Volume 6 Number 11 1979 Volume 6 Number 111979
Nucleic Nucleic Acids Acids Research
Research~
Structural investigations of DNA-histone complexes. A spin label study1 Birandra K.Sinha* and Colin F.Chignell National Institute of Environmental Health Sciences, Post Office Box 12233, Research Triangle Park, NC 27709, and Victorio T.Wee2 National Heart, Lung and Blood Institute, Bethesda, MD 20014, USA
Received 13 March 1979
ABSTRACT We have prepared two acridine spin labels, 6-chloro-9-[4-(2,2,6,6tetramethyl-l-piperidinyloxy)amino]-2-methoxyacridine (I) and 9-[4-(2,2, 6,6-tetramethyl-l-piperidinyloxy)amino]-acridine (II) and have used them to study the binding of lysine-rich histone (H1) to DNA using electron spin resonance (ESR). ESR spectra of I in the presence of DNA, polydApolydT and polydG-polydC were characteristic of highly innobilized radicals with maximum hyperfine splitting (2T ) of 59G, 62.5G and 59G respectively. However, the 2T values fN4 II in the same systems were 55.5G, 55.5G and 62.5G respectively. Addition of H1 at a low P/D released ionically bound I and II from DNA. In the presence of 0.1 M NaCl, which prevents ionic binding, H1 still caused a significant release of bound II but not I from DNA. At a high P/D (with or without NaCl) H1 caused no displacement of either I or II. Our findings suggest that H does not affect the intercalating sites and probably binds to one of tAe grooves of DNA, most probably the major groove, and specifically in the A-T-rich regions.
Introduction The association of DNA-binding proteins, such as histones, with nucleic acids must involve charge-charge interactions with the sugar phosphate backbone of the DNA. It is, therefore, believed that most native DNA-protein interactions involve a rather non-specific interaction between the protein side chains and the functional groups located in the major or the minor grooves of the DNA with a defined sequence of base pairs. Nucleohistones, complexes of DNA and histones, are thought to be involved in genetic regulation. The mechanism of this control may involve changes in the interactions and conformations of nucleohistones. The molecular structure of nucleohistone is not precisely known, however, considerable progress in this area has been made since the isolation of repetitive subunits called nucleosomes or nu-bodies.1 Each nucleosome contains about 140 base pairs of DNA complexed with a pair of each of C) Information Retrieval Limited 1 Falconberg Court London Wl V 5FG England
3703
Nucleic Acids Research the four histone types H2A, H2B, H3 and H4.24 Histone H1, composed of about 25% lysine residue, is not part of the nucleosome proper but is bound to the linker regions between adjacent particles.5'6 Histone H has been implicated in the maintenance of higher order structure of chromatin.2-7 Furthermore, acetylation, deacetylation and phosphorylation of H1 are believed to be involved in DNA replication.8 Recently, aflatoxin-B1 has been shown to increase the rate of histone deacetylation in the rat liver.9 In recent years spin labeled drug molecules have played an indreasingly important role in studies of interactions between drugs and their cellular targets or receptor molecules.1 0-12 The commonly employed spin label utilizes the nitroxide group, since this free radical is very stable in aqueous solution at physiological pH values. In our earlier papers, we have described procedures for synthesizing spin labeled analogs of 9-aminoacridinesl3 and electron spin resonance (ESR) studies of their binding with calf thymus DNA.14 In this paper, we describe the binding of lysine-rich histone (H1) to DNA using ESR techniques and spin labeled 9-aminoacridine analogs I and II.
Hl
=R1 = OCH3 R2=C I U =R1=R2= H Methods and Materials Spin labeled acridines were prepared by reacting 4-amino-2.2,6,6tetramethyl-l-piperidinyloxy with the appropriate 9-chloroacridines according to the published method.13 Calf thymus DNA, polydA-polydT, polydG-poly dC and lysine-rich histone were purchased from Sigma Biochemicals. Samples for ESR studies were prepared by the addition of a methanolic solution (5 x 103 ) M) of spin labels to an aqueous solu3704
Nucleic Acids Research tion of nucleic acid containing 5 mM phosphate buffer (pH 7.4) and, where indicated, 100 mM NaCl. ESR spectra were recorded with a Varian E-109 spectrometer operating at 9.4 GHz. Samples were introduced into an E-238 TMHo cavity in quartz flat cells at ambient temperature (250). Titration of DNA with H1 was carried out in the flat cell (provided with a syringe in the bottom for mixing) while monitoring the increase in the high field line. The binding of the labels I and II to DNA and the synthetic polymers were carried out over a wide range of P/D ratios by monitoring the amplitude of the low field line of the unbound spin labels. The binding of I and II with DNA in the presence of histone H1 were carried out by first preparing DNA-H1 complex in 10% NaCl and then slowly dialyzing the complex to the desired buffer according to the method of Johns, et al . 15
Results and Discussion The ESR spectra of I and II in dilute solution consisted of three sharp lines of equal intensity with a splitting of 17 G between adjacent hyperfine lines. The ESR spectra of I and II in the presence of calf thymus DNA contained broad line components (Figure 1) and resembled that of highly immobilized nitroxide radicals with maximal hyperfine splittings (2T11) of 59.0 G and 55.5 G respectively (Table 1). These observations indicate that when I and II bind to DNA, the nitroxide bearing piperidine rings become imiobilized. The high degree of immobilization of I and II suggest an intercalative binding mechanism, since it appears unlikely that the motion of the piperidine rings of I and II would be reduced so
IO
Figure 1. The ESR spectrum of spin label I (20 pM) bound to calf thymus DNA (1.0 mg/ml) in phosphate buffer (5 mM) pH 7.4. 3705
Nucleic Acids Research TABLE 1 THE HYPERFINE SPLITTINGS (2T11) OF I AND II BOUND TO DNA, PolydA-dT AND PolydG-dC
Label I I I II II II
Nucleic Acid
2T1l
(Gauss)
DNA
59.0
PolydA-polydT
62.5
PolydG-polydC
59.0
DNA PolydA-polydT
55.5 55.5 62.5
PolydG-polydC
drastically if these compounds bound at the sugar-phosphate groups on the outside of the DNA helix. Supporting evidence for intercalation have also come from sedimentation viscosityl3 and Tm measurements studies with these labels. While the addition of 0.1 M NaCl decreased the ionic binding of I and II to the phosphate groups of DNA, it did not affect 2T11 values.14 The observed difference in the 2Tll values also suggest that the binding of I and II to DNA may be different. In order to ascertain the nature of binding sites for I and II, studies were carried out with two synthetic polynucleotides, polydA-polydT and polydG-polydC. The 2T11 values obtained (Table 1) suggest that these labels occupy different binding sites. For example the 2T11 value for I bound to DNA is the same as that for polydG-polydC while the 2T11 value for II bound to DNA corresponds to that for polydA-polydT. In order to characterize further the binding of I and II to DNA, displacement studies were carried out with daunorubicin, an antitumor agent, known to bind to DNA by intercalation.16,17 Kersten, et al.18 have shown that while the complex formation with DNA is only slightly dependent on the G-C content of the DNA, the stablization of the double helix by daunorubicin is dependent on the A-T content of the DNA. Chandra, et al.,19 on the basis of polymerase inhibition studies using synthetic polymers, have also shown that daunorubicin has a high affinity for the A-T-rich regions of the DNA. When daunorubicin was added to complexes formed between I and DNA in 5 mM phosphate 30% of bound I was displaced. In contrast, the addition of daunorubicin to DNA-II complexes under similar conditions caused a 70% displacement of bound II 3706
Nucleic Acids Research
70 I-
60
/
z LbJ 50
tLiJU40
/
a- 2
CL30p1
/
v
/
20
e
10 0
1 0
5
1o
1 15
20
D/L Figure 2. Displacement of bound spin labels I and II (10 vM) from calf thymus DNA (0.5 mg/ml) by daunorubicin. I in 5 mM phosphate (A) I in 0.1 M NaCl (0), II in 5 mM phosphate ([]), II in 0.1 M NaCl
( x ).
from DNA. In the presence of 0.1 M NaCl, which decreases the binding of I and II at the phosphate groups of DNA, daunorubicin displaced only 12% of the bound I from DNA. However, the addition of daunorubicin caused a much larger displacement (46%) of bound II from DNA. A larger displacement of It from DNA coupled with rather small displacement of bound I by daunorubicin suggest that II occupies similar binding sites as daunorubicin in DNA; namely the A-T-rich regions of the DNA. Actinomycin D (AMD) has been shown to bind to DNA and the phenoxazone ring of AMD is intercalated preferentially between pdG-dC base pairs while the peptide subunits of AMD lie in the minor groove of the DNA and interact with deoxyguanosine residues in the minor groove through specific hydrogen bonds. 20 Addition of AMD at a low AMD/spin label ratio failed to displace either I or II from DNA. At a very high AMD/spin label ratio (120), 15% of the bound I from DNA could be displaced. In contrast, at this ratio, only 5% of the bound II was displaced. The binding of label I and II to the synthetic polymers, polydGpolydC and polydA-polydT, was measured over a wide range of P/D ratios by 3707
Nucleic Acids Research monitoring the amplitude of the high field line of the unbound labels. The Scatchard plots for I and II in 5 mM phosphate with the synthetic polymers and DNA are presented in Figure 3. The binding data for label II indicate that this label has a greater affinity for polydA-polydT and binds to these regions in DNA. When histone H1 was added to a complex formed between I and DNA at a low P/D ratio, some of the bound I was displaced into the solution (Figure 4). In the presence of 0.1 M NaCl, which precludes the ionic binding at the phosphate groups of the DNA helix, addition of H1 had little or no effect on the bound I. In contrast, H1 caused a significant displacement of the bound II from DNA in the presence of NaCl (Figure 5). At a high P/D ratio, where I and II are primarily bound by intercalation only, addition of H1 to DNA did not displace either I or II (Figure 6). The addition of NaCl did not alter this effect. The binding of I and II to DNA-H1 complexes (prepared in 10% NaCl followed by dialysis to 5 mM phosphate) over a wide range of P/D ratios was measured. The binding results obtained with II and DNA-H1 and DNA in 0.1 M NaCl are presented in Figure 7. The results suggest that while the addition of NaCl decreased the binding of II to DNA, the presence of H1 caused a large decrease in the binding of II from its high affinity sites on DNA. Under identical experimental conditions, H1 also
3T 3~~~~~
0~~~~~~~~~~~~~~~~~~~~ 0
.1
.2
.3
.4
r
.5
.6
.7
.8
0
.05
.1
.15
.2
.25 -.3
r
Figure 3. Scatchard plots of the binding of I (A) and II (B) to polydA-dA (0) polydG-dC (A), and calf thymus DNA ( [] ). y= moles of I and II bound per mole of nucleotides. C = concentration of free labels. 3708
Nucleic Acids Research
IF
Figure 4. The ESR spectra of the binding of H (50 ug) to calf thymus DNA (0.25 mg/ml) in the presence of label I PM). (1) I bound to DNA. (2) I bound to DNA in the presence of H . (3) I bound to DNA in 0.1 M NaCl. (4) I bound to DNA in 0.1 M NaCI in the presence of H1. Arrows represent the free spin label peaks.
610
I
A It
OG
H
Figure 5. The ESR spectra of the binding of H1 (50 iig) to calf thymus DNA (0.25 mg/ml) in the presence of label II (10 PM). (1) II bound to DNA. (2) II bound to DNA in 0.1 M NaCl. (3) II bound to DNA in the presence of H (4) II bound to DNA in 0.1 M NaCl in the presence of H1. Arrows represent the free spin label peaks. 3709
Nucleic Acids Research
IOG
Figure 6. The ESR titration of DNA (2.0 mg/ml) in the presence of (1) 20 9g. H1. (2) 40 ug H1, (3) 60 .g H1, (4) II (20 um) with H (5) 1 o -ig H1. 100 .g
Hi,
decreased the binding of I to DNA, however, considerably more of I was bound to DNA. A large decrease in the binding of II, as compared to I, to DNA in the presence of H1 suggests that H1 is binding to the same affinity sites as II in DNA. Our previous results14 confirm that I and II interact with DNA. 1.6 14 1.2 1.0 ..X
.8
u.6
Figure 7. Scatchard plots of the binding of II ( 0 ) to DNA (1 mg/ml) in the presence of 0.1 M NaCl, and (A) in the presence of H1 (0.5 mg/ml) in 5 mM phosphate. 3710
Nucleic Acids Research The binding of I and II to DNA has been characterized as one of intercalation of the planar acridine ring system into the bases of DNA. In addition to intercalation, ionic binding is also involved, as addition of NaCl decreased binding of I and II to DNA. The present studies suggest that I and II occupy different binding sites in native DNA. For example the 2T11 value for bound I to DNA is the same as that for polydG-polydC while the 2Tll value for bound II to DNA is the same as that for polydA-polydT. The suggestion that I and II occupy different binding sites is further supported by the observation that while I showed a dose dependent stimulation of DNA-dependent RNA polymerase, similar to the polyamines, II is a strong inhibitor of this enzyme.13 The high affinity of II for the A-T rich regions of DNA has been shown by a greater displacement caused by the addition of daunorubicin, an anthracycline antitumor agent. The binding of daunorubicin to DNA is well establishedl6'17 and involves intercalation of this drug into DNA. Furthermore, it has been shown that daunorubicin has high affinity for the A-T regions of DNA18'19 X-ray crystallographic16 studies of daunorubicin-DNA complexes have shown that the intercalation proceeds through the major groove of the DNA. The displacement studies with actinomycin D, an agent known to bind to G-C rich regions and in the minor groove of the DNA, show that I has higher affinity for the G-C regions. The relatively small displacement of both I and II by AMD may be due to the high affinity of the acridine labels for DNA. The binding of I and II to the synthetic polymers over a wide range of P/D ratios further substantiates the idea that II has higher binding affinity for the polydA-polydT and binds to these A-T rich regions in the native DNA. At low P/D ratios where both ionic (phosphate and groove) and intercalative binding siles are occupied, the addition of H1 caused displacement of bound I and II from DNA. When binding to the ionic sites at the phosphate groups was prevented by the addition of NaCl, only bound II could be displaced by H1. This finding suggest that II and H1 may be occupying the same binding sites. Our results on binding studies with I and II in the DNA-H1 complexes over different P/D ratios show that H1 decreased the binding of II to DNA. This indicated that H1 was binding to the same sites where II binds to DNA. At high P/D ratio, where I and II are bound only by intercalation, our results indicate that H1, does not affect or bind to these sites. Based on the thermal denaturation studies with nucleohistones, 3711
Nucleic Acids Research Ohba21 has suggested that histones interact with the A-T rich regions of DNA. Energy calculation studies on protein-DNA binding have shown that the interaction .of lysine residues with the A-T regions of DNA provides the most stable complex with slight stacking of the DNA bases.22 Furthermore, Shapiro, et al.23 have shown that poly-L-lysine interacts exclusively with the A-T bases of DNA. Recently, Goodwin, et al.24 have shown that the nonhistone proteins are located in the major groove of DNA. Our ESR results do not distinguish unequivocally between the minor or the major groove of DNA, thus binding of H1 could involve either groove. However, our ESR results do show that label II has a high affinity for polydA-polydT and binds to these regions in native DNA. In addition, we have shown that II and daunorubicin occupy the same binding sites in DNA. Since the binding of daunorubicin is well established and involves the A-T bases and the major groove of DNA, label II may also intercalate in the major groove. Since H1 displaces only bound II and reduced the binding of II with DNA, we believe that H1 must also bind to the A-T rich regions of DNA, the same binding sites occupied by II. Thus, the binding of H 1 with the native DNA must involve chargecharge interactions between the lysine residues of H1 and the functional groups of the major groups of the DNA with a defined sequence of A-T base pairs. These results are interesting in connection with the findings of Votavova, et al.25 who have suggested that H1 binds to the major groove of DNA.
*To whom all correspondence should be addressed
Presented in part at the 7th International Congress of Pharmacology, July 16-21, 1978, Paris, France.
2Proctor and Gamble Company, Ivorydale Technical Center, Professional and Regulatory Services Division, Cincinnati, Ohio 45217
REFERENCES 1. Oudet, M., Gross-Bellard, M. and Chambon, P. (1975) Cell 4: 281. 2. Elgin, S.C.R. and Weintraub, H. (1975) Ann. Rev. Biochem. 44: 725. 3. Kornberg, R.D. (1974) Science 184: 868. 3712
Nucleic Acids Research 4. Van Holde, K.E., Sharsrabuddhe, C.G. and Shaw, B.R. (1974) Nucleic Acids Res. 1: 1579. 5. Varshavsky, A.J., Bakayev, V.V. and Gieorgiev, G.P. (1976) Nucleic Acids Res. 3: 477. 6. Noll, M.FanU Kornberg, R.D. (1977) J. Mol. Biol. 109: 393. 7. Muller, V., Zentgraf, H., Eicken, TT.anidKeller, W(1978) Science 201: 406. 8. 7iU;ppers, R., Otto, B. and Bohme, R. (1978) Nucleic Acids Res. 5: 2113. 9. Edwards, G.S. and Allfrey, V.G. (1973) Biochem. Biophys. Acta 299: 354. 10. McConnell, H.M. and McFarland, B.R. (1970) W Rev. Biophys. 3: 91. 11. Griffith, O.H., Jost, P. (1972) in Methods in Pharmiacolo, Vol. 2, pp.223-276, Appleton-Century-Croft, or 12. Chignell, C.F., Starkweather, D.K. and Sinha, B.K. (1975) J. Biol. Chem. 250: 5622. 13. Sinha,T?K., Cysyk, R., Millar, D. and Chignell, C.F. (1976) J. Med. Chem. 19: 994. 14. Sinha,.K. and Chignell, C.F. (1975) Life Sciences 17: 1829. 15. Johns, E.W. and Butler, J.A.V. (1964) Nature204: 8W3. 16. Pigram, W.J., Fuller, W. and Hamilton,LT7.D(T7M2) Nature 235: 17. 17. Zunino, F., Gambetta, R.A., DiMarco, A. and Zaccara, A. (1972) Biochem. Biophys. Acta 277: 489. 18. Kersten, W. and Kersten, H. (1968) in "Molecular Association in iology, p. 289, Academic Press, New York. 19. Chandra, P., Zunino, F., Gotz, A., Gericke, D., Thorbeck, R., and DiMarco, A. 1972) FEBS Letters 21: 264. 20. Sobel, H.M. 1974) Canc Reports 58: 101. 21. Ohba, Y. (1976) Biochem. Biophs. Acta 234. 22. DeSantis, P., FoiiTaii W Rizzo, TR.(T74) Biopol ymers 13: 313. 23. Shapiro, J.T., Leng, M. and Felsenfeld, G. (1W969-) iBheminstry 8: 31219. 24. Goodwin, D.C. and Brahms, J. (1979) Nucleic Acids Res. 5: 835. 25. Votavova, H., Blah, K. and Sponar, J. (1978) StudiaBophysics 67: 79.
3713
Nucleic Nucleic Acids Acids Research
Research~
Structural investigations of DNA-histone complexes. A spin label study1 Birandra K.Sinha* and Colin F.Chignell National Institute of Environmental Health Sciences, Post Office Box 12233, Research Triangle Park, NC 27709, and Victorio T.Wee2 National Heart, Lung and Blood Institute, Bethesda, MD 20014, USA
Received 13 March 1979
ABSTRACT We have prepared two acridine spin labels, 6-chloro-9-[4-(2,2,6,6tetramethyl-l-piperidinyloxy)amino]-2-methoxyacridine (I) and 9-[4-(2,2, 6,6-tetramethyl-l-piperidinyloxy)amino]-acridine (II) and have used them to study the binding of lysine-rich histone (H1) to DNA using electron spin resonance (ESR). ESR spectra of I in the presence of DNA, polydApolydT and polydG-polydC were characteristic of highly innobilized radicals with maximum hyperfine splitting (2T ) of 59G, 62.5G and 59G respectively. However, the 2T values fN4 II in the same systems were 55.5G, 55.5G and 62.5G respectively. Addition of H1 at a low P/D released ionically bound I and II from DNA. In the presence of 0.1 M NaCl, which prevents ionic binding, H1 still caused a significant release of bound II but not I from DNA. At a high P/D (with or without NaCl) H1 caused no displacement of either I or II. Our findings suggest that H does not affect the intercalating sites and probably binds to one of tAe grooves of DNA, most probably the major groove, and specifically in the A-T-rich regions.
Introduction The association of DNA-binding proteins, such as histones, with nucleic acids must involve charge-charge interactions with the sugar phosphate backbone of the DNA. It is, therefore, believed that most native DNA-protein interactions involve a rather non-specific interaction between the protein side chains and the functional groups located in the major or the minor grooves of the DNA with a defined sequence of base pairs. Nucleohistones, complexes of DNA and histones, are thought to be involved in genetic regulation. The mechanism of this control may involve changes in the interactions and conformations of nucleohistones. The molecular structure of nucleohistone is not precisely known, however, considerable progress in this area has been made since the isolation of repetitive subunits called nucleosomes or nu-bodies.1 Each nucleosome contains about 140 base pairs of DNA complexed with a pair of each of C) Information Retrieval Limited 1 Falconberg Court London Wl V 5FG England
3703
Nucleic Acids Research the four histone types H2A, H2B, H3 and H4.24 Histone H1, composed of about 25% lysine residue, is not part of the nucleosome proper but is bound to the linker regions between adjacent particles.5'6 Histone H has been implicated in the maintenance of higher order structure of chromatin.2-7 Furthermore, acetylation, deacetylation and phosphorylation of H1 are believed to be involved in DNA replication.8 Recently, aflatoxin-B1 has been shown to increase the rate of histone deacetylation in the rat liver.9 In recent years spin labeled drug molecules have played an indreasingly important role in studies of interactions between drugs and their cellular targets or receptor molecules.1 0-12 The commonly employed spin label utilizes the nitroxide group, since this free radical is very stable in aqueous solution at physiological pH values. In our earlier papers, we have described procedures for synthesizing spin labeled analogs of 9-aminoacridinesl3 and electron spin resonance (ESR) studies of their binding with calf thymus DNA.14 In this paper, we describe the binding of lysine-rich histone (H1) to DNA using ESR techniques and spin labeled 9-aminoacridine analogs I and II.
Hl
=R1 = OCH3 R2=C I U =R1=R2= H Methods and Materials Spin labeled acridines were prepared by reacting 4-amino-2.2,6,6tetramethyl-l-piperidinyloxy with the appropriate 9-chloroacridines according to the published method.13 Calf thymus DNA, polydA-polydT, polydG-poly dC and lysine-rich histone were purchased from Sigma Biochemicals. Samples for ESR studies were prepared by the addition of a methanolic solution (5 x 103 ) M) of spin labels to an aqueous solu3704
Nucleic Acids Research tion of nucleic acid containing 5 mM phosphate buffer (pH 7.4) and, where indicated, 100 mM NaCl. ESR spectra were recorded with a Varian E-109 spectrometer operating at 9.4 GHz. Samples were introduced into an E-238 TMHo cavity in quartz flat cells at ambient temperature (250). Titration of DNA with H1 was carried out in the flat cell (provided with a syringe in the bottom for mixing) while monitoring the increase in the high field line. The binding of the labels I and II to DNA and the synthetic polymers were carried out over a wide range of P/D ratios by monitoring the amplitude of the low field line of the unbound spin labels. The binding of I and II with DNA in the presence of histone H1 were carried out by first preparing DNA-H1 complex in 10% NaCl and then slowly dialyzing the complex to the desired buffer according to the method of Johns, et al . 15
Results and Discussion The ESR spectra of I and II in dilute solution consisted of three sharp lines of equal intensity with a splitting of 17 G between adjacent hyperfine lines. The ESR spectra of I and II in the presence of calf thymus DNA contained broad line components (Figure 1) and resembled that of highly immobilized nitroxide radicals with maximal hyperfine splittings (2T11) of 59.0 G and 55.5 G respectively (Table 1). These observations indicate that when I and II bind to DNA, the nitroxide bearing piperidine rings become imiobilized. The high degree of immobilization of I and II suggest an intercalative binding mechanism, since it appears unlikely that the motion of the piperidine rings of I and II would be reduced so
IO
Figure 1. The ESR spectrum of spin label I (20 pM) bound to calf thymus DNA (1.0 mg/ml) in phosphate buffer (5 mM) pH 7.4. 3705
Nucleic Acids Research TABLE 1 THE HYPERFINE SPLITTINGS (2T11) OF I AND II BOUND TO DNA, PolydA-dT AND PolydG-dC
Label I I I II II II
Nucleic Acid
2T1l
(Gauss)
DNA
59.0
PolydA-polydT
62.5
PolydG-polydC
59.0
DNA PolydA-polydT
55.5 55.5 62.5
PolydG-polydC
drastically if these compounds bound at the sugar-phosphate groups on the outside of the DNA helix. Supporting evidence for intercalation have also come from sedimentation viscosityl3 and Tm measurements studies with these labels. While the addition of 0.1 M NaCl decreased the ionic binding of I and II to the phosphate groups of DNA, it did not affect 2T11 values.14 The observed difference in the 2Tll values also suggest that the binding of I and II to DNA may be different. In order to ascertain the nature of binding sites for I and II, studies were carried out with two synthetic polynucleotides, polydA-polydT and polydG-polydC. The 2T11 values obtained (Table 1) suggest that these labels occupy different binding sites. For example the 2T11 value for I bound to DNA is the same as that for polydG-polydC while the 2T11 value for II bound to DNA corresponds to that for polydA-polydT. In order to characterize further the binding of I and II to DNA, displacement studies were carried out with daunorubicin, an antitumor agent, known to bind to DNA by intercalation.16,17 Kersten, et al.18 have shown that while the complex formation with DNA is only slightly dependent on the G-C content of the DNA, the stablization of the double helix by daunorubicin is dependent on the A-T content of the DNA. Chandra, et al.,19 on the basis of polymerase inhibition studies using synthetic polymers, have also shown that daunorubicin has a high affinity for the A-T-rich regions of the DNA. When daunorubicin was added to complexes formed between I and DNA in 5 mM phosphate 30% of bound I was displaced. In contrast, the addition of daunorubicin to DNA-II complexes under similar conditions caused a 70% displacement of bound II 3706
Nucleic Acids Research
70 I-
60
/
z LbJ 50
tLiJU40
/
a- 2
CL30p1
/
v
/
20
e
10 0
1 0
5
1o
1 15
20
D/L Figure 2. Displacement of bound spin labels I and II (10 vM) from calf thymus DNA (0.5 mg/ml) by daunorubicin. I in 5 mM phosphate (A) I in 0.1 M NaCl (0), II in 5 mM phosphate ([]), II in 0.1 M NaCl
( x ).
from DNA. In the presence of 0.1 M NaCl, which decreases the binding of I and II at the phosphate groups of DNA, daunorubicin displaced only 12% of the bound I from DNA. However, the addition of daunorubicin caused a much larger displacement (46%) of bound II from DNA. A larger displacement of It from DNA coupled with rather small displacement of bound I by daunorubicin suggest that II occupies similar binding sites as daunorubicin in DNA; namely the A-T-rich regions of the DNA. Actinomycin D (AMD) has been shown to bind to DNA and the phenoxazone ring of AMD is intercalated preferentially between pdG-dC base pairs while the peptide subunits of AMD lie in the minor groove of the DNA and interact with deoxyguanosine residues in the minor groove through specific hydrogen bonds. 20 Addition of AMD at a low AMD/spin label ratio failed to displace either I or II from DNA. At a very high AMD/spin label ratio (120), 15% of the bound I from DNA could be displaced. In contrast, at this ratio, only 5% of the bound II was displaced. The binding of label I and II to the synthetic polymers, polydGpolydC and polydA-polydT, was measured over a wide range of P/D ratios by 3707
Nucleic Acids Research monitoring the amplitude of the high field line of the unbound labels. The Scatchard plots for I and II in 5 mM phosphate with the synthetic polymers and DNA are presented in Figure 3. The binding data for label II indicate that this label has a greater affinity for polydA-polydT and binds to these regions in DNA. When histone H1 was added to a complex formed between I and DNA at a low P/D ratio, some of the bound I was displaced into the solution (Figure 4). In the presence of 0.1 M NaCl, which precludes the ionic binding at the phosphate groups of the DNA helix, addition of H1 had little or no effect on the bound I. In contrast, H1 caused a significant displacement of the bound II from DNA in the presence of NaCl (Figure 5). At a high P/D ratio, where I and II are primarily bound by intercalation only, addition of H1 to DNA did not displace either I or II (Figure 6). The addition of NaCl did not alter this effect. The binding of I and II to DNA-H1 complexes (prepared in 10% NaCl followed by dialysis to 5 mM phosphate) over a wide range of P/D ratios was measured. The binding results obtained with II and DNA-H1 and DNA in 0.1 M NaCl are presented in Figure 7. The results suggest that while the addition of NaCl decreased the binding of II to DNA, the presence of H1 caused a large decrease in the binding of II from its high affinity sites on DNA. Under identical experimental conditions, H1 also
3T 3~~~~~
0~~~~~~~~~~~~~~~~~~~~ 0
.1
.2
.3
.4
r
.5
.6
.7
.8
0
.05
.1
.15
.2
.25 -.3
r
Figure 3. Scatchard plots of the binding of I (A) and II (B) to polydA-dA (0) polydG-dC (A), and calf thymus DNA ( [] ). y= moles of I and II bound per mole of nucleotides. C = concentration of free labels. 3708
Nucleic Acids Research
IF
Figure 4. The ESR spectra of the binding of H (50 ug) to calf thymus DNA (0.25 mg/ml) in the presence of label I PM). (1) I bound to DNA. (2) I bound to DNA in the presence of H . (3) I bound to DNA in 0.1 M NaCl. (4) I bound to DNA in 0.1 M NaCI in the presence of H1. Arrows represent the free spin label peaks.
610
I
A It
OG
H
Figure 5. The ESR spectra of the binding of H1 (50 iig) to calf thymus DNA (0.25 mg/ml) in the presence of label II (10 PM). (1) II bound to DNA. (2) II bound to DNA in 0.1 M NaCl. (3) II bound to DNA in the presence of H (4) II bound to DNA in 0.1 M NaCl in the presence of H1. Arrows represent the free spin label peaks. 3709
Nucleic Acids Research
IOG
Figure 6. The ESR titration of DNA (2.0 mg/ml) in the presence of (1) 20 9g. H1. (2) 40 ug H1, (3) 60 .g H1, (4) II (20 um) with H (5) 1 o -ig H1. 100 .g
Hi,
decreased the binding of I to DNA, however, considerably more of I was bound to DNA. A large decrease in the binding of II, as compared to I, to DNA in the presence of H1 suggests that H1 is binding to the same affinity sites as II in DNA. Our previous results14 confirm that I and II interact with DNA. 1.6 14 1.2 1.0 ..X
.8
u.6
Figure 7. Scatchard plots of the binding of II ( 0 ) to DNA (1 mg/ml) in the presence of 0.1 M NaCl, and (A) in the presence of H1 (0.5 mg/ml) in 5 mM phosphate. 3710
Nucleic Acids Research The binding of I and II to DNA has been characterized as one of intercalation of the planar acridine ring system into the bases of DNA. In addition to intercalation, ionic binding is also involved, as addition of NaCl decreased binding of I and II to DNA. The present studies suggest that I and II occupy different binding sites in native DNA. For example the 2T11 value for bound I to DNA is the same as that for polydG-polydC while the 2Tll value for bound II to DNA is the same as that for polydA-polydT. The suggestion that I and II occupy different binding sites is further supported by the observation that while I showed a dose dependent stimulation of DNA-dependent RNA polymerase, similar to the polyamines, II is a strong inhibitor of this enzyme.13 The high affinity of II for the A-T rich regions of DNA has been shown by a greater displacement caused by the addition of daunorubicin, an anthracycline antitumor agent. The binding of daunorubicin to DNA is well establishedl6'17 and involves intercalation of this drug into DNA. Furthermore, it has been shown that daunorubicin has high affinity for the A-T regions of DNA18'19 X-ray crystallographic16 studies of daunorubicin-DNA complexes have shown that the intercalation proceeds through the major groove of the DNA. The displacement studies with actinomycin D, an agent known to bind to G-C rich regions and in the minor groove of the DNA, show that I has higher affinity for the G-C regions. The relatively small displacement of both I and II by AMD may be due to the high affinity of the acridine labels for DNA. The binding of I and II to the synthetic polymers over a wide range of P/D ratios further substantiates the idea that II has higher binding affinity for the polydA-polydT and binds to these A-T rich regions in the native DNA. At low P/D ratios where both ionic (phosphate and groove) and intercalative binding siles are occupied, the addition of H1 caused displacement of bound I and II from DNA. When binding to the ionic sites at the phosphate groups was prevented by the addition of NaCl, only bound II could be displaced by H1. This finding suggest that II and H1 may be occupying the same binding sites. Our results on binding studies with I and II in the DNA-H1 complexes over different P/D ratios show that H1 decreased the binding of II to DNA. This indicated that H1 was binding to the same sites where II binds to DNA. At high P/D ratio, where I and II are bound only by intercalation, our results indicate that H1, does not affect or bind to these sites. Based on the thermal denaturation studies with nucleohistones, 3711
Nucleic Acids Research Ohba21 has suggested that histones interact with the A-T rich regions of DNA. Energy calculation studies on protein-DNA binding have shown that the interaction .of lysine residues with the A-T regions of DNA provides the most stable complex with slight stacking of the DNA bases.22 Furthermore, Shapiro, et al.23 have shown that poly-L-lysine interacts exclusively with the A-T bases of DNA. Recently, Goodwin, et al.24 have shown that the nonhistone proteins are located in the major groove of DNA. Our ESR results do not distinguish unequivocally between the minor or the major groove of DNA, thus binding of H1 could involve either groove. However, our ESR results do show that label II has a high affinity for polydA-polydT and binds to these regions in native DNA. In addition, we have shown that II and daunorubicin occupy the same binding sites in DNA. Since the binding of daunorubicin is well established and involves the A-T bases and the major groove of DNA, label II may also intercalate in the major groove. Since H1 displaces only bound II and reduced the binding of II with DNA, we believe that H1 must also bind to the A-T rich regions of DNA, the same binding sites occupied by II. Thus, the binding of H 1 with the native DNA must involve chargecharge interactions between the lysine residues of H1 and the functional groups of the major groups of the DNA with a defined sequence of A-T base pairs. These results are interesting in connection with the findings of Votavova, et al.25 who have suggested that H1 binds to the major groove of DNA.
*To whom all correspondence should be addressed
Presented in part at the 7th International Congress of Pharmacology, July 16-21, 1978, Paris, France.
2Proctor and Gamble Company, Ivorydale Technical Center, Professional and Regulatory Services Division, Cincinnati, Ohio 45217
REFERENCES 1. Oudet, M., Gross-Bellard, M. and Chambon, P. (1975) Cell 4: 281. 2. Elgin, S.C.R. and Weintraub, H. (1975) Ann. Rev. Biochem. 44: 725. 3. Kornberg, R.D. (1974) Science 184: 868. 3712
Nucleic Acids Research 4. Van Holde, K.E., Sharsrabuddhe, C.G. and Shaw, B.R. (1974) Nucleic Acids Res. 1: 1579. 5. Varshavsky, A.J., Bakayev, V.V. and Gieorgiev, G.P. (1976) Nucleic Acids Res. 3: 477. 6. Noll, M.FanU Kornberg, R.D. (1977) J. Mol. Biol. 109: 393. 7. Muller, V., Zentgraf, H., Eicken, TT.anidKeller, W(1978) Science 201: 406. 8. 7iU;ppers, R., Otto, B. and Bohme, R. (1978) Nucleic Acids Res. 5: 2113. 9. Edwards, G.S. and Allfrey, V.G. (1973) Biochem. Biophys. Acta 299: 354. 10. McConnell, H.M. and McFarland, B.R. (1970) W Rev. Biophys. 3: 91. 11. Griffith, O.H., Jost, P. (1972) in Methods in Pharmiacolo, Vol. 2, pp.223-276, Appleton-Century-Croft, or 12. Chignell, C.F., Starkweather, D.K. and Sinha, B.K. (1975) J. Biol. Chem. 250: 5622. 13. Sinha,T?K., Cysyk, R., Millar, D. and Chignell, C.F. (1976) J. Med. Chem. 19: 994. 14. Sinha,.K. and Chignell, C.F. (1975) Life Sciences 17: 1829. 15. Johns, E.W. and Butler, J.A.V. (1964) Nature204: 8W3. 16. Pigram, W.J., Fuller, W. and Hamilton,LT7.D(T7M2) Nature 235: 17. 17. Zunino, F., Gambetta, R.A., DiMarco, A. and Zaccara, A. (1972) Biochem. Biophys. Acta 277: 489. 18. Kersten, W. and Kersten, H. (1968) in "Molecular Association in iology, p. 289, Academic Press, New York. 19. Chandra, P., Zunino, F., Gotz, A., Gericke, D., Thorbeck, R., and DiMarco, A. 1972) FEBS Letters 21: 264. 20. Sobel, H.M. 1974) Canc Reports 58: 101. 21. Ohba, Y. (1976) Biochem. Biophs. Acta 234. 22. DeSantis, P., FoiiTaii W Rizzo, TR.(T74) Biopol ymers 13: 313. 23. Shapiro, J.T., Leng, M. and Felsenfeld, G. (1W969-) iBheminstry 8: 31219. 24. Goodwin, D.C. and Brahms, J. (1979) Nucleic Acids Res. 5: 835. 25. Votavova, H., Blah, K. and Sponar, J. (1978) StudiaBophysics 67: 79.
3713
