VIROLOGY

191, 9-18 (1992)

Serine Protein Kinase Associated with Varicella-Zoster

Virus ORF 47

TERESA I. NG AND CHARLES GROSE’ Departments

of Microbiology

and Pediatrics,

University

of Iowa College of Medicine,

Iowa City, Iowa 52242

Received May 5, 1992; accepted July 17, 1992

Varicella-zoster virus (VZV) ORF 47 lies in the unique long region of the VZV genome. Sequence homology studies have demonstrated that gene 47 possessed conserved protein kinase motifs. In this study, we investigated the properties of the ORF 47 product. First, a rabbit antiserum was raised against a protein generated from the fusion of the most antigenic ORF 47 domain with fscherichia co/i &galactosidase. The high-titer antiserum reacted specifically with ORF 47 polypeptides translated in vitro. When incubated with VZV-infected cell lysate, the antiserum immunoprecipitated a phosphoprotein of Mr 54,000, a size comparable with the predicted molecular mass. The precipitated viral protein was phosphorylated in a protein kinase assay; subsequent phosphoamino acid analysis indicated that the phosphotransferase associated with the ORF 47 protein was a serine protein kinase. Synthesis of the ORF 47 product in VZV-infected cell culture increased in the first and second days and plateaued after the third day of infection. The protein kinase activity associated with VZV ORF 47 had several distinctive biochemical properties: (i) its phosphotransferase activity was enhanced more by manganese than by magnesium, (ii) it utilized both ATP and GTP as donors of phosphate, and (iii) it phosphorylated both acidic and basic substrates. In summary, this report lends support to the computer homology data which predicted that VZV ORF 47 would encode a serine protein kinase. o 1992 Academic press. I~C.

1986; Smith and Smith, 1989). This novel class of phosphotransferases has drawn great attention because they may play a regulatory role in the life cycle of herpesviruses and also they may be selected as potential targets for antiviral therapy. Protein kinases encoded by members of the alphaherpesvirinae: HSV, VZV, and pseudorabies virus (PRV)were the first to be identified. McGeoch and Davison (1986) discovered alphaherpesviral genes with conserved protein kinase motifs, HSV gene Us3 and VZV gene 66, in the short unique (Us) DNA segments of these viruses. A similar protein kinase gene was found in the U, region of PRV (Zhang et a/., 1990). The PRV kinase activity had been recognized prior to the knowledge about the viral DNA sequences (Katan et al., 1985). A later homology study conducted by Smith and Smith (1989) revealed that kinase-related genes were not restricted to the U, regions of the herpesvirus genomes, but were also present in the long unique (UJ regions of two alphaherpesviruses: UL13 of HSV and ORF 47 of VZV. In this report, we describe the identification and initial characterization of a serine protein kinase associated with VZV ORF 47.

INTRODUCTION

Members of the herpesvirus family are divided into three subfamilies, the alpha-, beta-, and gammaherpesvirinae according to their biological properties and genomic structures (Roizman, 1991). VZV is an alphaherpesvirus that causes chicken pox in children. After primary infection, the virus becomes latent in the dorsal root ganglia where it remains for decades, until reactivation occurs and causes the disease herpes zoster (shingles). VZV is the smallest human herpesvirus, having a linear double-stranded DNA genome of 125 kbp with about 70 open reading frames (ORF) (Davison and Scott, 1986). Most studies have concentrated on the five glycoprotein ORFs (Grose, 1990). All five VZV glycoproteins have counterparts in herpes simplex virus (HSV). Although little is known about the functions of most nonglycoproteins of VZV, it is now possible to speculate on the identity of some VZV gene products by performing computer-assisted sequence comparisons with other members of the Herpesviridae because these viruses share a substantial degree of homology throughout their immediate early, early, and late genes. Based on the data from sequence homology studies, there are recent reports that some members of the alpha-, beta-, and gamma-herpesviruses have the potential to encode viral protein kinases (Chee et al., 1989; Hanks et a/., 1988; McGeoch and Davison,

MATERIALS Cells, viruses,

AND METHODS

and plasmids

Monolayer cultures of human melanoma cells (Mewo strain) were grown in Eagle’s minimum essential medium supplemented with 10% fetal bovine

’ To whom reprint requests should be addressed.

9

0042-6822192 $5.00 Copyright 0 1992 by Academic Press. Inc. All rights of reproduction in any form reserved.

NG AND GROSE

10

serum, 2 mhllglutamine, 1%I nonessential amino acids, penicillin (100 U/ml), and streptomycin (100 pg/ml). Stock virus used for all experiments was the VZV-32 strain (Grose et a/., 1979). The VZV genomic library was generated by Ecker and Hyman (1982) and provided to us by Hyman (Palo Alto, CA). The prokaryotic expression vector pTRB2 for fusion protein construction was derived from pUR292 vector (Ruther and Muller-Hill, 1983). Construction

of a ,&galIORF

47 fusion protein

The initial strategy involved the production of a fusion protein in order to produce antibody that specifically recognized ORF 47 protein. The computer program “Antigen” predicted that the most antigenic epitope of ORF 47 spanned amino acids 36-41. ORF 47 is located in the EcoRl L and K fragments, in the UL region of the VZV genome. The 172-bp fstl-HindIll fragment of ORF 47 was cloned in frame downstream to the IacZ gene in pTRB2 (Fig. 1). The cloned ORF 47 fragment encoded a peptide that carried the most antigenic site predicted by the “Antigen” program, but this polypeptide did not possess any conserved protein kinase motif. Thus, this cloning strategy created a protein that contained the N-terminus of the ORF 47 protein fused to the C-terminus of P-galactosidase (b-gal). Isotopic

labeling and immunoprecipitation

Because of the cell-associated nature of the virus, infections were carried out with inocula consisting of VZV-infected cells, as described by Grose et al., (1979). For in viva radiolabeling, culture medium in a 25cm2 culture flask of VZV- or mock-infected cells was replaced with medium containing 1.25 mCi of [32P]orthophosphate (Amersham) or 1 mCi of [35S]cysteine (Amersham) at 20 hr postinfection. The cells were incubated for an additional 48 hr at 32” and then harvested. Indirect immunoprecipitation was carried out in RlPAbuffer(0.01 MTris-HCI,pH7.4,0.15MNaCI, 1% deoxycholate, 1% Nonidet P-40) containing varying final concentrations of sodium dodecyl sulfate (SDS) from 0.1% to 0.4%. SDS-polyacrylamide gel electrophoresis (SDS-PAGE) of VZV protein was performed by published methods (Grose and Friedrichs, 1982). lmmunoprecipitation

of in vitro-translated

products

Indirect immunoprecipitation of in vitro-translated ORF 47 protein was based on methods which have been previously described (Grose and Friedrichs, 1982; Anderson and Blobel, 1983). Two conditions of stringency were tested. In the first, the translation mixture was adjusted to 3% SDS to solubilize the protein and diluted 10 times with RIPA buffer containing 0.1%

SDS. In the second, the translation mixture was solubilized directly in RIPA buffer (0.1% SDS). The precipitation mixture contained 170 ~1of diluted antigen and 5 ~1 of rabbit antiserum. In control experiments, ORF 66 was transcribed and translated and the gene product was immunoprecipitated in a manner identical to that of ORF 47. Protein kinase assay of immunoprecipitated proteins

viral

VZV- and mock-infected cell lysates solubilized in RIPA buffer (0.1 to 0.4% SDS) were immunoprecipitated with rabbit antiserum and protein A-Sepharose CL-4B beads (Pharmacia). Protein A beads with the antibody-bound ORF 47 protein were washed five times with PBS wash buffer (10 mM sodium phosphate, pH 7.2, 150 mM NaCI, 0.5% Nonidet P-40, 0.5% bovine serum albumin, 0.1% SDS, and 0.2% NaN,) and then once with kinase buffer (25 mM HEPES, pH 7.4, 10 mM MnCI,, and 50 mlLl KCI). The washed beads were suspended in 70 ~1 of kinase buffer. The protein kinase assay was initiated by the addition of 5 &i [T-~‘P]ATP (3000 Ci/mmol; Amersham). The reaction mixture was incubated at 30” for 30 min and the pellets were washed five times with PBS wash buffer. ORF 47 protein was subsequently eluted from the protein A beads by boiling for 5 min in reducing sample buffer (0.125 MTris-HCI, pH 6.8, 6% SDS, 20% glycerol, 10% 2-mercaptoethanol) and analyzed by SDS-PAGE. Protein kinase assay with general substrates Washed immunoprecipitates of ORF 47 protein were incubated with histone (Sigma, no. H7755) or casein (Sigma, no. C4765) and 5 &i [T-~~P]GTP. Three concentrations of each substrate were tested: 3 pg/ml, 0.3 pg/ml, and 0.03 pg/ml. After incubation at 30” for 30 min, the antigen-antibody complexes were removed by centrifugation and the substrates were precipitated from the supernatant with cold 20% trichloroacetic acid (TCA) by incubating on ice for 30 min. Precipitated proteins were washed two times with 20% TCA, two times with 95% ethanol, and dried under vacuum. The dried pellets were then suspended in reducing sample buffer and subjected to SDS-PAGE in 15% gels. Utilization

of GTP as phosphate

Phosphorylation GTP (30 Ci/mmol; dures described beled GTP, 5 PM (1 pnll) radioactive

donor

was carried out with 2 &i [T-~~P]Amersham) according to the proceabove. In the experiment with unlaunlabeled GTP was mixed with 2 &i GTP and this mixture was the phos-

VZV PROTEIN KINASE

11

L

C 1

II PP Ii 1111

NOG 11

IRS

8

A/

L

D I

11

I

K I

I

S

TRs

EF I

I

EcoR I gsnomic I I bray

I

\

H/M

vzv genomr

\

ORF 47

stu I

Hind III

Pstl

EcoR I

PSI I pTRB2

\

pBluescrip1

SK

stu I EcoRV -r

insert 4.6 kb

FIG. 1. Cloning strategies for VZV ORF 47. The long (L) and short the internal repeat sequence (IRS) and terminal repeat sequence protein, the 172-bp Pstl-HindIll fragment of ORF 47 was cloned promoter. For in vitro transcription and translation, the St&EcoRI sites of pBluescript. See text for details.

RESULTS of ORF 47-specific

T3

promoter

(S) unique sequences of the VZV genome are represented as open bars and (TRs) as slashed boxes, For the construction of a r!%galNZV ORF 47 fusion in pTRB2 in frame with the LacZ gene that is under the control of the P,,, segment of &coRl L fragment was inserted between the EcoRV and EcoRl

phate donor in the protein kinase assay. In the control experiments, VZV glycoprotein I (gpl) was immunoprecipitated and phosphorylated by casein kinase II in the presence of 2 PCi [T-~~P]GTP with or without 10 PM unlabeled GTP, using described methods (Grose et al., 1989). Purified casein kinase II was provided by J. Traugh, University of California, Riverside.

Production

t

antibody

To generate a large amount of ORF 47 fusion protein for immunization, the P-galIORF 47 fusion construct (Fig. 1) was transformed into E. co/i and fusion protein synthesis was induced by growth in medium containing 5 mM isopropyl-p-D-thio-galactopyranoside (IPTG).

The bacterial cells were lysed and the bacterial proteins were separated by SDS-PAGE. After the gel was stained with Coomassie blue, a novel induced protein that migrated slower than P-gal was observed. The induced fusion protein, which was the major product in the bacterial cell lysate, was excised from the gel and injected intramuscularly into two rabbits. VZV-specific antiserum was raised by each rabbit. When analyzed by indirect immunofluorescence, the rabbit anti-ORF 47 antibody bound mainly to the cytoplasm of VZV-infected cells (Fig. 2). The perinuclear concentration of immunostaining may indicate an accumulation of the VZV protein within large virus-filled vacuoles commonly found in this location (Grose et a/., 1979). Immune serum did not attach to mock-infected cells (Fig. 2). Preimmune serum also was reacted with infected cells, with negative findings similar to those shown in the lower panel of Fig. 2.

NG AND GROSE

12

FIG. 2. Reactivity of rabbit antiserum with VZV infected cells. (Upper panel) VZV-infected cells 2-days postinfection were reacted with anti-ORF 47 rabbit antiserum in an indirect immunofluorescence staining assay. A 1:200 dilution of antiserum was added to acetone-fixed cells for 30 min at 37”. After washing, the cells were reacted with goat anti-rabbit fluorescein conjugate. (Lower panel) Mock-infected cells were reacted with anti-ORF 47 rabbit antiserum, as described above.

In vitro transcription

of VZV genes

ORF 47 was transcribed and translated in vitro to provide antigen for testing the specificity of the rabbit antiserum. ORF 47 is not intact in either the EcoRl or the Hindlll VZV genomic libraries because both EcoRl and Hindlll cleave the sequence at sites within the gene. Since the EcoRl L fragment of the genomic library contains the largest portion of ORF 47, i.e., 879/o of the whole gene, the St&EcoRI segment of this fragment was inserted between the EcoRV and EcoRl

restriction sites of pBluescript plasmid (Stratagene) for in vitro transcription and translation (Fig. 1). In addition to the ORF 47 sequence, this insert contained a sequence of 283 bp upstream of the AUG start codon of ORF 47 but it lacked the 196 bp at the 3’ end of the gene. The cloned kinase sequence was linearized at EcoRl or Smal (Fig. 1) to generate two templates that specified ORF 47 polypeptides with different C-terminal deletions. With T7 RNA polymerase, the VZV DNA samples were transcribed in the presence of ribonucleoside triphosphates. As a control reagent, the sec-

VZV PROTEIN KINASE

A

B .

ORF 47

12 Sma I

3 EcoRl

* lmmunoprecipitation

Translation

RNA transcript

13

none

-EcoR

lmmunoppt

Translation

456789

:

-0RF66

12 I-+Sma

I-

ECOR v Bamti

34 I EcoR V BamH

I

FIG. 3. lmmunoprecipitation of truncated VZV ORF 47 and ORF 66 polypeptides translated in vitro. (A) Transcripts of VZV ORF 47 template cleaved at Smal (lane 1) and EcoRl (lane 2) restriction sites (see Fig. 1) were translated in rabbit reticulocyte lysates in the presence of [35S]methionine. Lane 3 represents the translation reaction containing no ORF 47 mRNA. The M, values of the radioactive VZV polypeptides are 35,000 (lane 1) and 47,000 (lane 2) based on migration relative to the following molecularweight marker proteins: phosphotylase b (97K), bovine serum albumin (66K), actin (45K), and carbonic anhydrase (29K). After in vitro translation, the EcoRl (lanes 4-6) and Smal (lanes 7-9) translated mixtures were immunoprecipitated with no rabbit serum (lanes 4 and 7) preimmune serum (lanes 5 and 8). or immune serum (lanes 6 and 9). (B) ORF 66 cleaved at fcoRV (lane 1) or BarnHI (lane 2) was translated in rabbit reticulocyte lysate in the presence of [35S]methionine. Lanes 3 and 4 represent the results of immunoprecipitations (immunoppt) of the two ORF 66 translated polypeptides by immune serum to ORF 47. A and B represent separate gels; exposure time of each gel was.2 days.

ond putative VZV protein kinase gene (ORF 66) was subcloned from the HindIll C fragment of the VZV genomic library into theXbal and EcoRV sites of pBluescript (cloning strategy not shown). VZV ORF 66 was linearized at EcoRV and BarnHI sites for in vitro transcription. lmmunoprecipitation

of translated

polypeptides

The two ORF 47 mRNAs were translated in rabbit reticulocyte lysate in the presence of [35S]methionine for subsequent analysis by SDS-PAGE and immunoprecipitation. The truncated radioactive polypeptides migrated according to their predicted molecular weights, a result which indicated that the in vitro-translated products were specific to the ORF 47 gene. Based on a comparison with standard molecular weight marker proteins, the lul, of the larger product was 47,000 and the smaller one was 35,000 (Fig. 3A, lanes 1 and 2). The translation mixture lacking viral mRNA did not yield any radioactive peptide (lane 3). To test the specificity of ORF 47 protein-specific antibody, the rabbit antiserum was reacted with the two in vifrotranslated VZV ORF 47 truncated polypeptides and it immunoprecipitated both proteins (Fig. 3A, lanes 6 and 9). The faint lower band detectable in lane 2 of Fig. 3A was also observed in lane 6. This product could be either a premature termination of translation from the 5’

AUG or an initiation of translation at a site other than the first methionine. Neither the preimmune serum nor the control mixture lacking antibody reacted with the radioactive viral polypeptides. The above immunoprecipitations were carried out in the presence of SDS final concentrations of either 0.1 or 0.4%; under both conditions, the precipitation profiles were identical. In control experiments, the ORF 66 mRNAs were translated and yielded two products (Fig. 3B, lanes l-2). The protein in lane 1 represents the entire gene product, with a predicted mass of 44,000 Da, while that in lane 2 represents the truncated gene product. In an immunoprecipitation reaction with anti-ORF 47 serum, the two ORF 66 products were not detected (Fig. 3B, lanes 3-4). These results confirmed that the rabbit immune serum contained specific antibody against an epitope found in the ORF 47 polypeptides. Detection

of ORF 47 protein

in VZV infected

cells

A unique feature of many protein kinases is their ability to autophosphorylate. As the word autophosphorylation implies, the protein kinase phosphorylates its own hydroxyamino acid residues or those of a neighboring identical protein. In addition, many kinases are themselves phosphorylated by other protein kinases. If ORF 47 protein were phosphorylated in cell culture

NG AND GROSE

14 3zp

1

2

-

3

-35s

4

5

Phosphorylation protein

-

6

7

8

FIG. 4. Detection of ORF 47 protein in VZV-infected cells. Mock-infected cells (lanes 1 and 2) and VA/-infected cells (lanes 3-8) were labeled with [3ZPlorthophosphate (lanes l-4) or [35S]cysteine (lanes 5-8) for 48 hr. Solubilized cell lysates were immunoprecipitated with either preimmune (lanes 1, 3, 5, and 7) or immune (lanes 2,4, 6, and 8) rabbit serum. Samples in lanes l-6 were immunoprecipitated in RIPA containing 0.1% SDS and those in lanes 7-8 were immunoprecipitated in RIPA containing 0.4% SDS. ORF 47 protein is designated by closed circles in lanes 4, 6, and 8. Lanes l-4 and lanes 5-8 represent separate gels.

under either circumstance, it would be possible to easily identify the viral protein kinase among numerous other phosphoproteins in an infected cell culture. To this end, a VZV-infected monolayer was incubated in the presence of [32P]orthophosphate and then solubilized for antigen preparation. The rabbit anti-ORF 47 antiserum specifically immunoprecipitated a phosphoprotein from the VZV-infected cells, while the preimmune serum failed to react with any protein (Fig. 4). Based on its migration relative to standard molecular weight marker proteins, the Mrof the authentic ORF 47 protein was 54,000. This value agreed well with the predicted molecular weight of the deduced amino acid sequence (Davison and Scott, 1986). To further assess the specificity of immunoprecipitation, a [35S]cysteine-labeled infected cell lysate was subjected to immunoprecipitation in a similar manner 32P-labeled lysate. The immunoprecipitation as the was performed in RIPA buffer containing either 0.1% SDS (lanes 5 and 6) or 0.4% SDS (lanes 7 and 8). The antiserum reacted with ORF 47 protein under both conditions. Although the binding in 0.1 O/O SDS was greater, the binding in 0.4% SDS demonstrated the high avidity of the antiserum to viral antigen. No radiolabeled protein coprecipitated with ORF 47. Results from Figs. 3 and 4, when taken together, establish the specificity of the rabbit antiserum to the ORF 47 protein.

of immunoprecipitated

ORF 47

Results in the previous section indicated that the VZV ORF 47 protein was 32P-labeled in cell culture. To test whether the same gene product could be phosphorylated in vitro, the protein was immunoprecipitated from VZV-infected cells and mixed with [r-“*PIATP in a protein kinase assay. The ORF 47 protein was phosphorylated and easily detected under these reaction conditions (Fig. 5, lane 3). The same phosphotylated protein was detected regardless of whether the immunoprecipitate was obtained under low (0.1% SDS) or high (0.49/o SDS) stringency conditions. The preimmune serum did not immunoprecipitate any protein that could be phosphorylated in a protein kinase assay nor was any phosphorylated protein detectable after immunoprecipitates from mock-infected cells were placed in a protein kinase assay. To determine the kinetics of ORF 47 protein synthesis in an infected cell culture, a time course experiment was conducted (Fig. 5, lanes 5-8). lmmunoprecipitates were obtained daily from infected cultures, and each immunoprecipitate was placed in a protein kinase assay. At 1 day postinfection, ORF 47 protein was barely detectable in the VZV-infected cell culture. It became more evident in a 2-day infection assay but was most abundant 3

FIG. 5. Phosphot-ylation of ORF 47 protein in virro and accumulation of ORF 47 protein in VZV-infected cells. Solubilized VZV-infected cell lysate was immunoprecipitated without rabbit serum (lane 1) and with preimmune serum (lane 2) or immune serum (lane 3); in lane 4, a mock-infected cell lysate was reacted with immune serum. Phosphorylation of each precipitate was carried out as described under Materials and Methods. As shown in lanes 5-9, one VZV-infected cell culture was harvested on Days 1 through 5 postinfection and ORF 47 protein was immunoprecipitated and identified after phosphorylation in a protein kinase assay. Lanes l-3 and lanes 4-9 represent separate gels.

VZV PROTEIN KINASE

and 4 days postinfection. Because of the inability to synchronize VZV infection in tissue culture (Grose er al., 1979) it was difficult to assess the temporal class to which ORF 47 should be assigned. When phosphonoacetic acid (PAA) was added to the culture medium at concentrations of 20-200 pug/ml, the amount of ORF 47 did not decrease in Days 3-4 postinfection (data not shown). This result suggested that the gene product belonged to either an early (beta) or early-late (beta-gamma) class. Optimum activity

reaction

conditions

for protein

kinase

Different protein kinases require different concentrations of magnesium (MS*‘) or manganese (Mn*+) as the divalent cation donor, e.g., HSV U,3 protein kinase and PRV-PK react best in the presence of Mg2+ (Katan et a/., 1985; Purves et a/., 1986). To test if the protein kinase associated with VZV ORF 47 had the same preference for Mg’+ over Mn *+ , different concentrations of Mg2+ and Mn2+ were included in the buffers prepared for the kinase assay of immunoprecipitated ORF 47 protein. The optimum cation concentration was then tested against various concentrations of KCI to define the best concentration of this salt. Concentrations of Mn2+, Mg2+, and KCI were selected based on known requirements of other protein kinases. After the kinase assays, one-tenth of each reaction mixture was measured in a scintillation counter and nine-tenths analyzed by SDS-PAGE. The efficiency of the reaction was calculated by comparing the amount of 32P incorporated into the substrate under different conditions, as reflected by the radioactivity of the counted reaction mixtures. The phosphorylated proteins separated by SDS-PAGE were excised from gels and their radioactivity quantified by scintillation counting to confirm the accuracy of the previous data. The results are tabulated in Table 1. Of the concentrations tested, Mn*+ was a better choice of cation donor; in fact, Mn2+ was found to be at least twice as efficient as Mg2+ as the divalent cation donor. This unusual requirement for manganese is different from that of the U, protein kinases of other herpesviruses, which react best in the presence of magnesium. The optimum KCI concentration had a broad range, similar to that observed with the PRV protein kinase (Katan et a/., 1985). Phosphorylation

of histone

and casein

Histone and casein are common substrates for protein kinases. To examine the ability of the protein kinase associated with VZV ORF 47 to catalyze the phosphorylation of exogenous substrates, both histone and casein were examined in protein kinase as-

15 TABLE 1

DETERMINATIONOF THE OPTIMUM REACTIONCONDITIONSFOR PHOSPHORYLATIONOF THE IMMUNOPRECIPITATEDORF 47 PROTEIN

Fw

Test No.

Mn2+ (mnn)

Mg2+ b-W

KC1 (mM)

(crx-4

%

1 2 3 4 5 6 7

10 0 50 0 10 10 10

0 10 0 50 0 0 0

0 0 0 0 50 250 500

10,861 1,113 4,490 686 15,546 8,366 14,190

70 7 29 4 100 54 91

Nore, [“PI indicates the amount of radioactivity incorporated into the substrate; maximum range of each mean value was f 15%. Percentage (%) recovery of radioactivity is based on the highest recovery being considered as 100%.

says. When immunoprecipitated ORF 47 protein was mixed with each substrate in the presence of [r-“‘PIATP and kinase buffer, both substrates were phosphorylated (Fig. 6). Both histone and casein are mixtures; therefore, there were more than one phosphorylated species in each sample. In a similar manner to Table l( the optimum conditions for phosphorylation of casein were defined and results were tabulated in Table 2. Mn2+ was the preferred divalent cation donor; in the presence of 10 mM Mn2+, phosphorylation was greatest with 50-250 mM KCI concentrations. This broad range may reflect the fact that neither enzyme nor substrate were purified molecules. As noted for other protein kinases, the KCI optimum may vary considerably with any change in substrate or buffer conditions (Hathaway and Traugh, 1983). Phosphoamino

acid analysis of ORF 47 protein

Members of the protein kinase family are classified into two categories according to the hydroxyamino acids which they phosphorylate: (i) kinases that phosphor-ylate serine or threonine and (ii) kinases that phosphorylate tyrosine. It is predicted by sequence homology that the ORF 47 protein would phosphorylate serine/threonine (Smith and Smith, 1989). To confirm which amino acid residues were phosphorylated, the in vitro-phosphorylated ORF 47 protein was subjected to phosphoamino acid analysis. Radiolabeled ORF 47 protein was excised from the gel, hydrolyzed, and analyzed by thin-layer electrophoresis, as previously described by Grose et al. (1989). The results indicated that VZV ORF 47 protein was predominantly phosphorylated at a serine residue with perhaps a small degree of phosphorylation at a minor threonine site (Fig. 7).

NG AND GROSE

16 C

H

c

H

FIG. 7. Phosphoamino acid analysis of phosphorylated ORF 47 protein kinase. ORF 47 protein, which had been phosphorylated in virro in the presence of [y-32P]ATP. was excised from an unfixed polyacrylamide gel. The radioactive protein was eluted, acid hydrolyzed, analyzed by thin-layer electrophoresis, and visualized by autoradiography. Unlabeled phosphoserine (P-SER), phosphothreonine (P-THR), and phosphotyrosine (P-TYR). which were electrophoresed in the lane adjacent to the hydrolyzed sample, were visualized with 0.2% ninhydrin.

PK : MOCK-INFECTED CELLS Ab :

ANTI-ORF

47

WV INFECTED CELLS ANTI-ORF

47

FIG. 6. Phospholylation of casein and histone by protein kinase associated with ORF 47 protein. Mock-infected (lanes 1 and 2) and VZV-infected (lanes 3 and 4) cell lysates were immunoprecipitated by anti-ORF 47 antiserum. These precipitates were the source of the protein kinase activity (PK). Casein (C) or histone (H) was incubated with each immunoprecipitate in a protein kinase assay as described under Materials and Methods. After the phosphorylation reactions, casein and histone samples were subjected to electrophoresis in separate 15% SDS-polyacrylamide gels. The lanes on the autoradiograms were grouped based on source of protein kinase.

GTP as a phosphate donor A small number of protein kinases, e.g., casein kinase II, utilize GTP as well as ATP as a phosphate TABLE 2 OPTIMUM REACTIONCONDITIONS FOR PHOSPHORYUTION OF SUBSTRATE CA~EIN BY PROTEINKINASEASSOCIATEDWITH ORF 47 PROTEIN

Test No.

Mn2+ (mm

1 2 3 4 5 6 7

10 0 50 0 10 10 10

[W 0 10 0 50 0 0 0

0 0 0 0 50 250 500

km)

%

26,877 3,468 5,140 10,574 39,034 38,945 27,083

69 9 13 27 100 100 69

Nore. [3*P] indicates the amount of radioactivity incorporated into the substrate; maximum range of each mean value was + 15%. Percentage (%) recovery of radioactivity is based on the highest recovery being considered as 100%.

donor (Hathaway and Traugh, 1983). Because all previous experiments had been carried out with ATP, protein kinase assays were repeated in the presence of GTP and results compared with ATP of the same specific radioactivity and concentration. Under these conditions, the ORF 47 product was phospholylated in vitro with GTP as the source of phosphate, although ATP appeared to be a better phosphate donor than GTP in the protein kinase assay (Fig. 8, lanes 4 and 5). To confirm the specificity of GTP as a phosphate donor, a mixture of radioactive and unlabeled GTP was substituted as the source of phosphate in the phosphotransferase assay. If the usage of GTP was specific, the unlabeled GTP in the reaction mixture should be able to compete with the radioactive GTP as phosphate donor. This assumption was confirmed when unlabeled GTP was found to reduce the amount of [T-~~P]GTP incorporated into the protein (lanes 5 and 6). As a control, we repeated an experiment from this laboratory which showed that VZV gpl could be phosphorylated by casein kinase II with GTP as the source of phosphate (Grose et a/., 1989). When unlabeled GTP was added into this reaction, the amount of [-y-32P]GTP incorporated into immunoprecipitated gpl was substantially reduced (lanes 1 and 2). In another assay, we demonstrated that the protein kinase associated with ORF 47 protein phosphorylated casein in the presence of [T-~‘P]GTP (data not shown). All these experiments substantiated the specificity of GTP as a phosphoryl donor for the kinase activity associated with ORF 47 protein.

DISCUSSION In this study, we provide an initial characterization of the VZV ORF 47 gene product. As described under

VZV PROTEIN KINASE

ClPl.

0 1ORF

47

FIG. 8. Utilization of GTP as a phosphate donor in the protein kinase assay. Lanes l-3 represent immunoprecipitates of WV glycoprotein gpl and lanes 4-6 immunoprecipitates of ORF 47 protein. Phosphorylation of the ORF 47 protein kinase was carried out in the presence of [y-32P]ATP (lane 4) [T-~*P]GTP (lane 5) or a mixture of unlabeled GTP and [Y-~*P]GTP (lane 6). As a control assay, immunoprecipitated gpl was phosphorylated by casein kinase II in the presence of [y-32P]GTP (lane 1) or a mixture of unlabeled GTP and [-r-3’P]GTP (lane 2). Lane 3 represents an additional control experiment with immunoprecipitated gpl incubated with radioactive GTP in the absence of casein kinase II.

Results, the immunoprecipitated ORF 47 protein has phosphotransferase properties. The question is whether this activity resides within the viral protein itself or, alternatively, represents a minor contaminating enzyme coprecipitating with the viral protein. We favor the first possibility for several reasons: (1) The kinase activity continues to be present when precipitation is carried out under increasingly stringent conditions of higher SDS concentrations, which usually disrupt noncovalent protein-protein interactions. (2) The kinase activity does not resemble that associated with common mammalian kinases. For example, although casein is a substrate, the kinase reacts best under conditions that do not resemble those for either casein kinase II or casein kinase I. Interestingly, the ability to use Mn2+ as a cation donor has been associated with regulation of autophosphorylation on several kinases (White and Kahn, 1986). (3) The kinase activity can utilize GTP as well as ATP as a phosphate donor. Few mammalian kinases have this property, with the notable exception of casein kinase II. Yet the kinase described herein bears no other resemblance to casein kinase II, which has two subunits (rJI, 44,000 and 28,000). If the ORF 47 related kinase activity is due to a contaminating enzyme, there is no obvious candidate protein kinase.

17

Lastly, the kinase activity closely resembles that recently found to be associated with HSV homolog U,13 protein. Cunningham et al. (1992) investigated the properties of the HSV U-1 3 protein, which they discovered to be a phosphoprotein in the nucleus of HSV-infected cells. Within a protein kinase assay, the phosphotransferase properties associated with a nuclear extract containing lJL1 3 included the usage of GTP, as well as ATP, and the ability to phosphorylate exogenous casein. They presented the dilemma that HSV-1 infection could induce a cellular enzyme which catalyzed the phosphorylation of Q-13 but they considered the evidence to be stronger that UL13 was itself a protein kinase, that was capable of autophosphorylation. When all the above issues are taken into consideration, we concur with Cunningham et al. (1992) and favor the possibility that the VZV ORF 47 product is a functional serine protein kinase, which can autophosphorylate itself or a neighboring ORF 47 molecule. A potential physiologic substrate for an alphaherpesviral U, protein kinase has been reported recently. Purves et a/. (199 1) discovered that phosphorylation of the HSV type 1 U,34 gene product was catalyzed by the HSV-1 U,3 protein kinase. In particular, the posttranslational processing of the U,34 phosphoprotein was dependent on the presence of a functional U,3 protein kinase. Gene product 34 is also modified by other protein kinases, but their identities have not yet been determined. Interestingly, the U,3 protein kinase does not appear to be essential for HSV-1 growth in cell culture because it can be deleted (Purves et a/., 1987; Longnecker and Roizman, 1987). Whether the alphaherpesviral U, protein kinase can ever substitute for the deleted U, protein kinase is an intriguing possibility which needs to be investigated as part of a broader examination of substrates for both U, and U, protein kinases. In his extensive review of the herpesviral genomes, McGeoch (1989) presented his concept of the evolution of the protoherpesvirus, certainly one of the most ancient viruses antedating humankind. McGeoch (1989) noted that there is a cluster of well-preserved genes common to most herpesviruses. Furthermore, these herpesviral genes display similarity in their deduced amino acid sequences to nonherpesviral genes and, therefore, may have been acquired, directlyorindirectly, from a cellular genome. Among these ancient and preserved genes, he includes the U, protein kinase. The U,- genes were assembled as part of the protoherpesvirus long before the more recent division into the alphaherpesvirinae 30-70 million years ago (Gentry et a/., 1988). By this reasoning, the U, protein kinase may fulfill a regulatory function central to the general replication cycle of herpesviruses, while the U,

18

NG AND GROSE

kinase has a function related specifically to alphaherpesviruses. One possible function is to catalyze the phosphorylation of a cellular substrate, an event which would change the environment within the infected cells so as to facilitate the viral infection (Leader and Purves, 1988). Another possibility is that the UL protein kinase may participate as an essential team player in the phosphorylation of one or more viral proteins (Traugh, 1989).

ACKNOWLEDGMENTS We thank Richard Hyman for supplying the VZV genomic library and Jolinda Traugh for providing purified casein kinase II and, in addition, for reviewing the manuscript. This research was supported by USPHS Grant A122795.

REFERENCES ANDERSON, D. J., and BLOBEL. G. (1983). lmmunoprecipitation of proteins from cell-free translations. Methods Enzymol. 96, 11 l-l 20. CHEE, M. S., LAWRENCE, G. L., and BARRELL, B. G. (1989). Alpha-, beta- and gammaherpesviruses encode a putative phosphotransferase. J. Gen. viral. 70, 1151-l 160. CUNNINGHAM, C., DAVISON, A. J., DOLAN, A., FRAME, M. C., MCGEOCH, D. J., MEREDITH, D. M., Moss, H. W. M., and ORR, A. C. (1992). The UL13 virion protein of herpes simplex virus type 1 is phosphorylated by a novel virus-induced protein kinase. J. Gen. viral. 73, 303-311. DAVISON, A. J., and Sco-rr, J. E. (1986). The complete DNA sequence of varicella-zoster virus. 1. Gen. Viral. 67, 1759-l 816. ECKER, J. R., and HYMAN, R. W. (1982). Varicella zoster virus DNA exists as two isomers. Proc. Nat/. Acad. Sci. USA 79, 156-l 60. GENTRY, G. A., LOWE, M.. ALFORD, G., and NEVINS, R. (1988). Sequence analyses of herpesviral enzymes suggest an ancient origin for human sexual behavior. Proc. /Vat/. Acad. SC;. USA 85, 26592661. GROSE, C., PERROTA, D. M., BRUNELL, P. A., and SMITH, G. C. (1979). Cell-free varicella-zoster virus in cultured human melanoma cells. J. Gen. Viral. 43, 15-27. GROSE, C., and FRIEDRICHS,W. E. (1982). lmmunoprecipitable polypeptides specified by varicella-zoster virus. virology 118, 86-95. GROSE,C., JACKSON,W., ~~~TWUGH, J. A. (1989). Phosphorylation of varicella-zoster virus glycoprotein gpl by mammalian casein kinase II and casein kinase I. J. Viral. 63, 3912-3918. GROSE, C. (1990). Glycoproteins encoded by varicella-zoster virus: Biosynthesis, phosphorylation and intracellular trafficking. Annu. Rev. Microbial. 44, 59-80.

HANKS, S. K., QUINN, A. M., and HUNTER.T. (1988). The protein kinase family: Conserved features and deduced phylogeny of the catalytic domains. Science 241, 42-52. HATHAWAY, G. M., and TRAUGH, 1. A. (1983). Casein kinase II. Methods Enzymol. 99, 317-331. KATAN, M., STEVELY,W. S., and LEADER, D. P. (1985). Partial purification and characterization of a new phosphoprotein kinase from cells infected with pseudorabies virus. Eur. /. Biochem. 152, 5765. LEADER, D. P., and PURVES, F. C. (1988). The herpesvirus protein kinase: a new departure in protein phosph’orylation? Trends Biochem.Sci. 151,244-246. LONGNECKER,R., and ROIZMAN, B. (1987). Clustering of genes dispensable for growth in culture in the S component of the HSV-1 genome. Science 236, 573-576. MCGEOCH, D. J., and DAVISON, A. J. (1986). Alphaherpesviruses possess a gene homologous to the protein kinase gene family of eukaryotes and retroviruses. Nucleic Acids Res. 14, 1765-l 777. MCGEOCH, D. J. (1989). The genomes of the human herpesviruses: Contents, relationships, and evolution. Annu. Rev. Microbial. 43, 235-265. PURVES,F. C., KATAN, M., STEVELY,W. S., and LEADER, D. P. (1986). Characteristics of the induction of a new protein kinase in cells infected with herpesviruses. 1. Gen. Viral. 67, 1049-l 057. PURVES, F. C., LONGNECKER,R. M., LEADER, D. P., and ROIZMAN, B. (1987). Herpes simplex virus 1 protein kinase is encoded by open reading frame U,3 which is not essential for virus growth in cell culture. 1. Viral. 61, 2896-2901. PURVES,F. C., SPECTOR,D., and ROIZMAN. B. (1991). The herpes simplex virus 1 protein kinase encoded by the U,3 gene mediates posttranslational modification of the phosphoprotein encoded by the U,34 gene. J. Viral. 65, 5757-5764. ROIZMAN, B. (1991). Herpesviridae: A brief introduction. In “Fundamental Virology” (B. N. Fields and D. M. Knipe, Eds.), 2nd ed., pp. 841-847. Raven Press, New York. RUTHER, U., and MULLER-HILL, B. (1983). Easy identification of cDNA clones. EMBO/. 2, 1791-1794. SMITH, R. F., and SMITH, T. F. (1989). Identification of new protein kinase-related genes in three herpesviruses, herpes simplex virus, varicella-zoster virus, and Epstein-Barr virus. /. Viral. 63, 450455. TRAUGH, 1. A. (1989). Approaches to examine the role of multiple serine protein kinases in the coordinate regulation of cell growth. Adv. Regul. Cell Growth 1, 173-202. WHITE, M. F., and KAHN, C. R. (1986). The insulin receptor and tyrosine phosphorylation. pp. 247-310. In “The Enzymes” (P. D. Boyer and E. G. Krebs. Eds.). Academic Press, Orlando, FL. ZHANG, G., STEVENS,R., and LEADER,D. P. (1990). The protein kinase encoded in the short unique region of pseudorabies virus: Description of the gene and identification of its product in virions and in infected cells. J. Gen. Viral. 71, 1757-1765.