Cell, Vol. 11, 719-727,

August

1977,

Copyright

0 1977 by MIT

Mouse L Cell Mitochondrial DNA Molecules Are Selected Randomly for Replication throughout the Cell Cycle Daniel Bogenhagen and David A. Clayton Laboratory of Experimental Oncology Department of Pathology Stanford University School of Medicine Stanford, California 94305

Summary The number of mitochondrial DNA molecules in a cell population doubles at the same rate as the cell generation time. This could occur by a random selection of molecules for replication or by a process that ensures the replication of each individual molecule in the cell. We have investigated the rate at which mouse L cell mitochondrial DNA molecules labeled with 3H-thymidine during one round of replication are reselected for a second round of replication. Mouse L cells were labeled with 3H-thymidine for 2 hr, chased for various periods of time and then labeled with 5-bromodeoxyuridine for 4 hr immediately before mitochondrial DNA isolation. A constant fraction of 3Hthymidine-labeled mitochondrial DNA incorporated B-bromodeoxyuridine after chase intervals ranging from 1.5-22 hr. This result demonstrates that mitochondrial DNA molecules replicated in a short time interval are randomly selected for later rounds of replication, and that replication of mitochondrial DNA continues throughout the cell cycle in mouse L cells. Introduction A mouse L cell contains approximately 1000 mitochondrial DNA (mtDNA) molecules (Nass, 1969; Bogenhagen and Clayton, 1974). Since the biosynthesis of an individual mtDNA molecule requires approximately 5% of the ceil generation time (Berk and Clayton, 1974; D. Bogenhagen and D. A. Clayton, manuscript in preparation), it is important to understand the mechanism whereby the mammalian cell coordinates the doubling of its mtDNA content during the cell cycle. It is conceivable that the cell may replicate each mtDNA molecule once and only once in one generation. Alternatively, a few molecules may be selected to replicate early in the cell cycle, and only their progeny may be involved in further rounds of replication. The latter situation could result if the complete replication apparatus were contained in only a few of a cell’s mitochondria or if some limiting factor necessary for replication were vertically transmitted to one daughter molecule. The experiments reported here determined whether recently synthesized mtDNA molecules are predisposed for or against re-repli-

cation, or whether selection of molecules for replication is a random process. The basic approach in these experiments, as outlined in Figure 1, is to label mouse LA9 cell mtDNA for 2 hr with 3H-thymidine, then to chase with nonradioactive thymidine for a varied amount of time and finally to label with 5bromodeoxyuridine (BrdUrd) for 4 hr. Buoyant density analyses can then be used to determine the fraction of radioactive mtDNA shifted in buoyant density by a second round of replication in the presence of BrdUrd. Since the time required for biosynthesis of a mtDNA molecule is approximately 1 hr, mtDNA molecules which begin replication during the final hr of BrdUrd labeling will not have completed replication before mtDNA isolation. Thus these molecules will not be isolated as closed circular molecules in the lower band of the propidium diiodideCsCl gradient. Thus the effective BrdUrd labeling time is approximately 3 hr in this experiment. In principle, this experiment is similar to a continuous BrdUrd labeling with measurement of the production of “heavy-heavy” molecules, as performed by Flory and Vinograd (1973). The experimental design used here, however, avoids toxicity associated with prolonged exposure of cells to BrdUrd (see Hakala, 1959; Simon, 1961). Moreover, as shown below, both 3H-thymidine and BrdUrd labeling can be performed discretely in LA9 cells without media changes and without interruption of exponential growth. Mathematical analysis of the results is straightforward, since the data can be fitted to a horizontal line rather than to a more complicated exponential expression. If time 0 is defined as the end of the 3H-thymidine labeling period (Figure l), the subsequent increase in the number of mtDNA molecules is given by N(t) = N(O)e”” (CY= In2/Tg; Tg = doubling time = 18 hr). During a 3 hr period of BrdUrd labeling, the increase of mtDNA content of the culture is AN = N(t = ,,,(0)eKt+3,1nZ/181 N(0)et1”2”8 = + 3) - N(t) N(t)(e31nn’is - 1 ) = O.l22N(t). Thus a constant fraction, 12.2%, of molecules present at time t will enter replication during a fixed 3 hr period of BrdUrd labeling. If molecules replicated previously in the short period of 3H-thymidine labeling are randomly reselected from the total mtDNA population for later rounds of replication, the fraction of radioactively labeled mtDNA shifted to hybrid density will be a constant 12.2% regardless of the duration of the intervening chase. The results of this analysis have a direct bearing on the question of cell cycle restrictions on mtDNA synthesis. If mtDNA molecules were replicated in a discrete portion of the cell cycle, one should observe a periodicity with respect to their re-replica-

Cell 720

Time,

hours

-2 f Chase I,//

Figure

1. Experimental

Design

$H-thymidine (50 &i/l, 50-70 CVmmole) was added to an asynchronous exponentially growing culture of mouse LA9 cells at t = -2. The chase was begun at time 0 with addition of 20 PM thymidine. At time = t, 50 PM BrdUrd was added for 4 hr. The lower half of the figure illustrates the sequence of replication events (depicted by arrows) required to produce mtDNA molecules containing both 3H-thymine and BrUra. Replication in the presence of 3H-thymidine (arrow 1) produces two molecules containing 3H-thymine in one strand (dashed lines). Later replication of such a molecule in the presence of BrdUrd (arrow 2) produces one molecule containing 3H-thymine in one strand and BrUra (heavy line) in the other, as well as a cold hybrid density molecule which is not detected experimentally. Arrows 3 and 4 show that an additional round of replication during the chase is also not detected experimentally.

tion. Correspondingly, there should be definable intervals in which few, if any, mtDNA molecules are re-replicated. In contrast, if mtDNA synthesis is independent of the cell cycle, then a constant number of mtDNA molecules should be in the process of replication in all cells at all times. The results reported here are consistent with this independent model. The accompanying Appendix demonstrates that the experimental design used would have obtained significantly different results if mtDNA synthesis were restricted to the S phase or to the S + G2 phases of the cell cycle.

Results and Discussion Several control experiments were performed to determine conditions permitting 3H-thymidine labeling and an effective chase with minimal disturbance of normal cellular events. In the first experiment, a spinner culture of LA9 cells was labeled with a single dose of 0.5 mCi/l of 3H-thymidine. Aliquots were removed at several times for mtDNA isolation. The rate of labeling of mtDNA was deter-

mined by normalization for a constant yield of longterm 14C-thymidine prelabel in mtDNA. Figure 2 shows that the rate of 3H-thymidine incorporation into closed circular DNA was linear for approximately 2 hr, but was greatly diminished during continued labeling. Addition of 20 FM cold thymidine after 2 hr abruptly decreased the labeling rate to ~6% of the pulse labeling rate without transfer of the cells to nonradioactive medium. This pattern of labeling is consistent with a rapid uptake and metabolism of high specific activity 3H-thymidine from the medium. Figure 3 shows that the half-life of carrier-free 3H-thymidine (50-71 Ci/mmole) in the medium is approximately 1 hr over the entire range of 0.05-2 mCi/l (approximately l-40 nM thymidine). This rate is only a measure of the uptake of 3Hthymidine into LA9 cells, but does serve as an indication of how rapidly isotope may be either incorporated or metabolized to molecules which are not effective precursors of DNA (Lang, Mijller and Maurer, 1966). Given such rapid uptake of 3H-thymidine, it was necessary to determine whether 3H-thymidine pro-

Random 721

Replication

of Mitochondrial

DNA

cl3-

3-

I

0

lil Figure

0,

2. Efficiency

2

q

LWiELlNE

TIME/

of the Chase

0

.!i

I

I .0 T I ME/

I

I

I .s HUUR5

2.0

2.5

Ii Figure

HOURS

of 3H-Thymidine

Labeling

A “C-thymidine-prelabled culture of LA9 cells was labeled with 0.5 mCi/l 3H-thymidine (60 Cilmmole). After 2 hr, the chase was initiated by addition of 20 PM thymidine. mtDNA was isolated at several time points during continuous labeling (O-O), and at 2 and 4 hr of chase (O----O). 3H incorporation into mtDNA is plotted after normalization to a constant yield of 3000 cpm 14C prelabel. The rate of residual incorporation in the chase is 56% of the pulse labeling rate. The dose of 3H-thymidine in this experiment was 10 times that used in re-replication experiments (for example, Figure 5).

duced toxic effects such as those observed by Whitmore and Gulyas (1966). These investigators observed a period of growth arrest beginning approximately 4 hr after addition of 3H-thymidine, suggesting that cells which incorporated 3H-thymidine into nuclear DNA in the S phase could complete the G2 phase of the cell cycle but could not divide at mitosis. This mitotic delay was presumed to be due to accumulation of chromosomal damage (Wimber, 1959). To our knowledge, a reinvestigation of these phenomena using the high specific activities of 3H-thymidine presently available commercially has not been reported. We monitored the growth in suspension of LA9 cells exposed to varied concentrations of carrier-free 3H-thymidine. Quantities of 3H-thymidine >0.2 mCi/l resulted in growth inhibition of dose-dependent severity. No significant growth lag was observed, however, in any of several experiments reported below using 0.05 mCi/l 3H-thymidine. At this concentration, approximately 20,000 cpm were incorporated into mtDNA in 2 hr per liter of cells at a density of 4 x 105/ml. In a study of BrdUrd incorporation into HeLa cell mtDNA, Flory and Vinograd (1973) isolated mtDNA using ethidium bromide-CsCl gradients. In that study, DNAase treatment of mitochondria and velocity sedimentation purification of mtDNA were

3. Depletion

of 3H-Thymidine

from

Medium

by LA9 Cells

The procedure for measurement of the rate of uptake of JHthymidine from growth medium is given in Experimental Procedures. The dosages of 3H-thymidine (71 Ci/mmole) were 2 mCi/l (A-A), 0.5 mCi/l (O--O), 0.05 mCi/l (O-O), 0.5 mCi/l plus 10 mM carrier thymidine (O----O). A single least-squares line is drawn through all data points from samples without carrier thymidine.

necessary for efficient purification of mtDNA from nuclear DNA-containing bromouracil (BrUra). These procedures were unnecessary in our experiments for three reasons. First, the use of propidium diiodide-CsCI gradients increased the separation of unsubstituted closed circular mtDNA from nuclear DNA by a factor of 1.8 relative to ethidium bromide-CsCI (Hudson et al., 1969), providing an absolute buoyant density difference of about 70-75 mg/ml. Second, nuclear DNA containing BrUra in both strands is not produced in these experiments since the duration of BrdUrd labeling is only 4 hr. Thus the theoretical maximum density shift for nuclear DNA containing 16% BrUra (that is, fully substituted in one strand) is only 64 mg/ml (Wake and Baldwin, 1962). Finally, complete substitution of BrUra for thymine is not expected under the conditions of these experiments, since the 20 PM thymidine added to chase 3H-thymidine label was not removed before addition of 80 PM BrdUrd to begin density labeling. Since a variable amount of cold thymidine was probably metabolized during the chase, the composition of the density labeling media was somewhat greater than 80% BrdUrd:20% thymidine. An example of the separation between unsubstituted mtDNA and hybrid nuclear DNA is shown in Figure 4. In these experiments, the chase period intervening between 3H-thymidine and BrdUrd labeling varied from 1.5-22 hr. Since it was necessary that all experimental procedures be performed while cultures were in logarithmic growth, 3H-thymidine

Cdl 722

Figure DNA

4. Isolation

of

mtDNA

from

BrUra-Substituted

Nuclear

mtDNA was isolated from LA9 cells labeled for 4 hr with 80 PM BrdUrd plus 20 PM thymidine. mtDNA was isolated in a propidium diiodide-CsCI gradient centrifuged 44 hr at 45,000 rpm, 2O”C, in a 75Ti rotor. SrUra-substituted nuclear DNA forms a discrete band intermediate in density between the lower band of closed circular unsubstituted mtDNA and the upper band of unsubstituted nuclear DNA.

labeling was routinely performed at a cell density of 4 x 105/ml. LA9 cells labeled under these conditions incorporate 3H-thymidine into an unstable 500 nucleotide single-strand sequence, 7s mtDNA, associated with the displacement loop of D mtDNA (Kasamatsu, Robberson and Vinograd, 1971; Berk and Clayton, 1974), as well as into full-length strands of mtDNA. If 7s mtDNA is incorporated into full-length strands of mtDNA by expansion synthesis in the presence of BrdUrd, then any 3Hlabeled 75 DNA serving as a primer would contribute to the amount of radioactivity shifted in density in these experiments. We have determined that after 2 hr of labeling, only 9% of the incorporated isotope in mtDNA is in 7s mtDNA, and that the mean lifetime of 7s mtDNA in LA9 cells is only 90 min (D. Bogenhagen and D. A. Clayton, manuscript in preparation). Thus the majority of 7s mtDNA strands are lost by turnover before the beginning of BrdUrd labeling and are not a factor in these experiments. Three sets of density shift experiments were performed with suspension cultures of 3-6 liters. The 3H-thymidine labeling and chase were performed without transfer of cells from the original culture vessel. One liter aliquots were transferred at various chase times to prewarmed spinners containing 24.6 mg BrdUrd. mtDNA was isolated 4 hr later. CsCl gradients of mtDNA from one set of these experiments are shown in Figure 5 (density gradients of the corresponding nuclear DNA samples are shown as insets). These experiments are repre-

sented by triangular data points in the scatter diagram of Figure 6, along with the results from the two other experimental sets, represented by square and x symbols. For comparison, Figure 6 also shows the fraction of radioactive nuclear DNA shifted in density (smooth curve) in the same experiment which provided the mtDNA samples of Figure 5. The fraction of nuclear DNA shifted is maximal at a chase time approximately equal to one generation time. This result is a further indication that the traverse of 3H-labeled cells through the cell cycle is not interrupted by the level of 3Hthymidine incorporation in these experiments. The results obtained with mtDNA are strikingly different from those with nuclear DNA. The fraction of mtDNA shifted was relatively constant at all chase times used, with a mean and standard deviation of 8.9 -C 1.7%. This fit to a horizontal slope satisfies the critical requirement of the random selection model. The mean value, however, is slightly lower than the 12.2% shift pred,icted above. This discrepancy is probably due to ‘the, catenation of molecules labeled with both 3H-thymine and BrUra to light density mtDNA molecules. Berk and Clayton (1976) have shown that ~2.5 hr is required for the equilibration of recently replicated molecules with the pool of catenated mtDNA, which represents 23% by mass of total mtDNA. Thus approximately 23% of 3H + BrdUrd-labeled mtDNA monomers may be shifted to a “quarter heavy” buoyant density intermediate between hybrid and light density mtDNA. This region of the CsCl gradients was not included in the calculation of the fraction of mtDNA at hybrid density. With this correction, a constant 9.4% shift is expected if molecules are randomly selected for replication, in good agreement with the observed result. A similar random selection for replication has been observed for several autonomously replicating bacterial plasmids and viruses [for example, Proteus mirabilis R factors NRI (Rownd, 1969), Rtsl (Terawaki and Rownd, 1972) and R12 (Morris et al., 1974); E. coli colicin El (Bazaral and Helinski, 1970); and Pseudomonas phage PM2 (Espejo, Canelo and Sinsheimer, 1971)]. In each of these systems, multiple copies of the circular DNA are present per cell. Rownd (1969) has emphasized that under this condition, there is presumably no stringent requirement for replication of any individual molecule, resulting in a “relaxed” control of replication. Thus this random selection pattern may be the direct result of replication with DNA template in excess, so that some other factor becomes ratelimiting. It is clear that the random reentry of mtDNA molecules into the replication process requires continuous synthesis of mtDNA throughout the LA9 cell cycle. Several studies of the synthesis of mtDNA in

Random 723

Replication

of Mitochondrial

DNA

1-i I0

20 FRRCTION

30 NUMBER

Id

q0

z?i7

: I0

20 FRRCT I UN

NLlM3i0ER

20 FRRCTION

30 NUMBER

90

/

6

I B

10

20 FRRCTllIlN

Figure

5. CsCl

Gradients

30

q0

NUMBER

of 3H-Thymine-BrUra-Labeled

I0

vz

DNA

The pulse-chase-density labeling protocol was as in Figure 1. The large figures are radioactivity profiles of buoyant CsCl gradients (45,000 rpm, 75Ti rotor) of mtDNA labeled with BrdUrd. The duration of the chase was 1.5 hr (A), 7.5 hr (B), 11.5 hr (C)and 20.5 hr (D). The insets are plots of ethidium bromide-CsCl gradients (30,000 rpm. SW50.1 rotor) of the corresponding nuclear DNA samples. Gradient density increases to the left. All data are plotted as “% total cpm” for ease of comparison. The total cpm in each case were (A) 16,700. inset 36,700; (B) 16,500, inset 353,700; (C) 16.200. inset 101,000; (D) 15,400, inset 69,160. Data from this and two similar experiments are combined in Figure 6

synchronized cell cultures have suggested, however, that mtDNA synthesis may not occur in some fraction of the animal cell cycle (Koch and Stokstad, 1967; Pica-Mattoccia and Attardi, 1972; Ley and Murphy, 1973). There are several reasons why these results cannot be considered definitive. Both Koch and Stokstad (1967) and Ley and Murphy (1973) made no attempt to highly purify mtDNA from nuclear DNA. This is of importance since both groups reported a maximal rate of mtDNA synthesis during the S period, which is the result which would be obtained if mtDNA were contaminated with nuclear DNA. In addition, the species considered to represent mtDNA in the early study of Koch and Stokstad (1967) was later found to be a nuclear DNA component (Koch, 1969). Pica-Mattoccia and Attardi (1972) observed mtDNA synthesis throughout the cell cycle in HeLa cells synchronized by

double-thymidine block. The rate of mtDNA synthesis was reduced in the Gl period, however, when synchronized cultures were obtained by mitotic selection. The cell cultures in their experiments were stored at 2°C for variable amounts of time, a procedure which has deleterious effects during the subsequent Gl period (Ehmann and Lett, 1972). Thus the depressed rate of incorporation of 3H-thymidine into mtDNA during this period may have resulted from cold shock to the mtDNA replication apparatus. We have developed a model to estimate the results which would be expected in our experiments if mtDNA were synthesized in only a portion of the cell cycle (see Appendix). If mtDNA synthesis occurred only in the nuclear S period, the graph of “% mtDNA shifted in density” versus “chase time” (Figure 6) would be similar to the markedly peaked

Cell

724

3H-thymidine labeling studies in stationary cell cultures (Vesco and Basilica, 1971; Levine, 1971) and in a BHK cell line which is temperature-sensitive for nuclear DNA synthesis (Burstin, Meiss and Basilice, 1974). Additional controls are necessary, however, to exclude the possibility of a contribution of specific activity artifacts to these results. Appendix

Figure 6. Fraction tion of the Chase

of DNA Shifted Time

in Buoyant

Density

as a Func-

The scattered symbols (x. A, n ) represent the results from mtDNA isolations from three different experiments. The triangular symbols (A mtDNA; A nuclear DNA) represent the data in Figure 5. The mtDNA sample corresponding to the nuclear DNA sample at a chase time of 16 hr was lost due to a defect in the centrifuge tube. A smooth curve was drawn through the data points for nuclear DNA using an empirical curve-fitting program. The horizontal line at 9.4% is the prediction of the random selection model.

curve for nuclear DNA. If an increasingly greater fraction of the cell cycle were allowed for mtDNA synthesis, this peaked curve would tend to flatten toward the horizontal line expected if mtDNA were synthesized continuously. This mathematical model has enabled us to estimate the results which would be expected if mtDNA synthesis did not continue during the Gl stage. We conclude that these expected results,.discussed in the Appendix, are significantly different from the observed results. In other words, our experimental protocol is designed such that these experiments would have detected discontinuous synthesis of mtDNA of the type proposed by Pica-Mattoccia and Attardi (1972). Continuous synthesis of mtDNA throughout the cell cycle has been observed in several other systems. Incorporation of 3H-thymidine into mtDNA has been observed using autoradiography in chick fibroblasts (Meyer and Ris, 1966), Tetrahymena pyriformis (Parsons and Rustad, 1968) and Physarum polycephalum (Guttes, Hanawalt and Guttes, 1967). The results with P. polycephalum were confirmed by centrifugal analysis and in independent studies by Braun and Evans (1969). Carter (1975) has recently reviewed experiments relating to the coordination of mtDNA and nuclear DNA synthesis in yeast. In addition, mtDNA synthesis has been found to continue in the absence of nuclear DNA synthesis in mouse L cells treated with 5-fluorodeoxyuridine or methotrexate (Bogenhagen and Clayton, 1976). A similar lack of close coordination of mtDNA and nuclear DNA synthetic rates has been inferred from

Model for Replication of mtDNA in S and 62 Phases of the Mouse L Cell Cycle On the basis of cell synchronization experiments, Pica-Mattoccia and Attardi (1972) have suggested that mtDNA replication may be restricted to the S and G2 phases of the HeLa cell cycle. Ley and Murphy (1973) have suggested an even greater restriction of mtDNA synthesis to only the S phase of CHO cells. The calculations presented below show that our experimental design would discriminate among three alternative models: replication throughout the cell cycle; replication restricted to S and G2 phases of the cell cycle: mtDNA replication during S phase only. It is important to note that each of these models assumes that recently replicated mtDNA molecules may be randomly selected from the total popuiation for another round of replication. Thus a molecule produced by replication in early S phase could be re-replicated later in S in any of these models. The durations of cell cycle compartments are assumed to be proportional to the values observed by Stanners and Till (1960) for mouse L cells, with an adjustment for the fact that we observe a doubling time, Tg, of 18 hr. Thus Gl = 8.1 hr, S = 6.3 hr and G2 + M = 3.6 hr. In this analysis, G2 and M are considered as one postS period designated G2’. Variability in the duration of cell cycle compartments is not considered in this Appendix. This variability has been considered in detail for general questions regarding cell cycle kinetics (Nachtwey and Cameron, 1968) and would probably become appreciable only after 12-18 hr of chase. We also assume that incorporation of 3H-thymidine or BrdUrd begins instantaneously upon addition of each compound and is strictly proportional to the rate of mtDNA synthesis regardless of cell cycle stage. This assumption therefore neglects possible variations during the cell cycle in the transport or phosphorylation of thymidine or of thymidylic acid. For any cell, a cell cycle parameter, 7, can be defined as the time elapsed since the last cell division. Thus a cell at the GllS boundary has 7 = Gl = 8.1 hr. This cell was produced by division 8.1 hr previously and will progress to mitosis in S + G2’ hr. The fraction of total cells at a given stage 7 of the cell cycle, n(T), is n(T) = 2ae-“‘, where a = InP/Tg. Note that the constant 201 is obtained since n(T) is a normalized function of 7 such that: Tg I0

n(T)dT = 1

(1)

The three models above correspond to three different kinetic patterns for the increase in the number of mtDNA genomes per cell, m(T). during the cell cycle. Regardless of the model, every cell must double its postmitotic content of mtDNA, c, before the next cell division. Each of these models can be characterized by two relations. First, a probability function r(T) is used to define the portion of the cell cycle during which mtDNA synthesis is permitted. The value of r(T) equals zero when mtDNA is not synthesized, and equals one when synthesis is allowed. Second, a constant, H, is defined for each model as the absolute rate of mtDNA synthesis during periods when mtDNA synthesis is allowed. For example, in model 1, since each cell must increase in mtDNA content from the postmitotic content, c, at 7 = 0 to 2c, at 7 = Tg, H = c/Tg. In contrast, in model 3, mtDNA synthesis is permitted only in S, so that the rate of synthesis during S must be correspondingly greater and H = c/S. These relationships are illustrated in Figure 7.

Random 725

Replication

Model

of Mitochondrial

DNA

number

Synthetic

period

G,,

mtDNA content/cell,

3

2

1

S

S, G2’

S, G2'

m(T):

2c

2c

C

C

I--

P 0. probability

0

-b

T

Tg T

7

of mtDNA synthesis,

r(T):

L

1

1

Lz OO

OO

Tg

0

Tg

Tg T

T

rate

of mtDNA synthesis

during

the

synthetic

H = c/Tg Figure

7. Definitions

of the Cell Cycle

Dependence

H = c/(S+G2')

of mtDNA

Content

In these calculations, the time scale is defined so that 3H-thymidine labeling begins at time t = -2 hr and continues until t = 0 (Figure 1). Each cell is identified throughout the experiment by its stage in the cell cycle, TV, at t = 0. For any value of TV, the total number of mtDNA molecules made during the 2 hr 3H-thymidine labeling (from t = -2 tot = 0) is dependent upon the total number of cells, N(t) n(T,,); the probability that mtDNA synthesis is occurring, r(T,, + t); and the rate of mtDNA synthesis, H, for each model. Thus the number of radioactive mtDNA molecules (strands) contained in cells with cell cycle parameter 70 is defined as P(T~): 0 R%) =

i -2

The total number of radioactive cells during the pulse is: jTg P(~,)dr, 0

(2)

N(t) n(TO) H r(TO + t) dt

= iTZ 1’ N(t) n(TJ 0 --2

mtDNA

molecules

H r(~ + t) dt ho

made

period:

by all

(3)

During the chase interval following =H-thymidine labeling, this distribution P(TJ is advanced to P(T) = P(T~ + t), where t = chase time. P(TJ and P(T,, + 3) are illustrated for model 3 in Figure 8. BrdUrd labeling begins at the end of the chase, \, and continues until t, + 3. During BrdUrd labeling, 3H-thymine-containing mtDNA molecules may be randomly selected for replication from the total mtDNA in a cell. Each such round of replication will produce a molecule labeled with 3H-thymine in one strand and ErUra in the other. In any cell, the number of radioactive/hybrid density molecules produced is determined by the following: first,

and Synthesis

for Each

H = c/S‘

Model

the probability that a given molecule contained in a cell at cycle stages 70 + t is radioactive; this is also the fraction of molecules which are radioactively labeled

wo

F(T~ + t) = ~~ N(t) nb)

+ 0

(4)

m(T)

where m(T) is defined in Figure 7 as the mtDNA content of a single cell at stage 7; second, the probability that this cell is able to synthesize mtDNA. r(TO + t); and third, the rate of mtDNA synthesis, H. In general, we define G(T,, + t) as the number of radioactive molecules contained in cells at stage 70 of the cell cycle at t = 0 shifted in density by BrUra incorporation: G(T~ + t) = jtci3 k Substituting

from

G(T,, + t) = jk+3 te Integrating cycle gives molecules:

F(T~ + t) N(t) n(T) H r(TO + t) dt equation y

(5)

four:

(6)

H r(TO + t) dt

over the total number of cells at each the total number of radioactive/hybrid

stage of the cell density mtDNA

Then at a given chase time, the fraction of all radioactive shifted to hybrid density by BrUra labeling, F(t), is:

mtDNA

Cell 726

\

I

q

I I

Ii

I2

\

I If3

TRU Figure

6. Example

of the Effect

of the Chase

The solid line, P(rO), is the distribution of radioactive mtDNA molecules within cells at various stages of the cell cycle at the end of 3H-thymidine labeling, as calculated for model 3. The broken line (-.-) illustrates the effect of a 3 hr chase. Cells containing radioactive mtDNA have advanced 3 hr through the cell cycle and some cells have undergone mitosis.

Figure 9. Mathematical Predictions for Three Models for the Cell Cycle sis

shifted

(6)

= +p I0

P (Q + t) dT,

This is the formal definition of the experimental results shown in Figure 6 of the text. With the use of a computer, the above relations were solved for each of the three models under consideration for chase times of 1-29 hr. The results of all three models are shown in Figure 9. The experimental data are in good agreement only with the predicted data for model 1, mtDNA synthesis throughout the cell cycle. For model 2, mtDNA synthesis in S + G2’, there is a prominent minimum at t = 7 hr which results from the large fraction of cells containing 3H-mtDNA which are in Gl for most or all of the BrdUrd labeling period. The initially high value of F(t) at t = 0.5 hr reflects cells which are in one S + G2’ period throughout the experimental protocol. We conclude that these expected results are in disagreement with our experimental data. The predictions of model 3, mtDNA synthesis only in S, are even more clearly excluded. It is important to consider what effects would be introduced by cell cycle-specific variations in the specific activity of mtDNA labeling. For example, Pica-Mattoccia and Attardi (1972) estimated that 3H-thymidine is more avidly incorporated into mitochondrial TTP pools in HeLa cells in the S phase of the cell cycle. If a similar variation occurs in LAS cells, the %H-thymine labeling of mtDNA in our experiments would overemphasize mtDNA synthesis in S phase cells. Such a situation would be comparable to postulating an increased rate of synthesis of mtDNA during S and would also tend to produce a deviation of the experimental data from the random re-replication model. We conclude that any great degree of variation in the rate of mtDNA synthesis during the cell cycle would be detectable in our experiments and would not be masked by specific activity effects. Such specific activity effects are relatively unimportant during the final BrUra labeling of mtDNA, since the 60 PM BrdUrd concentration used produces a large, fairly homogeneous density shift (Figure 5).

Results Synthe-

Equation (8) was used to calculate the expected fraction of radioactive mtDNA shifted to hybrid density at various chase times assuming one of three models: mtDNA synthesis throughout the cell cycle (-); mtDNA synthesis in S and G2’ (O-O); mtDNA synthesis only in the S phase of the cell cycle (A-A).

Experimental Fraction

of the Experimental Dependence of mtDNA

Procedures

Cell Growth and Labeling Methods for suspension culture of mouse LAS cells in Eagle’s minimal essential medium have been described (Bogenhagen and Clayton, 1974). Cell growth was monitored 2 or3 times daily using a Coulter counter. The cell doubling time was 18 f 2 hr during logarithmic growth. 3H-thymidine was obtained at 50-71 Ci/ mmole (SchwarzIMann). Uptake experiments were performed by removing 2 ml samples at various times from 100 to 200 ml suspension cultures labeled with the indicated concentrations of 3H-thymidine. Spinner flasks were tightly closed in the intervals between sampling to avoid pH variation. Cell samples were immediately chilled on ice and centrifuged for 3 min at 2000 rpm in a Sorvall GLC-1 centrifuge. Aliquots of lo-50 ~1 of the supernatant were pipetted onto Whatman GF/A filters, which were dried and immersed in 5 ml of a solution of 4 g Omnifluor (NEN) per liter toluene for radioactivity determination in a Beckman LS-230 scintillation counter. Radioactivity measurements of DNA samples were made similarly, except that aliquots of 80 ~1 or greater were placed on GF/B filters and washed with 5% CCI,COOH and ethanol. DNA

Isolation

and

Buoyant

Density

Analysis

mtDNA was isolated without sucrose gradient purification of mitochondria as described by Bogenhagen and Clayton (1974). Mitochondrial lysates were adjusted to contain 0.5 mglml propidium diiodide (Calbiochem) in a lo-11 ml vol of CsCl solution of p = 1.54. Gradients were formed by centrifugation for 244 hr in cellulose nitrate tubes at 45,000 rpm in a 75Ti rotor at 20°C. Regions of the gradient containing closed circular DNA were estimated either visually or by assay of radioactivity. Samples were freed of dye by passage through columns of Dowex-BOW cation-exchange resin (0.6 x 4 cm) without loss of mtDNA. CsCl was added to adjust the density to 1.70 in a final volume of 8-S ml. Centrifugation was performed as above. Fractions of 0.08 ml were collected through the tube bottom for determination of radioactivity.

Random 727

Replication

of Mitochondrial

DNA

Nuclear DNA samples were prepared from nuclei dispersed in 0.1 M NaCI, 50 mM Tris, 10 mM EDTA (pH 8.3), and lysed in 0.6% sodium dodecylsulfate. Nuclear DNA samples were sheared by several passages through a 22 gauge needle. The fraction of nuclear DNA radioactivity shifted in buoyant density by BrUra substitution was measured in buoyant CsCl (,J = 1.70) or in ethidium bromide-CsCI (0.3 mg/ml, p = 1.54) gradients. Ethidium bromide was added to permit visualization of the nuclear DNA bands. Adequate resolution was obtained in ethidium bromideCsCl gradients despite a 35% reduction in the buoyant separation of BrUra substituted from unsubstituted DNA relative to the separation in the absence of dye. Acknowledgments We thank P. A. Martens and J. Battey for assistance in the ration of the Appendix, and A. J. Berk and C. A. Smith for ing the manuscript prior to publication. This investigation supported by grants from the NIH and the American Society. D.B. is a Medical Scientist Training Program (NIH) and D.A.C. is a Faculty Research Awardee (ACS). Received

March

21, 1977;

revised

April

prepareviewwas Cancer Fellow

29, 1977

References Bazaral,

M. and Helinski,

D. R. (1970).

Berk,

A. J. and Clayton,

D.A.

Berk,

A. J. and Clayton,

D. A. (1976).

Biochemistry

(1974).

9,399-406.

J. Mol. Biol. 86, 801-824. J. Mol. Biol.

700, 85-102.

Bogenhagen, 7991-7995.

D. and Clayton,

D. A. (1974).

J. Biol.

Chem.

249,

Bogenhagen, 2938-2944.

D. and

D. A. (1976).

J. Biol.

Chem.

257,

Clayton,

Braun, R. and Evans, 511-522.

T. E. (1969).

Burstin, S. J., Meiss, 84, 397-407. Carter,

H. K. and Basilica,

8. L. A. (1975).

Ehmann,

U. K. and

Biophys.

C. (1974).

Exp.

P. J., Jr. and Vinograd,

J. (1973).

E. W., Hanawalt, P. C. and Acta 742, 181-194.

Hakala,

M. T. (1959).

J. Brol.

234,

S. (1967).

J. (1969).

J. and Stokstad,

Lang, W., 645-648. Levine,

Eur. J. Biochem.

Mtiller,

A. J. (1971).

Proc.

Ley, K. D. and Murphy, Meyer,

R. R. and Ris,

Nachtwey, Physiology, Parsons,

Maurer,

Nat. Acad.

H. (1966).

J. (1969).

J. (1971).

Eur. J. Biochem.

W. (1966).

M. M. (1973).

Morris, C. F., Hashimoto, Bacterial. 778, 855-866.

Nass,

Vinograd,

Sci.

3, l-6.

Cell Res.

J. Mol.

J. A. and Rustad,

Pica-Mattoccia, 484.

USA 68, 717-720.

Biol. 37, 76A.

S. and Rownd,

R. (1974).

J.

in Cell

Biol. 42, 521-528.

R. C. (1968).

L. and Attardi,

44,

J. Cell Biol. 58,340-345.

J. Cell

H., Mickel,

Exp.

D. S. and Cameron, I. L. (1968). In Methods 3, D. M. Prescott, ed., pp. 213-259.

M. M. K. (1969).

Proc.

11, 296-304.

E. L. R. (1967).

D. and

Biochim.

3072-3076.

H., Robberson, D. L. and Vinograd, Sci. USA 68, 2252-2257.

Koch,

J. Mol.

J. Mol. Biol. 74, 81-94. Guttes,

Chem.

74, 9-14.

R. L. (1971).

Hudson, B., Upholt, W. B., Devinny, J. and Proc. Nat. Acad. Sci. USA 62, 813-820.

Koch,

182,

J. Cell. Physiol.

Cell Res.

E. S. and Sinsheimer,

Guttes, Biophys.

Kasamatsu, Nat. Acad.

Acta

Cell 6, 259-268.

Lett, J. T. (1972).

Espejo, R. T., Canelo, Biol. 56, 597-621. Flory,

Biochim.

J. Cell Biol. 37,683-

G. (1972).

J. Mol.

Biol.

693. 64, 465-

Rownd,

R. (1969).

Simon,

J. Mol.

E. H. (1961).

Stanners, 406-419.

Biol. 3, 101-109.

C. P. and Till, J. E. (1960).

Terawaki,

Y. and Rownd,

Vesco,

C. and Basilica,

Wake,

R. G. and Baldwin,

Whitmore. Wimber,

Biol. 44, 387-402.

J. Mol.

R. (1972). C. (1971).

Proc.

Biophys.

J. Bacterial. Nature

R. L. (1962).

G. F. and Gulyas, D. E. (1959).

Biochim.

S. (1966). Nat. Acad.

Acta 37

709, 492-498.

229, 336-338.

J. Mol. Biol. 5,201-216. Science Sci.

757, 691-694.

USA 45, 839-846.