Cell,
Vol. 11, 859-869,
August
1977,
Copyright
0 1977 by MIT
Initiation Points for DNA Replication in Nontransformed and Simian Virus 40-Transformed Chinese Hamster Lung Cells Robert G. Martin and Ariella Oppenheim Laboratory of Molecular Biology National Institute of Arthritis, Metabolism and Digestive Diseases Bethesda, Maryland 20014
Summary Randomly growing Chinese hamster lung cells were pulse-labeled with 3H-thymidine, and the replicating forks of individual DNA fibers were visualized by autoradiography. When grown in complete medium, wild-type SV40-transformed cells had more forks per unit length of DNA than nontransformed cells. In isoleucine-depleted medium, wild-type SV40-transformed cells had fewer forks per unit length than those few nontransformed cells (l-3% of the population) which continued DNA replication. Cells transformed by a tsA mutant of SV40 when grown at the permissive temperature had more forks per unit length in complete medium and fewer forks per unit length in depleted medium than nontransformed cells, but when grown at the restrictive temperature, the tsA-transformed cells behaved like nontransformed cells.
Introduction The T antigen of simian virus 40 (SV40) is the product of the viral A gene (Alwine et al., 1975; Kuchino and Yamaguchi, 1975; Prives et al., 1975; Tegtmeyeretal.,1975;Tenenetal.,1975,1977;Rundell et al., 1977). Evidence for the crucial role of T antigen in the establishment and maintenance of transformation of hamster, mouse and rat cells was provided by studies using temperature-sensitive (ts) mutants in the A gene (Tegtmeyer, 1972, 1975; Butel, Brugge and Noonan, 1974; Martin et al., 1974; Osborn and Weber, 1974, 1975; Brugge and Butel, 1975; Kimura and Itagaki, 1975; Martin and Chou, 1975; Anderson and Martin, 1976; Tenen et al., 1977). In addition, T antigen has been shown to be required for the initiation of viral DNA synthesis and the induction of host DNA synthesis during lytic infection of monkey cells by SV40 (Tegtmeyer, 1972; Chou, Avila and Martin, 1974; Chou and Martin, 1975). We have therefore presented a model for the molecular basis of transformation by SV40, which proposes that T antigen also acts as an initiator of host DNA synthesis in transformed cells (Martin et al., 1974, 1976, 1977). Specifically, we have argued that normal cells proceed through the Gl phase of the cell cycle and enter a resting
phase, GO, outside of the cell cycle if the culture medium is depleted. In SV40-transformed cells, the T antigen acting at the encl of the Gl phase drives the cells into S phase even under suboptimal growth conditions (Martin and Stein, 1976). This model predicts that under optimal conditions, SV40-transformed cells initiate DNA synthesis both at the initiation sites of nontransformed cells (N sites) and at new sites activated by T antigen (T sites). In depleted medium, nontransformed cells do not enter S phase, whereas SV40-transformed cells still initiate DNA synthesis, but only at the T sites. We have therefore measured the number of initiation points for DNA synthesis in nontransformed cells and in SV40-transformed cells in complete and depleted medium. Unfortunately, there is no simple direct technique for estimating the number of origins of DNA replication in mammalian cells. The number of origins, however, should be inversely proportional to the average replicon size, since the genomic DNA content of any spc!cies is constant. The method of DNA fiber autorEdiography of Huberman and Riggs (1968), by which replicating forks are directly visualized, has bef?n widely applied to the estimation of replicon sixe. Using this method, Huberman and Riggs (lEl68) demonstrated that DNA replication proceeds bidirectionally from the initiation points in mammalian cells. When cells are labeled for short times, fcrks from origins which started replication prior to the addition of the label are preferentially labeled (Blumenthal, Kriegstein and Hogness, 1973). Sine? replication is bidirectional, two forks are seen for each replicon (see legend to Figure I), so that the mean distance between origins of replirxtion is twice the mean distance between replicating forks (Huberman and Riggs, 1968; Blumenthal et al., 1973; Callan, 1973). Using the method of Hub?rman and Riggs (1968), we have found a significant decrease in the average replicon size in transformed cells growing in complete medium. In cells trarxformed by atsA mutant, the decrease is temperature-sensitive, indicating that a functioning T antigeli is required. Additional studies have shown that transformed cells continue to synthesize DNA when grown in isoleucine-depleted medium; the replicons are longer, however, implying that not all the nitiation sites for DNA replication are operating irl these cells.
Cell 860
Results DNA Synthesis in Nontransformed and SV40Transformed Cells under Various Growth Conditions Randomly growing cells or cells incubated in isoleucine-depleted medium for 24 hr were pulse-labeled with 3H-thymidine. Approximately 50% of the nuclei of nontransformed cells and of cells transformed by wild-type SV40 or a temperature-sensitive A gene mutant were labeled by an 8 min pulse during growth at 33°C in complete medium (Table 1); at 4o”C, there was some decrease in the percentage of labeled cells in the nontransformed culture. In depleted medium, a dramatic decrease in the percentage of labeled nuclei was observed in the nontransformed cells at both temperatures and in the tsA239-transformed cells incubated at 40°C. These results corroborate our previous finding that nontransformed cells enter a resting state in isoleutine-depleted medium, whereas wild-type SV40transformed cells do not do so (Martin and Stein, 1976). The cells transformed by a tsA mutant of SV40 enter a resting state at the restrictive temperature (4o”C), but not at the permissive temperature (33°C). The isoleucine-depleted medium is not entirely free of isoleucine, so that cells in S phase at the time of depletion can complete the cell cycle and do not die (Martin and Anderson, 1976; Martin and Stein, 1976). Slides of the cells labeled in this experiment were prepared for DNA fiber autoradiography by a modification (see Experimental Procedures) of the procedure of Huberman and Riggs (1968). The slides were developed after 3 or 6 months of exposure. Each radioactive track seen by this procedure (Figure 1) corresponds to a replicating fork labeled during the pulse period (Huberman and Riggs, 1968; Blumenthal et al., 1973; Callan, 1973; Hand and Tamm, 1974). The mean distance between track centers was estimated by the two methods described below. Method A: Estimation of the Mean Distances between Track Centers by Counting Sets of Tracks This method involved counting the number Table
1. Percentage
of Nuclei
% Nuclei Complete
Labeled
in 8 Min by 3H-Thymidine
Labeled Medium
Depleted
Medium
Cell Line
33°C
40°C
33°C
40°C
CHL CHLWT15 CHLA239Ll
51 66 52
29 54 45
3 48 48
1 45 7
tracks in sets of tracks at least 130 p in length (for details of the counting procedure, see Experimental Procedures). The results are represented in Figure 2. A modal number of 6 tracks per 130 F was found for nontransformed cells growing at 33 or 40°C in complete medium. On the other hand, a modal number of 8 tracks per 130 p was found for cells transformed by wild-type SV40 in complete medium at both 33 and 40°C. The cells transformed by tsA239 also had a modal number of 8 tracks per 130 p at 33”C, but only 6 tracks per 130 p at 40°C in complete medium. The tsA239-transformed cells were thus more similar to transformed cells at the permissive temperature and to nontransformed cells at the restrictive temperature, suggesting that intertrack distances (the distances between the centers of adjacent tracks) are influenced by a functioning T antigen. We interpret these data as indicating that in complete medium, the modal intertrack distance in nontransformed cells is -22 p, whereas the modal intertrack distance in transformed cells is -16 p. In isoleucine-depleted medium at 33”C, the modal number of tracks per 130 p for nontransformed cells was -6, the same as in complete medium. We stress that the number of labeled tracks per slide for the nontransformed cells in depleted medium was very small, reflecting the fact that only 3% of the nuclei were labeled in the iso-
Figure
of
1. A Set of Tracks
Corresponding
to Six Replicating
Forks
Cells labeled for 8 min with 3H-thymidine were lysed, and the DNA fibers were spread on glass slides (Experimental Procedures). This slide was developed after 3 months of exposure. Because of the bidirectional nature of mammalian DNA replication (Huberman and Riggs, 1968), the replicating forks represented in this autoradiogram could have originated either at A, Band C, or at D, E and origins off the photograph. In either case, an average replicon size can be approximated by the sum of two adjacent distances between track centers. The tracks usually appear in pairs with alternating short and long distances (Huberman and Riggs, 1968; Hand and Tamm, 1974). In the estimation of replicon size by method B, an even number of intertrack distances (that is, equal number of long and short distances) was recorded for every set of tracks measured (see also Blumenthal et al., 1973).
Replicons 861
of Cl-IL cells
.-:.i... ,..i-. i I-. I-i I-.: ilII
Other experiments, in which the pulse labeling was for 4, 5 or 10 min, gave results similar to those presented. Several investigators have pointed out the danger of personal bias inherent in the interpretation of autoradiograms (Blumenthal et al., 1973; Hand and Tamm, 1974). We therefore decided to perform additional analyses using a “blind” procedure.
i
;...! ,-! ..I - I
5
-
3
6
3
6
i.:;...: I _,/.... .:
9
12
3
6
9
12
3
6
REPLICATION FORKS PER IjO,,
Figure 2. Histograms of the Number of Sets Forks as Opposed to the Number of Tracks
9
12
9 h)
of Tracks
with
n
The slides as in Figure 1 were all developed after 3 months of exposure, and the sets of tracks were counted by method A (Experimental Procedures). Approximately 100 sets of tracks were counted for each cell type under each growth condition, with the exception of the CHL cells in depleted medium at 4O”C, for which only 46 sets were counted. Nontransformed cells (CHL) (-), wild-type SV40-transformed cells (CHLWT15) (- --) and cells transformed by a temperature-sensitive A mutant of SV40 (CHLA239Ll) (. “) were examined in complete (upper panels) or depleted medium (lower panels) at 33°C (left-hand panels) or 40°C (right-hand panels).
leucine-depleted medium (see Table 1). Indeed, we could not find 100 sets of tracks for the analysis of the nontransformed cells grown in depleted medium at 40°C and believe that the small sample size may account for the apparent decrease in modal number (Figure 2). Accordingly, there do not appear to be significant differences in the distributions of the three cell types in isoleucine-depleted medium at 40°C. A dramatic decrease in the modal number of tracks per 130 p was observed in the wild-type SV40-transformed cells incubated in isoleucine-depleted medium. The modal number of tracks was 4 per 130 pas compared with 8 per 130 p in complete medium. The cultures at both 33 and 40°C contained many sets of tracks in contrast to the nontransformed cells. The tsA239-transformed cells also had a modal number of -4 tracks per 130 p when grown in isoleucine-depleted medium at 33°C. At 40X, however, these cells seemed to have a bimodal distribution of 4 and 6 tracks per 130 p. A modal number of 4 tracks per 130 p indicates a mean distance between tracks of -33 p.
Method B: Direct Measurement of the Distances between Track Centers The coded slides were photographed by one investigator, and a new code was then noted on the back of each photograph. At least 50 photographs of each cell type under each experimental condition (that is, over 600 photographs) were taken. The photographs were thoroughly shuffled, and the second investigator then chose which sets of tracks to measure in each photograph, according to the rules outlined in Experimental Procedures. After the measurements were made and recorded, the photograph was turned over and its code number was recorded. Only after all the data had been collected were the samples decoded. In Figure 3, the frequency of intertrack distances of length n, F,, is plotted against n for the three cell lines grown and labeled at 33°C in complete and depleted medium (with the exception of CHLA239Ll in depleted medium). F, is expressed as the percentage of the total number of intertrack distances measured. The cumulative frequency, I;,“F,, is plotted along with F, in each of the five panels for the different cell types and conditions. The five cumulative frequency curves are superimposed in the sixth panel. As previously observed by other investigators (Huberman and Riggs, 1968; Blumenthal et al., 1973; Hand and Tamm, 1974), the distributions of the intertrack distances are not Gaussian; a pronounced skewing of the curves to greater values of n is seen (Figure 3). The value of n at which Z;F, = 50% corresponds to the median of the distribution. Since the medians are considerably less than the arithmetic means for all the cell lines (Table 2), the distributions are clearly not “normal” in the statistical sense. Thus statistically it is not meaningful to calculate standard deviations of the means, as is often done. The curves in Figure 3 fall into three distinct groups. First, transformed cells, both wild-type and tsA, grown in complete medium, have the highest frequencies of short intertrack distances and hence the most steeply rising cumulative frequency curves (Figure 3, lower right-hand panel). Second, nontransformed cells, grown either in complete or in depleted medium, had intermediate frequencies of short intertrack distances and therefore intermediate initial slopes for the cumulative frequency
Cell 882
DISTANCE BETWEEN
TRACK
CENTERS
IMICRONS,
Figure 3. Frequency and Cumulative Frequency Profiles track Distances in Nontransformed and SV40-Transformed in Complete and Depleted Medium at 33°C
of InterCells
Slides of the various cell types in complete or depleted medium were developed after 6 months of exposure, and intertrack distances were measured and analyzed by method B (Experimental Procedures). The frequency, in percentages, of the number of intertrack distances of length n versus n (- --) is represented in each of the panels (except the lower right-hand panel). The cumulative frequency is represented in the same five panels (-). The number of inter-track distances measured for each cell type is given in Table 2. All cells were grown at 33°C in complete medium (left-hand panels) or isoleucine-depleted medium (upper two right-hand panels). The cumulative frequency curves are replotted in the lower right-hand panel (note that the abscissa has been expanded) and correspond to nontransformed CHL cells in complete medium (-) or depleted medium (---), wild-type SV40transformed cells in complete (, .) or depleted medium (xxxx), and tsA239 SV40-transformed cells in complete medium (0000).
curves. Third, transformed cells grown in depleted medium had the lowest frequencies of short intertrack distances and had the lowest initial slope for their cumulative frequency curve. The mean intertrack distances found by analysis using method B (Table 2) were in excellent agreement with the modal distances found using method A. In complete or depleted medium, the mean intertrack distances in nontransformed cells at both temperatures and in tsA239-transformed cells at 40°C clustered around 22 p (compare to -22 p by method A). In complete medium, the mean intertrack distances in wild-type-transformed cells at both temperatures and in tsA239-transformed cells at 33°C clustered around 17 p (compare to -16 p by method A). In depleted medium, the mean intertrack distances in wild-type-transformed cells at both temperatures was -30 p (compare to -33 p by method A). The mean intertrack distance in tsA239-transformed cells at 33°C in depleted medium was slightly less than for the wild-type-trans-
formed cells in depleted medium, but was considerably greater than for the same cells in complete medium. Statistical analyses of the data obtained by method 6 were made. Since the frequency distributions of intertrack distances were clearly skewed (Figure 3), we first examined whether the distributions could be approximated by a log-normal (rather than a normal) distribution. It was found by the x2 test that each of the twelve sets of data (five of which are illustrated in Figure 3) could be approximated by a log-normal distribution in a statistically significant way, provided the distances ~4.5 p were pooled in the analyses. This pooling is legitimate, since 4.5 p corresponds to a distance between tract centers equal to the mean track lengths (see below) and hence corresponds to the limit of resolution of our measurements. Since the data from each set could be approximated by a log-normal distribution, we felt justified in dividing the data at random into subsets from which we calculated log-normal means and standard deviations for each set of data (for details of the statistical analysis, see Experimental Procedures). Table 2 shows, as expected, that the geometric means (antilogs of the log-normal means) calculated by this method are smaller than the arithmetic means and approach the medians. However, the three parameters of the intertrack distances (the arithmetic means, medians and geometric means) for the various cell types and conditions follow essentially the same pattern and fall into three groups: the smallest values were found for the transformed cells-(wild type-transformed cells at 33 and 4o”C, and tsA239-transformed cells at 33°C) in complete medium (group I); the largest values for the same transformed cells in depleted medium (group II); and intermediate values for cells having the nontransformed phenotype (nontransformed cells and tsA239-transformed cells at 40°C) whether in complete or depleted medium (group Ill). The only exception was that the tsA239-transformed cells in depleted medium at 33°C had a somewhat smaller mean intertrack distance than the wild-type-transformed cells in depleted medium. Probability (P) values for the significance of the differences between the geometric means of the various sets of data are listed (pairwise) in Table 3. As can be seen, the differences in the intertrack distances clearly fall into the three groups indicated above and are highly significant (with the exception that the tsA239-transformed cells in depleted medium at 33°C are not significantly different from the nontransformed cells). Furthermore, when the combined means for the transformed cells in complete medium (group I) were compared
Replicons 863
Table
of CHL cells
2. lntertrack
Distances
in Nontransformed
and SV40-Transformed
Cell Line
Medium
Temperature
Number Distances
CHL
Complete
33°C
380
40°C
417
33°C
Depleted
CHLWT15
Complete
Depleted
CHLA239Ll
Complete
Depleted
Cells of Intertrack Measured
Median (+)
Arithmetic
Mean
Geometric
Mean
(CL)
(CL)
15.6
21.7
17.01
k 1.68
15.9
23.8
17.58
rt 1.71
429
15.8
22.2
17.14
t
40°C
400
14.8
21.9
16.14
2 1.64
33°C
354
11.0
16.1
13.10
2 1.58
40°C
453
12.7
18.2
14.28
+ 1.69
33°C
232
22.2
30.7
22.43
i 1.79
40°C
210
22.1
28.6
22.15
k 1.88
1.63
33°C
402
10.9
17.0
12.90
2 1.61
40°C
319
14.1
22.5
18.42
k 1.73
33°C
243
17.4
25.4
18.62
i 1.76
40°C
374
14.3
22.8
17.08
+ 1.56
with the combined means for the transformed ceils in depleted medium (group II), the difference was significant at P = 3.6 x 10e3. When group I was compared with the means for the cells having the nontransformed phenotype (group Ill), the difference was significant at P = 1 .O x 10p4. When group II and group III were compared, the difference in the intertrack distances was significant at only P = 0.075 because of the relatively small intertrack distance for the tsA239-transformed cells. A similar comparison between groups II and III excluding the &A239-transformed cells from group II was significant at P = 10m4. Measurement of the Rate of DNA Chain Elongation by Track Length The rate of DNA chain elongation should be represented by the length of tracks in the autoradiograms. We measured the lengths of the tracks in all the photographs obtained by method B above to determine the rate of chain elongation in nontransformed and transformed cells. Since these measurements were made at the same time as the distances between tracks were measured, this analysis also was “blind.” A typical distribution pattern is shown in Figure 4. Unlike the log-normal distributions of the distances between track centers, each of these sets of data could be approximated by a truncated normal distribution curve, indicating that the rate of chain elongation was approximately uniform in all forks. Means and standard errors were calculated (Table 4) on the sets and subsets. Very few significant differences (P < 0.05) in track lengths were detected when the sets were compared pairwise in all possible combinations. The tsA239-transformed cells grown at 33°C in depleted medium had a small mean track length (Table 4), which was significantly different from that of
most of the other cell lines grown under the different conditions. The mean track length of nontransformed cells grown in depleted medium at 40°C was significantly different from the mean track length of nontransformed and wild-type-transformed cells grown in complete medium at 4O”C, and tsA239-transformed cells grown in depleted medium at 40°C. All the other 54 pairwise t tests yielded no significant differences (P > 0.05) in track length. We do not understand the reason for the shorter track length in the tsA239-transformed cells grown in depleted medium at 33°C. It is the same cell line and growth condition that gave intertrack distances which were shorter than expected (see Tables 2 and 3). We were somewhat surprised that there was no consistent pattern of differences in track lengths, since we expected that the rate of chain elongation would be greater at 40 than at 33°C and possibly greater in complete than in depleted medium. We therefore grouped all the means of the cells grown at 40°C and compared them with the group consisting of all the means of the cells grown at 33°C. No significant difference in track length between the group at 40°C compared with the group at 33°C could be detected. When the means were grouped by growth medium (complete as opposed to depleted), there was also no significant difference between the track lengths of the cells grown in complete as opposed to depleted medium. Since all the cultures were pulse-labeled for 8 min, we calculated an average rate of chain elongation of 0.58 @min for all the cultures. Measurement of the Rate of Chain Elongation by the Method of Painter and Schaefer We wished to obtain an independent verification of the change in replicon size following transforma-
33°C 40°C 33°C 40°C
33°C 40°C 33°C 40°C
Complete
33°C 40°C 33°C 40°C
Complete
Depleted
Depleted
between
Temperature
t Tests
33°C >O.l”
40°C >O.l >O.l
33°C -co.01 O.l indicate no significant difference. bA significant difference, although expected in these pairwise tests, was not found. can be accounted for by the smaller than expected value for the intertrack distance cThese values are of borderline significance 0.1 > P > 0.05.
CHLA239Ll
CHLWT15
Depleted
Complete
Pairwise
CHL
from
Medium
3. P Values
Cell Type
Table
Vol. 11, 859-869,
August
1977,
Copyright
0 1977 by MIT
Initiation Points for DNA Replication in Nontransformed and Simian Virus 40-Transformed Chinese Hamster Lung Cells Robert G. Martin and Ariella Oppenheim Laboratory of Molecular Biology National Institute of Arthritis, Metabolism and Digestive Diseases Bethesda, Maryland 20014
Summary Randomly growing Chinese hamster lung cells were pulse-labeled with 3H-thymidine, and the replicating forks of individual DNA fibers were visualized by autoradiography. When grown in complete medium, wild-type SV40-transformed cells had more forks per unit length of DNA than nontransformed cells. In isoleucine-depleted medium, wild-type SV40-transformed cells had fewer forks per unit length than those few nontransformed cells (l-3% of the population) which continued DNA replication. Cells transformed by a tsA mutant of SV40 when grown at the permissive temperature had more forks per unit length in complete medium and fewer forks per unit length in depleted medium than nontransformed cells, but when grown at the restrictive temperature, the tsA-transformed cells behaved like nontransformed cells.
Introduction The T antigen of simian virus 40 (SV40) is the product of the viral A gene (Alwine et al., 1975; Kuchino and Yamaguchi, 1975; Prives et al., 1975; Tegtmeyeretal.,1975;Tenenetal.,1975,1977;Rundell et al., 1977). Evidence for the crucial role of T antigen in the establishment and maintenance of transformation of hamster, mouse and rat cells was provided by studies using temperature-sensitive (ts) mutants in the A gene (Tegtmeyer, 1972, 1975; Butel, Brugge and Noonan, 1974; Martin et al., 1974; Osborn and Weber, 1974, 1975; Brugge and Butel, 1975; Kimura and Itagaki, 1975; Martin and Chou, 1975; Anderson and Martin, 1976; Tenen et al., 1977). In addition, T antigen has been shown to be required for the initiation of viral DNA synthesis and the induction of host DNA synthesis during lytic infection of monkey cells by SV40 (Tegtmeyer, 1972; Chou, Avila and Martin, 1974; Chou and Martin, 1975). We have therefore presented a model for the molecular basis of transformation by SV40, which proposes that T antigen also acts as an initiator of host DNA synthesis in transformed cells (Martin et al., 1974, 1976, 1977). Specifically, we have argued that normal cells proceed through the Gl phase of the cell cycle and enter a resting
phase, GO, outside of the cell cycle if the culture medium is depleted. In SV40-transformed cells, the T antigen acting at the encl of the Gl phase drives the cells into S phase even under suboptimal growth conditions (Martin and Stein, 1976). This model predicts that under optimal conditions, SV40-transformed cells initiate DNA synthesis both at the initiation sites of nontransformed cells (N sites) and at new sites activated by T antigen (T sites). In depleted medium, nontransformed cells do not enter S phase, whereas SV40-transformed cells still initiate DNA synthesis, but only at the T sites. We have therefore measured the number of initiation points for DNA synthesis in nontransformed cells and in SV40-transformed cells in complete and depleted medium. Unfortunately, there is no simple direct technique for estimating the number of origins of DNA replication in mammalian cells. The number of origins, however, should be inversely proportional to the average replicon size, since the genomic DNA content of any spc!cies is constant. The method of DNA fiber autorEdiography of Huberman and Riggs (1968), by which replicating forks are directly visualized, has bef?n widely applied to the estimation of replicon sixe. Using this method, Huberman and Riggs (lEl68) demonstrated that DNA replication proceeds bidirectionally from the initiation points in mammalian cells. When cells are labeled for short times, fcrks from origins which started replication prior to the addition of the label are preferentially labeled (Blumenthal, Kriegstein and Hogness, 1973). Sine? replication is bidirectional, two forks are seen for each replicon (see legend to Figure I), so that the mean distance between origins of replirxtion is twice the mean distance between replicating forks (Huberman and Riggs, 1968; Blumenthal et al., 1973; Callan, 1973). Using the method of Hub?rman and Riggs (1968), we have found a significant decrease in the average replicon size in transformed cells growing in complete medium. In cells trarxformed by atsA mutant, the decrease is temperature-sensitive, indicating that a functioning T antigeli is required. Additional studies have shown that transformed cells continue to synthesize DNA when grown in isoleucine-depleted medium; the replicons are longer, however, implying that not all the nitiation sites for DNA replication are operating irl these cells.
Cell 860
Results DNA Synthesis in Nontransformed and SV40Transformed Cells under Various Growth Conditions Randomly growing cells or cells incubated in isoleucine-depleted medium for 24 hr were pulse-labeled with 3H-thymidine. Approximately 50% of the nuclei of nontransformed cells and of cells transformed by wild-type SV40 or a temperature-sensitive A gene mutant were labeled by an 8 min pulse during growth at 33°C in complete medium (Table 1); at 4o”C, there was some decrease in the percentage of labeled cells in the nontransformed culture. In depleted medium, a dramatic decrease in the percentage of labeled nuclei was observed in the nontransformed cells at both temperatures and in the tsA239-transformed cells incubated at 40°C. These results corroborate our previous finding that nontransformed cells enter a resting state in isoleutine-depleted medium, whereas wild-type SV40transformed cells do not do so (Martin and Stein, 1976). The cells transformed by a tsA mutant of SV40 enter a resting state at the restrictive temperature (4o”C), but not at the permissive temperature (33°C). The isoleucine-depleted medium is not entirely free of isoleucine, so that cells in S phase at the time of depletion can complete the cell cycle and do not die (Martin and Anderson, 1976; Martin and Stein, 1976). Slides of the cells labeled in this experiment were prepared for DNA fiber autoradiography by a modification (see Experimental Procedures) of the procedure of Huberman and Riggs (1968). The slides were developed after 3 or 6 months of exposure. Each radioactive track seen by this procedure (Figure 1) corresponds to a replicating fork labeled during the pulse period (Huberman and Riggs, 1968; Blumenthal et al., 1973; Callan, 1973; Hand and Tamm, 1974). The mean distance between track centers was estimated by the two methods described below. Method A: Estimation of the Mean Distances between Track Centers by Counting Sets of Tracks This method involved counting the number Table
1. Percentage
of Nuclei
% Nuclei Complete
Labeled
in 8 Min by 3H-Thymidine
Labeled Medium
Depleted
Medium
Cell Line
33°C
40°C
33°C
40°C
CHL CHLWT15 CHLA239Ll
51 66 52
29 54 45
3 48 48
1 45 7
tracks in sets of tracks at least 130 p in length (for details of the counting procedure, see Experimental Procedures). The results are represented in Figure 2. A modal number of 6 tracks per 130 F was found for nontransformed cells growing at 33 or 40°C in complete medium. On the other hand, a modal number of 8 tracks per 130 p was found for cells transformed by wild-type SV40 in complete medium at both 33 and 40°C. The cells transformed by tsA239 also had a modal number of 8 tracks per 130 p at 33”C, but only 6 tracks per 130 p at 40°C in complete medium. The tsA239-transformed cells were thus more similar to transformed cells at the permissive temperature and to nontransformed cells at the restrictive temperature, suggesting that intertrack distances (the distances between the centers of adjacent tracks) are influenced by a functioning T antigen. We interpret these data as indicating that in complete medium, the modal intertrack distance in nontransformed cells is -22 p, whereas the modal intertrack distance in transformed cells is -16 p. In isoleucine-depleted medium at 33”C, the modal number of tracks per 130 p for nontransformed cells was -6, the same as in complete medium. We stress that the number of labeled tracks per slide for the nontransformed cells in depleted medium was very small, reflecting the fact that only 3% of the nuclei were labeled in the iso-
Figure
of
1. A Set of Tracks
Corresponding
to Six Replicating
Forks
Cells labeled for 8 min with 3H-thymidine were lysed, and the DNA fibers were spread on glass slides (Experimental Procedures). This slide was developed after 3 months of exposure. Because of the bidirectional nature of mammalian DNA replication (Huberman and Riggs, 1968), the replicating forks represented in this autoradiogram could have originated either at A, Band C, or at D, E and origins off the photograph. In either case, an average replicon size can be approximated by the sum of two adjacent distances between track centers. The tracks usually appear in pairs with alternating short and long distances (Huberman and Riggs, 1968; Hand and Tamm, 1974). In the estimation of replicon size by method B, an even number of intertrack distances (that is, equal number of long and short distances) was recorded for every set of tracks measured (see also Blumenthal et al., 1973).
Replicons 861
of Cl-IL cells
.-:.i... ,..i-. i I-. I-i I-.: ilII
Other experiments, in which the pulse labeling was for 4, 5 or 10 min, gave results similar to those presented. Several investigators have pointed out the danger of personal bias inherent in the interpretation of autoradiograms (Blumenthal et al., 1973; Hand and Tamm, 1974). We therefore decided to perform additional analyses using a “blind” procedure.
i
;...! ,-! ..I - I
5
-
3
6
3
6
i.:;...: I _,/.... .:
9
12
3
6
9
12
3
6
REPLICATION FORKS PER IjO,,
Figure 2. Histograms of the Number of Sets Forks as Opposed to the Number of Tracks
9
12
9 h)
of Tracks
with
n
The slides as in Figure 1 were all developed after 3 months of exposure, and the sets of tracks were counted by method A (Experimental Procedures). Approximately 100 sets of tracks were counted for each cell type under each growth condition, with the exception of the CHL cells in depleted medium at 4O”C, for which only 46 sets were counted. Nontransformed cells (CHL) (-), wild-type SV40-transformed cells (CHLWT15) (- --) and cells transformed by a temperature-sensitive A mutant of SV40 (CHLA239Ll) (. “) were examined in complete (upper panels) or depleted medium (lower panels) at 33°C (left-hand panels) or 40°C (right-hand panels).
leucine-depleted medium (see Table 1). Indeed, we could not find 100 sets of tracks for the analysis of the nontransformed cells grown in depleted medium at 40°C and believe that the small sample size may account for the apparent decrease in modal number (Figure 2). Accordingly, there do not appear to be significant differences in the distributions of the three cell types in isoleucine-depleted medium at 40°C. A dramatic decrease in the modal number of tracks per 130 p was observed in the wild-type SV40-transformed cells incubated in isoleucine-depleted medium. The modal number of tracks was 4 per 130 pas compared with 8 per 130 p in complete medium. The cultures at both 33 and 40°C contained many sets of tracks in contrast to the nontransformed cells. The tsA239-transformed cells also had a modal number of -4 tracks per 130 p when grown in isoleucine-depleted medium at 33°C. At 40X, however, these cells seemed to have a bimodal distribution of 4 and 6 tracks per 130 p. A modal number of 4 tracks per 130 p indicates a mean distance between tracks of -33 p.
Method B: Direct Measurement of the Distances between Track Centers The coded slides were photographed by one investigator, and a new code was then noted on the back of each photograph. At least 50 photographs of each cell type under each experimental condition (that is, over 600 photographs) were taken. The photographs were thoroughly shuffled, and the second investigator then chose which sets of tracks to measure in each photograph, according to the rules outlined in Experimental Procedures. After the measurements were made and recorded, the photograph was turned over and its code number was recorded. Only after all the data had been collected were the samples decoded. In Figure 3, the frequency of intertrack distances of length n, F,, is plotted against n for the three cell lines grown and labeled at 33°C in complete and depleted medium (with the exception of CHLA239Ll in depleted medium). F, is expressed as the percentage of the total number of intertrack distances measured. The cumulative frequency, I;,“F,, is plotted along with F, in each of the five panels for the different cell types and conditions. The five cumulative frequency curves are superimposed in the sixth panel. As previously observed by other investigators (Huberman and Riggs, 1968; Blumenthal et al., 1973; Hand and Tamm, 1974), the distributions of the intertrack distances are not Gaussian; a pronounced skewing of the curves to greater values of n is seen (Figure 3). The value of n at which Z;F, = 50% corresponds to the median of the distribution. Since the medians are considerably less than the arithmetic means for all the cell lines (Table 2), the distributions are clearly not “normal” in the statistical sense. Thus statistically it is not meaningful to calculate standard deviations of the means, as is often done. The curves in Figure 3 fall into three distinct groups. First, transformed cells, both wild-type and tsA, grown in complete medium, have the highest frequencies of short intertrack distances and hence the most steeply rising cumulative frequency curves (Figure 3, lower right-hand panel). Second, nontransformed cells, grown either in complete or in depleted medium, had intermediate frequencies of short intertrack distances and therefore intermediate initial slopes for the cumulative frequency
Cell 882
DISTANCE BETWEEN
TRACK
CENTERS
IMICRONS,
Figure 3. Frequency and Cumulative Frequency Profiles track Distances in Nontransformed and SV40-Transformed in Complete and Depleted Medium at 33°C
of InterCells
Slides of the various cell types in complete or depleted medium were developed after 6 months of exposure, and intertrack distances were measured and analyzed by method B (Experimental Procedures). The frequency, in percentages, of the number of intertrack distances of length n versus n (- --) is represented in each of the panels (except the lower right-hand panel). The cumulative frequency is represented in the same five panels (-). The number of inter-track distances measured for each cell type is given in Table 2. All cells were grown at 33°C in complete medium (left-hand panels) or isoleucine-depleted medium (upper two right-hand panels). The cumulative frequency curves are replotted in the lower right-hand panel (note that the abscissa has been expanded) and correspond to nontransformed CHL cells in complete medium (-) or depleted medium (---), wild-type SV40transformed cells in complete (, .) or depleted medium (xxxx), and tsA239 SV40-transformed cells in complete medium (0000).
curves. Third, transformed cells grown in depleted medium had the lowest frequencies of short intertrack distances and had the lowest initial slope for their cumulative frequency curve. The mean intertrack distances found by analysis using method B (Table 2) were in excellent agreement with the modal distances found using method A. In complete or depleted medium, the mean intertrack distances in nontransformed cells at both temperatures and in tsA239-transformed cells at 40°C clustered around 22 p (compare to -22 p by method A). In complete medium, the mean intertrack distances in wild-type-transformed cells at both temperatures and in tsA239-transformed cells at 33°C clustered around 17 p (compare to -16 p by method A). In depleted medium, the mean intertrack distances in wild-type-transformed cells at both temperatures was -30 p (compare to -33 p by method A). The mean intertrack distance in tsA239-transformed cells at 33°C in depleted medium was slightly less than for the wild-type-trans-
formed cells in depleted medium, but was considerably greater than for the same cells in complete medium. Statistical analyses of the data obtained by method 6 were made. Since the frequency distributions of intertrack distances were clearly skewed (Figure 3), we first examined whether the distributions could be approximated by a log-normal (rather than a normal) distribution. It was found by the x2 test that each of the twelve sets of data (five of which are illustrated in Figure 3) could be approximated by a log-normal distribution in a statistically significant way, provided the distances ~4.5 p were pooled in the analyses. This pooling is legitimate, since 4.5 p corresponds to a distance between tract centers equal to the mean track lengths (see below) and hence corresponds to the limit of resolution of our measurements. Since the data from each set could be approximated by a log-normal distribution, we felt justified in dividing the data at random into subsets from which we calculated log-normal means and standard deviations for each set of data (for details of the statistical analysis, see Experimental Procedures). Table 2 shows, as expected, that the geometric means (antilogs of the log-normal means) calculated by this method are smaller than the arithmetic means and approach the medians. However, the three parameters of the intertrack distances (the arithmetic means, medians and geometric means) for the various cell types and conditions follow essentially the same pattern and fall into three groups: the smallest values were found for the transformed cells-(wild type-transformed cells at 33 and 4o”C, and tsA239-transformed cells at 33°C) in complete medium (group I); the largest values for the same transformed cells in depleted medium (group II); and intermediate values for cells having the nontransformed phenotype (nontransformed cells and tsA239-transformed cells at 40°C) whether in complete or depleted medium (group Ill). The only exception was that the tsA239-transformed cells in depleted medium at 33°C had a somewhat smaller mean intertrack distance than the wild-type-transformed cells in depleted medium. Probability (P) values for the significance of the differences between the geometric means of the various sets of data are listed (pairwise) in Table 3. As can be seen, the differences in the intertrack distances clearly fall into the three groups indicated above and are highly significant (with the exception that the tsA239-transformed cells in depleted medium at 33°C are not significantly different from the nontransformed cells). Furthermore, when the combined means for the transformed cells in complete medium (group I) were compared
Replicons 863
Table
of CHL cells
2. lntertrack
Distances
in Nontransformed
and SV40-Transformed
Cell Line
Medium
Temperature
Number Distances
CHL
Complete
33°C
380
40°C
417
33°C
Depleted
CHLWT15
Complete
Depleted
CHLA239Ll
Complete
Depleted
Cells of Intertrack Measured
Median (+)
Arithmetic
Mean
Geometric
Mean
(CL)
(CL)
15.6
21.7
17.01
k 1.68
15.9
23.8
17.58
rt 1.71
429
15.8
22.2
17.14
t
40°C
400
14.8
21.9
16.14
2 1.64
33°C
354
11.0
16.1
13.10
2 1.58
40°C
453
12.7
18.2
14.28
+ 1.69
33°C
232
22.2
30.7
22.43
i 1.79
40°C
210
22.1
28.6
22.15
k 1.88
1.63
33°C
402
10.9
17.0
12.90
2 1.61
40°C
319
14.1
22.5
18.42
k 1.73
33°C
243
17.4
25.4
18.62
i 1.76
40°C
374
14.3
22.8
17.08
+ 1.56
with the combined means for the transformed ceils in depleted medium (group II), the difference was significant at P = 3.6 x 10e3. When group I was compared with the means for the cells having the nontransformed phenotype (group Ill), the difference was significant at P = 1 .O x 10p4. When group II and group III were compared, the difference in the intertrack distances was significant at only P = 0.075 because of the relatively small intertrack distance for the tsA239-transformed cells. A similar comparison between groups II and III excluding the &A239-transformed cells from group II was significant at P = 10m4. Measurement of the Rate of DNA Chain Elongation by Track Length The rate of DNA chain elongation should be represented by the length of tracks in the autoradiograms. We measured the lengths of the tracks in all the photographs obtained by method B above to determine the rate of chain elongation in nontransformed and transformed cells. Since these measurements were made at the same time as the distances between tracks were measured, this analysis also was “blind.” A typical distribution pattern is shown in Figure 4. Unlike the log-normal distributions of the distances between track centers, each of these sets of data could be approximated by a truncated normal distribution curve, indicating that the rate of chain elongation was approximately uniform in all forks. Means and standard errors were calculated (Table 4) on the sets and subsets. Very few significant differences (P < 0.05) in track lengths were detected when the sets were compared pairwise in all possible combinations. The tsA239-transformed cells grown at 33°C in depleted medium had a small mean track length (Table 4), which was significantly different from that of
most of the other cell lines grown under the different conditions. The mean track length of nontransformed cells grown in depleted medium at 40°C was significantly different from the mean track length of nontransformed and wild-type-transformed cells grown in complete medium at 4O”C, and tsA239-transformed cells grown in depleted medium at 40°C. All the other 54 pairwise t tests yielded no significant differences (P > 0.05) in track length. We do not understand the reason for the shorter track length in the tsA239-transformed cells grown in depleted medium at 33°C. It is the same cell line and growth condition that gave intertrack distances which were shorter than expected (see Tables 2 and 3). We were somewhat surprised that there was no consistent pattern of differences in track lengths, since we expected that the rate of chain elongation would be greater at 40 than at 33°C and possibly greater in complete than in depleted medium. We therefore grouped all the means of the cells grown at 40°C and compared them with the group consisting of all the means of the cells grown at 33°C. No significant difference in track length between the group at 40°C compared with the group at 33°C could be detected. When the means were grouped by growth medium (complete as opposed to depleted), there was also no significant difference between the track lengths of the cells grown in complete as opposed to depleted medium. Since all the cultures were pulse-labeled for 8 min, we calculated an average rate of chain elongation of 0.58 @min for all the cultures. Measurement of the Rate of Chain Elongation by the Method of Painter and Schaefer We wished to obtain an independent verification of the change in replicon size following transforma-
33°C 40°C 33°C 40°C
33°C 40°C 33°C 40°C
Complete
33°C 40°C 33°C 40°C
Complete
Depleted
Depleted
between
Temperature
t Tests
33°C >O.l”
40°C >O.l >O.l
33°C -co.01 O.l indicate no significant difference. bA significant difference, although expected in these pairwise tests, was not found. can be accounted for by the smaller than expected value for the intertrack distance cThese values are of borderline significance 0.1 > P > 0.05.
CHLA239Ll
CHLWT15
Depleted
Complete
Pairwise
CHL
from
Medium
3. P Values
Cell Type
Table
