• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of jbacterPermissionsJournals.ASM.orgJournalJB ArticleJournal InfoAuthorsReviewers
J Bacteriol. Dec 2010; 192(23): 6160–6171.
Published online Sep 24, 2010. doi:  10.1128/JB.00141-10
PMCID: PMC2981206

Membrane Topology and Identification of Critical Amino Acid Residues in the Wzx O-Antigen Translocase from Escherichia coli O157:H4 [down-pointing small open triangle]

Abstract

Wzx belongs to a family of membrane proteins involved in the translocation of isoprenoid lipid-linked glycans, which is loosely related to members of the major facilitator superfamily. Despite Wzx homologs performing a conserved function, it has been difficult to pinpoint specific motifs of functional significance in their amino acid sequences. Here, we elucidate the topology of the Escherichia coli O157 Wzx (WzxEcO157) by a combination of bioinformatics and substituted cysteine scanning mutagenesis, as well as targeted deletion-fusions to green fluorescent protein and alkaline phosphatase. We conclude that WzxEcO157 consists of 12 transmembrane (TM) helices and six periplasmic and five cytosolic loops, with N and C termini facing the cytoplasm. Four TM helices (II, IV, X, and XI) contain polar residues (aspartic acid or lysine), and they may form part of a relatively hydrophilic core. Thirty-five amino acid replacements to alanine or serine were targeted to five native cysteines and most of the aspartic acid, arginine, and lysine residues. From these, only replacements of aspartic acid-85, aspartic acid-326, arginine-298, and lysine-419 resulted in a protein unable to support O-antigen production. Aspartic acid-85 and lysine-419 are located in TM helices II and XI, while arginine-298 and aspartic acid-326 are located in periplasmic and cytosolic loops 4, respectively. Further analysis revealed that the charge at these positions is required for Wzx function since conservative substitutions maintaining the same charge polarity resulted in a functional protein, whereas those reversing or eliminating polarity abolished function. We propose that the functional requirement of charged residues at both sides of the membrane and in two TM helices could be important to allow the passage of the Und-PP-linked saccharide substrate across the membrane.

Lipopolysaccharide (LPS), a major component of the outer membrane of Gram-negative bacteria, plays critical roles in bacterial cell physiology (36) and in disease (53). The structure of LPS is complex and consists at a minimum of lipid A and core oligosaccharide (OS) (42). Many Gram-negative bacteria also have an O-specific antigen polysaccharide (or O antigen) attached to one of the terminal residues of the core OS (42). The O antigen is the most variable portion of the LPS molecule and arises from the polymerization of discrete oligosaccharide units (42, 54).

The biosynthesis of LPS requires many enzymes and assembly proteins and generally involves two separate pathways. One pathway results in the synthesis of the lipid A-core OS (42), which is translocated across the inner membrane by the lipid A flippase MsbA, an ABC transporter (14, 15, 60). The other pathway involves the synthesis and assembly of the O-antigen polysaccharide, which also begins at the cytosolic side of the inner membrane resulting in the formation of a lipid-linked molecule that is further translocated across the inner membrane. The formation of a complete LPS molecule containing O antigen is catalyzed by the O-antigen ligase WaaL (41). LPS molecules are further translocated to the outer leaflet of the outer membrane by the Lpt transport system involving a number of inner membrane, periplasmic, and outer membrane proteins (44, 45, 48, 49).

There are at least three known mechanisms for the assembly and translocation of lipid-linked O antigens (42, 54). One of them involves a synthase protein that is homologous to processive glycosyltransferases for the synthesis of cellulose and chitin (24, 42). The other mechanism requires ATP hydrolysis for the translocation step, which is mediated by a two-component ABC transporter. This mechanism was initially described for homopolymeric O antigens (42) but also occurs with heteropolymeric O antigens (38). The third mechanism, known as the Wzy-dependent pathway (42, 54), requires three proteins: Wzx (O-antigen translocase), Wzy (O-antigen polymerase), and Wzz (regulator of O-antigen chain-length distribution). This mechanism, used primarily for the synthesis of heteropolymeric O antigens, differs from the other two in that each O unit is separately synthesized and individually translocated across the inner membrane, while the polymerization takes place at the periplasmic side of the membrane (42, 54). The O-antigen precursors are always synthesized as oligosaccharides covalently attached by a phospho-anhydride linkage to an isoprenoid lipid known as undecaprenyl phosphate (Und-P). The formation of the phospho-anhydride linkage is the first committed step toward the synthesis of O antigens and is catalyzed by two classes of membrane enzymes whose prototypes are WecA and WbaP (3, 26, 39, 46, 54). Remarkably, the involvement of an isoprenoid phosphate lipid for these reactions is a common theme in nature and also appears in the synthesis of glycan precursors for cell wall peptidoglycan and in protein glycosylation in bacteria and eukaryotic cells (9, 10). Furthermore, the Wzy-dependent pathway is functionally analogous to the initial steps of dolichol-PP-linked glycans at the endoplasmic reticulum, which are involved in protein N glycosylation (21, 54). Indeed, a membrane protein with roughly similar features as Wzx has been identified in eukaryotic cells as the dolichol-PP-linked glycan flippase and named Rft1 (22).

Our laboratory focuses on the characterization of the Wzy-dependent pathway, and we have previously shown that a single Und-PP-sugar is the minimal substrate for translocation (19, 33). Consistent with this notion, Wzx proteins appear to recognize the Und-PP-bound sugar of the O-antigen unit, irrespective of the composition and structure of the remainder O unit (19, 33). Based on these observations, Wzx proteins can be loosely separated among those that can function with Und-PP-linked N-acetylhexosamines versus those that can function with Und-PP-linked N-hexoses (33). However, comparisons among Wzx primary amino acid sequences do not provide any hints on putative functional residues conserved across the members of this family. It is generally accepted that the translocation process mediated by members of Wzx and Rft1 families does not involve ATP hydrolysis (21, 54), which agrees with the absence of features in the protein that are characteristic of ATP binding or hydrolysis domains. Another complication to investigate functionally the members of these families is the lack of solid topological models that accurately predict transmembrane helices and solvent-exposed loops. Currently, experimentally based topological models have only been established for the Salmonella enterica serovar Typhimurium group B Wzx protein (12), and the Wzx-like protein PssL from Rhizobium leguminosarum (35), which is involved in exopolysaccharide capsule production. However, these studies did not identify any regions or specific amino acids from the protein that could play a functional role in the translocation process. In the present study, we have experimentally characterized the topology of the Wzx protein from Escherichia coli O157 (WzxEcO157) and subjected this protein to extensive mutagenesis by alanine and serine replacements targeting native cysteines and most of the aspartic acid, arginine, and lysine residues. Complementation experiments measuring the ability of each mutant protein to restore O-antigen synthesis in an E. coli K-12 Δwzx mutant resulted in the identification of four charged residues that are required for function, two of which occur in transmembrane helices. Additional replacement mutagenesis revealed that charge but not the nature of the residue is important for Wzx function.

MATERIALS AND METHODS

Bacterial strains, plasmids, and reagents.

Strains and plasmids in the present study are described in Table Table1,1, and oligonucleotide primers are listed in Table Table2.2. Bacteria were cultured in Luria broth (LB) supplemented with antibiotics at the following final concentrations: 100 μg of ampicillin/ml, 30 μg of chloramphenicol/ml, 40 μg of kanamycin/ml, and 80 μg of spectinomycin/ml. Chemicals and antibiotics were purchased from Sigma Aldrich and Roche Diagnostics. Oligonucleotide primers were purchased from Invitrogen. Plasmids were introduced into electrocompetent cells by electroporation (16). Green fluorescent protein (GFP) fusions, PhoA fusions, plasmids, and single replacement constructs were confirmed by sequencing analysis (York University Core Molecular Biology Facility).

TABLE 1.
Strains and plasmids used in this study
TABLE 2.
Primers used in this study

Construction of an E. coli ΔaraCIBAD mutant.

The deletion of the araCIBAD chromosomal genes in W3110 was performed as described by Datsenko and Wanner (13). We generated primers composed of 40 to 45 nucleotides corresponding to regions adjacent to the genes targeted for deletion of the ara genes. The primers also contained 20 additional nucleotides that annealed to the template DNA from plasmid pKD4, which carries a kanamycin resistance gene flanked by FRT (FLP recognition target) sites. A PCR was carried out using the primers 3341 and 3342 (Table (Table2),2), and the DNA product was introduced by electroporation into E. coli W3110(pKD46) competent cells grown in LB containing 0.5% (wt/vol) arabinose. Ampicillin-sensitive (indicating loss of pKD46), kanamycin-resistant (presence of inserted cassette) colonies were screened by PCR using primers annealing to regions outside of the mutated genes (primers 3333 and 3335, Table Table2).2). The antibiotic gene was excised by introducing the plasmid pCP20 encoding the FLP recombinase. Plasmids pKD46 and pCP20 are both thermosensitive for replication, and they were cured at 42°C.

Construction of fusions to green fluorescent protein (GFP).

Plasmid pCM256, encoding WzxEcO157 fused to GFP, was constructed by PCR amplification of a 1.3-kb fragment encoding wzxEcO157 from plasmid pJV7 with the primers 252 and 3370. The PCR product was digested with EcoRI and StuI, while pBADGFP was digested with EcoRI and SmaI. The digested DNA products were ligated (rapid ligation kit; Roche Diagnostics), and the ligation mix was introduced by electroporation into DH5α-competent cells. To construct the plasmids encoding the WzxGFP deleted derivatives, PCR products were amplified by using pCM256 as a DNA template and primers containing StuI restriction sites (underlined in Table Table2).2). A combination of primer GFP and primers K244, T288, Q301, T308, D320, K331, and G397 (Table (Table2)2) were used. The PCR products were digested with StuI and self-ligated. Plasmid pBL1 was constructed by PCR amplification of a 1.3-kb fragment encoding wzxEcO157 from plasmid pJV7, using the primers 252 and 2317. The PCR product and the vector plasmid pBADHIS were digested with NheI and XhoI, ligated, and transformed into DH5α as described above. pBL1 was used as a template to replace all native cysteine residues in WzxEcO157 (residues 13, 102, 130, 434, and 449) with alanine, giving rise to pBL2, pBL3, pBL4, pML202, and pML203 (Table (Table1).1). The replacement of other residues in either pBL1 or pML203 was also performed by site-directed mutagenesis.

Construction of fusions to alkaline phosphatase (PhoA).

The phoA gene was amplified from E. coli JM109 genomic DNA by using the primers 4771 and 4770 (Table (Table2).2). The primers were designed so that only the portion of the gene encoding the mature PhoA protein was amplified. The PCR product and plasmid pBADNTF were both digested with PstI and HindIII and, after ligation, the mixture was introduced into E. coli DH5α cells, resulting in the isolation of pAH18 (Table (Table1).1). A fragment of the wzxEcO157 gene in pJV7 encoding amino acids 1 to K367 was amplified by using the primers 4769 and 4772. The PCR product and pAH18 were both digested with NheI and PstI, and ligation was performed for 15 min at room temperature (using the Roche rapid ligation kit). The ligation mixture was introduced by transformation into E. coli CC118 cells, and the transformants were plated on LB-ampicillin agar with 60 μg of BCIP (5-bromo-4-chloro-3-indolylphosphate [XP])/ml. Plasmids from blue colonies were isolated and sequenced with primers 252 and 4817 (Table (Table2)2) to verify the correct insertion of wzxEcO157-1-K367 into pAH18. The 4772 primer incorporated a PstI site to facilitate the cloning, which resulted in the addition of two amino acids between the K367 and the PhoA protein. One of these plasmids, encoding WzxEcO157-K367-PhoA(PstI), was designated pAH18-K367PhoA(PstI). This plasmid was used to produce further PhoA fusion constructs by inverse PCR amplifications using Pwo DNA polymerase (Roche Diagnostics), the same forward primer 4768, and the reverse primers 4764, 4766, 4767, and 4816 (Table (Table2).2). In this case, the primers were designed to remove the PstI site and the extra two amino acids used initially for constructing WzxEcO157-K367-PhoA(PstI). PCR products were phosphorylated for 30 min with polynucleotide kinase (Roche Diagnostics) and self-ligated for 15 min at room temperature (using a rapid ligation kit). The ligation mixtures were introduced into E. coli CC118 cells, and transformants were screened for blue-colony phenotypes in XP-ampicillin plates. DNA sequencing with the primers 252 and 4817 confirmed the presence of in-frame fusions.

Alkaline phosphatase assay.

To quantify alkaline phosphatase activity, 1:100-diluted overnight cultures were grown for 4 h at 37°C in LB with ampicillin and 0.02% (vol/vol) arabinose, harvested, lysed, and assayed as described previously (29) with the addition of 1 mM iodoacetamide in all buffers. E. coli CC118 cell lysates were used as a negative control.

Microscopy.

An overnight culture of CLM74 cells containing the plasmids encoding the GFP fusion proteins (Table (Table1)1) was diluted in LB to an optical density at 600 nm (OD600) of 0.15, and protein expression was induced with 0.1% (wt/vol) arabinose when the OD600 reached values between 0.6 and 0.7. After a 3-h induction, the culture was placed on ice for 60 min to facilitate GFP folding. Similar experiments were also performed without arabinose in the growth medium. Bacteria were visualized with no fixation using an Axioscope-2 (Carl Zeiss) microscope with an ×100/1.3 numerical aperture Plan-Neofluor objective lens and a 50-W mercury arc lamp with a GFP band-pass emission filter set (Chroma Technology) with a (470 ± 20)-nm excitation range and a (525 ± 25)-nm emission range. Images were digitally processed using the Northern Eclipse version 7.0 imaging analysis software (Empix Imaging, Mississauga, Ontario, Canada).

Labeling of cells and vesicles with sulfhydryl-reactive reagents.

Overnight cultures of DH5α cells containing the appropriate plasmids (Table (Table2)2) supplemented with 100 μg of ampicillin/ml were diluted in LB medium (250 ml) to an OD600 of 0.2; when the OD600 reached values between 0.6 and 0.7, protein expression was induced with 0.1% arabinose. After a 3-h induction, bacteria were harvested at 3,300 × g for 15 min, washed twice with 0.1 M sodium phosphate buffer (pH 7.2), and resuspended in 8 ml of the same buffer. At this point, the bacterial suspension was divided into two 4-ml aliquots to proceed to the biotinylation steps (8). One aliquot was pretreated with 0.5 mM [2-(trimethylammonium)ethyl] methanethiosulfonate bromide (MTSET; Toronto Research Chemicals, Inc., Toronto, Ontario, Canada) for 10 min at room temperature. Both aliquots were then treated with 0.5 mM Nα-(3-maleimidylpropionyl)biocytin (biotin maleimide, BM; Invitrogen) for 60 min at room temperature, with constant rocking. The reaction was terminated by addition of 350 μl of 2% (vol/vol) 2-mercaptoethanol in 0.1 M sodium phosphate buffer (pH 7.2). Treated cultures were then washed twice with 20 ml of phosphate buffer and resuspended in a final volume of 3 ml of the same buffer containing protease inhibitors (Complete Cocktail inhibitor; Roche Diagnostics). Cells were lysed in two passes at 15,000 lb/in2 using a French press, lysates were centrifuged at 39,000 × g for 15 min, and the supernatant was sedimented at 280,000 × g for 30 min. The pellet, containing total membranes, was resuspended in 80 μl of 20 mM Tris-HCl buffer (pH 8) plus protease inhibitor, and the membranes were solubilized for 4 h with 0.5% Triton X-100 in 20 mM Tris-HCl buffer (pH 8) and 8 M urea (pH 8). To label membrane vesicles, total membranes were resuspended in 2 ml of 0.1 M sodium phosphate buffer (pH 7.2), treated with 0.4 M dithiothreitol (DTT; Sigma Chemical Company) for 10 min, and centrifuged at 280,000 × g for 30 min. The membrane pellets were resuspended in 500 μl of 0.1 M sodium phosphate buffer (pH 7.2) and divided into two aliquots of 25 μl each. The volume was brought up to 1 ml with the same buffer, and one aliquot was pretreated for 30 min at room temperature with 125 mM MTSET. Both aliquots were incubated with 0.5 mM BM for 60 min as described for whole cells. The reaction was terminated by addition of 83 μl of 2% (vol/vol) 2-mercaptoethanol in 0.1 M sodium phosphate buffer (pH 7.2), and the membranes were sedimented again. Membrane proteins were solubilized in 0.5% Triton X-100 as described above.

Isolation of His-tagged protein.

Solubilized samples were spun at 39,000 × g for 45 min to remove insoluble material. The supernatant was mixed with 100 μl of Ni-bound chelating Sepharose Fast Flow resin (GE Healthcare), equilibrated with wash buffer (20 mM Tris-HCl, 300 mM NaCl, 100 mM imidazole, 8 M urea [pH 8]), and the mixture was incubated for 2 h at room temperature with gentle mixing. The resin was spun down at 1,000 × g for 2 min, the supernatant was aspirated, and the protein-loaded column was washed three times with 1 ml of wash buffer. WzxEcO157 was eluted by incubating the resin with 70 μl of elution buffer (20 mM Tris-HCl, 300 mM NaCl, 300 mM imidazole, 8 M urea, 0.5% Triton X-100 [pH 8]) for 15 min. Elution fractions were incubated for 30 min at 45°C, separated by SDS-14% PAGE, and transferred to nitrocellulose membranes that were reacted with anti-FLAG monoclonal antibodies (Sigma). Biotinylated proteins were detected by incubation with IRDye 800-streptavidin (Rockland, Gilbertsville, PA).

LPS analysis.

LPS was prepared as previously described (31) from cells grown on LB plates with 0.1% (wt/vol) arabinose, and the samples were separated using 14% (wt/vol) Tricine SDS-PAGE. Gels were stained with silver nitrate as described previously (31, 34). Densitometry of silver stained gels was performed using the program ImageJ (1). Three regions of the gel were considered for the quantitative analysis: the bands in the polymeric O antigen region, the band of lipid A-core OS plus one O antigen unit, and the lipid A-core OS band. The relative amount of O antigen expression was determined by adding the pixels corresponding to the polymeric O antigen region of the gel (previously subtracted from background pixels) to the pixels corresponding to the lipid A-core OS plus one O antigen unit and dividing this value by the pixels of the lipid A-core OS band. The results were expressed as percent values relative to the positive control (100%).

Protein analysis.

Total membranes were prepared from cells grown under the conditions described above for preparation of the LPS samples. Bacterial cells were suspended in 20 mM Na2PO4 (pH 7.2) plus protease inhibitors (Complete Tablets; Roche Diagnostics), and they were lysed by sonic disruption for two 15-s pulses (Branson). Total membrane fractions were obtained by centrifugation of the lysates for 40 min at 40,000 × g, and the pellet was resuspended in the same buffer. The protein concentration was measured by using a Bradford protein assay (Bio-Rad). Portions (20 to 40 μg) of total membrane proteins were incubated for 30 min at 45°C, separated on by SDS-14% PAGE, and transferred to nitrocellulose membranes that were reacted with one of the following: anti-FLAG polyclonal rabbit antibodies (Rockland), anti-FLAG monoclonal antibodies (Sigma), or anti-GFP monoclonal antibodies (Roche Diagnostics). The reacting bands were detected by determining the fluorescence with an Odyssey infrared imaging system (Li-Cor Biosciences) using IRDye800 CW affinity-purified anti-rabbit IgG antibodies (Rockland) and Alexa Flour 680-labeled anti-mouse IgG antibodies (Invitrogen).

RESULTS

Topological analysis of WzxEcO157 by the substituted cysteine accessibility method.

We attempted to establish a working topological model for WzxEcO157 using commonly used algorithms that predict the number and location of transmembrane helices and the orientation of the intervening loops (17, 37). The programs HMMTOP (52), MEMSAT (23), TMHMM (50), and Octopus (55) predicted 12 TM helices with six periplasmic and five cytoplasmic loops, while TOPPRED (11) predicted only 10 TM helices with five periplasmic and four cytoplasmic loops (Fig. (Fig.1).1). All of the programs predicted that the N terminus and C terminus of the protein were located in the cytoplasm. However, even with the programs predicting the same number of TM helices, each of the models obtained differed in the length and orientation of periplasmic and cytoplasmic loops, particularly in the region between amino acids 260 and 425 (Fig. (Fig.1,1, square). Given the difficulties in producing a reliable topological model for WzxEcO157 by computer predictions, we resorted to using SCAM (substituted cysteine accessibility method [8]). The native WzxEcO157 protein contains five cysteines at positions 13, 102, 130, 434, and 449, which were all replaced by alanine in a sequential manner, resulting in WzxEcO157 forms with one or more cysteine replacements. These proteins were engineered to contain an N-terminally fused FLAG epitope for detection by Western blotting and a C-terminal 7×His to facilitate their isolation by Ni2+-affinity chromatography after labeling with biotin maleimide (see below). To determine whether these proteins support O antigen expression, we prepared LPS samples from CLM17(pMF19) bacteria containing the appropriate plasmid constructs. We have demonstrated earlier that the WzxEcO157 protein can restore O16 antigen synthesis in the ΔwzxEcO16 mutant with the same efficiency as WzxEcO16 (33). The pMF19 plasmid contains the wbbL gene encoding a rhamnosyltransferase that allows for the completion of O16 LPS synthesis in the E. coli K-12 W3110 strain and its derivatives (19, 28, 33). The WzxEcO157 mutant proteins lacking one or more native cysteines (Fig. (Fig.2A,2A, lanes 3 to 7) supported O-antigen production with the same banding pattern characteristics as the one mediated by the parental WzxEcO157-FLAG-7xHis protein (Fig. (Fig.2A,2A, lane 2). Cells expressing mutant proteins containing quadruple or quintuple cysteine replacements also exhibited a small reduction in the amount of polymeric O antigen (Fig. (Fig.2A,2A, lanes 6 to 7). This reduction could be due instability of the mutated protein in the membrane. However, all of the proteins were found at roughly the same level in the membrane fractions (Fig. (Fig.2B,2B, lanes 2 to 7), and our method to prepare membrane fractions affords a mixture of outer and inner membrane proteins with negligible contamination of cytoplasmic proteins (4-6, 26, 32, 33, 56). Therefore, the experimental results demonstrate that replacing all of the native cysteines in WzxEcO157-FLAG-7×His does not significantly affect protein expression, membrane localization, and O-antigen production.

FIG. 1.
Graphical representation of the topological predictions by the various programs used here. The numbers indicate the position of the amino acids in WzxEcO157 (463 amino acids). The location of soluble segments (cytosolic or periplasmic) and the positions ...
FIG. 2.
LPS profiles (A) and protein expression (B) of CLM17(pMF19) cells expressing parental WzxEcO157 or the constructs containing the single (wzx-1xC), double (wzx-2xC), triple (wzx-3xC), quadruple (wzx-4xC), and quintuple ...

We used the WzxEcO157-Cys-less protein to introduce novel cysteine replacements at various positions for topological analysis (Table (Table22 and Fig. Fig.3).3). All replacements resulted in proteins that were detectable by Western blotting with anti-FLAG antibodies, confirming that the cysteine replacements are well tolerated and do not affect protein stability or targeting to the plasma membrane (data not shown). The accessibility of the novel cysteines to the sulfhydryl reactive reagent biotin maleimide was determined by incubating whole cells with the label, followed by treatment with excess β-mercaptoethanol before bacterial cell lysis. This treatment prevents the labeling of any cysteine that could become exposed to biotin maleimide during cell fractionation. Biotin maleimide is membrane permeable and reacts with thiol groups that are next to water molecules since the reaction with an ionized thiol group requires a water molecule as a proton acceptor (8). Therefore, cysteines buried in the core of the hydrophobic transmembrane segments are usually not labeled (8). Cysteine-substituted K144, K367, R298, H209, and D277 were accessible to biotin maleimide (Fig. (Fig.4A,4A, lanes 3, 7, 9, 11, and 13), suggesting that these residues are exposed to the labeling reagent (Fig. (Fig.22 and Table Table2).2). In contrast, the parental WzxEcO157-Cys-less and the cysteine-substituted forms at P313, F293, and P306 positions were not labeled (Fig. (Fig.4A,4A, lanes 1, 5, 15, and 17), indicating that these residues are not exposed to biotin and suggesting that they are in close proximity to the inner membrane or buried within the membrane bilayer. Similarly, cysteine substitutions at positions K244, D320, R324, and K331 were not labeled in whole cells (Table (Table22 and data not shown).

FIG. 3.
Topological model of WzxEcO157 by a combination of bioinformatics, GFP and PhoA deletion-fusion analyses, and substituted cysteine accessibility (SCAM) experiments. The model was originally derived according to the HHMTOP computer program and graphically ...
FIG. 4.
Labeling experiments with biotin-maleimide (BM) with (−) or without (+) MTSET pretreatment performed on cysteine replacement mutants of WzxEcO157. Labeled proteins were isolated by NTA-Ni2+ affinity chromatography and detected ...

To determine whether the labeled amino acids were located at the periplasmic face of the inner membrane we used MTSET, a charged thiol-specific probe that reacts with sulfhydryl groups under similar conditions as biotin maleimide but is impermeable to the cytoplasmic membrane due to its positive charge (8). Incubation of whole bacterial cells with MTSET prior to treatment with biotin maleimide prevents labeling of periplasmic cysteines (26). Accordingly, pretreatment with MTSET prevented the labeling with biotin maleimide of cysteine-substituted K144, K367, R298, His209, and D277 residues (Fig. (Fig.4A,4A, lanes 4, 8, 10, 12, and 14), suggesting that these residues are exposed to the periplasmic face of the inner membrane. These results agree with the control experiment using WecAS362C (not protected by MTSET) and WecAG181S (protected by MTSET), which contain cysteine replacements in cytosolic and periplasmic residues, respectively (26; data not shown).

Despite the fact that biotin maleimide is supposed to be membrane permeable, we could not confirm by labeling in whole cells the location of cysteine-substituted residues at positions K244, D320, R324, and K331, which were all expected to be in the cytosolic loops 3 and 4 (Fig. (Fig.3).3). To resolve the location of these residues, we performed the labeling experiment on isolated membrane vesicles. This strategy permitted us to biotinylate all periplasmic and cytoplasmic exposed residues. Cysteine-substituted proteins at D320, K331, K367, and R324 positions were labeled with biotin, but labeling was prevented by MTSET (Fig. (Fig.4B,4B, lanes 3 to 10), indicating that these residues are surface exposed. From these, the cysteine-substituted K367 residue was the only one that could be labeled in whole cells and also in membrane vesicle preparations, and in both cases labeling was prevented by pretreatment with MTSET (Fig. 4A and B, lanes 7 and 8). Therefore, this residue is unequivocally located in the predicted periplasmic loop 5 (Fig. (Fig.3).3). Because cysteine replacements at D320, K331, and R324 could not be labeled with biotin maleimide upon treatment of whole cells (Table (Table22 and data not shown) but were detectable in membrane vesicles (Fig. (Fig.4B,4B, lanes 3 to 6, 9, and 10), we concluded that these residues are exposed to the cytoplasmic face of the inner membrane. Vesicles prepared from cells expressing the Cys substitution at Phe264 did not react with biotin maleimide (Fig. (Fig.4B,4B, lanes 1 and 2), nor did vesicles containing Wzx proteins with replacements at P254, G286, F293, P306, and P313 (data not shown). From these, cysteines at positions P254, P306, and P313 were also not labeled in whole cells (Table (Table2),2), suggesting that they are buried in the membrane. G286 and F293 span an 18-amino-acid (from F279 to T296) region containing hydrophobic residues flanked by D277 and R298, which were both mapped to the periplasmic space (Fig. (Fig.3)3) since they strongly react with biotin maleimide and labeling is prevented by MTSET in whole cells (Fig. (Fig.4A,4A, lanes 9, 10, 13, and 14). Therefore, it is unlikely that this region can form a transmembrane helix, but it could still have a secondary structure that would prevent access to biotin maleimide for labeling, as we have previously shown for an analogous short hydrophobic region within the large cytoplasmic loop 5 of WecA (26). Together, the SCAM results allowed us to construct an experimentally supported topological model of WzxEcO157 (Fig. (Fig.33).

Additional topological analysis by targeted fusions to GFP and alkaline phosphatase.

To get additional evidence for the topological assignments from the SCAM, we constructed a WzxEcO157 derivative C-terminally fused to GFP (encoded by pCM256, Table Table1),1), a reporter that can only fluoresce if present in the cytosol (18). As before, the WzxEcO157 protein expressed from pCM256 also contains an N-terminal FLAG epitope. The WzxEcO157-GFP fusion protein afforded an ~80-kDa polypeptide band in Western blots of total membrane preparations reacted with anti-GFP and anti-FLAG antiserum, indicating that Wzx is correctly fused to the GFP protein and that is also in the membrane fraction (Fig. (Fig.5A,5A, lanes 4 and 6, black arrowheads). In contrast, the membrane fraction from E. coli cells expressing the parental WzxEcO157-FLAG-7xHis shows an ~53-kDa band that is only detectable with anti-FLAG antibodies (Fig. (Fig.5A,5A, lane 5, white asterisk), in agreement with the predicted mass of this protein. As a control, we detected a 27-kDa band corresponding to soluble GFP in the cell lysate from DH5/pBADGFP (Fig. (Fig.5A,5A, lane 1, asterisk). Additional bands of large molecular mass reacting with either anti-GFP or anti-FLAG antibodies are also detected (Fig. (Fig.5A,5A, lanes 1 and 4 to 6). These were likely oligomeric forms of Wzx (lanes 4 to 6) and GFP (lane 1) due to the incomplete denaturing conditions used to prepare the samples. Complete denaturation including treatment of the samples by heating at 100°C in SDS prevents the detection of membrane proteins in SDS-PAGE, since we have previously observed with various membrane proteins containing multiple transmembrane helices (26, 41, 51, 56), including several other Wzx proteins (33).

FIG. 5.
Analysis of wzxEcO157-GFP fusion constructs. For panels A and B, protein samples were prepared as indicated in the legend to Fig. Fig.1.1. (A) Expression of wild-type wzxEcO157-GFP fusion protein. (B) Expression of deleted wzxEcO157-GFP fusion ...

Using the wzxEcO157-gfp gene encoded in pCM256 as a starting point, various deletions were constructed that resulted in C-terminally truncated forms of WzxEcO157 remaining C-terminally fused to GFP. The deletion-fusion endpoints chosen corresponded to amino acids K244, T288, Q301, T308, D320, K331, and G397 of the native WzxEcO157 protein (Table (Table1).1). These endpoints were chosen to help resolve the topology of the most difficult part of WzxEcO157, where most of the prediction programs differed (Fig. (Fig.1,1, square). The expected fusion was produced by all plasmid constructs, as suggested by the detection with anti-GFP antibodies of polypeptide bands of decreasing molecular mass, ranging from 80 kDa for the parental WzxEcO157-GFP (Fig. (Fig.5B,5B, lane 1) to 55.3 kDa for WzxEcO157-K244-GFP (Fig. (Fig.5B,5B, lane 8).

The plasmids encoding parental and deleted versions of WzxEcO157-GFP fusions were introduced into the E. coli cells CLM74 by electroporation. CLM74 has a deletion of the araCIBAD chromosomal genes (Table (Table1),1), which reduces the toxic effects of arabinose and its metabolites on E. coli K-12 strains (43). We have observed that arabinose toxicity affects cell growth and leads to altered cell morphologies upon membrane protein expression driven by cloned genes under the control of arabinose-inducible promoters (data not shown). Bacterial cells expressing parental WzxEcO157-GFP displayed fluorescence around the cell periphery detectable after a 0.85-s exposure to UV light (Fig. (Fig.1D),1D), confirming the prediction that Wzx is in the membrane with the C terminus facing the cytosol (Fig. (Fig.3).3). Cells expressing WzxEcO157-K244-GFP, WzxEcO157-K331-GFP, and WzxEcO157-G397-GFP (Fig. (Fig.5C)5C) also displayed fluorescence around the periphery but with less intensity (samples exposed for 3 s to be able to detect a fluorescent signal), suggesting that residues K244, K331, and G397 are exposed to the cytosol. The cytosolic locations of K244, D320, and K331 agreed with the SCAM data (Fig. (Fig.33 and and4).4). Also, these results support the assignment of R324 at the cytoplasmic side of the membrane given its proximity to D320 and K331 (Fig. (Fig.33).

In CLM74 cells expressing WzxEcO157-D320-GFP, WzxEcO157-Q301-GFP, and WzxEcO157-T308-GFP the fluorescence was very faint, requiring 4-s exposures, and also present within the cell bodies, while cells expressing and WzxEcO157-T288-GFP did not fluoresce even after longer exposure (Fig. (Fig.5D5D and data not shown). We interpreted these results as an indication that residues D320, Q301, T308, and T288 are not exposed to the cytosol. To verify this, we performed similar deletion-fusion experiments using the alkaline phosphatase PhoA as a reporter for periplasmic localization. This protein can only fold properly and therefore becomes enzymatically active if exported to the periplasmic space (30). Although we originally intended to obtain protein fusions at endpoints identical to those for GFP fusions, this was only achievable in some cases, while in others only small in-frame deletions were recovered that resulted in fusions to nearby residues. Also, despite numerous attempts we could not obtain a WzxEcO157-PhoA C-terminal fusion, concluding that this fusion is probably toxic to E. coli cells (data not shown). E. coli CC118 cells (Table (Table1)1) containing plasmids encoding WzxEcO157-PhoA fusions with deletion-fusion endpoints at amino acids W139, T288, V360, and K367 gave a strong blue-colony phenotype in XP plates (data not shown). In agreement with these results, cell-free lysates of CC118 bacteria expressing PhoA fusions at W139, T288, V360, and K367 yielded 10,754 ± 390, 1,372 ± 75, 3,288 ± 237, and 3,134 ± 170 U of alkaline phosphatase activity, respectively, suggesting that these residues are exposed to the periplasmic space. This agrees with data obtained from the SCAM method for residues at the fusion endpoint (K367) or for residues near those experimentally determined as exposed to the periplasm (W139 near K144 and V360 near K367; Fig. Fig.3)3) and also with the absence of GFP-mediated fluorescence in the case of T288-GFP fusion. In contrast, expression of the WzxEcO157-F242-PhoA fusion only afforded white colonies, and cell-free lysates had no enzymatic activity. F242 is near K244, based on GFP fusion data, and SCAM can unequivocally be placed in cytoplasmic loop 3 (Fig. (Fig.3).3). Thus, the combination of fusion experiments with GFP and PhoA reporters and the SCAM methods confirmed the topological assignment of WzxEcO157 (Fig. (Fig.3)3) as a membrane protein with 12 TM helices, 6 periplasmic and 5 cytoplasmic loops, and the C terminus in the cytosol (Fig. (Fig.33).

Identification of functional residues in WzxEcO157.

We investigated whether any of the cysteine-substituted WzxEcO157 mutants was unable to support O-antigen surface expression in vivo using CLM17(pMF19). Bacteria containing WzxEcO157 proteins with cysteine replacements at K244, K331, K367, and K402 showed a moderate reduction in surface O-antigen production, ranging from 60 to 90% compared to cells expressing the parental cysteine-less WzxEcO157 (Table (Table33 and data not shown). Substitutions P313C, D320C, and K176C resulted in proteins greatly compromised in their ability to support O-antigen production, as demonstrated by a reduction in surface O antigen ranging from 30 to 50% relative to levels in the parental strain (Table (Table3).3). In the case of the WzxEcO157 mutant R298C, there was no O-antigen production (Table (Table3).3). The differences in O-antigen production observed in these mutant proteins were not due to lack of protein expression, as confirmed by Western blotting with anti-FLAG antibodies (data not shown). That the mutant proteins were detected in total membrane fractions suggests that they are sufficiently similar to the parental derivative to least be inserted in the membrane, although we cannot rule out localized changes in the secondary structure. Because the cysteineless form of WzxEcO157 mediated a somewhat reduced production of O antigen compared to the parental protein with its native cysteines (Fig. (Fig.2A,2A, compare lanes 2 and 7) and the replacement of a native amino acid by a cysteine at certain positions could lead to local structural changes due to differences in hydrophobicity and the oxidation state of the SH side chain, we reconstructed the cysteine substitutions that caused a functional defect in WzxEcO15 as alanine or serine replacements in the parental protein. Alanine and serine are usually much better tolerated since, because of their small mass, they generally cause only minor conformational changes in the protein structure (7). Also, serine favors the formation of loops (25) and therefore would not be expected to alter the extracellular loops of WzxEcO157. Using this approach, we confirmed that only WzxEcO157-D326A and WzxEcO157-R298A were functionally impaired in mediating O-antigen production showing 60% and no O antigen, respectively (Fig. (Fig.6A,6A, lanes 9 and 15), while the rest of the replacements did not compromise WzxEcO157 functionality.

FIG. 6.
O-antigen and WzxEcO157 protein expression in CLM17(pMF19) cells containing plasmids encoding protein constructs with alanine or serine amino acid replacements. (A) Amino acid replacements located on either periplasmic or cytosolic loops. (B) Amino acid ...
TABLE 3.
Properties of the replacement mutants of the cysteineless WzxEcO157

Charged residues are rarely found in the interior of transmembrane helices (57). Therefore, we also investigated the functional contribution of the charged residues D85, K159, D385, and K419 in transmembrane helices 2, 4, 10, and 11, respectively. We also investigated the functional role of Q223 and Q267 in transmembrane helices 6 and 7, since residues at analogous positions are important for the function of the LacY permease (2), a prototypic member of the major facilitator superfamily. As shown in Fig. Fig.6B,6B, lanes 3 and 8, only D85 and K419 are required for WzxEcO157 function since the mutant proteins WzxEcO157-D85A and WzxEcO157-K419S were unable to mediate O-antigen production. The lack of function was not due to loss of protein expression since all mutant proteins were produced at similar levels by Western blotting (data not shown). From these experiments, we concluded that residues D85, R298, D326, and K419 are required for the normal function of WzxEcO157.

To gain more information on the role of these residues, we constructed additional replacement mutants geared to cause charge modifications. Thus, D85 and D326 were also replaced with glutamic acid and arginine, R298 was replaced with aspartic acid and lysine, and K419 was replaced with arginine and aspartic acid. The plasmids expressing the replaced proteins were examined for their ability to support O antigen production in CLM17(pMF21) bacteria. Only substitutions preserving the net charge at each residue were functional (Fig. (Fig.7),7), while substitutions causing a charge reversal were not functional as with the control replacements to alanine or cysteine. From these experiments, we conclude that the net charge of these residues is crucial for WzxEcO157 activity.

FIG. 7.
O-antigen (upper panel) and wzxEcO157 (lower panel) expression of CLM17(pMF19) cells containing constructs with single amino acids replacements, as indicated, at positions R298, D85, D326, and K419. LPS and proteins were prepared and analyzed as described ...

DISCUSSION

A combination of deletion and fusions to the GFP and PhoA proteins as topology probes and biotin-maleimide labeling of cysteine replacement mutants allowed us to construct an experimentally validated topological model for WzxEcO157 that contains 12 TM helices and 6 periplasmic and 5 cytoplasmic loops. A cysteineless version of WzxEcO157 that remained functional and localized to the membrane indicated that the native cysteines are dispensable for Wzx function. Some interesting observations can be drawn from the topology of WzxEcO157. First, four TM helices (II, IV, X, and XI) contain polar residues (aspartic acid or lysine), which are atypical residues for a TM location. Second, some of the periplasmic loops contain stretches of hydrophobic residues flanked by polar residues (D277-R298 in periplasmic loop 4 and E358-K367 in periplasmic loop 5). De novo modeling of the D277-R298 segment predicts that the intervening amino acids form two β-strands in parallel orientation (data not shown). This arrangement could be important for interactions with the Und-PP-O unit substrate. R298, which is required for a functional WzxEcO157, could be involved in interactions with the phosphates of the Und-PP. Similarly, we have recently shown that positively charged periplasmic residues in the O-antigen ligase WaaL are prime candidates to interact with the phosphate residues of Und-PP-linked saccharides (41). Whether this is also true for WzxEcO157 will require additional experiments to test specific interactions of this protein with Und-PP-linked O-antigen precursors.

With the programs we used, the periplasmic loops containing hydrophobic residues correspond to the part of the protein that are most difficult to predict topologically in a consistent manner. Similar features appear in the topological models of Wzx proteins from S. enterica (WzxSe) (12) and R. leguminosarum (PssL) (35). In both proteins there are also four transmembrane domains, each containing lysine or arginine residues, and two periplasmic loops with short stretches of hydrophobic amino acids. The only major difference between these proteins and WzxEcO157 is that R. leguminosarum PssL has a larger cytosolic loop between TM helices 6 and 7, which is absent in the E. coli O157 and S. enterica homologs. Paulsen et al. (40) have classified proteins involved in lipid-linked saccharide transport into one family with two subclasses based on the presence of the large predicted cytosolic loop. However, it remains unclear whether this has any functional significance and, unfortunately, it has not been possible to obtain R. leguminosarum pssL mutants (35).

As an attempt to identify functional amino acids, we utilized alanine and/or serine replacement mutagenesis, taking advantage of the topological model. We were particularly interested in targeting the charged residues in TM helices and also residues in periplasmic and cytosolic loops. Each mutant protein was examined for expression and localization to the inner membrane. We have found with other membrane proteins, such as WecA and WaaL, that mutations affecting protein insertion in the membrane result in no detectable protein. By these criteria, we assumed that all of the replacement mutants behave identically to the parental WzxEcO15, although small local changes could not be identified. Only D85, R298, D326, and K419 yielded nonfunctional proteins when replaced by alanine. Furthermore, conservative replacements at these positions demonstrated that the charge, but not the nature of the targeted amino acid, is critical to retaining the ability of WzxEcO157 to support O-antigen production. These results suggest that residues at these positions could be involved in making contacts with substrates directly or via water molecules.

From our current data, we propose that Wzx proteins have similar features to transporters of the major facilitator superfamily. This is based on the following: (i) the presence of at least four TM helices with charged amino acids, suggesting the possibility of a tertiary structure such that a core of TM helices interact with each other and are located further from the lipid bilayer, in an arrangement similar to that found for LacY (47), and (ii) the functional requirement of charged residues at both sides of the membrane and in two TM helices, which could be important for creating an electrostatic cavity (2, 47) and perhaps even electrostatic interactions with the phosphate groups of Und-PP-linked sugars, which may in turn allow localized perturbation of the lipid bilayer to facilitate the movement of the Und-PP-linked saccharide substrate across the membrane. These possibilities are supported by previous results of in vitro studies of interactions between hydrophobic peptides with lipid vesicles containing isoprenoid phosphates and molecular modeling of isoprenoid phosphates in artificial membranes (58, 59). Further experiments involving additional mutagenesis, neighborhood analysis of TM helices, and Und-PP-saccharide substrate binding assays are required to unequivocally identify functional regions of WzxEcO157 based on an experimentally established topological model and to begin elucidating the mechanism of translocation.

Acknowledgments

We thank the coworkers referenced or mentioned in Table Table11 for strains and plasmids.

This study was supported by grants from the Canadian Institutes of Health Research and the Mizutani Foundation for Glycoscience. Summer Research Awards from the Natural Sciences and Engineering Research Council of Canada supported B.L. and M.L. M.A.V. holds a Canada Research Chair in Infectious Diseases and Microbial Pathogenesis.

Footnotes

[down-pointing small open triangle]Published ahead of print on 24 September 2010.

REFERENCES

1. Abramoff, M. D., P. J. Magelhaes, and S. J. Ram. 2004. Image processing with ImageJ. Biophotonics Int. 11:36-42.
2. Abramson, J., I. Smirnova, V. Kasho, G. Verner, H. R. Kaback, and S. Iwata. 2003. Structure and mechanism of the lactose permease of Escherichia coli. Science 301:610-615. [PubMed]
3. Alexander, D. C., and M. A. Valvano. 1994. Role of the rfe gene in the biosynthesis of the Escherichia coli O7-specific lipopolysaccharide and other O-specific polysaccharides containing N-acetylglucosamine. J. Bacteriol. 176:7079-7084. [PMC free article] [PubMed]
4. Amer, A. O., and M. A. Valvano. 2001. Conserved amino acid residues found in a predicted cytosolic domain of WecA (UDP-N-acetyl glucosamine:undecaprenol-phosphate N-acetylglucosamine-1-phosphate transferase) are implicated in the recognition of UDP-N-acetylglucosamine. Microbiology 147:3015-3025. [PubMed]
5. Amer, A. O., and M. A. Valvano. 2002. Conserved aspartic acids are essential for the enzymic activity of the WecA protein initiating the biosynthesis of O-specific lipopolysaccharide and enterobacterial common antigen in Escherichia coli. Microbiology 148:571-582. [PubMed]
6. Amer, A. O., and M. A. Valvano. 2000. The N-terminal region of the Escherichia coli WecA (Rfe) protein containing three predicted transmembrane helices is required for function but not for membrane insertion. J. Bacteriol. 182:498-503. [PMC free article] [PubMed]
7. Baase, W. A., A. E. Eriksson, X. J. Zhang, D. W. Heinz, U. Sauer, M. Blaber, E. P. Baldwin, J. A. Wozniak, and B. W. Matthews. 1992. Dissection of protein structure and folding by directed mutagenesis. Faraday Discuss. 1992:173-181. [PubMed]
8. Bogdanov, M., W. Zhang, J. Xie, and W. Dowhan. 2005. Transmembrane protein topology mapping by the substituted cysteine accessibility method (SCAM™): application to lipid-specific membrane protein topogenesis. Methods 36:148-171. [PMC free article] [PubMed]
9. Bugg, T. D., and P. E. Brandish. 1994. From peptidoglycan to glycoproteins: common features of lipid-linked oligosaccharide biosynthesis. FEMS Microbiol. Lett. 119:255-262. [PubMed]
10. Burda, P., and M. Aebi. 1999. The dolichol pathway of N-linked glycosylation. Biochim. Biophys. Acta 1426:239-257. [PubMed]
11. Claros, M. G., and G. von Heijne. 1994. TopPred II: an improved software for membrane protein structure predictions. Comput. Appl. Biosci. 10:685-686. [PubMed]
12. Cunneen, M. M., and P. R. Reeves. 2008. Membrane topology of the Salmonella enterica serovar Typhimurium group B O-antigen translocase Wzx. FEMS Microbiol. Lett. 287:76-84. [PubMed]
13. Datsenko, K. A., and B. L. Wanner. 2000. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. U. S. A. 97:6640-6645. [PMC free article] [PubMed]
14. Doerrler, W. T., and C. R. H. Raetz. 2002. ATPase activity of the MsbA lipid flippase of Escherichia coli. J. Biol. Chem. 277:36697-36705. [PubMed]
15. Doerrler, W. T., M. C. Reedy, and C. R. H. Raetz. 2001. An Escherichia coli mutant defective in lipid export. J. Biol. Chem. 276:11461-11464. [PubMed]
16. Dower, W. J., J. F. Miller, and C. W. Ragsdale. 1988. High efficiency transformation of Escherichia coli by high-voltage electroporation. Nucleic Acids Res. 16:6127-6145. [PMC free article] [PubMed]
17. Drew, D., D. Sjöstrand, J. Nilsson, T. Urbig, C.-N. Chin, J.-W. de Gier, and G. von Heijne. 2002. Rapid topology mapping of Escherichia coli inner-membrane proteins by prediction and PhoA/GFP fusion analysis. Proc. Natl. Acad. Sci. U. S. A. 99:2690-2695. [PMC free article] [PubMed]
18. Feilmeier, B. J., G. Iseminger, D. Schroeder, H. Webber, and G. J. Phillips. 2000. Green fluorescent protein functions as a reporter for protein localization in Escherichia coli. J. Bacteriol. 182:4068-4076. [PMC free article] [PubMed]
19. Feldman, M. F., C. L. Marolda, M. A. Monteiro, M. B. Perry, A. J. Parodi, and M. A. Valvano. 1999. The activity of a putative polyisoprenol-linked sugar translocase (Wzx) involved in Escherichia coli O antigen assembly is independent of the chemical structure of the O repeat. J. Biol. Chem. 274:35129-35138. [PubMed]
20. Guzman, L. M., D. Belin, M. J. Carson, and J. Beckwith. 1995. Tight regulation, modulation, and high-level expression by vectors containing the arabinose pBAD promoter. J. Bacteriol. 177:4121-4130. [PMC free article] [PubMed]
21. Helenius, J., and M. Aebi. 2002. Transmembrane movement of dolichol linked carbohydrates during N-glycoprotein biosynthesis in the endoplasmic reticulum. Semin. Cell Dev. Biol. 13:171-178. [PubMed]
22. Helenius, J., D. T. Ng, C. L. Marolda, P. Walter, M. A. Valvano, and M. Aebi. 2002. Translocation of lipid-linked oligosaccharides across the ER membrane requires Rft1 protein. Nature 415:447-450. [PubMed]
23. Jones, D. T., W. R. Taylor, and J. M. Thornton. 1994. A model recognition approach to the prediction of all-helical membrane protein structure and topology. Biochemistry 33:3038-3049. [PubMed]
24. Keenleyside, W. J., and C. Whitfield. 1996. A novel pathway for O-polysaccharide biosynthesis in Salmonella enterica serovar Borreze. J. Biol. Chem. 271:28581-28592. [PubMed]
25. Lahti, J. L., A. P. Silverman, and J. R. Cochran. 2009. Interrogating and predicting tolerated sequence diversity in protein folds: application to E. elaterium trypsin inhibitor-II cystine-knot miniprotein. PLoS Comput. Biol. 5:e1000499. [PMC free article] [PubMed]
26. Lehrer, J., K. A. Vigeant, L. D. Tatar, and M. A. Valvano. 2007. Functional characterization and membrane topology of Escherichia coli WecA, a sugar-phosphate transferase initiating the biosynthesis of enterobacterial common antigen and O antigen lipopolysaccharide. J. Bacteriol. 189:2618-2628. [PMC free article] [PubMed]
27. Lin, W.-J., and M.-J. Hwang. 1998. VHMPT: a graphical viewer and editor for helical membrane protein topologies. Bioinformatics 14:866-868. [PubMed]
28. Liu, D., and P. R. Reeves. 1994. Escherichia coli K-12 regains its O antigen. Microbiology 140:49-57. [PubMed]
29. Manoil, C. 1991. Analysis of membrane protein topology using alkaline phosphatase and beta-galactosidase gene fusions. Methods Cell. Biol. 34:61-75. [PubMed]
30. Manoil, C., and J. Beckwith. 1986. A genetic approach to analyzing membrane protein topology. Science 233:1403-1408. [PubMed]
31. Marolda, C. L., P. Lahiry, E. Vinés, S. Saldías, and M. A. Valvano. 2006. Micromethods for the characterization of lipid A-core and O-antigen lipopolysaccharide. Methods Mol. Biol. 347:237-252. [PubMed]
32. Marolda, C. L., L. D. Tatar, C. Alaimo, M. Aebi, and M. A. Valvano. 2006. Interplay of the Wzx translocase and the corresponding polymerase and chain length regulator proteins in the translocation and periplasmic assembly of lipopolysaccharide O antigen. J. Bacteriol. 188:5124-5135. [PMC free article] [PubMed]
33. Marolda, C. L., J. Vicarioli, and M. A. Valvano. 2004. Wzx proteins involved in O antigen biosynthesis function in association with the first sugar of the O-specific lipopolysaccharide subunit. Microbiology 150:4095-4105. [PubMed]
34. Marolda, C. L., J. Welsh, L. Dafoe, and M. A. Valvano. 1990. Genetic analysis of the O7-polysaccharide biosynthesis region from the Escherichia coli O7:K1 strain VW187. J. Bacteriol. 172:3590-3599. [PMC free article] [PubMed]
35. Mazur, A., M. Marczak, J. E. Król, and A. Skorupska. 2005. Topological and transcriptional analysis of pssL gene product: a putative Wzx-like exopolysaccharide translocase in Rhizobium leguminosarum bv. trifolii TA1. Arch. Microbiol. 184:1-10. [PubMed]
36. Nikaido, H. 2003. Molecular basis of bacterial outer membrane permeability revisited. Microbiol. Mol. Biol. Rev. 67:593-656. [PMC free article] [PubMed]
37. Nilsson, J., B. Persson, and G. von Heijne. 2000. Consensus predictions of membrane protein topology. FEBS Lett. 486:267-269. [PubMed]
38. Ortega, X., T. A. Hunt, S. Loutet, A. D. Vinion-Dubiel, A. Datta, B. Choudhury, J. B. Goldberg, R. Carlson, and M. A. Valvano. 2005. Reconstitution of O-specific lipopolysaccharide expression in the Burkholderia cenocepacia strain J2315 that is associated with transmissible infections in patients with cystic fibrosis. J. Bacteriol. 187:1324-1333. [PMC free article] [PubMed]
39. Patel, K. B., S. E. Furlong, and M. A. Valvano. Functional analysis of the C-terminal domain of the WbaP protein that mediates initiation of O antigen synthesis in Salmonella enterica. Glycobiology, in press. [PubMed]
40. Paulsen, I. T., A. M. Beness, and M. H. Saier. 1997. Computer-based analyses of the protein constituents of transport systems catalysing export of complex carbohydrates in bacteria. Microbiology 143(Pt. 8):2685-2699. [PubMed]
41. Pérez, J. M., M. A. McGarry, C. L. Marolda, and M. A. Valvano. 2008. Functional analysis of the large periplasmic loop of the Escherichia coli K-12 WaaL O-antigen ligase. Mol. Microbiol. 70:1424-1440. [PubMed]
42. Raetz, C. R. H., and C. Whitfield. 2002. Lipopolysaccharide endotoxins. Annu. Rev. Biochem. 71:635-700. [PMC free article] [PubMed]
43. Reiner, A. M. 1977. Xylitol and D-arabitol toxicities due to derepressed fructose, galactitol, and sorbitol phosphotransferases of Escherichia coli. J. Bacteriol. 132:166-173. [PMC free article] [PubMed]
44. Ruiz, N., D. Kahne, and T. J. Silhavy. 2006. Advances in understanding bacterial outer-membrane biogenesis. Nat. Rev. Microbiol. 4:57-66. [PubMed]
45. Ruiz, N., D. Kahne, and T. J. Silhavy. 2009. Transport of lipopolysaccharide across the cell envelope: the long road of discovery. Nat. Rev. Microbiol. 7:677-683. [PMC free article] [PubMed]
46. Saldías, M. S., K. Patel, C. L. Marolda, M. Bittner, I. Contreras, and M. A. Valvano. 2008. Distinct functional domains of the Salmonella enterica WbaP transferase that is involved in the initiation reaction for synthesis of the O antigen subunit. Microbiology 154:440-453. [PubMed]
47. Sorgen, P. L., Y. Hu, L. Guan, H. R. Kaback, and M. E. Girvin. 2002. An approach to membrane protein structure without crystals. Proc. Natl. Acad. Sci. U. S. A. 99:14037-14040. [PMC free article] [PubMed]
48. Sperandeo, P., G. Dehò, and A. Polissi. 2009. The lipopolysaccharide transport system of Gram-negative bacteria. Biochim. Biophys. Acta 1791:594-602. [PubMed]
49. Sperandeo, P., F. K. Lau, A. Carpentieri, C. De Castro, A. Molinaro, G. Dehò, T. J. Silhavy, and A. Polissi. 2008. Functional analysis of the protein machinery required for transport of lipopolysaccharide to the outer membrane of Escherichia coli. J. Bacteriol. 190:4460-4469. [PMC free article] [PubMed]
50. Stenberg, F., P. Chovanec, S. L. Maslen, C. V. Robinson, L. Ilag, G. von Heijne, and D. O. Daley. 2005. Protein complexes of the Escherichia coli cell envelope. J. Biol. Chem. 280:34409-34419. [PubMed]
51. Tatar, L. D., C. L. Marolda, A. N. Polischuk, D. van Leeuwen, and M. A. Valvano. 2007. An Escherichia coli undecaprenyl-pyrophosphate phosphatase implicated in undecaprenyl-phosphate recycling. Microbiology 153:2518-2529. [PubMed]
52. Tusnády, G. E., and I. Simon. 2001. The HMMTOP transmembrane topology prediction server. Bioinformatics 17:849-850. [PubMed]
53. Ulevitch, R. J., J. C. Mathison, and J. da Silva Correia. 2004. Innate immune responses during infection. Vaccine 22(Suppl. 1):S25-S30. [PubMed]
54. Valvano, M. A. 2003. Export of O-specific lipopolysaccharide. Front. Biosci. 8:s452-471. [PubMed]
55. Viklund, H., and A. Elofsson. 2008. OCTOPUS: improving topology prediction by two-track ANN-based preference scores and an extended topological grammar. Bioinformatics 24:1662-1668. [PubMed]
56. Vinés, E., C. L. Marolda, A. Balachandran, and M. A. Valvano. 2005. Defective O antigen polymerization in tolA and pal mutants of Escherichia coli in response to extracytoplasmic stress. J. Bacteriol. 187:3359-3368. [PMC free article] [PubMed]
57. Von Heijne, G. 2006. Membrane-protein topology. Nat. Rev. Mol. Cell. Biol. 7:909-918. [PubMed]
58. Zhou, G. P., and F. A. Troy. 2005. NMR study of the preferred membrane orientation of polyisoprenols (dolichol) and the impact of their complex with polyisoprenyl recognition sequence peptides on membrane structure. Glycobiology 15:347-359. [PubMed]
59. Zhou, G. P., and F. A. Troy II. 2005. NMR studies on how the binding complex of polyisoprenol recognition sequence peptides and polyisoprenols can modulate membrane structure. Curr. Protein Pept. Sci. 6:399-411. [PubMed]
60. Zhou, Z., K. A. White, A. Polissi, C. Georgopoulos, and C. R. Raetz. 1998. Function of Escherichia coli MsbA, an essential ABC family transporter, in lipid A and phospholipid biosynthesis. J. Biol. Chem. 273:12466-12475. [PubMed]

Articles from Journal of Bacteriology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...