ABSTRACT
In order to develop a photoaffinity labeling reagent for DNA polymerases,
including retroviral reverse transcriptase (RT), we utilized 2
'
,3
'
-dideoxy-
E
-5-[4-(3-trifluoromethyl-3
H
-diazirin-3-yl)styryl]UTP (TDSddUTP) as a substrate dTTP analog.
Photoaffinity labeling experiments with human immunodeficiency virus type-1 (HIV-1) RT using a radioactive labeling reagent ([
[gamma]
-
32
P]TDSddUTP) and poly(A)[middot]oligo(dT) as the template/primer yielded different results depending on
the concentration of Mg
2+
. In the presence of 0.025 mM Mg
2+
, photoaffinity labeling showed that TDSddUTP bound selectively to the dTTP
binding site in the 66 kDa subunit of the p66/p51 heterodimeric enzyme protein
when irradiated by near-UV light (365 nm). In the presence of 4 mM Mg
2+
or 0.05 mM Mn
2+
, TDSddUTP was incorporated into the 3'
-end of the primer strand due to RT activity and the resulting photolabile
primer bound to the 66 kDa subunit of HIV-1 RT on photoir- radiation. These results suggest that TDSddUTP could be a useful
tool for studying the substrate binding site(s) of DNA polymerases, including
HIV-1 RT, which show affinity for this compound.
Higher eukaryotic cells contain five species of DNA polymerases, [alpha], [beta], [gamma], [delta] and [epsilon], which play important roles in replication
and/or repair of the genome (
1
). Of these species, only the crystal structure of the DNA polymerase [beta] molecule has been successfully analyzed by X-ray crystallography (
2
-
4
). Comparison of the amino acid sequences of various DNA polymerases has allowed
speculative identification of the nucleotide substrates and/or DNA binding
domains of each enzyme molecule (
1
,
5
). Affinity labeling of these enzymes by substrate analogs may be a powerful
method for investigating the position(s) of their substrate binding domain(s) (
6
,
7
). When developing an antiviral agent, it is especially important to investigate
three-dimensional interactions between the antiviral agent in its active form
and the amino acid residue at the active site of the viral enzyme molecule and
to compare these reactions with those of eukaryotic enzymes. The reverse
transcriptases (RTs) of human immunodeficiency viruses (HIVs) 1 and 2 are the
pharmacological targets of treatment for acquired immunodeficiency syndrome
(AIDS) (
8
-
11
). Therefore, studies of the structures of the active sites of these enzymes and
of the enzymatic mechanisms by which RTs recognize their substrates are
urgently required. As the photoaffinity labeling reagents for HIV-1 RT, dTTP (
12
), 4-thio-2'-deoxyuridine 5'-triphosphate (
13
) and dCTP analogs bearing tetrafluorophenylazide (
14
) have been successfully used for active site labeling. Reactive
oligothymidylate derivatives containing an aldehyde group or 2-chloroethyl group have been reported as affinity labeling reagents (
15
). The resulting information about the substrate binding site and the crystal structure of HIV-1 RT (
16
,
17
) have made it possible to understand the basis for the antiviral activity of
nucleoside analogs such as 3'-azido-3'- deoxythymidine (AZT), 2',3'-dideoxycytidine (ddC) and 2',3'-
dideoxyinosine (ddI) and the mechanisms of drug resistance of these agents (
18
). As a typical example of a HIV-1 RT inhibitor, tritiated BI-RJ-70, an arylazide photoaffinity analog of nevirapine (BI-RG-587), has been successfully used for labeling HIV-1 RT, the binding site of this compound
within the enzyme molecule has been identified and the mechanism underlying the
anti-HIV-1 activity of nevirapine has been clarified (
19
).
Our current interest has focused on obtaining photoaffinity labeling reagents
which show selective and high affinity for each DNA polymerase, including HIV-1 RT, and are more reactive under near-UV irradiation but more stable for handling under ambient light
conditions than arylazide derivatives. As reported by Brunner
et al.
, the aryl(trifluoromethyl)diazirine skeleton is effective for labeling studies
and more easily handled than the azido functional group (
20
,
21
). Since the 5'-triphosphates of 2',3'-dideoxythymidine analogs, including AZT, 2',3'-dideoxythymidine (ddT) and
2',3'-didehydro-3'-deoxythymidine (D4T), have been
demonstrated to be potent inhibitors of HIV-1 RT (
22
), we synthesized a number of 2',3'-dideoxythymidine 5'-triphosphate (ddTTP) analogs to obtain lead
compounds which show high affinity for HIV-1 RT and evaluated their inhibitory effects on the activity of this
enzyme. Of these compounds, 2',3'-dideoxy-5-styryluridine 5'-triphosphate (StddUTP, I) (Fig.
1
) was shown to be a potent inhibitor of various enzymes, with
K
i
values of 0.05, 0.003 and 2 [mu]M for HIV-1 RT and DNA polymerases [gamma] and [beta] respectively when poly(A)[middot]oligo(dT) was used as the template/primer (
23
). Also, we have synthesized 2', 3'-dideoxy-
E
-5- [4-(3-trifluoromethyl-3
H
-diazirin-3-yl)styryl]UTP (TDSddUTP, II) (
24
,
25
). We examined the inhibitory effects of II on HIV-1 RT and reported the preliminary results of a photoaffinity labeling
study of HIV-1 RT using TDSddUTP (II). When poly(A)[middot]oligo(dT) was used as the template/primer in the dark, TDSddUTP
(II) inhibited RT activity competitively with respect to dTTP, its
K
i
value (0.075 [mu]M) for this enzyme being much smaller than the
K
m
for dTTP (14 [mu]M) (
24
,
25
). In addition to the high affinity of TDSddUTP (II) for HIV-1 RT, it is well known that 2',3'-dideoxynucleotides (ddNTPs) act as chain terminators.
The present paper describes the further characterization and application of
TDSddUTP (II) as a dTTP analog and chain terminator.
[
3
H]dTTP and [[gamma]-
32
P]ATP were purchased from Amersham. T4 polynucleotide kinase was purchased from
Nippon Gene. Recombinant HIV-1 RT was purchased from Seikagaku Kogyo. Sephadex NAP-10 columns and a silver staining kit were purchased from Pharmacia
Biotech. DE-52 resin was purchased from Whatman. dNTPs were purchased from Yamasa
Shoyu and dTTP was further purified by high performance liquid chromatography
(HPLC). The intensity of UV light was measured using a Spectronics Model DRC-100X digital radiometer and DIX-365 as a sensor. HPLC was performed using a Shimadzu LC-9A apparatus, with a flow rate of 1.0 ml/min.
Method A.
A YMC Pack ODS A-302 (YMC) reverse phase column (4.6 * 150 mm) was used. The solvent was 50 mM Et
3
NHOAc, pH 7.0, containing 26% (v/v) CH
3
CN.
Method B.
A TSK-GEL DEAE-2SW (Tohso) ion exchange column (4.6 * 250 mm) was used, with 180 mM potassium phosphate buffer, pH
6.95, containing 20% (v/v) CH
3
CN. The analyses were performed at 45oC.
TDSddUrd (V) was synthesized as reported previously (
25
) and was repeatedly purified by HPLC using a reverse phase column. The
aryl(trifluoromethyl)diazirine photoaffinity probe, [[gamma]-
32
P]TDSddUTP (II), was synthesized from the 5'-diphosphate (TDSddUDP, III) using the phosphate exchange reaction
as reported previously (
25
). Non-radioactive TDSddUTP was purified by HPLC (Method A). The purity of the
final product was >95% (Method B) and it was stored at -20oC in the dark.
Oligo(dT)
16
and its analog bearing a 2'-deoxy-[beta]-D-xylofuranosyl thymine residue at the 3'-end,
epi
-oligo(dT)
16
(Fig.
4
C) (
26
), were synthesized using an Applied Biosystems 391 DNA synthesizer and purified
by HPLC. The oligodeoxyribonucleotides were phosphorylated at the 5'-end using [[gamma]-
32
P]ATP and T4 polynucleotide kinase (
27
). The labeled oligodeoxyribonucleotides were purified by gel filtration using
Sephadex NAP-10 columns and were then condensed under reduced pressure at 35oC.
Assaying of HIV-1 RT activity was carried out as described previously using poly(A)[middot]oligo(dT)
12-18
as the template/primer (
28
).
The reaction mixture (7.5 [mu]l) comprised 50 mM Tris-HCl, pH 8.3, 1 mM dithiothreitol, 50 mM KCl, 80 [mu]g/ml poly(A), 16 [mu]g/ml oligo(dT)
16
, 0.05 [mu]M [[gamma]-
32
P]TDSddUTP (26 kBq/pmol), 40 ng (0.34 pmol) HIV-1 RT and 0.025 mM MgCl
2
or 0.5 mM MnCl
2
. When 5'-end-labeled primer was used, 0.5 [mu]g/ml 5'-
32
P end-labeled oligo(dT)
16
(15 kBq/pmol) and 0.5 [mu]M TDSddUTP, instead of 16 [mu]g/ml oligo(dT)
16
and radioactive TDSddUTP, were added. When the effect of dNTP was investigated,
various concentrations of dNTP were also included. Incubation was carried out
for 5 min at 25oC in the dark. After chilling, the reaction mixture was irradiated with a
100 W black light lamp (UVP Inc., model BA100AF) at a distance of 8 cm (11 000 [mu]W/cm
2
) for 15 min in an ice bath. When the effect of
epi
-oligo(dT)
16
on the photoaffinity labeling of the enzyme was investigated,
epi
-oligo(dT)
16
was added to the reaction mixture after the incubation for 5 min at 25oC in the dark. The mixture was incubated for a further 15 min at 37oC in the dark and irradiated with near-UV light as described above. The reaction mixture was mixed with
the same volume of a solution consisting of 0.1 M Tris-HCl, pH 6.8, 4% SDS, 20% (v/v) glycerol, 50 mM dithiothreitol and 0.02%
bromophenol blue, then heat denatured for 5 min at 95oC. The samples were subjected to SDS-PAGE. After detection of the protein bands by staining with
Coomassie brilliant blue, the gel was dried and exposed to X-ray film (Konica) with an intensifying screen for 2-24 h at -70oC. The radioactivity associated with the polypeptide was
determined by Cerenkov counting after excising the radioactive band from the
gel.
SDS-PAGE was carried out according to the method of Laemmli, using a 10%
separation gel and a 5% condensation gel (
29
).
The procedures were carried out at 0-4oC. After photolysis as described above, 1600 [mu]l irradiation mixture containing 20 [mu]g HIV-1 RT was diluted with the same volume of 60 mM Tris-HCl, pH 6.8, buffer containing 10% (v/v) glycerol,
then loaded onto a DEAE-cellulose column (0.48 * 4.5 cm) pre-equilibrated with buffer DE (25 mM Tris-HCl, pH 7.5, 0.5 mM EDTA, 10 mM 2-mercaptoethanol and 10% glycerol). After
washing the column with buffer DE (2.3 ml), the protein was eluted with a
linear gradient of KCl (0-0.8 M) in buffer DE (14 ml). Aliquots of each fraction were assayed and
analyzed by SDS-PAGE as described above.
In order to ensure the stability of TDSddUTP (II) for handling, 50 [mu]l aqueous solution of 50 [mu]M TDSddUTP in a 1.5 ml Eppendorf tube was exposed to ambient light on an
ice bed. No decrease in the TDSddUTP (II) concentration was observed on HPLC
analysis after exposure for 6 h. This reagent was thus stable enough to handle.
Photoirradiation was carried out for 15 min at 0oC in a photoaffinity labeling experiment using HIV-1 RT and TDSddUTP as described in our previous paper (
25
). In the absence of TDSddUTP (II), 98% of the original RT activity was
recovered after this irradiation procedure. This confirmed that
photoirradiation does not directly inactivate HIV-1 RT.
The radioactive analog [[gamma]-
32
P]TDSddUTP (II) was treated with HIV-1 RT and irradiated with near-UV light at 0oC. The effect of Mg
2+
ion concentration was examined using poly(A)[middot]oligo(dT)
16
. As shown in Figure
2
, radioactivity was selectively detected at 66 kDa in the presence of 0.025 mM
MgCl
2
(lane 2). This is consistent with the fact that of the 51 and 66 kDa subunits
of HIV-1 RT, only the latter subunit catalyzes the polymerase reaction. Approximately 3%
of the original radioactivity of [[gamma]-
32
P]TDSddUTP (II) was recovered from the 66 kDa band after photoaffinity labeling
(lane 2 in Fig.
2
). No radioactive 66 kDa polypeptide was obtained without photoirradiation or
primer [oligo(dT)
16
] (lanes 7 and 8 in Fig.
2
). The other template/primers tested, poly(dA)[middot]oligo(dT) and activated calf thymus DNA, did not result in the
production of a photoadduct (data not shown). Addition of dTTP inhibited the
photoaddition of [[gamma]-
32
P]TDSddUTP (II) to HIV-1 RT, while dATP, dGTP or dCTP had little effect when poly(A)[middot]oligo(dT) was used as the template/primer (Fig.
3
). A further photoaffinity labeling experiment was carried out using the 3'-terminated oligo(dT)
16
analog
epi
-oligo(dT)
16
in place of oligo(dT)
16
as the primer. As shown in Figure
4
, radioactive 66 kDa bands were detected in the presence of 5 mM MgCl
2
(lane 5) as well as 0.025 mM MgCl
2
(lane 4). Since
epi
-oligo(dT)
16
does not possess a 3'-down hydroxyl group and its 3'-up hydroxyl group might not attack the [alpha]-phosphate of [[gamma]-
32
P]TDSddUTP (II), incorporation of TDSddUMP (IV) and release of inorganic [
32
P]pyrophosphate due to nucleophilic attack by the 3'-hydroxyl group at the 3'-end of the primer did not occur. Although
photoaffinity labeling of HIV-1 RT was successfully accomplished in the presence of 5 mM Mg
2+
using
epi
-oligo(dT)
16
, the extent of photoaddition to the 66 kDa polypeptide was 55% when
epi
-oligo(dT)
16
was used in place of oligo(dT)
16
as the primer (Fig.
4
). This observation suggests that the 3'-down hydroxyl group of the primer strand might play an important
role in increasing the affinity of TDSddUTP (II) for its binding site on the
enzyme molecule.
In order to examine whether TDSddUTP (II) is a substrate for HIV-1 RT and acts as a chain terminator, non-radioactive TDSddUTP (II) was treated with HIV-1 RT using poly(A)[middot][
32
P]oligo(dT)
16
as the template/primer in the presence of 4 mM MgCl
2
or 0.5 mM MnCl
2
and was then photoirradiated at 0oC. Autoradiography of the radioactive products resolved by SDS-PAGE showed a shift in the molecular mass of the labeled enzyme
species to the 78 kDa position and no other radioactive bands were detected
between 51 and 78 kDa (lane 1 in Fig.
5
). As shown in Figure
5
, a divalent metal cation, poly(A) and TDSddUTP (II) are essential for
production of the 78 kDa radioactive polypeptide. These observations suggest
that the TDSddUMP moiety was incorporated into the 3'-terminus of the primer in the presence of divalent metal cations,
allowing this photolabile primer analog to bind covalently to the enzyme
molecule upon photoirradiation. Figure
6
shows the effect of a 3'-terminated oligo(dT)
16
analog (
epi
-oligo(dT)
16
) on the photoaffinity labeling reaction when poly(A)[middot]oligo(dT)
16
was used as the template/primer. When
epi
-oligo(dT)
16
was added to the mixture after preincubation (5 min, 25oC) and the mixture was incubated for a further 15 min at 37oC before the photoirradiation procedure, the yield of the 78 kDa
radioactive product decreased with increasing concentrations of
epi
-oligo(dT)
16
. This means that
epi
-oligo(dT)
16
competed to yield a photolabile oligodeoxyribonucleotide.
In order to clarify whether the 66 or 51 kDa subunit reacts with the photolabile
oligodeoxyribonucleotide, photolabeled product was chromatographed on a DEAE-cellulose column and the eluate was analyzed by SDS-PAGE (Fig.
8
). The amounts of the 51 and 66 kDa polypeptides were proportional to the RT
activity. On the other hand, fractions eluted from the column by a higher salt
eluent contained predominantly 51 and 78 kDa polypeptides in approximately
equimolar proportions and showed no polymerase activity. This observation
suggests that the 66 kDa subunit reacted with the photolabile
oligodeoxyribonucleotide.
Photolabile derivatives of substrates are potentially useful tools for
structural analyses of binding sites on enzymes or proteins at the molecular
level (
30
). Near-UV irradiation of a complex between a protein and a satisfactorily
photolabile substrate produces covalent linkages between the substrate and the
protein with minimal perturbation of the protein structure. Analyses of this
photoadduct provide information regarding specific contact points between the
relevant amino acid residues and the substrate. Several arylazide derivatives
of dATP and dTTP have been used for investigation of the interaction of dNTP
substrates with DNA polymerases. However, arylazides are known to have a number
of disadvantages (
31
). (i) Intermediates produced through irradiation have relatively long half-lives (
32
). (ii) Arylazides are inactivated by thiols (
33
). (iii) It is difficult to handle arylazides because of their high light
sensitivity. To circumvent these disadvantages, the carbene-generating aryl(trifluoromethyl)- diazirine has been adopted as a candidate. Perfluorophenylazides constitute a new and highly promising class of photoaffinity
labeling reagents. However, Brunner described that
aryl(trifluoromethyl)diazirines seem to undergo more efficient photolysis at a
slightly longer wavelength than perfluorophenylazides (
31
).
REFERENCES
Return