Vol. 58, No. 6INFECTION AND IMMUNITY, June 1990, p. 1545-15510019-9567/90/061545-07$02.00/0Copyright C 1990, American Society for Microbiology

Pilins from the B Serogroup of Bacteroides nodosus:Characterization, Expression, and Cross-Protection

THOMAS C. ELLEMAN,1* DAVID J. STEWART,2 KENNETH G. FINNEY,' PETER A. HOYNE,1AND COLIN W. WARD'

Divisions ofBiotechnology' and Animal Health,2 Commonwealth Scientific and Industrial Research Organisation,343 Royal Parade, Parkville 3052, Victoria, Australia

Received 30 October 1989/Accepted 22 February 1990

Sequences of pilin genes from four strains of serogroup B of the ovine pathogen Bacteroides nodosus havebeen determined. These sequences permit comparisons of amino acid sequence between pilins from differentsubtypes (B1, B2, B3, B4) of the B serogroup and assessment of intraserogroup variation. Pili of B. nodosusstrains 234 (Bi) and 183 (B2) were produced by Pseudomonas aeruginosa harboring a plasmid-borne B.nodosus pilin gene, and these pili were used in sheep vaccination trials. Pi from strain 183 (B2) were foundto be a senior antigen to pili from strains of other B subtypes, providing protection against footrot infectioncaused by strains of the other B subtypes. Pili of this strain are therefore the most suitable candidate forinclusion in a pilus-based vaccine. Pili of strain 234 from subtype Bi, the reference strain of the B serogroup,provided poor protection against infection with other subtypes.

Ovine footrot infection is restricted to the avascular epi-dermis of the interdigital skin and hoof, and no naturalimmunity is developed after infection. However, protectionis achieved through immunization of sheep with killed cellsof Bacteroides nodosus (8, 10), the causative organism ofovine footrot, and this immunity may be passively trans-ferred from immunized sheep to naive recipients by gammaglobulin (9). Antibodies appear to be able to diffuse from thevascular system to the infected area of the hoof to combatinfection.The pili of B. nodosus are the major host-protective

immunogen in killed, whole-cell vaccines against ovinefootrot (20, 31) and are responsible for the agglutination ofB.nodosus cells with antiserum raised against the organism.The agglutination reaction is the basis of systems for sero-grouping field isolates of B. nodosus (7), and nine sero-groups, designated A to I, have been defined with isolatesfrom Australian sheep (3-5). Purified pili alone are able toconfer protective immunity, through vaccination, against ahomologous strain (20, 34), but immunity does not extend tostrains of other serogroups (33). As a consequence, commer-cial whole-cell vaccines contain strains of B. nodosus fromthe nine serogroups.

Within individual serogroups, different subgroups (sub-types) can be recognized by the different degree of reciprocalcross-agglutination in tube agglutination tests and by posses-sion of both unique and common antigenic determinants, asdemonstrated by cross-absorption studies (3). The ability ofsubtypes to elicit cross-protective immunity through shareddeterminants is of interest to vaccine manufacture becausethe number of components of a multivalent vaccine is limitedby considerations of antigenic competition and economics.Eighteen subtypes have been recognized among the nineserogroups, a prohibitively high number for inclusion in avaccine.

In previous studies to determine the extent to whichserogrouping reflects cross-protection following vaccination,the two subtypes of serogroup A have been shown tomutually cross-protect against infection (3, 35). On the other

* Corresponding author.

hand, subtypes of serogroup H do not cross-protect (unpub-lished data). However, strains were assigned to subtypes ofserogroup H on the basis of low-level mutual cross-aggluti-nation with antisera raised in rabbits (5), but the absence ofcross-agglutination between H subtypes with sheep antiseraand the extent of differences between amino acid sequencesof pilins from H subtypes suggest that these subtypes shouldnot have been classified within a single serogroup (24). Itremains to be shown whether pili from subtypes in otherserogroups elicit the production of sufficient cross-reactiveantibody for serogroup protection.

Serogroup B of B. nodosus is the most abundant and themost diverse serogroup, with four subtypes recognized inisolates from Australian sheep (5). Strains from the differentsubtypes of serogroup B share a major agglutinating anti-genic determinant but differ in their pattern of cross-agglu-tination titers through the presence of other determinantswhich may or may not be shared with strains from othersubtypes (3). In the present study, we investigate the rela-tionship between pilins from different subtypes of serogroupB by sequence determination of pilin genes and throughvaccination studies with pili of these strains.

MATERIALS AND METHODSBacterial organisms and expression vectors. B. nodosus

strains 234 (subtype Bi), 183 (B2), 334 (B2), 235 (B3), and112 (B3) were from the Commonwealth Scientific and Indus-trial Research Organisation Division of Animal Health cul-ture collection, and a B4 isolate, VRS 54 (Veterinary Re-search Station, Glenfield, Sydney; designated VCS1125 bythe Department of Veterinary Clinical Studies, University ofSydney), was kindly provided by P. D. Claxton. A multipil-iate variant of Pseudomonas aeruginosa designated K/2PfS(2) was used in expression studies. DNA was prepared fromB. nodosus grown anaerobically in liquid culture as de-scribed previously (13). Plasmid pBR322 and Escherichiacoli RR1 were used in the initial cloning of large (about 7kilobase pairs [kbp]) DNA fragments carrying the pilin geneof B. nodosus. In construction of pilin expression plasmids,smaller fragments of the gene (about 1 kbp) were isolated byDraI digestion from the pBR322 derivatives, and these

1545

on March 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

1546 ELLEMAN ET AL.

fragments were inserted downstream of the PL promoter atthe HpaI site of plasmid pPL-lambda (Pharmacia, Uppsala,Sweden). E. coli POP2136 (F- endA thi hsdR malT X PRc1857 malPQ [obtained from 0. Raibaud]) was used as a hoststrain containing the c1857 gene for temperature-sensitiverepression of the PL promoter. A BamHI restriction frag-ment cartridge bearing the PL promoter and pilin gene fromthis construction was inserted into a plasmid vector suitablefor replication in P. aeruginosa (17, 27).

Identification of pilin genes and sequence determination.DNA fragments bearing the pilin gene of B. nodosus 183were identified in restriction enzyme digests of chromosomalDNA by using a hybridization probe prepared from a previ-ously isolated pilin gene of serogroup A (12, 15). Followingcharacterization of the pilin gene from strain 183 (B2), ahybridization probe carrying the complete coding sequenceof the pilin gene of this strain was prepared from a 0.8-kbpDraI fragment cloned in M13mp8 (29). This probe was usedon Southern transfers of restriction enzyme-digested chro-mosomal DNA to confirm the presence of a single gene copyin strain 183 (B2) and to identify the fragment of DNAharboring the pilin gene in strains 235 (B3), 112 (B3), andVRS 54 (B4). The same hybridization probes were used inthe identification of bacterial colonies which harbored thepilin genes following transformation of host cells with vec-tors carrying pilin genes. E. coli was made competent fortransformation by treatment with CaCl2 (25), and P. aerug-inosa was made competent by treatment with MgCl2 (17).The sequence of the pilin gene from strain 183 was

determined by cloning DNA fragments into M13mp8 forproduction of single-stranded template to use with a univer-sal primer (29). The sequences of pilin genes from strains235, 112, and 54 were determined by using alkali-denaturedplasmids (23) and primers synthesized on an Applied Bio-systems 381A DNA synthesizer to correspond to commonregions of the B serogroup pilin genes.

Preparation of vaccines. P. aeruginosa expressing B. no-dosus pili from plasmid-encoded genes (17, 18) was grown inbaffle shake flasks at 37°C. Pili detached during culture wereharvested from the culture supernatant by precipitation with0.1 M MgCl2 and then suspended in phosphate-bufferedsaline. Insoluble debris was removed by centrifugation.Polyacrylamide (15%) gel electrophoresis in sodium dodecylsulfate showed a single pilin band (Mr, about 17,000) inpurified preparations. Purified pili were quantified by amodification (22) of the method of Lowry et al. Pili in salinesolution were mixed with an equal volume of Alhydrogel(Superfos A/S, Vedbaek, Denmark) and then emulsified withincomplete Freund adjuvant (Difco Laboratories) at a ratioof 1:2. The vaccine (2 ml, containing 50 ,ug of pili per dose)was administered subcutaneously in two doses, 35 daysapart, in opposite sides of the neck. Following challengewith a pure culture of B. nodosus (strains 234 [Bi], 334 [B2],and 112 [B3]), the feet of all sheep were inspected andfootrot lesions were evaluated (34). Blood samples werecollected at the time of challenge, and the antibody titerswere assayed by agglutination tests with serial dilutions ofserum in microtiter trays (6).

RESULTS

Gene sequences. A hybridization probe prepared from thepilin gene of strain 183 (B2) identified a single pilin gene ineach of strains 183 (B2), 235 (B3), 112 (B3), and VRS 54 (B4).The pilin gene in all strains was present on a 7-kbp fragmentfrom SphI-digested DNA. E. coli harboring plasmids carry-

K -

1 2 3 4 5

FIG. 1. Immunoblot with sheep anti-B. nodosus 183 pili antise-rum following sodium dodecyl sulfate-15% polyacrylamide gelelectrophoresis of cellular proteins. Lanes: 1, B. nodosus 183 pili; 2,E. coli harboring pilin gene of B. nodosus 183; 3, E. coli harboringpilin gene of B. nodosus 235; 4, E. coli harboring pilin gene of B.nodosus VRS 54; 5, E. coli harboring pilin gene of B. nodosus 112.Sizes are indicated to the left (in kilodaltons).

ing the pilin gene (as an SphI fragment in pBR322) produceda protein (Mr, about 18,000) which cross-reacted in immu-noblotting with antiserum raised against pili of strain 183(Fig. 1). By analogy with the more fully characterizedexpression product of E. coli strains which harbor the pilingene of B. nodosus 198 (12, 14, 15), these proteins corre-spond to the encoded prepilins of the gene sequences.The pilin gene sequences from strains of different subtypes

showed extensive similarity (Fig. 2), while those of the B3subtype representatives were identical despite their differentorigins. A single polypeptide of 161 amino acid residues wasencoded by each sequence beginning at ATG362, the mostlikely choice as initiation codon from the ATGATG362sequence by the rules of Stormo et al. (36) and by analogywith genes of other N-methylphenylalanine pilins (11). Theencoded sequences (Fig. 3) had a high degree of similarityboth to each other and to the previously determined (28)amino acid sequence of pilin from strain 234 (Bi). Therelatively few differences found in the comparison of codingsequences were nucleotide substitutions with the singleexception of an insertion-deletion at the encoded pilin'sC-terminus which failed to alter the length of the encodedsequence. A seven-residue leader sequence preceded thephenylalanine residue which corresponded to the N-methyl-ated, amino-terminal residue of mature pilin; this sequencewas identical to the leader sequences of prepilins from otherB. nodosus strains determined to date (11).The proximal 3' downstream sequences (alignment posi-

tions 845 to 972 of Fig. 2) are identical in pilins from strains183 (B2), 235 (B3), 112 (B3), and VRS 54 (B4), but thislimited sequence shows no similarity to the limited publishedsequences of other B. nodosus pilin genes available to dateexcept in the transcription terminator region (11). Thiscontrasts with the near identity of sequence downstream ofpilin genes from strains 198 (Al) and 238 (Gl) or strains 265(Hi) and 340 (D) (12, 19, 21).The upstream sequences, on the other hand, were remark-

INFECT. IMMUN.

on March 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

B. NODOSUS SEROGROUP B PILINS 1547

B2 GCGCACGGTAAAAAGCCGCGGATTTCCGTTTGTCCTTCTGCTAACGCTGCTAATAATAAGGCGCGATGCGACATTGATTTATCGCCGCATATCGTTATTTB3 GCGCACGCTAAAAAGCCGCGGATTTCCGTTTGTCCTTCTGCTAACGTTGCTAATAATAAGGCGCGATGCGACATTGATTTATCGCCGCATATCGTTATTTB4 GCGCACGGTAAAAAGCCACGGATTTCCGTTTGTCCTTCCGCTAACGCTGCTAATAATAAGGCGCGATGCGACATTGATTTATCGCCGCATATCGTTATTT 100

B2 CGCCGGAAAGCGCAGAGACGGGCGCAGTGTGCCATATATTCGTCATCATCATTTCCTTTTATAAAACCAACGTTATTGTGCCGGAAAAATAGTAGAAAAA

B3 CGCCGGAAAGCGCTGAGACGGGCGCGGTGTGCCATATATT CATCATTATTCCTTTTTGTGAAACGGGCGTTATTGTGCCGGAAAAATAGTAGAAAAA

84 CGCCGGAAAGCGCTGAGACGGGCGCGATGTGCCATATATTCGTCATCATCATTTCCTTTTATAAAACCAACGTTATTGTGCCGGAAAAATAGTAGAAAAA 200

B2 ATGCAAAAAAAGCGCGTGTGTCAGAAAATAATTTTTTTAACTAATTGTTTTTAAATATAAAAACAATATTGGCATTGATGACGCATAATGAAAGGCGTCA

B3 ATGCAAAAAAAGCGCGTGCGCCAGAAAAATAATTTTTTAAATTATTGTTTTTAAATATAAAAACAATATTGGCATTGATGACGCATAATGAAAGGCGTCA

84 ATGCAAAAAAAGCGCGTGTGTCAGAAAAATAATTTTTTAACTTATTGTTTTTAAATATAAAAATAATGTTGGCATTGATGACGCATAATGAAAGGCGTCA 300

82 GGCAACTGACTCTAAACAAGATGATATTTAAATGTTCACATTCTTAATAGGAGAATATGATGAAAAGTTTACAAAAAGGTTTCACCTTAATCGMCTCAT

83 GGCAACTGACTCTAAACAAGATGATATTTAAATGTTCACATTCTTAATAGGAGAATATGATGAAAAGTTTACAAAAAGGTTTCACCTTAATCGMCTCAT

B4 GGCAACTGACTCTAAACAAGATGATATTTAAATGTTCACATTCTTAATAGGAGAATATGATGAAAAGTTTACAMAAGGTTTCACCTTAATCGAACTCAT 400

82 GATTGTAGTTGCAATTATCGGTATCTTAGCGGCTTTCGCTATCCCTGCATACMCGACTACATCGCTCGTTCACAAGCAGCTGMGGCGTMGCTTGGCT

83 GATTGTAGTTGCAATTATCGGTATCTTAGCTGCTTTCGCTATCCCTGCATATAACGACTACATCGCTCGTTCACAAGCAGCTGAAGGCGTAAGCTTGGCT

84 GATTGTAGTTGCAATTATCGGTATCTTAGCGGCTTTCGCTATCCCTGCATACAACGACTACATCGCTCGTTCACAAGCAGCTGAAGGCGTTAGTTTGGCT 500

82 GATGGTTTGAAAGTTCGTATCGCTGAAAACTTACAAGACGGCGAATGTAAAGGACCGGACGCAGATCCAGCATCTGGTGTTGTTGGCAACGAAGACAAAG

83 GATGGTTTGAAAGTTCGTATCGCTGAAAACTTACAAGACGGCGAATGTAAAGGACCGGACGCAGATCCACAATCTGGTGTTGTTGGCAACCAAGACAAAG

B4 GATGGTTTAAAAATCCGCATCGCTGAGAACTTGCAAGACGGCGAATGTAAAGGACCGGACGCAGATCCAGCATCTGGTGTTGTTGGCAACCAAGACAAAG 600

82 GTAAGTATGCCTTAGCAAAAATTGATGGTACCTATGATCAATCAAAGACAGAAGCTGGCGATCCGAATGGTTGTAAGGTCGAAATCACTTATGGTCAAGG

83 GTAAGTATGCCTTAGCAAAAATTGATGGTACCTATGATCAATCAAAGACAGAAGCTGGCGATCCGAATGGTTGTAAGGTCGAAATCACTTATGGTCAAGG84 GTAAGTATGCCTTAGCAGAAATTAAGGGTGACTATGATGAATCAAAGACAGACGCTGGCGATCCGAATGGTTGTAAGGTCGAAATCGCTTATGGTCAAGG 700

82 CACTGCAGAAGGTAAAATTTCTAAGCTGATCACTGGTAAAAAATTGGTTTTAGATCAATTGGTTAATGGTTCATTTATTGCAGGTGATGATACTGACTTA83 CACTGCAGMGGTAAAATTTCTAAGCTGATCACTGGTAAAAAATTGGTTTTAGAACAATTGGTTAATGGTTCATTTATTGCAGGTGCTGGTACAGACTTA84 CACTGCAGAAGGTAAAATTTCTAAGCTGATCACTGGTAAAAAATTGGTTTTAGACCMTTGGTTAATGGTTCATTTGTTCAAGGTGATGGTACTGACTTA 800

82 GCAGATAAATTTATCCCGAATGCAGTAAAA AAAGCTAAAAAATAGTATCTAGTGTAAACATTAGCTTACTTAAAAGCCTCTCTCTT

83 GCAGATAAATTTATCCCGAATGCAGTAMAGCTAAAAAGCCGTAAAAGCTAMAAATAGTATCTAGTGTAAACATTAGCTTACTTAAAAGCCTCTCTCTT84 GCAGATAAATTTATCCCGAATGCAGTAMAGTTAAAAAGCCGTAAAAGCTAAAAAATAGTATCTAGTGTAAACATTAGCTTACTTAAAAGCCTCTCTCTT 900

82 GAGAGGCTTTTTTATTTCACCATCAATCAGTAAAAAGCTATGAATAAACAACGATTTTTATTTGCTGCAAAA

83 GAGAGGCTTTTTTATTTCACCATCAATCAGTAAAAAGCTATGAATAAACMCGATTTTTATTTGCTGCAMA

B4 GAGAGGCTTTTTTATTTCACCATCAATCAGTAAAAAGCTATGAATAAACMCGATTTTTATTTGCTGCAAAA 972

FIG. 2. Nucleotide sequence of pilin genes from strains 183 (B2), 235 and 112 (B3), and VRS 54 (B4). Initiation and termination codonsare underlined, and arrows indicate a potential transcription termination sequence. Positions having substitutions are indicated by carets. TheGenBank accession numbers for the pilin gene sequences are as follows: strain 183, M32664; strain 235, M32602; strain VRS 54, M32601.

VOL. 58, 1990

on March 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

1548 ELLEMAN ET AL.

-7 1meF T

M K S L Q K G F TM K S L Q K G F TM K S L Q K G F T

10L I E L M I V V A IL I E L M I V V A IL I E L M I V V A IL I E L M I V V A I

I G I L AI G I L AI G I L AI G I L A

20A F A I P A Y NA F A I P A Y NA F A I P A Y NA F A I P A Y N

30D Y I A R S QD Y I A R S QD Y I A R S QD Y I A R S Q

40A A E G V S L AA A E G V S L AA A E G V S L AA A E G V - L A

60D G E C K G P D A A S G VD G E C K G P D A D P A S G VD G E C K G P D A D P j S G VD G E C K G P D A D P A SG V

70 80 90V G NED KG K Y[n]L A KID G[O]Y D[5JS K TV G N E D K G K Y A L A K I D G T Y D Q S K TV G Nj> D K G K Y A L A K I D G T Y D Q S K TV G N]D KG KY AL AMI[JGDY DgS K T

100D P N G C K V E I T Y GD P N G C K V E I T Y GD P N G C K V E I T Y GD P N G C K V E IN Y G

110Q G T A KQ G T A E G KQ G T A E G KQ G T A E G K

I S K L II S K L II S K L II S K L I

120T G K K L V L D Q L VT G K K L V L D Q L VT G K K L V L@ Q L VT G K K L V L D Q L V

130 140N G S F I A G D G T D LN G S F I A G D[T D LN G S F I A G[J G T D LN G S F 3[JG D G T D L

150B1 A D K F I P N A V K A K KEB2 A D K F I P N A V K[DE) KB3 A D K F I P N A V K A K K PB4 A D K F I P N A V K ] K K P

FIG. 3. Comparison of amino acid sequence of pilin from strain 234 (B1) with encoded prepilin sequences of strains 183 (B2), 235 (B3),and VRS 54 (B4). Substitutions relative to the majority occupancy are boxed. meF, N-Methylphenylalanine.

ably similar (>95% identity) to those of pilin genes fromother serogroups of B. nodosus. Closer examination of 5'upstream sequences from all B. nodosus pilin genes showeda segregation of sequences into two distinct groups based on

the 11 variable positions between nucleotides 141 and 170 ofFig. 2 (Table 1). This segregation of pilin gene sequences wasindependent of pilin classification criteria, e.g., subtypes B2and B3 fell into different groups, as did subtypes Al and A2(Table 1). These differences presumably arose through ho-mologous recombination between allelic regions upstream ofthe pilin gene subsequent to the divergence of pilins intosubtypes. Mixed infections, often present in a single foot,presumably presented an opportunity for generating thisfurther diversity.

Vaccination trials. The incidence and severity of infection,as seen from the control groups, varied due to differences inthe virulence of challenge strains (Table 2). Nonvaccinatedcontrol groups were severely affected by strains 234 (B1) and334 (B2), whereas infection by strain 112 (B3) was consider-ably less severe. Despite these variations in the severity ofchallenge, it was apparent that sheep vaccinated with pili ofstrains 183 (B2) exhibited high levels of protection againstchallenges with both the homologous subtype (B2) andheterologous subtypes (Bi or B3). These sheep showed farfewer infected feet (footscores .2 [see Table 2, footnote a,

for definition of scores]) and a significantly lower incidenceof severe infections (footscores -3c) with the more virulentchallenge strains than nonvaccinated groups. In contrast,sheep vaccinated with pili of strain 234 (Bi) were less wellprotected against challenge with strains of heterologoussubtypes (B2 and B3) and showed significant protection onlyagainst challenge by a homologous strain.

The magnitude of a serum agglutinin titer is sufficientlydependent on the quality of antigenic preparation, tech-nique, and endpoint determination that the absolute value ofa titer has little significance. Despite this limitation, thesignificant difference (P < 0.001) between the group meantiters to strain 234 antigen (Table 3) for the two vaccinegroups needs to be addressed, since both groups of sheepwere resistant to infection by strain 234 (Table 2). Similar"anomalous" lack of reciprocal cross-reactivity betweenstrains and antisera has been found frequently and maydepend on the relative immunogenicities of various antigenicdeterminants of the pilus (6).

In previous studies on sheep vaccinated and challengedwith the same strain of B. nodosus, a correlation betweengroup mean titer and resistance to infection (34, 38) wasdemonstrated. Thus, if agglutination constitutes a majormechanism of immunity, then the lower titer to the heterol-ogous strain in the present study may reflect a paucity ofcommon agglutinating determinants between strains 183 and234, which is nevertheless sufficient to exceed a thresholdlevel required for resistance to infection. Alternatively, theprotective antibody directed against the pilus may not be an

agglutinin, with agglutinin production and immunity beingseparate independent responses to vaccination with pili. Asimilar interpretation has been used to explain the lack ofcorrelation between resistance and the agglutinin responseof individual sheep to vaccination (34).

DISCUSSION

In the serological classification system of Claxton et al.(3-5) the assignment of B. nodosus isolates to a serogroup

TABLE 1. Nucleotide sequences located 5' upstream of the pilin genes of B. nodosusa

Strain (serogroup and subtype) Sequence

198 (Al), 183 (B2), VRS 54 (B4), 265 (Hi) TCGTCATCATCATTTCCTTTTATAAAACCAA286 (A2), 235 (B3), 112 (B3), 340 (D), 238 (Gl), 351 (H2) T-CATCATTATTCCTTTTTGTGAAACGGG

a Nucleotides 140 to 170 of Fig. 2.

Serot ypeB1B2B3B4

Strain(234)(183)(235)

(VRS544)

B1B2B3B4

D G LD G LD G LD G L

K V R I AK V R I AK V R I A

K[IR I A

50E N L QE N L QE N L QE N L Q

B1B2B3B4

E A GE A GE A GJA G

INFECT. IMMUN.

on March 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

B. NODOSUS SEROGROUP B PILINS 1549

TABLE 2. Analysis of footrot lesions by incidence or severityof infection in vaccinated sheepa

No. of sheep (no. of feet) with indicated footscoresafter challenge with strain (subtype):

Vaccine 234 (Bi) 334 (B2) 112 (B3)

.2 -3c .2 .3c .2 .3c

Strain 234 (Bi) 3 (5) ob (0) 7 (22) 4 (12) 2 (5) 2 (2)Strain 183 (B2) 5 (9) ob (0) 7 (13) ob (0) ob (0) 0 (0)Nonec 8 (29) 6 (14) 8 (32) 8 (25) 6 (8) 2 (3)

a Vaccines contained 50 I.g of pili per dose; groups of eight sheep receivedtwo doses of vaccine at a 35-day interval and were challenged with about 5 x108 B. nodosus organisms per foot after a further 12 days. Feet were inspected20 days after challenge. Footscores: .2 includes all categories of footrotincluding interdigital dermatitis, whereas .3c includes only more severelesions, with separation of the horn of the heel as the minimum lesion.

b p < 0.01 versus no vaccine.c The possibility of nonspecific protection (1) through vaccination has been

excluded in other studies in which pili of a different serogroup from thechallenge strain (18, 33, 35) or pili from another bacterial species (17) failed toelicit protection.

depends on the presence of common agglutinating antigenicdeterminants on pili of both the typing and test strains.Greater amino acid sequence identity might therefore beexpected between pilins from different subtypes within aserogroup than between pilins from different serogroups.This presumption is supported by the data presented above:the pilins of the B serogroup displayed from 89 to 95%identity in amino acid sequence (Fig. 3), but showed nogreater than 76% identity with pilins of other serogroups(11). By comparison, the maximum degree of identity be-tween pilins from different serogroups was only 81%. Thesequence variation in pilins from serogroup B is character-istic of genetic drift. It contrasts with the dramatic changesin antigenic profile of single strains of Neisseria gonorrhoeaeand Moraxella bovis, which is associated with recombina-tion between multiple pilin genes (or parts thereof) in thegenome (26, 30). Only a single pilin gene has been detectedin the B. nodosus genome (15, 16, 21).

In sequence comparisons of closely related pilins fromdifferent serogroups of B. nodosus, four hypervariable re-gions were noted (residues 57 to 64, 83 to 94, 117 to 124, and132 to 142) and were suggested as potential antigenic deter-minants (11). Residues at the C-terminus of the pilin mole-cule, present at or beyond residue 151, are also highlyvariable (11). Comparison of pilin sequences of the B sub-types shows that substitutions which occur outside theinterserogroup hypervariable regions are generally conserv-ative (e.g., V-45 to I, G-77 to A, T-103 to A, D-125 to E, E-71to Q) or compensatory G-110-D to EG), while nonconserva-

TABLE 3. Mean agglutinin titers in serum of vaccinated sheep attime of challenge in cross-agglutination reactions with strains 234

(Bi), 334 (B2), 112 (B3), and VRS 54 (B4) of B. nodosusa

Geometric mean titer with antigen:Vaccine

234 (Bi) 334 (B2) 112 (B3) VRS 54 (B4)

Strain 234 (Bi) pili 114,940 43,053 1,270 5,700Strain 183 (B2) pili 6,218 157,922 11,404 33,200None 37 127 14 26

a Geometric means from groups of 24 sheep. Titers of vaccinated sheepwere significantly greater than titers of unvaccinated sheep. The difference intiters between sheep vaccinated with the two pilus vaccines was alsosignificant (P < 0.001) for each antigen.

tive substitutions occur predominantly in the hypervariableregions and at the C-terminus. One hypervariable regionencompasses nonconservative substitutions found at resi-dues 62 (charge change) and 63 (imino/amino change) in pilinof subtype Bi relative to other B-subtype pilins. Anotherencompasses the substitutions at residue 84 (charge change)and 87 (charge or polarity change) in pilins of subtypes Biand B4 relative to pilins of B2 and B3. Cross-absorptionstudies of B subtypes suggest that the B2 and B3 subtypesshare a unique major agglutinating antigenic determinant inaddition to determinants common the all B subtypes (3), soit is possible that residues at 84, 87, and other positions inthis region (residues 80, 82, and 91) may constitute all or partof this determinant. A further hypervariable region encom-passes residues 133 to 137, where substitutions (mostlycharge changes) occur in pilins from all B subtypes, and inaddition the C-termini of pilins from all B subtypes differ.The hypervariable regions of B. nodosus pilin are unlikely

to contribute greatly to the structural integrity and functionof a pilus, but they may play major roles in determining theantigenic profile of a pilus. The few nonconservative substi-tutions occurring in the B subtype pilins at these putativeantigenic determinants may, by disrupting interface interac-tions between antigenic determinant and antibody, play amajor role in determining the differing serological cross-reactivities of B subtypes and the level of cross-protectionafforded by each subtype. No differences are present in afourth interserogroup hypervariable region (residues 117 to124), which may therefore represent part of a commondeterminant of the B serogroup.As noted in the introduction, Claxton et al. (3-5) were able

to accommodate all B. nodosus isolates from Australiansheep in nine serogroups, with some serogroups furtherdivided into subtypes. In an alternative classificationscheme, Day et al. (6) and Thorley and Day (37) assigned alarge number of British and European isolates to 17 sero-types without need for further subdivision into subtypes.The procedure used by Day et al. (6) emphasized uniquedeterminants on pili, as serotyping antisera were first preab-sorbed to remove all cross-reacting antibodies betweenserotypes. In contrast, the system of Claxton et al. (5) wasbased on both shared and unique antigenic determinants. Asa consequence of these differences, isolates classified as Bsubtypes in the Claxton et al. (5) system would be assignedto different groups (serotypes) by Day et al. (6). While Dayet al. (6) do concede that serotypes B, K, L, and N of theirscheme might be considered analogs of the B serogroup ofClaxton et al. (5), the sequence data presented in thiscommunication argue for the closer relationship suggestedby the classification scheme of Claxton et al. (5).

Serological data are of value in classifying B. nodosusisolates. However, the design of an effective multivalentfootrot vaccine requires information on the immunogenicityof strains. Of importance is the ability of pili prepared froma strain of one subtype to cross-protect against challenge bystrains from other subtypes of the serogroup. The presentwork has demonstrated that pili from strain 183 of subtypeB2 (produced in a surrogate host) afforded protection againstchallenge by strains of subtypes Bi, B2, and B3. A separatestudy (unpublished data) with indigenous B. nodosus 183 pilidemonstrated that protection also extended to challenge bystrain VRS 54 (B4 subtype). Since the immunological iden-tity of indigenous and recombinant-DNA-derived pili hasbeen demonstrated (17, 18), a pilus vaccine from a B2subtype alone should be able to protect sheep againstinfection by strains from other B subtypes. A pilus vaccine

VOL. 58, 1990

on March 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

1550 ELLEMAN ET AL.

from a Bi subtype was much less effective in protectingagainst challenge by strains from other B subtypes. Inwhole-cell vaccines, the presence of non-pilus, cross-protec-tive antigens (32, 33) might provide additional protectionagainst infection from strains of different subtypes, makingthe choice of subtype less important. However, in formulat-ing a multivalent, pilus-based vaccine, the choice of subtypemay be more critical than in whole-cell vaccines in order toelicit similar wide-spectrum protection.

ACKNOWLEDGMENTS

We are grateful to J. A. Vaughan, K. J. Burns, R. D. Edwards,S. B. McLaverty, P. A. Campbell, and D. von Ahlefeldt for skilledtechnical assistance and to P. Lehrbach for provision of P. aerugi-nosa K/2PfS producing B. nodosus 234 pili.

LITERATURE CITED1. Baccarini, M., F. Bistoni, P. Puccetti, and E. Garaci. 1983.

Natural cell-mediated cytotoxicity against Candida albicansinduced by cyclophosphamide: nature of the in vitro cytotoxiceffector. Infect. Immun. 42:1-9.

2. Bradley, D. E. 1974. The adsorption of Pseudomonas aerugi-nosa pilus-dependent bacteriophages to a host mutant withnonretractile pili. Virology 58:149-163.

3. Claxton, P. D. 1981. Serogrouping of Bacteroides nodosus inrelation to protection by vaccination, p. 16-34. In W. I. B.Beveridge and J. R. Egerton (ed.), Ovine footrot: report of aworkshop at the University of Sydney. University of Sydney,Sydney, Australia.

4. Claxton, P. D. 1986. Serogrouping of Bacteroides nodosusisolates, p. 131-134. In D. J. Stewart, J. E. Peterson, N. M.McKern, and D. L. Emery (ed.), Footrot in ruminants: proceed-ings of a workshop, Melbourne 1985. CSIRO Division of AnimalHealth, Sydney, Australia.

5. Claxton, P. D., L. A. Ribeiro, and J. R. Egerton. 1983. Classi-fication of Bacteroides nodosus by agglutination tests. Aust.Vet. J. 60:331-334.

6. Day, S. E. J., C. M. Thorley, and J. E. Beesley. 1986. Serotypingof Bacteroides nodosus: proposal for 9 further serotypes (J-R)and a study of the antigenic complexity of B. nodosus pili, p.147-159. In D. J. Stewart, J. E. Peterson, N. M. McKern, andD. L. Emery (ed.), Footrot in ruminants: proceedings of aworkshop, Melbourne 1985. CSIRO, Division of Animal Health,Sydney, Australia.

7. Egerton, J. R. 1973. Surface and somatic antigens of Fusiformisnodosus. J. Comp. Pathol. 83:151-159.

8. Egerton, J. R., and D. H. Burrell. 1970. Prophylactic andtherapeutic vaccination against ovine foot-rot. Aust. Vet. J.46:517-522.

9. Egerton, J. R., and G. C. Merritt. 1973. Serology of foot-rot:antibodies against Fusiformis nodosus in normal, affected, vac-cinated and passively immunised sheep. Aust. Vet. J. 49:139-145.

10. Egerton, J. R., and D. S. Roberts. 1971. Vaccination againstovine foot-rot. J. Comp. Pathol. 81:179-185.

11. Elieman, T. C. 1988. Pilins of Bacteroides nodosus: molecularbasis of serotypic variation and relationships to other bacterialpilins. Microbiol. Rev. 52:233-247.

12. Elieman, T. C., and P. A. Hoyne. 1984. Nucleotide sequence ofthe gene encoding pilin of Bacteroides nodosus, the causalorganism of ovine footrot. J. Bacteriol. 160:1184-1187.

13. Elieman, T. C., P. A. Hoyne, D. L. Emery, D. J. Stewart, andB. L. Clark. 1984. Isolation of the gene encoding pilin ofBacteroides nodosus (strain 198), the causal organism of ovinefootrot. FEBS Lett. 173:103-107.

14. EUeman, T. C., P. A. Hoyne, D. L. Emery, D. J. Stewart, andB. L. Clark. 1986. Expression of the pilin gene from Bacteroidesnodosus in Escherichia coli. Infect. Immun. 51:187-192.

15. EUeman, T. C., P. A. Hoyne, N. M McKern, D. L. Emery, D. J.Stewart, and B. L. Clark. 1986. Isolation and characterization of

the gene encoding pilin of Bacteroides nodosus (strain 198) andits expression in Escherichia coli, p. 257-261. In D. J. Stewart,J. E. Peterson, N. M. McKern, and D. L. Emery (ed.), Footrotin ruminants: proceedings of a workshop, Melbourne 1985.CSIRO, Division of Animal Health, Sydney, Australia.

16. Elieman, T. C., P. A. Hoyne, N. M. McKern, and D. J. Stewart.1986. Nucleotide sequence of the gene encoding the two-subunitpilin of Bacteroides nodosus 265. J. Bacteriol. 167:243-250.

17. Elleman, T. C., P. A. Hoyne, D. J. Stewart, N. M. McKern, andJ. E. Peterson. 1986. Expression of pili from Bacteroides no-dosus in Pseudomonas aeruginosa. J. Bacteriol. 168:574-580.

18. Elleman, T. C., and D. J. Stewart. 1988. Efficacy against footrotof a Bacteroides nodosus 265 (serogroup H) pilus vaccineexpressed in Pseudomonas aeruginosa. Infect. Immun. 56:595-600.

19. Elieman, T. C., and D. A. von Ahlefeldt. 1987. Nucleotidesequence of the pilin gene from Bacteroides nodosus strain 238(serogroup G). Nucleic Acids Res. 15:7189.

20. Every, D., and T. M. Skerman. 1982. Protection of sheep againstexperimental footrot by vaccination with pili purified fromBacteroides nodosus. N.Z. Vet. J. 30:156-158.

21. Finney, K. G., T. C. Elleman, and D. J. Stewart. 1988. Nucle-otide sequence of the pilin gene of Bacteroides nodosus 340(serogroup D) and implications for the relatedness of sero-groups. J. Gen. Microbiol. 134:575-584.

22. Hartree, E. F. 1972. Determination of protein: a modification ofthe Lowry method that gives a linear photometric response.Anal. Biochem. 48:422-427.

23. Hattori, M., and Y. Sakaki. 1986. Dideoxy sequencing methodusing denatured plasmid templates. Anal. Biochem. 152:232-238.

24. Hoyne, P. A., T. C. Elleman, N. M. McKern, and D. J. Stewart.1989. Sequence of pilin from Bacteroides nodosus 351 (sero-group H) and implications for serogroup classification. J. Gen.Microbiol. 135:1113-1122.

25. Lederberg, E. M., and S. N. Cohen. 1974. Transformation ofSalmonella typhimurium by plasmid deoxyribonucleic acid. J.Bacteriol. 119:1072-1074.

26. Marrs, C. F., W. W. Ruehl, G. K. Schoolnik, and S. Falkow.1988. Pilin gene phase variation of Moraxella bovis is caused byan inversion of the pilin genes. J. Bacteriol. 170:3032-3039.

27. Mattick, J. S., M. M. Bills, B. J. Anderson, B. Dalrymple, M. R.Mott, and J. R. Egerton. 1987. Morphogenetic expression ofBacteroides nodosus fimbriae in Pseudomonas aeruginosa. J.Bacteriol. 169:33-41.

28. McKern, N. M., D. J. Stewart, and P. M. Strike. 1988. Aminoacid sequences of pilins from serologically distinct strains ofBacteroides nodosus. J. Protein Chem. 7:157-164.

29. Messing, J. 1983. New M13 vectors for cloning. MethodsEnzymol. 101:20-78.

30. Meyer, T. F., and R. Haas. 1988. Phase and antigenic variationby DNA rearrangements in prokaryotes, p. 193-220. In A. J.Kingsman, K. F. Chater, and S. M. Kingsman (ed.), Transpo-sition. Symposium of the Society for General Microbiology, vol.43. Cambridge University Press, Cambridge.

31. Stewart, D. J. 1978. The role of various antigenic fractions ofBacteroides nodosus in eliciting protection against foot-rot invaccinated sheep. Res. Vet. Sci. 24:14-19.

32. Stewart, D. J., B. L. Clark, D. L. Emery, J. E. Peterson, andK. J. Fahey. 1983. A Bacteroides nodosus immunogen, distinctfrom the pilus, which induces cross-protective immunity insheep vaccinated against footrot. Aust. Vet. J. 60:83-84.

33. Stewart, D. J., B. L. Clark, J. E. Peterson, D. L. Emery, E. F.Smith, D. A. Griffiths, and I. J. O'Donnell. 1985. The protectiongiven by pilus and whole cell vaccines of Bacteroides nodosusstrain 198 against ovine foot-rot induced by strains of differentserogroups. Aust. Vet. J. 62:153-159.

34. Stewart, D. J., B. L. Clark, J. E. Peterson, D. A. Griffiths, andE. F. Smith. 1982. Importance of pilus-associated antigen inBacteroides nodosus vaccines. Res. Vet. Sci. 32:140-147.

35. Stewart, D. J., and T. C. Elleman. 1987. A Bacteroides nodosuspili vaccine produced by recombinant DNA for the preventionand treatment of footrot in sheep. Aust. Vet. J. 64:79-81.

INFECT. IMMUN.

on March 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

B. NODOSUS SEROGROUP B PILINS 1551

36. Stormo, G. D., T. D. Schneider, and L. M. Gold. 1982. Charac-terization of translational initiation sites in E. coli. NucleicAcids Res. 10:2971-2996.

37. Thorley, C. M., and S. E. J. Day. 1986. Serotyping survey of1296 strains of Bacteroides nodosus isolated from sheep andcattle in Great Britain and western Europe, p. 135-142. In D. J.Stewart, J. E. Peterson, N. M. McKern, and D. L. Emery (ed.),

Footrot in ruminants: proceedings of a workshop, Melbourne1985. CSIRO Division of Animal Health, Sydney, Australia.

38. Thorley, C. M., and J. R. Egerton. 1981. Comparison ofalum-absorbed or non-alum-absorbed oil emulsion vaccinescontaining either pilate or non-pilate Bacteroides nodosus cellsin inducing and maintaining resistance of sheep to experimentalfoot-rot. Res. Vet. Sci. 30:32-37.

VOL. 58, 1990

on March 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from