Characterization of In40 of Enterobacter aerogenes BM2688, a Class 1 Integron with Two New Gene Cassettes, cmlA2 and qacF

Abstract

Enterobacter aerogenes BM2688, which is resistant to multiple antibiotics, and its aminoglycoside-susceptible derivative BM2688-1 were isolated from the same clinical sample. Strain BM2688 harbored plasmid pIP833, which carries a class 1 integron, In40, containing (in addition to qacEΔ1 and sul1, which are characteristic of class 1 integrons) four gene cassettes: aac(6′)-Ib, qacF, cmlA2, and oxa-9. The cmlA2 gene had 83.7% identity with the previously described nonenzymatic chloramphenicol resistance cmlA1 gene. The qacF gene conferred resistance to quaternary ammonium compounds and displayed a high degree of similarity with qacE (67.8% identity) which, however, has been found as part of a cassette with a very different 59-base element. The oxa-9 gene was not expressed due to a lack of promoter sequences. Study of the antibiotic-susceptible derivative BM2688-1 indicated that a 3,148-bp deletion between the 3′ end of the aac(6′)-Ib gene and the 3′ conserved segment of In40 was responsible for the loss of resistance. The occurrence of this DNA rearrangement, which did not involve homologous sequences, suggests that the In40 integrase could promote recombination at secondary sites.

Integrons are genetic elements that can integrate, by site-specific recombination, gene cassettes, usually antibiotic resistance genes, between two conserved segments (14, 28). The 5′ conserved segment contains the int gene encoding the integrase which catalyzes site-specific recombination. In class 1 integrons, the 3′ conserved segment carries qacEΔ1, a functional deletion derivative of the qacE gene, which specifies resistance to antiseptics and disinfectants; the sul-1 gene, which confers sulfonamide resistance; and an open reading frame (ORF; ORF5) of unknown function (13, 22, 39). Integrons are often part of transposons or plasmids of various incompatibility groups. Cassettes are individual mobile units composed of a gene and of a short inverted repeat called the 59-base element, located at the 3′ end of the gene, which is recognized by the integrase IntI (4). Over the last 50 years the spread of antibiotic resistance genes occurred by integration and excision of cassettes into integrons (4, 5). Multiple cassette insertions can occur, and more than 40 distinct cassettes have been identified (28). Genes other than those conferring antibiotic resistance have been described, such as qacE, which encodes an exporter protein mediating resistance to antiseptics and disinfectants (22, 27). Cassettes are always integrated in the same orientation and are transcribed from a promoter located in the 5′ conserved segment; the cmlA1 gene, however, carries its own promoter (1, 37).

Enterobacter aerogenes BM2688, which is resistant to multiple antibiotics, and its aminoglycoside-susceptible derivative BM2688-1 were isolated from the same clinical sample. Since an aac(6′)-Ib gene was detected in both strains by PCR, we attempted to determine the genetic event responsible for the loss of resistance in BM2688-1. We report on the characterization of a new integron, In40, located on a large plasmid in E. aerogenes BM2688. This element contained four cassettes including two new genes, qacF and cmlA2. Analysis of E. aerogenes BM2688-1 indicated that the loss of aminoglycoside resistance resulted from an unusual recombination event in In40.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions.

The strains and plasmids used in the study are listed in Table 1. E. aerogenes BM2688 and BM2688-1 were isolated in 1992 from a human urine sample at the Saint-Michel Hospital in Paris, France. The bacteria were grown in brain heart infusion broth (Difco Laboratories, Detroit, Mich.) or on Mueller-Hinton agar (Sanofi Diagnostics Pasteur, Marnes-la-Coquette, France). Antibiotic susceptibility was determined by disk diffusion on Mueller-Hinton agar. The method of Steers et al. (36) with 10⁴ CFU per spot was used to determine the MICs. Induction of chloramphenicol resistance by pregrowth in the presence of 1 μg of the antibiotic per ml was performed as described previously (7). Incubations were done at 37°C.

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Relevant characteristic(s)a	Source or reference
E. coli
JM83	araφ (lac mutant proAB) rpsL [Φ80Δ(lacZ)M15]	42
J5-3	pro met rpoB	31
E. aerogenes
BM2688	Ak Km Su Qa Tc Tm	Wild strain
BM2688-1	Su Tc	Wild strain
Plasmids
pUC18	Tra− Mob− Ap	42
pAT671	Tra− Mob− Ak Ap Cm Km Qa Su Tm	6.5-kb BamHI fragment from BM2688-1 into pUC18
pAT672	Tra− Mob− Ak Ap Km Qa Tm	3.2-kb BamHI-SmaI fragment from pAT671 into pUC18
pAT673-1	Tra− Mob− Ap Cm	1,678-bp PCR fragment from pAT671 into pUC18
pAT673-2	Tra− Mob− Ap Cm	Differs from pAT673-1 by the insert orientation
pBGS18	Tra− Mob− Km	35
pAT674	Tra− Mob− Ak Cm Km Qa Su	6.5-kb BamHI fragment from BM2688-1 into pBGS18
pAT675	Tra− Mob− Ap Km Su	3.3-kb BamHI-SmaI fragment from pAT671 into pBGS18
pCRII	Tra− Mob− Ap Km	TA cloning kit (Invitrogen)
pAT676	Tra− Mob− Ap Km	951-bp HaeIII IPCR fragment from BM2688-1 into the pCRII vector
pIP833	Tra+ In40 Ak Cm Km Qa Su Tm	>100-kb plasmid from E. aerogenes BM2688
pIP833-1	Tra+ In40Δ Su	>100-kb plasmid from E. aerogenes BM2688-1

^aAbbreviations: Tra⁻, non-self-transferable; Mob⁻, nonmobilizable; Ak, amikacin resistance; Ap, ampicillin resistance, Cm, chloramphenicol resistance; Km, kanamycin resistance; Qa, quarternary compound resistance; Su, sulfonamide resistance; Tc, ticarcillin resistance; Tm, tobramycin resistance. 

Preparation and analysis of DNA.

Total DNA and small- and large-scale plasmid DNA preparations were prepared as described previously (30). DNA for PCR was obtained by boiling as described previously (16). Electrophoresis was performed in 0.8% agarose gel (Sigma Chemical Co., St. Louis, Mo.) with a Tris-borate-EDTA buffer system.

Genetic techniques.

Transformation of E. coli JM83 was performed as described previously (32). Selective antibiotic concentrations were as follows: ampicillin, 100 μg/ml; kanamycin, 20 μg/ml; rifampin, 250 μg/ml; and tobramycin, 10 μg/ml.

DNA techniques.

Fragments internal to the aac(6′)-Ib and qacF genes were obtained by PCR with the primers listed in Table 2. PCR products were separated by agarose gel electrophoresis, purified (Sephaglass BandPrep kit; Pharmacia, St Quentin-en-Yvelines, France), and radiolabeled by nick translation (29). For Southern hybridization, DNA was immobilized on Nytran membranes (Schleicher & Schuell, Dassel, Germany) as described previously (32). Prehybridization and hybridization were carried out for 5 and 15 h, respectively, at 65°C in 6× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) containing 0.5% sodium dodecyl sulfate and 0.05% nonfat dry milk (32). Inverted PCR (IPCR) was performed as described previously (41) with primers IL and IR directed outward (Table 2). Total DNA from BM2688-1 was digested with HaeIII, self-ligated, and used as a template for IPCR. The IPCR product was separated by agarose gel electrophoresis, purified, and cloned into pCRII with the TA cloning kit (Invitrogen, San Diego, Calif.). A 395-bp fragment of the aac(6′)-Ib gene was amplified by PCR with primers BL and BR (Table 2) as described previously (24). A 1,678-bp fragment containing the cmlA2 cassette was obtained by PCR with primers CL and CR (Table 2). Amplifications were performed in 50-μl reaction mixtures consisting of 1× Taq or Pfu DNA polymerase buffer, 200 μM deoxyribonucleoside triphosphates, 50 pmol of each primer, 1 U of Taq or Pfu DNA polymerase, and 25 ng of DNA prepared by boiling. PCR was performed in a DNA Thermal Cycler 480 (Perkin-Elmer Cetus, Norwalk, Conn.). Random amplified polymorphism DNA (RAPD) analysis was performed as described previously (25), and the ability to distinguish unrelated strains was tested against five E. aerogenes clinical isolates.

TABLE 2.

Oligodeoxyribonucleotides used for amplification

Primer	Sequence	Location	Strand	Size (bp) of product
BL	5′-TATGAGTGGCTAAATCGAT-3′	aac(6′)-Ib (28)a	+	395
BR	5′-CCCGCTTTCTCGTAGCA-3′	aac(6′)-Ib (422)	−
QL	5′-GTCATCGCAACTTCCGCACTG-3′	qacF (43)	+	245
QR	5′-CTGACGATAAGTCCCAT-3′	qacF (287)	−
IL	5′-GAGTTGCTGTTCAATGA-3′	aac(6′)-Ib (331)	+	948 (IPCR)
IR	5′-TAACTTTGTTTTAGGGC-3′	5′ conserved segment	−
L1	5′-GGCATCCAAGCAGCAAG-3′	5′ conserved segment	+
CL	5′-ATGGGACTTATCGTCAG-3′	End of qacF (271)	+	1,678
CR	5′-TAACGCTTGAGCTCAGCCGACCGA-3′	cmlA2 cassette (1345)	−

^aThe positions of the 5′ ends of the internal oligonucleotides are indicated (position numbers are in parentheses) relative to the putative start codon of the corresponding gene. 

Cloning and DNA sequencing.

A 6.5-kb BamHI fragment of BM2688-1 was cloned into the pUC18 vector by selecting Escherichia coli JM83 transformants on 10 μg of tobramycin per ml (42). Double-stranded DNA sequencing was carried out with synthetic oligonucleotides (Unité de Chimie Organique, Institut Pasteur). Sequencing reactions were performed by the dideoxynucleotide chain termination method (33) with a modified T7 DNA polymerase and [α-³⁵S]dATP. DNA fragments were resolved by electrophoresis on 8% vertical polyacrylamide gels containing 8 M urea.

Computer analysis of sequence data.

Nucleotide and amino acid sequences were analyzed and compared by use of GenBank, EMBL, and Swiss-Prot databases with the FASTA program (Genetics Computer Group software) (23).

Enzymes and chemicals.

T4 DNA ligase, restriction endonucleases (Amersham, Buckinghamshire, England), and the Sequenase version 2.0 DNA sequencing kit (United States Biochemical Corp., Cleveland, Ohio) were used according to the recommendations of the manufacturers. Lysozyme, cetyltrimethylammonium bromide (CTAB), and chloramphenicol were obtained from Sigma Chemical Co., and RNaseA (bovine pancreas) was obtained from Calbiochem-Behring (La Jolla, Calif.). Nick translation kits were obtained from Bethesda Research Laboratories, Inc. (Gaithersburg, Md.), and [α-³²P]dCTP and [α-³⁵S]dATP (400 Ci/mmol) were obtained from the Radiochemical Centre (Amersham, Buckinghamshire, England). Taq DNA polymerase was purchased from Bioprobe Systems (Montreuil-sous-Bois, France), and Pfu DNA polymerase was obtained from Stratagene (Stratagene Cloning Systems, La Jolla, Calif.). 2′-N-Ethylnetilmicin and 6′-N-ethylnetilmicin were kindly provided by the Schering-Plough Research Institute (Kenilworth, N.J.). Antibiotic disks were from Sanofi Diagnostics-Pasteur.

Nucleotide sequence accession number.

The nucleotide sequences of the qacF and cmlB genes have been deposited in the GenBank data library (GenBank, Los Alamos, N.M.) under accession no. AF034958.

RESULTS AND DISCUSSION

Antibiotic resistance of E. aerogenes BM2688 and BM2688-1.

E. aerogenes BM2688 and BM2688-1 were isolated from a human urine sample. Strain BM2688 was resistant to netilmicin, tobramycin, and amikacin but was susceptible to gentamicin. The activity of 2′-N-ethylnetilmicin against this strain was diminished compared to that of 6′-N-ethylnetilmicin, suggesting the production of a 6′-N-aminoglycoside acetyltransferase of type I [AAC(6′)-I] (34). The strain was also resistant to CTAB, sulfonamide, and ticarcillin. E. aerogenes BM2688-1 was susceptible to all aminoglycosides and CTAB but remained resistant to sulfonamide and ticarcillin. The two strains were isolated on the basis of susceptibility or resistance to aminoglycosides. BM2688 and BM2688-1 had indistinguishable RAPD profiles, in contrast to five E. aerogenes strains used as controls, which had distinguishable RAPD profiles (data not shown). These results suggest that E. aerogenes BM2688 and BM2688-1 are probably related to each other. Tobramycin resistance was transferred by conjugation from BM2688 to E. coli J5-3 at a frequency of 10⁻⁸. The transconjugants were resistant to the aminoglycosides that are modified by an AAC(6′)-I enzyme, CTAB, and sulfonamide; they were also resistant, albeit at a low level, to chloramphenicol (Table 3).

TABLE 3.

MICs of antibiotics and CTAB for the strains

Strain	MIC (μg/ml)
	Ampicillin	Chloramphenicol			Tobramycin	CTAB
		Noninduceda		Inducedb
		24 h	48 h
E. aerogenes
BM2688	>128	4	8	8	32	800
BM2688-1	>128	4	4	4	0.5	200
E. coli
J5-3	2	4	4	4	0.25	200
J5-3(pIP833)	2	12	24	24	16	400
JM83	4	4	4	4	0.5	200
JM83(pAT672)	>128	NDc	ND	ND	32	400
JM83(pAT673-1)	>128	16	16	ND	ND	ND
JM83(pAT673-2)	>128	8	8	ND	ND	ND
JM83(pAT674)	4	ND	ND	ND	ND	ND
JM83(pAT675)	32	ND	ND	ND	ND	ND

^a24 and 48 h, after 24 or 48 h of incubation at 37°C, respectively. 

^bInduction was with 1 μg of chloramphenicol per ml added to exponentially growing cells 60 min before plating. 

^cND, not determined. 

Comparative analysis of the plasmid contents of BM2688, a transconjugant, and BM2688-1 by agarose gel electrophoresis of crude bacterial lysates indicated that the acquisition of antibiotic resistance was due to the transfer of a large plasmid of ca. 100 kb, designated pIP833. The plasmid in BM2688-1 was designated pIP833-1. In addition, BM2688 and BM2688-1 harbored another plasmid mediating ticarcillin resistance.

Detection and cloning of the aac(6′)-Ib gene.

Amplification of BM2688 and BM2688-1 plasmid DNA with oligonucleotides BL and BR specific for the aac(6′)-Ib gene (Table 2) gave rise, in each case, to a fragment of 395 bp (data not shown). Plasmid DNA from the two strains was digested with BamHI and hybridized by the technique of Southern with the amplified fragment internal to aac(6′)-Ib. The gene was found to be part of a 6.5-kb fragment in BM2688 and a 3.5-kb fragment in BM2688-1. Plasmid DNAs from BM2688 and pUC18 were digested with BamHI, mixed, ligated, and introduced by transformation into E. coli JM83 with selection on ampicillin and tobramycin, and the transformants were screened for their plasmid contents by agarose gel electrophoresis. The smallest recombinant plasmid conferring tobramycin resistance, pAT671, contained a 6.5-kb BamHI insert.

Nucleotide sequence of the insert in pAT671.

The sequence of the 6.5-kb BamHI insert of pAT671 was determined and its genetic organization is shown in Fig. 1A. Eight ORFs were present. At the 5′ and 3′ ends of the insert, the sequence was identical to a portion of the 5′ and 3′ conserved segments of class 1 integrons (2). The 5′ conserved segment was identical to that of the class 1 integron which encodes a site-specific class 1 integron integrase (2). The 3′ conserved segment was composed of qacEΔ1 fused to the sul-1 gene and the first 233 bp of ORF5. These data indicate that the insert in plasmid pAT671 contains an integron, designated In40, which belongs to class 1, the most prevalent in clinical isolates.

FIG. 1.

Physical map of the 6.5-kb BamHI fragment internal to integron In40 from BM2688 (A) and of the 3.4-kb recombinant derivative from BM2688-1 (B). The 5′ and 3′ conserved segments are represented by thick lines. Genes cassettes are shown as thin lines with a filled circle to represent the 59-base element. Open arrows represent coding sequences. Nucleotides adjacent to the recombination site between aac(6′)-IbΔ and the 3′ conserved segment are indicated. Identity with the consensus GTTRRRY recombination site is underlined on the complementary strand. The CL, CR, IL, and IR amplification primers are indicated by arrowheads. The inserts of the constructed plasmids are shown.

The 5′ and 3′ conserved segments flanked four ORFs. At the 3′ end of each ORF, structures homologous to the 59-base element were present, suggesting that each ORF was part of a cassette (5, 38). The 59-base element is an imperfect inverted repeat sequence which acts as a recombination site. The recombination crossover occurs after the first guanine of the conserved GTTRRRY (R = purine, Y = pyrimidine) core site, located at one end of the 59-base element (12). In40 contained the aac(6′)-Ib cassette, which is wide spread among gram-negative bacteria (9), and the oxa-9 cassette found in Tn1331 (40).

The qacF cassette.

The second cassette contained an ORF of 345 nucleotides (Fig. 2). A putative initiation GTG codon at position 115 was preceded at 7 bp by a ribosome-binding site (RBS)-like sequence. The coding sequence, designated qacF, which could direct the synthesis of a 110-amino-acid protein, had 67.8% identity with the sequence of the qacE gene (22). The qacE gene specifies an exporter protein mediating resistance to intercalating dyes and quaternary ammonium compounds and has been found in the class 1 integron of transposon Tn402, later designated Tn5090 (27). The qacF and qacE cassettes diverge at their extremities and particularly at their 3′ ends; a 60-bp sequence was present downstream from qacF, whereas a 141-base element has been associated with qacE (27). The qacF 59-base element shared 88.3% identity with that of catB3 (3) (Fig. 3). The members of the family of 59-base elements vary in length, and the family includes members with longer imperfect inverted repeats, up to 141 bp, that retain similarity to the consensus at their termini and that are active in Int-mediated site-specific recombination (5, 38).

FIG. 2.

Comparison of the sequences of the qacE cassette from R751 and the cmlA1 cassette of integron In4 and the corresponding genes of In40 from BM2688. The nucleotide sequence of BM2688 is indicated in capital letters. The start and stop codons for qacF and cmlA2 are indicated in boldface. The putative RBS (wavy line) and promoter sequences (−35 and −10) are underlined. Stop codons are indicated by an asterisk; the plus signs indicate identical nucleotides in the two sequences. Dashes indicate gaps introduced to optimize similarity. The GTT sequence of the core sites for recombination are double underlined. The 59-base elements are overlined. The deduced amino acid sequences of QacF and CmlA2 are indicated below the corresponding nucleotide sequences. Amino acids that differ in the sequences are presented above the nucleotide sequences. Oligodeoxyribonucleotides were complementary to the integron In40 from positions 157 to 177 for QL, positions 385 to 401 for CL, positions 401 to 385 for QR, and positions 2062 to 2042 for CR.

FIG. 3.

Alignment of the 59-base elements from catB3 (3), qacE (27), and qacF cassettes. Asterisks indicate bases identical in all three sequences. Numbers indicate the number of bases in the central region of each 59 bp.

Hypotheses for the mechanism of formation of cassettes have been proposed (29), but the question of whether the genes and the 59-bp elements have independent origins remains to be elucidated. Closely related genes associated with closely related 59-bp elements have been described, e.g., catB3 and catB5 cassettes (3). By contrast, catB3 and qacF represent an interesting example of 59-bp elements associated with genes encoding different functions. A similar observation was made with the 90% sequence identity between the Vibrio cholerae repeated sequences and the 59-base element associated with blaP3, an integron-associated gene encoding the CARB-4 carbenicillinase (21). These data suggest that cassettes can exchange 59-bp elements via integrase-mediated recombination at the internal boundaries of the two 59-bp elements instead of at the normal position at the outer boundaries of the 59-bp elements (29).

To study the phenotype conferred by qacF, a 3.2-kb BamHI-SmaI fragment from pAT671 was subcloned into pUC18, generating pAT672. The MICs of CTAB for E. coli JM83 and JM83/pAT672 were 100 and 400 mg/liter, respectively, indicating that QacF confers resistance to quarternary compounds. The deduced protein, QacF, shares 75% identity with QacE (22), 37.6% with QacC, an antiseptic resistance protein from Staphylococcus aureus (19), and 70.1% with Ebr, an E. coli protein mediating resistance to ethidium bromide (26). These proteins form a family of small multidrug export proteins that use proton motive force to energize transport and mediate resistance to antiseptics and disinfectants (22).

The cmlA2 cassette.

The third cassette, spanning 1,548 nucleotides and extending from bases 512 to 2059 (Fig. 2), contains an ORF of 1,230 nucleotides starting at the TGA codon at position 660. A putative GTG initiation codon at position 718 was preceded at 8 bp by a RBS-like sequence. This coding sequence, designated cmlA2, shared 83.7% identity with the cmlA1 gene of the class 1 integron In4 in Tn1696 which confers nonenzymatic chloramphenicol resistance (1, 37). The gene was 27 bp shorter than cmlA1 as the result of a guanine insertion at position 1940 in cmlA2, which generated a frame shift leading to a stop codon at position 1945. A 1,678-bp PCR fragment obtained with primers CL and CR (Table 2 and Fig. 1), which include the cmlA2 cassette, was cloned into pUC18 in both orientations, and the resulting plasmids, pAT673-1 and pAT673-2, respectively, conferred chloramphenicol resistance to E. coli (Table 3). The deduced protein of 409 amino acids, CmlA2, shared 85.5% identity with CmlA1 and 50.3% identity with the polypeptide predicted from the Pasteurella piscicida flo gene which confers resistance to florfenicol (17). The cmlA2 59-bp element was 70 bp in length, and the sequence differs at four positions from that of the 59-bp element of cmlA1 (Fig. 2). The mechanism of resistance to chloramphenicol mediated by cmlA1 has not been established. CmlA1 is a hydrophobic polypeptide, and resistance is thought to be due to reduced uptake of chloramphenicol (1, 37).

Expression of the cmlA2 gene.

The cassettes are usually inserted in the same orientation and are under the control of the common promoter P_ant located in the 5′ conserved segment (6). Of the cassettes described to date, only cmlA1, qacE, and qacEΔ1 contain a promoter-like sequence (1, 11, 37). Analysis of the region upstream from cmlA2 showed a putative promoter consisting of −35 (TCGCGG) (positions 533 to 538) and −10 (TACGAT) motifs separated by 17 nucleotides (underlined nucleotides indicate identity with the consensus −10 and −35 promoter elements recognized by the E. coli ς⁷⁰ factor). The −10 motif was identical to that proposed for cmlA1, whereas the −35 motif differed by two base pairs. The 1,678-bp PCR product obtained with primers CL and CR (Table 2) containing the cmlA2 cassette was cloned into pUC18 in both orientations. The resulting plasmids, pAT673-1 and pAT673-2 (Table 1), conferred chloramphenicol resistance to E. coli JM83 at similar levels (Table 3). This indicates that the cmlA2 gene is expressed from its own promoter. The region upstream from cmlA2 contains a small ORF that could encode a 9-amino-acid peptide that is closely related to the leader peptides of cat genes and that differs from that of cmlA1 by Lys-6 instead of Asn. This region also contains inverted repeats capable of forming alternate stem-loop structures (10, 37). These features are similar to those found upstream from the inducible cat and ermC genes, which are regulated by transcriptional attenuation (10, 20). The cmlA2 gene in plasmid pIP833 conferred very low level inducible chloramphenicol resistance to E. coli J5-3 and E. aerogenes BM2688 (Table 3). Inducible expression of cmlA1 from plasmid R26 in E. coli K-12 by subinhibitory concentrations of chloramphenicol has been demonstrated previously (7). These data suggest that the regulation of cmlA2 could be similar to that of cmlA1.

Expression of the oxa-9 gene.

In In40, the oxa-9 cassette was located downstream from cmlA2 (Fig. 1), and analysis of the region upstream from oxa-9 did not reveal any promoter-like sequence. By contrast, in Tn1331 the oxa-9 gene is located immediately downstream from an aadA cassette deleted of most of the 59-base element. It has been shown that in this structure, the oxa-9 gene is transcribed not only from the common promoter P_ant but also from a second promoter located at the end of the aadA gene (40). To study the expression of oxa-9, the 6.5-kb BamHI insert of pAT671 was cloned into pBGS18, generating pAT674, which did not confer ampicillin resistance to E. coli JM83. In In40, the oxa-9 gene was located downstream from the cmlA2 cassette and could be transcribed from the cmlA2 promoter. By contrast, recombinant plasmid pAT675, in which the oxa-9 gene was placed under the control of the lacZ promoter by subcloning the 3.3-kb SmaI fragment of the pAT671 insert into pBGS18, conferred ampicillin resistance to E. coli JM83 (Table 3). Taken together, these results indicated that the oxa-9 gene encoded a functional β-lactamase which was not expressed in the In40 structure. That could result from the weakness of the cmlA2 promoter or from an attenuation effect of the cmlA2 59-base element (6, 7), or both.

Loss of antibiotic resistance by BM2688-1.

Strain BM2688-1 differs from BM2688 by the loss of resistance to aminoglycosides and to very low levels of chloramphenicol. A probe made from a 245-bp PCR fragment internal to the qacF gene obtained with primer QL and QR (Table 2) hybridized to the 6.5-kb BamHI fragment of plasmid pIP833 from BM2688 but not to pIP1833-1 DNA (data not shown). Amplification of pIP833-1 DNA with oligonucleotides L1 and BR (Table 2), which are complementary to the 5′ conserved segment (18) and aac(6′)-Ib, respectively, gave rise to a fragment of 579 bp (data not shown), suggesting that the aac(6′)-Ib cassette is also part of an integron in BM2688-1. Sequencing of the 3′ end of aac(6′)-Ib and of the adjacent region following IPCR (Table 1) indicated that In40 had lost 3,148 bp that included 37 bp at the 3′ end of aac(6′)-Ib, the downstream 59-base element, the qacF, cmlA2, and oxa9 cassettes, and the first 36 bp of the 3′ conserved element. Recombination within the aac(6′)-Ib gene led to a protein 35 amino acids longer than AAC(6′)-Ib. The first 161 amino acids of the fusion protein are those of AAC(6′)-Ib; the missing 11 amino acids therefore appeared to be critical for the expression of the resistance.

The deletion took place between nonhomologous recombining sites, which suggests a RecA-independent process. This event may result from slippage of DNA polymerase during replication or from integrase-mediated recombination between the GTTTGGC and GTTTGAA motifs located on the complementary strands (underlined nucleotides indicate identity with the consensus GTTRRRY recombination site) (Fig. 1B). IntI1 has been shown to catalyze recombination events involving a 59-base element or attI and a secondary site (8, 15, 30, 38). However, it is not known if recombination between secondary sites can occur. Such an event is likely to occur at an extremely low frequency and can thus remain undetected. Although the role of the integrase in the deletion observed in In40 has not been established, two observations are consistent with our hypothesis: (i) the high level of homology of the recombination sites with the 7-bp core site sequence and (ii) the fact that IntI1-mediated recombination involves a crossover between the G at the first T in the GTT triplet (38), which is consistent with the deletion observed in pIP833-1 (Fig. 1). Interestingly, the deletion had produced a new 3′ conserved segment in which the defective aac(6′)-Ib was fixed because of the loss of the 59-bp element; meanwhile, the attI1 was still present.

Conclusions.

Three classes of integrons and more than 40 gene cassettes have been identified so far. In40 is a class 1-associated integron borne by the self-transferable pIP833 plasmid in E. aerogenes BM2688. This element contained four cassettes including qacF and cmlA2 genes. The cmlA1 and cmlA2 cassettes were closely related and probably derive from a common ancestor. In contrast, the qacE and qacF genes, which are also closely related, are part of cassettes that contain distinct 59-base elements. This observation implies the independent genesis of two cassettes by acquisition of 59-base elements following an unexplained mechanism (29). Another peculiarity of In40 is the lack of oxa-9 expression from the cmlA2 promoter leading to the presence of a silent gene. In addition, this study has revealed a surprising genetic event: a deletion, possibly mediated by the In40 integrase, between secondary recombination sites. This excisive recombination was not selected by therapy since it resulted in the loss of antibiotic resistance. Finally, BM2688-1 harbored an integron with the attI motif as an unique recombination site for the integrase and a new 3′ conserved segment. The latter contains an altered aac(6′)-I gene resulting from recombination at an unusual site. A similar genetic event could have generated qacEΔ1, which is part of the class 1 integrons.