ABSTRACT
The roles of purified Int and Xis proteins of the conjugative transposon Tn916 in excision of a deletion derivative of the closely related element Tn1545 were investigated. At a low salt concentration (37.5 mM NaCl), Int alone was able to promote limited excision to produce a covalently closed circular form of the transposon, showing that Tn916 Int can catalyze both DNA cleavage and strand exchange. This reaction was stimulated by Xis. At higher salt concentrations (150 mM NaCl), excision by Int alone was reduced to barely detectable levels and Xis was required for excision. The low salt, Xis-stimulated reaction was ~8-fold more efficient than the high salt, Xis-dependent reaction. These results reflect invivo requirements for Int and Xis in excision.
Tn916 (1 -3 ), which encodes resistance to tetracycline, and its close relative, Tn1545 (4 ), which, in addition to tetracycline resistance, also encodes resistance to kanamycin and erythromycin, are members of a class of genetic elements called conjugative transposons (5 -7 ). These elements are found in both Gram-positive and Gram-negative bacteria, and, as their name implies, are able to promote their own transfer from a donor bacterium to a recipient during transposition. They are important determinants of antibiotic resistance in several human pathogens, and have a very broad host range, being able to transfer between bacteria belonging to different species and genera.
During transposition, Tn916 first excises and forms a circle (8 ,9 ). This circular intermediate then transfers to a recipient bacterium using an origin of conjugal transfer that is distinct from the ends of the transposon (10 ). Following transfer, the circular transposon integrates into the DNA of the recipient. Unlike other bacterial transposons, there is no duplication of the target sequence upon integration. Rather, the integrated transposon is usually flanked on one side by DNA belonging to the recipient target and on the other side by a short stretch of DNA carried by the transposon from the donor (9 ,11 ,12 ). These short stretches of DNA are called coupling sequences. Because the ends of Tn916 are rich in A-T base-pairs, and the transposon inserts into DNA that is rich in A-T base-pairs (12 ,13 ), it has been difficult to determine the length of the coupling sequences. In two cases where it was possible to distinguish 5 from 6 bp and 6 from 7 bp, respectively, the coupling sequences were found to be 6 bp (14 ).
It appears that excision is the rate limiting step in conjugative transposition, and that the coupling sequences can be major determinants of excision frequency (15 ,16 ). In a comparison of a series of donors that differed only in the coupling sequences flanking Tn916, Clewell found that the transposition frequency varied over a 1000-fold range. High frequency donors contained detectable levels of circular Tn916, while low frequency donors did not (16 ). Other studies have shown that in donor strains where the transposon is inserted at different sites, the amount of excised circular transposon is proportional to the frequency of conjugative transposition (15 ).
Two genes, close to one end of the transposon, called int and xis, play a role in conjugative transposition (17 -19 ). The int gene is essential for transposition and encodes a protein that is related to the lambda integrase family of site-specific recombinases (20 ). The int genes of Tn916 and Tn1545 differ by a single base, which leads to a conservative amino-acid difference between the two proteins (3 ,18 ,21 ). The ends of the two transposons are identical (11 ,22 ), and their int genes can complement each other (23 ,24 ).
The xis gene encodes a small basic protein (3 ,18 ). Measurements of the loss of Tn916 from Escherichia coli (21 ) and conjugative transposition in Enterococcus faecalis (19 ) suggest that xis is required for excision and transposition of Tn916. However, excision of Tn1545 by both Tn1545 Int (18 ) and Tn916 Int (24 ) can occur in E.coli in the absence of xis, although xis stimulates excision by Tn1545 Int (18 ).
Int protein is a bivalent DNA binding protein (25 ). The C-terminal domain of Int binds to the ends of Tn916 and flanking bacterial DNA, and protects ~40 bp, centered on the coupling sequence, from nuclease cleavage (25 ). This domain contains Arg, His and Tyr residues that are conserved within the integrase family of recombinases (20 ) and that have been shown to be essential for biochemical activity in other recombinases (26 -28 ). At the ends of the transposon, Int makes staggered cleavages flanking the coupling sequences, leaving a 5' hydroxyl group and forming a covalent protein-DNA complex at the 3' side of the site of cleavage similar to those formed by other integrase family recombinases (29 ). The N-terminal domain of Int binds to direct repeats called DR-2 close to the ends of the transposon, and protects ~40 bp of DNA from nuclease cleavage (29 ). These repeats are located 90 bp from one end of the transposon and 150 bp from the other (11 ).
It appears that the reactions catalyzed by Int protein during conjugative transposition are similar to the recombination reactions catalyzed by other integrase family recombinases. The coupling sequences are analogous to the overlap regions in lambda, Flp and XerC/D recombination. One difference between Tn916 and these other integrase family members is that in normal Tn916 excison and integration, the coupling sequences of the two partners are different and the resulting recombinant products contain heteroduplex DNA (9 ,30 ). In reactions catalyzed by other integrase family members, the presence of non-complementary sequences in the overlap regions of recombining molecules inhibits recombination (31 -35 ).
Xis protein binds to both ends of the transposon close to the DR-2 repeats and, at higher concentrations, produces a nuclease cleavage pattern that suggests that the DNA is wrapped around the protein (36 ). The role of Xis in conjugative transposition is currently unclear.
Having established how Int cleaves the DNA at the ends of the transposon (29 ), to learn more about the mechanism of Int-mediated recombination and the function of Xis we have established an in vitro recombination system focusing on excision because it appears to be the rate limiting step in conjugative transposition of Tn916. We report here that using a plasmid carrying a derivative of Tn1545 as substrate, Tn916 Int alone will catalyze a low level of excision under low-salt conditions. This reaction is stimulated by Xis. At higher salt concentrations, Xis is required for excision.
pUC18::Tn1545del4 (18 ) carries a deletion derivative of Tn1545. pUCTnJ4 was constructed by amplifying the joined transposon ends of Tn916 from circular transposon DNA present in DNA extracted from the E.faecalis strain FB31 containing pIP501orfA::Tn916 (12 ). The amplification was carried out using the primers 5'-TTTGAATTCCATATTTTTACTATCC-3' and 5'-GATCGGATCCGCTTGAATAAAGAGAAGC-3'. The resulting product was digested with EcoRI and BamHI and ligated to similarly digested pUC18.
The oligonucleotides 5'-TCAGAATTCTAAGGAGGTAAAATATTATGAAGCAGACTGACATTCC-3' and 5'-TTTGAAGCTTCTAGATTGCGTCCAATGTA-3' were used as primers in a PCR reaction to amplify the xis gene of Tn916. The first oligonucleotide contains a consensus ribosome binding site nine bases upstream of the initiation codon of xis. The resulting amplified DNA fragment was digested with EcoRI and HindIII and ligated to similarly digested pUC18 so that the cloned xis gene was expressed from the lac promoter of pUC18. The mixture was introduced into SG22094 lon clpP (37 ) by electroporation, and resulting colonies were screened for the presence of a plasmid containing an insert of the expected size. Bacteria that contained such a plasmid were then grown overnight in LB, and screened for the production of a protein of 8 kDa, the size expected for Xis. The nucleotide sequence of the insert in several clones was determined. One clone was recovered that contained an insert with a sequence identical to Tn916 xis. This clone, pUC18RBSXIS was used to produce Xis.
A 1 l culture of SG22094 containing pUC18RBSXIS in LB with ampicillin (50 µg/ml) and chloramphenicol (34 µg/ml) was inoculated from an overnight culture at an OD600 of 0.2 and allowed to grow at 37oC for 3.5 h. The cells were harvested by centrifugation and the resulting cell pellet was frozen at -80oC. The pellet was thawed on ice, and suspended in 30 ml buffer A (50 mM Tris-HCl pH 7.5, containing 0.15 M NaCl and 1 mM EDTA). PMSF was added to a final concentration of 2 mM. The cells, on ice, were then lysed by sonication using three 30 s pulses at maximum power. A 15 ml aliquot of ice cold buffer B (50 mM Tris-HCl pH 7.5, containing 3 M NaCl and 1 mM EDTA) was added, and the lysate, on ice, was stirred gently for 30 min. The lysate was then centrifuged at 20 000 r.p.m. in an SS34 rotor for 25 min. The supernatant was dialyzed against buffer A overnight at 4oC. After dialysis, the suspension was centrifuged at 10 000 r.p.m. for 10 min in an SS-34 rotor. The supernatant was loaded onto a Heparin Hitrap (Pharmacia) column equilibrated with buffer A. The column was eluted with a gradient of 0.15-1 M NaCl. Xis-containing fractions were pooled and concentrated using a Centricon 3.5 (Amicon). The concentrated Xis was then further purified by gel filtration chromatography using Superdex 75 (Pharmacia) in the presence of buffer A. The Xis protein was electroblotted onto PVDF membrane and its N-terminal sequence was determined using an Applied Biosystems pulsed-liquid sequencing system (38 ).
Reactions were carried out in 50 mM Tris-HCl pH 7.5, containing 1 mM EDTA, 15 mM MgCl2 and either 37.5 mM or 150 mM NaCl. Reactions contained as substrate pUC18::Tn1545del4 (50 ng/µl, 13.75 nM). Int purified from baculovirus-infected insect cells (29 ) was used at concentrations between 1 and 10 ng/µl (22.4-224 nM) and Xis was used at concentrations between 1 and 25 ng/µl (125 nM-3.12 µM). Samples were incubated at 37oC for different times, then SDS was added to a final concentration of 0.2%, and the samples were heated at 65oC for 20 min. Proteinase K (500 ng) was added to the samples, which were then incubated at 37oC for 2 h. The DNA in the samples was recovered by ethanol precipitation.
To detect excised circular transposon DNA, PCR reactions were carried out using as primers 5'-CTCGAAAGCACATAGAATAAGGC-3' and 5'-GAGTGGTTTTGACCTTGATA-3', and 5 µl of a 1:10 dilution of the DNA samples as template. PCR was carried out in the presence of [32P]dATP for 20 cycles (94oC, 1 min; 51oC, 1.5 min; 72oC, 1 min). Each experiment also included a series of reactions containing 10, 40, 160 and 640 pg of purified pUCTnJ4 plasmid DNA as template to construct a standard curve. Following PCR, the radiolabelled product was separated from template DNA by electrophoresis on a 5% polyacrylamide gel, and quantitated using a Molecular Dynamics phosphoimager.
Tn916 Xis was purified from E.coli as described in Materials and Methods by treatment of a cell lysate with 1 M NaCl, followed by heparin chromatography and gel filtration. Fractions eluted from the heparin and gel filtration columns were analyzed by SDS-polyacrylamide gel electrophoresis. The elution of Xis from a Hi-trap heparin column and a Superdex 75 column is shown in Figure 1 A and B, respectively. The purified protein had a mass of 8 kDa in SDS-PAGE (39 ), the size expected for Xis. The N-terminal amino acid sequence of the protein was found to be Met-Lys-Gln-Thr-Asp-Ile-Pro-Ile-X-Glu-Arg-Tyr in agreement with the nucleotide sequence of the xis gene.
We established conditions that would allow us to quantitate the amount of transposon excision using PCR. Figure 2 shows the combined results of 20 PCR reactions using as template different amounts of pUC18TnJ4, a plasmid containing the joined ends of Tn916 cloned from the circular form of the transposon. The reactions contained [32P]dATP, and after 20 cycles of amplification, the radiolabelled product was separated from the template on a 5% polyacrylamide gel and detected using a phosphoimager. To correct for differences in the labeling of the product and the time of exposure of the gel from experiment to experiment, for each experiment the value for the amount of product obtained with 10 pg of template was set to 1.0, and the rest of the data were normalized to this value. As shown in Figure 2 , the amount of product was proportional to the amount of template containing the joined ends up to 40 pg of template per reaction.
To determine the optimal concentration of NaCl required for excision of Tn1545del4, a series of reactions was carried out using 10 ng/µl Int and 10 mM MgCl2 (Fig. 3 A). We found that above 37.5 mM NaCl, the amount of excision decreased as the salt concentration increased so that at salt concentrations of 150 mM and higher, excision was barely detectable. We then performed two series of reactions with 10 ng/µl Int, and different concentrations of MgCl2, in 37.5 and 150 mM NaCl, respectively. As shown in Figure 3 B, at both NaCl concentrations, the maximum amount of excision was found at 15 mM MgCl2, which was the concentration used in all subsequent experiments.
Excision reactions with different concentrations of Int were performed in both 37.5 and 150 mM NaCl. As shown in Figure 4 , at 37.5 mM NaCl, the amount of excision of Tn1545del4 was approximately linear with time up to 4 h but there was little increase in product after this time. Comparison of the amount of excision after 4 h with different amounts of Int showed that excision of Tn1545del4 was proportional to the amount of Int added to the reaction over the range of Int concentrations used, which varied from 0.5 to 5 ng/µl. At 150 mM NaCl, the amount of excison was reduced to barely detectable levels, which were not increased using 10-fold higher concentrations of Int (data not shown).
To determine if Xis stimulated excision of Tn1545del4 at low salt concentrations, reactions were carried out in 37.5 mM NaCl using different concentrations of Xis and either 1 or 2 ng/µl of Int. As shown in Figure 5 A and B, a low concentration of Xis (1 ng/µl) had no effect on the amount of excision, but higher concentrations of Xis (5 and 25 ng/µl) stimulated excision at both concentrations of Int. The results in Figure 5 B show that higher concentrations of Xis stimulated the reaction up to ~8-fold. In addition, the higher concentrations of Xis accelerated the reaction, so that with 2 ng/µl of Int (Fig. 5 B) the majority of the product was produced within 1 h. In the presence of 2 ng/µl of Int and 25 ng/µl of Xis, ~9% of the substrate DNA underwent recombination. The molar ratio of Xis to Int under these conditions was 70:1.
We have found that in low salt conditions, Tn916 Int protein can catalyze a low level of excision of Tn1545del4 from a plasmid substrate. The assay we used detects only circular transposon DNA with one or both strands covalently closed. Therefore Int alone is capable of performing strand exchange in addition to DNA cleavage (29 ).
Depending upon the salt conditions, Xis either stimulates excision of Tn1545del4 (at 37.5 mM NaCl) or is required for the reaction (at 150 mM NaCl). Similar observations have been made for phage lambda (40 ). Excision of phage lambda can occur in the absence of Xis at low salt conditions (40 mM NaCl), and this reaction is stimulated by Xis. At higher salt concentrations (140 mM NaCl), Xis is required for lambda Int-mediated excision. Our results for Tn916 reflect observations made in vivo in different hosts, where int alone can promote excision, and xis either stimulates or is required for the reaction (18 ,19 ,21 ,24 ). The concentration of Xis necessary for stimulation of excision under low salt conditions is lower than the concentration required for excision to occur at higher salt concentrations. In both low and high salt conditions, stimulation of excision by Xis occurs at protein concentrations that are higher than those required for specific binding at each end of Tn916 (36 ). The high molar ratios of Xis to Int that were required for stimulation of recombination may mean that a significant fraction of the Xis preparation was inactive. However, the stimulatory concentrations of Xis are comparable to those previously observed to produce an extended pattern of regularly spaced, enhanced nuclease cleavages around the specific binding sites (36 ). Unless the stimulatory activity of Xis can be destroyed without affecting its DNA binding properties, this suggests that wrapping of the DNA around Xis may be important for its activity.
We are grateful to Jan Pohl of the Emory University Microchemical Facility for determining the N-terminal sequence of Xis. This work was supported by grant GM50376 from NIH.
REFERENCES