Metallodrug-protein interaction probed by synchrotron terahertz and neutron scattering spectroscopy

Abstract

This experimental work applied coherent synchrotron-radiation terahertz spectroscopy and inelastic neutron scattering to address two processes directly associated with the mode of action of metal-based anticancer agents that can severely undermine chemotherapeutic treatment: drug binding to human serum albumin, occurring during intravenous drug transport, and intracellular coordination to thiol-containing biomolecules (such as metallothioneins) associated with acquired drug resistance. Cisplatin and two dinuclear platinum (Pt)- and palladium (Pd)-polyamine agents developed by this research group, which have yielded promising results toward some types of human cancers, were investigated. Complementary synchrotron-radiation-terahertz and inelastic neutron scattering data revealed protein metalation, through S- and N-donor ligands from cysteine, methionine, and histidine residues. A clear impact of the Pt and Pd agents was evidenced, drug binding to albumin and metallothionein having been responsible for significant changes in the overall protein conformation, as well as for an increased flexibility and possible aggregation.

Significance

This study reports an investigation of drug interactions with human serum albumin and metallothioneins of unequivocal importance for understanding drug’s pharmacokinetics and pharmacodynamics, which are pivotal for overcoming chemotherapy handicaps such as low drug bioavailability and acquired resistance. New potential drugs—dinuclear Pt-Pd complexes—are probed. The use of leading-edge complementary techniques—synchrotron-radiation-terahertz and neutron-based spectroscopies—is an innovative approach, providing detailed molecular information on these types of metallodrugs and their interaction with vital biomolecules. The use of a high-flux terahertz coherent synchrotron radiation (at Diamond Light Source, Didcot, UK) coupled to neutron scattering spectroscopy (at ISIS Neutron Source, Oxfordshire, UK) ensured a very high quality of the acquired data and allowed the observation of all vibrational modes of the samples.

Introduction

Cancer is still a major public health problem worldwide, and thus, new and more efficient chemotherapeutic strategies are an urgent clinical need. Numerous cytostatic agents have been developed over the years, aiming at an improved antineoplastic activity coupled to decreased acquired resistance and deleterious side effects, including Pt- and Pd-based compounds introduced upon the discovery of cisplatin (cis-(NH₃)₂PtCl₂) in the late 1960s (1, 2, 3). Among these, Pt(II) and Pd(II) polynuclear chelates (i.e., comprising more than one metal center) with polyamines constitute a specific class of DNA-damaging anticancer agents, its cytotoxicity being mediated by selective covalent binding of the metal ion to DNA bases via an interplay not available to conventional mononuclear Pt drugs that may lead to an enhanced therapeutic effect (4,5). These compounds have been synthetized and investigated by the team following a multidisciplinary approach using vibrational spectroscopy, synchrotron-based x-ray absorption and biological assays. A comprehensive set of data has already been obtained on their activity toward several human cancer cells (6, 7, 8), interactions with DNA and glutathione (9, 10, 11, 12), and impact on cellular metabolism and intracellular water (13, 14, 15, 16). The dinuclear chelate Pd₂Spm (spermine [Spm]; H₂N(CH₂)₃NH(CH₂)₄NH(CH₂)₃NH₂), in particular, has yielded quite promising results toward human metastatic (triple-negative) breast cancer (6,13).

However, low bioavailability, toxicity, and acquired resistance, specifically regarding metallodrugs, are still severe handicaps in oncology practice that drastically restrict chemotherapy success and affect millions of patients (17). Therefore, the mechanisms underlying drug transport and resistance, which control the agent’s biodistribution and availability at its pharmacological target, must be thoroughly understood to overcome these drawbacks. Intracellular inactivation of the drug via coordination to biomolecules apart from the target (DNA) is one of such mechanisms, namely involving metallothioneins (MTs). These ubiquitous cysteine-rich proteins are linked to cellular protection against metal toxicity and have a high affinity for soft transition metal ions such as Pt²⁺ and Pd²⁺ (18), thus being involved in cisplatin’s resistance through the formation of Pt-thiolate clusters, leading to the drug’s intracellular capture and consequent decreased bioavailability at the therapeutic site (19,20). Additionally, MT sequestration can also trigger severe side effects (e.g., nephrotoxicity) and induce an enhanced drug efflux from the cell. Human serum albumin (HSA), in turn, is the most abundant protein in human serum (50–60% of the total plasma proteins), playing a key role in the transport and metabolism of both endogenous and exogenous compounds such as drugs. Pt-based agents, in particular cisplatin, are known to reversibly bind to HSA upon intravenous administration (65–98%, through multiple binding sites such as nitrogen and sulfur atoms from cysteine, methionine, and histidine), yielding stable adducts that circulate in the bloodstream (21, 22, 23, 24). The effects of Pt-drug binding to HSA, however, are controversial: although some authors agree that it decreases antitumor activity by lowering the drug’s bioavailability at the target (drug inactivation) (25,26), others suggest that the drug-protein conjugates may act as reservoirs for the active Pt species, mediating their transfer to DNA (27,28). Hence, detailed information is needed to thoroughly understand metallodrug-protein interaction, which to the best of the authors’ knowledge, has never been investigated at a molecular level. This will help to elucidate the drug’s pharmacokinetic profile, specifically the competition between S- or N-donor ligands from proteins (cysteine, methionine, or histidine) and the N7 atom at adenine and guanine from DNA (the preferred drug coordination sites) (27, 28, 29).

Fourier transform infrared (FTIR) spectroscopy coupled to synchrotron radiation constitutes an unmatched tool for studying biological systems, specifically for monitoring drug-biomolecule interactions (9, 10, 11, 12, 13,30,31). In addition, terahertz (THz) absorption spectroscopy, which emerged in the early 1990s, covering the 3–100 cm⁻¹ (0.1–3 THz) frequency range (32,33), is particularly sensitive to very-low-frequency vibrational modes arising from intermolecular interactions (e.g., hydrogen-bond close contacts) or from changes in global molecular conformational structure. Therefore, it may directly probe solvation of biomolecules as well as collective motions in the picosecond timescale, as compared with infrared, which monitors localized vibrational modes in the pico- to femtosecond timescale. THz spectroscopy has proved to be a very suitable method to study large-scale motions of secondary and tertiary structure elements such as amino acid residues, protein backbone and helices, or RNA and DNA loops, as well as other large-amplitude delocalized modes associated to conformational rearrangements and biomolecular function (22,34, 35, 36, 37, 38, 39, 40). The unique features of THz radiation, namely its nonionizing nature (avoiding damage of biological samples), ability to penetrate up to several millimeters inside nonmetallic samples, and unprecedented sensitivity to the water content of cells and tissues, have prompted its growing application in the biological and medical fields (41,42). THz measurements on proteins, in particular, are highly sensitive to intra- and intermolecular interactions such as protein-protein and protein-water hydrogen-bond-mediated close contacts, providing reliable information on protein’s conformational rearrangements, average flexibility, and functionality (39,43). Hence, this technique can deliver unique molecular information on the biopolymer’s behavior under specific conditions, namely upon drug-exposure-revealing drug-protein interactions and drug-elicited conformational rearrangements. The use of a high-flux THz coherent synchrotron-radiation (CSR) emission at the B22 infrared beamline of Diamond Light Source (Didcot, UK) (44,45) ensures a greatly enhanced intensity of the exciting beam (the gain being approximately three to four orders of magnitude higher than conventional synchrotron radiation regarding both flux intensity and flux delivered at the sample) which, coupled to highly sensitive broadband detectors (such as liquid helium cryogenically cooled silicon bolometer) provides a much better quality of the acquired data. In particular, the low-alpha mode available at Diamond allowed us to probe the very-low-frequency spectral range (10–65 cm⁻¹) with unmatched sensitivity (45,46). Furthermore, THz absorption spectroscopy via attenuated total reflectance (ATR) mode presents significant advantages for protein analysis because it does not require any type of sample preparation, avoiding further handling that might disturb the protein structure and limiting the number of artifacts generally associated with pelleting samples (powder form being suitable for ATR).

Inelastic neutron scattering (INS) spectroscopy, in turn, is a nonoptical technique complementary to THz spectroscopy that has proved to be extremely useful for monitoring a drug’s impact on biological samples (e.g., cells, proteins, and DNA) (9,12,47). INS enables the observation of all the fundamental vibrational modes, overtones, and combination bands and is especially sensitive to hydrogenous materials, particularly in the low-energy range (0–400 cm⁻¹), comprising the lattice modes and those that are characteristic of H-bond interactions. This good complementarity between THz absorption and neutron scattering allows to probe the same energy modes of a sample (in the very-low-frequency range) in two completely different ways–respectively via an optical method according to selection rules and via inelastic scattering according to nuclear cross section.

This study aimed at probing the interaction between the anticancer agents cisplatin, Pt₂Spm, and Pd₂Spm and the proteins HSA and MT at two levels: either conformational alterations undergone by the proteins upon drug binding or drug changes prompted by this interplay. Two crucial processes for drug activity were tackled, through the investigation of these drug-protein interactions: 1) drug transport through the blood stream for these types of intravenously administered chemotherapeutic agents—using a drug-HSA model, and 2) acquired resistance, due to drug binding to sulphur-containing intracellular proteins such as metallothioneins—using a drug-MT model. The high sensitivity and complementarity of the CSR-THz-ATR and INS spectroscopies allowed us to obtain comprehensive, unique, and very accurate information on the systems under analysis in the low-energy region of the vibrational spectrum encompassing the drug’s vibrations involving the metal ion as well as the modes ascribed to protein’s collective motions that reflect conformation and functionality. This is an innovative approach, the results thus obtained being pivotal for improving the insight of drug efficacy aiming at an enhanced bioavailability coupled to reduced acquired resistance and deleterious side effects.

Materials and methods

The list of chemicals as well as the details of the UV, Raman, FTIR-ATR, CSR-THz-ATR, and INS data acquisition, preprocessing, and analysis are described in the Supporting materials and methods.

Preparation of the drug-protein adducts

Aqueous solutions of cisplatin, Pt₂Spm and Pd₂Spm were prepared to obtain the exact drug amount per HSA or MT in the light of the available metal binding sites in each protein: HSA is known to bind five molecules of the mononuclear complex cisplatin per HSA molecule via histidine cysteine and methionine residues (22), whereas MT can coordinate up to seven divalent cations, mostly through cysteine residues (but also via histidine, glutamic, and aspartic acid moieties) (48,49).

The solid proteins were solubilized in aqueous solutions of each drug: 10 mg of HSA or 2 mg of MT were added to 7.53 × 10⁻⁷ mol of cisplatin and to 2.15 × 10⁻⁶ mol of either Pt₂Spm or Pd₂Spm—yielding drug-HSA (5:1) or drug-MT (7:1) adducts. The reaction mixtures were left overnight at 4–8°C with gentle stirring to ensure a complete drug coordination to the proteins, upon which the drug-protein conjugates were lyophilized and fully characterized by UV-visible, FTIR, and Raman spectroscopies (Fig. S1).

CSR-THz-ATR spectroscopy

The FTIR spectra using CSR THz data were acquired (at room temperature, for the lyophilized samples) in ATR mode, at the Multimode InfraRed Imaging and Microspectroscopy (MIRIAM) B22 beamline of Diamond Light Source (44, 45, 46), in a Bruker Vertex 80 V FTIR Interferometer (Billerica, MA), at room temperature, and in vacuum of ∼1 mbar (see details in the Supporting materials and methods).

INS spectroscopy

The INS measurements were carried out at the ISIS Pulsed Neutron and Muon Source of the Science and Technology Facilities Council Rutherford Appleton Laboratory (Oxfordshire, UK), using the time-of-flight, high resolution broad range spectrometer TOSCA (50, 51, 52, 53) (see details in the Supporting materials and methods). Apart from the drug-protein adducts, the isolated drugs (cisplatin, Pt₂Spm, and Pd₂Spm) and proteins (HSA and MT) were measured. All data were acquired at low temperature: ∼10 K).

Results and discussion

The UV profile obtained for the free HSA or MT proteins and the corresponding drug adducts clearly evidenced new absorption maxima specific for the drug-protein systems (at 205 and 212 nm) as well as a significant hypochromism of the characteristic protein absorption band at 278 nm (π-π^∗ transition from tyrosine, tryptophan, and phenylalanine chromophores) upon drug binding (Fig. S1 A). These spectral changes revealed drug-prompted conformational changes, as previously reported by other authors for different drug-HSA systems (54,55). In addition, the Raman signatures of both free and drug-bound proteins showed the disappearance of the vibrational modes associated to HSA’s cysteinic sulfhydryl groups (R–SH), namely ν(SH) at ∼2400 cm⁻¹ (Fig. S1 B; (49)), showing that virtually all cysteine residues in the protein are involved in coordination to the drug through its metal center (either Pt(II) or Pd(II)), with a high affinity for these sulfur soft ligands. Additionally, the ν(SS) bands from cystine (R–S-S–R), at 488 and 512 cm⁻¹ in the Raman profile of the free protein, were not detected in the spectra of the adducts where the presence of the S-bound metal ions hinders the formation of this disulfide bond. Also, metal-S stretching modes were observed at ∼230 cm⁻¹ in both cisplatin-HSA and Pd₂Spm-HSA. Furthermore, comparison of the FTIR spectra of the free proteins and their drug conjugates revealed a slight red shift of the amide I band (ν(C=O)) upon drug interaction. Moreover, conformational rearrangements of the polypeptide chains were detected in the drug adducts, as evidenced by variations in the CH stretching modes: appearance of a weak band at 2909 cm⁻¹ and changes in the intensity ratio 2959/2935 cm⁻¹ (Fig. S1 C).

Because drug-protein interactions are expected to induce some degree of conformational rearrangement in the biopolymer, the low-energy vibrational modes (mainly skeletal motions and intermolecular vibrations) associated with protein’s tertiary structure should be affected, providing unique information on the drug’s interplay with the two specific proteins under study: HSA and MT. Both CSR-THz and INS spectroscopies were applied to probe drug-prompted conformational changes as techniques of choice for monitoring the low wavenumber spectral region with high sensitivity.

CSR-THZ-ATR spectroscopy

The measurements performed along this study covered the 10–660 cm⁻¹ (0.3–20 THz) frequency range. This allowed us to probe specific vibrational features, namely H-bond signatures and collective modes from the protein, prone to be affected by drug binding, and low-frequency bands from the drugs, particularly those involving the metal center (as previously observed through the complementary vibrational techniques FTIR, Raman, and INS (47,56)) (Fig. 1). Special emphasis was placed in the low-energy region of the spectrum (below 65 cm⁻¹), which was measured with very high sensitivity using a dedicated low-alpha mode of operating Diamond and consequent CSR available at the B22 beamline (Diamond Light Source).

Figure 1.

Vibrational spectra—THz-ATR, Raman, and INS (below 600 cm⁻¹)—of the anticancer agents cisplatin, Pt₂Spm, and Pd₂Spm. To see this figure in color, go online.

The THz profiles obtained for the HSA and MT proteins were broad and featureless, lacking an identifiable structure and sometimes displaying a periodic profile, chiefly because of the high density of low-frequency states and to the damping from the strong coupling between different modes in these types of large and heterogeneous samples (Fig. 2). The oscillations observed in the spectra are probably Etalon fringes related to the penetration depth of the THz radiation within the thickness of the sample, even in the case of the evanescent wave of the ATR setup used for these experiments. The temporal coherence of the synchrotron radiation enhances the observation of such phenomena in the FTIR spectra. This is therefore a common feature of the THz signature of proteins. Also, in addition to the collective vibrational modes from secondary and tertiary structure elements, there is also a significant contribution from rotational motions of the protein side chains. Despite the lack of a narrow band profile, the THz response was still unique for each protein reflecting conformational preferences and degree of flexibility. This experimental signal trend was observed for all the replicates measured (at least two per sample), which validates the data. Although for MT the THz profile was found to rise almost linearly with frequency, for HSA, there was an increase up to ∼45 cm⁻¹, which turned into a plateau for the higher energies (45–65 cm⁻¹; Fig. 2 A), as previously reported for bovine serum albumin (57).

Figure 2.

CSR-THz-ATR spectra (acquired in low-alpha mode) of lyophilized proteins human serum albumin (HSA) and metallothionein (MT) (A), lyophilized versus powder forms of HSA (B), and MT (C). To see this figure in color, go online.

Because the THz absorbance from water may constitute a serious problem in the analysis of THz data from aqueous proteins because it is much stronger than the biomolecule’s signature that can therefore be masked (mainly below 300 cm⁻¹), lyophilized samples were currently used. However, once lyophilization might affect protein’s secondary structure (e.g., increase in β-sheet content and decrease of α-helix and unordered arrangements) (58,59), spectra were measured and compared for both the powder (commercial product, as received) and the lyophilized samples. It was found that the corresponding data nearly overlay in the very-low-frequency interval (0–65 cm⁻¹) (Fig. 2, B and C), mainly for MT, which is evidence of negligible structural changes due to the lyophilization process (in accordance with previous observations by other authors in mammalian MTs monitored by Raman spectroscopy (49)). Hence, the differences in the THz profile presently observed between each protein and its drug adducts can reliably be ascribed to drug-elicited variations in collective protein motions and/or protein’s intra- and intermolecular H-bond close contacts, as well as to possible changes in the vibrational profile of the drugs upon interaction with the polypeptide chains.

The global (large-scale) collective vibrational modes of the biopolymer, which are prone to be affected by the drug-protein interplay (e.g., upon protein’s aggregation or polymerization) (21,28), are strongly dependent on the protein’s overall conformation. Therefore, any changes in these conformational preferences will be clearly reflected in the corresponding THz profile (low-energy vibrational density of states), namely regarding its slope as a function of energy, as shown in Fig. 3, A and B, which depict the THz-ATR spectra of drug-HSA and drug-MT adducts. Although for the HSA systems, the absorbance raised with frequency up to ∼40 cm⁻¹, then becoming nearly constant, mainly for the Pt-containing systems (Fig. 3 A), for the MT conjugates, there was an almost linear increase from 20 to 65 cm⁻¹ (Fig. 3 B). This slope increment of the THz profile, which was found to be more pronounced for the drug-protein conjugates as compared with the free proteins (particularly for the Pd₂Spm adducts), evidences a conformational rearrangement of the protein upon drug binding consistent with an enhanced flexibility (36). In addition, it reveals a strong interaction between the drug and the MT, which is known to be associated to acquired resistance mediated by the formation of metal-thiolate species that are responsible for a decreased drug bioavailability at its target. This plasticity enhancement within a protein due to binding to an external ligand (protein softening effect) has been previously reported, namely regarding the interaction of the anticancer drug methotrexate with its target enzyme dihydrofolate reductase (60).

Figure 3.

CSR-THz-ATR spectra (acquired in low-alpha mode) of HSA and its drug adducts (A); MT and its drug adducts (B); and cisplatin, Pt₂Spm, and Pd₂Spm (C). To see this figure in color, go online.

Each of the currently tested metallodrugs induced distinct conformational changes in either HSA or MT (Figs. 3 and S2). The effect of the dinuclear agents Pt₂Spm and Pd₂Spm followed a similar trend for each of the biomolecules, with either a lower or higher intensity relative to free HSA and MT, respectively. In turn, the mononuclear agent cisplatin was found to have opposite impacts on HSA as compared with MT and always inverse to the effects of the Pt- and Pd-Spm compounds (Fig. 3, A and B). Also, Pd₂Spm showed a more significant influence on MT relative to HSA, and an overall stronger impact on both proteins as compared with its Pt counterpart.

The isolated drugs under analysis displayed characteristic THz signatures in the low-frequency range, with a noticeable discrimination between the mononuclear cisplatin and the dinuclear Pt₂Spm and Pd₂Spm (Fig. 3 C). These THz profiles are consistent with the drug’s FTIR spectra depicted in Fig. 1 (e.g., cisplatin displays an intense and broad infrared band below 100 cm⁻¹).

INS spectroscopy

After a previous vibrational spectroscopy study by the authors on drug-DNA (drug = cisplatin, Pt₂Spm, and Pd₂Spm), which constituted the first INS analysis of a drug’s interplay with a biopolymer (12), INS measurements were presently performed for the cisplatin-HSA and Pd₂Spm-HSA adducts (the Pt₂Spm-HSA conjugates and those with MT being unavailable in high enough amounts for INS). Thanks to the recent upgrade of the TOSCA spectrometer (53), high-sensitivity data were obtained for these systems, complementing the THz global dynamics information gathered for the same samples and providing clear evidence of drug-induced conformational changes. Although the INS spectra were obtained at low temperature to reduce the impact of the Debye-Waller factor on the observed spectral intensity (Eq. S1), these data are still comparable with the one provided by THz spectroscopy (at room temperature) because the decrease in temperature does not affect the samples’ vibrational modes (instead, it allows their accurate observation).

The INS spectra measured for HSA’s drug conjugates may be divided into 1) the very-low-frequency interval (<120 cm⁻¹), encompassing lattice vibrations from drugs and proteins; 2) the translational range (160–370 cm⁻¹), comprising vibrational modes from the drugs (mainly those involving the metal center(s)) and torsional vibrations from the proteins (e.g., τ(CH₃) at ∼240 cm⁻¹); 3) the interval spanning from ∼370 to 1750 cm⁻¹, containing typical protein bands as well as signals from the drugs (mostly from the ligands); and 4) the high wavenumber range (2500–3750 cm⁻¹), covering the typical ν(SH) mode from HSA (at 2380 cm⁻¹) and the CH and NH stretching vibrations from both the drugs and the proteins (at 2950–3000 and 3200–3300 cm⁻¹, respectively). The INS profiles of the cisplatin and Pd₂Spm metallodrugs are represented in Fig. 4 A, whereas Fig. 4, B–D depicts the spectra of the drug-protein adducts, as compared with those from free cisplatin, Pd₂Spm, and HSA (Fig. 4, C and D comprising expanded spectral windows of Fig. 4 B).

Figure 4.

INS spectra (at 10 K) of cisplatin and Pd2Spm (0–2000 cm⁻¹) (A); HSA, cisplatin-HSA and Pd2Spm-HSA (0–2000 cm⁻¹) (B); and HSA, Pd2Spm, and Pd2Spm-HSA (0–400 cm⁻¹) (C). (D) HSA, cisplatin, cisplatin-HSA, and difference spectrum between cisplatin-HSA and HSA (0–2500 cm⁻¹). (The main drug-elicited changes are highlighted by dashed lines.) To see this figure in color, go online.

The main signals from the drugs are clearly observed in Fig. 4, A and C – v(metal-nitrogen) (at ∼500 cm⁻¹), ρ(NH₂-NH₃) (at 750–800 cm⁻¹), and NH₂ and CH₂ deformations of the Spm complexes (above 850 cm⁻¹). Fig. 4 B shows characteristic protein features–from tyrosine and tryptophan aromatic residues (at ∼650 and 750 cm⁻¹, respectively), CH₂ deformations from the polypeptide chain (900–1550 cm⁻¹), and modes associated to the peptide bond (amide III, amide II, and amide I at 1250, 1540–1550, and 1650 cm⁻¹, respectively).

Clear changes were detected in the INS profile of the drug-HSA conjugates relative to both the free protein and the isolated Pd agent, revealing the perturbation on protein’s conformation due to drug binding, which takes place via metal coordination to the N and S sites from HSA’s histidine and methionine residues (4). A drug impact on HSA’s skeletal torsions (amide VII mode) and NH ^… O interactions within the peptide backbone, responsible for signals centered at ∼160 cm⁻¹, was evidenced in the cisplatin adducts for which these features virtually disappeared (Fig. 4 B)—this may be due to drug-elicited protein aggregation, as previously found for HSA in the presence of cisplatin (21), apart from conformational changes upon metal coordination. Additionally, the band from tryptophan (∼750 cm⁻¹) slightly lowered its intensity upon drug interaction, this variation being more significant has the Pd₂Spm-HSA system. The drug-triggered conformational rearrangements in the protein were also reflected in the changes observed for the methylene and amine deformation modes of the polypeptide chain–the bands at ca. 1300/1450 and 1650 cm^-1, assigned to δ(CH₂) and δ(NH₂), respectively, displayed an increased intensity in the adducts, as shown by the difference spectrum between the drug-HSA conjugate and free HSA (Fig. 4 D, see also difference spectrum between cisplatin-HSA and cisplatin; Fig. S3). Additionally, a noticeable variation detected in the protein’s vibrational spectra is ascribed to the SH stretching band (at 2380 cm⁻¹) that was found to disappear in the drug-protein adducts (Fig. 4 D (as also observed by Raman, Fig. S1 B), clearly revealing drug binding to the protein’s sulfur atoms. Moreover, the drug’s impact on protein’s ν(CH) modes was evidenced by a variation in the intensity ratio 3050:2945 cm⁻¹ in the drug adducts, mainly for cisplatin-HSA (Fig. S1 D), also observed by FTIR (Fig. S1 C). This may reveal a distinct drug effect on the protein’s saturated versus unsaturated carbon chains, which give rise to each of these signals (ν(CH) ∼2945 and 3050 cm⁻¹, respectively) because of drug-prompted oxidation reactions and/or selective coordination of the Pt and Pd agents to the polypeptide side chains through different patterns of protein metalation.

Particularly regarding the drugs’ vibrational profile, the distinctive γ(N–M–Cl), δ(Cl–M–Cl), δ(N–M–Cl), and ν(M–Cl) modes (M = Pt(II) or Pd(II)), which yield well-defined bands for the free drugs (namely ν(M–Cl) at 330 and 356 cm⁻¹, respectively, for cisplatin and Pd₂Spm (47,56) Fig. 4 A), were not detected in the spectra of the adducts (Fig. 4, B and C, as foreseen upon chloride hydrolysis undergone by the drug molecules before protein metalation (12). In addition, the NH₃-NH₂ rocking vibrations from cisplatin and Pd₂Spm (at 750–850 and 720–815 cm⁻¹, respectively) were found to be affected by drug coordination to albumin (Fig. 4, A and B).

Conclusions

Impaired transport and the development of acquired resistance are known to severely restrain the clinical use of metal-based anticancer agents. In this work, the role of proteins as potential drug transporters, intermediates, or reservoirs was investigated. The metalation of HSA and MT by cisplatin and two newly developed Pt- and Pd-dinuclear compounds was studied, with a view to attain a better understanding of the drug-protein interplay and its effect on the drug’s pharmacokinetic behavior, which determines its biodistribution and bioavailability at the target. The drug-HSA and drug-MT models were used to address drug transport, acquired resistance, and deleterious side effects. Two complementary cutting-edge approaches were applied as unique molecular fingerprint techniques for probing the low-energy region of the vibrational spectrum: synchrotron THz and INS spectroscopies, the former having been established as an invaluable tool to monitor biomolecular binding and conformational rearrangements, whereas the latter is a method of choice for analyzing structural changes with high sensitivity, particularly in hydrogenated samples.

The THz data currently gathered for the proteins and drug-protein adducts evidence different slopes, capable of discriminating among the systems under study. Similar to former observations on the effect of cisplatin-like metallodrugs on DNA (12), these experiments revealed a clear impact of the Pt and Pd agents on protein’s structure, conformational behavior, and overall flexibility, both for HSA and MT. These drug-elicited conformational changes and the consequent plasticity enhancement in drug-HSA and drug-MT conjugates are in accordance with THz data formerly reported for bovine serum albumin (highly homologous to HSA), which showed a similar spectral trend upon protein hydration (larger hydration values corresponding to a higher flexibility of the biopolymer) (57). In addition, these results agree with the impact of the same metal-based drugs on DNA, previously determined by the authors through INS and FTIR (12)–significant conformational changes in the biopolymer upon drug binding and disappearance of the vibrational modes ascribed to metal-chloride bonds in the Pt and Pd agents. Furthermore, drug effects on proteins (directly related to the degree of metalation) often lead to aggregation via cross-linking interactions (28,61), as previously reported for lysozyme and HSA (21,62). This type of protein aggregation has been identified as one of the pre-DNA binding mechanisms responsible for acquired drug resistance (63).

Coupled to previous data on the impact of Pt- and Pd-Spm compounds on human cancer cells and DNA (9, 10, 11, 12,15), these results are expected to contribute to a better understanding of the drug’s mode of action, particularly of the inactivation processes associated to drug-protein linkage before DNA binding, which underlie resistance and toxicity. This will assist in the design of improved platinum- and palladium-derived anticancer agents with higher bioavailability at the target, lower acquired resistance, and decreased adverse side effects, thus enhancing chemotherapy effectiveness.

Author contributions

L.A.E.B.d.C. performed experimental measurements (in THzs) and data analysis. A.P.M. performed sample preparation and experimental measurements (in THzs). A.L.M.B.d.C. performed sample preparation and experimental measurements (Raman and FTIR, in THzs). J.M. performed sample preparation and experimental measurements (UV). G.C. performed THz experimental measurements and data analysis. S.R. performed INS experimental measurements and data analysis. M.P.M.M. performed conceptualization, experimental measurements (in THzs), data analysis, and manuscript writing. All authors have read and agreed to the submitted version of the manuscript.

Editor: Jill Trewhella.

Supporting citations

References (64, 65, 66, 67) appear in the Supporting material.