DNA recombination with a heterospecific Cre homolog identified from co
http://www.100md.com
《核酸研究医学期刊》
1 Stowers Institute, 1000 E 50th Street, Kansas City, MO 64110 and 2 Department of Biochemistry and Molecular Biology, University of Kansas Medical Center, Kansas City, KS 66160, USA
* To whom correspondence should be addressed. Tel: +1 816 926 4432; Fax: +1 816 926 2068; Email: BLS@stowers-institute.org
+AY751747–AY751749 and AY753669
DDBJ/EMBL/GenBank accession nos+
ABSTRACT
Sequencing of the 7 kb immC region from four P1-related phages identified a novel DNA recombinase that exhibits many Cre-like characteristics, including recombination in mammalian cells, but which has a distinctly different DNA specificity. DNA sequence comparison to the P1 immC region showed that all phages had related DNA terminase, C1 repressor and DNA recombinase genes. Although these genes from phages P7, w39 and p15B were highly similar to those from P1, those of phage D6 showed significant divergence. Moreover, the D6 sequence showed evidence of DNA deletion and substitution in this region relative to the other phages. Characterization of the D6 site-specific DNA recombinase (Dre) showed that it was a tyrosine recombinase closely related to the P1 Cre recombinase, but that it had a distinct DNA specificity for a 32 bp DNA site (rox). Cre and Dre are heterospecific: Cre did not catalyze recombination at rox sites and Dre did not catalyze recombination at lox sites. Like Cre, Dre catalyzed both integrative and excisive recombination and required no other phage-encoded proteins for recombination. Dre-mediated recombination in mammalian cells showed that, like Cre, no host bacterial proteins are required for efficient Dre-mediated site-specific DNA recombination.
INTRODUCTION
The Cre recombinase of bacteriophage P1 is a member of the lambda integrase or tyrosine recombinase family of site-specific DNA recombinases. Because of its relatively simple biochemical requirements, its highly specific recognition of a 34 bp site not normally present in most genomes, and its ability to catalyze DNA recombination both in prokaryotes and in eukaryotes Cre has proven useful both in elucidating the mechanism of site-specific DNA recombination (1) and in developing powerful strategies for genetic engineering (2).
Large scale mutagenesis (3), in vitro evolution approaches (4–6) and the availability of several crystal structures for Cre (1), as well as comparative sequence analysis to other recombinases (7–9), have all contributed to an understanding of Cre's DNA specificity and mechanism of action. Still, a deeper appreciation of specificity and function could be had if close homologs of Cre recombinase were available for comparison. The elegant genetic dissection of specificity determination using the lambda and HK022 Int proteins attests to the power of such an approach (10,11). In this work, we sequenced the pac-c1 genomic regions of several P1-related phages to identify Cre homologs.
The P1 genome, the sequencing of which has recently been completed in its entirety (12), is relatively large for a temperate DNA bacteriophage: 95 kb. P1 is unusual among temperate bacteriophages in that it maintains itself as an extrachromosomal unit copy plasmid in the lysogenic state. Cre is expressed in P1 lysogens and its site-specific DNA recombination activity contributes to the stable maintenance of the P1 prophage during lysogeny. Cre resolves P1 dimers that arise by homologous recombination after DNA replication, thus helping to ensure segregation of a P1 monomer to each daughter cell at cell division (13). Because P1-related phages also maintain themselves as an extrachromosomal plasmid in the lysogenic state, we expected that they too would have a comparable recombinase function.
The P1 cre gene and its 34 bp recombination target site loxP lie in a relatively short interval of P1 DNA that includes two other phage functions with unusual features. To the left of cre is the immC immunity region of P1 that encodes the C1 repressor and several other immunity proteins that modify its action. ImmC, in turn, lies just to the right of the two genes for the P1 pacase or terminase and the pac site at which P1 DNA packaging begins. DNA packaging in P1 is unusual because protein recognition of the P1 DNA packaging site is regulated by DNA adenine methylation (dam). Understanding of P1 DNA packaging has led to the development of both high frequency P1 transducing phage strains (14) and an in vitro P1 packaging system for the construction of recombinant DNA libraries with large DNA inserts (15). Although immunity in P1 is orchestrated in a complex manner and includes several different immunity regions (16), including anti-repressor components, the C1 protein is itself unusual compared with other phage repressors in that it recognizes an asymmetric DNA-binding site (17). To the left of the P1 cre gene is c8, another immunity gene, followed by ref, a gene involved in the homologous recombination of short DNA repeats (18,19).
In this work, we sequenced and compared the 7 kb pac-c1 region from five P1-related phages, namely, the closely related phage P7 (20), the defective phage p15B (21), transducing phage w39 (22) and the P1-like transducing phage D6 isolated from Salmonella oranienburg (23). All exist as an extrachromosomal plasmid in Escherichia coli and, with the exception of D6, all are members of plasmid incompatibility group Y (22,24). P1, P7 and p15B genomic DNAs cross-hybridize (24), as do those of D6 and P1 (25). Interestingly, the immC regions of D6 and P1 do not cross-hybridize (25), suggesting that counterparts of P1 genes in this region of D6 may be diverged or absent. Sequence comparisons presented here showed that this region is similarly organized in all phages and encodes DNA packaging, repressor and recombinase genes. Sequence identity was very high for this 7 kb region amongst all the phages except for D6. Characterization of the D6 recombinase showed that although it is closely related to the P1 Cre recombinase, it exhibits a distinctly different DNA specificity that can be used either to remove a marker gene or to activate gene expression in vivo.
MATERIALS AND METHODS
Bacteria and phage
Bacterial strains and lysogens used in this work are listed in Table 1. Bacteria were propagated in Luria–Bertani broth (26) with appropriate antibiotics: streptomycin (10 μg/ml), ampicillin (100 μg/ml) and chloramphenicol (34 μg/ml). Zeocin (Invitrogen) was used at a concentration of 10 μg/ml in Lennox broth (27). Phages were titered, maintained and propagated in the presence of 5 mM CaCl2. Lysogenization of DH5lacU169 by selection for CmR after infection with P1 CM (a gift from N. Sternberg) generated BS610.
Table 1. Bacterial strains
In general, phage stocks were prepared and titered using the indicator strain Sh-16. Strains lysogenic for P1 CM, P7 c1.9, w39 and D6 (Table 1) all spontaneously released a low number of phage after culture overnight that gave small plaques on Sh-16, indicating that these P1-like phages were capable of both lysogenic and lytic growth. Phage plate stocks (9–10 h at 37°C) were prepared with a fresh overnight culture of the indicated donor strain (28), and used immediately. For transduction, 0.1 ml of a fresh overnight of the recipient strain DH5lacU169 was infected with an aliquot of the transducing stock. After pre-adsorption for 5 min at room temperature, the cells were diluted to 1 ml with LB broth + 5 mM CaCl2, incubated 50 min at 37°C and then plated on selective medium.
DNA sequencing and plasmid construction
Circular plasmid DNA of bacteriophages P1 CM, P7 c1.9, w39, p15B and D6 was prepared using the Qiagen Large Construct Kit (Valencia, CA). For phages P1 CM, P7 c1.9, w39 and p15B, we designed PCR primers based on the published sequences of the P1 genes between lpa (formerly gene 10) and ref. This region encompasses the DNA packaging genes pacA and pacB, the c1 gene along with several other immunity genes, and the gene for Cre recombinase. For all these phages, at least several P1 primer pairs were identified for amplification of this interval. PCR fragments generated from each phage were sequenced directly and the resulting sequence information was used to design additional primers. A combination of PCR fragment sequencing and primer walking by direct sequencing from phage DNA was then used to obtain the complete sequence for each phage of the 7 kb region from pacA to the beginning of the ref gene.
For phage D6, none of the PCR primer pairs used for P1 or the other P1-related phages produced an amplified product. We therefore constructed a D6 library by partial Sau3A digestion of D6 DNA, cloning into the BamHI site of pUC19 and transformation of DH5. To eliminate smaller clones from this library, it was digested with EcoRI, size-selected by agarose gel electrophoresis to contain inserts of 1.5 kb and religated. Shotgun sequencing identified one clone having strong similarity to the P1 pacB gene. Specific primers were designed and were used to sequence directly from D6 DNA by primer walking in both directions to obtain 7 kb of flanking sequence.
DNA sequences were assembled and analyzed using Vector NTI (Invitrogen) and then compared with the corresponding region of the P1 genome, GenBank accession no. AF234172 . The sequences of the immC regions of phages P7 c1.9 (6751 bp), w39 (7208 bp), p15B (7094 bp) and D6 (7644 bp) have been assigned GenBank accession numbers AY751747 , AY751748 , AY751749 and AY753669 , respectively.
Oligonucleotide sequencing primers and linkers were synthesized by Integrated DNA Technologies (Coralville, IA) and restriction enzymes were from New England Biolabs (Beverly, MA). Annealing of the oligo's 5'-CTA GAT AAC TTT AAA TAA TTG GCA TTA TTT AAA GTT AG-3' and 5'-GAT CCT AAC TTT AAA TAA TGC CAA TTA TTT AAA GTT AT-3' and cloning into the XbaI and BamHI sites of pUC19 generated the rox plasmid pBS1051 (D6 sequence underlined). Similarly, oligo's 5'-CTA GCT ATA ACT TCG TAT AAT GTA TGC TAT ACG AAzG TTG-3' and 5'-TCG ACA ACT TCG TAT AGC ATA CAT TAT ACG AAG TTA TAG-3' were cloned into a pBluescript II KS (Stratagene) derivative using NheI and SalI to give the lox plasmid pBS516. The one nucleotide difference (double underline) of this lox site from loxP does not affect Cre-mediated recombination (29). Digestion of pBS1051 with either XbaI + AlwNI or BamHI + AlwNI and ligation with the XbaI–BamHI zeo fragment from pZeoSV (Invitrogen) generated the rox2-zeo plasmid pBS1080. The analogous loxP2 zeo plasmid pBS890 was constructed by V. Petyuk of this laboratory by blunt-ending the 600 bp FokI–SalI zeo fragment from pZeoSV and cloning it into the SmaI site of the loxP2 vector pBS246 (30). The dre gene was amplified with Pfu DNA polymerase (Stratagene) from D6 DNA with the oligo's 5'-AGA TGG TAC CAG GAG GAT ATC AAT ATG AGT GAA TTA ATT ATC TCT GG-3' and 5'-CTT TAG TCT AGA TTC ATT ATG AAT CCA TCA AGC GGC-3' (D6 coding region underlined), digested with KpnI and XbaI and cloned into the CmR arabinose-inducible vector pBAD33 (31) to generate pBS1081. The analogous pBAD33-cre construct has been described previously (6). All cloned oligos and PCR products were confirmed by DNA sequencing.
The dre gene was placed under the control of the EF1 promoter for mammalian expression by replacing the KpnI–XbaI GFP fragment of pBS377 (32) with the KpnI–XbaI dre fragment from pBS1081. The EF1-cre expression plasmid pBS513 (33) and the control CMV-lacZ plasmid p324 (34) have been described previously. To construct the enhanced green fluorescent protein (EGFP) expression vector pBS504, the following three DNA fragments were cloned between the unique HindIII and EcoRI sites of pBS397 (35): the EcoRI–KpnI EF1 fragment of pBS377, the KpnI–XbaI EGFP fragment of pEGFP-1 (Clontech, Palo Alto, CA) and the XbaI–EcoRI fragment from pBS377 carrying the polyA signal. The EcoRI–HindIII rox2 zeo cassette from pBS1080 was blunt-end ligated into the EcoRV site lying between the EF1 promoter and the EGFP gene in pBS504 to generate the rox recombination reporter plasmid pBS1083.
DNA analysis
We performed PSI-BLAST searches (36) to determine closest homologs of identified genes. The bendability/curvature-propensity plots were calculated with the bend.it server, using DNase I-based bendability parameters (37) and the consensus bendability scale (38).
Site-specific DNA excision
Plasmids pBAD33, pBAD33-dre or pBAD33-cre were electroporated into DH5 cells containing either the reporter plasmid pBS890 or pBS1080, incubated for 1 h at 37°C in SOB media (6) and then plated, selecting for CmR ApR to ensure retention of both plasmids in the resulting transformants. To ensure that recombinase was not expressed at inordinately high levels, the pBAD plasmids were used without overt arabinose induction (9). The next day, colonies were individually tested for drug resistance markers by growth on appropriate antibiotic containing plates.
CHO-K1 cells were transfected using Polyfect (Qiagen) with 1.5 μg DNA per well of a 6 well dish as recommended by the manufacturer. Co-transfections used a 9:1 ratio of Cre, Dre or lacZ expression vector DNA to either pBS504 or pBS1083 as indicated. Fluorescence was monitored 2 days after transfection with a Leica DMR microscope mounted with an Optronics Magnafire digital camera, and DNA was then prepared for PCR analysis. Recombination was detected by PCR (31 cycles for 30 s at 94°C, 30 s at 60°C, 60 s at 72°C) using the sense EF1 primer KC315 5' GCTTGGCACTTGATGTAATTCTCCTTG 3' and the antisense EGFP primer BSBS382 5' GGTCAGCTTGCCGTAGGTGGC 3'. Predicted product sizes are 302 bp for pBS504, 1704 bp for pBS1083 (not observed because of the short cycling times used) and 389 bp for the Dre excision product of pBS1083.
RESULTS
Even though the immC region of phage P1 does not cross-hybridize to phage D6 DNA, it seemed likely that D6 would carry a site-specific DNA recombination system like that of P1, and that this recombinase/recombinase recognition site would lie in the D6 immC region. We therefore checked for the presence of a recombinase activity genetically. If there were a Cre-like D6 recombinase, then infection of the D6 clone plasmid library with phage D6 would ‘pick-up’ relevant plasmid clones as a result of low level site-specific integrative recombination between phage DNA and plasmids carrying a D6 recombination site. Thus, D6 would be able to transduce relevant ApR clones to a new bacterial recipient where they would be excised from the phage genome by the D6 site-specific recombinase to take up plasmid residency. A similar strategy was used previously to pick-up loxP plasmid clones using phage lambda carrying the loxP-cre region of P1 (39). Table 2 shows that D6 could transduce the ApR marker from a library of D6 clones, but not from a strain having an anonymous randomly chosen D6 cloned insert or from a non-plasmid strain, Sh-16. These results indicate that, like P1, D6 carries a site-specific DNA recombinase that catalyzes both integrative and excisive recombination.
Table 2. D6 ‘pick-up’ identification of a D6 recombination site
To compare the recombinase genes, as well as the adjacent c1 and pac genes, we sequenced the 7 kb immC region from four P1-related phages: P7 c1.9, w39, p15B and D6. For the first three of these phages, the gene organization of this region was identical to that of phage P1, whereas that of phage D6 had a similar structure (Figure 1). To confirm that this D6 region did indeed harbor the recombinase site that we had detected previously, we sequenced six randomly chosen plasmids transduced by D6, naming them Pick-Up clones PU-5, etc. All carried overlapping sequences from the same immC DNA region we had sequenced (Figure 1), strongly indicating that each carried a D6 DNA site for site-specific recombination.
Figure 1. Comparison of the pac-c1 regions of P1 and D6. Shown is a 7623 bp region of P1 that includes the region from lpa (gene 10) to ref. The equivalent lpa-c8 regions of P7, w39 and p15B showed the same genes and gene order as shown for P1 except that in P7 the imcA and imcB genes were fused. As discussed in the text, gene products of this region from these other phages were 89–100% similar to their P1 counterparts. Also shown is the corresponding 7644 bp region from D6. The degree of similarity between individual genes of P1 and D6 is indicated according to the color scale shown. P1 genes in white have no D6 equivalent in this region, and D6 genes in black have no P1 equivalent. Below and aligned with the D6 map is diagrammed the DNA insert of several PU clones obtained by site-specific recombination.
The overall gene organization of the sequenced region from D6 is similar to the pac-ref region of P1 despite low sequence identity. On the left (Figure 1), there is a gene showing 24% similarity to P1's lpa (late promoter activating protein) or gene 10. This is followed by DNA packaging genes (pacA and pacB), a c1 repressor gene and a D6 recombinase ‘dre’ similar to the P1 cre gene. We discuss these genes in more detail below. There are also several non-conserved open reading frames (ORFs) in this region of the D6 genome, and an insertion of a gene similar to the deoxyuridine 5'-triphosphate nucleotidohydrolase of the photosynthetic bacterium Rubrivivax gelatinosus (dut; Figure 1).
DNA packaging genes
P1 packages its DNA using a terminase composed of the PacA small subunit and the PacB large subunit. Table 3 compares the PacA's and PacB's for each of the 5 P1 family phages. Although PacA was very similar (97–100%) for phages P1, P7, w39 and p15B, the PacA protein of D6 was 57 amino acids longer and showed only 18% similarity to PacA from any of the other phages (see also Figure 2). D6 PacB was somewhat more similar to the PacB proteins of the other P1 family members (53–54%) and was slightly larger. Interestingly, the other four phages form two clear subgroups: P1 and P7 are nearly identical to each other, likewise w39 and p15B are nearly identical, with the two subgroups 89% similar to each other. In addition, the defective phage p15B carried a nonsense mutation that truncates the C-terminus of PacB.
Table 3. DNA packaging genes
Figure 2. Comparison of the P1 and D6 pacA genes. Dam methylation sites for both genes are shown as vertical lines above the gene. A blowup of the 162 bp P1 pac site that includes two clusters of dam sites is shown above the P1 pacA gene, with dam sites represented as black boxes, the IHF-binding site as a white rectangle and the region of cleavage as a vertical arrow. The 397 amino acid P1 PacA protein and the 454 amino acid D6 PacA protein are 18% similar: insertions into the 1194 bp P1 pacA gene relative to the 1365 bp D6 pacA gene are shown in yellow, insertions into D6 pacA relative to P1 pacA are shown in orange. An asterisk at nucleotide position 182 of D6 pacA marks the maximum of a curvature-propensity plot calculated with DNase I-based trinucleotide parameters (37).
P1 headful DNA packaging proceeds from a specific 162 bp pac site located within the 5' end of the pacA structural gene (Figure 2). The site consists of two clusters of hexanucleotide repeats flanking a central region where DNA cleavage occurs (40). Each of the hexamer repeats has a core 5'-GATC dam methylation site. PacA protein binds to the hexamer repeats in a DNA methylation-dependent manner and then associates with PacB to loop the two binding domains in an IHF and HU-dependent manner prior to DNA cleavage (41,42). In phages P7, w39 and p15B, this 162 bp region differed from that of P1 by 4, 9 and 2 bp, respectively, but the integrity of the seven hexamer repeats was completely preserved, suggesting that all four of these phages package DNA in the same manner. Similarly, the IHF-binding site adjacent to the region of DNA cleavage was identical in w39 and P1, although in P7 and p15B this site displayed a one nucleotide change from 5'-AAACAAAGAGTTA to 5'-AAACAGAGAGTTA (change underlined).
The low degree of similarity between D6 and P1 PacA proteins suggests that their DNA-binding specificities may differ. In addition, there was no region of clustered dam sites characteristic of the P1 pac site and no consensus IHF-binding site in the D6 pacA gene (Figure 2), further suggesting a difference in the DNA recognition specificities of the P1 and D6 terminases. Interestingly, though, there was a potentially curved DNA sequence in the 5' region off the D6 pacA gene (asterisk, Figure 2). The curvature-propensity plot, calculated with DNase I-based trinucleotide parameters, contained one peculiar maximum in this region, whose magnitude (14.7°/helical turn) exceeded the value calculated for Columba risoria bent satellite DNA (13.5°/helical turn). No such potentially curved DNA was detected in the pacA genes of P1, P7, w39 or p15B.
C1 repressor and immunity
There was little difference in the immunity genes of the immC region for the four phages P1, P7, w39 and p15B (Figure 1). The C1 repressors of P1 and P7 had previously been shown to be identical (43), so the three amino acid changes (A110V, P190L and D277S) in P7 c1.9 are likely specific to this P7 temperature-sensitive repressor. Aside from a K268R change in w39, the p15B and w39 C1 repressors were identical to that of P1. The Coi (c one inactivator) protein sequence was identical for all four of these phages except for an A62T substitution in p15B. The predicted imcA and imcB gene products were also identical except for an A27T difference in P1. Interestingly, these two genes were fused in P7. Some variation was seen for C8 among phages P1, P7, w39 and p15B, but all showed 85% similarity. On the other hand, the D6 C1 protein was only 16% similar to P1 C1. Moreover, in phage D6, the distance between c1 and the recombinase gene dre was much shorter than the corresponding region in P1, and the coi and imcA/imcB genes were missing.
Recombinase
Among the four phages P1, P7, w39 and p15B, the 343 amino acid Cre recombinase was highly conserved. The w39 Cre differed from P1 Cre by the single amino acid change T206A, P7 Cre differed from P1 Cre by the two changes A178S and G280D, and the p15B Cre differed from P1 Cre by three changes: P107L, A249S and A252P. Similarly conserved among these four phages was cra (putative cre associated function), an ORF of unknown function adjacent to cre originally designated as orf1 (44). The putative P1 Cra protein differed from that of P7, w39 and p15B by 3, 1 and 2 amino acids, respectively. In accord with the nearly identical Cre recombinases of these four phages, all of them carried an identical 34 bp lox site located midway between c1 and cre.
In contrast, D6 displayed a recombinase gene only 39% similar to the P1 cre gene and no 34 bp lox site anywhere in this 7.6 kb region of DNA (Figure 1). This suggested that the dre might recognize a recombination site distinct from lox. As noted above, the interval between c1 and dre was much shorter in D6 than the corresponding region in P1, and no ORF corresponding to cra was found. From the D6 pick-up experiments, we knew that a recombination site must lie within the D6 sequence present in the PU7 clone (Figure 1), a 2.1 kb fragment that includes both dre and the interval between c1 and dre. We therefore inspected the 394 bp c1-dre interval for a candidate recombination site for Dre.
The P1 lox site consists of two 13 bp inverted repeats flanking an asymmetrical 8 bp spacer region that imparts an overall directionality to the recombination site. In the c1-dre interval from D6, we detected two DNA sequences resembling this structure. Figure 3 shows a 69 bp portion of the c1-dre interval within which is a 32 bp sequence having two perfect 14 bp inverted repeats (solid arrows) separated by 4 bp. Just abutting this potential Dre recombination site was a similarly configured 18 bp DNA sequence with less perfect inverted repeats (dashed line arrows).
Figure 3. The D6 c1-dre integenic region. The 394 bp region from D6 between c1 and dre is diagrammed along with the single dam site of this region. The sequence shown is a 70 bp portion of this region suspected to include the D6 rox site because of the presence of inverted repeat elements. Repeat elements of a DNA sequence shown not to be the recombination site are indicated by dashed arrows, and the repeat elements of the actual rox site are indicated by solid arrows. The two DraI sites are shown in lower case.
To determine the involvement of either of these DNA sequences in Dre-mediated recombination, we constructed several test plasmids and evaluated their ability to recombine with D6 using the D6 plasmid transduction assay. Table 4 shows that D6 transduces PU7, but not pUC19, from a recA host to a recA recipient at high frequency. Taking advantage of a fortuitous DraI site present in each 14 bp inverted repeat (Figure 3, upper case), we mutated the 32 bp candidate site by deleting the 12 bp between the DraI sites to generate a PU7 derivative, PU7-Dra1. The lack of D6 transduction of this plasmid (Table 4) indicated that the integrity of the 32 bp region was necessary for recombination and suggested that the adjacent shorter imperfect inverted repeat region was not a recombination site. The 32 bp region might thus be the region of crossover (X-over) recombination or rox site for Dre recombinase. To test this, we placed this sequence on pUC19 and tested its ability to undergo recombination with D6. Transduction of pUC19 carrying the putative rox site occurred at high frequency (Table 4), confirming that the rox sequence was sufficient for recombination. Moreover, no transduction of a pUC19-lox plasmid was obtained, indicating that D6 cannot recombine with the P1 lox site.
Table 4. Plasmid transduction by phage D6
We wanted to establish whether or not the Dre recombinase was the only D6-encoded protein required for recombination at rox. As a first step in this direction, we placed the dre structural gene under the control of the arabinose-inducible promoter in plasmid pBAD33. SDS–PAGE and western blot analysis showed that this construct expressed a Cre-sized protein of 36 kDa that cross-reacted with a polyclonal antibody to Cre (Figure 4).
Figure 4. Western blot detection of Dre recombinase. Whole cell extracts were prepared from mid-log cultures of E.coli DH5 containing pBAD-dre, pBAD-cre or a pBAD vector with no insert after induction for 2 h with 0.2% L-arabinose. Following SDS–PAGE western blot analysis was with a polyclonal rabbit antibody prepared against purified Cre recombinase (45). Size markers in kDa are shown to the right.
We assayed Dre-mediate recombination using a rox2 zeo construct (Figure 5) that carries two directly repeated rox sites flanking zeo, a gene that confers resistance to the antibiotic zeocin. Cells carrying this construct become sensitive to zeocin upon loss of zeo by excisive recombination at the rox sites. Table 5 shows that transformation of the rox2 zeo strain with the compatible CmR plasmid pBAD33-dre resulted in loss of zeocin resistance in all transformants. No loss of zeo was seen with a control plasmid having no insert or with a construct expressing Cre recombinase. Conversely, a strain carrying a lox2 zeo construct was refractory to excisive recombination by the Dre-expressing construct but readily underwent excisive recombination with loss of zeo when tranformed with the Cre-expressing construct. Sequencing of the rox plasmid from several zeocin-sensitive colonies confirmed that precise excisive recombination had occurred at the rox site. Thus, Dre mimicked Cre's ability to perform excisive recombination, but the two recombinases were heterospecific, i.e. Cre did not catalyze recombination at rox and Dre did not catalyze recombination at lox.
Figure 5. Recombinase-mediated excision. Reporter plasmids were constructed by placing the zeo gene (diagonal bars) between two identical directly repeated recombination sites (black triangles). Shown are the sequences of the 32 bp rox site and the 34 bp loxP site. Horizontal arrows indicate the inverted repeat elements; positions of nucleotide identity between rox and loxP are indicated by the black boxes; and vertical arrows show the sites of Cre cleavage that define the 6 bp overlap region of the lox site.
Table 5. Dre-mediated excision in E.coli
Unlike many members of the tyrosine recombinase family, Cre recombinase of phage P1 requires no accessory phage or bacterial proteins for recombination. This characteristic of Cre is illustrated by Cre's ability to catalyze DNA recombination in a variety of eukaryotic cells (45,46). To determine whether Dre also had no requirement for accessory bacterial proteins, and thus would be able to catalyze DNA recombination in a mammalian cell, we tested Dre's ability to recombine rox sites in CHO cells. Dre recombination was assayed by using a reporter plasmid that would express EGFP in transfected cells only if activated by recombination between two directly repeated rox sites (Figure 6A). In the reporter plasmid, a rox2 zeo cassette is inserted between the EF1 promoter and the EGFP gene, positioning both an upstream zeo gene and a polyadenylation site act to block EGFP expression. Dre-mediated recombination at the flanking rox sites would remove these blocks. To express Dre, we constructed a second plasmid vector in which the dre gene was placed under the control of the EF1 promoter. Figure 6B shows that co-transfection of CHO cells with the rox reporter construct and the Dre expression vector produced a significant number of green fluorescent cells (panel i), whereas there was no detectable fluorescence when the reporter was co-transfected with a Cre expression plasmid (panel ii). The frequency of fluorescent cells from Dre-mediated activation of the EGFP gene was similar to the transfection efficiency of an EGFP reporter plasmid having no rox2 zeo cassette (panel iii). PCR confirmed that recombination at the rox sites in cells co-transfected with the Dre expression plasmid and that no recombination occurred at the rox sites in cells co-transfected with the Cre expression plasmid (Figure 6C). Thus, Dre-mediated recombination requires no bacterial proteins for efficient DNA recombination at rox.
Figure 6. Dre-mediated recombination in mammalian cells. (A) Recombination activation of gene expression by Dre. Expression of the EGFP gene from the EF1 promoter in the reporter plasmid pBS1083 is blocked by the interposed zeo gene and polyadenylation site (An). Dre-mediated recombination at the flanking rox sites (black triangles) removes this ‘STOP’ signal to allow EGFP expression. Horizontal arrows indicate the PCR primers used to produce the 389 bp fragment diagnostic of recombination. (B) Activation of EGFP expression by Dre. Epifluorescence (panels i–iii) and differential interference contrast (panels iv–vi) images of CHO cells 36 h after DNA transfection. Panels i and iv, Dre expression plasmid pBS1081 + the rox2 STOP EGFP plasmid pBS1083; panels ii and v, Cre expression plasmid pBS513 + pBS1083; and panels iii and vi, control lacZ plasmid p324 + the EGFP plasmid pBS504. (C) PCR detection of recombination. DNA from the transfected CHO cells shown in (B) was amplified using the primers shown in (A). The 389 bp fragment is diagnostic of Dre recombination at the rox sites of pBS1083. The same primers give a 302 bp on the parental plasmid having no rox cassette inserted between the EF1 promoter and the EGFP gene. Lane 1, CHO cells transfected with the Dre expression plasmid pBS1081 + the rox2 STOP EGFP plasmid pBS1083; lane 2, CHO cells transfected with the Cre expression plasmid pBS513 + pBS1083; lane 3, CHO cells transfected with the control lacZ plasmid p324 + the EGFP plasmid pBS504; lane 4, EGFP plasmid pBS504; and lane 5, rox2 STOP EGFP plasmid pBS1083.
DISCUSSION
Here, we compare the sequence of the 7 kb region encompassing the terminase, repressor and recombinase genes from five related bacteriophages: P1, P7, w39, p15B and D6. With the exception of phage D6, the respective genes of this region were nearly identical. The few differences that did exist indicated that P1 and P7 form one subgroup and that w39 and p15B form a second closely related subgroup. Although this region of D6 displayed the same overall gene organization, the amino acid sequences of the D6 proteins showed much less similarity to their P1 counterparts than did these proteins from the other P1-like phages. In addition, there were several insertions and deletions of ORFs.
The conservation of the PacA protein and its cognate DNA packaging site pac (to which PacA binds) in four P1-like phages indicated that their DNA packaging proceeds in a near identical manner. In contrast, the DNA-binding PacA protein of D6 was only distantly related to the P1 PacA protein and the hallmark features of the P1 pac site sequence, two clusters of potentially dam-methylated hexamer sites flanking an 90 bp segment carrying an IHF-binding site, were not found in the D6 pacA gene, although there was a potentially curved DNA sequence in this region found. This suggests that although D6 terminase is clearly related to the P1 enzyme, D6 DNA packaging will differ in detail from the process used by phage P1.
The near identity of the C1 repressors of phages P1, P7, w39 and p15B was paralleled by the high conservation of Coi, a protein that binds to C1 and negatively regulates its activity. In contrast, the C1 protein of D6 showed considerable divergence from that of P1 and no coi gene was found in this region. The low sequence similarity of the repressors suggests that they bind different DNA sequences and this may contribute to the observed heteroimmunity of phages P1 and D6 (25).
There were six amino acid differences amongst the four Cre recombinases of phages P1, P7, w39 and p15B, with no one protein differing from P1 Cre by more than three changes. Of these six amino acid differences, four lay in flexible regions of the Cre protein that tolerate pentapeptide insertion (Figure 7), the other two were conserved amino acid changes. The presence of an identical 34 bp loxP site in these four phages also indicated that these proteins are functionally equivalent. On the other hand, the D6 recombinase Dre was only 39% similar to Cre. Even so, Dre is more closely related to Cre than are other members of the tyrosine recombinase family (9). The two recombinases are clearly heterospecific: Dre catalyzed recombination at a distinct 32 bp rox recombination site that was not recognized by Cre. Conversely, Dre did not recognize lox, the 34 bp recombination site recognized by Cre.
Figure 7. Comparison of P1 Cre and D6 Dre. Sequence identity is shown in yellow, and conserved amino acid changes are shown in green. Below the sequences is shown the known secondary structure (-helices as dark blue cylinders, ?-strands as orange arrows) of Cre (50). Below the depiction of Cre secondary structure is a functional map of Cre based on pentapeptide insertion (9): blue bars represent points of insertion which have little or no effect on recombinase activity, and red bars indicate points at which insertion abolishes recombinase activity.
Both the lox and rox recombination sites (Figure 5) are structurally similar, exhibiting two perfect inverted repeat elements (13 bp for lox, 14 bp for rox) flanking a short asymmetric intervening sequence or spacer that is typically used by members of the tyrosine recombinase family to provide a homology sensor for productive pairing and recombination. Both recombination sites share sequence identity except for the spacer region and a 3 bp heterology in the middle of the inverted repeat elements. The distance between the repeat elements is 8 bp for lox but only 4 bp for rox. This is intriguing for several reasons. The distance between the top strand and bottom strand cleavages (the ‘overlap’ region) is generally 6–8 bp for the tyrosine recombinases (1). For Cre, the overlap region is 6 bp (Figure 5) and even a 1 nt shortening of the distance between the inverted repeat elements to which Cre binds abolishes recombination (unpublished data). Subsequently, for the Dre/rox recombination system, one possibility is that the rox overlap region would be an unusual one for the tyrosine recombinases: 4 bp. Alternatively, the overlap region for the Dre/rox system may be a more typical 6 or 8 bp. If so, however, the resulting residency of the cleavage sites within the inverted repeat elements would mean that the nucleotides adjacent to the cleavage sites are identical for both the top and bottom strand cleavages. This would be unlike the Cre system, for which it has been proposed that it is precisely these flanking nucleotide differences at the cleavage sites that determine which strand is first cleaved (47).
The large patches of high similarity between Cre and Dre (Figure 7) in general correlate well with functionally important regions identified from pentapeptide map insertional mutagenesis of Cre. Conversely, several regions permissive for pentapeptide insertion (the N-terminus and the I-J and J-K interhelical regions) are less well conserved between Cre and Dre. We note that one region of sequence dissimilarity between Cre and Dre that is not tolerant of pentapeptide insertion is the linker between the M and N helices, perhaps indicating that for this linker it is specific length rather than specific sequence that is critical.
Genetic and biochemical exploitation of the amino acid sequence similarities and differences between Cre and Dre should help to better understand sequence-specific DNA recognition of the recombination site. One point of interest readily apparent in contrasting the Cre and Dre systems is the 3 bp heterology near the center of the inverted repeats (Figure 5, sequence in right repeat: 5'-ACG in lox, 5'-TTA in rox). Mutation of this guanosine in the lox site blocks Cre binding and recombination (48), as does mutation of R259 of Cre (49), which contacts the guanosine of this sequence (50). Dre displays proline instead of an arginine at this amino acid position, consistent with an important role for this region in rox-lox discrimination. We note that the nearby E262 position of Cre is important in specific recognition of the conserved sequence 5'-AA just next to this G residue in lox (Figure 5) (6) and that this amino acid position is conserved in Dre.
In conclusion, comparison of the pac-c1 regions of P1 and four other P1-related phages showed that the gene organization of this region is conserved, and that the gene products of this region are nearly identical for all phages with the exception of D6. From sequence comparisons, the specificity of the D6 DNA packaging, immunity and site-specific DNA recombinase functions are likely to be different from those of P1. We have confirmed this heterospecificity for the DNA recombinase Dre of D6, and we have also shown that Dre can catalyze site-specific DNA recombination in mammalian cells. We anticipate that further work with Dre will not only provide insight into the mechanism of site-specific recombination but will also lead to the generation of new tools for genetic engineering (51).
ACKNOWLEDGEMENTS
We thank M. Yarmolinsky and F. Blattner for communication of results prior to publication and A. Mushegian for comments on the manuscript.
REFERENCES
Van Duyne,G.D. ( (2002) ) A structural view of tyrosine recombinase site-specific recombination. In Craig,N.L., Craigie,R., Gellert,M. and Lambowitz,A.M. (eds.), Mobile DNA II, ASM Press, Washington, DC, pp. 93–117.
Sauer,B. ( (2002) ) Chromosome manipulation by Cre-lox recombination. In Craig,N.L., Craigie,R., Gellert,M. and Lambowitz,A.M. (eds.), Mobile DNA II, ASM Press, Washington, DC, pp. 38–58.
Wierzbicki,A., Kendall,M., Abremski,K. and Hoess,R. ( (1987) ) A mutational analysis of the bacteriophage P1 recombinase Cre. J. Mol. Biol., , 195, , 785–794.
Buchholz,F. and Stewart,A.F. ( (2001) ) Alteration of Cre recombinase site specificity by substrate-linked protein evolution. Nat. Biotechnol., , 19, , 1047–1052.
Santoro,S.W. and Schultz,P.G. ( (2002) ) Directed evolution of the site specificity of Cre recombinase. Proc. Natl Acad. Sci. USA, , 99, , 4185–4190.
Rüfer,A.W. and Sauer,B. ( (2002) ) Non-contact positions impose site selectivity on Cre recombinase. Nucleic Acids Res., , 30, , 2764–2771.
Argos,P., Landy,A., Abremski,K., Egan,J.B., Ljungquist,E.H., Hoess,R.H., Kahn,M.L., Kalionis,B., Narayana,S.V.L., Pierson,L.S. et al. ( (1986) ) The integrase family of site-specific recombinases: regional similarities and global diversity. EMBO J., , 5, , 433–440.
Swalla,B.M., Gumport,R.I. and Gardner,J.F. ( (2003) ) Conservation of structure and function among tyrosine recombinases: homology-based modeling of the lambda integrase core-binding domain. Nucleic Acids Res., , 31, , 805–818.
Petyuk,V., McDermott,J., Cook,M. and Sauer,B. ( (2004) ) Functional mapping of Cre recombinase by pentapeptide insertional mutagenesis. J. Biol. Chem., , 279, , 37040–37048.
Dorgai,L., Yagil,E. and Weisberg,R.A. ( (1995) ) Identifying determinants of recombination specificity: construction and characterization of mutant bacteriophage integrases. J. Mol. Biol., , 252, , 178–188.
Yagil,E., Dorgai,L. and Weisberg,R.A. ( (1995) ) Identifying determinants of recombination specificity: construction and characterization of chimeric bacteriophage integrases. J. Mol. Biol., , 252, , 163–177.
Lobocka,M.B., Rose,D.J., Plunkett,G.,III, Rusin,M., Samojedny,A., Lehnherr,H., Yarmolinsky,M. and Blattner,F.R. ( (2004) ) The genome of bacteriophage P1. J. Bacteriol., , 186, , 7032–7068.
Austin,S., Ziese,M. and Sternberg,N. ( (1981) ) A novel role for site-specific recombination in maintenance of bacterial replicons. Cell, , 25, , 729–736.
Sternberg,N.L. and Maurer,R. ( (1991) ) Bacteriophage-mediated generalized transduction in Escherichia coli and Salmonella typhimurium. Meth. Enzymol., , 204, , 18–43.
Sternberg,N. ( (1990) ) Bacteriophage P1 cloning system for the isolation, amplification, and recovery of DNA fragments as large as 100 kilobase pairs. Proc. Natl Acad. Sci. USA, , 87, , 103–107.
Heinrich,J., Velleman,M. and Schuster,H. ( (1995) ) The tripartite immunity system of phages P1 and P7. FEMS Microbiol. Rev., , 17, , 121–126.
Eliason,J.L. and Sternberg,N. ( (1987) ) Characterization of the binding sites of c1 repressor of bacteriophage P1. Evidence for multiple asymmetric sites. J. Mol. Biol., , 198, , 281–293.
Laufer,C.S., Hays,J.B., Windle,B.E., Schaefer,T.S., Lee,E.H., Hays,S.L. and McClure,M.R. ( (1989) ) Enhancement of Escherichia coli plasmid and chromosomal recombination by the Ref function of bacteriophage P1. Genetics, , 123, , 465–476.
Lu,S.D., Lu,D. and Gottesman,M. ( (1989) ) Stimulation of IS1 excision by bacteriophage P1 ref function. J. Bacteriol., , 171, , 3427–3432.
Smith,H.W. ( (1972) ) Ampicillin resistance in Escherichia coli by phage infection. Nature New Biol., , 238, , 205–206.
Ikeda,H., Inuzuka,M. and Tomizawa,J.I. ( (1970) ) P1-like plasmid in Escherichia coli 15. J. Mol. Biol., , 50, , 457–470.
Yoshida,Y. and Mise,K. ( (1984) ) Characterization of generalized transducing phage w39 heteroimmune to phage P1 in Escherichia coli W39. Microbiol. Immunol., , 28, , 415–426.
Mise,K. and Suzuki,K. ( (1970) ) New generalized transducing bacteriophage in Escherichia coli. J. Virol., , 6, , 253–255.
Meyer,J., Stalhammar-Carlemalm,M., Streiff,M., Iida,S. and Arber,W. ( (1986) ) Sequence relations among the IncY plasmid p15B, P1, and P7 prophages. Plasmid, , 16, , 81–89.
Watkins,C.A. and Scott,J.R. ( (1981) ) Characterization of bacteriophage D6. Virology, , 110, , 302–317.
Sambrook,J., Fritsch,E.F. and Maniatis,T. ( (1989) ) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.
Lennox,E.S. ( (1955) ) Transduction of linked genetic characters of the host by bacteriophage P1. Virology, , 1, , 190–206.
Wall,J.D. and Harriman,P.D. ( (1974) ) Phage P1 mutants with altered transducing abilities for Escherichia coli. Virology, , 59, , 532–544.
Sauer,B., Whealy,M., Robbins,A. and Enquist,L. ( (1987) ) Site-specific insertion of DNA into a pseudorabies virus vector. Proc. Natl Acad. Sci. USA, , 84, , 9108–9112.
Sauer,B. ( (1993) ) Manipulation of the transgene by site-specific recombination: use of the cre recombinase. Meth. Enzymol., , 225, , 890–900.
Guzman,L.-M., Belin,D., Carson,M.J. and Beckwith,J. ( (1995) ) Tight regulation, modulation, and high-level expression by vectors containing the arabinose pBAD promoter. J. Bacteriol., , 177, , 4121–4130.
Gagneten,S., Le,Y., Miller,J. and Sauer,B. ( (1997) ) Brief expression of a GFPcre fusion gene in embryonic stem cells allows rapid retrieval of site-specific genomic deletions. Nucleic Acids Res., , 25, , 3326–3331.
Le,Y., Miller,J.L. and Sauer,B. ( (1999) ) GFPcre fusion vectors with enhanced expression. Anal. Biochem., , 270, , 334–336.
Eustice,D.C., Feldman,P.A., Colberg-Poley,A.M., Buckery,R.M. and Neubauer,R.H. ( (1991) ) A sensitive method for the detection of ?-galactosidase in transfected mammalian cells. BioTechniques, , 11, , 739–743.
Bethke,B. and Sauer,B. ( (1997) ) Segmental genomic replacement by Cre-mediated recombination: genotoxic stress activation of the p53 promoter in single-copy transformants. Nucleic Acids Res., , 25, , 2828–2834.
Altschul,S.F., Madden,T.L., Schaffer,A.A., Zhang,J., Zhang,Z., Miller,W. and Lipman,D.J. ( (1997) ) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res., , 25, , 3389–3402.
Brukner,I., Sanchez,R., Suck,D. and Pongor,S. ( (1995) ) Sequence-dependent bending propensity of DNA as revealed by DNase I: parameters for trinucleotides. EMBO J., , 14, , 1812–1818.
Gabrielian,A. and Pongor,S. ( (1996) ) Correlation of intrinsic DNA curvature with DNA property periodicity. FEBS Lett., , 393, , 65–68.
Hoess,R.H., Ziese,M. and Sternberg,N. ( (1982) ) P1 site-specific recombination: nucleotide sequence of the recombining sites. Proc. Natl Acad. Sci. USA, , 79, , 3398–3402.
Sternberg,N. and Coulby,J. ( (1987) ) Recognition and cleavage of the bacteriophage P1 packaging site (pac). II. Functional limits of pac and location of pac cleavage termini. J. Mol. Biol., , 194, , 469–479.
Skorupski,K., Sauer,B. and Sternberg,N. ( (1994) ) Faithful cleavage of the P1 packaging site (pac) requires two phage proteins, PacA and PacB, and two Escherichia coli proteins, IHF and HU. J. Mol. Biol., , 243, , 268–282.
Skorupski,K., Sternberg,N. and Sauer,B. ( (1994) ) Purification and DNA-binding activity of the PacA subunit of the bacteriophage P1 pacase enzyme. J. Mol. Biol., , 243, , 258–267.
Osborne,F.A., Stovall,S.R. and Baumstark,B.R. ( (1989) ) The c1 genes of P1 and P7. Nucleic Acids Res., , 17, , 7671–7680.
Sternberg,N., Sauer,B., Hoess,R. and Abremski,K. ( (1986) ) Bacteriophage P1 cre gene and its regulatory region: evidence for multiple promoters and for regulation by DNA methylation. J. Mol. Biol., , 187, , 197–212.
Sauer,B. ( (1987) ) Functional expression of the cre-lox site-specific recombination system in the yeast Saccharomyces cerevisiae. Mol. Cell. Biol., , 7, , 2087–2096.
Sauer,B. and Henderson,N. ( (1988) ) Site-specific DNA recombination in mammalian cells by the Cre recombinase of bacteriophage P1. Proc. Natl Acad. Sci. USA, , 85, , 5166–5170.
Lee,L. and Sadowski,P.D. ( (2003) ) Sequence of the loxP site determines the order of strand exchange by the Cre recombinase. J. Mol. Biol., , 326, , 397–412.
Hartung,M. and Kisters-Woike,B. ( (1998) ) Cre mutants with altered DNA binding properties. J. Biol. Chem., , 273, , 22884–22891.
Kim,S., Kim,G., Lee,Y. and Park,J. ( (2001) ) Characterization of Cre-loxP interaction in the major groove: hint for structural distortion of mutant Cre and possible strategy for HIV-1 therapy. J. Cell. Biochem., , 80, , 321–327.
Guo,F., Gopaul,D.N. and Van Duyne,G.D. ( (1997) ) Structure of Cre recombinase complexed with DNA in a site-specific recombination synapse. Nature, , 389, , 40–46.
Sauer,B. ( (1994) ) Site-specific recombination: developments and applications. Curr. Opin. Biotechnol., , 5, , 521–527.
Scott,J.R. ( (1968) ) Genetic studies on bacteriophage P1. Virology, , 36, , 564–574.
Wandersman,C. and Yarmolinsky,M. ( (1977) ) Bipartite control of immunity conferred by the related heteroimmune plasmid prophages, P1 and P7 (formerly phi Amp). Virology, , 77, , 386–400.(Brian Sauer1,2,* and Jeffrey McDermott1)
* To whom correspondence should be addressed. Tel: +1 816 926 4432; Fax: +1 816 926 2068; Email: BLS@stowers-institute.org
+AY751747–AY751749 and AY753669
DDBJ/EMBL/GenBank accession nos+
ABSTRACT
Sequencing of the 7 kb immC region from four P1-related phages identified a novel DNA recombinase that exhibits many Cre-like characteristics, including recombination in mammalian cells, but which has a distinctly different DNA specificity. DNA sequence comparison to the P1 immC region showed that all phages had related DNA terminase, C1 repressor and DNA recombinase genes. Although these genes from phages P7, w39 and p15B were highly similar to those from P1, those of phage D6 showed significant divergence. Moreover, the D6 sequence showed evidence of DNA deletion and substitution in this region relative to the other phages. Characterization of the D6 site-specific DNA recombinase (Dre) showed that it was a tyrosine recombinase closely related to the P1 Cre recombinase, but that it had a distinct DNA specificity for a 32 bp DNA site (rox). Cre and Dre are heterospecific: Cre did not catalyze recombination at rox sites and Dre did not catalyze recombination at lox sites. Like Cre, Dre catalyzed both integrative and excisive recombination and required no other phage-encoded proteins for recombination. Dre-mediated recombination in mammalian cells showed that, like Cre, no host bacterial proteins are required for efficient Dre-mediated site-specific DNA recombination.
INTRODUCTION
The Cre recombinase of bacteriophage P1 is a member of the lambda integrase or tyrosine recombinase family of site-specific DNA recombinases. Because of its relatively simple biochemical requirements, its highly specific recognition of a 34 bp site not normally present in most genomes, and its ability to catalyze DNA recombination both in prokaryotes and in eukaryotes Cre has proven useful both in elucidating the mechanism of site-specific DNA recombination (1) and in developing powerful strategies for genetic engineering (2).
Large scale mutagenesis (3), in vitro evolution approaches (4–6) and the availability of several crystal structures for Cre (1), as well as comparative sequence analysis to other recombinases (7–9), have all contributed to an understanding of Cre's DNA specificity and mechanism of action. Still, a deeper appreciation of specificity and function could be had if close homologs of Cre recombinase were available for comparison. The elegant genetic dissection of specificity determination using the lambda and HK022 Int proteins attests to the power of such an approach (10,11). In this work, we sequenced the pac-c1 genomic regions of several P1-related phages to identify Cre homologs.
The P1 genome, the sequencing of which has recently been completed in its entirety (12), is relatively large for a temperate DNA bacteriophage: 95 kb. P1 is unusual among temperate bacteriophages in that it maintains itself as an extrachromosomal unit copy plasmid in the lysogenic state. Cre is expressed in P1 lysogens and its site-specific DNA recombination activity contributes to the stable maintenance of the P1 prophage during lysogeny. Cre resolves P1 dimers that arise by homologous recombination after DNA replication, thus helping to ensure segregation of a P1 monomer to each daughter cell at cell division (13). Because P1-related phages also maintain themselves as an extrachromosomal plasmid in the lysogenic state, we expected that they too would have a comparable recombinase function.
The P1 cre gene and its 34 bp recombination target site loxP lie in a relatively short interval of P1 DNA that includes two other phage functions with unusual features. To the left of cre is the immC immunity region of P1 that encodes the C1 repressor and several other immunity proteins that modify its action. ImmC, in turn, lies just to the right of the two genes for the P1 pacase or terminase and the pac site at which P1 DNA packaging begins. DNA packaging in P1 is unusual because protein recognition of the P1 DNA packaging site is regulated by DNA adenine methylation (dam). Understanding of P1 DNA packaging has led to the development of both high frequency P1 transducing phage strains (14) and an in vitro P1 packaging system for the construction of recombinant DNA libraries with large DNA inserts (15). Although immunity in P1 is orchestrated in a complex manner and includes several different immunity regions (16), including anti-repressor components, the C1 protein is itself unusual compared with other phage repressors in that it recognizes an asymmetric DNA-binding site (17). To the left of the P1 cre gene is c8, another immunity gene, followed by ref, a gene involved in the homologous recombination of short DNA repeats (18,19).
In this work, we sequenced and compared the 7 kb pac-c1 region from five P1-related phages, namely, the closely related phage P7 (20), the defective phage p15B (21), transducing phage w39 (22) and the P1-like transducing phage D6 isolated from Salmonella oranienburg (23). All exist as an extrachromosomal plasmid in Escherichia coli and, with the exception of D6, all are members of plasmid incompatibility group Y (22,24). P1, P7 and p15B genomic DNAs cross-hybridize (24), as do those of D6 and P1 (25). Interestingly, the immC regions of D6 and P1 do not cross-hybridize (25), suggesting that counterparts of P1 genes in this region of D6 may be diverged or absent. Sequence comparisons presented here showed that this region is similarly organized in all phages and encodes DNA packaging, repressor and recombinase genes. Sequence identity was very high for this 7 kb region amongst all the phages except for D6. Characterization of the D6 recombinase showed that although it is closely related to the P1 Cre recombinase, it exhibits a distinctly different DNA specificity that can be used either to remove a marker gene or to activate gene expression in vivo.
MATERIALS AND METHODS
Bacteria and phage
Bacterial strains and lysogens used in this work are listed in Table 1. Bacteria were propagated in Luria–Bertani broth (26) with appropriate antibiotics: streptomycin (10 μg/ml), ampicillin (100 μg/ml) and chloramphenicol (34 μg/ml). Zeocin (Invitrogen) was used at a concentration of 10 μg/ml in Lennox broth (27). Phages were titered, maintained and propagated in the presence of 5 mM CaCl2. Lysogenization of DH5lacU169 by selection for CmR after infection with P1 CM (a gift from N. Sternberg) generated BS610.
Table 1. Bacterial strains
In general, phage stocks were prepared and titered using the indicator strain Sh-16. Strains lysogenic for P1 CM, P7 c1.9, w39 and D6 (Table 1) all spontaneously released a low number of phage after culture overnight that gave small plaques on Sh-16, indicating that these P1-like phages were capable of both lysogenic and lytic growth. Phage plate stocks (9–10 h at 37°C) were prepared with a fresh overnight culture of the indicated donor strain (28), and used immediately. For transduction, 0.1 ml of a fresh overnight of the recipient strain DH5lacU169 was infected with an aliquot of the transducing stock. After pre-adsorption for 5 min at room temperature, the cells were diluted to 1 ml with LB broth + 5 mM CaCl2, incubated 50 min at 37°C and then plated on selective medium.
DNA sequencing and plasmid construction
Circular plasmid DNA of bacteriophages P1 CM, P7 c1.9, w39, p15B and D6 was prepared using the Qiagen Large Construct Kit (Valencia, CA). For phages P1 CM, P7 c1.9, w39 and p15B, we designed PCR primers based on the published sequences of the P1 genes between lpa (formerly gene 10) and ref. This region encompasses the DNA packaging genes pacA and pacB, the c1 gene along with several other immunity genes, and the gene for Cre recombinase. For all these phages, at least several P1 primer pairs were identified for amplification of this interval. PCR fragments generated from each phage were sequenced directly and the resulting sequence information was used to design additional primers. A combination of PCR fragment sequencing and primer walking by direct sequencing from phage DNA was then used to obtain the complete sequence for each phage of the 7 kb region from pacA to the beginning of the ref gene.
For phage D6, none of the PCR primer pairs used for P1 or the other P1-related phages produced an amplified product. We therefore constructed a D6 library by partial Sau3A digestion of D6 DNA, cloning into the BamHI site of pUC19 and transformation of DH5. To eliminate smaller clones from this library, it was digested with EcoRI, size-selected by agarose gel electrophoresis to contain inserts of 1.5 kb and religated. Shotgun sequencing identified one clone having strong similarity to the P1 pacB gene. Specific primers were designed and were used to sequence directly from D6 DNA by primer walking in both directions to obtain 7 kb of flanking sequence.
DNA sequences were assembled and analyzed using Vector NTI (Invitrogen) and then compared with the corresponding region of the P1 genome, GenBank accession no. AF234172 . The sequences of the immC regions of phages P7 c1.9 (6751 bp), w39 (7208 bp), p15B (7094 bp) and D6 (7644 bp) have been assigned GenBank accession numbers AY751747 , AY751748 , AY751749 and AY753669 , respectively.
Oligonucleotide sequencing primers and linkers were synthesized by Integrated DNA Technologies (Coralville, IA) and restriction enzymes were from New England Biolabs (Beverly, MA). Annealing of the oligo's 5'-CTA GAT AAC TTT AAA TAA TTG GCA TTA TTT AAA GTT AG-3' and 5'-GAT CCT AAC TTT AAA TAA TGC CAA TTA TTT AAA GTT AT-3' and cloning into the XbaI and BamHI sites of pUC19 generated the rox plasmid pBS1051 (D6 sequence underlined). Similarly, oligo's 5'-CTA GCT ATA ACT TCG TAT AAT GTA TGC TAT ACG AAzG TTG-3' and 5'-TCG ACA ACT TCG TAT AGC ATA CAT TAT ACG AAG TTA TAG-3' were cloned into a pBluescript II KS (Stratagene) derivative using NheI and SalI to give the lox plasmid pBS516. The one nucleotide difference (double underline) of this lox site from loxP does not affect Cre-mediated recombination (29). Digestion of pBS1051 with either XbaI + AlwNI or BamHI + AlwNI and ligation with the XbaI–BamHI zeo fragment from pZeoSV (Invitrogen) generated the rox2-zeo plasmid pBS1080. The analogous loxP2 zeo plasmid pBS890 was constructed by V. Petyuk of this laboratory by blunt-ending the 600 bp FokI–SalI zeo fragment from pZeoSV and cloning it into the SmaI site of the loxP2 vector pBS246 (30). The dre gene was amplified with Pfu DNA polymerase (Stratagene) from D6 DNA with the oligo's 5'-AGA TGG TAC CAG GAG GAT ATC AAT ATG AGT GAA TTA ATT ATC TCT GG-3' and 5'-CTT TAG TCT AGA TTC ATT ATG AAT CCA TCA AGC GGC-3' (D6 coding region underlined), digested with KpnI and XbaI and cloned into the CmR arabinose-inducible vector pBAD33 (31) to generate pBS1081. The analogous pBAD33-cre construct has been described previously (6). All cloned oligos and PCR products were confirmed by DNA sequencing.
The dre gene was placed under the control of the EF1 promoter for mammalian expression by replacing the KpnI–XbaI GFP fragment of pBS377 (32) with the KpnI–XbaI dre fragment from pBS1081. The EF1-cre expression plasmid pBS513 (33) and the control CMV-lacZ plasmid p324 (34) have been described previously. To construct the enhanced green fluorescent protein (EGFP) expression vector pBS504, the following three DNA fragments were cloned between the unique HindIII and EcoRI sites of pBS397 (35): the EcoRI–KpnI EF1 fragment of pBS377, the KpnI–XbaI EGFP fragment of pEGFP-1 (Clontech, Palo Alto, CA) and the XbaI–EcoRI fragment from pBS377 carrying the polyA signal. The EcoRI–HindIII rox2 zeo cassette from pBS1080 was blunt-end ligated into the EcoRV site lying between the EF1 promoter and the EGFP gene in pBS504 to generate the rox recombination reporter plasmid pBS1083.
DNA analysis
We performed PSI-BLAST searches (36) to determine closest homologs of identified genes. The bendability/curvature-propensity plots were calculated with the bend.it server, using DNase I-based bendability parameters (37) and the consensus bendability scale (38).
Site-specific DNA excision
Plasmids pBAD33, pBAD33-dre or pBAD33-cre were electroporated into DH5 cells containing either the reporter plasmid pBS890 or pBS1080, incubated for 1 h at 37°C in SOB media (6) and then plated, selecting for CmR ApR to ensure retention of both plasmids in the resulting transformants. To ensure that recombinase was not expressed at inordinately high levels, the pBAD plasmids were used without overt arabinose induction (9). The next day, colonies were individually tested for drug resistance markers by growth on appropriate antibiotic containing plates.
CHO-K1 cells were transfected using Polyfect (Qiagen) with 1.5 μg DNA per well of a 6 well dish as recommended by the manufacturer. Co-transfections used a 9:1 ratio of Cre, Dre or lacZ expression vector DNA to either pBS504 or pBS1083 as indicated. Fluorescence was monitored 2 days after transfection with a Leica DMR microscope mounted with an Optronics Magnafire digital camera, and DNA was then prepared for PCR analysis. Recombination was detected by PCR (31 cycles for 30 s at 94°C, 30 s at 60°C, 60 s at 72°C) using the sense EF1 primer KC315 5' GCTTGGCACTTGATGTAATTCTCCTTG 3' and the antisense EGFP primer BSBS382 5' GGTCAGCTTGCCGTAGGTGGC 3'. Predicted product sizes are 302 bp for pBS504, 1704 bp for pBS1083 (not observed because of the short cycling times used) and 389 bp for the Dre excision product of pBS1083.
RESULTS
Even though the immC region of phage P1 does not cross-hybridize to phage D6 DNA, it seemed likely that D6 would carry a site-specific DNA recombination system like that of P1, and that this recombinase/recombinase recognition site would lie in the D6 immC region. We therefore checked for the presence of a recombinase activity genetically. If there were a Cre-like D6 recombinase, then infection of the D6 clone plasmid library with phage D6 would ‘pick-up’ relevant plasmid clones as a result of low level site-specific integrative recombination between phage DNA and plasmids carrying a D6 recombination site. Thus, D6 would be able to transduce relevant ApR clones to a new bacterial recipient where they would be excised from the phage genome by the D6 site-specific recombinase to take up plasmid residency. A similar strategy was used previously to pick-up loxP plasmid clones using phage lambda carrying the loxP-cre region of P1 (39). Table 2 shows that D6 could transduce the ApR marker from a library of D6 clones, but not from a strain having an anonymous randomly chosen D6 cloned insert or from a non-plasmid strain, Sh-16. These results indicate that, like P1, D6 carries a site-specific DNA recombinase that catalyzes both integrative and excisive recombination.
Table 2. D6 ‘pick-up’ identification of a D6 recombination site
To compare the recombinase genes, as well as the adjacent c1 and pac genes, we sequenced the 7 kb immC region from four P1-related phages: P7 c1.9, w39, p15B and D6. For the first three of these phages, the gene organization of this region was identical to that of phage P1, whereas that of phage D6 had a similar structure (Figure 1). To confirm that this D6 region did indeed harbor the recombinase site that we had detected previously, we sequenced six randomly chosen plasmids transduced by D6, naming them Pick-Up clones PU-5, etc. All carried overlapping sequences from the same immC DNA region we had sequenced (Figure 1), strongly indicating that each carried a D6 DNA site for site-specific recombination.
Figure 1. Comparison of the pac-c1 regions of P1 and D6. Shown is a 7623 bp region of P1 that includes the region from lpa (gene 10) to ref. The equivalent lpa-c8 regions of P7, w39 and p15B showed the same genes and gene order as shown for P1 except that in P7 the imcA and imcB genes were fused. As discussed in the text, gene products of this region from these other phages were 89–100% similar to their P1 counterparts. Also shown is the corresponding 7644 bp region from D6. The degree of similarity between individual genes of P1 and D6 is indicated according to the color scale shown. P1 genes in white have no D6 equivalent in this region, and D6 genes in black have no P1 equivalent. Below and aligned with the D6 map is diagrammed the DNA insert of several PU clones obtained by site-specific recombination.
The overall gene organization of the sequenced region from D6 is similar to the pac-ref region of P1 despite low sequence identity. On the left (Figure 1), there is a gene showing 24% similarity to P1's lpa (late promoter activating protein) or gene 10. This is followed by DNA packaging genes (pacA and pacB), a c1 repressor gene and a D6 recombinase ‘dre’ similar to the P1 cre gene. We discuss these genes in more detail below. There are also several non-conserved open reading frames (ORFs) in this region of the D6 genome, and an insertion of a gene similar to the deoxyuridine 5'-triphosphate nucleotidohydrolase of the photosynthetic bacterium Rubrivivax gelatinosus (dut; Figure 1).
DNA packaging genes
P1 packages its DNA using a terminase composed of the PacA small subunit and the PacB large subunit. Table 3 compares the PacA's and PacB's for each of the 5 P1 family phages. Although PacA was very similar (97–100%) for phages P1, P7, w39 and p15B, the PacA protein of D6 was 57 amino acids longer and showed only 18% similarity to PacA from any of the other phages (see also Figure 2). D6 PacB was somewhat more similar to the PacB proteins of the other P1 family members (53–54%) and was slightly larger. Interestingly, the other four phages form two clear subgroups: P1 and P7 are nearly identical to each other, likewise w39 and p15B are nearly identical, with the two subgroups 89% similar to each other. In addition, the defective phage p15B carried a nonsense mutation that truncates the C-terminus of PacB.
Table 3. DNA packaging genes
Figure 2. Comparison of the P1 and D6 pacA genes. Dam methylation sites for both genes are shown as vertical lines above the gene. A blowup of the 162 bp P1 pac site that includes two clusters of dam sites is shown above the P1 pacA gene, with dam sites represented as black boxes, the IHF-binding site as a white rectangle and the region of cleavage as a vertical arrow. The 397 amino acid P1 PacA protein and the 454 amino acid D6 PacA protein are 18% similar: insertions into the 1194 bp P1 pacA gene relative to the 1365 bp D6 pacA gene are shown in yellow, insertions into D6 pacA relative to P1 pacA are shown in orange. An asterisk at nucleotide position 182 of D6 pacA marks the maximum of a curvature-propensity plot calculated with DNase I-based trinucleotide parameters (37).
P1 headful DNA packaging proceeds from a specific 162 bp pac site located within the 5' end of the pacA structural gene (Figure 2). The site consists of two clusters of hexanucleotide repeats flanking a central region where DNA cleavage occurs (40). Each of the hexamer repeats has a core 5'-GATC dam methylation site. PacA protein binds to the hexamer repeats in a DNA methylation-dependent manner and then associates with PacB to loop the two binding domains in an IHF and HU-dependent manner prior to DNA cleavage (41,42). In phages P7, w39 and p15B, this 162 bp region differed from that of P1 by 4, 9 and 2 bp, respectively, but the integrity of the seven hexamer repeats was completely preserved, suggesting that all four of these phages package DNA in the same manner. Similarly, the IHF-binding site adjacent to the region of DNA cleavage was identical in w39 and P1, although in P7 and p15B this site displayed a one nucleotide change from 5'-AAACAAAGAGTTA to 5'-AAACAGAGAGTTA (change underlined).
The low degree of similarity between D6 and P1 PacA proteins suggests that their DNA-binding specificities may differ. In addition, there was no region of clustered dam sites characteristic of the P1 pac site and no consensus IHF-binding site in the D6 pacA gene (Figure 2), further suggesting a difference in the DNA recognition specificities of the P1 and D6 terminases. Interestingly, though, there was a potentially curved DNA sequence in the 5' region off the D6 pacA gene (asterisk, Figure 2). The curvature-propensity plot, calculated with DNase I-based trinucleotide parameters, contained one peculiar maximum in this region, whose magnitude (14.7°/helical turn) exceeded the value calculated for Columba risoria bent satellite DNA (13.5°/helical turn). No such potentially curved DNA was detected in the pacA genes of P1, P7, w39 or p15B.
C1 repressor and immunity
There was little difference in the immunity genes of the immC region for the four phages P1, P7, w39 and p15B (Figure 1). The C1 repressors of P1 and P7 had previously been shown to be identical (43), so the three amino acid changes (A110V, P190L and D277S) in P7 c1.9 are likely specific to this P7 temperature-sensitive repressor. Aside from a K268R change in w39, the p15B and w39 C1 repressors were identical to that of P1. The Coi (c one inactivator) protein sequence was identical for all four of these phages except for an A62T substitution in p15B. The predicted imcA and imcB gene products were also identical except for an A27T difference in P1. Interestingly, these two genes were fused in P7. Some variation was seen for C8 among phages P1, P7, w39 and p15B, but all showed 85% similarity. On the other hand, the D6 C1 protein was only 16% similar to P1 C1. Moreover, in phage D6, the distance between c1 and the recombinase gene dre was much shorter than the corresponding region in P1, and the coi and imcA/imcB genes were missing.
Recombinase
Among the four phages P1, P7, w39 and p15B, the 343 amino acid Cre recombinase was highly conserved. The w39 Cre differed from P1 Cre by the single amino acid change T206A, P7 Cre differed from P1 Cre by the two changes A178S and G280D, and the p15B Cre differed from P1 Cre by three changes: P107L, A249S and A252P. Similarly conserved among these four phages was cra (putative cre associated function), an ORF of unknown function adjacent to cre originally designated as orf1 (44). The putative P1 Cra protein differed from that of P7, w39 and p15B by 3, 1 and 2 amino acids, respectively. In accord with the nearly identical Cre recombinases of these four phages, all of them carried an identical 34 bp lox site located midway between c1 and cre.
In contrast, D6 displayed a recombinase gene only 39% similar to the P1 cre gene and no 34 bp lox site anywhere in this 7.6 kb region of DNA (Figure 1). This suggested that the dre might recognize a recombination site distinct from lox. As noted above, the interval between c1 and dre was much shorter in D6 than the corresponding region in P1, and no ORF corresponding to cra was found. From the D6 pick-up experiments, we knew that a recombination site must lie within the D6 sequence present in the PU7 clone (Figure 1), a 2.1 kb fragment that includes both dre and the interval between c1 and dre. We therefore inspected the 394 bp c1-dre interval for a candidate recombination site for Dre.
The P1 lox site consists of two 13 bp inverted repeats flanking an asymmetrical 8 bp spacer region that imparts an overall directionality to the recombination site. In the c1-dre interval from D6, we detected two DNA sequences resembling this structure. Figure 3 shows a 69 bp portion of the c1-dre interval within which is a 32 bp sequence having two perfect 14 bp inverted repeats (solid arrows) separated by 4 bp. Just abutting this potential Dre recombination site was a similarly configured 18 bp DNA sequence with less perfect inverted repeats (dashed line arrows).
Figure 3. The D6 c1-dre integenic region. The 394 bp region from D6 between c1 and dre is diagrammed along with the single dam site of this region. The sequence shown is a 70 bp portion of this region suspected to include the D6 rox site because of the presence of inverted repeat elements. Repeat elements of a DNA sequence shown not to be the recombination site are indicated by dashed arrows, and the repeat elements of the actual rox site are indicated by solid arrows. The two DraI sites are shown in lower case.
To determine the involvement of either of these DNA sequences in Dre-mediated recombination, we constructed several test plasmids and evaluated their ability to recombine with D6 using the D6 plasmid transduction assay. Table 4 shows that D6 transduces PU7, but not pUC19, from a recA host to a recA recipient at high frequency. Taking advantage of a fortuitous DraI site present in each 14 bp inverted repeat (Figure 3, upper case), we mutated the 32 bp candidate site by deleting the 12 bp between the DraI sites to generate a PU7 derivative, PU7-Dra1. The lack of D6 transduction of this plasmid (Table 4) indicated that the integrity of the 32 bp region was necessary for recombination and suggested that the adjacent shorter imperfect inverted repeat region was not a recombination site. The 32 bp region might thus be the region of crossover (X-over) recombination or rox site for Dre recombinase. To test this, we placed this sequence on pUC19 and tested its ability to undergo recombination with D6. Transduction of pUC19 carrying the putative rox site occurred at high frequency (Table 4), confirming that the rox sequence was sufficient for recombination. Moreover, no transduction of a pUC19-lox plasmid was obtained, indicating that D6 cannot recombine with the P1 lox site.
Table 4. Plasmid transduction by phage D6
We wanted to establish whether or not the Dre recombinase was the only D6-encoded protein required for recombination at rox. As a first step in this direction, we placed the dre structural gene under the control of the arabinose-inducible promoter in plasmid pBAD33. SDS–PAGE and western blot analysis showed that this construct expressed a Cre-sized protein of 36 kDa that cross-reacted with a polyclonal antibody to Cre (Figure 4).
Figure 4. Western blot detection of Dre recombinase. Whole cell extracts were prepared from mid-log cultures of E.coli DH5 containing pBAD-dre, pBAD-cre or a pBAD vector with no insert after induction for 2 h with 0.2% L-arabinose. Following SDS–PAGE western blot analysis was with a polyclonal rabbit antibody prepared against purified Cre recombinase (45). Size markers in kDa are shown to the right.
We assayed Dre-mediate recombination using a rox2 zeo construct (Figure 5) that carries two directly repeated rox sites flanking zeo, a gene that confers resistance to the antibiotic zeocin. Cells carrying this construct become sensitive to zeocin upon loss of zeo by excisive recombination at the rox sites. Table 5 shows that transformation of the rox2 zeo strain with the compatible CmR plasmid pBAD33-dre resulted in loss of zeocin resistance in all transformants. No loss of zeo was seen with a control plasmid having no insert or with a construct expressing Cre recombinase. Conversely, a strain carrying a lox2 zeo construct was refractory to excisive recombination by the Dre-expressing construct but readily underwent excisive recombination with loss of zeo when tranformed with the Cre-expressing construct. Sequencing of the rox plasmid from several zeocin-sensitive colonies confirmed that precise excisive recombination had occurred at the rox site. Thus, Dre mimicked Cre's ability to perform excisive recombination, but the two recombinases were heterospecific, i.e. Cre did not catalyze recombination at rox and Dre did not catalyze recombination at lox.
Figure 5. Recombinase-mediated excision. Reporter plasmids were constructed by placing the zeo gene (diagonal bars) between two identical directly repeated recombination sites (black triangles). Shown are the sequences of the 32 bp rox site and the 34 bp loxP site. Horizontal arrows indicate the inverted repeat elements; positions of nucleotide identity between rox and loxP are indicated by the black boxes; and vertical arrows show the sites of Cre cleavage that define the 6 bp overlap region of the lox site.
Table 5. Dre-mediated excision in E.coli
Unlike many members of the tyrosine recombinase family, Cre recombinase of phage P1 requires no accessory phage or bacterial proteins for recombination. This characteristic of Cre is illustrated by Cre's ability to catalyze DNA recombination in a variety of eukaryotic cells (45,46). To determine whether Dre also had no requirement for accessory bacterial proteins, and thus would be able to catalyze DNA recombination in a mammalian cell, we tested Dre's ability to recombine rox sites in CHO cells. Dre recombination was assayed by using a reporter plasmid that would express EGFP in transfected cells only if activated by recombination between two directly repeated rox sites (Figure 6A). In the reporter plasmid, a rox2 zeo cassette is inserted between the EF1 promoter and the EGFP gene, positioning both an upstream zeo gene and a polyadenylation site act to block EGFP expression. Dre-mediated recombination at the flanking rox sites would remove these blocks. To express Dre, we constructed a second plasmid vector in which the dre gene was placed under the control of the EF1 promoter. Figure 6B shows that co-transfection of CHO cells with the rox reporter construct and the Dre expression vector produced a significant number of green fluorescent cells (panel i), whereas there was no detectable fluorescence when the reporter was co-transfected with a Cre expression plasmid (panel ii). The frequency of fluorescent cells from Dre-mediated activation of the EGFP gene was similar to the transfection efficiency of an EGFP reporter plasmid having no rox2 zeo cassette (panel iii). PCR confirmed that recombination at the rox sites in cells co-transfected with the Dre expression plasmid and that no recombination occurred at the rox sites in cells co-transfected with the Cre expression plasmid (Figure 6C). Thus, Dre-mediated recombination requires no bacterial proteins for efficient DNA recombination at rox.
Figure 6. Dre-mediated recombination in mammalian cells. (A) Recombination activation of gene expression by Dre. Expression of the EGFP gene from the EF1 promoter in the reporter plasmid pBS1083 is blocked by the interposed zeo gene and polyadenylation site (An). Dre-mediated recombination at the flanking rox sites (black triangles) removes this ‘STOP’ signal to allow EGFP expression. Horizontal arrows indicate the PCR primers used to produce the 389 bp fragment diagnostic of recombination. (B) Activation of EGFP expression by Dre. Epifluorescence (panels i–iii) and differential interference contrast (panels iv–vi) images of CHO cells 36 h after DNA transfection. Panels i and iv, Dre expression plasmid pBS1081 + the rox2 STOP EGFP plasmid pBS1083; panels ii and v, Cre expression plasmid pBS513 + pBS1083; and panels iii and vi, control lacZ plasmid p324 + the EGFP plasmid pBS504. (C) PCR detection of recombination. DNA from the transfected CHO cells shown in (B) was amplified using the primers shown in (A). The 389 bp fragment is diagnostic of Dre recombination at the rox sites of pBS1083. The same primers give a 302 bp on the parental plasmid having no rox cassette inserted between the EF1 promoter and the EGFP gene. Lane 1, CHO cells transfected with the Dre expression plasmid pBS1081 + the rox2 STOP EGFP plasmid pBS1083; lane 2, CHO cells transfected with the Cre expression plasmid pBS513 + pBS1083; lane 3, CHO cells transfected with the control lacZ plasmid p324 + the EGFP plasmid pBS504; lane 4, EGFP plasmid pBS504; and lane 5, rox2 STOP EGFP plasmid pBS1083.
DISCUSSION
Here, we compare the sequence of the 7 kb region encompassing the terminase, repressor and recombinase genes from five related bacteriophages: P1, P7, w39, p15B and D6. With the exception of phage D6, the respective genes of this region were nearly identical. The few differences that did exist indicated that P1 and P7 form one subgroup and that w39 and p15B form a second closely related subgroup. Although this region of D6 displayed the same overall gene organization, the amino acid sequences of the D6 proteins showed much less similarity to their P1 counterparts than did these proteins from the other P1-like phages. In addition, there were several insertions and deletions of ORFs.
The conservation of the PacA protein and its cognate DNA packaging site pac (to which PacA binds) in four P1-like phages indicated that their DNA packaging proceeds in a near identical manner. In contrast, the DNA-binding PacA protein of D6 was only distantly related to the P1 PacA protein and the hallmark features of the P1 pac site sequence, two clusters of potentially dam-methylated hexamer sites flanking an 90 bp segment carrying an IHF-binding site, were not found in the D6 pacA gene, although there was a potentially curved DNA sequence in this region found. This suggests that although D6 terminase is clearly related to the P1 enzyme, D6 DNA packaging will differ in detail from the process used by phage P1.
The near identity of the C1 repressors of phages P1, P7, w39 and p15B was paralleled by the high conservation of Coi, a protein that binds to C1 and negatively regulates its activity. In contrast, the C1 protein of D6 showed considerable divergence from that of P1 and no coi gene was found in this region. The low sequence similarity of the repressors suggests that they bind different DNA sequences and this may contribute to the observed heteroimmunity of phages P1 and D6 (25).
There were six amino acid differences amongst the four Cre recombinases of phages P1, P7, w39 and p15B, with no one protein differing from P1 Cre by more than three changes. Of these six amino acid differences, four lay in flexible regions of the Cre protein that tolerate pentapeptide insertion (Figure 7), the other two were conserved amino acid changes. The presence of an identical 34 bp loxP site in these four phages also indicated that these proteins are functionally equivalent. On the other hand, the D6 recombinase Dre was only 39% similar to Cre. Even so, Dre is more closely related to Cre than are other members of the tyrosine recombinase family (9). The two recombinases are clearly heterospecific: Dre catalyzed recombination at a distinct 32 bp rox recombination site that was not recognized by Cre. Conversely, Dre did not recognize lox, the 34 bp recombination site recognized by Cre.
Figure 7. Comparison of P1 Cre and D6 Dre. Sequence identity is shown in yellow, and conserved amino acid changes are shown in green. Below the sequences is shown the known secondary structure (-helices as dark blue cylinders, ?-strands as orange arrows) of Cre (50). Below the depiction of Cre secondary structure is a functional map of Cre based on pentapeptide insertion (9): blue bars represent points of insertion which have little or no effect on recombinase activity, and red bars indicate points at which insertion abolishes recombinase activity.
Both the lox and rox recombination sites (Figure 5) are structurally similar, exhibiting two perfect inverted repeat elements (13 bp for lox, 14 bp for rox) flanking a short asymmetric intervening sequence or spacer that is typically used by members of the tyrosine recombinase family to provide a homology sensor for productive pairing and recombination. Both recombination sites share sequence identity except for the spacer region and a 3 bp heterology in the middle of the inverted repeat elements. The distance between the repeat elements is 8 bp for lox but only 4 bp for rox. This is intriguing for several reasons. The distance between the top strand and bottom strand cleavages (the ‘overlap’ region) is generally 6–8 bp for the tyrosine recombinases (1). For Cre, the overlap region is 6 bp (Figure 5) and even a 1 nt shortening of the distance between the inverted repeat elements to which Cre binds abolishes recombination (unpublished data). Subsequently, for the Dre/rox recombination system, one possibility is that the rox overlap region would be an unusual one for the tyrosine recombinases: 4 bp. Alternatively, the overlap region for the Dre/rox system may be a more typical 6 or 8 bp. If so, however, the resulting residency of the cleavage sites within the inverted repeat elements would mean that the nucleotides adjacent to the cleavage sites are identical for both the top and bottom strand cleavages. This would be unlike the Cre system, for which it has been proposed that it is precisely these flanking nucleotide differences at the cleavage sites that determine which strand is first cleaved (47).
The large patches of high similarity between Cre and Dre (Figure 7) in general correlate well with functionally important regions identified from pentapeptide map insertional mutagenesis of Cre. Conversely, several regions permissive for pentapeptide insertion (the N-terminus and the I-J and J-K interhelical regions) are less well conserved between Cre and Dre. We note that one region of sequence dissimilarity between Cre and Dre that is not tolerant of pentapeptide insertion is the linker between the M and N helices, perhaps indicating that for this linker it is specific length rather than specific sequence that is critical.
Genetic and biochemical exploitation of the amino acid sequence similarities and differences between Cre and Dre should help to better understand sequence-specific DNA recognition of the recombination site. One point of interest readily apparent in contrasting the Cre and Dre systems is the 3 bp heterology near the center of the inverted repeats (Figure 5, sequence in right repeat: 5'-ACG in lox, 5'-TTA in rox). Mutation of this guanosine in the lox site blocks Cre binding and recombination (48), as does mutation of R259 of Cre (49), which contacts the guanosine of this sequence (50). Dre displays proline instead of an arginine at this amino acid position, consistent with an important role for this region in rox-lox discrimination. We note that the nearby E262 position of Cre is important in specific recognition of the conserved sequence 5'-AA just next to this G residue in lox (Figure 5) (6) and that this amino acid position is conserved in Dre.
In conclusion, comparison of the pac-c1 regions of P1 and four other P1-related phages showed that the gene organization of this region is conserved, and that the gene products of this region are nearly identical for all phages with the exception of D6. From sequence comparisons, the specificity of the D6 DNA packaging, immunity and site-specific DNA recombinase functions are likely to be different from those of P1. We have confirmed this heterospecificity for the DNA recombinase Dre of D6, and we have also shown that Dre can catalyze site-specific DNA recombination in mammalian cells. We anticipate that further work with Dre will not only provide insight into the mechanism of site-specific recombination but will also lead to the generation of new tools for genetic engineering (51).
ACKNOWLEDGEMENTS
We thank M. Yarmolinsky and F. Blattner for communication of results prior to publication and A. Mushegian for comments on the manuscript.
REFERENCES
Van Duyne,G.D. ( (2002) ) A structural view of tyrosine recombinase site-specific recombination. In Craig,N.L., Craigie,R., Gellert,M. and Lambowitz,A.M. (eds.), Mobile DNA II, ASM Press, Washington, DC, pp. 93–117.
Sauer,B. ( (2002) ) Chromosome manipulation by Cre-lox recombination. In Craig,N.L., Craigie,R., Gellert,M. and Lambowitz,A.M. (eds.), Mobile DNA II, ASM Press, Washington, DC, pp. 38–58.
Wierzbicki,A., Kendall,M., Abremski,K. and Hoess,R. ( (1987) ) A mutational analysis of the bacteriophage P1 recombinase Cre. J. Mol. Biol., , 195, , 785–794.
Buchholz,F. and Stewart,A.F. ( (2001) ) Alteration of Cre recombinase site specificity by substrate-linked protein evolution. Nat. Biotechnol., , 19, , 1047–1052.
Santoro,S.W. and Schultz,P.G. ( (2002) ) Directed evolution of the site specificity of Cre recombinase. Proc. Natl Acad. Sci. USA, , 99, , 4185–4190.
Rüfer,A.W. and Sauer,B. ( (2002) ) Non-contact positions impose site selectivity on Cre recombinase. Nucleic Acids Res., , 30, , 2764–2771.
Argos,P., Landy,A., Abremski,K., Egan,J.B., Ljungquist,E.H., Hoess,R.H., Kahn,M.L., Kalionis,B., Narayana,S.V.L., Pierson,L.S. et al. ( (1986) ) The integrase family of site-specific recombinases: regional similarities and global diversity. EMBO J., , 5, , 433–440.
Swalla,B.M., Gumport,R.I. and Gardner,J.F. ( (2003) ) Conservation of structure and function among tyrosine recombinases: homology-based modeling of the lambda integrase core-binding domain. Nucleic Acids Res., , 31, , 805–818.
Petyuk,V., McDermott,J., Cook,M. and Sauer,B. ( (2004) ) Functional mapping of Cre recombinase by pentapeptide insertional mutagenesis. J. Biol. Chem., , 279, , 37040–37048.
Dorgai,L., Yagil,E. and Weisberg,R.A. ( (1995) ) Identifying determinants of recombination specificity: construction and characterization of mutant bacteriophage integrases. J. Mol. Biol., , 252, , 178–188.
Yagil,E., Dorgai,L. and Weisberg,R.A. ( (1995) ) Identifying determinants of recombination specificity: construction and characterization of chimeric bacteriophage integrases. J. Mol. Biol., , 252, , 163–177.
Lobocka,M.B., Rose,D.J., Plunkett,G.,III, Rusin,M., Samojedny,A., Lehnherr,H., Yarmolinsky,M. and Blattner,F.R. ( (2004) ) The genome of bacteriophage P1. J. Bacteriol., , 186, , 7032–7068.
Austin,S., Ziese,M. and Sternberg,N. ( (1981) ) A novel role for site-specific recombination in maintenance of bacterial replicons. Cell, , 25, , 729–736.
Sternberg,N.L. and Maurer,R. ( (1991) ) Bacteriophage-mediated generalized transduction in Escherichia coli and Salmonella typhimurium. Meth. Enzymol., , 204, , 18–43.
Sternberg,N. ( (1990) ) Bacteriophage P1 cloning system for the isolation, amplification, and recovery of DNA fragments as large as 100 kilobase pairs. Proc. Natl Acad. Sci. USA, , 87, , 103–107.
Heinrich,J., Velleman,M. and Schuster,H. ( (1995) ) The tripartite immunity system of phages P1 and P7. FEMS Microbiol. Rev., , 17, , 121–126.
Eliason,J.L. and Sternberg,N. ( (1987) ) Characterization of the binding sites of c1 repressor of bacteriophage P1. Evidence for multiple asymmetric sites. J. Mol. Biol., , 198, , 281–293.
Laufer,C.S., Hays,J.B., Windle,B.E., Schaefer,T.S., Lee,E.H., Hays,S.L. and McClure,M.R. ( (1989) ) Enhancement of Escherichia coli plasmid and chromosomal recombination by the Ref function of bacteriophage P1. Genetics, , 123, , 465–476.
Lu,S.D., Lu,D. and Gottesman,M. ( (1989) ) Stimulation of IS1 excision by bacteriophage P1 ref function. J. Bacteriol., , 171, , 3427–3432.
Smith,H.W. ( (1972) ) Ampicillin resistance in Escherichia coli by phage infection. Nature New Biol., , 238, , 205–206.
Ikeda,H., Inuzuka,M. and Tomizawa,J.I. ( (1970) ) P1-like plasmid in Escherichia coli 15. J. Mol. Biol., , 50, , 457–470.
Yoshida,Y. and Mise,K. ( (1984) ) Characterization of generalized transducing phage w39 heteroimmune to phage P1 in Escherichia coli W39. Microbiol. Immunol., , 28, , 415–426.
Mise,K. and Suzuki,K. ( (1970) ) New generalized transducing bacteriophage in Escherichia coli. J. Virol., , 6, , 253–255.
Meyer,J., Stalhammar-Carlemalm,M., Streiff,M., Iida,S. and Arber,W. ( (1986) ) Sequence relations among the IncY plasmid p15B, P1, and P7 prophages. Plasmid, , 16, , 81–89.
Watkins,C.A. and Scott,J.R. ( (1981) ) Characterization of bacteriophage D6. Virology, , 110, , 302–317.
Sambrook,J., Fritsch,E.F. and Maniatis,T. ( (1989) ) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.
Lennox,E.S. ( (1955) ) Transduction of linked genetic characters of the host by bacteriophage P1. Virology, , 1, , 190–206.
Wall,J.D. and Harriman,P.D. ( (1974) ) Phage P1 mutants with altered transducing abilities for Escherichia coli. Virology, , 59, , 532–544.
Sauer,B., Whealy,M., Robbins,A. and Enquist,L. ( (1987) ) Site-specific insertion of DNA into a pseudorabies virus vector. Proc. Natl Acad. Sci. USA, , 84, , 9108–9112.
Sauer,B. ( (1993) ) Manipulation of the transgene by site-specific recombination: use of the cre recombinase. Meth. Enzymol., , 225, , 890–900.
Guzman,L.-M., Belin,D., Carson,M.J. and Beckwith,J. ( (1995) ) Tight regulation, modulation, and high-level expression by vectors containing the arabinose pBAD promoter. J. Bacteriol., , 177, , 4121–4130.
Gagneten,S., Le,Y., Miller,J. and Sauer,B. ( (1997) ) Brief expression of a GFPcre fusion gene in embryonic stem cells allows rapid retrieval of site-specific genomic deletions. Nucleic Acids Res., , 25, , 3326–3331.
Le,Y., Miller,J.L. and Sauer,B. ( (1999) ) GFPcre fusion vectors with enhanced expression. Anal. Biochem., , 270, , 334–336.
Eustice,D.C., Feldman,P.A., Colberg-Poley,A.M., Buckery,R.M. and Neubauer,R.H. ( (1991) ) A sensitive method for the detection of ?-galactosidase in transfected mammalian cells. BioTechniques, , 11, , 739–743.
Bethke,B. and Sauer,B. ( (1997) ) Segmental genomic replacement by Cre-mediated recombination: genotoxic stress activation of the p53 promoter in single-copy transformants. Nucleic Acids Res., , 25, , 2828–2834.
Altschul,S.F., Madden,T.L., Schaffer,A.A., Zhang,J., Zhang,Z., Miller,W. and Lipman,D.J. ( (1997) ) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res., , 25, , 3389–3402.
Brukner,I., Sanchez,R., Suck,D. and Pongor,S. ( (1995) ) Sequence-dependent bending propensity of DNA as revealed by DNase I: parameters for trinucleotides. EMBO J., , 14, , 1812–1818.
Gabrielian,A. and Pongor,S. ( (1996) ) Correlation of intrinsic DNA curvature with DNA property periodicity. FEBS Lett., , 393, , 65–68.
Hoess,R.H., Ziese,M. and Sternberg,N. ( (1982) ) P1 site-specific recombination: nucleotide sequence of the recombining sites. Proc. Natl Acad. Sci. USA, , 79, , 3398–3402.
Sternberg,N. and Coulby,J. ( (1987) ) Recognition and cleavage of the bacteriophage P1 packaging site (pac). II. Functional limits of pac and location of pac cleavage termini. J. Mol. Biol., , 194, , 469–479.
Skorupski,K., Sauer,B. and Sternberg,N. ( (1994) ) Faithful cleavage of the P1 packaging site (pac) requires two phage proteins, PacA and PacB, and two Escherichia coli proteins, IHF and HU. J. Mol. Biol., , 243, , 268–282.
Skorupski,K., Sternberg,N. and Sauer,B. ( (1994) ) Purification and DNA-binding activity of the PacA subunit of the bacteriophage P1 pacase enzyme. J. Mol. Biol., , 243, , 258–267.
Osborne,F.A., Stovall,S.R. and Baumstark,B.R. ( (1989) ) The c1 genes of P1 and P7. Nucleic Acids Res., , 17, , 7671–7680.
Sternberg,N., Sauer,B., Hoess,R. and Abremski,K. ( (1986) ) Bacteriophage P1 cre gene and its regulatory region: evidence for multiple promoters and for regulation by DNA methylation. J. Mol. Biol., , 187, , 197–212.
Sauer,B. ( (1987) ) Functional expression of the cre-lox site-specific recombination system in the yeast Saccharomyces cerevisiae. Mol. Cell. Biol., , 7, , 2087–2096.
Sauer,B. and Henderson,N. ( (1988) ) Site-specific DNA recombination in mammalian cells by the Cre recombinase of bacteriophage P1. Proc. Natl Acad. Sci. USA, , 85, , 5166–5170.
Lee,L. and Sadowski,P.D. ( (2003) ) Sequence of the loxP site determines the order of strand exchange by the Cre recombinase. J. Mol. Biol., , 326, , 397–412.
Hartung,M. and Kisters-Woike,B. ( (1998) ) Cre mutants with altered DNA binding properties. J. Biol. Chem., , 273, , 22884–22891.
Kim,S., Kim,G., Lee,Y. and Park,J. ( (2001) ) Characterization of Cre-loxP interaction in the major groove: hint for structural distortion of mutant Cre and possible strategy for HIV-1 therapy. J. Cell. Biochem., , 80, , 321–327.
Guo,F., Gopaul,D.N. and Van Duyne,G.D. ( (1997) ) Structure of Cre recombinase complexed with DNA in a site-specific recombination synapse. Nature, , 389, , 40–46.
Sauer,B. ( (1994) ) Site-specific recombination: developments and applications. Curr. Opin. Biotechnol., , 5, , 521–527.
Scott,J.R. ( (1968) ) Genetic studies on bacteriophage P1. Virology, , 36, , 564–574.
Wandersman,C. and Yarmolinsky,M. ( (1977) ) Bipartite control of immunity conferred by the related heteroimmune plasmid prophages, P1 and P7 (formerly phi Amp). Virology, , 77, , 386–400.(Brian Sauer1,2,* and Jeffrey McDermott1)