The susceptibility trends for Bacteroides fragilis and related species against various antibiotics were determined using data from 3 years of surveillance (2010-2012) on 779 isolates referred by 7 medical centers. The antibiotic test panel included imipenem, ertapenem, meropenem, ampicillin-sulbactam, piperacillin-tazobactam, cefoxitin, clindamycin, moxifloxacin, tigecycline, linezolid, chloramphenicol and . MICs were determined using the agar dilution CLSI reference method. Carbapenem resistance remained low (range 1.1%-2.5%) and unchanged from 2008 to 9 through 2010-2012. Resistance also remained low to the beta-lactam/beta-lactamase inhibitor combinations (1.1%-4.4%). While resistance to clindamycin and moxifloxacin remained high; rates were lower for B. fragilis in 2010-12 (24% and 19% respectively) compared to the earlier time frame of 2008-9 (29% and 35% respectively for the earlier time frame). There were notable species and resistance associations which have been demonstrated previously. No resistance to metronidazole or chloramphenicol resistance was seen. These data demonstrate the continued variability in resistance among Bacteroides and Parabacteroides species, but do demonstrate that carbapenems and beta-lactam/beta-lactamase inhibitor combinations remain very active throughout the United States.

Salivary contamination of surgical wounds in clean-contaminated head and neck surgery with free flap reconstruction remains a major cause of infection and leads to significant morbidity. This study investigates the correlation between intraoral flora and surgical site infections (SSIs) among high-risk head and neck cancer patients undergoing resection and free flap reconstruction.

One hundred twenty-nine patients were identified as being at high risk for infective complications based on cancer stage, tumor size, comorbid factors, and extent of reconstruction. All patients had intraoral swab cultures before surgery. Patients with culture-confirmed SSI after surgery were chosen for analysis, using the κ index and its 95% confidence interval for concordance analysis. All patients received clindamycin and gentamicin for antibiotic prophylaxis for 5 days. Antibiotic susceptibility testing of all isolates was obtained and analyzed.

Thirty-seven patients experienced SSI, or an infection rate of 28.3%, occurring at a mean of 9.3 postoperative days. The overall concordance between oral flora and SSI was fair to moderate (κ index of 0.25), but detailed analysis shows a higher concordance for known and opportunistic pathogens, such as Pseudomonas aeruginosa and Enterococcus faecalis, compared to typical oral commensals. Antibiotic susceptibility tests show rapid and significant increases in resistance to clindamycin, indicating a need for a more effective alternative.

Predicting pathogens in SSI using preoperative oral swabs did not demonstrate a good concordance in general for patients undergoing clean-contaminated head and neck surgery, although concordance for certain pathogenic species seem to be higher than for typical intraoral commensals. The rapid development of resistance to clindamycin precludes its use as a prophylactic agent.

The susceptibility trends for the species of the Bacteroides fragilis group against various antibiotics were determined using data from 4 years [2006-2009] on 1957 isolates referred by 8 medical centers participating in a National Survey for the Susceptibility of B. fragilis. The antibiotic test panel included doripenem, ertapenem, imipenem, meropenem, ampicillin:sulbactam, piperacillin:tazobactam, cefoxitin, clindamycin, moxifloxacin, tigecycline, chloramphenicol and metronidazole. MICs were determined using agar dilution methods following CLSI recommendations. Genetic analysis of isolates from 2008 with elevated MICs (>2 μg/mL) to one or more of the carbapenems to detect presence of the cfiA gene was performed using PCR methodology. The results showed an increase in the resistance rates to the β-lactam antibiotics. High resistance rates were seen for clindamycin and moxifloxacin (as high as 60% for clindamycin and >80% for moxifloxacin), with relatively stable low resistance (5.4%) for tigecycline. For carbapenems, resistance in B. fragilis was 1.1%-2.5% in 2008-9. One isolate resistant to metronidazole (MIC 32 μg/mL) was observed as well as isolates with elevated MICs to chloramphenicol (16 μg/mL). Genetic analysis indicated that the cfiA gene was present in some but not all of the isolates with high MICs to the carbapenems. These data indicate that there continue to be changes in susceptibility over time, and that resistance can be seen among the carbapenems. High antibiotic resistance rates tend to be associated with specific species.

The bacteria of the Bacteroides fragilis group are considered important clinical pathogens and they are the most common anaerobes isolated from human endogenous infections. In this study, the susceptibility patterns to antibiotics and metals of 114 species of the B. fragilis group isolated from children with and without diarrhea were determined. Susceptibility was assayed by using an agar dilution method with Wilkins-Chalgren agar. All B. fragilis strains were resistant to lead and nickel, but susceptible to metronidazole and imipenem. beta-lactamase production was detected by using biological and nitrocefin methods, respectively, in 50% and 90.6% of the isolates of children with diarrhea and in 60% and 90% of the isolates of children without diarrhea. Our results show an increase of antibiotics and metals resistance in this microbial group, and a periodic evaluation of the antimicrobial susceptibility is needed. In Brazil, the contamination for antibiotics or metal ions is often observed, and it is suggested an increase the antimicrobial resistance surveillance of this microbial group, mainly those isolated from children's diarrhea.

Bacteroides fragilis has been isolated from several human and non-human monomicrobial and mixed infections. In this study, some virulence markers and the antimicrobial susceptibility of bacteria of the B. fragilis group isolated from children's stools were evaluated. All the 64 isolates showed the following characteristics: capsulated, beta-hemolytic, hydrophilic, and serum-resistant. Only, 24 (37.5%) strains were resistant at 60 masculine C, for 30 min, and among them, 12 (18.75%) were resistant at 60 masculine C, for 60 min. Also, none strain was resistant at 100 masculine C. Four strains were able to hemagglutinate erythrocytes and D-mannose, D-galactose, D-arabinose, and D-xylose inhibited hemagglutination in 2 B. fragilis strains (p76a, p76b). The hemagglutination in the strain B. uniformis p3-2 was inhibited by D-xylose and D-galactose. The bft gene detection and the enterotoxin production were observed only in 13 EF-enterotoxigenic species. Fragilysin activity was confirmed on HT-29 cells. The antimicrobial determination confirmed that both imipenem and metronidazole were efficient against B. fragilis species; all the strains were resistant to lead and nickel. Plasmids of 2.9, 4.4, 4.8, and 8.9 kb were observed in 6 tested strains. These results show the values of the species identification from clinical infections, as well as of the periodic evaluation of the resistance patterns of the B. fragilis group at Brazilian medical institutions.

Antibiotic-resistant anaerobic bacteria have become increasingly recognized as a confounding factor in the selection of therapeutic agents. The use of potent, broad-spectrum antibiotics as empirical therapy, along with appropriate adjunctive measures, has, in some ways, masked the magnitude of the antibiotic resistance problem that parallels that observed for nonanaerobic pathogens. The use of standardized testing methods that recognize resistance and an understanding of resistance mechanisms have become essential for the treatment of patients and the development of new agents.

Ertapenem, a new long acting beta-lactam with broad-spectrum antimicrobial coverage, was tested in vitro to compare its activity against 556 clinical isolates of anaerobes to other established agents using a broth microdilution method as recommended by the NCCLS. Against all anaerobes ertapenem inhibited 99.1% of isolates at 4 microg/mL, had a mode MIC of 0.12 microg/mL, and showed activity comparable to imipenem, meropenem, trovafloxacin, piperacillin-tazobactam, and metronidazole. Five of the B. fragilis group (4 B. fragilis, 1 B. vulgatus) isolates tested showed reduced susceptibility (>or=8 microg/mL; <1%) to ertapenem while all isolates of Prevotella, Porphyromonas, Fusobacterium, and Peptostreptococcus were susceptible. Only piperacillin-tazobactam had susceptible MIC's for all test isolates followed by metronidazole and the carbapenems exhibiting low resistance rates (<1%).

In vitro surveys of antimicrobial resistance among clinically important anaerobes are an important source of information that can be used for clinical decisions in the choice of empiric antimicrobial therapy. This study surveyed the susceptibilities of 556 clinical anaerobic isolates from four large medical centers using a broth microdilution method. Piperacillin-tazobactam was the only antimicrobial agent to which all the isolates were susceptible. Similarly, imipenem, meropenem, and metronidazole were highly active (resistance, <0.5%), whereas the lowest susceptibility rates were noted for penicillin G, ciprofloxacin, and clindamycin. For most antibiotics, blood isolates were less susceptible than isolates from intra-abdominal, obstetric-gynecologic, and other sources. All isolates of the Bacteroides fragilis group were susceptible to piperacillin-tazobactam and metronidazole, while resistance to imipenem and meropenem was low (<2%). For these same isolates, resistance rates (intermediate and resistant MICs) to ampicillin-sulbactam, cefoxitin, trovafloxacin, and clindamycin were 11, 8, 7, and 29%, respectively. Among the individual species of the B. fragilis group, the highest resistance rates were noted among the following organism-drug combinations: for clindamycin, Bacteroides distasonis and Bacteroides ovatus; for cefoxitin, Bacteroides thetaiotaomicron, B. distasonis, and Bacteroides uniformis; for ampicillin-sulbactam, B. distasonis, B. ovatus, and B. uniformis; and for trovafloxacin, Bacteroides vulgatus. For the carbapenens, imipenem resistance was noted among B. fragilis and meropenem resistance was seen among B. fragilis, B. vulgatus, and B. uniformis. With few exceptions all antimicrobial agents were highly active against isolates of Prevotella, Fusobacterium, Porphyromonas, and Peptostreptococcus. These data further establish and confirm that clinically important anaerobes can vary widely in their antimicrobial susceptibilities. Fortunately most antimicrobial agents were active against the test isolates. However, concern is warranted for what appears to be a significant increases in resistance to ampicillin-sulbactam and clindamycin.

Antimicrobial resistance, including plasmid-mediated resistance, among the species of the Bacteroides fragilis group is well documented. An analysis of the in vitro susceptibility of B. fragilis group species referred between 1995 and 1996 as well as during a 7-year (1990 to 1996), prospective, multicenter survey of over 4,000 clinical isolates of B. fragilis group species was undertaken to review trends in the percent resistance to and geometric mean MICs of the antibiotics tested. There was a trend toward a decrease in the geometric mean MICs of most beta-lactam antibiotics, while the percent resistance to most agents was less affected. Within the species B. fragilis, the geometric mean MICs showed significant (P < 0.05) decreases for piperacillin-tazobactam, ticarcillin-clavulanate, piperacillin, ticarcillin, ceftizoxime, cefotetan, and cefmetazole; a significant increase was observed for clindamycin and cefoxitin. For the non-B. fragilis species, a significant decrease in the geometric mean MICs was observed for meropenem, ampicillin-sulbactam, ticarcillin-clavulanate, piperacillin, ticarcillin, ceftizoxime, and cefmetazole; a significant increase was observed for cefoxitin. Significant increases in percent resistance were observed within the B. fragilis strains for ticarcillin and ceftizoxime and within the non-B. fragilis isolates for cefotetan. Significant increases in percent resistance among all B. fragilis group species were observed for clindamycin, while imipenem showed no significant change in resistance trends. The trend analysis for trovafloxacin was limited to 3 years, since the quinolone was tested only in 1994, 1995, and 1996. During the 7 years analyzed, there was no resistance to metronidazole or chloramphenicol observed. The data demonstrate that resistance among the B. fragilis group species has decreased in the past several years, the major exception being clindamycin. The majority of the resistance decrease has been for the beta-lactams in B. fragilis, compared to other species. The reasons for these changes are not readily apparent.

The Bacteroides fragilis conjugal plasmid pBFTM10 contains two genes, btgA and btgB, and a putative oriT region necessary for transfer in Bacteroides fragilis and Escherichia coli. The BtgA protein was predicted to contain a helix-turn-helix motif, indicating possible DNA binding activity. DNA sequence analysis of the region immediately upstream of btgA revealed three sets of inverted repeats, potentially locating the oriT region. A 304-bp DNA fragment comprising this putative oriT region was cloned and confirmed to be the functional pBFTM10 oriT by bacterial conjugation experiments using E. coli and B. fragilis. btgA was cloned and overexpressed in E. coli, and the purified protein was used in electrophoretic mobility shift assays, demonstrating specific binding of BtgA protein to its cognate oriT. DNase I footprint analysis demonstrated that BtgA binds apparently in a single-stranded fashion to the oriT-containing fragment, overlapping inverted repeats I, II, and III and the putative nick site.

Three small 5-nitroimidazole (5-Ni) resistance plasmids (pIP417, pIP419, and pIP421) from Bacteroides clinical isolates are transferable by a conjugative process during homologous or heterologous matings. The mobilization properties of pIP417 originated from strain BV-17 of Bacteroides vulgatus were studied. The plasmid was successfully introduced by in vitro conjugation into different strains of Bacteroides and Prevotella species and could be transferred back from these various strains to a plasmid-free 5-Ni-sensitive Bacteroides fragilis strain, indicating that in vivo spread of the resistance gene may occur. The transfer of plasmid pIP417 harbored by the Tc(r) strain BF-2 of B. fragilis was stimulated by low concentrations of tetracycline or chlorotetracycline. This suggests a possible role for coresident conjugative transposons in the dissemination of 5-Ni resistance among gram-negative anaerobes. The nucleotide sequence of the 2.1-kb DNA mobilization region was determined. It contains a putative origin of transfer (oriT) in an A+T-rich-region, including three inverted repeats, and two integration host factor binding sites. The two identified mobilization genes (mobA and mobB) are organized in one operon and were both required for efficient transfer. Southern blotting indicated that the mobilization region of plasmid pIP417 is closely related to that of both the erythromycin resistance plasmid pBFTM1O and the 5-Ni resistance plasmid pIP419 but not to that of the 5-Ni resistance plasmid pIP421.

Two standardized susceptibility test methods, a broth microdilution (BMD) and agar dilution (AD) method were performed on a total of 441 clinical isolates of anaerobes with ceftizoxime, cefotaxime, ceftriaxone, cefoxitin, piperacillin, and metronidazole. Against the 339 strains of the Bacteroides fragilis group BMD minimum inhibitory concentration (MIC) values were lower than those from AD testing for all the beta-lactams. Overall for the B. fragilis group and the beta-lactams, the mode MIC values were two- to 64-fold lower, and the MIC50 values two- to eightfold lower. Resistance rates were 11%-28% higher overall with AD results and were higher especially for non-B. fragilis species. For non-Bacteroides anaerobes no major discrepancies were noted for Prevotella species, Peptostreptococcus species, and Viellonella parvula. With Clostridium species and Eubacterium species, some differences were noted with ceftizoxime because of differences in cut-off points. These data illustrate the magnitude of differences in results produced by the two methods using essentially the same test medium for the B. fragilis group. Fortunately, such major discordant results were not widely noted with other groups of anaerobes.

Resistant populations of the Bacteroides fragilis group bacteria (two reference ones and two isolated from human and Callithrix penicillata marmoset) were obtained by the gradient plate technique, to clindamycin, penicillin G, metronidazole and mercuric chloride. All the four tested strains were originally susceptible to the four antimicrobial drugs at the breakpoint used in this study. MICs determination for the four cultures gave constant values for each antimicrobial, on the several steps by the gradient plate technique. The intestinal human B. fragilis strains showed three DNA bands, that could be representative of only two plasmids in the closed covalently circular (CCC) form with molecular weights of approximately 25 and 2.5 Md. The results do not permit an association between the presence of plasmid in the human strain with the susceptibility to the studied drugs. The four strains were beta-lactamase negative in the two methods used, and no particular chromosomal genetic resistance marker was demonstrated. The resistance (MIC) observed, after contact with penicillin G and mercuric chloride, were two-fold in the four tested strains.

The antibiotic resistance of the subgingival microbiota was studied by 3 approaches. First, we assessed the ability of subgingival isolates taken following therapy to grow on media containing tetracycline (TC). Higher percentages of TC-resistant organisms appeared at TC fiber-treated periodontal sites and within the saliva 1 week after treatment as compared with pre-treatment levels. By 1 month, the percentage of TC-resistant organisms had returned to levels comparable to those seen before treatment. In the second approach, subgingival isolates taken following therapy were grown on media without antibiotics, and isolates were selected for Gram-stain and cell morphology determination. This study indicated that subgingival sites became colonized with gram-positive cocci in the same time period that an increase of TC-resistant isolates was observed in the first study. This may account for the transient increase in TC resistance, because many gram-positive cocci are intrinsically resistant to TC. In the third approach, the antibiotic resistance of subgingival gram-negative species was determined. The predominant cultivable microbiota of 9 sites from 3 subjects were isolated immediately before and 6 months after TC fiber treatment. Gram-negative rods were characterized and tested for sensitivity to TC (minimum inhibitory concentration [MIC] 1-128 micrograms/ml), penicillin at 80 micrograms/ml, and erythromycin at 8 micrograms/ml. None of the gram-negative rods were resistant to TC (MIC greater than or equal to 16 micrograms/ml), either before or after treatment. Before treatment 98% of the gram-negative rods were susceptible to TC at 1-2 micrograms/ml and after therapy 88% were susceptible.(ABSTRACT TRUNCATED AT 250 WORDS)

The abilities of erythromycin and clindamycin to act as inducers of clindamycin resistance in the strain Bacteroides vulgatus RYC18F6 is evaluated in vivo (efficiency of plating of inhibitory clindamycin concentrations) and in vitro [efficiency of poly(U)-directed polypeptide synthesis by ribosomes]. Uninduced cells failed to grow during the first 72 h, even at a very low clindamycin concentration (0.1 microgram/ml); after induction with erythromycin or clindamycin, cells were able to form colonies at 32 micrograms/ml after 48 h. The in vitro polymerization test with B. vulgatus RYC18F6 ribosomes (S-30 extract) showed that ribosomes from uninduced cells were fully sensitive to the inhibitory effect of clindamycin. Ribosomes obtained from erythromycin- or clindamycin-induced cells presented a reduced sensitivity to clindamycin inhibition. These results show that resistance to clindamycin in B. vulgatus RYC18F6 is an inducible phenomenon involving a ribosomal modification, probably similar to that previously described for gram-positive bacteria.

Transferable plasmids play an important role in the dissemination of clindamycin-erythromycin resistance in Bacteroides fragilis. We previously described the isolation and properties of pBFTM10, a 14.9-kb ClnR transfer factor from B. fragilis TMP10. We also reported the isolation of a transfer-deficient deletion derivative of pBFTM10 contained in the B. fragilis-Escherichia coli shuttle vector pGAT400. In the present study we used pGAT400 and a similar shuttle vector, pGAT550, to characterize and sequence a region of pBFTM10 required for its transfer from B. fragilis to B. fragilis or E. coli recipients and for its mobilization by the broad-host-range plasmid R751 from E. coli donors to E. coli recipients. Deletion of certain BglII restriction fragments from pBFTM10 resulted in partial or complete loss of transfer ability. Tn1000 insertions into this same region also resulted in altered transfer properties. We used the sites of Tn1000 insertions to determine the DNA sequence of the transfer region. Two potential open reading frames encoding proteins of 23.2 and 33.8 kDa, corresponding to two genes, btgA or btgB, were identified in the sequence. Tn1000 insertions within btgA or btgB or deletion of all or portions of btgA or btgB resulted in either a transfer deficiency or greatly reduced transfer from B. fragilis donors and alterations in mobilization by R751 in E. coli. A potential oriT sequence showing similarity in organization to the oriT regions of the IncP plasmids was also detected. Thus, pBFTM10 encodes and requires at least two proteins necessary for efficient transfer from B. fragilis. These same functions are expressed in E. coli and are required for mobilization by R751.

Using a newly constructed Bacteroides fragilis-Escherichia coli cloning shuttle vector, pJST61, we have cloned the cefoxitin (FOX)-imipenem (IMP) resistance determinant from B. fragilis TAL2480. FOX-IMP resistance in this strain results from the production of a periplasmic, Zn2(+)-containing beta-lactamase which hydrolyzes carbapenems and cephamycins and whose activity is resistant to clavulanic acid but sensitive to Zn2(+)-binding reagents, including EDTA. The pJST61 vector permits efficient library construction in E. coli and allows for the transfer of the library to B. fragilis recipients for the screening or selection of specific phenotypes. The library clone containing the FOX-IMP resistance gene was detected after transfer to B. fragilis TM4000 (Fox-Imps) selecting for Foxr. One of the isolates carrying plasmid pJST241 is resistant to FOX and IMP and synthesizes a periplasmic protein with substrate and inhibitor properties identical to those of strain TAL2480. On the basis of deletion analysis, Tn1000 insertion mutations, and DNA sequencing, we have defined the 747-base cfiA (FOX-IMP resistance) gene within the 3.6-kilobase cloned insert in pJST241. The cfiA gene contains an open reading frame that could code for a precursor protein of 249 amino acids and with a molecular mass of 27,260 daltons. A potential signal sequence has been identified at the N terminus of this protein; cleavage within this sequence would result in a protein of 231 amino acids with a molecular mass of 25,249 daltons. The CfiA protein shows remarkable similarities to the exported, Zn2(+)-requiring, type II beta-lactamase Blm proteins from Bacillus cereus 569/H and 5/B/6. Although overall amino acid identity is only 32%, the Zn ligand-binding His and Cys residues are precisely conserved and the amino acids in the vicinity of these sites show strong similarities (greater than 80%) when the CfiA and Blm proteins are compared.

Conjugal transposons play an important role in the dissemination of antibiotic resistance determinants in the streptococci and have been postulated to exist in Bacteroides fragilis. To investigate the presence of conjugal transposons in B. fragilis, we employed a Tra- derivative of the transfer factor pBFTM10 contained in the chimeric plasmid pGAT400 delta BglII. We attempted to restore transferability to this plasmid from a series of transconjugants generated by crossing B. fragilis TMP230 containing the TET transfer factor with B. fragilis TM4000, a standard recipient. Transconjugant TM4.2321 transferred pGAT400 delta BglII to Escherichia coli HB101 at almost the same frequency as did the Tra+ parental plasmid, pGAT400. Analysis of the transferred plasmids revealed the presence of 9.6 kilobases of additional DNA in every case but at different positions in independent isolates. The presence of this DNA, designated Tn4399, allowed the pGAT400 delta BglII derivatives to retransfer from the TM4000 background to B. fragilis or E. coli recipients. DNA hybridization studies demonstrated the presence of one copy of Tn4399 in TMP230 and three copies at new sites in TM4.2321. Tn4399 is a new B. fragilis transposon with unique transfer properties that may play a role in the dissemination of drug resistance genes. It differs from previously described conjugal transposons by its ability to mobilize nonconjugal plasmids in cis.

The antimicrobial susceptibility of clinical isolates of the Bacteroides fragilis group obtained from nine large Belgian university and community hospitals was studied. Of the beta-lactam antibiotics tested, none were active against 100% of the isolates. Piperacillin was active against 93% of the strains at a breakpoint of 64 micrograms/ml. The percentage of isolates inhibited at 16 micrograms/ml (and 32 micrograms/ml) for the 7-alpha-methoxy antibiotics was: cefoxitin 84; latamoxef 87; and cefotetan 64. B. fragilis tended to be more susceptible to all beta-lactams than the other members of the group. Among the non beta-lactam antibiotics, the resistance rate for minocycline, doxycycline, clindamycin and erythromycin (at a breakpoint of 4 micrograms/ml) were 1%, 4%, 14% and 34%, respectively. Like the beta-lactams, clindamycin was less active against non B. fragilis species of the group. Chloramphenicol, metronidazole and tinidazole were still active against 100% of the isolates at their breakpoint values (8 micrograms/ml).

The susceptibility of C. difficile isolated at the Department of Medical Microbiology of the University of Zurich to a wide selection of antibacterial, antimycobacterial and antifungal agents was tested in vitro. Great differences in susceptibility were found against chloramphenicol, clindamycin, erythromycin, rifamycin, and tetracycline. Resistance to clindamycin and erythromycin could always be transferred jointly to a susceptible C. difficile strain by mixed culture on filters at low frequencies (1 X 10(-8) to 4 X 10(-8) per donor cell). Transfer of tetracycline resistance occurred at frequencies of 3 X 10(-7) to 5 X 10(-7). Chloramphenicol and rifamycin resistance could not be transferred in the system used (frequencies less than 10(-8)). Although a total of 38,000 colonies was screened by various methods known to affect plasmid replication, resistance to chloramphenicol, clindamycin, erythromycin, rifamycin, and tetracycline could not be eliminated. No connection between plasmid DNA and antimicrobial resistance could be established. Especially, no plasmid DNA was involved in the transfer of resistance determinants from resistant to susceptible strains.

Obligately anaerobic bacteria are important in terms of their role as medical pathogens as well as their degradative capacities in a variety of natural ecosystems. Two major anaerobic genera, Bacteroides and Clostridium, are examined in this review. Plasmid elements in both genera are reviewed within the context of conjugal transfer and drug resistance. Genetic systems that facilitate the study of these anaerobic bacteria have emerged during the past several years. In large part, these developments have been linked to work centered on extrachromosomal genetic systems in these organisms. Conjugal transfer of antibiotic resistance has been a central focus in this regard. Transposable genetic elements in the Bacteroides are discussed and the evolution and spread of resistance to lincosamide antibiotics are considered at the molecular level. Recombinant DNA systems that employ shuttle vectors which are mobilized by conjugative plasmids have been developed for use in Bacteroides and Clostridium. The application of transmission and recombinant DNA genetic systems to study these anaerobes is under way and is likely to lead to an increased understanding of this important group of procaryotes.

Bacteroides are Gram-negative, obligate anaerobes that are present in high concentrations within the intestinal tracts of humans and animals. Bacteroides are also important opportunistic pathogens of humans and animals. Methods for genetic manipulation of these important organisms have only recently begun to emerge. Shuttle vectors which can be transferred by conjugation between Escherichia coli to Bacteroides are now available. A method for transforming some strains of Bacteroides has been developed. Two Bacteroides transposons, Tn4351 and Tn4400, have been found and one of them, Tn4351, has been used for transposon mutagenesis of Bacteroides. Several different Bacteroides genes have now been cloned, including a gene that codes for resistance to clindamycin, genes that code for polysaccharidases (chondroitin lyase and pullulanase), and a gene that codes for a fimbrial subunit. These cloned genes have been used to study the organization and regulation of Bacteroides genes.

DNA sequence analysis of a portion of an EcoRI fragment of the Bacteroides fragilis R plasmid pBF4 has allowed us to identify the macrolide-lincosamide-streptogramin B resistance (MLSr) gene, ermF. ermF had a relative moles percent G + C of 32, was 798 base pairs in length, and encoded a protein of approximately 30,360 daltons. Comparison between the deduced amino acid sequence of ermF and six other erm genes from gram-positive bacteria revealed striking homologies among all of these determinants, suggesting a common origin. Based on these and other data, we believe that ermF codes for an rRNA methylase. Analysis of the nucleotide sequences upstream and downstream from the ermF gene revealed the presence of directly repeated sequences, now identified as two copies of the insertion element IS4351. One of these insertion elements was only 26 base pairs from the start codon of ermF and contained the transcriptional start signal for this gene as judged by S1 nuclease mapping experiments. Additional sequence analysis of the 26 base pairs separating ermF and IS4351 disclosed strong similarities between this region and the upstream regulatory control sequences of ermC and ermA (determinants of staphylococcal origin). These results suggested that ermF was not of Bacteroides origin and are discussed in terms of the evolution of ermF and the expression of drug resistance in heterologous hosts.

The emergence of beta-lactamase producing bacteria in the microflora in the oropharyngeal cavity was studied in ten healthy volunteers treated with 1 g phenoxymethylpenicillin b.i.d. for ten days. Beta-lactamase activity in saliva was also investigated. A significant increase in the number of beta-lactamase producing strains of Bacteroides species and Fusobacterium nucleatum was observed. One beta-lactamase producing Staphylococcus aureus strain was recovered in one of the volunteers before the penicillin administration started and three Staphylococcus aureus strains produced beta-lactamase after ten days of antibiotic treatment. Beta-lactamase-production in Haemophilus influenzae, Haemophilus parainfluenzae or Branhamella catarrhalis was not observed before, during or after the antibiotic treatment. Beta-lactamase activity was noted in the broth cultures from one volunteer colonized with a beta-lactamase producing Escherichia coli strain. Beta-lactamase activity in saliva was observed in all volunteers, the activity increasing significantly in parallel to the increase of beta-lactamase producing bacterial strains. Beta-lactamase activity in saliva was completely inhibited in vitro by clavulanic acid and p-chloromercurbenzoate and about 70-80 per cent of the activity was inhibited by cefoxitin. The increase of beta-lactamase producing bacteria in the oropharynx as a consequence of penicillin treatment raises doubt as to whether penicillin is the drug of choice in the treatment of tonsillitis caused by group A streptococci when previous treatment has failed.

Broth disk elution procedures represent one of the most practical means for clinical laboratories to perform routine antibiotic susceptibility tests on anaerobic bacteria. The accuracy of five disk elution test methods and media (including the one to be proposed by the National Committee for Clinical Laboratory Standards) was evaluated for the testing of newer beta-lactam antibiotics, including cefoperazone, cefotaxime, cefoxitin, ceftazidime, ceftizoxime, moxalactam, and piperacillin. Various numbers of antibiotic disks were used to achieve disk elution test concentrations which approximated the highest MIC termed susceptible by the Food and Drug Administration. A group of 88 anaerobes representing many different species was tested in parallel by the five disk elution methods and the National Committee for Clinical Laboratory Standards reference agar dilution procedure. Overall, full agreement between the reference agar dilution MICs and the disk elution category results was 88.3% for PRAS BHI, 84.5% for Schaedler, 85.7% for thioglycolate, and 87.4% for Wilkins-Chalgren broth. Essential agreement (+/- 1 twofold MIC increment from the disk elution concentration) was achieved with 94.6% of PRAS BHI tests, 94.3% of Schaedler tests, 93.6% of thioglycolate tests, and 95.7% of Wilkins-Chalgren tests. Due to growth failures with a number of isolates and difficulties in interpreting results, the use of Wilkins-West broth was discontinued after approximately one-half of the isolates had been tested. The majority of errors with all of the disk elution methods occurred with isolates (most notably members of the Bacteroides fragilis group) having MICs near the single test concentrations used in the disk methods. With the notable exception of tests for the B. fragilis group, the disk elution methods offered acceptable accuracy with the newer beta-lactam antibiotics tested in this study.

Erythromycin is a macrolide that acts by inhibiting the translocation reaction during protein synthesis. Erythromycin is inactive against the Enterobacteriaceae and Pseudomonas aeruginosa except under alkaline conditions. Erythromycin is active against most gram-positive bacteria; some gram-negative bacteria, including Neisseria, Bordetella, Brucella, Campylobacter, and Legionella; and Treponema, Chlamydia, and Mycoplasma. The emergence of resistance to erythromycin is closely associated with its use and is often plasmid mediated. After its oral or parenteral administration, erythromycin diffuses readily into intracellular fluids and is actively concentrated intracellularly by polymorphonuclear leukocytes and alveolar macrophages.

High-level clindamycin resistance in Bacteroides species was investigated by measuring zone sizes surrounding 2 micrograms clindamycin and 60 micrograms erythromycin discs, using a nonstandardized disc diffusion method, and by determining minimal inhibitory concentrations (MIC). The absence of a zone of inhibition surrounding either disc was predictive for all isolates having high-level resistance to both antibiotics (MIC greater than 256 micrograms/ml), characteristic of macrolide-lincosamide-streptogramin (MLS) cross-resistance. Although zone size could not be used as an absolute predictor of MIC, a clindamycin zone diameter of less than 17 mm was suggestive of strains with a moderate level of clindamycin resistance (MIC greater than or equal to 8 micrograms/ml), regardless of erythromycin zone size. Disc diffusion testing using a combination of clindamycin and erythromycin discs can be a useful screening method for detection of clindamycin-resistant Bacteroides species, occurring either alone or as part of the MLS resistance phenotype.

Results presented in this paper have shown that a widely distributed LM-resistance determinant is present on at least 3 distinct Bacteroides R-plasmids. In fact these plasmids bear no homology outside of the defined regions implicated in the LM resistance. This resistance determinant on pBF4, pBFTM10, and pBI136 is located within DNA segments bounded on each side by a directly repeated sequence of more than 500 bp. The intervening sequences of these 3 elements are variable, and range in size from about 3.7 kb to 7.2 kb (Fig. 7). Apart from the EcoRI/AvaI restriction sites which characterize the repeated sequence, there is a notable lack of common restriction sites within these elements. These results suggest that the elements do possess a certain degree of structural similarity but significant evolutionary divergence has occurred. The presence and location of the directly repeated sequences, their association with specific deletions, and their association with an antibiotic-resistance determinant, are features common to many antibiotic-resistance transposons described for other prokaryotes. In addition, these elements are highly mobile, being found on a number of R-plasmids. The unique relationship between pBF4, pBI106, and pBI136 described here is a clear indication of the potential for these DNA sequences to move from one molecule to another. Given the extensive dissemination and the genetic and structural characteristics described above, it seems likely that the LM-resistance determinant is carried on transposon-like elements present in pBF4, pBFTM10, and pBI136. However, further experimentation will be necessary to document the transposition event. Bacteroides strains such as B. fragilis V503, possess a transmissible LM-resistance determinant which does not appear to be associated with detectable extrachromosomal elements (5,9,10). Presently, a number of strains of this type have been found over a wide geographic area. The LM-resistance genes associated with these strains are apparently similar to the one carried on the Bacteroides R-plasmids because homology between the 2 has been observed. However, it is important to note that within the limits of Southern filter blot hybridization, neither V503 nor its transconjugants possess the directly repeated sequence found on the LM-resistance plasmids (Fig. 8). The elusive nature of the V503 LM-resistance elements presents an intriguing problem. One model that has been proposed is that these resistance determinants reside on a conjugative transposon similar to Tn916 of Streptococcus faecalis (3).(ABSTRACT TRUNCATED AT 400 WORDS)

Two different species of clindamycin-resistant Bacteroides were isolated from the same infection. One isolate contained a single 15-kb plasmid (pCP1) which encoded transferable clindamycin resistance. pCP1 appears similar to the Bacteroides clindamycin resistance plasmid pBFTM10 isolated independently by F.P. Tally, D.R. Snydman, M.J. Shimell, and M.H. Malamy (1982, J. Bacteriol. 151, 686-691). The second strain had a 10-kb plasmid (pCP2) but did not transfer resistance. DNA hybridization studies revealed that pCP1 shares a 5-kb region of homology with the B. fragilis R plasmid pBF4 studied by R.A. Welch and F.L. Macrina (1981, J. Bacteriol. 145, 867-872). This region in both plasmids was shown to be bounded by homologous direct repeats and contains the putative clindamycin resistance determinant. pCP1 and pCP2 were found to share extensive homology but sequences homologous to the clindamycin resistance region were missing from pCP2 and found instead in the whole cell DNA of the host strain. These results identify a transposon-like structure on Bacteroides clindamycin resistance plasmids.

Bacteroides ovatus IB106 contained two plasmids, pBI106 (46 kilobases) and pBI136 (82 kilobases). Transmissible clindamycin-erythromycin resistance (Ccr) was mediated by pBI136 , whose Ccr determinant was closely related to the determinant on the Bacteroides R plasmids pBF4 and pBFTM10 . Hybridization studies showed that pBI106 was not involved in Ccr transfer, but it shared extensive homology to pBF4 with the exception of the pBF4 region implicated in Ccr.

One hundred and eight strains of Clostridium perfringens were collected from polluted waters and tested for their drug resistance and bacteriocinogeny . Thirty of the strains were tetracycline resistant and 36 were bacteriocinogenic. Only one strain possessed both tetracycline resistance and bacteriocinogen . Transfer experiments on tetracycline resistance and bacteriocinogeny were carried out with several selected strains among these isolates. Both tetracycline resistance and bacteriocinogeny were shown to be transferable by conjugation-like procedures. Transfer experiments on these two properties revealed that there were two types of recipient strains among the isolates in regard to the acceptance of tetracycline resistance and bacteriocinogeny . In one type of the recipient strain, these two properties seemed to be incompatible although in the other, both tetracycline resistance and bacteriocinogeny were transferred by two-step mating. Transcipients of this two-step mating retransferred their tetracycline resistance but not bacteriocinogeny to another recipient strain. The existence of an incompatibility-like relationship between tetracycline resistance and bacteriocinogeny in one recipient strain, and the inability of these two properties to be co-transferred from the two-step transcipients , suggest a certain interrelationship between the factors responsible for tetracycline resistance and bacteriocinogeny .

High-level resistance to clindamycin can be accurately detected by the Wadsworth disk identification test. Of the 98 isolates of the Bacteroides fragilis group that were tested, 90 were inhibited by the 60-micrograms erythromycin disk and had clindamycin minimal inhibitory concentrations of less than or equal to 3.2 micrograms/ml. Of the remaining eight isolates, all were resistant to the erythromycin disk and had clindamycin minimal inhibitory concentrations of greater than or equal to 100 micrograms/ml.

Over the last decade, developments in recombinant deoxyribonucleic acid techniques and molecular biology have revolutionized bacterial genetics, creating vast, new potential uses of bacteria (as well as animal and plant cells) that were not even considered previously. Bacterial production of hormones is but one example. With bacterial species with well developed genetic systems, such as Escherichia coli, it is now possible genetically to "design" or "engineer" bacterial strains having specific characteristics. One reasonable future approach toward improvement of animal agriculture would be manipulation of the rumen ecosystem via the use of genetically modified ruminal bacteria, but significant obstacles exist with this approach. Genetic systems of ruminal and of anaerobic bacteria of the mammalian gastrointestinal tract, in general, have not been studied and are largely unknown. In this paper, the various criteria for possible establishment of recombinant deoxyribonucleic acid systems in ruminal bacteria are outlined. Secondly, applications for utilizing genetically engineered ruminal bacteria to control digestion of specific feedstuffs, to regulate specific fermentation products, and to control growth of specific bacterial species are discussed.

Resistance to clindamycin, erythromycin, and streptogramins in Clostridium difficile could be transferred to a susceptible strain by mixed culture at low frequencies (1 x 10(-8) to 4 - 10(-8) per donor cell). Transfer of tetracycline resistance occurred at frequencies of 3 x 10(-7) to 5 x 10(-7). No plasmid DNA involved in these transfers could be detected.

A rapid and easy final purification method is described for the isolation of plasmids from B. fragilis. Using RbTCA density gradient centrifugation in an airfuge ultracentrifuge ccc plasmid DNA can be separated from RNA, residual chromosomal DNA, linear and oc plasmid DNA. Pure ccc plasmid DNA is obtained from cultures of between 1 ml and 2 l in less than one day.

A systematic analysis of the plasmid content of intestinal Bacteroides spp. was made. Eight of fifteen clinical isolates and seven of nineteen normal rectal flora isolates examined contained small plasmids (less than 5 Mda). The majority of these small plasmids could be assigned to three homology classes by Southern hybridization. Further investigation demonstrated the presence of subclasses within one of the homologous classes. Class I plasmids were 1.8 Mda, Class III were 3.7 Mda, and Class II plasmids were of three different molecular weights: 2.6 Mda (IIA), 3.2 Mda (IIB), and 4.0 Mda (IIC). Among representative plasmids from each class there was remarkable sequence similarity based upon digestion patterns obtained using the restriction endonucleases, AluI and DdeI. In addition, similar polypeptide products were observed using purified Class I plasmids as template in an Escherichia coli in vitro coupled transcription-translation system. Small plasmids were found in seven of the ten recognized species of intestinal bacteroides, indicating that no species barrier exists for these plasmids. In addition, since plasmids from all three classes were found together in a single isolate, it was concluded that these plasmids are not incompatible.

Two clindamycin-erythromycin resistance transfer factors of Bacteroides fragilis, pBF4 and pBFTM10, were analyzed for regions of DNA homology. Although the plasmids were derived from different clinical isolates of B. fragilis and have different sizes, they showed homology in the clindamycin-erythromycin resistance region; no homology could be detected outside of this region.

The susceptibility of 105 strains of Bacteroides to arsenate, silver, nickel, cobalt, lead, cadmium, chromium, and mercuric ions was tested by using an agar dilution technique. All strains of Bacteroides were multiply resistant. Bacteroides fragilis was significantly more resistant to silver ions than other Bacteroides species and slightly more resistant to several other ions. Arsenate was the most active ion tested, with 75% of strains being susceptible. All 25 strains of B. fragilis tested against silver sulfadiazine were susceptible.

Clindamycin resistance in Bacteroides fragilis was examined in 507 strains isolated from 1973 to 1981. Three groups were recognized: highly susceptible (minimum inhibitory concentration [MIC] less than or equal to 0.125 microgram/ml), intermediately susceptible (MIC = 0.25 to 4 micrograms/ml), and highly resistant to (MIC greater than or equal to 8 microgram/ml). The incidence of high-level resistance (1.8%) had not changed during this period. Only 8 of 17 isolates reputed to be highly clindamycin resistant that were referred to our laboratory proved to be highly resistant (MICs greater than or equal to 32 microgram/ml), whereas the other 9 were intermediately susceptible. Analysis of 2- and 10-microgram clindamycin disks for determining the susceptibility of B. fragilis revealed a high false-resistance rate with the 2-microgram disk, most errors occurring with the intermediate group. There was no false resistance with the 10-microgram disk. When disk diffusion susceptibility of B. fragilis is employed, we recommend the 10-microgram disk to predict accurately the susceptibility of B. fragilis to clindamycin.

Cefoxitin resistance, but not resistance to clindamycin, erythromycin, lincomycin, or tetracycline, was transferred by a conjugation-like process from Bacteroides thetaiotaomicron UN101, a clinical isolate haboring four kinds of plasmids, to other Bacteroides species. Of a sample of 20 cefoxitin-resistant transconjugants, 8 contained all 4 plasmids, 10 contained 1 to 3 plasmids, and 2 contained no plasmids.

Bacteroides fragilis TMP10, which is clindamycin-erythromycin resistant (Clnr) and tetracycline resistant (Tetr), contains several plasmids and is capable of transferring drug resistance markers to suitable recipients. We were able to separate a 14.6-kilobase self-transmissible Clnr plasmid, pBFTM10, from the other plasmids of TMP10 in a tetracycline-sensitive recipient strain, B. fragilis TM4000. All Clnr transconjugants acquired an unaltered pBFTM10 and became plasmid donor strains. Transfer is proposed to occur by conjugation since it required to cell-to-cell contact of filter matings and was insensitive to DNase, but sensitive to chloroform treatment of donor cells. The efficiency of transfer of pBFTM10 in a Tets background (TM4003) was not affected by pretreatment of donor cells with clindamycin. A spontaneously occurring Clns derivative, pBFTM10 delta 1, suffered a deletion of DNA, which included a 4.4-kilobase EcoRI fragment. A complex interaction between the autonomous plasmid pBFTM10 and a tetracycline transfer element also present in strain TMP10 was observed since pretreatment of this donor with tetracycline or clindamycin resulted in a marked increase in transfer of both tetracycline and clindamycin resistance.

Bacteroides fragilis V479-1 (also designated strain 92) has previously been shown to contain a conjugative plasmid, pBF4, that specifies resistance to clindamycin (Cc). A report of inducible tetracycline (Tc) resistance in this strain suggested that this phenotype was also plasmid associated (G. Privitera et al., Nature [London] 278:657-659, 1979) and prompted further investigation. Mating experiments with V469-1 and a Cc-sensitive derivative of V479-1, V598, showed that Tc resistance transfer occurred by a conjugation-like event which was insensitive to DNase, was not mediated by donor culture cell-free filtrates, and required cell-to-cell contact. Results from transfer experiments with V479-1 indicated that Tc and Cc resistance determinants were not linked and segregated independently in matings. Progeny recovered from matings with the V479-1 or V598 donor strain were able to transfer the Tc resistance marker in secondary crosses. Tc resistance transfer from V479-1 or V598 was greatly stimulated by pregrowth in the presence of Tc but not Cc. pBF4-mediated Cc resistance transfer was not affected by pregrowth in the presence of Cc or Tc. Filter blot DNA hybridization studies revealed that pBF4 sequences were not involved in either the Tc resistant phenotype or its associated conjugal transfer properties. The Tc resistance transfer element was not associated with pBF4 or any other extrachromosomal DNA element.

Tetracycline resistance of three Bacteroides fragilis strains was shown to be inducible by subinhibitory concentrations of tetracycline. Tetracycline resistance markers could be transferred to another B. fragilis strain by filter mating. The transferability was inducible by subinhibitory concentrations of tetracycline and did not take place in the absence of tetracycline. The optimum concentration of tetracycline for induction of transfer was about 2 microgram/ml. The transfer was shown to be a conjugation-like process requiring cell-to-cell contact between donor and recipient. Screening of parental donor strains for the presence of plasmid DNA did not demonstrate any detectable plasmids in two of the strains. A 3.0-megadalton plasmid, designated pBY5, was present in the third donor strain. Mobilization of pBY5 by another plasmid (pBF4) showed that pBY5 did not carry the genes responsible for tetracycline resistance. It appears that the genes responsible for resistance to tetracycline as well as those responsible for conjugal transfer may be carried on the chromosome in all three donor strains.

Resistance to tetracycline and lincosamide antibiotics was transferred en bloc from a strain of Bacteroides fragilis (V503) to a plasmidless strain of Bacteroides uniformis (V528) during in vitro filter matings. Resistance transfer was detected at frequencies of 10(-5) to 10(-6) drug-resistant progeny per input donor cell and was dependent on cell-to-cell contact of donors and recipients. Transfer was insensitive to DNase and was not mediated by chloroform- or filter-sterilized donor broth cultures. A determinant for resistance to cefoxitin in V503 was not transferred in this system. V503 contained a 3.7 x 10(6)-dalton plasmid (pVA503). Drug-resistant progeny of V503 x V528 matings usually contained pVA503, but up to 20% of the total progeny of such crosses were plasmid free. Filter blot DNA hybridization studies (Southern method) confirmed that pVA503 was not integrated into the host chromosome of the plasmidless progeny. Drug-resistant progeny from V503 x V528 matings (with or without pVA503) conjugally transferred clindamycin resistance an tetracycline resistance to a suitable recipient strain. None of the drug resistance determinants of V503 were affected by treatment with standard plasmid curing regimens, and methods designed to detect very large plasmid molecules failed to suggest the involvement of extrachromosomal DNA in this resistance transfer system. The well-characterized Bacteroides R plasmid, pBF4 (conferring clindamycin resistance), was found to share hybridizing sequences with bulk cellular V503 DNA when examined by filter blot hybridization. Similarly sized sequences were found in drug-resistant progeny recovered from matings. Neither of the two pBF4 derivatives carrying deletions that abolished clindamycin resistance hybridized with V503 DNA.