Stem-loop SL4 of the HIV-1 psi RNA packaging signal exhibits weak affinity for the nucleocapsid protein. structural studies and implications for genome recognition.

Encapsidation of the genome of the human immunodeficiency virus type-1 (HIV-1) during retrovirus assembly is mediated by interactions between the nucleocapsid (NC) domains of assembling Gag polyproteins and a approximately 110 nucleotide segment of the genome known as the Psi-site. The HIV-1 Psi-site contains four stem-loops (SL1 through SL4), all of which are important for genome packaging. Recent isothermal titration calorimetry (ITC) studies have demonstrated that SL2 and SL3 are capable of binding NC with high affinity (K(d) approximately 140 nM), consistent with proposals for protein-interactive functions during packaging. To determine if SL4 may have a similar function, NC-interactive studies were conducted by NMR and gel-shift methods. In contrast to previous reports, we find that SL4 binds weakly to NC (K(d)=(+/-14 microM), suggesting an alternative function. NMR studies indicate that the GAGA tetraloop of SL4 adopts a classical GNRA-type fold (R=purine, N=G, C, A or U), a motif that stabilizes RNA tertiary structures in other systems. In combination with previously reported gel mobility studies of Psi-site deletion mutants, these findings suggest that SL4 functions in genome recognition not by binding to Gag, but by stabilizing the structure of the Psi-site. Differences in the affinities of NC for SL2, SL3 and SL4 stem-loops can now be rationalized in terms of the different structural properties of stem loops that contain GGNG (SL2 and SL3) and GNRA (SL4) sequences.     

High affinity nucleocapsid protein binding to the muPsi RNA packaging signal of Rous sarcoma virus

The genomes of all retroviruses contain sequences near their 5' ends that interact with the nucleocapsid domains (NC) of assembling Gag proteins and direct their packaging into virus particles. Retroviral packaging signals often occur in non-contiguous segments spanning several hundred nucleotides of the RNA genome, confounding structural and mechanistic studies of genome packaging. Recently, a relatively short, 82 nucleotide region of the Rous sarcoma virus (RSV) genome, called muPsi, was shown to be sufficient to direct efficient packaging of heterologous RNAs into RSV-like particles. We have developed a method for the preparation and purification of large quantities of recombinant RSV NC protein, and have studied its interactions with native and mutant forms of the muPsi encapsidation element. NC does not bind with significant affinity to truncated forms of muPsi, consistent with earlier packaging and mutagenesis studies. Surprisingly, NC binds to the native muPsi RNA with affinity that is approximately 100 times greater than that observed for other previously characterized retroviral NC-RNA complexes (extrapolated dissociation constant K(d)=1.9 nM). Tight binding with 1:1 NC-muPsi stoichiometry is dependent on a conserved UGCG tetraloop in one of three predicted stem loops, and an AUG initiation codon controvertibly implicated in genome packaging and translational control. Loop nucleotides of other stem loops do not contribute to NC binding. Our findings indicate that the structural determinants of RSV genome recognition and NC-RNA binding differ considerably from those observed for other retroviruses.     

NMR structure of the HIV-1 nucleocapsid protein bound to stem-loop SL2 of the psi-RNA packaging signal. Implications for genome recognition

The RNA genome of the human immunodeficiency virus type-1 (HIV-1) contains a approximately 120 nucleotide Psi-packaging signal that is recognized by the nucleocapsid (NC) domain of the Gag polyprotein during virus assembly. The Psi-site contains four stem-loops (SL1-SL4) that possess overlapping and possibly redundant functions. The present studies demonstrate that the 19 residue SL2 stem-loop binds NC with affinity (K(d)=110(+/-50) nM) similar to that observed for NC binding to SL3 (K(d)=170(+/-65) nM) and tighter than expected on the basis of earlier work, suggesting that NC-SL2 interactions probably play a direct role in the specific recognition and packaging of the full-length, unspliced genome. The structure of the NC-SL2 complex was determined by heteronuclear NMR methods using (15)N,(13)C-isotopically labeled NC protein and SL2 RNA. The N and C-terminal "zinc knuckles" (Cys-X(2)-Cys-X(4)-His-X(4)-Cys; X=variable amino acid) of HIV-1 NC bind to exposed guanosine bases G9 and G11, respectively, of the G8-G9-U10-G11 tetraloop, and residues Lys3-Lys11 of the N-terminal tail forms a 3(10) helix that packs against the proximal zinc knuckle and interacts with the RNA stem. These structural features are similar to those observed previously in the NMR structure of NC bound to SL3. Other features of the complex are substantially different. In particular, the N-terminal zinc knuckle interacts with an A-U-A base triple platform in the minor groove of the SL2 RNA stem, but binds to the major groove of SL3. In addition, the relative orientations of the N and C-terminal zinc knuckles differ in the NC-SL2 and NC-SL3 complexes, and the side-chain of Phe6 makes minor groove hydrophobic contacts with G11 in the NC-SL2 complex but does not interact with RNA in the NC-SL3 complex. Finally, the N-terminal helix of NC interacts with the phosphodiester backbone of the SL2 RNA stem mainly via electrostatic interactions, but does not bind in the major groove or make specific H-bonding contacts as observed in the NC-SL3 structure. These findings demonstrate that NC binds in an adaptive manner to SL2 and SL3 via different subsets of inter and intra-molecular interactions, and support a genome recognition/packaging mechanism that involves interactions of two or more NC domains of assembling HIV-1 Gag molecules with multiple Psi-site stem-loop packaging elements during the early stages of retrovirus assembly.     

Modulation of homo- and heterodimerization of Harvey sarcoma virus RNA by GACG tetraloops and point mutations in palindromic sequences

Retroviruses harbour a diploid genome of two plus-strand RNAs linked non-covalently at the dimer linkage structure. Co-packaging of two parental RNAs is a prerequisite for recombination in retroviruses, but formation of heterodimers has not been demonstrated directly in vivo. Here, we explore elements in Harvey sarcoma virus (HaSV) RNA involved in homodimerization and heterodimerization with RNA of Moloney (Mo) and Akv murine leukemia viruses (MLV).By an in vitro assay, we found that HaSV dimerization specificity could be modulated by mutations in a decanucleotide palindrome (Pal) probably folded into a kissing-loop. Autocomplementary and non-autocomplementary sequences introduced into the putative loop directed the specificity towards formation of homodimers and heterodimers, respectively. Two stem-loop (SL) structures, both exposing a GACG tetraloop, enhanced the formation of stable HaSV dimers.A similar decanucleotide palindrome has been implicated in homodimerization of MLVs. Heterodimers between HaSV RNA and Mo- or Akv MLV were unstable, but could be stabilized by introduction of two point mutations in the putative HaSV kissing-loop, creating exact complementarity with Mo/Akv MLV palindromes. Moreover, such changes increased the HaSV RNA affinity for the two MLV RNAs. Similar to HaSV RNA homodimers, formation of heterodimers with Mo- or Akv MLV RNAs was induced by the presence of GACG loops.On the basis of these results, we propose that palindromic sequences act as variable determinants of specificity and GACG tetraloops as conserved determinants in the formation of homodimers and heterodimers of gamma-retrovirus retroviral RNAs in vivo. The complementarity of loop sequences in the packaging signal upstream of the GACG tetraloops might therefore determine homo- and heterodimerization specificity and recombination activity of these viruses.     

The Nucleocapsid Domain Is Responsible for the Ability of Spleen Necrosis Virus (SNV) Gag Polyprotein To Package both SNV and Murine Leukemia Virus RNA

Murine leukemia virus (MLV)-based vector RNA can be packaged and propagated by the proteins of spleen necrosis virus (SNV). We recently demonstrated that MLV proteins cannot support the replication of an SNV-based vector; RNA analysis revealed that MLV proteins cannot efficiently package SNV-based vector RNA. The domain in Gag responsible for the specificity of RNA packaging was identified using chimeric gag-pol expression constructs. A competitive packaging system was established by generating a cell line that expresses one viral vector RNA containing the MLV packaging signal (Psi) and another viral vector RNA containing the SNV packaging signal (E). The chimeric gag-pol expression constructs were introduced into the cells, and vector titers as well as the efficiency of RNA packaging were examined. Our data confirm that Gag is solely responsible for the selection of viral RNAs. Furthermore, the nucleocapsid (NC) domain in the SNV Gag is responsible for its ability to interact with both SNV E and MLV Psi. Replacement of the SNV NC with the MLV NC generated a chimeric Gag that could not package SNV RNA but retained its ability to package MLV RNA. A construct expressing SNV gag-MLV pol supported the replication of both MLV and SNV vectors, indicating that the gag and pol gene products from two different viruses can functionally cooperate to perform one cycle of retroviral replication. Viral titer data indicated that SNV cis-acting elements are not ideal substrates for MLV pol gene products since infectious viruses were generated at a lower efficiency. These results indicate that the nonreciprocal recognition between SNV and MLV extends beyond the Gag-RNA interaction and also includes interactions between Pol and other cis-acting elements.     

An RNA Structural Switch Regulates Diploid Genome Packaging by Moloney Murine Leukemia Virus

Retroviruses selectively package two copies of their RNA genomes via mechanisms that have yet to be fully deciphered. Recent studies with small fragments of the Moloney murine leukemia virus (MoMuLV) genome suggested that selection may be mediated by an RNA switch mechanism, in which conserved UCUG elements that are sequestered by base-pairing in the monomeric RNA become exposed upon dimerization to allow binding to the cognate nucleocapsid (NC) domains of the viral Gag proteins. Here we show that a large fragment of the MoMuLV 5′ untranslated region that contains all residues necessary for efficient RNA packaging (Ψ^WT; residues 147-623) also exhibits a dimerization-dependent affinity for NC, with the native dimer ([Ψ^WT]₂) binding 12 ± 2 NC molecules with high affinity (K_d = 17 ± 7 nM) and with the monomer, stabilized by substitution of dimer-promoting loop residues with hairpin-stabilizing sequences (Ψ^M), binding 1-2 NC molecules. Identical dimer-inhibiting mutations in MoMuLV-based vectors significantly inhibit genome packaging in vivo (∼ 100-fold decrease), whereas a large deletion of nearly 200 nucleotides just upstream of the gag start codon has minimal effects. Our findings support the proposed RNA switch mechanism and further suggest that virus assembly may be initiated by a complex comprising as few as 12 Gag molecules bound to a dimeric packaging signal.     

Thermodynamic parameters for loop formation in RNA and DNA hairpin tetraloops.

We determined the melting temperatures (Tm) and thermodynamic parameters of 15 RNA and 19 DNA hairpins at 1 M NaCl, 0.01 M sodium phosphate, 0.1 mM EDTA, at pH 7. All these hairpins have loops of four bases, the most common loop size in 16S and 23S ribosomal RNAs. The RNA hairpins varied in loop sequence, loop-closing base pair (A.U, C.G, or G.C), base sequence of the stem, and stem size (four or five base pairs). The DNA hairpins varied in loop sequence, loop-closing base pair (C.G, or G.C), and base sequence of the four base-pair stem. Thermodynamic properties of a hairpin may be represented by nearest-neighbor interactions of the stem plus contributions from the loop. Thus, we obtained thermodynamic parameters for the formation of RNA and DNA tetraloops. For the tetraloops we studied, a free energy of loop formation (at 37 degrees C) of about +3 kcal/mol is most common for either RNA or DNA. There are extra stable loops with delta G degrees 37 near +1 kcal/mol, but the sequences are not necessarily the same for RNA and DNA. The closing base pair is also important; changing from C.G to G.C lowered the stability of several tetraloops in both RNA and DNA. These values will be useful in predicting RNA and DNA secondary structures.     

The packaging signal of MLV is an integrated module that mediates intracellular transport of genomic RNAs

Packaging of MLV genomes requires four cis-acting stem-loops. Stem-loops A and B are self-complementary and bind Gag in their dimeric form, while the C and D elements mediate loop-loop interactions that facilitate RNA dimerization. Packaging also requires nuclear export of viral genomes, and their cytoplasmic transport toward the plasma membrane. For MLV, this is mediated by Gag and Env, and occurs on endosomal vesicles. Here, we report that MLV Psi acts at several steps during the transport of genomic RNAs. First, deletion of stem-loop B or C leads to the accumulation of genomic RNAs in the nucleus, suggesting that these elements are involved in export. Second, in chronically infected cells, mutation of the C and D loops impairs endosomal transport. This suggests that RNA dimerization is essential for vesicular transport, consistent with its proposed requirement for Gag binding. Surprisingly, deletion of stem-loop A blocks vesicular transport, whereas removal of stem-loop B has no effects. This suggests that stem-loop A has unique functions in packaging, not predicted from previous in vitro analyses. Finally, in packaging cells that do not express any Psi-containing RNA, endosomal RNA transport becomes sequence-independent. This non-specific activity of Gag likely promotes packaging of cellular mRNAs.     

Solution properties of murine leukemia virus gag protein: differences from HIV-1 gag

Immature retrovirus particles are assembled from the multidomain Gag protein. In these particles, the Gag proteins are arranged radially as elongated rods. We have previously characterized the properties of HIV-1 Gag in solution. In the absence of nucleic acid, HIV-1 Gag displays moderately weak interprotein interactions, existing in monomer-dimer equilibrium. Neutron scattering and hydrodynamic studies suggest that the protein is compact, and biochemical studies indicate that the two ends can approach close in three-dimensional space, implying the need for a significant conformational change during assembly. We now describe the properties of the Gag protein of Moloney murine leukemia virus (MLV), a gammaretrovirus. We found that this protein is very different from HIV-1 Gag: it has much weaker protein-protein interaction and is predominantly monomeric in solution. This has allowed us to study the protein by small-angle X-ray scattering and to build a low-resolution molecular envelope for the protein. We found that MLV Gag is extended in solution, with an axial ratio of ～7, comparable to its dimensions in immature particles. Mutational analysis suggests that runs of prolines in its matrix and p12 domains and the highly charged stretch at the C terminus of its capsid domain all contribute to this extended conformation. These differences between MLV Gag and HIV-1 Gag and their implications for retroviral assembly are discussed.     

Comparative structural effects of HIV-1 Gag and nucleocapsid proteins in binding to and unwinding of the viral RNA packaging signal

The major RNA binding region of the HIV-1 Gag polyprotein is the nucleocapsid (NC) domain, which is responsible for the specific capture of the genomic RNA genome during viral assembly. The Gag polyprotein has other RNA chaperone functions, which are mirrored by the isolated NC protein after physiological cleavage from Gag. Gag, however, is suggested to have superior nucleic acid chaperone activity. Here we investigate the interaction of Gag and NC with the core RNA structure of the HIV-1 packaging signal (Ψ), using 2-aminopurine substitution to create a series of modified RNAs based on the Ψ helix loop structure. The effects of 2-aminopurine substitution on the physical and structural properties of the viral Ψ were characterized. The fluorescence properties of the 2-aminopurine substitutions showed features consistent with the native GNAR tetraloop. Dissociation constants (K(d)) of the two viral proteins, measured by fluorescence polarization (FP), were similar, and both NC and Gag affected the 2-aminopurine fluorescence of bases close to the loop binding region in a similar fashion. However, the influence of Gag on the fluorescence of the 2-aminopurine nucleotides at the base of the helix implied a much more potent helix destabilizing action on the RNA stem loop (SL) versus that seen with NC. This was further supported when the viral Ψ SL was tagged with a 5' fluorophore and 3' quencher. In the absence of any viral protein, minimal fluorescence was detected; addition of NC yielded a slight increase in fluorescence, while addition of the Gag protein yielded a large change in fluorescence, further suggesting that, compared to NC, the Gag protein has a greater propensity to affect RNA structure and that Ψ helix unwinding may be an intrinsic step in RNA encapsidation.     

Binding of an aminoacridine derivative to a GAAA RNA tetraloop

RNA tetraloops are common secondary structural motifs in many RNAs, especially ribosomal RNAs. There are few studies of small molecule recognition of RNA tetraloops although tetraloops are known to interact with RNA receptors and proteins, and to form nucleation sites for RNA folding. In this paper, we investigate the binding of neomycin, kanamycin, 2,4-diaminoquinazoline, quinacrine, and an aminoacridine derivative (AD1) to a GAAA tetraloop using fluorescence spectroscopy. We have found that AD1 and quinacrine bind to the GAAA tetraloop with the highest affinity of the molecules examined. The equilibrium dissociation constant of the AD1-GAAA tetraloop complex was determined to be 1.6 microM. RNase I and lead acetate footprinting experiments suggested that AD1 binds to the junction between the loop and stem of the GAAA tetraloop.     

Dissecting the protein-RNA and RNA-RNA interactions in the nucleocapsid-mediated dimerization and isomerization of HIV-1 stemloop 1

The specific binding of HIV-1 nucleocapsid protein (NC) to the different forms assumed in vitro by the stemloop 1 (Lai variant) of the genome's packaging signal has been investigated using electrospray ionization-Fourier transform mass spectrometry (ESI-FTMS). The simultaneous observation of protein-RNA and RNA-RNA interactions in solution has provided direct information about the role of NC in the two-step model of RNA dimerization and isomerization. In particular, two distinct binding sites have been identified on the monomeric stemloop structure, corresponding to the apical loop and stem-bulge motifs. These sites share similar binding affinities that are intermediate between those of stemloop 3 (SL3) and the putative stemloop 4 (SL4) of the packaging signal. Binding to the apical loop, which contains the dimerization initiation site (DIS), competes directly with the annealing of self-complementary sequences to form a metastable kissing-loop (KL) dimer. In contrast, binding to the stem-bulge affects indirectly the monomer-dimer equilibrium by promoting the rearrangement of KL into the more stable extended duplex (ED) conformer. This process is mediated by the duplex-melting activity of NC, which destabilizes the intramolecular base-pairs surrounding the KL stem-bulges and enables their exchange to form the inter-strand pairs that define the ED structure. In this conformer, high-affinity binding takes place at stem-bulge sites that are identical to those present in the monomeric and KL forms. In this case, however, the NC-induced "breathing" does not result in dissociation of the double-stranded structure because of the large number of intermolecular base-pairs. The different binding modes manifested by conformer-specific mutants have shown that NC can also provide low affinity interactions with the bulged-out adenine bases flanking the DIS region of the ED conformer, thus supporting the hypothesis that these exposed nucleotides may constitute "base-grips" for protein contacts during the late stages of the viral lifecycle.     

Dimeric assembly of human Suv3 helicase promotes its RNA unwinding function in mitochondrial RNA degradosome for RNA decay

Human Suv3 is a unique homodimeric helicase that constitutes the major component of the mitochondrial degradosome to work cooperatively with exoribonuclease PNPase for efficient RNA decay. However, the molecular mechanism of how Suv3 is assembled into a homodimer to unwind RNA remains elusive. Here, we show that dimeric Suv3 preferentially binds to and unwinds DNA–DNA, DNA–RNA, and RNA–RNA duplexes with a long 3′ overhang (≥10 nucleotides). The C‐terminal tail (CTT)‐truncated Suv3 (Suv3ΔC) becomes a monomeric protein that binds to and unwinds duplex substrates with ~six to sevenfold lower activities relative to dimeric Suv3. Only dimeric Suv3, but not monomeric Suv3ΔC, binds RNA independently of ATP or ADP, and is capable of interacting with PNPase, indicating that dimeric Suv3 assembly ensures its continuous association with RNA and PNPase during ATP hydrolysis cycles for efficient RNA degradation. We further determined the crystal structure of the apo‐form of Suv3ΔC, and SAXS structures of dimeric Suv3 and PNPase–Suv3 complex, showing that dimeric Suv3 caps on the top of PNPase via interactions with S1 domains, and forms a dumbbell‐shaped degradosome complex with PNPase. Overall, this study reveals that Suv3 is assembled into a dimeric helicase by its CTT for efficient and persistent RNA binding and unwinding to facilitate interactions with PNPase, promote RNA degradation, and maintain mitochondrial genome integrity and homeostasis.     

mRNA molecules containing murine leukemia virus packaging signals are encapsidated as dimers

Prior work by others has shown that insertion of psi (i.e., leader) sequences from the Moloney murine leukemia virus (MLV) genome into the 3' untranslated region of a nonviral mRNA leads to the specific encapsidation of this RNA in MLV particles. We now report that these RNAs are, like genomic RNAs, encapsidated as dimers. These dimers have the same thermostability as MLV genomic RNA dimers; like them, these dimers are more stable if isolated from mature virions than from immature virions. We characterized encapsidated mRNAs containing deletions or truncations of MLV psi or with psi sequences from MLV-related acute transforming viruses. The results indicate that the dimeric linkage in genomic RNA can be completely attributed to the psi region of the genome. While this conclusion agrees with earlier electron microscopic studies on mature MLV dimers, it is the first evidence as to the site of the linkage in immature dimers for any retrovirus. Since the Psi(+) mRNA is not encapsidated as well as genomic RNA, it is only present in a minority of virions. The fact that it is nevertheless dimeric argues strongly that two of these molecules are packaged into particles together. We also found that the kissing loop is unnecessary for this coencapsidation or for the stability of mature dimers but makes a major contribution to the stability of immature dimers. Our results are consistent with the hypothesis that the packaging signal involves a dimeric structure in which the RNAs are joined by intermolecular interactions between GACG loops.     

Electrostatic interactions drive membrane association of the human immunodeficiency virus type 1 Gag MA domain

The assembly of most retroviruses occurs at the plasma membrane. Membrane association is directed by MA, the N-terminal domain of the Gag structural protein. For human immunodeficiency virus type 1 (HIV-1), this association is mediated in part by a myristate fatty acid modification. Conflicting evidence has been presented on the relative importance of myristoylation, of ionic interactions between protein and membrane, and of Gag multimerization in membrane association in vivo. We addressed these questions biochemically by determining the affinity of purified myristoylated HIV-1 MA for liposomes of defined composition, both for monomeric and for dimeric forms of the protein. Myristoylation increases the barely detectable intrinsic affinity of the apo-protein for liposomes by only 10-fold, and the resulting affinity is still weak, similar to that of the naturally nonmyristoylated MA of Rous sarcoma virus. Membrane binding of HIV-1 MA is absolutely dependent on the presence of negatively charged lipid and is abrogated at high ionic strength. Forced dimerization of MA increases its membrane affinity by several orders of magnitude. When green fluorescent protein fusions of monomeric or dimeric MA are expressed in cells, the dimeric but not the monomeric protein becomes strongly membrane associated. Computational modeling supports these results and suggests a molecular mechanism for the modest effect of myristoylation on binding, wherein the membrane provides a hydrophobic environment for the myristate that is energetically similar to that provided by the protein. Overall, the results imply that the driving force for membrane association stems largely from ionic interactions between multimerized Gag and negatively charged phospholipids.     

Characterization of an RNA receptor motif that recognizes a GCGA tetraloop

Tertiary interactions between a new RNA motif and RNA tetraloops were analyzed to determine whether this new motif shows preference for a GCGA tetraloop. In the structural context of a ligase ribozyme, this motif discriminated GCGA loop from 3 other tetraloops. The affinity between the GCGA loop and its receptor is strong enough to carry out the ribozyme activity.     

In vitro identification and characterization of an early complex linking HIV-1 genomic RNA recognition and Pr55Gag multimerization

The minimal protein requirements that drive virus-like particle formation of human immunodeficiency virus type 1 (HIV-1) have been established. The C-terminal domain of capsid (CTD-CA) and nucleocapsid (NC) are the most important domains in a so-called minimal Gag protein (mGag). The CTD is essential for Gag oligomerization. NC is known to bind and encapsidate HIV-1 genomic RNA. The spacer peptide, SP1, located between CA and NC is important for the multimerization process, viral maturation and recognition of HIV-1 genomic RNA by NC. In this study, we show that NC in the context of an mGag protein binds HIV-1 genomic RNA with almost 10-fold higher affinity. The protein region encompassing the 11th alpha-helix of CA and the proposed alpha-helix in the CA/SP1 boundary region play important roles in this increased binding capacity. Furthermore, sequences downstream from stem loop 4 of the HIV-1 genomic RNA are also important for this RNA-protein interaction. In gel shift assays using purified mGag and a model RNA spanning the region from +223 to +506 of HIV-1 genomic RNA, we have identified an early complex (EC) formation between 2 proteins and 1 RNA molecule. This EC was not present in experiments performed with a mutant mGag protein, which contains a CTD dimerization mutation (M318A). These data suggest that the dimerization interface of the CTD plays an important role in EC formation, and, as a consequence, in RNA-protein association and multimerization. We propose a model for the RNA-protein interaction, based on previous results and those presented in this study.     

Mapping the RNA binding sites for human immunodeficiency virus type-1 gag and NC proteins within the complete HIV-1 and -2 untranslated leader regions.

Encapsidation of HIV-1 genomic RNA is mediated by specific interactions between the RNA packaging signal and the Gag protein. During maturation of the virion, the Gag protein is processed into smaller fragments, including the nucleocapsid (NC) domain which remains associated with the viral genomic RNA. We have investigated the binding of glutathione- S -transferase (GST) Gag and NC fusion proteins from HIV-1, to the entire HIV-1 and -2 leader RNAencompassing the packaging signal. We have mapped the binding sites at conditions where only about two complexes are formed and find that GST-Gag and GST-NC fusion proteins bind specifically to discrete sites within the leader. Analysis of the HIV-1 leader indicated that GST-Gag strongly associates with the PSI stem-loop and to a lesser extent with regions near the primer binding site. GST-NC binds the same regions but with reversed preferences. The HIV-1 proteins also interact specifically with the 5'-leader of HIV-2 and the major site of interaction mapped to a stem-loop, with homology to the HIV-1 PSI stem-loop structure. The different specificities of Gag and NC may reflect functionally distinct roles in the viral replication, and suggest that the RNA binding specificity of NC is modulated by its structural context.