SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Structure and function of core OS
  5. Genetic organization and regulation of the major LPS core OS biosynthesis operons
  6. Conservation in enzymes for inner core OS assembly
  7. Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases
  8. Assembly of the outer core OS — HexI substitution
  9. Assembly of the outer core OS — HexIII substitution
  10. A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4
  11. A ‘missing’ transferase for assembly of the E. coli R3 core OS
  12. What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?
  13. Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?
  14. Concluding remarks
  15. Acknowledgements
  16. References

Bacterial lipopolysaccharides (LPS) are unique and complex glycolipids that provide characteristic components of the outer membranes of Gram-negative bacteria. In LPS of the Enterobacteriaceae, the core oligosaccharide links a highly conserved lipid A to the antigenic O-polysaccharide. Structural diversity in the core oligosaccharide is limited by the constraints imposed by its essential role in outer membrane stability and provides a contrast to the hypervariable O-antigen. The genetics of core oligosaccharide biosynthesis in Salmonella and Escherichia coli K-12 have served as prototypes for studies on the LPS and lipo-oligosaccharides from a growing range of bacteria. However, despite the wealth of knowledge, there remains a number of unanswered questions, and direct experimental data are not yet available to define the precise mechanism of action of many gene products. Here we present a comparative analysis of the recently completed sequences of the major core oligosaccharide biosynthesis gene clusters from the five known core types in E. coli and the Ra core type of Salmonella enterica serovar Typhimurium and discuss advances in the understanding of the related biosynthetic pathways. Differences in these clusters reflect important structural variations in the outer core oligosaccharides and provide a basis for ascribing functions to the genes in these model clusters, whereas highly conserved regions within these clusters suggest a critical and unalterable function for the inner region of the core.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Structure and function of core OS
  5. Genetic organization and regulation of the major LPS core OS biosynthesis operons
  6. Conservation in enzymes for inner core OS assembly
  7. Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases
  8. Assembly of the outer core OS — HexI substitution
  9. Assembly of the outer core OS — HexIII substitution
  10. A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4
  11. A ‘missing’ transferase for assembly of the E. coli R3 core OS
  12. What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?
  13. Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?
  14. Concluding remarks
  15. Acknowledgements
  16. References

The general features of bacterial lipopolysaccharides (LPS) are now well established. Lipid A, the hydrophobic moiety of LPS, forms the outer leaflet of the outer membrane and is responsible for many of the biological properties attributed to endotoxic LPS (Raetz, 1996; Wyckoff et al., 1998). Attached to lipid A is a short core oligosaccharide (core OS) that is divided into two structurally distinct regions: the inner (lipid A proximal) and outer core OS. The core OS is the terminal part of the LPS molecule in rough LPS (R-LPS). However, in families such as the Enterobacteriaceae and the Pseudomonadaceae an antigenic, repeat-structure O-polysaccharide (O-PS) is attached to the distal core OS to form smooth LPS (S-LPS). Alternatively, in genera such as Haemophilus and Neisseria, the core OS may be modified by serospecific domains comprising several glycosyl residues to generate lipo-oligosaccharides (LOS). General features of lipid A and the inner region of the core OS are highly conserved (particularly within families), perhaps reflecting constraints imposed by their essential roles in outer membrane integrity. The outer parts of LPS are more likely to interact with the environment and, in some bacteria, with the host immune system. Although some variations can occur in the outer core OS structure, O-PS provide the most striking example of structural diversity, with ≈170 distinct antigens reported for Escherichia coli alone. From the viewpoint of the molecular evolution of surface antigen diversity, the structural limitations imposed on the core OS make it a very interesting counterpoint to the almost limitless variation in O-PS.

The core OS is assembled on preformed lipid A by sequential glycosyl transfer from nucleotide sugar precursors, and it is assumed that the process involves a co-ordinated complex of membrane-associated glycosyltransferases acting at the cytoplasmic face of the plasma membrane. Certainly, available sequence data predict that the majority of these enzymes are peripheral membrane proteins. Transfer of the completed lipid A-core molecule across the plasma membrane to the site of O-PS ligation (Whitfield et al., 1997) may involve the ATP-binding cassette (ABC)-transporter MsbA (Polissi and Georgopoulos, 1996). Current understanding of core OS assembly owes much to early studies on Salmonella enterica serovars (unless otherwise indicated, Salmonella is used in the following text to indicate S. enterica serovar Typhimurium). This work has been described in excellent past reviews (Mäkelä and Stocker, 1984; Schnaitman and Klena, 1993; Raetz, 1996).

The chromosomal waa region (formerly rfa; Table 1) contains the major core OS assembly operons, and E. coli K-12 represents the first waa region to be sequenced in its entirety. Although some functions have been biochemically characterized in E. coli K-12, characterization of others relies heavily on LPS structure and LPS-specific phage receptor data for Salmonella mutants and on heterologous complementation of those mutations. Comparative sequence analysis of the waa regions from type strains representing the known core OS types of E. coli and Salmonella reveals some expected similarities but also some intriguing differences in the spectrum of waa genes and the structures of predicted gene products. Homologues of core OS biosynthetic genes are now being identified in a variety of nonenteric bacteria and advances in whole-genome sequencing will result in a rapid expansion of data in this area. As many gene assignments currently rely heavily on E. coli and Salmonella prototypes, a consideration of the limitations of the functional assignments in these prototypes is critical.

Table 1. . Genes involved in the assembly and modification of the core OS of E. coli and Salmonella that are present in the waa locus of the chromosome. a. The new nomenclature system for bacterial polysaccharide synthesis genes is discussed in Reeves et al. (1996) and is also described on-line (http://www.microbio.su.oz.au/BPGD/default.htm).b. ND, sequence data not available but their presence is predicted; ST, S. enterica serovar Typhimurium; −, no gene or predicted activity; +, presence of gene and predicted activity.Thumbnail image of

Structure and function of core OS

  1. Top of page
  2. Abstract
  3. Introduction
  4. Structure and function of core OS
  5. Genetic organization and regulation of the major LPS core OS biosynthesis operons
  6. Conservation in enzymes for inner core OS assembly
  7. Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases
  8. Assembly of the outer core OS — HexI substitution
  9. Assembly of the outer core OS — HexIII substitution
  10. A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4
  11. A ‘missing’ transferase for assembly of the E. coli R3 core OS
  12. What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?
  13. Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?
  14. Concluding remarks
  15. Acknowledgements
  16. References

The core OS in E. coli and Salmonella share a common organization (Fig. 1), and the structures of many of the known core OS have been reviewed elsewhere (Holst and Brade, 1992). The core OS is not generally considered a virulence factor per se, although recent reports have implicated it in the adhesion of some bacteria to host cells (Jacques, 1996). The core OS does, however, have some indirect roles in virulence. For example, it provides the attachment site for O-PS (an accepted virulence factor in most of the Enterobacteriaceae). Also, the inner core OS plays a crucial role in establishing the essential barrier function of the outer membrane, which may explain the high degree of structural conservation in E. coli and Salmonella (Fig. 2). Limited structural variation in the core OS, in comparison to the O-antigens, has stimulated interest in the possibility of targeting the core OS for the generation of immunotherapeutic antibodies (Di Padova et al., 1993; Stanislavsky et al., 1997 and references therein).

image

Figure 1. . Generalized structure of the carbohydrate backbone of the LPS core OS (Holst and Brade, 1992). The inner core is highly conserved; it comprises three 3-deoxy-D-manno-oct-2-ulosonic acid (Kdo) and L-glycero-D-manno-heptose (Hep) residues and is often phosphorylated (Fig. 2). The outer core comprises a tri-hexose backbone modified with varying side-branch substitutions of hexose and acetamidohexose residues. Dashed line indicates non-stoichiometric substitution.

Download figure to PowerPoint

image

Figure 2. . Summary of current understanding of the inner core OS structure in E. coli and Salmonella. The structural data are described elsewhere (Holst and Brade, 1992). Dotted lines indicate the linkages formed by specific gene products and the basis for the assignments is documented in the text. Dashed arrows identify modifications that are either non-stoichiometric or that are confined to particular core OS types (as indicated). The R4 inner core region is not defined.

Download figure to PowerPoint

Until recently, only one wild-type core OS structure had been described in Salmonella spp. (chemotype Ra in serovar Typhimurium). However, a recent report has described a second structure in serovar Arizonae IIIa O62 (Olsthoorn et al., 1998) that differs from the Ra core OS in a single terminal side-branch in the outer core (Fig. 3). In E. coli there are five distinct core structures, termed K-12 and R1–R4, and these vary principally in the outer core OS regions (Fig. 3). Of these, the E. coli R1 core OS is the most prevalent in clinical isolates of E. coli (Gibb et al., 1992; Appelmelk et al., 1994). Structures identical to the E. coli R1, R3 and R4 types have been described in Shigella species (Holst and Brade, 1992 and references therein). Ironically, the K-12 core (formed by the actions of the best characterized waa gene products) is not detected at any appreciable frequency in clinical isolates (Gibb et al., 1992; Appelmelk et al., 1994; K. Amor, D. E. Heinrichs and C. Whitfield, unpublished).

image

Figure 3. . Structures of the outer core OS from the LPS of E. coli and Salmonella and their genetic determinants. Details of the structures are described elsewhere (Holst and Brade, 1992; Brade et al., 1996; Olsthoorn et al., 1998). All glycoses are in the α-anomeric configuration unless otherwise indicated. The genes whose products catalyse formation of each linkage are indicated and the basis for the assignments is discussed in the text. An asterisk denotes the residue of the core OS to which attachment of O-PS occurs. The genetic organization of the known core OS biosynthesis regions are shown. Sequence data for the core OS biosynthesis region of S. enterica serovar Arizonae IIIa are not available. The sequence of the E. coli K-12 core OS biosynthesis region has been reported previously (GenBank accession numbers X62530, M80599, M86935, AE000440, U00096, M86305, U00039 and M95398) as has the majority of the equivalent S. enterica serovar Typhimurium region (S56361, X53847, M73826, U06472, and AF026386). The regions shown for E. coli reference strains with the R1 (AF019746), R2 (AF019375), R3 (AF019745) and R4 (AF019747) core OS types were each amplified by polymerase chain reaction. The primers were based on the sequences of waaA and waaC that were known to be conserved by Southern hybridization analyses (Heinrichs et al., 1998). No sequence data are currently available for the regions upstream of waaC and downstream of waaA in the R-core strains and we have therefore omitted what are predicted to be conserved open reading frames (compared with E. coli K-12) from their chromosomal maps. The functions and predicted functions of gene products are summarized in Table 1. Further, the partial structure Glc I-1 [RIGHTWARDS ARROW] 3-HepII has not yet been published for the R4 core but is shown based on analogy to other known structures. This structure has been identified in the R1 core (D. E. Heinrichs et al., submitted).

Download figure to PowerPoint

In the laboratory, the minimal LPS structure required for E. coli and Salmonella viability consists of lipid A glycosylated with two Kdo residues, often referred to as Re LPS. This minimal structure may reflect a limitation for the successful translocation of LPS to form a viable outer membrane. In these bacteria, lipid A is not fully acylated in the absence of Kdo, and only conditional mutants of lipid A or Kdo synthesis have been isolated. This feature has been exploited in attempts to generate therapeutic antibacterial compounds (reviewed in Raetz, 1996; Wyckoff et al., 1998).

The heptose region of the core OS is crucial for outer membrane stability in E. coli and Salmonella, as phosphate substitution of HepI and HepII are proposed to facilitate both cross-linking of adjacent LPS molecules by divalent cations or polyamines, and interaction with positively charged groups on proteins. The inability to synthesize or incorporate Hep, or the loss of phosphoryl derivatives alone, gives rise to significant compositional and structural changes in the outer membrane and the pleiotropic phenotype known as ‘deep-rough’ (reviewed in Schnaitman and Klena, 1993; Raetz, 1996). ‘Deep-rough’E. coli and Salmonella strains often show an increase in phospholipid–protein ratios in the outer membrane that, in addition to the alterations just described, leads to the characteristic supersensitivity to hydrophobic compounds. The perturbation of their outer membrane structure also leads to a ‘leaky’ outer membrane that releases considerable amounts of periplasmic enzymes into the medium, and causes an induction of colanic acid expression (Parker et al., 1992). Further, lack of inner core OS phosphorylation results in an inactive form of secreted haemolysin in E. coli (Stanley et al., 1993; Bauer and Welch, 1997). These dramatic effects, together with the observation that some ‘deep-rough’ mutants show increased susceptibility to attack by lysosomal fractions of polymorphonuclear leucocytes and to phagocytosis by macrophages (Hammond et al., 1984) suggest that the heptose region of the LPS molecule may provide novel therapeutic targets in many bacteria. Interestingly, phosphorylation is not the only way to achieve a robust outer membrane because there is no phosphorylation of the core OS in Klebsiella pneumoniae— the negative charges instead being provided by galacturonic acid and Kdo residues (Severn et al., 1996; Süsskind et al., 1998). This feature may impact significantly on the outer membrane stability of Klebsiella and the biosynthetic processes leading to the modification of its inner core region provide a number of interesting avenues for further investigation.

Genetic organization and regulation of the major LPS core OS biosynthesis operons

  1. Top of page
  2. Abstract
  3. Introduction
  4. Structure and function of core OS
  5. Genetic organization and regulation of the major LPS core OS biosynthesis operons
  6. Conservation in enzymes for inner core OS assembly
  7. Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases
  8. Assembly of the outer core OS — HexI substitution
  9. Assembly of the outer core OS — HexIII substitution
  10. A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4
  11. A ‘missing’ transferase for assembly of the E. coli R3 core OS
  12. What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?
  13. Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?
  14. Concluding remarks
  15. Acknowledgements
  16. References

In E. coli and Salmonella, the core OS biosynthesis gene cluster consists of three operons (Fig. 3), located in the waa region of the chromosome, and mapping between cysE and pyrE at 81–82 min on the E. coli K-12 and Salmonella linkage maps. The gmhD, waaQ and waaA operons are defined by the first gene of each transcriptional unit.

In E. coli K-12, the gmhD gene is the first of a block of four genes (gmhD–waaFCL ) that form an operon required for inner core OS biosynthesis (Schnaitman and Klena, 1993; Raetz, 1996). The gmhD, waaF and waaC genes encode proteins involved in the biosynthesis and transfer of heptose (Table 1, Fig. 2[link]), whereas the waaL gene encodes a ligase enzyme that is required for attachment of various cell-surface polysaccharides (including O-PS) to lipid A core (Whitfield et al., 1997). One intriguing aspect of the regulation of the core OS biosynthesis region is that transcription of the gmhD operon in E. coli K-12 is regulated by a heat shock promoter, which may indicate that the heptose domain of LPS, at least in E. coli K-12, may be required for growth at elevated temperatures (Schnaitman and Klena, 1993; Raetz, 1996). The long central waaQ operon contains genes necessary for the biosynthesis of the outer core OS and for modification and/or decoration of core OS (1[link]Table 1; Figs 2 and 3). The waaA transcript contains the structural gene (waaA) for the bifunctional Kdo transferase (Raetz, 1996) and a gene (designated 18 k) encoding a polypeptide of unknown function. The waaQ and waaA operons are divergently transcribed and are separated by 400–500 bp of intervening DNA. A 5′ mRNA start site has been mapped for each operon, although concensus promoter sequences are not evident. A 39 bp sequence, which is designated JUMPStart (Just Upstream of Many Polysaccharide-associated gene Starts), is found in the untranslated region upstream of waaQ in the Salmonella and all of the E. coli core OS biosynthesis gene clusters. Part, or all, of this sequence also occurs in regulatory regions for several operons associated with cell-surface or exported structures, suggesting a common and potentially co-ordinate regulation. These include the hly (haemolysin synthesis and export), kps (E. coli group II capsule synthesis and export), tra (F-pilus synthesis and export) and most intriguingly the rfb (LPS O-antigen biosynthesis) operons. JUMPStart includes the conserved 8 bp region known as ops (operon polarity suppressor) that, together with RfaH (a NusG homologue), is required for operon polarity suppression. RfaH may interact with Rho and RNA polymerase at the transcribed ops element to form a more processive transcription complex that is capable of extending through certain termination elements. For a more detailed discussion of the features of RfaH/ops-regulated LPS biosynthesis operons, we refer the reader to Marolda and Valvano (1998) and references therein. In the case of LPS core OS biosynthesis, Salmonella and E. coli K-12 mutants deficient in RfaH produce truncated LPS molecules (reviewed in Schnaitman and Klena, 1993). Most, if not all, of the genes of the waaQ operon are regulated by RfaH, but the effects of an rfaH mutation are particularly evident in promoter-distal genes. Polar mutations argue against secondary promoters within the waaQ operon, and the presence of multiple termination sequences throughout this operon (Schnaitman and Klena, 1993) explains the requirement for RfaH/ops or an alternative antitermination system in core OS biosynthesis.

Although the general organizational features are conserved in E. coli K-12 and Salmonella, there are differences in the gene content of operons from different E. coli core types (Fig. 3). However, one striking difference is that the clinically predominant R1 and R4 types have their waaL gene as the last gene of the waaQ operon, rather than in their ‘usual’ position in the gmhD operon. It is currently unclear whether this has any significance in the biology of these bacteria. Other differences in individual open reading frames will be discussed below.

Conservation in enzymes for inner core OS assembly

  1. Top of page
  2. Abstract
  3. Introduction
  4. Structure and function of core OS
  5. Genetic organization and regulation of the major LPS core OS biosynthesis operons
  6. Conservation in enzymes for inner core OS assembly
  7. Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases
  8. Assembly of the outer core OS — HexI substitution
  9. Assembly of the outer core OS — HexIII substitution
  10. A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4
  11. A ‘missing’ transferase for assembly of the E. coli R3 core OS
  12. What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?
  13. Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?
  14. Concluding remarks
  15. Acknowledgements
  16. References

As might be predicted by the common structure of the inner core OS in the Enterobacteriaceae (Fig. 2), enzymes responsible for its formation are highly conserved. In E. coli, the Kdo transferase (WaaA, formerly KdtA) is bifunctional, adding two Kdo residues (Belunis and Raetz, 1992). The enzyme responsible for the non-stoichiometric addition of the third Kdo residue is currently unidentified. WaaA enzymes are highly conserved, and the amino acid sequences of the six homologues from E. coli and Salmonella share more than 96% identity.

Functional analysis of the biosynthesis of the heptose region of the core OS has been limited by the lack of purified precursor (ADP-L-glycero-D-manno-heptose). Recent progress has been made by using the alternative substrate ADP-mannose, and this approach has identified WaaC as heptosyltransferase I (Kadrmas and Raetz, 1998). The WaaC enzymes from E. coli K-12 and Salmonella are functionally interchangeable (Klena et al., 1992; Schnaitman and Klena, 1993), as might be expected because the WaaC proteins from Salmonella and the five E. coli core types show greater than 84% total similarity. The amino acid sequence similarity of the predicted WaaC proteins only begins to break down in the extreme C-terminal segment of the protein. Significantly, the 3′ end of the waaC gene lies adjacent to the more variable waaL gene (see below).

Although the waaF gene is implicated in the transfer of the HepII residue onto the LPS core OS molecule (Schnaitman and Klena, 1993), direct demonstration of this enzymatic activity is lacking. The Salmonella and E. coli K-12 waaF genes have both been cloned and sequenced, and the deduced polypeptide products share 95% similarity (Sirisena et al., 1994).

The core OS structures of E. coli K-12 and Salmonella contain a stoichiometric HepIII residue (Holst and Brade, 1992), and its addition requires the product of waaQ (Yethon et al., 1998). A non-polar waaQ::aacC1 insertion in an E. coli R1 core-type strain yields LPS totally devoid of HepIII. Sequence data are consistent with the assignment of waaQ as heptosyltransferase III. In E. coli K-12, WaaF shares ≈49% similarity with WaaC and 32% similarity with WaaQ, which reflects their similar enzymatic function and interaction with ADP-heptose. Most conserved residues occur in the C-terminal half of the WaaF and WaaQ proteins from E. coli K-12. The predicted WaaQ proteins from Salmonella LT2 and E. coli R1, R2, R3 and R4 share 75.6–99.6% total similarity.

It is generally accepted that WaaP is involved in decoration of the inner core OS with phosphoryl substituents, although other functions have been ascribed to WaaP in the literature. Deduced WaaP proteins of E. coli and Salmonella are highly conserved, with values for total similarity varying from 89.8% to 99.6%, and the WaaP proteins from Salmonella and E. coli R1 are functionally interchangeable (J. A. Yethon et al., unpublished), as are those from E. coli K-12 and Salmonella (Parker et al., 1992). Although basic BLASTP searches of the databases fail to identify significant similarities between WaaP and other proteins, PSI (Position-Specific Iterated) BLAST searches reveal that WaaP shares similarity with a vast number of kinases. The PSI-BLAST program aids in the identification of proteins with weak, but biologically relevant, sequence similarities in key positions throughout the proteins. Much of the chemical characterization of the waaP phenotype has been carried out using mutants of S. enterica serovars Typhimurium and Minnesota, and evaluation of the data is perhaps complicated by the fact that the prototype mutants are derived by chemical mutagenesis. As such, they potentially carry additional and unrecognized defects. Defined non-polar waaP ::aacC1 mutations in the E. coli R1 (Yethon et al., 1998) and Salmonella (J. A. Yethon and C. Whitfield, unpublished) core OS prototypes show a ‘deep-rough’ phenotype with respect to sensitivity to hydrophobic agents, despite the fact that the core OS is ‘capped’ with O-PS (J. A. Yethon et al., unpublished). However, the efficiency of ‘capping’ is reduced in the Salmonella and E. coli R1 waaP ::aacC1 mutants compared with wild type. This finding is in agreement with some, but not all, previously reported Salmonella waaP mutants. Given the similarities in protein sequences and inner core OS structures, radical differences between the biosynthesis of the heptose region in Salmonella and E. coli would be surprising. The inner core OS heptose region the E. coli R1 waaP ::aacC1 mutant is completely devoid of phosphate and 2-aminoethyl diphosphate (Yethon et al., 1998) as is the equivalent mutant in Salmonella (J. A. Yethon et al., unpublished).

Each of the waaQ operons from E. coli and Salmonella encode a well-conserved WaaY homologue (67.4–99.5% total similarity). The relative position of waaY is conserved in all but E. coli R3 (Fig. 3). Database searches reveal that WaaY shares no similarity with known glycosyltransferases and consistent with this, the carbohydrate structure of an E. coli R1 waaY ::aacC1 mutant is indistinguishable from the wild-type structure. PSI-BLAST searches of the databases reveal that WaaY, like WaaP, possesses characteristic features of kinase proteins. Moreover, WaaY proteins share limited similarity with WaaP proteins. Analysis of the phosphorylation of the R1 waaY ::aacC1 mutant revealed that while HepI remained phosphorylated with both phosphate and pyrophosphorylethanolamine, phosphoryl substituents on HepII were absent (Yethon et al., 1998). This would suggest that WaaY is the enzyme responsible for HepII phosphorylation and that the activity of the WaaP enzyme is a prerequisite to WaaY function.

Several important questions still remain about the phosphorylation of the heptose region. (i) What is the substrate donor of phosphoryl substituents? Soluble enzyme fractions from strains of S. enterica serovars Typhimurium and Minnesota both catalyse the transfer of [γ-32P]-ATP into S. enterica serovar Minnesota acceptor LPS (Schnaitman and Klena, 1993). There is also preliminary evidence (Hasin and Kennedy, 1982) that supports phosphatidylethanolamine as the direct donor of 2-aminoethyl phosphate (PEtN) residues into the inner core OS, but this would require prior addition of a phosphate to give PPEtN (Fig. 2). Does the diversity of substituents (P and PPEtN; Fig. 2) then arise by incomplete substitution or by a processing reaction as suggested by others (Schnaitman and Klena, 1993)? (ii) Is it possible that the same enzyme system that modifies Hep residues modifies KdoII as well? Modification of KdoII with PEtN has not been investigated in defined waaP or waaY mutants. (iii) There is no direct evidence that identifies WaaP or WaaY as a phosphorylating enzyme per se. Could it be that WaaP and WaaY are enzymes in a pathway including other, as yet unidentified, players?

Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases

  1. Top of page
  2. Abstract
  3. Introduction
  4. Structure and function of core OS
  5. Genetic organization and regulation of the major LPS core OS biosynthesis operons
  6. Conservation in enzymes for inner core OS assembly
  7. Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases
  8. Assembly of the outer core OS — HexI substitution
  9. Assembly of the outer core OS — HexIII substitution
  10. A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4
  11. A ‘missing’ transferase for assembly of the E. coli R3 core OS
  12. What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?
  13. Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?
  14. Concluding remarks
  15. Acknowledgements
  16. References

Detailed analysis of many of the outer core OS glycosyltransferases is limited by the lack of biochemical data derived from purified enzymes and structurally defined acceptors and/or reaction products. In fact, demonstration of enzyme activity is limited to WaaG in E. coli K-12 and WaaB and WaaI in Salmonella (reviewed in Schnaitman and Klena, 1993). In general, gene assignments result from chemical characterization of mutant LPS and heterologous complementation studies with cloned genes. This is not always possible or interpretable because of variations in outer core OS structure. The E. coli R1 prototype is the only system for which precise non-polar insertions have been made in each of the outer core glycosyltransferase genes and for which the resulting core OS structures have been determined (Heinrichs et al., 1998a). Using the expanded list of known enzymes and sequence features discussed below, assignment of HexII and HexIII transferases becomes possible. However, the existing nomenclature for the HexII and HexIII transferases is illogical. For example, the gene designations are the same in E. coli K-12 and Salmonella, despite the fact that the enzymes have different specificities. For this reason, we have proposed that the genes for novel specificities be given unique designations (Table 1). A new name has also been given to the E. coli K-12 gene that is responsible for the unique HexIII side-branch substitution for the same reason. Salmonella nomenclature is accorded priority because it was in this organism that the mutations were first characterized.

In all of the E. coli and Salmonella core OS, HexI is a Glc residue. Biochemical data, LPS structure and heterologous gene complementation experiments identify WaaG as the UDP-glucose:(heptosyl) lipopolysaccharide α1,3-glucosyltransferase in E. coli K-12 and Salmonella (reviewed in Schnaitman and Klena, 1993). As expected, the six known homologues of the HexI glycosyltransferase, WaaG, are highly conserved (> 85% identity).

There is biochemical evidence supporting the assignment of waaI as the structural gene for the HexII transferase (UDP-galactose:(glucosyl) lipopolysaccharide α1,3-galactosyltransferase) in Salmonella (Schnaitman, and Klena, 1993). Although WaaJ is accepted as the HexIII transferase, the existing data are equivocal. Genes in similar positions to waaI and waaJ in other clusters encode related proteins (Fig. 3). Non-polar insertion mutations in the genes for the HexII and HexIII transferases of E. coli R1 (waaO and waaT respectively) give the predicted truncated core OS structures and confirm these assignments (Heinrichs et al., 1998a; Fig. 3).

Relationships between the predicted HexII and HexIII transferases of E. coli K-12 and Salmonella were initially obscured by problems with the published DNA sequence, but resequencing of the waaIJ region in Salmonella identified genes with similar sizes to the equivalent genes in E. coli K-12 (Heinrichs et al., 1998b). It is now apparent that all of the putative HexII and HexIII transferases share significant similarities (collectively 15.4% identity, 41.3% total similarity), as might be expected because all bind a nucleotide diphosphohexose precursor and transfer to a hexose residue in the acceptor. When grouped separately, the HexII transferases are closer to themselves (35.2% identity; 68.3% similarity) than they are to the HexIII transferases (27.2%; 50.8%). Predictably, when transferases with identical acceptor and donor specificities are considered, these values range from as low as 42.6% identity (57.5% similarity) up to as high as 99.7% identity. Most of the HexII and HexIII transferases are similar in length (≈330–340 residues). The exception is the HexIII transferase of E. coli R3 (see below). All of the HexII and HexIII transferases contain characteristic motifs (Fig. 4) that are absent in WaaG proteins. Some of the core OS α-glycosyltransferases [including WaaG, WaaB and WaaK (Heinrichs et al., 1998b), as well as WaaA, WaaZ and WaaD, see below] contain the motif E–(X7)–E. This motif was originally thought to be characteristic of α-glycosyltransferases (Geremia et al., 1996), but this conclusion is complicated by the discovery of the motif in virtually all of the core OS biosynthesis enzymes, including WaaP and WaaY proteins, which are clearly not glycosyltransferases, and WaaX and WaaV, which are β-glycosyltransferases (see below). Putative heptosyltransferases (WaaC, WaaF and WaaU) also have copies of the motif and its role remains an interesting but unanswered question. Clearly, care should be taken in the interpretation of this motif.

image

Figure 4. . Consensus features of the WaaIJ family of α-glycosyltransferases involved in outer core OS biosynthesis. These motifs were identified by multiple sequence alignments (Heinrichs et al., 1998).

Download figure to PowerPoint

Assembly of the outer core OS — HexI substitution

  1. Top of page
  2. Abstract
  3. Introduction
  4. Structure and function of core OS
  5. Genetic organization and regulation of the major LPS core OS biosynthesis operons
  6. Conservation in enzymes for inner core OS assembly
  7. Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases
  8. Assembly of the outer core OS — HexI substitution
  9. Assembly of the outer core OS — HexIII substitution
  10. A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4
  11. A ‘missing’ transferase for assembly of the E. coli R3 core OS
  12. What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?
  13. Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?
  14. Concluding remarks
  15. Acknowledgements
  16. References

HexI substitution in the core OS of E. coli K-12, R2 and Salmonella occurs in the form of α1,6-linked galactose. Biochemical data identify WaaB as the UDP-galactose: (glucosyl)lipopolysaccharide α1,6-galactosyltransferase. The waaB mutation in Salmonella can be complemented with waaB genes from Salmonella and E. coli K-12 (Schnaitman and Klena, 1993). The three known WaaB proteins are highly conserved (75.5–95.3% total similarity) and share a similar location in their respective gene clusters (Fig. 3).

Assembly of the outer core OS — HexIII substitution

  1. Top of page
  2. Abstract
  3. Introduction
  4. Structure and function of core OS
  5. Genetic organization and regulation of the major LPS core OS biosynthesis operons
  6. Conservation in enzymes for inner core OS assembly
  7. Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases
  8. Assembly of the outer core OS — HexI substitution
  9. Assembly of the outer core OS — HexIII substitution
  10. A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4
  11. A ‘missing’ transferase for assembly of the E. coli R3 core OS
  12. What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?
  13. Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?
  14. Concluding remarks
  15. Acknowledgements
  16. References

The HexIII (Glc) residues of the Salmonella and E. coli R2 core OS are substituted with an α1,2-linked GlcNAc residue (Fig. 3). This substitution is non-stoichiometric in the E. coli R2 core OS (Heinrichs et al., 1998b). As might be expected, the WaaK enzymes from Salmonella and E. coli R2 are closely related (73.7% identity, 83.2% total similarity) and are capable of cross-complementation. The GlcNAc transferase activity has been confirmed directly for the E. coli R2 homologue (Heinrichs et al., 1998b). It is this gene that is presumably replaced to yield the novel core OS structure in S. enterica serovar Arizonae IIIa O62 (Fig. 3).

The corresponding HexIII residue in E. coli K-12 is partially substituted at the C-6 position with the terminal disaccharide β-D-GlcNAc-(1 [RIGHTWARDS ARROW] 7)-L-α-D-Hep (Fig. 3) (Brade et al., 1996). One would not expect the glycosyltransferase catalysing the HexIII substitution to be conserved between Salmonella (or E. coli R2; WaaK) and E. coli K-12 (WaaU) and this is indeed the case. BLASTP searches of the databases identify regions of local similarity shared by WaaU and a variety of heptosyltransferases (Heinrichs et al., 1998b), suggesting that WaaU is probably involved in the addition of the HepIV substitution in E. coli K-12. However, the possibility that the WaaU protein is involved in the substitution of the HepIV residue with a β-linked GlcNAc residue should not be ruled out, and we have illustrated this in Fig. 3. Clearly, more work is required to resolve which genes are involved in the attachment of the terminal sugars of the K-12 core OS. The HexIII position in E. coli R1 and R4 (Gal) is substituted by α1,2-Gal, and the glycosyltransferase involved has recently been determined to be WaaW by structural analysis of a waaW ::aacC1 mutant of R1. The WaaW proteins of R1 and R4 are 93% identical.

A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4

  1. Top of page
  2. Abstract
  3. Introduction
  4. Structure and function of core OS
  5. Genetic organization and regulation of the major LPS core OS biosynthesis operons
  6. Conservation in enzymes for inner core OS assembly
  7. Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases
  8. Assembly of the outer core OS — HexI substitution
  9. Assembly of the outer core OS — HexIII substitution
  10. A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4
  11. A ‘missing’ transferase for assembly of the E. coli R3 core OS
  12. What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?
  13. Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?
  14. Concluding remarks
  15. Acknowledgements
  16. References

The R1 and R4 core OS are identical with respect to backbone structure and the α1,2-Gal substituent on HexIII and, as would be expected, the relevant gene products all show > 95% identity. These two structures differ in the β-linked substituent at HexII (Fig. 3). The predicted products of waaV in E. coli R1 and waaX in E. coli R4 show only low levels of similarity (20.2% identity; 36.0% total similarity). However, they do show characteristic consensus sequence features and are members of two distinct families of β-glycosyltransferases (Heinrichs et al., 1998a). The core OS structure in an E. coli R1 waaV ::aacC1 mutant and those in the mutant expressing either waaV or waaX have been determined. These analyses definitively identified waaV and waaX as the structural genes for the β-glycosyltransferases that define the R1 and R4 outer core OS structure (Heinrichs et al., 1998a). The relatively low levels of similarity in the HexII-substituting β-glycosyltransferases is a marked contrast to the high levels of conservation in the rest of the R1 and R4 core OS biosynthesis genes. These data suggest that the structural diversity between R1 and R4 probably results from acquisition and replacement of one open reading frame, rather than by sequence drift. Interestingly, the addition of the β-linked substituent at HexII occurs as the last step in core OS assembly and is eliminated in a mutant lacking the HexIII substitution (Heinrichs et al., 1998a).

A ‘missing’ transferase for assembly of the E. coli R3 core OS

  1. Top of page
  2. Abstract
  3. Introduction
  4. Structure and function of core OS
  5. Genetic organization and regulation of the major LPS core OS biosynthesis operons
  6. Conservation in enzymes for inner core OS assembly
  7. Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases
  8. Assembly of the outer core OS — HexI substitution
  9. Assembly of the outer core OS — HexIII substitution
  10. A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4
  11. A ‘missing’ transferase for assembly of the E. coli R3 core OS
  12. What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?
  13. Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?
  14. Concluding remarks
  15. Acknowledgements
  16. References

The E. coli R3 core OS is of some interest because it is found in E. coli J5, which provides the LPS antigen for many significant (and sometimes controversial) immunological studies involving poly-and monoclonal antibodies. Sequence similarities in the predicted proteins readily facilitate identification of the HexII and HexIII transferases in E. coli R3. However, despite the fact that the E. coli R3 outer core OS has two side-group substitutions, only one additional transferase gene is found in the cluster (Fig. 3). The waaD gene occupies a position normally expected for an outer core OS function, and its product shares ≈70% total similarity (57% identity) with the GlcNAc transferase (WaaK) common to Salmonella and E. coli R2, strongly suggesting that WaaD forms the HexII substitution in E. coli R3. However, this interpretation is complicated by the presence of an additional partial GlcNAc modification of HepIII, which, to date, is unique to E. coli R3 among the E. coli core OS. The gene required for the additional outer core OS substituent in E. coli R3 (an α1,2-Glc substitution of HexIII) remains unidentified. There are two potential interpretations of these observations. One is clearly the existence of an additional gene located outside the waa cluster. Alternatively, one of the already identified enzymes may play a more complicated role. Interestingly, sequence data from the E. coli R3 prototype strain (F653) and from clinical R3 isolates revealed that the predicted waaJ gene product is truncated by ≈150 amino acid residues at the N-terminus compared with other HexII and HexIII transferases (D. E. Heinrichs and C. Whitfield, unpublished). This may result from a rearrangement leading to a reversal in the relative orders of waaJ and waaY (Fig. 3). The truncated R3 WaaJ protein lacks the first motif identified in the WaaIJ family of proteins (see Fig. 4), and it is conceivable that this change leads to a processive glycosyltransferase enzyme which adds two α1,2-linked Glc residues, instead of the typical single residue.

What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?

  1. Top of page
  2. Abstract
  3. Introduction
  4. Structure and function of core OS
  5. Genetic organization and regulation of the major LPS core OS biosynthesis operons
  6. Conservation in enzymes for inner core OS assembly
  7. Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases
  8. Assembly of the outer core OS — HexI substitution
  9. Assembly of the outer core OS — HexIII substitution
  10. A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4
  11. A ‘missing’ transferase for assembly of the E. coli R3 core OS
  12. What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?
  13. Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?
  14. Concluding remarks
  15. Acknowledgements
  16. References

Current data predict that the ligation reaction, which joins newly synthesized O-PS to lipid A core, occurs at the periplasmic face of the plasma membrane (reviewed in Whitfield et al., 1997). The mechanism involved has not been resolved and the waaL gene product is the only enzyme known to be required for ligation. The ligation-deficient phenotype of a waaL mutant was first established in Salmonella, then later in E. coli K-12, and has been confirmed for chromosomal waaL ::aacC1 insertion mutants in both E. coli R2 (Heinrichs et al., 1998b) and E. coli R1 (Heinrichs et al., 1998a). The ligase enzyme, or enzyme complex, is essentially a glycosyltransferase with a complex substrate requirement. However, the WaaL homologues contain none of the currently known glycosyltransferase features. This is not unexpected because their substrates are not nucleotide diphosphosugars. The WaaL primary sequences offer no obvious clues about function. In multiple alignments, the six known WaaL proteins collectively share only low levels of similarity in their primary sequences, but their higher order structures have common features. All of the WaaL proteins are predicted to be integral membrane proteins with eight or more membrane-spanning domains and their hydropathy profiles are virtually identical; they share hydrophobic domains of similar length and distribution. In contrast, outer core OS glycosyltransferases are all predicted to be peripheral membrane proteins. These features may reflect differences in their substrate specificities (i.e. nucleotide diphosphosugars for glycosyltransferases or undecaprenol-linked oligosaccharides for ligases).

The WaaL protein probably functions as part of a complex that involves highly specific interactions with O-PS biosynthetic intermediates on undecaprenol-P-P and lipid A-core acceptor. As part of this complex, WaaL proteins would be involved in specific protein–carbohydrate interactions, and this may explain the diversity in WaaL primary sequences. Ligase enzymes show no obvious specificity for the structure of the ligated polysaccharide, possibly because these may be presented for ligation in a common undecaprenol-P-P-linked form. This has led to the widespread success in cloning O-PS biosynthesis gene clusters from heterologous species and expressing the O-PS on the surface of E. coli K-12. The known core types of E. coli and Salmonella are all capable of being efficiently ‘capped’ by the same prototype O-PS (D-galactan I of Klebsiella pneumoniae O1) (D. E. Heinrichs and C. Whitfield, unpublished). Furthermore, E. coli WaaL is required for the attachment of type I capsular polysaccharide and one form of enterobacterial common antigen (ECA) to the same core OS (for further discussion see Whitfield et al., 1997).

The nature of the acceptor molecule is certainly important for ligation. Indeed, ligation activity is eliminated in an E. coli R2 waaK ::aacC1 mutant, in a Salmonella waaK mutant, and in an E. coli K-12 derivative lacking the equivalent HexIII substitution (MacLachlan et al., 1991; Klena et al., 1992; Heinrichs et al., 1998b). The HexIII side-branch residue is present in R2 core OS carrying ligated O-PS (Gamian et al., 1992), although the possibility that the HexIII side-branch residue may be trimmed away during the ligation step in some other core OS-type acceptor molecules has been suggested for S. enterica serovar Arizonae IIIa (discussed in Olsthoorn et al., 1998). The finding that partial ligation activity results from complementation of the E. coli K-12 waaL mutant with the R2 waaL gene indicates that the structural requirement for a terminal side-group can be fulfilled by different residues and argues against a role for the precise core OS backbone structure in determining acceptor specificity (Heinrichs et al., 1998b).

Attempts to establish relationships in WaaL sequences based on potential acceptor structures do not give a consistent pattern. The E. coli R2 and Salmonella WaaL homologues link an O-PS to the same substituted Glc (HexIII) residue (Fig. 3) and the proteins are closely related (65.8% identity, 81.1% total similarity) and functionally interchangeable (Heinrichs et al., 1998b). If the attachment site on core OS is conserved, the R3 and K-12 WaaL enzymes should attach O-PS to the same HexIII residue. However, the R3 homologue resembles those from R2 and Salmonella (50–66% identity) and shares minimal identity with the K-12 enzyme. In terms of function, the waaL gene of E. coli K-12 shows no complementation of the Salmonella waaL defect (Klena et al., 1992), suggesting that the Salmonella core OS is not a suitable acceptor for the K-12 ligase. In the converse experiments, E. coli K-12 and Salmonella core OS serve as acceptors with varying efficiencies for the E. coli R2 ligase (Heinrichs et al., 1998b). These data reflect the complex acceptor requirements for ligase activity. Recent studies in our laboratory have identified the HexII side-branch substituent (β1,3-linked Glc) as the attachment site for O-PS to an R1-type core OS. E. coli R1 waaV ::aacC1 mutants (lacking only the β1,3-linked Glc residue in the core OS) fail to ligate O-PS to lipid A core (Heinrichs et al., 1998a). This represents a fundamental difference from the O-PS attachment site of the core OS of Salmonella and E. coli R2. Although the E. coli R1 and R4 WaaL homologues are related (33.1% identity; 54% total similarity), the values are low given that (i) most of the Waa proteins in the E. coli R1 and R4 systems are virtually identical; and (ii) the E. coli R1 and R4 systems share the distinction of having the waaL gene as the last gene of the waaQ operon, rather than the terminus of the gmhD operon (Fig. 3). Differences between these proteins are expected, given that the R4 core OS does not contain a β1,3-linked Glc residue in its core OS. It is interesting to speculate that the linkage of O-PS to an R4-type core OS is the β1,4-linked Gal residue. The location of this residue as well as its anomeric configuration are features in common with the attachment site in R1 core OS molecules. Similarities within the R1 and R4 WaaL proteins may be attributable to these conserved features. Efforts to understand the broader structural requirements and mechanism of ligation are underway in this laboratory.

In Salmonella and E. coli K-12, putative functions have been assigned to many of the core biosynthesis genes based on their ability to complement known core defects, or by similarity to other genes of known function. Such analyses have not been informative for waaS and waaZ of E. coli K-12. Although similar in size and located at an identical position, the waaS and wabA genes from E. coli K-12 and R2, respectively, predict proteins with negligible similarity. The R2 WabA protein contains most of the features characteristic of the HexII, HexIII and related transferases (Fig. 4) and is therefore predicted to use a UDP-hexose precursor. The K-12 WaaS protein does not contain any identifiable glycosyltransferase motifs and the sequence variation in the predicted WaaS and WabA proteins could reflect different transferase specificities. Significantly, the E. coli K-12 and R2 core OS are distinguished by the presence of non-stoichiometric α1,5-Rha and α1,7-Gal substituents on KdoII respectively (Fig. 2). We believe that the WabA and WaaS proteins are involved in these modifications and structural studies are underway with LPS from an E. coli R2 wabA::aacC1 mutant to test this hypothesis.

The predicted WaaZ proteins, identified in Salmonella, and E. coli K-12 and R2 share 81.1–91.4% total similarity. The phenotype of a precise waaZ mutation has not yet been clearly established but the conservation in WaaZ, together with its presence in Salmonella argue against it being involved in variable substitutions of the core OS structure. WaaZ has no sequence features that provide any additional clues to its function. The only unassigned function in core OS assembly is the addition of KdoIII and it will be interesting to see if WaaZ plays any role in this.

Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?

  1. Top of page
  2. Abstract
  3. Introduction
  4. Structure and function of core OS
  5. Genetic organization and regulation of the major LPS core OS biosynthesis operons
  6. Conservation in enzymes for inner core OS assembly
  7. Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases
  8. Assembly of the outer core OS — HexI substitution
  9. Assembly of the outer core OS — HexIII substitution
  10. A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4
  11. A ‘missing’ transferase for assembly of the E. coli R3 core OS
  12. What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?
  13. Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?
  14. Concluding remarks
  15. Acknowledgements
  16. References

It has been suggested that WaaS and WaaZ together with WaaQ (heptosyltransferase III) play an integral role in the production of two forms of LPS in E. coli K-12: the classical LPS form that contains O-PS and an ‘LOS form’ distinct from R-LPS and that is destined to terminate without O-PS addition (Schnaitman and Klena, 1993). To date, this interesting proposal is based primarily on the banding patterns of E. coli K-12 R-LPS derivatives in SDS–PAGE analysis, and no firm conclusions can be made because of the lack of chemical and enzymatic data.

One might expect that an alternative system for the production of an ‘LOS form’ of LPS would not be confined to E. coli K-12. E. coli R2 and Salmonella are the only other core types that possess a waaZ homologue, and only E. coli R2 has an additional gene (wabA) in an equivalent position to waaS (Fig. 3). However, strains with the most clinically prevalent core types (E. coli R1 and R3) are missing the waaS, wabA and/or waaZ. They therefore lack these components of the putative ‘LOS pathway’. This raises important questions concerning the biological role of the alternative pathway. It is possible that an ‘LOS’ system is confined to E. coli K-12 and R2, although the existence of homologues of these genes outside the core OS biosynthesis gene cluster cannot be ruled out. Nevertheless, the possibility of an ‘LOS pathway’ remains an intriguing hypothesis and warrants further investigation.

Concluding remarks

  1. Top of page
  2. Abstract
  3. Introduction
  4. Structure and function of core OS
  5. Genetic organization and regulation of the major LPS core OS biosynthesis operons
  6. Conservation in enzymes for inner core OS assembly
  7. Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases
  8. Assembly of the outer core OS — HexI substitution
  9. Assembly of the outer core OS — HexIII substitution
  10. A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4
  11. A ‘missing’ transferase for assembly of the E. coli R3 core OS
  12. What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?
  13. Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?
  14. Concluding remarks
  15. Acknowledgements
  16. References

It should be apparent from the above discussion that knowledge of the enzymes involved in core OS assembly has progressed substantially in recent years. Although we can now assign roles to many of the gene products, their precise mechanisms of action remain elusive in the absence of biochemical data. These limitations should certainly be considered when the Enterobacteriaceae systems are used as a basis for interpretation of data from other bacterial species. Among the unanswered questions are the following. (i) Is there a mechanism that diverts specific lipid A-core acceptors out of an O-PS substitution pathway? If so, how is this regulated and what is its impact on pathogenicity? (ii) How does WaaL operate, and are there other unidentified components in the ligation reaction? (iii) How are the critical phosphoryl derivatives put in place in the core OS? (iv) How are the individual enzymes for core assembly organized into a functional complex? (v) How important are specific protein–protein interactions, and to what extent would the system tolerate heterologous enzymes? The reagents, techniques, and sequences of many waa homologues are now available to address these important and unresolved issues in the assembly of the bacterial cell envelope. Despite the open questions that remain, E. coli and Salmonella continue to provide a solid foundation for further work in other bacteria.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Structure and function of core OS
  5. Genetic organization and regulation of the major LPS core OS biosynthesis operons
  6. Conservation in enzymes for inner core OS assembly
  7. Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases
  8. Assembly of the outer core OS — HexI substitution
  9. Assembly of the outer core OS — HexIII substitution
  10. A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4
  11. A ‘missing’ transferase for assembly of the E. coli R3 core OS
  12. What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?
  13. Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?
  14. Concluding remarks
  15. Acknowledgements
  16. References

This work was supported by funding from the Natural Sciences and Engineering Research Council (NSERC) and the Canadian Bacterial Diseases Network (NCE program) awarded to C.W. D.E.H. gratefully acknowledges receipt of postdoctoral fellowships from NSERC and the Medical Research Council. J.A.Y. is the recipient of an NSERC postgraduate scholarship. Owing to space limitations, we are unable to fully reference many individual studies and, where possible, have cited previous reviews where that work has been discussed. We apologize to those whose contribution to this area was not discussed in more detail.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Structure and function of core OS
  5. Genetic organization and regulation of the major LPS core OS biosynthesis operons
  6. Conservation in enzymes for inner core OS assembly
  7. Assembly of the outer core OS — HexI, HexII and HexIII glycosyltransferases
  8. Assembly of the outer core OS — HexI substitution
  9. Assembly of the outer core OS — HexIII substitution
  10. A single glycosyltransferase change defines core structure specificity in E. coli R1 and R4
  11. A ‘missing’ transferase for assembly of the E. coli R3 core OS
  12. What is the mechanism involved in ligation of O-PS to lipid A-core acceptor?
  13. Is there a separable ‘LOS pathway’ in the Enterobacteriaceae?
  14. Concluding remarks
  15. Acknowledgements
  16. References
  • 1
    Appelmelk, B.J., An, Y.-Q., Hekker, T.A.M., Thijs, L.G., MacLaren, D.M., De Graaf, J. (1994) Frequencies of lipopolysaccharide core types in Escherichia coli strains from bacteraemic patients. Microbiology 140: 11191124.
  • 2
    Bauer, M.E. & Welch, R.A. (1997) Pleiotropic effects of a mutation in rfaC on Escherichia coli hemolysin. Infect Immun 65: 22182224.
  • 3
    Belunis, C.J. & Raetz, C.R.H. (1992) Biosynthesis of endotoxins: purification and catalytic properties of 3-deoxy-D-manno-octulosonic acid transferase from Escherichia coli. J Biol Chem 267: 99889997.
  • 4
    Brade, L., Grimmecke, H.-D., Holst, O., Brabetz, W., Zamojski, A., Brade, H. (1996) Specificity of monoclonal antibodies against Escherichia coli K-12 lipopolysaccharide. J Endotox Res 3: 3947.
  • 5
    Di Padova, F.E., Brade, H., Barclay, G.R., Poxton, I.R., Liehl, E., Schuetze, E., et al (1993) A broadly cross-protective monoclonal antibody binding to Escherichia coli and Salmonella lipopolysaccharides. Infect Immun 61: 38633872.
  • 6
    Gamian, A., Romanowska, E., Ulrich, J., Defaye, J. (1992) The structure of the sialic acid-containing Escherichia coli O104 O-specific polysaccharide and its linkage to the core region in lipopolysaccharide. Carbohydr Res 236: 195208.
  • 7
    Geremia, R.A., Petroni, E.A., Ielpi, L., Henrissat, B. (1996) Towards a classification of glycosyltransferases based on amino acid sequence similarities: prokaryotic α-mannosyltransferases. Biochem J 318: 133138.
  • 8
    Gibb, A.P., Barclay, G.R., Poxton, I.R., Di Padova, F. (1992) Frequencies of lipopolysaccharide core types among clinical isolates of Escherichia coli defined with monoclonal antibodies. J Infect Dis 166: 10511057.
  • 9
    Hammond, S.M., Lambert, P.A., Rycroft, A.N. (1984) The cell envelope in bacterial disease. In The Bacterial Cell Surface. Hammond, S.M., Lambert, P.A. and Rycroft, A.N. (eds). Washington, DC: Kapitan Szabo, pp. 147193.
  • 10
    Hasin, M. & Kennedy, E.P. (1982) Role of phosphatidylethanolamine in the biosynthesis of pyrophosphoethanolamine residues in the lipopolysaccharide of Escherichia coli. J Biol Chem 257: 1247512477.
  • 11
    Heinrichs, D.E., Yethon, J.A., Amor, P.A., Whitfield, C. (1998a) The assembly system for the outer core portion of R1 and R4-type lipopolysaccharides of Escherichia coli. The R1 core-specific β-glucosyltransferase provides a novel attachment site for O polysaccharides. J Biol Chem (in press).
  • 12
    Heinrichs, D.E., Monteiro, M.A., Perry, M.B., Whitfield, C. (1998b) The assembly system for the lipopolysaccharide R2 core-type of Escherichia coli is a hybrid of those found in Escherichia coli K-12 and Salmonella enterica. Structure and function of WaaK and WaaL homologs. J Biol Chem 273: 88498859.
  • 13
    Holst, O. & Brade, H. (1992) Chemical structure of the core region of lipopolysaccharides. In Bacterial Endotoxic Lipopolysaccharides. Morrison, D.C. and Ryan, J.L. (eds). Boca Raton, FL: CRC Press, pp. 134170.
  • 14
    Jacques, M. (1996) Role of lipo-oligosaccharides and lipopolysaccharides in bacterial adherence. Trends Microbiol 4: 408410.
  • 15
    Kadrmas, J.L. & Raetz, C.R.H. (1998) Enzymatic properties of lipopolysaccharide in Escherichia coli. Purification and properties of heptosyltransferase I. J Biol Chem 273: 27992807.
  • 16
    Klena, J.D., Ashford, R.S., II, Schnaitman, C.A. (1992) Role of Escherichia coli K-12 rfa genes and the rfp gene of Shigella dysenteriae 1 in generation of lipopolysaccharide core heterogeneity and attachment of O antigen. J Bacteriol 174: 72977307.
  • 17
    MacLachlan, P.R., Kadam, S.K., Sanderson, K.E. (1991) Cloning, characterization, and DNA sequence of the rfaLK region for lipopolysaccharide synthesis in Salmonella typhimurium LT2. J Bacteriol 173: 71517163.
  • 18
    Mäkelä, P.H. & Stocker, B.A.D. (1984) Genetics of lipopolysaccharide. In Handbook of Endotoxin. Rietschel, E.T. (ed.). Amsterdam: Elsevier Science Publishers, pp. 59137.
  • 19
    Marolda, C.L. & Valvano, M.A. (1998) The promoter region of the Escherichia coli O7-specific lipopolysaccharide gene cluster: Structural and functional characterization of an upstream untranslated mRNA sequence. J Bacteriol 180: 30703079.
  • 20
    Olsthoorn, M.M.A., Petersen, B.O., Schlecht, S., Haverkamp, J., Bock, K., Thomas-Oates, J.E., Holst, O. (1998) Identification of a novel core type in Salmonella lipopolysaccharide. Complete structural analysis of the core region of the lipopolysaccharide from Salmonella enterica sv. Arizonae O62. J Biol Chem 273: 38173829.
  • 21
    Parker, C.T., Kloser, A.W., Schnaitman, C.A., Stein, M.A., Gottesman, S., Gibson, B.W. (1992) Role of the rfaG and rfaP genes in determining the lipopolysaccharide core structure and cell surface properties of Escherichia coli K-12. J Bacteriol 174: 25252538.
  • 22
    Polissi, A. & Georgopoulos, C. (1996) Mutational analysis and properties of the msbA gene of Escherichia coli, coding for an essential ABC family transporter. Mol Microbiol 20: 12211233.
  • 23
    Raetz, C.R.H. (1996) Bacterial lipopolysaccharides: A remarkable family of bioactive macroamphiphiles. In Escherichia coli and Salmonella. Cellular and Molecular Biology. Neidhardt, F.C., Curtiss R. III, Ingraham, J.L., Lin, E.C.C., Low, K.B., et al. (eds). Washington, DC: American Society for Microbiology Press, pp. 10351063.
  • 24
    Reeves, P.R., Hobbs, M., Valvano, M.A., Skurnik, M., Whitfield, C., Coplin, D., et al (1996) Bacterial polysaccharide synthesis and gene nomenclature. Trends Microbiol 4: 495503.
  • 25
    Schnaitman, C.A. & Klena, J.D. (1993) Genetics of lipopolysaccharide biosynthesis in enteric bacteria. Microbiol Rev 57: 655682.
  • 26
    Severn, W.B., Kelly, R.F., Richards, J.C., Whitfield, C. (1996) Structure of the core oligosaccharide in the serotype O8 lipopolysaccharide from Klebsiella pneumoniae. J Bacteriol 178: 17311741.
  • 27
    Sirisena, D.M., MacLachlan, P.R., Liu, S.-L., Hessel, A., Sanderson, K.E. (1994) Molecular analysis of the rfaD gene, for heptose synthesis, and the rfaF gene, for heptose transfer, in lipopolysaccharide synthesis in Salmonella typhimurium. J Bacteriol 176: 23792385.
  • 28
    Stanislavsky, E.S., Makarenko, T.A., Kholodkova, E.V., Lugowski, C. (1997) R-form lipopolysaccharides (LPS) of Gram-negative bacteria as possible vaccine antigens. FEMS Immunol Med Microbiol 18: 139145.
  • 29
    Stanley, P.L.D., Diaz, P., Bailey, M.J.A., Gygi, D., Juarez, A., Hughes, C. (1993) Loss of activity in the secreted form of Escherichia coli haemolysin caused by an rfaP lesion in core lipopolysaccharide assembly. Mol Microbiol 10: 781787.
  • 30
    Süsskind, M., Brade, L., Brade, H., Holst, O. (1998) Identification of a novel heptoglycan of α1 [RIGHTWARDS ARROW] 2-linked D-glycero-D-manno-heptopyranose. Chemical and antigenic structure of lipopolysaccharides from Klebsiella pneumoniae ssp. pneumoniae rough strain R20 (O1:K20). J Biol Chem 273: 70067017.
  • 31
    Whitfield, C., Amor, P.A., Köplin, R. (1997) Modulation of the surface architecture of Gram-negative bacteria by the action of surface polymer:lipid A-core ligase and by determinants of polymer chain length. Mol Microbiol 23: 629638.
  • 32
    Wyckoff, T.J.O., Raetz, C.R.H., Jackman, J.E. (1998) Antibacterial and anti-inflammatory agents that target endotoxin. Trends Microbiol 6: 154159.
  • 33
    Yethon, J.A., Heinrichs, D.E., Monteiro, M.A., Perry, M.B., Whitfield, C. (1998) Involvement of waaY, waaQ, and waaP in the modification of Escherichia coli lipopolysaccharide, and their role in the formation of a stable outer membrane. J Biol Chem (in press).