J Med Microbiol NEW Faster Access
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via CrossRef
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hess, J. F.
Right arrow Articles by Angelos, J. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hess, J. F.
Right arrow Articles by Angelos, J. A.
Agricola
Right arrow Articles by Hess, J. F.
Right arrow Articles by Angelos, J. A.
J Med Microbiol 55 (2006), 443-449; DOI: 10.1099/jmm.0.46366-0
© 2006 Society for General Microbiology
ISSN 0022-2615

The Moraxella bovis RTX toxin locus mbx defines a pathogenicity island

John F. Hess1 and John A. Angelos2

Department of Cell Biology and Human Anatomy, School of Medicine, University of California, 3301 Tupper Hall1 and Department of Medicine and Epidemiology, School of Veterinary Medicine, University of California, 2108 Tupper Hall,2 Davis, CA 95616, USA

Correspondence
John A. Angelos
jaangelos{at}ucdavis.edu

Received 8 October 2005
Accepted 6 December 2005

To characterize flanking regions of the mbx operon in Moraxella bovis, DNA surrounding mbxCABDtolC was sequenced in haemolytic and nonhaemolytic strains of M. bovis. In two haemolytic strains of M. bovis, the mbx operon, including the adjacent M. bovis tolC orthologue, was flanked by approximately 700 bp imperfect repeats. Nonhaemolytic strains of M. bovis had only one or no such repeats, as well as ORFs identical to those flanking the repeats from haemolytic M. bovis. Two nonhaemolytic strains also contained ORFs with deduced amino acid sequence similarity to bacterial araJ genes. The G+C content of the mbxCABDtolC gene region was lower than the flanking regions. The genetic organization and G+C content of mbxCABDtolC genes, and flanking repeats in haemolytic M. bovis, as well as the presence or absence of flanking repeats in nonhaemolytic M. bovis, suggests that this RTX operon is located on a mobile genetic element, and supports the designation of this region as a pathogenicity island, which is believed to be the first such element demonstrated in M. bovis.

Abbreviations: PAI, pathogenicity island.

The GenBank/EMBL/DDBJ accession number for the complete RTX operon and flanking sequences comprising PAI ITifton I is AF205359.


   INTRODUCTION
TOP
 INTRODUCTION
METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
Moraxella bovis, a Gram-negative bacterium of the family Moraxellaceae, has been recognized as the aetiological agent of infectious bovine keratoconjunctivitis (‘pinkeye’; Henson & Grumbles, 1960), the most common ocular disease of cattle. Pathogenicity of M. bovis is associated with expression of pilin, which is essential for adhesion of bacteria to the corneal epithelium (Annuar & Wilcox, 1985; Lehr et al., 1985; Moore & Rutter, 1989), and a pore-forming cytotoxin (Clinkenbeard & Thiessen, 1991) that promotes the development of corneal ulcers by lysis of corneal epithelial cells (Beard & Moore, 1994) and host neutrophils, with subsequent leakage of neutrophil-derived degradative enzymes into the corneal stroma (Kagonyera et al., 1988, 1989; Rogers et al., 1987). Amino acid sequencing of a partially purified cytotoxin from M. bovis culture supernatants led to the identification of a gene encoding an ~100 kDa protein with amino acid sequence similarity to RTX (repeats in the structural toxin) toxins from human and animal pathogens (Angelos et al., 2001). These related toxins included: Escherichia coli HlyA (Felmlee et al., 1985), Mannheimia haemolytica LktA (Lo et al., 1987), Actinobacillus pleuropneumoniae ApxIIA (Chang et al., 1989) and Actinobacillus actinomycetemcomitans LtxA (Lally et al., 1989). We subsequently identified and sequenced genes encoding M. bovis RTX C, B and D proteins that together formed a classic RTX operon, designated the mbx operon, in the genome of haemolytic M. bovis (Angelos et al., 2003).

The hly operon of uropathogenic E. coli is currently the most well characterized RTX operon. Typical RTX operons contain four genes arranged 5'-CABD-3'. The structural cytotoxin molecule is encoded by an rtxA gene. Activation of RTX A occurs by acylation of lysine residues mediated by the product of the rtxC gene (Stanley et al., 1994). Secretion of an activated RTX A toxin is facilitated by the combined actions of the RTX B and D proteins (Koronakis et al., 1992; Schlor et al., 1997; Schulein et al., 1994), as well as the secretion accessory protein, TolC (Wandersman & Delepelaire, 1990). Previously we demonstrated that M. bovis was similar to Bordetella pertussis (Glaser et al., 1988) in having a close genetic linkage between rtx genes and tolC orthologues (Angelos et al., 2003); in both organisms, tolC is located immediately downstream of rtx CABD genes.

Southern blotting of multiple nonhaemolytic M. bovis strains revealed an absence of DNA that hybridized to individual mbxC, A, B or D gene probes (Angelos et al., 2003). In addition, the characterization of DNA upstream of the mbx operon from haemolytic M. bovis strain Tifton I revealed an ORF with similarity to bacterial transposases. To further characterize the region flanking the mbx operon, and to identify possible mechanisms by which M. bovis could become nonhaemolytic, we sequenced DNA flanking this region in haemolytic and nonhaemolytic M. bovis.


   METHODS
TOP
 INTRODUCTION
 METHODS
RESULTS AND DISCUSSION
 REFERENCES
 
Bacterial strains and culture conditions. Haemolytic M. bovis Tifton I and T+, as well as nonhaemolytic M. bovis strains T–, SFS9b and SFS100b, were generously provided by Lisle George, University of California, Davis. Nonhaemolytic strain IBH63 was a field strain originally isolated by Pugh et al. (1966); nonhaemolytic strain NDL68 (991R) was obtained from the National Animal Disease Laboratory, Ames, Iowa, USA. Cultures were grown in Luria–Bertani (LB) broth containing 1·5 mM calcium chloride or in tryptic soy broth. Genomic DNA was prepared from each strain (Qiagen DNeasy kit). E. coli DH5{alpha} (Invitrogen) was the host for recombinant plasmids containing cloned regions of M. bovis genomic DNA and was grown on either LB agar or LB broth containing ampicillin (100 µg ml–1).

Determination of mbx operon flanking sequences. Genomic DNA flanking the M. bovis mbx operon was isolated using a PCR-based approach. Genomic DNA was restriction endonuclease digested with HindIII or HpaII and purified (Qiagen PCR purification kit). Purified, cleaved DNA (100 ng) was ligated to pBluescript II KS (–) (Stratagene) and was digested with HindIII (for ligation to HindIII digested M. bovis DNA) or ClaI and EcoRI (for ligation to HpaII digested M. bovis DNA). Ligation was performed at room temperature, overnight, using either T4 DNA ligase (Invitrogen) or LigaFast (Promega). Ligated DNA was amplified by PCR using a primer specific for the pBluescript II KS (–) vector (T3 or T7; Table 1Go) and a primer specific for M. bovis genomic sequences. The primer specific for M. bovis sequences was oriented to point away from the mbx operon. DNA sequences upstream from mbxCABDtolC were amplified using primers T3 and Nested. DNA sequences downstream of mbxCABDtolC were amplified using primers T7 and TolC dn. Successful amplification between known sequences and the pBluescript II KS (–) ‘tag’ resulted in the isolation of additional flanking DNA sequences. PCRs for the isolation of flanking DNA were performed using 25 cycles of the following temperatures and times: 94 °C for 30 s, 60 °C for 30 s, 72 °C for 3 min, followed by a final 15 min step at 72 °C.


View this table:
[in this window]
[in a new window]
  		Table 1. Primers used to amplify regions flanking mbxCABDtolC

 
Amplification of nonhaemolytic M. bovis was performed using several oligonucleotide primer pairs that flanked the mbxCABDtolC region of haemolytic M. bovis. The primer pair spanning the largest region, used for sequence determinations reported here, was Mdh dn and Ap up (see Table 1Go). Left and right repeats were amplified using oligonucleotide primers Tsp dn and PreC up, and the following conditions repeated for 25 cycles: 94 °C for 30 s, 60 °C for 30 s and 72 °C for 30 s. Sequencing of the amplified DNA was performed by the University of California, Davis, DNA sequencing laboratory or by Davis Sequencing. Final complete sequences were assembled from multiple overlapping fragments with DNA analysis software (MacDNASIS and Lasergene Seqman; Hitachi Software Engineering and DNAStar, respectively).


   RESULTS AND DISCUSSION
TOP
 INTRODUCTION
 METHODS
 RESULTS AND DISCUSSION
REFERENCES
 
Isolation and characterization of regions flanking mbxCABDtolC in haemolytic M. bovis

Approximately 1·5 kb of sequence was isolated flanking the 5' end of mbxC and approximately 2·5 kb of sequence was isolated flanking the 3' end of tolC. A map detailing regions 5' and 3' to mbxCABDtolC is shown in Fig. 1Go. The complete DNA sequence of the flanking regions and previously determined mbxCABDtolC genes has been deposited with GenBank, accession no. AF205359. The DNA sequence of flanking regions revealed several ORFs along with a pair of approximately 700 bp repeats flanking mbxC and tolC (left repeat and right repeat, respectively; Fig. 1Go).


Figure 1
View larger version (16K):
[in this window]
[in a new window]
  		Fig. 1. Schematic diagram depicting the arrangement of DNA elements flanking mbxCABDtolC within M. bovis Tifton I. The top section shows approximately 13 kb of DNA, with the mbxCABDtolC genes flanked by ORFs tentatively identified as malate dehydrogenase (mdh) and aminopeptidase (ap). The bottom section of the figure represents an enlargement of the ORFs, and flanking left and right repeats, with scale bars (bp) indicated below the respective left and right hand regions; the hatched and checkerboard bars above the scale bars indicate sequences also found in common with nonhaemolytic strains, as depicted in Fig. 2Go. The white triangle on the left side of the left repeat represents the sequence 5'-AAATCCT-3', the inverse complement of which, 5'-AGGATTT-3', is found on the right side of the right repeat, also indicated by a white triangle. Within the left and right repeats of M. bovis Tifton I, at position 27 is a sequence difference, C in the left repeat and T in the right repeat, indicated above them. The black triangle over the left repeat indicates an 80 bp insertion that is not present in the right repeat. With regard to this schematic, M. bovis T+ has an identical structure, with the exception of a matching T at position 27 of the left and right repeats.

 
At the extreme 5' end of the DNA sequence isolated, an ORF with an ATG at position 6 through to a TAA stop codon at position 698 is present. Comparison of the peptide sequence predicted by this ORF to the GenBank database revealed similarity to malate dehydrogenases from numerous bacteria. The best match was Polaromonas sp. JS666 with 78 % identity (accession ZP_00364926). BLAST results further revealed that the M. bovis ORF is not complete, but most likely represents only the carboxy terminal ~20 % of the coding sequence.

Downstream from the partial coding sequence of the putative malate dehydrogenase, the sequence of the left repeat begins at position 946 and extends to position 1687. The DNA sequence of the left repeat contains an insertion of 80 bases (1395–1474) relative to the right repeat. Contained within the sequence of the left repeat is a single ORF extending from the beginning of the repeat to a TAG termination codon at positions 1672–1674. BLAST alignment of the peptide sequence predicted by this ORF with the GenBank database revealed similarity with numerous sequences all identified as putative transposases, including IS1548, a transposase of Streptococcus agalactiae (accession no. NP_688610).

At the 3' end of mbxCABDtolC is the second direct repeat (right repeat), which extends from position 11 617–12 278, and is 80 nucleotides shorter than the left repeat. An ORF extends from position 11 617 to 12 063; as with the left repeat, the predicted peptide sequence also exhibits similarity to transposases. This is followed by an ORF (position 12 601 to the end of the cloned sequence) tentatively identified as encoding an aminopeptidase, based on deduced amino acid sequence similarity to aminopeptidases of Psychrobacter sp. (GenBank accession no. ZP_00262417) and B. pertussis (GenBank accession NP_881024). Aminopeptidases have been reported to be synthesized by M. bovis (Frank & Gerber, 1981).

In addition to the presence of the repeats flanking mbxCABDtolC, a perfect 7 bp inverted repeat flanks the left and right repeats. On the 5' side of the left repeat is 5'-AAATCCT-3' and on the 3' side of the right repeat is the sequence 5'-AGGATTT-3'; these are indicated by the white triangles in Figs 1 and 2GoGo.


Figure 2
View larger version (31K):
[in this window]
[in a new window]
  		Fig. 2. The genomic organizations of M. bovis strains T+, T–, NDL68, IBH63, SFS9b and SFS100b between putative mdh and ap genes. (a) The relationship between M. bovis T+ and T–. The mbxCABDtolC region of T+ is indicated above thesequence in abbreviated form. (b) The organization of nonhaemolytic M. bovis strains NDL68, IBH63, SFS9b and SFS100b strains. The white triangles and black arrows delineate features as described for Fig. 1Go. Thehatched bar below each drawing indicates DNA sequence that is collinear with M. bovis Tifton I from positions 1 to 945 (as shown in Fig. 1Go). The checkerboard bar below each drawing indicates DNA sequence that is collinearwith M. bovis Tifton I (Fig. 1Go) from positions 12 278 (T– and NDL68), 12 534 (IBH63) and 11 179 (SFS9b, SFS100b) to 13 009. The araJ-like ORFs identified in strainsIBH63, SFS9b and SFS100b are oriented to the left. The asterisk above the araJ-like ORF in IBH63 indicates theapproximate location of a single nucleotide deletion that breaks the ORF. The white circle shown flanking the ORF inSFS9b and SFS100b indicates a 30 bp region that has no homology to M. bovis Tifton I sequences characterized thus far.

 
Genomic DNA from M. bovis T+, a haemolytic strain isolated from a naturally infected cow (Angelos et al., 2001) was also characterized by PCR. Individual amplification reactions with oligonucleotide primer pairs that were specific for each mbx operon gene were used to demonstrate the presence of each mbx operon gene and tolC. As expected from the haemolytic phenotype of M. bovis T+, PCR revealed the presence of mbxCABD genes as well as tolC. Further characterization by PCR revealed that the RTX operon was placed between the mdh and ap ORFs, identical to the genomic organization of M. bovis Tifton I. At the 5' end of the characterized region, the primers Mdh dn and MidC up#2 were used to produce an approximately 2 kb fragment. DNA sequencing of this fragment revealed near 100 % identity with the sequence determined for M. bovis Tifton I. The only changes were transitions from cytosine to thymidine at positions 455 and 972 of the Tifton I sequence.

At the 3' end of the M. bovis Tifton I characterized region, T+ DNA was amplified using oligonucleotides Ap up and TolC dn++, producing a nearly 3 kb fragment. This fragment was subject to DNA sequencing with multiple primers and the resulting sequence aligned with nearly 100 % identity to the corresponding region of M. bovis Tifton I. The only differences were a transition from cytosine to thymidine at position 11 627 and the reverse transition, from thymidine to cytosine at position 11 270. Both of these positions are located within the right repeat; the significance of the changes is unknown.

Characterization of nonhaemolytic M. bovis

Previously we reported the absence of hybridization between mbxCABD gene-specific probes and genomic DNA from nonhaemolytic M. bovis, and concluded that mbxCABD genes were deleted from nonhaemolytic M. bovis (Angelos et al., 2003). The existence of the repeats flanking mbxCABDtolC suggested that these genes might be absent because of mobilization of this region from the genome. To explore this possibility, genomic DNA from nonhaemolytic M. bovis was analysed by PCR using oligonucleotide primers within the putative malate dehydrogenase (mdh) and aminopeptidase (ap) sequences, as described above for M. bovis Tifton I. All nonhaemolytic M. bovis strains tested were positive for amplification by Mdh dn/Mdh up and Ap dn/Ap up primer pairs. Amplification of nonhaemolytic M. bovis T–, IBH63 and NDL68 genomic DNA with the Mdh dn/Ap up primer pair each produced a 2–3 kb amplicon; the amplified product from IBH63 was slightly larger than either T– or NDL68. Amplification of genomic DNA from SFS100b and SFS9b with Mdh dn and Ap up each produced a larger product of ~4 kb.

The identity of the amplified DNA was determined by DNA sequencing of PCR products using the original PCR primers as sequencing primers. After the initial DNA sequences were obtained, additional sequencing reactions were performed using sequencing primers internal to the amplified product. The DNA sequences of M. bovis T– and NDL68 had a single nucleotide difference, which corresponded to the same nucleotide difference that was found between the left and right repeats (at position 27; Fig. 2Go). Thus, while nonhaemolytic M. bovis T– lacks mbxCABDtolC, it does contain a repeat corresponding to the right repeat of M. bovis Tifton I and T+. DNA sequences flanking the single repeat are a composite of DNA sequences flanking the left and right repeats of haemolytic M. bovis, and are identical to malate dehydrogenase (5' side) and aminopeptidase (3' side) (Fig. 2aGo).

M. bovis NDL68 was identical to T–, except that the single remaining repeat contains a C at position 27. Thus, M. bovis NDL68 represents a situation wherein mbxCABDtolC were precisely deleted such that one repeat remains that contains sequence characteristics of both the left and right repeats found in haemolytic M. bovis Tifton I (Fig. 2bGo).

The amplified product from Mdh dn and Ap up primers with M. bovis strain IBH63 was ~2380 bp. The DNA sequences at the 5' end corresponded exactly to the putative malate dehydrogenase; identity between the amplified IBH63 sequences and Tifton I ended exactly at position 945 of the Tifton I sequence (left repeat starts at position 946). At the 3' end of the amplified sequence, IBH63 exactly matches the putative aminopeptidase of M. bovis Tifton I (positions 12 534–12 850); however, the intervening DNA had no sequence similarity to DNA flanking mbxCABDtolC in M. bovis Tifton I (Fig. 2bGo). Contained within this region were two ORFs each encoding a peptide of greater than 100 aa. A GenBank BLAST search of these ORFs versus the database revealed amino acid sequence similarity to numerous proteins identified as AraJ, a sugar efflux transporter. The best match was with the AraJ protein of ‘Mannheimia succiniciproducens’ (GenBank accession no. YP_088771).

M. bovis strains SFS9b and SFS100b produced the largest amplified product with the Mdh dn and Ap up primers. At the 5' end of the amplified product, DNA sequence analysis revealed the presence of malate dehydrogenase identical sequences continuing from the location of the Mdh dn primer downstream to Tifton I position 945, the location of the left repeat. At the 3' end of the amplified product, identity was found between SFS9b and SFS100b sequences, and the aminopeptidase sequence, extending upstream to include an entire right repeat and an additional ~450 bp 5' to the repeat. That is, a 1700 bp block of identity was identified between the 3' end of the SFS9b and SFS100b PCR products, and the M. bovis Tifton I sequence. Tifton I sequences from position 11 179 and downstream match SFS9b/SFS100b amplified sequences from position 1966 and downstream. The repeat amplified in SFS9b/SFS100b strains was determined to be a hybrid repeat, containing the 80 nt insertion characteristic of the left repeat, and T at position 27, characteristic of the Tifton I right repeat (Fig. 2bGo). Intervening between the sequences homologous to the left side (mdh region) and the right side (ap region) is a region of 1·1 kb with no similarity to the mbxCABDtolC region of M. bovis Tifton I. There was 98 % nucleotide sequence identity between this sequence and the corresponding region in IBH63. Likewise, this region contained an ORF; however, the ORF in SFS9b and SFS100b extended throughout the entire 1·1 kb, representing a 369 aa peptide that also showed homology to bacterial AraJ proteins. This peptide is predicted to represent an almost full-length AraJ (~93 % of ‘M. succiniciproducens’ AraJ).

Characterization of regions flanking the RTX operon of haemolytic and nonhaemolytic M. bovis has revealed an organization that supports the hypothesis that RTX genes within M. bovis are located on a mobile genetic element resembling a pathogenicity island (PAI), which therefore has been designated PAI ITifton I. Within haemolytic strains Tifton I and T+, we have identified repeated DNA sequences flanking mbxCABDtolC that are similar to bacterial transposases and are likely to delineate mobile elements. On each side of the repeated sequences, ORFs were identified that exhibited similarity to other bacterial gene products, and these ORFs were found in similar locations in the nonhaemolytic strains that we studied. Although isolation of these ORFs was not complete, we highly suspect that these coding regions represent bona fide M. bovis genes. Our data are consistent with a hypothesis that M. bovis Tifton I represents a typical haemolytic strain and that nonhaemolytic strains arise following deletion of mbxCABDtolC. Characterization of M. bovis T+ by PCR and limited DNA sequencing shows a gene organization identical to M. bovis Tifton I. It is possible that mbxCABDtolC will be identified at other chromosomal locations in other strains of haemolytic M. bovis and this would further support the mobility of this gene cluster.

The data presented support the following series of events by which RTX toxin gene acquisition could occur in M. bovis. First, nonhaemolytic M. bovis acquired the genes for RTX toxin production and secretion. This acquisition event could have been by insertion of the RTX genes (and associated tolC) into the locus we have identified in the Tifton I and T+ strains, or it is possible that RTX sequences were inserted into a different locus and later moved into the current location. The close association of mbxCABD and tolC between flanking repeats suggests that these genes were associated before they integrated into the regions flanked by the putative malate dehydrogenase and aminopeptidase coding sequences. Such close association between RTX toxin genes and necessary toxin efflux proteins has been described in B. pertussis (Glaser et al., 1988). Nonhaemolytic strains T– and NDL68 could have arisen by a single aberrant excision of the mobile element, yielding a genomic locus with only a single terminal repeat. Nonhaemolytic strains IBH63, SFS9b and SFS100b could have arisen by either a single aberrant deletion process that involved recombination with araJ sequences or a mbxCABDtolC deletion event followed by the introduction of araJ sequences.

The organized arrangement of genes for toxin production and secretion, bounded by direct repeats with ORFs exhibiting amino acid similarity to bacterial transposases, is similar to the genetic organization of a PAI. As described originally in E. coli, PAIs exhibit certain common features such as: (i) occupying large genomic DNA regions (>20 kb), (ii) carrying at least one virulence gene, (iii) insertion within or near a tRNA gene, (iv) containing direct repeats and mobility sequences, (v) having a G+C content different from the host bacteria (Schmidt & Hensel, 2004). An additional requirement of PAIs is that they be present in the pathogenic strain, but absent from non-pathogenic strains of the same species.

The ~11 kb region (positions 946–12 278) we describe is smaller than has normally been associated with PAIs (10–100 kb; Hacker & Kaper, 2000). However, the most well studied PAIs are those from uropathogenic E. coli, and such strains typically contain genes for adherence factors, cytotoxicity, iron uptake and serum resistance (Hacker et al., 1999). Because of its small size, the mbxCABDtolC region could be termed a pathogenicity islet, used to designate DNA regions that are too small to fit the classical definition of PAI (Schmidt & Hensel, 2004). In either case, island or islet, inasmuch as M. bovis MbxA (cytotoxin) is a requirement for pathogenesis, expression of mbxCABDtolC is expected to be essential. Indeed, previous investigators utilizing extracts of nonhaemolytic M. bovis have been unable to demonstrate disease or lytic effects with such nonhaemolytic strains (Beard & Moore, 1994; Gray et al., 1995; Hoien-Dalen et al., 1990). Our data indicate that deletions of the mbxCABDtolC region do occur and result in nonhaemolytic M. bovis. Similar deletions have been shown to occur in pathogenic E. coli, resulting in non-pathogenic phenotypes (Hacker et al., 1990).

BLAST searches of the DNA sequence of strain Tifton I within the mdh/mbxCABDtolC/ap region did not reveal any similarity to tRNA genes. In addition, searching the DNA sequence with the program TRNASCAN (http://www.genetics.wustl.edu/eddy/tRNAscan-SE/) did not reveal any tRNA genes and it appears that mbxCABDtolC PAI in M. bovis is not located near a tRNA gene.

The fourth characteristic cited above for characteristics of PAIs indicates the possession of mobility related sequences. This relates to the transposability of mobile genetic elements, and the ability of these elements to encode enzymes capable of recognizing specific sequences and catalysing breakage and rejoining of DNA strands during the excision of an element from one location and subsequent integration into another site. Flanking the mbxCABDtolC region are approximately 700 bp direct repeats; as we have noted, the left repeat is 80 nt longer than the right repeat. Other than this deletion, there is only 1 nt difference between the left and right repeats in Tifton I, and none in T+. BLAST searches of the left and right repeat nucleotide sequence revealed no similarity to other DNA sequences. However, comparison of the deduced amino acid sequences of translated left and right repeats revealed similarity to various documented transposase polypeptides. In addition, a short inverted repeat flanks the left (5'-AAATCCT-3') and right (5'-AGGATTT-3') repeats, and offers additional evidence that mbxCABDtolC is located on a mobile genetic element.

The last piece of evidence usually considered necessary to define a PAI is the G+C content of the proposed PAI as compared to non-PAI genomic DNA. The 5' flanking region, from the partial malate dehydrogenase coding region to the beginning of the left repeat, is 49 mol% G+C. The 3' flanking region, from the end of the right repeat to the end of the isolated DNA sequence is 44 mol% G+C. The entire mbxCABDtolC PAI from the beginning of the left repeat through to the end of the right repeat is 36 mol% G+C. Thus, the proposed PAI has a G+C content that is different from the flanking DNA, again suggesting this region represents a PAI.

With the exception of overall size and lack of involvement of tRNA genes, the mbxCABDtolC region conforms to general characteristics of a PAI. Characterization of the haemolytic M. bovis T+ strain and the related M. bovis T– strain, isolated in the laboratory, provides compelling evidence that the region is a mobile gene cluster and thus, is a PAI.

The presence of an insertion sequence associated with an RTX operon was previously reported in A. actinomycetemcomitans (He et al., 1999). In A. actinomycetemcomitans, a novel insertion sequence upstream of its RTX operon (ltx operon) is responsible for increased expression of leukotoxin in certain clinical isolates (He et al., 1999). Further analysis of multiple haemolytic M. bovis strains will be necessary to determine whether differences in the insertion sequences reported here are associated with differential expression of cytotoxin, and whether the PAI described between malate dehydrogenase and aminopeptidase sequences is found at different loci in other strains of pathogenic M. bovis.


   ACKNOWLEDGEMENTS
 
This work was supported by School of Veterinary Medicine Formula Funds, grant numbers AH-125 and 4979-H. During completion of this work, Dr Hess was supported by grant DAMD017-02-1-0664 from the US Army Medical Research and Material Command.


   REFERENCES
TOP
 INTRODUCTION
 METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
Angelos, J. A., Hess, J. F. & George, L. W. (2001). Cloning and characterization of a Moraxella bovis cytotoxin gene. Am J Vet Res 62, 1222–1228.[CrossRef][Medline]

Angelos, J. A., Hess, J. F. & George, L. W. (2003). An RTX operon in hemolytic Moraxella bovis is absent from nonhemolytic strains. Vet Microbiol 92, 363–377.[CrossRef][Medline]

Annuar, B. O. & Wilcox, G. E. (1985). Adherence of Moraxella bovis to cell cultures of bovine origin. Res Vet Sci 39, 241–246.[Medline]

Beard, M. K. & Moore, L. J. (1994). Reproduction of bovine keratoconjunctivitis with a purified haemolytic and cytotoxic fraction of Moraxella bovis. Vet Microbiol 42, 15–33.[CrossRef][Medline]

Chang, Y. F., Young, R. & Struck, D. K. (1989). Cloning and characterization of a hemolysin gene from Actinobacillus (Haemophilus) pleuropneumoniae. DNA 8, 635–647.[Medline]

Clinkenbeard, K. D. & Thiessen, A. E. (1991). Mechanism of action of Moraxella bovis hemolysin. Infect Immun 59, 1148–1152.[Abstract/Free Full Text]

Felmlee, T., Pellett, S. & Welch, R. A. (1985). Nucleotide sequence of an Escherichia coli chromosomal hemolysin. J Bacteriol 163, 94–105.[Abstract/Free Full Text]

Frank, S. K. & Gerber, J. D. (1981). Hydrolytic enzymes of Moraxella bovis. J Clin Microbiol 13, 269–271.[Abstract/Free Full Text]

Glaser, P., Sakamoto, H., Bellalou, J., Ullmann, A. & Danchin, A. (1988). Secretion of cyclolysin, the calmodulin-sensitive adenylate cyclase-haemolysin bifunctional protein of Bordetella pertussis. EMBO J 7, 3997–4004.[Medline]

Gray, J. T., Fedorka-Cray, P. J. & Rogers, D. G. (1995). Partial characterization of a Moraxella bovis cytolysin. Vet Microbiol 43, 183–196.[CrossRef][Medline]

Hacker, J. & Kaper, J. B. (2000). Pathogenicity islands and the evolution of microbes. Annu Rev Microbiol 54, 641–679.[CrossRef][Medline]

Hacker, J., Bender, L., Ott, M., Wingender, J., Lund, B., Marre, R. & Goebel, W. (1990). Deletions of chromosomal regions coding for fimbriae and hemolysins occur in vitro and in vivo in various extraintestinal Escherichia coli isolates. Microb Pathog 8, 213–225.[CrossRef][Medline]

Hacker, J., Blum-Oehler, G., Janke, B., Nagy, G. & Goebel, W. (1999). Pathogenicity islands of extraintestinal Escherichia coli. In Pathogenicity Islands and Other Mobile Virulence Elements, pp. 59–76. Edited by J. B. Kaper & J. Hacker. Washington, DC: American Society for Microbiology.

He, T., Nishihara, T., Demuth, D. R. & Ishikawa, I. (1999). A novel insertion sequence increases the expression of leukotoxicity in Actinobacillus actinomycetemcomitans clinical isolates. J Periodontol 70, 1261–1268.[CrossRef][Medline]

Henson, J. B. & Grumbles, L. C. (1960). Infectious bovine keratoconjunctivitis. I. Etiology. Am J Vet Res 21, 761–766.[Medline]

Hoien-Dalen, P. S., Rosenbusch, R. F. & Roth, J. A. (1990). Comparative characterization of the leukocidic and hemolytic activity of Moraxella bovis. Am J Vet Res 51, 191–196.[Medline]

Kagonyera, G. M., George, L. W. & Munn, R. (1988). Light and electron microscopic changes in corneas of healthy and immunomodulated calves infected with Moraxella bovis. Am J Vet Res 49, 386–395.[Medline]

Kagonyera, G. M., George, L. W. & Munn, R. (1989). Cytopathic effects of Moraxella bovis on cultured bovine neutrophils and corneal epithelial cells. Am J Vet Res 50, 10–17.[Medline]

Koronakis, V., Stanley, P., Koronakis, E. & Hughes, C. (1992). The HlyB/HlyD-dependent secretion of toxins by gram-negative bacteria. FEMS Microbiol Immunol 5, 45–53.[Medline]

Lally, E. T., Golub, E. E., Kieba, I. R., Taichman, N. S., Rosenbloom, J., Rosenbloom, J. C., Gibson, C. W. & Demuth, D. R. (1989). Analysis of the Actinobacillus actinomycetemcomitans leukotoxin gene. Delineation of unique features and comparison to homologous toxins. J Biol Chem 264, 15451–15456.[Abstract/Free Full Text]

Lehr, C., Jayappa, H. G. & Goodnow, R. A. (1985). Serologic and protective characterization of Moraxella bovis pili. Cornell Vet 75, 484–492.[Medline]

Lo, R. Y., Strathdee, C. A. & Shewen, P. E. (1987). Nucleotide sequence of the leukotoxin genes of Pasteurella haemolytica A1. Infect Immun 55, 1987–1996.[Abstract/Free Full Text]

Moore, L. J. & Rutter, J. M. (1989). Attachment of Moraxella bovis to calf corneal cells and inhibition by antiserum. Aust Vet J 66, 39–42.[Medline]

Pugh, G. W., Jr, Hughes, D. E. & McDonald, T. J. (1966). The isolation and characterization of Moraxella bovis. Am J Vet Res 27, 957–962.[Medline]

Rogers, D. G., Cheville, N. F. & Pugh, G. W., Jr (1987). Pathogenesis of corneal lesions caused by Moraxella bovis in gnotobiotic calves. Vet Pathol 24, 287–295.[Abstract]

Schlor, S., Schmidt, A., Maier, E., Benz, R., Goebel, W. & Gentschev, I. (1997). In vivo and in vitro studies on interactions between the components of the hemolysin (HlyA) secretion machinery of Escherichia coli. Mol Gen Genet 256, 306–319.[CrossRef][Medline]

Schmidt, H. & Hensel, M. (2004). Pathogenicity islands in bacterial pathogenesis. Clin Microbiol Rev 17, 14–56.[Abstract/Free Full Text]

Schulein, R., Gentschev, I., Schlor, S., Gross, R. & Goebel, W. (1994). Identification and characterization of two functional domains of the hemolysin translocator protein HlyD. Mol Gen Genet 245, 203–211.[Medline]

Stanley, P., Packman, L. C., Koronakis, V. & Hughes, C. (1994). Fatty acylation of two internal lysine residues required for the toxic activity of Escherichia coli hemolysin. Science 266, 1992–1996.[Abstract/Free Full Text]

Wandersman, C. & Delepelaire, P. (1990). TolC, an Escherichia coli outer membrane protein required for hemolysin secretion. Proc Natl Acad Sci U S A 87, 4776–4780.[Abstract/Free Full Text]




This article has been cited by other articles:


Home page
 jvdiHome page
J. A. Angelos and L. M. Ball
Differentiation of Moraxella bovoculi sp. nov. from other coccoid moraxellae by the use of polymerase chain reaction and restriction endonuclease analysis of amplified DNA
J Vet Diagn Invest, September 1, 2007; 19(5): 532 - 534.
[Abstract] [Full Text] [PDF]

Home page
 J. Bacteriol.Home page
T. E. Kehl-Fie and J. W. St. Geme III
Identification and Characterization of an RTX Toxin in the Emerging Pathogen Kingella kingae
J. Bacteriol., January 15, 2007; 189(2): 430 - 436.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via CrossRef
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hess, J. F.
Right arrow Articles by Angelos, J. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hess, J. F.
Right arrow Articles by Angelos, J. A.
Agricola
Right arrow Articles by Hess, J. F.
Right arrow Articles by Angelos, J. A.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
INT J SYST EVOL MICROBIOL J MED MICROBIOL MICROBIOLOGY J GEN VIROL ALL SGM JOURNALS