Genetics of capsule O-acetylation in serogroup C, W-135 and Y meningococci



Capsular polysaccharides of serogroup C, W-135 and Y meningococci were previously reported to be O-acetylated at the sialic acid residues. There is evidence that O-acetylation affects the immunogenicity of polysaccharide vaccines. We identified genes indispensable for O-acetylation of serogroup C, W-135 and Y meningococci downstream of the capsule synthesis genes siaA–D. The genes were co-transcribed with the sia operon as shown by reverse transcription polymerase chain reaction analysis. The putative capsular polysaccharide O-acetyltransferases were designated OatC and OatWY. The protein OatWY of serogroups W-135 and Y showed sequence homologies to members of the NodL–LacA–CysE family of bacterial acetyltransferases, whereas no sequence homology with any known protein in the different databases was found for the serogroup C protein OatC. In serogroup W-135 and Y meningococci, several clonal lineages either lacked OatWY or OatWY was inactivated by insertion of IS1301. For serogroup C meningococci, we observed in vitro phase variation of O-acetylation, which resulted from slipped-strand mispairing in homopolymeric tracts. This finding explains the observation of naturally occurring de-O-acetylated serogroup C meningococci. Our report is the first description of sequences of sialic acid O-acetyltransferase genes that have not been cloned from either other bacterial or mammalian organisms.


The Gram-negative bacterium Neisseria meningitidis is a major cause of bacterial meningitis and septicaemia worldwide. Meningococci express an extracellular polysaccharide capsule that is a prerequisite for meningococcal virulence (Vogel et al., 1996). Meningococcal capsular polysaccharides have been classified into 13 distinct serogroups. The capsules were defined by antibody reactivity and structural analysis. Clinically important serogroups are the serogroups A, B, C, W-135 and Y.

The serogroup B, C, W-135 and Y capsules are composed of sialic acid [N-acetylneuraminic acid (Neu5Ac)]. Sialic acids are a family of 9-carbon carboxylated sugars commonly found as the terminal sugar of eukaryotic oligosaccharides and not generally found in plants, most invertebrates or prokaryotes (Klein and Roussel, 1998; Traving and Schauer, 1998). Besides the meningococcal serogroups B, C, W-135 and Y, only Escherichia coli K-1 and K-92 and type III group B streptococci contain sialic acids in their capsular polysaccharides. More than 40 different types of naturally occurring sialic acids have been identified. This diversity results from substituents such as acetyl, methyl, lactyl, sulphate and phosphate groups. Most commonly, one or more of the hydroxyl groups in positions 4, 7, 8 and 9 are substituted by acetyl groups. The O-acetylation influences many properties of the sialic acid molecule, i.e. the size and the net charge. Furthermore, biological properties can be affected in O-acetylated eukaryotic glycoproteins including activation of the alternative complement pathway, the activities of enzymes involved in sialic acid metabolism, the specificity of recognition of sialylo-oligosaccharides by influenza viruses and the binding of antibodies to gangliosides (Varki and Kornfeld, 1980; Cheresh et al., 1984; Rogers et al., 1986). Genes encoding bacterial or mammalian sialic acid O-acetyltransferases have not been cloned until now.

The serogroup B meningococcal polysaccharide capsule is a homopolymer of α-2-8-linked N-acetylneuraminic acid, whereas the serogroup C meningococcal polysaccharide capsule is a homopolymer of α-2-9-linked N-acetylneuraminic acid (Bhattacharjee et al., 1975). The serogroup W-135 and Y capsular polysaccharides are heteropolymers, composed of α-2-6-linked sialic acid and galactose or glucose respectively (Bhattacharjee et al., 1976). The genes necessary for meningococcal capsule expression are clustered within the capsule gene complex (cps) that comprises five regions (A–E) (Frosch et al., 1989). The genes of regions A, B and C are responsible for capsule synthesis, modification and transport respectively. In meningococci with sialic acid-containing capsules, region A comprises the genes siaA–D with siaA–C encoding the enzymes required for activated sialic acid synthesis, and siaD encoding the polysialyltransferases of the different serogroups (Edwards et al., 1994; Claus et al., 1997; Swartley et al., 1997). The four genes constitute a transcriptional unit (Edwards et al., 1994). The A+T content of the capsule synthesis genes siaA–D is significantly higher than the A+T content of the remaining cps locus and of the whole genome, suggesting that region A was acquired by horizontal gene transfer.

O-acetylation patterns of meningococcal capsular polysaccharides have been investigated by nuclear magnetic resonance spectroscopy (NMR). 13C-NMR as well as 1H-NMR were used to analyse the capsular polysaccharide structures (Bhattacharjee et al., 1976; Jennings et al., 1977; Lemercinier and Jones, 1996). Of the sialic acid-containing capsules, the serogroup B capsule was not O-acetylated in contrast to the E. coli K-1 capsule which shares an identical composition of α-2-8-linked sialic acid. In serogroups C, W-135 and Y, the hydroxyl groups of the sialic acid glycerol residue are the sites of O-acetylation. In serogroup C, sialic acid is acetylated at positions C7 or C8 respectively. In serogroups W-135 and Y, sialic acid is acetylated at positions C7 or C9 respectively. Recently, in the UK, the proportion of serogroup C, W-135 and Y isolates with O-acetylated capsular polysaccharides was estimated to be 88% for the serogroup C isolates, 8% for the serogroup W-135 isolates and 79% for the serogroup Y isolates. A similar distribution was found in both carrier and case strains (Borrow et al., 2000; Longworth et al., 2002).

In bacteria, not only the capsular polysaccharides, but also the O-antigenic polysaccharide chains of lipooligosaccharides (LPS) and the peptidoglycan, are targets for O-acetylation. The peptidoglycan of some species of eubacteria, i.e. Proteus mirabilis and Neisseria gonorrhoeae, has been reported to be O-acetylated, which confers both intrinsic and complete resistance to lysozyme hydrolysis (Dupont and Clarke, 1991). Two families of proteins that participate in the acylation of exported carbohydrate moieties have been identified. (i) One family comprises integral membrane proteins that acylate macrolide antibiotics (Streptomyces spp.) and O-acetylate LPS O-antigen (Legionella pneumophila Lag-1, Salmonella typhimurium OafA, Shigella flexneri bacteriophage SF6 Oac) as well as Nod factors (Rhizobium leguminosarum NodX) (Verma et al., 1991; Hara and Hutchinson, 1992; Firmin et al., 1993; Arisawa et al., 1994; Slauch et al., 1996; Zou et al., 1999). Although many regions of these integral membrane proteins are similar, the homology is particularly striking in regions predicted to be transmembrane. Whether these highly conserved regions correspond to a conformational determinant necessary for the reaction or to a binding site for a common substrate is not clear. (ii) The second so-called NodL–LacA–CysE acetyltransferase family comprises cytoplasmic proteins that use acetyl coenzyme A (Ac-CoA) as the acetyl donor. Members of this family O-acetylate dissimilar substrates including thiogalactoside (E. coli LacA), serine (E. coli CysE), N-acetylglucosamine (Rhizobium leguminosarum NodL), antibiotics such as chloramphenicol (E. coli Cat) and streptogramin A (Staphylococcus aureus Vat and Enterococcus faecium Sat) and capsular polysaccharides (S. aureus Cap1G, Cap5H and Cap8J) (Hediger et al., 1985; Denk and Bock, 1987; Downie, 1989; Parent and Roy, 1992; Allignet et al., 1993; Lin et al., 1994; Sau et al., 1997). This family of O-acetyltransferases shares amino acid homology at the carboxy-terminus of the proteins, which includes the putative active site of the enzyme. Besides these two groups, other putative O-acetyltransferases are known, both cytosolic and integral membrane proteins, which share no homologies with the other two groups (i.e. Cps9vM and Cps9vO from Streptococcus pneumoniae; AlgI, AlgJ and AlgF from Pseudomonas aeruginosa) (Franklin and Ohman, 2002; van Selm et al., 2002).

The biological relevance of O-acetylation of bacterial surface polysaccharides is mostly unclear. A contribution of O-acetylation to pathogenicity was studied for some species. O-acetylation of alginate, the exopolysaccharide of P. aeruginosa, is a requirement for biofilm formation (Nivens et al., 2001). O-acetylation of alginate also maximizes the resistance of mucoid P. aeruginosa to antibody-independent opsonic killing and is the molecular basis for resistance to normally non-opsonic but alginate-specific antibodies found in the sera of infected cystic fibrosis patients (Nivens et al., 2001; Pier et al., 2001). A spontaneous Lag-1 mutant of L. pneumophila lacked O-acetyltransferase activity and failed to produce high-molecular-weight long-chain O-polysaccharide. However, Lag-1 mutation did not affect serum resistance and interaction with Acanthamoeba castellanii and macrophages respectively (Lück et al., 2001). O-acetylation of the O5 lipopolysaccharide antigen of Salmonella typhimurium gave rise to a slight increase in virulence in the animal infection model (Kim and Slauch, 1999). The biological effects of capsule O-acetylation on S. aureus serotype 5 virulence were examined in an in vitro opsonophagocytic assay and in a mouse model of bacteraemia and renal abscess formation. The data from both experiments (in vitro and in vivo) suggest that the O-acetylated polysaccharide may be more proficient than the de-O-acetylated polysaccharide in protecting the Staphylococcus from immune clearance (Bhasin et al., 1998).

O-acetyl groups have been shown to affect the antigenicity and immunogenicity of several bacterial polysaccharides. Monoclonal antibodies against the O-acetylated serotype 5 capsule of S. aureus were predominantly reactive with the acetyl substituents, demonstrating their immunodominant nature (Fattom et al., 1998). O-acetylated variants of the E. coli K-1 capsular polysaccharide were more immunogenic than variants with de-O-acetylated polysaccharide (Orskov et al., 1979). Furthermore, complete de-O-acetylation eliminated the immunogenicity of the Vi polysaccharide of Salmonella typhi (Szu et al., 1991). In contrast, the O-acetylation of the pneumococcal type 9 V polysaccharide was not required for the induction of opsonic antibodies (McNeely et al., 1998).

Although O-acetylation of meningococci has been known for decades, the biological role with regard to biological fitness or pathogenicity is unclear. On the one hand, this results from the lack of an appropriate animal infection model and, on the other hand, the genetic basis of O-acetylation was unknown so that isogenic mutants were not available. In vaccinees receiving de-O-acetylated serogroup C polysaccharide conjugate vaccines, levels of IgG directed against de-O-acetylated polysaccharide were twice as high as levels of IgG specific to O-acetylated polysaccharide, and the serum bactericidal activity was significantly higher against de-O-acetylated meningococcal strains (Richmond et al., 2001). Furthermore, it has been shown that asymptomatic carriers had higher titres of antibodies to the de-O-acetylated serogroup C variant, which was assumed to explain the fact that most isolates from patients with meningitis were O-acetylated (Arakere and Frasch, 1991). In contrast, it was shown recently that the immunogenicity of meningococcal serogroup A polysaccharide depended on the presence of O-acetyl groups (Berry et al., 2002). Data on the impact of O-acetylation on immunogenicity of serogroup W-135 or Y polysaccharides are not available yet, although conjugate vaccines including serogroups W-135 and Y are now progressing through trials (Campbell et al., 2002; Rennels et al., 2002).

In the present study, we describe for the first time genes required for O-acetylation of sialic acids. Furthermore, we investigated the distribution and regulation of the genes in the meningococcal population. Our findings will allow future studies on the impact of O-acetylation in meningococcal virulence and immunity.


Identification of the capsular O-acetyltransferase genes in serogroup C, W-135 and Y meningococci

In contrast to the genes required for capsule synthesis in meningococci, the genes encoding O-acetyltransferases are unknown. In the search for such genes, we assumed the meningococcal capsular O-acetyltransferase genes to be located in the vicinity of the capsule synthesis genes. Therefore, the region downstream of the sia operon was amplified by polymerase chain reaction (PCR) (Fig. 1), and the DNA sequence was obtained. In serogroup C strain 2120 (sequence type [ST]-11, Oac+), an open reading frame (ORF) was found 28 bp downstream of siaD, which was named oatC (EMBL accession number AJ243686). It was 1383 bp in length with an A+T content of 71.22%, which corresponds to the elevated A+T content of region A. OatC encoded a putative protein of 461 amino acids with a molecular mass of ≈ 53.28 kDa. Comparison of the sequences with the GenBank and the swissprot databases revealed no homologies with any known genes or proteins respectively. PCR of the siaD downstream region of serogroup W-135 strain 2144 (ST-169, Oac+) and serogroup Y strain 172 (ST-166, Oac+) resulted in a DNA fragment of identical length and sequence. A 636 bp ORF with an A+T content of 68.24% was located 80 bp downstream of the respective siaD genes. This ORF was named oatWY (EMBL accession number Y13969). The putative protein comprised 212 amino acids and had a predicted molecular mass of 23.28 kDa. When compared with sequences deposited in the swissprot database, OatWY exhibited 30–40% identity and 50–60% similarity over 100 amino acids at the carboxy-terminus to various bacterial acetyltransferases that belong to the NodL–LacA–CysE family of acetyltransferases with a bacterial transferase hexapeptide repeat ([LIV]-G-X(4)) (Fig. 2). According to the transmembrane prediction program tmhmm (Centre for Biological Sequence Analysis, Technical University of Denmark, Lyngby, Denmark), both OatC and OatWY are soluble proteins without any transmembrane domain.

Figure 1.

Schematic depiction of region A, the siaD downstream region and parts of region D (galE) of Neisseria meningitidis serogroups C, W-135 and Y respectively. Region A comprises the genes siaA–D (capsule synthesis). The order of the genes in the siaD downstream region was confirmed by DNA sequencing for serogroup C (strain 2120) and for serogroups W-135 and Y (strains 2140 [ST-11], 2225 [ST-11], 2227 [ST-183], 2232 [ST-185], 2235 [ST-189], 2239 [ST-174]). Oligonucleotides used for amplification of the siaD downstream region are presented. NMB0065, ORF 65 of the genome sequence of the meningococcal strain MC58; galU, UTP-glucose-1-phosphate uridyltransferase

Figure 2.

Amino acid homologies among the C-terminal regions of the deduced amino acid sequences of OatWY and seven related acetyltransferases. The bacterial transferase hexapeptide repeats are boxed. Rl NodL, Rhizobium leguminosarum lipooligosaccharide acetyltransferase (Downie, 1989); Ec LacA, E. coli thiogalactoside transacetylases (Hediger et al., 1985); Ec CysE, E. coli serine acetyltransferase (Denk and Bock, 1987); Ec Cat, E. coli chloramphenicol acetyltransferase (Parent and Roy, 1992); Sa Vat, S. aureus streptogramin A acetyltransferase (Allignet et al., 1993); Sa Cap1G, S. aureus capsule type 1 putative O-acetyltransferase (Lin et al., 1994); Sa Cap5H, S. aureus capsule type 5 putative O-acetyltransferase (Bhasin et al., 1998); Nm OatWY, N. meningitidis serogroup W-135 and Y putative capsular O-acetyltransferase (this study).

The oatC and oatWY genes were located downstream of the sia genes, which were shown to constitute a polycistronic operon (Edwards et al., 1994). Furthermore, they shared a high A+T content with the sia operons. Putative ribosomal binding sites were identified upstream of the oatC and oatWY genes, but no promoter sequence was detectable (oatC: TGGAG, 11 bp upstream of the start codon ATG; oatWY: AGGAA, 5 bp upstream of the start codon ATG). Therefore, we assumed that the O-acetyltransferase genes are co-transcribed together with the capsule synthesis genes. To verify this hypothesis, a reverse transcription (RT)-PCR was performed with total bacterial RNA and specific primers within the siaD genes and the O-acetyltransferase genes. The RT-PCR resulted in the same product as that amplified by a standard PCR using chromosomal DNA as a template. No PCR product was visible after a PCR run with RNA as a template, indicating that the RNA preparation was free of chromosomal DNA (Fig. 3). These data showed that the O-acetyltransferase genes are co-transcribed with the sia operon and may thus be considered as part of region A of the cps locus.

Figure 3.

RT-PCR demonstrating co-transcription of siaD and oatC or oatWY respectively. PCRs and RT-PCRs were performed with serogroup-specific primers for siaD and the respective oat gene. Lanes 1–4, serogroup C; lanes 5–8, serogroup W-135; lanes 9–12, serogroup Y; lanes 1, 5 and 9, PCR without a template; lanes 2, 6 and 10, PCR with chromosomal DNA as a template; lanes 3, 7 and 11, PCR with total RNA as a template; lanes 4, 8 and 12, RT-PCR with total RNA as a template. M, DNA fragment length marker.

To test whether the putative meningococcal O-acetyltransferases are functional, the oatC gene in strain 2120 and the oatWY gene in strains 2144 and 172 were mutated by insertional inactivation. Subsequently, the resulting mutants were complemented in trans by cloning the oatC gene under the control of an opa promoter and the oatWY gene under the control of an opc promoter. Correct recombination events were proved by Southern hybridizations with probes specific to the respective O-acetyltransferase genes, the hrtA locus and the kanamycin resistance cassette (kan) (data not shown). Capsule O-acetylation of the wild-type strains and the mutant strains of all serogroups was tested using specific antibodies by enzyme-linked immunosorbent assay (ELISA) and dot blot (Table 1). The oatC and oatWY knock-out mutants of the respective serogroups were shown to be de-O-acetylated. After complementation, the capsules of the mutants regained O-acetylation. These data demonstrated that oatC and oatWY are essential for capsule O-acetylation. For oatWY, this finding, together with the sequence homologies to the NodL–LacA–CysE family of acetyltransferases and the physical association to region A, provides a formal proof that the O-acetyltransferase gene was identified. For oatC, complementation analysis revealed the essential nature of the gene for O-acetylation. Final proof of the assumed function requires in vitro assays using purified OatC and capsular polysaccharide that will be carried out in future.

Table 1. Monoclonal antibodies specific to O-acetylation.
2016.3Serogroup C, Oac and Oac+ Rubinstein and Stein (1988)
2055.5Serogroup C, Oac+ Rubinstein and Stein (1988)
MnW1-3Serogroup W-135, Oac and Oac+P. Fernsten (Wyeth Research, USA)
MnW78-59Serogroup W-135, Oac+P. Fernsten (Wyeth Research, USA)
MnY4-1Serogroup Y, Oac and Oac+P. Fernsten (Wyeth Research, USA)
MnY38-49Serogroup Y, Oac+P. Fernsten (Wyeth Research, USA)

Sequence variations in homopolymeric tracts of oatC

Approximately 12% of serogroup C meningococci isolated in the UK have been shown to be de-O-acetylated (Borrow et al., 2000). We could show by PCR amplification and Southern hybridization that Oac strains harboured oatC (data not shown). Therefore, we sequenced the oatC gene amplified by PCR of O-acetylated (n = 16) and de-O-acetylated (n = 10) strains to investigate the reasons for de-O-acetylation (Table 2). In 14 O-acetylated strains, a poly(T) and a poly(A) homopolymeric tract with seven nucleotides each were found at positions 210–216 and 458–464 of the oatC gene respectively. In two Oac+ strains, we found non-synonymous single nucleotide polymorphisms within the poly(T) tract [(T)2C(T)4 instead of (T)7]. All 10 Oac strains showed sequence variations in either the poly(T) or the poly(A) tract respectively. All these mutations resulted in premature stop codons.

Table 2. Homopolymeric tracts in oatC.
Serogroup C
No. of
isolates (n)
Homopolymeric tract
Position 210–216Position 458–464
De-O-acetylated 7(T)7(A)6
(n = 10) 1(T)7(A)8
(n = 16) 2(T)2C(T)4(A)7

Next, we investigated the possibility of in vitro phase variation from the O-acetylated to the de-O-acetylated phenotype. For this purpose, a bacterial suspension of the O-acetylated serogroup C wild-type strain 2120 was treated with the bactericidal monoclonal antibody (mAb) 2055.5 and normal human serum as a complement source. mAb 2055.5 directed against the O-acetyl group of the serogroup C polysaccharide selectively killed O-acetylated variants in the bactericidal assay. Two successive bactericidal assays resulted in a 2000-fold reduction in colony-forming units (cfu), compared with treatment with normal human serum alone. After these treatments, every tenth colony was negative for capsular O-acetylation as demonstrated by immunocolony blot analysis. We therefore estimate the frequency of phase variation to the de-O-acetylated phenotype to be ≈ 1:20 000. De-O-acetylated variants were isolated after detection by immunocolony blots and, subsequently, the oatC gene was sequenced. De-O-acetylated variants exhibited a single nucleotide deletion at the poly(T) stretch at positions 210–216. This finding proved that naturally occurring variants of O-acetylation evolve by slipped-strand mispairing.

Population biology of oatWY

Recently, meningococcal serogroup W-135 and Y isolates from the UK have been investigated for O-acetylation. Eight per cent of the W-135 isolates and 79% of the Y isolates were found to be O-acetylated (Longworth et al., 2002). In this study, the genetic basis for de-O-acetylation was investigated using a collection of 34 serogroup W-135 and Y strains isolated between 1983 and 1997 in five countries (Table 3). All isolates were typed by multilocus sequence typing (MLST) showing that the collection was representative and genetically diverse. The siaD downstream region was amplified by PCR (HC53/HC46) and subsequently sequenced. Five variants of the siaD downstream region were identified (Fig. 4). In two variants, the oatWY gene was interrupted by IS1301 insertions at different positions within the gene. Furthermore, the IS1301 insertion exhibited an opposite direction at the two insertion sites, which proves the presence of two independent insertion events. Two variants harboured deletions in oatWY. A functional oatWY gene was present in most serogroup Y isolates, whereas the gene was knocked out in most serogroup W-135 isolates. These data correlated with the above-mentioned observation that serogroup Y isolates are usually O-acetylated, whereas serogroup W-135 isolates lack capsule O-acetylation (Longworth et al., 2002). Genotypic variants were linked to clonal lineages: six out of seven ST-23 isolates (serogroup Y) harboured the intact oatWY, in one isolate a point mutation within oatWY resulted in a premature stop codon; five out of six ST-11 strains exhibited a deletion in oatWY, whereas an IS1301 insertion was found in eight out of eight ST-22 isolates (serogroup W-135).

Table 3. Serogroup W-135 and Y meningococci and their characteristics.
  • a. 

    Numbering of the strain collection at the Institute of Hygiene and Microbiology, Würzburg, Germany.

  • b. 

    NRZM, German National Reference Centre for Meningococci; ATCC, American Type Culture Collection; MRU, Meningococcal Reference Unit, Manchester, UK.

  • c. 

    Resulting in a premature stop codon.

W-1352140251635MRU1997UK 11 11493 bp deletion
2141252316MRU,1996UK 11 11493 bp deletion
2525Z6351M. Achtman1995Mali 11 11493 bp deletion
 171M-603ATCC35559 USA 11 11Deleted
 304770276J. van Putten   22 22IS1301 insertion
2139251634MRU1997UK 22 22IS1301 insertion
2143253145MRU1996UK 22 22IS1301 insertion
2138250690MRU1997UK 22 22IS1301 insertion
2234  3192NRZM1991GER 22 22IS1301 insertion
2236  3798NRZM1994GER 22 22IS1301 insertion
2237  3799NRZM1994GER 22 22IS1301 insertion
2142252937MRU1996UK184 22IS1301 insertion
2238  3811NRZM1994GER1731 bp deletion
2239  3851NRZM1994GER1741 bp deletion
2232  3047NRZM1990GER185Functional+
2233  3048NRZM1990GER185Functional+
2235  3704NRZM1994GER189334493 bp deletion
Y 172M-112ATCC35561 USA166 11Functional+
2225  3651NRZM1994GER 11 11493 bp deletion
 325 Own collection1989GER 23 23Functional+
2219  3162NRZM1991GER 23 23Functional+
2222  3461NRZM1992GER 23 23Functional+
2223  3496NRZM1993GER 23 23Functional+
2224  3637NRZM1994GER 23 23Functional+
2226  3708NRZM1994GER 23 23Point mutationc
2227  3715NRZM1994GER183 23Functional+
2228  3795NRZM1994GER167Functional+
2221  3258NRZM1992GER168IS1301 insertion
 637 28506Own collection GER171 92Functional+
2220  3245NRZM1992GER172 92Functional+
2522Z3580M. Achtman1983Gambia175IS1301 insertion
2523Z5812M. Achtman1983Gambia175IS1301 insertion
2229  3849NRZM1994GER186 44Functional+
Figure 4.

Variants of the oatWY region in serogroup W-135 and Y meningococci. The oatWY gene is demonstrated by an arrow. Two de-O-acetylated variants harbour insertions of IS1301 at different sites and in opposite directions. Another two de-O-acetylated variants harbour deletions denoted by different shadings of the horizontal bars. The most abundant sequence types (STs) and serogroups (Sg) are given in bold letters.


In this study, the capsular O-acetyltransferase genes of serogroup C, W-135 and Y meningococci, i.e. oatC and oatWY, respectively, were identified downstream of siaD within the capsule synthesis region A of the cps locus. The A+T content of these genes (≈ 70%) was significantly higher than the A+T content of the whole genome and comparable to that of the sia genes, indicating that oatC and oatWY have the same origin as the sia genes. It could be shown by RT-PCR that the O-acetyltransferase genes were co-transcribed with the sia genes. We therefore propose that the oat genes belong to region A. A similar physical linkage of O-acetyltransferase genes and capsule biosynthesis genes has been shown in S. aureus and S. pneumoniae (Sau et al., 1997).

In serogroup W-135 and Y meningococci, identical O-acetyltransferase genes were found, in spite of their different capsular composition. The capsular polysaccharides are heteropolymers composed of α-2-6-linked sialic acid and galactose in serogroup W-135 meningococci and α-2-6-linked sialic acid and glucose in serogroup Y meningococci. Identical positions of the sialic acid residues are O-acetylated (Lemercinier and Jones, 1996). For the most likely assumption that the O-acetyltransferase acts not on activated sialic acid monomers but on polymeric sugars, this finding suggests a promiscuous mechanism of substrate recognition with regard to the alternate sugar residues, i.e. galactose in serogroup W-135 and glucose in serogroup Y. The encoded protein OatWY shared homology with the NodL–LacA–CysE family of O-acetyltransferase enzymes. Members of this family O-acetylate a variety of substrates ranging from antibiotics to bacterial lipooligosaccharides and capsular polysaccharides (Downie, 1989; Allignet and el Solh, 1995). The group shares amino acid homology at the carboxy-terminus of the protein, which includes the putative active site of the protein. The O-acetyltransferase OatC of serogroup C meningococci shared no homology with OatWY or any other known protein. This might not be surprising because several O-acetyltransferases are known neither to have homologies to the NodL–LacA–CysE family of cytoplasmic O-acetyltransferases nor to be related to the family of inner membrane transacylases.

As shown by NMR technology, the sialic acid residues of the serogroup C, W-135 and Y capsular polysaccharides are O-acetylated at two positions, either at position 7 or 8 in serogroup C or at position 7 or 9 in serogroups W-135 and Y (Jennings et al., 1977; Lemercinier and Jones, 1996). However, only one O-acetyltransferase gene was identified in each serogroup. Although no information is available on the genetics of capsule O-acetylation in E. coli K-1, which expresses a related capsular polysaccharide composed of α-2-8-linked sialic acid, the acetyl-coenzyme A:polysialosyl O-acetyltransferase has already been partially purified from O-acetylated isolates (Higa and Varki, 1988). In vitro analyses of O-acetylation revealed that the majority of the polysaccharide is O-acetylated at position 9 of the sialic acid, whereas only 25% was 7-O-acetylated. A migration of O-acetyl groups from position 7 to position 9 was observed during the experiment. O-acetyl group migration from position 7 to position 9 has also been observed in serogroup W-135 polysaccharide (Lemercinier and Jones, 1996). Furthermore, stability studies of meningococcal serogroup C conjugate vaccines revealed acetyl group migration from position 8 to position 7 (Ravenscroft et al., 1999; Ho et al., 2001). It can be assumed that the O-acetyltransferases encoded by oatC and oatWY mediate O-acetylation at different positions of the sialic acid residue, at position 8 in serogroup C meningococci and at position 7 in serogroup W-135 and Y meningococci. This step might be followed by a non-enzymatic isomerization to position 7 or position 9 respectively. In vitro studies using purified enzymes are needed to analyse the enzymatic specificities.

As mentioned above, the prevalence of capsule O-acetylation in serogroup C, W-135 and Y meningococci has been estimated (Borrow et al., 2000; Longworth et al., 2002). The regulatory mechanisms of O-acetylation of extracellular polysaccharides are as yet unknown in most cases. E. coli K-1 isolates have a prevalent expression of O-acetylated or de-O-acetylated capsular polysaccharide, showing a high reversion rate (1–3 × 10−2) to the opposite phenotype (Orskov et al., 1979). No information is available on the mechanism of phase variation. By in vitro analysis, we could prove in this study that slipped-strand mispairing is the cause of phase variation of O-acetylation in serogroup C meningococci. This kind of regulation of gene expression is reminiscent of frameshifts in other mostly virulence-associated genes of N. meningitidis (Sarkari et al., 1994; van der Ende et al., 1995; Hammerschmidt et al., 1996a; Jennings et al., 1998; 1999). Therefore, we provided new experimental evidence for phase variation by slipped-strand mispairing of a neisserial gene.

O-acetylation status in serogroup W-135 and Y meningococci correlated with clonal lineages. The ST-23 complex is the most abundant lineage associated with serogroup Y. Most strains of this lineage harboured an intact copy of oatWY. The ST-22 complex and the ST-11 complex are linked to serogroup W-135. In those lineages, oatWY was either deleted or rendered non-functional by the insertion of IS1301. Thus, the recently observed close associations of serogroups W-135 and Y with de-O-acetylation and O-acetylation, respectively (Longworth et al., 2002), were at least partially the result of clonal expansion. Whether the O-acetylation status contributes to the fitness of the lineages, or whether it simply represents a random and clonal association of siaD genes and variants of the oatWY gene, is unclear. Epidemiological evidence for either hypothesis might be obtained by longitudinal analyses of global strain collections of ST-22 and ST-23 isolates.

It is unclear whether widespread use of either O-acetylated or de-O-acetylated conjugate vaccines will favour the emergence of either de-O-acetylated or O-acetylated serogroup C strains as a consequence of immune evasion. Before the introduction of the meningococcal serogroup C conjugate vaccines in the UK (Miller et al., 2001), ≈ 12% of disease-causing serogroup C isolates from 1998 (14/111 isolates) and 1999 (20/164 isolates) were de-O-acetylated (Borrow et al., 2000). Two of the vaccines licensed in the UK were based on O-acetylated serogroup C polysaccharide and one on de-O-acetylated polysaccharide. After vaccination, 21.95% (18/32) of serogroup C case isolates were de-O-acetylated in 2000, 27.9% (12/43) in 2001 and 15.4% (2/13) in 2002 (Balmer et al., 2002). Therefore, because of the natural fluctuations in the level of de-O-acetylated isolates and the reduced numbers of serogroup C isolates after vaccination, the O-acetylation status of disease-causing serogroup C isolates in the UK does not appear to have been influenced since the introduction of vaccines. Furthermore, bactericidal assays using isogenic O-acetylation mutants of the serogroup C strain 2120 and sera obtained from vaccinees who received either O-acetylated or de-O-acetylated meningococcal serogroup C polysaccharide conjugate vaccines did not provide evidence for a protection by O-acetylation from bactericidal antibodies (R. Borrow, H. Claus and U. Vogel, unpublished observation). However, these results are still preliminary and will require further experiments.

In conclusion, we identified the genes responsible for O-acetylation of meningococcal capsular polysaccharides composed of sialic acid. The regulation and population biology of the genes have been defined. Future work is now possible to analyse the role of capsule O-acetylation for immunogenicity and pathogenicity. The sequence information provided in this report may also support the in silico identification of other bacterial or mammalian sialic acid O-acetyltransferases.

Experimental procedures

Strains, culture media and antibodies

Meningococcal strains used in this study are listed in Tables 3 and 4. All strains were grown at 37°C either in an atmosphere of 5% CO2 on GC agar (BD Difco™) or in proteose peptone broth (Difco). Both media were supplemented with PolyViteX (bioMerieux). E. coli strains DH5α (Hanahan, 1983) and XL1-blue (Stratagene) were used for propagation of recombinant DNA constructs and were grown at 37°C on Luria–Bertani agar (Invitrogen). When appropriate, the following antibiotics were used: ampicillin (100 µg ml−1 for E. coli), erythromycin (30 µg ml−1 for E. coli, 7 µg ml−1 for meningococci) and kanamycin (30 µg ml−1 for E. coli, 100 µg ml−1 for meningococci). The monoclonal antibodies (mAbs) used for ELISA and dot blot are listed in Table 1 and were kindly provided by K. E. Stein, Food and Drug Administration, Bethesda, MD, USA (serogroup C), and P. Fernsten, Wyeth Research, West Henrietta, NY, USA (serogroups W-135 and Y). The specificity of these antibodies has been documented by Rubinstein and Stein (1988) and Longworth et al. (2002).

Table 4. .N. meningitidis strains used for functional analyses of oatC and oatWY.
GenotypePlasmids used
for transformation
  • a. 

    As determined by ELISA with specific mAbs (see Table 1).

2120C 11 oatC O-ac
2948C 11 oatC knock-outpHC15Kande-O-ac
2950C 11 oatC knock-out, complemented in transpHC15, pHC16Kan, EryO-ac
2144W-135169 oatWY O-ac
3149W-135169 oatWY knock-outpHC21Kande-O-ac
3150W-135169 oatWY knock-out, complemented in transpHC21, pHC23Kan, EryO-ac
 172Y166 oatWY O-ac
3073Y166 oatWY knock-outpHC21Kande-O-ac
3138Y166 oatWY knock-out, complemented in transpHC21, pHC23Kan, EryO-ac


Internal fragments of seven housekeeping genes (abcZ, adk, aroE, fumC, gdh, pdhC and pgm) were amplified and sequenced on both strands according to the instructions on the Neisseria MLST website (, which is located and managed at the Peter Medawar Building for Pathogen Research, University of Oxford, UK (Maiden et al., 1998). Assignments of new alleles, sequence types and clonal complexes were done by the curator of the website.

Dot blot

The dot blot for detection of capsule O-acetylation in serogroups W-135 and Y was performed as described by Longworth et al. (2002).


The detection of capsule O-acetylation in serogroup C was performed by ELISA as described by Vogel et al. (1998). The monoclonal antibodies 2016.3 and 2055.5 were diluted 1:9000.

Selection of de-O-acetylated serogroup C meningococci

O-acetylated serogroup C meningococci (107 cfu) were incubated with 1 µl of the bactericidal mAb 2055.5 directed against the capsular O-acetyl groups and 20% normal human serum in veronal-buffered saline with 0.5% bovine serum albumin for 30 min at 37°C (Vogel et al., 1997). Subsequently, the bacteria were plated in 10-fold dilutions on GC agar and incubated at 37°C in an atmosphere of 5% CO2 overnight. De-O-acetylated serogroup C meningococci were detected as non-reacting colonies with mAb 2055.5 by colony blotting as described recently (Hammerschmidt et al., 1996b).

Recombinant DNA techniques

Restriction enzymes and DNA-modifying enzymes were purchased from New England Biolabs. Chromosomal DNA of Neisseria meningitidis was purified with the Qiagen® Genomic-tip system according to the manufacturer's instruction. Southern blot hybridizations were performed as described previously with digoxigenin-labelled probes (Hilse et al., 1996). Recombinant plasmids were isolated with the QIAprep Spin miniprep kit (Qiagen). Transformation of meningococci was performed as described previously (Frosch et al., 1990). Oligonucleotides were purchased from Sigma-ARK and are listed in Table 5. PCR was performed on a thermal cycler obtained from Biometra. The thermostable DNA polymerase AmpliTaq was purchased from Applied Biosystems. Automated DNA sequencing was performed on an Applied Biosystems model 377 using the dye terminator cycle method with AmpliTaq. Nucleotide sequence data were analysed with the lasergene sequence analysis software. DNA and protein sequences were compared with the GenBank and swissprot databases on the blast server hosted by the National Center for Biotechnology Information (NCBI, Bethesda, MD, USA). The transmembrane prediction program tmhmm was used for the determination of transmembrane regions.

Table 5. Oligonucleotides used in this study.
TargetPosition as referred
to accession no.
  • a. 

    Number of nucleotides downstream of siaD in strain 172.

  • Restriction sites used for subsequent cloning of the PCR products are underlined.

HC51 ATGTCAATCAATACGTTTG oatC   29–47 AJ243686 This study
HC73 GCATGATTTTAGAATTGCAGAG oatC downstream region1235–1214 L09188 Hammerschmidt et al. (1994)
HC53 CATTACTCCTTATCATCCTG siaD (serogroups2665–2684 Y13969 Claus et al. (1997)
W-135 and Y)1286–1267a Y13969 Claus et al. (1997)
region (serogroups
W-135 and Y)
HC68 GATAAAGATAAAATGATTTGG siaD (serogroup C)1411–1432 U75650 Swartley et al. (1997)
W-135 and Y)
3041–3059 Y13969 Claus et al. (1997)
HC358 GAGCGACAATCCTGTGTCTG oatWY 4206–4187 Y13969 Claus et al. (1997)
HC285 gcgcgcGGATCCGCTGGAGTCTTATTCATCATG oatC   11–31 AJ243686 This study
HC286 gcgcgcCTGCAGACGCTTTTTTGATTGGATTTAAC oatC downstream1407–18 nt AJ243686 This study
HC368 gcgcgcGCATGCTCATAGGAACTCACATGTATTC oatWY 3332–3353 Y13969 Claus et al. (1997)
HC342 gcgcgcCTGCAGTTGATGATTTTGCTTCCTAAGC oatWY 4010–3989 Y13969 Claus et al. (1997)
HC47 CAGATTTCCCGTGACG hrtA  356–371 AJ000250 Claus et al. (1998)
HC48 TGCCGGCAAACTGCTGCG hrtA  865–848 AJ000250 Claus et al. (1998)
GH5 GGGGCGCTGAGGTCTGpUC4K 430–445 X06404 Taylor and Rose (1988)
GH6 ATATGGGTATAAATGGGCTCpUC4K1467–1448 X06404 Taylor and Rose (1988)
HC379 TGATTGATTCGATACACCATACsiaD (serogroup C)1049–1070 U75650 Swartley et al. (1997)
HC381 CGCATATCAGGATTGGAATAG oatC 1389–1369 AJ243686 This study
W-135 and Y)
2661–2681 Y13969 Claus et al. (1997)
HC384 GGTCATCATACATAGTTGAATG oatWY 3955–3934 Y13969 Claus et al. (1997)


Total RNA from strains 2120 (serogroup C), 2144 (serogroup W-135) and 172 (serogroup Y) was isolated with the RNeasy Midi kit (Qiagen) according to the manufacturer's instructions followed by DNase digestion. A siaD- (HC379 for serogroup C and HC382 for serogroups W-135 and Y) and an oat-specific primer (HC381 for serogroup C and HC384 for serogroups W-135 and Y) were used to amplify a transcript covering both the siaD and the oat gene for each RT-PCR using the Qiagen OneStep RT-PCR system. The reverse transcription was performed at 50°C for 30 min until the reverse transcriptase was inactivated at 95°C for 15 min. Simultaneously, by incubation at 95°C, the cDNA template was denatured, and the HotStartTaq DNA polymerase was activated. The parameters for the subsequent PCR were as follows: denaturation at 94°C for 1 min, annealing at 55°C for 1 min and extension at 72°C for 2 min. This cycle was repeated 35 times. In a standard PCR using the same primers as in the RT-PCR, the correct fragment length was controlled using chromosomal DNA as a template, and the absence of DNA was verified by a negative result using the RNA preparation as a template.

Insertional inactivation and complementation of the O-acetyltransferase genes

The PCR products HC68/HC73 (serogroup C) and HC357/HC358 (siaD downstream region of serogroups W-135 and Y) were cloned into the PCR product cloning vector pCR-ScriptTM SK(+) (Stratagene), resulting in the plasmids pHC13 and pHC20 respectively (Table 6). The kanamycin resistance cassette of pUC4K (Amersham Biosciences) was inserted into the EcoRI site of oatC (pHC15) or into the blunted StyI site of oatWY (pHC21) respectively. Finally, the meningococcal wild-type strains were transformed with the appropriate plasmids. For complementation of the serogroup C O-acetyltransferase gene, the PCR product HC285/HC286 (oatC) was ligated between the BamHI and PstI sites of the vector pEWI-1 under the control of the opa promoter upstream of an erythromycin resistance cassette within the hrtA locus (Claus et al., 1998). The hrtA locus (NMB1875–NMB1878) was chosen as the chromosomal insertion site in order to take advantage of the high number of transformants achieved even in strains with a low level of in vitro competence. Knock-out mutation of NMB1877 did not affect the growth rate of the bacteria, serum resistance or virulence in the infant rat model of meningococcal infection (U. Vogel and H. Claus, unpublished observation). Subsequently, the resulting plasmid pHC16 was transformed into the serogroup C oatC mutant (strain 2948). For complementation of the serogroup W-135/Y O-acetyltransferase genes, the PCR product HC368/HC342 (oatWY) was cloned between the SphI and PstI sites of the plasmid pMD33.1 under the control of the opc promoter resulting in pHC22. Both the promoter and the oatWY gene were excised by BamHI–EcoRV restriction, blunted by T4 DNA polymerase and ligated into the StuI site of pMD34 downstream of an erythromycin resistance cassette within the hrtA locus. Finally, the resulting plasmid pHC23 was transformed into the serogroup W-135 and Y mutants (strains 3149 and 3073 respectively). Correct insertion of recombinant DNA was controlled for all mutants by PCR and Southern blot. For Southern blot hybridizations, chromosomal DNA of wild-type and mutant strains was restricted with XmnI for serogroup C and with StyI for serogroups W-135 and Y and probed with the respective oat, kan and hrtA (data not shown). Additionally, the mutants were analysed by ELISA.

Table 6. Plasmids used in this study.
pCRScript™ Amp SK(+) pBluescript II SKPCR product cloning vector Cloning vectorStratagene
pUC4KVector containing an aminoglycoside 3′-phosphotransferase gene conferring resistance to kanamycinAmersham Biosciences
pUC18Cloning vectorAmersham Biosciences
pEWI-1 hrtA cloned into pUC18 is disrupted by an omega fragment-flanked insertion, which is composed of an opa promoter, a multiple cloning site and an ermC cassetteE. Wintermeyer and M. Frosch
pMD33.1An opc promoter is inserted into pBluescript II SK. The opc promoter and a cloned gene of interest are excised from this vector and inserted into pMD34 for expression (see below)E. Wintermeyer and M. Frosch
pMD34 hrtA cloned into pUC18 is disrupted by an omega fragment-flanked insertion, which is composed of a multiple cloning site and an ermC cassetteE. Wintermeyer and M. Frosch
pHC13PCR product HC68/HC73 comprising oatC cloned into pCRScript™ Amp SK(+)This study
pHC15Insertion of the kanamycin resistance cassette of pUC4K into the EcoRI site of oatC in pHC13This study
pHC16PCR product HC285/HC286 comprising oatC cloned between the BamHI and the PstI sites of pEWI-1This study
pHC20PCR product HC357/HC358 comprising oatWY cloned into pCRScript™ Amp SK(+)This study
pHC21Insertion of the kanamycin resistance cassette of pUC4K into the StyI site of oatWY in pHC20This study
pHC22PCR product HC368/HC342 comprising oatWY cloned between the SphI and the PstI sites of pMD33-1This study
pHC23Restriction fragment BamHI–EcoRV of pHC22 cloned into the StuI site of pMD34This study


This work was supported by the priority programme (SPP) 1047 (grant VO718/3-4 to U.V. and M.F.) and by the Sonderforschungsbereich (SFB) 479 (project B2 to M.F. and U.V.) of the Deutsche Forschungsgemeinschaft. Phil Fernsten (Wyeth Research, West Henrietta, NY, USA) and Kathryn E. Stein (Food and Drug Administration, Bethesda, MD, USA) are gratefully acknowledged for providing monoclonal antibodies essential to this research. Eva Wintermeyer is thanked for providing plasmids pEWI-1, pMD33.1 and pMD34. We are indebted to Rita Gerardy-Schahn and Oliver Kurzai for helpful discussions.