Molecular biology of extended-spectrum b-lactamase-producing Enterobacteriaceae responsible for digestive tract colonization
N. Moustaoui
a, A. Soukri
b, N. Elmdaghri
a,c, M. Boudouma
d, M. Benbachir
a,c,*
aMicrobiology laboratory, IbnRochd University Hospital, Casablanca, Morocco
bFaculte´ des Sciences AinChock, Casablanca, Morocco
cFaculte´ de Me´decine, Laboratoire de Microbiologie, Casablanca, Morocco
dFaculte´ des Sciences BenMsick, Casablanca, Morocco
Received 29 October 2003; accepted 8 January 2004 Available online 28 May 2004
KEYWORDS Extended spectrumb- lactamase; Digestive tract colonization
Summary Twenty-nine extended-spectrum b-lactamase (ESBL)-producing Enterobacteriaceae strains (14 Klebsiella pneumoniae, 10 Escherichia coli and fiveCitrobacter diversus) isolated from April to July 1996 from faecal carriers in a surgical intensive care unit at the university hospital of Casablanca (Morocco) were studied. Plasmid content and DNA macrorestric- tion polymorphism determined by pulsed-field gel electrophoresis (PFGE) were used to compare the strains. Restriction profiles of total genomic DNAs cleaved byXbaI and compared by PFGE revealed nine, four and two clones in K. pneumoniae,E. coliandC. diversus, respectively. Plasmid profile analysis of ESBL-producing strains of K. pneumoniae showed that only seven of 14 isolates had a plasmid; four different plasmid profiles were observed. Three different plasmid profiles were observed in E. coliand two in C. diversus.
Plasmids responsible for ESBL production could be transferred by conjugation to E. coli K12 J53-2 from all E. coli isolates and from four of seven K. pneumoniae. No plasmid transfer could be obtained from C. diversus strains. Restriction enzyme digests of plasmids from transconjugants (four transconjugants ofK. pneumoniaeand five transconjugants ofE. coli) showed different patterns. In the surgical intensive care unit where the survey was conducted, the dissemination of ESBLs was due to a mix of strain spread and strain diversity rather than to plasmid dissemination.
Q 2004 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved.
Introduction
Extended-spectrumb-lactamase-producing Entero- bacteriaceae (ESBLPE) are important agents of
Journal of Hospital Infection (2004)57, 202–208
www.elsevierhealth.com/journals/jhin
0195-6701/$ - see front matterQ2004 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.jhin.2004.01.007
*Corresponding author. Address: Faculte´ de Me´decine, Laboratoire de Microbiologie, 19 rue Tarik Bnou Zyad, BP 9154 Casablanca, Morocco. Tel.:þ212-22-26-90-57.
E-mail address:mbenbachir@wanadoo.net.ma
nosocomial infections and are responsible for out- breaks, which occur mainly in intensive care units.1 – 3The dissemination of ESBL is due to clonal spread,2,4,5 to plasmid dissemination among6,7 or between species,8 or to both mechanisms.3,9 – 11 Genes encoding for ESBL may also be found on transposons.1,11
Digestive tract colonization is a prerequisite for infection by ESBLPE.2,12In Casablanca a prospective study conducted in 1996 in an intensive care unit of the IbnRochd University Hospital13 revealed that digestive tract colonization with ESBLPE was fre- quent, occurred shortly after admission to the unit, and was transient. Bacteria responsible for this colonization belonged to five species; Klebsiella pneumoniae was the most frequent (42.4%) fol- lowed by Escherichia coli (30.3%) andCitrobacter diversus(15.2%).
To analyse the mode of spread of ESBLs (plasmid and/or clone spread) in the intensive care unit, we studied the ESBPLE isolates by macrorestriction polymorphism of DNA determined by pulsed-field gel electrophoresis (PFGE) and plasmid DNA analysis.
Materials and methods
Bacterial strainsThe most prevalent species of the 33 ESBLPE isolated in the colonization survey were studied:
K. pneumoniae ðN¼14Þ; E. coli ðN¼10Þ and C. diversusðN¼5Þ:Three of the 10E. coliisolates were isolated sequentially from the same patient.
Four patients were simultaneously colonized by two species: K. pneumoniaeþE. coli (two patients), K. pneumoniaeþKlebsiella oxytoca(one patient), K. pneumoniaeþC. diversus(one patient).
Genomic fingerprinting by PFGE
Agarose plugs containing chromosomal DNA were prepared by a method similar to that described by Yuan et al.5 The chromosomal DNA was digested overnight with 30 U ofXbaI (Boehringer, Mannheim, Germany). The PFGE was run in a CHEF-DRIII apparatus (Bio-Rad, Richmond, CA, USA) at 6 V/cm, 13.58C during 26 h forE. coliand 28 h for K. pneumoniae and C. diversus. The pulse times were 5 – 40 s. The banding patterns were analysed with the Biogene software (Vilbert-Lourmat, Marne-la-Valle´e, France) and were interpreted according to the criteria of Tenover et al.14 The PFGE experiments were performed by one of
the authors (NM) at the microbiology laboratory of the Ramon y Cajal Hospital of Madrid, Spain (Dr F. Baquero).
Plasmid isolation and resistance transfer Plasmids were extracted by the method of Kado and Liu,15 the alkaline lysis method as described by Sambrook et al.16 and by the High Pure Plasmid Isolation Kit (Boehringer Mannheim, Germany).
Transfer of plasmids by conjugation was performed as described previously.17 A rifampicin-resistant strain ofE. coli K12 J53-2, was used as recipient.
Transconjugants were selected on MacConkey agar containing rifampicin (250 mg/L) and amoxicillin (100 mg/L).
Restriction endonuclease analysis
Plasmid DNA from the transconjugants was digested withEcoRI (Sigma Bio-Sciences, St Louis, MO, USA) and Hind III (Roche Molecular Biochemicals, Mann- heim, Germany) according to the manufacturer’s recommendations.
Antibiotic susceptibility testing
The antibiotic susceptibilities were determined by disk diffusion on Mueller-Hinton agar. The following agents were tested: gentamicin, tobramycin, netil- micin, amikacin, trimethoprim-sulfamethoxazole, tetracycline, chloramphenicol and imipenem. The minimal inhibitory concentrations (MICs) of cefo- taxime and ceftazidime alone and combined with clavulanic acid (4 mg/L) were determined by agar dilution with an inoculum of 104cfu/spot.18E. coli ATCC 25922 and E. coli ATCC 35218 were used as controls.
Results
Macrorestriction polymorphism of DNA determined by PFGE distinguished nine clones among the 14 K. pneumoniaestrains (Figure 1). The 10E. coliand fiveC. diversus isolates divided into four and two clones, respectively (Table I). In each of the three species, some pulsotypes were detected only once whereas others were shared by two or three patients.
For each of the three species, patients carrying the same clone were analysed in an attempt to determine the links between them. Cases carrying a given clone were either imported from another ward or due to the presence in the ICU of a patient
already colonized. In some instances no link could be established between cases with the same clone.
During the study five new clones were introduced in the ICU by patients transferred from three different hospital wards. Three clones detected in patients on admission to the unit were already present in it.
Plasmid DNA analysis of ESBL producing isolates of K. pneumoniae showed that only seven of 14 isolates had a plasmid and four plasmid profiles were found. Three clones (C, F, I) had the same plasmid profile, but the restriction enzyme studies performed in clones C and F revealed different patterns. Three different plasmid profiles were observed in E. coli and two in C. diversus. Three clones of E. coli (J, K, and M) shared a common plasmid pattern, but the enzyme restriction profiles were different (Table I).
Plasmid transfer by conjugation to E. coli K12
J53-2 occurred with all the E. coli isolates. This transfer was observed for only four of seven K. pneumoniaeisolates. Plasmid restriction analysis from the transconjugants obtained with four K. pneumoniae and five E. coli strains showed three and five patterns, respectively (Figure 2). No plasmids could be transferred from theC. diversus isolates.
Different antibiotic resistance profiles were noted in isolates belonging to the same pulsotype (pulsotype L of E. coli and pulsotype A of K. pneumoniae). The two isolates of E. coli pulsotype L had a different plasmid restriction pattern.
Among the 29 ESBLPE isolates, 10 phenotypes of resistance were detected. Most isolates, except
one of C. diversus, were multiresistant (Table I).
The antibiotics most frequently involved in resist- ance were gentamicin, tobramycin, cotrimoxazole, tetracycline and chloramphenicol. The MICs of ceftazidime were frequently higher than those of cefotaxime suggesting a ceftazidimase activity of the ESBLs.19Clavulanic acid at a concentration of 4 mg/L substantially reduced the MICs of both ceftazidime and cefotaxime (Table I), indicating the presence of an ESBL.
Discussion
Molecular biology methods have improved the understanding of ESBL epidemiology.20 – 22 Specific identification of the enzyme subtype has revealed the considerable diversity of ESBLs,21their spread between bacterial species,20,21the coexistence and turnover of ESBLs in the same centre20 and their variable distribution according to geographic area.20,22 Precise identification of ESBL variant is very challenging21,22 as the reference method is nucleotide sequencing.21 For the investigation of strain relatedness, PFGE is usually used.2,5,7,12,24
The analysis of ESBLPE responsible for digestive tract colonization by DNA macrorestriction poly- morphism determined by PFGE and plasmid analysis Figure 1 PFGE ofXba1-digested total DNA from ESBL-
producingK. pneumoniaeresponsible for digestive tract colonization in 14 patients.
Figure 2 EcoRI restriction patterns of plasmids from transconjugants from ESBL-producing K. pneumoniae.
Lane 1, lamda DNA/EcoRIþHind III. Lanes 2 and 3 transconjugants from clone C (isolates 9 and 11). Lane 4 transconjugant from clone F (isolate 6). Lane 5 transcon- jugant from clone G (isolate 10).
N. Moustaouiet al.
204
Table I Characteristics of extended-spectrumb-lactamase-producingKlebsiella pneumoniae,Escherichia coliandCitrobacter diversusresponsible for digestive tract colonization Species Number of isolates PFGE type Plasmid profile Plasmid digestion pattern MIC (mg/L) phenotype of resistance Phenotype of resistance
CTX CTX-CLA CAZ CAZ-CLA
K. pneumoniaeðN¼14Þ 1 A NPD ND 4 ,0.06/4 32 ,0.06/4 GeToNeAkTsChTe
2 A NPD ND 2 ,0.06/4 32 ,0.06/4 GeToTsChTe
3 B NPD ND 2 ,0.06/4 4 ,0.06/4 GeToTsChTe
2 C d a 4 ,0.06/4 64 – 128 ,0.06/4 ToNeAkTe
1 D b ND 4 ,0.06/4 .128 ,0.25/4 ToNeAkTsChTe
1 E c ND 2 ,0.06/4 64 ,0.25/4 GeToTsChTe
1 F d b 2 ,0.06/4 32 ,0.06/4 GeToTsChTe
1 G a c 16 ,0.06/4 64 ,0.06/4 GeToTsTe
1 H NPD ND 4 ,0.06/4 128 0.25/4 GeToNeTsChTe
1 I d ND 2 ,0.06/4 16 ,0.06/4 GeToTsChTe
E. coliðN¼10Þ 5 J e d 8 ,0.06/4 16 ,0.06/4 GeToTs
2 K e e 2 ,0.06/4 64 0.12/4 GeToTsChTe
1 L f f 2 ,0.06/4 16 ,0.06/4 ToTs
1 L g i 2 ,0.06/4 16 ,0.06/4 Ts
1 M e k 4 ,0.06/4 32 ,0.06/4 ToNeAkTsChTe
C. diversusðN¼5Þ 4 N h ND 128 0.25/4 32 1/4 GeToNeAkTsChTe
1 O i ND 4 ,0.06/4 4 ,0.06/4
CTX,cefotaxime; CTX-CLA, cefotaximeþclavulanic acid; CAZ, ceftazidime; CAZ-CLA, ceftazidimeþclavulanic acid; Ge, gentamicin; To, tobramycin; Ne, netilmicin; Ak, amikacin; Ts, trimethoprim-sulfamethoxazole; Ch, chloramphenicol; Te, tetracycline; NPD, no plasmid detected; ND, not determined.
ularbiologyofESBLproducingEnterobacteriaceae205
allowed the study of the mode of spread of the ESBLs in an intensive care unit during the four- month study, in which seven of 29 isolates had no detectable plasmid and eight isolates failed to transfer their plasmids by conjugation. Although only a limited number of plasmids could be analyzed, the prevalent mechanism of ESBL dis- semination appeared to be the existence of multiple clones and the limited spread of a few clones rather than a plasmid spread as shown by the different plasmid digestion profiles observed for each species. Outbreaks caused by ESBLPE have been reported as due to the dissemination of a single strain,2,4to horizontal transfer of plasmid7or to concurrent dissemination of plasmids and strains.9,11
The frequencies, origin and the temporal distri- bution of the observed clones were studied. The high number of clones suggests that ESBL pro- duction has become endemic rather than epidemic.
For each species studied some clones were observed only once whereas a few were isolated from two or three patients. This finding may be explained by differences in isolation procedures among colonized patients and by the fact that the colonization was transient.13 Similar studies of ESBPLE responsible for digestive tract colonization revealed different results depending on the epide- miology in the unit. During an outbreak the same strain was responsible for most infections and for digestive tract colonization.12After an outbreak of K. oxytocainfections in an intensive care unit,11a screening programme of gastrointestinal coloniz- ation identified by PFGE a predominant clone and 15 unrelated isolates. Cukieret al.23reported that the ESBL-producingE. colistrain responsible for an outbreak of urinary tract infections in a geriatric ward had a unique ribotype, whereas after the outbreak ESBL-producingE. colistrains responsible for colonization belonged to four different ribotypes.
The number of clones detected in this study was higher inK. pneumoniae(nine clones in 14 isolates) than in E. coli (four clones in 10 isolates) and in C. diversus(two clones in five isolates). This finding confirms that the spread of ESBL production in K. pneumoniaeis due to multiple clones rather than to the spread of a single epidemic clone. A multi- centre study of K. pneumoniae with ESBLs con- ducted in European intensive care units5 reported 85 distinct strains in 220 isolates. Major variations in antibiotic resistance and plasmid patterns were noticed within strains, with some intra-strain vari- ations in b-lactamase subtype.5 This result is also supported by the important variety of plasmids and types of ESBL reported in K. pneumoniae.5,6,24,25
This circulation of multiple clones at one time was found by Bingen et al.,9 Nouvellon et al.10 and Essack et al.24 who reported on multiple strains prevalent simultaneously in single ICUs.
Colonization by ESBLPE could be due to clones already present in the unit, or to imported clones from other hospital wards. ESBL production may have appeared de novo in the unit, although their pre-existence cannot be ruled out as screening for ESBLPE was not performed in the patients who were present at the beginning of the survey. Five new clones were introduced in the ICU by patients transferred from three separate wards, demon- strating the spread of ESBLPE in our hospital. This finding confirms the observations that the problem of ESBLPE is usually of hospital-wide importance19 and that spread occurs through transfer of patients between wards.26These results support the recom- mendations for screening of patients from at risk areas on admission26and documenting the carriage of multi-resistant bacteria in case of transfer of patients between wards.23
Temporal analysis of cases due to the same clone was used to identify the link between patients. In some cases no link could be found between the cases raising the question of persistence of the agents within the unit. The sensitivity of the method used for the detection of the ESBPLE may have been insufficient and may explain the failure to detect a colonized patient who could have served as a reservoir. Although the environment is not considered a major reservoir for ESBLPE, one outbreak strain was isolated from equipment.3 Transmission from asymptomatic colonized staff is unlikely.26
Isolates belonging to the same clone may have a different plasmid digestion profile (e.g. clone L of E. coli), which can be explained by the possible existence of transposons or integrons or an instable plasmid that may change easily.27 Conversely, isolates of the same clone may have different antibiotic resistance phenotypes,10 which may be due to gain, or loss of plasmids or fragments of plasmids.5This finding confirms that the antibiotic resistance profile is not a valuable marker.
In this study, despite using three methods of extraction, plasmids could be detected in only seven of 14 K. pneumoniae strains, and of these, only four transferred their plasmids by conjugation.
In these isolates, the conjugative plasmid was responsible for ESBL production. The absence of plasmid transfer by conjugation has been reported,6,19,27although in other studies all isolates tested had detectable plasmids4,26or could transfer their plasmids.9,25 The absence of plasmids in N. Moustaouiet al.
206
ESBLPE may be explained by the fact that genes coding for ESBL may be on transposons.1,11
Transfer of plasmids responsible for ESBL pro- duction between two species present simul- taneously in the same patient is one mode of spread of ESBLs.8,11 In the four patients who harboured two ESBLPE simultaneously this transfer could not be demonstrated because, in each case, one of the species did not have a detectable plasmid or did not transfer its plasmid.
Macrorestriction polymorphism of DNA deter- mined by PFGE and plasmid analysis are useful tools for the study of ESBLPE responsible for digestive tract colonization in a hospital unit. The mix of strain diversity and strain dissemination, the existence of multiple different plasmids, the importation of ESBLPE from different wards of the hospital and antibiotic multi-resistance all confirm the complexity of the epidemiology of ESBLPE.5,9,24 Consideration of these factors is important in making recommendations for preven- tion of spread of ESBLPE. Current recommendations include screening on admission to the Unit,26 improvement of infection control measures22 and rational use of antibiotics.22
Acknowledgements
The authors thank Dr F. Baquero (Hospital Ramon y Cajal, Madrid, Spain) for supporting this study and Dr J. Wilimas (St Jude Children’s Research Hospital, Memphis, TN, USA) for English revision of the manuscript. This work was partially supported by a grant from the Programme The´matique d’Appui a` la Recherche Scientifique (PROTARS P1T2/04).
References
1. DeChamps C, Sirot D, Chanal C, Poupart MC, Dumas MP, Sirot J. Concomitant dissemination of three extended-spectrum b-lactamases among different Enterobacteriaceae isolated in a French hospital. J Antimicrob Chemother 1991;7:
441—457.
2. Pena C, Pujol M, Ardanuy C, et al. Epidemiology and successful control of a large outbreak due toK. pneumoniae producing extended-spectrum b-lactamases. Antimicrob Agents Chemother1998;42:53—58.
3. Prodinger WM, Fille M, Bauernfeind A, et al. Molecular epidemiology of K. pneumoniaeproducing SHV-5 b-lacta- mase: parallel outbreaks due to multiple plasmid transfer.
J Clin Microbiol1996;34:564—568.
4. Decre´ D, Gachot B, Lucet JC, Arlet G, Bergone-Be´re´zin E, Re´gnier B. Clinical and bacteriological epidemiology of extended-spectrum beta-lactamase-producing strains of K.
pneumoniaein medical intensive care unit.Clin Infect Dis 1998;27:834—844.
5. Yuan M, Aucken H, Hall LM, Pitt TL, Livermore DM.
Epidemiological typing ofK. pneumoniae with extended- spectrumb-lactamases from European intensive care units.
J Antimicrob Chemother1998;41:527—539.
6. Legakis NJ, Tzouvelekis LS, Hadzoudis G,et al. K. pneumo- niaeinfections in Greek hospitals. Dissemination of plasmids encoding an SHV-5 typeb-lactamase.J Hosp Infect1995;31:
177—187.
7. Liu PYF, Tung JC, Ke SC, Chen SL. Molecular epidemiology of extended-spectrumb-lactamase-producing K. pneumoniae isolates in a district hospital in Taiwan. J Clin Microbiol 1998;36:2759—2762.
8. D’Agata E, Venkataraman L, DeGirolami P, Wiegel L, Samore M, Tenover F. The molecular and clinical epidemiology of Enterobacteriaceae-producing extended-spectrum b-lacta- mase in a tertiary care hospital.J Infect1998;36:279—285.
9. Bingen EH, Desjardins P, Arlet G,et al.Molecular epidemiol- ogy of plasmid spread among extended-broad-spectrumb- lactamase-producingK. pneumoniaeisolates in a pediatric hospital.J Clin Microbiol1993;31:179—184.
10. Nouvellon M, Pons JL, Syros D, Combe ML, Lemond HF. Clonal outbreaks of extended-spectrum b-lactamase producing strains of K. pneumoniae demonstrated by antibiotic susceptibility testing, b-lactamase typing and multilocus enzyme electrophoresis. J Clin Microbiol 1994;32:
2625—2627.
11. Venezia RA, Scarano FJ, Preston KE, et al. Molecular epidemiology of an SHV-5 extended-spectrumb-lactamase in Enterobacteriaceae isolated from infants in a neonatal intensive care unit.Clin Infect Dis1995;21:915—923.
12. Lucet JC, Chevret S, Decre´ DI. Outbreak of multiply resistant Enterobactericeae in an intensive care unit epidemiology and risk factors for acquisition. Clin Infect Dis 1996;22:
430—436.
13. Moustaoui N, Bensghir R, Mjahed K,et al.Digestive tract colonization with extended-spectrum b-lactamase produ- cing Enterobacteriaceae in a surgical intensive care unit in Casablanca.J Hosp Infect2000;46:238—240.
14. Tenover F, Arbeit RD, Goering RV, et al. Interpreting chromosomal DNA restriction patterns produced by pulsed- field gel electrophoresis: criteria for bacterial strain typing.
J Clin Microbiol1995;33:2233—2239.
15. Kado CL, Liu S. Rapid procedure for detection and isolation of large and small plasmids. J Bacteriol 1981;145:
1367—1373.
16. Sambrook J, Fritch EF, Mandates T. Molecular cloning: a laboratory manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press; 1989.
17. Tzouvelekis LS, Tzelepi E, Mentis AF, Takris A. Identification of a novel plasmid mediatedb-lactamase with chromosomal cephalosporinase characteristics from K. pneumoniae.
J Antimicrob Chemother1993;31:645—654.
18. National Committee for Clinical Laboratory Standards.
Performance standards for dilution antimicrobial suscepti- bility tests for bacteria that grow aerobically. Approved Standards M7-A4. Wayne, PA: National Committee for Clinical Laboratory Standards; 1997.
19. Gniadkowsky A, Palucha A, Grzesiowsky P, Hryniewicz W.
Outbreak of ceftazidime-resistant K. pneumoniae in a pediatric hospital in Warsaw, Poland: clonal spread of the TEM-47 extended spectrum b-lactamase (ESBL)-producing strain and transfer of a plasmid carrying the SHV5-like ESBL- encoding gene. Antimicrob Agents Chemother 1998;42:
3079—3085.
20. Gniadkowsky M. Evolution and epidemiology of extended- spectrumb-lactamases (ESBLs) and ESBL-producing micro- organisms.Clin Microbiol Infect2001;7:597—608.
21. Bradford PA. Extended-spectrumb-lactamases in the 21st century: characterization, epidemiology and detection of this important resistance threat.Clin Microbiol Rev2001;
14:933—951.
22. Sturenburg E, Mack D. Extended-spectrum b-lactamases:
implications for the clinical microbiology laboratory, therapy, and infection control.J Infect2003;47:273—295.
23. Cukier L, Lutzler P, Bizen A, Avril JL. Investigation d’une e
´pide´mie a` E. coli a` be´ta-lactamase a` spectre e´tendu dans un service de ge´riatrie.Pathol Biol1999;47:440—444.
24. Essack SY, Hall LMC, Pillay DG, McFayden ML, Livermore DM.
Complexity and diversity of K. pneumoniae strains with extended-spectrumb-lactamases isolated in 1994 and 1996 at a teaching hospital in Durban, South Africa.Antimicrob Agents Chemother2001;45:88—95.
25. Shannon K, Stapleton P, Xiang X,et al.Extended-spectrum b-lactamase-producingK. pneumoniaestrains causing noso- comial outbreaks of infections in the United Kingdom.J Clin Microbiol1998;36:3105—3110.
26. Shannon K, Fung K, Stapleton P, Anthony R, Power E, French G. A hospital outbreak of extended-spectrumb-lactamase producingK. pneumoniaeinvestigated by RAPD typing and analysis of the genetics and mechanisms of resistance.J Hosp Infect1998;39:291—300.
27. Siu LK, Lu PL, Hsueh PR,et al.Bacteraemia due to extended- spectrumb-lactamase-producingE. coliandK. pneumoniae in a pediatric oncology ward: clinical features and identifi- cation of different plasmids carrying both SHV-5 and TEM-1 genes.J Clin Microbiol1999;37:4020—4027.
N. Moustaouiet al.
208
