conservation and diversity of sap homologues and their ... · homologies in the sap island include...

INFECTION AND IMMUNITY, Mar. 2004, p. 1715–1724 Vol. 72, No. 30019-9567/04/$08.00�0 DOI: 10.1128/IAI.72.3.1715–1724.2004Copyright © 2004, American Society for Microbiology. All Rights Reserved.

Conservation and Diversity of sap Homologues and TheirOrganization among Campylobacter fetus Isolates

Zheng-Chao Tu,1,2 John Hui,1 and Martin J. Blaser1,2,3*Departments of Medicine1 and Microbiology,2 New York University School of Medicine, and

Department of Veterans Affairs Medical Center,3 New York, New York 10016

Received 5 September 2003/Returned for modification 24 October 2003/Accepted 23 November 2003

Campylobacter fetus surface layer proteins (SLPs), encoded by sapA homologues, are important in virulence.In wild-type C. fetus strain 23D, all eight sapA homologues are located in the 54-kb sap island, and SLPexpression reflects the position of a unique sapA promoter in relation to the sapA homologues. The extensivehomologies in the sap island include both direct and inverted repeats, which allow DNA rearrangements,deletion, or duplication; these elements confer substantial potential for genomic plasticity. To better under-stand C. fetus sap island diversity and variation mechanisms, we investigated the organization and distributionof sapA homologues among 18 C. fetus strains of different subspecies, serotypes, and origins. For all type Astrains, the boundaries of the sap island were relatively consistent. A 187-bp noncoding DNA insertion near theupstream boundary of the sap island was found in two of three reptile strains studied. The sapA homologueprofiles were strain specific, and six new sapA homologues were recognized. Several homologues from reptilestrains are remarkably conserved in relation to their corresponding mammalian homologues. In total, theobserved differences suggest that the sap island has evolved differing genotypes that are plastic, perhapsenabling colonization of varied niches, in addition to antigenic variation.

Campylobacter fetus, a microaerophilic spiral gram-negativebacterium, has been recognized as an important pathogen inboth animals and humans (6, 19, 23, 33, 35, 36, 38). C. fetus hasbeen isolated from numerous hosts including mammals (ungu-lates, swine, monkeys, and humans), birds, and reptiles (6, 33),and has been divided into two closely related subspecies, fetusand venerealis (1, 31). All subsp. venerealis strains are serotypeA (type A), whereas subsp. fetus may be either serotype A orserotype B (type B) (25, 28, 29).

As with many other bacterial species (34), C. fetus possessesan outermost crystalline surface layer of regular closely packedhigh-molecular-weight protein subunits (S-layer proteins[SLPs]) (9, 10, 17, 24, 26). In C. fetus, the SLPs are encoded byfive to nine sapA homologues in each strain (11, 18, 40). The C.fetus SLPs have been shown to play a critical role in C. fetusvirulence (2, 4, 5, 21, 22, 27) by protecting the bacterium fromphagocytosis and serum killing (3). In addition, SLP phasevariation results in antigenic variation and thus allows thebacterium to escape from host immune defenses (5, 8, 9, 20,21).

In wild-type C. fetus subsp. fetus strain 23D, nine (eightcomplete and one partial) sapA homologues, the unique sappromoter, and the SLP secretion system genes are clustered ona 54-kb chromosomal region termed the sap island, as de-scribed previously (41) (Fig. 1). To better understand the con-servation and diversity of the C. fetus sap island and to gainfurther insights into C. fetus antigenic variation mechanisms,we investigated the distribution and profiles of the sapA ho-mologues and the sap island boundaries among C. fetus strainsdiffering in subspecies and serotype. In the course of the

present study, we also identified six new sap homologues thatfurther our understanding of sap homologue evolution.

MATERIALS AND METHODS

Bacterial strains. The 18 C. fetus strains examined have been extensivelystudied (40, 45) (Table 1). The strains were grown on Brucella Broth (BBLMicrobiology Systems, Cockeysville, Md.) supplemented with 7 U of polymyxinB/ml, 10 �g of vancomycin/ml, 50 �g of nalidixic acid/ml, and 10 �g of trimetho-prim lactate/ml. Frozen stocks were stored at �80°C supplemented with 15% glyc-erol. Escherichia coli strains were grown on Luria-Bertani medium or in Luria-Bertani broth (Difco), and for plasmid selection 100 �g of ampicillin/ml was used.

DNA isolation and manipulation. After 48 h of bacterial growth on two agarplates, C. fetus chromosomal DNA was prepared from cells of each strain asdescribed previously (47). Plasmid DNA was isolated from E. coli by using aQIAgen Spin Miniprep kit (Qiagen, Inc., Valencia, Calif.). Standard protocolswere used for cloning, transformation, restriction digestion, and ligation of plas-mid DNA (32).

PCR. Amplifications of the eight sapA homologues in the 18 C. fetus strainswere performed by using two sets of homologue-specific primers (Table 2). Todetect the presence of Cf0002 and to define the sap island upstream boundaries,PCRs were performed with primer MF paired with MR or paired with AR orBR, respectively (Table 2 and Fig. 2D). The presence of the sap island down-stream gene Cf0031 and the sap island downstream boundary were examined byPCRs by using primer DF paired with DR and primer DR paired with AbF or3F1, respectively (Table 2 and Fig. 3). Cf0009 presence and its location relatedto sapA homologues were investigated by PCRs with primers HF paired with HRand primer HR paired with AR or BR (Fig. 4A). The Cf0020 and sapC statuswithin the sap island were identified by PCRs with primer TR paired with 2F andwith primer CR paired with AR or BR, respectively (Table 2 and Fig. 4E and G).The positions of sapF with sap homologues were examined by using a PCR withprimer FF paired with AR or BR (Table 2 and Fig. 4J). Amplifications wereperformed with 100 ng of chromosomal DNA, 20 pM (each) forward and reverseprimers, 350 nM deoxynucleoside triphosphates, 5 �l of the provided buffer, and0.5 U of Taq polymerase (Qiagen) in a final volume of 50 �l per reaction mixture.PCR conditions consisted of denaturation at 94°C for 1 min, annealing at 5°Cbelow the predicted melting temperature of the primers for 1 min, and extensionat 72°C for 1 min kb�1.

DNA sequencing and analysis. DNA sequencing was carried out by using adideoxy dye termination method on an ABI sequencer (Perkin-Elmer Cetus) bythe Rockefeller University Core facility. Sequence analyses were performed withthe Genetics Computer Group programs (Madison, Wis.). The nucleotide se-

* Corresponding author. Mailing address: Department of Medicine,New York University School of Medicine, 550 First Ave., New York,NY 10016. Phone: (212) 263-6394. Fax: (212) 263-7700. E-mail: [email protected].

1715

on July 23, 2019 by guesthttp://iai.asm

.org/D

ownloaded from

http://iai.asm.org/

quences of sapA8, sapB9, sapA10, sapB11, sapA12, and sapA13 have been de-posited in GenBank under accession numbers AY450397 to AY450402.

Southern hybridization. Chromosomal DNA was digested with HindIII, elec-trophoresed on a 1% agarose gel, and transferred to a positively charged nylonmembrane. The eight sapA homologue-specific, Cf0002-specific, and Cf0009-specific probes were PCR-amplified products labeled by using the RenaissanceChemiluminescence kit (NEN Research Products, Boston, Mass.).

SDS-PAGE and immunoblotting. The sodium dodecyl sulfate (SDS)-poly-acrylamide gel electrophoresis and immunoblotting procedures were used asdescribed previously (40). For immunological detection, recombinant SLP pro-teins were detected by immunoblotting them on 7% SDS-polyacrylamide gels,with polyclonal rabbit serum against the 97-kDa SLP from C. fetus strain 82-40LP, as described previously (26). The goat anti-rabbit immunoglobulin G alkalinephosphatase conjugate was used as secondary antibody at a 1:1,000 dilution.

RESULTS

Upstream boundaries of the sap islands. Cf0002 (mtfB),homologous to other genes encoding mannosyltransferase B,has been shown to be located upstream of the sap island instrain 23D (41). To identify the upstream boundaries of theother 17 C. fetus strains, we first investigated whether mtfB ispresent. PCR with mtfB specific primers indicated that mtfBexists in all 11 type A strains, but not in any of the type Bstrains (Fig. 2A), and Southern hybridization with the mtfB-specific PCR product as a probe verified this result (Fig. 2B).The two different bands in reptile strain 85-389 (Fig. 2B, lane

8) may suggest two copies of mtfB; this analysis was repeatedand the repeat blot showed the same result (data not shown).The various band sizes in different strains indicate substantialpolymorphism in the location of the HindIII restriction sites.

FIG. 1. Schematic representation and genomic organization of the sap island and its flanking regions in C. fetus strain 23D. For ease of reading,selected ORFs (02, 09, 20, 31, 44, and A to Ap8) representing Cf0002, Cf0009, Cf0020, Cf0031, Cf0044, and sapA to sapAp8, are labeled, respec-tively. Each arrow represents the ORF orientation.

TABLE 1. Wild-type C. fetus strains used in this studya

Strainno.

Straindesignation Source

Major SLP No. of saphomologuesTypeb Size (kDa)

1 80-109 Human A 127 82 82-40 Human A 97 83 83-94 Human A 97 54 84-32 Bovine A 97 85 84-86 Human A 97 86 84-92 Bovine A 97 77 85-388 Reptile A 97 98 85-389 Reptile A 149 89 84-112 Bovine A 149 810 99-256 Bovine A 97 811 99-257 Human A 97 712 85-387 Reptile A/B 97 813 84-87 Human B 97 814 84-90 Bovine B NDc 715 84-91 Human B 97 816 84-94 Human B 127 817 84-104 Monkey B 97 718 84-107 Human B 97 7

a As determined elsewhere (40, 45).b That is, the SLP type (45).c ND, not determined.

TABLE 2. PCR primers used in this study

Primer Sequence (5�-3�) Gene Orien-tationa

5� Posi-tionb

A0F AACTCAGTCATCATAACTAC sapA F 1903A0R TTCCATCATCAACTACAACA sapA R 28090F1 AGCTTATTACAGTGAAACTA sapA F 20140R1 GATCTAGCGTACCTGAAA sapA R 2775A1F GGTAGTGATGATACTGTAAA sapA1 F 1953A1R AATCCAGCAAGCTTAATCAA sapA1 R 27201F1 TCAAACTGCAGCTAGTAAAA sapA1 F 12401R1 ATTGTATCATTTCCTTCACC sapA1 R 17002F CTAACACTATAACCGTTACT sapA2 F 2741A2R AGCATCAACAGTGTCATTAA sapA2 R 32912F1 GATGATGCATTAACAATAATA sapA2 F 28892R1 GCAGTGTCTGGAGTAACG sapA2 R 3211A3F TTTCAACGGTGCTAAGCTTA sapA3 F 2496A3R GTCTTATGGGTGGGTTAAAAG sapA3 R 29733F1 TAAATGCTGCGGACTAAA sapA3 F 20493R1 AGATAGATCTATGGCTACAC sapA3 R 2426A4F ATCTGCGTTTGATACGATAA sapA4 F 1921A4R CTCCAAGAGTTAAATTAGCTA sapA4 R 24674F1 GATCTGTAGATGCGCTAA sapA4 F 9834R1 GTGAATGAATCGTTAGCCA sapA4 R 1553A5F ATGTAAGCGTAGAGAATAAG sapA5 F 2618A5R GTAAGGCTATCTAAATCAAC sapA5 R 32905F1 GTTACTGCTGCTGCTAAAATAG sapA5 F 2971A6F ATCTTACAGCTATCGATATCA sapA6 F 2783A6R TTGCATCTTTTGTTGTGCTA sapA6 R 35626R1 CCTATGGCTATCTTATCTATA sapA6 R 29616F1 CTAGCGCTTCACTAAAATTAG sapA6 F 2183A7F GGACTAGAAGTAGGAAATA sapA7 F 2044A7R AAGAAGCTATAGTTGAAGCA sapA7 R 22437F1 ATGTGAAGGTACAAAAGGAA sapA7 F 33627R1 CGTCTATAAGCTGTAGATTAA sapA7 R 3850MF ACCACTAGCATCAAATCTTA mtfB F 37MR CCAACTATAACTAGATCGATA mtfB R 734AR ATCAAGATCACTAGCACTA sapA R 531BR TCAACACTACTACTATTACTA sapB R 525HF AGCAGGAGGATTTGTTGAA Cf0009 F 651HR CAATAGCATTTGCATTTGTG Cf0009 R 1737IF CTTTGTTTTGTTTATTCATTGAA sapAx F �261IR GCATTAGAATAGCATTAATACT sapAx R �112AbF GATTTTATTTTATTTTATTAAGGA sapA(B) F �32AbR TATCTACTATTTATTAAGGTTTGG sapA(B) R �184FF ACTATTAGAAATTTAGAAAGAG sapF F 1268DF AGCTGAGCTAAATACACA Cf0031 F 50SF GATTTTATTTTATTTTATTAAGGA sap F �32SR TATCTACTATTTATTAAGGTTTGG sap R 185DR CAGTTCCTTCTCTTTGATAT Cf0031 R 362CR TCTGCTTGCGTATCTATACA sapC R 671TR CGTCGTTTTCATGATCGA Cf0020 R 321

a F, forward; R, reverse.b Based on the position in the specified gene (accession no. AY211269).

1716 TU ET AL. INFECT. IMMUN.


.org/D

ownloaded from

http://iai.asm.org/

Next, PCRs with primers MF and AR (or BR) indicated thatmtfB is located upstream of a sapA homologue in type A strainsbut, as expected, not in type B strains (Fig. 2C). The productsizes were conserved in all of the type A strains, except for85-388 and 85-389, in which larger bands were seen, suggestingan insertion in these two strains (Fig. 2C). Sequence analysis ofthe product from strain 85-388 indicated a 187-bp polymorphicregion, with 24 (TTT) trimers beginning 293 bp and ending107 bp upstream of the nearest sapA homologue (Fig. 2D). Asearch of GenBank failed to reveal any significant homology tothe 187-bp noncoding sequence. The 187-bp insertion frag-

ment present in strain 85-388 appeared only to be present inthe other type A reptile strain (85-389), as determined by PCRwith primers IF and IR (Fig. 2E).

In strain 23D, the homologue adjacent to mtfB is sapA4. Todetermine whether this relation is conserved in the other typeA strains, we performed PCR with MF and A4R. The resultshowed that sapA4 is the homologue closest to the boundary inonly 4 (including strain 23D) of 11 strains (data not shown).

Downstream boundaries of the sap islands. Since the down-stream boundary of the sap island in strain 23D is flanked byCf0031, we performed PCRs with a Cf0031-specific reverse

FIG. 2. Identification of the sap island upstream boundary in 18 C. fetus strains. (A) Detection of the presence of Cf0002 (mtfB) by PCR withprimers MF and MR. (B) Southern hybridization of Cf0002 with HindIII-digested genomic DNA. The probe is a 707-bp mtfB fragment amplifiedby PCR from strain 23D. (C) PCR detection of the proximity of Cf0002 to sapA homologues with primers MF and AR (lanes 1 to 11) or primersMF and BR (lanes 12 to 18). (D) Schematic representation of the sap island upstream boundary and the strain 85-388 insertion site and sequence.The black box represents the 5� conserved regions of the sapA or sapB homologues. The PCR primers and orientations are designated by arrows.(E) PCR amplification of the insertion fragment found between Cf0002 and sapA homologue in strain 85-388 with primers IF and IR.

VOL. 72, 2004 C. FETUS sap LOCUS AND ANTIGENIC DIVERSITY 1717


.org/D

ownloaded from

http://iai.asm.org/

primer (DR) paired with either its forward specific primer DFor a sapA/B-specific forward primer AbF (Fig. 3A). The resultsshow that Cf0031 is present and is adjacent to the sap island inall 18 strains (Fig. 3B). The PCR with a sapA3-specific primerand Cf0031-specific primer indicate that sapA3 is adjacent toCf0031 in 10 of the 18 strains (Fig. 3D), a finding consistentwith their presence in these strains (see Fig. 5, sapA3-I). Theseobservations indicate that the downstream border of the sapisland is conserved, in contrast to the upstream boundary,where the specific homolog present is not conserved.

Internal boundaries within the sap island. In strain 23D, theeight sapA homologues were separated by internal regionsvarying in size from 0.7 to 12.4 kb. Our previous results showedthat each sapA homologue could potentially change position inthe sap island due to DNA rearrangement mediated by homol-ogy between their 5� conserved regions (41). However, forsapA4, the 3� internal region was stable in strain 23D (41). Toinvestigate the positions of the sapA homologues relative to thenon-sapA genes within the island, we sought to characterize theCf0009, Cf0020, sapC, and sapF locations. The PCR and South-ern hybridization results indicate that Cf0009 is present in allmammalian but not reptilian strains and that it is locatedupstream of sapA or sapB homologues in each case (Fig. 4A toD). Cf0020 is located downstream of sapA2 or sapB2 in 14 ofthe 15 strains in which either are present (Fig. 4E and F). Thus,for most strains, sapA2 (or sapB2) appears to be at the internalboundary of the homologues and nonhomologues within zone

2 of the island (Fig. 1). The PCR results indicate that withinzone 2, sapC always is located upstream of a sapA or a sapB (ina type B strain) homologue (Fig. 4G, H, and I). As in strain23D, sapF is located upstream of sap homologues, with theexception of strain 14, in which the expected PCR product wasabsent (Fig. 4K).

Conservation of the sapA homologues among the 18 C. fetusstrains. In C. fetus strain 23D, all eight sapA homologues (sapAand sapA1 to sapA7) showed at least a 553-bp 5� conservedregion, partially conserved midregions, and substantially diver-gent 3� regions. To assess the presence and genetic diversity ofthese eight sapA homologues in the other 17 C. fetus strains, weperformed PCR with sap homologue-specific primers locatedwithin the divergent 3� regions. All eight sapA homologues arepresent in six strains (strains 1, 2, 4, 5, 6, and 10), but one ormore homologues were absent in the others (Fig. 5). In par-ticular, of the three reptile strains, strains 7 and 8 showed onlysapA and sapA7 PCR products, and strain 12 showed only thesapA7 PCR product. For each PCR product, variation onlyencompassed homologue presence or absence in a strain; therewas no size variation. For each homologue, we performed twoindependent PCRs to determine whether a negative result wasdue to primer mismatches or to the absence of the homologue.In each case, except for sapA3, sapA4, and sapA7, results wereconsistent for the two sets of primers; the observed differencescould be due to polymorphisms in the primer regions or sug-gest new close homologues.

FIG. 3. Identification of the sap island downstream boundary. (A) Strategy for detection of Cf0031 location relative to the sap island. Theprimers are shown as arrows, and the black area indicates the 5� conserved regions of the sapA or sapB homologues. (B) PCR amplification forCf0031 in the 18 C. fetus strains with primers DF and DR. (C) PCR identification of the proximity of Cf0031 with sapA or sapB homologues withprimers AbF and DR. (D) PCR identification of the proximity of cf0031 with sapA3 with primers 3F1 and DR.

FIG. 4. Analyses of Cf0009, Cf0020, sapC, and sapF locations related to sapA or sapB homologues within the sap island of 18 C. fetus strains.(A) Schematic representation of PCR for Cf0009 analyses, with primers indicated by the arrows; (B) PCR for detection of the presence of Cf0009;(C) Southern hybridization for Cf0009; (D) PCR amplifications for the location of Cf0009 related to the sapA homologues with primer HR pairedwith AR (for type A strains) or BR (for type B strains); (E) schematic representation of PCR for Cf0020 analysis, with the primers indicated bythe arrows; (F) detection of the location of Cf0020 in relation to sapA2 (or sapB2) by PCR with primers 2F and TR; (G) schematic representationof PCR for sapC analysis, with the primers indicated by arrows; (H) PCR for sapC amplification and orientation with primers AR and CR; (I) PCRfor sapC analysis with primers BR and CR; (J) schematic representation of PCR for sapF analysis, with the primers indicated by arrows; (K) PCRfor sapF amplification and orientation with primers AR or BR and FF.



.org/D

ownloaded from

http://iai.asm.org/



.org/D

ownloaded from

http://iai.asm.org/

Southern hybridizations were performed with one of the saphomologue-specific PCR products as a probe for each homo-logue (Fig. 6). Most of the sap homologue-specific hybridiza-tions, with the exception of reptile strains 7, 8, and 13, revealedthe same size bands, indicating a close genomic relationshipsacross the strains. The hybridizations with the sapA2 and sapA5probes showed extra bands in some of the type B strains,suggesting the presence of other gene copies or close homo-logues.

Six new sapA homologues and their encoded SLPs. OnlysapA and sapA7 homologues were detected in the reptilestrains, and yet the general architecture of their sap islandsindicated conserved features (Fig. 3 and 4), suggesting thepresence of other sapA homologues. To address this hypothe-sis, we performed PCR for strains 85-387 and 85-388 by usingthe sap noncoding (upstream) region conserved forward prim-

er SF paired with the reverse primer SR (Table 2), which flankthe expressed homologues in strain 23D (41). The productswere cloned into pGEM-T-Easy, and 32 clones were selectedfor subsequent study from each strain. Based on HindIII di-gestion of chromosomal DNA from several of these coloniesindicating variation, DNA sequencing was performed. Thefour homologues identified in strain 85-387 were potentialsapA1 and sapA7 (based on partial sequence; data not shown)and new sap B11 and sapA12. Thus, we now provide directgenetic evidence for the basis of the A/B chimerism (40) ofstrain 85-387. In strain 85-388, the three sapA homologuesidentified are sapA (based on partial sequence; data not shown)and new sapA10 and sapA13 (Fig. 7). By using a similar meth-od, we also identified sapB9 and sapA8 in strains 84-104 and82-40, respectively (Fig. 7). All of the new sapA homologueswere expressed in E. coli and recognized by polyclonal rabbit

FIG. 5. Distribution of eight sapA homologues in 18 C. fetus strains by using two differing sets (I or II) of sap homologue-specific primers (Table2). The numbers in parentheses indicate the primer locations in the specified genes. The lane numbers representing the strains are the same asin Table 1; lane C represents the no-DNA control.



.org/D

ownloaded from

http://iai.asm.org/

serum against the 97-kDa SLP expressed by strain 82-40 (Fig.7B), with the exception of sapB11, the type B SLP from reptilestrain 85-387. A phylogeny of the now 16 known sapA/sapB ho-mologues (7, 12, 41) shows a major dichotomy (Fig. 7C). Thehomologues on the top branch all encode proteins of �97 kDa,whereas the sizes are �111 kDa on the lower branch. Earlierwork showed that 97-kDa SLPs had hexagonal crystallinestructure, whereas the larger products formed tetragonal crys-tals (17).

DISCUSSION

Prior genetic studies indicated that C. fetus strains containfive to nine sapA homologues that are clustered in a narrowchromosomal DNA region (11, 18, 40). In strain 23D, all eightsapA homologues are located in the 54-kb sap island. That 18of 28 predicted open reading frame (ORFs) in the sap island,representing 86% of the coding region length, encode surface-associated proteins and actual or putative protein secretion

FIG. 6. Southern analysis of the distribution of eight sap homologues among C. fetus strains. Each lane contains HindIII-digested genomic DNAfrom the 18 different C. fetus strains, hybridized with each of eight sap homologue-specific probes and amplified by one set of the PCR productsfrom strain 23D shown in Fig. 5. The numbers in parentheses indicate the probe locations within the specified genes.



.org/D

ownloaded from

http://iai.asm.org/

systems (37, 42), suggests that the sap island plays an im-portant role in the interaction between the pathogen andhost. The genomic constituent and DNA structure analysescomparing the sap island with its flanking region and the saphomologues with non-sap homologues suggest that the sapisland was not acquired by recent horizontal gene transferbut is an ancient C. fetus genomic constituent (40, 41, 42).However, we had not determined whether the sap island existsas a fixed region in strains that differ in subspecies, serotype,and origin.

Our present studies indicate that all 18 C. fetus sap islandsshow a consistent downstream boundary adjacent to Cf0031and that all type A C. fetus sap islands are located downstreamof a potential lipopolysaccharide (LPS) locus. Type B strainsdo not possess mtfB, a finding which is consistent with thecompositional differences between type A and type B LPS (25,48). Our results do not exclude the possibility that sap islandsin type B C. fetus strains share a consistent upstream boundaryor that they could be located downstream of their LPS locus. Intotal, these data provide further evidence that the sap islandentered the C. fetus genome before the different subspeciesand serotypes diverged and that the type A and type B sapAhomologue divergence must have occurred after the island hadbeen present. That Cf0009, encoding a putative high-molec-ular-weight surface protein (41), only exists in the 15 mam-malian, but not reptilian, strains suggests that Cf0009 mighthave entered the C. fetus genome after the divergence be-tween mammalian and reptilian strains, a finding consistentwith analyses of G�C content and dinucleotide signatures(42).

The substantial global direct (18.0%) and inverted (16.5%)repeats in the sap island (42) can mediate DNA rearrangementand lead to instability of the island (14, 30, 39, 43, 44). South-ern hybridization with the 5� conserved region as a probeshowed different profiles among the strains, suggesting that thesap islands are relatively variable (40). That all eight sapAhomologues in C. fetus strain 23D can switch sap island posi-tion with one another at high (10�1 to 10�2) frequency due toDNA recombination mediated by their conserved 5� noncod-ing and coding regions is consistent with the plasticity of thesap island (41). The stable existence of sap island internalgenes Cf0009, Cf0020, and sapC among the different strainssuggests that the island components and boundaries may beconserved, but we cannot exclude the existence of furthersmaller polymorphisms.

Reptile and mammalian C. fetus isolates have a number ofdifferences that can be shown in phylogenetic studies (40). Ofthe eight sapA homologues examined, sapA and sapA7 homo-logues were detected in strains 85-388 and 85-389, but no sapA

homologues were detected in strain 85-387 by PCR or bySouthern hybridization. However, the sequence analyses of thecloned sapA homologues in strain 85-387 indicate that sapA1and sapA7 are present. The reasons for these dichotomousresults reflect sequence divergence between the sapA homo-logues in strains 23D and 85-387. The latter sequence failed tobe amplified by the PCR or demonstrated by Southern hybrid-ization due to primer mismatching and internal HindIII sites,respectively (data not shown).

C. fetus strains may be either type A or type B based on theLPS structure and SLP type. The finding that sapA and sapBhomologues coexist in strain 85-387 confirms our previousstudy (40) that strain 85-387 is a type A/B chimera. We ex-cluded the possibility that the strain is a mixture of type A andtype B strains. If the sample tested reflected a mixture of twodifferent strains, we would find relatively equal numbers ofsapA and sapB bands. However, the probe of the 85-387 ge-nome with the sapA and sapB conserved regions showed sixsapB bands and only one sapA band (40). These results indi-cate that 85-387 indeed represents a single strain. SLPs bindspecifically to LPS molecules from homologous (type A or B)but not heterologous cells (49). What the LPS type is for thisstrain and how the different SLPs bind to LPS remains un-known.

In each C. fetus strain, the unique sapA promoter potentiallypermits transcription of all of the sapA homologues, resultingin the expression of different SLP antigens (13–16, 39, 41). C.fetus antigenic variation involving the C. fetus SLPs has beenobserved in vivo and in vitro (21, 30, 45, 46). That different C.fetus strains possess new sap homologues and that their en-coded SLPs are antigenically cross-reactive with one anotherindicate an extensive family of related proteins. The phyloge-netic analyses showed no greater difference between mamma-lian and reptile isolates than between two mammalian strains.The SLP encoded by sapB11 in strain 85-387 did not showcross-antigenicity using the polyclonal antibody against the 97-kDa SLP encoded by the sapA homologue in type A C. fetusstrain 82-40LP, indicating that they do not share major epi-topes. This result suggests that the design of potential C. fetusvaccines should include a pool of different antigens. Finally,the deep branching of the phylogeny of the homologues cor-responds exactly to the observed differences in protein size andin crystalline structure (17). These results both confirm theutility of the distinction and suggest that an ancient gene du-plication event led to the major branching. That the sapB ho-mologues are present on separate branches suggests recombi-nation events to explain this homoplasy.

FIG. 7. Sequence and antigenicity of sapA homologues identified in the present study. (A) Schematic representation of the structures of newsapA homologues compared to their closest homologues in strain 23D. The different colors indicate regions of sequence identity, the white boxesrepresent diverse sequences, and the red outlined box in sapA10 represents a deletion in the (gray) semiconserved region compared to sapA4 andsapB9. (B) Immunoblot of recombinant SLPs probed with polyclonal rabbit antiserum to the 97-kDa SLP of type A strain 84-20LP. The SLPs areencoded by sapA8, sapA12, sapB11, sapA13, sapB9, sapA, and sapA10, respectively. An immunoreactive product was observed in each case, exceptfor sapB11. (C) Phylogenetic tree constructed from the nucleotide sequences of 16 sapA or sapB homologues, including the six new homologues.The tree was constructed by using PAUP 4.0bs neighbor-joining method based on Kimura’s two-parameter model distance matrices. The size ofthe deduced or experimentally determined SLP encoded by the homologue is shown at the left. The major branching perfectly conforms to thedichotomy between 97-kDa SLP (hexagonal) and �97-kDa SLP (tetragonal) in crystalline structure (17).



.org/D

ownloaded from

http://iai.asm.org/



.org/D

ownloaded from

http://iai.asm.org/

ACKNOWLEDGMENTS

This study was supported in part by R01 AI24145 from the NationalInstitutes of Health and by the Medical Research Service of the De-partment of Veterans Affairs.

REFERENCES

1. Berg, R. L., J. W. Jutila, and B. D. Firehammer. 1971. A revised classificationof Vibrio fetus. Am. J. Vet. Res. 32:11–22.

2. Blaser, M. J., P. F. Smith, J. A. Hopkins, I. Heinzer, J. H. Bryner, and W. L.Wang. 1987. Pathogenesis of Campylobacter fetus infections: serum resis-tance associated with high-molecular-weight surface proteins. J. Infect. Dis.155:696–706.

3. Blaser, M. J., P. F. Smith, J. E. Repine, and K. A. Joiner. 1988. Pathogenesisof Campylobacter fetus infections. Failure of encapsulated Campylobacterfetus to bind C3b explains serum and phagocytosis resistance. J. Clin. Inves-tig. 81:1434–1444.

4. Blaser, M. J., and Z. Pei. 1993. Pathogenesis of Campylobacter fetus infec-tions: critical role of high-molecular-weight S-layer proteins in virulence.J. Infect. Dis. 167:372–377.

5. Blaser, M. J., E. Wang, M. K. Tummuru, R. Washburn, S. Fujimoto, and A.Labigne. 1994. High-frequency S-layer protein variation in Campylobacterfetus revealed by sapA mutagenesis. Mol. Microbiol. 14:453–462.

6. Blaser, M. J. 1998. Campylobacter fetus emerging infection and model systemfor bacterial pathogenesis at mucosal surfaces. Clin. Infect. Dis. 27:256–258.

7. Casademont, I., D. Chevrier, and J. L. Guesdon. 1998. Cloning of a sapBhomologue (sapB2) encoding a putative 112-kDa Campylobacter fetus S-layerprotein and its use for identification and molecular genotyping. FEMS Im-munol. Med. Microbiol. 21:269–281.

8. Corbeil, L. B., G. G. Schurig, P. J. Bier, and A. J. Winter. 1975. Bovineveneral vibriosis: antigenic variation of the bacterium during infection. In-fect. Immun. 11:240–244.

9. Dubreuil, J. D., M. Kostrzynska, J. W. Austin, and T. J. Trust. 1990. Anti-genic differences among Campylobacter fetus S-layer proteins. J. Bacteriol.172:5035–5043.

10. Dubreuil, J. D., S. M. Logan, S. Cubbage, D. N. Eidhin, W. D. McCubbin,C. M. Kay, T. J. Beveridge, F. G. Ferris, and T. J. Trust. 1988. Structural andbiochemical analyses of a surface array protein of Campylobacter fetus. J.Bacteriol. 170:4165–4173.

11. Dworkin, J., M. K. Tummuru, and M. J. Blaser. 1995. A lipopolysaccharide-binding domain of the Campylobacter fetus S-layer protein resides within theconserved N terminus of a family of silent and divergent homologs. J. Bac-teriol. 177:1734–1741.

12. Dworkin, J., M. K. Tummuru, and M. J. Blaser. 1995. Segmental conserva-tion of sapA sequences in type B Campylobacter fetus cells. J. Biol. Chem.270:15093–15101.

13. Dworkin, J., and M. J. Blaser. 1996. Generation of Campylobacter fetusS-layer protein diversity utilizes a single promoter on an invertible DNAsegment. Mol. Microbiol. 19:1241–1253.

14. Dworkin, J., and M. J. Blaser. 1997. Nested DNA inversion as a paradigm ofprogrammed gene rearrangement. Proc. Natl. Acad. Sci. USA 94:985–990.

15. Dworkin, J., and M. J. Blaser. 1997. Molecular mechanisms of Campy-lobacter fetus surface layer protein expression. Mol. Microbiol. 26:433–440.

16. Dworkin, J., O. L. Shedd, and M. J. Blaser. 1997. Nested DNA inversion ofCampylobacter fetus S-layer genes is recA dependent. J. Bacteriol. 179:7523–7529.

17. Fujimoto, S., A. Takade, K. Amako, and M. J. Blaser. 1991. Correlationbetween size of the surface array protein and morphology and antigenicity ofthe Campylobacter fetus S layer. Infect. Immun. 59:2017–2022.

18. Fujita, M., T. Morooka, S. Fujimoto, T. Moriya, and K. Amako. 1995.Southern blotting analyses of strains of Campylobacter fetus using the con-served region of sapA. Arch. Microbiol. 164:444–447.

19. Garcia, M. M., G. M. Ruckerbauer, M. D. Eaglesome, and W. E. Boisclair.1983. Detection of Campylobacter fetus in artificial insemination bulls with atransport enrichment medium. Can. J. Comp. Med. 47:336–340.

20. Garcia, M. M., C. L. Lutze-Wallace, A. S. Denes, M. D. Eaglesome, E. Holst,and M. J. Blaser. 1995. Protein shift and antigenic variation in the S-layer ofCampylobacter fetus subsp. venerealis during bovine infection accompaniedby genomic rearrangement of sapA homologs. J. Bacteriol. 177:1976–1980.

21. Grogono-Thomas, R., J. Dworkin, M. J. Blaser, and D. G. Newell. 2000.Roles of the surface layer proteins of Campylobacter fetus subsp. fetus inovine abortion. Infect. Immun. 68:1687–1691.

22. Grogono-Thomas, R., M. J. Blaser, M. Ahmadi, and D. G. Newell. 2003. Roleof S-layer protein antigenic diversity in the immune responses of sheepexperimentally challenged with Campylobacter fetus subsp. fetus. Infect. Im-mun. 71:147–154.

23. Guerrant, R. L., R. G. Lahita, W. C. Winn, Jr., and R. B. Roberts. 1978.

Campylobacteriosis in man: pathogenic mechanisms and review of 91 blood-stream infections. Am. J. Med. 65:584–592.

24. McCoy, E. C., D. Doyle, K. Burda, L. B. Corbeil, and A. J. Winter. 1975.Superficial antigens of Campylobacter (Vibrio) fetus: characterization of anantiphagocytic component. Infect. Immun. 11:517–525.

25. Moran, A. P., D. T. O’Malley, T. U. Kosunen, and I. M. Helander. 1994.Biochemical characterization of Campylobacter fetus lipopolysaccharides. In-fect. Immun. 62:3922–3929.

26. Pei, Z. R., T. D. Ellison, R. V. Lewis, and M. J. Blaser. 1988. Purification andcharacterization of a family of high molecular weight surface-array proteinsfrom Campylobacter fetus. J. Biol. Chem. 263:6414–6420.

27. Pei, Z., and M. J. Blaser. 1990. Pathogenesis of Campylobacter fetus infec-tions: role of surface array proteins in virulence in a mouse model. J. Clin.Investig. 85:1036–1043.

28. Perez-Perez, G. I., and M. J. Blaser. 1985. Lipopolysaccharide characteristicsof pathogenic campylobacters. Infect. Immun. 47:353–359.

29. Perez-Perez, G. I., M. J. Blaser, and J. H. Bryner. 1986. Lipopolysaccharidestructures of Campylobacter fetus are related to heat-stable serogroups. In-fect. Immun. 51:209–212.

30. Ray, K. C., Z. C. Tu, R. Grogono-Thomas, D. G. Newell, S. A. Thompson, andM. J. Blaser. 2000. Campylobacter fetus sap inversion occurs in the absenceof RecA function. Infect. Immun. 68:5663–5667.

31. Salama, S. M., M. M. Garcia, and D. E. Taylor. 1992. Differentiation of thesubspecies of Campylobacter fetus by genomic sizing. Int. J. Syst. Bacteriol.42:446–450.

32. Sambrook, J. E. F., Fritsch, and T. Maniatis. 1989. Molecular cloning: alaboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, ColdSpring Harbor, N.Y.

33. Skirrow, M. B. 1990. Campylobacter and Helicobacter infections of man andanimals, p. 531–545. In M. T. Parker, and L. H. Collier (ed.), Principles ofbacteriology, virology, and immunity, 8th ed., vol. 2. Edward Arnold, Lon-don, England.

34. Sleytr, U. B., and T. J. Beveridge. 1999. Bacterial S-layers. Trends Microbiol.7:253–260.

35. Smibert, R. M. 1978. The genus Campylobacter. Annu. Rev. Microbiol.32:673–709.

36. Thompson, S. A., and M. J. Blaser. 2000. Pathogenesis of Campylobacterfetus infections, p. 321–347. In I. Nachamkin and M. J. Blaser (ed.), Campy-lobacter, 2nd ed. American Society for Microbiology, Washington, D.C.

37. Thompson, S. A., O. L. Shedd, K. C. Ray, M. H. Beings, J. P. Jorgensen, andM. J. Blaser. 1998. Campylobacter fetus surface layer proteins are transportedby a type I secretion system. J. Bacteriol. 180:6450–6458.

38. Tremblay, C., C. Gaudreau, and M. Lorange. 2003. Epidemiology and anti-microbial susceptibilities of 111 Campylobacter fetus subsp. fetus strains iso-lated in Quebec, Canada, from 1983 to 2000. J. Clin. Microbiol. 41:463–466.

39. Tu, Z. C., K. C. Ray, S. A. Thompson, and M. J. Blaser. 2001. Campylobacterfetus uses multiple loci for DNA inversion within the 5� conserved regions ofsap homologs. J. Bacteriol. 183:6654–6661.

40. Tu, Z. C., F. E. Dewhirst, and M. J. Blaser. 2001. Evidence that the Campy-lobacter fetus sap locus is an ancient genomic constituent with origins beforemammals and reptiles diverged. Infect. Immun. 69:2237–2244.

41. Tu, Z. C., T. M. Wassenaar, S. A. Thompson, and M. J. Blaser. 2003.Structure and genotypic plasticity of the Campylobacter fetus sap locus. Mol.Microbiol. 48:685–698.

42. Tu, Z. C., D. W. Ussery, D. T. Pride, and M. J. Blaser. 2003. Genomiccharacteristics of the Campylobacter fetus sap island. Genome Lett. 2:40–46.

43. Tummuru, M. K., and M. J. Blaser. 1992. Characterization of the Campy-lobacter fetus sapA promoter: evidence that the sapA promoter is deleted inspontaneous mutant strains. J. Bacteriol. 174:5916–5922.

44. Tummuru, M. K., and M. J. Blaser. 1993. Rearrangement of sapA homologswith conserved and variable regions in Campylobacter fetus. Proc. Natl. Acad.Sci. USA 90:7265–7269.

45. Wang, E., M. M. Garcia, M. S. Blake, Z. Pei, and M. J. Blaser. 1993. Shiftin S-layer protein expression responsible for antigenic variation in Campy-lobacter fetus. J. Bacteriol. 175:4979–4984.

46. Wesley, I. V., and J. H. Bryner. 1989. Antigenic and restriction enzymeanalysis of isolates of Campylobacter fetus subsp. venerealis recovered frompersistently infected cattle. Am. J. Vet. Res. 50:807–813.

47. Wilson, K. 1995. Preparation of genomic DNA from bacteria, p. 2.4.1–2.4.5.In F. M. Ausubel, R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman,J. A. Smith, et al. (ed.), Current protocols in molecular biology, vol. 1. JohnWiley & Sons, Inc., New York, N.Y.

48. Winter, A. J., E. C. McCoy, C. S. Fullmer, K. Burda, and P. J. Bier. 1978.Microcapsule of Campylobacter fetus: chemical and physical characterization.Infect. Immun. 22:963–971.

49. Yang, L. Y., Z. H. Pei, S. Fujimoto, and M. J. Blaser. 1992. Reattachment ofsurface array proteins to Campylobacter fetus cells. J. Bacteriol. 174:1258–1267.

Editor: W. A. Petri, Jr.



.org/D

ownloaded from

http://iai.asm.org/

conservation and diversity of sap homologues and their ... · homologies in the sap island include...

Documents