Innate Immunity

Innate Immunity http://ini.sagepub.com

Review: Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa Jerry D. King, Dana Kocíncová, Erin L. Westman and Joseph S. Lam Innate Immunity 2009; 15; 261 originally published online Aug 26, 2009; DOI: 10.1177/1753425909106436 The online version of this article can be found at: http://ini.sagepub.com/cgi/content/abstract/15/5/261

Published by: http://www.sagepublications.com

On behalf of: International Endotoxin & Innate Immunity Society

Additional services and information for Innate Immunity can be found at: Email Alerts: http://ini.sagepub.com/cgi/alerts Subscriptions: http://ini.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.co.uk/journalsPermissions.nav Citations http://ini.sagepub.com/cgi/content/refs/15/5/261

Downloaded from http://ini.sagepub.com at UNIV OF GUELPH on October 6, 2009

Review

Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa

15(5) (2009) 261–312 ß SAGE Publications 2009 ISSN 1753-4259 (print) 10.1177/1753425909106436

Jerry D. King, Dana Kocıńcova´, Erin L. Westman, Joseph S. Lam Department of Molecular and Cellular Biology, University of Guelph, Guelph, Ontario, Canada

Pseudomonas aeruginosa causes serious nosocomial infections, and an important virulence factor produced by this organism is lipopolysaccharide (LPS). This review summarizes knowledge about biosynthesis of all three structural domains of LPS – lipid A, core oligosaccharide, and O polysaccharides. In addition, based on similarities with other bacterial species, this review proposes new hypothetical pathways for unstudied steps in the biosynthesis of P. aeruginosa LPS. Lipid A biosynthesis is discussed in relation to Escherichia coli and Salmonella, and the biosyntheses of core sugar precursors and core oligosaccharide are summarised. Pseudomonas aeruginosa attaches a Common Polysaccharide Antigen and O-Specific Antigen polysaccharides to lipid A-core. Both forms of O polysaccharide are discussed with respect to their independent synthesis mechanisms. Recent advances in understanding O-polysaccharide biosynthesis since the last major review on this subject, published nearly a decade ago, are highlighted. Since P. aeruginosa O polysaccharides contain unusual sugars, sugar-nucleotide biosynthesis pathways are reviewed in detail. Knowledge derived from detailed studies in the O5, O6 and O11 serotypes is applied to predict biosynthesis pathways of sugars in poorly-studied serotypes, especially O1, O4, and O13/O14. Although further work is required, a full understanding of LPS biosynthesis in P. aeruginosa is almost within reach. Keywords: lipid A, lipopolysaccharide, O antigen, core, Pseudomonas aeruginosa

PSEUDOMONAS AERUGINOSA Pseudomonas aeruginosa is a Gram-negative, motile, rod-shaped bacterium, which is widely distributed in the environment. In the 1890s, it was isolated from the wounds of two patients in France; because of the characteristic blue-green pus associated with these infections, it was called Bacillus pyocyaneus (‘blue pus bacillus’). Early reports from the 19th century had already associated this organism with hospital epidemics1 and today it remains one of the major causes of nosocomial infections.2 P. aeruginosa is an opportunistic pathogen and causes disease by exploiting defects in normal host defences. In humans, this pathogen can cause life-threatening infections in individuals who are compromised because of immune deficiencies, and those suffering from burn wounds, cancer or cystic fibrosis.3–5

The success of P. aeruginosa as a pathogen is due partly to its metabolic diversity but also to its arsenal of virulence factors (reviewed by Lyczak et al.6). Furthermore, once a P. aeruginosa infection is established, treatment can be difficult due to its high intrinsic antibiotic resistance. This innate resistance is due, in part, to a large number of multidrug efflux systems,7 and to the low permeability of the P. aeruginosa outer membrane, which is estimated to be two orders of magnitude lower than the outer membrane of Escherichia coli for small hydrophilic compounds.8,9 Further resistance to multiple drugs can be acquired in P. aeruginosa by step-by-step accumulation of multiple mutations that reduce drug uptake, increase efflux, or alter targets within the cell. Alone, these individual mutations may have modest effects on antibiotic resistance, but the cumulative effect can significantly reduce the efficacy of clinically important drugs without

Received 4 March 2009; Revised 21 April 2009; Accepted 24 April 2009 Correspondence to: Dr Joseph S. Lam, Department of Molecular and Cellular Biology, University of Guelph, Ontario N1G 2W1, Canada. Tel: þ1 519 824 4120 ext 53823; Fax: þ1 519 837 1802; E-mail: [email protected] Downloaded from http://ini.sagepub.com at UNIV OF GUELPH on October 6, 2009

262

King, Kocıńcova´, Westman, Lam

any need for the strain to acquire exogenous antibioticresistance enzymes.10 The emergence of multiple-drug resistant strains is a contributing factor to the poor prognoses associated with P. aeruginosa infections once they are established.11 The biofilm mode of growth also contributes to P. aeruginosa resistance to antibiotics. Biofilms, defined as complex bacterial communities within an extracellular matrix, are more resistant than non-biofilm grown bacteria, because antibiotics may bind to the extracellular matrix, and have reduced activity in areas of biofilm that are characterised by poor oxygenation or low pH.12

LIPOPOLYSACCHARIDE An important virulence factor for P. aeruginosa is lipopolysaccharide (LPS), a complex glycolipid which is the major component of the outer leaflet of the outer membrane in Gram-negative bacteria. Lipopolysaccharide constitutes a physical barrier protecting the bacterium from host defences, mediates direct interactions with host cell receptors and antibiotics and, as endotoxin, is itself a potent signalling molecule which initiates some of the events leading to host tissue damage and much of the pathology associated with bacteraemia. Due to the biological significance of this molecule, it has been the subject of intense study and a great deal is known about its biosynthesis. It is now nearly a decade since the last major review on P. aeruginosa LPS biosynthesis,13 which focused specifically on O polysaccharides; since that time, many significant advances have been made in our understanding of enzymes and pathways involved in the biosynthesis of this molecule in P. aeruginosa. The purpose of this review is to reflect the current knowledge of the biology and chemistry of LPS biosynthesis in P. aeruginosa. Lipopolysaccharide has three domains Lipopolysaccharide is often described as a molecule with three domains. The first domain is called lipid A: it contains a disaccharide backbone, to which are attached several fatty acid chains that anchor LPS into the outer membrane. In all wild-type strains, the lipid A is attached to a nine or ten-sugar, branched oligosaccharide known as the core. A proportion of the LPS molecules on the surface of any given cell has only these two domains. Such molecules are sometimes referred to as lipid A-core. The third LPS domain consists of a repetitive carbohydrate polymer, which is covalently attached to the core, and can be referred to as the O antigen, O polysaccharide or O chain. Pseudomonas aeruginosa

cells can simultaneously produce two types of O polysaccharide in the same cell, which are distinct structurally, serologically and according to the mechanisms of their biosynthesis. Historically, these have been called A-band and B-band O polysaccharides; however, because these are not intuitively meaningful terms, we propose the following nomenclature for these structures. As the B-band polysaccharides are the basis for the P. aeruginosa O-serotyping scheme, these structures will be called ‘O-Specific Antigens’, or ‘O Antigens’. The A-band O polysaccharide has previously been called ‘Common Polysaccharide Antigen’, or ‘Common O Polysaccharide’ and these terms usefully distinguish this structure from the O-Specific Antigens and also, for example, from common outer-membrane protein antigens.14 The Common Polysaccharide Antigen is a homopolymer of D-rhamnose (D-Rha). It is produced by the majority of, but not all, P. aeruginosa strains, and elicits a relatively weak antibody response. There is some heterogeneity in the rhamnan polymer lengths produced in one cell, but Common O-Polysaccharide chains are usually approximately 70 sugars long. The O-Specific Antigens are heteropolymers in which different sugars are organised into repetitive O units. In contrast to the Common O Polysaccharide, the O-Specific Antigen polysaccharides are highly immunogenic, eliciting strong antibody responses. The nature and number of sugars in the O-Antigen repeating unit varies from strain to strain, and these are the structures recognised by O-specific antibodies. The International Antigenic Typing Scheme (IATS) for P. aeruginosa classifies strains into 20 major serotypes, O1 to O20, reflecting the diversity of O-Antigen structures within the species. In comparison to the Common O Polysaccharides, the O-Specific Antigen polymers are added to lipid A-core with a wide range of different chain lengths, some considerably longer than Common O-Polysaccharide chains, and some containing a single repeat unit. Not all P. aeruginosa strains are O-serotypable, however, and it is common, especially in clinical isolates, for the ability to produce O-Antigen to be lost. Bacteria devoid of O Antigen are known as rough strains because, in some genera, mutations which abrogated O Antigen production result in ‘rough’ colony morphology, compared with the ‘smooth’ colonies of O Antigen-producing strains. The rough and smooth terminology is now used to refer to Opolysaccharide or O-polysaccharideþ strains regardless of colony morphology, and is sometimes even used to describe LPS molecules which either are or are not substituted with O-polysaccharides. For example, lipid A-core can be called ‘rough-type’ (R) LPS. O-Antigen mutants have also been described which only decorate lipid A-core with a single O unit, but do not produce O-Antigen polymers. Such strains are called


Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa ‘semi-rough’ (SR), and it is also common to refer to semi-rough LPS molecules (or core þ 1 LPS), which normally make up a proportion of the LPS present on the surface of smooth strains. The three-domain notion of LPS has been very successful because different properties of the molecule segregate more-or-less along the domain boundaries. For example, O polysaccharides and lipid A-core are synthesised by independent biosynthetic pathways, and the lipid A domain is responsible for the biological properties of LPS associated with its alternative name of endotoxin. In the following section, we briefly outline the biological functions of each domain; then, we shall describe the chemical structures and consider the biosynthesis of LPS, taking each domain in turn. The biological significance of specific LPS domains Role of lipid A in pathogenesis

Lipid A is the domain of LPS which mediates inflammatory response-induced endotoxicity;15 however, the extent of that toxicity varies widely between different bacterial species. E. coli and Neisseria meningitidis lipidA structures each contain six 14- or 12-carbon lipids, and are generally considered to be optimal for eliciting the maximal inflammatory response via human Toll-like receptor 4 (TLR4).16 As described in detail below, the major lipid A produced by laboratory strains of P. aeruginosa differs from these optimal structures both in the number and length of acyl chains. Therefore, the inflammatory response elicited by P. aeruginosa lipid A would be expected to be less than that of enterobacterial LPS, and this is indeed the case.17 This being said, some clinical isolates of P. aeruginosa, from cystic fibrosis patients with severe pulmonary disease, synthesise a hyperacylated lipid A which induces an enhanced inflammatory response. In these cases, lipid A structures produced by the cystic fibrosis airway-adapted P. aeruginosa probably contribute to the severity of lung damage.18,19 Many bacteria make modifications of their lipid A structures, in response to environmental stimuli, which increase resistance to the bactericidal effects of cationic antimicrobial peptides. These peptides are a diverse, wide-spread component of the innate immune system (reviewed by Hancock and Diamond20). They are abundant both inside neutrophils and on mucosal surfaces and other sites of potential infection; while they contribute to host defences through a range of different mechanisms, some (for example, polymyxin B, human defensin, and cecropin) function directly by interaction with bacterial membranes. By modification of the patterns of lipid A acylation, or addition of positively-charged substituents to the phosphate groups, bacteria reduce the permeability

263

of the membrane or the affinity of lipid A for these cationic peptides.21 Pseudomonas aeruginosa also modifies lipid A in response to environmental signals, resulting in greater resistance to some antimicrobial peptides22 and, in part, these modifications can be induced by exposure to such peptides.23 Role of core in pathogenesis

Binding and internalisation of respiratory antigens by epithelial cells, followed by desquamation may be an important mechanism for clearing bacteria from the lung.24 Internalisation of P. aeruginosa by epithelial cells correlates with membrane expression of the cystic fibrosis transmembrane conductance regulator (CFTR) that is defective (for instance, the F508 allele) in cystic fibrosis patients. When an airway epithelial cell line expressing a functional wild-type human CFTR was tested for ingestion of P. aeruginosa, the internalisation of P. aeruginosa cells could be inhibited by addition of exogenous LPS core oligosaccharides from a strain producing a complete core, but oligosaccharides from a strain with a truncated core did not inhibit. This indicated that the bacterial ligand for the wild-type CFTR receptor which mediates epithelial-cell internalisation of P. aeruginosa is core oligosaccharide.24 Pseudomonas aeruginosa is the most common cause of bacterial keratitis associated with the use of contact lenses.25 It is able to enter and survive in corneal cells in an in vitro cell line and mouse or rabbit models.26,27 As with the epithelial cell line experiments described above, it appears that the bacterial ligand which mediates uptake into corneal cells is the core oligosaccharide. Strains with core structure defects were ingested more slowly than the control with complete core. Ingestion could again be inhibited by addition of exogenous core oligosaccharides, but only if the core was complete. Since the core oligosaccharide structures in these experiments were only characterised using sugar composition analyses, the core structure equating to the ligand responsible for interaction with the host cells is unknown.28 Later, Zaidi and co-workers29 showed, in an animal model, that homozygous F508 cftr mice were almost completely resistant to P. aeruginosa corneal infection. Thus, wild-type CFTR-mediated internalisation of P. aeruginosa by corneal epithelial cells is a mechanism which causes disease in experimental eye infection,29 while in the airways of humans or animals, CFTRmediated P. aeruginosa uptake by epithelial cells appears to be protective, probably by enhancing clearance of bacteria.30,31 Role of O polysaccharides in pathogenesis

O polysaccharides extend outward from the outer cell membrane, and are thus involved in many of the


264


interactions between the bacterium and the environment or host. During infections caused by P. aeruginosa, the O-Specific Antigen protects the bacteria from phagocytosis and confers resistance to complement-mediated killing.32 Whole cell agglutination tests suggest that long-chain O-Specific Antigen conceals the shorter Common O Polysaccharide on cells producing both forms of O polysaccharide,33 and strains that express only the Common O Polysaccharide were not protected from lysis by complement.32 When present in a host, P. aeruginosa is faced with immune defence systems producing a variety of reactive oxygen species, such as H2O2. A recent study showed that O polysaccharide protects against oxidative stress. The sensitivity to hydrogen peroxide of a P. aeruginosa waaL mutant, which cannot ligate O polysaccharide to the LPS core, was tested. This strain had a 4-fold reduction in survival when challenged with H2O2, compared to the wild type.34 Pseudomonas aeruginosa waaL mutants in PAO1 and PA14 have also been shown to be defective in twitching and swimming motility.35 Since these forms of motility are important for initial surface colonisation, O polysaccharides may play a role in surface attachment. In the context of chronic P. aeruginosa infections, such as those common in cystic fibrosis patients, the Common Polysaccharide Antigen becomes the major LPS antigen over time, because there are apparently long-term selective pressures for loss of high molecular mass O-Specific Antigen.36 The production of O Antigen has been restored to clinical isolates by transformation with a plasmid encoding all or select genes from the O-Antigen biosynthesis locus, suggesting that loss of this structure was due to mutations in the biosynthetic genes rather than due to regulatory changes.37 O Antigen is highly immunogenic, and elicits a strong antibody response from the infected host. In chronic P. aeruginosa infections, it is common for strains to reduce O-Antigen production and then to lose it altogether. These changes may be driven by the selective pressures on the bacteria to evade the antiO-Antigen immune response. Another selective pressure for loss of O Antigen may be exposure of the bacteria to repeated antibiotic therapy, since defects in synthesis of either O polysaccharide reduce ionic binding of the aminoglycoside antibiotic gentamicin to the cell surface and the strains are, therefore, less prone to antibioticinduced killing.38 This being said, there have been no reports of selective pressure for loss of Common Polysaccharide Antigen in chronic infections. A longitudinal study of lung infections in cystic fibrosis patients showed that the appearance of Common Polysaccharide Antigen-specific antibodies in host serum correlates with decreased pulmonary function and extended duration of P. aeruginosa infection.33

The exopolysaccharide alginate can also protect against anti-Common Polysaccharide Antigen antibodymediated opsonophagocytosis.39 In vitro assays with monolayers of cultured human bronchial epithelial cells showed that a defined Common Polysaccharide Antigendeficient P. aeruginosa mutant had a defect in adherence. The numbers of mutant cells attached to the epithelial cells were lower than those of wild type, and the mutant was attached as isolated bacteria rather than microcolonies like the wild-type. This suggests that, despite being apparently masked by the long-chain O-Antigen, the Common O polysaccharide plays a role in adherence of P. aeruginosa to epithelial cells under certain conditions.40 Infection studies using animal models have shown that O polysaccharides are required for virulence. In a burned-mouse model, the LD50 for an O Antigendeficient mutant was at least 1000-fold higher than that of the wild-type strain. This was attributed to the inability of the mutant to colonise burned skin effectively and to avoid serum killing.41 The status of Common O Polysaccharide production in this strain was not tested, but visualisation of the results from LPS analysis by SDS-PAGE and silver staining in this report indicated that the strain used in this study did produce Common Polysaccharide Antigen. In several studies, mutants producing truncated core structures, that are defective as acceptors for either O polysaccharide, have been used to assess the role of O polysaccharides in virulence. Some of the mutations have pleiotropic effects, but the loss of O-polysaccharide production likely contributed significantly to the decrease in virulence observed. For example, a galU mutant that cannot produce UDP-D-Glc was significantly attenuated in a corneal infection model. The same mutant had higher LD50 than wild-type cells in an acute pneumonia model, but still caused mortality and severe pneumonia despite minimal systemic spread.42 A separate study43 used a neonatal mouse model to compare the virulence of P. aeruginosa PAO1 to that of a PAO1 algC mutant, defective in phosphomannomutase (PMM) and phosphoglucomutase (PGM) activity. The algC gene is required for the synthesis of UDP-D-Glc, GDP-D-Man, and dTDP-L-Rha, which are precursors for LPS core, Common Polysaccharide Antigen, or the two exopolysaccharides important for biofilm formation, alginate and Psl. Due to the defect in LPS core, the algC mutant is unable to produce LPS core with either O polysaccharide.44 In an in vitro binding assay, the algC mutant was significantly less adherent to human respiratory epithelial cells, transformed tracheal epithelial cells, or to mouse adenoma cells. When mice were challenged with the algC mutant, the bacteria were readily cleared from the respiratory tract and the bacteraemia and mortality rates were significantly lower than those with PAO1.43


Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa (a)

LIPID A

OH

O HO HO

Lipid A structures in P. aeruginosa

P O O

O

GlcN

O

O

Lipopolysaccharide is present on the bacterial cell surface as a heterogeneous mixture of related structures, and much of this heterogeneity is contributed by variations in the lipid A moiety. While relatively little work has been done to characterise the biosynthetic steps for lipid A in P. aeruginosa, the structures themselves have been extensively studied (reviewed by Knirel et al.45). The backbone of P. aeruginosa lipid A consists of doubly-phosphorylated diglucosamine (phosphate! 40 -b-D-GlcNII-(1 ! 6)-a-D-GlcNI-1 ! phosphate) and the lipid A produced by most laboratory strains in standard media, has a basic penta- or hexa-acylated form (Fig. 1A).46 In the hexa-acyl form, which comprises about 25% of lipid A in these strains, the diglucosamine backbone is symmetrically acylated, each sugar having an N-linked 12 : 0(3-OH) [(R)-3-hydroxylauroyl] group at the 2 position, an O-linked 10 : 0(3-OH) [(R)-3hydroxydecanoyl] substituent at the 3 position and a secondary 12 : 0 (lauroyl) chain. The secondary lauroyl chains are non-stoichiometrically 2-hydroxylated (Fig. 1, green). In about 75% of lipid A prepared from laboratory strains, the 3 position of GlcNI is unsubstituted, and this constitutes the penta-acyl form (Fig. 1A, red). When lipid A from clinical isolates has been characterised, novel structures were found, arising from the loss and/or addition of particular structural elements (Fig. 1B). Some of these modifications are regulated in response to environmental stimuli, allowing P. aeruginosa to modulate its interactions with a host organism, or to resist the bactericidal effects of antimicrobial agents.22 In isolates of P. aeruginosa from non-cystic fibrosis infections (such as blood infection), a penta-acyl lipid A was found lacking the secondary acyl chain on GlcNI (Fig. 1B, red secondary chain).19,22 Pseudomonas aeruginosa isolated from cystic fibrosis infections exhibits lipid A with six or seven acyl chains, similar to the laboratory penta- and hexa-acyl structures but with an additional secondary palmitoyl group (16 : 0) in an ester linkage to the 3-hydroxylauroyl at C3 on GlcNII (Fig. 1B, blue).19,22,47 Non-stoichiometric addition of 4-amino4-deoxy-L-arabinose (L-Ara4N) at either or both lipid-A phosphates has also been observed in clinical isolates as well as in laboratory strains grown under particular conditions, such as low magnesium (Fig. 1B, blue).47–49 Biosynthesis of lipid A in P. aeruginosa The biosynthesis of lipid A has been studied in detail for enteric bacteria, particularly E. coli, and this pathway is assumed to be generally conserved in P. aeruginosa.

265

NH

HO O

O

HO O

NH P O HO OH

O

HO OH

10

O

3

O O

O

GlcN

O O

12

OH

10

12 12 12

(b) NH2 O HO

OH

O

OH O HO

NH2 O

P

O

O O

HO

GlcN

O

OH

3'

O

NH O

O

O

HO O

GlcN 2

O

O

O

O

O

HO

O OH

10

NH P O HO O

O O

12

OH

10 16

12 12 12

(c)

O-Kdo2

O HO HO

P

O

O O

GlcN

O

O NH

O

O

O

HO O

GlcN

O

O

O

O

HO

O

O NH P O HO OH

HO

14 14 14 14

14 12

Fig. 1. Lipid A structures. (A) Typical lipid A produced by laboratoryadapted strains in standard laboratory media. The 3-hydroxydecanoyl chain at the 3-position is absent in approximately 75% of LPS molecules. Secondary acyl chains can exhibit 2-hydroxylation. (B) Clinical isolates of P. aeruginosa commonly produce lipid A with various modifications of the basic structure shown in (A). These include loss of secondary acylation at the 2-position, secondary palmitoylation at position 30 , and addition of L-Ara4N to either of the phosphates. (C) The hexa-acyl lipid A structure of E. coli.


266


This assumption is based largely on the identification of homologues of the E. coli genes in the P. aeruginosa genome (Table 1), but the majority of lipid A biosynthetic steps have not been directly investigated. In the following section, we concentrate on steps in P. aeruginosa lipid A biosynthesis for which experimental data exist. The reader is referred to two excellent reviews50,51 for the general scheme of lipid A biosynthesis. Transfer of primary acyl chains: lpxA and lpxD

The lengths of the primary O- or N-linked lipid chains differ between bacterial species (in E. coli they are all 14 : 0(3-OH), whereas P. aeruginosa has 10 : 0(3-OH) and 12 : 0(3-OH) groups). The distribution of these fatty acids on lipid A is determined by the species-specific hydrocarbon-length preferences of their respective O-acyltransferase (LpxA)52 and the N-acyltransferase (LpxD)53 enzymes. The first step in lipid A biosynthesis is the reversible 3-O-acylation of UDP-D-GlcNAc (1)

catalysed by LpxA (Fig. 2). By making use of the anticipated preference of P. aeruginosa LpxA for the shorter chain lengths, compared with the E. coli enzyme, the P. aeruginosa lpxA gene was identified in a 3-hydroxydecanoyl acyltransfer screen of a PAO1 library expressed in E. coli.52 The P. aeruginosa gene was then cloned and shown to complement an E. coli lpxA temperature sensitive mutation, restoring viability to the cell when grown at the non-permissive temperature, but it caused incorporation of 3-hydroxydecanoic acid [10 : 0(3-OH)] rather than the 3hydroxymyristoic acid [14 : 0(3-OH)] usually found in E. coli lipid A. LpxA enzymes use acyl carrier protein (ACP)-linked fatty acids as substrates; in a transferase assay, the P. aeruginosa enzyme showed a 41000-fold preference for 3-hydroxydecanoyl-ACP over 3-hydroxymyristoyl-ACP, which is the reverse of the E. coli LpxA preference.52 These enzymes exquisitely discriminate between different fatty chain lengths, using what have been called ‘precise hydrocarbon rulers’, and their chain-length preferences can be interchanged by single amino acid substitutions.54

Table 1. Homologues of lipid-A biosynthesis genes found in the PAO1 genome. Sequence conservation relates to alignments with the E. coli K-12, or B. bronchiseptica RB50 sequences E. coli /P. aeruginosa gene

Related proteins (% identity)

Proposed function

Key reference*

lpxA/PA3644

54% E. coli LpxA

52

lpxC/PA4406 lpxD/PA3646

57% E. coli LpxC 51% E. coli LpxD

lpxH/PA1792 lpxB/PA3643 lpxK/PA2981 lpxL PA3242

47% E. coli 53% E. coli 45% E. coli 47% E. coli

lpxL PA0011 lpxO PA4512

arnA/PA3554 arnB/PA3552 arnC/PA3553 arnD/PA3555 arnE/PA3557 arnF/PA3558 arnT/PA3556

29% E. coli LpxL 60% LpxO S. enterica sv. Typhimurium 50% LpxO S. enterica sv. Typhimurium 29% B. bronchiseptica PagL 69% E. coli ArnA (K-12) 62% E. coli ArnB 64% E. coli ArnC 60% E. coli ArnD 54% E. coli ArnE 36% E. coli ArnF 41% E. coli ArnT

3-Hydroxydecanoyl-[acyl carrier protein]! UDP-GlcNAc 3-O-acyltransferase UDP-3-O-acyl-GlcNAc 2-N-deacetylase 3-Hydroxylauroyl-[acyl carrier protein]!UDP3-O-acyl-GlcN 2-N-acyltransferase UDP-diacyl-GlcN pyrophosphatase Lipid A precursor disaccharide synthase Lipid A precursor disaccharide kinase Lauroyl-[acyl carrier protein]!lipid IVA secondary lauroyl chain transferase Secondary lauroyl chain 2-hydroxylase Secondary lauroyl chain 2-hydroxylase

ugd/PA3559

28% E. coli Ugd

lpxO PA1936 pagL/PA4661

LpxH LpxB LpxK LpxL

53 53

59 59

Lipid A GlcNI 3-O-deacylase Lipid A GlcNI 3-O-deacylase

65

Undecaprenyl-phosphate-L-Ara4N synthesis Undecaprenyl-phosphate-L-Ara4N synthesis Undecaprenyl-phosphate-L-Ara4N synthesis Undecaprenyl-phosphate-L-Ara4N synthesis Undecaprenyl-phosphate-L-Ara4N export Undecaprenyl-phosphate-L-Ara4N export Undecaprenyl-phosphate-L-Ara4N!lipid A L-Ara4N transferase UDP-Glc 6-dehydrogenase (undecaprenyl-phosphate-linked L-Ara4N synthesis)

23

*References are cited if they describe experimental investigation of the P. aeruginosa gene. Downloaded from http://ini.sagepub.com at UNIV OF GUELPH on October 6, 2009

69

Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa The assay data described above, from ectopically expressed P. aeruginosa LpxA, mimicked the substrate preference observed in an assay previously conducted to assess the abilities of P. aeruginosa crude extracts to modify UDP-GlcNAc with ACPlinked acyl chains of different lengths.53 The same approach was used to judge the specificity of the reaction in which the N-linked acyl chain of lipid A is attached, by incubating crude cell extracts with in situ-generated UDP-3-O-[(R)-3-hydroxymyristoyl]GlcN as the acyl acceptor. The compound added into these incubations was UDP-3-O-[(R)-3-hydroxymyristoyl]-GlcNAc, and the authors relied on the ability of crude extracts to de-acetylate this compound (via the LpxC-catalysed reaction) to generate the free amine substrate. In this case, P. aeruginosa extracts showed a preference for 3-hydroxylauroyl-ACP, in contrast to E. coli extracts, which used 3-hydroxymyristoyl-ACP most efficiently.53 The enzyme catalysing this reaction in P. aeruginosa is presumably the P. aeruginosa homologue of E. coli LpxD, encoded by PA3646.

Biosynthesis of the lipid IVA analogue in P. aeruginosa: lpxH, lpxB and lpxK

Lipid IVA is an important intermediate lipid structure in the biosynthesis of E. coli and Salmonella enterica sv. Typhimurium lipid A.50 To our knowledge, none of the three steps leading to biosynthesis of the lipid IVA analogue have been investigated experimentally in P. aeruginosa, but the nucleotidase (LpxH), disaccharide synthase (LpxB) and kinase (LpxK) reactions are presumed to be catalysed by the P. aeruginosa homologues of the E. coli enzymes in a conserved manner (Fig. 3; Table 1). The end-product of these reactions is the P. aeruginosa analogue of lipid IVA (7) (Fig. 3).

O

HO HO

Addition of secondary lauroyl chains: lpxL1 and lpxL2

In E. coli and S. enterica sv. Typhimurium, lipid A biosynthesis cannot be completed without addition of the first core sugar, 3-deoxy-D-manno-oct-2-ulosonic acid (Kdo). The E. coli acyl transferase enzymes are called LpxL (HtrB) and LpxM (MsbB), and they catalyse addition of the secondary lauroyl group and then the myristoyl group, respectively, to complete synthesis of Kdo2-lipid A.55,56 The E. coli enzymes have a Kdocontaining substrate specificity.55–57 The enzymes which catalyse transfer of the two secondary lauroyl chains in P. aeruginosa apparently do not share the same specificity because even if Kdo biosynthesis is inhibited (by an antibacterial agent), P. aeruginosa produces lipid A without Kdo, but fully acylated.58 The lack of an absolute Kdo-requirement was confirmed by in vitro assays using dilute cytoplasmic extracts or membrane preparations from P. aeruginosa with lauroyl-ACP (12 : 0-ACP) and E. coli lipid IVA as the donor and acceptor substrates, respectively. In a similar experiment, soluble E. coli extracts are unable to lauroylate lipid IVA, but added two lauroyl groups to Kdo2-lipid IVA. In contrast, each of the P. aeruginosa extracts catalysed the addition of a single lauroyl group, and was able to do so more efficiently without Kdo present in the substrate.59 The P. aeruginosa gene or genes which encode this activity have not been determined, but there are two candidates in the PAO1 genome, PA3242 (lpxL1) and PA0011 (lpxL2). These genes encode putative acyltransferases homologous to LpxL and LpxM from E. coli, but both are more closely related to the lauroyltransferase LpxL. Presumably, these enzymes each transfer one of the secondary lauroyl groups found in most mature P. aeruginosa lipid A molecules (8) (Table 1). It is not known why only one secondary lauroyl group was transferred by the P. aeruginosa extracts. It could be that the lauroyltransferase to act second has a preference for its native P. aeruginosa

OH

OH + 3-OH-C10

NH UDP O

LpxA

HO O O HO

OH O

–CH3CO2H

NH UDP O

LpxC

267

HO O O

OH O

H2N

HO

+ 3-OH-C12 LpxD UDP

O

HO O O O

HO

NH UDP

HO UDP-D-GlcNAc

10

10

10

12

1

2

3

4

Fig. 2. Early steps in the lipid A biosynthesis pathway. The starting material UDP-D-GlcNAc (1) is a common precursor for synthesis of LPS sugars, lipid A and peptidoglycan. Activity relating to all three steps has been detected in P. aeruginosa lysates.52,53 ACP, acyl carrier protein; 3-OH-C10 (R)-3hydroxydecanoate [10 : 0(3-OH)]; 3-OH-C12, (R)-3-hydroxylaurate [12 : 0(3-OH)]. Downloaded from http://ini.sagepub.com at UNIV OF GUELPH on October 6, 2009


268

OH O

HO O O O

HO

NH UDP

HO 10 12

–UMP

O

HO O

–UDP

LpxB

NH O

OH

HO

O

HO O

HO O

O

HO

NH Pi

NH Pi

+ ATP

O O

O

O

LpxK

NH O HO

O

HO

NH Pi

HO

10

12

O

HO O

O

HO

HO

10

O

O

HO O

O

HO

Pi

O

O

LpxH

OH

OH

4

12 10

10 12

12

HO 10 12

5

6

7 (“lipid IVA”)

Fig. 3. Proposed pathway for the biosynthesis of the P. aeruginosa lipid IVA analogue. This scheme is predicted by analogy to the well-described pathway for lipid IVA biosynthesis in E. coli.50 Well-conserved homologues for the E. coli genes are present in the PAO1 genome. Pi, phosphate.

substrate in which the primary acyl chains are shorter than those in E. coli-derived substrate used in these experiments.59 Some non-cystic fibrosis clinical isolates of P. aeruginosa, for example from blood infections, lack one of the secondary acyl chains usually present in isolates from other environments (Fig. 1B, red secondary chain).19 It is unknown whether this structure arises because the function of one of the lauroyltransferase enzymes is down-regulated or lost and the acyl chain was never added to these molecules, or whether this acyl chain is perhaps removed after completion of lipid A synthesis by a novel deacylase enzyme similar to PagL (see below). Both scenarios are conceptually possible.

When lpxO is inactivated in this organism, the (S)-2hydroxymyristoyl [14 : 0(2-OH)] group is lost63 and when lpxO is introduced to E. coli on a plasmid, its lipid A gains 2-hydroxylation of myristate.61 Salmonella enterica sv. Typhimurium LpxO is a membrane protein which incorporates 18O from 18O2 into lipid A61 and, in vitro, membranes containing LpxO catalysed the hydroxylation of Kdo2-hexa-acyl-lipid A in the presence of Fe2þ, O2, and a-ketoglutarate.62 Genes encoding LpxO homologues are found in bacteria which produce 2-hydroxylacyl-containing lipid A and two homologues are present in the PAO1 genome – PA4512 and PA1936.61 Probably the enzymes encoded by these genes each hydroxylate one of the two acyl chains exhibiting this modification, but neither gene has yet been investigated experimentally.

Hydroxylation of secondary lauroyl chains: lpxO1 and lpxO2

Analyses of P. aeruginosa lipid A structures reveal that both secondary lauroyl chains can be 2-hydroxylated. This is a feature of P. aeruginosa lipid A which is not found in E. coli, but 2-hydroxylation of one secondary myristoyl chain is observed in S. enterica sv. Typhimurium lipid A.60 The enzyme responsible for this in S. enterica sv. Typhimurium is LpxO.61,62

Modifications of lipid A in P. aeruginosa In this section, we describe alterations to lipid A structures which occur after transport of LPS across the inner membrane. For a general description of lipid-A modification systems, the reader is referred to an excellent review by Trent et al.21


Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa Deacylation: pagL

As described above, the majority of lipid A molecules produced by laboratory strains of P. aeruginosa lack acylation at the 3 position of GlcNI (Fig. 1, red primary acyl chain). From work in S. enterica sv. Typhimurium, it was known that such species can be synthesised by removal of the 3-acyl chain in a reaction catalysed by an outer membrane enzyme known as PagL.64 The pagL gene in P. aeruginosa shares little sequence identity with the S. enterica sv. Typhimurium pagL but was identified using an iterative BLAST methodology.65 As judged by mass spectrometry (MS) and gas chromatography (GC), a P. aeruginosa pagL mutant lost the predominant pentaacylated lipid A of its parental strain (PAO1) and expressed instead a hexa-acylated lipid A.66 When over-expressed in E. coli, the P. aeruginosa PagL caused lipid A to be deacylated, indicating that P. aeruginosa PagL can remove the longer 14 : 0(3-OH) chains of E. coli lipid A as well as the 10 : 0(3-OH) chains of its native species. PagL from P. aeruginosa has been crystallised and the crystal structure revealed that this enzyme has a serine hydrolase-type catalytic triad consisting of His126, Ser128 and Glu140 (amino acid numbering is for the mature protein).67

269

by which bacteria reduce their susceptibility to the bactericidal properties of cationic antimicrobial peptides. The pathway for addition of L-Ara4N to lipid A is well understood from studies in E. coli and S. enterica sv. Typhimurium. It requires the products of the ugd gene and the arnBCADTEF gene cluster; the pathway has been reviewed.68 Briefly, L-Ara4N is transferred to lipid A from an undecaprenyl-phosphate lipid carrier on the periplasmic face of the inner membrane, in a reaction catalysed by ArnT. UDP-Glc 6-dehydrogenase (Ugd), ArnA, ArnB, ArnC, and ArnD catalyse biosynthesis of the undecaprenol-phosphate-linked L-Ara4N in the cytoplasm, and ArnE and ArnF are required for its efficient transport across the inner membrane. Homologues of the E. coli genes are all found in a single cluster in P. aeruginosa – PA3552-PA3559 (or arnB-ugd) (Table 1, Fig. S1). To our knowledge, none of the P. aeruginosa genes have been investigated experimentally except for ugd which was confirmed to encode a UDP-glucose 6-dehydrogenase. PA3552 is one of two ugd genes in PAO1 but the polymyxin B-susceptible phenotype specific to the PA3552 mutant was characteristic of strains defective for aminoarabinoylation of lipid A suggesting that the ugd in the arn operon has a specific role in lipid A modification.69

Regulation of PagL activity

Mass spectrometrical and GC analysis of lipid-A structures indicated that, in acute infection and environmental isolates of P. aeruginosa, the 3-O-deacylase activity attributed to PagL was regulated by magnesium concentration and temperature, being less active at low [Mg2þ] and low temperatures.66 In contrast to these observations, when PagL protein expression levels were analysed in laboratory-adapted P. aeruginosa stains, a lack of response to Mg2þ concentration was observed.65 This may indicate that P. aeruginosa strains lose Mg2þregulation of PagL activity as a part of their adaptation to the laboratory, or that these researchers did not test a wide-enough range of Mg2þ concentrations. A third possibility is that PagL activity is regulated by some mechanism other than control of protein expression levels. Ernst and co-workers66 showed that a third of P. aeruginosa isolates that they examined from cystic fibrosis patients with severe lung disease produced hexaand hepta-acylated lipid A, which appeared to have arisen by loss of 3-O-deacylase activity. The hypothesis that these lipid A structures were due to loss of PagL function was confirmed by complementation of this phenotype by introduction of pagL on a plasmid. Addition of L-Ara4N: the arn locus

Partial masking of the negative charges of the lipid A phosphate substituents by addition of positively charged-moieties such as L-Ara4N is one mechanism

Regulation of L-Ara4N addition to lipid A: phoP, phoQ, pmrA and pmrB

The expression of arn genes in P. aeruginosa is controlled by two two-component regulatory systems, PmrA–PmrB and PhoP–PhoQ. Both of these systems respond to Mg2þ-limiting conditions and gene transcript analysis shows that both also regulate expression of the arn operon.70 With use of quantitative PCR and an arnB : lux reporter fusion, McPhee and co-workers70 showed that arn gene expression is induced by low Mg2þ in a manner which is partially dependent on PhoP. arn genes are also induced by a number of cationic antimicrobial peptides, including polymyxin B, in a manner which was partially dependent on PmrA–PmrB system.23 The role of the PmrA–PmrB system in resistance to cationic antimicrobial peptides has also been examined at the level of lipid A structures: single nucleotide mutations in pmrB can cause constitutive activation of this two-component system resulting in resistance to polymyxin B and other cationic peptides. MS analysis of purified lipid A from these mutants showed that they added L-Ara4N to lipid A in conditions where the parental strain does not do so.49 While it has not been directly demonstrated in P. aeruginosa, the addition of L-Ara4N is very likely an important mechanism through which PmrA–PmrB mediates cationic peptide resistance as this modification correlates well with resistance across a broad range of bacterial species.71–75


270


Transfer of secondary palmitoyl chain: PagP-like activity

Laboratory adapted P. aeruginosa strains add a secondary palmitoyl (16 : 0) fatty acid to lipid A at the 30 position (Fig. 1B, blue) when they are grown in low Mg2þ concentrations, a condition which activates PhoP– PhoQ.22 However, in a survey of P. aeruginosa strains obtained from different sources, there was a perfect correlation between isolation of strains from a cystic fibrosis patient lung infection and their stable palmitoylation of lipid A, irrespective of Mg2þ concentration.47 In other bacteria, palmitoylation of lipid A is catalysed by homologues of the E. coli and S. enterica sv. Typhimurium enzyme PagP.60,76,77 These are outer membrane proteins which catalyse the transfer of a palmitoyl chain to lipid A using acyl groups derived from phospholipids.78,79 PagP activity is generally regulated in response to environmental stimuli76,77 and confers greater resistance to cationic antimicrobial peptides in some bacteria,60,76 but not all.77 It is required for full virulence in certain infection models.76,77 Differences in regulation and expression of PagP, even between the closely related Bordetellae (B. pertussis, B. parapertussis and B. bronchiseptica), suggest that specific bacteria–host relationships apply diverse selective pressures on the function of this enzyme.80,81 This point is underlined by the contrast between complete inactivation of pagP in the acute, human respiratory tract pathogen B. pertussis, with apparently rapid selection for constitutive PagPlike activity in P. aeruginosa which is typically associated with chronic lung infections in cystic fibrosis. Thus far, no PagP homologue has been identified in P. aeruginosa, but its existence is inferred from the regulated palmitoylation of P. aeruginosa lipid A (which takes place at the 30 -position, just as in PagP-mediated palmitoylation in B. bronchiseptica) and also indicated by reports of an unpublished observation of PagP-like acyltransferase activity detected in P. aeruginosa membranes.51

CORE OLIGOSACCHARIDE Structure of core oligosaccharide The central domain of LPS that is interposed between lipid A and O polysaccharide is the core oligosaccharide. Core oligosaccharide structure of wild-type P. aeruginosa PAO1

The core oligosaccharide of PAO1 is the most-studied among P. aeruginosa strains. As with other species, the P. aeruginosa core structure is commonly divided, conceptually, into two parts – inner and outer core.

The inner core structure is similar to that of E. coli and S. enterica sv. Typhimurium, composed of two residues of Kdo (KdoI and KdoII) and two residues of L-glyceroI II D-manno-heptose (Hep and Hep ). Both heptose residues serve as phosphorylation sites (Fig. 4). Among Gram-negative bacteria, P. aeruginosa has the most phosphorylated core. Phosphorylation of LPS has been associated with intrinsic resistance to antibiotics, such as novobiocin.82 Ionic interactions between negativelycharged phosphate substituents on heptoses and divalent cations may stabilise the outer membrane by crosslinking LPS molecules,83 and this mechanism may be particularly important for P. aeruginosa because it is hypersusceptible to lysis by agents that chelate divalent cations (such as EDTA).84,85 This hypersusceptibility to metal chelators appears to be somehow related to the high phosphate content of P. aeruginosa LPS.86 Pseudomonas aeruginosa core has three phosphorylation sites: positions 2 and 4 of HepI and position 6 of HepII (Fig. 4).87–90 In addition, the phosphorylation position 2 of HepI has been found to be non-stoichiometrically occupied by diphosphoethanolamine (Fig. 4).90 Another modification in the inner core structure is the stoichiometric presence of an O-carbamoyl substituent at position 7 of HepII.91 The P. aeruginosa outer core is composed of three sugar types: D-glucose (Glc), L-rhamnose (L-Rha), and 2-amino-2-deoxy-D-galactose (D-galactosamine, GalN). The GalN is further modified by alanine (Ala). Pseudomonas aeruginosa cells simultaneously produce two structurally similar outer-core glycoforms, which are present in comparable amounts. These are glycoform 1(O), which is never substituted with O polysaccharide and glycoform 2(Oþ) which is the core glycoform which can be further substituted by O polysaccharides. Glycoform 1(O) has sometimes been called the ‘uncapped’ core glycoform, and glycoform 2(Oþ) is also known as the ‘capped’ core glycoform. Both glycoforms share the same inner core structure and four sugars, GalN and GlcI–GlcIII, of the outer core (Fig. 4). The glycoforms differ in the position of L-Rha and only glycoform 1(O) has a fourth glucose, GlcIV.88,89,92–95 Glycoform 2(Oþ) has L-RhaB (1 ! 3)-linked to GlcI, whereas the glycoform 1(O) has L-RhaA (1 ! 6)-linked to GlcII (Fig. 4). No core glycoform with two Rha residues has been found, suggesting that the attachment of either L-Rha residue precludes attachment of the other. Strain-dependent variations in core structures

Core structures have been characterised for a number of P. aeruginosa strains; while the basic structure is conserved, variation is seen among the peripheral structural features, including the terminal GlcIV and many of the non-carbohydrate substituents.



271

Glycoform 1 b-D-GlcIV 1

2 L-RhaA 1

1 4 6 b-D-GlcI 1

Pi

Pi

Pi

6

4

wapH*

migA* D-GlcIII 1

Pi 6 D-GlcII

3 D-GalN 1 2

3

waaP*

HepII 1

3

7

wapG L-Ala

HepI 1

5

KdoI 2

waaC*

waaF*

Cm

lipid A

4

2

waaA 2

Pi

KdoII

EtnPi Glycoform 2 O polysaccharide

3 L-RhaB 1

1

wapR* 3

D-GlcIII 1

6

Pi

Pi

Pi

Pi

6

4

D-GlcII

b-D-GlcI 1

wapH* 4

3 D-GalN 1 2

3

waaP*

HepII 1

wapG L-Ala

3

7

HepI 1

waaF*

Cm

5

KdoI 2

lipid A

4

2

waaC*

Pi

waaA 2

KdoII

EtnPi Fig. 4. Structures of core glycoform 1 and glycoform 2. Sugars have the a configuration unless otherwise indicated. Dotted lines indicate nonstoichiometric substitutions, or substitutions not present in all strains. Genes involved or putatively involved in enzymatic steps are indicated and an asterisk denotes the existence of some experimental evidence for the functions of these genes. Ala, alanine; Cm, carbamoyl; Etn, ethanolamine; GalN, 2-amino-2-deoxy-galactose; Glc, glucose; Hep, L-glycero-D-manno-heptose; Kdo, 3-deoxy-D-manno-oct-2-ulosonic acid; Rha, rhamnose.

Phosphorylation

It is very commonly observed that core phosphorylation patterns differ between P. aeruginosa strains. Some of this variation is probably due to genuine strain-to-strain differences, but it should be borne in mind that chemical treatment of LPS prior to analysis may cause underestimation of the true phosphorylation state. Acid hydrolysis is a common procedure used prior to structural analysis of LPS oligosaccharide domains, because it removes lipids and makes the molecules readily watersoluble. However, some pyrophosphate bonds may be hydrolysed in the process.94 Each phosphorylation site of the inner-core heptose residues may be occupied by monophosphate, diphosphate or even triphosphate substitutions.90,94,96 A rare phosphate substitution was found at position 4 of HepII in a P. aeruginosa cystic fibrosis isolate (the strain 2192).88 Most of the analyzed inner cores of P. aeruginosa also contain phosphoethanolamine or diphosphoethanolamine, present in non-stoichiometric amounts.94 Early studies on phosphorylation of LPS showed that even ethanolamine triphosphate might be present in LPS.97,98 In some strains, no ethanolamine could be detected in the core; these include strain 170021 (serotype O4), an algC mutant (derived from serotype O3) which has a truncation within the outer core, a wbjE mutant (derived from

serotype O11) which lacks O Antigen and produces several truncated core structures, and strain 2192, which is a cystic fibrosis isolate.90,94,99 Addition of GlcIV

Not all strains add the fourth glucose (GlcIV) to glycoform 1(O). In studies which determined outer core structures of different P. aeruginosa strains by NMR and MS, most of the analyzed serotypes had three Glc residues in the glycoform 1(O) core (serotypes: O1, O3, O4, O6, O9, O11, O12, O13, O15, O17, O19).88,89,94,100 The O10 serotype contained only the four-Glc glycoform 1(O) core and in the serotype O8 and O2 strains, the major glycoform contained four Glc residues and a three-Glc core was present as a minor species.94 These data from full structural determination of core structures correlate well with earlier immunochemical characterisation of core oligosaccharides by ELISA (enzymelinked immunosorbent assay) and Western blotting using an anti-glycoform 1(O) core monoclonal antibody (5C-101) that recognizes the terminal GlcIV.101 Several IATS reference strains reacted with mAb 5C-101, including O2, O5, O7, O8, O10, O16, O18, O19, and O20; while LPS from O1, O3, O4, O6, O9, O11, O12, O13, O14, O15, O17 reference strains were not recognised by this antibody.101


272


Acetylation

Acetylation of outer-core sugars is relatively common in P. aeruginosa strains and up to five O-acetyl groups have been found. Similar to phosphate and ethanolamine pyrophosphates, O-acetyl groups might be lost during mild acid delipidation of LPS prior to structural analysis; thus, the O-acetylation of core may be underestimated.45 The extent of core acetylation detected in different IATS reference strains was as follows: no acetylation was found in O4, O5, and O10; one acetylation site in O8, O12, and O13; at least one acetylation site is present in O6; one or two acetylation sites in O1 and O19; three acetylation sites in O3; four acetylation sites in a cystic fibrosis isolate; and up to 5 acetylation sites in O9, O11, O15, and O17.88,89,95 Although the O-acetyl groups are randomly distributed among sugars of the outer core, the most frequently acetylated residue is the terminal RhaA of glycoform 1(O).95 O-Acetylation occurs to a much higher extent when RhaA is not substituted with GlcIV or if RhaB in glycoform 2 is not substituted with O Antigen, for example in clinical isolates that have lost O Antigen expression.88,89,95 Unlike the wild type, PAO1, some strains that have truncated outer core such as wapR and rmlC mutants or a migA rmd double mutant, were found to have an O-acetyl group on terminal GlcII residue.102 The functions of these genes with respect to core biosynthesis are discussed below. The wbjE mutant of strain PA103 (serotype O11) produced various truncated forms of outer core, some of which had acetylated GalN, instead of the N-linked alanyl group.99 Similarly, the P. aeruginosa H4 strain that produces a truncated core, also had the terminal GalN acetylated rather than modified with the alanyl group.87

crude extracts of the P. aeruginosa algC mutant, and the algC gene cloned from P. aeruginosa complemented a pgm mutant of E. coli, restoring its ability to ferment galactose, which is dependent on the presence of PGM activity.44 The algC mutants of P. aeruginosa PAO1 (serotype O5) and PAC1R (serotype O3) produced truncated core structures missing all Glc and Rha residues as determined by NMR and MS.90 Conversion of Glc-1-phosphate to UDP-D-Glc, the proposed nucleotide-activated precursor for core Glc OH O

HO HO

OH

OH

D-Glc 9 + ATP

Glk Pi

O

HO HO

10

OH

OH

AlgC OH

OH

GalU

O

HO HO

+ UTP

OH

11

RmlA

Pi

O

HO HO

OH

UDP-D-Glc 12

+ dTTP

UDP

OH O

HO HO

Biosynthesis of core sugars

OH

The predominant sugar in the outer core oligosaccharide is D-Glc. In E. coli and other bacteria, glucose is transported as glucose-6-phosphate (10) (Fig. 5).103 Otherwise, glucose can be converted to glucose-6phosphate by Glk, a glucokinase. The glk gene function was demonstrated by measurements of Glk enzyme activity in cell lysates from a glk mutant and from complemented strains harbouring glk on plasmids.104 The next step in this pathway is conversion of glucose6-phosphate to glucose-1-phosphate (11) catalyzed by AlgC (Fig. 5). As mentioned earlier, the algC gene encodes a bifunctional enzyme that has phosphoglucomutase (PGM) and phosphomannomutase (PMM) activities. The PGM activity converts glucose-6-phosphate to glucose-1-phosphate,44 which is a common intermediate between the biosynthetic pathways for D-Glc and L-Rha. Phosphoglucomutase activity could not be detected in

dTDP

13

Biosynthesis of D-glucose (D-Glc): glk, algC and galU RmlB O

O HO

14

OH dTDP

RmlC dTDP

O

RmlD

dTDP

O HO

O

+ NADPH

HO

OH

15

HO

OH

dTDP-L-Rha 16

Fig. 5. Proposed UDP-D-glucose and dTDP-L-rhamnose biosynthesis pathways. The phosphoglucomutase activity of AlgC is required for the biosynthesis of both UDP-D-Glc (12) and dTDP-L-Rha (16). Pi, phosphate. 10, Glc-6-Pi; 11, Glc-1-Pi; 13, dTDP-D-Glc; 14, dTDP6-deoxy-D-xylo-4-hexulose; 15, dTDP-6-deoxy-L-lyxo-4-hexulose.


Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa residues, is catalyzed by GalU, a UDP-D-glucose pyrophosphorylase (Fig. 5). A galU mutant of PA103 (serogroup O11) was constructed and showed truncation in the core oligosaccharide and absence of O polysaccharide as judged by silver staining.105 By NMR analysis, the galU mutant core oligosaccharide was shown to lack all Glc and Rha residues.96 Transforming the mutants with the galU gene in trans restored O-polysaccharide production, but did not restore the wild-type mobility of lipid A-core bands in SDS-PAGE for reasons that are unknown.105 Biosynthesis of L-rhamnose (L-Rha): rmlA, rmlB, rmlC and rmlD

Synthesis of dTDP-L-Rha, the core L-Rha precursor, has been characterised at the biochemical level in other bacteria, such as S. enterica sv. Typhimurium.106,107 Accordingly, conversion of Glc-1-phosphate and dTTP to dTDP-L-Rha requires four reaction steps catalyzed by RmlA (glucose-1-phosphate thymidylyltransferase), RmlB (dTDP-D-glucose 4,6-dehydratase), RmlC (dTDP-6-deoxy-D-xylo-4-hexulose 30 ,50 -epimerase) and RmlD (dTDP-6-deoxy-L-lyxo-4-hexulose reductase; Fig. 5). The P. aeruginosa genome contains rmlB, rmlD, rmlA and rmlC in this order in the PA5161–PA5164 gene cluster (Table 2, Fig. S1). The three-dimensional

structure of RmlA from P. aeruginosa as well as the enzymatic mechanism of this protein has been determined.108,109 There are clear structural relationships between RmlA and SpsA, a glycosyltransferase from Bacillus subtilis110 and the N-terminal domain of GlmU, N-acetylglucosamine-1-phosphate uridyltransferase. Based on the contacts with substrates and absolute sequence conservation amongst sugar nucleotidyltransferases, several catalytically important residues have been identified (Arg15, Lys25, Asp110, Lys162 and Asp225). Pseudomonas aeruginosa RmlA exhibits feedback inhibition by dTDP-L-Rha, the final product of the RmlABCD pathway.111 The structure of P. aeruginosa RmlC in complex with product and product mimics has been recently solved.112 Although RmlC enzymes from different organisms have considerable diversity in sequence, they seem to share a common enzyme mechanism. Dong and co-workers112 suggested that the order of the double epimerisation (C3 and C5) of the substrate (dTDP-6-deoxy-D-xylo-4hexulose; 14) catalysed by RmlC of P. aeruginosa is first C5 followed by C3. The structural and biochemical data were consistent with a mechanism in which His65 and Tyr134 of P. aeruginosa RmlC are catalytic residues involved in the epimerisation steps.112 The role of rmlBDAC in LPS biosynthesis in P. aeruginosa was confirmed by genetic experiments.

Table 2. Genes involved or putatively involved in the biosynthesis of core oligosaccharide sugars Gene


eno/PA3635 pyrG/PA3637 kdsA/PA3636 kdsB/PA2979 kdsC/PA4458 kdsD/PA4457

76% – – 54% 48% 54%

gmhA/PA4425 gmhB/PA0006 hldE/PA4996

41% E. coli GmhA 35% E. coli GmhB 57% E. coli HldE

hldD/PA3337

53% E. coli HldD

glk/PA3193 algC/PA5322 galU/PA2023

– – –

rmlA/PA5163 rmlB/PA5161 rmlC/PA5164 rmlD/PA5162

– 61% S. enterica RmlB – 52% S. enterica RmlD

E. coli Eno

E. coli KdsB E. coli KdsC E. coli KdsD

273

Proposed/demonstrated function CMP-Kdo biosynthesis Enolase CTP synthase Kdo-8-phosphate synthase CMP-Kdo synthetase Kdo-8-phosphate phosphatase D-Arabinose 5-phosphate isomerase ADP-L-b-D-heptose biosynthesis Sedoheptulose 7-phosphate isomerase D-a,b-D-heptose 1,7-bisphosphate phosphatase D-b-D-heptose 7-phosphate kinase/ D-b-D-heptose 1-phosphate adenylyltransferase ADP-D-b-D-heptose epimerase UDP-D-glucose biosynthesis Glucokinase Phosphoglucomutase/phosphomannomutase UDP-D-glucose pyrophosphorylase dTDP-L-rhamnose biosynthesis Glucose-1-phosphate thymidylyltransferase dTDP-D-glucose 4,6-dehydratase dTDP-6-deoxy-D-xylo-4-hexulose 30 ,50 -epimerase dTDP-6-deoxy-L-lyxo-4-hexulose reductase

*References are cited if they describe experimental investigation of the P. aeruginosa gene. Downloaded from http://ini.sagepub.com at UNIV OF GUELPH on October 6, 2009

Key reference*

115 115

104 44 105 108 112

274


A non-polar mutation of rmlC in P. aeruginosa abrogates the production of complete glycoform 1(O) core (not recognised by the 5C-101 antibody) as well as the addition of O polysaccharides to glycoform 2 (not recognised by the anti-O polysaccharide antibodies).113 NMR and MS analyses of rmlC mutants showed complete absence of Rha residues within the outer core.102,113 Biosynthesis of Kdo: kdsD, kdsA, kdsC and kdsB

In E. coli, the biosynthesis of CMP-Kdo, the precursor for Kdo residues in LPS, involves four sequentiallyacting enzymes (Fig. 6): KdsD, a D-arabinose 5-phosphate isomerase; KdsA, a 3-deoxy-D-mannooctulosonate 8-phosphate (Kdo 8-P) synthase that utilises D-arabinose 5-phosphate (A5P; 17) and phosphoenolpyruvate (PEP) as substrates; KdsC, a Kdo 8-phosphate phosphatase; and KdsB, a cytidine 50 monophosphate-Kdo synthetase (reviewed by Raetz114). Our group showed that in P. aeruginosa, PA3636 encodes a functional homologue of KdsA.115 This kdsA gene (PA3636) was able to rescue growth of S. enterica sv. Typhimurium strain with a temperature-sensitive mutation in kdsA at 42 C. Moreover, the crude cell lysate from an E. coli strain overexpressing kdsA from P. aeruginosa catalysed the formation of Kdo-8-phosphate from arabinose-5-phosphate and phosphoenolpyruvate (PEP).115 The kdsA gene is located within a gene cluster containing pyrG and eno. Eno (enolase) is structurally conserved in organisms which metabolise sugars and catalyses the formation of PEP from phosphoglycerate during glycolysis.116 PyrG (CTP synthase) catalyzes the

transfer of ammonia to UTP to form cytidine triphosphate (CTP)117 that is required for the CMP-Kdo synthesis step.118 The pyrG gene of P. aeruginosa was able to complement growth of the E. coli pyrG mutant that was auxotrophic for cytidine, on a minimal medium lacking cytidine.115 A KdsB homologue is present in P. aeruginosa, encoded by PA2979. Its function has not been proven, but it shares 54% identity with characterized E. coli KdsB. KdsC and KdsD homologues are encoded by adjacent ORFs in the P. aeruginosa genome, PA4458 and PA4457, respectively. These two genes from P. aeruginosa have not been characterised, but their products share 48% and 54% identity to the characterised E. coli homologues. Biosynthesis of L-glycero-D-manno-heptose (Hep): gmhA, hldE, gmhB and hldD

In E. coli, ADP-L-glycero-b-D-manno-heptose (ADPL-D-Hep) is the substrate for heptosyltransferase reactions.119 The initial substrate for ADP-L-D-Hep biosynthesis is D-sedoheptulose-7-phosphate (Fig. 7; 21) that is an intermediate in the pentose phosphate pathway.120 Four enzymes catalyse the biosynthesis ADP-L-D-Hep from D-sedoheptulose-7-phosphate (Fig. 7): (i) GmhA, a sedoheptulose-7-phosphate isomerase: (ii) HldE, (formerly RfaE), a bifunctional enzyme that functions as a D-D-heptose-7-phosphate kinase and a D-D-heptose 1phosphate adenylyltransferase; (iii) GmhB, a D-D-heptose1,7-bisphosphate phosphatase; and (iv) HldD (formerly RfaD), an ADP-D-b-D-heptose epimerase (Fig. 7).121,122

CH2OH O OH OH Pi

Ribulose–5–Pi KdsD

Pi

O HO

OH

OH

OH

KdsA

17

–Pi

Pi

HO

O

HO

OH

O Pi

O

OH

OH

KdsC

18

–Pi

OH

HO

OH

KdsB

O

HO

OH

OH

O

OH

Kdo 19

+ CTP

HO

OH O

HO

CMP O

OH

CMP–Kdo 20

PEP Fig. 6. Proposed CMP-Kdo biosynthesis pathway. Pi, phosphate. PEP, phosphoenolpyruvate. 17, D-arabinose-5-Pi; 18, 3-deoxy-D-manno-oct-2-ulosonate 8-Pi. Downloaded from http://ini.sagepub.com at UNIV OF GUELPH on October 6, 2009

Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa None of the enzymes of the ADP-L-D-Hep pathway has been characterised in P. aeruginosa, but homologues of the ADP-L-D-heptose synthesis genes from E. coli are scattered though the P. aeruginosa genome (Table 2).

occur within each of the families. Glycosyltransferases use either an inverting or a retaining mechanism, in which the stereochemistry of the substrate sugar’s anomeric bond is inverted or retained in the transfer reaction. Frequently, members of a given GT family have the same type of mechanism so that, in theory, the CAZy family classification is useful for discriminating the possible functions of uncharacterised glycosyltransferases. However, there can be exceptions and, in at least one example, a glycosyltransferase that possessed the closest homology to inverting GT-2 family was found to be mechanistically a retaining glycosyltransferase,124 showing that GT family classification does not perfectly predict glycosyltransferase mechanism.

Location of transferase genes involved in the core oligosaccharide biosynthesis

Most of the transferase genes involved in assembly of the core are located in a large gene cluster that contains 17 ORFs (PA4996–PA5012) (Fig. S1). The 50 -end of the cluster is terminated by PA4996/hldE, which is likely involved in the biosynthesis of heptose (as mentioned in the previous biosynthetic section). The locus also contains two genes (msbA and waaL) involved in LPS transport and ligation of O polysaccharide to core (see below). Organisation and sequence (77–100% amino acid identity) of the core gene cluster is well conserved among different, sequenced P. aeruginosa strains. This is not unexpected, since the composition of the core is also conserved among various P. aeruginosa strains. Pseudomonas aeruginosa mutants lacking inner core Hep or phosphates have never been isolated or constructed, suggesting that Hep-linked phosphates are essential for P. aeruginosa viability. Thus, transferase genes involved in the assembly of inner core biosynthesis have been identified by cross-complementation experiments of Salmonella or E. coli mutants instead of creating knock-outs in P. aeruginosa.

Kdo transferase: waaA

The biochemical activity of the Kdo transferase has not been characterised in P. aeruginosa, but PA4988 encodes a homologue of WaaA, which catalyzes the sequential addition of two Kdo sugars to a molecule of lipid A in E. coli.125 PA4988 is close to, but outside of, the large core biosynthesis gene cluster and the PA4988 gene product shares 55% identity to E. coli WaaA. While the E. coli WaaA contains a transmembrane domain,126 no transmembrane domain was predicted for the P. aeruginosa homologue, by topology prediction programs such as TMHMM or SOSUI.127,128 Whether this reflects a genuine difference between the E. coli and P. aeruginosa WaaA proteins is unknown. PA4988 is classified as a member of GT-30 family, members of which are predicted to use an inverting mechanism. This is consistent with the glycosidic linkages found in the core structure.

Glycosyltransferase-encoding genes Numerous glycosyltransferases are required for core biosynthesis (Table 3). Glycosyltransferases have been classified into 91 different families on the basis of BLAST analysis and they are listed in the regularly updated CAZy database (5http://www.cazy.org4),123 and the same three-dimensional fold is expected to Pi HO HO HO

O

GmhA

OH OH

sedoheptulose-7-phosphate 21

HO HO HO

Pi OH O

22

Heptosyltransferases: waaC and waaF

Pseudomonas aeruginosa core oligosaccharide contains heptose residues that have identical linkages to HepI and

HldE OH

275

+ AT P

HO HO HO

Pi OH O

GmhB Pi

23

HO HO HO

OH OH O

– Pi

HldD ADP

ADP-L-glycero-β-D-manno-heptose

HO HO HO

OH OH O Pi

24 HldE

+ ATP

HO

OH OH O

HO HO

ADP

25

26 Fig. 7. Proposed ADP-L-glycero-b-D-manno-heptose biosynthesis pathway. Pi, phosphate. 22, D-glycero-D-manno-heptose-7-Pi; 23, D-glycero-b-D-mannoheptose-1,7-diphosphate; 24, D-glycero-b-D-manno-heptose-1-Pi; 25, ADP-D-glycero-b-D-manno-heptose. Downloaded from http://ini.sagepub.com at UNIV OF GUELPH on October 6, 2009

276


Table 3. Genes involved or putatively involved in core biosynthesis Gene

Related proteins/domains (% identity)

Proposed/demonstrated function

waaA/PA4988 waaC/PA5011 waaF/PA5012 migA/PA0705 wapR/PA5000 wapG/PA5010 wapH/PA5004 PA5001 wapQ/PA5007 wapP/PA5008 waaP/PA5009 PA5006

55% E. coli WaaA – – – – 51% E. coli WaaG 32% E. coli WaaG 24% WaaG GT domain – – – 30% APH/ChoK kinase domain 28% LPS kinase domain 25% APH/ChoK kinase domain 55% Pseudomonas putida Ttg8 31% Streptomyces clavuligerus CmcH 24% PIG-L domain 26% S. enterica sv. Typhimurium Mig-14

Glycosyltransferase (GT-30) of Kdo Glycosyltransferase (GT-9) of HepI Glycosyltransferase (GT-9) of HepII Glycosyltransferase (GT-2) of RhaA Glycosyltransferase (GT-2) of RhaB Glycosyltransferase (GT-4) possibly of GalN Glycosyltransferase (GT-4) possibly of GlcII Glycosyltransferase (GT-4) Heptose kinase Heptose kinase Heptose kinase: position 4 of HepI Kinase

PA4998 wapO/PA5005 PA5002 PA5003

Key reference*

131 131 102 102 40 82 82 141

Kinase Carbamoyltransferase Unknown Unknown Unknown

Glycosyltransferase (GT) family classification is taken from the CAZy database.123 *References are cited if they describe experimental investigation of the P. aeruginosa gene.

HepII present in Salmonella core oligosaccharide. In Salmonella, these residues are transferred by two enzymes, heptosyltransferase I and heptosyltransferase II, encoded by the waaC and waaF genes, respectively.129,130 The P. aeruginosa waaC (PA5011) and waaF (PA5012) possess 56% and 54% amino acid identity, respectively, to their Salmonella homologues and complemented S. enterica sv. Typhimurium waaC and waaF mutants.131 The complemented strains regained the ability to produce O polysaccharide and lost sensitivity to phage FFM phage, which is specific for deep-rough Salmonella strains (having LPS with severe defect in core oligosaccharide). Attempts to construct knockout mutants of waaC and waaF in P. aeruginosa were unsuccessful, despite using a method that has proven to be effective for making mutants in other P. aeruginosa genes. This indicates that these genes may be essential for cellular viability.131 Despite carrying out similar functions, P. aeruginosa WaaC and WaaF have a limited level of identity to each other (24%). Both enzymes are classified as GT-9 family members and are, therefore, predicted to have an inverting mechanism of catalysis. This is consistent with the core oligosaccharide structure. Possible GalN-transferase: wapG

WapG, encoded by PA5010, shares 51% identity with WaaG, which transfers the first sugar of the outer core in E. coli, the D-Glc which is (1 ! 3)-linked to HepII.132 In P. aeruginosa, the first outer core sugar is the

(1 ! 3)-linked N-(L-alanyl)-D-GalN. Because wapG is the closest waaG homologue in the P. aeruginosa genome, it has been suggested that WapG is responsible for addition of the GalN residue. There are also conserved wapG homologues in P. fluorescens, P. stutzeri and P. syringae genomes, supporting its hypothetical role as a GalN transferase, since the GalN residues are also present in the outer cores of these three Pseudomonas species.133–135 Although many different strategies were tried, a wapG mutant in P. aeruginosa was not obtained (M.J. Matewish and J.S. Lam, unpublished observations). Therefore, there is so far no biochemical or genetic evidence for the proposed function of WapG in outer core biosynthesis. The structure of the sugar-nucleotide precursor of GalN in the core is not known. In Klebsiella pneumoniae, the core contains a similar aminohexose, 2-amino2-deoxy-D-glucose, (GlcN). Since UDP-D-GlcN was not found in cell lysates, it is believed that the K. pneumoniae GlcN residue is derived from an initial transfer of D-GlcNAc catalysed by WabH which uses UDP-D-GlcNAc as substrate.136 This sugar is then subsequently deacetylated, on the nascent core oligosaccharide, in a reaction catalysed by the WabN enzyme.137 A similar scheme is conceptually possible for the first outer core sugar in P. aeruginosa; however, it should be noted that, although WapG is a homologue of K. pneumoniae WabH (23% identity), this is much lower sequence conservation than that shared with E. coli WaaG (51%) and thus far we have been unable


Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa to identify a wabN homologue in the P. aeruginosa genome. In any case, WapG is classified as a member of CAZy family GT-4 and the predicted retaining mechanism would be consistent with the proposed role of adding the a-D-GalN residue from a likely nucleotidea-D-sugar precursor. Possible GlcII-transferase: wapH

WapH, encoded by PA5004, is another core glycosyltransferase. It shares 32% identity with WaaG of E. coli and is a member of the GT-4 family of retaining glycosyltransferases. A non-polar knockout mutant of wapH produced incomplete core oligosaccharide showing that WapH participates in LPS core assembly.40 The mutant did not synthesize glycoform 1(O), because the LPS core was not recognised by the 5C-101 antibody. Inactivation of wapH also influenced synthesis of the core glycoform 2(Oþ), which was shown by the complete loss of Common Polysaccharide Antigen and low molecular mass forms of the O-Specific Antigen from LPS, including the lipid A-core þ 1 form. However, the mutant was not completely devoid of O polysaccharide and still produced some high molecular mass O Antigen. Based upon the SDS-PAGE migration of wapH LPS bands in comparison with LPS from wild-type and other truncated core mutants, it was suggested that the wapH gene might code for a glycosyltransferase that adds GlcII in (1 ! 4)-linkage to N-(L-alanyl)-galactosamine. Absence of GlcII would preclude addition of L-RhaA and GlcIV, and may make the growing oligosaccharide a weak acceptor for addition of L-RhaB which is necessary for attachment of O polysaccharides. The predicted retaining mechanism of WapH would be in agreement with transfer of GlcII to the core by WapH. Another piece of evidence that indirectly supports the proposed role of WapH in P. aeruginosa core biosynthesis is the absence of close wapH homologues in the P. stutzeri genome, since the core of a P. stutzeri strain does not contain (1 ! 4)-linked D-glucose.134 The core structure of the wapH mutant has not been determined by NMR and MS, so the proposed WapH function remains hypothetical. L-Rhamnosyltransferases: migA and wapR

Two genes encoding putative rhamnosyltransferases, wapR, and migA, are involved in the outer core oligosaccharide biosynthesis. They share 35% identity with each other and are both members of the CAZy family GT-2. The predicted inverting mechanism based on this CAZy classification is consistent with the use of dTDP-b-L-Rha as sugar donor to form either the a-LRhaA or a-L-RhaB in P. aeruginosa core. MigA is thought to catalyse formation of the (1 ! 6)-linkage of A L-Rha in core glycoform 1(O) and WapR to catalyse

277

formation of the (1 ! 3)-linkage of L-RhaB in glycoform 2(Oþ). Their function was characterised by genetic experiments and substantiated by elucidation of the chemical structure of core prepared from mutant strains. Interruption of the migA gene led to the truncation of core glycoform 1(O) and the core was not recognised by mAb 5C-101.102,138 NMR and MS analyses of LPS from migA mutants showed the presence of core glycoform 2(Oþ) (with RhaB), ruling out the possibility that MigA catalyses formation of the (1 ! 3)-bond of RhaB or addition of any other core glycoform 2(Oþ) sugars (GalN, GlcI, GlcII, GlcIII). Neither anti-O-polysaccharide or anti core þ 1 monoclonal antibodies reacted with LPS from the wapR mutant, showing that the WapR function is required for O-polysaccharide attachment. Data from NMR and MS analyses of LPS from the wapR mutant showed the presence of glycoform 1(O) core structure (with L-RhaA), ruling out the possibility that WapR is responsible for the formation of the (1 ! 6)-L-RhaA linkage or addition of any other core glycoform 1(O) sugars (GalN, GlcI, GlcII, GlcIII, GlcIV).102 Since both enzymes are suggested to utilise the same dTDP-L-Rha donor and nascent lipid A-core acceptor, a certain level of competition for substrates might be expected to exist between these two proteins, which could determine the ratio between glycoform 1(O) and glycoform 2(Oþ) LPS molecules. Complementation of the migA mutant phenotype by introducing a functional gene carried on a multicopy plasmid also led to loss of core þ 1 LPS, as assayed by Western blotting. This observation might support the MigA/WapR competition hypothesis, since overproduction of MigA due to the multicopy plasmid likely causes rapid consumption of substrates for the WapR transferase reaction, which is critical for the attachment of one repeat unit of O-Antigen. This provides evidence that regulation of migA or wapR expression may be a mechanism used by P. aeruginosa to determine the ratio of core glycoform 1(O) and glycoform 2(Oþ) produced. At present, nothing is known about regulation of wapR expression, but the migA gene was initially described as a mucus-inducible gene, up-regulated in the presence of cystic fibrosis mucus.139 In a separate study, it was shown by Northern blotting with RNA collected from cystic fibrosis sputum that migA is expressed in this environment.138 The transcription start of migA was identified by primer extension, and it did not change in the presence or absence of cystic fibrosis sputum. A lacZ promoter fusion showed that the expression of migA is cell-density dependent, which is characteristic of genes under the control of quorum sensing. Furthermore, las box-like sequences, (CT-(N11)-AG), recognition sites for quorum sensing regulation were found upstream of the migA gene. Two quorum sensing systems in P. aeruginosa were known at the time of these experiments, RhlI–RhlR


278


and LasI–LasR. Each is composed of two components the first of which is an inducer locus, lasI or rhlI, that controls the synthesis of the diffusible auto-inducer molecules N-3-oxo-dodecanoyl-homoserine lactone (3OC12-HSL) or N-butanoyl homoserine lactone (C4HSL), respectively. The second is a response locus, lasR or rhlR, that encode transcription factors which become active upon binding to the respective auto-inducers. At high cell density, migA expression levels were much lower in either the rhlI or lasI rhlI double mutant background than in wild-type strain, suggesting that migA is regulated by the RhlI–RhlR system. Furthermore, migA expression in a lasI rhlI double mutant background was increased by the addition of C4HSL, an auto-inducer that positively stimulates the RhlI/ RhlR system.138 Putative glycosyl/glucosyltransferase: PA5001

PA5001 encodes a putative glycosyltransferase sharing 24% identity with a conserved glycosyltranferase domain of WaaG proteins, which are classified into the GT-4 family. It has not been previously suggested which enzyme might be responsible for the transfer of a-GlcIII, b-GlcI or b-GlcIV. The GT-4 family contains retaining glycosyltransferases and thus, theoretically, PA5001 might be responsible for transfer of a-GlcIII, but this prediction must be tested experimentally. Addition of phosphates to core: waaP, wapP, and wapQ

Three genes encoding the inner core kinases and a putative kinase have been identified – waaP (PA5009), wapP (PA5008) and wapQ (PA5007).82 WaaP was recognised as a putative heptose kinase due to a high (53%) sequence identity to the inner core heptose kinase WaaP of E. coli (Table 3). Although WapP and WapQ are not homologues of WaaP from E. coli or S. enterica sv. Typhimurium, they contain conserved kinase domains and were hypothesised to be involved in the inner core oligosaccharide heptose phosphorylation.82 Mutants in waaP and wapP genes could not be constructed, indicating that the genes may be essential, but the authors were able to create the wapQ mutant. The wapQ mutant did not exhibit any particular change in LPS phenotype based on silver-stained SDS-PAGE gels and Western immunoblotting analyses, and results from NMR analysis of the wapQ core structure were inconclusive. However, this mutant exhibited a reduced MIC (minimal inhibition concentration) for sodium dodecylsulphate (SDS), when compared with wildtype. In general, mutants affected in phosphorylation of inner core have a destabilized outer membrane that leads to a higher susceptibility to SDS and a higher permeability to hydrophobic compounds such as novobiocin. Thus, the SDS susceptibility data suggested that the

wapQ mutant might have a subtle core phosphorylation defect, which was not readily apparent in other analyses. However, the wapQ function as a kinase has not been proven. waaP from P. aeruginosa (waaPPa) complemented a S. enterica sv. Typhimurium waaP mutant as assayed by NMR analysis; the complemented mutant restored phosphate and ethanolamine diphosphate substitutions at the 4 position of HepI. Mutation of the chromosomal waaP was only achieved in P. aeruginosa when a copy of the wild-type gene was present in trans, indicating that the mutation is probably lethal. The viability of the P. aeruginosa mutant could also be maintained by a plasmid containing the E. coli waaP gene. These data suggest that WaaP of P. aeruginosa, like its homologues from E. coli and S. enterica sv. Typhimurium, catalyses transfer of phosphate to the 4 position of HepI. Alignment of WaaPPa with eukaryotic protein kinases, such as the a-form of bovine brain protein kinase C (PKC-a), reveals potential conserved residues, among them, Lys69 being a putative catalytic residue involved in the proton transfer in the phosphotransfer reaction and Asp163 being putatively involved in the ATP-binding. The site-directed mutants K69A and D163A were constructed, cloned into a vector and tested for ability to complement the E. coli waaP mutant in terms of its susceptibility to SDS and novobiocin. Neither of these mutant constructs was able to restore the wild-type phenotype, indicating that the residues might be important for the function of WaaPPa. In addition, WaaPPa was shown by Western blotting to contain phosphorylated amino acids and data from MS analysis indicated that all eight tyrosine residues in this protein were phosphorylated. By developing a chemiluminescence-based ELISA method to determine the activity of WaaP, our group showed that this protein was able to autophosphorylate and also that WaaP has a sugar-kinase activity when given dephosphorylated LPS as a substrate.140,141 WaaP was localized to the cytoplasmic cell fraction indicating that core phosphorylation occurs before transport across the inner membrane and therefore before O-Antigen units are attached to the core in the periplasm. The function of WapP as a heptose kinase was first assigned indirectly, by showing that a S. enterica sv. Typhimurium waaP mutant complemented with wapP and waaPPa resulted in increased MICs for novobiocin and SDS above the Salmonella mutant complemented with waaPPa alone. Immunochemical evidence was obtained using a similar chemiluminescence-based ELISA method as that used to study WaaP, which showed that WapP could phosphorylate dephosphorylated LPS.142 Further evidence from MS and Western blotting analyses revealed that WapP contains phosphorylated serine consistent with a kinase function.142


Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa Addition of carbamoyl: wapO

Another putative transferase-encoding gene present in the core biosynthesis cluster is wapO (PA5005). WapO shares 31% amino acid sequence identity with CmcH (Table 3), an O-carbamoyltransferase from Streptomyces clavuligerus143 suggesting that WapO may catalyse transfer of the carbamoyl group found at the 7 position on HepII. Core locus genes encoding proteins with unknown functions

The product of the PA5006 gene has not yet been characterised, but the sequence contains a LPS kinase family conserved domain (Pfam06293), and an APH/ ChoK kinase family conserved domain (cd05120).144 The PA4998 gene product also contains a APH/ChoK conserved domain, and shares 55% identity with the toluene tolerance protein of P. putida, whose mechanism and function remain unclear. In general, substrates for the members of the APH/ChoK family are small molecule substrates such as ethanolamine.145 Therefore, it is tempting to speculate that the PA5006 or PA4998 gene product is an ethanolamine kinase involved in synthesis of the ethanolamine-phosphate moieties found in the inner core. The PA5002 protein contains a putative transmembrane domain and shares 24% identity with the PIG-L protein domain (Pfam02585) and the PA5003 gene product shares 26% identity with Mig-14 of S. enterica sv. Typhimurium, an inner-membrane associated virulence factor with uncharacterised function.146 The PA5006, PA5003, PA5002 genes with unknown functions are well conserved in all the genomes of P. aeruginosa strains that have been sequenced as well as in the genomes of all other Pseudomonas species that have been sequenced (P. entomophila, P. fluorescens, P. mendocina, P. putida, P. stutzeri, P. syringae) suggesting an important role in Pseudomonas biology. Unknown mechanisms of core modification Addition of diphosphoethanolamine

In E. coli, kinetic experiments indicated that phosphatidylethanolamine (PE) is the precursor of both the phosphate and the ethanolamine of phosphoethanolamine. It has been proposed that the diphosphoethanolamine substituent of LPS arises by transfer of the phosphoethanolamine moiety from PE to phosphate on HepI.147 As mentioned before, in P. aeruginosa ethanolamine can be part of phosphoethanolamine, diphosphoethanolamine or even triphosphoethanolamine. The steps leading to tri-, di- or monophosphoethanolamine on the P. aeruginosa core are presently unclear. There is an indication that WaaP may contribute to the substitution

279

of heptose with diphosphoethanolamine, since complementation of the S. enterica sv. Typhimurium waaP mutant by P. aeruginosa waaP restored both phosphate and diphosphoethanolamine on the inner core heptose.82 However, data from this experiment did not directly demonstrate the ability of WaaP to transfer di- or monophosphoethanolamine, since the diphosphoethanolamine in Salmonella is found at position 4 of HepI, whereas the diphosphoethanolamine position in P. aeruginosa core oligosaccharide is at the 2 position of HepI. In S. enterica sv. Typhimurium, the cptA gene is necessary for phosphoethanolamine addition to the inner core as demonstrated by analysis of mutant LPS by NMR and MS.148 However, there is no direct biochemical evidence of the CptA phosphoethanolamine transferase function. The P. aeruginosa genome contains three homologues of cptA – PA4517, PA1972 and PA3310 – which all encode proteins with four or five predicted transmembrane helices and a long hydrophilic loop at the C terminus, similar to CtpA. PA4517 is the closest homologue, and shares 59% amino acid identity with Salmonella CtpA. Transfer of Ala residue to GalN

The enzyme that is responsible for attachment of Ala to the GalN-residue has not been identified. Substitution of core oligosaccharide with an amino acid is not rare. Some bacteria, such as N. meningitidis, Haemophilus influenzae and Shigella flexneri,149–151 have core substituted with a glycine residue. The mechanism of the transfer is not known. A specific tRNA was identified that is a glycine donor and might be the substrate for an unknown transferase.152 However, no further study identifying the potential transferase or confirming role of a tRNA in the transfer has been published. O-Acetylation of core

The gene(s) responsible for O-acetylation of core have not been identified in P. aeruginosa. In H. influenzae, the HI0392 gene is required for O-acetylation of the core and was originally found as a homologue of S. enterica sv. Typhimurium O Antigen acetylase encoding gene oafA.153 In PAO1, the PA5238 gene is a H. influenzae HI0392 homologue encoding a protein with 30% identity. Transport of lipid A-core: msbA Compelling genetic data from studies in E. coli indicate that MsbA is an essential ABC (ATP-binding cassette) transporter which is required for export of lipid A across the inner membrane. When MsbA is disabled, newly synthesised lipid A accumulates on the cytoplasmic face of this membrane.154 Furthermore, single amino acid


280


substitutions in E. coli MsbA allow export of lipid A precursor molecules to the cell surface.155 In PAO1, msbA (PA4997) is localised to the large core cluster and encodes a homologue of the E. coli MsbA, sharing 40% sequence identity. Characterisation of purified P. aeruginosa MsbA, reconstituted into membranes, showed that this protein has intrinsic ATPase activity, which is stimulated by lipid A-core. The stimulatory effect was diminished when the experiment was performed with S. enterica sv. Typhimurium LPS, a P. aeruginosa lipid A-truncated core lacking L-Rha and IV D-Glc (see Fig. 4) or with a P. aeruginosa lipid A-core from which the phosphates had been removed by hydrofluoric acid solvolysis.156 These data are consistent with assignment of this protein as the LPS transporter, which catalyses ATP-dependent transport of lipidA-core across the inner membrane. The specificity observed for complete lipid-A core would prevent MsbA from prematurely exporting unfinished LPS molecules.

O POLYSACCHARIDES O Antigen serotyping schemes In the International Antigenic Typing System (IATS), P. aeruginosa strains are classified into 20 serotypes on the basis of serological reactivity of major heat-stable antigens.157,158 Other schemes exist, and may have certain advantages. In particular, the La`nyi–Bergan classification scheme emphasises serological relatedness of strains based on analysis using the classical precipitin method in gel, which detects identity, and partial identity (that is, common cross-reactivities and minor antigens) of antigenic epitopes between any two LPS types.159,160 The La`nyi–Bergan classification involves a smaller number (13) of serogroups, but most are subdivided, which results in a total of more than 31 distinct subtypes. The La`nyi–Bergan system is favoured by researchers analysing the chemical structures of the O polysaccharides, because this system exquisitely reflects structural relationships between O-Antigen repeat units.45 In this review, however, we will mainly use the IATS classification because the genetics of O-Antigen biosynthesis has been investigated exclusively with reference to this scheme. Some of the 20 IATS serotypes are sometimes collected into groups, which are related serologically,157,158 structurally45 and genetically.161 These are the O2/O5/O16/O18/O20, O7/O8, O10/O19 and O13/O14 serogroups. Structures of the O-Antigen repeat units produced by the 20 IATS type strains are shown in Table 4.

Common Polysaccharide Antigen and O-Specific Antigen O polysaccharides The terms ‘A band’ and ‘B band’, which have been used to describe these structures, were coined by Rivera and co-workers162 to describe LPS subpopulations which they identified in LPS from P. aeruginosa PAO1 derivatives. Gel filtration chromatography in a deoxycholate-containing buffer and extremely slow flow rate was able to resolve LPS fractions which when subjected to SDS-PAGE revealed two subpopulations of silver-stained LPS molecules. The upper (A) cluster of bands contained a polymer rich in neutral sugars which was not recognised by the O serotypespecific monoclonal antibody (anti-503). The lower (B) cluster of bands contained the serotype-specific epitope, and its higher electrophoretic mobility toward the anode is probably due to the negative charge of the PAO1 O-Antigen repeat. As mentioned above, the O-Specific Antigen and Common Antigen O polysaccharides are distinct structurally, serologically and according to the mechanisms of their biosynthesis. The Common O Polysaccharide, which is produced by most P. aeruginosa isolates, is a homopolymer of D-Rha. In contrast, the O-Specific Antigens contain different sugars that are organised into repetitive O units. Models of O-polysaccharide biosynthesis pathways All the available data suggest that the Common and O-Specific polysaccharides are assembled by different mechanisms according to the two most common models for O-polysaccharide assembly. Several studies of LPS biosynthesis in P. aeruginosa, for example de Kievit and Lam on Wzy,163 and Rochetta and Lam on Wzm and Wzt,164 have provided evidence to substantiate the models, both of which were proposed earlier by Whitfield.165 In brief, Common O-Polysaccharide synthesis follows the ‘ABC transporter-dependent’ pathway and O-Specific Antigen synthesis follows the ‘Wzydependent’ pathway. These models have been reviewed by Raetz and Whitfield.50 According to both models, O-polysaccharide sugars are assembled on an isoprenyl lipid carrier by cytoplasmic glycosyltransferases, and the completed O-polysaccharide chains are ligated to the lipid-A core in the periplasm. However, the two pathways use distinct mechanisms for the steps in-between, such as polymerisation of the O chain, and export of lipid-linked glycans across the inner membrane. In the ABC transporterdependent pathway, which is apparently followed by Common Polysaccharide Antigen biosynthesis, the O chain is fully assembled on the cytoplasmic side of the



281

Table 4. O-Antigen repeating units of the IATS reference strains IATS serotype

O-Antigen repeat

O1 O2 O5 O16 O18 O20 O3 O4 O6 O7 O8 O9 O10 O19 O11 O12 O13 O14 O15 O17

!4)-D-GalNAc-(1 ! 4)-b-D-GlcNAc3NAcA-D-(1 ! 3)-D-FucNAc-(1 ! 3)-D-QuiNAc-(1! !4)-b-D-ManNAc3NAmA-(1 ! 4)-L-GulNAc3NAcA-(1 ! 3)-b-D-FucNAc-(1! !4)-b-D-ManNAc3NAmA-(1 ! 4)-b-D-ManNAc3NAcA-(1 ! 3)-b-D-FucNAc-(1! !4)-b-D-ManNAc3NAmA-(1 ! 4)-b-D-ManNAc3NAcA-(1 ! 3)-b-D-FucNAc-(1! !4)-L-GulNAc3NAmA-(1 ! 4)-b-D-ManNAc3NAcA-(1 ! 3)-b-D-FucNAc-(1! !4)-L-GulNAc3NAmA-(1 ! 4)-b-D-ManNAc3NAcA-(1 ! 3)-b-D-FucNAc4OAc-(1! !2)-L-Rha3OAc-(1 ! 6)-D-GlcNAc-(1 ! 4)-L-GalNAc4OAcA-(1 ! 3)-b-D-QuiNAc4NSHb-(1! !2)-L-Rha-(1 ! 3)-L-FucNAc-(1 ! 3)-L-FucNAc-(1 ! 3)-D-QuiNAc-(1! !3)-L-Rha-(1 ! 4)-D-GalNAc3OAcAN-(1 ! 4)-D-GalNFoA-(1 ! 3)-D-QuiNAc-(1! !4)-Pse4OAc5NRHb7NFo-(2 ! 4)-b-D-Xyl-(1 ! 3)-D-FucNAc-(1! !4)-Pse4OAc5NAc7NFo-(2 ! 4)-b-D-Xyl-(1 ! 3)-D-FucNAc-(1! !30 )-Pse4OAc5NAc7NRHb-(2 ! 4)-D-FucNAc-(1 ! 3)-b-D-QuiNAc-(1!* !3)-L-Rha2OAc-(1 ! 4)-L-GalNAcA-(1 ! 3)-D-QuiNAc-(1! !3)-L-Rha-(1 ! 4)-L-GalNAcA-(1 ! 3)-D-QuiNAc-(1! !2)-b-D-Glc-(1 ! 3)-L-FucNAc-(1 ! 3)-D-FucNAc-(1! !8)-8eLeg5NAc7NAc-(2 ! 3)-L-FucNAm-(1 ! 3)-D-QuiNAc-(1! !2)-L-Rha-(1 ! 3)-L-Rha-(1 ! 4)-D-GalNAc3OAcA-(1 ! 3)-b-D-QuiNAc-(1! !2)-[D-Glc-1 ! 3)]-L-Rha-(1 ! 3)-L-Rha-(1 ! 4)-D-GalNAc3A-(1 ! 3)-b-D-QuiNAc-(1! !2)-b-D-Ribf-(1 ! 4)-D-GalNAc-(1! !4)-b-L-Rha-(1 ! 3)-D-ManNAc-(1!

Adapted from Knirel et al.45 O-Antigen structures are clustered to reflect structural relationships. Anomeric conformations are a unless marked b. Sugars have the pyranose form except ribose in O15. Ac, acetyl; 8eLeg, 5,7-diamino-3,5,7,9-tetradeoxy-Lglycero-D-galacto-non-2-ulosonic (8-epilegionaminic) acid; Fo, formyl; FucN, 2-amino-2,6-dideoxy-galactose; GalN, 2-amino2-deoxy-galactose; GlcNA, 2-amino-2-deoxy-glucuronic acid; GulNA, 2-amino-2-deoxy-guluronic acid; ManN 2-amino-2deoxy-mannose; ManNA, 2-amino-2-deoxy-mannuronic acid; N, amino; NAc, acetamido; NAm, acetamidino; OAc, O-acetyl; Pse, 5,7-diamino-3,5,7,9-tetradeoxy-L-glycero-L-manno-non-2-ulosonic (pseudaminic) acid; QuiN, 2-amino-2,6-dideoxyglucose; Rha, rhamnose; RHb (R)-3-hydroxybutanoyl; Rib, ribose; SHb, (S)-3-hydroxybutanoyl. In the polymer, QuiNAc is linked to the pseudaminic acid via the 3-hydroxybutanoyl group.

inner membrane and, after it is complete, is exported by an ABC transporter to the periplasmic face of the inner membrane. In the Wzy-dependent pathway, which is followed by O-Specific Antigen biosynthesis, the Orepeating units are individually flipped across the inner membrane and polymerised in the periplasm in a reaction which requires the product of the wzy gene.

Initiation of O Antigen biosynthesis: wbpL For both the Wzy-dependent and ABC transporterdependent pathways, initiation of O-polysaccharide synthesis is accomplished by transfer of a sugar-1-phosphate to undecaprenyl-phosphate. This reaction generates an undecaprenyl-pyrophosphoryl-linked glycan which can be extended by the action of downstream glycosyltransferases. The initiating sugar-1-phosphate transferases have been identified in S. enterica (WbaP, formerly RfbP) and E. coli (WecA, formerly Rfe). WbaP is a transferase that catalyses the transfer of galactose-1-phosphate from UDP-galactose to

undecaprenyl-phosphate.166 WecA transfers either 2acetamido-2-deoxy-glucose-1-phosphate (GlcNAc1-Pi)167 or 2-acetamido-2-deoxy-galactose-1-phosphate (GalNAc-1-Pi).168 The initiating sugar-1-phosphate transferase in P. aeruginosa is encoded by the wbpL gene which is in the wecA family of 2-acetamido-2-deoxyhexose-1-phosphate transferases. wbpL is located in the O-Specific Antigen biosynthesis gene cluster (described below), but is necessary for synthesis of both Common Polysaccharide Antigen and O-Specific Antigen. wbpL mutants made in O2, O5, O6 and O16 strains were all completely devoid of O polysaccharides.169,170 The structures of the biological repeating units have been elucidated for most of the P. aeruginosa IATS serotype O-Specific Antigens and each has a 2-acetamido-2-deoxy-D-fucose (D-FucNAc), a 2-acetamido-2-deoxy-D-quinovose (DQuiNAc), or a derivative of these at the reducing terminus.94 The presence of these different sugars at the reducing termini raised the question of the specificity of WbpL enzymes from different serotypes for the sugar substrate. Therefore, the ability of WbpL enzymes from


282


O5 and O6 strains to initiate O-Antigen synthesis using either of these two sugars was tested in cross-complementation experiments. Serotype O5 has D-FucNAc at the reducing terminus and O6 has D-QuiNAc. The wbpLO5 gene could fully complement O-Antigen biosynthesis in an O6 serotype wbpL mutant, showing that the WbpLO5 enzyme is capable of transferring either sugar-1-phosphate. On the other hand, the wbpLO6 gene restored O-Antigen biosynthesis in an O5 serotype wbpL mutant, but only at reduced levels.169 This suggests that the WbpLO6 enzyme has a preference for transferring QuiNAc-1-phosphate as opposed to FucNAc-1-phosphate. In two P. aeruginosa O serotypes (O1 and O9) both D-FucNAc and D-QuiNAc are present in the O-Antigen repeating unit, but the initiating sugar in both cases is D-QuiNAc suggesting that WbpLO1 and WbpLO9 also have a strong specificity for this sugar.94 Although the repeating unit of the Common Polysaccharide Antigen has been well characterised,171 the full structure of the Common O polysaccharide has not yet been determined (see below), hence, the identity of the initiating sugar involved in Common O-polysaccharide biosynthesis is unknown. In attempts to compare the substrate specificities of WbpL and E. coli WecA, a series of cross-complementation experiments was performed. Expression of wecA in trans restored Common O Polysaccharide, but was unable to restore O-Specific Antigen synthesis in O5or O6-background wbpL mutants.169,170 In the converse experiment, wbpL was able to restore synthesis of the K. pneumoniae O1 O-Antigen expression in an E. coli K-12 wecA mutant strain, but only at very reduced levels.170 These data indicate that wecA is not functional as a FucNAc-1-phosphate or QuiNAc-1-phosphate transferase, and that Common O-Polysaccharide and O Specific-Antigen synthesis are initiated with different sugars. Klebsiella pneumoniae O polysaccharides are synthesized with a GlcNAc residue at the reducing terminus,172 but it is not possible to discern from these experiments whether this is the same for P. aeruginosa Common Polysaccharide Antigen. Indeed, without knowing the substrate specificity of the downstream glycosyltransferase it is not possible to know from the available data whether polysaccharide synthesis is initiated using the same sugar in the cross-complementation experiments as is usually used in the wild-type. As an example to illustrate this problem, broad substrate specificity in both WecA and in the subsequent mannosyltransferase reaction led to the original assignment of WecA as a Glc-1-phosphate transferase on the basis of in vitro reactions.173,174 Later, however, WecA was shown to transfer GlcNAc-1-phosphate to undecaprenyl phosphate in vivo175,176 and to transfer GalNAc-1-phosphate to initiate synthesis of Yersinia enterocolitica O8 O Antigen synthesis in E. coli K-12.168

To establish the in vivo function of WbpL in initiation of Common O-polysaccharide biosynthesis it will probably be necessary to characterise the linkage of the Drhamnan to LPS core, as has been done for K. pneumoniae LPS,172 or to elucidate the full structure of the undecaprenol linked precursor. Composition and glycosyl linkage analysis of a Common Polysaccharide Antigen-containing preparation from a P. aeruginosa galU mutant (which is unable to add O polysaccharides to LPS due to truncation of the outer core) indicated the presence of 4-linked GlcNAc suggesting this may be the sugar initiating Common O polysaccharide.96 Other analyses of Common O-Polysaccharide preparations failed, however, to detect this sugar.171,177

COMMON O POLYSACCHARIDE The Common Polysaccharide Antigen is produced by the majority of P. aeruginosa isolates. Of the IATS type strains the serotype O7, O12, O13, O14, O15 and O16 strains lack Common O Polysaccharide in their LPS.33 Structure of Common O Polysaccharide The Common O Polysaccharide contains a trisaccharide chemical repeating unit: [ ! 3)-a-D-Rha-(1 ! 3)-a-DRha-(1 ! 2)-a-D-Rha-(1!]n. This rhamnan structure has been characterised several times with good agreement between different studies.96,171,177 However, monosaccharide composition analyses have identified small amounts of other sugars, with less consistency between reports. As well as rhamnan repeating unit and probable lipid A-core sugars, the following have been identified: 3-O-methyl-6-deoxyhexose and xylose;177 3-O-methylrhamnose, ribose, mannose and 3-O-methylhexose;171 or mannose, GlcNAc and small amounts of O- and N-acetylation.96 The disagreement between these reports may be due to genuine strain differences, the details of preparation, or to the sensitivity and accuracy of the methods employed. The Common Polysaccharide Antigen biosynthetic locus Functions required specifically for biosynthesis of the P. aeruginosa Common O Polysaccharide are encoded in the Common Polysaccharide Antigen biosynthesis gene cluster. Attention has previously focused exclusively on eight contiguous genes from rmd (PA5454) to wbpZ (PA5447) which are probably co-transcribed from a promoter upstream of rmd (Fig. S2).178 A cosmid clone (pFV3) which contained these genes was able to restore Common Polysaccharide Antigen expression to five of the six IATS type strains which otherwise lack


Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa this structure.179 While preparing this review, we analysed the region upstream of rmd and identified five more coding sequences which were not all present on pFV3 but are related to polysaccharide biosynthesis/ modification genes in other species and may, therefore, also function in the synthesis and processing of the Common O Polysaccharide. These five open reading frames, PA5455–PA5459, are apparently transcribed divergently from rmd (Table 5). It is also known that Common Polysaccharide Antigen production requires functions encoded outside of this 13-gene cluster. All of these genes are described below.

Biosynthesis of D-rhamnose (D-Rha) The Common Polysaccharide Antigen biosynthetic cluster contains three genes encoding functions in the biosynthesis of GDP-D-Rha, the sugar-nucleotide precursor for the Common O polysaccharide homopolymer. These genes code for WbpW, GMD and RMD. A fourth enzyme, AlgC which is encoded outside of this locus, also participates in this pathway (Fig. 8). wbpW

In E. coli, the biosynthesis of GDP-D-Man from D-fructose-6-phosphate (27) is encoded by three genes manA, manB, and manC. In P. aeruginosa, wbpW

283

encodes a bifunctional enzyme which catalyses the first and third (ManA- and ManC-catalysed) of these steps (Fig. 8). WbpW catalyses both D-mannose-6phosphate isomerase (phosphomannoisomerase, PMI) and GDP-D-Man pyrophosphorylase (GMP) reactions, which was demonstrated by the ability of wbpW to complement E. coli manA (PMI) and manC (GMP) mutants in K30 capsule biosynthesis. wbpW mutants express reduced quantities of Common Polysaccharide Antigen-containing LPS but, when cultures are grown for extended periods of time, it is possible to detect Common Polysaccharide Antigen production.178 Two wbpW functional homologues are present in the PAO1 genome and the presence of these genes probably accounts for the production of the Common O Polysaccharide in the wbpW mutant. The first homologue, algA (PA3551, 46% amino acid identity with WbpW) is located in the locus responsible for alginate biosynthesis and AlgA is also bifunctional.180 Mature alginate is derived from a homopolymer of mannuronic acid and GDP-D-Man is an intermediate in the alginate biosynthetic pathway (reviewed by Remminghorst and Rehm181). The provision of algA in trans is able to boost Common Polysaccharide Antigen production in an algA wbpW double mutant, a greater effect being seen with a multicopy algA-containing plasmid as opposed to a single copy algA chromosomal insertion.178 This indicates that, in these strains, Common Polysaccharide

Table 5. Genes involved or potentially involved in Common Polysaccharide Antigen biosynthesis Gene



Key reference*

wbpZ/PA5447 wbpY/PA5448 wbpX/PA5449

52% E. coli O9a WbdC 34% E. coli O9a WbdB 33% E. coli O9a WbdA C-terminal domain (over 301 amino acids) 25% E. coli O9a WbdA N-terminal domain (over 262 amino acids) 61% E. coli O8 Wzt 56% E. coli O8 Wzm 46% P. aeruginosa AlgA

Glycosyltransferase (GT-4) Glycosyltransferase (GT-4) Glycosyltransferase (GT-4)

170 170 170

ABC transporter ABC transporter D-Man-6-phosphate isomerase / GDP-D-Man pyrophosphorylase

164 164 178

GDP-D-Man 4,6-dehydratase GDP-D-Rha synthase Glycosyltransferase (GT-4) Glycosyltransferase (GT-4) Methyltransferase Acetyltransferase Methyltransferase Phosphomannomutase/phosphoglucomutase

187 187

wzt/PA5450 wzm/PA5451 wbpW/PA5452

gmd/PA5453 rmd/PA5454 PA5455 PA5456 PA5457 PA5458 PA5459 algC/PA5322

60% P. aeruginosa PslB 47% E. coli GMD 33% Aneurinibacillus thermoaerophilus RMD

20% E. coli O8 WbdD (over 149 amino acids) 24% Staphylococcus aureus OatA 23% E. coli O8 WbdD (over 139 amino acids) 31% E. coli ManB

Glycosyltransferase (GT) family classification is taken from the CAZy database.123 *References are cited if they describe experimental investigation of the P. aeruginosa gene. Downloaded from http://ini.sagepub.com at UNIV OF GUELPH on October 6, 2009

184


284

Pi Pi

OH

O HO OH

WbpW

HO HO

OH O

HO HO

OH

OH

D-fructose-6-phosphate

AlgC

OH OH O

WbpW Pi

28

HO HO

OH OH O

+ GTP

GDP

29

GDP-D -Man

27

30

Gmd

HO HO

OH O

O

Rmd GDP

GDP-D -Rha

OH O

HO

+ NADPH

31

GDP

32 Fig. 8. The GDP-D-rhamnose biosynthesis pathway. Except for the second step in this pathway, all of the enzymes are encoded in the Common Polysaccharide Antigen biosynthesis gene cluster. The gene encoding AlgC is in the alginate locus. Pi, phosphate; 28, mannose-6-Pi; 29, a-mannose-1-Pi; 31, GDP-4-keto-D-Rha.

Antigen synthesis is limited, in part, by the availability of the product of the reactions catalysed by these enzymes. The second wbpW homologue is pslB (PA2232). pslB, which was originally called ORF488, encodes a bifunctional enzyme that shares 60% sequence identity with WbpW,182 and is in the locus responsible for production of a cell surface polysaccharide known as Psl. Psl is a galactose- and mannose-rich exopolysaccharide;183 therefore, GDP-D-Man is also presumably a precursor for synthesis of this polymer. algC

The step interposed between the two reactions catalysed by WbpW/AlgA/PslB is encoded by an alginate-locus gene, algC. This appears to be the only gene in P. aeruginosa encoding a phosphomannomutase (PMM)184 as algC mutation results in loss of detectable cell-lysate PMM activity. As mentioned above in the core biosynthesis section, AlgC also functions as a phosphoglucomutase (PGM) as shown by loss of this activity in the lysates of algC P. aeruginosa mutants, and the ability of algC to complement a defined E. coli PGM mutant.44 In contrast to the degeneracy in wbpW/algA/ pslB where each polysaccharide biosynthesis cluster possesses its own PMI/GMP, AlgC catalyses the mutase steps required for synthesis of multiple cell surface glycans. Common Polysaccharide Antigen,44 alginate,184 and presumably Psl synthesis require the AlgC PMM function, and LPS core,44,90 and rhamnolipids185 require its PGM function.

restored Common Polysaccharide Antigen expression to the Common O Polysaccharide strain rd7513.186 The function of the purified protein has also been characterised in vitro by capillary electrophoresis and NMR analysis of enzyme-catalysed reaction products187 and the X-ray crystal structure has been solved.188 The protein product of rmd catalyses the final step in the GDP-D-Rha biosynthesis pathway, the dinucleotide cofactor-dependent reduction of the GMD reaction product, GDP-4-keto-D-Rha (31) producing GDP-DRha (32). Disruption of the rmd gene on the P. aeruginosa chromosome abrogates Common O Polysaccharide synthesis.178 RMD has also been studied in purified form, using in vitro assays.187 In vitro assays showed that, like some homologues from other species, P. aeruginosa GMD is bifunctional, and capable of catalysing the same reaction as RMD.187 While this second GMD function may be important in regulation of enzyme function or allowing GMD to change the oxidation state of its bound co-factor, it is unlikely to be significant for in vivo GDP-D-Man synthesis. The fact that rmd mutations prevent Common Polysaccharide Antigen production indicates that GMD alone cannot complement a defect in the specific reductase (RMD) function. Co-incubation of GMD, RMD and GDP-D-Man inhibits the GMD reaction,187 indicating that this pathway may be subject to feedback inhibition. This could be a mechanism by which the cell regulates consumption of GDP-D-Man, which as noted above, is a precursor for several cell surface polysaccharides.

gmd and rmd

gmd encodes a GDP-D-mannose 4,6-dehydratase, which catalyses the conversion of GDP-D-mannose (30) to GDP-6-deoxy-D-lyxo-hexos-4-ulose (GDP-4keto-D-Rha; 31). Expression of this gene in trans

Glycosyltransferases, wbpX, wbpY and wbpZ

wbpX, wbpY, and wbpZ encode putative glycosyltransferases required for biosynthesis of the Common Polysaccharide Antigen. All encode proteins with the


Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa Pfam00534 glycosyltransferase conserved motif and generation of non-polar mutants showed that each is necessary for Common O-Polysaccharide production. The IATS O15 P. aeruginosa type strain appears to lack Common Polysaccharide Antigen as a result of a mutation in wbpX since introduction of wbpX in trans restores Common Polysaccharide Antigen synthesis in this strain.170 The closest, characterised homologues of these P. aeruginosa genes are annotated as a-mannosyltransferases from the E. coli O9a O Antigen biosynthesis (wbd) locus.189 The E. coli O9a antigen is a homopolymer of a-(1 ! 2)- and a-(1 ! 3)-linked mannose which differs from rhamnose only in the presence of the C-6 hydroxyl. WbpX, WbpY and WbpZ are homologous with WbdA, WbdB and WbdC, respectively. The E. coli genes were cloned from what was originally reported to be an O9 strain, and when they were introduced to E. coli K-12 cells on plasmids, membranes prepared from those cells were able to incorporate mannose from GDP-[14C]mannose in vitro. On the basis of these data, a model was proposed for the function of WbdA, WbdB and WbdC in biosynthesis of the E. coli O9 mannan.189 Based upon this model, precise functions were also assigned for WbpX, WbpY and WbpZ from P. aeruginosa.170 For example, WbpZ was proposed as the transferase which adds the first rhamnose to the lipid carrier, because it is 52% identical with WbdC, the enzyme required for initiation of the O9a mannan.189 This is a remarkably high level of sequence conservation for glycosyltransferase enzymes and it is interesting to speculate that since small amounts of mannose have twice been identified in Common O-Polysaccharide preparations,96,171 WbpZ may add mannose to initiate Common O-Polysaccharide synthesis. Since the original report, however, Kido and colleagues190 have reported further characterisation of the E. coli system and it is rather difficult to reconcile some of the newer data with the original model, particularly with respect to the roles of the two mannosyltransferases proposed to elongate the chain: WbdB being suggested to create a-(1 ! 3) glycosyl linkages and WbdA responsible for the a(1 ! 2) linkages. First, E. coli strain F719, from which these mannosyltransferases were cloned has been reclassified as serotype O9a with a shorter repeating unit than originally thought.190 Second, WbdA has two domains and truncation of WbdA resulted in a polymer with exclusively a-(1 ! 2) linkages.191 In other words, the C-terminal domain of WbdA was found to be necessary for formation of a-(1 ! 3) linkages. Further work is required to establish the precise functions of WbdA and WbdB. It is also important to note that while wbpX and wbdA are homologous to each other, wbpX encodes a much shorter protein of only 460 amino acids with only one predicted glycosyltransferase domain compared with the 815 amino acids and two domains

285

of WbdA. BLAST analysis aligns WbpX with both the WbdA C-terminal and N-terminal domains (Table 5). At this stage, it is unclear whether the P. aeruginosa proteins function in a manner precisely analogous to their homologues in the E. coli model system. Export and processing of Common O polysaccharide ABC transporter: wzm and wzt

The Common Polysaccharide Antigen biosynthesis gene cluster contains two genes, wzm and wzt, which encode the integral membrane protein and ATP-binding protein components of the ABC O-polysaccharide exporter, respectively. Allele replacement mutagenesis of either coding sequence prevents export of Common O Polysaccharide to the cell surface. Western blot analysis of whole cell lysates and immunoelectron microscopy indicated that these mutants still produced the Common O Polysaccharide, but that it had a higher electrophoretic mobility in SDS-PAGE and was located in the cytoplasm.164 These data are consistent with the hypothesis that the ABC transporter mutants were defective in translocation of the O polymer across the inner membrane, so that the polymer cannot be ligated to lipid-A core (which can only occur in the periplasm), but remains attached to an isoprenyl-pyrophosphate lipid carrier. Cuthbertson and colleagues192,193 identified a subclass of ABC transporters with an extended Wzt C-terminal domain which specifically recognises modifications of the non-reducing terminus of their polysaccharide chain transport substrates. In E. coli O8, the terminal mannose is methylated, whereas in E. coli O9a, the terminal sugar is both phosphorylated and methylated. While the Wzm ABC transporter components are interchangeable between the two systems, the Cterminal domains of the Wzt proteins are not, as these recognise the serotype-specific modifications of the nonreducing terminal residue.192 The P. aeruginosa Common Polysaccharide Antigen locus Wzt has an extended C-terminus like that of E. coli O8 and O9a Wzt proteins. Furthermore, the E. coli O8 Wzt is the closest characterised homologue, sharing 61% sequence identity overall, with 69% identity over the Nterminal 254 amino acids and, perhaps most meaningfully, 50% identity over the extended C-terminal domain. The E. coli O9a Wzt is also a homologue of the P. aeruginosa transporter protein with 64% identity in the N-terminus, but only 21% in the C-terminal domain. This is comparable to the low conservation between the E. coli O8 and O9a Wzt C-termini.193 On the basis of this sequence analysis, we tentatively predict that the Common O-polysaccharide translocation across the inner membrane is dependent on recognition of specific


286


modification of the non-reducing terminal residue, and that this modification is methylation, and not phosphorylation. Determination of chain length and other modifications: PA5455–PA5459

In the E. coli O8 and O9a systems, the non-reducing terminus of the O polysaccharide is modified by a protein called WbdD. As well as introducing the specific terminal modifications for recognition of the polysaccharide by the ABC transporter as described above, the WbdD-catalysed reaction terminates chain extension thereby regulating the O-Antigen chain length. In E. coli O9a, WbdD has both kinase and methyltransferase domains, and chain extension is terminated by transfer of a phosphate group (followed by transfer of a methyl group) prior to translocation of the complete polymer across the inner membrane. The E. coli O8 WbdD lacks the kinase domain and terminates chain extension by transfer of a methyl group to the nonreducing terminus.194 In the P. aeruginosa eight-gene Common Polysaccharide Antigen cluster previously described, no wbdD is present, but BLAST interrogation of the PAO1 genome identifies two homologues of WbdDE. coliO8: PA5457 and PA5459 encoded in the putative operon immediately upstream of rmd (rmd is PA5454; Table 5). The same analysis using the E. coli O9a WbdD sequence as probe fails to identify any significant homologues in PAO1. PA5457 and PA5459 share 46% identity with each other and both contain an S-adenosylmethionine (SAM)-dependent methyltransferase domain (Pfam08242). Thus, sequence analysis suggests a role for one or both of these encoded proteins in regulation of O-chain length. The mechanisms by which these systems regulate methylation may differ, however, since the P. aeruginosa cluster contains two putative methyltransferases and the PA5457 and PA5459 proteins are both shorter than WbdD from E. coli O8. WbdD has an extended C-terminus containing a predicted coiled-coil domain.194 The methytransferase domain of PA5459 is also homologous to the N-terminus of WsaE from Geobacillus stearothermophilus. G. stearothermophilus NRS 2004/3a produces an L-Rha homopolymer195 by a pathway believed to be related to the biosynthesis of ABC transporter-dependent O polysaccharides.196 WsaE methylates the non-reducing terminus of L-Rha oligomers, preventing further elongation of the chain. In this system, the chain-terminating methyltransferase domain is fused to two glycosyltransferase domains which function in elongation of the L-rhamnan.196 Of the other genes in this locus, PA5455 and PA5456 both encode putative glycosyltransferases having partial

glycosyltransferase motifs (Pfam00534). They are also homologous to each other. The detection of small quantities of non-D-Rha sugars in Common OPolysaccharide preparations (as described above) suggests the possibility that there may be an oligosaccharide linker or adapter present at the reducing end of the D-Rha homopolymer. The homopolymeric O polysaccharide of Bordetella bronchiseptica appears to be assembled on a five sugar oligosaccharide which links the O homopolymer to the lipid carrier.197,198 PA5455 and PA5456 could function in the assembly of such a linker. Finally, PA5458 has an acetyltransferase-3-family conserved domain (Pfam01757) and shares 24% amino acid sequence identity with OatA, the peptidoglycan O-acetylase which O-acetylates C-6 of N-acetyl muramic acid (MurNAc).199 The acetyltransferase domain and homology with a polysaccharide acetylase support the hypothesis that the PA5458 enzyme is responsible for the small amounts of O-acetylation detected in one compositional analysis of the Common O polysaccharide.96 None of the genes in the PA5455–PA5459 locus have been previously investigated experimentally, though this work is now under way in our laboratory.

Conservation of the Common Polysaccharide Antigen biosynthetic locus in P. aeruginosa strains

Southern hybridisation identified each of the eight genes from rmd to wbpZ in all 20 IATS serotype strains’ genomic DNA.164,170,178,200 The majority of these strains showed the same hybridisation patterns, with a few exceptions. In particular, the IATS-type strains for O12, O13, O15 and O16 showed altered Southern hybridisation patterns with probes containing more than one of these genes. These are four of the IATS strains which do not produce Common Polysaccharide Antigen. IATS strains O7 and O14 also do not produce Common Polysaccharide Antigen,33 but showed normal hybridisation patterns. In each of the Common-Polysaccharide Antigen IATS reference strains except for the O12 type strain, the recombinant cosmid pFV3 (which contains the eight-gene Common O-Polysaccharide biosynthesis operon) was able to restore Common Polysaccharide Antigen LPS production. This indicates that even in the O7 and O14 reference strains which have normal Southern hybridisation profiles, the defect in Common Polysaccharide Antigen is due to a mutation in this region of the chromosome, potentially frameshift mutations or point mutations affecting catalytic residues. As mentioned above, in the IATS O15 reference strain the defect was localised to a mutation in wbpX by complementation experiments.170 Not all O12 strains lack Common Polysaccharide Antigen.200


Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa Of the P. aeruginosa strains whose genome sequences are currently available at Genbank,201 PAO1, PA7, PA14, 2192, C3719, PACS2 (and we also checked LESB58) all have the Common Polysaccharide Antigen biosynthesis locus including the five genes which we identified upstream of rmd. The organisation of these genes is also conserved in all of these strains except for PA7, in which the five-gene subcluster has been apparently transposed and flipped in a re-arrangement event and now lies on the same strand as the other eight genes and 9.8-kb upstream of the rmd homologue.

287

made by Bystrova et al.94 that the IATS O15 and O17 O-Antigen structures (La`nyi–Bergan serogroups O15 and O14, respectively) are exceptional because they do not have the usual initiating sugar (a FucNAc, or a QuiNAc derivative) in the O repeat and their lipopolysaccharides lack the semi-rough (core þ 1 O repeat) forms. It is believed, therefore, that biosynthesis of the O15 and O17 O polysaccharides is encoded elsewhere in the genome. Biosynthesis of O-Antigen repeat units

O-SPECIFIC ANTIGEN O Antigen gene clusters Early work on O-Antigen biosynthesis in P. aeruginosa demonstrated that, in each of IATS serotype O5, O6 and O11 strains, a large cluster of O-Antigen biosynthesis genes was located in the same region of the genome (Fig. S3).169,202,203 Each serotype has a different set of genes located between the highly conserved flanking sequences of the himD/ihfB gene upstream of the locus and wbpM at the downstream end of the locus. High stringency Southern blotting analysis with a wbpM probe indicated that this gene is common to all P. aeruginosa serotypes.202 Taking advantage of the conserved flanking sequences on either side of the O-Antigen locus, Maynard Olson’s group used a yeast-recombinational cloning method and cloned the himD/ihfB–wbpM region from all 20 IATS serotypes and sequenced the intervening DNA.161 They identified 11 divergent O-Antigen clusters. Four of these clusters were found in more than one IATS strain, with a very high degree of sequence conservation. If strains are clustered according to shared sets of genes in this locus, there is excellent agreement with the grouping of strains suggested by serological157–160 and structural94 properties of O-Antigen. In other words, almost identical sets of genes are present within the O2/O5/O16/O18/O20, O7/O8, O10/O19 and O13/O14 serogroups. The two exceptions to this rule are the IATS O15 and O17 strains. The IATS O17 himD/ihfB–wbpM cluster is closely related to the O11 O-Antigen locus but contains two insertion-sequence elements and a major deletion, and the locus does not function to produce the O11 O-Specific polysaccharide.204 Of the two IATS O15 strains examined by Raymond and co-workers,161 one had a cluster resembling the O3 locus, but containing three insertion-sequence elements, and the other was completely devoid of an O-Antigen cluster in this region. Furthermore, wbpM can be mutated in O15 or O17 backgrounds without affecting O-Antigen production.205 This observation was later explained by the observations

wbpM – a common gene in all serotypes

wbpM is conserved in all 20 serotypes of P. aeruginosa and is found at the distal end of the O-Antigen biosynthetic locus. Its presence in all serotypes suggested that it may be involved in synthesis of the 6-deoxy sugars D-QuiNAc and D-FucNAc, and their derivatives, found in the O repeat in almost all serotypes. Consistent with this hypothesis, knockout mutation of wbpM abrogated O-Antigen synthesis in O5,202 O6,169 O7,206 O10, O3,205 and O11,105 but is not required for synthesis of the Common Polysaccharide Antigen, or the O15 and O17 O-Antigens. O15 and O17 O repeats do not contain 205 D-QuiNAc or D-FucNAc. Furthermore, subsequent biochemical analysis indicated that WbpM is a UDP207 D-GlcNAc 4,6-dehydratase (Fig. 9). Reduction of the product of the WbpM-catalysed reaction (33) produces UDP-D-QuiNAc (35) or UDP-D-FucNAc (34) depending on which face of the keto sugar accepts the hydride. WbpM is a short-chain dehydrogenase/reductase (SDR) family protein, and it has an unusual catalytic triad composed of Thr428, Met438 and Lys442, in which a methionine takes the place and apparently performs the function of the usual SDR catalytic triad tyrosine.208 Note that Ser427 was originally reported as the conserved Ser/Thr member of the triad, but Reverse Position-Specific (RPS)-BLAST alignment of the WbpM sequence with Pfam01370 identified Thr428 as the conserved residue. WbpM is a member of a subfamily of SDR enzymes which have an extended N-terminus that includes several transmembrane helices. Topological analysis by generation of PhoA and LacZ fusions showed that WbpM has four transmembrane helices, and the catalytic C-terminal domain is located cytoplasmically. The N-terminal membrane-spanning domains are not required for enzyme function, but appear to play a role in LPS biosynthesis nevertheless: a construct called S262 is a 786-bp 50 -truncation of wbpM and it encodes a soluble enzyme that can be expressed and restore O-Antigen expression in a wbpM mutant. S262 is, therefore, catalytically active in vivo. However, the LPS produced by this complemented mutant lacks the usual cluster of high molecular mass O chains


288


suggesting that the N-terminal region of WbpM is necessary for proper regulation of O-Antigen chain length, possibly through interactions with other components of the hypothetical O-Antigen biosynthetic complex.209 WbpM has homologues in many medically important bacterial strains. Of the homologues of comparable length (approximately 600 amino acids) and with similar predicted membrane-spanning domains, the genes encoding WbgZ from Plesiomonas shigelloides (47% identical), WlbL from the LPS trisaccharide biosynthesis cluster of Bordetella pertussis (44% identical), and Cap8D from Staphylococcus aureus (34% identical) can all complement a wbpM mutation in P. aeruginosa.205 PglF (Cj1120c) from Campylobacter jejuni also has the same catalytic function.210 There are also a number of homologues of wbpM which encode shorter,

soluble enzymes, notably flaA1 from Helicobacter pylori (HP0840, recently renamed pseB), pseB from C. jejuni (Cj1293),211 and wbjB from the P. aeruginosa O11 O-Antigen cluster.212 For many years, WbpM was misannotated as a bifunctional UDP-D-GlcNAc 4,6-dehydratase/C4-reductase. Part of the confusion was due to early cross-complementation experiments that showed that in multicopy, flaA1 (which was also misannotated) could restore O-Antigen production in a wbpM mutant205 and the implication was drawn that the two enzymes were functionally equivalent.208 The confusion now appears to have been resolved, however, thanks in part to developments in NMR methodology and instrumentation.212 WbpM and all currently characterised members of the longer, membrane-associated subfamily of WbpM homologues are simple (retaining) UDP-GlcNAc 4,6-dehydratases and the purified product

OH

WbpM

O

HO HO

OH

O

WbpK

O

O HO

HO AcNH

AcNH

UDP

UDP-D-GlcNAc 1

+ NAD(P)H

AcNH

UDP

UDP

UDP-D-FucNAc 34

33

WbjB

OH

WbjB

WbpV + NAD(P)H OH

O O

AcNH

O

HO HO

HO

AcNH

UDP

36

UDP

UDP-D-QuiNAc 35

WbjC

UDP

O O

HO NHAc

37

UDP

O

WbjC + NADPH

O

WbjD

HO OH

NHAc

38

UDP NHAc

HO OH

UDP-L-FucNAc 39 LfnA + ATP? + NH3? O HO OH

H N

UDP

HN

UDP-L-FucNAm 49 Fig. 9. Sugar-nucleotide biosynthesis pathways initiated by UDP-GlcNAc 4,6-dehydratase enzymes. UDP-D-GlcNAc 4,6-dehydratase enzymes such as WbpM catalyse the synthesis of only one product, UDP-4-keto-D-QuiNAc (33). UDP-GlcNAc 4,6-dehydratase-5-epimerases such as WbjB catalyse the production of UDP-6-deoxy-4-keto-L-IdoNAc (36), which is slowly racemised to produce (33). Conserved domains present in LfnA suggest that it may catalyse amidotransfer to the 2-acetamido group of L-FucNAc to form the L-FucNAm residue in P. aeruginosa O11,226 but there is no direct evidence that the amidotransfer occurs onto the sugar-nucleotide as shown here. 38, UDP-L-PneNAc. Downloaded from http://ini.sagepub.com at UNIV OF GUELPH on October 6, 2009

Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa of a reaction catalysed by S262, the soluble truncated form of WbpM, was identified by NMR analysis to be UDP-2-acetamido-2,6-dideoxy-D-xylo-4-hexulose (UDP-4-keto-D-QuiNAc,33; Fig. 9).207 An analogous experiment with PglF detected the same product. On the other hand, current data suggest that the subfamily of short WbpM homologues, i.e. FlaA1, PseB and WbjB, catalyse a 4,6-dehydration of UDP-D-GlcNAc, by a mechanism involving inversion of the stereochemistry at C5. The X-ray crystal structure of FlaA1213 and mechanistic studies on PseB from C. jejuni214 have elucidated details of the inverting dehydratase reaction catalysed by these enzymes. The initial product of the inverting dehydratase reaction is UDP-2-acetamido-2,6-dideoxyL-arabino-4-hexulose (UDP-6-deoxy-4-keto-LIdoNAc), which is a very labile compound and so was not detected in early experiments. Subsequently, a second product appears in these enzyme–substrate incubations, which is identical to the WbpM-catalysed reaction product, UDP-4-keto-D-QuiNAc (33). This second product appears as a result of racemisation of the C5 stereocentre.210,212 Because of the slow kinetics for the C5 racemisation, it has been suggested that it is probably not physiologically relevant.213 Slow racemisation of the FlaA1 product may account, however, for the ability of flaA1 to complement a wbpM mutant when expressed from a multi-copy plasmid (discussed by Schoenhofen et al.210).

Biosynthesis of the P. aeruginosa O5 O Antigen

Pseudomonas aeruginosa PAO1 is the type-strain for the species,215 hence its genome was the first in this species to be sequenced and annotated.216 Pseudomonas O Antigen is one of the best characterised (Table 6). The O Antigen of P. aeruginosa serotype O5 contains 2-acetamido-3-acetamidino-2,3-dideoxyD-mannuronic acid (D-ManNAc3NAmA), 2,3-diacetamido-2,3-dideoxy-D-mannuronic acid (DManNAc3NAcA), and 2-acetamido-2,6-dideoxyD-galactose (N-acetyl-D-fucosamine; D-FucNAc); (Table 4).217 Based on similar serology and O Antigen structures, serotype O5 is in the O2/O5/O16/O18/O20 serogroup. Members of this serogroup have nearly identical O Antigen biosynthesis clusters containing 20 ORFs (Fig. S3).161 Biosynthesis of 2-acetamido-2-deoxy-D-fucose (D-FucNAc): wbpM, and wbpK

The biosynthesis of UDP-D-FucNAc (34) is thought to use UDP-D-GlcNAc (1) as the initial substrate and require wbpM and wbpK (Fig. 9). Currently, WbpK is postulated to be a dinucleotide co-factor-dependent 4-reductase enzyme which catalyses reduction of UDP-4-keto-D-QuiNAc (33, product of the WbpM-catalysed reaction) to generate UDP-D-FucNAc (Fig. 9). A non-polar knockout mutation of wbpK abrogated

Table 6. Genes in the O-Antigen biosynthesis cluster of P. aeruginosa IATS O5 Gene


Key reference*

wzz/PA3160 wbpA/PA3159 wbpB/PA3158 wbpC/PA3157 wbpD/PA3156 wbpE/PA3155 wzy/PA3154 wzx/PA3153 hisH2/PA3152 hisF2/PA3151 wbpG/PA3150 wbpH/PA3149 wbpI/PA3148 wbpJ/PA3147 wbpK/PA3146

O-Antigen chain length regulator UDP-D-GlcNAc 6-dehydrogenase UDP-D-GlcNAcA 3-oxidase Possible O-acetyltransferase UDP-D-GlcNAc3NA 3-acetyltransferase UDP-3-keto-D-GlcNAcA 3-transaminase O Antigen a-polymerase O unit flippase Amidotransferase glutaminase domain?

268 221 223

wbpL/PA3145 PA3144 PA3143 PA3142 wbpM/PA3141

289

Amidotransferase Glycosyltransferase (GT-4) UDP-D-GlcNAc3NAcA 2-epimerase Glycosyltransferase (GT-4) UDP-4-keto-D-QuiNAc 4-reductase (UDP-D-FucNAc forming) Initiating glycosyl-1-phosphate transferase Insertion sequence Insertion sequence Insertion sequence UDP-D-GlcNAc 4,6-dehydratase

Glycosyltransferase (GT) family classification is taken from the CAZy database.123 *References are cited if they describe experimental investigation of the P. aeruginosa gene. Downloaded from http://ini.sagepub.com at UNIV OF GUELPH on October 6, 2009

223 223 259 261

13 225 169

207

290


O-Antigen synthesis169 confirming that it has a role in O-Antigen biosynthesis, but there is no experimental evidence to support, specifically, annotation of the enzyme as a 4-reductase. That being said, WbpK is a member of the SDR family, with 36% amino acid identity with the UDP-4-keto-D-QuiNAc 4-reductase enzyme from the O6 locus, WbpV (see below) and these bioinformatic data are consistent with its proposed role. D-FucNAc residues, or closely related sugars, are also present in the O units of serotypes O1, O9, and O11 and the serogroups O2/O5/O16/O18/O20 and O7/O8. Except for serotypes O1 and O9, the O-Antigen biosynthesis clusters in all of these serotypes possess wellconserved homologues of WbpK with greater than 50% amino acid identity (Table 7). This strongly suggests that for most D-FucNAc-producing P. aeruginosa serotypes, the WbpM–WbpK pathway is conserved. The O units in serotypes O1 and O9 are the only ones to contain both D-FucNAc and its 4-epimer, D-QuiNAc. Together with the lack of a well-conserved WbpK homologue, this leads us to speculate that D-FucNAc synthesis may follow a different pathway in these serotypes, possibly involving stereo-non-specific reduction of UDP-4-keto-D-QuiNAc, or 4-epimerisation of UDP-D-QuiNAc (35).

Biosynthesis of 2,3-diacetamido-2,3-dideoxy-D-mannuronic acid (D-ManNAc3NAcA): wbpA, wbpB, wbpE, wbpD, and wbpI

Biosynthesis of UDP-D-ManNAc3NAcA begins with the common precursor, UDP-D-GlcNAc, and requires five genes – wbpA, wbpB, wbpE, wbpD, and wbpI (Fig. 10). A knockout mutant lacking any one of these genes is unable to produce O Antigen.218–220 The first enzyme of the pathway, WbpA, is a 6-dehydrogenase that catalyzes the conversion of UDP-D-GlcNAc to UDP-2-acetamido-2-deoxy-D-glucuronic acid (UDP-DGlcNAcA, 40) using NADþ as a co-enzyme.221 The next steps are thought to be an oxidation reaction catalyzed by WbpB, forming UDP-2-acetamido-2deoxy-D-ribo-hex-3-uluronic acid (UDP-3-keto-DGlcNAcA, 41), followed by a WbpE-catalysed transamination reaction creating UDP-2-acetamido-3-amino2,3-dideoxy-D-glucuronic acid (UDP-D-GlcNAc3NA, 42). A similar two-step oxidation and transamination of UDP-D-GlcNAc to UDP-D-GlcNAc3N has been demonstrated in Acidithiobacillus ferrooxidans lipid A biosynthesis.222 WbpB contains putative oxidoreductase motifs and a Rossman-fold region for binding NAD(P). A wbpE knockout is complemented by the B. pertussis wlbC gene, indicating that they have the same

Table 7. Homologues of WbpKO5 in other D-FucNAc-producing serotypes of P. aeruginosa O-Antigen cluster

O1y

O1y

O2/O16/O18/O20

O7/O8

O9y

O9y

O11

ORF Identity (%)

orf14 21

orf16 38

orf18 100

orf18 50

orf14 21

orf16 38

orf12 51

y

Serotypes that produce both D-FucNAc and D-QuiNAc.

OH HO HO

OH

WbpA

O

AcNH

O

UDP

WbpB

O

HO HO

+ 2NAD+

OH O

O

40

UDP-D-GlcNAc 1

WbpE

O

HO

+ NAD+

AcNH UDP

OH

AcNH

UDP

+L-Glu

O

O

HO H2N

AcNH

41

UDP

42 WbpD + Ac-CoA

O HO HN

OH NHAc O

H N

WbpG

O

+ ATP?, NH3? UDP

UDP-D -ManNAc3NAmA 45

HO AcHN

OH NHAc O

UDP

UDP-D-ManNAc3NAcA 44

OH

WbpI

O

O

HO AcHN

AcNH

UDP

43

Fig. 10. UDP-D-ManNAc3NAcA and proposed UDP-D-ManNAc3NAmA biosynthesis pathways. No direct evidence exists for the precise role proposed here for WbpG. 40, UDP-GlcNAcA; 41, UDP-3-keto-GlcNAcA; 42, UDP-GlcNAc3NA; 43, UDP-GlcNAc3NAcA. Downloaded from http://ini.sagepub.com at UNIV OF GUELPH on October 6, 2009

Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa function,220 and WlbC is a typical type-2 aminotransferase based on structural and spectroscopic data.223 The reactions catalysed by WbpB and WbpE were characterised by incubation of both purified proteins with UDP-GlcNAcA (40). The anticipated product of the dual-enzyme reaction, UDP-GlcNAc3NA (42), was purified and its structure confirmed by NMR.223 WbpB produces UDP-3-keto-GlcNAcA and uses a-ketoglutarate to regenerate the NADþ co-factor bound in the enzyme’s active site. a-ketoglutarate is a product of the downstream, WbpE-catalysed aminotransfer reaction, which generates UDP-GlcNAc3NA.224 UDP-D-GlcNAc3NA (42) was long suspected to be the substrate for WbpD, which was proposed to be an Nacetyltransferase.219 Early studies showed that WbpD belongs to the superfamily of hexapeptide acyltransferases, forms trimers, and utilizes acetyl-CoA as a cofactor.219 After advances in our understanding of WbpB and WbpE, a four-step synthesis of UDP-2,3-diacetamido-2,3-dideoxy-D-glucuronic acid (UDP-DGlcNAc3NAcA, 43) was accomplished using WbpA, WbpB, WbpE, and WbpD in a single reaction and the final product was verified by NMR.223 This demonstrated that WbpD is a 3-N-acetyltransferase that converts UDPD-GlcNAc3NA (42) into UDP-D-GlcNAc3NAcA (43) using acetyl-CoA as the co-factor.223 The final step in this pathway requires wbpI, which encodes a 2-epimerase that interconverts UDP-DGlcNAc3NAcA (43) and UDP-D-ManNAc3NAcA (44).225 This reaction was studied using a live-reaction NMR approach that allowed the conversion to be followed in real-time.225 Biosynthesis of 2-acetamido-3-acetamidino-2,3-dideoxy-Dmannuronic acid (D-ManNAc3NAmA): wbpG

UDP-D-ManNAc3NAcA (44) is thought to undergo an amidotransfer reaction catalysed by WbpG to produce UDP-D-ManNAc3NAmA (45) (Fig. 10). The amino acid sequence of WbpG has conserved domains common to ATP-binding enzymes and amidotransferases; this prediction is supported by genetic evidence since a WbpG knockout lacks O-Antigen polymers, but does appear to produce the semi-rough LPS (lipid A core þ 1 O unit).13,202 The LPS phenotype of the wbpG knockout, therefore, resembles that of an O-Antigen polymerase (wzy) knockout strain (Wzy is described below). The wbpG phenotype is consistent with its proposed function if it is supposed that three undecaprenyl-pyrophosphatelinked O-Antigen sugars are processed and transferred to lipid-A-core as normal, with a second ManNAc3NAcA in the normal place of ManNAc3NAmA; but that the presence of the acetamidino (NAm)-containing nonreducing terminal sugar is required by the O-Antigen polymerase in order to polymerise the O units.13

291

Homologues of WbpG are found in other P. aeruginosa serotypes and other bacterial species that synthesize sugars with acetamidino groups. P. aeruginosa O12 and E. coli O145 synthesize 2,6-dideoxy-2-acetamidinoL-galactose (L-FucNAm) as part of the O Antigen and the wbpG homologue lfnA is required to synthesize the acetamidino moiety in this sugar in serotype O12 (Fig. 9).226 Campylobacter jejuni produces 5-acetamido-7-acetamidino-3,5,7,9-tetradeoxy-L-glycero-a-Lmanno-nonulosonic acid (Pse5NAc7Am) for the flagellin glycan, and synthesis of the acetamidino moiety in this sugar requires the wbpG homologue, pseA.227 Thus, there is an inductive suggestion that wbpG is an amidotransferase responsible for generating DManNAc3NAmA. Glycosyltransferases: wbpH and wbpJ

The O-Antigen biosynthesis cluster of P. aeruginosa PAO1 contains two putative glycosyltransferases, wbpH and wbpJ. Both are assigned by the CAZy database into the GT-4 family. Due to the similarity of their predicted substrates (UDP-D-ManNAc3NAcA and UDP-DManNAc3NAmA), sequence analysis is not sufficient to predict which glycosyltransferase adds each residue. Biosynthesis of the P. aeruginosa O6 O Antigen The P. aeruginosa O6 O Antigen contains D-QuiNAc, two sugars related to 2-amino-2-deoxy-D-galacturonic acid (GalNA), and L-Rha (Table 4). The O6 O-Antigen biosynthesis cluster contains 12 ORFs (Fig. S3), some of which have been demonstrated to encode biosynthetic enzymes required to form these sugars (Table 8). Biosynthesis of 2-acetamido-2-deoxy-D-quinovose (D-QuiNAc): wbpM and wbpV

The UDP-4-keto-D-QuiNAc (33) product of the WbpMcatalysed 4,6-dehydration of UDP-D-GlcNAc (1), is thought to be reduced to UDP-D-QuiNAc (35) in a reaction catalysed by WbpV (Fig. 9). WbpV is predicted to have a NAD(P)-binding Rossmann fold, and is in the SDR family of oxidoreductase enzymes. This reaction will therefore use a NADH or NADPH co-factor as the hydride donor in reduction of the keto-sugar. wbpV is essential for O-Antigen synthesis, as a knockout mutant in P. aeruginosa lost O-Antigen expression.169 In Rhizobium etli, a homologue of wbpV, lpsQ, (the two proteins share 46% identity) is required for synthesis of D-QuiNAc in the R. etli LPS core. lpsQ mutants synthesise LPS with 4-keto-D-QuiNAc in the usual place of D-QuiNAc,228 and expression of wbpV in trans complemented the R. etli lpsQ mutation.207 This suggests that LpsQ and WbpV are functionally homologous and catalyse the 4-reduction of UDP-4-keto-D-QuiNAc (33)


292


Table 8. Genes present in the O-Antigen biosynthesis cluster of P. aeruginosa IATS serotype O6 strain Gene



Key reference*

wzz wbpO

53% P. aeruginosa O5 Wzz 71% E. coli WbqA O121 71% S. dysenteriae 7 Gna 68% E. coli O121 WbqB 68% S. dysenteriae 7 Gne No significant homologues (by BLAST analysis)

Chain length regulator UDP-GlcNAc 6-dehydrogenase

230

UDP-GlcNAcA 4-epimerase

230

wbpP (wbpQ) wzx wbpR wbpS

wbpV

32% P. aeruginosa O13/O14 Orf10 54% E. coli O121 WbqG 54% S. dysenteriae 7 WbpS 30% B. subtilis AsnB (over 631 amino acids) 46% E. coli O121 WbqH 46% S. dysenteriae 7 WbnK 52% E. coli O121 WbqI 52% S. dysenteriae 7 WbnL 88/87% P. aeruginosa O13/O14 Orf11 46% R. etli LpsQ

wbpL wbpM

61% P. aeruginosa O5 WbpL 98% P. aeruginosa O5 WbpM

wbpT wbpU

Und-PP-O unit flippase Glycosyltransferase (GT-4) Glutamine-dependent amidotransferase –synthesis of UDP-GalNAcAN Glycosyltransferase (GT-4) probably of a GalNA sugar derivative Glycosyltransferase (GT-4), probably of a GalNA sugar derivative UDP-4-keto-D-QuiNAc 4-reductase (UDP-QuiNAc synthesising) Initiating glycosyl-1-phosphate transferase UDP-D-GlcNAc 4,6-dehydratase

207 169 169


to produce UDP-D-QuiNAc (35). WbpV shares 36% sequence identity with its homologue encoded in the O5 O-Antigen locus, WbpK. While WbpK is also a putative UDP-4-keto-D-QuiNAc 4-reductase, wbpV does not complement a P. aeruginosa O5 wbpK mutation, and wbpK does not complement a P. aeruginosa O6 wbpV knockout,169 indicating that these two 4-reductases are stereospecific, catalysing the donation of hydride to opposite faces of the substrate hexose ring. D-QuiNAc residues are also found in the O-Antigen repeat units of serotypes O1, O4, O6, O9, O10, O12, O13, O14 and O19. In all of these serotypes the WbpM– WbpV pathway for UDP-D-QuiNAc (35) biosynthesis appears to be conserved, since close WbpV homologues, with greater than 55% amino acid identity, are present in their O-Antigen clusters (Table 9). Biosynthesis of 2-amino-2-deoxy-D-galacturonic acid (D-GalNA) derivative sugars

The IATS O6 type strain, and strains with serologically related O Antigens are distinguished by subgroup classifications in the La`nyi–Bergan serotyping scheme (Table 10). The O-specific repeats in all of these O-Antigen subtypes contain galactosaminuronic acid (GalNA)-derived sugars. One is 2-N-formylated, and in

different strains this sugar is present to varying degrees in the uronamide form. The second is 2-N-acetylated and is always present as the uronamide. One or other of these sugars can be 3-O-acetylated depending on the strain. The biosynthetic steps leading to the 3-Oacetylation and 2-N-formylation are unknown, and probably catalysed by enzymes encoded outside of the O-Antigen gene cluster. Biosynthesis of 2-acetamido-2-deoxy-D-galacturonic acid (D-GalNAcA): wbpO and wbpP

The first steps in the biosynthesis of the GalNAderivative sugars found in the O6 O repeat are thought to be the WbpO and WbpP-catalysed reactions which convert UDP-D-GlcNAc (1) to UDP-GalNAcA (47) (Fig. 11). Knockout mutations made in either of wbpO or wbpP abrogate O-Antigen biosynthesis.169 WbpO is a member of the sugar-nucleotide dehydrogenase family, which also includes UDP-Glc and GDPMan dehydrogenases. Members of this enzyme family catalyse the NADþ-dependent 2-fold oxidation of an alcohol to an acid without the release of an aldehyde intermediate. Consistent with this classification, capillary electrophoresis (CE), CE-MS, tandem MS and NMR analysis of reaction products confirm that WbpO



293

Table 9. Homologues of WbpVO6 in other D-QuiNAc-producing serotypes of P. aeruginosa O-Antigen cluster

O1y

O1y

O4

O9y

O9y

O10/O19

O12

O13

O14

ORF Identity (%)

orf14 39%

orf16 60%

orf13 55%

orf14 39%

orf16 60%

orf13 58%

orf24 55%

orf12 82%

orf12 83%

y

Serotypes that produce both D-FucNAc and D-QuiNAc.

Table 10. La`nyi–Bergan O6 serogroup O-Antigen repeat structures Other typing schemes

La`nyiBergan serogroup

IATS O6

6a 6a,6b 6a,6c 6a,6d

Fisher 1 Homma G

O-Antigen repeat

!3)-L-Rha-(1 ! 4)-D-GalNAc3OAcAN-(1 ! 4)-D-GalNFoA-(1 ! 3)-D-QuiNAc-(1! !2)-L-Rha-(1 ! 4)-D-GalNAcAN-(1 ! 4)-D-GalNFo3OAcA(N)-(1 ! 3)-D-QuiNAc-(1! !3)-L-Rha-(1 ! 4)-D-GalNAcAN-(1 ! 4)-D-GalNFoA-(1 ! 3)-D-QuiNAc-(1! !3)-L-Rha-(1 ! 4)-D-GalNAc3OAcAN-(1 ! 4)-D-GalNFoA(N)-(1 ! 3)-b-D-QuiNAc-(1! !2)-L-Rha-(1 ! 4)-D-GalNAc3OAcAN-(1 ! 4)-D-GalNFoAN-(1 ! 3)-D-QuiNAc-(1! !3)-L-Rha-(1 ! 4)-D-GalNAcAN-(1 ! 4)-D-GalNFoA(N)-(1 ! 3)-D-QuiNAc-(1!

GalNFoAN amidation

– 10% – 20% 100% 50%

Adapted from Knirel et al.45

is a dehydrogenase which catalyses the NAD(P)þdependent 2-fold oxidation of hexoses in sugarnucleotide substrates to generate uronic acids.229,230 WbpP is a NADþ-binding SDR enzyme.231 On the basis of capillary electrophoresis, and spectrophotometric assays, WbpP is a sugar-nucleotide 4-epimerase230,232 which received detailed biochemical and structural scrutiny because it was the first sugarnucleotide 4-epimerase enzyme shown to have a substrate preference for UDP-D-GlcNAc/UDP-D-GalNAc over UDP-D-Glc/UDP-D-Gal.231,233 Due to somewhat relaxed substrate specificities in both enzymes, WbpO and WbpP can operate in either order to catalyse the overall conversion of UDP-DGlcNAc (1) to UDP-D-GalNAcA (47; Fig. 11). However, measurement of the kinetic and equilibrium parameters for all of the possible reactions catalysed by these enzymes indicated that in the major pathway in vivo, WbpO acts first, producing UDP-D-GlcNAcA (40). This is then followed by subsequent WbpP-catalysed 4-epimerisation to produce UDP-D-GalNAcA (47; Fig. 11).230 Biosynthesis of uronamides: wbpS

Uronamides are uncommon in nature, and the mechanism for biosynthesis of these sugars is unknown. They are present in the O polysaccharides of Shigella dysenteriae type 7,234 Francisella tularensis ssp. tularensis Schu S4,235,236 Bordetella spp.198 and E. coli O121.237 In three of these species, the O polysaccharide contains D-GalNAcAN, and the gene cluster contains a close homologue of the P. aeruginosa O6 O-Antigen locus gene wbpS (at the amino acid level, S. dysenteriae

WbpS, 54%; E. coli WbqG, 54%; F. tularensis WbtH, 48%). As WbpS has conserved glutamine-dependent amidotransferase domains, Feng and co-workers234 proposed that WbpS catalyses the transfer of ammonia to D-GalNAcA, producing D-GalNAcAN. Our laboratory has confirmed the role of WbpS in O-Antigen biosynthesis: a non-polar knockout mutation of wbpS abrogates O-Antigen production (J.D. King, V. Tran and J.S. Lam, unpublished data). Furthermore, the Nterminus of WbpS is homologous to the glutaminehydrolysing domain of asparagine synthase B.238 Point mutation of conserved amino acids in WbpS showed that this region functions to provide ammonia for O-Antigen synthesis. The point-mutant wbpS genes could fully complement the P. aeruginosa wbpS mutant when it was grown in ammonia-rich medium, but could not do so in minimal medium without ammonia (J.D. King and J.S. Lam, unpublished data). Biosynthesis of D-GalNAcA-like sugars in other serotypes

The O units of serotypes O13 and O14 contain D-GalNAc(3OAc)A. Since the O13/O14 O-Antigen cluster contains close homologues of WbpO and WbpP, sharing 81% and 69% identity with the O6 genes, respectively; the pathway for production of UDP-D-GalNAcA appears to be a conserved strategy in these serotypes. Characterisation of the O Antigen of a Homma K serogroup strain indicated that a proportion of O13 D-GalNAc3OAcA residues can be present as the uronamide (Homma is the name of a serotyping scheme commonly used in Japan, Homma K corresponds to IATS O13).239 However, no WbpS homologue was identified in the IATS O13 reference strain’s



294

OH O

WbpO

AcNH

OH HO HO

40

+ 2NAD+

O

O

HO HO

WbpP UDP

OH O

OH

WbpS

O

HO AcNH

UDP

UDP-D-GlcNAc 1

+ WbpP

OH

AcNH

2NAD+

OH

WbpO

+ ATP?, Gln/NH3

OH O

NH2

AcNH

UDP

UDP-D-GalNAcA 47

O

HO UDP

UDP-D-GalNAcAN 48

O HO AcNH

UDP

46 Fig. 11. UDP-D-GalNAcA and proposed UDP-D-GalNAcAN biosynthesis pathways. WbpO can catalyze the dehydrogenation of UDP-D-GlcNAc (1) or UDP-D-GalNAc (46); and WbpP is an epimerase that can catalyze the interconversion of UDP-D-GlcNAc (1) and UDP-D-GalNAc (46) or UDP-DGlcNAcA (40) and UDP-D-GalNAcA (47). The physiologically most relevant of the alternative pathways is first oxidation, then epimerisation, with compound 40 as intermediate.230 It has been proposed WbpS catalyses an amidotransfer reaction to generate the uronamide, GalNAcAN 234, but it is unknown whether this transformation occurs at the sugar nucleotide level, forming UDP-D-GalNAcAN (48) as shown here.

O-Antigen cluster. This may indicate that this O13 strain has an amidotransferase gene outside of this locus, or it may be that the sequenced O13 O-Antigen cluster was from a strain which does not produce uronamides. Biosynthesis of L-rhamnose (L-Rha) L-Rha

is a component of the LPS core oligosaccharide, and the L-Rha found in the O6 O Antigen is, therefore, presumably synthesised by products of the rml gene cluster as described in the section on core biosynthesis. This is probably the case for all serotypes which have LRha in their O units (O3, O4, O6, O10, O13, O14, O17 and O19). Glycosyltranferases: wbpR, wbpT and wbpU

In addition to the initiating glycosyl-1-phosphate transferase gene, wbpL (described above), there are three glycosyltransferase-encoding genes present in the O6 O-Antigen cluster, wbpR, wbpT and wbpU, and none of these have been investigated experimentally. The CAZy database123 classifies the products of all three of these genes in glycosyltransferase family GT-4. Based on the properties of characterised members of this family, GT-4 enzymes are believed to share the GT-B type fold and to catalyse the sugar transfer using a retaining mechanism, i.e. one in which the donor sugar stereochemistry at the anomeric centre is conserved in the product glycan.240 The O6 O-repeat structure contains three internal aglycosyl linkages. The enzyme which catalyses the addition of L-Rha in an a-(1 ! 4) linkage must be an inverting glycosyltransferase (L-Rha is derived from the activated precursor dTDP-b-L-Rha). The other internal linkages will be formed by retaining glycosyltransferases. Either the inverting L-Rha transferase is encoded outside of this gene cluster or this structure illustrates the limitations in the use of CAZy classifications to predict

glycosyltransferase properties. WbpT and WbpU both have relatively well-conserved homologues encoded in the E. coli O121 O-Antigen locus, WbqH (46% identical to WbpT) and WbqI (52% identical to WbpU). The E. coli O121 O-Antigen repeat has the structure !4)-DGalNAc(3OAc)AN-(1 ! 4)-D-GalNAcA-(1 ! 3)-DGlcNAc-(1 ! 3)-b-D-Qui4NGlyAc-(1!,237 which contains two very similar glycosyl bonds to those in the P. aeruginosa O6 O Antigen (in bold). On this basis, it is likely that WbpT and WbpU are responsible for catalyzing addition of the D-GalNAc3OAcAN-(1 ! 4)D-GalNFoA-(1 ! 3) disaccharide. However, this prediction must be verified by experimental investigation.

Biosynthesis of the P. aeruginosa O11 O Antigen O11 was the second P. aeruginosa serotype for which the O-Antigen biosynthesis cluster was sequenced (Table 11).203 It is the shortest of the P. aeruginosa O-Antigen loci (Fig. S3), reflecting the relative simplicity of the O repeat. The repeating unit contains D-Glc and both stereoisomers of FucNAc. As D-Glc is a component of LPS core, the biosynthesis of the sugar-nucleotide precursor for this sugar is discussed above. Therefore in this section we focus on the biosynthesis of D- and L-FucNAc. Biosynthesis of 2-acetamido-2-deoxy-D-fucose (D-FucNAc): wbpM and wbjF

The wbpM gene in the O11 O-Antigen locus encodes a protein which shares 96% sequence identity with the PAO1 sequence, and is required for O11 O-Antigen production. Mutation of wbpM in this serotype appears to place the cells under stress, because they apparently pick up second-site mutations with high frequency.105



295

Table 11. Genes present in the O-Antigen biosynthesis cluster of P. aeruginosa IATS serotype O11 strain Gene wzz wzx wbjA wzy wbjB

wbjC wbjD wbjE wbjF wbpL wbpM


78% 66% 37% 40% 54% 40% 54% 53% 20% 52% 60% 96%

E. coli O26 FnlA S. aureus Cap5E H. pylori FlaA1 C. jejuni 11168 PseB E. coli O26 FnlB S. aureus Cap5F E. coli O26 FnlC S. aureus Cap5G S. aureus Cap5L P. aeruginosa O5 WbpK P. aeruginosa O5 WbpL P. aeruginosa O5 WbpM


Key reference*

Chain-length regulator O-Antigen repeat unit flippase Glycosyltransferase (GT-2) O-Antigen polymerase UDP-D-GlcNAc 4,6-dehydratase/5-epimerase

203 212

UDP-6-deoxy-4-keto-L-IdoNAc 3-epimerase/4-reductase

249

UDP-PneNAc/UDP-FucNAc 2-epimerase

249

Glycosyltransferase (GT-4) UDP-4-keto-D-FucNAc 4-reductase (UDP-FucNAc forming) Initiating glycosyl-1-phosphate transferase UDP-D-GlcNAc 4,6-dehydratase

105


The O11 wbjF gene is believed to encode a functional homologue of WbpK from serotype O5. This is hypothesised because P. aeruginosa O11 O-unit biosynthesis is initiated with D-FucNAc94 and because the WbjF sequence is closer to WbpKO5 (52% identity) than WbpVO6 (40%). If this is the case, then UDP-D-FucNAc (34) is synthesised in serotype O11 in a scheme like that of serotype O5 (Fig. 9) The function of the product encoded by wbjF has not, however, been investigated experimentally. Biosynthesis of 2-acetamido-2-deoxy-L-fucose (L-FucNAc): wbjB, wbjC and wbjD

The biosynthesis of L-FucNAc has been studied in detail, because it is a constituent of O-Antigen or capsule polysaccharides in a number of pathogenic bacteria including P. aeruginosa serotypes O4, O11,241 S. aureus serotypes 5242 and 8,243 Streptococcus pneumoniae capsule type 4,244 and E. coli O4,245 O26246 and O172.247 Specifically, the pathways in P. aeruginosa O11 and S. aureus type 5, have been investigated both using genetics and biochemistry (Fig. 9).210,212,248,249 The O11 genes wbjB, wbjC and wbjD constitute a conserved UDP-L-FucNAc (39) biosynthesis cassette, with homologues present in other L-FucNAc-producing bacteria. In E. coli, these genes are annotated as fnlA, fnlB and fnlC, or fnl1, fnl2 and fnl3; and in S. aureus as capE, capF and capG, often with the capsular serotype of the originating strain indicated; for example, cap5E is capE from S. aureus type 5. The E. coli genes are the closest to their P. aeruginosa homologues, FnlA, FnlB, and FnlC sharing 78%, 52–54% and 52–54% sequence identity with WbjB, WbjC and WbjD, respectively.

wbjB is essential for O-Antigen production in P. aeruginosa O11.248 cap5E encodes an enzyme which shares 66% sequence identity with WbjB, and is functionally equivalent, as demonstrated by both crosscomplementation of the wbjB mutation with cap5E, and by capillary electrophoresis (CE), and CE-MS analysis of in vitro, enzyme–substrate incubation products.248 The WbjB-catalysed reaction is a C5-inverting 4,6dehydration in which UDP-D-GlcNAc is converted to UDP-2-acetamido-2,6-dideoxy-L-arabino-4-hexulose (UDP-6-deoxy-4-keto-L-IdoNAc; 36); (Fig. 9). As discussed above, in the section on wbpM, this is the same reaction as is catalysed by H. pylori FlaA1 (37% identity with WbjB) and C. jejuni PseB (40% identity with WbjB) in the pseudaminic acid (Pse, 5,7-diacetamido3,5,7,9-tetradeoxy-L-glycero-L-manno-nonulosonic acid) biosynthesis pathway.211 In a side reaction, all of these functional homologues also catalyse racemisation of the C5 stereocentre in the product hexulose moiety producing UDP-4-keto-D-QuiNAc (33),210,212 but this reaction is not relevant to UDP-L-FucNAc (39) biosynthesis. WbjB is a member of the SDR family and, like WbpM, has an unusual TMK catalytic triad (the WbjB triad was originally reported as SMK, but re-alignment with the Pfam01370 motif using BLAST identifies Thr124 rather than Ser123 as the conserved active site residue). WbjB requires NADPþ to catalyse the dehydratase reaction.248 WbjC is a bifunctional enzyme catalysing C3-epimerisation of the WbjB-catalysed reaction product (36), and then reduction of the C4 ketone to an alcohol (38). The product of the second step is UDP-2-acetamino-2,6dideoxy-L-talose (UDP-L-PneNAc, 38) as determined by


296


NMR analysis of the purified reaction product (Fig. 9).249 WbjC is another SDR family protein, and requires a reduced-form dinucleotide co-factor to catalyse the ketone reduction without any specific preference for either NADH or NADPH.248 WbjC catalyses the same reaction as Cap5F of S. aureus (40% sequence identity) in vitro.248 The final step in the pathway is catalysed by WbjD. Cap5G (53% identity with WbjD) is functionally equivalent in vitro but, in this case, the S. aureus enzyme was inactive at 37 C.248 These enzymes have the sugarnucleotide 2-epimerase conserved domain (Pfam02350), and catalyse the epimerisation of UDP-L-PneNAc (38) to UDP-L-FucNAc (39; Fig. 9). As is typical of epimerisation reactions, total conversion of the starting material does not occur, and the two isomers attain equilibrium.248 In vitro, the proportion of UDP-LFucNAc at equilibrium can be maximised by performing the reaction at pH 9, an observation which facilitated purification and NMR characterisation of the final product of the pathway by NMR.249 L-FucNAc is also found in the O-unit of serotype O4. The O-Antigen biosynthesis cluster of serotype O4 has three genes orf9, orf10 and orf11, which encode homologues of the O11 UDP-L-FucNAc biosynthesis enzymes WbjB, WbjC, and WbjD with very high identity (88%, 74%, and 83%, respectively).

Glycosyltranferases: wbjA and wbjE

In addition to WbpLO11, which probably initiates OAntigen synthesis by transfer of D-FucNAc-1-phosphate to undecaprenyl-phosphate, the O11 O-Antigen gene cluster encodes two putative glycosyltransferases. wbjA encodes a protein classified in the CAZy family GT-2. To date, characterised members of this family use an inverting mechanism and the GT-A three-dimensional fold.240 wbjE encodes a CAZy GT-4 family protein which would suggest that WbjE has the GT-B fold and retains the anomeric configuration of the transferred sugar. As formation of the two internal glycosyl bonds in the O11 repeat (namely the L-FucNAc-a-(1! linkage from a UDP-b-L-FucNAc donor, and D-Glc-b-(1! linkage from a presumed UDP-a-D-Glc donor) requires two inverting glycosyltransferases, the CAZy classification of these enzymes does not help distinguish the functions of WbjA and WbjE. WbjE and Cap5L, a putative glycosyltransferase from the S. aureus type 5 capsule locus, share rather limited sequence conservation (20% identity); on this basis, we very tentatively suggest that these two enzymes may transfer the LFucNAc present in each of their respective polysaccharides.

Biosynthesis of sugars in other serotypes While the O5, O6 and O11 O-Antigen clusters have been studied in some detail, very little direct experimental investigation has been conducted into the functions of genes in the remaining O-Antigen loci. To our knowledge, only one report has been published, which concerns the function of lfnA in serotype O12.226 Based on comparison of O-Antigen cluster gene sequences and the repeat unit structures between serotypes, it seems that where the same sugar appears in more than one serotype, the biosynthetic apparatus is also conserved. The identification of homologues of characterised O5, O6 and O11 genes in the other serotypes, therefore, allows us to annotate the other O-Antigen clusters intelligently. In this section, we illustrate how our understanding of the biosynthetic pathways for O-Antigen sugars in O5, O6, and O11 provides a useful basis on which to generate hypotheses regarding the biosynthesis of some sugars required for the other serotypes. Biosynthesis of 2-acetamidino-2-deoxy-L-fucose (L-FucNAm): lfnA

The P. aeruginosa serotype O12 O repeat contains a 2,6dideoxy-2-acetamidino-L-galactose (L-FucNAm) sugar. Since the O12 O-Antigen biosynthesis cluster contains close homologues of the serotype O11 genes wbjB, wbjC and wbjD (sharing 87%, 71% and 83% amino acid identity, respectively), it is likely that the L-FucNAm biosynthetic pathway involves initial synthesis of UDPL-FucNAc (39), and subsequent modification of the acetamido group by an amidotransfer reaction. Our group has investigated the genetics of this modification reaction and shown that, without a functional copy of the lfnA gene, P. aeruginosa O12 produces an O Antigen which contains L-FucNAc rather than L-FucNAm. The lfnA mutant can also be complemented by in trans expression of the E. coli O145 gene wbuX (LfnA and WbuX share 71% identity).226 Furthermore, LfnA contains conserved domains consistent with a role activating its substrate to accept the ammonia nucleophile in an amidotransfer reaction, including a partial alignment with Pfam00733 (C-terminal domain of asparagine synthase). While these experiments demonstrated that lfnA is necessary for synthesis of the acetamidino moiety, they do not prove that LfnA is sufficient to catalyse this reaction by itself. Indeed, introduction of lfnA alone into other serotypes of P. aeruginosa which have L-FucNAc as a component of the O repeat (O4 and O11) apparently did not result in changes to the O-Antigen structure.226 One of the possible reasons for this result may be that other proteins encoded in the large O12 O-Antigen locus are required


Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa for acetamidino group synthesis. The O12 locus orf10 is a homologue of Campylobacter coli ptmE gene (40% identity at amino acid level). It has been postulated that ptmE was required for acetamidino-modification of one of the C. coli flagellar glycan, though it should be noted that the sugars in question were only tentatively identified by MS.250 Furthermore, many amidotransferase enzymes have a second domain which hydrolyses glutamine to provide ammonia, but sequence analysis predicts that LfnA has no such domain. However, a glutamine hydrolysis domain (Pfam00117, glutamine amidotransferase class I) is apparently encoded, by orf18 which is immediately downstream of lfnA. orf10 and orf18 may be required, in addition to lfnA, for acetamidino group synthesis. Since mutation in the lfnA homologue, PseA, in C. jejuni results in loss of acetamidino-containing species from the sugar-nucleotide pool,227 the current working hypothesis is that the amidotransfer reaction which requires lfnA also occurs at the sugar-nucleotide level and produces UDP-L-FucNAm (49) as the precursor to the L-FucNAm found in the O repeat (Fig. 6). Further work is required to verify (or refute) this hypothesis. Biosynthesis of nonulosonic acids

Serotypes O7, O8, O9 and O12 contain nonulosonic acids at the non-reducing termini of their O-Antigen repeats. In O7, O8 and O9 these residues are derivatives of pseudaminic acid (Pse); in O12 this is a derivative of 5,7-diamino-3,5,7,9-tetradeoxy-L-glycero-D-galactonon-2-ulosonic acid (8-epilegionaminic acid, 8eLeg). The biosyntheses of these sugars have not been investigated in the context of P. aeruginosa LPS but probably follow pathways analogous to the Pse pathway in C. jejuni211 and the Leg pathway in Legionella pneumophila.251 In these pathways, after biosynthesis of a UDP-activated aminohexose sugar, the distinctive enzymatic steps are: 1. Removal of the UDP. In the pseudaminic acid pathway this is catalysed by a nuclotidase PseG,211 in the neuraminic acid and legionaminic acid pathway this is accomplished by a homologue of NeuC which

297

catalyses both inversion of the stereochemistry at C2 and also removal of the UDP moiety.251,252 2. Condensation of the aminohexose derived from UDPGlcNAc, with phosphoenolpyruvate (PEP), catalysed by a NeuB homologue (PseI in pseudaminic acid biosynthesis). 3. CMP-activation of the resulting nonulosonic acid with consumption of CTP, which is catalysed by NeuA in E. coli K1, or its homologue PseF in the pseudaminic acid pathway. BLAST analysis of P. aeruginosa sequences using the sequences of characterised enzymes identifies candidates for these genes in the O7, O8, O9 and O12 O-Antigen biosynthesis clusters (Table 12). Biosynthesis of 2-acetamido-4-O-acetyl-2-deoxy-D-fucose (D-FucNAc4OAc): wbpC

Modification of the D-FucNAc residue with a 4-O-acetyl substituent is observed in P. aeruginosa serotype O20. The gene which encodes this modification is currently unknown; a candidate is the wbpC gene within the O-Antigen biosynthesis cluster, which shows similarity to acyltransferases. The WbpC sequence has an acyltransferase 3 superfamily conserved domain (Pfam01757).253 wbpC is also present in the O-Antigen clusters of other members of the O2/O5/O16/O18/O20 serogroup. Knockout mutation of wbpC in P. aeruginosa serotype O5 made no noticeable change to the LPS phenotype; however, because O20 is the only serotype which produces D-FucNAc4OAc, no phenotype would be predicted (C. Wenzel and J.S. Lam, unpublished data). Reverse-transcriptase PCR detected wbpC mRNA in both serotypes O5 and O20 indicating that, if WbpC is indeed responsible for 4-O-acetylation of D-FucNAc in O20, the reason this activity is not seen in O5 is not because the wbpC gene is not transcribed (E.L. Westman and J.S. Lam, unpublished data). Additional research is required to clarify the role of wbpC; construction of a wbpC knockout in P. aeruginosa serotype O20 would be the next logical experiment.

Table 12. Homologues of distinctive nonulosonic acid biosynthesis enzymes in P. aeruginosa O-Antigen gene clusters BLAST query

Homologues from O7/O8 cluster

Homologues from O9 cluster

C. jejuni PseG E. coli K1 NeuC E. coli K1 NeuB E. coli K1 NeuA

Orf11 (25%)

Orf 9 (29%)

Orf13 (35%) Orf10 (24%)

Orf11 (31%) Orf8 (27%)

Homologues from O12 cluster

Orf7 (32%) Orf 8 (39%) Orf12 (29%)

Amino acid sequence identities are listed in parentheses. Alignments were produced using the PSI-BLAST algorithm, interrogating P. aeruginosa sequences at Genbank. Downloaded from http://ini.sagepub.com at UNIV OF GUELPH on October 6, 2009

298


Biosynthesis of 2,3-diacetamido-2,3-dideoxy-L-guluronic acid (L-GulNAc3NAcA) and derivatives: hypothetical AlgG-like activity

same scheme that has been described in P. aeruginosa PAO1.

The O-Antigen repeats of P. aeruginosa serotypes O2, O18, and O20 contain a guluronic acid derivative, 2,3-acetamido-2,3-dideoxy-L-guluronic acid or 2-acetamido-3-acetamidino-2,3-dideoxy-L-guluronic acid (LGulNAc3NAcA or L-GulNAc3NAmA). The L-gulo configuration can be produced by 5-epimerisation of the corresponding D-manno-hexose. For example, alginate, an exopolysaccharide of P. aeruginosa, is synthesised as a polymer of D-mannuronic acid before epimerisation of some of the residues to L-guluronic acid by the mannuronan 5-epimerase, AlgG.254 Thus, a UDP-D-ManNAc3NAcA 5-epimerase could produce UDP-L-GulNAc3NAcA, or a putative AlgG-like 5epimerase could act on D-ManNAc3NAcA residues in the assembled O-Antigen polymer. No candidate 5epimerase genes are present in the O-Antigen biosynthesis cluster, so this function may be encoded outside of this locus.

New annotation of selected O-Antigen biosynthesis clusters

Biosynthesis of D-xylose (D-Xyl): serogroup O7/O8 orf16 (arnA homologue)

Serotypes O7 and O8 can be treated together from the perspective of O-Antigen sugar biosynthesis, since their O-Antigen biosynthesis clusters contain highly conserved sets of genes.161 The O-Antigen repeating units of O7 and O8 both contain D-Xyl.45 UDP-D-Xyl could be synthesised from UDP-D-GlcA in a pathway involving a decarboxylase reaction. The C-terminal domain of ArnA from E. coli catalyzes the oxidation and decarboxylation of UDP-GlcA to UDP-4-keto-arabinose. P. aeruginosa PAO1 also encodes an ArnA homologue in the arn locus (see lipid A section). orf16 of the serogroup O7/O8 OAntigen locus encodes an SDR enzyme with 29% amino acid identity with the C-terminal region of ArnAPAO1. Possibly, orf16 encodes the necessary decarboxylase, but the functions of SDR enzymes cannot be predicted from percentage identity alone.255 Biosynthesis of 2,3-diacetamido-2,3-dideoxy-D-glucuronic acid (D-GlcNAc3NAcA): wbpA, wbpB, wbpE, and wbpD

The O Antigen of serotype O1 contains DGlcNAc3NAcA. UDP-D-GlcNAc3NAcA (43) is produced as an intermediate in P. aeruginosa serotype O5; the O Antigen of serotype O5 contains the 2-epimer, DManNAc3NAcA, for which the pathway has been described (Fig. 10). The predicted protein sequences of Orf6, Orf7, Orf8, and Orf9 from the serotype O1 O-Antigen cluster show 475% identity to WbpA, WbpB, WbpD, and WbpE of serotype O5 (Table 13). Thus, these genes are expected to be involved in the biosynthesis of UDP-D-GlcNAc3NAcA (43), using the

While the majority of P. aeruginosa O-Antigen biosynthesis genes has not been directly investigated, sufficient information has been provided by the characterisation of genes and their products from O5, O6 and O11 (see previous sections) to allow rational prediction of the pathways for synthesis of most sugars found in the O repeats of other serotypes. With this notion, we outline below, a preliminary annotation of the O-Antigen biosynthesis clusters of O1, O4 and O13/O14. These annotations will form a basis for understanding O-Antigen biosynthesis in these serotypes. Serotype O1

This serotype is clinically prevalent, constituting, for example, 21% of P. aeruginosa clinical isolates identified in a recent study.256 The O-Antigen cluster of serotype O1 (Fig. S4) contains 15 ORFs with similarity to genes from the well-studied serotypes O5 and O6 (Table 13). Our assignment of genes is based on homology to characterised genes from serotypes O5 and O6, conserved motifs and predicted transmembrane helices (Table 13). For example, there are two ORFs of serotype O1 that have 19–20% identity to Wzx of serotype O6 (WzxO6, 474 amino acids); however, orf5 encodes only 91 amino acids and so, if it encodes any function at all, it is very unlikely to encode the same function as WzxO6. Therefore the gene encoding a fulllength homologue of WzxO6 was annotated as wzx. The wzy annotation was based on 30% amino acid identity to Wzy of E. coli O42 and multiple predicted transmembrane helices. One challenge in annotation of the serotype O1 gene cluster arises because the O-Antigen repeat unit contains D-GalNAc, D-FucNAc, and D-QuiNAc. In other serotypes, all of these sugars are synthesized from UDP-DGlcNAc in reactions catalysed by specific SDR enzymes in addition to WbpM (Figs 9 and 11). Enzymes are identifiable as SDRs through the conservation of several motifs.257 However, members of this large protein family catalyse such a wide range of chemical reactions, including oxidation, reduction, epimerisation, dehydration and decarboxylation, and typically share 15–25% identity with other SDR enzymes, including those that share the same catalytic chemistry, and those that do not. Enzyme chemistry is determined by the presence or absence of a small number of amino acid side chains, which are often not readily identifiable in sequence alignments.255 This means that, where well-conserved



299

Table 13. Genes present in the O-Antigen biosynthesis cluster of P. aeruginosa IATS serotype O1 strain Gene/ORF


Proposed function

wzz/orf4 orf5* wbpA/orf6 wbpB/orf7 wbpD/orf8 wbpE/orf9 orf10 wzx/orf11 wzy/orf12 orf13

69% 19% 85% 76% 82% 76%

O-Antigen chain length regulator

orf14

orf15

wbpV/orf16

wbpL/orf17 wbpM/orf18

P. aeruginosa O5 Wzz P. aeruginosa O6 Wzx P. aeruginosa O5 WbpA P. aeruginosa O5 WbpB P. aeruginosa O5 WbpD P. aeruginosa O5 WbpE

20% P. aeruginosa O6 WzxO6 30% E. coli O42 Wzy 37% P. aeruginosa O5 MigA (rhamnosyltransferase) 98% P. aeruginosa O9 Orf14 23% P. aeruginosa O11 WbjF 23% P. aeruginosa O6 WbpV 21% P. aeruginosa O5 WbpK 28% P. aeruginosa O5 WbpZ 27% P. aeruginosa O6 WbpU 26% P. aeruginosa O6 WbpT 98% P. aeruginosa O9 Orf16 60% P. aeruginosa O6 WbpV 37% P. aeruginosa O5 WbpK 25% P. aeruginosa O6 WbpP 63% P. aeruginosa O5 WbpL 97% P. aeruginosa O5 WbpM

UDP-D-GlcNAc 6-dehydrogenase UDP-D-GlcNAcA 3-oxidase UDP-D-GlcNAc3NA 3-acetyltransferase UDP-3-keto-D-GlcNAcA 3-transaminase glycosyltransferase (GT-4) probably of D-GalNAc or D-FucNAc O-unit flippase O-Antigen polymerase Glycosyltransferase (GT-2) probably of D-GlcNAc3NAcA SDR probably involved in production of D-GalNAc or D-FucNAc

Glycosyltransferase (GT-4) probably of D-GalNAc or D-FucNAc

UDP-4-keto-D-QuiNAc 4-reductase (UDP-D-QuiNAc synthesising)

Initiating glycosyl-1-phosphate transferase UDP-D-GlcNAc 4,6-dehydratase

*Truncated sequence (91 vs 474 amino acids in P. aeruginosa O6 Wzx) Glycosyltransferase (GT) family classification is taken from the CAZy database.123

homologues do not exist, experimental characterisation is the only option for annotation of SDR enzymes. Despite the hypothetical requirement for three SDR enzymes to produce UDP-D-GalNAc, UDP-D-FucNAc, and UDP-D-QuiNAc, the O-Antigen cluster of serotype O1 encodes only two putative SDRs in addition to WbpM: orf14 and orf16. orf16 encodes a close homologue of WbpV (60% identity), which is the 4-reductase of serotype O6 involved in producing UDP-D-QuiNAc, suggesting that UDP-D-QuiNAc could be produced in the common manner (Fig. 9). However, the other SDR enzyme has limited similarity (525% identity) to any other P. aeruginosa gene except for orf14 of serotype O9, which has an unknown function (Table 13). Therefore, precise prediction of the enzyme function encoded by orf1401 is premature given the available information, but it is likely involved in the production of UDP-D-GalNAc and/or UDP-D-FucNAc. Identification of conserved genes putatively involved in the biosynthesis of common sugars among P. aeruginosa serotypes has allowed nearly complete annotation of the sugar-nucleotide pathways of serotype O1 (Table 13). Experimentation is required to clearly delineate the roles of the putative SDR enzymes in this serotype, particularly with respect to orf14.

Serotype O4

Serotype O4 strains have not been well-studied, but have recently been identified in Hungary with multidrug resistance and country-wide distribution.258 The serotype O4 O-Antigen cluster orf6 did not share significant similarity with any proteins in BLAST analysis. However, since orf6 encodes a protein with multiple predicted transmembrane helices and other encoded membrane proteins in the O4 cluster were all assigned other functions in our analysis, we annotated orf6 as a putative wzy gene (Table 14). The serotype O4 cluster (Fig. S4) contains several genes with homology (425% identity) to genes of serotype O11 (Table 14), which encode creation of L-FucNAc residues (Fig. 9); therefore, the functions of these genes are very likely conserved in the serotype O4 homologues. Based on the presence of these conserved genes and the production of conserved sugars, we have been able to present a complete annotation of the serotype O4 O-Antigen locus (Table 14). Serotypes O13/O14

Serotypes O13 and O14 can be treated together from the perspective of O-Antigen sugar biosynthesis, since their O-Antigen biosynthesis clusters contain highly


300


conserved sets of genes.161 Serotypes O13/O14 produce an O repeat unit that is very similar to that of serotype O6: all three serotypes produce O Antigens that contain L-Rha, D-QuiNAc, and D-GalNAc(3OAc)A (Table 4). Accordingly, many ORFs in the serotype O13/O14 gene cluster have a good degree of identity (432%) to genes of serotype O6 (Table 15). As with serotype O4, a putative wzy gene was identified based on the presence of multiple predicted transmembrane helices and by eliminating other candidates for this function within the cluster. As is often the case with Wzy sequences, no significant similarity to

other proteins in P. aeruginosa or other species was found by BLAST analysis of the protein encoded by this gene (Table 15). In serotype O14, a frameshift mutation has apparently occurred which likely renders the putative wzy non-functional; but polymerized O Antigens are still produced. These observations are consistent with the findings of Kaluzny et al.,259 that O2 and O16 O-specific polysaccharide polymerisation is encoded outside of the O-Antigen cluster, which must also be the case in O6, where there is no Wzy gene within the O-Antigen cluster.169 The putative wzy with frameshift in the O-Antigen cluster of serotype O14 thus

Table 14. Genes present in the O-Antigen biosynthesis cluster of P. aeruginosa IATS serotype O4 strain Gene/ORF


Proposed function

wzz/orf4 wzx/orf5 wzy/orf6

48% P. aeruginosa O5 Wzz 20% P. aeruginosa O6 Wzx Contains predicted transmembrane domains

Chain length regulator Und-PP-O unit flippase O-Antigen polymerase

orf7 orf8 wbjB/orf9 wbjC/orf10

88% P. aeruginosa O11 WbjB 74% P. aeruginosa O11 WbjC

wbjD/orf11 wbjE/orf12 wbpV/orf13

83% P. aeruginosa O11 WbjD 25% P. aeruginosa O11 WbjE 55% P. aeruginosa O6 WbpV



Glycosyltransferase (GT-2) probably of D-QuiNAc or L-FucNAc Glycosyltransferase (no CAZy family assigned) probably of D-QuiNAc or L-FucNAc UDP-D-GlcNAc 4,6-dehydratase and 5-epimerase UDP-2-acetamido-2,6-dideoxy-b-L-arabino-4-hexulose 3-epimerase and 4-reductase UDP-L-PneNAc 2-epimerase Glycosyltransferase (GT-4) probably of L-FucNAc UDP-4-keto-D-QuiNAc 4-reductase (UDP-D-QuiNAc synthesising) Initiating glycosyl-1-phosphate transferase UDP-D-GlcNAc 4,6-dehydratase

Glycosyltransferase (GT) family classification is taken from the CAZy database.123

Table 15. Genes present in the O-Antigen biosynthesis cluster of P. aeruginosa IATS serotype O13/O14 strain Gene/ORF


Proposed function

wzz/orf4 wbpO/orf5 wbpP/orf6 wzx/orf7 wzy/orf8

Chain length regulator UDP-D-GlcNAc/UDP-D-GalNAc 6-dehydrogenase UDP-D-GlcNAcA/UDP-D-GalNAcA 4-epimerase Und-PP-O unit flippase O-Antigen polymerase

wbpV/orf12

54% P. aeruginosa O5 Wzz 81% P. aeruginosa O6 WbpO 69% P. aeruginosa O6 WbpP 25% P. aeruginosa O5 Wzx Contains predicted transmembrane domains 26% P. aeruginosa O5 MigA 32% P. aeruginosa O6 WbpR 88% (O13) or 87% (O14) P. aeruginosa O6 WbpU 82% P. aeruginosa O6 WbpV



migA/orf9 wbpR/orf10 wbpU/orf11

Glycosyltransferase (GT-2), probably of L-Rha Glycosyltransferase (GT-4), probably of L-Rha Glycosyltransferase (GT-4), probably of D-GalNAc(3OAc)A UDP-4-keto-D-QuiNAc 4-reductase (UDP-D-QuiNAc synthesising) Initiating glycosyl-1-phosphate transferase UDP-D-GlcNAc 4,6-dehydratase

Glycosyltransferase (GT) family classification is taken from the CAZy database.123 Downloaded from http://ini.sagepub.com at UNIV OF GUELPH on October 6, 2009

Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa is likely complemented by another wzy encoded elsewhere in the genome. Unlike serotype O13, serotype O14 produces OAntigen with a lateral D-Glc residue (Table 4), which is probably derived from UDP-D-Glc and produced by enzymes involved in core biosynthesis. Collectively, these predictions have allowed us to present a complete annotation of the genes putatively required for sugarnucleotide biosynthesis in the serotype O13/O14 O-Antigen locus (Table 15). Glycosyltransferases encoded in the O-Antigen clusters

Assignment of the putative glycosyltransferases is problematic in all serotypes. Their classification into glycosyltransferase families provides some clues, but annotation of their specific substrates will require additional study. Little diversity of glycosyltransferase families is observed within the P. aeruginosa O-Antigen biosynthesis clusters, all belonging to either CAZy family GT-2 or GT-4. According to the CAZy database,123 GT-2 enzymes are expected to use the inverting mechanism and have the GT-A fold, while GT-4 proteins are predicted to use retaining chemistry and have the GT-B fold. The ability to predict accurately the inverting or retaining mechanism of glycosyltransferases based on their families would be very useful in assigning roles to individual gene products. For example, in serotype O1, the chemistry of the sugar linkages (Table 4) suggests that UDP-D-GlcNAc3NAcA should be transferred by an inverting glycosyltransferase to form the b-(1 ! 3) linkage, while the others are expected to be transferred by retaining glycosyltransferases. Orf10, Orf13, and Orf15 are candidate glycosyltransferases in serotype O1 (Table 13), but only Orf13 belongs to an inverting family (GT-2). Orf10 and Orf15 belong to a retaining family, GT-4. Thus, this analysis suggests that Orf13 transfers D-GlcNAc3NAcA, and either Orf10 or Orf15 transfer D-GalNAc or D-FucNAc. Based on its role as the initiating glycosyltransferase, WbpLO1 is expected to transfer D-QuiNAc, the first sugar of the repeating O-unit in serotype O1. However, in other serotypes, the glycosyltransferase family assignments appear to be inconsistent with the Orepeat structures, casting some general doubt on the utility of the CAZy group assignments. For example, serotype O4 appears to require three inverting glycosyltransferases to add the three L-sugars in a-linkages (Table 4). However, of the putative glycosyltransferases Orf7, Orf8, and wbjE04 (Table 14), only Orf7 belongs to an inverting glycosyltransferase family (GT-2). Orf8 has not been placed in any CAZy family, and wbjE04 belongs to the retaining glycosyltransferase family, GT-4. Despite the CAZy assignment of wbjE to a retaining

301

glycosyltransferase family, its 25% sequence identity to WbjEO11 (a proposed L-FucNAc transferase of serotype O11) and its position downstream of genes required for UDP-L-FucNAc biosynthesis (wbjB, wbjC, and wbjD; Fig. S4) lead us to propose that wbjE is an inverting LFucNAc transferase. In general, each of the O-Antigen clusters contains the correct number of putative glycosyltransferases as would be predicted from the number of sugar residues in the Orepeat. Serotype O14 is an exception, which produces a pentasaccharide repeat unit (Table 4) but contains only four putative glycosyltransferases in the O-Antigen biosynthesis cluster (Table 15). This is an unusual case because the fifth sugar, the lateral Glc, is not present in the core þ 1 O-unit (semi-rough LPS); this observation has not yet been explained. However, in all serotypes except O14, the agreement between the number of putative glycosyltransferase-encoding genes and the number of required glycosyltransferases indicates that assignment of the function as glycosyltransfer is probably correct, even if bioinformatics alone cannot determine the substrate or mechanism. This highlights one of the shortcomings of sequence analysis and underlines the importance of functional studies. Export, polymerisation and chain-length regulation of O-Specific Antigen After synthesis of the O-repeat units on their undecaprenyl pyrophosphate carriers, they must be polymerised and transported before ligation to lipid A-core. The proteins encoded by three genes, wzx wzy and wzz, are required for these processes. Transport of O-Antigen units across the inner membrane: wzx

The first studies of wzx were conducted in S. enterica and demonstrated that a strain carrying a wzx mutation in a plasmid-borne copy of the O-Antigen biosynthesis locus accumulated undecaprenol-pyrophosphate linked to a single O-Antigen repeat unit on the cytoplasmic side of the inner membrane.260 Due to this phenotype, Wzx is proposed to catalyze transmembrane transport of undecaprenol-pyrophosphate linked O units, and has been called the O-unit flippase. Burrows and Lam261 identified a putative wzx gene in P. aeruginosa PAO1 (serotype O5) based on the high hydrophobicity and large number of expected transmembrane spanning segments in the predicted protein sequence. Analysis of chromosomal knockout mutants of wzx showed that the mutants did not produce O Antigen. Strangely, the wzx knockouts also demonstrated a marked delay in the production of the Common Polysaccharide Antigen, which may be due to effects on WbpL since providing


302


an additional copy of wbpL in trans could complement the delay.261 Wzx proteins from some O-Antigen systems in the Enterobacteriaceae appear to have specificity for the first sugar attached to undecaprenoyl-phosphate,262 regardless of the structure of the remainder of the subunit.263 E. coli K-12/O16 produces LPS in which the first sugar is D-GlcNAc; in these experiments, wzx from S. enterica LT2 (which has D-Gal as the initial sugar) could not complement a knockout in wzx of E. coli K-12/ O16.262 However, wzx from S. flexneri 2a or E. coli serotypes that use D-GlcNAc or D-GalNAc as the initial sugar could complement the K-12/O16 wzx knockout. Of particular interest is the finding that wzx of P. aeruginosa PAO1, which uses D-FucNAc as the first sugar residue, could complement the wzx knockout in E. coli K-12/O16.262 This suggests that P. aeruginosa PAO1 Wzx belongs to the proposed subfamily of Wzx proteins that recognizes O units initiated by N-acetylhexosamines.262 Since P. aeruginosa serotypes mostly use either D-FucNAc or D-QuiNAc as the initiating sugar, we compared protein sequences to see if any correlation between Wzx sequence and the initiating sugar could be detected. If any correlation were detected, it would suggest that closely related Wzx proteins may share an adaptation for transport of substrates with the same initiating sugar. PSI-BLAST analysis was used to identify homologues of Wzx from serotype O6 (WxzO6), which was selected as an example of D-QuiNAc-initiated repeat units (Table 16). Almost all of the P. aeruginosa serotypes that produce D-QuiNAcinitiated repeat units were found to be homologous to WzxO6. Only serotypes O10/O19 (which have identical Wzx sequences) were not aligned with WxzO6 and instead were shown through separate searches to share 27% identity to Wzx of serotype O11, which has D-FucNAc as the initial sugar. When the same strategy was used to identify homologues of Wzx from PAO1 (serotype O5; WzxO5) as an example of D-FucNAcinitiated repeat units, only the closely related serogroup

O2/O5/O16/O18/O20 Wzx sequences were identified (identity 99–100%). Serotypes O7/O8 Wzx sequences are identical to each other, but produced no significant hits to any other Wzx of P. aeruginosa. Thus, the predicted Wzx amino acid sequences cluster into four groups: O5-like (serogroup O2/O5/O16/O18/O20), O6like (all strains with D-QuiNAc initiated O units except O10/O19), O11-like (O11 and O10/O19), and O7/O8. Of these, only the O11-like group contains members that use different initial sugars. Experimentation is required to determine whether the conservation of primary sequence between P. aeruginosa Wzx protein sequences genuinely indicates any substrate specificity. O-Antigen polymerase: wzy

A knockout of wzy produces semi-rough LPS (LPS with only one O-repeat unit). This phenotype has been described in knockout studies in P. aeruginosa PAO1 (serotype O5)163 and P. aeruginosa PA103 (serotype O11).203 Since wzy is necessary for polymerization of O-Antigen, it has been called the O-Antigen polymerase gene despite the lack of biochemical evidence for polymerase activity of the encoded protein from any bacterial species. The wzy (formerly rfc) genes from PAO1 and PA103 show no significant similarity at the nucleotide or protein level, but are predicted to share very high hydrophobicity and topology with 9–12 transmembrane spanning segments. Both characteristics are typical of other bacterial O-Antigen polymerases.264,265 Of all the well-studied serotypes, only O6 does not have a putative wzy gene within the O-Antigen cluster;169 presumably, the gene is located elsewhere in the genome. The closely related serogroup O2/O5/O16/O18/O20 have nearly identical O-Antigen biosynthesis clusters that contain a single wzy gene (Table 6).161 This serogroup produces O units that are similar in sugar composition, but with different linkages between the repeat units: serotypes O5, O18, and O20 have a-linked O units, whereas serotypes O2 and O16 have b-linked O units.217 The wzy gene of the O-Antigen biosynthesis

Table 16. Amino acid sequence identities to WzxO6 among P. aeruginosa serotypes with D-QuiNAc (or derivatives) as the initial sugar Serotype

ORF in O-antigen biosynthesis cluster

Initial sugar

Sequence identity with WzxO6

O1 O3 O4 O9 O10/O19 O12 O13/O14

Orf5 Orf7 Orf5 Orf5 Orf7 Orf15 Orf7

D-QuiNAc

21% over 147 amino acids 19% over 204 amino acids 20% over 240 amino acids 19% over 393 amino acids –, (27% with WzxO11) 13% over 375 amino acids 25% over 312 amino acids

D-QuiNAc4NSHb D-QuiNAc D-QuiNAc D-QuiNAc D-QuiNAc D-QuiNAc


Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa cluster (PA3154) has, therefore, sometimes been designated wzya to emphasise the proposed role of the encoded protein in polymerizing a-linked O-repeat units. An alternate form of Wzy, encoded by wzyb elsewhere in the genome, has been proposed to be responsible for the polymerization of O units with b-linkages.259 Reverse-transcriptase-PCR showed that the wzya gene in the O-Antigen biosynthesis cluster of P. aeruginosa serotypes O2 and O16 is transcribed, despite the fact that these two serotypes have b-linked O-units. The predicted protein sequence for wzya of serotypes O2 and O16 is 100% identical to that of O5 (which has a-linked O-units) and any of the three genes are sufficient to complement a wzya knockout in serotype O5.259 Serotypes O2 and O16 were each found to contain a wzyb gene encoding identical Wzyb proteins that is located outside of the O-Antigen biosynthesis cluster. The O16 wzyb gene was sufficient to cause the production of b-linked O-Antigen repeat units when expressed in an O5 wzya knockout strain.259 The expressed and functional Wzya of serotype O2 and O16 does not synthesize a-linked repeat units in these strains because of the presence of an inhibitor. This inhibitor of the a-polymerase, iap, was initially described as part of a three-component cassette from bacteriophage D3 that was required for serotype conversion of P. aeruginosa from O5 to O16 and O-acetylation of the D-FucNAc moiety.266 Chromosomal copies of Iap were located in serotypes O2 and O16, suggesting that iap inactivates Wzya while leaving Wzyb functional.259 O Antigen chain-length regulator: wzz

The O Antigen could theoretically contain any number of repeat units if polymerisation were not regulated. In wild-type P. aeruginosa serotype O5, a characteristic modal distribution of O-Antigen lengths exists, with O chains of approximately 12–16, 22–30, or 40–50 repeat units being the most common. The wzz genes of P. aeruginosa were initially studied in serotypes O5 and O16; knockouts of the O-Antigen cluster wzz synthesize O Antigens with a different distribution of chain lengths than the wild-type strains.267 The wzz mutant in serotype O5 produced proportionally more short O chains of approximately 1–12 repeat units, and wzz mutants in both O5 and O16 produced less midlength O chains of roughly 12–30 repeat units. Thus, the average O-Antigen chain length was altered in these wzz knockouts, but it seemed that some chain length modulation remained. The predicted Wzz protein of O5 has roughly 20% identity with Wzz of E. coli, S. enterica, and S. flexneri, and was able to function in E. coli to alter polysaccharide chain lengths when expressed in trans.267 The mode of action of Wzz proteins is

303

unknown but they all share similar topology and are required for regulation of the chain lengths of both O Antigens and capsules. Pseudomonas aeruginosa PAO1 possesses wzz1 (PA3160) and wzz2 (PA0938) that encode two distinct functional Wzz proteins.268 Each Wzz protein governs the production of LPS with different modal lengths; Wzz1 imparts LPS with a bimodal distribution of 12–16 or 22–30 repeat units, while Wzz2 imparts LPS with longer modal lengths (40–50 repeat units).268 A strain lacking wzz2 and unable to produce normal core oligosaccharide (wzz2 rmlC double mutant) was shown to lose wild-type modality and produce only undecaprenollinked O Antigen.268 When this strain was complemented with wzz2, the O Antigen was still undecaprenollinked due to the absence of core oligosaccharide acceptors for O polysaccharide caused by the rmlC knockout, but wild-type modality of the undecaprenollinked bands was restored. Thus, Wzz directs the preferential polymerization of modal chain lengths before the O Antigen is ligated to the core oligosaccharide.268 Wzz proteins were also found to co-precipitate with undecaprenol-linked O Antigens, indicating a direct or indirect interaction with the O-Antigen chain.268 A wzz1 homologue was identified using the GENEMARK algorithm;161 in all of the functional O-Antigen biosynthesis clusters. Using polyclonal antisera raised against Wzz2 of PAO1, which is located outside the O-Antigen cluster, Wzz2 was detected in all serotypes.268 The proposed wzz1 homologues share 440% identity to the wzz1 from PAO1, with the poorest conservation (41% identity) from serotype O11. Recent studies of P. aeruginosa PA103 (serotype O11) revealed the sequence of wzz2 of PA103, which shares 92% amino acid similarity with wzz2 of PAO1.269 Knockout mutants lacking wzz1 or wzz2 were constructed in PA103, and were shown to lack wild-type chain length distributions. Specifically, wzz1 was determined to produce O Antigens of ‘long’ chain length, and wzz2 was determined to produce O Antigens of ‘very long’ chain length,269 which is consistent with the observations made in an earlier study of these genes in PAO1.268 The PA103 wzz1 mutant demonstrated reduced virulence than either the wild-type PA103 or the wzz2 mutant, based on serum sensitivity assays and a mouse pneumonia model study, suggesting that the ‘long’ chain length of the O Antigen is particularly important for pathogenesis.269 Ligation of O Antigen to lipid A-core: waaL By different mechanisms, the Common Polysaccharide Antigen and O-specific polysaccharides are assembled and delivered as complete polymers to the periplasmic side of the inner membrane. The final step in


304


O-polysaccharide biosynthesis is the ligation of these O chains to lipid A-core, in a reaction which requires the product of the waaL gene. WaaL is known as the Opolysaccharide ligase (reviewed by Raetz and Whitfield50). There is little primary sequence conservation between the enzymes which perform this function, but identification of the P. aeruginosa waaL was possible because these proteins do share very similar hydropathy profiles, which are characterised by the presence of approximately 11 predicted transmembrane helices and a 50–100-amino acid loop near the C terminus, normally predicted to be in the periplasm. In PAO1, the waaL gene (PA4999) is located in the core cluster described above. Pseudomonas aeruginosa waaL mutants do not produce Common Polysaccharide Antigen-containing or O-Specific Antigen-containing LPS, or the semi-rough form. In Western blotting analysis of whole cell lysates of these mutants, polymers are present which are recognised by Common Polysaccharide Antigen-specific and O-specific antibodies. However, in aqueous phenol LPS extractions these polymers are not detected, suggesting that these bands are O-polysaccharide chains linked to an isoprenylpyrophosphate lipid carrier, that are probably hydrolysed in hot aqueous phenol.35 Site-directed mutagenesis of conserved amino acids in the probable periplasmic loop of P. aeruginosa WaaL showed that His303 is important for ligase function in complementation experiments, but that another positively-charged amino acid side chain can be placed in this position without loss of function. The P. aeruginosa WaaL enzyme was purified and shown to hydrolyse ATP in vitro suggesting that the O-Antigen ligation reaction may be an ATP-dependent process. Site-directed mutagenesis studies showed that conserved amino acids in what had been proposed as potential ATP binding/hydrolysis motifs,270 were not essential for ligation by the E. coli WaaL.271 Therefore, further work is required to establish the link between ATP hydrolysis and the ligation reaction. Escherichia coli K-12 waaL cannot cross-complement a P. aeruginosa waaL mutation.270 Since the outer core structures of E. coli and P. aeruginosa are different, the failure to cross-complement is probably due to specificity of WaaL enzymes for structural features of their cognate core oligosaccharide acceptors.272

CONCLUDING REMARKS Pseudomonas aeruginosa LPS biosynthesis has proven to be an excellent biological system in which to study polysaccharide biosynthesis and many details are now known about the biogenesis of this macromolecule. In the last decade, the publication of the O-Antigen cluster sequences, the annotation of the PAO1 genome, and the

very detailed characterisation of LPS structures from a large number of wild-type and mutant P. aeruginosa strains have provided a valuable opportunity to advance our understanding of LPS genetics. Many features of P. aeruginosa LPS structures and pathways are conserved in surface polysaccharides of pathogenic bacteria, so the work summarised in this review also illuminates the biosynthesis of virulence factors in other medically important organisms.

ACKNOWLEDGEMENTS Research in our laboratory is supported by a grant from the Canadian Institutes of Health Research (CIHR; MOP 14687) and an operating grant from the Canadian Cystic Fibrosis Foundation (CCFF). ELW holds a CIHR Canada Graduate Scholarship Doctoral Research Award, DK holds a CCFF Postdoctoral Fellowship, and JSL holds a Canada Research Chair in Cystic Fibrosis and Microbial Glycobiology, funded jointly by the Canadian Foundation for Innovation and the Ontario Innovation Trust.

REFERENCES 1. Villavicencio RT. The history of blue pus. J Am Coll Surg 1998; 187: 212–216. 2. Cardo D, Horan T, Andrus M et al. National Nosocomial Infections Surveillance (NNIS) System Report, data summary from January 1992 through June 2004, issued October 2004. Am J Infect Control 2004; 32: 470–485. 3. Govan JR, Deretic V. Microbial pathogenesis in cystic fibrosis: mucoid Pseudomonas aeruginosa and Burkholderia cepacia. Microbiol Rev 1996; 60: 539–574. 4. Lang AB, Horn MP, Imboden MA, Zuercher AW. Prophylaxis and therapy of Pseudomonas aeruginosa infection in cystic fibrosis and immunocompromised patients. Vaccine 2004; 22 (Suppl 1): S44–S48. 5. Taneja N, Emmanuel R, Chari PS, Sharma M. A prospective study of hospital-acquired infections in burn patients at a tertiary care referral centre in North India. Burns 2004; 30: 665–669. 6. Lyczak JB, Cannon CL, Pier GB. Establishment of Pseudomonas aeruginosa infection: lessons from a versatile opportunist. Microbes Infect 2000; 2: 1051–1060. 7. Poole K. Efflux-mediated multiresistance in Gram-negative bacteria. Clin Microbiol Infect 2004; 10: 12–26. 8. Yoshimura F, Nikaido H. Permeability of Pseudomonas aeruginosa outer membrane to hydrophilic solutes. J Bacteriol 1982; 152: 636–642. 9. Nikaido H, Hancock REW. Outer membrane permeability of Pseudomonas aeruginosa. In: Sokatch JR. (ed). The Bacteria, a Treatise on Structure and Function. Orlando, FL: Academic Press, 1986; 145–193. 10. El’Garch F, Jeannot K, Hocquet D, Llanes-Barakat C, Plesiat P. Cumulative effects of several nonenzymatic mechanisms on the resistance of Pseudomonas aeruginosa to aminoglycosides. Antimicrob Agents Chemother 2007; 51: 1016–1021.


Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa 11. Giamarellou H. Prescribing guidelines for severe Pseudomonas infections. J Antimicrob Chemother 2002; 49: 229–233. 12. Foweraker J. Recent advances in the microbiology of respiratory tract infection in cystic fibrosis. Br Med Bull 2009; 89: 93–110. 13. Rocchetta HL, Burrows LL, Lam JS. Genetics of O-antigen biosynthesis in Pseudomonas aeruginosa. Microbiol Mol Biol Rev 1999; 63: 523–553. 14. Stanislavsky ES, Lam JS. Pseudomonas aeruginosa antigens as potential vaccines. FEMS Microbiol Rev 1997; 21: 243–277. 15. Galanos C, Lu¨deritz O, Rietschel ET et al. Synthetic and natural Escherichia coli free lipid A express identical endotoxic activities. Eur J Biochem 1985; 148: 1–5. 16. Alexander C, Rietschel ET. Bacterial lipopolysaccharides and innate immunity. J Endotoxin Res 2001; 7: 167–202. 17. Backhed F, Normark S, Schweda EK, Oscarson S, RichterDahlfors A. Structural requirements for TLR4-mediated LPS signalling: a biological role for LPS modifications. Microbes Infect 2003; 5: 1057–1063. 18. Hajjar AM, Ernst RK, Tsai JH, Wilson CB, Miller SI. Human Toll-like receptor 4 recognizes host-specific LPS modifications. Nat Immunol 2002; 3: 354–359. 19. Ernst RK, Hajjar AM, Tsai JH, Moskowitz SM, Wilson CB, Miller SI. Pseudomonas aeruginosa lipid A diversity and its recognition by Toll-like receptor 4. J Endotoxin Res 2003; 9: 395–400. 20. Hancock RE, Diamond G. The role of cationic antimicrobial peptides in innate host defences. Trends Microbiol 2000; 8: 402–410. 21. Trent MS, Stead CM, Tran AX, Hankins JV. Diversity of endotoxin and its impact on pathogenesis. J Endotoxin Res 2006; 12: 205–223. 22. Ernst RK, Yi EC, Guo L et al. Specific lipopolysaccharide found in cystic fibrosis airway Pseudomonas aeruginosa. Science 1999; 286: 1561–1565. 23. McPhee JB, Lewenza S, Hancock RE. Cationic antimicrobial peptides activate a two-component regulatory system, PmrAPmrB, that regulates resistance to polymyxin B and cationic antimicrobial peptides in Pseudomonas aeruginosa. Mol Microbiol 2003; 50: 205–217. 24. Pier GB, Grout M, Zaidi TS, Goldberg JB. How mutant CFTR may contribute to Pseudomonas aeruginosa infection in cystic fibrosis. Am J Respir Crit Care Med 1996; 154: S175–S182. 25. Willcox MD. Pseudomonas aeruginosa infection and inflammation during contact lens wear: a review. Optom Vis Sci 2007; 84: 273–278. 26. Fleiszig SM, Zaidi TS, Fletcher EL, Preston MJ, Pier GB. Pseudomonas aeruginosa invades corneal epithelial cells during experimental infection. Infect Immun 1994; 62: 3485–3493. 27. Fleiszig SM, Zaidi TS, Pier GB. Pseudomonas aeruginosa invasion of and multiplication within corneal epithelial cells in vitro. Infect Immun 1995; 63: 4072–4077. 28. Zaidi TS, Fleiszig SM, Preston MJ, Goldberg JB, Pier GB. Lipopolysaccharide outer core is a ligand for corneal cell binding and ingestion of Pseudomonas aeruginosa. Invest Ophthalmol Vis Sci 1996; 37: 976–986. 29. Zaidi TS, Lyczak J, Preston M, Pier GB. Cystic fibrosis transmembrane conductance regulator-mediated corneal epithelial cell ingestion of Pseudomonas aeruginosa is a key component in the pathogenesis of experimental murine keratitis. Infect Immun 1999; 67: 1481–1492. 30. Pier GB, Grout M, Zaidi TS. Cystic fibrosis transmembrane conductance regulator is an epithelial cell receptor for clearance of Pseudomonas aeruginosa from the lung. Proc Natl Acad Sci USA 1997; 94: 12088–12093.

305

31. Schroeder TH, Reiniger N, Meluleni G, Grout M, Coleman FT, Pier GB. Transgenic cystic fibrosis mice exhibit reduced early clearance of Pseudomonas aeruginosa from the respiratory tract. J Immunol 2001; 166: 7410–7418. 32. Dasgupta T, de Kievit TR, Masoud H et al. Characterization of lipopolysaccharide-deficient mutants of Pseudomonas aeruginosa derived from serotypes O3, O5, and O6. Infect Immun 1994; 62: 809–817. 33. Lam MY, McGroarty EJ, Kropinski AM et al. Occurrence of a common lipopolysaccharide antigen in standard and clinical strains of Pseudomonas aeruginosa. J Clin Microbiol 1989; 27: 962–967. 34. Berry MC, McGhee GC, Zhao Y, Sundin GW. Effect of a waaL mutation on lipopolysaccharide composition, oxidative stress survival, and virulence in Erwinia amylovora. FEMS Microbiol Lett 2009; 291: 80–87. 35. Abeyrathne PD, Daniels C, Poon KK, Matewish MJ, Lam JS. Functional characterization of WaaL, a ligase associated with linking O-antigen polysaccharide to the core of Pseudomonas aeruginosa lipopolysaccharide. J Bacteriol 2005; 187: 3002–3012. 36. Hancock RE, Mutharia LM, Chan L, Darveau RP, Speert DP, Pier GB. Pseudomonas aeruginosa isolates from patients with cystic fibrosis: a class of serum-sensitive, nontypable strains deficient in lipopolysaccharide O side chains. Infect Immun 1983; 42: 170–177. 37. Evans DJ, Pier GB, Coyne MJJr, Goldberg JB. The rfb locus from Pseudomonas aeruginosa strain PA103 promotes the expression of O antigen by both LPS-rough and LPS-smooth isolates from cystic fibrosis patients. Mol Microbiol 1994; 13: 427–434. 38. Kadurugamuwa JL, Lam JS, Beveridge TJ. Interaction of gentamicin with the A band and B band lipopolysaccharides of Pseudomonas aeruginosa and its possible lethal effect. Antimicrob Agents Chemother 1993; 37: 715–721. 39. Hatano K, Goldberg JB, Pier GB. Biologic activities of antibodies to the neutral-polysaccharide component of the Pseudomonas aeruginosa lipopolysaccharide are blocked by O side chains and mucoid exopolysaccharide (alginate). Infect Immun 1995; 63: 21–26. 40. Matewish MJ. The functional role of lipopolysaccharide in the cell envelope and surface proteins of Pseudomonas aeruginosa. PhD thesis. University of Guelph, Ontario, Canada, 2004. 41. Cryz SJJr , Pitt TL, Furer E, Germanier R. Role of lipopolysaccharide in virulence of Pseudomonas aeruginosa. Infect Immun 1984; 44: 508–513. 42. Priebe GP, Dean CR, Zaidi T et al. The galU gene of Pseudomonas aeruginosa is required for corneal infection and efficient systemic spread following pneumonia but not for infection confined to the lung. Infect Immun 2004; 72: 4224–4232. 43. Tang HB, DiMango E, Bryan R et al. Contribution of specific Pseudomonas aeruginosa virulence factors to pathogenesis of pneumonia in a neonatal mouse model of infection. Infect Immun 1996; 64: 37–43. 44. Coyne MJJr, Russell KS, Coyle CL, Goldberg JB. The Pseudomonas aeruginosa algC gene encodes phosphoglucomutase, required for the synthesis of a complete lipopolysaccharide core. J Bacteriol 1994; 176: 3500–3507. 45. Knirel YA, Bystrova OV, Kocharova NA, Za¨hringer U, Pier GB. Conserved and variable structural features in the lipopolysaccharide of Pseudomonas aeruginosa. J Endotoxin Res 2006; 12: 324–336. 46. Kulshin VA, Za¨hringer U, Lindner B, Ja¨ger KE, Dmitriev BA, Rietschel ET. Structural characterization of the lipid A component


306

47.

48.

49.

50. 51. 52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

King, Kocıńcova´, Westman, Lam of Pseudomonas aeruginosa wild-type and rough mutant lipopolysaccharides. Eur J Biochem 1991; 198: 697–704. Ernst RK, Moskowitz SM, Emerson JC et al. Unique lipid A modifications in Pseudomonas aeruginosa isolated from the airways of patients with cystic fibrosis. J Infect Dis 2007; 196: 1088–1092. Bhat R, Marx A, Galanos C, Conrad RS. Structural studies of lipid A from Pseudomonas aeruginosa PAO1: occurrence of 4-amino4-deoxyarabinose. J Bacteriol 1990; 172: 6631–6636. Moskowitz SM, Ernst RK, Miller SI. PmrAB, a two-component regulatory system of Pseudomonas aeruginosa that modulates resistance to cationic antimicrobial peptides and addition of aminoarabinose to lipid A. J Bacteriol 2004; 186: 575–579. Raetz CR, Whitfield C. Lipopolysaccharide endotoxins. Annu Rev Biochem 2002; 71: 635–700. Trent MS. Biosynthesis, transport, and modification of lipid A. Biochem Cell Biol 2004; 82: 71–86. Dotson GD, Kaltashov IA, Cotter RJ, Raetz CR. Expression cloning of a Pseudomonas gene encoding a hydroxydecanoyl-acyl carrier protein-dependent UDP-GlcNAc acyltransferase. J Bacteriol 1998; 180: 330–337. Williamson JM, Anderson MS, Raetz CR. Acyl-acyl carrier protein specificity of UDP-GlcNAc acyltransferases from Gramnegative bacteria: relationship to lipid A structure. J Bacteriol 1991; 173: 3591–3596. Wyckoff TJ, Lin S, Cotter RJ, Dotson GD, Raetz CR. Hydrocarbon rulers in UDP-N-acetylglucosamine acyltransferases. J Biol Chem 1998; 273: 32369–32372. Clementz T, Bednarski JJ, Raetz CR. Function of the htrB high temperature requirement gene of Escherichia coli in the acylation of lipid A: HtrB catalyzed incorporation of laurate. J Biol Chem 1996; 271: 12095–12102. Clementz T, Zhou Z, Raetz CR. Function of the Escherichia coli msbB gene, a multicopy suppressor of htrB knockouts, in the acylation of lipid A. Acylation by MsbB follows laurate incorporation by HtrB. J Biol Chem 1997; 272: 10353–10360. Brozek KA, Raetz CR. Biosynthesis of lipid A in Escherichia coli. Acyl carrier protein-dependent incorporation of laurate and myristate. J Biol Chem 1990; 265: 15410–15417. Goldman RC, Doran CC, Kadam SK, Capobianco JO. Lipid A precursor from Pseudomonas aeruginosa is completely acylated prior to addition of 3-deoxy-D-manno-octulosonate. J Biol Chem 1988; 263: 5217–5223. Mohan S, Raetz CR. Endotoxin biosynthesis in Pseudomonas aeruginosa: enzymatic incorporation of laurate before 3-deoxy-Dmanno-octulosonate. J Bacteriol 1994; 176: 6944–6951. Guo L, Lim KB, Poduje CM et al. Lipid A acylation and bacterial resistance against vertebrate antimicrobial peptides. Cell 1998; 95: 189–198. Gibbons HS, Lin S, Cotter RJ, Raetz CR. Oxygen requirement for the biosynthesis of the S-2-hydroxymyristate moiety in Salmonella typhimurium lipid A. Function of LpxO, A new Fe2þ/alpha-ketoglutarate-dependent dioxygenase homologue. J Biol Chem 2000; 275: 32940–32949. Gibbons HS, Reynolds CM, Guan Z, Raetz CR. An inner membrane dioxygenase that generates the 2-hydroxymyristate moiety of Salmonella lipid A. Biochemistry 2008; 47: 2814–2825. Gibbons HS, Kalb SR, Cotter RJ, Raetz CR. Role of Mg2þ and pH in the modification of Salmonella lipid A after endocytosis by macrophage tumour cells. Mol Microbiol 2005; 55: 425–440. Trent MS, Pabich W, Raetz CR, Miller SI. A PhoP/PhoQ-induced lipase (PagL) that catalyzes 3-O-deacylation of lipid A precursors in membranes of Salmonella typhimurium. J Biol Chem 2001; 276: 9083–9092.

65. Geurtsen J, Steeghs L, Hove JT, van der Ley P, Tommassen J. Dissemination of lipid A deacylases (PagL) among Gramnegative bacteria: identification of active-site histidine and serine residues. J Biol Chem 2005; 280: 8248–8259. 66. Ernst RK, Adams KN, Moskowitz SM et al. The Pseudomonas aeruginosa lipid A deacylase: selection for expression and loss within the cystic fibrosis airway. J Bacteriol 2006; 188: 191–201. 67. Rutten L, Geurtsen J, Lambert W et al. Crystal structure and catalytic mechanism of the LPS 3-O-deacylase PagL from Pseudomonas aeruginosa. Proc Natl Acad Sci USA 2006; 103: 7071–7076. 68. Yan A, Guan Z, Raetz CR. An undecaprenyl phosphateaminoarabinose flippase required for polymyxin resistance in Escherichia coli. J Biol Chem 2007; 282: 36077–36089. 69. Hung RJ, Chien HS, Lin RZ et al. Comparative analysis of two UDP-glucose dehydrogenases in Pseudomonas aeruginosa PAO1. J Biol Chem 2007; 282: 17738–17748. 70. McPhee JB, Bains M, Winsor G et al. Contribution of the PhoPPhoQ and PmrA-PmrB two-component regulatory systems to Mg2þ-induced gene regulation in Pseudomonas aeruginosa. J Bacteriol 2006; 188: 3995–4006. 71. Nummila K, Kilpelainen I, Za¨hringer U, Vaara M, Helander IM. Lipopolysaccharides of polymyxin B-resistant mutants of Escherichia coli are extensively substituted by 2-aminoethyl pyrophosphate and contain aminoarabinose in lipid A. Mol Microbiol 1995; 16: 271–278. 72. Helander IM, Kato Y, Kilpelainen I et al. Characterization of lipopolysaccharides of polymyxin-resistant and polymyxinsensitive Klebsiella pneumoniae O3. Eur J Biochem 1996; 237: 272–278. 73. Gunn JS, Lim KB, Krueger J et al. PmrA-PmrB-regulated genes necessary for 4-aminoarabinose lipid A modification and polymyxin resistance. Mol Microbiol 1998; 27: 1171–1182. 74. McCoy AJ, Liu H, Falla TJ, Gunn JS. Identification of Proteus mirabilis mutants with increased sensitivity to antimicrobial peptides. Antimicrob Agents Chemother 2001; 45: 2030–2037. 75. Derzelle S, Turlin E, Duchaud E et al. The PhoP-PhoQ twocomponent regulatory system of Photorhabdus luminescens is essential for virulence in insects. J Bacteriol 2004; 186: 1270–1279. 76. Robey M, O’Connell W, Cianciotto NP. Identification of Legionella pneumophila rcp, a pagP-like gene that confers resistance to cationic antimicrobial peptides and promotes intracellular infection. Infect Immun 2001; 69: 4276–4286. 77. Preston A, Maxim E, Toland E et al. Bordetella bronchiseptica PagP is a Bvg-regulated lipid A palmitoyl transferase that is required for persistent colonization of the mouse respiratory tract. Mol Microbiol 2003; 48: 725–736. 78. Brozek KA, Bulawa CE, Raetz CR. Biosynthesis of lipid A precursors in Escherichia coli. A membrane-bound enzyme that transfers a palmitoyl residue from a glycerophospholipid to lipid X. J Biol Chem 1987; 262: 5170–5179. 79. Bishop RE, Gibbons HS, Guina T, Trent MS, Miller SI, Raetz CR. Transfer of palmitate from phospholipids to lipid A in outer membranes of Gram-negative bacteria. EMBO J 2000; 19: 5071–5080. 80. El Hamidi A, Novikov A, Karibian D, Perry MB, Caroff M. Structural characterization of Bordetella parapertussis lipid A. J Lipid Res 2009; 50: 854–859. 81. MacArthur I, Mann PB, Harvill ET, Preston A. IEIIS Meeting minireview: Bordetella evolution: lipid A and Toll-like receptor 4. J Endotoxin Res 2007; 13: 243–247. 82. Walsh AG, Matewish MJ, Burrows LL, Monteiro MA, Perry MB, Lam JS. Lipopolysaccharide core phosphates are required for



83. 84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

viability and intrinsic drug resistance in Pseudomonas aeruginosa. Mol Microbiol 2000; 35: 718–727. Schnaitman CA, Klena JD. Genetics of lipopolysaccharide biosynthesis in enteric bacteria. Microbiol Rev 1993; 57: 655–682. Eagon R, Carson K. Lysis of cell walls and intact cells of Pseudomonas aeruginosa by ethylenediamine tetraacetic acid and by lysosozyme. Can J Microbiol 1965; 11: 193–201. Gray GW, Wilkinson SG. The effect of ethylenediaminetetraacetic acid on the cell walls of some Gram-negative bacteria. J Gen Microbiol 1965; 39: 385–399. Wilkinson SG. Cell walls of Pseudomonas species sensitive to ethylenediaminetetraacetic acid. J Bacteriol 1970; 104: 1035–1044. Sanchez Carballo PM, Rietschel ET, Kosma P, Za¨hringer U. Elucidation of the structure of an alanine-lacking core tetrasaccharide trisphosphate from the lipopolysaccharide of Pseudomonas aeruginosa mutant H4. Eur J Biochem 1999; 261: 500–508. Knirel YA, Bystrova OV, Shashkov AS et al. Structural analysis of the lipopolysaccharide core of a rough, cystic fibrosis isolate of Pseudomonas aeruginosa. Eur J Biochem 2001; 268: 4708–4719. Bystrova OV, Shashkov AS, Kocharova NA et al. Structural studies on the core and the O-polysaccharide repeating unit of Pseudomonas aeruginosa immunotype 1 lipopolysaccharide. Eur J Biochem 2002; 269: 2194–2203. Kooistra O, Bedoux G, Brecker L et al. Structure of a highly phosphorylated lipopolysaccharide core in the Delta algC mutants derived from Pseudomonas aeruginosa wild-type strains PAO1 (serogroup O5) and PAC1R (serogroup O3). Carbohydr Res 2003; 338: 2667–2677. Beckmann F, Moll H, Ja¨ger KE, Za¨hringer U. Preliminary communication 7-O-carbamoyl-L-glycero-D-manno-heptose: a new core constituent in the lipopolysaccharide of Pseudomonas aeruginosa. Carbohydr Res 1995; 267: C3–C7. Sadovskaya I, Brisson JR, Lam JS, Richards JC, Altman E. Structural elucidation of the lipopolysaccharide core regions of the wild-type strain PAO1 and O-chain-deficient mutant strains AK1401 and AK1012 from Pseudomonas aeruginosa serotype O5. Eur J Biochem 1998; 255: 673–684. Sadovskaya I, Brisson JR, Thibault P, Richards JC, Lam JS, Altman E. Structural characterization of the outer core and the Ochain linkage region of lipopolysaccharide from Pseudomonas aeruginosa serotype O5. Eur J Biochem 2000; 267: 1640–1650. Bystrova OV, Knirel YA, Lindner B et al. Structures of the core oligosaccharide and O-units in the R- and SR-type lipopolysaccharides of reference strains of Pseudomonas aeruginosa Oserogroups. FEMS Immunol Med Microbiol 2006; 46: 85–99. Bystrova OV, Lindner B, Moll H et al. Structure of the lipopolysaccharide of Pseudomonas aeruginosa O-12 with a randomly O-acetylated core region. Carbohydr Res 2003; 338: 1895–1905. Choudhury B, Carlson RW, Goldberg JB. The structure of the lipopolysaccharide from a galU mutant of Pseudomonas aeruginosa serogroup-O11. Carbohydr Res 2005; 340: 2761–2772. Drewry DT, Gray GW, Wilkinson SG. Low-molecular-weight solutes released during mild acid hydrolysis of the lipopolysaccharide of Pseudomonas aeruginosa. Biochem J 1972; 130: 289–295. Wilkinson SG. 31P N.M.R. evidence for the presence of triphosphate residues in lipopolysaccharides from Pseudomonas aeruginosa. Biochem J 1981; 199: 833–835. Choudhury B, Carlson RW, Goldberg JB. Characterization of the lipopolysaccharide from a wbjE mutant of the serogroup O11 Pseudomonas aeruginosa strain, PA103. Carbohydr Res 2008; 343: 238–248.

307

100. Bystrova OV, Lindner B, Moll H et al. Structure of the biological repeating unit of the O-antigen of Pseudomonas aeruginosa immunotype 4 containing both 2-acetamido-2,6-dideoxy-Dglucose and 2-acetamido-2,6-dideoxy-D-galactose. Carbohydr Res 2003; 338: 1801–1806. 101. de Kievit TR, Lam JS. Monoclonal antibodies that distinguish inner core, outer core, and lipid A regions of Pseudomonas aeruginosa lipopolysaccharide. J Bacteriol 1994; 176: 7129–7139. 102. Poon KK, Westman EL, Vinogradov E, Jin S, Lam JS. Functional characterization of MigA and WapR: putative rhamnosyltransferases involved in outer core oligosaccharide biosynthesis of Pseudomonas aeruginosa. J Bacteriol 2008; 190: 1857–1865. 103. Postma PW, Lengeler JW, Jacobson GR. Phosphoenolpyruvate:carbohydrate phosphotransferase systems of bacteria. Microbiol Rev 1993; 57: 543–594. 104. Temple L, Cuskey SM, Perkins RE et al. Analysis of cloned structural and regulatory genes for carbohydrate utilization in Pseudomonas aeruginosa PAO. J Bacteriol 1990; 172: 6396–6402. 105. Dean CR, Goldberg JB. Pseudomonas aeruginosa galU is required for a complete lipopolysaccharide core and repairs a secondary mutation in a PA103 (serogroup O11) wbpM mutant. FEMS Microbiol Lett 2002; 210: 277–283. 106. Giraud MF, Naismith JH. The rhamnose pathway. Curr Opin Struct Biol 2000; 10: 687–696. 107. Graninger M, Nidetzky B, Heinrichs DE, Whitfield C, Messner P. Characterization of dTDP-4-dehydrorhamnose 3,5epimerase and dTDP-4-dehydrorhamnose reductase, required for dTDP-L-rhamnose biosynthesis in Salmonella enterica serovar Typhimurium LT2. J Biol Chem 1999; 274: 25069–25077. 108. Blankenfeldt W, Asuncion M, Lam JS, Naismith JH. The structural basis of the catalytic mechanism and regulation of glucose-1-phosphate thymidylyltransferase (RmlA). EMBO J 2000; 19: 6652–6663. 109. Blankenfeldt W, Giraud MF, Leonard G et al. The purification, crystallization and preliminary structural characterization of glucose-1-phosphate thymidylyltransferase (RmlA), the first enzyme of the dTDP-L-rhamnose synthesis pathway from Pseudomonas aeruginosa. Acta Crystallogr D Biol Crystallogr 2000; 56: 1501–1504. 110. Charnock SJ, Davies GJ. Structure of the nucleotide-diphosphosugar transferase, SpsA from Bacillus subtilis, in native and nucleotide-complexed forms. Biochemistry 1999; 38: 6380–6385. 111. Melo A, Glaser L. The nucleotide specificity and feedback control of thymidine diphosphate D-glucose pyrophosphorylase. J Biol Chem 1965; 240: 398–405. 112. Dong C, Major LL, Srikannathasan V et al. RmlC, a C30 and C50 carbohydrate epimerase, appears to operate via an intermediate with an unusual twist boat conformation. J Mol Biol 2007; 365: 146–159. 113. Rahim R, Burrows LL, Monteiro MA, Perry MB, Lam JS. Involvement of the rml locus in core oligosaccharide and O polysaccharide assembly in Pseudomonas aeruginosa. Microbiology 2000; 146 (Pt 11): 2803–2814. 114. Raetz CR. Biochemistry of endotoxins. Annu Rev Biochem 1990; 59: 129–170. 115. Walsh AG, Burrows LL, Lam JS. Genetic and biochemical characterization of an operon involved in the biosynthesis of 3deoxy-D-manno-octulosonic acid in Pseudomonas aeruginosa. FEMS Microbiol Lett 1999; 173: 27–33. 116. Lebioda L, Stec B. Crystal structure of enolase indicates that enolase and pyruvate kinase evolved from a common ancestor. Nature 1988; 333: 683–686.


308


117. Weng M, Makaroff CA, Zalkin H. Nucleotide sequence of Escherichia coli pyrG encoding CTP synthetase. J Biol Chem 1986; 261: 5568–5574. 118. Ray PH, Benedict CD, Grasmuk H. Purification and characterization of cytidine 50 -triphosphate:cytidine 50 -monophosphate-3deoxy-D-manno-octulosonate cytidylyltransferase. J Bacteriol 1981; 145: 1273–1280. 119. Gronow S, Oertelt C, Ervela E et al. Characterization of the physiological substrate for lipopolysaccharide heptosyltransferases I and II. J Endotoxin Res 2001; 7: 263–270. 120. Eidels L, Osborn MJ. Lipopolysaccharide and aldoheptose biosynthesis in transketolase mutants of Salmonella typhimurium. Proc Natl Acad Sci USA 1971; 68: 1673–1677. 121. Valvano MA, Marolda CL, Bittner M, Glaskin-Clay M, Simon TL, Klena JD. The rfaE gene from Escherichia coli encodes a bifunctional protein involved in biosynthesis of the lipopolysaccharide core precursor ADP-L-glycero-D-mannoheptose. J Bacteriol 2000; 182: 488–497. 122. Valvano MA, Messner P, Kosma P. Novel pathways for biosynthesis of nucleotide-activated glycero-manno-heptose precursors of bacterial glycoproteins and cell surface polysaccharides. Microbiology 2002; 148: 1979–1989. 123. Cantarel BL, Coutinho PM, Rancurel C, Bernard T, Lombard V, Henrissat B. The Carbohydrate-Active EnZymes database (CAZy): an expert resource for glycogenomics. Nucleic Acids Res 2009; 37: D233–D238. 124. Flint J, Taylor E, Yang M et al. Structural dissection and highthroughput screening of mannosylglycerate synthase. Nat Struct Mol Biol 2005; 12: 608–614. 125. Belunis CJ, Raetz CR. Biosynthesis of endotoxins. Purification and catalytic properties of 3-deoxy-D-manno-octulosonic acid transferase from Escherichia coli. J Biol Chem 1992; 267: 9988–9997. 126. Pradel E, Parker CT, Schnaitman CA. Structures of the rfaB, rfaI, rfaJ, and rfaS genes of Escherichia coli K-12 and their roles in assembly of the lipopolysaccharide core. J Bacteriol 1992; 174: 4736–4745. 127. Hirokawa T, Boon-Chieng S, Mitaku S. SOSUI: classification and secondary structure prediction system for membrane proteins. Bioinformatics 1998; 14: 378–379. 128. Krogh A, Larsson B, von Heijne G, Sonnhammer EL. Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. J Mol Biol 2001; 305: 567–580. 129. Roantree RJ, Kuo TT, MacPhee DG. The effect of defined lipopolysaccharide core defects upon antibiotic resistances of Salmonella typhimurium. J Gen Microbiol 1977; 103: 223–234. 130. Chatterjee AK, Sanderson KE, Ross H. Influence of temperature on growth of lipopolysaccharide-deficient (rough) mutants of Salmonella typhimurium and Salmonella minnesota. Can J Microbiol 1976; 22: 1540–1548. 131. de Kievit TR, Lam JS. Isolation and characterization of two genes, waaC (rfaC) and waaF (rfaF), involved in Pseudomonas aeruginosa serotype O5 inner-core biosynthesis. J Bacteriol 1997; 179: 3451–3457. 132. Yethon JA, Vinogradov E, Perry MB, Whitfield C. Mutation of the lipopolysaccharide core glycosyltransferase encoded by waaG destabilizes the outer membrane of Escherichia coli by interfering with core phosphorylation. J Bacteriol 2000; 182: 5620–5623. 133. Knirel YA, Helbig JH, Za¨hringer U. Structure of a decasaccharide isolated by mild acid degradation and dephosphorylation of the lipopolysaccharide of Pseudomonas fluorescens strain ATCC 49271. Carbohydr Res 1996; 283: 129–139.

134. Leone S, Izzo V, Silipo A et al. A novel type of highly negatively charged lipooligosaccharide from Pseudomonas stutzeri OX1 possessing two 4,6-O-(1-carboxy)-ethylidene residues in the outer core region. Eur J Biochem 2004; 271: 2691–2704. 135. Zdorovenko EL, Vinogradov E, Zdorovenko GM et al. Structure of the core oligosaccharide of a rough-type lipopolysaccharide of Pseudomonas syringae pv. phaseolicola. Eur J Biochem 2004; 271: 4968–4977. 136. Frirdich E, Vinogradov E, Whitfield C. Biosynthesis of a novel 3-deoxy-D-manno-oct-2-ulosonic acid-containing outer core oligosaccharide in the lipopolysaccharide of Klebsiella pneumoniae. J Biol Chem 2004; 279: 27928–27940. 137. Regue´ M, Izquierdo L, Fresno S et al. The incorporation of glucosamine into enterobacterial core lipopolysaccharide: two enzymatic steps are required. J Biol Chem 2005; 280: 36648–36656. 138. Yang H, Matewish M, Loubens I, Storey DG, Lam JS, Jin S. migA, a quorum-responsive gene of Pseudomonas aeruginosa, is highly expressed in the cystic fibrosis lung environment and modifies low-molecular-mass lipopolysaccharide. Microbiology 2000; 146: 2509–2519. 139. Wang J, Lory S, Ramphal R, Jin S. Isolation and characterization of Pseudomonas aeruginosa genes inducible by respiratory mucus derived from cystic fibrosis patients. Mol Microbiol 1996; 22: 1005–1012. 140. Zhao X, Lam JS. WaaP of Pseudomonas aeruginosa is a novel eukaryotic type protein-tyrosine kinase as well as a sugar kinase essential for the biosynthesis of core lipopolysaccharide. J Biol Chem 2002; 277: 4722–4730. 141. Zhao X, Wenzel CQ, Lam JS. Nonradiolabeling assay for WaaP, an essential sugar kinase involved in biosynthesis of core lipopolysaccharide of Pseudomonas aeruginosa. Antimicrob Agents Chemother 2002; 46: 2035–2037. 142. To T. Purification and characterization of WapP of Pseudomonas aeruginosa, a putative lipopolysaccharide kinase. MSc thesis. University of Guelph, Ontario, Canada, 2006. 143. Coque JJ, Perez-Llarena FJ, Enguita FJ, Fuente JL, Martin JF, Liras P. Characterization of the cmcH genes of Nocardia lactamdurans and Streptomyces clavuligerus encoding a functional 3-hydroxymethylcephem O-carbamoyltransferase for cephamycin biosynthesis. Gene 1995; 162: 21–27. 144. Scheeff ED, Bourne PE. Structural evolution of the protein kinase-like superfamily. PLoS Comput Biol 2005; 1: e49. 145. Tian Y, Jackson P, Gunter C, Wang J, Rock CO, Jackowski S. Placental thrombosis and spontaneous fetal death in mice deficient in ethanolamine kinase 2. J Biol Chem 2006; 281: 28438–28449. 146. Brodsky IE, Ghori N, Falkow S, Monack D. Mig-14 is an inner membrane-associated protein that promotes Salmonella typhimurium resistance to CRAMP, survival within activated macrophages and persistent infection. Mol Microbiol 2005; 55: 954–972. 147. Hasin M, Kennedy EP. Role of phosphatidylethanolamine in the biosynthesis of pyrophosphoethanolamine residues in the lipopolysaccharide of Escherichia coli. J Biol Chem 1982; 257: 12475–12477. 148. Tamayo R, Choudhury B, Septer A, Merighi M, Carlson R, Gunn JS. Identification of cptA, a PmrA-regulated locus required for phosphoethanolamine modification of the Salmonella enterica serovar Typhimurium lipopolysaccharide core. J Bacteriol 2005; 187: 3391–3399. 149. Cox AD, Li J, Richards JC. Identification and localization of glycine in the inner core lipopolysaccharide of Neisseria meningitidis. Eur J Biochem 2002; 269: 4169–4175.


Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa 150. Lundstro¨m SL, Li J, Deadman ME, Hood DW, Moxon ER, Schweda EK. Structural analysis of the lipopolysaccharide from nontypeable Haemophilus influenzae strain R2846. Biochemistry 2008; 47: 6025–6038. 151. Molinaro A, Silipo A, Castro CD et al. Full structural characterization of Shigella flexneri M90T serotype 5 wild-type R-LPS and its delta galU mutant: glycine residue location in the inner core of the lipopolysaccharide. Glycobiology 2008; 18: 260–269. 152. Gamian A, Krzyzaniak A, Barciszewska MZ, Gawronska I, Barciszewski J. Specific incorporation of glycine into bacterial lipopolysaccharide. Novel function of specific transfer ribonucleic acids. Nucleic Acids Res 1991; 19: 6021–6025. 153. Fox KL, Yildirim HH, Deadman ME, Schweda EK, Moxon ER, Hood DW. Novel lipopolysaccharide biosynthetic genes containing tetranucleotide repeats in Haemophilus influenzae, identification of a gene for adding O-acetyl groups. Mol Microbiol 2005; 58: 207–216. 154. Doerrler WT, Gibbons HS, Raetz CR. MsbA-dependent translocation of lipids across the inner membrane of Escherichia coli. J Biol Chem 2004; 279: 45102–45109. 155. Mamat U, Meredith TC, Aggarwal P et al. Single amino acid substitutions in either YhjD or MsbA confer viability to 3-deoxyD-manno-oct-2-ulosonic acid-depleted Escherichia coli. Mol Microbiol 2008; 67: 633–648. 156. Ghanei H, Abeyrathne PD, Lam JS. Biochemical characterization of MsbA from Pseudomonas aeruginosa. J Biol Chem 2007; 282: 26939–26947. 157. Liu PV, Matsumoto H, Kusama H, Bergan TOM. Survey of heatstable, major somatic antigens of Pseudomonas aeruginosa. Int J Syst Bacteriol 1983; 33: 256–264. 158. Liu PV, Wang S. Three new major somatic antigens of Pseudomonas aeruginosa. J Clin Microbiol 1990; 28: 922–925. 159. Akatova NS, Smirnova NE. [Serological classification of Pseudomonas aeruginosa]. Zh Mikrobiol Epidemiol Immunobiol 1982; 87–91. 160. La`nyi B, Bergan T. Serological characterization of Pseudomonas aeruginosa. Methods Microbiol 1978; 10: 93–168. 161. Raymond CK, Sims EH, Kas A et al. Genetic variation at the O-antigen biosynthetic locus in Pseudomonas aeruginosa. J Bacteriol 2002; 184: 3614–3622. 162. Rivera M, Bryan LE, Hancock RE, McGroarty EJ. Heterogeneity of lipopolysaccharides from Pseudomonas aeruginosa: analysis of lipopolysaccharide chain length. J Bacteriol 1988; 170: 512–521. 163. de Kievit TR, Dasgupta T, Schweizer H, Lam JS. Molecular cloning and characterization of the rfc gene of Pseudomonas aeruginosa (serotype O5). Mol Microbiol 1995; 16: 565–574. 164. Rocchetta HL, Lam JS. Identification and functional characterization of an ABC transport system involved in polysaccharide export of A-band lipopolysaccharide in Pseudomonas aeruginosa. J Bacteriol 1997; 179: 4713–4724. 165. Whitfield C. Biosynthesis of lipopolysaccharide O antigens. Trends Microbiol 1995; 3: 178–185. 166. Jiang XM, Neal B, Santiago F, Lee SJ, Romana LK, Reeves PR. Structure and sequence of the rfb (O antigen) gene cluster of Salmonella serovar Typhimurium (strain LT2). Mol Microbiol 1991; 5: 695–713. 167. Alexander DC, Valvano MA. Role of the rfe gene in the biosynthesis of the Escherichia coli O7-specific lipopolysaccharide and other O-specific polysaccharides containing Nacetylglucosamine. J Bacteriol 1994; 176: 7079–7084. 168. Zhang L, Radziejewska-Lebrecht J, Krajewska-Pietrasik D, Toivanen P, Skurnik M. Molecular and chemical characterization of the lipopolysaccharide O antigen and its role in the

169.

170.

171.

172.

173.

174.

175.

176.

177.

178.

179.

180.

181.

182.

309

virulence of Yersinia enterocolitica serotype O:8. Mol Microbiol 1997; 23: 63–76. Be´langer M, Burrows LL, Lam JS. Functional analysis of genes responsible for the synthesis of the B-band O antigen of Pseudomonas aeruginosa serotype O6 lipopolysaccharide. Microbiology 1999; 145: 3505–3521. Rocchetta HL, Burrows LL, Pacan JC, Lam JS. Three rhamnosyltransferases responsible for assembly of the A-band D-rhamnan polysaccharide in Pseudomonas aeruginosa: a fourth transferase, WbpL, is required for the initiation of both A-band and B-band lipopolysaccharide synthesis. Mol Microbiol 1998; 28: 1103–1119. Arsenault TL, Hughes DW, MacLean DB, Szarek WA, Kropinski AMB, Lam JS. Structural studies on the polysaccharide portion of ‘A-band’ lipopolysaccharide from a mutant (AK14O1) of Pseudomonas aeruginosa PAO1. Can J Chem 1991; 69: 1273–1280. Vinogradov E, Frirdich E, MacLean LL et al. Structures of lipopolysaccharides from Klebsiella pneumoniae. Elucidation of the structure of the linkage region between core and polysaccharide O chain and identification of the residues at the nonreducing termini of the O chains. J Biol Chem 2002; 277: 25070–25081. Jann K, Goldemann G, Weisgerber C, Wolf-Ullisch C, Kanegasaki S. Biosynthesis of the O9 antigen of Escherichia coli. Initial reaction and overall mechanism. Eur J Biochem 1982; 127: 157–164. Weisgerber C, Jann K. Glucosyldiphosphoundecaprenol, the mannose acceptor in the synthesis of the O9 antigen of Escherichia coli. Biosynthesis and characterization. Eur J Biochem 1982; 127: 165–168. Meier-Dieter U, Barr K, Starman R, Hatch L, Rick PD. Nucleotide sequence of the Escherichia coli rfe gene involved in the synthesis of enterobacterial common antigen. Molecular cloning of the rfe-rff gene cluster. J Biol Chem 1992; 267: 746–753. Rick PD, Hubbard GL, Barr K. Role of the rfe gene in the synthesis of the O8 antigen in Escherichia coli K-12. J Bacteriol 1994; 176: 2877–2884. Yokota S, Kaya S, Kawamura T, Araki Y, Ito E. The structure of the O-specific chain of lipopolysaccharide from Pseudomonas aeruginosa IID 1008 (ATCC 27584). J Biochem 1986; 99: 1551–1561. Rocchetta HL, Pacan JC, Lam JS. Synthesis of the A-band polysaccharide sugar D-rhamnose requires Rmd and WbpW: identification of multiple AlgA homologues, WbpW and ORF488, in Pseudomonas aeruginosa. Mol Microbiol 1998; 29: 1419–1434. Lightfoot J, Lam JS. Molecular cloning of genes involved with expression of A-band lipopolysaccharide, an antigenically conserved form, in Pseudomonas aeruginosa. J Bacteriol 1991; 173: 5624–5630. Shinabarger D, Berry A, May TB, Rothmel R, Fialho A, Chakrabarty AM. Purification and characterization of phosphomannose isomerase-guanosine diphospho-D-mannose pyrophosphorylase. A bifunctional enzyme in the alginate biosynthetic pathway of Pseudomonas aeruginosa. J Biol Chem 1991; 266: 2080–2088. Remminghorst U, Rehm BH. Bacterial alginates: from biosynthesis to applications. Biotechnol Lett 2006; 28: 1701–1712. Lee HJ, Chang HY, Venkatesan N, Peng HL. Identification of amino acid residues important for the phosphomannose isomerase activity of PslB in Pseudomonas aeruginosa PAO1. FEBS Lett 2008; 582: 3479–3483.


310


183. Ma L, Lu H, Sprinkle A, Parsek MR, Wozniak DJ. Pseudomonas aeruginosa Psl is a galactose- and mannose-rich exopolysaccharide. J Bacteriol 2007; 189: 8353–8356. 184. Zielinski NA, Chakrabarty AM, Berry A. Characterization and regulation of the Pseudomonas aeruginosa algC gene encoding phosphomannomutase. J Biol Chem 1991; 266: 9754–9763. 185. Olvera C, Goldberg JB, Sanchez R, Soberon-Chavez G. The Pseudomonas aeruginosa algC gene product participates in rhamnolipid biosynthesis. FEMS Microbiol Lett 1999; 179: 85–90. 186. Lightfoot J, Lam JS. Chromosomal mapping, expression and synthesis of lipopolysaccharide in Pseudomonas aeruginosa: a role for guanosine diphospho (GDP)-D-mannose. Mol Microbiol 1993; 8: 771–782. 187. King JD, Poon KK, Webb NA et al. The structural basis for catalytic function of GMD and RMD, two closely related enzymes from the GDP-D-rhamnose biosynthesis pathway. FEBS J 2009; 276: 2688–2700. 188. Webb NA, Mulichak AM, Lam JS, Rocchetta HL, Garavito RM. Crystal structure of a tetrameric GDP-D-mannose 4,6-dehydratase from a bacterial GDP-D-rhamnose biosynthetic pathway. Protein Sci 2004; 13: 529–539. 189. Kido N, Torgov VI, Sugiyama T et al. Expression of the O9 polysaccharide of Escherichia coli: sequencing of the E. coli O9 rfb gene cluster, characterization of mannosyl transferases, and evidence for an ATP-binding cassette transport system. J Bacteriol 1995; 177: 2178–2187. 190. Kido N, Morooka N, Paeng N et al. Production of monoclonal antibody discriminating serological difference in Escherichia coli O9 and O9a polysaccharides. Microbiol Immunol 1997; 41: 519–525. 191. Kido N, Sugiyama T, Yokochi T, Kobayashi H, Okawa Y. Synthesis of Escherichia coli O9a polysaccharide requires the participation of two domains of WbdA, a mannosyltransferase encoded within the wb* gene cluster. Mol Microbiol 1998; 27: 1213–1221. 192. Cuthbertson L, Kimber MS, Whitfield C. Substrate binding by a bacterial ABC transporter involved in polysaccharide export. Proc Natl Acad Sci USA 2007; 104: 19529–19534. 193. Cuthbertson L, Powers J, Whitfield C. The C-terminal domain of the nucleotide-binding domain protein Wzt determines substrate specificity in the ATP-binding cassette transporter for the lipopolysaccharide O-antigens in Escherichia coli serotypes O8 and O9a. J Biol Chem 2005; 280: 30310–30319. 194. Clarke BR, Cuthbertson L, Whitfield C. Nonreducing terminal modifications determine the chain length of polymannose O antigens of Escherichia coli and couple chain termination to polymer export via an ATP-binding cassette transporter. J Biol Chem 2004; 279: 35709–35718. 195. Scha¨ffer C, Wugeditsch T, Ka¨hlig H, Scheberl A, Zayni S, Messner P. The surface layer (S-layer) glycoprotein of Geobacillus stearothermophilus NRS 2004/3a. Analysis of its glycosylation. J Biol Chem 2002; 277: 6230–6239. 196. Steiner K, Novotny R, Werz DB et al. Molecular basis of S-layer glycoprotein glycan biosynthesis in Geobacillus stearothermophilus. J Biol Chem 2008; 283: 21120–21133. 197. King JD. Characterisation of polysaccharide biosynthesis genes in Bordetella bronchiseptica. PhD thesis. University of Cambridge, UK, 2006. 198. Preston A, Petersen BO, Duus JO et al. Complete structures of Bordetella bronchiseptica and Bordetella parapertussis lipopolysaccharides. J Biol Chem 2006; 281: 18135–18144. 199. Bera A, Herbert S, Jakob A, Vollmer W, Go¨tz F. Why are pathogenic Staphylococci so lysozyme resistant? The peptidoglycan O-acetyltransferase OatA is the major determinant for

200.

201. 202.

203.

204.

205.

206.

207.

208.

209.

210.

211.

212.

213.

214.

215. 216.

lysozyme resistance of Staphylococcus aureus. Mol Microbiol 2005; 55: 778–787. Currie HL, Lightfoot J, Lam JS. Prevalence of gca, a gene involved in synthesis of A-band common antigen polysaccharide in Pseudomonas aeruginosa. Clin Diagn Lab Immunol 1995; 2: 554–562. Benson DA, Karsch-Mizrachi I, Lipman DJ, Ostell J, Wheeler DL. GenBank. Nucleic Acids Res 2005; 33: D34–D38. Burrows LL, Charter DF, Lam JS. Molecular characterization of the Pseudomonas aeruginosa serotype O5 (PAO1) B-band lipopolysaccharide gene cluster. Mol Microbiol 1996; 22: 481–495. Dean CR, Franklund CV, Retief JD et al. Characterization of the serogroup O11 O-antigen locus of Pseudomonas aeruginosa PA103. J Bacteriol 1999; 181: 4275–4284. Dean CR, Goldberg JB. The wbpM gene in Pseudomonas aeruginosa serogroup O17 resides on a cryptic copy of the serogroup O11 O antigen gene locus. FEMS Microbiol Lett 2000; 187: 59–63. Burrows LL, Urbanic RV, Lam JS. Functional conservation of the polysaccharide biosynthetic protein WbpM and its homologues in Pseudomonas aeruginosa and other medically significant bacteria. Infect Immun 2000; 68: 931–936. DiGiandomenico A, Matewish MJ, Bisaillon A, Stehle JR, Lam JS, Castric P. Glycosylation of Pseudomonas aeruginosa 1244 pilin: glycan substrate specificity. Mol Microbiol 2002; 46: 519–530. Miller WL, Lam JS. Molecular biology of cell-surface polysaccharides in Pseudomonas aeruginosa: from gene to protein function. In: Cornelis P. (ed) Pseudomonas: Genomics and Molecular Biology. Horizon Scientific Press, 2007, 87–128. Creuzenet C, Urbanic RV, Lam JS. Structure-function studies of two novel UDP-GlcNAc C6 dehydratases/C4 reductases. Variation from the SYK dogma. J Biol Chem 2002; 277: 26769–26778. Creuzenet C, Lam JS. Topological and functional characterization of WbpM, an inner membrane UDP-GlcNAc C6 dehydratase essential for lipopolysaccharide biosynthesis in Pseudomonas aeruginosa. Mol Microbiol 2001; 41: 1295–1310. Schoenhofen IC, McNally DJ, Vinogradov E et al. Functional characterization of dehydratase/aminotransferase pairs from Helicobacter and Campylobacter: enzymes distinguishing the pseudaminic acid and bacillosamine biosynthetic pathways. J Biol Chem 2006; 281: 723–732. Schoenhofen IC, McNally DJ, Brisson JR, Logan SM. Elucidation of the CMP-pseudaminic acid pathway in Helicobacter pylori: synthesis from UDP-N-acetylglucosamine by a single enzymatic reaction. Glycobiology 2006; 16: 8C–14C. McNally DJ, Schoenhofen IC, Mulrooney EF et al. Identification of labile UDP-ketosugars in Helicobacter pylori, Campylobacter jejuni and Pseudomonas aeruginosa: key metabolites used to make glycan virulence factors. Chembiochem 2006; 7: 1865–1868. Ishiyama N, Creuzenet C, Miller WL et al. Structural studies of FlaA1 from Helicobacter pylori reveal the mechanism for inverting 4,6-dehydratase activity. J Biol Chem 2006; 281: 24489–24495. Morrison JP, Schoenhofen IC, Tanner ME. Mechanistic studies on PseB of pseudaminic acid biosynthesis: a UDP-N-acetylglucosamine 5-inverting 4,6-dehydratase. Bioorg Chem 2008; 36: 312–320. Holloway BW, Cooper GN. Lysogenic conversion in Pseudomonas aeruginosa. J Bacteriol 1962; 84: 1321–1324. Stover CK, Pham XQ, Erwin AL et al. Complete genome sequence of Pseudomonas aeruginosa PAO1, an opportunistic pathogen. Nature 2000; 406: 959–964.


Lipopolysaccharide biosynthesis in Pseudomonas aeruginosa 217. Knirel YA, Kochetkov NK. The structure of lipopolysaccharides of Gram-negative bacteria. III. The structure of O-antigens: a review. Biochemistry 1994; 59: 1325–1382. 218. Burrows LL, Pigeon KE, Lam JS. Pseudomonas aeruginosa Bband lipopolysaccharide genes wbpA and wbpI and their Escherichia coli homologues wecC and wecB are not functionally interchangeable. FEMS Microbiol Lett 2000; 189: 135–141. 219. Wenzel CQ, Daniels C, Keates RA, Brewer D, Lam JS. Evidence that WbpD is an N-acetyltransferase belonging to the hexapeptide acyltransferase superfamily and an important protein for Oantigen biosynthesis in Pseudomonas aeruginosa PAO1. Mol Microbiol 2005; 57: 1288–1303. 220. Westman EL, Preston A, Field RA, Lam JS. Biosynthesis of a rare di-N-acetylated sugar in the lipopolysaccharides of both Pseudomonas aeruginosa and Bordetella pertussis occurs via an identical scheme despite different gene clusters. J Bacteriol 2008; 190: 6060–6069. 221. Miller WL, Wenzel CQ, Daniels C, Larocque S, Brisson JR, Lam JS. Biochemical characterization of WbpA, a UDP-Nacetyl-D-glucosamine 6-dehydrogenase involved in O-antigen biosynthesis in Pseudomonas aeruginosa PAO1. J Biol Chem 2004; 279: 37551–37558. 222. Sweet CR, Ribeiro AA, Raetz CR. Oxidation and transamination of the 30 -position of UDP-N-acetylglucosamine by enzymes from Acidithiobacillus ferrooxidans. Role in the formation of lipid A molecules with four amide-linked acyl chains. J Biol Chem 2004; 279: 25400–25410. 223. Westman EL, McNally DJ, Charchoglyan A, Brewer D, Field RA, Lam JS. Characterization of WbpB, WbpE, and WbpD, and reconstitution of a pathway for the biosynthesis of UDP-2,3-diacetamido-2,3-dideoxy-D-mannuronic acid in Pseudomonas aeruginosa. J Biol Chem 2009; 284: 11854–11862. 224. Larkin A, Imperiali B. Biosynthesis of UDP-GlcNAc(3NAc)A by WbpB, WbpE, and WbpD: Enzymes in the Wbp Pathway Responsible for O-antigen Assembly in Pseudomonas aeruginosa PAO1. Biochemistry 2009: In press. 225. Westman EL, McNally DJ, Rejzek M et al. Identification and biochemical characterization of two novel UDP-2,3-diacetamido-2,3-dideoxy-alpha-D-glucuronic acid 2-epimerases from respiratory pathogens. Biochem J 2007; 405: 123–130. 226. King JD, Mulrooney EF, Vinogradov E, Kneidinger B, Mead K, Lam JS. lfnA from Pseudomonas aeruginosa O12 and wbuX from Escherichia coli O145 encode membrane-associated proteins and are required for expression of 2,6-dideoxy-2acetamidino-L-galactose in lipopolysaccharide O antigen. J Bacteriol 2008; 190: 1671–1679. 227. McNally DJ, Hui JP, Aubry AJ et al. Functional characterization of the flagellar glycosylation locus in Campylobacter jejuni 81– 176 using a focused metabolomics approach. J Biol Chem 2006; 281: 18489–18498. 228. Forsberg LS, Noel KD, Box J, Carlson RW. Genetic locus and structural characterization of the biochemical defect in the Oantigenic polysaccharide of the symbiotically deficient Rhizobium etli mutant, CE166. Replacement of N-acetylquinovosamine with its hexosyl-4-ulose precursor. J Biol Chem 2003; 278: 51347–51359. 229. Zhao X, Creuzenet C, Belanger M, Egbosimba E, Li J, Lam JS. WbpO, a UDP-N-acetyl-D-galactosamine dehydrogenase from Pseudomonas aeruginosa serotype O6. J Biol Chem 2000; 275: 33252–33259. 230. Miller WL, Matewish MJ, McNally DJ et al. Flagellin glycosylation in Pseudomonas aeruginosa PAK requires the O-antigen biosynthesis enzyme WbpO. J Biol Chem 2008; 283: 3507–3518.

311

231. Ishiyama N, Creuzenet C, Lam JS, Berghuis AM. Crystal structure of WbpP, a genuine UDP-N-acetylglucosamine 4epimerase from Pseudomonas aeruginosa: substrate specificity in UDP-hexose 4-epimerases. J Biol Chem 2004; 279: 22635–22642. 232. Creuzenet C, Be´langer M, Wakarchuk WW, Lam JS. Expression, purification, and biochemical characterization of WbpP, a new UDP-GlcNAc C4 epimerase from Pseudomonas aeruginosa serotype O6. J Biol Chem 2000; 275: 19060–19067. 233. Demendi M, Ishiyama N, Lam JS, Berghuis AM, Creuzenet C. Towards a better understanding of the substrate specificity of the UDP-N-acetylglucosamine C4 epimerase WbpP. Biochem J 2005; 389: 173–180. 234. Feng L, Tao J, Guo H et al. Structure of the Shigella dysenteriae 7 O antigen gene cluster and identification of its antigen specific genes. Microb Pathog 2004; 36: 109–115. 235. Vinogradov EV, Shashkov AS, Knirel YA et al. Structure of the O-antigen of Francisella tularensis strain 15. Carbohydr Res 1991; 214: 289–297. 236. Prior JL, Prior RG, Hitchen PG et al. Characterization of the O antigen gene cluster and structural analysis of the O antigen of Francisella tularensis subsp tularensis. J Med Microbiol 2003; 52: 845–851. 237. Parolis H, Parolis LA, Olivieri G. Structural studies on the Shigella-like Escherichia coli O121 O-specific polysaccharide. Carbohydr Res 1997; 303: 319–325. 238. Larsen TM, Boehlein SK, Schuster SM et al. Three-dimensional structure of Escherichia coli asparagine synthetase B: a short journey from substrate to product. Biochemistry 1999; 38: 16146–16157. 239. Kaya S, Araki Y, Ito E. The structure of the O-specific chain of lipopolysaccharide from Pseudomonas aeruginosa IID 1012 (ATCC 27588). J Biochem 1989; 105: 29–34. 240. Schuman B, Alfaro JA, Evans SV. Glycosyltransferase structure and function. In: Peters T. (ed). Bioactive Conformation I. Berlin: Springer, 2007; 217–257. 241. Knirel YA, Kochetkov NK. [Structure of lipopolysaccharides from Gram-negative bacteria. III. Structure of O-specific polysaccharides]. Biokhimiia 1994; 59: 1784–1851. 242. Moreau M, Richards JC, Fournier JM, Byrd RA, Karakawa WW, Vann WF. Structure of the type 5 capsular polysaccharide of Staphylococcus aureus. Carbohydr Res 1990; 201: 285–297. 243. Fournier JM, Vann WF, Karakawa WW. Purification and characterization of Staphylococcus aureus type 8 capsular polysaccharide. Infect Immun 1984; 45: 87–93. 244. Jones C, Currie F, Forster MJ. N.M.R. and conformational analysis of the capsular polysaccharide from Streptococcus pneumoniae type 4. Carbohydr Res 1991; 221: 95–121. 245. Schmidt MA, Jann B, Jann K. Cell-wall lipopolysaccharide of the urinary-tract-infective Escherichia coli O4:K12:H–. Structure of the polysaccharide chain. Eur J Biochem 1983; 137: 163–171. 246. Manca MC, Weintraub A, Widmalm G. Structural studies of the Escherichia coli O26 O-antigen polysaccharide. Carbohydr Res 1996; 281: 155–160. 247. Landersjo¨ C, Weintraub A, Widmalm G. Structural analysis of the O-antigen polysaccharide from the Shiga toxin-producing Escherichia coli O172. Eur J Biochem 2001; 268: 2239–2245. 248. Kneidinger B, O’Riordan K, Li J, Brisson JR, Lee JC, Lam JS. Three highly conserved proteins catalyze the conversion of UDP-N-acetyl-D-glucosamine to precursors for the biosynthesis of O antigen in Pseudomonas aeruginosa O11 and capsule in Staphylococcus aureus type 5. Implications for the UDP-Nacetyl-L-fucosamine biosynthetic pathway. J Biol Chem 2003; 278: 3615–3627.


312


249. Mulrooney EF, Poon KK, McNally DJ, Brisson JR, Lam JS. Biosynthesis of UDP-N-acetyl-L-fucosamine, a precursor to the biosynthesis of lipopolysaccharide in Pseudomonas aeruginosa serotype O11. J Biol Chem 2005; 280: 19535–19542. 250. Guerry P, Doig P, Alm RA, Burr DH, Kinsella N, Trust TJ. Identification and characterization of genes required for posttranslational modification of Campylobacter coli VC167 flagellin. Mol Microbiol 1996; 19: 369–378. 251. Glaze PA, Watson DC, Young NM, Tanner ME. Biosynthesis of CMP-N,N’-diacetyllegionaminic acid from UDP-N,N’-diacetylbacillosamine in Legionella pneumophila. Biochemistry 2008; 47: 3272–3282. 252. Vann WF, Daines DA, Murkin AS et al. The NeuC protein of Escherichia coli K1 is a UDP N-acetylglucosamine 2-epimerase. J Bacteriol 2004; 186: 706–712. 253. Marchler-Bauer A, Anderson JB, Derbyshire MK et al. CDD: a conserved domain database for interactive domain family analysis. Nucleic Acids Res 2007; 35: D237–D240. 254. Franklin MJ, Chitnis CE, Gacesa P, Sonesson A, White DC, Ohman DE. Pseudomonas aeruginosa AlgG is a polymer level alginate C5-mannuronan epimerase. J Bacteriol 1994; 176: 1821–1830. 255. King JD, Harmer NJ, Preston A et al. Predicting protein function from structure - the roles of short-chain dehydrogenase/reductase enzymes in Bordetella O-antigen biosynthesis. J Mol Biol 2007; 374: 749–763. 256. Faure K, Shimabukuro D, Ajayi T, Allmond LR, Sawa T, Wiener-Kronish JP. O-antigen serotypes and type III secretory toxins in clinical isolates of Pseudomonas aeruginosa. J Clin Microbiol 2003; 41: 2158–2160. 257. Persson B, Kallberg Y, Oppermann U, Jornvall H. Coenzymebased functional assignments of short-chain dehydrogenases/ reductases (SDRs). Chem Biol Interact 2003; 143/144: 271–278. 258. Libisch B, Balogh B, Fuzi M. Identification of two multidrugresistant Pseudomonas aeruginosa clonal lineages with a countrywide distribution in Hungary. Curr Microbiol 2009; 58: 111–116. 259. Kaluzny K, Abeyrathne PD, Lam JS. Coexistence of two distinct versions of O-antigen polymerase, Wzy-alpha and Wzy-beta, in Pseudomonas aeruginosa serogroup O2 and their contributions to cell surface diversity. J Bacteriol 2007; 189: 4141–4152. 260. Liu D, Cole RA, Reeves PR. An O-antigen processing function for Wzx (RfbX): a promising candidate for O-unit flippase. J Bacteriol 1996; 178: 2102–2107.

261. Burrows LL, Lam JS. Effect of wzx (rfbX) mutations on A-band and B-band lipopolysaccharide biosynthesis in Pseudomonas aeruginosa O5. J Bacteriol 1999; 181: 973–980. 262. Marolda CL, Vicarioli J, Valvano MA. Wzx proteins involved in biosynthesis of O antigen function in association with the first sugar of the O-specific lipopolysaccharide subunit. Microbiology 2004; 150: 4095–4105. 263. Feldman MF, Marolda CL, Monteiro MA, Perry MB, Parodi AJ, Valvano MA. The activity of a putative polyisoprenol-linked sugar translocase (Wzx) involved in Escherichia coli O antigen assembly is independent of the chemical structure of the O repeat. J Biol Chem 1999; 274: 35129–35138. 264. Daniels C, Vindurampulle C, Morona R. Overexpression and topology of the Shigella flexneri O-antigen polymerase (Rfc/ Wzy). Mol Microbiol 1998; 28: 1211–1222. 265. Jiang SM, Wang L, Reeves PR. Molecular characterization of Streptococcus pneumoniae type 4, 6B, 8, and 18C capsular polysaccharide gene clusters. Infect Immun 2001; 69: 1244–1255. 266. Newton GJ, Daniels C, Burrows LL, Kropinski AM, Clarke AJ, Lam JS. Three-component-mediated serotype conversion in Pseudomonas aeruginosa by bacteriophage D3. Mol Microbiol 2001; 39: 1237–1247. 267. Burrows LL, Chow D, Lam JS. Pseudomonas aeruginosa B-band O-antigen chain length is modulated by Wzz (Rol). J Bacteriol 1997; 179: 1482–1489. 268. Daniels C, Griffiths C, Cowles B, Lam JS. Pseudomonas aeruginosa O-antigen chain length is determined before ligation to lipid A core. Environ Microbiol 2002; 4: 883–897. 269. Kintz E, Scarff JM, DiGiandomenico A, Goldberg JB. Lipopolysaccharide O-antigen chain length regulation in Pseudomonas aeruginosa serogroup O11 strain PA103. J Bacteriol 2008; 190: 2709–2716. 270. Abeyrathne PD, Lam JS. WaaL of Pseudomonas aeruginosa utilizes ATP in in vitro ligation of O antigen onto lipid A-core. Mol Microbiol 2007; 65: 1345–1359. 271. Pe´rez JM, McGarry MA, Marolda CL, Valvano MA. Functional analysis of the large periplasmic loop of the Escherichia coli K-12 WaaL O-antigen ligase. Mol Microbiol 2008; 70: 1424–1440. 272. Heinrichs DE, Monteiro MA, Perry MB, Whitfield C. 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 the R2 WaaK and WaaL homologs. J Biol Chem 1998; 273: 8849–8859.
