Alpha-helicoidal HEAT-like Repeat Proteins (αRep) Selected as Interactors of HIV-1 Nucleocapsid Negatively Interfere with Viral Genome Packaging and Virus Maturation

A new generation of artificial proteins, derived from alpha-helicoidal HEAT-like repeat protein scaffolds (αRep), was previously characterized as an effective source of intracellular interfering proteins. In this work, a phage-displayed library of αRep was screened on a region of HIV-1 Gag polyprotein encompassing the C-terminal domain of the capsid, the SP1 linker and the nucleocapsid. This region is known to be essential for the late steps of HIV-1 life cycle, Gag oligomerization, viral genome packaging and the last cleavage step of Gag, leading to mature, infectious virions. Two strong αRep binders were isolated from the screen, αRep4E3 (32 kDa; 7 internal repeats) and αRep9A8 (28 kDa; 6 internal repeats). Their antiviral activity against HIV-1 was evaluated in VLP-producer cells and in human SupT1 cells challenged with HIV-1. Both αRep4E3 and αRep9A8 showed a modest but significant antiviral effects in all bioassays and cell systems tested. They did not prevent the proviral integration reaction, but negatively interfered with late steps of the HIV-1 life cycle: αRep4E3 blocked the viral genome packaging, whereas αRep9A8 altered both virus maturation and genome packaging. Interestingly, SupT1 cells stably expressing αRep9A8 acquired long-term resistance to HIV-1, implying that αRep proteins can act as antiviral restriction-like factors.


Selection of αRep binders of the HIV-1 Gag C-terminal domain.
Our viral bait, abbreviated CA 21 -SP1-NC, spanned from residue L343 to F433 of the Pr55Gag sequence (Fig. 1a). GST-fused CA 21 -SP1-NC was used for the selection of specific binders from a phage-displayed, random library of artificial αRep proteins generated by combinatorial methods 26,27 . After three rounds of phage biopanning, 30 individual clones out of an original library of ~1.7 × 10 9 independent bacteriophages, were randomly picked and analyzed by ELISA. Eighteen clones were found to react positively with the immobilized bait, and six of them, αRep4D4, -4E3, -4F2, 5A12, -5F1 and -9A8, showed a significantly higher signal compared to the twelve other clones. The genes of these six αRep were sub-cloned into the pQE-31 expression vector for the production of His-tagged αRep proteins in E. coli. After purification by affinity chromatography on Ni 2+ -column, the six selected αRep proteins were analyzed for their binding activity to GST-CA 21 -SP1-NC in vitro, using an indirect ELISA method. Out of the six clones, αRep4E3 and αRep9A8 showed the highest binding reaction with the viral target (not shown), and were therefore kept for further characterization. DNA sequencing showed that αRep4E3 and αRep9A8 had in common a His-tag at their N-terminus, the flanking N-terminal cap (N-cap) and C-terminal cap (C-cap) sequences, and the constant regions of the internal repeat modules (Fig. 1b). αRep4E3 had seven internal repeat modules, 284 amino acids (including N-cap and C-cap) and a molecular mass of 32 kDa, while αRep9A8 had six internal repeats, 253 amino acids and a molecular mass of 28 kDa (Fig. 2).
Protein pull-down assays. The interaction of αRep4E3 and αRep9A8 with the viral target and the mapping of their binding sites on CA 21 -SP1-NC was investigated using protein pull-down assays. Purified His-tagged αRep proteins were incubated with the full-length viral target CA 21 -SP1-NC, or its carboxy-terminal deletants, GST-CA 21 -SP1-NCΔZF2 (lacking the downstream ZF2), and GST-CA 21 -SP1 (lacking the whole NC domain; see Supplementary Table S1). Protein complexes were isolated on glutathione-coated agarose beads, and analyzed by SDS-PAGE and Western blotting. Both αRep4E3 and αRep9A8 bound equally well to CA 21 -SP1-NC, but no binding was detectable with the NC-deleted target (Fig. 3a). Deletion of ZF2 significantly decreased the binding of both αRep, suggesting an overlap of their binding sites on the Gag target. However, differences were observed between the two binding patterns. The amounts of αRep complexed with GST-CA 21 -SP1-NCΔZF2 were 5-fold lower for αRep4E3 and 2-fold lower for αRep9A8, compared to their complex with full-length GST-CA 21 -SP1-NC (Fig. 3b). This suggested that the major binding determinants of αRep4E3 mapped to ZF2, with minor determinants in ZF1, whereas for αRep9A8, the binding determinants were distributed between the ZF1 and the ZF2 domains.
Effects of αRep4E3 and αRep9A8 proteins on HIV-1 VLP assembly and extracellular release: quantitative aspects. The effects of the two αRep on HIV-1 assembly were explored using a baculovirus-insect cell system of production of HIV-1 virus-like particles (VLPs). In this system, Sf9 cells were infected with AcMNPV gag , a recombinant vector derived from the baculovirus AcMNPV (Autographa californica Multiple Nuclear Polyhedrosis Virus). AcMNPV gag expresses the full-length, wild-type, N-myristoylated HIV-1 Gag polyprotein in Sf9 cells, resulting in the assembly and extracellular release of membrane-enveloped VLPs at high yields. The high number of particles produced per cell in this heterologous model system of VLP assembly and egress, allowed quantitative analyses of morphologically normal VLPs compared to aberrant particles 31,33,[48][49][50][51] .
Expression of αRep in insect cells. The αRep4E3 and αRep9A8 genes were fused to the GFP gene and to the coding sequence of a six-histidine tag at the C-terminus (Fig. 1c). The coding sequence of an N-myristoylation signal 22,52 was inserted to the 5′-end of the αRep-GFP genes (Fig. 1c). The resulting genetic constructs were transfected into Sf9 cells, and the two Sf9-derived cell lines thus obtained, Sf/(Myr+)αRep4E3-GFP and Sf/ (Myr+)αRep9A8-GFP, stably expressed N-myristoylated (Myr+)αRep4E3-GFP and (Myr+)αRep9A8-GFP proteins under the control of the baculoviral OpIE2 promoter (see Supplementary Fig. S1). The aim of the spanning residues L343 to F433 of HIV-1 Pr55Gag, with its amino acid sequence (HIV-1 LAI isolate) shown underneath. Amino acid residues of the SP1 domain are in bold; the cysteine and histidine residues of the zinc fingers (ZF1 and ZF2) responsible for the Zn coordinates are in red; the residues of the basic motif separating ZF1 and ZF2 are in blue. (b) Three-dimensional model of an αRep molecule. The constant regions N-cap and C-cap are shown in red, and the variable region, comprising of four internal alpha-repeats in this particular model (n = 4), is in yellow and green. (c) Schematic representation of an αRep-GFP fusion protein, with the N-terminus modified to carry a N-myristoylation signal (Myr), and a His-tag at the C-terminus. The respective number of amino acids (aa), and molecular weight of αRep4E3, αRep9A8 and GFP domains are indicated in kDa.
N-myristoylation of αRep4E3-GFP and αRep9A8-GFP was to compensate for the difference in cell content between the N-myristoylated Pr55Gag polyprotein, expressed at high levels under the control of the strong polyhedrin promoter 31,33,[48][49][50][51] , and αRep proteins, expressed at lower levels under the control of the weaker OpIE2 promoter. Addressing (Myr+)αRep proteins to the plasma membrane was meant to promote the interaction between Pr55Gag and αRep proteins at the assembly sites of membrane-enveloped VLPs.
Quantitative effect of (Myr+)αRep4E3 and (Myr+)αRep9A8 on VLP assembly. Control, nonmodified Sf9 cells, Sf/(Myr+)αRep4E3-GFP and Sf/(Myr+)αRep9A8-GFP cells were infected with AcMNPV gag , and the amounts of VLPs assembled and released into the culture medium were estimated by SDS-PAGE and Western blot analysis of pelletable, extracellular Pr55Gag [31][32][33] . The intensity of the Pr55Gag band on immunoblots did not radically differ between the three samples (data not shown). A quantitative, luciferase-based assay of extracellular VLPs was then performed. The principle of this assay resides in the co-packaging of Vpr (and Vpr-fused proteins) with Pr55Gag into membrane-enveloped VLPs banding at a density of 1.18. This co-packaging is mediated by the specific interaction between the Gag p6 domain and Vpr 53,54 . The co-expression of the Luciferase-Vpr fusion protein together with Pr55Gag by co-infection of Sf9 with the recombinant vectors AcMNPV luc-vpr and AcMNPV gag showed that the activity of luciferase co-incorporated with Pr55Gag into VLPs directly reflected the amounts of VLPs released in the extracellular medium 33 .
Control Sf9 cells, Sf/(Myr+)αRep4E3-GFP and Sf/(Myr+)αRep9A8-GFP cells were co-infected with AcMNPV gag and AcMNPV luc-vpr , and harvested at 48 hrs pi. Cell lysates and gradient fractions were assayed for luciferase activity, and the results of VLP-associated luciferase activity were normalized to the cellular luciferase levels 33 . A peak of luciferase activity was detected at the density of VLPs in all three culture media and at similar levels (see Supplementary Fig. S2), which implied minor differences in the amounts of VLPs recovered from the different cell lines. This suggested that (Myr+)αRep4E3 and (Myr+)αRep9A8 were not able to completely block the assembly and extracellular release of VLPs from AcMNPV gag -infected insect cells.
Effects of αRep4E3 and αRep9A8 on HIV-1 VLP assembly and extracellular release: qualitative aspects. The morphology of VLPs released from αRep-expressing cells were examined under the electron microscope (EM). From our previous studies 48,50,51 , membrane-enveloped VLPs released from the plasma membrane of AcMNPV gag -infected Sf9 cells appeared as quasi-spherical in shape, and homogeneous in size, with a diameter ranging from 110 to 130 nm (Fig. 4). By contrast, Sf/(Myr+)αRep4E3-GFP cells infected with  Aliquots of bacterial cell lysates containing the full-length target GST-CA 21 -SP1-NC, or the carboxy-truncated mutants GST-CA 21 -SP1-NC∆ZF2 or GST-CA 21 -SP1 were incubated with purified αRep4E3 or αRep9A8, and protein complexes, isolated on glutathione-coated agarose beads, were analyzed by SDS-PAGE and Western blotting. (a) Blot reacted with rabbit anti-αRep antibody, followed by alkaline phosphatase-conjugated goat anti-rabbit IgG antibody. Asterisks indicate the immunoreactive bands of αRep4E3 and αRep9A8 proteins. Note the doublet band of αRep4E3 protein, suggesting some degree of proteolytic cleavage. (b) Bar graph of the scanning of αRep4E3 and αRep9A8 proteins bound to the different viral targets. The level of αRep binding to the full-length target GST-CA 21 -SP1-NC was attributed the 100% value. AcMNPV gag produced a majority of aberrant VLPs, irregular in size and shape, with the frequent occurrence of more than one single immature Gag protein core within the same membrane bud (Fig. 5). Likewise, Sf/(Myr+) αRep9A8-GFP cells infected with AcMNPV gag showed aberrant VLPs (Fig. 6a), or arrested budding morphologies at their surface (Fig. 6b). Interestingly, a peculiar type of pattern was observed in (Myr+)αRep9A8-expressing cells, consisting of (i) the accumulation of αRep4E3 and αRep9A8 material at the inner leaflet of the plasma membrane (Figs 6b and 7a), and (ii) intravesicular budding of VLPs which seemed to be morphologically normal ( Fig. 7a,b). This suggested that one of the consequences of the (Myr+)αRep9A8-mediated blockade of normal VLP budding at the plasma membrane was the redirection of Gag to the internal membranal compartment, and the budding of VLPs into intracytoplasmic vesicles.
Taken together, the results of these analyses suggested that both αRep4E3 and αRep9A8 negatively interfered with the assembly of Gag proteins into immature HIV-1 particles. These negative effects were not quantitative, as there was no significant difference in the yields of extracellular VLPs compared to control cells. The effects of αRep4E3 and αRep9A8 on Gag assembly were qualitative, as shown by the alteration of the structure of VLPs and the change in their cellular sites of assembly. In addition, EM analysis suggested two distinct phenotypes for αRep4E3 and αRep9A8.
Co-encapsidation of Gag and N-myristoylated αRep proteins into VLPs. Since defective VLP assemblies were observed in the presence of αRep4E3 and αRep9A8, we hypothesized that these αRep proteins could be co-packaged with Gag precursor, resulting in altered VLP structure. To test this hypothesis, VLPs released from AcMNPV gag -infected Sf/(Myr+)αRep4E3-GFP and Sf/(Myr+)αRep9A8-GFP cells were purified by ultracentrifugation, and analyzed by SDS-PAGE and immunoblotting, using anti-Gag antibody to detect the Gag polyprotein, and histidine antibody to detect the His-tagged αRep-GFP proteins. An anti-histidine immunoreactive protein was detected in both types of VLP samples, at 59 and 55 kDa for αRep4E3-GFP and αRep9A8-GFP, respectively. These values corresponded to the respective molecular weights of the GFP-fused αRep proteins. The variations in the signal intensity of the 59-kDa and 55-kDa proteins throughout the gradient fractions followed the same pattern as that of the Pr55Gag band, with a maximum intensity in fractions of density ρ = 1.18 (see Supplementary Fig. S4). This suggested that N-myristoylated αRep4E3-GFP and αRep9A8-GFP proteins were able to interact with Pr55Gag at the VLP assembly sites, and to be copackaged with Pr55Gag into

Antiviral activity of αRep4E3 and αRep9A8 proteins in human T-cells challenged with HIV-1.
The genes encoding αRep4E3-GFP, αRep9A8-GFP and αRep9C2-GFP (GFP-fused irrelevant αRep used as negative control) were transferred into SupT1 cells. Three αRep-expressing cell lines were successfully established, which were referred to as SupT1/αRep4E3-GFP, SupT1/αRep9A8-GFP and SupT1/αRep9C2-GFP cells, respectively (see Supplementary Fig. S5). Of note, the non-N-myristoylated versions of αRep proteins were expressed in SupT1 cells to allow them to diffuse and exert their possible antiviral activity in all cell compartments. The cells were infected with HIV-1 and the effect of the αRep proteins on HIV-1 infection was assessed by different methods, as described below. As shown by fluorescence microscopy and flow cytometry, the expression of αRep-GFP molecules was maintained at levels close to 100% throughout the time of the viral challenge (21 days; Fig. 8a).
Kinetics of HIV-1 particle production. CAp24-ELISA was used to evaluate the quantities of extracellular viral particles released from HIV-1-infected cells at days (D) 3, 7, 14, and 21 pi (Fig. 8b). In control cells, CAp24 was detected as early as at D7, with a peak at D14 followed by a decrease at D21 pi (Fig. 8b). However, CAp24 was barely detectable in samples from SupT1/αRep4E3-GFP and SupT1/αRep9A8-GFP cell cultures at D14 pi. The amounts detected were about 10-fold lower compared to that of control samples, and the CAp24 reached the levels observed in control D14-samples only at D21 pi. This indicated that in the cells expressing αRep4E3 and αRep9A8, there was a significant delay in the kinetics of HIV-1 replication and viral progeny production.
Extracellular release of HIV-1 genomes. The viral genome copies in the culture media followed the same kinetics as CAp24, with a similar delay observed in the presence of αRep4E3 and αRep9A8 (Fig. 8c). The viral load was maximal at D14 pi in control samples, and decreased at D21 due to cell apoptosis. In culture media of SupT1/ αRep4E3-GFP and SupT1/αRep9A8-GFP cells, the virus loads slowly increased with time, and reached the maximal levels of control samples at D21 pi (Fig. 8c). A net difference between the αRep-induced inhibitory effects was observed at D14 pi, with a more pronounced negative effect with αRep4E3 (ca. 8-fold) compared to αRep9A8 (ca. 2-fold; Fig. 8c). Cytopathic effects of HIV-1 on αRep-expressing SupT1 cells. HIV-1-infected SupT1 cells were maintained in culture for six weeks. They were monitored by phase microscopy for possible morphological alterations at different time points after infection (Fig. 9a). Formation of syncytia and premature cell death were clearly visible in control cell cultures, in which HIV-1 drastically reduced the proportion of live cells. These cellular alterations were delayed in SupT1/αRep4E3-GFP and SupT1/αRep9A8-GFP cultures (Fig. 9a). Cells samples were also collected at 3-day intervals, and controlled for cell count and cell viability. As expected, the cell concentration and viability decreased rapidly in control SupT1 cell cultures after D7 pi, and there was no more viable cells detected at D21 pi (Fig. 9b,c). The diminution in living cell numbers was delayed and more progressive in SupT1/αRep4E3-GFP and SupT1/αRep9A8-GFP cultures, with a decline starting only after D14 pi. A total cell loss was observed at D28 pi in SupT1/αRep4E3-GFP cultures, but a rebound in cell growth was observed after D28 pi in SupT1/ αRep9A8-GFP cultures (Fig. 9b,c).

Proviral DNA status in HIV-1-infected aRep-expressing SupT1 cells. The viral genome integration in HIV-
1-infected SupT1 cells, evaluated at D14 pi by quantitative PCR, showed that Ct values were not significantly different between control and αRep-expressing cells (see Supplementary Table S2). This suggested that resident, intracellular αRep4E3 and αRep9A8 did not prevent the viral genome integration in SupT1 cells, and that the antiviral effects of αRep4E3 and αRep9A8 proteins did not depend on viral integration. Molecular mechanisms of the αRep4E3and αRep9A8-mediated antiviral effects. Two major parameters influence the infectivity of HIV-1: (i) the degree of proteolytic processing of Pr55Gag which controls the transition from immature to condensed, mature core; (ii) the ability of the virus to encapsidate its genomic RNA during its morphogenetic process, and thus the genome content of viral particles. It has been commonly considered that HIV-1, like most retroviruses, produce large numbers of defective particles lacking all or part of the viral genome, and that the infectivity index is usually in the low range of values, ≤1:1,000. However, the ratio of infectious to defective HIV-1 particles has been recently re-evaluated to 1:8 55 . Both Gag maturation and genome content were assayed in particles produced by αRep-expressing cells infected with HIV-1. Proteolytic maturation of the Pr55Gag precursor. The viral protease-mediated processing of HIV-1 Pr55Gag is a sequential and high-order process. The initial cleavage occurs between SP1 and the NC domain, and the secondary (e.g. MA-CA) and tertiary (e.g. CAp24-SP1) cleavages occur at approximately 10-fold and 400-fold lower rates, respectively, than the initial cut 56,57 . The extent of Gag processing was assayed using HB-8975, a monoclonal antibody which recognized the C-terminal epitope DTGHSSQVSQNY only on free MAp17 protein, and not on intact Pr55Gag. After cleavage of the MA-CA junction, the exposed epitope can compete with a synthetic peptide of the same sequence for binding to HB-8975. The degree of competition in ELISA correlates with the extent of MA-CA cleavage, and provided a quantitative evaluation of the degree of Pr55Gag maturation, as shown in previous studies [58][59][60] . The results indicated that more than 85% viral particles which egressed from control, HIV-1-infected cells, and 75% from SupT1/αRep4E3-GFP cells, consisted of mature particles (Fig. 10a). This proportion was only 35% for SupT1/αRep9A8-GFP cells (Fig. 10a), suggesting that αRep9A8, but not αRep4E3, negatively interfered with the virus maturation.
Viral genome encapsidation. The efficiency of viral genome packaging in the different SupT1 cell lines was measured by the ratio of extracellular genome concentration to CAp24 concentration at D14 pi, and the results normalized to 25 µg CAp24 per sample, a value which corresponded to 5 × 10 7 to 7 × 10 7 infectious particles 61,62 , and 4 × 10 8 to 6 × 10 8 total particles 55 . The results showed that the RNA packaging function was altered in both SupT1/αRep4E3-GFP and SupT1/αRep9A8-GFP cell lines, but to different levels: a 10-to 13-fold reduction in RNA packaging efficiency was observed in αRep4E3-expressing cells, versus a 3-to 4-fold reduction in αRep9A8-expressing cells (Fig. 10b). This implied that the antiviral activity associated with αRep4E3 was predominantly directed against the viral genome packaging.
Infectivity of the viral progeny. The infectious titer of the viral progeny released from the different SupT1 cell lines was determined using Jurkat-GFP cells. This Jurkat T cell-based reporter cell line expresses CD4 and CXCR4 receptors, and enhanced GFP under the control of the HIV-1 LTR promoter, as a direct and quantitative marker of HIV-1 infection and Tat expression 63 . The results obtained at D10 post-reinfection (Fig. 11, leftmost bars) indicated that the infectious titers of viruses which egressed from SupT1/αRep4E3-GFP (1.4%) and SupT1/ αRep9A8-GFP (1.25%) were significantly lower (15-to 17-fold), compared to that from the two control cell lines, SupT1 (24%) and SupT1/αRep9C2-GFP cells (19.5%). To evaluate the long-term effects of αRep proteins on virus titers, the reinfected Jurkat-GFP cells were maintained in culture, and analyzed by flow cytometry at D14 post-reinfection (Fig. 11, rightmost bars). The virus titers found in SupT1/αRep4E3-GFP samples were in the same range of values as those measured in control cells at D10. However, a significant difference (~2-fold) was observed between Jurkat-GFP cells reinfected with SupT1/αRep4E3-GFP (27.8%) and SupT1/αRep9A8-GFP supernatants (15.4%) at D14 post-reinfection. This indicated that αRep9A8 had a more pronounced negative effect on virus infectivity, compared to αRep4E3. Of note, the apparently low infectious titers of culture supernatants from control SupT1 cells at D14 post-reinfection were due to the massive cell death induced by HIV-1 in Jurkat-GFP cells at this time point, and no conclusion could be deduced from these values.

Resident αRep9A8 protein in SupT1 cells conferred resistance to HIV-1 infection. Cells which
survived from the HIV-1-infected SupT1/αRep9A8-GFP cultures (refer to Fig. 9b,c) were collected at D38 pi, and were found to be healthy under the light and fluorescence microscopes (Fig. 12a). Samples from SupT1/ αRep9A8-GFP culture supernatant at D38 pi were then assayed for extracellular virus yields. Viral particles and genomes were found in these samples, at concentrations almost equivalent to the values determined at D14 pi (Fig. 12b,c). Of note, these values were about 10-fold lower for CAp24 and 3-fold lower for the viral genome copies, compared to the corresponding parameters in control cell cultures at D14 pi (compare Fig. 12b and Fig. 8b, and Fig. 12c and Fig. 8c, respectively). The infectious virus titer of SupT1/αRep9A8-GFP culture supernatants at D38 pi was then determined using Jurkat-GFP indicator cells. The results showed that most of the extracellular HIV-1 particles released from SupT1/αRep9A8-GFP at this time pi were noninfectious (Fig. 12d).

Discussion
In recent studies, we established the proof-of-concept that artificial proteins based on repeat motif scaffolds, could be used as intracellular antiviral agents, and we demonstrated their therapeutic potential as novel inhibitors of HIV-1 22,23,25 . In the present work, we applied this concept to another library of highly thermoresistant protein scaffolds, made up of alpha-helicoidal HEAT-like repeats, called αReps 26, 27,29 . The bait used for the screening of our phage-displayed αRep library was the HIV-1 Pr55Gag-derived polypeptide CA 21 SP1-NC. It contained Gag determinants essential for achieving the late steps of the HIV-1 life cycle, i.e. the alpha-helical spacer peptide SP1 and its two flanking domains, the C-terminal domain of the CA protein and the full-length NC domain. Both SP1 [30][31][32][33][34][35] and NC 44,45 have been used as targets for antiviral drugs against HIV-1.  Two αRep molecules, αRep4E3 and αRep9A8, with high binding activity to CA 21 SP1-NC were isolated from our screen. Both were found to exert a modest but significant inhibition on HIV-1, as shown by various biological assays. Their antiviral effects were independent of the proviral integration and occurred at the post-integration step. Differences were observed in their antiviral properties, as schematized in the model of Fig. 13. The αRep4E3 showed a marked negative effect on the viral genome packaging, while αRep9A8 interfered negatively with Gag maturation and its corollary, virus infectivity. More interesting, αRep9A8-expressing SupT1 cells acquired a long-term resistance to HIV-1, as shown by the absence of detectable infectious particles from the culture medium of αRep9A8-expressing SupT1 cells collected five weeks after HIV-1 challenge.
The mapping of the binding determinants of αRep4E3 and αRep9A8 on the viral target provided clues to their antiviral functions. The antiviral activity of αRep4E3, which mainly concerned the viral genome packaging, was consistent with the localization of its major binding site to the downstream zinc finger ZF2. The binding site of αRep9A8 spanned a wider NC domain, encompassing both ZF1 and ZF2 with a possible SP1 overlapping. The masking of SP1 cleavage sites would explain the negative interference of αRep9A8 with the PR-mediated Gag processing.
Since no binding of αRep9A8 to the GST-CA 21 -SP1 target was detected in our pull-down assays, it could be argued that the removal of the NC domain had altered the structure of SP1, and transformed its α-helical structure 64,65 into a non-structured region. Even though this hypothesis cannot be totally excluded, it was not supported by the observation that isolated SP1 peptide in aqueous solution spontaneously folds into an α-helix without the aid of the NC, in a concentration-dependent manner. This α-helix forms at millimolar protein concentrations, i.e. at values equivalent to those found in virions [66][67][68] . Our results suggested that the negative interference of αRep9A8 with the viral PR resulted from the masking of the SP1 domain, indirectly via steric hindrance, rather than via direct binding.
The binding of αRep9A8 to a relatively large protein domain such as the NC raised a number of issues concerning the structure of the two partners. NMR studies of the three-dimensional structure of the HIV-1 NC protein have shown that the two ZF form a single folded protein domain characterized by spatial contacts between residues belonging to ZF1 and ZF2 69,70 . This globular structure is favoured by a highly conserved proline residue (P408 in HIV-1 LAI sequence) generating a bend in the linker separating the two ZF, and is stabilized by hydrophobic and aromatic interactions between residues belonging to ZF1 and ZF2 69,70 . The binding surface of an αRep with six internal repeats, such as αRep9A8, could be estimated to measure between 1,200 and 1,500 Å 2 , based on structural studies of αRep proteins co-crystallized with their target 26 . Such a surface area would theoretically be sufficient to interact with amino acid residues belonging to the two zinc fingers, and even to the adjacent SP1 linker.
The role of the NC and its two zinc fingers in the temporal control of reverse transcription, viral genome replication and packaging have been extensively documented, as well as their contribution to the structure of the viral core [36][37][38][39][40][71][72][73] . The observed antiviral effects of NC-interactors such as αRep4E3 and αRep9A8 might result from the blockage of one or several of these multiple functions. The application of real-time observation of fluorescent-tagged Gag in live cells 74 to αRep-expressing live SupT1 cells should help us to dissect the kinetics and molecular mechanisms of the αRep-mediated antiviral effects, in particular their interference with Gag-RNA interaction and Gag assembly. NC also plays a role in the budding and release of HIV-1 particles by recruiting proteins of the cellular machinery necessary for viral budding, such as ALIX/AIP1 46,47 . Interestingly however, αRep4E3 and αRep9A8 showed no detectable negative effect on virus particle egress.
In terms of antiviral activity, the present study confirmed the feasibility of artificially produce antiviral restriction-like factors 22,23,25 . However, the immunogenenicity of such molecules will need to be assessed before envisaging any therapeutic applications. As a preliminary evaluation, we have performed the T-cell epitope prediction search for αRep4E3 and αRep9A8 (http://tools.iedb.org/processing/help/), and found negative MHC score values for both of them (unpublished data). The results of our comparative study of αRep4E3 and αRep9A8 properties suggested that, even though αRep4E3 was more detrimental to viral genome packaging compared to αRep9A8, the therapeutic advantage would be in favour of the virus maturation-interfering αRep9A8. In addition, the finding that αRep9A8 had the capability to block the production of infectious viral particles, and prolong the survival of HIV-1-infected cells, could significantly influence future strategies and choice of antivirals in anti-HIV-1 cell therapy.   (1); the antiviral effects of the αRep4E3 and αRep9A8 proteins are displayed in pathways (2) and (3). αRep4E3 (blue symbol) negatively interferes with Gag-RNA interaction and genome packaging; αRep9A8 (pink symbol) acts as a virus maturation inhibitor. Of note, the model presented in this figure was built by the first author, using a combination of elements available from the Servier Medical Art site http://smart.servier.com/. Servier Medical Art by Servier is licensed under a Creative Commons Attribution 3.0 Unported License (https://creativecommons.org/licenses/ by/3.0/). As specified, users are free to share, copy and redistribute the material in any medium or format, and to adapt, remix, transform, and build upon the material for any purpose, even commercially.

Eukaryotic cells. Insect cells. Spodoptera frugiperda (Sf9) cells (Life Technologies
SCIenTIFIC REPoRTS | 7: 16335 | DOI:10.1038/s41598-017-16451-w Bacterial cells and plasmid vectors. XL-1 Blue bacteria (Stratagene) were used as the host strain for generating the αRep phage library and for phage amplification. M15[pREP4] bacteria (Qiagen) were used to produce H 6 -tagged αRep proteins. E. coli BL21 was used as the host strain for producing GST-tagged proteins of HIV-1 Gag domains. The pQE-31 plasmid (Qiagen) was used for the production of H 6 -tagged recombinant αRep proteins. The pGEX-4T1 plasmid (GE Healthcare Life Sciences) was used to produce GST-tagged recombinant proteins at high levels.

Viral vectors. Baculoviral vector. Autographa california multiple nuclear polyhedrosis virus (AcMNPV)
has been used to construct the recombinant baculoviral vector AcMNPV gag expressing the full-length, N-myristoylated wild-type HIV-1 Gag polyprotein, abbreviated Pr55Gag. The production of recombinant Pr55Gag and VLPs in AcMNPV gag -infected Sf9 cells have been descrived in detail in previous studies 48,50,51 .

Generation of HIV-1 target proteins. Our viral targets consisted of various glutathione-S-transferase
(GST)-fused polyproteins derived from the HIV-1 GagPr55 precursor (LAI isolate). GST-CA 21 -SP1-NC included the 21 amino acids of the C-terminal domain of the capsid protein (CA), fused to the SP1 linker and the whole nucleocapsid (NC) domain, encompassing its two zinc fingers (ZF) until residue F433 (Fig. 1a). A panel of C-truncated mutants of GST-CA 21 SP1-NC were constructed by PCR amplification, using specific oligonucleotide primers (see Supplementary Table S1). The amplification products were purified by GeneJET PCR Purification kit, and a first step gene cloning was performed using InsTAclone PCR cloning kit (ThermoFisher Scientific, Waltham, MA). XL-1 Blue competent cells were transformed by the ligation products, and colonies were picked for PCR amplification, restriction enzyme analysis using BamHI and XhoI, followed by DNA sequencing. Each construct was isolated from the pTZ57R/T vector and transferred to pGEX-GST expression vector using the same restriction sites, BamHI and XhoI. The step of clonal identification was performed as previously described 21,22 .
Three pGEX-GST-gag plasmids were constructed: (i) pGEX-GST-CA 21 -SP1-NC; (ii) pGEX-GST-CA 21 -SP1-NCΔZF2; and (iii) pGEX-GST-CA 21 -SP1. BL21 cells were used for recombinant Gag protein production. BL-21 harboring each of the recombinant pGEX-GST-gag plasmids were cultured in 200 mL LB broth containing 100 µg/mL ampicillin and 1% (w/v) glucose, at 37 °C with shaking. When OD 600 reached 0.8, protein expression was induced by addition of 0.1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG), and maintained in culture at 30 °C overnight with shaking. Cells were pelleted by centrifugation at 1,200 × g for 30 min at 4 °C. Cell pellets were resuspended in PBS containing a cocktail of protease inhibitors (Roche Diagnostics GmbH), then lysed by sonication. Soluble proteins were purified by affinity chromatography on glutathione-agarose gel (ThermoFisher Scientific), following the protocol of the manufacturer, and analyzed by SDS-PAGE and Western blotting.
Screening of the αRep library on the viral target. The construction of the αRep phage library 2.1 has been described in a previous study 27 . In brief, the αRep library was constructed by polymerization of synthetic microgenes corresponding to individual HEAT-like repeats, and the αRep proteins were expressed at the surface of M13-derived filamentous phages (phage display). In terms of diversity, our αRep library is estimated to contain 1.7 × 10 9 independent clones.
For the αRep library screening, purified GST-CA 21 SP1-NC protein which represented our viral target was diluted at 5 µg/mL in coating buffer (TBST: 20 mM Tris/HCl, pH 8.0, 150 mM NaCl, 0.05% Tween-20) was immobilized on glutathione-coated microtiter ELISA plate (ThermoFisher Scientific) by incubation overnight at 4 °C in a moisture chamber. The coated wells were washed four times with TBST, and saturated with blocking solution (2% BSA in TBST; 200 µL/well) for 1 hr, after which an aliquot of the phage library was added to the GST-CA 21 SP1-NC coated wells, and incubated at room temperature (RT) for 1 hr with shaking. Several washes of the wells were then performed with TBST, and bound phages were eluted by three successive rounds of adsorption/elution. Phage elution was performed by acidic solution (0.1 M glycine-HCl buffer, pH 2.5) in the first two rounds. The last phage elution consisted of specific ligand elution 76,77 , by using a large excess of GST-CA 21 SP1-NC as the competitor. The population of αRep-displayed phages eluted from the GST-CA 21 SP1-NC bait was then amplified and subcloned in XL-1 Blue cells, as described for ankyrin scaffolds 21,22 .
Individual phage clones were selected and amplified as previously described 21,22 , and their respective binding activity towards the GST-CA 21 SP1-NC target was determined by ELISA. 100 µL-aliquots of purified GST-CA 21 SP1-NC (5 µg protein/mL) were diluted in PBS and dispensed into the wells of a glutathione-coated microtiter plate, then incubated overnight at 4 °C. The coated plate was washed four times with TBST, then blocked with TBST-BSA (200 µL per well) at RT for 1 hr with shaking. After a washing step, 100 µL-aliquots of each phage culture supernatant were added to the wells and incubated at RT for 1 hr, followed by HRP-conjugated mouse anti-M13 (GE Healthcare Life Sciences) diluted to 1:5,000 in TBST-BSA (100 µL-aliquot per well), and incubation proceeded at RT for an extra 1 hr. The wells were washed again, prior to the addition of 100 µL SureBlue ™ TMB Microwell Substrate (KPL, Gaithersburg, MD). Reaction was stopped with 1 N HCl, and absorbance measured at 450 nm. Phage clones showing a high binding activity towards the immobilized target were sequenced and kept for cytoplasmic expression of individual αRep proteins.
Expression and purification of αRep proteins. The αRep genes corresponding to strong binders to the GST-CA 21 -SP1-NC target were sub-cloned in pQE-31 plasmids, and used for transforming the bacterial strain M15[pREP4] strain, as previously described 21,22

SDS-PAGE and Western blotting.
Samples normalized to equal amounts of total protein were separated by electrophoresis in SDS-containing 12%-or 15%-polyacrylamide gel. Gels were stained with PageBlue TM protein staining solution (ThermoFisher Scientific), or used for Western blotting (WB). For WB, proteins separated on SDS-gels were electrically transferred to PVDF membrane (GE Healthcare, UK). After blocking with 5% skimmed milk in TBST, the PVDF membranes were incubated with various primary and secondary antibodies. GST-tagged proteins were detected using rabbit anti-GST antibody (Applied Biological Materials Inc., Richmond, Canada [ABM]). Gag proteins were detected using rabbit anti-Pr55Gag antibody (laboratory-made [31][32][33]. αRep proteins were detected using mouse monoclonal anti-H 6 tag (ABM), in the conditions recommended by the manufacturer, or rabbit polyclonal anti-αRep protein (laboratory-made). Secondary antibodies consisted of HRP-or phosphatase-conjugated goat anti-mouse or anti-rabbit IgG antibody (KPL, Gaithersburg, MD), followed by specific chromogen reaction and color development. The protein bands on blots were scanned using the Photo Scan Lite application and the average relative band intensity values were obtained using Image Studio Lite Software Version 5.2.5.
Construction of Sf9-derived cell lines stably expressing N-myristoylated αRep proteins. The integrative plasmid pIB/V5-His-TOPO vector (Life Technologies) was used for stable expression of GFP-tagged and N-myristoylated αRep proteins in Sf9 cells, under the control of the OplE2 promoter. The sequence coding for a N-myristoylation signal 22,25 was inserted at the 5′-end of the αRep-GFP genes to obtain N-myristoylated recombinant proteins, abbreviated (Myr+)αRep-GFP (Fig. 1c). Sf9 cells were transfected with the pIBV5-His-TOPO-based vectors (2.5 µg/10 6 cells), using the DOTAP Liposomal Transfection Reagent (Roche Diagnostics GmbH). Transfected cells were maintained in complete Grace's insect cell medium containing blasticidin (ThermoFisher Scientific) at 100 µg/mL. Two Sf9-derived cell lines were thus generated, referred to as Sf/ (Myr+)4E3-GFP and Sf/(Myr+)9A8-GFP (see Supplementary Fig. S1). The expression of these proteins was monitored by fluorescence microscopy, and confirmed by SDS-PAGE and Western blot analysis using rabbit polyclonal anti-αRep antibody (laboratory-made).

Isolation of extracellular HIV-1 virus-like particles (VLPs). Sf9 cells were infected with AcMNPV gag
at an input multiplicity of 10 PFU/cell (MOI 10). Cells were harvested at 48 hrs postinfection (pi), and the culture supernatants were clarified by low-speed centrifugation. VLPs were recovered from the clarified supernatants using a two-step procedure comprising (i) a sucrose-step gradient centrifugation, followed by (ii) isopycnic ultracentrifugation [31][32][33]78 . In step (i), VLPs were pelleted through a sucrose cushion (20%, w:v, in TNE buffer; TNE: 100 mM NaCl, 10 mM Tris-HCl pH 7.4, 1 mM Na 2 EDTA) at 30 krpm for 1 hr at 15 °C in a Beckman SW55 rotor. For step (ii), pelleted VLPs were gently resuspended in 0.20-0.25 mL PBS, and centrifuged in a linear sucrose-D 2 O gradient. Linear gradients (10-mL total volume, 30-50%, w:v) were centrifuged for 18 hrs at 28 krpm in a Beckman SW41 rotor. The 50% sucrose solution was prepared in D 2 O buffered to pH 7.2 with NaOH, and the 30% sucrose solution was prepared in 10 mM Tris-HCl, pH 7.2, 150 mM NaCl, 5.7 mM Na 2 EDTA. Aliquots of 0.5 mL were collected from the top, and fractions analyzed for protein content by SDS-PAGE and immunoblotting, and by luciferase assays, as detailed below.
Luciferase-based quantitative assay for HIV-1 Gag assembly and extracellular release of VLPs. The co-expression of the Luc-Vpr fusion protein by AcMNPV LucVpr results in the incorporation of Luc-Vpr into Gag VLPs via the interaction of Vpr with the Gag p6 domain 33 . The levels of luciferase activity present in the pelletable fraction from the culture supernatant represented the amount of VLPs released. Aliquots (10 6 ) of Sf9 (control cells), Sf/(Myr+)4E3-GFP and Sf/(Myr+)9A8-GFP cells were co-infected with two recombinant baculoviruses, AcMNPV gag and AcMNPV LucVpr , at equal multiplicity of infection each (MOI 10). Co-infected cells were harvested at 48 hrs pi, and the amounts of extracellular VLPs released in the culture medium were quantitated using luciferase assays. VLPs were pelleted and lysed in KDT buffer (KDT: 0.1 M potassium phosphate buffer, pH 7.8, 1 mM DTT, containing 0.2% Triton × 100) for 30 min at 37 °C with vortexing every 10 min. The VLP-associated luciferase activity was measured using luciferin-ATP substrate (VivaGlo ™ Luciferin, Promega) and the Lumat LB-9501 luminometer (Berthold Technologies, Bad Wildbad, Germany), as previously described 33 , and results expressed as relative light units (RLU) per μg protein. In order to correct for possible variations in the cellular expression of LucVpr, the values obtained with VLPs were normalized to the luciferase activity measured in the corresponding cell lysates, also expressed as RLU/μg protein.
Grids were examined under a Jeol JEM-1400 electron microscope, equipped with an ORIUS ™ digitalized camera (Gatan France, 78113-Grandchamp). For statistical EM analyses, a minimum of 50 grid squares containing 10 to 20 cell sections each were examined for counting VLPs budding at the cell surface, released in the external milieu or remaining in intracellular compartments [31][32][33]48 .
Flow cytometry. The efficacy of cell transfection or transduction by lentiviral vectors was evaluated by flow cytometry. Cell samples were harvested at 48 hrs post-transfection or transduction. Cells were washed once with PBS, then with PBS containing 1% FBS-0.02% NaN 3 . After the washing steps, the cells were resuspended in 1% paraformaldehyde in PBS and analyzed by flow cytometry, using a BD Accuri TM C6 flow cytometer (BD Biosciences).

Antiviral activity of αRep proteins in HIV-1-infected SupT1 cells. Control SupT1 cells and SupT1
cells stably expressing αRep4E3, αRep9A8 or αRep9C2 were maintained in growth medium for at least 4 weeks before incubation with HIV-1 NL 4-3 virus inoculum at MOI 1 for 16 hrs. After rinsing with serum-free medium, the cells were resuspended in fresh growth medium, and maintained in culture for several weeks. HIV-1 particle yields were determined in the cell culture supernatants collected at 3-day intervals, using CAp24-ELISA (Bio-Rad, Marnes-la-Coquette, France), and viral genome copy numbers were assayed using the COBAS ® AmpliPrep/COBAS ® TaqMan HIV-1 Test v2.0 (Roche Molecular Systems, Branchburg, NJ).
The evaluation of viral genome integration could not be performed by using conventional Alu-Gag PCR assays on αRep-expressing SupT1 cells. These stable cell lines were generated by using the third-generation lentiviral vectors, which contain part of the U5 region required for the integration of the αRep gene into the cell chromosome. As a consequence, this region would give a positive signal in Alu-Gag PCR assays, even with noninfected cells. To avoid this drawback, viral genome integration was evaluated using SYBR green-based qRT-PCR assays performed on high molecular weight DNA isolated from infected cell lysates, with a pair of primers specific to an internal sequence of the pol-prt gene (see Supplementary Table S2). The kit used in these assays was the High Pure PCR Template Preparation Kit (Roche Product No. 11796828001), which has been developed to recover high molecular weight DNA from cellular genomic DNA, ranging from 30 to 50 kbp and free of contaminations with unintegrated viral cDNA.
The cell viability was evaluated using Trypan blue dye exclusion assays and the Countess ™ Automated Cell Counter (Invitrogen, Thermo Fisher Scientific).
SCIenTIFIC REPoRTS | 7: 16335 | DOI:10.1038/s41598-017-16451-w ELISA-based HIV-1 Gag processing assay. The ELISA-based assay used to quantitate the extent of Gag processing relied on HB-8975, an anti-matrix (MA) monoclonal antibody issued from the hybridoma clone MHSVM33C9/ATCC HB-8975, and obtained from the American Type Culture Collection (ATCC, Manassas, VA). HB-8975 fails to react with its epitope (DTGHSSQVSQNY) located immediately upstream to the MA-CA junction when the MA domain is embedded in the Pr55Gag polyprotein precursor. After PR-mediated cleavage of the MA-CA junction, this epitope becomes fully exposed and accessible to HB-8975, and can compete with a free synthetic peptide mimicking its amino acid sequence. The degree of competition in ELISA correlates with the extent of MA-CA cut, and provides a quantitative evaluation of the degree of Pr55Gag maturation, as shown in previous studies [58][59][60] .
Microtiter plates were coated with 100 µL anti-MA HB-8975, diluted at 5 µg/mL in coating buffer (1 M NaHCO 3 pH 9.6), and left overnight at 4 °C in a moisture chamber. Wells were washed 3 times with PBS-T, and incubated with 200 µL blocking solution (2% BSA in PBS) at RT for 1 hr to prevent non-specific binding. Next, 60 µL-aliquots of cell culture medium, normalized to the same CAp24 titer (80 µg/mL), were pre-incubated with Triton X-100 (1% final concentration) and 10 ng/mL biotinylated synthetic MAp17 C-terminal peptide at RT for 10 min. The mixtures were then added to the wells and incubation proceeded for 1hr. The binding reaction was revealed by incubation with HPR-conjugated streptavidin at RT for 1 hr, followed by addition of 100 µL-aliquots of TMB microwell peroxidase substrate. The reaction was stopped with 1 N HCl, and the optical density measured at 450 nm.
Infectivity assays. Infectious titers were determined using JLTRG cells, a CD4/CXCR4-expressing Jurkat T cell-based reporter cell line, which harbors the enhanced GFP gene under the control of the HIV-1 LTR promoter. JLTRG cells (referred to as Jurkat-GFP cells in the present study) are Tat-dependent indicator T-cells which express GFP upon infection with HIV-1 63 . Samples from culture supernatants were adjusted to equal amounts of CAp24 (15 µg/mL) by dilution into fresh C-RPMI. Jurkat-GFP cells were seeded at 2.5 × 10 5 cells/ well in 96-well culture plate containing 8 µg/mL Polybrene (Sigma-Aldrich, St. Louis, MO). 100 μL-aliquots of the CAp24-normalized culture supernatants were added to the wells, and further incubated at 37 °C and 5% CO 2 for 16 hrs. At 16 hrs pi, 100 μL-aliquots of fresh C-RPMI was added to the wells. Jurkat-GFP cells were harvested at D14 pi and assayed for GFP-positivity by flow cytometry 63,[79][80][81] .