Activation of Trans Geometry in Bifunctional Mononuclear Platinum Complexes by a Piperidine Ligand

A paradigm for the structure-pharmacological activity relationship of bifunctional platinum antitumor drugs is that the trans isomer of antitumor cisplatin (transplatin) is clinically ineffective. To this end, however, several new complexes of the trans structure have been identified that exhibit cytotoxicity in tumor cells that is even better than that of the analogous cis isomers. We reported recently (Kasparkova, J., Marini, V., Najajreh, Y., Gibson, D., and Brabec, V. (2003) Biochemistry 42, 6321–6332) that the replacement of one ammine ligand by the heterocyclic ligand, such as piperidine, piperazine, or 4-picoline in the molecule of transplatin resulted in a radical enhancement of its cytotoxicity. We examined oligodeoxyribonucleotide duplexes bearing a site-specific cross-link of the transplatin analogue containing the piperidine ligand by biochemical methods. The results indicate that in contrast to transplatin, trans-(PtCl2(NH3)(piperidine)) forms stable 1,3-intrastrand cross-links in double-helical DNA that distort DNA and are not readily removed from DNA by nucleotide excision repair system. Hence, the intrastrand cross-links of trans-(PtCl2(NH3)(piperidine)) could persist for a sufficiently long time, potentiating its toxicity toward tumor cells. trans-(PtCl2(NH3)(piperidine)) also forms in DNA minor interstrand cross-links that are similar to those of transplatin so that these adducts appear less likely candidates for genotoxic lesion responsible for antitumor effects of trans-(PtCl2(NH3)(piperidine)). Hence, the role of structurally unique intrastrand cross-links in the anti-tumor effects of transplatin analogues in which one ammine group is replaced by a heterocyclic ligand may predominate.

The more widespread clinical applicability of cis-diamminedichloroplatinum(II) (cisplatin) 1 (Fig. 1A) and its ana-logue cis-diamminecyclobutanedicarboxylatoplatinum(II) is limited to a relatively narrow range of tumors (1-3); some tumors have natural resistance to these platinum drugs, whereas others develop resistance after the initial treatment. Cisplatin also has limited solubility in aqueous solution and is administered intravenously. There are also significant problems in terms of inducing severe side effects (especially kidney damage and vomiting/nausea). The drawbacks coupled with cisplatin and cis-diamminecyclobutanedicarboxylatoplatinum(II) toxicity have been the impetus for the development of improved platinum drugs.
The paradigm for structure-pharmacological activity relationship of platinum complexes is that the trans isomer of cisplatin (transplatin) (Fig. 1A) is clinically ineffective. To this end, however, several new complexes of the trans structure have been identified that exhibit an enhanced cytotoxicity in tumor cell lines, such that cytotoxicity is equivalent or even better than that of the analogous cis isomers and, indeed, cisplatin itself (for reviews, see Refs. 4 -6). Examples of such new antitumor transplatinum compounds are (i) analogues containing planar amine ligand of general structure trans-(PtCl 2 (NH 3 )(L)), where L represents planar amine such as quinoline or thiazole; (ii) analogues containing iminoether groups of the general formula trans-(PtCl 2 (E-iminoether) 2 ) (trans-EE); and (iii) analogues with asymmetric aliphatic ligands and trans-(PtCl 2 (NH 3 )(L)), where L represents cyclohexylamine.
We have recently reported (7)(8)(9) that the replacement of one ammine ligand by the heterocyclic ligand, such as piperidine (pip) (Fig. 1A), piperazine, or 4-picoline in the molecule of transplatin results in a radical enhancement of its activity in tumor cell lines both sensitive and resistant to cisplatin. For instance, the IC 50 (the concentration of the compound that afforded 50% cell killing) of trans-(PtCl 2 (NH 3 )(pip)) in the A2780 cell line was more than 40 times lower than that of transplatin. Since DNA is considered the major pharmacological target of antitumor platinum drugs (10), it appears important that due to this replacement, the analogues of transplatin alter the properties of DNA in a markedly different way than the parent compound. This is an intriguing finding, because the same replacement in the molecule of antitumor cisplatin (Fig.  1A) results in a reduced activity of the drug in both sensitive and resistant cell lines (for instance, the IC 50 measured in A2780 cell line was ϳ12 times higher than that of cisplatin (9)). Thus, the latter observation represents an additional example of activation of trans geometry in bifunctional mononuclear platinum(II) compounds, although the reasons for this activation have not been completely clarified.
Preliminary mechanistic studies using the analogues of transplatin containing piperidine, piperazine, or 4-picoline nonleaving ligand (9) suggest that one strategy how to activate trans geometry in antitumor bifunctional platinum(II) compounds may consist in a chemical modification of the ineffective transplatin, resulting in an increased stability of its intrastrand cross-links (CLs) in double-helical DNA and/or in an increased efficiency to form interstrand CLs. Hence, the latter suggestion is consistent with the concept of designing new platinum drugs (acting by a new mechanism and with activity complementary to drugs such as cisplatin) based on the observation that the replacement of the ammine ligand in transplatin can modulate their DNA binding mode and consequently their activity in cancer cell lines (11).
Initial data on calf thymus and plasmid DNA globally modified in cell-free media by analogues of transplatin in which one ammine ligand was replaced by piperidine, piperazine, or 4-picoline have been reported in our previous work (9). The results have demonstrated that these analogues form on DNA mainly intrastrand and interstrand CLs (for instance trans-(PtCl 2 (NH 3 )(pip)) forms ϳ26% interstrand and ϳ59% intrastrand CLs if DNA is incubated with this compound for 48 h). These results are of fundamental importance, because the clinical ineffectivity of transplatin has been proposed to be associated with a low stability of its intrastrand CLs in double-helical DNA and in general with its reduced capability to form in double-helical DNA bifunctional adducts (12,13). Because structural details of individual DNA adducts formed by these analogues of transplatin are not yet available, it remains uncertain how these CLs affect conformation of DNA and how these alterations are further processed in the cells. Therefore, in order to shed light on the mechanism that underlies activity of transplatin analogues containing a heterocyclic ligand we examine in the present work in detail short oligodeoxyribonucleotide duplexes containing single, site-specific intrastrand or interstrand CL of the transplatin analogue containing the nonplanar piperidine ligand (Fig. 1A). The piperidine analogue of transplatin was chosen as the representative of this class of new platinum compounds. This choice was made because the changes in the activity in cancer cell lines and some features of its DNA binding mode in a cell-free medium due to the replacement of one ammine ligand by the heterocyclic group in the parent compound were most pronounced (9). We investigated how the CLs affect the local conformation of DNA (in particular, bending and unwinding) and how these adducts are stable in double-helical DNA and further processed by some cellular components in cell-free media.
Antitumor activity of platinum compounds is also affected by the factors that do not operate directly at the level of DNA adducts. Among these factors are also those that affect the amount of platinum complex that can reach target DNA in cancer cells by changing the cell accumulation of the complexes. Therefore, we also compared the cellular uptake of transplatin and its analogue containing the piperidine ligand and determined the amount of platinum bound to DNA in the cells treated with these platinum compounds.

EXPERIMENTAL PROCEDURES
Chemicals-trans-(PtCl 2 (NH 3 )(pip)), where pip represents piperidine (Fig. 1A), was prepared by the methods described in detail previously (8). Cisplatin and transplatin (Fig. 1A) were obtained from Sigma. The stock solutions of platinum compounds were prepared at the concentration of 5 ϫ 10 Ϫ4 M in 10 mM NaClO 4 and stored at 4°C in the dark. The synthetic oligodeoxyribonucleotides (Fig. 1B) were synthesized and purified as described previously (14). Human recombinant replication protein A (RPA) was purified from Escherichia coli (15) and was a kind gift of John J. Turchi. Human XPA protein was prepared and purified as in our previous paper (16). Restriction endonucleases, Klenow fragment of E. coli DNA polymerase I deficient in 3Ј 3 5Ј proofreading exonuclease activity (KF Ϫ ) and T4 polynucleotide kinase were purchased from New England Biolabs (Beverly, MA). Reverse transcriptase from human immunodeficiency virus type 1 (RT HIV-1) was from Calbiochem. Nonidet P-30 was from Fluka (Prague, Czech Republic). Acrylamide, bis(acrylamide), urea, and NaCN were from Merck. Radioactive products were from Amersham Biosciences. Proteinase K, RNase A, and ATP were from Roche Applied Science.
Measurements of Platinum Accumulation in CH1cisR Cells and Determination of Platinum Binding to DNA in Culture Cells-All details of these experiments were recently published (17,18).
Platinations of Oligonucleotides-The duplexes containing single, intrastrand CL of cisplatin, transplatin, or trans-(PtCl 2 (NH 3 )(pip)) in the top strand (the duplexes in Fig. 1B that contained two guanine (G) residues in the top strand) were prepared as described (9,19,20). The interstrand cross-linked duplexes were also prepared and characterized in the same way as described previously (21,22).
Hydroxyl Radical Footprinting of Interstrand CLs-Platinated (or unplatinated) oligodeoxyribonucleotide duplexes (their concentration was 6 nM) that had either the top or bottom strand 32 P-labeled at the 5Ј-end were dissolved in the medium of 50 mM NaCl, 10 mM Tris-HCl, pH 7.5. The cleavage of the phosphodiesteric bonds was performed by incubating the duplexes in 0.04 mM Fe(NH 4 ) 2 (SO 4 ) 2 , 0.08 mM EDTA, 0.03% H 2 O 2 , and 2 mM sodium ascorbate for 5 min at 20°C. The reaction was stopped by adding 15 mM thiourea, 3 mM EDTA, 0.3 M sodium acetate, and 0.3 mg tRNA/ml. After precipitation, the samples were loaded onto a 24% denaturing polyacrylamide (PAA)/8 M urea gel. Maxam-Gilbert sequencing reactions were run in parallel.
Inhibition of DNA Polymerization-The 23-or 30-mer templates (see Fig. 4) containing a single 1,3-GTG intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) or 1,2-GG intrastrand CL of cisplatin were prepared in the same way as described (9,19,20). 8-or 13-mer DNA primers whose sequences are also shown in Fig. 4 were complementary to the 3Ј termini of the 23-or 30-mer templates, respectively. The DNA substrates were formed by annealing templates and 5Ј-end-labeled primers at a molar ratio of 3:1. All experiments using KF Ϫ and RT HIV-1 were performed at 25°C in a volume of 50 l in a buffer containing 50 mM Tris-HCl (pH 7.4), 10 mM MgCl 2 , 0.1 mM dithiothreitol, 50 g/ml bovine serum albumin, 0.1% Nonidet P-30, 25 M dATP, 25 M dCTP, 25 M dGTP, and 25 M TTP and 0.5 unit of KF Ϫ . The experiments with RT HIV-1 were performed at 37°C using the same conditions, except that the nucleoside triphosphates were at a concentration of 100 M and 1.0 unit of RT HIV-1 was used. Reactions were terminated by the addition of EDTA, so that its resulting concentration was 20 mM, and heating at 100°C for 30 s. Products were resolved by denaturing 24% PAA/8 M urea gel and then visualized and quantified by using the FUJIFILM bioimaging analyzer and AIDA image analyzer software.
Gel Mobility Shift Assay-Reactions with RPA and XPA. 32 P-Labeled DNA substrates (0.5 nM) and the amounts of RPA or XPA indicated were incubated at 20°C in reactions of 20 l containing 25 mM HEPES-KOH (pH 8.3), 30 mM KCl, 4 mM MgCl 2 , 1 mM EDTA, 0.9 mM dithiothreitol, 45 g/ml bovine serum albumin, and 10% glycerol. The reaction with XPA was still supplemented by unlabeled and unplatinated competitor duplex (20-bp, 70 nM). To assess binding at equilibrium, reactions were stopped after 30 min by cooling the samples to 0°C. Following the addition of gel loading buffer (4 l) containing 100 mM Tris-HCl (pH 8.3), 10% glycerol, and 0.05% Orange G, the extent of binding was determined on 6% native PAA gel. The electrophoresis was performed for 50 min at 4°C, gels were dried and visualized by using the FUJIFILM bio-imaging analyzer, and the radioactivities associated with bands were quantitated with the AIDA image analyzer software.
Nucleotide Excision Assay-The 148-bp substrates containing single, central intrastrand or interstrand CL were assembled from three oligonucleotide duplexes as described previously (23,24).
Oligonucleotide excision reactions were performed in cell-free extracts (CFEs) prepared from the HeLa S3 and CHO AA8 cell lines as described (25,26). These extracts were kindly provided by J. T. Reardon and A. Sancar (University of North Carolina, Chapel Hill, NC). In vitro repair was measured with excision assay using these CFEs and 148-bp linear DNA substrates (see above) in the same way as described previously (26).

Cellular Platinum
Complex Uptake-Platinum complexes of novel structure may have an altered pharmacology associated with drug uptake that may affect their cytotoxicity. Therefore, we examined first how the replacement of the ammine group in transplatin by piperidine ligand affects its accumulation in tumor cell lines.
Results in CH1cisR cells (this cancer cell line was among those used for testing activity of the novel transplatin analogues in which trans-(PtCl 2 (NH 3 )(pip)) exhibited a markedly enhanced activity in comparison with transplatin (9)) show that the replacement of the ammine group by the piperidine ligand considerably reduces the amount of platinum associated with the cells (Fig. 2A). Given that significantly less trans-(PtCl 2 (NH 3 )(pip)) appears to be accumulated in cells, the question arises as to how the piperidine ligand affects the binding of transplatin analogues to DNA in cells. Results in Fig. 2B demonstrate that trans-(PtCl 2 (NH 3 )(pip)) binds to DNA in CH1cisR cells treated with this transplatin analogue more slowly and to a lesser extent than the parent complex. This is in contrast to the data showing that transplatin and trans-(PtCl 2 (NH 3 )(pip)) bind to DNA in a cell-free medium with approximately the same rate (9). Hence, it is reasonable to suggest that the slower binding rate and the lesser binding level of trans-(PtCl 2 (NH 3 )(pip)) to DNA in culture cells is mainly a consequence of its lower intracellular concentration. Fig. 2 shows that the markedly enhanced cytotoxicity of trans-(PtCl 2 (NH 3 )(pip)) in comparison with transplatin (9) is not due to either enhanced cellular accumulation or higher levels of DNA platination in the cells. Since the cytotoxicity cannot be correlated with the levels of DNA modification, it is intuitively appealing to suggest that the enhanced cytotoxicity may be related to the nature of the adducts that are formed with the DNA and to the effects of the altered DNA properties on downstream cellular events. Intrastrand Cross-links-Unlike cisplatin, due to steric reasons, transplatin and its analogues cannot form intrastrand CLs in double-helical DNA between adjacent base residues. The trans compounds can cross-link two bases on the same strand only if they are separated by at least one intervening base, forming mostly 1,3-GNG intrastrand CLs (where N represents adenine, cytosine, or thymine). These adducts formed by the antitumor analogues of transplatin, in which the ammine group was replaced by the heterocyclic ligand, are stable in double-helical DNA and represent major DNA adducts of this class of antitumor trans compounds. It was, therefore, of great interest to examine how single, site-specific 1,3-GTG intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) affects conformation of double-helical DNA and what are some subsequent "downstream" effects of this type of DNA damage, such as recognition by some damaged DNA-binding proteins and repair capacity for this lesion of eukaryotic nucleotide excision repair (NER) system. Important structural motifs induced in DNA by antitumor platinum compounds that play a significant role in the mechanism underlying their antitumor activity are the bending and unwinding of the helix axis (27). For DNA intrastrand adducts of cisplatin and transplatin, the structural details responsible for bending and unwinding have been elucidated (28,29). In this work we further performed studies on the bending and unwinding induced by single, site-specific intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) using electrophoretic retardation as a quantitative measure of the extent of planar curvature.
The oligodeoxyribonucleotide duplexes TGTGT(19 -22) (19 -22 bp long, whose sequences were identical or similar to that of the duplex TGTGT (21) shown in Fig. 1B; the 19-and 20-bp duplexes had one or two marginal C⅐G pairs deleted, respectively, whereas one additional T⅐A pair was added to one end in the 22-bp duplex) were used for the bending and unwinding studies of the present work. The ligation products of these unplatinated or CL containing duplexes were analyzed on native PAA electrophoresis gel. Experimental details of these studies are given in our recent reports (20,30,31). The DNA bending toward the minor groove and unwinding due to one 1,3-intrastrand adduct of trans-(PtCl 2 (NH 3 )(pip)) has been found 30 Ϯ 2 and 8 Ϯ 2°, respectively (the direction of the bend was determined using the duplex (TGTGTϩ(A/T) 5 )(32) (Fig.  1B) in the same way as in our recent papers (21,32,33)). Moreover, the ligation of the 21-and 22-bp duplexes containing 1,3-intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) resulted in the formation of circles, suggesting that the 1,3-intrastrand CLs of this transplatin analogue increased the flexibility of the double helix (29,34).
Further studies of the present work were focused on analysis of the distortion induced by the 1,3-intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) by chemical probes of DNA conformation. The duplex TGTGT (Fig. 1B) containing a single, site-specific adduct was treated with several chemical agents that are used as tools for monitoring the existence of conformations other than canonical B-DNA. These agents included KMnO 4 , diethyl pyrocarbonate, and bromine. They react preferentially with single-stranded DNA and distorted double-stranded DNA (34,35). We used for this analysis exactly the same methodology described in detail in our recent papers dealing with DNA adducts of various antitumor platinum drugs (21,31,34,36,37). Therefore, these experiments are only described in more detail in the Supporting material (see "Chemical Probes" and Fig. S2). The results schematically summarized in Fig. 3A indicate that 1,3-intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) induces in DNA the distortion that extends over at least 6 bp and is localized mainly at the base pair between the platinated G residues and the base pairs on its 5Ј side.
It has been demonstrated that various DNA secondary structures have significant effects on processivity of a number of prokaryotic, eukaryotic, and viral DNA polymerases (38,39). Interestingly, with DNA templates containing site-specifically placed adducts of various platinum compounds, a number of prokaryotic and eukaryotic DNA polymerases were blocked but could also traverse through platinum adducts, depending on their character and conformational alterations induced in DNA. It is therefore of great interest to examine whether DNA polymerases, processing DNA substrates containing either the 1,3-intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) or 1,2-GG intrastrand CL of cisplatin, could reveal potential differences in conformational alterations imposed on DNA by these two adducts. We investigated in the present work DNA polymerization using the templates site-specifically modified by trans-(PtCl 2 (NH 3 )(pip)) or cisplatin by two DNA polymerases, which differ in processivity and fidelity. In the first series of our experiments, we used Klenow fragment of DNA polymerase I as a model enzyme frequently used in the studies aimed at understanding the processes in which nucleic acid polymerases take part.
We constructed the 8-mer/23-mer primer-template duplexes TGTGT(KF) and TGGT(KF) (Fig. 4A)  or in 1,2-GG CL of cisplatin on the template strand was located at its 13th position from the 3Ј terminus (Fig. 4A). After annealing the 8-nucleotide primer to the 3Ј terminus of the un-platinated or platinated template strand (positioning the 3Јend of the primer five bases before the adduct in the template strand), we examined DNA polymerization through the single 1,3-CL of trans-(PtCl 2 (NH 3 )(pip)) or 1,2-intrastrand CL of cisplatin on the template by KF Ϫ in the presence of all four deoxyribonucleoside 5Ј-triphosphates. The reaction was stopped at various time intervals, and the products were analyzed using a sequencing gel (Fig. 4A). Polymerization using the template containing the CL of cisplatin proceeded rapidly up to the nucleotide preceding and at the sites opposite the CL, such that the 12-and 13-nucleotide products accumulated to a significant extent (shown in Fig. 4A,  lanes 6-10). There was only a slight accumulation of larger DNA intermediates, whereas no intermediate products were seen with the 23-mer control template as the full-length product was being formed (shown in Fig. 4A, lanes 1-5). The full-length products were also noticed with the 23-mer template containing the CL of cisplatin, although in a smaller amount. This result is in an agreement with previously published work (40) in which T7 DNA polymerase and RT HIV-1 were used and confirms that 1,2-GG intrastrand CL of cisplatin inhibits DNA synthesis (38), but translesion synthesis may occur. In contrast, under the same experimental conditions, DNA polymerization by KF Ϫ using the template containing the 1,3-intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) proceeded up to the nucleotide preceding and at the site opposite the 3Ј G involved in the CL (Fig. 4A, lanes 11-15). There was almost no accumulation of shorter and larger DNA intermediates, and importantly, no fulllength products accumulated. This result indicates that the character of the 1,3-GTG intrastrand CLs of trans-(PtCl 2 (NH 3 )(pip)) and alterations induced in DNA by this adduct are distinctly different from the features of the major adduct of cisplatin so that the adducts of trans-(PtCl 2 (NH 3 )(pip)) could potentially impede elongation of DNA to a higher extent than the major adducts of cisplatin (Fig. 4A).
We have also examined the effects of the 1,3-intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) on polymerization by RT HIV-1. This enzyme also possesses DNA template-dependent DNA polymerase activity but relatively low processivity and fidelity (41). In these studies, elongation of the 17-mer/30-mer primer-template duplexes was tested. As is demonstrated in Fig. 4B, we confirmed also by using this DNA polymerase showing a different mechanism underlying its catalytic activity than KF Ϫ that the 1,3-intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) constitutes a fairly strong block to DNA synthesis catalyzed by both DNA polymerases. Since there is a high degree of structural and sequence conservation of the domains among eukaryotic, prokaryotic, and viral polymerases (42), insights gleaned from studies of the KF Ϫ and RT HIV-1 should be also applicable to other DNA polymerases (43)(44)(45). Hence, the repercussion of stronger inhibition of DNA polymerization by the 1,3-intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) in comparison with the major adduct of cisplatin adds a new dimension to the impact of the activated trans geometry in platinum compounds on biological processes, possibly including replication or DNA repair.
An important feature of the mechanism that underlies the antitumor activity of cisplatin and its analogues is that the major adducts of these drugs (1,2-GG intrastrand CLs) are recognized by proteins containing HMG domains (27). Importantly, DNA modified by transplatin or monodentate platinum(II) compounds, such as chlorodiethylenetriamineplatinum(II) chloride or (PtCl(NH 3 ) 3 )Cl, is not recognized by these cellular proteins. We examined whether also the replacement of the ammine group in transplatin by piperidine (resulting in the enhancement of cytotoxicity in tumor cell lines) also affects affinity of HMG-box proteins to the intrastrand adduct of this transplatin analogue. The interactions of the rat HMGB1 domain A (HMGB1a) and HMGB1 domain B (HMGB1b) with the 1,3-GTG intrastrand CLs of trans-(PtCl 2 (NH 3 )(pip)) were investigated using a gel mobility shift assay (46,47). In these experiments (described in more detail in Supplemental Mate-rial under "Recognition by HMGB1 proteins" (Fig. S4)), the 20-bp duplex TGTGT (Fig. 1B) was modified so that it contained a single, site-specific 1,3-intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)). The binding of the HMGB1a and HMGB1b to these DNA probes was detected by retardation of the migration of the radiolabeled 20-bp probes through the gel (not shown) under identical conditions described in detail in our recent papers (21,33).
Consistent with the previous reports (21,46), a shifted band due to the incubation of the duplex containing 1,2-GG intrastrand CL of cisplatin with both HMGB1a and HMGB1b was observed, indicating that both proteins recognize the duplex containing the major adduct of cisplatin (Fig S4, lanes 5 and 6). These proteins exhibited under the same experimental conditions no binding to the 20-bp duplex unplatinated or containing the 1,3-intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) (Fig. S4,  lanes 8 and 9). These data indicate that HMGB1 proteins do not bind the probe containing the 1,3-GTG intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)).
NER is a pathway used by human cells for the removal of damaged nucleotides from DNA (48,49). In mammalian cells, this repair pathway is an important mechanism for the removal of bulky, helix distorting DNA adducts, such as those generated by various chemotherapeutics including cisplatin (50). Efficient repair of 1,2-GG or 1,3-GNG intrastrand CL of cisplatin has been reported by various NER systems including human and rodent excinucleases (26,(51)(52)(53)(54)(55). The results presented in Fig. 5A, lanes 4 and 8, are consistent with these reports. The major excision fragment contains 28 nucleotides, and other primary excision fragments are 23-27 nucleotides in length (26,56). Importantly, the 1,3-GTG intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) was also repaired by both human and rodent excinucleases, but with a markedly lower efficiency than the intrastrand CLs of cisplatin (shown in Fig. 5A, lane 6, and in Fig. 5B for the adduct repaired by rodent excinuclease).
The initial and rate-determining step of NER is the recognition of the damaged DNA (48,57). This recognition process involves multiple protein components, and, for example, RPA and XPA belong to the initial damage-sensing factors of eukaryotic excision nuclease initiating repair (16,58). These recognition proteins preferentially bind to damaged DNA so that binding of RPA and/or XPA to damaged DNA may be a sensitive indicator to predict whether mammalian NER could be effective in the removal of damaged nucleotides.
The binding of RPA is thought to proceed via the denaturation of the DNA substrate followed by high affinity binding to the single-stranded DNA (15,16,59). On the other hand, XPA binds most efficiently to rigidly bent DNA but not to singlestranded structures (16). To determine the effect of 1,3-intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) on RPA and XPA binding, the 20-bp TGTGT duplexes were prepared so that they contained a single, site-specific 1,3-GTG intrastrand CL of this compound. RPA or XPA binding to this duplex was assessed in a gel mobility shift assay. The duplex with blunt ends was purified so that it contained no contaminating single-stranded DNA, as was evident from analysis on native PAA gel (not shown). The 20-bp duplex TGGT containing single, site-specific 1,2-GG intrastrand CL of cisplatin was run as a positive control. The results of the gel mobility shift assay analysis (Fig. 6) demonstrated that increasing RPA or XPA concentrations resulted in the increasing amounts of these proteins bound to the 1,2-GG intrastrand CL of cisplatin, consistent with the previous observations (16). On the other hand, the same analysis performed with the substrate containing the single, site-specific 1,3-intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) revealed lower binding of RPA and no binding of XPA protein (Fig. 6).
These results are consistent with the view that the low efficiency of the mammalian NER system to remove the 1,3-intrastrand CLs of trans-(PtCl 2 (NH 3 )(pip)) is associated with inca-pability of the NER system to recognize this type of platinum damage.
Interstrand Cross-linking-Frequent, although not major, adducts formed by bifunctional antitumor analogues of transplatin containing a heterocyclic ligand are interstrand CLs (9,60). Interestingly, transplatin forms these CLs preferentially between G and complementary C residues (19), whereas quite surprisingly its analogues, such as trans-(PtCl 2 (NH 3 ) (quinoline)) or trans-(PtCl 2 (NH 3 )(thiazole)) (i.e. the trans complexes containing a planar heterocyclic ligand) form these CLs between G residues in the 5Ј-GC/5Ј-GC sequences (22) (i.e. "cisplatin-like" interstrand CLs) (61). Therefore, it was of interest to unambiguously identify the residues involved in the interstrand CL formed by trans-(PtCl 2 (NH 3 )(pip)). Diamminedichloroplatinum(II) complexes react with DNA in a twostep process (62). Monofunctional adducts are formed preferentially at N-7 atoms of G residues. These lesions subsequently close to bifunctional CLs (intrastrand and/or interstrand). Considering this fact, we have designed synthetic oligodeoxyribonucleotide duplexes (duplexes CGC, TGC, and TGT in Fig. 1B). The pyrimidine-rich top strands of these duplexes contain a unique G residue at which the monofunctional adduct of transplatin or its piperidine analogue was formed. Thus, the choice of this nucleotide allowed for a cross-linking study under competitive conditions (i.e. interstrand CLs were in principle possible: in the CGC duplex, between the central G in the top strand and either complementary C or adjacent 5Ј or 3Ј Gs on the opposite strand; in the case of the TGC duplex, between the central G in the top strand and complementary C or adjacent 5Ј G in the bottom strand; in the case of the TGT duplex, between the central G in the top strand and complementary C. The top strands of the duplexes containing the monofunctional adduct of transplatin or trans-(PtCl 2 (NH 3 )(pip)) were hybridized with their complementary (bottom), 5Ј-end 32 P-labeled strands. The mixtures were incubated at 37°C in 0.1 M NaClO 4 , and the aliquots were withdrawn at various time intervals and subjected to gel electrophoresis under denaturing (strand-separat- ing) conditions to separate and quantify the interstrand cross-linked duplexes. It was found (not shown) that the halftimes of the interstrand cross-linking reactions of trans-(PtCl 2 (NH 3 )(pip)) in the duplexes CGC, TGC, and TGT were 13, 15, and 26 h, respectively, whereas these half-times found for the cross-linking by transplatin were 1.5 times higher.
The interstrand cross-linked samples obtained after 48 h were further analyzed by hydroxyl radical footprinting (63). From the observation that all fragments were detected corresponding to the cleavage by hydroxyl radicals from the 5Ј-end up to the interstrand CL and separated according to size on a PAA gel (shown for the duplex TGC in Fig. 7), the exact location of the bases involved in the interstrand CL was deduced (63). The interstrand adducts formed by transplatin or trans-(PtCl 2 (NH 3 )(pip)) in all duplexes were unambiguously identified as the interstrand adducts involving the central G site in the top strand and its complementary cytosine residue in the bottom strand. Thus, substitution of NH 3 by a nonplanar piperidine produces "transplatin-like" interstrand CLs (i.e. this replacement does not change the sites involved in the CL in contrast to the substitution by quinoline or thiazole ligand) (22).
For DNA interstrand adducts of cisplatin and transplatin, the structural details responsible for bending and unwinding have recently been elucidated (34, 64 -66). We further performed studies on the bending and unwinding induced by single, site-specific interstrand CL of trans-(PtCl 2 (NH 3 )(pip)) using electrophoretic retardation as a quantitative measure of the extent of planar curvature in the same way as described in the present work for 1,3-intrastrand CL (see below).
The oligodeoxyribonucleotide duplexes TGCT(19 -22) (19 -22 bp long) were used for the bending and unwinding studies of the present work. The DNA unwinding due to one interstrand adduct of trans-(PtCl 2 (NH 3 )(pip)) has been found to be 20 Ϯ 4°. Moreover, the interstrand CL of trans-(PtCl 2 (NH 3 )(pip)) bends DNA by about 26 -30°toward the minor groove (the direction of the bend was determined using the duplex (TGCϩ(A/T) 5 )(32) (Fig. 1B) in the same way as in our recent papers (32)). Thus, the bending and local unwinding induced by the interstrand CL of trans-(PtCl 2 (NH 3 )(pip)) are very similar to those afforded by the interstrand CL of transplatin using the same experimental procedure (34). Further studies of the present work were focused on analysis of the distortion induced by the interstrand CL of trans-(PtCl 2 (NH 3 )(pip)) by chemical probes of DNA conformation also in the same way as described in the present work for 1,3intrastrand CL (see above). The results described in more detail in the Supplemental Material (see "Chemical Probes" and Fig. S3) and summarized in Fig. 3, B and C, indicate that this adduct induces in DNA the distortion that extends over at least 5 bp and is localized mainly at the platinated base pairs. The distortion induced by the CL of trans-(PtCl 2 (NH 3 )(pip)) was stronger than that induced by the parent compound.
We also examined how the interstrand CL of trans-(PtCl 2 (NH 3 )(pip)) is recognized by DNA-binding proteins, such as HMGB1 domain and RPA and XPA proteins using the same experimental approach as demonstrated in the present work for 1,3-intrastrand CL of this transplatin analogue. The data indicate that HMGB1a, HMGB1b (Fig. S4, lanes 11 and 12), and RPA proteins (Fig. 6A) do not bind the probe containing the interstrand CL of trans-(PtCl 2 (NH 3 )(pip)). XPA exhibited affinity to the probe containing this lesion, but it was markedly lower than that to the probe containing the 1,2-GG intrastrand CL of cisplatin (ϳ50%) (Fig. 6B).
Excision repair substrates containing a site-specific interstrand CL of trans-(PtCl 2 (NH 3 )(pip)) or transplatin were also prepared and analyzed using both human and rodent excinucleases. No excision products were detected under conditions when 1,2-or 1,3-intrastrand CLs were readily excised (shown in Fig. 5A, lanes 10 and 12 for the CLs treated with rodent excinuclease).

DISCUSSION
The results of the present work (Fig. 2) demonstrate that the replacement of one of the ammine group in transplatin results in the reduced cellular accumulation of this drug in comparison with the parent compound, which, however, does not correlate with its markedly enhanced activity in the cancer cell lines (9). It is therefore reasonable to expect that there are other biochemical factors dominating the mechanism of action of transplatin analogues containing a heterocyclic nonleaving ligand, such as piperidine in tumor cells. DNA is a major pharmacological target of platinum compounds (10). Hence, among these factors might be also those associated with the modulation of the platinum-DNA interaction, with subsequent effects on further "downstream" effects of damaged DNA, such as for instance repair capacity for the platinum-DNA lesions (11).
The results of the interstrand cross-linking assay (Fig. 7) and transcription mapping experiments (Fig. S1) are consistent with the view that the replacement of one NH 3 nonleaving group in transplatin by piperidine has not significantly altered base sequence selectivity of the parent platinum drug and the spectrum of its DNA adducts. On the other hand, the results of our recent work (9) suggest that the rate of the formation of the bifunctional adducts is considerably enhanced by this replacement.
It has been also demonstrated (67)  stable intrastrand CLs, and this property of transplatin has been related to its clinical inefficiency (13,68). We have demonstrated in our recent work (9) that the replacement of one ammine ligand in "classical" transplatin by piperidine, piperazine, or 4-picoline ligand results in a distinctively enhanced stability of the 1,3-GNG intrastrand CLs formed by these compounds in several sequence contexts in short oligodeoxyribonucleotide duplexes. This result correlates with the markedly enhanced activity of these transplatin analogues in tumor cell lines.
The characteristics of this most frequent 1,3-intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) are summarized and compared with the interstrand CLs of this compound and transplatin and 1,2-GG intrastrand CL of cisplatin in Table I. There are several reports (12,28,29,69) describing properties of short oligodeoxyribonucleotide duplexes containing single, site-specific 1,3intrastrand CLs of transplatin. These CLs were, however, formed in the sequence TGTGT in which these adducts are unstable in double-helical DNA and readily isomerize in interstrand CLs (9,67). Thus, no reliable data on 1,3-intrastrand CLs of transplatin in double-helical DNA are available so that a comparison of the 1,3-intrastrand CLs of trans-(PtCl 2 (NH 3 )(pip)) and transplatin is impossible.
The bending experiments were carried out with the oligodeoxyribonucleotide duplexes containing the unique intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) in their central sequence. The phasing assay has revealed that the 1,3-GTG intrastrand CL results in a directional bending of helix axis (30°toward the minor groove) and a relatively small duplex unwinding (8°). In addition to these bending and unwinding effects, the 1,3-intrastrand CLs formed by trans-(PtCl 2 (NH 3 )(pip)) create rather extensive local conformational distortions revealed by the chemical probes extending over 6 bp (Fig. 3A).
It has been suggested (46,70) that HMG domain proteins play a role in sensitizing cells to cisplatin. It has been shown that HMG domain proteins recognize and bind to DNA CLs formed by cisplatin between bases in neighboring base pairs (46,70,71). The molecular basis for this recognition is still not entirely understood, although several structural details of the 1:1 complex formed between HMG domain and the duplex containing 1,2-GG intrastrand CL of cisplatin were recently elucidated (46). The details of how the binding of HMG domain proteins to cisplatin-modified DNA sensitize tumor cells to cisplatin are also still not completely resolved, but possibilities such as shielding cisplatin-DNA adducts from excision repair or that these proteins could be recruited from their native transcriptional regulatory function have been suggested (6,70,72,73) as clues for how these proteins are involved in the antitumor activity.
An important structural motif recognized by HMG domain proteins on DNA containing the major 1,2-GG intrastrand CL of cisplatin is a stable, directional bend of the helix axis toward the major groove. As demonstrated in the present work (Table  I) the 1,3-GTG intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) bends the helix axis almost as efficiently as the intrastrand CLs of cisplatin (29,74). However, no recognition of DNA intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) by HMGB1 proteins was observed in the present work. A plausible explanation of this observation may be that the bending due to the 1,3-intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) is in the opposite direction from that due to the 1,2-intrastrand CL of cisplatin. Hence, it is possible that trans-(PtCl 2 (NH 3 )(pip)) in the 1,3-intrastrand CL prevents the DNA bending toward the major groove required for its accommodation in the complex with HMGB1 protein. Thus, from the results of the present work, it is clear that the DNA intrastrand CLs of antitumor trans-(PtCl 2 (NH 3 )(pip)) are not a substrate for recognition by HMG domain proteins (Fig.  S4). From these considerations and from the fact that also interstrand CLs of trans-(PtCl 2 (NH 3 )(pip)) are not recognized by HMG domain proteins, we could conclude that the mechanism of antitumor activity of trans-(PtCl 2 (NH 3 )(pip)) does not involve recognition of its DNA adducts by HMG domain proteins as a crucial step, in contrast to the proposals for cisplatin and its direct analogues (70).
Several reports have demonstrated (26,52,53) that NER is a major mechanism contributing to cisplatin resistance. The examinations of excision repair of 1,3-intrastrand CL of trans-(PtCl 2 (NH 3 )(pip)) have revealed that these adducts cannot be removed so readily by excision repair as intrastrand adducts of antitumor cisplatin (Fig. 5). Hence, the intrastrand CLs of bifunctional trans-(PtCl 2 (NH 3 )(pip)) would not have to be shielded by damaged DNA recognition proteins, such as those containing HMG domains, as efficiently as 1,2-intrastrand CLs of cisplatin to prevent their repair. It is reasonable to suggest that intrastrand CLs of trans-(PtCl 2 (NH 3 )(pip)) could persist for a sufficiently long time even without being shielded by HMG box proteins, which would potentiate its toxicity toward tumor cells sensitive to this drug. trans-(PtCl 2 (NH 3 )(pip)) also forms in DNA minor interstrand CLs. Their basic characteristics are also summarized in Table  I and compared with those of transplatin. The properties of interstrand CLs of trans-(PtCl 2 (NH 3 )(pip)) and clinically ineffective transplatin investigated in the present work are very similar. The only more pronounced difference consists in a higher efficiency of trans-(PtCl 2 (NH 3 )(pip)) to form this type of the adducts. Despite this difference, the interstrand CLs remain minor adducts of trans-(PtCl 2 (NH 3 )(pip)). Hence, the interstrand CLs are probably less likely candidates for genotoxic lesion responsible for antitumor effects of this compound. Nonetheless, the cytotoxic effects of trans-(PtCl 2 (NH 3 )(pip)) may realistically be due to a cumulative effect of the structurally heterogeneous adducts produced by this drug, but the role of structurally unique intrastrand CLs in the antitumor effects of transplatin analogues in which one ammine group is replaced by a heterocyclic ligand may predominate.