Short sequence-paper

Genetic analysis of the gene cluster for the synthesis of serotypea-speci¢c polysaccharide antigen in Actinobacillus

actinomycetemcomitans1

Nao Suzuki, Yoshio Nakano *, Yasuo Yoshida, Hiroshi Nakao, Yoshihisa Yamashita,Toshihiko Koga

Department of Preventive Dentistry, Kyushu University Faculty of Dental Science, Fukuoka 812-8582, Japan

Received 1 August 2000; received in revised form 4 September 2000; accepted 5 September 2000

Abstract

The serotype a-specific polysaccharide antigen (SPA) of Actinobacillus actinomycetemcomitans consists of 6-deoxy-D-talose. A gene clusterassociated with the biosynthesis of SPA was cloned and sequenced from the chromosomal DNA of A. actinomycetemcomitans SUNYaB 75(serotype a). This cluster consisted of 14 open reading frames. Insertional inactivation of eight genes in this cluster resulted in loss ofthe ability of A. actinomycetemcomitans SUNYaB 75 cells to produce the polysaccharide. A protein database search revealed that the11 sequential genes containing these eight genes were not found in SPA-associated gene clusters of the other serotypes ofA. actinomycetemcomitans. These results suggest that the gene cluster is unique to serotype a and is essential to the synthesis of theSPA. ß 2000 Elsevier Science B.V. All rights reserved.

Keywords: Serotype antigen; Polysaccharide; Actinobacillus actinomycetemcomitans

Actinobacillus actinomycetemcomitans is a non-motile,Gram-negative, capnophilic, fermentative coccobacillusthat has been implicated in the etiology and pathogenesisof localized juvenile periodontitis [1^3], adult periodontitis[4], and severe non-oral human infections [5]. A. actino-mycetemcomitans strains can be divided into ¢ve serotypes(a, b, c, d, and e) [6^8]. The serologic speci¢city is de¢nedby the polysaccharides on the surface of the organism [9]and the serotype speci¢c polysaccharide antigens (SPAs)are the immunodominant antigens in the organism [10^14]. The serotype a antigen has a unique structural feature,consisting solely of a 6-deoxyhexose, 6-deoxy-D-talose [15].Bacterial extracellular polysaccharides consisting solely of6-deoxytalose are rare, and no gene involved in the syn-thesis of 6-deoxy-D-talan has been reported. Structuralanalysis of serotype a antigen indicated that the polymerwas composed of a repeating unit: -3)-6-deoxy-K-D-talose-(1-2)-6-deoxy-K-D-talose-(1- [15,16].

Previously, we reported the gene clusters responsible for

the synthesis of SPAs of A. actinomycetemcomitans sero-types b, c, d, and e, and that the gene clusters are locatedin the same locus in these strains [17^20]. In contrast tothese serotypes, no genes responsible for SPA were foundaround this locus. In addition, the biosynthetic pathwayfor GDP-6-deoxy-D-talose, which is the activated nucleo-tide sugar form of 6-deoxy-D-talose, is predicted to bequite di¡erent from the pathways for the precursors ofserotype b-, c-, d- and e-speci¢c polysaccharide antigens[21]. Therefore, the genes speci¢c to serotype a cannot beisolated by hybridization with SPA genes from other sero-type strains. In this study, we constructed SPA-de¢cientmutants by transposon mutagenesis, and isolated the genecluster responsible for the biosynthesis of serotype a-spe-ci¢c polysaccharide antigen.

To detect SPA-de¢cient mutants of A. actinomycetem-comitans SUNYaB 75 (serotype a), transposon IS903Pkanwas used [22]. A. actinomycetemcomitans SUNYaB 75Rifwas isolated as a spontaneous rifampicin-resistant (75Wg/ml) mutant of strain SUNYaB 75, and used as a recip-ient strain. Escherichia coli KD59 was used as a donor inmatings with strain SUNYaB 75Rif. Conjugal transferwas performed by the method of Thomson et al. [22].SPA-de¢cient mutants were screened from a library of

0167-4781 / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved.PII: S 0 1 6 7 - 4 7 8 1 ( 0 0 ) 0 0 2 2 9 - 3

* Corresponding author. Fax: +81-92-642-6354;E-mail : yosh@dent.kyushu-u.ac.jp

1 The nucleotide sequence data reported in this paper have been sub-mitted to the DDBJ with accession number AB046360.

BBAEXP 91479 6-12-00

Biochimica et Biophysica Acta 1517 (2000) 135^138

www.elsevier.com/locate/bba

ca. 200 kanamycin-resistant transformants by colony im-munoblotting with rabbit antiserum directed against wholecells of A. actinomycetemcomitans ATCC 29523 (serotypea) [15]. One colony was an SPA-de¢cient derivative ofSUNYaB 75 strain, and the mutant strain was designatedSN-T1. Western blotting with rabbit antiserum againstwhole cells of A. actinomycetemcomitans ATCC 29523 (se-rotype a) showed that this mutant did not produce a highmolecular mass smear (data not shown). Southern blothybridization analysis with the digoxigenin-labeled kangene in IS903Pkan as a probe showed a 5-kb HindIII frag-ment hybridized from A. actinomycetemcomitans SN-T1chromosomal DNA. This HindIII fragment extractedfrom an agarose gel was cloned into plasmid vectorpMCL210 [23] and the resultant plasmid was designatedpSAA201 (Fig. 1). DNA sequencing of pSAA201 revealedthat the transposon was inserted in a mannosyltransferase-homologue gene.

Three DNA fragments around this position were clonedfrom A. actinomycetemcomitans SUNYaB 75 by chromo-some walking with the probes indicated in Fig. 1. A probefor the mannosyltransferase-homologue gene (mt probe)was ampli¢ed with the following primers: forward primer,5P-CCATTAGTATTAGTTGGTGATCAGGGAT-3P ; re-verse primer, 5P-CCTCAGTTGAATGGGGATTGAC-TAAAAC-3P. A probe for the putative ABC transporter,

abc probe, was ampli¢ed (forward primer, 5P-CAAC-TAATGTCGCGGTTGGTTCAA-3P ; reverse primer, 5P-ACTTTCCGAGGTTCGGTTAAAACG-3P).

DNA sequencing templates were prepared using theGPS-1 Genome Priming System (New England Biolabs)according to the manufacturer's instructions. E. coliDH5K was used for chemical transformation and thetransformants were plated onto 2UTY medium with ka-namycin (25 Wg/ml) and chloramphenicol (20 mg/ml). Thenucleotide sequence was determined by the dideoxy chaintermination technique of Sanger et al. [24] with a BigDyeTerminator Cycle Sequencing kit, FS and an ABI 373ADNA sequencer or an ABI PRISM 310 (Perkin-Elmer,Applied Biosystems Division).

The nucleotide sequences of pSAA211, pSAA212 andpSAA213 revealed that the 15.9-kb EcoRI^HindIII frag-ment contained a gene cluster of 14 open reading frames(ORFs). All the ORFs started with the ATG codon. Thespacing between the ORFs varied between 33 (i.e. threeoverlapping bases) and 89 bp, while the spaces betweenORF1 and ORF2 (185 bp) and between ORF13 andORF14 (181 bp) were relatively large. A computer-aidedsearch for promoters revealed a promoter-like motif ofc70-like principal sigma factor between ORF1 andORF2. ORF1 through ORF12 had the same orientation,while ORF13 and ORF14 were on the opposite strand.

Fig. 1. Restriction map and genetic organization of the gene cluster responsible for production of the SPA of A. actinomycetemcomitans SUNYaB 75(serotype a). Closed arrows indicate ORFs. The functions of the gene products predicted by homology search, the G+C content of each ORF, and theSPA phenotypes caused by speci¢c insertion mutants are shown in descending order below the restriction map. A £ag indicates the putative promoter.The horizontal lines show the DNA fragments inserted into pMCL210 used for nucleotide sequencing or preparing the autoclaved extracts. Restrictionenzyme abbreviations: H, HindIII; E, EcoRI; A, Acc65I; Pa, PacI; Pm, PmeI; Tld, a putative GDP-4-keto-6-deoxy-D-mannose reductase; Ac-TRase,acetyltransferase; Gmd, GDP-mannose dehydratase; XylR, xylose operon regulatory protein.

BBAEXP 91479 6-12-00

N. Suzuki et al. / Biochimica et Biophysica Acta 1517 (2000) 135^138136

The G+C contents (25.1^37.8%) of ORFs 2^12 werelower than the average G+C content (47.6%) of the genes£anking them. The genes encoding basic cellular functionsin A. actinomycetemcomitans are reported to have an aver-age G+C content of 48% [25]. These ¢ndings stronglysuggest the interspeci¢c transfer of these genes from otherspecies with low G+C content to A. actinomycetemcomi-tans [26].

A protein database search revealed that the products of11 genes, ORF2^ORF12, were homologous to bacterialgene products involved in the biosynthesis of extracellularpolysaccharides (Fig. 1). ORF2 had homology to themanC gene in E. coli [27]. The manC product is a man-nose-1-phosphate guanylyltransferase, which convertsGTP and mannose-1-phosphate to GDP-mannose. Theproteins encoded by ORF3 and ORF4 showed high iden-tities (64% and 73%, respectively) to ABC transport pro-teins encoded by ORFs in the clusters responsible for syn-thesis of the SPAs in other serotypes of A. actinomycetem-comitans. ORF5, ORF6, ORF10, and ORF11 had 25^45%homologies to the genes that encoded mannosyltransferaseof E. coli [28,29], Brucella meltensis [30], and Klebsiellapneumoniae [31]. The ORF7 product shared 28% identitywith the rmd product of Pseudomonas aeruginosa [32]. Thermd product is a reductase that reduces GDP-4-keto-6-deoxy-D-mannose to GDP-D-rhamnose. We suspect thatthe ORF7 gene encodes GDP-4-keto-6-deoxy-D-mannosereductase. Characterization of the gene product is in pro-gress. ORF8 shared 24% identity with the acetyltransferasegene from Salmonella typhimurium [33]. Since the O-2 po-sition of 1,3-linked 6-deoxy-D-talose in the SPA is acety-lated [16], the ORF8 product may participate in the ace-tylation of the polysaccharide. The ORF9 product hadhomology to GDP-mannose 4,6-dehydratase in Yersiniapseudotuberculosis (GenBank, accession no. AJ251712),which produces GDP-6-deoxy-D-lyxo-4-hexulose fromGDP-mannose. The ORF12 product was related to thegalactosyltransferase of Salmonella enterica, whose O-anti-gen biosynthesis is GlcNAc-transferase dependent [34].ORF13 had 71% homology to the gene encoding the xy-lose operon regulatory protein of E. coli (Protein Identi-¢cation Resource, S47790). Analysis of the nucleotide se-quence in the A. actinomycetemcomitans genome project atthe University of Oklahoma's Advanced Center for Ge-nome Technology (http://www.genome.ou.edu/act.html)revealed that the ORF13-homologue gene of serotype bwas £anked by the ORF1-homologue gene, and ORF2^ORF12 were absent from this region.

To determine which regions are essential for the biosyn-thesis of serotype a-speci¢c polysaccharide in A. actino-mycetemcomitans SUNYaB 75, the ORFs in this clusterwere inactivated by insertion of a kanamycin-resistancegene. Insertional mutagenesis of 13 genes (Fig. 1) wasperformed by homologous recombination of a kanamy-cin-resistance gene from pGPS1.1 into the respective chro-mosomal locus. Linearized plasmids containing genes dis-

rupted by insertion of the kanamycin-resistance gene wereintroduced by electroporation into A. actinomycetemcomi-tans SUNYaB 75 resuspended in ice-cold 10% glycerol[35]. The electroporation was performed in a Gene Pulserapparatus (Bio-Rad) set to 12.5 kV/cm, and the cells wereplated onto THY (Todd-Hewitt supplemented with 1.0%yeast extract)-kanamycin (25 Wg/ml) medium.

Cell suspensions of A. actinomycetemcomitans SUNYaB75, mutant strains derived from SUNYaB 75 and E. coliDH5K transformant in phosphate-bu¡ered saline (0.12 MNaCl, 0.01 M Na2HPO4, 5 mM KH2PO4, pH 7.5) wereautoclaved at 121³C for 20 min and centrifuged at 4³C.After centrifugation, the supernatants were collected andused as autoclaved extracts. SPA was puri¢ed from theautoclaved extract of A. actinomycetemcomitans ATCC29523 (serotype a) as described by Amano et al. [13]. Im-munodi¡usion analysis was carried out in 1.0% agarose(Gibco BRL) in phosphate-bu¡ered saline.

Insertional inactivation of ORF2, ORF3, and ORF7through ORF12, but not of ORF1, ORF4, ORF5,ORF6, and ORF13, resulted in loss of the ability of A.actinomycetemcomitans SUNYaB 75 to produce SPA (Fig.2). This cluster has four mannosyltransferase-homologuegenes, ORF5, ORF6, ORF10 and ORF11. It seems rea-sonable that only two of the four mannosyltransferase-homologues are essential to produce SPA, because twotalosyltransferases are required to synthesize the repeatingunit of SPA polysaccharide, which contains two D-talosyllinkages. The insertion of ORF4 was at the 3P end site,and the resistant truncated product may have some enzy-matic activity. Based on these results, it is reasonable topredict that the region containing ORF2^ORF12 is indis-pensable for the biosynthesis of SPA of A. actinomycetem-comitans SUNYaB 75. However, further detailed analysis

Fig. 2. Immunodi¡usion reaction of anti-A. actinomycetemcomitansATCC 29523 (serotype a) serum with autoclaved extracts prepared fromA. actinomycetemcomitans SUNYaB 75 (serotype a) and E. coli trans-formants. The center wells contain anti-A. actinomycetemcomitansATCC 29523 serum. The outer wells contain the autoclaved extractsfrom A. actinomycetemcomitans SUNYaB 75 (well 1), SPA puri¢edfrom A. actinomycetemcomitans ATCC 29523 (1 Wg) (well 2), A. actino-mycetemcomitans Y4 (serotype b) (well 3), A. actinomycetemcomitansNCTC 9710 (serotype c) (well 4), A. actinomycetemcomitans IDH 781(serotype d) (well 5), A. actinomycetemcomitans IDH 1705 (serotype e)(well 6), E. coli DH5K containing pMCL210 (well 10), E. coli DH5Kcontaining pSAA214 (well 11), ORF1-de¢cient (well 7), ORF2-de¢cient(well 8), ORF3-de¢cient (well 9), ORF4-de¢cient (well 12), ORF5-de¢-cient (well 13), ORF6-de¢cient (well 14), ORF7-de¢cient (well 15),ORF8-de¢cient (well 16), ORF9-de¢cient (well 17), ORF10-de¢cient(well 18), ORF11-de¢cient (well 19), ORF12-de¢cient (well 20), andORF13-de¢cient (well 21) mutants derived from A. actinomycetemcomi-tans SUNYaB 75.

BBAEXP 91479 6-12-00

N. Suzuki et al. / Biochimica et Biophysica Acta 1517 (2000) 135^138 137

which rules out polar e¡ects may reveal that the numberof essential genes is less than indicated in the above experi-ment.

The EcoRI^PacI fragment of pSAA213, the PacI^PmeIfragment of pSAA212 and the PmeI^HindIII fragment ofpSAA211 were combined and subcloned into pMCL210,and the resultant plasmid was designated pSAA214. Theautoclaved extract from E. coli DH5K containingpSAA214 did not produce any precipitin lines with rabbitantiserum against whole cells of serotype a A. actinomy-cetemcomitans (Fig. 2). E. coli transformants harboringthe gene cluster of biosynthetic genes for serotype-speci¢cpolysaccharide from A. actinomycetemcomitans Y4 (sero-type b) [17], NCTC 9710 (serotype c) [18], or IDH 1705(serotype e) [19] acquired the ability to produce each SPA,but IDH 781 (serotype d) [20] and SUNYaB 75 did not. Inserotypes a and d, other gene products may be requiredfor the production of SPAs in E. coli DH5K, or the genesmay not be expressed in E. coli.

We determined the gene cluster responsible for the bio-synthesis of SPA of A. actinomycetemcomitans in thisstudy. This should help to unravel the di¡erences amongserotypes of this organism.

This work was supported in part by a Grant-in-Aid forthe Encouragement of Young Scientists 11771157 (Y.N.)and Grants-in-Aid Scienti¢c Research (A) 10307054(T.K.), (B) 11470452 (T.K.) and (B) 09470474 (Y.Y.)from the Ministry of Education, Science, Sports and Cul-ture, Tokyo, Japan, and a research grant from the UeharaMemorial Foundation (Y.N.).

References

[1] S. Asikainen, C.-H. Lai, S. Alaluusua, J. Slots, Oral Microbiol. Im-munol. 6 (1991) 115^118.

[2] J.L. Ebersole, D. Cappelli, M.-N. Sandoval, J. Clin. Periodontol. 21(1994) 65^75.

[3] D.H. Meyer, P.M. Fives-Taylor, Trends Microbiol. 5 (1997) 224^228.[4] J. Slots, L. Bragd, M. Wikstro«m, G. Dahlen, J. Clin. Periodontol. 13

(1986) 570^577.[5] A.H. Kaplan, D.J. Weber, E.Z. Oddone, J.R. Perfect, Rev. Infect.

Dis. 11 (1989) 46^63.[6] R. Gmu«r, H. McNabb, T.J.M. van Steenbergen, P. Baehni, A. Mom-

belli, A.J. van Winkelho¡, B. Guggenheim, Oral Microbiol. Immu-nol. 8 (1993) 116^120.

[7] J.J. Zambon, J. Slots, R.J. Genco, Infect. Immun. 41 (1983) 19^27.[8] M. Saarela, S. Asikainen, S. Alaluusua, L. Pyha«la«, C.-H. Lai, H.

Jousimies-Somer, Oral Microbiol. Immunol. 7 (1992) 277^279.

[9] T. Koga, T. Nishihara, K. Amano, T. Takahashi, K. Nakashima, Y.Ishihara and N. Shibuya, in: S. Hamada, S.C. Holt and J.R. McGhee(Eds.), Periodontal Disease: Pathogenesis and Host Immune Re-sponses, Quintessence Publishing Co., Ltd., Tokyo, 1991, pp. 117^127.

[10] H. Lu, J.V. Califano, H.A. Schenkein, J.G. Tew, Infect. Immun. 61(1993) 2400^2407.

[11] R.C. Page, T.J. Sims, L.D. Engel, B.J. Moncla, B. Bainbridge, J.Stray, R.P. Darveau, Infect. Immun. 59 (1991) 3451^3462.

[12] J.V. Califano, H.A. Schenkein, J.G. Tew, Oral Microbiol. Immunol.6 (1991) 228^235.

[13] K. Amano, T. Nishihara, N. Shibuya, T. Noguchi, T. Koga, Infect.Immun. 57 (1989) 2942^2946.

[14] M.E. Wilson, R.E. Schi¡erle, Infect. Immun. 59 (1991) 1544^1551.[15] N. Shibuya, K. Amano, J. Azuma, T. Nishihara, Y. Kitamura, T.

Noguchi, T. Koga, J. Biol. Chem. 266 (1991) 16318^16323.[16] M.B. Perry, L.M. MacLean, J.-R. Brisson, M.E. Wilson, Eur.

J. Biochem. 242 (1996) 682^688.[17] Y. Yoshida, Y. Nakano, Y. Yamashita, T. Koga, Infect. Immun. 66

(1998) 107^114.[18] Y. Nakano, Y. Yoshida, Y. Yamashita, T. Koga, Biochim. Biophys.

Acta 1442 (1998) 409^414.[19] Y. Yoshida, Y. Nakano, N. Suzuki, H. Nakao, Y. Yamashita, T.

Koga, Biochim. Biophys. Acta 1489 (1999) 457^461.[20] Y. Nakano, Y. Yoshida, N. Suzuki, Y. Yamashita, T. Koga, Bio-

chim. Biophys. Acta 1493 (2000) 259^263.[21] V.N. Shibaev, Adv. Carbohydr. Chem. Biochem. 44 (1986) 277^339.[22] V.J. Thomson, M.K. Bhattacharjee, D.H. Fine, K.M. Derbyshire,

D.H. Figurski, J. Biol. Chem. 181 (1999) 7298^7307.[23] Y. Nakano, Y. Yoshida, Y. Yamashita, T. Koga, Gene 162 (1995)

157^158.[24] F. Sanger, S. Nicklen, A.R. Coulson, Proc. Natl. Acad. Sci. USA 74

(1977) 5463^5467.[25] J.B. Kaplan, D.H. Fine, FEMS Microbiol. Lett. 163 (1998) 31^36.[26] L. Wang, L.K. Romana, P.R. Reeves, Genetics 130 (1992) 429^443.[27] G. Stevenson, K. Andrianopoulos, M. Hobbs, P.R. Reeves, J. Bac-

teriol. 178 (1996) 4885^4893.[28] T. Sugiyama, N. Kido, Y. Kato, N. Koide, T. Yoshida, T. Yokochi,

J. Bacteriol. 180 (1998) 2775^2778.[29] J. Drummelsmith, C. Whit¢eld, Mol. Microbiol. 31 (1999) 1321^

1332.[30] F. Godfroid, B. Taminiau, I. Danese, P. Denoel, A. Tibor, V. Weyn-

ants, A. Cloeckaert, J. Godfroid, J.-J. Letesson, Infect. Immun. 66(1998) 5485^5493.

[31] S. Merino, M. Altarriba, L. Izquierdo, M.M. Nogueras, M. Regue,J.M. Tomas, Infect. Immun. 68 (2000) 2435^2440.

[32] H.L. Rocchetta, J.C. Pacan, J.S. Lam, Mol. Microbiol. 29 (1998)1419^1434.

[33] J.M. Slauch, A.A. Lee, M.J. Mahan, J.J. Mekalanos, J. Bacteriol. 178(1996) 5904^5909.

[34] D. Liu, A.M. Haase, L. Lindqvist, A.A. Lindberg, P.R. Reeves,J. Bacteriol. 175 (1993) 3408^3413.

[35] P.K. Sreenivasan, D.J. LeBlanc, L.N. Lee, P. Fives-Tayllor, Infect.Immun. 59 (1991) 4621^4627.

BBAEXP 91479 6-12-00

N. Suzuki et al. / Biochimica et Biophysica Acta 1517 (2000) 135^138138