- Email: [email protected]
Protein interactions with platinum–DNA adducts: from structure to function
JOURNAL OF
Inorganic Biochemistry Journal of Inorganic Biochemistry 98 (2004) 1551–1559 www.elsevier.com/locate/jinorgbio
Protein interactions with platinum–DNA adducts: from structure to function Stephen G. Chaney *, Sharon L. Campbell, Brenda Temple, Ekaterina Bassett, Yibing Wu, Mihir Faldu Department of Biochemistry and Biophysics, Lineberger Comprehensive Cancer Center and Curriculum in Toxicology, CB #7260 Mary Ellen Jones Building, University of North Carolina, Chapel Hill, NC 27599-7260, USA Received 13 February 2004; received in revised form 15 April 2004; accepted 17 April 2004 Available online 15 June 2004
Abstract Because of the efficacy of cisplatin and carboplatin in a wide variety of chemotherapeutic regimens, hundreds of platinum(II) and platinum(IV) complexes have been synthesized and evaluated as anticancer agents over the past 30 years. Of the many third generation platinum compounds evaluated to date, only oxaliplatin has been approved for clinical usage in the United States. Thus, it is important to understand the mechanistic basis for the differences in efficacy, mutagenicity and tumor range between cisplatin and oxaliplatin. Cisplatin and oxaliplain form the same types of adducts at the same sites on DNA. The most abundant adduct for both compounds is the Pt-GG intrastrand diadduct. Cisplatin-GG adducts are preferentially recognized by mismatch repair proteins and some damage-recognition proteins, and this differential recognition of cisplatin- and oxaliplatin-GG adducts is thought to contribute to the differences in cytotoxicity and tumor range of cisplatin and oxaliplatin. A detailed kinetic analysis of the insertion and extension steps of dNTP incorporation in the vicinity of the adduct shows that both pol b and pol g catalyze translesion synthesis past oxaliplatin-GG adducts with greater efficiency than past cisplatin-GG adducts. In the case of pol g, the efficiency and fidelity of translesion synthesis in vitro is very similar to that previously observed with cyclobutane TT dimers, suggesting that pol g is likely to be involved in error-free bypass of Pt adducts in vivo. This has been confirmed for cisplatin by comparing the cisplatininduced mutation frequency in human fibroblast cell lines with and without pol g. Thus, the greater efficiency of bypass of oxaliplatin-GG adducts by pol g is likely to explain the lower mutagenicity of oxaliplatin compared to cisplatin. The ability of these cellular proteins to discriminate between cisplatin and oxaliplatin adducts suggest that there exist significant conformational differences between the adducts, yet the crystal structures of the cisplatin- and oxaliplatin-GG adducts were very similar. We have recently solved the solution structure of the oxaliplatin-GG adduct and have shown that it is significantly different from the previously published solution structures of the cisplatin-GG adducts. Furthermore, the observed differences in conformation provide a logical explanation for the differential recognition of cisplatin and oxaliplatin adducts by mismatch repair and damage-recognition proteins. Molecular modeling studies are currently underway to analyze the mechanistic basis for the differential bypass of cisplatin and oxaliplatin adducts by DNA polymerases. Ó 2004 Elsevier Inc. All rights reserved. Keywords: Platinum; Cisplatin; Oxaliplatin; DNA polymerase; NMR
1. Introduction Cis-diamminedichloroplatinum(II) (cisplatin) and cisdiammine-1,1-cyclobutane dicarboxylate (carboplatin) *
Corresponding author. Tel.: +1-919-966-3286; fax: +1-919-9662852. E-mail address: [email protected] (S.G. Chaney). 0162-0134/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.jinorgbio.2004.04.024
are widely used in chemotherapy, and are particularly effective in the treatment of testicular, ovarian, head, neck and non-small cell lung cancer. However, cisplatin and carboplatin have significant toxicity and are mutagenic in cell culture and animal model systems [1,2]. Resistance is also a major limitation of cisplatin and carboplatin chemotherapy, with many tumors displaying intrinsic resistance or acquiring resistance during the
1552
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559 O
H3N
Cl
H3 N
Cl
H 3N
Pt H3N
O Pt
cisplatin
NH2 Cl Cl Pt
O C Pt
O O
carboplatin
G
H3N
O
NH2 NH2
O C
NH2 Cl Cl
O
oxaliplatin
ormaplatin
NH2
Pt H3N
G Pt
G
cis-diammine Pt (CP) adducts
NH2
G
(trans-RR)-diaminocyclohexane Pt (OX) adducts
Fig. 1. Selected platinum compounds and their DNA adducts.
course of treatment [3–5]. Cisplatin and carboplatin share the cis-diammine carrier ligands (Fig. 1). They form the same Pt–DNA dducts in vivo and are generally not effective in cell lines or tumors that have developed resistance to either agent. (Trans-R,R)1,2-diaminocyclohexaneoxalatoplatinum (II) (oxaliplatin) is a second generation platinum complex that is often effective in cisplatin-resistant cell lines and tumors [6,7] and appears to be less mutagenic than cisplatin [8,9]. Oxaliplatin has recently been approved for the treatment of colon cancer in the US. Although oxaliplatin has been studied for over 30 years, it is still not clear why it is so much more effective than the hundreds of platinum analogs that have been evaluated over that time span. The elucidation of the molecular basis for the efficacy of oxaliplatin in cisplatin resistant tumors would represent a major advance that might allow the design of even more effective platinum drugs and/or the development of prognostic indicators capable of identifying those tumors most likely to respond to oxaliplatin-based therapy. Oxaliplatin and the related compound ormaplatin form Pt–DNA adducts with the (trans-R,R)1,2-diaminocyclohexane carrier ligand (Fig. 1). For simplicity the cis-diammine-Pt adducts will be referred to as cisplatin adducts even though they can be formed by both cisplatin and carboplatin, and the (trans-R,R)1,2-diaminocyclohexane-Pt adducts will be referred to as oxaliplatin adducts even though they can be formed by both oxaliplatin and ormaplatin. The cytotoxicity of platinum compounds is thought to result primarily from the formation of Pt–DNA adducts. The effectiveness of oxaliplatin in cisplatinresistant cell lines is thought to be due to repair or damage-recognition processes that discriminate between cisplatin and oxaliplatin adducts. This has been best established for mismatch repair. For example, the binding of the mismatch repair complex appears to
increase the cytotoxicity of Pt–DNA adducts [10–13], either by activating downstream signaling pathways that lead to apoptosis [14,15] or by causing ‘‘futile cycling’’ during translesion synthesis past Pt–DNA adducts [16]. These effects appear to be specific for cisplatin adducts. Thus, defects in mismatch repair increase net translesion synthesis past cisplatin DNA adducts [16], decrease cisplain-induced expression of damage-response genes [14], and increase cellular resistance to cisplatin adducts [10–12,17], but have no effect on oxaliplatin adducts [16]. As one might predict from these biological differences, hMSH2 [11] and MutS [18] bind with greater affinity to cisplatin adducts than to oxaliplatin adducts. Some damage-recognition proteins, especially those of the HMG-domain family, also bind more tightly to cisplatin adducts than to oxaliplatin adducts [19,20]. The biological consequences of these effects are less clear, but the binding of damage-recognition proteins to cisplatin–DNA adducts has been postulated to inhibit nucleotide excision repair [21,22] and/or translesion DNA synthesis [23,24]. Finally, since many of the damage-recognition proteins are low abundance transcription factors, some of the cytotoxicity of cisplatin– DNA adducts has been proposed to be due to their ability to ‘‘hijack’’ the transcription factors away from their natural promoters [20,25]. To the extent that these damage-recognition proteins bind with different affinities to cisplatin and oxaliplatin adducts, they could obviously influence the differential cytotoxicity of cisplatin and oxaliplatin adducts. For example, HMG1 has been shown to inhibit translesion synthesis past cisplatin adducts to a greater extent than past oxaliplatin adducts [24]. Finally, both pol g and pol b appear to discriminate between cisplatin and oxaliplatin adducts based on the kinetics of dCTP incorporation opposite the Pt-GG adduct [24,26,27]. However, the previous studies may
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559
have been incomplete because they only measured dNTP incorporation opposite the adducts, and recent studies have shown that extension from the adduct is also strongly inhibited [28]. The ability of translesion polymerases to discriminate between cisplatin and oxaliplatin adducts could help explain the greater mutagenicity of cisplatin compared to oxaliplatin. However, the significance of these observations with respect to translesion synthesis in vivo has not been previously described. We [29] and others [30,31] have found that cisplatin and oxaliplatin form the same types of adducts at the same sites on the DNA. Both cisplatin and oxaliplatin form approximately 60–65% intrastrand GG, 25–30% intrastrand AG, 5–10% intrastrand GNG, and 1–3% interstrand GG diadducts [32]. X-ray crystallographic structures have been reported for both the cisplatin-GG [33] and oxaliplatin-GG [34] adducts in the same dodecamer DNA sequence. The two structures were virtually identical [34]. However, all of the mismatch repair proteins, DNA polymerases, and damage-recognition proteins that discriminate between cisplatin and oxaliplatin adducts and for which structural information is available appear to bind DNA primarily in the minor groove and bend DNA in the direction of the major groove [35–38]. Thus, the basis for the differential recognition of cisplatin and oxaliplain adducts by these proteins was not clear from the existing crystal structures of the cisplatin and oxaliplatin adducts.
2. Results 2.1. The role of pol g and pol b in bypass of platinum adducts Several previous studies had suggested that a number of DNA polymerases bypassed oxaliplatin-GG adducts with greater efficiency than cisplatin-GG adducts [24,26,27]. However, there are three essential steps in translesion synthesis past a platinum-GG adduct: insertion of a dNTP opposite the 30 G, insertion of a dNTP opposite the 50 G, and extension from the 50 G. The earlier studies had only characterized the insertion steps opposite the 30 and 50 Gs, but more recent data suggested that extension from the 50 G of a platinum-GG adduct might also be impaired [28]. Thus, we have characterized the insertion and extension kinetics for both cisplatin and oxaliplatin adducts in the AGG sequence context by human pol g and pol b [39]. Table 1 shows the efficiency of insertion opposite the 30 G, 50 G, and 50 A for DNA containing cisplatin and oxaliplatin-AGG adducts compared to insertion opposite the same sequence in undamaged DNA. For pol g the platinum adducts have very little effect on insertion of dCTP opposite either the 30 or 50 G. Most of the inhibition of translesion synthesis
1553
Table 1 Efficiency of translesion DNA synthesis past Pt-AGG adducts by pol g and pol b 30 G (frel )a
50 G (frel )a
50 A (fext )b
Overall efficiencyc
Pol g Control Cisplatin Oxaliplatin
1 0.9 1
1 1 1
1 0.26 0.39
1 0.23 0.39
Pol b Control Cisplatin Oxaliplatin
1 0.011 0.022
1 0.022 0.026
1 0.29 0.19
1 0.7 104 1.1 104
a
frel refers to the insertion efficiency (kcat =Km ) for dCTP opposite the 3 G and 50 G of Pt-AGG adducts relative to the insertion efficiency opposite the same bases on undamaged DNA. b fext refers to the extension efficiency (kcat =Km ) for dTTP opposite the 50 A of Pt-AGG adducts relative to the insertion efficiency opposite the same base on undamaged DNA. c Overall efficiency of translesion synthesis (frel frel fext ) past PtAGG adducts compared to DNA synthesis over the same sequence on undamaged DNA. 0
occurs at the extension step opposite the 50 A, and the net efficiency of translesion synthesis for cisplatin- and oxaliplatin-GG adducts is 0.23 and 0.39, respectively. However, for pol b there is significant inhibition of dNTP insertion at every step, and the overall efficiency of translesion synthesis is 0.7–1.1 104 compared to undamaged DNA. Both pol g and pol b bypass oxaliplatin adducts better than cisplatin adducts. For pol g this discrimination occurs primarily at the extension step, while for pol b the discrimination occurs primarily at the insertion step opposite the 30 G. We also investigated the in vitro fidelity of translesion synthesis past Pt-GG adducts by pol b and pol g. For pol g the overall error frequency was 2 104 , which is probably sufficient for a translesion polymerase. For pol b, misinsertion was only observed opposite the 30 G and the overall error frequency was 1 106 . It is instructive to compare these data with previous studies where data exist for both the kinetics of translesion synthesis past various adducts by pol g in vitro and the role of pol g in bypassing those adducts in vivo (Table 2). For example, the reported efficiency of translesion synthesis past cyclobutane TT dimers by pol g in vitro ranges from 0.1 [40] to 0.6 [41,42], which is very comparable to our data for the efficiency of bypass of cisplatin and oxaliplatin GG adducts by pol g in vitro. The fidelity of translesion synthesis past cyclobutane TT dimers by pol g in vitro is 1 104 [43], which is also comparable to what we observed with cisplatin- and oxaliplatin-GG adducts [39]. Human pol g is known to bypass TT dimers in an error-free manner in vivo [44]. These data suggest that pol g is likely to be involved in relatively error-free translesion synthesis past Pt adducts in vivo. We have recently confirmed this hypothesis by measuring the cisplatin-induced mutation frequency in
1554
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559
Table 2 The relative efficiency of insertion and extension by pol g on templates damaged with various types of adducts compared to that on undamaged DNA Organism
Adduct
Insertion at the 30 base (frel )a
Insertion at the 50 base (frel )a
Extension from the 50 base (fext )b
Overallc
In vivo translesion synthesisd
Yeast Human Human Human Human Yeast
8-Oxo-G Cb TT dimer OX-GG CP-GG Cb TT dimer [6] TT
1 [58] 1 [41] 1 [39] 0.9 [39] 0.6 [40] 0.01 [59](pol g)
N/A 1 [41] 1 [39] 1 [39] 0.4 [40] 0.4 [59](pol f)
1 [58] 0.6 [42] 0.4 [39] 0.26 [39] 0.6 [39] N/A
1 0.6 0.4 0.23 0.1 0.004
Human
BPDE
0.01 [61]
N/A
0.01 [61]
0.0001
Error-free [58] Error-free [44] ? ? Error-free [44] Error-prone [60](pol g + pol f) None [45]
a
frel refers to the insertion efficiency (kcat =Km ) opposite the 30 and 50 bases of diadducts or the only base of monoadducts relative to the insertion efficiency opposite the same bases on undamaged DNA. b fext refers to the extension efficiency (kcat =Km ) from the 50 base of the adduct relative to the extension efficiency from the same base on undamaged DNA. c Overall efficiency of translesion synthesis (frel frel fext ) past the adducts compared to DNA synthesis over the same sequence on undamaged DNA. d The ability of pol g to catalyze either error-free or error-prone translesion DNA synthesis past that adduct in vivo.
XPV (pol g )) and normal (pol g +) human fibroblasts. The CP-induced mutation frequency was higher in the XPV (pol g )) cells than in the normal (pol g +) cells (Bassett et al., manuscript in preparation), which is similar to the response of these cells to UV irradiation [44]. For pol g the overall efficiency and fidelity of translesion synthesis past oxaliplatin adducts was greater than for cisplatin adducts [27], suggesting that oxaliplatin may be less mutagenic than cisplatin. This hypothesis is currently being evaluated by mutagenesis experiments in the same cell lines. In contrast, the efficiency of translesion synthesis past Pt GG adducts by pol b in vitro is only 0.7–1.1 104 , which is comparable to the efficiency of translesion synthesis past BPDE adducts by pol g in vitro (Table 2). Since pol g does not appear to be able to bypass BPDE adducts in vivo [45], these data suggest that pol b is much less likely to bypass Pt adducts in vivo at normal levels of expression. However, these data do not preclude the possibility that pol b could become involved in bypass of Pt adducts when it is overexpressed, as has been suggested by Canitrot and colleagues [46,47]. 2.2. NMR solution structure of the oxaliplatin-GG adduct As described above, the crystal structures of the oxaliplatin-GG and cisplatin-GG adducts were virtually identical, and did not offer a mechanistic rationale for the differential recognition of those adducts by mismatch repair proteins, damage-recognition proteins and DNA polymerases. Thus, we have determined the NMR solution structure of the oxaliplatin-GG adduct (Wu et al., manuscript submitted). The 15 lowest energy NMR structures are shown in Fig. 2. The average final structure obtained for the oxaliplatin-GG adduct was analyzed by 3-DNA [48] and MADBEND [49]. Based on sugar pucker, Zp and Chi (data not shown), the DNA structure appears to be mostly B-like throughout, con-
Fig. 2. The 15 lowest energy solution structures consistent with the NMR data for the oxaliplatin-GG 12-mer DNA adduct in the CCTCAGGCCTCC sequence context.
sistent with the previous NMR solution structures of the cisplatin-GG adducts [50,51]. However, the crystal structures of both the cisplatin-GG adduct [33] and the oxaliplatin-GG [34] adduct were mostly A-like at the 50 end, with a transition to more B-like at the 30 end. When the structure of the central 4 base pairs of the oxali-
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559
platin-GG adduct was compared to the same 4 base pairs in B-DNA, it was apparent that the binding of oxaliplatin to the N7 atoms of the central Gs introduces a number of structural distortions in the DNA. For example, at the A5 T20/G6 C19 base pair step there is a significant decrease in shift, slide and twist compared to B DNA. At the G6 C19/G7 C18 base pair step there is a significant decrease in slide and twist and a large positive roll compared to B-DNA. Finally, at the G7 C18/C8 G17 base pair step there is an increased slide and decreased shift compared to B-DNA. Selected structural parameters for our average structure are compared with the same parameters obtained from the two most recent cisplatin-GG solution structures [50,51] in Fig. 3. The minor groove width for all 2.0 1.5
Slide (Dy)
1.0 0.5 0.0 -0.5 -1.0 -1.5
Twist (Ω))
40 30 20
1555
three structures is compared in Fig. 4. Our average solution structure for the oxaliplatin-GG adduct appears to have less distortion at the 50 side of the adduct than the solution structures of the cisplatin-GG adducts, particularly the structure described by Marzilli et al. [50]. This can be best seen by comparing the slide for the A5 T20/G6 C19 base pair step of our oxaliplatin-GG adduct with the slide for the comparable base pair step in the structure reported for the cisplatin-GG adduct by Marzilli et al. [50]. Our average solution structure of the oxaliplatin-GG adduct also differs from the solution structures of all cisplatin-GG adducts in that: (1) there is a smaller GG dihedral angle and roll and more twist at the G6 C19/G6 C18 base pair step, (2) the minor groove is smaller, and (3) the overall bend angle is less. For example, the dodecamer sequence of our oxaliplatin-GG adduct is identical to the dodecamer sequence used for the cisplatin-GG adduct characterized by Gelasco and Lippard [51] except for the bases on either side of the adduct (AGGC vs TGGT). However, the bend angle is 31° for the oxaliplatin-GG adduct and 82° for the cisplatin-GG adduct. The differences in conformation at the A5 T20/G6 C19 and G6 C19/G7 C18 base pair steps are directly related to the differences in the NMR data (Wu et al., manuscript submitted). The differences in overall bend angle and minor groove width are not directly derived from the NMR data, but are consistent with the smaller roll and dihedral angle. It is possible that some of differences between our oxaliplatin-GG structure and previous NMR cisplatin-GG structures could be due to differences in sequence context. Thus, efforts are underway to determine the solution structure of the cisplatin-GG adduct in the same sequence context.
10
22
Minor Groove Width
0 60
Roll (ρ)
40 20 0 -20
20 18 16 14 12 10
-40 0
2
4
6
8
10
12
Base pair step Fig. 3. DNA estimations of slide, twist and roll for our average solution structure of the oxaliplatin-GG adduct and two previous solution structures of cisplatin-GG adducts: ( ) 3DNA parameters for our oxaliplatin-GG solution structure; ( ) DNA parameters for the cisplatin-GG solution structure reported by Gelasco and Lippard [51]; ( ) 3DNA parameters for the cisplatin-GG solution structure reported by Marzilli et al. [50].
2
4
6
8
10
Base pair step Fig. 4. DNA estimations of minor groove width of our average solution structure of the oxaliplatin-GG adduct and two previous solution structures of cisplatin-GG adducts: ( ) Minor groove width of our oxaliplatin-GG solution structure; ( ), Minor groove width of the cisplatin-GG solution structure reported by Gelasco and Lippard [51]; ( ) Minor groove width of the cisplatin-GG solution structure reported by Marzilli et al. [50].
1556
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559
Our average solution structure of the oxaliplatin-GG adduct is similar to the previously reported crystal structure of the oxaliplatin-GG adduct in the shift, slide, GG dihedral angle, and roll at the G6 C19/G7 C18 base pair step and the overall degree of bending. Our structure was also similar to the crystal structure of the oxaliplatin-GG adduct in that the diaminocyclohexane ring was in the chair conformation. However, our average solution structure is different from the crystal structure in that: (1) it is mostly B-like in conformation, (2) the minor groove is narrower, and (3) none of the diaminocyclohexane amino protons are hydrogen bonded to the DNA. The narrower minor groove probably reflects the fact that the conformation is more B-like in solution and more A-like in the crystal structures. 2.3. Molecular dynamics modeling of the oxaliplatin-GG adduct We also hypothesize that the bulky diaminocyclohexane ring may restrict the conformational flexibility of the oxaliplatin-GG adduct compared to the cisplatinGG adduct and those differences in conformational flexibility may also affect protein binding to the oxaliplatin and cisplatin DNA adducts. Molecular dynamics simulation is one approach that has been successfully used to estimate conformational flexibility of proteins and DNA in solution [52–54]. Elizondo-Riojas and Kozelka [52] have recently performed an AMBER simulation of the cisplatin-GG adduct and reported significant conformational flexibility on the 50 side of the adduct. A 2 ns unrestrained MD simulation of the same 12-mer oxaliplatin-GG adduct studied by NMR spectroscopy (Section 2.2) was performed with the SANDER module of the AMBER program version 7.0 and the extended AMBER parm99 force field (data not shown). The average structure over the last 1.5 ns of simulation was in excellent agreement with the NMR data. A comparison of our results with that reported by Elizondo-Riojas and Kozelka [52] shows consistency in mobility of one of the adducted guanines. ElizondoRiojas and Kozelka [52], working with a cisplatin adduct in a CGGC sequence context, reported mobility in the 50 G C base pair, with the base-pair sliding back and forth relative to the C G base pair along a direction roughly perpendicular to the local helix axis. All 3 Watson–Crick hydrogen bonds were maintained during this sliding motion. We modeled our average NMR structure of an oxaliplatin adduct in the AGGC sequence context. The Watson–Crick hydrogen bonds in the 50 G C base pair were not disrupted during the course of the simulation, as indicated by the trajectory of the G6 H1 to C19 H41 distance in Fig. 5. Also shown in Fig. 5 are coordinate trajectories for two of the interproton distances between the A5 T20 and G6 C19
base pairs. These are representative of a group of 9 highly correlated interproton distances between the two base pairs, which is consistent with the mobility of the 50 G–C base pair reported by Elizondo-Riojas and Kozelka [52]. We have not yet made a quantitative comparison of the mobility of the oxaliplatin 50 G C base pair with the mobility of the cisplatin 50 G C base pair. We also see correlation in the trajectories of other movements, which we intend to analyze further. Finally, in the crystal structure of the oxaliplatin adduct in the TGGT sequence context, Springler et al. [34] reported that the 30 amine of the diaminocyclohexane group formed a hydrogen bond with the O6 of the 30 G, but we were unable to detect that hydrogen bond by NMR (Section 2.2). A careful analysis of the trajectory shows that the closest proton of the 30 amine is within hydrogen bonding distance of the 30 G O6 only around 27% of the time, which is consistent with our NMR data and suggests that the crystal structure may have captured a minor conformation. NMR studies are also underway to characterize the conformational dynamics of oxaliplatin- and cisplatin-GG adducts. 2.4. Molecular dynamics simulation of the oxaliplatin-GG adduct in the active site of pol b As described above, the mechanistic basis for the differences in bypass of cisplatin- and oxaliplatin-GG adducts by pol b and g is not currently clear. As an approach to obtaining a better understanding of the mechanistic basis for these differences, we have begun a molecular dynamics simulation of oxaliplatin- and cisplatin-GG adducts in the active site of pol b. Pol b was a logical choice for these studies because of an abundance of crystal structures of pol b wih undamaged DNA in various conformations [38], and because of our extensive kinetic data comparing the bypass of cisplatin- and oxaliplatin-GG adducts by pol b [24,26,55]. Currently, we have completed a two ns unrestrained MD simulation of gapped oxaliplatin–DNA bound to pol b in the open conformation. The protein maintained secondary structure and interactions with the DNA throughout the simulation. The cyclohexane ring fits into the active site of pol b without any steric clashes, and could potentially be stabilized by van der Waals interaction with Ile33. A superimposition of pol b from the original crystal structure and the MD simulation showed that the only major difference was a change in the position of the 8 kDa domain relative to the rest of the protein. Fig. 6 shows a close-up view of the AGG sequence in the active site. As expected from the NMR structure of the oxaliplatin-GG adduct, the 50 (templating)G is very distorted because of the roll and twist induced by oxaliplatin (blue is undamaged DNA and red is oxaliplatin DNA). (For the colours, see the online version of the paper.) The distortions seen here for the oxalipl-
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559
1557
Fig. 5. Trajectories of selected interproton distances of the A5 G6:T20 C19 base pair during molecular dynamics simulation of the oxaliplatin-GG adduct. The A5 G6:T20 C19 base pair is shown, along with selected interproton trajectories during the last 1.5 ns of simulation. (A) The trajectory of distances between the G6 H1 and C19 H41 protons. (B) The trajectory of distances between the A5 H2 and G6 H10 protons. (C) The trajectory of distances between the C19 H10 and T20 H6 protons.
Fig. 6. The effect of the oxaliplatin-GG adduct on the DNA conformation in the pol b-DNA complex in the open conformation. The conformation of DNA in the active site of the pol b-DNA complex in the open conformation [38] (pdb ¼ 1bpx) is shown in dark blue. The conformation of DNA containing the oxaliplatin-GG adduct (green) following a 2 ns molecular dynamics simulation is shown in red. Only the template strand is shown, and the AGG sequence of the template strand is oriented with the 50 G as the templating base. Pol b bound to undamaged DNA is shown in light blue and pol b bound to the oxaliplatin-GG adduct is shown in cyan. The His 34 residue that stacks with the 50 A is also shown in light blue and cyan. (For a colour version of the figure see the online paper.)
atin-GG adduct in the active site of pol b are very similar to those observed in the recently reported crystal structure of the cyclobutane TT dimer in the active site of Dpo4 [56]. The position of the downstream 50 A is also affected by the oxaliplatin adduct, as is the His34 that normally stacks with the downstream base in the template strand [38]. This may contribute to the movement of the 8 kDa domain, but a more detailed analysis is planned. Preliminary analysis indicates that most contacts between the protein and the template strand of undamaged DNA are also maintained between pol b and oxaliplatin–DNA, with the exception of Lys 230 and Lys234, which normally interact with the template strand. This might affect the binding affinity of pol b for the damaged DNA, but is unlikely to affect catalysis. However, because of the distortion of the templating base by the oxaliplatin adduct, one would predict that a number of critical amino acid interactions and the orientation of the incoming dNTP could be affected in the closed conformation. Molecular dynamic simulations are currently being performed for the oxaliplatin-GG adduct in the closed conformation of pol b. Simulations are also being performed for the cisplatin-GG adduct in
1558
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559
both the open and closed conformations of pol b. Hopefully, a detailed comparison of the conformational distortions imposed on the active site of pol b by cisplatin- and oxaliplatin-GG adducts will provide insight into the preferential bypass of oxaliplatin-GG adducts by this polymerase.
3. Discussion Our solution structure of the oxaliplatin-GG adduct is relatively similar to the crystal structure of the oxaliplatin-GG adduct, but very different from all previous solution structures of the cisplatin-GG adducts. However, the crystal structure of the cisplatin-GG adduct [33] is also very different from the solution structures of the cisplatin-GG adduct in that it is less bent and has a narrower minor groove width [50,51]. This has been interpreted as suggesting that the crystal structure of the cisplatin-GG adduct may be influenced by crystal constraints that prevent it from bending to the extent seen in solution [50,51]. The similarity between the crystal structures of the oxaliplatin-GG and cisplatin-GG adducts also does not provide an explanation for the biological differences between those adducts. Thus, we believe that the comparison between our solution structure of the oxaliplatin-GG adduct and the previous solution structures of the cisplatin-GG adduct is more relevant for understanding the biological differences between the platinated adducts. For example, when comparing solution structures, the oxaliplatin-GG adduct bends the DNA by 31° in the direction of the major groove and has a relatively narrow minor groove, while the cisplatin-GG adducts bend the DNA by 60–80° and have a much wider minor groove [50,50,51]. The HMG domains of chromosomal HMG proteins such as HMG1 generally bind to prebent DNA and form complexes with a DNA bending angle in the range 70–120° [57]. Thus, HMG1 may bind preferentially to cisplatin adducts because the presence of the cisplatin adduct bends to DNA to approximately the same extent as in the complex that is formed when HMG1 binds to undamaged DNA. In addition, most of the interactions between the HMG1 DNA-binding domain and the DNA (both unplatinated DNA and DNA containing the cisplatin-GG adduct) occur in the minor groove and are facilitated by the hydrophobic face of the widened minor groove seen with the cisplatin-GG solution structures [37] In the case of HMG1 Domain A, the bending of the DNA appears to be caused by intercalation of Phe 37 between the two central Gs. The Phe 37 is inserted from the minor groove side of the DNA and binding of HMG1 to the DNA is stabilized by interactions between the N termini of both helices I and II with the sugar phosphate backbone of the minor groove on the 30 side of the adduct.
Similarly, hMSH2 [11] and MutS [18] have been shown to bind to cisplatin-GG adducts with greater affinity than to oxaliplatin-GG adducts, and the ability of the mismatch repair system to discriminate between cisplatin and oxaliplatin DNA adducts is thought to explain the greater efficacy of oxaliplatin in mismatch repair-deficient cell lines and tumors [10–13]. The crystal structure of MutS bound to DNA containing a mismatch [35] shows that the binding is primarily to the minor groove and results in 60° bend towards the major groove. In this case the bending of the DNA is caused by intercalation of Phe 39 into the site of the mismatch. Once again, the Phe 39 is inserted into the minor groove and the binding is stabilized by interaction between a b-sheet region of the protein with the minor groove on both sides of the adduct. Thus, the ability of HMGdomain DNA binding proteins and mismatch repair proteins to discriminate between cisplatin and oxaliplatin DNA adducts appears to correlate very well with the structural differences observed between our solution structure of the oxaliplatin-GG adduct and the previous solution structures of cisplatin-GG adducts [50,51]. Finally, we have also shown that several DNA polymerases are inhibited to a greater extent by cisplatin-GG adducts than by oxaliplatin-GG adducts [24,26,27]. The mechanistic basis for this differential inhibition of translesion DNA synthesis is not known. However, it is worth noting that most DNA polymerases appear to primarily interact with the minor groove and bend the DNA in the direction of the major groove. NMR or X-ray structures of the DNA polymerase– Pt–DNA complex would be most valuable for defining the mechanistic basis for the bypass of Pt–DNA adducts. However, such structural information is currently unavailable. Thus, molecular modeling studies to clarify the mechanistic basis for the differential inhibition of translesion DNA synthesis by cisplatin-GG and oxaliplatin-GG adducts are currently underway. Now that solution structures are available for both the oxaliplatin and cisplatin adducts, we feel that molecular modeling studies such as those already underway will provide a much better understanding of how the differences in conformation of cisplatin and oxaliplatin adducts affect their recognition by various cellular proteins. This information, in turn, should finally provide a clear mechanistic explanation for the differences in cytotoxicity, mutagenicity and tumor range between oxaliplatin and cisplatin.
4. Abbreviations Cisplatin Oxaliplatin
cis-diamminedichloroplatinum(II) (trans-R,R)1,2-diaminocyclohexaneoxalatoplatinum(II)
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559
Acknowledgements The authors thank Sanofi-Synthelabo for providing oxaliplatin and Pt(dach)PtCl2 ; Dr. Sam Wilson for providing DNA polymerase b; Dr. Fumio Hanaoka for providing DNA polymerase g; and Jody Havener for preparing the cisplatin- and oxaliplatin–DNA templates used in these studies. Support for this work was provided by USPHS Grant CA84480. References [1] M.H. Greene, J. Nat. Cancer Inst. 84 (1992) 306–312. [2] B.J.S. Sanderson, L.R. Ferguson, W.A. Denny, Mutat. Res.Fundam. Mol. Mech. Mut. 355 (1996) 59–70. [3] S.W. Johnson, R.F. Ozols, T.C. Hamilton, Cancer 71 (1993) 644– 649. [4] B.T. Hill, Int. J. Oncol. 9 (1996) 197–203. [5] M. Kartalou, J.M. Essigmann, Mutat. Res. 478 (2001) 23–43. [6] S.G. Chaney, A. Vaisman, J. Inorg. Biochem. 77 (1999) 71–81. [7] O. Rixe, W. Ortuzar, M. Alvarez, R. Parker, E. Reed, K. Paull, T. Fojo, Biochem. Pharmacol. 52 (1996) 1855–1865. [8] Y. Kidani, Drugs Future 14 (1989) 529–532. [9] W.R. Leopold, R.P. Batzinger, E.C. Miller, J.A. Miller, R.H. Earhardt, Cancer Res. 41 (1981) 4368–4377. [10] S. Aebi, B. Kurdihaidar, R. Gordon, B. Cenni, H. Zheng, D. Fink, R.D. Christen, C.R. Boland, M. Koi, R. Fishel, S.B. Howell, Cancer Res. 56 (1996) 3087–3090. [11] D. Fink, S. Nebel, S. Aebi, H. Zheng, B. Cenni, A. Nehme, R.D. Christen, S.B. Howell, Cancer Res. 56 (1996) 4881–4886. [12] D. Fink, H. Zheng, S. Nebel, P.S. Norris, S. Aebi, T.-P. Lin, A. Nehme, R.D. Christen, M. Haas, C.L. MacLeod, S.B. Howell, Cancer Res. 57 (1997) 1841–1845. [13] J.T. Drummond, A. Anthoney, R. Brown, P. Modrich, J. Biol. Chem. 271 (1996) 19645–19648. [14] A. Nehme, R. Baskaran, S. Aebi, D. Fink, S. Nebel, B. Cenni, J.Y.J. Wang, S.B. Howell, R.D. Christen, Cancer Res. 57 (1997) 3253–3257. [15] A. Nehme, R. Baskaran, S. Nebel, D. Fink, S.B. Howell, J.Y.J. Wang, R.D. Christen, Br. J. Cancer 79 (1999) 1104–1110. [16] A. Vaisman, M. Varchenko, A. Umar, T.A. Kunkel, J.I. Risinger, J.C. Barrett, T.C. Hamilton, S.G. Chaney, Cancer Res. 58 (1998) 3579–3585. [17] D.A. Anthoney, A.J. Mcilwrath, W.M. Gallagher, A.R.M. Edlin, R. Brown, Cancer Res. 56 (1996) 1374–1381. [18] Z.Z. Zdraveski, J.A. Mello, C.K. Farinelli, J.M. Essigmann, M.G. Marinus, J. Biol. Chem. 277 (2002) 1255–1260. [19] M. Wei, S.M. Cohen, A.P. Silverman, S.J. Lippard, J. Biol. Chem. 276 (2001) 38774–38780. [20] X. Zhai, H. Beckmann, H.-M. Jantzen, J.M. Essigmann, Biochemistry 37 (1998) 16307–16315. [21] J.C. Huang, D.B. Zamble, J.T. Reardon, S.J. Lippard, A. Sancar, Proc. Natl. Acad. Sci. USA 91 (1994) 10394–10398. [22] M.M. Mcanulty, S.J. Lippard, Mutat. Res.-DNA Repair 362 (1996) 75–86. [23] J.S. Hoffmann, D. Locker, G. Viliani, M. Leng, J. Mol. Biol. 270 (1997) 539–543. [24] A. Vaisman, S.E. Lim, S.M. Patrick, W.C. Copeland, D.C. Hinkle, J.J. Turchi, S.G. Chaney, Biochemistry 38 (1999) 11026–11039. [25] D.K. Treiber, X.Q. Zhai, H.M. Jantzen, J.M. Essigmann, Proc. Natl. Acad. Sci. USA 91 (1994) 5672–5676. [26] A. Vaisman, S.G. Chaney, J. Biol. Chem. 275 (2000) 13017–13025.
1559
[27] A. Vaisman, C. Masutani, F. Hanaoka, S.G. Chaney, Biochemistry 39 (2000) 4575–4580. [28] C. Masutani, R. Kusumoto, S. Iwai, F. Hanaoka, EMBO J. 19 (2000) 3100–3109. [29] J.D. Page, I. Husain, A. Sancar, S.G. Chaney, Biochemistry 29 (1990) 1016–1024. [30] M.M. Jennerwein, A. Eastman, A.R. Khokhar, Chem.-Biol. Interact. 70 (1989) 39–49. [31] J.M. Woynarowski, W.G. Chapman, C. Napier, M.C.S. Herzig, P. Juniewicz, Mol. Pharmacol. 54 (1998) 770–777. [32] A. Eastman, Pharmacol. Ther. 34 (1987) 155–166. [33] P.M. Takahara, C.A. Frederick, S.J. Lippard, J. Am. Chem. Soc. 118 (1996) 12309–12321. [34] B. Spingler, D.A. Whittington, S.J. Lippard, Inorg. Chem. 40 (2001) 5596–5602. [35] M.H. Lamers, A. Perrakis, J.H. Enzlin, H.H.K. Winterwerp, N. de Wind, T.K. Sixma, Nature 407 (2000) 711–717. [36] G. Obmolova, C. Ban, P. Hsieh, W. Yang, Nature 407 (2000) 703– 710. [37] U.M. Ohndorf, M.A. Rould, Q. He, C.O. Pabo, S.J. Lippard, Nature 399 (1999) 708–712. [38] M.R. Sawaya, R. Prasad, S.H. Wilson, J. Kraut, H. Pelletier, Biochemistry 36 (1997) 11205–11215. [39] E. Bassett, A. Vaisman, J.M. Havener, C. Masutani, F. Hanaoka, S.G. Chaney, Biochemistry 42 (2003) 14197–14206. [40] R. Kusumoto, C. Masutani, S. Iwai, F. Hanaoka, Biochemistry 41 (2002) 6090–6099. [41] R.E. Johnson, M.T. Washington, S. Prakash, L. Prakash, J. Biol. Chem. 275 (2000) 7447–7450. [42] M.T. Washington, R.E. Johnson, L. Prakash, S. Prakash, Proc. Natl. Acad. Sci. USA 98 (2001) 8355–8360. [43] T. Matsuda, K. Bebenek, C. Masutani, F. Hanaoka, T.A. Kunkel, Nature 404 (2000) 1011–1013. [44] V.M. Maher, L.M. Ouellette, R.D. Curren, J.J. McCormick, Nature 261 (1976) 593–595. [45] M. Cordeiro-Stone, J.C. Boyer, B.A. Smith, W.K. Kaufmann, Carcinogenesis 7 (1986) 1783–1786. [46] V. Bergoglio, Y. Canitrot, L. Hogarth, L. Minto, S.B. Howell, C. Cazaux, J.S Hoffmann, Oncogene 20 (2001) 6181–6187. [47] Y. Canitrot, C. Cazaux, M. Frechet, K. Bouayadi, C. Lesca, B. Salles, J.-S. Hoffmann, Proc. Natl. Acad. Sci. USA 95 (1998) 12586–12590. [48] X.J. Lu, W.K. Olson, Nucleic Acids Res. 31 (2003) 5108–5121. [49] A. Barbic, D.M. Crothers, J. Biomol. Struct. Dyn. 21 (2003) 89– 97. [50] L.G. Marzilli, J.S. Saad, Z. Kuklenyik, K.A. Keating, Y. Xu, J. Am. Chem. Soc. 123 (2001) 2764–2770. [51] A. Gelasco, S.J. Lippard, Biochemistry 37 (1998) 9230–9239. [52] M.A. Elizondo-Riojas, J. Kozelka, J. Mol. Biol. 314 (2001) 1227– 1243. [53] C.M. Reyes, P.A. Kollman, J. Mol. Biol. 297 (2000) 1145–1158. [54] M. Wu, S. Yan, D.J. Patel, N.E. Geacintov, S. Broyde, Nucleic Acids Res. 30 (2002) 3422–3432. [55] A. Vaisman, M.W. Warren, S.G. Chaney, J. Biol. Chem. 276 (2001) 18999–19005. [56] H. Ling, F. Boudsocq, B.S. Plosky, R. Woodgate, W. Yang, Nature 424 (2003) 1083–1087. [57] A. Travers, Curr. Opin. Struct. Biol. 10 (2000) 102–109. [58] L. Haracska, S.L. Yu, R.E. Johnson, L. Prakash, S. Prakash, Nature Gen. 25 (2000) 458–461. [59] R.E. Johnson, L. Haracska, S. Prakash, L. Prakash, Mol. Cell. Biol. 21 (2001) 3558–3563. [60] A. Bresson, R.P. Fuchs, EMBO J. 21 (2002) 3881–3887. [61] Y. Zhang, X. Wu, D. Guo, O. Rechkoblit, N.E. Geacintov, Z. Wang, Mutat. Res. 510 (2002) 23–35.
Inorganic Biochemistry Journal of Inorganic Biochemistry 98 (2004) 1551–1559 www.elsevier.com/locate/jinorgbio
Protein interactions with platinum–DNA adducts: from structure to function Stephen G. Chaney *, Sharon L. Campbell, Brenda Temple, Ekaterina Bassett, Yibing Wu, Mihir Faldu Department of Biochemistry and Biophysics, Lineberger Comprehensive Cancer Center and Curriculum in Toxicology, CB #7260 Mary Ellen Jones Building, University of North Carolina, Chapel Hill, NC 27599-7260, USA Received 13 February 2004; received in revised form 15 April 2004; accepted 17 April 2004 Available online 15 June 2004
Abstract Because of the efficacy of cisplatin and carboplatin in a wide variety of chemotherapeutic regimens, hundreds of platinum(II) and platinum(IV) complexes have been synthesized and evaluated as anticancer agents over the past 30 years. Of the many third generation platinum compounds evaluated to date, only oxaliplatin has been approved for clinical usage in the United States. Thus, it is important to understand the mechanistic basis for the differences in efficacy, mutagenicity and tumor range between cisplatin and oxaliplatin. Cisplatin and oxaliplain form the same types of adducts at the same sites on DNA. The most abundant adduct for both compounds is the Pt-GG intrastrand diadduct. Cisplatin-GG adducts are preferentially recognized by mismatch repair proteins and some damage-recognition proteins, and this differential recognition of cisplatin- and oxaliplatin-GG adducts is thought to contribute to the differences in cytotoxicity and tumor range of cisplatin and oxaliplatin. A detailed kinetic analysis of the insertion and extension steps of dNTP incorporation in the vicinity of the adduct shows that both pol b and pol g catalyze translesion synthesis past oxaliplatin-GG adducts with greater efficiency than past cisplatin-GG adducts. In the case of pol g, the efficiency and fidelity of translesion synthesis in vitro is very similar to that previously observed with cyclobutane TT dimers, suggesting that pol g is likely to be involved in error-free bypass of Pt adducts in vivo. This has been confirmed for cisplatin by comparing the cisplatininduced mutation frequency in human fibroblast cell lines with and without pol g. Thus, the greater efficiency of bypass of oxaliplatin-GG adducts by pol g is likely to explain the lower mutagenicity of oxaliplatin compared to cisplatin. The ability of these cellular proteins to discriminate between cisplatin and oxaliplatin adducts suggest that there exist significant conformational differences between the adducts, yet the crystal structures of the cisplatin- and oxaliplatin-GG adducts were very similar. We have recently solved the solution structure of the oxaliplatin-GG adduct and have shown that it is significantly different from the previously published solution structures of the cisplatin-GG adducts. Furthermore, the observed differences in conformation provide a logical explanation for the differential recognition of cisplatin and oxaliplatin adducts by mismatch repair and damage-recognition proteins. Molecular modeling studies are currently underway to analyze the mechanistic basis for the differential bypass of cisplatin and oxaliplatin adducts by DNA polymerases. Ó 2004 Elsevier Inc. All rights reserved. Keywords: Platinum; Cisplatin; Oxaliplatin; DNA polymerase; NMR
1. Introduction Cis-diamminedichloroplatinum(II) (cisplatin) and cisdiammine-1,1-cyclobutane dicarboxylate (carboplatin) *
Corresponding author. Tel.: +1-919-966-3286; fax: +1-919-9662852. E-mail address: [email protected] (S.G. Chaney). 0162-0134/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.jinorgbio.2004.04.024
are widely used in chemotherapy, and are particularly effective in the treatment of testicular, ovarian, head, neck and non-small cell lung cancer. However, cisplatin and carboplatin have significant toxicity and are mutagenic in cell culture and animal model systems [1,2]. Resistance is also a major limitation of cisplatin and carboplatin chemotherapy, with many tumors displaying intrinsic resistance or acquiring resistance during the
1552
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559 O
H3N
Cl
H3 N
Cl
H 3N
Pt H3N
O Pt
cisplatin
NH2 Cl Cl Pt
O C Pt
O O
carboplatin
G
H3N
O
NH2 NH2
O C
NH2 Cl Cl
O
oxaliplatin
ormaplatin
NH2
Pt H3N
G Pt
G
cis-diammine Pt (CP) adducts
NH2
G
(trans-RR)-diaminocyclohexane Pt (OX) adducts
Fig. 1. Selected platinum compounds and their DNA adducts.
course of treatment [3–5]. Cisplatin and carboplatin share the cis-diammine carrier ligands (Fig. 1). They form the same Pt–DNA dducts in vivo and are generally not effective in cell lines or tumors that have developed resistance to either agent. (Trans-R,R)1,2-diaminocyclohexaneoxalatoplatinum (II) (oxaliplatin) is a second generation platinum complex that is often effective in cisplatin-resistant cell lines and tumors [6,7] and appears to be less mutagenic than cisplatin [8,9]. Oxaliplatin has recently been approved for the treatment of colon cancer in the US. Although oxaliplatin has been studied for over 30 years, it is still not clear why it is so much more effective than the hundreds of platinum analogs that have been evaluated over that time span. The elucidation of the molecular basis for the efficacy of oxaliplatin in cisplatin resistant tumors would represent a major advance that might allow the design of even more effective platinum drugs and/or the development of prognostic indicators capable of identifying those tumors most likely to respond to oxaliplatin-based therapy. Oxaliplatin and the related compound ormaplatin form Pt–DNA adducts with the (trans-R,R)1,2-diaminocyclohexane carrier ligand (Fig. 1). For simplicity the cis-diammine-Pt adducts will be referred to as cisplatin adducts even though they can be formed by both cisplatin and carboplatin, and the (trans-R,R)1,2-diaminocyclohexane-Pt adducts will be referred to as oxaliplatin adducts even though they can be formed by both oxaliplatin and ormaplatin. The cytotoxicity of platinum compounds is thought to result primarily from the formation of Pt–DNA adducts. The effectiveness of oxaliplatin in cisplatinresistant cell lines is thought to be due to repair or damage-recognition processes that discriminate between cisplatin and oxaliplatin adducts. This has been best established for mismatch repair. For example, the binding of the mismatch repair complex appears to
increase the cytotoxicity of Pt–DNA adducts [10–13], either by activating downstream signaling pathways that lead to apoptosis [14,15] or by causing ‘‘futile cycling’’ during translesion synthesis past Pt–DNA adducts [16]. These effects appear to be specific for cisplatin adducts. Thus, defects in mismatch repair increase net translesion synthesis past cisplatin DNA adducts [16], decrease cisplain-induced expression of damage-response genes [14], and increase cellular resistance to cisplatin adducts [10–12,17], but have no effect on oxaliplatin adducts [16]. As one might predict from these biological differences, hMSH2 [11] and MutS [18] bind with greater affinity to cisplatin adducts than to oxaliplatin adducts. Some damage-recognition proteins, especially those of the HMG-domain family, also bind more tightly to cisplatin adducts than to oxaliplatin adducts [19,20]. The biological consequences of these effects are less clear, but the binding of damage-recognition proteins to cisplatin–DNA adducts has been postulated to inhibit nucleotide excision repair [21,22] and/or translesion DNA synthesis [23,24]. Finally, since many of the damage-recognition proteins are low abundance transcription factors, some of the cytotoxicity of cisplatin– DNA adducts has been proposed to be due to their ability to ‘‘hijack’’ the transcription factors away from their natural promoters [20,25]. To the extent that these damage-recognition proteins bind with different affinities to cisplatin and oxaliplatin adducts, they could obviously influence the differential cytotoxicity of cisplatin and oxaliplatin adducts. For example, HMG1 has been shown to inhibit translesion synthesis past cisplatin adducts to a greater extent than past oxaliplatin adducts [24]. Finally, both pol g and pol b appear to discriminate between cisplatin and oxaliplatin adducts based on the kinetics of dCTP incorporation opposite the Pt-GG adduct [24,26,27]. However, the previous studies may
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559
have been incomplete because they only measured dNTP incorporation opposite the adducts, and recent studies have shown that extension from the adduct is also strongly inhibited [28]. The ability of translesion polymerases to discriminate between cisplatin and oxaliplatin adducts could help explain the greater mutagenicity of cisplatin compared to oxaliplatin. However, the significance of these observations with respect to translesion synthesis in vivo has not been previously described. We [29] and others [30,31] have found that cisplatin and oxaliplatin form the same types of adducts at the same sites on the DNA. Both cisplatin and oxaliplatin form approximately 60–65% intrastrand GG, 25–30% intrastrand AG, 5–10% intrastrand GNG, and 1–3% interstrand GG diadducts [32]. X-ray crystallographic structures have been reported for both the cisplatin-GG [33] and oxaliplatin-GG [34] adducts in the same dodecamer DNA sequence. The two structures were virtually identical [34]. However, all of the mismatch repair proteins, DNA polymerases, and damage-recognition proteins that discriminate between cisplatin and oxaliplatin adducts and for which structural information is available appear to bind DNA primarily in the minor groove and bend DNA in the direction of the major groove [35–38]. Thus, the basis for the differential recognition of cisplatin and oxaliplain adducts by these proteins was not clear from the existing crystal structures of the cisplatin and oxaliplatin adducts.
2. Results 2.1. The role of pol g and pol b in bypass of platinum adducts Several previous studies had suggested that a number of DNA polymerases bypassed oxaliplatin-GG adducts with greater efficiency than cisplatin-GG adducts [24,26,27]. However, there are three essential steps in translesion synthesis past a platinum-GG adduct: insertion of a dNTP opposite the 30 G, insertion of a dNTP opposite the 50 G, and extension from the 50 G. The earlier studies had only characterized the insertion steps opposite the 30 and 50 Gs, but more recent data suggested that extension from the 50 G of a platinum-GG adduct might also be impaired [28]. Thus, we have characterized the insertion and extension kinetics for both cisplatin and oxaliplatin adducts in the AGG sequence context by human pol g and pol b [39]. Table 1 shows the efficiency of insertion opposite the 30 G, 50 G, and 50 A for DNA containing cisplatin and oxaliplatin-AGG adducts compared to insertion opposite the same sequence in undamaged DNA. For pol g the platinum adducts have very little effect on insertion of dCTP opposite either the 30 or 50 G. Most of the inhibition of translesion synthesis
1553
Table 1 Efficiency of translesion DNA synthesis past Pt-AGG adducts by pol g and pol b 30 G (frel )a
50 G (frel )a
50 A (fext )b
Overall efficiencyc
Pol g Control Cisplatin Oxaliplatin
1 0.9 1
1 1 1
1 0.26 0.39
1 0.23 0.39
Pol b Control Cisplatin Oxaliplatin
1 0.011 0.022
1 0.022 0.026
1 0.29 0.19
1 0.7 104 1.1 104
a
frel refers to the insertion efficiency (kcat =Km ) for dCTP opposite the 3 G and 50 G of Pt-AGG adducts relative to the insertion efficiency opposite the same bases on undamaged DNA. b fext refers to the extension efficiency (kcat =Km ) for dTTP opposite the 50 A of Pt-AGG adducts relative to the insertion efficiency opposite the same base on undamaged DNA. c Overall efficiency of translesion synthesis (frel frel fext ) past PtAGG adducts compared to DNA synthesis over the same sequence on undamaged DNA. 0
occurs at the extension step opposite the 50 A, and the net efficiency of translesion synthesis for cisplatin- and oxaliplatin-GG adducts is 0.23 and 0.39, respectively. However, for pol b there is significant inhibition of dNTP insertion at every step, and the overall efficiency of translesion synthesis is 0.7–1.1 104 compared to undamaged DNA. Both pol g and pol b bypass oxaliplatin adducts better than cisplatin adducts. For pol g this discrimination occurs primarily at the extension step, while for pol b the discrimination occurs primarily at the insertion step opposite the 30 G. We also investigated the in vitro fidelity of translesion synthesis past Pt-GG adducts by pol b and pol g. For pol g the overall error frequency was 2 104 , which is probably sufficient for a translesion polymerase. For pol b, misinsertion was only observed opposite the 30 G and the overall error frequency was 1 106 . It is instructive to compare these data with previous studies where data exist for both the kinetics of translesion synthesis past various adducts by pol g in vitro and the role of pol g in bypassing those adducts in vivo (Table 2). For example, the reported efficiency of translesion synthesis past cyclobutane TT dimers by pol g in vitro ranges from 0.1 [40] to 0.6 [41,42], which is very comparable to our data for the efficiency of bypass of cisplatin and oxaliplatin GG adducts by pol g in vitro. The fidelity of translesion synthesis past cyclobutane TT dimers by pol g in vitro is 1 104 [43], which is also comparable to what we observed with cisplatin- and oxaliplatin-GG adducts [39]. Human pol g is known to bypass TT dimers in an error-free manner in vivo [44]. These data suggest that pol g is likely to be involved in relatively error-free translesion synthesis past Pt adducts in vivo. We have recently confirmed this hypothesis by measuring the cisplatin-induced mutation frequency in
1554
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559
Table 2 The relative efficiency of insertion and extension by pol g on templates damaged with various types of adducts compared to that on undamaged DNA Organism
Adduct
Insertion at the 30 base (frel )a
Insertion at the 50 base (frel )a
Extension from the 50 base (fext )b
Overallc
In vivo translesion synthesisd
Yeast Human Human Human Human Yeast
8-Oxo-G Cb TT dimer OX-GG CP-GG Cb TT dimer [6] TT
1 [58] 1 [41] 1 [39] 0.9 [39] 0.6 [40] 0.01 [59](pol g)
N/A 1 [41] 1 [39] 1 [39] 0.4 [40] 0.4 [59](pol f)
1 [58] 0.6 [42] 0.4 [39] 0.26 [39] 0.6 [39] N/A
1 0.6 0.4 0.23 0.1 0.004
Human
BPDE
0.01 [61]
N/A
0.01 [61]
0.0001
Error-free [58] Error-free [44] ? ? Error-free [44] Error-prone [60](pol g + pol f) None [45]
a
frel refers to the insertion efficiency (kcat =Km ) opposite the 30 and 50 bases of diadducts or the only base of monoadducts relative to the insertion efficiency opposite the same bases on undamaged DNA. b fext refers to the extension efficiency (kcat =Km ) from the 50 base of the adduct relative to the extension efficiency from the same base on undamaged DNA. c Overall efficiency of translesion synthesis (frel frel fext ) past the adducts compared to DNA synthesis over the same sequence on undamaged DNA. d The ability of pol g to catalyze either error-free or error-prone translesion DNA synthesis past that adduct in vivo.
XPV (pol g )) and normal (pol g +) human fibroblasts. The CP-induced mutation frequency was higher in the XPV (pol g )) cells than in the normal (pol g +) cells (Bassett et al., manuscript in preparation), which is similar to the response of these cells to UV irradiation [44]. For pol g the overall efficiency and fidelity of translesion synthesis past oxaliplatin adducts was greater than for cisplatin adducts [27], suggesting that oxaliplatin may be less mutagenic than cisplatin. This hypothesis is currently being evaluated by mutagenesis experiments in the same cell lines. In contrast, the efficiency of translesion synthesis past Pt GG adducts by pol b in vitro is only 0.7–1.1 104 , which is comparable to the efficiency of translesion synthesis past BPDE adducts by pol g in vitro (Table 2). Since pol g does not appear to be able to bypass BPDE adducts in vivo [45], these data suggest that pol b is much less likely to bypass Pt adducts in vivo at normal levels of expression. However, these data do not preclude the possibility that pol b could become involved in bypass of Pt adducts when it is overexpressed, as has been suggested by Canitrot and colleagues [46,47]. 2.2. NMR solution structure of the oxaliplatin-GG adduct As described above, the crystal structures of the oxaliplatin-GG and cisplatin-GG adducts were virtually identical, and did not offer a mechanistic rationale for the differential recognition of those adducts by mismatch repair proteins, damage-recognition proteins and DNA polymerases. Thus, we have determined the NMR solution structure of the oxaliplatin-GG adduct (Wu et al., manuscript submitted). The 15 lowest energy NMR structures are shown in Fig. 2. The average final structure obtained for the oxaliplatin-GG adduct was analyzed by 3-DNA [48] and MADBEND [49]. Based on sugar pucker, Zp and Chi (data not shown), the DNA structure appears to be mostly B-like throughout, con-
Fig. 2. The 15 lowest energy solution structures consistent with the NMR data for the oxaliplatin-GG 12-mer DNA adduct in the CCTCAGGCCTCC sequence context.
sistent with the previous NMR solution structures of the cisplatin-GG adducts [50,51]. However, the crystal structures of both the cisplatin-GG adduct [33] and the oxaliplatin-GG [34] adduct were mostly A-like at the 50 end, with a transition to more B-like at the 30 end. When the structure of the central 4 base pairs of the oxali-
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559
platin-GG adduct was compared to the same 4 base pairs in B-DNA, it was apparent that the binding of oxaliplatin to the N7 atoms of the central Gs introduces a number of structural distortions in the DNA. For example, at the A5 T20/G6 C19 base pair step there is a significant decrease in shift, slide and twist compared to B DNA. At the G6 C19/G7 C18 base pair step there is a significant decrease in slide and twist and a large positive roll compared to B-DNA. Finally, at the G7 C18/C8 G17 base pair step there is an increased slide and decreased shift compared to B-DNA. Selected structural parameters for our average structure are compared with the same parameters obtained from the two most recent cisplatin-GG solution structures [50,51] in Fig. 3. The minor groove width for all 2.0 1.5
Slide (Dy)
1.0 0.5 0.0 -0.5 -1.0 -1.5
Twist (Ω))
40 30 20
1555
three structures is compared in Fig. 4. Our average solution structure for the oxaliplatin-GG adduct appears to have less distortion at the 50 side of the adduct than the solution structures of the cisplatin-GG adducts, particularly the structure described by Marzilli et al. [50]. This can be best seen by comparing the slide for the A5 T20/G6 C19 base pair step of our oxaliplatin-GG adduct with the slide for the comparable base pair step in the structure reported for the cisplatin-GG adduct by Marzilli et al. [50]. Our average solution structure of the oxaliplatin-GG adduct also differs from the solution structures of all cisplatin-GG adducts in that: (1) there is a smaller GG dihedral angle and roll and more twist at the G6 C19/G6 C18 base pair step, (2) the minor groove is smaller, and (3) the overall bend angle is less. For example, the dodecamer sequence of our oxaliplatin-GG adduct is identical to the dodecamer sequence used for the cisplatin-GG adduct characterized by Gelasco and Lippard [51] except for the bases on either side of the adduct (AGGC vs TGGT). However, the bend angle is 31° for the oxaliplatin-GG adduct and 82° for the cisplatin-GG adduct. The differences in conformation at the A5 T20/G6 C19 and G6 C19/G7 C18 base pair steps are directly related to the differences in the NMR data (Wu et al., manuscript submitted). The differences in overall bend angle and minor groove width are not directly derived from the NMR data, but are consistent with the smaller roll and dihedral angle. It is possible that some of differences between our oxaliplatin-GG structure and previous NMR cisplatin-GG structures could be due to differences in sequence context. Thus, efforts are underway to determine the solution structure of the cisplatin-GG adduct in the same sequence context.
10
22
Minor Groove Width
0 60
Roll (ρ)
40 20 0 -20
20 18 16 14 12 10
-40 0
2
4
6
8
10
12
Base pair step Fig. 3. DNA estimations of slide, twist and roll for our average solution structure of the oxaliplatin-GG adduct and two previous solution structures of cisplatin-GG adducts: ( ) 3DNA parameters for our oxaliplatin-GG solution structure; ( ) DNA parameters for the cisplatin-GG solution structure reported by Gelasco and Lippard [51]; ( ) 3DNA parameters for the cisplatin-GG solution structure reported by Marzilli et al. [50].
2
4
6
8
10
Base pair step Fig. 4. DNA estimations of minor groove width of our average solution structure of the oxaliplatin-GG adduct and two previous solution structures of cisplatin-GG adducts: ( ) Minor groove width of our oxaliplatin-GG solution structure; ( ), Minor groove width of the cisplatin-GG solution structure reported by Gelasco and Lippard [51]; ( ) Minor groove width of the cisplatin-GG solution structure reported by Marzilli et al. [50].
1556
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559
Our average solution structure of the oxaliplatin-GG adduct is similar to the previously reported crystal structure of the oxaliplatin-GG adduct in the shift, slide, GG dihedral angle, and roll at the G6 C19/G7 C18 base pair step and the overall degree of bending. Our structure was also similar to the crystal structure of the oxaliplatin-GG adduct in that the diaminocyclohexane ring was in the chair conformation. However, our average solution structure is different from the crystal structure in that: (1) it is mostly B-like in conformation, (2) the minor groove is narrower, and (3) none of the diaminocyclohexane amino protons are hydrogen bonded to the DNA. The narrower minor groove probably reflects the fact that the conformation is more B-like in solution and more A-like in the crystal structures. 2.3. Molecular dynamics modeling of the oxaliplatin-GG adduct We also hypothesize that the bulky diaminocyclohexane ring may restrict the conformational flexibility of the oxaliplatin-GG adduct compared to the cisplatinGG adduct and those differences in conformational flexibility may also affect protein binding to the oxaliplatin and cisplatin DNA adducts. Molecular dynamics simulation is one approach that has been successfully used to estimate conformational flexibility of proteins and DNA in solution [52–54]. Elizondo-Riojas and Kozelka [52] have recently performed an AMBER simulation of the cisplatin-GG adduct and reported significant conformational flexibility on the 50 side of the adduct. A 2 ns unrestrained MD simulation of the same 12-mer oxaliplatin-GG adduct studied by NMR spectroscopy (Section 2.2) was performed with the SANDER module of the AMBER program version 7.0 and the extended AMBER parm99 force field (data not shown). The average structure over the last 1.5 ns of simulation was in excellent agreement with the NMR data. A comparison of our results with that reported by Elizondo-Riojas and Kozelka [52] shows consistency in mobility of one of the adducted guanines. ElizondoRiojas and Kozelka [52], working with a cisplatin adduct in a CGGC sequence context, reported mobility in the 50 G C base pair, with the base-pair sliding back and forth relative to the C G base pair along a direction roughly perpendicular to the local helix axis. All 3 Watson–Crick hydrogen bonds were maintained during this sliding motion. We modeled our average NMR structure of an oxaliplatin adduct in the AGGC sequence context. The Watson–Crick hydrogen bonds in the 50 G C base pair were not disrupted during the course of the simulation, as indicated by the trajectory of the G6 H1 to C19 H41 distance in Fig. 5. Also shown in Fig. 5 are coordinate trajectories for two of the interproton distances between the A5 T20 and G6 C19
base pairs. These are representative of a group of 9 highly correlated interproton distances between the two base pairs, which is consistent with the mobility of the 50 G–C base pair reported by Elizondo-Riojas and Kozelka [52]. We have not yet made a quantitative comparison of the mobility of the oxaliplatin 50 G C base pair with the mobility of the cisplatin 50 G C base pair. We also see correlation in the trajectories of other movements, which we intend to analyze further. Finally, in the crystal structure of the oxaliplatin adduct in the TGGT sequence context, Springler et al. [34] reported that the 30 amine of the diaminocyclohexane group formed a hydrogen bond with the O6 of the 30 G, but we were unable to detect that hydrogen bond by NMR (Section 2.2). A careful analysis of the trajectory shows that the closest proton of the 30 amine is within hydrogen bonding distance of the 30 G O6 only around 27% of the time, which is consistent with our NMR data and suggests that the crystal structure may have captured a minor conformation. NMR studies are also underway to characterize the conformational dynamics of oxaliplatin- and cisplatin-GG adducts. 2.4. Molecular dynamics simulation of the oxaliplatin-GG adduct in the active site of pol b As described above, the mechanistic basis for the differences in bypass of cisplatin- and oxaliplatin-GG adducts by pol b and g is not currently clear. As an approach to obtaining a better understanding of the mechanistic basis for these differences, we have begun a molecular dynamics simulation of oxaliplatin- and cisplatin-GG adducts in the active site of pol b. Pol b was a logical choice for these studies because of an abundance of crystal structures of pol b wih undamaged DNA in various conformations [38], and because of our extensive kinetic data comparing the bypass of cisplatin- and oxaliplatin-GG adducts by pol b [24,26,55]. Currently, we have completed a two ns unrestrained MD simulation of gapped oxaliplatin–DNA bound to pol b in the open conformation. The protein maintained secondary structure and interactions with the DNA throughout the simulation. The cyclohexane ring fits into the active site of pol b without any steric clashes, and could potentially be stabilized by van der Waals interaction with Ile33. A superimposition of pol b from the original crystal structure and the MD simulation showed that the only major difference was a change in the position of the 8 kDa domain relative to the rest of the protein. Fig. 6 shows a close-up view of the AGG sequence in the active site. As expected from the NMR structure of the oxaliplatin-GG adduct, the 50 (templating)G is very distorted because of the roll and twist induced by oxaliplatin (blue is undamaged DNA and red is oxaliplatin DNA). (For the colours, see the online version of the paper.) The distortions seen here for the oxalipl-
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559
1557
Fig. 5. Trajectories of selected interproton distances of the A5 G6:T20 C19 base pair during molecular dynamics simulation of the oxaliplatin-GG adduct. The A5 G6:T20 C19 base pair is shown, along with selected interproton trajectories during the last 1.5 ns of simulation. (A) The trajectory of distances between the G6 H1 and C19 H41 protons. (B) The trajectory of distances between the A5 H2 and G6 H10 protons. (C) The trajectory of distances between the C19 H10 and T20 H6 protons.
Fig. 6. The effect of the oxaliplatin-GG adduct on the DNA conformation in the pol b-DNA complex in the open conformation. The conformation of DNA in the active site of the pol b-DNA complex in the open conformation [38] (pdb ¼ 1bpx) is shown in dark blue. The conformation of DNA containing the oxaliplatin-GG adduct (green) following a 2 ns molecular dynamics simulation is shown in red. Only the template strand is shown, and the AGG sequence of the template strand is oriented with the 50 G as the templating base. Pol b bound to undamaged DNA is shown in light blue and pol b bound to the oxaliplatin-GG adduct is shown in cyan. The His 34 residue that stacks with the 50 A is also shown in light blue and cyan. (For a colour version of the figure see the online paper.)
atin-GG adduct in the active site of pol b are very similar to those observed in the recently reported crystal structure of the cyclobutane TT dimer in the active site of Dpo4 [56]. The position of the downstream 50 A is also affected by the oxaliplatin adduct, as is the His34 that normally stacks with the downstream base in the template strand [38]. This may contribute to the movement of the 8 kDa domain, but a more detailed analysis is planned. Preliminary analysis indicates that most contacts between the protein and the template strand of undamaged DNA are also maintained between pol b and oxaliplatin–DNA, with the exception of Lys 230 and Lys234, which normally interact with the template strand. This might affect the binding affinity of pol b for the damaged DNA, but is unlikely to affect catalysis. However, because of the distortion of the templating base by the oxaliplatin adduct, one would predict that a number of critical amino acid interactions and the orientation of the incoming dNTP could be affected in the closed conformation. Molecular dynamic simulations are currently being performed for the oxaliplatin-GG adduct in the closed conformation of pol b. Simulations are also being performed for the cisplatin-GG adduct in
1558
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559
both the open and closed conformations of pol b. Hopefully, a detailed comparison of the conformational distortions imposed on the active site of pol b by cisplatin- and oxaliplatin-GG adducts will provide insight into the preferential bypass of oxaliplatin-GG adducts by this polymerase.
3. Discussion Our solution structure of the oxaliplatin-GG adduct is relatively similar to the crystal structure of the oxaliplatin-GG adduct, but very different from all previous solution structures of the cisplatin-GG adducts. However, the crystal structure of the cisplatin-GG adduct [33] is also very different from the solution structures of the cisplatin-GG adduct in that it is less bent and has a narrower minor groove width [50,51]. This has been interpreted as suggesting that the crystal structure of the cisplatin-GG adduct may be influenced by crystal constraints that prevent it from bending to the extent seen in solution [50,51]. The similarity between the crystal structures of the oxaliplatin-GG and cisplatin-GG adducts also does not provide an explanation for the biological differences between those adducts. Thus, we believe that the comparison between our solution structure of the oxaliplatin-GG adduct and the previous solution structures of the cisplatin-GG adduct is more relevant for understanding the biological differences between the platinated adducts. For example, when comparing solution structures, the oxaliplatin-GG adduct bends the DNA by 31° in the direction of the major groove and has a relatively narrow minor groove, while the cisplatin-GG adducts bend the DNA by 60–80° and have a much wider minor groove [50,50,51]. The HMG domains of chromosomal HMG proteins such as HMG1 generally bind to prebent DNA and form complexes with a DNA bending angle in the range 70–120° [57]. Thus, HMG1 may bind preferentially to cisplatin adducts because the presence of the cisplatin adduct bends to DNA to approximately the same extent as in the complex that is formed when HMG1 binds to undamaged DNA. In addition, most of the interactions between the HMG1 DNA-binding domain and the DNA (both unplatinated DNA and DNA containing the cisplatin-GG adduct) occur in the minor groove and are facilitated by the hydrophobic face of the widened minor groove seen with the cisplatin-GG solution structures [37] In the case of HMG1 Domain A, the bending of the DNA appears to be caused by intercalation of Phe 37 between the two central Gs. The Phe 37 is inserted from the minor groove side of the DNA and binding of HMG1 to the DNA is stabilized by interactions between the N termini of both helices I and II with the sugar phosphate backbone of the minor groove on the 30 side of the adduct.
Similarly, hMSH2 [11] and MutS [18] have been shown to bind to cisplatin-GG adducts with greater affinity than to oxaliplatin-GG adducts, and the ability of the mismatch repair system to discriminate between cisplatin and oxaliplatin DNA adducts is thought to explain the greater efficacy of oxaliplatin in mismatch repair-deficient cell lines and tumors [10–13]. The crystal structure of MutS bound to DNA containing a mismatch [35] shows that the binding is primarily to the minor groove and results in 60° bend towards the major groove. In this case the bending of the DNA is caused by intercalation of Phe 39 into the site of the mismatch. Once again, the Phe 39 is inserted into the minor groove and the binding is stabilized by interaction between a b-sheet region of the protein with the minor groove on both sides of the adduct. Thus, the ability of HMGdomain DNA binding proteins and mismatch repair proteins to discriminate between cisplatin and oxaliplatin DNA adducts appears to correlate very well with the structural differences observed between our solution structure of the oxaliplatin-GG adduct and the previous solution structures of cisplatin-GG adducts [50,51]. Finally, we have also shown that several DNA polymerases are inhibited to a greater extent by cisplatin-GG adducts than by oxaliplatin-GG adducts [24,26,27]. The mechanistic basis for this differential inhibition of translesion DNA synthesis is not known. However, it is worth noting that most DNA polymerases appear to primarily interact with the minor groove and bend the DNA in the direction of the major groove. NMR or X-ray structures of the DNA polymerase– Pt–DNA complex would be most valuable for defining the mechanistic basis for the bypass of Pt–DNA adducts. However, such structural information is currently unavailable. Thus, molecular modeling studies to clarify the mechanistic basis for the differential inhibition of translesion DNA synthesis by cisplatin-GG and oxaliplatin-GG adducts are currently underway. Now that solution structures are available for both the oxaliplatin and cisplatin adducts, we feel that molecular modeling studies such as those already underway will provide a much better understanding of how the differences in conformation of cisplatin and oxaliplatin adducts affect their recognition by various cellular proteins. This information, in turn, should finally provide a clear mechanistic explanation for the differences in cytotoxicity, mutagenicity and tumor range between oxaliplatin and cisplatin.
4. Abbreviations Cisplatin Oxaliplatin
cis-diamminedichloroplatinum(II) (trans-R,R)1,2-diaminocyclohexaneoxalatoplatinum(II)
S.G. Chaney et al. / Journal of Inorganic Biochemistry 98 (2004) 1551–1559
Acknowledgements The authors thank Sanofi-Synthelabo for providing oxaliplatin and Pt(dach)PtCl2 ; Dr. Sam Wilson for providing DNA polymerase b; Dr. Fumio Hanaoka for providing DNA polymerase g; and Jody Havener for preparing the cisplatin- and oxaliplatin–DNA templates used in these studies. Support for this work was provided by USPHS Grant CA84480. References [1] M.H. Greene, J. Nat. Cancer Inst. 84 (1992) 306–312. [2] B.J.S. Sanderson, L.R. Ferguson, W.A. Denny, Mutat. Res.Fundam. Mol. Mech. Mut. 355 (1996) 59–70. [3] S.W. Johnson, R.F. Ozols, T.C. Hamilton, Cancer 71 (1993) 644– 649. [4] B.T. Hill, Int. J. Oncol. 9 (1996) 197–203. [5] M. Kartalou, J.M. Essigmann, Mutat. Res. 478 (2001) 23–43. [6] S.G. Chaney, A. Vaisman, J. Inorg. Biochem. 77 (1999) 71–81. [7] O. Rixe, W. Ortuzar, M. Alvarez, R. Parker, E. Reed, K. Paull, T. Fojo, Biochem. Pharmacol. 52 (1996) 1855–1865. [8] Y. Kidani, Drugs Future 14 (1989) 529–532. [9] W.R. Leopold, R.P. Batzinger, E.C. Miller, J.A. Miller, R.H. Earhardt, Cancer Res. 41 (1981) 4368–4377. [10] S. Aebi, B. Kurdihaidar, R. Gordon, B. Cenni, H. Zheng, D. Fink, R.D. Christen, C.R. Boland, M. Koi, R. Fishel, S.B. Howell, Cancer Res. 56 (1996) 3087–3090. [11] D. Fink, S. Nebel, S. Aebi, H. Zheng, B. Cenni, A. Nehme, R.D. Christen, S.B. Howell, Cancer Res. 56 (1996) 4881–4886. [12] D. Fink, H. Zheng, S. Nebel, P.S. Norris, S. Aebi, T.-P. Lin, A. Nehme, R.D. Christen, M. Haas, C.L. MacLeod, S.B. Howell, Cancer Res. 57 (1997) 1841–1845. [13] J.T. Drummond, A. Anthoney, R. Brown, P. Modrich, J. Biol. Chem. 271 (1996) 19645–19648. [14] A. Nehme, R. Baskaran, S. Aebi, D. Fink, S. Nebel, B. Cenni, J.Y.J. Wang, S.B. Howell, R.D. Christen, Cancer Res. 57 (1997) 3253–3257. [15] A. Nehme, R. Baskaran, S. Nebel, D. Fink, S.B. Howell, J.Y.J. Wang, R.D. Christen, Br. J. Cancer 79 (1999) 1104–1110. [16] A. Vaisman, M. Varchenko, A. Umar, T.A. Kunkel, J.I. Risinger, J.C. Barrett, T.C. Hamilton, S.G. Chaney, Cancer Res. 58 (1998) 3579–3585. [17] D.A. Anthoney, A.J. Mcilwrath, W.M. Gallagher, A.R.M. Edlin, R. Brown, Cancer Res. 56 (1996) 1374–1381. [18] Z.Z. Zdraveski, J.A. Mello, C.K. Farinelli, J.M. Essigmann, M.G. Marinus, J. Biol. Chem. 277 (2002) 1255–1260. [19] M. Wei, S.M. Cohen, A.P. Silverman, S.J. Lippard, J. Biol. Chem. 276 (2001) 38774–38780. [20] X. Zhai, H. Beckmann, H.-M. Jantzen, J.M. Essigmann, Biochemistry 37 (1998) 16307–16315. [21] J.C. Huang, D.B. Zamble, J.T. Reardon, S.J. Lippard, A. Sancar, Proc. Natl. Acad. Sci. USA 91 (1994) 10394–10398. [22] M.M. Mcanulty, S.J. Lippard, Mutat. Res.-DNA Repair 362 (1996) 75–86. [23] J.S. Hoffmann, D. Locker, G. Viliani, M. Leng, J. Mol. Biol. 270 (1997) 539–543. [24] A. Vaisman, S.E. Lim, S.M. Patrick, W.C. Copeland, D.C. Hinkle, J.J. Turchi, S.G. Chaney, Biochemistry 38 (1999) 11026–11039. [25] D.K. Treiber, X.Q. Zhai, H.M. Jantzen, J.M. Essigmann, Proc. Natl. Acad. Sci. USA 91 (1994) 5672–5676. [26] A. Vaisman, S.G. Chaney, J. Biol. Chem. 275 (2000) 13017–13025.
1559
[27] A. Vaisman, C. Masutani, F. Hanaoka, S.G. Chaney, Biochemistry 39 (2000) 4575–4580. [28] C. Masutani, R. Kusumoto, S. Iwai, F. Hanaoka, EMBO J. 19 (2000) 3100–3109. [29] J.D. Page, I. Husain, A. Sancar, S.G. Chaney, Biochemistry 29 (1990) 1016–1024. [30] M.M. Jennerwein, A. Eastman, A.R. Khokhar, Chem.-Biol. Interact. 70 (1989) 39–49. [31] J.M. Woynarowski, W.G. Chapman, C. Napier, M.C.S. Herzig, P. Juniewicz, Mol. Pharmacol. 54 (1998) 770–777. [32] A. Eastman, Pharmacol. Ther. 34 (1987) 155–166. [33] P.M. Takahara, C.A. Frederick, S.J. Lippard, J. Am. Chem. Soc. 118 (1996) 12309–12321. [34] B. Spingler, D.A. Whittington, S.J. Lippard, Inorg. Chem. 40 (2001) 5596–5602. [35] M.H. Lamers, A. Perrakis, J.H. Enzlin, H.H.K. Winterwerp, N. de Wind, T.K. Sixma, Nature 407 (2000) 711–717. [36] G. Obmolova, C. Ban, P. Hsieh, W. Yang, Nature 407 (2000) 703– 710. [37] U.M. Ohndorf, M.A. Rould, Q. He, C.O. Pabo, S.J. Lippard, Nature 399 (1999) 708–712. [38] M.R. Sawaya, R. Prasad, S.H. Wilson, J. Kraut, H. Pelletier, Biochemistry 36 (1997) 11205–11215. [39] E. Bassett, A. Vaisman, J.M. Havener, C. Masutani, F. Hanaoka, S.G. Chaney, Biochemistry 42 (2003) 14197–14206. [40] R. Kusumoto, C. Masutani, S. Iwai, F. Hanaoka, Biochemistry 41 (2002) 6090–6099. [41] R.E. Johnson, M.T. Washington, S. Prakash, L. Prakash, J. Biol. Chem. 275 (2000) 7447–7450. [42] M.T. Washington, R.E. Johnson, L. Prakash, S. Prakash, Proc. Natl. Acad. Sci. USA 98 (2001) 8355–8360. [43] T. Matsuda, K. Bebenek, C. Masutani, F. Hanaoka, T.A. Kunkel, Nature 404 (2000) 1011–1013. [44] V.M. Maher, L.M. Ouellette, R.D. Curren, J.J. McCormick, Nature 261 (1976) 593–595. [45] M. Cordeiro-Stone, J.C. Boyer, B.A. Smith, W.K. Kaufmann, Carcinogenesis 7 (1986) 1783–1786. [46] V. Bergoglio, Y. Canitrot, L. Hogarth, L. Minto, S.B. Howell, C. Cazaux, J.S Hoffmann, Oncogene 20 (2001) 6181–6187. [47] Y. Canitrot, C. Cazaux, M. Frechet, K. Bouayadi, C. Lesca, B. Salles, J.-S. Hoffmann, Proc. Natl. Acad. Sci. USA 95 (1998) 12586–12590. [48] X.J. Lu, W.K. Olson, Nucleic Acids Res. 31 (2003) 5108–5121. [49] A. Barbic, D.M. Crothers, J. Biomol. Struct. Dyn. 21 (2003) 89– 97. [50] L.G. Marzilli, J.S. Saad, Z. Kuklenyik, K.A. Keating, Y. Xu, J. Am. Chem. Soc. 123 (2001) 2764–2770. [51] A. Gelasco, S.J. Lippard, Biochemistry 37 (1998) 9230–9239. [52] M.A. Elizondo-Riojas, J. Kozelka, J. Mol. Biol. 314 (2001) 1227– 1243. [53] C.M. Reyes, P.A. Kollman, J. Mol. Biol. 297 (2000) 1145–1158. [54] M. Wu, S. Yan, D.J. Patel, N.E. Geacintov, S. Broyde, Nucleic Acids Res. 30 (2002) 3422–3432. [55] A. Vaisman, M.W. Warren, S.G. Chaney, J. Biol. Chem. 276 (2001) 18999–19005. [56] H. Ling, F. Boudsocq, B.S. Plosky, R. Woodgate, W. Yang, Nature 424 (2003) 1083–1087. [57] A. Travers, Curr. Opin. Struct. Biol. 10 (2000) 102–109. [58] L. Haracska, S.L. Yu, R.E. Johnson, L. Prakash, S. Prakash, Nature Gen. 25 (2000) 458–461. [59] R.E. Johnson, L. Haracska, S. Prakash, L. Prakash, Mol. Cell. Biol. 21 (2001) 3558–3563. [60] A. Bresson, R.P. Fuchs, EMBO J. 21 (2002) 3881–3887. [61] Y. Zhang, X. Wu, D. Guo, O. Rechkoblit, N.E. Geacintov, Z. Wang, Mutat. Res. 510 (2002) 23–35.