1. IntroductionNascent viral particles of human immunodeficiency virus type-1 (HIV-1) released from HIV-1-producing cells are immature and non-infectious [1,2]. The particles contain Gag and GagPol polyproteins, regulatory viral proteins (Vif, Vpr, Vpu, Tat, Rev, and Nef), and two molecules of single-stranded genomic RNAs with host proteins [2,3,4]. Gag and GagPol polyproteins and the regulatory viral proteins are encoded in a single genomic RNA. GagPol polyproteins are produced by a −1 nucleotide (nt) ribosomal frameshifting event during the translation of the Gag protein [5]. A ratio of translated Gag to GagPol is generally known as 20:1 in HIV-1-producing cells [6]. The maintenance of the ratio is important for HIV infectivity [7]. Gag and GagPol target the plasma membrane via the myristoylated Gly of matrix protein (MA) at the N terminus of the polyprotein. Gag consists of MA, capsid (CA), nucleocapsid (NC), the spacer peptides p2 and p1, and p6 domains; GagPol comprises MA, CA, NC, p6*, p2, p1, protease (PR), reverse transcriptase (RT), RNase H (RH), and integrase (IN) [1]. Each domain is concatenated via the sequences of PR cleavage site to form the polyprotein chains [1]. Gag and GagPol assemble into a hexagonal lattice via its CA domain, recruiting viral proteins and the genomic viral RNAs. During the assembly process, Gag and GagPol translocated to the membrane form buds with the curved Gag lattices on the cell surface. These structures are formed corresponding to the bending of the membrane at the site, followed by the recruitment of the endosomal sorting complex III (ESCRT-III) components required for transport by the p6 domain of Gag, and induce membrane scission and the release of nascent virus particles [2,8,9,10,11].The nascent immature HIV-1 particles mature into infectious viruses through a sequential processing step of the enzymatic cleavage of Gag and GagPol proteins by the mature PR [2,3]. The newly translated, embedded PR in the GagPol polyprotein is not a fully functional enzyme. However, it is converted into a mature functional enzyme via self-cleavage in an intramolecular manner (cis) [12]; this step is called autoprocessing. Once the protein is converted to its active form, it cleaves the Gag and GagPol polyproteins at the cleavage sites and scissors each functional protein inside the virus particle during budding or after releasing [1]. The initiation of the autoprocessing step is primed at the junction between p2 and NC [12,13,14]. This initial processing step is caused via an intramolecular digestion (the embedded PR cleaves the polyprotein at the initial cleavage site on the same polyprotein chain) [12]. The first digestion triggers a cascade of enzymatic reactions to release the functional mature PR. The order of the cleavage sequence is a strict regulation [12,13,14]. The autoprocessing step is suppressed by PR inhibitors [15,16] or regulated by truncated IN [17,18,19,20], amino acid substituted IN [21,22] or RH [22]; thus, the autoprocessing step is considered a target for developing novel anti-HIV drugs.A genome-wide association study of HIV genomic RNA demonstrated that a total of 14 non-synonymous single nucleotide polymorphisms (SNPs) were correlated with plasma viral load (VL) in anti-retrovirus treatment-naive HIV-infected patients [23]. The mutations were located on CA, RH, IN, envelope, and Nef. We previously investigated the roles of each SNP in viral fitness in vitro using a laboratory-adapted HIV clone, HIVNL4.3, as a parent strain, and site-directed mutagenesis was conducted to induce individual or a combination of the SNPs. We found that a mutant containing a Met-to-Ile change at codon 50 of IN (IN:M50I) was impaired. The mutant suppressed the virus release to 0.3% of wild-type HIVNL4.3 (WT) and inhibited the GagPol autoprocessing step. These results indicated that IN regulates virus release and autoprocessing. Other groups reported that IN regulated maturation/processing using IN-deleted/truncated HIV mutants [17,18,19,21]. Therefore, our finding using aa-substituted IN consists of the previous reports. The defect was rescued by the other co-existing VL-associated SNP—Ser-to-Asn change at codon 17 of IN (IN:S17N) [22]. Interestingly, an Asn-to-Ser change at codon 79 of RH (RH:N79S) also restored virus release, autoprocessing, and replication competence of the defective IN:M50I virus [22]. These results indicated that RH is also involved in virus release and autoprocessing, thereby affecting virus fitness.We investigated a population of HIV variants containing the IN:M50I mutation in patients participating in NIAID clinical trials. Surprisingly, the plasma VL level in patients infected with HIV containing IN:M50I without the compensatory mutations (either IN:S17N or RH:N79S) was 1207–201,353 copies/mL [24], suggesting that the clinical isolates containing IN:M50I were not impaired in the virus released from HIV-producing cells, unlike our previous in vitro study. To clarify this discrepancy between the in vitro and in vivo data, we performed a comparative analysis of the RNA genome of the entire pol region. We then found that the recombinant HIV(IN:M50I) virus contained a Val-to-Ile mutation at codon 151 of IN (IN:V151I) in the backbone [24]. IN:V151I is an endogenous aa substitution in HIVNL4.3 [25], and is known as a polymorphic mutation associated with drug resistance [26,27,28,29,30]. Given the reports that a recombinant consensus B HIVNL4.3 IN protein-containing M50I and V151I is functional [31], while our recombinant virus possessing the mutations was defective in virus release and autoprocessing, we hypothesized that backbone mutations in the HIV(M50I/V151I) may regulate virus release, autoprocessing, and replication activities in in vitro studies.

In the current study, we further investigated the mechanism of the defects in virus release and autoprocessing in HIV(M50I/V151I) and reported that the C-terminal domains of RH and IN are associated with the deficiency. These results provide a potential target for the development of anti-HIV drugs/therapy.

4. DiscussionWe have previously demonstrated that a combination of M50I and V151I mutations in IN suppresses virus release by accumulating the abnormal size of buds on the surface of HIV-producing cells. In the presence of the combination, the processing of Gag and GagPol was inhibited at the initiation of autoprocessing [22,24]. RH:N79S or IN:S17N rescued the defect; however, the mechanism of the defective virus is not yet clear. Given the information, in the current study, using truncated mutants, we further investigated the role of each RH and IN domain in the impaired virus. We found that RH:D109 and IN:D288 negatively regulate the defects occurring in the M50I/V151I setting. RH:D109 is a component to form the active pocket of RH and is highly conserved [36]. The sequential deletion of RH in HIV(IN:M50I/V151I) demonstrated that the α4 domain containing RH:D109 regulates the defect of HIV(IN:M50I/V151I). The RH:D109N change in HIV(IN:M50I/V151I) rescued the processing of Gag and GagPol, but significantly suppressed virus fitness. Therefore, a combination of RH:D109 with IN:M50I/V151I induces a defect in the initiation of autoprocessing. Although we illustrated that RH:D109N in HIV(IN:WT) was replication-incompetent, to precisely elucidate the role of the domain in HIV replication, we need further studies using HIV(M50I/V151I) and HIV(IN:WT).A series of HIV(IN:M50I/V15I) variants containing aa substitutions at IN:D288 illustrated that six aa changes (Gly, Asn, Lys, Ala, Glu, Tyr) could play a role as compensatory mutations for virus release, autoprocessing, and virus fitness of HIV(IN:M50I/V151I) without the previously described rescue mutations. In order to define the clinical settings, we conducted a population analysis using NIAID HIV sequences and the LANL HIV database. We previously reported that 17 sequences in LANL (14 subtype B and 3 subtype C) contained IN:M50I/V151I mutations without the rescue mutations [24]. We expected that the 17 sequences contained aa changes at D288; however, only one out of 17 contained D288N mutation, and the others had no change in the sequence. It appears that the 16 sequences may carry other uncharacterized compensatory mutation(s).The results of WB from HIV(ΔIN) and HIV(IN:M50I/V151/ΔCTD) demonstrated that both mutants contained less PR than HIV(WT). By contrast, HIV(IN:M50I/V151I/Δ14aa) and HIV(IN:M50I/V151I/D288G) carried a comparable amount of PR to HIV(WT) (Figure 4). The band intensity of p24 in the WB results from all mutants was similar to that of HIV(WT). Therefore, the loaded protein amounts of the mutants in the WB assay were similar, indicating that HIV(ΔIN) and HIV(IN:M50I/V151/ΔCTD) particles might contain less incorporated GagPol. Both mutants commonly lack aa residues 212–274; thus, the region may be associated with GagPol incorporation in the particles. The IN-deletion studies report that mutations in CCD (aa residues 58–202) regulate GagPol incorporation [19,43,45,46]. Uncharacterized aa residues in the 212–274 region may regulate GagPol incorporation in the virus particles with a similar mechanism by the mutations in the CCD.The sequential deletion of aa residues from the C-terminal of IN(WT) revealed that the IN C-terminal tail region (aa residues 270–288) is required for HIV-1 infection, reverse transcription, and maturation [17,20]. A functional study of IN using recombinant protein identified that Lys273 in the tail region takes part in virus RNA binding with other residues [47]. Our findings provide a further insight into the regulation of virus release and maturation by amino-acid-substituted IN and the tail region of HIV(IN:M50I/V151I). Inhibitors of HIV maturation have been considered as the next-generation anti-virus drugs [48,49]; thus, our finding may provide new insight into developing new drugs.Autoprocessing is initiated at the junction between p2 and NC in intramolecular mechanisms [12]. In our previous work, we demonstrated that HIV(IN:M50I/V151I) suppressed this step without affecting GagPol dimerization [22], indicating that GagPol-containing IN:M50I/V151I may directly or indirectly suppress the autoprocessing after the dimerization. Due to its flexibility and poor solubility of the C-terminal domain of IN, the structure analyses were performed using the tail region-deleted IN, not the intact, full-length of IN [35,38,42,43,44]. To understand the mechanism of the defect of the initiation of autoprocessing in HIV(IN:M50I/V151I), we need to structure an analysis of the full-length of GagPol. Our previous study illustrated that HIV(IN:M50I/V151I), with a modification at the junction between PR and RT in the GagPol, restored the processing [22], suggesting that the tail domain of IN: M50I/V151I may directly or indirectly interact with the junction and induce a conformation change in the GagPol structure. Subsequently, it may inhibit the initiation of autoprocessing. HIV-1 particles incorporate a large array of host proteins during assembly and budding [50,51,52,53,54,55], and those proteins may positively or negatively regulate virus life cycles. The M50I/V151I particles may incorporate a unique host protein via the C-terminal domain of IN:M50I/V151I and interfere with the initiation of autoprocessing, or may lack the incorporation of the host proteins requiring budding [55] via the IN, subsequently suppressing virus release. Currently, a comparative proteomic analysis of HIV particles between HIV(WT) and HIV(IN:M50I/V151I) is underway, which may reveal the regulatory mechanism(s) of this defect.

Using RH-truncated mutants in HIV(IN:WT) and HIV(IN:M50I/V151I), our study demonstrated that the RH domain is involved in virus release; ΔRH suppressed HIV(IN:WT) release, but ΔRH partially rescued HIV(IN:M50I/V151I) release. In contrast, the WB results demonstrated that the amounts of Gag/GagPol processing in both ΔRH mutants were comparable to that of HIV(WT). It appears that the RH domain is not involved in regulating the initiation of autoprocessing. IN:D288G/K/E/Y fully restored virus release, autoprocessing, and viral replication. The IN:D288N change also nearly fully recovered the processing, however, it partially rescued the virus release of HIV(IN:M50I/V151I) to 48.6 ± 6.3% of that of HIV(WT) (p < 0.05, n = 5). In addition, a mutant lacking codon 288, HIV(IN:M50I/V151I/Δ288), was able to restore virus release and autoprocessing, but virus replication was 10~20% of HIV(WT) on day 7. Therefore, the virus release, autoprocessing, and virus replication may be independently regulated in a C-terminal aa-dependent manner.

HIV particles are budding and released from the cell surface by recruiting the host ESCRT system [9,56,57,58,59]. Near 20 ESCRT proteins and other host factors are involved in the process [9,10,11,60,61,62]. Thus, the GagPol-containing IN:M50I/V151I may differentially regulate the ESCRT system and suppress virus release. Recent studies highlight the accessory functions of matured IN [47,63,64]: IN stimulates RT [63], IN interacts with RNA [47,65,66], the acetylation of CTD of IN is associated with the defects in proviral transcription [65]. We now demonstrated a new potential role of IN domain in GagPol polyprotein.We previously observed that HIV(IN:M50I/V151I) was defective in virus release and autoprocessing followed by maturation; thus, the mutant was replication-incompetent [22,24]. The current study focused on determining the amino acid residue(s) that regulate the defects. We found that the C-terminal domains of RH and IN play a pivotal role in the impaired virus. Further studies are required to define the role of each aa in virus replication. However, site-directed mutagenesis study in the M50I/V151I context demonstrated that some mutations partially restored virus release and rescued a nearly complete level of the processing. These results imply that virus release is tightly correlated with autoprocessing (the initiation of processing), and autoprocessing may trigger virus release. Recently, Tabler et al. demonstrated that the embedded PR in GagPol is activated during assembly and budding prior to virus release [16], indicating that PR activation is related with the virus release process. Our results are in line with their report. Further investigations need to reveal the regulatory mechanism of virus release and budding, focusing on the C-terminal domains of RH and IN, which may provide a feasible strategy for the suppression of virus transmission, and may provide a novel insight into developing therapy.