lpt6, a Gene Required for Addition of Phosphoethanolamine to Inner-Core Lipopolysaccharide of Neisseria meningitidis and Haemophilus influenzae

Abstract

We previously described a gene, lpt3, required for the addition of phosphoethanolamine (PEtn) at the 3 position on the β-chain heptose (HepII) of the inner-core Neisseria meningitidis lipopolysaccharide (LPS), but it has long been recognized that the inner-core LPS of some strains possesses PEtn at the 6 position (PEtn-6) on HepII. We have now identified a gene, lpt6 (NMA0408), that is required for the addition of PEtn-6 on HepII. The lpt6 gene is located in a region previously identified as Lgt-3 and is associated with other LPS biosynthetic genes. We screened 113 strains, representing all serogroups and including disease and carriage strains, for the lpt3 and lpt6 genes and showed that 36% contained both genes, while 50% possessed lpt3 only and 12% possessed lpt6 only. The translated amino acid sequence of lpt6 has a homologue (72.5% similarity) in a product of the Haemophilus influenzae Rd genome sequence. Previous structural studies have shown that all H. influenzae strains investigated have PEtn-6 on HepII. Consistent with this, we found that, among 70 strains representing all capsular serotypes and nonencapsulated H. influenzae strains, the lpt6 homologue was invariably present. Structural analysis of LPS from H. influenzae and N. meningitidis strains where lpt6 had been insertionally inactivated revealed that PEtn-6 on HepII could not be detected. The translated amino acid sequences from the N. meningitidis and H. influenzae lpt6 genes have conserved residues across their lengths and are part of a family of proven or putative PEtn transferases present in a wide range of gram-negative bacteria.

The lipopolysaccharides (LPS) of the gram-negative bacteria Neisseria meningitidis and Haemophilus influenzae lack the repeating O antigens that are characteristic of the LPS of enteric bacteria. The core LPS molecule of these bacteria is heterogeneous both within and among strains and plays a role in the commensal and pathogenic behavior of each species. N. meningitidis and H. influenzae are host-restricted bacteria found exclusively in humans. They normally reside in the nasopharynx of humans but can spread contiguously within the respiratory tract or disseminate systemically into the blood to cause serious diseases such as pneumonia, septicemia, and meningitis.

N. meningitidis LPS has been classified into 12 distinct immunotypes (L1 to L12), on the basis of monoclonal antibody (MAb) reactivity (42), the basis of which has been further defined by structural analysis as detailed in the following references: L1 and L6, references 11 and 49; L2, reference 13; L3, reference 36; L4 and L7, reference 29; L5, reference 35; L9, reference 22. N. meningitidis LPS is based on a diheptose backbone, which is attached via one of two 3-deoxy-d-manno-2-octulosonic acid (Kdo) residues to the lipid A portion. Additions occur to the first heptose (HepI), and extension past the proximal glucose (Glc) residue is classed as the outer-core structure (Fig. 1). The second heptose (HepII) is invariably substituted by an N-acetylglucosamine (GlcNAc) residue. Additions of Glc and phosphoethanolamine (PEtn) to HepII within the inner-core region also vary among immunotypes. Although the LPS of H. influenzae has many similarities with that of N. meningitidis, notable differences include a conserved triheptose inner-core structure linked via a single Kdo residue to the lipid A (Fig. 1). Oligosaccharide chains of the outer core can extend from each heptose, with the pattern and degree of substitution with noncarbohydrate components varying among strains (32). The lipid A of both bacterial species contributes significantly to the pathogenesis of disease caused by these organisms.

FIG. 1.

Schematic representation of the LPS structures in N. meningitidis and H. influenzae. (A) N. meningitidis structure. The immunotypes L2, L3, and L4 are differentiated on the basis of the specific additions at the 3 and 6 positions of HepII. (B) H. influenzae structure of strain Rd. Abbreviations: Gal, galactose; Hep, heptose; GalNAc, N-acetyl galactosamine; P, phosphate; PC, phosphorylcholine; Glc, glucose; Kdo, 3-deoxy-d-manno-2-octulosonic acid; GlcNAc, N-acetylglucosamine; PETn, phosphoethanolamine. The products of the genes responsible for these additions are indicated, with alternative names for rfaK and rfaF given in parentheses. The product of the gene galE is an enzyme involved in converting activated Glc to Gal, and mutation of this gene prevents any addition to the Hep backbone past the first Glc, i.e., just the inner-core LPS structure is produced. *, lpt6 gene identified in this study (boldface).

A number of sugars and their linkages in the LPS of N. meningitidis are similar to those in H. influenzae LPS, and, in both species, the majority of genes responsible for the biosynthesis of the LPS are known. Many genes required to add the same sugars in the same linkage of the respective LPS are homologous, examples being the genes for the α-galactosyltransferase LgtC and the β-glucosyltransferase LgtF, found in both N. meningitidis (15, 28, 47) and H. influenzae (18, 20). A notable characteristic of the LPS of both N. meningitidis and H. influenzae is phase variation, the high-frequency, reversible loss and gain of expression of outer-core structures mediated by slippage of DNA repeat tracts present in LPS biosynthetic genes (19, 24). These DNA repeat tracts are absent from the inner-core biosynthetic genes in both species, so that these structures are relatively conserved and are candidates for inclusion in vaccines for the prevention of disease caused by these bacteria.

The presence and location of noncarbohydrate substituents of LPS, such as PEtn, are critical for the immune recognition of N. meningitidis LPS by antibodies. Our previous work on LPS structure and immune responses led to the identification of a gene (lpt3) responsible for the addition of PEtn to the 3 position of HepII (PEtn-3) (31). MAb L3B5 was used to detect PEtn-3, for which binding of the MAb has an absolute requirement, in 70% of epidemiologically diverse N. meningitidis strains (38). Loss of PEtn-3 from N. meningitidis LPS resulted in relative resistance to bactericidal killing and opsonophagocytosis by MAb L3B5 in vitro (31). Structural analysis of LPS has shown that alternative inner-core LPS structures exist, for example, where PEtn is added via the 6 position to HepII (PEtn-6) either concurrently with PEtn-3 or alone (8). For H. influenzae LPS, PEtn-6 attached to HepII has been found in all strains studied to date (32).

In a previous study in our laboratory to investigate the genetic basis of alternative linkages of PEtn, a search of the N. meningitidis serogroup B genome sequence (46) with the Lpt3 sequence identified the homologous gene NMB1638, or lptA. LptA is a transferase for addition of PEtn to the lipid A moiety (9). Such homology searches, however, identified no gene responsible for the addition of PEtn to the 6 position of HepII.

In the present study we utilized a MAb (L2-16) that specifically recognizes PEtn at the 6 position in N. meningitidis LPS (14) to identify the gene required for the addition of PEtn-6. We have found that this gene, designated lpt6, has significant homology to a gene in H. influenzae that carries out the related function in this bacterium.

MATERIALS AND METHODS

Strains and culture conditions.

The N. meningitidis and H. influenzae strains used in this study are listed in Table 1. All strains were grown overnight at 37°C on brain heart infusion (BHI) medium (Oxoid), supplemented with Levinthal's reagent (10%, vol/vol) and solidified with agar (1% [wt/vol]; Bioconnections), in an atmosphere of 5% CO₂. For selection of strains following transformation, kanamycin (75 μg ml⁻¹) or erythromycin (6 μg ml⁻¹) was added to the culture medium for N. meningitidis and kanamycin (10 μg ml⁻¹) was added for H. influenzae. Escherichia coli strain DH5α was used to propagate recombinant DNA constructs and was grown at 37°C on Luria-Bertani medium supplemented with kanamycin (50 μg ml⁻¹), erythromycin (300 μg ml⁻¹), or ampicillin (50 μg ml⁻¹) when appropriate.

TABLE 1.

List of N. meningitidis and H. influenzae strains and plasmids used in this study

Strain or plasmid	Description	Source or reference
N. meningitidis
35E	L2 immunotyping reference strain
35EgalE	galE gene disrupted, truncated LPS	This study
35Elpt6	lpt6 gene disrupted	This study
35EgalElpt6	galE and lpt6 genes disrupted, truncated LPS	This study
89I	L4 immunotyping reference strain
89IgalE	galE gene disrupted, truncated LPS	This study
89Ilpt6	lpt6 gene disrupted	This study
89IgalElpt6	galE and lpt6 genes disrupted, truncated LPS	This study
126E	L1 immunotyping reference strain
MC58	L3 immunotyping reference strain	34
M981	L5 immunotyping reference strain
M992	L6 immunotyping reference strain
6155	L7 immunotyping reference strain
M978	L8 immunotyping reference strain
120M	L9 immunotyping reference strain
7880	L10 immunotyping reference strain
7889	L11 immunotyping reference strain
3200	L12 immunotyping reference strain
2996	L3 immunotype	37
8047	L3 immunotype	37
M986	L3 immunotype	37
BZ232	L2 immunotype
C751 (Z2491)	Serogroup A genome sequence strain	10
H. influenzae
Rd	Genome sequence strain
Rdlpt6	lpt6 gene disrupted	This study
375	Nontypeable H. influenzae strain	6
375lpt6	lpt6 gene disrupted	This study
285	Nontypeable H. influenzae strain	6
285lpt6	lpt6 gene disrupted	This study
Plasmids
pT7Blue	Cloning vector	Novagen
pER2	Source of erythromycin resistance cassette	25
pUC4-kan	Source of kanamycin resistance cassette
pT7-galE-ery	Disrupted galE gene from N. meningitidis, cloned into pT7Blue	This study
pFA1	Disrupted lpt6 gene from N. meningitidis, cloned into pT7Blue	This study
pFB5, pFB6	Disrupted lpt6 gene from H. influenzae, cloned into pT7Blue	This study

DNA techniques.

Recombinant DNA techniques were performed as described by Sambrook et al. (41). Restriction endonuclease and DNA-modifying enzymes were obtained from Boehringer Mannheim or New England Biolabs and used according to the manufacturer's instructions. Oligonucleotide primers were synthesized by Sigma-Genosys. Standard PCR amplifications were performed in 50-μl reaction volumes (final concentrations: 20 mM Tris-HCl [pH 8.4], 50 mM KCl, 2.5 mM MgCl₂, 0.4 μM forward primer, 0.4 μM reverse primer, 0.4 mM deoxynucleoside triphosphates) with 1.25 U of Taq recombinant polymerase (Invitrogen) in a Master-Cycler (Eppendorf) gradient thermal cycler. Thirty cycles of PCR were performed, each consisting of 1 min of denaturation at 94°C, 1 min of annealing at typically 5°C below the melting temperature, and 1 min of extension at 72°C, with a final prolonged extension of 10 min at 72°C. Chromosomal DNA was prepared from H. influenzae and N. meningitidis strains as described previously (1). DNA sequences were determined with the Big Dye terminator, version 3.1, cycle sequencing kit (Applied Biosystems). Sequencing reactions were performed in a Master-Cycler gradient thermal cycler with 25 cycles of amplification, each consisting of 30 s at 95°C, 15 s at 50°C, and 4 min at 60°C. Samples were run on an ABI Prism 377 DNA sequencer. A list of oligonucleotide primers used in this study is given in Table 2.

TABLE 2.

List of oligonucleotide primers used in this study

Primer name	Sequence (5′ → 3′)	Gene in which primer binds	Binding position within gene (bp)a
galE-f	TCA GCC CCC ATA CGA GAT AG	galE (NMB0064)	429-469
galE-r	CAG TCA AAG ATT CCA AAT TGC C	galE (NMB0064)	2271-2249
275-a	TTC ACG ATT GCG GCA GTG	lpt6 (HI0275)	28-45
275-b	GCA TTC AGC TTA CGG TGC	HI0276	585-602
arb-4	GGC CAC GCG TCG ACT AGT AC N10 GAC G	Random	—
arb-6	GGC CAC GCG TCG ACT AGT AC N10 ACG C	Random	—
arb-2	GGC CAC GCG TCG ACT AGT AC	Random	—
ext-lt	GCT TCA GGG TTG AGA TGT G	Transposon RC325	31-13
ext-rt	GAT GGT TGA GAT GTG TAT AAG	Transposon RC325	1193-1213
int2-lt	AGC ATT ACG CTG ACT TGA CG	Transposon RC325	233-215
int-rt	GTT GGA ATT TAA TCG CGG C	Transposon RC325	1029-1047
lgtG-F	GAA CTC GTC CCC GTA TTT	mtr (NMA0409)	1216-1234
lgtG-H	GAT GTC TTC AAT CAT ACC	gmhX (NMA0405)	330-312
408-A	TTG TAA CGG CGG CGT TGG	lpt6 (NMA0408)	29-47
408-B	TTT CCG ACC AAC TCC CAG	lpt6 (NMA0408)	1610-1592
408-C	TTG TTT GCG GCG GCG TTC	lpt6 (NMA0408)	403-420
408-D	TCC TGC CGG CAA ACT GCG	lpt6 (NMA0408)	1429-1412
408-E	TCC AAG TAC CCG TGT TCG	lpt6 (NMA0408)	538-556
408-F	ACT TTC CAC CGC TTT GTC	lpt6 (NMA0408)	1287-1269
408-Cr	GAA CGC CGC CGC AAA CAA	lpt6 (NMA0408)	420-403
408-Dr	CGC AGT TTG CCG TCA GGA	lpt6 (NMA0408)	1412-1429
408-Er	CGA ACA CGG GTA CTT GGA	lpt6 (NMA0408)	556-538
408-Fr	GGA CAA AGC GGT GGA AAG	lpt6 (NMA0408)	1269-1287
408-G	TGA ACG AGT TTT CAG ACG GC	lpt6 (NMA0408)	1706-1686
408-H	AGT GAT GCG CCT GTG CTG C	lpt6 (NMA0408)	−87-−68
kan-if	AGC CAT ATT CAA CGG GAA AC	kan (pUC4-kan)	139-159
kan-ir	TTT GCT TTG CCA CGG AAC	kan (pUC4-kan)	538-520

^aThe binding position is given with respect to the first base of the relevant reading frame. —, not applicable.

Antibodies.

The MAb L2-16, raised against N. meningitidis inner-core LPS structures was prepared from ascites fluid as described previously (14). The MAbs L6A9, E12, D2, C1, B2, 4C4, 6E4, 9D8, 7E7, MAHI 2, and MAHI 3 were raised against H. influenzae LPS structures (M. A. J. Gidney, unpublished data; 3, 5).

Construction of a library of N. meningitidis transposon mutants and colony immunoblotting.

The transposon RC325, consisting of a kanamycin resistance cassette flanked by inverted repeats, was used for in vitro transposition to create a random-mutant library in chromosomal DNA (45) isolated from N. meningitidis strain 89I. This chromosomal DNA was then used to transform N. meningitidis strain 89IgalE. A 1-μl loopful of bacteria grown overnight on BHI agar plates was resuspended in 100 μl of phosphate-buffered saline (PBS). On a fresh BHI agar plate, 5 μl of PBS was added, followed by 10 μl of the bacterial suspension and 6 μl of DNA from the random-mutant library, and the constituents were briefly mixed with a loop. The plate was incubated for 5 to 6 h, the growth was then scraped from the plate surface and resuspended in 500 μl of PBS, and aliquots of 100 μl were plated onto BHI agar containing kanamycin. Following overnight incubation, the colonies were transferred to nitrocellulose filters (Schleicher & Schuell) and air dried. These filters were probed with MAb L2-16, as described previously (33), nonreactive colonies were selected and rescreened with MAb L2-16, and those confirmed as showing no reactivity were frozen and selected for further study.

Arbitrary PCR.

The DNA sequence flanking the transposon insertion sites was determined by arbitrary PCR in a method adapted from that of Caetano-Annoles (4). A nested-PCR amplification was performed with a primer specific to the ends of the transposon in combination with a primer of random sequence, which can anneal to chromosomal sequences flanking the transposon (Table 2). In the first round of PCR a primer unique to one end of the transposon (int-lt or int-rt) and an arbitrary primer (either arb-4 or arb-6) were used in a 50-μl PCR mixture (containing [final concentrations] 20 mM Tris-HCl [pH 8.4], 50 mM KCl, 1.5 mM MgCl₂, 0.4 μM transposon-specific primer, 0.8 μM random primer) with 1.25 U of Taq recombinant polymerase. The first-round reaction conditions were as follows: (i) 5 min at 94°C; (ii) 6 cycles consisting of 1 min at 94°C, 1 min at 30°C, and 1 min at 72°C; (iii) 30 cycles consisting of 1 min at 94°C, 1 min at 50°C, and 1 min at 72°C; and (iv) 10 min at 72°C. The second-round PCR used 1 μl of the first-round product as the template DNA and used a transposon-specific primer (either ext-lt or ext-rt) and a primer (arb-2) designed to bind to either of the arbitrary primers used in the first round. The reactions for the second round were performed under the standard PCR conditions described above except that both primers were used at a concentration of 0.4 μM and an annealing temperature of 50°C was used. The products from the second round of PCR were separated on a 3% Metaphor agarose gel (FMC Bioproducts). Products that were unique to the transposon-inserted DNA, compared to the products from DNA from the parental strain, were excised and purified with the Qiaex gel extraction kit (QIAGEN) as instructed by the manufacturer. To increase the amount of DNA present, the extracted DNA was then used as template in a repeat of the second-round PCR amplification. The DNA sequence was obtained with the transposon-specific primer ext-lt or ext-rt, as appropriate.

Southern blotting and hybridization.

Bacterial genomic DNA digested with either ClaI, HindIII, or EcoRI was separated on 1% agarose gels and then transferred to a Hybond-N nylon membrane (Amersham International), as described by Sambrook et al. (41). The kanamycin resistance cassette was prepared as a hybridization probe after amplifying the DNA with primers kan-if and kan-ir, with pUC4-kan plasmid DNA as the template. PCR-amplified products were labeled with digoxigenin (DIG) using the DIG luminescence detection kit for nucleic acids (Boehringer Mannheim). Hybridization was performed at 42°C for 16 h under conditions described for DIG (Boehringer Mannheim). The blots were then washed as follows: twice for 15 min at room temperature with 2× SSC (300 mM NaCl, 30 mM sodium citrate)-0.1% sodium dodecyl sulfate (SDS) and then twice for 15 min at 65°C with 0.1× SSC (15 mM NaCl, 1.5 mM sodium citrate)-0.1% SDS. The blots were developed with the DIG detection protocol and exposed to autoradiograph film according to the manufacturer's instructions.

T-SDS-PAGE.

Whole-cell lysates were prepared from H. influenzae and N. meningitidis strains grown overnight, by harvesting and resuspending cells in PBS and then adding the equivalent amount of dissociation buffer and heating at 100°C for 5 min. Samples were separated by Tricine-SDS-polyacrylamide gel electrophoresis (T-SDS-PAGE) using 16.5% gels run at 30 mA and 4°C for 18 h (30) and were then visualized by staining with silver according to the manufacturer's instructions (Amersham Biosciences).

Disruption of genes and construction of mutant strains.

To mutate the N. meningitidis galE gene, the gene was first amplified by PCR, at an annealing temperature of 55°C, from strain MC58 chromosomal DNA with oligonucleotide primers galE-f and galE-r and cloned into plasmid pT7Blue (Novagen). An erythromycin resistance cassette was excised from pER2 (25) by digestion with BssHII and inserted into the BssHII site within the cloned galE gene. The resulting construct, pT7-galE-ery, was used to transform N. meningitidis strains 89I and 35E as described above, and transformants were selected on erythromycin.

To make N. meningitidis lpt6 mutants, NMA0408 was first amplified by PCR from N. meningitidis strain C751 chromosomal DNA with oligonucleotide primers 408-A and 408-B and cloned into pT7Blue. A kanamycin resistance (Kan^r) cassette was excised from pUC4-kan by digestion with HincII and inserted into the EagI site within the cloned NMA0408 (N. meningitidis lpt6) gene. The resulting construct, pFA1, was used to transform N. meningitidis strains 89I, 89IgalE, 35E, and 35EgalE.

To make H. influenzae lpt6 mutants, HI0275 was amplified by PCR from H. influenzae strain Rd chromosomal DNA with oligonucleotide primers 275-a and 275-b and cloned into pT7Blue. A Kan^r cassette, excised from pUC4-kan by digestion with BamHI, was inserted to replace a 490-bp BglII fragment within the cloned HI0275 (H. influenzae lpt6) gene. The resulting constructs, pFB5 (Kan^r cassette in opposite orientation to HI0275) and pFB6 (Kan^r cassette in same orientation as HI0275), were used to transform strain Rd and NTHi isolates 285 and 375 by the method of Herriott et al. (17).

Analysis of the Lgt-3 region.

The Lgt-3 region from N. meningitidis strains was amplified by PCR using Bio-X-act Long polymerase (Bioline) with oligonucleotide primers lgtG-F and lgtG-H according to the manufacturer's instructions in a Master-Cycler gradient thermal cycler. The DNA sequence of lpt6 was obtained with the primers 408-C, 408-Cr, 408-D, 408-Dr, 408-Er, 408-Fr, 408-G, and 408-H.

Resistance to the bactericidal effects of human sera.

Bacteria cultured on BHI plates were assayed in pooled human serum as described previously (21).

LPS structural analysis.

LPS samples were obtained from N. meningitidis and H. influenzae strains grown overnight as described previously (38). Typically, N. meningitidis bacteria were scraped from 40 BHI plates and resuspended in 30 ml of 0.05% phenol in PBS, while H. influenzae bacteria were harvested from 5 liters of BHI broth by centrifugation at 4,000 × g and then resuspended in 30 ml of 0.05% phenol in PBS. LPS was extracted by the hot phenol-water method of Westphal and Jann (50). Briefly, contaminating RNA was removed from the crude LPS following the first phenol extraction by incubation with 1,000 U of RNase (Promega) for 1 h at 37°C. A second phenol extraction was performed before precipitation of the LPS overnight with 0.5 M NaCl in 3 volumes of ethanol. The LPS was recovered by centrifugation at 4,000 × g at 4°C for 25 min, washed with 70% ethanol, and then freeze-dried for 6 h. O-deacylated LPS was prepared as described previously (37). Capillary electrophoresis (CE)-mass spectrometry (MS), tandem MS (CE-MS/MS), and nuclear magnetic resonance analyses of O-deacylated LPS samples were carried out as described previously (8).

RESULTS

Use of in vitro transposon mutagenesis to identify genes required for expression of the PEtn-6 epitope.

To investigate the genetic basis for the addition of PEtn-6 to HepII of N. meningitidis LPS, we screened a library of transposon mutants in a galE mutant background of strain 89I (immunotype L4) for reactivity with MAb L2-16 (14). Because MAb L2-16 reactivity is dependent on expression of the PEtn-6 epitope, the lack of MAb L2-16 reactivity would identify colonies in which transposon insertions caused disruption of PEtn-6 incorporation into the LPS. We used the strain 89I galE mutant because its reactivity with MAb L2-16 was stronger than that of wild-type strain 89I, thus facilitating the visual screening of nonreacting colonies. Eighteen colonies nonreactive with MAb L2-16 were identified from approximately 20,000 N. meningitidis colonies of the library of transposon insertional mutants. These colonies were picked and regrown to confirm their lack of reactivity with MAb L2-16 (Fig. 2).

FIG. 2.

(A and B) Screening of strain 89IgalE transformed with the random transposon library and then colony immunoblotted with MAb L2-16. Shown are colonies from independent transformations. Arrows, examples of nonreacting colonies. (C) Southern blot analysis of DNA isolated from representative transformants that were nonreactive with MAb L2-16 following transformation. Chromosomal DNA was digested with EcoRI, and a PCR-amplified fragment of the kanamycin resistance cassette was used as a probe. Lanes 1 to 9, single insertion of the transposon in the DNA from individual nonreacting colonies; lane 10, parental DNA from strain 89IgalE.

A lack of MAb L2-16 reactivity to transformants with transposon insertions could be caused by functional loss of one or more genes required for the addition of PEtn to the 6 position of HepII or interruption of any of the genes required for the synthesis of the PEtn-6-containing LPS epitope. Thus, chromosomal DNA isolated from the MAb L2-16-nonreactive mutants was used in Southern blot analyses. We observed single hybridizing bands, indicating that, for each mutant, a single transposon had inserted within the chromosome and that a range of mutants with independent transposon insertions had been obtained (Fig. 2C).

DNA adjacent to the different transposon insertion sites of each of the 18 mutants was amplified by the arbitrary-PCR method (4), and the sequence was determined. The DNA sequence was then used to search the completed bacterial genome sequences (open reading frames and intergenic regions) available at The Institute for Genome Research (http://tigrblast.tigr.org/cmr.blast/) using the BLASTN algorithm. For several of the nonreacting mutant strains, PCR amplification resulted in several products and sequences, probably the result of amplification with the arbitrary primers alone.

Sites of transposon insertion correspond to known LPS genes.

Of the 18 mutant strains that were nonreactive with MAb L2-16, 9 had transposon insertions in genes known to be related to LPS biosynthesis: rfaE or hldE (putative ADP-heptose synthase gene), rfaD or hldD (ADP-l-glycero-d-manno-heptose-6-epimerase gene), rfaF or lsi (ADP-heptose-LPS heptosyltransferase II gene), and rfaK or icsA (α-1,2-N-acetylglucosamine transferase gene). RfaD and RfaE are involved in the heptose biosynthetic pathway (12, 47). RfaF is the transferase required for the addition of HepII to HepI (23), and RfaK is the transferase that adds a GlcNAc to HepII (27, 47). Inactivation of each of the rfaD, rfaE, and rfaF genes would be expected to prevent formation of the diheptose backbone of the LPS inner core, to which PEtn is attached. Inactivation of the rfaK gene would prevent MAb L2-16 from reacting, as the GlcNAc is included as part of the target epitope. The gel profiles of LPS extracted from these insertion mutants were each consistent with the expected phenotype that would result from disruption of the corresponding LPS genes (Fig. 3). Thus, in comparison to LPS from the parental strain, that from each of the rfaD and rfaE mutant strains was highly truncated. The rfaF mutants had higher-molecular-weight LPS, corresponding to the addition of a single heptose, and LPS from the rfaK and galU mutants were of lower molecular weight than that from the corresponding galE mutant, indicating incomplete inner-core synthesis and no further oligosaccharide extension.

FIG. 3.

Profiles of LPS isolated from transformants of strain 89IgalE that were nonreactive with MAb L2-16 following T-SDS-PAGE and silver staining. Lanes 1 and 12, strain 89I; 11 and 22, 89IgalE; lanes 2 to 10 and 13 to 21, transformants with transposon insertion sites identified in the indicated genes. Lanes: 2, 4.1, rfaE (ADP-heptose synthase, putative); 3, 7.1, rfaE; 4, 8.1, NMA0408 (putative integral membrane protein); 5, 9.1, NMA0408; 6, 9.2, rfaD (ADP-l-glycero-d-mannoheptose-6-epimerase); 7, 9.3, galU (UTP-glucose-1-phosphate uridyltransferase]; 8, 10.1, NMA0408; 9, 10.2, rfaK (alpha-1,2-N-acetylglucosamine transferase]; 10, 10.3, rfaD; 13, 10.4, mtr (tryptophan transporter); 14, 10.5, rfaD; 15, 10.6, mtr; 16, 10.7, rfaD; 17, 10.8, rfaF (ADP-heptose-LPS heptosyltransferase II); 18, 10.9, lgtG (lipopolysaccharide glycosyl transferase); 19, 10.10, NMA0408; 20, 10.11, mtr; 21, 10.12, rfaF. The insertion of the transposon in the rfaD gene results in a “leaky” phenotype with some LPS bands appearing to have the same banding profile as the parental 89IgalE strain. As the LPS decreases in size to the highly truncated LPS structures, the staining becomes noticeably weaker.

Based on these findings, we conclude that the screening of our library of MAb L2-16 insertional mutants identified several relevant genes, thus validating the experimental approach. No nonreacting colonies were identified with the transposon inserted in the glucosyltransferase gene lgtF/icsB, required to add glucose to HepI (28, 47). However, galU, encoding a UTP-glucose-1-phosphate uridyltransferase, required to provide activated glucose for LPS synthesis (19), produced a MAb L2-16-nonreactive phenotype when mutated, and thus one might assume that the glucose residue attached to HepI is part of the epitope that reacts with MAb L2-16. The function of the galU gene in N. meningitidis, with respect to LPS biosynthesis, had not been previously investigated.

Some of the sites of transposon insertion correspond to genes of the Lgt-3 region, including a candidate gene for the addition of the PEtn-6 epitope.

The remaining eight transposon insertion sites, resulting in mutants nonreactive with MAb L2-16, were found to be clustered in a region of the genome previously designated Lgt-3, one of three N. meningitidis chromosomal regions containing LPS-related genes that had been investigated by Zhu and colleagues (51). In the present study, transposon insertions resulting in nonreactivity with MAb L2-16 occurred independently in four NMA0408 mutants, three mtr mutants, and one lgtG mutant.

In the two completed and annotated N. meningitidis genome sequences, those of strains MC58 (serogroup B) and Z2491[C751](serogroup A), Lgt-3 is flanked by mtr (NMB2031/NMA0409) and NMB2033/NMA0405, designated tryptophan transporter and amino acid transporter genes, respectively (Fig. 4). Shih and coworkers (43) identified NMA0405/NMB2033 as the gene gmhX. GmhX is a novel enzyme required for the incorporation of l-glycero-d-manno-heptose into LPS and is proposed to be a phosphatase in the heptose biosynthesis pathway. Zhu et al. (51) showed that, depending on the N. meningitidis strain, none, one, or two of the genes lgtG, NMA0408, and NMA0407 may be present between the mtr and NMB2033/NMA0405 genes. PCR analysis of the Lgt-3 region in strain 89I showed that the region comprised only lgtG and NMA0408, conforming to one of the described Lgt-3 regions, designated the type I arrangement. Of particular interest was NMA0408. A search of The Institute for Genome Research complete microbial genome database with the translated amino acid sequence using the BLAST algorithm identified homology (72.5% similarity and 68.5% identity) to a single gene (HI0275) from the H. influenzae Rd genome sequence, annotated as encoding a hypothetical protein. H. influenzae and N. meningitidis are both known to have PEtn 1→6 Hep linkages in their LPS. We hypothesized that NMA0408 and HI0275 might encode PEtn-6 transferases. These proteins, 550 and 551 amino acids respectively, are similar in size to the other PEtn transferases previously identified from N. meningitidis. When the translated amino acid sequences from NMA0408 and HI0275 were aligned with the sequences of the characterized PEtn transferases, Lpt3 and LptA, it was found that they are of similar length and have a number of conserved residues over the entire lengths of the proteins (Fig. 5). Lpt3 and LptA have previously been shown to belong to a family of predicted PEtn transferases present in a wide range of gram-negative bacteria (9). When the NMA0408 and HI0275 translated products are added to alignments of the compiled sequences from other organisms, it is found that a number of residues are conserved across the family of proteins, located at intervals spanning the entire lengths of the sequences (data not shown).

FIG. 4.

Alternative arrangements of genes found in N. meningitidis at the Lgt-3 region of the chromosome. Types I to IV were identified by Zhu et al. (51), while type V was identified from the N. meningitidis serogroup C genome sequence of strain FAM18. The positions of the primers LgtG-F and LgtG-H, which bind within the genes flanking the Lgt-3 region, mtr and gmhX, respectively, and which were used to amplify across the region, are indicated. A screen of 113 N. meningitidis strains was performed by PCR amplification to identify which Lgt-3 type was present and whether the strain also possessed a copy of the lpt3 gene (31). Three strains gave atypical product sizes or no product.

FIG. 5.

Alignment of the translated sequences from NMA0408 and HI0275, identified in this study, and previously identified PEtn transferase genes (lptA and lpt3) in N. meningitidis (9, 31). Similarity and identity values for each translated product versus that of NMA0408 are, respectively, as follows: HI0275 product, 73 and 69%; Lpt3, 43 and 33%; LptA, 41 and 30%.

Mutations in NMA0408 and HI0275 result in loss of the PEtn-6 from the LPS.

To investigate the function of the NMA0408 and HI0275 gene products, each gene was inactivated by insertion of a kanamycin resistance cassette into the reading frame. Plasmid constructs containing the interrupted NMA0408 or HI0275 gene were transformed into N. meningitidis (89I, 89IgalE, 35E, and 35EgalE) or H. influenzae (Rd and nontypeable 375 and 285) strains, respectively. Electrophoretic profiles of LPS isolated from the NMA0408 mutant N. meningitidis strains and parental strains appeared to be similar when analyzed by T-SDS-PAGE (Fig. 6), although there was a consistent difference in the quality of silver staining of LPS, a phenomenon which has been seen previously associated with mutations in both the lptA and lpt3 genes (9, 31). Western blot analysis showed that reactivity to MAb L2-16 was significantly reduced. Electrophoretic profiles of LPS from H. influenzae strain Rd and isolates 285 and 375, in which HI0275 was inactivated, showed some shift in the banding pattern of the LPS compared to that for the relevant wild-type strains (Fig. 6).

FIG. 6.

Profiles of LPS isolated from parental and lpt6 mutant strains separated by T-SDS-PAGE and stained with silver. (A) Pairs of parental and lpt6 mutant N. meningitidis strains. Lanes: 1, 89I; 2, 89Ilpt6; 3, 89IgalE; 4, 89IgalElpt6; 5, 35E; 6, 35Elpt6; 7, 35EgalE; 8, 35EgalElpt6. (B) Western blot with MAb L2-16. Lanes 1 to 4 are as in panel A. (C) Pairs of wild-type and lpt6 H. influenzae mutants. Lanes: 1, Rd; 2, Rdlpt6; 3, 375; 4, 375lpt6; 5, 285; 6, 285lpt6.

The MAb L2-16 is specific for N. meningitidis and does not react with H. influenzae LPS; however, when a panel of MAbs raised against H. influenzae LPS was used to screen the HI0275 mutants, some differences between wild-type and mutant strains in reactivity with only a few MAbs were observed. For MAbs MAHI2 and MAHI3, the strain Rd HI0275 mutants reacted better with both MAbs than did the wild type. The strain 375 HI0275 mutant reacted with MAHI3, whereas the wild type did not react. MAb MAHI2 has a pattern of reactivity similar to that of MAHI3, which has been characterized as being an antibody that recognizes inner-core structures on H. influenzae LPS (3). MAHI3 has been shown to enhance bloodstream clearance of H. influenzae in an infant rat model of bacteremia (3). To investigate whether or not this MAb showed altered functional activity against the HI0275 mutant and wild-type strains, isogenic Rd strains were tested in the serum bactericidal assay using MAHI3 and a human complement source from pooled human sera. The MAb showed no functional activity against the wild-type strain and no activity against the HI0275 mutants that displayed increased binding to MAHI3 (data not shown). For both H. influenzae and N. meningitidis, a comparison of isogenic wild-type and respective HI0275 and NMA0408 mutant strains showed no difference in the killing effect in pooled normal human serum (data not shown).

O-deacylated LPS was used to further characterize by MS the structure of inner-core LPS from N. meningitidis strains 89IgalE and 35E as well as H. influenzae strain Rd and the respective N. meningitidis and H. influenzae mutants (Table 3). A consistent finding was that the mutants displayed a loss of PEtn from the inner-core region of the LPS. PEtn was still present in the lipid A region of the N. meningitidis mutants, and, as was observed previously in the H. influenzae mutant, both Kdo-phosphate (Kdo-P) and Kdo-P-PEtn glycoforms were observed. This confirms that the gene products of NMA0408 and HI0275 were specific for the transfer of PEtn-6 and not involved in the general biosynthesis or presentation of PEtn. Proof of the presence of PEtn on the Kdo-P moiety in the H. influenzae mutant was obtained from MS/MS studies in negative-ion mode on the triply charged ion at m/z 989, which revealed the characteristic singly charged ion for the P-PEtn moiety at m/z 220 (Fig. 7A). The absence of PEtn at the 6 position of HepII was also indicated by MS studies in positive-ion mode, whereby a high orifice voltage gave a series of fragment ions that revealed the absence of the characteristic ion at m/z 316 for a Hep-PEtn group and a series of fragment ions consistent with the absence of PEtn from the HepII location (Fig. 7B) (40). Additionally, nuclear magnetic resonance studies confirmed that the characteristic signals diagnostic for PEtn presence at the 6 position of HepII in both N. meningitidis (8) and H. influenzae (40) were absent from the mutant O-deacylated-LPS spectra (data not shown). These genes were named lpt6 (LPS PEtn transferase at position 6) and are referred to as either N. meningitidis lpt6 (NMA0408) or H. influenzae lpt6 (HI0275) in this paper.

TABLE 3.

Negative-ion CE-MS and CE-MS-MS data and proposed compositions of O-deacylated LPS from N. meningitidis strains 89IgalElpt6 and 35Elpt6 and H. influenzae strain Rd/pt6^a

Strain	Observed m/z for:		Molecular mass (Da)		Molecular mass (amu) of:		Proposed composition
	(M-3H)3−	(M-2H)2−	Observed	Calculated	Lipid A-OHb	Core OSc
N. meningitidis
89I	1,027.3		3,085	3,083.8	952	2,133	Sial, 3Hex, 2HexNAc, 2Hep, 2Kdo, PEtn, lipid A-OH
89Ilpt6	889		2,670	2,669.5	952	1,720	3Hex, 2HexNAc, 2Hep, 2Kdo, lipid A-OH
	986.3		2,962	2,960.7	952	2,010	Sial, 3Hex, 2HexNAc, 2Hep, 2Kdo, lipid A-OH
89IgalE	754.8	1,132.3	2,266	2,265.1	952	1,314	Hex, HexNAc, 2Hep, 2Kdo, PEtn, lipid A-OH
89IgalElpt6	713	1,070	2,142	2,142	952	1,190	Hex, HexNAc, 2Hep, 2Kdo, lipid A-OH
35E	930		2,793	2,792.5	952	1,841	3Hex, 2HexNAc, 2Hep, 2Kdo, PEtn, Lipid A-OH
	943		2,832	2,831.7	952	1,880	4Hex, 2HexNAc, 2Hep, 2Kdo, lipid A-OH
	984		2,955	2,954.5	952	2,003	4Hex, 2HexNAc, 2Hep, 2Kdo, PEtn, lipid A-OH
	1,081		3,246	3,245.7	952	2,294	Sial, 4Hex, 2HexNAc, 2Hep, 2Kdo, PEtn, lipid A-OH
35Elpt6	889		2,670	2,669.5	952	1,720	3Hex, 2HexNAc, 2Hep, 2Kdo, lipid A-OH
	943		2,833	2,831.7	952	1,881	4Hex, 2HexNAc, 2Hep, 2Kdo, lipid A-OH
	1,040.5		3,124	3,122.9	952	2,172	Sial, 4Hex, 2HexNAc, 2Hep, 2Kdo, lipid A-OH
H. influenzae
Rdlpt6	718	1,078	2,156	2,155.8	952	1,204	Hex, 3Hep, Kdo-P, PCho, lipid A-OH
	759	1,139	2,279	2,278.9	952	1,327	Hex, 3Hep, Kdo-P, PEtn, PCho, lipid A-OH
	826	1,240	2,480	2,480.1	952	1,528	3Hex, 3Hep, Kdo-P, PCho, lipid A-OH
	867	1,301	2,603	2,603.2	952	1,661	3Hex, 3Hep, Kdo-P, PEtn, PCho, lipid A-OH
	880	1,321	2,642	2,642.3	952	1,690	4Hex, 3Hep, Kdo-P, PCho, lipid A-OH
	921	1,382	2,765	2,765.4	952	1,813	4Hex, 3Hep, Kdo-P, PEtn, PCho, lipid A-OH
	948	1,423	2,845	2,845.5	952	1,893	4Hex, HexNAc, 3Hep, Kdo-P, PCho, lipid A-OH
	989	1,484	2,968	2,968.5	952	2,016	4Hex, HexNAc, 3Hep, Kdo-P, PEtn, PCho, lipid A-OH

^aAverage mass units (amu) were used for calculation of molecular mass based on proposed composition as follows: Hex (hexose), 162.15 Da; Hep (heptose), 192.17 Da; HexNAc (N-acetylhexosamine), 203.19 Da; Kdo, 220.18 Da; PEtn, 123.05 Da; PCho (phosphocholine), 165.05 Da; Sial (sialic acid), 291.18 Da.

^bOnly the basal O-deacylated lipid A species (952 amu) is detailed for simplicity. For all N. meningitidis strains examined all lipid A species commonly encountered (952, 1,032, 1,075, 1,155 and 1,278 amu) were observed. For H. influenzae, a lipid A species of 952 amu is the only species seen.

^cAs deduced from lipid A-OH size determination. OS, oligosaccharide.

FIG. 7.

(A) Negative-ion MS/MS product ion spectrum for the triply charged ion at m/z 989³⁻ of H. influenzae strain Rdlpt6. Ions indicative of P-PEtn (m/z 220⁻), core oligosaccharide (OS; m/z 1007²⁻), core OS with loss of CO₂ (m/z 985²⁻), and lipid A-OH species (m/z 952⁻ and 475²⁻) are as indicated. (B) Positive-ion MS spectrum of Rdlpt6 at a high orifice voltage. A fragmentation pattern consistent with absence of PEtn at HepII and other ions is indicated. The spectrum for the parental strain Rd has been previously published by Risberg et al. (40).

Characteristics of the chromosomal location and prevalence of lpt6 in natural isolates of N. meningitidis and H. influenzae.

The gene N. meningitidis lpt6 is located within a region of the N. meningitidis chromosome required for LPS biosynthesis, the Lgt-3 region. This region of the genome, which is heterogeneous among strains, can be found in any one of four arrangements, termed types I to IV (51). A fifth arrangement, whereby all three genes, lgtG, NMA0408, and NMA0407, are present within the Lgt-3 region has been identified in the N. meningitidis serogroup C genome sequence strain FAM18, designated here type V (Fig. 4).

PCR amplification was performed to determine if the genes lpt6 and lpt3 were present in the N. meningitidis LPS immunotyping reference L1 to L10 strains, for each of which the LPS structure has been characterized (11, 13, 22, 29, 35, 36, 49). The lpt6 gene was found in the L2, L4, L6, and L10 strains, and lpt3 was present in the L1, L3, L7, L8, and L9 strains. The distribution of the two genes was consistent with what would be expected from the known LPS structures, i.e., those strains with PEtn-3 possessed only lpt3, while those strains with PEtn-6 possessed only lpt6. The number of strains screened by PCR amplification was increased to include a more diverse population both in disease profile and geographical location. These additional studies used a carriage collection from the Czech Republic (26), an invasive disease collection from the United Kingdom (16), and other laboratory N. meningitidis strains (total number = 113). This screen identified 14 strains that were found to possess the lpt6 gene only, 57 strains that possessed the lpt3 gene only, 41 strains that possessed both the lpt3 and lpt6 genes, and only one strain that contained neither lpt3 nor lpt6.

PCR was used to characterize the Lgt-3 region from the set of 113 N. meningitidis strains. Oligonucleotide primers lgtG-F and lgtG-H, which bound within the genes mtr and NMB2033/NMA0405, flanking the Lgt-3 region, were used to amplify by PCR across the Lgt-3 region to assess the distribution of the different Lgt-3 types (Fig. 4). The Lgt-3 types II, IV, and V were present in approximately equal numbers (22, 28, and, 24%, respectively) of strains, while types I and III were found in about one-half the number of strains (12 and 13% respectively), compared to the other types. The N. meningitidis lpt6 genes from nine representative strains that had each been shown to contain one of the three different Lgt-3 types containing the lpt6 gene were partially sequenced. DNA sequences obtained from the majority of these genes showed little difference from the Z2491[C751] genome sequence. (http://www.sanger.ac.uk/Projects/N_meningitidis/seroC/seroC.shtml) (data not shown).

Our data also provide compelling evidence that H. influenzae lpt6 encodes the PEtn-6 transferase. Furthermore, previous work has indicated that all strains of H. influenzae investigated possess PEtn at the 6 position of HepII (32). HI0275 is located within the strain Rd genome sequence between the genes HI0274, encoding a glutamyl-tRNA synthetase, and HI0276, encoding an RNase (Fig. 8A). To test for the prevalence of lpt6 in the H. influenzae population, a PCR screen of strains was undertaken using primers based on sequences within the reading frames of the HI0275 and HI0276 genes. Seventy H. influenzae strains, representing all capsular serotypes and nonencapsulated (nontypeable) strains from both carriage and disease, were investigated. We found that lpt6 was invariably present in these H. influenzae strains and was adjacent to HI0276 in all cases.

FIG. 8.

Alignment of regions of DNA upstream and downstream of lpt6 genes from H. influenzae and N. meningitidis. (A) Schematic representation of genes surrounding H. influenzae lpt6 and N. meningitidis lpt6. NMA, NMC, and H. influenzae (Hi) regions are taken from the relevant genome sequences. (B) Alignment using CLUSTALX of the intergenic regions upstream of the start codon of N. meningitidis lpt6 in the genomes of NMA and NMC strains and 200 bp upstream of the start codon of H. influenzae lpt6. (C) Alignment using CLUSTALX of the intergenic regions downstream of the stop codon of N. meningitidis lpt6 in the genomes of NMA and NMC strains and 200 bp downstream of the stop codon of H. influenzae lpt6.

Comparison of the DNA sequences immediately up- and downstream of the HI0275 and NMA0408 genes indicated large regions of homology between the sequences from both H. influenzae and N. meningitidis (Fig. 8B). There is homology between the N. meningitidis intergenic regions and the H. influenzae intergenic region and beyond into the neighboring tRNA-Val gene in H. influenzae. Similarly, the alignment of the downstream intergenic regions shows homology among the N. meningitidis intergenic regions extending into the downstream H. influenzae RNase gene (Fig. 8C). This observation raises the possibility that one of the highly related lpt6 genes found in these two species has been obtained from the other species by lateral transfer.

DISCUSSION

This study has identified an LPS PEtn transferase gene (lpt6) responsible for addition of PEtn at the 6 position of HepII within the inner cores of both N. meningitidis and H. influenzae LPS. The transposon mutagenesis approach utilized in this study identified 18 transformant N. meningitidis colonies that were unable to react with MAb L2-16, a MAb that specifically recognizes PEtn-6 in N. meningitidis LPS. The transposon insertion sites in over half of these 18 nonreacting colonies were found to be in known genes related to the biosynthesis of the LPS and would have prevented formation of the diheptose backbone and other elements of the LPS which constitute the inner-core epitope that reacts with MAb L2-16. This has allowed the LPS epitope to which MAb L2-16 binds to be substantially identified directly by genetic methods. The transposon was not found inserted in any of the genes required for steps in LPS synthesis preceding addition of HepI to the LPS, i.e., in lipid A synthesis and addition of Kdo. Although it has been reported that N. meningitidis bacteria can unusually survive when lipid A is absent if capsule is present (44), it may be that the reduced growth rates in these mutants would result in their production of colonies of insufficient size to be screened effectively in the colony immunoblotting procedure.

The remaining eight transposon insertion sites were clustered within Lgt-3, one of three regions of the N. meningitidis genome containing LPS biosynthesis genes, studied by Zhu and colleagues (51). The Lgt-3 locus bounded by genes mtr and gmhX can contain up to three possible genes in arrangements termed type I to V, found in different N. meningitidis strains (Fig. 4). The Lgt-3 region of N. meningitidis strain 89I is of type I, containing only lpt6 and lgtG, encoding a phase-variable glycosyltransferase, which is responsible for addition of Glc at the 3 position (Glc-3) of HepII and which is thought to compete with the gene product from lpt3. However, the inner-core LPS of this strain does not contain Glc-3, as the homopolymeric tract of lgtG predominantly is of a length that results in a truncated and inactive LgtG protein.

Transposon insertions were identified within the lgtG gene (one hit), mtr (three hits), and NMA0408 (four hits). Homology searches with the translated sequence of NMA0408 revealed only one significant homologue, which was HI0275, in H. influenzae. We had expected to find a homologue of lpt6 within H. influenzae, as all LPS analyzed to date from this organism have the PEtn-6 substitution on HepII (32). To confirm that NMA0408 and HI0275 encoded the likely PEtn transferases, insertional mutations with a kanamycin resistance cassette were made in each gene. MS analysis of LPS isolated from both N. meningitidis and H. influenzae lpt6 mutant strains confirmed a loss of PEtn from the inner core in both species. Biochemical confirmation of the Lpt6 function remains to be carried out.

The other MAb L2-16-nonreactive strain transposon hits in the Lgt-3 region were in the genes lgtG and mtr. For the insertion in the lgtG gene, when the Lgt-3 region was mapped with various combinations of primers and PCR amplification, it was shown that there had been a rearrangement (data not shown). The gene mtr proved impossible to mutate in a directed manner with an antibiotic resistance cassette (data not shown). A search of the N. meningitidis genome sequences did not reveal any other open reading frames) that were likely to act as tryptophan transporters in this organism. Some possible polar effect of mutation of mtr on the lpt6 gene cannot be ruled out, and the details of the genetic regulation of this region require further investigation.

The N. meningitidis and H. influenzae lpt6 genes appear to be related to two PEtn transferases specific for N. meningitidis LPS previously identified in our laboratory. Lpt3 and LptA are transferases adding PEtn to the 3 position of HepII (31) and to the lipid A of N. meningitidis LPS (9), respectively, and are part of an apparent family of PEtn transferases present in a wide range of gram-negative bacteria. When the translated amino acids from this family of genes are aligned, it is found that conserved residues are present across the entire length.

A search of recently completed genome sequences from related species has identified homologues of genes encoding the PEtn-6 transferase in Haemophilus somnus and Neisseria gonorrhoeae. H. somnus is known to have PEtn-6 on HepII in the inner-core structure of its LPS (7), while the structure for the N. gonorrhoeae strain FA1090 used in the genome sequencing project has not yet been elucidated. The possible transfer of the lpt6 gene between species has been inferred by the homology of the intergenic regions surrounding NMA0408 in N. meningitidis and HI0275 in H. influenzae.

In H. influenzae, all strains tested contain the HI0275 gene, and each analyzed LPS derived from H. influenzae to date contains PEtn-6. This is the only stoichiometric PEtn substitution in H. influenzae LPS, but partial substitution of Kdo-phosphate by PEtn has been identified in all strains investigated, while a partial substitution of HepIII with PEtn is found in some strains (32). The genetic determinants of these partial substitutions remain to be elucidated. The situation is somewhat different in N. meningitidis, where stoichiometric PEtn substitution on HepII of the inner core is known to occur at different sites depending on the strain. The gene lpt3 is responsible for substitution of the alternative PEtn-3 on HepII (31). In some N. meningitidis strains it is present but contains a large deletion (∼800 bp), making the product apparently nonfunctional. Screening a large number of strains for the presence of full-length copies of the lpt3 and lpt6 genes by PCR has shown that 36% of strains tested (n = 113) have both genes, while 50% have lpt3 only, 12% have lpt6 only, and only 1 strain has neither. In structural analyses of N. meningitidis LPS to date, very few N. meningitidis strains have been identified as producing LPS comprising mixtures of glycoforms with either PEtn-3 or PEtn-6 or with a di-PEtn substitution at HepII, whereas the potential for making this structure is in fact present in over one-third of strains tested. In those strains whose LPS has been analyzed and found to contain both PEtn-3 and PEtn-6, the proportion of glycoforms containing each substituent has been varied (8, 39). It is interesting that strain BZ157 has a greater number of LPS molecules with PEtn-3 only, while strain 2220Y has stoichiometric distribution of both PEtn-3 and PEtn-6. The gene lgtG, also located in the Lgt-3 region, is responsible for addition of Glc-3 and is subject to phase-variable expression via a homopolymeric tract located within the reading frame (2). It has been postulated that LgtG outcompetes Lpt3 for occupancy of the 3 position of HepII when LgtG is expressed, resulting in no PEtn-3 incorporation. It may be more than coincidence that both lpt6 and lgtG are located adjacent in the same chromosomal region. The details of the balance in LPS synthesis between Lpt3, LgtG, and Lpt6 expression and how this influences the resulting LPS structure remain to be fully elucidated.

Previous studies in our laboratory have shown that PEtn-3, when present in N. meningitidis LPS, can have a pronounced effect on the susceptibility of N. meningitidis bacteria to opsonophagocytic and bactericidal killing mediated by MAb L3B5 (31). In the present study we have been unable to demonstrate any clear functional role for PEtn-6 on either N. meningitidis or H. influenzae bacteria, although this is partly due to not having suitable MAbs which demonstrate bacterial killing available to allow us to carry out the relevant studies.

In summary, we have identified and characterized a third gene encoding a PEtn transferase in N. meningitidis LPS synthesis, which has a homologue carrying out the equivalent function in H. influenzae. The contribution of PEtn-3, PEtn-6, and PEtn on lipid A to the biology of the LPS of these bacteria can now be studied in detail.

ADDENDUM IN PROOF

This gene has also been identified by C. M. Kahler, A. Datta, R. W. Carlson, Y.-L. Tzeng, L. Martin, and D. Stephens, and their results were presented at the 14th International Pathogenic Neisseria Conference 2004.