Polysaccharide Copolymerase WzzB/WzzE Chimeras Reveal that the Transmembrane 2 Region of WzzB Is Important for Interaction with WzyB

Bacteria synthesize complex polysaccharide chains at a controlled number of repeating units; this has wide implications for a range of bacterial activities involved in virulence. Examples of complex polysaccharide chains include the Oag component of lipopolysaccharide and the ECA; both of these examples are predominantly synthesized by their own independent Wzy-dependent pathway.

ABSTRACT

The ability of bacteria to synthesize complex polysaccharide chains at a controlled number of repeating units has wide implications for a range of biological activities that include symbiosis, biofilm formation, and immune system avoidance. Complex polysaccharide chains, such as the O antigen (Oag) component of lipopolysaccharide and the enterobacterial common antigen (ECA), are synthesized by the most common polysaccharide synthesis pathway used in bacteria, known as the Wzy-dependent pathway. The Oag and ECA are polymerized into chains via the inner membrane proteins WzyB and WzyE, respectively, while the respective copolymerases WzzB and WzzE modulate the number of repeat units in the chains, or the “modal length” of the polysaccharide, via a hypothesized interaction. Our data, show for the first time, “cross talk” between Oag and ECA synthesis in that WzzE is able to partially regulate Oag modal length via a potential interaction with WzyB. To investigate this, one or both of the transmembrane regions (TM1 and TM2) of WzzE and WzzB were swapped, creating six chimera proteins. Several chimeric proteins showed significant increases in Oag modal length control, while others reduced control. Additionally, copurification experiments show an interaction between WzyB and WzzB for the first time without the use of a chemical cross-linker, and a novel interaction between WzyB and WzzE. These results suggest the TM2 region of Wzz proteins plays a critical role in Oag and ECA modal length control, presumably via the interaction with respective Wzy proteins, thus providing insight into the complex mechanism underlying the control of polysaccharide biosynthesis.

IMPORTANCE Bacteria synthesize complex polysaccharide chains at a controlled number of repeating units; this has wide implications for a range of bacterial activities involved in virulence. Examples of complex polysaccharide chains include the Oag component of lipopolysaccharide and the ECA; both of these examples are predominantly synthesized by their own independent Wzy-dependent pathway. Our data show, for the first time, “cross talk” between Oag and ECA synthesis and identify novel physical protein-protein interactions between proteins in these systems. These findings further the understanding of how the system functions to control polysaccharide chain length, which has great implications for novel biotechnologies and/or the combat of bacterial diseases.

INTRODUCTION

Shigella flexneri bacteria have the ability to synthesize many complex polysaccharide chains that are critical for virulence and structural stability. Complex polysaccharide chains such as the O antigen (Oag) element of the lipopolysaccharide (LPS), and the unrelated enterobacterial common antigen (ECA), are synthesized by variants of the Wzy-dependent pathway (1, 2).

S. flexneri LPS is comprised of three domains, including (i) lipid A, a hydrophobic lipid that anchors the molecule to the outer membrane; (ii) core oligosaccharides, a nonrepeating domain; and (iii) Oag, an oligosaccharide repeat unit (RU) that varies in number of repeats per LPS molecule. The Oag repeat units are attached to the lipid A component via the core sugars (3). In S. flexneri Y serotype strains, the Oag is comprised of tetrasaccharide RUs which contain N-acetylglucosamine (GlcNAc) and rhamnose (Rha) (4). ECA is a polysaccharide repeat unit that is comprised of three different sugars and is conserved among Enterobacterales. There are three forms of ECA, two of which are lipid anchored to the outer membrane (ECA_PG and ECA_LPS), while the third form is located in the periplasm (ECA_cyc) (2, 5–7).

The Wzy-dependent polysaccharide synthesis mechanism is the most commonly used system in bacteria to produce Oag, which takes place in the inner membrane and requires a set of integral membrane proteins (8). Oag is synthesized by the Oag polymerase WzyB. The individual RUs bound to the lipid carrier undecaprenyl diphosphate (Und-PP) are transported across the inner membrane from the cytoplasm into the periplasm by the Wzx flippase/transporter. The WaaL protein ligates the Oag to the previously synthesized core plus lipid A molecules. Oag chain length is regulated by the copolymerase protein WzzB. LPS lacking the Oag polysaccharide is termed rough LPS, LPS structures without Oag and outer core sugars are termed deep-rough LPS, and the completed structure with all components is termed smooth LPS. In S. flexneri, the ECA polysaccharide is also synthesized by a unique homologue set of the aforementioned integral membrane proteins named WzyE and WzzE, which function similarly in the ECA biosynthetic pathway.

Despite there being several studies on S. flexneri WzyB and its paralogues, there is little known about how WzyB functions in combination with the copolymerase WzzB to polymerize and control Oag length. The S. flexneri WzyB protein is approximately 43.7 kDa with 12 transmembrane (TM) domains, 6 periplasmic loops, and 5 cytoplasmic loops (9). The proposed “catch-and-release” model suggests that the Und-PP-Oag repeat unit (RU) binds a site on the third periplasmic loop and is then transferred to a second site on the fifth periplasmic loop, mediated by the loop differences in pI, basic versus acidic, respectively (10). Conversely in S. flexneri, the pI of the loops appear to play less of a role as both have a similar basic pI (11).

In S. flexneri, WzzB confers the short-Oag (S-Oag) (∼10 to 17 RUs) modal length and is a 36-kDa inner membrane protein that is anchored by two TM domains which are located near the N and C termini of the protein; however, the majority of the protein resides in the periplasmic space. WzzB is a member of the polysaccharide copolymerase group 1 (PCP1) family of proteins (12). Despite low sequence identity among members of the PCP1 family, the monomeric structures of the proteins are remarkably similar. Structural studies on PCP1 proteins show that each form a distinct “bell”-shaped structure with different numbers of monomers. The number of monomers that make up each protein is controversial, as formation of both hexamers and octamers has been reported (13, 14). Additionally, in situ WzzB has been shown to exist in multiple oligomeric forms at equilibrium with the monomeric form (15, 16). The full-length oligomeric structure of WzzB^ST from Salmonella enterica serovar Typhimurium was resolved using cryo-electron microscopy experiments, revealing a potential binding region for Wzy (17). Recently, a biochemical interaction between WzzB and WzyB has been shown through the use of in vivo chemical cross-linking using dithiobis (succinimidyl propionate) (DSP) (18). This interaction is thought to be required for modal length control, and although this interaction occurred between WzyB and WzzB, it is hypothesized that WzyE and WzzE also have a similar interaction.

Several studies have revealed genetic interactions between the Oag and ECA pathways. The Oag and ECA biosynthesis pathways share the same initial glycosyl transferase (WecA) which adds GlcNAc to Und-PP; thus, disruptions to WecA result in loss of both ECA and Oag (5). It has also been previously reported that certain mutations in ECA genes (e.g., wzzE) had an effect on the Oag of Escherichia coli strain O25 (19). However, no other effect of ECA gene mutations on Oag has been reported.

In light of this and the fact that despite having a low sequence homology, WzzB and WzzE monomeric three-dimensional structures are remarkably similar (20), we investigated whether WzzE could function in the Oag system and, vice versa, if WzzB could function in the ECA system.

In this study, we found that WzzE partially controlled Oag modal length. We then created chimera proteins of the unrelated PCP proteins WzzB and WzzE by swapping the TM regions. The WzzE chimera with TM2 of WzzB showed a significant increase in activity compared to WT WzzE, which was determined via LPS analysis by silver-stained SDS-PAGE gels and colicin E2 assays. Pulldown experiments involving the purification of FLAG-tagged WzyB revealed an interaction between WzyB and WzzB as well as an interaction between WzyB and WzzE, without the use of a chemical cross-linker.

RESULTS

WzzE is able to partially complement a wzzB mutant.

Due to the previously reported links between the biosynthetic pathways of Oag and ECA, as well as the monomeric three-dimensional structural similarities between WzzE and WzzB, we investigated if the ECA copolymerase, WzzE, was able to function in the modal length control of Oag. The LPS of parent RMA2162 (S. flexneri Y Δwzz^pHS2) and RMA4662 (S. flexneri Y ΔwzzE-Δwzz^pHS2-wzzB::Kan^r), transformed with pQE30, pQE30-wzzE-His₆ or pQE30-wzzB-His₆, was analyzed via SDS-PAGE and silver staining (Fig. 1). As expected when His₆-WzzB (referred to here as WzzB) was expressed, the LPS profile was comparable to that of the parent strain with a pronounced S-Oag modal length (10 to 17 Oag RUs) (Fig. 1, lanes 5 and 1, respectively). Interestingly, when His₆-WzzE (referred to here as WzzE) was expressed, there was a distinct increase in the level of S-Oag compared to the mutant and vector-only controls (Fig. 1, lanes 4, 2, and 3, respectively). These results suggested that WzzE was able to partially act on the Oag Wzy-dependent system by regulating the length of the Oag chains of LPS molecules.

FIG 1.

Analysis of LPS profile conferred by triple copolymerase mutant in the presence of either WzzB or WzzE. S. flexneri Y strains were grown, and the LPS was isolated and detected from whole-cell lysates of parent S. flexneri Y Δwzz^pHS2 (lane 1) or S. flexneri Y ΔwzzE-Δwzz^pHS2-wzzB::Kan^r (lane 2) harboring pQE30 control (lane 3) or expressing WzzE protein (lane 4) or WzzB protein (lane 5). Strains grown to log phase were collected (1 × 10⁹ cells) and lysed in lysis buffer in the presence of proteinase K. Samples were then electrophoresed on an SDS-15% (wt/vol) PAGE gel and silver stained (see Materials and Methods). The numbers of Oag RUs are shown on the left-hand side.

Functional analysis of the transmembrane regions within the copolymerases.

To further investigate which regions of the copolymerases may be interacting with WzyB, the TM regions of the WzzE and WzzB were selected for mutagenesis on the basis that the majority of WzyB exists in the membrane and the conserved GGXXG motifs of the Wzz proteins are present in TM region 2. Six unique copolymerase chimeric proteins (WzzB^ETM1, WzzB^ETM2, WzzB^ETM1&2, WzzE^BTM1, WzzE^BTM2, and WzzE^BTM1&2) were designed to exchange either TM region 1, TM region 2, or both between WzzB and WzzE; these were generated via inverse PCR and in vitro oligonucleotide annealing (as described in Materials and Methods) (Fig. 2 and Table 1; see also Fig. S1 and S2 in the supplemental material). The aligned amino acid sequences of the TM regions of WzzB and WzzE are shown in Fig. 2c. The LPS profile of S. flexneri Y ΔwzzE-Δwzz^pHS2-wzzB::Kan^r, transformed with pQE30 vectors expressing each of these chimera proteins (pVLRM16-21) (Table 1), was analyzed via silver-stained SDS-PAGE (Fig. 3a). The presence of WzzE and, additionally, the presence of WzzE^BTM1 increased the level of S-Oag compared to the empty vector and controls (Fig. 3, lanes 4, 5, 3, and 2, respectively). Strains with WzzE^BTM2 and WzzE^BTM1&2 had LPS with a further increase of modal length control compared to the vector control (Fig. 3, lanes 6, 7, and 3, respectively), suggesting the presence of the WzzB TM region 2 is important for modal length control and potential interaction with WzyB. Reciprocally, the addition of WzzE’s TM2 region to WzzB (i.e., WzzB^ETM2) reduced the modal length control of Oag when comparing to the parent strain and wild-type WzzB, further supporting the importance of WzzB’s TM2 region for function/interaction (Fig. 3, lanes 10, 1, and 8, respectively). The replacement of WzzB TM1 (i.e., WzzB^ETM1) with WzzE TM1 had little effect on the LPS profile (Fig. 3, lane 9).

FIG 2.

Graphical representation of the polysaccharide copolymerase chimera proteins. Schematic representing 6 chimeric polysaccharide copolymerase proteins generated via inverse PCR cloning. (a) Topology representation. (b) Graphical representation of protein maps compositions (IM, inner membrane; TM, transmembrane). A yellow box indicates a His₆ tag. (c) Amino acid sequence alignment of WzzB and WzzE transmembrane regions 1 and 2. An asterisk indicates a fully conserved residue, and a period indicates conservation between groups of weakly similar properties.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Description	Source
Strains
RMA2162	S. flexneri PE860 Y serotype; this strain lacks the virulence plasmid and pHS-2 (parent)	Laboratory stock
RMA4662	S. flexneri PE860 Y serotype wzzB::Kanr, ΔwzzE	Laboratory stock
Plasmids
pQE30	IPTG inducible, expression vector; Ampr	Qiagen
pQE30-His6-wzzB	pQE30 with Shigella flexneri wzzB gene; Ampr	GenScript
pQE30-His6-wzzE	pQE30 with Shigella flexneri wzzE gene; Ampr	GenScript
pVLRM18	pQE30 encoding WzzE(WzzBTM1)-His6; Ampr	This study
pVLRM16	pQE30 encoding WzzE(WzzBTM2)-His6; Ampr	This study
pVLRM20	pQE30 encoding WzzE(WzzBTM1 & 2)-His6; Ampr	This study
pVLRM19	pQE30 encoding WzzB(WzzETM1)-His6; Ampr	This study
pVLRM17	pQE30 encoding WzzB(WzzETM2)-His6; Ampr	This study
pVLRM21	pQE30 encoding WzzB(WzzETM1 & 2)-His6; Ampr	This study
pBAD33	Arabinose inducible, expression vector; Cmlr
pVLRM11	pBAD33 encoding WzyB-Flag	This study

FIG 3.

Analysis of the LPS profiles of strains expressing chimeric Wzz proteins. (a) Silver-stained PAGE separation of the LPS from parent S. flexneri Y Δwzz^pHS2 (lane 1) and ΔwzzE-wzzB::Kan^r (lane 2), harboring pQE30 (lane 3), encoding either WT Wzz proteins (lanes 4 or 8) or a chimeric Wzz (lanes 5 to 7 or 9 to 11). Strains grown to log phase were collected (1 × 10⁹ cells) and lysed in lysis buffer in the presence of proteinase K. Samples were electrophoresed on a SDS-15% (wt/vol) PAGE gel and silver stained. (b) Whole-cell lysates (5 × 10⁸ cells) of indicated strains were separated by SDS-PAGE, transferred onto a nitrocellulose membrane, and probed with either polyclonal rabbit WzzE_SF or WzzB_SF antibodies. (c) Analysis of polymerization of S-Oag via densitometry (Image Lab). The degree of polymerization of S-Oag is represented as the densitometry of Oag RUs 10 to 17 (indicated by bracket) as a percentage relative to the parent. Data represent 3 independent experiments with SEM shown, and significance is calculated with a one-way analysis of variance (ANOVA). *, P < 0.033; **, P < 0.0021; ***, P < 0.0002; ****, P < 0.0001.

We then investigated the expression of the chimeras. Western immunoblotting with either anti-WzzE_SF or anti-WzzB_SF (21) detected bands consistent with the size of WzzE (∼40 kDa) and of WzzB (∼38 kDa), respectively, for all strains expressing each of the WT and chimera copolymerase proteins (Fig. 3b).

To quantify the relative degree of polymerization of S-Oag produced by strains expressing the chimera proteins, densitometry was performed on three biological replicates (Fig. 3b and Fig. S3) of silver-stained SDS-PAGE gels (Fig. 3c). The degree of polymerization is presented as the densitometry of S-Oag LPS molecules normalized to the parent as a percentage. This analysis provides insight into a particular strain’s ability to regulate Oag length, whereby strains without the presence of WzzB have a lower proportion of overall LPS with 10 to 17 RUs. In the presence of either WzzE or WzzE^TM1, the relative degree of polymerization of S-Oag appeared higher than that of the vector control. Additionally, in the presence of either WzzE^BTM2 or WzzE^BTM1&2, the relative degree of polymerization of S-Oag was significantly higher than that of both the vector control and WzzE. Furthermore, LPS from the strain expressing WzzB^ETM2 had a significant decrease in S-Oag compared to the WzzB-expressing strain. Collectively, these results further support the importance of WzzB TM2 in Oag modal length control and presumably the interaction with WzyB.

The effect of the chimeric proteins on resistance to colicin E2.

A colicin E2 spot sensitivity assay was used as an indirect measure to determine the S-Oag production of the strains expressing the chimera proteins (22). Log-phase cultures of S. flexneri Y ΔwzzE-Δwzz^pHS2-wzzB::Kan^r transformed with pQE30 vectors encoding each of WT or chimeric Wzz proteins (pVLRM16-21) were tested against decreasing concentrations of purified His₆-colicin E2 (Fig. 4). As anticipated, when WzzE was expressed, the strain was more resistant to colicin E2 than the vector control. When WzzE^BTM2 was expressed, the strain was significantly more resistant to colicin E2 than both the vector control and WzzE. When WzzE^BTM1&2 was expressed, the strain was significantly more resistant than the vector control (Fig. 4a). Furthermore, when WzzB^ETM1 was expressed, the strain was slightly less resistant to colicin E2 than the WzzB control strain; however, this was nonsignificant. When either WzzB^ETM2 or WzzB^ETM1&2 was expressed, the strains showed a significant reduction in colicin E2 resistance compared to the WzzB control strain. Combined, these results suggested that the TM1 region of Wzz proteins was not likely to be important for function and/or interaction with WzyB, while TM2 was likely critical for function and/or interaction with WzyB.

FIG 4.

Colicin E2 sensitivity of strains producing Wzz chimeric proteins. S. flexneri Y strains expressing the indicated proteins from pQE30 are indicated on the x axis. The MIC of colicin E2 required to generate a clear zone of bacterial inhibition for each strain was normalized (shown as percentage) to the parent strain MIC. (a) WzzE/WzzE chimeric protein-expressing strains. (b) WzzB/WzzB chimeric protein-expressing strains. Data represent 4 independent experiments with SEM shown, and significance is calculated with a one-way ANOVA. *, P < 0.033; **, P < 0.0021; ***, P < 0.0002; ****, P < 0.0001.

The effect of the chimeric proteins on ECA production and modal length control.

We then investigated the effect of the Wzz chimeric proteins’ ability to produce ECA, by Western immunoblotting. ECA modal length is controlled by WzzE (23), and, as such, when S. flexneri Y ECA was analyzed, banding could be seen between approximately 16 and 23 kDa (Fig. 5, lane 1), while the mutant S. flexneri Y ΔwzzE-Δwzz^pHS2-wzzB::Kan^r and vector control strains produced unregulated ECA chains ranging from 16 to 40 kDa (Fig. 5, lane 2). When WT WzzE was expressed, the modal length of ECA was controlled; however, it was shorter than that of the parent strain (Fig. 5, lanes 4 and 1, respectively), suggesting that the absence of WzzB or Wzz^pHS2 may be altering ECA production and/or modal length control. Similarly, when WzzE^BTM1 was expressed, the ECA modal length was controlled; however, again, it was shorter than the parent strain (Fig. 5, lanes 5 and 1, respectively). Despite the relatively shorter modal length, WzzE^BTM1 was still able to control the ECA modal length, suggesting that the presence of WzzB TM1 was equivalent to WzzE TM1, as no negative effect on ECA production or control was detected. When either WzzE^BTM2 or WzzE^BTM1&2 was expressed, the ECA banding size was less controlled and comparable to the mutant and vector control (Fig. 5, lanes 6, 7, 2, and 3, respectively). Protein expression of these samples can be seen in Fig. 3b, as the same samples were used in LPS and ECA analysis. These results suggest that the WzzE TM2 region was likely to be critical for function and/or interaction with WzyE. Strains expressing WzzB and all the WzzB chimeras did not have any effect on ECA modal length control (Fig. 5, lanes 8 to 10).

FIG 5.

Functional analysis of the transmembrane regions within the copolymerases via Western immunoblotting. Silver-stained PAGE separation and Western immunoblot of ECA from parent S. flexneri Y Δwzz^pHS2 (lane 1) and ΔwzzE-wzzB::Kan^r (lane 2), harboring pQE30 (lane 3) encoding either WT Wzz proteins (lanes 4 or 8) or a chimeric Wzz protein (lanes 5 to 7 or 9 to 11). Whole-cell lysates (1 × 10⁹ cells) of the indicated strains were treated with proteinase K, separated by SDS-PAGE, transferred onto a nitrocellulose membrane, and probed with polyclonal rabbit ECA antibodies. The migration positions of the protein molecular mass standards are indicated on the left-hand side.

Interaction of the Wzz chimera proteins with WzyB.

To investigate the Wzz chimeric proteins’ ability to physically interact with WzyB, a new WzyB expression construct was made. The S. flexneri wzyB coding region was PCR amplified from pRMPN1 (24) to include a FLAG tag and cloned into pBAD33, creating a WzyB-FLAG expression construct which possesses a strong arabinose promoter for protein expression, named pVLRM11 (Table 1). S. flexneri Y ΔwzzE-Δwzz^pHS2-wzzB::Kan^r strains expressing the WT and chimeric Wzz proteins from pQE30 plasmids were also transformed with pVLRM11 (pBAD33-wzyB-FLAG) and the vector controls. WzzB wild-type (WT)- and WzzE WT-expressing strains, coexpressing WzyB, were grown, and whole cells were incubated in the presence or absence of DSP, followed by copurification experiments using anti-DYKDDDDK G (FLAG) affinity resin (see Materials and Methods). The purified samples were separated on an SDS-12% (wt/vol) polyacrylamide gel and subjected to Western immunoblotting using either anti-FLAG, anti-WzzB_SF, or anti-WzzE_SF antibodies (Fig. 6). WzyB was purified at comparable levels in all the separate copurifications in the presence of the different Wzz proteins, and 3 bands at ∼47 to ∼148 kDa were detected (Fig. 6a). As expected when probing with anti-WzzB, WzzB was detected in both the presence and absence of DSP (Fig. 6b). Similarly, WzzE was detected via anti-WzzE in both the presence of and absence of DSP (Fig. 6c). As expected, no bands were detected when probing with either WzzB or WzzE when either WzyB or Wzz protein was absent from the copurification experiment. Previously, copurification experiments of WzzB and WzyB without the presence of DSP were unable to detect copurification of WzzB (18).

FIG 6.

Copurification of WzyB and Wzz/Wzz chimeric proteins using anti-FLAG resin. Purification of WzyB-FLAG from pVLRM11 (Table 1) from S. flexneri Y ΔwzzE-wzzB::Kan^r expressing either WzzB or WzzE proteins. Whole-cell treatment with (+) and without (−) 1 mM DSP (as described in Materials and Methods). Samples were electrophoresed on SDS-12% (wt/vol) polyacrylamide gels followed by Western immunoblotting with polyclonal FLAG antibody (a), polyclonal WzzB_SF antibody (b), or polyclonal WzzE_SF antibody (c). The migration positions of the molecular mass standards are indicated on the left-hand side.

Next, we performed copurification experiments in strains expressing each of the Wzz chimeric proteins followed by Western immunoblotting using either anti-FLAG or anti-His antibodies. Anti-His antibodies were used so WzzB and WzzE chimeric proteins could be detected equally. WzyB was purified at comparable levels in all the separate copurifications in the presence of the different Wzz chimeric proteins, as bands ranging from ∼47 to ∼148 kDa were detected (Fig. 7a). All chimeric proteins were detected using anti-His in both the presence and absence of DSP treatment. Copurifications performed with no WzyB expressed (vector control) produced no WzyB bands in both anti-FLAG and anti-His Western immunoblots, as expected (two nonspecific bands can be seen at ∼110 and ∼90 kDa) (Fig. 7). Furthermore, copurifications performed in the presence of WzyB and pQE30 vector control did not detect any bands via anti-WzzE and anti-WzzB Western immunoblotting, as expected (Fig. 7). Most notably, copurifications in the presence of WzzE^BTM2 and absence of DSP show markedly more intense bands than all other the other WzzE chimeric proteins (Fig. 7b). Copurification experiments in strains expressing WzzE and all WzzE chimeric proteins in the absence of DSP were further subjected to Western immunoblotting using anti-WzzE antibodies for direct comparison of relative interaction with WzyB (Fig. 8). The strain expressing WzzE^TM2 produced significantly darker banding than the WT WzzE and all other WzzE chimeric proteins, suggesting that the TM2 region of WzzB increased the interaction between WzzE and WzyB. Solubilized membrane fractions prior to purification were also analyzed by Western immunoblotting and confirmed that WzyB and all Wzz proteins were expressed at comparable levels before copurification (Fig. S4).

FIG 7.

Copurification of WzyB and Wzz chimeric proteins using anti-FLAG resin. Purification of WzyB-FLAG expressed by pVLRM11 (Table 1) from S. flexneri Y ΔwzzE-wzzB::Kan^r expressing WT or chimeric WzzE or WzzB proteins. Whole-cell treatment with and without 1 mM DSP indicated with plus or minus signs, respectively (as described in Materials and Methods). Samples were electrophoresed on SDS-12% (wt/vol) polyacrylamide gels followed by Western immunoblotting with polyclonal FLAG antibody (a) or monoclonal His-Tag antibody (b). The migration positions of the molecular mass standards are indicated on the left-hand side.

FIG 8.

Comparison of copurification of WzyB and WzzE chimeric proteins using anti-FLAG resin comparison using anti-WzzE antibodies. Purification of WzyB-FLAG expressed by pVLRM11 (Table 1) from S. flexneri Y ΔwzzE-wzzB::Kan^r expressing WT or chimeric WzzE proteins (as described in Materials and Methods). Samples were electrophoresed on SDS-12% (wt/vol) polyacrylamide gels followed by Western immunoblotting with polyclonal FLAG antibody (a) or polyclonal WzzE antibody (b). The migration positions of the molecular mass standards are indicated on the left-hand side.

DISCUSSION

In this study, we characterized a series of WzzB/WzzE TM chimeric proteins, and analysis of their activity indicated that the TM2 region of Wzz proteins is likely to be involved in a direct interaction with their system’s Wzy partner. Previous mutagenesis studies suggested that the WzzB_SF TM2 region, particularly the GGXXG motif, was important for function, as a G306A/G311A double mutant conferred a very short-chain Oag (21, 25). However, this mutation altered modal length but did not decrease modal length control. Additionally, structural studies have also pointed toward the TM region being the likely region in which WzyB and WzzB interact (17). In light of this, the TM regions of the two Wzz proteins were targeted to be swapped to investigate their roles in function and interaction with WzyB.

The three-dimensional structures of WzzB and WzzE are remarkably similar despite only sharing approximately 22% overall amino acid homology. The conserved GGXXG motif is present in both and is located in the TM2 region. The chimeras were designed to substitute both the TM regions of the Wzz proteins, and particular care was taken with the TM2 region (23 amino acids [aa]) to ensure the GGXXG motifs were in the same position within the TM2 region of the chimera proteins. In fact, the TM1s (23 aa) between WzzB and WzzE share approximately 30% amino acid identity, and the TM2s share approximately 42% identity; thus, these regions both share a higher average homology than the overall homology of the full-length protein. In the case of TM2, only 14 amino acids were substituted (out of 23), which did not include the conserved GGXXG motif. These surrounding amino acids were able to increase the WzzE proteins function in Oag modal length control, which suggests that the nonmotif amino acids are either directly involved with protein interaction with WzyB, or the amino acids aid in correctly positioning the important GGXXG motif for full activity and/or interaction.

The results of this study can be found summarized in Table 2. Interestingly, WzzE alone was able to partially complement a wzzB mutant, as it partially controlled Oag modal length, as seen by SDS-PAGE and silver staining, and also by an increase in the strains’ resistance to colicin E2 (Fig. 1 and 4). The addition of the TM2 region of WzzB into WzzE produced the most interesting result, as the presence of the WzzB TM2 region was able to significantly increase the modal length control of Oag while also significantly increasing colicin E2 resistance compared to the WT WzzE protein (Fig. 3a, lanes 6 and 4, respectively, and Fig. 4).

TABLE 2.

Summary of results

Protein	LPS profilea	S-Oag densitometry (mean) normalized to parent (%)b	Colicin E2 MIC (mean) normalized to parent (%)c	ECA profiled	WzyB interaction with or without DSPe
WzzB	Regulated Oag	102	97	Unregulated ECA, darker banding	+
WzzE	Partially regulated Oag	69	18	Regulated ECA, lower modal length	++
WzzEBTM1	Partially regulated Oag	71	6	Regulated ECA, lower modal length	+
WzzEBTM2	Moderately regulated Oag	87	35	Unregulated ECA	++++
WzzEETM1&2	Partially regulated Oag	82	27	Unregulated ECA	++
WzzBETM1	Regulated Oag	102	78	Unregulated ECA, darker banding	+
WzzBETM2	Partially regulated Oag	85	17	Unregulated ECA, darker banding	+
WzzBETM1&2	Moderately regulated Oag	90	18	Unregulated ECA	+
pQE30 control	Unregulated Oag	30	5	Unregulated ECA	−

^aBased on Fig. 3a.

^bBased on Fig. 3b.

^cBased on Fig. 4.

^dBased on Fig. 5.

^eBased on Fig. 7 and 8 (WzzE). + indicates interaction with WzyB (the number of plus signs indicates the relative level of interaction with WzyB); − indicates no interaction with WzyB detected.

Strains expressing the WzzB proteins with the WzzE TM2 region produced LPS with decreased levels of S-Oag compared to the WT WzzB-expressing strain, resulting in a 50% increase in colicin E2 sensitivity (Fig. 3a, lanes 10 and 8, respectively, and Fig. 4). These strains were still able to partially control Oag modal length compared to the mutant and the empty vector strains, highlighting that these proteins were still functional. These data again suggest that the TM2 region of WzzB is critical for Oag modal length control.

The WzzE TM2 region was shown to be important in ECA modal length control. When WzzB TM2 was present in the WzzE protein (WzzE^BTM2 or WzzE^BTM1&2), the strains expressing these proteins lost the ability to control the modal length of ECA. Since the three-dimensional structures of WzzB and WzzE are highly conserved (20), it is likely that the TM2 region of WzzE may interact with WzyE (ECA polymerase). Although not investigated, we speculate that WzzE^BTM2 has decreased interaction with WzyE. WzzB chimeras with WzzE TM regions were not able to restore modal length control of ECA, suggesting that the TM2 region may not be the sole interacting region in the case of ECA. This was expected, as the WT WzzB protein was unable to partially control the modal length of ECA, unlike WT WzzE, which was able to partially regulate Oag modal length. Interestingly, when either WzzE or WzzE^BTM1 was used in complementation, the ECA modal length was controlled, but this was shorter than the parent strain (Fig. 5). This is likely a result of increased wzzE copy numbers in this strain compared to the single copy in the parent strain. Furthermore, when WzzB or its chimeric variants were expressed, there was increased ECA banding intensity as detected in the anti-ECA Western immunoblot, suggesting increased ECA production was occurring, which may be tied with the Oag system. We are unable to explain this phenomenon.

Many groups have hypothesized a potential interaction between WzyB and WzzB. However, our group was able to show through the use of the cross-linker DSP, followed by mass spectrometry, that, indeed, an interaction between these proteins does occur (18). We hypothesized that this interaction is required for function, and thus, further investigation into this interaction is essential. In this study, not only were we able to recreate this interaction via the use of copurification and Western immunoblotting but we did so in the absence of the cross-linker DSP, proving a native interaction. We feel the new WzyB-FLAG construct which lacks green fluorescent protein (GFP) (18) contributed to detecting this interaction. Additionally, WzzE was shown, for the first time, to be able to interact with WzyB; however, this may be detectable as a result of the absence of other PCP1 proteins in the strains used. The results show that the addition of the WzzB TM2 region not only increased the activity of the protein in vivo but was shown to increase WzzE interaction with WzyB greater than the baseline levels of WzzE interaction. We speculate that Wzz proteins may bind to WzyB in a competitive nature in that WzzB binds more strongly than Wzz^pHS2, which binds more strongly than WzzE. Previous studies suggest that decreasing the level of WzyB removed the presence of very long Oag (90 RUs) (controlled by Wzz^pHS2), suggesting WzzB is able to outcompete Wzz^pHS2 for WzyB interaction (26).

In conclusion, our findings show that the TM2 regions of WzzB and WzzE are important for function in regulating their respective polysaccharide chains and that WzzE is able to partially act on the Oag pathway by controlling modal length. Furthermore, we have shown that WzzE with only the WzzB’s TM2 region is able to increase Oag modal length control significantly compared to the WT WzzE. We have also shown a direct interaction between WzzB without the use of a cross-linker and a novel interaction of WzzE with WzyB detected in the absence of other PCP proteins, and this interaction was increased when the TM2 of WzzB was substituted into the protein. These findings pave the way for future studies that focus on the TM2 region in order to fully elucidate which amino acids are involved in binding to Wzy proteins.

MATERIALS AND METHODS

Ethics statement.

The WzzB_SF and WzzE_SF antibodies were produced under the National Health and Medical Research Council Australian Code of Practice for the Care and Use of Animals for Scientific Purposes and was approved by the University of Adelaide Animal Ethics Committee.

Bacterial strains, growth media, and growth conditions.

The bacterial strains used in this study are shown in Table 1. Strains were routinely grown on lysogeny broth (LB) agar (10 g tryptone liter⁻¹, 5 g yeast extract liter⁻¹, 5 g NaCl liter⁻¹, and 15 g agar liter⁻¹) or in LB. Strains carrying pQE30 constructs requiring induction were grown in LB at 37°C with aeration for 16 h, subcultured 1/20 into fresh broth with 0.01 mM isopropyl-β-d-thiogalactopyranoside (IPTG), and grown for another 4 h. Strains carrying pBAD33 constructs requiring induction were grown in LB at 37°C with aeration for 16 h, subcultured 1/20 into fresh broth with 0.2% (wt/vol) arabinose, and grown for another 4 h. Antibiotics were used at the following concentrations: 100 mg ampicillin (Amp) ml⁻¹, 25 mg chloramphenicol (Cml) ml⁻¹, and 50 mg kanamycin (Kan) ml⁻¹.

Chimera construction by inverse PCR.

The oligonucleotides used to amplify pQE30-wzzB-His₆ without the TM1 region were 5′-ACGCCACAACTGCACTAG-3′ and 5′-AAGGAGAAATGGACGTCAACA-3′ (see Table S1 in the supplemental material). The complementary forward and reverse strands of DNA encoding the TM1 region of wzzE were synthesized (Integrated DNA Technologies) (Table S1) and then annealed in vitro. First, 0.02 nM DNA was resuspended in annealing buffer (100 mM potassium acetate, 30 mM HEPES, and 2 mM magnesium acetate) heated to 95°C for 4 min followed by 70°C for 10 min and finally followed by slowly cooling the DNA down to 4°C. The annealed oligonucleotide’s 5′ ends were then phosphorylated using T4 polynucleotide kinase (NEB) before being mixed at a 3:1 ratio with previously amplified plasmid. The DNA mix was then ligated using T4 ligase (NEB) and transformed into competent E. coli DH5α strains. This process was repeated to construct all chimeric wzz genes in pQE30 using the oligonucleotides listed in Table S1 and were all sequencing confirmed via Sanger sequencing (Australian Genome Research Facility).

DNA methods.

The plasmids used in this study are shown in Table 1. Plasmid constructs were extracted from E. coli DH5α strains using a QIAprep Spin miniprep kit (Qiagen). Preparation of electro-competent cells and the electroporation method were performed as described previously (27).

LPS SDS-PAGE and silver staining.

Bacteria were grown and induced as described above before 1 × 10⁹ cells were harvested by centrifugation, resuspended in 2× lysis buffer (28), and heated at 100°C for 10 min prior to incubation with 2.5 mg/ml proteinase K (Sigma-Aldrich) for 2 h at 56°C followed by electrophoresis on 15% SDS-PAGE gels for 13 h at 12 mA. Silver staining on LPS was performed as described previously (28).

Protein SDS-PAGE.

Bacteria were grown and induced as described above before 5 × 10⁸ cells were harvested by centrifugation, resuspended in 2× sample buffer (29), and heated at 100°C or 37°C for 5 min prior to SDS-PAGE on 12% gels for 1 h at 200 V. Purified proteins were mixed 1:1 with 2× sample buffer (29) with or without β-mercaptoethanol and heated to 37°C for 5 min prior to SDS-PAGE on 15% gels for 1 h at 200 V. ECA samples were prepared as described above (“LPS SDS-PAGE and silver staining”), followed by SDS-PAGE on 12% gels for 1 h at 200 V.

Western immunoblotting.

Western transfers were performed at 400 mA for 1 h. Protein gels were then subjected to Western immunoblotting on nitrocellulose membrane (Bio-Rad) with either polyclonal WzzB_SF or WzzE_SF rabbit antibodies (21) at 1:500 dilution, polyclonal ECA rabbit antibodies (made in-house) at 1:500 dilution, monoclonal His₆ mouse antibodies (GenScript) at 1:50,000 dilution, or polyclonal FLAG rabbit antibodies (Sigma) at 1:2,000 dilution in 2.5% (wt/vol) skim milk. Detection was performed with either goat anti-rabbit horseradish peroxidase-conjugated antibodies (KPL) or goat anti-mouse horseradish peroxidase-conjugated antibodies (KPL) and chemiluminescence reagent (Sigma). We used 5 μl of BenchMark SeeBlue Plus2 prestained protein ladder (Invitrogen) as the molecular mass standard.

Colicin spot assay.

For spot sensitivity assays, bacteria (∼5 × 10⁸ cells) were spread onto selective 25 ml LB agar plates, and plates were spotted with 5 μl purified colicin E2 protein (22) in Milli-Q (MQ) water at the following concentrations: 0.50, 1.0, 2.0, 4.0, 8.0, 12, 16, 20, 24, 28, 32, and 36 μg ml⁻¹. Plates were incubated at 37°C for 16 h, and the MIC, defined here as the lowest concentration that generated a clear zone of inhibition, was recorded.

Purification of WzyB from S. flexneri strains.

Cells were harvested from 100 ml induced culture by centrifugation (Beckman Coulter Avanti J-26 XPI centrifuge; 9,600 × g, 10 min, 4°C) and the cell pellet was resuspended in sonication buffer (50 mM Tris-HCl, 150 mM NaCl, pH 7.5) followed by cell lysis via sonication (Branson B15). Cell debris was removed by centrifugation (Thermo Scientific Labofuge 400 R centrifuge; 3,500 × g, 10 min, 4°C), and the whole-membrane (WM) fraction was collected by ultracentrifugation (Beckman Coulter Optima L-100 XP ultracentrifuge; 250,000 × g, 1 h, 4°C). The WM fraction was then resuspended in 500 μl of MQ water and 500 μl of 2× solubilization buffer (100 mM Tris-HCl, 300 mM NaCl, and 2% [wt/vol] N-dodecyl-β-d-maltopyranoside [DDM] [Anatrace]) and was mixed for 16 h at 4°C. Insolubilized material was removed by ultracentrifugation (Beckman Coulter Optima Max-XP tabletop ultracentrifuge; 160,000 × g, 1 h, 4°C), and the solubilized supernatant was incubated with 100 μl of anti-DYKDDDDK G (FLAG) affinity resin (GenScript) preequilibrated with equilibration buffer (50 mM Tris-HCl, 150 mM NaCl, and 0.02% [wt/vol] DDM, pH 8) for 1.5 h at room temperature (RT). The loaded FLAG affinity resin was then washed with wash buffer (50 mM Tris-HCl, 150 mM NaCl, and 0.02% [wt/vol] DDM, pH 8) several times (10 ml total). Finally, WzyB-FLAG was eluted in 100 μl of elution buffer (0.1 M glycine, 0.02% [wt/vol] DDM, pH 3.5) at RT and neutralized with 1 M Tris, pH 9.0.

In vivo protein cross-linking.

In vivo cross-linking with DSP was performed prior to cell lysis via sonication as above. Cells were harvested from 100 ml induced culture by centrifugation (Beckman Coulter Avanti J-26 XPI centrifuge; 9,600 × g, 10 min, 4°C), and the cell pellet was washed in DSP cross-linking buffer (20 mM sodium phosphate buffer [Na₂PO₄-NaH₂PO₄], 150 mM NaCl, pH 7.5). The pellets were then resuspended in 5 ml DSP cross-linking buffer followed by incubation with 1 mM DSP (Thermo Fisher Scientific) (treated samples [T]) for 30 min at 37°C. A duplicate sample was also incubated without DSP [untreated samples (UT)]. Excess DSP was quenched with 20 mM Tris-HCl, pH 7.5, and left for 10 min at RT before resuspension in sonication buffer and purification (as described above).