The Major Herpes Simplex Virus Type-1 DNA-binding Protein Is a Zinc Metalloprotein”

The primary amino acid sequence of the major herpes simplex virus type 1 (HSV-1)-infected cell polypeptide 8 (ICPS) deduced from the DNA sequence of the unique long open reading frame 29 (UL29 ORF) contains a potential metal-binding domain of the form cys-X~-~- C ~ S - X ~ - ~ ~ - A - X ~ - ~ - A where A may be either histidine or cysteine and X is any amino acid. The putative metal- binding sequence in ICPB encompasses residues 499-Atomic absorption analysis of several preparations of ICPB indicates the presence of 1 mol of zinc/mol of protein. The zinc is resistant to removal by dialysis against concentrations of EDTA which deplete zinc from alcohol dehydrogenase. The bound zinc can be removed by reaction with the reversible sulfhydryl reagent p-hydroxymercurimethylsulfonate and the zinc-depleted protein transiently retains DNA binding activity. Digestion of both native and zinc-depleted ICPB with VB protease indicates that the bound zinc is required for the structural integrity of the protein. The major herpes simplex virus 1 (HSV-1)’ DNA-binding protein commonly designated infected cell polypeptide 8 (ICP8) is one of seven HSV-1 gene products required for origin-dependent replication of HSV DNA (1, 2). A large body of work with conditional lethal mutants mapping in the gene encoding ICP8

The major herpes simplex virus type 1 (HSV-1)' DNAbinding protein commonly designated infected cell polypeptide 8 (ICP8) is one of seven HSV-1 gene products required for origin-dependent replication of HSV DNA (1, 2). A large body of work with conditional lethal mutants mapping in the gene encoding ICP8 indicates an absolute requirement for ICP8 in the replication process (3-5). The nucleotide sequence of the ICP8 gene (6) indicates that the protein is composed of 1,196 amino acids and has a predicted molecular weight of 128,341. This value is in good agreement with previous estimates of 120,000-130,000 for the molecular weight of ICP8 ICP8 binds tightly and preferentially to single-stranded DNA. This interaction is cooperative based on electron microscopic analysis of ICP8. DNA complexes prepared with sub-saturating amounts of protein (8-10) and by Scatchard analysis (11). ICP8 also binds to duplex DNA and polyriboadenylic acid acid but with a lower affinity than that observed * This work was supported by United States Public Health Service Grant A122468 from the National Institute for Allergy and Infectious Disease. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisemerzt" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
with single-stranded DNA, as determined by salt elution from DNA-cellulose matrices and/or by competitive filter binding assays (12)(13)(14). ICP8 has also been shown to have the ability to melt duplexes of complementary homopolymers and to protect bound single-stranded DNA from nuclease digestion (15, 16). These properties are reminiscent of those displayed by the T4 gene 32 protein and the Escherichia coli SSB, both of which also bind cooperatively to single-stranded DNA and are required for DNA replication (17,18). Further evidence that ICP8 is an analogue of these prokaryotic proteins comes from recent work by Hernandez and Lehman (19) who have shown that ICP8 is required for the complete processive synthesis of singly primed @X-174 DNA in the presence of the gene products of the HSV origin dependent replication system.
Following its synthesis in the cytoplasm, ICP8 moves to the nucleus where it intially interacts with the nuclear matrix at discrete prereplicative sites. With the onset of DNA replication, ICP8 moves into regions of the nucleus which have been dubbed "replication compartments." More recently it has been shown that ICP8 is required for the organization of DNA replication proteins at discrete sites within the cell nucleus (20). Thus ICP8 appears to be responsible for the organization of the proposed HSV DNA replication complex. Finally, studies with temperature sensitive mutants indicate that ICP8 is involved in the regulation of expression of late genes during HSV infection. Specifically this protein appears to down-regulate expression of such genes by an as yet unknown mechanism (21).
Examination of the predicted amino acid sequence of ICP8 shows that residues 499-512 conform to the consensus sequence (Cys-X2-5-Cys2.16-Cys/His-X~~4-Cys/His) for a divalent metal coordination site or "zinc finger" described by Berg (22). This sequence is -C-N-L-C-T-F-D-T-R-H-A-C-V-Hand occurs near the midpoint of the primary structure of the molecule (Fig. 1). We report here on experiments which show that ICP8 is a zinc metalloprotein and that the zinc atom appears to be required for the overall structural stability of ICP8.

MATERIALS AND METHODS
Cells and Viruses-Vero and U-35 cell lines were propagated in monolayer culture as described by Orberg and Schaffer (23) using Dulbecco's modified Eagle's medium suplemented with 2% fetal calf serum and 8% Serum Plus (Hazelton Laboratories). Stocks of HSV-1 strain mP were propagated and titered in Vero cell monolayers.
Purification of ICP8 from HSV-1-infected U-35 Cells"ICP8 was purified from U-35 cells infected with HSV-1 strain mP at a multiplicity of infection of 1. This cell line is stably transformed and contains multiple copies of the ICP8 gene integrated into the cellular genome. The U-35 cell line has been used by several laboratories to generate large quantities of homogenous protein (16, 24). ICP8 was purified from U-35 cell pellets using protocols similar to those developed for purification of this protein from infected Vero cells (25). Briefly, 2 x IO9 infected U-35 cells which had been stored as pellets frozen at -70 "C were resuspended in 200 ml of lysis buffer (1.7 M KCL, 50 mM Tris hydrochloride (pH 7.5), 5 mM EDTA, 0.5 mM ditiothreitol, 0.2% Nonidet P-40 (Sigma), 20 pg/ml phenylmethylsulfonyl fluoride, and 2.5 Fg/ml leupeptin (Sigma)). The cells were lysed by gentle agitation, and the resulting viscous solution was clarified by centrifugation for 6 h at 24,000 rpm in a Beckman SW28 rotor. Subsequent dialysis against buffer containing 0.15 M KCl, and purification by DNA-agarose chromatography and heparin-Sepharose chromatography were carried out as described previously (25). The final yields of homogeneous ICP8 (homogeneity defined as the migration of the protein as a single band by SDS-PAGE) from two separate preparations were 8 and 10 mg. The protein was dialyzed against buffer containing 0.15 M KCI, 10 mM Tris-C1 (pH 7.6), 1 mM EDTA, 0.1 mM dithiothreitol, and 50% glycerol and stored a t -20 "C prior to use. Protein concentrations were determined using the method of Lowry (26) following precipitation of protein with 5% tricholoracetic acid and redissolution in 1 N NaOH.
Reaction of ICP8 with PMPS-Reaction of ICP8 with the reversible sulfhydryl reagent PMPS followed the procedure of Giedroc et al. (27). PMPS was dissolved in 100 mM NaCI, 10 mM Tris-C1 (pH 7.6) (TN buffer). The PMPS was added in 40-fold molar excess to ICP8 dialyzed in the same buffer, and the resulting mixture was incubated at 4 "C for 60 min. Following incubation the reaction mixture was divided and dialyzed against a variety of buffers as described under "Results and Discussion." A control sample of ICP8 was diluted without PMPS treatment and dialyzed in a similar fashion.
Zinc Analysis of ICP8-All buffers used for zinc analysis of ICP8 were treated by passage through a 200-ml Chelex-100 column (Bio-Rad). Prior to use, the Chelex-100 gel was activated by successive treatment with 4 volumes of 1.0 N HC1,4 volumes of deionized water, 4 volumes of 1.0 N NaOH, and an additional 4 volumes of deionized water (28). The buffers were Chelex-treated at 10 X concentration and stored in sterile plasticware (Bellco) which was found to be free of contaminating zinc. Zinc analyses were performed essentially as described by Foote and Delves (29) using a Perkin-Elmer model 2100 atomic absorption spectrometer equipped with an HGA-700 graphite furnace. A standard curve, linear from 0.1 to 1.0 pmol of zinc, using three independently diluted zinc standards (Johnson and Mathey Co., Seabrook, NH) was obtained for each set of experiments. Individual volumes and dilutions. Zinc content of horse liver alcohol dehydro-samples were assayed for the presence of zinc using a variety of genase (Sigma) and E. coli SSB (Sigma) were determined as a positive and negative control, respectively.
Nitrocellulose Filter Biding Assays-Nitrocellulose filter binding assays assessing the DNA binding activity of ICP8 were carried out as described previously in this laboratory (25). Briefly, 38 ng of heatdenatured X phage DNA which had been labeled with 3H to a specific activity of 2.8 X lo6 dpm/pg were mixed with increasing amounts of ICP8. The mixtures were brought to a final volume of 100 p1 by addition of TE buffer (10 mM Tris-CI, pH 7.6, 1 mM EDTA) and incubated on ice for 10 min. 80-p1 aliqouts were then withdrawn and passed through nitrocelluose filters using a Hoeffer multichannel filtration manifold. Filtration, washing, and counting of the filters was carried out as described previously (14).
V8 Protease Treatment of ICPB-Both control and zinc-depleted ICP8 were adjusted to a final concentration of 125 pg/ml. Two pl of 250 pg/ml V8 protease (Sigma) solution was then added to 20O-pl aliquots of each of the ICP8 samples, which were chilled in an ice bath. The reaction was begun by shifting the mixtures to 37 "C, and 2 0 4 aliquots were removed at 10-min intervals. Immediately upon removal the aliquots were mixed with an equal volume of 2 X SDS-PAGE sample buffer containing 0.2% SDS and heated at 100 "C for 2 min. The samples were held on ice until all time points had been collected. The protein products were then separated by SDS-PAGE using 9% acrylamide gels and visualized by staining with Coomassie Brilliant Blue.

RESULTS AND DISCUSSION
ICP8 derived from two separate preparations was dialyzed overnight against three 1-liter changes of zinc-free TN buffer, 0.1 mM dithiothreitol. The final concentration of ICP8 ranged from 125 to 300 bg/ml. The samples were then divided in half, and dialysis was continued against either TN buffer or against T N buffer containing 1.0 mM EDTA or 10 mM EDTA (0.1 mM dithiothreitol added to all samples). Horse liver alcohol dehydrogenase which is known to contain 4 mol of zinc/l mol of protein of molecular weight 80,000 (30) and the SSB, which does not contain zinc (27), were also dialyzed against zincfree buffers and used as controls. Samples were removed upon completion of dialysis, and the zinc content of the proteins was determined as described under "Materials and Methods." The results are presented in Table I. ICP8 from both preparations contained, on the average, 1 mol of zinc/mol of protein.
The zinc was resistant to exhaustive dialysis against EDTA concentrations of up to 10 mM in the presence of 0.1 mM dithiothreitol. In contrast, the zinc atoms bound to horse liver alcohol dehydrogenase were partially removed upon dialysis against 1.0 mM EDTA in the presence of dithiothreitol.
Giedroc et al. (27,31) had considerable success in addressing the role played by zinc in the activity of the T4 gene 32 protein using the reversible and strongly dissociating sulfhydryl reagent PMPS. Reaction with this reagent resulted in the loss of zinc from the T4 gene 32 protein and allowed reconsitution with zinc and other divalent metals upon removal of PMPS. Our treatment of ICP8 with PMPS followed the scheme outlined by these authors as described under "Materials and Methods." The results are shown in Table I1 and indicate that reaction with PMPS and subsequent reconstitution of free sulfhydryl groups with dithiothreitol resulted in the efficient removal of zinc from ICP8. This result is in contrast with those obtained with the T4 gene 32 protein where complete removal of zinc required dialysis against both

FIG.
2. Nitrocellulose filter binding assays using fraction I (+), fraction 2 (A), and fraction 3 (0) from the PMPS experiment descibed in the text and Table 11. All reaction mixes contained 38 ng of heat denatured .'H-Iaheled h DNA. dithiothreitol and EDTA (27). The positive control (sample 1) was dialyzed under nitrogen and in the presence of dithiothreitol and EDTA, and the results indicate that these procedures do not dissociate the bound zinc atom.
The single-stranded DNA binding activity of the zincdepleted ICP8 was assessed by the nitrocellulose filter binding assay. The DNA-binding activity of the zinc-depleted protein remained intact in experiments carried out immediately after treatment with dithiothreitol ( Fig. 2) but was gradually lost (as compared with native ICP8) after several days storage at 4 "C (Fig. 3). Attempts to reconstitute zinc-depleted ICP8 by dialysis against zinc-containing buffers were not successful.
Although zinc was bound, the molar ratios varied from 0.1 t o 3.2 mol of zinc/mol of protein, and no DNA binding activity was observed. These results suggest that the zinc atom in ICP8 is most likely not required for DNA binding but may play a role in maintaining the native tertiary structure of the protein. Thus loss of zinc would result in a destahilization of the protein tertiary st.ructure ultimately leading to an irreversible loss of activity.
In order to further examine this possibility, native and zincdepleted ICP8 were treated with V8 protease as described under "Materials and Methods." Time courses of reaction were carried out and the products analyzed by SDS-PAGE. The results show that the zinc-depleted ICI'8 was highly susceptihle to cleavage by the protease, whereas the native protein was relatively untouched (Fig.  4). This outcome is consistent with the idea that zinc is involved in the stabilization of t h e overall tertiary structure of ICP8 and suggests that the gradual loss of DNA binding activity noted above was most likely due to a slow denaturation of the protein.
In summary we have shown that ICP8 is a zinc-metalloprotein and contains 1 mol of zinc/mol of protein. The zinc atom is tightly bound and is resistant to removal by extensive dialysis against 10 mM EDTA. The zinc can be removed by reaction with the reversible sulfhydryl reagent PMPS and appears to be required for structural stability of ICP8 but not directly involved in the DNA binding activity of this protein as assessed by the nitrocellulose filter binding assay.
The question t.hen arises as to whether the zinc atom is actually located in the putative zinc finger encompassing residues 499-512. While no direct evidence is available, data gathered by several investigators suggest that this region ( a ) is important for the overall function of ICP8 in the replication of the virus and (6) is separate from the apparent DNAbinding domain of ICP8. Gao, Knipe, and co-workers (32,33) have shown that mutation of the ICP8 gene resulting in simultaneous substitution of cysteine with glycine at positions 499 and 502 gives rise to a protein which is 9 0 7 insoluble in infected cell extracts. Only one-third of the remaining soluble protein was capable of binding to single-stranded DNA cellulose matrices. This mutant designated pml was incapable of supporting virus growth on Vero cells. These investigators have also engineered a deletion mutant of IC1'8 ( d l 0 l ) which lacks 546 amino acids from t.he amino-terminal portion of ICP8 including the zinc finger region. The protein product encoded by the dl01 strain is also predominantly insoluble (78%) in infected cell extracts, but the remaining soluble portion did show DNA hinding. Although the size of the deletion is such that it is difficult to draw specific conclusions. the data are compatible with a model in which the zinc and  I~n m 2 A and I H contain molecular weight standards llG,250,92,500,  66,200, 45,000. and 30.000).
DNA-binding sites are separate.
Finally, data from several laboratories based upon work with the above deletion mutants (33), proteolytic fragments (16), and specific portions of ICPS expressed from recombinant plasmids (34) indicate that the DNA-binding domain of ICP8 most likely lies between residues 564 and 849. Thus the retention of DNA binding activity upon removal of zinc from ICPS presented here is again consistent with the evidence that the putative zinc-binding domain is outside the DNAbinding domain. Site-specific mutagenesis of the zinc finger region of ICPS such as that carried out by Gauss et al. (35) in the case of the gene 32 protein and zinc analysis of the resulting mutant proteins may be useful in the study of this problem and such work is currently underway. However, unambiguous location of the zinc atom in ICP8 will require x-ray crystallographic analysis of the protein.