Quantum Dot Photoactivation of Pt(IV) Anticancer Agents: Evidence of

Article pubs.acs.org/JPCC

Quantum Dot Photoactivation of Pt(IV) Anticancer Agents: Evidence of an Electron Transfer Mechanism Driven by Electronic Coupling Ivan Infante,*,†,# Jon M. Azpiroz,†,# Nina Gomez Blanco,‡ Emmanuel Ruggiero,‡ Jesus M. Ugalde,† Juan C. Mareque-Rivas,‡,§ and Luca Salassa*,‡ †

Kimika Fakultatea, Euskal Herriko Unibertsitatea and Donostia International Physics Center (DIPC), P.K. 1072 Donostia, Euskadi, Spain ‡ CIC biomaGUNE, Paseo Miramón 182, 20009 Donostia, Euskadi, Spain § Ikerbasque, Basque Foundation for Science, 48011 Bilbao, Euskadi, Spain S Supporting Information *

ABSTRACT: Herein we elucidate the mechanism of photoreduction of the Pt(IV) complex cis,cis,trans-[Pt(NH3)2(Cl)2(O2CCH2CH2CO2H)2] (1) into Pt(II) species (among which is cisplatin) by quantum dots (QDs), a process which holds potential for photodynamic therapy. Density functional theory (DFT) and time-dependent density functional theory (TDDFT) methodologies, integrated by selected experiments, were employed to study the interaction and the light-induced electron transfer (ET) process occurring between two QD models and 1. Direct adsorption of the complex on the nanomaterial surface results in large electronic coupling between the LUMO (lowest unoccupied molecular orbital) of the excited QD* and the LUMO+1 of 1, providing the driving force to the light-induced release of the succinate ligands from the Pt derivative. As confirmed by photolysis experiments performed a posteriori, DFT highlights that QD photoactivation of 1 can favor the formation of preferred Pt(II) photoproducts, paving the way for the design of novel hybrid Pt(IV)−semiconductor systems where photochemical processes can be finely tuned.

■

INTRODUCTION Semiconductor quantum dots1 (QDs) have become in recent years one of the most fascinating and promising type of materials with technological applications ranging from photovoltaics2−6 and optoelectronic devices7 to biosensors8 and bioimaging agents.9 The attractiveness of QDs lies in their high optical extinction coefficients, sharp emission spectrum, carrier multiplication ability, and high photo- and thermal stability,10,11 features which can be tuned by varying QD size, shape, and composition. A key ubiquitous process in QD chemistry is electron transfer (ET). For example, efficient charge separation at the interface between a QD (or a film of QDs) and an oxide semiconductor material, e.g., TiO2, is mandatory to obtain highly performing photovoltaic cells,6 while field-effect transistor devices engineered from thin film of QDs rely on efficient charge hopping between adjacent QD units.12 QDs are also being explored as light-induced ET activators with very promising results in other research areas, as, for example, photoactivatable protein inhibitors, photodynamic therapy (PDT),13,14 and photocatalysis.15,16 Thorough understanding of the ET mechanism in these systems, in particular the role played by parameters such as reorganization energy, Gibbs free energies, and electronic coupling, is therefore paramount for researchers involved in the advance of QD-based applications. © 2014 American Chemical Society

Density functional theory (DFT) is a powerful tool to achieve such goals as recently demonstrated in the ET mechanism elucidation of QD−metal oxide17 and QD− fullerene systems.18 Indeed DFT is able to provide a good description of QDs electronic structure and model their interaction with molecules and material surfaces, ultimately providing useful insights into the thermodynamic and kinetic factors ruling ET. In this work we present a combination of DFT calculations and selected experiments in the context of rationalizing the ET process from a core−shell CdSe@ZnS QD into a Pt(IV) anticancer agent, namely, cis,cis,trans[PtIV(NH3)2(Cl)2(O2CCH2CH2CO2H)2] (1). Pt(IV) complexes have been extensively studied as prodrugs whose activity can be switched on in vitro and in vivo by biological reductants19−24 or by light excitation.25−28 Nanoparticlemediated photoactivation of anticancer complexes is now becoming a hot topic as demonstrated by several groundbreaking results reported lately.29−32 Mareque and co-workers have recently employed core-only and core−shell QDs to control Pt(IV) → Pt(II) reduction and Received: February 10, 2014 Revised: March 26, 2014 Published: March 27, 2014 8712

dx.doi.org/10.1021/jp501447q | J. Phys. Chem. C 2014, 118, 8712−8721

The Journal of Physical Chemistry C

Article

Chart 1. DFT Optimized Structures of the Models Investigated in This Study (PBE0/def2-SV(P) Level)

Table 1. Selected Bond Distances (Å) for the X-ray and DFT-Optimized Geometries of 1, 1a, and 1b, Together with Pt Atomic Charges Calculated with Different Methods PBE0/def2-SV(P) complex 1 X-ray GS S1a ll-T ll-T2 complex 1−

a

Pt−O4

Pt−O9

Pt−N14

Pt−N18

Pt−Cl2

Pt−Cl3

1.993(4) 2.014 2.278 2.301 1.998

2.008(4) 2.016 2.291 2.320 1.996

2.050(5) 2.057 2.081 2.078 2.384

2.065(5) 2.049 2.043 2.071 2.065

2.3108(15) 2.344 2.291 2.308 2.353

2.3194(15) 2.333 2.322 2.319 2.488

1a 1b charges

2.380 2.048 1-GS

2.386 2.045 1-S1

2.059 2.372 1-ll-T

2.045 2.046 1-ll-T2

2.369 2.378 1−(1a)

2.363 2.669 1−(1b)

Mulliken Hirshfeld NPA

0.444 0.467 0.841

0.414 0.481 0.792

0.387 0.439 0.729

0.530 0.981 0.888

0.228 0.642 0.642

0.307 0.726 0.726

S1 geometry was obtained using the CAM-B3LYP functional. All attempts to optimize the structure of S1 with PBE0 functional failed.

deliver cisplatin,33,34 synthesizing a QD−Pt composite with potential value as a theranostic agent for cancer therapy and multimodal imaging.34 In such a system, a core−shell CdSe@ ZnS QD (ca. 5 nm diameter), complex 1, and the γ-emitter radioactive fac-[99mTc(H2O)3(CO)3]+ complex are transported by a high-stability micelle to the target site. The 99mTc compound gives SPECT (single-photon emission computed tomography) imaging capability while the Pt(IV) and the QDs are exploited for therapy, with QDs being suitable for optical imaging as well. In such composite, QDs can efficiently photoactivate the Pt(IV) complex via ET using visible light up to 630 nm, a wavelength currently in use for clinical PDT. These hybrid systems have the potential to overcome the poor absorption properties of metal complexes in the visible, particularly in the therapeutic window (630−700 nm). In this study, we initially investigate the behavior of 1 under UV light irradiation and elucidate how dicarboxylato Pt(IV) prodrugs are transformed under chemically reducing conditions. Successively, modeling QD-1 adducts we show how the

energy alignment between the QD and 1 favors the injection of the photoexcited electron from the former to the latter, with the concomitant reduction of the Pt(IV) complex into Pt(II). In particular, DFT indicates that QD injects electrons not only to the LUMO (lowest unoccupied molecular orbital) level of 1, but also and more efficiently to the LUMO+1, favoring the release of selected ligands and the formation of the preferred photoproducts (among which is cisplatin). Photolysis experiments performed a posteriori support such outcome, highlighting that QD surfaces can be tailored for photoactivation of specific target states of metal complexes, potentially opening new opportunities to control a wide range of photochemical processes for different applications. Structures of all the models investigated in our study are reported in Chart 1.

■

COMPUTATIONAL SECTION All calculations were performed with the Gaussian 09 program.35 Ground- and excited-state properties of all the systems employed in this work were analyzed with DFT and 8713



Article

Figure 1. (a) Comparison between experimental (black) and calculated (blue) absorption spectra for 1 in water (PCM) calculated at the PBE0/ def2-SV(P) level. Singlet−singlet transitions are shown as vertical bars with heights equal to their oscillator strengths. The theoretical curves were obtained using GAUSSSUM 2.2 (fwhm = 3000 cm−1). Inset: PL spectrum of QDs used in this work. (b) Selected EDDMs of singlet−singlet electronic transitions and molecular orbitals for 1 in water (PCM) calculated at the PBE0/def2-SV(P) ground-state geometry. In the EDDMs, green indicates a decrease in electron density, while orange indicates an increase. (c) Energy diagram for complexes 1, 1a, 1b, 2, and 3 calculated at the PBE0/def2-SV(P) level in water (PCM). For complexes 1a and 1b the spin density surface is reported in the figure (isovalue 0.0005).

Figure 2. (a) Selected regions of the 1H NMR spectrum of 1 in H2O/D2O (5:1, [1] = 1 mM) in the dark (bottom) and after irradiation with 385 nm light (top, 2 min, ∼40 mW·cm−2); Pt 4f XPS spectra of 1 in the dark (bottom) and after irradiation (top) with 385 nm light (15 min, ∼40 mW· cm−2).

approximate the micelle environment. The nature of all stationary points was confirmed by normal-mode analysis. For the QD-1 complexes, the basis set superposition error (BSSE) was considered by means of the counterpoise method.39,40 TDDFT calculations on a reduced space (vide infra) were carried out with Q-Chem 4.0.41 For analysis and visualization of computational results, the software packages GAUSSSUM 2.2,42 AOMix,43,44 and Chimera45 were employed. Details on materials and experimental methods are reported in the Supporting Information.

time-dependent DFT (TDDFT), using the PBE0/def2-SV(P) combination. Such functional and basis set was chosen after careful benchmarking (Supporting Information) against available experimental data of 136 and because it has been recently demonstrated to perform extremely well for the QDs studied here.37 Calculations on 1 and its derivatives were performed using the polarized continuum model (PCM) with water as implicit solvent38 (as this is the environment in which their photophysical and photochemical properties were studied), while calculations including the QD alone and the QD with complex 1 were performed using as implicit solvent hexane, to 8714



■

Article

RESULTS AND DISCUSSION Complex 1: Photochemistry and Redox Properties. Ground- and excited-state structure and electronic properties of cis,cis,trans-[PtIV(NH3)2(Cl)2(O2CCH2CH2CO2H)2] (1) and related species (vide infra) were investigated by DFT and TDDFT to characterize the photochemical and redox behavior of the complex (Table 1 and Figure 1a−c). The DFT groundstate (GS) geometry of 1 displays bond distances in good agreement with the X-ray values. Similarly, the TDDFT simulated spectrum (Figure 1a) correctly reproduces the experimental absorption in the 275−450 nm range. Electron density difference maps (EDDMs), which graphically describe changes in electron density (Figure 1b), indicate the singlet− singlet low-energy transitions in the UV region are mixed d−d/ LMCT (ligand-to-metal charge transfer) and have dissociative nature due to significant contributions from the σ-antibonding orbitals LUMO and LUMO+1 (Supporting Information, Tables S2 and S3).46 Optimization of excited-state structures (singlet with TDDFT method and triplet with unrestricted Kohn−Sham method) was performed to obtain snapshots of transient species and characterize the excited-state dynamics of 1 upon UV light excitation. In the S1 excited state, 1 shows significantly elongated Pt−O distances (ca. 0.25 Å) compared to the GS, while the other Pt−L bond lengths are only slightly changed (Table 1). It is reasonable that population of S1, which has major contributions from the LUMO+1 orbital, leads to succinate release, whereas S2 and S3, mainly involving the LUMO orbital, should promote the formation of free Cl− and NH3. Triplet-state optimizations gave similar results affording two lowest-lying triplet geometries, ll-T and ll-T2 (ΔE = 0.29 eV), suggesting dissociation of succinate as well as Cl− and NH3 ligands could take place. In particular, ll-T geometry resembles the S1 geometry with elongated Pt−O bond lengths (ca. 0.3 Å) with respect to the GS, whereas ll-T2 shows elongated Pt−Cl (2.384 Å) and Pt−N (2.488 Å) bond lengths. In principle, photosubstitution via these excited-state species can involve a single monodentate ligand as well as the simultaneous release of two ligands. Furthermore, subsequent isomerization reactions should not be ruled out, indicating that the formation of several photoproducts is likely to occur under light irradiation. Photolysis experiments on 1 (Figure 2a) clearly show appearance of the free succinate signal in the NMR spectrum within a few minutes of UV light irradiation (385 nm, ∼40 mW·cm−2). Moreover, the NH3 multiplet of 1 (1J14N‑1H = 54 Hz, 2J195Pt‑1H = 27 Hz) at 6.3 ppm decreases in intensity while peaks corresponding to free NH3 (1J14N‑1H = 52 Hz) and a number of Pt(IV) photoproducts appear nearby, consistently with the possibility of multiple photochemical pathways. However, it is significant that XPS (X-ray photoelectron spectroscopy) measurements performed on UV-irradiated solution of 1 (Figure 2b) show little or no formation of Pt(II) species compared to the dark control (a fraction of Pt(II) is present due to direct reduction of the metal under the X-ray beam as visible in the control spectrum). Consistently with the prevalent photogeneration of inert Pt(IV), binding experiments with GMP (guanosine 5′-monophosphate) exclude the formation of Pt(II)−GMP adduct and the presence of cisplatin-like complexes (Supporting Information, Figure S13). Pt charge calculations performed with different methods (Table 1) confirm these experimental results.47−49 One of the

most likely photoproducts formed upon release of the succinate could be the complex cis,cis,trans-[PtIV(NH3)2(Cl)2(OH)2] in agreement with the NMR data and the in vitro activity observed for solutions of 1 in the dark and after UV irradiation (Supporting Information, Figure S14 and S15). For the sake of comparison with the excited-state computational characterization and to gain further insights, the first step in the reduction of 1 was investigated by performing geometry optimization calculations on the complex cis,cis,trans[PtIII(NH3)2(Cl)2(O2CCH2CH2CO2H)2]−, which results from the addition of one electron (1e) to 1. Chemical reduction of Pt(IV) complexes is a fundamental process in the biological mechanism of action of this family of prodrugs.19,20,50 Conversion of octahedral Pt(IV) to square planar Pt(II) can be promoted by endogenous reducing agents and occur through simultaneous ligand dissociation pathways.19,50,51 Two geometries were obtained corresponding to electronic structures where the extra electron is localized either on the LUMO+1 (1a) or on the LUMO (1b) (ΔE = 0.15 eV). Consistently with the shape of the orbitals hosting the unpaired electron (see spin density surface in Figure 1c), 1a is characterized by elongated Pt−O bonds (Table 1), similarly to S1 and ll-T. Instead 1b shows one longer Pt−Cl and Pt−N distance compared to 1 (Table 1), in agreement with the ll-T2 geometry of 1. Plausibly, release of succinate occurs from 1a, while Cl− and NH3 can dissociate from 1b. Adopting 1a, 1b, and the GS structures as a starting point, the 2e-reduced structure of 1 was also optimized employing the aforementioned approach. As shown in Figure 1c, two different Pt(II) reduction products were obtained: cis-[PtII(NH3)2(Cl)2] (2, cisplatin) + 2 O2CCH2CH2CO2H from 1a, and trans[PtII(NH3)(Cl)(O2CCH2CH2CO2H)2] (3) + Cl− and NH3 from 1b and the GS. Such result suggests the presence of distinct reductive pathways and formation of different Pt(II) species19,50,51 and is in agreement with cell toxicity tests.34 QD-1: The Model and the Interaction. As shown earlier for complex 1 and other related compounds,33,34 an efficient approach to promote Pt(IV) to Pt(II) reduction using (more penetrating and less harmful) visible light is by excitation of a tailored-synthesized QD in the proximity of the Pt complex. In such a way, the photoactivated QD* can transfer the excited electron from its conduction band to the Pt(IV) complex, (e.g., 1) which rapidly undergoes ligand dissociation and formation of Pt(II) toxic species at the target cancer cells. Full modeling of the interaction between 1 and the QD is not straightforward due to the complexity and heterogeneity of the system (even more when the composite is designed to reach the target site inside water-soluble poly(ethylene glycol) micelles). Nevertheless, QD-1 models where the complex is directly anchored on the surface of the nanomaterials are a valid approximation, able to provide a satisfactory description of the electronic phenomena taking place on the QD surface. This is supported by recent results showing that Ru metal complexes bearing COOH-functionalized ligands can directly be attached onto the QD surface despite the presence of bulky organic ligands.16,52,53 Indeed, photoluminescence (PL) quenching experiments were carried out (Figure 3 and Supporting Information, Figure S16) and indicate that 1 is adsorbed on the QD surface (K ∼ 5 × 104 M−1) as demonstrated by better fitting to the Langmuir isotherm (surface adsorption) compared to the Stern−Volmer expression (collisional dynamic quenching). 8715



Article

32.1 kcal/mol, respectively. The larger stabilization in the core−shell complex is most likely associated with the larger surface area of adsorption with respect to the much smaller (CdSe)12 complex, where the strain forces to accommodate complex 1 are augmented. Bond decomposition analysis reveals that the nature of QD-1 interaction is about 60% electrostatic and 40% covalent for both the bare and core−shell QDs. In particular, complex 1 transfers about 0.09 electrons to the bare (CdSe)12 and 0.29 electrons to (CdSe)12@(ZnS)48. Such charge transfer has important effects on the electronic structure of (CdSe)12-1 and (CdSe)12@(ZnS)48-1. In Figure 4a and b, the density of states (DOS) and molecular orbital (MO) energy levels of both adducts are plotted.

Figure 3. Core−shell CdSe@ZnS PL quenching plots as a function of the concentration of 1. (a) Data fitting with the Langmuir absorption isotherm (R-square = 0.99) and (b) with the Stern−Volmer model for dynamic collisional quenching (R-square = 0.80). Fitting details are reported in the Supporting Information. Error bars were determined by comparing three PL spectra for each point and propagating the error in an additive fashion.

On the bases of such experimental findings, we explored QD1 models where the metal complex is directly anchored to the surface of the nanomaterial. Furthermore, being the composites developed by Mareque and co-workers made of a core−shell CdSe@ZnS QDs of about 5 nm (ca. 2000−3000 atoms),34 a size untreatable by current theoretical approaches like DFT, QD dimensions were scaled down as much as possible but still kept them large enough to reflect the instantaneous interactions between the QD and 1. At first, we chose a small bare CdSe composed of 24 atoms, i.e., a (CdSe)12 (Chart 1), and then we added a ZnS layer to provide a model for the core−shell system. The final QD is composed by a (CdSe)12 core covered by a shell of (ZnS)48, for a size of approximately 1.5 nm (Chart 1). Structural optimization for the (CdSe) 1 2 -1 and (CdSe)12@(ZnS)48-1 adducts shows that in both cases complex 1 is attached to the QD with three anchoring groups, the two carboxylates and one of the chlorine atoms. The chlorine and the oxygen of the carboxy groups are directly adsorbed on the Cd atoms (or Zn atoms in the core−shell adduct), while the O−H group of the succinates form hydrogen bonds with the Se atoms (or S atoms in the core−shell). The instantaneous interaction energy between complex 1 and the QD is stabilizing by about 32.5 kcal/mol in (CdSe)12-1 and 78.4 kcal/mol in (CdSe)12@(ZnS)48-1. When the counterpoise correction is applied, the interaction energies are 22.6 and 60.0 kcal/mol, respectively. If preparation energies are also included (i.e., the energy required to deform the fragments from their equilibrium structure to the geometry they adopt in the interacting complex), the interaction energies reduce further to 12.3 and

Figure 4. MOs energy levels and DOS for the (CdSe)12-1 (a) and (CdSe)12@(ZnS)48-1 (b) compounds. The MOs for the QDs are always depicted in black, while the MOs of 1 are in red. The MOs of the final complex are shown in the center and are split into those mostly localized on the QD (in black) and those mostly localized on 1 (in red). Solid (dashed) arrows refer to HOMO−LUMO (TDDFT) gaps. For the core@shell model, the shaded region corresponds to the projection of the DOS on the core atoms.

As it is immediately clear, the charge donation from complex 1 to any of the two QDs raises the MOs energy levels of the QDs and lowers those of complex 1. The latter indeed becomes slightly more cationic, with the consequence of increasing its ionization energy (IP), roughly approximated to the negative value of the HOMO (highest occupied molecular orbital) energy.54,55 The HOMO (or IP) of 1 passes from −7.62 eV in the isolated conformation to −7.77 eV in (CdSe)12-1 and to −8.28 eV in (CdSe)12@(ZnS)48-1. The larger IP of the latter is a consequence of the larger charge transfer to the core−shell complex. On the other hand, the QD becomes more negative 8716



Article

organic ligands is present inside the micelle, we can assume that this shell behaves as a solvent with a very low dielectric, and therefore λEXT can be expected to be negligible. The ET activation energy is thus expressed as

and more prone to ionization (HOMO level raised). The HOMO of the isolated (CdSe)12 is indeed shifted from −6.46 to −6.39 eV, whereas that of (CdSe)12@(ZnS)48 moves from −6.25 to −6.02 eV. One might think that these effects on the MO energy levels are dampened in a much larger QD than the one employed in the calculation; however, in a realistic system is also more likely that each QD absorbs more than one molecule of complex 1, maintaining in this way the trend shown here. Moreover, it is well-known that ligands affect the QD absorption spectrum by shifting their excitation energies.56,57 For this reason, the first excitation energy of (CdSe)12-1, associated with the HOMO−LUMO transition within the bare (CdSe)12, is blue-shifted by about 0.1 eV, from 3.27 to 3.38 eV upon bond formation at the TDDFT level of theory. Note that the TDDFT HOMO−LUMO gap localized in the QD has been computed within a reduced single-excitation subspace to avoid the interference of low-lying charge transfer states induced by the Pt complex. Remarkably, the HOMO− LUMO gap estimation taken from the MO energies (Kohn− Sham orbitals) provides the same trend with a similar blue shift from 4.00 to 4.09 eV. From these numbers, it is, however, clear that the absolute value of the HOMO−LUMO gap within the QD is not a good approximation to the actual excitation energy, as it does not take into account relaxation effects induced by the formation of the electron−hole pair. This effect is included in the TDDFT calculation which gives a value for the first excitation energy of (CdSe)12 much lower than the HOMO− LUMO gap of 4.00 eV. For the larger (CdSe)12@(ZnS)48, the first excitation energy decreases to 3.00 eV with respect to 3.27 eV of the (CdSe)12 QD. Considering that the core has the same size as the bare (CdSe)12 cluster, the effect can be attributed largely to the ZnS shell. TDDFT could not be employed to compute the HOMO−LUMO excitation energy of the QD in the(CdSe)12@(ZnS)48-1 complex due to its large size. However, assuming that the trends are well reproduced by the simple Kohn−Sham orbitals energy level, we can estimate a red shift of only 0.03 eV upon interaction with complex 1. QD- 1: The ET Mechanism. Under the assumption the electron donor is the excited QD* and the acceptor is complex 1, two mechanisms for the QD-mediated photoactivation of 1 are possible: (a) electronic energy transfer (EET) or (b) direct electron transfer (ET). EET processes (Dexter and Forster) can be ruled out since no donor−acceptor spectral overlap is present, thus leaving direct ET as the only valid alternative. The standard theoretical framework for ET processes is the Marcus theory,57 which in the nonadiabatic and high-temperature limits assumes the form kET =

⎡ −(ΔG + λ)2 ⎤ π |HDA|2 exp⎢ ⎥ 4λk bT ℏλk bT ⎣ ⎦

ΔG⧧ =

(ΔG + λINT)2 4λINT

(2)

Notably, λINT and ΔG have opposite signs; hence, the largest ET rate is reached at values of ΔG ∼ λINT. For values of ΔG < λINT the Marcus theory predicts that the ET rate increases with the driving force, while for ΔG > λINT a decrease of ET rate occurs at increased driving forces (inverted regime). In Table 2 Table 2. Reorganization Energies, Gibbs Free Energy Differences (Absolute Values), and Electronic Couplings Calculated for the QD-1 Nanocomposites Studied, Along with the Corresponding Electron Transfer Rates (s−1) Calculated by Means of the Marcus Equationa (CdSe)12 (CdSe)12@(Zns)48 1a 1b

λQD + QD*

λQD + QD*

0.22 0.23 − −

0.33 0.50 − −

0.89 0.59 − 0.82 0.58 − − − 1.64 − − 1.21 Marcus Parameters

(CdSe)12−1a CdSe)12−1b (CdSe)12@(ZnS)48−1a (CdSe)12@(ZnS)48−1b a

−

+

λQD QD

λQD QD*

λ11−

ΔG

ΔG‡

λINT

HDA

1.46 0.98 1.82 1.34

0.27 0.23 0.09 0.08

2.17 1.71 2.22 1.77

0.235 0.064 0.227 0.134

λ11

− − 1.48 1.00 kET

6.7 2.5 2.9 8.3

× × × ×

1013 1012 1014 1013

All energies are expressed in eV.

we present all the computed parameters of eq 1 for two processes, one associated with the electron injection from the excited QD* (bare and core−shell) to complex 1, to form QD+ and 1a, and the other to form complex 1b (see Supporting Information for details on the method). Internal Reorganization Energy. In the weak interaction limit the reorganization energy is computed as +

λINT =

QD*→ QD+ λINT

+

1 → 1− λINT

=

QD QD * λQD + + λ QD *

2

−

λ 1− + λ11 + 1 2

(3) +

−

→QD where λQD and λ1→1 INT* INT are the reorganization energies of the excited QD* and complex 1 (whether as 1a or 1b), respectively, during the ET mechanism, and λAB indicates the deformation energy of adduct A (QD*, QD+, 1, 1−) to attain the geometry of B (QD+, QD*, 1−, 1). As shown in Table 2, the (CdSe)12 and (CdSe)12@(ZnS)48 QDs present deformation →QD+ energies λQD of 0.33 and of 0.50 eV, respectively. These INT* values are likely overestimated because in the real environment the QDs are passivated by ligands that dampen the structural rearrangements. However, the structural deformations of complex 1a and 1b are well described and show larger values due to the population of the σ-antibonding Pt−X orbitals (vide supra). The elongation of the Pt−O bond distances is more effective and leads to a deformation energy of 1.64 eV for the 1a complex, as compared to 1.21 eV for the 1b, where the Pt− Cl bond is stretched. Overall, the total reorganization energies λINT after electron injection, computed from eq 3, are therefore quite large,

(1)

where kb is the Boltzmann constant, T is the temperature, and ℏ is the reduced Planck constant. According to 1, the kET depends on three tunable variables: (i) the electron coupling term HDA; (ii) the driving force ΔG for the charge separation process; and (iii) the reorganization energy λ, which quantifies the deformation energy of the donor and acceptor upon ET. This latter term is usually decomposed into internal reorganization energy, λINT, which reflects the response of the molecular donor and acceptor systems to ET, and external reorganization energy, λEXT, associated instead with the rearrangement of the solvent upon ET. Because an excess of 8717



Article

Figure 5. Selected regions of the 1H NMR spectrum of 1 in H2O/D2O (5:1, [1] = 1 mM) in the dark (bottom), under irradiation at 385 nm (middle, 2 min, ∼ 40 mW·cm−2), and under irradiation at 630 nm in the presence of PEGylated CdSe@ZnS QDs (top, [QD] = 100 nM, 3 h, ∼ 20 mW·cm−2).

ranging from 2.17 eV in (CdSe)12−1a to 1.71 eV in (CdSe)12− 1b. Gibbs Free Energy: Difference ΔG. The formula to evaluate ΔG is the following (Supporting Information for further details) +

QD * QD − 1 → QD QD − 1 → QD *−1 ΔG = −[(ω(QD − λQD ) − (ω(QD − 1) − 1) +

the supermolecular complex QD−1. According to Table 2, in both types of QDs, the electron coupling is ca. two times larger when the ET induces the release of the succinate ligands to form the 1a complex. The electronic coupling enters the Marcus equation with the square of its value; therefore, only according to Fermi’s golden rule, the kinetics of ET to form the 1a complex is four times larger than to form 1b. Furthermore, the electronic coupling is much smaller than the reorganization energy, validating the weak interaction limit of the Marcus equation. Table 2 provides estimates for the kinetics of ET calculated as in eq 1. In absolute values, these terms might be far from the real experimental value because of the high sensitivity of the Marcus equation to the exponential term. Nevertheless, trends are qualitatively well described and show how electron injection from both types of QD to the Pt complex more favorably leads to 1a than 1b, i.e., to release the succinate ligand and form Pt(II) species, among them cisplatin. To verify this DFT result, we followed by 1H NMR the photolysis (λexc = 630 nm) of 1 in the presence of CdSe@ZnS core−shell QDs. NMR shows that the photoreduction is much cleaner under such conditions compared to UV-light irradiation (Figure 5). In particular, reduction of Pt(II) and dissociation of succinate appears to be favored compared to formation of the Pt(IV) photoproducts responsible for the set of peaks at 5.5−7 ppm. The presence of cisplatin and cisplatin-like species was confirmed by GMP binding experiments (Supporting Information, Figures S17 and S18) and previously by XPS.34 Calculations and experiments point out that interaction of 1 with the semiconductor surface can affect the photochemical and redox behavior of this Pt(IV) derivative. Selection of photochemical pathways through thermodynamic and kinetic control of the QD−metal complex interaction has the potential to create new fascinating scenarios for future developments on hybrid nanosystems coupled with metal complexes for a range of application (e.g., phototherapy, photocatalysis).

− 1−

−

QD − (λQD + λ11 ))]

(4)

*−1 where ωQD−1→QD is the TDDFT vertical excitation energy of (QD−1) the QD−1a (or QD−1b) complex associated with a localized + −1− QD → QD* HOMO−LUMO excitation; ωQD−1→QD is the (QD−1) TDDFT vertical excitation energy associated with the chargeseparation process and is usually the lowest computed excited state. The choice of the PBE0 functional is supposed to minimize the statistical error inherent in TDDFT for the estimation of charge transfer states. From Table 2, we can infer that the computed ΔG values are smaller than λINT and, therefore, the ET process occurs always in the normal regime of Marcus theory. For the smaller (CdSe)12 QD, the ΔG of injection to form complex 1a (1.46 eV) is larger by about 0.50 eV than to form 1b (0.98 eV). However, the corresponding activation energy ΔG‡ is more favorable for the creation of the 1b complex, owing to a lower reorganization energy to rearrange the QD*−1 complex (see eq 2 above). The same trend is found for the core−shell QD, which shows a larger ΔG value of injection to form 1a, but lower activation energy. Interestingly, the absolute ΔG values of injection from (CdSe)12@(ZnS)48 to 1 are higher than in the smaller cluster: 1.82 (1a) and 1.34 (1b) eV versus 1.46 (1a) and 0.98 (1b) eV. The consequence is that the overall ET activation energies are smaller for the core−shell systems, with the effect that the larger QD might inject electrons more rapidly into the Pt complex than (CdSe)12. Electronic Coupling and Kinetics of Charge Separation. The most significant results obtained in this work involve electronic coupling (HDA). HDA is computed by evaluating the charge transfer integrals between the LUMO of the QD, corresponding to the donor orbital of QD*, and the LUMO+1 (for 1a) or LUMO (for 1b) of complex 1, both representing the acceptor orbitals. The HDA is calculated on the geometry of

■

CONCLUSIONS In summary, combination of DFT/TDDFT with selected experiments was used to gain new insights into the photochemistry and reduction behavior of 1 and to investigate the QD-mediated photoactivation of the complex, a precursor of 8718



Article

the Department of Education, Universities and Research of the Basque Country (grants PI-2012-33 and IT588-13). L.S. and E.R. are supported by the MICINN of Spain with the Ramón y Cajal Fellowship RYC-2011-07787 and by the MC CIG fellowship UCnanomat4iPACT (grant n. 321791). J.M.A. thanks the Spanish Ministry of Education for funding through a FPU fellowship (AP2009-1514). The SGI/IZO-SGIker UPV/ EHU is gratefully acknowledged for generous allocation of computational resources. We also deeply thank Serge Gorelski for the prompt implementation of the charge transfer integrals within AOMIX and for his help. L.S. and E.R. thank Dr. L. Yate for his support with XPS and members of the European COST Action CM1105 for stimulating discussions.

cisplatin. Our study allows to draw the following key conclusions: (a) Excitation of 1 with UV light leads to the generation of dissociative singlet and triplet excited states which can evolve toward the formation of different Pt(IV) photoproducts, among which a likely product is the cytotoxic cis,cis,trans[PtIV(NH3)2(Cl)2(OH)2] complex. As confirmed by XPS data, direct UV light excitation does not seem to efficiently promote Pt(IV) → Pt(II) conversion. On the contrary, 1 can be transformed into at least two distinct Pt(II) complexes upon chemical reduction, including cisplatin. (b) QD-mediated light activation can promote the formation of Pt(II) photoproducts as a result of the interaction between 1 and the surface of the semiconductor material. Fitting of quenching data shows that the complex is adsorbed onto the QD surface. Consistently, DFT calculations on two different QD-1 models gives strong stabilization energies for such adducts. (c) The electron transfer (ET) occurring from the QD to 1 falls in the normal regime of Marcus theory (ΔG < λINT) where increase in driving force favors the process. Irrespective of the QD model, the ΔG values calculated predict the formation of 1a to be favored with respect to 1b; however, a higher activation energy ΔG‡ needs to be surmounted due to an increased reorganization energy. ET appears to be favored in the core−shell model, as shown by the increase of ΔG and the decrease of ΔG‡ relative to the core-only model. Crucially, computation of the electronic coupling HDA highlights that electrons are better transferred from the QD to the LUMO+1 of 1 to generate a transient intermediate such as 1a, hence the release of the succinate ligands. The second possible pathway involving the LUMO and leading to 1b is indeed less probable, consistently with a cleaner photolysis reaction in the presence of QDs with respect to UV excitation. Such findings suggests that activation through QDs might also drive the formation of the preferred photoproduct (e.g., cisplatin) compared to simple chemical reduction or direct light excitation, a fine prospect for developing anticancer agents acting through novel and selected mechanisms of actions.

■

■

(1) Murphy, C. J.; Coffer, J. L. Quantum Dots: A Primer. Appl. Spectrosc. 2002, 56, 16−27. (2) Hetsch, F.; Xu, X.; Wang, H.; Kershaw, S. V.; Rogach, A. L. Semiconductor Nanocrystal Quantum Dots as Solar Cell Components and Photosensitizers: Material, Charge Transfer, and Separation Aspects of Some Device Topologies. J. Phys. Chem. Lett. 2011, 2, 1879−1887. (3) Kamat, P. V. Quantum Dot Solar Cells. Semiconductor Nanocrystals as Light Harvesters. J. Phys. Chem. C 2008, 112, 18737−18753. (4) Sargent, E. H. Colloidal Quantum Dot Solar Cells. Nat. Photonics 2012, 6, 133−135. (5) Kamat, P. V. Meeting the Clean Energy Demand: Nanostructure Architectures for Solar Energy Conversion. J. Phys. Chem. C 2007, 111, 2834−2860. (6) Kamat, P. V. Boosting the Efficiency of Quantum Dot Sensitized Solar Cells through Modulation of Interfacial Charge Transfer. Acc. Chem. Res. 2012, 45, 1906−1915. (7) Nann, T.; Skinner, W. M. Quantum Dots for Electro-Optic Devices. ACS Nano 2011, 5, 5291−5295. (8) Hildebrandt, N. Biofunctional Quantum Dots: Controlled Conjugation for Multiplexed Biosensors. ACS Nano 2011, 5, 5286− 5290. (9) Medintz, I. L.; Uyeda, H. T.; Goldman, E. R.; Mattoussi, H. Quantum Dot Bioconjugates for Imaging, Labelling, and Sensing. Nat. Mater. 2005, 4, 435−446. (10) Alivisatos, A. P. Perspectives on the Physical Chemistry of Semiconductor Nanocrystals. J. Phys. Chem. 1996, 100, 13226−13239. (11) Alivisatos, A. P. Semiconductor Clusters, Nanocrystals, and Quantum Dots. Science 1996, 271, 933−937. (12) Mora-Sero, I.; Bertoluzzi, L.; Gonzalez-Pedro, V.; Gimenez, S.; Fabregat-Santiago, F.; Kemp, K. W.; Sargent, E. H.; Bisquert, J. Selective Contacts Drive Charge Extraction in Quantum Dot Solids via Asymmetry in Carrier Transfer Kinetics. Nat. Commun. 2013, 4, 2272. (13) Maldonado, C. R.; Salassa, L.; Gomez-Blanco, N.; MarequeRivas, J. C. Nano-Functionalization of Metal Complexes for Molecular Imaging and Anticancer Therapy. Coord. Chem. Rev. 2013, 257, 2668− 2688. (14) Garaikoetxea Arguinzoniz, A.; Ruggiero, E.; Habtemarian, A.; Hernández-Gil, J.; Salassa, L.; Mareque-Rivas, J. C. Light Harvesting and Photoemission by Nanoparticles for Photodynamic Therapy. Part. Part. Syst. Charact. 2014, 31, 46−75. (15) Peterson, M. D.; Holbrook, R. J.; Meade, T. J.; Weiss, E. A. Photoinduced Electron Transfer from PbS Quantum Dots to Cobalt(III) Schiff Base Complexes: Light Activation of a Protein Inhibitor. J. Am. Chem. Soc. 2013, 135, 13162−13167. (16) Tseng, H.-W.; Wilker, M. B.; Damrauer, N. H.; Dukovic, G. Charge Transfer Dynamics between Photoexcited CdS Nanorods and Mononuclear Ru Water-Oxidation Catalysts. J. Am. Chem. Soc. 2013, 135, 3383−3386.

ASSOCIATED CONTENT

S Supporting Information *

Full account of DFT and TDDFT data, materials and experimental methods, NMR and cytotoxicity results, and electron transfer kinetics calculation details. This material is available free of charge via the Internet at http://pubs.acs.org.

■

REFERENCES

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Author Contributions #

These authors contributed equally.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Our work in this area was supported by Spanish Ministry of Economy and Competitiveness (grants CTQ2012-38496-C0501, CTQ2011-22723, CTQ2012-39315, and PRI−PIBIN-20110812), the Department of Industry of the Basque Country (grant ETORTEK and grant SAIOTEK S-PC12UN003), and 8719



Article

(17) Long, R.; Prezhdo, O. V. Ab Initio Nonadiabatic Molecular Dynamics of the Ultrafast Electron Injection from a PbSe Quantum Dot into the TiO2 Surface. J. Am. Chem. Soc. 2011, 133, 19240−19249. (18) Chaban, V. V.; Prezhdo, V. V.; Prezhdo, O. V. Covalent Linking Greatly Enhances Photoinduced Electron Transfer in Fullerene− Quantum Dot Nanocomposites: Time-Domain Ab Initio Study. J. Phys. Chem. Lett. 2012, 4, 1−6. (19) Varbanov, H. P.; Jakupec, M. A.; Roller, A.; Jensen, F.; Galanski, M.; Keppler, B. K. Theoretical Investigations and Density Functional Theory Based Quantitative Structure−Activity Relationships Model for Novel Cytotoxic Platinum(IV) Complexes. J. Med. Chem. 2013, 56, 330−344. (20) Gramatica, P.; Papa, E.; Luini, M.; Monti, E.; Gariboldi, M.; Ravera, M.; Gabano, E.; Gaviglio, L.; Osella, D. Antiproliferative Pt(IV) Complexes: Synthesis, Biological Activity, and Quantitative Structure−Activity Relationship Modeling. J. Biol. Inorg. Chem. 2010, 15, 1157−1169. (21) Mukhopadhyay, S.; Barnés, C. M.; Haskel, A.; Short, S. M.; Barnes, K. R.; Lippard, S. J. Conjugated Platinum(IV)−Peptide Complexes for Targeting Angiogenic Tumor Vasculature. Bioconjugate Chem. 2008, 19, 39−49. (22) Graf, N.; Bielenberg, D. R.; Kolishetti, N.; Muus, C.; Banyard, J.; Farokhzad, O. C.; Lippard, S. J. αVβ3 Integrin-Targeted PLGA-PEG Nanoparticles for Enhanced Anti-Tumor Efficacy of a Pt(IV) Prodrug. ACS Nano 2012, 6, 4530−4539. (23) Dhar, S.; Daniel, W. L.; Giljohann, D. A.; Mirkin, C. A.; Lippard, S. J. Polyvalent Oligonucleotide Gold Nanoparticle Conjugates as Delivery Vehicles for Platinum(IV) Warheads. J. Am. Chem. Soc. 2009, 131, 14652−14653. (24) Dhar, S.; Liu, Z.; Thomale, J.; Dai, H.; Lippard, S. J. Targeted Single-Wall Carbon Nanotube-Mediated Pt(IV) Prodrug Delivery Using Folate as a Homing Device. J. Am. Chem. Soc. 2008, 130, 11467−11476. (25) Farrer, N. J.; Salassa, L.; Sadler, P. J. Photoactivated Chemotherapy (PACT): the Potential of Excited-State d-Block Metals in Medicine. Dalton Trans. 2009, 10690−10701. (26) Schatzschneider, U. Photoactivated Biological Activity of Transition-Metal Complexes. Eur. J. Inorg. Chem. 2010, 1451−1467. (27) Farrer, N. J.; Woods, J. A.; Salassa, L.; Zhao, Y.; Robinson, K. S.; Clarkson, G.; Mackay, F. S.; Sadler, P. J. A Potent Trans-Diimine Platinum Anticancer Complex Photoactivated by Visible Light. Angew. Chem., Int. Ed. 2010, 49, 8905−8908. (28) Westendorf, A. F.; Woods, J. A.; Korpis, K.; Farrer, N. J.; Salassa, L.; Robinson, K.; Appleyard, V.; Murray, K.; Gruenert, R.; Thompson, A. M.; Sadler, P. J.; Bednarski, P. J. Trans,trans,trans[PtIV(N3)2(OH)2(py)(NH3)]: A Light-Activated Antitumor Platinum Complex That Kills Human Cancer Cells by an ApoptosisIndependent Mechanism. Mol. Cancer Ther. 2012, 11, 1894−1904. (29) Burksa, P. T.; Ford, P. C. Quantum Dot Photosensitizers. Interactions with Transition Metal Centers. Dalton Trans. 2012, 41, 13030−13042. (30) Ruggiero, E.; Habtemariam, A.; Yate, L.; Mareque-Rivas, J. C.; Salassa, L. Near Infrared Photolysis of a Ru Polypyridyl Complex by Upconverting Nanoparticles. Chem. Commun. 2014, 50, 1715−1718. (31) Min, Y.; Li, J.; Liu, F.; Yeow, E. K. L.; Xing, B. Near-Infrared Light-Mediated Photoactivation of a Platinum Antitumor Prodrug and Simultaneous Cellular Apoptosis Imaging by Upconversion-Luminescent Nanoparticles. Angew. Chem., Int. Ed. 2014, 53, 1012−1016. (32) Dai, Y.; Xiao, H.; Liu, J.; Yuan, Q.; Ma, P.; Yang, D.; Li, C.; Cheng, Z.; Hou, Z.; Yang, P.; Lin, J. In Vivo Multimodality Imaging and Cancer Therapy by Near-Infrared Light-Triggered trans-Platinum Pro-Drug-Conjugated Upconverison Nanoparticles. J. Am. Chem. Soc. 2013, 135, 18920−18929. (33) Blanco, N. G.; Maldonado, C. R.; Mareque-Rivas, J. C. Effective Photoreduction of a Pt(IV) Complex with Quantum Dots: A Feasible New Light-Induced Method of Releasing Anticancer Pt(II) Drugs. Chem. Commun. 2009, 5257−5259. (34) Maldonado, C. R.; Gomez-Blanco, N.; Jauregui-Osoro, M.; Brunton, V. G.; Yate, L.; Mareque-Rivas, J. C. QD-Filled Micelles

Which Combine SPECT and Optical Imaging with Light-Induced Activation of a Platinum(IV) Prodrug for Anticancer Applications. Chem. Commun. 2013, 49, 3985−3987. (35) Frisch, M. J.; et al.. Gaussian 09, revision B.01; Gaussian, Inc.: Wallingford, CT, 2009. (36) Reithofer, M. R.; Valiahdi, S. M.; Jakupec, M. A.; Arion, V. B.; Egger, A.; Galanski, M.; Keppler, B. K. Novel Di- and Tetracarboxylatoplatinum( IV) Complexes. Synthesis, Characterization, Cytotoxic Activity, and DNA Platination. J. Med. Chem. 2007, 50, 6692−6699. (37) Azpiroz, J. M.; Ugalde, J. M.; Infante, I. Benchmark Assessment of Density Functional Methods on Group II-VI MX (M = Zn, Cd; X = S, Se, Te) Quantum Dots. J. Theor. Comput. Chem. 2014, 10, 76−89. (38) York, D. M.; Karplus, M. A Smooth Solvation Potential Based on the Conductor-Like Screening Model. J. Phys. Chem. A 1999, 103, 11060−11079. (39) Simon, S.; Duran, M.; Dannenberg, J. J. How Does Basis Set Superposition Error Change the Potential Surfaces for HydrogenBonded Dimers? J. Chem. Phys. 1996, 105, 11024. (40) Boys, S. F.; Bernardi, F. The Calculation of Small Molecular Interactions by the Differences of Separate Total Energies. Some Procedures with Reduced Errors. Mol. Phys. 1970, 19, 553. (41) Krylov, A. I.; Gill, P. M. W. Q-Chem: An Engine for Innovation. WIREs Comput. Mol. Sci. 2013, 3, 317−326. (42) O’Boyle, N. M.; Tenderholt, A. L.; Langner, K. M. cclib: a Library for Package-Independent Computational Chemistry Algorithms. J. Comput. Chem. 2008, 29, 839−845. (43) Gorelsky, S. I. AOMix: Program for Molecular Orbital Analysis, 6.81 ed.; University of Ottawa: Ottawa, Canada, 2012. (44) Gorelsky, S. I.; Lever, A. B. P. Electronic Structure and Spectra of Ruthenium Diimine Complexes by Density Functional Theory and INDO/S. Comparison of the Two Methods. J. Organomet. Chem. 2001, 635, 187−196. (45) Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.; Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E. UCSF ChimeraA Visualization System for Exploratory Research and Analysis. J. Comput. Chem. 2004, 25, 1605−1612. (46) Garino, C.; Salassa, L. The Photochemistry of Transition Metal Complexes Using DFT. Philos. Trans. R. Soc. A 2013, 20120134. (47) Mulliken, R. S. Electronic Population Analysis on LCAO-MO Molecular Wave Functions. I. J. Chem. Phys. 1955, 23, 1833−1840. (48) Hirshfeld, F. L. Bonded-Atom Fragments for Describing Molecular Charge Densities. Theor. Chem. Acc. 1977, 44, 129−38. (49) Foster, J. P.; Weinhold, F. Natural Hybrid Orbitals. J. Am. Chem. Soc. 1980, 102, 7211−7218. (50) Nemirovski, A.; Vinograd, I.; Takrouri, K.; Mijovilovich, A.; Rompel, A.; Gibson, D. New Reduction Pathways for ctc[PtCl2(CH3CO2)2(NH3)(Am)] Anticancer Prodrugs. Chem. Commun. 2010, 46, 1842−1844. (51) Zhang, J. Z.; Wexselblatt, E.; Hambley, T. W.; Gibson, D. Pt(IV) Analogs of Oxaliplatin That Do Not Follow the Expected Correlation between Electrochemical Reduction Potential and Rate of Reduction by Ascorbate. Chem. Commun. 2012, 48, 847−849. (52) Koposov, A. Y.; Cardolaccia, T.; Albert, V.; Badaeva, E.; Kilina, S.; Meyer, T. J.; Tretiak, S.; Sykora, M. Formation of Assemblies Comprising Ru−Polypyridine Complexes and CdSe Nanocrystals Studied by ATR-FTIR Spectroscopy and DFT Modeling. Langmuir 2011, 27, 8377−8383. (53) Sykora, M.; Petruska, M. A.; Alstrum-Acevedo, J.; Bezel, I.; Meyer, T. J.; Klimov, V. I. Photoinduced Charge Transfer between CdSe Nanocrystal Quantum Dots and Ru−Polypyridine Complexes. J. Am. Chem. Soc. 2006, 128, 9984−9985. (54) Koopmans, T. Ü ber die Zuordnung von Wellenfunktionen und Eigenwerten zu den Einzelnen Elektronen Eines Atoms. Physica 1934, 1, 104−113. (55) Theoretically, there is still some debate whether Koopman’s Theorem hold within DFT. However, in this contest, it provides useful information on IP trends. 8720



Article

(56) Shalom, M.; Ruhle, S.; Hod, I.; Yahav, S.; Zaban, A. Energy Level Alignment in CdS Quantum Dot Sensitized Solar Cells Using Molecular Dipoles. J. Am. Chem. Soc. 2009, 131, 9876−9877. (57) Kilina, S.; Ivanov, S.; Tretiak, S. Effect of Surface Ligands on Optical and Electronic Spectra of Semiconductor Nanoclusters. J. Am. Chem. Soc. 2009, 131, 7717−7726.

8721


Quantum Dot Photoactivation of Pt(IV) Anticancer Agents: Evidence of

Recommend Documents