Polymerization of the triphosphates of AraC, 2',2'-difluorodeoxycytidine (dFdC) and OSI-7836 (T-araC) by human DNA polymerase alpha and DNA primase.

Article abstract of DOI:10.1016/j.bcp.2004.07.042

OSI-7836 (4'-thio-araC, T-araC) is a nucleoside analogue that shows efficacy against solid tumor xenograft models. We examined how the triphosphates of OSI-7836 (T-araCTP), cytarabine (araCTP), and gemcitabine (dFdCTP) affected the initiation of new DNA strands by the pol alpha primase complex. Whereas dFdCTP very weakly inhibited primase, both T-araCTP and araCTP potently inhibited this enzyme. Primase polymerized T-araCTP and araCTP more readily than its natural substrate, CTP, and incorporation resulted in strong chain termination. dFdCTP, araCTP, and T-araCTP inhibited pol alpha competitively with respect to dCTP. When exogenously added primentemplates were used, pol alpha incorporated all three analogues into DNA, and incorporation caused either weak chain termination (dFdCTP), strong termination (araCTP), or extremely strong termination (T-araC). Furthermore, pol alpha polymerized T-araCTP only nine-fold less well than dCTP, whereas it polymerized araCTP and dFdCTP 24- and 83-fold less well, respectively. The presence of these three analogues in the template strand resulted in significant pausing by pol alpha, although the site and severity of pausing varied between the analogues. During the elongation of primase-synthesized primers, a reaction that is thought to mimic the normal sequence of events during the initiation of new DNA strands, pol alpha polymerized all three compounds. However, incorporation of araCTP and dFdCTP resulted in minimal chain termination, while incorporation of T-araCTP still caused extremely strong termination. The implications of these results with respect to how these compounds affect cells are discussed.

Full text of DOI:10.1016/j.bcp.2004.07.042

Biochemical Pharmacology 68 (2004) 2337–2346
www.elsevier.com/locate/biochempharm
Polymerization of the triphosphates of AraC, 20,20-difluorodeoxycytidine
(dFdC) and OSI-7836 (T-araC) by human DNA polymerase a
and DNA primase
c
b
a,
b
*
Katherine A. Richardson , Tanya P. Vega , Frank C. Richardson , Chad L. Moore ,
John C. Rohloffd, Blake Tomkinsona, Raymond A. Bendelea, Robert D. Kuchtab
aOSI Pharmaceuticals Inc., 2860 Wilderness Place Boulder, CO 80301, USA
bDepartment of Chemistry and Biochemistry, University of Colorado, Boulder, CO 80309, USA
cMolecular Biology Experimental Design Inc., Louisville, CO 80027, USA
d605 Meadowbrook Drive, Boulder, CO 80303, USA
Received 17 March 2004; accepted 2 July 2004
Abstract
OSI-7836 (40-thio-araC, T-araC) is a nucleoside analogue that shows efficacy against solid tumor xenograft models. We examined how
the triphosphates of OSI-7836 (T-araCTP), cytarabine (araCTP), and gemcitabine (dFdCTP) affected the initiation of new DNA strands by
the pol a primase complex. Whereas dFdCTP very weakly inhibited primase, both T-araCTP and araCTP potently inhibited this enzyme.
Primase polymerized T-araCTP and araCTP more readily than its natural substrate, CTP, and incorporation resulted in strong chain
termination. dFdCTP, araCTP, and T-araCTP inhibited pol a competitively with respect to dCTP. When exogenously added primen-
templates were used, pol a incorporated all three analogues into DNA, and incorporation caused either weak chain termination (dFdCTP),
strong termination (araCTP), or extremely strong termination (T-araC). Furthermore, pol a polymerized T-araCTP only nine-fold less well
than dCTP, whereas it polymerized araCTP and dFdCTP 24- and 83-fold less well, respectively. The presence of these three analogues in
the template strand resulted in significant pausing by pol a, although the site and severity of pausing varied between the analogues. During
the elongation of primase-synthesized primers, a reaction that is thought to mimic the normal sequence of events during the initiation of
new DNA strands, pol a polymerized all three compounds. However, incorporation of araCTP and dFdCTP resulted in minimal chain
termination, while incorporation of T-araCTP still caused extremely strong termination. The implications of these results with respect to
how these compounds affect cells are discussed.
# 2004 Elsevier Inc. All rights reserved.
Keywords: Gemzar1; (gemcitabine; dFdC; 2020-difluorodeoxycytidine); 1-beta-D-arabinofuranosylcytosine (cytosine arabinoside, araC); OSI-7836 (T-araC,
40-thio-araC, 40-thio-beta-D-arabinofuranosylcytosine); polymerase alpha; human; DNA primase
Nucleoside analogues are an important class of cancer
chemotherapeutics that are used for the treatment of var-
ious malignancies. These compounds can target DNA
replication both directly and indirectly. For example, 5-
fluorouracil indirectly targets replication by altering dNTP
levels, while cytarabine triphosphate (araCTP) directly
affects DNA synthesis by interacting with DNA poly-
merases to slow the rate of polymerization. Of particular
interest to our group was the nucleoside analog OSI-7836
(40-thio-araC, T-araC), a structural analog of araC (Fig. 1)
currently in Phase I development for the treatment of solid
tumors [1-3]. T-araC is more efficacious in the treatment of
solid tumor xenografts than either gemcitabine or araC, a
compound that exhibits little activity ([4-7], and unpub-
lished data). Activity of T-araC may be dependent on
Abbreviations: araC, cytosine arabinoside,1-b-D-arabinofuranosylcyto-
sine; dFdC, gemcitabine,20,20-difluorodeoxycytidine; OSI-7836, 4-thio-
araC,T-araC,40-thio-beta-D-arabinofuranosylcytosine; BSA, bovine serum
albumin; TBE, Tris-borate-EDTA; Tris, tris[hydroxymethyl]amino-
methane; EDTA, ethylenediaminetetraacetic acid; pol a, human DNA
polymerase alpha; pol a; primase, human polymerase alpha primase com-
plex
* Corresponding author. Tel.: +1 303 546 7764; fax: +1 303 444 0672.
E-mail address: frichardson@osip.com (F.C. Richardson).
0006-2952/$ – see front matter # 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.bcp.2004.07.042

2338
K.A. Richardson et al. / Biochemical Pharmacology 68 (2004) 2337–2346
Fig. 1. Structures of the normal dC and the modified nucleosides, araC, OSI-7836, and dFdC. The differences as compared to dC of each modified nucleoside
are in bold and larger font.
conversion to the triphosphate form inside the cell, which
has been shown to occur over a relatively long period of
time as compared to dFdC or araC [5]. Therefore, its
efficacy in solid tumor xenografts may be due to this slow
conversion to the triphosphate form and subsequent longer
availability of this tripohosphate for use in DNA replica-
tion [5]. Similar to other nucleoside analogues, T-araC is
incorporated into DNA, suggesting that its antitumor
activity most likely involves its effects on DNA replication
and/or its incorporation into DNA. Eukaryotic DNA repli-
cation involves a host of different proteins that include at
least four different nucleotide polymerizing enzymes, pol
a, pol d, pol e, and DNA primase [8,9]. DNA pol a and
primase form a tightly bound complex that is responsible
for initiating the synthesis of all new strands of DNA. On
single-stranded DNA, primase synthesizes a short RNA
primer, approximately 8–10 nucleotides long, and then
moves intramolecularly to the pol a active site [10]. Pol a
then elongates this RNA primer via dNTP polymerization
and polymerizes 25 or so dNTPs before dissociating. This
product is further elongated by pol d and/or pol e, which
perform the bulk of both leading and lagging strand
synthesis [9].
translesion synthesis by pol a when one of these lesions is
in the template strand.
1. Materials and methods
1.1. Materials
Cytosine arabinosidetriphosphate (araCTP)was obtained
from Sigma Chemical Company. dATP, dGTP, dTTP, and
dCTP were obtained from United States Biochemicals. T-
araCTP was generously provided by Southern Research
Institute (SRI). AraC phosphoramidite was purchased from
Glen Research. Synthetic oligonucleotides of defined
sequence that contained only the four natural deoxynucleo-
tides were obtained from either Oligos, etc. or Operon
Technologies, Inc. The 35 bp araC template and 35 bp
normal template were synthesized by Qiagen/Operon Tech-
nologies. The 35 bp templates containing dFdC or T-araC
and the 20 bp primer were synthesized at OSI Pharmaceu-
ticals using an Applied Biosystem Incorporated synthesizer.
Acetylated BS Awas obtained from Promega. Polymerase a
was expressed and purified using a baculovirus expression
system [11]. The cloned human DNA pol a-primase (four-
subunitcomplex)andthetwosubunit primasecomplexwere
expressed and purified as previously described [12,13].
Currently, there are no data on the effects of T-araC on
pol a catalyzed elongation of DNA primer:templates; on
how araCTP, dFdCTP, or T-araC affect primase or the pol
a-primase coupled reactions; or on the effects of an
incorporated (T-araC) on translesion synthesis. Therefore,
to better understand the potential mechanisms of T-araC
antitumor activity, we examined the effects of these com-
pounds on some of the individual components of DNA
replication primarily responsible for the initiation of DNA
synthesis. Specifically, we compared: (a) the utilization of
araCTP, dFdCTP and T-araCTP by polymerase a as sub-
stitutes for normal nucleotides in DNA synthesis using
defined primentemplates; (b) the effects of these analogues
on the pol a-primase coupled reaction and; (c) the extent of
1.2. Synthesis of T-araC phosphoramidite (Fig. 2)
The OSI-7836 CE-phosphoramidite (Fig. 2) monomer
was synthesized by analogy to the preparation of araC
methyl-phosphoramidite by Beardsley, et al. [14].
1.3. Synthesis of dFdCphosphoramidite
dFdC phosphoramidite was synthesized from gemcita-
bine following previously published methods [15].

K.A. Richardson et al. / Biochemical Pharmacology 68 (2004) 2337–2346
2339
and T-araC. For these modified nucleotide phosphorami-
dites, the coupling time was elongated to 6 min. One
exception was the synthesis of the araC-containing 35
mer (Table 1), which was synthesized by Qiagen-Operon
Technologies, using their conditions. Oligonucleotides
containing the normal dC or any of the modified nucleo-
sides were digested to nucleosides using DNAse I, snake
venom phosphodiesterase, and bacterial alkaline phospha-
tase and standard reaction conditions [16]. The nucleosides
were then separated by HPLC using a Supelco C-18-S
reverse phase column, an Hewlett Packard HPLC with a
diode array detector, and a isocratic 50 mM KPO4 (pH 6.5),
8% MeOH buffer.
1.6. Labeling and annealing
The primer (1.0 nmols) was end-labeled using 120 mCi
of [g-32P]dATP (3000 Ci/mmol, NEN Life Sciences) and
50 U of T4 polynucleotide kinase (10 U/ml, Boehringer
Mannheim) in a final volume of 100 ml of 1ꢀ kinase buffer
(Boehringer Mannheim). After 30 min at 37 8C the unin-
corporated nucleotides were removed using Centri Sep
spin columns (Princeton Separations).
Fig. 2. Synthesis of OSI-7836 phosphoramidite. TIPDSCl2: l,3-dichloro-
1,1,3,3-tetraisopropyldisilane; Ac2O: acetic anhydride; NH4OH: ammo-
nium hydroxide; TBAF: tetrabutyl ammonium fluoride; DMTCI: 4.40-
dimethoxytrityl chloride; TCA: trichloroacetic acid; IPr2NP(Cl)OCH2
CH2CN: 2-cyanoethyl diisopropylchlorophosphoramidite.
Labeled primer and template (1:1.5) were annealed in
H2O at a primer concentration of 5–10 mM.
1.4. Primers and templates
For KM and Vmax determinations, the samples were
incubated at 90 8C for 5 min in a heat block. The heat
block was then turned off and the samples were allowed to
cool slowly to room temperature.
The primer and templates used for the pol a assays are
shown in Table 1. All reactions used to determine the KM
and Vmax for the enzymes were running start reactions, with
the modified nucleotide being the third nucleotide incor-
porated from the end of the primer. For the DNA #1, two
dAMPs are inserted before the insertion of the modified
nucleotide. Those primers and templates used in experi-
ments to determine the ability of pol a to polymerize over
an incorporated, modified nucleoside (translesion synth-
esis, DNA #2) are also shown in Table 1.
For translesion polymerization experiments, the samples
were incubated at 70 8C for 5 min in a heat block, placed
on ice for 5 min and then placed at room temperature
overnight. The annealed samples were stored at ꢁ20 8C
until use, but not past 3 weeks post end-labeling.
1.7. Incorporation reaction conditions
To determine the KM and Vmax for the various modified
nucleosides, the annealed DNA #1 (Table 1) was incubated
with human polymerase a under the following conditions:
Five pmol of a specific primer/template in a total reaction
volume of 5.0 ml were incubated at 37 8C for 8 min in
50 mM Tris–HCl (pH 7.5), 10 mM MgCl2, 1 mM dithio-
1.5. Synthesis and digestion of oligonucleotides
Synthesis of all oligonucleotides was performed on an
ABI Synthesizer using standard reagents and coupling
times with exceptions for phosphoramidites of araC, dFdC
Table 1
Sequences of primers and templates
Primer/template
DNA #1
50-GCCGCATAATTCCAACAAAG
30-CGGCGTATTAAGGTTGTTTCTT G CCAAGCTCGTAC
DNA #2
50-CCCAAACGATATTCACAAAG
30-GGGTTTGCTATAAGTGTTTCAA X GGAAGCTCGTAC
d(T3G)15
d(C2G)20
d(C3G)15
TTTGTTTGTTTGTTTGTTTGTTTGTTTGTTTGTTTGTTTGTTTGTTTGTTTGTTTGTTTG
CCGCCGCCGCCGCCGCCGCCGCCGCCGCCGCCGCCGCCGCCGCCGCCGCCGCCGCCGCCG
CCCGCCCGCCCGCCCGCCCGCCCGCCCGCCCGCCCGCCCGCCCGCCCGCCCGCCCGCCCG
G indicates site of incorporation of dCTP, araCTP dFdCTP or OSI-7836TP (T-araCTP) across from this base to determine KM and Vmax. X indicates the presence
of dC, araC, dFdC, or T-araC in the DNA strand. Assays were performed as described under Section 1.

2340
K.A. Richardson et al. / Biochemical Pharmacology 68 (2004) 2337–2346
threitol, and 1 mg/mL acetylated BSA. Each reaction also
contained: 0.1 units pol a, 80 mM of dATP as the running
start nucleotide, and various concentrations of the normal
or modified nucleotide.
merization of CTP versus T-araCTP could be determined
using the methods described in [19].
1.11. Primase coupled pol a assays
Methods developed by Boosalis, 1987 [17] to measure
the KM and Vmax for each of these nucleotides using pol a
were utilized in these studies.
Reactions (10 mL) were performed as previously
described and typically contained pol a-primase, 50 mM
Tris–HCl, pH 7.5, 5 mM MgCl2, 60 mM template DNA
(total nucleotide), 0.1 mg/mL BSA, 0.2 mM NTPs, and 5–
10 mM [a-32P]dNTPs [13]. Reactions were initiated by
adding enzyme and incubated at 37 8C for 1 h. After
quenching the reactions with gel loading buffer, products
were separated and analyzed as described above for the
primase assays. The oligonucleotides used in these assays
are shown in Table 1.
1.8. Elongation reactions
Elongation reactions were performed as for the incor-
poration reactions, but included appropriate dNTPs to
allow elongation past the site of araC, dFdC, or T-araC
incorporation.
1.9. Translesion synthesis conditions
2. Results
To determine translesion synthesis, DNA #2 (Table 1)
was annealed as above. HPLC analysis of these oligos
indicated the presence of the normal or modified nucleo-
side was present in the correct molar ratios (data not
shown). The reaction conditions for translesion synthesis
using pol a were the same as those for incorporation
reactions, except that each reaction mixture contained a
concentration range of 0.1–400 mM of a mixture of all four
deoxynucleotides. All reactions were quenched by addition
of a 2ꢀ dye stock solution consisting of 98% formamide,
20 mM EDTA, 0.025% SDS, 0.05% (w/v) bromophenol
blue, 0.05% (w/v) xylene cyanol, and 0.05% (w/v) orange
G. Products were separated by denaturing polyacrylamide
gel electrophoresis (16% acrylamide, 8 M urea in 1ꢀ
TBE.) The amount of radiolabeled products was visualized
using an Amersham Storm 840 Imager and quantified
using the ImageQuant software program.
2.1. DNA primase
In order to measure the effects of cytosine-based nucleo-
tide analogs on primase, we designed the templates
1.10. Primase assays
Reactions (10 ml) were performed as previously
described and typically contained 50 mM Tris–HC1, pH
7.9, 5 mM MgCl2, 60 mM DNA template (total nucleo-
tide), 0.05 mg/mL BSA, 1 mM dithiothreitol, and 100–
800 mM [a-32P]NTPs [18]. Reactions were initiated by
adding enzyme and incubating at 37 8C for 1 h. Assays
were quenched by adding 2.5 volumes of gel-loading
buffer (90% formamide), and the products separated by
denaturing polyacrylamide gel electrophoresis (20% poly-
acrylamide, 8 M urea) and analyzed by phosphorimagery
(Molecular Dynamics). The oligonucleotides used in these
assays are shown in Table 1.
Fig. 3. Inhibition of primase activity by araCTP, T-araCTP and dFdCTP on
the template d(C3G)15. Assays were performed as described under Section
1. Briefly, the 10 ml reactions were incubated for 1 h at 37 8C and contained
50 mM Tris–HCl, pH 7.9, 5 mM MgCl2, 60 mM d(C3G)15 template (total
nucleotide), 0.05 mg/mL BSA, 1 mM dithiothreitol, and 200 mM
[a-32P]ATP [18] and contained 200 mM NTPs and; no further addition
(C); 2.5, 5, 7.5, 10, 15, or 20 mM araCTP or T-araCTP, or; 25, 50, 75, 100,
150, or 200 mM dFdCTP. The lane marked C contained no analogue. The
length of products is noted to the right of the image.
The relative Vmax/KM for polymerization of CTP versus
T-araCTP was determined as described previously [19].
Assays contained the template d(C2G)20, 200 mM NTPs
[a-32P]GTP], and increasing concentrations of either
araCTP or T-araCTP. Since incorporation of these com-
pounds result in chain termination, the Vmax/KM for poly-

K.A. Richardson et al. / Biochemical Pharmacology 68 (2004) 2337–2346
2341
d(T3G)15, d(C2G)20, and d(C3G)15 (Table 1). Each of these
templates contain either 2 or 3 pyrimidines in order to
accommodate primase’s strong preference to initiate pri-
mer synthesis opposite pyrimidines, followed by a deox-
yguanylate to measure the effects of the analogues under
conditions where it could form a correct base-pair. Fig. 3
shows that d(C3G)is supports high levels of primase activ-
ity in the presence of the cognate NTPs (lane C). Similar
levels of primase activity were observed on d(T3G)is and
d(C2G)20 (data not shown).
AraCTP, T-araCTP, and dFdCTP inhibited primase
activity on the templates d(C3G)15 and d(T3G)15, but with
vastly different potency (Fig. 3 and data not shown).
Whereas low concentrations of T-araCTP potently inhib-
ited primase, much higher concentrations of dFdCTP were
required to obtain similar effects. A detailed kinetic ana-
lysis of inhibition by these three analogues revealed that
they inhibited primase activity competitively with respect
to NTPs and T-araCTP was the most potent inhibitor
(Table 2).
The data in Fig. 3 also indicate that primase readily
polymerizes araCTP, T-araCTP, and dFdCTP. Including
either compound in assays results in products of altered
mobility as compared to assays containing only the natural
NTPs. Significant incorporation of araCTP and T-araCTP
occur even when the concentration of analogue is 10-fold
lower than the NTP concentration, indicating that both
compounds are very good substrates for primase. In con-
trast, primase only polymerizes dFdCTP at relatively high
concentrations.
Fig. 4. Incorporation of T-araCTP and AraCTP by primase. Assays were
performed as described under, Section 1. Briefly, the 10 ml reactions were
incubated for 1 h at 37 8C and contained 50 mM Tris–HCl, pH 7.9, 5 mM
60 mM d(T3G)15 template (total nucleotide), 0.05 mg/mL BSA, 1 mM
dithiothreitol, and 200 mM [a-32P] ATP, and the indicated concentrations
of either T-araCTP or araCTP. The lane marked (–) lacked enzyme and the
lane marked (C) contained no analogue. The mobility of the dinucleotide
and trinucleotide are noted on the side of the gel. The large arrow notes the
product due to incorporation of the analogue.
Primase incorporates araCTP and T-araCTP opposite
template deoxyguanylates, and incorporation results in
strong chain termination. In assays containing d(T3G)15
and only ATP, primase synthesized products 2 and 3
nucleotides long (Fig. 4), indicating that the enzyme
encountered a template dG and then dissociated. Upon
adding either araCTP or T-araCTP, a new product that
exhibited altered electrophoretic mobility appeared, indi-
cating that primase polymerized the analogues opposite the
deoxyguanylate. Since ATP is present in all reactions, these
data also indicate that polymerization of araCTP and T-
araCTP results in strong chain termination. If incorporation
of araCTP and T-araCTP did not result in strong chain
termination, large amounts of longer products should have
been synthesized since the assays contained the next
correct NTP (ATP). Since only trace amounts of longer
products are synthesized, it is concluded that incorporation
of araCTP and T-araCTP results in strong chain termina-
tion.
The fact that incorporation of araCTP and T-araCTP
resulted in chain termination allowed us to measure how
efficiently primase polymerizes T-araCTP as compared to
CTP for the template d(C2G)20. In the presence of CTP and
GTP, primase readily synthesizes long primers, and almost
no trinucleotide is formed. When either T-araCTP or
araCTP is also present in the reaction, primase now has
a choice after generating the pppGpG dinucleotide. It can
either incorporate CTP, which will lead to the synthesis of
even longer products via continued NTP polymerization,
or it can incorporate the analogue, which will result in
chain termination and consequent accumulation of a tri-
nucleotide. Thus, by measuring the amount of analogue-
terminated trinucleotide synthesized at various ratios of
CTP/analogue, the relative Vmax/KM for CTP versus ana-
logue polymerization can be obtained [18]. Primase pre-
ferred to polymerize T-araCTP rather than CTP by a factor
of 1.7, and preferred to polymerized araCTP rather than
CTP by a factor of 1.4.
Table 2
Inhibition of primase by NTP analogues
Template
Ki (mM)
dFdCTP
AraCTP
T-araCTP
d(C3G)15
d(T3G)15
300 ꢂ 60
240 ꢂ 40
50 ꢂ 25
60 ꢂ 30
25 ꢂ 10
20 ꢂ10
Assays were performed as described under Section 1. Briefly, 10 ml reac-
tions containing 50 mM Tris–HCl, pH 7.9, 5 mM MgCl2, 60 mM DNA
template, 0.05 mg/mL BSA, 1 mM dithiothreitol, 200 mM [a-32P]GTP], pol
a primase and increasing concentrations of either dFdCTP, araCTP or T-
araCTP were incubated at 37 8C for 1 h. The reactions were analyzed by
PAGE and phosphoimagery.

2342
K.A. Richardson et al. / Biochemical Pharmacology 68 (2004) 2337–2346
Table 3
2.2. DNA polymerase a
KM, Vmax, and mutation frequency for pol a on DNA #1 (Table 1)
Compound
KM
Vmax
Vmax/KM
Misinsertion
frequency
Initial studies focused on determining the effects of
araCTP, dFdCTP, or T-araCTP on polymerization by pol
a. In each case, polymerization was measured using the
‘‘running start’’ assay developed by Boosalis et al. [17]. In
accordance with previous studies [20,21], araC insertion
resulted in very strong chain termination; incorporation of
dFdC caused a major pause at the site just after the
incorporation of dFdC; but elongation could occur in
the presence of elevated levels of dGTP and dTTP (Fig.
5). Like araC, incorporation of T-araC caused very strong
chain termination at the site of incorporation (Fig. 5).
dCTP
0.64 ꢂ 0.2
2.03 ꢂ 0.4
3.4 ꢂ 0.6
1.5 ꢂ 0.5
2.34
–
araCTP
0.2 ꢂ 0.04
0.1 ꢂ 0.05
0.47 ꢂ 0.36
0.098
0.029
0.27
0.042
0.012
0.12
dFdCTP
OSI-7836TP
1.72 ꢂ 0.15
Methods developed by Boosalis, 1987 [17] were used to determine the KM
and Vmax for each nucleoside as described in Section 1. Briefly, the annealed
DNA #1 (Table 1) was incubated in a total volume of 5 ml at 37 8C for 8 min.
Using 0.1 units pol a, 80 mM of dATP as the running start nucleotide, and
various concentrations of the normal or modified nucleotide. Misinsertion
frequency = Vmax/KM (modified)/Vmax/KM (normal).
Table 3 summarizes the steady-state kinetic parameters
and misinsertion frequencies (MF) for all nucleotides
tested. All three compounds exhibit increased KM and
decreased Vmax as compared to dCTP. When compared
to control dCTP values, T-araCTP (T-araCTP) had the least
effect on either the KM (2.7-fold increase) or the Vmax (3.2-
fold decrease). For araC, there was a 3.2-fold increase in
the KM and a 7.5-fold decrease in the Vmax. For dFdC there
was a 5.3-fold increase in the KM and a 15-fold decrease in
the Vmax. The calculated misinsertion frequencies for
polymerase a indicated that the order of misinsertion is
T-araC > araC > dFdC.
2.3. Primase-coupled polymerase a activity
We examined the effects of araCTP, T-araCTP and
dFdCTP on the pol a-catalyzed elongation of primase-
synthesized primers (i.e. primase-coupled pol a activity).
This reaction is of major interest because it recapitulates
the primary function of pol a-primase in vivo, the initiation
of new DNA strands. Initially, the ability of each com-
pound to inhibit the synthesis of elongated primers was
measured in the presence of the template d(T3G)15, ATP,
CTP, dATP and dCTP.
Fig. 5. Ability of pol a to continue synthesis following incorporation of dC,
dFdC, T-araC, or AraC reactions containing DNA #1 (Table 1) for insertion
of dCTP, dFdCTP, OSI-7836TP or araCTP across from the third base (G)
following the end of the primer. Reaction conditions are described in
Section 1. Briefly, 5 pmol of 50-end labeled primer/template were incubated
at 37 8C for 8 min in 50 mM Tris–HCl (pH 7.5), 10 mM MgCl2, 1 mM
dithiothreitol, and 1 mg/mL acetylated BSA. Each reaction also contained:
0.1 units pol a, 25 mM of dATP as the running start nucleotide, 50 mM of the
normal dCTP or modified nucleotise and 100 mM of dGTP and 100 mM of
TTP for elongation past the site of incorporation of the modified nucleoside.
The sequence of the synthesized strand is shown on the left-hand side of the
figure. The ‘‘X’’ represents position of incorporation of a modified nucleo-
tide across from a G in the template strand. Lane 1: control synthesis using
insertion of the normal Watson–Crick basepairing nucleotide (dC) across
from the G in the template strand. Lane 2: complete reaction mixture
without the pol a. Lane 3: reaction mixture containing only 25 mM of dATP,
the running start nucleotide. Lanes 4, 6, and 8: reaction mixture containing
only 25 mM dATP and 50 mM of dFdCTP, T-araCTP, or araCTP; showing
the ability of pol a to incorporate the lesion into the DNA. Lanes 5, 7, and 9:
reaction mixture containing 25 mM dATP, 100 mM each of dGTP and dTTP
and either 50 mM of dFdCTP, T-araCTP or araCTP with Lane 5 showing the
ability of pol a to incorporate dFdC and then continue synthesis, whereas,
Lanes 7 and 9 indicate the chain terminating effect of incorporation of T-
araC (Lane 7) or araC (Lane 9).
All three compounds inhibited product synthesis, but
with very different potency (Fig. 6). In assays containing
500 mM NTPs and 10 mM dNTPs, the IC50s were 2.1 mM
(araCTP), 0.65 mM (T-araCTP), and 140 mM (dFdCTP). In
addition, each compound had very different effects on the
size distribution of products. As compared to products
generated in control reactions, T-araCTP resulted in sig-
nificantly shorter products, araCTP had only very slight
effects on product length while dFdCTP had no detectable
effect (data not shown).
In order to further explore polymerization of each
analogue, we examined the effects of the analogues on
primase-coupled pol a activity on the template d(T3G)15 in
the absence of dCTP. Under these conditions, primase
synthesizes
a primer and pol a can polymerize
[a-32P]dATP opposite the template thymidylates. Upon
encountering a template deoxyguanylate, for which pol
a lacks the required dCTP, the enzyme complex will

K.A. Richardson et al. / Biochemical Pharmacology 68 (2004) 2337–2346
2343
Fig. 6. Inhibition of primase-coupied pol a activity by araCTP, T-araCTP,
and dFdCTP on the template d(T3G)15 Assays were performed as described
under Section 1. Briefly, 10 ml reactions were incubated for 1 h at 37 8C and
contained pol a-primase, 50 mM Tris–HCl, pH 7.5, 5 mM MgCl2, 60 mM
d(T3G)15 template (total nucleotide), 0.1 mg/mL BSA, 5–10 mM
[a-32P]dATP, 0.2 mM NTPs [13] and either no inhibitor (C), or 0.5,
1,2,4, 8 or 16 mM of the indicated compound. The lane marked C contained
no analogue. The lengths of products are noted to the right of the image.
polymerize CTP [13,22]. Fig. 7 shows that in the presence
of both dATP and dCTP, products approximately 60
nucleotides long are synthesized. Omitting dCTP results
in products up to approximately 30 nucleotides long, 8–10
of which comprise the primer. Adding 10 mM araCTP to
the assays results in products of greater length and of
altered electrophoretic mobility, indicating that araCTP
was polymerized opposite the template deoxyguanylates
and that chain termination did not occur. Including
dFdCTP markedly increased product length to near that
of those synthesized in the assays containing dCTP, indi-
cating that pol a incorporated dFdCTP and that incorpora-
tion did not inhibit subsequent polymerization events. In
contrast, adding T-araCTP resulted in significantly shorter
products of altered electrophoretic mobility, indicating that
pol a-primase incorporated the analogue and chain termi-
nation ensued. Similar results were obtained on the tem-
plate d(C3G)15 upon the omission of dCTP (data not
shown).
Fig. 7. Incorporation of dFdCTP, araCTP and T-araCTP during primase-
coupled pol a activity on the template d(T3G)15. Assays were performed as
described under Section 1. Briefly, 10 ml reactions were incubated for 1 h at
37 8C and contained pol a-primase, 50 mM Tris–HCl, pH 7.5, 5 mM MgCb,
60 mM d(T3G)15 template (total nucleotide), 0.1 mg/mL BSA, 5–10 mM
[a-32P]dATP, 0.2 mM NTPs [13]. Assays contained ATP, CTP, dATP and
either C (10 mM dCTP): (1) no further addition; (2) 10 mM araCTP; (3)
10 mM T-araCTP; or (4) 10 mM dFdCTP.
ing for control; Fig. 8) and, synthesis past the dFdC lesion
was similar to the control from 50 to 400 mM dNTP (2.5–
5% pausing). There was pronounced pausing of translesion
synthesis at the base just before the araC lesion at dNTP
concentrations of 2 mM and above (20–30% versus 12–
20% pausing for control; Fig. 8), and elongation was
concentration dependent.
For the T-araC template, pausing of synthesis occurred
across from T-araC and was concentration dependent from
10 to 400 mM dNTP (5–14% versus 3–4% pausing for
control; Fig. 8). The amount of elongation past the T-araC
lesion was reduced as compared to all other templates (2%
as compared to 4–5% elongation; Fig. 8).
2.4. Translesion synthesis by DNA polymerase a
In an effort to elucidate the effects of internal incorpora-
tion of araC, dFdC, or T-araC, the ability of pol a to
polymerize dNTPs opposite templates containing either
araC, T-araC or dFdC at the third base following the end of
a DNA primer was determined (Table 1). There was slight
pausing of synthesis across from the dFdC lesion compared
to control at 10–400 mM dNTP (4–8% versus 3–4% paus-
3. Discussion
We have examined the effects of dFdCTP, araCTP and T-
araCTP on human DNA primase and pol a. T-araCTP was

2344
K.A. Richardson et al. / Biochemical Pharmacology 68 (2004) 2337–2346
chain termination. Presumably, the more efficient poly-
merization and chain termination reflects the altered che-
mical properties of sulfur as compared to oxygen at the 40-
position. It has previously been demonstrated that although
primase polymerizes both araATP and 2-fluoro-araATP
more efficiently than ATP, incorporation results in very
strong chain termination [19,23,24]. The similarity of the
effects of three different araNTPs indicates that NTPs
containing a 20-hydroxyl in the ara configuration should
generally be both good substrates and potent chain termi-
nators for primase.
dFdCTP interacted with primase very similarly to how
dCTP interacts with primase [19], even though the
chemical properties of the 20-position are extremely dif-
ferent between these two compounds. Both compounds
are relatively weak inhibitors and poor substrates of
primase. The weak polymerization of these two com-
pounds emphasizes the importance of the 20-hydroxy,
and suggests that primase makes a specific contact with
the 20-hydroxyl.
Studies with pol a have demonstrated that araATP,
araCTP, and dFdCTP potently inhibit the elongation of
exogenously added DNA primer:templates and inhibition
is competitive with respect to the natural substrate
[19,21,25]. In addition to binding the analogues, pol a
readily polymerizes them with incorporation of either
araCTP or araATP resulting in immediate, strong chain
termination [21,25-28]. Interestingly, after polymerization
of dFdCTP, pol a polymerizes one additional dNTP, at
which point further synthesis is strongly inhibited [21]. In
our studies, similar results for araC and dFdC were
observed. In comparison, T-araC, like araC, caused
immediate and strong chain termination, suggesting that
the 20-ara-hydroxyl plays an important role in the chain
termination aspect of these compounds.
Fig. 8. Translesion synthesis by pol a for control and lesion-containing
DNA translesion synthesis as described in Section 1 using DNA #2 from
Table 1. Briefly, five pmol of 50-end labeled primer/template were incubated
at 37 8C for 8 min in 50 mM Tris–HCl (pH 7.5), 10 mM MgCl2, 1 mM
dithiothreitol, and 1 mg/mL acetylated BSA. Each reaction also contained:
0.1 units pol a, and 0.1–400 mM of a mixture of all four deoxynucleotides.
The percent of radioactivity incorporated is depicted as a function of
the concentration of an equimolar mixture of dGTP, dATP, dCTP, dTTP:
(A) incorporation of the first two bases prior to the position of the control
(blue), dFdC (red), araC (It. blue), T-araC (yellow) in the template; (B)
Incorporation across from the control (blue), dFdC (red), araC (It. blue),
T-araC (yellow) in the template (3rd bp); and (C) Incorporation past the site
of control (blue), dFdC (red), araC (It. blue), T-araC (yellow) in the
template.
In contrast to results using pol a alone during primase-
coupled pol a activity, both araCTP and dFdCTP were
readily incorporated into the growing RNA–DNA strands
when primase-coupled pol a complex was used. Impor-
tantly, pol a continued polymerizing additional dNTPs
after incorporation of these two analogues. Why pol a
should exhibit such divergent properties depending upon
the source of the primer, remains a mystery. However,
these data are consistent with previous studies showing that
calf thymus pol a-primase readily incorporates araATP
into internucleotide linkages during the elongation of
primase synthesized primers but not when elongating
exogenously added primer templates [22]. Importantly,
pol a continued polymerizing additional dNTPs after
incorporation of these two analogues. The rapid incorpora-
tion of these analogues into internucleotide linkages during
primase coupled pol a activity would account for the
observation in whole cells, that the majority of these
analogues are both incorporated into DNA and are found
at internucleotide linkages [29,30]. In contrast, incorpora-
tion of these analogues results in either moderate
the most potent inhibitor of primase, and incorporation of
T-araCTP resulted in chain termination. Similarly, T-
araCTP was the most potent inhibitor of pol a. Whereas
polymerization of dFdCTP and araCTP resulted in mod-
erate and strong chain termination, respectively, polymer-
ization of T-araCTP resulted in extremely strong chain
termination. This was true both when pol a elongated
exogenously added primentemplates as well as primase-
generated primentemplates. When present in the template,
all three nucleotide analogues caused pausing by pol a.
However, the strength and position of the pause varied for
each analogue.
T-araCTP was the strongest inhibitor of primase, while
dFdCTP was the weakest inhibitor. Compared to araCTP,
primase polymerized T-araCTP approximately 20% more
readily than araCTP, but incorporation resulted in stronger

K.A. Richardson et al. / Biochemical Pharmacology 68 (2004) 2337–2346
2345
(dFdCTP) or strong (araCTP) chain termination when pol
Acknowledgement
a elongates exogenously added primentemplates. These
data are consistent with previous studies showing that calf
thymus pol a-primase readily incorporates araATP into
internucleotide linkages during the elongation of primase
synthesized primers but not when elongating exogenously
added primer:templates.
This work was partially supported by National Institutes
of Health Grant GM54194 to R.D.K. T.P.V. was supported
in part by a NIH Predoctoral Training Grant T32
GM08759.
In comparison, T-araC still appeared to cause a signifi-
cant reduction in the elongation by the primase-coupled
pol a complex. Despite this reduction in elongation,
studies in cells and tumors demonstrate that the vast
majority of T-araC present in DNA following treatment
is also located in intranucleotide linkages [31].
References
[1] Whistler RL, Doner LW, Nuyak UG. 4-Thio-D-arabinofuranosylpyr-
imidine nucleosides. J Org Chem 1971;36:108–20.
[2] Siu LL, Goss GD, Powers J, Waterfield B, MacLean M, Adams LM, et
al. A phase I trial of the novel nucleoside analog OSI-7836 given on
days 1 and 8 every 3 weeks. Eur J Cancer 2003;1(Suppl 5):S165.
[3] de Jonge M, O’Donnell A, Verweij J, Chick J, Hanwell J, Schothorst
K, et al. Phase I Evaluation of OSI-7836—a New Nucleoside Analo-
gue—Using Two Administration Schedules. AACR-NCI-EORTC
Meeting, November 2003.
A key question is how large amounts of T-araCTP
becomes incorporated into internucleotide linkages in
whole cells with only small amounts found at the 30-
terminus of DNA. In contrast to the results with araCTP,
T-araCTP remained a very strong chain terminator even
when using a primase-synthesized primer. Potentially,
pol d and/or pol e, two additional DNA polymerases
involved in replication, might incorporate T-araCTP into
internucleotide linkages, thereby accounting for the
results observed in whole cells. Alternatively, accessory
proteins with which pol a likely interacts in vivo might
greatly alter the properties of pol a, such that it can
continue polymerizing dNTPs after incorporating a T-
araCTP. Work with DNA synthesomes has suggested that
these complexes of DNA replication machinery may
more accurately reflect the manner in which cells will
handle DNA analogs and have demonstrated elongation
with araC and dFdC [32]. It would be interesting to see
how T-araC would behave in such a system. Alternately
or in addition, the relative absence of T-araC at the 30-
terminus of DNA could result from rapid excision of
these residues via cellular exonuclease activity. For
example, if pol a incorporated T-araC into DNA with
consequent chain termination, pol a might dissociate and
a cellular exonuclease could then excise the T-araC.
Whatever the mechanism, these data demonstrate that
while the use of isolated systems are essential to under-
standing the effects of the analogs on DNA polymerases
and DNA synthesis, they are also limited in their
ability to approximate what occurs in whole cells and
tissues.
[4] Merriman RL, Hertel LW, Schultz RM, Houghton PJ, Rutherford PG,
Tanzer LR, et al. Comparison of the antitumor activity of gemcitabine
and AraC in a panel of human breast, colon, lung, and pancreatic
xenograft models. Invest New Drugs 1996;14:243–7.
[5] Parker WB, Shaddix SC, Rose LM, Waud WR, Shewach DS, Tiwari
KN, et al. Metabolism of 40-thio-B-D-arabinosylcytosine in CEM cells.
Biochem Pharmacol 2000;60:1925–32.
[6] Tiwari KN, Shortnacy-Fowler AT, Cappellacci L, Parker WB, Waud
WR, Montgomery JA, et al. Synthesis of 40-thio-B-D-arabinofurano-
sylcytosine and comparison of its anticancer activity with that of AraC.
Nucleosides Nucleotides Nucleic Acids 2000;19:329–40.
[7] Waud WR, Gilbert KS, Shepherd RV, Montgomery JA, Secrist JA..
Preclinical antitumor activity of 40-thio-beta-D-arabinofuranosylcyto-
sine. Cancer Chemother Pharmacol 2003;51:422–6.
[8] Kornberg A, Baker T. DNA Replication. San Francisco: W.H. Free-
man; 1992.
[9] Budd ME, Cambell JE. DNA polymerases delta and epsilon are
required for chromosomal replication in Saccharomyces cerevisae.
Mol Cel Biol 1993;13:496–505.
[10] Sheaff R, Kuchta RD, Ilsley D. Calf thymus DNA polymerase a-
primase: ‘‘communication’’ and primen template movement between
the two active sites. Biochemistry 1994;33:2247–54.
[11] Copeland WC, Wang TSF. Catalytic subunit of human DNA poly-
merase alpha overproduced from baculovirus-infected insect cells. J
Biol Chem 1991;266:22739–48.
[12] Stadlbauer S, Brueckner A, Rehfuess C, Eckerskorn C, Lottspeich F,
Forster V, et al. DNA replication in vitro by recombinant DNA
polymerase alpha-primase. Eur J Biochem 1994;222:781–93.
[13] Zerbe LK, Goodman MF, Efrati E, Kuchta RD. Abasic template
lesions are strong chain terminators for DNA primase but not for
DNA polymerase alpha during the synthesis of a new DNA strand.
Biochemistry 1999;38:12908–14.
The results from the translesion syntheses studies
demonstrated that the presence of dFdC, araC, or T-araC
in internucleotide linkages alters pol a synthesis. Since the
concentration range of the normal nucleotide triphosphates
used in these studies encompassed the physiological range
and these modified nucleosides are present in internucleo-
tide linkages in vivo following therapeutic doses of the
compound, the effects on polyermase a and the pol a-
primase complex may occur in cells and tissues. This
pausing of translesion synthesis in turn may be enough
to trigger pathways in the cell that would lead to apoptosis
observed in xenograft models.
[14] Beardsley GP, Mikita T, Klaus MM, Nussbaum AL. Chemical synth-
esis of DNA oligomers containing cytosine arabinoside. Nucleic Acids
Res 1988;16:9166–76.
[15] Richardson FC, Richardson KK, Kroin JS, Hertel LW. Synthesis and
restriction enzyme analysis of oligodeoxyribonucleotides containing
the anti-cancer drug 20,20difiuoro-20-deoxycytidine. Nucleic Acids Res
1992;20:1763–8.
[16] Richardson FC, Zhang C, Lehrman SR, Koc H, Swenberg JA, Richard-
son KA, et al. Quantification of 20-fluoro-20-deoxyuridine and 20-
fluoro-20-deoxycytidine in DNA and RNA isolated from rats and
woodchucks using LC–MS/MS. Chem Res Toxicol 2002;15:922–6.
[17] Boosalis MS, Petruska J, Goodman MF. DNA polymerase insertion
fidelity: gel assay for site-specific kinetics. J Biol Chem 1987;262:
14689–96.

2346
K.A. Richardson et al. / Biochemical Pharmacology 68 (2004) 2337–2346
[18] Sheaff R, Kuchta RD. The mechanism of calf thymus DNA primase:
slow initiation, rapid polymerization and intelligent termination.
Biochemistry 1993;32:3027–37.
sphate and its potentiation by 6-mercaptopurine ribonucleoside-50-
monophosphate. Biochemistry 1980;19:215–23.
[26] Reid R, Mar EC, Huang ES, Topal MD. Insertion and extension of
acyclic, dideoxy and ara nucleosides by herpesviridae, human alpha
and human beta polymerases. J Biol Chem 1988;263:3898–904.
[27] Mikita T, Beardsley GP. Functional consequences of the arabinosyl-
cytosine structural lesion in DNA. Biochemistry 1988;27:4698.
[28] Perrino FW, Mekosh HL. Incorporation of cytosine arabinoside mono-
phosphates into DNA at internucleotide linkages by human DNA
polymerase alpha. J Biol Chem 1992;267:23043–51.
[19] Kuchta RD, Ilsley D, Kravig KD, Schubert S, Harris B. Inhibition of
DNA primase and polymerase alpha by arabinofuranosylnucleoside
triphosphates and related compounds. Biochemistry 1992;31:4720–8.
[20] Kufe DW, Major PE, Egan M, Beardsley GP. Correlation of cytotoxi-
city with incorporation of AraC into DNA.
J Biol Chem
1980;255:8997–9000.
[21] Huang P, Chubb S, Hertel LW, Grindey GB, Plunkett W. Action of
20,20-difluorocytidine on DNA synthesis. Cane Res 1991;51:6110–7.
[22] Thompson H, Kuchta RD. Arabinofuranosyl nucleotides are not chain-
terminators during initiation of new strands of DNA by DNA poly-
merase a-primase. Biochemistry 1995;34:11198–203.
[29] Major PP, Egan EM, Beardsley GP, Minden MD, Kufe DW. Lethality
of human myeloblasts correlates with the incorporation of arabinofur-
anosylcytosine into DNA. Proc Natl Acad Sci USA 1981;78:3235–9.
[30] Kufe DW, Major PP, Munroe D, Egan M, Herrick D. Relationship
between Incorporation of 9-beta-D-arabinofuranosyladenine in L1210
DNA and cytoxicity. Cane Res 1983;43:2000–4.
[23] Kuchta RD, Willhelm L. Inhibition of DNA primase by 9-beta-D-
arabinofuranosyladenosinetriphosphate. Biochemistry 1991;30:797–
803.
[31] Richardson F, Black C, Richardson K, Franks A, Wells E, Karimi S, et
al. Incorporation of OSI-7836 into DNA of Calu-6 and H460 Xeno-
graft Tumors. Cancer Chemother Pharmacol 2003, in press.
[32] Wills PW, Hickey R, Malkas L. AraC differentially affects multi-
protein forms of human cell DNA polymerase. Cancer Chemother
Pharmacol 2000;46:193–203.
[24] Catapano CV, Perrino FW, Fernandes DJ. Primer RNA chain termina-
tion induced by arabinofuranosyl-2-fluoroadenine 50-triphosphate. J
Biol Chem 1993;268:7179–85.
[25] Lee MYWT, Byrnes JJ, Downey KM, So AG. Mechanism of inhibition
of DNA synthesis by 1-beta-D-arabinofuranosyladenosine tripho-