- Email: [email protected]
Interactions of mammalian proteins with cisplatin-damaged DNA
www.elsevier.nl/locate/jinorgbio Journal of Inorganic Biochemistry 77 (1999) 83–87
Interactions of mammalian proteins with cisplatin-damaged DNA John J. Turchi *, Karen M. Henkels, Ingrid L. Hermanson, Steve M. Patrick Department of Biochemistry and Molecular Biology, Wright State University School of Medicine, Dayton, OH 45435, USA
Abstract We have undertaken the systematic isolation and characterization of mammalian proteins which display an affinity for cisplatin-damaged DNA. Fractionation of human cell extracts has led to the identification of two classes of proteins. The first includes proteins that bind duplex DNA in the absence of cisplatin damage and retain their affinity for DNA in the presence of cisplatin–DNA adducts. The DNA-dependent protein kinase (DNA-PK) falls into this class. The inhibition of DNA-PK phosphorylation activity by cisplatin-damaged DNA has led to the hypothesis that cisplatin sensitization of mammalian cells to ionizing radiation may be mediated by DNA-PK. The second class of proteins identified are those which display a high relative affinity for cisplatin-damaged DNA and a low affinity for undamaged duplex DNA. Proteins that fall into this class include high mobility group 1 protein (HMG-1), replication protein A (RPA) and xeroderma pigmentosum group A protein (XPA). Each protein has been isolated and purified in the lab. The interaction of each protein with cisplatin-damaged DNA has been assessed in electrophoretic mobility shift assays. A series of DNA binding experiments suggests that RPA binds duplex DNA via denaturation and subsequent preferential binding to the undamaged DNA strand of the partial duplex. DNA substrates prepared with photo-reactive base analogs on either the damaged or undamaged DNA strand have also been employed to investigate the mechanism and specific protein–DNA interactions that occur as each protein binds to cisplatin-damaged DNA. Results suggest both damage and strand specificity for RPA and XPA binding cisplatin-damaged DNA. q1999 Elsevier Science Inc. All rights reserved. Keywords: Mammalian proteins; DNA; Cisplatin-damaged DNA
1. Isolation of mammalian cisplatin-damaged DNA binding proteins The repair of cisplatin-induced DNA damage has received much attention, in part as a result of findings demonstrating that the inability to repair cisplatin–DNA adducts results in hypersensitivity to cisplatin. Repair of cisplatin–DNA damage proceeds via the nucleotide excision repair (NER) pathway which requires the coordinated action of greater than 30 proteins. The first step in the NER pathway is recognition of the damaged DNA and this can be accomplished by two mechanisms. The first couples repair to transcription and provides a mechanism to assure that active genes, and more specifically, the transcribed strand of active genes are efficiently repaired [1]. The second pathway for recognizing damaged DNA relies on specific proteins or protein complexes that display a greater affinity for binding damaged DNA compared to undamaged duplex DNA. A number of proteins have been implicated in this process and the elucidation of the damaged DNA binding activity has been * Corresponding author. Tel.: q1-937-775-2853; fax: q1-937-775-3730; e-mail: [email protected]
reported for many of them [2]. We have undertaken the systematic isolation and characterization of cisplatin–DNA binding proteins from human cells focusing not only on proteins that display a higher affinity for cisplatin-damaged DNA but on proteins that have a duplex DNA binding activity which is not altered by the presence of cisplatin–DNA adducts. Fractionation of cell-free extracts is accomplished initially using the scheme depicted in Fig. 1. A cisplatindamaged DNA-Sepharose column is prepared using calf thymus DNA which was treated in vitro with cisplatin, purified from unreacted cisplatin and then covalently linked to Sepharose beads. The matrix is packed into a column and cell-free extracts applied to the column under low ionic strength. The unbound protein is washed through the column with low ionic strength buffer and then bound proteins are eluted with buffer containing 0.75 M NaCl. The ability to bind cisplatin-damaged DNA is assessed using electrophoretic mobility shift assays (EMSAs). A representative electrophoretic assay is depicted in Fig. 2 and reveals the presence of two protein–DNA complexes with different relative mobilities and characteristics with respect to binding duplex DNA.
0162-0134/99/$ - see front matter q1999 Elsevier Science Inc. All rights reserved. PII S 0 1 6 2 - 0 1 3 4 ( 9 9 ) 0 0 1 4 5 - 2
Thursday Nov 25 03:15 PM
StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)
Article: 6243
84
J.J. Turchi et al. / Journal of Inorganic Biochemistry 77 (1999) 83–87
Fig. 1. Fractionation of cisplatin–DNA binding proteins. Calf thymus DNA was treated with cisplatin, purified and covalently linked to activated Sepharose beads. The matrix was washed and packed into a column. HeLa cell-free extracts were loaded on the column in buffer containing 0.1 M NaCl and the column washed extensively. Bound proteins were eluted in buffer containing 0.75 M NaCl and then assayed for DNA binding activity and further fractionated by column chromatography.
as no binding is observed in the absence of cisplatin. Following the purification of DRP-2, identification of the protein responsible for this cisplatin–DNA binding activity revealed it was HMG-1 [3]. HMG-1 is a non-histone chromosomal associated protein with a tripartite organization [4]. Two tandem HMG-1 boxes (A and B) are responsible for DNA binding and are followed by a C-terminal acidic region. The HMG-1 protein, other proteins with the HMG-1 box domains and isolated HMG-1 boxes have been shown to bind to cisplatin-damaged DNA in a number of assays [5,6]. A recent study has demonstrated that the rate of association of the Bbox of HMG-1 with a duplex DNA containing a single cisplatin 1,2-d(GpG) adduct is near diffusion limited [7]. In addition, HMG-1 proteins can serve to shield cisplatin adducts from incision by the NER machinery [8], which is likely to proceed via blocking of the initial recognition of the DNA damage [9]. Fig. 2. EMSA identification of cisplatin–DNA binding proteins. A 44-bp duplex DNA was treated with varying concentrations of cisplatin, and the platinum-damaged DNA was purified. Protein binding to the DNA substrates containing varying degrees of cisplatin–DNA damage was assessed by EMSA [3]. (Reproduced with permission from Biochemistry. Copyright: 1996, American Chemical Society.)
2. HMG-1 The damaged-DNA recognition protein-2 (DRP-2) requires cisplatin damage to promote binding to duplex DNA
Thursday Nov 25 03:15 PM
3. DNA-PK and a possible role in cisplatin sensitization to ionizing radiation The protein responsible for the DRP-1 binding activity was also purified and was identified as the Ku subunits of DNAdependent protein kinase [10]. The dramatic finding was that, while Ku bound to cisplatin damaged DNA, phospho-
StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)
Article: 6243
J.J. Turchi et al. / Journal of Inorganic Biochemistry 77 (1999) 83–87
rylation of target proteins by the catalytic subunit on DNAPK (DNA-PKcs) was drastically reduced. Inhibition was observed using a variety of cisplatin-damaged DNA substrates which was proportional to the degree of DNA damage. Kinetic analyses revealed a minimal increase in Km for Ku binding and a decrease in kcat for kinase turnover [11]. The inability of DNA-PKcs in vitro to catalyze phosphorylation of target proteins using cisplatin-damaged DNA as a cofactor is especially interesting in light of the requirement of this activity for DSB repair and V(D)J recombination [12]. In addition, cells that contain mutations in any of the three DNAPK subunits display hypersensitivity to ionizing radiation (IR) [12]. Considering that cisplatin is used clinically in conjunction with IR for treating numerous cancers [13] we hypothesize that the mechanism of cisplatin sensitization to IR is via the inhibition of DNA-PK-catalyzed phosphorylation of target proteins. Consistent with this hypothesis, we have also shown that transplatin-damaged DNA inhibits DNA-PK-catalyzed phosphorylation of target proteins in vitro (Fig. 3). Transplatin is also an effective radiosensitizer [13], and, as transplatin is relatively non-toxic, it demonstrates that the cytotoxic activity of the platinum analogs is not required for sensitization to IR. The presence of cisplatin– DNA adducts positioned near double-strand breaks would then enable the Ku subunits to bind while being unable to support kinase activity. The decrease in DNA-PK-catalyzed phosphorylation of target proteins results in a defect in DSB repair. Very recent data, in fact, demonstrate that the repair of DSBs is decreased in cells treated with cisplatin prior to irradiation [14]. Experiments are under way to directly test the hypothesis that the disruption of the DSB repair pathway in cisplatin-treated cells is via inhibition of DNA-PK.
4. RPA and XPA: identifying damaged DNA for NER In addition to DNA-PK and HMG-1, two other proteins were identified by retention on the cisplatin–DNA Sepharose column, RPA and XPA. As measured by immunoblot analysis, approximately 50% of the XPA from HeLa cells was retained on the cisplatin-Sepharose column. The bound XPA was eluted and purified using a combination of ion-exchange, adsorption and affinity matrices. The final pool contained approximately 30 mg of protein at about 50% purity. These results are consistent with those demonstrating the low level of XPA in mammalian cells. Considering the low level of expression and the finding that the NER process can be regulated by post-translational modification, we sought to express recombinant human XPA in a eukaryotic expression system. Recombinant baculovirus was prepared expressing the human XPA gene as an N-terminal [His]6 tagged fusion protein. Infection of Sf-21 insect cells with the recombinant baculovirus results in the production of greater than 0.5 mg of protein approximately 95% pure from a 200 ml culture of cells using a two-column purification procedure [15].
Thursday Nov 25 03:15 PM
85
Fig. 3. Effect of different platinum analogs on DNA-PK phosphorylation activity. The 75-mer duplex DNA [3] was left untreated (filled circles) or treated with equal concentrations of either cisplatin (open circles) or transplatin (filled triangles). The DNA substrates were purified and assessed for their ability to support DNA-PK phosphorylation activity as previously described [10,11].
A novel migrating protein–DNA complex initially identified as DRP-3 was purified and was subsequently identified as RPA [9]. Characterization of the interaction of RPA with cisplatin-damaged DNA revealed that, under reaction conditions containing 50 mM NaCl and 2 mM MgCl2, preferential binding to a duplex cisplatin-damaged DNA was achieved. In a series of competition binding experiments combining RPA and HMG-1, we demonstrated that HMG-1 binding to cisplatin-damaged DNA precluded the binding of RPA. These results provide evidence that shielding of cisplatin-damaged DNA by HMG-1 interferes with the initial recognition of the damage [9]. These results are consistent with the ability of HMG proteins to inhibit in vitro incision of cisplatin-damaged DNA [8] and suggest that the mechanism of inhibition is by blocking the initial recognition [9]. Results obtained from kinetic analyses of binding in competition assays also suggested that the rate of association of HMG-1 with cisplatin-damaged DNA was greater than that of RPA for the same DNA substrate. Further analyses of the damaged DNA binding ability of RPA revealed differential recognition of site-specific cisplatin–DNA adducts [16]. The adduct bound most efficiently is the 1,3-d(GpXpG) adduct, with the 1,2-d(GpG) adduct also being bound preferentially compared to the undamaged control, but approximately 30% less compared to the 1,3 adduct. A DNA substrate containing a single cisplatin-interstrand crosslink was bound least efficiently, and in fact, the binding was less than that observed with the undamaged control DNA. The differential binding was correlated with the relative thermal instability of the duplex DNA. The 1,3 adduct results in localized denaturation around the adduct, whereas the 1,2 adduct lowers the melting temperature of the duplex, while all the bases remain hydrogen bonded. The interstrand adduct dramatically increases the Tm as denaturation is abrogated by the crosslinking of the two complementary DNA strands. The correlation of RPA binding and the instability of the duplex
StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)
Article: 6243
86
J.J. Turchi et al. / Journal of Inorganic Biochemistry 77 (1999) 83–87
suggests that RPA may bind to these DNA substrates via the generation of transient single-stranded regions. RPA has been shown to denature DNA in an ATP-independent manner [17,18]. In a series of experiments, the ability to bind duplex DNA correlated directly with the ability to denature the duplex and generate single-stranded DNA. This two-step mode of recognizing duplex-damaged DNA, low affinity binding to the duplex, generation of single-stranded DNA followed by high affinity binding to the single-stranded DNA could account for the slower rate of association with cisplatindamaged DNA compared to HMG-1. Also, consistent with these results, a recent report demonstrated that the rate of association of an HMG-1 box with cisplatin-damaged DNA was near diffusion limited [7]. The preferential binding of RPA to undamaged singlestranded DNA suggests that, in the context of NER, RPA would bind the undamaged stand with a specific polarity such that the N-terminus of the 70 kDa subunit is 59 and the Cterminus is 39 on the undamaged strand. Preliminary experiments using a cisplatin-damaged DNA substrate containing site-specific zero-distance crosslinkers confirmed the ability of RPA to bind the undamaged strand of the cisplatin-damaged duplex. A model for the interaction of XPA and RPA with damaged DNA is reproduced in Fig. 4 and is based on the preferential binding of RPA to the undamaged DNA strand and XPA to the damaged DNA strand. The polarity of RPA binding single-stranded DNA is also incorporated into the model where the N-terminus is located 59 on the DNA and the C-terminus is 39 [19]. The assembly of the NER incision proteins is also included and the positioning is based on the interaction of XPG, which incises 39 on the damaged strand, with the N-terminus of the RPA p70 subunit and the interaction of XPF/ERCC1, which incises 59 on the damaged strand, with the XPA protein [20]. Following incision, the RPA protein is positioned for supporting DNA polymerasecatalyzed gap-filling synthesis with the p34 subunit in close proximity to the 39 OH that is to be extended [21].
5. Conclusions The study of mammalian cisplatin–DNA binding proteins has resulted in a series of important observations. Results supporting the hypothesis that proteins not directly involved in NER can alter the repair of specific cisplatin adducts has direct implications on increasing the cytotoxic efficacy of platinum-based drugs. Despite the wealth of information on specific mechanisms of mammalian NER, there are still many unanswered questions including how global genomic cisplatin damage is recognized. While numerous experiments implicate both RPA and XPA, there is convincing evidence that the XPC-HHR23B protein is also involved in the process [22]. In addition, there is the possibility that different types of DNA damage are recognized by different sub-assemblies of the NER proteins. Our data support the involvement of RPA in recognizing DNA adducts that result in the thermo-
Thursday Nov 25 03:15 PM
Fig. 4. Model for RPA-directed recognition of cisplatin-damaged DNA. (Reproduced with permission from the Journal of Biological Chemistry. Copyright: 1999, ASBMB.)
dynamic instability in the duplex DNA such that the formation of single-stranded DNA is generated. Other types of DNA damage do not result in a decrease in the melting temperature of duplex DNA. The repair of these adducts may be initiated by other proteins including XPA and the XPCHHR23B complex. The differential rates of repair for different adducts and those with different chromosomal positions may then represent the number of pathways that lead to the identification of the damage. Thus far, transcription coupling, RPA–XPA and XPC–HHR23B all remain viable pathways to recognize DNA damage. A very interesting finding was that proteins involved in the mismatch repair (MMR) pathway also display an affinity for cisplatin-damaged DNA. Specifically, human MutS a can bind to DNA containing a 1,2-intrastrand d(GpG) adduct [23]. This is especially interesting in the light of more recent data demonstrating that MMR deficient cells display a decreased sensitivity to cisplatin [24]. In addition, MMR deficient cells have been reported to be defective in the transcription-coupled NER pathway [25]. Definitive in vivo determination of the subset of proteins that are involved in recognizing cisplatin-induced DNA damage and how non-NER proteins affect recognition awaits further investigation.
Acknowledgements This work was supported by a National Institutes of Health Award CA64374 to J.J.T.
References [1] P.C. Hanawalt, DNA Damage and Repair, vol. 2, DNA Repair in Higher Eukaryotes, Humana Press, Totowa, NJ, 1998, pp. 1–8. [2] J. Zlatanova, J. Yaneva, S.H. Leuba, FASEB J. 12 (1998) 791–799. [3] J.J. Turchi, M. Li, K.M. Henkels, Biochemistry 35 (1996) 2992– 3000. [4] A.D. Baxevanis, D. Landsman, Nucleic Acids Res. 23 (1995) 1604– 1613. [5] P.M. Pil, C.S. Chow, S.J. Lippard, Proc. Natl. Acad. Sci. USA 90 (1993) 9465–9469. [6] E.N. Hughes, B.N. Engelsberg, P.C. Billings, J. Biol. Chem. 267 (1992) 13520–13527. [7] E.R. Jamieson, M.P. Jacobson, C.M. Barnes, C.S. Chow, S.J. Lippard, J. Biol. Chem. 274 (1999) 12346–12354. [8] J.C. Huang, D.B. Zamble, J.T. Reardon, S.J. Lippard, A. Sancar, Proc. Natl. Acad. Sci. USA 91 (1994) 10394–10398.
StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)
Article: 6243
J.J. Turchi et al. / Journal of Inorganic Biochemistry 77 (1999) 83–87 [9] S.M. Patrick, J.J. Turchi, Biochemistry 37 (1998) 8808–8815. [10] J.J. Turchi, K.M. Henkels, J. Biol. Chem. 271 (1996) 13861–13867. [11] J.J. Turchi, S.M. Patrick, K.M. Henkels, Biochemistry 36 (1997) 7586–7593. [12] G.C. Smith, S.P. Jackson, Gene Dev. 13 (1999) 916–934. [13] A.C. Begg, F.A. Stewart, L. Dewit, H. Bartelink, in: B.T. Hill, A.S. Bellamy (Eds.), Antitumor Drug–Radiation Interactions, CRC Press, Boca Raton, FL, 1990, pp. 153–170. [14] J.A. Dolling, D.R. Boreham, D.L. Brown, R.E.J. Mitchel, G.P. Raaphorst, Int. J. Radiat. Biol. 74 (1998) 61–69. [15] I.L. Hermanson, J.J. Turchi, Overexpression and purification of human XPA using a baculovirus expression system, Protein Expression and Purification, submitted for publication. [16] S.M. Patrick, J.J. Turchi, J. Biol. Chem. 274 (1999) 14972–14978. [17] A. Georgaki, U. Hubscher, Nucleic Acids Res. 21 (1993) 3659–3665. [18] K. Treuner, U. Ramsperger, R. Knippers, J. Mol. Biol. 259 (1996) 104–112.
Thursday Nov 25 03:15 PM
87
[19] W.L. de Laat, E. Appeldoorn, K. Sugasawa, E. Weterings, N.G.J. Jaspers, J.H.J. Hoeijmakers, Gene Dev. 12 (1998) 2598–2609. [20] W.L. de Laat, N.G. Jaspers, J.H. Hoeijmakers, Gene Dev. 13 (1999) 768–785. [21] O.I. Lavrik, H.P. Nasheuer, K. Weisshart, M.S. Wold, R. Prasad, W.A. Beard, S.H. Wilson, A. Favre, Nucleic Acids Res. 26 (1998) 602– 607. [22] K. Sugasawa, J.M.Y. Hg, C. Masutani, S. Iwai, P.J. van der Spek, A.P.M. Eker, F. Hanaoka, D. Bootsma, J.H.J. Hoeijmakers, Mol. Cell 2 (1998) 223–232. [23] D. Duckett, J. Drummond, A. Murchie, J. Reardon, A. Sancar, D. Lilley, P. Modrich, Proc. Natl. Acad. Sci. USA 93 (1996) 6443–6447. [24] J. Drummond, A. Anthoney, R. Brown, P. Modrich, J. Biol. Chem. 271 (1996) 19645–19648. [25] I. Mellon, D.K. Rajpal, M. Koi, C.R. Boland, G.N. Champe, Science 272 (1996) 557–560.
StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)
Article: 6243
Interactions of mammalian proteins with cisplatin-damaged DNA John J. Turchi *, Karen M. Henkels, Ingrid L. Hermanson, Steve M. Patrick Department of Biochemistry and Molecular Biology, Wright State University School of Medicine, Dayton, OH 45435, USA
Abstract We have undertaken the systematic isolation and characterization of mammalian proteins which display an affinity for cisplatin-damaged DNA. Fractionation of human cell extracts has led to the identification of two classes of proteins. The first includes proteins that bind duplex DNA in the absence of cisplatin damage and retain their affinity for DNA in the presence of cisplatin–DNA adducts. The DNA-dependent protein kinase (DNA-PK) falls into this class. The inhibition of DNA-PK phosphorylation activity by cisplatin-damaged DNA has led to the hypothesis that cisplatin sensitization of mammalian cells to ionizing radiation may be mediated by DNA-PK. The second class of proteins identified are those which display a high relative affinity for cisplatin-damaged DNA and a low affinity for undamaged duplex DNA. Proteins that fall into this class include high mobility group 1 protein (HMG-1), replication protein A (RPA) and xeroderma pigmentosum group A protein (XPA). Each protein has been isolated and purified in the lab. The interaction of each protein with cisplatin-damaged DNA has been assessed in electrophoretic mobility shift assays. A series of DNA binding experiments suggests that RPA binds duplex DNA via denaturation and subsequent preferential binding to the undamaged DNA strand of the partial duplex. DNA substrates prepared with photo-reactive base analogs on either the damaged or undamaged DNA strand have also been employed to investigate the mechanism and specific protein–DNA interactions that occur as each protein binds to cisplatin-damaged DNA. Results suggest both damage and strand specificity for RPA and XPA binding cisplatin-damaged DNA. q1999 Elsevier Science Inc. All rights reserved. Keywords: Mammalian proteins; DNA; Cisplatin-damaged DNA
1. Isolation of mammalian cisplatin-damaged DNA binding proteins The repair of cisplatin-induced DNA damage has received much attention, in part as a result of findings demonstrating that the inability to repair cisplatin–DNA adducts results in hypersensitivity to cisplatin. Repair of cisplatin–DNA damage proceeds via the nucleotide excision repair (NER) pathway which requires the coordinated action of greater than 30 proteins. The first step in the NER pathway is recognition of the damaged DNA and this can be accomplished by two mechanisms. The first couples repair to transcription and provides a mechanism to assure that active genes, and more specifically, the transcribed strand of active genes are efficiently repaired [1]. The second pathway for recognizing damaged DNA relies on specific proteins or protein complexes that display a greater affinity for binding damaged DNA compared to undamaged duplex DNA. A number of proteins have been implicated in this process and the elucidation of the damaged DNA binding activity has been * Corresponding author. Tel.: q1-937-775-2853; fax: q1-937-775-3730; e-mail: [email protected]
reported for many of them [2]. We have undertaken the systematic isolation and characterization of cisplatin–DNA binding proteins from human cells focusing not only on proteins that display a higher affinity for cisplatin-damaged DNA but on proteins that have a duplex DNA binding activity which is not altered by the presence of cisplatin–DNA adducts. Fractionation of cell-free extracts is accomplished initially using the scheme depicted in Fig. 1. A cisplatindamaged DNA-Sepharose column is prepared using calf thymus DNA which was treated in vitro with cisplatin, purified from unreacted cisplatin and then covalently linked to Sepharose beads. The matrix is packed into a column and cell-free extracts applied to the column under low ionic strength. The unbound protein is washed through the column with low ionic strength buffer and then bound proteins are eluted with buffer containing 0.75 M NaCl. The ability to bind cisplatin-damaged DNA is assessed using electrophoretic mobility shift assays (EMSAs). A representative electrophoretic assay is depicted in Fig. 2 and reveals the presence of two protein–DNA complexes with different relative mobilities and characteristics with respect to binding duplex DNA.
0162-0134/99/$ - see front matter q1999 Elsevier Science Inc. All rights reserved. PII S 0 1 6 2 - 0 1 3 4 ( 9 9 ) 0 0 1 4 5 - 2
Thursday Nov 25 03:15 PM
StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)
Article: 6243
84
J.J. Turchi et al. / Journal of Inorganic Biochemistry 77 (1999) 83–87
Fig. 1. Fractionation of cisplatin–DNA binding proteins. Calf thymus DNA was treated with cisplatin, purified and covalently linked to activated Sepharose beads. The matrix was washed and packed into a column. HeLa cell-free extracts were loaded on the column in buffer containing 0.1 M NaCl and the column washed extensively. Bound proteins were eluted in buffer containing 0.75 M NaCl and then assayed for DNA binding activity and further fractionated by column chromatography.
as no binding is observed in the absence of cisplatin. Following the purification of DRP-2, identification of the protein responsible for this cisplatin–DNA binding activity revealed it was HMG-1 [3]. HMG-1 is a non-histone chromosomal associated protein with a tripartite organization [4]. Two tandem HMG-1 boxes (A and B) are responsible for DNA binding and are followed by a C-terminal acidic region. The HMG-1 protein, other proteins with the HMG-1 box domains and isolated HMG-1 boxes have been shown to bind to cisplatin-damaged DNA in a number of assays [5,6]. A recent study has demonstrated that the rate of association of the Bbox of HMG-1 with a duplex DNA containing a single cisplatin 1,2-d(GpG) adduct is near diffusion limited [7]. In addition, HMG-1 proteins can serve to shield cisplatin adducts from incision by the NER machinery [8], which is likely to proceed via blocking of the initial recognition of the DNA damage [9]. Fig. 2. EMSA identification of cisplatin–DNA binding proteins. A 44-bp duplex DNA was treated with varying concentrations of cisplatin, and the platinum-damaged DNA was purified. Protein binding to the DNA substrates containing varying degrees of cisplatin–DNA damage was assessed by EMSA [3]. (Reproduced with permission from Biochemistry. Copyright: 1996, American Chemical Society.)
2. HMG-1 The damaged-DNA recognition protein-2 (DRP-2) requires cisplatin damage to promote binding to duplex DNA
Thursday Nov 25 03:15 PM
3. DNA-PK and a possible role in cisplatin sensitization to ionizing radiation The protein responsible for the DRP-1 binding activity was also purified and was identified as the Ku subunits of DNAdependent protein kinase [10]. The dramatic finding was that, while Ku bound to cisplatin damaged DNA, phospho-
StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)
Article: 6243
J.J. Turchi et al. / Journal of Inorganic Biochemistry 77 (1999) 83–87
rylation of target proteins by the catalytic subunit on DNAPK (DNA-PKcs) was drastically reduced. Inhibition was observed using a variety of cisplatin-damaged DNA substrates which was proportional to the degree of DNA damage. Kinetic analyses revealed a minimal increase in Km for Ku binding and a decrease in kcat for kinase turnover [11]. The inability of DNA-PKcs in vitro to catalyze phosphorylation of target proteins using cisplatin-damaged DNA as a cofactor is especially interesting in light of the requirement of this activity for DSB repair and V(D)J recombination [12]. In addition, cells that contain mutations in any of the three DNAPK subunits display hypersensitivity to ionizing radiation (IR) [12]. Considering that cisplatin is used clinically in conjunction with IR for treating numerous cancers [13] we hypothesize that the mechanism of cisplatin sensitization to IR is via the inhibition of DNA-PK-catalyzed phosphorylation of target proteins. Consistent with this hypothesis, we have also shown that transplatin-damaged DNA inhibits DNA-PK-catalyzed phosphorylation of target proteins in vitro (Fig. 3). Transplatin is also an effective radiosensitizer [13], and, as transplatin is relatively non-toxic, it demonstrates that the cytotoxic activity of the platinum analogs is not required for sensitization to IR. The presence of cisplatin– DNA adducts positioned near double-strand breaks would then enable the Ku subunits to bind while being unable to support kinase activity. The decrease in DNA-PK-catalyzed phosphorylation of target proteins results in a defect in DSB repair. Very recent data, in fact, demonstrate that the repair of DSBs is decreased in cells treated with cisplatin prior to irradiation [14]. Experiments are under way to directly test the hypothesis that the disruption of the DSB repair pathway in cisplatin-treated cells is via inhibition of DNA-PK.
4. RPA and XPA: identifying damaged DNA for NER In addition to DNA-PK and HMG-1, two other proteins were identified by retention on the cisplatin–DNA Sepharose column, RPA and XPA. As measured by immunoblot analysis, approximately 50% of the XPA from HeLa cells was retained on the cisplatin-Sepharose column. The bound XPA was eluted and purified using a combination of ion-exchange, adsorption and affinity matrices. The final pool contained approximately 30 mg of protein at about 50% purity. These results are consistent with those demonstrating the low level of XPA in mammalian cells. Considering the low level of expression and the finding that the NER process can be regulated by post-translational modification, we sought to express recombinant human XPA in a eukaryotic expression system. Recombinant baculovirus was prepared expressing the human XPA gene as an N-terminal [His]6 tagged fusion protein. Infection of Sf-21 insect cells with the recombinant baculovirus results in the production of greater than 0.5 mg of protein approximately 95% pure from a 200 ml culture of cells using a two-column purification procedure [15].
Thursday Nov 25 03:15 PM
85
Fig. 3. Effect of different platinum analogs on DNA-PK phosphorylation activity. The 75-mer duplex DNA [3] was left untreated (filled circles) or treated with equal concentrations of either cisplatin (open circles) or transplatin (filled triangles). The DNA substrates were purified and assessed for their ability to support DNA-PK phosphorylation activity as previously described [10,11].
A novel migrating protein–DNA complex initially identified as DRP-3 was purified and was subsequently identified as RPA [9]. Characterization of the interaction of RPA with cisplatin-damaged DNA revealed that, under reaction conditions containing 50 mM NaCl and 2 mM MgCl2, preferential binding to a duplex cisplatin-damaged DNA was achieved. In a series of competition binding experiments combining RPA and HMG-1, we demonstrated that HMG-1 binding to cisplatin-damaged DNA precluded the binding of RPA. These results provide evidence that shielding of cisplatin-damaged DNA by HMG-1 interferes with the initial recognition of the damage [9]. These results are consistent with the ability of HMG proteins to inhibit in vitro incision of cisplatin-damaged DNA [8] and suggest that the mechanism of inhibition is by blocking the initial recognition [9]. Results obtained from kinetic analyses of binding in competition assays also suggested that the rate of association of HMG-1 with cisplatin-damaged DNA was greater than that of RPA for the same DNA substrate. Further analyses of the damaged DNA binding ability of RPA revealed differential recognition of site-specific cisplatin–DNA adducts [16]. The adduct bound most efficiently is the 1,3-d(GpXpG) adduct, with the 1,2-d(GpG) adduct also being bound preferentially compared to the undamaged control, but approximately 30% less compared to the 1,3 adduct. A DNA substrate containing a single cisplatin-interstrand crosslink was bound least efficiently, and in fact, the binding was less than that observed with the undamaged control DNA. The differential binding was correlated with the relative thermal instability of the duplex DNA. The 1,3 adduct results in localized denaturation around the adduct, whereas the 1,2 adduct lowers the melting temperature of the duplex, while all the bases remain hydrogen bonded. The interstrand adduct dramatically increases the Tm as denaturation is abrogated by the crosslinking of the two complementary DNA strands. The correlation of RPA binding and the instability of the duplex
StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)
Article: 6243
86
J.J. Turchi et al. / Journal of Inorganic Biochemistry 77 (1999) 83–87
suggests that RPA may bind to these DNA substrates via the generation of transient single-stranded regions. RPA has been shown to denature DNA in an ATP-independent manner [17,18]. In a series of experiments, the ability to bind duplex DNA correlated directly with the ability to denature the duplex and generate single-stranded DNA. This two-step mode of recognizing duplex-damaged DNA, low affinity binding to the duplex, generation of single-stranded DNA followed by high affinity binding to the single-stranded DNA could account for the slower rate of association with cisplatindamaged DNA compared to HMG-1. Also, consistent with these results, a recent report demonstrated that the rate of association of an HMG-1 box with cisplatin-damaged DNA was near diffusion limited [7]. The preferential binding of RPA to undamaged singlestranded DNA suggests that, in the context of NER, RPA would bind the undamaged stand with a specific polarity such that the N-terminus of the 70 kDa subunit is 59 and the Cterminus is 39 on the undamaged strand. Preliminary experiments using a cisplatin-damaged DNA substrate containing site-specific zero-distance crosslinkers confirmed the ability of RPA to bind the undamaged strand of the cisplatin-damaged duplex. A model for the interaction of XPA and RPA with damaged DNA is reproduced in Fig. 4 and is based on the preferential binding of RPA to the undamaged DNA strand and XPA to the damaged DNA strand. The polarity of RPA binding single-stranded DNA is also incorporated into the model where the N-terminus is located 59 on the DNA and the C-terminus is 39 [19]. The assembly of the NER incision proteins is also included and the positioning is based on the interaction of XPG, which incises 39 on the damaged strand, with the N-terminus of the RPA p70 subunit and the interaction of XPF/ERCC1, which incises 59 on the damaged strand, with the XPA protein [20]. Following incision, the RPA protein is positioned for supporting DNA polymerasecatalyzed gap-filling synthesis with the p34 subunit in close proximity to the 39 OH that is to be extended [21].
5. Conclusions The study of mammalian cisplatin–DNA binding proteins has resulted in a series of important observations. Results supporting the hypothesis that proteins not directly involved in NER can alter the repair of specific cisplatin adducts has direct implications on increasing the cytotoxic efficacy of platinum-based drugs. Despite the wealth of information on specific mechanisms of mammalian NER, there are still many unanswered questions including how global genomic cisplatin damage is recognized. While numerous experiments implicate both RPA and XPA, there is convincing evidence that the XPC-HHR23B protein is also involved in the process [22]. In addition, there is the possibility that different types of DNA damage are recognized by different sub-assemblies of the NER proteins. Our data support the involvement of RPA in recognizing DNA adducts that result in the thermo-
Thursday Nov 25 03:15 PM
Fig. 4. Model for RPA-directed recognition of cisplatin-damaged DNA. (Reproduced with permission from the Journal of Biological Chemistry. Copyright: 1999, ASBMB.)
dynamic instability in the duplex DNA such that the formation of single-stranded DNA is generated. Other types of DNA damage do not result in a decrease in the melting temperature of duplex DNA. The repair of these adducts may be initiated by other proteins including XPA and the XPCHHR23B complex. The differential rates of repair for different adducts and those with different chromosomal positions may then represent the number of pathways that lead to the identification of the damage. Thus far, transcription coupling, RPA–XPA and XPC–HHR23B all remain viable pathways to recognize DNA damage. A very interesting finding was that proteins involved in the mismatch repair (MMR) pathway also display an affinity for cisplatin-damaged DNA. Specifically, human MutS a can bind to DNA containing a 1,2-intrastrand d(GpG) adduct [23]. This is especially interesting in the light of more recent data demonstrating that MMR deficient cells display a decreased sensitivity to cisplatin [24]. In addition, MMR deficient cells have been reported to be defective in the transcription-coupled NER pathway [25]. Definitive in vivo determination of the subset of proteins that are involved in recognizing cisplatin-induced DNA damage and how non-NER proteins affect recognition awaits further investigation.
Acknowledgements This work was supported by a National Institutes of Health Award CA64374 to J.J.T.
References [1] P.C. Hanawalt, DNA Damage and Repair, vol. 2, DNA Repair in Higher Eukaryotes, Humana Press, Totowa, NJ, 1998, pp. 1–8. [2] J. Zlatanova, J. Yaneva, S.H. Leuba, FASEB J. 12 (1998) 791–799. [3] J.J. Turchi, M. Li, K.M. Henkels, Biochemistry 35 (1996) 2992– 3000. [4] A.D. Baxevanis, D. Landsman, Nucleic Acids Res. 23 (1995) 1604– 1613. [5] P.M. Pil, C.S. Chow, S.J. Lippard, Proc. Natl. Acad. Sci. USA 90 (1993) 9465–9469. [6] E.N. Hughes, B.N. Engelsberg, P.C. Billings, J. Biol. Chem. 267 (1992) 13520–13527. [7] E.R. Jamieson, M.P. Jacobson, C.M. Barnes, C.S. Chow, S.J. Lippard, J. Biol. Chem. 274 (1999) 12346–12354. [8] J.C. Huang, D.B. Zamble, J.T. Reardon, S.J. Lippard, A. Sancar, Proc. Natl. Acad. Sci. USA 91 (1994) 10394–10398.
StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)
Article: 6243
J.J. Turchi et al. / Journal of Inorganic Biochemistry 77 (1999) 83–87 [9] S.M. Patrick, J.J. Turchi, Biochemistry 37 (1998) 8808–8815. [10] J.J. Turchi, K.M. Henkels, J. Biol. Chem. 271 (1996) 13861–13867. [11] J.J. Turchi, S.M. Patrick, K.M. Henkels, Biochemistry 36 (1997) 7586–7593. [12] G.C. Smith, S.P. Jackson, Gene Dev. 13 (1999) 916–934. [13] A.C. Begg, F.A. Stewart, L. Dewit, H. Bartelink, in: B.T. Hill, A.S. Bellamy (Eds.), Antitumor Drug–Radiation Interactions, CRC Press, Boca Raton, FL, 1990, pp. 153–170. [14] J.A. Dolling, D.R. Boreham, D.L. Brown, R.E.J. Mitchel, G.P. Raaphorst, Int. J. Radiat. Biol. 74 (1998) 61–69. [15] I.L. Hermanson, J.J. Turchi, Overexpression and purification of human XPA using a baculovirus expression system, Protein Expression and Purification, submitted for publication. [16] S.M. Patrick, J.J. Turchi, J. Biol. Chem. 274 (1999) 14972–14978. [17] A. Georgaki, U. Hubscher, Nucleic Acids Res. 21 (1993) 3659–3665. [18] K. Treuner, U. Ramsperger, R. Knippers, J. Mol. Biol. 259 (1996) 104–112.
Thursday Nov 25 03:15 PM
87
[19] W.L. de Laat, E. Appeldoorn, K. Sugasawa, E. Weterings, N.G.J. Jaspers, J.H.J. Hoeijmakers, Gene Dev. 12 (1998) 2598–2609. [20] W.L. de Laat, N.G. Jaspers, J.H. Hoeijmakers, Gene Dev. 13 (1999) 768–785. [21] O.I. Lavrik, H.P. Nasheuer, K. Weisshart, M.S. Wold, R. Prasad, W.A. Beard, S.H. Wilson, A. Favre, Nucleic Acids Res. 26 (1998) 602– 607. [22] K. Sugasawa, J.M.Y. Hg, C. Masutani, S. Iwai, P.J. van der Spek, A.P.M. Eker, F. Hanaoka, D. Bootsma, J.H.J. Hoeijmakers, Mol. Cell 2 (1998) 223–232. [23] D. Duckett, J. Drummond, A. Murchie, J. Reardon, A. Sancar, D. Lilley, P. Modrich, Proc. Natl. Acad. Sci. USA 93 (1996) 6443–6447. [24] J. Drummond, A. Anthoney, R. Brown, P. Modrich, J. Biol. Chem. 271 (1996) 19645–19648. [25] I. Mellon, D.K. Rajpal, M. Koi, C.R. Boland, G.N. Champe, Science 272 (1996) 557–560.
StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)
Article: 6243