Studies on the Penetration of Mammalian Cells by Deoxyribonucleoside-5'-Phosphates

'

M. ANWAR WAQAR,2 ROBERT L. TABER AND JOEL A. HUBERMAN Department of Viral Oncology, Roswell Park Memorial Institute, Buffalo, New York 14263

ABSTRACT We have tested the ability of [5'-32Pl-deoxyribonucleoside monophosphates (dNMPs) to penetrate living mouse fibroblast L cells and human HeLa cells. Under the conditions of our experiments, small numbers of apparently intact dNMP molecules appeared to penetrate into the interior of L cells and be incorporated into DNA. This incorporation was not due to mycoplasma contamination nor to extracellular hydrolysis of the dNMPs followed by resynthesis inside the cell. Under these same conditions, penetration of HeLa cells by intact dNMPs did not occur to a significant extent. However, HeLa cells were capable of hydrolyzing extracelluar dNMPs to Pi and deoxyribonucleosides a t a much faster rate than L cells. These experiments provide a starting point for attempts to specifically label the DNA in intact, living eukaryotic cells with [32P1-dNMPs.

I

For investigation of certain aspects of DNA structure arid replication, i t would be useful to be able to label DNA in intact, living cells monophoswith [5'-32P1-deoxyribonucleoside phates (dNMPs). For example, the demonstrations that ribonucleotides are covalently linked to deoxyribonucleotides in the DNA of mammalian cells and papovaviruses (Magnusson et al., '73; Tseng and Goulian, '75; Waqar and Huberman, '75; Anderson et al., '77) have been carried out by nearest neighbor analysis (Lehman e t al., '58) only in cell-free in vitro systems. This demonstration could be extended to living cells were it possible to label t h e nascent DNA of living cells with [5'-32P1dNMPS. Although it has been "common knowledge" t h a t phosphorylated compounds such as nucleotides cannot enter living cells (Roll et al., '56a,b; Leibman and Heidelberger, '55; Reichard, '53; Kerr et al., '51), there is reason to think that a t least a small portion of exogenous nucleotide may be able to enter eukaryotic cells and participate in RNA or DNA synthesis. Mammalian cells can be infected at low efficiency by intact viral nucleic acids (Holland et al., '59). A mutant of yeast has been obtained which incorporates 5'-dTMP into DNA (Wickner, '74). Plunkett et al. ('73, '741, &hen and Plunkett ('75) and Plunkett and &hen ('77) have obtained evidence for uptake of inJ. CELL. PHYSIOL. (1979)101: 251-260

tact 5'-dAMP and 5'-araAMP into mouse fibroblasts (L cells) and for subsequent incorporation of these nucleotides into DNA. They point out that previous workers used such low specific activities of [32P1-nucleotidesthat incorporation a t the efficiencies measured for dAMP and araAMP would have gone undetected. Because still higher specific activities of PP1-nucleotides are now commercially available, we thought it would be worthwhile to check the uptake of [5'-32P1-dNMPsby mammalian cells. This paper reports our finding that although a small portion of intact dNMP is incorporated into L cells, little or no intact dNMP is incorporated into HeLa cells. The major route of incorporation for HeLa cells is by hydrolysis of the dNMP (by cell surface phosphatases) to deoxynucleoside and Pi. Both deoxynucleosides and Pi are readily incorporated. MATERIALS AND METHODS

HeLa S3 cells (from Drs. Paul Atkinson and Sheldon Penman) and mouse fibroblast L cells (from Dr. Robert Hughes) were maintained Received Feb. 12, '79. Accepted June 18, '79. IThis study was supported by Grant PCM77-14451 from the National Science Foundation and by Grant CA-14801 from the National Cancer Institute. 2Author to whom requests for reprints should he addressed.

251

252

M. A. WAQAR, R. L. TABER AND J. A. HUBERMAN

free of mycoplasma in suspension culture a t 37°C in Joklik-modified Minimum Essential Medium (Grand Island Biological Co.), supplemented with 10%fetal calf serum (Microbiological Associates). Serum was pretreated at 56°C for 20 hours before use in the medium. This treatment destroyed most phosphatase activity in the serum, but did not affect its growth-stimulating properties. Cells in exponential growth were centrifuged and resuspended in fresh, prewarmed medium a t concentrations from 2.0 X lo5 to 3.0 x 1 0 6 cells/ml. Cells were incubated with constant stirring in final volumes from 5-50 ml, and 1M HEPES buffer, pH 7.0, was added to give a final concentration of 10 mM. Radioactive dNMPs or dThd were added a t concentrations indicated in figure and table legends. Aliquots were taken as indicated in figure and table legends. [5'-32P]-dTMP,[5'-32P1-dCMP,[5'-32PIdAMP, and [5'-3ZPI-dGMP were prepared from the respective triphosphates in the following way. [a-32PI-dNTPs(4-9 X lo1' cpm/pmole) were purchased from Amersham-Searle Corp. and were degraded to [5'-32P1-dNMPs with venom phosphodiesterase (Boeringer-Mannheim) in the presence of 10 mM MgCl,, 1 mM CaCl,, 10 mM Tris-HC1, pH 9.2, at 37°C for 30 minutes. The reaction was stopped by adding EDTA to a final concentration of 100 mM. The product [5'-32P1-dNMPswere purified by paper (Whatman 3 MM) chromatography in isobutyric acid: 1M NH,OH : 0.1 M EDTA (100:60: 1.6, v/v). After chromatography for 18 hours at room temperature, the chromatogram was dried, the positions of the nucleotides were located under UV lights and the nucleotide bands were cut out, washed twice with isopropanol, then once with ether:acetone (1:l) and dried. The nucleotides were then eluted from the paper with distilled water. (18 Ci/mmole), [meth[Meth~l-~Hl-dTMP ~ l - ~ H I - d T (47 h d Ci/mmole) and 32Piwere purchased from Amersham-Searle Corp. [Methyl-3HI-dTMP was further purified by chromatography as described above. RESULTS

Uptake of 3,P from [32PI-dNMPsinto DNA of L and HeLa cells In order to test whether significant uptake of intact nucleotides occurs in mammalian cells, we grew large numbers (over lo8) of HeLa and L cells in the uresence of each of the [5'-32P1-dNMPs(-1 X- 10" cpmlpmole) for

two hours. We also grew both kinds of cells in the presence of 32Pi (final specific activity, after dilution by the cold Pi in the medium, was 2.6 X lo5cpm/pmole) and in the presence of r3H1-dThd (4 x lo9 cpm/pmole). Nuclei were then isolated from the cells, and DNA was purified from the nuclei, then digested to 5'-dNMPs by DNase I and venom phosphodiesterase. The data obtained for L cells are shown in figure 1 and table 1, while the data for HeLa cells are in figure 2 and table 1. There are several striking features of these results. First, a significant proportion of the incorporated 32Pis found in the 5'-dNMP used for labeling in the case of L cells, but not in the case of HeLa cells. This suggests that incorporation of intact nucleotides may be occurring in L cells. Additional data consistent with this possibility are presented later in this paper. Second, in the case of L cells and especially in the case of HeLa cells, some 32Pis found i n 5'-dNMPs not used for labeling. This suggests that some breakdown of the P2PI-dNMPused for labeling occurred during the 2-hour incubation, according to the formula: (1) [5-32P1-dNMP + H,O -deoxynucleoside

+ 32Pi.

Presumably the resulting 32Pientered the cells' Pi pool and was subsequently incorporated into DNA by normal means. Additional data consistent with this possibility are presented later. The breakdown of the [32P1dNMPs contrasts with the complete recovery of L3HI-dThdin [3HI-dTMP (figs. lF, 2F). Third, the incubation carried out with 32Pi for two hours (figs. lE, 2E) did not produce a n equilibrium distribution of 32Pin the four dNMPs. This unequal labeling is probably due to differences in the pool sizes of the nucleotides and to differences in the rates of the various biochemical reactions needed to incorporate Pi into the four dNTPs. It is well known t h a t longer incubation times (up to 24 hours) are necessary to achieve equilibrium labeling with 32Pi.

Breakdown of dTMP by HeLa cells and L cells In order to test whether the incorporation of 32Pfrom specific [5'-32Pl-dNMPsinto other dNMPs (figs. 1, 2) was due to breakdown of the labeled dNMP (formula (1)above), we incubated medium, serum, HeLa cells, and L cells with L3HI-dTMP for varying lengths of time. We then tested for conversion of [3HIdTMP to L3H1-dThd in the extracellular medium by paper chromatography. The results

PENETRATION OF CELLS BY DEOXYNUCLEOTIDES

253

E n 0

X

cm from Origin Fig. 1 Chromatographic separation of the 5-dNMPs from the DNA of labeled L cells. Exponentially growing L cells were pelleted, then resuspended to a final concentration of 9.0 X lo6 cells/ml in prewarmed medium containing 10 mM HEPES buffer, pH 7.0. The cells were incubated in portions of 50 ml each a t 37" for two hours in the presence of either (A) 18 nM [W-dAMP (8.4 X 1 O ' O cpm/pmole), (B) 14 nM 13TI-dGMP(8.4 x 10'O cpm/pmole), (C) 0.4 nM 13TI-dCMP (1.8 x 10" cpm/pmole), (D) 15 nM 13T1-dTMP (8.4 x 1O'O cmp/pmole), (E) 9.6 mM 3 T i (2.6 X lo5cpm/pmole), or (F) 34 nM IW-dThd (3.8 X lO'cpm/pmole). The cells were then pelleted and washed twice with 50 ml of 10 mM Tris, 10 mM EDTA, 10 mM NaC1, pH 7.5. Their nuclei were prepared, lysed with detergent and treated with Proteinase K and KOH as previously described (Waqar and Huberman, '75). The DNA was purified a s previously described (Waqar and Huberman, '751, then dissolved in 10 mM Tris-HC1, pH 7.0 and digested with DNase I and venom phosphodiesterase. The resulting 5'-dNMPs were separated by paper chromatography (Whatman 3 MM in water-saturated ( N H J ~ 0 , : l . OM sodium acetate:isopropanol, (80:18:2, V/V) (Richardson, '65)).One-centimeter strips were cut from the chromatogram and counted. In section F only 1/200 of the total sample was applied to the paper chromatogram.

presented in table 2 indicate that dTMP is stable in medium alone for up to two hours (1-2% of the label was in dThd in the starting preparation of C3HI-dTMP),but that heat-treated serum must contain a low-level phosphatase activity since a n additional 4% of the dTMP is hydrolyzed to dThd after two hours. HeLa cells contribute a large amount of phosphatase activity: in the presence of serum 88%of the recovered 3H is found in dThd after two

hours. Notice, however, that -40% of the 3His not recovered on the chromatogram after two hours. This low recovery is due to uptake of 3H (as 3H-dThd)by the HeLa cells with resultant loss of 3H from the extracellular medium. The data in table 2 appear at first sight to suggest that L cells stabilize dTMP because, even in the presence of serum, there is no increase in C3H1-dThd in the extracellular medium with time. However, L cells, like HeLa

254

M. A. WAQAR, R. L. TABER AND 3. A. HUBERMAN TABLE 1

Distribution of cpm among the 5'-dNMPs found in the DNA of labeled HeLa or L cells Percent of total cpm in Cells

HeLa

L

Cells labeled with

Total cpm in dNMPs

[3T1-dAMP E3T1-dCMP E3TI-dGMP [3Tl-dTMP I3zpl-PO4 E3H1-dThd [3T1-dAMP [3T1-dCMP E3T1-dGMP [3T1-dTMP mi -PO, [%I-dThd

4,604 1,465 3,643 4,645 4,978 5,284 8,086 534 3,115 3,177 11,905 3,606

dAMP

dCMP

dGMP

29 17 20 23 21 0 92 10 4 11 22

13 20 8 8 8 0 2 42 2 7 15 0

12 10 18 13 15 0 1 2 87 7 14 0

0

dTMP

47

__

53

55 56 56 100 5 45 7 76 50 100

Data are taken from figure 1 (L cells) or figure 2 (HeLa cells).

TABLE 2

Degradation of L3H1-dTMPunder various conditions of incubation Time of incubation, minutes 5

Cells

Serum

Total JH cpm

-

-

+

10,221 10,637

HeLa HeLa L

-

9,837 10,434 11,404 11,132

L 0.2

120 %total 3H

+

-

+

%total 3H

dTMP

dThd

Total 3H cpm

98.7 98.0 85.2 77.3 99.0 98.8

1.3 2.0 14.8 22.7 1.0 1.2

12,149 11,184 6,179 7,746 13,131 11,189

dTMP

dThd

98.3 94.6 23.0 11.6 98.7 98.6

1.7 5.4 77.0 88.4 1.3 1.4

rM IX-dTMP was incubated in Joklik-modified MEM a t 37" for the times indicated, with the additions indicated. Cells,

when present, were at 9.6 X lo5cells/ml. Heat-treated serum, when present, was at 10%.At the indicated times, 0.2-ml samples were taken and the cells were centrifuged. Unlabeled dThd and dTMP were added to the supernatant as optical density markers, and the supernatant was chromatographed on Whatman 3 MM paper as described in MATERIALS AND METHODS. 3Hradioactivity was detected only in the dTMP and dThd repions. Data are presented as the total JH cpm recovered on each chromatogram, and as the percent of that total found in dTMP or dThd.

cells, readily incorporate 3H-dThd (figs. lF, 2F). In this experiment, the L3H1-dThd produced by serum phosphatases and L cell phosphatases (if any) was incorporated into the cells as rapidly as it was produced. The overall rate of hydrolysis of dTMP was obviously lower than in the case of HeLa cells because there was no detectable decrease in the total amount of 3H in the extracellular medium. Uptake of apparently intact nucleotides by L and HeLa cells The data in table 2 provide a rational explanation for the differences between L and HeLa cells demonstrated in figures 1and 2 and table 1.In the case of HeLa cells, breakdown of exo-

genous [32PI-dNMPsto deoxynucleosides and 32Piis very rapid. The 32Piis diluted by the 9.6 mM cold Pi in the medium, then taken up by the cells and incorporated into all four dNTPs which in turn are incorporated into DNA. Thus the 32Pis eventually recovered in all four dNMPs after hydrolysis of the DNA. Cellular uptake of intact 132PI-dNMPsis difficult to detect due to the high background of uptake of 32Piand due to the fact that the extracellular concentration of intact dNMPs is continuously decreasing. In the case of L cells, breakdown of exogenous [32PI-dNMPsis much slower. However, it does occur, and, as with HeLa cells, it produces a background of 32Pin all four dNMPs after

255

PENETRATION OF CELLS BY DEOXYNUCLEOTIDES

8-

24

6-

I

A

4-

-

18 -

I,

12 -

c 8

16

24

32

A

6-

L Y 0

T

40

G

1

I

8

16

24

32

4

cm from Origin Fig. 2 Chromatographic separation of the 5-dNMPs from the DNA of labeled HeLa cells. Exponentially growing HeLa cells were pelleted, then resuspended to a final concentration of 7.8 x lo6 celldm1 in prewarmed medium containing 10 mM HEPES buffer, pH 7.0. All other experimental procedures were identical to those in figure 1.

hydrolysis of the DNA. In the case of L cells, however, apparent uptake of intact nucleotides is sufficiently rapid that significant levels of 32P are recovered in the dNMP used for labeling (fig. 1, table 1).When [32P1-dCMP is used for labeling (fig. 1C), 45%of the recovered 32Pis found in dTMP. The most likely explanation for this large transfer of label from dCMP specifically to dTMP is deamination of dCMP to dUMP, which is then methylated to dTMP (Scarano et al., '63; Plagemann et al., '78). thus the 32Pabove background recovered in dTMP, as well a s that recovered in dCMP, must be included in estimating uptake of intact dCMP. In table 3 we have estimated the number of

apparently intact dNMP molecules incorporated per cell during the 2-hour incubation by subtracting background and then correcting for specific activity and cell number. The fact that only 0.8 molecules of dCMP were incorporated into each L cell (table 3) despite the striking increase in cpm in dCMP and dTMP when [32P1-dCMPwas used for labeling (fig. 1C) suggests that incorporation is concentration dependent. If we assume a linear dependence of uptake on concentration (we emphasize that this assumption is not based on experimental evidence), then the efficiency of dCMP uptake by L cells is comparable to that of the other nucleotides (third column of table 3).

256

M. A. WAQAR, R. L. TABER AND J. A. HUBERMAN TABLE 3

Estimated amounts of intact dNMPs incorporated by HeLa or L cells during a 2-hour incubation Incorporation during 2 ~ h o u rincubation

cells dNMP

HeLa

L

dAMP dCMP dGMP dTMP dAMP dCMP dGMP dTMP

Total cpm

Moleculeslcell

MoleculeslcelllnM

2 0.5 0.7 0 39

0.11 1.3 0.05 0 2.2 1.9 1.0 0.6

368 176 109

0 7,197 336 2,648 1,620

0.8 14 9

Data are taken from figures 1 (L cells) and 2 (HeLa cells) and from table 1. In each case t h e contribution of ?Ti incorporation was estimated by comparison of the pattern of label incorporated into the dNMPs not used for labeling with the pattern of label incorporated directly from T i . We assumed t h a t ?Ti contributed to the same extent to labeling of the nucleotide (in DNA) which was used for labeling. After subtraction of this contribution, t h e incorporated cpm were corrected for specific activity and cell number to provide a n estimate of the number of nucleotides incorporatedlcell. Because incorparation probably depends on external concentration, we have normalized t h e number of nucleotides incorporated per cell to a 1 nM external concentration of nucleotide. assuming a linear relationship between external concentration and incorporation. All specific activities were c o r ~ rected for the 14-day half-life of yT. In the case of labeling with dCMP, 32Pabove background in both dCMP and dTMP was included in the estimate (see text).

The background of 32P-incorporationinto all four dNMPs by HeLa cells is so high that the significance of the values calculated for HeLa cells in table 3 is questionable. It is clear, however, that HeLa cells are much less efficient than L cells at apparent uptake of intact dNMPs. Intracellular synthesis of dNMPs by deoxynucleoside kinases does not contribute significantly to the apparent uptake of intact dNMPs by L cells A possible explanation of the apparent uptake of intact nucleotides demonstrated above is the resynthesis of intact [32P1-dNMP by the appropriate deoxynucleoside kinase inside the cell, from the deoxynucleoside and 32Pi (after conversion to [y32P1-ATP) generated by extracellular hydrolysis of the [32P1dNMP used for labeling. Specific resynthesis of the dNMP used for labeling would occur because the intracellular pool of the corresponding deoxynucleoside would be increased by uptake of the exogenous deoxynucleoside. The reactions involved are, first, extracellular hydrolysis of the dNMP (see formula (111, followed by transport of the resulting deoxynucleoside and 32Pi: (2) Deoxynucleoside (extracellular) -Deoxynucleoside (intracellular) (3) 32Pi (extracellular) 3zPi(intracellular).

-

A portion of the intracellular 32Pi would be converted to [ Y - ~ ~ P I - A Tand P , this could be used by deoxynucleoside kinases to synthesize

[32PI-dNMPs.Because cellular pools of deoxynucleosides are ordinarily very low (Plagemann and Erbe, '74; Waqar, unpublished), the deoxynucleoside entering the cell by reactions (1) and (2) would be a favored substrate for kinase action: (4) Deoxynucleoside (intracellular)

deoxynucleoside kinase

*

+ [y-32P1-ATP

I3T1-dNMP.

There are several reasons for concluding that the mechanism described in formulae (1)(4)cannot account for the apparent uptake of intact dNMPs by L cells. First, HeLa cells are much more efficient than L cells a t extracellular hydrolysis of dNMPs (reaction (1)). Thus HeLa cells generate larger quantities of the substrates for reaction (4).Nevertheless, HeLa cells are much less efficient than L cells at apparent uptake of intact nucleotides. Second, the mechanism predicts that, if one were to provide L cells with 32Piand a specific deoxynucleoside, then the corresponding [32P1dNMP would be generated in excess over noncoresponding [32PI-dNMPs.We have tested this prediction and have found that it does not hold true. The data in table 4 show that, when L cells are incubated with 32Piand dCyd a t 10 pM, only slightly more [32PI-dCMPis recoverable from the cells' DNA than when incubation is carried out without dCyd. A similar insignificant increase in [32P1-dTMPis seen when incubation is carried out in the presence of 10 p M dThd. Another prediction of the mechanism is that, if a different deoxynucleoside were

257

PENETRATION OF CELLS BY DEOXYNUCLEOTIDES TABLE 4

Effect of exogenous deoxynucleosides on the incorporation o f 32Piinto DNA by mouse L cells Percent of incorporated cpm in Added deoxynucleoside

cpm incorporated

-

30,023 38,191 39,227

dCyd (10 p M ) dThd (10 pMf

dAMP

dCMP

dGMP

dTMP

20

15 21 16

21 17 18

44 44 47

18 19

Exponentially growing mouse L cells were centrifuged and resuspended in three portlons of 50 ml a t 5.3 x lo6 cellslml. To each portion was added 3'Tito give a final concentration of 9.6 mM and a final specific activity of 1.4 x lo5 cpmlwnole. Deoxynucleosides (unlabeled) were added or not added as indicated. All other conditions were as in figures 1 and 2. TABLE 5

Effect o f exogenous deoxyguanosine on the uptake o f [32Pl-dAMPby mouse L cells Percent of incorporated cpm in Added dGuo

cpm incorporated

dAMP

dCMP

dGMP

dTMP

0 30 nM 10 p M

27,605 25,525 38,121

65 61 50

8 8 14

6 7 11

21 24 25

Exponentially growing mouse L cells were centrifuged and resuspended in three portions of 50 ml a t 5.0 X lo6cellslml. [$TIdAMP (4.31 x 10" cprnifimole) was added to 30 nM, and dGuo was added to t h e indicated concentrations. Incubation and all other conditions were as in figures 1 and 2.

added during incubation of cells with a particular [32P1-dNMP,then a significant proportion of 32Pwould be recovered in the dNMP corresponding to the added deoxynucleoside as well as in the dNMP used for labeling. The data in table 5 show that this prediction also does not hold true. When dGuo at 0.03 p M or a t 10 p M was added to L cells incubated with P2P1-dAMP (0.03 p M ) , there was no significant increase in labeling of the dGMP recovered from the cells' DNA. Because the mechanism postulated in formulae (1)-(4)is not capable of explaining the data in figure 1 and table 1, the simplest explanation for these data is that small amounts of intact dNMPs are capable of entering L cells. DISCUSSION

The data in this paper suggest that, as previously observed by Plunkett and &hen ('771, mouse L cells are capable of incorporating intact, exogenously supplied dAMP molecules into their DNA. In addition, the data show t h a t the other dNMPs can also be incorporated by L cells. A different cell line, HeLa, is incapable of incorporating intact dNMPs with the same efficiency as L cells. Whether HeLa cells can incorporate intact nucleotides at all is not clear from the data presented here. We think it is likely that the apparent up-

take of intact dNMPs into DNA by L cells (fig. 1, tables 1, 3) is a result of passage of the intact dNMPs through the membranes of healthy living cells and is not due to incorporation by damaged cells. The fact that different dNMPs were incorporated with different efficiencies (table 3) suggests a selective barrier to incorporation, either at the membrane level or a t the internal nucleotide pool level. Plunkett and Cohen ('77) also cite a reason for thinking that incorporation is carried out by the majority of healthy cells in the population, rather than a minority of damaged ones: the entire L cell population was susceptible to the toxic action of araAMP and ddAMP in their hands (under conditions where araAMP is not dephosphorylated and where ddAdo is non-toxic). The plasma membrane of healthy cells can be rendered reversibly permeable to nucleotides and other small molecules by treating the cells with salt (Castellot et al., '78) or with ATP (Rozengurt e t al., '77).It is possible that, under the conditions used in our experiments (which include buffering with HEPES and high cell concentration), L cells are rendered slightly permeable whereas HeLa cells are not. If the uptake of intact dNMPs by L cells under our growth conditions is due to a general weakening of permeability barriers, then it should be possible to detect migration of

258

M. A. WAQAR, R. L. TABER AND J. A. HUBERMAN

prelabeled dNMPs or other small molecules out of L cells under these conditions. We think i t is unlikely that the apparent uptake of intact dNMPs was due to contamination of L cells by mycoplasma. Both the L cell and HeLa cell cultures were tested for mycoplasma contamination by the uridine:uracil ratio method (Schneider et al., '74); no contamination was found. In addition, DNA was extracted from purified nuclei in all experiments; mycoplasma should have been removed. If we assume that incorporation of intact dNMPs is linearly proportional to dNMP concentration (there is no evidence for or against this assumption), then the extents of dNMP uptake detected by us (table 3) are in reasonable agreement with the value for dAMP uptake obtained by Plunkett and &hen ('77) of 3.4 x lo6 molecules of dAMP per L cell in a 4-hour incubation (starting concentration of 100 pM). The data presented in this paper show that HeLa cells and possibly (to a much smaller extent) L cells have surface or exogenous phosphatases which gradually hydrolyze exogenous dNMPs to deoxynucleosides and Pi. Although we only tested hydrolysis of dTMP directly, we can infer from the fact that the four [32PI-dNMPseach produce labeling of all four dNMPs in DNA (table 1) that all four dNMPs are hydrolyzed. This lack of specificity for dNMPs is consistent with the hydrolysis being carried out by 5'-nucleotidase, an enzyme found in high proportions on the outer surface of cell plasma membranes (Essner et al., '58; DePierre and Karnovsky, '73). If so, then these data suggest that the activity of 5'nucleotidase per cell can vary from cell line to cell line. These experiments provide a starting point for attempts to label DNA in mammalian cells with [32P1-dNMPs.They show that both uptake and hydrolysis of exogenous dNMPs are cell dependent. Therefore, attempts should be made to find a cell line demonstrating maximum uptake and minimum hydrolytic activity. To suppress incorporation of 32Pireleased by hydrolysis, experiments should be carried out in medium containing as high a concentration as possible of cold Pi. Finally, both the concentration and the specific activity of the added [32PI-dNMPshould be as high as possible. If i t is important that uptake occur without general permeabilization of the cell membrane, then tests should be done for each cell

line of interest to determine whether labeled nucleotides from internal cellular pools can leave the cell. LITERATURE CITED Anderson, S., G. Kaufmann and M. L. DePamphilis 1977 RNA primers in SV40 DNA replication: identification of transient RNA-DNA covalent linkages in replicating DNA. Biochemistry, 16: 4990-4998. Castellot, J. J., Jr., M. R. Miller and A. B. Pardee 1978 Animal cells reversibly permeable to small molecules. Proc. Natl. Acad. Sci. (U.S.A.), 75: 351-355. Cohen, S. S., and W. Plunkett 1975 The utilization of nucleotides by animal cells. Ann. N.Y. Acad. Sci., 255: 269-284. DePierre, J., and M. L. Karnovsky 1973 Plasma membranes of mammalian cells. J. Cell Biol., 56: 275-303. Essner, E., A. B. Novikoff and B. Masek 1958 Adenosine triphosphatase and 5'-nucleotidase activities in t h e plasma membrane of liver cells a s revealed hy electron microscopy. J. Biophys. Biochem. Cytol., 4: 711-716. Holland, J., L. McLaren and J. Syverton 1959 The mammalian cell virus relationship: IV. Infection of naturally insusceptible cells with enterovirus ribonucleic acid. J. Exp. Med., ZZO: 65-80. Kerr, S. E., K. Seraidarian and G. B. Brown 1951 On the utilization of purines and their ribose derivatives by yeast. J. Biol. Chem., 188: 207-216. Lehman, I. R., M. J. Bessman, E. S. Simms and A. Kornherg 1958 Enzymatic synthesis of deoxyribonucleic acid: I. Preparation of substrates and partial purification of a n enzyme from Escherichia coli. J. Biol. Chem., 233: 163-170. Leibman, K. C., and C. Heidelberger 1955 The metabolism of 32P-laheled ribonucleotides in tissue slices and cell suspensions. J. Biol. Chem., 216: 823-830. Magnusson, G.,V. Pigiet, E. L. Winnacker, R. Abrams and P. Reichard 1973 RNA-linked short DNA fragments during polyoma replication. Proc. Natl. Acad. Sci. (U.S.A.), 70: 412-415. Plagemann, P. G. W., and R. Erbe 1974 The deoxyribonucleoside transport systems of cultured Novikoff rat hepatoma cells. J. Cell. Physiol., 83: 337-344. Plagemann, P. G . W., R. Marz and R. M. Wohlhueter 1978 Transport and metabolism of deoxycytidine and 1-p-Darabinofuranosylcytosine into cultured Novikoff rat hepatoma cells, relationship to phosphorylation, and regulation of triphosphate synthesis. Cancer Res., 38: 978-989. Plunkett, W., and S. S. Cohen 1977 Penetration of mouse fibroblasts by 2-deoxyadenosine 5'-phosphate and incorporation of the nucleotide into DNA. J. Cell. Physiol., 91; 261-270. Plunkett, W., L. Lapi, P. J. Ortiz and S. S. Cohen 1973 Cellular penetration and metabolism of 9-6-D-arabinofuranosy1 adenine 5'-monophosphate faraAMPf. Proc. Amer. Assn. Cancer Res., 14: 32. 1974 Penetration of mouse fibroblasts by the 5'phosphate of 9-P-D-arabinofuranosyladenine and incorporation of the nucleotide into DNA. Proc. Natl. Acad. Sci. (U.S.A.), 71: 73-77. Reichard, P. 1953 Incorporation of cytidylic acids a and b into liver pentose nucleic acid. Acta Chem. Scand., 7: 862-865. Richardson, C. C. 1965 Phosphorylation of nucleic acid by enzyme from T4 bacteriophage-infected Escherichia coli. Proc. Natl. Acad. Sci. (U.S.A.), 54: 158-165. Roll, P. M., H. Weinfeld and E. Carroll 1956a The utilization of nucleotides by the mammal: V. Metabolism of pyrimidine nucleotides. J. Biol. Chem., 220: 455-465.

PENETRATION OF CELLS BY DEOXYNUCLEOTIDES Roll, P. M., H. Weinfeld, E. Carroll and G. B. Brown 1956h The utilization of nucleotides by the mammal: IV. Triply labeled purine nucleotides. J. Biol. Chem., 220; 439-454. Rozengurt, E., L. A. Heppel and I. Friedberg 1977 Effect of exogenous ATP on the permeability properties of transformed cultures of mouse cell lines. J. Biol. Chem., 252: 4584-4590. Scarano, E., G. Gerazi, A. Polzella and E. Campanile 1963 The enzymatic aminohydrolysis of 4-amino pyrimidine deoxyribonucleotides. J. Biol. Chem., 238: PC1556-1557. Schneider, E. L., E. J. Stanbridge and C. J. Epstein 1974 Incorporation of 3H uridine and 3H-uracilinto RNA: a sim-

259

ple technique for the detection of mycoplasma contamination of cultured cells. Exp. Cell Res., 84; 311-318. Tseng, B. Y., and M. Goulian 1975 Evidence for covalent association of RNA with nascent DNA in human lymphocytes. J. Mol. Biol., 99: 339-346. Waqar, M. A., and J. A. Huberman 1975 Covalent attachment of RNA to nascent DNA in mammalian cells. Cell, 6; 551-557. Wickner, R. B. 1974 Mutants of Saccharomyces cerevisiae that incorporate deoxythymidine 5'-monophosphate into deoxyribonucleic acid in vivo. J. Bact., 117: 252-260.