Photoaffinity Labeling by 4-Thiodideoxyuridine Triphosphate of the HIV-1 Reverse Transcriptase Active Site during Synthesis

The active site of HIV-1 reverse transcriptase (HIV-1 RT) was investigated by photoaffinity labeling based on catalytic competence. A stable ternary elongation complex was assembled containing enzyme, DNA template (RT20), DNA primer molecule (P12), and the necessary dNTPs (one of which was α-32P-labeled) needed for primer elongation. The photoaffinity probe 4-thiodideoxyuridine triphosphate was incorporated uniquely at the 3′ terminus of the 32P-labeled DNA product. Upon photolysis, the p66 subunit of a HIV-1 RT heterodimer (p66/p51) was uniquely cross-linked to the DNA product and subsequently digested by either trypsin or endoproteinase Lys-C. The labeled HIV-1 RT peptide was separated, purified, and finally subjected to Edman microsequencing. A unique radioactive hexapeptide (V276RQLCK281) was identified and sequenced. Our photoaffinity labeling results were positioned on the HIV-1 RT·DNA·Fab complex x-ray crystallography structure and compared with the suggested aspartic triad active site.

The enzyme reverse transcriptase (RT) 1 derived from the HIV-1 virus is a heterodimer composed of two subunits p66 and p51, which are derived from the same sequence.
A variety of experimental techniques have been directed at the elucidation of the active site and the mechanism of catalysis of the RT DNA polymerase activity, to assist in developing a strategy for treating HIV infection. Kinetic studies have established that an ordered sequential assembly of components forms a ternary complex, which then conducts a processive polymerization during the elongation phase (1)(2)(3)(4)(5)(6)(7). Genetic substitution experiments have shown that several single amino acid interchanges D110Q, D185H, or D186N in the p66 subunit produce an inactive HIV-1 RT enzyme (8,9). Studies involving specific amino acid derivatizations have suggested that Lys 263 (10) and Arg 277 (11) are critically involved at the active site. Photoaffinity labeling studies have yielded additional suggestions for the nucleotide binding site components: Lys 73 (12), residues 288 -307 (13), and residues 288 -423 (14). Several x-ray structures have been solved for the HIV-1 RT unliganded enzyme (Rogers et al. (15) reported a structure at 3.2 Å and Hsiou et al. (16) at 2.7 Å resolution), and for various ligands complexed with the holoenzyme; Kohlstaedt et al. (17) reported a nevirapine-RT enzyme structure at a resolution of 3.5 Å, and Jacobo-Molina et al.. (18) reported a RT⅐dsDNA⅐Fab x-ray structure at a resolution of 3.0 Å. Model building efforts based primarily on the RT⅐dsDNA⅐Fab x-ray structure have yielded a detailed mechanistic proposal for the HIV-1 RT enzyme (19). The model is consistent with the critical involvement of the Asp-triad residues and two Mg 2ϩ , as originally proposed by Steitz et al. (20).
The results reported in this paper utilized the photoaffinity probe S 4 -ddUTP to derivatize the active site of HIV-1 RT during productive synthesis involving a ternary complex. The chain terminating probe is located specifically at the 3Ј-OH end of the nascent product. The hexapeptide VRQLCK of the p66 subunit was the only target peptide that was detected. A general discussion of the possible mechanistic implications of these results is directed at making all the known topologic information compatible.

EXPERIMENTAL PROCEDURES
Materials-Recombinant HIV-1 reverse transcriptase purified from an Escherichia coli clone was kindly supplied by Dr. Christine Debouck and Dr. Jeffrey Culp from SmithKline Beecham Pharmaceuticals. The template and primer were a gift from Dr. Xiaolin Zhang at the Nucleic Acid Facility of the University of Pennsylvania Cancer Center. The sequence of the DNA templates RT19 and RT20 were 3Јd-[GCGCG-GGGCGCGGTGTGTA]-5Ј and 3Јd-[GCGCGGGGCGCGGTGTTGTA]-5Ј. The sequence of the DNA primer P12 was 5Јd-[CGCGCCCCGCGC]-3Ј. The nonradioactive deoxynucleotides (dATP, dCTP, and TTP) and Pronase E were purchased from Sigma. Radioactive nucleotide triphosphate [␣-32 P]dCTP (3,000 Ci/mmol), reflection autoradiography film, and reflection intensifying screen were purchased from NEN Life Science Products. The nucleoside 4-thiodideoxyuridine was prepared and kindly provided by Dr. Robert Coleman of Ohio State University. The corresponding triphosphate was prepared from the nucleoside according to the method of Ruth and Cheng (21). Exonuclease III was purchased from Amersham Life Science. Sequencing grade endoproteinase Lys-C was purchased from Promega. Sequencing grade modified trypsin was purchased from Boehringer Mannheim. HinP1I restriction endonuclease was purchased from New England Biolabs. All reagents for gel electrophoresis were purchased from Bio-Rad. Spectra/Por 6 molecular porous dialysis membrane (M r cut-off ϭ 1,000) was purchased from Spectrum. Mini ProBlott™ membranes was purchased from Applied Biosystems. The high intensity black lamp (model B-100A, long wave UV, which peaks at 365 nm) was purchased from Eastern Corp. Computer program Insight II was purchased from Biosym Technologies.
Photoaffinity Labeling of HIV-1 RT-The standard reaction mixture (100 l) for photoaffinity labeling was: 20 mM Tris-HCl, pH 7.9, 6 mM MgCl 2 , 1.0 M HIV-1 RT heterodimer (M r ϭ 117,000), 10 M DNA template, and 10 M DNA primer. The nonradioactive substrates dATP, dCTP, and photoprobe (S 4 -ddUTP) were each added to a final concentration of 100 M. The radioactive substrate [␣-32 P]dCTP (3,000 Ci/ mmol) was added to achieve a specific activity within the range of (1,000ϳ10,000 cpm/pmol). The template primer complex was prepared by heating a solution of 20 M DNA template and 20 M DNA primer in the reaction buffer to 100°C for 3 min. Upon cooling (1 h), the HIV-1 RT was added and the mixture was incubated at 37°C for 5 min. Immediately upon adding the substrates, the reaction mixture was placed in a depression well of an aluminum foil-covered, temperature-regulated aluminum block. The aluminum block was covered with an inverted Petri dish to prevent extensive evaporation of the reaction mixtures. The UV lower wavelength cut-off of a Pyrex Petri dish is 290 nm. An additional Petri dish, filled with water, was positioned on top of the first one to provide cooling. The lamp was positioned to shine the light from a distance of about 2 cm onto the top of the water-filled Petri dish. The reaction mix was irradiated with UV (365 nm) for 60 min at 37°C. To minimize the damage from the heat generated by the UV lamp, the water in the top Petri dish was replaced with cool water at 20 and 40 min.
A small aliquot of the photoaffinity labeling reaction mix was removed and analyzed for protein and DNA product. Radioactive labeled and unlabeled HIV-1 RT was analyzed by 10% SDS-PAGE (22). The DNA product of the HIV-1 RT was analyzed by 20% urea-PAGE. In both cases, the radioactive bands were detected by autoradiography. For a quantitative analysis, radioactive bands of interest were excised for Cerenkov counting.
DNA Product Analysis-The DNA products were separated on the basis of molecular weight using 20% polyacrylamide (acrylamide:bisacrylamide ϭ 19:1) gel electrophoresis containing 7 M urea (dimensions: 170 ϫ 140 ϫ 1.7 mm). The electrophoresis buffer was 0.89 M Tris, 0.89 M boric acid, 20 mM EDTA, pH 8.3. The gel was preelectrophoresed with bromphenol blue and xylene cyanol dye markers in deionized formamide for 1 h prior to loading the samples. Sample mixtures of 15 l were applied and electrophoresed for 3.5-4 h at 600 V until the xylene cyanol dye marker was about 5 cm from the bottom of the gel. The DNA products were visualized by autoradiography. The radioactive bands were excised and counted in Eppendorf tubes by the Cerenkov method.
Protein Analysis-For protein labeling analysis, samples from the photoaffinity labeling experiments were mixed with 2 ϫ Laemmli buffer containing the bromphenol blue dye marker (22) at 1:1 (v/v) ratio and boiled at 100°C for 3 min. The total samples were loaded onto a 10% polyacrylamide (acrylamide:bisacrylamide ϭ 30:0.8) gel containing 0.1% SDS (dimensions: 275 ϫ 140 ϫ 7 mm). The electrophoresis buffer was 25 mM Tris glycine, pH 8.3. The electrophoresis was carried out at 150 V for 1 h and 300 V for another 4.5 h. The radioactive bands were visualized by autoradiography. For quantitative analysis, the radioactive bands were excised and counted in Eppendorf tubes by the Cerenkov method.
Cerenkov Counting Calibration-The same size gel bands containing variable amounts of 32 P radioactivity were excised from a 20% urea-PAGE (or 10% SDS-PAGE) and counted in Eppendorf tubes once. Each band was then transferred to a scintillation vial and minced. The scintillation vial was filled with scintillation fluid and counted in the 32 P channel for 5 min. The data from the scintillation counting was plotted versus the data from the Cerenkov counting. The points were fitted with a linear function. The slope is the calibration factor for Cerenkov counting. The calibration factor for gel bands from 20% urea-PAGE is 1.3724. The calibration factor for gel bands from 10% SDS-PAGE is 1.7245.
Proteolytic Digestion of the Labeled HIV-1 RT and Amino Acid Sequencing of the Labeled Proteolysis Product-The photoaffinity labeling reaction (total volume 10 ml containing 10 nmol of HIV-1 RT and 100 nmol each of DNA template (RT20) and DNA primer (P12)) was performed as described above. After UV irradiation, 1.0 ml of 10 ϫ exonuclease III (ExoIII) digestion buffer was added to the reaction mixture. This reaction mixture was immediately digested by 5,000 units of ExoIII (100 unit/l) at 37°C for 30 min (1 unit of ExoIII will catalyze the release of 1.0 nmol of acid-soluble nucleotide in 30 min at 37°C in 1 ϫ buffer). After the ExoIII digestion, the reaction mixture was precipitated by trichloroacetic acid. A 20-ml glass vial was used as a container, and a stir bar constantly stirred the reaction mixture while the trichloroacetic acid was added. Ice-cold 50% trichloroacetic acid (2.5 ml) was added to the ExoIII reaction mixture to a final 1:4 (v/v) ratio. After the precipitation step, the mixture was distributed repetitively into two Eppendorf tubes and was subjected to centrifugation (13,000 rpm, microcentrifuge at 4°C for 5 min). The supernatant was removed from the precipitated protein. Multiple centrifugations were required to accumulate the precipitated protein. After centrifugation, the pellet was washed once with cold (Ϫ20°C) 100% acetone and centrifuged at 4°C for another 5 min. The pellet was air-dried and immediately dissolved into a buffer suitable for proteolytic digestion to form peptides (trypsin or Lys-C).
In the case of trypsin digestion, the pellet was recovered by dissolving it into 100 l of denaturation buffer 1 (800 mM ammonium bicarbonate, pH 7.9, and 8 M urea). The solution was incubated at 50°C for 30 min, and then it was diluted 8-fold with H 2 O. Trypsin was added at a ratio of 1:20 (w/w). The trypsin digestion was incubated at 37°C overnight. In the case of Lys-C digestion, the pellet was recovered by dissolving it into 100 l of denaturation buffer 2 (125 mM Tris-HCl, pH 7.7, 5 mM EDTA, and 5 M urea). The solution was incubated at 50°C for 30 min and then was diluted 5-fold with H 2 O. The Lys-C (1 g/l) was added at a ratio of 1:60 (w/w). The Lys-C digestion was incubated at 37°C overnight.
To remove small radioactive nucleotides, excess salts, and urea, the peptide digestion solution was dialyzed against H 2 O at 4°C overnight, then concentrated to 30 l by SpeedVac. The molecular weight cut-off of the dialysis membrane was 1,000. The concentrated peptide solution was mixed with buffer (98% formamide, 0.1% bromphenol blue) at a ratio of 1:1 (v/v), and then loaded on to a 20% urea-PAGE. Electrophoresis was conducted at 600 V for 4 h. The radioactive bands in the gel were visualized by autoradiography. A small amount of labeled peptide sample, which was digested with Pronase, was also loaded on to the same 20% urea-PAGE. The Pronase-sensitive radioactive band was the labeled peptide band of interest. This band was excised, quantified by Cerenkov counting, minced in a 1.5-ml Eppendorf tube, and then soaked in 1.0 ml of H 2 O overnight to elute the labeled peptides. The eluate was separated from the gel by centrifugation. The elution was repeated two more times so that the recovery of labeled peptide reached 90%. The eluates were combined, dialyzed against H 2 O, and concentrated again by SpeedVac.
The recovered peptide (about 20 pmol) was dissolved into 100 l of H 2 O. To purify the peptide adduct, a "gel band shift" procedure was conducted using the endonuclease HinP1I. The DNA template (RT20) was added to the peptide solution at a ratio of 80:1 (template/peptide). The standard reaction buffer was also added to the peptide solution to give a final peptide solution containing 1 ϫ HIV-1 RT reaction buffer. The DNA template and DNA product double helix (total volume was 178 l) were annealed by boiling the mixture at 100°C for 3 min and cooling it down gradually (30 min). About 20 l of 10 ϫ NEBuffer 2 (100 mM Tris-HCl, pH 7.9, 500 mM NaCl, 100 mM MgCl 2 , and 10 mM dithiothreitol) and 2 l of HinP1I (5 unit/l) were added to the mixture. The final volume was 200 l. HinP1I digestion was performed at 37°C for 30 min. The reaction mixture was then dialyzed against H 2 O at 4°C overnight. The molecular weight cut-off of the dialysis membrane was 1,000. The dialyzed solution was then concentrated to 15 l by Speed-Vac. The sample was analyzed on a 20% urea-PAGE, and the shifted labeled peptide was visualized by autoradiography. The gel was removed from the electrophoresis cell and soaked in the electroblotting buffer for 5 min. The electroblotting buffer was 10 mM CAPS, pH 11, and 20% HPLC-grade methanol. The Mini ProBlott™ membrane was prewetted with methanol for a few seconds and then was placed in a dish containing blotting buffer. The transblotting sandwich was then assembled and electroblotting was conducted at a constant current of 200 mA (50 -75 V), at 4°C for 60 min. The Mini ProBlott™ membrane was removed from the transblotting sandwich and was rinsed with deionized water prior to air drying. The labeled peptide on the Mini ProBlott™ membrane was located by autoradiography. The peptide area was cut out and subjected to Edman microsequencing on a Perkin Elmer 494 automated protein sequencer equipped with a 6-mm microcartridge. Sequence interpretation was performed on a DEC Alpha (23).

RESULTS
Photoaffinity Labeling of HIV-1 RT with S 4 -ddUTP-A previous photoaffinity labeling study employing S 4 -dUTP as a probe for the active site subunit of HIV-1 RT has been reported by Sheng and Dennis (14). The basic features of that study were repeated using the chain terminator S 4 -ddUTP, and similar results were obtained in that the labeling of the p66 subunit was light-and photoprobe-dependent and the efficiency of labeling was ϳ2%. The basic strategy for substituting S 4 -ddUTP in place of S 4 -dUTP in the current study was to ensure that no additional nucleotides could be added after the photoprobe was inserted at the 3Ј-terminal position of the elongated primer. In our previous study (14), the isolation of a small labeled peptide was not achieved and a large 16-kDa (288 -423) fragment was isolated and partially sequenced (288 -313). In our present study, the isolation and purification of the peptide derived from the Lys-C proteolysis of the labeled p66 subunit enabled us to obtain a single radioactive hexapeptide for sequencing.
Generation of Labeled HIV-1 RT Peptide by Trypsin-The photoaffinity labeling reaction mixture was treated with the 3Ј 3 5Ј DNA specific nuclease ExoIII and precipitated with trichloroacetic acid to remove the majority of radioactive oligonucleotide (and mononucleotides). The radioactive elongated primer containing the photoprobe at the 3Ј terminus (and covalently linked to the p66 subunit) is resistant to the nuclease since its 3Ј end is blocked. In Fig. 1, the results of the ExoIII digestion can be seen by comparing lanes 1 and 2 in panel A or B. The radioactive component at the origin is the band of interest, whereas the product (n) and the photodimer containing the template cross-linked to the radioactive elongated primer (T-n) are essentially removed from the trichloroacetic acid precipitate. As seen in lane 3, after treatment with either trypsin (panel A) or Lys-C (panel B), a single radioactive peptide (n-p) from the p66 subunit is produced in addition to the product (n), which persists as a detectable contaminant. The shift of n-p to a lower position is produced upon treatment with Pronase, which cleaves most (but not all) of the peptide from the cross-linked n fragment (lane 4 in A and B). A steric restraint may prevent Pronase from hydrolyzing the peptide bonds close to the derivatization point between the peptide and the 3Ј end of the oligonucleotide product. Notice that the oligonucleotide product located at n is not altered by Pronase, and therefore does not contain a peptide.
The efficiency of labeling was calculated to be 2%, by measuring the total combined radioactivity of the p66 subunit isolated as p1 and p2 bands on a 10% SDS-PAGE (14) and assuming that all of the HIV-1 RT enzyme added to the reaction was catalytically active. The reaction mixture produced a total of 200 pmol of derivatized p66 subunit. The treatment with ExoIII followed by trichloroacetic acid precipitation yielded 150 pmol of p66 subunit. The labeled peptide recovered from the (n-p) gel band after the 20% urea-PAGE procedure was 45 pmol.
Purification of n-p by HinP1I Band Shift Method-To purify the labeled peptide n-p (Fig. 1, lane 3), the radioactive band was excised from the 20% urea-PAGE gel, minced, and eluted into a HinP1I reaction buffer. An excess of the DNA template RT20 was added and annealed to the n-p peptide. The specific endonuclease HinP1I was then added and incubated for 10 min to cleave a 9-base oligonucleotide fragment from the 5Ј end of n-p to generate nЈ-p. This mixture was separated by rerunning in another 20% urea-PAGE, and that autoradiogram is shown in Fig. 2 for the peptides generated by trypsin (panel A) or Lys-C (panel B). The "band shift" achieved by digestion with HinP1I nuclease repositions the radioactive band n-p to a unique faster moving position denoted nЈ-p. This operation (removal of a 9-mer oligonucleotide from n-p) nicely separates the peptide of interest from any nonradioactive peptide contaminants that might have comigrated with n-p. The nЈ-p radioactive band contained 15 pmol of peptide in the case of the trypsin-produced peptide (panel A) and 24 pmol in the case of the Lys-C-produced peptide (panel B).
The "2" indicates the cleaving sites of HinP1I on template DNA RT20 and product DNA n. The cleavage generated nЈ-p

FIG. 1. Generation of labeled HIV-1 RT peptide by trypsin (panel A) or Lys-C (panel B).
The autoradiogram of electrophoretically separated components in a 20% urea-PAGE after specific treatments and purification of the photoaffinity labeling reaction mixture. See procedure for details of the ExoIII digestion, trichloroacetic acid precipitation, trypsin or Lys-C hydrolysis, and Pronase digestion. (CGCCACAACAS 4 ddU-peptide). The cleavage site for HinP1I endonuclease on the target adduct is shown below. n-p: 5Ј-CGCGCCCCG2CGCCACAACAS 4 ddU-peptide Template RT20: 3Ј-GCGCGGGGCGC2GGTGTTGTA-5Ј

SEQUENCE 1
Sequence Analysis of the nЈ-p Fragments-The nЈ-p fragments resulting from HinP1I treatment of the n-p component shown in Fig. 2 were transferred by electroblotting to a PVDF membrane and autoradiographed to visualize the band of interest, which was excised and a portion subjected to a Edman microsequencing. The total sample was calculated to be 12-20 pmol. The analytical data for several sequence analysis are presented in Table I for either the trypsin-or Lys-C-generated nЈ-p samples.
The sequence analysis for the nЈ-p peptide derived from a trypsin digestion sample gave evidence of a tetrapeptide that did not allow the identification of the amino acids located in either position 1 or 3. This result did not allow distinction between two possible tetrapeptide sequences ELNK (positions 79 -82) or QLCK (positions 278 -281). The sequence analysis for the nЈ-p derived from a Lys-C digestion sample was therefore conducted to make the selection between these choices since the nonomer peptide LVDFRELNK would be produced in place of the ELNK trypsin-generated tetrapeptide and the hex-  amer peptide VRQLCK would be produced in placed of the QLCK trypsin-generated tetrapeptide. The photoaffinity-labeled amino acid sequence was found to be the hexapeptide VRQLCK (positions 276 -281), which resulted from the Lys-C digestion of the p66 subunit. The data for several different experiments is collected in Table I. Note that a second template (RT19) was also used to generate the same labeled peptide.
Several attempts were made to alkylate the Cys 280 residue (just prior to the Edman microsequencing) to pinpoint the possible positions of derivatization with the photoprobe. Since we were unable to successfully achieve the alkylation, we suggest that the Cys 280 is the most likely choice of the amino acid that is covalently linked with the photoprobe.

DISCUSSION AND CONCLUSION
In our previous study of the photoaffinity labeling of the HIV-1 reverse transcriptase using S 4 -dUTP as a photoprobe (14), we isolated a 16-kDa labeled fragment (288 -423) from the p66 subunit and sequenced the N-terminal portion (288 -313). In our present study using S 4 -ddUTP as the photoprobe, we have isolated a unique labeled hexapeptide (276 -281). The photoprobe in our present studies was positioned at (and only at) the 3Ј end of the primer terminus of an actively synthesizing ternary complex and was incorporated as a chain terminator. Exposure of the photoprobe to ultraviolet light 360 nm resulted in derivatization of the amino acid Cys 280 as an accessible target, presumably in close proximity to the active site. The isolation of a target hexapeptide that is somewhat different from our initial study (14) could reflect the possible different binding options available for the different photoprobes. For example, if translocation occurs prior to photolysis, the S 4 -dUTP photoprobe could engage in a binding interaction at the 3Ј-hydroxy binding site for the next phosphodiester bond forming event. The S 4 -ddUTP would have no such new binding option since it does not have a 3Ј-hydroxyl group.
Extensive studies have been conducted to elucidate the location of the active site and the mechanism of the DNA polymerase activity of HIV-1 RT (24). Attention has been focused on the role of the Asp-triad in catalysis, since initial genetic substitution experiments by Larder et al. (8,9) and Boyer et al. (25) showed that HIV-1 RT containing either of the mutants D110Q, D185E, D186E, D185H, D110E, or D186N (and several others) were essentially inactive. This aspartate triad is a strongly conserved feature of many nucleic acid polymerases (26).
The availability of several x-ray structures of HIV-1 RT has greatly stimulated a detailed consideration of the active site of the polymerase activity (17,18). The structure of RT⅐dsDNA⅐ Fab at a resolution of 3.0 Å (18) showed that the ␣ phosphate of a modeled incoming nucleotide triphosphate could be positioned in close proximity to the Asp triad, which is located near the 3Ј-OH terminus of the primer strand of the complexed dsDNA. This observation has promoted several detailed mechanistic proposals and suggestions for a two/three divalent metal Asp triad mechanism (17,19,20,27) for DNA polymerases (HIV-1 RT, Klenow, T7 DNA polymerase, as well as DNA polymerase ␤).
Kinetic descriptions of the catalytic events have assisted in the formation of mechanistic and structural proposals in that a binary complex involving the complexation of the templateprimer with the enzyme appears to be required prior to the binding of the dNTP substrate. A conformational change occurs in the HIV-1 RT⅐DNA complex coincident with the binding of the substrate (28). The formation of this binary complex is greatly enhanced when the single stranded template extends to 7 or more nucleotides upstream from the 3Ј primer terminus (29).
Amino acid derivitizations have been conducted to implicate specific amino acids in the RT polymerization event. Pyridoxal phosphate was complexed with Lys 263 and reduced to form a stable covalent derivative, which produced an inactive RT enzyme (10). Phenyl glyoxal was used to form a unique derivative with Arg 277 , which also inactivated the polymerization activity of the RT enzyme (11).
Photoaffinity labeling studies of HIV-1 RT in various stages of assembly or catalysis have been conducted to implicate various amino acids. The photolysis of the holoenzyme HIV-1 RT in the presence of dTTP produced a derivatized Lys 73 , which was suggested to be at or near the dNTP substrate binding site (12). The photolysis of RT in the presence of short oligonucleotide primers bound to template yielded derivatized enzyme, which was linked to Leu 289 -Thr 290 or Leu 295 -Thr 296 (13,30).
The catalytic competent ternary complex has been derivatized using the photoprobe S 4 -dUTP (14) or the photoprobe (FABdCTP) (31). A large peptide containing the derivative was reported by Sheng and Dennis (14). Our present study using S 4 -ddUTP as a photoprobe with a catalytically competent strategy, photoaffinity labeled the p66 subunit of HIV-1 RT and allowed the isolation of the hexapeptide Val 276 -Lys 281 with Cys 280 as the derivatized amino acid.
We have collected the data from many diverse experimental approaches to the elucidation of the active site of HIV-1 RT and attempted to integrate the information. In Fig. 3 (A-C), we have examined the topographic locations of the C␣ carbons of the various targeted amino acids as they would be positioned in the reported 3.0-Å x-ray structure of the RT⅐dsDNA⅐Fab complex (18).
A cluster of ␣ carbons of certain targeted amino acids suggested to be involved at the active site of the polymerase are positioned at about a 90°clockwise rotation from the terminal phosphate. This rotation would correspond to about 2-3 translocations during synthesis (assuming ϳ36°/dNTP added) if the aspartic 185 served as a fixed marker for the catalytic site of the polymerase event. The location of Lys 73 is about 90°in the opposite direction (counter clockwise) and might correspond to a location of the single-stranded template upstream from the active site (Fig. 4A).
The large distances between these targeted ␣ carbons and the aspartic 185 suggested to be at the catalytic site are problematic (Fig. 4B). The static "snapshot" of the enzyme complexed with the duplex template-primer moiety might misrepresent the location of the 3Ј end of the product as being positioned in the double helix when, in fact, the dynamic synthesizing ternary complex might contain a segment (3 or more nucleotides long) that is quite flexible and not fixed in the final duplex helix.
The active site of a polymerase also presents problems with respect to topographic assignments of function to structure since derivatization (e.g. by a photoprobe substrate) could occur either before or after translocation. A substrate probe containing a photoreactive group in the binding recognition loci (4-thio moiety) might effectively target the substrate binding site prior to translocation but would be located at a very different site after translocation. In contrast, a substrate probe containing a photoreactive group in the 3Ј moiety would be positioned at the catalytic site only after bond formation (involving the 5Ј-phosphate), followed by translocation to position the 3Ј-hydroxyl for formation of the next phosphodiester bond.
Kinetic studies have indicated certain conformational changes in HIV-1 RT and one could consider that the metal/Asp triad loci not only functionally masks the negative charge of the incoming nucleotide triphosphate but actually escorts the pyrophosphate product away from the newly formed phosphodiester bond at the catalytic site. We are currently investigating a chain terminating photoprobe, which appears to derivatize at the catalytic site of the polymerase, since the probe is located at the 3Ј-position of the substrate analogue.