Exposed Hydrophobic Residues in Human Immunodeficiency Virus Type 1 Vpr Helix-1 Are Important for Cell Cycle Arrest and Cell Death

The human immunodeficiency virus type 1 (HIV-1) accessory protein viral protein R (Vpr) is a major determinant for virus-induced G2/M cell cycle arrest and cytopathicity. Vpr is thought to perform these functions through the interaction with partner proteins. The NMR structure of Vpr revealed solvent exposed hydrophobic amino acids along helices 1 and 3 of Vpr, which could be putative protein binding domains. We previously showed that the hydrophobic patch along helix-3 was important for G2/M blockade and cytopathicity. Mutations of the exposed hydrophobic residues along helix-1 were found to reduce Vpr-induced cell cycle arrest and cell death as well. The levels of toxicity during virion delivery of Vpr correlated with G2/M arrest. Thus, the exposed hydrophobic amino acids in the amino-terminal helix-1 are important for the cell cycle arrest and cytopathicity functions of Vpr.


Introduction
The HIV-1 accessory protein Vpr has many well-characterized functions and properties during HIV-1 infection that appear to be dependent on binding to partner molecules [1,2]. Vpr is incorporated into budding HIV-1 virions [3], and within cells, Vpr mostly localizes to the nucleus [4]. These two properties are thought to allow Vpr to facilitate infection of non-dividing cells, by mediating the nuclear translocation of the viral pre-integration complex [5,6]. This is accomplished through an interaction with a member of the nuclear transport pathway, importin-a [7]. However, the necessity for Vpr during infection of macrophages is controversial because a recent study has shown no difference in the infectivity of viruses lacking Vpr versus wild-type viruses [8]. In addition, in chemically growth-arrested cells, Vpr was found to be non-essential for infection [9]. Vpr has been reported to increase the transcriptional activity from the HIV-1 long terminal repeat (LTR) promoter, as well as several other promoters, through the binding of cellular transcription factors, such as Sp1 [10,11,12]. The HIV-1 mutation rate during reverse transcription is reduced by a binding of Vpr with uracil DNA glycosylase 2 and the incorporation of the latter into viral particles [13,14]. However, besides bringing UNG2 into viral particles, Vpr is also able to reduce cellular levels of the enzyme by directly delivering it to an E3 ubiquitin ligase containing DCAF-1, DDB1, and Cul4 [15,16,17]. Vpr is a major effector of HIV-1 cytopathicity [18,19,20,21]. Finally, Vpr is able to halt the proliferation of infected CD4 + T cells by causing cell cycle arrest at the G2/M phase [22,23,24,25].
Vpr is a very small protein of only 14 kD (96 amino acids) and has no known enzymatic activity, no recognizable domains in the NCBI database, and no structural homologues in the Protein Data Bank (PDB), so the molecular basis of its actions is not completely understood [26]. Vpr functions are thought to depend on binding with partner proteins [1,2]. The NMR structure of full length Vpr provides some insights into the possible mechanism of Vpr activity [27]. Vpr forms a bundle of three a-helices folded around a hydrophobic core with flexible termini. The structure reveals solvent exposed hydrophobic amino acids along helices 1 and 3 of Vpr [27]. In principle, hydrophobic patches seek to be solvent inaccessible and therefore we conjecture that a hydrophobic substance, perhaps another protein, likely interacts with these patches. Therefore, we hypothesized that these hydrophobic patches serve as protein interaction sites that might be important for Vpr functions.
Separate regions within Vpr have been shown to be critical to many Vpr functions [28]. Incorporation into budding virions has been mapped to the amino (N)-terminal half of Vpr, particularly the amphipathic helix-1 [29,30,31,32,33]. Vpr helices 1 and 2 have also been shown to be important in the nuclear localization of Vpr [29,30,33,34,35], which depends on an interaction between Vpr helix-1 and importin-a, an integral component of the nuclear import pathway [7]. However, a number of studies have highlighted a role for the carboxyl (C)-terminal portion of Vpr in nuclear localization, as well [4,36,37]. The ability of Vpr to increase transcription from the viral LTR promoter was mapped to the C-terminal half [38,39]. Historically, the region containing helix-3 and the flexible C-terminus of Vpr has been shown to be necessary for G2/M cell cycle arrest [29,30,40]. Thus these distinct portions of Vpr appear to differentially regulate Vpr functions, likely through the binding of different partner molecules.
Recently, we published a structure-function based study exploring the role of the exposed hydrophobic residues along Vpr helix-3 in cell cycle arrest, nuclear localization, Vpr dimerization, and cell death [18]. This first look at the role of this hydrophobic patch found that these residues are indeed important for G2/M arrest. However, the role of these amino acids in cytotoxicity is less clear. The levels of death were proportional to the levels of cell cycle arrest induced by hydrophobic patch mutants during virion delivery of Vpr. However, when these same mutants were expressed during HIV-1 infection, the levels of cell death were similar to wild type Vpr, likely due to a G1 arrest of the infected cells as opposed to G2/ M blockade [18].
Although the N-terminal helix-1 of Vpr has not been previously associated with cell cycle arrest, we hypothesized that the exposed hydrophobic patch along this a-helix could serve as a protein interaction region with a partner protein necessary for this function. We performed a similar structure-function based approach to test the involvement of the helix-1 exposed hydrophobic residues in cell cycle arrest and cytopathicity [18]. Mutation of these amino acids reduced G2/M arrest and cytopathicity during virion delivery of Vpr. Similar to our previous study [18], cell cycle arrest and cell death were correlated when Vpr was delivered by non-replicative virions. These data suggest that the hydrophobic patch along Vpr helix-1 is important during cell cycle arrest and cytopathicity, probably serving as a protein binding region.

HIV virus stock and virion delivery of Vpr
HIV-1 viral stocks were produced in 293T cells by transfection using ExGen 500 according to the manufacturer's instructions (Fermentas). Virion delivery of Vpr has been previously described [18,41,43,44]. A plasmid encoding a reverse transcriptase mutant with the vpr gene deleted (pNL4-3 e-n-GFP RT m , VprD22-86; a gift from E. Freed, National Cancer Institute, NIH) and a plasmid encoding the VSV-G envelope protein (pLVSV-G) were transfected into 293T cells to produce VSV-G pseudotyped virions. Plasmids encoding WT and mutant Vpr were co-transfected in order to transcomplement the Vpr-deficient virus. Jurkat cells were infected in 12-well plates in the presence of Polybrene (5 mg/ml; Sigma-Aldrich). Virus was adsorbed for 30 min at 37uC in 5% CO 2 , and then the plates were centrifuged for 30 min at 8006g at room temperature.

Transfection
An Electro-Cell manipulator (BTX) apparatus was used to transiently transfect Jurkat cells by electroporation. A total of 4610 6 cells were resuspended to a concentration of 10610 6 cells/ml with 10 to 15 mg of DNA in a 4-mm gap cuvette (Bio-Rad Laboratories) and electroporated at 260 V and 1,060 mF. Transfected cells were transferred to fresh supplemented RPMI and assayed after 3 days. A plasmid encoding the green fluorescent protein (GFP) was cotransfected at a 1:5 ratio to the Vpr expression plasmid to identify transfected cells. The human codonoptimized Vpr plasmid, hVpr [18,41], was used to express WT Vpr. Mutations of W18 and L22 were generated using a PfuUltra II (Stratagene) site-directed mutagenesis protocol.
Assays for cell cycle and cell viability DNA content analysis was performed by propidium iodide (PI) staining. Cells were fixed with 1% paraformaldehyde in phosphate-buffered saline (PBS) for 10 min at room temperature, washed in PBS, and incubated in 70% ethanol for at least 30 min. Cells were washed again in PBS and stained with DNA staining solution (5 mg of PI/ml, 50 mg of RNase/ml, and 0.45 mg of sodium citrate/ml in PBS) at room temperature for 30 min. Stained cells were examined using a FACSCalibur flow cytometer (Becton Dickinson), and a constant number of cells were measured. Viability of infected cells was assessed by the exclusion of the vital dye propidium iodide, measured for a constant period of time (30 sec) per sample. All flow cytometric data were analyzed by using FlowJo software (Tree Star, Inc.).

Immunoblotting
Jurkat cells, 293T cells, and concentrated virus stocks were lysed in a 2% sodium dodecyl sulfate (SDS) buffer (2% SDS and 10% glycerol in 60 mM Tris-HCl [pH 7.5] with 1 U/ml of DNase [Benzonase nuclease; Novagen] and Complete protease inhibitor cocktail [Roche]) for at least 30 min at 4uC. A bicinchoninic acid assay (Pierce) was used to determine protein concentration of the lysates. Equal masses of protein were loaded onto a 4 to 20% Bis-Tris SDS gel (Bio-Rad Laboratories) for SDS-polyacrylamide gel electrophoresis, and protein was transferred to nitrocellulose using a semidry transfer apparatus (Bio-Rad Laboratories). Nitrocellulose blots were blocked in 5% nonfat milk in 0.1% PBS-Tween 20 (PBS-T). Blots were probed with primary antibody overnight at 4uC, followed by secondary antibody conjugated to horseradish peroxidase diluted 1:5,000. All antibodies were diluted in 5% nonfat milk in PBS-T. After each antibody incubation, blots were washed three times in PBS-T. The bands were imaged by using enhanced chemiluminescent or SuperSignal West Dura substrates (Pierce). Densitometry was performed using ImageJ software (NIH). The primary antibodies used include Vpr antiserum (a gift from K. Strebel), p24-capsid (AIDS Research and Reference Reagent Program; 6457), and b-actin (Sigma-Aldrich; A1978).

Results
The core of the Vpr protein (amino acids [AA] 17-77) is comprised of three a-helices folded around a hydrophobic core [27]. Figure 1A shows the boundaries of Vpr helix-1 (AA 17-33), helix-2 (AA 38-50), and helix-3 (AA 56-77) [27]. The animo (N)-and carboxyl (C)-termini were found to be flexible and unstructured (Figure 1). Along helices 1 and 3 of Vpr are several solvent-exposed hydrophobic amino acids (Figure 1, red for amino acids of helix-1 and blue for those of helix-3). The rest of the Vpr surface is hydrophilic, with only a few sporadically spaced hydrophobic amino acids partially exposed to the solvent environment ( Figure 1C, E, and G). Using a previously successful strategy [18], we employed a structurally based mutagenesis approach to disrupt the hydrophobic patch on Vpr helix-1. The residues tryptophan (W) 18 and leucine (L) 22 were mutated to alanine (A) (small, non-polar), serine (S) (polar), or glutamic acid (E) (charged, highly hydrophilic) in order to disrupt the chemical nature of the hydrophobic patch to varying degrees. Although L26 is also exposed to the surface of the protein, the side chain has lipophilic interactions with the core of the protein as well [27] ( Figure 1C, E, and G). Therefore, mutations of L26 would probably destabilize the entire protein. We attempted to minimize any possible disruption to the secondary and tertiary structure of Figure 1. The NMR structure of Vpr shows solvent exposed hydrophobic residues along helices 1 and 3. (A) A wire and box diagram of Vpr. The gray boxes represent the 3 helices of Vpr, and the upper set of numbers denotes the boundaries of each helix. The lower set of numbers identifies exposed hydrophobic residues along helix 1 and helix 3. (B) A tube diagram of the side view of the nuclear magnetic resonance (NMR) structure of HIV-1 Vpr (PDB: 1M8L) indicating hydrophobic amino acids exposed on the apparently solvent exposed surfaces of helix 1 (red) and helix 3 (blue). (C) A space filling diagram of the side view of HIV-1 Vpr using a hydrophobicity color scheme. The exposed hydrophobic residues along helices 1 and 3 are colored as in (B). Positively charged or polar amino acids with high hydrophilicity are colored light blue. Purple residues are negatively charged with high hydrophilicity. Polar amino acids with low to neutral hydrophobicity are gray, and highly hydrophobic, nonpolar residues are brown. The a-  Vpr by only mutating amino acids with outward facing side groups.
In the previous analysis, some of the Vpr helix-3 mutants could not be expressed [18]. Therefore, we first tested the expression of the helix-1 mutants by western blot analysis as previously described [41]. All Vpr mutants expressed at similar levels as WT Vpr in HEK293T cells (Figure 2A). Since helix-1 of Vpr is responsible for the incorporation into budding virions by binding to the p6 region of Gag [30,31,32,45], we also studied the helix-1 mutants for Vpr incorporation by producing non-replicative virus using a reverse transcriptase mutant and Vpr deficient strain of HIV-1 (NL4-3 e-n-GFP RT m , VprD22-86) and trans-complementing Vpr with separate expression plasmids as previously described [18,41]. Although expression was comparable in the 293T producer cells, WT and mutant Vpr were differentially packaged into virions ( Figure 2B-C). We calculated the percent of Vpr that is incorporated into the virions (relative to WT Vpr) by comparing the relative intensities (by densitometry analysis of the western blots) of the Vpr protein bands to the p24 capsid structural protein.
All of the Vpr helix-1 mutants were packaged into virions at approximately 40-80% of the level of WT, except for L22E, which was only ,20% compared to WT ( Figure 2C).
We first characterized the cell cycle arrest activity of the helix-1 mutants by using virion delivery of Vpr (Vpr v ) as previously described [18,41]. We titrated the amount of WT Vpr v to provide matched control levels of Vpr for the mutants. All of the mutants were delivered into the Jurkat cells at similar levels as the medium dose of WT Vpr v (G2,M/G1 = 4.6) by western blot analysis ( Figure 3A). Interestingly, mutation of W18 to any of the substitute residues (A, S, or E) greatly reduced the Vpr-induced G2/M arrest ( Figure 3B). Glutamic acid had the greatest effect (G2,M/ G1 = 1.4), as predicted, since the negatively charged glutamic acid should disrupt the chemical nature of the hydrophobic patch to the greatest degree. W18 may assist the binding to a partner protein through stacking interactions of the aromatic ring structure, which could explain why all mutations of this residue had a pronounced effect. The alanine substitution of L22 had no effect on Vpr v cell cycle arrest. However, the L22S (G2,M/ G1 = 3.2) and the L22E (G2,M/G1 = 1.7) mutations reduced the G2/M arrest ( Figure 3A).
Transient transfection can also be used to evaluate Vpr-induced cell cycle arrest [18,41]. Equivalent expression of the mutants was confirmed by western blot analysis, and all of the mutants were expressed at a similar level as the WT sample relative to the loading control ( Figure 4A). The same cell cycle trend was found with transfection as with virion delivery. All substitutions for W18 reduced cell cycle arrest. Whereas the L22A mutation had minimal effect, mutations to S and E both showed less accumulation of cells in G2/M than WT ( Figure 4B). These results indicate that the exposed hydrophobic patch on Vpr helix-1 is important for cell cycle arrest function, corroborating the viriondelivered Vpr v system.
The previous study on the helix-3 hydrophobic patch (I63, I67, and I74) found a correlation between cell cycle arrest and cytopathicity for the mutants using virion delivered Vpr v [18]. We examined the helix-1 mutants for a similar correlation. WT and mutant Vpr was delivered into Jurkat cells, and the cell cycle profile and viability of these cells were assessed over time as previously described ( Figure 5) [18]. Similar to the previous report, we found that wild type Vpr v reduced the cell viability to approximately 65% over the course of infection ( Figure 5B). G2/ M arrest correlated with the toxicity of the helix-1 mutants during virion delivery ( Figure 5C). The L22A mutant, which has comparable cell cycle arrest function as WT, also showed similar cell death activity ( Figure 5). The other mutants exhibited reduced cytopathicity proportional to the reduction in cell cycle arrest. The W18E and L22E mutants showed the most pronounced reduction in toxicity, corresponding with the most attenuation of G2/M arrest ( Figure 5). Although cell cycle arrest correlated with cell death among the Vpr mutants, some discrepancies still existed. For example, Vpr L22S caused more arrest than both W18A and W18S ( Figure 5A); yet L22S caused less death than W18A/S ( Figure 5B). These results suggest that the exposed hydrophobic residues along Vpr helix-1 are important for cytopathicity, and that G2/M arrest frequently correlates with cell death, similar to the results with the helix-3 mutants [18].

Discussion
Vpr likely functions by binding with partner proteins [1,2,28], and our lab has previously shown a role for the exposed hydrophobic patch along Vpr helix-3, an expected protein-protein interaction site, in the G2/M cell cycle arrest and cell death functions [18]. Here, we present an analysis of another expected protein binding domain, the hydrophobic patch of Vpr helix-1, for Vpr-induced G2/M arrest and cytopathicity, as well as virion incorporation. Although, the expected defect in virion incorporation was observed with the mutations we created, this is the first examination specifically of the hydrophobic patch residues. The L22E mutant was particularly impaired in virion incorporation, suggesting that the introduction of a negative charge at this site inhibits the binding of Vpr to the p6 region of the Gag polyprotein. We found that the exposed hydrophobic amino acids W18 and L22 are important for both G2/M arrest and cytopathicity, and that these two properties of Vpr are correlated.
We found that specific mutants of W18 and L22 reduced cell cycle arrest to varying degrees in both virion delivery of Vpr and transient transfection. The differences between the mutations to A, S, and E at each position were most apparent with Vpr v due to the higher degree of cell cycle arrest caused by this technique. However, even though the differences were smaller using transfection, the same trend was seen between the mutants (Figures 4 and 5A). The correlation between G2/M arrest and Vpr-induced death caused by Vpr v is similar to our previous study of Vpr helix-3 [18]. However, analysis of the helix-3 mutants in the context of HIV-1 infection did not show reduced cytopathicity. This was due to a G1 arrest caused by the mutant Vpr. Even though the infected cells exhibited a more normal DNA content profile, the cells were still inhibited from proliferating [18]. It would be interesting to test whether the helix-1 mutants relieve G2/M arrest but promote a G1 arrest during HIV-1 infection, and thus still kill the infected cell.
Vpr is thought to mediate all of its known functions through the binding of partner molecules [26,28]. We and others have shown that Vpr binds to the 14-3-3 family of scaffold proteins [42,46]. This interaction induces a large molecular complex nucleated on the 14-3-3 protein that consists of a number of cell cycle regulatory proteins [42]. Recently, multiple groups have found that Vpr binds to an E3 ubiquitin ligase comprised of DCAF1/VprBP, DDB1, Cul4, and Roc1 [16,47,48,49,50,51,52]. Vpr is proposed to recruit an unknown but critical cell cycle component into the E3 ligase complex and induce the abnormal degradation of that unknown cell cycle protein. This mechanism has already been shown to occur for the UNG2 protein [15,16,17]. Perhaps the hydrophobic patch of Vpr helix-1 and/or helix-3 participates in the interaction between Vpr and any of these partner proteins. Thus, it would be interesting to examine the role of exposed hydrophobic residues along these a-helices (W18, L22, I63, L67, and I74) in Vpr-dependent molecular complexes.
A close examination of our data and the results of the helix-3 study reveals that none of the mutants completely abrogated G2/M arrest [18]. However, all of these mutants only substitute a single residue. It is possible that a single change in either of the hydrophobic patches is not enough of a disruption of the chemical nature of the putative protein interaction site. Therefore, testing combinations of mutants, such as W18A and L22A, may further reduce cell cycle arrest and cell death cause by Vpr. Another possibility is that the exposed hydrophobic amino acids along both Vpr helices could actually comprise one protein interaction site. The hydrophobic side groups do orient to the same outward ''half'' of the core of Vpr (Figure 1). However, the ''ridge'' of arginines along helix-3 ( Figure 1C, E, and G) would likely disrupt the continuity of a single large hydrophobic patch. Perhaps mutations in both patches are necessary to completely inhibit the G2/M arrest function.
We have previously shown that HIV-1 causes a necrotic death of the host CD4 + T cells and T cell lines [53,54]. The accessory proteins Vpr and Vif are independently able to cause this cytotoxicity [19]. Interestingly, both Vpr and Vif are able to induce a G2/M cell cycle arrest, suggesting that cell cycle arrest may be the cause of HIV-induced necrosis [19,42]. In fact, mutants of the exposed hydrophobic residues of Vpr helix-3 cause a G1 cell cycle arrest in the context of HIV-1 infection. This G1 arrest still leads to the death of the host cell, indicating that any inhibition of proliferation of HIV-1 infected cells will cause the death of the cell [18]. It would be interesting to test whether the helix-1 mutants also induced a G1 arrest in the context HIV-1 infection leading to cell death. Further investigation into the exact mechanism linking cell cycle arrest to cytopathicity could provide possible targets for therapies that would reduce the depletion of CD4 + T cells during HIV-1 infection.
We found that the exposed hydrophobic amino acids W18 and L22 are important for both G2/M arrest and cytopathicity, and that these two properties of Vpr are correlated. Although this is not the first study to implicate the amino-terminal region of Vpr in cell cycle arrest, there are unresolved questions in the previous mutagenesis findings. Mutations of hydrophobic core residues, such as A30L [29] and H33R [40], would likely cause misfolding of the protein around the core [27]. Introduction of proline residues into the first a-helix, such as the E21,24P mutation [30], would destroy the helical structure of that region, and probably alter the entire global structure of Vpr [26]. We cannot be certain that the W18 and L22 mutations that we introduced did not significantly alter Vpr structure without further structural examination. However, these residues do not interact with the hydrophobic core of the protein, and disruption of the local secondary structure by the substitute residues is unlikely. Therefore, this is the first study showing a role for the helix-1 hydrophobic patch in Vpr-induced G2/M arrest and cytotoxicity.