The viral protein R (Vpr) is encoded by the human immunodeficiency viruses types 1 and 2 (HIV-1/HIV-2), the simian immunodeficiency viruses (SIV) and primate lentiviruses 1 2 . This accessory protein facilitates transport of the pre-integration complex into the nucleus of non-dividing cells 3 and fulfils multiple functions in the viral life cycle, including increase of viral replication in non-dividing host cells, induction of G2 cell-cycle arrest 4 5 , apoptosis 6 7 and transduction through cell membranes 8 (The multiple functions of Vpr are reviewed in 9 10 11 ). Vpr interacts with several cellular factors, including the human peptidyl prolyl isomerase cyclophilin A (CypA) 12 13 14 . One of the main problems with existing antiretroviral therapy is that the viruses can develop drug resistance, which necessitates identification of new potential drug targets that overcome this problem. One approach that recently has received increased attention, is targeting host factors essential for the pathogen life cycle, rather than pathogen components directly 15 16 17 . CypA could be such a target as it is dispensable for cell viability 18 19 , and viral replication of HIV-1 is determined to be effectively inhibited by use of selective inhibitors of CypA 20 21 22 23 24 25 .

Recently, we investigated the interactions of CypA with the N-terminus of Vpr at atomic resolution 14 . Prolyl cis/trans isomerization of the highly conserved proline residues Pro-5, -10, -14 and -35 of Vpr are catalyzed by human CypA and require only very low concentrations of the isomerase relative to that of the peptide substrates. However, of the N-terminal peptides of Vpr investigated, only those containing Pro-35, which appears to be vital for manifold functions of Vpr, bind to CypA in surface plasmon resonance (SPR) biosensor experiments. Extensive analysis revealed that the binding region of N-terminal Vpr to CypA consisted of the heptapeptide motif RHFPRIW centered at Pro-35 14 .

The biological significance of the interaction of Vpr with CypA, including the extensively studied interaction of CypA with HIV-1 capsid (CA), that is crucial for viral replication 26 27 , is still not completely understood (Reviewed in 25 28 ). However, specific inhibitors of the prolyl cis/trans isomerase activity of CypA, such as cyclosporine A and SDZ-NIM811, inhibit HIV-1 replication 20 21 22 23 24 25 and a possible role of CypA in both entry and postentry events of the viral life cycle of HIV-1 has been indicated 29 . The interaction of HIV-1 Vpr with CypA is known to occur in vitro and in vivo 12 13 25 , although the biological consequences thereof are disputed. The original reports concluded that CypA had significance for the de novo synthesis of Vpr, as the Vpr-mediated cell cycle arrest in HIV-1 infected T cells appeared to be eliminated in the absence of CypA activity 13 . More recently it has been suggested that the interaction of Vpr with CypA is independent of the ability of Vpr to induce G2 cell cycle arrest 12 .

Independent of these data, previous studies have unambiguously shown the N-terminal residues from Ala-30 to Phe-34, which are adjacent to or incorporated in the N-terminal binding region of Vpr to CypA, comprised of ³²RHFPRIW³⁸ centered at Pro-35, are crucial for the ability of Vpr to induce G2 cell cycle arrest 30 31 32 . Similarly, the ability of Vpr to induce G2 cell cycle arrest is also influenced by residues in the C-terminal region 71-90 30 31 33 34 35 36 37 .

All studies are in agreement that mutation of Pro-35 with Ala disrupts the interaction of Vpr with CypA 12 13 , which is consistent with a conformational change in the hydrophobic core 38 . On the other hand, intriguingly, Ardon et al. reported that mutation of the C-terminal residue Arg-80 with Ala also prevented coimmunoprecipitation of Vpr with CypA 12 . However, the molecular explanation for the latter C-terminal mutation of Vpr to interfere with the interaction of the protein with CypA has remained elusive 12 . Mutation of Arg-80 with Ala may cause a change in the folded structure of full-length Vpr or, could in theory, alter the structure of a specific novel C-terminal binding region of Vpr to CypA. The Pro residues of HIV-1_NL4-3 Vpr are located in the N-terminal domain of the protein, namely at positions 5, 10, 14 and 35. All previously determined binding domains of various proteins interacting with CypA have hitherto required at least one Pro residue as part of the binding region.

However, a limited number of studies have shown that CypA is able to bind to peptides that do not contain Pro residues through an alternative non-Pro-dependent binding region 39 . According to Demange et al., CypA exhibits two binding sites, of which the S1' site requires a Pro-containing substrate whereas the S2'-S3' site facilitates binding to peptides that do not contain Pro 39 . Saphire et al. reported that a Vpr-CypA fusion protein which has no isomerase activity and no capacity to bind to CA also rescues HIV-1 replication, which may indicate an important role of CypA associated with Vpr 40 .

These facts, together with the observation of considerable amounts of CypA in virions 41 prompted us to study the interaction of Vpr with CypA in order to identify potential novel binding sites. This paper presents SPR binding studies of the interactions of CypA with full-length synthetic Vpr (sVpr) and C-terminal sVpr peptides. A novel specific non-Pro-containing C-terminal binding region of Vpr facilitating strong binding to CypA has been identified. The contribution of this novel C-terminal binding region together with the previously characterized N-terminal binding region RHFPRIW centered at Pro-35 has been demonstrated through characterization and quantification of the interaction of full-length Vpr with CypA.

To investigate the interaction between full-length Vpr and CypA in more detail, and to find an explanation for the previously undefined role of the C-terminus of Vpr with respect to this interaction, we performed interactions studies by SPR spectroscopy between highly pure recombinant CypA and full-length Vpr and C-terminal Vpr peptides. Additionally, possibly structural changes in the C-terminal region caused by substitution of Arg-80 by Ala were investigated by nuclear magnetic resonance (NMR) spectroscopy.

Recently we determined that the N-terminal binding region of Vpr to CypA is comprised of the heptapeptide ³²RHFPRIW³⁸ centered at Pro-35 14 . Mutation of the central Pro residue (P35A) causes a loss of binding through disruption of the Pro-dependent N-terminal heptapeptide binding region. In the context of the full length molecule this mutation also results in merging of helix 1 and 2, which in turn causes a substantial change in the hydrophobic core of the protein as the anti-parallel folded structure of the N-terminal and central helices is no longer possible 38 .

In this paper we have characterized a C-terminal region of Vpr, comprising the 16 residues ⁷⁵GCRHSRIGVTRQRRAR⁹⁰, with high binding affinity for CypA. This region of Vpr does not contain any Pro residues, but binds much more strongly to CypA than to the N-terminal binding region (Figure 1). The fact that the mutant peptide Vpr^75-90 R80A binds considerably weaker to CypA than the wt peptide (Figure 2) confirmed that Arg-80 is a key residue in the C-terminal binding region. Arg-80 is considered to be located in proximity to, but not included in, the well-defined helix 3 (residues 55-77) of the structured protein formed under membranous conditions at physiological pH 45 . The secondary structure NOEs of residues 78-90 of full-length Vpr found by Morellet et al. 45 are mainly restricted to those between NH-NH_{(i, i+1)}, in addition to a few Hα-NH_{(i, i+2)} and Hα-NH_{(i, i+3)}, indicating that these residues comprise a region with a relatively weak α-helical structure. Indeed, our NMR data of wt Vpr^75-90 and Vpr^75-90 R80A confirmed this interpretation and revealed that the propensity for weak helical structure of the C-terminal binding region is essentially unaffected by the mutation of Arg-80 to Ala (Figure 3). Thus, the mutation prevents binding of Vpr to CypA through a local change in the C-terminal binding region rather than any change in the tertiary structure of Vpr.

In aqueous solution Vpr is present as high order aggregates (~decamers) with a lower percentage of higher multimers 8 . According to Fritz et al. oligomerization is mediated by the Vpr hydrophobic core but not by the flexible N- and C-terminal domains 47 . The C-terminal binding domain of Vpr to CypA is located beyond the residues that have been shown experimentally to be involved in oligomerization of Vpr. In agreement with this we used the statistical mechanics algorithm TANGO, which identifies aggregation-prone regions of peptides and denatured proteins using a set of balanced physico-chemical parameters 48 49 . According to the TANGO algorithm, a score of ≤ 0.02% indicates no aggregation, 0.02-5.0% indicates moderate aggregation, and ≥ 5.0% indicates high aggregation propensities. Application of this program predicted a region of 16 residues (Ala-55 to Ile-70) populating the oligomerization state to more than 5% per residue (5.09 - 35.56% per residue). Four further residues were predicted with much lower score, namely Thr-53 (1.30%), Trp-54 (3.50%), His-71 (0.84%) and Phe-72 (0.80%). The fact that the dissociation constant of full-length Vpr is in the same order of magnitude as the dissociation constant of the C-terminal binding domain Vpr^75-90 does not indicate that oligomerization of Vpr influences the interaction with CypA significantly.

Previous studies have shown that the mutants Vpr P35A and Vpr R80A were key residues for coimmunoprecipitation of Vpr and CypA, and when replaced abrogated the coimmunoprecipitation to CypA 12 . Although, these authors concluded that residues beyond Vpr^1-40 also are important for binding to CypA, no further explanation was provided. Our data independently confirm the importance of these residues for the interactions and indicate that the C-terminal binding is quantitatively stronger than the N-terminal binding to CypA (Figure 1). However, the fact that the decapeptide Vpr^75-84 fails to bind to CypA (Figure 2) clearly demonstrates that Arg-80 and its nearby surrounding residues are insufficient for the interaction to occur. Reduced binding affinity was also observed for a mutant Vpr^75-90 peptide where Arg-80 was conserved, demonstrating the importance of an intact 16 residue C-terminal binding domain of Vpr.

A simultaneous interaction of both Vpr sites with CypA was suggested as SPR sensorgrams of full-length Vpr interacting with immobilized CypA gave an optimal fit with a bivalent analyte model (Figure 5). Taken together, these data indicate that a simultaneous binding of the N- and C-terminal domains is required for full-length Vpr to interact with CypA. The dissociation constant (K_D) of full-length Vpr was found to be approximately 0.32 μM (Table 1), which implies that the Vpr-CypA interaction may be stronger than the well characterized interaction of HIV-1 CA with CypA, which has been determined previously by SPR spectroscopy (K_D 16 ± 4 μM) 50 . This suggests that the CypA-Vpr interaction has a functional role in the relationship between the host and pathogen.

The access to detailed information of the interaction of full-length Vpr with CypA, based on the experimental characterization of the N- and C-terminal binding domains of Vpr, now allows us to visualize the interaction of full-length Vpr with CypA. As prerequisites, we have used the NMR structure of Vpr 45 and X-ray structure of CypA 51 as rigid units, together with knowledge of the key residues of the Pro-dependent and Pro-independent binding domains of CypA 39 . A cartoon illustrating the Vpr-CypA complex based on these parameters was generated with the ZDOCK algorithm 52 and is shown in Figure 7. This crude model suggests that a folded Vpr structure is required to provide N- and C-terminal binding regions sufficiently close in space in the Vpr-CypA complex (Figure 7) for the cooperative interaction with the two binding sites of CypA. Currently one should keep in mind that the model has its limitations but clearly rationalizes the findings presented here. The structure of Vpr used, which is the potentially variable/flexible component of the model, is the limiting structure at low pH in 30% aqueous acetonitrile 45 . However, a more flexible structure of Vpr, which would be expected under the hydrophilic conditions at physiological pH 8 53 used in the Biacore experiments performed to characterize the interaction, would be beneficial for improving the model and, thus, provide more accurate details of the interactions of the Vpr-CypA complex at the atomic level.

Most residues of the N- and C-terminal binding domains of Vpr are highly conserved among HIV-1 strains (with the exception of Ile-37 of the N-terminal binding domain and Arg-77, Val-83, Thr-84, Arg-85, Gln-86 and Ala-89 of the C-terminal binding 54 ), indicating that maintenance of these structural regions of the protein is important for the viral life cycle.

The N-terminal Vpr binding domain is located between helix 1 and 2, and the C-terminal Vpr binding domain is located in an arginine-rich region in proximity to the third α-helical domain of Vpr (Figure 7). Although the biological importance of the binding of Vpr to CypA remains elusive, multiple functions of Vpr are connected to the two regions of the protein which bind cooperatively to CypA, including apoptosis 55 56 57 58 59 60 61 62 63 , reverse transcriptase (RT) activity 36 64 65 66 , replication of R5 tropic HIV-1 38 67 , nuclear localization of the protein 30 31 68 69 70 , G2 cell cycle arrest 30 31 32 33 34 35 36 71 and binding of Vpr to DNA and RNA, which is linked to the ability of activating the ATR (ataxia-telangiectasia and Rad3-related) pathway leading to G2 arrest 72 73 74 75 (Figure 8). The importance of both domains to achieve binding of full-length Vpr to CypA suggests that the biological relevance of the interaction may be associated with other functional interactions of Vpr involving both the N- and C-terminal regions (Figure 8). Previous studies have presented contradictory results regarding the possibility that the interaction of Vpr with CypA is involved in induction of G2 arrest or that it has any effect on Vpr expression 12 13 . However, other studies have independently shown that Vpr induced G2 arrest is dependent on residues belonging to both the N- and C-terminal binding regions of Vpr to CypA 30 31 32 33 34 35 36 71 . Residues within the C-terminal domain of Vpr including Arg-80, which was found essential for retaining a strong binding to CypA in this study, are also considered to be important for increasing RT activity 36 and Vpr induced apoptosis 56 60 61 . Furthermore, residues in the N-terminal binding region of Vpr are important for Vpr induced apoptosis (Figure 8) 57 58 62 , indicating that, both the N- and C-terminal CypA binding domains are involved in this function of Vpr. Moreover, the nuclear localization of Vpr and replication of R5 tropic HIV-1 are related to residues included in both N- and C-terminal CypA binding domains of Vpr (Figure 8). The identification of N- and C-terminal binding domains of Vpr, which cooperatively bind to CypA (Figure 7), should encourage further investigations into the biological relevance of the interaction of Vpr with the host cellular factor CypA, as it is likely that this strong interaction is of importance in the viral life cycle of HIV-1.

CypA may represent a potential key molecule in future antiretroviral therapy since inhibition of CypA suppresses HIV-1 replication. All previous studies of protein-protein interactions involving human CypA have hitherto been limited to Pro-containing substrates. The fact that a non-proline-containing 16-residue region of a protein is able to bind with high affinity to CypA, presented for the first time in this manuscript, changes the view on how CypA is able to interact with other proteins. It is interesting to note that several previously reported key functions of HIV-1 are associated with the identified N- and C-terminal binding domains of the protein to CypA.

SMØS, VW and TF participated in planning the experimental work. SPR measurements were performed by SMØS, OH and analyzed by SMØS, TF and AJR. PH and PH synthesized the peptides. ITC measurement and analysis were performed by MIF and SMØS. US provided the highly pure recombinant CypA that was used in the experiments. MN performed and analyzed MS data. Docking analysis was performed by AJR. SMØS and TF planned and performed the NMR studies and SMØS, VW and TF wrote the manuscript. All authors read and approved the final manuscript.