manipulating chromosome structure and ... - Semantic Scholar

J. Cell Sci. 73, 159-186 (1985)

159

Printed in Great Britain © The Company of Biologists Limited 1985

MANIPULATING CHROMOSOME STRUCTURE AND METAPHASE STATUS WITH ULTRAVIOLET LIGHT AND REPAIR SYNTHESIS INHIBITORS ANN M. MULLINGER AND ROBERT T. JOHNSON Cancer Research Campaign Mammalian Cell DNA Repair Group, Department of Zoology, University of Cambridge, Downing Street, Cambridge CB2 3EJ, U.K.

SUMMARY DNA repair occurs in metaphase-arrested cells in response to ultraviolet irradiation. In the presence of the repair synthesis inhibitors hydroxyurea and 1-fS-D-arabinofuranosylcytosine the chromosomes of such cells, as seen in Carnoy-fixed preparations, are decondensed. The extent of decondensation is related to both the u.v. dose and the duration of incubation in the presence of inhibitors. For any particular cell type there is a reasonable correlation between the amount of decondensation and the number of single-strand DNA breaks generated by the repair process under the same inhibitory conditions, though the chromosome changes continue after the number of single-strand breaks has reached a plateau. The dose response of chromosome decondensation varies between different cell types but is in general correlated with differences in levels of single-strand breaks accumulated under comparable inhibitory conditions. Decondensation can be detected after 0-5 Jm~ 2 in repair-competent human cells. In human cells defective in excision repair there is much less chromosome decondensation in response to the same u.v. dose and time of repair inhibition. However, a simian virus 40transformed muntjac cell displays pronounced chromosome decondensation but has limited incision ability. Both chromosome decondensation and single-strand break accumulation in the presence of inhibitors are reversed when DNA precursors are provided, but reversal after higher u.v. doses and longer periods of incubation leads to recondensed chromosomes that are fragmented. Elution of the DNA from such cells through polycarbonate filters under non-denaturing conditions reveals that double-strand DNA breaks are generated during the period of incubation with inhibitors. Although the chromosomes of repair-inhibited metaphase cells are decondensed in fixed preparations, their morphology appears normal in intact cells. The cells also retain a capacity to induce prematurely condensed chromosomes (PCC) when fused with interphase cells: compared with control mitotic cells, the speed of induction is sometimes reduced but the final amount of PCC produced is similar.

INTRODUCTION

The apparent inertness of the DNA of mitotic chromosomes, judged by their inability to transcribe or replicate, is not evident when DNA repair is considered. For instance, ultraviolet light induces considerable repair activity in a wide range of mitotic cells, both human and rodent (Schor, Johnson & Waldren, 1975; Burg, Collins & Johnson, 1977; Johnson & Sperling, 1978; Johnson & Collins, 1978). Nitrosamides (Collins, Ord & Johnson, 1981) and X-rays (Rydberg, 1983) also Key words: cnromosome decondensation, damage and fragmentation, DNA repair inhibition.

160

A. M. Mullinger and R. T. Johnson

induce mitotic repair, suggesting that a number of different repair pathways can be mobilized in the mitotic cell. Our interest in metaphase chromosome repair arose out of an observation that very high levels of u.v. irradiation (in excess of 200 Jm~2) resulted in decondensation of metaphase chromosomes and that this effect was potentiated in the presence of hydroxyurea (HU) or 1-fi-D-arabinofuranosylcytosine (araC) (Schor et al. 1975). HU and araC are inhibitors of u.v.-induced excision-repair synthesis, acting after the initial incision step and resulting in the accumulation of single-strand (SS) breaks in the DNA. Further work suggested that a quantitative relationship existed between the number of SS breaks accumulated in the presence of HU and araC and the degree of chromosome decondensation (Collins, Schor & Johnson, 1977; Johnson & Collins, 1978) and it seemed likely that enlargement of the DNA breaks into gaps was necessary to produce this decondensation (Johnson & Collins, 1978). Although the cells were exposed to supralethal doses of u.v., the disintegration of chromosome structure was surprisingly easily reversed by providing DNA precursors following a period of inhibited repair. Recondensation of chromosomes coincided with the removal of single-strand breaks/gaps (Johnson & Collins, 1978). In the present work we have re-examined chromosome decondensation using repair synthesis inhibitors much more stringently than before (i.e. using HU and araC together, essentially as done by Squires, Johnson & Collins, 1982). This has allowed us to use very low, biologically relevant amounts of u.v. to disturb chromosome structure. The results indicate that chromosome decondensation is at least as sensitive as any other repair assay in current use, and can be used to discriminate between cells differing in their u.v. sensitivity. In addition we show that prolonged repair inhibition in metaphase results in chromosome breakage, i.e. irreversible damage; fragmentation is dependent on u.v. dose and duration of inhibition and is associated with the appearance of double-strand (DS) breaks in the DNA. In this paper we have also explored the state of some other cellular characteristics of repair-inhibited metaphase cells. We show that the decondensed chromosomal organization revealed in standard fixed, spread preparations is not apparent in the intact cell: Moreover, the repair-inhibited cell retains chromosome-condensing factors, although the latter may be somewhat reduced in potency compared with unirradiated metaphase cells. MATERIALS AND METHODS

Cell culture and synchronization Cells were grown in monolayer in Eagle's Minimal Essential Medium (MEM) supplemented with non-essential amino acids and serum (Gibco Europe; 10 or 15 % foetal calf serum was used for primary cells and 5 or 10% of a mixture of foetal and newborn calf serum for all other cells). The sources of the cells were as follows: human embryonic lung fibroblasts, passages 7-12 from Gibco; xeroderma pigmentosum (XP) cells belonging to complementation groups C (XP4BR), D (XP1BR) and G (XP2BI), a gift from Dr A. Lehmann, MRC Mutation Unit, University of Sussex, Brighton; HT 1080 fibrosarcoma cells, gift from Dr R. Baker, Massachusetts Institute of Technology; HD2 hybrids produced by fusion between XPD and HeLa cells (Johnson, Squires,

Manipulating chromosome structure

161

Elliott, Koch & Rainbow, 1985); simian virus 40 (SV40)-transformed cells of the Indian muntjac Munljac muntiacus kindly supplied by Dr. K. Sperling, Free University of Berlin; 10a2b cells produced by spontaneous transformation in vitro of embryonic mouse fibroblasts (Elliott & Johnson, 1983). All mitotic cells were accumulated by growing cultures for 9 h in high-pressure nitrous oxide (5-4 atm; 5-066X 105 N/m 2 ; Johnson, Mullinger & Downes, 1978); in some cases cells were also presynchronized by growth for 20 h in 2-5 mM-thymidine prior to mitotic arrest (Johnson et al. 1978). Detached, mitotic cells were collected as floaters in the medium either immediately or after gentle shaking of the dishes; in some cases mitotic cells were mixed with trypsinized interphase cells in order to produce enough material to handle in experiments. When prelabelled mitotic cells were required, cultures were grown for 24 h prior to synchronization in the presence of [methyl3 H]thymidine (0-1 (iCi/ml, 48 Ci/mmol, Radiochemical Centre, Amersham).

Experimental procedure Mitotic cells were spun down, resuspended in prewarmed medium with inhibitors (unless otherwise stated, 10~z M-hydroxyurea and 10~4M-l-p-D-arabinofuranosylcytosine) and incubated for 30 min in plastic Petri dishes. In some experiments araC was used at other concentrations or aphidicolin, usually at 10/ig/ml, replaced araC; details are given in Results. Colcemid (005^.g/ml) was present throughout the experimental procedure except at the irradiation step. After pre-incubation cells were spun down, resuspended in prewarmed phosphate-buffered saline (PBS), with inhibitors, at a density of not more than 2xlO 5 /cells per ml, care being taken to ensure that cells were well dispersed and not clumped, u.v. irradiation, of 10-ml samples in 90 mm plastic Petri dishes, was carried out with a germicidal tube emitting predominantly at 254 nm at a dose rate of 0 02-l-0Jm~ 2 s~'. Irradiated cells were spun down and resuspended in fresh medium with inhibitors and post-incubated for !/z, 1 or 2 h; in one series of experiments with HeLa cells [me//iy/-3H]thymidine (5/xCi/ml, 48 Ci/mmol) was also present in the post-incubation medium. For control samples the u.v. irradiation or the inhibitors were omitted. In some experiments cells were given a second period of post-incubation after centrifugation and resuspension in normal medium, without inhibitors and either with or without the four deoxyribonucleosides (dXs) each at 10~ 4 M. Cell fusions were carried out with u.v.-inactivated Sendai virus, as described by Rao & Johnson (1972).

Chromosome preparations After post-incubation chromosome preparations were made by standard procedures: cells were hypotonically swollen in Hanks' balanced salt solution (Hanks & Wallace, 1949) diluted 1:4 with distilled water at 37°C for 7-20 min, depending on the cell type, and fixed in three changes of Carnoy's fixative (methanol:acetic acid, 3:1, v/v). Fixed material was dropped onto clean microscope slides and air dried before staining in 6% Giertisa. For autoradiography chromosome preparations on slides were extracted with trichloroacetic acid at 4°C and coated with Ilford K2 nuclear emulsion diluted 1 :2 (v/v) with 2 % glycerol, exposed for 3-6 days at 4CC and developed for 3 min at 20°C in Ilford D19 developer.

Measurement of single- and double-strand DNA breaks The accumulation of single-strand breaks in metaphase cells in the presence of inhibitors was measured as described earlier (Collins, 1977; Collins & Johnson, 1981). Briefly, mitotic cells prelabelled by incubation in [me/-3H]thymidine (Fig. I F ) . Although there is necessarily a certain amount of subjective judgement in the assignment of chromosome spreads to the various decondensation categories, the scheme allows useful discriminations and comparisons to be made, and is particularly reliable for samples within one experiment and for one cell type. Some caution must be exercised in its use both for comparing different experiments on the same cell type and for comparing different cell types, where the pattern of condensation may be slightly different. The latter problem arises most acutely in the case of the Indian muntjac, which has a small number of very large chromosomes. Chromosdme decondensation here differs from HeLa in two chief respects. First, all the chromosomes in a spread from one cell tend to appear decondensed to the same extent. Second, in the initial stages of decondensation the spiral chromosome axis is more distinct; with further attenuation the axis remains particularly sharp but loses its spiral nature. Nevertheless it is possible to accommodate the muntjac pattern within the HeLa scale (Fig. 2). Chromosome decondensation in a range of different cell types Repair-competent human cells. Fig. 3 shows the amount of chromosome decondensation produced in HeLa cells in response to different amounts of u.v. irradiation and different post-incubation periods in the presence of DNA synthesis

Manipulating chromosome structure

2A

165

B

-•

1

Fig. 2. Light micrographs of muntjac chromosome spreads illustrating progressive stages of decondensation (A-D). A. Chromosomes just beginning to decondense at the telomeres (category 2); B, chromosomes elongated and paler, spirals of chromatid axes extended (category 4); C, axes of chromatids almost completely uncoiled and some expanding laterally (category 7); D, chromosomes extensively decondensed and merged together into a homogeneous mass (category 10). u.v. doses and lengths of postincubation with inhibitors were for A-D, respectively: A, 0-5Jm~ 2 , 90min; B, 2Jm~ 2 , 90min; c, 2Jm~ 2 , 90min; D, 5Jm~ 2 , 90min. Giemsa stained. X900.

inhibitors. For u.v. doses in the range 0-5—90Jm 2 chromosome decondensation increases with time of incubation from 0-30 min and again from 30-90 min; however, there is very little further increase in the period from 90-120 min, at least for the two u.v. doses tested (0-5 and S Jm" 2 ). Incubation of unirradiated cells in the presence of inhibitors does not lead to decondensation. Surprisingly, a response at the chromosomal level can be detected at u.v. doses as low as 0-5—1 Jm" 2 ,

166

A. M. Mullinger and R. T. Johnson A

0 miij

'30 min

90 min

120 min

100-

50-

•rfm.

0 100-

50-

0. 100-

£•

ft

30

30

30

90

90

90

10

10

10

100

100

100

50-

0^ 100'

50-

0-Lta 100-

50-

100-

50-

PI I I I I I I I

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Chromosome decondensation category Fig. 3 (first part)

Manipulating chromosome structure A

0 min

15 min

167 120 min

90 min

100-

50-

0 100'

0-5

0-5

0-5

ory

50-

an t>

sc •3 t Q.

0 100

50

nftl

t£ 0 100

50-

JfL

0 100-

50-

11

111

11

i n

11

i i 11

i 11

11

11111

111111

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 1 0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 C h r o m o s o m e d e c o n d e n s a t i o n category Fig. 3. Histograms showing the percentage of HeLa chromosome spreads in each decondensation category (see Fig. 1) for a range of u.v. doses (shown at top right-hand corner of each histogram in Jm ) and times of post-incubation (shown above each panel in min) in the presence (A) or absence (B) of inhibitors. At least 100 spreads were scored for each histogram. Data for two separate experiments at 5 Jm~ 2 are given, to illustrate the variation typical of these procedures.

168

A. M. Mullinger and R. T. Johnson

although this requires incubation beyond 30 min and careful examination of preparations to detect the slight effect. With increasing u.v. the amount of decondensation also increases, though the response saturates at about 10—20Jm , regardless of whether incubation is for 30 or 90 min. Omission of inhibitors during incubation prevents decondensation at u.v. doses below 10 Jm~ 2 ; under these inhibitor-free conditions the chromosomal response to a u.v. dose of 100 Jm~ 2 is similar to that for 1 Jm~ 2 in the presence of inhibitors (Fig. 3). The pattern of chromosome decondensation in another human tumour cell, the near diploid fibrosarcoma HT 1080 (data not shown) and also in normal human fibroblasts (Fig. 4A) is very similar to that of HeLa in terms of dose and time dependence; the effect saturates by 90 min and can be detected at u.v. doses as low as 0-5 Jm . The variations between histogram profiles for these different human cells are within the range for repeat experiments for any one cell type (e.g. HeLa, 5 Jm- 2 ; Fig. 3). Repair-defective human cells. The chromosomal responses of several repairefective human cells, including u.v.-sensitive xeroderma pigmentosum fibroblasts belonging to complementation groups C, D and G, are shown in Fig. 4. The extent of decondensation in all these cells is greatly reduced compared with that in normal human fibroblasts but there is variation between the response of the different groups. For groups C and G there is no detectable decondensation even after 20 Jm~ 2 and 90min incubation; in group D cells, a dose of 5 Jm~ 2 can be detected at 90 min though there is no significant increase in the extent of decondensation for a dose of 30 Jm~ 2 , nor for a more extensive period of incubation up to 120 min. A similar correlation between low repair capacity and reduced chromosome decondensation is also seen in another cell type, the permanent hybrid cell line (HD2) produced by fusion between XP(D) and HeLa. HD2 cells have a repair capacity close to that of the XP (D) parent (Johnson et al. 1985) and the response at the chromosome level is also broadly similar (Fig. 4). Decondensation is just detectable in the hybrid after 3 Jm~ 2 and 90 min incubation. Other cell lines. The chromosomal response of a number of animal cell lines differs from that of HeLa and normal human fibroblasts in several respects. For SV40-transformed Indian muntjac fibroblasts, chromosome decondensation continues for at least 120 min after irradiation and by this time a u.v. dose as low as 0-2 Jm~ 2 is detectable (Fig. 5). This enhanced response of muntjac chromosomes is also seen at higher u.v. doses, particularly if incubation is continued for 120 min (Fig. 5). Per unit of fluence, therefore, muntjac cells are more sensitive, so far as chromosome decondensation is concerned, than any of the human cells studied despite the fact that in terms of ability to accumulate DNA breaks they are relatively poor (see below). The extent of chromosome decondensation in a second animal cell shows some similarities to the muntjac cell line; 10a2b is a transformed mouse cell with rather poor ability to accumulate DNA breaks (Elliott & Johnson, 1983). Chromosome decondensation does not saturate until after 90 min but is much less extensive, dose

Manipulating chromosome structure

90 min

90 min

90 min

90-

E 60-

60--i

30-

300-

0 20

60300 60-

60-

30-

30-

i: o o

0

«H •8

S

II

60300 6030-

itL

0

60

60-

30

30"

Q.

0

0' 60-

60

30-

30-j

0 30

10

60-

10

jru

20

30-

0

0

60

169

0 D

60

30

30

0

0 60-

D

20

300

1 2 3 4 5 6 7 8 9 10

h,.,,,.

1 2 3 4 5 6 7 8 9 10

1 2 3 4 5 6 7 8 9 10

Chromosome dccondensation category

Fig. 4. Histograms showing the percentage of chromosome spreads in each decondensation category for various human cells with different incision capacities: normal human fibroblasts (A); xeroderma pigmentosum group C (B), group D (c) and group G (D); and the HeLa/XPD hybrid HD2 (E). U.V. doses are shown in the top right-hand corner of each histogram in Jm~ 2 . All data refer to 90 min post-incubation with inhibitors. At least 100 spreads were scored for each histogram.

170

A. M. Mullinger and R. T. Johnson

for dose, than in muntjac cells. For example, after exposure to 5 Jm~2, chromosomal effects are detectable after 120min incubation but not at 90min (Fig. 5). Incision capacity of various cell lines in relation to chromosome decondensation Comparison of observations on chromosome decondensation with measurements of SS breaks in cells that have been u.v. irradiated and incubated in the presence of HU and araC under comparable conditions, suggests that there is a general correlation between these two parameters. For random populations of HeLa cells the number of breaks accumulated increases with both the u.v. dose and length of incubation up to a plateau level (Johnson & Collins, 1978; Squires et al. 1982; Table 1), and this is also the case for chromosome decondensation. With respect to u.v. dose the plateau level is quite similar for both phenomena. However, the time courses of DNA break accumulation and chromosome behaviour do not show such good correspondence. Break accumulation at any particular u.v. dose is rapid, reaching about 70% of its final value by 30 min (Table 1; Squires et al. 1982; Squires, personal communication), while major chromosome decondensation continues between 30 and 90 min of incubation. The more extensively decondensed chromosomes (categories 9 and 10) are found only with saturating levels of SS DNA breaks (about 15 breaks/109 daltons). For each of the other cell types there is a similar broad relationship between SS break accumulation in random populations and chromosome decondensation, with respect to u.v. dose and length of incubation (Table 1). In all cases chromosome decondensation takes longer to reach a plateau than break accumulation, and in those cells where decondensation continues beyond 90 min the time taken to reach the maximum number of SS breaks is also slightly longer (Tablel). Comparison of the dose response of chromosome decondensation in different cell types reveals a broad relationship with levels of SS break accumulation, though there are some interesting anomalies. For example, normal human fibroblasts accumulate two to three times as many SS breaks as HeLa when exposed to the same conditions of irradiation and incubation with inhibitors (Squires et al. 1982; Table 1), and yet the chromosomal response in these two cells is not significantly different in terms of dose/time response and levels of sensitivity. On this basis HeLa chromosomes appear to be more sensitive to decondensation for given numbers of SS breaks. However, these SS break data refer to random populations and, although other work has shown that levels of SS breaks are only slightly reduced in metaphase HeLa compared with cells in G\ and S phase (Downes & Collins, 1982), it seemed

Fig. 5. Histograms showing the percentage of SV40-muntjac (A) and mouse 10a2b (B) chromosome spreads in each decondensation category for a range of u.v. doses (shown at top right-hand corner of each histogram in Jm~ ) and times of post-incubation (shown above each panel in min) in the presence of inhibitors. At least 100 spreads were scored for each histogram.

Manipulating chromosome structure 0 min

60- 1 40-

15 min

90 min

0.

0

0

0

1

20.

171

120 min

I

"L

0-2

604020-

1 h-t-ru

060-

°'5

1

40-

0-

2

604020-

060-

H

£

40-

*

20H

0-

20

1L

2

2

5

20-

1 a-

0-5

rwrffTU

I 40o 8

JL

1

20-

0-5

Th

20

20 r

lKVu_

100- -l

50-

0.

100'

50-

2*3'4' 5*6' 7* 8*9 io' 123\'s Chromosome decondensation category Fig. 5

6 7 8 910

***

120

8.5

2.8 7.8

20

2.5 3.0 3.8

13.0

1.4 3.8 9.8

30

0.65

40

12.1

4.6 9.0

60

2-8 3.4

75

tExtcnt of chromosome decondensation indicated on star rating: 0, not detectable; f B d on data for random populntions.

Mouse lOn2b

HeLa/XPD hybrid HD2

XPD XPDIBR

0.25

9.1

0.5 2.4 6.3

15

1.14

80

120 Squires, Johnson & Collins (1982); Squires (personal communication)

Source of S S b r u k data

Elliott &Johnson (1983)

Pillidge & J o h m n (unpublished)

Johnson, el al. (1985); Johnson (unpublished)

*, just detectable; **-, fully decondenscd.

6.6 6.8 7.9

15.8

13.6

90

Squirm & Johnson (unpublished)

**w

**I

** ***

90

XPC XP4BR

**

** **

0 0

30

No. of S S b r e n h per lo9 dnltons DNA after varioua post-incubation t i m a (min)f

Squircs cl a/. (1982); Squires (peraonnl communication)

0.5 1.O 4.0 5.0 8.0 30.0

U.V.

dose (Jm-') 15

Extent of chromosome decondensation after various post-incubation timm h i m

Comparison of SS break accumulafion levels with extent of c h m o s o m e decondensation for vanous cell types

Normal humnn fibmblmts

Cell t YPe

Table 1.

Manipulatimg chromosome structure

173

Table 2. Comparison of SS break accumulation levels for metaphase-arrested and interphase populations of human fibroblasts u.v. dose

(Jm-2) 1 1 2 2 4 4 10 10

Time of post-incubation (min)

Metaphase : interphase ratio of SS breaks per 109 daltons DNA

20 40 20

0-63 0-29 0-40 0-46 0-25 0-38 1-18 0-50

40 20 40

20 40

Metaphase and interphase cells taken from cultures that had been in nitrous oxide for 9 h without a previous thymidine block.

possible that normal fibroblasts might show a greater cell cycle differential in this parameter, as was found, for Chinese hamster cells (Collins, Downes & Johnson, 1980). We have therefore measured the accumulation of SS breaks in metaphasearrested diploid fibroblasts and find that the levels are much lower than for a random population (Table 2) and similar to those of mitotic HeLa cells. Thus, from the viewpoint of SS breaks the chromosomal responses of normal human fibroblasts and HeLa cells are much the same; in both cells the threshold level for detection of chromosome decondensation corresponds to about one to two SS breaks per 109 daltons DNA. Turning to the repair-defective human cells, break accumulation data are available only for random cultures. Nevertheless, for the three XP complementation groups and the hybrid HD2, there is a correlation between break accumulation and chromosome decondensation (Table 1). Interestingly, for all these cells the threshold conditions for detectable chromosomal decondensation correspond to SS break levels close to those for normal human fibroblasts at threshold (i.e. 4-5 per 109 daltons), if data from random cultures are compared in each case.For XP (C) and (B) cells this threshold level was not attained under the conditions used in our experiments (20 Jm~2 and 90 min incubation) and no chromosome decondensation was observed. Per unit of u.v. fluence, muntjac cells show enhanced chromosomal behaviour compared with human cells. By contrast, SS break accumulation in these cells is poor (Pillidge & Johnson, unpublished data; Table 1). At the threshold for chromosome decondensation the break levels correspond, by extrapolation from random populations at higher doses, to less than 1 per 109 daltons. For the mouse transformed cell line 10a2b numbers of SS breaks for random populations are similar to those of muntjac. However, in terms of chromosomal sensitivity to absolute break levels, 10a2b cells are closer to HeLa (Table 1).

A. M. Mullinger and R. T. Johnson

174

Recovery of condensed chromosomes We have previously shown (Johnson & Collins, 1978) that the changes in chromosome morphology produced by the inhibition of DNA repair following high doses of u.v. can to some extent be reversed if the inhibition is removed by addition of deoxyribonucleosides (dXs) to the post-incubation medium. Here, we examine the process of recondensation in more detail over a longer time course and at much lower u.v. doses. In these experiments mitotic HeLa cells were irradiated with up to 20 Jm~2, but after 30—90 min incubation with inhibitors the cells were incubated for a further 60min in the presence of dXs. Examination of chromosomes at various times during this procedure shows that the extensive decondensation, associated with repair inhibition, is reversed and that condensed chromosomes reappear. The chromosome decondensation/recondensation cycle is correlated with the appearance of singlestrand breaks in metaphase DNA and the subsequent loss of most of these when inhibitory conditions are reversed (Fig. 6). After the low u.v. doses (up to 5 Jm~2) and with short periods of inhibition before reversal (e.g. 30min), the appearance of chromosomes after a 60min incubation with dXs is indistinguishable from normal metaphase. After higher doses (10 Jm~2) and longer periods of inhibition, i.e. conditions that induce extensively decondensed chromosomes, the condensed metaphase morphology also returns although the time

5< O

I

ffl

30

60

90

120

150

Time after u.v.(min) Fig. 6. Single-strand DNA break accumulation in metaphase-arrested HeLa cells as a function of time of post-incubation with inhibitors. Data are shown for 0 5 (A—A), 1 ( • — • ) and 5(B—B)Jm~ 2 u.v. At 90min after irradiation inhibitors were removed from a parallel 5 Jm~ 2 sample, which was incubated for a further 60 min in the presence of the four deoxyribonucleosides (dXs) each at 10~4M ( • •).

Manipulating chromosome structure

175

Table 3. Relationship between u.v. dose, duration of repair inhibition and number of HeLa chromosome fragments produced on reversal of repair inhibition

Duration of post-incubation with HU and araC — (min)

u.v. dose

Number of chromosome pieces after 60 mm further incubation with dXs* Mean

S.D.

0 2 10 20

30

65-5 65-5 685 71-1

2-8 3-6 3-2 4-9

0 2 S 10 20

90

66-8 65-8 73-2 162-0 202-5

2-6 2-8 7-9 48-7 70-3

•Counts include whole chromosomes and detached chromosome fragments, and were made on photographic prints. At least 40 cells were scored for each sample.

required for complete recondensation is somewhat longer. The recondensed chromosomes produced under such conditions are also markedly different from their normal metaphase counterparts in one respect, namely that they are fragmented (Fig. 7). The extent of fragmentation is related to the u.v. dose and to the duration of the period of repair inhibition (Table 3). For example, with 60 min inhibition

. .^:££r-^--

:X.v.

7A

B Fig. 7. Light micrographs of HeLa chromosome spreads, showing reversal of decondensation leading to fragmented chromosomes. Metaphase cells were irradiated with 20Jm~ 2 u.v., incubated with inhibitors for 90min, followed by a further 60min period either in the continued presence of inhibitors (A) or in the absence of inhibitors and presence of 10~3M-dXs (B). Giemsa stained. X750.

176

A. M. Mullinger and R. T. Johnson

before reversal, the average number of fragments per cell in excess of the normal chromosome number increases from about 8 after 5 Jm~ 2 to about 140 after a dose of 20Jm~ . If the inhibition period before reversal is reduced to 30 min, fragmentation is detected only after 20Jm~ 2 . The marked increase in the number of fragments produced if inhibition continues for 60 rather than 30 min before reversal contrasts with the decline in rate of appearance of single-strand gaps in mitotic HeLa cells in the 30-60 min period (Fig. 6); it accompanies, however, the continuing chromosome decondensation that occurs beyond the first 30 min period.

6 9 12 Elution time (h)

15

6 9 12 Elution time (h)

15

Fig. 8. Effect of repair inhibitors on the elution behaviour of u.v.-irradiated metaphase DNA under 'neutral' (pH9 - 6) conditions, A. U.V. dose response after 150 min postincubation with HU and araC: ( • ) no u.v.; (A) 10Jm~ 2 ; ( • ) 20Jm~ 2 ; ( • ) 4€Jm~ 2 . B. Post-incubation time course after 20Jm~ 2 u.v.: (A) 30min; ( • ) 90min; ( • ) ISO min; ( • ) no u.v., 150 min.

Manipulating chromosome structure

111

The appearance of metaphase chromosome fragments (and they are seen also if aphidicolin replaces araC), is temporally associated with the appearance of doublestrand breaks in the DNA. We have detected such breaks in metaphase-arrested HeLa cells using the sensitive technique of neutral filter elution (Bradley & Kohn, 1979; Coquerelle & Weibezahn, 1981). The appearance of DS breaks is related both to u.v. dose and to the duration of incubation in the presence of inhibitors (Fig. 8). After 20Jm~2, for example, the amount of radioactivity eluting from the filter increases above control values only when incubation with inhibitors is continued beyond 30min. The progressive increase in DS break frequency at later times possibly correlates with the continuing chromosome changes in these cells, u.v. followed by incubation without inhibitors does not increase the rate of DNA elution over the same time course (Fig. 8), while addition of deoxyribonucleosides for the last 60min of a 150 min post-incubation period results in a reduction in the amount of radioactivity eluted from the filter (2 experiments, data not shown). The status of the repair-inhibited metaphase cell So far we have examined the characteristics of repair-inhibited metaphase cells in terms of their DNA lesions and of their chromosomal morphology seen in hypotonically swollen, fixed, spread cells. A question therefore arises as to whether these changes can be detected in the intact cell, both structurally and functionally. State of chromosomes in unfixed repair-inhibited metaphase cells. When viewed by Nomarsky interference optics the chromosomes of repair-inhibited HeLa cells are clearly visible as distinct structures and are indistinguishable from their normal metaphase counterparts. This applies even after high u.v. doses and long periods of incubation with inhibitors (e.g. 10Jm~2 and 90min), i.e. conditions that produce extensive attenuation in standard light-microscrope preparations of fixed cells. Similarly, repair-inhibited chromosomes also look 'normal' after incubation of cells with ethidium bromide (20/u.g/ml) and examination by fluorescence microscopy. Thus, the repair-inhibited state does not appear to disrupt the higher-order structure of metaphase chromosomes in the intact cell but is revealed or potentiated by changes introduced after hypotonic treatment, fixation and spreading. Mitotic inducing factors in repair-inhibited metaphase cells. One of the features of normal metaphase cells is the ability to induce prematurely condensed chromosomes (PCC) from the nuclei of interphase cells with which they are fused (Johnson & Rao, 1970). We have investigated whether repair-inhibited metaphase cells are competent in this respect. HeLa cells incubated with inhibitors for 90min after 20 or 300 Jm~2 u.v. induce PCC in interphase HeLa cells when fused in the continued presence of inhibitors (Fig. 9). In general (3 out of 4 experiments) the speed and extent of induction is the same as for control cells but induction can be slower, as revealed in a fourth experiment (Table 4). In this case, the number of nuclei converted to PCC after 20 min of fusion is much less for interphase cells fused with repair-inhibited cells

Ill

A. M. Mullinger and R. T. Johnson

Table 4. Comparison of PCC-inducing capacity of irradiated and unirradiated metaphase HeLa cells incubated with inhibitors u.v. dose given to metaphase cells

Duration of incubation at 37 °C (min)f

State of fused interphase cellsj Uninduced nuclei (%)

G, and G2 phase PCC (%)

S phase PCC (%)

20 60

52-9

3-8

45-2 57-7

1-9 38-5

20 20

20 60

90-9 10-9

9-1 61-4

0 27-7

300 300

20 60

89-S 130

10-5 60-2

26-8

(Jm- 2 )» 0

a

0

A sample of 2xlO 6 prelabelled metaphase cells, u.v. irradiated at the dose shown (•) and incubated with inhibitors ( l O - ^ - H U , 10~4M-araC) for 90min, was mixed with lXlO 6 random cells in 0'Sml Hanks' balanced salt solution, also containing inhibitors. After addition of Sendai virus the cell suspension was kept at 4°C for lOmin and then incubated at 37°C for the times indicated (f), after which portions were removed for chromosome preparations. The state of chromatin condensation of at least 100 fused interphase cells was scored for each sample from autoradiographs (%); fused cells having a ratio of interphase to metaphase partners greater than 2 were excluded from the counts. (Repeat experiments showed no difference in PCC-inducing capacity between irradiated and unirradiated cells.)

v

v{?

^Y^~

9 Fig. 9. For legend see p. 180

Manipulating chromosome structure

179

:j

10A

B

»»

D Fig. 10. For legend see p. 180

",..-' •v

w ^ » v.1 *M *H ~

-

_i

180

A. M. Mullinger and R. T. Johnson

than for unirradiated cells incubated with inhibitors (10% as against about 47%, respectively); moreover 5 phase PCC, which are more difficult to induce than G\ or Gz (Johnson & Rao, 1970), are found only in the control sample. If fusion is continued for 60min, however, the amount of PCC induced is only slightly lower with the repair-inhibited cells than in the control, the difference being accounted for by less 5 phase induction in the former case. There is no detectable difference in induction capacity between the cells irradiated with 20 and 300 Jm~2. In order to investigate whether the decondensed state of repair-inhibited chromosomes is directly related to reduced levels of chromosome-condensing factors we have fused repair-inhibited with normal metaphase cells. As shown in Fig. 10A, fusion in the presence of araC and HU (for up to 90 min) has no detectable effect on the state of condensation of the chromosomes of either partner. However, if inhibitors are removed from the time of fusion the repair-inhibited chromosomes recondense, an effect that is accelerated by the presence of deoxyribonucleosides in the medium. As in previous restitution experiments, recondensed chromosomes in this situation are also fragmented (Fig. 10 B—D).

DISCUSSION

The inhibition of DNA repair by agents such as araC and HU has a profound effect on several biological parameters including cell killing and the frequency of chromosome aberrations. Unfortunately the molecular mechanisms underlying the chromosome lesions are by no means fully understood (for reviews, see Kihlman, 1971; Johnson, Collins & Waldren, 19826; Collins, Downes & Johnson, 1984). Part of the problem lies in the time elapsed between damaging.the DNA, inhibiting the induced repair and finally, hours later, collecting the mitotic cells for chromosome examination. We have therefore concentrated on experimental situations where chromosomes can be inspected within a relatively short period after exposure to DNA-damaging agents, using the techniques of premature chromosome condensation (Waldren & Johnson, 1974; Hittelman & Rao, 1974; Schor et al. 1975;

Fig. 9. Light micrograph of a chromosome spread showing G] PCC (arrow) induced by a repair-inhibited metaphase cell. Metaphase HeLa cells irradiated with 300 Jm~ 2 u.v. and post-incubated with inhibitors for 90 min were fused with interphase HeLa cells in the continued presence of inhibitors, and processed 60 min after the start of fusion. Giemsa stained. X800. Fig. 10. Chromosome spreads from the products of fusion between irradiated, repairinhibited metaphase cells and unirradiated metaphase HeLa. Cells previously irradiated with 10Jm~ 2 u.v. and post-incubated with inhibitors for 90 min were fused with unirradiated metaphase cells. After 30 min of fusion the products were plated out and incubated in medium for a further 30 min (B) or 60 min (A,C,D). A. Inhibitors present during fusion, and subsequent incubation; irradiated chromosomes (arrow) remain decondensed, despite the presence of condensed, unirradiated chromosomes in the cell. B,C,D. Inhibitors not present during fusion and subsequent incubation but 10~4 dXs added; chromosomes of irradiated partner (arrows) partially condensed in (B) and (c) and almost fully condensed, though fragmented, in (D). Giemsa stained. X800.

Manipulating chromosome structure

181

Johnson et al. 19826; Collins et al. 1981) and metaphase cell repair (Schor et al. 1975; Johnson & Collins, 1978). In the present paper we have exploited the metaphase repair system further so as to explore the relationship between chromosome structure and DNA lesions, making use of the profound effect that inhibition of u.v. repair synthesis has on higher-order chromosome packing. Our studies suggest that one of the contributing factors in chromosome decondensation is the presence of abundant SS breaks in the DNA; 10Jm~2 of u.v. irradiation alone produces no detectable effect either on chromosome morphology in HeLa cells or on the continuity of DNA (Johnson & Collins, 1978; this paper), whereas incubation in the presence of repair inhibitors such as HU and araC, following u.v. doses as low as 0-5 Jm~2, leads to the accumulation of SS breaks in DNA, and also promotes chromosome decondensation. The fact that both chromosomal changes and DNA breaks increase with u.v. dose and time of incubation, supports the connection between these two phenomena. Our data in this paper confirm the general correlation between chromosomal response per unit of u.v. fluence and amount of SS break accumulation. However, there are exceptions. In particular, muntjac cells show far more chromosome disruption than other cell types for a given level of SS breaks. Whether the extreme sensitivity of muntjac chromosomes is associated with their large size remains to be answered. In a range of human cells of differing repair capacities the threshold level of DNA breaks that is detectable as producing significant chromosome decondensation is roughly constant, even though the u.v. dose required to produce the effect varies markedly according to cell type. This threshold level is of the order of 1—2 SS breaks (SSB) per 109 daltons DNA, corresponding to about 15 or 75 total breaks for the smallest and largest human chromosomes, respectively (assuming a range of DNA content of 30X109 to 150X109 daltons per chromosome; Lewin, 1974). Higher levels of breaks, around 15 or more SSB per 109 daltons (450-2000 per chromosome), are associated with the more extensively decondensed states. So far, in discussing chromosome decondensation in relation to the accumulation of SSBs we have referred to data based on hydroxyapatite chromatography, which strictly measures SS and DS breaks. There are, however, indications that DNA breaks alone are insufficient to produce decondensation. X-irradiation, even up to 20 krad, does not result in chromosome decondensation (Johnson & Collins, 1978), and therefore it seems likely that single-strand gaps in the DNA generated by excision and held open by inhibition of the resynthesis step are required. The existence of gaps rather than breaks in this material has been indicated (Johnson & Collins, 1978). Another indication that additional factors are involved in chromosome decondensation comes from the difference in time course of SSB accumulation and chromosome effects. Breaks saturate within about an hour of irradiation despite the existence of remaining substrate (Squires et al. 1982; this paper) and the continuing availability of repair endonuclease (Downes, 1984). Chromosome decondensation on the other hand is gradual and progressive, much of the massive attenuation occurring after the saturation of single-strand breakage. Thus secondary changes taking place in the chromatin after repair gaps have been created may promote decondensation.

182

A: M. Mullinger and R. T. Johnson

The fact that chromosomes can be made to recondense when HU and araC are removed suggests that the chromatin has not changed irreversibly during the period of repair inhibition. But recondensation to an apparently normal state occurs only when the period of inhibited repair has been quite short; if it is more prolonged, chromosome recondensation is accompanied by fragmentation, and this is correlated with the appearance of double-strand DNA breaks. It is, however, only after 30min incubation with inhibitors following 20Jm~ 2 u.v. that DS breaks begin to be detected in HeLa cells, i.e. when the accumulation of SS breaks is beginning to saturate. Correspondingly, reversal after only a 30min period of inhibition leads to few signs of chromosome fragmentation. The appearance of DS breaks (DSBs) appears therefore to correlate with the more extensive types of chromosome decondensation. Not unexpectedly, in the absence of inhibitors, u.v. alone (up to 40 J m ) generates few if any DSBs in mitotic cells over an extended time course when assessed by neutral elution. At much higher u.v. doses (100-300Jm~2) DSBs have been reported to accumulate in the absence of inhibitors (Bradley & Taylor, 1981). It is interesting that after similar amounts of u.v. there is extensive decondensation of prematurely condensed G\ chromosomes visualized by cell fusion (Waldren & Johnson 1974; Schor et al. 1975), though fragmentation has not been reported under these circumstances, probably because reversal with DNA precursors was not attempted. Although the present work provides clear evidence for the relationship between DSBs and chromosome breaks, quantitation of the DSBs by reference to the behaviour of X-irradiated DNA eluted under the same conditions is hazardous because of the non-linear nature of the X-ray elution profiles (Bradley & Kohn, 1979; Mullinger & Johnson, data not shown). Therefore the question of equivalence between DSBs and the number of chromosome fragments per cell must remain speculative. In other systems removal of the DNA synthesis inhibitors is followed by repair of DSBs (Bryant, 1983, 1984) and it is possible, therefore, that the recondensed fragments we see after DNA precursors have been added represent residual unrepaired DSBs. Though our knowledge of chromosome organization is still poor it seems likely on the basis of available models (e.g. see Paulson & Laemmli, 1977; Mullinger & Johnson, 1980) that not all unrepaired DSBs will result in chromosome breakage: for example, breaks in lateral loops are less likely to cause fragmentation than those in more axial regions of the chromosome. The question remains about the origin of the DSBs. Presumably only a few DSBs are generated by overlap of dimer repair patches in opposite strands (Bradley & Taylor, 1981) since the majority of DSBs appear after SSBs have saturated. Many will probably arise through a different mechanism. Recently, viscoelastometric studies have shown that inhibitor accumulated SS breaks are precursors for the production of DSBs (Filatov & Noskin, 1983). Other workers have also shown that exogenously provided single-strand-specific endonuclease can convert existing single to double-strand lesions, and that this results in a change in the nature of the chromosome aberrations (Natarajan et al. 1980; Bradley & Taylor, 1981). More recently, Ahnstrom & Bryant (1982) reported the conversion of single-strand lesions

Manipulating chromosome structure

183

to DSBs by endogenous endonuclease action, postulating that the substrate in the Xirradiated cells could be base damage. The endogenous production of DSBs reported in the present paper and also by Filatov & Noskin (1983) may reflect the presence and abundance of non-dimer photochemical lesions in the strand opposite the partially completed repair site. Chromosome decondensation may also be related to another secondary consequence of SSB accumulation in the presence of inhibitors. Though breaks accumulate in the presence of inhibitors, repair synthesis is apparently not suppressed; in some cases incorporation of radioactive thymidine is increased compared with the level in the uninhibited, repairing cell (Mullinger et al. 1983). This paradoxical situation may be explained in terms of an increase in the average patch length of newly synthesized DNA, for which there is some evidence (Clarkson, 1978), and it has been suggested (Collins, 1983; Mullinger, Collins & Johnson, 1983) that in the presence of synthesis inhibitors, exonuclease activity is maintained ahead of the resynthesis, creating an ever-longer gap and resulting repair patch with the extra possibility of generating overlapping gaps in the DNA duplex. Perhaps such a distortion of the repair process promotes instability of higher-order chromatin structures. The demonstration of decondensed chromosomes from repair-inhibited metaphase cells raises the possibility that the state of the chromatin has shifted in some respects towards that of interphase, despite the fact that the changes in chromosome morphology are not detectable in intact cells. There is indeed some evidence for interphase-like properties in repair-inhibited cells. For example, the ability of extracts from repair-inhibited mitotic cells to induce germinal vesicle breakdown and chromosome condensation when injected into Xenopus oocytes is greatly reduced compared with normal mitotic extracts (Adlakha et al. 1984). In addition, these authors found that repair-inhibited metaphase cells were unable to induce PCC when fused with interphase cells. Our results, however, do not show such a major change in PCC-inducing ability. In the majority of our experiments repairinhibited cells are as capable as control mitotic cells, though a reduced potency is suggested in one experiment by a delayed time course of PCC induction, particularly in the case of S phase nuclei (recognized as much more resistant to induction under control conditions than nuclei in G\ or Gi\ Johnson & Rao, 1970). Even in this experiment, however, repair-inhibited metaphase cells finally induce almost as much PCC as their unirradiated counterparts. We are not able to explain the variation between results of different experiments and different authors. Moreover, it is not yet clear whether any reduced potency in repair-inhibited cells is due to lessabundant or less-active mitotic inducing factors or even to the appearance of an inhibitor of these factors such as that described for G\ cells when chromosome decondensation occurs (Adlakha, Sahasrabuddhe, Wright & Rao, 1983). Whatever the state of chromosome condensation factors in repair-inhibited cells, the decondensed appearance of their chromosomes is probably not directly related to changes in levels of such factors. This is shown by our observations that fusion of repair-inhibited with normal metaphase cells does not lead to recondensation when

184

A. M. Mullinger and R. T. Johnson

inhibitors of repair are still present. The recondensation effect reported by others on the basis of similar fusion experiments (Adlakha et al. 1984) is presumably largely due to the absence of inhibitors during fusion since, as we have shown, reversal of decondensation can be produced without fusion and is accelerated by the presence of precursors of DNA synthesis in the incubation medium, whether or not the repairinhibited cells have been fused with normal metaphase cells. The nature of the changes that underly decondensation of chromosomes in metaphase cells, u.v.-irradiated and incubated with DNA synthesis inhibitors, must remain speculative. Presumably, though, since the changes are revealed by standard karyotype preparation they are associated with hypotonic swelling and, or, acid fixation. The latter step removes certain chromosomal proteins (Burkholder & Duczek, 1982). In repair-inhibited chromosomes perhaps the spectrum of proteins removed is shifted. The relationship between accumulated single-strand repair sites (and later DSBs) and changes in the nature of chromatin must also remain speculative. For example, by analogy with inhibited repair in interphase cells, we might expect nicotinamide adenine dinucleotide levels to fall (Jacobson, Antol, Juarez-Salinas & Jacobson, 1983) and, by extrapolation, poly(ADP) ribosylation of chromatin-associated proteins to rise (Shall, 1984). We believe that repair-inhibited chromosomes will provide a rich source of material to investigate, among other processes, the reversible modification of chromatin associated with excision repair. We thank Dr Andrew Collins for valuable comments on the manuscript and Roger Northfield for assistance. This work was supported by the Medical Research Council and the Cancer Research Campaign of which R.T.J. is a Research Fellow. REFERENCES ADLAKHA, R. C , SAHASRABBUDHE, C. G., WRIGHT, D. A. & RAO, P. N. (1983). Evidence for the

presence of inhibitors of mitotic factors during G\ period in mammalian cells. J. Cell Biol. 97, 1707-1713. ADLAKHA, R. C , WANG, Y. C , WRIGHT, D. A., SAHASRABUDDHE, C. G., BIGO, H. & RAO, P. N.

(1984). Inactivation of mitotic factors by ultraviolet irradiation of HeLa cells in mitosis. J. Cell Sci. 65, 279-295. AHNSTROM, G. & BRYANT, P. E. (1982). DNA double strand breaks generated by the repair of Xray damage in Chinese hamster cells. Int. J. radiat. Biol. 41, 671-676. BRADLEY, M. O. & KOHN, K. W. (1979). X-ray induced DNA double strand break production and repair in mammalian cells as measured by neutral filter elution. Nucl. Acids Res. 7, 793-804. BRADLEY, M. O. & TAYLOR, V. I. (1981). DNA double-strand breaks induced in normal human cells during the repair of ultraviolet light damage. Proc. natn. Acad. Sci. U.SA. 78, 3619-3623. BRYANT, P. E. (1983). 9-p"-D-arabinofuranosyladenine increases the frequency of X-ray induced chromosome abnormalities in mammalian cells. lnt.J. radiat. Biol. 43, 459-464. BRYANT, P. E. (1984). Effects of ara A and fresh medium on chromosome damage and DNA double-strand break repair in X-irradiated stationary cells. Br.J. Cancer 49 (Suppl. VI), 61-65. BURG, K., COLLINS, A. R. S. & JOHNSON, R. T . (1977). Effects of ultraviolet light on

synchronized Chinese hamster ovary cells; potentiation by hydroxyurea. J. Cell Sci. 28, 29-48. BURKHOLDER, G. D. & DUCZEK, L. L. (1982). The effect of chromosome banding techniques on the proteins of isolated chromosomes. Chromosoma 87, 425-435. CLARKSON, J. M. (1978). Enhancement of repair replication in mammalian cells by hydroxyurea. Mutat. Res. 52, 273-284.

Manipulating

chromosome

structure

185

COLLINS, A. R. S. (1977). DNA damage in ultraviolet-irradiated HeLa and CH0-K1 cells examined by alkaline lysis and hydroxyapatite chromatography. Biochim. biophys. Ada 478, 461-473. COLLINS, A. R. S. (1983). DNA repair in ultraviolet-irradiated HeLa cells is disrupted by aphidicolin. The inhibition of repair need not imply the absence of repair synthesis. Biochim. biophys. Ada 741, 341-347. COLLINS, A. R. S., DOWNES, C. S. & JOHNSON, R. T . (1980). Cell cycle-related variations in UV

damage and repair capacity in Chinese hamster (CHO-K1) cells. J. cell. Physiol. 103, 179-191. COLLINS, A. R. S., DOWNES, C. S. & JOHNSON, R. T. (1984). Introduction: an integrated view of

inhibited repair. In DNA Repair and its Inhibition (ed. A. Collins, C. S. Downes & R. T . Johnson), Nucleic Acids Symp. Series, no. 13, pp. 1-11. Oxford, Washington, DC: IRL Press. COLLINS, A. R. S. & JOHNSON, R. T . (1981). The use of metabolic inhibitors in repair studies. In DNA Repair. A Laboratory Manual of Research Procedures, vol. 1 (ed. E. C. Friedberg & P. C. Hanawalt), pp. 341-360. New York, Basel: Marcel Dekker, Inc. COLLINS, A. R. S., ORD, M. J. & JOHNSON, R. T . (1981). Correlations of DNA damage and repair with nuclear and chromosomal damage in HeLa cells caused by methylnitrosamides. Cancer Res. 4, 5176-5187. COLLINS, A. R. S., SCHOR, S. L. & JOHNSON, R. T . (1977). The inhibition of repair in UV

irradiated human cells. Mutal. Res. 42, 413-432. COQUERELLE, T . M. & WEIBEZAHN, K. F. (1981). Rejoining of DNA double strand breaks in human fibroblasts and its impairment in one ataxia telangiectasia and two Fanconi strains. J. supraniolec. Struct, cell. Biochem. 17, 369-376. DOWNES, C. S. (1984). Approaches to the quantitative analysis of repair through the use of inhibitors. In DNA Repair and its Inhibition (ed. A. Collins, C. S. Downes & R. T . Johnson), Nucleic Acids Symp. Series, no. 13, pp. 231-254. Oxford, Washington, DC: IRL Press. DOWNES, C. S. & COLLINS, A. R. S. (1982). Effects of DNA replication inhibitors on UV excision repair in synchronised human cells. Nucl. Acids Res. 10, 5357-5368. ELLIOTT, G. C. & JOHNSON, R. T. (1983). DNA repair in mouse embryo fibroblasts. I. Decline in ultraviolet-induced incision rate accompanies cell transformation. J . Cell Sci. 60, 267-288. FILATOV, M. V. & NOSKIN, L. A. (1983). Sensitization of human cells by inhibitors of DNA synthesis following the action of DNA-damaging agents. Mutat. Res. 110, 393-399. HANKS, J. H. & WALLACE, R. E. (1949). Relation of oxygen and temperature in the preservatio of tissues by refrigeration. Proc. Soc. exp. Biol. Med. 71, 196—200. HITTELMAN, W. R. & RAO, P. N. (1974). Premature chromosome condensation. I. Visualization of X-ray induced chromosome damage in interphase cells. Mutat. Res. 23, 251-258. JACOBSON, E. L., ANTOL, K. M., JUAREZ-SALINAS, H. & JACOBSON, M. K. (1983). Poly(ADP-

ribose) metabolism in ultraviolet irradiated human fibroblasts. J . biol. Chem. 258, 103-107. JOHNSON, R. T. & COLLINS, A, R. S. (1978). Reversal of the changes in DNA and chromosome structure which follow the inhibition of UV-induced repair in human cells. Biochem. biophys. Res. Commun. 80, 36K369. JOHNSON, R. T., COLLINS, A. R. S., DOWNES, C. S. & SQUIRES, S. (1982a). DNA-synthesis

inhibitors and the analysis of UV repair. In Progress in Mutation Research, vol. 4 (ed. A. T . Natarajan, G. Obe & H. Altmann), pp. 357-373. Amsterdam: Elsevier Biomedical Press. JOHNSON,

R. T., COLLINS, A. R. S. & WALDREN, C. A. (19826). Premature

condensed

chromosomes and the analysis of DNA and chromosome lesions. In Premature Chromosome Condensation. Application in Basic, Clinical, and Mutation Research (ed. P. N. Rao, R. T . Johnson & K. Sperling), pp. 253-308. New York, London: Academic Press. JOHNSON, R. T.', MULLINGER, A. M. & DOWNES, C. S. (1978). Human minisegregant cells. In

Methods in Cell Biology, vol. 20 (ed. D. M. Prescott), pp. 255-314. New York, London: Academic Press. JOHNSON, R. T. & RAO, P. N. (1970). Mammalian cell fusion: Induction of premature chromosome condensation in interphase nuclei. Nature, Lond. 226, 717-722. JOHNSON, R. T. & SPERLING, K. (1978). Pattern of ultra-violet-light-induced repair in metaphase and interphase chromosomes. Int. J. radial. Biol. 34, 575-582. JOHNSON, R. T., SQUIRES, S., ELLIOTT, G. C , KOCH, G. L. E. & RAINBOW, A. J. (1985). J. Cell

Sci. (in press).

186

A. M. Mullinger and R. T. Johnson

KIHLMAN, B. (1971). Molecular mechanisms of chromosome breakage and rejoining. In Adv. Cell molec. Biol., vol. 1 (ed. E. J. Du Praw), pp. 59-107. New York: Academic Press. LEWIN, B. (1974). Gene expression, vol. 2. London, New York: John Wiley & Sons. MULLINGER, A. M., COLLINS, A. R. S. & JOHNSON, R. T. (1983). Cell growth state determines susceptibility of repair DNA synthesis to inhibition by hydroxyurea and 1-3-D-arabinofuranosylcytosine. Carcinogenesis 4, 1039-1043. MULLINGER, A. M. & JOHNSON, R. T. (1980). Packing DNA into chromosomes.7. Cell Sci. 46, 61-86. NATARAJAN, A. T., OBE, G., VAN ZEELAND, A. A., PALITTI, F., MEIJERS, M. & VERDEGAAL-

IMMERZEEL, E. A. M. (1980). Molecular mechanisms involved in the production of chromosomal aberrations. II. Utilization of Neumspora endonuclease for the study of aberration production by X-rays in G l and G2 stages of the cell cycle. Mutat. Res. 69, 293-305. PAULSON, J. R. & LAEMMLI, U. K. (1977). The structure of histone-depleted metaphase chromosomes. Cell 12, 817-828. RAO, P. N. &JOHNSON, R. T. (1972). Cell fusion and its application to studies on the regulation of the cell cycle. In Methods in Cell Physiology, vol. 5 (ed. D. M. Prescott), pp. 75-126. New York, London: Academic Press. RYDBERG, B. (1983). Can DNA double strand breaks in mammalian cells be repaired without help from a homologous DNA molecule? Proc. 7th Internal. Congress of Radiat. Res. B2-31. Amsterdam: Martinus Nijhoff. SCHOR, S. L., JOHNSON, R. T . & WALDREN, C. A. (1975). Changes in the organization of

chromosomes during the cell cycle: response to ultraviolet light. J . Cell Sci. 17, 539-555. SHALL, S. (1984). Inhibition of DNA repair by inhibitors of nuclear ADP-ribosyl transferase. In DNA Repair and its Inhibition (ed. A. Collins, C. S. Downes & R. T . Johnson), Nucleic Acids Symp. Series, no. 13, pp. 143-191. Oxford, Washington, DC; IRL Press. SQUIRES, S., JOHNSON, R. T. & COLLINS, A. R. S. (1982). Initial rates of DNA incision in UV-

irradiated human cells. Differences between normal, xeroderma pigmentosum and tumour cells.

Mutat. Res. 95, 389-404. WALDREN, C. A. & JOHNSON, R. T . (1974). Analysis of interphase chromosome damage by means of premature chromosome condensation after X- and ultraviolet-irradiation. Proc. natn. Acad.

Sci. U.SA. 71, 1137-1141.

(Received 2 July 1984 - Accepted 20 July 1984)