Varicella-Zoster Virus ORF47 Protein Serine Kinase: Characterization of a Cloned, Biologically Active Phosphotransferase and Two Viral Substrates, ORF62 and ORF63

Abstract

Varicella-zoster virus (VZV) codes for a protein serine kinase called ORF47; the herpes simplex virus (HSV) homolog is UL13. No recombinant alphaherpesvirus serine kinase has been biologically active in vitro. We discovered that preservation of the intrinsic kinase activity of recombinant VZV ORF47 required unusually stringent in vitro conditions, including physiological concentrations of polyamines. In this assay, ORF47 phosphorylated two VZV regulatory proteins: the ORF62 protein (homolog of HSV ICP4) and the ORF63 protein (homolog of HSV ICP22). Of interest, ORF47 kinase also coprecipitated ORF63 protein from the kinase assay supernatant.

Varicella-zoster virus (VZV) contains two genes, the ORF47 gene and the ORF66 gene, with classical serine/threonine kinase motifs (26). Two VZV immediate-early proteins, ORF62 and ORF63, are functional homologs of herpes simplex virus (HSV) regulatory proteins ICP4 and ICP22, respectively. ORF62 is the major VZV transcriptional regulatory phosphoprotein during primary infection (22). VZV ORF62 is also present in significant and readily detectable quantities in the viral tegument (12). ORF63 is expressed as a major transcript and protein during VZV latency (2, 3, 9, 10, 13), frequently in conjunction with the transcripts for ORF4, ORF21, ORF29, and ORF62 (2, 10). The ORF63 protein has been shown to be phosphorylated by casein kinase II (CKII) in vitro (30).

Previously, all in vitro data concerning ORF47 have been obtained with infected cell lysate immunoprecipitations or recombinant viruses. With ORF47 and ORF62 proteins purified from VZV-infected MeWo cells, ORF47 phosphorylated ORF62 (18). Under these conditions, the possibility existed that a cellular protein kinase bound to ORF47 or ORF62. In recombinant VZV where stop codons truncated ORF47, ORF47 was not necessary for virus replication in cultured cells and not required for in vivo phosphorylation of VZV ORF63 (1). However, ORF47 is required for virus replication in fetal skin and thymus implants in the SCID-hu mouse model and for efficient replication in human T lymphocytes (15, 27).

Previous attempts to express an active recombinant alphaherpesvirus kinase have been unsuccessful (17). Herein, we define in vitro conditions for the VZV ORF47 protein kinase and document that two VZV regulatory proteins are authentic substrates.

ORF47.12 required polyamines in the protein kinase reaction.

ORF47 has often been compared to CKII, so conditions that increase CKII phosphorylation may stimulate ORF47 kinase activity (4, 7, 18). To investigate the effect of basic molecules on ORF47.12 kinase activity, various polyamines were added at a final concentration of 1 mM to ORF47 in vitro kinase reactions (8). To aid recovery, the 3B3 epitope of VZV gE was inserted into ORF47 (GenBank accession number AAK19253) by PCR mutagenesis, and the new construct was called ORF47.12 (6, 25). The sequence of the ORF47.12 gene was confirmed at the University of Iowa DNA facility. ORF47.12 was subcloned into pCAGGS (19).

HeLa cells (5 × 10⁵) (ATCC CCL2) seeded into 35-mm tissue dishes were transfected with 2 μg of plasmid DNA with Lipofectin (Gibco BRL) and variations (21). Sixteen hours later, after 15 min of pretreatment with 1 ng of okadaic acid (Sigma)/ml, the cells were lysed in radioimmunoprecipitation assay buffer with 50 mM NaF, 1 mM Na₃VO₄, 1 mM PMSF, 1 mM benzamidine, 1 mM leupeptin, and 0.025 trypsin inhibition unit each of aprotinin and I-S soybean trypsin inhibitor (Sigma) and immunoprecipitated with monoclonal antibody (MAb) 3B3 (25, 32). Immunoprecipitates were washed twice with lysis buffer and twice with kinase buffer (17). The reaction mixture contained 25 μl of kinase buffer, 1 μl of a 30 mM stock of the indicated polyamine, and, where indicated, 1 μl of a 125-mg/ml poly-dl-lysine stock (Sigma). Radiophosphate (0.5 μCi of [γ-³²P]ATP; Amersham) initiated the kinase reaction (1 h, 30°C). The reaction was halted by adding ice-cold lysis buffer. After being washed, the samples were boiled for 5 min in reducing sample buffer, electrophoresed on sodium dodecyl sulfate–10% polyacrylamide gel electrophoresis gels (17), and analyzed with an HP InstantImager. Data are presented as counts per minute per square millimeter minus counts per minute per square millimeter of background. The stabilizing conditions described herein were the result of numerous experiments defining exact conditions under which ORF47.12 was an active kinase.

All polyamines stimulated ORF47.12 autophosphorylation above background levels (Fig. 1, lanes 1 to 4). Trivalent spermidine stimulated ORF47.12 kinase autophosphorylation by 18-fold, while the tetravalent spermine increased autophosphorylation by 15-fold. Stimulation of ORF47.12 autophosphorylation was not merely due to the presence of a cation, as Mg²⁺ does not stimulate ORF47 kinase and kinase assays using Ca²⁺ did not stimulate the kinase (data not shown and reference 17). We interpret this data to mean that the polyamines stabilized either the tertiary structure of ORF47.12 or the formation of dimer complexes of ORF47.12, thus facilitating auto- or allophosphorylation. Polyamines are basic cations synthesized in cells from ornithine. In a cell, positively charged polyamines stabilize protein-protein interactions and protein complex formation (8).

FIG. 1.

Polyamines in the ORF47.12 protein kinase assay. Vector pCAGGs-transfected (diagonally striped columns and lower gel slice) or pCAGGs-ORF47.12-transfected (black columns and upper gel slice) lysates were immunoprecipitated with MAb 3B3 and reacted in ORF47 kinase buffer, which is supplemented with manganese (17). The indicated polyamine (lane 1, no polyamines; lane 2, putrescine; lane 3, spermidine; lane 4, spermine) was added to a final concentration of 1 mM in each kinase reaction mixture. The gel slice shows the area at 54 kDa, the molecular mass of ORF47 (17), and the graph above depicts the amount of radioactivity as quantified with the HP InstantImager and presented as counts per minute per square millimeter minus background (bk). In the results shown in the last three lanes of the gel, 1 μl of a 125-mg/ml poly-dl-lysine stock was added to the polyamine-stimulated kinase reactions (lane 5, putrescine and lysine; lane 6, spermidine and lysine; lane 7, spermine and lysine). The first three cells of the table below the gel slice show the fold increase of ORF47.12-transfected radioactivity incorporation in counts per minute per square millimeter minus background above that of vector-transfected lysate (ORF47.12/V). The last three cells show percent decrease due to adding poly-dl-lysine to the polyamine-stimulated kinase reaction mixture, for example, 100% − (putrescine + lysine/putrescine only).

Poly-dl-lysine also stimulates CKII kinase activity (7). When poly-dl-lysine alone was added to the ORF47.12 in vitro kinase reactions, phosphorylation levels remained undetectable (data not shown). When poly-dl-lysine was added to ORF47.12 in vitro kinase reactions with polyamines, autophosphorylation decreased by an average of 48% (Fig. 1, lanes 5 to 7), and the most polyamine-stimulated samples were reduced the most by poly-dl-lysine. Poly-dl-lysine may be competing for the same epitope to which the polyamines bind on ORF47.12, yet the poly-dl-lysine did not stimulate the kinase autophosphorylation. Because poly-dl-lysine is much larger than the polyamine compounds, this reduction in ORF47.12 stimulation may be due to steric hindrance. ORF47.12 is a monomer; therefore, ORF47.12 may not have the conformational flexibility of CKII, which is a tetramer stimulated by poly-dl-lysine.

PCR mutagenesis of the invariant lysine reduced ORF47.12 protein kinase autophosphorylation.

Kinases specify an invariant lysine upstream of the kinase activation domain that is required for kinase activity (26). To definitively attribute the phosphotransferase activity to ORF47, the invariant lysine (codon 169) near the kinase catalytic domain was replaced with an aspartic acid by PCR mutagenesis, and the mutant was designated ORF47.12d (31). VZV gE was selected as a positive control in these experiments because VZV gE coprecipitates with and is heavily phosphorylated by CKII (20). Ten percent of the total lysate was immunoblotted with primary MAb 3B3 with SuperSignal West Pico Chemiluminescent Substrate (Pierce) (data not shown and reference 25). Equivalent amounts of transfected VZV gE, ORF47.12, and ORF47.12d proteins were produced.

As assessed by the in vitro kinase assay, the lysine substitution in ORF47.12d reduced autophosphorylation by 63% compared to autophosphorylation of wild-type ORF47.12 (Fig. 2, lanes 2 and 3). The observed protein kinase activity cannot be attributed to contaminating CKII because (i) kinase activity was not inhibited by heparin (references14 and 17 and data not shown), (ii) CKII requires magnesium for activity (7) and the ORF47 kinase buffer supplied only manganese, and (iii) mutation of the invariant lysine in ORF47 drastically reduced the biological activity of the cloned and expressed kinase.

FIG. 2.

Substitution of the invariant lysine (K169) in the ORF47.12 protein kinase. For both the gels and the graph, HeLa cells were transfected with pCAGGS-vector (lane 1), pCAGGS-ORF47.12 (lane 2), pCAGGS-ORF47.12d (lane 3), or pTM1-VZV gE (lane 4). Ninety percent of each lysate was immunoprecipitated with MAb 3B3 and reacted with [γ-³²P] ATP in an in vitro kinase assay in the presence of 1 mM spermine in a suitable kinase, ORF47, or CKII buffer, which was supplemented with magnesium and separated by sodium dodecyl sulfate–10% polyacrylamide gel electrophoresis (20). The remaining 10% of the whole-cell lysate was immunoblotted with MAb 3B3 (data not shown). The gel slice for lanes 1 to 3 shows that the molecular mass of ORF47.12 and ORF47.12d is 54 kDa, and that for lane 4 shows that the molecular mass of VZV gE is 98 kDa. Radioactivity was quantified with an HP InstantImager. Results are shown as counts per minute per square millimeter minus background (bk) in the same units.

The ORF47.12 protein kinase phosphorylated exogenous, recombinant baculovirus-expressed VZV ORF62.

Previously, wild-type ORF47 immunoprecipitated from VZV-infected cells was found to phosphorylate coimmunoprecipitated VZV ORF62 (18). Expression and purification of VZV ORF62 in Sf21 insect cells by a recombinant baculovirus vector have been described (28). Using the stringent reaction conditions, we added 2 μg of ORF62 to the ORF47.12 in vitro kinase reactions (Fig. 3) and 1 μl of a 1-μg/μl heparin stock (Sigma) where indicated to inhibit any contaminating CKII. A 30-μl aliquot of the kinase reaction supernatant was removed before washing and precipitated in 80% acetone (−20°C).

FIG. 3.

Phosphorylation of the VZV ORF62 protein by ORF47.12 and CKII. HeLa cells were transfected with pCAGGS-vector (lane 1), pCAGGS-ORF47.12 (lane 2), or pTM1-VZV gE (lanes 3 and 4). MAb 3B3 immunoprecipitates were reacted with exogenous baculovirus-produced ORF62. The samples were incubated in suitable kinase buffer (ORF47 buffer or CKII buffer) with spermine and 1 mM heparin where indicated (+hep). The gel slice shows radiolabeled ORF62. The graph above depicts radioactivity in the gel slice as quantified by the HP InstantImager.

The ORF47.12 kinase heavily phosphorylated the full-length ORF62. Phosphorylated ORF62 was recovered from the supernatant at levels of radioactivity 10-fold above background. In control samples, CKII phosphorylated ORF62 at levels threefold higher than those in the ORF47.12 samples. Heparin reduced CKII phosphorylation of ORF62 drastically, as expected. With another set of in vitro kinase assays, we utilized truncation mutants of ORF62 that included only the first 42 amino acids and therefore the potential phosphorylation site at S16. These ORF62 constructs were derived from the first half of the acidic transcription activation domain (23). In these experiments, ORF47.12 failed to phosphorylate the mutants or glutathione S-transferase (GST) alone (data not shown). Thus, this site either was not utilized or was inaccessible to the ORF47 kinase under these conditions.

ORF47.12 protein kinase phosphorylated VZV ORF63.

ORF63 was excised from the plasmid pCMV63 (a gift of Paul Kinchington, University of Pittsburgh), inserted into pGEX-4T-1 (Pharmacia), expressed in Escherichia coli, and purified by glutathione-Sepharose 4B affinity chromatography (Pharmacia). A 2-μg aliquot of GST-ORF63 was added to ORF47.12 in vitro kinase assays.

ORF47.12 phosphorylated GST-ORF63 and, even after extensive washing, precipitated GST-ORF63 from the kinase reaction supernatant and retained it in the pellet (Fig. 4). Indeed, 71% of the total radiolabeled ORF63 was found in the reaction pellet (lane 2). CKII also phosphorylated GST-ORF63 (lane 6), though phosphorylation observed with CKII was substantially less than that observed with ORF47.12. The addition of 1 or 2 mM heparin to the CKII in vitro kinase reactions reduced CKII phosphorylation of GST-ORF63 by 73 and 72%, respectively (lanes 7 and 8).

FIG. 4.

Phosphorylation of the VZV ORF63 protein by ORF47.12 and CKII. HeLa cells were transfected with pCAGGS-vector (lanes 1 and 5), pCAGGS-ORF47.12 (lanes 2 to 4), or pTM1-gE (as a source of CKII) (lanes 6 to 8). MAb 3B3 immunoprecipitates were reacted with 2 μg of GST-ORF63 (all lanes) and 1 mM (lanes 3 and 7) or 2 mM (lanes 4 and 8) heparin. The upper gel slice shows ORF63 in the reaction mixture supernatant, while the lower gel slice shows ORF63 retained in the kinase reaction pellet. (Lanes 1 through 4) The horizontally striped columns on the graph show ORF47.12-phosphorylated GST-ORF63 in the kinase reaction mixture supernatant. The solid black columns show the amount of ORF47.12-phosphorylated GST-ORF63 precipitated by the kinase even after extensive washing. Lane 1 (negative control) was reacted under ORF47 kinase conditions. (Lanes 5 through 8) The gray areas of the columns show CKII-phosphorylated GST-ORF63 in the kinase reaction mixture supernatant. The pixilated columns show the amount of CKII-phosphorylated GST-VZV ORF63 in the kinase reaction pellet. Lane 5 (negative control) was reacted under CKII kinase conditions. The table shows the amounts of incorporated radioactivity as quantified with the HP InstantImager.

When 1 or 2 mM heparin was added to ORF47.12 in vitro kinase assays to inhibit any contaminating CKII, ORF47.12 phosphorylation of GST-ORF63 increased (Fig. 4, lanes 3 and 4). We hypothesize that the ORF47.12 was allowed greater access to the exogenous GST-ORF63, and GST-ORF63 phosphorylation due to ORF47.12 increased by 2.5-fold. Also, in the presence of 1 mM heparin, the coprecipitation of GST-ORF63 by ORF47.12 increased from 71 to 75%. When heparin was increased to 2 mM, phosphorylation of the exogenous GST-ORF63 was very nearly the same as for the 1 mM sample, though more of the phosphorylated GST-ORF63 was retained in the kinase reaction pellet (Fig. 4).

Using this in vitro system that preserved the biological activity of cloned ORF47.12, we determined with certainty that the VZV ORF62 protein (homolog of HSV ICP4) and the ORF63 protein (homolog of HSV ICP22) were authentic viral substrates of the ORF47.12 protein kinase. Phosphorylation-dephosphorylation events have long been considered to be important modifications in transcriptional regulatory proteins (5). In addition, the observation that ORF47 bound so tightly to ORF63 that ORF47 precipitated ORF63 from the kinase reaction supernatant was an unusual finding. Although kinases may often be identified by precipitation of substrate with concurrent coprecipitation of kinase, only rarely can phosphorylated substrates be identified by precipitation of the relevant kinase (24). Typically, the substrate is released immediately after phosphorylation. Thus, the latter result suggested that ORF47 and its substrate ORF63 existed as a complex. Subsequent experiments with addition of exogenous heparin demonstrated that ORF47 and the cellular kinase CKII competed for ORF63 binding, a possible regulatory interaction.

The fact that ORF62, ORF63, and ORF47 are present in the viral tegument is of great interest because HSV phosphorylation analyses have demonstrated that protein phosphorylation and release from the tegument occur concurrently (11, 12, 16, 29). Presumably, therefore, at least one functional site of the ORF47 kinase is within the tegument. Finally, these studies add more information about the likely phylogeny of the herpesvirus protein serine/threonine kinases and their relatedness to mammalian protein kinases (5).