INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The in vivo role of human immunodeficiency virus (HIV) and simian immunodeficiency virus (SIV) Nef proteins has gained increasing consideration in the past few years, both in experimental models (e.g., HIV-infected [35, 38] or nef transgenic SCID mice [25], SIV-infected monkeys [28]) and in humans, in which the presence of HIV genomes expressing heavily mutated or even truncated forms of the nef gene was correlated with an impaired progression of the disease (16, 29, 44). Of note, it has been proposed that the pathogenicity of full-length HIV or SIV strains may be, at least in part, the consequence of the effects that Nef elicits directly on immune system (13).

On the other hand, conflicting results about the role of Nef in in vitro HIV replication have been reported (11, 24, 37, 50). There is now agreement that Nef induces the downregulation of both CD4 and major histocompatibility complex class I molecules (3, 21, 47) and dramatically improves the HIV replication efficiency in "resting" peripheral blood lymphocytes (PBLs). The stimulating effect of Nef in the HIV replication cycle acts in the producer cells at the stage of virus formation and could be appreciated in the target cells at the step of proviral DNA synthesis (2, 12, 46). Conversely, a much smaller enhancing effect on viral replication has been observed in both activated PBLs and immortalized cell lines (11, 37, 50). Furthermore, Nef interacts with several types of cellular kinases (for a review, see reference 43). The enhancement of HIV replication and the CD4 downregulation are the most accurately characterized functions of Nef.

From the provirus of a naturally occurring, nonproducer, and interfering HIV-1 variant, we have recently cloned and characterized an HIV-1 nef allele (F12-HIV nef) whose expression in stably transfected cells fails to downregulate the CD4 receptors and, more originally, induces an antiviral state (15). Interestingly, despite the dramatic phenotypic differences with respect to wild-type (wt) forms, this Nef mutant shows only three amino acid substitutions (9) never detected at the same time in any nef allele sequenced so far. Preliminary studies indicated that the block of infecting HIV induced in cells stably expressing F12-HIV Nef occurs at the stage of viral assembly and/or release (15). In view of the uniqueness of its phenotype, we decided to look more deeply at the molecular mechanism of the HIV inhibition induced by F12-HIV Nef. We analyzed the phenotype of a chimeric HIV molecular clone where the wt nef was replaced by the F12-HIV gene (NL4-3/chi). We established that in-cis expression of F12-HIV nef induces a complete block of viral replication through a mechanism acting at the step of viral assembly and/or release. Our efforts to determine both cellular and viral targets of the F12-HIV Nef inhibitory effect allowed us to obtain strong indications that the CD4-F12-HIV Nef interaction is involved in the F12-HIV Nef-induced viral retention. This was consistently observed by either transfecting or infecting cells lacking the expression of the CD4 intracytoplasmic domain. Furthermore, data obtained by cotransfecting an F12-HIV Nef expression vector together with diverse env mutant HIV molecular clones highlight the Env gp41 intracytoplasmic tail as a major viral target of the F12-HIV Nef inhibitory effect.

A model of HIV inhibition induced by in-cis expression of a mutated nef allele is described here for the first time. This seems only in part reminiscent of the negative trans dominance previously reported for other structural or regulatory HIV protein mutants (i.e., Gag, Rev, and Tat) (20, 33, 51). Furthermore, our results allow us to propose novel molecular interactions among Nef, the CD4 intracytoplasmic tail, and the HIV Env gp41 glycoprotein.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

HIV molecular clones and expressing vectors. In order to obtain the NL4-3/chi chimeric construct, the F12-HIV nef gene was amplified by PCR from the pcDNAI/F12-HIV nef construct (15) with oligoprimers carrying the MluI (5' end) and ClaI (3' end) restriction sites and overlapping the nef initiation and stop codons, respectively. The amplified F12-HIV nef gene was then inserted into a derivative of the pNL4-3 plasmid (1) carrying the MluI (present in a linker inserted between the env stop and the nef start codons) and ClaI (downstream of the nef stop codon, by mutagenizing the ATC GAG sequence to ATC GAT) sites. pNL4-3 molecular clones defective in nef (nef) (11), env (env) (48), the Env gp120 CD4 binding domain (pNL-A1[CD4]) (54), the Env gp41 intracytoplasmic tail (HIV Tr712Env) (53), or the Nef myristoylation signal (pMD) (11) were as previously described. A Nef myristoylation-defective (myr) NL4-3/chi molecular clone was obtained by reproducing the experimental strategy pursued to recover the NL4-3/chi construct, except that the oligoprimer overlapping the F12-HIV nef start codon used for the PCR amplification was designed by including nucleotide changes leading to Ala-Ala consensus in place of the two N-terminal glycines. An env-defective (env) NL4-3chi molecular clone was obtained by replacing the SalI/BamHI region of the NL4-3/chi provirus with the homologous fragment from the wt env HIV molecular clone (48).

Cell cultures and transfections. C8166 and CEMss cells were grown in RPMI 1640 (Life Technologies, Gaithersburg, Md.) supplemented with 10% heat-inactivated fetal calf serum (FCS). 293 cells, HeLa cells and derivatives thereof were grown in Dulbecco modified minimal essential medium supplemented with 10% of heat-inactivated FCS.

HIV infections and detection. Supernatants from transiently transfected 293 cells were used as the source of wt, nef, or NL4-3/chi strains. For the infection experiments, supernatants were concentrated by ultracentrifugation as described earlier (10). Infections were performed by adsorbing the virus on cells in a small volume for 1 h at 37°C with occasional shaking. Cells were then extensively washed and refed. Virus detection in supernatants of infected cells was performed by reverse transcriptase (RT) assay (41). RT activities of supernatants were measured as counts per minute per milliliter and normalized for 10⁶ cells after background subtraction. Viral titrations were performed either by scoring the syncytium number on C8166 cells 5 days after infection (17) or by evaluating the number of blue cells 2 days after the infection of HeLaCD4-LTR-gal cells (7).

Protein analyses. To determine the pattern of intracellular HIV structural proteins, Western blot assays of on infected cells were performed by the enhanced chemiluminescence method (Amersham Buckinghamshire, United Kingdom) as previously described (15) by using a strongly reactive HIV-positive human serum. Either specific polyclonal rabbit antisera or monoclonal antibodies (MAbs) (all obtained by the AIDS Research and Reference Program) were used in order to detect HIV regulatory proteins in infected or transfected cells. Detection of virionic proteins in supernatants of infected cells was performed as described earlier (6) by ultracentrifuging clarified (22-µm-pore-size filtration) supernatants at 100,000 × g, for 45 min at 4°C. Viral pellets were then loaded onto a 20% sucrose cushion, ultracentrifuged at 95,000 × g for 150 min at 4°C, and finally dissolved in a Western blot loading buffer (45).

Immunofluorescence analyses. CD4 and CD8 membrane markers were detected by direct immunofluorescence analyses as previously described (4) by using, respectively, phycoerythrin (PE)-conjugated Leu3a and Leu2a MAbs (Becton Dickinson, Mountain View, Calif.).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Production and titration of NL4-3/chi viral particles. We previously demonstrated that CD4⁺ cells stably transfected with a vector expressing the F12-HIV nef gene maintain levels of CD4 receptors similar to those in control cells and became resistant to the HIV infection (15). To explore the mechanism of action underlying the F12-HIV nef-induced inhibition of HIV release, we constructed the pNL4-3/chi chimeric molecular clone by replacing the nef gene of the pNL4-3 molecular clone with the F12-HIV nef allele. In this manner, large amounts of F12-HIV Nef protein could be coexpressed with the remainder of the HIV protein pattern, excluding wt Nef. In most of our experiments, both wt and Nef-defective (nef) NL4-3 HIV infectious molecular clones or viral strains were used as controls.

View larger version (9K): [in this window] [in a new window]   FIG. 1.   Amounts (A) and infectivity (B) of HIV particles released by 293 cells transfected with pNL4-3, nef, or pNL4-3/chi HIV molecular clones (3 µg in semiconfluent 6-cm dishes). In panel A, RT activities in 1 ml of supernatant from each transfected cell culture are reported as the mean values ± the standard deviation (SD) from eight independent experiments. In panel B, fold reductions of infectivity (as measured by scoring either the formation of syncytia in C8166 cells infected with serial dilutions of supernatants or the relative number of blue cells after the infection of HeLaCD4-LTR-gal cells) with respect to volumes of supernatants normalized for RT activity from pNL4-3 transfected 293 cells are reported as the average values ± the SD from four separate experiments. Viral titers of supernatants from pNL4-3 transfected cells ranged from 1 × 106 to 3.5 × 106 50% tissue culture infective doses/ml. Columns: , wt; , nef; , NL4-3/chi.

No viral particles could be detected in the supernatants of CD4⁺ cells infected with the NL4-3/chi HIV strain. In order to better delineate the virological features of NL4-3/chi virions, infection experiments were performed by challenging either CEMss or PHA-stimulated CD8-depleted PBLs with a large dose (i.e., MOI of 0.5, as determined by titration on HeLaCD4-LTR-gal cells) of NL4-3, nef, or NL4-3/chi viral particles. At 24 h after the challenge, infected cells were extensively washed and reseeded, and the supernatants were tested for RT activity at different days postinfection. The data reported in Fig. 2 clearly demonstrate that strain NL4-3/chi is unable to replicate in either CEMss cells or PHA-stimulated CD8-depleted human PBLs despite the very high MOI used. Infection experiments performed by utilizing a lower viral input produced qualitatively superimposable results (not shown).

View larger version (10K): [in this window] [in a new window]   FIG. 2.   RT activities at different days after infection with nef or NL4-3/chi HIV strains of either CEMss cells (A) or CD8-depleted PHA-stimulated human PBLs (B) (MOI, 0.5). Data from one representative of five (CEMss cells) or two (CD8-depleted PBLs) independent experiments are reported. Symbols: , wt; , nef; , NL4-3/chi.

NL4-3/chi infection leads to a viral protein pattern highly reminiscent of that from nef NL4-3-infected cells. As demonstrated by infecting HeLaCD4-LTR-gal cells, both the viral entry efficiency and the ability to transactivate the -galactosidase gene appear to be indistinguishable between the nef and NL4-3/chi viral strains. In order to assess whether the block of the viral release observed in NL4-3/chi-infected cells originated from a defect in structural viral protein synthesis, CEMss cells were infected (MOI, 0.5) and Western blot analyses were carried out 48 h after challenge. As shown in Fig. 3A, the intracellular HIV protein pattern appears to be essentially indistinguishable between nef- and NL4-3/chi-infected cells. As a control, the amounts of Nef protein on the same cell lysates were also determined (Fig. 3B). The apparent reduced levels of Nef production in NL4-3/chi with respect to wt HIV-infected cells is the likely consequence of an impaired ability of the anti-Nef polyclonal antibody used to bind the F12-HIV Nef protein. This was also observed by testing a large panel of poly- or monoclonal anti-Nef antibodies under different experimental conditions (not shown).

View larger version (49K): [in this window] [in a new window]   FIG. 3.   Western blot analyses performed with 20 µg of proteins from CEMss cells lysed 2 days after infection with either wt (lane A), nef (lane B), or NL4-3/chi (lane D) HIV strains (MOI, 0.5). Cell lysates from uninfected CEMss cells (lane C) were included as a negative control. Protein revelations were performed by using a strong HIV-reactive serum from an AIDS patient (A) or a polyclonal anti-Nef rabbit serum (B). HIV proteins are indicated on the right side of each panel. Molecular size markers (in kilodaltons) are given on the left side.

The block of viral release from cells infected by NL4-3/chi strain is not relieved by the coexpression of wt nef and is not mediated by soluble factors. To test whether the presence of wt Nef protein could in some way relieve the F12-HIV Nef-induced inhibition of viral assembly and release, 293/CD4 cells were transfected with the pcDNA3/wt nef-expressing vector 24 h after infection (MOI, 1) with the NL4-3/chi or nef NL4-3 strain. After an additional 48 h, the RT activity was measured in cell supernatants. As shown in Fig. 4, whereas the expression of wt nef in nef NL4-3-infected cells favors, as expected, HIV production, no increase in RT activity could be conversely seen in supernatants from NL4-3/chi-infected cells. From these data we deduced that the possible positive action of wt Nef protein expression on HIV replication is not sufficient to overcome the block of viral replication imposed by the expression of F12-HIV Nef.

View larger version (17K): [in this window] [in a new window]   FIG. 4.   RT activities in supernatants of 293/CD4 cells infected with either nef or NL4-3/chi HIV and, after 24 h, transfected with different amounts of a pcDNA3-wt nef expressing vector. Viral input (MOI, 1) was adsorbed in 200 µl for 1 h at 37°C on 2 × 105 cells seeded in a 12-well plate. Cells were then extensively washed and refed. After 24 h, transfections with either 1 () or 2 () µg of pcDNA3-wt nef vector were performed. After an additional 48 h, supernatants were collected and RT activities were measured. Supernatants from mock-transfected infected cultures () were included as a control. Data from one representative of two independent experiments are shown.

Membrane association is essential for the F12-HIV Nef inhibitory effect. Attempting to distinguish possible molecular targets of the F12-HIV Nef action, we first decided to monitor whether the F12-HIV Nef membrane association is important for the above-described viral retention. The N-terminal myristoylation allows Nef protein to localize in the inner side of the cytoplasmic membranes and, in particular, of the cell membrane (55). It has been demonstrated that Nef myristoylation is necessary for the interaction with CD4 at the cell membrane and, consequently, for the CD4 downregulation (27). F12-HIV Nef does not show amino acid substitutions in the Nef myristoylation signal (18) (i.e., GGXXS at the N-terminal end), and thus it presumably retains its capacity to be myristoylated and localized at the inner side of the membranes. We questioned whether the F12-HIV Nef membrane association was necessary to induce the block of the NL4-3/chi viral cycle. CEMss cells were then infected with either the parental or myr forms of NL4-3/chi and, as controls, with wt or myr NL4-3 HIV strains. In Fig. 5, the reversion of the NL4-3/chi to an infectious strain as the consequence of the lack of the Nef N-terminal myristoylation is clearly demonstrated. This result indicates that the myristoylation and the consequent membrane association are essential for the HIV-inhibitory effect of F12-HIV Nef protein.

View larger version (12K): [in this window] [in a new window]   FIG. 5.   RT activities of supernatants recovered on different days after infection of CEMss cells with myr NL4-3 () or myr NL4-3/chi () HIV (MOI, 0.5). As controls, infections with strains NL4-3 () and NL4-3/chi () were replicated. Data from one representative of two independent experiments are reported.

F12-HIV Nef may interact with CD4. (i) Inhibition of the CD4 downregulation induced by wt Nef. The data presented above strongly suggest that the block of the HIV replication cycle takes place during viral assembly and/or release. Furthermore, we showed that the membrane association of F12-HIV Nef is required for the HIV inhibitory effect. The cell membrane is the cellular compartment where a major Nef molecular target (i.e., CD4 receptor) is localized in its biologically active conformation. Moreover, it is well documented that wt Nef actively interacts with the CD4 intracytoplasmic domain (22, 23, 26, 40). We tried to establish whether F12-HIV Nef retains the ability to interact with CD4. This is conceivable, since amino acidic sequences involved in binding of the CD4 intracytoplasmic domain are well conserved in F12-HIV Nef (9). We also sought to determine, if this is true, whether F12-HIV Nef-CD4 interaction is involved in the block of HIV release.

View larger version (25K): [in this window] [in a new window]   FIG. 6.   FACS analyses on 293/CD4 cells transfected (A) or cotransfected (B) with pcDNA3 vectors expressing either the wt or the F12-HIV nef alleles. Transfections were performed on 2 × 105 cells seeded in a 12-well plate. (A) Levels of CD4 receptors were measured by using the Leu3A anti-CD4 PE-conjugated MAb on mock-transfected 293/CD4 cells (b) and on cells transfected with 1 µg of either wt (c)- or F12-HIV nef (d)-expressing vectors. (B) CD4 levels were measured on 293/CD4 cells transfected with 1.5 µg of empty pcDNA3 vector (b), 0.5 µg of pcDNA3-wt nef plus 1 µg of empty pcDNA3 vector (c), 0.5 µg of pcDNA3-wt nef plus 0.5 µg of pcDNA3-F12-HIV nef vectors (d), or 0.5 µg of pcDNA3-wt nef plus 1 µg of pcDNA3-F12-HIV nef vectors (e). Transfected cells labeled with PE-conjugated nonspecific mouse immunoglobulin G isotype were used as a control (lane a in both panels).

(ii) Influence on CD4 distribution at the cell membrane. It was previously reported that the expression of wt Nef-GFP induces a CD4 redistribution from a uniform to a distinct punctate pattern at the cell margin (22). The ability of wt Nef protein to induce a CD4 clusterization was interpreted as a consequence of the interaction with the CD4 intracytoplasmic tail and considered as an intermediate step in CD4 downregulation (22). In order to highlight the ability of F12-HIV Nef to interact with CD4 more directly, we labeled 293/CD4 cells with an anti-CD4 MAb after transfection with vectors expressing F12-HIV Nef-GFP or, as a control, wt Nef/GFP or GFP alone. The analyses with a confocal microscope demonstrated that cells actively expressing F12-HIV Nef-GFP fusion protein show a clearly distinguishable punctate pattern of CD4 at the cell membrane (Fig. 7C). However, in contrast to cells expressing GFP alone (where a uniformly labeled cell margin is detectable [Fig. 7A]) and in cells expressing wt Nef-GFP protein (which show a massive lack of CD4-specific staining [Fig. 7B]), these zones of accumulation overlay a background of CD4 positivity. When the observations were carried out on 293 cells stably expressing CD4 molecules truncated in the intracytoplasmic domain (293/CD44x), a uniformly distributed CD4 pattern could be observed in all tested conditions (Fig. 7D to F). Of note, remarkable differences could be observed in the intracytoplasmic distributions between the two Nef proteins. In fact, whereas wt Nef localizes in discrete punctate aggregations, F12-HIV Nef (in conditions of similar protein expression levels [not shown]) appears to accumulate in larger amounts and in a more compartmentalized intracellular region (Fig. 7C and E).

View larger version (18K): [in this window] [in a new window]   FIG. 7.   Confocal microscopy analyses of 293/CD4 (A to C) and 293/CD44x (D and E) cells labeled with a PE-conjugated anti-CD4 MAb (see legend to Fig. 6) 48 h after transfection with vectors expressing GFP alone (A and D), wt Nef-GFP (B and E), and F12-HIV Nef-GFP fusion proteins (C and F). In panel C, the arrows indicate some of the CD4 accumulation zones. The panel C and F images were produced by using a laser power lower than that applied to obtain the remainder of the panels. This allowed a better discrimination of both F12-HIV Nef intracellular localization and membrane CD4 accumulation zones.

The F12-HIV Nef-induced block of HIV release correlates with the presence of both the CD4 intracytoplasmic tail and the HIV Env products. We questioned whether the CD4-F12-HIV Nef interaction could be involved in the block of viral production. In addition, attempting to distinguish a viral target(s) of the F12-HIV Nef inhibitory effect, we also tried to determine whether viral Env products are involved in the F12-HIV Nef-induced viral retention. The expression of the Env products appears to be dispensable for viral release (19, 48, 49) but may play a role in the polarization of HIV budding (30). 293 cells stably expressing either CD4, CD44x, CD884 (a CD8-based chimeric receptor bearing the CD4 intracytoplasmic tail) (3), or CD88x (a CD8 receptor truncated in its intracytoplasmic domain) (5) were cotransfected with env wt or env NL4-3/chi HIV molecular clones together with vectors expressing either HIV or murine leukemia virus amphotropic Env proteins in a molar ratio of 1:10. All cell populations were scored in advance by fluorescence-activated cell sorter (FACS) analyses as >95% positive for the expression of either CD4 or CD8 membrane expression (not shown). As reported in Table 1, a strong impairment of HIV spread was detected in cells expressing the CD4 intracytoplasmic tail (whatever the extracellular domain) and transfected with the NL4-3/chi molecular clone expressing HIV Env in trans. These results indicate that (i) both the CD4 intracytoplasmic tail and the HIV Env products are required for the F12-HIV Nef-induced viral retention and (ii) the CD4 extracellular domain does not seem to be involved in the F12-HIV Nef-induced inhibition of HIV release.

View this table: [in this window] [in a new window]   TABLE 1.   RT activities on supernatants of 293 cells stably expressing CD4, CD44x, CD884, or CD88x receptors 48 h after the cotransfection of env forms of either wt or pNL4-3/chi molecular clones with either T-tropic HIV or MLV Env-expressing vectorsa

View larger version (17K): [in this window] [in a new window]   FIG. 8.   RT activities in supernatants of either 293/CD4 or 293/CD44x cells 5 days after infection with two different MOIs (0.5 and 1; panels A and B, respectively) of wt (), nef (), or NL4-3/chi () HIV strains. Data from one representative of three independent experiments are shown.

The HIV Env gp41 intracytoplasmic tail is required for F12-HIV Nef-induced viral inhibition. In order to define the region(s) of HIV Env products required for F12-HIV Nef-induced viral retention, we took advantage of the evidence that through transient cotransfection of an F12-HIV nef expressing vector and a pNL4-3 molecular clone, the F12-HIV Nef inhibitory effect has been fairly reproduced (15). 293/CD4 cells were cotransfected with HIV molecular clones defective for either the Env gp120 CD4 binding domain (pNL-A1 [CD4]) (54) or the Env gp41 intracytoplasmic tail (HIV Tr712Env) (53), together with the pcDNA3-F12-HIV nef vector at different molar ratios (from 1:0 to 1:10). As controls, similar cotransfections were performed by using either the full infectious molecular clone pNL4-3 or an HIV molecular clone defective for the expression of the whole env gene (env wt HIV). At 48 h after the cotransfections, supernatants were collected, and the RT levels therein were determined. The results (Fig. 9) clearly show that the F12-HIV Nef inhibitory effect does not take place in the absence of either the whole HIV Env protein or of the Env gp41 intracytoplasmic tail. Conversely, similar viral release inhibitions have been detected in supernatants of cells transfected with pNL-A1 (CD4) or wt HIV molecular clones. These results allow us to propose the Env gp41 intracytoplasmic domain as a major viral target of the F12-HIV Nef inhibitory effect.

View larger version (17K): [in this window] [in a new window]   FIG. 9.   RT activities in supernatants of 293/CD4 cells 48 h after the cotransfection at different molar ratios (from 1:0 to 1:10) of either pNL4-3 (), env wt (), pNLA1 (CD4) (), or Tr712 Env () HIV molecular clones (0.1 µg in a semiconfluent 24-well plate) with pcDNA3-F12-HIV nef vector. In the 1:0 points, the RT activities in the supernatants of cells transfected with each HIV molecular clone only are reported. Importantly, cotransfections with empty pcDNA3 vector did not significantly alter the RT levels in the supernatants of cells transfected with different HIV molecular clones. For example, values from supernatants of cells cotransfected with pNL4-3 and pcDNA3 plasmids () were included. Values from one representative of two independent experiments are reported.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The strong impairment in the replication of wt HIV induced by F12-HIV nef expression (15) is a unique feature among HIV or SIV nef alleles characterized so far. By infecting cell clones stably expressing F12-HIV nef, we previously demonstrated that the inhibitory effect acts at a very late step of the viral replication cycle, presumably during the viral assembly and/or release process (15). We also observed that cells expressing F12-HIV Nef resist infection with nef HIV (15; R. Bona, unpublished observations). We thus tentatively excluded the possibility that the F12-HIV Nef inhibitory effect could be the consequence of a negative trans-dominant effect, as has already been described for Tat (20) and Rev (33) mutants. In order to gain an experimental model in which an optimal coexpression of F12-HIV Nef may be achieved together with the full viral protein complement of a replication-competent HIV, but in the absence of wt Nef protein (that in most in vitro models is dispensable for HIV replication), the NL4-3/chi HIV clone was constructed. In this manner, we were able to make a closer discrimination of the effects of F12-HIV Nef on the HIV replicative cycle.

We demonstrate here that the expression of the F12-HIV nef gene is itself sufficient to transform the highly infectious NL4-3 HIV strain in a virus exclusively able to infect target cells abortively. Nef, as well as Vpr and Vpu, has been shown to be dispensable for in vitro HIV replication (11), except for the infection of resting PBLs (37, 50). Thus, the observation reported here that a slightly mutated nef gene may have such a dramatic effect on the HIV replication cycle appears to be both intriguing and original. More generally, our findings may enforce the hypothesis of a correlation between the high frequency of recovery of nef-mutated HIV genomes in long-term nonprogressor AIDS patients and the delay in disease development (16, 29, 44).

The observation that cells infected with the NL4-3/chi strain express an apparently unmodified pattern of viral proteins strongly suggests that F12-HIV Nef-induced HIV inhibition occurs at a very late replication step. This result appears to be consistent with that already observed by infecting cells stably expressing the F12-HIV nef gene (15). Furthermore, we found that the functional defect(s) of the chimeric virus could not be relieved by in-trans expression of wt nef. This finding indicates that the inhibitory effect(s) of F12-HIV Nef protein could overcome the wt Nef function(s) favoring HIV replication.

The amino acid sequences involved in Nef-induced CD4 downregulation have been mapped in both the CD4 intracytoplasmic tail and the Nef central core (23). The presence of a dileucine motif was demonstrated in both CD4 (3) and Nef (14) to be necessary for the CD4 downregulation. Through the utilization of either CD8 or CD4-Nef chimeric molecules, it has also been shown that Nef possesses an intrinsic ability to induce receptor internalization by means of a direct or indirect connection with the endocytic machinery (34). Although F12-HIV Nef fails to induce CD4 downregulation, no amino acidic substitutions have been detected in the sequences involved in the Nef-CD4 interaction (9). The possibility that F12-HIV Nef maintains the ability to interact with CD4 may be inferred from our competition assay that demonstrated the capacity of F12-HIV Nef protein to displace wt Nef from the CD4 molecule. However, we cannot formally exclude the possibility that such a result is the consequence, at least in part, of molecular events other than competition for CD4, e.g., the formation of wt-F12-HIV Nef dimers or polymers lacking the ability to bind and/or internalize CD4 receptors. An additional support for the hypothesis that F12-HIV Nef retains the ability to interact with CD4 comes from our fluorescence microscopy analyses. In fact, the coexpression of full-length CD4 and F12-HIV Nef leads to a CD4 redistribution at the cell membrane similar to that already observed for wt Nef (22). On the contrary, this phenomenon does not occur in cells expressing CD4 truncated in its intracytoplasmic domain. The lack of CD4 downregulation is likely at the basis of the much stronger CD4-specific membrane labeling in F12-HIV nef-transfected 293/CD4 cells with respect to those transfected with wt nef.

Both transfection and infection experiments done with 293/CD44x cells strongly support the hypothesis that a direct or indirect interaction between F12-HIV Nef and the intracytoplasmic tail of CD4 receptors is at the basis of the F12-HIV Nef-induced block of the viral release. This possibility is very intriguing, considering that neither nef expression (11) nor the CD4 intracytoplasmic domain (34) is necessary for completion of the HIV life cycle. In addition, from the results obtained by Western blot analyses, we may exclude defects in synthesis or processing of HIV proteins. Also noteworthy, from the transfection experiments carried out with the HIV molecular clones with either deletions or mutations in the env gene, is the suggestion that a critical role of the HIV Env gp41 intracytoplasmic domain in the F12-HIV Nef inhibitory effect may be inferred.

It has been reported that HIV Env constitutively undergoes internalization (42). This process appears to be mediated by the interaction with clathrin adapter molecules (36). The association of Env gp41 protein with Gag-derived viral core components at the cell membrane may prevent Env internalization by interfering with the binding of clathrin adapter molecules, leading to viral assembly and/or release (36). On the other hand, it was well demonstrated that Nef interacts with components of the clathrin coat at the plasma membrane, including clathrin and  subunits of adapter protein complex-2 (a component of the endocytotic machinery) (22, 27). We may hypothesize that CD4-F12-HIV Nef binding induces the recruitment of molecules from the endocytic machinery, as already shown for wt Nef. In this regard, it should be noted that F12-HIV Nef retains the dileucine motif (9) needed to address the cellular sorting machinery (14). However, the lack of CD4 downregulation may alter the physiological recycling of clathrin-adaptin molecules which thus may bridge the CD4-F12-HIV Nef complex with the Env gp41 intracytoplasmic tail. This could lead to an abnormal accumulation and/or an altered spatial distribution of clathrin molecules that may hinder their displacement from Env gp41 by HIV Gag proteins, thus inhibiting correct viral assembly and/or release. We consistently observed that the lack of either CD4 or Env gp41 intracytoplasmic domains renders ineffective the antiviral action of F12-HIV Nef. Our finding that removal of the F12-HIV Nef myristoyl group correlates with the reversion of NL4-3/chi to an infectious strain fulfills the hypothesis that crucial events leading to the viral block occur in the neighborhood of intracellular membranes or, more likely, of the inner side of the cell membrane. At the moment, a detailed description of both the molecular basis of the F12-HIV Nef, CD4, and Env interactions and their effects on correct HIV morphogenesis represents an attractive experimental challenge.

In summary, our data delineate a model of inhibition of HIV replication that has, to the best of our knowledge, no counterpart in the HIV field. HIV genomes that in vivo express nef alleles with a similar phenotype may have a role in the containment of HIV spread in that they show a reduced replication capacity and also could behave as interfering HIV genomes, as demonstrated for the whole F12-HIV genome (18). The findings we obtained while attempting to elucidate the mechanism of action of F12-HIV Nef-induced HIV inhibition are also encouraging from the perspective of anti-HIV gene therapy strategy. It is worth noting that the F12-HIV Nef inhibitory effect acts through the major component needed for viral entry, thus guaranteeing the block of viral spread from the cells attacked by HIV infection.