Synapsis and DNA cleavage in C31 integrase-mediated site-specific reco
http://www.100md.com
《核酸研究医学期刊》
Institute of Genetics, University of Nottingham, Queens Medical Centre, Nottingham NG7 2UH, UK and 1 Centre of Biomolecular Sciences, University of Nottingham, University Park, Nottingham, NG7 2RD, UK
*To whom correspondence should be addressed at Institute of Medical Sciences, University of Aberdeen, Foresterhill, Aberdeen AB25 2ZD, UK. Tel: +44 1224 555739; Fax: +44 1224 555844; Email: maggie.smith@abdn.ac.uk
ABSTRACT
The Streptomyces phage C31 encodes an integrase belonging to the serine recombinase family of site-specific recombinases. The well studied serine recombinases, the resolvase/invertases, bring two recombination sites together in a synapse, and then catalyse a concerted four-strand staggered break in the DNA substrates whilst forming transient covalent attachments with the recessed 5' ends. Rotation of one pair of half sites by 180° relative to the other pair occurs, to form the recombinant configuration followed by ligation of the DNA backbone. Here we address the nature of the recombination intermediates formed by C31 integrase when acting on its substrates attP and attB. We have identified intermediates containing integrase covalently attached to cleaved DNA substrates, attB or attP, by analysis of complexes in gels and after treatment of these complexes with proteinases. Using a catalytically inactive integrase mutant, S12A, the synaptic complexes containing integrase, attP and attB were identified. Furthermore, we have shown that attB mutants containing insertions or deletions are blocked in recombination at the stage of strand cleavage. Thus, there is a strict spacing requirement within attB, possibly for correct positioning of the catalytic serine relative to the scissile phosphate in the active site. Finally, using integrase S12A we confirmed the inability of attL and attR or other combinations of sites to form a stable synapse, indicating that the directionality of integrative recombination is determined at synapsis.
INTRODUCTION
Many temperate bacteriophages encode systems that enable site-specific recombination of the phage genome with the host chromosome, bringing about integration during the establishment of lysogeny and excision during induction into lytic growth. The recombinase can belong either to a family of proteins related to phage integrase (the tyrosine recombinases) or to an evolutionary and mechanistically unrelated family which includes the resolvase/invertases (the serine recombinases) (1). The integrases belonging to the serine recombinase class form part of a diverse group of proteins that range in size from about 450 amino acids to greater than 750 amino acids, and all contain an N-terminal domain similar to the resolvase/invertase catalytic domain (1). Because of their much greater mass compared to the resolvase/invertases this group of proteins have been called the large serine recombinases and their biological functions include deletion (SpoIVCA in Bacillus subtilis sporulation and XisF in heterocyst development in cyanobacteria), transposition (TnpX and TndX in Tn4451 and Tn5397, respectively) and phage integration (shown originally in the lactococcal phage TP901-1 and Streptomyces phage R4) (2–7). The two systems that have been studied in most detail, however, are Streptomyces phage C31 integrase and mycobacteriophage Bxb1 integrase (8–12). A CLUSTALW alignment of approximately 30 serine integrases revealed conserved regions outside the catalytic domain but their function in recombination is still largely uncharacterized (1).
Phage integrases act on pairs of recombination sites that are different in sequence; attP and attB for integration and attL and attR for excision. For integration of C31 attP and attB are both short, <50 bp (11,13), and similar length recombination sites have also been reported for other systems that use serine integrases (9,14,15). All the recombination sites characterized to date that are acted on by large serine recombinases contain a stretch of identical sequence between 2 and 12 bp at which recombination occurs. Both C31 and Bxb1 integrases catalyse integrative recombination in vitro between attP and attB in the absence of other proteins or high energy cofactors and with supercoiled or linear substrates (9,11).
The mechanism of recombination by the large serine recombinases is thought to resemble that of the resolvase/invertases. Biochemical studies have demonstrated that the resolvase/invertases act via a concerted, four-strand cleavage and rejoining mechanism. A staggered break is generated at a 2 bp core sequence and transient phospho-serine linkages are formed to the recessed 5' ends (16,17). A 180° rotation of one pair of half sites is then thought to occur to configure the DNA molecules into the recombinant format prior to ligation of the DNA backbone. Several of the essential catalytic residues equivalent to, for example, Arg8, Ser10 and Arg68 using the resolvase nomenclature have been shown to be required for activity in the transposases TnpX and TndX (5,6). In Bxb1 and C31 integrases the proposed catalytic serines are required (8,10). The importance of the 2 bp at the crossover site has been demonstrated recently in both C31 and Bxb1 integrases (8,12). Mutations in these 2 bp can completely switch the polarity of the att sites leading to incorrectly joined recombination products, i.e. the left side of one site becomes joined to the left side of the other site and the reciprocal right side to right side joining. Moreover, in Bxb1 the precise location of the 2 bp staggered break in both attP and attB has also been demonstrated (8). The conclusions from these papers indicate that integrases apparently synapse freely in either of two forms that, if the sequences at the staggered break are complementary, lead to correct or incorrect recombination products. Furthermore, synapsis activates initiation of strand exchange but if the products cannot form due to mismatches, integrase iterates rotation of the half sites leaving a topological footprint but no recombination products.
The directionality of the integrative versus excisive recombination reactions has been studied with purified C31 integrase (11). In vitro C31 integrase cannot recombine pairs of sites other than attP and attB and this has led to the proposal that C31 encodes an Xis or directionality factor that enables excisive recombination between attL and attR. Experiments performed to investigate the mechanism of this site selectivity indicated that, whilst C31 integrase binds attP, attB, attL and attR with approximately equal affinities, putative recombination intermediates (such as those detected in assays containing attP/attB) could not be detected when non-permissive pairs of sites, such as attP/attP, attL/attR, were used in a reaction (11). This implied that the block in recombination between non-permissive sites is at an earlier step, i.e. synapsis.
The aim of this study was to characterize the recombination intermediates generated in the C31 integration reaction and to test the assertion that C31 integrase uses a similar mechanism of recombination as the resolvase/invertases. Using non-denaturing gels we have demonstrated the presence of complexes containing the cleaved intermediates covalently bound to protein. We have also characterized a putative synaptic complex that accumulates in reactions containing a mutant integrase containing an S12A mutation in the catalytic serine residue. Furthermore, mutant attB sites have been created that can apparently synapse but cannot efficiently activate recombination. Finally, we confirmed that combinations of att sites other than attP and attB do not form synaptic complexes.
MATERIALS AND METHODS
Bacterial strains and plasmids
Growth, maintenance and transformation of Escherichia coli were performed as described previously (18). Escherichia coli DH5 was used as the routine host for amplification and preparation of plasmids throughout this study. The methods used for preparation of plasmids and in vitro manipulation of DNA were described previously (18). Plasmids pRT600 and pRT700 containing 50 bp attB and attP sites, respectively, in pGEM7 were constructed previously (11). The screen for mutants in attB and attP was performed as follows: pRT504 contains attP and attB sites in the same orientations flanking an 5 kb fragment encoding the lac operon. pRT504 was constructed by inserting an EcoRI fragment encoding attP into pZMR100 (19) (a plasmid derived from the replication functions of phage ) followed by insertion of a PstI–BamHI fragment encoding attB derived from pHS23 (10) to give pHS40. The 5.8 kb NruI–ScaI fragment encoding the lac operon from pSKS106 (20) was inserted into the unique BamHI site in pHS40 via BglII linkers to form pRT504. This plasmid was introduced into XL1Red according to the manufacturer’s instructions (Stratagene). The transformants from each plate were pooled and plasmids were prepared from an overnight culture and used to transform DH5 containing pHS62 and expressing integrase. Ampicillin, kanamycin resistant colonies that retained expression of ?-galactosidase were picked, re-streaked to confirm the continued expression of ?-galactosidase and the plasmids isolated. The pRT504 derivatives containing the mutated att sites were purified and the att sites sequenced to reveal the mutations. The mutant sites were subcloned into pGEM7 to form pRT518 and pRT519.
Plasmid pMSX7 was constructed by annealing oligonucleotides PRMS74 (5'-AGCTTATCGATTTCGGGAGTACGCGCCCGGGGAGCCCAAGGGGCACGCCCTGGCACCCGCACCGCGGAATTCC-3'; Invitrogen) and PRMS75 (5'-TCGAGGAATTCCGCGGTGCGGGTGCCAGGGCGTGCCCCTTGGGCTCCCCGGGCGCGTACTCCCGAAATCGATA-3') which were then ligated to pGEM7 digested with XhoI and HindIII and introduced into DH5 by transformation. Plasmid pHS2116 was prepared by inserting a PCR fragment amplified using primers B+1G (5'-TCCTCTAGACTCGAGGAATTCCGCGGTGCGGGTGCCAGGGCGTGCCCTTGGgGCTCCCCGGGCGCGTACTC-3') and SP6 with pRT600 as a template, digesting this with XhoI and HindIII and ligating the product with XhoI and HindIII cut pGEM7. pHS2110 was prepared as for pHS2116 except that the primer used was B+ 10 (5'-TCCTCTAGACTCGAGGAATTCCGCGGTGCGGGTGCCAGGGCGTGCCCAGATCTTGTACAGGGCTCCCCGGGCGCGTACTC-3'). Mutations were verified by DNA sequencing.
Construction of pMSX6 containing the S12A mutant of integrase was performed by replacing the NdeI-EcoRI fragment at the 5' end of the int gene in pHS62. PRMS64 (5'-TATGGACACGTACGCGGGTGCTTACGACCGTCAGGCGCGCGAGCGCGAG-3') and PRMS65 (5'-AATTCTCGCGCTCGCGCGCCTGACGGTCGTAAGCACCCGCGTACGTGTCCA-3') were annealed and ligated to pHS62 digested with NdeI and EcoRI. The ligated DNAs were transformed into DH5 and pMSX6 was verified by sequencing.
Preparation of DNA fragments
For the supershift assays the 200 bp cold partner fragments were prepared by PCR amplification with SP6 and T7 primers of pRT600, pRT700, pMSX7 and pHS2116 as substrates for attB, attP, attBMSX7 and attB2116, respectively. The 94 bp cold partner attP site was prepared by PCR amplification of pRT700 with primers PRMS66 (5'-TCTAGACTCGAGGAATTAGT-3') and PRMS67 (5'-TCTCCGGATCCAAGCTTATCG-3'). The PCR templates for attL and attR were prepared from pRT602700, the product of recombination between attB (pRT602) and attP (pRT700). pRT602700 was restricted with ScaI and the resulting fragments were separated by electrophoresis and purified to yield a 2 kb fragment encoding attR and a 3.5 kb fragment encoding attL. PCR amplification of attR and attL was carried out using these fragments as templates and primers SP6 and T7. The resulting 200 bp fragments were separated by agarose gel electrophoresis (1.5% TAE) and purified.
The labelled fragments for all the DNA binding assays were prepared as follows: pRT600, pRT700, pMSX7 or pHS2116 were digested with XhoI and HindIII and the desired fragments of 76, 75, 77 and 77 bp containing attB, attP, attBMSX7 and attB2116, respectively, were purified. Fragments containing attL (71 bp) and attR (74 bp) were prepared by triple digestion of the 200 bp attL and attR cold partner fragments (MspI, HindII, EcoRI for attL and XhoI, HindIII, AatII for attR). The fragments were end-labelled with Klenow fragment in the presence of dCTP and dATP.
DNA binding and recombination assays
DNA affinity and supershift assays were performed as described previously (11). Briefly, assays were performed with 1.5 ng of labelled probe in binding buffer (BB) containing 20 mM Tris–HCl pH 8.0, 0.1 mM EDTA, 50 mM KCl, 5% glycerol, 70 μg/ml of sonicated salmon sperm DNA, 70 μg/ml bovine serum albumin (BSA) and integrase added to final concentrations 0, 66, 33, 16, 8, 4 and 2 nM. Reactions with no integrase contained instead 1 μg of BSA. Reactions were incubated at 30°C for 30 min prior to electrophoresis. Samples were separated on 5% acrylamide gels run at 200 V, 5 W for 2 h. For the supershift assays, 0.9–1.5 ng of labelled probe was mixed with 20 ng of cold partner fragment and 66 nM integrase in BB buffer. Reactions were incubated at 30°C for 2 h prior to electrophoresis. Where samples required proteinase treatment, samples were first heat inactivated by incubation at 72°C for 10 min and then treated either with 0.1 μg of subtilisin (incubated at 30°C for 15 min) or proteinase K (1 μg in 5 mM EDTA, 0.5% SDS; incubated at 50°C for 30 min). Finally, samples were incubated at 72°C for an additional 10 min to inactivate the proteinase prior to electrophoresis.
For the two-dimensional (2D) gels a reaction containing labelled attB or mutant attB, cold partner attP and integrase was set up as for the supershift analysis, loaded onto a non-denaturing gel and subjected to electrophoresis. The lane was then excised and immersed in 5 ml of proteinase K buffer (50 μg/ml proteinase K, 5 mM EDTA pH 8.0, 0.5% SDS) at 50°C for 30 min. The gel slice was then laid over a second non-denaturing gel containing one large well to accommodate the gel slice and one small well for the markers. The marker lane consisted of an identical reaction to that on the gel slice and preformed previously (Fig. 2, lane 4) only omitting the first electrophoresis step. The proteinase K treated DNAs were separated, the gel was dried and the radioactivity detected by a Molecular Dynamics phosphorimager.
Figure 2. Susceptibility of complexes to subtilisin and release of cleaved substrates. (A) Representation of the two radiolabelled substrates attB and attP. The bases filled in by the Klenow reaction are shown in grey and those in bold are radiolabelled. The black arrows indicate the proposed positions of cleavage by integrase and concomitant formation of a covalent bond to the recessed 5' ends. The grey arrows on the attB sequence are the positions of cleavage by StyI. (B) Disruption of complexes by subtilisin (lanes 4 and 6) and sizing of the cleaved substrates using StyI cut attB as a marker (lane 5). Labelled attB (lanes 1–5) and attP (lanes 6–9) and cold partner fragments (192 bp attP, lanes 3 and 4; 193 bp attB, lanes 6 and 7) were used as substrates for integrase (lanes 2–4 and 6–8).
Recombination assays were performed exactly as described previously (12). 150 ng of each plasmid substrate were mixed in R buffer (10 mM Tris pH 7.5, 1 mM EDTA pH 8, 100 mM NaCl, 5 mM DTT, 5 mM spermidine, 4.5% glycerol and 0.5 mg/ml bovine serum albumin). A 2-fold dilution series of integrase was used for each plasmid pair containing final concentrations of 0, 100, 50, 25 and 13 nM integrase. Reactions were incubated at 30°C for 30 min prior to heating (70°C, 10 min), restriction with HindIII and then electrophoresed at 3 V/cm for 16 h. Bands were visualized by staining with ethidium bromide.
Purification of integrase
Wild-type and S12A mutant integrases were purified using E.coli BL21-SI as the expression host (Life Technologies). pHS62 (10) encoding the wild-type integrase and its derivative, pMSX6 encoding the S12A mutant integrase were introduced into BL21-SI by transformation and plated on LB-0N agar (i.e. NaCl-free LB; 10 g/l tryptone, 5 g/l yeast extract) containing carbenicillin at 50 μg/ml. Overnight cultures grown at 20°C were used to inoculate fresh LB-0N/carbenicillin and cells were grown at 20°C to late log phase (OD600nm = 0.8). Expression was induced by addition of 200 mM NaCl (final concentration) and continued incubation overnight. Pellets were resuspended in 20 ml of 20 mM Tris, 2 mM DTT, 2 mM MgCl2, 150 mM NaCl, 10% sucrose pH 7.8, and sonicated on ice using 5 x 15 s pulses interspersed with 30 s intervals on ice. Soluble cytoplasm was collected after centrifugation at 36 000 g for 30 min. Benzonase (Novagen) was then added (1 U/ml) and was incubated at room temperature for 45 min. The solution was loaded directly onto a HiPrep Heparin 16/10 column (Amersham Biosciences) pre-equilibrated in TED pH 7.8 (20 mM Tris, 2 mM EDTA, 1 mM DTT), followed by washing until absorbance reached baseline. The column was then subjected to a salt wash using 15% buffer B (TED + 2 M NaCl pH 7.8). A gradient was then applied from 15 to 35% buffer B over 150 ml. Fractions containing integrase were pooled and protein was precipitated using (NH4)2SO4 to a saturation of 60% on ice. Precipitate was collected by centrifugation at 36 000 g for 30 min. The pellet was re-dissolved in 4 ml of TED + 1 M NaCl pH 7.8, and applied to a Superdex S200 26/60 gel filtration column (Amersham Biosciences) pre-equilibrated in TED + 400 mM NaCl pH 7.8. Fractions containing integrase were pooled, and kept as aliquots in 50% glycerol at –70°C until required. Protein was essentially pure as assessed by SDS–PAGE, and the final yield was 4 mg/l of culture.
RESULTS
Identification of a covalent Int-DNA intermediate and cleaved att sites
In reactions driven by C31 integrase, recombination occurs at the 5'TT core sequence (21,22). Based on the mechanism of other serine recombinases, it is likely that C31 integrase cleaves 3' of this sequence and forms a covalent attachment via the active site serine at position 12 to the recessed 5' end (23,24). Both the parental attB and attP sites are predicted to undergo concerted cleavage prior to 180° rotation of one pair of half sites relative to the other to configure the DNA into a recombinant format and, when there is complementary base pairing, joining of the DNA backbone. We have shown previously that if sites are used where the sequences involved in crossover are non-complementary, such as attB x attPTA containing a T to A mutation within the core sequence of attP, ligation of the backbone is severely inhibited and integrase iterates the strand exchange to reform the parental substrates (12). We set out to isolate both the synaptic complex and the integrase covalently attached to a cleaved att site. We expect to obtain both of these intermediates in reactions with wild-type sites, attB x attP, and with the sites containing a mismatched core sequence, attB x attPTA.
In reactions containing integrase and two DNA fragments each containing an att site but only one of which is labelled, we aimed to trap reaction intermediates in non-denaturing gels. To facilitate this analysis we have identified buffer conditions (BB) that are sub-optimal for recombination and in which intermediates in recombination are greatly enriched compared to the reaction in optimal conditions (R buffer; see Materials and Methods for the composition of BB and R buffer and Supplementary Material Fig. 1 for band shifts performed under these two conditions). In a reaction with wild-type sites, in addition to the complexes attributable to free and shifted attB, attL and attR, two major supershifted complexes were clearly visible in addition to the shifted attL and attR products formed during the reaction (Fig. 1). In subsequent experiments described below these were shown to be covalent complexes and synaptic complexes and hence are labelled CC and SC, respectively. With the mismatched sites no shifted or free attL and attR fragments were observed, but both CC and SC were formed (Fig. 1). Treatment of the complexes with subtilisin removed all of the shifted complexes including CC and SC and yielded free attB, attL and attR with the wild-type sites and free attB with the mismatched sites (Fig. 2). In addition, a band smaller than attB was visible, which we propose is the cleaved attB, derived from the covalently attached integrase to the attB site during recombination (Fig. 2).
Figure 1. Detection of recombination intermediates by C31 integrase and susceptibility to proteinase K. Labelled attB fragment (lanes 5–11) was incubated without (lane 5) or with integrase (lanes 6–11) with no further additions (for 30 min and 2 h; lanes 6 and 7, respectively) or with 192 bp cold attP (for 30 min and 2 h; lanes 8 and 9, respectively) or 192 bp cold attPTA (for 30 min and 2 h; lanes 10 and 11, respectively) under suboptimal conditions for recombination. The shifted attB fragment in the presence of integrase alone is labelled Int:attB II (see also Fig. 4). The positions of the recombination products either free (attL in lane 1; attR in lane 3) or complexed with integrase (Int:attL in lane 2 and Int:attR in lane 4) are also shown. When attB, integrase and the cold partner fragment were present, supershifts (labelled CC and SC) and recombination products (free and shifted) were observed in addition to free attB and shifted attB.
In order to determine the size of this putatively cleaved attB site and to show that attP is also cleaved, reactions with either labelled attB or labelled attP were treated with subtilisin to remove the bound integrase and the fragments were subjected to electrophoretic separation against a marker fragment, i.e. labelled attB cut with StyI (Fig. 2). StyI cuts attB to give two 36 bp fragments each with a four-base overhang. Integrase-mediated cleavage of the labelled attB fragment should yield two 37 bp fragments, both containing a 2 bp 3' overhang and a serine residue, whilst cleavage of the labelled attP fragment should yield two fragments, 35 and 38 bp, both containing a 2 bp 3' overhang and a serine residue (Fig. 2A). The observed bands obtained after treatment of recombination reactions with subtilisin, were consistent with these predictions; in addition to free attL and attR, a band migrating slower than the StyI cut attB was obtained when attB was labelled, and when attP was labelled, two bands, one slower and one faster than StyI cut attB, were observed (Fig. 2B).
Thus, attP and attB are cleaved by integrase, presumably concomitant with the formation of a covalent bond via the catalytic serine residue. To identify which of the two major supershifted bands in our gels contained the covalent complexes, we performed a 2D analysis on the reaction components (Fig. 3). An integration reaction was carried out containing labelled attB, attP and integrase, run in a non-denaturing gel (as shown in Fig. 1, lane 8), the lane isolated as a gel slice, treated with proteinase K and then run in a second dimension at 90° to the first, again in a non-denaturing gel. We expected that only one of the two major supershifted complexes would release the cleaved attB fragment. In order to provide markers for the cleaved attB fragment, the free attB and the recombinant products attL and attR, an identical recombination reaction was performed and treated with proteinase K but without separation in the first dimension (as shown in Fig. 2, lane 4). From this 2D gel analysis it was clear that only one of the bands, CC, contained cleaved attB intermediates (Fig. 3). The other major complex, SC, only released free attB fragment on treatment with proteinase K.
Figure 3. Identification of the covalently bound cleaved intermediate of attB. A reaction containing labelled attB, integrase and 192 bp cold partner attP was loaded on a non-denaturing gel (as shown for lane 9, Fig. 1), the lane excised and incubated with proteinase K. The gel slice was then inserted horizontally along the top of a second non-denaturing gel and subjected to electrophoresis. On the same gel a reaction containing the same components, i.e. labelled attB, integrase, 192 bp cold partner attP was incubated, treated with proteinase and then run in a marker lane (M) next to the gel slice. No correction was made for the additional gel matrix from the gel slice so there is a shift in mobility of free attB, cleaved attB and attL and attR (shown with arrows) compared to the marker lane.
Detection of a synapse by C31 integrase
In order to provide evidence that SC is a synaptic complex, we employed a mutant integrase containing an S12A mutation, which renders the integrase catalytically inactive. Integrase S12A should be able to form a synapse but should be incapable of forming the cleaved intermediates. The S12A mutant protein was purified, during which it behaved almost identically to the wild-type protein. The affinities of the S12A protein for wild-type attP and attB sites were assayed and found to be 2-fold lower than the wild-type (Fig. 4A). In binding assays with integrase S12A in the presence of a labelled attB fragment and the 200 bp cold attP fragment, a major complex with the same mobility as the designated SC complex was observed in addition to the shifted attB site (Fig. 4B; compare lanes 5 and 11). We further demonstrated that the mobility of the SC complex can be increased or decreased using small or larger cold fragments (Fig. 4B). This demonstrates that formation of the complex is dependent on integrase and the cold fragment. Despite the lower affinities for the att sites these synaptic complexes were more abundant compared to those seen with the wild-type integrase (Fig. 4B). The accumulation of the putative synaptic complexes with the S12A mutant is consistent with the inability of the mutant integrase to process the synapse into cleaved att sites.
Figure 4. Interaction with attP and attB by the integrase mutant S12A. (A) Affinity of IntS12A for attB and attP compared to the wild-type integrase. The major complexes are labelled ‘Int:attB II’ and ‘Int:attP II’, the minor complexes ‘Int:attB I’ and Int:attP I’ . (B) Supershift complexes formed with Int S12A. Labelled attB was incubated with wild-type (lanes 2–9) and mutant integrases (lanes 10–15) in the presence of no cold attP (lanes 2 and 3), 75 bp cold attP (lanes 4, 7, 10 and 13), 94 bp cold attP (lanes 5, 8, 11 and 14), or 192 bp cold attP (lanes 6, 9, 12 and 15) and then untreated (lanes 2, 4–6 and 10–12) or treated (lanes 3, 7–9 and 13–15) with proteinase K. The mobilities of the different complexes and recombination products changes with the size of the cold attP fragment added. The SC and CC complexes are shown by asterisks and filled circles, respectively. The shifted attL and/or attR fragments are shown by open circles and the recombinant products attL and attR are shown by filled and empty triangles, respectively.
Spacing mutants in attB are blocked in DNA cleavage but can form a synapse
The identification of intermediates enabled dissection of the recombination reaction into different stages, i.e. synapsis, cleavage of substrates with concomitant formation of a covalent complex and product formation. Thus information on the mechanism of the integrase reaction could be gained by elucidation of the stage at which the reaction is blocked with defective att sites or mutant integrases. We therefore isolated and analysed a particular group of attB mutants, containing insertions or deletions that were completely inactive for recombination.
Using a reporter system, we screened for mutants of attB that were defective in recombination. A plasmid, pRT504, was constructed containing attP and attB flanking the lacZ gene. Co-transformation of E.coli with pRT504 and pHS62, an integrase expression construct, resulted in 100% white colonies in the presence of X-Gal indicating efficient recombination and separation of lacZ from the plasmid replicon. To obtain mutants in either attP or attB, pRT504 was passaged through the mutator strain XL1Red (Stratagene) and the pooled plasmids used in transformations with pHS62. Blue colonies containing non-recombined plasmids were isolated and the attP and attB sites were sequenced. Three types of mutants were isolated; mutants in the attB site contained a deletion of a single base pair on either side of the core sequence, i.e. C to the left of the core sequence and G to the right (Fig. 5A), or a mutation was found in attP with a deleted T residue from the string of three Ts that includes the core sequence (data not shown). No further work was performed with the mutant attP site. The attB mutant sites containing C to the left of the core sequence and G to the right of the core were subcloned into pGEM7 to form pRT518 and pRT519, respectively. These plasmids were used as substrates with attP in vitro and no recombination was observed (Fig. 5B). These deletion mutants imply that there is a strict spacing requirement, presumably between sequences within attB that are recognized by integrase. We proceeded to investigate this further by generating mutants in attB with single insertions at the attB core sequence. pMSX7 contained an additional C in the run of Cs on the left side of the core sequence and pHS2116 contained an additional G on the right side. We also generated pHS2110, which contained an additional 5 bp on either side of the core sequence. None of these attB sites were active in vitro. Affinity assays of integrase with pMSX7, pHS2116 (Fig. 5C) and pHS2110 (data not shown) indicated little change in affinity by integrase for the attB site.
Figure 5. Insertions or deletions in attB abolish recombination but have little effect on affinity of integrase. (A) Sequence of attB and the mutant attB sites used in this study. The plasmids that encode the sites are also shown. (B) Recombination reactions by wild-type and mutant sites. Reactions containing two supercoiled plasmids, one encoding attP (pRT700 or pRT702) and the other encoding attB (wild-type or mutant sites) were incubated for 1 h at 30°C with varying integrase concentrations, heat treated and then digested with HindIII. Linearized parental plasmids and fragments containing recombination products were separated by agarose gel electrophoresis. (C) Affinity of wild-type integrase for the mutant sites attBMSX7 and attB2116.
The complexes formed by the mutant sites, attBMSX7 and attB2116 were studied with a view to understanding which stage of the recombination reaction was blocked. attBMSX7and attB2116 behaved similarly in the supershift assays, both giving an aberrant pattern of complexes compared to the behaviour of wild-type attB (Fig. 6A). The abundance of the covalent complex, CC, was very much reduced whilst the abundance of the putative synapse, SC, was consistently enhanced compared to the attB x attP reaction. We also observed an enhancement of the faint bands (Int:attB I) that run just below the shifted att complexes (Int:attB II). After proteinase K treatment these complexes disappeared leaving just free attB or attP and a small amount of the cleaved att sites (Fig. 6B). 2D gel analysis was performed on the complexes obtained with attB2116 in order to ascertain which complexes contained cleaved attB fragments. This confirmed that the CC intermediate did indeed contain some cleaved attB and the SC band did not (Fig. 6C). This gel also showed that the faint Int:attB I band also contained some cleaved attB fragment. The nature of this band is discussed further below. Our data suggested that the insertion mutants bind to integrase normally and can form a synapse with attP, but the next step in recombination, i.e. cleavage and strand exchange is severely reduced.
Figure 6. Insertion mutants attBMSX7 and attB2116 can synapse with attP and integrase but are inhibited in strand cleavage. (A) Enrichment of the SC complex in reactions containing 194 bp attBMSX7 and attB2116 as cold partner fragments. Labelled attP was incubated in the absence (lane 1) or presence of wild-type integrase (lanes 2–9) or IntS12A (lanes 10 and 11). Partner fragments (200 bp fragments) were added as shown and either not treated further or treated with protease. (B) Labelled substrate attB2116 also accumulated SC. (C) 2D gel analysis of the intermediates obtained with labelled attB2116. The experiment was performed as described for Figure 3.
attL and attR cannot form a synapse with the mutant integrase, S12A
We reported previously that wild-type integrase could not form supershifted complexes with non-permissive pairs of recombination sites (i.e. sites that do not recombine to form products) in vitro. As the nature of the supershifted complexes were not characterized at that time, it was not possible to draw any conclusion as to the stage at which the recombination process was blocked. We have revisited this experiment using the integrase S12A mutant, which can only proceed in the reaction as far as synapsis. As the synaptic complexes accumulate with this protein, this assay is the most sensitive method we have at present in vitro for detection of the synapse. Using the same conditions for synapsis of attP and attB we did not detect formation of complexes with combinations of non-permissive att sites that were indicative of synapse formation (Fig. 7).
Figure 7. Use of IntS12A to detect SC with all combinations of att sites. AttB, attP, attL and attR were labelled and incubated without (lanes 1, 6, 11 and 16) or with (lanes 2–5, 7–10, 12–15 and 17–20) integrase S12A and all four 200 bp partner fragments as shown. The positions of the free probe (attP, attB, attR or attL), the shifted att sites (Int:att) and the SC are shown with arrows. Only combinations of attP and attB generated SC.
DISCUSSION
The data presented here provide compelling evidence that the C31 integrase driven recombination reaction is mechanistically similar to that of the resolvase/invertases. We have identified complexes containing covalently attached C31 integrase and cleaved att sites. The sizes of the cleaved att sites are consistent with cleavage at the crossover sequence in both attP and attB. We also showed that both attB and attP are cleaved in the same reaction. Previously, we presented indirect evidence that C31 integrase can synapse and initiate strand exchange under circumstances when the recombinants cannot join due to mismatched crossover sequences (12). These circumstances could be due to a mismatch in the crossover sequence or due to the formation of an incorrect or ‘anti-parallel’ synapse between wild-type att sites. When the core sequence contained a mismatch, integrase generated catenated substrates from attB and attP indicating that integrase could iterate the rotation of one pair of half sites until the parental sites can be rejoined. In this paper we provide direct evidence that the mismatched sites do indeed synapse and cleave the DNA. The covalently attached protein–DNA complex was identified using 2D gel analysis in which complexes were separated, treated with proteinase K and then run in a second dimension. This identified directly the covalent intermediate in the recombination reaction.
We also identified in non-denaturing gels, synaptic complexes containing integrase, attP and attB. We used a mutant integrase, S12A, with the putative active site serine changed to an alanine in the presence of attP and attB to demonstrate the presence of synaptic complexes. A high molecular weight complex was observed that accumulated with the S12A mutant but was present in very low abundance with the wild-type integrase. We propose that with wild-type integrase there is rapid progression to att site cleavage, a reaction that cannot occur with the mutant integrase. The formation of this complex with either the wild-type or mutant integrase is dependent on the presence of attP and attB in the reaction.
The identification of intermediates in the recombination reaction enabled dissection of the recombination reaction into its different stages, a useful tool in the analysis of mutants in the components of recombination. We therefore isolated and studied defective attB sites containing either a single base pair insertion or deletion flanking the crossover site. The attB insertion mutants were shown to have very similar affinities for integrase as wild-type attB, produced very little cleaved intermediate and produced an abundance of the putative synaptic complexes. Thus, the block in recombination appeared to be largely in the progression from synapse to cleaved intermediates. In these reactions cleavage of the partner site, wild-type attP, was also inhibited suggesting that cleavage is concerted, as in the resolvase/invertases. Furthermore, close scrutiny of the complexes obtained with the mutant attB suggests that when strand exchange is initiated, there is collapse of the recombination complex. This is suggested by the appearance of cleaved mutant attB2116 that co-migrated with one of the shifted complexes (Int:attB I) obtained with the attB and integrase alone. Previously, we showed that the mobility shifts by integrase with all of the att sites produce a major complex (labelled Int:att II) and a minor band running just ahead (Int:att I). We proposed that the major band is a dimer of integrase bound to an att site. Here we showed that the concentration of the minor faster moving band is enhanced when the mutant attB sites are used either as a probe or as the cold fragment (Fig. 6). Moreover, in the 2D gel using attB2116 as a probe, the DNA component of this complex appears to contain both intact and cleaved attB. We propose that this minor complex contains an integrase monomer bound to an att site. With the mutant attB2116 this band may also contain cleaved half sites with an integrase monomer covalently attached. Band shift assays with integrase and a half site (attB cut with StyI) migrate with approximately the same mobility as this band (data not shown). Thus, it would appear that insertion of a single base pair in attB does not inhibit synapse formation but imposes a significant block on the formation of the cleaved intermediates and leads to greater protein–protein instability during strand exchange. Taken together, the insertion/deletion mutants indicate that there is a spacing requirement between sequences in attB recognized by integrase that positions the catalytic serine correctly with the scissile phosphate.
Previous work has shown that integrase and attP and attB can form two types of synapse; a ‘parallel’ synapse, which after strand exchange leads to the formation of correct recombinant products attL and attR, and an ‘anti-parallel’ synapse, which cannot form products after strand exchange due to mismatches at the crossover site. Unfortunately, we were not able to resolve these two types of synapse in our experiments, even using the S12A mutants. However, using this mutant integrase we were able to confirm the specificity of the formation of the synapse that we had hinted at previously (11). In the presence of S12A only attP and attB can form a synapse that we can easily visualize in non-denaturing gels. It may be that other synapses can form but they are too unstable to persist during electrophoresis. This synapse specificity supports our hypothesis that integrase takes on different conformations on binding to attP and attB and only these conformations present interfaces that permit the stable synapse to form. Of course for excision of the phage genome, attL and attR must be able to synapse. We have proposed that integrase employs a phage encoded Xis or directionality factor to switch the reaction from integration to excision. We have not yet isolated Xis from C31 but genes that encode these functions from other systems have been identified (25–27). These studies indicate that Xis is a small protein and highly variable in sequence. It is not obvious where Xis may act. It could be interacting with DNA or with integrase or both. However Xis acts it seems likely that the end result is to enable integrase to form a stable synapse with attL and attR sites.
SUPPLEMENTARY MATERIAL
ACKNOWLEDGEMENT
This work was funded by a BBSRC project grant, EGH16108
REFERENCES
Smith,M.C.M. and Thorpe,H.M. (2002) Diversity in the serine recombinases. Mol. Microbiol., 44, 299–307.
Kunkel,B., Losick,R. and Stragier,P. (1990) The Bacillus subtilis gene for the development transcription factor sigma K is generated by excision of a dispensable DNA element containing a sporulation recombinase gene. Genes Dev., 4, 525–535.
Christiansen,B., Brondsted,L., Vogensen,F.K. and Hammer,K. (1996) A resolvase-like protein is required for the site-specific integration of the temperate lactococcal bacteriophage TP901-1. J. Bacteriol., 178, 5164–5173.
Carrasco,C.D., Ramaswamy,K.S., Ramasubramanian,T.S. and Golden,J.W. (1994) Anabaena xisF gene encodes a developmentally regulated site-specific recombinase. Genes Dev., 8, 74–83.
Wang,H. and Mullany,P. (2000) The large resolvase TndX is required and sufficient for integration and excision of derivatives of the novel conjugative transposon Tn5397. J. Bacteriol., 182, 6577–6583.
Crellin,P.K. and Rood,J.I. (1997) The resolvase/invertase domain of the site-specific recombinase TnpX is functional and recognizes a target sequence that resembles the junction of the circular form of the Clostridium perfringens transposon Tn4451. J. Bacteriol., 179, 5148–5156.
Matsuura,M., Noguchi,T., Yamaguchi,D., Aida,T., Asayama,M., Takahashi,H. and Shirai,M. (1996) The sre gene (ORF469) encodes a site-specific recombinase responsible for integration of the R4 phage genome. J. Bacteriol., 178, 3374–3376.
Ghosh,P., Kim,A.I. and Hatfull,G.F. (2003) The orientation of mycobacteriophage Bxb1 integration is solely dependent on the central dinucleotide of attP and attB. Mol. Cell, 12, 1101–1111.
Kim,A.I., Ghosh,P., Aaron,M.A., Bibb,L.A., Jain,S. and Hatfull,G.F. (2003) Mycobacteriophage Bxb1 integrates into the Mycobacterium smegmatis groEL1 gene. Mol. Microbiol., 50, 463–473.
Thorpe,H.M. and Smith,M.C.M. (1998) In vitro site-specific integration of bacteriophage DNA catalyzed by a recombinase of the resolvase/invertase family. Proc. Natl Acad. Sci. USA, 95, 5505–5510.
Thorpe,H.M., Wilson,S.E. and Smith,M.C.M. (2000) Control of directionality in the site-specific recombination system of the Streptomyces phage phiC31. Mol. Microbiol., 38, 232–241.
Smith,M.C.A., Till,R. and Smith,M.C.M. (2004) Switching the polarity of a bacteriophage integration system. Mol. Microbiol., 51, 1719–1728.
Groth,A.C., Olivares,E.C., Thyagarajan,B. and Calos,M.P. (2000) A phage integrase directs efficient site-specific integration in human cells. Proc. Natl Acad. Sci. USA, 97, 5995–6000.
Breuner,A., Brondsted,L. and Hammer,K. (2001) Resolvase-like recombination performed by the TP901–1 integrase. Microbiology, 147, 2051–2063.
Olivares,E.C., Hollis,R.P. and Calos,M.P. (2001) Phage R4 integrase mediates site-specific integration in human cells. Gene, 278, 167–176.
Hatfull,G. and Grindley,N. (1988) Resolvases and DNA-invertases: a family of enzymes active in site-specific recombination. In Kucherlapati,R. and Smith,G. (eds), Genetic Recombination. American Society for Microbiology, Washington, pp. 357–396.
Johnson,R.C. and Bruist,M.F., (1989) Intermediates in Hin-mediated DNA inversion: a role for Fis and the recombinational enhancer in the strand exchange reaction. EMBO J., 8, 1581–1590.
Sambrook,J. and Russell,D.W. (2001) Molecular Cloning: A Laboratory Manual, 3rd Edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York.
Rogers,M., Ekaterinaki,N., Nimmo,E. and Sherratt,D. (1986) Analysis of Tn7 transposition. Mol. Gen. Genet., 205, 550–556.
Shapira,S.K., Chou,J., Richaud,F.V. and Casadaban,M.J. (1983) New versatile plasmid vectors for expression of hybrid proteins coded by a cloned gene fused to lacZ gene sequences encoding an enzymatically active carboxy-terminal portion of beta-galactosidase. Gene, 25, 71–82.
Combes,P., Till,R., Bee,S. and Smith,M.C.M. (2002) The streptomyces genome contains multiple pseudo-attB sites for the (phi)C31-encoded site-specific recombination system. J. Bacteriol., 184, 5746–5752.
Kuhstoss,S. and Rao,R.N. (1991) Analysis of the integration function of the streptomycete bacteriophage phic31. J. Mol. Biol., 222, 897–908.
Grindley,N.D.F. (2002) The movement of Tn3-like elements: transposition and cointegrate resolution. In Craig,N.L., Craigie,R., Gellert,M. and Lambowitz,A.M. (eds), Mobile DNA II. ASM Press, Washington, DC, pp. 272–302.
Johnson,R.C. (2002) Bacterial site-specific DNA inversion systems. In Craig,N.L., Craigie,R., Gellert,M. and Lambowitz,A.M. (eds), Mobile DNA II. ASM Press, Washington, DC, pp. 230–271.
Bibb,L.A. and Hatfull,G.F. (2002) Integration and excision of the Mycobacterium tuberculosis prophage-like element, phiRv1. Mol. Microbiol., 45, 1515–1526.
Breuner,A., Brondsted,L. and Hammer,K. (1999) Novel organisation of genes involved in prophage excision identified in the temperate lactococcal bacteriophage TP901-1. J. Bacteriol., 181, 7291–7297.
Ramaswamy,K.S., Carrasco,C.D., Fatma,T. and Golden,J.W. (1997) Cell-type specificity of the Anabaena fdxN-element rearrangement requires xisH and xisI. Mol. Microbiol., 23, 1241–1249.(Matthew C. A. Smith, Rob Till, Kevin Bra)
*To whom correspondence should be addressed at Institute of Medical Sciences, University of Aberdeen, Foresterhill, Aberdeen AB25 2ZD, UK. Tel: +44 1224 555739; Fax: +44 1224 555844; Email: maggie.smith@abdn.ac.uk
ABSTRACT
The Streptomyces phage C31 encodes an integrase belonging to the serine recombinase family of site-specific recombinases. The well studied serine recombinases, the resolvase/invertases, bring two recombination sites together in a synapse, and then catalyse a concerted four-strand staggered break in the DNA substrates whilst forming transient covalent attachments with the recessed 5' ends. Rotation of one pair of half sites by 180° relative to the other pair occurs, to form the recombinant configuration followed by ligation of the DNA backbone. Here we address the nature of the recombination intermediates formed by C31 integrase when acting on its substrates attP and attB. We have identified intermediates containing integrase covalently attached to cleaved DNA substrates, attB or attP, by analysis of complexes in gels and after treatment of these complexes with proteinases. Using a catalytically inactive integrase mutant, S12A, the synaptic complexes containing integrase, attP and attB were identified. Furthermore, we have shown that attB mutants containing insertions or deletions are blocked in recombination at the stage of strand cleavage. Thus, there is a strict spacing requirement within attB, possibly for correct positioning of the catalytic serine relative to the scissile phosphate in the active site. Finally, using integrase S12A we confirmed the inability of attL and attR or other combinations of sites to form a stable synapse, indicating that the directionality of integrative recombination is determined at synapsis.
INTRODUCTION
Many temperate bacteriophages encode systems that enable site-specific recombination of the phage genome with the host chromosome, bringing about integration during the establishment of lysogeny and excision during induction into lytic growth. The recombinase can belong either to a family of proteins related to phage integrase (the tyrosine recombinases) or to an evolutionary and mechanistically unrelated family which includes the resolvase/invertases (the serine recombinases) (1). The integrases belonging to the serine recombinase class form part of a diverse group of proteins that range in size from about 450 amino acids to greater than 750 amino acids, and all contain an N-terminal domain similar to the resolvase/invertase catalytic domain (1). Because of their much greater mass compared to the resolvase/invertases this group of proteins have been called the large serine recombinases and their biological functions include deletion (SpoIVCA in Bacillus subtilis sporulation and XisF in heterocyst development in cyanobacteria), transposition (TnpX and TndX in Tn4451 and Tn5397, respectively) and phage integration (shown originally in the lactococcal phage TP901-1 and Streptomyces phage R4) (2–7). The two systems that have been studied in most detail, however, are Streptomyces phage C31 integrase and mycobacteriophage Bxb1 integrase (8–12). A CLUSTALW alignment of approximately 30 serine integrases revealed conserved regions outside the catalytic domain but their function in recombination is still largely uncharacterized (1).
Phage integrases act on pairs of recombination sites that are different in sequence; attP and attB for integration and attL and attR for excision. For integration of C31 attP and attB are both short, <50 bp (11,13), and similar length recombination sites have also been reported for other systems that use serine integrases (9,14,15). All the recombination sites characterized to date that are acted on by large serine recombinases contain a stretch of identical sequence between 2 and 12 bp at which recombination occurs. Both C31 and Bxb1 integrases catalyse integrative recombination in vitro between attP and attB in the absence of other proteins or high energy cofactors and with supercoiled or linear substrates (9,11).
The mechanism of recombination by the large serine recombinases is thought to resemble that of the resolvase/invertases. Biochemical studies have demonstrated that the resolvase/invertases act via a concerted, four-strand cleavage and rejoining mechanism. A staggered break is generated at a 2 bp core sequence and transient phospho-serine linkages are formed to the recessed 5' ends (16,17). A 180° rotation of one pair of half sites is then thought to occur to configure the DNA molecules into the recombinant format prior to ligation of the DNA backbone. Several of the essential catalytic residues equivalent to, for example, Arg8, Ser10 and Arg68 using the resolvase nomenclature have been shown to be required for activity in the transposases TnpX and TndX (5,6). In Bxb1 and C31 integrases the proposed catalytic serines are required (8,10). The importance of the 2 bp at the crossover site has been demonstrated recently in both C31 and Bxb1 integrases (8,12). Mutations in these 2 bp can completely switch the polarity of the att sites leading to incorrectly joined recombination products, i.e. the left side of one site becomes joined to the left side of the other site and the reciprocal right side to right side joining. Moreover, in Bxb1 the precise location of the 2 bp staggered break in both attP and attB has also been demonstrated (8). The conclusions from these papers indicate that integrases apparently synapse freely in either of two forms that, if the sequences at the staggered break are complementary, lead to correct or incorrect recombination products. Furthermore, synapsis activates initiation of strand exchange but if the products cannot form due to mismatches, integrase iterates rotation of the half sites leaving a topological footprint but no recombination products.
The directionality of the integrative versus excisive recombination reactions has been studied with purified C31 integrase (11). In vitro C31 integrase cannot recombine pairs of sites other than attP and attB and this has led to the proposal that C31 encodes an Xis or directionality factor that enables excisive recombination between attL and attR. Experiments performed to investigate the mechanism of this site selectivity indicated that, whilst C31 integrase binds attP, attB, attL and attR with approximately equal affinities, putative recombination intermediates (such as those detected in assays containing attP/attB) could not be detected when non-permissive pairs of sites, such as attP/attP, attL/attR, were used in a reaction (11). This implied that the block in recombination between non-permissive sites is at an earlier step, i.e. synapsis.
The aim of this study was to characterize the recombination intermediates generated in the C31 integration reaction and to test the assertion that C31 integrase uses a similar mechanism of recombination as the resolvase/invertases. Using non-denaturing gels we have demonstrated the presence of complexes containing the cleaved intermediates covalently bound to protein. We have also characterized a putative synaptic complex that accumulates in reactions containing a mutant integrase containing an S12A mutation in the catalytic serine residue. Furthermore, mutant attB sites have been created that can apparently synapse but cannot efficiently activate recombination. Finally, we confirmed that combinations of att sites other than attP and attB do not form synaptic complexes.
MATERIALS AND METHODS
Bacterial strains and plasmids
Growth, maintenance and transformation of Escherichia coli were performed as described previously (18). Escherichia coli DH5 was used as the routine host for amplification and preparation of plasmids throughout this study. The methods used for preparation of plasmids and in vitro manipulation of DNA were described previously (18). Plasmids pRT600 and pRT700 containing 50 bp attB and attP sites, respectively, in pGEM7 were constructed previously (11). The screen for mutants in attB and attP was performed as follows: pRT504 contains attP and attB sites in the same orientations flanking an 5 kb fragment encoding the lac operon. pRT504 was constructed by inserting an EcoRI fragment encoding attP into pZMR100 (19) (a plasmid derived from the replication functions of phage ) followed by insertion of a PstI–BamHI fragment encoding attB derived from pHS23 (10) to give pHS40. The 5.8 kb NruI–ScaI fragment encoding the lac operon from pSKS106 (20) was inserted into the unique BamHI site in pHS40 via BglII linkers to form pRT504. This plasmid was introduced into XL1Red according to the manufacturer’s instructions (Stratagene). The transformants from each plate were pooled and plasmids were prepared from an overnight culture and used to transform DH5 containing pHS62 and expressing integrase. Ampicillin, kanamycin resistant colonies that retained expression of ?-galactosidase were picked, re-streaked to confirm the continued expression of ?-galactosidase and the plasmids isolated. The pRT504 derivatives containing the mutated att sites were purified and the att sites sequenced to reveal the mutations. The mutant sites were subcloned into pGEM7 to form pRT518 and pRT519.
Plasmid pMSX7 was constructed by annealing oligonucleotides PRMS74 (5'-AGCTTATCGATTTCGGGAGTACGCGCCCGGGGAGCCCAAGGGGCACGCCCTGGCACCCGCACCGCGGAATTCC-3'; Invitrogen) and PRMS75 (5'-TCGAGGAATTCCGCGGTGCGGGTGCCAGGGCGTGCCCCTTGGGCTCCCCGGGCGCGTACTCCCGAAATCGATA-3') which were then ligated to pGEM7 digested with XhoI and HindIII and introduced into DH5 by transformation. Plasmid pHS2116 was prepared by inserting a PCR fragment amplified using primers B+1G (5'-TCCTCTAGACTCGAGGAATTCCGCGGTGCGGGTGCCAGGGCGTGCCCTTGGgGCTCCCCGGGCGCGTACTC-3') and SP6 with pRT600 as a template, digesting this with XhoI and HindIII and ligating the product with XhoI and HindIII cut pGEM7. pHS2110 was prepared as for pHS2116 except that the primer used was B+ 10 (5'-TCCTCTAGACTCGAGGAATTCCGCGGTGCGGGTGCCAGGGCGTGCCCAGATCTTGTACAGGGCTCCCCGGGCGCGTACTC-3'). Mutations were verified by DNA sequencing.
Construction of pMSX6 containing the S12A mutant of integrase was performed by replacing the NdeI-EcoRI fragment at the 5' end of the int gene in pHS62. PRMS64 (5'-TATGGACACGTACGCGGGTGCTTACGACCGTCAGGCGCGCGAGCGCGAG-3') and PRMS65 (5'-AATTCTCGCGCTCGCGCGCCTGACGGTCGTAAGCACCCGCGTACGTGTCCA-3') were annealed and ligated to pHS62 digested with NdeI and EcoRI. The ligated DNAs were transformed into DH5 and pMSX6 was verified by sequencing.
Preparation of DNA fragments
For the supershift assays the 200 bp cold partner fragments were prepared by PCR amplification with SP6 and T7 primers of pRT600, pRT700, pMSX7 and pHS2116 as substrates for attB, attP, attBMSX7 and attB2116, respectively. The 94 bp cold partner attP site was prepared by PCR amplification of pRT700 with primers PRMS66 (5'-TCTAGACTCGAGGAATTAGT-3') and PRMS67 (5'-TCTCCGGATCCAAGCTTATCG-3'). The PCR templates for attL and attR were prepared from pRT602700, the product of recombination between attB (pRT602) and attP (pRT700). pRT602700 was restricted with ScaI and the resulting fragments were separated by electrophoresis and purified to yield a 2 kb fragment encoding attR and a 3.5 kb fragment encoding attL. PCR amplification of attR and attL was carried out using these fragments as templates and primers SP6 and T7. The resulting 200 bp fragments were separated by agarose gel electrophoresis (1.5% TAE) and purified.
The labelled fragments for all the DNA binding assays were prepared as follows: pRT600, pRT700, pMSX7 or pHS2116 were digested with XhoI and HindIII and the desired fragments of 76, 75, 77 and 77 bp containing attB, attP, attBMSX7 and attB2116, respectively, were purified. Fragments containing attL (71 bp) and attR (74 bp) were prepared by triple digestion of the 200 bp attL and attR cold partner fragments (MspI, HindII, EcoRI for attL and XhoI, HindIII, AatII for attR). The fragments were end-labelled with Klenow fragment in the presence of dCTP and dATP.
DNA binding and recombination assays
DNA affinity and supershift assays were performed as described previously (11). Briefly, assays were performed with 1.5 ng of labelled probe in binding buffer (BB) containing 20 mM Tris–HCl pH 8.0, 0.1 mM EDTA, 50 mM KCl, 5% glycerol, 70 μg/ml of sonicated salmon sperm DNA, 70 μg/ml bovine serum albumin (BSA) and integrase added to final concentrations 0, 66, 33, 16, 8, 4 and 2 nM. Reactions with no integrase contained instead 1 μg of BSA. Reactions were incubated at 30°C for 30 min prior to electrophoresis. Samples were separated on 5% acrylamide gels run at 200 V, 5 W for 2 h. For the supershift assays, 0.9–1.5 ng of labelled probe was mixed with 20 ng of cold partner fragment and 66 nM integrase in BB buffer. Reactions were incubated at 30°C for 2 h prior to electrophoresis. Where samples required proteinase treatment, samples were first heat inactivated by incubation at 72°C for 10 min and then treated either with 0.1 μg of subtilisin (incubated at 30°C for 15 min) or proteinase K (1 μg in 5 mM EDTA, 0.5% SDS; incubated at 50°C for 30 min). Finally, samples were incubated at 72°C for an additional 10 min to inactivate the proteinase prior to electrophoresis.
For the two-dimensional (2D) gels a reaction containing labelled attB or mutant attB, cold partner attP and integrase was set up as for the supershift analysis, loaded onto a non-denaturing gel and subjected to electrophoresis. The lane was then excised and immersed in 5 ml of proteinase K buffer (50 μg/ml proteinase K, 5 mM EDTA pH 8.0, 0.5% SDS) at 50°C for 30 min. The gel slice was then laid over a second non-denaturing gel containing one large well to accommodate the gel slice and one small well for the markers. The marker lane consisted of an identical reaction to that on the gel slice and preformed previously (Fig. 2, lane 4) only omitting the first electrophoresis step. The proteinase K treated DNAs were separated, the gel was dried and the radioactivity detected by a Molecular Dynamics phosphorimager.
Figure 2. Susceptibility of complexes to subtilisin and release of cleaved substrates. (A) Representation of the two radiolabelled substrates attB and attP. The bases filled in by the Klenow reaction are shown in grey and those in bold are radiolabelled. The black arrows indicate the proposed positions of cleavage by integrase and concomitant formation of a covalent bond to the recessed 5' ends. The grey arrows on the attB sequence are the positions of cleavage by StyI. (B) Disruption of complexes by subtilisin (lanes 4 and 6) and sizing of the cleaved substrates using StyI cut attB as a marker (lane 5). Labelled attB (lanes 1–5) and attP (lanes 6–9) and cold partner fragments (192 bp attP, lanes 3 and 4; 193 bp attB, lanes 6 and 7) were used as substrates for integrase (lanes 2–4 and 6–8).
Recombination assays were performed exactly as described previously (12). 150 ng of each plasmid substrate were mixed in R buffer (10 mM Tris pH 7.5, 1 mM EDTA pH 8, 100 mM NaCl, 5 mM DTT, 5 mM spermidine, 4.5% glycerol and 0.5 mg/ml bovine serum albumin). A 2-fold dilution series of integrase was used for each plasmid pair containing final concentrations of 0, 100, 50, 25 and 13 nM integrase. Reactions were incubated at 30°C for 30 min prior to heating (70°C, 10 min), restriction with HindIII and then electrophoresed at 3 V/cm for 16 h. Bands were visualized by staining with ethidium bromide.
Purification of integrase
Wild-type and S12A mutant integrases were purified using E.coli BL21-SI as the expression host (Life Technologies). pHS62 (10) encoding the wild-type integrase and its derivative, pMSX6 encoding the S12A mutant integrase were introduced into BL21-SI by transformation and plated on LB-0N agar (i.e. NaCl-free LB; 10 g/l tryptone, 5 g/l yeast extract) containing carbenicillin at 50 μg/ml. Overnight cultures grown at 20°C were used to inoculate fresh LB-0N/carbenicillin and cells were grown at 20°C to late log phase (OD600nm = 0.8). Expression was induced by addition of 200 mM NaCl (final concentration) and continued incubation overnight. Pellets were resuspended in 20 ml of 20 mM Tris, 2 mM DTT, 2 mM MgCl2, 150 mM NaCl, 10% sucrose pH 7.8, and sonicated on ice using 5 x 15 s pulses interspersed with 30 s intervals on ice. Soluble cytoplasm was collected after centrifugation at 36 000 g for 30 min. Benzonase (Novagen) was then added (1 U/ml) and was incubated at room temperature for 45 min. The solution was loaded directly onto a HiPrep Heparin 16/10 column (Amersham Biosciences) pre-equilibrated in TED pH 7.8 (20 mM Tris, 2 mM EDTA, 1 mM DTT), followed by washing until absorbance reached baseline. The column was then subjected to a salt wash using 15% buffer B (TED + 2 M NaCl pH 7.8). A gradient was then applied from 15 to 35% buffer B over 150 ml. Fractions containing integrase were pooled and protein was precipitated using (NH4)2SO4 to a saturation of 60% on ice. Precipitate was collected by centrifugation at 36 000 g for 30 min. The pellet was re-dissolved in 4 ml of TED + 1 M NaCl pH 7.8, and applied to a Superdex S200 26/60 gel filtration column (Amersham Biosciences) pre-equilibrated in TED + 400 mM NaCl pH 7.8. Fractions containing integrase were pooled, and kept as aliquots in 50% glycerol at –70°C until required. Protein was essentially pure as assessed by SDS–PAGE, and the final yield was 4 mg/l of culture.
RESULTS
Identification of a covalent Int-DNA intermediate and cleaved att sites
In reactions driven by C31 integrase, recombination occurs at the 5'TT core sequence (21,22). Based on the mechanism of other serine recombinases, it is likely that C31 integrase cleaves 3' of this sequence and forms a covalent attachment via the active site serine at position 12 to the recessed 5' end (23,24). Both the parental attB and attP sites are predicted to undergo concerted cleavage prior to 180° rotation of one pair of half sites relative to the other to configure the DNA into a recombinant format and, when there is complementary base pairing, joining of the DNA backbone. We have shown previously that if sites are used where the sequences involved in crossover are non-complementary, such as attB x attPTA containing a T to A mutation within the core sequence of attP, ligation of the backbone is severely inhibited and integrase iterates the strand exchange to reform the parental substrates (12). We set out to isolate both the synaptic complex and the integrase covalently attached to a cleaved att site. We expect to obtain both of these intermediates in reactions with wild-type sites, attB x attP, and with the sites containing a mismatched core sequence, attB x attPTA.
In reactions containing integrase and two DNA fragments each containing an att site but only one of which is labelled, we aimed to trap reaction intermediates in non-denaturing gels. To facilitate this analysis we have identified buffer conditions (BB) that are sub-optimal for recombination and in which intermediates in recombination are greatly enriched compared to the reaction in optimal conditions (R buffer; see Materials and Methods for the composition of BB and R buffer and Supplementary Material Fig. 1 for band shifts performed under these two conditions). In a reaction with wild-type sites, in addition to the complexes attributable to free and shifted attB, attL and attR, two major supershifted complexes were clearly visible in addition to the shifted attL and attR products formed during the reaction (Fig. 1). In subsequent experiments described below these were shown to be covalent complexes and synaptic complexes and hence are labelled CC and SC, respectively. With the mismatched sites no shifted or free attL and attR fragments were observed, but both CC and SC were formed (Fig. 1). Treatment of the complexes with subtilisin removed all of the shifted complexes including CC and SC and yielded free attB, attL and attR with the wild-type sites and free attB with the mismatched sites (Fig. 2). In addition, a band smaller than attB was visible, which we propose is the cleaved attB, derived from the covalently attached integrase to the attB site during recombination (Fig. 2).
Figure 1. Detection of recombination intermediates by C31 integrase and susceptibility to proteinase K. Labelled attB fragment (lanes 5–11) was incubated without (lane 5) or with integrase (lanes 6–11) with no further additions (for 30 min and 2 h; lanes 6 and 7, respectively) or with 192 bp cold attP (for 30 min and 2 h; lanes 8 and 9, respectively) or 192 bp cold attPTA (for 30 min and 2 h; lanes 10 and 11, respectively) under suboptimal conditions for recombination. The shifted attB fragment in the presence of integrase alone is labelled Int:attB II (see also Fig. 4). The positions of the recombination products either free (attL in lane 1; attR in lane 3) or complexed with integrase (Int:attL in lane 2 and Int:attR in lane 4) are also shown. When attB, integrase and the cold partner fragment were present, supershifts (labelled CC and SC) and recombination products (free and shifted) were observed in addition to free attB and shifted attB.
In order to determine the size of this putatively cleaved attB site and to show that attP is also cleaved, reactions with either labelled attB or labelled attP were treated with subtilisin to remove the bound integrase and the fragments were subjected to electrophoretic separation against a marker fragment, i.e. labelled attB cut with StyI (Fig. 2). StyI cuts attB to give two 36 bp fragments each with a four-base overhang. Integrase-mediated cleavage of the labelled attB fragment should yield two 37 bp fragments, both containing a 2 bp 3' overhang and a serine residue, whilst cleavage of the labelled attP fragment should yield two fragments, 35 and 38 bp, both containing a 2 bp 3' overhang and a serine residue (Fig. 2A). The observed bands obtained after treatment of recombination reactions with subtilisin, were consistent with these predictions; in addition to free attL and attR, a band migrating slower than the StyI cut attB was obtained when attB was labelled, and when attP was labelled, two bands, one slower and one faster than StyI cut attB, were observed (Fig. 2B).
Thus, attP and attB are cleaved by integrase, presumably concomitant with the formation of a covalent bond via the catalytic serine residue. To identify which of the two major supershifted bands in our gels contained the covalent complexes, we performed a 2D analysis on the reaction components (Fig. 3). An integration reaction was carried out containing labelled attB, attP and integrase, run in a non-denaturing gel (as shown in Fig. 1, lane 8), the lane isolated as a gel slice, treated with proteinase K and then run in a second dimension at 90° to the first, again in a non-denaturing gel. We expected that only one of the two major supershifted complexes would release the cleaved attB fragment. In order to provide markers for the cleaved attB fragment, the free attB and the recombinant products attL and attR, an identical recombination reaction was performed and treated with proteinase K but without separation in the first dimension (as shown in Fig. 2, lane 4). From this 2D gel analysis it was clear that only one of the bands, CC, contained cleaved attB intermediates (Fig. 3). The other major complex, SC, only released free attB fragment on treatment with proteinase K.
Figure 3. Identification of the covalently bound cleaved intermediate of attB. A reaction containing labelled attB, integrase and 192 bp cold partner attP was loaded on a non-denaturing gel (as shown for lane 9, Fig. 1), the lane excised and incubated with proteinase K. The gel slice was then inserted horizontally along the top of a second non-denaturing gel and subjected to electrophoresis. On the same gel a reaction containing the same components, i.e. labelled attB, integrase, 192 bp cold partner attP was incubated, treated with proteinase and then run in a marker lane (M) next to the gel slice. No correction was made for the additional gel matrix from the gel slice so there is a shift in mobility of free attB, cleaved attB and attL and attR (shown with arrows) compared to the marker lane.
Detection of a synapse by C31 integrase
In order to provide evidence that SC is a synaptic complex, we employed a mutant integrase containing an S12A mutation, which renders the integrase catalytically inactive. Integrase S12A should be able to form a synapse but should be incapable of forming the cleaved intermediates. The S12A mutant protein was purified, during which it behaved almost identically to the wild-type protein. The affinities of the S12A protein for wild-type attP and attB sites were assayed and found to be 2-fold lower than the wild-type (Fig. 4A). In binding assays with integrase S12A in the presence of a labelled attB fragment and the 200 bp cold attP fragment, a major complex with the same mobility as the designated SC complex was observed in addition to the shifted attB site (Fig. 4B; compare lanes 5 and 11). We further demonstrated that the mobility of the SC complex can be increased or decreased using small or larger cold fragments (Fig. 4B). This demonstrates that formation of the complex is dependent on integrase and the cold fragment. Despite the lower affinities for the att sites these synaptic complexes were more abundant compared to those seen with the wild-type integrase (Fig. 4B). The accumulation of the putative synaptic complexes with the S12A mutant is consistent with the inability of the mutant integrase to process the synapse into cleaved att sites.
Figure 4. Interaction with attP and attB by the integrase mutant S12A. (A) Affinity of IntS12A for attB and attP compared to the wild-type integrase. The major complexes are labelled ‘Int:attB II’ and ‘Int:attP II’, the minor complexes ‘Int:attB I’ and Int:attP I’ . (B) Supershift complexes formed with Int S12A. Labelled attB was incubated with wild-type (lanes 2–9) and mutant integrases (lanes 10–15) in the presence of no cold attP (lanes 2 and 3), 75 bp cold attP (lanes 4, 7, 10 and 13), 94 bp cold attP (lanes 5, 8, 11 and 14), or 192 bp cold attP (lanes 6, 9, 12 and 15) and then untreated (lanes 2, 4–6 and 10–12) or treated (lanes 3, 7–9 and 13–15) with proteinase K. The mobilities of the different complexes and recombination products changes with the size of the cold attP fragment added. The SC and CC complexes are shown by asterisks and filled circles, respectively. The shifted attL and/or attR fragments are shown by open circles and the recombinant products attL and attR are shown by filled and empty triangles, respectively.
Spacing mutants in attB are blocked in DNA cleavage but can form a synapse
The identification of intermediates enabled dissection of the recombination reaction into different stages, i.e. synapsis, cleavage of substrates with concomitant formation of a covalent complex and product formation. Thus information on the mechanism of the integrase reaction could be gained by elucidation of the stage at which the reaction is blocked with defective att sites or mutant integrases. We therefore isolated and analysed a particular group of attB mutants, containing insertions or deletions that were completely inactive for recombination.
Using a reporter system, we screened for mutants of attB that were defective in recombination. A plasmid, pRT504, was constructed containing attP and attB flanking the lacZ gene. Co-transformation of E.coli with pRT504 and pHS62, an integrase expression construct, resulted in 100% white colonies in the presence of X-Gal indicating efficient recombination and separation of lacZ from the plasmid replicon. To obtain mutants in either attP or attB, pRT504 was passaged through the mutator strain XL1Red (Stratagene) and the pooled plasmids used in transformations with pHS62. Blue colonies containing non-recombined plasmids were isolated and the attP and attB sites were sequenced. Three types of mutants were isolated; mutants in the attB site contained a deletion of a single base pair on either side of the core sequence, i.e. C to the left of the core sequence and G to the right (Fig. 5A), or a mutation was found in attP with a deleted T residue from the string of three Ts that includes the core sequence (data not shown). No further work was performed with the mutant attP site. The attB mutant sites containing C to the left of the core sequence and G to the right of the core were subcloned into pGEM7 to form pRT518 and pRT519, respectively. These plasmids were used as substrates with attP in vitro and no recombination was observed (Fig. 5B). These deletion mutants imply that there is a strict spacing requirement, presumably between sequences within attB that are recognized by integrase. We proceeded to investigate this further by generating mutants in attB with single insertions at the attB core sequence. pMSX7 contained an additional C in the run of Cs on the left side of the core sequence and pHS2116 contained an additional G on the right side. We also generated pHS2110, which contained an additional 5 bp on either side of the core sequence. None of these attB sites were active in vitro. Affinity assays of integrase with pMSX7, pHS2116 (Fig. 5C) and pHS2110 (data not shown) indicated little change in affinity by integrase for the attB site.
Figure 5. Insertions or deletions in attB abolish recombination but have little effect on affinity of integrase. (A) Sequence of attB and the mutant attB sites used in this study. The plasmids that encode the sites are also shown. (B) Recombination reactions by wild-type and mutant sites. Reactions containing two supercoiled plasmids, one encoding attP (pRT700 or pRT702) and the other encoding attB (wild-type or mutant sites) were incubated for 1 h at 30°C with varying integrase concentrations, heat treated and then digested with HindIII. Linearized parental plasmids and fragments containing recombination products were separated by agarose gel electrophoresis. (C) Affinity of wild-type integrase for the mutant sites attBMSX7 and attB2116.
The complexes formed by the mutant sites, attBMSX7 and attB2116 were studied with a view to understanding which stage of the recombination reaction was blocked. attBMSX7and attB2116 behaved similarly in the supershift assays, both giving an aberrant pattern of complexes compared to the behaviour of wild-type attB (Fig. 6A). The abundance of the covalent complex, CC, was very much reduced whilst the abundance of the putative synapse, SC, was consistently enhanced compared to the attB x attP reaction. We also observed an enhancement of the faint bands (Int:attB I) that run just below the shifted att complexes (Int:attB II). After proteinase K treatment these complexes disappeared leaving just free attB or attP and a small amount of the cleaved att sites (Fig. 6B). 2D gel analysis was performed on the complexes obtained with attB2116 in order to ascertain which complexes contained cleaved attB fragments. This confirmed that the CC intermediate did indeed contain some cleaved attB and the SC band did not (Fig. 6C). This gel also showed that the faint Int:attB I band also contained some cleaved attB fragment. The nature of this band is discussed further below. Our data suggested that the insertion mutants bind to integrase normally and can form a synapse with attP, but the next step in recombination, i.e. cleavage and strand exchange is severely reduced.
Figure 6. Insertion mutants attBMSX7 and attB2116 can synapse with attP and integrase but are inhibited in strand cleavage. (A) Enrichment of the SC complex in reactions containing 194 bp attBMSX7 and attB2116 as cold partner fragments. Labelled attP was incubated in the absence (lane 1) or presence of wild-type integrase (lanes 2–9) or IntS12A (lanes 10 and 11). Partner fragments (200 bp fragments) were added as shown and either not treated further or treated with protease. (B) Labelled substrate attB2116 also accumulated SC. (C) 2D gel analysis of the intermediates obtained with labelled attB2116. The experiment was performed as described for Figure 3.
attL and attR cannot form a synapse with the mutant integrase, S12A
We reported previously that wild-type integrase could not form supershifted complexes with non-permissive pairs of recombination sites (i.e. sites that do not recombine to form products) in vitro. As the nature of the supershifted complexes were not characterized at that time, it was not possible to draw any conclusion as to the stage at which the recombination process was blocked. We have revisited this experiment using the integrase S12A mutant, which can only proceed in the reaction as far as synapsis. As the synaptic complexes accumulate with this protein, this assay is the most sensitive method we have at present in vitro for detection of the synapse. Using the same conditions for synapsis of attP and attB we did not detect formation of complexes with combinations of non-permissive att sites that were indicative of synapse formation (Fig. 7).
Figure 7. Use of IntS12A to detect SC with all combinations of att sites. AttB, attP, attL and attR were labelled and incubated without (lanes 1, 6, 11 and 16) or with (lanes 2–5, 7–10, 12–15 and 17–20) integrase S12A and all four 200 bp partner fragments as shown. The positions of the free probe (attP, attB, attR or attL), the shifted att sites (Int:att) and the SC are shown with arrows. Only combinations of attP and attB generated SC.
DISCUSSION
The data presented here provide compelling evidence that the C31 integrase driven recombination reaction is mechanistically similar to that of the resolvase/invertases. We have identified complexes containing covalently attached C31 integrase and cleaved att sites. The sizes of the cleaved att sites are consistent with cleavage at the crossover sequence in both attP and attB. We also showed that both attB and attP are cleaved in the same reaction. Previously, we presented indirect evidence that C31 integrase can synapse and initiate strand exchange under circumstances when the recombinants cannot join due to mismatched crossover sequences (12). These circumstances could be due to a mismatch in the crossover sequence or due to the formation of an incorrect or ‘anti-parallel’ synapse between wild-type att sites. When the core sequence contained a mismatch, integrase generated catenated substrates from attB and attP indicating that integrase could iterate the rotation of one pair of half sites until the parental sites can be rejoined. In this paper we provide direct evidence that the mismatched sites do indeed synapse and cleave the DNA. The covalently attached protein–DNA complex was identified using 2D gel analysis in which complexes were separated, treated with proteinase K and then run in a second dimension. This identified directly the covalent intermediate in the recombination reaction.
We also identified in non-denaturing gels, synaptic complexes containing integrase, attP and attB. We used a mutant integrase, S12A, with the putative active site serine changed to an alanine in the presence of attP and attB to demonstrate the presence of synaptic complexes. A high molecular weight complex was observed that accumulated with the S12A mutant but was present in very low abundance with the wild-type integrase. We propose that with wild-type integrase there is rapid progression to att site cleavage, a reaction that cannot occur with the mutant integrase. The formation of this complex with either the wild-type or mutant integrase is dependent on the presence of attP and attB in the reaction.
The identification of intermediates in the recombination reaction enabled dissection of the recombination reaction into its different stages, a useful tool in the analysis of mutants in the components of recombination. We therefore isolated and studied defective attB sites containing either a single base pair insertion or deletion flanking the crossover site. The attB insertion mutants were shown to have very similar affinities for integrase as wild-type attB, produced very little cleaved intermediate and produced an abundance of the putative synaptic complexes. Thus, the block in recombination appeared to be largely in the progression from synapse to cleaved intermediates. In these reactions cleavage of the partner site, wild-type attP, was also inhibited suggesting that cleavage is concerted, as in the resolvase/invertases. Furthermore, close scrutiny of the complexes obtained with the mutant attB suggests that when strand exchange is initiated, there is collapse of the recombination complex. This is suggested by the appearance of cleaved mutant attB2116 that co-migrated with one of the shifted complexes (Int:attB I) obtained with the attB and integrase alone. Previously, we showed that the mobility shifts by integrase with all of the att sites produce a major complex (labelled Int:att II) and a minor band running just ahead (Int:att I). We proposed that the major band is a dimer of integrase bound to an att site. Here we showed that the concentration of the minor faster moving band is enhanced when the mutant attB sites are used either as a probe or as the cold fragment (Fig. 6). Moreover, in the 2D gel using attB2116 as a probe, the DNA component of this complex appears to contain both intact and cleaved attB. We propose that this minor complex contains an integrase monomer bound to an att site. With the mutant attB2116 this band may also contain cleaved half sites with an integrase monomer covalently attached. Band shift assays with integrase and a half site (attB cut with StyI) migrate with approximately the same mobility as this band (data not shown). Thus, it would appear that insertion of a single base pair in attB does not inhibit synapse formation but imposes a significant block on the formation of the cleaved intermediates and leads to greater protein–protein instability during strand exchange. Taken together, the insertion/deletion mutants indicate that there is a spacing requirement between sequences in attB recognized by integrase that positions the catalytic serine correctly with the scissile phosphate.
Previous work has shown that integrase and attP and attB can form two types of synapse; a ‘parallel’ synapse, which after strand exchange leads to the formation of correct recombinant products attL and attR, and an ‘anti-parallel’ synapse, which cannot form products after strand exchange due to mismatches at the crossover site. Unfortunately, we were not able to resolve these two types of synapse in our experiments, even using the S12A mutants. However, using this mutant integrase we were able to confirm the specificity of the formation of the synapse that we had hinted at previously (11). In the presence of S12A only attP and attB can form a synapse that we can easily visualize in non-denaturing gels. It may be that other synapses can form but they are too unstable to persist during electrophoresis. This synapse specificity supports our hypothesis that integrase takes on different conformations on binding to attP and attB and only these conformations present interfaces that permit the stable synapse to form. Of course for excision of the phage genome, attL and attR must be able to synapse. We have proposed that integrase employs a phage encoded Xis or directionality factor to switch the reaction from integration to excision. We have not yet isolated Xis from C31 but genes that encode these functions from other systems have been identified (25–27). These studies indicate that Xis is a small protein and highly variable in sequence. It is not obvious where Xis may act. It could be interacting with DNA or with integrase or both. However Xis acts it seems likely that the end result is to enable integrase to form a stable synapse with attL and attR sites.
SUPPLEMENTARY MATERIAL
ACKNOWLEDGEMENT
This work was funded by a BBSRC project grant, EGH16108
REFERENCES
Smith,M.C.M. and Thorpe,H.M. (2002) Diversity in the serine recombinases. Mol. Microbiol., 44, 299–307.
Kunkel,B., Losick,R. and Stragier,P. (1990) The Bacillus subtilis gene for the development transcription factor sigma K is generated by excision of a dispensable DNA element containing a sporulation recombinase gene. Genes Dev., 4, 525–535.
Christiansen,B., Brondsted,L., Vogensen,F.K. and Hammer,K. (1996) A resolvase-like protein is required for the site-specific integration of the temperate lactococcal bacteriophage TP901-1. J. Bacteriol., 178, 5164–5173.
Carrasco,C.D., Ramaswamy,K.S., Ramasubramanian,T.S. and Golden,J.W. (1994) Anabaena xisF gene encodes a developmentally regulated site-specific recombinase. Genes Dev., 8, 74–83.
Wang,H. and Mullany,P. (2000) The large resolvase TndX is required and sufficient for integration and excision of derivatives of the novel conjugative transposon Tn5397. J. Bacteriol., 182, 6577–6583.
Crellin,P.K. and Rood,J.I. (1997) The resolvase/invertase domain of the site-specific recombinase TnpX is functional and recognizes a target sequence that resembles the junction of the circular form of the Clostridium perfringens transposon Tn4451. J. Bacteriol., 179, 5148–5156.
Matsuura,M., Noguchi,T., Yamaguchi,D., Aida,T., Asayama,M., Takahashi,H. and Shirai,M. (1996) The sre gene (ORF469) encodes a site-specific recombinase responsible for integration of the R4 phage genome. J. Bacteriol., 178, 3374–3376.
Ghosh,P., Kim,A.I. and Hatfull,G.F. (2003) The orientation of mycobacteriophage Bxb1 integration is solely dependent on the central dinucleotide of attP and attB. Mol. Cell, 12, 1101–1111.
Kim,A.I., Ghosh,P., Aaron,M.A., Bibb,L.A., Jain,S. and Hatfull,G.F. (2003) Mycobacteriophage Bxb1 integrates into the Mycobacterium smegmatis groEL1 gene. Mol. Microbiol., 50, 463–473.
Thorpe,H.M. and Smith,M.C.M. (1998) In vitro site-specific integration of bacteriophage DNA catalyzed by a recombinase of the resolvase/invertase family. Proc. Natl Acad. Sci. USA, 95, 5505–5510.
Thorpe,H.M., Wilson,S.E. and Smith,M.C.M. (2000) Control of directionality in the site-specific recombination system of the Streptomyces phage phiC31. Mol. Microbiol., 38, 232–241.
Smith,M.C.A., Till,R. and Smith,M.C.M. (2004) Switching the polarity of a bacteriophage integration system. Mol. Microbiol., 51, 1719–1728.
Groth,A.C., Olivares,E.C., Thyagarajan,B. and Calos,M.P. (2000) A phage integrase directs efficient site-specific integration in human cells. Proc. Natl Acad. Sci. USA, 97, 5995–6000.
Breuner,A., Brondsted,L. and Hammer,K. (2001) Resolvase-like recombination performed by the TP901–1 integrase. Microbiology, 147, 2051–2063.
Olivares,E.C., Hollis,R.P. and Calos,M.P. (2001) Phage R4 integrase mediates site-specific integration in human cells. Gene, 278, 167–176.
Hatfull,G. and Grindley,N. (1988) Resolvases and DNA-invertases: a family of enzymes active in site-specific recombination. In Kucherlapati,R. and Smith,G. (eds), Genetic Recombination. American Society for Microbiology, Washington, pp. 357–396.
Johnson,R.C. and Bruist,M.F., (1989) Intermediates in Hin-mediated DNA inversion: a role for Fis and the recombinational enhancer in the strand exchange reaction. EMBO J., 8, 1581–1590.
Sambrook,J. and Russell,D.W. (2001) Molecular Cloning: A Laboratory Manual, 3rd Edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York.
Rogers,M., Ekaterinaki,N., Nimmo,E. and Sherratt,D. (1986) Analysis of Tn7 transposition. Mol. Gen. Genet., 205, 550–556.
Shapira,S.K., Chou,J., Richaud,F.V. and Casadaban,M.J. (1983) New versatile plasmid vectors for expression of hybrid proteins coded by a cloned gene fused to lacZ gene sequences encoding an enzymatically active carboxy-terminal portion of beta-galactosidase. Gene, 25, 71–82.
Combes,P., Till,R., Bee,S. and Smith,M.C.M. (2002) The streptomyces genome contains multiple pseudo-attB sites for the (phi)C31-encoded site-specific recombination system. J. Bacteriol., 184, 5746–5752.
Kuhstoss,S. and Rao,R.N. (1991) Analysis of the integration function of the streptomycete bacteriophage phic31. J. Mol. Biol., 222, 897–908.
Grindley,N.D.F. (2002) The movement of Tn3-like elements: transposition and cointegrate resolution. In Craig,N.L., Craigie,R., Gellert,M. and Lambowitz,A.M. (eds), Mobile DNA II. ASM Press, Washington, DC, pp. 272–302.
Johnson,R.C. (2002) Bacterial site-specific DNA inversion systems. In Craig,N.L., Craigie,R., Gellert,M. and Lambowitz,A.M. (eds), Mobile DNA II. ASM Press, Washington, DC, pp. 230–271.
Bibb,L.A. and Hatfull,G.F. (2002) Integration and excision of the Mycobacterium tuberculosis prophage-like element, phiRv1. Mol. Microbiol., 45, 1515–1526.
Breuner,A., Brondsted,L. and Hammer,K. (1999) Novel organisation of genes involved in prophage excision identified in the temperate lactococcal bacteriophage TP901-1. J. Bacteriol., 181, 7291–7297.
Ramaswamy,K.S., Carrasco,C.D., Fatma,T. and Golden,J.W. (1997) Cell-type specificity of the Anabaena fdxN-element rearrangement requires xisH and xisI. Mol. Microbiol., 23, 1241–1249.(Matthew C. A. Smith, Rob Till, Kevin Bra)