- Email: [email protected]
An improved method of alkaline sucrose density gradient sedimentation to detect less than one lesion per 1 Mb DNA
DNA Repair
ELSEVIER
Mutation Research 364 (1996) 125-131
An improved method of alkaline sucrose density gradient sedimentation to detect less than one lesion per 1 Mb DNA Kouichi Yamada *, Yumi Kameyama, Shuji Inoue Division of Geriatric Health Science, The National Institute of Health and Nutrition, Shinjuku-ku, Received 22 December
1995; revised 24 May 1996; accepted
Tokyo 162, Japan
14 June 1996
Abstract We improved
alkaline
sucrose
density
gradient
sedimentation
to detect
very
long
single-strand
DNA at the megabase
4 Mb). Hitherto, these have not sedimented correctly due to some artifacts. One artifact was aggregation of sticky DNA and proteins formed in the gradient. Then, in some gradients, biphasic distribution was observed, level (from
less than
1 to about
the major peak of which was reasonable as a result of random scission by X-rays, but the minor, fast-sedimenting population was another artifact resulting from incomplete denaturation of the DNA. We mainly reduced the centrifugal force and used a solution for cell lysis with a high concentration of salt. By means of this procedure, DNA single-strand breaks induced by relatively low doses of X-rays and subsequent repair processes can be measured in human fibroblasts. The protocol is also applicable to the study of DNA damage accompanied by strand scission, such as by UV or dimethyl sulfate as well as their repair. The technique is sensitive enough to detect even single-strand breaks induced by 0.1 J/m’ UV and sufficiently reproducible that breaks induced by increasing UV dosages were dose dependent. Thus, this technique was proven to be very sensitive, reliable and simple to perform. Therefore, this improvement will be extremely useful to investigators studying DNA repair. Keywords: X-ray: Ultraviolet; DNA repair
Dimethyl
sulfate: Single-strand
break; Alkaline
1. Introduction DNA strand scission caused by DNA-damaging agents and the subsequent repair processes in cells are usually detected by several methods, including
Abbreviations: ASDG. alkaline sucrose density gradient; CMF-PBS, Ca’+,Mg ‘+-free phosphate-buffered saline: DMS, dimethyl sulfate; ssDNA, single-stranded DNA: SSB, single-strand breaks; UV, ultraviolet * Corresponding author. Tel.: (8 1) (3) 3203-5723; Fax: (8 1) (3) 3203-0335. 0921-8777/96/$15.00 Copyright PI1 SO921-8777(96)00033-X
0 1996. Published
sucrose density gradient
centrifugation;
Human fibroblast;
alkaline elution (for review see Kohn et al., 1981) and alkaline sucrose density gradient (ASDG) sedimentation (for review see Lett (1981)). Alkaline elution is far more sensitive than ASDG, as it can detect even single-strand breaks (SSB) caused by 0.5 or 1 Gy X-rays (Kohn et al., 1981). However, the method has some shortcomings; it is not quantitative, the reproducibility is poor and it requires special skills and apparatus. On the other hand, ASDG needs only an ultracentrifuge and a rotor. For ASDG centrifugation, the cells were originally lysed in tubes and layered onto the gradient by,
by Elsevier Sciencee B.V.
K. Yamadaet
126
d/Mutation
for example, a capillary or pipette tip. As a result, the DNA became smaller than about 0.3 Mb owing to shearing forces at work during mixing and layering. McGrath and Williams (1966) reported an improved method, in which the cells were applied directly onto the gradient. Since then, longer singlestranded DNA (ssDNA) fragments have been detectable (Lett et al., 1967). However, in this case, a maximum of 0.6 Mb was still the limit, because longer DNA was not sedimented correctly, mainly due to two artifacts. First, sticky DNA and proteins often aggregated in the gradient (Lett. 1981). In the other case, DNA from both unirradiated cells and cells irradiated with low doses of X-rays sometimes sedimented in a biphasic profile (Dugle and Gillespie, 1975). The larger, slow-sedimenting peak was the result of random breakage by X-rays, but the smaller, quickly sedimenting population was a second artifact due to incomplete strand separation of DNA from loaded cells. We have improved the method and have found the optimum conditions, in which these artifacts are hardly formed and ssDNA fragments as long as 4 Mb can be detected.
2. Materials
and methods
2.1. Cell culture Normal human fibroblasts, NB 1RGB (RIKEN Cell Bank, Tsukuba, Japan) were maintained in monolayers in Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum. 2.2. X-ray irradiation,
UV irradiation
and exposure
to dimethyl sulfate
Confluent cells were trypsinized and seeded into culture dishes (1 X lo5 tells/60-mm-diameter dish). After 3 days of culture, the cells were labeled with 0.2 t&i/ml [methyl-‘4C]thymidine (Amersham, 57 mCi/mmol) for 24 h, then chased in non-radioactive medium overnight. For X-ray irradiation, the medium was removed, then the dishes were floated on ice-water and irradiated using a soft X-ray generator OM-150R (Ohmic, Tokyo) at 3.6 Gy/min (100 kV, 10 mA, with 0.2 mm Al filter). The dose rate was checked using a electrometer model 500 (Victoreen, Ohio).
Research364 (19%) 125-131
For UV irradiation, the medium was removed and the cells were exposed to UV light from a germicidal lamp (Toshiba GL15) at 0.05 or 0.5 J/m’/s. The dose rate was checked using a DRC-100X Digital Radiometer (Spectronics, Westbury, NY) In UV experiments, 3 mM hydroxyurea was always added 30 min before UV irradiation. And immediately after the irradiation, medium containing 10 (*g/ml aphidicolin and 3 mM hydroxyurea was added in order to inhibit repair synthesis (Yamada et al., 1985; Yamada and Itoh, 1994). Dimethyl sulfate (DMS) was dissolved in Ca”,Mg*+ -free phosphate-buffered saline (CMFPBS) to a concentration of 10 mM, and added to the medium. 2.3. Alkaline sucrose density gradient centrifugation The cells containing uniformly prelabeled DNA, were exposed to DNA-damaging agents, harvested with trypsin, then layered on the ASDG as follows. First, 100 ~1 of cell-lysis solution (0.6 M KOH, 2.0 M KCl, 10 mM EDTA, and 1% N-lauroylsarcosine), then 50 pl of 1% sucrose in CMF-PBS, followed by the cell suspension (about 4 X lo4 cells in 50 ~1 of CMF-PBS) were layered in series on top of a preformed 4.35 ml alkaline 5-20% sucrose gradient (0.3 M KOH, 2.0 M KCl, 1 mM EDTA, and 0.1% N-lauroylsarcosine) with 0.4 ml of alkaline 80% sucrose as a cushion at the bottom. The gradient was centrifuged at 6000 rpm (4320 X g) for 15.3 h at 15°C in a Beckman SW50.1 rotor. (A slow accelerator is preferable.) The gradient was fractionated from the bottom by drop-counting onto 3MM paper circles (Whatman) using a peristalic pump. The paper circles were dried, immersed in cold 5% trichloroacetic acid, washed 3 times with ethanol and once with acetone, dried, then the radioactivity was measured in a toluene-based scintillator. The percentage of the total cpm in each fraction was calculated taking the input cpm as 100%. The recovery of the radioactivity was 85-100%. These experiments were repeated at least twice, and essentially the same profiles were obtained. 2.4. High molecular weight markers Escherichia
medium,
coli (JMl09)
labeled
with
was grown in 2 X TY 0.2 pCi/ml [methyl-
K. Yatnada et al./Mutation
“C]thymidine (Amersham) for 24 h. Saccharomyces cererlisiae (Sigma) was cultured in YPD medium at 30°C with 0.1 pCi/ml [8-‘4C]deoxyadenosine (NEN, 57.4 mCi/mmol) for 24 h. then chased in non-radioactive medium overnight. E. coli (Kl2) was infected with T4 phage in A-medium (0.1 M K-PO,, pH 7.3. 20 mM (NH,),SO,. 1 mM MgSO,, 2 pM Fe (NH,),@O,),, 5% glucose, 5 pg/ml thymine) with 20 p.g/ml L-Trp, then cultured for 2-3 h with 50 FM [methyl‘“Clthymine (1 mCi/mmol) (Moravek). The E. coli was lysed with chloroform, then T4 phage particles were purified by digestion with DNase and RNase followed by centrifugation (5000 X g for 10 min to
1
121
Research 364 (19961 I?_%131
i FRACTION
NtiMBEfi
Fig. 2. ASDG profiles of DNA in X-ray-irradiated NBlRGB cells (X-ray dose-response). NBlRGB cells were labeled with [‘JC]thymidine and irradiated at 0 CO), 2.5 (0). 5.0 (0 ). 10 (O), 20 ( + ) and 40 ( X ) Gy X-rays. These cells were analyzed by ASDC centrifugation as described in Materials and methods.
remove the E. coli pellet and 35 000 X g for 25 min to sediment the T4 phage particles) (Hirose, 1973). These cells or particles were directly placed onto the cell-lysis layer.
3. Results 3.1. Mod$cations introduced in order to improoe alkaline sucrose dens&v gradient centrifugation
Fig. 1. ASDG profiles of high MW DNA markers, E. coli (column B), S. cerevisiae (column C) and phage T4 (column D). These whole cells or particles were directly placed onto the cell-lysis layer, of which DNAs were labeled in viva as described in Materials and methods. Column A is sucrose density and the volume of each fraction after centrifugation.
To avoid artificial aggregation of the DNA and protein in the gradient, we tested some combinations of salt and detergent in the cell-lysis solution. SDS is a powerful detergent and soluble in high concentrations of NaCl and NaOH. It was not suitable, however, because the drop size decreased markedly near the top fractions during fractionation. Triton X-100 made the salts insoluble in the lysis solution. NLauroylsarcosine is soluble in 0.6 M KOH and 2.0 M KCl, does not decrease the drop size (Fig. 1) and allows higher MW DNA to sediment without the aggregation (Figs. 2-5). (2.0 M KC1 prevented this artifact better than 1.4 M KCI.) The conditions of centrifugation were also checked to reduce the shearing force as much as possible. We used [ “C]thymidine, not [ 3H]thymidine, because [3H]thymidine inhibits cell growth (Pollack et
128
K. Yamda
et nl./Mutation
Research 364 (1996) 125-131
al., 1979). Consequently, it was difficult to label the DNA to the required specific activity (> 3000 cpm/4 X lo4 cells) in NBlRGB cells. The tritium decay of the incorporated thymidine also causes SSB in DNA (Cleaver et al., 1972). Previously we had used [3H]thymidine to prelabel the DNA in HeLa cells (Yamada et al., 1985). In those ASDG profiles of DNA from non-damaged cells, there was a broad distribution of counts in the middle of the gradient. Using [ “C]thymidine, they disappeared (Figs. 2-5). To measure the size of DNA fragments sedimented in this ASDG, high MW DNA markers were applied onto the gradients (Fig. 1). Labeled E. coli DNA (about 4 Mb), 5. cerer!isiae DNA (the chromo-
FRACTION
20
I
5
0 10
= a 0 --I Q c 0 c
6
LE
a
"
15
J 4 c z
10
LL 0
5
b:
5
0 0 (BOTTOM)
5
10 FRACTION
LL 0 w 0 4 + z w u uz w a
NUMBER
30
Fig. 4. ASDG profiles of DNA in UV-irradiated NBIRGB cells (UV dose-response). NBlRGB cells were labeled with [“Clthymidine, incubated at 37°C for 30 min in medium containing 3 mM hydroxyurea, irradiated at (A) 0 (0). 0.1 (0). 0.3 (A ), 0.6(o), l.O(OJor(BJ 1.3(aJ,2.5(r>J, 5.0(X), lO(+JJ/m’ UV, then incubated for 1 h in medium containing 3 mM hydroxyurea and 10 kg/ml aphidicolin. These cells were analyzed by ASDG centrifugation as described in Materials and methods.
0 5
0 5
0
somes range from 0.25 to 2.2 Mb) and T4 phage DNA (0.166 Mb) sedimented near the bottom, near the top and in the top fractions, respectively.
5
0 (BOTTOM)
20 NUMBER
10
FRACTION
20
31
NUMBER
Fig. 3. ASDG profiles of DNA in X-ray-irradiated NB 1RGB cells (time course for rejoining of SSBJ. NBlRGB cells were labeled with [I4Clthymidine. Non-irradiated cells (column A). Twenty Gy-irradiated cells (column B) were incubated at 37°C for 15 min (column CJ, 30 min (column DJ, 2 h (column EJ or 4 h (column FJ. These cells were analyzed by ASDG centrifugation as described in Materials and methods.
3.2. Analysis scission
of X-ray-induced
DNA single-strand
NBlRGB cells were irradiated with 2.5-40 Gy of X-rays and SSB in the cells were analyzed by our modified procedure (Fig. 2). DNA in non-irradiated cells sedimented in the bottom fractions. X-rays (2.5 Gy) caused single-strand breaks, which appeared as a
K. Yamada et al. /Mutation Research 364 (1996) 125-131
129
broad mound in the middle fractions of the gradient. As the X-ray dose increased, the peak became larger and steeper and the position shifted towards the top of the gradient. X-rays (20 Gy)-irradiated NB 1RGB cells were incubated at 37°C for 15-240 min, and the SSB gradually disappeared (Fig. 3). A steep peak, which was seen in the cells immediately after the irradiation, became broader within 15 min and thereafter gradually shifted towards the bottom. Four hours later, the SSB were almost rejoined and the profile became similar to that of non-irradiated cells.
DMS, an alkylating reagent, caused accumulation of SSB in the absence of these DNA synthesis inhibitors (Fig. 5). The DNA in 25 PM DMS-treated cells was widely distributed in the middle fractions, SSB were markedly accumulated by 50 PM DMS, which resembled the profile of the 10 Gy of X-rayirradiated cells shown in Fig. 2. However, the sizes of the resulting ssDNA fragments in the former were somewhat larger than the latter.
3.3. Analysis sulfate
An improved method of ASDG centrifugation was presented for measuring high MW (l-4 Mb) ssDNA fragments, which were difficult to detect until now (Fig. 1). Using this procedure, we showed the profiles of DNA scission by X-rays at a physiological dosage and the following rejoining (Fig. 2 and Fig. 3). which could hardly be detected in reasonable manner so far. ASDG was modified mainly by reducing the centrifugal force and by using a cell-lysis solution with a high salt concentration. Another improvement was that the cells were chilled completely during X-ray irradiation, because rejoining of X-ray-induced SSB is very fast, according to our results (Fig. 3) and those of others (Fornace and Little, 1980). In fact, SSB caused by lower dosages of X-rays are ligated within a few minutes (data not shown). Using our procedure described here, SSB can be detected in a highly dose-dependent manner (Fig. 2 and Fig. 4). DNA from undamaged cells sedimented to the bottom of the gradient, and that from cells treated with damaging agents sedimented in the middle fractions. As shown in Fig. 4A, when cells were exposed to low doses of UV, peak area of accumulated SSB was nearly in proportion to the UV dose. Then, the higher doses of UV made the peak steeper (Fig. 4B). The similar tendency was shown in Xray-produced SSB (Fig. 2). To define the reliability of this method, the sizes of the ssDNA fragments produced by X-rays were calculated from the above results. From the position of the T4 DNA marker, the size of the 20 Gy-induced ssDNA fragments was estimated at 1.7 Mb (most frequent value). To compare our calculations with those of others, we extrapolated from the data
of SSB caused
by UV or dimethyl
NB 1RGB cells were exposed to 0.1- 10 J/m’ of UV, and incubated at 37°C in medium containing 3 mM hydroxyurea and 10 kg/ml aphidicolin for 1 h (Fig. 4 in Da>. SSB did not accumulate in DNA from hydroxyurea and aphidicolin-treated cells that were not UV irradiated, indicating these reagents alone do not cause SSB. In their presence, a few SSB induced by only 0.1 J/m’ UV were detectable. Increasing the UV dose resulted in the dose-dependent accumulation of SSB. Finally SSB levels induced by 5 or 10 J/m2 UV slowly reached a plateau.
-
(BOTTOM)
1lJ
FRACTfON
20
30
NUMBER
Fig. 5. ASDG profiles of DNA in DMS-treated NBlRGB cells (DMS dose-response). NBIRGB cells were labeled with [‘4C]thymidine and incubated at 37°C for 20 min in medium containing 0 CO), 25 (0) or 50 (+) pM DMS. These cells were analyzed by ASDG centrifugation as described in Materials and methods.
4. Discussion
130
K. Yamada et al./Muiation Research 364 (19%) 125-131
of Lett et al. (1967) that 20 Gy X-rays induce about one DNA strand scission per 5 X lo* Da (about 1.5 Mb) (their results did not include doses under 100 Gy due to the problem described in Introduction), which is in good agreement with our calculations. We also estimated the amount of energy necessary to cause one breakage from the X-ray dose and the size of ssDNA fragments. First, 20 Gy = 1.25 X lOi eV/g. And, this value was divided by N,/MW, for [ NA/MW] molecules of ssDNA are contained in 1 g of DNA, leading to the amount of energy given to one DNA strand. Here, NA is Avogadro’s number and MW is the molecular weight of intact DNA (in daltons). The resulting value was divided by the number of SSB, which is MW/(1.7 X 3.3 X lo’), leading to about 116 eV per break. This estimation agrees with that of Cleaver et al. (1972) who measured SSB in frozen cells, though the value is larger than one scission per 44-66 eV reported by others (Dean et al., 1969; Lehmann and Ormerod, 1970). Using this technique, we also showed the profiles of SSB accumulation in UV-exposed cells in the presence of hydroxyurea and aphidicolin (Fig. 4). The combination of these reagents inhibits UV-induced repair-synthesis almost completely. In their absence, breaks are not detected (Yamada et al., 1985; Yamada and Itoh, 1994) probably because the incision stages of repair occur slowly and sporadically, whereas polymerization and ligation proceed rapidly. Our method was proven to be extremely sensitive, because SSB induced by 0.1 J/m’ UV were detected. Furthermore. when the cells were exposed to a higher concentration of aphidicolin (for example, 30 pg/ml) together with 3 mM hydroxyurea for a longer incubation (for example, 8 h), even the SSB induced by 0.03 J/m’ UV were measurable (data not shown). We also applied this protocol to DMS-induced DNA strand breaks (Fig. 5). DMS methylates DNA bases, mainly G and A, which causes the release of bases enzymatically and non-enzymatically in cells. DNA strands are broken at the resulting apurinic/apyrimidinic sites in alkaline solution by P-elimination. Such lesions are also counted in ASDG. Our improved ASDG can be applied to study all DNA damage accompanied by strand breaks at very low levels. Here we only used NB 1RGB cells, but the same
profiles were also obtained from HeLa cells. This method is also applicable to cells cultured in suspension, such as FM3A (mouse mammary carcinoma cells) or lymphocytes, which can be labeled after phytohemagglutinin stimulation or after transformation in lymphoblasts. In addition, our procedure of ASDG is sensitive enough to detect the small number of SSB that occur in UV-repair-deficient cells, such as xeroderma pigmentosum cells. Thus, our improved ASDG centrifugation is reproducible and highly sensitive. Moreover it is easy to perform, because the protocol requires no special skills or apparatus. Accordingly the improved method will be highly valuable to studies of DNA repair. Recently, in order to observe SSB in cells, single-cell microgel electrophoresis assay (comet assay) has been examined, and it provides new possibilities of detection (for review see McKelvey-Martin et al., 1993). Various alkali unwinding techniques are also widely used. In addition to these, ASDG sedimentation will be frequently used by our modification.
Acknowledgements This work was supported by Grant for Nuclear Research from Science and Technology Agency, Japan.
References Cleaver, J.E.. G.H. Thomas and H.J. Burki (1972) Biological damage from intranuclear tritium; DNA strand breaks and their repair. Science. 177, 996-998. Dean, C.J., M.G. Ormerod, R.W. Serianni and P. Alexander (1969) DNA strand breakage in cells irradiated with X-rays. Nature, 222, 1042- 1044. Dugle, D.L. and C.J. Gillespie (1975) Kinetics of the single-strand repair mechanism in mammalian cells, in: P.C. Hanawalt and R.B. Setlow (Eds.), Molecular Mechanisms for Repair of DNA, Part A, Plenum Press, New York, pp. 685-687. Fornace. A.J., Jr. and J.B. Little (1980) Normal repair of DNA single-strand breaks in patients with ataxia telangiectasia. Biochim. Biophys. Acta, 607, 432-437. Hirose, S. (1973) Preparation of bacterial and phage DNA. in: A. Ishihama, R. Okazaki, Y. Kyogoku and S. Nishimura (Eds.) supplement of Protein, Nucleic Acids and Enzyme. Method in Nucleic Acid Research, Kyoritu Shuppan. Tokyo. pp. 4-8. Kohn. K.W.. R.A.G. Ewig, L.C. Erickson and L.A. Zwelling
K. Yamada et al/Mutation
(198 11 Measurement
of strand breaks and cross-links by alkaline elution. in: E.C. Friedberg and P.C. Hanawalt (Eds.), DNA repair - A Laboratory Manual of Research Procedures, Vol. 1, Marcel Dekker, New York, pp. 379-401. Lehmann, A.R. and M.G. Ormerod (19701 The replication of DNA in murine lymphoma cells (L5178Y1, Biochim. Biophys. Acta, 204, 128-143. Lett, J.T. (198 I ) Measurement of single strand breaks by sedimentation in alkaline sucrose gradients, in: E.C. Friedberg and P.C. Hanawalt (Eds.1, DNA repair - A Laboratory Manual of Research Procedures, Voi. 1, Marcel Dekker. New York, pp. 363-378. Lett, J.T.. I. Caldwell. C.J. Dean and P. Alexander (1967) Rejoining of X-ray induced breaks in the DNA of leukaemia cells. Nature, 214, 790-792. McCrath, R.A. and R.W. Williams (19661 Reconstitution in vivo
Research 364 (1996) 125-131
131
of irradiated Escherichia coli deoxyribonucleic acid; the rejoining of broken pieces, Nature, 2 12. 534-535. McKelvey-Martin, V.J., M.H.L. Green, P. Schmezer. B.L. PoolZobel, M.P. De Meo and A.R. Collins (19931 The single cell gel electrophoresis assay (comet assay): a European review. Mutation Res.. 288. 47-63. Pollack. A., C.B. Bagwell and G.L. Irvin, III (19791 Radiation from tritiated thymidine perturbs the cell cycle progression of stimulated lymphocytes, Science, 203. 1025-1027. Yamada, K., F. Hanaoka and M. Yamada (19851 Effects of aphidicolin and/or 2’3.dideoxythymidine on DNA repair induced in HeLa cells by four types of DNA-damaging agents, J. Biol. Chem., 260, 10412-10417. Yamada. K. and R. Itoh (19941 Involvement of DNA polymerase 6 and/or E in joining UV-induced DNA single strand breaks in human fibroblasts, Biochim. Biophys. Acta. 1219, 302-306.
ELSEVIER
Mutation Research 364 (1996) 125-131
An improved method of alkaline sucrose density gradient sedimentation to detect less than one lesion per 1 Mb DNA Kouichi Yamada *, Yumi Kameyama, Shuji Inoue Division of Geriatric Health Science, The National Institute of Health and Nutrition, Shinjuku-ku, Received 22 December
1995; revised 24 May 1996; accepted
Tokyo 162, Japan
14 June 1996
Abstract We improved
alkaline
sucrose
density
gradient
sedimentation
to detect
very
long
single-strand
DNA at the megabase
4 Mb). Hitherto, these have not sedimented correctly due to some artifacts. One artifact was aggregation of sticky DNA and proteins formed in the gradient. Then, in some gradients, biphasic distribution was observed, level (from
less than
1 to about
the major peak of which was reasonable as a result of random scission by X-rays, but the minor, fast-sedimenting population was another artifact resulting from incomplete denaturation of the DNA. We mainly reduced the centrifugal force and used a solution for cell lysis with a high concentration of salt. By means of this procedure, DNA single-strand breaks induced by relatively low doses of X-rays and subsequent repair processes can be measured in human fibroblasts. The protocol is also applicable to the study of DNA damage accompanied by strand scission, such as by UV or dimethyl sulfate as well as their repair. The technique is sensitive enough to detect even single-strand breaks induced by 0.1 J/m’ UV and sufficiently reproducible that breaks induced by increasing UV dosages were dose dependent. Thus, this technique was proven to be very sensitive, reliable and simple to perform. Therefore, this improvement will be extremely useful to investigators studying DNA repair. Keywords: X-ray: Ultraviolet; DNA repair
Dimethyl
sulfate: Single-strand
break; Alkaline
1. Introduction DNA strand scission caused by DNA-damaging agents and the subsequent repair processes in cells are usually detected by several methods, including
Abbreviations: ASDG. alkaline sucrose density gradient; CMF-PBS, Ca’+,Mg ‘+-free phosphate-buffered saline: DMS, dimethyl sulfate; ssDNA, single-stranded DNA: SSB, single-strand breaks; UV, ultraviolet * Corresponding author. Tel.: (8 1) (3) 3203-5723; Fax: (8 1) (3) 3203-0335. 0921-8777/96/$15.00 Copyright PI1 SO921-8777(96)00033-X
0 1996. Published
sucrose density gradient
centrifugation;
Human fibroblast;
alkaline elution (for review see Kohn et al., 1981) and alkaline sucrose density gradient (ASDG) sedimentation (for review see Lett (1981)). Alkaline elution is far more sensitive than ASDG, as it can detect even single-strand breaks (SSB) caused by 0.5 or 1 Gy X-rays (Kohn et al., 1981). However, the method has some shortcomings; it is not quantitative, the reproducibility is poor and it requires special skills and apparatus. On the other hand, ASDG needs only an ultracentrifuge and a rotor. For ASDG centrifugation, the cells were originally lysed in tubes and layered onto the gradient by,
by Elsevier Sciencee B.V.
K. Yamadaet
126
d/Mutation
for example, a capillary or pipette tip. As a result, the DNA became smaller than about 0.3 Mb owing to shearing forces at work during mixing and layering. McGrath and Williams (1966) reported an improved method, in which the cells were applied directly onto the gradient. Since then, longer singlestranded DNA (ssDNA) fragments have been detectable (Lett et al., 1967). However, in this case, a maximum of 0.6 Mb was still the limit, because longer DNA was not sedimented correctly, mainly due to two artifacts. First, sticky DNA and proteins often aggregated in the gradient (Lett. 1981). In the other case, DNA from both unirradiated cells and cells irradiated with low doses of X-rays sometimes sedimented in a biphasic profile (Dugle and Gillespie, 1975). The larger, slow-sedimenting peak was the result of random breakage by X-rays, but the smaller, quickly sedimenting population was a second artifact due to incomplete strand separation of DNA from loaded cells. We have improved the method and have found the optimum conditions, in which these artifacts are hardly formed and ssDNA fragments as long as 4 Mb can be detected.
2. Materials
and methods
2.1. Cell culture Normal human fibroblasts, NB 1RGB (RIKEN Cell Bank, Tsukuba, Japan) were maintained in monolayers in Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum. 2.2. X-ray irradiation,
UV irradiation
and exposure
to dimethyl sulfate
Confluent cells were trypsinized and seeded into culture dishes (1 X lo5 tells/60-mm-diameter dish). After 3 days of culture, the cells were labeled with 0.2 t&i/ml [methyl-‘4C]thymidine (Amersham, 57 mCi/mmol) for 24 h, then chased in non-radioactive medium overnight. For X-ray irradiation, the medium was removed, then the dishes were floated on ice-water and irradiated using a soft X-ray generator OM-150R (Ohmic, Tokyo) at 3.6 Gy/min (100 kV, 10 mA, with 0.2 mm Al filter). The dose rate was checked using a electrometer model 500 (Victoreen, Ohio).
Research364 (19%) 125-131
For UV irradiation, the medium was removed and the cells were exposed to UV light from a germicidal lamp (Toshiba GL15) at 0.05 or 0.5 J/m’/s. The dose rate was checked using a DRC-100X Digital Radiometer (Spectronics, Westbury, NY) In UV experiments, 3 mM hydroxyurea was always added 30 min before UV irradiation. And immediately after the irradiation, medium containing 10 (*g/ml aphidicolin and 3 mM hydroxyurea was added in order to inhibit repair synthesis (Yamada et al., 1985; Yamada and Itoh, 1994). Dimethyl sulfate (DMS) was dissolved in Ca”,Mg*+ -free phosphate-buffered saline (CMFPBS) to a concentration of 10 mM, and added to the medium. 2.3. Alkaline sucrose density gradient centrifugation The cells containing uniformly prelabeled DNA, were exposed to DNA-damaging agents, harvested with trypsin, then layered on the ASDG as follows. First, 100 ~1 of cell-lysis solution (0.6 M KOH, 2.0 M KCl, 10 mM EDTA, and 1% N-lauroylsarcosine), then 50 pl of 1% sucrose in CMF-PBS, followed by the cell suspension (about 4 X lo4 cells in 50 ~1 of CMF-PBS) were layered in series on top of a preformed 4.35 ml alkaline 5-20% sucrose gradient (0.3 M KOH, 2.0 M KCl, 1 mM EDTA, and 0.1% N-lauroylsarcosine) with 0.4 ml of alkaline 80% sucrose as a cushion at the bottom. The gradient was centrifuged at 6000 rpm (4320 X g) for 15.3 h at 15°C in a Beckman SW50.1 rotor. (A slow accelerator is preferable.) The gradient was fractionated from the bottom by drop-counting onto 3MM paper circles (Whatman) using a peristalic pump. The paper circles were dried, immersed in cold 5% trichloroacetic acid, washed 3 times with ethanol and once with acetone, dried, then the radioactivity was measured in a toluene-based scintillator. The percentage of the total cpm in each fraction was calculated taking the input cpm as 100%. The recovery of the radioactivity was 85-100%. These experiments were repeated at least twice, and essentially the same profiles were obtained. 2.4. High molecular weight markers Escherichia
medium,
coli (JMl09)
labeled
with
was grown in 2 X TY 0.2 pCi/ml [methyl-
K. Yatnada et al./Mutation
“C]thymidine (Amersham) for 24 h. Saccharomyces cererlisiae (Sigma) was cultured in YPD medium at 30°C with 0.1 pCi/ml [8-‘4C]deoxyadenosine (NEN, 57.4 mCi/mmol) for 24 h. then chased in non-radioactive medium overnight. E. coli (Kl2) was infected with T4 phage in A-medium (0.1 M K-PO,, pH 7.3. 20 mM (NH,),SO,. 1 mM MgSO,, 2 pM Fe (NH,),@O,),, 5% glucose, 5 pg/ml thymine) with 20 p.g/ml L-Trp, then cultured for 2-3 h with 50 FM [methyl‘“Clthymine (1 mCi/mmol) (Moravek). The E. coli was lysed with chloroform, then T4 phage particles were purified by digestion with DNase and RNase followed by centrifugation (5000 X g for 10 min to
1
121
Research 364 (19961 I?_%131
i FRACTION
NtiMBEfi
Fig. 2. ASDG profiles of DNA in X-ray-irradiated NBlRGB cells (X-ray dose-response). NBlRGB cells were labeled with [‘JC]thymidine and irradiated at 0 CO), 2.5 (0). 5.0 (0 ). 10 (O), 20 ( + ) and 40 ( X ) Gy X-rays. These cells were analyzed by ASDC centrifugation as described in Materials and methods.
remove the E. coli pellet and 35 000 X g for 25 min to sediment the T4 phage particles) (Hirose, 1973). These cells or particles were directly placed onto the cell-lysis layer.
3. Results 3.1. Mod$cations introduced in order to improoe alkaline sucrose dens&v gradient centrifugation
Fig. 1. ASDG profiles of high MW DNA markers, E. coli (column B), S. cerevisiae (column C) and phage T4 (column D). These whole cells or particles were directly placed onto the cell-lysis layer, of which DNAs were labeled in viva as described in Materials and methods. Column A is sucrose density and the volume of each fraction after centrifugation.
To avoid artificial aggregation of the DNA and protein in the gradient, we tested some combinations of salt and detergent in the cell-lysis solution. SDS is a powerful detergent and soluble in high concentrations of NaCl and NaOH. It was not suitable, however, because the drop size decreased markedly near the top fractions during fractionation. Triton X-100 made the salts insoluble in the lysis solution. NLauroylsarcosine is soluble in 0.6 M KOH and 2.0 M KCl, does not decrease the drop size (Fig. 1) and allows higher MW DNA to sediment without the aggregation (Figs. 2-5). (2.0 M KC1 prevented this artifact better than 1.4 M KCI.) The conditions of centrifugation were also checked to reduce the shearing force as much as possible. We used [ “C]thymidine, not [ 3H]thymidine, because [3H]thymidine inhibits cell growth (Pollack et
128
K. Yamda
et nl./Mutation
Research 364 (1996) 125-131
al., 1979). Consequently, it was difficult to label the DNA to the required specific activity (> 3000 cpm/4 X lo4 cells) in NBlRGB cells. The tritium decay of the incorporated thymidine also causes SSB in DNA (Cleaver et al., 1972). Previously we had used [3H]thymidine to prelabel the DNA in HeLa cells (Yamada et al., 1985). In those ASDG profiles of DNA from non-damaged cells, there was a broad distribution of counts in the middle of the gradient. Using [ “C]thymidine, they disappeared (Figs. 2-5). To measure the size of DNA fragments sedimented in this ASDG, high MW DNA markers were applied onto the gradients (Fig. 1). Labeled E. coli DNA (about 4 Mb), 5. cerer!isiae DNA (the chromo-
FRACTION
20
I
5
0 10
= a 0 --I Q c 0 c
6
LE
a
"
15
J 4 c z
10
LL 0
5
b:
5
0 0 (BOTTOM)
5
10 FRACTION
LL 0 w 0 4 + z w u uz w a
NUMBER
30
Fig. 4. ASDG profiles of DNA in UV-irradiated NBIRGB cells (UV dose-response). NBlRGB cells were labeled with [“Clthymidine, incubated at 37°C for 30 min in medium containing 3 mM hydroxyurea, irradiated at (A) 0 (0). 0.1 (0). 0.3 (A ), 0.6(o), l.O(OJor(BJ 1.3(aJ,2.5(r>J, 5.0(X), lO(+JJ/m’ UV, then incubated for 1 h in medium containing 3 mM hydroxyurea and 10 kg/ml aphidicolin. These cells were analyzed by ASDG centrifugation as described in Materials and methods.
0 5
0 5
0
somes range from 0.25 to 2.2 Mb) and T4 phage DNA (0.166 Mb) sedimented near the bottom, near the top and in the top fractions, respectively.
5
0 (BOTTOM)
20 NUMBER
10
FRACTION
20
31
NUMBER
Fig. 3. ASDG profiles of DNA in X-ray-irradiated NB 1RGB cells (time course for rejoining of SSBJ. NBlRGB cells were labeled with [I4Clthymidine. Non-irradiated cells (column A). Twenty Gy-irradiated cells (column B) were incubated at 37°C for 15 min (column CJ, 30 min (column DJ, 2 h (column EJ or 4 h (column FJ. These cells were analyzed by ASDG centrifugation as described in Materials and methods.
3.2. Analysis scission
of X-ray-induced
DNA single-strand
NBlRGB cells were irradiated with 2.5-40 Gy of X-rays and SSB in the cells were analyzed by our modified procedure (Fig. 2). DNA in non-irradiated cells sedimented in the bottom fractions. X-rays (2.5 Gy) caused single-strand breaks, which appeared as a
K. Yamada et al. /Mutation Research 364 (1996) 125-131
129
broad mound in the middle fractions of the gradient. As the X-ray dose increased, the peak became larger and steeper and the position shifted towards the top of the gradient. X-rays (20 Gy)-irradiated NB 1RGB cells were incubated at 37°C for 15-240 min, and the SSB gradually disappeared (Fig. 3). A steep peak, which was seen in the cells immediately after the irradiation, became broader within 15 min and thereafter gradually shifted towards the bottom. Four hours later, the SSB were almost rejoined and the profile became similar to that of non-irradiated cells.
DMS, an alkylating reagent, caused accumulation of SSB in the absence of these DNA synthesis inhibitors (Fig. 5). The DNA in 25 PM DMS-treated cells was widely distributed in the middle fractions, SSB were markedly accumulated by 50 PM DMS, which resembled the profile of the 10 Gy of X-rayirradiated cells shown in Fig. 2. However, the sizes of the resulting ssDNA fragments in the former were somewhat larger than the latter.
3.3. Analysis sulfate
An improved method of ASDG centrifugation was presented for measuring high MW (l-4 Mb) ssDNA fragments, which were difficult to detect until now (Fig. 1). Using this procedure, we showed the profiles of DNA scission by X-rays at a physiological dosage and the following rejoining (Fig. 2 and Fig. 3). which could hardly be detected in reasonable manner so far. ASDG was modified mainly by reducing the centrifugal force and by using a cell-lysis solution with a high salt concentration. Another improvement was that the cells were chilled completely during X-ray irradiation, because rejoining of X-ray-induced SSB is very fast, according to our results (Fig. 3) and those of others (Fornace and Little, 1980). In fact, SSB caused by lower dosages of X-rays are ligated within a few minutes (data not shown). Using our procedure described here, SSB can be detected in a highly dose-dependent manner (Fig. 2 and Fig. 4). DNA from undamaged cells sedimented to the bottom of the gradient, and that from cells treated with damaging agents sedimented in the middle fractions. As shown in Fig. 4A, when cells were exposed to low doses of UV, peak area of accumulated SSB was nearly in proportion to the UV dose. Then, the higher doses of UV made the peak steeper (Fig. 4B). The similar tendency was shown in Xray-produced SSB (Fig. 2). To define the reliability of this method, the sizes of the ssDNA fragments produced by X-rays were calculated from the above results. From the position of the T4 DNA marker, the size of the 20 Gy-induced ssDNA fragments was estimated at 1.7 Mb (most frequent value). To compare our calculations with those of others, we extrapolated from the data
of SSB caused
by UV or dimethyl
NB 1RGB cells were exposed to 0.1- 10 J/m’ of UV, and incubated at 37°C in medium containing 3 mM hydroxyurea and 10 kg/ml aphidicolin for 1 h (Fig. 4 in Da>. SSB did not accumulate in DNA from hydroxyurea and aphidicolin-treated cells that were not UV irradiated, indicating these reagents alone do not cause SSB. In their presence, a few SSB induced by only 0.1 J/m’ UV were detectable. Increasing the UV dose resulted in the dose-dependent accumulation of SSB. Finally SSB levels induced by 5 or 10 J/m2 UV slowly reached a plateau.
-
(BOTTOM)
1lJ
FRACTfON
20
30
NUMBER
Fig. 5. ASDG profiles of DNA in DMS-treated NBlRGB cells (DMS dose-response). NBIRGB cells were labeled with [‘4C]thymidine and incubated at 37°C for 20 min in medium containing 0 CO), 25 (0) or 50 (+) pM DMS. These cells were analyzed by ASDG centrifugation as described in Materials and methods.
4. Discussion
130
K. Yamada et al./Muiation Research 364 (19%) 125-131
of Lett et al. (1967) that 20 Gy X-rays induce about one DNA strand scission per 5 X lo* Da (about 1.5 Mb) (their results did not include doses under 100 Gy due to the problem described in Introduction), which is in good agreement with our calculations. We also estimated the amount of energy necessary to cause one breakage from the X-ray dose and the size of ssDNA fragments. First, 20 Gy = 1.25 X lOi eV/g. And, this value was divided by N,/MW, for [ NA/MW] molecules of ssDNA are contained in 1 g of DNA, leading to the amount of energy given to one DNA strand. Here, NA is Avogadro’s number and MW is the molecular weight of intact DNA (in daltons). The resulting value was divided by the number of SSB, which is MW/(1.7 X 3.3 X lo’), leading to about 116 eV per break. This estimation agrees with that of Cleaver et al. (1972) who measured SSB in frozen cells, though the value is larger than one scission per 44-66 eV reported by others (Dean et al., 1969; Lehmann and Ormerod, 1970). Using this technique, we also showed the profiles of SSB accumulation in UV-exposed cells in the presence of hydroxyurea and aphidicolin (Fig. 4). The combination of these reagents inhibits UV-induced repair-synthesis almost completely. In their absence, breaks are not detected (Yamada et al., 1985; Yamada and Itoh, 1994) probably because the incision stages of repair occur slowly and sporadically, whereas polymerization and ligation proceed rapidly. Our method was proven to be extremely sensitive, because SSB induced by 0.1 J/m’ UV were detected. Furthermore. when the cells were exposed to a higher concentration of aphidicolin (for example, 30 pg/ml) together with 3 mM hydroxyurea for a longer incubation (for example, 8 h), even the SSB induced by 0.03 J/m’ UV were measurable (data not shown). We also applied this protocol to DMS-induced DNA strand breaks (Fig. 5). DMS methylates DNA bases, mainly G and A, which causes the release of bases enzymatically and non-enzymatically in cells. DNA strands are broken at the resulting apurinic/apyrimidinic sites in alkaline solution by P-elimination. Such lesions are also counted in ASDG. Our improved ASDG can be applied to study all DNA damage accompanied by strand breaks at very low levels. Here we only used NB 1RGB cells, but the same
profiles were also obtained from HeLa cells. This method is also applicable to cells cultured in suspension, such as FM3A (mouse mammary carcinoma cells) or lymphocytes, which can be labeled after phytohemagglutinin stimulation or after transformation in lymphoblasts. In addition, our procedure of ASDG is sensitive enough to detect the small number of SSB that occur in UV-repair-deficient cells, such as xeroderma pigmentosum cells. Thus, our improved ASDG centrifugation is reproducible and highly sensitive. Moreover it is easy to perform, because the protocol requires no special skills or apparatus. Accordingly the improved method will be highly valuable to studies of DNA repair. Recently, in order to observe SSB in cells, single-cell microgel electrophoresis assay (comet assay) has been examined, and it provides new possibilities of detection (for review see McKelvey-Martin et al., 1993). Various alkali unwinding techniques are also widely used. In addition to these, ASDG sedimentation will be frequently used by our modification.
Acknowledgements This work was supported by Grant for Nuclear Research from Science and Technology Agency, Japan.
References Cleaver, J.E.. G.H. Thomas and H.J. Burki (1972) Biological damage from intranuclear tritium; DNA strand breaks and their repair. Science. 177, 996-998. Dean, C.J., M.G. Ormerod, R.W. Serianni and P. Alexander (1969) DNA strand breakage in cells irradiated with X-rays. Nature, 222, 1042- 1044. Dugle, D.L. and C.J. Gillespie (1975) Kinetics of the single-strand repair mechanism in mammalian cells, in: P.C. Hanawalt and R.B. Setlow (Eds.), Molecular Mechanisms for Repair of DNA, Part A, Plenum Press, New York, pp. 685-687. Fornace. A.J., Jr. and J.B. Little (1980) Normal repair of DNA single-strand breaks in patients with ataxia telangiectasia. Biochim. Biophys. Acta, 607, 432-437. Hirose, S. (1973) Preparation of bacterial and phage DNA. in: A. Ishihama, R. Okazaki, Y. Kyogoku and S. Nishimura (Eds.) supplement of Protein, Nucleic Acids and Enzyme. Method in Nucleic Acid Research, Kyoritu Shuppan. Tokyo. pp. 4-8. Kohn. K.W.. R.A.G. Ewig, L.C. Erickson and L.A. Zwelling
K. Yamada et al/Mutation
(198 11 Measurement
of strand breaks and cross-links by alkaline elution. in: E.C. Friedberg and P.C. Hanawalt (Eds.), DNA repair - A Laboratory Manual of Research Procedures, Vol. 1, Marcel Dekker, New York, pp. 379-401. Lehmann, A.R. and M.G. Ormerod (19701 The replication of DNA in murine lymphoma cells (L5178Y1, Biochim. Biophys. Acta, 204, 128-143. Lett, J.T. (198 I ) Measurement of single strand breaks by sedimentation in alkaline sucrose gradients, in: E.C. Friedberg and P.C. Hanawalt (Eds.1, DNA repair - A Laboratory Manual of Research Procedures, Voi. 1, Marcel Dekker. New York, pp. 363-378. Lett, J.T.. I. Caldwell. C.J. Dean and P. Alexander (1967) Rejoining of X-ray induced breaks in the DNA of leukaemia cells. Nature, 214, 790-792. McCrath, R.A. and R.W. Williams (19661 Reconstitution in vivo
Research 364 (1996) 125-131
131
of irradiated Escherichia coli deoxyribonucleic acid; the rejoining of broken pieces, Nature, 2 12. 534-535. McKelvey-Martin, V.J., M.H.L. Green, P. Schmezer. B.L. PoolZobel, M.P. De Meo and A.R. Collins (19931 The single cell gel electrophoresis assay (comet assay): a European review. Mutation Res.. 288. 47-63. Pollack. A., C.B. Bagwell and G.L. Irvin, III (19791 Radiation from tritiated thymidine perturbs the cell cycle progression of stimulated lymphocytes, Science, 203. 1025-1027. Yamada, K., F. Hanaoka and M. Yamada (19851 Effects of aphidicolin and/or 2’3.dideoxythymidine on DNA repair induced in HeLa cells by four types of DNA-damaging agents, J. Biol. Chem., 260, 10412-10417. Yamada. K. and R. Itoh (19941 Involvement of DNA polymerase 6 and/or E in joining UV-induced DNA single strand breaks in human fibroblasts, Biochim. Biophys. Acta. 1219, 302-306.