A chromosomal region of Mycoplasma agalactiae containing vsp-related genes undergoes in vivo rearrangement in naturally infected animals

Abstract

A Mycoplasma agalactiae genomic fragment carrying four vsp-related genes (designated avg: agalactiae variable genes) was cloned, sequenced and compared to the vspA gene of Mycoplasma bovis. The following features were revealed: (i) the presence of a highly conserved vsp 5′ upstream region; (ii) a highly homologous vsp N-terminal end encoding a putative lipoprotein signal sequence; (iii) sequence divergence of the rest of the mature proteins. By using avg specific probes in Southern blot analysis of genomic DNAs of M. agalactiae strains as well as of isolates from infected animals, marked DNA polymorphism of avg fragments was demonstrated. In addition, the avg genomic fingerprints were monitored for a period of 7 months, in isolates of M. agalactiae from an individual chronically infected animal. The results provided evidence that a chromosomal region of M. agalactiae, carrying vsp-related genes, undergoes rearrangements in vivo in the natural animal host during the course of infection.

avg, vsp, Genomic rearrangement, Mycoplasma agalactiae, Mycoplasma bovis

1 Introduction

The mycoplasmas (class Mollicutes) are the smallest known organisms capable of self-replication, totally lacking a cell wall and carry a remarkably small genome that in some species consists of only 600 bp. [1,2]. Nonetheless, these microorganisms infect a diverse range of animal hosts, and are often implicated in disease [3]. Mycoplasma agalactiae is one of the most significant pathogenic mycoplasmas of small ruminants and has been isolated worldwide [4]. Contagious Agalactia, the disease caused by M. agalactiae, is expressed clinically as mastitis, arthritis and keratoconjunctivitis [4] and is of considerable economic importance, particularly in the Mediterranean basin. The disease is endemic in Israel [5]. M. agalactiae is phylogenetically closely related to Mycoplasma bovis, an important bovine pathogen causing mastitis, arthritis and respiratory disease [6]. Despite the fact that there are frequent serological reactions between the two mycoplasmas [6,7] and a high degree of homology in the 16S rRNA genes [8], less than 40% homology between the genomes was reported [9].

The ability of pathogenic mycoplasmas to overcome the host immune defense mechanisms and to cause chronic infections, have been attributed in recent years to the capability of these microorganisms to rapidly change the antigenic repertoire of their cell surface [1,10,11]. The generation within the population of antigenically distinct cells (heterotypes), in which high-frequency variation of key surface antigens has occurred, enables efficient circumvention of the host immune system during infection [12–14].

Recently, a genomic locus containing a cluster of 13 genes encoding abundant amphilic variable surface lipoproteins (designated Vsps) in M. bovis has been identified and characterized [15]. Members of the Vsp family were shown: (i) to be anchored in membrane via a lipid moiety on an N-terminal Cys residue; (ii) to undergo high-frequency non-coordinated phase variation between ON and OFF expression states; (iii) to undergo high-frequency size variation; (iv) to possess a highly conserved promoter region as well as a highly homologous N-terminal encoding region; and (v) to be composed of extensive repetitive units [15,16]. In addition, we have shown that the spontaneously high rate of Vsp phenotypic switching involves DNA rearrangements that occur at high frequency within the M. bovis vsp locus [17].

The relatedness between M. agalactiae and M. bovis led us to explore the possibility that a genetic system analogous to the vsp gene family of M. bovis may also be present in the M. agalactiae genome. We have previously shown that multiple genomic fragments homologous to the M. bovis vspA gene are present in the M. agalactiae chromosome and that these vsp-related fragments exhibit marked DNA polymorphism among strains [18]. However, the vsp-related genes of M. agalactiae have not yet been cloned and characterized. Another novel aspect of this study was to gain experimental evidence linking the DNA polymorphism of the vsp-related genomic fragments with possible genomic rearrangement events occurring in vivo in the vsp-related locus of M. agalactiae strains isolated from naturally infected animals.

2 Materials and methods

2.1 Mycoplasma strains, cultivation and identification

Type strain M. agalactiae PG2 was originally obtained from Dr. D.G.ff. Edward, Wellcome Research Laboratories, Beckenham, Kent, UK. M. bovis PG45 type strain was obtained from the collection of the Institute for Microbiology and Infectious Diseases of Animals, School of Veterinary Medicine, Hannover, Germany. M. agalactiae strain GM139 is the prototype USA isolate described by DaMassa [19]. Strain designations of M. agalactiae isolates in Israel, some of which have been previously described [18], are summarized in Table 1.

1

Origin of M. agalactiae strains isolated in Israel

Strain designation	Animal host	Clinical sign	Site of isolation	Date of isolation
Baka	sheep	C. A.a	milk	1970
15445	sheep	C. A.	milk	12/91
5292	goat	C. A.	milk	4/95
80346/1	sheep	C. A.	milk	3/95
80346/2	lamb	keratoconjunctivitis	eye	3/95
13605	sheep	C. A.	milk	6/92
5854	sheep	C. A.	milk	10/92
5512	sheep	C. A.	milk	1/92
5026	Sheep	polyarthritis	joint	1/93
55003/1	goat	C. A.	milk	3/94
5089/1–3b	sheep	C. A.	milk	2/96
5675/1–4b	sheep	C. A.	milk	9/96
5303/1–3b	sheep	C. A.	milk	4/96
627/3–15c	sheep	asymptomatic	milk	1997–1998

^aC. A.=Contagious Agalactia.

^bIsolation from different animals at the same sampling time.

^cCorresponding to lanes 2–14 in Fig. 5. Isolation from the same animal at sequential sampling times.

Isolation of mycoplasmas was by standard methods using mycoplasma medium based on Heart Infusion broth (Difco) containing 0.5% yeast extract (Difco) and 20% inactivated horse serum. Clinical samples, usually milk, were plated directly on mycoplasma agar and in parallel inoculated into a mycoplasma broth culture for enrichment. Species identification of mycoplasma colonies was by direct or indirect immunofluorescence with antiserum specific for M. agalactiae and M. bovis. In addition, all M. agalactiae isolates were examined genetically for their specific rRNA fingerprints using the ribosomal probe pMC5 and were shown to display a typical pattern for M. agalactiae, identical in all strains tested and distinct from that of M. bovis, as we have previously shown [18].

2.2 DNA preparation, genomic library construction and manipulations

Genomic DNA from M. agalactiae strains and clinical isolates was extracted and purified as described previously for M. bovis[15]. The DNA was digested by restriction enzymes and electrophoresed as previously described [20]. A recombinant phage library was constructed in the phage vector Lambda ZAP II (Stratagene, Cedar Creek, TX, USA) using partially digested Sau3A chromosomal fragments from M. agalactiae PG2 type strain. About 4×10⁴ PFU (plaque forming units) were subjected to in situ hybridization [21] using the M. bovis vspA gene as a probe [15]. Positive phages were identified and purified as described elsewhere [15]. In vivo excision of positive phages was done according to the recommendation of the manufacturer.

2.3 Labeling of oligonucleotide and hybridization conditions

avg sequence specific oligonucleotides were synthesized at the interdepartmental facility of the Hebrew University-Hadassah Medical School on a model 380B DNA synthesizer (Applied Biosystems, Inc., Foster City, CA, USA). The sequence of the 23-bp sig-1 oligonucleotide was: 5′-CCTTTTGTAGCTGCTAAGTGTGG-3′; that of the 18-bp a-1 oligonucleotide was: 5′-CTGAAGGTGGTAGCGATC-3′A; that of the 20 bp b-1 oligonucleotide was: 5′-AACTGAAGGTGGCGGTAATC-3′. About 100 ng oligonucleotide was ³²P-labeled with 25 U of T4 polynucleotide kinase at 37°C for 1 h in 25 μl of a reaction mixture containing 40 mM Tris, pH 7.5, 10 mM MgCl₂, 5 mM DTT, and 2.5 μl of [γ³²P]ATP (3000 Ci/nmol). The DNA hybridization conditions were described elsewhere [15,20].

2.4 DNA sequence analysis

DNA sequence analysis of both strands was performed by the dideoxy chain termination method, using the automatic sequencer Dye-terminator cycle sequencing model ABI Prism 337 (Perkin-Elmer, Foster City, CA, USA). Overlapping sets of deletion mutants were generated from the recombinant phagemid carrying the avg genes (depicted in Fig. 1) by graded directional exonuclease III digestion using the Erase-A-Base deletion kit (Promega). As the sequencing primers, the T7 promoter sequence located on the cloning vector as well as avg-related sequences were used. Sequence data were analyzed using the genetic computer software: AssemblyLIGN and MacVector 6.0 (Oxford Molecular Group, Oxford, UK).

1

Schematic representation, restriction map and organization of the avgA, avgB, avgC and avgD genes of M. agalactiae PG2 type strain. The solid thick line represents a 6.6-kb M. agalactiae DNA insert obtained from the recombinant bacteriophage λMA-5. The position of HindIII (H) and EcoRI (E) restriction sites are marked. The locations and the directions of four Avg ORFs are indicated by shaded arrows. An additional non-Avg-related ORF (ORF-1) is indicated by an open arrow. Highly homologous regions 5′ of each avg gene and highly conserved N-terminal region encoding a lipoprotein signal peptide are indicated by hatched boxes. The location of two synthetic oligonucleotides representing sequences complementary to distinct sequences of the avgA or avgB structural genes, designated a-1 and b-1, are marked by arrows. A third oligonucleotide (designated sig-1), representing sequences complementary to the highly conserved N-terminal encoding region common to all avg genes, is also indicated by an arrow. The two HindIII fragments of 1 kb and of 5.3 kb in size, that were mentioned within Section 3, are shown by a broken line.

2.5 In vivo case study analysis

A longitudinal case study was carried out in an intensively managed (zero grazing) dairy sheep flock located in a village in northern Israel. M. agalactiae was isolated in this flock from outbreaks of Contagious Agalactia in 1997 and early 1998. During these episodes animals with severe clinical disease were culled from the flock whereas some ewes which displayed mild disease signs returned to milk production and remained in the flock. At the beginning of the study, eight animals in which M. agalactiae was isolated from milk samples were selected and tested individually at 2–4 week intervals over a period of 7 months. There was no introduction of animals from other flocks during the course of this study. Sporadic clinical manifestations were observed in the flock but none were seen in the animals being tested.

2.6 Nucleotide sequence accession numbers

The nucleotide sequence data in this report have been been deposited in the GenBank database under accession numbers: AF112466 (avgA), AF112467 (avgB), AF205063 (avgC), AF205064 (avgD).

3 Results

3.1 The M. agalactiae genome possesses a gene cluster displaying homology to the vsp gene family of M. bovis

To identify M. agalactiae genomic fragments homologous to the vsp gene family of M. bovis, a genomic library of M. agalactiae PG2 type strain was screened with the M. bovis vspA gene [15] as a probe. A recombinant phage, designated λMA-5, hybridizing strongly with the vspA gene, was isolated and analyzed further. An about 7-kb DNA insert from the resultant phagmid was digested with the restriction enzyme HindIII to give three authentic genomic fragments of about 0.3, 1.0 and 5.3 kb (Fig. 1). Within the cloned insert fragment, a cluster of five open reading frames (ORFs), not all similarly oriented, was deduced to exist from the nucleotide sequences analyzed (Fig. 1). Examination of the nucleotide sequences and of the deduced amino acid (aa) sequences of these ORFs revealed four vsp-related genes, designated as avgA–D. The fifth ORF (designated ORF-1), did not show homology to any vsp gene. All four avg genes are preceded by a highly conserved 5′ non-coding sequence that can be divided into two regions. The first, a 72-bp region upstream of the ATG initiation codon, exhibits about 92% homology among the four-avg genes as well as 92% identity to the corresponding region of the vspA gene of M. bovis. The second region, of about 80 bp, is more divergent (Fig. 2a). The N-terminal region of 31 aa is also a highly homologous domain showing 93% homology among avgA–D, as well as 90% homology to the N-terminal region of the vspA protein of M. bovis (Fig. 2b). The Avg N-terminal region contains a prokaryotic lipoprotein signal sequence which begins with a sequence containing three positively charged Lys residues followed by an adjoining core of 19 hydrophobic aa terminating with the tetrapeptide Ala-Ala-Lys-Cys. In contrast to the high conservation of the avg 5′ region and of the N-terminal end among AvgA–D, and in comparison to the VspA protein, the rest of the mature Avg proteins exhibit sequence divergence. The most obvious difference is the absence of reiterated coding sequences within the four Avg-coding regions, a motif that comprises the major portion of the Vsp molecules in M. bovis[15].

2

a: Sequence alignment of the 5′ regions of four M. agalactiae avg genes in comparison to the corresponding region of the vspA gene of M. bovis. The alignment was done using the MacVector 6.0 software. The name of each gene is indicated on the left of each row. Identical nucleotides are shaded. Nucleotides representing a putative ribosome-binding site (SD) are underlined. The Avg initiation codon (ATG) is marked by an arrow. b: Alignment of deduced aa sequences of the N-terminal region of AvgA, AvgB, AvgC and AvgD M. agalactiae molecules in comparison to the corresponding region of the VspA protein of M. bovis. Identical aa residues are shown by dark shaded boxes and similar residues by light shaded boxes. The name of each protein is indicated on the left of each row. The Cys residue within the lipoprotein box is indicated by an arrowhead.

The presence of multiple avg genes within the chromosome of M. agalactiae was also demonstrated by a Southern blot experiment. An oligonucleotide, representing sequences complementary to the highly conserved avg-signal peptide-encoding region (designated sig-1, Fig. 1), was used as a probe against HindIII-digested genomic DNAs of M. agalactiae reference strains and a representative cross-section of clinical isolates (Table 1), chosen to represent different regions in Israel and clinical presentations (Fig. 3). Multiple genomic fragments of different intensity and size (between 6 and 12 fragments depending on the isolate) hybridized to the conserved N-terminal oligonucleotide probe, suggesting the presence of a large avg gene family. Interestingly, marked DNA polymorphism was detected, as demonstrated by the existence of different avg hybridization profiles among all the 13 M. agalactiae isolates tested (Fig. 3). The difference from the GM139 strain isolated in the USA [19] is particularly marked (Fig. 3, lane 2). Notably, the difference in intensity of the avg genomic fragments, obtained using the conserved sig-I oligonucleotide probe, was attributed not only to the number of avg genes present on each fragment but also to nucleotide substitutions among the signal peptide-encoding region of the various avg genes. For example, the sig-I oligonucleotide derived from the avgC gene, differs in two nucleotides from that of the avgA gene, resulted in a weak 1.0-kb HindIII fragment carrying the avgA gene of the PG2 type strain (Fig. 3, lane 1).

3

Identification of multiple M. agalactiae vsp-related genomic fragments. Chromosomal DNA (4 μg) from a representative cross-section of M. agalactiae reference strains and clinical isolates was digested to completion with restriction endonuclease HindIII and subjected to Southern blot hybridization with a ³²P-labeled synthetic oligonucleotide (sig-1) which represents sequences complementary to the highly homologous Avg-N-terminal region. M. agalactiae strains include: M. agalactiae PG2 type strain (lane 1), GM139 (lane 2), BAKA (lane 3), 15445 (lane 4), 5292 (lane 5), 80346/1 (lane 6), 80346/2 (lane 7), 13605 (lane 8), 5854 (lane 9), 5512 (lane 10), 5026 (lane 11), 55003/1 (lane 12), and 5089/1 (lane 13). Molecular size markers are indicated.

3.2 In vivo rearrangements of the avg-related genomic fragments in naturally infected animals

The marked DNA polymorphism of avg-related genomic fragments, observed among the M. agalactiae clinical isolates (Fig. 3), raises the intriguing possibility that such DNA polymorphism may reflect genomic rearrangements occurring within the avg locus, as was demonstrated for the vsp locus of M. bovis[15]. To gain experimental evidence supporting this hypothesis, we isolated M. agalactiae from several individual animals in each of three separate flocks (see Table 1). The isolates were made at the acute stage of clinical Contagious Agalactia from milk samples submitted for diagnosis. HindIII-digested genomic DNAs from these clinical isolates were subjected to Southern blot hybridization using the oligonucleotide sig-1 probe (Fig. 4). Comparison of the avg genomic profiles among the three flocks showed a relatively similar pattern among the isolates within each flock in marked contrast to the high degree of heterogeneity among isolates from different outbreaks (Fig. 3). Differences in avg profile among isolates in the same flock showed an interesting pattern. For example, two HindIII fragments of about 3 and 4.7 kb in size, present in two isolates from flock #3 (Fig. 4, lanes 8 and 10, indicated by open arrows), were absent in another isolate of the same flock (Fig. 4, lane 9). However, the isolate in lane 9 contains two other HindIII fragments of about 6.5 and 2.1 kb in size, which were not detected in the flock-mates (Fig. 4, lane 9, indicated by solid arrows). All the other avg-related genomic fragments remained unchanged in all three isolates of flock #3.

4

Genomic rearrangement of avg-related genomic fragments among M. agalactiae clinical isolates. Isolates were obtained from milk samples of ewes with Contagious Agalactia in three different flocks indicated at the top by numbered brackets. Chromosomal DNA (4 μg) from these isolates was digested to completion with restriction endonuclease HindIII and subjected to Southern blot hybridization with sig-1 which represents sequences complementary to the highly homologous Avg-N-terminal region. M. agalactiae strains include: 5303/3 (lane 1), 5303/2 (lane 2), 5303/1 (lane 3), 5675/4 (lane 4), 5675/3 (lane 5), 5675/2 (lane 6), 5675/1 (lane 7), 5089/4 (lane 8), 5089/3 (lane 9) and 5089/2 (lane 10). Bands present only in isolate 5089/3 in lane 9 (solid arrows) or only in isolates 5089/4 and 5089/2 in lanes 8 and 10 (open arrows) are indicated. Molecular size markers are indicated.

3.3 In vivo rearrangements of the avg genomic fragments in a chronically infected asymptomatic animal

To further demonstrate that rearrangement events occur in vivo within the avg genomic region, we carried out a longitudinal case study in closed dairy sheep flock which had undergone recurrent outbreaks of Contagious Agalactia, suggesting the presence of persistent infection in the flock. Preliminary testing at the onset of the study in July 1997 indicated that M. agalactiae was present in milk samples from some of the asymptomatic animals which had previously undergone clinical disease. We followed the avg genomic fingerprints in one naturally infected animal (#627), in which M. agalactiae was consistently isolated at each sampling time (2–4 weekly intervals over a period of 7 months). HindIII-digested genomic DNAs from these sequential isolates were subjected to Southern blot hybridization using the highly conserved sig-1 probe (Fig. 5A). In vivo rearrangements of avg-related genomic fragments occurring in the same animal during the course of infection were clearly evident. During the period of 7 months several distinct avg genomic profiles were identified. For example, M. agalactiae strains isolates at sampling time 3 (627/3) and sampling time 4 (627/4), 2 weeks apart at the beginning of the study, show identical patterns (Fig. 5a, lanes 2 and 3) while strains 627/5 and 627/6 (Fig. 5a, lanes 4 and 5), isolated subsequently at 1-month intervals, display two distinct avg genomic profiles.

5

In vivo rearrangements of M. agalactiae-avg-related genomic fragments in successive isolates from an asymptomatic ewe. HindIII-digested genomic DNA (4 μg) from 13 M. agalactiae isolates was subjected to Southern blot hybridization. The isolates (lanes, 2–14) were obtained from a single naturally infected animal, designated #627, over a period of 7 months in which M. agalactiae was consistently isolated at each sampling time (2–4 weekly intervals). The ³²P-labeled sig-1, a-1, and b-1 oligonucleotides represent sequences complementary to the highly Avg-homologous N-terminal region, to a distinct sequence of the avgA gene or to a distinct sequence of the avgB gene (a–c respectively), were used as probes. M. agalactiae PG2 type strain is also shown in lane 1. Molecular size markers are indicated.

Two genomic fragments of the PG2 type strain that were part of the cloned avg locus (depicted in Fig. 1) were shown to vary in size in the sequential clinical isolates. The first was a 1.0-kb HindIII fragment bearing the entire avgA gene and the 3′-end of the avgB gene and the second was a 5.3-kb HindIII fragment bearing part of the avgB gene and the complete avgC and avgD genes (Fig. 5a, lane 1, see also Fig. 1).

In order to monitor the corresponding genomic fragments in the chromosome of sequential clinical isolates, oligonucleotides representing sequences complementary to distinct sequences of the avgA or avgB structural genes, designated a-1 and b-1 respectively (see Fig. 1), were used as probes in Southern blot hybridization against the isolates depicted in Fig. 5a. The a-1 probe recognized only the 1.0-kb HindIII fragment in the genome of the PG2 strain (Fig. 5b, lane 1). The avgA-bearing fragment was shown to undergo rearrangement in the sequential clinical isolates during the period of 7 months of infection, yielding three differently size fragments of 1.8 kb (Fig. 5b, lanes, 2 and 3), 1.1 kb (Fig. 5b, lanes 4, 12 and 13), and 6.0 kb (Fig. 5b, lanes 5–11 and lane 14). Sequence analysis of the 1.8-kb (lane 3) and of the 1.1-kb (lane 4) genomic fragments in comparison to the 1.0-kb fragment of the PG2 type strain (lane 1) has shown in these isolates that the avgB gene was inverted and replaced in two different locations within the avg locus (data not shown). Rearrangement of the avgB gene gave rise to changes in the position of the two HindIII restriction sites present within the avgB structural gene thus, affecting the HindIII hybridization pattern.

The b-1 probe recognized four genomic fragments in the PG2 type strain including the 5.3-kb fragment (Fig. 5c, lane 1). While three genomic fragments, recognized by the b-1 probe, remained unchanged in all isolates including the PG2 strain (Fig. 5c), rearrangement of the 5.3-kb fragment in the sequential clinical isolates during the period of 7 months was clearly evident. Rearrangement of the 5.3-kb fragment apparently gave rise to the generation of three new fragments of about 1.8 kb (Fig. 5c, lanes 2 and 3), 2.4 kb (Fig. 5c, lanes 4, 12 and 13) and 2.1 kb (Fig. 5c, lanes 5–11 and lane 14). Taken together, these results provide evidence indicating that a M. agalactiae specific genomic region comprised of several avg genes undergoes in vivo rearrangements in the natural animal host.

4 Discussion

Two noteworthy findings were revealed in this study. First, the chromosome of M. agalactiae, a pathogen of small ruminants, contains a family of multiple avg genes. These avg genes display a significant homology at their promoter region as well as in their N-terminal encoding end to the corresponding regions of the vsp genes of the bovine pathogen M. bovis encoding Vsps [15]. Interestingly, the two species share a unique property of the lipoprotein signal sequence. While the presence of the Ala and Cys residues is consistent with a prokaryotic prolipoprotein signal peptidase recognition sequence [22], the presence of the Lys residue preceding the Cys residue within the lipoprotein box of four Avg proteins of M. agalactiae as well as in the Vsp lipoproteins of M. bovis is unique [15]. The Cys residue in the lipoprotein box is the only one occurring in each ORF and represents a predicted acylation site and point of anchorage of a mature processed prokaryotic lipoprotein [22].

In a previous study, we showed that synthetic oligonucleotides representing distinct repetitive sequences, comprising a significant portion of the vsp structural gene region, failed to react in Southern blot hybridization with the M. agalactiae avg genes [18]. Sequence analysis of four avg genes (avgA–D) cloned in this study, confirmed these observations. This indicates the lack of vsp-like repetitive domains on one hand and the complete absence of any avg-reiterated coding sequences on the other hand. A significant portion of the vsp genes is composed of repetitive coding sequences [15], thus enabling the Vsp molecules to undergo high-frequency size variation [16]. The lack of repetitive elements within four members of the avg gene family represents a fundamental difference between the two gene families, and suggests that at least four Avg proteins are not subject to size variation.

The second notable finding in this study emerged when avg profiles of sequential M. agalactiae isolates from a chronically infected animal during the period of about 7 months (Fig. 5) were compared. Hybridizations with avg specific oligonucleotide probes present clear evidence for genomic rearrangement of the 1.0-kb HindIII genomic fragment and of the 5.3-kb fragment), for instance at sampling times 4, 5 and 6 (Fig. 5, lanes 3–5). There is a possibility that animal #627 was simultaneously infected by several M. agalactiae strains differing in their avg profiles, and that the putative genomic rearrangement is an artifact resulting from isolation of ‘unrelated’ strains. In fact, previous studies have pointed out the pitfalls of selecting clonal variants within a highly variable population, and thereby obtaining a homogeneous population that does not reflect the authentic in vivo situation [23]. For this reason, we have purposely not cloned isolates from infected animals, in order to examine the predominant ‘strain’, which apparently results from in vivo selection within the animal host. Results with the avgA specific probe (a-1), which hybridized with a single genomic fragment, clearly indicate that isolates at each sampling time in animal #627 were homogeneous with respect to this avg genomic fragment (Fig. 5b).

Rearrangement events within the avg locus resemble a key feature of the Vsp family of M. bovis. Phenotypic switching of the M. bovis Vsp genes was shown to be linked with high-frequency chromosomal rearrangements occurring within the vsp locus [17]. Chromosomal rearrangements have been shown to be associated with phenotypic switching in many bacterial systems including in the mycoplasmas [13,24,25]. The highly homologous 5′ region common to all known avg genes may serve as potential recombination site for homologous recombination, gene conversions, DNA inversions and gene duplications [27]. Such mechanisms are frequently employed for regulating genes encoding surface antigens in other bacterial systems [12,14,26]. One or more of these mechanisms are likely to be involved in the regulation of the avg genes.

The variation in the avg genomic profiles observed among isolates of M. agalactiae isolated from chronically infected animals provide evidence demonstrating rearrangements of genes, potentially involved in mycoplasma antigenic variation, which occur in vivo in the natural host. Experiments are underway to fully characterize the avg genomic locus, to elucidate the precise nature of rearrangement events within it and to assess whether these genomic changes are associated with phenotypic variation.

Acknowledgements

The authors would like to express their appreciation to Dr. J.G. Tully (Frederick, MD, USA), Dr. G. Jones (Edinburgh, UK) and Dr. K. Sachse (Jena, Germany) for supplying specific antisera used for identification of M. agalactiae and M. bovis isolates. This work was supported by a Grant (IS-2540-95R) from the United States–Israel Binational Agricultural Research and Development Fund (BARD).