- Email: [email protected]
Interactions of cisplatin and transplatin with proteins
Journal of Inorganic Biochemistry 91 (2002) 306–311 www.elsevier.com / locate / jinorgbio
Interactions of cisplatin and transplatin with proteins. Comparison of binding kinetics, binding sites and reactivity of the Ptprotein adducts of cisplatin and transplatin towards biological nucleophiles Tal Peleg-Shulman, Yousef Najajreh, Dan Gibson*
,1
Department of Medicinal Chemistry and Natural Products, School of Pharmacy, PO Box 12065, The Hebrew University of Jerusalem, Jerusalem 91120, Israel Received 27 September 2001; received in revised form 26 November 2001; accepted 18 December 2001
Abstract In this manuscript we report on the interactions of cis-DDP (cisplatin, cis-diamminedichloroplatinum(II)) and trans-DDP (transplatin, trans-diamminedichloroplatinum(II)) with two model proteins, ubiquitin (Ub) and horse heart myoglobin (Mb), and attempt to answer the question whether proteins that have methionine-Pt adducts can transfer the platinum to biological nucleophiles and particularly to DNA. Our study shows that cisplatin and transplatin form different adducts with ubiquitin: transplatin forms one major adduct, trans[Pt(Ub)(NH 3 ) 2 Cl], while cisplatin forms four distinct adducts, [Pt(Ub)(NH 3 ) 2 Cl], [Pt(Ub)(NH 3 ) 2 (H 2 O)], [Pt(Ub)(NH 3 ) 2 ], and [Pt(Ub)(NH 3 )]. When binding ubiquitin, Met1 is the preferred binding site of cisplatin, but not of transplatin. Cisplatin binds faster than transplatin to both ubiquitin and horse heart myoglobin. Both cisplatin and transplatin adducts form stable ternary adducts when reacted with 59-guanosine monophosphate (59-GMP) or a tetranucleotide. No transfer of the Pt moiety from the proteins to the nucleotides was observed. Glutathione efficiently removes the platinum from preformed adducts of both cisplatin and transplatin with ubiquitin. 2002 Elsevier Science Inc. All rights reserved. Keywords: Cisplatin; Transplatin; Protein binding; Electrospray ionization mass spectrometry (ESI-MS)
1. Introduction Cisplatin (Fig. 1a) is a widely used anti-tumor agent that is used in the clinic to treat testicular and ovarian cancers
Fig. 1. (a) cisplatin; (b) transplatin.
*Corresponding author. Tel.: 1972-2-675-8702; fax: 1972-2-6757076. E-mail address: [email protected] (D. Gibson). 1 Affiliated with the David R. Bloom Center for Pharmacy at The Hebrew University of Jerusalem, Jerusalem, Israel.
[1]. Cisplatin is believed to induce apoptosis in cancer cells by covalently modifying the DNA [2]. Its geometric isomer, trans-[Pt(NH 3 ) 2 Cl 2 ] (Fig. 1b), has no cytotoxic activity [3]. The difference in antitumor activity between the two isomers is attributed to the inability of the trans isomer to form 1,2-GpG intrastrand crosslinks due the 1808 angle between its two semi-labile chloride ligands. Cellular proteins such as the HMG box proteins recognize cisplatin modified DNA and bind to it. The binding of the proteins to the platinated DNA can serve to protect the DNA from being repaired by nuclear excision repair enzymes, or alternatively, hijack the proteins from their normal function [4,5]. Transplatin modified DNA is not recognized by those same cellular proteins [6,7]. Cisplatin is administered intravenously, and within 1 day, 65–98% of the drug is bound to blood plasma proteins [8]. While Pt-DNA adducts are believed to be responsible for the drug’s cytotoxicity, the exact role that Pt-protein adducts play in the mechanism of action of the drug is yet to be elucidated. It has been postulated that in addition to drug inactivation, cisplatin binding to proteins may be the cause of many of the drug’s side-effects [9,10].
0162-0134 / 02 / $ – see front matter 2002 Elsevier Science Inc. All rights reserved. PII: S0162-0134( 02 )00362-8
T. Peleg-Shulman et al. / Journal of Inorganic Biochemistry 91 (2002) 306 – 311
Other reports suggest that Pt-HSA (human serum albumin) adducts may be important for the activity of the drug [11]. When preformed Pt-HSA adducts were administered clinically they increased the survival time of the patients. Also, patients with low levels of HSA did not respond well to cisplatin based chemotherapy [12]. It is clear that the formation of Pt-protein adducts, that effectively competes with formation of the cytotoxic Pt-DNA lesions, can reduce the efficacy of Pt antitumor agents. The efficient binding of platinum complexes to proteins and peptides is not surprising since platinum has a high affinity for sulfur containing ligands such as methionine and cysteine [13]. The abundance of extra- and intracellular platinophiles makes it difficult to understand how Pt even reaches the DNA. Model studies suggest that a platinum moiety that is bound to a methionine thioether can be transferred to the N7 of the guanine of single stranded and double stranded oligonucleotides (but not from the cysteine thiolate) [14–17]. On the basis of these studies it was suggested that proteins that form Ptmethionine adducts could act as a platinum reservoir for subsequent DNA platination [13]. Many studies on the reaction of platinum complexes with amino acids and proteins have been reported [18–20], but there are few high-resolution reports on the interactions of cis-DDP (cisplatin, cis-diamminedichloroplatinum(II)) or transDDP (transplatin, trans-diamminedichloroplatinum(II)) with proteins. We have recently shown that electrospray ionization mass spectrometry is an extremely useful technique for studying the interactions of cisplatin with proteins [21], and provides information on the nature of the Pt-protein adducts that are formed. [ 1 H, 15 N] heteronuclear single quantum coherence (HSQC) NMR spectroscopy has been utilized in the analysis of 15 N-labeled platinum ammine complexes providing information on the nature of the ligands that are trans to the 15 N-labeled ammine [22]. The combination of the two techniques provides a powerful tool for simultaneously studying the modifications of the protein (electrospray ionization mass spectrometry, ESIMS) and the changes in the platinum coordination sphere (HSQC) [23]. Since Pt-protein adducts are important in defining the therapeutic profiles of the drugs, it is important to understand the basic principles that govern the formation and reactivity of these adducts. In this manuscript we report on the interaction of cis- and trans-DDP with model proteins, ubiquitin (Ub) and horse heart myoglobin (Mb). These proteins were chosen because both are well characterized, have methionines and histidines (but not cysteines) and are good candidates to try and answer the question whether proteins that have methionine-Pt adducts can transfer the Pt to biological nucleophiles and particularly to DNA. The binding kinetics, the nature of the adducts formed, the binding sites of the complexes and the reactivity of the Pt-protein adducts towards biologically relevant nu-
307
cleophiles such as glutathione and 59-guanosine monophosphate are reported. 2. Experimental
2.1. Materials Cis- and trans-DDP, ubiquitin and horse heart myoglobin were all purchased from Sigma–Aldrich (Israel) and were used without further purification.
2.2. Platination reactions Platination reactions were carried out at 1–2 mM concentrations, in 10 mM phosphate buffer, pH 6.4, 37 8C. Excess platinum was removed by ultra-filtration using Microcon YM-3 centrifugal filter devices at 4 8C and 12,000 rpm, prior to all NMR and adduct-reactivity studies. Kinetics by ESI-MS were measured directly on the reaction mixtures following ZipTip姠 (C 18 , Millipore) treatment.
2.3. Reactions of Pt-protein adducts with nucleophiles A 2 mM concentration of Ub-Pt adducts in 10 mM phosphate buffer, pH 6.4, free of any unreacted platinum (removed by ultrafiltration), was reacted with a five-fold excess of the relevant biological nucleophile or oligonucleotide. The reactions were at 37 8C. Kinetics by ESIMS were measured directly on the reaction mixtures following ZipTip姠 (C 18 , Millipore) treatment.
2.4. ESI-MS Electrospray ionization mass spectrometry was measured on a ThermoQuest Finnigan LCQ-Duo in the positive ion mode. Elution was in 49:49:2 water: methanol:acetic acid at a flow rate of 15 ml / min. Samples of the platination reactions and adduct reactivity studies were diluted 100-fold prior to ESI-MS analysis. Data were processed using ThermoQuest Finnigan’s Xcalibur姠 Biomass Calculation and Deconvolution software.
2.5. Oxidation of ubiquitin Performic acid oxidation of the Met1 residue of ubiquitin was performed according to the method of Breslow et al. [24]. 3. Results and discussion
3.1. The types of adducts formed by cis- and trans-DDP with ubiquitin Electrospray ionization mass spectrometry is a soft
308
T. Peleg-Shulman et al. / Journal of Inorganic Biochemistry 91 (2002) 306 – 311
ionization technique that is very useful in detection of protein modifications and since Pt complexes form covalent adducts with the proteins, these can be observed by ESI-MS. The reaction of one equivalent of cisplatin with one equivalent of ubiquitin was monitored by ESI-MS. The four adducts (I–IV) that are formed appear in the mass spectrum depicted in Fig. 2a. These adducts have been previously assigned on the basis of their masses as: I, cis-[Pt(Ub)(NH 3 ) 2 Cl]; II, [Pt(Ub)(NH 3 ) 2 (H 2 O)]; III, [Pt(Ub)(NH 3 ) 2 ]; and IV, [Pt(Ub)(NH 3 )] [21]. The mass assignment also corresponds to the expected binding properties of cisplatin where first the monofunctional adduct (I) is formed followed by aquation (II) and subsequent chelate formation (III) [21,23]. The trans isomer forms primarily one adduct, the monofunctional adduct trans-[Pt(Ub)(NH 3 ) 2 Cl] (Fig. 2b). In contrast to the cis-[Pt((Ub)(NH 3 ) 2 Cl] adduct, the trans-[Pt(Ub)(NH 3 ) 2 Cl] does not undergo aquation even after 3 weeks at 37 8C and does not form chelates with the protein. The results of this study demonstrate that cisplatin and transplatin form different spectra of adducts with ubiquitin.
3.2. The binding sites of cis- and trans-[ Pt( NH3 )2 Cl2 ] on ubiquitin Ubiquitin is a small tightly folded protein of 76 amino acids (MW 8565 Da) that has two chemically favorable platinum binding sites: Met1 and His68 [25]. Met1 is slightly buried under the protein surface while His68 is exposed on the surface of the protein (Fig. 3). Kinetic studies with linear and cyclic peptides containing both methionine and histidine show that Pt(II) complexes bind faster to methionine than to histidine and eventually form macrochelates with the peptides [26]. In
Fig. 2. (a) ESI-MS of the reaction mixture of 1:1 cisplatin and ubiquitin at 37 8C after 24 h. The four adducts are labeled I–IV. (b) ESI-MS of the reaction mixture of 1:1 transplatin and ubiquitin at 37 8C after 24 h.
Fig. 3. X-ray crystal structure of ubiquitin [25]. The potential binding sites Met1 and His68 are drawn in boldface and labeled. The four carboxy-terminal amino acids are annotated.
order to see whether the same kinetic preference holds true for proteins, the thioether of Met1 was selectively oxidized to the sulfone thereby preventing any Pt binding to Met1. Ubiquitin was oxidized in a quantitative manner, as determined by the single peak in the mass spectrum at 8597 amu that corresponds to the addition of two oxygen atoms to the sulfur (32 amu, results not shown). The binding kinetics of cisplatin and transplatin to the native ubiquitin and the oxidized protein were compared. The results are shown in Fig. 4. Cisplatin binds to the native protein (Ub) significantly faster than to the oxidized protein (Fig. 4a) suggesting that Met1 is a kinetically favored binding site for cisplatin. For transplatin, there is no significant difference between the binding profiles with the native versus the oxidized protein (Fig. 4b) suggesting that Met1 is not a preferred binding site. The nature of the adducts that cisplatin and transplatin form with Ub and UbOx, as observed by mass spectrometry, lead us to conclude that Met1 of ubiquitin is the preferred binding site for cisplatin but not for transplatin. Based on previous results obtained with amino acids and peptides [27,28], it is not surprising that the initial binding of cisplatin is to the thioether of Met1 to yield the monofunctional cis-[Pt(Ub)(NH 3 ) 2 Cl]. In the interactions of cis-[Pt(Am) 2 Cl 2 ] with amino acids and peptides, the monofunctional adduct converts to the chelate (or macrochelate) [18]. Free amino acids can chelate the two cis positions of Pt(II) using the a-amino group and the side chain nucleophile (such as the sulfur atom in Met and
T. Peleg-Shulman et al. / Journal of Inorganic Biochemistry 91 (2002) 306 – 311
309
Fig. 4. (a) Binding curves of cisplatin to native ubiquitin (Ub, –d–) and to oxidized ubiquitin (UbOx, –♦–); (b) binding curves of transplatin to native ubiquitin (Ub,) and to oxidized ubiquitin (UbOx,). Reactions were carried out at 1–2 mM concentrations, in 10 mM phosphate buffer, pH 6.4, 37 8C.
Cys). In proteins, it is mainly the side chain nucleophiles that are available for Pt binding. In the case of Ub, Met1 can easily chelate the Pt(NH 3 ) 2 moiety with its free N-terminal amine and its thioether, acting like a free amino acid. The thioether can then trans labilize the ammine leading to the formation of a tri-functional adduct between the Pt moiety and ubiquitin (Fig. 2a, species IV). Transplatin, however, forms a very stable trans[Pt(Ub)(NH 3 ) 2 Cl] adduct; no aquated species are observed in the ESI-MS over a reaction period of 21 days. This would be unlikely if the Pt were bound to the thioether due the strong trans effect of the sulfur. It is not clear why Met1 is not a preferred binding site for transplatin since transplatin is known to bind with high affinity to sulfur containing ligands. The explanation may have to do with the steric accessibility of this specific site.
Fig. 5. Binding curves of cisplatin (–j–) and transplatin (–m–) to ubiquitin (a) and to horse heart myoglobin (b). Reactions were carried out at 1–2 mM concentrations, in 10 mM phosphate buffer, pH 6.4, 37 8C. The ordinate depicts the percentage of all the adducts formed in the reactions.
this trend is maintained when the reaction kinetics of the binding of cis- and trans-DDP to horse heart myoglobin are monitored (Fig. 5b). A possible explanation could be the preferred binding of cisplatin to the soft thioether ligand of Met1, which is not the case for transplatin (see Section 3.2 above on binding sites), thus making transplatin binding kinetically less favorable.
3.3. The binding kinetics of cis- and trans-DDP to ubiquitin and horse heart myoglobin
3.4. Reaction of preformed ubiquitin-cis-Pt ubiquitintrans-Pt adducts with 59 -guanosine monophosphate (59 GMP)
The binding kinetics of cis- and trans-DDP to Ub are depicted in Fig. 5a. Cisplatin binds to Ub more efficiently than transplatin. The reaction with cisplatin proceeds more rapidly, and a greater percentage of the native protein is platinated in the presence of cisplatin. This is contrary to the expectation based on the fact that the trans effect of the chloride is stronger than that of the ammine ligand. Yet,
Ubiquitin was reacted with cisplatin or transplatin and the reactions were monitored by ESI-MS. The reactions were terminated when |60–70% of the native protein was converted to Ub-Pt adducts. The unreacted platinum complexes were removed by ultrafiltration and the mixture of unreacted Ub and Ub-Pt adducts was reacted with a five-fold excess of 59-GMP. Both the cis- and trans-Pt-Ub
310
T. Peleg-Shulman et al. / Journal of Inorganic Biochemistry 91 (2002) 306 – 311
adducts formed ternary complexes where the Pt atom covalently bridges the protein and the nucleotide. The ternary adducts were stable for 2 weeks. Since methioninecis-Pt lesions are the preferred adducts for the reaction of cisplatin with Ub, any transfer of the Pt moiety from the protein to the 59-GMP would have resulted in a decrease in the concentration of the ternary adduct and an increase in the concentration of the native Ub. Since this is not observed (Fig. 6), it seems that Ub does not transfer the Pt from Met1 to 59-GMP. No transfer of platinum was observed in the case of transplatin either. An interesting phenomenon was detected in the reaction of transplatin-Ub with 59-GMP. Several days into the reaction, we observed a new peak in the mass spectrum at 8131.5 amu. This mass, which is lower than the mass of the native protein, can only be explained by the loss of several amino acids from the protein. A closer look at the structure of the protein (Fig. 3) reveals that while most of the protein is tightly folded, the four carboxy-terminal amino acids (LRGG, LeuArgGlyGly, annotated in Fig. 3), that are involved in the ubiquitination of proteins, protrude towards the solution. The mass loss corresponds to the loss of these four carboxy-terminal amino acids of the protein (LRGG). As the reaction continues, the ‘decapitated’ protein becomes the dominant species, although the native ubiquitin is observed throughout. This may serve as a further indication that cisplatin and transplatin do not bind the same sites on the protein, and hence the loss of the carboxy-terminal residues occurs only in the case of transplatin. When cisplatin-platinated horse heart myoglobin (Mb) was reacted in a similar fashion with a tetranucleotide, d(TGGT), a similar trend was observed; ternary complexes of the form Mb-Pt-d(TGGT) were formed and remained stable over 14 days (results not shown).
Fig. 7. Reaction of glutathione with the Ub-Pt adducts of cisplatin (–d–) and transplatin (–m–). The reactivity studies were carried out at a 2 mM concentration, in 10 mM phosphate buffer, pH 6.4, 37 8C.
form ternary complexes where the protein is tethered to the peptide via the Pt moiety. With time, the concentration of the ternary adducts decreases with a concomitant increase in the concentration of the native Ub indicating that the Pt moiety is removed from the protein by GSH, probably due to the strong trans effect of the thiolate of GSH which is able to discharge the Pt moiety from the protein. The ternary complex with cisplatin forms more rapidly within the first several days of the reaction, compared to the transplatin ternary complex formation (Fig. 7). However, in both cases, over a period of 14 days, all platinum adducts are removed, yielding the protein in its original native form. The efficient removal of the platinum complexes from the protein by glutathione, along with the lack of platinum transfer from the protein to either the nucleotide or the short oligonucleotide, renders it unlikely that platinated amino acids on the surface of these proteins act as a reservoir for the sustained release of the drug towards cellular DNA platination.
3.5. Reaction of preformed ubiquitin-cis-Pt ubiquitintrans-Pt adducts with glutathione ( GSH) 4. Conclusions Preformed adducts of cisplatin and transplatin with Ub were reacted with a five-fold excess of GSH to initially
Fig. 6. Binding curves of 59-GMP to the Ub-Pt adducts of cisplatin (–s–) and transplatin (–m–). The reactivity studies were carried out at a 2 mM concentration, in 10 mM phosphate buffer, pH 6.4, 37 8C.
Cisplatin and transplatin form different spectra of adducts with ubiquitin. Transplatin forms one major adduct with Ub which is a monofunctional adduct trans[Pt(Ub)(NH 3 ) 2 Cl], while cisplatin forms four distinct adducts; two monofunctional [Pt(Ub)(NH 3 ) 2 Cl] and [Pt(Ub)(NH 3 ) 2 (H 2 O)], one bifunctional [Pt(Ub)(NH 3 ) 2 ] and one trifunctional [Pt(Ub)(NH 3 )]. When binding ubiquitin, Met1 is the preferred binding site of cisplatin, but not of transplatin. Cisplatin binds faster than transplatin to both ubiquitin and horse heart myoglobin. Both cisplatin and transplatin form stable ternary adducts where the platinum moieties bridge ubiquitin and 59-GMP. These adducts are stable for weeks at 37 8C. Similar results have been obtained with a short oligonucleotide. Glutathione efficiently removes the platinum from preformed adducts of both cisplatin and transplatin with ubiquitin.
T. Peleg-Shulman et al. / Journal of Inorganic Biochemistry 91 (2002) 306 – 311
5. Abbreviations amu cis-DDP trans-DDP Da ESI-MS GMP GSH HSA HSQC Mb MW NMR Ub UbOx
atomic mass units cisplatin, cis-diamminedichloroplatinum(II) transplatin, trans-diamminedichloroplatinum(II) dalton electrospray ionization mass spectrometry 59-guanosine monophosphate glutathione human serum albumin heteronuclear single quantum coherence horse heart myoglobin molecular weight nuclear magnetic resonance ubiquitin oxidized ubiquitin
Acknowledgements T.P.-S. is grateful to The Alex Grass Center for Drug Design and Synthesis of Novel Therapeutics at The School of Pharmacy, for partial support. Y.N. thanks the David R. Bloom Center for Pharmacy at The Hebrew University of Jerusalem, Israel for financial support. D.G. thanks the Israel Cancer Association (grant 20010031-B) for support of this project. The authors would like to thank Liliana Balter for technical assistance.
References [1] K.M. Comess, S.J. Lippard, in: S. Neidle, M. Waring (Eds.), Molecular Aspects of Platinum–DNA Interactions, Molecular Aspects of Anticancer Drug–DNA Interactions, Vol. 1, Macmillan, London, 1993, pp. 134–168. [2] E.R. Jamieson, S.J. Lippard, Chem. Rev. 99 (1999) 2467–2498. [3] B. Lippert, Met. Ions Biol. Syst. 33 (1996) 105–141.
311
[4] P.M. Takahara, A.C. Rosenzweig, C.A. Frederick, S.J. Lippard, Nature 377 (1995) 649–652. [5] M. Kartalou, J.M. Essigmann, Mutat. Res. Fundam. Mol. Mech. Mutagen. 478 (2001) 1–21. [6] J.M. Perez, M.A. Fuertes, C. Alonso, C. Navarro-Ranninger, Crit. Rev. Oncol. Hematol. 35 (2000) 109–120. [7] G. Natile, M. Coluccia, Coord. Chem. Rev. 216–127 (2001) 383– 410. [8] R.C. Deconti, B.R. Toftness, R.C. Lange, W.A. Creasy, Cancer Res. 33 (1973) 1310–1315. [9] T.G. Appleton, Coord. Chem. Rev. 166 (1997) 313–359. [10] J.J. Gullo, C.L. Litterst, P.J. Maguire, B.J. Sikis, D.F. Holth, P.V. Woodley, Cancer Chemother. Pharmacol. 5 (1980) 21–26. [11] E. Espinosa, J. Feliu, P. Zamora, M.J. Baron, J.J. Sanchez, A. Ordonez, J. Espinosa, Lung Cancer 12 (1995) 67–76. [12] J.D. Holding, W.E. Lindup, C. Van Laer, C.G.M. Vreeburg, V. Schiling, J.A. Wilson, P.M. Stell, Br. J. Clin. Pharmacol. 33 (1992) 75–81. [13] J. Reedijk, Chem. Rev. 99 (1999) 2499–2510. [14] K.J. Barnham, M.I. Djuran, P.d.-S. Murdoch, P.J. Sadler, J. Chem. Soc., Chem. Commun. 6 (1994) 721–722. [15] K.J. Barnham, M.I. Djuran, P.d.-S. Murdoch, J.D. Ranford, P.J. Sadler, J. Chem. Soc., Dalton Trans. 22 (1995) 3721–3726. [16] J.-M. Teuben, J. Reeijk, J. Biol. Inorg. Chem. 5 (2000) 463–468. [17] V. Marchan, V. Moreno, E. Pedroso, A. Grandas, Chem. Eur. J. 7 (2001) 808–815. [18] M. Hahn, M. Kleine, W.S. Sheldrick, J. Biol. Inorg. Chem. 6 (2001) 556–566. [19] O. Heudi, A. Cailleux, P. Allain, J. Inorg. Biochem. 71 (1998) 61–69. [20] A.I. Ivanov, J. Christodoulou, J.A. Parkinson, K.J. Barnahm, A. Tucker, J. Woodrow, P.J. Sadler, J. Biol. Chem. 273 (1998) 14721– 14730. [21] D. Gibson, C.E. Costello, Eur. Mass Spectrom. 5 (1999) 501–510. [22] S.J. Berners-Price, P.J. Sadler, Coord. Chem. Rev. 151 (1996) 1–40. [23] T. Peleg-Shulman, D. Gibson, J. Am. Chem. Soc. 123 (2001) 3171–3172. [24] S. Bamezai, M.A.T. Banez, E. Breslow, Biochemistry 29 (1990) 5389–5396. [25] S. Vijay-Kumar, C.E. Bugg, W.J. Cook, J. Mol. Biol. 194 (1987) 525–538. [26] D. Wolters, W.S. Sheldrick, J. Chem. Soc., Dalton Trans. 7 (1999) 1121–1129. [27] T.G. Appleton, J.W. Connor, J.R. Hall, Inorg. Chem. 27 (1988) 130–137. [28] E.L.M. Lempers, K. Inagaki, J. Reedijk, Inorg. Chim. Acta 152 (1988) 201–207.
Interactions of cisplatin and transplatin with proteins. Comparison of binding kinetics, binding sites and reactivity of the Ptprotein adducts of cisplatin and transplatin towards biological nucleophiles Tal Peleg-Shulman, Yousef Najajreh, Dan Gibson*
,1
Department of Medicinal Chemistry and Natural Products, School of Pharmacy, PO Box 12065, The Hebrew University of Jerusalem, Jerusalem 91120, Israel Received 27 September 2001; received in revised form 26 November 2001; accepted 18 December 2001
Abstract In this manuscript we report on the interactions of cis-DDP (cisplatin, cis-diamminedichloroplatinum(II)) and trans-DDP (transplatin, trans-diamminedichloroplatinum(II)) with two model proteins, ubiquitin (Ub) and horse heart myoglobin (Mb), and attempt to answer the question whether proteins that have methionine-Pt adducts can transfer the platinum to biological nucleophiles and particularly to DNA. Our study shows that cisplatin and transplatin form different adducts with ubiquitin: transplatin forms one major adduct, trans[Pt(Ub)(NH 3 ) 2 Cl], while cisplatin forms four distinct adducts, [Pt(Ub)(NH 3 ) 2 Cl], [Pt(Ub)(NH 3 ) 2 (H 2 O)], [Pt(Ub)(NH 3 ) 2 ], and [Pt(Ub)(NH 3 )]. When binding ubiquitin, Met1 is the preferred binding site of cisplatin, but not of transplatin. Cisplatin binds faster than transplatin to both ubiquitin and horse heart myoglobin. Both cisplatin and transplatin adducts form stable ternary adducts when reacted with 59-guanosine monophosphate (59-GMP) or a tetranucleotide. No transfer of the Pt moiety from the proteins to the nucleotides was observed. Glutathione efficiently removes the platinum from preformed adducts of both cisplatin and transplatin with ubiquitin. 2002 Elsevier Science Inc. All rights reserved. Keywords: Cisplatin; Transplatin; Protein binding; Electrospray ionization mass spectrometry (ESI-MS)
1. Introduction Cisplatin (Fig. 1a) is a widely used anti-tumor agent that is used in the clinic to treat testicular and ovarian cancers
Fig. 1. (a) cisplatin; (b) transplatin.
*Corresponding author. Tel.: 1972-2-675-8702; fax: 1972-2-6757076. E-mail address: [email protected] (D. Gibson). 1 Affiliated with the David R. Bloom Center for Pharmacy at The Hebrew University of Jerusalem, Jerusalem, Israel.
[1]. Cisplatin is believed to induce apoptosis in cancer cells by covalently modifying the DNA [2]. Its geometric isomer, trans-[Pt(NH 3 ) 2 Cl 2 ] (Fig. 1b), has no cytotoxic activity [3]. The difference in antitumor activity between the two isomers is attributed to the inability of the trans isomer to form 1,2-GpG intrastrand crosslinks due the 1808 angle between its two semi-labile chloride ligands. Cellular proteins such as the HMG box proteins recognize cisplatin modified DNA and bind to it. The binding of the proteins to the platinated DNA can serve to protect the DNA from being repaired by nuclear excision repair enzymes, or alternatively, hijack the proteins from their normal function [4,5]. Transplatin modified DNA is not recognized by those same cellular proteins [6,7]. Cisplatin is administered intravenously, and within 1 day, 65–98% of the drug is bound to blood plasma proteins [8]. While Pt-DNA adducts are believed to be responsible for the drug’s cytotoxicity, the exact role that Pt-protein adducts play in the mechanism of action of the drug is yet to be elucidated. It has been postulated that in addition to drug inactivation, cisplatin binding to proteins may be the cause of many of the drug’s side-effects [9,10].
0162-0134 / 02 / $ – see front matter 2002 Elsevier Science Inc. All rights reserved. PII: S0162-0134( 02 )00362-8
T. Peleg-Shulman et al. / Journal of Inorganic Biochemistry 91 (2002) 306 – 311
Other reports suggest that Pt-HSA (human serum albumin) adducts may be important for the activity of the drug [11]. When preformed Pt-HSA adducts were administered clinically they increased the survival time of the patients. Also, patients with low levels of HSA did not respond well to cisplatin based chemotherapy [12]. It is clear that the formation of Pt-protein adducts, that effectively competes with formation of the cytotoxic Pt-DNA lesions, can reduce the efficacy of Pt antitumor agents. The efficient binding of platinum complexes to proteins and peptides is not surprising since platinum has a high affinity for sulfur containing ligands such as methionine and cysteine [13]. The abundance of extra- and intracellular platinophiles makes it difficult to understand how Pt even reaches the DNA. Model studies suggest that a platinum moiety that is bound to a methionine thioether can be transferred to the N7 of the guanine of single stranded and double stranded oligonucleotides (but not from the cysteine thiolate) [14–17]. On the basis of these studies it was suggested that proteins that form Ptmethionine adducts could act as a platinum reservoir for subsequent DNA platination [13]. Many studies on the reaction of platinum complexes with amino acids and proteins have been reported [18–20], but there are few high-resolution reports on the interactions of cis-DDP (cisplatin, cis-diamminedichloroplatinum(II)) or transDDP (transplatin, trans-diamminedichloroplatinum(II)) with proteins. We have recently shown that electrospray ionization mass spectrometry is an extremely useful technique for studying the interactions of cisplatin with proteins [21], and provides information on the nature of the Pt-protein adducts that are formed. [ 1 H, 15 N] heteronuclear single quantum coherence (HSQC) NMR spectroscopy has been utilized in the analysis of 15 N-labeled platinum ammine complexes providing information on the nature of the ligands that are trans to the 15 N-labeled ammine [22]. The combination of the two techniques provides a powerful tool for simultaneously studying the modifications of the protein (electrospray ionization mass spectrometry, ESIMS) and the changes in the platinum coordination sphere (HSQC) [23]. Since Pt-protein adducts are important in defining the therapeutic profiles of the drugs, it is important to understand the basic principles that govern the formation and reactivity of these adducts. In this manuscript we report on the interaction of cis- and trans-DDP with model proteins, ubiquitin (Ub) and horse heart myoglobin (Mb). These proteins were chosen because both are well characterized, have methionines and histidines (but not cysteines) and are good candidates to try and answer the question whether proteins that have methionine-Pt adducts can transfer the Pt to biological nucleophiles and particularly to DNA. The binding kinetics, the nature of the adducts formed, the binding sites of the complexes and the reactivity of the Pt-protein adducts towards biologically relevant nu-
307
cleophiles such as glutathione and 59-guanosine monophosphate are reported. 2. Experimental
2.1. Materials Cis- and trans-DDP, ubiquitin and horse heart myoglobin were all purchased from Sigma–Aldrich (Israel) and were used without further purification.
2.2. Platination reactions Platination reactions were carried out at 1–2 mM concentrations, in 10 mM phosphate buffer, pH 6.4, 37 8C. Excess platinum was removed by ultra-filtration using Microcon YM-3 centrifugal filter devices at 4 8C and 12,000 rpm, prior to all NMR and adduct-reactivity studies. Kinetics by ESI-MS were measured directly on the reaction mixtures following ZipTip姠 (C 18 , Millipore) treatment.
2.3. Reactions of Pt-protein adducts with nucleophiles A 2 mM concentration of Ub-Pt adducts in 10 mM phosphate buffer, pH 6.4, free of any unreacted platinum (removed by ultrafiltration), was reacted with a five-fold excess of the relevant biological nucleophile or oligonucleotide. The reactions were at 37 8C. Kinetics by ESIMS were measured directly on the reaction mixtures following ZipTip姠 (C 18 , Millipore) treatment.
2.4. ESI-MS Electrospray ionization mass spectrometry was measured on a ThermoQuest Finnigan LCQ-Duo in the positive ion mode. Elution was in 49:49:2 water: methanol:acetic acid at a flow rate of 15 ml / min. Samples of the platination reactions and adduct reactivity studies were diluted 100-fold prior to ESI-MS analysis. Data were processed using ThermoQuest Finnigan’s Xcalibur姠 Biomass Calculation and Deconvolution software.
2.5. Oxidation of ubiquitin Performic acid oxidation of the Met1 residue of ubiquitin was performed according to the method of Breslow et al. [24]. 3. Results and discussion
3.1. The types of adducts formed by cis- and trans-DDP with ubiquitin Electrospray ionization mass spectrometry is a soft
308
T. Peleg-Shulman et al. / Journal of Inorganic Biochemistry 91 (2002) 306 – 311
ionization technique that is very useful in detection of protein modifications and since Pt complexes form covalent adducts with the proteins, these can be observed by ESI-MS. The reaction of one equivalent of cisplatin with one equivalent of ubiquitin was monitored by ESI-MS. The four adducts (I–IV) that are formed appear in the mass spectrum depicted in Fig. 2a. These adducts have been previously assigned on the basis of their masses as: I, cis-[Pt(Ub)(NH 3 ) 2 Cl]; II, [Pt(Ub)(NH 3 ) 2 (H 2 O)]; III, [Pt(Ub)(NH 3 ) 2 ]; and IV, [Pt(Ub)(NH 3 )] [21]. The mass assignment also corresponds to the expected binding properties of cisplatin where first the monofunctional adduct (I) is formed followed by aquation (II) and subsequent chelate formation (III) [21,23]. The trans isomer forms primarily one adduct, the monofunctional adduct trans-[Pt(Ub)(NH 3 ) 2 Cl] (Fig. 2b). In contrast to the cis-[Pt((Ub)(NH 3 ) 2 Cl] adduct, the trans-[Pt(Ub)(NH 3 ) 2 Cl] does not undergo aquation even after 3 weeks at 37 8C and does not form chelates with the protein. The results of this study demonstrate that cisplatin and transplatin form different spectra of adducts with ubiquitin.
3.2. The binding sites of cis- and trans-[ Pt( NH3 )2 Cl2 ] on ubiquitin Ubiquitin is a small tightly folded protein of 76 amino acids (MW 8565 Da) that has two chemically favorable platinum binding sites: Met1 and His68 [25]. Met1 is slightly buried under the protein surface while His68 is exposed on the surface of the protein (Fig. 3). Kinetic studies with linear and cyclic peptides containing both methionine and histidine show that Pt(II) complexes bind faster to methionine than to histidine and eventually form macrochelates with the peptides [26]. In
Fig. 2. (a) ESI-MS of the reaction mixture of 1:1 cisplatin and ubiquitin at 37 8C after 24 h. The four adducts are labeled I–IV. (b) ESI-MS of the reaction mixture of 1:1 transplatin and ubiquitin at 37 8C after 24 h.
Fig. 3. X-ray crystal structure of ubiquitin [25]. The potential binding sites Met1 and His68 are drawn in boldface and labeled. The four carboxy-terminal amino acids are annotated.
order to see whether the same kinetic preference holds true for proteins, the thioether of Met1 was selectively oxidized to the sulfone thereby preventing any Pt binding to Met1. Ubiquitin was oxidized in a quantitative manner, as determined by the single peak in the mass spectrum at 8597 amu that corresponds to the addition of two oxygen atoms to the sulfur (32 amu, results not shown). The binding kinetics of cisplatin and transplatin to the native ubiquitin and the oxidized protein were compared. The results are shown in Fig. 4. Cisplatin binds to the native protein (Ub) significantly faster than to the oxidized protein (Fig. 4a) suggesting that Met1 is a kinetically favored binding site for cisplatin. For transplatin, there is no significant difference between the binding profiles with the native versus the oxidized protein (Fig. 4b) suggesting that Met1 is not a preferred binding site. The nature of the adducts that cisplatin and transplatin form with Ub and UbOx, as observed by mass spectrometry, lead us to conclude that Met1 of ubiquitin is the preferred binding site for cisplatin but not for transplatin. Based on previous results obtained with amino acids and peptides [27,28], it is not surprising that the initial binding of cisplatin is to the thioether of Met1 to yield the monofunctional cis-[Pt(Ub)(NH 3 ) 2 Cl]. In the interactions of cis-[Pt(Am) 2 Cl 2 ] with amino acids and peptides, the monofunctional adduct converts to the chelate (or macrochelate) [18]. Free amino acids can chelate the two cis positions of Pt(II) using the a-amino group and the side chain nucleophile (such as the sulfur atom in Met and
T. Peleg-Shulman et al. / Journal of Inorganic Biochemistry 91 (2002) 306 – 311
309
Fig. 4. (a) Binding curves of cisplatin to native ubiquitin (Ub, –d–) and to oxidized ubiquitin (UbOx, –♦–); (b) binding curves of transplatin to native ubiquitin (Ub,) and to oxidized ubiquitin (UbOx,). Reactions were carried out at 1–2 mM concentrations, in 10 mM phosphate buffer, pH 6.4, 37 8C.
Cys). In proteins, it is mainly the side chain nucleophiles that are available for Pt binding. In the case of Ub, Met1 can easily chelate the Pt(NH 3 ) 2 moiety with its free N-terminal amine and its thioether, acting like a free amino acid. The thioether can then trans labilize the ammine leading to the formation of a tri-functional adduct between the Pt moiety and ubiquitin (Fig. 2a, species IV). Transplatin, however, forms a very stable trans[Pt(Ub)(NH 3 ) 2 Cl] adduct; no aquated species are observed in the ESI-MS over a reaction period of 21 days. This would be unlikely if the Pt were bound to the thioether due the strong trans effect of the sulfur. It is not clear why Met1 is not a preferred binding site for transplatin since transplatin is known to bind with high affinity to sulfur containing ligands. The explanation may have to do with the steric accessibility of this specific site.
Fig. 5. Binding curves of cisplatin (–j–) and transplatin (–m–) to ubiquitin (a) and to horse heart myoglobin (b). Reactions were carried out at 1–2 mM concentrations, in 10 mM phosphate buffer, pH 6.4, 37 8C. The ordinate depicts the percentage of all the adducts formed in the reactions.
this trend is maintained when the reaction kinetics of the binding of cis- and trans-DDP to horse heart myoglobin are monitored (Fig. 5b). A possible explanation could be the preferred binding of cisplatin to the soft thioether ligand of Met1, which is not the case for transplatin (see Section 3.2 above on binding sites), thus making transplatin binding kinetically less favorable.
3.3. The binding kinetics of cis- and trans-DDP to ubiquitin and horse heart myoglobin
3.4. Reaction of preformed ubiquitin-cis-Pt ubiquitintrans-Pt adducts with 59 -guanosine monophosphate (59 GMP)
The binding kinetics of cis- and trans-DDP to Ub are depicted in Fig. 5a. Cisplatin binds to Ub more efficiently than transplatin. The reaction with cisplatin proceeds more rapidly, and a greater percentage of the native protein is platinated in the presence of cisplatin. This is contrary to the expectation based on the fact that the trans effect of the chloride is stronger than that of the ammine ligand. Yet,
Ubiquitin was reacted with cisplatin or transplatin and the reactions were monitored by ESI-MS. The reactions were terminated when |60–70% of the native protein was converted to Ub-Pt adducts. The unreacted platinum complexes were removed by ultrafiltration and the mixture of unreacted Ub and Ub-Pt adducts was reacted with a five-fold excess of 59-GMP. Both the cis- and trans-Pt-Ub
310
T. Peleg-Shulman et al. / Journal of Inorganic Biochemistry 91 (2002) 306 – 311
adducts formed ternary complexes where the Pt atom covalently bridges the protein and the nucleotide. The ternary adducts were stable for 2 weeks. Since methioninecis-Pt lesions are the preferred adducts for the reaction of cisplatin with Ub, any transfer of the Pt moiety from the protein to the 59-GMP would have resulted in a decrease in the concentration of the ternary adduct and an increase in the concentration of the native Ub. Since this is not observed (Fig. 6), it seems that Ub does not transfer the Pt from Met1 to 59-GMP. No transfer of platinum was observed in the case of transplatin either. An interesting phenomenon was detected in the reaction of transplatin-Ub with 59-GMP. Several days into the reaction, we observed a new peak in the mass spectrum at 8131.5 amu. This mass, which is lower than the mass of the native protein, can only be explained by the loss of several amino acids from the protein. A closer look at the structure of the protein (Fig. 3) reveals that while most of the protein is tightly folded, the four carboxy-terminal amino acids (LRGG, LeuArgGlyGly, annotated in Fig. 3), that are involved in the ubiquitination of proteins, protrude towards the solution. The mass loss corresponds to the loss of these four carboxy-terminal amino acids of the protein (LRGG). As the reaction continues, the ‘decapitated’ protein becomes the dominant species, although the native ubiquitin is observed throughout. This may serve as a further indication that cisplatin and transplatin do not bind the same sites on the protein, and hence the loss of the carboxy-terminal residues occurs only in the case of transplatin. When cisplatin-platinated horse heart myoglobin (Mb) was reacted in a similar fashion with a tetranucleotide, d(TGGT), a similar trend was observed; ternary complexes of the form Mb-Pt-d(TGGT) were formed and remained stable over 14 days (results not shown).
Fig. 7. Reaction of glutathione with the Ub-Pt adducts of cisplatin (–d–) and transplatin (–m–). The reactivity studies were carried out at a 2 mM concentration, in 10 mM phosphate buffer, pH 6.4, 37 8C.
form ternary complexes where the protein is tethered to the peptide via the Pt moiety. With time, the concentration of the ternary adducts decreases with a concomitant increase in the concentration of the native Ub indicating that the Pt moiety is removed from the protein by GSH, probably due to the strong trans effect of the thiolate of GSH which is able to discharge the Pt moiety from the protein. The ternary complex with cisplatin forms more rapidly within the first several days of the reaction, compared to the transplatin ternary complex formation (Fig. 7). However, in both cases, over a period of 14 days, all platinum adducts are removed, yielding the protein in its original native form. The efficient removal of the platinum complexes from the protein by glutathione, along with the lack of platinum transfer from the protein to either the nucleotide or the short oligonucleotide, renders it unlikely that platinated amino acids on the surface of these proteins act as a reservoir for the sustained release of the drug towards cellular DNA platination.
3.5. Reaction of preformed ubiquitin-cis-Pt ubiquitintrans-Pt adducts with glutathione ( GSH) 4. Conclusions Preformed adducts of cisplatin and transplatin with Ub were reacted with a five-fold excess of GSH to initially
Fig. 6. Binding curves of 59-GMP to the Ub-Pt adducts of cisplatin (–s–) and transplatin (–m–). The reactivity studies were carried out at a 2 mM concentration, in 10 mM phosphate buffer, pH 6.4, 37 8C.
Cisplatin and transplatin form different spectra of adducts with ubiquitin. Transplatin forms one major adduct with Ub which is a monofunctional adduct trans[Pt(Ub)(NH 3 ) 2 Cl], while cisplatin forms four distinct adducts; two monofunctional [Pt(Ub)(NH 3 ) 2 Cl] and [Pt(Ub)(NH 3 ) 2 (H 2 O)], one bifunctional [Pt(Ub)(NH 3 ) 2 ] and one trifunctional [Pt(Ub)(NH 3 )]. When binding ubiquitin, Met1 is the preferred binding site of cisplatin, but not of transplatin. Cisplatin binds faster than transplatin to both ubiquitin and horse heart myoglobin. Both cisplatin and transplatin form stable ternary adducts where the platinum moieties bridge ubiquitin and 59-GMP. These adducts are stable for weeks at 37 8C. Similar results have been obtained with a short oligonucleotide. Glutathione efficiently removes the platinum from preformed adducts of both cisplatin and transplatin with ubiquitin.
T. Peleg-Shulman et al. / Journal of Inorganic Biochemistry 91 (2002) 306 – 311
5. Abbreviations amu cis-DDP trans-DDP Da ESI-MS GMP GSH HSA HSQC Mb MW NMR Ub UbOx
atomic mass units cisplatin, cis-diamminedichloroplatinum(II) transplatin, trans-diamminedichloroplatinum(II) dalton electrospray ionization mass spectrometry 59-guanosine monophosphate glutathione human serum albumin heteronuclear single quantum coherence horse heart myoglobin molecular weight nuclear magnetic resonance ubiquitin oxidized ubiquitin
Acknowledgements T.P.-S. is grateful to The Alex Grass Center for Drug Design and Synthesis of Novel Therapeutics at The School of Pharmacy, for partial support. Y.N. thanks the David R. Bloom Center for Pharmacy at The Hebrew University of Jerusalem, Israel for financial support. D.G. thanks the Israel Cancer Association (grant 20010031-B) for support of this project. The authors would like to thank Liliana Balter for technical assistance.
References [1] K.M. Comess, S.J. Lippard, in: S. Neidle, M. Waring (Eds.), Molecular Aspects of Platinum–DNA Interactions, Molecular Aspects of Anticancer Drug–DNA Interactions, Vol. 1, Macmillan, London, 1993, pp. 134–168. [2] E.R. Jamieson, S.J. Lippard, Chem. Rev. 99 (1999) 2467–2498. [3] B. Lippert, Met. Ions Biol. Syst. 33 (1996) 105–141.
311
[4] P.M. Takahara, A.C. Rosenzweig, C.A. Frederick, S.J. Lippard, Nature 377 (1995) 649–652. [5] M. Kartalou, J.M. Essigmann, Mutat. Res. Fundam. Mol. Mech. Mutagen. 478 (2001) 1–21. [6] J.M. Perez, M.A. Fuertes, C. Alonso, C. Navarro-Ranninger, Crit. Rev. Oncol. Hematol. 35 (2000) 109–120. [7] G. Natile, M. Coluccia, Coord. Chem. Rev. 216–127 (2001) 383– 410. [8] R.C. Deconti, B.R. Toftness, R.C. Lange, W.A. Creasy, Cancer Res. 33 (1973) 1310–1315. [9] T.G. Appleton, Coord. Chem. Rev. 166 (1997) 313–359. [10] J.J. Gullo, C.L. Litterst, P.J. Maguire, B.J. Sikis, D.F. Holth, P.V. Woodley, Cancer Chemother. Pharmacol. 5 (1980) 21–26. [11] E. Espinosa, J. Feliu, P. Zamora, M.J. Baron, J.J. Sanchez, A. Ordonez, J. Espinosa, Lung Cancer 12 (1995) 67–76. [12] J.D. Holding, W.E. Lindup, C. Van Laer, C.G.M. Vreeburg, V. Schiling, J.A. Wilson, P.M. Stell, Br. J. Clin. Pharmacol. 33 (1992) 75–81. [13] J. Reedijk, Chem. Rev. 99 (1999) 2499–2510. [14] K.J. Barnham, M.I. Djuran, P.d.-S. Murdoch, P.J. Sadler, J. Chem. Soc., Chem. Commun. 6 (1994) 721–722. [15] K.J. Barnham, M.I. Djuran, P.d.-S. Murdoch, J.D. Ranford, P.J. Sadler, J. Chem. Soc., Dalton Trans. 22 (1995) 3721–3726. [16] J.-M. Teuben, J. Reeijk, J. Biol. Inorg. Chem. 5 (2000) 463–468. [17] V. Marchan, V. Moreno, E. Pedroso, A. Grandas, Chem. Eur. J. 7 (2001) 808–815. [18] M. Hahn, M. Kleine, W.S. Sheldrick, J. Biol. Inorg. Chem. 6 (2001) 556–566. [19] O. Heudi, A. Cailleux, P. Allain, J. Inorg. Biochem. 71 (1998) 61–69. [20] A.I. Ivanov, J. Christodoulou, J.A. Parkinson, K.J. Barnahm, A. Tucker, J. Woodrow, P.J. Sadler, J. Biol. Chem. 273 (1998) 14721– 14730. [21] D. Gibson, C.E. Costello, Eur. Mass Spectrom. 5 (1999) 501–510. [22] S.J. Berners-Price, P.J. Sadler, Coord. Chem. Rev. 151 (1996) 1–40. [23] T. Peleg-Shulman, D. Gibson, J. Am. Chem. Soc. 123 (2001) 3171–3172. [24] S. Bamezai, M.A.T. Banez, E. Breslow, Biochemistry 29 (1990) 5389–5396. [25] S. Vijay-Kumar, C.E. Bugg, W.J. Cook, J. Mol. Biol. 194 (1987) 525–538. [26] D. Wolters, W.S. Sheldrick, J. Chem. Soc., Dalton Trans. 7 (1999) 1121–1129. [27] T.G. Appleton, J.W. Connor, J.R. Hall, Inorg. Chem. 27 (1988) 130–137. [28] E.L.M. Lempers, K. Inagaki, J. Reedijk, Inorg. Chim. Acta 152 (1988) 201–207.