Genomic analysis of the erythromycin resistance element Tn5398 from Clostridium difficile

Kylie A. Farrow1, Dena Lyras1 and Julian I. Rood1

Bacterial Pathogenesis Research Group, Department of Microbiology, PO Box 53, Monash University, Victoria 3800, Australia1

Author for correspondence: Julian I. Rood. Tel: +61 3 9905 4825. Fax: +61 3 9905 4811. e-mail: Julian.Rood{at}med.monash.edu.au


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Clostridium difficile is a nosocomial pathogen that causes a range of chronic intestinal diseases, usually as a result of antimicrobial therapy. Macrolide-lincosamide-streptogramin B (MLS) resistance in C. difficile is encoded by the Erm B resistance determinant, which is thought to be located on a conjugative transposon, Tn5398. The 9630 bp Tn5398 element has been cloned and completely sequenced and its insertion site determined. Analysis of the resultant data reveals that Tn5398 is not a classical conjugative transposon but appears to be a mobilizable non-conjugative element. It does not carry any transposase or site-specific recombinase genes, nor any genes likely to be involved in conjugation. Furthermore, using PCR analysis it has been shown that isolates of C. difficile obtained from different geographical locations exhibit heterogeneity in the genetic arrangement of both Tn5398 and their Erm B determinants. These results indicate that genetic exchange and recombination between these determinants occurs in the clinical and natural environment.

Keywords: Erm determinants, conjugative transposons, mobilization

Abbreviations: DR, direct repeat; MLS, macrolide-lincosamide-streptogramin B

The GenBank accession number for the Tn5398 element and flanking sequence is AF109075.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Clostridium difficile is a Gram-positive, spore-forming, anaerobic bacterium that causes a range of chronic gastrointestinal syndromes, including antibiotic-associated diarrhoea and colitis. The most severe form of these infections is pseudomembranous colitis, a potentially lethal infection that primarily occurs in hospital patients that have been treated with antimicrobial agents such as cephalosporins, penicillins or macrolides (Kelly & LaMont, 1998 ). C. difficile is recognized as the major cause of nosocomial diarrhoea in the USA (Gorbach, 1999 ) and is a significant pathogen in both British (Wilcox, 1998 ) and Australian hospitals (Riley et al., 1995 ).

The C. difficile isolates that cause antibiotic-associated diarrhoea are usually not resistant to the antibiotic responsible for the onset of infection, suggesting that an important factor in pathogenesis is the elimination of the normal flora of the intestine by antibiotic therapy. The end result is the provision of an ecological niche for the germination and growth of C. difficile spores that originate in the hospital environment (Banerjee & LaMont, 2000 ). However, because of the strong association between antimicrobial therapy and the onset of C. difficile disease, antibiotic resistance in this organism has been extensively studied (Chow et al., 1985 ; Levett, 1988 ; Roberts et al., 1994 ; Wüst & Hardegger, 1988 ). In addition, studies have shown that resistance to clindamycin increases the risk of C. difficile-associated diarrhoea (Johnson et al., 1999 ).

Clindamycin and erythromycin, both of which are members of the macrolide-lincosamide-streptogramin B (MLS) group of antibiotics, have often been implicated in the onset of C. difficile-associated disease. The most common mechanism of resistance to these antibiotics involves N6-dimethylation of a specific adenine residue of the 23S rRNA molecule (Leclercq & Courvalin, 1991 ). This alteration of the antibiotic target site is invariably catalysed by an rRNA methyltransferase that is encoded by an erm gene. Numerous erm genes have been characterized and divided into distinct classes based on their level of sequence similarity (Roberts et al., 1999 ). In general, each of the classes is loosely associated with a particular bacterial genus, with the exception of the Erm B class of determinants, which have been detected in a wide variety of bacterial genera, indicating their potential for intergeneric transfer.

Hybridization analysis has indicated that MLS-resistant strains of C. difficile carry erm(B) genes (Farrow et al., 2000 ; Hächler et al., 1987 ). The Erm B determinant carried by C. difficile strain 630 has been shown to be transferred by a conjugation-like mechanism to C. difficile (Wüst & Hardegger, 1983 ), Staphylococcus aureus (Hächler et al., 1987 ) and Bacillus subtilis (Mullany et al., 1995 ). Transfer has been shown to occur in the absence of detectable plasmid DNA. The B. subtilis transconjugants could transfer the Erm B determinant back to C. difficile, with integration of the determinant occurring at a specific site on the C. difficile chromosome. By contrast, integration was not site-specific in B. subtilis. Because of these observations it was proposed that the Erm B determinant from C. difficile resides on a conjugative transposon, Tn5398 (Mullany et al., 1995 ). This element has not been completely sequenced or characterized, although we have shown (Farrow et al., 2000 ) that it carries two identical erm(B) genes that are separated by a copy of the direct repeat (DR) sequence that is found on either side of the erm(B) gene from Clostridium perfringens (Berryman & Rood, 1995 ). These C. perfringens DR sequences are two directly repeated segments of DNA that primarily consist of an ORF, ORF298, which shows low levels of identity at the amino acid level to Soj and ParA proteins, flanked by highly palindromic sequences, palA and palB. In C. difficile strain 630 the two erm(B) genes are separated by a copy of the DR sequence and are flanked downstream by a variant of this DR sequence from which ORF298 has been deleted and flanked upstream by a DR variant that contains only the sequence downstream of palB (Farrow et al., 2000 ).

The aim of this research was to characterize the putative transposon and to examine the arrangement and distribution of both the erm(B) genes and Tn5398 in C. difficile strains from diverse geographical locations. The results showed that there was considerable genetic heterogeneity in the organization of the erm(B) gene region and that Tn5398 was an unusual genetic element in that it did not contain any discernible recombinase or mobilization genes.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Strains, growth conditions and molecular methods.
All C. difficile strains (Table 1) were grown in BHIS medium (Smith et al., 1981 ), supplemented with 50 µg erythromycin ml-1 or 20 µg rifampicin ml-1, at 37 °C in an anaerobic glove chamber (Coy Laboratories) in an atmosphere of 80% N2, 10% H2, 10% CO2. DNA was prepared from C. difficile strains by dye buoyant density gradient ultracentrifugation at 260000 g for 20 h at 20 °C (Abraham & Rood, 1985 ). Recombinant strains were derivatives of Escherichia coli DH5{alpha} (BRL) and were grown in 2x YT medium (Miller, 1972 ). Plasmid DNA was purified from E. coli cells by a modified mini alkaline lysis/polyethylene glycol precipitation procedure (Applied Biosystems). Unless otherwise stated molecular manipulations were carried out as described by Sambrook et al. (1989) .


View this table:
[in this window]
[in a new window]
 
Table 1. Bacterial strains and plasmids

 
Cloning of Tn5398 from C. difficile.
Chromosomal DNA from C. difficile strain 630 was digested overnight with XbaI and ligated at 15 °C to XbaI-digested DNA from the low-copy-number E. coli plasmid vector pWSK29 (Wang & Kushner, 1991 ). E. coli DH5{alpha} transformants were selected on medium containing erythromycin (150 µg ml-1).

DNA sequencing and computer analysis.
DNA sequencing was carried out using the ABI Prism Big Dye Terminator Cycle Sequencing Ready Reaction kit and analysed using an Applied Biosystems 373 DNA sequencer. Both DNA strands were sequenced using sequence-specific oligonucleotide primers. Nucleotide and amino acid comparisons were accomplished using the National Center for Biotechnology Information BLAST server at http://www.ncbi.nlm.nih.gov/BLAST/. The SOSUI program developed by the Mitaku Group, Department of Biotechnology, Tokyo University of Agriculture and Technology, Japan (http://sosui.proteome.bio.tuat.ac.jp/cgi-bin/sosui.cgi?/sosui_submit.html) was used to predict the structure of putative transmembrane proteins.

DNA dot blots.
DNA dot blots were performed by transferring 8 µg chromosomal DNA from each respective strain to a Hybond-N+ (Amersham Pharmacia Biotech) nylon membrane using a Minifold I Dot Blotter (Schleicher & Schuell). The DNA was cross-linked to the membrane by exposure to UV light at 312 nm for 3 min. The DNA was denatured by prehybridization for a minimum of 3 h in a solution containing SDS. The blots were then probed at high stringency with digoxigenin (DIG)-labelled probes prepared by PCR with the oligonucleotide primers listed in Table 2 as follows. The probes included a 688 bp erm(B)-specific probe (PCR primers: 2980 and 2981), a 399 bp ORF298-specific probe (4538 and 4451), a 339 bp palA-specific probe (4191 and 4537), a 984 bp ilvD-specific probe (6018 and 6278), a 933 bp hydD-specific probe (6339 and 6940), a 792 bp ORF13-specific probe (6019 and 6785), a 1166 bp effD-specific probe (9069 and 10237) and a 1124 bp ispD-specific probe (11546 and 11864). Following high stringency washes the blots were developed using the chemiluminescent substrate CDP-Star (Roche Molecular Biochemicals).


View this table:
[in this window]
[in a new window]
 
Table 2. Sequences of oligonucleotide primers

 
PCR conditions.
PCR analysis of C. difficile strains was carried out on a Perkin Elmer GeneAmp PCR System 2400. Each reaction contained 0·24 µg chromosomal DNA, 50 µM mixed deoxynucleotide triphosphates, 1x PCR buffer (Roche Molecular Biochemicals), 0·2 µM each oligonucleotide primer and 2·5 units Taq polymerase in a 100 µl final volume. Reactions were incubated for 1 cycle of 95 °C for 3 min, 70 °C for 1 min; 30 cycles of 95 °C for 1 min, 50 °C for 2 min, 72 °C for 3 min; 1 cycle of 50 °C for 2 min, 72 °C for 5 min and were then held at 4 °C. Following PCR, 10 µl each reaction was run on a 0·8% agarose gel to detect the PCR products.

Filter matings.
C. difficile cultures were grown overnight on BHIS agar at 37 °C. Single colonies of the donor and recipient were separately used to inoculate 20 ml BHIS broth and the cultures were grown until mid-exponential phase (OD600 ~0·45). The cells were then harvested by centrifugation at 1500 g for 10 min at room temperature and the cell pellets resuspended in 1 ml BHIS broth. Aliquots (100 µl) of the donor and recipient suspensions were mixed together on 0·45 µm pore-size nitrocellulose filters on BHIS agar. After incubation for 24 h at 37 °C, the filters were then removed, vigorously washed with 1 ml BHIS broth and 100 µl aliquots spread onto BHIS agar supplemented with the appropriate antibiotics and incubated for 48 h.


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Delineation and comparative analysis of Tn5398
To clone the entire Tn5398 element, XbaI-digested chromosomal DNA from C. difficile strain 630 was cloned into the low-copy-number vector pWSK29 and an erythromycin-resistant DH5{alpha} transformant was isolated and analysed. This strain carried a recombinant plasmid, pJIR1790, which contained a 19·5 kb insert. Approximately 13·5 kb of pJIR1790 was sequenced on both strands using a primer walking approach, beginning with primers within the Erm leader peptide and erm2(B) genes (Farrow et al., 2000 ). Analysis of the resultant data indicated that pJIR1790 carried a potentially novel genetic element.

In addition to duplicated erm(B)–ORF3 genes and ORF298, which were previously identified by Farrow et al. (2000) , the gene region contained nine other potential genes (Fig. 1). Upstream of the Erm B determinant two complete ORFs and one partial ORF were detected. The latter appeared to encode a protein with 52% amino acid sequence identity to IlvD from B. subtilis (Sorokin et al., 1996 ) and had greater than 50% identity to IlvD proteins from many other organisms. IlvD is a dihydroxy-acid dehydratase that is involved in the synthesis of the amino acids isoleucine and valine.



View larger version (29K):
[in this window]
[in a new window]
 
Fig. 1. Genetic organization of Tn5398. (a) Schematic representations of Tn5398, as observed in the donor and transconjugant strains, shown to scale. ORFs and their direction of transcription, are represented as blocked arrows. The region encompassed by Tn5398 is represented by a cross-hatched box and is further indicated by the scale below the diagram. Regions encompassing DR sequences are indicated by black boxes. The location and direction of oligonucleotides used in the delineation of Tn5398 are shown above the diagram. The target site in the recipient strain CD37 is also shown. The location of each of the ends of the transposon and the target sequence is indicated. (b) The nucleotide sequences of the left and right ends of Tn5398 and of the target region. Nucleotides included in Tn5398 are indicated in bold and are underlined. The nucleotides that represent the Tn5398 target site in C. difficile strain CD37 are represented in bold italic type.

 
The potential gene product of the ORF located downstream of ilvD had 25% identity to a transcription regulator of the TetR family from Aquifex aeolicus (Deckert et al., 1998 ), 22% identity to a regulatory protein, IfeR, from Agrobacterium tumefaciens (Palumbo et al., 1998 ) and 19% identity to probable transcriptional regulators from Pseudomonas aeruginosa and Listeria innocua (Stover et al., 2000 ). The next ORF encoded a putative product with 22% identity to the hydrolase IpbD from Pseudomonas putida (Eaton & Timmis, 1986 ) and 21% to the PcbD hydrolase from Archaeoglobus fulgidus (Klenk et al., 1997 ). Therefore, these two ORFs were designated as hydR and hydD, respectively.

Downstream of the Erm B determinant there are three ORFs, designated ORF13, ORF9 and ORF7 (Fig. 1) because their putative products share 20, 43 and 20% amino acid sequence identity, respectively, to the equivalent ORFs from the conjugative transposon Tn916 from Enterococcus faecalis (Flannagan et al., 1994 ). The function of these ORFs in Tn916 is not known.

The effR and effD genes were also located downstream of the Erm B determinant (Fig. 1). EffR shows low-level (15–20%) identity to several repressor genes, including the repressor of the marRAB operon from Salmonella typhimurium, which is involved in multiple antibiotic resistance (Sulavik et al., 1997 ). EffD showed some similarity to integral membrane proteins from a variety of organisms, including the hyperthermophile Pyrococcus horikoshii (Kawarabayasi et al., 1998 ) and the alkaliphilic Bacillus sp. C-125 (Takami et al., 1999 ). Computer analysis using the membrane prediction program SOSUI suggested that EffD was a membrane protein with 12 membrane-spanning helices.

The putative protein encoded by the gene located 3' of ORF7, ispD, had 48–52% identity to intracellular serine proteases from B. subtilis (Koide et al., 1986 ), Bacillus amyloliquefaciens and Bacillus polymyxa (Surova et al., 1994 ; Takekawa et al., 1991 ). In B. subtilis the homologous protease ISP-1 has been postulated to have a critical role in sporulation, possibly through the turnover of intracellular proteins, the processing of spore coat protein precursors and the inactivation of transcarbamylase and several other enzymes (Koide et al., 1986 ).

The last ORF identified in this study was designated flxD because it appeared to encode a protein with 39–45% identity to flavodoxin proteins from Clostridium beijerinckii (Tanaka et al., 1974a ), Megasphaera elsdenii and Treponema pallidum (Fraser et al., 1998 ; Tanaka et al., 1974b ). Flavodoxins are low-molecular-mass proteins that function as electron transfer agents in a variety of microbial metabolic processes (Simondsen & Tollin, 1980 ).

Since ilvD, hydR, hydD, ispD and flxD were probably housekeeping genes we postulated that Tn5398 extended from a region downstream of hydD to the intergenic ORF7–ispD region (Fig. 1). To examine this hypothesis a series of DIG-labelled probes specific for ilvD, hydD, erm(B), ORF13, effD and ispD were used in dot blots to examine chromosomal DNA from the wild-type C. difficile erythromycin-resistant strain 630, the susceptible C. difficile recipient CD37 and four independently derived transconjugants. The latter were derivatives of CD37 and were the result of conjugative transfer of erythromycin resistance from strain 630. If a particular gene was of a generic or housekeeping nature, we would expect a gene-specific probe to bind to DNA from all of the strains tested. If the gene was located on Tn5398 we would expect the probe to hybridize with DNA from the wild-type and the transconjugants but not the recipient. The results showed that each of the predicted housekeeping genes hybridized to all of the strains tested (Fig. 2). By contrast, the erm(B), ORF13 and effD probes hybridized only to the wild-type and transconjugant strains, indicating that these genes are likely to be located within Tn5398.



View larger version (54K):
[in this window]
[in a new window]
 
Fig. 2. DNA dot blots of Tn5398 derivatives. Hybridization analysis was carried out on chromosomal DNA from strains 630 (A2), CD37 (A3) and the transconjugant strains JIR1162 (B1), JIR1164 (B2), JIR1181 (B3) and JIR1184 (B4) using ilvD-, hydD-, erm(B)-, ORF13-, effD-, ispD-specific probes.

 
A PCR-based strategy was used to more precisely delineate the potential transposon by sequentially moving across the sequenced gene region (Fig. 1, Table 3). Amplification of chromosomal DNA from the wild-type strain 630, the recipient strain CD37 and the four transconjugant strains, confirmed (Table 3) the results of the DNA dot blots and indicated that the erm(B) genes, ORF13, effR, effD, ORF9 and ORF7 were located within Tn5398. Based on the products that were amplified from strain 630, but not from strain CD37, the left end was localized to the region between oligonucleotides 6940 and 6604 and the right end was localized to the region between oligonucleotides 9493 and 11662 (Fig. 1).


View this table:
[in this window]
[in a new window]
 
Table 3. Delineation of the conjugative transposon Tn5398 by PCR analysis*

 
To precisely define the ends of the putative element, the regions encompassing the ends of the transposon were amplified from strain 630 and the transconjugants and sequenced. The target region from C. difficile strain CD37 was amplified using oligonucleotides 6260 and 12143 and was also sequenced. The results showed that the left end of Tn5398 was located in the intergenic space between hydD and the region encoding the Erm leader peptide, at a site 272 bp downstream from the hydD stop codon. The right end was shown to be within the coding sequence of ORF7, 84 bp upstream of the stop codon (Fig. 1). Both the right and left ends of the element were extremely AT-rich. The left end consisted of the palindromic sequence TTTTTATATAAAAA, while the right hand consisted of the imperfect palindrome TTTATAATAAAAA. These ends had no significant similarity to the ends of any known conjugative transposon in the databases. The target site in strain CD37 was in the intergenic space between the hydD and ispD genes, 227 bp downstream of the hydD stop codon and 208 bp upstream of the ispD start codon (Fig. 1). The target site was also extremely AT-rich and consisted of the palindromic sequence TTTTTATATAAAAA (Fig. 1), which is identical to the palindromic sequence located at the left end of the element and differs from the right end by the omission of one of the central A nucleotides. Based on these data it was concluded that Tn5398 was 9630 bp in length.

Variation in the genetic organization of erm(B) gene regions of C. difficile strains from different geographical locations
Prior to this study we had reported that the duplicated arrangement of the erm(B) genes in C. difficile strain 630 was novel when it was compared to Erm B MLS resistance determinants from other bacterial species (Farrow et al., 2000 ). To determine if this arrangement was common in C. difficile we analysed 27 erythromycin-resistant C. difficile strains from a range of geographical locations and clinical sources (Table 1). DNA dot blots were carried out on chromosomal DNA from each isolate using an erm(B)-specific probe. The results showed that nine of these strains, including all five of the Japanese isolates, did not carry an erm(B) gene (Fig. 3). Probing with an ORF298-specific probe showed that an additional four isolates contained an erm(B) gene but did not contain the complete DR sequence that contains ORF298 (Fig. 3). Subsequent analysis, using a probe specific for the palA-like sequence that is located on either side of ORF298, revealed that the French isolate, 685, which did not contain ORF298, had a DR sequence with an internal deletion (Fig. 3), as found previously in other bacteria (Berryman & Rood, 1995 ; Farrow et al., 2000 ).



View larger version (43K):
[in this window]
[in a new window]
 
Fig. 3. DNA dot blots on C. difficile isolates from different geographical locations. Chromosomal DNA from the strains indicated was probed with erm(B)-, ORF298- or palA-specific probes. Strain key: A1, 630; A2, CD37; A3, L289; A4, 662; A5, AM140; A6, AM480; B1, AM1180; B2, AM1182; B3, AM1185; B4, SGC0545; B5, B1; B6, KZ1604; C1, KZ1610; C2, KZ1614; C3, KZ1623; C4, KZ1655; C5, 660/2; C6, 685; D1, 24/5-507; D2, R5948; D3, J9/5602; D4, J9/5610; D5, J9/5627; D6, J9/4478; E1, J9p2/5644; E2, J9p2/5650; E3, J7/4224; E4, J7/4290; E5, B1/832.

 
Therefore, in addition to strain 630, these studies revealed 15 isolates that could carry variants of Tn5398. A series of eight PCR amplifications were conducted on each isolate to determine the arrangement of the Erm B determinant in these strains. The first reaction was designed to detect the presence of an erm leader peptide upstream of an erm(B) gene. If the arrangement was the same as in strain 630, we would expect a fragment of 610 bp. Four of the isolates had this profile (Table 4). The remaining isolates had a 388 bp product, which as revealed by sequence analysis, consisted of the same region but without the leader peptide sequence, as is found upstream of the C. perfringens erm(B) gene (Berryman & Rood, 1995 ) (Fig. 4).


View this table:
[in this window]
[in a new window]
 
Table 4. PCR analysis of the genetic organization of the Erm B determinants from C. difficile isolates from different geographical locations

 


View larger version (25K):
[in this window]
[in a new window]
 
Fig. 4. PCR analysis of the genetic organization of the Erm B determinants from C. difficile isolates from different geographical locations. The regions encompassed by each reaction are shown in relation to the arrangements of the Erm B determinants from C. difficile strain 630 (Fig. 1) and C. perfringens strain CP592 (Berryman & Rood, 1995 ). Results of the PCR analysis are shown in Table 4.

 
The remaining PCR reactions were designed to step sequentially across the erm(B) gene region and to detect the presence and location of the Tn5398 gene, effD. The combined results of these PCR experiments (Table 4) allowed us to divide the strains into five groups based on the arrangement of the erm(B) gene region (Fig. 5). The first group of nine isolates were resistant to erythromycin but did not contain an erm(B) gene (Fig. 5a). This group included all of the Japanese isolates, three of the Australian isolates and a British isolate. The second group of three isolates, from the UK, Belgium and the USA, had an erm(B) gene but did not have either complete or incomplete DR sequences (Fig. 5b). Eleven strains, nine of which were from the USA, had a complete DR sequence that was located downstream of the erm(B) gene (Fig. 5c). The two non-USA strains in this group also carried an effD gene but it was not associated with the erm(B) gene. Strain 685 was very similar except that it had an erm leader peptide, the erm(B) gene was followed by an incomplete DR sequence and the effD gene was associated with the erm(B) gene (Fig. 5d).



View larger version (27K):
[in this window]
[in a new window]
 
Fig. 5. The different arrangements of the Erm B determinants. A diagrammatic representation of each genetic organization is shown to scale. The isolates that have this organization are listed beside each diagram.

 
PCR primers designed to detect the presence of two complete DR sequences flanking an erm(B) gene, as is observed in the arrangement of the Erm B determinant from C. perfringens, were included in these experiments (Fig. 4). No product was amplified from any of the C. difficile strains, although a product was observed when DNA from C. perfringens strain CP592 was included as a positive control.

Three C. difficile isolates had the same arrangement as observed in strain 630 and appeared to have a complete copy of Tn5398. That is, they had two erm(B) genes, one located upstream of a complete DR sequence and the other upstream of an incomplete DR sequence, and had a genetically linked ORF13–effD gene region (Fig. 5e). This group was geographically diverse as it included two Australian isolates from different hospitals and a French isolate, as well as the prototype Swiss isolate, strain 630.


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Tn5398 appears to be a mobilizable but non-conjugative genetic element
Previous studies suggested that the erythromycin resistance determinant in C. difficile strain 630 was located on a conjugative transposon, Tn5398 (Mullany et al., 1995 ). Conjugative transposons are discrete DNA elements that are normally integrated into the bacterial chromosome and are characterized by their ability to encode their own movement from one bacterial cell to another by a process requiring cell to cell contact. Conjugative transposition involves excision of the element from the chromosome to form a non-replicating covalently closed circular intermediate, which can either integrate elsewhere in the genome or transfer by conjugation to another cell where it integrates into the recipient’s genome. To carry out these reactions conjugative transposons generally encode site-specific recombinases that are responsible for the excision and integration of the element and other proteins that are required for conjugation (Salyers et al., 1995 ).

Tn5398 does not appear to encode genes that are involved in either excision, integration or conjugation. We have shown that it is 9·6 kb in size and encodes two copies of erm(B)–ORF3 and one copy of ORF298 (Farrow et al., 2000 ). In addition, there is only one incomplete and four complete ORFs located within the putative transposon. The proteins encoded by the effR and effD genes are unlikely to be involved in the transposition of the putative transposon, as they appear to encode a potential efflux protein and its associated regulator. The remaining ORFs, ORF13, ORF9 and ORF7, have similarity to equivalent ORFs from the prototype conjugative transposon, Tn916. The ORF7 protein shows limited homology to sigma factors (Flannagan et al., 1994 ). It has been postulated to have a regulatory role in the mobility of Tn916 because in the presence of tetracycline, increased ORF7 expression leads to increased transcription of ORF7, ORF8, xis, int and other genes (Celli & Trieu-Cuot, 1998 ). The finding that the right end of Tn5398 is internal to ORF7 may have implications for the level of excision, transfer and integration of the element. After excision from the donor chromosome, Tn5398 would leave behind part of ORF7, resulting in an incomplete ORF7 gene in the circular intermediate. The result could be altered levels of transcription of other genes involved in transfer of the element. The end of ORF7 also appears to be the target sequence for the element in recipient strains such as CD37. It is this region of the ORF7 protein that has identity to the helix–turn–helix motif of sigma factors, which is involved in DNA binding. Fusion with this region may provide a selective advantage for recombination of the circular intermediate at the target site.

In Tn916 ORF9 has been predicted to be a putative transcriptional repressor; however, the role of this repressor in the mobility of Tn916 has not been determined (Celli & Trieu-Cuot, 1998 ). The ORF13 protein has no known role in the mobility of Tn916. If these Tn916 homologues have any role at all in the movement of the Tn5398 element, it would appear that they are most likely to encode proteins that are involved in the regulation of transposition events, rather than proteins that are involved in excision, mobilization, transposition or integration.

ORF298 is the only other ORF that could encode a protein involved in Tn5398 mobility. The putative ORF298 protein has some similarity to replication proteins and proteins from the ParA and Soj families. ParA and Soj proteins generally have a role in the partitioning of plasmids and chromosomes during the replication cycle (Easter et al., 1998 ; Sharpe & Errington, 1996 ). It appears unlikely that ORF298 has a role in either the excision or integration of Tn5398 but this possibility cannot be completely eliminated.

In addition to being capable of catalysing its own conjugative transposition, Tn916 is capable of enhancing the transfer of another homologous conjugative transposon that is co-resident in the cell (Flannagan & Clewell, 1991 ) and of mobilizing non-conjugative plasmids (Jaworski & Clewell, 1995 ; Showsh & Andrews, 1999 ). Based on its small size and our comparative analysis of the genes carried on Tn5398, we postulate that it is more likely to be a non-conjugative but mobilizable element rather than a conjugative transposon.

Mobilization of non-conjugative plasmids by Tn916 does not appear to be dependent on the presence of a functional mobilization, or mob, gene on the plasmid, but does require the presence of an origin of transfer, or oriT, sequence. It is postulated that the same protein or proteins involved in the nicking of the Tn916 circular intermediate at its oriT site nicks similar sequences present on co-resident mobilizable plasmids. Once nicked, the plasmid then assumes a relaxed form, which is capable of being transferred during conjugation (Showsh & Andrews, 1999 ).

In addition to Tn5398, strain 630 carries a second conjugative transposon, Tn5397 (Mullany et al., 1990 ), which is closely related to Tn916 (Roberts et al., 2001 ). Comparison of Tn5397 and Tn916 reveals that they have very similar conjugation regions (ORF15–ORF23), but different insertion and excision modules (Roberts et al., 2001 ). In addition, the oriT sites present on both elements are identical and the putative ORF23 mobilization proteins have greater than 90% identity. Analysis of Tn5398 reveals the presence of two potential oriT sites, which are located within the coding sequence of ORF298 and in the intergenic space between ORF3 and ORF13 (Fig. 6). The nic sites (TGGTGT) of these two potential oriT sites are identical to the nic sites found on Tn916 and Tn5397. We postulate that Tn5398 is excised from the chromosome either by TndX, the site-specific recombinase responsible for the excision of Tn5397, or by another large resolvase encoded on the chromosome of strain 630. The resultant circular intermediate of Tn5398 would then be nicked at one of the oriT sites by the same protein responsible for nicking the Tn5397 circular intermediate and subsequently transferred to a recipient cell by a Tn5397-dependent process. Once in the recipient the element may be either integrated into the chromosome by means of the TndX protein or by another large resolvase encoded on the chromosome of the recipient cell. Attempts to verify this hypothesis by using outward firing PCR primers to amplify the putative circular intermediate were unsuccessful. However, in the absence of a positive chromosomal control, little can be concluded from this experiment.



View larger version (17K):
[in this window]
[in a new window]
 
Fig. 6. Alignment of the oriT sites from Tn916, Tn5397 and the two potential oriT sites located on Tn5398. Regions of sequence identity are boxed in black. The nic site is marked with a black triangle.

 
Note that there is an alternative explanation for these results. It is possible that the region excised from the chromosome is much larger than that proposed and includes genes upstream of the proposed left end (ilvD, hydR and hydD) and downstream of the proposed right end (ispD, flxD). After transfer of this region, which contains Tn5398, it could be integrated into the recipient chromosome by RecA-dependent homologous recombination. The end result would be the integration of Tn5398 and some of the genes flanking the element. This hypothesis is consistent with the experimental data.

Tn5398 is present in C. difficile strains from diverse sources
To determine if Tn5398, with its duplicated erm(B) genes, was common in C. difficile we carried out a series of comparative PCR and dot blot studies. The results showed that the arrangement of the Erm B determinants in 15 C. difficile isolates from diverse sources was not the same, with four major genetic variants being detected. The simplest variant was that of a single erm(B) gene and the most complex was represented by the Erm B determinant carried by Tn5398 (Fig. 5). Tn5398 was present in strain 630, which was originally from Switzerland, and three other C. difficile isolates, AM1180 and AM1185, isolated from different Australian hospitals, and the French isolate 660/2. The identification of these strains provides evidence that the duplicated erm(B) gene organization carried by Tn5398 is widespread and is therefore likely to be transferred between C. difficile isolates.

In general, each of the genetic variants was represented by isolates from a wide variety of geographical sources, with the exception of the group that carried a single erm(B) gene flanked by a downstream complete DR sequence (Fig. 5c). This group of isolates, with the exception of strains 662 and B1, were all isolated in the USA. With one exception, AM140, these USA isolates were isolated from large outbreaks of diarrhoea that occurred in four hospitals located in different parts of the country (Johnson et al., 1999 ). Analysis of these strains by restriction digestion, PCR and PFGE had previously determined that these isolates were actually derivatives of the same strain (Johnson et al., 1999 ), which was referred to as the epidemic strain. The isolates were all highly resistant to clindamycin and were shown to account for approximately 30–66% of the C. difficile strains isolated at these hospitals during the period 1989 to 1992 (Johnson et al., 1999 ). We previously showed that each of these isolates contained an erm(B) gene (Johnson et al., 1999 ). Our finding that they contain Erm B determinants with the same genetic organization supports the conclusion that they are derivatives of an epidemic strain. It was previously suggested (Johnson et al., 1999 ) that the erm(B) gene present in these isolates was associated with Tn5398. This study reveals that this is unlikely as the Tn5398-specific gene, effD, was not detected in these isolates.

We previously proposed that the C. perfringens Erm B determinant, which consists of an erm(B) gene flanked by two complete DR sequences, represents the Erm B progenitor and that Erm B determinants in other bacteria probably evolved by homologous recombination events that removed part of the DR sequences (Berryman & Rood, 1995 ). The different genetic arrangements of the various C. difficile Erm B determinants observed in this study are all consistent with this hypothesis. However, despite the fact that many erm(B) genes are located on mobile genetic elements that can move freely between different bacterial species, C. perfringens still appears to be the only species that contains an erm(B) gene associated with two intact copies of the DR sequence.


   ACKNOWLEDGEMENTS
 
We thank the Australian National Health and Medical Research Council for its research support and Peter Mullany for helpful discussions. K.A.F. was the recipient of an Australian Postgraduate Award.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Abraham, L. J. & Rood, J. I. (1985). Molecular analysis of transferable tetracycline resistance plasmids from Clostridium perfringens. J Bacteriol 161, 636-640.[Medline]

Banerjee, S. & LaMont, J. T. (2000). Non-antibiotic therapy for Clostridium difficile infection. Curr Opin Infect Dis 13, 215-219.[Medline]

Berryman, D. I. & Rood, J. I. (1995). The closely related ermB-ermAM genes from Clostridium perfringens, Enterococcus faecalis (pAMß1), and Streptococcus agalactiae (pIP501) are flanked by variants of a directly repeated sequence. Antimicrob Agents Chemother 39, 1830-1834.[Abstract]

Borriello, S. P., Davies, H. A. & Barclay, F. E. (1988). Detection of fimbriae amongst strains of Clostridium difficile. FEMS Microbiol Lett 49, 65-67.

Celli, J. & Trieu-Cuot, P. (1998). Circularization of Tn916 is required for expression of the transposon-encoded transfer functions: characterization of long tetracycline-inducible transcripts reading through the attachment site. Mol Microbiol 28, 103-117.[Medline]

Chow, A. W., Cheng, N. & Bartlett, K. H. (1985). In vitro susceptibility of Clostridium difficile to new ß-lactam and quinolone antibiotics. Antimicrob Agents Chemother 28, 842-844.[Medline]

Deckert, G., Warren, P. V., Gaasterland, T. & 12 other authors (1998). The complete genome of the hyperthermophilic bacterium Aquifex aeolicus. Nature 392, 353–358.[Medline]

Easter, C. L., Schwab, H. & Helinski, D. R. (1998). Role of the parCBA operon of the broad-host range plasmid RK2 in stable plasmid maintenance. J Bacteriol 180, 6023-6030.[Abstract/Free Full Text]

Eaton, R. W. & Timmis, K. N. (1986). Characterization of a plasmid-specified pathway for catabolism of isopropylbenzene in Pseudomonas putida RE204. J Bacteriol 168, 123-131.[Medline]

Farrow, K. A., Lyras, D. & Rood, J. I. (2000). The macrolide-lincosamide-streptogramin B resistance determinant from Clostridium difficile 630 contains two erm(B) genes. Antimicrob Agents Chemother 44, 411-413.[Abstract/Free Full Text]

Flannagan, S. E. & Clewell, D. B. (1991). Conjugative transfer of Tn916 in Enterococcus faecalis: trans activation of homologous transposons. J Bacteriol 173, 7136-7141.[Medline]

Flannagan, S. E., Zitzow, L., Su, Y. A. & Clewell, D. B. (1994). Nucleotide sequence of the 18-kb conjugative transposon Tn916 from Enterococcus faecalis. Plasmid 32, 350-354.[Medline]

Fraser, C. M., Norris, S. J., Weinstock, G. M. & 30 other authors (1998). Complete genome sequence of Treponema pallidum, the syphilis spirochete. Science 281, 375–388.[Abstract/Free Full Text]

Gorbach, S. (1999). Antibiotics and Clostridium difficile. N Engl J Med 341, 1690-1691.[Medline]

Hächler, H., Berger-Bächi, B. & Kayser, F. H. (1987). Genetic characterization of a Clostridium difficile erythromycin-clindamycin resistance determinant that is transferable to Staphylococcus aureus. Antimicrob Agents Chemother 31, 1039-1045.[Medline]

Hayter, P. M. & Dale, J. W. (1984). Detection of plasmids in clinical isolates of Clostridium difficile. Microbios Lett 27, 151-156.

Jaworski, D. D. & Clewell, D. B. (1995). A functional origin of transfer (oriT) on the conjugative transposon Tn916. J Bacteriol 177, 6644-6651.[Abstract]

Johnson, S., Samore, M. H., Farrow K. A. & 9 other authors (1999). Epidemics of diarrhea caused by a clindamycin-resistant strain of Clostridium difficile in four hospitals. N Engl J Med 341, 1645–1651.[Abstract/Free Full Text]

Kawarabayasi, Y., Sawada, M., Horikawa, H. & 22 other authors (1998). Complete sequence and gene organization of the genome of a hyper-thermophilic archaebacterium, Pyrococcus horikoshii OT3. DNA Res 5, 55–76.[Medline]

Kelly, C. & LaMont, J. T. (1998). Clostridium difficile infection. Annu Rev Med 49, 375-390.[Medline]

Koide, Y., Nakamura, A., Uozumi, T. & Beppu, T. (1986). Cloning and sequencing of the major intracellular serine protease gene of Bacillus subtilis. J Bacteriol 167, 110-116.[Medline]

Klenk, H. P., Clayton, R. A., Tomb, J. F. & 48 other authors (1997). The complete genome sequence of the hyperthermophilic, sulphate-reducing archaeon Archaeoglobus fulgidus. Nature 390, 364–370.[Medline]

Leclercq, R. & Courvalin, P. (1991). Bacterial resistance to macrolide, lincosamide, and streptogramin antibiotics by target modification. Antimicrob Agents Chemother 35, 1267-1272.[Medline]

Levett, P. N. (1988). Antimicrobial susceptibility of Clostridium difficile determined by disc diffusion and breakpoint methods. J Antimicrob Chemother 22, 167-173.[Abstract]

Miller, J. H. (1972). Experiments in Molecular Genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Mullany, P., Wilks, M., Lamb, I., Clayton, C., Wren, B. W. & Tabaqchali, S. (1990). Genetic analysis of a tetracycline resistance element from Clostridium difficile and its conjugal transfer to and from Bacillus subtilis. J Gen Microbiol 136, 1343-1349.[Medline]

Mullany, P., Wilks, M. & Tabaqchali, S. (1995). Transfer of macrolide-lincosamide-streptogramin B (MLS) resistance in Clostridium difficile is linked to a gene homologous with toxin A and is mediated by a conjugative transposon, Tn5398. J Antimicrob Chemother 35, 305-315.[Abstract]

Nakamura, S., Yamakawa, K., Nakashio, S., Kamiya, S. & Nishida, S. (1987). Correlation between susceptibility to chloramphenicol, tetracycline and clindamycin, and serogroups of Clostridium difficile. Med Microbiol Immunol 176, 79-82.[Medline]

Palumbo, J. D., Kado, C. I. & Phillips, D. A. (1998). An isoflavonoid-inducible efflux pump in Agrobacterium tumefaciens is involved in competitive colonization of roots. J Bacteriol 180, 3107-3113.[Abstract/Free Full Text]

Riley, T., Codde, J. & Rouse, I. (1995). Increased length of hospital stay due to Clostridium difficile associated diarrhoea. Lancet 345, 455-456.[Medline]

Roberts, A. P., Johanesen, P. A., Lyras, D., Mullany, P. & Rood, J. I. (2001). Comparison of Tn5397 from Clostridium difficile, Tn916 from Enterococcus faecalis, and the CW459 tet(M) element from Clostridium perfringens shows that they have similar conjugation regions but different insertion and excision modules. Microbiology 147, 1243-1251.[Abstract/Free Full Text]

Roberts, M. C., McFarland, L. V., Mullany, P. & Mulligan, M. E. (1994). Characterization of the genetic basis of antibiotic resistance in Clostridium difficile. J Antimicrob Chemother 33, 419-429.[Abstract]

Roberts, M. C., Sutcliffe, J., Courvalin, P., Jensen, L. B., Rood, J. I. & Seppala, H. (1999). Nomenclature for macrolide and macrolide-lincosamide-streptogramin B antibiotic resistance determinants. Antimicrob Agents Chemother 43, 2823-2830.[Free Full Text]

Salyers, A., Shoemaker, N., Stevens, A. M. & Li, L. (1995). Conjugative transposons: an unusual and diverse set of integrated gene transfer elements. Microbiol Rev 59, 579-590.[Abstract]

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Sharpe, M. E. & Errington, J. (1996). The Bacillus subtilis soj-spoOJ locus is required for a centromere-like function involved in prespore chromosome partitioning. Mol Microbiol 21, 501-509.[Medline]

Showsh, S. A. & Andrews, R. E.Jr (1999). Analysis of the requirement for a pUB110 mob region during Tn916-dependent mobilization. Plasmid 41, 179-186.[Medline]

Simondsen, R. P. & Tollin, G. (1980). Structure–function relations in flavodoxins. Mol Cell Biochem 33, 13-24.[Medline]

Smith, C. J., Markowitz, S. M. & Macrina, F. L. (1981). Transferable tetracycline resistance in Clostridium difficile. Antimicrob Agents Chemother 19, 997-1003.[Medline]

Sorokin, A., Azevedo, V., Zumstein, E., Galleron, N., Ehrlich, S. D. & Serror, P. (1996). Sequence analysis of the Bacillus subtilis chromosome region between the serA and kdg loci cloned in a yeast artificial chromosome. Microbiology 142, 2005-2016.[Abstract]

Stover, C. K., Pham, X. Q., Erwin, A. L. & 28 other authors (2000). Complete genome sequence of Pseudomonas aeruginosa PAO1, an opportunistic pathogen. Nature 406, 959–964.[Medline]

Sulavik, M. C., Dazer, M. & Miller, P. F. (1997). The Salmonella typhimurium mar locus: molecular and genetic analyses and assessment of its role in virulence. J Bacteriol 179, 1857-1866.[Abstract]

Surova, I. A., Revina, L. P., Ianonis, V. V., Kolesnikova, L. A. & Stepanov, V. M. (1994). Primary structure of the intracellular serine proteinase from Bacillus amyloliquefaciens. III. Amino acid sequence of peptides obtained by hydrolysis with a Glu, Asp-specific proteinase. Reconstruction of the entire amino acid sequence of the proteinase. Bioorg Khim 20, 1310-1326.[Medline]

Takami, H., Nakasone, K., Hirama, C., Takaki, Y., Masui, N., Fuji, F., Nakamura, Y. & Inoue, A. (1999). An improved physical and genetic map of the genome of alkaliphilic Bacillus sp. C-125. Extremophiles 3, 21-28.[Medline]

Takekawa, S., Uozumi, N., Tsukagoshi, N. & Udaka, S. (1991). Proteases involved in generation of ß- and {alpha}-amylases from a large amylase precursor in Bacillus polymyxa. J Bacteriol 173, 6820-6825.[Medline]

Tanaka, M., Haniu, M., Yasunobu, F. & Mayhew, S. G. (1974a). The amino acid sequence of the Clostridium MP flavodoxin. J Biol Chem 249, 4393-4396.[Abstract/Free Full Text]

Tanaka, M., Haniu, M., Yasunobu, F., Mayhew, S. G. & Massey, V. (1974b). Correction of the amino acid sequence of Peptostreptococcus elsdenii flavodoxin. J Biol Chem 249, 4397.[Free Full Text]

Wang, R. F. & Kushner, S. R. (1991). Construction of versatile low-copy-number vectors for cloning, sequencing and gene expression in Escherichia coli. Gene 100, 195-199.[Medline]

Wilcox, M. H. (1998). Treatment of Clostridium difficile infection. J Antimicrob Chemother 41, 41-46.[Abstract/Free Full Text]

Wren, B. W., Mullany, P., Clayton, C. & Tabaqchali, S. (1988). Molecular cloning and genetic analysis of a chloramphenicol acetyltransferase determinant from Clostridium difficile. Antimicrob Agents Chemother 32, 1213-1217.[Medline]

Wüst, J. & Hardegger, U. (1983). Transferable resistance to clindamycin, erythromycin, and tetracycline in Clostridium difficile. Antimicrob Agents Chemother 23, 784-786.[Medline]

Wüst, J. & Hardegger, U. (1988). Studies on the resistance of Clostridium difficile to antimicrobial agents. Zentbl Bakteriol Mikrobiol Hyg (A) 267, 383-394.

Received 21 May 2001; revised 2 July 2001; accepted 3 July 2001.