| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Institute of Medical Microbiology and Hygiene, Technical University of Dresden, Fetscherstrasse 72, Dresden, Germany
Correspondence
Herbert Schmidt
hschmidt{at}uni-hohenheim.de
Received 10 April 2006
Accepted 25 October 2006
4 mg l–1, 18 strains had the same MIC for ceftazidime, and nine strains had this MIC of >4 mg l–1 for both antibiotics. The ESBL phenotypes often coincided with ciprofloxacin or gentamicin resistance.
Abbreviations: CAZ, ceftazidime; CFP, cefepime; CIPR, ciprofloxacin; CTX, cefotaxime; ESBL, extended-spectrum beta-lactamase; GENT, gentamicin; ICU, intensive care unit; IMIP, imipenem; PIP, piperacillin; PITA, piperacillin/tazobactam.
Present address: Institute of Food Science and Biotechnology, Department of Food Microbiology, University of Hohenheim, Garbenstrasse 28, D-70599 Stuttgart, Germany. 
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Transfer of resistance determinants.
In order to investigate the conjugative transfer of resistance determinants, broth mating experiments were performed. The plasmid-free, rifampicin-resistant E. coli strain A15 R– was used as an acceptor (Elwell & Falkow, 1986). This strain was kindly provided by Branka Bedeni
, ‘A. 
tampar’ Medical School, University of Zagreb, Croatia. Overnight Luria–Bertani (LB) broth cultures of the acceptor and the respective donor strain were mixed in a ratio of 1 : 2 and incubated for 18 h at 37 °C without shaking. The selection of transconjugants was performed by spreading 100 µl of the mixture onto MacConkey agar plates containing 200 mg rifampicin l–1 as well as 2 mg CAZ l–1 or cefotaxime (CTX). As some K. pneumoniae strains developed rifampicin resistance during mating, colonies were subcultured on citrate agar (Simmons citrate agar, Oxoid) to detect citrate-positive, rifampicin-resistant K. pneumoniae donor strains.
Preparation of plasmid DNA. Plasmid DNA was obtained using a modified alkaline-lysis method (Sambrook et al., 1989), with slight modifications. Briefly, a further purification step with phenol/chloroform was added before DNA precipitation in 2-propanol. In some strains, the quality of this preparation was not sufficient for the subsequent restriction analysis, probably because of the presence of capsular polysaccharides. In these cases, bismuth (III) pentahydrate (Sigma-Aldrich) and sodium salicylate were added to the LB medium before incubation in final concentrations of 0.5 and 2.5 mM, respectively, as described by Domenico et al. (1992). After incubation and before alkaline lysis, EDTA (pH 10) was added to a final concentration of 5 mM to bind bismuth ions. Unlike the method of Domenico et al. (1992), the treatment with a detergent was omitted. Plasmids were separated on a 0.8 % (w/v) agarose gel in Tris/borate/EDTA buffer for 1.5 h at 120 V.
Restriction of plasmid DNA. Aliquots of 10 µl of the plasmid DNA preparation were subjected to restriction with 1 µl EcoRI (10 U µl–1; Fermentas), 3 µl Buffer EcoRI+ (Fermentas), and 16 µl nuclease-free deionized water for 2 h at 37 °C. DNA was separated on a 1 % (w/v) agarose gel for 17 h at 30 V together with the molecular mass standard peqGOLD DNA-Sizer IX (Peqlab Biotechnologie).
PCR. In order to detect genes encoding Ambler class A beta-lactamases (bla), a standard PCR was performed with plasmid DNA as a template. For amplification of genes encoding SHV beta-lactamases (blaSHV), primers MN I (5'-CGC CGG GTT ATT CTT ATT TGT CGC-3') and MN II (5'-TCT TTC CGA TGC CGC CGC CAG TCA-3') were used (Nüesch-Inderbinen et al., 1996). The thermal cycling conditions used for the PCR with these primers included 30 cycles at 94 °C for 30 s, 68 °C for 60 s, and 72 for 60 s, with a final extension of 72 °C for 300 s. The 1016 bp PCR product contained the entire ORF. For amplification of genes encoding TEM beta-lactamases (blaTEM), primers TEM-F (5'-ATA AAA TTC TTG AAG ACG AAA-3') and TEM-R (5'-GAC AGT TAC CAA TGC TTA ATC A-3') were used with thermal cycling conditions as described elsewhere (Wu et al., 2001). The 1080 bp amplicon reached from 214 bp upstream of the start codon to the stop codon. In order to detect genes for CTX-M beta-lactamases (blaCTX-M), a PCR was performed with primers and conditions described elsewhere (Edelstein et al., 2003). The amplicon is a 544 bp intragenic fragment. Ten microlitres of PCR product was electrophoresed in a 1 % (w/v) agarose gel for 1 h at 140 V together with a Smartladder DNA standard (Eurogentec).
DNA sequencing. Since the above-mentioned MN and TEM primers are specific for all blaSHV and blaTEM, including the non-ESBL variants, it was necessary to sequence the PCR products. For this purpose, PCR products were purified with a QIAquick PCR purification kit (Qiagen). Sequencing reactions were performed with the primers used for PCR detection using a BigDye Terminator Cycle Sequencing Ready Reaction Premix (Applied Biosystems). After purification with AutoSeq-G50 columns (Amersham Biosciences), products of the sequencing reaction were subjected to direct sequencing on an ABI Prism 377 automated sequencer (Applied Biosystems). The sequences were analysed using the program Bioedit (Hall, 1999). Chromatograms were visually inspected for double peaks as signs of the presence of different bla genes from the same type.
Differentiation between blaESBL-SHV and blanon-ESBL-SHV in a PCR product. In the case of double peaks within the chromatograms of blaSHV sequences, a 40 µl aliquot of the corresponding PCR product was digested with 5 µl NheI (10 U µl–1; Fermentas) and 5 µl 10x Buffer Y+/Tango (Fermentas) with BSA for 3 h at 37 °C. Digested DNA was separated on 1.5 % (w/v) agarose gels for 2 h at 140 V. As described elsewhere (Nüesch-Inderbinen et al., 1997), most blaESBL-SHV genes contain the NheI restriction site, in contrast to blanon-ESBL-SHV genes. Thus, both fragments could be separated from each other. The blanon-ESBL-SHV fragment was extracted from the gel under visualization with UV light. The extraction was performed using a QIAquick Gel Extraction kit (Qiagen), according to the guidelines of the manufacturer. The product was directly subjected to sequencing, as described above, without any further purification. In this way, mixed sequence signals at characteristic positions inside ORFs of some SHV sequences could be assigned to ESBL and non-ESBL enzymes, respectively.
RFLP analysis of blaCTX-M. A further subgrouping of blaCTX-M was performed according to Edelstein et al. (2003). For this purpose, 10 µl of the products of CTX-M-specific PCR were digested with 0.9 µl PstI (10 U µl–1), 0.2 µl PvuII (20 U µl–1), 2 µl 10-fold NEBuffer 3 (enzymes and buffer, New England BioLabs) and 6.9 µl nuclease-free distilled water at 37 °C for 3 h. Restriction fragments were electrophoresed in a 3 % (w/v) agarose gel for 2 h at 140 V together with the DNA standard Smartladder SF (Eurogentec).
Determination of MICs. In order to phenotypically characterize the resistance properties of the strains, MICs were determined for CTX (Aventis), CAZ (GlaxoSmithKline), cefepime (CFP; Bristol-Myers Squibb), piperacillin (PIP; Ratiopharm) in combination with tazobactam (Wyeth Pharma), imipenem (IMIP; Merck, Sharp and Dohme), gentamicin (GENT; Merck) and ciprofloxacin (CIPR; Bayer). MICs were determined by the agar dilution method on Mueller–Hinton agar plates containing antibiotics in concentrations of a geometric series according to the DIN 58940-6 guidelines of the German Institute for Standardization (Deutsches Institut für Normung, 1989). MIC determination for the PIP/tazobactam (PITA) combination was tested using a fixed concentration of 4 mg tazobactam l–1 and varying concentrations of PIP.
|
RESULTS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Most strains (19) were isolated from clinical specimens originating from patients' airways, followed by urine (12), skin (three), body cavities (three) and blood (two). All 39 strains belonged to the family Enterobacteriaceae.
|
A further four raw DNA sequences contained mixed signals in the respective electropherograms at characteristic positions, all at amino acid position 238 according to the numbering provided by Ambler (Ambler et al., 1991). This problem was resolved by NheI restriction endonuclease digestion, as described above. It resulted in the detection of two SHV alleles in a single strain, as follows: blaSHV-2 paired with blaSHV-1 twice; blaSHV-2 paired with blaSHV-28 once; blaSHV-12 paired with blaSHV-1 once. The sequence of K. pneumoniae KP01 revealed a non-functional start codon, damaged by a T to G mutation; apart from that it was identical to the deduced SHV-1 sequence. In conclusion, the ESBL phenotype could be explained by SHV ESBL production in 24 strains only, in spite of a positive SHV PCR result in 26 cases. Ent. aerogenes EA05 (single blaSHV-1), K. pneumoniae KP12 (single blaSHV-28) and KP01 (defective start codon) putatively express other resistance mechanisms in addition to the production of SHV-type enzymes. Coding and non-coding genetic polymorphisms are summarized in Table 1. Since all ESBL genes from the same SHV type were identical among themselves, only one representative example is shown for all strains of the same ESBL type.
|
CTX-M-specific PCR and RFLP analysis of PCR products
As ESBLs from type CTX-M have been increasing in frequency in Western and Central Europe for some years (Alobwede et al., 2003), further screening for this type was performed. With respect to the genetic heterogeneity within the group of blaCTX-M, a subgrouping into five clusters (CTX-M-1, CTX-M-2, CTX-M-8, CTX-M-9 and recently discovered CTX-M-25) was performed using sequence homology criteria (Bonnet, 2004). To detect the presence of CTX-M ESBL genes in the strains used in this study, a primer pair sensitive to all blaCTX-M genes independent of cluster affiliation was chosen. However, this primer pair produces false-positive results for the species-specific chromosomally encoded beta-lactamase (bla) genes of some Gram-negative strains (Edelstein et al., 2003). Among the species used in this work, only the K. oxytoca strains with their chromosomal beta-lactamase K1 encoded by blaOXY were affected. In fact, an amplicon could be obtained by this PCR analysis from DNA preparations from all these strains. In addition, a positive result could be obtained from eight E. coli, two K. pneumoniae and the Ent. aerogenes strains. For all non-K. oxytoca strains, RFLP analysis of PCR products was performed in order to obtain information about the cluster affiliation of the genes found. Restriction patterns were compared with the expected ones described in the original publication (Edelstein et al., 2003). For the new cluster of CTX-M-25-like enzymes (CTX-M-25 and CTX-M-26), which was not discovered at the date of development of the method, a restriction map was created with Bioedit. The digestion of such a CTX-M-25-like PCR product would result in fragments of 422 and 120 bp in length, which can be distinguished from the restriction patterns of the other clusters. The PCR products of all strains, except two E. coli strains, displayed the fragment of a blaCTX-M-1-like gene. Two E. coli isolates, EC25 and EC27, carried a gene belonging to the CTX-M-9 cluster. The PCR products of K. oxytoca strains were sequenced using the same primer pair as for PCR detection. This procedure was performed to identify mixed signals in sequence chromatograms that would have occurred in the case of the simultaneous presence of a blaOXY encoding K1 and a blaCTX-M. All sequences of K. oxytoca strains were clear cut in encoding K1. Table 2
shows the detected enzyme genes, the plasmid restiction patterns and the MIC values for all strains.
Mating experiments
Transconjugants could be obtained from 11 strains only. If a mating experiment was unsuccessful, it was repeated with CTX or CAZ as selective agent. Transconjugant plasmid DNA was obtained and PCRs were performed in the same way as in the case of donor strains. Two SHV-5 producers, three SHV-12 producers, four of the strains producing a CTX-M-1-like enzyme and both strains with the blaCTX-M-9-like gene transferred their ESBL gene to E. coli A15 R–. Five of the 11 donors carried an additional TEM-1, but only one of them co-transferred it to the acceptor strain. SHV PCR of plasmid DNA from transconjugants of two CTX-M producers with an additional blanon-ESBL-SHV was negative.
Analysis of plasmids
In order to detect the presence of common restriction bands among plasmids of strains with the same ESBL type and among their transconjugants, plasmid DNA was isolated and digested with EcoRI. Five different restriction patterns were found among SHV-2 producers, the band sizes of which ranged from approximately 70 to 130 kb. Four of the restriction patterns appeared to be paired. Two of the pairs with identical plasmid restriction patterns were isolated from the same departments within hospital A. In the case of the other two paired restriction patterns, one pair of strains was isolated at hospital A and one at hospital B. Analysis of unrestricted plasmid DNA of SHV-5 producers revealed at least two single bands, which allowed us to infer the presence of at least two independent plasmids. All restriction patterns of SHV-5-producing strains plasmid DNA had common elements, which can be inferred from Fig. 1. Strains EC20 and KP28 were two SHV-5 producers from which transconjugants could be obtained. All restriction bands of the donor pattern were present in transconjugants in both cases (110 and 160 kb). SHV-5-producing strains were isolated at both hospitals. Among SHV-12-producing strains, no common plasmid DNA restriction pattern could be observed, other than a common 6 kb band in KP10, CF18 and KO26. Plasmid sizes ranged from 45 to 160 bp. The CTX-M producers EC07, EC16 and EC17 had a common restriction pattern, probably because of three common bands in electrophoresis of their undigested plasmid DNA. In addition to these three bands there were a further one at EC07, two at EC16 and four at EC17. All three strains were isolated from the same patient of hospital A on two different dates. The two strains EC25 and EC27 producing a CTX-M-9-like beta-lactamase had an identical plasmid restriction profile (
140 kb). Both were isolated from intensive care units (ICUs) of hospital A. As was the case with the SHV-12 producers, there was no common restriction pattern observable upon gel electrophoresis of digested plasmid DNA of the other CTX-M-producing strains. Total plasmid DNA of the strains ranged from approximately 40 to 90 kb in size.
|
. Depending on their MICs, a resistance type was assigned to CTX (MIC for CAZ greater than twofold the MIC for CTX) and CAZ (MIC for CAZ greater than fourfold the MIC for CTX) (Bedeni et al., 2001). The resistance type correlated with ESBL genotype as follows. All strains producing SHV-2 and CTX-M beta-lactamases were assigned to a CTX resistance type, while SHV-5- and SHV-12-producing strains displayed a CAZ type. KP24 and the transconjugant from EC20, R+20, were exceptions. Surprisingly 10 strains had a MIC 2 mg l–1 for CTX and 12 had the same for CAZ. The SHV-2 producer KP37 and the K1 carrier KO36 did not grow at this concentration, either on CTX or CAZ. In summary, 20 strains appeared to be susceptible to CTX (MIC 4 mg l–1) and 19 to CAZ (MIC 4 mg l–1). Nine strains were susceptible to both antibiotics according to the German standard tests. CFP was revealed to be more effective against ESBL producers; only four strains were resistant. The MICs towards PITA were distributed bimodally: 20 strains had a MIC between 2+4 and 16+4 mg l–1, while 17 strains still grew at concentrations of 512+4 mg l–1. Twelve strains were resistant towards GENT (MIC 
4 mg GENT l–1). KP28 transferred this resistance together with an ESBL to the donor strain as well as reduced susceptibility towards CIPR (MIC 2 mg CIPR l–1), in common with 12 other strains. All strains were completely susceptible to IMIP (MIC 2 mg IMIP l–1). |
DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
) as the one first described (Nüesch-Inderbinen et al., 1996) and all the other sequenced genes of this type published in the NCBI database. Amongst the SHV-12 producers analysed in this study we found no common plasmid restriction patterns. Probably, blaSHV-12 is part of a worldwide-disseminating mobile element. Besides K. pneumoniae, C. freundii (CF18), E. coli (EC29) and K. oxytoca (KO26) also contain SHV-12 genes. Moreover, the SHV-12 gene has been found in Serratia liquefaciens (GenBank accession no. AY273807), Enterobacter cloacae (AF462395) and Acinetobacter baumanii (AY259163). Although first identifications of CTX-M-type ESBLs occurred about 15 years ago (Bauernfeind et al., 1990), for years, dissemination has only been known in endemic areas of Asia, Southern America and Eastern Europe (Bonnet, 2004). Recently, dissemination of this ESBL type has been observed (Coque et al., 2002; Alobwede et al., 2003). Detection of CTX-M-type enzymes in Dresden underlines this trend. Most of the strains carried genes for a CTX-M-1-like enzyme. For CTX-M-3, the most important representative of this group, epidemic occurrence in Poland (Gniadkowski et al., 1998b) and Russia (Edelstein et al., 2003) is known. In those two studies, the two CTX-M-9-like carriers EC25 and EC27 had identical plasmid restriction profiles, in contrast to the CTX-M-1-like enzyme carriers found in this study. Both were isolated from ICUs of hospital A, implicating cross-infection. Their 160 kb plasmid could be transferred in vitro, suggesting further dissemination of this enzyme type. The most disseminated representative of this enzyme type, CTX-M-14, has already been found in Spain (Coque et al., 2002), France and the UK (Alobwede et al., 2003). Most of the strains able to transfer their resistance in vitro were SHV-12 or CTX-M producers, implicating the role of transconjugation in dissemination, while transconjugants could not be obtained from SHV-2 producers. A similarly low transconjugation success was reported from a molecular survey of SHV-producing Klebsiella strains in Switzerland (Nüesch-Inderbinen et al., 1996). In that study, 26 of 37 blaSHV-carrying strains were unable to transfer ESBL genes. In our survey, 19 of 24 strains did not transfer the ESBLs. In contrast, six of 11 CTX-M carriers were shown to transconjugate the CTX-M genes. Probably, clonal expansion is a more important mechanism in the dissemination of SHV-2 and SHV-5 genes.
All but one of the non-ESBL SHV sequences (blaSHV-1 in KP23) carried silent mutations that were different to those found in ESBL genes, making co-evolution improbable (Table 1). SHV-28 had already been identified from China (GenBank accession nos AF538324 and AY299299). This allele differs from SHV-1 by an amino acid substitution at position 7 according to Ambler's standard numbering scheme for class A beta-lactamases (Ambler et al., 1991). Since the mature protein starts with Ambler position 26 compared to SHV-1 (Péduzzi et al., 1989), there should not be any difference either in biochemical properties or in isoelectric point. The detected SHV-1 from Ent. aerogenes EA05 must have been plasmid encoded, because from this species only a chromosomally encoded Ambler class C beta-lactamase (AmpC) is known, which is produced only in small amounts, but is inducible. The same applies to C. freundii. The beta-lactamase K1 is the chromosomally encoded class A beta-lactamase of K. oxytoca. It is encoded by two genes, blaOXY-1 and blaOXY-2 (Mammeri et al., 2001). Intrinsically, K1 mediates only moderate resistance towards penicillins. However, at present, up to 20 % of isolates hyperproduce this enzyme in Europe. The consequence is a significant resistance towards ceftriaxone, CTX and aztreonam. This must be the case with the K. oxytoca strains in this study, as one can observe typical signs of K1 hyperproduction in the resistance patterns, such as resistance to PITA and at least reduced susceptibility to CTX (Table 2) (Livermore, 1995). Hyperproduction is more usual with blaOXY-2-encoded enzymes, and can be caused by point mutations within the promoter sequence (Jeong et al., 2001; Fournier et al., 1995). In the case of K. oxytoca KO26, in which an SHV-12 beta-lactamase had already been identified, K1 was encoded by blaOXY-1. In the case of the other four K. oxytoca strains, K1 was encoded by blaOXY-2. With respect to the Vitek 1 ESBL test as performed in our department, it is known that it cannot distinguish between a K1 hyperproducer and a ‘classical’ ESBL carrier (Leverstein-van Hall et al., 2002). Although K1 was assigned to group 2be (ESBL) within the functional beta-lactamase classification scheme of Bush–Jacoby–Medeiros (Bush et al., 1995), in fact, ESBL detection in K1 hyperproducers represents a false-positive result, because it is not necessary to regard such strains as resistant to all cephalosporins independently of their resistance patterns, as is the case with ESBL carriers. Moreover, these strains cannot transfer their resistance horizontally. KO26 demonstrates that K. oxytoca strains are potential ESBL producers, and false-negative ESBL test results should be regarded as more dangerous than false-positive results in that regard. Paterson et al. (2001) have described the higher risk of therapy failure in the treatment by oxyimino-cephalosporins of infections with ESBL-producing strains, even if the strains appear to be susceptible in vitro. The number of strains having a MIC defined as susceptible to CAZ or CTX underlines the necessity to test with both antibiotics and cut-offs below 4 mg l–1. Experts suggest 2 mg l–1 (Witte & Mielke, 2003; Geiss et al., 2003). CFP seems to be more effective than CTX and CAZ. One should consider the possibility of the decreasing efficacy of this beta-lactam due to inoculum effects and changes in the outer-membrane protein profile during therapy, particularly in the case of infections that are connected with high bacterial inocula in vivo, such as endocarditis, meningitis, septic arthritis, osteomyelitis, abscesses and other deep-seated infections (Thomson & Moland, 2001). The bimodal distribution of MICs towards PITA has recently been observed and investigated (Babini et al., 2003). It is caused by titration effects when testing with fixed concentrations of tazobactam. The distribution becomes unimodal if the ratio of piperacillin to tazobactam is kept constant at 1 : 8. Babini et al. (2003) emphasize the general susceptibility of ESBLs towards inhibition with tazobactam. If strains are susceptible to PITA in vitro, it should be considered as a therapeutic option. The prevalence of ESBL-producing strains decreases if hospitals reduce their use of oxyimino-cephalosporins in favour of beta-lactam inhibitor plus piperacillin (Gniadkowski, 2001; Bradford, 2001). In our survey ESBL production coincided with fluoroquinolone (28 % reduced susceptibility to CIPR) and aminoglycoside resistance (30 % GENT resistance in this survey), similar to results seen elsewhere. Thus, fewer therapeutic options remain for treatment, and that is one reason why the emergence of ESBLs should be controlled in the future. In this context, the importance of molecular diagnostics will increase, as they are a more reliable method than phenotypic testing.
|
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
, B., Randegger, C., Boras, A. & Hächler, H. (2001). Comparison of five different methods for detection of SHV extended-spectrum ß-lactamases. J Chemother 13, 24–33.[CrossRef][Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
