Studies in Mouse L-cells on the Incorporation of 1-ß-D-
Arabinofuranosylcytosine into DNA and on Inhibitionof DNA Polymerase by 1-0-o-Arabinofuranosylcytosine5'-Triphosphate*
F. L. Graham2 and G. F. Whitmore
Department of Medical Biophysics, University of Toronto, and the Ontario Cancer Institute, Toronto, Ontario, Canada
SUMMARY
Studies have been carried out in mouse L-cells on theincorporation of l-|3-D-arabinofuranosylcytosine (ara-C) intoDNA and on the inhibition of DNA polymerase by the5'-triphosphate of ara-C (ara-CTP) to determine whether either
of the two current models, incorporation into DNA orinhibition of DNA polymerase, could account for ara-C action.With a modification of the McGrath-Williams technique, it wasfound that ara-C was initially incorporated into small(Okazaki) pieces of DNA but shifted into longer DNA strandswhen cells were washed and incubated in a medium free ofara-C-3H. On degradation of DNA from ara-C-3 H-labeled cells
with micrococcal nuclease and spleen phosphodiesterase, itwas found that most of the ara-C appeared to be ininternucleotide rather than terminal linkages, suggesting thatchain elongation is not stopped by the addition of ara-C to agrowing strand. Studies on ara-C incorporation into nucleicacids failed to show any correlation between the amount ofincorporation and the degree of lethality. With crude extractsof L-cells, it was found that ara-CTP was a competitiveinhibitor of DNA polymerase, and values of 9.0 ±4.3 and 8.7±5.2 X 10~6M were obtained for the Michaelis-Menten
constants of dCTP and ara-CTP, respectively. Calculationsbased on these values and on measured values of the dCTPand ara-CTP concentrations in vivo indicated that thepredicted inhibition of DNA synthesis was significantlysmaller than that actually observed in whole cells. Nevertheless, an evaluation of all of the available data suggeststhat the most plausible model for the action of ara-C is thatDNA synthesis is inhibited by inhibition of DNA polymerase.
INTRODUCTION
The previous paper (13) described studies on viability,growth, and DNA synthesis of mouse L-cells exposed to
ara-C.3 Our results were consistent with a model in which
inhibition of DNA synthesis is the result of inhibition ofDNA polymerase by ara-CTP and were inconsistent withinhibition being the result of incorporation into DNA. Thepresent paper contains the results of further studies designedto determine whether inhibition of DNA synthesis was theresult of inhibition of DNA polymerase or ara-C incorporation into DNA. Momparler (23) has obtained evidence whichsuggests that inhibition of DNA synthesis might be the resultof incorporation of ara-C into the 3'-hydroxyl terminal of
the newly synthesized strand. He found that the ara-Cincorporated by a partially purified calf thymus DNA polymerase was confined almost exclusively to the 3'-hydroxyl
terminal of the DNA, suggesting that such incorporationblocked further elongation of the chain.
We have undertaken a series of experiments to determinewhether a similar observation could be made in whole cellsand to determine whether any correlation could be madebetween incorporation of ara-C into nucleic acids and loss ofviability.
Furth and Cohen (10) have shown that ara-CTP is acompetitive inhibitor of partially purified calf thymus DNApolymerase and measured the Michaelis-Menten constants fordCTP (Km) and ara-CTP (K¡).In order to determine whetherthe inhibition observed in whole cells could be predictedfrom the inhibition of DNA polymerase observed in vitro,we felt that is was necessary to measure not only Km and K¡but also the concentrations of dCTP and ara-CTP in ara-C-treated cells.
MATERIALS AND METHODS
Materials
All experiments to be described in this paper were performed with mouse L-cells, strain L60T (36). Techniques for
'Supported by the National Cancer Institute of Canada.2Fellow of the National Cancer Institute of Canada.
Received February 19, 1970; accepted July 2, 1970.
3The abbreviations used are: ara-C, l-(3-D-arabinofuranosylcytosine;ara-CTP, the 5'-triphosphate of ara-C; TCA, trichloroacetic acid; PBS,
the maintenance and treatment of cell cultures have beendescribed (13). Deoxycytidine-3H, labeled in the 5-positionwith specific activity 15.5 Ci/mmole and thymidine-3H
labeled in the methyl group with specific activity 17.4Ci/mmole were purchased from the Radiochemical Centre,Amersham, England. ara-C-3H, labeled generally with a
specific activity of 11 Ci/mmole, was purchased from NewEngland Nuclear Corp., Boston, Mass. Thymidine-2-14C, 54mCi/mmole, and ara-C-3H, labeled nominally (more than
95%) in the 5-position with a specific activity of 26Ci/mmole, were obtained from Amersham/Searle Corp., DesPlaines, 111.ara-C-3H was always purified less than 1 week
prior to use by descending paper chromatography with1-butanol: water (86:14) (20) (System 1). dCTP-5-3H(specific activity, 27.3 Ci/mmole) and TTP-methyl-3H
furanosylcytosine hydrochloride was purchased from theSigma Chemical Co., St. Louis, Mo. All other nucleosideswere purchased from General Biochemicals, Inc., ChagrinFalls, Ohio. The 5'-monophosphate of ara-C was synthesized
from ara-C and 2-cyanoethyl phosphate by the Tenermethod (35), and ara-CTP was synthesized from the 5'-
monophosphate of ara-C with the partially purified rat liverkinase preparation described by Maley et al. (19). All othernucleotides were obtained from P—L Biochemicals, Inc.,Milwaukee, Wis. Micrococcal nuclease (Staphylococcusaureus), spleen phosphodiesterase (bovine spleen), alkalinephosphatase (Escherichia coif), RNase-free DNase (DNase I,bovine pancreas), and RNase (bovine pancreas) were purchased from Worthington Biochemical Corp., Freehold, N.J.Bacteriophage X DNA was the generous gift of Dr. M. Goldand Mr. S. McClure.
Velocity Sedimentation of DNA
Method 1. One preliminary experiment (Chart 1) wasperformed using a variation of the McGrath-Williamstechnique (22) described by Sambrook et al. (29). Approximately IO6 labeled cells were layered onto alkaline sucrose
gradients, which were incubated overnight (14 to 16 hr) at4°,then centrifuged for 6 hr at 76,000 X g in an SW 25.1
rotor. The gradients were then fractionated into 1.0-mlfractions, which were precipitated with ice-cold 5% TCA,filtered onto Whatman FG/C glass fiber filters, washed with95% ethanol, dried, and counted in a toluene-based scintillation fluid on a liquid scintillation counter.
Method 2. The size distribution of DNA labeled withara-C-3 H was studied with a variation of Method 1, details of
which will be published elsewhere (M. McBurney, F. L.Graham, and G. F. Whitmore, in preparation). By allowingvery gentle lysis of cells after they have been layered on thegradient, this technique permits the sedimentation of high-molecular-weight DNA (approximately 400 to 500 S, compared with 100 to 120 S obtained by Method 1). Labeledcells (approximately IO6 cells/gradient) were layered onto
alkaline sucrose gradients which were incubated overnight asin Method 1 and centrifuged for 150 min at 95,000 X g in
an SW 27 rotor. The gradients were then fractionated, filtered,and counted as in Method 1.
The sedimentation coefficient of the rapidly sedimentingDNA was obtained from the equation (2)
ß-DS20,w (rpm)2,
where D is the distance sedimented and t is time and wherethe constant ßwas determined for our gradients bycentrifuging bacteriophage X DNA and using Studier's value
of 40.1 S for the sedimentation coefficient of Xdg DNA (33)in alkali.
Incorporation of Labeled Compounds
The determination of incorporation of labeled nucleosidesinto acid-soluble intracellular pools and into acid-insolublematerial was made in the following way. Labeled cells werecentrifuged and washed twice with ice-cold PBS (9), and thepellet was extracted 3 times with 0.5 ml of ice-cold 0.2 NHC104. The supernatants from these 3 extractions were thenpooled and neutralized with 2.0 N KOH, and aliquots wereanalyzed by descending chromatography on Whatman No.3MM paper with 95% ethanol:! M ammonium acetate(75:30), pH 7.5 (36), (System 2). The radioactive compounds were then located and identified as previouslydescribed (13). The acid-insoluble pellet was washed 2 moretimes with 0.2 N HC1O4, dissolved in 1.0 ml of 0.5 N NaOH,incubated for 36 hr at 37°to hydrolyze the RNA, cooled on
ice, and acidified by the addition of 75 ¿dof 12 N HC1O4,and the resulting precipitate was removed by centrifugation.This precipitate was sensitive to DNase and insensitive toRNase. The supernatant (RNA) was decanted, and theprecipitate (DNA) was redissolved in 0.5 N NaOH, reprecipi-tated with HC1O4, filtered onto GF/C glass fiber filters,washed with ice-cold 5% TCA, dried, and counted. Chromatography of acid-soluble extracts from ara-C-3H-labeled cells
was also carried out with 95% ethanol: l M ammoniumacetate saturated with sodium tetraborate (75:30) (System3). In one experiment, the acid-soluble extract from cellslabeled for 4 hr with ara-C-3H was heated for 15 min at 80°
in 0.5 N HC104, neutralized and further hydrolyzed withalkaline phosphatase, and chromatographed on System 1.More than 96% of the radioactivity moved with the same Rpas ara-C and was well separated from deoxycytidine, cyti-dine, cytosine, deoxyuridine, uracil, and thymidine. In oneexperiment, the precipitates resulting from acidification ofthe alkaline incubation mixture were digested with micro-coccal nuclease and spleen phosphodiesterase by the methodof Josse et al. (14). Aliquots of the resulting digest werethen chromatographed on System 2 to determine theamount of radioactivity released in the nucleoside form, andthe remainder was further hydrolyzed by alkaline phosphatase and chromatographed for 72 hr with 1-butanol:5%sodium tetraborate in water (86:14) (18) (System 4) toidentify the incorporated radioactivity with ara-C.
Measurements on total endogenous dCTP pools in L-cellswere made as follows. L-cells were centrifuged, washed twicewith ice-cold PBS, and extracted 3 times with 1.0 ml ofice-cold 0.2 N HC104. Along with the first 1.0 ml of HC1O4was added 1.0 juCi of dCTP-3H (specific activity, 27.3
Ci/mmole) to serve as a chromatography marker and as arecovery standard. The acid-soluble extract was neutralizedwith 2.0 N KOH immediately after decanting from theacid-insoluble pellet, and, after removal of the insolubleKC104, it was chromatographed on a Dowex 1-X8 (C03)(200 to 400 mesh) column, eluting with triethylammoniumbicarbonate (19). The dCTP fraction was then evaporated todryness, heated in 0.5 N HC1O4 at 80° for 15 min,
neutralized with KOH, and treated with alkaline phosphataseto hydrolyze it further to the deoxynucleoside. A finalchromatography was performed on System 1 to separatedeoxycytidine from any other deoxynucleosides which maynot have been eliminated by the column chromatography.Finally, the deoxycytidine was assayed for total deoxycytidine content with the microbiological assay describedbelow and assayed for radioactivity by counting a smallaliquot on a scintillation counter. Thus, from the finalspecific activity and the specific activity and amount ofdCTP-3H added initially, the total amount of dCTP original
ly present in the cells could be calculated by isotopedilution.
The microbiological assay for deoxycytidine was carriedout with Lactobacillus acidophilus R26 (ATCC 11506) asdescribed by Siedler et al. (31), except that the assay volumewas 0.5 ml and growth was measured by counting on aModel A Coulter Counter with a 30-/n aperture (CoulterElectronics, Hialeah, Fla.). With this modification, the assaywas capable of measuring 2 X 10~12 mole of deoxycytidineand was linear from 0 to 100 X lo"12 mole/assay tube.
DNA Polymerase Assay
DNA polymerase activity was assayed in cell-free lysates ofL-cells by the methods described by Gold and Helleiner (12).The reaction mixture contained 20 Amóles of phosphatebuffer, pH 7.5; 2 Amólesof 2-mercaptoethanol; 2 /nmoles ofMgCl2; 120 Mg of heat-denatured calf thymus DNA; 60mpmoles each of dATP, dGTP, and TTP-3H (IO4cpm/m/itmole); varying amounts of dCTP and ara-CTP; and0.05 to 0.2 mg of extract protein in a total volume of 0.3ml. The reaction mixture was incubated at 37°for 30 min,
then the reaction was stopped by the addition of 10 ¿/molesof Na4P207 and ice-cold 5% TCA, and the mixture wasfiltered onto glass fiber filters, which were dried and countedon an Ansitron scintillation counter.
RESULTS
Incorporation of ara-C into DNA. In spite of evidencepresented in the previous paper (13) that ara-C was probablynot irreversibly inhibiting DNA synthesis by incorporation, itwas of interest to examine in more detail the incorporation
of ara-C into nucleic acids. Recently, Momparler (23) haspresented evidence that, in vitro, ara-C incorporationappeared to be limited to the 3'-hydroxyl position of the
DNA strand, possibly implying that the further addition ofnucleotides to ara-C-terminated strands is blocked. Considerable evidence now exists that DNA is initiallysynthesized in the form of short (Okazaki) pieces (26, 28,30, 34) which are joined together at a later stage. If, in vivo,ara-C incorporation prevented further addition of deoxy-nucleotides to newly synthesized strands, then it might beexpected to block the elongation of Okazaki pieces, and theincorporated ara-C would never be found in large DNA. Oneway to determine whether this was true was to determinethe size distribution of DNA containing ara-C-3H.
The results shown in Chart 1 indicate that in L-cells a shortpulse of thy nudine-3 H results in incorporation primarily into
small DNA (20 to 30 S), which is later converted into largematerial (120 S). To measure the effect of ara-C, cells werelabeled with ara-C-3 H for 2 hr, centrifuged, either washedwith ice-cold PBS and stored at 0°or resuspended in fresh
medium free of ara-C and containing deoxycytidine, andincubated for various times to allow DNA synthesis toresume and Okazaki pieces to elongate, if possible. Finally,all cell samples were layered onto alkaline sucrose gradients,centrifuged, and fractionated as described in "Materials and
1
400
300
200
100
0
a)
10 15 20 25 30
b)12,000
9,000
6,000
3,000
Chart 1. The sedimentation properties of (a) L-cell DNA labeled by a1-min pulse with thymidine- H, 1.0 nCi/ml (specific activity, 17.4Ci/mmole) and (o) Inceli DNA from cells labeled 1 min withthymidine-3 H, then centrifuged and incubated for 2 hr in fresh medium
containing deoxycytidine and thymidine at 3 mM. Labeled cells werelayered onto alkaline sucrose gradients according to Method 1, thencentrifuged for 5 hr at 23,000 rpm in an SW 25.1 rotor. Fractions werecollected from the top, precipitated with ice-cold 5% TCA, filtered, andcounted.
Methods" (Method 2). Two ara-C concentrations were usedin these experiments: 2.5 X 10~6 M, which is nonlethal
during a 2-hr treatment, although DNA synthesis is severelyinhibited (see Charts 3 and 7 of Ref. 13), and 2.5 X IGT5 M,
which induces loss of viability of S phase cells within 2 hr(Chart 6 of Ref. 13). The results of these experiments areshown on Charts 2 and 3. For purposes of comparison,Chart 2a shows the sedimentation profile of DNA fromcontrol cells labeled for 16 hr with thymidine-2-14C in theabsence of ara-C. Chart 26 shows the sedimentation profileof DNA labeled by a 2-hr treatment with ara-C-3H at 2.5 XIO"6 M (specific activity, 1.4 Ci/mmole). Almost all of the
counts are found at the top of the gradient, as expected ifara-C is initially incorporated into small DNA strands and ifthe continued presence of ara-C blocks further DNAsynthesis. Chart 2c shows the result of washing the cells freeof unincorporated ara-C after a 2-hr treatment andincubating them for 4 hr at 37° in medium containing
deoxycytidine (pulse-chase) before layering the cells onto thealkaline sucrose gradient. Within 4 hr most of the incorporated radioactivity has shifted into material with asedimentation coefficient similar to the control. Chart 3shows a similar experiment carried out at 2.5 X 105 M
4.000
2.000
I
20Fraction number
Bottom
Chart 2. The sedimentation properties of (a) L-cell DNA from cellslabeled for 16 hr with thymidine-2-' 4C, (ft) DNA from cells labeled for2 hr with ara-C-3H at 2.5 X IO"6 M (specific activity, 1.4 Ci/mmole),
and (c) DNA from cells labeled 2 hr with ara-C as above, thencentrifuged and resuspended in fresh medium containing deoxycytidineat 10 mM and incubated for an additional 4 hr. Labeled cells werelayered onto alkaline sucrose gradients by Method 2 and centrifuged at23,000 rpm in an SW 27 rotor. Fractions were collected from the top,precipitated with ice-cold 5% TCA, filtered, and counted.
1200
eoo
400
o
300
200
100
o
ISO
no
50
o)
b)
c)
••••^^•^^^
oTop
10 20
Fraction number
30Bottom
Chart 3. Alkaline sucrose sedimentation of (a) DNA from cells labeledfor 2 hr with 2.5 X 10~s M, ara-C-3H (specific activity, 1.4 Ci/mmole),(ft) DNA from cells labeled with ara-C-3H as in a, then centrifuged,
resuspended, and incubated for 2 hr at 37 in fresh medium containingdeoxycytidine at 10 mM, and (c) DNA from cells labeled withara-C- H, then incubated for 6 hr in fresh medium containing 10 mMdeoxycytidine. Labeled cells were layered onto gradients andcentrifuged, and the fractions were collected, filtered, and counted as inChart 2.
ara-C, a lethal treatment, which gave qualitatively similarresults.
Thus both at a lethal and a nonlethal concentrationincorporation of ara-C into Okazaki pieces did not preventthe subsequent elongation of a large fraction of these piecesonce the cells were removed from the presence of ara-C andDNA synthesis was allowed to resume. The only apparentdifference between results obtained at 2.5 X 1(T6and at 2.5X 10~s M is that the initial distribution of counts is more
sharply distributed towards the top of the gradient for 2.5 X10"s M and may take slightly longer to shift to faster
sedimenting material. Some of the radioactivity remaining atthe top of the gradients may be due to ara-C-3 H in
corporated into RNA which has been only partially degradedby the alkali.
The above results strongly suggested that DNA strandsterminating in ara-C could be further elongated by theaddition of nucleotides to the 3' end, but there is at least one
other possible interpretation. For instance, the shift ofradioactivity from very short pieces into high molecularweight DNA could be the result of the joining together ofOkazaki pieces not terminating in ara-C with the 5' end of astrand containing ara-C at its 3' terminus, as illustrated in
Chart 4. A possible mechanism by which ara-C- H incorporated intoOkazaki pieces could be chased into high-molecular-weight DNA in theabsence of addition of nucleotides to ara-C-terminated strands.
To determine more directly whether nucleotides could beadded to ara-C-terminated DNA, the following experimentwas performed. A culture of cells was exposed to ara-C-3Hat 1.0 X IO6 M (specific activity, 1.4 Ci/mmole), and after
4 hr the culture was split, and 1 part was centrifuged,washed in ice-cold PBS, and set aside for subsequent assay.The other part was centrifuged, resuspended in fresh mediumcontaining 1CT4M deoxycyUdine, and incubated at 37°.At
10 hr, 6 hr after washing the cells, the remaining cells werecentrifuged and washed with PBS and, along with the cellsremoved at 4 hr, extracted with HC1O4 and treated withNaOH as described in "Materials and Methods." The DNA
which precipitated on addition of HC104 to the alkalinesolution was finally digested with micrococcal nuclease andspleen phosphodiesterase according to the method of Josseet al. (14). If the incorporated ara-C was confined to the3'-hydroxyl position of the DNA, then it would be expected
that all the radioactivity would be released in the nucleosideform. Instead, when aliquots of the digests were chroma-tographed on System 3, it was found that only 4% of theradioactivity from the cells harvested at 4 hr was released asara-C and only 2% from cells harvested at 10 hr. To ensurethat the radioactivity was ara-C-3 H and to determine
whether the nuclease digestion had gone to completion, theremainder of the hydrolysate resulting from the micrococcalnuclease and spleen phosphodiesterase digestion was furtherdegraded with alkaline phosphatase and chromatographed onSystem 4. The data from one of the resulting chromatogramsare illustrated in Chart 5, and a comparison of the radioactivity in the nucleoside peak with the total radioactivityindicates that the digestion was approximately 73%complete. The radioactivity in the nucleoside form can beidentified as ara-C. Thus, at least 70% of the incorporatedara-C was not located at the 3'-hydroxyl terminus of the
DNA but rather was incorporated in internucleotide linkage.The results of the enzymatic digestion described above aresummarized in Table 1. The observed incorporation of ara-Cinto internucleotide linkage was expected from the experiments on Okazaki pieces, but is not in agreement withobservations made in vitro by Momparler (23).
One important difference between the conditions in wholecells and those of Momparler's experiments in vitro is that
whole cells contain dCTP at significant concentrations (3 to4 m¿imoles/108 cells), as will be shown later, while in vitro,
1200
1000
800
600
400
200
.CR,
.UR.
COR UOR TdR
10 15 2O 25Distance from origin! inches)
30 35
Chart 5. Chromatography of products resulting from the successivedigestion of ara-C- H-labeled DNA by micrococcal nuclease, spleenphosphodiesterase, and alkaline phosphatase. An aliquot of the finaldigest was chromatographed on Whatman No. 3MM paper for 72 hr,with l-butanol:5% sodium tetraborate in water (86:14), and thechromatogram was cut into strips and counted as previously described(13). Horizontal bars, ara-C, cytosine (CR), deoxycytidine (CdR),deoxyuridine (UdR), thymidine (TdR), uridine (UR), cytosine (C),and uracil (U). Radioactivity remaining at the origin presumablyrepresents oligonucleotides resulting from incomplete digestion bymicrococcal nuclease and spleen phosphodiesterase.
Table 1
Enzymatic digestion of ara-C-3H-labeled DNAl^cells were treated for 4 hr with 1.0 tiM ara-C-3H (specific activity,
1.4 Ci/mmole) or treated for 4 hr followed by a 6-hr chase with lö~4Mdeoxycytidine, then centrifuged, washed with ice-cold PBS, andextracted with 0.2 N HC1O4.The acid-insoluble fraction was incubatedfor 36 hr at 37 in 0.5 N NaOH. Finally, the DNA was digested withmicrococcal nuclease and spleen phosphodiesterase, according to themethod of Josse et al. ( 14). Aliquots were chromatographed on System2 to determine the proportion of the radioactivity in the nucleosideform, and some of the digest was further hydrolyzed with alkalinephosphatase and chromatographed on System 4 to identify the radioactivity as ara-C and to determine the efficiency of degradation bymicrococcal nuclease and spleen phosphodiesterase.
nucleotide linkage within the polydeoxynucleotide chain if asmall amount of dCTP is provided during the reaction orshortly after.
On the basis of the results shown in Charts 2 and 3 and thefact that ara-C is not confined to the 3'-hydroxyl terminal of
DNA, it seems unlikely that inhibition of DNA synthesis isdue to incorporation. It could still be responsible for thelethal effect of ara-C by altering the DNA in some detrimental way. For instance, it has been shown that the 2'
position of the nucleoside sugar may be an important factorin the determination of the conformation of nucleic acids (1,21), and therefore it is possible that the 2'-hydroxyl of the
arabinose sugar could distort the DNA molecule sufficientlyto be lethal. Since it was known that exposure to concentrations below 4 X IO"6 M ara-C over a 4-hr period wasnonlethal for L-cells, while concentrations above 7 X 10~6 M
were toxic for S phase cells within 2 hr (13), an attempt wasmade to observe a correlation between incorporation andlethality.
When cells were exposed to ara-C-3 H at various con
centrations for 4 hr, the incorporation into DNA and RNA,as illustrated in Chart 6, was already quite appreciable atconcentrations of only 0.1 and 0.2 X IGT6M relative to the
incorporation observed at lethal concentrations. Yet in theprevious paper (Ref. 13, Chart 1), it was shown that, at 0.3X IO"6 M ara-C, L-cells were able to multiply for many
weeks, with no sign of any cumulative damage, althoughtheir doubling time was increased. At the higher ara-Cconcentrations shown in Chart 6 of the present paper, theincorporation into DNA levels off and even appears todecrease at 2.0 X 10~s M, a concentration which is rapidly
lethal for S phase cells. Thus the amount of ara-C incorporation into nucleic acids is apparently not directlycorrelated with lethality.
O.24
- 0.20 «
- 0.16
- 0.12
- 0.06
- 0.04
22
Inhibition of DNA Polymerase. Since it appeared thatincorporation of ara-C could not account for the inhibitionof DNA synthesis, the possibility that it was acting byinhibiting DNA polymerase was investigated. In this study, 2sets of parameters were determined: the Michaelis-Mentenconstants (Km and K¡)of DNA polymerase for dCTP andara-CTP (necessarily measured in vitro) and the concentrations of dCTP and ara-CTP in whole cells.
The measurements on inhibition of DNA polymerase werecarried out with the procedures described by Gold andHelleiner (12). The reaction mixture contained dGTP, dATP,and TTP-3H at saturating concentrations; the dCTP concentration was varied in the absence of ara-CTP to determineKm and varied with ara-CTP present to determine K¡.In theabsence of any dCTP the rate of incorporation of TTP-3H
was approximately 25% of the rate found in the presence ofall 4 nucleotides and was unaffected by the presence ofara-CTP. This was presumably due to the terminal additionof nucleotides and was subtracted before attempting todetermine Km and K¡.The results of a typical experimentare plotted as a double reciprocal plot in Chart 7. Sincestraight lines can be drawn passing through the same pointon the ordinate, the inhibition is apparently competitive, ashas been shown previously for partially purified calf thymus
-OX)8 -0.04 O O04l/dCTP(IOSM)
0.08 0.12
Chart 6. Incorporation of ara-C- H (specific activity, 11 Ci/mmole)into ara-CTP, RNA, and DNA during a 4-hr treatment at variousconcentrations of ara-C.
Chart 7. Inhibition by ara-CTP of DNA polymerase activity incrude lysates of I^cells. The reaction mixture (0.3 ml) contained 20Mmoles of phosphate buffer, pH 7.5; 2 /¿molesof 2-mercaptoethanol;2 junóles of MgCl2; 120 ng of heat-denatured calf thymus DNA; 60mamóles each of dATP, dGTP, and TTP-3H (IO4 cpm/mfimole); 0.15
mg of extract protein; and varying amounts of dCTP. After 30 minat 37 , the reaction was terminated, and incorporation into acid-insoluble material was determined as described in "Materials andMethods." The incorporation in the absence of dCTP has been sub
tracted from these data. »,no ara-CTP; A, 63 nM ara-CTP; A, 126 MMara-CTP.
DNA polymerase (10). From 7 determinations of Km and6 determinations of K¡values of 9.0 ±4.3 and 8.7 ±5.2X IO"6 M (mean ±95% confidence limits), respectively, were
obtained. Furth and Cohen (10) obtained values of 3 and 1X IO"6 M, respectively, for Km and Kj with calf thymus
DNA polymerase. These values seem significantly lower thanour values obtained with L-cell DNA polymerase.
The intracellular ara-CTP concentration was determined bytreating cells with ara-C-3H and chromatographing aliquots
of the acid-soluble fraction. The data of Chart 6 show thatat low concentrations of ara-C after a 4-hr treatment theintracellular ara-CTP was proportional to the ara-C concentration in the medium, but it deviated from linearityabove 10~s M. Usually 80 to 90% of the acid-soluble pool of
ara-C-containing compounds in cells washed with PBS consisted of ara-CTP. Measurement of the intracellular pool ofdCTP was carried out as described in "Materials andMethods" with approximately IO8 untreated cells or 10scells which had been treated with 2.0 X lo"5 M ara-C for 4hr. The values obtained were 3.2 mamóles of dCTP/108control cells and 4.1 m/nnoles/lO8 ara-C-treated cells, values
which are similar to the pool si/e of total deoxycytidinenucleotides in L5178Y cells (24). Treatment with ara-C, andthe resulting inhibition of DNA synthesis, did not appreciably alter the intracellular dCTP pool, although theseamounts of dCTP are sufficient to provide for only 3 to 4min of DNA synthesis.
If Michaelis-Menten kinetics hold within cells, then from aknowledge of the constants Km and K¡and the concentrations of dCTP and ara-CTP it should be possible to predictwhat inhibition of DNA synthesis would be expected in cellstreated with ara-C. In carrying out the calculation, it wasassumed that ara-CTP and dCTP were uniformly distributedover the total cell volume, which from the volume of apacked cell pellet and from pulse height analysis of L-cellsuspensions was found to be 0.15 ml/108 cells. Since the
data described in the preceding paragraph indicated only aslight variation of dCTP concentration with changes in ara-Cconcentrations, it was also assumed that the concentration ofdCTP was independent of the concentration of ara-CTP.
The inhibition actually observed in .cells treated for 4 hrwith ara-C was determined from experiments on thy-midine-3H incorporation of the type described in the
previous paper (See Ref. 13, Chart 7). Chart 8 of this paperillustrates the predicted and observed inhibitions plotted as afunction of the ara-C concentration in the medium. Asidefrom the log-log scale, which was chosen to permit the use ofdata obtained over a wide range of ara-C concentrations, thevalues are plotted according to the method of Dixon (7). Ifthe inhibition follows Michaelis-Menten kinetics, then:
where v and v¡are the control and inhibited rates of DNAsynthesis, respectively, and S and / are the concentrations ofsubstrate (dCTP) and inhibitor (ara-CTP). The curve for theobserved inhibition of DNA synthesis is approximatelyparallel to the predicted inhibition suggesting that in wholecells inhibition of DNA synthesis, expressed as (v/i>/) - 1, is
Chart 8. A comparison between the observed inhibition of DNAsynthesis in l^cells exposed to ara-C and the inhibition predictedfrom in vitro measurements of inhibition of DNA polymerase byara-CTP. The observed inhibition (X) was calculated from the rate ofincorporation of thymidine- H into acid-insoluble material in cellsexposed for 4 hr to various concentrations of ara-C. Thymidine- Hincorporation was measured as described in Chart 9 of Ref. 13. Thepredicted inhibition of DNA synthesis (*) was calculated with thevalues determined in vitro for Km and Kj and the values determinedin vivo for dCTP and ara-CTP, assuming Michaelis and Mentenkinetics is valid in whole cells. - - -, maximum inhibition which wouldbe predicted within the experimental errors of Km and K¡,i.e., Km =9.0 ±4.3 nM and K¡= 8.7 - 5.2 jiM.
proportional to the ara-CTP concentration; however, theobserved inhibition is 60 times greater than that predictedfrom inhibition of DNA polymerase in vitro. The dashedcurve of Chart 8 indicates that even the maximum inhibitionwhich could be expected within the experimental values ofKm and Ki; i.e.. Km = (9.0 + 4.3) X 1(T6 M, Kj = (8.7-5.2) X 10~6 M, could not completely account for the
observed inhibition. Possible reasons for this discrepancy willbe considered in the discussion to follow.
DISCUSSION
Incorporation of ara-C into DNA. The results of our studieson incorporation of ara-C into the DNA of mouse L-cellsand the studies reported in the previous paper (13) suggestthat incorporation of ara-C into DNA is not the cause ofinhibition of DNA synthesis. As evidence we cite thefollowing facts.
ara-C can severely inhibit DNA synthesis without affectingviability. Thus inhibition must be reversible, since synthesis
must resume in order for the cells to divide and formcolonies.
Incorporation of ara-C into newly synthesized strands ofDNA (Okazaki pieces) does not prevent the subsequentelongation of these pieces when the cells are incubated inthe absence of ara-C.
When DNA labeled with ara-C-3 H was digested with micro-
coccal nuclease and spleen phosphodiesterase, it was foundthat more than 70% of the radioactivity was released as a3'-phosphate and was therefore incorporated into inter-
nucleotide linkage. It is not obvious how ara-C incorporationwould affect DNA synthesis if it does not block the additionof nucleotides to ara-C-terminated strands.
Although ara-C incorporation does not appear to inhibitDNA synthesis, it is still conceivable that incorporationmight be lethal through the introduction of some deleteriousalteration in the DNA structure. However, our resultsindicate that incorporation of ara-C is not correlated withlethality. It has also been suggested that the lethal effect ofara-C might be due to its incorporation into RNA (5). Sinceour results indicate that the incorporation into RNA at lethalconcentrations is not strikingly different from that at non-lethal concentrations, and since ara-C is specifically toxic forS phase cells, while RNA synthesis occurs in all parts of thecell cycle except mitosis (8, 27), it appears that incorporation of ara-C into RNA is not the lethal event.
Inhibition of DNA Polymerase. Of the three current modelsproposed to explain the inhibition of DNA synthesis byara-C, inhibition of CDP reducÃase (3), inhibition of DNAsynthesis by incorporation (23, 32), and inhibition of DNApolymerase (10), the model that ara-C inhibits DNAsynthesis via inhibition of DNA polymerase appears to bemost consistent with the following facts. First, inhibition ofDNA synthesis is apparently reversible, since cells survive andform colonies after an inhibition of DNA synthesis as severeas 97%. Second, ara-C does not appear to interfere with thesynthesis of either dCTP or TIP (at least not from exog-enously supplied deoxynucleosides), yet inhibits theirincorporation into DNA. Third, ara-CTP does inhibit DNApolymerase in vitro. Finally, the concentration of ara-CTPwithin ara-C-treated cells is appreciable relative to the dCTPconcentration. One serious difficulty, however, is thediscrepancy between the observed and predicted inhibitionof DNA synthesis illustrated in Chart 8. The observedinhibition in whole cells is approximately 60 times moresevere than that predicted from our in vitro studies oninhibition of DNA polymerase by ara-CTP. There are anumber of possible explanations for this difference.
ara-C may be inhibiting DNA synthesis by some mechanismother than inhibition of DNA polymerase. As we havementioned previously, ara-C does not appear to block DNAsynthesis by interfering with the production of either dCTPor TTP. The possibility that ara-C affects the synthesis ofdeoxypurines has not been investigated and cannot becompletely disregarded, although we know of no evidencesuggesting that this occurs. The work of Moore and Cohen(25) on the effect of arabinonucleotides on ribonucleotidereduction would suggest that at least ara-C does not interferewith the reduction step in purine biosynthesis.
ara-C: Incorporation-Inhibition Studies
The conditions of the in vitro assay for DNA polymeraseactivity may be so unlike the conditions in vivo that nocomparison can legitimately be made. For instance, in the invitro experiments the concentrations of the 3 deoxynucleo-tides, dGTP, dATP, and TTP, and the concentration of DNAprimer were at saturating levels, while in vivo this conditionmay not hold. (The dCTP concentration has been found tobe approximately 3 X IO"5 M in I^cells. If the concentra
tions of dGTP, dATP, and TTP are comparable or lower,then they would be far below the 2 X 10" M concentration
used in vitro.) In view of this, it may be naive to expectquantitative agreement between in vivo and in vitro observations.
The properties of the DNA polymerase enzyme in vitromay be quite different from its properties in vivo. It hasbeen suggested that in vivo DNA replication may require anenzyme complex containing nuclease, polymerase, and ligaseactivities (11), and it is possible that DNA polymerase mightbe altered by dissocation from this complex. Indeed, recentstudies (6, 16) on E. coli could be interpreted as suggestingthat E. coli DNA polymerase is not responsible for the majorsynthesis of DNA, but may only act as a repair enzyme.
Because of the discrepancy between the observed andpredicted inhibition of DNA synthesis, the model in whichara-C acts by inhibiting DNA polymerase cannot be considered proven. At least part of the observed inhibition ofDNA synthesis, however, can be accounted for on the basisof inhibition of DNA polymerase.
At this point, it seems appropriate to summarize thepossible action of ara-C in terms of a model in whichinhibition of DNA synthesis is the result of inhibition ofDNA polymerase.
ara-C must be converted to ara-CTP in order to inhibitDNA synthesis. The ability of deoxycytidine to protect cellsagainst the effects of ara-C is presumably due, at least inpart, to the prevention of formation of ara-CTP from ara-Cby competition at the kinase level (15, 17). Thus resistanceto ara-C could result from a decrease in activity of deoxycytidine kinase, as has often been observed (4).
ara-CTP inhibits DNA polymerase and, consequently, DNAsynthesis. Since this inhibition is competitive with dCTP, theability of deoxycytidine to protect against ara-C-inducedinhibition of DNA synthesis may be in part due to competition at the polymerase-binding site.
Inhibition of DNA synthesis can cause some type ofdamage (as yet unknown) to cells synthesizing DNA at thetime of induction of the block which permanently preventscell proliferation after removal of the block. The inductionof this damage evidently depends on the severity of theinhibition.
Therefore, it appears that inhibition of DNA polymerasecould account for all of the observed actions of ara-C, and,while additional work is required to prove the model, thisseems the most satisfactory hypothesis at the present time.
ACKNOWLEDGMENTS
We thank Dr. M. Gold for his interest in this work and his helpfuladvice and Dr. J. E. Till and Dr. W. R. Bruce for their comments onthe manuscript.
1. Adler, A. J., Grossman, L., and Pasman, G. D. Circular Dichro-ism of Cytosine Dinucleoside Monophosphates Containing Arabi-nose, Ribose, and Deoxyribose. Biochemistry, 7: 3836-3843,1968.
2. Huii'i. E., and Hershey, A. D. Sedmentation Rate as a Measure ofMolecular Weight of DNA. Biophys. J., 3: 309-321, 1963.
3. Chu, M. Y., and Fischer, G. A. A Proposed Mechanism of Actionof l-/3-D-Arabinofuranosyl-Cytosine as an Inhibitor of the Growthof Leukemic Cells. Biochem. Pharmacol., //: 423-430, 1962.
4. Chu, M. Y., and Fischer, G. A. Comparative Studies of LeukemicCells Sensitive and Resistant to Cytosine Arabinosidc. Biochem.Pharmacol., 14: 333-341, 1965.
5. Chu, M. Y., and Fischer, G. A. The Incorporation of H3-Cytosine
Arabinoside and Its Effect on Murine Leukemic Cells (L5178Y).Biochem. Pharmacol., 17: 753-767, 1968.
6. DeLucia, P., and Cairns, J. Isolation of an E. coli Strain with aMutation Affecting DNA Polymerase. Nature, 224: 1164-1166,1969.
7. Dixon, M. The Determination of Enzyme Inhibitor Constants.Biochem. ¡.,55: 170-171, 1953.
8. Doida, Y., and Okada, S. Determination of Switching-off andSwitching-on Time of Overall Nuclear RNA Synthesis. Nature,216: 272-273, 1967.
9. Dulbecco, R., and Vogt, M. Plaque Formation and Isolation ofPure Lines with Poliomyelitis Viruses. J. Exptl. Med., 99:167-182, 1954.
10. Furth, J. J., and Cohen, S. S. Inhibition of Mammalian DNAPolymerase by the 5'-Triphosphate of 1-0-D-Arabinofuranosyl-cytosine and the 5'-Triphosphate of 9-(3-D-Arabinofuranosyl-
adenine. Cancer Res., 28: 2061-2067, 1968.11. Ganesan, A. T. Studies on the in Vitro Synthesis of Transforming
DNA. Proc. Nati. Acad. Sei. U. S., 61: 1058-1065, 1968.12. Gold, M., and Helleiner, C. W. Deoxyribonucleic Acid Polymerase
in L-cells. I. Properties of the Enzyme and Its Activity inSynchronized Cell Cultures. Biochim. Biophys. Acta, 80:193-203, 1964.
13. Graham, F. L., and Whitmore, G. F. The Effect of 1-0-D-Arabinofuranosylcytosine on Growth, Viability, and DNASynthesis of Mouse L-cclls. Cancer Res., 30: 2627-2635, 1970.
14. Josse, J., Kaiser, A. D., and Kornberg, A. Enzymatic Synthesis ofDeoxyribonucleic Acid. VIII. Frequencies of Nearest NeighborBase Sequences in Deoxyribonucleic Acid. J. Biol. Chem., 236:864-875, 1961.
15. Kaplan, A. S., Brown, M., and BenPorat, T. Effect of 1-/3-D-Arabinofuranosylcytosine on DNA Synthesis. I. In NormalRabbit Kidney Cell Cultures. Mol. Pharmacol., 4: 131-138,1968.
16. Kelly, R. B., Cozzarelli, N. R., Deutscher, M. P., Lehman, I. R.,and Kornberg, A. Enzymatic Synthesis of Deoxyribonucleic Acid.XXXII. Replication of Duplex Deoxyribonucleic Acid byPolymerase at a Single Strand Break. J. Biol. Chem., 245:39-45, 1970.
17. Kessel, D. Some Observations on the Phosphorylation of Cytosine Arabinoside. Mol. Pharmacol., 4: 402-410, 1968.
18. Khym, J. X. The Use of the Borate Complex. In: S. P. Colowickand N. O. Kaplan (eds.), Methods in Enzymology, Vol. 12, Part A,pp. 93-101. New York: Academic Press, Inc., 1967.
19. Maley, F., Maley, G. F., and McGarrahan, J. F. A SimplifiedEnzymic Procedure for the Preparation of Nucleoside Di- andTriphosphates. Anal. Biochem., 19: 265-271, 1967.
20. Markham, R., and Smith, J. D. Chromatographie Studies ofNucleic Acids. I. Technique for the Identification and Estimationof Purine and Pyrimidine Bases, Nucleosides and Related Substances. Biochem. J., 45: 294-298, 1949.
21. Maurizot, J. C., Brahms, J., and Eckstein, F. Forces Involved inthe Conformational Stability of Nucleic Acids. Nature, 222:559-561, 1969.
22. McGrath, R. A., and Williams, R. W. Reconstruction in Vivo ofIrradiated Escherichia coli Deoxyribonucleic Acid: The Rejoiningof Broken Pieces. Nature, 2/2: 534-535, 1966.
23. Momparler, R. L. Effect of Cytosine Arabinoside 5'-Triphosphate
on Mammalian DNA Polymerase. Biochem. Biophys. Res.Commun., 34: 465-471, 1969.
24. Momparler, R. L., Chu, M. Y., and Fischer, G. A. Studies on aNew Mechanism of Resistance of L5178Y Murine Leukemia Cellsto Cytosine Arabinoside. Biochim. Biophys. Acta, 161: 481-493,1968.
25. Moore, E. C., and Cohen, S. S. Effect of Arabinonucleotides onRibonucleotide Reduction by an Enzyme System from RatTumor. J. Biol. Chem., 242: 2116-2118, 1967.
26. Painter, R. B., and Schaefer, A. State of Newly SynthesizedHeLa DNA. Nature, 227. 1215-1217,1969.
27. Pfeiffer, S. E., and Tolmach, L. J. RNA Synthesis in Synchronously Growing Populations of HeLa S3 Cells. I. Rate of TotalRNA Synthesis and Its Relationship to DNA Synthesis. J.Cellular Comp. Physiol., 71: 77-94, 1968.
28. Sakabe, K., and Okazaki, R. A Unique Property of theReplicating Region of Chromosomal DNA. Biochim. Biophys.Acta, 729: 651-654, 1966.
29. Sambrook, J., Westphal, H., Srinivasan, P. R., and Dulbecco, R.The Integrated State of Viral DNA in SV40-transformed Cells.Proc. Nati. Acad. Sei. U. S., 60: 1288-1295, 1968.
30. Schandl, E. K., and Taylor, J. H. Early Events in the Replicationof DNA into Mammalian Chromosomes. Biochem. Biophys. Res.Commun., 34: 291-300, 1969.
31. Siedler, A. J., Nayder, F. A., and Schweigen, B. S. Studies onImprovements in the Medium for Lactobacillus acidophilus in theAssay for Deoxyribonucleic Acid. J. Bacterio!., 73: 670-675,1957.
32. Silagi, S. Metabolism of l-(J-D-Arabinofuranosylcy tosine inL-cells. Cancer Res., 25: 1446-1453, 1965.
33. Studier, F. W. Sedimentation Studies of the Size and Shape ofDNA. J. Mol. Biol., 11: 373-390, 1965.
34. Sugimoto, K., Okazaki, T., and Okazaki, R. Mechanism of DNAChain Growth. II. Accumulation of Newly Synthesized ShortChains in E. coli Infected with Ligase-defective T4 Phages. Proc.Nati. Acad. Sei. U. S., 60: 1356-1362, 1968.
35. Tener, G. M. 2-Cyanoethyl Phosphate and Its Use in the Synthesis of Phosphate Esters. J. Am. Chem. Soc., 83: 159-168,1961.
36. Till, J. E., Whitmore, G. F., and Gulyas, S. DeoxyribonucleicAcid Synthesis in Individual L-strain Mouse Cells. II. Effects ofThymidine Starvation. Biochim. Biophys. Acta, 72: 277-289,1963.
