Apr 22, 2018 - From the Department of Chemistry, State University of New York, Buffalo, New York ..... of RNA-dependent DNA polymerase activity of RT were also made with ..... a covalent bond between the inhibitor and the enzyme causes.
Vol. 269,No. 16,Issue of April 22,pp. 12024-12031,
THEJOURNAL OF Blomlcu. CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc
1994
Printed in U.S.A.
Design of Structure-based Reverse Transcriptase Inhibitors* (Received forpublication, October 28, 1993,and in revised form, February 4, 1994)
Insug Kang and Jui
H.Wan&
From the Department of Chemistry, State University of New York, Buffalo, New York 14214-3094
Based on the crystallographic structureof the active scription experiments, it seems possible to synthesize a funcsite in the reverse transcriptase (RT) of human immu- tionally specific inhibitor of RT by using either poly(A) or olinodeficiency virus (HIV),a group of hydrophobic polyadgo(A) as the backbone and attaching largehydrophobic groups enylic acid (5') derivatives were designed and synthe- to its 2'-OH groups through ether linkages (so that the bulkier, sized as inhibitors of the enzyme.Thesecompounds hybridized, or single-stranded poly(A) or oligo(A) derivative were foundto inhibit all six of the RTs tested, with IC, could occupy the long binding cleft in a similar manner). If = 10-"-1O4 M, but did not inhibit either RNA polymerase 2,Cdinitrophenyl groups are attached in this way, the resulting I1 (even at lo-, M) or DNA polymerase I up to lo-(' M in- compound, poly-2'-0-(2,4-dinitrophenyl)-poly(A)or -oligo(A) hibitorconcentration. The underivatizedpoly(A)did (DNP-poly(A) or DNP-oligo(A)) could function as a potent renot inhibit anyof the RTs tested under the same condi- versible inhibitor of RT.If 3-fluoro-4,6-dinitrophenyl groups are tions. In aqueous solutions of purified HIV-1 RT, poly-2'attached in this way, the resulting compound, poly-2'-0-(30-(2,4-dinitrophenyl)-oligo(A)was found to inhibit the or -oligo(A) (FDNP-poly(A) or enzyme reversibly and compete with the primer-tem- fluoro-4,6-dinitrophenyl)-poly(A) FDNP-oligo(A)), could function as an affinity-labeling reagent plate poly(A)-(dT),,, whereas poly-2'-0-(3-fluor0-4,6-diniof RT. In either case,since the tropheny1)-poly(A) was found to inactivate HIV-1 RT ir- and potent irreversible inhibitor reversibly by covalent labeling. A comparison of bound inhibitor would fill the entire binding cleft of the RT molecule, it would be very difficult for the virus to avoid the physicochemicalpropertiesof the hybridspoly(A)poly(dT) anddinitrophenyl-poly(A)-poly(dT) shows that inhibition by a few point mutations. The bindingof this typeof the hydrophobic dinitrophenyl groups stabilize double inhibitor could be prevented if the mutationsabolish the whole RT is already inachelical structures. These inhibitors werealso found to binding cleft. But in that case, the mutant be effective in keeping susceptible lymphocytes viable tive. In addition, their design suggests that these inhibitors in the presence of HIV-1(wild type). The effective inhibi- may inhibit RTs more effectively than the otherenzymes and tor concentrations(EC,,) were foundto be 0.2-2.6 pB/ml. hence may exhibit low toxicity to the hostcells. No toxic effect on the host cells was found even at 100The synthesis and characterization of DNP-poly(A), DNP1000-fold higher inhibitor concentrations. oligo(A),FDNP-poly(A), and FDNP-oligo(A) are reported in this paper. These compounds were found to be potent inhibitors of all of the six different viral RTs tested in aqueous solution. They can keep susceptible T4 lymphocytes viable in the presThe major obstacle to the development of an effective anti- ence of HIV-1 in microculture plates. FDNP-poly(A) is also a AIDS drug or vaccine is the hypermutability of the human potent irreversible inactivator of all of the mammalian and immunodeficiency virus (HIV)' (1,2). For example, nevirapine bacterial ribonucleases that we tested and,hence, is not hydrowas found to be a potent inhibitor of the reverse transcriptase lyzed by any of them. The data on the inhibition of ribonucle(RT) of HIV-1 with IC,, = 90 nM (3)and could be used to combat ases by these inhibitors will be presented in a separate report. AIDS. But the replacementof either Tyr-181 by Ile or "188 by Leu in HIV-1 RT via point mutation resulted in an enzyme EXPERIMENTALPROCEDURES that remained active in the presence of nevirapine (4). It would Materials-Recombinant HIV-1 RT was purchased from Worthington be desirable to design a function-specific but species-nonspe- and stored in 10 m~ potassium phosphate (pH 7.1) with 1 m~ dithiocific inhibitor that inactivatesall viable strains of HIV RT. threitol and50%glycerol (v/v)at -20 "C. Avian myeloblastosisvirus RT Recent crystallographic data show that the polymerase ac- waspurchasedfromU. S. BiochemicalCorp.,andmurineleukemia tive site in HIV-1 RT consists of a long binding cleft that islong virus RT was purchased from Life Technologies, Inc. DNA-dependent RNA polymerase I1 was isolated from fresh calf thymus (7, 8). DNAenough to accommodate a segment of hybridized RNA 25-30 6). In view of the widespread use of dependent DNA polymerase I was also from Worthington. The potasresidues in length (5, sium salt of polyadenylicacid (5') (poly(A),average M, 1 x lo'), poly(A)-(dT),, as a primer-template for in vitro reverse tranpoly(dT) (average M, 1 x lo6), 1,5-difluoro-2,4-dinitrobenzene, l-fluoro-2,4-dinitrobenzene, and other reagentsand enzymes were purchased from Sigma. [3H]d'ITP was from DuPont NEN, p~ly-[~HIadenylic *This work was supported in part by Grant GM 41610from the acid was from Amersham Corp., and oligo(A) (8000 < M,< 10,000) was by the U/B Foundation. The costs of National Institutes of Health and prepared in our laboratory from ADP by the polynucleotide phosphopublication of this article were defrayedin part by the payment of page rylase reaction at pH 9.0 and 37 "C and subsequently separated by charges. This article must therefore be hereby marked "aduertisement" dialysis with two membranes (SpectraPor 7 with M,cut off at 8000 and in accordance with 18 U.S.C. Section 1734 solelyto indicate this fact. was prepared by equilibrating 10,000,respectively).[14ClPoly(A) $ To whom correspondence should be addressed: Dept. of Chemistry, poly(A) with[I4C]ADP at room temperaturein the presence of Acheson Hall, SUNY,Buffalo, NY 14214-3094. The abbreviations used are: HIV, human immunodeficiency virus, polynucleotide phosphorylase for 2 days, followed by gel filtration to RT, reverse transcriptase; DNP, 2,4-dinitrophenyl; DNP-poly(A), poly- remove V4C1ADP. Synthetic Procedures-The potassium salt of FDNP-poly(A) of for2'-0-(2,4-dinitrophenyl)-poly(A);DNP-oligo(A), poly-2'-0-(2,4-dinitrophenyl)-oligo(A); FDNP, 3-fluoro-4,6-dinitrophenyl; FDNP-poly(A), mulaK,(FDNP),(A),, with average n = 270, m = 56,FW = 1.1 x lo', was synthesized by dissolving 10 mg of the potassium salt of polyadenylic poly-2'-O-(3-fluoro-4,6-dinitrophenyl)-poly(A); FDNP-oligo(A1,poly-2'O-(3-fluoro-4,6-dinitrophenyl)-oligo(A); DFDB, 1,3-difluoro-2,4-dinitro- acid (5')(Sigma, MW 1.0 x 10') in 0.5 ml of water and 0.1 ml of 0.1 M &C03 + 2.0 M KHCO, solution (pH8.8). The stirred solution was mixed benzene; XlT,2,3-bis[2-methoxy-4-nitro-5-sulfophenyll-5-~~phenylamiwith 0.4 ml of acetone containing a 5-fold excess of 1,3-difluoro-2,4no)carbonyll-2H-tetrazoliumhydroxide.
-
12024
-
12025
Reverse Danscriptase Znhibitors
I
A
1.04
1.0
1'
0.8-
B
0.0
0.6-
0.40.4
0 0
0.20.2
0
0.0-8 0.
0
0
0
I
I
IO
20
'
I
- I
C
'"1 . a4
0.21
'0
40
30
0
4 IO
FIG.1. Effect of FDNP-poly(A)on the polymerase activity of viral reverse transcriptases and mammalian RNA polymerase 11. A , 54 PM HIV-1 RT (wild type); B , 0.92 IIM avian myeloblastosis virus RT; C, 0.76 MI murine leukemia virus RT and D , 100 IIM DNA-dependentRNA polymerase 11 from calf thymus. In each experiment the enzyme was preincubated with FDNP-poly(A)at theindicated concentration for 10 min at 25 "C and then assayed as described under "Experimental Procedures." A, and A denote the enzyme activities without inhibitor and with inhibitor, respectively, at a preincubation time of 10 min. [ X I denotes the total concentration of the inhibitor (including both bound and free inhibitor molecules). dinitrobenzene (DFDB) but was added in 4 aliquots at 6-h intervals. During the reaction the pH of reaction mixture was checked and readjusted to 8.8 by adding &COS and KHCO,. The progress of the reaction was monitored by thin layer chromotography analysis of the reaction mixture as a function of time on an Eastman Kodak Co. cellulose plate with a fluorescent indicator, using (20 nm &HPO, + 20 mM KH,PO,)/ CH,CN(2:1,v/v) as the developing solvent. Initially, the thin layer chromatography showed only a poly(A) band at R , = 0.38 and a DFDB band near the solvent front. As the reaction progressed, the poly(A) band continued to fade, and a new band due to FDNP-poly(A)at R, = 0.77 grew in intensity. Other peaks due to hydrolysis products also appeared near thesolvent front. Eventually,all of the poly(A)peak at RF = 0.38 was converted to the R, = 0.77 peak, which normally took 24h at room temperature. The excess DFDB was then removed by repeated extraction with a 7:l (v/v) mixture of methylene chloride and dimethyl sulfoxide. The resulting aqueous solution was centrifugally gel-filtered through Sephadex G-50 to Sephadex G-80 or dialyzed and subsequently lyophilized. The above procedure was first developed in our laboratory for synthesizing 5-fluoro-2,4-dinitrophenyl-ADP ether (9). In that case 'H NMR analysis of the principal thin layer chromotography-separated product showedit to be 3'-0-(5-fluoro-2,4-dinitrophenyl)-ADP ether. In the present reaction system, since poly(A) has only one terminal 3'-OH group and 270 2'-OH groups, the principal reaction product can only be 2"O-FDNP derivatives. Thin layer chromotography analysis indicated that poly(A) was quantitatively converted to a sharply defined polymeric product band. Unfortunately, both the 'HNMR and "FNMR peaks of FDNP-poly(A) were too broad to yield detailed structural information. The molar ratio of FDNP to adenine groups in the product FDNP-poly(A)was calculated from the observed absorbance at 330 and 259 nm to be 1:4.8, using 3-fluoro-4,6-dinitrophenyl ethyl ether (ezsg= 3900, eZw= 6300) and AMP = 15,400)as thestandards. This FDNPI adenine ratio of 1:4.8 was confirmed by "F NMR and 31P NMR measurements, using a 400-MHz Varian instrument. The observed l9FP'P
ratio in FDNP-poly(A)was also 1:4.8. DNP-poly(A) was synthesized by a similar procedure with l-fluoro2,4-dinitrobenzenereplacing DFDB. In the synthesis of FDNP-poly(A), the reaction with DFDB had to be terminated after 24 h to avoid appreciable hydrolysisof the 5-flUOrO group, and theFDNP group/adenine residue molar ratio of the product was 1:4.8. In the synthesis of DNPpoly(A), poly(A) was allowed to react with excess l-fluoro-2,4-dinitrobenzene for 48 h at room temperature, and the DNP group/adenine residue molar ratio in the product DNP-poly(A)was L1.5. FDNP-oligo(A) (of the FDNP/adenine molar ratio 1:4.8) and DNPoligo(A) (ofthe DNP/adenine molar ratio 1:1.5) wereprepared similarly - 9000) replacing poly(A) (M,- 1x lo5). The oligo(A) with oligo(A)(M, was prepared by the polymerization ofADPcatalyzed by polynucleotide phosphorylase. Potassium salt of ADP (210 mg)was dissolved in water. About 110 mgof &COS and 1pl of 1.0 M MgC1, solution were added to the ADP solution (pH adjusted to 9.0). Five units of polynucleotide phosphorylase-agarose (Sigma) were added t o the mixture and incubated at 37 "C for 48 h. The resulting mixture was dialyzed against water through a pair of dialysis bags (inside membrane with cutoff M, at 10,000 and outside membrane with cutoff M, at 8,000). These fractions were lyophilized separately. The middle fraction, with 8000 < M, < 10,000, was used as oligo(A). Assay of Enzyme Activities-The RNA-dependent DNA polymerase activity of RT was assayed by injecting a 3-pl solution of the enzyme into 27 pl of an assay mixture at 37 "C containing 125 n m Tris-HC1(pH 8.2), 1nm MgCl,, 125 nm KC1,250 p~ [3Hldl"P as the substrate, and 23 IIM poly(A)-(dT),, as the template. After the reaction had proceeded for 10 min at 37 "C, the amount of [3Hlpoly(dT) formed was determined by precipitation with 10% trichloroacetic acid, removal of the liquid with glass microfiber filters (Whatman, GF/C), and a wash of the precipitate with cold 10%trichloroaceticacid and 50 nm sodium pyrophosphate followed by ethanol, and determination of the radioactivity in the washed precipitate was by liquid scintillation counting. Similar assays
12026
Reverse Danscriptase Inhibitors 1.2
1.2
A I
c . c
c .
0
~
1.0
'
0.8
-
0.6
-
0.4
-
03
-
0
B 0
C
0.0 1 0
[DNP-oligo A] ( nM )
2
1
[DNP-01ig0 A] ( 1.2
c .
. e 0
0
C
C
4
3 pM)
D
0.6-
0.4
-
03
-
0.0
I
1
0
20 30 40 [FDNP-oligo A] ( nM )
10
50
0
1
2
[FDNP-dig0 AI (
3
pM 1
FIG. 2. Comparison Of the inhibitionefficiencies of DNF"oligo(A) and FDNP-oligo(A)on HW-1 RT and onE. DNA polymerase I. A, inhibition of HIV-1 RT (1.16 m) by DNP-oligo(A) (with DNPIA = 1:1.5), IC,, = 14 m; E , inhibition of DNA polymerase I (37.5 m) by DNP-oligo(A) (with DNP/A = 1:l.Q IC, = 12 w; C, inhibition of HIV-1 RT (1.39 m) by FDNP-oligo(A) (with FDNP/A = 1:4.8), IC, = 6 m;D, inhibition of DNA polymerase I (37.5 m) by FDNP-oligo(A) (with FDNP/A = 1:4.8), IC, = 11 w. In all cases the enzyme was preincubated with the inhibitor for 10 min at 25 "C and assayed at 37 "C.
of RNA-dependent DNA polymerase activityof RT were also made with RESULTS AND DISCUSSION poly(C)-(dG),,,, as the template andL3HldGTP as the substrate. Class-specific Inhibition of Different Reverse Banscriptuses The DNA-dependentRNA polymerase I1 activity was assayedby the procedure of Freund andMcGuire (10).The DNA polymerase I activity in Solution-Using poly(A)-(dT),, as the template, we found was assayed by the procedure of Harwood et al. ( l l ) , using that theRNA-dependent DNA polymerase activityof HIV-1 RT, poly(dA).poly(dT) as the template with E3H1"P and dATP as the sub- avian myeloblastosis virus RT, and murine leukemia virus RT strates. were all inhibited by a 20-min preincubation of the enzyme Inhibition of Purified Enzymes inSolution-In each inactivationex- with 10-ll-lO-lo M FDNP-poly(A). The dependence of the obperiment, aknown amount of the RT was preincubated with the inhibiserved ratio of the catalytic activity of inhibited enzyme (A) to tor at a given concentration for a certain length of time a t 25 "C. The incubated mixture was immediately injected intoa n assay solution at that of the uninhibited enzyme (A,) upon total concentration of the inhibitor (E111is illustrated in Fig. 1. The data in Fig. 1, 37 "C and assayed as described above. Detection of in Vitro Anti-HN Activity-The in vitro testing of anti- A-C, show that FDNP-poly(A) is a potent inhibitor of RT withHIV activity was conducted by the Antiviral Evaluations Branch,De- out species specificity and, hence, might be mutation-insensivelopmental Therapeutics Program of the Division of Cancer Treatment tive. On theotherhand, M, Fig. lD shows that even at of the National Cancer Institute, in their automated screening program. FDNP-poly(A) does not inhibit DNA-dependent RNA polymIn these in vitro tests, susceptible T4 lymphocytes (CEM-SS cell line) erase 11. were mixed with various concentrations of the inhibitor in the presence Fig. 2 shows that DNP-oligo(A) and FDNP-oligo(A)do inhibit and absence of HIV in microculture plates. The plates were subsequently incubated at 37 "C in a 5% CO, atmosphere for 6 days (12). Escherichia coli DNA polymerase I but at effective concentrations 103-foldhigher than those for HIV-1 RT. These observaThen, the tetrazolium salt X'IT was added to all wells, and the cultures were incubated to allow the development of formazan color by viable tions suggest that these inhibitors may be sufficiently functioncells. Individual wells were analyzed spectrometrically to quantitate specific so that a suitable concentration range may be formazan production and also were inspected under the microscope for determined for them to inhibit the RT of HIV effectively withdetection of viable cells toconfirm the protection of the T4 lymphocytes out causing serious damage to the host cells. against HIV. Results from drug-treated, virus-infected cells were comof Fig. 1, A and C also shows that the longer A comparison pared with thosefrom various controls onthe same plate,i.e. untreated infected cells, untreated noninfected cells, and drug-containing wells inhibitor, FDNP-poly(A) (M, 1.1 x lo5), is much more potent than the shorterinhibitor, FDNP-oligo(A)(M, 1 x by a without cells.
-
-
lo4),
Reverse Inhibitors Danscriptase 0.4
A
12027
15000
0
2
4
6
10
8
Preincubation Time (min)
0
2
4
6
8
1
0
4 0.0
0.2
0.4
0.6
1.0
0.8
l/[FDNP-Poly A] ( nM
1.2
- 1
)
FIG.4. Irreversible competitiveinhibition of HIV-1 RT by FDNP-poIy(Af.A, rate of affinity labeling of HIV-1RT.HW-1 RT (28.5 n ~was ) preincubated in solution containing 125 m~ Tris-HC1 (pH 8.21, 1 m~ MgCl,, 125 rn KC1, and 7.6 (0)or 13.3 (a)1 1 FDNP-poly(A) ~ at 25 "C. Aliquots of the mixture were removed at certain time intervals and added into the assay mixture for the measurements of polymerase activity at 37 "C;B,competitive irreversible inhibition. The preincubation mixture contained 125 m v Tris-HC1(pH 8.2),1r m MgCl,, 125 m w 0 2 4 6 8 1 0 KCl, 1-5 IZM F D ~ - ~ l y (as A indicated j and 0 (Of,34.5(Of, or 75.9 (A) Fractionnumber XIM poly(Aj-dT~~ a t 25 "C. At different time intervals, aliquots of the FIG.3. C ~ m a ~ ~ a pdemonstration h i c of the covalent label- sample were taken out andassayed for enzymeactivity a t 37 "C. ing of HIV-1RT by ~ ~ - [ ' * C l ~ l y (About A ) . 50 mg of oligo (dTfcellulose (Sigma) were equilibrated with the binding buffer (10 m~ factor of a t least lo3.The total molar concentration of FDNPTris-HC1 (pH7.5),0.5 KC0 for 30 min. The slurry was packed into a poly(A) ([XI]) in Fig. lA is indeed lower than thetotal concen1-cc syringe column and washed with the binding buffer. A, HW-1 RT tration of RT (54PM). Since the concentra~on of free ligand [I] (1.41Wj was incubated with FDNP-[l*Clpoly(A)(4.0p ~ in ) 100 111 of is unknown, the observed values of MA,, are plotted versus [XI] reaction buffer (20m Tris-HC1 (pH 7.5),1m~ MgCl,, and 60 m KC11 should ) be for 1 h a t 0 "C. The resulting solution was mixed with additional KC1 in Fig. lA. The actual IC,, value of F D ~ - p o l ~ A and diluted to 200 pl with final [KC11 = 0.5 M. The mixture was then even lower than the apparent value read from this plot. The loaded onto the oligo(dT)-cellulose column, washed with the binding data in Fig. lA also show that each long FDNP-~ly(A) ligand buffer (fractions 2 and 3),and then eluted with an elution buffer (10m~ can inactivate several HIV-1RT molecules. For example, a t [XI] Tris-HC1 (pH 7.5),without KCI) (fractions 4-7). B, HIV-1 RTwas incu= 10 PM, about 75% of the enzyme or about 40 PM RT are bated as in A but with ['*Clpol~A)replacing F D ~ - [ ' * C ] ~ l y ( AThe ). resulting mixture was loaded onto the oligo(dTj-cellulose columnand inactivated. Similarly, Fig. 1,B and C , shows that each FDNPwashed with the binding buffer, which eluted all of the protein ffrac- poly(Af molecule can inhibit a large number b10) of avian tions 1 3 ) . All of the radioactivity was subsequently eluted with the myeloblastosis virus RT or murine leukemia virus RT molelution buffer (fractions 4-81. C, same sample solution as in B was loaded onto the column, but it was washed with the elution buffer that ecules. On the other hand,the concentration of F D ~ - o l i g ~ A ) eluted all of the radioactivity (fractions 1 4 1 . All of the protein was in Fig. 2C is generally higher than thatof the RT (1.39nM), and, subsequeRtly eluted with the binding buffer. hence, the observed values of MA,, are plotted versus [I] in the
usual way.
12028
Reverse Dunscriptuse Inhibitors HIV-1 RT ( wild )
2.0 't
1
Inhibitor 1 ( nM )
[Z: Inhibitor] ( nM )
Mutant HIV-1 RT(Nevlrapine-resis~anf)
Mutant HIV-I RT ( AZT-resistant )
1.51 1.o
D
1.0 0
a a
. 0
<
