Biochimica et Biophysica Acta 1489 (1999) 445^451 www.elsevier.com/locate/bba
Short sequence-paper
Molecular characterisation of In51, a class 1 integron containing a novel aminoglycoside adenylyltransferase gene cassette, aadA6, in Pseudomonas aeruginosa1 Thierry Naas a
a;b;
*, Laurent Poirel a , Patrice Nordmann
a
Service de Bacte¨riologie-Virologie, Hoªpital de Biceªtre, Assistance Publique-Hoªpitaux de Paris, 78 rue du Ge¨ne¨ral Leclerc, 94275 Le Kremlin-Biceªtre Ce¨dex, France b Service de Bacte¨riologie-Virologie, Hoªpital Antoine Be¨cle©re, Assistance Publique-Hoªpitaux de Paris, 92141 Clamart, Faculte¨ de Me¨decine Paris-Sud, Paris, France Received 3 June 1999; received in revised form 27 September 1999; accepted 8 October 1999
Abstract Polymerase chain reaction-amplification and subsequent sequencing of the variable region of a novel integron, In51, from Pseudomonas aeruginosa revealed the presence of a novel aminoglycoside adenylyltransferase gene, aadA6, together with an open reading frame of unknown function, orfD. AADA6 enzyme has only 75% amino acid identity with AADA1 and is able to confer high level resistance to streptomycin and spectinomycin in Escherichia coli. ß 1999 Elsevier Science B.V. All rights reserved. Keywords: Integron; aadA; Aminoglycoside resistance; (Pseudomonas aeruginosa)
Bacterial resistance to aminoglycosides is often due to enzymatic inactivation by acetyltransferases, nucleotidyltransferases (adenylyltransferases) or phosphotransferases [1]. The ANT(3Q)(9) or AADA aminoglycoside resistance enzymes, encoded by aadA genes, are adenylyltransferases that confer resistance to streptomycin and spectinomycin [1]. These enzymes add an adenyl-group to the 3Q-hydroxyl group on the amino-hexose III ring of streptomycin and to the 9-hydroxyl group on the actinamine ring of spectinomycin [1]. So far, only two major and distinct versions of aadA genes are known. They have been * Corresponding author. Address a. Fax: +33-1-45-21-63-40; E-mail:
[email protected] 1 The GenBank accession number of the published sequence is AF140629.
designated aadA1, found in Tn7 and Tn21, and aadA2, found in Tn1696 and on plasmid pSA [2]. The DNA sequences of aadA1 [3^5] and aadA2 [6,7] genes share 88% identity. Most of the aadA genes described are point mutant derivatives of either aadA1 or aadA2 [2,8]. Besides these genes, four other aadA genes have been described in single bacterial isolates: aadA3, isolated once from Escherichia coli, has 97% identity with aadA2 (unpublished, GenBank accession number AF047479); aadA4 (formerly aadA3), isolated once from E. coli has 57% identity with aadA1 (unpublished, GenBank accession number Z50802); aadA5, isolated once from E. coli has 95% identity with aadA4 (unpublished, GenBank accession number AF169041); and aadAsc, isolated from Salmonella choleraesuis, which shares only 43% identity with aadA1 [9]. AADA enzymes are
0167-4781 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 7 8 1 ( 9 9 ) 0 0 2 0 2 - X
BBAEXP 93344 1-12-99
446
T. Naas et al. / Biochimica et Biophysica Acta 1489 (1999) 445^451
reported in most Gram-negative species, including Enterobacteriaceae and Pseudomonas aeruginosa [8], and in some Gram-positive species; i.e., Corynebacterium glutamicum [10] and recently, Enterococcus faecalis [11]. Aminoglycoside-resistant strains often emerge as a result of acquiring plasmid-borne genes encoding aminoglycoside-modifying enzymes [12]. Many of these genes are associated with transposons, which help in their rapid spread across bacterial species boundaries. In addition, most aadA genes are encoded in gene cassettes that are present on the variable region of integrons [2,13]. Integrons are genetic structures capable of integrating individual gene cassettes encoding antibiotic resistance genes and are themselves located on either plasmids or transposons or both [2,13,14]. Gene cassettes comprise a gene, often an antibiotic resistance gene, and a recombination site, known as 59-base element (59-be), located downstream of the integrated genes [13]. The most highly conserved features of 59-bes are a 7-bp core site with the consensus `GTTRRRY' motif located at the right-hand end of the element (furthest from the 3P-end of the cassette-encoded gene) and an inverse core site with the consensus `RYYYAAC' motif at the left-hand end [2,13]. All the aadA genes isolated so far, except aadAsc from S. choleraesuis, are genecassette encoded [9]. The two major gene cassettes, aadA1 and aadA2, have distinguishable 59-bes and 5P core sites (GTTAAAC for aadA1 and GTTAGAC for aadA2) (Fig. 1C). A recent study has shown that aadA genes are the most frequently integronassociated resistance genes found among clinical enterobacterial isolates in Europe [8]. The integrons most commonly characterised from antibiotic-resistant clinical isolates from members of the Enterobacteriaceae and Pseudomonad families belong to class 1 [2,8,13]. They commonly possess two conserved regions located on either side of the integrated gene cassettes. The 5P-conserved segment (5P-CS) includes a gene, intI1, encoding the integrase; attI1, the cassette integration site; and the promoters that are responsible for expression of cassette genes [2,13]. The 3P-conserved segment (3P-CS) includes along with orf5, an open reading frame (ORF) of unknown function, the disinfectant (qacEv1) and the sulfonamide (sulI) resistance determinant [2,13]. Primers to these conserved segments are usually used for
the ampli¢cation of the variable regions of the class 1 integrons. In addition, de¢ning a cassette when it is the only integrated cassette is normally done by identifying the known boundaries of the 5P-CS and the 3P-CS. Pseudomonas aeruginosa JES was isolated in August 1998, at the Hoªpital de Biceªtre, Le KremlinBiceªtre (suburb of Paris), France, from an urinary tract infection of a 52-year-old woman hospitalised in the orthopaedic ward [15]. This patient had likely imported the strain from Thailand where she was previously hospitalised [15]. Previous studies have shown that P. aeruginosa JES was multi-resistant to most antibiotics and that it harboured In50, a class 1 integron encoding, within its variable region, the gene for the extended-spectrum M-lactamase VEB-1 [15]. In an attempt to further characterise the integron content of P. aeruginosa JES, a polymerase chain reaction (PCR) approach using class 1 speci¢c integron primers was used along with integrase-speci¢c hybridisation experiments. PCR ampli¢cation using the 5P-CS and the 3P-CS primers, which allow the ampli¢cation of the entire variable region of class 1 integrons, (Table 1) revealed two amplicons of 1.2 kb and 5 kb, the latter being the known In50 that contains the extendedspectrum L-lactamase gene blaVEBÿ1 [15,16]. The 1.2-kb amplicon was puri¢ed using Qiaquick PCR puri¢cation kit (Qiagen) and subsequently sequenced on both strands, using an Applied Biosystem sequencer (ABI 373), revealing the presence of a su¤ciently large ORF of 842 bp encoding a novel aminoglycoside resistance gene, aadA6 (Fig. 1A,B). The overall GC content of aadA6 gene was 53%, which did not lie within the expected range of GC content (60.1^69.5%) for P. aeruginosa genes [17]. The value of GC content is more typical of Enterobacteriaceae genes. The 1.2-kb PCR product containing aadA6 was cloned into pCR-ScriptCam (Stratagene), resulting in the recombinant plasmid, pAADA6, which was transformed into E. coli XL1 blue electro-competent cells. Plasmid DNA extraction, analysis and electroporation were performed as described elsewhere [15,16]. Plasmid pAADA6 conferred in E. coli a similar resistance phenotype to pLPO-1, which expresses AADA2 (data not shown). The minimum inhibitory concentrations (MICs) were determined by an agar dilution technique on Mueller^Hinton plates
BBAEXP 93344 1-12-99
T. Naas et al. / Biochimica et Biophysica Acta 1489 (1999) 445^451
447
Fig. 1. (A) Schematic representation of In51 containing the aadA6 and orfD gene cassettes. The intI1 gene, which codes for the integrase, is included in the 5P-CS (conserved segment). The 3P-CS found downstream of the integrated gene cassettes includes the sulfonamide resistance gene, sul1, and the disinfectant resistance determinant, qacEv1. Inserted genes are indicated by boxes and the arrows within the boxes indicate their translational orientation. Numbered arrowheads below In51 represent the primers used in this study for integron characterisation as described in Table 1. (B) Schematic representation of the variable region of In51. The coding regions are boxed and relevant nucleotides for the coding sequence, gene cassettes and integron features are also shown. AadA6 starts with an `ATG' codon and ends with an `TAA' codon while orfD starts with an `CTG' codon and ends with an `TAA' codon. `//' indicates an interruption of the DNA sequence. The underlined and double-underlined sequences represent respectively the core sites and inverse core sites for the two gene cassettes. Boundaries of aadA6 and orfD gene cassettes are represented below the sequence by divergent arrows. The composite-attI1 recombination site and the composite-cassette-associated recombination sites known as composite-59-be are represented above the sequence by divergent arrows. Parts of the 5P-CS and of the 3P-CS are represented by open arrows. The entire nucleotide sequence will appear under the GenBank accession number AF140629. (C) Comparison of the 59-be sequences associated with aadA genes. The sequences of the 59-bes are in a circular gene cassette con¢guration, which combines the end of the aadA gene cassette with the start of the same cassette. Sequence di¡erences are indicated below the 59-be sequence of aadA1 gene cassette, as found in Tn7 and Tn21 [3^5] while sequence identities are represented by dashes. The boxes represent four regions (1L, 2L, 2R and 1R) found to be conserved within the 59-bes [23]. The references and GenBank accession numbers are as follows: aadA1 (GenBank accession number X02340, [3]) ; aadA1b (M95287, [19]); aadA2 (X68227 [7]); aadA3 (AF047479, unpublished); aadA4 (formerly aadA3, Z50802, unpublished); aadA5 (AF169041, unpublished).
with a Steers multiple inoculator and an inoculum of 1U104 CFU per spot [17]. For E. coli strains harbouring pAADA6 and pLPO-2 harbouring aadA2 from S. typhimurium [18], MICs for spectinomycin and streptomycin were s 2000 Wg/ml. The MIC for gentamicin was lower than 4 Wg/ml for both recombinant E. coli strains. Analysis of the genetic environment of aadA6 revealed key signatures of gene cassettes. The aadA6 gene cassette has a recombinational core site `GTTA-
GAC' (Fig. 1B,C) and, within the 3P-end of the aadA6 coding sequence, an inverse core site `GTCTAAC' followed by the remainder of a 59-be (Fig. 1B). The aadA6 gene cassette is 861 bp long and its 59-be is 60 bp long. Comparison of the 59be sequences from various aadA genes indicate a strong conservation in length and in sequence (Fig. 1C). The aadA6 59-be has 4 and 8 nucleotide di¡erences with those of aadA1 and aadA2, respectively. A second gene cassette, orfD, which codes for a poly-
BBAEXP 93344 1-12-99
448
T. Naas et al. / Biochimica et Biophysica Acta 1489 (1999) 445^451
peptide of unknown function, was identi¢ed to have the same orientation as aadA6 (Fig. 1B). This orfD cassette was identical to previously identi¢ed orfD sequences found in the sequence of the integron insert of In1, which is located on plasmid R46 [19]. Interestingly, the sequence between the aadA6 and orfD gene cassettes was relatively short and both cassettes contain their inverse core site within the 3P-end of their coding sequence. In1 contains also an aadA1 gene but is separated from orfD by an oxa2 cassette [19]. So far, orfD was found only to be associated with oxa2 gene cassettes. This is the ¢rst description of an aadA gene located immediately next to an orfD. The DNA sequences of aadA1 [3] and aadA2 [7] genes share 88% identity, and the cassettes have distinguishable 59-bes and 5P core sites (GTTAAAC for aadA1 and GTTAGAC for aadA2) (Fig. 1C). AadA3 is a hybrid gene cassette because the coding sequence is related to aadA2 while its 59-be resembles the one of aadA1 (unpublished, GenBank accession number AF047479). Aad5 has a high degree of sequence divergence with the other aadA genes and its 59-be is signi¢cantly di¡erent from the previous 59-bes (Fig. 1C) (unpublished, GenBank accession number AF169041). AadA6 gene shares at most 70% DNA sequence identity with the other aadA genes (aadA1), and its 59-be, even though di¡erent, belongs to the same sub-family of 59-bes. Thus aadA6 represents a novel version of gene cassette that has probably diverged from the other members of this family earlier in evolution. Multiple sequence alignments of AADA6 with six AADA enzymes revealed signi¢cant sequence homol-
ogy (Fig. 2). These enzymes vary very little in length (263 or 262 aa) except for AADA6, which contains 19 additional amino acids. This extension of the Cterminal end of AADA6 is likely due to a single nucleotide insertion at the very end of the gene resulting in a frameshift (data not shown). Sequencing of various independent PCR products con¢rmed this C-terminal extension, thus indicating that the C-terminus can accommodate changes in length without a¡ecting noticeably the activity of the enzyme. The exact translation start of aadA gene cassettes still remains unclear. In fact, two putative translation start sites may be used in most cassettes. The ¢rst one corresponds to an `ATG' codon that is present immediately after the cassette recombination core site, while the second one, a `GTG' codon, is 12 bp downstream of the `ATG'. DNA sequence analysis for putative ribosomal binding sites immediately upstream of the two codons revealed, in both cases, purine-rich stretches of DNA. Therefore, both may serve as a translation start site in bacteria and only N-terminal sequencing of the aadA gene products may be able to discriminate between them. However, for numbering the amino acids of AADA enzymes, the ¢rst possible translation start codon, the `ATG' codon, will be used (Fig. 2). The AADA6 enzyme shares 75% amino acid identity with AADA1, AADA2, AADA3, and about 60% with AADA4, AADA5 and AADAsc (Table 2). AADA6 protein is therefore a novel type of AADA enzyme. Sequence comparison between the di¡erent AADA proteins reveals blocks of conservation, which may be good
Table 1 Primers used in the PCR reactionsa Primers
Numbering as in Fig. 1A
Sequence (5P to 3P)
5PCS 3PCS AADA6-1 AADA6-2 INT1-B INT1-F SUL1-B
1 2 3 4 5 6 7
ggc aag ttg cag cag tcg gca
atc cag tcc aga cgc gtc agg
caa act aag att atc tcc cgg
gca tga ata tct aag acg aaa
Accession number gca cct agc tta cgg cat ccc
agc gat ctg acc tga cg gcg
U12338 U12338 cc AF140629 aag AF140629 gc AF133699 AF133699 cc AF133699
a
Position of primer
Ref.
1416^1433 4831^4814 762^743 509^529 27^46 794^778 6271^6252
[14] [14] This work This work [15] [15] [15]
Genomic DNA extraction of P. aeruginosa JES and PCR reactions were as described [15,24]. The 5P-CS and the 3P-CS primers were used, in order to determine the integron content. INT1-B and INT1-F were used to amplify a 767-bp intragenic fragment of intI1. INT1-B with AADA6-1 primers were used to demonstrate the collinearity of the integrase with the resistance genes and to sequence the promoter region of the cassette genes, while SUL1-B primers together with AADA6-2 primer were used to analyse the genetic content of the 3P-CS.
BBAEXP 93344 1-12-99
T. Naas et al. / Biochimica et Biophysica Acta 1489 (1999) 445^451
449
Fig. 2. Protein sequence alignment. The amino acid sequence of AADA6 was aligned with those of six AADA enzymes. All the sequences were compared to AADA1 taken as a reference AADA enzyme [3]. Dashes represent sequence identities and di¡erences are represented by the appropriate amino acid. Numbering starts with the ¢rst methionine immediately after the recombination core site `GTTARRC' for aadA gene cassettes. Stars below the alignment indicate perfect sequence conservation. The references of the enzymes included in the alignment are listed in the legend of Fig. 1C except for AADAsc (GenBank accession number X68089 [9]).
candidates for structure^function studies of this type of enzyme (Fig. 2). No promoter sequences were found within the gene cassettes, indicating that the intI1 promoters are likely to be responsible for cassette gene expres-
sion, as for most of the gene cassettes. Sequencing of the PCR-ampli¢ed products using intI1 and aadA6 speci¢c primers revealed the presence of an intact integrase gene in the 5P-CS of In51. In addition, analysis of the promoter region located within the inte-
BBAEXP 93344 1-12-99
450
T. Naas et al. / Biochimica et Biophysica Acta 1489 (1999) 445^451
Table 2 Percentage of amino acid identities between the di¡erent AADA enzymes AADA1 AADA1 AADA2 AADA3 AADA4 AADA5 AADA6 AADAsc
100
AADA2 86 100
AADA3 87 98 100
grase gene intI1, revealed that the gene cassettes are likely to be expressed from a weak P1 promoter (335; TGGACA and 310; TAAGCT) and that the accessory P2 promoter, found in some class 1 integrons, is in its inactive form [8]. Similarly, PCR ampli¢cation using sulI- and aadA6-speci¢c primers indicated that the qacEv1 and sulI genes were present in the 3P-CS of the integron (Table 1). Plasmid DNA extractions from P. aeruginosa JES failed despite repeated attempts using four di¡erent extraction methods [17]. Furthermore, electroporation of ¢ve independent plasmid DNA preparations failed to produce streptomycin or spectinomycin-resistant E. coli clones. In addition, PFGE of unrestricted genomic DNA of P. aeruginosa JES revealed only one fragment corresponding to the chromosomal DNA, indicating the absence of any plasmid and thus, the chromosomal origin of In51 (data not shown). However, lack of plasmid detection in P. aeruginosa JES does not preclude a plasmidic origin. Indeed, plasmids from other species often transfer to P. aeruginosa but then cannot replicate and become chromosomally associated [20]. When a PFGE gel, as previously described [18], containing SpeI and XbaIrestricted P. aeruginosa genomic DNA, was transferred and subsequently hybridised with an intI1 speci¢c probe, three distinct bands of 25, 30 an 35 kb in size were observed with the XbaI-restriction while in the case of the SpeI-restriction, only one band of 55 kb with about three times the intensity of the XbaI bands was seen (data not shown). These results seemed to indicate that a third integrase gene, besides those of In50 and In51, may be present on the chromosome of P. aeruginosa JES and that all the three IntI1 genes may clustered on the same 55-kb SpeIfragment. Clustering of integrons has recently been
AADA4 57 54 54 100
AADA5 56 53 53 95 100
AADA6 75 74 74 59 58 100
AADAsc 43 43 43 43 42 59 100
observed for Salmonella typhimurium DT104, where two distinct integrons, each containing di¡erent gene cassettes, are present on a 14-kb genomic DNA fragment [21]. In50 and In51 may also be located close together on the chromosome of P. aeruginosa JES. The third integrase gene may belong to another integron for which either the 5P-CS and/or 3P-CS primer binding sites are absent or mutated or too far apart from each other to allow its ampli¢cation under our PCR conditions. Indeed, some class 1 integrons are known to lack the 3P-CS [13,22]. P. aeruginosa JES may therefore contain at least two integrons In50 and In51, which both seem to be chromosomally-located. Both may have originated from Enterobacteriaceae and may have been transferred through plasmids and/or transposons to P. aeruginosa JES. In addition, both integrons carry novel resistance determinants (In50: blaVEBÿ1 and In51: aadA6) underlining the ongoing evolution of resistance genes in di¡erent geographic areas. Moreover, comparison of the amino acid sequence of the AADA enzymes described so far indicate that they, like other resistance genes, may present sequence variability with conserved enzymatic functions. The knowledge of di¡erent AADA sequences may give some insights into the structure^function relationship of this class of enzymes. This work was funded by the Universite¨ Paris XI, Faculte¨ de Me¨decine Paris-Sud, France (UPRES, JE 2227).
References [1] K.J. Shaw, P.N. Rather, R.S. Hare, G.H. Miller, Molecular genetics of aminoglycoside resistance genes and familial re-
BBAEXP 93344 1-12-99
T. Naas et al. / Biochimica et Biophysica Acta 1489 (1999) 445^451
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
lationships of the aminoglycoside-modifying enzymes, Microbiol. Rev. 57 (1993) 138^163. R.M. Hall, C.M. Collis, Mobile gene cassettes and integrons: capture and spread of genes by site speci¢c recombination, Mol. Microbiol. 15 (1995) 593^600. S. Hollingshead, D. Vapnek, Nucleotide sequence analysis of a gene encoding a streptomycin/spectinomycin adenylyltransferase, Plasmid 13 (1985) 17^20. L. Sundstroem, P. Radstroem, G. Swedberg, O. Skoel, Sitespeci¢c recombination promotes linkage between trimethoprim and sulfonamide resistance genes. Sequence characterization of dhfrV, sulI and a recombination active locus of Tn21, Mol. Gen. Genet. 213 (1988) 191^201. M.E. Fling, J. Kopf, C. Richards, Nucleotide sequence of the transposon Tn7 gene encoding an aminoglycoside-modifying enzyme, 3Q(9)-O-nucleotidyltransferase, Nucleic Acids Res. 13 (1985) 7095^7106. R.C. Tait, H. Rempel, R.L. Rodriguez, C.I. Kado, The aminoglycoside-resistance operon of the plasmid pSA: nucleotide sequence of the streptomycin-spectinomycin resistance gene, Gene 36 (1985) 97^104. A. Bito, M. Susani, Revised analysis of aadA2 gene of plasmid pSA, Antimicrob. Agents Chemother. 38 (1997) 1172^ 1175. P. Martinez-Freijo, A.C. Fluit, F.J. Schmitz, J. Verhoef, M.E. Jones, Many class I integrons comprise distinct stable structures occurring in di¡erent species of Enterobacteriaceae isolated from widespread geographic regions in Europe, Antimicrob. Agents Chemother. 43 (1999) 686^689. K.Y. Leung, S.R. Ruschkowski, B.B. Finlay, Isolation and characterization of the aadA aminoglycoside-resistance gene from Salmonella choleraesuis, Mol. Microbiol. 6 (1992) 2453^2460. J. Nesvera, J. Hochmannova, M. Patek, An integron of class 1 is present on the plasmid pCG4 from gram-positive bacterium Corynebacterium glutamicum, FEMS Microbiol. Lett. 169 (1998) 391^395. N.C. Clark, O. Olsvik, J.S. Swenson, C.A. Spiegel, F.C. Tenover, Detection of a streptomycin/spectinomycin adenylyltransferase gene (aad) in Enterococcus faecalis, Antimicrob. Agents Chemother. 43 (1999) 157^160. P. Courvalin, C. Carlier, Resistance towards aminoglycoside-aminocyclins antibiotics in bacteria, J. Antimicrob. Chemother. 8 (1981) 57^69. G.D. Recchia, R.M. Hall, Gene cassettes: a new class of mobile elements, Microbiology 141 (1995) 3015^3027.
451
[14] C.M. Collis, R.M. Hall, Site-speci¢c deletion and rearrangement of integron insert genes catalysed by the integron DNA integrase, J. Bacteriol. 174 (1997) 1574^1585. [15] T. Naas, L. Poirel, A. Karim, P. Nordmann, Molecular characterization of In50, a class 1 integron encoding the gene for the extended-spectrum L-lactamase VEB-1 in Pseudomonas aeruginosa, FEMS Microbiol. Lett. 176 (1999) 411^ 419. [16] L. Poirel, T. Naas, M. Guibert, E.B. Chaibi, R. Labia, P. Nordmann, Molecular and biochemical characterization of VEB-1, a novel class A extended-spectrum L-lactamase encoded by an Escherichia coli integron gene, Antimicrob. Agents Chemother. 43 (1999) 573^581. [17] L.N. Philippon, T. Naas, A.T. Bouthors, V. Barakett, P. Nordmann, OXA-18, a class D clavulanic acid-inhibited extended-spectrum L-lactamase from Pseudomonas aeruginosa, Antimicrob. Agents Chemother. 41 (1997) 2188^2195. [18] L. Poirel, M. Guibert, S. Bellais, T. Naas, P. Nordmann, Integron- and Carbenicillinase-mediated reduced susceptibility to amoxicillin-clavulanic acid in isolates of multidrugresistant Salmonella enterica serotype Typhimurium DT104 from French patients, Antimicrob. Agents Chemother. 43 (1999) 1098^1104. [19] H.W. Stokes, R.M. Hall, The integron In1 in plasmid R46 includes two copies of the oxa2 gene cassette, Plasmid 28 (1992) 225^234. [20] G.A. Jacoby, Resistance plasmids of Pseudomonas, in: J.R. Sokatch (Ed.), Bacteria: A Treatise on Structure and Function, vol. X, Academic Press, San Diego, CA, 1986, pp. 265^ 294. [21] C.E. Briggs, P.M. Fratamico, Molecular characterization of the antibiotic gene cluster of Salmonella typhimurium DT104, Antimicrob. Agents Chemother. 43 (1999) 846^849. [22] T. Naas, W. Sougako¡, A. Casetta, P. Nordmann, Molecular characterization of OXA-20, a novel class D L-lactamase and of its integron from Pseudomonas aeruginosa, Antimicrob. Agents Chemother. 42 (1998) 2074^2083. [23] H.W. Stokes, D.B. O'Gorman, G.D. Recchia, M. Parsekhian, R.M. Hall, Structure and function of 59-base element recombination sites associated with mobile gene cassettes, Mol. Microbiol. 26 (1997) 731^745. [24] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning : A Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1989.
BBAEXP 93344 1-12-99
