Correlation between tRNALys3 aminoacylation and its incorporation into HIV-1.

During human immunodeficiency virus type 1 (HIV-1) assembly, tRNA(Lys) is selectively packaged into the virus, where tRNA(Lys3) serves as the primer for reverse transcription. Lysyl-tRNA synthetase is also selectively incorporated into HIV-1 and is therefore a strong candidate for being the signal by which viral proteins interact with tRNA(Lys) isoacceptors. Previously, mutations in the tRNA(Lys3) anticodon have been shown to strongly inhibit the charging of tRNA(Lys3) by lysyl-tRNA synthetase in vitro, and we show here that in vivo aminoacylation is also inhibited by anticodon changes. The order of decreasing in vivo aminoacylation for tRNA(Lys3) anticodon mutants is: wild-type SUU (where S = mcm(5)S(2)U) 100%) --> SGU (49%) --> CGU (40%) --> SGA (0%) and CGA (0%). We found that the ability of these tRNA(Lys3) anticodon variants to be aminoacylated in vivo is directly correlated with their ability to be packaged into HIV-1. These data showed that the anticodon is a major determinant for tRNA(Lys3) packaging and support the conclusion that its productive interaction with lysyl-tRNA synthetase is important for tRNA(Lys3) incorporation into HIV-1.

As described previously, LysRS is selectively packaged into HIV-1 (4). During HIV-1 assembly (see Ref. 5 for review), the major structural precursor protein, Gag, is assembled at the plasma membrane. Also, part of the assembly complex is viral genomic RNA and another precursor protein, Gag-Pol, which contains the sequences for the viral enzymes, protease, reverse transcriptase (RT), and integrase. During or after budding of the virion from the cell, the viral protease converts these precursor proteins to the processed proteins found in the mature virion. However, the incorporation of tRNA Lys occurs independently of both precursor protein processing and genomic RNA packaging (6). Gag alone is sufficient to produce extracellular Gag particles, and LysRS is packaged into these particles (4), but the additional presence of Gag-Pol is required for tRNA Lys incorporation as well (6). Gag-Pol might be stabilizing the Gag/LysRS/tRNA Lys complex since Gag-Pol interacts with both Gag (7,8) and tRNA Lys (9,10).
The recognition and binding of aminoacyl-tRNA synthetases with their cognate tRNAs involve interactions with the acceptor stem and, in most cases, the anticodon arms of the tRNAs (11). Based on studies of the human tRNA Lys3 /human LysRS interaction, the anticodon of tRNA Lys3 plays an important role in LysRS recognition and binding (12). In contrast, the enzyme is relatively insensitive to mutations in the acceptor stem domain both in vivo (13) and in vitro (12). Another recent report indicated that the N-terminal domain of hamster LysRS, which is adjacent to the anticodon binding domain, although not essential for aminoacylation, improves the docking of the acceptor arm of tRNA Lys3 into the active site of the enzyme (14).
Previous work has indicated that a certain variability in the anticodon sequence is tolerated for tRNA Lys packaging into virions, i.e. not only do tRNA Lys3 (anticodon SUU, where S ϭ mcm 5 S 2 U) 2 and tRNA Lys1,2 (anticodon CUU) appear to be packaged with equal efficiency, but a mutant tRNA Lys3 with a CUA anticodon is packaged, although aminoacylated to only 40% of wild-type levels in vivo (15). However, mutations at U35 have been shown to have a more severe effect on aminoacylation catalytic efficiency in vitro (12,16), and changes at this position have not yet been tested for their effect on tRNA Lys3 aminoacylation in vivo nor for packaging into virions. In this report, we constructed different tRNA Lys3 genes mutated at position U35 as well as at other anticodon positions, and we expressed these genes in COS7 cells also transfected with HIV-1 proviral DNA. We show that the anticodon is indeed a major determinant for tRNA Lys3 packaging. Moreover, the ability of mutant tRNA Lys3 molecules to be aminoacylated in vivo correlates directly with their ability to be packaged into HIV-1.

EXPERIMENTAL PROCEDURES
Plasmid Construction-SVC21.BH10, a simian virus 40-based vector containing wild-type HIV-1 proviral DNA, was a gift of E. Cohen, University of Montreal. SVC12.BH10Lys3 UUU contains the HIV-1 proviral DNA plus a wild-type tRNA Lys3 gene with the anticodon DNA sequence TTT. SVC12.BH10Lys3 CGA , SVC12.BH10Lys3 CGU , SVC12.BH10Lys3 SGU , and SVC12.BH10Lys3 SGA contain the HIV-1 proviral DNA plus a mutant tRNA Lys3 gene where the anticodon DNA sequence has been changed from TTT to CGA, CGT, TGT, and TGA, respectively. Mutant tRNA Lys3 genes were created by PCR mutagenesis (11). The amplified products were cloned into the Hpa-I site of SVC21.BH10, which is upstream of the HIV-1 proviral DNA sequence. Mutations were confirmed by DNA sequencing.
Production of Wild-type and Mutant HIV-1 Virus-Transfection of COS7 cells with the above plasmids by the calcium phosphate method was as described previously (6). Viruses were isolated from COS7 cell culture medium 63 h posttransfection or from the cell culture medium of infected cell lines. The virus-containing medium was first centrifuged in a Beckman GS-6R rotor at 3,000 rpm for 30 min, and the supernatant was then filtered through a 0.2-m filter. The viruses in the filtrate were then pelleted by centrifugation in a Beckman Ti45 rotor at 35,000 rpm for 1 h. The viral pellet was purified by centrifugation with a Beckman SW41 rotor at 26,500 rpm for 1 h through 15% sucrose onto a 65% sucrose cushion.
RNA Isolation and Analysis-Total cellular or viral RNA was extracted from cell or viral pellets by the guanidinium isothiocyanate procedure (17) and dissolved in 5 mM Tris buffer, pH 7.5. Dot-blots of cellular or viral RNA were hybridized with DNA probes complementary to tRNA Lys3 and tRNA Lys1,2 (1), genomic RNA (18), and ␤-actin mRNA (DNA probe from Ambion). Two-dimensional PAGE of [ 32 P]Cp 3Ј-endlabeled viral RNA was carried out as described previously (1).
Measurement of Wild-type and Mutant tRNA Lys3 Using RNA-DNA Hybridization-To measure the total amount of tRNA Lys3 (combined wild type and mutant) present in cellular or viral RNA, we used an 18-mer DNA oligonucleotide complementary to the 3Ј 18 nucleotides of tRNA Lys3 (5Ј-TGGCGCCCGAACAGGGAC-3Ј). Previously, this probe has been shown to hybridize specifically with tRNA Lys3 (14) and was hybridized to Dot-blots on Hybond N (Amersham Biosciences) containing known amounts of purified in vitro transcript of tRNA Lys3 and either cellular tRNA or viral RNA produced in cells transfected with either SVC21.BH10 alone or SVC21.BH10 containing a wild-type or mutant tRNA Lys3 gene. The DNA oligomer was first 5Ј-end-labeled using T4 polynucleotide kinase and [␥-32 P]ATP (3000 Ci/mMol, Dupont Canada), and specific activities of from 10 8 to 10 9 cpm/g were generally reached. Approximately 10 7 cpm of oligomer was used per blot in hybridization reactions.
Measurement of in Vivo Aminoacylation-To measure the extent of in vivo aminoacylation of tRNA Lys3 , the isolation of cellular or viral RNA was performed using acidic conditions required for stabilizing the aminoacyl-tRNA bond described previously (19,20). To measure the extent of in vivo aminoacylation of tRNA Lys3 , the isolation of cellular or viral RNA was performed at low pH conditions required for stabilizing the aminoacyl-tRNA bond. The guanidinium isothiocyanate procedure for isolating RNA was modified by including 0.2 M sodium acetate, pH 4.0, in solution D, and the phenol used was equilibrated in 0.1 M sodium acetate, pH 5.0. The final isopropanol-precipitated RNA pellet was dissolved in 10 mM sodium acetate, pH 5.0, and stored at Ϫ70°C until electrophoretic analysis. RNA was mixed with one volume loading buffer (0.1 M sodium acetate, pH 5.0, 8 M urea, 0.05% bromphenol blue, and 0.05% xylene cyanol) and electrophoresed in a 0.5-mm thick polyacrylamide gel containing 8 M urea in 0.1 M sodium acetate, pH 5.0. The running buffer was 0.1 M sodium acetate, pH 5.0, and electrophoresis was carried out at 300 V for 15-18 h at 4°C in a Hoefer SE620 electrophoretic apparatus. RNA was electroblotted onto Hybond N filter paper (Amersham Biosciences) using an electrophoretic transfer cell (Bio-Rad) at 750 mA for 15 min. using 1ϫ TBE buffer (0.09 M Tris borate, pH 8.0, 2 mM EDTA). Hybridization of the blots with probes for wild-type and mutant tRNA Lys3 was performed as described above. Deacylated tRNA was produced by treating the RNA sample with 0.1 M Tris-HCl, pH 9.0, at 37°C for 3 h to hydrolyze the aminoacyl linkage and provide an uncharged electrophoretic marker.
Western Blotting-Western blot analysis was performed using 300 g of cytoplasmic or nuclear proteins as determined by the Bradford assay (21). Cytoplasmic and nuclear extracts were resolved by SDS-PAGE followed by blotting onto nitrocellulose membranes (Gelman Sciences). Detection of the nuclear transcription factor YY1 on the Western blot utilized monoclonal antibodies to YY1 (Santa Cruz Biotechnology). Western blots were analyzed by enhanced chemiluminescence (ECL kit, Amersham Biosciences) using anti-mouse (Amersham Biosciences) as a secondary antibody. The sizes of the detected protein bands were estimated using prestained high molecular weight protein markers (Invitrogen).
Cell Fractionation-The cytoplasmic supernatant and nuclear extract were prepared from the COS7 cells as described previously (22). Western blot analysis was performed as above using anti-YY1.

Expression of Wild-type and Mutant tRNA Lys3 and Their
Incorporation into Virions-The anticodon is a major determinant for aminoacylation of human tRNA Lys3 by human LysRS (12). The ability of tRNA Lys3 to be aminoacylated in vitro was shown to be particularly sensitive to changes at anticodon position U35 (12). To determine whether a correlation exists between the ability of tRNA Lys3 to be aminoacylated and incorporated into HIV-1, we have transfected COS7 cells with a plasmid containing both HIV-1 proviral DNA and a wild-type or mutant tRNA Lys3 gene. As shown previously, this results in more tRNA Lys3 being synthesized in the cytoplasm, and in the case of wild-type tRNA Lys3 and tRNA Lys3 variants examined to date, increased viral packaging has also been observed (23). However, these previously tested tRNAs did not contain changes at U35. Because the middle anticodon position is critical for aminoacylation, the different tRNA Lys3 variants examined in this work all contained a U35G mutation in addition to other possible anticodon mutations such as S34C or U36A (Fig. 1).
We have measured the ability of mutant tRNA Lys3 to be incorporated into virions using two different hybridization probes, which are shown in Fig. 1. We have measured changes in total tRNA Lys3 packaged into virions (Fig. 2) using a hybridization probe complementary to the 3Ј-terminal 18 nucleotides of tRNA Lys3 , which detects both wild-type and mutant tRNA Lys3 . We have also monitored the incorporation into virions of specific mutant tRNA Lys3 using hybridization probes complementary to the anticodon arm, i.e. probes that are specific for each mutant tRNA Lys3 (see Fig. 4). As shown below, both types of probes yield similar conclusions, and both probes were used (see Fig. 6) to measure aminoacylation of either total tRNA Lys3 or of specific wild-type or mutant tRNA Lys3 .
The data in Fig. 2 show both the expression of total (wildtype and mutant) tRNA Lys3 in the cytoplasm of the transfected COS7 cells and their incorporation into HIV-1. Fig. 2A shows Dot-blots of cellular or viral RNA hybridized with a radioactive 18-nucleotide DNA oligomer complementary to the 3Ј-terminal 18 nucleotides of tRNA Lys3 . The top strip represents increasing amounts of synthetic tRNA Lys3 , and the hybridization results are plotted as a standard curve in Fig. 2B. The bottom two strips in Fig. 2A show Dot-blots of RNA isolated from either cell lysates containing equal amounts of ␤-actin mRNA (cell) or viral lysates containing equal amounts of viral genomic RNA (viral). Dot-blots for determining ␤-actin mRNA and genomic RNA amounts are not shown. The relative total tRNA Lys3 /␤actin mRNA ratios are plotted in Fig. 2C, normalized to the value obtained in COS7 cells transfected with HIV-1 proviral DNA alone (BH10). Transfection with the wild-type tRNA Lys3 gene or the mutant tRNA Lys3 genes results in an approximately 2-fold increase in the cytoplasmic concentration of total tRNA Lys3 . However, as shown in Fig. 2D, these cytoplasmic increases in tRNA Lys3 did not all result in increases in tRNA Lys3 incorporation into virions. The maximum increase in tRNA Lys3 incorporation into virions occurred with excess wildtype tRNA SUU Lys3 (1.85-fold). tRNA SGU Lys3 and tRNA CGU Lys3 increased packaging 1.4-and 1.3-fold, respectively. tRNA CGA Lys3 showed no increase in tRNA Lys3 incorporation, and tRNA SGA Lys3 actually showed a small decrease in packaging as compared with wildtype tRNA SUU Lys3 . The changes in tRNA Lys3 incorporation into virions upon transfection with plasmids encoding wild-type tRNA Lys3 or anticodon variants were also visualized by two-dimensional PAGE. Fig. 3A shows the electrophoresis pattern of low molecular weight viral RNA in wild-type virions. Previously, we have identified spot 3 as tRNA Lys3 and spots 1 and 2 as belonging to the tRNA Lys1,2 isoacceptor family (1). Transfection of COS7 cells with a plasmid containing both the wild-type tRNA Lys3 gene and HIV-1 proviral DNA results in virions containing an increase in tRNA Lys3 and a decrease in tRNA Lys1,2 (Fig. 3B). Similar observations were reported previously (23). When COS7 cells are transfected with HIV-1 DNA and either tRNA SGU Lys3 or tRNA CGU Lys3 , which have an intermediate ability to be packaged into virions (Fig. 2D), the two-dimensional PAGE patterns show an increased ratio of tRNA Lys3 /tRNA Lys1,2 in the virions (Fig. 3,  panels C and D). On the other hand, tRNA SGA Lys3 and tRNA CGA Lys3 , which the data indicate are not packaged into virions (Fig. 2D), do not show a significant change in the tRNA Lys3 /tRNA Lys1,2 ratio relative to the wild-type control (Fig. 3, panels E and F). The tRNA Lys3 /tRNA Lys1,2 ratios, determined by phosphorimaging, are listed beneath each panel. Taken together, the data presented in Figs. 2 and 3 show that the ability of tRNA Lys3 anticodon variants to be packaged into HIV is directly correlated with the tRNA Lys3 /tRNA Lys1,2 ratio.
The experiments in Figs. 2 and 3 measure total tRNA Lys3 (wild-type and mutant) in the cytoplasm and in the virion. We have also used hybridization probes specific for each tRNA Lys3 anticodon mutant to examine their specific expression in the cytoplasm and incorporation into virions. The Dot-blots shown in Fig. 4A measure the amount of a specific tRNA Lys3 variant isolated from cell or viral lysates containing equal amounts of ␤-actin mRNA or genomic RNA, respectively. The top strip in panel A (SUU) shows the amounts of tRNA Lys3 in cytoplasm and viruses from cells either transfected with HIV-1 alone (BH10) or transfected with HIV-1 and a wild-type tRNA Lys3 gene (BH10Lys3). The remaining four strips in panel A show the amount of tRNA Lys3 in cytoplasm and viruses from cells transfected with HIV-1 and different mutant tRNA Lys3 genes whose anticodon sequence is listed to the left of the strip. In each of these strips, the wild-type tRNA Lys3 transcript was used as a control to show that the mutant anticodon probes are not detecting wild-type tRNA Lys3 . For each tRNA, a standard hybridization curve was generated and used to calculate the amount (nanograms) present in cell lysate or virus, thereby taking into account any differences in efficiencies of hybridization between the probes.
The relative tRNA Lys3 /␤-actin mRNA ratios are plotted in Fig. 4B, normalized to the value found for cells transfected with HIV-1 alone (BH10). The results are very similar to those shown in Fig. 2 using a DNA hybridization probe that measures total tRNA Lys3 . Wild-type tRNA Lys3 expression is increased significantly when cells are transfected with a wildtype tRNA Lys3 gene. The expression of each mutant tRNA Lys3 in the cytoplasm is similar and results in an approximately 2-fold increase in total tRNA Lys3 (endogenous wild type and mutant) when probed with an oligonucleotide complementary to the 3Ј-end of tRNA Lys3 (Fig. 2C). The tRNA Lys3 /genomic RNA ratios in virions are shown in Fig. 4C, normalized to the value found for cells transfected with HIV-1 proviral DNA alone (BH10). The results with wild-type tRNA SUU Lys3 also match the results shown in Fig. 2D. The incorporation of this tRNA into virions increased the tRNA Lys3 /genomic RNA ratio to 1.87, indicating a relative incorporation of exogenous tRNA Lys3 as FIG. 5. Cytoplasmic expression of wild-type and mutant tRNA Lys3 . COS7 cells were transfected with a plasmid containing HIV-1 proviral DNA and a wildtype or mutant tRNA Lys3 gene, and differential centrifugation was used to separate nuclei and cytoplasm. A, a radioactive DNA probe complementary to the 3Ј-terminal 18 nucleotides present in all forms of tRNA Lys3 studied here was hybridized to Northern blots containing either increasing amounts of an in vitro tRNA Lys3 transcript (left side) or cytoplasmic RNA samples that contain equal amounts of ␤-actin mRNA (right side). On the right side, none refers to the absence of transfection of a gene coding for tRNA Lys3 , whereas the wild-type or mutant tRNA Lys3 anticodon of transfected tRNA Lys3 -encoding genes is listed above each of the remaining lanes. B, graphic representation of the total cytoplasmic tRNA Lys3 /␤-actin mRNA obtained from the data in panel A and normalized to the value obtained in HIV-1-transfected cells not transfected with a tRNA Lys3 -encoding gene. C, Western blot of nuclear and cytoplasmic fractions of transfected cells, labeled as in panel A. Blots were probed with antibody to YY1, a nuclear transcription factor. N, nuclear fraction; C, cytoplasmic fraction. compared with endogenous tRNA Lys3 of 0.87. In contrast, the relative incorporation of tRNA SGU Lys3 and tRNA CGU Lys3 was only 0.50 and 0.37, respectively, whereas tRNA CGA Lys3 and tRNA SGA Lys3 showed relative incorporations of 0.013 and 0.029 (Fig. 4C). The data in Fig. 4 indicate that while wild-type or mutant tRNA Lys3 is expressed at approximately equal levels in the total cell lysate, they are incorporated into virions to variable extents. One explanation could be that some mutant forms of tRNA Lys3 are not exported out of the nucleus with equal efficiency. To test this possibility, we lysed cells and separated nuclei from cytoplasm by low speed centrifugation. A radioactive DNA probe complementary to the 3Ј-terminal 18 nucleotides present in all forms of tRNA Lys3 studied here was hybridized to Northern blots containing both increasing amounts of an in vitro tRNA Lys3 transcript (Fig. 5, left side of panel A) and cytoplasmic RNA samples that contain equal amounts of ␤-actin mRNA (Fig. 5, right side of panel A). The standard curve generated on the left side of panel A shows that the cytoplasmic tRNA Lys3 hybridization signals obtained are within the linear range of the standard curve. The relative total cytoplasmic tRNA Lys3 /␤-actin mRNA obtained from these Northern blot hybridizations is shown graphically in Fig. 5B. These experiments indicate that both wild-type and tRNA Lys3 variants are expressed at approximately equal amounts in the cytoplasm, consistent with the data shown in Figs. 2 and 4. To ensure that effective separation of nuclear and cytoplasmic fractions was achieved in our experiments, we demonstrated that the transcription factor YY1, which concentrates in the nucleus, is only detected in the nuclear fraction (Fig. 5C).
Aminoacylation of Wild-type and Mutant tRNA Lys3 in Vivo-We next determined the aminoacylation state of the wild-type and mutant forms of tRNA Lys3 examined here. The electrophoretic mobility of acylated tRNA in acid-urea PAGE is reduced relative to the deacylated form, and this property can be used to determine the extent of tRNA aminoacylation (23). Fig. 6 shows Northern blots of cellular and viral RNA samples electrophoresed in acid-urea gels, blotted onto Hybond N filter paper, and hybridized with radioactive DNA probes specific for tRNA Lys3 . In panel A, cellular tRNA was hybridized with the 18-nucleotide DNA oligomer complementary to the 3Ј-18 nucle-FIG. 6. Electrophoretic detection of acylated and deacylated tRNA Lys3 . Cellular RNA was isolated, and fractions containing equal concentrations of ␤-actin mRNA were electrophoresed under acidic conditions as described under "Results." Northern blots of the cellular RNA were hybridized with either the 3Ј-terminal DNA probe, which hybridizes to all forms of tRNA Lys3 (A), or with anticodon probes specific for each mutant tRNA Lys3 (B-E). The first lane in each panel (lanes 1, 8, 11, 14, and 17) represents cellular RNA that was first exposed to alkaline pH to deacylate the tRNA (see "Experimental Procedures").  otides of tRNA Lys3 , whereas in panels B-E, the cellular tRNA was hybridized with the anticodon probes specific for different tRNA Lys3 mutants. Lane 1 in panel A shows the mobility of wild-type tRNA Lys3 deacylated in vitro at alkaline pH. As reported previously (23), in cells either transfected with the wildtype tRNA Lys3 gene (lane 2) or not transfected with any tRNA Lys3 gene (lane 3), the tRNA Lys3 detected is entirely in the aminoacylated form. As also shown in panel A, a majority of the total tRNA Lys3 is aminoacylated in cells transfected with genes encoding tRNA CGU Lys3 (lane 5) and tRNA SGU Lys3 (lane 6). In contrast, a larger proportion of total tRNA Lys3 is in the deacylated form in cells transfected with genes encoding tRNA CGA Lys3 (lane 4) and tRNA SGA Lys3 (lane 7). Although the data in Fig. 6A (lanes 4 -7) suggest that the tRNA Lys3 anticodon mutants are defective in in vivo aminoacylation, the total tRNA Lys3 probed in these experiments consists of both endogenous wild-type tRNA Lys3 and exogenous wild-type or mutant tRNA Lys3 . Thus to more directly probe the capability of the tRNA Lys3 mutants to be aminoacylated, we used anticodon DNA probes specific to the different tRNAs ( Fig. 6, panels B-E). Lanes 8,11,14, and 17 represent mutant tRNA Lys3 samples that have been deacylated in vitro at alkaline pH, whereas lanes 9, 12, 15, and 18 contain cellular RNA from cells transfected with HIV-1 proviral DNA, as well as the various tRNA genes. To demonstrate the hybridization specificity of the anticodon probes, we also probed cellular RNA from cells transfected only with HIV-1 proviral DNA (lanes 10, 13, 16, and 19). Based on the data presented in these panels, we conclude that tRNA SGU Lys3 (lane 9) and tRNA CGU Lys3 (lane 12) are aminoacylated to a significantly greater extent in vivo than tRNA CGA Lys3 (lane 15) and tRNA SGA Lys3 (lane 18). The latter two variants are present exclusively in the uncharged state. The percentage of wild-type or mutant tRNA Lys3 present in the aminoacylated state is shown graphically in panel F. Fig. 7 shows a plot of the relative tRNA Lys3 incorporation into HIV-1 for each of the anticodon variants tested in this work versus the percentage of aminoacylated tRNA detected in vivo. The linear correlation is striking and shows that the ability of tRNA Lys3 to be incorporated into HIV-1 is closely correlated with its ability to be aminoacylated. The efficiency of aminoacylation by aminoacyl-tRNA synthetases may be described by an overall specificity constant (k cat /K m ), which contains a catalytic rate constant (k cat ), as well as a binding parameter (K m ). Previous studies indicated that the U35G anticodon mutation in human tRNA Lys3 abolished in vitro aminoacylation (Ͼ3000-fold decrease in k cat /K m ) and that this change affected both LysRS binding and catalysis (12). Despite this dramatic effect on the ability to be aminoacylated with lysine in vitro, we report here that mutants with changes only in U35 (tRNA SGU Lys3 ) or in both S34 and U35 (tRNA CGU Lys3 ) are still aminoacylated in vivo to significant levels (Fig. 6). These mutations change the anticodon to the sequence found in tRNA Thr . The anticodon is a major recognition element for Escherichia coli threonyl-tRNA synthetase in vitro (24). Moreover, a change in the anticodon can switch the identity of E. coli elongator methionine tRNA (tRNA Metm ) from methionine to threonine, showing that it is also a major identity determinant in vivo (25). Assuming that human threonyl-tRNA synthetase maintains this strong anticodon recognition, it may be that the XGU anticodon variants tested here are at least partially charged with threonine in vivo.

DISCUSSION
Based on in vitro studies, a single U36A change in human tRNA Lys3 affects primarily the k cat parameter with only a 3-fold increase in K m and an overall 182-fold decrease in catalytic efficiency. Coupling the U36A mutation with a U35G change abolished in vivo aminoacylation and packaging into HIV-1 (Fig. 7). Both the UGA and the CGA anticodons correspond to sequences normally found in tRNA Ser isoacceptors. It has been shown that human seryl-tRNA synthetase does not use the anticodon as a recognition element but instead requires a long variable extra arm (26,27). Thus, it is unlikely that human tRNA SGA Lys3 or tRNA CGA Lys3 would be aminoacylated by serine in vivo, consistent with our observations that in vivo aminoacylation is abolished in these variants.
In this work, the data indicate that the anticodon of human tRNA Lys3 is a major determinant for packaging into HIV-1 (Fig.  7) with the caveat that we do not yet know whether altering the anticodon sequence affects modifications elsewhere in the tRNA molecule. However, a clear correlation between aminoacylation and packaging has been established (Fig. 7). Based on these data, we cannot determine whether aminoacylation itself is a prerequisite for packaging or whether the major factor is LysRS binding. To address this question, the identity of the amino acid attached to the mutant tRNAs that are incorporated to significant extents (i.e. tRNA SGU Lys3 and tRNA CGU Lys3 ) must be determined, and these experiments are planned.
The data presented in this work support a model in which the tRNA Lys3 /LysRS interaction is important for tRNA Lys3 incorporation into viruses. We also show that the anticodon is a major determinant for packaging. However, the anticodon sequence has also been shown to contribute to the in vitro binding of mature reverse transcriptase in studies using either native tRNA Lys3 (28) or unmodified tRNA Lys3 transcripts (29,30). Other data suggest that RT sequences in the precursor Gag-Pol polyprotein interact with tRNA Lys3 during its incorporation into virions (6,9). Thus, an anticodon mutation might also weaken this tRNA Lys3 /Gag-Pol interaction. However, there is no clear evidence to date that the interaction between RT sequences in Gag-Pol and the anticodon of tRNA Lys3 is involved in its packaging into HIV-1, and in fact, mutations in RT that prevent the enzyme from interacting with the tRNA Lys3 antiodon in vitro (31) were reported to have no effect on viral packaging of tRNA Lys3 (9).