaloyMSS2 LT2, ΔwbaP::cat, CmR Saldías et al. 2008Escherichia coliDH5α F− ϕ80lacZ M15 endA recA hsdR(rK

−mK−)

supE thi gyrA relA? Δ(lacZYA-argF)U169Laboratory stock

MV501 E. coli O7, wecA::Tn10 Alexander and Valvano 1994Plasmidsp0381-FLAG pDA12 derivative encoding BCAL0381

C-terminally fused to the FLAG epitopeAubert et al. 2008

pBAD24 Cloning vector inducible by arabinose, AmpR Guzman et al. 1995pBADFLAG pBAD24 for C-terminal FLAG fusions Amer and Valvano 2000pBADNTF pBAD24 for N-terminal FLAG fusionspEB-T7 Cloning vector, IPTG-inducible, for N-terminal

6xHis fusions AmpRE.H. Ball

pKP2 pBADFLAG derivative encoding the S. enterica LT2 WbaPM1–R354

C-terminally fused to the FLAG epitopeSaldías et al. 2008

pKP10 pBADNTF derivative encoding the S. enterica Ty2 WbaPM1–I144/F258–Y476

N-terminally fused to the FLAG epitopeSaldías et al. 2008

pKP12 pBADNTF derivative encoding the S. enterica Ty2 WbaPF258–Y476N-terminally fused to the FLAG epitope

Saldías et al. 2008

pKP41 pEB-T7 derivative encoding the S. enterica LT2 WbaPF258–Y476N-terminally fused to 6xHis–TrxA

This study

pKP42 pBAD24 derivative encoding the S. enterica LT2 WbaPF258–Y476N-terminally fused to 6xHis–TrxA

This study

pKP50 pKP42 encoding WbaP-R319A This studypKP52 pKP42 encoding WbaP-R377A This studypKP54 pKP42 encoding WbaP-R401A This studypKP55 pKP42 encoding WbaP-D458A This studypKP56 pKP42 encoding WbaP-R377K This studypKP57 pKP42 encoding WbaP-R401K This studypKP58 pKP42 encoding WbaP-D446A This studypKP59 pKP42 encoding WbaP-Y449A This studypKP60 pKP42 encoding WbaP-R319Q This studypKP61 pKP42 encoding WbaP-R401Q This studypKP63 pKP42 encoding WbaP-R377Q This studypKP64 pKP42 encoding WbaP-Y449F This studypKP65 pKP42 encoding WbaP-Y449S This studypKP66 pKP42 encoding WbaP-R368A This studypKP67 pKP42 encoding WbaP-D458E This studypKP68 pKP42 encoding WbaP-Q368A This studypKP69 pKP42 encoding WbaP-K331R This studypKP70 pKP42 encoding WbaP-R333A This studypKP71 pKP42 encoding WbaP-D382A This studypKP73 pKP42 encoding WbaP-G393A This studypKP74 pKP42 encoding WbaP-E383A This studypKP75 pKP42 encoding WbaP-K331A This studypKP80 pBAD24 encoding WbaPF258–Y476 with a C-terminal GFP This studypKP81 pKP80 encoding an N-terminal 6xHis–TrxA fusion This studypKP84 pKP42 encoding WbaP-R319K This studypKP85 pKP42 encoding WbaP-R331Q This studypKP87 pKP42 encoding WbaP-D382E This studypKP88 pKP42 encoding WbaP-D382N This studypKP89 pKP42 encoding WbaP-E383D This studypKP90 pKP42 encoding WbaP-E383Q This studypKP91 pKP42 encoding WbaP-D458N This studypKP92 pKP42 encoding WbaP-I352A This studypKP93 pKP42 encoding WbaP-V372A This studypKP94 pKP42 encoding WbaP-F388A This studypKP95 pKP42 encoding WbaP-E409A This studypKP96 pKP42 encoding WbaP-W453A This studypKP97 pKP42 encoding WbaP-L462A This studypSEF4 pBADNTF derivative encoding S. enterica LT2 WbaPI105–Y476

N-terminally fused to the FLAG epitopeThis study

pWaldo-GFPe Cloning vector, IPTG-inducible, for C-terminal GFP–6xHis fusions, KnR Drew et al. 2006

aCm, chloramphenicol; Amp, ampicillin; Kn, kanamycin.

KB Patel et al.


way by PCR amplification of a 657-bp fragment using LT2DNA as template and primers 3674 and 3675. This fragmentwas digested with KpnI and BamHI and ligated to these sitesof pWaldo-GFPe. Plasmid pKP80 was constructed usingpKP43 as template for PCR with primers 897 and 2991. The1508-bp fragment was digested with SmaI and HindIII and li-gated to these sites of pBAD24. The plasmid pKP81 wasconstructed by PCR amplification of the trxA gene from thepEB-T7 plasmid using primers 3896 and 3897. The 442-bp trxAfragment was digested with SmaI and ligated to this site ofpKP80. Site-directed mutagenesis was performed using theQuikChange Site-Directed Mutagenesis Kit from Stratagene(Santa Clara, California), as recommended by the supplier. Plas-mid pKP42 was used as a template in the PCR reactions for theconstruction of the plasmids in Table I. Primers were designedwith 15–20 nucleotides flanking each side of the targeted muta-tion. The resulting PCR products were digested with DpnI andintroduced into E. coli DH5α by transformation using the cal-cium chloride method (Hanahan 1983). WbaP constructs andreplacement mutants were confirmed by sequencing.

Isolation of spheroplastsBacterial cultures were grown overnight in 5 mL of LB, dilutedto an initial OD600 of 0.2, and incubated at 37°C for 2 h untilcultures reached an OD600 of 0.6. At this point, arabinose wasadded to a final concentration of 0.2% (w/v), and cells were in-cubated for an additional 3 h until reaching an OD600 ofapproximately 0.8–1.0. The cells were placed on ice for20 min, and 7.5 mL volumes were centrifuged at 16,100 × gfor 1 min. The pellet was washed in 1 mL of cold sucrose buffer(18% sucrose, 100 mM Tris–HCl, pH 8.5) and resuspended in1 mL sucrose buffer, 50 μL EDTA (0.5 M), and 50 μL lysozyme(5 mg/mL). Cells were incubated on ice for 1 h and observed us-ing light microscopy after 30 min. Once 90% of cells wereobserved to be spheroplasts, the cells were separated into250 μL aliquots.

Protease treatmentTrypsin accessibility experiments were performed using a pro-tocol adopted from Lee and Manoil (1997). Spheroplasts(250 μL) and membrane vesicles (60 μg of total membrane pro-tein) were treated with 25 μL and 7 μL of trypsin (300 μg/mL),respectively, for 30 min at room temperature. The reactions

were terminated with the addition of one sample volume ofSTI (1 mg/mL) and phenylmethylsulfonyl fluoride (PMSF)(20 mg/mL). Spheroplasts were washed in 250 μL of sucrosebuffer containing 25 μL STI and 5 μL PMSF, and cells wereresuspended in 600 μL of freeze–thaw buffer (10 mM TrispH 8, 10 μL STI [1 mg/mL], 10 μL PMSF [20 mg/mL]). Cellswere lysed with four freeze–thaw cycles using an ethanol–dryice bath and then treated with 20 μL of MgCl2 (1 M), 6 μL ofDNase (1 mg/mL) and protease inhibitor cocktail (Roche Diag-nostics). For TEV cleavage experiments, spheroplasts wereprepared as mentioned above. Samples (150 μL) were treatedwith 2 μL of AcTEV protease (Invitrogen) and 1.5 μL of0.1 M dithiothreitol. After treatment, cells were centrifuged at5900 × g and the supernatant, representing the periplasmic frac-tion in intact spheroplast samples, was collected for analysis.

Total membrane preparation and immunoblottingFor visualization of epitope-tagged WbaP constructs, bacterialcultures were grown and induced with arabinose as describedabove. Cells were then harvested by centrifugation at10,000 × g for 10 min at 4°C. The bacterial pellet was sus-pended in Tris/NaCl (20 mM Tris–HCl [pH 8.5] + 300 mMNaCl) and protease inhibitor cocktail (Roche diagnostics) andthe suspension lysed using a French Press cell. Cell debris wereremoved by centrifugation (15,000 × g for 15 min at 4°C), andthe clear supernatant was centrifuged at 30,000 × g for 30 min at4°C. The pellet, containing total membranes, was suspended inTris/NaCl. The protein concentration was determined by theBradford assay (Bio-Rad, Hercules, California). SDS–PAGE,protein transfers to nitrocellulose membranes, and immunoblotswere performed as described (Perez et al. 2008). For detectionof 6xHis-, FLAG-, and GFP-tagged proteins, membranes wereincubated with a 1:10,000 dilution of anti-His IgG2a monoclo-nal antibodies (Amersham, Piscataway, New Jersey), anti-FLAG monoclonal antibodies (Sigma, St. Louis, Missouri),or anti-GFP monoclonal antibodies (Roche), respectively. Forquantitative immunoblotting, a 6xHis-tagged lambda phageLT was purified and various amounts (50 to 200 ng) of thisstandard were loaded on the same gel as test samples. The pixeldensity of the gel bands was analyzed with ImageJ software(Abramoff et al. 2004) (W.S. Rasband, U.S. National Institutesof Health, Bethesda, MD; http//rsb.info.nih.gov/ij/).

MicroscopyOvernight cultures of MSS2 cells containing pKP80 or pKP81,which express WbaPCT–GFP or TrxA–WbaPCT–GFP undercontrol of the arabinose-inducible pBAD promoter (Table I),were diluted in LB medium to obtain an OD600 of 0.2, and pro-tein expression was induced with arabinose as describedabove. After 2 h of induction, the culture was placed on icefor 1 h to facilitate GFP folding. Similar experiments were per-formed without arabinose in the growth medium. Bacteria werevisualized with no fixation using an Axioscope 2 (Carl Zeiss)microscope with an ×100/1.3 numerical aperture Plan-Neo-fluor objective and a 50-W mercury arc lamp with a GFPband pass emission filter set (Chroma Technology) with exci-tation at 470 ± 20 nm and emission at 525 ± 25 nm. Imageswere digitally processed using the Northern Eclipse imaginganalysis software (version 6.0; Empix Imaging, Mississauga,Ontario, Canada).

1399

Table II. Oligonucleotide primers used in this study

Name Sequence

258 5′-GACCGCTTCTGCGTTCTGAT-3′591 5′-CTCCCCGGGAATGGATAATATTGATAATAAG-3′897 5′-CTCCCCGGGTTTAGCCATGAAGTTATGTTATTAAGG-3′1152 5′-CTCTCTAGATTAATACGCACCATCTCGCCG-3′2159 5′-GACTGTCGACTCCCTGGCAATAGGATCGTTAG-3′2991 5′-GCTAGTTATTGCTCAGCGG-3′3164 5′-CTACCCGGGATAGCGTTTACAAAATGGCAG-3′3599 5′-CGAAAGCTTGTTTAGCCATGAAGTTAT-3′3601 5′-AGGGGAATTGTGAGCGGATA-3′3602 5′-CCTTTCGGGCTTTGTTAGCA-3′3674 5′-CCAGGTACCTTTAGCCATGAAGTTATGT-3′3675 5′-CTATAAGGATCCATACGCACCATCTCGCCG-3′3896 5′-CTACCATGGGCAGCAGCCATCAT-3′3897 5′-CTACCCGGGTCCCTGAAAATACAG-3′



LPS analysisCulture samples were adjusted to OD600 2.0 in a final volumeof 100 μL. Then, proteinase-K-digested whole-cell lysateswere prepared as described elsewhere (Marolda et al. 1990)and LPS was separated on 14% acrylamide gels using a Tri-cine/SDS buffer system (Lesse et al. 1990). Gel loading wasnormalized so that each sample represented the same numberof cells. Each well was loaded with approximately 1 × 108 cfu.Gels were silver stained by a modification of the procedure ofTsai and Frasch (1982).

In vitro transferase assayTotal membranes (containing WbaP and endogenous Und-P)were isolated from strains expressing WbaP constructs orpBAD24. The reaction mixture for the transferase assaycontained the membrane fraction, with 0.025 μCi radiolabeledUDP-14C-Gal (specific activity: 307 mCi/mmol, AmershamBiosciences) in 250 μL buffer (40 mM Tris–HCl, pH 8.5,0.5 mM EDTA, specified MgCl2). After incubation at 37°Cfor 30 min, the lipid-associated material was extracted twicewith 250 μL 1-butanol. The combined 1-butanol extracts werewashed once with 500 μL distilled water, and the radioactivecounts of the 1-butanol fraction were determined with a Beck-man liquid scintillation counter. Enzyme activity wasexpressed as pmol 14C-Gal incorporated per mg of total mem-brane protein−1, or radioactive counts were normalized using a6xHis-tagged protein standard (see above) and expressed aspmol 14C-Gal incorporated per mg protein. Assays using exog-enous Und-P (Institute of Biochemistry and Biophysics of thePolish Academy of Sciences, Warsaw, Poland) were carried outas described above with 50 μM Und-P and 0.1% 3-[(3-chola-midopropyl)dimethylammonio]-1-propanesulfonate in thereaction mixture.

Computer techniquesClustalW was used for sequence alignments. For amino acidconservation, the server SAM-T02 was used (Karplus et al.2003) with the input peptide sequence spanning G310–Y476(Figure 6). Molecular models of WbaPK305–Y476 and homologswere constructed using de novo modeling provided by the Ro-betta server (http://robetta.bakerlab.org) of the University ofWashington. Structure visualization was performed using theprogram Pymol Molecular Graphics System (Delano ScientificLLC; http://pymol.sourceforge.net).

Acknowledgements

The authors thank the colleagues referenced or mentioned inTable I for strains and plasmids, M.S. Saldías for criticalreading of the manuscript, and Ewa Ciepichal for supplyingthe Und-P. This work was supported by a grant from theCanadian Institutes of Health Research to M.A.V. K.P.received an Ontario Graduate Scholarship in Science andTechnology. M.A.V. holds a Canada Research Chair inInfectious Diseases and Microbial Pathogenesis.

Conflict of interest statement

None declared.

Abbreviations

Gal, galactose; GFP, green fluorescent protein; IMVs, invertedmembrane vesicles; LB, Luria–Bertani; LPS, lipopolysaccha-ride; LT, lytic transglycosylase; PCR, polymerase chainreaction; PHPT, polyisoprenyl-phosphate hexose-1-phosphatetransferase; PL2, periplasmic loop 2; PMSF, phenylmethyl-sulfonyl fluoride; PNPT, polyisoprenyl-phosphate N-acetylhexosamine-1-phosphate transferase; PSI-PRED, proteinstructure prediction; SAM, Sequencing Alignment and Model-ing; SDS–PAGE, sodium dodecylsulfate polyacrylamide gelelectrophoresis; STI, soybean trypsin inhibitor; TEV, tobaccoetch virus; TM, transmembrane; TrxA, thioredoxin; UDP, uri-dine diphosphate; Und-P, undecaprenyl phosphate; UV,ultraviolet light; WbaPCT, WbaP C-terminus.

References

Abramoff MD, Magelhaes PJ, Ram SJ. 2004. Image processing with ImageJ.Biophoto Int. 11:36–42.

Alexander DC, ValvanoMA. 1994. Role of the rfe gene in the biosynthesis of theEscherichia coli O7-specific lipopolysaccharide and other O-specific polysac-charides containing N-acetylglucosamine. J Bacteriol. 176:7079–7084.

Amer AO, Valvano MA. 2000. The N-terminal region of the Escherichia coliWecA (Rfe) protein containing three predicted transmembrane helices is re-quired for function but not for membrane insertion. J Bacteriol. 182:498–503.

Amer AO, Valvano MA. 2002. Conserved aspartic acids are essential forthe enzymic activity of the WecA protein initiating the biosynthesis ofO-specific lipopolysaccharide and enterobacterial common antigen inEscherichia coli. Microbiology. 148:571–582.

Arakawa Y, Wacharotayankun R, Nagatsuka T, Ito H, Kato N, Ohta M. 1995.Genomic organization of the Klebsiella pneumoniae cps region responsiblefor serotype K2 capsular polysaccharide synthesis in the virulent strain Che-did. J Bacteriol. 177:1788–1796.

Aubert DF, Flannagan RS, Valvano MA. 2008. A novel sensor kinase-responseregulator hybrid controls biofilm formation and type VI secretion systemactivity in Burkholderia cenocepacia. Infect Immun. 76:1979–1991.

Cartee RT, Forsee WT, Bender MH, Ambrose KD, Yother J. 2005. CpsE fromtype 2 Streptococcus pneumoniae catalyzes the reversible addition ofglucose-1-phosphate to a polyprenyl phosphate acceptor, initiating type 2capsule repeat unit formation. J Bacteriol. 187:7425–7433.

Dai K, Xu Y, Lutkenhaus J. 1996. Topological characterization of the essentialEscherichia coli cell division protein FtsN. J Bacteriol. 178:1328–1334.

Drew D, Lerch M, Kunji E, Slotboom DJ, de Gier JW. 2006. Optimization ofmembrane protein overexpression and purification using GFP fusions. NatMethods. 3:303–313.

Drummelsmith J, Whitfield C. 1999. Gene products required for surface ex-pression of the capsular form of the group 1 K antigen in Escherichia coli(O9a:K30). Mol Microbiol. 31:1321–1332.

Feilmeier BJ, Iseminger G, Schroeder D, Webber H, Phillips GJ. 2000. Greenfluorescent protein functions as a reporter for protein localization in Es-cherichia coli. J Bacteriol. 182:4068–4076.

Geertsma ER, Groeneveld M, Slotboom DJ, Poolman B. 2008. Quality controlof overexpressed membrane proteins. Proc Natl Acad Sci USA. 105:5722–5727.

Guzman LM, Belin D, Carson MJ, Beckwith J. 1995. Tight regulation, mod-ulation, and high-level expression by vectors containing the arabinosePBAD promoter. J Bacteriol. 177:4121–4130.

Hanahan D. 1983. Studies on transformation of Escherichia coli with plasmids.J Mol Biol. 166:557–580.

Huber D, Cha MI, Debarbieux L, Planson AG, Cruz N, Lopez G, Tasayco ML,Chaffotte A, Beckwith J. 2005. A selection for mutants that interfere withfolding of Escherichia coli thioredoxin-1 in vivo. Proc Natl Acad Sci USA.102:18872–18877.

Kadam SK, Rehemtulla A, Sanderson KE. 1985. Cloning of rfaG, B, I, and Jgenes for glycosyltransferase enzymes for synthesis of the lipopolysaccha-ride core of Salmonella typhimurium. J Bacteriol. 161:277–284.

Karchin R, Cline M, Mandel-Gutfreund Y, Karplus K. 2003. Hidden Markovmodels that use predicted local structure for fold recognition: Alphabets ofbackbone geometry. Proteins. 51:504–514.

1400

KB Patel et al.


Karplus K, Karchin R, Draper J, Casper J, Mandel-Gutfreund Y, Diekhans M,Hughey R. 2003. Combining local-structure, fold-recognition, and new foldmethods for protein structure prediction. Proteins. 53(Suppl 6): 491–496.

Katzen F, Ferreiro DU, Oddo CG, Ielmini MV, Becker A, Puhler A, Ielpi L.1998. Xanthomonas campestris pv. campestris gum mutants: Effects onxanthan biosynthesis and plant virulence. J Bacteriol. 180:1607–1617.

LaVallie ER, DiBlasio EA, Kovacic S, Grant KL, Schendel PF, McCoy JM.1993. A thioredoxin gene fusion expression system that circumvents inclu-sion body formation in the E. coli cytoplasm. Nat Biotechnol. 11:187–193.

Lee MH, Manoil C. 1997. Engineering trypsin-sensitive sites in a membranetransport protein. Protein Eng. 10:715–723.

Lehrer J, Vigeant KA, Tatar LD, Valvano MA. 2007. Functional characteriza-tion and membrane topology of Escherichia coli WecA, a sugar-phosphatetransferase initiating the biosynthesis of enterobacterial common antigen andO-antigen lipopolysaccharide. J Bacteriol. 189:2618–2628.

Lesse AJ, Campagnari AA, Bittner WE, Apicella MA. 1990. Increased resolu-tion of lipopolysaccharides and lipooligosaccharides utilizing tricine-sodiumdodecyl sulfate-polyacrylamide gel electrophoresis. J Immunol Methods.126:109–117.

Lloyd AJ, Brandish PE, Gilbey AM, Bugg TD. 2004. Phospho-N-acetyl-mur-amyl-pentapeptide translocase from Escherichia coli: Catalytic role ofconserved aspartic acid residues. J Bacteriol. 186:1747–1757.

Maniatis T, Fritsch EF, Sambrook J. 1982. Molecular Cloning: A LaboratoryManual. Cold Spring, New York: Cold Spring Harbor Laboratory Press.

Marolda CL, Vicarioli J, Valvano MA. 2004. Wzx proteins involved in bio-synthesis of O antigen function in association with the first sugar of theO-specific lipopolysaccharide subunit. Microbiology. 150:4095–4105.

Marolda CL, Welsh J, Dafoe L, Valvano MA. 1990. Genetic analysis of theO7-polysaccharide biosynthesis region from the Escherichia coli O7:K1strain VW187. J Bacteriol. 172:3590–3599.

Morona R, Van Den Bosch L, Daniels C. 2000. Evaluation of Wzz/MPA1/MPA2proteins based on the presence of coiled-coil regions.Microbiology. 146(Pt 1):1–4.

Olsen JV, Ong SE, Mann M. 2004. Trypsin cleaves exclusively C-terminal toarginine and lysine residues. Mol Cell Proteomics. 3:608–614.

Osborn MJ, Rosen SM, Rothfield L, Horecker BL. 1962. Biosynthesis of bac-terial lipopolysaccharide. I. Enzymatic incorporation of galactose in amutant strain of Salmonella. Proc Natl Acad Sci USA. 48:1831–1838.

Pedersen LB, Angert ER, Setlow P. 1999. Septal localization of penicillin-bind-ing protein 1 in Bacillus subtilis. J Bacteriol. 181:3201–3211.

Perez JM, McGarry MA, Marolda CL, Valvano MA. 2008. Functional analysisof the large periplasmic loop of the Escherichia coli K-12 WaaL O-antigenligase. Mol Microbiol. 70:1424–1440.

Raetz CR, Whitfield C. 2002. Lipopolysaccharide endotoxins. Annu RevBiochem. 71:635–700.

Rath A, Glibowicka M, Nadeau VG, Chen G, Deber CM. 2009. Detergentbinding explains anomalous SDS-PAGE migration of membrane proteins.Proc Natl Acad Sci USA. 106:1760–1765.

Reeves P. 1993. Evolution of Salmonella O antigen variation by interspecificgene transfer on a large scale. Trends Genet. 9:17–22.

Riley JG, Menggad M, Montoya-Peleaz PJ, Szarek WA, Marolda CL, ValvanoMA, Schutzbach JS, Brockhausen I. 2005. The wbbD gene of E. colistrain VW187 (O7:K1) encodes a UDP-Gal: GlcNAcα-pyrophosphate-Rβ1, 3-galactosyltransferase involved in the biosynthesis of O7-specific lipo-polysaccharide. Glycobiology. 15:605–613.

Ruiz N, Kahne D, Silhavy TJ. 2009. Transport of lipopolysaccharide acrossthe cell envelope: The long road of discovery. Nat Rev Microbiol. 7:677–683.

Saldías MS, Patel K, Marolda CL, Bittner M, Contreras I, Valvano MA. 2008.Distinct functional domains of the Salmonella enterica WbaP transferasethat is involved in the initiation reaction for synthesis of the O antigen sub-unit. Microbiology. 154:440–453.

Smith DB, Johnson KS. 1988. Single-step purification of polypeptides ex-pressed in Escherichia coli as fusions with glutathione S-transferase.Gene. 67:31–40.

Sonezaki S, Kondo A, Oba T, Ishii Y, Kato Y, Nakayama H. 1994. Overpro-duction and purification of Lon protease from Escherichia coli using amaltose-binding protein fusion system. Appl Microbiol Biotechnol.42:313–318.

Sperandeo P, Deho G, Polissi A. 2009. The lipopolysaccharide transport sys-tem of Gram-negative bacteria. Biochim Biophys Acta. 1791:594–602.

Steiner K, Novotny R, Patel K, Vinogradov E, Whitfield C, Valvano MA,Messner P, Schaffer C. 2007. Functional characterization of the initiationenzyme of S-layer glycoprotein glycan biosynthesis in Geobacillus stear-othermophilus NRS 2004/3a. J Bacteriol. 189:2590–2598.

Stevenson G, Andrianopoulos K, Hobbs M, Reeves PR. 1996. Organization ofthe Escherichia coli K-12 gene cluster responsible for production of the ex-tracellular polysaccharide colanic acid. J Bacteriol. 178:4885–4893.

Toh E, Kurtz HD Jr, Brun YV. 2008. Characterization of the Caulobactercrescentus holdfast polysaccharide biosynthesis pathway reveals signifi-cant redundancy in the initiating glycosyltransferase and polymerasesteps. J Bacteriol. 190:7219–7231.

Tsai CM, Frasch CE. 1982. A sensitive silver stain for detecting lipopolysac-charides in polyacrylamide gels. Anal Biochem. 119:115–119.

Valvano MA. 2003. Export of O-specific lipopolysaccharide. Front Biosci. 8:s452–s471.

Wang L, Liu D, Reeves PR. 1996. C-terminal half of Salmonella entericaWbaP (RfbP) is the galactosyl-1-phosphate transferase domain catalyzingthe first step of O-antigen synthesis. J Bacteriol. 178:2598–2604.

White AP, Gibson DL, Collinson SK, Banser PA, Kay WW. 2003. Extracellularpolysaccharides associated with thin aggregative fimbriae of Salmonella en-terica serovar enteritidis. J Bacteriol. 185:5398–5407.

Wollin R, Creeger ES, Rothfield LI, Stocker BA, Lindberg AA. 1983. Salmonel-la typhimurium mutants defective in UDP-D-galactose:lipopolysaccharidealpha 1, 6-D-galactosyltransferase. Structural, immunochemical, and enzy-mologic studies of rfaB mutants. J Biol Chem. 258:3769–3774.

Xayarath B, Yother J. 2007. Mutations blocking side chain assembly, polymer-ization, or transport of a Wzy-dependent Streptococcus pneumoniae capsuleare lethal in the absence of suppressor mutations and can affect polymertransfer to the cell wall. J Bacteriol. 189:3369–3381.

1401



Functional analysis of the C-terminal domain of the WbaP protein ...

Documents

Transcript of Functional analysis of the C-terminal domain of the WbaP protein ...