The Herpes Simplex Virus Type 1 Alkaline Nuclease and Single-Stranded DNA Binding Protein Mediate Strand Exchange In Vitro

Abstract

The replication of herpes simplex virus type 1 (HSV-1) DNA is associated with a high degree of homologous recombination. While cellular enzymes may take part in mediating this recombination, we present evidence for an HSV-1-encoded recombinase activity. HSV-1 alkaline nuclease, encoded by the UL12 gene, is a 5′→3′ exonuclease that shares homology with Redα, commonly known as λ exonuclease, an exonuclease required for homologous recombination by bacteriophage lambda. The HSV-1 single-stranded DNA binding protein ICP8 is an essential protein for HSV DNA replication and possesses single-stranded DNA annealing activities like the Redβ synaptase component of the phage lambda recombinase. Here we show that UL12 and ICP8 work together to effect strand exchange much like the Red system of lambda. Purified UL12 protein and ICP8 mediated the complete exchange between a 7.25-kb M13mp18 linear double-stranded DNA molecule and circular single-stranded M13 DNA, forming a gapped circle and a displaced strand as final products. The optimal conditions for strand exchange were 1 mM MgCl₂, 40 mM NaCl, and pH 7.5. Stoichiometric amounts of ICP8 were required, and strand exchange did not depend on the nature of the double-stranded end. Nuclease-defective UL12 could not support this reaction. These data suggest that diverse DNA viruses appear to utilize an evolutionarily conserved recombination mechanism.

Herpes simplex virus type 1 (HSV-1) is a double-stranded DNA (dsDNA) virus with a 152-kb linear genome. Replication of HSV-1 DNA takes place in the host nucleus. The first step of viral replication involves the circularization of the genome (13, 43). Shortly thereafter, replication intermediates appear as longer-than-unit-length head-to-tail concatemers (19) that have undergone genomic inversion (2, 27, 47, 58). The genome concatemers are not linear but rather consist of a mixture of complex structures such as Y- and X-shaped branches, replication bubbles, and tangled masses (20, 28, 48, 49). The presence of these structures and the inversion of the L and S genome segments suggest that recombination plays a role in the replication of HSV-1 DNA. In fact, high levels of recombination are known to accompany HSV infection (3, 10, 11, 46, 54). While cellular recombinases may be involved in mediating some of these processes (57), the possibility exists that HSV-1 encodes recombinases that can also participate.

Herpes simplex virus type 1 (HSV-1) encodes a 5′-to-3′ exonuclease (17, 23, 32, 52) termed alkaline nuclease, the product of the UL12 open reading frame (29). Recently, computer database searches have revealed that the HSV-1 UL12 gene shares homology with bacteriophage lambda Redα, commonly known as λ exonuclease (1, 36). The Redα protein is a 5′→3′ exonuclease which is part of the λ Red recombinase previously shown to be required for recombination by bacteriophage lambda in a RecA⁻ host (8, 9, 50). Redα operates in conjunction with a single-stranded DNA (ssDNA) binding protein, lambda Redβ, which promotes ssDNA annealing (34). The lambda Red recombinase is functionally similar to Escherichia coli RecE/RecT (24). These proteins are a paradigm for a class of recombinases that employ a strand-annealing protein and an exonuclease and do not require a high-energy cofactor. The model for recombination mediated by these proteins proposes that the exonuclease degrades DNA from a double-stranded end in the 5′→3′ direction, exposing a 3′ single-stranded tail. This tail is bound by the ssDNA binding protein, which assembles a nucleoprotein filament that mediates annealing to a complementary ssDNA sequence (24). In the presence of recA, the lambda Red system can also participate in strand invasion reactions (50). Thus, lambda phage apparently uses a complex interaction between virus and host recombinases to carry out replication and recombination. The homology of UL12 to lambda Redα suggests that UL12 itself could be part of a recombinase.

Genetic and biochemical evidence point to a role for HSV-1 UL12 in the replication and processing of HSV-1 DNA, but its precise role is still unclear. While UL12 is not an essential gene, it is needed for efficient production of viral progeny (56). Near normal amounts of viral DNA are produced in UL12 mutant virus-infected cells, but this DNA exists in a different form, one that is more susceptible to breakage than wild-type HSV-1 DNA. The DNA is packaged into capsids that fail to exit the nuclei of infected cells, leading to low viral yields (28). Using point mutations in UL12 that eliminate its exonuclease activity, the in vivo function of UL12 was shown to depend upon its exonuclease activity (14). Interestingly, evidence for genomic inversion was observed in cells infected with the UL12-null mutant (28). This may be due to the activity of host cell recombination enzymes. The aberrant structure of replicating DNA which accumulates in cells infected with the UL12 mutant indicates, however, that proper DNA replication and processing is dependent on the viral enzyme. We propose that, like bacteriophage lambda, HSV may rely on a complex interaction between viral and host recombination systems to carry out DNA replication and recombination.

Other similarities between lambda and HSV-1 also support the possibility that HSV-1 UL12 could function as a recombinase in a manner analogous to that of lambda Redα/β and RecE/T. The UL12 gene product, like Redα and RecE, is a 5′→3′ exonuclease (17, 23, 32, 52). Analogous to the interaction of lambda Redα and the ssDNA binding protein (SSB) lambda Redβ, UL12 interacts with the SSB of HSV-1, ICP8 (53, 55). Furthermore, ICP8 possesses strand-melting (4) and strand-annealing activities (12) and has been reported to mediate limited strand exchange (6, 38). A recent study has also shown that ICP8 can promote strand invasion (37).

In this paper we demonstrate that UL12 and ICP8 mediate strand exchange in vitro by using an agarose gel-based strand exchange assay. Both UL12 and ICP8 were required in order to detect joint molecule and gapped circle formation in this assay. Strand exchange was most efficient under conditions that were suboptimal for UL12 nuclease activity, but a mutant form of UL12 that was devoid of exonuclease activity could not support this reaction. The fact that UL12 shares homology with a component of a class of recombinases argues strongly for the functional relevance of this complex in HSV-1 biology. The similarity of these activities with those of the lambda Red system shows that mechanisms of recombination from diverse DNA viruses appear to be evolutionarily conserved.

MATERIALS AND METHODS

Materials.

[γ-³²P]ATP was from Dupont or Amersham. [Thymidine-methyl-³H]DNA (E. coli) was from Dupont. Low-melting-point agarose (NuSieveGTG) was from BioWhittaker Molecular Applications. All other materials were reagent grade.

DNA.

M13mp18 replicative form was purified from infected E. coli UT481 [Δ(lac-pro)hsdS(r⁻m⁻)lacI^q lacZ] cells by using the Qiagen maxi plasmid kit. M13mp18 ssDNA was from New England Biolabs or was purified from M13 phage-infected UT481 cells according to standard protocols (44). DNA fragments were purified from agarose gels with the GeneClean Spin kit (Bio-101).

Enzymes and proteins.

Restriction endonucleases were from New England Biolabs. T4 polynucleotide kinase and T4 DNA ligase were from Life Technologies. Proteinase K was from Roche. Protein concentrations were determined by the Bradford method (7).

The UL12 and UL12_D340E proteins used in this paper were purified as described previously and were estimated to be greater than 95% pure (14). The UL12 protein has an activity of 0.2 ng of DNA degraded/min/ng of protein, assayed as the release of acid-soluble counts from an E. coli [³H]DNA substrate under conditions optimal for UL12 nuclease activity (14). The nuclease activity of UL12 at standard strand exchange assay conditions (see below) was found to be 0.06 ng of DNA degraded/min/ng of protein. The UL12_D340E mutant protein is devoid of exonuclease activity (14).

ICP8 was purified from Spodoptera frugiperda (Sf21) cells infected with recombinant baculovirus AcUL29 (51). Cells were collected 3 days after infection, pelleted, quick-frozen, and stored at −80°C. Three grams of frozen cells (wet weight) was resuspended in 30 ml of swelling buffer (10 mM Tris · Cl [pH 7.5], 10 mM KCl, 1.5 mM MgCl₂) with 200 μl of Sigma protease inhibitor cocktail. Cells were incubated on ice for 30 min and then homogenized in a Dounce homogenizer. Nuclei were pelleted and resuspended in 20 ml of extraction buffer (swelling buffer with 1.2 M NaCl and protease inhibitors) to extract nuclear proteins. After a 40-min incubation on ice, the mixture was centrifuged for 40 min at 30,000 rpm at 4°C in a Beckmann Ti70 rotor. The supernatant was dialyzed against 1.5 liters of buffer A (20 mM HEPES [pH 7.5], 10% [wt/vol] glycerol, 0.1 mM NaCl, 0.1 mM EDTA, 1 mM dithiothreitol [DTT]) for 16 h at 4°C. During dialysis, a white precipitate formed which was removed by centrifugation and by filtration through a Millex-HV low-protein-binding 0.45-μm-pore-size syringe filter (Millipore). The cleared extract was loaded onto an SP Sepharose HiLoad 16/10 column (Pharmacia) with buffer A and washed with 3 column volumes of buffer A. The protein was eluted by using a linear gradient from 0.1 to 1 M NaCl over 50 ml. ICP8 eluted at 0.3 to 0.4 M NaCl. The ICP8 peak fractions were pooled and dialyzed against 1.5 liters of 20 mM HEPES (pH 7.5), 10% (wt/vol) glycerol, 0.5 mM EDTA, and 0.5 mM dithiothreitol. The protein concentration was determined by the Bradford method and by UV absorbance at 280 nm (extinction coefficient, 82,720 M⁻¹cm⁻¹) (5). The protein concentration was 1.8 mg/ml (total yield, 12.6 mg) by both methods. The purity of the protein as estimated by Coomassie brilliant blue-stained gels was 95%. A nuclease assay (described below) was used to determine whether the purified ICP8 had any nuclease contaminants. The specific nuclease activity was found to be 8.6 × 10⁻⁶ ng of DNA released/min/ng of protein. Thus, the nuclease contamination of the ICP8 preparation was negligible.

Labeled double-stranded substrate for strand exchange assay.

M13mp18 replicative form DNA was digested with BsrGI and gel purified. The purified DNA fragment was end labeled by the exchange reaction by using T4 polynucleotide kinase, [γ-³²P]ATP, and the exchange reaction buffer supplied by the manufacturer. The labeled fragment was then religated with T4 DNA ligase and then cleaved by PstI. The labeled 7.25-kb fragment (full-length M13) was gel purified.

Strand exchange assay.

The reaction was carried out in a final volume of 20 μl and consisted of 100 ng of circular M13mp18 ssDNA (2 nM), 100 ng of linear ³²P-labeled double-stranded M13mp18 (1 nM), 18.8 ng of UL12 (13.9 nM), 4.5 μg of ICP8 (1.75 μM), 20 mM Tris · Cl (pH 7.5), 40 mM NaCl, 1 mM MgCl₂, and 1 mM dithiothreitol, or as indicated in figure legends. The reaction mixture was incubated at 37°C for the times indicated in the figure legends and stopped by adding 5 μl of 5× stop buffer (50% glycerol, 50 mM EDTA, 1% sodium dodecyl sulfate [SDS], 0.2% bromphenol blue). Samples were electrophoresed on a 1% agarose gel with 0.7 μg of ethidium bromide/ml with TAE buffer (0.04 M Tris-acetate, 0.001 M EDTA). Gels were dried and exposed to phosphorimager screens (National Diagnostics). The ImageQuant, version 5.0, software package was used for quantification of the results. Adobe Photoshop (version 6.0) and Adobe Illustrator (version 7.0) were used in the preparation of figures.

Southern blots.

Strand exchange assays were performed and loaded onto 1% agarose gels as described above. Following electrophoresis, the DNA was blotted onto GeneScreen Plus membranes (Dupont) according to the manufacturer's suggested protocols. The oligonucleotide probes used to detect the M13 DNA strands were end labeled with T4 polynucleotide kinase and [γ-³²P]ATP with the forward reaction buffer supplied by the manufacturer (Life Technologies). The sequences of the two probes are as follows: 5′-GTCGGTGACGGTGATAATTCACCTTTAATG, for detection of the pairing, or minus, strand, and 5′-CATTAAAGGTGAATTATCACCGTCACCGAC, for detection of the displaced, or plus, strand.

Nuclease assay.

Total unlabeled chromosomal DNA from E. coli was isolated from late-log-phase UT481 cells by phenol extraction and ethanol precipitation essentially as described previously (14). [Thymidine-methyl-³H]DNA (E.coli) was mixed with unlabeled chromosomal E. coli DNA to provide a substrate with the desired specific radioactivity. The nuclease assay was performed in a 50-μl volume, with 250 ng of [³H]DNA as the substrate. UL12 (47 ng, 13.9 nM) and ICP8 (11.25 μg, 1.75 μM) were assayed for nuclease by using the same concentrations of these proteins and the same assay buffer as were used in the strand exchange assay. Reaction mixtures were incubated for 10 min at 37°C and then stopped with 150 μl of 0.5% yeast RNA and 200 μl of 20% (wt/vol) trichloroacetic acid. After 10 min on ice, samples were centrifuged for 10 min at 14,000 × g, and the radioactivity in 200 μl of the supernatant fraction was determined by scintillation counting. Results presented are averages of duplicate determinations.

RESULTS

HSV-1 alkaline nuclease (UL12) and ssDNA binding protein (ICP8) promote strand exchange.

An agarose gel-based strand exchange assay was used to test the ability of UL12 and ICP8 to promote strand exchange in vitro between linear dsDNA and circular ssDNA of M13mp18. The dsDNA substrate was internally labeled at a single site on both strands of the molecule. This 7.25-kb fragment had ³²P-labeled nucleotides positioned 5.25 kb from the 5′ end of the pairing strand and 2 kb from the 5′ end of the strand that would be displaced or degraded during strand exchange (Fig. 1, top line). Since UL12 is a 5′→3′ exonuclease, it was important to position the label at a distance from the 5′ end. The internal location of the label also prevented its loss to any potential 3′→5′ exonuclease contaminants in the protein preparations. In addition, since each strand had only one labeled nucleotide, this simplified quantification of the products of the reaction.

FIG. 1.

Model for strand exchange by UL12 and ICP8. A schematic representation of the strand exchange reaction is presented. Sigma, alpha, and gapped circle forms represent strand exchange products at different stages of the reaction. The asterisk marks the internal ³²P label, which is 2 kb from the 5′ end of the displaced strand and 5.25 kb from the 5′ end of the pairing strand.

The strand exchange catalyzed by UL12 and ICP8 is shown in Fig. 2. Upon incubation of UL12 and ICP8 with the DNA substrates, more-slowly migrating DNA molecules were formed (Fig. 2, lanes 5 to 12). The appearance of these more-slowly migrating forms is reminiscent of the joint molecules observed by Hall and Kolodner upon incubation of the E. coli RecE and RecT proteins with similar substrates (16). The conclusion that these more-slowly migrating forms seen in the presence of RecE/T represent homologous pairing and strand exchange was confirmed by electron microscopy. The simplest explanation for the more-slowly migrating forms seen in Fig. 2 is that they represent joint molecules which result from a strand exchange reaction. Little or no slowly migrating forms were observed when dsDNA was incubated alone with UL12 and ICP8 (data not shown). Incubation of ICP8 alone with both DNA substrates did not lead to the formation of joint molecules, even after an 80-min incubation (Fig. 2, lane 2). Similarly, joint molecules were not produced upon incubation with UL12. Incubation with UL12 merely led to the gradual degradation of the DNA substrates (Fig. 2, lane 3, and data not shown). The reaction products demonstrated the same mobility whether they were loaded onto the gel with a buffer containing SDS (as shown in Fig. 2 to 7) or whether they were treated with proteinase K prior to loading (data not shown). Therefore, the slowly migrating products were not the result of protein-DNA complexes with retarded mobility on agarose gels. When a loading buffer without SDS was used, most of the DNA failed to exit the well. This indicates that protein-DNA complexes formed during the assay are disrupted by the SDS loading buffer.

FIG. 2.

Time course of joint molecule formation catalyzed by UL12 and ICP8. Strand exchange reactions were carried out using the ³²P-labeled linear M13 dsDNA and unlabeled circular M13 ssDNA substrates as described in Materials and Methods. Left panel: phosphorimager image of dried gel. Right panel: photograph of ethidium bromide-stained gel. Incubations were at 37°C for the times indicated. Lane 1, control reaction, no proteins added; lane 2, strand exchange with ICP8 alone; lane 3, strand exchange with UL12 alone; lanes 4 to 12, strand exchange by UL12 and ICP8. jm, joint molecules; ds, linear M13 dsDNA; ss, circular M13 ssDNA.

FIG. 7.

Strand exchange by UL12, UL12_D340E, and ICP8. A photograph of the ethidium bromide-stained gel is shown. Strand exchange was carried out as described in Materials and Methods, with 4.5 μg of ICP8, 18.8 ng of UL12, and 20 ng of UL12_D340E, as indicated. Reaction mixtures were incubated at 37°C for the times indicated.

The progression of the strand exchange reaction over time is shown in Fig. 2. During the reaction, two species of reaction products are produced: a slowly migrating species consisting of joint molecules and a rapidly migrating species that will be discussed below. The joint molecules produced are a heterogeneous population, probably including sigma, alpha, and possibly gapped circular forms, as shown in Fig. 1. The early joint molecule products migrate more slowly than those from later time points, reflecting the higher percentage of sigma forms with long tails assumed to be present at this stage. The expected products of a complete strand exchange reaction would be a gapped circle and a displaced linear single strand. In order to determine whether the expected gapped circles were produced, we compared the migration of the strand exchange products to the migration of known DNA controls. Nicked circular, linear double-stranded, and linear single-stranded forms are produced when the linear dsDNA and circular ssDNA substrates are boiled and allowed to reanneal slowly (Fig. 3, lane 7). The joint molecules produced after 40 min of incubation with ICP8 and UL12 (Fig. 3, lane 3) migrate slightly faster than the nicked circle seen in Fig. 3, lane 7, suggesting that these joint molecules may in fact be gapped circles and that some strand exchange reactions may have gone to completion.

FIG. 3.

Analysis of strand exchange products. Both panels represent phosphorimager images of dried gels. Left panel, strand exchange was performed as described in Materials and Methods with ³²P-labeled dsDNA and unlabeled ssDNA substrates. Lane 1, control reaction, no proteins added, 40-min incubation; lanes 2 and 3, strand exchange with ICP8 and UL12, 20- and 40-min time points, respectively. Lanes 4 to 7 represent various DNA-only controls. DNAs (double stranded only in lanes 4 to 5, both double stranded and single stranded in lanes 6 to 7) were boiled for 2 min in strand exchange buffer and either quickly cooled on ice (lanes 4 and 6) or slowly cooled to allow strands to reanneal (lanes 5 and 7). Right panel, strand exchange reactions were performed as in the left panel (lanes 1 to 3) but were electrophoresed on a 1% low-melting-point agarose gel. Three gel slices were cut from each lane: A, containing joint molecules; B, containing remaining double-stranded substrate; and C, containing low-molecular-weight products. The positions of A, B, and C gel slices are indicated on the left panel. The gel slices were melted at 65°C and divided into two portions. One portion was loaded directly into the well of a second 1% agarose gel (shown in right panel). The other portion was boiled for 2 min prior to loading. Electrophoresis was performed as for the strand exchange assay. jm, joint molecules; nc, nicked circle; ds, dsDNA; ss, ssDNA.

The second expected product of a complete strand exchange reaction is a displaced single strand. The displaced single strand should migrate rapidly, and therefore, we investigated the rapidly migrating strand exchange products to see if such a displaced strand could be detected. At early time points (Fig. 2, lanes 5 to 7, and Fig. 3, lane 2), the rapidly migrating band is found at a position between the original double-stranded substrate and the ssDNA. Therefore, it appears to represent linear dsDNA that has been shortened by the UL12 nuclease. At later time points, the rapidly migrating species is found near or slightly below the position of the circular ssDNA (Fig. 2, lanes 8 to 12, and Fig. 3, lane 3). This band could represent dsDNA that was shortened further or ssDNA that was displaced during strand exchange. In order to investigate these possibilities, the samples were run on a low-melting-point agarose gel and bands containing the dsDNA substrate and products of the strand exchange reaction were isolated. This was done to determine whether the products of the strand exchange reaction, particularly the putative displaced single strand, were double or single stranded. The gel slices containing the individual species were melted, and a portion of each sample was boiled to separate the DNA strands, as described in the legend to Fig. 3. These samples were loaded on a second agarose gel. This procedure was effective at separating the strands, as shown by the migration of the dsDNA substrate (1B) with and without boiling (Fig. 3, lanes 8 to 9). The DNA strands of the various reaction products were also separated by boiling. Boiling of the joint molecule products (2A and 3A) released the strands that were paired with the circular ssDNA (Fig. 3, lanes 11 and 17). We can see that the pairing strands are shortened over time (Fig. 3, compare lanes 11 to 17), most likely due to UL12 digestion. Interestingly, when the joint molecule products (Fig. 3, lanes 10 and 16) were run on the gel, a portion of the DNA did not exit the well. We presume that this is because when the gel slices were melted at 65°C, the DNA strands of the joint molecules were able to breathe, allowing for reannealing between joint molecule species and creating a complex which was unable to enter the gel. This complex was disrupted by boiling, as no DNA is evident in the wells when the joint molecules were boiled (Fig. 3, lanes 11 and 17). This experiment also demonstrates that the rapidly migrating species (2C) from the 20-min reaction is double-stranded, as the migration pattern changes after boiling (Fig. 3, lanes 14 to 15). This fragment is most likely made up of double-stranded substrate molecules that were shortened by UL12. In contrast, the rapidly migrating species (3C) from the 40-min reaction appears to be single stranded, since its migration is unchanged by boiling (Fig. 3, lanes 20 to 21). We thus believe that this rapidly migrating species is the displaced single strand. This result suggests that after 40 min of incubation, some of the strand exchange reactions have gone to completion, forming a gapped circular molecule and a displaced single strand.

To further validate the identity of the displaced strand, we analyzed the strand exchange reaction by using a Southern blot (Fig. 4). Reactions were performed with unlabeled substrates, and duplicate samples originating from the same tubes (with the exception of lanes 7 to 8) were loaded on a single agarose gel. The two halves of the membrane were probed with ³²P-end-labeled oligonucleotide probes, corresponding to nucleotides 2616 to 2645 of M13mp18, a position that is equidistant from the two ends of the PstI-cut M13 dsDNA. The left side of the membrane (Fig. 4, lanes 1 to 5) was probed with the oligonucleotide probe recognizing the pairing (minus) strand, while the right side (Fig. 4, lanes 6 to 12) was probed with the oligonucleotide recognizing the M13 circular ssDNA and the displaced strand (plus strand). Both probes hybridized with the slowly migrating strand exchange products. The rapidly migrating species seen after 40 min of incubation was only recognized by the displaced strand probe (Fig. 4). Therefore, this species is not likely to be a dsDNA fragment shortened by UL12 action because then both probes should have recognized it. Furthermore, the putative displaced strand does not appear to be a degraded ssDNA substrate because similar bands are not seen when UL12 and ICP8 are incubated with the single-stranded substrate alone (Fig. 4, lanes 7 to 8). The displaced strand is less intense at the 50- and 60-min time points (Fig. 4, lanes 11 to 12), presumably due to degradation by UL12. Taken together, these data indicate that UL12 and ICP8 mediate a true strand exchange reaction, one that includes both annealing and displacement.

FIG. 4.

Southern blot of strand exchange reaction. Unlabeled dsDNA and ssDNA were used in strand exchange reactions as described in Materials and Methods and incubated for the times indicated. The gel was loaded with duplicate samples originating from the same reaction tubes (except for lanes 7 to 8), and the DNA was blotted onto a GeneScreen Plus membrane according to the manufacturer's protocol. The membrane was cut into two, and each half was hybridized with its respective oligonucleotide probe. jm, joint molecules. The arrow marks the position of the putative displaced strand.

Reaction conditions for strand exchange and UL12 nuclease activity.

Strand exchange by UL12/ICP8 and UL12 nuclease activity were assayed under different conditions of pH and Na⁺ and Mg²⁺ concentration, and the results are presented in Fig. 5. The extent of strand exchange was determined by measuring the amount of DNA that had migrated as a high-molecular-weight species at 20 min of incubation. We chose this time point because at 20 min generally little or no radioactivity had been lost due to nuclease activity, thereby simplifying comparisons. The results show a correlation of high strand exchange activity with moderate UL12 nuclease activity. The UL12 nuclease, also known as the alkaline nuclease, exhibits a pH optimum of 9 to 10 (17, 32, 52). Strand exchange activity was highest at pH 7.5 to 8.5, a pH range that supported an intermediate amount of UL12 nuclease activity. At pH levels below 7.5, both the nuclease and strand exchange activities were abolished. This was not a buffer effect, as nuclease activity and strand exchange were the same whether HEPES or Tris buffers were used at pH 7.5 (data not shown). At high NaCl concentrations, both UL12 nuclease activity and strand exchange were inhibited. In contrast, although UL12 nuclease was active in the absence of NaCl, significant strand exchange was not observed in the absence of NaCl. UL12 nuclease activity is optimal at higher magnesium concentrations. However, at these concentrations, more of the double-stranded substrate was degraded than exchanged (Fig. 4 and data not shown). In the absence of magnesium, UL12 nuclease was inactive and strand exchange did not occur. Optimal strand exchange was achieved at low magnesium concentrations, which allowed for a moderate amount of UL12 nuclease activity. Low magnesium concentrations also promote branch migration, which would be expected to potentiate strand exchange (41). The optimal conditions for the reaction were intermediate between the optima for strand melting and strand annealing by ICP8. For true strand exchange to occur, ICP8 must be able to mediate both strand melting and strand annealing. ICP8's annealing activity is optimal at 6 mM MgCl₂ and 80 mM NaCl and is significantly reduced at low levels of MgCl₂ and NaCl (12). In contrast, ICP8's helix-destabilizing activity is highest at low MgCl₂ levels (0 to 1 mM) and in the total absence of salt (4). Both strand-melting and strand-annealing activities of ICP8 are functional at pH 7 to 9 (4, 12). Thus, the conditions found to be optimal for strand exchange were those that allowed for moderate levels of all three activities involved in this reaction: nuclease, strand melting, and strand annealing.

FIG. 5.

Strand exchange and UL12 nuclease activity at different conditions of [Mg²⁺], [Na⁺], and pH. Strand exchange and nuclease assays were performed as described in Materials and Methods under the various conditions shown. Open circles, strand exchange; closed squares, nuclease activity. Unless indicated otherwise, the conditions were pH 7.5, 1 mM MgCl₂, and 40 mM NaCl. Strand exchange assay mixtures were incubated for 20 min. Percent strand exchange was calculated as the percentage of radioactivity in joint molecule products out of the total radioactivity in the lane. The buffers used were HEPES-NaOH, pH 6.5 and 7.0; Tris-Cl, pH 7.5, 8.0, and 8.5; and glycine-NaOH, pH 9.0 and 9.5. The nuclease activity of UL12 was assayed by using the [³H]DNA E. coli substrate and is represented as the amount of DNA (in nanograms) digested by 47 ng of UL12 (13.9 nM) in a 10-min assay at 37°C.

In other experiments, we tested the ability of either manganese or zinc to replace magnesium in the strand exchange assay. Manganese could support some strand exchange, but zinc could not (data not shown). This is consistent with previous reports on the effect of divalent cations on UL12 activity (17, 32). In addition, ATP was not required for strand exchange nor did it enhance the activity (data not shown).

The double-stranded fragment in our standard strand exchange assay was cut with PstI, which leaves 4-base 3′ overhangs. In order to test whether the nature of the double-stranded end was important in our reaction, we used double-stranded substrates that had been cut with SmaI, which leaves blunt ends, and BamHI, which leaves 4-base 5′ overhangs. Since all three sites are clustered in the polylinker region of M13mp18, any differences between them should be attributable to the nature of the end and not to local sequence context. No differences were noted in the ability of UL12/ICP8 to promote strand exchange using substrates with the different types of overhanging ends (data not shown). In order to make the assay more sensitive, a competition experiment was used. The strand exchange assay was performed with a mixture of ³²P-labeled PstI-cut dsDNA (50 ng) and unlabeled dsDNA (125 ng) cut with either PstI, BamHI, or SmaI. If the new substrates are utilized either more or less efficiently than the PstI-cut substrate, a change in the amount of labeled PstI substrate that undergoes strand exchange would be expected. Neither of the new substrates caused a change, suggesting that the three substrates were utilized equally (data not shown).

For complete coverage of the 100 ng of M13 ssDNA used in the strand exchange assay, 3.7 μg of ICP8 are required (15, 40). Our assays were done with an amount slightly in excess (4.5 μg) of this minimal amount. When increasing amounts of ssDNA were added to the strand exchange assay, such that the amount of ICP8 was insufficient for full coverage, strand exchange was reduced accordingly (Fig. 6, lanes 6 to 8). When the amount of ICP8 was increased to correlate with the increases in ssDNA (Fig. 6, lanes 9 to 14), strand exchange was restored. Therefore, efficient strand exchange appeared to require stoichiometric amounts of ICP8. This experiment also demonstrates that moderate excess of ICP8 does not inhibit strand exchange (Fig. 6, lane 3).

FIG. 6.

Titration of ssDNA and ICP8 in the strand exchange assay. Strand exchange reactions were performed as described in Materials and Methods, with 20-min incubations. UL12 and dsDNA were added according to standard conditions while the amounts of ssDNA and ICP8 used are indicated on the figure. A photograph of the ethidium bromide-stained gel is shown. Lanes 1 to 2 are no-protein controls. jm, joint molecules; ds, dsDNA; ss, ssDNA.

Active UL12 is required for strand exchange.

A mutant UL12 protein, UL12 _D340E, was previously purified and characterized in our laboratory (14). This protein has a single point mutation that has eliminated its exonuclease activity. When included in the strand exchange assay, this protein was unable to promote strand exchange (Fig. 7, lanes 5 to 7), both at the standard UL12 concentration (Fig. 7) and when present at 10 times the standard concentration (data not shown). This deficiency cannot be attributable to a global effect of the mutation on UL12, as excess mutant protein still retained the ability inhibit wild-type UL12, both in the strand exchange assay and in the nuclease assay (data not shown). The mechanism of inhibition is probably through competition for the DNA substrate, as the inhibition of the nuclease activity was seen only at low substrate DNA concentrations (data not shown).

DISCUSSION

The data presented in this report demonstrate that the HSV-1 alkaline nuclease (UL12) and ssDNA binding protein (ICP8) work together to carry out a strand exchange reaction in vitro. This reaction occurs under conditions of low UL12 nuclease activity and requires both proteins for strand exchange to occur. These results differ from those of previous studies (6, 38), which found that ICP8 alone could effect a measure of strand exchange. This difference could be due to the fact that different substrates, methods of detection, and assay conditions were used. Our assay employed a full-length M13 fragment, whereas shorter DNA fragments (2 kb or less) were used by the other investigators. If short regions of DNA were annealed by ICP8 alone in our assay, these small stretches might not have been enough to stabilize the joint molecule, which would have a long tail due to the use of the 7.25-kb M13 fragment, and would therefore have gone undetected. In the study by Nimonkar et al., double-stranded substrates with 9- and 12-base overhangs were used to allow for ICP8 loading. The assay was performed in two stages, with a different concentration of Mg²⁺ at each stage, in order to allow for both strand melting and strand annealing by ICP8. The low level of joint molecule formation produced by ICP8 alone in this assay was potentiated by the addition of HSV-1 helicase-primase complex (38). In the study by Bortner et al., electron microscopy was used to detect joint molecule formation. This method may have enabled the detection of joint molecules with limited paired regions that could not be visualized by the gel assay. Another possible explanation for the apparent discrepancy is that the ICP8 preparation used in the Bortner study may have contained some UL12 nuclease. In that study, ICP8 was prepared from cells that were infected with HSV-1. Since very little UL12 is required for strand exchange to occur (our standard conditions use a 125:1 molar ratio of ICP8 to UL12), a trace amount of UL12 could have been sufficient to promote the reaction.

The need for active UL12 is consistent with data from similar strand exchange protein pairs, notably RecE/T and Redα/β. Although some joint molecules could be produced by incubation of RecT alone with a preresected substrate (16), another study with an in vivo recombination assay found that recombination was obtained only when the respective partners (RecE/T and Redα/β) were expressed together. Mixing and matching RecE with Redβ or Redα with RecT did not lead to strand exchange (35). This result implies that specific protein-protein interactions are involved in recombination of this type. Consistent with this model, ICP8 interacts with UL12 (53, 55). Interestingly, Mikhailov et al. have reported that the baculovirus alkaline nuclease associates with the DNA binding protein LEF-3, suggesting that baculovirus may also encode a two-subunit recombinase (30).

The activities shown here for UL12/ICP8 suggest that it may be a member of the family of two-component viral recombinases comprised of an alkaline exonuclease and an associated ssDNA annealing protein. This family of proteins can mediate strand exchange in the absence of a high-energy cofactor. Several other proteins found in different organisms share the ability to promote strand exchange in the absence of a nucleotide cofactor and may be related to this family. The eukaryotic proteins Sep-1 and Rrp-1 (21, 45) embody both strand exchange and nuclease activities in a single polypeptide. The human HPP-1 protein mediates strand exchange but appears to function without a cognate exonuclease (31). Rad52 is also able to promote strand annealing and shares structural and functional similarities with RecT and Redβ (22, 42). These proteins are highly effective at mediating single-strand annealing, but some are also capable of strand invasion (26, 39). It is possible that UL12/ICP8 also mediates strand invasion in conjunction with other viral or host proteins. This widens the possibilities for UL12/ICP8 involvement in recombination associated with HSV-1 replication.

Recombination occurring during replication of HSV-1 DNA could be of several types. A strand-annealing mechanism could be used by the virus to generate genomic concatemers. Since the HSV-1 genome has direct repeats at its ends, concatemerization through single-strand annealing could proceed through a mechanism similar to that used by bacteriophage lambda (50). UL12/ICP8 could potentially be the mediator of such a mechanism. Another intriguing possibility is that strand invasion could be used by HSV-1 to prime DNA replication. A recent study has shown that ICP8 can mediate strand invasion of a supercoiled substrate by a homologous single-stranded oligonucleotide (37). Alternatively, because the HSV-1 DNA contains numerous gaps (18) (N. B. Reuven and S. K. Weller, unpublished results), it is possible that DNA replication could be primed by an invading strand annealing to a single-stranded gapped region without requiring true invasion. The importance of recombination-dependent replication has been recognized as playing a critical role in both prokaryotes and eukaryotes (25). Replication of T4 phage DNA is a paradigm for this type of replication (33), which involves the use of the 3′ end of an invading strand as a primer for DNA replication on the new template. This mechanism leads to branched intermediates, consistent with what is found in replicating HSV-1 DNA. UL12/ICP8 might also take part in recombination of this type. The HSV-1 helicase-primase complex has also been shown to participate with ICP8 in mediating a strand exchange reaction that uses resected substrates (38). The ability of the HSV-1 proteins UL12, ICP8, and the helicase-primase to participate in strand transfer reactions is interesting in light of the tight linkage between DNA replication and recombination in this virus. Furthermore, these results identify HSV-1 as an ideal model system for investigating the interface between these processes in eukaryotes.