Resolvase-like recombination performed by the TP901-1 integrase

Anne Breüner1, Lone Br{oslash}ndsteda,1 and Karin Hammer1

Department of Microbiology, Technical University of Denmark, DK-2800 Lyngby, Denmark1

Author for correspondence: Karin Hammer. Tel: +45 45 25 24 96. Fax: +45 45 88 26 60. e-mail: karin.hammer{at}biocentrum.dtu.dk


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The site-specific recombination system of temperate lactococcal bacteriophage TP901-1 is unusual in several respects. First, the integrase belongs to the family of extended resolvases rather than to the {lambda} integrase family and second, in the presence of this integrase, a 56 bp attP fragment is sufficient for efficient recombination with the chromosomal attB site in the host Lactococcus lactis subsp. cremoris MG1363. In the present work, this attB site was analysed and a 43 bp attB region was found to be the smallest fragment able to participate fully in recombination. In vitro studies showed that the TP901-1 integrase binds this 43 bp attB fragment, the 56 bp attP and a larger attP fragment with equal affinity. Mutational analysis of the 5 bp common core region (TCAAT) showed that the TC dinucleotide is essential for recombination, but not for binding of the integrase, whereas none of the last three bases are important for recombination. When a number of attL sites, obtained by recombination between an attB site containing a mutation in this TC dinucleotide and a wild-type attP site, were sequenced, a mix of sites with the wild-type or the mutated sequence was obtained. These results are consistent with the hypothesis that the TC dinucleotide constitutes the TP901-1 overlap region. A 2 bp overlap region has been observed in recombination reactions catalysed by all other members of the resolvase/invertase family tested so far. By selecting for attB sites with a decreased ability to participate in recombination, two bases located outside the core region of attB were shown to be involved in the in vitro binding of the TP901-1 integrase.

Keywords: lactococcal bacteriophage, extended resolvase, site-specific recombination, mechanism of recombination, chromosomal attachment site

Abbreviations: Ap, ampicillin; Cm, chloramphenicol; Ery, erythromycin; Kn, kanamycin

The GenBank accession number for the sequence reported in this paper is Y15043.

a Present address: Department of Dairy and Food Science, Rolighedsvej 30, The Royal Veterinary and Agricultural University, DK-1958 Frederiksberg, Denmark.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
In most temperate bacteriophages, the phage genome is integrated in the host chromosome to ensure stable inheritance during lysogenic growth. Integration is obtained by site-specific recombination between the attachment sites attB and attP, located on the bacterial and the phage genomes respectively. In the resulting lysogenic strain, the prophage is integrated on the host chromosome between attL and attR, which are the recombination products of attB and attP.

TP901-1 is a temperate bacteriophage, induced by UV light from the lysogenic strain Lactococcus lactis subsp. cremoris 901. The attachment sites of TP901-1 before and after recombination are shown in Fig. 1(b). A 13 bp identical region with a mismatch at position 6 was identified by comparison of the attachment sites. Since the mismatch from attB was consistently found in attR and not attL, the core region, the region of homology between the attachment sites, was defined as being the 5 bp TCAAT region (Christiansen et al., 1994 ).



View larger version (26K):
[in this window]
[in a new window]
 
Fig. 1. (a) attB fragments tested in the deletion analysis of the attB region of the temperate bacteriophage TP901-1. The ability of each fragment to participate in site-specific recombination with attP is shown in percentages, 100% indicating that recombination takes place in all cells. (b) Nucleotide sequence of the minimal attB and attP sites, and of the corresponding attL and attR sites. The core and identical region (IR) are boxed. The repeats identified in attB (this work) and in attP (Br{oslash}ndsted & Hammer, 1999 ) are shown as arrows. The sites are written as: attB, B-core-B’; attP, P-core-P'; attL, B-core-P'; attR, P-core-B'. (c) Map of the plasmid pAB105, which is used in the attB analyses. attB can be inserted in the EcoRI and KpnI sites as shown. attB and attP are shown as black lines, with the orientation of the sites indicated, erm as a broken line. Recombination between attB and attP leads to attL being present on the plasmid and loss of erm.

 
The TP901-1 integrase, the only phage-encoded protein required for site-specific integration, belongs to the new protein family of extended resolvases (Christiansen et al., 1996 ). The extended resolvases contain a region with homology to the catalytic domain of the members of the resolvase family of resolvases and invertases, including the serine residue forming a covalently bound intermediate, and it is therefore likely that these proteins perform recombination in a similar way (resolvase-type recombination). During this type of recombination, double-stranded cuts are introduced into the DNA backbone, leading to 2 bp single-stranded overhangs. The strands are then rotated relative to one another, the overhangs are paired with their new partner, and the strands are religated. The overlap region of the resolvase family, the region between the cleavage sites on opposite strands, which consists of heteroduplex DNA after recombination, is thus 2 bp (Stark et al., 1989 ).

Seven members of the family of extended resolvases have been investigated functionally. These proteins are involved in the following site-specific recombination events: excision of transposons (TnpX, Crellin & Rood, 1997 ; TndX, Wang et al., 2000 ), excision of fragments from the chromosome (CisA, Sato et al., 1990 ; Popham & Stragier, 1992 ; XisF, Carrasco et al., 1994 ) and phage integration ({phi}C31 Int, Kuhstoss & Rao, 1991 ; Thorpe & Smith, 1998 ; TP01-1 Int, Christiansen et al., 1996 ; R4 Sre, Matsuura et al., 1996 ). For three of the recombinases (TnpX, {phi}C31 Int and R4 Sre) the conserved serine residue in the N-terminal part of the protein has been shown to be vital for the catalysis, and for TnpX excision a 2 bp overlap region has been demonstrated. These findings support the hypothesis that the extended resolvases perform resolvase-type recombination.

In contrast to the extended resolvase integrases, most phage integrases identified to date are related to the {lambda} integrase and are thus expected to perform integrase-type recombination, which involves a covalently bound tyrosine–DNA intermediate. Four integrase monomers bind to the substrate DNA sites, two of which are responsible for the introduction of specific single-stranded breaks in the DNA backbone of both substrate sites. The DNA strands are ligated to the new partners, creating a Holliday junction located in the approximate centre of the overlap region. Isomerization of the Holliday intermediate activates the second pair of integrase monomers, allowing cleavage and exchange of the second pair of DNA strands (Azaro & Landy, 1997 ; Gopaul & Van Duyne, 1999 ; Guo et al., 1999 ).

To ensure proper directionality of the recombination processes, the recognition sites attB and attP for phage integration are always of different sequence, in contrast to the substrate DNA sites for resolution or inversion processes. However, in the family of extended resolvases the {phi}C31 attB and attP sites are of similar size (minimum sizes 34 and 39 bp, respectively; Groth et al., 2000 ), and the {phi}C31 integrase has the same affinity for the two sites (Thorpe et al., 2000 ). In contrast, the minimal attB of Escherichia coli phage {lambda} is 23 bp in length, and thus much shorter than the 235 bp minimal attP site (Mizuuchi & Mizuuchi, 1985 ). The affinity of the {lambda} integrase for the attB site is much lower than for the attP site (Richet et al., 1988 ).

The TP901-1 minimal attP was found to be 56 bp long (Br{oslash}ndsted & Hammer, 1999 ), and is thus substantially smaller than the minimal attP regions reported for bacteriophages performing integrase-type recombination. To support the hypothesis that the TP901-1 integrase performs resolvase-type recombination further, we decided to investigate the boundaries of the TP901-1 attB site and the affinity of the integrase for the attP and attB sites. Furthermore, single bases involved in recombination were identified by mutational analysis, and the importance of some of these bases for the in vitro binding of the TP901-1 integrase was examined. Our data strongly support the hypothesis that the TP901-1 integrase performs resolvase-type recombination.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bacterial strains and culture conditions.
The bacterial strains used are listed in Table 1. Lactococcus lactis subsp. cremoris MG1363 was grown without shaking at 30 °C in M17 broth (Terzaghi & Sandine, 1975 ) supplemented with 0·5% (w/v) glucose (GM17). When required, erythromycin (Ery) was added to a final concentration of 2 mg l-1 and chloramphenicol (Cm) to 5 mg l-1. Escherichia coli cells were propagated at 37 °C with shaking in Luria–Bertani (LB) broth (Miller, 1972 ), and Ery was added to a final concentration of 150 mg l-1, Cm to 25 mg l-1, kanamycin (Kn) to 50 mg l-1 and ampicillin (Ap) to 100 mg l-1. IPTG was added to a final concentration of 1 mM. To prepare plates, all media were solidified by adding 1·5% Bacto agar.


View this table:
[in this window]
[in a new window]
 
Table 1. Bacterial strains and plasmids

 
DNA preparation.
Plasmid DNA was isolated from both lactococcal and E. coli cells by the alkaline lysis technique (Sambrook et al., 1989 ). To promote lysis, lactococcal cells were treated with lysozyme at a final concentration of 20 g l-1 at 37 °C with shaking before the addition of NaOH. If required, the DNA was further purified on Qiagen columns. Chromosomal DNA from lactococcal cells was prepared as described for E. coli (Sambrook et al., 1989 ), except that the cells were frozen for 30 min at -80 °C after harvesting, thawed to room temperature and treated with lysozyme at a final concentration of 20 g l-1 for 30 min at 37 °C to promote lysis. Phage DNA was prepared according to Sambrook et al. (1989) .

Recombinant DNA techniques.
DNA manipulations were performed by standard techniques (Sambrook et al., 1989 ). Enzymes and corresponding buffers were supplied by Pharmacia Biotech or New England Biolabs. The AmpliTaq DNA polymerase was used for PCR amplification of DNA fragments (Perkin Elmer Cetus). DNA sequences were determined as described by Sanger et al. (1977) , modified according to the Thermo Sequenase Radiolabelled terminator cycle sequencing kit (Amersham Life Science).

Oligonucleotides used in this study.
Oligonucleotides were supplied by T-A-G-Copenhagen, Symbion, Copenhagen, Denmark or by Pharmacia Biotech. Degenerate oligonucleotides were supplied by P. Hobolth, Lyngby, Denmark.

The sequences of the primers used were (written 5'–3'): T7, TAATACGACTCACTATAGGG; #1201, AACAGCTATGACCATG; #1211, GTAAAACGACGGCCAGT; perm, GTTACACGTTACTAAAGGG; perm1, GCAAGTCACGAACAC; pINT-1, CCCCGGATCCAGAAATGAGGTACAAAAAC; pINT-2, CCCCTCGAGTCGACGCAATTAAGCGAGTTGGAATTT; p1attB, CCCCGAATTCGATCCAACTCATAAAGTT; p3attB, CCCCGAATTCTGATAATTGCCAACAC; p4attB, GGGGGGGTACCATTTACCTTGATTGAGATG; p5attB, CCCCGAATTCATCTCAATCAAGGTAAATG; p6attB, AATTCATCTCAATCAAGGTAAATGTAC; p7attB, ATTTACCTTGATTGAGATG; p8attB, AAT-TACACAATTAACATCTCAATCAAGGTAAATGTAC; p9attB, ATTTACCTTGATTGAGATGTTAATTGTGT; pG10C, AATTCTGATAATTCCCAACACAATTAACATCTCAATCAAGGTAAATGGTAC; pG10C-rev, CATTTACCTTGATTGAGATGTTAATTGTGTTGGGAATTATCAG; pC33T, AATTCTGATAATTGCCAACACAATTAACATCTCAATTAAGGTAAATGGTAC; pC33T-rev, CATTTACCTTAATTGAGATGTTAATTGTGTTGGCAATTATCAG; pT28M, AATTCTGATAATTGCCAACACAATTAACATCMCAATCAAGGTAAATGGTAC; pC29D, AATTCTGATAATTGCCAACACAATTAACATCTDAATCAAGGTAAATGGTAC; pA30B, AATTCTGATAATTGCCAACACAATTAACATCTCBATCAAGGTAAATGGTAC; pA31B, AATTCTGATAATTGCCAACACAATTAACATCTCABTCAAGGTAAATGGTAC; pT32V, AATTCTGATAATTGCCAACACAATTAACATCTCAAVCAAGGTAAATGGTAC; pattB-Low and pattB-High, AATTCTGATAATTGCCAACACAATTAACATCTCAATCAAGGTAAATGGTAC; pattB-Low-rev and pattB-High-rev, CATTTACCTTGATTGAGATGTTAATTGTGTTGGCAATTATCAG. Bases in italics constitute the overhangs after annealing, underlined bases are changed relative to the wild-type sequence: M, A or C; D, G, A or T; B, G, T or C; V, G, A or C; pattB-Low, pattB-High, and pattB-High-rev are degenerate. Except at the bases constituting the single-stranded overhangs the level of degeneracy is reported later in this section.

Transformation of E. coli and L. lactis.
Plasmid DNA was introduced into E. coli cells by making the cells competent with CaCl2 and transforming as described by Sambrook et al. (1989) . To introduce DNA into lactococcal cells, they were made electrocompetent and electroporated (Holo & Nes, 1989 ).

Construction of pAB105 and deletion analysis of attB.
The plasmids used and constructed in this work are shown in Table 1 and the attB fragments are listed in Table 2. To construct plasmid pAB105 (Fig. 1c), pBF12 was digested with EcoRI, the larger fragment was religated, and the 1·1 kb erm cassette obtained by digesting pUC7,erm with BamHI was inserted in the BamHI site.


View this table:
[in this window]
[in a new window]
 
Table 2. attB fragments

 
Plasmids pAB121, pAB123, pAB124, pAB125 and pAB127 contain attB fragments 1, 3, 4, 5 and 7, respectively, inserted in the EcoRI–KpnI sites in pAB105. The fragments were obtained by PCR with pAB107 as template and digested with EcoRI and KpnI (primers: Table 2). pAB107 was obtained by digesting pBC193 (attB224) with HindIII and religating, resulting in attB1 being present. The attB8 and attB9 fragments in plasmids pAB128 and pAB129 were obtained by annealing pairs of oligonucleotides (Table 2), resulting in double-stranded attB fragments surrounded by EcoRI and KpnI overhangs. The sequence of all attB fragments was verified before further analysis.

Introducing site-specific mutations in attB.
When introducing base-substitutions in attB, again complementary oligonucleotides were annealed and the resulting fragment was inserted in the EcoRI–KpnI sites in pAB105. For the fragments G10C and C33T both oligonucleotides were of the desired sequence. The remaining attB fragments, which were obtained by introducing site-specific mutations in attB, were mutated in the core region (T28C, T28A, C29G C29A, C29T, A30G, A30T, A30C, A31G, A31T, A31C, T32G, T32C, T32A). To obtain the mutations by using a minimum of oligonucleotides, one of the pair of oligonucleotides annealed was a mix of primers with all bases except the original one at the desired position, while the other was wild-type.

Introducing random single-base-pair mutations in attB.
One procedure for obtaining random mutations in attB involved selecting for non-functional sites. One of the oligonucleotides annealed, pattB-Low, was degenerate except in the regions constituting the single-stranded overhangs, while the second primer was not (pattB-Low-rev). To select for non-functional sites, the pAB105-attBmut ligation mix was introduced directly into E. coli strain DH5{alpha}/pLB81 and plated directly on Ery plates. The nucleotide sequence was determined of all the non-functional attB sites to identify fragments with single-base-pair substitutions. Four such attB fragments were obtained: C17A, C17G, A25T and T28G. The remaining non-functional fragments contained multiple substitutions and/or deletions (data not shown).

The oligonucleotide pattB-Low is 2·4% degenerate. Since P(X), the probability of an oligonucleotide of length N containing X mutations at a level of degeneracy C, equals (N!/[X!(NX!)]!)xCX(1–C)N–X, P(0), the probability of one pattB-Low oligonucleotide being of the wild-type sequence is 35·0%, P(1) is 37·2%, and thus the probability of pattB-Low containing more than one mutation is 27·8%. Since pattB-Low was annealed to pattB-Low-rev, which is not degenerate, the probability of obtaining mutations is approximately half this.

In the second strategy for obtaining random single-base-pair mutations in attB, no selection for non-functional sites was involved, and therefore the oligonucleotides annealed had to be more degenerate than the pattB-Low-primers to avoid obtaining a high number of non-mutated sites. Both pattB-High and pattB-High-rev are 7·2% degenerate; thus P(0) is 4·0%, P(1) is 13·4%, P(2) is 21·8%, and the probability of pattB-High or pattB-High-rev containing more than two mutations is 60·8%. The sequence of a number of attB fragments inserted in pAB105 was determined, and two were found to contain single mutations (G3A and T9A); the remaining fragments analysed contained either no or multiple substitutions (data not shown).

Construction of remaining plasmids.
Plasmid pAB131 contains attBmin and attPmin separated by the erm gene, and was obtained by ligating the 150 bp AccI–HindIII fragment of pAB30 to the 4 kb AccI–HindIII fragment of pAB130. Plasmid pAB30 is identical to pBF30, except that erm is inserted in the opposite orientation by digesting pBF30 with BamHI and reannealing. The pAB130 plasmid was obtained by inserting erm on a 1·1 kb cassette from pUC7,erm in the BamHI site in pAB120. Plasmid pAB120 contains attBmin, inserted in the EcoRI–KpnI sites in pGEM-3Zf(-). The attBmin fragment was obtained by annealing pattB-Low and pattB-Low-rev. pattB-Low is degenerate, but a wild-type fragment was obtained after sequencing attB in two plasmids.

The integrase overexpression plasmid pAB202-4 was constructed by inserting the 1·4 kb BamHI–SalI fragment of pAB201, containing TP901-1 orf1, in BamHI–SalI in pUHE23-2. To construct pAB201, the TP901-1 orf1 gene and SD sequence was PCR amplified with primers pINT-1 and pINT-2 and TP901-1 DNA as template. The 1·4 kb PCR fragment was digested with BamHI and SalI, inserted in BamHI–SalI in pGEM3-Zf(-) and sequenced.

Investigation of the frequency of site-specific recombination.
Plasmid DNA of the pAB105 derivatives was introduced into DH5{alpha}/pLB81, in which the TP901-1 integrase is present and site-specific recombination between attP and attB can take place. The transformants were selected on LB plates containing Cm and Ap and, after growth overnight at 37 °C, a number of colonies were streaked both on LB with Cm and Ap and on LB with Cm, Ap and Ery. After another overnight growth the number of cells which had retained resistance to Ery could be determined. If zero to five single colonies were present on the plate with Ery, the colony was considered erythromycin sensitive (EryS); otherwise the colony was considered to be resistant (EryR). We estimate that approximately 104 cells are transferred in the streaking from the transformation plate without Ery to the plate with Ery, and thus this level of discrimination means that colonies which contain one, or fewer, erm-containing cell per 2000 cells will be considered EryS.

For the deletion analysis, 100 colonies were tested for each attB fragment (1, 3–5 and 7–9), and for the remaining fragments 50 colonies were tested.

Overexpression of the TP901-1 integrase in E. coli.
Overproduction of the TP901-1 integrase was induced in a culture of NF1830/pAB202-4 cells by addition of IPTG. The cells were resuspended in elution buffer (20 mM Tris-HCl, pH 7·5; 1 mM EDTA, pH 7·5; 100 mM NaCl; 5 mM MgCl2; 1 mM DTT; 5%, v/v, glycerol) and sonicated until lysis on ice. The extract was desalted on a HiTrap Desalting Column (Pharmacia Biotech) and eluted with 5x400 µl elution buffer. To monitor overexpression of the integrase, the extracts were applied on a SDS-PAGE gel (Laemmli, 1970 ). A protein of approximately 55 kDa, the predicted size of the TP901-1 integrase, was visible in the extract prepared from cells grown with IPTG, but absent in the extract prepared from cells grown without IPTG (results not shown). The integrase was estimated to constitute approximately 10% of the total protein content, which was determined to be 9·4 g l-1 by Lowry analysis, using bovine serum albumin as reference.

Gel mobility-shift assays.
Fragments were labelled by PCR with [{alpha}-32P]dATP. The following templates and primers were used: attPmin, template pBF30, primers perm1 and T7; attBmin, T28G and C17A, template pAB105 with the relevant fragment, primers perm and #1211. Binding was carried out in binding buffer (elution buffer with 40 mg bovine serum albumin l-1) with competitor DNA added to a final concentration of 0·5 mg l-1 (the erm cassette amplified by PCR on pUC7,erm plasmid DNA with primers #1201 and #1211). Protein extracts were added and the reaction was incubated for 10–15 min at room temperature. Labelled fragment was added to a final concentration of 0·06 nM and allowed to bind for another 25–30 min at room temperature (final volume: 10 µl). The reactions were run on a 5% polyacrylamide gel at ~7 V cm-1. After vacuum-drying onto 3 MM Whatman paper, the gel was acquired and analysed using a Packard Instant Imager.


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Strategy for analysis of attB
To analyse the TP901-1 attB site the E. coli vector pAB105 was constructed. This plasmid carries a functional attP site of 333 bp positioned next to the erm gene, and a polylinker region for insertion of a DNA fragment containing the attB sequences of interest on the other side of erm, relative to attP, resulting in the gene order attP-erm-attB (Fig. 1c). Thus, recombination between the attP and attB sites will result in excision of the erm gene and in erythromycin sensitivity of the strain carrying the plasmid. Since pAB105 also contains the bla gene, this plasmid can be selected for by ampicillin (Ap) independently of recombination. After insertion of a fragment containing the attB sequences of interest into pAB105, the plasmid is introduced into the E. coli strain DH5{alpha} already containing plasmid pLB81, from which the TP901-1 integrase is expressed. The ability of the inserted attB fragment to participate in recombination is then determined by testing the sensitivity of the ApR transformants towards Ery. If all transformants are EryS, recombination between attP and attB has taken place in all cells (100% recombination), while no recombination results in all cells being EryR.

Identification of the minimal attB
This system was used to investigate the boundaries of attB by testing the ability of attB sites of different lengths to recombine with attP (Fig. 1a). Both the 187 bp attB1 fragment and the 68 bp attB3 fragment are fully able to recombine with attP. By deletion of additional 23 bp (attB5) only 20% recombination is observed. Thus, these 23 bp stimulate, but are not required for, recombination. When 25 bp are deleted from the right end of attB1, attB3 and attB5, giving rise to attB4, attB7 and attB8, respectively, attB4 and attB7 are still able to recombine 100%, whereas no recombination is observed for attB8. Thus, the 43 bp attB7 fragment is the smallest attB fragment identified giving rise to full recombination, and is therefore termed the minimal attB or attBmin. With attB9, in which 13 bp are deleted from the left of attB7, 36% recombination is obtained. Thus, the 13 bp located between the left ends of attB9 and attB7 are required for full recombination.

As mentioned, pAB105 contains an attP site of 333 bp. To ensure that the size of the attP site does not affect the results of the deletion analysis, we constructed a plasmid (pAB131) containing the 43 bp attBmin and the 56 bp attPmin separated by the erm gene. When recombination was tested with this plasmid, all cells tested (100) were EryS, confirming that the minimal attB and attP fragments do contain the information required for recombination.

Functional analysis of the minimal attB in L. lactis
To investigate recombination between TP901-1 attP and attB in L. lactis, E. coli plasmids pAB130 (erm, attBmin) and pLB54 (erm, an attB fragment of 224 bp), were introduced into L. lactis subsp. cremoris MG1363 which already contained the plasmid pLB58 carrying a lactococcal origin of replication, a functional TP901-1 attP site and orf1 encoding the TP901-1 integrase. Since the plasmids pAB130 and pLB54 contain no lactococcal origin of replication, EryR transformants will only be obtained if they are integrated in pLB58 or the host chromosome. When pAB130 or pLB54 were introduced into MG1363/pLB58, the transformation frequencies were four orders of magnitude higher than when the control plasmid pBC144, containing no attB sequences, was introduced (Table 3). Thus, the attB sequences of pAB130 and pLB54 are responsible for the increase in transformation efficiency. Recombination was verified by PCR analysis on plasmid DNA prepared from EryR MG1363/pLB58 transformants (results not shown). In contrast, the presence of attB in pAB130 and pLB54 did not increase the transformation frequency compared to pBC144 in strain MG1363/pCI372, which contains neither attP nor the integrase. Thus, the increased frequency of transformation was only observed when both functional attP and attB sites are present in the cell, as well as the phage integrase.


View this table:
[in this window]
[in a new window]
 
Table 3. Transformation frequencies of L. lactis subsp. cremoris MG1363 strains

 
The efficiency of transformation of MG1363/pLB58 was very similar using the attB plasmids pLB54 or pAB130 (Table 3). Thus, the minimal attB fragment in pAB130 participates in integrative recombination with the same efficiency as the 224 bp attB fragment in pLB54. This confirms that the attB fragment in pAB130 contains the information required for efficient recombination by the TP901-1 integrase.

Mutational analysis of the attB core
Since sequencing of a large number of recombination products has revealed that the sequences of the attL and attR sites are always as shown in Fig. 1(b), recombination must take place in the 5 bp core (Christiansen et al., 1994 , and results not shown). To define the precise size of the overlap region, we performed an analysis of the importance of the bases of the attB core for recombination. Each of the five bases was changed by site-specific mutagenesis, and the effect on in vivo recombination was investigated in E. coli strain DH5{alpha} carrying pLB81 expressing the integrase, as described for the deletion analysis. The results are presented in Fig. 2(a).



View larger version (45K):
[in this window]
[in a new window]
 
Fig. 2. attB fragments containing single-base-pair mutations and the ability of these fragments to participate in recombination in E. coli. The 43 bp attBmin is shown at the top, bases numbered 1 to 43, core and identical region in bold. The changed bases are highlighted. The numbers to the right indicate the percentage of the isolates tested in which complete recombination had taken place. (a) attB fragments containing single-base-pair mutations in the core region. B: any base but A, V: any base but T. (b) attB fragments containing single-base-pair mutations outside the core region.

 
When the first T in the core region is changed to G or C (T28G and T28C), recombination is abolished completely, but for T28A 10% recombination can take place. Changing C29 also reduces recombination, although not as profoundly as changing the first base. The remaining three bases of the core, AAT, can be changed to any of the other three bases without any effect on recombination. Thus, only the two first bases of the core region were observed to be important for recombination, indicating that this TC dinucleotide could constitute the TP901-1 overlap region.

Identification of the TP901-1 overlap region
To obtain further evidence that the TC dinucleotide of the TP901-1 attachment sites does constitute the overlap region, we used the plasmids containing attB sites T28A, C29G, C29A and C29T, since the core mutation in these sites does not prevent recombination completely (Fig. 2a). The plasmids were introduced in the integrase-producing E. coli strain DH5{alpha}/pLB81, also used in the previous experiments, to allow recombination between the mutated attB and wild-type attP. If the mutation is located within the overlap region, recombination will lead to the pairing of two non-complementary bases. DNA replication or mismatch repair will correct this mismatch, enabling both bases to be present after recombination. The sequences of the attL site retained on the plasmids after recombination were determined from a number of EryS isolates, and are summarized in Table 4.


View this table:
[in this window]
[in a new window]
 
Table 4. Product after recombination between wild-type attP and mutated attB

 
For all four mutated attB sites tested, attL sites of both wild-type or mutated sequence were identified after recombination, as would be expected if the two bases are part of the overlap. For the three attB fragments with C29 mutated, an approximately 50–50 distribution of the wild-type and changed base was obtained (Table 4), as would be expected if the mismatch created during recombination is corrected by a random mechanism. In contrast, for the plasmid containing T28A, the attL sites of plasmids from 12 EryS isolates had to be sequenced before a wild-type attL site was obtained.Thus, there seem to a strong bias towards the base originating from attB being part of the attL site after recombination (Table 4).

Even though the bases to the right of this TC do not seem to be important for recombination (Fig. 2a), they could still be part of the overlap region. We therefore repeated the above experiment using the plasmids with the A30G and A31G attB sites. We sequenced the attL sites of 18 and 17 plasmids obtained after recombination, respectively, and found the wild-type sequence TCAAT in all cases (Table 4). The number of sequences determined rules out the possibility that the wild-type sequence obtained could be the result of mismatch repair. Thus, recombination takes place to the left of these bases, and the data are consistent with the hypothesis that the TC dinucleotide constitutes the overlap region.

Analysis of bases outside the core region
To identify those bases in the attB region outside the core that are important for recombination, the attB site was mutagenized randomly, and non-functional sites were selected as described in Methods. Approximately 450 attB fragments were tested in this way, of which 21 displayed a reduced ability to participate in recombination. Sequence analysis showed that four of these contained single-base-pair substitutions (Fig. 2b). In two of the fragments C17 was mutated, and it was observed that an A at this position affected the ability of the fragment to participate in recombination more profoundly than a G. In the third fragment A25 was changed to a T, which had only a slight effect on recombination. The fourth fragment obtained by this procedure, the T28G attB fragment, contained a substitution in the core region and was discussed in the previous section (Fig. 2a).

To identify single bases that are not important for recombination, attB was mutagenized without selecting for a non-functional site. Four attB fragments with single-base-pair substitutions were isolated, which were mutated in either G3, T9, G10 or C33 (Fig. 2b).

Binding of the integrase to the minimal attP and attB sites
To investigate the binding of the TP901-1 integrase to the attachment sites, we decided to perform gel mobility-shift assays, using a crude extract containing the TP901-1 integrase (see Methods). In Fig. 3, binding of the integrase to the 56 bp attPmin site and the 43 bp attBmin site is shown. No retardation of the fragments was observed with the control extract containing no integrase, nor did the integrase bind to a labelled fragment of similar size containing no attP or attB sequences (data not shown).



View larger version (38K):
[in this window]
[in a new window]
 
Fig. 3. Binding of the TP901-1 integrase to the minimal attP and attB fragments. Gel retardation experiments with labelled attPmin (a) and attBmin (b) fragments and the TP901-1 integrase extract. Relative amounts of integrase extract added are indicated. Undiluted extract corresponds to a final concentration of total protein in the assay of 1·9 g l-1. The positions of the bound and free fragments on the gels are indicated. (c) Percentage of the fragments bound (relative to total amount of labelled fragment) as a function of the amount of protein added. {blacktriangleup}, attPmin; {blacksquare}, attBmin.

 
The integrase binds to attBmin and attPmin with nearly the same affinity (Fig. 3). The affinity of the integrase for the two fragments was further compared in a competition assay, in which both fragments were present in the same binding reaction. The two fragments were discriminated in size by elongation of one with approximately 200 bp of the flanking erm cassette. This experiment verified that the affinity of the integrase for the two fragments is identical (results not shown).

An additional binding assay showed that the integrase binds a 333 bp attP fragment with the same affinity as attPmin and attBmin (data not shown). This confirms that the minimal attachment sites contain all the information required for binding of the integrase.

Effect of attB mutations on integrase binding
A number of mutated attB fragments with reduced ability to participate in recombination have been isolated. The reduction could be due either to an altered binding of the integrase to the fragment, or to an impairment later in the recombination process. To discriminate between these possibilities, the in vitro binding of the integrase to selected, mutated attB fragments was investigated.

The T28G attB fragment is not able to participate in recombination at all (Fig. 2a). However, the integrase binds this fragment with the same affinity as wild-type attB (Fig. 4a, b). Thus, the inability of the T28G attB fragment to participate in recombination is not due to a decreased affinity of the integrase for the fragment. In contrast, binding of the integrase to C17A, which can participate in recombination at a level of 21% relative to wild-type attB (Fig. 2b) is severely reduced (Fig. 4c, d), and binding to the A25T fragment, which leads to a less pronounced reduction in recombination, is slightly reduced (data not shown). Thus, the substitutions in the C17A and A25T fragments result in reduced affinities of the TP901-1 integrase for the fragments, and the relative reduction in affinity correlates well with the relative decrease in the ability of the fragments to participate in recombination.



View larger version (39K):
[in this window]
[in a new window]
 
Fig. 4. Binding of the TP901-1 integrase to the T28G and C17A fragments, compared to binding to the attBmin fragment. (a) Gel retardation experiments with the T28G fragment and the TP901-1 integrase extract. Relative amounts of integrase extract added, and the positions of the bound and free fragments on the gels, are indicated. (b) Percentage of the T28G fragment bound (relative to total amount of labelled fragment) as a function of the amount of protein added ({blacktriangleup}). The similar results for the wild-type attBmin fragment are included for comparison ({blacksquare}). (c) Gel retardation experiments with the C17A fragment and a crude extract containing the TP901-1 integrase. (d) Percentage of the C17A fragment bound ({bullet}) compared to the wild-type attBmin ({blacksquare}).

 
Repeats in the minimal attB and attP sites
The nucleotide sequences of the 56 bp minimal attP and the 43 bp minimal attB were inspected for repeated sequences, since these could be integrase recognition sites. Three copies of a 9 bp sequence (TGTTAATTG), with a number of mismatches, were observed in the minimal attB (Fig. 1b). These repeats are almost identical to the two P1 repeats previously reported in the minimal attP (Br{oslash}ndsted & Hammer, 1999 ). If these repeats are integrase recognition sites, it would be expected that mutations that decrease the affinity of the integrase for the fragment are located in these repeats. And indeed, the two bases that have been found to be required for proper binding of the integrase (C17 and A25) are located at both ends of the middle repeat. However, mutations have been introduced in the two other repeats without an effect on recombination (B1: G3A, T9A, G10C; B3: A30B, A31B, T32V, C33T).

Location of attB on the lactococcal chromosome
Many temperate bacteriophages integrate their genomes in intergenic sequences or in the 3' end of tRNA genes (for review see Campbell, 1992 ); however, it has earlier been reported that this is not the case for TP901-1 (Christiansen et al., 1994 ). We determined the nucleotide sequence of the region of the lactococcal chromosome surrounding the TP901-1 attB site (GenBank accession number: Y15043), and found that attB is located within what appears to be an operon, the first three genes of which encode the proteins Orf311, Orf283 and Orf125. These proteins show very high homology to the ComYA, ComYB and ComYC proteins of Streptococcus gordonii, and to the ComGA, ComGB and ComGC proteins of Bacillus subtilis, which are involved in the development of natural competence in both strains (Lunsford & Roble, 1997 ; Chung & Dubnau, 1998 ). Integration of TP901-1 in attB results in orf125 being disrupted, but since L. lactis is not known to be able to develop natural competence, there exists no obvious way to investigate the function of this putative protein or the physiological effects of disruption of the ORF.


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Identification of the TP901-1 minimal attB
Recombination by the TP901-1 integrase can take place in E. coli as well as in Lactococcus. This capacity to function in hosts other than the original one has previously been described for another member of the family of extended resolvases, the integrase of Streptomyces temperate bacteriophage {phi}C31 (E. coli: Thorpe & Smith, 1998 ; Homo sapiens: Groth et al., 2000 ). It has furthermore been shown that only the purified {phi}C31 integrase is required for recombination in vitro (Thorpe et al., 2000 ).

The smallest fully functional attB site identified is 43 bp, and is termed the minimal attB or attBmin (Fig. 1a). This fragment includes 27 bp upstream and 11 bp downstream of the 5 bp core. When 23 bp are deleted from the upstream part of this fragment, no recombination is possible. Curiously, the presence of an additional 25 bp downstream of attB8 restores recombination to a level of 20%, even though these bases are not part of the minimal attB. Upon inspection of the sequence it is observed that these 25 bp contain two T-stretches of 5 and 7 bp, respectively, separated by a GC dinucleotide. One possible explanation for the stimulatory effect of this region on recombination could be the formation of some special conformation that might enhance binding of a protein required for the process – possibly the integrase.

The TP901-1 overlap region
A substitution of either of the first two bases (TC) of the 5 bp attB core reduces the frequency of in vivo recombination. Furthermore, when a number of attL products, obtained after recombination between such a mutated attB fragment and a wild-type attP, were analysed, a mix of plasmids with and without the substitution was obtained. In contrast, when the attB site contains a mutation in A30 or A31, located just downstream of the TC dinucleotide, only wild-type recombination products were obtained. The number of recombinants analysed was high enough that we feel safe to conclude that recombination does take place just downstream of the TC dinucleotide. With the T28A mutant only one wild-type recombination was obtained of the 12 plasmids sequenced (Table 4). However, the inhibitory effect on recombination of any kind of substitution in this position strongly suggests that T28 belongs to the overlap region. Furthermore it was shown that the T28G substitution did not affect binding of the integrase to attB; hence it has to affect the catalytic activity. Taken together, our results strongly suggest that the TC dinucleotide constitutes the overlap region, which again confirms the hypothesis that the TP901-1 integrase performs resolvase-type recombination.

TC/GA overlap regions have been described both for the Gin invertase of bacteriophage Mu (Mertens et al., 1988 ) and for the extended resolvase TnpX, which is involved in site-specific excision of the conjugative transposon Tn4451 (Crellin & Rood, 1997 ; Bannam et al., 1995 ). A mutation analysis revealed that changing the T of the Tn4451 overlap region has a more profound effect on recombination than changing the C, in agreement with the observations in the present study.

Binding of the TP901-1 integrase to the attP and attB sites
The TP901-1 integrase binds the minimal attB and attP sites with equal affinity, resulting in a single retarded band in gel mobility-shift assays. Thus, a likely mechanism for bringing together the DNA species in recombination is that integrase proteins bind to both attP and attB and the DNA–protein complexes are then paired. This mechanism for DNA binding resembles that of the resolvases and invertases, rather than that of the {lambda} integrase (Richet et al., 1988 ). Likewise the integrase of Streptomyces phage {phi}C31, which is another extended resolvase, has also been shown to bind attP and attB with similar affinity (Thorpe et al., 2000 ).

The TP901-1 integrase binds to the T28G fragment, which contains a mutation in the overlap region that prevents recombination completely, as well as to the wild-type attB (Fig. 4a, b). This observation confirms the suggestion that this TC dinucleotide constitutes the TP901-1 overlap region. Both of the mutations outside the core which affect recombination were found also to lead to a decrease in the affinity of the integrase for the fragment [C17A: Fig. 4(a, b); A25T: results not shown), suggesting that the decrease in recombination is the result of impaired binding of the integrase. Both C17 and A25 are situated in the central part of attB, at both ends of the B2 repeat (Fig. 1b). The suggestion that the repeats identified in attB are involved in integrase binding is further strengthened by the observations that the deletion of the B1 repeat in attB fragment 9 reduces recombination to 36% (Fig. 1a), and that related sequences are found in attP (Fig. 1b).

Other proteins binding to the attachment sites
The small size of both of the TP901-1 attachment sites suggests that the number of proteins binding to the attachment sites is limited. This observation correlates well with the suggestion that the integrase of TP901-1 performs resolvase-type recombination, since neither the resolvases nor the invertases require the binding of additional proteins to the res or inv sites for recombination. However, if host-encoded proteins are required to bind the attachment sites during integration, they must be present both in E. coli and L. lactis, since recombination can take place in both hosts. Another candidate for a protein binding to the attachment sites is the TP901-1 excisionase, encoded by TP901-1 orf7, which is required for efficient excisive recombination (Breüner et al., 1999 ). In the TP901-1 prophage, binding of Orf7 to either attL or attR could stimulate excision by inducing a change in the conformation of the DNA, facilitating excisive recombination.


   ACKNOWLEDGEMENTS
 
We sincerely appreciate the expert technical assistance of Lotte Bredahl and Lise S{oslash}rensen. This work was supported by the FØTEK program through The Centre of Advanced Food Studies and by grants from the EC STARLAB program (BIO4-CT96-0402) and the Carlsberg Foundation.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Azaro, M. A. & Landy, A. (1997). The isomeric preference of Holliday junctions influences resolution bias by lambda integrase. EMBO J 16, 3744-3755.[Abstract/Free Full Text]

Bannam, T. L., Crellin, P. K. & Rood, J. I. (1995). Molecular genetics of the chloramphenicol resistance transposon Tn4451 from Clostridium perfringens: the TnpX site-specific recombinase excises a circular transposon molecule. Mol Microbiol 16, 535-551.[Medline]

Bonekamp, F., Clemmensen, K., Karlström, O. & Jensen, K. F. (1984). Mechanism of UTP-modulated attenuation at the pyrE gene of Escherichia coli: an example of operon polarity control through the coupling of translation to transcription. EMBO J 3, 2857-2861.[Abstract]

Breüner, A., Br{oslash}ndsted, L. & Hammer, K. (1999). Novel organization of genes involved in prophage excision identified in the temperate lactococcal bacteriophage TP901-1. J Bacteriol 181, 7291-7297.[Abstract/Free Full Text]

Br{oslash}ndsted, L. & Hammer, K. (1999). Use of the integration elements encoded by the temperate lactococcal bacteriophage TP901-1 to obtain chromosomal single-copy transcriptional fusions in Lactococcus lactis. Appl Environ Microbiol 65, 752-758.[Abstract/Free Full Text]

Campbell, A. M. (1992). Chromosomal insertion sites for phages and plasmids. J Bacteriol 174, 7495-7499.[Medline]

Carrasco, C. D., Ramaswamy, K. S., Ramasubramanian, T. S. & Golden, J. W. (1994). Anabaena xisF gene encodes a developmentally regulated site-specific recombinase. Genes Dev 8, 74-83.[Abstract]

Christiansen, B. (1995). Site-specific integration of the lactococcal temperate phage TP901-1. PhD thesis, Technical University of Denmark.

Christiansen, B., Johnsen, M. G., Stenby, E., Vogensen, F. K. & Hammer, K. (1994). Characterization of the lactococcal temperate phage TP901-1 and its site-specific integration. J Bacteriol 176, 1069-1076.[Abstract]

Christiansen, B., Br{oslash}ndsted, L., Vogensen, F. K. & Hammer, K. (1996). A resolvase-like protein is required for the site-specific integration of the temperate lactococcal bacteriophage TP901-1. J Bacteriol 178, 5164-5173.[Abstract]

Chung, Y. S. & Dubnau, D. (1998). All seven comG open reading frames are required for DNA binding during transformation of Bacillus subtilis. J Bacteriol 180, 41-45.[Abstract/Free Full Text]

Crellin, P. K. & Rood, J. I. (1997). The resolvase/invertase domain of the site-specific recombinase TnpX is functional and recognizes a target sequence that resembles the junction of the circular form of the Clostridium perfringens transposon Tn4451. J Bacteriol 179, 5148-5156.[Abstract]

Gasson, M. J. (1983). Plasmid complements of Streptococcus lactis NCDO 712 and other lactic streptococci after protoplast-induced curing. J Bacteriol 154, 1-9.[Medline]

Gopaul, D. N. & Van Duyne, G. D. (1999). Structure and mechanism in site-specific recombination. Curr Opinion Struct Biol 9, 14-20.[Medline]

Groth, A. C., Olivares, E. C., Thyagarajan, B. & Calos, M. P. (2000). A phage integrase directs efficient site-specific integration in human cells. Proc Natl Acad Sci USA 97, 5995-6000.[Abstract/Free Full Text]

Guo, F., Gopaul, N. & Van Duyne, G. D. (1999). Asymmetric DNA bending in the Cre-loxP site-specific recombination synapse. Proc Natl Acad Sci USA 96, 7143-7148.[Abstract/Free Full Text]

Hayes, F., Daly, C. & Fitzgerald, G. F. (1990). Identification of the minimal replicon of Lactococcus lactis subsp. lactis UC317 plasmid pCI305. Appl Environ Microbiol 56, 202-209.

Holo, H. & Nes, I. (1989). High-frequency transformation, by electroporation, of Lactococcus lactis subsp. cremoris grown with glycine in osmotically stabilized media. Appl Environ Microbiol 55, 3119-3123.

Israelsen, H., Madsen, S. M., Vrang, A., Hansen, E. B. & Johansen, E. (1995). Cloning and partial characterization of regulated promoters from Lactococcus lactis Tn917-lacZ integrants with the new promoter probe vector, pAK80. Appl Environ Microbiol 61, 2540-2547.[Abstract]

Kuhstoss, S. & Rao, R. N. (1991). Analysis of the integration function of the streptomycete bacteriophage {phi}C31. J Mol Biol 222, 897-908.[Medline]

Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685.[Medline]

Lunsford, R. D. & Roble, A. G. (1997). comYA, a gene similar to comGA of Bacillus subtilis, is essential for competence-factor-dependent DNA transformation in Streptococcus gordonii. J Bacteriol 179, 3122-3126.[Abstract]

Lutz, R. & Bujard, H. (1997). Independent and tight regulation of transcriptional units in Escherichia coli via the LacR/O, the TetR/O and AraC/I1-I2 regulatory elements. Nucleic Acids Res 25, 1203-1210.[Abstract/Free Full Text]

Matsuura, M., Noguchi, T., Yamaguchi, D., Aida, T., Asayama, M., Takahashi, H. & Shirai, M. (1996). The sre gene (ORF469) encodes a site-specific recombinase responsible for integration of the R4 phage genome. J Bacteriol 178, 3374-3376.[Abstract]

Mertens, G., Klippel, A., Fuss, H., Blöcker, H., Frank, R. & Kahmann, R. (1988). Site-specific recombination in bacteriophage Mu: characterization of binding sites for the DNA invertase Gin. EMBO J 7, 1219-1227.[Abstract]

Miller, J. H. (1972). Experiments in Molecular Genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Mizuuchi, M. & Mizuuchi, K. (1985). The extent of DNA sequence required for a functional bacterial attachment site of phage {lambda}. Nucleic Acids Res 13, 1193-1208.[Abstract]

Popham, D. L. & Stragier, P. (1992). Binding of the Bacillus subtilis spoIVCA product to the recombination sites of the element interrupting the {sigma}K-encoding gene. Proc Natl Acad Sci USA 89, 5991-5995.[Abstract]

Richet, E., Abcarian, P. & Nash, H. A. (1988). Synapsis of attachment sites during lambda integrative recombination involves capture of a naked DNA by a protein-DNA complex. Cell 52, 9-17.[Medline]

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Sanger, F., Nicklen, S. & Coulson, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci USA 74, 5463-5467.[Abstract]

Sato, T., Samori, Y. & Kobayashi, Y. (1990). The cisA cistron of Bacillus subtilis sporulation gene spoIVC encodes a protein homologous to a site-specific recombinase. J Bacteriol 172, 1092-1098.[Medline]

Stark, W. M., Boocock, M. R. & Sherrat, D. J. (1989). Site-specific recombination by Tn3 resolvase. Trends Genet 5, 304-309.[Medline]

Terzaghi, B. E. & Sandine, W. E. (1975). Improved medium for lactic streptococci and their bacteriophages. Appl Microbiol 29, 807-813.

Thorpe, H. M. & Smith, M. C. M. (1998). In vitro site-specific integration of bacteriophage DNA catalyzed by a recombinase of the resolvase/invertase family. Proc Natl Acad Sci USA 95, 5505-5510.[Abstract/Free Full Text]

Thorpe, H. M., Wilson, S. E. & Smith, M. C. M. (2000). Control of directionality in the site-specific recombination system of the Streptomyces phage {phi}C31. Mol Microbiol 38, 232-241.[Medline]

Wang, H., Roberts, A. P., Lyras, D., Rood, J. I., Wilks, M. & Mullany, P. (2000). Characterization of the ends and target sites of the novel conjugative transposon Tn5397 from Clostridium difficile: excision and circularization is mediated by the large resolvase, TndX. J Bacteriol 182, 3775-3783.[Abstract/Free Full Text]

Received 21 December 2000; revised 19 March 2001; accepted 11 April 2001.