INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Human immunodeficiency virus type 1 (HIV-1)¹ is a member of the lentivirus family. HIV-1 has a complex viral life cycle and utilizes multiple cellular and virally encoded regulatory proteins to tightly control its replication (1). The essential retroviral enzymes, reverse transcriptase, ribonuclease H, protease (PR), and integrase, lack cellular counterparts and have been used as targets for developing agents that inhibit virus replication (2-4). Despite considerable advances in anti-reverse transcriptase and PR therapy, it is obvious that ongoing genetic changes of the virus can confer drug resistance (5). This problem has led to proposals for alternative therapeutic strategies.

Specific cellular compartment localization of therapeutic moieties influences their desired functions. Cytoplasmically localized anti-HIV-1 integrase single chain variable fragments inhibit HIV-1 infection of T-lymphocytes more potently than anti-integrase single chain variable fragments that are concentrated in nuclei (6). As well, functional ribozymes have also been tested by targeting these moieties into murine leukemia virus virions in comparison with cytoplasmically localized ribozymes (7). The concept of incorporation of foreign proteins into retrovirus particles has previously been reported, utilizing fusion with HIV-1 Gag and Vpr (8-12). Sorting of most cellular proteins into specific cellular compartments is determined by protein-protein interactions through specific domains (7). Membrane protein docking with target protein signal domains can incorporate target proteins into either the Golgi apparatus or mitochondria (7). Applying this target protein docking model for further molecular analyses is now possible due to the development of the phage display system.² Short peptide libraries can be generated, and panning against any target protein for searching specific binding peptides can be accomplished. By engineering the target protein binding peptide into specific protein moieties, it should be possible to deliver the binding peptide protein fusion moieties into specific cellular or viral compartments.

HIV-1 encodes proteins in addition to Gag, Pol, and Env that are packaged into virus particles. These include Vpr, which is present in all primate lentiviruses (14). The virion-associated protein, Vpr, has been studied extensively with respect to understanding its role in lentivirus infection. Vpr is expressed relatively late in the viral life cycle and encodes a 14-kDa protein (15) that is predominantly localized in the nucleus of infected cells (16). Vpr has been reported to be incorporated into viral particles at molar quantities (16, 17). The carboxyl-terminal domain (p6 region) of the Gag polyprotein precursor, p55, plays a role in the packaging of Vpr into virions (18-21). None of these publications have shown evidence for a direct interaction between Vpr and p6 proteins. Recently, several groups have shown a direct interaction between Vpr and the nucleocapsid protein, p7 (22, 23). Roques and co-workers (23) show that the interaction between Ncp7 and Vpr occurs in vitro by a recognition mechanism requiring the zinc fingers of Ncp7 and the last 16 amino acids of Vpr (23). The authors suggest that Ncp7 cooperates, possibly with p6, to induce Vpr encapsidation in mature HIV-1 particles (23). Several biological functions of Vpr have been defined. It has been shown that Vpr is essential for optimal infection of macrophages (24-26). Vpr has been reported to influence the nuclear transport of the viral preintegration complex (27). Vpr activates transcription from the HIV-1 long terminal repeat (28-30), influences terminal differentiation of some cell types (31), plays a role in the reactivation of viral gene expression, as demonstrated by addition of exogenous Vpr to cultures of latently infected T cell lines (32, 33), and causes blockage of cells in the G₂ stage of the cell cycle (34-37). Recently, Chen and co-workers (38) showed that Vpr is capable of inducing apoptosis after cell cycle arrest. Other recent studies indicate that the Vpr protein can also associate with cellular proteins, such as glucocorticoid receptors (39), the transcription factors Sp1 and TFIIB (28, 30), or the uracil DNA glycosylase (UDG) enzyme involved in cellular DNA repair (40).

Since the highly conserved HIV-1 Vpr protein can be packaged in quantities within virions similar to those of the major structural proteins, this accessory protein may be used as a docking target to deliver anti-viral agents or other foreign proteins to progeny virus. In an attempt to understand how Vpr can interact with different viral and cellular proteins that are implicated in diverse virological and cell biological pathways, we used the "peptide phage-display" methodology to determine whether any common binding motif is shared among those proteins. We demonstrate that most of the peptides that bind to Vpr contain a common motif, WXXF. In this report, the domain-specific intracellular interactions of the WXXF with Vpr were further confirmed by utilizing a variety of complementary systems. Finally, to evaluate the feasibility of this peptide-based docking strategy, we have constructed a WXXF-CAT fusion protein and analyzed the ability of this fusion construct to be packaged into HIV-1 particles through interaction with Vpr. Our results show that the WXXF-CAT fusion protein is incorporated into HIV-1 particles through a Vpr-dependent docking mechanism. This study represents a new approach in the targeting of anti-viral agents with the ability to potentially interfere with HIV-1 replication.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Plasmids-- The HIV-1 molecular clones used in this study included pNL_4-3 (41) and pNL_4-3VPR. These strains were obtained from the AIDS reagent repository (NIH). pNL_4-3VPR was created by a deletion of a 115-base pair (bp) (nucleotides 5621-5736) fragment in the Vpr open reading frame. The murine leukemia virus-based retroviral expression vector, pSLXCMV-CAT, was used in these experiments and has been described previously (42). pSLXCMV-VPR-CAT was constructed by inserting a 304-bp BamHI-MluI fragment containing the vpr gene via BamHI-MluI sites of the polylinker region of the pSLXCMV vector. The stop codon of the vpr gene was removed and replaced by a MluI site to permit the fusion of the 677-bp MluI-BamHI fragment containing the CAT gene into pSLXCMV-VPR* in MluI-BglII sites. The pSLXCMV-dWF-CAT was constructed by inserting a 76-bp BamHI-MluI fragment containing the WXXF dimer (dWF) into pSLXCMV-VPR-CAT in the BamHI-MluI sites. The dWF fragment was synthesized by the polymerase chain reaction (PCR) using WF-dimer-1 as a sense primer and WF-dimer-2 as an antisense primer. The construction of plasmids, GAL4 DNA binding domain (pGBT10), and GAL4 activation domain (pGADGH) have been described elsewhere (43). The UDG gene was used as a PCR template, corresponding to the cDNA UDG described previously (40). This gene was amplified by PCR using a 5' primer, UDG1, and a 3' primer, UDG2, containing EcoRI and SalI restriction sites, respectively, and cloned in-frame into the corresponding sites of pGBT10 to obtain the plasmid pGB-UDG. Mutagenesis of the UDG gene was performed with complementary PCR oligonucleotides that contained the region of the gene with the desired mutations. Two rounds of PCR using the primers UDG222, UDG225, and UDG152, were used to construct pUDGW222G, pUDGF225G, and pUDGQ152L, respectively. The PCR products were then cloned back into EcoRI and SalI restriction sites of pGBT10. All the DNA fragments that correspond to the peptides (peptides 1-10) were amplified by PCR from purified phage DNA and then cloned in-frame into the corresponding sites of pGADGH (EcoRI-BglII) to obtain the plasmids pGAD-peptides-(1-10). All new constructs were sequenced to verify reading frames and mutations.

His-tag Vpr Purification-- Baculovirus expression vector, pACHis-SV-A-Vpr, was used to generate the recombinant Vpr-baculovirus.² SF9 cells were infected with Vpr baculoviruses (1 × 10⁸ colony-forming units/ml). SF9 cells expressing His-tag Vpr were harvested by centrifugation. The cells were suspended in binding buffer (20 mM Tris-HCl, pH 7.9, 0.5 M NaCl, 5 mM imidazole). After sonication, the cellular debris, which contains most of the expressed Vpr, was collected by centrifugation at 20,000 × g for 15 min and washed with binding buffer. The pellet was then dissolved in binding buffer containing 6 M guanidine and incubated at room temperature for 1 h. Insoluble material was removed by centrifugation (20,000 × g, 20 min), and the supernatant was mixed with 1 ml of pre-equilibrated His-Bind resin (Novagen) and incubated at 4 °C for 1 h with gentle rotation. Nonbound substances were washed from the resin with binding buffer containing 6 M guanidine, and the binding protein was eluted with binding buffer containing 6 M guanidine and 0.5 M imidazole. After dialysis against water, the eluted purified His-tag Vpr was precipitated and resuspended in phosphate-buffered saline.

GST Fusion Protein Expression and Purification-- The pGEX-GST-VPR was constructed by inserting a 304-bp BamHI-HindIII fragment containing the vpr gene via BamHI-HindIII sites of pGEX-KG (44). The expression of the GST-Vpr fusion protein was performed as described previously (45). Protein expression was induced with 0.1 mM isopropyl-1-thio--D-galactopyranoside for 1 h. Cells were harvested by centrifugation and suspended in 10 ml of PBST (20 mM phosphate buffer, pH 7.3, 150 mM NaCl, 1% Triton X-100) containing 2 mM EDTA, 0.4 mM phenylmethylsulfonyl fluoride, 2 µg/ml leupeptin, 50 µg/ml 1-chloro-3-tosylamido-7-amino-2-heptanone, and 100 µg/ml L-1-tosylamido-2-phenylethyl chloromethyl ketone. Cells were lysed by sonication. The lysate was centrifuged at 10,000 × g for 10 min at 4 °C. The supernatant was applied to 1 ml of glutathione agarose column (Sigma) pre-equilibrated with TBST (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 0.5% Tween 20). After washing with phosphate-buffered saline, the bound GST-Vpr was eluted by 5 mM glutathione in 50 mM Tris-HCl, pH 8.0, 150 mM NaCl.

Phage-display Peptide Screening-- A phage-display peptide library kit (New England Biolabs, Beverly, MA) was used to screen binding peptides. The kit contained a heptapeptide phage-display library. For phage panning, 50 mg of purified GST or GST-Vpr fusion protein was directly applied to 50 ml of glutathione-agarose gel. After a 30-min incubation at 4 °C, the gel was washed with TBST. The phage library (2 × 10¹¹ plaque-forming units in 250 ml of TBST) was added to the GST-containing gel, mixed, and incubated at room temperature for 1 h, then centrifuged at low speed for 2 min. Supernatant was washed with 250 ml of TBST. The nonbound phage was added to GST or GST-Vpr fusion protein at equal amounts. The mixtures were incubated at room temperature for 1 h then centrifuged and washed extensively with TBST to remove the nonbinding phage. The binding phage was eluted with 100 ml of 5 mM reduced glutathione in TBS (50 mM Tris-HCl, pH 7.5, 150 mM NaCl). The binding specificity between GST and GST-Vpr was titered, as suggested by the protocol from this kit.

Yeast Two-hybrid System-- The yeast reporter strain, HF7c (40) containing two GAL4-inducible reporter genes, HIS3 and LacZ, was co-transformed with plasmids, pGBUDG and pGA-Vpr. Double transformants were plated on tryptophan-, leucine-, and histidine-deficient synthetic medium, and growth was continued for 3 days. A liquid culture assay for -galactosidase activity was performed as described (40) using SFY526 as a yeast reporter strain (46) and O-nitrophenyl--D-galactosidase as a substrate. Each assay was performed in triplicate.

Transfections-- 293T cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum at 37 °C. Transfections were performed by a standard calcium phosphate transfection method (Promega, Inc.). 1 × 10⁶ 293T cells were plated in 10-cm dishes, co-transfected with 10 µg of pNL_4-3 or pNL_4-3VPR DNA and 10 µg of pSLXCMV-CAT, pSLXCMV-VPR-CAT, or pSLXCMV-dWF-CAT, and then incubated in Dulbecco's modified Eagle's medium containing 10% fetal calf serum (growth medium) for 7 h post-transfection. This medium was then removed and replaced with fresh Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, and virus-containing supernatants and cells were collected at 48 h. The amount of virus present in transfected cell supernatants was determined by measuring HIV-1 p24 antigen levels using an enzyme-linked immunosorbant assay (Dupont).

CAT Assays-- CAT assays were performed as described previously (39). Briefly, 1 × 10⁶ transfected cells were harvested by centrifugation. The pelleted cells were lysed in 0.9 ml of CAT lysis buffer (Promega, Inc.) and 50 µl of supernatants normalized for protein content and used for standard CAT assays. Virions obtained from transfected 293T supernatants were concentrated by ultracentrifugation (20,000 × g for 3 h in a A-621 Sorval rotor) and purified on 20%, 65% sucrose gradients (40,000 × g for 16 h in a TH-641 Sorval rotor). Each fraction was quantified by measuring HIV-1 p24 antigen levels. Virions from the peak were normalized for HIV-1 p24 antigen and finally lysed with 0.3 ml of Promega CAT lysis buffer at room temperature for 15 min then at 65 °C for 15 min. Lysed virions were assayed for CAT activity. CAT activity was detected by thin-layer chromatographic separation of [¹⁴C]chloramphenicol from its acetylated derivatives and was quantitated by radioactivity counting in liquid scintillation. One hundred microliters of the lysed virions (p24 in ng/ml) was used in each CAT assay, with a fixed time of 3 h.

Western Blot Analysis-- Yeast cells growing with similar density were lysed in radioimmunoprecipitation assay buffer, and immunoprecipitation of lysates was performed with a mouse anti-GAL4-DNA binding domain antibody (Tebu) and protein A-Sepharose beads. After washes in radioimmunoprecipitation assay buffer, bound proteins were subjected to 12% SDS-polyacrylamide gel electrophoresis, electrotransferred to a polyvinylidene difluoride membrane (Amersham Pharmacia Biotech), and immunodetected with a rabbit anti-GAL4 binding domain antibody (Tebu) followed by horseradish peroxidase-linked swine anti-rabbit immunoglobulin (Dako). Antibody binding was demonstrated with ECL Western blotting detection reagents (Amersham).

Oligonucleotides-- Sequences of oligonucleotides used in this study are presented. Regions in boldface indicate restriction sites for each construct. Underlined regions indicate base changes from point mutations.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Mapping of Vpr-specific Binding Peptides-- Phage display was used to identify peptides that bind to the HIV-1 Vpr protein. Purified and denatured His-tag Vpr produced from baculovirus was used by coating plates as the target protein to screen the binding phage for Vpr interactions. The His-tag Vpr showed a high titer of bound phage compared with bovine serum albumin alone after second round panning (data not shown). To increase the binding specificity of phage to Vpr, a third round of panning was performed and showed a high titer binding phage (P/n = 24.5). The positive clones (for binding with Vpr) were amplified. Phage DNA was purified according to the manufacturer's protocol and was then sequenced using the 96gIII sequencing primer. Fig. 1A illustrates the sequencing results. Analysis of these sequences demonstrated that 90% of these peptides contained a consensus motif, WXXF.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Biopanning a heptapeptide library against His-tag Vpr and GST-Vpr. Three rounds of selection/amplification were carried out at constant stringency (0.5% Tween 20). A, peptides specific for the His-tag Vpr binding were selected by eluting with 5 mM reduced glutathione in Tris-buffered saline, as described under "Materials and Methods." B, peptides specific for the GST-Vpr binding were selected by eluting with 5 mM reduced glutathione in Tris-buffered saline. Phage DNA was purified according to the provided protocol and then sequenced using the 96gIII primer. The results illustrated here are the amino acid sequences corresponding to each phage DNA.

View larger version (10K): [in this window] [in a new window]   Fig. 2.   Specific interaction of Vpr with peptides containing the WXXF motif in the yeast two-hybrid system. The HF7c reporter strain expressing pairs of hybrid proteins fused to GAL4BD and to GAL4AD was analyzed for histidine auxotrophy. Double transformants were patched on selective medium without histidine. GAL4AD alone was used as a negative control. +++, strong interaction; ++, moderate interaction; + or ±, weak interaction.

Binding of Vpr and UDG Mutants-- We recently reported that Vpr was able to bind tightly to and be coimmunoprecipitated with UDG (40), the major uracil DNA glycosylase in human cells (47). Examination of the human UDG amino acid sequence indicates that this protein contains a WXXF motif (amino acids 222-225). The region of UDG protein containing the WXXF motif is conserved between different species, as shown in Table I by alignment of protein sequences of uracil DNA glycosylases.

View this table: [in this window] [in a new window]   Table I Comparison of protein sequences of uracil-DNA glycosylases Alignment of uracil-DNA glycosylase from man (HUM), yeast (YEA, UNG1, E. coli (Eco, ung gene (59)), herpes simplex virus type 1 (HSV1 gene UL2 (60)), and herpes simplex virus type 2 (HSV2, UL2 (61)). The WF motif shared in all sequences is boxed.

View larger version (28K): [in this window] [in a new window]   Fig. 3.   In vivo interaction of UDG mutants with Vpr in the yeast two-hybrid system. Mutant forms of UDG were subcloned into GAL4BD. Yeast strain HF7c (for histidine selection) or SFY526 (for -galactosidase activity) were co-transformed with each GAL4BD-UDG mutant and the GAL4AD-Vpr fusion expression plasmid. A, expression of UDG mutant forms. Yeast cells were lysed in radioimmunoprecipitation assay buffer and immunoprecipitated with a mouse anti-GAL4BD antibody and protein A-Sepharose beads. Western blots were performed with a rabbit anti-GAL4BD antibody. The arrow represents the size corresponding to GAL4BD-UDG. cont, control. B, double transformants were patched on selective medium with or without histidine. Growth in the absence of histidine is an indication of the interaction between hybrid proteins. C, double transformants were analyzed for -galactosidase activity in a liquid culture assay. Values are means of triplicate determinations made on three independent transformants. The background level is approximately 7% and corresponds to yeast co-transformed with either the GAL4AD-Vpr or the GAL4BD-UDG expression plasmids and either the GAL4BD or GAL4AD expression plasmids, respectively. Interaction between Vpr and wild-type UDG was used as a positive control.

Detection of CAT Activity in HIV-1 Virions Based on Binding of WXXF-CAT Chimeras to Vpr-- We next tested the hypothesis that WXXF might interact with Vpr and dock fusion proteins containing this motif into HIV-1 virions. This was first tested by fusion of WXXF to the CAT gene. Of note, retroviral vectors are powerful tools for the transfer of new genetic information into target cells due to their high transduction efficiency (49-51). The plasmids used in this experiment are represented in Fig. 4. The pSLXCMV-VPR-CAT plasmid was used as a positive control. Incorporation of foreign proteins into HIV-1 particles has previously been reported by direct fusion with Vpr (8, 9). These authors demonstrated the capability of HIV-1 Vpr to direct the packaging of foreign proteins such as CAT into HIV-1 virions when expressed as heterologous fusion molecules. Virion-associated CAT fusion proteins remained enzymatically active. Based on our results in the yeast two-hybrid system and the studies of WXXF motif in the interactions of UDG and Vpr, we constructed a dimer of WXXF and dWF fused to the CAT gene. The amino acid sequence of dWF was based on the amino acid sequence of the peptides that bind to Vpr, located by peptide phage display (Fig. 1B). These two WXXF domains were joined by a linker (GGGS) to allow flexibility and proper folding of these domains. The dWF was amplified by PCR using WF-dimer-1 and -2 as primers and then fused in-frame with CAT. The final constructs were named pSLXCMV-dWF-CAT (Fig. 4).

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Schematic maps of the plasmids used to express CAT, Vpr-CAT, and dWF-CAT gene products. Underlined regions indicate restriction sites used for cloning. Regions in bold face (amino acid sequence) indicate the WXXF motif. Underlined regions indicate the amino acid sequence for the linker. aa seq, amino acid sequence.

4-3VPR

VPR

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Virion association of enzymatically active CAT fusion proteins. A, HIV-1 virions collected from the culture supernatants of 293T cells co-transfected with pSLXCMV-CAT, pSLXCMV-VPR-CAT, or pSLXCMV-dWF-CAT were sedimented in gradients of 20-65% sucrose. Fractions of 0.7 ml were collected and analyzed for HIV-1 p24 antigen. B, CAT enzyme activity was determined for either the fraction corresponding to the peak (Fraction 9, lanes 5-8) or the cell extract (lanes 1-4) by standard methods. Lanes 1 and 5, NL4-3WT/pSLXCMV-CAT (WT/CAT); Lanes 2 and 6, NL4-3WT/pSLXCMV-VPR-CAT (WT/Vpr-CAT); lanes 3 and 7, NL4-3WT/pSLXCMV-dWF-CAT (WT/dWF-CAT); lanes 4 and 8, NL4-3VPR/pSLXCMV-dWF-CAT (VPR/dWF-CAT). CAT activity was detected by thin-layer chromatographic separation of [14C]chloramphenicol from its acetylated derivatives and was quantitated by radioactivity counting in liquid scintillation. The experiment was repeated three times with similar results in each replicate.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The highly conserved HIV-1 Vpr protein expressed in the late stage of viral production (15) and incorporated into virions (16) was an ideal target for testing the docking protein model for intravirion protein delivery. HIV-1 Vpr functional studies also show that this protein has a diverse phenotype via interactions with multiple cellular or viral structural proteins (22-28, 30, 31, 34-37, 39, 40). These multiple interactions indicate the possibility of Vpr-protein binding through specific domains, which may be carried in the different target proteins.

To identify such potential Vpr-specific binding domains, random peptide libraries constructed in a phage-display system were used for initial screening. Subsequently, each of the identified Vpr-binding peptides were further confirmed by using the yeast two-hybrid system for intracellular interactions of Vpr and peptides. More than 32 different Vpr binding phages were identified after a third round of panning using both denatured His-Vpr or native GST-Vpr fusion proteins in different panning systems.² Interestingly, the great majority of the peptides binding to Vpr revealed by these assays contained a consensus motif, WXXF. Intracellular binding of Vpr and peptides was subsequently confirmed from 10 selected Vpr-binding phage-displayed peptides by using the yeast two-hybrid system. Computer searching with the WXXF motif against protein data bases revealed several cellular proteins containing the WXXF motif. One of these, UDG, the major uracil DNA glycosylase in human cells (47), has been reported in our recent work and has demonstrated that UDG specifically binds to Vpr (40). More importantly, we now demonstrate that the WXXF motif of UDG is implicated in the interaction of UDG with the Vpr protein, as mutants of the UDG motif bind differentially to Vpr. The mutant F225G lost its ability to interact with Vpr, and the mutant W222G lost more than 60% of the binding with Vpr. However, we could not exclude the participation of the first or second X of the WXXF motif in the interaction, since an appropriate conformation of the WXXF motif may be important. It is unclear at present why Vpr interacts with UDG, a DNA repair enzyme. A recent report excludes the involvement of UDG in contributing to G₂ arrest of cells. Mutational analysis of Vpr showed that binding to UDG is neither necessary nor sufficient for its effect on the cell cycle (52).

UDGs and deoxyuraciltriphosphate pyrophosphatases (dUTPases) are thought to prevent the misincorporation of deoxyuracil into DNA during DNA synthesis. Since UDG is not involved in the G₂ checkpoint, the hypothesis that association of Vpr and UDG may perform a role similar to that played by the dUTPases of non-primate lentiviruses is quite possible, i.e. the reduction of uracil misincorporation into proviral DNA. Recently, Turelli et al. (53) reported that dUTPase minus caprine arthritis encephalitis virus accumulates G-to-A substitutions in vivo (53). Mansky (54) reported that the vpr gene partially accounts for the lower than predicted in vivo mutation rate of HIV-1. A vpr-negative shuttle vector had an overall mutation rate as much as 4-fold higher than that of the parental vector (54).

We recently reported that TFIIB, a basal transcription factor, binds to Vpr (28). The portion of Vpr that interacts specifically with TFIIB ranges from amino acids 15 to 77. Also, it was indicated that the NH₂-terminal domain of TFIIB is required for this interaction. Interestingly, we located the WXXF motif in the NH₂-terminal domain of TFIIB. Preliminary data indicate that mutants of TFIIB, which have a point mutation in the WXXF motif, lose their ability to interact with Vpr (data not illustrated).

Vpr has many different potential roles in HIV-1 expression. Vpr is able to transactivate several heterologous viral promoters lacking a common DNA sequence element (29). It has been shown that the addition of exogenous Vpr can reactivate HIV-1 replication in latently infected cell lines, indicating that Vpr could play a role in increasing HIV-1 expression through transcriptional or translational events (32, 33). Also, Vpr has been reported to influence the regulation of some cellular functions. Vpr induces terminal differentiation of rhabdomyosarcoma cells (31) and causes arrest in the G₂/M phase of the cell cycle (34-37), and recently this viral protein has been shown to be capable of inducing apoptosis after cell cycle arrest (38). All of these effects by Vpr are probably mediated by interactions with cellular proteins. Our findings could assist in the screening of proteins that are implicated in these functions and have the motif WXXF. We have already demonstrated the potential implications of this motif (WXXF) for UDG and TFIIB in their interactions with Vpr. Because of the strict conformation of UDG, it is formally possible that other regions are important for the interaction with Vpr. UDG is an enzyme that is extraordinarily specific in its function and so must have a very strict and defined three-dimensional conformation (48).

Recently, Chen and co-workers (55) demonstrated a weak interaction between Vpr and the cellular DNA repair protein, HH23A. The authors show that the carboxyl-terminal 45-amino acid region of HH23A interacts with Vpr. Interestingly, this carboxyl-terminal region has 2 Phe residues that are conserved between different species. This may explain the weak interaction of this protein with Vpr (55). Of interest, we show in Fig. 3 that the UDG mutant, W222G, lost 60% of its interaction with Vpr.

Since the Vpr protein can be packaged into virions in quantities similar to those of the major structural proteins (16, 17), this protein has be used to target fusion proteins to progeny virus. Serio et al. (56) have generated a chimeric protein based on Vpr utilizing the conserved protease cleavage site sequences from Gag and Gag-Pol precursor polyproteins as fusion partners. Kappes and co-workers (8, 9, 13, 57) have generated chimeric proteins based on HIV-1 Vpr and HIV-2 Vpx, utilizing CAT, staphylococcal nuclease, wild-type and mutated HIV-1 protease, integrase, and reverse transcriptase. All of these reports use Vpr or Vpx as partners for chimeric proteins. As HIV-1 Vpr may functionally block the cell cycle, possibly involving the cellular apoptosis pathways, fusion proteins with Vpr for therapeutic purposes will be difficult to utilize, as this retroviral protein may (i) perturb viral replication, (ii) lead to difficulties in delivering a fusion protein in vivo because of competition with wild-type Vpr (or natural Vpr) and thus lead to viral escape, and (iii) interfere in cellular functions. In this study, a new strategy is proposed. HIV-1 Vpr was targeted as a docking protein, by which one could target anti-viral agents or any foreign proteins fused to the WXXF motif into HIV-1 virions. This is based on our demonstration that the WXXF motif in fusion with CAT is (i) capable of delivering the CAT protein into the HIV-1 virion by interaction with Vpr, and (ii) the CAT protein retains its enzymatic activity upon fusion with WXXF.

Our choice in utilizing two copies instead of one of the WXXF domain was based on our CAT assay comparison between the dWF and WF motif of UDG. Virions from 293T cells co-transfected with HIV-1_WT/WF_UDG demonstrated differences in CAT activity compared with the virions from 293T cells co-transfected with HIV-1_VPR/WF_UDG but significantly less than with dWF (data not shown).

Development of target protein-specific docking delivery systems have broad potential applications for viral or cellular functional studies. Rapid generation of target protein-specific-binding peptide motifs using the phage-display system will provide critical peptides in different proteins. In this report, HIV-1 intravirion delivery was used as an example, but the same principle can be used for a variety of viral diseases. Utilizing combination protein-based intracellular immunization strategies, intracellular docking, and intravirion docking systems may dramatically enhance efficiency and specificity of these approaches.

These data indicate that Vpr binds to the WXXF motif, and by interaction with this motif, one can deliver a fusion protein into the HIV-1 virion through a new docking strategy. The interesting features of the chimeric proteins generated here are the minimal addition of residues (22 amino acids) and the use of Vpr not as a partner for the fusion protein but as a receptor for docking to deliver this fusion protein into virions. This study illustrates a potential new strategy in the targeting of anti-viral agents into virions that may permit a feasible avenue to interfere with HIV-1 replication in vivo.