J. gen. Virol. (1983), 64, 523-532.

Printed in Great Britain 523

Key words: HSV/antiviral/thymidine kinase/resistant mutants

Characterization of Abnormal Thymidine Kinases Induced by

Drug-resistant Strains of Herpes Simplex Virus Type 1
By B R E N D A N A. L A R D E R , 1 Y U N G - C H I C H E N G 2
AND G R A H A M D A R B Y 1.
1Division of Virology, Department of Pathology, Cambridge University, Cambridge, U.K. and
2Department of Pharmacology and Medicine, School o f Medicine, University of North Carolina,
Chapel Hill, North Carolina 27514, U.S.A.

(Accepted 12 October 1982)

SUMMARY
Two TK + acyclovir-resistant variants of herpes simplex virus (HSV) (S1 and Tr7)
and one TK + BVdU-resistant variant (B3) induce abnormal thymidine kinases with
impaired ability to phosphorylate the drugs used in their isolation. These enzymes have
been purified and their properties compared with those of the wild-type (wt) parent,
SC16. The enzyme induced by S1 differed markedly from the other three in both its
responses to salt and to pH. B3 T K recognized the enzyme's natural substrates,
thymidine, deoxycytidine, dTMP and ATP as well as the wt enzyme. In contrast, Tr7
and S 1 TKs failed to bind deoxycytidine and bind thymidine less well than wt. Tr7 and
S1 TKs had affinities for dTMP similar to those of B3 and the wt enzymes. ATP
binding to wt, Tr7 and B3 enzymes was similar but this substrate bound only weakly to
S1 TK. Each mutant displayed a characteristically distinct pattern of affinities for a
range of nucleoside analogue substrates, suggesting that they will show some cross-
resistance to drugs which have a similar mechanism of action to acyclovir and BVdU.

INTRODUCTION
Several nucleoside analogues have recently been recognized as potent inhibitors of herpes
simplex virus (HSV) replication (for review, see De Clercq, 1981), and two which are attracting
considerable interest, acyclovir [ACV or 9-(2-hydroxyethoxymethyl)guanine](Elion et al., 1977;
Schaeffer et al., 1978) and BVdU [E-5(2-bromovinyl)-2'-deoxyuridine] (De Clercq et aL, 1979)
are now being used to treat herpes infections in humans (e.g. Brigden et aL, 1981 ; De Clercq et
al., 1980a). Each compound has to be activated to its triphosphate form in the infected cell and it
is this form which disrupts virus DNA synthesis (Elion et al., 1977; Furman et al., 1979; Derse et
aL, 1981 ; Allaudeen et al., 1981). The enzyme, thymidine kinase (TK), which is encoded by the
virus, has an important role in their activation since it is this enzyme which selectively converts
each drug to its monophosphate (Fyfe et al., 1978; Cheng et al., 1981 a). Virus variants which fail
to induce TK (TK-) are resistant to these compounds (Coen & Schaffer, 1980; Schnipper &
Crumpacker, 1980; Field et al., 1980; De Clercq et al., 1980 b). However, although such variants
are viable in tissue culture systems, many recent studies have shown them to be avirulent in
animal models and they appear to establish latent infections only with difficulty (Field & Wildy,
1978; Field & Darby, 1980; Tenser & Dunstan, 1979; Tenser et al., 1979).
There appear to be two strategies by which HSV may acquire resistance to these compounds
and at the same time retain pathogenicity, both strategies leading to retention of the T K +
phenotype. The first mechanism to be recognized is mutation of the DNA polymerase gene so
that DNA replication becomes insensitive to inhibition by the analogue triphosphates (Coen &
Schaffer, 1980; Schnipper & Crumpacker, 1980; Field et al., 1980; Furman et al., 1981). The
second mechanism, recognized more recently, is mutation in the TK gene resulting in the
induction of an enzyme with altered substrate specificity so that its ability to phosphorylate the

0022-1317/83/0000-5404$02.000 1983 SGM

524 B . A . LARDER, Y.-C. CHENG AND G. DARBY

drug is severely i m p a i r e d whilst at the s a m e t i m e its ability to phosphorylate thymidineis

retained ( D a r b y et al., 1981 ; L a r d e r & D a r b y , 1982).
T h r e e drug-resistant m u t a n t s of HSV-1 recently isolated in this l a b o r a t o r y ( D a r b y et al., 1981 ;
Field & N e d e n , 1982; B. A. Larder, u n p u b l i s h e d results) all h a v e T K a c t i v i t y and r e t a i n
pathogenicity. Studies w i t h purified e n z y m e s e n c o d e d by these v a r i a n t s i n d i c a t e t h a t t h e y
induce a p p a r e n t l y n o r m a l D N A p o l y m e r a s e s but that there are lesions in t h e i r T K genes w h i c h
probably account for their drug resistance ( L a r d e r ' e t al., 1983).
In this p a p e r we describe the c h a r a c t e r i z a t i o n of the T K s , purified by affinity c o l u m n
c h r o m a t o g r a p h y f r o m m u t a n t virus-infected T K - cells. W e c o m p a r e b o t h physical and k i n e t i c
properties o f these e n z y m e s w i t h purified T K i n d u c e d by w t virus. W e p r o v i d e clear e v i d e n c e
that the m u t a n t e n z y m e s differ, not only f r o m wt, but also f r o m e a c h other. T h e c h a n g e s in T K
are discussed in relation to the n o r m a l functions o f the e n z y m e , a n d t h e i r likely i m p l i c a t i o n s for
cross-resistance to o t h e r anti-herpesvirus drugs are considered.

METHODS
Cells and tissue culture. The cells used in this study were baby hamster kidney cells (BHK-21) and 5-hromo-2'-
deoxyuridine (BUdR)-resistant BHK cells, which express no cellular TK (BU-BHK). Cells were maintained in
Glasgow modified Eagle's medium containing 10% newborn calf serum and 10~ tryptose-phosphate broth.
Viruses. The parental HSV-1 strain, SC16 (Hill et al., 1975) and drug-resistant mutants made from it were used.
The isolation of SC16 S1 (S1), an ACV-resistant strain, has been described by Darby et al. (1981). The BVdU-
resistant mutant, SC16 B3 (B3) which was kindly provided by Dr H. J. Field, Department of Pathology,
Cambridge, U.K., was isolated by passage of the parent virus (SC 16) in B HK cells in the presence of 30 txM-BVdU
(Field & Neden, 1982). SC16 Tr7 (Tr7), an ACV-resistant strain, was isolated from BHK TK1 cells (BU-BHK
cells biochemically transformed to the TK ÷ phenotype with cloned HSV-1 TK gene) infected with SC16 in the
presence of 4 IxM-ACV.
Virus stocks were all cloned twice by plaque purification and grown in BHK cells using low multiplicity
infections.
Virus infection o f BU-BHK cells. Confluent monolayers of BU-BHK cells (about 2 × 10s) were infected with
either the parental (SC16) or drug-resistant strains of virus at 10 p.f.u./cell. After 18 h infection, cells were
harvested into the medium, pelleted and frozen at - 7 0 °C until required.
Extraction and purification o f virus thymidine kinase. Infected cell pellets were thawed and resuspended in 4 vol. of
extraction buffer containing 250 mM-potassium phosphate buffer pH 7.5, 2 mM-dithiothreitol (DTT), 1 mM-
EDTA, 1 mM-phenylmethylsulphonyl fluoride (PMSF), 50 ~tM-thymidine and 20~ glycerol. Cells were disrupted
by ultrasonic vibration and the sonicate was centrifuged at 15000 rev/min in a Sorvall RC 2B rotor. The
supernatant was Passed through two DEAE-cellulose columns, equilibrated with 250 raM- and 50 raM-phosphate
buffer (pH 7-5) respectively. Fractions containing TK activity recovered in the flow-through from the second
DEAE-cellulose column were pooled and solid ammonium sulphate (to 60~) was added to this material. The
precipitate formed was dissolved in affinity column buffer A (0.05 M-Tris-HC1 pH 7.5, 2 mM-DTT and 10%
glycerol) and TK was purified by affinity column chromatography as described previously using a thymidine-
Sepharose affinity matrix (Kowal & Markus, 1975; Cheng & Ostrander, 1976).
Enzymes were stored at - 2 0 °C in 30 ~ glycerol, 0.5 mg/ml bovine serum albumin (BSA) and 100 to 200 IxM-
thymidine. When stored in this way, TK activity remained stable for at least 5 months.
The TK preparations were desalted by Sephadex G-25 filtration prior to use.
Enzyme assays. TK was assayed essentially as described by Klemperer et al. (1967). However, the reaction
mixture for the purified enzymes also contained 2 mM-DTT and 0.5 mg/ml BSA in addition to 0.02 M-sodium
phosphate buffer pH 6, 5 mM-ATP, 5 n~-MgCl2 and 50 ~tM-[14C]thymidine (1-6 lxCi/ml). When thymidine was
replaced by alternative substrates the following concentrations were used: [14C]deoxycytidine ' 2 mM (3 laCi/ml)
and [14C]thymidine-5'-monophosphate (dTMP), 100 ~tM (3 ~tCi/ml).
Reaction products from thymidylate kinase assays were separated from substrate by thin-layer chromatography
on Polygram Ce1300 PEI/UV254 sheets (Macherey-Nagel & Co., F.R.G.), pre-spotted with unlabelled dTMP and
thymidine-5'-diphosphate (dTDP) as markers. The chromatograms were developed in 0.5 M-LiC1, 2 M-acetic acid.
The dTDP band was located, cut out and radioactivity determined by liquid scintillation counting.
A unit of TK activity is defined as the amount of enzyme catalysing the conversion of 1 pmol substrate/min at
37 °C under the standard reaction conditions described.
Radiolabelled substrates were obtained from Amersham International.
Antiviral compounds and chemicals. The following drugs were gifts: ACV from Dr G. B. Elion, Burroughs
WeUcome Co., Research Triangle Park, N.C., U.S.A.; E-5(2-bromovinyl)-, E-5(2-chlorovinyl)- and E-5(2-
 Abnormal TKs o f drug-resistant H S V variants 525
iodovinyl)-2'-deoxyuridines and Z-5(2-bromovinyl)-2'-deoxyuridine from Dr E. De Clercq, Rega Institute for
Medical Research, Katholieke Universiteit, Leuven, Belgium; l-(2"-deoxy-2'-fluoro-fl-D-arabinofuranosyl)-5-
methyluracil (FMAU) and 1-(2'-deoxy-2'-fluoro-fl-I>-arabinofuranosyl)-5-iodocytosine(FIAC) from Dr J. Fox,
Sloan-Kettering Institute for Cancer Research, Memorial Sloan-Kettering Cancer Center, New York, N.Y.,
U.S.A. ; 1-fl-~arabinofuranosyl thymine (Ara-T) from Dr G. A. Gentry, Department of Microbiology, School of
Medicine, University of Mississippi, Jackson, Mississippi, U.S.A. ; E-5-propenyl-2'-deoxyuridine and 5-propyl-2'-
deoxyuridine were prepared by Dr D. E. Bergstrom, Department of Chemistry, University of North Dakota,
Grand Forks, N.D., U.S.A.
Other nucleosides and nucleotides. These were purchased from Sigma.
Isoelectricfocusing. Isoelectric focusing was performed at 2 °C using an LKB 8101 column (110 ml) with 2.24%
ampholine (pH range 3-5 to 10) in a 0 to 60% glycerol gradient. The gradient also contained 2 mM-DTT and 50 ~tM-
thymidine. The column was run at 5 W constant power, with voltage not exceeding 1500 V, until the current had
decreased to and stabilized at a minimum (about 20 h). Fractions (2 ml) were collected from the gradient, the pH
was measured and a sample was removed from each to assay for TK activity using the standard assay described.
Protein determinations. Protein concentration was determined by the method of Bradford (1976).

RESULTS
ACV-resistant mutants, Tr7 and S1, and the BVdU-resistant mutant, B3, were all derived
from the same wt parental strain, SC16. The T K induced by each m u t a n t was purified, the final
stage in the purification being thymidine affinity column chromatography. Fuller details of the
purification procedure will be published elsewhere; however, it is relevant to point out that the
final yields and specific activities o f the purified wt and m u t a n t enzymes (all between 1 x 105
and 3.3 x l0 s units/mg protein) were very similar. The specific activities were also similar to
those obtained previously with purified wt HSV-1 T K (Cheng & Ostrander, 1976).
Representative purification d a t a for SC16 are summarized in Table 1.
Our initial experiments were designed to investigate whether there were any major
differences in the enzyme activities resulting in changes in their response to ionic strength or
pH.

Effect o f ionic strength on enzyme activity

The activities of the three mutant enzymes and wt were tested at various salt concentrations.
The results are shown in Fig. 1. B3, Tr7 and wt T K s all showed m a x i m u m activity with 0.25 M-
N a C I in the reaction mixture and their responses to increased salt concentration were very
similar. The enzyme induced by S1 was quite different, showing m a x i m u m activity at very low
salt concentrations and decreasing activity with increased salt.

Effect o f p H on enzyme activity

The effect of varying the p H of the assay system on the amount o f activity measured was
assessed in the range p H 5.8 to 8.0. Once again, the enzyme induced by S1 was different from the
other three. Increasing p H correlated with increasing activity in the case o f S1 T K , whereas
other enzymes were less active at higher p H (Fig. 2). The difference in p H dependence of the S1
T K activity suggested a possible alteration in a charged a m i n o acid species at or affecting the
active site of the enzyme. In fact, we had suggested earlier on the basis o f studies with crude
enzymes that S1 T K might differ in charge from the wt (Larder & Darby, 1982). This possibility
was further investigated by use of isoelectric focusing.

Table 1. Purification o f SC16 T K

Volume Units/mg Yield Purification Total
Procedure (ml) Units/ml protein (%) (fold) protein (%)
Cell sonicate 3-5 4 × 104 2-4 x 103 100 1 100
1st DEAE-
cellulose column 7-5 1.6 x I(P 4-2 x l03 86 1.75 49
2nd DEAE-
cellulose column 40 3 x 103 6 x liP 86 25 3-4
Affinity column 6 3.5 x 103 3"3 x l0 s 15 137 0-1
526 B. A. L A R D E R , Y . - C . C H E N G A N D G. D A R B Y

~" 200 J I
150
Z
~ 150
~ 100
lO0

~ 50
~ 50

,.~ 0
E- 0.25 0.5 0.75 1 6 7
NaC1 concentration (M) pH
Fig. 1 Fig. 2
Fig. 1. Effect of NaCI on the rate of thymidine phosphorylation. Activity of TK purified from SC16
and mutant virus-infected cells was measured at different concentrations of NaC1 in the standard TK
assay reaction mixture (described in Methods). Reaction velocities were measured and are expressed as
a percentage of the velocity of the control to which no NaCI was added, m, SC16; O, S1 ; O, B3; A,
Tr7.
Fig. 2. Effect of pH on the rate of thymidine phosphorylation. Activity of the purified TKs was
measured over the pH range of 5-8 to 8.0. Thymidine phosphorylation rates were measured at different
pH values of the reaction mixture, the pH being adjusted using a phosphate buffer system. TK activities
are expressed as a percentage of the activity measured at pH 6. II, SC16; O, S1; O, B3; A, Tr7.

Isoelectric focusing of S1 and wt TKs

It was hoped that we might demonstrate a difference in charge between S 1 and wt TKs by
isoelectric focusing, since the loss or gain of a charge on the mutant enzyme would have resulted
in a change in the isoelectric point. In fact, the isoelectric points for the enzymes were not
significantly different being 7.0 and 6.8 for SC16 and S1 T K respectively.

Phosphorylation of natural substrates

HSV-specific T K is a multi-functional enzyme which recognizes several different natural
substrates. The enzyme is able to phosphorylate both pyrimidine deoxynucleosides (thymidine
and deoxycytidine) and also the nucleoside monophosphate, d T M P (Kit & Dubbs, 1965;
Klemperer et al., 1967; Jamieson et al., 1974; Jamieson & Subak-Sharpe, 1974; C h e n & Prusoff,
1978; Chen et al., 1979). We firstly looked at the phosphorylation of radiolabelled substrates.
High substrate concentrations were used and under these conditions it is possible to compare the
phosphorylation rates of different substrates for a particular enzyme. The data from these
experiments are shown in Table 2. In each case the rate of phosphorylation is compared to the
phosphorylation rate for thymidine. The wt enzyme phosphorylated deoxycytidine at a similar
rate to thymidine, although this was only true at high deoxycytidine concentration and may not
reflect the situation in the infected cell. d T M P was phosphorylated at a rate double that at which
thymidine was phosphorylated but once again higher concentrations were required to achieve
this. Neither of the ACV-resistant mutants (S1 or Tr7) induced a T K with detectable
deoxycytidine kinase activity, although B3 TK, like wt, phosphorylated deoxycytidine as well as
thymidine. Data with d T M P were variable, both S1 and B3 T K s phosphorylating the nucleotide
less well than thymidine, and Tr7 T K phosphorylating it more efficiently.
We next measured Michaelis-Menten constants (Kin values) using radiolabeUed substrates, to
provide an indication of binding affinity of the enzymes for these substrates. These data are
shown in Fig. 3 and are also summarized in Table 2. In addition to the phosphate acceptors, we
measured the Km for ATP, the phosphate donor. Several interesting points emerged from these
experiments. The TKs of the ACV-resistant mutants, S1 and Tr7, both had increased Km values
for thymidine (10 and 2.5 ~tM respectively), whereas that induced by B3 had a similar Km (0"3 p.M)
 Abnormal TKs of drug-resistant H S V variants 527
0-1- 0-1-

o.o5-
(b)
/ (a)

~_ 0.05-

--5 0 ; 10 1'5 -'4 0 ; 1'0

1/thymidine (pM-~) 1/deoxycytidine(mM-1)
0.1-
(e) (d)
0.01"

0005
J
0.05-

-0.1 0 Ol 02 0:3-d.l 0 0'.1 012

I/dTMP (pM-1) 1/ATP-Mg 2+ (~tM-1)
Fig. 3. Reaction kinetics for natural substrates of thymidine kinase induced by SC16 (m), S1 (O), B3
(O) and Tr7 (Ak). Initial reaction rates were measured at different concentrations of substrate, with
saturating levels of ATP-Mg2+ (5 mM) when measuring thymidine, deoxycytidine and dTMP
pliosphorylation, and saturating levels of thymidine (150 ~tM) when ATP-Mg2+ Km values were
determined. Lineweaver-Burk plots (reciprocal of velocity versusreciprocal of substrate concentration)
are shown for (a) thymidine, (b) deoxycytidine, (c) dTMP and (d) ATP-Mg2+ Kmdeterminations. The
plots shown were constructed to illustrate differences in reaction kinetics between the different
enzymes; however, Km values (shown in Table 2) were calculated from this data redrawn on more
appropriate scales. V = pmol substrate converted per min per ml enzyme.

Table 2. Relative phosphorylation rates and kinetic constants for natural substrates of TK
Phosphorylation (9/o)relative
to thymidine* Km ~U)l"
A A
r • r •

Substrate SC16 S1 B3 Tr7 SC16 S1 B3 Tr7

Thymidine 100 100 100 100 0.2 10 0.3 2.5
Deoxycytidine 103 NI,:~ 96 NP 500 NI' 500 NI'
dTMP 196 66 37 330 10 10 10 8
ATP-Mg2+ . . . . 20 1000 20 12
* Phosphorylation rates by purified TK were measured using the standard reaction mixture described in
Methods, containing either thymidine (50 IXM),deoxycytidine (2 rnM)or dTMP (100 pM) as substrate. Reaction
velocities are expressed as a percentage of the thymidine phosphorylating rate for each enzyme.
I"Km values were evaluated for thymidine, deoxycytidine, dTMP and ATP Mg2+ from the Lineweaver-Burk
plots, as described in the legend to Fig. 3.
:~NP, No phosphorylation (phosphorylation rate < 1~); therefore, Km values could not be evaluated.

to wt enzyme (0.2 ~tM). The T K of B3 also had a similar Km for deoxycytidine (500 pM) to that of
the wt enzyme (500 pM). As no phosphorylation of deoxycytidine by S1 or Tr7 T K s could be
detected, Km values could not be determined. One other major difference between the enzymes
was seen when the Km values for A T P were measured. In this case B3 and Tr7 T K s were very
similar (Kin values of 20 VtMand 12 pM respectively) but S1 showed a 50-fold increase (Kin 1 mM)
compared to wt (Kin 20 pM). Surprisingly, all enzymes had the same Km for d T M P (10 pM),
further confirmation that this substrate probably binds to a different site on the enzyme than
528 B. A. L A R D E R , Y . - C . C H E N G A N D G. D A R B Y

Table 3. Ki values for nucleoside analogues

gi (~M)*
Nucleoside analogue SC16 S1 B3 Tr7
Thymidinet 0-2 10(50) 0.3(1.5) 2.5(12)
ACV 200 2200(11) 225(1-1) 2500(12)
BVdU 0.10 5(50) 4.5(45) 1.5(15)
Ara-T 3-6 4 2 0 ( 1 1 6 ) 6-6(1.8) 60(17)
IUdR 0-14 6(43) 1-1(7.8) 0.72(5-1)
FIAC 0-50 138(276) 2-9(5.8) 29(58)
FMAU 0-24 50(208) 1.3(5-4) 1.6(6.7)
* Purified TKs were used from wt virus- and mutant virus-infected cells. Initial reaction velocities of thymidine
phosphorylation were measured in the presence and absence of fixed concentrations of inhibitor at a saturating
concentration of ATP-Mg2+ (5 mM). K, values were either evaluated from Lineweaver-Burk plots or by the
procedure of Cheng & Prusoff (1973). Numbers in parentheses indicate the fold increase compared with the value
for SC16.
t Thymidine Km was included for comparison.

Table 4. Ki values for 5-substituted 2"-deoxyuridine analogues*

Nucleoside analogue SC16 B3 Fold increase
5-propyl-2'-deoxyuridine 0.2 1.1 5-5
E-5-propenyl-2'-deoxyuridine 0-064 1.5 23
E-5-chlorovinyl-2'-deoxyuridine 0-11 3.4 31
E-5-bromovinyl-2'-deoxyuridine 0.10 4.5 45
E-5-iodovinyl-2'-deoxyuridine 0.10 6.2 62
Z-5-bromovinyl-2'-deoxyuridine 0.86 10.2 12
* Purified TK from the wt virus (SC16) and the BVdU-resistant mutant, B3, was used. Inhibition of thymidine
phosphorylation was determined by measuring initial reaction velocities in the presence and absence of fixed
concentrations of inhibitor at a saturating level of ATP-Mg2+(5 mM).Ki values were evaluated by the procedure of
Cheng & Prusoff (1973).

that binding thymidine. Conflicting results have been obtained elsewhere with the T K induced
by S1, since Dr J. A. Fyfe at Burroughs Wellcome (personal communication) has been unable to
detect d T M P phosphorylating activity in purified preparations of this enzyme.

Inhibition studies using nucleoside analogues

We investigated the implications of these alterations in T K with respect to the recognition of
other nucleoside analogue inhibitors of HSV. It was felt that such studies might throw light on
possible patterns of cross-resistance. All analogues tested are known substrates of HSV T K
(Cheng et al., 1981a, b; Larder & Darby, 1982). The interactions of these analogues with T K
were assessed by measuring the extent to which they inhibited the phosphorylation o f
thymidine. The Ki values for these compounds are shown in Table 3. The T K s of the A C V -
resistant variants, S1 and Tr7, showed similar binding patterns, with increased Ki values
(decreased binding affinities) for all analogues relative to the wt enzyme. The Ki values for A C V
were very similar, but with all other analogues the increases were greater for S1 T K than for Tr7.
In fact, with S1 T K the smallest increase in Ki was seen with A C V (11-fold), all other K~ values
increasing by at least 40-fold.
The situation with T K induced by the BVdU-resistant mutant, B3, was different. There was a
large decrease in affinity for BVdU (45-fold compared to wt) but the enzymes' affinities for A C V
and Ara-T showed only minor changes. In addition to BVdU, significant increases in Ki were
seen with I U d R (8-fold), F I A C (6-fold) and F M A U (5-fold).
Since the alteration in binding affinity of nucleoside analogues to B3 T K seemed to be
somewhat more specific than the changes in S 1 and Tr7 enzymes we thought it worthwhile to
look at the affinity of this enzyme for a family of compounds related to BVdU. The affinities
were compared with those for the wt enzyme (Table 4). The first notable feature is that these
 Abnormal TKs o f drug-resistant H S V variants 529
compounds bound to wt TK with similar affinities (Ki values 0.06 to 0.2 ~tM). The only exception
was the Z-5-bromo compound which bound more weakly (Ki 0-9 p.M). These results are
consistent with earlier observations using T K purified from HSV-l-infected cells (Cheng, 1976;
Cheng et al., 1980, 1981 a). In each case, the binding of these compounds to the mutant enzyme
was weaker and a clear pattern emerged. Both the absolute Kt values and the differences in K~
values between the mutant and wt enzymes increased in magnitude as the size of the trans-
substituted halogen on the 2-position increased. Smaller changes were seen with a methyl
substituent at that position (E-5-propenyl-2'-deoxyuridine). The smallest change was seen with
5-propyl-2'-deoxyuridine. In this case, the double bond is removed and a methyl substituent is
present on the 2-position.

DISCUSSION
Three TKs induced by drug-resistant mutants of HSV have been purified and their properties
compared with those of the wt enzyme. All the viruses are at least 100-fold less sensitive than wt
virus to the drug used in their selection (Darby et al., 1981 ; Field & Neden, 1982) as judged by
their EDso values determined by plaque-reduction assay in BHK cells, and data will be
presented elsewhere showing that these viruses induce apparently normal DNA polymerases. A
consequence of their TK ÷ phenotype is that they remain pathogenic in animal model systems
and are also capable of establishing latent infections (Darby et al., 1981 ; Field & Neden, 1982).
Experiments designed to assess the response of these enzymes to changes in salt concentration
and pH revealed the TKs induced by Tr7 and B3 to be very similar to that of wt. S1 TK, in
contrast, behaved differently, showing sensitivity to increased salt concentration and an
increase in activity with increased pH. Although we felt that the changes in the S1 enzyme might
best be explained by an amino acid substitution at the active site thus affecting the charge on the
protein, we could detect no significant difference in the isoelectric points of S 1 and wt TKs. The
isoelectric point for wt (pI 7.0) is close to values reported previously for HSV T K from crude
infected extracts (Leung et al., 1975) but in contrast to other work using purified T K where
broad, multicomponent peaks were seen (Chen& Prusoff, 1978) we observed a single peak of
enzyme activity.
Since TK is a multifunctional enzyme it was of considerable interest to investigate the ability
of each enzyme to perform its 'normal' functions. The TK from the mutant B3 was competent in
this respect. The Km values for thymidine, deoxycytidine and dTMP were very similar to those
of the wt enzyme, indicating similar affinities for these substrates and apart from a significant
decrease in the relative rate of phosphorylation of dTMP little difference could be seen between
this enzyme and the wt. In view of these results it would be expected that the biological
properties of B3 would be very similar to those of the wt, and preliminary results in this
laboratory support this conclusion (Field & Neden, 1982).
A different pattern was seen with the ACV-resistant mutants Tr7 and S1. Most striking was
the failure of either enzyme to phosphorylate deoxycytidine. However, the Km of the wt T K for
deoxycytidine (500 ktM) suggests that this nucleoside is a poorer substrate for the enzyme than
thymidine or any of the analogues tested. This may indicate that the ability of wt T K to
phosphorylate deoxycytidine is purely fortuitous, a result of the broad spectrum of substrate
recognition inherent in the structure of the enzyme. Although the relative rates ~ of
phosphorylation of thymidine and deoxycytidine were similar, these experiments were carried
out in substrate excess, 50 ttM in the case of thymidine but 2 mM for deoxycytidine. In view of the
difference in concentrations of substrates required to achieve these rates it is unlikelyithat
similar rates would be achieved in vivo where the concentration Of deoxycytidine would almost
certainly be considerably lower than the Kin. However, these arguments may be spurious. It is
interesting that the only major discrepancy in kinetic values obtained using purified wt T K
rather than crude infected cell extracts was in the Km for deoxycytidine. The value of 500 ~tM i s
considerably higher than that measured when using crude infected cell extracts (Kin 20 ~tM)
(Larder & Darby, 1982). This anomaly, which is consistent with observations reported by others
(Jamieson & Subak-Sharpe, 1974; Chen& Prusoff, 1979), suggests that purification results in a
decrease in !~heaffinity of the enzyme for this substrate. In view of these differences in K,, values
530 B.A. LARDER, Y.-C. CHENG AND G. DARBY
observed using crude and purified TK preparations, the significance of the deoxycytidine kinase
activity in vivo must remain in some doubt.
The failure of Tr7 and S1 TKs to phosphorylate deoxycytidine was probably a result of the
general decrease in affinities of these enzymes for nucleoside substrates. [It is worth noting in
this context that we have been unable to demonstrate phosphorylation of deoxycytidine using
crude enzymes where stronger binding might be expected (Larder & Darby, 1982).] When a
wide range of analogues were tested with these enzymes each showed decreased affinity
compared to wt, the changes apparently reflecting the changes in affinity for thymidine. The
correlation was not absolute but a clear trend emerged with S1, showing lower affinity than Tr7
in all cases. The only exception to this rule was with ACV binding where the affinities were
similar. It is attractive to speculate on the basis of these kinetic data that S1 and Tr7 will show a
degree of cross-resistance to all nucleoside analogues whose selective antiviral activity is
dependent upon phosphorylation by HSV TK. This possibility is currently being investigated.
We should not conclude on the basis of the data obtained with S1 and Tr7 TKs that all TK
mutants exhibiting a TK ÷ phenotype will exhibit the pattern of resistance suggested by these
variants. In fact, data obtained with B3 argue against this idea. Although B3 TK has a
considerably reduced affinity for BVdU its affinity for ACV is unchanged and its affmity for
Ara-T is only 2-fold lower than that of the wt virus. In fact, B3 has already been shown to be
sensitive to ACV both in vitro and in vivo in the mouse (Field & Neden, 1982).
The more subtle changes in TK observed with B3 prompted us to look at the affinity of this
enzyme for a number of compounds related to BVdU. In a series of experiments using 2-
substituted vinyl compounds where the bromine atom of BVdU was replaced by other
substituents, a clear pattern emerged. All compounds had similar affinity for the wt enzyme.
However, the decrease in affinity of the mutant TK relative to that of wt enzyme for each
substrate depended on the nature of the substituent. Where the halogen was replaced by a
methyl group the decrease was 23-fold. Greater decreases in affinity were observed with CI, Br
or I substituents, the decrease increasing in magnitude as the size of the halogen atom was
increased. It therefore seems possible that the change in this enzyme has a specific effect on the
topology of the active site affecting entry of these substrates, the magnitude of the effect
depending on both the size and electronegativity of the substituent.
In conclusion, we have described three TK mutants of HSV-1 which all exhibit a TK ÷
phenotype. They were selected for resistance to either ACV or BVdU and the resistance
appeared in each case to be due to impaired ability to phosphorylate the appropriate drug. Each
variant exhibits a characteristically distinct pattern of recognition of natural substrates and
nucleoside analogues. Since they combine drug resistance with pathogenicity and ability to
establish latent infections they may be the first representatives of a large and important group of
HSV mutants.

We wouldlike to thank Hugh Field for providingus with the mutant B3. This workwas supported in part by the
Medical ResearchCouncil, U.K., and by Research Project Grant CH-29from the American Cancer Society.One
of us (B.A.L.) was in recepit of an SRC CASE Award in collaboration with the Wellcome Foundation Ltd.,
Beckenham, U.K.

REFERENCES
ALLAUDEEN, H. S., KOZARICH, J. W., BERTINO, J. R. & DE CLERCQ, E. (1981). On the m e c h a n i s m of selective i n h i b i t i o n
of herpes virus replication by (E)-5-(2-bromovinyl)-2'-deoxyuridine.Proceedings of the National Academy of
Sciences of the United States of America 78, 2698-2702.
BRADFORD,M. M. (1976). A rapid and sensitive method for the quantitation of microgram quantities of protein
utilizing the principle of protein-dye binding. Analytical Biochemistry 72, 248-254.
BRIGDEN, D., FIDDIAN, P., ROSLING, A. E. & RAVENSCROFT, T. (1981). Acyclovir- a reviewof the preclinical and early
clinical data of a new antiherpes drug. Antiviral Research 1, 203-212.
CHEN,~i~ S. & PRUSOFr,W. n. (1978). Association of thymidylate kinase activity with pyrimidine deoxyribo-
nucleoside kinase induced by herpes simplex virus. Journal of Biological Chemistry 253, 1325-1327.
CHEN,M. S. &PRUSOFF,W..It. (1979). Phosphorylationof 5-iodo-5'-amino-2'-5'-dideoxyuridineby herpes simplex
virus type 1 encoded thymidine kinase. Journal of Biological Chemistry 254, 10449-10452.
 Abnormal TKs o f drug-resistant H S V variants 531

CHEN, M. S., SUMMERS,W. P., WALKER,J, SUMMERS,W. C. & FRUSOFF, W. H. (1979). Characterization of pyrimidine
deoxyribonucleoside kinase (thymidine kinase) and thymidylate kinase as a multifunctional enzyme in cells
transformed by herpes simplex virus type 1 and in cells infected with m u t a n t strains of herpes simplex virus.
Journal of Virology 30, 942-945.
CHENG, Y.-C. (1976). Deoxythymidine kinase induced in H e L a T K - cells by herpes simplex virus type I and type 2.
II. Substrate specificity and kinetic behaviours. Biochimica et biophysica acta 452, 370-381.
CHENG, Y.-C. & OSTRANDER,~l. (1976). Herpes simplex type 1 and type 2 specific thymidine kinase. I. Induction,
purification, and general properties. Journal of Biological Chemistry 251, 2605-2610.
CHENO, Y.-C. & FRUSOFF, W. n. (1973). Relationship between the inhibition constant (Ki) and the concentration of
inhibitor which causes 50 per cent inhibitiot,/(Is0) of an enzymic reaction. Biochemical Pharmacology 22,
3099-3108.
CHENG, Y.-C., GRILL, S., RUTH, J. & BERGSTROM,D. E. (1980). Anti-herpes simplex virus and a n t i - h u m a n cell growth
activity of E-5-propenyl-2'-deoxyuridine and the concept of selective protection in antivirus chemotherapy.
Antimicrobial Agents and Chemotherapy 18, 957-961.
CHENG, Y.-C., DUTSCHMAN,G., DE CLERCQ, E., JONES, A. S., RAHIM, S. G., VERHELST, G. & WALKER, R. T. (1981a).
Differential affinities of 5-(2-halogenovinyl)-2'-deoxyuridines for deoxythymidine kinases of various origins.
Molecular Pharmacology 20, 230-233.
CHENG, Y.-C., DUTSCHMAN,G., FOX, 1.1., WATANABE,K. A. & MICHIDA, H. (1981 b). Differential activity o f potential
antiviral nucleoside analogs on herpes simplex virus-induced and h u m a n cellular thymidine kinases.
Antimicrobial Agents and Chemotherapy 20, 420-423.
COEN, D. M. & SCHAFFER,P. A. (1980). Two distinct loci confer resistance to acycloguanosine in herpes simplex virus
type 1. Proceedings of the National Academy of Sciences of the United States of America 77, 2265-2269.
DARBY, G., FIELD, H. J. & SALISBURY,S. A. (1981). Altered substrate specificity of herpes simplex virus thymidine
kinase confers acyclovir-resistance. Nature, London 289, 81-83.
DE CLERCQ, E. (1981). Nucleoside analogues as antiviral agents. Acta microbiologica Academiae scientiarum
hungaricae 28, 289-306.
DE CLERCQ, E., DESCAMPS,I., DE SOMER, P., EARR, P. J., JONES, A. S. & WALKER, R. T. (1979). (E)-5-(2-bromovinyl)-2'-
deoxyuridine: a potent and selective anti-herpes agent. Proceedingsof the National Academy of Sciences of the
United States of America 76, 2947-2951,
DE CLERCQ, E., DE GREEF, H., WILDIERS, J , DE JONGE, G., DROCHMANS,A., DESCAMPS,J. & DE SOMER, P. (1980a). Oral
(E)-5-(2-bromovinyl)-2'-deoxyuridine in severe herpes zoster. British Medical Journal 281, 1178.
DE CLERCQ, E., DESCAMPS,1, VERHELST, G., WALKER, R. T., JONES, A. S., TORRENCE, P. F. & SHUGAR, D. (1980b).
Comparative efficacy of antiherpes drugs against different strains of herpes simplex virus. Journal of
Infectious Diseases 141, 563-574.
DERSE, D., CHENG, Y.-C., FURMAN,P. A., ST. CLAIR, i . H. & ELION, G. B. (1981). Inhibition of purified h u m a n and
herpes simplex virus-induced D N A polymerases by 9-(2-hydroxyethoxymethyl)guanine triphosphate. Effects
on primer-template function. Journal of Biological Chemistry 256, 11447-11451.
ELION, G. B., FURMAN,P. A., FYFE, J. A., DE MIRANDA,P., BEAUCHAMP,L. & SCHAEFFER,H. 1. (1977). Selectivity of an
antiherpetic agent, 9-(2-hydroxyethoxymethyl)guanine. Proceedingsof the NationalAcademy of Sciences of the
United States of America 74, 5716-5720.
FIELD, H. J. & DARBY,G. (1980). The pathogenicity for mice of strains of herpes simplex virus which are resistant to
acyclovir in vitro and in vivo. Antimicrobial Agents and Chemotherapy 17, 209-216.
FIELD, H. J. & NEDEN, J. (1982). Isolation of bromovinyldeoxyuridine-resistant strains of herpes simplex virus and
successful chemotherapy of mice infected with one such strain by using acyclovir. Antiviral Research 2, 243-
254.
FmLD, H. J. & WILDY, P. (1978). The pathogenicity of thymidine kinase-deficient m u t a n t s o f herpes simplex virus in
mice. Journal of Hygiene 81, 267-277.
FIELD, H. J, DARBY, G. & WILDY, P. (1980). Isolation and characterization of acyclovir-resistant m u t a n t s of herpes
simplex virus. Journal of General Virology 49, 115-124.
FURMAN, P. A., ST. CLAIR,M. H., FYFE, J. A., RIDEOUT, J. L., KELLER, P. M. & ELION, G. B. (1979). Inhibition of herpes
simplex virus-induced D N A polymerase activity and viral D N A replication by 9-(2-hydroxyethoxymethyl)-
guanine and its triphosphate. Journal of Virology 32, 72-77.
FURMAN, P. A., COEN, D. M., ST. CLAIR,M. H. & SCHAFFER,P. A. (1981). Acyclovir-resistant m u t a n t s of herpes simplex
virus type 1 express altered D N A polymerase or reduced acyclovir phosphorylating activities. Journal of
Virology 40, 936-941.
FYFE, J. A., KELLER,P. M., FURMAN,P. A., MILLER,R. L. & ELION, 6. B. (1978). Thymidine kinase from herpes simplex
virus phosphorylates the new antiviral compound 9-(2-hydroxyethoxymethyl)guanine. Journal of Biological
Chemistry 253, 8721-8727.
HILL, T. J., FIELD, H. J. & BLYTH,W. A. (1975). Acute and recurrent infection with herpes simplex virus in the mouse:
a model for studying latency and recurrent disease. Journal of General Virology 28, 341-353.
JAMESON, A. T. & SURAK-SHAaPE, J. H. (1974). Biochemical studies on the herpes simplex virus-specified
deoxypyrimidine kinase activity. Journal of General Virology 24, 481-492.
JAMIESON,A. X., GENTRY, G. A. & SUBAK-SHARPE,J. H. (1974). Induction of both thymidine and deoxycytidine kinase
activity by herpes virus. Journal of General Virology 24, 465-480.
KIT, S. & DURBS,D. R. (1965). Properties of deoxythymidine kinase partially purified from non-infected and virus-
infected mouse fibroblast cells. Virology 26, 16-27.
532 B. A. L A R D E R , Y . - C . C H E N G A N D G. D A R B Y

KLEMPERER,H. G., HAYNES,G. R., SHEDDEN,W. I. H. & WATSON,D. H. (1967). A virus-specific thymidine kinase in
BHK21 cells infected with herpes simplex virus. Virology 31, 120-128.
KOWAL, E. P. & MARKUS, G. (1975). Affinity chromatography of mammalian thymidine kinase. Federation
Proceedings 34, 700.
LARDER,B. A. &DARBY, G. (1982). Properties of a novel thymidine kinase induced by an acyclovir-resistant herpes
simplex virus type 1 mutant. Journal of Virology 42, 649~158.
LARDER,B. A., DERSE, D., CHENG,Y.--C.& DARBY,G. (1983). Properties of purified enzymes induced by pathogenic
drug-resistant mutants of herpes simplex virus. Evidence for virus variants expressing normal DNA
polymerase and altered thymidine kinase. Journal of Biological Chemistry (in press).
LEUNG, W.-C., DUBBS,D. R., TRKULA,D. & KIT, S. (1975). Mitochondrial and herpesvirus-specificdeoxypyrimidine
kinases. Journal of Virology 16, 486-497.
SCHAEFFER,H. J., BEAUCHAMP,L., DE MIRANDA,P., ELION,G. B., BAUER,D. I. & COLLINS,P. (1978). 9-(2-Hydroxy-
ethoxymethyl)guanine activity against viruses of the herpes group. Nature, London 272, 583-585.
SCHNIPPER, L. E. & CRUMPACKER,C. S. (1980). Resistance of herpes simplex virus to acycloguanosine: the role of
viral thymidine kinase and DNA polyraerase loci. Proceedings of the Nat~anal Academy of Sciences of the
United States of America 77, 2270-2273.
TENSER, R. B. & DUNSTAN,M. E. (1979). Herpes simplex virus thymidine kinase expression in infection of the
trigeminal ganglion. Virology 99, 417-422.
TENSER,R. B., MILLER,R. L. &RAPP,F. (1979). Trigeminal ganglion infection by thymidine kinase-negative mutants
of herpes simplex virus. Science 205, 915-917.

(Received 23 August 1982)