Previous Article | Next Article ![]()
Antimicrobial Agents and Chemotherapy, May 2002, p. 1402-1409, Vol. 46, No. 5
0066-4804/02/$04.00+0 DOI: 10.1128/AAC.46.5.1402-1409.2002
Copyright © 2002, American Society for Microbiology. All Rights Reserved.
Département de Biochimie et de Microbiologie, Faculté des Sciences et de Génie, Université Laval, Sainte-Foy, Québec, Canada G1K 7P4,1 Centre de Recherche en lnfectiologie, CHUQ Pavillon CHUL, Sainte-Foy, Québec, Canada G1V 4G2,2 Facultad de Medicina de la Universidad de Buenos Aires, 1121 Capital Federal, Argentina3
Received 23 August 2001/ Returned for modification 16 October 2001/ Accepted 3 January 2002
|
|
|---|
|
|
|---|
The origins of the cassette-associated genes are not known (9, 27). It has been proposed that cassettes may have originated from transcripts that were converted to DNA, involving the activity of an as-yet-unidentified reverse transcriptase (RT) (27). The palindromic attC sites (59-base elements) found at the end of each cassette either may have been present within the original transcripts, for example, as transcription terminators, or may have been added at a later stage (27). It has been also speculated that secondary sites for IntI1, defined as the degenerate pentanucleotide GWTMW, can facilitate the association of new genes into integrons and may have been the substrate for the creation of attC sites during evolution (6, 9).
The goal of our work is to elucidate the role of the antimicrobial resistance mechanisms carried on integrons in a multiresistant Serratia marcescens strain. We describe a novel ant(3")-Ii-aac(6')-IId gene cassette, inserted into the variable region of a class I integron, that codes for a bifunctional ANT(3")-I-AAC(6')-II mechanism. The role of the integrase activity in the formation of this fused cassette is discussed. The role of a group II intron located between the ant(2")-Ia resistance gene and its attC site is discussed.
|
|
|---|
Media and culture conditions. Bacterial strains were grown on Luria-Bertani agar and in Luria-Bertani broth supplemented with one or more of the following antibiotics, as appropriate: ampicillin (50 µg/ml), gentamicin (25 µg/ml), nalidixic acid (50 µg/ml), or streptomycin (25 µg/ml).
Antibiotic susceptibility. Determination of the aminoglycoside resistance profile (AGRP) (30) was done by using aminoglycoside disks (Schering-Plough Corporation, Kenilworth, N.J.) (17) on Mueller-Hinton agar. The 11 aminoglycosides included fortimycin, amikacin, tobramycin, apramycin, dibekacin, gentamicin, isepamicin, netilmicin, 5-episisomicin, 2'-N-ethylnetilmicin, and 6'-N-ethylnetilmicin. Assignment of enzymatic mechanisms to each strain was based on the unique pattern of resistance to these aminoglycosides (17). In addition, the MICs in liquid medium were determined with an inoculum of 105 to 106 CFU/ml.
DNA techniques. Preparation of DNA was performed as described by Sambrook et al. (29). Plasmid DNA was prepared by using a Wizard purification kit supplied by Promega. DNA was digested with various restriction enzymes, and the fragments were separated in horizontal gels of 0.8 to 1.0% (wt/vol) agarose dissolved in 0.4 M Tris-acetate-0.01 M EDTA. Gels were stained with ethidium bromide, and the DNA was visualized by UV transillumination at 302 nm. Transformation procedures were performed as described by Sambrook et al. (29). Conjugation was carried out with brain heart infusion agar for 4 h at 37°C, and the transconjugants were selected on Mueller-Hinton agar medium containing ampicillin (50 µg/ml).
PCR amplifications were carried out with 100-µl volumes containing 10 ng of DNA, 10 µl of 10x PCR buffer, 10 µl of 10x deoxynucleoside triphosphate mix (2 mM each dATP, dCTP, dGTP, and dTTP), 10 µl of each primer stock solution (2.5 pmol of each primer per µl), and 60 µl of sterile distilled water. Each reaction mixture was covered with 75 µl of mineral oil. Taq DNA polymerase from Promega was added (1 µl of a 3-U/µl diluted solution) after 12 min at 94°C (hot-start method). To amplify the DNA in the thermocycler (Perkin-Elmer Cetus, Emeryville, Calif.), we used a three-step profile described previously (14). The gene probes and/or primers used for PCRs were specific for the following nucleotide sequences: ß-lactam resistance genes blaCTX-M-2 (ATGACTCAGAGCATTCGC and TCACTTTATCGGGACCAC), blaPER-2 (1), and blaTEM (14); aminoglycoside resistance genes aac(6')-Ib, ant(3")-Ia, aac(3)-Ia, aac(3)-IIa, and ant(2")-Ia (14); trimethoprim resistance gene dfrA1 (14); and the 5' conserved segment (5' CS) and the integrase gene of class I integrons (Sulpro3) (14). DNA was transferred to nylon filters by Southern blotting and hybridized with appropriate probes as described by Sambrook et al. (29). Autoradiography was done with Kodak X-Omat AR film.
The 5.0-kb HindIII fragment of pSm909 containing integron 3 and Tn1696-related sequences, the 2.3-kb BamHI fragment of pSm909 containing the integrase gene, the ant(2")-Ia gene, and part of the group II intron from integron 2, and the 4.2-kb HincII fragment containing the remainder of the group II intron, ant(3")-Ii-aac(6')-IId, open reading frame (ORF) O, blaOXA-10
, IS1, and unknown sequences from pSm909 were cloned into the corresponding sites of pUC19.
DNA sequencing and analysis. DNA was purified by using a QIAquick kit according to the manufacturer's instructions (Qiagen Inc., Studio City, Calif.). Sequencing was done on both DNA strands by using an ABI 373 sequencer. Internal oligonucleotide primers were used when necessary to ensure that both strands were sequenced. The nucleotide sequences were analyzed by using Genetics Computer Group programs.
Nucleotide sequence accession numbers. The nucleotide sequences of the variable regions of integrons 2 and 3 in S. marcescens SCH909 have been deposited in GenBank under accession numbers AF453998 and AF453999, respectively.
|
|
|---|
|
View this table: [in a new window] |
TABLE 1. MICs of antibiotics for S. marcescens SCH909, E. coli C600 transconjugant, and control E. coli ATCC 25922
|
![]() ![]() View larger version (74K): [in a new window] |
FIG. 1. Mapping of integrons of S. marcescens SCH909. (a) General structures of integrons 1, 2, and 3 found in S. marcescens SCH909. The arrows show the direction of transcription. (b) PCR amplification from SCH909. The PCR products were separated by electrophoresis through 1.0% agarose. Lanes 1 and 14, 1-kb DNA ladder; lanes 2 to 13, SCH909 with the following primers: lane 2, Sulpro3 and 3'-CS; lane 3, Sulpro3 and dfrA1; lane 4, Sulpro3 and ant(3")-Ia; lane 5, Sulpro3 and ant(3")-Ii; lane 6, Sulpro3 and aac(6')-Ib; lane 7, Sulpro3 and ant(2")-Ia; lane 8, Sulpro3 and aac(3)-Ia; lane 9, aac(3)-Iacooh and 3'-CS; lane 10, ant(3")-Iacooh and 3'-CS; lane 11, aac(3)-Iacooh and ant(3")-Ia; lane12, ant(2")-Iacooh and aac(6')-Ib; and lane 13, tnpR and BLATEM. Primer Sulpro3 is a rightward primer in the 5' conserved segment. Primers with gene names are leftward primers near the beginning of their respective cassettes. Primers with "cooh" are rightward primers near the end of their respective cassettes. Primer 3'-CS is a leftward primer in the 3' conserved sequence. These primers are used to determine gene order in integrons (14).
|
Analysis of the three variable regions of the integrons. The nucleotide sequences of the 1.5-, 5.2-, and 3.1-kb regions from the variable regions of the three integrons found in SCH909 were determined (see Materials and Methods). Integron 1 contains the dfrA1 and ant(3")-Ia genes, and its variable region is identical to that of pLMO229 (32).
Integron 2 (GenBank accession no. AF453998) contains an ant(2")-Ia gentamicin resistance gene which is separated from its attC site by a 1,971-bp insert containing an ORF on the complementary strand (Fig. 1a). While the ant(2")-Ia gene was identical, the ant(2")-Ia attC site possessed only 84.5% identity with that of the ant(2")-Ia gene already described (4). The complementary strand of the insert showed 48.1% identity with a group II intron from Pseudomonas alcaligenes (36). The product of this ORF, called Smtr, from the bottom strand showed 39.2% identity with the maturase of this intron, although the ORF length for the putative maturase and/or RT differed. The putative RT from SCH909 possessed all seven polymerase-like domains characteristic of RT activity (Fig. 2). In addition to the seven RT-like domains and like the maturase from P. alcaligenes (intron xln6), a domain, z, which is characteristic of non-long terminal repeat retroelements (19), was detected at the N-terminal end of the RT from SCH909. Also, domain x was found downstream of the seven RT domains in Smtr. Domain x has been suggested to be involved in binding of the intron RNA during reverse transcription and splicing. A noncoding region immediately downstream of the putative RT could fold into typical V and VI RNA secondary structures (Fig. 3). 3 Domain V is the only component other than domain I that is absolutely essential for group II ribozyme activity (16). A 1,971-bp sequence (from bp 534 to bp 2504 of integron 2) could fold into the typical RNA secondary structure of a central wheel with six spokes that define the six major ribozyme domains of group II introns (data not shown).
![]() View larger version (59K): [in a new window] |
FIG. 2. Protein alignment of the putative maturase, Smtr, in SCH909 with eubacterial group II intron-encoded proteins from an Acinetobacter sp. (Acin.sp.) (J. H. Yum et al., unpublished; GenBank accession number AF369871; ORF II), N. europaea (N. euro) (JGI, unpublished), P. alcaligenes (PaOrX6) (36), Streptococcus pneumoniae (S. pneu) (5), and Lactococcus lactis (LlltrA) (18). Consensus amino acids (four or more identical in the six sequences) are in bold letters. Domains conserved among intron-encoded proteins are denoted by the lines above the alignment. RT1 through RT7, RT-like domains (19); z, domain of undetermined function in non-long terminal repeat retroelements; x, maturase-specific domain (19). Numbers in parentheses indicate numbers of amino acids not shown.
|
![]() ![]() View larger version (27K): [in a new window] |
FIG. 3. General features of Smtr and its insertion site. (a) Secondary structure model of domains V and VI of the group II intron, Smtr, within integron 2 from SCH909. The boundary of the intron and the ant(2")-Ia attC site is indicated by a vertical line. The bulging adenine residue involved in lariat formation is indicated by an asterisk. (b) Nucleotide sequence alignment of the ant(2")-Ia stop codon and its attC site (59-base element) (4) with the ant(2")-Ia gene and Smtr-intron boundary from SCH909). ant(2")-Ia stop codons are indicated by asterisks. Exon sequences are shown in bold letters. (c) Nucleotide sequence alignment of the group II introns of SCH909 and N. europaea. The latter is adjacent to an attC site on one end; there is no recognizable ORF on the other end.
|
Integron 3 (GenBank accession no. AF453999) contains another aminoglycoside resistance gene cassette, aac(3)-Ia, followed by two identical ORF P cassettes, a different ORF Q cassette, and ant(3")-Ia (Fig. 1). The sequence to the left of the integrase of integron 3 shows that this integron is part of a Tn1696-like transposon (data not shown). ORF P and ORF Q are identical to ORF X(A) and ORF X(B) of Klebsiella oxytoca plasmid pACM1 (24), which has only one copy of the former.
Sequencing of the 5' conserved segments of integrons 1 and 2 revealed integrase genes identical to that of Tn21 (13), while that of integron 3 is like that of Tn1696. Unlike the situation in Tn1696, the 19-bp duplication which permits the expression of the aac(3)-Ia gene as a translational fusion which begins before the attI site (35) is absent. In SCH909, the aac(3)-Ia gene probably uses an initiation codon internal to the cassette.
Characterization of a new aminoglycoside resistance gene, the ant(3")-Ii-aac(6')-IId gene cassette. Plasmid DNA from strain SCH909 was digested with HincII. The fragments were ligated to the HincII site of pTZ19r and introduced by transformation into E. coli NM522. The transformants selected on ampicillin-gentamicin were screened for inserts by agarose gel electrophoresis, and a plasmid containing an insert of 4.2 kb was chosen (pGM172). The fragment from a Sau3AI partial digest was subcloned into the BamHI site of pTZ19r to yield plasmid pGM172-7, which contained a 2.3-kb insert that conferred resistance to gentamicin, streptomycin, and spectinomycin on E. coli NM522. The AGRP of pGM172 and pGM172-7 revealed the combination of ANT(3")-I plus AAC(6')-II mechanisms (Tables 2 and 3). We found an ORF spanning 1,392 nucleotides from the start codon ATG at position 2563 to the stop codon TAA at positions 3952 to 3954 in the sequence of integron 2 (GenBank accession number AF453998). The first 789 nucleotides of this ORF possessed 99.2 and 82.7% identities, respectively, with the novel ant(3")-If (aadA6) (22) and ant(3")-Ig (aadA7) (15) over the entire length and 76.4 and 77.1% identities, respectively, with ant(3")-Ia and ant(3")-Ib over a 785-bp length. There is no stop codon at the position corresponding to the ant(3")-Ia and ant(3")-Ib stop codons. We called this part of this ORF ant(3")-Ii (aadA9).
|
View this table: [in a new window] |
TABLE 2. Disk susceptibility of host strain, clones, and S. marcescens parent straina
|
|
View this table: [in a new window] |
TABLE 3. Susceptibility to streptomycin and spectinomycina
|
Downstream of the ant(3")-Ii sequence, there is a sequence from positions 3457 to 3954 constituting a gene that was previously called aac(6')-Ib' but that we renamed (because of the AGRP) aac(6')-IId, since the AGRP of the protein is that of an AAC(6')-II, i.e., amikacin sensitive and gentamicin resistant. The ant(3")-Ii-aac(6')-IId gene confers resistance to streptomycin, spectinomycin, and gentamicin, while the cloned ant(3")-Ii sequence confers resistance only to streptomycin and spectinomycin. Nucleotide sequence determination indicated total identity between the aac(6')-IId sequence from SCH909 and the aac(6')-Ib' sequence from Pseudomonas aeruginosa BM2687 (12), which confers resistance to gentamicin but not to amikacin. Like other aac(6')-Ib' genes, the aac(6')-IId sequence has a T-to-C transition that results in a leucine-to-serine substitution at position 90. This point mutation is responsible for the altered substrate specificity (25).
The ant(3")-Ii-aac(6')-IId gene cassette extends from positions 2563 to 3954. At the 3' end of the ant(3")-Ii-aac(6')-IId gene, there is a typical attC site that has a 74-nucleotide sequence (positions 3949 to 4022) with an approximate 20-base similarity at each end that is related to the attC site (59-base element) consensus sequence (31). This 74-bp sequence showed 91.9% identity with the aac(6')-Ib' attC site from P. aeruginosa BM2687 (12) and 89.2% identity with the attC site of aac(6')-Ib from Tn1331 (33).
|
|
|---|
At least two hypotheses can be proposed for the origin of the fused protein: the bifunctional protein could have been created by cassette fusion, or it could have been encoded by an ancestral aminoglycoside resistance gene from which several others evolved. The recent finding of the ant(3")-If (aadA6) gene (22) favors the former possibility, since translation of this gene continues well into its attC site. Indeed, the part of the ant(3")-Ii structural gene before the (missing) attC site differs from ant(3")-If by only six nucleotides. The close relationship of ant(3")-Ia and ant(3")-Ib suggested that they have evolved either from one another or from a common ancestral gene after the formation of a cassette, since their attC sites are also very similar. However, the attC site of ant(3")-If is also very similar to those of ant(3")-Ia and ant(3")-Ib, but the ca. 23% difference in the structural genes points to separate events of cassette formation rather than evolution after acquisition of an attC site by an ancestral gene. The sequence downstream of ant(3")-Ii corresponds to the aac(6')-IId gene, which has total identity with the aac(6')-Ib' gene from P. aeruginosa BM2687 (12). However, their attC sites are not identical (91.9% identity), again raising the possibility of independent events of cassette formation. The spacer sequence between the presumed mutation removing the ant(3")-Ii stop codon (positions 3352 to 3354) and a possible internal aac(6')-IId start codon (position 3457) includes the sequence CTAAAACAAA, identical to 10 bases at the end of the attI site from the 5' conserved segment of the integrons.
The fused cassette could have been created by an illegitimate integrase-mediated intermolecular recombination event involving the inverted core site of the attC site of ant(3")-Ii (CTAA'CAATT) and the end of an attI site (CTAA'AACAAAGTT) preceding aac(6')-IId. An identical event might have been responsible for the creation of the fused ant(3")-Ia-blaOXA-9 cassette in Tn1331 (33). The crossing-over point in attI would correspond to a seven-nucleotide stagger with G'TT and cutting 3' of two A residues, as seen with XerC-XerD recombination and as proposed for the integron integrase. A similar intramolecular excision event involving attI and the inverted core site of an attC site has been observed (F. Gagnon and P. H. Roy, unpublished results). The gene aac(6')-IId is probably the ancestral gentamicin resistance gene from which the amikacin resistance gene aac(6')-Ib emerged by the mutation resulting in Ser119Leu. aac(6')-Ib often uses protein fusions for optimal expression: fusion with TEM-1 in Tn1331 (33), fusion with the catB4
product in Serratia sp. strain 45 (34), and fusion with the 5' conserved sequence in pCFF04, which can use a 19-bp repeat in a manner identical to that used for the expression of aac(3)-Ia in Tn1696 (35).
An alternate explanation for the origin of the bifunctional protein is that ant(3")-Ii-aac(6')-IId was an ancestral aminoglycoside gene from which aac(6')-Ib, aac(6')-Ib', aac(6')-IIa, aac(6')-IIb, ant(3")-Ia, and ant(3")-Ib evolved by deletion and cassette formation due to selective pressure. There is another case of a bifunctional enzyme described in the aminoglycoside resistance protein family, that encoded by aac(6')-Ie-aph(2") and confined to gram-positive species.
In integron 2, the ant(2")-Ia gene is separated from its attC site by a group II intron that we call Smtr. At least two possible routes that would lead to a group II intron within a cassette can be envisaged. The simplest explanation is a direct insertion of Smtr, mediated by one of the two retrotransposition pathways, retrohoming or ectopic insertion (8, 16, 18, 19), at the inverted core site at the junction of ant(2")-Ia and its attC site. The SCH909 group II intron Smtr begins with GTACG and ends with GAT, in accord with consensus sequences for the ends of group II introns. The target DNA needed for reverse splicing to occur in Ll.LtrB, the only bacterial group II intron that has been shown to be functional for splicing and mobility in vivo, is rather specific and contains the target site TG'GTTA (20). In contrast, although the complete recognition site of group II intron xln6 from P. alcaligenes has not been determined, it contains TTGT'TA (36). The SCH909 group II intron Smtr forms a subgroup with xln6 and also with a group II intron found in 14 copies in genomic DNA of Pseudomonas putida, with a site specificity of TTTTTGT'T. The complement of the consensus inverted core site of the attC site (59-base element) is AATTGT'TAGGC, where the prime corresponds to the point of intron insertion. Notably, in the integron context, the group II intron is "upside down," so that the attC site is exon 1 while the ant(2")-Ia structural gene is exon 2. It is unknown what effect transcription arriving from exon 2, in contrast to the situation in Ll.LtrB, where external transcription occurs from exon 1, has on intron functions.
Taking into account the special features of mobility of group II introns, splicing (to remove the intron), retrohoming (in which RNA invades double-stranded DNA), and ectopic insertion (in which RNA invades RNA) (5a, 8, 16, 18, 19), an alternative explanation for the presence of Smtr could involve a role in one or more steps in the formation of cassettes from preexisting structural genes and attC recombination sites. It was proposed that cassettes could be formed by reverse transcription (10) mediated by RT activity (27), although no role for introns was mentioned by these authors. A possible mechanism for cassette formation would involve two independent transpositional events placing one intron immediately downstream of a structural gene (such as a resistance gene) and another adjacent to an attC site. The next event would involve recombination, possibly recA mediated, between the two introns. The next step would involve splicing out of the intron, followed by final reverse transcription to conserve the newly formed cassette at the DNA level. An interesting example of one of these possible intermediates was found in the partial genome sequence of Nitrosomonas europaea (Joint Genome Institute [JGI], unpublished data), in which one copy of a novel group II intron which we call Netr is precisely inserted at the inverted core site of an attC site, in the same manner as in SCH909 (Fig. 3c). The other end of this element is not associated with the stop codon of a structural gene; thus, it may represent the intron-attC intermediate. A second copy of Netr is adjacent to a putative gene cassette, and there is no attC site at its other end; thus, it may be an example of a gene-intron intermediate. The target specificity of the intron, as mentioned above, may explain the precise juxtaposition in many cassettes in integrons of the structural gene stop codon and the inverted core site of the attC site. In SCH909, the ant(2")-Ia gene is identical to that of pDGO100 (4),whereas the ant(2")-Ia attC site shows only 84.5% identity with ant(2")-Ia attC sites already described (3). The SCH909 sequence may represent an unspliced intermediate in an independent event of ant(2")-Ia cassette formation.
Recently, three integron sequences submitted to GenBank and containing a gene for ß-lactamase VIM-2 from Acinetobacter spp. (J. H. Yum et al., unpublished data; accession number AF369871), P. aeruginosa (K. Lee et al., unpublished data; accession number AY029772), and S. marcescens (K. Lee et al., unpublished data; accession number AY030343) indicated the presence of a group II intron (identical among these three but distinct from that of SCH909) between ant(3")-Ia and its attC site (in the first two) or between the qacF gene and an attC site identical to that of ant(3")-Ia rather than that of qacF (in the third). These may be further examples of independent events of cassette formation.
A few scattered attC sites (59-base elements), apparently not associated with structural genes and not making up parts of integrons, are found in some partially sequenced genomes, such as those of Shewanella putrefaciens (The Institute for Genomic Research, unpublished data) and N. europaea (JGI, unpublished). These genomes also contain functional integron integrases (70; G. Léon and P. H. Roy, unpublished data), and closely related integrases have been found in other strains of these species (28). These environmental organisms may serve as reservoirs of integron components.
Group II introns have been described, until now, as being inserted close to mobile elements in eubacteria (11, 21). Moreover, as mobile elements by themselves and located, in the case of SCH909, within antimicrobial resistance gene cassettes, they contribute to DNA rearrangements leading to R plasmid evolution. The elucidation of the mechanism of the formation of integron cassettes remains a key to the understanding of their relationship with the group II introns and their role in aminoglycoside resistance gene evolution and in the accumulation of resistance genes by integrons. We are attempting to determine whether the intron RNA can be spliced to yield a template in which cassette formation could be completed by reverse transcription. If splicing occurs, intron mobility will be tested to determine whether it can be targeted to an attC site within a cassette, e.g., ant(2")-Ia (retrohoming), to an attC site not associated with a cassette, and to the region of the stop codon of a resistance gene. In N. europaea, the two copies of a group II intron are in the latter two contexts, which may be earlier intermediates in cassette formation. Recombination between them would produce a putative later intermediate like that in SCH909, and this process is also being attempted.
This work was supported by grant MT-13564 from the Medical Research Council of Canada to P.H.R. D.C. was a CONICET postdoctoral fellow from Argentina.
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»