In Vitro Characterization of the Interaction between HIV-1 Gag and Human Lysyl-tRNA Synthetase*

Human immunodeficiency virus type 1 (HIV-1) viral assembly is mediated by multiple protein-protein and protein-nucleic acid interactions. Human tRNALys3 is used as the primer for HIV reverse transcription, and HIV Gag and GagPol are required for packaging of the tRNA into virions. Human lysyl-tRNA synthetase (LysRS) is also specifically packaged into HIV, suggesting a role for LysRS in tRNA packaging. Gag alone is sufficient for packaging of LysRS, and these two proteins have been shown to interact in vitro using glutathione S-transferase pull-down assays. In vitro pull-down assays using truncated constructs have also revealed that residues important for homodimerization of Gag and LysRS are critical for the Gag/LysRS interaction. In this work, we report further in vitro characterization of the interaction between HIV Gag and human LysRS using affinity pull-down assays, fluorescence anisotropy measurements and gel chromatography. An equilibrium binding constant of 310 ± 80 nm was measured for the Gag/LysRS interaction. We also show that capsid alone binds to LysRS with a similar affinity as full-length Gag. Point mutations that disrupt the homodimerization of LysRS and Gag in vitro do not affect their interaction. These results suggest that dimerization of each protein per se is not required for the interaction but that residues involved in forming the homodimer interfaces contribute to heterodimer formation. Gel chromatography studies further support the formation of a Gag/LysRS heterodimer.

the single-stranded viral RNA is reverse-transcribed into double-stranded DNA. Host-encoded tRNA Lys3 is the primer used by HIV-1 for initiation of reverse transcription (2). Newly synthesized proviral DNA translocates into the nucleus and is integrated into the host chromosomal DNA. Transcription of viral DNA yields spliced and unspliced mRNAs, progeny RNA genomes, and viral proteins. Among these viral proteins are two large precursor proteins, Gag and GagPol, both of which are translated from the same full-length viral RNA (3). During viral maturation, Gag is processed into the mature viral proteins: matrix, capsid (CA), and nucleocapsid. GagPol processing yields the mature HIV-1 enzymes: protease, reverse transcriptase, and integrase. In the last step of the viral life cycle, Gag, GagPol, genomic RNA, and specific host cell components assemble at plasma membranes and are eventually released from the cell.
Human tRNA Lys3 , one of the three major tRNA Lys isoacceptors, is used as the primer for reverse transcription in HIV-1, although all three isoacceptors are selectively packaged into the newly forming virion (2,4). Previous work has shown that packaging of tRNA Lys requires GagPol (5), as well as human lysyl-tRNA synthetase (LysRS), which is also selectively packaged into HIV-1 (6). LysRS can be packaged into viral-like particles composed only of Gag, suggesting that interactions with Gag are both necessary and sufficient for LysRS incorporation (6).
Human LysRS is a class II synthetase, belonging to a closely related subgroup (IIb) with aspartyl-and asparginyl-tRNA synthetases (7,8). The crystal structures of Escherichia coli LysRS (9) and Thermus thermophilus LysRS (10) have been solved. LysRS is a homodimer, with each monomer consisting of an N-terminal anticodon binding domain, a dimerization domain formed by motif 1, and motifs 2 and 3 that together constitute the aminoacylation active site. LysRS is one of the most highly conserved synthetases, and sequence alignments suggest that prokaryotic and eukaryotic LysRSs are structurally very related (11).
Using wild-type (WT) and truncated HIV-1 Gag and LysRS variants, the regions critical for the protein-protein interaction have been mapped using in vitro glutathione S-transferase pulldown assays and in vivo LysRS packaging studies (12). The Gag/ LysRS interaction depends on Gag sequences within the C-terminal domain (CTD) of CA and amino acids 208 -259 in motif 1 of LysRS. Interestingly, these two regions contain elements involved in the formation of the dimerization interface of each protein. This suggests either that dimerization is critical for the interaction or that monomer units of each protein interact to form a heterodimeric Gag/LysRS complex.
In this report, we attempt to distinguish between these alternative models by constructing point mutants to disrupt homodimerization of LysRS and Gag. We characterize the interaction of WT and mutant proteins using in vitro affinity pull-down assays, fluorescence anisotropy, and gel chromatography. We also characterize the binding between LysRS and CA. Taken together, the results of this study lead to a refined model for the HIV-1 tRNA Lys -packaging complex.
The plasmid encoding histidine-tagged human TrpRS on a pET20b vector was a gift from Dr. Paul Schimmel (Scripps Research Institute, La Jolla, CA). BL21(DE3) cells transformed with the TrpRS expression vector were grown in LB containing 100 g/ml ampicillin at 37°C. Protein expression was induced at an A 600 nm of 0.4 -0.6 by the addition of isopropyl-1-thio-␤-D-galactopyranoside to a final concentration of 1 mM for 4 h at 37°C. Cells were harvested by centrifugation, and the purification was performed on Ni 2ϩ -nitrilotriacetic acid (NTA) resin (Qiagen) equilibrated with 20 mM Tris⅐HCl, pH 8. TrpRS was eluted from the column using an imidazole gradient (30 -250 mM) with TrpRS eluting at 100 mM imidazole. Purified TrpRS was then dialyzed against phosphate-buffered saline, concentrated, and diluted to 50% glycerol for storage at Ϫ80°C. The final storage conditions were 10 mM NaPO 4 , pH 7.4, 1.5 mM KPO 4 , 3 mM KCl, 137 mM NaCl, and 50% glycerol.
The plasmid encoding HIV-1 CA (WISP98-85) was a gift from Dr. Wesley Sundquist (University of Utah, Salt Lake City, UT). The expression and purification of CA was adapted from Sundquist and co-workers (17). BL21(DE3) cells transformed with the CA expression vector were grown in medium containing 100 g/ml ampicillin at 37°C. CA protein expression was induced at an A 600 nm of 0.8 by the addition of isopropyl-1-thio-␤-D-galactopyranoside to a final concentration of 1 mM for 4 h. The cells were harvested by centrifugation, and the remaining steps were performed at 4°C. Cells were resuspended in buffer B containing 20 mM KMOPS, pH 6.9, 5 mM ␤-ME, and two Complete mini protease inhibitor tablets (Roche Applied Science). Following sonication, cells were centrifuged at 10,000 ϫ g for 30 min. CA was precipitated with 20% ammonium sulfate and exchanged into buffer B by dialysis using a 10,000 molecular weight cut off Slide-A-Lyzer cassette (Pierce). Purification was achieved by cation-exchange chromatography using SP-Sepharose resin (Amersham Biosciences). CA was eluted using a 250-ml linear salt gradient (0 -1 M NaCl) in buffer B. The purified protein, which typically eluted at 300 mM NaCl, was exchanged into 20 mM Tris⅐HCl, pH 8.0, and 5 mM ␤-ME. The concentrations of all proteins were determined by the Bradford method using the Bio-Rad protein assay kit and bovine serum albumin as the standard (18).
Fluorophore Labeling of Proteins-A solution of Alexa Fluor 488 C 5 maleimide (AF) from Molecular Probes was prepared in dimethyl sulfoxide, and the concentration was determined using the extinction coefficient supplied by the manufacturer (⑀ 493 nm ϭ 77,100 M Ϫ1 cm Ϫ1 ). Prior to labeling, LysRS, Mut-LysRS, or TrpRS were first incubated at room temperature for 15 min followed by elution through a 1-ml G50 Sephadex spin column to remove dithiothreitol. The proteins (40 M) were labeled with AF at a 10:1 AF:protein ratio for 5 min on ice in 40 mM HEPES, pH 7.5, and 40 mM KCl. The reaction was quenched, and unreacted dye was removed by passing the labeling reaction through two 1-ml G50 Sephadex spin columns.
HIV-1 CA was labeled similarly, using fluorescein isothiocyanate (FITC). The concentration of the stock solution of FITC in dimethyl formamide was determined at pH Ͼ9 using the extinction coefficient ⑀ 494 nm ϭ 73,000 M Ϫ1 cm Ϫ1 . The final labeling reaction was incubated for 15 min at room temperature. The final labeling stoichiometries for LysRS:AF, TrpRS:AF, Mut-LysRS:AF, and CA:FITC were estimated to be 2:1, 2:1, 1.5:1, and 1:1.6 fluorophore:protein, respectively. These values were determined by measuring the absorbance at 494 nm and using the extinction coefficients of the free fluorophores. Protein concentrations were estimated using the Bradford method (18). Samples were also subjected to 10% SDS-polyacrylamide gel electrophoresis. Ultraviolet illumination of the gels confirmed that our final labeled products had little or no free fluorophore.
tRNA Lys3 Preparation and Aminoacylation Activity-Unmodified WT tRNA Lys3 was prepared by in vitro transcription from FokI-digested plasmid pLYSF119 using T7 RNA polymerase as described previously (11). The extinction coefficient of 604,000 M Ϫ1 cm Ϫ1 was used to determine the concentration of all the tRNAs. Aminoacylation assays to test the activity of unlabeled and labeled LysRS and Mut-LysRS proteins were conducted as described previously (11). For WT LysRS and LysRS: AF488, assays contained 25 nM LysRS and tRNA Lys3 concentrations of 0.5 and 1.0 M, respectively. For Mut-LysRS, protein concentrations ranged from 0.025 to 1 M, and tRNA Lys3 concentrations ranging from 0.5 to 2 M were used.
Electrophoretic Band-shift Assay-The interaction between WT LysRS or Mut-LysRS and human tRNA Lys3 was measured using an electrophoretic band-shift assay. Dephosphorylation and 5Ј-32 P-end labeling of the tRNA Lys3 was performed using a modification of the method described earlier (19). Prior to 5Ј-end labeling, 20 g of unmodified tRNA Lys3 was denatured for 2 min at 90°C, cooled on ice, and incubated for 1 h at 60°C with four units of bacterial alkaline phosphatase in 25 mM Tris⅐HCl, pH 8, 0.1% SDS, and 25% formamide. The dephosphorylated tRNA was then phenol/chloroformextracted, ethanol-precipitated, and 5Ј-32 P-end-labeled by incubation with 30 units of phage T4 polynucleotide kinase (New England Biolabs) and 100 Ci of [␥-32 P]ATP for 30 min at 37°C in the buffer supplied by the manufacturer. Labeled tRNA was purified by electrophoresis on a 12% denaturing polyacrylamide gel.
Labeled tRNA Lys3 (4.5 M) was incubated with different concentrations of either WT LysRS or Mut-LysRS (0.15, 0.375, 0.75, or 1.5 M) in 20 mM Tris⅐HCl, pH 7.4, 50 mM NaCl, 0.1 mM EDTA, and 5% glycerol for 20 min at room temperature and analyzed by electrophoresis at 4°C on a 5% Tris-borate-EDTA non-denaturing polyacrylamide gel prepared with 0.25ϫ Trisborate-EDTA (22.5 mM Tris⅐borate and 0.25 mM EDTA) and 0.5ϫ Tris-borate-EDTA running buffer. The gels were visualized using a Bio-Rad molecular imager FX, and band densities of free and bound tRNA Lys3 were determined using Quantity One software. Equilibrium binding constants were determined by fitting these data to a standard hyperbolic curve using Kaleidagraph.
Gel Chromatography-The oligomeric states of WT LysRS and Mut-LysRS in the absence of Gag were initially monitored by sizeexclusion chromatography using a Superose 12 column (Amersham Biosciences) attached to an Amersham Biosciences fast performance liquid chromatography (FPLC) system. The mobile phase was 50 mM NaPO 4 , pH 7, 150 mM NaCl, and 10 mM ␤-ME. The column was calibrated using molecular mass markers ranging from 13.7 to 440 kDa (Amersham Biosciences). A 2.5 M solution of WT LysRS or Mut-LysRS was loaded onto the column, and protein absorbance at 280 nm was monitored.
The monomer-dimer equilibrium of WT LysRS and Mut-LysRS was analyzed by elution through a Bio-Sil 250-5 sizeexclusion column (Bio-Rad) attached to a Beckman high performance liquid chromatography (HPLC) system. The mobile phase was 50 mM NaPO 4 , pH 7.5, and 150 mM NaCl. The column was calibrated using gel chromatography molecular weight mass ranging from 1.35 to 670 kDa (Bio-Rad). To esti-mate monomer-dimer equilibrium dissociation constants, solutions of each protein ranging from 0.5 to 20 M were loaded onto the column, and protein absorbance at 280 nm was monitored. The areas under the monomer and dimer peaks were quantified using Beckman PACE 32 Karat software version 4.0, and the peak ratios were plotted versus Mut-LysRS concentration. The K d was assumed to be the concentration at which we observed a 1:1 ratio of monomer to dimer peaks.
The interaction between LysRS and Gag was analyzed by size-exclusion FPLC carried out at 4°C using a Superose 12 column (Amersham Biosciences). The mobile phase was 50 mM NaPO 4 , pH 7, and 150 mM NaCl. Solutions of LysRS alone (15 M), Gag alone (40 M), and LysRS/Gag (15 M/40 M) were incubated in mobile phase buffer at room temperature for 2 h. Each sample (100 l) was injected separately on the column, and protein elution was monitored at 280 nm. The traces were digitized using Origin 7 software (Microcal). Fractions (250 l) were collected every 30 s and analyzed by 10% SDS-polyacrylamide gel electrophoresis. Gels were stained with Coomassie Brilliant Blue and scanned to visualize the bands.
Circular Dichroism Analysis-CD spectra were measured at room temperature using a J-710 spectropolarimeter (Jasco) with a 0.1-cm path-length cuvette. Prior to analysis, proteins were dialyzed into 10 mM NaPO 4 , pH 7.5, and diluted to a concentration of 5 M. Spectra were accumulated over six scans.
In Vitro Pull-down Assays-Purified human LysRS, Mut-LysRS, human TrpRS, human ProRS, HIV-1 Gag, WM-Gag, and CA were dialyzed against binding/dialysis buffer containing 25 mM NaPO 4 , pH 7, 125 mM NaCl, 2.5 mM ␤-ME at 4°C. The histidine-tagged LysRS or TrpRS (250 g) were incubated with 25 l of Ni 2ϩ -NTA resin in a volume of 125 l for 1 h at room temperature. The resin was washed once

HIV-1 Gag and Lysyl-tRNA Synthetase Interaction
with binding/dialysis buffer to remove any unbound LysRS, TrpRS, or ProRS. The dialyzed Gag, WM-Gag, or CA (125 g in a volume of 150 l) was added to the LysRS or TrpRS bound to Ni 2ϩ -NTA and incubated for 2 h at room temperature. As a control, a similar binding experiment was performed in the absence of bound synthetase. The resin was then washed once with binding/dialysis buffer to remove any unbound protein followed by six washes with binding/dialysis buffer containing 10 mM imidazole to remove any protein non-specifically bound to the resin. To elute the bound protein, the resin was incubated in 150 l of binding/dialysis buffer containing 200 mM imidazole for 1 h at room temperature, eluted from the resin by centrifugation, and diluted with 2ϫ gel loading buffer containing 50 mM Tris⅐HCl, pH 6.8, 100 mM ␤-ME, 2% SDS, 0.1% bromphenol blue, and 10% glycerol.
Samples were run on a 10 or 12% SDS-polyacrylamide gel followed by transfer onto a polyvinylidene fluoride Immobilon-P membrane (Millipore). Gag, WM-Gag, and CA were detected using a polyclonal antibody for HIV-1 CA from Pocono Rabbit Farm and Laboratory. A goat anti-rabbit horseradish peroxidase conjugate (Bio-Rad) was used as the secondary antibody. Detection was performed by enhanced chemiluminescence using Western blotting detection reagents from Amersham Biosciences.
Fluorescence Anisotropy Measurements-Equilibrium dissociation constants were determined by measuring the fluorescence anisotropy of 50 nM fluorescently labeled protein (LysRS:AF, TrpRS:AF, Mut-LysRS:AF, or CA:FITC) as a function of increasing concentrations of an unlabeled protein (LysRS, TrpRS, Mut-LysRS, Gag, or WM-Gag). The labeled protein was incubated with varying amounts of the desired unlabeled protein for 30 min at room temperature in 40 mM HEPES, pH 7.5, and 50 mM NaCl. Anisotropy measurements were made on a Photon Technology International spectrofluorimeter (Model QM-2000). The excitation and emission wavelengths were 490 and 520 nm, respectively (slit widths ϭ 5 nm).
Anisotropy was measured using the time-based function for 30 s (integration time ϭ 1 s; resolution ϭ 8 s), and the data were averaged. All measurements were carried out at least three times. The titration curves were fit to the following equation, which assumes a 1:1 binding stoichiometry (20 -22), where A is the measured anisotropy at a particular total concentration of the unlabeled protein (S) and the labeled protein (Y), A min is the minimum anisotropy, A max is the final maximum anisotropy, and K d is the dissociation constant.

Design and Characterization of Monomeric LysRS and Gag-
To investigate whether monomeric forms of human LysRS and Gag could interact, we designed variants containing point mutations in the known dimerization motif of each protein using the three-dimensional structure of E. coli LysRS as a model for the human synthetase (9). Fig. 1 shows the residues in the motif 1 dimerization domain, Arg-247, Glu-265, Phe-283, that were mutated to alanine to generate Mut-LysRS. These residues are highly conserved in known LysRS sequences and participate in interactions that appear to stabilize the dimer interface. The triple mutant was successfully overexpressed and purified. CD analysis indicated that the secondary structure of Mut-Lys-RS was similar to that of WT LysRS ( Fig. 2A). Class II synthetases require homodimerization for aminoacylation function (23). Thus, if the mutations do in fact disrupt dimerization, Mut-LysRS should show little or no aminoacylation activity. Fig. 2B shows that even at high concentrations of Mut-LysRS, the triple mutant lacks aminoacylation activity, consistent with this expectation. Using an electrophoretic band-shift assay, Mut-LysRS was also tested for its ability to bind tRNA Lys3 . Although it lacked aminoacylation activity, the triple mutant still bound tRNA with a similar affinity (ϳ670 nM) as WT LysRS (ϳ430 nM) (Fig. 2B, inset).
Gel chromatography experiments performed with WT LysRS and Mut-LysRS were also consistent with a shift in the equilibrium distribution toward monomeric species for the triple mutant (Fig. 2C). By carrying out gel chromatography as a function of Mut-LysRS concentration, we estimated the dissociation constant for the monomer-dimer equilibrium to be 24.5 M (Fig. 2D). Similar experiments performed with WT LysRS showed that it remained primarily dimeric at all concentrations tested (data not shown). Based on these data, the monomerdimer equilibrium dissociation constant of WT LysRS is Ͻ200 nM. Taken together, these data support successful disruption of interactions that are critical for dimerization in the triple mutant.
Individual mutations of W184A and M185A in the CTD of CA have been reported to prevent CA dimerization in vitro (24). In addition, the same individual mutations in the context of Gag, W317A and M318A, inhibit Gag-Gag interactions (25). In this study, we simultaneously mutated both of these residues to alanine to generate WM-Gag. Fig. 1 (inset) shows the location of these residues in the CTD of CA. Using sedimentation equilibrium experiments, the monomer association constants were determined to be 5.5 M for WT ⌬p6-Gag and Ͼ20 M for WM-Gag. 4 These results support significant disruption of Gag-Gag interactions in the double mutant.
Characterization of the Fluorescently Labeled Proteins-To confirm that the fluorescently labeled synthetases were properly folded and active, we measured the aminoacylation activity of labeled and unlabeled proteins. No detectable difference in the efficiency of aminoacylation relative to unlabeled protein was observed for LysRS:AF or TrpRS:AF (data not shown). Although we were unable to test an enzymatic activity for CA:FITC, CD analysis showed no difference in the secondary structure of CA:FITC relative to unlabeled CA (data not shown).
In Vitro Affinity Pull-down Assays-Purified histidine-tagged LysRS, Mut-LysRS, TrpRS, or ProRS were absorbed to Ni 2ϩ -NTA resin and incubated with Gag, WM-Gag, or CA. Following extensive washing to remove non-specifically bound protein, the material on the resin was eluted with 200 mM imidazole and subjected to SDS-polyacrylamide gel electrophoresis and immunoblot analysis.
LysRS pulls down both Gag and WM-Gag to similar extents (Fig. 3A, lanes 2 and 6). Negative control experiments per-  4 and 8) of LysRS. Lane 9 is a Gag standard. The Wash lanes correspond to the final wash fraction in each experiment. The Resin lanes correspond to a negative control in which Gag or WM-Gag was incubated with just the Ni 2ϩ -NTA resin to test for nonspecific binding. B, interaction between Gag and WT LysRS (lanes 1 and 2), TrpRS (lanes 3 and 4), ProRS (lanes 5 and 6), and no synthetase as a negative control (lanes 7 and 8). Lane 9 is a Gag standard. C, interaction between CA and WT LysRS (lanes 2 and 3) or TrpRS (lanes 4 and 5). Lane 1 is a CA standard. The Resin lanes correspond to a negative control in which CA was incubated with just Ni 2ϩ -NTA resin to test for nonspecific binding (lanes 6 and 7). D, interaction between Mut-LysRS and Gag (lanes 1 and 2), WM-Gag (lanes 3 and 4), or CA (lanes 5 and 6). The positions of protein standards are indicated on the left of the gel.
formed with Gag or WM-Gag in the absence of LysRS did not yield a Gag signal (Fig. 3A, lanes 4 and 8). This demonstrates that Gag and WM-Gag do not bind to the Ni 2ϩ -NTA resin in the absence of LysRS. The final wash fractions were also examined by SDS-polyacrylamide gel electrophoresis and immunoblot analysis and contained no protein (Fig. 3A, lanes 1, 3,  5, and 7).
A pull-down experiment designed to test the specificity of the LysRS/Gag interaction was carried out using human TrpRS and human ProRS (Fig. 3B). We observe that both LysRS and TrpRS were able to bind and pull down Gag, although the interaction with TrpRS appeared to be weaker (Fig. 3B, lanes 2 and  4). No signal was observed in the case of ProRS (Fig. 3B, lane 6).
The critical region for the LysRS interaction was previously mapped to amino acids 323-362 of Gag, located in the CTD of CA (12). To investigate whether the CA domain of Gag was sufficient for an interaction with LysRS, a pull-down experiment was performed using immobilized WT LysRS or TrpRS and CA (Fig. 3C ). Both LysRS (lane 3) and TrpRS (lane 5) could pull down CA, showing that this domain of Gag is sufficient for the interaction. Negative control experiments performed with CA in the absence of synthetase did not show significant binding (Fig. 3C, lanes 6 and 7 ).
Mut-LysRS was also tested for its ability to pull down Gag, WM-Gag, and CA. Based on the concentration of Mut-LysRS applied to the column (25 M), at least 50% is expected to be monomeric. The results shown in Fig. 3D (lanes 2, 4, and 6) show that Mut-LysRS is able to interact with all three proteins tested.
Characterization of Binding Affinity-Fluorescence anisotropy experiments were used to determine equilibrium dissociation constants (K d ) for LysRS and Mut-LysRS binding to Gag, WM-Gag, and CA. The binding curve obtained for the WT LysRS:AF/Gag complex is shown in Fig. 4A. The data were fit to a binding model that assumes 1:1 binding stoichiometry, and an apparent K d of 310 Ϯ 80 nM was determined. An ϳ5-fold higher K d of 1550 Ϯ 290 nM was measured for the TrpRS:AF/Gag complex (Table 1). We were unable to detect an interaction between ProRS:AF and Gag using this assay.
Both LysRS and TrpRS also bound to CA, but once again, a 4-fold higher K d was measured for the TrpRS/CA complex relative to the LysRS/CA complex ( Table 1). The affinity of the CA:FITC/LysRS interaction (420 Ϯ 10 nM, Fig. 4B) is similar to that of the LysRS:AF/Gag interaction, in accord with the finding that the Gag/LysRS interaction is dependent on residues in the CA domain of Gag (12). Interestingly, binding of CA to LysRS was independent of the concentration of monovalent cations over the range 50 -500 mM NaCl (data not shown). This suggests that the protein-protein interaction between LysRS and CA is not ionic in nature.
Packaging of LysRS into HIV-1 or virus-like particles occurs independently of tRNA Lys packaging (6). To test whether tRNA affects the LysRS/CA interaction, we added an equal molar given under "Experimental Procedures." Only a representative data set is shown, but all measurements were carried out at least three times with the standard deviation indicated (see also Table 1).  Fig. 4C) was similar to the affinity measured for the WT proteins. Based on the known strength of the homodimer interactions, these experiments were conducted under conditions in which both proteins are expected to be monomeric (50 nM Mut-LysRS and Յ12 M WM-Gag). In addition, the affinities of monomeric LysRS or Gag for the corresponding WT proteins (Mut-LysRS/Gag, Mut-LysRS/CA, and LysRS/WM-Gag) were ϳ2-fold greater (770 Ϯ 100 nM, 530 Ϯ 60 nM, and 590 Ϯ 70 nM, respectively, Table 1) than the affinities between the two monomeric proteins or the two WT proteins.
Gel Chromatography Analysis of Heterodimer Formation-The stoichiometry of the LysRS/Gag interaction was further investigated by gel chromatography. Fig. 5A shows the results of three separate experiments performed following a 2-h incubation of the proteins either alone or together. LysRS alone (MW dimer ϭ 138 kDa) results in the appearance of a major peak eluting at 25 min, likely corresponding to a LysRS dimer ( peak 1, apparent molecular mass ϭ 160 kDa). Gag alone (MW dimer ϭ 100 kDa) results in the appearance of a single major peak eluting at 33 min, likely corresponding to a Gag dimer ( peak 3, apparent molecular mass ϭ 110 kDa). When both LysRS and Gag were incubated together at room temperature for 2 h, an additional species appeared at 31 min ( peak 2, apparent molecular mass ϭ 125 kDa). The predicted molecular mass of a LysRS/Gag heterodimer (119 kDa) is consistent with peak 2. Importantly, there is no appearance of a new species with molecular mass greater than the LysRS dimer, ruling out the formation of a LysRS/Gag ␣ 2 ␤ 2 complex.
To confirm that both proteins were present in peak 2, fractions from each of the three samples were collected every 30 s and analyzed by 10% SDS-polyacrylamide gel electrophoresis. Fig. 5B shows the fractions that eluted at 29.5-32 min. Although the LysRS-only (Fig. 5B, top) and the Gag-only (Fig.  5B, bottom) samples contain very little protein in the 31-min fraction, both proteins are clearly present in this fraction in the sample in which they were incubated together (Fig. 5B, middle). Thus, peak 2 is not the result of a shift in the elution of one of the individual proteins, and this result strongly supports the formation of a heterodimeric LysRS/Gag complex.

DISCUSSION
The proposed model for human tRNA Lys packaging into HIV-1 involves an assembly complex formed between genomic RNA, Gag, Gag-Pol, tRNA Lys , and human LysRS (26). The details of the molecular interactions between the components of this packaging complex are not known. As a first step toward this goal, we have begun to characterize the interaction between human LysRS and Gag, which occurs independent of the other components. The strength of the interaction between the WT proteins (K d ϭ 310 Ϯ 80 nM) is unaffected by the presence of tRNA. The interaction was previously local-ized to the CTD of CA (12), and indeed, our data show that the CA protein interacts with LysRS with a similar affinity as Gag.
The Gag/LysRS interaction depends on the dimerization motif of each protein (12). We show here that dimerization of each of the interacting partners can be disrupted by site-directed mutagenesis without significantly affecting the strength of the protein-protein interaction. These results raise the possibility that there is a slow interconversion of monomers and dimers, with the heterodimer formed only from LysRS and/or

HIV-1 Gag and Lysyl-tRNA Synthetase Interaction
Gag in the monomeric state. Since our K d calculations used the concentration of total Gag and LysRS present in solution, the calculated value represents an upper limit for the dissociation constant, which may be significantly lower for monomeric wild-type proteins. Nevertheless, taken together, these results suggest that homodimerization per se is not critical for the Gag/LysRS interaction but that a similar, although probably not identical, interface may be used for the hetero-protein interaction. Consistent with this idea, recent structural studies of a mammalian SCAN domain dimer, a homolog of the HIV-1 CA CTD, suggest that the interaction domain of CA is somewhat plastic and can adopt a different dimerization interface by swapping the major homology region element between monomers (27). This plasticity is likely advantageous as CA must participate in a wide variety of interactions throughout the viral life cycle (28).
Gel chromatography studies performed here in the absence of RNA are consistent with a Gag/LysRS complex size corresponding to a heterodimer. No evidence for a higher molecular weight complex was obtained (e.g. a dimer of dimers). The interaction between CA monomers has been reported to be relatively weak (K d ϭ 18 M) (24). The interaction between Gag monomers is similarly weak. 4 In contrast, the interaction between monomers in the LysRS homodimer is quite strong (K d Ͻ 200 nM). The precise mechanism of heterodimer formation is not clear and may involve other factors that serve to help disrupt the LysRS homodimer.
The specificity of the Gag/LysRS interaction was previously probed in vivo by examining the packaging of other synthetases into HIV-1 virions (29). Of the eight synthetases examined, only LysRS was found in HIV-1 (29). TrpRS was not detected in HIV-1. Conversely, in avian Rous sarcoma virus, which uses tRNA Trp as a primer, TrpRS is found, whereas LysRS is not present (30 -32). Whether this incorporation specificity in HIV-1 can be accounted for by the 5-fold reduction in the affinity of TrpRS for HIV-1 Gag when compared with LysRS is not certain, and other parameters, such as cellular factors or cellular compartmentalization, may contribute to the specificity of incorporation of LysRS into HIV-1. We were unable to detect an interaction between human ProRS and Gag in our in vitro studies. This synthetase is not found in HIV-1 (4) and is also not packaged into murine leukemia virus despite the fact that tRNA Pro is the primer in the murine system (32). The molecular interaction between motif 1 of LysRS and the CTD of CA represents a new target for anti-retroviral therapy, and studies to more finely map the residues involved in the protein-protein interaction are currently underway.