(Received for publication, September 17, 1996, and in revised form, December 10, 1996)
From the Department of Biological Chemistry and
Molecular Pharmacology, Harvard Medical School, Boston, Massachusetts
02115 and the § Laboratory of Molecular and Cellular
Biology, NIDDK, National Institutes of Health, Bethesda, Maryland
20892-0830
Bacteriophage T7 gene 2.5 single-stranded DNA-binding protein and gene 4 DNA helicase together
promote pairing of two homologous DNA molecules and subsequent polar
branch migration (Kong, D., and Richardson, C. C. (1996) EMBO
J. 15, 2010-2019). In this report, we show that gene 2.5 protein
is not required for the initiation or propagation of strand transfer
once a joint molecule has been formed between the two DNA partners, a
reaction that is mediated by the gene 2.5 protein alone. A mutant gene
2.5 protein, gene 2.5-21C protein, lacking 21 amino acid residues at
its C terminus, cannot physically interact with gene 4 protein.
Although it does bind to single-stranded DNA and promote the formation
of joint molecule via homologous base pairing, subsequent strand
transfer by gene 4 helicase is inhibited by the presence of the gene
2.5-
21C protein. Bacteriophage T4 gene 32 protein likewise inhibits
T7 gene 4 protein-mediated strand transfer, whereas Escherichia
coli single-stranded DNA-binding protein does not. The 63-kDa
gene 4 protein of phage T7 is also a DNA primase in that it catalyzes the synthesis of oligonucleotides at specific sequences during translocation on single-stranded DNA. We find that neither the rate nor
extent of strand transfer is significantly affected by concurrent
primer synthesis. The bacteriophage T4 gene 41 helicase has been shown
to catalyze polar branch migration after the T4 gene 59 helicase
assembly protein loads the helicase onto joint molecules formed by the
T4 UvsX and gene 32 proteins (Salinas, F., and Kodadek, T. (1995)
Cell 82, 111-119). We find that gene 32 protein alone
forms joint molecules between partially single-stranded homologous DNA
partners and that subsequent branch migration requires this
single-stranded DNA-binding protein in addition to the gene 41 helicase
and the gene 59 helicase assembly protein. Similar to the strand
transfer reaction, strand displacement DNA synthesis catalyzed by T4
DNA polymerase also requires the presence of gene 32 protein in
addition to the gene 41 and 59 proteins.
Single-stranded DNA (ssDNA)1 binding proteins, a class of ubiquitous proteins, not only are essential in DNA replication (1-3) but also play key roles in DNA recombination and repair (4-11). A number of studies have shown that the ssDNA-binding proteins encoded by Escherichia coli and its phages T4 and T7 exert significant function in maintaining the normal level of DNA recombination in vivo. Mutations that alter E. coli, T4, or T7 ssDNA-binding protein, the products of the ssb gene of E. coli, gene 32 of phage T4, or gene 2.5 of phage T7, can significantly depress DNA recombination frequencies in vivo (5, 6, 10, 11).
Although ssDNA-binding proteins are clearly involved in recombination, it has proven difficult to define their precise roles since they not only bind to DNA but also physically interact with other proteins of DNA metabolism to modulate their reactions. In addition, their involvement in both replication and recombination has made specific assignments of function in vivo speculative. One property, however, their ability to stimulate the annealing of complementary DNA strands (12-15), does provide a function that is essential to the overall process of general recombination. Both the T4 gene 32 and T7 gene 2.5 proteins have been shown to be essential for this early step in recombination (10, 12, 13).
Recombination proteins have been identified from a number of organisms
that expose single-stranded regions in order for ssDNA-binding proteins
of the type discussed above to facilitate annealing of the
complementary sequences (reviewed in Ref. 16). We have shown (15) that
the T7 gene 2.5 ssDNA-binding protein and the gene 4 helicase mediate
both homologous base pairing and subsequent polar branch migration
between a single-stranded circle and a linear duplex with a short
single-stranded tail created by 5 to 3
digestion by the T7 gene 6 exonuclease (Fig. 1). Strand transfer driven by the T7 gene 4 helicase
is coupled to hydrolysis of nucleoside 5
-triphosphate and proceeds at
a high rate (>120 nucleotides/s) in a polar 5
to 3
direction with
respect to the invading strand.
Proteins such as the RecA protein of E. coli (17), the UvsX protein of phage T4 (18-20), and the RAD 51 protein of Saccharomyces cerevisiae (21) mediate homologous base pairing and subsequent strand exchange between ssDNA and a fully duplex partner. The overall process of strand exchange by these RecA-like proteins is stimulated by ssDNA-binding proteins (17-20, 22, 23). Recent experiments with the T4 recombination system (24) showed that the T4 UvsY protein promotes homologous pairing by the UvsX protein but inhibits subsequent UvsX-catalyzed branch migration. Polar branch migration is instead driven by the T4 gene 41 helicase, which is loaded on the gene 32 protein-covered DNA by the T4 gene 59 helicase assembly protein.
In this report we have used the joint molecule formed between a single-stranded circle and linear duplex with a short single-stranded tail (Fig. 1) to compare the roles of the T7 and T4 helicases and ssDNA-binding proteins in polar branch migration. We show that the T4 helicase in the presence of T4 gene 32 protein and gene 59 protein catalyzes strand transfer in a manner similar to that previously described for the T7 helicase (15), thus allowing us to examine the role of the gene 32 protein in reactions devoid of the T4 UvsX and UvsY proteins.
One role of gene 2.5 protein in the overall T7 reaction is to promote
the formation of the joint molecule, a requisite intermediate for
strand transfer by the helicase. In this role the gene 2.5 protein can
be distinguished from the T4 or E. coli ssDNA-binding proteins in that its affinity (Ka = 1-4 × 106/M) for ssDNA (25) is approximately 50-fold
lower than that of the latter two proteins and it facilitates the
renaturation of ssDNA much more efficiently provided that either
Mg2+ or a relatively high concentration of NaCl is
present.2 In the simplified T4 reaction
described above, the gene 32 protein and gene 59 protein are required
to form the joint molecule and load the helicase. An unresolved
question with regard to both systems, one addressed by the current
study, is whether or not the ssDNA-binding proteins play a direct role
in the strand transfer reaction per se. For example, both
the T7 gene 4 and T4 gene 41 proteins share a requirement for a
relatively short single-stranded 5-tail to initiate translocation and
display helicase activity on a duplex DNA molecule (26, 27). It is
likely that there is sufficient nonprotein branch migration at the
junction between single-stranded and duplex DNA to provide for this
requirement, but it is not known if an ssDNA-binding protein is
required to stabilize this partially unwound structure. Likewise, it is
not known if helicase interactions with ssDNA-binding protein on the invading strand are required or if the binding of ssDNA-binding protein
to the single-stranded region created behind the translocating helicase
is essential to maintain unidirectional strand transfer. Our earlier
finding (15) that a mutant gene 2.5 protein, gene 2.5-
21C protein,
could not support strand transfer by the T7 gene 4 protein is most
likely due to its inability to physically interact with the gene 4 protein (28). The lack of an interaction between the two proteins could
exert its negative effect either by an inability of the gene 2.5-
21C
protein to load the gene 4 protein onto the joint molecule or to a lack
of stimulation during the actual strand transfer. Alternatively, a
specific interaction between the two proteins may be required for the
gene 4 protein to displace gene 2.5 protein bound to ssDNA. The same
explanations can also be extended to the inability of the T4 gene 32 protein to substitute for the T7 gene 2.5 protein in the T7 helicase
reaction.
In this report we show that the T7 gene 2.5 protein is not required for polar branch migration catalyzed by the T7 gene 4 helicase. In contrast, T4 gene 32 protein is required both for base pairing and for extensive strand exchange in the T4 gene 41 helicase-mediated reaction. This requirement for T4 gene 32 protein during recombination is in accord with its requirement for extensive strand displacement synthesis by the T4 DNA replication system in vitro (29).
T7 gene 2.5 (15) and T7 gene 2.5-21C (28) were
overexpressed, and the proteins were purified to apparent homogeneity (>98% pure) from E. coli cells as described. The 63-kDa
gene 4 protein (>98% pure), a species of gene 4 protein that has both helicase and primase activities, was prepared by B. Beauchamp (Harvard
Medical School) as described (30). T4 gene 41 DNA helicase (31) and
gene 59 protein (32) were purified to apparent homogeneity (>98%
pure) as described. The purity of proteins was determined by Coomassie
Blue staining of the proteins on polyacrylamide gels containing sodium
dodecyl sulfate. T4 gene 32 protein, E. coli ssDNA-binding
(SSB) protein, T7 gene 6 exonuclease, and restriction enzymes were
purchased from U.S. Biochemical Corp.
M13mp18 circular, ssDNA (33)
and M13mp18 RF1 DNA (34) were prepared as described previously. To
prepare linear, duplex M13mp18 bearing single-stranded 3-termini of
approximately 100 nucleotides, M13mp18 RF1 DNA was cut by restriction
enzyme SmaI and then hydrolyzed with T7 gene 6 5
to 3
exonuclease to a limited extent as described previously (15). All
nucleotides were purchased from Pharmacia Biotech Inc. Concentrations
of DNA are expressed in nucleotide equivalents.
Joint DNA molecules consisting of circular,
single-stranded M13 DNA annealed to the single-stranded region of
homologous linear, duplex DNA having 3 single-stranded termini were
prepared by incubating circular, single-stranded M13mp18 DNA (10 µM as nucleotide equivalents) and linear, duplex M13mp18
DNA having single-stranded termini (20 µM) in a standard
reaction containing 25 mM Tris-HCl (pH 7.5), 5 mM dithiothreitol, 10 mM MgCl2 (15 mM MgCl2 when gene 32 protein was present), and
2 µM T7 gene 2.5 protein or T4 gene 32 protein. The
reaction was incubated at 32 °C for 30 min and then stopped by the
addition of EDTA, SDS, and proteinase K to 50 mM, 0.5%,
and 0.6 mg/ml, respectively. After an additional incubation at 32 °C
for 10 min, the reaction mixture was extracted sequentially by phenol,
phenol/chloroform, and chloroform to remove the gene 2.5 or 32 proteins. The DNA was precipitated with ethanol and resuspended in TE
(10 mM Tris-HCl, pH 7.5, 1 mM EDTA) buffer. Agarose gel analysis showed that approximately 40-80% of the
substrates were converted to joint molecules. Further purification of
the joint molecule to eliminate circular ssDNA or linear dsDNA is not
necessary, since the two species do not affect either the reaction or
the detection of products of strand transfer with agarose gel
electrophoresis.
Reactions in which joint molecules were used directly to measure strand exchange contained 25 mM Tris-HCl, pH 7.5, 5 mM dithiothreitol, 15 mM MgCl2, 10-15 µM joint molecules, 5 mM dTTP or ATP (for T7 gene 4 protein) or ATP (for T4 gene 41 protein), and 0.2 µM T7 gene 4 protein unless stated otherwise. After incubation at 32 °C for 20 min, the reactions were stopped by the addition of EDTA, SDS, and proteinase K to 50 mM, 0.5%, and 0.6 mg/ml, respectively, and incubated at 32 °C for additional 10 min. Samples were analyzed by 0.8% agarose gel electrophoresis as described below. If strand transfer reactions were catalyzed by T4 gene 32, 59, and 41 proteins, the reaction conditions were as described above except that gene 32 (1.5 µM), 59 (0.2 µM), and 41 (0.2 µM monomer) proteins were substituted for T7 gene 4 protein.
In some reactions catalyzed by the T4 gene 32, 41, and 59 proteins, the formation of joint molecules and strand transfer were carried out in two steps. In the first step, joint molecules were formed in a reaction with only gene 32 protein, using the conditions described above. In the second step, 0.2 µM each (monomer) of the gene 41 and 59 proteins and 5 mM ATP were added to initiate the reaction. Reactions were then incubated at 32 °C for 20 min, and the products were analyzed by agarose gel electrophoresis to detect the product of strand transfer.
DNA ReplicationStrand displacement DNA synthesis by the T4
DNA replication system at 37 °C was carried out, and the products
were analyzed by alkaline-agarose gel electrophoresis, as described
(32). The primer-template was M13mp2 annealed to an 84-base
oligonucleotide in which only the 3 34-base are complementary, leaving
a 50-base tail (32).
0.8% agarose gel electrophoresis was used to detect the joint molecules or products of strand transfer. Electrophorsis was carried out in TAE (0.04 M Tris acetate, 0.001 M EDTA, pH 8.0) buffer at 0.6 V/cm for 15 h. After electrophoresis, the gel was stained in TAE buffer containing 0.5 µg/ml ethidium bromide, and DNA bands were illuminated by ultraviolet light and photographed.
In our previous report (15) we showed that T7 gene 2.5 protein
promotes the formation of joint molecules consisting of circular, single-stranded M13 DNA annealed to homologous linear duplex DNA bearing single-stranded termini. In the presence of T7 gene 4 helicase,
strand transfer proceeds at a rate of >120 nucleotides/s in a polar 5
to 3
direction with respect to the invading strand, resulting in the
production of circular duplex M13 DNA (15). Inasmuch as no strand
transfer occurred when T4 gene 32 protein was substituted for T7 gene
2.5 protein or T4 gene 41 helicase for T7 gene 4 helicase, it seemed
likely that specific protein interactions were important. This belief
was strengthened by the observation that gene 2.5-
21C protein, an
altered protein that does not physically interact with gene 4 protein,
and gene 4 protein together did not mediate strand transfer.
To
more precisely define the role of gene 2.5 protein in the T7
helicase-mediated strand transfer, we prepared joint molecules free of
gene 2.5 protein. With such a DNA substrate, the ability of gene 4 protein alone to catalyze strand transfer can be examined. As shown in
Fig. 2A, joint molecules were prepared by
incubating circular, single-stranded M13 DNA molecules and homologous
linear, duplex M13 DNA bearing single-stranded 3-termini with gene 2.5 protein. Gene 2.5 protein was then removed by extraction with phenol,
and the protein-free joint molecules were used as substrate for T7 gene
4 helicase.
When joint DNA molecules were incubated with gene 4 and 2.5 proteins, circular, duplex DNA molecules, the product of complete strand transfer, were observed (Fig. 2B, lane 3). As shown in Fig. 2B, lane 2, gene 4 protein alone mediates strand transfer equally well as the combination of gene 2.5 and 4 proteins. Thus, it appears that gene 2.5 protein is not required for either loading the gene 4 protein onto the proper strand or for the subsequent strand transfer reaction.
Effect of Other ssDNA-binding Proteins on T7 Helicase-mediated Strand TransferThe above finding that gene 2.5 protein is not
required for strand transfer implies strongly that the lack of strand
transfer previously observed (15) with the combination of gene 4 protein and mutant gene 2.5-21C protein reflects an inhibitory
property of this protein. To examine this point, increasing amounts of gene 2.5-
21C protein were added to a reaction containing joint molecules and gene 4 protein (Fig. 3A). As
the amount of the mutant gene 2.5 protein increases, there is a
progressive decrease in strand transfer as measured by the decrease in
circular, duplex DNA product. Since gene 2.5 protein binds to ssDNA
with a stoichiometry of 7 nucleotides bound/monomer of gene 2.5 protein
(25) the ssDNA in the joint molecule should be completely coated at a
concentration of 1 µM. At this concentration of gene
2.5-
21C protein, essentially no strand transfer occurs (Fig.
3A, lane 6). Gene 2.5-
21C binds to ssDNA with
the same affinity of wild-type protein but does not physically interact
with gene 4 protein (28), thus suggesting that the inhibition arises
from an inability of the gene 4 protein to displace the bound gene 2.5 protein in a step requiring a specific interaction.
In a similar manner, we have also examined the effect of the ssDNA-binding proteins of E. coli and phage T4 on gene 4 protein-mediated strand transfer (Fig. 3B). At a concentration of T4 gene 32 protein sufficient to coat all of the ssDNA, there is a complete inhibition of strand transfer (lane 3), whereas a similar concentration of E. coli SSB protein was without effect (lane 4). Presumably, the T7 gene 4 protein has evolved such that it can tolerate the relatively abundant host SSB protein present upon phage infection although the host protein cannot replace the gene 2.5 protein in DNA replication (3).
Strand Transfer Catalyzed by T4 Gene 32, 41, and 59 ProteinsThe UvsX protein, the bacteriophage T4 RecA analog, catalyzes strand exchange, albeit at a relatively slow rate (18-20). However, Salinas and Kodadek (24) recently demonstrated that the T4 gene 41 helicase can drive polar branch migration at a high rate in a multiprotein complex consisting of the products of the UvsX, UvsY, gene 32, gene 41, and gene 59 proteins. Thus, heteroduplex formation in vivo is must likely mediated by the T4 helicase system. In these studies they showed that joint molecules formed by the action of the UvsX and gene 32 proteins and then deproteinized were substrates for the gene 41 and 59 proteins but that strand transfer was inefficient.
We have used the same joint molecule described above for the T7
helicase-mediated strand transfer reaction to examine the basic
requirements for the actual strand transfer reaction and to examine the
protein specificity of the reaction. Like the T7 gene 2.5 protein,
stoichiometric amounts of the T4 gene 32 protein promote the formation
of stable joint molecules consisting of circular, single-stranded M13
DNA annealed to a complementary region of linear, duplex M13 DNA
bearing single-stranded 3 termini (Fig. 4A).
After formation, the preparation of joint molecules was deproteinized
and used as substrate for the T4 helicase-mediated strand transfer
reaction. As shown in Fig. 4B, lane 2, no strand transfer occurs in the presence of T4 gene 32 protein and gene 41 protein. However, if gene 59 protein is also present (lane 3), then strand transfer occurs, reflecting the requirement of gene 59 protein to load the helicase onto gene 32 protein-coated DNA
(32, 35-37). This result differs strikingly from that found with the
T7 helicase where the presence of its ssDNA-binding protein, gene 2.5 protein, does not impede loading of the T7 gene 4 protein, and hence a
loading factor such as the T4 gene 59 protein is not required.
Although gene 59 protein is clearly necessary for loading in the
presence of gene 32 protein, the question remains as to whether or not
gene 41 protein by itself, or with gene 59 protein, can drive strand
transfer in the absence of gene 32 protein. As shown in Fig.
4C, in the absence of gene 32 protein, the gene 41 protein is much less efficient in mediating strand exchange in the absence (lane 2) or presence (lane 3) of gene 59 protein.
Clearly, with the DNA substrate used here, T4 gene 41 protein requires
both gene 32 protein and gene 59 protein for it to mediate strand
transfer (lane 4). Strand transfer catalyzed by gene 32, 41, and 59 proteins requires nucleoside triphosphate hydrolysis, proceeds
from 5 to 3
with regard to the invading strand, and also takes place in physiological concentrations of salts (200-400 mM
potassium glutamate) (data not shown).
The above results show that efficient strand exchange
requires the presence of T4 gene 32, 41, and 59 proteins. A physical interaction among the three proteins appears necessary, since no strand
transfer occurred in the absence of gene 59 protein. To confirm that
there is a specific interaction among gene 32, 41, and 59 proteins, we
performed reactions with T7 gene 2.5 or 4 proteins replacing T4 gene 32 or 41 proteins, respectively. As shown in Fig.
5A, lane 1, T7 gene 4 protein
cannot replace T4 gene 41 protein. When gene 32 protein was replaced by
gene 2.5 protein, strand transfer was significantly inhibited as shown (Fig. 5B, lane 3). Strand transfer does not occur
in the presence of T7 gene 2.5 and T4 gene 41 proteins. The results
confirm that specific interactions among gene 32, 41, and 59 proteins
are essential for strand transfer.
Requirement for Gene 32 Protein for Strand Displacement Synthesis by the T4 Replication System
Previous studies (reviewed in Ref.
29) have shown that gene 32 protein is required for the slow rate of
strand displacement synthesis that occurs in the absence of the gene 41 helicase. If the helicase is present, there is some strand displacement synthesis from preformed fork templates in the absence of gene 32 protein (38, 39), but the ssDNA-binding protein markedly increases the
fraction of the template elongated at the high rate dependent on the
helicase. Gene 59 protein decreases the time needed to load the
helicase on the template (32, 36). In agreement with our demonstration
(Figs. 4 and 5) that gene 32 protein is required for efficient strand
transfer mediated by the gene 41 helicase and gene 59 helicase assembly
protein, we find that, if gene 32 protein is omitted, there is very
little strand displacement synthesis from a preformed fork template in
reactions with the 41 and 59 proteins (Fig. 6). In the
absence of gene 59 protein (reaction 1), there are shorter
strand displacement products elongating at the same rate as in
reactions without the helicase (not shown), and after 4 min, there are
longer products dependent on the helicase. Helicase dependent synthesis
is evident at the earliest time when both gene 59 and 32 proteins are
present (reaction 2), but is almost eliminated in the
absence of gene 32 protein (reaction 3). We conclude that
gene 32 protein facilitates gene 41-catalyzed unwinding of dsDNA during
both strand transfer and strand displacement DNA synthesis.
Effect of RNA Primer Synthesis on Strand Transfer
One
property that distinguishes the bacteriophage T7 helicase used in these
studies from that of the T4 helicase is the presence of a DNA primase
domain within the same polypeptide that constitutes the T7 helicase.
Gene 4 of phage T7 encodes two co-linear proteins, a 56-kDa protein
that has only helicase activity and a 63-kDa protein that has both
helicase and primase activity (30); the 63-kDa gene 4 protein is the
molecular weight species used in the current studies. The 56-kDa
protein also catalyzes strand transfer (15). The 63-kDa gene 4 protein
catalyzes the synthesis of tetraribonucleotides (pppACCC/A and pppACAC)
in a template-mediated reaction at specific pentanucleotide recognition
sites as it translocates 5 to 3
on DNA. In vivo (40) and
in vitro (30), these oligoribonucleotides are used as
primers for T7 DNA polymerase to initiate lagging strand DNA synthesis.
Debyser et al. (41) showed that primer synthesis on the
lagging strand decreased the rate of leading strand DNA synthesis,
presumably due to pausing of the 63-kDa gene 4 protein as it
synthesized a primer and hence reduced its effectiveness as a
helicase.
To see if concurrent primer synthesis by the T7 gene 4 protein affects
strand transfer in a manner similar to that observed with DNA synthesis
on a preformed replication fork, we have measured strand transfer in
the presence of ATP and CTP or ATP alone, the ribonucleotide precursors
for oligoribonucleotide synthesis. In addition, potassium glutamate was
included in the reaction, since it significantly stimulates the
synthesis of oligoribonucleotides by the 63-kDa gene 4 protein.3 As shown in Fig.
7, conditions for primer synthesis had no significant effect on either the extent (panel A) or the rate
(panel B) of reaction. In addition, the presence of T7 gene
2.5 protein that stimulates oligoribonucleotide synthesis by the gene 4 protein (30) had no effect either (Fig. 7A, lane
3). Analysis of the product of the primase reaction by 25%
polyacrylamide gel electrophoresis confirmed that dimer, trimer,
tetramer, and pentamer oligonucleotide were efficiently synthesized and
that the presence of gene 2.5 protein greatly stimulated synthesis of
these oligonucleotides in the reactions where ATP and CTP were present
(data not shown).
In a previous report (15) we reported that two proteins encoded by bacteriophage T7 work together to mediate DNA strand transfer. Stoichiometric amounts of gene 2.5 ssDNA-binding protein promoted the annealing of complementary DNA strands to form a joint molecule, the intermediate in DNA recombination. Subsequently, the T7 gene 4 helicase, in an energy-requiring reaction, mediated a polar exchange of one strand of the duplex portion of the joint molecule for a ssDNA partner. In bacteriophage T4, Salinas and Kodadek (24) showed that the gene 41 helicase of bacteriophage T4 was responsible for strand transfer in a multiprotein complex consisting of the products of the gene UvsX, UvsY, 32, 59, and 41 of the phage T4. Although these two studies demonstrated the novel role of DNA helicases in general recombination, several aspects of the overall reactions could not be compared directly due to differences in the formation and structure of the intermediate joint molecules. In the T7 system the formation of joint molecules between a single-stranded circle and a linear duplex with a single-stranded tail was accomplished by the gene 2.5 protein, a protein that enhances greatly the annealing of complementary DNA strands (15).4 In the T4 system, the UvsX, UvsY, and gene 32 proteins, which carry out a search for homology and may also stabilize the resulting joint molecule (reviewed in Ref. 16), were used to form joint molecules between a single-stranded circle and a fully duplex linear DNA (24). In the present study we have prepared joint molecules by incubating the appropriate complementary single-stranded circular and tailed duplex DNA molecules in the presence of either the T4 or T7 ssDNA-binding protein, and then isolated the joint molecules free of protein. With this DNA intermediate, we have addressed the relative roles of ssDNA-binding protein and helicase in the actual process of strand transfer with both the T4 and T7 proteins. Inasmuch as these same proteins are also involved in reactions at the replication fork, the information obtained is pertinent to the roles of these proteins in DNA replication.
We find that T7 gene 4 helicase mediates strand transfer within
preformed joint molecules equally well in the presence or absence of
gene 2.5 protein. The ability of T7 helicase to mediate strand transfer
in the absence of gene 2.5 protein demonstrates that the gene 4 protein
can load onto the strand to be displaced without accessory proteins
(Fig. 8). Since the gene 4 protein requires at least 17 nucleotides of ssDNA to which it can bind to initiate 5 to 3
translocation (26), it seems likely that this structure must arise as a
result of branch migration at the single-stranded duplex junction. This
result also necessitated a reexamination of our earlier observation
that a truncated form of gene 2.5 protein, gene 2.5-
21C protein,
could not substitute for the gene 2.5 protein in the overall strand
transfer reaction although it mediated the formation of joint molecules
as well as did the wild type gene 2.5 protein (15). Gene 2.5-
21C
protein, lacking 21 amino acid residues at its C terminus, binds to
ssDNA normally but cannot physically interact with itself to form
dimers or with the T7 DNA polymerase or gene 4 protein to stimulate
their activities (28). Rather than the physical interaction between gene 2.5 protein and gene 4 protein playing a positive role in enhancing helicase activity, our results suggested that the interaction may be necessary in order for the two proteins to exist together on
ssDNA. This interpretation appears to be correct, since we find that
gene 2.5-
21C protein is quite inhibitory to the T7 helicase-mediated
strand transfer. Therefore, we believe that the gene 2.5 protein bound
to ssDNA must physically interact, through its acidic C-terminal
region, with gene 4 protein in order for the helicase to displace it
and hence load onto ssDNA. In this regard, the T4 gene 32 protein also
has an acidic C-terminal region that mediates specific interaction with
many T4 proteins (42, 43). The fact that gene 32 protein is also
inhibitory to the T7 helicase-mediated reaction illustrates the
specificity of the C-terminal interaction. Interestingly, E. coli SSB protein has no effect on the T7 helicase-mediated
reaction, perhaps not a surprising result since T7 is faced with a
relatively large intracellular pool of this protein upon infection of
its host. In any case, the inability of gene 4 protein to load onto
ssDNA in the presence of T7 gene 2.5-
21C protein provides a
molecular basis for the essential nature of the C-terminal region of
the protein (28). However, the interaction of the C-terminal region of
the gene 2.5 protein with other T7 proteins may serve other functions,
since in the case of T7 DNA polymerase the T7 gene 2.5-
21C protein
does not stimulate the polymerase reaction as does wild-type protein,
but neither does it inhibit the reaction (28).
The replication system of bacteriophage T4 is more complex than that of bacteriophage T7, and the helicase-mediated recombination system proves no exception. In contrast to the results obtained with T7 gene 4 protein, T4 gene 41 helicase could not mediate strand transfer within a preformed joint molecule. This result is not unanticipated, however, since T4 gene 41 protein is known to require an accessory protein, T4 gene 59 protein, for efficient loading onto DNA (32, 35-37, 44, 45). In fact, gene 32 protein-bound DNA is practically inaccessible to gene 41 protein, necessitating the presence of gene 59 protein. Salinas and Kodadek (24) reported that T4 gene 41 protein and gene 59 protein together mediated strand transfer on deproteinized joint molecules, but the reaction was much less efficient than in the complete multiprotein complex with the T4 gene 32, UvsX, and UvsY proteins. In the present study, we observe little if any complete strand transfer by the two proteins alone, a difference that most probably reflects the fact that in our studies strand transfer must proceed for 7.1 kilobase pairs, whereas in the studies by Salinas and Kodadek (24) strand transfer was scored after only 2.2 kilobase pairs of transfer. However, when T4 gene 32 protein is present along with gene 59 protein and gene 41 helicase, rapid and extensive strand transfer occurs. Thus, while the T4 UvsX and UvsY proteins are required to form joint molecules if one of the DNA partners is totally duplex, the T4 gene 41 helicase can drive polar branch migration on joint molecules made without the UvsX and UvsY recombination proteins. Since gene 59 protein has been shown to bind to both the gene 32 and 41 proteins (35-37), it seems likely that gene 41 protein is loaded onto the joint molecule through interactions of the gene 59 protein with both the gene 32 protein bound to ssDNA and the gene 41 protein (Fig. 8). This interpretation is in agreement with that of Salinas and Kodadek (24), who showed that a mutant gene 32 protein, gene 32-A protein lacking its C-terminal domain, cannot substitute for gene 32 protein in the strand transfer reaction although it binds to ssDNA.
The requirement for T4 gene 32 protein during strand transfer and strand displacement synthesis is likely to be a consequence of its role in stabilizing a DNA structure needed to load the T4 gene 41 helicase onto joint molecules, and DNA replication forks. Both the T7 gene 4 protein and the T4 gene 41 protein are essential for replication, where they serve as the helicase for leading strand synthesis. The structure of the strand transfer intermediate with the helicase assembled on the displaced strand is not unlike that of a replication fork. Polymerization of nucleotides on the leading strand of the fork is replaced by the sequential annealing of complementary DNA sequence as the duplex region is processively unwound by the helicase. In both the T7 and T4 DNA replication systems, the helicase binds preferentially to a preformed fork to unwind the DNA ahead of the polymerase. T7 DNA polymerase and the gene 4 helicase together are sufficient to carry out this synthesis efficiently (46). In the T4 system, in contrast, omitting gene 32 protein greatly decreases the fraction of molecule copied at the high rate dependent on the gene 41 helicase (29, 38, 39), even when the helicase is loaded on the preformed fork by the gene 59 protein (Fig. 6). In these reactions, gene 32 protein appears to be affecting the initiation of unwinding rather than the rate of unwinding, because it increases the number of rapidly replicating molecules but does not change the rate at which the leading strand is elongated on DNA with the helicase. A similar conclusion has been reached by Tarumi and Yonesaki (44), who showed that gene 32 protein could stimulate the unwinding of a partial duplex by the gene 41 helicase if gene 59 protein was present to load the helicase on the gene 32 protein-covered DNA. In this view, the binding of gene 32 protein stabilizes a partially unwound DNA structure attractive to the helicase, and gene 59 protein is then essential to load the helicase onto the gene 32 protein-covered DNA. In the absence of the helicase, the binding of gene 32 protein allows strand displacement synthesis at a much slower rate by T4 DNA polymerase and its accessory proteins. The T4 ssDNA-binding protein is required for this reaction, even on templates containing a preformed fork (29).
The joint molecules used in the present study were constructed using the T7 gene 2.5 ssDNA-binding protein to promote the annealing of complementary ssDNA regions between the two partners. The T7 gene 2.5 protein is clearly involved in recombination, since mutations in gene 2.5 lower recombination frequencies (11). T7 gene 2.5 protein differs from the T4 gene 32 protein and E. coli SSB protein in that it is the most efficient in mediating homologous base pairing (25).4 In phage T7-infected E. coli cells, it is likely that the gene 2.5 protein is responsible for the annealing of homologous regions to form joint molecules, since recombination is not decreased in the absence of the host RecA pathway (48, 49), and thus far no RecA-type protein has been identified in T7 phage. On the other hand, it seems likely that the T4 UvsX protein plays a major role in homologous base pairing in that, in addition to facilitating the annealing of complementary strands, it can carry out a true search for homology.
The T7 gene 4 protein is unique in that one form of the gene 4 protein, the 63-kDa gene 4 protein, has full primase activity as well as helicase activity residing within the same polypeptide chain (30). Consequently, the T7 63-kDa gene 4 protein is capable of synthesizing oligoribonucleotides as it translocates along the displaced strand. We find, however, that the addition of ATP and CTP to strand transfer reactions containing the T7 gene 63-kDa gene 4 protein, enables primer synthesis but has no measurable effect on the rate or extent of strand transfer. Our inability to detect an effect of primer synthesis on strand transfer may be due to the limited number of primase recognition sites (eight major recognition sites) on the M13 DNA molecule. Nonetheless, in vivo, it would not be surprising to find that RNA-primed DNA synthesis does occur on the recombination intermediates. Such reactions may be important in the known role of recombination in the initiation of DNA replication in T4 phage-infected cells (11, 47).
We are grateful to T. Kusakabe for help in the examination of primer synthesis.