CurrentGenetics
Current Genetics 2, 175-180 (1980)
© by Springer-Verlag 1980
The Effect of Hydroxyurea on the Mechanism of DNA Synthesis in the Yeast Saccharomyces cerevisiae Leland H. Johnston Microbiology Division, National Institute for Medical Research, Mill Hill, London, NW7 1AA, England
Summary. Newly synthesised DNA molecules the same size as replicons (7 m i l l i o n - 6 0 million daltons) accumulate in yeast cells treated with hydroxyurea. During prolonged incubation in low concentrations o f the drug, there is a large accumulation o f these molecules without any corresponding increase in their molecular weight. On release from the inhibtion the molecules are converted to large molecular weight DNA. These observations are consistent with an inhibition b y hydroxyurea o f the joining of completed replicons. In addition, newly synthesised DNA molecules the size o f yeast Okazaki fragments also accumulate in cells treated with hydroxyurea.
fect of HU on the in vivo mechanism of DNA synthesis in this organism.
Key words: Yeast - Hydroxyurea - DNA synthesis
HU Treatment and Radioactive Labelling o f Cells for Sedimentation. Strain D273 was grown in YEPD containing 0.07 /zCi (2.6 kBq)/mi [2-14C] uracil (Amersharn; 62 mCi (2.3 GBq)/mmol) for approximately 20 h to 107 celis/ml (mid-log phase). A 2 M solution of HU was added to 8 ml of culture to give the required final I-IUconcentration and incubation was continued for 30 min. The cells were then centrifuged and resuspended in 1 ml of fresh YEPD containing the same concentration of HU. The required amount of [8-3H] adenine (Amersharn; 22 Ci (0.8 TBq)/mmol) was added and the cells were labelled for the desired period. Labelling was terminated by mixing with an equal volume of stop-mix (Johnston and Williamson 1978) and the cells were placed on ice.
Introduction The antitumour agent h y d r o x y u r e a (HU) has been found to inhibit DNA synthesis in a number of organisms, including bakers' yeast, Saccharomyces cerevisiae (for a review see Timson 1975). In yeast, indeed in most organisms, the action o f HU is specifically on DNA synthesis, there being little or no effect on RNA or protein synthesis (Slater 1973; HartweU 1976). It is generally agreed that this inhibition o f DNA synthesis is related to the supply o f deoxyribonucleotide precursors, possibly b y affecting the enzyme ribonucleotide reductase (Elford 1968; Sinha and Snustad 1972). However, the effects of the drug on t h e i n vivo mechanism of DNA synthesis have never been examined. We have previously used pulse-labelling o f cells followed b y sedimentation of the DNA in alkaline sucrose gradients to characterise in vivo DNA synthesis in yeast (Johnston and Williamson 1978). Here, the same technique has been used to study the ef-
Materials and Methods Strains and Culture Conditions. The petite strain of D273-11a adel his1 trp2 has been described previously (Johnston and Williamson 1978). In this publication this strain is simply referred to as D273. Liquid cultures were incubated with shaking in YEPD medium (Johnston and Game 1978) and a temperature of 25 °C was used for all experiments.
Sucrose Gradient Sedimentation. These procedures have already been described in detail (Johnston and Williamson 1978). Essentially, the stop-mix treated cells were washed and treated with zymolase. The resultant spheroplasts were lysed in alkaline sarkosyl and the lysate was heated to denature the DNA before loading onto the gradients. The previously published procedure was modified by increasing the concentration of dithiothreitol to 50 mM in the buffer used during the zymolase treatment. This resulted in the recovery of considerably larger 14C-prelabelled DNA from the gradients. The lysate was layered on a 5%-20% alkaline sucrose gradient, with a 0.5 ml cushion of 80% sucrose. The gradients were centrifuged in a Beckman SW41 rotor at 19,000 rpm at 20 °C for 16 h. Fractions were colieeted by suction from the bottom of the gradient, dripped into tubes, made 0.5 N with NaOH and, 0172-8083/80/0002/0175/$01.20
176
L.H. Johnston: Effect of Hydroxyurea on DNA Synthesis in Yeast tions arising from incorporation into this DNA, a petite strain was used from which the mitochondrial DNA had been eliminated. We have previously shown that, in respect to single-stranded intermediates in DNA synthesis, such strains behave in the same way as do respiratory competent strains (Johnston and Williamson 1978). Furthermore they respond to HU in a similar fashion to grandes (see below).
~4 e-
> .m
The Effect of HU on Gross DNA Synthesis
I r,,
E O. u
-v2 ¢o
HU
1
2
3
Hours
Fig. 1. Effect of HU on gross DNA synthesis in D273.6/~Ci/ml [8-3H] adenine were added to mid-log phase cells. After 1 h the first sample was taken. One hour later the culture was divided into four sub-cultures, appropriate amounts of HU were added to three of them and sampling was continued. All samples were placed in an alkaline stop-solution and after incubation at 37 °C for 24 h, the alkali-stable, acid-precipitable radioactivity was determined (Johnston and Game 1978). The results are expressed as the increase in incorporation relative to the first sample. (i ~) control; (A~----=)0.025 M HU; (~ ~) 0.05 M HU; (o--o) 0.1MHU
after incubation at 37 °C for 24 h, the alkali-stable, acid-precipitable radioactivity was determined. Molecular weights were determined using the equation of Levin and Hutchinson (1973) with a coefficient a of 0.36 (Jolmston and Williamson 1978). Phage X DNA was used as the marker. Results
General Experimental Considerations The DNA of the cells was labelled with low levels of 14C for some 20 h before sedimentation in alkaline sucrose gradients. This prelabel was used as an indication of the size of the chromosomal DNA and also to monitor the recovery of the DNA from the gradients. The shorter periods of labelling with 3H used in this study are those referred to as 'pulse-labelling'. In S. cerevisiae the mitochor~drial DNA comprises some 15% of the total cellular DNA. To avoid complica-
There was a very rapid inhibition of DNA synthesis with concentrations of HU as low as 0.025 M (Fig. 1). However, at this concentration, some synthesis occurred and even at 0.05 M, an escape from the inhibition was clearly evident after 45 min incubation in the drug. Using the higher concentration of 0.1 M HU, there was an unexplained initial loss of incorporated label, then no incorporation for some 45 min followed by a gradual escape from inhibition. A similar escape from inhibition has been observed previously (Hartwell 1976). Furthermore, in agreement with the observations of Slater (1973), there was little effect of these concentrations o f HU on protein synthesis (results not shown).
Effect of Various Concentrations of HU on the Pattern of Pulse-labelled DNA Observed in Alkaline Sucrose Gradients In Fig. 2 is shown a series of alkaline sucrose gradients prepared from cells pulse-labelled in the presence of concentrations of HU between 0.025 M and 0.125 M. The most striking result, when compared with the control (Fig. 2A), was the failure to synthesise large-sized DNA in the presence of HU. Even at 0.025 M HU (Fig. 2B) there was an accumulation of pulse-labelled molecules between fractions 11 and 22 with a peak at fraction 18. There was also a slight accumulation of small molecules of DNA at the top of the gradient. The result with 0.05 M HU (Fig. 2C) was remarkably similar. There was again a broad peak of pulse4abel in the middle of the gradient with most label in fraction 18 but at this HU concentration there was a more pronounced accumulation of pulse-label appearing at the top of the gradient. With higher concentrations of HU this trend of an increased proportion of the pulse-label appearing at the top of the gradient continued until with 0.125 M (Fig. 2F) the majority (67.3%) was recovered in the top five fractions. Indeed, this latter gradient resembled that of a yeast DNA ligase mutant pulse4abelled at the restrictive temperature where the joining of Okazaki fragments is defective (Johnston and Nasmyth 1978).
L. H. Johnston: Effect of Hydroxyurea on DNA Synthesis in Yeast
A.O
177
B. 0 ' 0 2 5 M
20
15
15
10
10
c. 0- 05M
E
Q. U
1.J O
10
,
20
30
10
20
10
30
20
30
O E. 0.1M
D. 0-075 M
F, 0,125 M
¢) Q..
15
10
10
5
10
20
30
10
20
30
10
20
30
F r a c t i o n number
Fig. 2. A-F. Effect of HU on the pattern of pulse-labelled DNA in alkaline sucrose gradients. A culture grown in medium containing 14C-uracil to 107eells/ml was divided into six parts and exposed to HU at the concentrations shown for 30 rain, Each was then pulselabelled for 40 rain, in the presence of the same concentration of HU, with 3H-adenine at: (A) and (B) 40 #Ci/ml, (C) 50 #Ci/ml, (D) 60 /~Ci/ml, (E) 80 t~Ci/ml, (F) 100 #Ci/ml. The DNA was extracted and sedimented on alkaline sucrose gradients as described in Methods. Sedimentation is from right to left. The total 14C-labelrecovered per gradient varied from 648 to 736 cpm, while the 3H-label recovered was: (A) 34,007 cpm, (B) 2,968 cpm, (C) 3,529 cpm, (D) 2,175 cpm, (E) 1,938 cpm, (F) 2,268 cpm. (~-v) 14C-label; (~ ~) 3H-label
In Fig. 2 incorporation in the presence of HU is low, only some 2%-8% o f the wild-type. It is, therefore, possible that DNA synthesis in only a small sub-population of the cells was being observed, rather than a general effect on the mechanism of DNA synthesis. This is unlikely for two reasons. Firstly, the observed DNA synthesis was abnormal (Fig. 2), so we are not simply dealing with a small population of cells resistant to the effects of HU. Secondly, when cells were preincubated in 0.05 M HU for 1.5 h, to allow 'escape' from the gross inhibition by HU (Fig.l), and then pulse-labelled, a gradient essentially identical to that in Fig. 2C was obtained (result not shown). In this case, however, the incorporation was
40% of the wild-type, of fivefold more than that in Fig. 2C. Thus, it seems likely that it is the residual synthesis in a substantial proportion of the cells which is being examined rather than a large amount of synthesis in a small proportion of the cells. Untreated cells can be pulsed for a shorter time, some 4 - 8 min, to give a level of incorporation similar to that in Fig. 2 B - F . This produces a tritiated profile on gradients (Johnston and Williamson 1978)resembling that of cells treated with 0.05 M HU, suggesting the effect of HU may merely be to slow down the rate of replication. This interpretation is ruled out by a series of pulses at 0.05 M HU and between 30 and 120 min duration. Desome
178
L. H. Johnston: Effect of Hydroxyurea on DNA Synthesis in Yeast
-
!
o
10
20
~
10
~
~
F. 4 0 rain
E. 2 0 rain
XlO
10
10
20
~
20
30
10
~
G, 60rain
1C
E u
51
10
2O
10
30
F r action
number
Fig. 3. A-H. Effect on DNA synthesis of prolonged incubation in 0.05 M HU. Ceils grown in the presence of 14C-uracil to 107/ml were exposed to 0.05 M HU for 30 rain and then pulse-labelled in the same concentration of HU with 50 ~Ci/ml 3H-adenine for (A, E) 20 rain, (B, F) 40 rain, (C, G) 60 rain, (D, H) 80 rain. The DNA was extracted and sedimented on alkaline sucrose gradients as described in Methods. Sedimentation is from right to left. In A - D the epm in each fraction are plotted as a percentage of the total cpm, as in Fig. 2, while in E-H the pulse-label from the same results is plotted as absolute cpm. The total 14C-label per gradient varied between 844 cpm and 861 cpm and the total 3H-label was (A) 1,776 cpm, (B) 5,731 cpm, ((2) 11,855 epm, (D) 20,273 cpm (o---e) 14C-label, (e---e) 3H-label
spite a greater than tenfold increase in incorporation during this time, all the piilses give essentially identical tritiated profiles on gradients (Fig. 3 A - D and unpublished observations). In Figs. 2 C - F minor peaks of pulse-label are present in fractions 9 and/or 11. These are a previously undescribed species of newly synthesised DNA and may represent the joining together of adjacent, completed replicons within a small cluster (manuscript in preparation). When an experiment similar to that in Fig. 2 was done, using a strain containing mitochondrial DNA, essentially identical results were obtained (results not shown). Therefore the results described above are not an anomaly arising from the use of a strain without mitochondfial DNA.
of label in this peak without any corresponding increase in molecular weight. This was found to be the case. Cells were pulse4abelled for periods of 2 0 - 8 0 min in 0.05 M HU and, despite an 11 A-fold increase in incorporation in this time, there was very little change in the relative distribution of pulse-label throughout the gradients (Fig. 3 A - D ; for a control see Fig. 2A). During the labelling period, however, there was a large increase in the absolute amount of pulse-label between fractions 22 and 11 and also in the top five fractions of the gradient (Fig. 3 E - H ) . During this period some high molecular weight DNA was also being synthesised (Fig. 3 E - H , bottom five fractions) in spite of the lack of any change in the molecular weight distribution o f the DNA between fractions 1 1 - 2 2 (Fig. 3 A - D ) . This further supports the idea that these molecules are completed replication intermediates and also indicates that some residual joining of them was occurring.
Effect of Prolonged Incubation in HU In Fig. 2B and 2C a reproducible 'peak' of pulse-label was present between fractions 22 and 11 with most label in fraction 18. If this peak represented an accumulation of completed replication intermediates, an increase in the period of pulse-labelling ought to'increase the amount
Release from HU Inhibition When cells pulse-labelled in the presence of HU were removed from the drug and incubated in medium free of HU, the pulse-labelled DNA was converted to the size of
L. H. Johnston: Effect of Hydroxyurea on DNA Synthesis in Yeast
179 radioactivity from the five fractions around the peak in fraction 18 (16-20). After the labelling in HU there were 4,407 cpm in these fractions (Fig. 4A) while after the incubation free of HU this figure was 2,095 cpm (Fig. 4B). Similarly, considering the top six fractions of the gradient, there was a decrease from 4,152 cpm to 2,890 cpm.
20
Discussion
E D_ U
10
20
30
10
20
30
c
u 2C c]_
15
10-
5
Fraction number
Fig. 4. A and B. Release from HU inhibition. Cells grown in the presence of 14C-uracil to 107/ml were exposed to 0.05 M HU for 30 rain and then pulseqabelled in the HU for 40 rain with 100 ~Ci/ml 3H-adenine. One half of the culture (A) was then placed in stop-mix, while the other half (B) was filtered, washed with YEPD, resuspended in 5 ml fresh YEPD without HU or label but containing 50 #g/ml unlabelled adenine and incubated a further 80 rain. The DNA was then extracted and sedirnented as described in Methods. Sedimentation is from right to left. The total 14C-labelwas 993 cpm and 1076 cpm, while the 3H-label was 15,968 cpm and 53,074 cpm respectively. (c--o) 14C-label, (~ ~) 3H-label
the prelabeUed DNA (Fig. 4). In this experiment the cells were simultaneously removed from labelled medium and the further incubation was in a medium containing an excess of unlabelled adenine. In yeast, however, it is not possible rapidly to terminate ('chase')pulse-labelling of DNA by this means so that there was a 3.3-fold increase in incorporation during the incubation in drugfree medium. Nevertheless there was an absolute loss of
The observed effects of HU on the mechanism of DNA synthesis are consistent with a partial inhibition of the joining of completed replicons. This hypothesis is supported by several lines of evidence. Firstly, the size distribution of the pulse-labelled molecules between fractions 22 and 11, corresponding to molecular weights of 7 - 6 0 million daltons, resembles exactly the size range of yeast replicons (Petes and Williamson 1975). Secondly, on increasing the pulse length in HU there was a dramatic increase in the amount of pulse-label in the replicons without any corresponding increase in their molecular weight (Fig. 3). This suggests that in the presence of HU completed replicons accumulate. Finally, on release from HU inhibition, there was an absolute loss of pulselabel from the peak of replicons, concomitant with the appearance of label in high molecular weight DNA (Fig. 4). The fact that the loss of radioactivity was less than complete reflects the known difficulty in pulse-chase experiments with yeast arising from the presence of high internal precursor pools. As well as the replicons, similar results were apparent with the population of small pulse-labelled molecules appearing at the top of the gradients (Figs. 2 - 4 ) . These approximately correspond in size to Okazaki fragments in yeast (Johnston and Nasmyth 1978) and it is, therefore, tempting to suppose that HU also inhibits the joining of these molecules. However, it is not certain that Okazaki fragments are real replication intermediates (Tye et al. 1977) and a number of artefacts are known to mimic the appearance of these fragments (for example see Anderson 1978). In the absence of clean pulsechase experiments it is, therefore, difficult to draw firm conclusions concerning the real nature of these molecules and discussion of the effects of HU on them are probably premature. The mechanism of inhibition of joining of replicons by HU is not clear. It could be affecting DNA ligase; however this would not explain the general inhibition of DNA synthesis, since a conditional yeast ligase mutant undergoes a complete round of DNA synthesis at the restrictive temperature (Culotti and Hartwell 1971; Johnston and Nasmyth 1978). Alternatively the drug may inhibit a DNA polymerising activity responsible for filling gaps left between adjacent completed replicons.
180 Since HU is known to inhibit ribonucleotide reductase (Elford 1968; Sinha and Snustad 1972), an effect on DNA polymerases could occur at the level of precursor supply, as suggested for the effect of HU on DNA repair (Collins 1977; Burg et al. 1977; Collins et al. 1977). This interpretation would provide sufficient flexibility to account for the experimental observations, assuming that there are DNA polymerising activities in the cell with different KmS. Lowered precursor pools would account for the general depression in DNA synthesis (Fig. 1). While the accumulation of replicons would occur if the synthesis of intermediates preceding replicons (in effect, gross DNA synthesis) was carried out by a polymerase with a low Km able to draw more efficiently on decreased deoxyribonucleotide pools than the polymerase responsible for filling gaps between replicons. In fact in vitro the yeast DNA polymerases have somewhat similar Kms (Wintersburger 1974; Chang 1977). However in association with other proteins their Kras may be altered, as has been observed for the interaction between DNA binding protein and polymerase in Ustilago (Yarranton et al. 1976). In the light of these results it is instructive to reconsider our previous studies on the mechanism of DNA synthesis in yeast (Johnston and Williamson 1978). There is a striking coincidence in the size distribution between the completed replicons accumulating in cells treated with HU (Figs. 2B-C and 3) and the DNA molecules sedimenting in fractions 22-11 after pulseqabeUing of (untreated) log-phase cells for 4 - 8 rain (Fig. 2 in Johnston and Williamson 1978). This suggests that the molecules evident in the untreated cells are completed entities rather than growing intermediates. This implies that, in yeast mitotic DNA synthesis, there must be a lag between the completion of a replicon and its joining to
L.H. Johnston: Effect of Hydroxyurea on DNA Synthesis in Yeast adjacent completed replicons. We were unable to detect such a lag using synchronous cells (Johnston and Williamson 1978) but possibly it is very brief and the degree of synchrony may have been insufficient to detect it.
Acknowledgements. I thank Don Williamson for critical reading of the manuscript.
References Anderson MLM (1978) Nucleic Acids Res 5:4343-4354 Burg K, Collins ARS, Johnson RL (1977) J Cell Sci 28:29-48 Chang LMS (1977) J Biol Chem 252:1873-1880 Collins ARS (1977) Biochim Biophys Acta 478:461-473 Collins ARS, Schor SL, Johnson RL (1977) Mutat Res 42:413432 Culotti J, Hartwell LH (1971) Exp Cell Res 67:389-901 Elford HL (1968) Biochem Biophys Res Commun 33:129-135 HartweU LH (1976) J Mol Bio1104:803-817 Johnston LH, Game JC (1978) Mol Gen Genet 61:205-214 Johnston LH, Nasmyth KA (1978) Nature 274:891-893 Johnston LH, Williamson DH (1978) Mol Gen Genet 164:217225 Levin D, Hutchinson F (1973) J Mot Bio175:495-502 Petes TD, Williamson DH (1975) Exp Cell Res 95:103-110 Sinha NK, Snustad DP (1972) J Bacteriol 112:1321-1334 Slater ML (1973) J Bacteriol 113:263-270 Timson J (1975) Murat Res 32:115-132 Tye B, Nyman P, Lehman IR, Hochhauser S, Weiss B (1977) Proc Natl Acad Sci USA 74:154-157 Wintershurger E (1974) Eur J Biochem 50:41-47 Yarranton GT, Moore PD, Spanos A (1976) Mol Gen Genet 145: 215-218
Communicated by B. S. Cox Received June 20, 1980
Current Genetics 2, 175-180 (1980)
© by Springer-Verlag 1980
The Effect of Hydroxyurea on the Mechanism of DNA Synthesis in the Yeast Saccharomyces cerevisiae Leland H. Johnston Microbiology Division, National Institute for Medical Research, Mill Hill, London, NW7 1AA, England
Summary. Newly synthesised DNA molecules the same size as replicons (7 m i l l i o n - 6 0 million daltons) accumulate in yeast cells treated with hydroxyurea. During prolonged incubation in low concentrations o f the drug, there is a large accumulation o f these molecules without any corresponding increase in their molecular weight. On release from the inhibtion the molecules are converted to large molecular weight DNA. These observations are consistent with an inhibition b y hydroxyurea o f the joining of completed replicons. In addition, newly synthesised DNA molecules the size o f yeast Okazaki fragments also accumulate in cells treated with hydroxyurea.
fect of HU on the in vivo mechanism of DNA synthesis in this organism.
Key words: Yeast - Hydroxyurea - DNA synthesis
HU Treatment and Radioactive Labelling o f Cells for Sedimentation. Strain D273 was grown in YEPD containing 0.07 /zCi (2.6 kBq)/mi [2-14C] uracil (Amersharn; 62 mCi (2.3 GBq)/mmol) for approximately 20 h to 107 celis/ml (mid-log phase). A 2 M solution of HU was added to 8 ml of culture to give the required final I-IUconcentration and incubation was continued for 30 min. The cells were then centrifuged and resuspended in 1 ml of fresh YEPD containing the same concentration of HU. The required amount of [8-3H] adenine (Amersharn; 22 Ci (0.8 TBq)/mmol) was added and the cells were labelled for the desired period. Labelling was terminated by mixing with an equal volume of stop-mix (Johnston and Williamson 1978) and the cells were placed on ice.
Introduction The antitumour agent h y d r o x y u r e a (HU) has been found to inhibit DNA synthesis in a number of organisms, including bakers' yeast, Saccharomyces cerevisiae (for a review see Timson 1975). In yeast, indeed in most organisms, the action o f HU is specifically on DNA synthesis, there being little or no effect on RNA or protein synthesis (Slater 1973; HartweU 1976). It is generally agreed that this inhibition o f DNA synthesis is related to the supply o f deoxyribonucleotide precursors, possibly b y affecting the enzyme ribonucleotide reductase (Elford 1968; Sinha and Snustad 1972). However, the effects of the drug on t h e i n vivo mechanism of DNA synthesis have never been examined. We have previously used pulse-labelling o f cells followed b y sedimentation of the DNA in alkaline sucrose gradients to characterise in vivo DNA synthesis in yeast (Johnston and Williamson 1978). Here, the same technique has been used to study the ef-
Materials and Methods Strains and Culture Conditions. The petite strain of D273-11a adel his1 trp2 has been described previously (Johnston and Williamson 1978). In this publication this strain is simply referred to as D273. Liquid cultures were incubated with shaking in YEPD medium (Johnston and Game 1978) and a temperature of 25 °C was used for all experiments.
Sucrose Gradient Sedimentation. These procedures have already been described in detail (Johnston and Williamson 1978). Essentially, the stop-mix treated cells were washed and treated with zymolase. The resultant spheroplasts were lysed in alkaline sarkosyl and the lysate was heated to denature the DNA before loading onto the gradients. The previously published procedure was modified by increasing the concentration of dithiothreitol to 50 mM in the buffer used during the zymolase treatment. This resulted in the recovery of considerably larger 14C-prelabelled DNA from the gradients. The lysate was layered on a 5%-20% alkaline sucrose gradient, with a 0.5 ml cushion of 80% sucrose. The gradients were centrifuged in a Beckman SW41 rotor at 19,000 rpm at 20 °C for 16 h. Fractions were colieeted by suction from the bottom of the gradient, dripped into tubes, made 0.5 N with NaOH and, 0172-8083/80/0002/0175/$01.20
176
L.H. Johnston: Effect of Hydroxyurea on DNA Synthesis in Yeast tions arising from incorporation into this DNA, a petite strain was used from which the mitochondrial DNA had been eliminated. We have previously shown that, in respect to single-stranded intermediates in DNA synthesis, such strains behave in the same way as do respiratory competent strains (Johnston and Williamson 1978). Furthermore they respond to HU in a similar fashion to grandes (see below).
~4 e-
> .m
The Effect of HU on Gross DNA Synthesis
I r,,
E O. u
-v2 ¢o
HU
1
2
3
Hours
Fig. 1. Effect of HU on gross DNA synthesis in D273.6/~Ci/ml [8-3H] adenine were added to mid-log phase cells. After 1 h the first sample was taken. One hour later the culture was divided into four sub-cultures, appropriate amounts of HU were added to three of them and sampling was continued. All samples were placed in an alkaline stop-solution and after incubation at 37 °C for 24 h, the alkali-stable, acid-precipitable radioactivity was determined (Johnston and Game 1978). The results are expressed as the increase in incorporation relative to the first sample. (i ~) control; (A~----=)0.025 M HU; (~ ~) 0.05 M HU; (o--o) 0.1MHU
after incubation at 37 °C for 24 h, the alkali-stable, acid-precipitable radioactivity was determined. Molecular weights were determined using the equation of Levin and Hutchinson (1973) with a coefficient a of 0.36 (Jolmston and Williamson 1978). Phage X DNA was used as the marker. Results
General Experimental Considerations The DNA of the cells was labelled with low levels of 14C for some 20 h before sedimentation in alkaline sucrose gradients. This prelabel was used as an indication of the size of the chromosomal DNA and also to monitor the recovery of the DNA from the gradients. The shorter periods of labelling with 3H used in this study are those referred to as 'pulse-labelling'. In S. cerevisiae the mitochor~drial DNA comprises some 15% of the total cellular DNA. To avoid complica-
There was a very rapid inhibition of DNA synthesis with concentrations of HU as low as 0.025 M (Fig. 1). However, at this concentration, some synthesis occurred and even at 0.05 M, an escape from the inhibition was clearly evident after 45 min incubation in the drug. Using the higher concentration of 0.1 M HU, there was an unexplained initial loss of incorporated label, then no incorporation for some 45 min followed by a gradual escape from inhibition. A similar escape from inhibition has been observed previously (Hartwell 1976). Furthermore, in agreement with the observations of Slater (1973), there was little effect of these concentrations o f HU on protein synthesis (results not shown).
Effect of Various Concentrations of HU on the Pattern of Pulse-labelled DNA Observed in Alkaline Sucrose Gradients In Fig. 2 is shown a series of alkaline sucrose gradients prepared from cells pulse-labelled in the presence of concentrations of HU between 0.025 M and 0.125 M. The most striking result, when compared with the control (Fig. 2A), was the failure to synthesise large-sized DNA in the presence of HU. Even at 0.025 M HU (Fig. 2B) there was an accumulation of pulse-labelled molecules between fractions 11 and 22 with a peak at fraction 18. There was also a slight accumulation of small molecules of DNA at the top of the gradient. The result with 0.05 M HU (Fig. 2C) was remarkably similar. There was again a broad peak of pulse4abel in the middle of the gradient with most label in fraction 18 but at this HU concentration there was a more pronounced accumulation of pulse-label appearing at the top of the gradient. With higher concentrations of HU this trend of an increased proportion of the pulse-label appearing at the top of the gradient continued until with 0.125 M (Fig. 2F) the majority (67.3%) was recovered in the top five fractions. Indeed, this latter gradient resembled that of a yeast DNA ligase mutant pulse4abelled at the restrictive temperature where the joining of Okazaki fragments is defective (Johnston and Nasmyth 1978).
L. H. Johnston: Effect of Hydroxyurea on DNA Synthesis in Yeast
A.O
177
B. 0 ' 0 2 5 M
20
15
15
10
10
c. 0- 05M
E
Q. U
1.J O
10
,
20
30
10
20
10
30
20
30
O E. 0.1M
D. 0-075 M
F, 0,125 M
¢) Q..
15
10
10
5
10
20
30
10
20
30
10
20
30
F r a c t i o n number
Fig. 2. A-F. Effect of HU on the pattern of pulse-labelled DNA in alkaline sucrose gradients. A culture grown in medium containing 14C-uracil to 107eells/ml was divided into six parts and exposed to HU at the concentrations shown for 30 rain, Each was then pulselabelled for 40 rain, in the presence of the same concentration of HU, with 3H-adenine at: (A) and (B) 40 #Ci/ml, (C) 50 #Ci/ml, (D) 60 /~Ci/ml, (E) 80 t~Ci/ml, (F) 100 #Ci/ml. The DNA was extracted and sedimented on alkaline sucrose gradients as described in Methods. Sedimentation is from right to left. The total 14C-labelrecovered per gradient varied from 648 to 736 cpm, while the 3H-label recovered was: (A) 34,007 cpm, (B) 2,968 cpm, (C) 3,529 cpm, (D) 2,175 cpm, (E) 1,938 cpm, (F) 2,268 cpm. (~-v) 14C-label; (~ ~) 3H-label
In Fig. 2 incorporation in the presence of HU is low, only some 2%-8% o f the wild-type. It is, therefore, possible that DNA synthesis in only a small sub-population of the cells was being observed, rather than a general effect on the mechanism of DNA synthesis. This is unlikely for two reasons. Firstly, the observed DNA synthesis was abnormal (Fig. 2), so we are not simply dealing with a small population of cells resistant to the effects of HU. Secondly, when cells were preincubated in 0.05 M HU for 1.5 h, to allow 'escape' from the gross inhibition by HU (Fig.l), and then pulse-labelled, a gradient essentially identical to that in Fig. 2C was obtained (result not shown). In this case, however, the incorporation was
40% of the wild-type, of fivefold more than that in Fig. 2C. Thus, it seems likely that it is the residual synthesis in a substantial proportion of the cells which is being examined rather than a large amount of synthesis in a small proportion of the cells. Untreated cells can be pulsed for a shorter time, some 4 - 8 min, to give a level of incorporation similar to that in Fig. 2 B - F . This produces a tritiated profile on gradients (Johnston and Williamson 1978)resembling that of cells treated with 0.05 M HU, suggesting the effect of HU may merely be to slow down the rate of replication. This interpretation is ruled out by a series of pulses at 0.05 M HU and between 30 and 120 min duration. Desome
178
L. H. Johnston: Effect of Hydroxyurea on DNA Synthesis in Yeast
-
!
o
10
20
~
10
~
~
F. 4 0 rain
E. 2 0 rain
XlO
10
10
20
~
20
30
10
~
G, 60rain
1C
E u
51
10
2O
10
30
F r action
number
Fig. 3. A-H. Effect on DNA synthesis of prolonged incubation in 0.05 M HU. Ceils grown in the presence of 14C-uracil to 107/ml were exposed to 0.05 M HU for 30 rain and then pulse-labelled in the same concentration of HU with 50 ~Ci/ml 3H-adenine for (A, E) 20 rain, (B, F) 40 rain, (C, G) 60 rain, (D, H) 80 rain. The DNA was extracted and sedimented on alkaline sucrose gradients as described in Methods. Sedimentation is from right to left. In A - D the epm in each fraction are plotted as a percentage of the total cpm, as in Fig. 2, while in E-H the pulse-label from the same results is plotted as absolute cpm. The total 14C-label per gradient varied between 844 cpm and 861 cpm and the total 3H-label was (A) 1,776 cpm, (B) 5,731 cpm, ((2) 11,855 epm, (D) 20,273 cpm (o---e) 14C-label, (e---e) 3H-label
spite a greater than tenfold increase in incorporation during this time, all the piilses give essentially identical tritiated profiles on gradients (Fig. 3 A - D and unpublished observations). In Figs. 2 C - F minor peaks of pulse-label are present in fractions 9 and/or 11. These are a previously undescribed species of newly synthesised DNA and may represent the joining together of adjacent, completed replicons within a small cluster (manuscript in preparation). When an experiment similar to that in Fig. 2 was done, using a strain containing mitochondrial DNA, essentially identical results were obtained (results not shown). Therefore the results described above are not an anomaly arising from the use of a strain without mitochondfial DNA.
of label in this peak without any corresponding increase in molecular weight. This was found to be the case. Cells were pulse4abelled for periods of 2 0 - 8 0 min in 0.05 M HU and, despite an 11 A-fold increase in incorporation in this time, there was very little change in the relative distribution of pulse-label throughout the gradients (Fig. 3 A - D ; for a control see Fig. 2A). During the labelling period, however, there was a large increase in the absolute amount of pulse-label between fractions 22 and 11 and also in the top five fractions of the gradient (Fig. 3 E - H ) . During this period some high molecular weight DNA was also being synthesised (Fig. 3 E - H , bottom five fractions) in spite of the lack of any change in the molecular weight distribution o f the DNA between fractions 1 1 - 2 2 (Fig. 3 A - D ) . This further supports the idea that these molecules are completed replication intermediates and also indicates that some residual joining of them was occurring.
Effect of Prolonged Incubation in HU In Fig. 2B and 2C a reproducible 'peak' of pulse-label was present between fractions 22 and 11 with most label in fraction 18. If this peak represented an accumulation of completed replication intermediates, an increase in the period of pulse-labelling ought to'increase the amount
Release from HU Inhibition When cells pulse-labelled in the presence of HU were removed from the drug and incubated in medium free of HU, the pulse-labelled DNA was converted to the size of
L. H. Johnston: Effect of Hydroxyurea on DNA Synthesis in Yeast
179 radioactivity from the five fractions around the peak in fraction 18 (16-20). After the labelling in HU there were 4,407 cpm in these fractions (Fig. 4A) while after the incubation free of HU this figure was 2,095 cpm (Fig. 4B). Similarly, considering the top six fractions of the gradient, there was a decrease from 4,152 cpm to 2,890 cpm.
20
Discussion
E D_ U
10
20
30
10
20
30
c
u 2C c]_
15
10-
5
Fraction number
Fig. 4. A and B. Release from HU inhibition. Cells grown in the presence of 14C-uracil to 107/ml were exposed to 0.05 M HU for 30 rain and then pulseqabelled in the HU for 40 rain with 100 ~Ci/ml 3H-adenine. One half of the culture (A) was then placed in stop-mix, while the other half (B) was filtered, washed with YEPD, resuspended in 5 ml fresh YEPD without HU or label but containing 50 #g/ml unlabelled adenine and incubated a further 80 rain. The DNA was then extracted and sedirnented as described in Methods. Sedimentation is from right to left. The total 14C-labelwas 993 cpm and 1076 cpm, while the 3H-label was 15,968 cpm and 53,074 cpm respectively. (c--o) 14C-label, (~ ~) 3H-label
the prelabeUed DNA (Fig. 4). In this experiment the cells were simultaneously removed from labelled medium and the further incubation was in a medium containing an excess of unlabelled adenine. In yeast, however, it is not possible rapidly to terminate ('chase')pulse-labelling of DNA by this means so that there was a 3.3-fold increase in incorporation during the incubation in drugfree medium. Nevertheless there was an absolute loss of
The observed effects of HU on the mechanism of DNA synthesis are consistent with a partial inhibition of the joining of completed replicons. This hypothesis is supported by several lines of evidence. Firstly, the size distribution of the pulse-labelled molecules between fractions 22 and 11, corresponding to molecular weights of 7 - 6 0 million daltons, resembles exactly the size range of yeast replicons (Petes and Williamson 1975). Secondly, on increasing the pulse length in HU there was a dramatic increase in the amount of pulse-label in the replicons without any corresponding increase in their molecular weight (Fig. 3). This suggests that in the presence of HU completed replicons accumulate. Finally, on release from HU inhibition, there was an absolute loss of pulselabel from the peak of replicons, concomitant with the appearance of label in high molecular weight DNA (Fig. 4). The fact that the loss of radioactivity was less than complete reflects the known difficulty in pulse-chase experiments with yeast arising from the presence of high internal precursor pools. As well as the replicons, similar results were apparent with the population of small pulse-labelled molecules appearing at the top of the gradients (Figs. 2 - 4 ) . These approximately correspond in size to Okazaki fragments in yeast (Johnston and Nasmyth 1978) and it is, therefore, tempting to suppose that HU also inhibits the joining of these molecules. However, it is not certain that Okazaki fragments are real replication intermediates (Tye et al. 1977) and a number of artefacts are known to mimic the appearance of these fragments (for example see Anderson 1978). In the absence of clean pulsechase experiments it is, therefore, difficult to draw firm conclusions concerning the real nature of these molecules and discussion of the effects of HU on them are probably premature. The mechanism of inhibition of joining of replicons by HU is not clear. It could be affecting DNA ligase; however this would not explain the general inhibition of DNA synthesis, since a conditional yeast ligase mutant undergoes a complete round of DNA synthesis at the restrictive temperature (Culotti and Hartwell 1971; Johnston and Nasmyth 1978). Alternatively the drug may inhibit a DNA polymerising activity responsible for filling gaps left between adjacent completed replicons.
180 Since HU is known to inhibit ribonucleotide reductase (Elford 1968; Sinha and Snustad 1972), an effect on DNA polymerases could occur at the level of precursor supply, as suggested for the effect of HU on DNA repair (Collins 1977; Burg et al. 1977; Collins et al. 1977). This interpretation would provide sufficient flexibility to account for the experimental observations, assuming that there are DNA polymerising activities in the cell with different KmS. Lowered precursor pools would account for the general depression in DNA synthesis (Fig. 1). While the accumulation of replicons would occur if the synthesis of intermediates preceding replicons (in effect, gross DNA synthesis) was carried out by a polymerase with a low Km able to draw more efficiently on decreased deoxyribonucleotide pools than the polymerase responsible for filling gaps between replicons. In fact in vitro the yeast DNA polymerases have somewhat similar Kms (Wintersburger 1974; Chang 1977). However in association with other proteins their Kras may be altered, as has been observed for the interaction between DNA binding protein and polymerase in Ustilago (Yarranton et al. 1976). In the light of these results it is instructive to reconsider our previous studies on the mechanism of DNA synthesis in yeast (Johnston and Williamson 1978). There is a striking coincidence in the size distribution between the completed replicons accumulating in cells treated with HU (Figs. 2B-C and 3) and the DNA molecules sedimenting in fractions 22-11 after pulseqabeUing of (untreated) log-phase cells for 4 - 8 rain (Fig. 2 in Johnston and Williamson 1978). This suggests that the molecules evident in the untreated cells are completed entities rather than growing intermediates. This implies that, in yeast mitotic DNA synthesis, there must be a lag between the completion of a replicon and its joining to
L.H. Johnston: Effect of Hydroxyurea on DNA Synthesis in Yeast adjacent completed replicons. We were unable to detect such a lag using synchronous cells (Johnston and Williamson 1978) but possibly it is very brief and the degree of synchrony may have been insufficient to detect it.
Acknowledgements. I thank Don Williamson for critical reading of the manuscript.
References Anderson MLM (1978) Nucleic Acids Res 5:4343-4354 Burg K, Collins ARS, Johnson RL (1977) J Cell Sci 28:29-48 Chang LMS (1977) J Biol Chem 252:1873-1880 Collins ARS (1977) Biochim Biophys Acta 478:461-473 Collins ARS, Schor SL, Johnson RL (1977) Mutat Res 42:413432 Culotti J, Hartwell LH (1971) Exp Cell Res 67:389-901 Elford HL (1968) Biochem Biophys Res Commun 33:129-135 HartweU LH (1976) J Mol Bio1104:803-817 Johnston LH, Game JC (1978) Mol Gen Genet 61:205-214 Johnston LH, Nasmyth KA (1978) Nature 274:891-893 Johnston LH, Williamson DH (1978) Mol Gen Genet 164:217225 Levin D, Hutchinson F (1973) J Mot Bio175:495-502 Petes TD, Williamson DH (1975) Exp Cell Res 95:103-110 Sinha NK, Snustad DP (1972) J Bacteriol 112:1321-1334 Slater ML (1973) J Bacteriol 113:263-270 Timson J (1975) Murat Res 32:115-132 Tye B, Nyman P, Lehman IR, Hochhauser S, Weiss B (1977) Proc Natl Acad Sci USA 74:154-157 Wintershurger E (1974) Eur J Biochem 50:41-47 Yarranton GT, Moore PD, Spanos A (1976) Mol Gen Genet 145: 215-218
Communicated by B. S. Cox Received June 20, 1980
