Effect of Lysyl-tRNA Synthetase on the Maturation of HIV-1 Reverse Transcriptase.

In human immunodeficiency virus-1 (HIV-1), reverse transcriptase (RT) is encoded as a 66 kDa protein, p66, in the Gag-Pol polyprotein. This protein is proteolytically cleaved by HIV-1 protease (PR) to finally generate a mature RT that is a heterodimer, composed of a p66 subunit and a p66-derived 51 kDa subunit, p51. In our prior work, we demonstrated that tRNALys3 binding to p66/p66 facilitates efficient cleavage of p66 to p51 by PR. However, tRNALys3 is known to be recruited to the virus by forming a complex with lysyl-tRNA synthetase (LysRS). Herein, we tested whether LysRS can have an effect on RT maturation in vitro. Importantly, our data show no significant differences in RT maturation in the presence of LysRS. Furthermore, no apparent p66/66 interaction with LysRS was observed. Although PR cleaved LysRS, it did not immediately release tRNALys3 from LysRS. Thus, we conclude that a free fraction of tRNALys3, which is in equilibrium with a LysRS-bound form, interacts with p66/p66 without any additional mechanism involving release of tRNALys3 from LysRS. Given that only transient tRNALys3-p66/p66 interaction is needed for efficient RT maturation, a small amount of free tRNA may be sufficient for this process. These studies reveal molecular level insights into RT maturation and will be useful for the design of cellular/viral experiments to better understand the role of tRNA in HIV-1 replication.

Introduction

Maturation of viral polyproteins is essential for the replication and production of viruses. In human immunodeficiency virus-1 (HIV-1), reverse transcriptase (RT) is expressed as a part of the viral Gag-Pol polyprotein, which is cleaved by HIV-1 protease (PR) to form a mature RT heterodimer composed of 66 (p66) and 51 kDa (p51) subunits (p66/p51).[1,2] The p51 subunit is generated upon removal of most of the ribonuclease H (RNH) domain from p66.[3−5] The maturation of RT from the Gag-Pol polyprotein is complex and varies depending on the environment and experimental conditions.[6−15]

At a molecular level, studies on RT maturation and its dependence on the protein concentration and dimerization have shown that efficient RT maturation requires p66 homodimer formation, p66/p66, prior to PR processing at the p51-RNH site.[8,16,17] We recently showed that such maturation of RT is significantly enhanced in the presence of tRNALys3in vitro and at the cellular level.[18,19] Because the p51-RNH cleavage site in p66/p66 is buried within the RNH domain in the mature RT, the presence of such an enhancer explains how the RT precursor can cross the energy barrier for p66/p51 formation. tRNALys3 is known to be recruited to the virus in complex with lysyl–tRNA synthetase (LysRS), a 68 kDa protein,[20−23] and we have shown that knockdown of LysRS results in the accumulation of p66, reducing the production of p66/p51 at the cellular level.[19] Because tRNALys3 serves as a primer to initiate reverse transcription and is known to interact with mature RT in complex with viral RNA,[2] we considered it important to further characterize whether LysRS might play a direct role in our model of tRNALys3-facilitated maturation of RT or not.

LysRS is a 68 kDa protein, exists primarily as a homodimer in solution, and is structured from residue 70 to 576.[24] LysRS interacts with tRNALys3 with a dissociation constant (KD) of 60–900 nM[25−27] and with a Michaelis constant (Km) for aminoacylation of ∼100 μM.[25] Aminoacylation does not affect incorporation of tRNALys3 to the virus.[23] The KD of tRNALys3 dissociation from LysRS is compatible with the apparent KD of tRNALys3 dissociation from p66/p66, ∼70 nM, and with that of the p66 homodimer dissociation, ∼4 μM.[18,28,29] In the virus, LysRS has been proposed to interact with Gag and Gag-Pol.[30−36] Specifically, LysRS has been suggested to interact with the connection and RNase H regions in the RT domain of the Gag-Pol polyprotein.[32] However, whether LysRS affects RT maturation is unknown.

Here, we present data from in vitro RT maturation assays, size exclusion chromatography (SEC), SEC-multiangle light scattering (SEC-MALS), and gel-mobility shift assays to investigate the effect of LysRS on in vitro RT maturation. Briefly, a LysRS–p66/p66 interaction was not detected, and RT maturation was not suppressed by LysRS in vitro. We conclude that a free tRNALys3 fraction in the LysRS–tRNALys3 equilibrium is sufficient to enhance RT maturation.

Results

LysRS–tRNALys3 Interaction

Prior to studying the LysRS effect on RT maturation, we first employed SEC-MALS to characterize the interaction between purified LysRS and tRNALys3. LysRS at two different injection concentrations, 10 and 40 μM (Figure A,B), gave maximum molecular masses of 127.3 ± 2.4 and 128.4 ± 3.4 (kg/mol), respectively, both similar to the expected molecular mass of a LysRS homodimer, 136 kDa. Each mass value, as listed in Table , is an average of those obtained within the selected region of the respective elution peak (i.e., symbols in Figure ). This analysis provides qualitative estimates of the species observed in each elution profile.

Figure 1

SEC-MALS of (A) 10 μM LysRS, (B) 40 μM LysRS, (C) 40 μM LysRS and 5 μM tRNALys3, and (D) 40 μM LysRS and 20 μM tRNALys3. The concentration is the injection concentration. Each graph indicates the UV 280 nm elution profile (solid line with the left vertical scale) and estimated molecular mass from MALS (symbols with the right vertical scale). In the protein and RNA mixed samples, the entire mass is shown by black symbols, and each mass component, either protein or RNA, is shown by gray symbols (see Table ).

Table 1

Deconvolutiona of the LysRS–tRNALys3 SEC-MALS Data Obtained in Figure D

elution volume (mL)	entire mass (kg/mol)	protein mass (kg/mol)	tRNA mass (kg/mol)	estimated species
∼11.2	202.5	140.5	62.1	LysRS dimer–
	(±11.6)	(± 9.7)	(± 2.0)	2 tRNA
∼12.0	146.9	116.4	30.5	LysRS dimer–
	(±1.0)	(±3.4)	(±3.9)	1tRNA
∼12.8	129.4	123.8	b	LysRS dimer
	(±2.7)	(±2.6)b
∼14.0	27.9	c	27.9	tRNA
	(±1.1)		(±1.1)c

Data were analyzed assuming a protein–nucleic acid complex. Note, tRNA and LysRS homodimer molecular weights are 25 and 136 kDa, respectively.

Fit indicated that tRNA did not exist.

Fit indicated that protein did not exist. The entire mass was attributed only to tRNA.

Next, SEC-MALS experiments were performed by injecting 40 μM LysRS mixed with 5 μM tRNALys3 (Figure C). The elution profile exhibited a fraction of a tRNALys3–LysRS complex at 1:1 tRNALys3–LysRS stoichiometry, with molecular masses at 26.9 ± 2.8 and 123.9 ± 1.5 (kg/mol), respectively, which is consistent with their known molecular weights, 25 and 136 kDa, respectively. When a mixture of 40 μM LysRS with 20 μM tRNALys3 was injected, two additional fractions were observed. One, the earliest elution peak, was from the tRNALys3–LysRS complex at 2:1 tRNALys3–LysRS stoichiometry, and the other, the last elution peak, was a free tRNALys3 fraction (Figure D and Table ). We verified these observations by recording additional SEC experiments in which the elution was detected using UV at 254 and 280 nm wavelengths (Figure S1). Based on the known KD of the tRNALys3–LysRS complex, 0.93 ± 0.32 μM,[27] observation of a free tRNALys3 fraction is reasonable at ∼μM protein and RNA concentrations (assuming 10-fold dilution in the column). Note, because LysRS has only two Trp residues, its extinction coefficient at UV 280 nm is 43,780 M–1 cm–1 and is approximately 10 times smaller than that of the tRNALys3, ∼350,000 M–1 cm–1. Thus, although the intensity of the free tRNALys3 in the UV elution profile, as shown in Figure D, is much higher than those of LysRS–tRNALys3 complexes, the actual population of the free tRNALys3 is approximately equivalent to the bound form.

Effect of LysRS on In Vitro RT Maturation

We next performed a set of in vitro RT maturation assays with or without LysRS. Production of p51 subunit from p66 was monitored by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) following incubation of purified p66 with HIV-1 PR for varying amounts of time.[18] Generation of equivalent amounts of p66 and p51 is indicative of p66/p51 heterodimer production. As shown previously,[18] incubation of p66 alone with PR did not generate significant p66/p51 heterodimer (Figure A) while incubation in the presence of tRNALys3 did (Figure B). A similar result was obtained in the presence of LysRS (Figure C,D): p51 band generation was more significant in the presence of tRNALys3 than in its absence. The enhancement of p66/p51 production in the presence of tRNALys3 is not due to the PR–tRNALys3 interaction, as we demonstrated previously.[18] Whether LysRS was present or not, 30–40% of p66 was cleaved to p66/p51 at the 20 min time-point in the presence of tRNALys3 (Figure E). Quantification of p66/p51 band intensity at the 20 min time-point suggested slightly higher p66/p51 production in the presence of LysRS, compared to the absence of LysRS. However, the reaction rates (measurements taken at initial time points) were not substantially different between these two conditions (Figure E). Thus, we conclude that LysRS does not have an impact on p66/p51 production neither alone nor in the presence of tRNALys3.

Figure 2

RT maturation assay: time-dependent proteolytic cleavage of p66 by HIV-1 PR monitored by SDS-PAGE. Cleavage experiments were conducted in the (A,B) absence or (C,D) presence of LysRS, and/or (B,D) presence or (A,C) absence of tRNALys3. Panel (E) indicates p66/p51 production, calculated based on the p51 band intensities of three repeated data sets (additional data in Figure S2). The concentrations of LysRS, p66, tRNALys3, and PR were 4, 8, 4, and 1 μM, respectively.

LysRS Does Not Interact with RT

To understand the underlying mechanism of the above-mentioned RT maturation results, we conducted molecular-interaction studies. We performed SEC experiments with LysRS alone (Figure A), p66 alone (Figure B), and LysRS mixed with p66 (Figure C). The elution profile of the mixed sample was explained well by superimposing it with those of the two isolated proteins (Figure C), demonstrating that LysRS and p66 do not interact in a substantial way when present at μM protein concentrations. We next performed SEC-MALS experiments of LysRS and p66 in the presence of tRNALys3. We used SEC-MALS, instead of SEC, to enable identification of each elution peak component as much as possible. A mixture containing 5 μM tRNALys3, 40 μM LysRS, and 40 μM p66 generated the following SEC-MALS elution peaks: (i) free p66, (ii) superimposed peaks of p66–tRNALys3 complex and of p66/p66, (iii) superimposed peaks presumably of p66/p66–tRNALys3 and of LysRS (could not be deconvoluted because of three different species), and (iv) the LysRS–tRNALys3 complex (Figure D). The same experiment with a higher tRNALys3 concentration generated a molecular mass that corresponded to a LysRS homodimer bound to two tRNALys3 (Figure E). The fraction at ∼11 mL showed a gradient of molecular masses, including a high molecular mass component, perhaps tetramer of LysRS. Importantly, an explicit elution peak indicating the presence of the p66/p66–tRNALys3–LysRS complex was not observed (Figure E). Thus, we conclude that both in the presence and absence of tRNALys3, direct interaction of LysRS dimer and p66 (or p66/p66) is not detectable at the μM concentration range.

Figure 3

(A–C) SEC and (D,E) SEC-MALS elution profiles showing no interaction between LysRS and p66 in the (A–C) absence and (D,E) presence of tRNALys3: (A) 30 μM LysRS, (B) 30 μM p66, (C) 30 μM LysRS and 30 μM p66, (D) 40 μM LysRS, 40 μM p66, and 5 μM tRNALys3, and (E) 40 μM LysRS, 40 μM p66, and 20 μM tRNALys3. In (C), dotted and dashed lines are taken from (A,B). In (D,E), the entire mass is shown by black symbols, individual mass component, either protein or RNA, is shown by gray symbols, and those that could not be deconvoluted are shown with “X.” In (D), mixed species are annotated with gray characters (see the details, as listed in Table ).

Table 2

Deconvolutiona of the LysRS–tRNALys3 SEC-MALS Data Obtained in Figure E

elution volume (mL)	entire mass (kg/mol)	protein mass (kg/mol)	tRNA mass (kg/mol)	estimated species
∼11.2	270.5	197.6	62.1	LysRS dimer–2 tRNA and a higher molecular mass
	(±14.6)	(±12.3)	(±2.0)
∼12.0	160.7	128.8	30.5	LysRSdimer–1 tRNA and p66/p66–tRNA
	(±2.0)	(±1.1)	(±3.9)
∼12.8	137–152b	b	b	X (LysRS dimer and p66/p66–tRNA)
∼13.6	89.3	66.0	23.3	p66–tRNA
	(±14.9)	(±7.2)	(±2.4)
∼14.6	69.5	59.6	c	p66
	(±1.2)	(±0.5)

Data were analyzed assuming a protein–nucleic acid complex. Note that tRNA, LysRS homodimer, p66 molecular weights are 25, 136, and 66 kDa, respectively.

Fit was not obtained presumably due to the mixture of LysRS and p66/p66–tRNA species.

Fit indicated that tRNA did not exist.

To verify the above observation, we studied tRNALys3 interactions with each individual protein and then examined competition between LysRS and p66 for tRNALys3 binding using a gel-based shift mobility assay (Figures and S3). In the experiment for binary complex formation with tRNALys3 and LysRS or p66, the tRNALys3 band clearly shifted with increasing concentrations of proteins (Figure A,B). Consistent with a previous observation,[26] LysRS was observed to bind one or two tRNALys3 molecules: at lower LysRS concentrations (1–2 μM), two tRNALys3 molecules bound the LysRS dimer, whereas at higher LysRS concentrations (4–8 μM), LysRS dimer bound one tRNALys3 (Figure A). Note, because LysRS dimer bound to two tRNA is more negatively charged than LysRS bound to one tRNA, it migrates faster in a native gel.[26] In the tRNALys3–p66 interaction study, the well-defined free tRNALys3 band was present at low p66 concentrations and shifted to a diffuse band in the presence of higher p66 concentrations (>2 μM p66), reflecting generation of p66–tRNALys3 and p66/p66–tRNALys3 complexes (Figure B), which is in agreement with our previous observation.[18]

Figure 4

Competitive gel-shift assay to characterize molecular interactions of (A,B) 0.5 μM tRNALys3 with (A) LysRS or (B) p66 at different protein concentrations, (C) 0.5 μM tRNALys3 and 8 μM LysRS with escalating p66 concentrations, or (D) 0.5 μM tRNALys3 and 8 μM p66 with LysRS at escalating protein concentrations. Binding experiments were performed in 25 mM Bis–Tris buffer, containing 100 mM NaCl, at pH 7.0. Reactions were monitored by native polyacrylamide gel, and the gel was stained with SYBR gold to detect tRNALys3. The repeated gel (Figure S3), run with similar conditions, reproduced the data.

In competition experiments, when p66 was added to a preincubated solution of LysRS and tRNALys3, a fraction of the LysRS–tRNALys3 complex decreased while fractions of p66–tRNALys3 and p66/p66–tRNALys3 increased (Figure C and S3C). When LysRS was added to the preincubated solution of p66 and tRNALys3, a LysRS–tRNALys3 complex band was observed (Figures D and S3D). These data indicate that binding of tRNALys3 to p66 or p66/p66 is with an apparent affinity similar to that of tRNALys3 binding to LysRS, and neither is extremely high.

Effect of PR Processing of LysRS

Interestingly, PR cleaved LysRS as well as p66 (Figure D). This raised the possibility that LysRS cleavage releases tRNALys3 from the LysRS–tRNALys3 complex, preventing our observation of any effect of LysRS–tRNALys3 on RT maturation. To test this hypothesis, we first investigated whether cleavage of LysRS releases tRNALys3 from the LysRS–tRNALys3 complex by monitoring (i) the mobility of tRNALys3 in native gels of LysRS–tRNALys3 in the presence of PR (Figure A) and (ii) the formation of LysRS cleavage products in denatured SDS-PAGE of LysRS in the presence of PR (Figure B). The samples for both assays were obtained from the same time course experiments. Our results demonstrate that tRNALys3 remains bound to LysRS even after PR cleaves the protein (Figures A, 5B, and S4).

Figure 5

Processing of LysRS by HIV-1 PR in the presence of tRNALys3, (A) monitored by native-PAGE, (B) monitored by SDS-PAGE, and (C) in the absence or presence of tRNALys3, monitored by SDS-PAGE. Panels (A,B) were obtained from a single set of experiments, performed using 1 μM PR, 8 μM LysRS, and 2 μM tRNALys3. Panel (C) in the presence of tRNALys3 corresponds to a repeat of (A) but using 4 μM tRNALys3 and the control for the data in the absence of tRNALys3. Repeated data for (A,B) are also shown in Figure S4. The full images and the repeated data for (C) are shown in Figure S7.

To identify the cleavage site, we performed mass spectrometry experiments. The data indicate that two fragments of LysRS were generated: (i) a fragment created by cleavage at an N-terminal D12-G13 site and (ii) a fragment generated by cleavage at a C-terminal A536-L537 site (Table and Figure S5). The cleavage at D12-G13 was most likely chemically induced by formic acid, which is known to be generated during the sample preparation for the mass spectrometry experiments, and D12-G13 is not a PR-recognized site.[37−39] Indeed, this peptide cleavage was observed in the samples of LysRS alone as well as those treated with PR (Table and Figure S5). On the other hand, the C-terminal A536-L537 site is a peptide sequence that aspartic proteases are predicted to cleave.[38,39] Thus, we conclude that PR cleaves LysRS at A536-L537 to generate a 61.5 kDa product from the 68.2 kDa full-length LysRS. The A536-L537 site is not located at a random coil region but is in a short α-helix, near the loop region (Figures S5A and S6, discussed later). Importantly, the PR-cleaved C-terminal β-strand is located within a seven-stranded β-sheet that is sandwiched between α-helices. This may be one of the reasons why LysRS remained bound to tRNALys3 (Figure A) even after PR cleavage at the C-terminal β-strand (Figure B).

Table 3

Mass Spectrometry Data of LysRS in the Absence or Presence of PR, Examined with/without DTTa

protein condition	experimental mass (Da)	theoretical mass	estimated amino acid sequence
LysRS	68,162.3	68,162.17	LysRS full-lengthb
	66,892.3	66,891.70	LysRS from G13 to the C-terminal end
LysRS + DTT	68,162.4	68,162.17	LysRS full-lengthb
	66,892.0	66,891.70	LysRS from G13 to the C-terminal end
LysRS + PR	68,161.4	68,162.17	LysRS full-lengthb
	66,891.4c	66,891.70	LysRS from G13 to the C-terminal end
	61,521.5	61,521.61	LysRS from the N-terminal end to A536
LysRS + DTT + PR	68,161.9	68,162.17	LysRS full-lengthb
	66,891.6c	66,891.70	LysRS from G13 to the C-terminal end
	61,521.3	61,521.61	LysRS from the N-terminal end to A536

The convoluted mass profiles are shown in Figure S5.

This is the molecular mass with the N-terminal Gly and the C-terminal Gly.

This is due to chemical cleavage by formic acid[37] and is seen in the samples without PR too.

Because tRNALys3 binding to LysRS could potentially delay fragmentation of LysRS by PR, as observed from the intensity plot (Figure S7C), we next compared the time course of LysRS cleavage by PR in the absence or presence of tRNALys3. Interestingly, cleavage of LysRS by PR was similar both in the absence and presence of tRNALys3 (Figure C). Thus, the surrounding β-sheet and α-helices support the cleaved β-strand to fold the protein, and tRNALys3 binding is not a major factor in maintaining the fold of the cleaved LysRS. Note, the fraction of full-length and truncated forms of LysRS decreased with time in the presence of PR, indicating that PR can digest other sites in LysRS but at a rate slower than the A536-L537 site. Overall, a drastic dissociation of tRNALys3 from the LysRS–tRNALys3 complex did not occur upon PR cleavage of LysRS.

Discussion

We previously found that the interaction of tRNALys3 with p66/p66 enhances specific cleavage by PR.[18,19] However, tRNALys3 is not recruited to the virus by itself and is known to be packaged in a complex with LysRS.[20,21,40,41] Thus, we tested whether LysRS affects RT maturation in vitro. Our observation of the LysRS interaction with tRNALys3 by SEC-MALS (Figure ) and the gel shift (Figure A) are consistent with the previously published data.[26] Interestingly, we did not observe significant delay of RT maturation in the presence of LysRS (Figure ). Consistent with this, we also did not observe p66/p66–LysRS interactions (Figure ), indicating that a ternary complex of LysRS–tRNALys3–p66/p66 is not formed in the conditions and concentration ranges that we examined. Similarly, we did not observe any effect of LysRS cleavage on tRNALys3 release (Figure ), indicating that LysRS digestion by PR is not a major mechanism to release tRNALys3 from LysRS. Instead, the apparent affinities of tRNALys3 to p66 and LysRS were similar to each other (Figure ). Taken together, because of a similar moderate, ∼μM, apparent dissociation constants of LysRS–tRNALys3 and p66/p66–tRNALys3, a fraction of the free tRNALys3 may bind p66/p66 and enhance RT maturation (Figure ). Here, “apparent” includes the effect of the monomer–dimer equilibrium of p66. Such a conclusion is consistent with a model in which tRNALys3 transiently interacts with p66/p66 to promote maturation, rather than only stabilizing the produced p66/p51 form.

Figure 6

Cartoon indicating the equilibrium between p66 (blue spheres) and tRNALys3 (red triangles), and between LysRS (pink oval) and tRNALys3. Based on our observation, we propose that a fraction of free tRNALys3 in the equilibrium of LysRS–tRNALys3 might be sufficient for RT maturation.

Our observations also provide information on the biophysical characteristics of human LysRS. LysRS is known to form a dimer in solution. Although elution peaks of both dimer and tetramer fractions have been shown in the previous gel-filtration chromatography studies (at a 1.7 μM loading concentration),[24] we did not see a significant tetramer fraction at a 30 μM loading concentration. Instead, the molecular masses estimated from MALS, 127.3 ± 2.4 and 128.4 ± 3.4 (kg/mol) for 10 and 40 μM injections, respectively, were ∼5% smaller than the expected molecular mass of a LysRS homodimer, 136 kDa (Figure ). These molecular masses may indicate that the protein is in a monomer–dimer equilibrium, rather than a dimer–tetramer equilibrium. Our observation is consistent with the fact that light scattering indicates a hydrodynamic diameter for LysRS of ∼9.5 nm, which is smaller than the diameter estimated from a crystal structure of the dimer, 10.2 nm.[35]

The C-terminal A536-L537 cleavage site in LysRS is not located in a random coil region but in a short α-helix (residues 530–539) near the loop region (Figure S6A). This site is buried by the N-terminal domain (residues 1–215) of the other subunit in the homodimer crystal structure (Figure S6B).[24] Because PR is not expected to be able to access such a buried site, our observation suggests that the N-terminal domain undergoes domain motion. This is consistent with the previous hydrogen–deuterium exchange and small-angle X-ray scattering study, showing flexibility of the N-terminal domain.[26] Importantly, the short α-helix (residues 530–539) directly interacts with the side chain of S207, the phosphorylation of which is critical for the release of LysRS from the multi-tRNA synthetase complex and the conformational change for it.[26,42] We also suggest that the helix itself may be more mobile, or shorter, to allow PR interaction with the A536-L537 site in solution.

With regard to the significance and consistency of our findings relative to previously described HIV-1 replication studies, we consider the following. Kleiman’s group previously showed truncation of LysRS in the virus from ∼70 to ∼62 kDa.[20] This cleavage is likely the same C-terminal truncation that we observed from 68.2 to 61.5 kDa (determined by mass spectrometry). Importantly, our data indicate that LysRS can carry tRNALys3 even after the cleavage. Thus, the truncation does not interfere with LysRS recruitment of tRNALys3 to the virus.[21−23] Our observation is also consistent with what Kleiman found regarding the LysRS truncation in their virus experiments.[20] Based on our data, because the binding affinities of tRNALys3 with LysRS and p66 are moderate and these proteins do not directly interact with each other, p66/p66 could be matured with assistance from free tRNALys3 in a trimolecule equilibrium (Figure ).

While only two tRNALys3 molecules are known to be annealed to the viral RNA,[43] 20–25 tRNALys3 molecules, including the tRNALys3 isoacceptors, are in the virus.[21,44,45] Such a difference in the total number of tRNALys3 molecules present in the virus and the need for many tRNALys3 molecules may be explained if one considers the population of the free tRNALys3 molecules in the LysRS–tRNALys3 equilibrium: LysRS in the virus may store tRNALys3 and release it for use during RT maturation and formation of the reverse transcription initiation complex. However, other molecules are also known to interact with tRNALys3 in the virus.[46,47] Further studies are needed to address the molecular interactions that occur in the virus.

Our study clarified the relative interactions of tRNALys3 with LysRS and p66/p66 and the effects of PR cleavage on these proteins. Interestingly, we found that LysRS does not directly interact with p66/66 and does not greatly influence the tRNALys3 effect on RT maturation. PR cleavage of LysRS does not release tRNALys3 either. Given that the apparent dissociation constants of LysRS–tRNALys3 and p66/p66–tRNALys3 are moderate, ∼μM, and similar to each other, we propose that free tRNALys3 is present in an amount sufficient to enhance RT maturation, further supporting our proposed RT maturation model.[18,19] Our results are also significant in that they support previous structural studies,[25−27] as discussed above.

Materials and Methods

Protein and tRNA Expression and Purification

We purchased the human KARS-1 clone, UniProtKB—Q15046 (SYK_HUMAN), in pD441-NHT vector (ATUM, Newark, CA) and expressed the protein in Escherichia coli BL21 (DE3). KARS is the gene name of LysRS; we use LysRS to refer to the protein in order to be consistent with publications from the Musier-Forsyth and Kleiman groups.[20−23] Protein was purified by a HisTrap HP column (GE Healthcare, USA) using the published protocol,[48] followed by overnight cleavage with TEV protease. Processed LysRS protein was collected in the flow-through after a second pass through a HisTrap HP column and finally purified using the HiTrap Q HP column (GE Healthcare) in 25 mM HEPES buffer, pH 7.5, containing 10% glycerol and 0.02% azide, with 1 M NaCl in elution buffer. Purified protein was buffer exchanged to 25 mM Bis–Tris buffer at pH 7.0, containing 250 mM NaCl and 50% glycerol, and stored at −80 °C. Throughout the process of protein purification, we added fresh 5 mM β-mercaptoethanol to the buffers immediately before their use. The protein concentration was determined by measuring absorbance at 280 nm with an extinction coefficient of 43,780 M–1 cm–1. Both gene sequencing and mass spectrometry consistently confirmed that the protein contained an N-terminal Gly at the TEV protease cleavage site and one Gly at the C-terminus.

HIV-1 p66 protein was prepared, as described previously.[18] tRNALys3 was prepared by in vitro transcription of its DNA using NTPs and T7 RNA polymerase, as described previously.[19] Note, tRNALys3 does not have amino-acyl modification. HIV-1 protease was expressed and purified from the inclusion body using gel-filtration in denaturing conditions and reverse-phase chromatography.[49,50]

SEC-MALS Experiments

SEC-MALS experiments were performed for the following samples: 10 μM LysRS; 40 μM LysRS; 40 μM LysRS with 5 μM tRNALys3; 40 μM LysRS with 20 μM tRNALys3; 40 μM LysRS and 40 μM p66 with 5 μM tRNALys3; 40 μM LysRS and 40 μM p66 with 20 μM tRNALys3. All SEC-MALS experiments were performed by injecting a 100 μL sample, pre-equilibrated and eluted with 25 mM Bis–Tris, pH 7.0, 100 mM NaCl, 0.02% sodium azide buffer at a flow rate of 0.5 mL/min at room temperature using an analytical Superdex 200 Increase 10/300 column (GE Healthcare) with in-line MALS, refractive index (Wyatt Technology, Inc., Santa Barbara, CA), and UV/vis (Waters Corporation, USA) detectors.

The molecular masses of the eluted protein species were determined using the ASTRA V.7.1.2 program (Wyatt Technologies). After the SEC-MALS elution profiles were obtained, we selected each elution region for the mass ASTRA calculation and determined the average and standard deviation of the mass for each selected region. For protein–RNA complex analysis, we implemented the protein conjugate analysis method within the ASTRA program (Wyatt Technologies). Briefly, the analysis requires that the individual constituents within a complex have unique UV extinction coefficients and/or refractive index increments. After selecting peaks of interest, we used the specific refractive index increment for protein, 0.185 mL/g, and that for polynucleotides, 0.17 mL/g; as well as the specific UV extinction coefficients for tRNALys3 and LysRS or p66, respectively.

SEC Experiments

SEC experiments were performed for 30 μM LysRS, 30 μM p66/p66, and their mixture using a 24 mL analytical Superdex 200 Increase 10/300 GL column (GE Healthcare), equilibrated with 25 mM Bis–Tris buffer, pH 7.0, containing 100 mM NaCl with 0.02% sodium azide, at a flow rate of 0.5 mL/min. In each experiment, a 50 μL of the sample was injected, and protein elution was monitored by UV absorbance at 254 and 280 nm.

Gel-Shift Mobility Assays

Gel-shift assays to characterize the tRNALys3 interaction with LysRS and p66 were performed by preincubating each protein at different concentrations, with 0.5 μM tRNALys3 in 25 mM Bis–Tris, pH 7.0, 100 mM NaCl buffer at room temperature. In a subset of experiments, 8 μM p66 or 8 μM LysRS was also included during the preincubation period of tRNALys3 with LysRS or p66, respectively. The samples were resolved by native-PAGE using precast 4–15% Tris-glycine gels (Bio-Rad) with 0.5× TBE buffer. Glycerol was added to the reaction mixtures prior to loading to obtain a final 10% glycerol concentration. Gels were stained with SYBR Safe DNA Gel Stain (Invitrogen, USA) and analyzed with an Amersham Imager 600 (GE Healthcare Life Sciences).

In Vitro RT Maturation Experiments

The time course of proteolytic processing of 8 μM p66 by 1 μM PR, in the absence or presence of 4 μM of tRNALys3, was carried out in 20 mM sodium acetate, pH 5.2, at 37 °C. In parallel, the same experiments were conducted in the presence of 4 μM LysRS. Aliquots at different time points were collected, quenched by the addition of tricine sample loading buffer (Bio-Rad Laboratories, Berkeley, CA) and monitored by SDS-PAGE stained with Bio-Safe Coomassie (Bio-Rad), as described previously.[18] For quantification, these gel-based experiments were repeated at least three times.

Proteolysis of LysRS by PR

Two types of gel-based LysRS proteolysis experiments were performed. First, 8 μM LysRS in the presence of 2 μM tRNALys3 in 25 mM Bis–Tris, pH 7.0, 100 mM NaCl buffer was incubated at 37 °C with 1 μM HIV PR for different periods of time. The samples were then aliquoted and either (i) quenched with 100 μM protease inhibitor, darunavir, for detection of the LysRS–tRNALys3 interaction in native gels, or (ii) quenched by the addition of tricine sample loading buffer and denatured at 95 °C to detect LysRS cleavage by HIV-1 PR in SDS-gels. Second, to investigate the effect of tRNALys3 on PR-directed LysRS proteolysis, time course experiments were performed using 8 μM LysRS and 1 μM PR in the absence or presence of 4 μM tRNALys3. Cleavage was monitored using SDS-PAGE, as described earlier.

The proteolysis site within LysRS was assessed by mass-spectrometry. LysRS solution, at 10 μM, was incubated for 30 min in 25 mM Bis–Tris, pH 7.0, 100 mM NaCl buffer, either with or without dithiothreitol (DTT), or with or without 1 μM PR. In each, 2 μL of the sample solution was diluted to an 18 μL solution, containing 5% acetonitrile, 0.01% trifluoroacetic acid, and 0.1% formic acid, at pH 2.65 and injected into a Bruker Compact QTOF LC–MS/MS system (Bruker Daltonics Inc., Billerica, MA).