JAC Advance Access originally published online on September 3, 2007
Journal of Antimicrobial Chemotherapy 2007 60(5):1142-1145; doi:10.1093/jac/dkm327
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Transfer of an ampicillin resistance gene between two Escherichia coli strains in the bowel microbiota of an infant treated with antibiotics
1 Department of Clinical Bacteriology, Göteborg University, Sweden 2 Department of Cellular and Molecular Medicine, University of Bristol, UK 3 Department of Paediatrics, Göteborg University, Sweden
* Correspondence address. Department of Clinical Bacteriology, Guldhedsgatan 10A, SE-413 46 Göteborg, Sweden. Tel: +46-31-3424729; Fax: +46-31-3424975; E-mail: nahid.karami{at}microbio.gu.se
Received 5 April 2007; returned 18 July 2007; accepted 31 July 2007
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResults and discussionFundingTransparency declarationsReferences |
|---|
Objectives: To investigate the presumed acquisition of ampicillin resistance by an Escherichia coli strain residing in the gut of an infant.
Methods: E. coli strains were quantified in faecal samples obtained at regular intervals from an infant followed from birth to 12 months of age and their resistance profiles were determined. ß-Lactamases were identified by isoelectric focusing and genes by PCR followed by DNA sequencing. Plasmids were characterized by restriction fragment analysis and Southern-blot hybridization, and tested for conjugative transfer.
Results: The infant carried two E. coli strains, termed 29A and 29B, simultaneously in the microbiota during the first month of life. All isolates of 29A were resistant to ampicillin, whereas strain 29B, which was initially ampicillin susceptible, acquired resistance following treatment of the infant with ampicillin/amoxicillin because of urinary tract infection. Acquisition of resistance by strain 29B was associated with acquisition of a blaTEM-1b-encoding plasmid, pNK29, which was also present in strain 29A. Transfer of plasmid pNK29 could be replicated by conjugation from strain 29A to strain 29B in vitro. Strain 29A also adapted to ampicillin treatment by mutation of the blaTEM-1b promoter gene to yield a higher level of resistance.
Conclusions: This is an unequivocal demonstration of gene transfer between two strains co-residing in the human gut, as the donor, recipient and transconjugant strains were isolated. The results suggest the dynamic adaptation by commensal bacteria in response to antibiotic treatment may occur readily.
Keywords: blaTEM-1b gene , plasmid , pNK29
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionFundingTransparency declarationsReferences |
|---|
Antimicrobial resistance genes are often associated with conjugative plasmids or transposons, which encode the proteins necessary to initiate and complete their transfer to new hosts.
The human large intestine has been proposed as a suitable environment for gene exchange. Recently, transfer of a plasmid-mediated ACC-1 ß-lactamase from a Klebsiella pneumoniae strain to an Escherichia coli strain during antibiotic treatment of an infant was reported.1 Facultative bacteria, such as E. coli, reach much higher population levels in the gut flora of infants than in adults, as many strict anaerobes do not establish until individuals are several years old.2 This fact may enhance the possibility of gene exchange between facultative bacteria in the infantile microbiota.
Here, we demonstrate that a plasmid carrying a ß-lactamase gene appears to have been transferred from an ampicillin-resistant E. coli strain to an initially susceptible strain during their co-residence in the infantile gut.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionFundingTransparency declarationsReferences |
|---|
Study protocol
The infant studied here participated in the ALLERGYFLORA study, which examines the relationship between intestinal colonization pattern and allergy development. His medical history, including antibiotic treatment, was recorded by the parents and the study nurse. Data were also obtained from hospital records. Informed consent was obtained from the parents and the study was approved by the Medical Ethics Committee of Göteborg University.
A rectal swab was obtained at 2–3 days of age and stool samples were collected at 1, 2 and 4 weeks of age and at 2, 6 and 12 months of age and cultivated quantitatively for all major groups of facultative and anaerobic bacteria. E. coli were quantified on Drigalski agar and various colony types were enumerated separately and speciated using API 20E identification strips (bioMérieux, Marcy-l'Étoile, France). All E. coli isolates were identified to the strain level using random amplified polymorphic DNA (RAPD) analysis.3 Confirmation of strain identity was performed using PFGE and the strains were subjected to complete (O:K:H) serotyping at the Statens Serum Institut (Copenhagen, Denmark). Strains were also characterized with respect to phylogenetic group identity (A, B1, B2 or D) and virulence genes by PCR as described previously.3,4
Antibiotic susceptibility, isoelectric focusing and DNA sequencing
E. coli isolates were tested for susceptibility to: ampicillin, amoxicillin/clavulanic acid, piperacillin, mecillinam, cefadroxil, ceftazidime, cefuroxime, cefoxitin, chloramphenicol, gentamicin, tobramycin, streptomycin, nitrofurantoin, nalidixic acid, tetracycline, trimethoprim and sulphonamide using the disc diffusion method (Oxoid, Sollentuna, Sweden).5 The MIC of ampicillin was determined by Etest.5 ß-Lactamases were characterized by isoelectric focusing at The Swedish Institute for Infectious Disease Control (SMI), as previously described.6 The presence of the blaTEM gene was determined by PCR,7,8 followed by DNA sequencing of the products. The promoters of ampC genes were also sequenced.8
Total plasmid DNA was prepared (Plasmid Mini Kit, Qiagen) and transformed into E. coli DH5
(Invitrogen, Carlsbad, CA, USA). Transformants were selected on agar plates containing ampicillin (50 mg/L). Plasmid DNA was treated with HindIII and separated on a 0.7% agarose gel, stained with ethidium bromide, and visualized under UV-light.
For Southern-blot hybridization, cleavage products were transferred onto a nylon membrane (Amersham, Uppsala, Sweden), hybridized with a digoxigenin-labelled TEM probe (Roche Diagnostics GmbH, Mannheim, Germany), detected with anti-digoxigenin-AP and visualized with a colorimetric substrate (Roche Diagnostics).
Transfer of resistance by conjugation
Conjugation experiments were performed by the broth method, using E. coli DH5 as the recipient and ampicillin-resistant E. coli isolates obtained from the infant studied as donors. To investigate in vitro plasmid transfer from strain 29A to 29B, a high-level streptomycin-resistant mutant of strain 29B was selected. Transconjugants were selected on LB agar containing ampicillin (50 mg/L) and nalidixic acid (50 mg/L) or streptomycin (1 g/L), while donor frequency was estimated by plating a dilution of the conjugation mixture onto LB agar containing ampicillin.
|
Results and discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionFundingTransparency declarationsReferences |
|---|
Detection of transfer of antibiotic resistance and confirmation of strain identities
During screening of 272 E. coli strains from 128 infants in the ALLERGYFLORA birth-cohort study for antibiotic resistance, we identified an E. coli strain that became resistant during the course of colonization of a newborn infant's bowel. The boy was born in 1999 by normal vaginal delivery. He was admitted to hospital at 8 days of age, due to a suspected urinary tract infection, and trimethoprim was administered for 5 days. A urinary culture yielded E. coli at 100 000 cfu/mL. Due to the presence of enterococci in addition to E. coli in a second urinary sample, antibiotic treatment was switched to intravenous ampicillin for 5 days, followed by amoxicillin perorally for an additional 8 days. Finally, trimethoprim was administered prophylactically for 7 months according to clinical routine.
The infant carried two faecal E. coli strains during the first month of life. One strain, termed 29A, colonized the infant's bowel from 2 days until 1 year of age and was retrieved from seven consecutive stool samples. All isolates of strain 29A were highly resistant to ampicillin (MIC 
256 mg/L). Although all isolates were resistant to piperacillin (zone diameter 14 mm), the isolates obtained at 32 days and at 2, 6 and 12 months of age had a smaller zone diameter of 6 mm than those obtained earlier. All isolates of strain 29A were fully susceptible to the other antibiotics tested. Strain 29A had the O134:K1:H31 serotype and all isolates were identical according to RAPD and PFGE typing, virulence factor gene carriage (fimA, neuB) and phylogenetic group identity (D).
Another E. coli strain, termed 29B, was detected in the faecal samples obtained from the boy at 9, 16 and 32 days of age. The isolates of strain 29B obtained on the first two occasions were susceptible to all antibiotics tested (termed 29BS), whereas the isolate obtained at 32 days of age (termed 29BR) was resistant to ampicillin (MIC 256 mg/L) and piperacillin (zone diameter 14 mm). Strain 29B had the O15:K52:H1 serotype, which is typical of a globally spread pyelonephritogenic clone. A specific PCR assay confirmed that all isolates of strain 29B belonged to this clonal group.9 All isolates of strain 29B were homogenous with respect to RAPD pattern and virulence gene profile (fimA, iutA) and belonged to phylogenetic group D.
Characterization of ß-lactamases
Isolates of strain 29A, 29BS and 29BR were analysed by isoelectric focusing to detect ß-lactamases that may have been responsible for the observed resistance. Strain 29A was found to harbour two ß-lactamases, whose pIs were 5.4 and 9.0, respectively. A pI of 5.4 is characteristic of enzymes of the TEM-1 type, whereas a pI of 8.6/9.0 is instead characteristic of AmpC ß-lactamases. The ampC gene is ubiquitous in E. coli, but is normally expressed at low levels, insufficient to mediate phenotypic resistance. The isolate 29BS contained a single ß-lactamase with a pI of 8.6. Since this isolate was not phenotypically resistant, and since the pI was typical of an AmpC ß-lactamase, we assumed that strain 29B carried a chromosomal ampC gene expressed at a low level. The isolate 29BR carried two ß-lactamases, one of pI 8.4/8.6 (assumed to represent the AmpC ß-lactamase) and another of pI 5.4, typical of TEM-1-type ß-lactamases. Sequence analysis of ampC promoters did not show any changes in the different isolates of strain A or B, indicating that AmpC up-regulation did not occur in these isolates.
PCR and DNA sequencing revealed the presence of the same molecular variant of blaTEM-1b in isolates 29A and 29BR, but not in 29BS. We also analysed the promoter region of the blaTEM amplicons. A weak promoter (P3) and two strong overlapping (Pa/Pb) promoters have been described for blaTEM genes.10 The weak P3 promoter was found in the strain 29A isolate recovered at 2 days of age and the 29BR isolate. The strong Pa/Pb promoter was found in isolates of strain 29A recovered at 32 days and 6 months of age after treatment of the colonized infant with ampicillin and amoxicillin. The occurrence of the stronger promoter coincided with a decreased piperacillin zone and decreased susceptibility to amoxicillin/clavulanic acid. Thus, the TEM-1 promoter underwent changes in strain 29A during persistence of the strain in the bowel, after it had transferred its blaTEM gene to strain 29B and while the infant received treatment with ampicillin and amoxicillin.
We examined plasmid similarity between strain 29A and strain 29B by restriction fragment length polymorphism analysis. Three plasmid DNA fragments, 
5, 10 and 15–20 kb in size derived from a plasmid of 40 kb in size (termed pNK29), were detected in the 29A and 29BR isolates, but were absent in the 29BS isolate. The same three fragments were also detected in E. coli DH5 transformed with 29A plasmid DNA (Figure 1a). Plasmid DNA fragments were subjected to Southern-blot hybridization with a blaTEM-specific probe, which hybridized to a 10 kb band in the plasmid DNA digests from isolates 29A, 29BR and the transformant, but not to DNA obtained from the isolate 29BS (Figure 1b).
|
Transconjugation experiments demonstrated that pNK29 could be transferred from isolates 29A and 29BR to E. coli DH5
at frequencies of 4 x 10–8 and 8 x 10–7 transconjugants per donor, respectively. pNK29 could also be transferred from strain 29A to a high-level streptomycin-resistant mutant of 29BS at a frequency of 3 x 10–10 transconjugants per donor. The lower transfer frequency of pNK29 from strain 29A to strain 29BS than to E. coli DH5 suggests that transmission of pNK29 from strain 29A to strain 29B might only occur during strongly predisposing conditions, i.e. treatment of the infant with ampicillin and amoxicillin. Microbiota population levels of individual E. coli strains
The population counts of all strains on all culture occasions were determined by separately enumerating each colony and assessing its strain identity by RAPD. Strains 29A and 29B had similar faecal population counts (108.3 and 108.6 cfu/g) at 9 days of age. At 16 days of age, after 1 week of trimethoprim treatment, the population levels of both strains were reduced (106.4 and 103.9 cfu/g), especially that of strain 29B. As the child was switched to ampicillin and amoxicillin, the faecal population counts of strain 29A rose dramatically to 1011 cfu/g at 32 days of age, a level rarely obtained under normal conditions (only 3 out of 272 strains had this high population level). The increase in E. coli faecal population counts coincided with a drop in anaerobic, especially bifidobacterial population levels. Strain 29B, which appeared to have acquired ampicillin resistance between 16 and 32 days of age, also had a very high population density, i.e. 1010.3 cfu/g of faeces at 32 days of age. The high population counts of strain 29A may have facilitated its ability to act as a donor of a resistance plasmid. When conjugation occurred, the presence of ampicillin in the intestinal microbiota would confer an immediate advantage on the recipient of pNK29. This was clearly reflected in the counts of strain 29B.
These observations provide conclusive evidence of gene transfer in the human commensal bowel microbiota in a child treated with ampicillin. Our results could indicate that antibiotics may not only select for resistant bacteria, but may also increase their potential as donors of resistance genes by increasing their population counts, a serious consequence that has previously received little consideration. This suspicion may indicate that closer attention should be paid to the ecological effects of antibiotic treatment on the commensal microbiota.
|
Funding |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionFundingTransparency declarationsReferences |
|---|
The study was supported by grants from the Medical Faculty of Göteborg University, the Swedish Strategic Programme for the Rational Use of Antimicrobial Agents and Surveillance of Resistance, the Magnus Bergvall Foundation, the Wilhelm and Martina Lundgren Foundation and the May Flower Foundation.
|
Transparency declarations |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionFundingTransparency declarationsReferences |
|---|
None to declare.
|
Acknowledgements |
|---|
We thank L. Vincent Collins, C. Wennerås, E. R. B. Moore and P. Larsson for critical reading of the manuscript, Barbro Olssson-Liljequist, Swedish Institute for Infectious Disease Control, for isoelectric focusing of ß-lactamases and Charles Hannoun for analysis of gene sequences.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionFundingTransparency declarationsReferences |
|---|
1 Bidet P, Burghoffer B, Gautier V, et al. In vivo transfer of plasmid-encoded ACC-1 AmpC from Klebsiella pneumoniae to Escherichia coli in an infant and selection of impermeability to imipenem in K. pneumoniae. Antimicrob Agents Chemother (2005) 49:3562–5.
2 Adlerberth I, Hanson LA, Wold AE. The ontogeny of the intestinal flora. In: Development of the Gastrointestinal Tract—Sanderson I, ed. (1999) Hamilton, Ontario: B. C. Decker. 279–92.
3 Nowrouzian F, Hesselmar B, Saalman R, et al. Escherichia coli in infants' intestinal microflora: colonization rate, strain turnover, and virulence gene carriage. Pediatr Res (2003) 54:8–14.[CrossRef][Web of Science][Medline]
4
Clermont O, Bonacorsi S, Bingen E. Rapid and simple determination of the Escherichia coli phylogenetic group. Appl Environ Microbiol (2000) 66:4555–8.
5 The Swedish Reference Group for Antibiotics. (2005) www.srga.org; revised 15 March 2007 (22 August 2007, date last accessed).
6
Matthew M, Harris AM. Identification of ß-lactamases by analytical isoelectric focusing: correlation with bacterial taxonomy. J Gen Microbiol (1976) 94:55–67.
7
Brinas L, Zarazaga M, Saenz Y, et al. ß-Lactamases in ampicillin-resistant Escherichia coli isolates from foods, humans, and healthy animals. Antimicrob Agents Chemother (2002) 46:3156–63.
8 Olesen I, Hasman H, Aarestrup FM. Prevalence of ß-lactamases among ampicillin-resistant Escherichia coli and Salmonella isolated from food animals in Denmark. Microb Drug Resist (2004) 10:334–40.[CrossRef][Web of Science][Medline]
9
Johnson JR, Owens K, Sabate M, et al. Rapid and specific detection of the O15:K52:H1 clonal group of Escherichia coli by gene-specific PCR. J Clin Microbiol (2004) 42:3841–3.
10
Lartigue MF, Leflon-Guibout V, Poirel L, et al. Promoters P3, Pa/Pb, P4, and P5 upstream from blaTEM genes and their relationship to ß-lactam resistance. Antimicrob Agents Chemother (2002) 46:4035–7.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. E. Galen, J. Y. Wang, M. Chinchilla, C. Vindurampulle, J. E. Vogel, H. Levy, W. C. Blackwelder, M. F. Pasetti, and M. M. Levine A New Generation of Stable, Nonantibiotic, Low-Copy-Number Plasmids Improves Immune Responses to Foreign Antigens in Salmonella enterica Serovar Typhi Live Vectors Infect. Immun., January 1, 2010; 78(1): 337 - 347. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Trobos, C. H. Lester, J. E. Olsen, N. Frimodt-Moller, and A. M. Hammerum Natural transfer of sulphonamide and ampicillin resistance between Escherichia coli residing in the human intestine J. Antimicrob. Chemother., January 1, 2009; 63(1): 80 - 86. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Schjorring, C. Struve, and K. A. Krogfelt Transfer of antimicrobial resistance plasmids from Klebsiella pneumoniae to Escherichia coli in the mouse intestine J. Antimicrob. Chemother., November 1, 2008; 62(5): 1086 - 1093. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Karami, C. Hannoun, I. Adlerberth, and A. E. Wold Colonization dynamics of ampicillin-resistant Escherichia coli in the infantile colonic microbiota J. Antimicrob. Chemother., October 1, 2008; 62(4): 703 - 708. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. G. Tristram, M. J. Pitout, K. Forward, S. Campbell, S. Nichols, and R. J. Davidson Characterization of extended-spectrum {beta}-lactamase-producing isolates of Haemophilus parainfluenzae J. Antimicrob. Chemother., March 1, 2008; 61(3): 509 - 514. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||