- Email: [email protected]
DNA-protein crosslinking by heavy metals in Novikoff hepatoma
ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 251, No. 2, December, pp. 39’7-402,1986
INVITED PAPER DNA-Protein
Crosslinking
by Heavy Metals in Novikoff Hepatoma’
ANDRZEJ WEDRYCHOWSKI; WARREN N. SCHMIDT, AND LUBOMIR S. HNILICAt Departments of Biochemistry and Pathology, the A.B. Hancock, Jr. Memorz‘al Laboratory of the Vanderbilt University Cancer Center, and Center in Molecular Toxicology, Vanderbilt University School of Medicine, Nashville, Tennessee 37232 Received February
19, 1986, and in revised form June 20,1986 DEDICATION
It is with sincere regret that we report that Lubomir S. Hnilica, member of the editorial board of Archives and Biophysics, Professor of Biochemistry and Pathology, and Mary Geddes Stahlman Professor of Cancer Research at the Vanderbilt University School of Medicine was killed in a traffic accident on Sunday, March 31. A native of Czechoslovakia, Lubosh, as most people knew him, devoted his whole professional life to the biochemistry of histone and nonhistone proteins. We all are indebted to him for his basic pioneering observations in this field. His students, associates, colleagues and friends will miss his ever-present scientific interest and his enthusiasm linked with warm patience, consideration, and generosity.
ofBiochemi&y
Crosslinking of proteins to DNA was studied in live intact Novikoff ascites hepatoma cells exposed in vitro to salts of chromium VI, III, and II, nickel II, cadmium II, and to CoC12,As203, and AlK(SO&. DNA-protein complexes were separated by high-speed centrifugation of cells solubilized in buffered 4% sodium dodecyl sulfate and a.ssayedby polyacrylamide gel electrophoresis. Hexavalent chromium compounds formed DNAprotein complexes very efficiently. The trivalent, poorly soluble, cupric chromite was nearly as efficient crosslinker as hexavalent Cr, perhaps because phagocytosis facilitated its entry into the cells. The more basic divalent form produced hardly any crosslinks. Most of the crosslinked proteins were common to all of the chromium salts employed. Nickel salts formed DNA-protein crosslinks less efficiently. Most proteins crosslinked by this metal had a high molecular weight ranging from 94,000 to 200,000. There was little qualitative difference between the crosslinked protein patterns for several various nickel (II) salts. Similar results were obtained for cells incubated with cadmium salts. Most of the proteins crosslinked by cadmium had high molecular weights and were similar to those crosslinked by nickel (II). Relatively weak, but significant, crosslinking was also observed when the Novikoff hepatoma cells were exposed to CoC12,AsZ03, or AIK( SO&. 0 1986 Academic Press, Inc. The carcinogenic and toxic effects of chromium and other metals (l-3) as well as the ability of the trivalent chromium to react avidly with DNA to form DNA-nu-
clear protein complexes in vitro are well documented (4-10). We have shown that when live intact Novikoff ascites hepatoma cells (NAH)3 were incubated with the
1 This research was supported by NIH Grants CA 36479 and ES 00267. 2To whom correspondence should be addressed at The Department of Biochemistry, Vanderbilt University School of Medicine, Nashville, Tenn. 37232. t Deceased March 31,1986.
’ Abbreviations used: PMSF, phenylmethylsulfonyl fluoride; SDS, sodium dodecyl sulfate; NAH, Novikoff ascites hepatoma; PBS, phosphate-buffered saline (10 mM N&HPOI, 150 mM NaCl); 4% SDS buffer, 4% NaDodSO,-50 mM Tris-HCl, pH 7.5; urea SDS buffer, 4% NaDodSO&O mM Tris-HCl, pH 7.5,5 M urea; and cis-DDP, cis-diamminedichloroplatinum II. 397
0003-9861/86 $3.00 Copyright 0 1986 by Academic Press, Inc. All rights of reproduction in any form reserved.
398
WEDRYCHOWSKI,
SCHMIDT, AND HNILICA
hexavalent form of chromium, a number of chromosomal proteins became crosslinked to DNA. Some of these proteins belonged to nuclear and cytokeratin fractions (11-13). Other metals, especially Pt and Ni, were also shown to associate with the cell nucleus and produce extensive DNA-protein crosslinking (14-24). We compare here the crosslinking effects of chromium compounds in different oxidation states as well as several salts of Ni, Cd, Al, Co, and As. The extent of crosslinking varied considerably between the different metals and, to a lesser extent, between different chemical forms of the same metal. MATERIALS
AND METHODS
Materials. All the metals used in this report (K&raO,, Cr08, 2CuO~CrzOs, CrCle, NiCOs, Ni(CH&O&, NiSO*, NiClr, CdCla, Cd(CH&O&, CdSe, AlK(SO&, CoCls, and Asx08), all of highest purity, were purchased from Aldrich Chemical Company (Milwaukee, Wise.). Electrophoresis supplies were from BioRad Laboratories, Inc. (Richmond, Calif.). Novikoff ascites hepatoma (NAH) cells were passed in male Sprague-Dawley rats (Harlan Industries, Indianapolis, Ind.) every 5 days. Cells were harvested, resuspended in ice-cold 0.17 M NH&l to lyse contaminating erythrocytes, washed in PBS, and used immediately in the crosslinking experiments. Crosslinking experiments The same procedure was employed for crosslinking with all the metals. The length of incubation of the NAH cells with the metals depended upon their viability; therefore, if the viability fell below 85%,the experiment was terminated. Cells were exposed for 6 h to KzCrz07, Cr08, 2CuO. CrOs, and CrC12, or for 8 h to NiCOx, Ni(CH&O&, NiSOd, NiClx, AlK(SO&, CdClx, Cd(CH&O&, CdSe, CoClx, and Asa08. Washed NAH cells were resuspended in 20 vol (packed cell/buffer, v/v) of Hanks’ balanced salt solution containing metal at indicated concentration and incubated at 37°C for the above time intervals. During incubation the cells were gently shaken every 20 min and at each collection time, viability was determined by trypan blue exclusion assay. Next, the cells were washed twice in ice-cold PBS and solubilized in 4% SDS-50 mM Tris-HCl, pH 7.5 (4% SDS buffer). From this step on, all the solutions also contained 1 mM phenylmethylsulfonyl fluoride (PMSF). The solution was stirred slowly at room temperature for 4 h and then gently homogenized in a glass homogenizer with a loosely fitting Teflon pestle (19). To remove insoluble cellular debris the homogenized solution was centrifuged at 7’70~for 10 min. The pellets were discarded and the supernatant was
centrifuged at lC@,OoOg for 16 h. The resulting DNA pellets were rinsed with 4% SDS buffer and resuspended in 5 M urea (volume equal to the original 4% SDS solution). After stirring at 4°C for 3 h the solution was gently homogenized, SDS was added to the final concentration of 4% (urea-SDS buffer), and the solution was stirred slowly at room temperature for 1 h. After additional centrifugation at 100,909gfor 16 h, the final pellets were rinsed with urea-SDS buffer and then resuspended by sonication in 2 mMTris-HCl buffer, pH 7.5 (three 20-s bursts followed by intermittent cooling on ice), precipitated with ice-cold acetone, and resuspended in 2 mM Tris-HCl-1 mM MgClx buffer, pH 7.5. Next, DNase I was added (25 pg/ml, sp act 1872 u/mg) and the samples were incubated at 37°C for 1 h. Polyacryhmide gel electrophor&. The nuclease digested samples were made (final concentration) 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, 0.0625M Tris-HCl, pH 6.8,boiled for 5 min, and electrophoresed as described by Laemmli (25) using a 3.0% stacking gel and a 7.5% running gel. The separated proteins were either visualized with the silver staining technique of Wray et al. (26) directly or modified as follows:’ the gels were first immersed for 24 h in 10% acetic acid with three changes on a rocking platform. Next the gels were washed with 50% methanol for 3 h with three changes and then allowed to expand for 20 min in distilled water to wash away the glycine. The wash with methanol and expansion in water were repeated two more times. Next the gels were washed 3 h in 50% methanol and the exact technique of Wray et ol (26) was followed. RESULTS
Our results demonstrate that every one of the tested metals produced at least some DNA-protein crosslinks when incubated with NAH cells. Figure 1, lanes l-4, presents the DNA-protein complexes induced by K2Cr207, CrOS,2CuO * Crz03, and CrClz , respectively, during incubation for 6 h. The 1 mM concentration, has been previously shown (12) to be optimal for DNA-protein complex formation induced by chromium. It is evident from Fig. 1 that the most efficient crosslinker was K&r207 (lane 1). Another hexavalent chromium compound, Cr03 (lane 2), crosslinked fewer proteins. A majority of the proteins in lanes 1 and 2 are, however, the same. We also used trivalent and divalent forms of chromium 4 Wedrychowski, A., Olinski, R., and Hnilica, L. S. Manuscript in preparation.
METAL-MEDIATED
123
4
DNA-PROTEIN
567
FIG. 1. DNA-protein crosslinking by chromium, NiCl*, and AIK(SO&. Equal numbers of NAH cells were incubated for 6 h with 1.0 mM concentration of K&&O,, CrOs, 2CuO - CraOa, and CrCla, lanes 1-4, respectively; with 1.0 rnY concentration of NiCla, lane 5; or with 0.5 mM concentration of AlK(SO&, lane 6. Lane ‘I-molecular weight standard (BioRad) (myosin M, = 200,000; @-galactosidase, M, = 116,000; phosphorylase-8, M, = 94,000; bovine serum albumin, M, = 64,000, ovalbumin, M, = 43,000). Gels were loaded with 4 pg as DNA/lane (lane l), 8 pg (lanes 2 and 3), 15 pg (lane 4), 30 pg (lane 5), and 40 pg (lane 6).
(lanes 3 and 4, respectively). These eompounds carry a negative charge (especially Cr II) which greatly reduces their traversibility through the cell membrane. The trivalent form of chromium, 2CuO * CrzO,, did crosslink many proteins (lane 3), while the divalent one (lane 4) produced crosslinks only marginally. It is conceivable that the trivalent chromium in a virtually insoluble complex with copper was capable of transversing the cell membrane more easily than other trivalent chromium forms such as CrzO,. Figure 1 also shows that different chemical forms of hexavalent chromium were not equally efficient in crosslink formation (compare lanes 1 and 2). This may be due either to a different cell membrane permeability for these compounds or to the ease with which they may be reduced by the cellular cytochrome P-450 electron transport system. Figure 1 (lanes 5 and 6) also shows DNAprotein complexes induced by NiClz,
COMPLEXES
399
and AlK(SO&, respectively. Additionally, NAH cells were incubated with 1.0 mM concentrations of three other salts of hexaaquanickel (II), i.e., NiCOa, NiS04 and Ni( CH&O& . As expected, all produced DNA-protein crosslinking patterns nearly identical to that shown in lane 5 (data not presented). The optimal concentration for AlC& has been previously established (13) as 0.5 mM; hence, we used the same concentration for AlK(SO&. The majority of the crosslinked proteins were within high molecular weight range (Mr = 64,000200,000), with one major polypeptide at approximately M, = 64,000. This crosslinking pattern is different from that obtained for the chromium compounds. In the latter case, there were many crosslinked proteins below 64 kDa (compare lanes l-3 with lanes 5 and 6). The control containing equivalent sample of NAH cells incubated in the absence of added metal is shown in Fig. 3, lane 5. Figure 2 depicts NAH cells incubated with CdC12,Cd(CH&O& CdSe, CoC12,and Asz03. This gel was purposely overdeveloped to bring out most of the very weakly stained protein bands. It is evident from this figure that the majority of proteins crosslinked by cadmium, cobalt, and arsenic were between 94 and 200 kDa and appeared nearly identical in all five lanes. Very few proteins were crosslinked below 94 kDa and the major protein band at 64 kDa evident in lanes 5 and 6 of Fig. 1 was not present. Figure 3 shows similarities and differences in the crosslinking patterns of K&raO,, NiC03, and CdClz (lanes 1-3, respectively), each representing a unique crosslinking pattern. Lanes 2 and 3 (loaded, respectively, with 30 and 40 pg as DNA/ lane) serve better to indicate the quantitative crosslinking differences between nickel and cadmium compounds, since these two lanes were developed for the same length of time. It is evident that cadmium crosslinked much smaller amounts of protein than nickel. It is also clear that K&ra07 (lane 1) loaded with only 4 pg as DNA is a much stronger crosslinker than the other two metals. Proteins of molecular weights between 94 and 200 kDa are com-
400
WEDRYCHOWSKI.
SCHMIDT.
123456 FIG. 2. DNA-protein crosslinking by cadmium, CoC12, and A&O,. Equal numbers of NAH cells were incubated for 8 h with 1.0 mM concentration of CdCla, Cd(CH&O&, CdSe, CoCla, and A%Oa (lanes l-5, respectively). Lane 6-molecular weight standards (see Fig. 1). All lanes were loaded with 40 pg as DNA/ lane. The gel was overstained to bring out weak bands. Compare molecular weight markers in Fig. 1 with those in Fig. 2. In the latter the long gel-developing procedure resulted in the visualization of the contaminating bands of molecular weight markers even though the amount of molecular weight markers in both figures is the same.
AND
HNILICA
polyclonal antisera or monoclonal antibodies, we have shown that heavy metal salts such as K2Crz04, c&DDP, CuS04, Pb(NO&, HgC12, and AICls were capable of crosslinking nuclear matrix proteins, nuclear lamins A, B, and C, chromosomal proteins, and cytokeratins to DNA. We have also reported (13) that DNA-protein crosslinking efficiency depended upon metal salt concentration. The crosslinking pattern and metal concentration dependence, however, remained the same irrespective of whether whole live cells or isolated nuclei were exposed to heavy metals. Our findings, presented here, extend our previous results indicating that some, but not all, of the chromosomal proteins can be crosslinked to DNA by a variety of metals. The evidence that different chemical forms of the same metal crosslink essentially the same proteins in tivo can be ex-
mon to all the three metals. Below 94 kDa, however, there are hardly any DNA-protein complexes formed by CdC12, while NiCos did crosslink some proteins below 94 kDa with a major protein band at 64 kDa. K2Cr207 induced crosslinking over the entire molecular weight spectrum. DISCUSSION
Although several of the metals and their salts shown here have been reported to form DNA-protein crosslinks, the extent and qualitative analysis of this phenomenon have not been compared systematically. Alkaline elution, the principal tool used by most investigators to study the effects of metal on DNA-protein crosslink formation, does not provide information about the heterogeneity of the crosslinked proteins. We have addressed in our earlier studies the identity of some of the crosslinked proteins (ll-13,19). With the aid of
12345 FIG. 3. Comparison of the DNA-protein crosslinking by K&r.& NiC08, and CdCl,. Equal numbers of NAH cells were incubated for 6 h with K&r,O, and 8 h with NiCO,, and CdCll (1 mM concentration for all three metals). Lanes 1, 2, and 3 were loaded, respectively, with 4,30, and 40 pg as DNA/lane. Lane 4-molecular weight standards (see Fig. 1). Lane B-control NAH cells incubated for 8 h with no metal (loaded with 40 pg as DNA/lane).
METAL-MEDIATED
DNA-PROTEIN
petted. Once in the cell, the metal becomes dissociated and the crosslinking pattern will depend on the reactivity and crosslinking distance of its ions. The differences seen for the various chromium compounds most likely arise from their ability to cross cellular membranes. It is well documented that while hexavalent chromium can readily cross cellular membranes, the trivalent form cannot penetrate (4-6). Therefore, we used the trypan blue dye exclusion test to gain information about the physiological state of cellular membranes during the incubation of NAH cells with metallic compounds. While the metal concentrations used in our crosslinking experiments are unquestionably lethal over long exposure periods, the cells remained viable, with impermeable membranes for the duration of our experiments. Another difference in the reactivity and crosslinking patterns of the chromium compounds likely arises from the rate at which hexavalent chromium is reduced, in the cytoplasm, to its trivalent form (notice the DNA-equivalent load between lanes 1 and 2 in Fig. 1). It has been shown that while hexavalent chromium can enter the living cell, it is the trivalent form which can react with DNA (4-10) and form DNAprotein crosslinks (8,9, 11). Contrary to our expectations, trivalent chromium in the form of 2CuO. Crz03 formed DNA-protein crosslinks with similar efficiency as CrOs (compare lanes 2 and 3 in Fig. 1). It is unlikely that the CuO component of this complex facilitated its crosslinking activity, especially since this compound is practically insoluble in aqueous solvents. However, because of their size (approximately 3-4 pm) the 2CuO. Crz03 particles must have entered NAH cells through phagocytosis where they became chemically modified and reactive. This kind of behavior has been described in the literature and was well documented for various water insoluble nickel compounds by Costa and Mollenhauer (27). Particles of (yNi&, crNiS, aNi$ez (mean diameters between 2.2 and 4.3 pm) were easily phagocytized by CHO cells causing their morphological transformation. It is also noteworthy that another water insol-
COMPLEXES
401
uble compound, CdSe, exhibited DNA-protein crosslinking behavior similar to that of Cd(CH&Oz)z which is very soluble in aqueous solvents (Fig. 3, lanes 2 and 3). The average size of CdSe particles used in our experiments was 3 pm. With some exceptions, each metal, regardless of its chemical form, exhibited a unique DNA-protein crosslinking pattern. Although we do not know the chemical nature of these crosslinks, the differences between the individual metals may arise from their reactivities with the macromolecules involved and from the distances they can crosslink. Although the biological significance (presently under study in our and other laboratories) of the metal-mediated DNA-protein crosslinks is not known, it is noteworthy that such crosslinks are repaired at much slower rates than any direct damage to the DNA alone (28,29). ACKNOWLEDGMENT We thank Ms. Doris Harris for her outstanding sistance in preparation of this manuscript.
as-
REFERENCES 1. ENTERLINE, P. E. (1974) J. Occup. Med 16, 523526. 2. SUNDERMAN, W. F., JR. (1934) Ann Clin Lab. Sci 14,93-122. 3. NORSETH, T. (1981) Environ Health Perspect. 40, 121-130. 4. RO~LE, H. (1975) Environ Res. 10,39-53. 5. LEVIS, A. G., BIANCHI, V., TAMINO, G., ANDPEGORARO, B. (1978) Brit. J. Cancer 37,386-391. 6. GRUBER, J. E., AND JENNETTE, K. W. (1978) Bidem Biophys. Res. Cornmun. 82.386-396. 7. ROSSMAN, T. G. (1981) Environ Health Perspect. 40,189-195. 8. FORNACE, A. J., SERES, D. S., LECHNER, J. F., AND HARRIS, C. C. (1981) Chem-Biol Interact. 36, 345-354. 9. TSAPAKOS, M. J., HAMPTON, T. M., AND WETTERHAHN-J-, K. W. (1981) J. Bid C&m 256, 3623-3626. 10. ROBINSON, S. H., AM) COSTA, M. (1982) Cancer L& l&35-40. 11. WEDRYCHOWSKI, A., WARD, S. W., SCHMIDT, W. N., AND HNILICA, L. S. (1985) J. Bid Chem 260,7150-7155. 12. WEDRYCHOWSKI, A., SCHMIDT, W. N., WARD, W. S., AND HNILICA, L. S. (1986) Biochemistry 25,109.
402
WEDRYCHOWSKI,
SCHMIDT,
13. WEDRYCHOWSKI, A., SCHMIDT, W. N., AND HNILICA, L. S. (1986) J. Biol Chem 261.3370-33'76. 14. ZWELLING, L. A., ANDERSON, T., ANDKOHN, K. W. (1979) Cancer Res. 39,365-369. 15. ZWELLING, L. A., BRADLEY, M. O., SHARKEY, N. A., ANDERSON, T., AND KOHN, K. W. (1979) Mutat.
Res. 67,211-280. 16. LIPPARD, S. J., AND HOESCHELE, J. D. (1979) Pmt.
NatL Acad Sci. USA 76,6091-6095. 17. FILIPSKI, J., KOHN, K. W., AND BONNER, W. M. (1983) FEBS I&t. 152,105-108. 18. BANJAR, Z. M., HNILICA, L. S., BRIGGS, R. C., STEIN, J., AND STEIN, G. (1984) Biochemistry 23,19211926. 19. WARD, W. S., SCHMIDT, W. N., SCHMIDT, C. A., AND HNILICA, L. S. (1984) Proc. NatL Acad. Sci USA 81,419-423. 20. CICCARELLI, R. B., HAMPTON, T. H., AND JENNE’ITE, K. W. (1981) Cancer b&t. 12,349-354.
AND
HNILICA
21. CICCARELLI, R. B., AND WE~ERHAHN,
K. E. (1982)
Cancer Res. 42,3544-3549. 22. CICCARELLI, R. B., AND WETTERHAHN, K. E. (1985)
Chem-BioL Interact. 52,347-360. 23. CICCARELLI, R. B., AND WETTERHAHN, K. E. (1984)
Cancer Res. 44,3892-3897. 24. LEE, J. E., CICCARELU, R. B., AND JENNETTE, K. W. (1982) Biochemistry 21,771-778. 25. LAEMMLI, LJ. K. (1970) Nature (Lmdm) 227,680685. 26. WRAY, W., BOIJLIKAS, T., WRAY, V. P., AND HANCOCK, R. (1981) AmL B&hem 118,179-203. 27. COSTA, M., AND MOLLENHAUER, H. H. (1980) in Nickel Toxicology (Brown, S. S., and Sunderman, F. W., Jr., eds.), pp. 43-46, Academic Press, New York. 28. TSAPAKOS, M. J., HAMPTON, T. H., AND WEITERHAHN, K. E. (1983) Cancer Res., 43,5662-5667. 29. CUPO, D. Y., AND WEWERHAHN, K. E. (1984) Carcinogenc.sti 5,1705-1708.
INVITED PAPER DNA-Protein
Crosslinking
by Heavy Metals in Novikoff Hepatoma’
ANDRZEJ WEDRYCHOWSKI; WARREN N. SCHMIDT, AND LUBOMIR S. HNILICAt Departments of Biochemistry and Pathology, the A.B. Hancock, Jr. Memorz‘al Laboratory of the Vanderbilt University Cancer Center, and Center in Molecular Toxicology, Vanderbilt University School of Medicine, Nashville, Tennessee 37232 Received February
19, 1986, and in revised form June 20,1986 DEDICATION
It is with sincere regret that we report that Lubomir S. Hnilica, member of the editorial board of Archives and Biophysics, Professor of Biochemistry and Pathology, and Mary Geddes Stahlman Professor of Cancer Research at the Vanderbilt University School of Medicine was killed in a traffic accident on Sunday, March 31. A native of Czechoslovakia, Lubosh, as most people knew him, devoted his whole professional life to the biochemistry of histone and nonhistone proteins. We all are indebted to him for his basic pioneering observations in this field. His students, associates, colleagues and friends will miss his ever-present scientific interest and his enthusiasm linked with warm patience, consideration, and generosity.
ofBiochemi&y
Crosslinking of proteins to DNA was studied in live intact Novikoff ascites hepatoma cells exposed in vitro to salts of chromium VI, III, and II, nickel II, cadmium II, and to CoC12,As203, and AlK(SO&. DNA-protein complexes were separated by high-speed centrifugation of cells solubilized in buffered 4% sodium dodecyl sulfate and a.ssayedby polyacrylamide gel electrophoresis. Hexavalent chromium compounds formed DNAprotein complexes very efficiently. The trivalent, poorly soluble, cupric chromite was nearly as efficient crosslinker as hexavalent Cr, perhaps because phagocytosis facilitated its entry into the cells. The more basic divalent form produced hardly any crosslinks. Most of the crosslinked proteins were common to all of the chromium salts employed. Nickel salts formed DNA-protein crosslinks less efficiently. Most proteins crosslinked by this metal had a high molecular weight ranging from 94,000 to 200,000. There was little qualitative difference between the crosslinked protein patterns for several various nickel (II) salts. Similar results were obtained for cells incubated with cadmium salts. Most of the proteins crosslinked by cadmium had high molecular weights and were similar to those crosslinked by nickel (II). Relatively weak, but significant, crosslinking was also observed when the Novikoff hepatoma cells were exposed to CoC12,AsZ03, or AIK( SO&. 0 1986 Academic Press, Inc. The carcinogenic and toxic effects of chromium and other metals (l-3) as well as the ability of the trivalent chromium to react avidly with DNA to form DNA-nu-
clear protein complexes in vitro are well documented (4-10). We have shown that when live intact Novikoff ascites hepatoma cells (NAH)3 were incubated with the
1 This research was supported by NIH Grants CA 36479 and ES 00267. 2To whom correspondence should be addressed at The Department of Biochemistry, Vanderbilt University School of Medicine, Nashville, Tenn. 37232. t Deceased March 31,1986.
’ Abbreviations used: PMSF, phenylmethylsulfonyl fluoride; SDS, sodium dodecyl sulfate; NAH, Novikoff ascites hepatoma; PBS, phosphate-buffered saline (10 mM N&HPOI, 150 mM NaCl); 4% SDS buffer, 4% NaDodSO,-50 mM Tris-HCl, pH 7.5; urea SDS buffer, 4% NaDodSO&O mM Tris-HCl, pH 7.5,5 M urea; and cis-DDP, cis-diamminedichloroplatinum II. 397
0003-9861/86 $3.00 Copyright 0 1986 by Academic Press, Inc. All rights of reproduction in any form reserved.
398
WEDRYCHOWSKI,
SCHMIDT, AND HNILICA
hexavalent form of chromium, a number of chromosomal proteins became crosslinked to DNA. Some of these proteins belonged to nuclear and cytokeratin fractions (11-13). Other metals, especially Pt and Ni, were also shown to associate with the cell nucleus and produce extensive DNA-protein crosslinking (14-24). We compare here the crosslinking effects of chromium compounds in different oxidation states as well as several salts of Ni, Cd, Al, Co, and As. The extent of crosslinking varied considerably between the different metals and, to a lesser extent, between different chemical forms of the same metal. MATERIALS
AND METHODS
Materials. All the metals used in this report (K&raO,, Cr08, 2CuO~CrzOs, CrCle, NiCOs, Ni(CH&O&, NiSO*, NiClr, CdCla, Cd(CH&O&, CdSe, AlK(SO&, CoCls, and Asx08), all of highest purity, were purchased from Aldrich Chemical Company (Milwaukee, Wise.). Electrophoresis supplies were from BioRad Laboratories, Inc. (Richmond, Calif.). Novikoff ascites hepatoma (NAH) cells were passed in male Sprague-Dawley rats (Harlan Industries, Indianapolis, Ind.) every 5 days. Cells were harvested, resuspended in ice-cold 0.17 M NH&l to lyse contaminating erythrocytes, washed in PBS, and used immediately in the crosslinking experiments. Crosslinking experiments The same procedure was employed for crosslinking with all the metals. The length of incubation of the NAH cells with the metals depended upon their viability; therefore, if the viability fell below 85%,the experiment was terminated. Cells were exposed for 6 h to KzCrz07, Cr08, 2CuO. CrOs, and CrC12, or for 8 h to NiCOx, Ni(CH&O&, NiSOd, NiClx, AlK(SO&, CdClx, Cd(CH&O&, CdSe, CoClx, and Asa08. Washed NAH cells were resuspended in 20 vol (packed cell/buffer, v/v) of Hanks’ balanced salt solution containing metal at indicated concentration and incubated at 37°C for the above time intervals. During incubation the cells were gently shaken every 20 min and at each collection time, viability was determined by trypan blue exclusion assay. Next, the cells were washed twice in ice-cold PBS and solubilized in 4% SDS-50 mM Tris-HCl, pH 7.5 (4% SDS buffer). From this step on, all the solutions also contained 1 mM phenylmethylsulfonyl fluoride (PMSF). The solution was stirred slowly at room temperature for 4 h and then gently homogenized in a glass homogenizer with a loosely fitting Teflon pestle (19). To remove insoluble cellular debris the homogenized solution was centrifuged at 7’70~for 10 min. The pellets were discarded and the supernatant was
centrifuged at lC@,OoOg for 16 h. The resulting DNA pellets were rinsed with 4% SDS buffer and resuspended in 5 M urea (volume equal to the original 4% SDS solution). After stirring at 4°C for 3 h the solution was gently homogenized, SDS was added to the final concentration of 4% (urea-SDS buffer), and the solution was stirred slowly at room temperature for 1 h. After additional centrifugation at 100,909gfor 16 h, the final pellets were rinsed with urea-SDS buffer and then resuspended by sonication in 2 mMTris-HCl buffer, pH 7.5 (three 20-s bursts followed by intermittent cooling on ice), precipitated with ice-cold acetone, and resuspended in 2 mM Tris-HCl-1 mM MgClx buffer, pH 7.5. Next, DNase I was added (25 pg/ml, sp act 1872 u/mg) and the samples were incubated at 37°C for 1 h. Polyacryhmide gel electrophor&. The nuclease digested samples were made (final concentration) 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, 0.0625M Tris-HCl, pH 6.8,boiled for 5 min, and electrophoresed as described by Laemmli (25) using a 3.0% stacking gel and a 7.5% running gel. The separated proteins were either visualized with the silver staining technique of Wray et al. (26) directly or modified as follows:’ the gels were first immersed for 24 h in 10% acetic acid with three changes on a rocking platform. Next the gels were washed with 50% methanol for 3 h with three changes and then allowed to expand for 20 min in distilled water to wash away the glycine. The wash with methanol and expansion in water were repeated two more times. Next the gels were washed 3 h in 50% methanol and the exact technique of Wray et ol (26) was followed. RESULTS
Our results demonstrate that every one of the tested metals produced at least some DNA-protein crosslinks when incubated with NAH cells. Figure 1, lanes l-4, presents the DNA-protein complexes induced by K2Cr207, CrOS,2CuO * Crz03, and CrClz , respectively, during incubation for 6 h. The 1 mM concentration, has been previously shown (12) to be optimal for DNA-protein complex formation induced by chromium. It is evident from Fig. 1 that the most efficient crosslinker was K&r207 (lane 1). Another hexavalent chromium compound, Cr03 (lane 2), crosslinked fewer proteins. A majority of the proteins in lanes 1 and 2 are, however, the same. We also used trivalent and divalent forms of chromium 4 Wedrychowski, A., Olinski, R., and Hnilica, L. S. Manuscript in preparation.
METAL-MEDIATED
123
4
DNA-PROTEIN
567
FIG. 1. DNA-protein crosslinking by chromium, NiCl*, and AIK(SO&. Equal numbers of NAH cells were incubated for 6 h with 1.0 mM concentration of K&&O,, CrOs, 2CuO - CraOa, and CrCla, lanes 1-4, respectively; with 1.0 rnY concentration of NiCla, lane 5; or with 0.5 mM concentration of AlK(SO&, lane 6. Lane ‘I-molecular weight standard (BioRad) (myosin M, = 200,000; @-galactosidase, M, = 116,000; phosphorylase-8, M, = 94,000; bovine serum albumin, M, = 64,000, ovalbumin, M, = 43,000). Gels were loaded with 4 pg as DNA/lane (lane l), 8 pg (lanes 2 and 3), 15 pg (lane 4), 30 pg (lane 5), and 40 pg (lane 6).
(lanes 3 and 4, respectively). These eompounds carry a negative charge (especially Cr II) which greatly reduces their traversibility through the cell membrane. The trivalent form of chromium, 2CuO * CrzO,, did crosslink many proteins (lane 3), while the divalent one (lane 4) produced crosslinks only marginally. It is conceivable that the trivalent chromium in a virtually insoluble complex with copper was capable of transversing the cell membrane more easily than other trivalent chromium forms such as CrzO,. Figure 1 also shows that different chemical forms of hexavalent chromium were not equally efficient in crosslink formation (compare lanes 1 and 2). This may be due either to a different cell membrane permeability for these compounds or to the ease with which they may be reduced by the cellular cytochrome P-450 electron transport system. Figure 1 (lanes 5 and 6) also shows DNAprotein complexes induced by NiClz,
COMPLEXES
399
and AlK(SO&, respectively. Additionally, NAH cells were incubated with 1.0 mM concentrations of three other salts of hexaaquanickel (II), i.e., NiCOa, NiS04 and Ni( CH&O& . As expected, all produced DNA-protein crosslinking patterns nearly identical to that shown in lane 5 (data not presented). The optimal concentration for AlC& has been previously established (13) as 0.5 mM; hence, we used the same concentration for AlK(SO&. The majority of the crosslinked proteins were within high molecular weight range (Mr = 64,000200,000), with one major polypeptide at approximately M, = 64,000. This crosslinking pattern is different from that obtained for the chromium compounds. In the latter case, there were many crosslinked proteins below 64 kDa (compare lanes l-3 with lanes 5 and 6). The control containing equivalent sample of NAH cells incubated in the absence of added metal is shown in Fig. 3, lane 5. Figure 2 depicts NAH cells incubated with CdC12,Cd(CH&O& CdSe, CoC12,and Asz03. This gel was purposely overdeveloped to bring out most of the very weakly stained protein bands. It is evident from this figure that the majority of proteins crosslinked by cadmium, cobalt, and arsenic were between 94 and 200 kDa and appeared nearly identical in all five lanes. Very few proteins were crosslinked below 94 kDa and the major protein band at 64 kDa evident in lanes 5 and 6 of Fig. 1 was not present. Figure 3 shows similarities and differences in the crosslinking patterns of K&raO,, NiC03, and CdClz (lanes 1-3, respectively), each representing a unique crosslinking pattern. Lanes 2 and 3 (loaded, respectively, with 30 and 40 pg as DNA/ lane) serve better to indicate the quantitative crosslinking differences between nickel and cadmium compounds, since these two lanes were developed for the same length of time. It is evident that cadmium crosslinked much smaller amounts of protein than nickel. It is also clear that K&ra07 (lane 1) loaded with only 4 pg as DNA is a much stronger crosslinker than the other two metals. Proteins of molecular weights between 94 and 200 kDa are com-
400
WEDRYCHOWSKI.
SCHMIDT.
123456 FIG. 2. DNA-protein crosslinking by cadmium, CoC12, and A&O,. Equal numbers of NAH cells were incubated for 8 h with 1.0 mM concentration of CdCla, Cd(CH&O&, CdSe, CoCla, and A%Oa (lanes l-5, respectively). Lane 6-molecular weight standards (see Fig. 1). All lanes were loaded with 40 pg as DNA/ lane. The gel was overstained to bring out weak bands. Compare molecular weight markers in Fig. 1 with those in Fig. 2. In the latter the long gel-developing procedure resulted in the visualization of the contaminating bands of molecular weight markers even though the amount of molecular weight markers in both figures is the same.
AND
HNILICA
polyclonal antisera or monoclonal antibodies, we have shown that heavy metal salts such as K2Crz04, c&DDP, CuS04, Pb(NO&, HgC12, and AICls were capable of crosslinking nuclear matrix proteins, nuclear lamins A, B, and C, chromosomal proteins, and cytokeratins to DNA. We have also reported (13) that DNA-protein crosslinking efficiency depended upon metal salt concentration. The crosslinking pattern and metal concentration dependence, however, remained the same irrespective of whether whole live cells or isolated nuclei were exposed to heavy metals. Our findings, presented here, extend our previous results indicating that some, but not all, of the chromosomal proteins can be crosslinked to DNA by a variety of metals. The evidence that different chemical forms of the same metal crosslink essentially the same proteins in tivo can be ex-
mon to all the three metals. Below 94 kDa, however, there are hardly any DNA-protein complexes formed by CdC12, while NiCos did crosslink some proteins below 94 kDa with a major protein band at 64 kDa. K2Cr207 induced crosslinking over the entire molecular weight spectrum. DISCUSSION
Although several of the metals and their salts shown here have been reported to form DNA-protein crosslinks, the extent and qualitative analysis of this phenomenon have not been compared systematically. Alkaline elution, the principal tool used by most investigators to study the effects of metal on DNA-protein crosslink formation, does not provide information about the heterogeneity of the crosslinked proteins. We have addressed in our earlier studies the identity of some of the crosslinked proteins (ll-13,19). With the aid of
12345 FIG. 3. Comparison of the DNA-protein crosslinking by K&r.& NiC08, and CdCl,. Equal numbers of NAH cells were incubated for 6 h with K&r,O, and 8 h with NiCO,, and CdCll (1 mM concentration for all three metals). Lanes 1, 2, and 3 were loaded, respectively, with 4,30, and 40 pg as DNA/lane. Lane 4-molecular weight standards (see Fig. 1). Lane B-control NAH cells incubated for 8 h with no metal (loaded with 40 pg as DNA/lane).
METAL-MEDIATED
DNA-PROTEIN
petted. Once in the cell, the metal becomes dissociated and the crosslinking pattern will depend on the reactivity and crosslinking distance of its ions. The differences seen for the various chromium compounds most likely arise from their ability to cross cellular membranes. It is well documented that while hexavalent chromium can readily cross cellular membranes, the trivalent form cannot penetrate (4-6). Therefore, we used the trypan blue dye exclusion test to gain information about the physiological state of cellular membranes during the incubation of NAH cells with metallic compounds. While the metal concentrations used in our crosslinking experiments are unquestionably lethal over long exposure periods, the cells remained viable, with impermeable membranes for the duration of our experiments. Another difference in the reactivity and crosslinking patterns of the chromium compounds likely arises from the rate at which hexavalent chromium is reduced, in the cytoplasm, to its trivalent form (notice the DNA-equivalent load between lanes 1 and 2 in Fig. 1). It has been shown that while hexavalent chromium can enter the living cell, it is the trivalent form which can react with DNA (4-10) and form DNAprotein crosslinks (8,9, 11). Contrary to our expectations, trivalent chromium in the form of 2CuO. Crz03 formed DNA-protein crosslinks with similar efficiency as CrOs (compare lanes 2 and 3 in Fig. 1). It is unlikely that the CuO component of this complex facilitated its crosslinking activity, especially since this compound is practically insoluble in aqueous solvents. However, because of their size (approximately 3-4 pm) the 2CuO. Crz03 particles must have entered NAH cells through phagocytosis where they became chemically modified and reactive. This kind of behavior has been described in the literature and was well documented for various water insoluble nickel compounds by Costa and Mollenhauer (27). Particles of (yNi&, crNiS, aNi$ez (mean diameters between 2.2 and 4.3 pm) were easily phagocytized by CHO cells causing their morphological transformation. It is also noteworthy that another water insol-
COMPLEXES
401
uble compound, CdSe, exhibited DNA-protein crosslinking behavior similar to that of Cd(CH&Oz)z which is very soluble in aqueous solvents (Fig. 3, lanes 2 and 3). The average size of CdSe particles used in our experiments was 3 pm. With some exceptions, each metal, regardless of its chemical form, exhibited a unique DNA-protein crosslinking pattern. Although we do not know the chemical nature of these crosslinks, the differences between the individual metals may arise from their reactivities with the macromolecules involved and from the distances they can crosslink. Although the biological significance (presently under study in our and other laboratories) of the metal-mediated DNA-protein crosslinks is not known, it is noteworthy that such crosslinks are repaired at much slower rates than any direct damage to the DNA alone (28,29). ACKNOWLEDGMENT We thank Ms. Doris Harris for her outstanding sistance in preparation of this manuscript.
as-
REFERENCES 1. ENTERLINE, P. E. (1974) J. Occup. Med 16, 523526. 2. SUNDERMAN, W. F., JR. (1934) Ann Clin Lab. Sci 14,93-122. 3. NORSETH, T. (1981) Environ Health Perspect. 40, 121-130. 4. RO~LE, H. (1975) Environ Res. 10,39-53. 5. LEVIS, A. G., BIANCHI, V., TAMINO, G., ANDPEGORARO, B. (1978) Brit. J. Cancer 37,386-391. 6. GRUBER, J. E., AND JENNETTE, K. W. (1978) Bidem Biophys. Res. Cornmun. 82.386-396. 7. ROSSMAN, T. G. (1981) Environ Health Perspect. 40,189-195. 8. FORNACE, A. J., SERES, D. S., LECHNER, J. F., AND HARRIS, C. C. (1981) Chem-Biol Interact. 36, 345-354. 9. TSAPAKOS, M. J., HAMPTON, T. M., AND WETTERHAHN-J-, K. W. (1981) J. Bid C&m 256, 3623-3626. 10. ROBINSON, S. H., AM) COSTA, M. (1982) Cancer L& l&35-40. 11. WEDRYCHOWSKI, A., WARD, S. W., SCHMIDT, W. N., AND HNILICA, L. S. (1985) J. Bid Chem 260,7150-7155. 12. WEDRYCHOWSKI, A., SCHMIDT, W. N., WARD, W. S., AND HNILICA, L. S. (1986) Biochemistry 25,109.
402
WEDRYCHOWSKI,
SCHMIDT,
13. WEDRYCHOWSKI, A., SCHMIDT, W. N., AND HNILICA, L. S. (1986) J. Biol Chem 261.3370-33'76. 14. ZWELLING, L. A., ANDERSON, T., ANDKOHN, K. W. (1979) Cancer Res. 39,365-369. 15. ZWELLING, L. A., BRADLEY, M. O., SHARKEY, N. A., ANDERSON, T., AND KOHN, K. W. (1979) Mutat.
Res. 67,211-280. 16. LIPPARD, S. J., AND HOESCHELE, J. D. (1979) Pmt.
NatL Acad Sci. USA 76,6091-6095. 17. FILIPSKI, J., KOHN, K. W., AND BONNER, W. M. (1983) FEBS I&t. 152,105-108. 18. BANJAR, Z. M., HNILICA, L. S., BRIGGS, R. C., STEIN, J., AND STEIN, G. (1984) Biochemistry 23,19211926. 19. WARD, W. S., SCHMIDT, W. N., SCHMIDT, C. A., AND HNILICA, L. S. (1984) Proc. NatL Acad. Sci USA 81,419-423. 20. CICCARELLI, R. B., HAMPTON, T. H., AND JENNE’ITE, K. W. (1981) Cancer b&t. 12,349-354.
AND
HNILICA
21. CICCARELLI, R. B., AND WE~ERHAHN,
K. E. (1982)
Cancer Res. 42,3544-3549. 22. CICCARELLI, R. B., AND WETTERHAHN, K. E. (1985)
Chem-BioL Interact. 52,347-360. 23. CICCARELLI, R. B., AND WETTERHAHN, K. E. (1984)
Cancer Res. 44,3892-3897. 24. LEE, J. E., CICCARELU, R. B., AND JENNETTE, K. W. (1982) Biochemistry 21,771-778. 25. LAEMMLI, LJ. K. (1970) Nature (Lmdm) 227,680685. 26. WRAY, W., BOIJLIKAS, T., WRAY, V. P., AND HANCOCK, R. (1981) AmL B&hem 118,179-203. 27. COSTA, M., AND MOLLENHAUER, H. H. (1980) in Nickel Toxicology (Brown, S. S., and Sunderman, F. W., Jr., eds.), pp. 43-46, Academic Press, New York. 28. TSAPAKOS, M. J., HAMPTON, T. H., AND WEITERHAHN, K. E. (1983) Cancer Res., 43,5662-5667. 29. CUPO, D. Y., AND WEWERHAHN, K. E. (1984) Carcinogenc.sti 5,1705-1708.