Plasmids-Mediated Antibiotic Resistance in E. coli

Plasmids-Mediated Antibiotic Resistance in E. coli (Received: 08.12.1998) Nariman A. H. Aly*; Amany Tharwat** and Wesam F. El-Baz** *Microbial Genetics Dept., Genetic Engineering and Biotechnology Div., National Research Centre, Cairo, Egypt. **Departments of Microbiology & Immunology and Internal Medicine, Faculty of Medicine, Ain Shams Univ., Cairo, Egypt. ABSTRACT In this study, E. coli strains were isolated from thirty hospitalized patients receiving antimicrobials, thirty not on antimicrobials and thirty healthy persons were used as controls. Antimicrobial susceptibility test was done to thirteen of the commonly used antimicrobials in hospitals. There was a significant increase in the percentage of E. coli resistant to most of the antimicrobials on comparing the patients receiving antimicrobials and control group. Among the 14 E. coli strains tested, plasmid profile analysis of resistant strains was done and plasmid numbers ranged from 0 to 8 indigenous plasmids with variable sizes. Curing and transformation experiments revealed the existence of CTX r, CRO r, CFB r and OFX r genes on plasmids. Among these plasmids, the OFX r plasmids of strains 1 and 14 were found to be conjugative. Key words: Cephalosporins, Antibiotic resistance genes, Conjugative plasmids. INTRODUCTION B acterial resistance to antimicrobial agents is a threat to public health throughout the world. During the last years, outbreaks of diseases due to multiresistant strains of enteric bacteria pathogens have occurred. The resistance patterns exhibited by these organisms have included those antibiotics used most heavily at the time of the outbreak, as well as older agents. The consequences of resistance affect not only the ability to treat the infection, but also the cost and duration of treatment (Ismaeel, 1993). Cephalosporin C is an important antibiotic which attacks the cell wall by inhibition of peptidoglycan synthesis. It possesses a β-lactam ring structure similar to the penicillins. Both side chains of cephalosporin C have been changed by chemical treatment to yield successive generations of cephalosporins with vastly improved properties, including greater antimicrobial activity, broader range of organisms inhibited, improved β-lactamase resistance and different pharmacological properties. Extended spectrum β-lactamases (ESBLs) exhibit an enhanced ability to hydrolyze the expanded-spectrum β-lactams. The rapid spread of ESBLs caused significant threats to the therapy for infections and usage of the expandedspectrum β-lactams. Thus, the challenge to clinicians and microbiologists to recognize

susceptibility patterns indicative of the presence of specific β-lactamases, such as the extendedspectrum β-lactamases, will become even more important, as this genus acquires additional antimicrobial resistance mechanisms (Pitout et. al., 1998). Resistance to expanded-spectrum cephalosporins usually emerges in Enterobacter spp. due to a mutation in an ampd gene, that normally prevents high-level expression of this organism's chromosomal β-lactamase (Sanders, 1992). In addition, resistance to cephalosporins has arisen in E. coli via the acquisition of plasmids containing the chromosomally encoded ampc found in Enterobacter spp., Pseudomonas aeruginosa, and Citrobacter spp. Human intestinal flora play an important role in the development of the immune system, resistance to colonization by exogenous pathogenic microorganisms. Also it constitutes a reservoir of potentially pathogenic bacteria in close contact with the host (Tancrede, 1992). There have been a few data on the natural frequency of antibiotic resistance genes in the normal non-pathogenic flora of hospitalized individuals. Previous studies revealed a high prevalence of resistant bacteria in the gut flora of hospitalized individuals, whether or not they were receiving antibiotics (Ismaeel, 1993). The aim of this work is to study the effect of hospital admission with or without antibiotic therapy on emergence of resistant strains of E. coli, which represent a part of normal faecal flora, and to investigate whether it is plasmid-mediated or not. MATERIALS AND METHODS Patients and E. coli strains E. coli strains were isolated from ninety persons classified into three groups, 30 persons each. The first group was under therapy with different antimicrobial agents, while the second group was not under antimicrobial therapy. The third group included healthy persons as a control group. All specimens were cultured on MacConky agar plates. The isolated colonies were identified by the API 20E system (Bio Merieux, mary, I E Toil, France). E. coli DH5α (Hanahan, 1983) was used as a recipient strain in transformation and conjugation experiments. Antimicrobial susceptibility The Kirby-Bauer disc diffusion method for antimicrobial susceptibility test was used (NCCLS, 1992). Four different classes of antimicrobials were used (Oxoid antimicrobial susceptibility test discs): 1- Cephalosporins: cephradine (CPH), cepfalexin (CPF), cefotaxime sodium (CTX), ceftriaxone disodium (CRO), cefo-perazone (CFP), or cefpodoxim (CFX). 2- Quinolones: ofloxacin (OFX) or norfloxacin (NOR). 3- Aminoglycosides: amikacin (AMK) or gentamicin (GEN). 4- Other Beta lactam drugs: aztreonam (AZT), imipenen (IMP), sulbactam (SUL) or ampicillin (AMP). The isolates that showed multiresistance against more than two groups of the tested antimicrobials were subjected to plasmid analysis. Plasmid manipulations The plasmids content and their relative sizes were determined using a modification of the inwell lysis procedure of Eckhardt (Plaziniski et al., 1985) using 0.65% agarose gel electrophoresis.

Plasmids were isolated using the method of Rodriguez and Tait (1983). Plasmid curing The tested E. coli isolates were cured from their own plasmids by growing them in elevated temperature (Toama et al., 1983). Then, an appropriate dilution was spread on LB plates and incubated at 37 0 C. Five random single colonies were picked up and tested for their sensitivity against CFP, CTX, CRO or OFX. Bacterial transformation E. coli DH5α was used as a recipient of plasmids isolated from the E. coli isolates following the method of Mandel and Higa (1970). Transformants were selected by their ability to grow with the appropriate antibiotics at the following concentrations 1.6, 0.12 and 40 and 3 µg /ml for OFX, CTX, CRO and CFP, respectively. Bacterial conjugation In conjugation, recipient strain has to be resistant to a factor while the donor strain is sensitive to the same factor. Only two E. coli isolates were applicable for this test, i.e. No. 1 and No. 14. Conjugation was performed between E. coli DH5α and each of the two E. coli isolates, using the plate method as described by Abdel-Salam and Klingmüller (1987). The conjugation mixtures were plated on LB plates with chloramphenicol (Cm) 35 µg/ml, to counterselect the donor strain, and supplemented with CRO or OFX. RESULTS AND DISCUSSION Resistance profiles The resistance percentages of the two patient groups as well as the healthy persons as controls were summarized in Table (1). Patients of Group I showed highly significant differences, with respect to the resistance to CPF, CRO, CFP, GEN, AZT and SUL/AMP, when compared with the control. The same group showed significant differences in resistance to CPH, CTX, CFX and NOR. Moreover, the highest resistance percentages of Group I were obtained when CPF, SUL/AMP and CPH were used with values of 95.8, 75 and 70.8%, respectively. However, there were no significant differences as regards the resistance to IMP and AMK. Patients of Group II showed significant differences, with respect to the resistance to CFP, AZT, GEN and SUL/AMP, when compared with the control. The other antibiotics used showed no significant differences. Further, patients of Group I showed highly significant differences when compared with Group II with respect to the resistance to CPF and SUL/AMP, while there was significant difference in resistance to OFX only. However, no significant difference was observed between the two Groups I and II in resistance to the other antibiotics examined (Table 1). Moreover, data revealed that all E. coli strains isolated from Group I and Group II were susceptible to IMP. Similar results were obtained among many different species such as Acinetobacter strains (Abul-Ella et al., 1997) and Enterobacter aerogenes strains (Pitout et al., 1998). Plasmid patterns Fourteen E. coli isolates were selected to analyze their indigenous plasmids as well as to study the plasmids role regarding the antibiotic resistance. As shown in Table (2), four resistance patterns, P 1 to P 4 were categorized, of which P 1 included seven strains which were multiresistant to the five antibiotics tested, and P 4 had one strain, i.e. No. 9, that exhibited multisusceptiblity to the

above five antibiotics. Furthermore, the intimately resistance pattern P 2 is characterized by the resistance to four antibiotics, CFP, CTX, CRO and OFX, and susceptiblity to CM. This patteren comprised three strains, while the intimately susceptible pattern P 3 comprised four strains; two of them, i.e. DH5α and 8, were resistant to two different antibiotics and susceptible to the other three. Other strains, i.e. 7 and 11, were susceptible to four antibiotics, CM, CFP, CTX and OFX, while resistant to CRO. These results are in agreement with previous findings on other cephalosporins drugs. For example, Rasheed et al., (1997) found nine isolates of E. coli susceptible to extendedspectrum cephalosporins and four strains gave a positive disc potentiation when CTX, CRO, AZT and/or ceftazidime discs were used. Moreover, Pitout et. al., (1998) stated that CTX, ceftazidime, AZT, and cefoxitin exhibited two resistance phenotypes; Ec1 and Ec2 in E. coli. Plasmid profile analysis in Fig. (1) revealed a variety of different plasmids with different sizes; (low, medium and high molecular weights). Each of the tested isolates showed a unique plasmid pattern, with the number of plasmids ranging from 0 in strain No. 2 to 8 in the two strains 8 and 9. The association between the plasmid patterns and the resistance profile patterns was reported by Pitout et al. (1998), who grouped the isolate strains of E. aerogenes to different plasmid profiles such as B consisting of five plasmids ranging from 50 to 10 kb, and F consisting of three plasmids ranging from 60 to 10 kb, and both were susceptible to gentamicin, while profile H did not contain plasmids and was resistant to ceftazidime, AZT, trimethoprim-sulfamethoxazole and GEN. Moreover, it is interesting to notice that the majority of the isolated strains that carried a low number of plasmids such as 2, 3, 5, 6, 13, 14 and 15 were definitely resistant to CFP, CTX, CRO and OFX. On the other hand, the strains carried the high numher of plasmids, such as 9, was susceptible to the above mentioned antibitotics (Fig. 1 and Table 2). Table (1): Resistance percentage of the two patient groups (Group I and II) and the healthy persons (Group III). Antibiotic classes Antibiotics tested Control Patients on Patients not on (Group III) antibiotics (Group I) antibiotics (Group II) CPF 52.2 95.8** 52.6 CPH 34.8 70.8* 58.1 CTX 4.3 16.7* 12.2 I CRO 4.3 37.5** 16.1 CFP 4.3 50.0** 22.6* CFX 8.4 41.7* 32.6 OFX 8.7 20.8 8.7 II NOR 8.7 41.7* 16.1 AMK 4.2 8.2 16.1 III GEN 4.3 37.5** 25.8* AZT 0.0 41.7** 22.6* IV IMP 0.0 3.2 3.2 SUL/AMP 21.0 75.0** 54.8*

II Total average 9.5 36.5 23.2 ** Highly significant P value when less than 0.005. * Significant at 0.05. Fig. (1): Photocopy (a) and diagram (b) of plasmid patterns of some E. coli isolates. Lanes 1 to 9 present E. coli isolates 1 to 9. Table (2): Antibiotics resistance patterns of 14 studied E. coli strains and the number of plasmids detected by electrophoresis. Antibitotics tested No. of E. coli Resistance plasmids strains Patterns CM CFP CTX CRO OFX Detected 2 + + + + + 0 3 + + + + + 2 4 + + + + + 6 5 P 1 (Multiresistance) + + + + + 2 6 + + + + + 2 10 + + + + + 3 13 + + + + + 1 1 - + + + + 4 P 2 14 - + + + + 2 (Intimately resistance) 15 - + + + + 1

II DH 5α + + - - - 0 8 P3 - + - + - 8 7 (Intimately susceptible) - - - + - 6 11 - - - + - 3 9 P 4 (Multisusceptible) - - - - - 8 Location of antibiotic resistance genes To locate the genes encoding the resistance for the tested antibiotics, plasmid curing experiments were performed. Five random single colonies were picked out, after growing their parents at elevated temperatures, and testing for their sensitivity against CFP, CTX, CRO or OFX. Different antibiotic resistance patterns were detected (Table 3). No antibiotic sensitive progenies were found after treatment of strains 4 and 6, indicating that the resistance genes are probably on the chromosome or on very high molecular weight plasmids, which are very difficult to be eliminated as found by Hynes et al. (1989). Results indicated that plasmid encoding CFP r probably exists in strains 2 and 10, while plasmid encoding CTX r is present in strains 2, 3 and 13; plasmid encoding CRO r is in strains 1, 3, 10 and 13; plasmid encoding OFX r is in strains 2, 5, 10 and 13. The presence of progeny sensitive to more than one antibiotic, i.e. CFP and OFX in strains 2 and 15, and CFP and CTX in strains 14 and 15, may indicate the existence of two separate R-plasmids or that one plasmid is carrying both antibiotic resistance gene(s). The triple sensitivity to CFP, OFX and CTX found among the progeny of strains 14 and 15 is also probably due to different R-plasmids or a multiple resistance plasmid in these strains. The inability to find indigenous plasmids in strain 2 by gel electrophoresis (Fig. 1), although curing experiments revealed the probability of existence of at least three plasmids carrying CFP r, OFX r or CTX r genes, could be explained by the low copy number of these plasmids, which is very difficult to be recognized by gel electrophoresis. Pitout et al. (1998) reported that plasmids encoding extended-spectrum β-lactam antibiotics may also encode resistance to other classes of antibiotics, such as aminoglycosides and trimethoprim-sulfamethoxazole. Moreover, Sirot (1995) reported that such plasmids are limiting the options of physicans treating infections caused by organisms producing enzymes such as SHVs. Therefore, factors leading to the selection and spread of strains producing ESBLs need to be identified and, where possible, eliminated (Sanders and Sanders, 1997). Makawi and Youssef (1998) studied the antibiotic resistance patterns of 89 E. coli isolates. Seventy one isolates showed a multiple resistance patterns and 13 isolates were susceptible to all antibiotics tested. The number of plasmids in each of these isolates ranged from one to six, of different sizes. E. coli transformation All plasmids were isolated from different E. coli strains and used to transform E. coli DH5α. Transformants were selected by their growth on LB plates supplemented with CRO, OFX and CRO / OFX. Results obtained were presented in (Table 4 and Fig. 2). All of the obtained CRO r and OFX r transformants were replica plated on the other antibiotics. However, not all the transformants were able to grow on such antibiotics. Moreover, results in Table (4) indicated the existence of three types of antibiotic resistant plasmids. The CRO r

gene is located in an indigenous plasmid in strains 1, 2, 3, 4, 5 and 6. The OFX r gene is located in a plasmid in all tested strains, except strains 3 and 13. These results also indicated that only one plasmid could be transferred to the recipient cell as reported by Maniatis et al. (1982). The double resistant CRO and OFX transformants (Table 4) in strains 1, 4, 5, 6, 13, 14 and 15 indicated that both resistance genes were located in the same plasmid in such strains. The transformation results confirmed the existence of two different indigenous plasmids in strain 2, i.e. CRO r and OFX r. This finding is supported by Yang et al. (1998) who reported that multidrug resistance plasmids may carry genes encoding resistance to other antibiotics such as aminoglycosides. In addition, Pitout et al. (1998) reported that it seems surprising that an organism with an inducible cephalo-sporinase would acquire an extended-spectrum β-lactamase, resistance to other agents such as the aminoglycosides and trimethoprimsulfamethoxazole, which is encoded on the same plasmid as the extended-spectrum β-lactamase. Table (3): Antibiotic sensitivity of five random colonies obtained after elevated temperature growth of some tested E. coli strains. Number of antibiotic sensitive colonies Strains Tested CFP s CTX s CRO s OFX s CFP s /OFX s CFP s / CTX s CFP s / CTX s /CRO s 1 - - 2 - - - - 2 1 1-1 1 - - 3-1 1 - - - - 4 - - - - - - - 5 - - - 1 - - - 6 - - - - - - - 10 1-1 2 - - - 13-2 1 1 - - - 14 - - - - - 2 1 15 - - - - 1 1 1 Table (4): E. coli DH 5a tansformants. Donor Transformants* Strains CRO r OFX r CRO r /OFX r 1 + + + 2 + + 3 + 4 + + + 5 + + + 6 + + + 10 +

13 + 14 + + 15 + + * Obtained by selection on LB supplemented with CRO OFX and CRO/OFX. Fig. (2): Photocopy (I) and diagram (II) of plasmid patterns of some E. coli strains Lane A, E. coli DH5a (recipient); Lane C, strain 6; Lane D, strain 4; Lane 1, strain 1; Lane B, transformant 6-E; Lane E and F, transformants 4-A and 4-B, respectively; Lane G and H transformants 1-F and 1-A, respectively. Chr.=Chromosomal DNA band.. Conjugation Conjugation tests were carried out between E. coli DH5α as the recipient and each of the strains 1 or 14 as doners. Results indicated that the obtained transconjugants were E. coli DH5α resistant to OFX, while resistance to CRO could not be obtained in both cases. Moreover, a new plasmid in E. coli DH5α transconjugants was confirmed by gel electrophoresis in both strains 1 and 14 (Fig. 3).

These results indicated that OFX r gene could be transferred via conjugation, which suggests that it is in a conjugative plasmid, while the CRO r gene could not be transferred to strain DH5α through conjugation, which suggests that it may be in a non-conjugative plasmid. Carmen et al. (1998) transconjugated a K. pneumoniae strain that was highly resistant to expanded-spectrum cephalosporins and monobactams to E. coli J53-2 strain. Moreover, Pitout et al. (1998) were able to transfer, by conjugation, the different β-lactamases resembling genes SHV-3, SHV-4, and SHV-5 as well as genes of resistance to gentamicin and trimethoprim-sulfamethoxa-zole into E. coli C600N, and this was accompanied by a plasmid of approximately 50 kb. However, Nordmann et al. (1993) was not be able to transfer IMP resistance into E. coli JM109 and the plasmid DNA was not detected in E. cloacae. Finally, it is possible that the continuous use of these drugs was responsible for selecting plasmid-mediated resistance to extendedspectrum cephalosporins in E. coli. Furthermore, it is important that such strains should be recognized, because they can be responsible for institutional spread of resistance genes. In conclusion, it has been shown that there is an increase in prevalence of resistant bacteria in the normal non-pathogenic flora of hospitalized individuals, especially those taking antibioics. Such resistant bacteria would represent a constant pool of resistance genes potentially transferable, directly or indirectly to human pathogens. Restricting the use of antibiotics in hospitals may play a role in decreasing the emergence of newly resistant bacterial strains. These resistant strains may be responsible for outbreaks of nasocomial infections. Fig. (3): Photograph (A) and diagram (B) represent a new plasmid of E. coli DH5a transconjugation. Lanes 1, 2, 5 & 6 transconjugates. Lanes, 3 & 7 the donor strains 14 & 1, respectively. Lanes, 4 & 8, the recipient E. coli DH5a.

REFERENCES Abdel-Salam, M.S. and Klingmüller, W. (1987). Transposon mutagenesis in Azospirillum lipoferum: isolation of indole acetic acid mutants. Mol. Gen. Genet. 210: 165-170. Abul-Ella, M.A., Zaghloul, W., Rizk, M. S. and Elmoadawy, M. (1997). Isolation and characterization of small plasmids encoding multiple drug resistance from Acinetobacter hospital strains. Egypt. J. Med. Microb., 6 (12): 205-212. Carmen, A., Josefina, L., Marya, A. D., Vicenta, J. B. and Martynez, L. (1998). Outer membrane profiles of clonally related Klebsiella pneumoniae isolates from clinical samples and activities of cephalosporins and carbapenems. Antimicrobial Agents and Chemotherapy, 42 (7): 1636-1640. Hanahan, D. (1983). Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol., 166: 557. Hynes, M.F., Quandt, J., O Connell, M. P. and Pühler, A. (1989). Direct selection for curing and deletion of Rhizobium plasmids using transposons carrying the Bacillus subtilis sac B gene. Gene, 78: 111-120. Ismaeel, N.A. (1993). Resistance of bacteria from human faecal flora to antimicrobial agents. J. Trop. Med. Hyg., 96 (1): 51-55. Makawi, H.I. and Youssef, M. T. (1998). The antibiotic susceptibility testing and plasmid profiles of Escherichia coli isolated from diarrhoeal patients. J. Trop. Pediatr. 44 (3): 128-132. Mandel, M. and Higa, A. (1970). Calciumdependent bacteriophage DNA infection. J. Mol. Biol., 53: 159. Maniatis, T., Fritsch, E. F. and Sambrook, J. (1982). Molecular cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. National committee for clinical laboratory standards (NCCLS) (1992). Performance standards for antimicrobial susceptibility testing. Fourth Information suppl. M100-S4 Villanova. Nordmann, P., Mariotte, S., Naas, T., Labia, R. and Nicolas, M. H. (1993). Biochemical properties of a carbapenem-hydrolyzing betalactamase from Enterobacter cloacae and cloning of the gene into Escherichia coli. Antimicrobial Agents and Chemotherapy, 37 (5): 939-946. Pitout, J.D.D., Thomson, K.S., Hanson, N.D., Ehrhardt, A.F., Moland, E.S. and Sanders, C.C. (1998). Beta-Lactamases responsible for resistance to expanded-spectrum cephalosporins in Klebsiella pneumoniae, Escherichia coli and Proteus mirabilis isolates recovered in South Africa. Antimicrobial Agents and Chemotherapy, 42 (6): 1350-1354. Plaziniski, J., Cen, Y.H. and Rolfe, B.G. (1985). Identification of plasmids by gel electrophoresis. Appl. Environ. Microbiol., 48 (4): 1001-1003. Rasheed, J.K., Jay, C., Metchock, B., and Tenover, F.C. (1997). Evolution of extendedspectrum β-lactam resistance SHV-8 in a strain of E. coli during multiple episodes of bacteremia. Antimicrobial Agents and Chemotherapy. 41 (3): 647-653. Rodriguez, R.L. and Tait, R. C. (1983). Rapid isolation of plasmid DNA. In: Recombinant DNA Techniques, An Introduction: The Benjamin/Cummings Publishing Inc.

IO Sanders, C.C. (1992). β-lactamases of gram negative bacteria: new challenges for new drugs. Clin. Infect. Dis., 14:1089-1099. Sanders, W.E. and Sanders, C.C. (1997). Enterobacter spp.: pathogens poised to flourish at the turn of the century. Clin. Microbiol. Rev., 10:220-241. Sirot, D. (1995). Extended-spectrum β- lactamases. J. Antimicrob. Chemother. 36 (Suppl. A): 19-34. 31. C. F. Weischer, M.; Schumacher, H. and Kolmos, H. J. 1994: Resistance characteristics of blood culture isolates of Enterobacter cloacae with special reference to beta-lactamases and relation to preceeding antimicrobial therapy. APMIS, 102: 356-366. Tancrede, C. (1992). Role of human microflora in health and disease. Eur. J. Clin. Microbiol. Inf. Dis., 11 (11): 1012-1015. Toama, M.A., El-Kersh, T. A. and Ahmady, A. M. (1983). Antibiotic resistant pattern and resistance curing of E. coli locally isolated. Bull. Fac. Pharm., Cairo Univ. 22 (1): 129. Yang, Y., Bhachech, N., Bradford, P. A., Jett, B. D., Sahm, D.F. and Bush, K. (1998). Ceftazidime-resistant Klebsiella pneumoniae and E. coli isolates producing TEM-10 and TEM-43 β-lactamases from St. Louis. Antimicrobial Agents and Chemotherapy, 42 (12): 1671-1676.?????????????????????????????T????????????O?p????????G?????????????????????????????????????????????????????????????????????????????????F??????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????F????????????????»??????»????????»???????????????????????????????????????????????????F?????»????????????????????????????????????F??????????????????????????????????????????????????????????????????????????????????????????????????????????? I????????????????????????????????????????????F????????????????????????F??????????????????????????????????????????????????????????????????????????????????????????????????????????????»??????»???»????»??????????????????????????????????????????F????????????????????????????????????????????????????????????? CTX, CRO, CFB, OFX??????????????????????????????????? I? I???????????? OFX