Site-specific insertion of genes into integrons: role of the 59-base element and determination of the recombination cross-over point

From examination of published DNA sequences of genes found inserted at a specific site in integrons, all genes are shown to be associated, at their 3' ends, with a short imperfect inverted repeat sequence, a 59-base element or relative of this element. The similarity of the arrangement of gene inserts in the integron and in the Tn7 transposon family is described. A refined consensus for the 59-base element is reported. Members of this family are highly diverged and the relationship of a group of longer elements to the 59-base elements is demonstrated. The ability of 59-base elements of different length and sequence to act as sites for recombination catalysed by the integron-encoded DNA integrase is demonstrated, confirming that elements of this family have a common function. The ability of elements located between gene pairs to act as recombination sites has also been demonstrated. The recombination cross-over point has been localized to the GTT triplet which is conserved in the core sites, GTTRRRY, found at the 3' end of 59-base elements. Recombination at the core site found in inverse orientation at the 5' end of the 59-base elements was not detected, and the sequences responsible for orientation of the recombination event appear to reside within the 59-base element. A model for site-specific insertion of genes into integrons and Tn7-like transposons is proposed. Circular units consisting of a gene associated with a 59-base element are inserted into an ancestral element which contains neither a gene nor a 59-base element.(ABSTRACT TRUNCATED AT 250 WORDS)     

Integrons: Novel DNA elements which capture genes by site-specific recombination

Integrons are unusual DNA elements which include a gene encoding a site-specific DNA recombinase, a DNA integrase, and an adjacent site at which a wide variety of antibiotic resistance and other genes are found as inserts. One or more genes can be found in the insert region, but each gene is part of an independent gene cassette. The inserted genes are expressed from a promoter in the conserved sequences located 5 to the genes, and integrons are thus natural expression vectors. A model for gene insertion in which circular gene cassettes are inserted individually via a single site-specific recombination event has been proposed and verified experimentally. The gene cassettes include a gene coding region and, at the 3 end of the gene an imperfect inverted repeat, a 59-base element. The 59-base elements are a diverse family of elements which function as sites recognized by the DNA integrase. Site-specific insertion of individual genes thus represents a further mechanism which contributes to the evolution of the genomes of Gram-negative bacteria and their plasmids and transposons.Members of the most studied class of integrons, which include thesulI gene in the conserved sequences, are believed to be mobile DNA elements on the basis that they are found in many independent locations, and a discrete boundary is found at the outer end of the 5-conserved segment. However, the length of the 3-conserved segment is variable in the integrons examined to date, and it is likely that this variability has arisen as the result of insertion and deletion events. Though the true extent of the 3-conserved segment remains to be determined, it seems likely that these integrons are mobile DNA elements. The second known class of integrons comprises members of the Tn7 transposon family.     

Genetic elements involved in Tn21 site-specific integration, a novel mechanism for the dissemination of antibiotic resistance genes.

Transposon Tn21 codes for a site-specific integration system, which is probably a novel recombination mechanism, responsible for the acquisition of resistance genes in this widespread family of transposons. Using insertion and deletion mutagenesis we have identified the genetic loci of the various recombination hot-spots (RHS) and of the gene product (the integrase) that catalyses the reaction. The site of recombination has been localized in two of the RHSs to the DNA sequence GTTAG, which is present at the 3' termini of a loosely conserved palindromic sequence of approximately 59 bp. This 59 bp sequence, which flanks the inserted genes in a number of naturally occurring transposons, is the only element required in cis for the recombination reaction.     

The structure of a partial duplication in the integron of plasmid pDGO100

A family of novel potentially mobile DNA elements called integrons, has recently been described (H. W. Stokes and R. M. Hall, 1989, Mol. Microbiol. 3, 1669-1683). The integrons present in the plasmids pDGO100 and pSa are unusual in that they include a duplication of the sulI gene which is located in one of the two conserved segments that make up these elements. In order to define the nature of the duplication in pDGO100, we have sequenced the sulI gene region located between the aadB and the dhfr genes of pDGO100. This region includes the first 1355 bases of the 2026-base 3'-conserved segment present in the integrons of Tn21, R46 and R388, and the sequence identity in pDGO100 ceases 24 bases beyond the end of the sulI gene. This position corresponds to the center of a 59-base element, a remnant of which is located at the end of sulI. This finding suggests that the 59-base element may have been involved in the event which gave rise to the partial duplication.     

Structural and functional characterization of tnpI, a recombinase locus in Tn21 and related beta-lactamase transposons

A novel discrete mobile DNA element from Tn21 from the plasmid R100.1 is described, and its mobilization function was confirmed experimentally. In addition, the element behaves as a recombinase-active locus (tnpI) which facilitates insertions of antibiotic resistance genes as modules or cassettes at defined hot spots or integration sites. A similar tnpI sequence was detected by DNA hybridization in a series of beta-lactamase transposons and plasmids and localized on their physical maps. The genetic function of the locus cloned from Tn21 into pACYC184 was tested for conduction and integration into the plasmids R388 and pOX38Km, and the results suggested recombinase-integrase activity and recA independence. DNA sequence analysis of the tnpI locus revealed no inverted or direct terminal repeats or transposition features of class I and class II transposons. The coding capacity revealed three putative open reading frames encoding 131, 134, and 337 amino acids. Orf3 encoded a putative polypeptide product of 337 amino acids that shared highly significant identity with the carboxyl region of integrase proteins. A comparison and an alignment of the tnpI locus from Tn21 and its flanking sequences identified similar sequences in plasmids and in transposons. The alignment revealed discrete nucleotide changes in these tnpI-like loci and a conserved 3' and 5' GTTA/G hot spot as a duplicated target site. Our data confirm the remarkable ubiquity of tnpI associated with antibiotic resistance genes. We present a model of transposon modular evolution into more complex multiresistant units via tnpI and site-specific insertions, deletions, and DNA rearrangements at this locus.     

Integrons found in different locations have identical 5' ends but variable 3' ends.

The positions of the outer boundaries of the 5'- and 3'-conserved segment sequences of integrons found at several different locations have been determined. The position of the 5' end of the 5'-conserved segment is the same for six independently located integrons, In1 (R46), In2 (Tn21), In3 (R388), In4 (Tn1696), In5 (pSCH884), and In0 (pVS1). However, the extent of the 3'-conserved segment differs in each integron. The sequences of In2 and In0 diverge first from the conserved sequence, and their divergence point corresponds to the 3'-conserved segment endpoint defined previously (H.W. Stokes and R.M. Hall, Mol. Microbiol. 3:1669-1683, 1989), which now represents the endpoint of a 359-base deletion in In0 and In2. The sequence identity in In3, In1, In4, and In5 extends beyond this point, but each sequence diverges from the conserved sequence at a different point within a short region. Insertions of IS6100 were identified adjacent to the end of the conserved region in In1 and 123 bases beyond the divergence point of In4. These 123 bases are identical to the sequence found at the mer end of the 11.2-kb insertion in Tn21 but are inverted. In5 and In0 are bounded by the same 25-base inverted repeat that bounds the 11.2-kb insert in Tn21, and this insert now corresponds to In2. However, while In0, In2, and In5 have features characteristic of transposable elements, differences in the structures of these three integrons and the absence of evidence of mobility currently preclude the identification of all of the sequences associated with a functional transposon of this type.     

Characterization of In0 of Pseudomonas aeruginosa plasmid pVS1, an ancestor of integrons of multiresistance plasmids and transposons of gram-negative bacteria.

Many multiresistance plasmids and transposons of gram-negative bacteria carry related DNA elements that appear to have evolved from a common ancestor by site-specific integration of discrete cassettes containing antibiotic resistance genes or sequences of unknown function. The site of integration is flanked by conserved segments coding for an integraselike protein and for sulfonamide resistance, respectively. These segments, together with the antibiotic resistance genes between them, have been termed integrons (H. W. Stokes and R. M. Hall, Mol. Microbiol. 3:1669-1683, 1989). We report here the characterization of an integron, In0, from Pseudomonas aeruginosa plasmid pVS1, which has an unoccupied integration site and hence may be an ancestor of more complex integrons. Codon usage of the integrase (int) and sulfonamide resistance (sul1) genes carried by this integron suggests a common origin. This contrasts with the codon usage of other antibiotic resistance genes that were presumably integrated later as cassettes during the evolution and spread of these DNA elements. We propose evolutionary schemes for (i) the genesis of the integrons by the site-specific integration of antibiotic resistance genes and (ii) the evolution of the integrons of multiresistance plasmids and transposons, in relation to the evolution of transposons related to Tn21.     

The region of the IncN plasmid R46 coding for resistance to beta-lactam antibiotics, streptomycin/spectinomycin and sulphonamides is closely related to antibiotic resistance segments found in IncW plasmids and in Tn21-like transposons.

The nucleotide sequence of a 2.5 kb segment of the pKM101 (R46) genome has been determined. The 1.3 kb from a BamHI site at 153 to base 1440 differs by only 2 bases from a part of the published sequence of the aadB (gentamicin resistance) gene region including the coding region for the N-terminal 70 amino acids of the predicted aadB product. The same sequence has been found 5'-to the dhfrII gene of R388 and to the aadA gene of Tn21 (R538-1). Three open reading frames are located in this region, two on the same strand as the resistance genes and one on the complementary strand. The latter predicts a polypeptide of 337 amino acids, whose N-terminal segment is 40% homologous to the predicted product of an open reading frame of 179 amino acids located next to the dhfrI gene of Tn7. The oxa2 (oxacillin resistance) gene predicts a long polypeptide commencing with (the N-terminal) 70 amino acids of the aadB product. A similar arrangement is found in the aadA gene of R538-1. The N-terminal segment of an aadA gene is located 3'- to oxa2, separated by 36 bases. Sequences surrounding the BamHI site are identical to sequences 5'- to the tnpM gene of Tn21 and homology ceases where homology between Tn21 and Tn501 commences. The possibility that this antibiotic resistance segment is a discrete mobile DNA element is discussed.     

Occurrence of Tn4371-related mobile elements and sequences in (chloro)biphenyl-degrading bacteria

Tn4371, a 55-kb transposable element involved in the degradation and biphenyl or 4-chlorobiphenyl identified in Ralstonia eutropha A5, displays a modular structure including a phage-like integrase gene (int), a Pseudomonas-like (chloro)biphenyl catabolic gene cluster (bph), and RP4- and Ti-plasmid-like transfer genes (trb) (C. Merlin, D. Springael, and A. Toussaint, Plasmid 41:40-54, 1999). Southern blot hybridization was used to examine the presence of different regions of Tn4371 in a collection of (chloro)biphenyl-degrading bacteria originating from different habitats and belonging to different bacterial genera. Tn4371-related sequences were never detected on endogenous plasmids. Although the gene probes containing only bph sequences hybridized to genomic DNA from most strains tested, a limited selection of strains, all beta-proteobacteria, displayed hybridization patterns similar to the Tn4371 bph cluster. Homology between Tn4371 and DNA of two of those strains, originating from the same area as strain A5, extended outside the catabolic genes and covered the putative transfer region of Tn4371. On the other hand, none of the (chloro)biphenyl degraders hybridized with the outer left part of Tn4371 containing the int gene. The bph catabolic determinant of the two strains displaying homology to the Tn4371 transfer genes and a third strain isolated from the A5 area could be mobilized to a R. eutropha recipient, after insertion into an endogenous or introduced IncP1 plasmid. The mobilized DNA of those strains included all Tn4371 homologous sequences previously identified in their genome. Our observations show that the bph genes present on Tn4371 are highly conserved between different (chloro)biphenyl-degrading hosts, isolated globally but belonging mainly to the beta-proteobacteria. On the other hand, Tn4371-related mobile elements carrying bph genes are apparently only found in isolates from the environment that provided the Tn4371-bearing isolate A5.     

Absence of DNA sequences homologous to transposable element Tn5 (Kan) in the chromosome of Escherichia coli K-12.

The DNA hybridization procedure of Southern has been used to search for homology between the transposable kanamycin resistance determinant Tn5 and sequences in the chromosome of Escherichia coli K-12. No homology was detected under conditions in which a segment homologous to 5% or more of the 5,300-base pair Tn5 element would have been seen.     

Combinatorial events of insertion sequences and ICE in Gram-negative bacteria

The emergence of antibiotic and antimicrobial resistance in Gram-negative bacteria is incremental and linked to genetic elements that function in a so-called 'one-ended transposition' manner, including ISEcp1, ISCR elements and Tn3-like transposons. The power of these elements lies in their inability to consistently recognize one of their own terminal sequences, while recognizing more genetically distant surrogate sequences. This has the effect of mobilizing the DNA sequence found adjacent to their initial location. In general, resistance in Gram-negatives is closely linked to a few one-off events. These include the capture of the class 1 integron by a Tn5090-like transposon; the formation of the 3' conserved segment (3'-CS); and the fusion of the ISCR1 element to the 3'-CS. The structures formed by these rare events have been massively amplified and disseminated in Gram-negative bacteria, but hitherto, are rarely found in Gram-positives. Such events dominate current resistance gene acquisition and are instrumental in the construction of large resistance gene islands on chromosomes and plasmids. Similar combinatorial events appear to have occurred between conjugative plasmids and phages constructing hybrid elements called integrative and conjugative elements or conjugative transposons. These elements are beginning to be closely linked to some of the more powerful resistance mechanisms such as the extended spectrum β-lactamases, metallo- and AmpC type β-lactamases. Antibiotic resistance in Gram-negative bacteria is dominated by unusual combinatorial mistakes of Insertion sequences and gene fusions which have been selected and amplified by antibiotic pressure enabling the formation of extended resistance islands.     

Transposition genes of the Bacteroides mobilizable transposon Tn4555: role of a novel targeting gene

Conjugative transposons have been identified in several bacterial species, most notably the Gram-positive Enterococci and the Gram-negative Bacteroides. In Bacteroides species, these elements encode a complete conjugative machinery, which mediates their own intercellular transfer, and they can mobilize in trans co-resident elements. One such mobilizable element is the antibiotic resistance transposon, Tn4555, which was previously found to integrate into a specific genome target site via a site-specific recombination mechanism. In this work, we demonstrate that three Tn4555 genes were involved in integration of the element. These were int encoding a lambda-type integrase, which was absolutely required for integration of the transposon, and two accessory genes, which increased the frequency of integration. Interestingly, one of these accessory gene products, TnpA, directed the insertion of Tn4555 into the genome target site; in the absence of tnpA, the insertion pattern was essentially random. This is the first example of a site-specific recombinase that uses a specific targeting protein.     

Characterization of a Bacteroides mobilizable transposon, NBU2, which carries a functional lincomycin resistance gene

The mobilizable Bacteroides element NBU2 (11 kbp) was found originally in two Bacteroides clinical isolates, Bacteroides fragilis ERL and B. thetaiotaomicron DOT. At first, NBU2 appeared to be very similar to another mobilizable Bacteroides element, NBU1, in a 2.5-kbp internal region, but further examination of the full DNA sequence of NBU2 now reveals that the region of near identity between NBU1 and NBU2 is limited to this small region and that, outside this region, there is little sequence similarity between the two elements. The integrase gene of NBU2, intN2, was located at one end of the element. This gene was necessary and sufficient for the integration of NBU2. The integrase of NBU2 has the conserved amino acids (R-H-R-Y) in the C-terminal end that are found in members of the lambda family of site-specific integrases. This was also the only region in which the NBU1 and NBU2 integrases shared any similarity (28% amino acid sequence identity and 49% sequence similarity). Integration of NBU2 was site specific in Bacteroides species. Integration occurred in two primary sites in B. thetaiotaomicron. Both of these sites were located in the 3' end of a serine-tRNA gene NBU2 also integrated in Escherichia coli, but integration was much less site specific than in B. thetaiotaomicron. Analysis of the sequence of NBU2 revealed two potential antibiotic resistance genes. The amino acid sequences of the putative proteins encoded by these genes had similarity to resistances found in gram-positive bacteria. Only one of these genes was expressed in B. thetaiotaomicron, the homolog of linA, a lincomycin resistance gene from Staphylococcus aureus. To determine how widespread elements related to NBU1 and NBU2 are in Bacteroides species, we screened 291 Bacteroides strains. Elements with some sequence similarity to NBU2 and NBU1 were widespread in Bacteroides strains, and the presence of linA(N) in Bacteroides strains was highly correlated with the presence of NBU2, suggesting that NBU2 has been responsible for the spread of this gene among Bacteroides strains. Our results suggest that the NBU-related elements form a large and heterogeneous family, whose members have similar integration mechanisms but have different target sites and differ in whether they carry resistance genes.     

Structure and function of hot spots providing signals for site-directed specific recombination and gene expression in Tn21 transposons

Tn21- and Tn3-related transposons are widespread and carry various resistance determinants. The insertion points of different resistance genes were precisely defined in Tn2424, Tn1696, Tn2410, Tn4000 and its derivatives and compared to the corresponding sites in Tn7, pSA, R388, R46, Tn2603, Tn1331 and in Tn3-related elements. Insertional 'hot spots' located at the 3' end of different genes comprised 55 nucleotides and yielded more than 90% homology to the corresponding consensus sequence, termed hs1. Elements of this class were found to direct recA-independent generation of deletions. Flanking the 5' ends, hs2 (CTAAAACAAAGTTA) comprised the terminal nucleotides of hs1. Functional properties of hot spots as recognition sites for site-specific recombination and regulation of gene expression indicate that they might be involved in transfer, stable inheritance and expression of prokaryotic genes.     

IntI2 integron integrase in Tn7

Integrons can insert and excise antibiotic resistance genes on plasmids in bacteria by site-specific recombination. Class 1 integrons code for an integrase, IntI1 (337 amino acids in length), and are generally borne on elements derived from Tn5090, such as that found in the central part of Tn21. A second class of integron is found on transposon Tn7 and its relatives. We have completed the sequence of the Tn7 integrase gene, intI2, which contains an internal stop codon. This codon was found to be conserved among intI2 genes on three other Tn7-like transposons harboring different cassettes. The predicted peptide sequence (IntI2*) is 325 amino acids long and is 46% identical to IntI1. In order to detect recombination activity, the internal stop codon at position 179 in the parental allele was changed to a triplet coding for glutamic acid. The sequences flanking the cassette arrays in the class 1 and 2 integrons are not closely related, but a common pool of mobile cassettes is used by the different integron classes; two of the three antibiotic resistance cassettes on Tn7 and its close relatives are also found in various class 1 integrons. We also observed a fourth excisable cassette downstream of those described previously in Tn7. The fourth cassette encodes a 165-amino-acid protein of unknown function with 6.5 contiguous repeats of a sequence coding for 7 amino acids. IntI2*179E promoted site-specific excision of each of the cassettes in Tn7 at different frequencies. The integrases from Tn21 and Tn7 showed limited cross-specificity in that IntI1 could excise all cassettes from both Tn21 and Tn7. However, we did not observe a corresponding excision of the aadA1 cassette from Tn21 by IntI2*179E.     

Evolutionary changes in the structural and regulatory regions of aminoglycoside-3'-phosphotransferase type I gene of E. coli

To investigate the evolutionary relationships between the aph(3') genes from different plasmids, the nucleotide sequence of the aph(3') gene from the E. coli R plasmid was determined and compared with the known aph(3') genes of Tn903 and Tn4352. Three point mutations in the structural part of the cloned aph(3') gene caused amino acid changes in the enzyme molecule at positions 19, 27 and 48 beginning from the start codon. The structural part of the gene was followed by two stop codons and a long DNA region containing no nucleotide sequences homologous to the sequences of Tn903 or Tn4352. Both the cloned aph(3') gene and Tn4352 were limited on the left by the spacer sequence and the insertion sequence IS176. Twenty one base pairs deletion abolished the -35 sequence of the promoter suggested for the aph(3') gene of Tn4352 and resulted in formation of a fusion promoter utilizing the -35 box of IS176 and the -10 box of the aph(3') gene. The distance between the -35 and -10 sequences changed from 18 to 17 bp. Changes in the cloned aph(3') gene and the flanking DNA regions resulted in formation of a new promoter and loss of the right IS176 element.     

Structure and function of 59-base element recombination sites associated with mobile gene cassettes

The integration of gene cassettes into integrons is effected by site-specific recombination catalysed by an integrase, IntI, encoded by the integron. The cassette-associated recombination sites, 59-base elements, are not highly conserved and vary in length from 57 to 141 bp. They can be identified by their location and the relationship of over 20 bp at their outer ends to consensus sequences that are imperfect inverted repeats of one another. The recombination cross-over occurs close to one end of the 59-base element, within a conserved core site with the consensus sequence GTTAGGC or GTTRRRY. By introducing single-base changes at each of these positions in the aadB 59-base element, bases that are critical for site activity were identified. The recombination cross-over was also localized to a unique position between the adjacent G and T residues. Changes introduced in the conserved AAC of the inverse core site (GCCTAAC or RYYYAAC) located at the opposite end of the 59-base element also reduced site activity but to a lesser extent. Sequences of rare recombinants revealed an alternative position for strand exchange and led to the conclusion that 59-base elements comprise two simple sites, analogous to those recognized by other integrases, with each simple site made up of a pair of inversely oriented IntI binding domains separated by a spacer of 7 or 8 bp. Re-examination of the sequences of all known 59-base elements revealed that this simple site configuration was present at both the left and right ends in all 59-base elements. The identity of bases in the spacer is not required for efficient recombination and the cross-over is located at one end of the spacer, suggesting that during IntI1-mediated recombination only one strand exchange occurs.