Nonnucleoside inhibitor binding affects the interactions of the fingers subdomain of human immunodeficiency virus type 1 reverse transcriptase with DNA

Abstract

Site-directed photoaffinity cross-linking experiments were performed by using human immunodeficiency virus type 1 (HIV-1) reverse transcriptase (RT) mutants with unique cysteine residues at several positions (i.e., positions 65, 67, 70, and 74) in the fingers subdomain of the p66 subunit. Since neither the introduction of the unique cysteine residues into the fingers nor the modification of the SH groups of these residues with photoaffinity cross-linking reagents caused a significant decrease in the enzymatic activities of RT, we were able to use this system to measure distances between specific positions in the fingers domain of RT and double-stranded DNA. HIV-1 RT is quite flexible. There are conformational changes associated with binding of the normal substrates and nonnucleoside RT inhibitors (NNRTIs). Cross-linking was used to monitor intramolecular movements associated with binding of an NNRTI either in the presence or in the absence of an incoming deoxynucleoside triphosphate (dNTP). Binding an incoming dNTP at the polymerase active site decreased the efficiency of cross-linking but caused only modest changes in the preferred positions of cross-linking. This finding suggests that the fingers of p66 are closer to an extended template in the "open" configuration of the enzyme with the fingers away from the active site than in the closed configuration with the fingers in direct contact with the incoming dNTP. NNRTI binding caused increased cross-linking in experiments with diazirine reagents (especially with a diazirine reagent with a longer linker) and a moderate shift in the preferred sites of interaction with the template. Cross-linking occurred closer to the polymerase active site for RTs modified at positions 70 and 74. The effects of NNRTI binding were more pronounced in the absence of a bound dNTP; pretreatment of HIV-1 RT with an NNRTI reduced the effect of dNTP binding. These observations can be explained if the binding of NNRTI causes a decrease in the flexibility in the fingers subdomain of RT-NNRTI complex and a decrease in the distance from the fingers to the template extension.

FIG.1.

(A) Superimposition of four structures of HIV-1 RT. The solid ribbons show the positions of the β3-β4 loop positions in all four structures. The binary complex with DNA (9) (cyan), the ternary complex (red loop and lavender Cα trace) with incoming dNTP (green) and DNA (yellow) (18), the NNRTI HBY 097 complex with RT (magenta, both the inhibitor and the loop) (16), and unliganded RT (17) (navy blue) are shown. The side chains of amino acid residues of the polymerase active site D110, D185, and D186 of the p66 subunit from the ternary complex are shown. The superimposition was performed by using PROFIT program with five points fixed: two in the polymerase active site, residues 100 to 200 and 400 to 420 of p51, and the RNase H subdomain of p66 (29). The position of the extended template strand was modeled based on the distance constraints that are in agreement with the recent structural data (Tuske et al., unpublished) and the data of our cross-linking experiments. (B) Portion of the ternary structure of HIV-1 RT (Tuske et al., unpublished) with tenofovir showing the portion of the extended template and β3-β4 loop. Tenofovir, DNA, and side chains of the active site residues and K65, D67, K70, and L74 are shown in ball-and-stick form colored according to atom type. Portions of the RT are shown as ribbon diagrams colored by subdomain: blue, fingers; green, connection; and red, palm.

FIG. 2.

Enzymatic activities of wild-type and mutant HIV-1 RT with or without photocrosslinker modification. (A) Polymerase activity of wild-type (WT) HIV-1 RT and the K65C mutant (unmodified) and modified with photocrosslinkers APTP and BATDHP. The polymerase assay was performed with a 5′-labeled primer and a 200-base RNA template (shown schematically in panel B) at the top (see also Materials and Methods). The reaction was stopped by the addition of formamide loading buffer at the indicated times (0.25 to 30 min), and the products were fractionated on a 6% denaturing polyacrylamide gel. (B) At the top of the panel is a schematic diagram of the polymerase substrate used in the experiment shown in panel A. In the middle of panel B is the RNase H substrate; the presence of the ³²P label in the substrates is indicated by an asterisk. At the bottom of panel B are the results of the RNase H assay. The RNase H activities of wild-type HIV-1 RT (WT), the K65C mutant, the APTP-modified K65C mutant, the D67C mutant, the APTP-modified D67C mutant, the K70C mutant, and APTP-modified K70C mutant are compared. Reactions were performed for periods of time between 0.5 and 10 min. The uncleaved RNA (shown in the “no RT” lane) migrates to a position of about 80 bases. The −17 cleavage produces products about 50 bases long; the −8 cleavage products are about 40 bases long. The reaction was stopped by the addition of formamide loading buffer at the indicated times (0.25 to 30 min), and the products were fractionated on a 15% denaturing polyacrylamide gel.

FIG. 3.

Photocrosslinking and cleavage reactions of RT and DNA with APTP, BATDHP, and PTHBEDS. These reagents can react specifically with the SH group of a cysteine on RT. In the presence of DNA and UV light, the modified RT can react with DNA. The modified DNA can be released from RT by treatment with DTT in the case of APTP or PTHBEDS, or the vicinal hydroxyls in BATDHP can be selectively cleaved with periodate (IO₄⁻). A schematic representation of the reaction of a photocrosslinker with RT and an extended template of a dsDNA is shown below the formulas. The DNA is shown as two lines of unequal length. To monitor the cross-linking reaction, the 5′ end of the template was labeled with ³²P (✽).

FIG. 4.

SH-selective biotinylation of single Cys mutant RTs before and after modification with photocrosslinkers (a “−” symbol stands for a sample biotinylated before modification). The samples biotinylated after modification with photocrosslinking reagents were APTP (lanes a), BATDHP (lanes b), and PTHBEDS (lanes c), respectively. The numbers 65, 67, 70, and 74 represent the K65C, D67C, K70C, and L74C mutant RTs, respectively.

FIG. 5.

Yield of photocrosslinking as a function of template extension length. The average of seven independent experiments is plotted, and error is calculated as the standard deviation. The relative yield was calculated as follows: (the percent cross-linking in the RT-TP+11 complex minus the percent cross-linking in a complex of interest/the percent cross-linking in the RT-TP+11 complex) × 100.