Advertisement

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Tolerance mechanisms in polysaccharide biosynthesis: Implications for undecaprenol phosphate recycling in Escherichia coli and Shigella flexneri

  • Jilong Qin ,

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing

    Jilong.qin@qut.edu.au, makrina.totsika@qut.edu.au

    Affiliations Centre for Immunology and Infection Control, School of Biomedical Sciences, Queensland University of Technology, Brisbane, QLD, Australia, Max Planck Queensland Centre, Queensland University of Technology, Queensland, Australia

    ⨯
  • Yaoqin Hong,

    Roles Formal analysis, Investigation, Methodology, Resources, Writing – review & editing

    Current address: Biomedicine and Molecular Microbiology, College of Public Health, Medical and Veterinary Sciences, James Cook University, Townsville, Australia

    Affiliations Centre for Immunology and Infection Control, School of Biomedical Sciences, Queensland University of Technology, Brisbane, QLD, Australia, Max Planck Queensland Centre, Queensland University of Technology, Queensland, Australia

    ⨯
  • Nicholas T. Maczuga,

    Roles Methodology, Resources, Writing – review & editing

    Affiliation School of Biological Sciences, Department of Molecular & Biomedical Sciences, Research Centre for Infectious Diseases, University of Adelaide, Adelaide, Australia

    ⨯
  • Renato Morona,

    Roles Formal analysis, Methodology, Resources, Writing – review & editing

    Affiliation School of Biological Sciences, Department of Molecular & Biomedical Sciences, Research Centre for Infectious Diseases, University of Adelaide, Adelaide, Australia

    ⨯
  • Makrina Totsika

    Roles Formal analysis, Funding acquisition, Project administration, Resources, Writing – review & editing

    Jilong.qin@qut.edu.au, makrina.totsika@qut.edu.au

    Affiliations Centre for Immunology and Infection Control, School of Biomedical Sciences, Queensland University of Technology, Brisbane, QLD, Australia, Max Planck Queensland Centre, Queensland University of Technology, Queensland, Australia

    ⨯

Abstract

Bacterial polysaccharide synthesis is catalysed on the universal lipid carrier, undecaprenol phosphate (UndP). The cellular UndP pool is shared by different polysaccharide synthesis pathways including peptidoglycan biogenesis. Disruptions in cytosolic polysaccharide synthesis steps are detrimental to bacterial survival due to effects on UndP recycling. In contrast, bacteria can survive disruptions in the periplasmic steps, suggesting a tolerance mechanism to mitigate UndP sequestration. Here we investigated tolerance mechanisms to disruptions of polymerases that are involved in UndP-releasing steps in two related polysaccharide synthesis pathways: that for enterobacterial common antigen (ECA) and that for O antigen (OAg), in Escherichia coli and Shigella flexneri. Our study reveals that polysaccharide polymerisation is crucial for efficient UndP recycling. In E. coli K-12, cell survival upon disruptions in OAg polymerase is dependent on a functional ECA synthesis pathway and vice versa. This is because disruptions in OAg synthesis lead to the redirection of the shared lipid-linked sugar substrate UndPP-GlcNAc towards increased ECA production. Conversely, in S. flexneri, the OAg polymerase is essential due to its limited ECA production, which inadequately redirects UndP flow to support cell survival. We propose a model whereby sharing the initial sugar intermediate UndPP-GlcNAc between the ECA and OAg synthesis pathways allows UndP to be redirected towards ECA production, mitigating sequestration issues caused by disruptions in the OAg pathway. These findings suggest an evolutionary buffering mechanism that enhances bacterial survival when UndP sequestration occurs due to stalled polysaccharide biosynthesis, which may allow polysaccharide diversity in the species to increase over time.

Author summary

Enzymes involved in bacterial polysaccharide biosynthesis have substrate specificity to ensure the correct polysaccharide is produced at the appropriate place and time. However, this specificity poses a challenge for the diversification of polysaccharide structure and hence function, as the acquisition of a novel oligosaccharide repeating unit would likely disrupt the synthesis pathways, leading to sequestration of the essential universal lipid carrier UndP and ultimately cause cell death. We investigated how cells tolerate disruptions in polysaccharide synthesis pathways and provide evidence that suggests that sharing a common substrate between the synthesis pathways of two common enteric bacterial surface polysaccharides (ECA and OAg), can redirect the flow of UndP. Our study provides insights into the mechanism of how bacteria alleviate sequestration issues, thereby enhancing cell survival which may allow them additional capacity for polysaccharide diversification.

Citation: Qin J, Hong Y, Maczuga NT, Morona R, Totsika M (2025) Tolerance mechanisms in polysaccharide biosynthesis: Implications for undecaprenol phosphate recycling in Escherichia coli and Shigella flexneri. PLoS Genet 21(1): e1011591. https://doi.org/10.1371/journal.pgen.1011591

Editor: Jue D. Wang, University of Wisconsin-Madison, UNITED STATES OF AMERICA

Received: October 2, 2024; Accepted: January 23, 2025; Published: January 30, 2025

Copyright: © 2025 Qin et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All data generated or analysed during this study were included in this article and raw values for bacterial growth kinetics were included in supplementary file S1 Table.

Funding: This work is funded by an Early Career Research Ideas Grant (ECRIG2024) from the Faculty of Health, Queensland University of Technology (Australia) to JQ, and in part by an Australian Research Council project grant (DP210101317 to MT), the Max Planck Queensland Centre on the Materials Science of Extracellular Matrices to MT, The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: I have read the journal’s policy and the authors of this manuscript have the following competing interests: MT is an employee of the GSK group of companies. All remaining authors declare no competing interests. This research was conducted in the absence of any commercial or financial relationships that could be constructed as a potential conflict of interest.

Introduction

Surface polysaccharides confer enteric bacteria with resistance to host antimicrobials [1,2] and the ability to colonise various host niches [3]. The O antigen (OAg) polysaccharide attached to lipid A–core oligosaccharide of lipopolysaccharide molecules is made of oligosaccharide repeating units (RUs) and represents an important virulence factor for Gram-negative bacteria contributing to host colonisation [4] and virulence modulation [5]. The structure of OAg is under strong selection by host immunity and host niche-residing bacteriophages, giving rise to over 180 diverse OAg structures characterised to date in Escherichia coli (including Shigella strains) [6], which form the molecular basis of the O-typing scheme [7].

The biosynthesis of OAg relies on the universal lipid carrier undecaprenol phosphate (UndP or C55-P), which is shared with the biosynthesis of other polysaccharides, such as enterobacterial common antigen (ECA) and peptidoglycan in the cell (Fig 1A). However, cells only produce a finite amount of UndP (<1% of total membrane lipids [8]) through de novo synthesis (Fig 1A), and the demand for UndP during synthesis of different polysaccharides is critically ensured through efficient recycling (Fig 1A). In the biosynthesis cycle of OAg and ECA for most E. coli and S. flexneri strains, UndP is engaged by the initial transferase, N-acetylglucosamine-1-phosphate transferase WecA, to form UndPP-GlcNAc in an enzymatically reversible manner [9] (Fig 1B). UndP is then committed to either the OAg or ECA synthesis cycle in the subsequent step catalysed by the second glycosyltransferase, WbbL (for OAg in E. coli K-12 [10]) or WecG (for ECA), and remains occupied by RUs during their assembly (Fig 1B), including the subsequent membrane-translocation step by flippases WzxB and WzxE, respectively (Fig 1A). UndPP is then released in the periplasm from UndPP-OAg (by OAg polymerase WzyB and ligase WaaL in a competitive manner [11]), and UndPP-ECA (by ECA polymerase WzyE and ligase WaaL) (Fig 1A). While both released ECA and OAg RUs can be ligated onto LPS by WaaL to form ECALPS and smooth LPS (S-LPS), respectively, ECA can also be ligated to phospholipids to generate phosphatidylglycerol-linked ECA (ECAPG) on the bacterial cell surface, or resides in the periplasm in a cyclic form (ECAcyc), through a process that remains to be defined [12]. For peptidoglycan synthesis, UndP is engaged after RU assembly by MraY, translocated by MurJ, and released by penicillin-binding proteins (PBPs) and proteins belonging to the SEDS (shape, elongation, division, and sporulation) family (Fig 1A). The released UndPP is further dephosphorylated into UndP by pyrophosphatases and translocated back into the cytosolic leaflet for subsequent rounds of synthesis (Fig 1A).

thumbnail
Download:
Fig 1. Biogenesis of bacterial polysaccharides on the universal lipid carrier UndP.

A) Schematic representation of UndP de novo synthesis, glycan RU engagement, release and recycling during OAg, ECA and peptidoglycan biosynthesis in E. coli. PP, periplasm; IM inner membrane; CP, cytoplasm. B) Schematic representation of the biogenesis of ECA and OAg RUs on the UndP lipid carrier with the shared initial saccharide GlcNAc. Glycosyltransferases responsible for saccharide assembly are shown in orange.

https://doi.org/10.1371/journal.pgen.1011591.g001

The enzymes collectively responsible for the biosynthesis and assembly of different polysaccharides recognise the substrates with a common feature, UndPP-linked glycans. To ensure that the different polysaccharides are synthesised and assembled correctly, synthesis pathways have specificities towards glycan moieties on the lipid carrier in three different processes: step-wised assembly of UndPP-RUs through glycosyltransferases, IM translocation of UndPP-RUs through flippases, and polymerisation of RUs through polymerases. The specificity of glycosyltransferases is also towards acceptors (UndPP-glycans), and disruptions in glycosyltransferases will abolish or stall the RU assembly on UndPP [13]. Flippase specificity ensures that only the complete UndPP-RUs are efficiently translocated across the IM, leaving any incomplete RUs in the cytoplasmic leaflet of the IM [14]. Polymerases (Wzy) only efficiently polymerise the specific RUs for a given pathway [15]. Therefore, genes coding for OAg glycosyltransferases, flippases and polymerases have high sequence diversity and are often specific to an individual OAg gene cluster, a feature that has been exploited as a molecular serotyping method for rapid strain identification and clinical detection [16].

Although enzyme specificities are crucial to the quality control of polysaccharide biosynthesis, they may represent a limitation in the evolutionary diversification process for their polysaccharide RU structures for serotype switching. We have previously shown that expression of the mono-rhamnosyltransferase WbbL in S. flexneri that is defective in di-rhamnosyltransferase RfbG resulted in cell lysis [13]. This is because the incomplete OAg RU intermediate UndPP-GlcNAc-Rha catalysed by WbbL is not recognised neither by the last glycosyltransferase RfbF nor the flippase WzxB in S. flexneri, causing sequestration of UndP in OAg biosynthesis, thereby limiting the availability of UndP for peptidoglycan synthesis. This finding is consistent with previous studies [17,18], which suggest that cell shape abnormalities resulting from defects in peptidoglycan synthesis—caused by disruptions in the late steps of OAg biogenesis—are due to a limited supply of UndP. These abnormalities can be alleviated by overexpressing UppS, which increases the de novo synthesis of UndPP. Consequently, direct genetic disruptions in the late steps of polysaccharide biogenesis, including ECA, OAg, and capsule polysaccharide [19–21], may be difficult to achieve due to potential lethal stress, which can select for secondary suppressor mutations. These observations suggested that the evolution of a new OAg RU type (serotype switching) comes with a risk of deleterious effects, in that the newly incorporated functional glycosyltransferase may stall the existing OAg synthesis due to specificities of polysaccharide enzymes, leading to UndP sequestration and cell death. Therefore, understanding bacterial tolerance to UndP sequestration, particularly in the context of disruptions to synthesis pathways, can provide insights into the evolution of polysaccharide diversification, a widespread biological process in bacteria.

Herein, we studied the survival and tolerance of E. coli and S. flexneri to UndP sequestration in mutants defective in UndPP-releasing steps during both OAg and ECA synthesis. We showed that E. coli and S. flexneri have different tolerance to the disruption of OAg polymerases. The high tolerance to OAg polymerase disruption in E. coli was due to increased redirection of the substrate UndPP-GlcNAc towards ECA synthesis, whereas the low tolerance to OAg polymerase disruption in S. flexneri was due to the low ECA synthesis capacity of this bacterium. Our data suggest a buffering mechanism, where through sharing the initial substrate UndPP-GlcNAc between OAg and ECA synthesis, E. coli can alleviate UndP sequestration stress during OAg synthesis by redirecting UndP flow to the ECA synthesis cycle and ensure increased cell survival.

Results

The OAg polymerase WzyB contributes to rapid UndP recycling

Biosynthesis of OAg occupies a pool of UndP [22]. In E. coli K-12 strain MG1655, OAg synthesis is inactive due to the disruption of wbbL gene encoding the glycosyltransferase responsible for the committed step of OAg RU assembly (Fig 1B). Consequently, expression of WbbL in MG1655, or restoration of OAg synthesis in E. coli K-12 strain MG1655-S by removal of IS element in wbbL gene [23] increased sensitivity towards bacitracin, an antibiotic targeting the pyrophosphatases for UndP recycling (Fig 1A), without affecting cell viability (Fig 2A). Limiting the level of UndP by sequestering it in OAg or ECA synthesis has a negative impact on cell survival as it limits the synthesis of essential cell wall component peptidoglycan. In E. coli K-12 strain MG1655, switching on OAg production by inducing the committed-step glyosyltransferase WbbL in a flippase wzxB mutant background inhibits cell growth (Fig 2B). Disruption of the UndPP-OAg flippase wzxB has been shown to sequester UndP in UndPP-OAg intermediates by radioactive labelling [22], and disruptions in late glycosyltransferases WbbK and WbbJ, which also sequester UndP in incomplete UndPP-OAg intermediates were documented with cell shape deformities [18], a hallmark of peptidoglycan synthesis defects. The cell shape abnormalities in the late glycosyltransferase mutants were shown to be fully rescued by overexpression of UppS (to increase UndPP de novo synthesis) or MurA (to increase competition for UndP in peptidoglycan synthesis), suggesting that the cell death was due to UndP shortage in peptidoglycan synthesis pathways [18]. These results collectively confirmed that sequestration of UndP in the UndPP-linked OAg RU intermediates in the cytosolic face of IM is lethal. In contrast, in the WbbL-complemented MG1655 strain background (Fig 2C), disruption of the OAg ligase WaaL, known to sequester UndP in the UndPP-OAg intermediates in the periplasmic leaflet [22,23], is not lethal. We reasoned that the polymerisation reaction by WzyB in the absence of WaaL releases enough UndPP to be recycled for peptidoglycan synthesis (Fig 2D), and therefore only the disruption of both WaaL and WzyB would be lethal. Indeed, while the absence of either WaaL or WzyB had no impact on cell growth, lack of both WzyB and WaaL inhibited cell growth (Fig 2C). These results suggested that the OAg polymerase WzyB contributes to rapid UndP recycling. Our findings are in line with a previous report showing that disruption of wzyB accumulated a high level of UndPP-OAg RUs [22].

thumbnail
Download:
Fig 2. Synthetic lethality of E. coli ΔwaaLΔwzyB mutant due to complete stalling of UndPP-OAg intermediates.

A) Bacitracin sensitivity assay for E. coli MG1655 and its derivative strains. Overnight bacterial cultures were 10-folded diluted and spotted on LB agar plates supplemented without or with 1 mg/ml bacitracin. B-C) Growth curves of E. coli K-12 strains harbouring plasmids without or with wbbL cultured in LB media supplemented with or without anhydrotetracycline (AnTet). The status of OAg production is labelled as OAgON in red, or OAgOFF in green. D) Schematic representation of enzymatic reactions catalysed by WaaL and WzyB in releasing UndPP from UndPP-OAg RUs. n = 17–21 in E. coli K-12 for O16-OAg.

https://doi.org/10.1371/journal.pgen.1011591.g002

Disruption of wzyB in S. flexneri is lethal

Interestingly, gene encoding OAg polymerase WzyB in S. flexneri, together with a previously confirmed essential gene rfbF [13] were reported to completely lack transposon insertions in a dense transposon insertion library, unlike genes up- and down-stream (rfbBDAC and rfbJ, respectively) [24], strongly indicating that WzyB is essential in S. flexneri. Here, we found that attempting to construct a direct wzyB deletion in S. flexneri resulted in a few small and slow-growing colonies (Fig 3A). By a wzyB-specific PCR, it was confirmed that the putative wzyB mutant colonies contained a wzyB duplication, which likely acts as a suppressor mutation (Fig 3B). Since WzyB contributes to rapid UndP recycling, we suspected that disruption of wzyB may cause UndP sequestration in S. flexneri, impacting its growth. We therefore made the wzyB deletion in an initial transferase ΔwecA mutant background (which is disrupted for UndP engagement in both ECA and OAg synthesis) to avoid potential UndP sequestration and confirmed this to be the case by successful deletion of wzyB in this background (Fig 3A–3B). Strikingly, complementation of WecA expression in the ΔwecAΔwzyB double mutant resulted in cell lysis with the release of cellular DNA into culture supernatant (Fig 3C). These results suggest that WzyB is essential in S. flexneri, implying that the deletion of wzyB causes UndP sequestration (via OAg RU buildup) leading to cell death.

thumbnail
Download:
Fig 3. WzyB is essential in S. flexneri.

A) Colony number and morphology of putative wzyB inactivation mutants in either the background of WT or wecA-null S. flexneri. B) wzyB-specific PCR for wzyB inactivation with primers targeting coding sequences of WzyB. Successful insertions of neo in wzyB is indicated as wzyB::neo. C) Amount of cellular DNA detected by ethidium bromide (EtBr) in S. flexneri ΔwecAΔwzyB culture supernatant released upon pWecA induction. Cell lysis was imaged as loss of turbidity in culture media. The status of OAg production is labelled as either OAgON in Red or OAgOFF in Green.

https://doi.org/10.1371/journal.pgen.1011591.g003

Redirection of UndPP-GlcNAc between ECA and OAg to mitigate UndP sequestration

Besides the OAg polymerase WzyB being essential in S. flexneri, the ECA polymerase WzyE was also found to be essential in E. coli K-12 strains lacking OAg due to inactivation of the gene encoding for the committed-step glycosyltransferase WbbL. This is potentially due to the sequestration of UndP in the UndPP-ECA intermediates, since we have generated evidence suggesting that polymerase Wzy is crucial for a rapid UndPP recycling. Indeed, accumulation of dead-end intermediates in the ECA pathway were shown to induce cell morphological defects through UndP sequestration, adversely affecting peptidoglycan synthesis [17]. In addition, accumulation of UndPP-ECA intermediates in strains with disruptions of wzxE was shown previously to be lethal [25]. We also showed here that the induction of WecG in the MG1655 ΔwecGΔwzxE double mutant lacking OAg resulted in cell death (Fig 4A), confirming that the sequestration of UndP in ECA synthesis route in E. coli K-12 is also lethal. Interestingly, wzyE was not essential in E. coli strains producing OAg (ST131) [26]. Moreover, wzyE could be deleted in E. coli K-12 strain MG1655-S restored for OAg production (Table 1). We also successfully deleted wzyE in MG1655-SΔwecA (disruption in engaging UndP for both OAg and ECA synthesis). In this strain, switching on both ECA and OAg production by inducing WecA in a ΔwecAΔwzyE double mutant revealed no growth defects, suggesting that when OAg is produced, WzyE is not essential (Fig 4B). We reasoned that this is because, in most E. coli strains, OAg shares the same first sugar N-acetylglucosamine (GlcNAc) with ECA, allowing the common substrate UndPP-GlcNAc to be redirected into making OAg (Fig 1B) when UndP is sequestered in a wzyE mutant. This prompted us to examine the essentiality of the OAg polymerase WzyB in the OAg-restored E. coli K-12 strain with ECA synthesis inactivated (MG1655-SΔwecAΔwecGΔwzyB). Intriguingly, when ECA synthesis is inactivated through the disruption of committed step catalysed by WecG (Fig 1B), induction of WecA in MG1655-SΔwecAΔwecGΔwzyB resulted in cell lysis (Fig 4C–4D), suggesting that the OAg polymerase WzyB is essential when ECA synthesis is inactivated. Disruptions of genes responsible for late glycosyltransferases wbbK and wbbJ and flippase wzxB in an OAg producing MG1655 strain background were well characterised previously with cell morphological changes due to sequestration of UndP in the UndPP-linked OAg RU intermediates [18], and the cell abnormalities in these mutants were further dampened when ECA biosynthesis was inactivated. However, the growth kinetics for these mutant strains were not previously characterised. To confirm that sequestration of UndPP-linked incomplete OAg RU intermediates in MG1655-S is lethal when ECA is inactivated, we generated the triple mutant MG1655-SΔwecAΔwecGΔwbbJ and confirmed that the induction of WecA in this mutant causes stalled cell growth followed by cell lysis indicated by a decrease in cell culture turbidity (Fig 4E). Therefore, lethality of MG1655-SΔwecGΔwzyB (Fig 4C) is also likely to be due to the sequestration of UndP in UndPP-linked OAg intermediates. Indeed, increasing the de novo synthesis of UndPP by overexpression of UppS rescued the lethality of MG1655-SΔwecGΔwzyB (Fig 4F). These results strongly suggests that cell lethality observed in strains with disruptions in late OAg synthesis steps were due to sequestration of UndP. Together, these results led us to propose a model that by sharing a common initiating sugar GlcNAc between ECA and OAg, bacteria could gain increased tolerance to UndP sequestration through redirecting their common substrate UndPP-GlcNAc into remaining functional synthesis pathways to ensure rapid UndP recycling and cell viability.

thumbnail
Download:
Fig 4. Interdependence of wzy essentiality between OAg and ECA synthesis pathways.

A-C) Growth curves of indicated E. coli K-12 strains harbouring plasmids either without or with wecG (A) or wecA (B, C, E &F) cultured in LB media supplemented either with or without arabinose (ara). D) Release of cellular DNA of MG1655-SΔwecAΔwecGΔwzyB in culture supernatant upon pWecA induction detected by ethidium bromide (EtBr). Cell lysis was imaged as loss of turbidity in culture media. The status of OAg production is labelled as either OAgON in Red or OAgOFF in Green.

https://doi.org/10.1371/journal.pgen.1011591.g004

Low tolerance to UndP sequestration in S. flexneri due to limited ECA production

OAg in both WbbL-restored E. coli K-12 and S. flexneri initiates with GlcNAc, yet showed different degrees of cell viability due to UndP sequestration upon wzyB disruption (Fig 2C&3C). This prompted us to examine ECA production in both strains. We first confirmed that disruptions of both wecG and wzyE abolished the detection of ECA in both MG1655 and MG1655-S with anti-ECA antibodies, showing that the antibody is specific. Consistent with our above-proposed model, E. coli K-12 lacking OAg (MG1655) produced a high level of ECA in comparison to its OAg restored strain MG1655-S (Fig 5A). In contrast, in S. flexneri strain 2457T, inactivation of OAg biosynthesis by disruption of rmlD, involved in the synthesis of the precursor for the second sugar, L-Rhamnose within the OAg RU, only marginally increased the ECA production level (Fig 5A). Consistent with a previous study done in S. flexneri [27], disruption of wzyB in MG1655-S was also found with a slight increase in ECA production (Fig 5A). Disruption of waaL in both MG1655 and MG1655-S retained ECA production detected by Western immunoblotting, albeit with decreased signal intensity (Fig 5A). This is consistent with a previous report demonstrating that disruption of waaL in multiple bacterial strains greatly reduced reactivity with anti-ECA antiserum in whole cell agglutination experiment [28].

ECA exists in three different forms, i.e. ECALPS, ECAPG on cell surface and ECAcyc in the periplasm [12]. The formation of ECALPS is dependent on the O-antigen ligase WaaL [29]. Given that it is not feasible to detect ECAcyc molecule by Western immunoblotting due to its low molecular weight (~2,513 Da) [30], our polyclonal anti-ECA antibodies hence recognises both ECAPG and ECALPS. Mature ECA is predominantly located on the bacterial cell surface, and we therefore examined ECA surface production in both E. coli and S. flexneri. Consistent with the literature that surface labelling of ECA does not work in bacterial strains with LPS capped with OAg [31], the production of OAg on the cell surface masks ECA detection by anti-ECA antibodies for both E. coli K-12 MG1655-S and S. flexneri 2457T (Fig 5B). In contrast, inactivation of OAg production unexpectedly revealed that surface ECA could only be detected on approximately 30% of the cells in the S. flexneri ΔrmlD population, in comparison to 95% for E. coli K-12, albeit with similar surface ECA levels detected in ECA-positive bacterial cell for both strains (Fig 5B–5D). These results in part may explain the different tolerance to wzyB disruption between E. coli K-12 and S. flexneri (Fig 2C&3C), in that the capacity to make ECA in the S. flexneri ΔrmlD cell population is limited (30%) in comparison to E. coli MG1655 cell population (95%), leading to inadequate redirection of UndPP-GlcNAc into making ECA when wzyB is disrupted for the majority of the cell population (~70%). This limited ECA production in S. flexneri when OAg is disrupted renders it more susceptible to UndP sequestration, and ultimately cell death when WzyB is disrupted.

thumbnail
Download:
Fig 5. Different ECA biosynthesis levels revealed between E. coli K-12 and S. flexneri upon OAg inactivation.

A) Western immunoblots of ECA of whole bacterial lysis samples from indicated bacterial strains. Detection of periplasmic marker SurA was used as a loading control. The status of OAg and ECA biogenesis for each strain is marked as OFF and ON. ECA detection levels were normalised against SurA and the ratio of ECA to SurA band intensities in MG1655 were defined as 100% to normalise all data. B) Surface ECA immunodetection via Epifluorescence microscopy. Scale bar shown as 5 μm. C) Quantification of ECA stained fluorescence intensity of whole bacteria, mean fluorescence intensity across each bacteria was used to perform quantification, n = 40. D) Quantification of ECA surface stained bacteria in the population of indicated bacterial strains quantified from three independent micrographs, at least 200 bacteria were counted per micrograph.

https://doi.org/10.1371/journal.pgen.1011591.g005

To test this hypothesis, we ectopically over-expressed WecG in S. flexneri ΔrmlD and showed by Western immunoblotting that the ECA levels were increased by approximately 2-fold (Fig 6A). Interestingly, while surface ECA levels remained similar (Fig 6B–6C), the percentage of ECA-positive cells in S. flexneri ΔrmlD expressing WecG were increased to 64.5% in comparison to S. flexneri ΔrmlD carrying the vector control (Fig 6D). Intriguingly, expression of WecG greatly reduced the cell lysis of S. flexneriΔwecAΔwzyB mutant upon induction of WecA expression (Fig 6E). In addition, expression of UppS also rescued the lethality of the strain (Fig 6E). Taken together, these results suggest that limited ECA production in S. flexneri can account for its low tolerance to WzyB polymerase disruption and the lethality was due to the sequestration of UndP in UndPP-linked OAg intermediates.

thumbnail
Download:
Fig 6. Expression of WecG increased ECA production and tolerance to UndP sequestration in S. flexneri ΔwzyB mutant.

A) Western immunoblots of ECA in whole bacterial lysates from indicated bacterial strains. Detection of periplasmic marker SurA was used as a loading control. ECA detection levels were normalised against SurA and the ratio of ECA to SurA band intensities in 2457TΔrmlD was defined as 100% to normalise all data. B) Surface ECA immunodetection via Epifluorescence microscopy. Scale bar shown as 5 μm. C) Quantification of ECA stained fluorescence intensity of whole bacteria, mean fluorescence intensity across each bacterial cell was used to perform quantification, n = 30. D) Quantification of ECA stained bacteria in the population of indicated bacterial strains quantified as percentages from three micrographs. At least 90 bacterial cells were counted per micrograph. E) Release of cellular DNA of PE860ΔwecAΔwzyB with or without plasmids expressing WecG or UppS in culture supernatant upon pWecA induction detected by ethidium bromide (EtBr). Cell lysis was imaged as loss of turbidity in culture media.

https://doi.org/10.1371/journal.pgen.1011591.g006

Discussion

Polysaccharide synthesis requires a cellular pool of universal lipid carrier UndP. Indeed, the production of OAg, while having no observable impact on growth, was shown to sequester a low level of UndP via UndPP-RU intermediates [22]. This was also supported here by the increased sensitivity of OAg-producing E. coli K-12 to bacitracin, an antibiotic that limits the availability of functional UndP. When the OAg synthesis pathway was disrupted beyond the committed steps, UndP was shown to be sequestered in UndPP-RU intermediates at a high level [22], severely affecting bacterial survival [13] due to deleterious effects on peptidoglycan synthesis [18]. Therefore, rapid assembly of polysaccharide RU on UndP carrier is favoured as it would lower the cellular demand for UndP at any given time.

RU-synthesis controlled system reveals cell lethality outcomes

Disruption of OAg ligase waaL was shown previously to sequester high levels of UndPP-RU in OAg-producing bacterial strains [22]. However, an OAg-producing E. coli K-12 with waaL deletion had no growth impact here albeit was shown to induce a mild cell morphological defect previously [18] and was confirmed to lack suppressor mutations by whole genome sequencing. We have revealed here that in the OAg-producing E. coli K-12 strain MG1655-S, the polymerase WzyB which consolidates the RUs into polymeric forms and therefore rapidly releases UndPP in the absence of WaaL in the periplasm is crucial for cell viability. In a previous study [22], the level of C14 labelled OAg RUs on the isolated UndP pool was quantified in both wzyB and waaL mutants, with higher signal level of UndPP-linked OAg RU detected in the wzyB mutant. In the wzyB mutant, the stoichiometric relationship between OAg RU and UndP is 1:1, whereas it is (17–21):1 in the waaL mutant. Therefore, this suggests that a wzyB mutant sequestered a substantially higher level of UndP than a waaL mutant, suggesting that WzyB contributes to the rapid release of UndP from the OAg synthesis pathway. Supporting this notion, we have also found that in S. flexneri, deletion of wzyB is lethal. However, two wzyB mutants in S. flexneri have been reported previously. One of the only two reported S. flexneri wzyB mutants was acquired through Sf6c phage selection (S. flexneri Y OAg as its primary receptor) [32]. However, this mutant could not be fully complemented to confer resistance to Colicin E2 to WT level [33], a toxin whose entry is blocked by polymerised OAg [34]. Moreover, this mutant was later also used to derive a ΔwzyBΔwaaL double mutant in S. flexneri with no reported impact on growth [35], whereas it was shown here that deletion of both wzyB and waaL is lethal due to completely stalled UndPP-OAg intermediates. The other previously reported S. flexneri wzyB mutant [27] was acquired through direct allelic exchange mutagenesis after numerous attempts in the laboratory. The colony morphology of those putative mutants on the mutagenesis selection plate was similar as reported here, being overall small and heterogeneous in size. Like the other S. flexneri wzyB mutant that resulted from phage selection, this wzyB mutant could not be complemented to produce S-LPS to WT level [27]. Therefore, we believe that the two previously reported wzyB mutants likely contain suppressor mutations that have tuned down the committed step of OAg synthesis to reduce UndP sequestration for fitness, similar to suppressor mutations in rml genes reported previously in an OAg-producing E. coli K-12, making OAg at a reduced level [23].

While in a previous study [18], it was demonstrated that OAg dead-end intermediates due to genetic interruptions in OAg synthesis steps including flippase WzxB and late glycosyltransferases WbbJ and WbbK resulted in cell morphological abnormalities including loss of cell shape with bulges around cell envelope. The cell morphological defects in these mutants were due to the undersupply of UndP in peptidoglycan synthesis (which was sequestered in disrupted OAg synthesis pathways) as overexpression of UppS or MurA rescued these phenotypes [18]. Here we also generated evidence that expressing UppS could rescue the lethal phenotypes due to disruptions of WzyB in both S. flexneri and E. coli K-12 MG1655-S with inactive ECA production. Similarly, over-expression of WecG, the committed glycosyltransferase for ECA RU assembly, also rescued the cell morphological defects in mutants with disruptions in OAg biosynthesis [18] and showed here to alleviate the cell lysis in S. flexneri ΔwzyB mutant, consistent with our model here that the shared substrate UndPP-GlcNAc could be redirected into ECA synthesis pathway when OAg synthesis pathway is disrupted to allow adequate UndP recycling. However, whether these mutant strains with direct genetic disruptions incur cell lysis during various growth phases was not characterised. We argue that such mutants are highly likely to be genetically unstable, and this is well supported by work carried out by Nikaido and colleagues in 1969 [20], where studies with disruptions of OAg late glycosyltransferase made strains genetically unstable with unstable phenotypes, and the authors also reasoned that such mutation with UndP sequestration could result in additional peptidoglycan synthesis issue thus select for suppressor mutations that allow cell survival (revertant) with various phenotypes. Therefore, we avoided direct genetic disruptions of OAg synthesis pathways in MG1655-S as it may inevitably involve secondary selection of suppressor mutations to allow cell survival, producing phenotypes that may complicate the interpretation of results.

Together, these results caution future investigations on the impacts of potential UndP sequestration when disrupting genes responsible for polysaccharide synthesis, highlighting the necessity of conducting genetic studies on these genes using an expression controlled system at initial transferase or committed-step glycosyltransferase [13].

Cell death requires high level of UndP sequestration

In an OAg production-controlled system (through tightly controlled expression of WbbL) employed in this work, we showed that the accumulations of OAg dead-end intermediates in OAg-producing E. coli K-12 ΔwaaL and ΔwzxB mutants although were all documented with cell shape changes previously [18], only deletions of wzxB or wzyB/waaL in the OAg-producing E. coli K-12 caused cell lethality, which for the first time further distinguished the outcomes of UndP sequestration at different levels. Single deletions of wzyB or waaL was shown previously to accumulate UndPP-OAg intermediates [22], however only when disrupted together is the cell growth completely stalled. This is because UndPP-OAg-1RU is a common substrate competed by both WzyB and WaaL [11]. Under normal conditions, WaaL incorporates approximately 11% OAg-1RU into lipid A-core oligosaccharide [36]. However, in the absence of WzyB, WaaL ligates substantially more OAg-1RU onto lipid-A core oligosaccharide to form semi-rough LPS (SR-LPS, Fig 5A), thereby releasing adequate UndPP for recycling. In contrast, when both OAg ligase WaaL and polymerase WzyB are absent, the UndP employed in the common substrate UndPP-OAg-1RU would be sequestered fully, completely blocking the recycling path in the OAg synthesis pathway. This blockage results in lethality, consistent with the rationale for the cell death observed in wzxB deletion mutants.

Different cell death outcomes were previously documented in the ECA biosynthesis pathway, where only the gene deletion of wzxE caused lethality [25]. In two previous reports, disruptions of rmlA (in ECA gene cluster) [37] and wecF [38] that would accumulate intermediates, UndPP-GlcNAc-ManNAc, were successfully constructed, implying that such accumulation is not lethal. However, it is worth mentioning that the rmlA deletion could be in part functionally complemented by rmlA from the OAg gene cluster. This is supported in their work where a traceable amount mature ECA can still be detected [37], in which case, the ECA intermediate is not completely stalled to explain its viability. The wecF mutant was generated through direct transposon insertion [38] and whether it harbours secondary mutations is unclear. Nevertheless, the level of UndP sequestration in the situation where the incomplete RU is the dead-end intermediate (e.g. UndP-GlcNAc-ManNAc in a wecF deletion mutant) is predicted to be less than in the situation where the complete RU is the dead-end intermediate (e.g. UndP-GlcNAc-ManNAc- Fuc4NAc in a wzxE deletion mutant). This is because in the wzxE deletion mutants, UndP may be requested in both incomplete and complete ECA-RU dead-ends. This is clearly demonstrated in a previous report [39] where the ΔwzxE mutant sequestered the highest level of UndP in comparison to all disruptions in preceding steps of ECA synthesis. Therefore, it is likely that a UndP sequestration threshold exists to drastically impair peptidoglycan synthesis to induce cell death phenotype.

We showed here for the OAg synthesis pathway in MG1655-S, a ΔwbbJ deletion in the absence of ECA production caused cell death. We reasoned that the intermediates accumulated in a ΔwbbJ mutant would also completely stall the UndP sequestered in the disrupted pathway. This is consistent with previous studies showing that disruptions in late glycosyltransferases in OAg synthesis pathways in Salmonella enterica (whose OAg initiates with galactose rather than GlcNAc, therefore preventing redirection of the common substrate into the ECA pathway when OAg synthesis is disrupted) resulted in the accumulation of 10-fold more UndPP-linked intermediates [40], leading in cell death [41]. Therefore, the lack of redirection of the common substrate UndPP-GlcNAc into ECA synthesis pathway could result in low tolerance towards disruptions in OAg synthesis. In contrast, the accumulation of UndPP-GlcNAc-ManNAc intermediates in the two mutants (ΔrmlA and ΔwecF) mentioned above requires efficient supply of its donor substrate UDP-ManNAc. The substrate UDP-ManNAc was reported previously to be also utilised in the synthesis of a novel glycan for phage N4 infection [42]. This indicates an additional level of redirection of the common substrate UDP-ManNAc away from forming the dead-end intermediate UndP-GlcNAc-ManNAc, thus alleviate the severity of sequestration. Aside from ECALPS, ECA polymers are also diverged to give ECAcyc and ECAPG. While it may be tempting to correlate our phenotype to the intricate fate of ECA substrate, this would be overly speculative as the biosynthetic detail of the non-LPS forms remained unclear. Nevertheless, these data collectively support our model that common substrate could be redirected into intact synthesis pathways to limits the accumulations in the disrupted polysaccharide synthesis pathway.

Increased ECALPS may in part account for the increased ECA production when OAg synthesis is disrupted

Contrasting to the essentiality of wzyB reported here in S. flexneri, a wzyB mutant was found to have no impact on growth in an OAg-producing E. coli K-12, and was reported previously with no suppressor mutations [23], hence being genetically stable. Both OAg of S. flexneri and E. coli K-12 are initiated with GlcNAc, which is the same initiating sugar for ECA. We generated evidence suggesting that E. coli K-12 redirects its UndPP-GlcNAc into ECA synthesis when OAg synthesis is disrupted, thereby alleviating the UndP sequestration stress. This was strongly supported with further experiments where inactivation of ECA synthesis rendered wzyB essential in OAg producing E. coli K-12. In contrast, we found that disruptions of OAg synthesis in S. flexneri only marginally increased overall ECA production, suggesting that it was unable to redirect adequate UndPP-GlcNAc from the disrupted OAg synthesis pathways, resulting in cell death due to UndP sequestration.

Interestingly, ECA was detected in very low levels in both MG1655ΔwaaL and MG1655-SΔwaaL mutants through Western immunoblotting in comparison to MG1655 and MG1655-S. Disruption of waaL would abolish formation of ECALPS, redirecting UndPP-ECA to the synthesis of other forms of ECA including ECAcyc. Given that it is not feasible to detect ECAcyc by Western immunoblotting due to its low molecular weight, the production of ECAcyc is unknown when waaL is disrupted. Therefore, it is possible that ECA production in MG1655ΔwaaL and MG1655-SΔwaaL mutants could be severely underestimated. The detailed analysis of ECAcyc is beyond the scope of this work as it is WzzE-dependent [30], and would warrant a separate study with new mutant combinations. Nevertheless, we could detect an increase in ECA production by Western immunoblotting in MG1655-SΔwzyB in comparison to MG1655-S. These results therefore support our model that production of ECA is increased when OAg biosynthesis is disrupted.

ECALPS was known to be present in E. coli R1, R2, R4 and K-12 LPS core type, while not in R3 core type [31,43]. S. flexneri has LPS R3 core type, a non-permissive core type to form ECALPS. Hence it is likely that the surface ECA stained in S. flexneri ΔrmlD mutant is the only other surface form, ECAPG. The result therefore is tempting to suggest that S. flexneri has limitations in producing ECALPS due to baring a non-permissive LPS core type, thereby having low capacity in redirecting UndPP-GlcNAc into making more ECA. However, although there is lack of characterisation of ECA forms in S. flexneri to confirm the absence of ECALPS, ECALPS was reported to only exist in small amount (less than 5%) in rough LPS E. coli with different core types [31] and was proposed to be a by-product when OAg synthesis is inactive [43]. Indeed, silver staining of MG1655 LPS sample detected no ECA substituted LPS, albeit the sensitivity of the staining is at ng range [44]. Therefore, it is likely that the limitation of ECA production is not entirely due to the potential inability in ECALPS synthesis in S. flexneri. Supporting to this notion, we have also shown that overexpression of WecG in S. flexneriΔrmlD increased both the ECA detection in Western immunoblotting, and the percentage of cells stained with ECA, and rescued the lethality in S. flexneriΔwzyB mutant. Regardless of different forms of ECA may be detected by our ECA antibodies, the results presented here support our model in that increased ECA production leads to increased tolerance to the disruptions in OAg synthesis pathways.

ECA is distributed heterogeneously in S. flexneri cell population

Interestingly, we unexpectedly found only a limited S. flexneri cell population (30%) decorated their cell surface with ECA upon disruptions of OAg synthesis. Since all bacterial cells were grown from a single bacterial colony for all experimental repeats, it is tempting to suggest that the surface ECA decoration in S. flexneri may be phase variable. This is the first example of surface ECA being heterogeneously displayed on the surface of S. flexneri cells in the population, but it remains unclear what mechanism underpins this. Since polysaccharide synthesis including the ECA synthesis pathway, lacks a feedback control mechanism, whereby disruptions in late steps of biosynthesis cause UndP sequestration leading to cell death [39], it is therefore likely to be regulated at the committed biosynthesis step, i.e. the committed glycosyltransferase WecG and enzymes (WecB and WecC) responsible for the synthesis of its nucleotide sugar substrate [42]. However, through sequence analysis between S. flexneri and E. coli, we were unable to identify any genetic regions that would potentially account for regulatory differences. Nevertheless, it may also be regulated through an additional molecular mechanism that is currently not known.

Disruption of Wzy polymerases induces lethal phenotypes in polysaccharide biosynthesis pathways

Wzy polysaccharide polymerases for ECA in E. coli K-12 [26,45] was also reported to be essential. However, ECA polymerase wzyE was found not essential in uropathogenic E. coli OAg-producing strains [2,26], and here with an OAg restored E. coli K-12, highlighting the redirection of the common substrate UndPP-GlcNAc between ECA and OAg synthesis pathways as a mechanism to mitigate UndP sequestration stress. In addition, wzyE was found also essential in OAg producing Salmonella Typhimurium strains [26]. This provides further supporting evidence, because the OAg for Salmonella Typhimurium initiates with galactose [46], different to the ECA initiating sugar GlcNAc, and is thereby unable to redirect UndP-GlcNAc into synthesis of OAg.

Although numerous groups have characterised different enteric bacteria with saturated transposon insertional mutant libraries, we argue that the results may not reflect the gene essentiality accurately for the polysaccharide biosynthesis pathway. This is because to retain the mutant that may result in slow-growing phenotype, the mutant library may only be allowed for growing to early exponential phase, while that the cell lysis due to disruptions in polysaccharide synthesis genes may not occur immediately. This is evident in our results where the disruption of polysaccharide genes at late steps only starting to show cell lysis at mid-exponential phase at approx. OD600 of 0.4. Therefore, this growth-phase dependent lysis may not be well-captured in those studies with exhaustive mutagenesis. In particular, disruption of wzxE was shown here and previously [25] to be lethal in MG1655, while was shown not essential in other studies in mutant libraries [47,48]. Another complication in studies through direct inactivation of polysaccharide genes was unintentional selection of potential secondary suppressor mutations allowing survival of the mutant with targeted mutation in polysaccharide genes, similar to the selection of a S. flexneri ΔwzyB duplication suppressor mutation demonstrated in this study.

It is worth noting that cell death may not be the only outcome when late biosynthesis steps are disrupted for other polysaccharides. In our study, the mechanism of cell death was specifically attributed to the sequestration of UndP, which limits its availability for the peptidoglycan biosynthesis pathway. However, several alternative mechanisms can mitigate cell death when UndP is sequestered in the biosynthesis pathway: 1) In polysaccharide pathways where production is generally less active during the exponential growth phase and more active during the stationary phase—when peptidoglycan biosynthesis is also less active—or at lower temperatures, which slow cell growth and reduce the demand for UndP in peptidoglycan biosynthesis, as seen in exopolysaccharide biosynthesis [49]; 2) In cells that carry genes with redundant functions, such as the rml genes present in both the ECA and OAg pathways; and 3) In certain bacterial species, where the availability of UndP may be higher. However, cell death outcomes have been widely observed for other polysaccharides, including capsule polysaccharide [21], exopolysaccharide succinoglycan [50] and protein O-linked glycan [51], when late biosynthesis steps are disrupted, as previously summarised [52]. Therefore, the essentiality of polysaccharide genes must be carefully studied within a polysaccharide synthesis-controlled system to avoid the complications mentioned above.

Sharing the initial substrate with ECA biosynthesis pathway provides benefit in OAg diversification adaptation process

The diversification of the OAg RU structure presents significant challenges due to the high specificity of the enzymes involved in RU assembly and processing, such as glycosyltransferases, flippases, and polymerases. When OAg structure is modified at the cytoplasmic leaflet of the IM, these existing enzymes may not recognise the altered UndPP-RU, leading to stalled synthesis. This disruption in the pathway can result in UndP sequestration and ultimately cause cell lysis, which potentially imposes evolutionary constraints for further diversification processes. Interestingly, in contrast to E. coli, the diversification of S. flexneri OAg structures is primarily restricted to modifications introduced by bacterial phage elements, which encode enzymes modifying OAg structures on the periplasmic face of the inner membrane [53]. This approach allows S. flexneri to bypass the specificity of the enzymes responsible for OAg RU biosynthesis. Our data suggest that, in addition to the pathogenic importance of S. flexneri OAg structures, a reduced capacity for ECA synthesis could pose an evolutionary constraint to its OAg diversification. Specifically, this limitation could impede the ability of S. flexneri to adapt its OAg structures efficiently through genetic recombination to acquire a new glycosyltransferase or other alterations leading to the evolution of a novel OAg structure, in that the limited tolerance to disrupted flows in OAg synthesis would quickly eliminate the recombinants before the acquisition of adaptive changes to evolve glycosyltransferase and flippase compatibility. We argue this may explain the restricted OAg structural diversity in S. flexneri compared to other E. coli lineages.

Most E. coli, Shigella and other Enterobacteriaceae have their OAg RU initiates with GlcNAc (or subsequently modified to from GlcNAc to GalNAc by epimerase on the lipid carrier in a reversible manner [9]) as their first sugar by the initial transferase WecA [6] encoded in the ECA gene cluster, thereby sharing the first RU intermediate UndPP-GlcNAc with ECA RU synthesis. We propose a model in which sharing the initial sugar, GlcNAc, between OAg and ECA RUs allows for the redirection of UndPP-GlcNAc when one pathway encounters stalling or disruptions. This mechanism helps mitigate the stress caused by UndP sequestration (Fig 7A). Therefore, ECA synthesis with a high level of production would enhance the cell’s tolerance to UndP sequestration caused by disruptions in OAg synthesis. This increased production allows the cell to temporarily decorate its surface with ECA, which is crucial for maintaining cell survival while enabling further diversification of the new OAg (Fig 7B). Indeed, E. coli O14 (ECA) reference strain Su4411-41 decorating cell surface with ECA was found with OAg gene cluster deleted via homologous recombination [54]. The same situation was found with other rough LPS E. coli strains devoid of OAg, where these strains decorate their surface with O14 (ECA) [31,43]. In contrast, limited ECA production results in low tolerance to disruptions in OAg biosynthesis, leading to cell death and potentially imposing constraints on the evolution process (Fig 7C).

thumbnail
Download:
Fig 7. A model of the buffering mechanism that redirects UndP into ECA biosynthesis to maintain rapid UndP recycling during OAg pathway stalling or disruption.

A) OAg and ECA RU assembly shares the initial substrate UndPP-GlcNAc which is catalysed by initial transferase WecA to engage the lipid carrier UndP. The availability of UndP to peptidoglycan biosynthesis is crucial for cell survival. When OAg biosynthesis is disrupted beyond the committed steps, UndPP-RU intermediates accumulate on the IM, locking UndP into the OAg biosynthesis pathway. B) This can be mitigated by redirecting UndPP-GlcNAc into a robust ECA biosynthesis pathway to maintain rapid UndP recycling rates. This cell survival supports further diversification of the new OAg. C) when the ECA pathway has limited biosynthesis capacity, disruptions in OAg biosynthesis may result in cell death halting further evolution process of OAg.

https://doi.org/10.1371/journal.pgen.1011591.g007

Conclusion

Our work is of great importance as only by revealing the death outcomes of these disruptions in late steps of OAg biogenesis, could we understand the significant hurdles of which OAg diversification may have to face, i. e. disruptions in late steps of OAg biosynthesis may cause UndP sequestration leading to cell death. This was completely overlooked due to the lack of characterisation of growth phenotypes of mutant strains with disruptions in late steps of polysaccharide synthesis. Our work here thus pointed out once again the importance of using an initial/committed-step glycosyltransferase-controlled system studying OAg biosynthesis that involved in genetic disruptions of subsequent assembly steps to report phenotype for interpretation based on genetically stable mutants strains, and provided with further evidence that robust ECA synthesis could help in redirection of the common substrate UndPP-GlcNAc of ECA and OAg, thereby alleviating the degree of UndP sequestration and allowing continuation of further genetic adaptation.

Methods and materials

Bacterial strains and plasmids

Bacterial strains and plasmids used in this work are listed in Table 1. Single colonies grown on Lysogeny Broth (LB)-Lennox [55] agar plates were picked and grown overnight in LB at 37°C for all experiments. Where appropriate, LB media were supplemented with ampicillin (Amp, 100 μg/mL), kanamycin (Kan, 50 μg/mL), chloramphenicol (Chl, 25 μg/mL), anhydrotetracycline (AhTet 50 ng/mL), or arabinose (Ara, 10 mM).

thumbnail
Download:
Table 1. Strains, plasmids, and oligonucleotides.

https://doi.org/10.1371/journal.pgen.1011591.t001

Bacterial mutagenesis via allelic exchange

Mutagenesis was performed as previously described [59] with modifications [64]. Bacterial strains with plasmid pKD46 were grown overnight in 10 mL LB at 30°C and diluted 1 in 100 into 10 mL LB. Lambda Red protein expression was induced with 50 mM L-arabinose when OD600 reached 0.3 and continued for 1 hour. Cells were then centrifuged (5,000 ◊g), washed with ice-cold water, and resuspended in 100 μL of 10% ice-cold glycerol for electroporation. The cat or neo gene was PCR-amplified from pKD3 or pKD4, respectively, using primers with 40–50 bp of homologous sequences (Table 1). The purified PCR product (1.5 μg) was introduced into electrocompetent cells by electroporation. After recovery in 3 mL LB at 37°C for 2 hours, cells were plated on LB agar with Chl or Kan and incubated at 37°C for 16 hours. Mutants were then confirmed by PCR screening.

Bacitracin sensitivity assay

Bacterial survival spotting assays were conducted as previously described [23]. Overnight bacterial cultures were adjusted to an OD600 of 1.0, then serially diluted 10-fold to 10−7 in fresh LB media. A 4 μL aliquot of each dilution was spotted onto LB agar plates, with or without 1 mg/ml bacitracin (Sigma, B0125).

Bacterial growth kinetic assay

Bacterial growth kinetics were recorded as described previously [23]. Overnight bacterial cultures were diluted 1:200 in fresh LB media, with or without 10 mM arabinose and/or 50 ng/mL anhydrotetracycline, in a 96-well plate. The plate was incubated at 37°C with aeration, and OD600 was measured every 10 minutes for 18 hours using a CLARIOstar plate reader (BMG, Australia).

Bacterial cell lysis assay

To measure bacterial cell lysis, WecA and WbbL were induced in bacterial cells grown to an OD600 of 0.8–1 and incubated for an additional 30 minutes at 37°C. Cultures showing reduced turbidity due to cell lysis were imaged. The supernatants were then collected by centrifugation at 20,000 × g and mixed with 5 μg/mL Ethidium bromide (EtBr, BioRad). Fluorescence was measured using a CLARIOstar plate reader (BMG, Australia) with excitation at 525 nm and emission at 615 nm.

LPS silver staining

For LPS silver staining, bacterial cells (109) were harvested via centrifugation, and lysed in 50 μl of SDS sample buffer. Samples were then heated at 100°C for 10 min, then treated with 20 μg/ml proteinase K (NEB, #P8107S) overnight at 60°C for 18 hours. Samples (3–5 μl) were then separated by SDS-Tris/Glycine gel (BioRad, #4568095) electrophoresis and subjected to LPS silver staining as detailed previously [44].

Western immunoblotting

For western immunoblotting of ECA, the above samples prepared for LPS silver staining were separated by SDS-Tris/Glycine gel (BioRad, #4568095) electrophoresis and subsequently transferred onto nitrocellulose membrane and detected with rabbit polyclonal anti-ECA antibodies [27]. For loading control, lysed whole cell bacterial samples without proteinase K treatment were separated by SDS-PAGE and subsequently immunoblotted with anti-SurA antibodies (gifted by Carol Gross, University of California). For comparisons of ECA detection between samples, densitometry analysis was performed by using Image Lab (Version 6.1.0, BioRad). Background adjusted band intensity volume of detected SurA signals were used to normalise the band intensity volume of detected ECA signals.

Surface ECA immunostaining

For ECA surface labeling, bacteria (108 cells) in mid-exponential phase were harvested by centrifugation (16,000 × g, 1 min), fixed with 3.7% (wt/vol) formaldehyde in PBS for 20 minutes at room temperature, and then washed with PBS. A 5 μL suspension of fixed bacteria was centrifuged onto coverslips precoated with 0.01% (wt/vol) poly-L-lysine (Sigma) in a 24-well tray (16,000 × g, 1 min). Coverslips were incubated sequentially with rabbit anti-ECA pAbs (1:100), followed by anti-rabbit Alexa Fluor 488 (Invitrogen, 1:100) in PBS with 10% (vol/vol) FBS (Gibco), with PBS washes in between. Coverslips were mounted with ProLong Diamond Antifade Mountant (Invitrogen) and imaged using a ZEISS Axio Vert.A1 microscope. The experiment was carried out twice, and the mean intensity of ECA signals across each whole bacterium and the percentage of ECA-positive bacteria was quantified using three micrographs in ZEISS ZEN blue (Version 3.6, Carl Zeiss).

Supporting information

S1 Table. Raw data for growth kinetics.

Recorded values of optical density at 600 nm of bacterial cultures growing at 37°C.

https://doi.org/10.1371/journal.pgen.1011591.s001

(XLSX)

Acknowledgments

The Ian Potter Foundation sponsored the CLARIOStar high-performance microplate reader (BMG, Australia) and epifluorescence microscope.

References

  1. 1. Chaput C, Spindler E, Gill RT, Zychlinsky A. O-antigen protects gram-negative bacteria from histone killing. PloS one. 2013;8(8):e71097. Epub 2013/08/21. pmid:23951089; PubMed Central PMCID: PMC3738592.
  2. 2. Phan MD, Peters KM, Sarkar S, Lukowski SW, Allsopp LP, Gomes Moriel D, et al. The serum resistome of a globally disseminated multidrug resistant uropathogenic Escherichia coli clone. PLoS Genet. 2013;9(10):e1003834. Epub 2013/10/08. pmid:24098145; PubMed Central PMCID: PMC3789825.
  3. 3. Smith HW. Survival of orally administered E. coli K 12 in alimentary tract of man. Nature. 1975;255(5508):500–2. Epub 1975/06/05. pmid:1094297.
  4. 4. Day CJ, Tran EN, Semchenko EA, Tram G, Hartley-Tassell LE, Ng PS, et al. Glycan:glycan interactions: High affinity biomolecular interactions that can mediate binding of pathogenic bacteria to host cells. Proceedings of the National Academy of Sciences of the United States of America. 2015;112(52):E7266–75. pmid:26676578; PubMed Central PMCID: PMC4702957.
  5. 5. West NP, Sansonetti P, Mounier J, Exley RM, Parsot C, Guadagnini S, et al. Optimization of virulence functions through glucosylation of Shigella LPS. Science. 2005;307(5713):1313–7. pmid:15731456.
  6. 6. Liu B, Furevi A, Perepelov AV, Guo X, Cao H, Wang Q, et al. Structure and genetics of Escherichia coli O antigens. FEMS microbiology reviews. 2020;44(6):655–83. Epub 2019/11/30. pmid:31778182; PubMed Central PMCID: PMC7685785.
  7. 7. Kauffmann F. The serology of the coli group. J Immunol. 1947;57(1):71–100. Epub 1947/09/01. pmid:20264689.
  8. 8. Barreteau H, Magnet S, El Ghachi M, Touze T, Arthur M, Mengin-Lecreulx D, et al. Quantitative high-performance liquid chromatography analysis of the pool levels of undecaprenyl phosphate and its derivatives in bacterial membranes. J Chromatogr B Analyt Technol Biomed Life Sci. 2009;877(3):213–20. Epub 2008/12/27. pmid:19110475.
  9. 9. Rush JS, Alaimo C, Robbiani R, Wacker M, Waechter CJ. A novel epimerase that converts GlcNAc-P-P-undecaprenol to GalNAc-P-P-undecaprenol in Escherichia coli O157. The Journal of biological chemistry. 2010;285(3):1671–80. Epub 2009/11/20. pmid:19923219; PubMed Central PMCID: PMC2804325.
  10. 10. Liu D, Reeves PR. Escherichia coli K12 regains its O antigen. Microbiology (Reading). 1994;140 (Pt 1):49–57. Epub 1994/01/01. pmid:7512872.
  11. 11. Hong Y, Hu D, Verderosa AD, Qin J, Totsika M, Reeves PR. Repeat-Unit Elongations To Produce Bacterial Complex Long Polysaccharide Chains, an O-Antigen Perspective. EcoSal Plus. 2023:eesp00202022. Epub 2023/01/10. pmid:36622162.
  12. 12. Rai AK, Mitchell AM. Enterobacterial Common Antigen: Synthesis and Function of an Enigmatic Molecule. Mbio. 2020;11(4). Epub 2020/08/14. pmid:32788387; PubMed Central PMCID: PMC7439462.
  13. 13. Qin J, Hong Y, Totsika M. Determining glycosyltransferase functional order via lethality due to accumulated O-antigen intermediates, exemplified with Shigella flexneri O-antigen biosynthesis. Appl Environ Microbiol. 2024;90(6):e0220323. Epub 2024/05/15. pmid:38747588; PubMed Central PMCID: PMC11218652.
  14. 14. Hong Y, Cunneen MM, Reeves PR. The Wzx translocases for Salmonella enterica O-antigen processing have unexpected serotype specificity. Mol Microbiol. 2012;84(4):620–30. Epub 2012/04/14. pmid:22497246.
  15. 15. Hong Y, Morcilla VA, Liu MA, Russell ELM, Reeves PR. Three Wzy polymerases are specific for particular forms of an internal linkage in otherwise identical O units. Microbiology (Reading). 2015;161(8):1639–47. Epub 2015/05/20. pmid:25987464.
  16. 16. Wang L, Reeves PR. Organization of Escherichia coli O157 O antigen gene cluster and identification of its specific genes. Infect Immun. 1998;66(8):3545–51. Epub 1998/07/23. pmid:9673232; PubMed Central PMCID: PMC108385.
  17. 17. Jorgenson MA, Kannan S, Laubacher ME, Young KD. Dead-end intermediates in the enterobacterial common antigen pathway induce morphological defects in Escherichia coli by competing for undecaprenyl phosphate. Mol Microbiol. 2016;100(1):1–14. Epub 2015/11/26. pmid:26593043; PubMed Central PMCID: PMC4845916.
  18. 18. Jorgenson MA, Young KD. Interrupting Biosynthesis of O Antigen or the Lipopolysaccharide Core Produces Morphological Defects in Escherichia coli by Sequestering Undecaprenyl Phosphate. J Bacteriol. 2016;198(22):3070–9. Epub 2016/10/23. pmid:27573014; PubMed Central PMCID: PMC5075036.
  19. 19. Le Bas A, Clarke BR, Teelucksingh T, Lee M, El Omari K, Giltrap AM, et al. Structure of WzxE the lipid III flippase for Enterobacterial Common Antigen polysaccharide. Open biology. 2025;15(1):240310. Epub 2025/01/08. pmid:39772807; PubMed Central PMCID: PMC11706664.
  20. 20. Yuasa R, Levinthal M, Nikaido H. Biosynthesis of cell wall lipopolysaccharide in mutants of Salmonella. V. A mutant of Salmonella typhimurium defective in the synthesis of cytidine diphosphoabequose. J Bacteriol. 1969;100(1):433–44. Epub 1969/10/01. pmid:4899003; PubMed Central PMCID: PMC315411.
  21. 21. Xayarath B, Yother J. Mutations blocking side chain assembly, polymerization, or transport of a Wzy-dependent Streptococcus pneumoniae capsule are lethal in the absence of suppressor mutations and can affect polymer transfer to the cell wall. J Bacteriol. 2007;189(9):3369–81. Epub 2007/02/27. pmid:17322316; PubMed Central PMCID: PMC1855910.
  22. 22. Liu D, Cole RA, Reeves PR. An O-antigen processing function for Wzx (RfbX): a promising candidate for O-unit flippase. J Bacteriol. 1996;178(7):2102–7. Epub 1996/04/01. pmid:8606190; PubMed Central PMCID: PMC177911.
  23. 23. Qin J, Hong Y, Morona R, Totsika M. O antigen biogenesis sensitises Escherichia coli K-12 to bile salts, providing a plausible explanation for its evolutionary loss. PLoS Genet. 2023;19(10):e1010996. Epub 2023/10/04. pmid:37792901; PubMed Central PMCID: PMC10578602.
  24. 24. Freed NE, Bumann D, Silander OK. Combining Shigella Tn-seq data with gold-standard E. coli gene deletion data suggests rare transitions between essential and non-essential gene functionality. BMC Microbiol. 2016;16(1):203. Epub 2016/09/08. pmid:27599549; PubMed Central PMCID: PMC5011829.
  25. 25. Rick PD, Barr K, Sankaran K, Kajimura J, Rush JS, Waechter CJ. Evidence that the wzxE gene of Escherichia coli K-12 encodes a protein involved in the transbilayer movement of a trisaccharide-lipid intermediate in the assembly of enterobacterial common antigen. The Journal of biological chemistry. 2003;278(19):16534–42. Epub 2003/03/07. pmid:12621029.
  26. 26. Ghomi FA, Jung JJ, Langridge GC, Cain AK, Boinett CJ, Abd El Ghany M, et al. High-throughput transposon mutagenesis in the family Enterobacteriaceae reveals core essential genes and rapid turnover of essentiality. Mbio. 2024:e0179824. Epub 2024/08/31. pmid:39207104.
  27. 27. Maczuga N, Tran ENH, Qin J, Morona R. Interdependence of Shigella flexneri O Antigen and Enterobacterial Common Antigen Biosynthetic Pathways. J Bacteriol. 2022;204(4):e0054621. Epub 2022/03/17. pmid:35293778; PubMed Central PMCID: PMC9017295.
  28. 28. Marx A, Petcovici M, Nacescu N, Mayer H, Schmidt G. Demonstration of enterobacterial common antigen by bacterial agglutination. Infect Immun. 1977;18(3):563–7. Epub 1977/12/01. PubMed Central PMCID: PMC421272. pmid:338482
  29. 29. Schmidt G, Mannel D, Mayer H, Whang HY, Neter E. Role of a lipopolysaccharide gene for immunogenicity of the enterobacterial common antigen. J Bacteriol. 1976;126(2):579–86. Epub 1976/05/01. PubMed Central PMCID: PMC233189. pmid:57114
  30. 30. Kajimura J, Rahman A, Rick PD. Assembly of cyclic enterobacterial common antigen in Escherichia coli K-12. J Bacteriol. 2005;187(20):6917–27. Epub 2005/10/04. pmid:16199561; PubMed Central PMCID: PMC1251615.
  31. 31. Kuhn HM, Meier-Dieter U, Mayer H. ECA, the enterobacterial common antigen. FEMS microbiology reviews. 1988;4(3):195–222. Epub 1988/09/01. pmid:3078744.
  32. 32. Morona R, Mavris M, Fallarino A, Manning PA. Characterization of the rfc region of Shigella flexneri. J Bacteriol. 1994;176(3):733–47. D—NLM: PMC205111 EDAT- 1994/02/01 MHDA- 1994/02/01 00:01 CRDT- 1994/02/01 00:00 PST—ppublish. pmid:7507920; PubMed Central PMCID: PMC205111.
  33. 33. Nath P, Tran EN, Morona R. Mutational analysis of the Shigella flexneri O-antigen polymerase Wzy: identification of Wzz-dependent Wzy mutants. J Bacteriol. 2015;197(1):108–19. Epub 2014/10/15. pmid:25313393; PubMed Central PMCID: PMC4288684.
  34. 34. Tran EN, Papadopoulos M, Morona R. Relationship between O-antigen chain length and resistance to colicin E2 in Shigella flexneri. Microbiology. 2014;160(Pt 3):589–601. Epub 2014/01/16. pmid:24425769.
  35. 35. Ascari A, Tran ENH, Eijkelkamp BA, Morona R. Identification of the Shigella flexneri Wzy Domain Modulating Wzz(pHS-2) Interaction and Detection of the Wzy/Wzz/Oag Complex. J Bacteriol. 2022;204(9):e0022422. Epub 2022/08/19. pmid:35980183; PubMed Central PMCID: PMC9487639.
  36. 36. Bastin DA, Stevenson G, Brown PK, Haase A, Reeves PR. Repeat unit polysaccharides of bacteria: a model for polymerization resembling that of ribosomes and fatty acid synthetase, with a novel mechanism for determining chain length. Mol Microbiol. 1993;7(5):725–34. Epub 1993/03/01. pmid:7682279.
  37. 37. Rick PD, Wolski S, Barr K, Ward S, Ramsay-Sharer L. Accumulation of a lipid-linked intermediate involved in enterobacterial common antigen synthesis in Salmonella typhimurium mutants lacking dTDP-glucose pyrophosphorylase. J Bacteriol. 1988;170(9):4008–14. Epub 1988/09/01. pmid:2842298; PubMed Central PMCID: PMC211403.
  38. 38. Danese PN, Oliver GR, Barr K, Bowman GD, Rick PD, Silhavy TJ. Accumulation of the enterobacterial common antigen lipid II biosynthetic intermediate stimulates degP transcription in Escherichia coli. J Bacteriol. 1998;180(22):5875–84. Epub 1998/11/13. pmid:9811644; PubMed Central PMCID: PMC107660.
  39. 39. Eade CR, Wallen TW, Gates CE, Oliverio CL, Scarbrough BA, Reid AJ, et al. Making the Enterobacterial Common Antigen Glycan and Measuring Its Substrate Sequestration. ACS Chem Biol. 2021;16(4):691–700. Epub 2021/03/20. pmid:33740380; PubMed Central PMCID: PMC8080848.
  40. 40. Hong Y, Reeves PR. Sequestration of dead-end undecaprenyl phosphate-linked oligosaccharide intermediate. bioRxiv. 2024:2024.09.22.614228.
  41. 41. Hong Y, Reeves PR. Model for the Controlled Synthesis of O-Antigen Repeat Units Involving the WaaL Ligase. Msphere. 2016;1(1). Epub 2016/06/16. pmid:27303678; PubMed Central PMCID: PMC4863624.
  42. 42. Sellner B, Prakapaite R, van Berkum M, Heinemann M, Harms A, Jenal U. A New Sugar for an Old Phage: a c-di-GMP-Dependent Polysaccharide Pathway Sensitizes Escherichia coli for Bacteriophage Infection. Mbio. 2021;12(6):e0324621. Epub 2021/12/15. pmid:34903045; PubMed Central PMCID: PMC8669472.
  43. 43. Maciejewska A, Kaszowska M, Jachymek W, Lugowski C, Lukasiewicz J. Lipopolysaccharide-Linked Enterobacterial Common Antigen (ECA(LPS)) Occurs in Rough Strains of Escherichia coli R1, R2, and R4. Int J Mol Sci. 2020;21(17). Epub 2020/08/26. pmid:32839412; PubMed Central PMCID: PMC7504096.
  44. 44. Tsai CM, Frasch CE. A sensitive silver stain for detecting lipopolysaccharides in polyacrylamide gels. Anal Biochem. 1982;119(1):115–9. Epub 1982/01/01. pmid:6176137.
  45. 45. Gerdes SY, Scholle MD, Campbell JW, Balazsi G, Ravasz E, Daugherty MD, et al. Experimental determination and system level analysis of essential genes in Escherichia coli MG1655. J Bacteriol. 2003;185(19):5673–84. Epub 2003/09/18. pmid:13129938; PubMed Central PMCID: PMC193955.
  46. 46. Wang L, Andrianopoulos K, Liu D, Popoff MY, Reeves PR. Extensive variation in the O-antigen gene cluster within one Salmonella enterica serogroup reveals an unexpected complex history. J Bacteriol. 2002;184(6):1669–77. Epub 2002/03/02. pmid:11872718; PubMed Central PMCID: PMC134871.
  47. 47. Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, et al. Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Syst Biol. 2006;2:2006 0008. Epub 2006/06/02. pmid:16738554; PubMed Central PMCID: PMC1681482.
  48. 48. Goodall ECA, Robinson A, Johnston IG, Jabbari S, Turner KA, Cunningham AF, et al. The Essential Genome of Escherichia coli K-12. Mbio. 2018;9(1). Epub 2018/02/22. pmid:29463657; PubMed Central PMCID: PMC5821084.
  49. 49. Sardari RR, Kulcinskaja E, Ron EY, Bjornsdottir S, Friethjonsson OH, Hreggviethsson GO, et al. Evaluation of the production of exopolysaccharides by two strains of the thermophilic bacterium Rhodothermus marinus. Carbohydr Polym. 2017;156:1–8. Epub 2016/11/16. pmid:27842803.
  50. 50. Reuber TL, Walker GC. Biosynthesis of succinoglycan, a symbiotically important exopolysaccharide of Rhizobium meliloti. Cell. 1993;74(2):269–80. Epub 1993/07/30. pmid:8343955.
  51. 51. Jebeli L, McDaniels TA, Ho DTT, Tahir H, Kai-Ming NL, Mcgaw M, et al. The Late-Stage Steps of Burkholderia cenocepacia Protein O-Linked Glycan Biosynthesis Are Conditionally Essential. bioRxiv. 2024:2024.11.21.624715.
  52. 52. Hong Y, Liu MA, Reeves PR. Progress in Our Understanding of Wzx Flippase for Translocation of Bacterial Membrane Lipid-Linked Oligosaccharide. J Bacteriol. 2018;200(1). Epub 2017/07/12. pmid:28696276; PubMed Central PMCID: PMC5717161.
  53. 53. Sadredinamin M, Yazdansetad S, Alebouyeh M, Yazdi MMK, Ghalavand Z. Shigella Flexneri Serotypes: O-antigen Structure, Serotype Conversion, and Serotyping Methods. Oman Med J. 2023;38(4):e522. Epub 2023/09/19. pmid:37724320; PubMed Central PMCID: PMC10505279.
  54. 54. Jensen SO, Reeves PR. Deletion of the Escherichia coli O14:K7 O antigen gene cluster. Can J Microbiol. 2004;50(4):299–302. Epub 2004/06/24. pmid:15213754.
  55. 55. Lennox ES. Transduction of linked genetic characters of the host by bacteriophage P1. Virology. 1955;1(2):190–206. Epub 1955/07/01. pmid:13267987.
  56. 56. Qin J, Doyle MT, Tran ENH, Morona R. The virulence domain of Shigella IcsA contains a subregion with specific host cell adhesion function. PloS one. 2020;15(1):e0227425. Epub 2020/01/08. pmid:31910229; PubMed Central PMCID: PMC6946128.
  57. 57. Van den Bosch L, Morona R. The actin-based motility defect of a Shigella flexneri rmlD rough LPS mutant is not due to loss of IcsA polarity. Microb Pathog. 2003;35(1):11–8. pmid:12860454.
  58. 58. Cronan JE. A family of arabinose-inducible Escherichia coli expression vectors having pBR322 copy control. Plasmid. 2006;55(2):152–7. Epub 2005/09/06. pmid:16139359.
  59. 59. Datsenko KA, Wanner BL. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proceedings of the National Academy of Sciences of the United States of America. 2000;97(12):6640–5. WOS:000087526300074. pmid:10829079
  60. 60. Larue K, Ford RC, Willis LM, Whitfield C. Functional and structural characterization of polysaccharide co-polymerase proteins required for polymer export in ATP-binding cassette transporter-dependent capsule biosynthesis pathways. The Journal of biological chemistry. 2011;286(19):16658–68. Epub 2011/04/02. pmid:21454677; PubMed Central PMCID: PMC3089508.
  61. 61. Hong Y, Reeves PR. Diversity of o-antigen repeat unit structures can account for the substantial sequence variation of wzx translocases. J Bacteriol. 2014;196(9):1713–22. Epub 2014/02/18. pmid:24532778; PubMed Central PMCID: PMC3993327.
  62. 62. Maczuga NT. Enterobacterial Common Antigen Biosynthesis in Shigella flexneri 2022.
  63. 63. Maczuga N, Tran ENH, Morona R. Subcellular localization of the enterobacterial common antigen GT-E-like glycosyltransferase, WecG. Mol Microbiol. 2022;118(4):403–16. Epub 2022/08/26. pmid:36006410; PubMed Central PMCID: PMC9804384.
  64. 64. Qin J, Hong Y, Pullela K, Morona R, Henderson IR, Totsika M. A method for increasing electroporation competence of Gram-negative clinical isolates by polymyxin B nonapeptide. Scientific reports. 2022;12(1):11629. Epub 2022/07/09. pmid:35804085; PubMed Central PMCID: PMC9270391.