Communication pubs.acs.org/IC
Cisplatin−Protein Interactions: Unexpected Drug Binding to N‑Terminal Amine and Lysine Side Chains Irene Russo Krauss,† Giarita Ferraro,† and Antonello Merlino*,† †
Department of Chemical Sciences, University of Naples Federico II, Via Cintia, I-80126 Naples, Italy S Supporting Information *
monoplatinated, bisplatinated, and even pluriplatinated protein adducts can be formed, and (d) cisplatin can bind to protein side chains in a monodentate or bidentate fashion.39 To obtain further information on the binding of cisplatin to proteins and, in particular, to unveil undiscovered binding sites or modes of interaction, we have investigated by X-ray crystallography the reactivity of this drug with thaumatin from Thaumatococcus daniellii (thaumatin), which is a model protein for crystallization studies.40 Thaumatin has a well-known 3D structure, which has been solved even at atomic resolution.41 We have used this protein because it has an abnormally low count of His, Met, and Cys residues (it does not contain any His and free Cys residues and possesses just one buried Met). This allows the possibility of cisplatin binding to other residues to be explored. This molecule has already been used for metalation studies.42 Titration experiments carried out by UV−vis and intrinsic fluorescence revealed that the protein binds cisplatin in solution (see Supporting Information). Crystals of the adduct between thaumatin and cisplatin were obtained by soaking experiments, where native protein crystals were incubated with an excess of the drug, with a protein-to-metal ratio of about 1:6. Thaumatin−cisplatin adduct crystals diffract X-rays at 1.45 Å resolution. The structure was solved using the PDB code 3QY5, without ligands, as the starting model43 and anisotropically refined to Rfactor and Rfree values of 0.112 and 0.153 (0.151 and 0.175 when refined isotropically), respectively (Table S1). The overall structure of the adduct (Figure 1) is very similar to that of the drug-free protein used as the starting model. The rootmean-square deviation in the positions of Cα atoms between the thaumatin−cisplatin adduct and drug-free protein is 0.14 Å. Pt binding sites were identified by analyzing the Fourier difference and anomalous electron density maps (Figures 1 and 2).
ABSTRACT: Literature studies carried out by mass spectrometry and X-ray crystallography have demonstrated that cisplatin is able to bind proteins mainly close to Met, His, and free Cys side chains. To identify possible alternative modes of cisplatin binding to proteins at the molecular level, here we have solved the high-resolution Xray structure of the adduct formed in the reaction between the drug and the model protein thaumatin, which does not contain any His and free Cys residues and possesses just one buried Met. Our data reveal unexpected cisplatin binding sites on the protein surface that could have general significance: cisplatin fragments −[Pt(NH3)2Cl]+, −[Pt(NH3)Cl2], and −[Pt(NH3)2(OH2)]2+ bind to a protein N-terminus and close to Lys and Glu side chains.
C
isplatin [cis-diamminedichloroplatinum(II)] is a highly effective anticancer therapeutic in wide clinical use for the treatment of testicular, ovarian, oral, head, neck, stomach, smallcell lung, and bladder cancers.1,2 Since the discovery of its cytotoxic activity in the 1960s,3 research has been largely focused on the characterization of DNA−cisplatin adduct formation. These studies indicated that N7 atoms of guanines are the preferential platination sites.4 Adducts can involve two adjacent nucleobases or two residues of the paired DNA strands.4,5 The formation of these stable adducts is thought to be a DNA damage event that affects proliferation of cancer cells. Nevertheless, N7 atoms of guanines are not the only biological targets of cisplatin because Pt can efficiently interact with sulfur and amine groups of peptides and proteins.6 The interaction between cisplatin and proteins has been studied in solution using many model systems including hen egg white lysozyme (HEWL),7 RNase A,8 cytochrome c (cyt c),9,10 transferrin,11 hemoglobin,12 human serum albumin (HSA),13,14 and other proteins of interest like the copper chaperone Atox-1,15 Menkens and Wilson’s disease proteins,16 ubiquitin,17,18 RNA polymerase,19 metallothionein,20 and superoxide dismutases (SODs).21,22 In most of these studies, mass spectrometry (MS) has been used to confirm the binding of cisplatin to specific protein sites.23 X-ray structures of cisplatin bound to HEWL,7,24−30 Atox-1,31 bovine and human SODs,32,33 RNase A,34,35 bovine heart cyt c,36 horse spleen ferritin,37 and HSA38 have also been solved. These structural data have recently been reviewed.39 This survey indicates that (a) cisplatin can selectively platinate His, Met, and free Cys side chains, (b) Pt(II) can bind proteins with different occupancies, generally retaining ammonia ligands, although exceptions to this general rule are possible and ammonia ligands could also be released, (c) © XXXX American Chemical Society
Figure 1. Cartoon representation of the thaumatin−cisplatin structure (PDB code 5L4R). Cisplatin binding sites are also shown along with Pt and its ligands. Anomalous e.d. map for Pt sites is shown at 2σ level. Received: June 1, 2016
A
DOI: 10.1021/acs.inorgchem.6b01234 Inorg. Chem. XXXX, XXX, XXX−XXX
Communication
Inorganic Chemistry
were modeled, although on both sites, one of them does not have well-defined electron density maps for its ligands. Hydrolysis of [Pt(NH3)2Cl]+ produces [Pt(NH3)2(OH2)]2+, which binds to the Glu168 side chain (Figure 2C). Interestingly, the structure also reveals the presence of a [Pt(NH3)Cl2] fragment bound to the Lys49 side chain (Figure 2E). In this site, the diffuse anomalous map also suggests the presence of two alternative conformations of the cisplatin fragment; here, platinum ligands interact only with solvent water molecules. Pt ions bound to Lys67, Lys139, and Lys19 (Figures 2H and S3) were also added to the model, but in these cases, because of very low occupancy (
