HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via CrossRef Citing Articles via Google Scholar Google Scholar Articles by Spigaglia, P. Articles by Mastrantonio, P. Search for Related Content PubMed PubMed Citation Articles by Spigaglia, P. Articles by Mastrantonio, P. Agricola Articles by Spigaglia, P. Articles by Mastrantonio, P.

Comparative analysis of Clostridium difficile clinical isolates belonging to different genetic lineages and time periods

Department of Infectious, Parasitic and Immunomediated Diseases, Istituto Superiore di Sanità, Viale Regina Elena 299, 00161 Rome, Italy

Correspondence Paola Mastrantonio pmastran{at}iss.it

Received March 23, 2004
Accepted June 25, 2004

Recent studies have shown that Clostridium difficile strains with variant toxins and those with resistance to macrolide–lincosamide–streptogramin B (MLSB) are increasingly causing severe disease and outbreaks in hospital settings. Here, the pathogenicity locus (PaLoc), the acquisition of binary toxin, and the genotypic and phenotypic characteristics of antibiotic resistance of 74 C. difficile clinical strains isolated from symptomatic patients in Italy during different time periods were studied. These strains were found to belong to two different lineages, and those isolated before 1991 were genetically unrelated to the more recent strains. The majority of recent C. difficile strains showed variations in toxin genes and in the toxin negative regulator (tcdC) and had the binary toxin. In 62 % of them, variations in tcdC and the presence of the binary toxin were associated. Five classes of susceptibility/resistance pattern (EC-a to -e) for erythromycin and clindamycin were identified in all strains studied. Most of the recent isolates belonged to EC-d and EC-e and, although erythromycin-resistant in vitro, did not harbour the commonly associated ermB determinant. Interestingly, two strains of the EC-d class were resistant to clindamycin only after induction with subinhibitory concentrations of the antibiotic. A decrease in tetracycline and chloramphenicol MIC values was also observed in the recently isolated strains, associated with less frequent detection of the catD and tetM genes. Two tetM-positive strains were resistant in vitro only after induction with subinhibitory concentrations of the antibiotic. The acquisition of the binary toxin, the possible increase in toxin production due to a mutated negative regulator and a decrease in the fitness cost as a result of lower levels of antibiotic resistance or other mechanisms may have led to the successful establishment of these new phenotypes, with potentially serious clinical implications.

	INTRODUCTION

TOP INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
Recently, the involvement of Clostridium difficile strains with variant toxins, as well as strains with resistance to macrolide–lincosamide–streptogramin B (MLSB), in causing severe diseases and outbreaks has been demonstrated (Alfa et al., 2000; Johnson et al., 1999; Pituch et al., 2003; Sambol et al., 2000; Spigaglia & Mastrantonio, 2002).

C. difficile toxin A (TcdA, an enterotoxin) and toxin B (TcdB, a cytotoxin) are encoded by two genes, tcdA and tcdB, which, together with three accessory genes, tcdE, tcdD and tcdC, form a pathogenicity locus (PaLoc) (Braun et al., 1996; Hammond & Johnson, 1995). TcdD and TcdC have been indicated as the positive and negative regulators of toxin A and B expression, respectively, whereas TcdE seems to have a holin function (Hammond et al., 1997; Hundsberger et al., 1997; Tan et al., 2001). Much heterogeneity has been observed in the toxin A and B genes and currently 20 groups of variant C. difficile strains have been recognized and defined as toxinotypes I–XX (Rupnik et al., 1997, 1998, 2003). Some studies have also focused on the production of a third toxin, binary toxin CDT, by several C. difficile isolates (Perelle et al., 1997; Stubbs et al., 2000). Binary toxin CDT is produced by the majority of strains with mutations in the tcdA and tcdB genes and its detection has been suggested as a method for the rapid identification of these variant strains (Rupnik et al., 2001). Mutations have been also detected in PaLoc accessory genes. Three different alleles of tcdC, which encodes the negative toxin regulator, have been identified and it has been hypothesized that these variants could have some influence on the amounts of toxins A and B released (Soehn et al., 1998; Spigaglia & Mastrantonio, 2002).

MLSB-resistant C. difficile strains have been demonstrated to be the cause of epidemics of diarrhoea in different hospitals (Johnson et al., 1999; Pituch et al., 2003). The erm genes, in particular the ermB class, are the main known cause of resistance to MLSB in C. difficile and Clostridium perfringens (Berryman & Rood, 1989; Farrow et al., 2000; Roberts, 1995). In C. difficile, this determinant is located on a conjugative transposon called Tn5398. Recent studies have demonstrated much heterogeneity in the genetic arrangement of this element (Farrow et al., 2001; Spigaglia & Mastrantonio, 2003). Furthermore, an association between different levels of resistance to erythromycin, different alleles and the number of ermB copies has been observed. Resistance to MLSB is frequently found in C. difficile strains that are also resistant to other antibiotics (Ackermann et al., 2003; Barbut et al., 1999; Delmee & Avesani, 1988; Roberts et al., 1994), underlining the importance of monitoring their circulation.

In a previous study, we analysed C. difficile clinical isolates from different Italian hospitals by PCR ribotyping and PFGE (Spigaglia et al., 2001). In the present study, the PaLoc genes, the presence of the binary toxin, and the genotypic and phenotypic characteristics of antibiotic resistance in several toxinogenic C. difficile clinical isolates were analysed with the aim of identifying characteristics differentiating older C. difficile strains from more recently isolated strains.

	METHODS

TOP INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
Bacterial strains and PCR ribotyping.

Seventy-four toxinogenic C. difficile strains from sporadic symptomatic cases or representative of outbreaks were examined in this study. These isolates, from diarrhoeic adults and children, were collected in five hospitals during the course of the study period and sent to our laboratory for further characterization. Twenty isolates were selected from 53 strains isolated in period I (before 1990), 34 from 90 strains in period II (1991–1999) and 20 from 56 strains isolated in period III (2000–2001). Thirty-five C. difficile isolates were selected from 100 strains collected from 1985 to 1999 and typed in a previous study (Spigaglia et al., 2001), while 39 isolates were selected from 110 strains isolated between 1991 and 2001 and typed during this study. PCR ribotyping was performed on these strains as previously described (Spigaglia et al., 2001). Briefly, 10 µl crude DNA was used as a template. Amplification was performed in a final volume of 100 µl with a reaction mixture containing buffer (10 mM Tris/HCl, 50 mM KCl and 1.5 mM MgCl₂), 200 mM each deoxynucleoside triphosphate, 100 pmol each primer and 2.5 U Takara Ex Taq (Takara Shuzo). The primers were complementary to conserved regions at the 3' and 5' ends of the 16S and 23S rRNA genes, respectively, as previously described by Kostman et al. (1992). DNA was amplified for 30 cycles, each consisting of 1 min at 94 °C, 1 min at 55 °C and 1 min at 72 °C. After amplification, 10 µl of the product was electrophoresed on a 1.5 % agarose gel. The genomic DNA fingerprinting patterns produced by PCR ribotyping were analysed using Molecular Analyst Fingerprinting Plus Software, version 1.0 (Bio-Rad). A similarity analysis was performed using Dice's coefficient and clustering was performed using the unweighted pair group mean association.

C. difficile 630 (Wüst & Hardegger, 1983), C250 (Wren et al., 1988) and F17 (Spigaglia & Mastrantonio, 2003) were used as control strains in multiplex PCRs to detect antibiotic-resistance determinants and in PCR-RFLP of ermB genes.

Multiplex PCR for tcdA and tcdB detection and PCR for the repetitive region of tcdA.

Multiplex PCR for toxin A and B genes, amplifying fragments of 642 and 412 bp, respectively, were performed as previously described (Spigaglia & Mastrantonio, 2002). The amplification of a 3.1 kb fragment of the repetitive region of tcdA was performed by PCR with primers A3C and A4N, as reported by Rupnik et al. (1997), using 2 µl purified DNA as the template in each PCR. DNA was purified using the NUCLEOBOND Buffer Set III and NUCLEOBOND cartridges AXG 20.

Multiplex PCR for the PaLoc accessory genes (tcdC, tcdD and tcdE) and detection of binary toxin gene cdtB.

Multiplex PCR for tcdC, tcdD and tcdE gene amplification were performed using the primer pairs PaL15/PaL16, PaL11/PaL12 and PaL13/PaL14, respectively (Table 1). Primers were designed based on the C. difficile PaLoc sequence (GenBank accession no. X92982). Five microlitres of crude DNA was used as the template, and was denatured for 5 min at 94 °C and amplified for 30 cycles consisting of 1 min at 94 °C, 1 min at 50 °C and 1 min at 72 °C. Ten microlitres of the PCR product was electrophoresed on a 1.5 % agarose gel. The expected sizes for PCR fragments were approximately 670 bp for tcdC, 470 bp for tcdD and 350 bp for tcdE. As suggested by M. Rupnik and others (www.uni-lj.si/bfbcdiff), binary toxin was detected by amplification of the cdtB gene encoding the binary-toxin-binding component. A PCR for an internal region of cdtB was performed as previously described (Stubbs et al., 2000).

PCR to detect antibiotic-resistance determinants.

Multiplex PCR for tetM, ermB and catD gene amplification were performed using the primer pairs TETMd/TETMr (Marchese et al., 1998), E5/E6 and CL1/CL2 (Table 1), respectively. E5/E6 and CL1/CL2 primer pairs were designed from ermB (GenBank accession nos AF109075 and U18931) and catD (GenBank accession no. AF226276) sequences, respectively. Five microlitres of crude DNA were denatured for 5 min at 94 °C and amplified for 30 cycles consisting of 1 min at 94 °C, 1 min at 50 °C and 1.5 min at 72 °C. Ten microlitres of the PCR products were electrophoresed on a 1.5 % agarose gel. The expected sizes of the PCR fragments were approximately 1.0 kb for tetM, 0.6 kb for ermB and 0.5 kb for catD.

Specific primers for rRNA methylases, classes A, C, F and Q, and the mefA gene, encoding a membrane-bound efflux protein, have been reported by M. C. Roberts (http://faculty.washington.edu/marilynr/). PCRs were performed as described previously (Chung et al., 1999; Luna et al., 2000).

PCR-RFLP of ermB genes.

Twenty microlitres of PCR products, obtained using the primers E5 and E6 and the same PCR conditions as for the multiplex PCR to detect antibiotic-resistance determinants, were digested with 40 U PvuII restriction enzyme. This enzyme has a cutting site in the C. difficile 630 ermB sequence (SZ-type sequence) but no site in the C. perfringens CP592 ermB sequence (SP-type sequence). The digestion of ermB PCR products with a sequence similar to C. difficile 630 ermB would thus produce two fragments of approximately 589 and 122 bp, whereas ermB PCR products with a sequence similar to C. perfringens CP592 should remain undigested. The fragments were separated on a 2 % agarose gel and visualized by ethidium bromide staining.

Susceptibility tests and induction of resistance.

MIC values for erythromycin, clindamycin, tetracycline and chloramphenicol were determined by the E-test (AB Biodisk), following the manufacturer's instructions. Wilkins–Chalgren agar plates (Unipath) were incubated anaerobically at 37 °C for 24 h for erythromycin, tetracycline and chloramphenicol and for 48 h for clindamycin. The cut-off values were 4 mg l^–1 for erythromycin and clindamycin, 8 mg l^–1 for tetracycline and 16 mg l^–1 for chloramphenicol (National Committee for Clinical Laboratory Standards, 1993). Induction of resistance to clindamycin, tetracycline and chloramphenicol in C. difficile isolates was evaluated by pre-growth for 18 h on blood agar plates containing subinhibitory concentrations of each antibiotic. Erythromycin, at a subinhibitory concentration of 0.05 mg l^–1, was used to induce MLSB resistance as described previously (Giovanetti et al., 1999). The subinhibitory concentration of 0.01 mg l^–1 was used for both tetracycline and chloramphenicol (Doherty et al., 2000; Wren et al., 1988). MICs were then determined by the E-test as described above.

	RESULTS

TOP INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
Detection of toxins A and B and PCR ribotyping results

All C. difficile isolates were confirmed as toxinogenic by multiplex PCR for toxin A and B gene detection (data not shown).

Comparison of all 74 strains included in this work showed that they clustered into two main lineages, previously described for some by Spigaglia et al. (2001). We have named these lineages 1 and 2. Five of the 13 PCR ribotypes (A, D, E, F and C) belonged to lineage 1 and eight (B, H, I, L, O, P, R and U) to lineage 2.

The main PCR ribotypes were A (21 strains), D (13 strains) and R (15 strains). The other PCR ribotypes each contained between one and seven C. difficile strains.

The majority (95 %) of C. difficile strains isolated in period I (before 1990) were PCR ribotypes of lineage 1, whereas the majority (90 %) of isolates from period III (2000–2001) were PCR ribotypes of lineage 2. Strains isolated in period II (1991–1999) were equally distributed in both clusters.

PCR ribotype A represented 52 % of the strains belonging to lineage 1 (21/40), whereas PCR ribotype R contained 44 % of those belonging to lineage 2 (15/34).

Analysis of PaLoc genes and detection of the binary toxin gene cdtB

A PCR specific for the toxin A repetitive region identified six isolates with noticeable deletions (Table 2). Five of them, four of lineage 2 (PCR ribotype R) and one of lineage 1 (PCR ribotype E), had a deletion of 0.8 kb and were very similar to toxinotype VII described by Rupnik et al. (1998, 2003), whereas one isolate of lineage 1, belonging to PCR ribotype E, had a deletion of 1.9 kb and could be recognized as a toxin A^– toxin B⁺ strain (toxinotype VIII).

None of the strains belonging to lineage 1 showed any variations in the tcdC, tcdD and tcdE accessory genes, whereas 15 C. difficile isolates of lineage 2 showed a different size for the PCR fragment obtained by amplification of the regulatory gene tcdC. Thirteen strains of PCR ribotype R had a tcdC gene type A, with a deletion of 39 bp, whereas the two strains of PCR ribotype L had a tcdC gene type B or C, with a deletion of 18 bp (Fig. 1 and Table 2).

View larger version (26K): [in this window] [in a new window]   Fig. 1. Multiplex PCR for tcdC, tcdD and tcdE detection in C. difficile isolates. The expected size for the tcdC PCR fragment was approximately 670 bp. tcdC type A showed a deletion of 39 bp, whereas tcdC type B and type C both showed a deletion of 18 bp. Lanes: 1, DNA Molecular Weight Marker VIII (Roche); 2, 3, 8 and 11, isolates with a classic tcdC; 4–6, isolates with a tcdC gene type A; 7, DNA Molecular Weight Marker IX (Roche); 9 and 10, isolates with a tcdC gene type B or C.

No isolate of lineage 1 was positive for the binary toxin gene cdtB, whereas in lineage 2 all strains of PCR ribotype U and R were positive for this gene, suggesting the presence of minor variations in toxin A and B genes.

Detection of antibiotic-resistance genes and antimicrobial susceptibility

The results obtained by multiplex PCR for ermB, tetM and catD gene detection and MIC values for tetracycline and chloramphenicol are shown in Fig. 2 and Table 3, respectively, while MIC values for erythromycin and clindamycin are shown in Table 4.

View larger version (24K): [in this window] [in a new window]   Fig. 2. Multiplex PCR for tetM, ermB and catD detection in C. difficile isolates. Lanes: 1 and 4, isolates positive for tetM, ermB and catD; 2 and 6, isolates positive for tetM and ermB; 3 and 7, isolates positive for tetM; 5, DNA Molecular Weight Marker IX (Roche); 9, isolate positive for ermB; 8, susceptible isolate.

Three PCR ribotypes of lineage 1 (A, D and E) and three of lineage 2 (B, L and R) contained all the antibiotic-resistant C. difficile strains.

Twenty-five of the 40 lineage 1 strains were found to have at least one resistance determinant. In particular, 19 isolates, belonging to PCR ribotype A, were positive for ermB, tetM and catD genes, four strains (two of PCR ribotype A, one of PCR ribotype D and one of PCR ribotype E) had both ermB and tetM genes and two isolates (one of PCR ribotype D and one of PCR ribotype E) had only an ermB gene. Of the 34 lineage 2 strains, one PCR ribotype R was positive for ermB, tetM and catD genes, one of the same PCR ribotype had both ermB and tetM genes, nine isolates (eight PCR ribotype R and one L) harboured only a tetM gene and two (one PCR ribotype B and one R) only an ermB gene.

catD genes have always been found with ermB and tetM genes.

C. difficile isolates belonging to lineage 1 showed MICs for erythromycin between 0.094 and 256 mg l^–1 and for clindamycin between 0.047 and 256 mg l^–1.

Twenty-six isolates of lineage 1 were resistant to both antibiotics. Erythromycin and clindamycin resistance patterns could be distinguished in four classes designated EC-a, EC-b, EC-c and EC-d (Tables 3 and 4). EC-a was characterized by susceptibility to both erythromycin (MIC 0.032–2 mg l^–1) and clindamycin (MIC 0.047–3 mg l^–1), class EC-b by high levels of resistance to both erythromycin and clindamycin (MIC 256 mg l^–1), class EC-c by low levels of resistance to erythromycin (MIC 12–48 mg l^–1) and high resistance to clindamycin (MIC 256 mg l^–1) and class EC-d by a high level of resistance to erythromycin (MIC 256 mg l^–1) and susceptibility to clindamycin (MICs 1.5–3 mg l^–1). Sixty-five per cent (17/26) of the C. difficile isolates of lineage 1, resistant to both antibiotics, showed a class EC-c phenotype, 31 % (8/26) a class EC-b phenotype and 4 % (1/26) a class EC-d phenotype. This last strain was ermB-negative. C. difficile isolates belonging to lineage 2 showed an MIC for erythromycin between 0.032 and 256 mg l^–1 and for clindamycin between 0.047 and 256 mg l^–1. Nine strains were resistant to erythromycin and seven to clindamycin. Four strains had an ermB gene and showed a class EC-b phenotype and five were ermB-negative. Three of the latter showed a phenotype with a high level of resistance to erythromycin (MIC 256 mg l^–1) and low to clindamycin (MIC 4–12 mg ml^–1), designated class EC-e, and two strains showed a class EC-d phenotype.

All C. difficile isolates were erythromycin-resistant but ermB-negative isolates were also PCR-negative for the ermA, ermC, ermF, ermQ and mefA genes.

C. difficile isolates of lineage 1 with a tetM gene were resistant to tetracycline with MICs ranging from 12 to 256 mg l^–1 (MIC₅₀ = 32 mg l^–1, MIC₉₀ = 64 mg l^–1). Five of the 11 C. difficile isolates of lineage 2 that were tetM-positive were susceptible to this antibiotic, with MICs between 0.023 and 4 mg l^–1 (MIC₅₀ = 3 mg l^–1, MIC₉₀ = 12 mg l^–1); one of these strains was also ermB-positive but erythromycin-susceptible. All isolates of lineage 1 that were catD-positive were resistant to chloramphenicol, whereas the only catD-positive isolate of lineage 2 was susceptible to chloramphenicol in vitro.

Induction of antibiotic resistance

All C. difficile strains in phenotypic classes EC-d and EC-e were analysed for the induction of resistance to clindamycin. Two isolates of the EC-d class, belonging to PCR ribotype R, showed induced resistance in vitro with MICs of 4 and 12 mg l^–1, respectively.

All tetM-positive strains that were susceptible in vitro were analysed for induction of tetracycline resistance. Two strains, belonging to PCR ribotype R, showed induced resistance with MICs of 8 and 12 mg l^–1, respectively.

No C. difficile isolate was found that showed induced resistance towards both clindamycin and tetracycline.

The only C. difficile isolate with a catD gene that was susceptible to chloramphenicol in vitro (PCR ribotype R) did not show induced resistance.

ermB characterization

All ermB genes were analysed to define the nucleotide sequence type (Table 3 and Fig. 3). Twenty-one of the 25 isolates of lineage 1 (84 %) with an ermB gene showed an SZ-type sequence, similar to C. difficile 630, and the remaining four strains had an SP-type sequence, similar to C. perfringens CP592. All ermB genes detected in isolates of lineage 2 had an SP-type sequence. C. difficile isolates showing a class EC-b phenotype had an ermB SZ-type sequence, whereas the majority of isolates (67 %) showing a class EC-a phenotype had an ermB SP-type sequence.

View larger version (22K): [in this window] [in a new window]   Fig. 3. PCR-RFLP of the ermB gene using PvuII. Genes with the SP-type sequence remained undigested, whereas genes with the SZ-type sequence were digested into two fragments of approximately 450 and 200 bp. Lanes: 1, 2, 5, 6, 10, 11 and 14, ermB with SP-type sequence; 3, 4, 8, 9, 12 and 13, ermB with SZ-type sequence; 7, DNA Molecular Weight Marker VIII (Roche).

	DISCUSSION

TOP INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
In this study we performed a comparative analysis of C. difficile clinical isolates belonging to two distinct genetic lineages, focusing our investigations on analysis of PaLoc genes, detection of binary toxin and antibiotic-resistance patterns.

The results showed that the majority of strains of recent isolation grouped in lineage 2 and had unusual characteristics that differed from those of older isolates belonging to lineage 1.

Various studies of C. difficile isolates with variant toxins have been published (Barbut et al., 2002; Rupnik et al., 2001, 2003), but there is no report of a comparison between C. difficile strains of older and more recent isolation. The results of this study seem to indicate an increase in C. difficile strains with variant PaLoc genes over time. The majority of these strains were identified as PCR ribotype R and represented 56 % of all isolates of lineage 2, whereas in lineage 1, they represented only 5 % and belonged to PCR ribotype E. Most of the recent strains with variant PaLoc genes (81 %) showed variations in toxin genes, as suggested by the presence of the binary toxin. Only six strains, four of lineage 2 and two of lineage 1, had major deletions in tcdA. Together with the acquisition of the binary toxin, mutations in the negative toxin regulator of toxins A and B (tcdC) differentiated the recent strains from the older ones. Seventy-one per cent of isolates with a variant PaLoc showed mutations in the tcdC gene; in particular, the allele tcdC-A was prevalent. tcdC-A encodes a truncated protein of only 61 amino acids that probably has an altered function, determining an increase in toxin production. It was interesting to note that 62 % of the isolates with a variant PaLoc showed both a mutated negative regulator of toxins A and B and the binary toxin.

Different characteristics were also found in the C. difficile isolates of the two lineages when resistance to MLSB, tetracycline and chloramphenicol was investigated. All multi-resistant strains belonged to lineage 1, in particular to PCR ribotype A, and represented 47 % of C. difficile isolates of that lineage. No multi-resistant strain was found in lineage 2. The only strain of PCR ribotype R that was positive for all the antibiotic-resistance determinants examined was resistant to erythromycin only in vitro. Barbut et al. (1999) described a similar trend in recently isolated strains, which were more susceptible to this panel of antibiotics than those isolated in the past, due to a major change in the serogroup distribution.

Five different classes were identified on the basis of erythromycin and clindamycin resistance patterns. Interestingly, all C. difficile isolates showing an EC-d and EC-e phenotype were ermB-negative, whereas the EC-b and EC-c phenotypes were always associated with the presence of an ermB gene. In particular, strains of the EC-c class were always associated with an allele showing an SZ-type sequence, whereas 67 % of those of the EC-b class were associated with an allele showing an SP-type sequence. As we observed in a previous study (Spigaglia & Matrantonio, 2003), these phenotypes were also associated with a different number of ermB copies: EC-c strains were associated with one copy of a gene with an SZ-type sequence, whereas EC-b strains had one or two copies of a gene with an SP-type sequence or two copies of a gene with an SZ-type sequence. In lineage 1, the predominant phenotypic class was EC-b, which contained 73 % of the strains resistant to erythromycin. In contrast, this phenotype was not found in lineage 2, where the predominant classes were EC-d and EC-e containing 56 % of the strains resistant to erythromycin. Recent studies have described ermB-negative C. difficile strains with high levels of resistance to clindamycin and/or erythromycin (Ackermann et al., 2003; Pituch et al., 2003). In our study, only C. difficile isolates with susceptibility or low levels of resistance to clindamycin were ermB-negative. These strains were also PCR-negative for the ermA, ermC, ermF, ermQ and mefA genes. Induction of resistance to clindamycin indicated that two strains of class EC-d phenotype were inducibly resistant. All these data suggest that erm genes different from those tested, or different mechanisms, may be involved in MLSB resistance in strains with an EC-d or EC-e phenotype.

As far as tetracycline resistance was concerned, we observed a decrease in MIC₅₀ and MIC₉₀ values in strains belonging to lineage 2. Furthermore, an inducible resistance to tetracycline was observed in two of the four isolates susceptible in vitro to this antibiotic, but tetM-positive.

The detection of inducible resistance to antibiotics in recent C. difficile isolates could be clinically relevant and further analysis should be performed to characterize these strains further and monitor their circulation.

A decrease in the presence of the catD gene and in chloramphenicol MIC values was observed on comparison of isolates from lineage 1 and 2.

In conclusion, strains belonging to lineage 2, in contrast to strains belonging to lineage 1, showed the presence of binary toxin, variations in toxins A and B and in their negative regulator tcdC, and the absence of multi-drug resistance.

Taken together, these results provide evidence of a recent circulation of C. difficile strains with characteristics that could increase their pathogenic potential. It may be hypothesized that additional virulence factors, such as the binary toxin and a possible increase in toxins A and B production due to a mutated negative regulator, could enhance C. difficile virulence (Soehn et al., 1998; Spigaglia & Mastrantonio, 2002). Moreover, a decrease in the biological fitness cost necessary to maintain constitutive expression and high levels of antibiotic resistance (Andersson & Levin, 1999; Bjorkman & Andersson, 2000; Johanesen et al., 2001) could be relevant for bacterial survival and for its competitive performance. Further studies will be necessary to confirm these hypotheses, but the possible clinical implications following the spread of C. difficile strains with these characteristics demand the monitoring of the molecular features in these clinical isolates.

	ACKNOWLEDGEMENTS

TOP INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
This work was partially supported by the European Community's Fifth Framework Programme ‘Quality of Life and Management of Living Resources', contract no. QLK2-CT-2002-00843-ARTRADI. We are grateful to Tonino Sofia for editing the manuscript and to Valentina Carucci for technical support.

	REFERENCES

TOP INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGEMENTS REFERENCES