This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Mattis, A. N.
Right arrow Articles by Gardner, J. F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Mattis, A. N.
Right arrow Articles by Gardner, J. F.

 Previous Article  |  Next Article 

Journal of Bacteriology, September 2008, p. 5781-5796, Vol. 190, No. 17
0021-9193/08/$08.00+0     doi:10.1128/JB.00170-08
Copyright © 2008, American Society for Microbiology. All Rights Reserved.

Purification and Characterization of Bacteriophage P22 Xis Protein{triangledown}

Aras N. Mattis,1 Richard I. Gumport,2 and Jeffrey F. Gardner3*

Department of Pathology, University of California, San Francisco, San Francisco, California,1 Department of Biochemistry and College of Medicine, University of Illinois at Urbana-Champaign, Urbana, Illinois,2 Department of Microbiology, University of Illinois at Urbana-Champaign, Urbana, Illinois3

Received 3 February 2008/ Accepted 15 May 2008


arrow
ABSTRACT
 
The temperate bacteriophages {lambda} and P22 share similarities in their site-specific recombination reactions. Both require phage-encoded integrase (Int) proteins for integrative recombination and excisionase (Xis) proteins for excision. These proteins bind to core-type, arm-type, and Xis binding sites to facilitate the reaction. {lambda} and P22 Xis proteins are both small proteins ({lambda} Xis, 72 amino acids; P22 Xis, 116 amino acids) and have basic isoelectric points (for P22 Xis, 9.42; for {lambda} Xis, 11.16). However, the P22 Xis and {lambda} Xis primary sequences lack significant similarity at the amino acid level, and the linear organizations of the P22 phage attachment site DNA-binding sites have differences that could be important in quaternary intasome structure. We purified P22 Xis and studied the protein in vitro by means of electrophoretic mobility shift assays and footprinting, cross-linking, gel filtration stoichiometry, and DNA bending assays. We identified one protected site that is bent approximately 137 degrees when bound by P22 Xis. The protein binds cooperatively and at high protein concentrations protects secondary sites that may be important for function. Finally, we aligned the attP arms containing the major Xis binding sites from bacteriophages {lambda}, P22, L5, HP1, and P2 and the conjugative transposon Tn916. The similarity in alignments among the sites suggests that Xis-containing bacteriophage arms may form similar structures.


arrow
INTRODUCTION
 
Bacteriophage P22 is a temperate lambdoid bacteriophage that, like {lambda}, undergoes both a lytic and a lysogenic life cycle. P22 uses site-specific recombination (SSR) to integrate its chromosome into the Salmonella enterica serovar Typhimurium chromosome and to excise itself during induction of lytic growth to form phage progeny. The integrase (Int) and integration host factor (IHF) binding sites in the P22 attP site have been mapped by footprinting techniques (19, 30) and the Int cleavage sites have been determined (30). P22 excisionase (Xis) is absolutely required for excision of the phage (20). P22 Xis lacks significant similarity in primary structure to {lambda} Xis. For example, {lambda} Xis contains a modified winged-helix motif (28), whereas Dodd et al. predicted that P22 Xis contains a helix-turn-helix DNA binding domain (12). Indeed, Lewis and Hatfull classified Xis proteins into families with {lambda} and P22 Xis proteins into different ones (21). To better understand how P22 Xis functions in excision, we developed a purification method and characterized its properties in vitro.

Bacteriophage {lambda} SSR is well characterized both genetically and biochemically and is the paradigm for a large class of related systems, including the P22 system. In {lambda} SSR, the Int protein is the catalytic component in both the integrative and excisive reactions. During integration, the phage-encoded Int binds specifically to the DNA attachment (att) sites and catalyzes the integration of the phage attP site into the host attB site, forming the recombinant attL and attR sites (for reviews, see references 5 and 17). The overall reaction requires the formation of a multiprotein-DNA complex known as the integrative intasome and comprising the host-encoded IHF protein, Int, and the attP DNA site.

In the excision process, when the phage enters into the lytic cycle, Int catalyzes the reaction between attL and attR to re-form attB and attP. In many SSR systems, the reaction is highly regulated, and an accessory protein is required to perform the excision reaction. {lambda} Xis is a DNA-binding protein that binds as a trimer to its DNA binding site (1, 36) and interacts cooperatively with {lambda} Int to recruit it to the P2 arm-type site of attR (7, 24). {lambda} Int binding to the P2 site is required to form the excisive multiprotein DNA complex or excisive intasome (7). {lambda} Xis bends DNA approximately 140 degrees (32) and can function as a multiprotein complex with the FIS (factor for inversion stimulation) protein when the concentration of Xis is limiting (6, 33). The {lambda} Xis is the best characterized of the Xis proteins because its residues involved in DNA binding (8, 27) and in cooperative interactions with Int (7, 9, 24, 31, 34) have been identified, its structure has been solved by nuclear magnetic resonance spectroscopy (28), and a DNA-Xis cocrystal structure has been solved (26, 27). Bacteriophage L5 Xis and P2 Xis (11, 22, 23, 37) and the HP1 Cox protein (13-15) have also been studied.


arrow
MATERIALS AND METHODS
 
Media, chemicals, and enzymes. Antibiotics were obtained from Sigma and were added to the media as follows: ampicillin to 100 µg ml–1, kanamycin to 50 µg ml–1, and chloramphenicol to 20 µg ml–1. Ammonium persulfate, L-arabinose, dithiothreitol (DTT), EDTA disodium salt, glycine sodium salt, HEPES, isopropyl-β-D-thiogalactopyranoside, N-methyl-N'-nitro-N-nitrosoguanidine, phenylmethylsulfonyl fluoride, sodium dodecyl sulfate (SDS), N-2-hydroxy-1, Tris(2-carboxyethyl)phosphine hydrochloride, N'N'N'N'-tetramethylethylenediamine, 1-bis(hydroxymethyl)ethyl-glycine (Tricine), and Tris base were obtained from Sigma. Polyacrylamide gels were prepared from 40% acrylamide/bis solution (37.5:1) purchased from Bio-Rad (Hercules, CA). Ethylene imine polymer [poly(ethylene imine) (PEI)] solution was obtained from Fluka. [{gamma}-32P]ATP was purchased from PerkinElmer, Inc., (Boston, MA). T4 DNA ligase and restriction endonucleases were purchased from Invitrogen (Carlsbad, CA) or New England Biolabs (Beverly, MA) and reactions were performed to manufacturers' specifications. Taq DNA polymerase was obtained from Invitrogen. RQ1 RNase-free DNase was obtained from Promega (Madison, WI).

Construction of plasmids. The wild-type P22 xis gene was cloned by PCR amplification from plasmid pIX-P22, which was constructed previously (10). For protein purification, the xis gene was cloned into plasmid pET27b (+) after PCR by using oligodeoxyribonucleotides P22Xis-NdeI [d(GGA ATT CCA TAT GGA ATC ACA CAG CCT CAC)] and P22Xis-HindIII [d(CCT CAA GCT TTC AGC TTG TCA TGA AGC TCT G)]. The 371-bp PCR fragment was digested with NdeI and HindIII and ligated into the unique NdeI and HindIII sites of pET27b (+) (Novagen, Madison, WI) to create pANM107. The expression of xis from this plasmid is under the control of the T7 RNA polymerase. For DNA bending assays, plasmid pANM110 was created. Complementary oligodeoxyribonucleotides of XisDBSBend (see Fig. 5) were annealed and inserted by linker tailing into the XbaI site of pBEND2 as described previously (18). The sequences of all cloned genes were confirmed by sequencing both DNA strands. All synthetic DNA was purchased from either Operon (Alameda, CA) or IDT (Coralville, IA).


Figure 5
View larger version (14K):
[in this window]
[in a new window]

  		FIG. 5. DNA sequences tested in EMSA with P22 Xis. All experiments were performed with annealed, double-stranded DNA oligonucleotides. Quantitation was performed as described in Materials and Methods. The DNA sequence protected by the P22 Xis DNase I footprint is at the top. For orientation, the direct repeats found in the footprint sequences have been boxed (from left to right, X1 to X4).

Growth media and strain constructions. All strains were grown in Luria-Bertani (LB) broth (Difco Laboratories, Detroit, MI) or on solid LB medium supplemented with agar (Difco). Escherichia coli strain DH5{alpha} was used for cloning and plasmid maintenance. E. coli BL21-CodonPlus (DE3)-RIL (Stratagene, Cedar Creek, TX) contained plasmid pANM107 (strain ANM110) and was used for P22 Xis protein overproduction.

Overproduction and purification of P22 Xis. Strain ANM110 was used to inoculate 50 ml of LB supplemented with kanamycin and chloramphenicol. After overnight growth on a shaker at 300 rpm at 37°C, the stationary-phase culture was added to a 1-liter portion of LB with no supplements and grown with shaking at 300 rpm at 37°C. The overproduction of P22 Xis was induced with the addition of isopropyl-β-D-thiogalactopyranoside to a final concentration of 0.1 mM at a cell optical density at 595 nm of between 0.60 and 0.80. The cultures were incubated an additional 3.5 h with shaking at 37°C. After growth, the cultures were harvested and the cells pelleted by centrifugation. The pellets were frozen in an ethanol/dry ice bath and stored at –80°C. Protein overproduction in cell extracts was confirmed by SDS-polyacrylamide gel electrophoresis (PAGE) on precast 10 to 20% Tricine gels (Novex).

After confirmation of the overproduction of P22 Xis, cell pellets were thawed on ice and resuspended in buffer HEG (50 mM HEPES [pH 7.0], 10 mM EDTA [pH 8.0], 5% glycerol, 1 mM DTT) supplemented with 50 mM NaCl. Twenty milliliters of buffer was used per cell pellet resulting from 1 liter of culture. Phenylmethylsulfonyl fluoride was added to a final concentration of 2.5 mM. Cells were lysed by sonication and the cell extract was clarified by centrifugation at 43,000 x g for 30 min at 4°C. The supernatant fraction was recovered and moved to clean tubes. A PEI precipitation was performed by addition of 10% PEI, pH 7.9. An amount of PEI equal to 0.035 of the cell suspension volume was added dropwise with stirring at 4°C. The precipitation mixture was allowed to continue being stirred for an additional 15 min at 4°C and then clarified by centrifugation at 30,500 x g for 15 min at 4°C.

The supernatant was loaded onto column I at a rate of 5 ml/min with buffer HEG supplemented to 50 mM NaCl. Column I consisted of two column elements in tandem. The first column element comprised four prepacked 5-ml HiTrap Q high-performance (HP) columns (Amersham Biosciences, Piscataway, NJ) followed immediately in-line with the second column element, consisting of four prepacked 5-ml HiTrap SP HP columns (Amersham Biosciences). The Xis protein thus passed through the HiTrap Q HP resin first, followed by binding to the HiTrap Sepharose (SP) HP resin. We had determined that P22 Xis has no affinity for the HiTrap Q HP resin in this buffer and passes through the resin to the HiTrap SP HP resin, where it is tightly bound. After being loaded, column I was washed with 5 column volumes of buffer HEG supplemented with 50 mM NaCl. Since the Xis bound the HiTrap SP HP column, the HiTrap Q HP resin column was removed at this step. The remaining HiTrap SP HP column was washed with an additional 10 column volumes of buffer.

Protein was eluted from the HiTrap SP HP column with a gradient in 15 column volumes of 50 mM to 500 mM NaCl in buffer HEG. Since we had determined that extracts containing overproduced Xis bind to attR DNA but not to attL DNA in electrophoretic mobility shift assays (EMSA), we used this assay to monitor purification. Specifically, extracts of overproduced P22 Xis or fractions containing partially purified P22 Xis were subjected to EMSA (EMSA conditions and attR and attL DNA products are described in the EMSA section below). During initial trials, we found that extracts of overproduced P22 Xis were able to shift P22 attR or attP DNA specifically out to a large dilution, whereas extracts of induced vector control DNA did not shift the attR DNA specifically. In contrast, P22 Xis was not able to specifically shift P22 attL DNA. Column fractions were thus tested for their abilities to shift P22 attR DNA forming at least a single specific complex. As a negative control, all fractions were also tested in EMSA of P22 attL DNA.

Fractions containing Xis were identified, pooled, and dialyzed into buffer TEG (50 mM Tricine [pH 8.4], 10 mM EDTA [pH 8.0], 5% glycerol, 1 mM DTT) supplemented with 20 mM NaCl. Protein was loaded onto a MonoS HR 5/5 column with length-by-internal-diameter dimensions of 5 cm by 5 mm (Amersham Biosciences) at a rate of 1 ml/min in buffer TEG supplemented with 20 mM NaCl. The column was washed with 10 column volumes of buffer TEG supplemented with 20 mM NaCl, and the protein was eluted in a gradient with 10 column volumes containing from 20 mM to 500 mM NaCl in buffer TEG. Fractions containing P22 Xis were identified, pooled, and concentrated using an Amicon Ultra-15 centrifugal filter unit with Ultracel-3 membrane according to the manufacturer's protocol.

Concentrated protein (4.5 ml) from the Mono S column was loaded onto a HiLoad 16/60 Superdex 75 preparative-grade column (60 cm [length] by 16 mm [internal diameter]) in buffer HN (10 mM HEPES [pH 7.0], 150 mM NaCl, 5% glycerol, 1 mM DTT). P22 Xis eluted as a 25,000 Mr dimer. Fractions containing Xis were buffer exchanged and concentrated into 10 mM HEPES and 2 mM Tris(2-carboxyethyl)phosphine hydrochloride by use of an Amicon Ultra-15 centrifugal filter unit with an Ultracel-3 membrane according to the manufacturer's protocol and frozen on ethanol/dry ice. The purified protein was more than 95% pure as judged by SDS-PAGE. Purified P22 Xis was subjected to matrix-assisted laser desorption ionization-time of flight mass spectrometry, which was performed by the UIUC Protein Sciences facility. The theoretical pI and molecular weight of the protein were calculated with Compute pI/Mw at http://us.expasy.org/tools.

EMSA of P22 Xis. EMSA was used both to purify P22 Xis and to evaluate the DNA binding properties of purified P22 Xis. The binding buffer contained 50 mM Tricine (pH 8.4), 1 mM EDTA, 50 mM NaCl, 300 µg ml–1 bovine serum albumin (BSA), 12% glycerol, 100 ng ml–1 poly(dI-dC)·poly(dI-dC), and 1 mM DTT. Purified P22 Xis and protein extract dilutions were made in 10 mM HEPES (pH 7.0) with 1 mM DTT. DNA products were made by PCR or annealing of oligodeoxyribonucleotides. The attP product (–160 to +190) was made by PCR using template pRA114 and oligodeoxyribonucleotides P22attR-up [d(CTG ATT GCT AAG TGG TTT GGG)] and P22attL-down [d(CGA TTT TTG GTA CTT CTG TCC C)]. The attL product was made by PCR using template pRA114 and oligodeoxyribonucleotides P22attL-up [d(AGC TCA GTT GGT AGA GCA GCG C)] and P22attL-down. The attR product was made by PCR using template pRA114 and oligodeoxyribonucleotides P22attR-up and P22attR-down [d(GAA ATG AGG TTG TAC ATA AGT GAT TG)]. The attB product (+10 to +109) was made by PCR using template pRA114 and oligodeoxyribonucleotides P22attR-down and P22attL-up. The XISDBS17 product was made by PCR using template pRA114 and oligodeoxyribonucleotides 801up [d(AAG GTA CCA TTC GGA AAG GTC TGA AGT GTA GC)] and 806do [d(AAG TCG ACC CAT CTT CGA AAG ACA TGC)]. The XISDBS34 product was made by PCR using template pRA114 and oligodeoxyribonucleotides 807up [d(AAG GTA CCC AAA TCT TTG CAT CGG TTT GC)] and 807do [d(GGG ACG TGT GAG CGC AGG TAT GAC GTC GAC TT)]. The XISDBS49 product was made by PCR using template pRA114 and oligodeoxyribonucleotides 802up [d(AAG GTA CCT TGT TTT GAT CGA TAC AAG C)] and 808do [d(TTT GTA TGT CCC ATT TTT GTC C)].

DNA was labeled with [{gamma}-32P]ATP and T4 polynucleotide kinase according to manufacturer's protocol (Invitrogen). Labeled DNA was deproteinated by extraction with phenol, precipitated with ethanol, and resuspended in 10 mM Tris (pH 8.0). The concentration of labeled DNA used in the reaction mixtures was 0.04 nM. Labeled DNA was incubated with various concentrations of P22 Xis for 40 min at room temperature in P22 Xis DNA binding buffer and loaded onto gels operating at 4°C. Tris-glycine (25 mM Tris base, 147 mM glycine, 1 mM EDTA; pH 8.3) nondenaturing polyacrylamide gels were used in EMSA at polyacrylamide concentrations of 5%, 10%, and 12%. The gel dimensions used were 16 cm in width by 16 cm in height by 3 mm in thickness. The gel field was set at 180 V for 30 to 90 min before samples were loaded. Gels were then set to a constant current of 35 mA and reaction mixtures were loaded. After the last lane was loaded, gel current was continued for an additional 10 min, and the field current was turned down to a 20-mA constant current. Electrophoresis was continued until the xylene cyanol marker had migrated halfway down the gel. The gels were dried and visualized on a Fuji FLA3000 phosphorimager (Fuji). Radioactivity was visualized and quantitated using Fuji Image Gauge version 3.4 software (Fuji). Apparent Kd (dissociation constant) values were estimated by quantitation of complexes or unbound DNA. The fraction of DNA bound was plotted versus the protein concentration and a curve was fit by nonlinear regression analysis with GraphPad Prism software, version 4.0c. Calculations were made assuming no cooperativity and 100% active protein based on the difficulty in calculating absolute activity given known cooperative interactions. Kd values were also estimated by refitting the data using the Hill coefficient (35).

DNase I footprinting of P22 Xis. To make each labeled DNA fragment, one primer was labeled with [{gamma}-32P]ATP as described above. Oligodeoxyribonucleotide purification was performed by column chromatography on Sephadex G-50 as previously described (29). The primers were then used to amplify DNA from plasmid pRA114, containing P22 attP. Footprinting was performed on two different DNA products. PCR product XISDBS29 was amplified from P22 attP (plasmid pRA114) by using Platinum Pfx DNA polymerase according to the Invitrogen protocol with 32P-labeled primers 802up and 806do. PCR product P22attP367 was amplified from P22 attP (plasmid pRA114) by use of Platinum Pfx DNA polymerase according to Invitrogen's protocol with 32P-labeled primers Alt2attPup [d(GCA TGT TAA TTG ATC GTT GTT ACC G)] and Alt1attPdown [d(CGC CGA TAT GCT CAT CTG GCA CC)].

For each reaction, the PCR product was purified from 5% or 8% native polyacrylamide gels as described previously (4). The DNase I footprinting reactions were performed according to Promega's core footprinting system protocol with the binding buffer conditions listed below (Promega, Madison, WI). The binding buffer contained 50 mM Tricine (pH 8.4), 50 mM NaCl, 300 µg ml–1 BSA, 12% glycerol, and 1 mM DTT. Purified P22 Xis dilutions were made in 10 mM HEPES (pH 7.0) with 1 mM DTT. Calcium (2.5 mM) and magnesium (5 mM) were added to the reaction mixtures as described in the Promega core footprinting system protocol. Reaction components were separated on 6% or 8% denaturing sequencing gels as described previously (4). Sequencing ladders were made using either a USB sequenase PCR product sequencing kit (USB, Cleveland, OH) or Maxam-Gilbert sequencing (3). Gels were dried and visualized on a phosphorimager after a 48-h exposure. Alternatively, gels were exposed to Kodak BioMax MS film by use of a Kodak BioMax Transcreen HE intensifying screen at –80°C.

DNA bending assay. Circularly permuted DNA fragments were created by single restriction enzyme digestions of plasmid pANM110 with the following enzymes: MluI, BglII, NheI, SpeI, XhoI, DraI, EcoRV, PvuII, SmaI, StuI, NruI, SspI, and BamHI. The digested products were labeled with [{gamma}-32P]ATP, separated on a 2% agarose gel, and purified by use of a QIAquick gel extraction kit (Qiagen). DNA bending assays were performed by EMSA, using protein concentrations that led to approximately 50% of the DNA being bound. Electrophoresis was done on 5% nondenaturing polyacrylamide gels, and the experiments were repeated in triplicate. Calculation of the DNA binding angle was done according to the method of Thompson and Landy (32).

Gel filtration chromatography. Protein, DNA, and their complexes were separated at room temperature on a fast protein liquid chromatography system (Pharmacia) coupled to an HR 10/30 (10-mm-by-30-cm) Superdex S200 column. For stoichiometry studies, the running buffer contained 50 mM Tricine (pH 8.4), 1 mM EDTA, 50 mM NaCl, 1 mM MgCl2, and 5% glycerol. Absorption of the eluate was monitored at A280 with samples of protein only and at A254 for stoichiometry studies where protein and DNA were present. Samples contained 1 to 18 µM P22 Xis and/or 1 to 4 µM of 50-bp or 28-bp duplex DNA oligonucleotide. Oligonucleotides were annealed and purified according to the procedure described by Sam et al. (26). Samples were incubated at room temperature for 30 min prior to injection. Injection volumes were 100 µl. The partition coefficient and Mr were calculated as described in reference 2. Sizing standards (and their molecular masses and elution volumes), all obtained from Sigma, for Superdex S200 were as follows: blue dextran (2,000 kDa; 7.7 ml), gamma globulin (300 and 150 kDa; 10.3 and 12.1 ml), BSA (134 and 67 kDa; 12.0 and 13.8 ml), RNase A (13.5 kDa; 20.5 ml), creatine kinase (86 kDa; 13.8 ml), and B-lactoglobulin (35.6 kDa; 16.4 ml). All standards were diluted to give peaks with absorbances of ~0.01 absorbance units at 280 nm.

Protein cross-linking. The protein-protein interactions of purified P22 Xis were evaluated by glutaraldehyde cross-linking. The reaction buffer contained 50 mM Tricine (pH 8.4), 1 mM EDTA, 50 mM NaCl, 12% glycerol, and 1 mM DTT. Proteins were incubated for 40 min at room temperature in the reaction buffer with and without BSA as indicated, and then glutaraldehyde was added to a final concentration of 0.0037% or 0.01%. Cross-linking was continued for 2 hours at room temperature, after which the reactions were quenched by the addition of Tris (pH 7.4) to a final concentration of 100 mM. Complexes were resolved on 10 to 20% Tricine SDS-PAGE gels as described above. Gels were stained by Coomassie blue G-250 stain followed by staining with the SilverQuest silver staining kit, as indicated by the manufacturer's instructions.


arrow
RESULTS
 
P22 Xis binds P22 attR specifically. P22 Xis protein was overproduced and purified to more than 95% purity as described in Materials and Methods. Figure 1 shows that the protein monomer migrates at ~14.4 kDa on denaturing SDS-PAGE gels. The identity and purity (more than 99% pure) of P22 Xis was confirmed by matrix-assisted laser desorption ionization-time of flight mass spectrometry (observed Mr, 12755.02; predicted Mr, 12754.55).


Figure 1
View larger version (21K):
[in this window]
[in a new window]

  		FIG. 1. Overproduction and purification of P22 Xis. Lanes: 1, molecular weight markers of Mr values of 16.9K, 14.4K, 10.7K, and 8.1K as indicated; 2 to 6, purified P22 Xis fractions 1 to 5, respectively, from final column. Samples were analyzed on 10 to 20% Tricine SDS-PAGE gels as described in Materials and Methods.

We used EMSA and DNase I footprinting to show that P22 Xis is a DNA binding protein and that it specifically protects one DNA region on the P22 attR site. Initially, the binding region for P22 Xis was localized by EMSA using large DNA fragments (Fig. 2A). P22 Xis retarded the migration of DNA that contained attP or attR sites but not that containing attL or attB sites. Further experiments were done with successively smaller DNA products. At lower protein concentrations, P22 Xis shifted only fragments containing the segment of DNA spanning the P2 and P3 arm-type sites. For example, when a fragment containing DNA including P1, P2, and P3 sites (XISDBS17) was used (Fig. 2B, gel A) a slowly migrating complex is observed at higher protein concentrations (lanes 1 to 3). As the protein is diluted in lanes 5 to 7 of that same gel, faster-migrating complexes are seen as the Xis concentration was lowered to a Xis concentration of 62 nM. In gel B of Fig. 2B, a different DNA fragment (XISDBS49) that does not contain the DNA beyond the P2 site was used. Here, complexes are present at high protein concentration but no complexes are seen at low protein dilutions. Comparing the data in these two gels allows us to make two conclusions. First, P22 Xis binds with high affinity to the region between the P2 and P3 sites. Second, at high protein concentrations, P22 Xis can form complexes with larger attP DNA with or without the high-affinity site.


Figure 2
Figure 2
View larger version (54K):
[in this window]
[in a new window]

  		FIG. 2. (A) Summary of electrophoretic shift analysis of DNA PCR products. For construct details, see Materials and Methods. Dark lines signify host DNA and thin lines signify bacteriophage P22 DNA. The key describes the symbols. DNA numbering is based on the original nomenclature of Smith-Mungo et al. (30). (B) EMSA of P22 Xis and XISDBS17 in gel A and XISDBS49 in gel B on a 5% polyacrylamide gel. Decreasing concentrations of P22 Xis were incubated with double-stranded PCR XISDBS17 in gel A and PCR XISDBS49 in gel B. Protein concentrations for both gels were equivalent for lanes 0 to 8 and are shown above the lane designations. Free DNA and complex 1 are indicated by arrows. Larger complexes termed complex 2 are indicated by the brace. The regions of attP DNA contained in the PCR products are shown in panel A. (C) EMSA of P22 Xis and XISDBS34 DNA on a 5% polyacrylamide gel. Decreasing concentrations of P22 Xis were incubated with double-stranded PCR XISDBS34. Protein concentrations are indicated. Free DNA is indicated by the large arrow. The DNA-protein complex is indicated by the brace. The region of attP DNA contained in the PCR product is shown in panel A. (D) EMSA analysis of P22 Xis and XISDBS52 DNA on a 10% polyacrylamide gel. Decreasing concentrations of P22 Xis were incubated with double-stranded synthetic XISDBS52 DNA. Free DNA, complex 1, and complex 2 are indicated by large black arrows. The XISDBS52 DNA sequence is shown in Fig. 5. (E) EMSA of P22 Xis and XISDBS52 on a 12% polyacrylamide gel. Decreasing concentrations of P22 Xis were incubated with double-stranded synthetic XISDBS52. Free DNA, complex 1A, complex 1B, and complex 2 (in the wells) are indicated by large black arrows. The XISDBS52 DNA sequence is shown in Fig. 5.

A smaller DNA fragment containing 52 bp of DNA from the right of the P2 site extending to P3 (XISDBS34) was analyzed as shown in Fig. 2C. Here, only one complex per lane is seen, with small changes in the migration of the DNA as the protein is diluted. At a concentration of 15 nM, no DNA is bound. At a concentration of 62 nM, nearly all of the XISDBS34 DNA is in a single complex. Similar experiments with the same DNA were performed as shown in Fig. 2D and E, except higher-percentage polyacrylamide gels were used in an attempt to resolve as many complexes as possible. In Fig. 2D, XISDBS52 DNA (see Fig. 5), which contains 52 bp of double-stranded DNA between the P2 and P3 sites, was analyzed on a 10% gel. In this case, two complexes were resolved: complex 1 and complex 2. When this experiment was repeated with a 12% gel, complex 1 was resolved into two complexes which we call complexes 1A and 1B (Fig. 2E). Based on these gel shifts, we hypothesize the existence of at least two independent binding sites represented by complexes 1A and 1B. Complex 2 might represent the situation when both 1A and 1B are fully bound by Xis. Binding of XISDBS52 was highly cooperative because complex 2 was formed at Xis concentrations where complexes 1A and 1B were also present. In a control experiment, the addition of unlabeled target DNA of a different length did not lead to additional complexes, thereby ruling out the possibility of bimolecular sandwich complexes containing two separate DNA molecules per complex (data not shown).

To map in detail the DNA bases bound by P22 Xis DNA binding, we performed DNase I footprint analysis of P22 Xis bound to 32P-end-labeled attP DNA. The footprints are shown in Fig. 3A, B, and C, with an analysis of the data given in Fig. 3D. Figure 3A shows protection of the top strand in the region between the P2 and P3 sites of attR. P22 Xis concentrations ranged from 1,500 nM to less than 1 nM. As the Xis concentration increased, DNA protection was first visible at a concentration of 10 to 20 nM P22 Xis, with hypersensitive DNase I cleavage enhancements already present at 10 nM. With a further increase in protein concentration to 80 nM P22 Xis, the hypersensitive sites increased to a maximum intensity, especially the sites within the protected region. As the P22 Xis concentration was increased to 1,500 nM, the protected region increased in size while the internal hypersensitive sites decreased in intensity. At the highest protein concentration, some very faint secondary protection sites could be seen near the top of the gel. Interestingly, the four hypersensitive sites are each the middle "T" of the direct repeat ("TTTGCA" for X1 to X3 and "TTTCGA" for X4) found in the sequence (see below).


Figure 3
Figure 3
View larger version (187K):
[in this window]
[in a new window]

  		FIG. 3. (A) DNase I footprint of P22 Xis on P22 attP DNA (XISDBS29), top strand. Concentrations of P22 Xis are indicated above the gel. Lanes 0 contain no Xis protein. P22 attP DNA binding sites P1, P2, and P3 are labeled on the left. A protected region, illustrated by the bracket, contains direct repeats X1, X2, X3, and X4. Enhanced cleavages are indicated by the arrows. T+C and G+A ladders are shown in the leftmost two lanes. (B) DNase I footprint of P22 Xis on P22 attP DNA (XISDBS29), bottom strand. Concentrations of P22 Xis are indicated above. Lanes 0 contain no Xis protein. P22 attP DNA binding sites P1, P2, and P3 are labeled on the left. Direct repeats (X1 to X4) are labeled and are located in the area that was clearly protected on the complementary strand. Enhanced cleavages are indicated by the arrows. T+C and G+A ladders are shown for the same DNA fragment (left lanes). Lanes 0 contain no Xis protein. (C) DNase I footprint of P22 Xis on P22 attP (P22attP367, top strand) DNA showing secondary footprint sites. Lanes contain increasing concentrations of P22 Xis as indicated. Lane 0 contains no Xis protein. A DNA sequencing ladder is shown on the right to identify the protected bases, which are represented as A, C, G, and T. To the right is a map of P22 attP. The secondary protected regions cover the C, C', H, and H' DNA binding sites. (D) Summary of DNA footprinting results. Bases specifically protected by secondary footprints. The P22 attP DNA binding sites are shown and labeled as follows: light blue, arm-type binding sites P1, P2, and P3; green, core Int binding sites C and C'; and gray, IHF binding sites H and H'. The P22 Xis DNA binding sites protected at low protein concentration are shown by solid bars above the strand. Sites protected by highest concentration of P22 Xis (including secondary sites) are shown by open bars above the strand. Direct repeats in the P22 Xis primary footprint site are shown in red and labeled as X1, X2, X3, and X4. Hypersensitive sites are marked with asterisks.

Figure 3B shows the footprint of P22 Xis on the bottom strand of the P22 attP DNA. The footprint was repeated several times with various electrophoresis run times to optimize resolution. No clear protection pattern was seen. However, upon analysis of multiple different exposures, we observed hypersensitive sites at a P22 Xis concentration of 320 nM and higher with limited bases protected. The hypersensitive sites are near the flanking regions of the direct repeats. A secondary footprint is also visible near the bottom of the gel.

In Fig. 3C, footprint analysis of the top strand was repeated with a longer DNA fragment (see Materials and Methods). In this gel, P22 Xis unexpectedly protected secondary sites upstream and downstream of the primary binding site. At low protein concentrations of up to 185 nM, no secondary binding was present. At 370 nM P22 Xis, secondary sites, which we call beta and gamma, were visible. The beta protection site contains the entire core region and at very high protein concentrations extends to cover the H site IHF binding site. The gamma site includes the H' IHF site. In multiple footprinting experiments on P22 attP DNA, additional protected sites at roughly every 40 to 70 bp of the DNA were found at higher protein concentrations. The secondary sites occurred regularly along the attP DNA and included protection of the H, C, C', H', P'1, and P'2 sites. Secondary footprints are also present upstream of P1. The binding of secondary sites likely explains the EMSA behavior of Xis shown in Fig. 2B, where with each increase in P22 Xis concentration, the complexes appear to decrease in mobility, forming discrete yet slower-migrating complexes. The footprinting data are summarized in Fig. 3D.

Features of the protected site. At the primary binding site, P22 Xis protects 41 bases on the top strand (Fig. 3D) with only limited bases protected on the bottom strand at a high protein concentration. Sites hypersensitive to DNase cleavage were observed in both the top and bottom strands. As the protein concentration was increased, the footprint on the top strand increased to 55 bases with further changes in hypersensitive sites.

The P22 Xis footprint on the top strand overlaps the P22 Int P2 arm-type site footprinted previously with DNase I and neocarcinostatin (30) and nearly completely protects it at high protein concentrations. The P3 site is significantly overlapped on the top strand at high protein concentrations (Fig. 3A and D). The footprints include multiple cleavage enhancements, of which a majority flank the primary binding site. Especially interesting are the cleavage enhancements in the P1, P2, and P3 sites, indicating these sites are being rendered accessible to the DNase I by Xis binding.

Inspection of the protected DNA sequence revealed a 6-bp "TTTGCA" motif that is directly repeated three times (X1, X2, and X3) and imperfectly repeated once as "TTTCGA" (X4) (Fig. 3D). These repeats are separated by 4-bp spacers, which place them on the same side of the DNA helix. A cleavage enhancement is present at the middle "T" of every repeat of the "TTT" sequence. This enhancement is not equivalent in intensity at each repeat but is much more enhanced at the second direct repeat (X2). Furthermore, this enhancement is more prominent at lower protein concentrations and becomes less prominent at high protein concentrations, suggesting conformational DNA changes with the increased protein concentration (Fig. 3A).

Quantitative analysis of P22 Xis binding. P22 Xis DNA binding was further analyzed by EMSA both with DNA derived from PCR and with synthetic, duplex oligodeoxyribonucleotides containing the primary binding region (see Fig. 5). Quantitative binding of Xis to XISDBS52 DNA, which contains the X1, X2, X3, and X4 sites, was analyzed by phosphorimager analysis (see Materials and Methods). The derived binding curve showed that P22 Xis bound XISDBS52 DNA with an approximate binding constant (KD) of 70 nM when analyzed assuming no cooperativity in binding. A Hill plot of the data gave a Hill coefficient of 2.2 (Fig. 4). The analysis indicates that P22 Xis binds with positive cooperativity to its cognate site.


Figure 4
View larger version (3K):
[in this window]
[in a new window]

  		FIG. 4. Hill plot analysis of P22 Xis binding XISDBS52. The plot was performed with data from Fig. 2D. The derived Hill coefficient was 2.2.

Determination of minimal DNA binding site. The apparent cooperativity in binding and the resolution of three complexes by EMSA suggested that at least two protomers of Xis bound the DNA containing the entire footprint region. We tried to identify which bases might be important for recognition and binding, as this sequence-related information might provide a clue to the stoichiometry of P22 Xis binding to its cognate site. We altered part of the sequence of X1, X2, X3, or X4 independently in four different experiments in an attempt to disrupt the binding of at least one molecule of Xis. The X1, X2, and X3 sites were each changed from "TTTGCA" to "TTACGA" (the change is underlined). Because the X4 site was different, it was changed from "TTTCGA" to "TTATGA." Disruption of X1, X3, or X4 failed to alter DNA binding, as judged by EMSA. However, changing X2 to "TTACGA" enhanced the formation of complex 2 at lower protein concentrations, making it the dominant complex over a wide range of protein concentrations (data not shown). The results suggest that the direct-repeat sequences are not all equally important for DNA binding. Interestingly, the X2 site also contains the most intense enhanced DNase I cleavage site among the direct-repeat sequences. This raises a number of questions, including whether the changes at X2 produced a better or worse DNA binding site. We favor a model where the new sequence destroys an optimal binding site. Since the altered X2 site is now worse, the cooperatively bound multiprotein DNA complex is favored.

We also tested shortened DNA fragments containing various regions of the X1, X2, X3, and X4 sites to identify the minimal binding site length (Fig. 5). In different experiments, the left and right sides of the footprint region were truncated to remove part or all of the "TTTGCA" repeat sequence. In further experiments, both sides were deleted. We determined the apparent KD values in an attempt to identify half-sites as well as the minimal recognition sequence (listed in Fig. 5).

Most truncations that removed any of the X1-to-X4 region within the footprint reduced the binding affinity. For example, Xis binding to XISDBS100 and XISDBS101, which each have right-side truncations, reduced the KDapp with respect to XISDBS52, which contains more of the footprint. Although the reduction from a KD of ~70 nM to one of ~200 nM is significant, Xis retains appreciable affinity for the DNA site. Truncations that removed a section of the right or left side of the footprint but retained at least two intact direct repeats had an intermediate binding affinity in the 200 to 400 nM range (XISDBS103 contains X1 and X2, while XISDBS108 contains X3 and X4). The XISDBS104, XISDBS109, and XISDBS110 variants that contained fewer than two full direct repeats bound DNA in the micromolar range.

To define more precisely the sequence boundaries for tight binding to the DNA target site, incremental DNA truncations of the right, left, or both sides of the target sequence were tested. The approximate endpoint of required DNA sequence on the left side for tight binding was found to be present in XISDBS132, that is, 6 bp upstream of X1. The required DNA sequence on the right side for tight binding was sequence available in XISDBS136, that is, 10 bp downstream from X4. Removing sequences from both sides within those boundaries resulted in reduced binding affinity. However, truncation sites that contained a longer segment on one side when a truncation was introduced on the opposite side supported tight DNA binding, indicating that additional sequence beyond the recognized site enhances binding. This finding implies that specific nucleotide sequences on either side of the specificity determinants may contribute to maximal binding, possibly through the cooperative binding of further Xis protomers.

In summary, truncations that had fewer than two full "TTTGCA" repeats bound DNA with a Kd in the µM range. If a minimum of two or more "TTTGCA" repeats was present, the apparent dissociation binding constant was approximately 400 nM or lower. Sequences that contained all four direct repeats and most of the footprint region bound with the highest affinity.

P22 Xis bends DNA. Fragments generated by restriction enzyme digestions of plasmid pBEND2, containing P22 Xis sites, were used in EMSA to determine if P22 Xis binding bends DNA. The DNA fragment, XisDBSBend, was subcloned into the XbaI site of pBEND2 (Fig. 5) (16). Single digests at restriction enzyme sites flanking the Xis DNA binding site were used to generate a series of fragments with circularly permuted ends. This procedure positions the Xis binding sites at different distances from the ends of the fragments while maintaining a constant fragment length (32). Concentrations of P22 Xis that bound approximately 50% of the probe were analyzed. The DNA alone was not intrinsically bent, as all unbound DNA permutations had similar mobilities during electrophoresis (Fig. 6). For each EMSA reaction, multiple complexes were observed, consistent with the EMSA analysis discussed previously. Using the method of Thompson and Landy, we calculated that the P22 Xis complexes bent DNA approximately 137 degrees when Xis was bound to the full site (32). The approximate middle of the bend was located within the X2 repeat (Fig. 6).


Figure 6
View larger version (27K):
[in this window]
[in a new window]

  		FIG. 6. Analysis of DNA bending of its cognate DNA binding site by P22 Xis. Lanes 1 to 13 contain gel-purified restriction fragments from subclone pBEND2-XisDBSBend and 143 nM purified P22 Xis. Restriction fragments generated by MluI (lane 1), BglII (lane 2), NheI (lane 3), SpeI (lane4), XhoI (lane 5), DraI (lane 6), EcoRV (lane 7), PvuII (land 8), SmaI (lane 9), StuI (land 10), NruI (lane 11), SspI (lane 12), and BamHI (lane 13) (35). The 60-bp fragment containing the footprint diagramed at the top was subcloned into the XbaI site shown as "C" above. The direct repeats are underlined and labeled X1 to X4 in the sequence. The numbered black bar above the gel shows the relative positions of endonuclease restriction sites in pBEND2. Numbers in the top diagram match lanes in the gel. Three fragments, i.e., 1, 7, and 13, are indicated as examples, and the position of the Xis binding site is shown by "C." The top arrow indicates the wells in the gel. The bottom arrow indicates the migration position of free DNA. The bracket shows the different migration positions of DNA-Xis complexes bound at circularly permuted sites. The bend center is indicated in the top diagram. Reactions were performed with approximately 50% of DNA bound (see Materials and Methods).

P22 Xis cross-links to itself. Purified P22 Xis was subjected to dilute glutaraldehyde treatment in the absence of DNA and was found to cross-link to itself (Fig. 7). As a control, BSA was added to the reaction mixtures to show the specificity of cross-linking. The cross-linked complexes appear to contain two, three, four, or more molecules of Xis, with possible higher-order complexes also present but unquantifiable. Cross-linking in the presence of the Xis DNA binding site failed to alter the cross-linking pattern even when different ratios of protein to DNA were used (data not shown). Cross-linking Xis at different concentrations also did not affect the complexes formed (data not shown). Thus, P22 Xis forms higher-order complexes in solution both with and without DNA.


Figure 7
View larger version (36K):
[in this window]
[in a new window]

  		FIG. 7. Silver-stained SDS-PAGE analysis of P22 Xis cross-linking on a 10 to 20% Tricine polyacrylamide gel. Lanes 1 to 3 contain BSA only, lanes 4 to 6 contain Xis only, and lanes 7 to 9 contain BSA and Xis. Reactions done in the absence of glutaraldehyde are in lanes 1, 4, and 7. Reaction mixtures containing 0.0037% glutaraldehyde are in lanes 2, 5, and 8, and reaction mixtures containing 0.01% glutaraldehyde are in lanes 3, 6, and 9. Lanes 10 and 11 contain a myoglobin molecular mass marker and a benchmark protein ladder, respectively.

Stoichiometry of P22 Xis DNA complexes by gel filtration. Double-strand oligonucleotide DNA binding sites of 50 and 28 base pairs (GF50 and its complements [50-mers] and GF28 and its complements [28-mers] with single 3' base overhangs) (Fig. 5) were used to test P22 Xis binding in vitro under conditions similar to those used for the EMSA experiments described above (see Materials and Methods). Reaction mixtures were analyzed by gel filtration chromatography. Multiple reactions were conducted where P22 Xis was titrated with increasing protein concentrations at a constant DNA concentration (Fig. 8 and Table 1). For the first sample set, the GF50 DNA containing the X1, X2, X3, and X4 sites was used. The complex migrated with a retention volume of 12.2 ml and an observed DNA molecular mass of 131,000 Da. The complex between Xis and the GF50 DNA had a retention volume of 11.7 ml when a ratio of four Xis molecules to one DNA molecule was tested. This elution volume corresponds to an apparent molecular mass of 172,000 Da or a calculated complex of 3.2 Xis protomers and one 50-mer double-stranded DNA (see Discussion). Experiments using GF50 DNA with protein-to-DNA ratios of 2 and 17.6 gave similar results (Fig. 8A and Table 1). It is important to note that DNA molecular weight estimates based on elution volumes of proteins of known size during calibration should not be taken literally. However, these data may provide an estimate of the number of stable complexes and their sizes relative to one another (see Discussion).


Figure 8
View larger version (23K):
[in this window]
[in a new window]

  		FIG. 8. (A) Gel filtration chromatography of P22 Xis protein and GF50 DNA. DNA-to-protein ratios are shown in the key above. Reaction conditions are described in Materials and Methods. (B) Gel filtration chromatography of P22 Xis protein and GF28 DNA. DNA-to-protein ratios are shown in the key above. Reaction conditions are described in Materials and Methods.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Data from gel filtration chromatography experiments on complexes between GF50 DNA and P22 Xis and between GF28 DNA and P22 Xis proteina

The same experiment was conducted on GF28 DNA, which contains the X3 and X4 sites (28 bp of DNA) (Fig. 5). GF28 DNA alone migrated with a retention volume of 14.0 ml and an observed DNA molecular mass of 58,000 Da. The complex between Xis and the GF28 DNA molecule had a retention volume of 13.1 ml when a ratio of two Xis molecules to one DNA molecule was used. This corresponds to an apparent molecular mass of 87,000 Da, or 2.3 Xis protomers per DNA. At higher ratios of protein to DNA, no significant further changes in the complex were detected (Fig. 8B and Table 1). A ratio higher than two to one of P22 Xis protein to DNA was necessary to completely shift the free GF28 DNA into the complex. This was expected, as we had shown that P22 Xis binds the truncated footprint site more weakly than the larger site represented by GF50.

The solution behavior of short DNA fragments during chromatography is expected to be that of stiff rods, and the elution volumes of the DNA do not accurately correspond to values derived from protein molecular weight standards. Although the study of DNA and protein complexes by this method is subject to error, estimates of the complex size can be obtained.

Our study provides evidence for a stoichiometry of up to four protomers binding to the P22 Xis footprint site. Over various protein-to-DNA ratios, only one multiprotein-DNA complex was isolated for both the GF50 and GF28 DNA fragments. The increase in protein-to-DNA ratios produced no significant changes in the migration of the largest complex above a ratio of four to one for GF50 and above a two-to-one ratio for GF28. At a two-to-one ratio of protein to DNA for GF50, the largest complex is present with just under one-half of the total DNA in the complex, suggesting that the largest possible complex contains four protomers. At less than a two-to-one ratio for GF50, the great majority of the DNA elutes as free DNA, reaffirming the cooperative nature of the complex, where at a low protein-to-DNA ratio, the complex is disfavored. As {lambda} Xis has been shown to bind as a trimer to 35 bases (1), it seems reasonable that four molecules of P22 Xis could bind over the span of the footprint (40 to 50 bases depending on the DNA tested and the protein concentration).


arrow
DISCUSSION
 
Models for P22 Xis-DNA interactions. Smith-Mungo et al. predicted that, by analogy with {lambda} Xis interactions with its DNA binding site, P22 Xis might bind to two direct repeats in the P arm at positions –118 to –81 between the P2 and P3 Int arm-type sites (30). We confirmed that P22 Xis indeed binds to that region. However, based on our footprinting results, P22 Xis bound a larger primary site with secondary lower-affinity complexes forming upstream and downstream of the primary binding site at high protein concentrations. Titration experiments showed the formation of at least three complexes. Furthermore, the Hill coefficient of 2.2 suggested a minimum of two binding sites (35) and P22 Xis indeed bound one-half of the primary footprint site with an affinity poorer than that of binding to the whole site. Therefore, the binding of two to four protomers is required to explain the pattern of binding to the DNA observed using EMSA.

The P22 Xis footprint contains four similar direct, helically repeated sequences, three of which are identical. Since P22 Xis was predicted to contain a helix-turn-helix domain, and the well-characterized {lambda} Xis makes use of a winged-helix DNA binding motif, we developed two models for DNA recognition of its DNA binding site, which are shown in Fig. 9. Head-to-tail (directly repeated sequences) and head-to-head (palindromic sequences) binding possibilities were considered.


Figure 9
View larger version (8K):
[in this window]
[in a new window]

  		FIG. 9. Schematic showing P22 Xis footprint sequence and possible recognition sequences along with models of Xis binding. In the diagram at the top, Int arm-type binding sites P1, P2, and P3 and direct repeats (X1 to X4) are shown. Model 1 shows the direct repeats, which are within the footprint and are labeled X1, X2, X3, and X4. Model 2 shows a set of imperfect palindromes within the P22 Xis footprint. The large black arrows represent possible binding orientations of P22 Xis.

The simplest palindromic (head-to-head) model would require that two homodimers recognize two palindromes containing a total of four half-sites. This model is shown as number two in Fig. 9 and fits both the minimal binding study and gel filtration data. This model also explains the EMSA shifts in Fig. 2D and E but does not explain the gel shifts in Fig. 2B and C or the cross-linking data. In Fig. 2E, complexes 1A and 1B would arise from the binding of a dimer to either the first or to the second palindrome. Complex 2 could then form when both palindromes are bound by Xis. Since the two palindromes contain sequence differences, they might be expected to have different binding affinities. The palindromes might promote protein-protein interactions by increasing the overall affinity to the entire site through cooperative DNA binding effects. In contrast, complexes in Fig. 2B and C migrate progressively slower with each increase in protein concentration, which is not easily explained by a palindromic model.

The direct-repeat model (X1 to X4) would involve head-to-tail binding. A simple model involves a protomer binding each repeat, yielding a stoichiometry of four protomers in the primary footprinted region. As evidence, the footprint of the top strand in Fig. 3A shows a distinct large protected region with a hypersensitive site at each "T" of the direct repeat. In contrast, the footprint of the bottom strand in Fig. 3B shows only limited protection of individual bases with flanking hypersensitive sites. The direct repeats suggest that P22 Xis binds in a head-to-tail fashion, rather than in the palindromic head-to-head arrangement. In a palindromic model, one might expect footprinting to bottom and top strands to be mirror reflections of one another. This was not observed, and the very large difference between footprinting patterns for the top and bottom strands lends credence to the direct-repeat model. A recent crystal structure of {lambda} Xis shows that it binds as a microfilament of three Xis proteins on one side of the helix. It is possible that, in general, proteins of the Xis class form single-sided microfilaments around which the DNA is wrapped (1, 25). The direct-repeat model also suggests that monomers could bind cooperatively, forming successive dimers, trimers, and longer nucleoprotein filaments. The palindromic model suggests that dimers could bind independently and would probably form even-numbered complexes (of two, four, or six) in protein-protein cross-linking. Given that our studies indicate that Xis can behave as a monomer, dimer, trimer, and larger multimers in solution and that the footprint shows that the strands are not recognized equally, the evidence, in total, favors the direct-repeat model.

P22, L5, and {lambda} Xis attP DNA binding site organization and function. Lewis and Hatfull showed that bacteriophage L5 Xis recruits L5 Int to its P2 site, stimulating excision, and that the L5 P3 site is not required (22). When the DNA binding sites of P22 Int and Xis are aligned with those from the SSR system of mycobacteriophage L5, their organizations are nearly identical (Fig. 10). A detailed review of the X1-to-X4 sequences of the P22 Xis DNA binding site shows a similarity to the L5 Xis binding sites in spacing and repetition of four helically repeated sequences (Fig. 10). Both proteins appear to bind with a stoichiometry of four protomers to the primary site. The P22 repeat is 6 bp, with 4-bp spacers between repeats, and the L5 repeat is 7 bp with 3-bp spacers. The distances between the P2 and P3 arm-type DNA binding sites are similar, with L5 containing two fewer base pairs. Not only are the Int arm-type DNA binding sites situated in the same relative locations, but the Xis footprints also map to the same areas relative to the position of the crossover site.


Figure 10
View larger version (15K):
[in this window]
[in a new window]

  		FIG. 10. Alignment of attP sites from lambda, P22, L5, HP1, P2, and Tn916. The sites are aligned at the right, along the middle of the two Int cleavage sites. DNA core sites are in green and their footprints in yellow. All other Int binding sites not at the point of cleavage are in blue. Xis proposed or known binding recognition sites are shown in red. The Xis footprint-protected bases are marked by the black line below the red recognition site for the strand shown. The second black line below the first represents protected bases for the cDNA strand, which is not shown (HP1 and P2 only). IHF or host factor binding sites are shown in gray. The mIHF protection site in L5 is shown in purple because no recognition sequence near the protection site has been proposed.

In the {lambda} SSR system, {lambda} Xis recruits the {lambda} Int to the P2 Int arm-type DNA binding site. By doing so, {lambda} Xis regulates excision by stimulating Int to favor the P2 arm-type binding site on attR and thus directing it toward the excision pathway. A comparison of the {lambda} and P22 attR sites shows that they have somewhat different organizations (Fig. 10). {lambda} attR contains two Int arm-type binding sites, with the P2 abutting the Xis binding site. P22 attR contains three P22 Int arm-type binding sites. The P2 site is adjacent to one side of the primary Xis binding site and the P3 site is on the other side. Though the organizations of the {lambda} and P22 attR sites are different, these sites may in fact function in similar manners. In both cases, the Xis protein binds immediately downstream of an arm-type binding site (P2 site in both cases) which is nearly equidistant from the core and cleavage region. Although the P1 site in {lambda} attR is further away from the core than the P22 P1 site and P22 contains the extra P3 site, neither difference may be important for excision if the common P2 and Xis DNA binding sites are the key regulatory attR binding sites for regulating excision. The similar localization of Xis binding sites relative to the crossover sites, the ability to bend DNA by Xis, the conservation of Xis direct repeats, and the DNase I hypersensitivity of the P22 P2 site suggest that P22 Int and P22 Xis might simultaneously bind the P22 P2 site and P22 Xis DNA binding sites, respectively, as in the case of {lambda} and L5 SSR systems.

P22 Xis binding to secondary sites might inhibit reintegration. The footprinting results from this study showed that P22 Xis binds to secondary sites upstream and downstream of the primary footprint site. At high protein concentrations, P22 Xis binds secondary DNA binding sites. EMSA experiments carried out under similar protein concentrations showed a single slower-migrating complex. Control experiments using EMSA ruled out simple DNA looping, although they could not rule out DNA wrapping around a protein filament. Comparison of the primary binding site with several secondary binding sites (Fig. 3D) with high Xis protein concentration revealed a possible degenerate recognition sequence, "TTGCAT." A recent study by Papagiannis et al. of intracellular concentrations of {lambda} Xis showed that the protein increases to over 15,000 copies per cell, or a concentration of greater than 30 µM, during excision (25). Therefore, if P22 Xis reaches such an intracellular level during excision, binding of secondary sites would become relevant in vivo.

The secondary sites also provide a possible mechanism for inhibiting reintegration after excision. Early during the expression of Xis, lower concentrations of the protein could stimulate excision, as discussed above. As the protein concentration continues to increase, binding to secondary sites could inhibit the reverse integration reaction. Indeed, the P22 Xis secondary binding sites specifically occupied the P22 Int DNA binding sites, including P2, C, and C', that would be critical for integration. This binding pattern would then serve to insure that the switch to lytic growth is irreversible by preventing nonproductive integration events during lytic growth. Future studies will be needed to clarify this hypothesis.

Organization of Xis DNA binding sites among different SSR families shows a common theme. We aligned the attP sites of several SSR systems based on the longest and most significant primary Xis binding site. If one compares the DNA binding sites of {lambda}, L5, HP1, P2, Tn916, and P22 excision systems by aligning the att sites containing the major Xis binding sites and the Int cleavage sites in the core regions (Fig. 10), common loci of Xis binding sites relative to Int arm-type, Int core-type, and host factor binding sites emerge. Two different types of Xis systems are revealed by the alignment. We define {lambda}, L5, HP1, and P22 as host factor-dependent (type I) Xis binding systems because their attP DNAs each contain at least one host factor binding site between the major Xis binding site and the Int core site. Among the type I Xis systems, the Xis recognition site features vary, but the footprints appear to cover similar lengths of DNA (Fig. 10). We define the P2 and Tn916 excision systems as host factor-independent (type II) Xis binding systems because their attP DNAs lack host factor binding sites between the Xis binding site and the Int core site. The type II Xis sites and footprints cover more DNA between the Int arm-type site and the core site; they extend to the Int core-type sites. Type II Xis binding sites also cover the aligned DNA region that contains a host factor site in the type I systems. Furthermore, type II systems may use a larger number of Xis protomers to achieve this larger footprint. We suggest that the larger number of Xis protomers may substitute the function provided by the host factor in bending.

The {lambda} Xis binding has been described as a micronucleoprotein filament and, indeed, all Xis proteins appear to footprint a large segment of DNA (Fig. 10). In both type I and type II systems, the Xis proteins could stimulate the excision reaction by similar mechanisms. They may form nucleoprotein filaments that, by bending DNA, help form the active architecture of the excisive intasome. In addition, they could act by inhibiting Int binding to another close-by arm-type binding site on the same att arm, as in the case of HP1 Cox blocking binding of HP1 Int to IBS5 (14, 15), or by cooperatively recruiting Int, as in the case of {lambda} and L5 (22, 24, 31). Finally, it will be interesting to see in future studies if the Xis class of proteins can all form microfilaments, stimulating excision at low protein concentrations, and extend their filaments at high protein concentrations, inhibiting reintegration.


arrow
ACKNOWLEDGMENTS
 
This work was supported by grant GM28717 from the National Institutes of Health.

We thank Jannette Carey for helpful discussions and advice and Sandra Szegedi and Chad Thomas for their insights on protein purification.


arrow
FOOTNOTES
 
* Corresponding author. Mailing address: Department of Microbiology, University of Illinois at Urbana-Champaign, Urbana, IL 61801. Phone: (217) 333-7289. Fax: (217) 244-6697. E-mail: jeffgard{at}uiuc.edu Back

{triangledown} Published ahead of print on 23 May 2008. Back


arrow
REFERENCES
 
    1
  1. Abbani, M. A., C. V. Papagiannis, M. D. Sam, D. Cascio, R. C. Johnson, and R. T. Clubb. 2007. Structure of the cooperative Xis-DNA complex reveals a micronucleoprotein filament that regulates phage lambda intasome assembly. Proc. Natl. Acad. Sci. USA 104:2109-2114.[Abstract/Free Full Text]
    2. 2
  2. Amersham Biosciences. 2002. Gel filtration principles and methods. Amersham Biosciences, Piscataway, NJ.
    3. 3
  3. Ausubel, F. M. 1987. Current protocols in molecular biology. Greene Publishing Associates, Media, PA.
    4. 4
  4. Ausubel, F. M. 1999. Short protocols in molecular biology: a compendium of methods from current protocols in molecular biology, 4th ed. Wiley, New York, NY.
    5. 5
  5. Azaro, M. A., and A. Landy. 2002. Lambda integrase and the lambda Int family, p. 118-148. In N. L. Craig, R. Craigie, M. Gellert, and A. M. Lambowitz (ed.), Mobile DNA II. ASM Press, Washington, DC.
    6. 6
  6. Ball, C. A., and R. C. Johnson. 1991. Multiple effects of Fis on integration and the control of lysogeny in phage lambda. J. Bacteriol. 173:4032-4038.[Abstract/Free Full Text]
    7. 7
  7. Bushman, W., S. Yin, L. L. Thio, and A. Landy. 1984. Determinants of directionality in lambda site-specific recombination. Cell 39:699-706.[CrossRef][Medline]
    8. 8
  8. Cho, E. H., R. Alcaraz, Jr., R. I. Gumport, and J. F. Gardner. 2000. Characterization of bacteriophage lambda excisionase mutants defective in DNA binding. J. Bacteriol. 182:5807-5812.[Abstract/Free Full Text]
    9. 9
  9. Cho, E. H., R. I. Gumport, and J. F. Gardner. 2002. Interactions between integrase and excisionase in the phage lambda excisive nucleoprotein complex. J. Bacteriol. 184:5200-5203.[Abstract/Free Full Text]
    10. 10
  10. Cho, E. H., C. E. Nam, R. Alcaraz, Jr., and J. F. Gardner. 1999. Site-specific recombination of bacteriophage P22 does not require integration host factor. J. Bacteriol. 181:4245-4249.[Abstract/Free Full Text]
    11. 11
  11. Cores de Vries, G., X. S. Wu, and E. Haggard-Ljungquist. 1991. Genetic analysis of the DNA recognition sequence of the P2 Cox protein. J. Virol. 65:4665-4669.[Abstract/Free Full Text]
    12. 12
  12. Dodd, I. B., M. R. Reed, and J. B. Egan. 1993. The Cro-like Apl repressor of coliphage 186 is required for prophage excision and binds near the phage attachment site. Mol. Microbiol. 10:1139-1150.[CrossRef][Medline]
    13. 13
  13. Esposito, D., and J. J. Scocca. 1994. Identification of an HP1 phage protein required for site-specific excision. Mol. Microbiol. 13:685-695.[CrossRef][Medline]
    14. 14
  14. Esposito, D., and J. J. Scocca. 1997. Purification and characterization of HP1 Cox and definition of its role in controlling the direction of site-specific recombination. J. Biol. Chem. 272:8660-8670.[Abstract/Free Full Text]
    15. 15
  15. Esposito, D., J. S. Thrower, and J. J. Scocca. 2001. Protein and DNA requirements of the bacteriophage HP1 recombination system: a model for intasome formation. Nucleic Acids Res. 29:3955-3964.[Abstract/Free Full Text]
    16. 16
  16. Kim, J., C. Zwieb, C. Wu, and S. Adhya. 1989. Bending of DNA by gene-regulatory proteins: construction and use of a DNA bending vector. Gene 85:15-23.[CrossRef][Medline]
    17. 17
  17. Landy, A. 1989. Dynamic, structural, and regulatory aspects of lambda site-specific recombination. Annu. Rev. Biochem. 58:913-949.[Medline]
    18. 18
  18. Lathe, R., M. P. Kieny, S. Skory, and J. P. Lecocq. 1984. Linker tailing: unphosphorylated linker oligonucleotides for joining DNA termini. DNA 3:173-182.[Medline]
    19. 19
  19. Leong, J. M., S. Nunes-Duby, C. F. Lesser, P. Youderian, M. M. Susskind, and A. Landy. 1985. The phi 80 and P22 attachment sites. Primary structure and interaction with Escherichia coli integration host factor. J. Biol. Chem. 260:4468-4477.[Abstract/Free Full Text]
    20. 20
  20. Leong, J. M., S. E. Nunes-Duby, A. B. Oser, C. F. Lesser, P. Youderian, M. M. Susskind, and A. Landy. 1986. Structural and regulatory divergence among site-specific recombination genes of lambdoid phage. J. Mol. Biol. 189:603-616.[CrossRef][Medline]
    21. 21
  21. Lewis, J. A., and G. F. Hatfull. 2001. Control of directionality in integrase-mediated recombination: examination of recombination directionality factors (RDFs) including Xis and Cox proteins. Nucleic Acids Res. 29:2205-2216.[Abstract/Free Full Text]
    22. 22
  22. Lewis, J. A., and G. F. Hatfull. 2003. Control of directionality in L5 integrase-mediated site-specific recombination. J. Mol. Biol. 326:805-821.[CrossRef][Medline]
    23. 23
  23. Lewis, J. A., and G. F. Hatfull. 2000. Identification and characterization of mycobacteriophage L5 excisionase. Mol. Microbiol. 35:350-360.[CrossRef][Medline]
    24. 24
  24. Numrych, T. E., R. I. Gumport, and J. F. Gardner. 1992. Characterization of the bacteriophage lambda excisionase (Xis) protein: the C-terminus is required for Xis-integrase cooperativity but not for DNA binding. EMBO J. 11:3797-3806.[Medline]
    25. 25
  25. Papagiannis, C. V., M. D. Sam, M. A. Abbani, D. Yoo, D. Cascio, R. T. Clubb, and R. C. Johnson. 2007. Fis targets assembly of the Xis nucleoprotein filament to promote excisive recombination by phage lambda. J. Mol. Biol. 367:328-343.[CrossRef][Medline]
    26. 26
  26. Sam, M. D., D. Cascio, R. Johnson, and R. T. Clubb. 2003. Crystallization and preliminary X-ray crystallographic analysis of the excisionase-DNA complex from bacteriophage lambda. Acta Crystallogr. D 59:1238-1240.[CrossRef][Medline]
    27. 27
  27. Sam, M. D., D. Cascio, R. C. Johnson, and R. T. Clubb. 2004. Crystal structure of the excisionase-DNA complex from bacteriophage lambda. J. Mol. Biol. 338:229-240.[CrossRef][Medline]
    28. 28
  28. Sam, M. D., C. V. Papagiannis, K. M. Connolly, L. Corselli, J. Iwahara, J. Lee, M. Phillips, J. M. Wojciak, R. C. Johnson, and R. T. Clubb. 2002. Regulation of directionality in bacteriophage lambda site-specific recombination: structure of the Xis protein. J. Mol. Biol. 324:791-805.[CrossRef][Medline]
    29. 29
  29. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.
    30. 30
  30. Smith-Mungo, L., I. T. Chan, and A. Landy. 1994. Structure of the P22 att site. Conservation and divergence in the lambda motif of recombinogenic complexes. J. Biol. Chem. 269:20798-20805.[Abstract/Free Full Text]
    31. 31
  31. Swalla, B. M., E. H. Cho, R. I. Gumport, and J. F. Gardner. 2003. The molecular basis of co-operative DNA binding between lambda integrase and excisionase. Mol. Microbiol. 50:89-99.[CrossRef][Medline]
    32. 32
  32. Thompson, J. F., and A. Landy. 1988. Empirical estimation of protein-induced DNA bending angles: applications to lambda site-specific recombination complexes. Nucleic Acids Res. 16:9687-9705.[Abstract/Free Full Text]
    33. 33
  33. Thompson, J. F., L. Moitoso de Vargas, C. Koch, R. Kahmann, and A. Landy. 1987. Cellular factors couple recombination with growth phase: characterization of a new component in the lambda site-specific recombination pathway. Cell 50:901-908.[CrossRef][Medline]
    34. 34
  34. Warren, D., M. D. Sam, K. Manley, D. Sarkar, S. Y. Lee, M. Abbani, J. M. Wojciak, R. T. Clubb, and A. Landy. 2003. Identification of the lambda integrase surface that interacts with Xis reveals a residue that is also critical for Int dimer formation. Proc. Natl. Acad. Sci. USA 100:8176-8181.[Abstract/Free Full Text]
    35. 35
  35. Weiss, J. N. 1997. The Hill equation revisited: uses and misuses. FASEB J. 11:835-841.[Abstract]
    36. 36
  36. Yin, S., W. Bushman, and A. Landy. 1985. Interaction of the lambda site-specific recombination protein Xis with attachment site DNA. Proc. Natl. Acad. Sci. USA 82:1040-1044.[Abstract/Free Full Text]
    37. 37
  37. Yu, A., and E. Haggard-Ljungquist. 1993. The Cox protein is a modulator of directionality in bacteriophage P2 site-specific recombination. J. Bacteriol. 175:7848-7855.[Abstract/Free Full Text]

Journal of Bacteriology, September 2008, p. 5781-5796, Vol. 190, No. 17
0021-9193/08/$08.00+0     doi:10.1128/JB.00170-08
Copyright © 2008, American Society for Microbiology. All Rights Reserved.




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Mattis, A. N.
Right arrow Articles by Gardner, J. F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Mattis, A. N.
Right arrow Articles by Gardner, J. F.

Copyright © 2009 by the American Society for Microbiology.   For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»