JOURNAL OF CELLULAR PHYSIOLOGY 142:365-371(19901

Inhibition of Replicon Cluster Ligation Into Chromosomal DNA at Elevated Temperatures RAYMOND L. WARTERS" AND BRADLEY W. LYONS Department of Rdd/o/ogy, UniversiLy of Utah Health Science5 Center, Sah Lake City, Utah 84732 The rate-limiting enzymatic step for DNA replication in HeLa cells incubated al 43.5"C was the ligation of clusters o l replicons into the cell's genome. At 43.5"C: thc rcciprocal slope for inhibition of DNA chain (replicon) initiation, or of the ligation of replicon clusters into the genome, was 18 or 7 min, respectively. The failure of replicon clusters to be ligated into chromosomal DNA was not a consequence of the failure of histone proteins to be deposited onto replicating DNA, or of chromatin replicated at 43.5"C to be organized into fully condensed chromatin. In addition it was not due to the failure of fully active topoisomerase I1 to be deposited at a normal frequency along replicating chromatin DNA. The failure of replicon clusters to be ligated into the genome resulted in the persistence of single, but not double, DNA strand breaks in the cell's genome 24 hours after cell heating.

Exposure of mammalian cells to elevated (hyperthermic) temperatures (43-48°C) has been demonstrated to inhibit a number of subcellular metabolic processes, including DNA, RNA, and protein synthesis (Hahn, 1982; Roti Roti and Laszlo, 1988). Exposure of cells to elevated temperatures inhibits the semiconservative replication of DNA measured either as the inhibition of pulse incorporation of radiolabeled DNA precursors (Mondovi et al., 1969; Henle and Leeper, 1982a,b),the inhibition of the initiation of new replicons (Wong and Dewey, 1982; Warters and Stone, 1983, 1984), the inhibition of the reorganization of replicating DNA into mature, bulk chromatin structure (Warters and Roti Roti, 1981), and the processing of replicating DNA a t the nuclear matrix (Warters, 1988). Previously hyperthermic inhibition of de novo replicon initiation was considered to be the most heat-sensitive DNA replicative process, and thus rate limiting for DNA synthesis at elevated temperatures (Wong and Dewey, 1982; Warters and Stone, 1984). However large, replicated DNA fragments accumulate during and after incubation of mammalian cells a t temperatures a s high as 45°C (Warters and Stone, 1983; Wong et al., 1988). Preliminary observations indicated that the conversion of these clusters of replicons into the cell's chromosomes was more sensitive to hyperthermic temperature exposure than the initiation of new replicons (Warters, 1988). These observations were taken to indicate that the inhibition of the ligation of fully polymerized replicon clusters into the mammalian cell genome was the most heat-sensitive step during DNA replication a t elevated temperatures, and thus was rate limiting for DNA replication a t these temperatures. The results reported here were a n attempt to characterize further the effect of heat shock on the ligation of replicon clusters into the chromosome. 0 1990 WILEY-LISS. INC

MATERIALS AND METHODS

Cell culture and labeling HeLa S3 cells (doubling time ca. 20 hr) were maintained in exponential growth in monolayer culture with McCoy's medium 5A (Grand Island Biological Co., Grand Island, NY? supplemented with 5% fetal bovine and 5% calf serum. For uniform labeling of DNA, monolayer cultures of cells were exposed to whole medium containing (methyl-14C)-thymidine (sp. act. 50 Ciimmol; New England Nuclear, Boston, MA) at 0.05 pCiiml for 20 hr. Pulse labeling of DNA was performed with (meth~l-~H)-thyrnidine (sp. act. 50-60 Ciimmol; New England Nuclear) a t 1.0 kCi/ml.

Cell heating Cells prelabeled in their parental DNA with 14C-thymidine were exposed to fresh, prewarmed, nonradioactive medium at 37°C for 60 min in monolayer culture. The cells were pulse-labeled in replicating DNA by exposure to 3H-thymidine-containing medium. For heating, pulse-labeled cells were exposed to prewarmed medium and placed directly into a prewarmed (41-45°C) precision-controled water bath. After various heating times the flasks were removed from the water bath, and the cell monolayer treated as described in the text.

DNA strand break and size determinations The alkaline filter elution technique used in these studies to estimate single-stranded DNA (ssDNA)size

Received July 24, 1989; accepted October 9, 1989 *To whom reprint requestsicorrespondence should be addressed.

366

WARTERS AND LYONS

was as previously described (Kohn et al., 1974; Warters and Stone, 1983). The size distribution of singlestranded DNA was determined by sedimentation through alkaline sucrose gradients in a manner similar to that previously described (Warters and Brizgys, 1988; Warters et. al., 1988). In some experiments the presence of double-strand breaks in cell DNA was determined by filter elution a t pH 7.2, as previously described (Warters et al., 1988). Double-stranded (ds) DNA length determinations To determine the length of large double-stranded DNA fragments, DNA was analyzed by transverse alternating field gel electrophoresis. Cells were resuspended to lo7 cells per ml in ice-cold lysis buffer (100 mM EDTA, 20 mM NaCI, 10 mM Tris, pH 7.8) and mixed with a n equal volume of low-melting-point agar (1%)in lysis buffer) previously held a t 50°C. This cell suspension was poured into 0.164 ml cylindrical chambers in a plastic mold and placed at 4°C until the agar solidified. These agar cylinders were removed from the mold and treated a t 50°C for two 24 h r periods with 2 washes with 5 vol of lysis buffer containing 1 mgiml proteinase K and 1%SDS. The treated agar cylinders were stored a t 4°C in TE buffer (1 mM EDTA, 10 mM Tris, pH 8.0). The agar cylinders were cut into 0.5 cm3 plugs and embedded in a 0.7 x 7.5 x 10 cm, 1%agar gel. DNA in the plug was electrophoresed for 18 h r a t 10°C through 9 cm of the 1%agar gel in a Beckman Instruments "Geneline" electrophoresis system with a transverse electrical field (constant current of 150 mA) alternating every 60 sec. The gel was removed, stained for 60 min with 0.8 pgiml ethidium bromide, and photographed under UV illumination. Saccharomyces cereuisiae chromosomal DNA (Beckman Instruments) and a ladder of lambda phage DNA (prepared a s described by Bernards et al., 1986) were used a s DNA molecular length standards.

0.0

,

0

1

I

1

1

2 4 6 8 T I M E ( H R I A T TEMPERATURE

Fig. 1. Recovery of DNA synthesis in heated cells. HeLa cells prelabeled with 14C-thymidinewere heated at 435°C for 45 min (triangles) or left unheated (circles).The cells were then placed either a t 37" (open symbols) or a t 41" (closed symbols). At the times indicated the cells were exposed to medium containing 3H-thymidine for 15 min at 37°C. 3H-radioactivity incorporated into acid 115% TCAI-insoluble material was determined. Plotted is the ratio of 3H- t~-'~C-radioactivity in heated relative to unheated cells pulse-labeled a t the times indicated. The error bars in this and subsequent figures represent the avcragc ? 1 standard deviation.

RESULTS When HeLa cells were heated at 43.5"C for 45 min and exposed to 'H-thymidine for 15 min at 37"C, total pulse incorporation into acid-insoluble material was depressed to 20% of that observed in control cells (Fig. 1). Subsequent pulse-incorporation of 'H-thymidine into DNA recovered toward control cell levels with a half-time of approximately 5 h r (Fig. I, open triangles). Pulse-incorporation of 3H-thymidine is usually a n accurate estimate of the level of ongoing replicon initiaIsolation and nuclease digestion of nuclei tion. To verify this the DNA from cells pulse-labeled Nuclei were isolated from HeLa cells and suspended with 'H-thymidine was analyzed by alkaline sucrose in TMN buffer to a concentration of 30-50 pg of DNA/ gradient sedimentation as previously described ml as estimated from the absorbance of 260 nm light of (Warters and Stone, 1984). Incorporation of 'H-thyminuclei placed in 0.2 N KOH (Warters and Roti Roti, dine into replicon-sized DNA immediately after the 1981). The specific activity of the DNA in the nuclei heat shock was observed to be 16% of that observed in was also determined by liquid scintillation counting. control cells. Pulse-incorporation of 'H-thymidine into Equal quantities of control or treated nuclei were replicon-sized DNA was 24, 28, 35, 49 and 75% of that placed in separate tubes. To the nuclei 0.05 to 50 units observed in unheated cells after 1, 3, 4,5, and 8 h r of of micrococcal nuclease (MNase; Worthington Bio- recovery a t 37°C. Thus the recovery of both 'H-thymichemical Corp.) per microgram of DNA was added. The dine pulse-incorporation into whole cell DNA, and of enzyme-nuclei suspension was placed a t 37°C for 60 replicon initiation during recovery from a 43.5"C heat min. The reaction was halted by the addition of ice-cold shock, was simultaneous. While pulse-incorporation of 3H-thymidine into cell TCA to 15%. The TCA-insoluble material was recovered by centrifugation and the total radioactivity con- DNA remained unaffected, relative to control cells, tained in the supernatant (acid soluble) or pellet (acid during continuous incubation at 41°C (Fig. 1, closed insoluble) was determined by scintillation counting. circles), incubation of 43.5"C heated cells at 41°C inDKA fragments produced by the enzyme were ex- hibited both recovery of 'H-thymidine pulse-incorporatracted from nuclei and prepared as previously de- tion (Fig. 1, closed triangles) and replicon initiation scribed (Warters et al., 1980). These DNA fragments (results not shown) to the same extent. Thus while were electrophoresed through a 10 cm composite 3% 41°C incubation alone did not inhibit DNA replication, acrylamide, 0.5% agarose gel at 5 Vicm for 4 hr. DNA incubation of previously heated cells at 41°C (stepmolecular length standards used included a HAE I11 down heating) thoroughly blocked recovery from the digest of plasmid pBR322 DNA and a 123 bp, DNA nuclear alteration(s) responsible for the inhibition of DNA synthesis. ladder (Bethesda Research Laboratories).

367

DNA REPLICATION AT ELEVATED TEMPERATURES

BOTTOM

p r

8

41"

a a

/T

I

I

0

CT _1

+ 4 + 0

LL

0

005

2

+

V

4

cc LL

1

0.00

I

0.1

1

5

I

1

I

I

10

15

20

25

FRACTION NUMBER

0 0-10

I

I

I

1

Fig. 2. Chain elongation of pulse-labeled DNA at hyperthermic temperatures. Cells prelabeled with 14C-thymidine were pulse-labeled with "H-thymidine a t 37°C for 15 min and placed into nonradioactive medium a t 37°C (open circles), 41°C (open squares),42°C (open triangles), 43.5"C (closed circles), or 45°C (closed squares). At the times indicated the cells were recovered and the elutability of radioactivity from polycarbonate filters determined as described in the Materials and Methods. Plotted is the ratio of the fraction of 3H-radioactivity remaining on the filters, divided by the fraction of '*C-radioactivity remaining on the filters, after elution with 26 ml of pH 12.2 buffer.

Fig. 3. Size distribution of replicating DNA during 43.5-C incubation. Cells prelabeled with 14C-thymidinewere pulse-labeled at 37°C for 15 min with 'H-thymidine (open circles) and either placed at 435°C for 120 min (open and closed triangles) or placed a t 37°C in 100 &ml cycloheximide for 120 min (open squares). The cells were recovered by trypsinization and the distribution of "H-labeled DNA determined hy sedimentation through alkaline sucrose gradients. Alternately, cells were lysed on polycarhonate filters and the DNA in the first 1.5 ml of material eluted a t pH 12.2 recovered onto, and sedimented through, alkaline sucrose gradients (closed triangles) as previously described (Warters et al., 1987).Plotted is the fraction of total 3H-radioactivity observed in various gradient fractions. 14C-parental DNA sedimented to the bottom of the gradients. Sedimentation is from left to right (total 1?'t = 8.2 x 10"' rad' sec-').

Alkaline filter elution has been used to quantitate DNA chain elongation (Kohn et al., 19741, since incompletely elongated, single stranded DNA with a size less than 10' Da will elute from a polycarbonate filter a t pH 12.2 (Kohn et al., 1974; Erickson et al., 1979; Warters and Stone, 1984). When HeLa cells were pulse-labeled with 'H-thymidine for 15 min at 37"C, approximately 20% of the incorporated 3H-radioactivity was trapped on the filters (Fig. 2). When these pulse-labeled cells were placed in new 37°C medium, the fraction of incorporated 3H-radioactivity retained on the filter was estimated, a s previously described (Warters and Stone, 1983), to increase with a half-time of 110 min (Fig. 2 , open circles). The rate of DNA chain elongation was the same, or slightly faster, in cells continuously incubated at 41°C (Fig. 2, open squares). The rate of DNA chain elongation was significantly slowed during continuous incubation of pulse-labeled cells a t 42°C (Fig. 2, open triangles), was virtually halted during continuous 435°C incubation (Fig. 2, closed circles) and apparently was reversed somewhat during continuous 45°C incubation (Fig. 2 , closed squares). The relative rate of chain elongation a t 435°C versus 37"C, as measured by the alkaline filter elution assay, was estimated by dividing the total 3H-DNA radioactivity collected onto polycarbonate filters during consecutive 15 min incubation periods at 435°C relative to 37°C. The 'H-DNA radioactivity collected onto the filters in 43.5"C incubated cells was found to be 13, 7, and 6% of

that observed in control, 37°C incubated, cells during the first, second, and third 15 min incubation periods, respectively. When fractional chain elongation at 43.5"C was plotted versus incubation time at 43.5"C, chain elongation was found to be inhibited a t 43.5"C with a reciprocal slope 5 7 min. This apparent inhibition of DNA chain elongation during continuous 435°C incubation was inconsistent with our previous finding that replicating DNA is rapidly elongated into large, replicon-cluster-sized (ca. 120-150 S) DNA during continuous 435°C (Warters, 1988) o r 45°C (Warters and Stone, 1983) incubation. The apparent discrepancy between these two observations was resolved by determining the size of 'Hlabeled DNA in cells during continuous 435°C incubation. DNA pulse-labeled with 'H-thymidine for 15 rnin at 37°C was predominantly of replicon size (Fig. 3, open circles). HeLa nuclear DNA cut by the restriction endonuclease HAE I11 to 50 kilobase pairs, as determined by pulse field gel electrophoresis, sedimented to fractions 5 and 6 in these alkaline sucrose gradients. By 30 min of 37°C incubation the majority (80%) of incorporated 3H-radioactivity was observed in a 120 to 150 S size range, and by 90 t o 120 min the majority (ca. 80%) of 3H-radioactivity was too large for accurate size estimation and sedimented with parental DNA to the bottom of the alkaline sucrose gradients. In contrast, by 120 min of 435°C incubation 80% of the 'H-labeled DNA was the size (120 t o 150 S) of a cluster of replicons

30 60 90 120 T I M E ( m i n 1 AT TEMPERATURE

WARTERS ANT) LYONS

368

slope of approximately 18 min during 435°C incubation. A qualitatively similar collection of replicating DNA into replicon clusters occurs in cells continuously incubated in the presence of the drug cycloheximide (Warters and Stone, 1983). To verify this previous observation HeLa cells were pulse-labeled with 3H-thymidine at 37°C for 15 min and then continuously incubated at 37°C in medium containing 100 pgiml cycloheximide. The fraction of 'H-labeled DNA which became trapped on the polycarbonate filters when analyzed by pH 12.2 filter elution was, respectively, 19, 29, 33, 35, and 38% a t 0.60, 120,180, and 240 rnin of cycloheximide exposure. Whole cell, or pH 12.2elutable, DNA from these cells was predominantly in the 120 to 150 S or replicon cluster size range (Fig. 3, open squares). A similar phenomenon also was observed in cells exposed to the drug novobiocin. When HeLa cells were exposed to 315, 630, or 945 p,M novobiocin for 60 min, pulse-labeled with 3H-thymidine for 15 min, and incubated a t 37°C in medium containing the same drug concentration for 180 min, only 75, 58, lo2 and 45% of pulse-labeled replicon clusters became ligated into the chromosome (i.e., were collected onto Fig. 4. Size-distribution of DNA pulse-labeled a t 43.5%. Cells pre- polycarbonate filters during pH 12.2 elution). These labeled with 14C-thymidinewere left a t 37°C (open circles) or incu- results indicated that treatment of cells with either bated a t 43.5"C for 0 (closed circles), 15 (open squares), or 30 (closed cycloheximide or novobiocin resulted in inhibition of squares) min. Then the cells were pulse-labeled with 3H-thymidine at 37°C (open circles) or at 43.5"C (all other symbols) for 15 min. The replicon cluster ligation into chromosomal DNA. distribution of DNA sizes in the cells was determined by sedimentaThe potential involvement of depressed histone protion through alkaline sucrose gradients as described in Figure 3. tein synthesis or altered chromatin structure in these Fractions 6-10, where peak radioactivity occurs in cells pulsed-laphenomena was determined by digestion of nuclear beled at 37"C, were assumed to contain replicon-sized DNA. chromatin with the enzyme micrococcal nuclease. In pulse-labeled chromatin (Seale, 1975; Warters and Roti Roti, 1981) or chromatin replicated during exposure to (Fig. 3, open triangles). The 'H-labeled DNA persisted cycloheximide (Weintraub, 1976; Warters and Stone, in this size range for up to 240 min incubation a t this 1983) DNA is more sensitive than parental DNA to temperature. When the 'H-DNA which eluted from digestion by exogenous nucleases due to a decrease in polycarbonate filters a t pH 12.2 in cells which had been the frequency of histone proteins (nucleosome particontinuously incubated a t 43.5% for 120 or 240 min cles) associated with DNA replicated under these conwas sedimented through alkaline sucrose gradients, ditions. The nuclease digestibility of DNA pulse100% of this 'H-labeled DNA was in the replicon clus- labeled a t 25°C for 5 min (Fig. 5, closed triangles) was ter size range (Fig. 3, closed triangles). Thus during significantly greater at all enzyme concentrations than continuous 435°C incubation pulse-labeled DNA was of 14C-labeled-parental DNA (Fig. 5, open circles). In rapidly elongated into a replicon cluster size range and contrast, DNA sequences pulse-labeled at 37°C for 15 remained a t this DNA size for the duration of 43.5"C min, (Fig. 5, closed circles) had a nuclease sensitivity incubation. intermediate between 24°C-pulse-labeled and parental Chain initiation (3H-thymidine, pulse-incorporation DNA sequences. Thus the 15 rnin pulse-labeling period into acid-insoluble, replicon-sized DNA) during contin- at 37°C allowed a significant fraction of "-labeled uous 43.5OC incubation was determined (Fig. 4) as pre- DNA to acquire a nearly normal nucleosome freviously described (Warters, 1988). The total, acid-insol- quency. The nuclease sensitivity of DNA pulse-labeled uble 3H-thymidine incorporation into DNA assumed to at 37°C for 15 min and than incubated a t 37°C or 435°C be replicon-sized (i.e., into 40 to 60s or 2 to 4 x l o 7 Da for 120 min in the absence of radioactivity (Fig. 5, open DNA in fractions 6 to 10 of the alkaline gradients in squares and triangles) was identical to that of parental Fig. 4) was determined during a 15 min incubation a t DNA. In contrast the nuclease sensitivity of pulse-la37°C or during the first, second, and third 15 min pe- beled DNA incubated for 120 min a t 37°C in cyclohexriods of a continuous 435°C incubation. When the frac- imide (Fig. 5, closed squares) was similar to the DNA tional pulse incorporation during 43.5"C incubation pulse-labeled at 37°C. In cells incubated at 435°C for ('H-radioactivity incorporated during each consecutive 120 min the distribution of 3H- and '*C-radioactivity 15 min 43.5'32 incubation period divided by the 3H- within nucleosomal DNA electrophoresed through radioactivity incorporated during 15 min a t 37°C) was acrylamide gels was the same, indicating that acquisiplotted versus time a t 43.5"C, chain initiation was ob- tion of a full complement of nucleosome particles and served to be 44, 20 and 9%, respectively, of that ob- mature chromatin structure occurred in replicating served in control cells during the first, second, and DNA during 43.5"C incubation. Since during 43.5OC third 15 min periods of 435°C incubation, indicabing incubation, pulse-labeled DNA acquired a normal nuthat chain initiation was inhibited with a reciprocal cleosome frequency and a mature, interphase chroma1

i

5

F R A C T I O N NUMBER 10 15 20

25

DNA REPLICATION AT ELEVATED TEMPERATURES

-

5

04-

n 0

2

03-

y

02

a z n

m

3

00 7 -I 05 10 50 100 0025 005 01 E N Z Y M E CONCENTRATION I u n i t s / g q DNA1

1 25 0

Fig. 5 . Nuclease sensitivity of pulse-labeled UNA. Cells prelabeled in their parental DNA with 14C-thymidine(open circles) were pulsclabeled in their replicating DNA with 'H-thymidine at 25°C for 5 min (closed triangles) or at 37°C for 15 min (closed circles). Cells pulselabeled at 37°C were then placed at 37°C for 120 min (open squares), placed a t 37°C for 120 min in the presence of 100 p&ml cycloheximide (closed squares) or placed a t 43.5"C for 120 min (open triangles). Nuclei were isolated from the cells, as described in Materials and Methods, and exposed to increasing concentrations of the enzyme micrococcal nuclease (as indicated) a t 37°C for 60 min. Plotted is the fraction of total 'H- or I4C-DNA radioactivity made soluble in 15% ice-cold TCA.

tin structure, we conclude that neither a n inhibition of histone protein synthesis nor a failure of replicating chromatin to be organized into mature chromatin structure can explain the inhibition of replicon cluster ligation into the chromosome during 43.5"C incubation. A significant fraction of those pulse-labeled DNA sequences which fail to be ligated into the chromosome during hyperthermic exposure persist in this unligated state for 24 to 36 h r afterward (Warters and Stone, 1983). When HeLa cells were pulse-labeled at 25°C for 5 min and placed a t 37"C, all the incorporated 3H-radioactivity became trapped on the polycarbonate filters by 24 h r (Table 1).When cells were heated a t 43.5"C for 45 min immediately after being pulse-labeled, 46% of the pulse-labeled DNA eluted from polycarbonate filters (i.e., had failed to be ligated into the cell's genome) 24 h r later (Table 1).The fraction of pulse-labeled DNA which ultimately failed to be ligated into the genome was increased by 50% by three treatments-either by step-down heating at 41"C, or by heating in the presence of either 0.79 mM novobiocin or 15 mM caffeine. No pulse-labeled DNA eluted from polycarbonate filters a t pH 7.2 24 h r after cell heating (results not shown), indicating that all of the "damage" persisting in the cell's genome at this time existed as single, but not double, DNA strand breaks. Since both novobiocin and caffeine inhibit topoisomerase I1 (Warters et al., 1988) we tested the possibility that 43.5"C incubation inhibits this enzyme activity. We made use of the drug 4'-(9-acridinylamino)methanesulfon-m-anisidide(m-AMSA). This DNAintercalating drug has been found to interact with

369

topoisomerase 11, inhibiting its DNA strand passage activity and consequently freezing the enzyme in a n enzyme-associated, DNA double-strand break (a cleavable complex) (Nelson et al., 1984; Minford et al., 1986). Thus at every site in nuclear DNA a t which topoisomerase I1 interacts with this drug a DNA double-strand break is produced, and the activity of the enzyme can be determined by estimating the frequency of DNA-strand breaks (Zwelling et al., 1981; Warters et al., 1988). Topoisomerase 11 activity is saturated (i.e., maximum DNA strand breakage) in this HeLa cell line by exposure to 50-100 pM m-AMSA (Warters et al., 1988). When HeLa cells were exposed to 100 pM m-AMSA and their DNA sedimented through alkaline sucrose gradients, their 14C-labeled,parental DNA was found to be reduced to a size (Fig. 6A, closed squares) slightly larger than that for pulse-labeled replicons (Fig. 6A, open circles). Pulse field gel electrophoresis revealed the dsDNA in these cells to be concentrated at lengths between 100 and 400 kilobase pairs, with 80% of the DNA being less than 1,000 kilobase pairs in length. dsDNA with a length of 50-100 kbp as determined by pulsed field gel electrophoresis exhibited a distribution on alkaline sucrose gradients similar to that of pulse labeled replicons (i.e., sedimented into fractions 5-8).'4C-labeled parental DNA in non-drugtreated cells sedimented to the bottom of the sucrose gradients (Fig. 6A, closed circles) and had a dsDNA length in excess of lo7 base pairs as measured by pulsed field gel electrophoresis. When the cells were pulse-labeled for 15 min at 37"C, placed into fresh medium a t 37°C for 30 min (results not shown) or 120 min (Fig. 6A, open squares), and then exposed to 100 pM m-AMSA for 60 min a t 37"C, the size of the pulselabeled DNA was reduced to that observed for parental DNA in drug-treated cells. When pulse-labeled cells were incubated a t 37°C in the presence of 100 pgiml cycloheximide (Fig. 6B, open squares) or a t 435°C (Fig. 6B, open triangles) for 120 min prior to exposure t o 100 p M m-AMSA under the same conditions, the pulselabeled DNA (open triangles in Fig. 3) was reduced in size (open triangles in Fig. 6B). This size change constitutes a decrease in ssDNA length of from about 7 x lo5 nucleotides to 2.5 x lo5 nucleotides. 14C-labeled parental DNA in cycloheximide-treated cells was reduced by m-AMSA exposure to a similar size (Fig. 6B, closed squares). Parental DNA in the 43.5"C incubated cells (Fig. 6B, closed triangles) was reduced to a significantly larger DNA size during exposure to 100 pm m-AMSA. With pulsed field gel electrophoresis 80% of the parental DNA in 435°C heated cells when exposed to m-AMSA had a length greater than 1,000 kilobase pairs. Thus topoisomerase 11 appeared to be fully active in DNA replicated a t 43.5"C for 120 min. Since parental dsDNA was ordinarily reduced to a n average length of about 250 kbp during exposure to 100 pM m-AMSA, but was reduced to a n average length in excess of 1,000 kbp during incubation at 43.5"C, topoisomerase I1 exhibited only 25%, o r less, of its expected activity in the parental DNA of cells exposed to 43.5% for 120 min.

DISCUSSION Hyperthermic temperature shock of mammalian cells inhibits a number of DNA replicative processes. Previously the inhibition of DNA replication in heated

370

WARTERS AND LYONS

TABLE 1. Alkaline clutability of DNA in previously heated cells Treatment' 1. 37°C for 24 h r 2. 43.5"C, 45 min + 37°C for 24 hr 3. 43.5"C, 45 min 1 41"C, 5 hr + 37°C for 24 hr 4. 43.5"C, 45 min in 0.79 mM novobiocin t 37°C for 24 hr 5. 43.5"C, 45 min in 15 mM caffeine 37°C for 24 hr

Relative fraction of pulse-labeled DNA' 0.0 0.463 I 0 . 0 6 4 0.786 i 0.031

0.712 ? 0.041 0.789 k 0.043

+

'HeLa cells were pulse-labeled with 'H-thymidine a t 25°C for 5 min, healed at 43.5"C for 45 min in the presence, or absence, of drugs as described, incubated in fresh medium for 24 hr a t 37°C and then analysed by alkaline (pH 12.2) filter elution as described in Materials and Methods. The cells which were not heated after being pulse labeled (group 1)are considered control cells in which 100% of the pulse labeled DNA is expectcd to elongate into chromosome-sized DNA. 'The fraction of %labeled DNA which eluted from polycarbonate filters (i.e., had a mass less than 10' Da) during exposure to 26 ml of pH 12.2 elution buffer was determined in all groups. The relative fraction of' elutable 'H-DNA in the treated groups was calculated as the ratio of the fraction of "H-radioactivity eluting in the treated groups divided by the fraction of 3H-radioactivity eluting in the control group (group number 1: usually 104 or less of the total radioactivity). The average 2 1 standard deviation from three repeats is given.

022t'

t 0151 A

I

5

10

I5

20 I 5 F R A C T I O N NUMBER

I0

15

20

25

Fig. 6 . Topoisomerase I1 cleavage of replicating DNA. Cells prelabeled with I4C-thymidine (closed symbols) were pulse-labeled with ,"H-thymidineat 37°C for 15 min (open symbols). A Cells were pulselabeled only (circles) or pulsed-labeled and placed a t 37°C for 120 min (squares). B: Pulse-labeled cells were either placed at 37°C for 120 min in the presence of 100 pg/ml cycloheximide (squares) or placed at 43.5"C for 120 min (triangles), Cells incubated under the various experimental conditions described above were then exposed to 100 pM m-AMSA for 60 min under similar conditions. The size distribution of DNA in the cells was determined by sedimentation through alkaline sucrose gradients as described in Figure 3.

cells was thought to result from direct inhibition of replicon initiation (Wong and Dewey, 1982; Warters and Stone, 1984). While depression of replicon initiation is a n important step in hyperthermic inhibition of DNA replication, the results reported here indicate that this is a secondary consequence of hyperthermic inhibition of the ligation of replicon clusters into the chromosome. The reciprocal slope for inhibition of replicon initiation (Fig. 4) and replicon cluster ligation into the genome (Fig. 2) was found to be approximately 18 min and 7 min, or less, respectively. Thus the ligation of replicon clusters into the genome is a more heat-sensi-

tive process than is replicon initiation, and is more likely to be the rate-limiting step for DNA replication a t elevated temperatures. Since during 435°C incubation previously initiated replicons with a length of 510 x lo4 nucleotides are elongated at a normal, or near normal, rate into replicon clusters with a n average length of 7 x lo5 nucleotides (i.e., are elongated through presumably uninitiated, adjacent replicons within a cluster of replicons) (Warters, 19881, we assume that heat shock does not block replicon initiation. Rather, the failure of replicon clusters to be ligated into the genome blocks, by a feedback mechanism, the initiation of replicon synthesis within adjacent replicon clusters. This series of events ultimately results in a depression in the number of replication forks polymerizing DNA, and thus in the depression in radioactive precursor incorporation most commonly measured. This failure of replicon clusters to be ligated into the genome in heated cells appears to persist in some cases for a t least 24-36 h r (Warters and Stone, 1983; Table 1).Although these DNA single-strand break-type lesions are not converted into DNA double-strand breaks within this time period, they may serve a s the molecular basis for the chromosome aberrations unique to heated S phase cells (Dewey et al., 1978) and thus also may serve as a molecular basis for the extraordinary heat sensitivity of S phase cells (Fox et al., 1985). No macromolecular explanation has been found for this hyperthermic inhibition of DNA replication. If in fact heat shock inhibits DNA replication primarily by blocking the ligation of replicon clusters into the genome, it is reasonable to expect that this inhibition ultimately must involve the failure of a DNA ligase activity to close a gap a t the end of fully polymerized replicon clusters. However, since at elevated temperatures DNA chain elongation, which must require a DNA ligase activity, proceeds at a normal rate within previously initiated replicon clusters (Warters, 1988), the majority of DNA ligase enzymes are fully active and functional in mammalian cells at elevated temperatures. Thus the failure of DNA ligase to close Dh'A gaps a t the end of replicon clusters must either result from the inhibition of a unique DNA ligase enzyme in a unique nuclear location, or be secondary to some other heat-induced nuclear lesion. Hyperthermic temperatures do inhibit protein synthesis (Mondovi et al., 1969; Henle and Leeper, 1982a,b).Since histone protein synthesis also is inhibited in heated cells (Warters and Stonc, 19841, the failure of these proteins to become associated with replicated DNA might have resulted in the inhibition of replicon cluster ligation into the genome. This appears not to be the case, since chromatin replicated a t 43.5"C acquires a full nucleosome complement and becomes fully condensed into normal interphase chromosome structure (Fig. 5). Maturation of replicating DNA containing single-strand gaps into mature interphase chromatin structure may promote the observed longterm persistence of these lesions (Table 1)by obscuring them from DNA replicative or repair processes having the potential subsequently to close them. Exposure of cells to the drug novobiocin sensitizes cells to hyperthermic cytotoxicity (Warters and Brizgys, 1988) and inhibits ligation of replicon clusters into the chromosome. This suggested to u s that this drug

DNA REPLICATION AT ELEVATED TEMPERATURES

might also sensitize DNA replication to heat shock. This, in fact, was observed (Table 11, indicating that inhibition of replicon cluster ligation into the genome a t elevated temperatures may involve topoisomerase I1 inhibition or a n ATP dependent activity. This inhibition does not appear to be due to hyperthermic inhibition of topoisomerase I1 in replicating DNA since this enzyme appears to be fully active and distributed normally along chromatin replicated at 43.5"C (Fig. 6). The finding that I4C-labeled parental DNA was cut less frequently by topoisomerase I1 in 43.5"C incubated cells indicates either that topoisomerase I1 in parental DNA is inactivated more readily a t 43.5"C; that there is a reduced accessibility of DNA in parental chromatin to drug intercalation; or that topoisomerase I1 in the parental chromatin of these cells interacts abnormally with intercalated drug. The failure of topoisomerase I1 to function normally in a n adjacent, a s yet unreplicated, replicon cluster might preclude ligation of fully polymerized replicon clusters into the genome. Our results indicate that a failure of fully polymerized replicon clusters to be ligated into chromosomesized DNA is the rate-limiting step for the continuation of DNA replication a t 435°C. Since ongoing chain elongation results in replicon cluster-sized DNA, the majority of ligase enzymes within any replicon cluster must remain active during 43.5% incubation. The failure at 435°C of these replicon clusters t o be ligated into chromosome-sized DNA must be due to a direct thermal effect (e.g., inhibition or inactivation) on a minority of DNA ligase enzymes, or on some ligase-dependent enzyme or structure, located a t the junction between replicon clusters. Our results do not provide a molecular explanation for thermal inhibition of this DNA replicative step. However, they do indicate that thermal effects on neither the majority of replicative enyzmes, nucleosome formation, nor the processing of replicating DNA into mature, interphase chromosome structure is involved in this inhibition. It seems more likely that thermal damage within chromatin adjacent to replicating clusters of replicons inhibits completion of their ligation into the genome. One nuclear change which may participate in this hyperthermic effect is the failure of topoisomerase I1 to function in chromatin adjacent to fully polymerized replicon clusters.

ACKNOWLEDGMENTS The authors thank Dr. Louis R. Barrows for critical review of the manuscript and Martha S. Knowlton and Janet Holman for assistance with the preparation of the manuscript. This work was supported by grant CA 25957 from the National Cancer Institute, DHHS. LITERATURE CITED Bernards, A,, Kooter, J.M., Michels, P.A.M., Moberts, R.M.P., and Borst, P. (1986) Pulsed-field gradient electrophoresis of UNA digested in agarose allows the sizing ofthe large duplication unit of a surface antigen gene in trypanosomes. Gene, 42'2:313-322. Dewey, W.C., Sapareto, S.A., and Betten, D.A. (1978) Hyperthermic

371

radiosensitization of synchronous Chinese hamster cells: Relationship between lethality and chromosome aberrations. Radiat. Res., 76:48-59. Erickson, L.C., Ross, W.E., and Kohn, K.W. (1979) Isolation and purification of large quantities of DNA replication intermediates by pH step alkaline elution. Chromosoma, 74t12.5-139. Pox, M.H., Read, R.A.. arid Bedford. J.S. (1986) The cell cycle dependence of thcrmotolerancc. 111. HeLa cells heatcd at 45°C. Radiat. Res., 104t429-442. and Cancer, Plenum Press, New Hahn. G.N. (1982) Hyperthermia _. York, 1982. Henle. K.J.?and Leeper, D.K. (1982a) Combinations of hyperthermia (40°,45'Ci with radiation. Radiology, 121:451-454. Henle, K.J., and Lccper, D.B. (198213) The effects of hyperthermia (45") on macromolecular synthesis in Chinese hamster ovary cells. Radiat. Res., 89t166-175. Kohn, K.W., Friedman, C.A., Ewig, R.A.G., and Iqbal, Z.M. 11974) DNA chain growth during replication of asynchronous I,] 210 cells. Alkaline elution of large DNA sections from cells lysed on filters. Biochemistry, 13t4134- 4139. Minford, J., Pommier, Y., Pilipski, H., Kohn, K.W., Kerrigan, D., Mattern, M., Michaels, S., Schwartz, R.: and Zwelling, L.A. (1986) Isolation o f intcrcalator-dependent protein-linked DNA strand cleavage activity from cell nuclei and identification as topoisomerase 11. Biochemistry. 25t9-16. Mondovi, B., Finazzi Agro, A,, Rutilio, G., Strom, R. Moricca, G., and Rossi Fanelli, A. (1969) The biochemical mechanism of selective heat sensitivity of cancer cells. 11. Studies on nucleic acids and protein synthesis. Eur. J. Cancer Clin. Oncol., 5:137-146. Nelson, E.M., Tewey, K.M., and Liu, L.F. 11984) Mechanism of antitumor drug action: Poisoning nf mammalian DNA topoisomerase I1 on DNA by 4'-(9-acridinylamino1-methanesulfon-n-anisidide. Proc. Natl. Acad. Sci. U.S.A., 81t1361-1365. Roti Roti, J.L., and Laszlo, A. (1988) The effect of hyperthermia on cellular macromolecules. In: Hyperthermia and Oncology, Vol. I. M. Urano, ed. VST Scientific Press, Zeist, The Netherlands. Seale, R.L. (1975)Assembly of DNA and protein during replication in HeLa cells. Nature, 255t247-249. Warters, R.L. 11988) Hypcrthermia blocks DNA processing a t the nuclear matrix. Radiat. Res., 115:258-272. Warters, R.L., and Roti Roti, J.L. (1981)The effect of hyperthermia on replicating chromatin. Radiat. Res., XBt69-78. Warters, R.L., and Stone, O.L. (19831Effects ofhyperthermia on DNA replication in HeLa cells. Radiat. Res., 93:71-84. Warters, R.L., and Stone. O.L. (1984) Histone protein and DNA synthesis in HeLa cells after thermal shock. J . Cell. Physiol., 118: 153-160. Warters, R.L., and Brizgys. L.M. (1988) Effect of topoisomerase I1 inhibitors on hyperthermic cytotoxicity. Cancer Res., 48: 3932-3938. Warters. R.L., Roti Roti, cJ.L.,and Winward, R.T. (1980) Nucleosome structure in chromatin from heated cells. l-ladiat. Res., 84t504-513. Warters, R.L., Lyons, B.W., Chiu, S.M., and Oleinick, N.L. (1987) Induction of DNA strand breaks in transcriptionally active DNA sequences of mouse cells by low doses of ionizing radiation. Mutat. Res., 180:21-29. Warters, R.T,., Lyons, B.W., Kennedy. K., and Li. T.M. (1988) Toooisomerase activity in irradiated mammalian celis. Mutat. Res., 216: 43-55. Weintraub, H. 11976)Cooperative alignment of nu bodies during chromosome replication in the presence of cycloheximide. Cell, 9t419422. Wong, R.S.L., and Dewey, W.C. (1982) Molecular studies on the hyperthermic inhibition of DNA synthesis in Chinese hamster ovary cells. Radiat. Res., 92:370-396. Wong, R.S.L..Thompson, L.L., and Dewey, W.C. (1988) Kecoveryfrom effects of heat on DNA synthesis in Chinese hamster ovary cells. Radiat. Res., 114t125-737. Zwelling, L.A., Michaels, S., Erickson, L.C., Ungerleider, R.S., Nichols, M., and Kohn, K.W. (1981) Protein-associated deoxyribonucleic acid strand breaks in L1210 cells treated with the deoxyribonucleic acid intercalating agents 4'(9-acridinylamino)-methanesulfon-m-anisidide and adriamycin. Biochemistry, 20.5553 -6563.