INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

Site-specific recombination by bacteriophage  is a complex process that requires the formation of nucleoprotein complexes composed of specific DNA sites in the phage and bacterial chromosomes and host- and phage-encoded proteins. Integrative recombination between specific attachment sites, attP on the phage DNA and attB on the bacterial chromosome, generates recombinant attR and attL sites flanking the prophage DNA (14). Excision of the prophage is accomplished by site-specific recombination between the attR and attL sites to regenerate the intact host and phage genomes. Both reactions are catalyzed by the phage-encoded protein integrase (Int), assisted by accessory proteins. The host-encoded integration host factor (IHF) is a protein required for both reactions. Excision requires an additional phage-encoded protein called excisionase (Xis). Excision is stimulated by the factor for inversion stimulation (FIS) supplied by the host. The directionality of recombination is determined by the amount of Xis present in the cell. During prophage induction, enough active Xis is present to promote excision (7, 26). The amount of Xis protein present during establishment of lysogeny is limited due to instability of the protein (26). The integration reaction is inhibited by Xis in vitro (1, 17).

Xis is a sequence-specific DNA binding protein of 72 amino acids (1). Xis recognizes two direct, imperfect 13-bp repeats, designated X1 and X2, on the attR site (see Fig. 1) (28). The FIS binding site, designated F, partly overlaps the X2 site (24, 25). Xis binds DNA cooperatively at the X1 and X2 sites (4) and also binds to X1 cooperatively with FIS at the F site (19, 24). Both Xis and FIS bend DNA upon binding to their specific sites (23). In addition, occupation of X1 by Xis facilitates binding of Int to the P2 site that lies adjacent to the X1 site, presumably through protein-protein interactions (4, 19, 27). The P2 site is a relatively weak Int binding site and is required for excision but not for integration (2, 25).

Xis protein is multifunctional despite its small size. It binds to DNA and interacts with FIS and Int. However, little structural information on the Xis protein is available. Numrych et al. (19) carried out an extensive mutational analysis of Xis and isolated amino acid substitution mutants of Xis with decreased DNA binding affinity in the bacteriophage P22 challenge-phage assays. Their mutations are located in the amino-terminal half of the protein. In contrast, other mutations resulting in defective interaction with Int are at the carboxyl end of the protein (19, 27). No mutants that bound DNA but failed to interact with FIS were isolated. Of the 15 amino acid substitution mutations that decrease DNA binding, 8 change polar residues. Four of these substitutions are clustered between glutamic acid 19 and arginine 26. In addition, another substitution mutant, E40K, showed decreased DNA binding in a challenge-phage assay. Because these polar residues may interact specifically with DNA or FIS, we constructed alanine substitution mutants of each and analyzed their DNA binding properties and their ability to promote excision in vivo.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Bacterial and phage strains. Escherichia coli strain DH5 [supE44 lacU169 (80 lacZM15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1] was used for cloning and purification of variant Xis proteins. BL21(DE3)[F ompT hsdS_B(r_Bm_B) gal( cI857 ind1 Sam7 nin5 lacUV5-T7gene1) dcm] was used for purification of wild-type His-tagged Xis protein. The E. coli strain LE292 {HfrH argE(Am) rpoB galT::([int-FII])} and its fis::kan derivative were used for the red colony test. LE292 (fis::kan) was constructed by generalized transduction using phage P1vir grown on E. coli strain MO (fis::kan) (19). A crude extract containing the wild-type Xis protein was prepared from RJ1529 (fis::kan) harboring pPS2-3RS (19). The challenge phage P22xis2B and its derivatives containing variant X1-X2-F sites were used as templates for preparation of DNA in gel-shift assays (see Fig. 1) (18). As a nonspecific DNA for gel-shift assays, a challenge phage, P22 P'123(II), was used (15).

Media, chemicals, and enzymes. The media and buffers were described previously (19). Antibiotics (Sigma) were added to the media as follows: ampicillin to 50 µg/ml, spectinomycin to 100 µg/ml, and kanamycin to 50 µg/ml. For the red-colony test, timetin (SmithKline Beecham Pharmaceuticals) was used instead of ampicillin at a concentration of 50 µg/ml to prevent the growth of ampicillin-sensitive satellite colonies. Isopropyl--D-thiogalactopyranoside (IPTG) was obtained from Sigma and used at the indicated concentrations. T4 DNA ligase, T4 polynucleotide kinase, and restriction endonucleases were obtained from Bethesda Research Laboratories or New England Biolabs. Taq DNA polymerase was obtained from Promega.

Plasmid constructions. PCR was used to isolate and amplify the xis gene from phage  DNA. The upstream primer Xis-1 contained an NheI restriction site preceding the initiation codon of the xis gene, and the downstream primer Xis-2 carried an EcoRI site following the xis translation termination codon (Table 1). PCR was carried out using Taq DNA polymerase in a solution containing 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl₂, 0.01% gelatin, and 200 µM concentrations of each deoxynucleoside triphosphate. The fragment was digested with NheI and EcoRI and ligated into the plasmid pET28a (Novagen), which was predigested with the same set of enzymes. After transformation into E. coli strain DH5, a clone harboring the insert was confirmed by sequencing (22). The resulting clone, pYL-XisHP, was transformed into E. coli BL21(DE3) for expression and purification of the Xis protein.

Site-directed mutagenesis. The plasmids containing alanine substitution mutants, pXisE19A, pXisR22A, pXisR23A, pXisR26A, pXisE27A, and pXisE40A, were constructed via site-specific mutagenesis using the modified megaplasmid PCR method (12). Ten picomoles of the mutagenic oligonucleotide and Xis-2 primer (Table 1) were added to the template pYL-XisHR DNA. This reaction amplified short fragments containing desired base-pair changes, which in turn would be used as a megaprimer in subsequence amplifications. After 10 cycles, 12.5 pmol of upstream primer, PCKR-p, was added, and DNA was amplified for 10 more cycles. In the next 10 cycles, 12.5 pmol of Xis-2 was supplied to finish the amplification. The amplified DNA was digested with EcoRI and cloned into the EcoRI backbone of the plasmid pYL-XisHR. Each mutation was confirmed by DNA sequencing.

Expression and purification of His-tagged Xis and its derivatives. E. coli strains DH5, harboring derivatives of the pYL-XisHR plasmid, and BL21(DE3), with the pYL-XisHP plasmid, were used for expression of the Xis mutants and the wild-type Xis protein, respectively. Cells were grown to mid-log phase in Luria-Bertani medium containing the appropriate antibiotics. Production of His-tagged proteins was induced by adding IPTG to a final concentration of 1 mM, followed by growth for 3 h at 37°C. After centrifugation, harvested cells were disrupted by treatment in a French pressure cell in 50 mM sodium phosphate buffer, pH 7.4, with 300 mM NaCl. Prior to disruption, 250 µl of protease inhibitor cocktail for bacterial cell extracts (Sigma product no. P8465) was added per gram of cell mass. Cell lysates were centrifuged at 12,000 × g for 30 min, and the supernatants were mixed with nickel-nitrilotriacetic acid column material (Qiagen) saturated in 50 mM sodium phosphate buffer, pH 7.4, with 300 mM NaCl and 10% glycerol. The column was washed extensively, and the His-tagged proteins were eluted by adding 250 mM imidazole in the same buffer. The total protein concentrations were measured using the dye-binding assay (3). The amount of Xis in each protein sample was determined by measuring peaks after scanning the Coomassie blue-stained sodium dodecyl sulfate-polyacrylamide gel.

Preparation of crude extracts. An E. coli strain, RJ1529, harboring the plasmid pPS2-3RS (18) was grown at 37°C to mid-exponential phase, and expression of Xis protein was induced by IPTG at a concentration of 1 mM for 1 h. After centrifugation, cell pellets were collected and resuspended in a solution containing 20 mM Tris-HCl (pH 7.4), 100 mM EDTA, 20 mM NaCl, and 10% glycerol. The cell suspension was sonicated using Branson Sonifier Cell Disruptor 200 and centrifuged at 20,000 × g for 1 h. The cleared sonic extract was used as a source of crude Xis protein.

Gel-shift assays. The DNA fragments used in gel-shift assays were amplified by PCR using phage P22xis2B or its derivatives as templates (Fig. 1). The phage P22 P'123(II), which contained the attL Int arm-type binding sites of  instead of the Xis binding sites, was used as a source of labeled, nonspecific DNA. A pair of synthetic oligodeoxyribonucleotides, Omnt and -Omnt (Table 1), were used as primers for amplification. They were annealed to the sequences encompassing the Xis binding sites or Int arm-type binding sites. One of the primers was labeled with [-³²P]ATP and polynucleotide kinase prior to amplification (21). The PCRs were carried out using Taq DNA polymerase (Promega). The DNA fragments were purified on a Micro Bio-Spin chromatography column 30 (Bio-Rad) and were quantified with a PhosphorImager (Molecular Dynamics).

View larger version (12K): [in this window] [in a new window]   FIG. 1.   The attR site of bacteriophage  and nucleotide sequences of wild-type and variant Xis binding sites. The two arm-type binding sites, P1 and P2, and the two core-type sites, C and B', on the attR region are recognized by Int. IHF binds to the sites designated as H1 and H2. Xis binds to the X1 and X2 sites. FIS binds to the F site. The isolation of the mutant Xis binding sites 2B10, 2B29, and 2B22 has been described by Numrych et al. (18). , deletion of the indicated base.

Excision assay. The red-colony test (11) was used to assay the ability of Xis derivatives to promote excisive recombination in vivo. The galT gene of E. coli strain LE292 contains a  prophage that lacks the int and xis genes. When the Int and Xis proteins are supplied from plasmids, the inserted prophage is excised from the chromosome and renders the cells Gal positive. Plasmids containing the wild-type or mutant xis genes were transformed into strain LE292 containing pIntB1, and the cells were spread on MacConkey galactose-timetin-spectinomycin plates containing 1 mM IPTG. The plates were incubated at 30°C, and the time required for the colonies to turn red was used as a measurement of in vivo excisive recombination activity.

Secondary structure predictions. The secondary structure of the Xis protein was predicted using various algorithms available at the Network Protein Sequence @nalysis web server (http://www.pbil.ibcp.fr/NPSA). The methods used to analyze the structure were DPM (5), DSC (13), GOR IV (9), PHD (20), Predator (8), SIMPA96 (16), and SOPM (10).

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

His-tagged Xis retains the binding specificity and functions of Xis. In order to simplify the purification of Xis and its mutant derivatives, we constructed a modified gene that encoded a His tag at the amino terminus of the protein. The xis gene from phage  was amplified and cloned into the NheI-EcoRI backbone of plasmid pET28a as described in Materials and Methods. The resulting clone, pYL-XisHP, was transformed into E. coli BL21(DE3) for expression and subsequent purification of the His-tagged Xis. After isolating the wild-type His-tagged protein, we analyzed its sequence-specific DNA binding properties in vitro by gel-shift assays. The DNA fragments used in the binding reactions contained either the wild-type X1-X2-F sites or the variant sites shown in Fig. 1. The mutant site 2B10 carries a 4-bp deletion in the X2 site. The other two variant sites, 2B29 and 2B22, have a single-nucleotide substitution in X1 and X2, respectively. Variants were isolated as mutants defective for Xis binding using the challenge-phage system (18).

View larger version (35K): [in this window] [in a new window]   FIG. 2.   DNA binding specificity of His-tagged Xis (A) and the wild-type Xis (B). The 32P-labeled DNA fragments were about 180 bp long and were amplified from P22 challenge phages containing either wild-type or variant Xis binding sites. A 1X amount of Xis (A) corresponds to 0.1 µM protein. The wild-type Xis protein (B) was in crude extracts prepared from an E. coli strain, RJ1529, containing the plasmid pPS2-3RS (18). Each reaction contained 2.5 nM labeled DNA, and where indicated, 12.5 ng (A) or 7.6 ng (B) of FIS was present. As a nonspecific competitor, 0.25 µg (A) or 1 µg (B) of sonicated calf thymus DNA was added. Electrophoresis was performed at room temperature. A contaminating fragment (probably single-stranded DNA) occurring as an artifact of the PCR migrated between the Xis-DNA complex and the FIS-DNA complex.

Amino acid residues from Leu 18 to Glu 27 may form an amphipathic -helix. The carboxyl-terminal region of Xis is required for cooperative binding of Int, presumably through protein-protein interactions (19, 27). A nonsense mutant was isolated that encodes an Xis protein containing the amino-terminal 53 amino acids. It bound to DNA containing the X1-X2-F sites and interacted cooperatively with the FIS protein (19). The precise regions of Xis involved in DNA binding or FIS interaction were not localized. Secondary structure prediction algorithms indicated that three regions of the Xis protein could form -helices (5, 8, 9, 10, 13, 16, 20). Amino acids from residues 5 to 10 were predicted to form the first helix, and the region spanning leucine 18 to glutamic acid 27 was predicted to form the second helix. The third postulated helix was in the region proposed by Numrych et al. (19) to be involved in cooperative interactions with Int. However, Xis lacked recognizable DNA binding motifs, for example, a helix-turn-helix motif (6). We observed that although Xis does not form a canonical helix-turn-helix motif, the second helical region could form a typical amphipathic helix (Fig. 3). Furthermore, Numrych et al. (19) found that several substitutions for the hydrophilic residues in this region resulted in a loss of Xis function in vivo. Those substitutions included the changing of glutamic acid 19 to a lysine, arginine 22 to a histidine, arginine 23 to a glutamine, and arginine 26 to a tryptophan. Those results are consistent with the hypothesis that this region forms an -helix and the surface-exposed hydrophilic residues may be in direct contact with DNA or FIS.

View larger version (21K): [in this window] [in a new window]   FIG. 3.   Helical wheel projection of the putative -helix spanning amino acid residues 18 to 27 in Xis. The residues named in the boxes are the amino acids for which substitutions resulted in defective excision in vivo (19).

Substitutions for polar residues in the putative amphipathic helix change DNA binding. To gain more information on the function of each of the polar side chains on the putative amphipathic helix, we constructed alanine substitution mutants of the hydrophilic residues. Glutamic acid 19, arginines 23 and 26, and glutamic acid 27 were individually replaced with an alanine residue. The proteins were designated E19A, R23A, R26A, and E27A, respectively. We also constructed a mutant with an alanine substitution at position 22, but we could not detect the protein on a sodium dodecyl sulfate-polyacrylamide gel. This mutant was not analyzed further. Each of the Xis variants was tested for DNA binding and ability to promote excision in vivo. Gel-shift assays were performed using partially purified proteins as described in Materials and Methods.

View larger version (36K): [in this window] [in a new window]   FIG. 4.   DNA binding of XisR23A and XisR26A. The final concentration of wild-type Xis was 0.1 µM, and that of the mutant Xis proteins was 5 µM. Each reaction contained 2.5 nM labeled DNA and 50 ng of sonicated calf thymus DNA. Where indicated, 6.3 ng of FIS was present. Electrophoresis was performed at 4°C.

View larger version (58K): [in this window] [in a new window]   FIG. 5.   DNA binding patterns of wild-type Xis (A), XisE19A (B), and XisE40A (C). The specific DNA fragments contained the Xis and FIS binding sites, X1-X2-F. The band indicated by an arrow in panel A corresponds to a specific Xis-DNA complex. Nonspecific DNA contained the Int arm-type sites P'1, P'2, and P'3 in place of the Xis and FIS binding sites. The flanking sequences of nonspecific DNA were the same as those of the specific DNA fragment. The final concentration of labeled DNA in each reaction was 1 nM. Nonspecific competitor DNAs were absent from the reactions. Electrophoresis was performed at room temperature. A contaminating fragment migrating slower than free DNA is believed to be single-stranded because it was not retarded by the nonspecific DNA binding mutants.

Mutant with a substitution of an alanine for glutamic acid 40 binds to nonspecific DNA. Numrych et al. (19) also isolated a mutant in which glutamic acid 40 was replaced by a lysine. To study the function of this residue in DNA recognition in more detail, we made a His-tagged construct of it. Gel-shift assays showed that the E40A protein, like the E19A protein, bound to both specific and nonspecific DNA fragments (Fig. 5C, lanes 5 to 8). However, the binding pattern for the fragment with specific sequences was different from that for the fragment with nonspecific DNA sequences (Fig. 5C). Although we do not understand the cause of the difference, it might indicate that the E40A protein distinguishes the specific Xis binding sites from random DNA to some extent. The fact that the E40A mutant promoted excision and showed cooperativity with FIS (Table 2) supports this hypothesis. As discussed above for the E19A protein, the E40A protein may also bend DNA when bound to a specific Xis binding site and provides a functional substrate for excision. However, the sequence specificity of the wild-type Xis was significantly relaxed by the substitution for glutamic acid 40, indicating that this residue may also participate, either directly or indirectly, in sequence-specific DNA recognition.