Virulence regions and virulence factors of the ovine footrot pathogen, Dichelobacter nodosus

Virulence regions and virulence factors of the ovine footrot pathogen, Dichelobacter nodosus

ELSEVIER FEMS Microbiology Letters 145 (1996) 147-156 MiniReview Virulence regions and virulence factors of the ovine footrot pathogen, Dichelobact...

975KB Sizes 0 Downloads 68 Views

ELSEVIER

FEMS Microbiology Letters 145 (1996) 147-156

MiniReview

Virulence regions and virulence factors of the ovine footrot pathogen, Dichelobacter nodosus Stephen J. Billington l, Joanne L. Johnston,

Julian I. Rood *

Department of Microbiology, Monash University, Clayton, Victoria 3168, Australia

Received 21 June 1996; revised 9 September 1996; accepted 11 September 1996

Abstract Ovine footrot is a debilitating and highly infectious disease that is primarily caused by the Gram-negative, anaerobic bacterium Dichelobacter nodosus. The major antigens implicated in virulence are the type IV fimbriae and extracellular proteases. The fimbriae show sequence and structural similarity to other type IV fimbriae, this similarity extends to genes that are involved in fimbrial biogenesis. Several acidic and basic extracellular serine proteases are produced by both virulent and benign isolates of D. nodosus. Subtle functional differences in these proteases appear to be important in virulence. In addition, there are two chromosomal regions that have a genotypic association with virulence. The partially duplicated and rearranged vup regions appear to have arisen from the insertion of a plasmid into a tRNA gene via an integrase-mediated site-specific insertion event. The 27 kb vrl region has several genes often found on bacteriophages and has inserted into an ssrA gene that may have a regulatory role in the cell. The determination of the precise role that each of these genes and gene regions has in virulence awaits the development of methods for the genetic analysis and manipulation of D. nodosus. Keywords: Footrot;

Virulence;

Fimbria;

Protease;

vap region;

vrl region;

1. Introduction Dichelobacter nodosus is the essential causative agent of ovine footrot, a disease that is of major economic significance in temperate climates. The disease is characterised by a mixed bacterial infection that results in the creation of an anaerobic environment and the growth of the aerotolerant anaerobe D. nodosus in the lesion. The resultant infection leads to the separation of the horn of the hoof from the un* Corresponding author. Tel.: +61 (3) 9905 4825; Fax.: +61 (3) 9905 4811; E-mail: [email protected]

1 Present address: Department of Veterinary Science, University of Arizona, Tucson, AZ 85721, USA. 0378-1097/96/$12.00 Copyright PIISO378-1097(96)00379-5

0 1996 Federation

of European

Dichelobacter nodosus

derlying soft tissue, which results in lameness and loss of body condition [I]. Virulent, intermediate and benign footrot are three generally recognised forms of the disease which vary in severity depending on the nature of the causative D. nodosus isolate. D. nodosus is a relatively slow growing Gram-negative rod. The organism was previously known as ‘Bacteroides’ nodosus but comparative sequence analysis of its 16s rRNA revealed that it belonged in the gamma subgroup of the Proteobacteria and was more closely related to genera such as Escherichia and Pseudomonas than it was to members of the genus Bacteroides [2,3]. It was therefore placed into the new genus Dichelobacter, which means ‘rod of the cloven hoof, in the family Cardiobacteriaceae Microbiological

Societies. Published

by Elsevier Science B.V.

S.J. Billmgton

148

e’f ul. I FEMS

Class I

Miuohiolog)~

Lrttrm

145 (IYY6j

147-156

e-P timA

ffmB

c/p6

Class spcific genes \

/ Class II

+-

__ aroA

ffmA

-+ fimC

Fig. 1. Genetic arrangement of the fimbrial gene regions from class frames is shown by black arrows. Class specific genes are indicated

[3]. Although there has been a great deal of molecular work done on D. nodosus in recent years, most of these studies have been carried out in Escherichia coli and Pseudomonas aeruginosa since there are no defined or recombinant methods of genetic analysis available for D. nodosus. The objective of this review is to provide an overview of these molecular studies, with an emphasis on virulence-associated gene regions and virulence factors.

2. Polar type IV fimbriae The fimbriae of D. nodosus are long proteinaceous appendages composed of polymers of a single pilin subunit that is encoded by thefimA gene. They are highly immunogenic, with agglutination reactions involving the pilus antigen providing the basis for the classification of D. nodosus isolates into nine major serogroups, designated A-1 [4]. Vaccination with whole cells or purified fimbriae protects against disease, but this protection is serogroup specific [5]. Sequencing of the genes of fimbrial subunits from the nine major serogroups has allowed the division of D. nodosus isolates into two major classes based on structural variation within the FimA protein and the genetic organisation of the fimbrial gene region [6,7]. Class I isolates (serogroups A-C, E-G and I) contain an additional gene, jimB, downstream of the fimbrial subunit gene. By contrast, strains in class II (serogroups D and H) possess three additional genes, fimC, jimD and jimZ, adjacent to JimA (Fig. 1). TheJimB gene encodes a potential 29.5 kDa membrane protein, which is postulated to play a role in interactions with FimA during pilin export. The product of$mD, a protein of 45 kDa, may be functionally analogous to FimB in class II strains of D. nodosus [6]. The jimC gene product, a putative

ffmD

fimZ

c/p6

I and class 11 isolates of D. nodosus. The extent of the open reading

inner membrane protein, has sequence similarity to TraX, a protein required for Na-acetylation of the pilin subunit of the conjugative F pilus [8]. FimC may be involved in the acetylation of FimA subunits in class II strains, however, classical type IV fimbrial subunits from other bacteria are all N-methylated, not N-acetylated [9]. The third class II specific protein, FimZ, has 50% identity to the FimA proteins of class II strains [6]. FimZ has a typical type IV signal sequence and is thought to represent a redundant fimbrial subunit. However, FimZ monomers are not likely to be assembled into fimbriae in D. nodosus, since they are not assembly competent when expressed in P. aeruginosa (Johnston, J.L., Billington, S.J., Mattick, J.S. and Rood, J.I., unpublished results). In both classes, fimA is transcribed from rpoN-dependent promoters [6]. The class-specific genes appear to be cotranscribed with$mA on low level transcripts which read through the jimA transcriptional terminator to the end of the class-specific region. This region is defined by the end of the clpB gene, which encodes the regulatory subunit of an ATPdependent protease [6] (Fig. 1). The aroA gene, located upstream ofjmA, defines the other end of the class-specific gene regions (Fig. 1). The FimA subunits of class I and class II strains are highly homologous over the amino-terminal onethird of the protein [7]. This region contains a short, positively charged leader sequence and a hydrophobic domain that is highly conserved in the pilin subunits of all type IV fimbriate bacteria. Along with the presence of an N-methylphenylalanine residue as the N-terminal amino acid, a polar distribution, and an association with twitching motility, these characteristics provide the basis for the designation of the fimbriae into the type IV or N-MePhe group [9]. The class-specific fimbrial subunits diverge over the

S.J. Billington et al. IFEMS Microbiology Letters 145 (1996) 147-156

Dn

< .~I I

pilA

fimP

pi/B

pi/C

0RF197

pi/D ORFX

Pa

PP Ng& Nm

Ah

a-D

WPC

wpD ORFX

vv

Fig. 2. Organisation of the D. nodosus fimbrial gene region and similar gene regions from other bacteria. The direction and extent of the open reading frames are shown. Related genes are indicated by similar shading. The ORFX homologue from P. aeruginosa has only been partially sequenced. An ORFX homologue from A. hydrophila was identified from the nucleotide sequence downstream of tapD. vvpC, vvpD and ORFX from Vibrio vulnificus were identified from database searches with the related D. nodosus protein sequences. Dn (D. nodosus), Pa (P. aeruginosa). Pp (P. putida). Ng (N. gonorrhoeae), Nm (N. meningitidis). Ah (A. hydrophila), Vv (V vulniJicus).

remainder of the protein whereas class I subunits lack the characteristic carboxy-terminal disulfide bridge present in all other type IV fimbrial subunits 171. FimA monomers of D. nodosus are assembled into fimbriae when expressed in P. aeruginosa, suggesting that the fimbrial biogenesis systems of D. nodosus and P. aeruginosa are related [lo]. The fimbrial subunit of P. aeruginosa is cleaved and methylated prior to assembly by the type IV prepilin peptidase, PilD [9]. FimP, a D. nodosus homologue of PilD, processes D. nodosus prepilin to the mature form in E. coli and is probably responsible for processing of the FimA subunit in D. nodosus [l 11. FimP can also complement PilD for protein secretion and may form part of a protein export pathway in D. nodosus (Johnston, J.L., Billington, S.J., Haring, V. and Rood, J.I., unpublished results). Two additional genes, jimN and jim0, were identified directly upstream of jimP [l l] (Fig. 2). The product of the$mN gene shows similarity to a group of cytoplasmic nucleotide-binding proteins, including PilB from P. aeruginosa, which are involved in fimbrial assembly, protein export and DNA uptake [12].

149

The FimO protein belongs to a family of proteins thought to reside in the cytoplasmic membrane [12]. EpsE, a Vibrio cholerae homologue of FimN, is required for protein secretion and has been shown to interact with EpsL to form a complex that is associated with the cytoplasmic membrane [13]. It is likely that FimN may interact with FimO or an EpsL homologue in a similar manner in D. nodosus. Although FimN and FimO are thought to be functionally analogous to their P. aeruginosa homologues, PilB and PilC, complementation experiments in P. aeruginosa suggest that they lack the structural domains which are required for interaction in the P. aeruginosa fimbrial biogenesis system (Johnston, J.L., Billington, S.J., Haring, V. and Rood, J.I., unpublished results). The JimNOP gene region has a similar genetic organisation to the equivalent regions in P. aeruginosa and Aeromonas hydrophila (Fig. 2) except that in D. nodosus the fimbrial subunit gene is located at a different genomic location with respect to the accessory genes [14]. The JimN, Jim0 and jmP genes appear to comprise an operon that is transcribed from a 07’ type promoter located upstream of ORFM, a putative gene thought to be involved in tryptophan biosynthesis (Fig. 2). In P. aeruginosa, the pilBCD genes appear to have their own promoters, although piZC and pilD may be are cotranscribed. 0RF197, which is located downstream offimP, is also cotranscribed with jimNOP. In Neisseria gonorrhoeae, pilD and ORFX, a homologue of ORF197, are transcriptionally linked. This conservation of the organization of the fimbrial gene region provides evidence for horizontal gene transfer between these bacterial species

[ill. 3. Extracellular proteases studies have shown that isolates of D. footrot produce four extracellular acidic serine protease isoenzymes (Vl-V3 and V5, with pl values of 5.2-5.6) and a basic protease (BprV, pZ 9.5) [15,16]. Five acidic proteases are produced by isolates that cause benign footrot but these proteases (Bl-B5) may have different electrophoretic mobilities, are relatively more heat labile, and have decreased elastase activity on elastin agar Biochemical

nodosus that cause virulent

150

S.J. Billington

et al. IFEMS

Microbiology

_l_

Letters

145 (1996)

147-156

Integration of pJIR896 at the end of a tRNA gene ISI ,&

1.

.. . ... . ... . .

tRNA &A vapC v9E

. .. .. . . .. .. .

w_uQ vapC v9A vapD vapB orfl18 I

vapl my375 orfli7

IS1253 transposition/deletion and duplication of vap genes

. .. .. ..

..*... .. .. ..

.. .. .. .

tRNA id

vapH v9E vapG

v(s9 wpC v9A vapl vapA ‘VqpFvapE’ v9D vapB orfl I8 vapG

vap region

v9D

l/3

Fig. 3. Model for the evolution of vap region l/3. Our model for the evolution of vap region l/3 involves the integration of a plasmid such as pJIR896 (shown here with ORFs indicated by black arrows) at an attachment site within the 3’ end of a tRNA gene. Subsequent deletion or non-replicative transposition of IS1253 (indicated by the horizontal lines) results in at least the initial duplication event which gives rise to the duplicated region of hap region 113 (diagonal lines).

medium [l]. Benign isolates also secrete a basic protease, BprB, that is closely related to BprV (97% identity), but has a lower pl of 8.6 [ 17,181. The precise role of all of these enzymes in the disease process has not been determined but they are believed to be important virulence factors [ 11. The first protease gene from D. nodosus to be cloned and sequenced was the bpr V gene (1809 bp) from strain A198, which encodes the 603 residue basic protease precursor [ 151. Analysis of the deduced amino acid sequence revealed that it has recognisable amino-terminal pre- and pro-peptide do-

mains and a carboxy-terminal extension. All of these domains are cleaved to form the mature BprV protease of 344 amino acid residues. The prepeptide domain consists of a potential 21 amino acid hydrophobic signal sequence whereas the propeptide and carboxy-terminal domains have 111 and 127 residues, respectively. Comparative analysis revealed that BprV has 18-38% amino acid sequence identity to members of the subtilisin protease family although it is 70 residues longer [15]. It also contains conserved residues that have been shown to be catalytic in other subtilisins. BprV is of similar length

S.J. Billington et al. I FEMS Microbiology Letters 145 (1996) 147-156

and has significant sequence identity (49%) to the serine protease from Xanthomonas campestris. This similarity extends to the respective propeptides. The homologous gene, bprB, from the benign strain 305 has also been cloned and sequenced [ 181. The sequence of the acidic protease V5 was determined by direct amino acid sequencing and shown to have 64% amino acid sequence identity to BprV and 53% identity to the X cumpestris serine protease [19]. It also is a member of the subtilisin family. The genes, aprV5 and aprB5, encoding the V5 protease from strain Al98 and the homologous B5 enzyme from strain 305, respectively, were subsequently cloned and sequenced [16]. These genes have a similar structure to bprV in that they encode comparable prepropeptides and carboxy-terminal extensions. Comparison of aprV5 and aprB5 reveals that they have 99% identity, resulting in only two different amino acids, one in the propeptide and one in the mature protein. Biochemical analysis of the other proteases indicates that the Vl-V3 and Bl-B4 acidic proteases are closely related and distinct from the V5 and B5 enzymes [19-211. It therefore appears that there are three types of related extracellular proteases in D. nodosus, as typified by the V2, V5 and BprV proteases. Note that recent studies have shown that upr V5 is located approx. 1 kb upstream of bpr V (RiiIkin, M. and Stewart, D.J., personal communication).

4. Virulence-associated

gene regions

4.1. Isolation and relationship to virulence The previous sections of this review have dealt with virulence factors identified by their phenotypic association with virulence. The remainder of this review will deal with loci identified by their genotypic association with virulence in D. nodosus. As such these loci do not necessarily represent true virulence factors but gene regions that have an association with virulence and that can be used as an indicator of virulence. These loci were identified from comparative hybridization studies carried out on plasmid libraries from the virulent isolate A198, using labelled chromosomal DNA from strain Al98 and the prototype benign isolate C305 [22]. Two regions

151

of DNA, which appeared to be preferentially associated with virulent isolates of D. nodosus, were identified and subsequently designated as the vup (virulence-associated protein) regions [23] and the vrl or virulence-related locus [24]. Hybridization analysis of over 800 D. nodosus isolates has indicated the presence of vap sequences in 95% of virulent/high intermediate isolates and 88% of intermediate strains but in only 38% of low intermediate/ benign strains [22,25]. By contrast, vrl sequences are present in 77% of virulent/high intermediate strains, but in only 13 and 7% of intermediate and low intermediate/benign strains, respectively [22,25]. Therefore, while the vup sequences are present in the majority of virulent strains, the presence of vrl is a more likely indication that a D. nodosus isolate is virulent. Other workers have repeated the initial studies but each of the virulenceassociated probes identified [26] have been shown to be part of either the vup or vrl loci [25], suggesting that these loci represent the major differences between the genomes of virulent and benign D. nodosus strains. 4.2. The vap gene regions The vup regions are virulence-associated sequences present in single or multiple copies in the genomes of the majority of D. nodosus isolates [22]. In the strain Al98 chromosome there are three copies of the vap sequence, designated as vup regions l-3 [22,27]. Most vap-related studies have revolved around vup regions 1 and 3, which are closely linked on the Al98 chromosome [27]. Using a series of probes from across vup regions 1 and 3 it was shown that while there were small regions of DNA within the locus that were not virulence associated, these two regions did in fact encompass a single virulence associated locus of 11.8 kb [28]. Nucleotide sequencing of the entire vap region l/3 indicated that it had a genetic organization similar to that of prophages [28]. At the extreme left end of the locus is a putative integrase gene, intA, with similarity to a number of bacteriophage integrase genes. In addition, the non-virulence associated sequences to the left of intA show similarity to tRNA-ser genes, suggesting that the vap sequences have integrated into the 3’ end of a tRNA gene, a favoured site

S.J. Billington

152

et al. I FEMS

Microbiology

Letters

I45 (1996)

147-156

Strain Al 98 (virulent)

%_ %\

--._ %*

____---___--___--___--___--___--___--___--___--,I’ ___--/’ attl

--._ %_

*.

m ssrA

-\

‘*.,, virulence ‘, ‘\ ‘, ‘,

Strain C305 (benign)

related locus (27kb)

“*Y,

intervening

WV

region (3kb)

ssrA Fig. 4. Comparison of the W/ region of the virulent 13. nodosus strain A198 with the corresponding region from the benign strain C305. The 27 kb r,rl locus of virulent D. nodosus strains (shown in white) is substituted in benign strains with an unrelated region of DNA of approx. 3 kb (shown in black). Sequences to the right and left of vrl. and the benign intervening region, are indicated by the horizontal hatching. The sequence of the 3’ end of the WA gene, indicated by the wavy line, is determined by WI region in the virulent strain and by the intervening region in the benign strain, as indicated by the white and black arrowheads, respectively. The sequence at the left junction point between non-virulence associated DNA and m-1DNA is shown and the putative atfL site is indicated by the shading.

for the integration of bacteriophages. To complete the similarity to prophages, vap region l/3 is bound by 19 bp repeats with 89% identity, similar to the duplicated attachment (att) sequences of prophages

M. In addition to intA, 15 ORFs have been identified within vap region l/3, namely vapA-I, vapA’, vupG’, vupE’ and orflZ8 [23,28] (Fig. 3). It appears that duplication and rearrangements of a portion of vap region 1 have resulted in the generation of variant copies of vapA, vapG and vupE, designated as vupA’, vapG’ and vapE’, respectively [28], and a second identical copy of vupD within vup region 3 [27]. Despite the structural similarity between vup region l/3 and prophage genomes, many of the vap-encoded genes have similarity to genes encoded on bacterial plasmids. For example, VapD has strong similarity to the products of ORFS from the N. gonorrhoeae cryptic plasmid [22] and an ORF from an Actinoba-

cillus uctinotnyceterncomitans

rolling circle plasmid [27]. VapD has been shown to be expressed by D. nodosus isolates that carry vup sequences, but antibodies to VapD do not react with its N. gonorrhoeae homologue [23]. In addition, the nucleotide sequence of vupD shows similarity to a sequence on pTD 1, a cryptic plasmid from Treponema denticola. VapB and VapC show similarity to putative products encoded in the trbH region of the F plasmid [23], and putative plasmid maintenance proteins, VagC and VagD [29], encoded by the Salmonella dublin virulence plasmid (Cheetham, B.F. and Whittle, G., personal communication). The VapE and VapE’ proteins have similarity to the product of 0RF2 from the cyanobacterial plasmid pMA 1. Our recent database searches have revealed intriguing similarity between ORF118 and the killer protein HigB [30] which is associated with the plasmid Rtsl killer gene system. Interestingly, VapA, VapA’

S.J. Billington et al. IFEMS

Microbiology Letters 145 (1996) 147-154

and VapI, which all have amino acid sequence similarity (Cheetham, B.F., personal communication), all show similarity to the antidote protein of this system, HigA [30]. Examination of the sequence upstream of vapl indicates the presence of a second ORF encoding a protein with similarity to HigB (Katz, M.E. and Cheetham, B.F., personal communication). Therefore, it appears that vup region l/3 carries at least two complete copies of a plasmid maintenance function. The similarity of vup-encoded products to plasmid-encoded proteins, in addition to the presence of a putative origin of replication within vap region l/3 [28], suggests that the vup sequences may have evolved from the site-specific insertion of an integrative plasmid rather than a bacteriophage. This hypothesis has been given credence by the recent discovery that a region of DNA that is similar to vup region l/3 can replicate as a plasmid in D. nodosus strain AC3577 [31]. This plasmid, pJIR896 (Fig. 3), appears to be a circular form of vap region l/3 with a single att site, as would be expected prior to the insertion and duplication of the att site. However, pJIR896 lacks the duplicated vapA’, vapF, vupG’, vapE’ genes and the second copy of vapD. In place of these regions is a 1.7 kb putative insertion sequence, ISZ2.53, which has similarity to a family of unusual insertion sequences [31]. The presence of a putative IS element in the plasmid at the site where duplication is predicted to have occurred in vup region l/3 implicates IS1253 in this duplication event and thus in the evolution of the chromosomal sequence. IS1253 is found in strain Al98 at a site adjacent to the ompl locus, which is not associated with vup region l/3. It is postulated that vup region l/3 originated from the insertion of an integrative plasmid such as pJIR896 into the 3’ end of a tRNA gene. The subsequent deletion or transposition of IS1253 may have resulted in at least the initial duplication event, giving rise to the vapA’-vapE’ region and the second copy of vupD (Fig. 3). The arrangement of these genes in vap regions 1 and 3 suggests that this duplication was followed by rearrangements, resulting in vup region l/3 as present in strain Al98 [28]. The evolution of vup region 2 is more difficult to deduce as less information is available. The order and arrangement of genes in vap region 2 appears

153

similar to vup region 1 with the exception that vup region 2 has an additional 600bp sequence located between vapC and vapD [27]. It is clear that minor variations, including insertions or deletions, can occur between vap sequences of different strains [25]. It was originally proposed that vup region 2 arose from a duplication event [27]. However, it now seems likely that this region is derived from a separate integration event since vap region 2 and vap region l/3 are integrated within the 3’ end of different tRNA molecules [32]. 4.3. The vrl locus The vrl locus is a large, 27 kb region of virulenceassociated DNA [24,33], which is present in a single copy, primarily in virulent D. nodosus isolates, including strain Al98 [22]. The nucleotide sequences which lie to the left and right of vrl sequences in strain Al98 are not adjacent to each other in the benign strain C305 [33], or in eight other strains we have tested (Billington, S.J. and Rood, J.I., unpub: lished data), but are separated by an unrelated sequence of approx. 3 kb (Fig. 4). These results suggest that vrl may not have arisen in virulent strains by a simple insertion event, but by an exchange mechanism involving the replacement of intervening sequences present in benign strains. Examination of the sequence at the left junction of vrl provides some interesting clues as to the evolution of the vrl region. As for the vap sequences, there is some evidence that vrl may have arisen by a sitespecific integrase-mediated event. Immediately upstream of the left end of vrl is a putative attachment site, attL, with similarity to the attachment sites of various prophages [33] (Fig. 4). This attachment site is found within the 3’ end of a gene with similarity to the E. coli ssrA gene, which encodes a regulatory 1OSa RNA molecule [34-361. Like tRNA genes the 3’ ends of these genes are targets for the integration of bacteriophages. There is a sharp shift in %G+C content between the adjacent ssrA sequence (43%) and the left end of vrl (57%). Additional sequence data suggests that there are local regions within vrl which can have a G+C content as high as 75-80% (Billington, S.J., Huggins, A.S. and Rood, J.I., unpublished results), consistent with the presence of a greater number of restriction sites for enzymes with

154

S.J. Billington

et al. IFEMS

Microbiology

Letters

I45 (1996)

147-156

GC-rich recognition sequences. This differential G+C content suggests that the vrl locus has originated in an organism other than D. nodosus, which has a G+C content of 45%. While the attL site and the shift in G+C content at the left end of vrl favour a site-specific integration mechanism for the insertion of vrl, there is no corresponding attachment site at the right end of the sequence. This observation suggests that the end of the virulence-associated region may not be identical to the end of the integrated element. This conclusion is supported by the almost identical G+C content of vrl and non-vrl sequences at this end [33]. Sequence information from within the vrl locus has identified some ORFs with similarities to proteins encoded on bacteriophages or plasmids. Proteins that have similarity to the DEAH family of helicase-related proteins and DNA methylases as well as proteins involved in resistance to bacteriophage infection appear to be encoded within vrl (Billington, S.J., Huggins, A.S. and Rood, J.I., unpublished data). However, at this stage no ORFs have been identified that can be directly related to the virulence of D. nodosus isolates. Hybridization analysis of over 800 D. nodosus isolates has failed to identify a single isolate that contains the vrl region but not the vap sequences [22,25]. Although it is possible that these results simply mirror the higher incidence of vap-containing strains, they may reflect a more fundamental interaction between the vap and vrl sequences. It is possible either that the vap integrase is required for vrl insertion, since sequencing of the vrl locus has not yet not identified a specific vrl-encoded integrase and there is sequence similarity between the att sites at the end of the vrl and vap loci, or that other vap-encoded products may be needed for either the integration or maintenance of vrl.

proteases are major factors involved in the virulence of D. nodosus. However, definitive proof has not been obtained because of the lack of a defined system for the genetic manipulation of this organism. The role of the vup sequences in virulence has been difficult to assess for similar reasons. There is no clear evidence of similarity between the vup genes and virulence genes. It is possible that the association of the vup regions with virulent strains represents an example of coevolution, with the vup sequences playing no direct role in virulence but providing a genetic marker for virulent strains. The correlation between virulence and the presence of vrl is much stronger than that between virulence and the vup regions [22-251. However, the reason that the vrl region correlates with virulence remains unknown. The vrl locus may carry an as yet unidentified virulence factor or may have an indirect effect on genes involved in virulence. The later explanation appears more likely and a possible effector molecule is the ssrA gene. In E. coli, ssrA encodes a 1OSa RNA molecule whose structure has regulatory effects on several different genes. In particular, modification of the 3’ end of ssrA leads to induction of an alternative lon protease [34], increased repression of the lac and gal operons and immunity to various h derivatives [35,36]. In D. nodosus, the insertion of the vrl region also changes the 3’ end of the ssrA gene [33]. It is postulated that this alteration affects the expression of 1OSa RNA-regulated genes that are involved in virulence. This hypothesis will only be verified or disproved after the development of both transformation methods for D. nodosus and shuttle vectors suitable for the transfer of recombinant molecules between D. nodosus and E. coli. The identification of a plasmid in D. nodosus [31] hopefully represents the first important step in this process.

5. Future perspectives

Acknowledgments

Vaccines prepared against purified fimbriae will protect against infection by strains with homologous fimbrial subunits. There is also an excellent correlation between virulence and the presence of more heat-stable protease isoenzymes. These data strongly suggest that the type IV fimbriae and extracellular

The work carried out in this laboratory has been supported by the Australian Government and Australian woolgrowers through research grants from the Australian Wool Research and Development Corporation and the Australian Research Council. J.J. is the recipient of an Australian Postgraduate

S.J. Billington et al. IFEMS

Microbiology

Award. We thank Brian Cheetham, Margaret Katz and David Stewart for providing information prior to publication.

References

111Stewart,

D.J. (1989) Footrot of sheep. In: Footrot and Foot Abscess of Ruminants (Egerton, J.R., Yong, W.K. and Riffkin, G.G., Eds.), pp. 545, CRC Press, Boca Raton, FL. PI La Fontaine, S. and Rood, J.I. (1990) Evidence that Bacteroides nodosus belongs in subgroup gamma of the Proteobacteria not in the genus Eacteroides: Partial sequence analysis of a B. nodosus 16s ribosomal RNA gene. Int. J. Syst. Bacterial. 40, 154159. [31 Dewhirst, F.E., Paster, B.J., La Fontaine, S. and Rood, J.I. (1990) Transfer of Kingella indologenes (Snell and Lapage 1976) to the genus Suttonella gen. nov. as Suttonella indologenes comb. nov.; transfer of Bacteroides nodosus (Beveridge 1941) to the genus Dichelobacier gen. nov. as Diehelobacter nodosus comb. nav.; and assignment of the genera Cardiobacterium, Dichelobacter, and Suttonella to Cardiobacteriaceae fam. nav. in the gamma division of Proteobacteria based on 16s ribosomal ribonucleic acid sequence comparisons. Int. J. Syst. Bacterial. 40, 426433. [41 Claxton, P.D. (1989) Antigenic classification of Bacteroides nodosw. In: Footrot and Foot Abscess of Ruminants (Egerton, J.R., Yong, W.K. and RiBkin, G.G., Eds.), pp. 155-166, CRC Press, Boca Raton, FL. [51 Elleman, T.C. (1988) Pilins of Bacteroides nodosus: Molecular basis of serotypic variation and relationships to other bacterial pilins. Microbial. Rev. 52, 233-247. B.P., Cox, P.T., Livingstone, S.P., PI Hobbs, M., Dahymple, Delaney, S.F. and Mattick, J.S. (1991) Organization of the fimbrial gene region of Bacteroides nodosus: class I and class II strains. Mol. Microbial. 5, 543-560. B.P., 171 Mattick, J.S., Anderson, B.J., Cox, P.T., Dahyrnple, Bills, M.M., Hobbs, M. and Egerton, J.R. (1991) Gene sequences and comparison of the fimbrial subunits representative of Bacteroides nodosus serotypes A to I: class I and class II strains. Mol. Microbial. 5, 561-573. PI Firth, N. and Skurray, R.A. (1995) A protein family associated with filament biogenesis in bacteria. Mol. Microbial. 17: 1218-1219. [9] Strom, MS. and Lory, S. (1993) Structure-function and biogenesis of the type IV pili. Annu. Rev. Microbial. 47, 565-596. [lo] Mattick, J.S., Bills, M.M., Anderson, B.J., Dalrymple, B., Mott, M.R. and Egerton, J.R. (1987) Morphogenetic expression of Bacteroides nodosus timbriae in Pseudomonas aeruginosa. J. Bacterial. 169, 33-41. [l l] Johnston, J.L., Billington, S.J., Haring, V. and Rood, J.I. (1995) Identification of fimbrial assembly genes from Dichelobatter nodosus: evidence that f%nP encodes the type IV prepilin peptidase. Gene 161, 21-26. [12] Hobbs, M. and Mattick, J.S. (1993) Common components in the assembly of type 4 fimbriae, DNA transfer systems, fila-

Letters 145 (1996)

147-156

155

mentous phage and protein-secretion apparatus: a general system for the formation of surface-associated protein complexes Mol. Microbial. 10, 233-243. [13] Sandkvist, M., Bagdasarian, M., Howard, S.P. and DiRita, V.J. (1995) Interaction between the autokinase EpsE and EpsL in the cytoplasmic membrane is required for extracellular secretion in Vibrio cholerae. EMBO. J. 14, 16641673. [14] La Fontaine, S. and Rood, J.I. (1996) Physical and genetic map of the chromosome of Dichelobacter nodosus A198. Gene (in press). [15] Lilley, G.G., Stewart, D.J. and Kortt, A.A. (1992) Amino acid and DNA sequences of an extracellular basic protease of Dichelobacter nodosw show that it is a member of the subtilisin family of proteases. Eur. J. Biochem. 210, 13-21. [I61 RiBkin, M.C., Focareta, A., Edwards, R.D., Stewart, D.J. and Kortt, A.A. (1993) Cloning, sequence and expression of the gene (aprV5) encoding extracellular serine acidic protease V5 from Dichelobacter nodosus. Gene 137, 259-264. R., u71 Kortt, A.A., Caldwell, J.B., Lilley, G.G., Edwards, Vaughan, J. and Stewart, D.J. (1994) Characterization of a basic serine protease (~1~9.5) secreted by virulent strains of Dichelobacter nodosus and identification of a distinct, but closely related, proteinase secreted by benign strains. Biochem. J. 299, 521-525. WI Lilley, G.G., RiBkin, M.C., Stewart, D.J. and Kortt, A.A. (1995) Nucleotide and deduced protein sequence of the extracellular, serine basic protease gene (bprB) from Dichelobacter nodosus strain 305: comparison with the basic protease gene (bprV) from virulent strain 198. Biochem. Mol. Biol. Int. 36, 101-111. 1191 Kortt, A.A., Riffkin, M.C., Focareta, A. and Stewart, D.J. (1993) Amino acid sequence of extracellular acidic protease V5 of Dichelobacter nodosus, the causative organism of ovine footrot. Biochem. Mol. Biol. Int. 29, 989-998. m Kortt, A.A., Bums, J.E., Vaughan, J.A. and Stewart, D.J. (1994) Purification of the extracellular acidic proteases of Dichelobacter nodosus. B&hem. Mol. Biol. Int. 34, 1157-l 166. Pll Kortt, A.A. and Stewart, D.J. (1994) Properties of the extracellular acidic proteases of Dichelobacter nodosus. Stability and specificity of peptide bond cleavage. Biochem. Mol. Biol. Int. 34, 1167-1176. P21 Katz, M.E., Howarth, P.M., Yong, W.K., RiBkin, G.G., Depiazzi, L.J. and Rood, J.I. (1991) Identification of three gene regions associated with virulence in Dichelobacter nodosus, the causative agent of ovine footrot. J. Gen. Microbial. 137, 2117-2124. 1231 Katz, M.E., Strugnell, R.A. and Rood, J.I. (1992) Molecular characterization of a genomic region associated with virulence in Dichelobacter nodosus. Infect. Immun. 60, 45864592. ~241 Rood, J.I., Wright, C.L., Hating, V. and Katz, M.E. (1994) Molecular analysis of virulence associated gene regions from the ovine footrot pathogen, Dichelobacier nodosus. In: Molecular Mechanisms of Bacterial Virulence (Kado, C.J. and Crosa, J.H., Eds.), pp. 615624, Kluwer, Dordrecht. [251 Rood, J.I., Howarth, P.A., Haring, V., Billington, S.J., Yong, W.K., Liu, D., Palmer, M.A., Pitman, D.A., Links, I., Stewart, D.A. and Vaughan, J.A. (1996) Comparison of gene

156

[26]

[27]

[28]

[29]

[30]

S.J. Billington et ul. IFEMS

Muohiology

probe and conventional methods for the differentiation of ovine footrot isolates of Dichelobacter nodosus. Vet. Microbiol. (in press). Liu, D. and Yang, W.K. (1993) Dichelobucter nodosus: differentiation of virulent and benign strains by gene probe based dot blot hybridization. Vet. Microbial. 38, 71-79. Katz, M.E., Wright, C.L., Gartside, TX, Cheetham, B.F.. Doidge, C.V., Moses, E.K. and Rood, J.I. (1994) Genetic organization of the duplicated rup region of the Dichelohnctrr nodosus genome. J. Bacterial. 176. 266332669. Cheetham, B.F., Tattersall, D.B., Bloomfield, G.A., Rood. J.1. and Katz, M.E. (1995) Identification of a bacteriophagerelated integrase gene in a mp region of the genome of D nodosus. Gene 162, 53-58. Pullinger, G.D. and Lax, A.J. (1992) A Sulmoneila dub/in virulence plasmid locus that affects bacterial growth under nutrient-limited conditions. Mol. Microbial. 6, 1631-1643. Tian, Q.B., Ohnishi, M., Tabushi, A. and Terawaki, Y. (1996) A new plasmid-encoded proteic killer gene system: cloning, sequencing and analyzing locus of plasmid Rtsl. Biochem. Biophys. Res. Commun. 220, 280-284.

Letters 145 (1996) 147 156

[31] Billington, S.J., Sinistaj, M.. Cheetham, B.F., Moses, E.K.. Katz, M.E. and Rood, J.I. (1996) Identification of a native Dichelobuctrr nodosus plasmid and implications for the evolution of the vup regions. Gene 172, 111-l 16. [32] Cheetham, B.F. and Katz. M.E. (1995) A role for bacteriophages in the evolution and transfer of bacterial virulence determinants. Mol. Microbial. 18, 201-206. [33] Haring, V., Billington, S.J., Wright, C.L., Huggins, AS.. Katz, M.E. and Rood, J.I. (1995) Delineation of the virulence related locus vrl of Dichelobacter nodosus. Microbiology 141. 2081~2091. [34] Kirby, J.E., Trempy, J.E. and Gottesman, S. (1994) Excision of a P4-like cryptic prophage leads to Alp protease expression in Escherichia roli. J. Bacterial. 176, 2068-2081. [35] Retallack, D.M., Johnson, L.L. and Friedman, D.l. (1994) Role for IOSa RNA in the growth of lambda-P22 hybrid phage. J. Bacterial. 176, 2082-2089. [36] Retallack, D.M. and Friedman, D.I. (1995) A role for a small stable RNA in modulating the activity of DNA-binding proteins. Cell 83. 227-235.