BAG3, a Host Cochaperone, Facilitates Varicella-Zoster Virus Replication

Abstract

Varicella-zoster virus (VZV) establishes a lifelong latent infection in the dorsal root ganglia of the host. During latency, a subset of virus-encoded regulatory proteins is detected; however, they are excluded from the nucleus. ORF29p, a single-stranded DNA binding protein, is one of these latency-associated proteins. We searched for cell proteins that interact with ORF29p and identified BAG3. BAG3, Hsp70/Hsc70, and Hsp90 colocalize with ORF29p in nuclear transcription/replication factories during lytic replication of VZV. Pharmacological intercession of Hsp90 activity with ansamycin antibiotics or depletion of BAG3 by small interfering RNA results in inhibition of virus replication. Replication in BAG3-depleted cell lines is restored by complementation with exogenous BAG3. Alteration of host chaperone activity provides a novel means of regulating virus replication.

As obligate intracellular parasites, viruses replicate inside living cells using the metabolic machinery of the host. Because of their limited coding potential, viruses have evolved means to hijack the host cell machinery and exploit it to their advantage.

Varicella-zoster virus (VZV) is an alphaherpesvirus that invades the dermis and epidermis during primary lytic infection and causes chicken pox (varicella). After viral DNA replication and production of infectious progeny, the virus moves into the dorsal root ganglia to establish latency (2, 3, 12, 48). Although DNA replication, late gene expression, and virion production cease during latency, a limited repertoire of immediate-early and early gene products, including the protein products of open reading frames (ORFs) 4, 21, 29, 62, 63, and 66, are still detected, and these latency-associated proteins (LAPs) are aberrantly localized in the cytoplasm (13-16, 19, 32, 37, 49, 56).

Although VZV can infect both epidermal and neuronal cells, the outcome of infection and the localization of a subset of viral proteins, the LAPs, depend on the infected cell type. This suggests that cell-specific pathways, in addition to the interactions between cellular and viral proteins, can determine whether virus replication is lytic or latent.

We and others have suggested that in response to certain stimuli, the LAPs translocate from the cytoplasm to the nucleus and the virus exits latency. This switch is likely regulated by a change in the host cell milieu. Therefore, elucidation of the molecular mechanisms governing nuclear targeting and exclusion of the LAPs and unraveling of their web of interactions with host cell proteins may shed light on the mechanisms controlling the replication and reactivation processes of the virus.

VZV ORF29 encodes a single-stranded DNA binding protein (ORF29p) that is assumed to function during viral DNA replication. Similar to the other LAPs, ORF29p is predominantly nuclear during lytic infection and reactivation but is excluded from the nucleus during latency (39, 49). Nuclear import of ORF29p occurs in the absence of other VZV-encoded proteins via its nonclassical nuclear localization signal (NLS) (71), and the localization of the protein is cell type specific and correlates with its stability (70). These observations suggest that the host affects the rate of ORF29p degradation and alters its localization pattern. However, the molecular mechanisms governing these processes are unknown.

In eukaryotic cells the ubiquitin-proteasome system is the main pathway for recycling polypeptides and eliminating misfolded or mutated proteins (34). It is well established that the native state of newly synthesized and stress-denatured proteins is attained by the ATP hydrolysis-driven function of molecular chaperones or heat shock proteins (27). The same machinery is also used for the recognition of folding-incompetent proteins that should be polyubiquitinated and targeted for degradation. Thus, there is a functional link between the folding activity of molecular chaperones and proteasomal degradation.

These evolutionarily conserved activities of chaperones also augment DNA replication. DnaK and DnaJ from Escherichia coli were first identified as proteins that are necessary for bacteriophage DNA replication (80). They specifically associate with a multicomponent preinitiation replication complex and are required for the initiation of DNA replication (81). In addition, eukaryotic Hsp70 interacts with Orc4p of Saccharomyces cerevisiae to prevent oligomerization of its N-terminal domain (29). Components of the replication machinery of animal DNA viruses, including the polyoma virus T antigen and papillomavirus helicase replication initiator protein E1, associate with mammalian chaperone proteins (10, 46). However, unlike bacteria, in these cases the functional requirements for chaperone and cochaperone proteins during host and virus replication remain unknown.

Several cell proteins are known to interact with the chaperone-client protein complexes and alter their function. Among these, members of the BAG family of proteins were shown to interact through their BAG domain with the N-terminal ATPase domain of Hsp70/Hsc70 (75), affecting the rate of ATP/ADP exchange and regulating their chaperone activity (4, 35).

In this report, we identify novel interactions of ORF29p with host proteins. The roles of these interactions in the degradation and localization of ORF29p and the ability of VZV to grow in cultured cells are addressed. We provide biochemical evidence that ORF29p interacts with the cochaperone BAG3 and forms a complex with at least BAG3 and Hsp70/Hsc70 in vivo in both transiently transfected and VZV-infected cells. Our results reveal that the ATPase activity of Hsp90 is required for stabilization and nuclear localization of ORF29p and virus replication. We show that VZV redistributes BAG3 and its partners Hsp70 and Hsp90 into nuclear replication/transcription foci in infected cells, suggesting that the virus exploits the highly conserved functions of the host heat shock proteins to efficiently complete its life cycle. Finally, we provide genetic evidence that BAG3 is required for efficient virus growth. In contrast, while herpes simplex virus (HSV) replication is also inhibited by ansamycins (7), its replication in MeWo cells depleted of BAG3 is unaffected. We propose that regulators of chaperone protein activity modulate VZV replication, raising the possibility that this pathway is required for controlling replication of animal viruses.

MATERIALS AND METHODS

Mammalian cells.

Human melanoma (MeWo) and human 293T fibroblasts were maintained and infected as previously described (70, 71).

To generate stable cell lines expressing small interfering RNAs (siRNAs) targeting BAG3 mRNA, MeWo cells were infected with retroviruses and selected in growth medium containing 1 μg/ml puromycin. Single-cell colonies were selected, expanded, and used for further analysis.

Transfections.

All transfections were performed using Lipofectamine PLUS in Opti-MEM medium (Invitrogen, Carlsbad, CA).

Viruses. (i) VZV.

Jones VZV, a wild-type clinical isolate, was propagated as described previously (33). Cell-free virus was obtained by infecting confluent monolayers of MeWo cells in 100-mm-diameter dishes. When cytopathic effect was present, cells were washed three times with cold phosphate-buffered saline (PBS; 1 mM KH₂PO₄, 10 mM Na₂HPO₄, 137 mM NaCl, 2.7 mM KCl, pH 7.4) and incubated with 0.1% EDTA in PBS for 5 min. The cells were detached from the plate by pipetting and were harvested by centrifugation at 500 × g for 5 min at 4°C. The cell pellet was resuspended in 1 ml per 100-mm dish of fresh phosphate sucrose glutamate calf serum buffer [5% sucrose, 0.1% l-(+)-glutamic acid, 10% fetal calf serum in PBS] and sonicated for three 30-s intervals. Cell debris was removed by centrifugation at 250 × g for 2 min at 4°C, and the supernatant was filtered through a 5-μm-pore-size filter as described previously (38). Cell-free virus was stored at −150°C, and the titer was determined by plaque assay on MeWo cells.

(ii) Adenoviruses.

Adenovirus AdBAG3 expressing flag-tagged BAG3 under the HSV-1 thymidine kinase promoter was constructed using pCK-mFLAG-BAG3 and a pBHGfrtΔE1,E3FLP system (Microbix Biosystems, Toronto, Ontario, Canada) (59). Adenovirus AdORF29 is described elsewhere (70). Adenovirus AdORF63, expressing VZV ORF63p, was constructed by M. S. Walters (M. S. Walters and S. J. Silverstein, unpublished data).

(iii) Retroviruses.

Retroviruses were constructed by transient cotransfection of 293T cells with the proviral vector pSuper.retro.puro (6), pCK-siBAG3-737, or pCK-siBAG3-2235 and pgag-polgpt (52) and pHCMV-G (79).

(iv) HSV.

HSV was grown and titrated as previously described (60).

Preparation of VZV and cell DNA.

VZV nucleocapsids were prepared as described previously (72). Virus DNA was isolated from nucleocapsids suspended in TE (10 mM Tris-HCl, pH 8.0, 1 mM EDTA) and digested in Proteinase K cocktail (100 mM NaCl, 10 mM EDTA, 50 mM Tris-HCl, pH 7.4, 200 μg/ml proteinase K) at 50°C for 3 h. Virus DNA was extracted with phenol and precipitated with isopropanol.

MeWo DNA was prepared using the DNAZOL reagent (Invitrogen, Carlsbad, CA) by following the manufacturer's instructions.

Drug treatment.

Cells were treated with geldanamycin or 17-N, N-dimethyl ethylene diamine-geldanamycin (17DMAG) (InvivoGen, San Diego, CA) from 1-mg/ml stocks in dimethyl sulfoxide (DMSO) and water, respectively, at the concentrations and times indicated in the figure legends.

Plasmid construction. (i) BAG3 plasmids.

The 3′ region of BAG3 cloned in the cDNA library was amplified with Pfu Turbo polymerase (Stratagene, La Jolla, CA) using the cDNA phagemid as template and the primers M13rev (5′-GAGCGGATAACAATTTCACACAGG-3′) and 3′Hind-BAG3 (5′-GGAAGCTTTACAGGGCAGAGGCTACGGTG-3′). The PCR product was cloned in pCR2.1 TOPO TA to yield pCK-BAG3par. The 5′ region of BAG3 cDNA was amplified from MeWo total RNA with the C. Therm. Polymerase One-Step RT-PCR System (Roche, Mannheim, Germany) using the primers BAG3-5-F (5′-CAGACCCCAACCCAGCATGAG-3′) and BAG3-5-R (5′-CCGCTGCCACCTGTCCACAC-3′) and cloned in pCR2.1 TOPO TA to yield pCK-BAG3-5. pCK-BAG3fl was constructed by digesting pCK-BAG3-5 with PpuMI and XhoI and ligating the released fragment to PpuMI- and XhoI-digested pCK-BAG3par. The full-length BAG3 cDNA was amplified from pCK-BAG3fl using Pfu Turbo polymerase, 5-Bam-BAG3 (5′-GGGGATCCAGCATGAGCGCCGCCACCCA-3′), and 3-Hind-BAG3 and cloned in pCR2.1 TOPO TA to yield pCK-BAG3.

An EcoRI/HindIII digestion fragment from pCK-Bag3par was cloned into EcoRI/HindIII-digested pALEX (61) to yield pCK-GST-BAG3par. pCK-GST-BAG3par was digested with SalI/XhoI and self-ligated to create pCK-GST-BAG3ag. pCK-GST-BAG3 was prepared by cloning an EcoRI and HindIII digestion fragment from pCK-BAG3fl into EcoRI- and HindIII-digested pALEX.

pCK-FLAG-BAG3 was constructed by cloning a BamHI/HindIII digestion fragment from pCK-BAG3 into BamHI/HindIII-digested pCMV-Tag2B (Stratagene). A NotI-digested, Klenow-filled, and HindII-digested fragment from pCK-FLAG-BAG3 was ligated to EcoRI-digested, Klenow-filled, and HindIII-digested pDC516 (Microbix Biosystems Inc.) to yield pCK-mFLAG-BAG3. The mCMV promoter of this plasmid was replaced by the 105 bp of the HSV-1 thymidine kinase promoter by ligating a BamHI/BglII-digested and Klenow-filled-in fragment from pLS115/105 (24, 55) into XbaI/NcoI-digested and Klenow-filled pCK-mFLAG-BAG3 to yield pCK-tkFLAG-BAG3.

(ii) BAG3 siRNA plasmids.

The siRNA oligonucleotides targeting BAG3 mRNA were designed using SVM RNA interference (RNAi) 3.6 (http://www.changbioscience.com/stat/sirna.html), which uses a collection of rules to predict functional siRNAs (30, 36, 64). To generate pCK-siBAG3-737 and pCK-siBAG3-2235, the annealed oligonucleotide pair 737-TS (5′-GATCCCCCCACTCAGCCAGATAAACATTCAAGAGATGTTTATCTGGCTGAGTGGTTTTTGGAAA-3′) and 737-BS (5′-AGCTTTTCCAAAAACCACTCAGCCAGATAAACATCTCTTGAATGTTTATCTGGCTGAGTGGGGG) or 2235-TS (5′-GATCCCCGAAGTTGCTTGTTGTTTGATTCAAGAGATCAAACAACAAGCAACTTCTTTTTGGAAA-3′) and 2235-BS (5′-AGCTTTTCCAAAAAGAAGTTGCTTGTTGTTTGATCTCTTGAATCAAACAACAAGCAACTTCGGG-3′) was ligated into pSuper.retro.puro (6).

(iii) ORF29p plasmids.

pZErO29, pET29, and pET29(1-345) were described previously (71). pCK-X29(346-1203) was constructed by releasing an EcoRV/NotI fragment from pZErO29 and ligating it into BamHI-digested, Klenow polymerase-filled, and NotI-digested pET-21c(+) (Novagen).

(iv) ORF63p plasmids.

Full-length ORF63 was amplified by PCR using the oligonucleotides 5-Eco-63 (5′-GGGAATTCATGTTTTGCACCTCACCGGCT-3′) and 3-Xho-63 (5′-GGCTCGAGTAAAGACTTCACGCCATGGGG-3′), Pfu Turbo polymerase, and VZV strain Jones DNA as template. The PCR product was cloned in pCR2.1 TOPO TA to create pCK-ORF63. pCK-X-ORF63 expressing ORF63 with a 6-His tag was constructed by digesting pCK-ORF63 with EcoRI and XhoI and cloning the resulting DNA fragment in EcoRI- and XhoI-digested pET-21a(+) (Novagen).

(v) HSP70 plasmids.

A full-length Hsp70 gene (HSPA1A) was amplified from MeWo DNA using Thermococcus kodakaraensis DNA polymerase (Novagen, San Diego, CA) and 5-Hind-Hsp70 (5′-GGAAGCTTAGAGAGCAGCGAACCGGCAT-3′) and 3-Xho-Hsp70 (5′-CGCTCGAGTTGGAAAGGCCCCTGATCTAC-3′). The PCR product was cloned in pCR2.1 TOPO TA to yield pCK-HSP70. Hsp70 was excised as a HindIII/XhoI fragment and subcloned into HindIII/XhoI-digested pALEX to create pCK-GST-HSP70, which expresses Hsp70 as a glutathione S-transferase (GST) fusion under the control of the T7 promoter.

All primers used were manufactured by Proligo LLC (Boulder, CO), and all vector inserts were verified by DNA sequencing.

Antibodies.

Rabbit polyclonal antibodies against amino acids (aa) 1086 to 1201 of ORF29p and aa 1 to 265 of ORF63p were described previously (49).

A portion of the BAG3 cDNA (pCK-GST-BAG3ag) encoding aa 134 to 299 was cloned in the bacterial expression vector pALEX (61). The GST fusion protein was overexpressed in E. coli, strain BL21(DE3), and purified to apparent homogeneity by affinity chromatography on a glutathione Sepharose column (68). The protein was used to immunize rabbits, and BAG3-specific antibodies were purified by affinity chromatography using BAG3 immobilized on cyanogen bromide-activated Sepharose 4B after removal of the GST tag (Amersham Biosciences, Uppsala, Sweden).

Mouse monoclonal antibodies to VZV gE and ORF62p were from ViroStat (Portland, ME). Mouse monoclonal antibodies to HSP90α/β and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were from Santa Cruz Biotechnology (Santa Cruz, CA), and antibodies to HSP70 were from United States Biological (Swampscott, MA). Alexa Fluor 488-conjugated anti-mouse and Alexa Fluor 546-conjugated anti-rabbit antibodies were from Molecular Probes (Carlsbad, CA). Goat anti-rabbit and anti-mouse antibodies conjugated to horseradish peroxidase for immunoblotting were from KPL (Gaithersburg, MD). HSV gC antibody was purchased from the Rumbaugh-Goodwin Institute (Plantation, FL). Antibody to ICP0 has been described already (47).

Indirect immunofluorescence (IF) microscopy.

Cells on glass coverslips were fixed and stained with antibody and Hoechst as previously described (70, 71). All samples were visualized with a Zeiss Axiovert 200 M inverted microscope (Carl Zeiss Microimaging Inc., Thornwood, NY), and images were acquired with a Hamamatsu C4742-80-12AG digital CCD camera (Hamamatsu Photonics, Hamamatsu City, Japan) using Openlab 5 software (Improvision, Lexington, MA). Images were deconvolved when necessary using Openlab 5 and assembled in Photoshop CS (Adobe Systems, San Jose, CA).

SDS-PAGE and Western blotting.

Infected or transfected cells were washed twice with cold PBS, scraped from tissue culture dishes, resuspended in radioimmunoprecipitation lysis buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% NP-40, 0.5% deoxycholate, 0.1% sodium dodecyl sulfate [SDS], 50 mM NaF) plus Complete protease inhibitor cocktail (Roche, Mannheim, Germany), and incubated on ice for 30 min. The lysate was clarified by centrifugation at 22,500 × g for 10 min in a Tomy MX-160 high-speed refrigerated microcentrifuge. Total protein concentration was measured using the Bio-Rad protein assay (Hercules, CA) (5). The appropriate amount of 5× SDS sample buffer (250 mM Tris-HCl, pH 6.8, 500 mM dithiothreitol, 10% SDS, 0.5% bromophenol blue, 50% glycerol) was added to the samples before boiling for 10 min and SDS-polyacrylamide gel electrophoresis (PAGE) analysis (41). The proteins were transferred to nitrocellulose membranes with a Bio-Rad Semi-Dry apparatus before Western blotting. After blocking the membrane in 5% nonfat milk in PBS containing 0.1% Tween 20 (PBST), immobilized proteins were reacted with ORF29p or BAG3 antibodies at a 1:1,000 dilution and HSP90, HSP70, or GAPDH antibody at a 1:2,000 dilution in 1% nonfat milk in PBST. The membrane was washed three times for 5 min each with PBST, incubated with an anti-rabbit or anti-mouse antibody conjugated to horseradish peroxidase at a 1:5,000 dilution, washed again three times for 5 min with PBST, and washed twice with PBS, and antibodies were visualized by addition of the LumiGLO substrate (KPL) and exposure to X-ray film.

Far-Western blotting.

GST, GST-BAG3, GST-BAG3par, GST-BAG3ag, and GST-Hsp70 were overexpressed in E. coli, strain BL21(DE3), and purified by affinity chromatography on a glutathione Sepharose column (68). The concentration of the purified proteins was determined, and the proteins were subjected to SDS-PAGE and transferred to nitrocellulose membranes using a Bio-Rad Semi-Dry apparatus. Proteins on the membrane were washed once for 10 min with TBST (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.05% Triton X-100). The membranes were then blocked for 2 h at room temperature with HBB (20 mM HEPES [pH 7.4], 5 mM MgCl₂, 1 mM KCl) containing 5% nonfat milk, incubated for 2 h with HBB containing 1% nonfat milk and 20 μl of in vitro-translated ORF29p, ORF29p(1-345), ORF29p(346-1203), and ORF63p, washed three times with TBST for 10 min at room temperature, dried, and exposed to a phosphorimager screen and X-ray film.

Immunoprecipitation.

Infected, transfected, or radioactively labeled cells were harvested and lysed as described above. Proteins were immunoprecipitated overnight at 4°C with 25 μl of GammaBind Plus Sepharose beads (Amersham Biosciences, Piscataway, NJ) conjugated with BAG3 or ORF29 antibody or 25 μl of anti-FLAG M2 agarose matrix (Sigma). The beads were collected by centrifugation at 400 × g at 4°C in a Tomy MX-160 high-speed refrigerated microcentrifuge and washed five times for 5 min each with radioimmunoprecipitation assay (RIPA) buffer at 4°C. Bound complexes were released from the beads by boiling for 10 min in 50 μl 1.5× SDS sample buffer. The released proteins were subjected to SDS-PAGE.

Pulse-chase labeling.

Proteins were radiolabeled after washing the cell cultures three times with PBS and incubation in Met⁻, Cys⁻ Dulbecco's modified Eagle's medium (DMEM) (GIBCO-BRL) for 30 min. Starvation medium was replaced with labeling medium (modified DMEM supplemented with 1% dialyzed calf serum and 500 μCi/ml trans-³⁵S label (ICN, Irvine, CA). After a 1-h pulse, cells were washed twice with chase medium (normal DMEM supplemented with 10% fetal calf serum, 2 mM Met, and 4 mM Cys) and chased for the indicated time periods. Cell lysates were prepared and ORF29p was immunoprecipitated. The bound material was subjected to SDS-PAGE. The gel was dried and exposed to X-ray film. ORF29p levels were quantified with ImageJ (http://rsb.info.nih.gov/ij/; NIH).

In vitro translation.

[³⁵S]Met-labeled ORF29p, ORF29p(1-345), ORF29p(281-1203), ORF29p(346-1203), and ORF63p were synthesized by coupled in vitro transcription and translation using the TNT coupled reticulocyte system (Promega, Madison, WI) with pET29, pET29(1-345), pCK-X29(346-1203), or pCK-X-ORF63.

RNA isolation and cDNA expression library construction.

mRNA extracted from MeWo cells was used to construct an expression cDNA library in the lambda ZAP vector (66). Total RNA was isolated from MeWo cells using the TRIzol reagent (Invitrogen). Poly(A)⁺ RNA was purified from total RNA using the Oligotex mRNA purification protocol (QIAGEN, Valencia, CA). Five micrograms of mRNA was used to construct a cDNA library using the ZAP Express cDNA cloning kit (Stratagene, La Jolla, CA) according to the manufacturer's instructions. The cDNAs were cloned in lambda ZAP (66), and phage DNA was packaged using the Gigapack III Gold packaging extract (Stratagene). The library was titrated and amplified once in Escherichia coli, strain XL1-Blue MRF′.

Screening of cDNA library and expression cloning.

The cDNA library was screened using a modification of methods previously described (50, 51, 67). Approximately 10⁶ PFU were used to infect E. coli, strain XL1-Blue MRF′, and then plated on 20 NZY agar petri dishes (150 mm). The plates were incubated for 4 h at 42°C and then overlaid with nitrocellulose filters impregnated in 10 mM isopropyl-β-d-thiogalactopyranoside and incubated overnight at 37°C. The filters were then removed and washed for 15 min with TBST, blocked with HBB buffer with 5% nonfat milk for 8 h at 4°C, and then incubated with 0.8 ng/ml of in vitro-translated [³⁵S]Met-ORF29p in HBB with 1% nonfat milk overnight at 4°C. All membranes were washed three times with TBST for 10 min at room temperature, dried, and exposed to phosphorimager screens. Positive clones were picked, and those phages demonstrating enrichment on subsequent screenings were plaque purified. The cDNA inserts were excised in vivo as phagemids.

RESULTS

BAG3 is a cell protein that interacts with ORF29p.

Host proteins that interact with ORF29p were identified from a bacteriophage lambda cDNA expression library constructed using polyadenylated RNA from actively growing human fibroblasts (MeWo cells). Screening of 10⁶ clones of the library for binding to in vitro-translated [³⁵S]methionine-labeled ORF29p yielded three positive clones that were plaque purified and used to extract phagemids. DNA sequencing of the phagemids revealed that all three clones coded for aa 134 to 575 of BAG3.

BAG3 is a predominantly cytoplasmic, 74-kDa member of an evolutionarily conserved family of proteins that contain at least one BAG domain that is responsible for binding to the ATPase domain of Hsp70/Hsc70 (76) and Bcl-2 (42). Via this interaction, these proteins can modulate the activity of the aforementioned chaperones and are thus characterized as cochaperones (23, 74, 76).

ORF29p interacts with BAG3 in vitro.

To verify that ORF29p interacts with BAG3 in vitro and to exclude the possibility that the known BAG3 interaction partner Hsp70 (HSPA1A) bridged the interaction between the proteins, we performed a far-Western blot. Purified GST, GST-BAG3, GST-BAG3par, GST-BAG3ag (Fig. 1A), and GST-Hsp70 were subjected to SDS-PAGE, transferred to nitrocellulose membranes, and renatured in situ (Fig. 1C). The membranes were probed with in vitro-translated ORF29p, ORF29p(1-345), ORF29p(346-1203), or ORF63p, which is another VZV LAP (Fig. 1B and D). Full-length ORF29p, but not the fragment of the protein that contains the N-terminal NLS, interacted with full-length BAG3 (Fig. 1E and F). In contrast, the C-terminal fragment that lacks the NLS was bound by BAG3 (Fig. 1G). BAG3ag lacks the BAG domain and was not bound by either ORF29p or ORF29p(346-1203) (Fig. 1E and G). However, ORF29p did not interact with Hsp70 in this assay, demonstrating that HSPA1A does not bridge ORF29p and BAG3 (Fig. 1E). We cannot exclude the possibility, though, that another rabbit reticulocyte protein has this role. Finally, ORF63p did not interact with BAG3 or Hsp70, showing that association with BAG3 is not a general property of the LAPs (Fig. 1H).

FIG. 1.

ORF29p and BAG3 interact in vitro. (A and B) Schematic diagram of BAG3 and ORF29p. (C) GST, GST-BAG3, GST-BAG3par, GST-BAG3ag, and GST-Hsp70 were overexpressed in E. coli and purified using affinity chromatography. Four micrograms of each protein was analyzed by SDS-PAGE and stained with Coomassie brilliant blue R-250. (D) Plasmids encoding ORF29p, ORF29p(1-345), ORF29p(346-1203), and ORF63p were used in coupled transcription/translation reactions to prepare [³⁵S]Met-labeled proteins. Two microliters of each reaction was analyzed by SDS-PAGE and autoradiography. (E to H) One microgram of each purified protein was subjected to SDS-PAGE, transferred to nitrocellulose membranes, renatured in situ, and blotted with (E) ORF29p, (F) ORF29p(1-345), (G) ORF29p(346-1203), or (H) ORF63p. The membranes were dried and exposed to X-ray film. ssDNA, single-stranded DNA.

ORF29p interacts with BAG3 in vivo.

We next asked if ORF29p and BAG3 interact in vivo. 293T cells were transiently transfected with plasmid constructs expressing flag-tagged BAG3, ORF29p, or both. Forty-eight hours posttransfection the cells were lysed, equal amounts of total protein were incubated with anti-flag M2 affinity matrix, and the bound material was subjected to SDS-PAGE. Western blotting with anti-ORF29p-specific antibodies demonstrated that ORF29p was only immunoprecipitated from cells cotransfected with plasmids expressing both Flag-tagged BAG3 and ORF29p (Fig. 2A).

FIG. 2.

ORF29p interacts with BAG3 and Hsp70/Hsc70 in vivo. (A) 293T cells were transiently transfected with plasmids expressing ORF29p, flag-tagged BAG3, or both. Equal amounts of total protein were incubated with anti-flag M2 matrix, and the bound material was analyzed by Western blotting with anti-ORF29p antibodies. (B) Confluent MeWo cells were either mock infected or infected with cell-free VZV at a multiplicity of infection of 0.01. Three days postinfection, cells were lysed and equal amounts of total protein were incubated with beads containing anti-BAG3 or anti-ORF29p antibodies. Western blotting of the bound materials was performed with anti-ORF29p or anti-Hsp70/Hsc70 antibodies. IP, immunoprecipitation.

To provide evidence that endogenous BAG3 interacts with ORF29p expressed during virus infection, confluent MeWo monolayers were infected with cell-free VZV. Three days postinfection the infected cells were lysed, and equal amounts of total protein were incubated with antibodies specific for BAG3. Analysis of the bound complexes by Western blotting with an anti-ORF29p antibody revealed that ORF29p was only present in the bound material from infected cells (Fig. 2B). Consistent with previous reports, Hsp70/Hsc70 but not Hsp90 were present in the complexes precipitated with the anti-BAG3 antibody (Fig. 2B and data not shown) (22). Moreover, immunocapture with an anti-ORF29p antibody showed that Hsp70/Hsc70, the known binding partners of BAG3, are among the bound proteins in VZV-infected cells (Fig. 2B). These results suggest that, during VZV infection, complexes composed of ORF29p, BAG3, and Hsp70/Hsc70 are formed.

ORF29p is an Hsp90 client for proteasomal degradation.

As discussed in the introduction, besides peptide folding, heat shock proteins can also control protein turnover (53, 54). Hsp90 is thought to be a key regulator of this balance between folding and polypeptide turnover, depending on its association with other chaperone and cochaperone proteins (58, 62). BAG3 can abrogate protein degradation mediated by Hsp70-Hsp90 chaperone complexes (22). Because ORF29p interacts with BAG3, we asked if this pathway controls ORF29p levels. Geldanamycin and other ansamycin antibiotics interact with and inhibit the ATPase domain of Hsp90, shifting the balance of Hsp90 activity from protein folding to degradation of its clients (65, 78). Therefore, we asked if geldanamycin affected the stability of ORF29p. The t_1/2 of ORF29p in MeWo cells infected with an adenovirus expressing ORF29p was examined by pulse-chase analysis in untreated and drug-treated cells. Twenty-four hours postinfection (hpi), DMSO or geldanamycin was added to the growth media. After 36 h the infected cells were washed, starved for 30 min in labeling medium containing DMSO or geldanamycin, and then labeled with ³⁵S for 1 h. The labeled proteins were chased for various times, and cell extracts were prepared and normalized for total protein concentration before capture with antibody to ORF29p bound to agarose beads. Bound proteins were subjected to SDS-PAGE and visualized by autoradiography (Fig. 3). Protein levels were determined relative to the 0-h chase point. In DMSO-treated cells, the t_1/2 of ORF29p exceeded the 6-h chase time point. However, the kinetics of protein degradation was faster and ORF29p was significantly less stable (t_1/2 ≤ 1 h) in geldanamycin-treated cells. Thus, the ATPase activity of Hsp90 is required for stabilization of ORF29p.

FIG. 3.

Half-life of ORF29p. MeWo cells were infected with an adenovirus expressing ORF29p at a multiplicity of infection of 10. At 24 hpi, DMSO or 1.5 μM geldanamycin (GM) was added to the medium. At 36 hpi, cells were labeled with 500 μCi/ml trans-³⁵S label for 1 h in the presence of DMSO (lanes 2 to 5) or geldanamycin (lanes 6 to 9). The labeling medium was replaced with DMEM supplemented with 10% fetal calf serum, 2 mM methionine, 4 mM cysteine, and DMSO or 1.5 μM geldanamycin for the indicated chase times. Cells were harvested and lysed, and ORF29p levels were determined following immunoprecipitation and SDS-PAGE. Proteins were visualized by autoradiography, and band intensity was quantified using ImageJ. The percentage of ORF29p was calculated relative to the amount present at the 0-h chase time point.

ORF29p localizes to the cytoplasm of geldanamycin- and 17DMAG-treated MeWo cells.

We previously showed that the intracellular localization of ORF29p correlated with its stability. The t_1/2 of the protein is shorter in cell lines where it is localized in the cytoplasm, such as U373MG, compared to cell lines, such as MeWo cells, where it localized in the nucleus (70). Therefore, we asked if the decreased half-life caused by geldanamycin treatment prevented nuclear localization of ORF29p in MeWo cells.

MeWo cells were infected with adenovirus expressing ORF29p. After adsorption, the medium containing the virus was removed and replaced with medium containing DMSO, geldanamycin, or 17DMAG (a water-soluble analog). The cells were fixed at 48 hpi, and ORF29p localization was examined. Following DMSO treatment, ORF29p remained in the nuclei of MeWo cells (Fig. 4A). However, in the presence of the drugs the number of cells with detectable ORF29p was significantly decreased, and the protein was excluded from the nucleus of the few cells where expression was detected (Fig. 4B and C). Similar results were obtained with shorter drug treatment.

FIG. 4.

Localization of ORF29p and ORF63p in cells treated with ansamycin antibiotics. MeWo cells were infected with an adenovirus expressing ORF29p (A to C) or ORF63p (D to F). After virus adsorption, medium was replaced with DMEM supplemented with 2% fetal calf serum and DMSO (A and D), 1.5 μM geldanamycin (GM) (B and E), or 0.5 μM 17DMAG (C and F). At 48 hpi, cells were fixed and ORF29p or ORF63p was detected by indirect IF microscopy after reaction with specific antibodies. Images were viewed and captured using a 63× objective.

To demonstrate that nuclear exclusion resulting from drug-induced inhibition of Hsp90 ATPase activity was specific for ORF29p and not from a nonspecific blockade of the cell's nuclear import machinery, the localization of ORF63p was examined. MeWo cells were infected with an adenovirus expressing ORF63p, treated as described above, and fixed, and the protein localization was examined (Fig. 4D to F). Unlike ORF29p, geldanamycin or 17DMAG did not affect the number of cells expressing ORF63p or its localization. This experiment demonstrates that steady-state levels and nuclear localization of ORF29p are modulated by the ATPase activity of Hsp90.

Inhibition of Hsp90 abolishes VZV plaque formation.

ORF29p is indispensable for viral growth (11). Therefore, we would predict that pharmacologically induced nuclear exclusion of ORF29p would inhibit VZV replication. Two experiments were done to investigate the ability of VZV to grow in the presence of Hsp90 ATPase inhibitors.

Confluent MeWo cells were infected with cell-free VZV. After adsorption, varying concentrations of geldanamycin or 17DMAG were added to the infected cell monolayers. At 96 hpi the cells were fixed and stained with crystal violet, and plaques were counted (Fig. 5A). The number of plaques in each well treated with drugs was compared to the number arising in untreated controls. This analysis revealed that drug-induced inhibition of Hsp90 ATPase activity blocked plaque formation. In the range of concentrations used for these experiments, 17DMAG was a more potent inhibitor of plaque formation than geldanamycin (Fig. 5A).

FIG. 5.

VZV plaque and spread assays in cells treated with ansamycin antibiotics. (A) Confluent MeWo monolayers were infected with cell-free VZV. After virus adsorption the medium was replaced with DMEM supplemented with the indicated concentrations of geldanamycin or 17DMAG. The cells were fixed and stained with crystal violet, and plaques were counted after 4 days. (B) Confluent MeWo cell cultures were infected with cell-free VZV. After virus adsorption, medium was replaced with DMEM supplemented with 1.5 μM geldanamycin (GM). The cells were fixed at the indicated times postinfection, and the presence and intracellular localization of ORF29p and gE were determined by indirect IF microscopy after reaction with specific antibodies. ORF29p was visualized after reaction with goat anti-rabbit antibody conjugated to AlexaFluor 546 (red) and gE with goat anti-mouse conjugated to AlexaFluor 488 (green). Images were viewed and captured using a 10× objective.

To ascertain the effect on virus spread, we infected confluent MeWo cells with cell-free VZV. After adsorption the cells were overlaid with medium containing DMSO or drugs. The cells were fixed at 24, 48, or 72 hpi, and the expression and intracellular localization of ORF29p and gE, a late glycoprotein, were determined. We show that in cells treated with DMSO, the virus has started to spread to neighboring cells at 24 hpi. By 48 hpi, the cells started to fuse and gE was expressed at high levels. By 72 hpi, large plaques were apparent (Fig. 5B). At all time points in cells treated with geldanamycin, a limited number of cells expressed viral early and late proteins, but the infection did not spread to neighboring cells (Fig. 5B). Similar findings were obtained with 17DMAG (data not shown). Although the nuclear import of ORF29p is inhibited by ansamycins, we cannot exclude the possibility that cellular clients of Hsp90 are also required for VZV replication and are affected by drug treatment.

Hsp90, Hsp70/Hsc70, and BAG3 are redistributed during VZV infection and are localized in nuclear replication-transcription foci.

We established that ORF29p forms complexes with BAG3 and Hsp70/Hsc70 in vivo, and its degradation is dependent on Hsp90. We next determined the intracellular localization of these proteins during VZV infection.

MeWo cells infected with VZV were fixed 24 hpi, and the intracellular localization of ORF29p, ORF62p, Hsp90, Hsp70/Hsc70, and BAG3 was monitored by indirect IF microscopy. As previously shown, ORF29p localized predominantly in the nuclei of cells infected with VZV (39, 71). The protein was diffuse in the nucleus of some infected cells, whereas in others it was localized in discrete regions of the nucleus. The localization pattern of ORF29p was very similar to that of its HSV-1 homolog, ICP8, which exhibits diffuse nuclear staining at early times but accumulates in sites of DNA replication later in infection (9, 63). Moreover, we showed that ORF62p, a transcription regulator, localized in a manner similar to ORF29p and its HSV-1 homolog, ICP4 (40). Specifically, ORF62p was predominantly in the nucleus of infected cells, showing a diffuse staining early but concentrating in globular structures later in infection. Furthermore, the merged images demonstrated that ORF29p and ORF62p colocalized, indicating that sites of replication and transcription form in the nuclei of infected cells, similar to what was shown for HSV-1 (21) (Fig. 6A).

FIG. 6.

Localization of ORF29p, ORF62p, Hsp90, Hsp70/Hsc70, and BAG3 in cells infected with VZV. MeWo cells were infected with cell-free VZV at a multiplicity of infection of approximately 0.01. After 24 h cells were fixed, and the localization of the indicated proteins was visualized after reaction with specific antibodies. Images were captured with a 100× objective and analyzed by volume deconvolution.

The intracellular localization of the ORF29p-associated chaperone proteins Hsp90 and Hsp70/Hsc70 was then investigated. Both proteins were predominantly localized in the cytoplasm of uninfected cells. However, during VZV infection, a fraction of these proteins was redistributed and colocalized with ORF29p in the nucleus (Fig. 6B and C).

Finally, we showed that the cochaperone protein BAG3 also redistributed during virus infection. BAG3 was diffusely localized predominantly in the cytoplasm of uninfected cells, but in infected cells a fraction of the protein colocalized with ORF62p, concentrating in discrete nuclear structures (Fig. 6D).

These data demonstrate that in infected cells the host chaperone and cochaperone machinery redistribute and localize to intranuclear structures that contain ORFs 29p and 62p.

BAG3 is essential for efficient VZV replication.

To evaluate the functional requirement of BAG3 for the replication of VZV, we constructed cell lines stably expressing siRNAs targeting the BAG3 mRNA. Pseudotyped retroviruses were constructed and used to transduce MeWo cells (Fig. 7A). After selection in puromycin and colony isolation, intracellular BAG3 levels were evaluated by Western blotting (Fig. 7B). siRNA 737 had almost no effect on the level of endogenous BAG3, whereas cells transduced with siRNA 2235 had only 8% of the level present in cells transduced with an empty vector. The resulting cell lines were tested for the ability of VZV to replicate using plaque and spread assays.

FIG. 7.

VZV and HSV plaque assay in BAG3 siRNA MeWo cells. (A) Schematic diagram of BAG3 mRNA, coding sequence, and siRNA targeting constructs. (B) Empty, si737, and si2235 MeWo cells and si2235 cells infected with AdBAG3 at a multiplicity of infection of 20 were lysed in RIPA buffer, and equal amounts of total protein were analyzed by Western blotting with anti-BAG3 and anti-GAPDH antibodies. The protein levels were quantified using ImageJ, and the percentage of remaining BAG3 was calculated after normalization to the GAPDH levels. (C) Confluent monolayers of empty, si737, si2235, and si2235 cells preinfected with AdBAG3 at a multiplicity of infection of 20 were infected with cell-free VZV (yellow). Monolayers of empty, si737, and si2235 cells were infected with HSV-1 (blue). Three or two days postinfection, respectively, cells were fixed and plaques were stained and counted. In each experiment, serial 10-fold dilutions of virus were titrated on the previously described monolayers. The plaque numbers in the different cell lines were then normalized to the number of plaques that formed in the empty cell line. The error bars represent the standard deviation of the mean for three representative analyses.

Monolayers of confluent empty, si737, and si2235 cells were infected with cell-free VZV. At 72 hpi the cells were fixed and stained, and then plaques were counted. The number of plaques in the siRNA cell lines was normalized to the number formed in the empty cell line. Reduction of BAG3 levels resulted in a 60-fold decrease in virus titer compared to the cell lines containing either the empty vector or the nonfunctional targeting sequence. Thus, BAG3 levels dramatically and specifically affect the efficiency of VZV plaque formation.

The defect in virus growth caused by the reduction of BAG3 levels was further characterized by IF microscopy to examine the presence and localization of ORF63p and gE, an immediate-early and a late virus protein, respectively. At 48 hpi the virus replicated and spread to neighboring cells in the empty (Fig. 8, rows 1 and 2) and si737 control cell lines (data not shown). The nuclei of the infected cells formed ring-shaped structures within the polykaryocytes, and the virus proteins stained as a continuous layer, which is indicative of VZV-directed syncytia formation (77). By 96 hpi almost all cells in the monolayer expressed virus proteins, and large plaques adjacent to the condensed nuclei were seen. Although virus proteins were expressed in the cell line with decreased BAG3 levels, the virus spread to neighboring cells was slow and inefficient (Fig. 8, rows 3 and 4). At 48 hpi, the average size of the infected foci was strikingly reduced and protein staining was detected only in individual cells, demonstrating that cell fusion and virus spread had not occurred. At 96 hpi, although the virus was able to spread, the ring-shaped nuclear structures did not form. Thus, in the absence of BAG3, virus replication and spread were both delayed and defective.

FIG. 8.

VZV spread assay in BAG3 siRNA MeWo cells. Confluent monolayers of empty, si737, si2235, and si2235 cells preinfected with AdBAG3 at a multiplicity of infection of 20 were infected with cell-free VZV. The cells were fixed at the indicated time points, and the presence and localization of gE and ORF63p were determined. Images were viewed and captured using a 10× objective.

To confirm that the altered plaque phenotype, decreased plaquing efficiency, and lower yield (data not shown) of VZV resulted from decreased levels of BAG3, an adenovirus vector expressing an siRNA-resistant form of BAG3 that lacked the si2235 target sequence was used to complement si2235 cells. Western blot analysis revealed that BAG3 levels were increased to 56% of that found in the empty cell line (Fig. 7B). Preinfection of si2235 cells with AdBAG3 significantly restored the defective phenotype as measured by VZV plaque formation. The number of plaques formed increased by 16-fold (Fig. 7C). Moreover, IF microscopy of infected cells revealed ring-shaped nuclear structures and enhanced virus spread compared to si2235 cells (Fig. 8, rows 5 and 6).

From these data we conclude that BAG3 is required for efficient VZV replication and plaque formation and that decreased levels of BAG3 expression can be complemented by exogenous expression from an adenovirus vector.

BAG3 is not required for HSV replication.

To evaluate BAG3's specificity for VZV replication, we asked if growth of another alphaherpesvirus, HSV, was altered in the siRNA knockdown cell lines using plaque and spread assays.

Monolayers of confluent empty, si737, and si2235 cells were infected with HSV. At 48 hpi cells were fixed and stained, and plaques were counted. In contrast to VZV, plaquing efficiency of HSV was unaffected by decreased BAG3 levels (Fig. 7C).

The growth of HSV in the knockdown cell lines was further characterized using spread assays. Confluent empty and si2235 cells were infected with HSV and fixed at 48 hpi. The presence and localization of an immediate-early protein, ICP0, and a late glycoprotein, gC, were examined by IF microscopy. Unlike VZV, virus spread and the morphology of virus-induced syncytia were similar in control and knockdown cell lines (Fig. 9).

FIG. 9.

HSV spread assays in BAG3 siRNA MeWo cells. Confluent monolayers of empty and si2235 cells were infected with 200 PFU of HSV. The cells were fixed at 48 hpi, and the presence and localization of ICP0 and gC were determined. Images were viewed and captured using a 10× objective.

Thus, our data indicate that BAG3 is a host protein that is specifically required for efficient replication of VZV and not HSV.

DISCUSSION

In this report we have identified and studied the role of BAG3, a host protein that interacts with VZV ORF29p. A lambda cDNA expression library screen identified BAG3 as a host protein that interacts with ORF29p. Far-Western blotting showed that ORF29p associates in vitro with full-length BAG3 and fragments that contain the BAG domain. The region of ORF29p that is required to form the complex lacks the NLS, suggesting that the direct biological role of the interaction is not the association of ORF29p with the nuclear import machinery. The far-Western assay also demonstrated that ORF29p does not directly interact with Hsp70, a known BAG3 partner (Fig. 1E). Thus, the interaction of the protein with BAG3 is either direct or bridged by other proteins present in the rabbit reticulocyte lysate.

Unlike other human alphaherpesviruses, VZV expresses a subset of its genome during latency. It is believed that the LAPs may passively maintain the latent state because they fail to accumulate in the nucleus and thus render the virus unable to replicate. However, a more active role is also possible. Successful maintenance of latency requires the survival of infected neurons. Bcl-2, a well-described oncoprotein and potent inhibitor of apoptosis (18), heterodimerizes with members of the Bcl-2 protein family, including the proapoptotic factor Bax. BAG3 and other BAG family members interact with Bcl-2 and synergize with it to prevent Bax-induced cell death (1, 42). The interaction of ORF29p with BAG3 raises the possibility that expression of this LAP in latently infected neurons can regulate apoptosis.

Since BAG3 abrogates protein degradation mediated by heat shock proteins (22), we hypothesized that its interaction partner ORF29p is a client for Hsp90-directed proteasomal degradation. We show that inhibition of Hsp90 activity with drugs that associate with the ATP binding domain of Hsp90 and “lock” the protein in its ADP-bound conformation (31, 73) results in the rapid degradation of the protein (Fig. 3). We have previously reported that restriction of ORF29p to the cytoplasm results in its rapid degradation. However, the protein is stabilized when proteasome inhibitors are added, and it can then accumulate in the nucleus (70). Here we show that the converse is also true. Pharmacological intervention of Hsp90 activity destabilizes ORF29p and causes it to accumulate in the cytoplasm of cells, where it is normally localized to the nucleus (Fig. 4). Interestingly, the localization of ORF63p is not affected by the same treatment. Although both ORF29p and ORF63p are expressed during latency in infected dorsal root ganglia and accumulate in the nucleus during reactivation, we demonstrate that the mechanisms governing the degradation and the nuclear/cytoplasmic switch for the two proteins are distinct. This result suggests that the nuclear import of the individual LAPs during the reactivation process may involve unique cellular pathways that are triggered in response to certain stimuli.

It is well known that molecular chaperones recognize and associate with all nascent polypeptides as they exit the ribosome to prevent illegitimate interactions between exposed hydrophobic surfaces and to assist proteins in adopting their native conformation (25). Although molecular chaperones can also direct protein turnover, unlike the folding of newly synthesized proteins, only a subset of the cellular polypeptides is degraded by this pathway. We suggest that recognition mechanisms exist to differentially target proteins, such as ORF29p, to the proteasome but spare others, such as ORF63p. It is likely that the interaction of cochaperones, e.g., BAG3, with their clients alters the activity of the chaperones, resulting in differential targeting.

Drug attenuation of Hsp90's ATPase activity efficiently inhibits VZV plaque formation (Fig. 5). Geldanamycin treatment also inhibits replication of HSV-1 (43) by inhibiting nuclear transport of the virus DNA polymerase (7). It is likely that inhibition of VZV replication is a result of interfering with both cell and virus targets. As far as VZV targets are concerned, we demonstrate that the drug affects at least ORF29p localization and stability. Ansamycin antibiotics are currently in clinical trials for their anti-tumor activities (57). We propose that the same compounds can be used as antiviral drugs targeting the replication and spread of alphaherpesviruses.

We also show that protein complexes containing ORF29p, Hsp90, Hsp70/Hsc70, and BAG3 assemble at distinct intranuclear sites in VZV-infected cells. Furthermore, our results demonstrate that ORF62p colocalizes with these proteins in virus replication compartments. This indicates that replication and transcription coincide in discrete compartments inside the nuclei of VZV-infected cells. The redistribution of heat shock proteins to the nucleus of infected cells is a conserved characteristic of human alphaherpesviruses and was described for HSV-1 (7, 8). However, the function of chaperone and cochaperone proteins during virus infection is not yet understood.

Bacterial chaperones are essential for lytic replication of bacteriophage lambda. DnaK, the prokaryotic homolog of Hsp70, is required for release of the P protein from the preprimosomal complex and generation of the large multienzyme complex on the origin of DNA replication (44). Furthermore, the GroES/EL chaperone system is required for folding and multimerization of the connector complex, which is similar to the portal structure of HSV-1 (28). The replication machinery of DNA animal viruses, such as papillomaviruses and polyomaviruses, also interacts with eukaryotic chaperone proteins (10, 46). However, the function of such interactions has not yet been determined. It is likely that the evolutionarily conserved activities of mammalian chaperones are required for efficient replication of animal viruses, such as alphaherpesviruses. The greatly increased macromolecular crowding in the nuclei of cells infected with herpesviruses and the efficient and rapid assembly of the replication and transcription complexes are likely to require heat shock proteins. Furthermore, proteolytic activity is also associated with these proteins. Given that a number of host proteins are degraded in the nucleus during alphaherpesvirus infection, molecular chaperones are candidates to provide the driving force for virus-directed protein degradation (20, 26).

To provide evidence for the functional significance of the virus-induced redistribution of the chaperone and cochaperone proteins in the nuclei of infected cells, we evaluated virus replication and spread in cell lines with reduced levels of BAG3. VZV is able to infect these cells and express immediate-early and late proteins. However, the infection inefficiently spreads when intracellular BAG3 levels are decreased. Nevertheless, in BAG3-depleted cells ORF29p accumulates in the nucleus when autonomously expressed, and following VZV infection heat shock proteins localize to the nucleus and virus transcription/replication sites form (data not shown). This suggests that localization of the chaperone proteins in the nucleus is not sufficient for successful replication of the viral genome. Regulation of heat shock protein activity by BAG3 and/or other cochaperone proteins is also required.

Because BAG3 interacts with ORF29p, a component of the VZV replication machinery, we posit that it facilitates virus genome replication. In the absence of BAG3, inefficient replication results in a temporal delay in the accumulation of viral progeny and drastically decreased levels of late glycoproteins that are required for cell fusion (17), thus limiting the spread of VZV to adjacent cells.

Although the functional dependence of the Hsp70/Hsc70 complexes on BAG3 is unknown, structural data suggest that members of the BAG family function as nucleotide exchange factors (69). Thus, they may promote the release of Hsp70/Hsc70 substrates and exchange bound ADP for ATP, similar to the prokaryotic protein GrpE (45). Based on this hypothesis, only small amounts of BAG3 would be required for certain cellular processes. This is consistent with our results, as knockdown of BAG3 does not block nuclear accumulation of virus and host proteins, allowing for formation of transcription/replication bodies but at the cost of a dramatically slowed and defective VZV infection cycle.

Interestingly, HSV-1 growth is not inhibited in cell lines with reduced BAG3 levels (Fig. 7C and 9). This may reflect the differences in distribution of chaperones and cochaperones in the nuclei of cells infected with HSV versus those infected with VZV. During HSV infection, Hsp70 localizes adjacent to but not within the globular replication factories where virus-specific proteins and Hsp90 reside (7). In contrast, Hsps 70 and 90 are uniformly distributed throughout replication sites formed in VZV-infected cells (Fig. 6). Thus, although these closely related viruses redistribute the host chaperone machinery inside the nuclei of infected cells, the functional requirements of their replication/transcription machinery for cochaperones differ.

Our results demonstrate that VZV replication in cultured cells depends on regulation of the activity of chaperone proteins, including Hsp90 and Hsp70/Hsc70. Virus infection of permissive cells results in commandeering the cell's chaperone machinery to sites of replication/transcription to promote virus growth and spread. In contrast, drug inhibition of Hsp90 function or reduction of the levels of the Hsp70 activity regulator BAG3 result in greatly diminished virus replication. These findings suggest that VZV, and probably other herpesviruses, exploit the heat shock protein machinery to sense the cellular environment and regulate their life cycle. Because the levels and activity of chaperone proteins are altered in response to different stimuli and intracellular conditions and alphaherpesviruses exit from latency in response to various stress conditions, it is tempting to speculate that heat shock proteins serve as a component of a molecular switch that regulates the shift between lytic and latent infection.