Comparison of the structure–activity relationships of the integron-associated recombination sites attI3 and attI1 reveals common features

Christina M. Collis and Ruth M. Hall{dagger}

CSIRO Molecular Science, Riverside Life Sciences Centre, Riverside Corporate Park, North Ryde, NSW 2113, Australia

Correspondence
Ruth Hall
Ruth.Hall{at}mmb.usyd.edu.au


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Incorporation of gene cassettes into integrons occurs by IntI-mediated site-specific recombination between a 59-base element (59-be) site in the cassette and an attI site in the integron. While the 59-be sites share common features and are recognized by several different IntI recombinases, the sequences of attI sites are not obviously related and are preferentially recognized by the cognate IntI. To determine the features of attI sites that are required for recombination proficiency, the structure–activity relationships of a second attI site, the attI3 site from the class 3 integron, were examined. The attI3 site was confined to within a region consisting of 68 bp from the integron backbone and 15 bp from the adjacent cassette. This region includes four IntI3-binding sites, as assessed by gel shift and methylation interference studies. Two of the binding sites are inversely oriented and constitute a simple site that includes the recombination crossover point. The two additional binding sites appear to be directly oriented and one of them is essential for efficient recombination of the attI3 site with a 59-be, but not for recombination with a second full-length attI3 site, which occurs at 100-fold lower frequency. The fourth site enhances attI3 with 59-be recombination 10-fold. The finding that the organization and overall properties of attI3 are very similar to those of attI1 indicates that these features are likely to be common to all attI sites.


Abbreviations: 59-be, 59-base element; F-IntI1/3, IntI1/3 with an N-terminal FLAG extension

{dagger}Present address: School of Molecular and Microbial Biosciences, Biochemistry and Microbiology Building G08, University of Sydney, Sydney 2006, Australia.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Gene cassettes are small mobile elements that generally include a single gene and a recombination site known as a 59-be (59-base element) (Hall et al., 1991; Recchia & Hall, 1995, 1997). They are normally found in an integron in arrays of one, a few or many cassettes at a specific position which corresponds to the integron-associated attI recombination site (Stokes & Hall, 1989; Hall & Collis, 1995, 1998; Hall, 2002). The attI site lies adjacent to the intI gene which produces the IntI site-specific recombinase responsible for the movement of gene cassettes. Incorporation of cassettes into an integron occurs by IntI-mediated recombination between the 59-be in a circular cassette and the attI site in the integron (Collis et al., 1993; Hall & Collis, 1995; Collis et al., 2002a). In contrast, excision of cassettes from an integron can occur via either recombination between two 59-be or between the attI site and a 59-be (Collis & Hall, 1992b; Bunny et al., 1995; Collis et al., 2002a; Drouin et al., 2002; Hansson et al., 2002).

The IntI recombinases encoded by integrons form a distinct family within the tyrosine recombinase superfamily (Nunes-Düby et al., 1998) and contain an additional motif that appears to be a signature for IntI family members (Messier & Roy, 2001; Nield et al., 2001). Over 20 distinct IntI have been identified to date (Collis et al., 2002b), and in most cases, the intI gene is in close proximity to one or more identifiable gene cassettes, indicating that it is indeed part of an integron. Some of these integrons are part of the mobile gene pool (Stokes & Hall, 1989; Arakawa et al., 1995; Recchia & Hall, 1995; Hochhut et al., 2001), some are a feature of a bacterial chromosome (Rowe-Magnus et al., 2001; Drouin et al., 2002) and the location of others has not been established (Nield et al., 2001; Vaisvila et al., 2001). Known members of the IntI family share as little as 35 % sequence identity (see Collis et al., 2002b for a compilation), indicating a long evolutionary history for this natural gene cloning system.

Members of the tyrosine recombinase or integrase superfamily use a tyrosine residue as the nucleophile in their strand exchange reactions and share similarities in a large C-terminal domain of ~180 aa which contains six conserved residues that are part of the catalytic site (Nunes-Düby et al., 1998; Grindley, 1997; Yang & Mizuuchi, 1997; Grainge & Jayaram, 1999). A few members of the tyrosine recombinase superfamily have been extensively studied. The Cre recombinase of phage P1, which circularizes the phage genome and resolves multimers, FLP, which inverts a segment of the 2 µm plasmid of yeast, and XerC and XerD, which act together to resolve chromosomal dimers formed during replication, each recombine two copies of the same simple site using two-strand exchange reactions that occur in an ordered fashion (Sadowski, 1986; Stark et al., 1992; Gopaul & Van Duyne, 1999). The simple sites (also called core sites or the functional core) are approximately 30 bp long and consist of a pair of inverted repeats that are Int-binding sites, separated by a spacer (also called an overlap or central region) of 6–8 bp between the positions of the top and bottom strand exchanges (Sadowski, 1986; Stark et al., 1992; Nash, 1996). In the case of the phage {lambda} and HK022 Int recombinases, a simple site in the bacterial chromosome recombines with a complex site in the phage genome on integration and two complex sites at the boundaries of the integrated phage genome recombine on excision (Nash, 1996; Weisberg et al., 1999). These reactions involve an additional domain in the Int protein and its binding sites as well as accessory binding factors and their sites. These additional binding sites are part of the complex site structure. Other examples where complex site arrangements and accessory factors are required include the resolution of plasmid multimers by XerCD, which is constrained to a resolution pathway by the accessory factors (Colloms et al., 1996, 1998; Alen et al., 1997).

Both the 59-be sites in gene cassettes and the attI1 site in class 1 integrons are also complex sites that include both a discernable simple site and additional IntI-binding sites (Fig. 1a), suggesting that in this system further molecules of IntI act as accessory factors when bound to the additional sites. The cassette-associated 59-be sites have a set of common features that permit them to be identified, despite substantial sequence diversity and differences in length (Hall et al., 1991; Collis & Hall, 1992a; Stokes et al., 1997). Each 59-be appears to consist of two pairs of inversely oriented IntI-binding sites, each making up a simple site, separated by a segment that has a variable length and sequence but generally includes an inverted repeat (Stokes et al., 1997). These features (Fig. 1a) appear to be generally recognized by IntI-type tyrosine recombinases, as the same 59-be act as sites for several different IntI that are less than 50 % identical at the amino acid sequence level (Martinez & de la Cruz, 1990; Hall et al., 1991; Stokes et al., 1997; Collis et al., 2001, 2002a; Drouin et al., 2002; Hansson et al., 2002). However, IntI recombinases can distinguish between the two simple sites and recombination normally occurs exclusively within the right-hand simple site (Martinez & de la Cruz, 1990; Hall et al., 1991; Stokes et al., 1997). IntI recombinases also appear to differ from other tyrosine recombinases in that they catalyse only a single strand exchange which occurs either between the G and TT in the right hand core site (vertical arrow in Fig. 1) or between the AA and C on the complementary strand (Hansson et al., 1997; Stokes et al., 1997). By analogy with the XerCD system (Colloms et al., 1996), strand exchange on the complementary strand seems more likely.



View larger version (52K):
[in this window]
[in a new window]
 
Fig. 1. (a). Schematic structure of a 59-be and the attI1 recombination site. Horizontal arrows indicate the IntI binding sites (GTTRRRYNNN) with the designation for the individual sites (1L, 2L etc.) marked above and the positions of simple sites below. A small vertical arrow marks the recombination crossover point. (b). Alignment of attI site sequences. The IntI1-binding sites 1–4 in attI1 are marked by horizontal arrows above the sequence. In all attI, potential 7-bp core sites (related to GTTRRRY) at positions corresponding to sites 1 and 2 and separated by 4 or 5 bp, are in bold typescript. Lower-case lettering denotes sequence derived from the first cassette and can vary according to the cassette present. Colons indicate identical bases in adjacent pairs. Sequences of the attI sites were obtained from the following GenBank database entries: attI1/oxa2, M95287; attI3/blaIMP-1, AF416297; attI6, AF314191; attI7, AF314190; Xanthomonas campestris attI (Xca), AF324483; Pseudomonas alcaligenes attI (Pal), AY038186; attI9/dfrA1, AY035340; Vibrio salmonicida plasmid pRVS1 attI (attI10), AJ277063; attI4, AF055586; Vibrio mimicus attI (attI5), AF180939; Vibrio metschnikovii attI (Vme), AY014398; Vibrio parahaemolyticus attI (Vpa), AY014399; Listonella pelagia attI (Lpe), AY014401; attI2/dfrA1, AJ001816; Shewanella putrefaciens attI (Spu), AF324211; Vibrio fischeri attI (Vfi), AY014400.

 
In contrast to the 59-be type sites, the attI sites associated with different intI genes share very little obvious sequence similarity with one another (Fig. 1b). Nor do they share identifiable features with 59-be other than the 7 bp (GTTRRRY) core site (Stokes & Hall, 1989) found at the recombination crossover point, 6 bp of which is usually derived from the 59-be of the first integrated cassette and a potential inverse core site in the position needed to form a simple site (Fig. 1b). Only the attI1 site from class 1 integrons has been characterized (Recchia et al., 1994; Hansson et al., 1997; Collis et al., 1998; Gravel et al., 1998; Hall et al., 1999; Partridge et al., 2000). Detailed studies have revealed that attI1 includes four IntI-binding sites each beginning with GTTRRRY or variations thereof (Fig. 1a) with two in the vicinity of the recombination crossover position in inverse orientation, as expected for the simple site of a tyrosine recombinase, and two in direct orientation (Collis et al., 1998; Gravel et al., 1998; Partridge et al., 2000). This organization is in contrast to the arrangement found in 59-be, which contain two simple sites (Stokes et al., 1997). A requirement for the two directly oriented binding sites, one of which is strong and the other weak, is only observed when the attI1 site recombines with a 59-be (Recchia et al., 1994; Partridge et al., 2000). The strong binding site (3 in Fig. 1a), which is closest to the simple site, is essential for recombination with a 59-be partner and the weaker binding site (4 in Fig. 1a) further enhances recombination efficiency. The simple site is sufficient for recombination with a second complete attI1 site (Hansson et al., 1997; Partridge et al., 2000).

We have recently examined the properties of the only known example of a class 3 integron (Arakawa et al., 1995) and demonstrated that both integrative and excisive recombination between two 59-be or between attI3 and a 59-be is catalysed by the class 3 recombinase, IntI3 (Collis et al., 2002a). The general properties of IntI3 were the same as those of IntI1. However, we have also shown that IntI1 and IntI3 do not recognize the non-cognate attI2 and attI3 or attI1 and attI2 sites in integrative reactions with a 59-be (Collis et al., 2002b). The stringent requirement of each IntI for the cognate attI site contrasts with their shared ability to recognize many different 59-be. This and the complex and quite distinct architecture of attI1 and 59-be recombination sites raises questions as to how these sites are recognized by the IntI recombinases and how they come together to participate in productive recombinational exchanges. Here, we have sought to identify the general features of attI sites by examining the features of a second attI site. The structure–activity relationships of the attI3 site from the class 3 integron were determined using IntI3-catalysed cointegration assays and binding of purified IntI3 to attI3-containing fragments of different lengths. The binding sites were further localized by examining interference with binding caused by methylation of bases in the major groove.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bacterial strains and plasmids.
UB1637 is F his lys trp recA56 rpsL and DH5{alpha} is supE44 {Delta}lacU169 (f80 lacZ {Delta}M15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1. Plasmids are described in Table 1. pRMH859 (Fig. 2) is a derivative of R388 containing only one recombination site, the orfA/qacE 59-be, and was generated by digestion of R388 with BamHI, religation and selection for transformants resistant to sulphamethoxazole. Transformants were then screened for sensitivity to trimethoprim. Digestion of pRMH859 with SphI+HindIII was used to check that the orientation of the orfA/qacE 59-be was the same as in R388. pRMH961 (Fig. 2) was derived from pRMH956 (Collis et al., 2002b) by IntI3-mediated excision of the gene cassettes.


View this table:
[in this window]
[in a new window]
 
Table 1. Plasmids

 


View larger version (16K):
[in this window]
[in a new window]
 
Fig. 2. Structure of integron regions in the conjugative plasmids. Conserved integron regions are represented by an open box containing a number indicating the class of integron from which it is derived and a bar represents the 25 bp inverted repeat, IRi, at the left-hand boundary of the original class 1 integron in R388. The dotted line in pRMH956 and pRMH961 represents 190 bp of pACYC184 vector sequence. attI sites are tall open boxes and 59-be filled boxes, and the positions of Pc and Pc3, the promoters for transcription of the cassette genes, are indicated by bent arrows. Genes are indicated by arrows. B indicates the BamHI restriction site.

 
Plasmids containing parts of the attI3 site listed in Table 1 were mostly constructed by PCR-amplifying fragments of the complete attI3 site in pMAQ545 and cloning these fragments into pACYC184 and pUC19 as follows. Fragments cloned in plasmids pRMH841, pRMH842 and pRMH843 were amplified using primer RH69 together with primers HS273, HS274 or HS245, respectively, and that in pRMH874 was amplified using RH244 with RH245. PCR fragments were gel-purified, digested with BamHI+HindIII and ligated with BamHI/HindIII-digested pACYC184. The BamHI–HindIII fragments from pRMH841, pRMH842 and pRMH843 were subsequently transferred into BamHI/HindIII-digested pUC19 to form pRMH847, pRMH848 and pRMH849, respectively. Similarly, pRMH852, pRMH855 and pRMH889 were constructed by PCR amplification of pMAQ545 using primer pairs RH220/RH69, RH235/RH69 and RH238/RH245, respectively, and cloning the purified products into pUC19 between the BamHI/HindIII sites for pRMH852 and pRMH855, or the BamHI/HincII sites for pRMH889. The BamHI–HindIII fragment of pRMH855 was recloned in pACYC184 to form pRMH871. pRMH964 was made by annealing primers HS245 and RH263 and end-filling using Klenow fragment of DNA polymerase I, followed by digestion with BamHI+HindIII and cloning between the BamHI/HindIII sites of pUC19. Sequences of the cloned fragments in all plasmids constructed using PCR or pairs of annealed oligonucleotides were verified by DNA sequencing.

pRMH860, pRMH861, pRMH862 and pRMH872, which contain attI3 linked to the 198-bp qacE fragment (Table 1), were recovered from cointegrates formed between pRMH841, pRMH842, pRMH843 or pRMH871, respectively and the orfA/qacE 59-be element in pMAQ495 (Collis et al., 1998) by religation of the appropriate HindIII fragment. pRMH962 and pRMH963 were constructed by end-filling the small BamHI–ClaI fragments from pRMH862 and pRMH872, respectively, and cloning them into the EcoRV site of pACYC184 in orientation 2 (Collis et al., 2001). To make pRMH894 and pACYC184{Delta}, the large AvaI fragment of pRMH836 or the large BamHI–EcoRV fragment of pACYC184, respectively, were religated after end-filling.

Culture conditions.
Bacteria were routinely cultured as described previously (Collis et al., 2001) in LB medium or on LB agar supplemented as appropriate with ampicillin (100 µg ml–1), chloramphenicol (Cm, 25 µg ml–1), kanamycin (5 µg ml–1), nalidixic acid (Nx, 25 µg ml–1), spectinomycin (25 µg ml–1), streptomycin (Sm, 25 µg ml–1); or sulphamethoxazole (Su, 25 µg ml–1). Antibiotics were obtained from Sigma–Aldrich.

Conduction assays.
The frequency with which a pACYC184-based plasmid (CmR) containing a cloned attI3 fragment was transferred from the donor strain UB1637 (RecA SmR) to the recipient DH5{alpha} (RecA NxR) after IntI3-catalysed cointegration with a conjugative plasmid based on R388 (Fig. 2) was determined as described previously (Collis et al., 2001). IntI3 was supplied in trans using pRMH851. Transconjugants were selected on agar containing Su and Nx and recombinants on Cm and Nx, and the conduction frequency, which is equivalent to the cointegration frequency, was expressed as the ratio of CmR to SuR transconjugants. Cointegrates were mapped using digestion with either BamHI+HindIII or with NcoI.

Oligonucleotides.
Oligonucleotides were synthesized by Geneworks (Adelaide). Oligonucleotides from the class 3 integron conserved segment are listed below with positions in GenBank accession no. AF416297 in parentheses, extensions are shown in bold type and relevant restriction sites (BamHI or HindIII) are underlined. HS245, 5'-CGGGATCCCGGTGCCGTGCGACTTTGTTTAAC-3' (1778–1800); HS273, 5'-CGGGATCCATTTACAGGATTGATTTCAAAC-3' (1747–1767); HS274, 5'-CGGGATCCATTTGTGGGTATCCGGTGTTTG-3' (1809–1828); RH235, 5'-CGGGATCCGGTGTTTGGTCAGAT-3' (1817–1835); RH238, 5'-GATACCCACAACCGTGGTCGTTAAAC-3' (1819–1794); RH245, 5'-CGGGATCCCGACCTGTTCATGCATCCAGT-3' (1721–1739); RH263, 5'-CGAAGCTTGAACACCGGATACCCACAACCGTGGTCGTTAAACAAAGTCGCA-3' (1826–1785). RH69, complementary to sequence in the qacE cassette, is 5'-GCTGTGAGCAATTATaaGCTTAGTGC-3', equivalent to 5394–5369 in M95287 where aa is GT. Oligonucleotides complementary to the pACYC184 sequence are RH220 (5'-ACCATACCCACGCCGAAACAAGCGCT-3') and RH244 (5'-TTCTCATGTTTGACAGCTTATCATCG-3') found at positions 2014–1989 and 1496–1521, respectively, in X06403.

DNA procedures.
DNA for gel shift assays was prepared using the Wizard maxipreps (Promega) or Jetstar (Genomed) DNA purification systems. DNA fragments labelled at both ends were prepared by digesting 1·5 µg DNA from a pUC-derived plasmid containing the cloned fragment with BamHI+HindIII and end-filling with ({alpha}-32P) dATP (Amersham Biosciences) and 100 µM dGTP, dCTP and dTTP and Klenow fragment of DNA polymerase I (Roche Diagnostics). The appropriate fragment was then recovered by electrophoresis through a 2·5 % agarose gel and purified using a Geneclean II kit (Bio 101). Plasmid DNA for automated sequencing was purified using Wizard minicolumns (Promega). Sequencing was carried out at the Macquarie Sequencing Facility, Department of Biological Sciences, Macquarie University Sydney, Australia, using an ABI Prism 377 DNA sequencer (PE Biosystems) and Big Dye terminator mixes. Cointegrate DNA for restriction analysis was isolated using an alkaline lysis method as described previously (Collis et al., 2001).

DNA binding assays.
IntI3 with an N-terminal FLAG extension (F-IntI3) was purified as described previously for F-IntI1 (Collis et al., 1998, 2002b). Briefly, F-IntI3 expression was induced in a 200 ml culture of DH5{alpha} containing pMAQ487 by the addition of IPTG to 0·5 mM. After growth for a further 2 h, a cell lysate was made and the soluble fraction passed through a column containing the M2 anti-FLAG affinity resin (International Biotechnologies), washed and eluted using the FLAG peptide (International Biotechnologies). After further purification using DEAE-Sephacel, F-IntI3 was concentrated using a Centricon 10 filter (Amicon) and stored at –70 °C in 20 mM Tris/HCl, pH 7·5, 0·1 M KCl, 4 mM dithiothreitol, 0·05 % (v/v) Tween 20, 10 % glycerol. The preparation included some GroEL, but the unfolded fraction of IntI3 was estimated to be less than 10 %.

Each 15 µl binding reaction contained 33 mM Tris/HCl, pH 7·5, 50 mM KCl, 9 % (v/v) glycerol, 7 µg poly(dG-dC) ml–1 (Amersham Biosciences), 100 nM (approx.) F-IntI3 and 0·2–2 ng end-labelled DNA fragment. After incubation at room temperature for 15 min, the samples were electrophoresed through a 5 % non-denaturing polyacrylamide gel at 200 V for 1·5 h at room temperature, using 23 mM Tris-borate, 0·25 mM EDTA, pH 8, in the gel and as running buffer. After fixing and drying, gels were exposed to X-ray film overnight with an intensifying screen or to a PhosphorImager screen (Molecular Dynamics).

Methylation interference.
The attI3 fragments were labelled at one 3' end, partially methylated at guanine residues using dimethyl sulphate and purified as described previously (Collis et al., 1998). This DNA was used as the substrate for a binding reaction with F-IntI3 as described above, scaled up four- to fivefold. After electrophoresis the bands were excised from the gel and the DNA was recovered by elution into water and precipitated with ethanol. Pellets were heated in 100 µl 1 M piperidine at 90 °C for 30 min and the DNA fragments were separated by electrophoresis on a 12 % acrylamide/7 M urea denaturing gel.


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Activity of the attI3 site
We have previously shown that the attI3 site is included within 131 bp of the conserved integron backbone, together with a further short region from the adjacent blaIMP or qacE cassette (Collis et al., 2002a). The sequence of the region that includes the attI3/qacE site is shown in Fig. 3. To more accurately localize the attI3 recombination site, further deletions of the integron portion were constructed (see Fig. 3 for end points) and attI3 with 59-be recombination examined. The assay measures the ability of a small non-conjugative plasmid containing the attI3 fragment to form cointegrates with a conjugative plasmid (pRMH859; Fig. 2) that contains only a single 59-be (the orfA/qacE 59-be) in the presence of a third plasmid that produces IntI3 (pRMH851). Fragments containing more than 68 bp of the conserved integron region all exhibited full activity (Table 2) and the values obtained for longer fragments were consistent with those reported previously using a different conjugative plasmid (Collis et al., 2002a). Removal of a further 31 bp (–37 to +198 fragment) reduced activity 10-fold, suggesting that an IntI3-binding site needed for maximal site activity had been lost. When the next 8 bp were deleted (–29 to +198 fragment), the activity was a further 100-fold lower. This reduction brings the activity to a level that is 1000-fold below the activity of the full site and close to the background levels obtained with the vector plasmid (last line, Table 2). This result indicates that a further binding site that is critical for attI3 activity is removed or disrupted by deletion of bp –37 to –29. Removal of all but 15 bp of the qacE region (–68 to +15 fragment) did not reduce activity, indicating that no more than 15 bp to the right of the crossover is required. Overall, these findings are similar to those obtained previously with the attI1 site (Recchia et al., 1994; Partridge et al., 2000), where progressive removal of the weak and strong binding sites reduced, then effectively abolished, site activity in reactions with a 59-be site.



View larger version (10K):
[in this window]
[in a new window]
 
Fig. 3. attI3/qacE fragments. The attI3/qacE sequence is shown above with potential 7-bp core site sequences boxed and the extent of the simple site indicated by a bar above. A small vertical arrow marks the recombination crossover point and numbering is from this position. The extent of the cloned fragments used are shown as lines, with numbers indicating the first and last base pair, inclusively. Lower-case letters represent vector-derived sequences that flank the cloned fragments and are also present in fragments used in binding and methylation interference assays.

 

View this table:
[in this window]
[in a new window]
 
Table 2. Effect of attI3 length on IntI3-mediated recombination with the orfA/qacE 59-be

 
Background IntI3-catalysed cointegrate formation is known to be due to recombination of the orfA 59-be with secondary sites in the vector (Collis et al., 2002a). However, when several recombinants formed using either the –29 to +200-bp fragment or a –29 to +15-bp fragment were analysed by restriction mapping, all cointegrates in each group were identical and had arisen by recombination of attI3 with the orfA 59-be. This indicates that the attI3 simple site is recognized in preference to available secondary sites. Furthermore, the orientation of the integrated plasmid in these cointegrates indicated that in all of them the crossover had occurred at the usual position, i.e. within site 1 (Fig. 1a), and therefore that the determinants of orientation specificity lie within the simple site region.

Recombination of two attI3 sites
Two attI1 sites are able to recombine (Hansson et al., 1997; Partridge et al., 2000), but the frequency is about 100-fold lower than for attI1 with 59-be events. To facilitate detection of these events with attI3, a conjugative plasmid that contains only the full-length attI3 site (pRMH961; Fig. 2) was constructed and used together with plasmids that have attI3 fragments cloned in orientation 2 (as defined by Collis et al., 2001) to permit recombination with the attI3 site in pRMH961. Using the full-length attI3 site, attI3 with attI3 recombination occurred (Table 3) at a frequency equivalent to approximately 1 % of that observed for attI3 recombination with a 59-be (Table 2). Mapping of 12 cointegrates confirmed that the recombination observed involved two attI3 sites. A fragment containing only the simple site portion of the attI3 site also recombined with the complete site in pRMH961 at a similar, though slightly lower, frequency. The background level of recombination between the attI3 site in pRMH961 and secondary sites in pACYC184 was also at least 10-fold lower than for a 59-be with secondary sites.


View this table:
[in this window]
[in a new window]
 
Table 3. Effect of attI3 length on IntI3-mediated recombination with attI3

 
Binding of F-IntI3 to attI3
From the activity assays, it is possible to surmise that attI3 resembles attI1 in structure, if not in sequence, and consists of a simple site and two further IntI3-binding domains, one of which is essential for, while the other enhances, the activity of the attI3 site in recombination with a 59-be. This was confirmed by gel shift analysis using purified F-IntI3 (hereafter IntI3) that has a 17-aa N-terminal extension including the FLAG octapeptide. F-IntI3 was previously shown to have the same activity as IntI3 in a variety of in vivo assays (Collis et al., 2002a, b). The DNA fragments used included various lengths of the integron portion and 15 bp derived from the adjacent qacE cassette, as well as short flanking regions derived from the oligonucleotides used to construct them (Fig. 3). The complete attI3 fragment (–131 to +15) exhibited three to four shifted bands (Fig. 4a). At the higher IntI3 concentrations, the most intense bands potentially correspond to two or three molecules of IntI3 bound and the band corresponding to four bound monomers is faint, though reproducibly present. Thus co-operative binding to form a single complex of attI3 with four bound IntI3 molecules did not occur.



View larger version (19K):
[in this window]
[in a new window]
 
Fig. 4. Gel retardation of attI3-containing fragments by IntI3. End-labelled DNA was incubated with IntI3 and then electrophoresed through an acrylamide gel. (a) The –131 to +15 fragment with increasing concentrations of IntI3 (nM, approx.) which are shown above the lanes. The positions of free DNA (F) and retarded complexes (I, II, III and IV) corresponding to one, two, three and four molecules of IntI3 bound are marked on the right. (b) Binding of attI3/qacE fragments that include increasing lengths of the conserved integron region. The position of the left boundary of each fragment and the presence or absence of IntI3 (approx. 100 nM) are indicated above the lanes. (c) Binding of attI3 half sites. The left hand (LH) fragment includes base pairs –125 to –27, the right hand (RH) fragment is –29 to +15 and the complete fragment (LH+RH) is –131 to +15.

 
The longer fragments all exhibit four shifted bands at the IntI3 concentration used, though the bands corresponding to one and four bound monomers are faint (Fig. 4b). The –68 to +15 fragment gave a pattern identical to that seen with longer fragments, indicating that all of the binding sites in attI3 lie within this span. This is consistent with the activity assays (Table 2). The –37 to +15 fragment formed one strong band corresponding to two molecules of IntI3 bound (Fig. 4b) and much weaker bands at positions corresponding to one and three bound IntI3 (not visible in Fig. 4b). The absence of significant levels of complex III, which is relatively abundant with longer fragments, indicates that a binding site, part or all of which is located between –68 and –37 (site 4), is likely to be involved in the formation of complex III.

The –29 to +15 fragment, which includes only the simple site, shows only very weak retarded bands potentially corresponding to one and two IntI3 monomers bound. This is consistent with the loss of a further binding site, at least part of which is located in the –37 to –30 span, which appears to be important for the much stronger binding seen with longer fragments. Binding to a fragment containing only the left-hand portion of attI3 was also examined. A fragment that includes the –125 to –27 span formed two retarded products (Fig. 4c) and a –68 to –20 fragment gave a similar pattern (data not shown). Binding to these fragments was stronger than binding to the right-hand fragment and a strong band (one IntI3 bound) and a weaker band (two IntI3 bound) were observed, consistent with the presence of a strong and a weak binding site. Together these results localize site 3 to between positions –37 and –27 and suggest that it is the strongest binding site.

Correlating these gel shift data with the activity data in Table 2, it is possible to conclude that, as for attI1, at least three binding sites corresponding to the simple site (sites 1 and 2) and the adjacent binding site 3 are absolutely required for attI3 to recombine with a 59-be site. The fourth site corresponds to site 4 of attI1 and enhances activity a further 10-fold.

Methylation interference of IntI3 binding
Methylation interference was used to confirm the binding stoichiometry and to localize the IntI3-binding sites more accurately. DNA fragments were labelled at one end, G residues were methylated using dimethyl sulphate and DNA recovered from retarded bands was examined for underrepresentation of residues where methylation interferes with the binding of IntI3. Complexes I and II formed with the –99 to +15 fragment were examined using the bottom strand only (Fig. 5a). In complex I, G residues at –42, –43 and –46 were underrepresented in comparison to the free DNA, indicating that this complex consists mainly of one IntI3 monomer bound to site 4. The G residues that interfered with binding are adjacent to a discernable 7 bp core site, GTTTAAC, in the sequence at –52 to –46 that is likely to form the left end of site 4 (Fig. 5c) and defines its orientation, as a 7-bp core site is normally found at the beginning of IntI-binding sites (Collis et al., 1998; Gravel et al., 1998; Stokes et al., 1997). In complex II, residues that were underrepresented were in the vicinity of site 2, the predicted inversely oriented site, and site 3, indicating that binding to these two sites predominates in this species, which is normally far more abundant than complex I (Fig. 4). No complex III was recovered.



View larger version (29K):
[in this window]
[in a new window]
 
Fig. 5. Methylation interference with IntI3 binding to attI3/qacE. The –99 to +15 bp fragment from pRMH847, 3'-end-labelled on the bottom strand (a) and the –37 to +15 bp fragment from pRMH848, 3'-end-labelled on the top strand (b) were partially methylated at guanine residues, mixed with IntI3 and protein-bound complexes separated from free DNA. The unshifted DNA (F) and DNA in complexes with one and two molecules of IntI3 bound (I and II) was recovered, cleaved at methylated residues and the products separated by acrylamide gel electrophoresis. Nucleotide positions are shown on the left and the recombination crossover point is marked by a horizontal arrow. Positions of potential IntI3-binding sites 1–4 are shown by vertical arrows on the right. Open circles mark nucleotides at which methylation interferes with binding in complex I and filled circles mark those at which methylation interferes with binding in complex II. (c) The attI3 sequence showing positions of methylation interference (filled circles) and predicted positions of binding sites 1–4 (arrows).

 
For both the top (Fig. 5b) and bottom (data not shown) strand of the –37 to +15 fragment isolated from the complex corresponding to two bound IntI3, the residues where methylation interfered with binding were also located in the vicinity of binding sites 2 and 3 and for the bottom strand were as observed for complex II of the longer –99 to +15 fragment. Residues within site 1 were not significantly underrepresented. Together these findings localize site 3 to a region that includes bp –33 to –26. On the basis of these findings, together with a comparison of the sequences in the regions of the four sites in attI3, we have placed 11 bp IntI3-binding sites in the positions shown in Fig. 5(c). The orientation of site 3 was deduced from the position of the 7-bp sequence most closely matching the core site consensus, and of site 2 by assuming that an inverse orientation is needed to form a simple site. The data also suggest that binding to sites 2 and 3 may be co-operative, as interference with binding to only these two sites is observed in the complexes with two bound IntI3, but they are not detected in complex I.


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The properties of the attI3 site identified in this study were very similar to those of the attI1 site (Recchia et al., 1994; Hansson et al., 1997; Collis et al., 1998; Gravel et al., 1998; Partridge et al., 2000). Four IntI3-binding sites were detected in gel shift assays and the positions of three determined directly using methylation interference. The position of the remaining binding site was inferred because the site that includes the position of the strand exchange reaction (site 1) must bind IntI3 for the reaction to occur. These four binding sites appear to be arranged in a similar way to the IntI1-binding sites in attI1 with two directly oriented sites located to the left of the pair of inversely oriented sites that comprise the simple site (Fig. 5c). As with attI1, at least three of these IntI3-binding sites are essential for integrative recombination with a 59-be. The fourth binding site enhanced site activity by a further 10-fold. Recombination with a second complete attI3 site occurred, but the frequency for this event was 100-fold less than that observed for attI3 with 59-be recombination and only the simple site was needed for the reaction. The methylation interference data presented here confirm the location of site 2 inferred from similarity to the 7 bp core sequence and assuming that the simple site consists of inversely oriented binding sites as is the case for attI1 (Partridge et al., 2000) and all other tyrosine recombinases studied to date, including Cre-lox for which the orientation is known from crystal structures (Gopaul & Van Duyne, 1999). The essential accessory binding site, site 3, was precisely located by both deletion and methylation interference studies and oriented by reference to the TGTGGGT (two differences from the GTTRRRY) at its left-hand end. The second accessory site, site 4, which is not critical for activity, was less conclusively defined, but the extent and orientation shown in Fig. 5(c) are consistent with all of the available information. However, further work is needed to confirm this. None of the species analysed by methylation interference contained IntI3 bound to site 1, where the recombination crossover occurs, and this site appears to have the lowest affinity for IntI3. An alignment of attI1 and attI3 showing binding sites in attI1 determined previously (Collis et al., 1998; Gravel et al., 1998) and the sites in attI3 determined in this study reveals that sites 3 and 4 are in equivalent positions with respect to the simple site (Fig. 6).



View larger version (5K):
[in this window]
[in a new window]
 
Fig. 6. Comparison of the structure of attI1 and attI3 with att5401 Int-binding sites. Arrows represent 11-bp binding sites for IntI1 (Collis et al., 1998; Gravel et al., 1998) and IntI3 (this study), and 12-bp binding sites for Int5401 (Baum, 1995). The position of the first base in each binding site relative to the cleavage point in attI1 and attI3 is marked below.

 
A number of features of the binding data, together with the methylation interference data permit some tentative conclusions to be drawn about how the attI3 complex with IntI3 might be formed. Binding is weak when only sites 1 and 2 are present and, while stronger binding is observed when only sites 3 and 4 are present, the predominant complex corresponds to one molecule of bound IntI3. However, when both sites 2 and 3 are present in the substrate DNA, complex I is barely visible and the predominant complex has IntI3 bound to both sites 2 and 3. Together these findings indicate that binding to site 3 is likely be strongest and perhaps to occur first and that binding to sites 2 and 3 is co-operative. Formation of complex III is likely to involve further binding to site 4 as this complex is not formed when site 4 had been deleted. With this scenario, binding to site 1 where the strand exchange occurs would occur last. However, further DNase I protection and interference data are needed to confirm this.

It remains to be established whether binding of IntI1 to attI1 proceeds in a similar way. In our study of IntI1 binding (Collis et al., 1998), only binding to sites 3 and 4 was observed, though patterns equivalent to those seen here with IntI3/attI3 were seen occasionally (C. M. Collis, unpublished). A pattern of four bands with IntI1/attI1 has also been reported by others (Gravel et al., 1998). As IntI1 binds only weakly to the attI3 site and IntI3 does not bind significantly to the attI1 site (Collis et al., 2002b), the ability of IntI3 and IntI1 to distinguish between the attI1 and attI3 sites and therefore to catalyse integrative recombination between a 59-be and the cognate attI site (Collis et al., 2002b) appears to involve differences in the sequences of one or more of the binding sites that differentially influence their affinity for the IntI enzymes. These differences are likely to reside in sites 2 and 3.

A similar arrangement of four integrase-binding sites, a simple site and two directly oriented sites, has also been found in the site recognized by the TnpI tyrosine recombinase of Tn5401 which is responsible for resolution of cointegrates formed during transposition (Baum, 1995). However, the Tn5401 resolution site is longer than the attI sites and the position of the directly oriented accessory sites with respect to the simple site is not the same (Fig. 6). In att5401 a binding site is not found at the position corresponding to site 3 of attI1 and attI3 and the fourth site is located further away from the simple site. Further differences are that the reaction catalysed involves two att5401 sites and that all four binding sites are required for resolution and to form a complex detectable in gel shift experiments. With the full-length site, a single complex corresponding to four bound TnpI is seen, indicating that TnpI binding is cooperative. Thus while the TnpI5401 recombinase is weakly related to members of both the IntI and Xer families (20–27 % identity) and the IntI and TnpI systems share the feature of using IntI binding to additional sites as accessory functions, comparisons to this system appear to shed little light on the structure–activity relationships for attI sites.

The site-specific recombination systems specified by integrons are already known to differ from the well-studied paradigm systems in a number of quite fundamental ways. One is the complexity of the recombination sites described here, which contrasts with the simple sites normally recognized by Cre, FLP and XerCD, and the {lambda} Int attB site. Another difference is that while the recombination events catalysed by {lambda} Int, Cre, FLP, XerCD and Int916 (Int1545) all involve two strand exchanges (top and bottom) that occur in an ordered fashion at positions staggered by 6, 7 or 8 bp, there is good evidence that IntI1 catalyses only a single strand exchange (Hansson et al., 1997; Stokes et al., 1997). IntI recombinases also include an additional conserved motif that is required for activity (Messier & Roy, 2001) but whose role in the recombination process is yet to be determined. Studies of further members of this important recombinase family as well as further in-depth studies on the better-characterized IntI1 and IntI3 are needed to determine how productive synaptic complexes are formed and how the reaction is constrained to a single strand exchange.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Alen, C., Sherratt, D. J. & Colloms, S. D. (1997). Direct interaction of aminopeptidase A with recombination site DNA in Xer site-specific recombination. EMBO J 16, 5188–5197.[Abstract/Free Full Text]

Arakawa, Y., Murakami, M., Suzuki, K., Ito, H., Wacharotayankun, R., Ohsuka, S., Kato, N. & Ohta, M. (1995). A novel integron-like element carrying the metallo-{beta}-lactamase gene blaIMP. Antimicrob Agents Chemother 39, 1612–1615.[Abstract]

Baum, J. A. (1995). TnpI recombinase: identification of sites within Tn5401 required for TnpI binding and site-specific recombination. J Bacteriol 177, 4036–4042.[Abstract]

Bunny, K. L., Hall, R. M. & Stokes, H. W. (1995). New mobile gene cassettes containing an aminoglycoside resistance gene, aacA7, and a chloramphenicol resistance gene, catB3, in an integron in pBWH301. Antimicrob Agents Chemother 39, 686–693.[Abstract]

Collis, C. M. & Hall, R. M. (1992a). Gene cassettes from the insert region of integrons are excised as covalently closed circles. Mol Microbiol 6, 2875–2885.[Medline]

Collis, C. M. & Hall, R. M. (1992b). Site-specific deletion and rearrangement of integron insert genes catalysed by integron DNA integrase. J Bacteriol 174, 1574–1585.[Abstract]

Collis, C. M., Grammaticopoulos, G., Briton, J., Stokes, H. W. & Hall, R. M. (1993). Site-specific insertion of gene cassettes into integrons. Mol Microbiol 9, 41–52.[Medline]

Collis, C. M., Kim, M.-J., Stokes, H. W. & Hall, R. M. (1998). Binding of the purified integron DNA integrase IntI1 to integron- and cassette-associated recombination sites. Mol Microbiol 29, 477–490.[CrossRef][Medline]

Collis, C. M., Recchia, G. D., Kim, M.-J., Stokes, H. W. & Hall, R. M. (2001). Efficiency of recombination reactions catalyzed by class 1 integron integrase IntI1. J Bacteriol 183, 2535–2542.[Abstract/Free Full Text]

Collis, C. M., Kim, M.-J., Partridge, S. R., Stokes, H. W. & Hall, R. M. (2002a). Characterization of the class 3 integron and the site-specific recombination system it determines. J Bacteriol 184, 3017–3026.[Abstract/Free Full Text]

Collis, C. M., Kim, M.-J., Stokes, H. W. & Hall, R. M. (2002b). Integron-encoded IntI integrases preferentially recognize the adjacent cognate attI site in recombination with a 59-be site. Mol Microbiol 46, 1415–1427.[CrossRef][Medline]

Colloms, S. D., McCulloch, R., Grant, K., Neilson, L. & Sherratt, D. J. (1996). Xer-mediated site-specific recombination in vitro. EMBO J 15, 1172–1181.[Abstract]

Colloms, S. D., Alen, C. & Sherratt, D. J. (1998). The ArcA/ArcB two-component regulatory system of Escherichia coli is essential for Xer site-specific recombination at psi. Mol Microbiol 28, 521–530.[CrossRef][Medline]

Datta, N. & Hedges, R. W. (1972). Trimethoprim resistance conferred by W plasmids in Enterobacteriaceae. J Gen Microbiol 72, 349–355.[Medline]

Drouin, F., Mélançon, J. & Roy, P. H. (2002). The IntI-like tyrosine recombinase of Shewanella oneidensis is active as an integron integrase. J Bacteriol 184, 1811–1815.[Abstract/Free Full Text]

Gopaul, D. N. & Van Duyne, G. D. (1999). Structure and mechanism in site-specific recombination. Curr Opin Struct Biol 9, 14–20.[CrossRef][Medline]

Grainge, I. & Jayaram, M. (1999). The integrase family of recombinases: organization and function of the active site. Mol Microbiol 33, 449–456.[CrossRef][Medline]

Gravel, A., Fournier, B. & Roy, P. H. (1998). DNA complexes obtained with the integron integrase IntI1 at the attI1 site. Nucleic Acids Res 26, 4347–4355.[Abstract/Free Full Text]

Grindley, N. D. (1997). Site-specific recombination: synapsis and strand exchange revealed. Curr Biol 7, R608–R612.[CrossRef][Medline]

Hall, R. M. (2002). Gene cassettes and integrons: moving single genes. In Horizontal Gene Transfer, pp. 19–28. Edited by M. Syvanen & C. I. Kado. London: Academic Press.

Hall, R. M. & Collis, C. M. (1995). Mobile gene cassettes and integrons: capture and spread of genes by site-specific recombination. Mol Microbiol 15, 593–600.[Medline]

Hall, R. M. & Collis, C. M. (1998). Antibiotic resistance in gram-negative bacteria: the role of gene cassettes and integrons. Drug Resist Updates 1, 109–119.

Hall, R. M., Brookes, D. E. & Stokes, H. W. (1991). Site-specific insertion of genes into integrons: role of the 59-base element and determination of the recombination cross-over point. Mol Microbiol 5, 1941–1959.[Medline]

Hall, R. M., Collis, C. M., Kim, M.-J., Partridge, S. R., Recchia, G. D. & Stokes, H. W. (1999). Mobile gene cassettes and integrons in evolution. Ann N Y Acad Sci 870, 68–80.[Abstract/Free Full Text]

Hansson, K., Sköld, O. & Sundström, L. (1997). Non-palindromic attI sites of integrons are capable of site-specific recombination with one another and with secondary targets. Mol Microbiol 26, 441–453.[Medline]

Hansson, K., Sundström, L., Pelletier, A. & Roy, P. H. (2002). IntI2 integron integrase in Tn7. J Bacteriol 184, 1712–1721.[Abstract/Free Full Text]

Hochhut, B., Lotfi, Y., Mazel, D., Faruque, S. M., Woodgate, R. & Waldor, M. K. (2001). Molecular analysis of antibiotic resistance gene clusters in Vibrio cholerae O139 and O1 SXT constins. Antimicrob Agents Chemother 45, 2991–3000.[Abstract/Free Full Text]

Martinez, E. & de la Cruz, F. (1990). Genetic elements involved in Tn21 site-specific integration, a novel mechanism for the dissemination of antibiotic resistance genes. EMBO J 9, 1275–1281.[Abstract]

Messier, N. & Roy, P. H. (2001). Integron integrases possess a unique additional domain necessary for activity. J Bacteriol 183, 6699–6706.[Abstract/Free Full Text]

Nash, H. A. (1996). Site-specific recombination: integration, excision, resolution and inversion of defined DNA segments. In Escherichia coli and Salmonella Cellular and Molecular Biology, pp. 2363–2376. Edited by F. C. Neidhardt. Washington, DC: American Society for Microbiology.

Nield, B. S., Holmes, A. J., Gillings, M. R., Recchia, G. D., Mabbutt, B. C., Nevalainen, K. M. H. & Stokes, H. W. (2001). Recovery of new integron classes from environmental DNA. FEMS Microbiol Lett 195, 59–65.[CrossRef][Medline]

Nunes-Düby, S. E., Kwon, H. J., Tirumalai, R. S., Ellenberger, T. & Landy, A. (1998). Similarities and differences among 105 members of the Int family of site-specific recombinases. Nucleic Acids Res 26, 391–406.[Abstract/Free Full Text]

Partridge, S. R., Recchia, G. D., Scaramuzzi, C., Collis, C. M., Stokes, H. W. & Hall, R. M. (2000). Definition of the attI1 site of class 1 integrons. Microbiology 146, 2855–2864.[Abstract/Free Full Text]

Recchia, G. D. & Hall, R. M. (1995). Gene cassettes: a new class of mobile element. Microbiology 141, 3015–3027.[Medline]

Recchia, G. D. & Hall, R. M. (1997). Origins of the mobile gene cassettes found in integrons. Trends Microbiol 5, 389–394.[CrossRef][Medline]

Recchia, G. D., Stokes, H. W. & Hall, R. M. (1994). Characterisation of specific and secondary recombination sites recognised by the integron DNA integrase. Nucleic Acids Res 22, 2071–2078.[Abstract]

Rose, R. E. (1988). The nucleotide sequence of pACYC184. Nucleic Acids Res 16, 355.[Medline]

Rowe-Magnus, D. A., Guerout, A.-M., Ploncard, P., Dychinco, B., Davies, J. & Mazel, D. (2001). The evolutionary history of chromosomal super-integrons provides an ancestry for multiresistant integrons. Proc Natl Acad Sci U S A 98, 652–657.[Abstract/Free Full Text]

Sadowski, P. (1986). Site-specific recombinases: changing partners and doing the twist. J Bacteriol 165, 341–347.[Medline]

Stark, W. M., Boocock, M. R. & Sherratt, D. J. (1992). Catalysis by site-specific recombinases. Trends Genet 8, 432–439.[Medline]

Stokes, H. W. & Hall, R. M. (1989). A novel family of potentially mobile DNA elements encoding site-specific gene-integration functions: integrons. Mol Microbiol 3, 1669–1683.[Medline]

Stokes, H. W., O'Gorman, D. B., Recchia, G. D., Parsekhian, M. & Hall, R. M. (1997). Structure and function of 59-base element recombination sites associated with mobile gene cassettes. Mol Microbiol 26, 731–745.[Medline]

Vaisvila, R., Morgan, R. D., Posfai, J. & Raleigh, E. A. (2001). Discovery and distribution of super-integrons among Pseudomonads. Mol Microbiol 42, 587–601.[CrossRef][Medline]

Weisberg, R. A., Gottesmann, M. E., Hendrix, R. W. & Little, J. W. (1999). Family values in the age of genomics: comparative analyses of temperate bacteriophage HK022. Annu Rev Genet 33, 565–602.[CrossRef][Medline]

Yang, W. & Mizuuchi, K. (1997). Site-specific recombination in plane view. Structure 5, 1401–1406.[Medline]

Received 25 June 2003; revised 8 December 2003; accepted 2 February 2004.



This Article
Abstract
Full Text (PDF)
Alert me when this article is cited
Alert me if a correction is posted
Citation Map
Services
Email this article to a friend
Similar articles in this journal
Similar articles in PubMed
Alert me to new issues of the journal
Download to citation manager
Google Scholar
Articles by Collis, C. M.
Articles by Hall, R. M.
Articles citing this Article
PubMed
PubMed Citation
Articles by Collis, C. M.
Articles by Hall, R. M.
Agricola
Articles by Collis, C. M.
Articles by Hall, R. M.


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
INT J SYST EVOL MICROBIOL MICROBIOLOGY J GEN VIROL
J MED MICROBIOL ALL SGM JOURNALS
Copyright © 2004 Society for General Microbiology.