Enterobacterial Common Antigen: Synthesis and Function of an Enigmatic Molecule

The outer membrane (OM) of Gram-negative bacteria poses a barrier to antibiotic entry due to its high impermeability. Thus, there is an urgent need to study the function and biogenesis of the OM. In Enterobacterales, an order of bacteria with many pathogenic members, one of the components of the OM is enterobacterial common antigen (ECA). We have known of the presence of ECA on the cell surface of Enterobacterales for many years, but its properties have only more recently begun to be unraveled.

D iverse environmental conditions on Earth (e.g., heat, pH, salinity, pressure, and osmotic activity) immensely affect the function of the cell, necessitating adaptation through structural modification. Gram-negative bacteria have an impermeable and strengthened outer membrane (OM) that allows them to withstand stress brought about by environmental factors, including other bacteria, antibiotics, and chemical stresses. The cell envelope structure of Gram-negative bacteria consists of the inner membrane, the periplasm containing the peptidoglycan cell wall, and the OM (1). The lipids of the OM form a barrier that is impermeable to large hydrophilic and hydrophobic molecules (2). Lipopolysaccharide (LPS) facilitates the formation of this barrier though (i) the high number of fatty acyl substituents per lipid molecule, which form a gel-like structure enhancing the rigidity of membrane (3,4), (ii) strong lateral interaction between LPS molecules mediated by salt bridges with divalent cations (5), and (iii) modification of LPS structure in response to different environmental conditions (6). For example, in Salmonella, the PhoPQ two-component system causes antimicrobial peptide resistance after induction by divalent cation starvation by activating PagP (7). PagP facilitates the addition of palmitate chain to lipid A, altering the fluidity of the LPS molecules in the OM (7).
OM proteins (OMPs) are integral membrane proteins present in the membranes of Gram-negative bacteria, mitochondria, and chloroplasts. These proteins adopt a ␤-barrel architecture arranged in the membrane in anti-parallel patterns (8). Some of these proteins (porins) can form pores in the OM (2). These OMPs regulate the movement of small hydrophilic molecules across the outer membrane, such as nutrients, water, ions, and some small hydrophilic antibiotics (2). In fact, in Escherichia coli, ␤-lactam antibiotics, tetracyclines, chloramphenicol, and fluoroquinolones quickly diffuse through OmpF (9)(10)(11)(12). Specific porins can also transport amphipathic substrates. For instance, transportation of long-chain fatty acid is facilitated by the lipid transporter FadL (13). Beyond its role controlling the entry of molecules into the cell, the OM plays a structural role, providing protection against mechanical and osmotic stresses (14,15).
The Gram-negative OM is coated in highly variable molecules that can cause immune activation, known as antigens. Bacteria are divided into serotypes based on different antigen combinations (16). The three major types of antigens present on the cell surface are O (somatic), K (capsular), and H (flagellar) (17,18). These antigens can play roles in motility (H-antigen), protection from a hostile environment (K-antigens and O-antigen), interaction with the environment (K-antigens and O-antigen), and increasing the ability of the OM to provide structural support to the cell (O-antigen) (15,(19)(20)(21). The outer leaflet of the OM is made mainly of lipopolysaccharide (LPS), which consists of lipid-A, the core polysaccharide, and O-antigen (1). O-antigen is a highly variable chain of carbohydrates and thus is serotype specific. K-antigens are the capsule, a coat on the surface of bacteria outside the cell envelope. They generally consist of high-molecular-weight polysaccharides with some exceptions (e.g., K-88 and K-99 of E. coli, which are protein antigens) (22)(23)(24). The H-antigen is a protein antigen based on flagellar structure (25). Enterobacterales is a bacterial order that is defined in part by the presence of an antigen known as enterobacterial common antigen (ECA) (26). ECA, a carbohydrate antigen, is located in the outer leaflet of the OM and in the periplasm (27)(28)(29)(30). Although Enterobacterales express various antigens (e.g., K, O, and H) (31,32), ECA is unique in that it is restricted to one order and in which it is invariant (Fig. 1A) allowing cross-reactivity among the members of Enterobacterales (33).  the rough R4 strain disguised by the production of K7 capsular antigen (51). In combination with the production of immunogenic ECA LPS by the R1, R4, and K-12 strains with rough LPS (43), this clarified that ECA LPS is produced in significant amounts only by rough strains that do not make O-antigen. In nonimmunogenic strains, including O-antigen-producing smooth strains and rough strains with incomplete LPS cores or mutations in waaL, the O-antigen ligase, ethanol-soluble ECA PG is the predominant form of ECA on the cell surface (35). While some early studies suggested purified ECA PG could induce an antibody response (52), this was only true in strains that also produced significant amounts of ECA LPS (53), suggesting that the antibody production may have resulted from contaminating ECA LPS . With our current knowledge of immunology, it can now be appreciated how the differences in structure between ECA LPS and ECA PG would lead to differences in their immunogenicity. Antibody production is not efficiently stimulated without innate immune signaling (54). ECA LPS possesses an intrinsic adjuvant to stimulate antibody production, as LPS is recognized by toll-like receptor 4 (TLR4), leading to the production of proinflammatory cytokines (55). Nevertheless, as a carbohydrate antigen, ECA LPS mainly stimulates the production of IgM low-affinity, high-avidity antibodies (56). In contrast, the production of high-affinity IgG antibodies requires a protein antigen (57). As a proof of concept of the potential immunogenicity of ECA PG , purified ECA PG that contains proteins in addition to ECA can generate an immune response (52). Likewise, a conjugate of ECA and tetanus toxoid, a classic adjuvant, produces ECA antibodies mainly of the IgG isoform (56). Many of the initial studies on ECA immunogenicity were carried out with heat-killed bacteria (50). The difference in immunogenicity of ECA LPS and ECA PG is less in live bacteria (50), likely because of the many pathogen-associated molecular patterns (PAMPs) linked with an active infection and an increased production of proinflammatory cytokines (58).
Prevalence of ECA antibodies. Many early studies have reported a low titer of ECA antibodies present in human serum (33,59,60), with the caveat that these studies were conducted before the availability of an ECA knockout strain and so may report the combined titer of both ECA antibodies and antibodies to protein antigens shared among Enterobacterales (e.g., OmpA). These antibodies have been found in both healthy donors and, at higher levels, in patients with chronic urinary tract infections (61). The titers of ECA antibodies present in the blood have been reported to increase with age (62). A maternal vaccination study reported that the cord blood of a child has lower amounts of ECA antibodies than maternal serum, showing some low level of maternal transfer (61). In other mammals such as cats, dogs, horses, pigs, and mice, ECA antibodies have been reported in blood sera with the exception of rabbits, where no ECA antibody is found. The likely reason for this is due to the high colonization of rabbits by Gram-positive bacteria and low prevalence of E. coli (60). Among several strains of mice, C57B1/6HA mice have ample ECA antibody titers after responding to ECA LPS immunogens (63-65).
Estimates of ECA antibody prevalence in various types of infections have indicated variable titers of ECA antibodies (35). In a few diseases such as enteritis, bacteremia, and acute urinary tract infections, low titers of ECA antibodies were detected. However, higher ECA antibody titers were found in shigellosis (61,66), peritonitis (53,67), and chronic urinary tract infections (61,68,69). Surprisingly consistent levels of ECA antibodies were observed in a rabbit pyelonephritis model immunized by ECA LPS , irrespective of the route of infection and diagnosis (70)(71)(72).
The presence of ECA antibodies has been detected in human serum after infection from pathogenic bacteria such as E. coli, Yersinia enterocolitica O3 strains (26,(73)(74)(75), and in humans suffering from arthritis associated with Proteus mirabilis strains (76). The ECA immunogenicity is mainly due to ECA LPS (26), with exception to the Rc mutant R4/O28 of P. mirabilis in which ECA PG provides immunogenicity. Therefore, the ECA can be used as a tool for serodetection in various infections caused by the members of the Enterobacterales order (77).
Role of ECA antibodies in host-pathogen interactions. Several studies have attempted to determine whether ECA antibodies play a role, either protective or pathogenic, in various disease contexts. Under experimental conditions, an appreciable amount of ECA antibodies was found in mice immunized by E. coli O-antigen-negative strains and later challenged by pathogenic bacteria (63)(64)(65)78), except in Swiss albino mice (65). However, protection from these immunizations was partial and temporary. The same partial protection was also observed in mice by passive immunization with serum from rabbits inoculated with E. coli O14 (63). Furthermore, a clinical study has demonstrated that passive immunization with a human monoclonal ECA antibody had no protective effect during sepsis caused by Enterobacterales (79).
Bridge et al. (80) reported that, in a mouse model of salmonellosis, infection with the ΔwecA strain of Salmonella enterica serovar Typhimurium SL1344 via the oral or intraperitoneal route provided cross-protection against this infection by the production of IgG antibodies. In addition, Huang et al. (81) reported cross-protection against heterologous Salmonella strains in mice by downregulating the expression of O-antigen (rmlB [rfbB]) and ECA (rmlB ECA [rffG]) biosynthesis genes, allowing production of proteinrecognizing antibodies. These studies suggest that, at least for Salmonella, ECA antibodies may not be protective and that ECA may "distract" the immune system from more efficacious targets. It should be mentioned that some studies have correlated the presence of antibodies to Enterobacterales, including to ECA, with rheumatoid arthritis (82)(83)(84); however, the causal relationships leading to the antibody production remains unclear.

FORMS AND BIOSYNTHESIS OF ECA
Over the years, several researchers have developed methods for isolating ECA (Table 2). However, the chemical composition of the antigen was initially difficult to identify, due in part to the existence of ECA in three forms (26) (Fig. 1B). These forms are ECA linked to diacylglycerol through phosphodiester linkage (ECA PG ), LPS-linked ECA (ECA LPS ), and periplasmic cyclized ECA (ECA CYC ) (43,47,48). Observations made by Kunin, Neter, and Suzuki initially demonstrated that ECA occurred in two forms in the immunogenic E. coli O14. One of the forms was soluble in aqueous ethanol and was able to be separated from the LPS (ECA PG ), while the other did not dissolve in ethanol and could not be separated from LPS (ECA LPS ) (35,37). ECA PG is the predominant surface-exposed form of ECA, while ECA LPS, the immunogenic form that allows antibody production, is predominantly found in rough mutants that do not make O-antigen (35). Later studies identified a third form of ECA, ECA CYC (48), which is now thought to be present in all species that make ECA (48,(85)(86)(87)(88). Initial studies done by Männel and Mayer in 1978 (89) reported that ECA consists of N-acetyl-D-glucosamine (GlcNAc) and N-acetyl-D-mannosaminuronic acid (ManNAcA). But in 1983, another component of ECA, 4-acetamido-4,6-dideoxy-D-galactose (Fuc4NAc), was identified in Shigella by Lugowski et al. (56), demonstrating that ECA consists of a trisaccharide repeating unit (Fig. 1A).  (Fig. 2). The genes necessary for many steps in the synthesis of ECA are located within an operon known as the wec operon. The wec operon begins at position 85.4 centisomes on the E. coli K-12 chromosome (90), and the functions of each gene of the operon have been analyzed (90)(91)(92)(93)(94)(95)(96)(97)(98).
As for many other extracytoplasmic glycans (99), the assembly of the ECA trisaccharide repeat unit is carried out on an isoprenoid lipid carrier, undecaprenylphosphate (Und-P), a 55-carbon molecule made of isoprenoid units (93,(100)(101)(102). The assembly occurs on the inner side of the plasma membrane (103)(104)(105)(106)(107). The first step involves the formation of GlcNAc-pyrophosphoryl-undecaprenol, which is also known as lipid I ECA (108). This step uses UDP-GlcNAc as a substrate to attach GlcNAc-1phosphate to Und-P and is catalyzed by WecA (93,101,102). In silico predictions,  (109,110). Furthermore, mutational studies have shown that several conserved aspartate residues in cytoplasmic loops 2 and 3 are important for WecA catalytic activity and divalent cation binding (110). Fluorescence microscopy for green fluorescent protein (GFP)-tagged WecA revealed that it is localized to punctate regions on the cell surface (110). The punctate localization of WecA suggests that ECA and/or O-antigen biosynthesis is localized to discrete cellular regions. The GlcNAc residue of ECA is nonstoichiometrically O-acetylated by WecH, an inner membrane O-acetylase (111). WecB and WecC are responsible for synthesizing UDP-ManNAcA from UDP-GlcNAc (103,112,113). Specifically, WecB (UDP-N-acetylglucosamine 2-epimerase) reversibly epimerizes at carbon position 2 between UDP-GlcNAc and UDP-N-acetylmannosamine (112,113). Campbell et al. (114) solved the structure of WecB at a 2.4-Å resolution. This homodimeric enzyme is comprised of two similar sandwich domains with the active site located in the deep cleft at the domain interface. Several basic residues in the active site may have a role in proton transfer at the C-2 position or stabilization of the oxy-carbonium ions in the transition state (114). Residues in the active site have been found to be important for allosteric regulation of WecB as well as for binding and catalysis (115). WecC oxidizes UDP-N-acetylmannosamine in the presence of NAD ϩ to form UDP-ManNAcA (116). The UDP-ManNAcA is the substrate to attach ManNAcA to the lipid I ECA , a reaction carried out by WecG (101,117). This process results in ManNAcA-GlcNAc-pyrophosphoryl-undecaprenol, also known as lipid II ECA .
TDP-glucose acts as a substrate for second enzyme, RmlB ECA (dTDP-glucose 4,6dehydratase 2). RmlB ECA converts dTDP-glucose to dTDP-4-keto-6-deoxy-D-glucose. This second reaction is a complex reaction that involves dehydrogenation, dehydration, and rereduction in the presence of cofactor NAD ϩ (121). Several active-site residues important for the function of RmlB ECA have been identified based on similarity to UDP-galactose-4-epimerase and mutational analysis (122,123).
The synthesis of lipid III ECA occurs on the inner leaflet of the cytoplasmic membrane facing the cytosol (103); however, polymerization of the trisaccharide repeat unit to form polysaccharide chains occurs on the outer leaflet of the cytoplasmic membrane (127,128). The flipping of lipid III ECA to the periplasmic face of the membrane is mediated by a "flippase," WzxE (127,128). WzxE is a member of the polysaccharidespecific transporter family of proteins, which flip polysaccharides, including some O-antigens and capsular polysaccharides, across the inner membrane (IM) (128,129). After translocation across the membrane, the ECA chain is polymerized by WzyE. The final chain length of the ECA polymer is determined by WzzE, the chain length regulator (91). These steps result in the formation of an ECA polymer attached to the isoprenoid carrier (Fig. 2).
Several studies have offered insights into the mechanism through which WzzE might control the chain length of ECA. Genetic evidence and cross-linking data support the notion that the flippase, polymerase, and chain length regulator work together as a complex (130,131). Several structural studies have been performed with WzzE and other members of the class 1 polysaccharide copolymerase family (PCP-1) (132). A crystal structure of the periplasmic domain of WzzE, solved at 2.4 Å, revealed a bell-shaped homo-octameric structure (133); however, reports have suggested various oligomeric states for other PCP-1 family members depending on whether full-length protein was used and the method of analysis (134)(135)(136)(137).
This structural inconsistent may be due to the lack of interactions with other complex members (i.e., WzyE). However, recent studies have again suggested octameric structure for both WzzE and WzzB (an E. coli O-antigen PCP-1) (136). The most recent structural data for PCP-1 proteins suggest a mechanism for PCP-1 chain length regulation where polymerization begins when the polymerase and PCP-1 form a complex with the growing polysaccharide chain wrapping either over the surface or through the cavity of the PCP-1 (137). The polymerization could then be terminated either when the PCP-1 and polymerase disassociate (a "stop-watch" mechanism) or when the polysaccharide-binding capacity of the PCP-1 is reached (a "molecular ruler" mechanism) (138)(139)(140).
Synthesis of the three forms of ECA. The three forms of ECA are made from the ECA polymer. ECA PG is the dominant membrane-associated form of ECA and constitutes about 0.2% of the cellular dry weight of E. coli K-12 (26,76,140). This form is present in all Enterobacterales (26). The polysaccharide chain is transferred from the isoprenoid carrier to an unidentified lipid to form ECA linked to diacylglycerol through phosphodiester linkage (ECA PG ) (47). In this molecule, ECA is the head group of the phospholipid (47). The newly synthesized ECA PG is then translocated to the outer membrane (27,29). The genes and mechanisms involved in the synthesis and translocation of ECA PG remain unknown (87). ECA LPS is synthesized by transferring the linear ECA polysaccharide chain to the core oligosaccharide of LPS (26,46). This step is catalyzed by WaaL, which is also responsible for attaching O-antigen to the core of LPS (43,51). However, data suggest that the method for attaching ECA to LPS in Yersinia enterocolitica may be different, allowing ECA and O-antigen to coexist (141). The last form of ECA, ECA CYC , is a cyclic molecule consisting of polymerized ECA trisaccharide repeat units, and it is water soluble (26,48). The ECA CYC is localized in the periplasm (88). This polymer has a variable number of repeat units (4-6) depending on the species (91), and the chain length regulator, WzzE, is necessary for its synthesis (48,87,88). A cyclase has not been identified. In-depth structural analysis by crystallography, nuclear magnetic resonance (NMR), and molecular dynamics have suggested that ECA CYC can exist in two threedimensional conformations (142)(143)(144). In contrast to ECA CYC , ECA PG and ECA LPS consist of 1 to 14 repeat units, with a modal value of 5 to 7 units in E. coli K-12 (91). In addition, different modal chain lengths have been observed depending on the growth temperature (145).
Interactions with other biosynthetic pathways. The use of isoprenoid carriers, such as Und-P, for the synthesis of extracytoplasmic glycans is highly conserved across the domains of life (99,146,147). Furthermore, these carriers are often utilized for the synthesis of multiple glycans in the same species. For instance, in E. coli, Und-P is used for the production of ECA, O-antigen, peptidoglycan, and the colanic acid capsule (43,(148)(149)(150)(151). Thus, Und-P is a universal lipid carrier required for the synthesis of numerous glycan polymers (152), and this can lead to complex interactions between biosynthetic pathways.
Disruption of one Und-P pathway may lead to indirect consequences on other glycans by altering the amount of Und-P and other precursors available for their synthesis. For example, obstructing the O-antigen pathway in E. coli compromises peptidoglycan biosynthesis by sequestering Und-P (153). In relation to ECA, it was first observed that disruption of later steps in ECA biosynthesis that lead to the accumulation of lipid II ECA causes detergent sensitivity (154) and bile salt sensitivity (155). It was then found that these mutations also lead to the activation of extracytoplasmic stress responses, including Cpx, E , and Rcs (156)(157)(158). Interestingly, in E. coli and Salmonella enterica, these stress responses are only activated with mutations that cause lipid II ECA accumulation (147,156), but in Serratia marcescens, Rcs activation has been reported even in strains with mutations early in ECA biosynthesis (158). The link of detergent sensitivity and stress response activation to defects in peptidoglycan biosynthesis was conclusively established by the observation that mutations leading to the accumulation of lipid II ECA cause changes in cell shape due to sequestering of Und-P (159). While the accumulation of lipid II ECA is deleterious to the cell, in E. coli, the accumulation of lipid III ECA has been shown to be lethal (88,128). This has been observed with both loss of WzyE, the ECA polymerase, and loss of the capacity to flip lipid III ECA across the inner membrane (see below) (88,128). Avoiding lipid III ECA accumulation by disrupting an earlier step in ECA biosynthesis prevents this lethality (88).
The ECA and O-antigen biosynthesis pathways are homologous in Enterobacterales. All members of Enterobacterales utilize the wec locus for the biosynthesis of ECA (102,160). However, many Enterobacterales with GlcNAc as their first O-antigen residue (e.g., Salmonella enterica serovar Minnesota [161], Salmonella enterica serovar Montevideo [162], and E. coli [43]) also utilize wecA (rfe), the first gene in ECA biosynthesis, for the production of O-antigen chains (163). Therefore, disruptions of wecA result in loss of both O-antigen and ECA biosynthesis (43,164). In addition, the functions of RmlA ECA and RmlB ECA , which function in the synthesis of dTDP-Fuc4NAc for ECA biosynthesis, are at least partially redundant with RmlA and RmlB, respectively, two enzymes involved in the synthesis of dTDP-L-rhamnose for O-antigen biosynthesis (118).
The absence of all Wzx family flippases (Wzx O16 , WzxC, and WzxE) in E. coli K-12 leads to a lethal accumulation of lipid III ECA , which can be prevented by the expression of any of the flippases or by prevention of ECA synthesis at an earlier step (88,128). However, with wild-type expression of wzx O16 and wzxC, deletion of wzxE is not lethal, and approximately wild-type levels of ECA CYC are produced (88). These data can be explained by the complex specificity of Wzx family flippases, which has recently been thoroughly reviewed (165). Recent work from Liu et al. (166) suggests "structure-specific triggering," where flipping is triggered by recognition of specific structural element(s) and other "incorrect" O-antigens may trigger flipping at low frequency when the "correct" substrate is absent (166).
Under normal conditions, the expression of Wzx flippases is low (167), and these flippases show specificity to their canonical substrates (166,(168)(169)(170)(171)(172). With various Wzx proteins, the specificity that triggers flipping has been found to be the first sugar residue attached to the Und-PP carrier (168,169), the presence of terminal side branch residues (170,171), or the identity or linkage of the terminal residue of the oligosaccharide (166,171,172). In the absence of their canonical substrate, the flippases can transport other Und-PP-linked oligosaccharides with various degrees of efficiency depending on their structural similarity, and this transport can be increased by overexpression (166,170,171). Thus, in E. coli K-12, which is O-antigen negative but would produce O-antigen with an initial GlcNAc and which does not produce colanic acid unless stressed, the O-antigen and colanic acid Wzx proteins can be utilized with enough efficiency by ECA to prevent a lethal accumulation of lipid III ECA and to produce a measurable accumulation of ECA (88), although the process may be inefficient.
Finally, WaaL, the O-antigen ligase, is responsible for attaching both O-antigen and ECA to LPS (43). Thus, the level of ECA LPS greatly depends on the presence or absence of O-antigen, with very little ECA LPS produced in O-antigen plus strains in E. coli (35). However, Yersinia enterocolitica and Proteus mirabilis produce significant amounts of ECA LPS even in the presence of O-antigen (76,105). Thus, the interactions between ECA biosynthesis, O-antigen, and peptidoglycan biosynthesis pathways are highly complex. One consequence of these complex interactions is that it confounds the interpretation of high-throughput genetic screens identifying functions of genes in ECA biosynthesis (173)(174)(175)(176).

BIOLOGICAL SIGNIFICANCE OF ECA
The order Enterobacterales contains many pathogens important to living organisms, including human beings. Early studies showed that when bacteria are subcultured for prolonged periods under laboratory conditions, the capability to synthesize O-chains diminishes, but there are no effects on the stability of ECA (37). Thus, ECA must have a vital role in Enterobacterales (Table 3). However, efforts to clarify the specific biological role of ECA for the enterobacterial cell have failed, partially because of the many interactions between the O-antigen, peptidoglycan (PG), and ECA biosynthesis pathways (43,144,(156)(157)(158)(159). Due to this, very few unambiguously assigned functions have been ascribed to ECA.
There has been no difference observed in both the autolysis and the viability in cells having ECA and their counterparts lacking the antigen (177). Barua et al. (178) have shown in the Shiga-toxin-producing E. coli strain O157:H7 that mutants in wecA and wecB resulting in the loss of ECA and O-antigen or ECA alone, respectively, are sensitive to acetic acid. In addition, there is an increased sensitivity to some antibiotics in mutants lacking ECA, especially aminoglycosides, for example, kanamycin and gentamicin (177). A similar observation was made in E. coli where a wecE mutant was found to be sensitive to gentamicin (179). In addition, Girgis et al. (180) observed that a wzxE mutant has no phenotype in neutral agar media but, in the presence of nalidixic acid and amikacin, the mutant is sensitive compared to the wild type. A large-scale chemical genetics screen suggested that a wecA mutant of E. coli K-12 lacking all ECAs was sensitive to azidothymidine, CHIR-90, minocycline, procaine, puromycin, triclosan, and trimethoprim and resistant to fusidic acid, isoniazid, amdinocillin, vancomycin, and polymyxin (181). A comparison between the ECA PG and the lipopolysaccharide of Salmonella Montevideo showed that the ECA PG has a higher Ca 2ϩ -to-Mg 2ϩ ratio than lipopolysaccharide, hence the suggestion that enterobacterial common antigen is important for the supply of calcium ions in the cell (52). Random-transposon mutagenesis experiments performed in Salmonella enterica revealed that disruption of six of the ECA operon genes (wecB, wecC, wecD, wecE, wecG, and wzxE) led to increased speed of lysis by bacteriophage P22 (176). As no effect was observed for disruption of wecA, it is possible that this effect is due to the peptidoglycan stress caused by these mutations. One of the critical roles of ECA is the pathogenicity of Enterobacterales, which has been found in some studies. For instance, Salmonella enterica lacking ECA (with mutations in either wecD or wecA) becomes less virulent and more sensitive to bile salts (147), although it does not use the wecA gene for O-antigen biosynthesis, as do many Enterobacterales. The phenotype of the wecD mutant was more severe than that for the wecA mutation, likely due to an accumulation of a Und-P-linked ECA precursor disrupting peptidoglycan synthesis. Gilbreath et al. (182) further validated this result. In vitro they found that a wecA null mutant of S. Typhimurium is deficient in ECA production but fully competent for O-antigen production and lacks stress response activation caused by peptidoglycan biosynthesis disruption. This mutant was highly attenuated in mice, causing a persistent low-level infection that did not kill the mice (182).
Interestingly, mutants defective in ECA biosynthesis trigger Rcs stress response activation in Serratia marcescens regardless of whether peptidoglycan biosynthesis is disrupted, suggesting that ECA may play an especially important role in envelope integrity for this species (150). A role in envelope integrity is also suggested by the overproduction of outer membrane vesicles (OMV) in the absence of ECA in Serratia marcescens, suggesting an instability in the OM (183). In contrast, a screen in E. coli K-12 revealed differences in OMV production in ECA biosynthesis mutants, but these phenotypes varied greatly depending on which gene was mutated, suggesting that the results may be indirect (184). Phan et al. (174) found that seven genes of the ECA operon are essential for serum resistance in E. coli; however, these effects may have been the result of the loss of O-antigen and/or isoprenoid carrier effects.
Recently, we determined that one of the forms of ECA, ECA CYC , plays a significant role in the OM permeability barrier (145). We found that loss of ECA CYC disrupts the OM permeability barrier, causing detergent and bile salt sensitivity. Furthermore, we determined that ECA CYC genetically interacts with a protein of unknown function, YhdP, to carry out this activity. When yhdP is deleted, ECA CYC takes on aberrant activity that damages the OM, despite greatly lowered ECA CYC levels. Careful screening of different ECA mutants and screening for hallmarks of peptidoglycan stress allowed us to eliminate peptidoglycan stress as a cause of these phenotypes (145).
ECA has been considered as a vaccination against infections stemming from enterobacterial strains due to its prevalence within the order. In a mouse model, salmonellosis leads to the development of ECA antibodies. However, no insight into the role of ECA antibodies in protection from infection was provided (185). Results regarding the protective nature of ECA antibodies have been mixed (63-65, 78, 79, 81). Further investigation into the protective nature of ECA antibodies against Enterobacterales species is warranted, as ECA antibodies have the potential to protect against all enterobacterial pathogens. In an ECA vaccine candidate, it would likely be important that the O-acetylation of ECA be maintained. O-Acetylation of surface-exposed polysaccharide has been shown to be important for the generation of protective antibodies for many pathogens (186)(187)(188)(189)(190)(191). Kajimura et al. (111) have reported that the absence of O-acetylation in ECA CYC decreases the immunoreactivity of this ECA form toward a rabbit-generated antibody. A recent study has determined that a partially O-acetylated polysaccharide may be highly advantageous for vaccine use due to epitope exposure and hydrophobicity profiles (192). In this aspect, ECA may be ideal given its nonstoichiometric acetylation of GlcNAc (111).

CONCLUDING REMARKS
Since the discovery of ECA in 1962, the work of many investigators has elucidated considerable information about the structure and biogenesis of ECA, and yet our understanding of this fascinating molecule remains incomplete. Whereas many carbohydrate antigens on the bacterial cell surface are highly variable allowing for escape from immune surveillance, ECA remains invariant despite its presence in many pathogenic species. This suggest that the function of ECA must be incredibly important for the cell. Yet, this function is largely unknown. In the past, investigations of the function of ECA have been hampered by the complex genetic and biosynthetic interactions between ECA and other cytoplasmic glycans (i.e., peptidoglycan and O-antigen). Now that these interactions have been characterized, more in-depth studies of ECA functions will be possible.
It is quite likely that the functions of ECA will be found to vary between the types of ECA and to occur on the level of cellular function as well as interaction with the environment. For instance, it has become clear that ECA CYC and ECA PG both play roles in maintaining the OM permeability barrier (145); however, given that antibiotic and detergent susceptibility differs between the loss of these two molecules and the different cellular location of the ECA forms, it is likely that the function of each is distinct. ECA LPS , as of yet, has had no cellular function described. While it is clear that ECA is important for pathogenesis, at least in Salmonella, it has not yet been determined whether this is due to an alteration of Salmonella cellular function or an alteration of Salmonella's interaction with the host.
Nevertheless, from the earliest studies of ECA it has been shown that ECA interacts with the host immune system, leading to the production of broadly cross-reactive antibodies. Yet, it is not known whether these ubiquitous antibodies play a role in the pathogenesis or protection from enterobacterial pathogens. Further evidence of the importance of ECA interactions with the environment can be gleaned from the regulation of ECA expression or chain length by temperature (141,145,193). Future investigations into the functions of ECA will lend important insights into the cellular function and host-pathogen interactions of an important group of bacteria.