Review pubs.acs.org/CR
Quantum Mechanical Studies on the Photophysics and the Photochemistry of Nucleic Acids and Nucleobases Roberto Improta,*,† Fabrizio Santoro,*,‡ and Lluís Blancafort*,§ †
Istituto di Biostrutture Biommagini (IBB-CNR), CNRConsiglio Nazionale delle Ricerche, Via Mezzocannone 16, I-80134, Napoli, Italy ‡ Area della Ricerca di Pisa, Istituto di Chimica dei Composti Organo Metallici (ICCOM-CNR), CNRConsiglio Nazionale delle Ricerche, Via G. Moruzzi 1, I-56124 Pisa, Italy § Institut de Química Computacional i Catàlisi and Departament de Química, Universitat de Girona, Campus de Montilivi, 17071 Girona, Spain ABSTRACT: The photophysics and photochemistry of DNA is of great importance due to the potential damage of the genetic code by UV light. Quantum mechanical studies have played a key role in interpretating the results of modern time-resolved pump−probe spectroscopy, and in elucidating the main photoactivated reactive paths. This review provides a concise, complete picture of the computational studies carried out, approximately, in the past decade. We start with an overview of the photophysics of the nucleobases in the gas phase and in solution. We discuss the proposed mechanisms for ultrafast decay to the ground state, that involve conical intersections, consider the role of triplet states, and analyze how the solvent modulates the photophysics. Then we move to larger systems, from dinucleotides to single- and double-stranded oligonucleotides. We focus on the possible role of charge transfer and delocalized or excitonic states in the photophysics of these systems and discuss the main photochemical paths. We finish with an outlook on the current challenges in the field and future directions of research.
CONTENTS 1. Introduction 2. Methods 2.1. Electronic Methods 2.1.1. Configuration Interaction and Propagator-Based Approaches. ADC(2) 2.1.2. Coupled-Cluster Based Approaches. CCn and EOM-CCSD 2.1.3. Multireference Methods. CASSCF, CASPT2, and MR-CI 2.1.4. Time-Dependent Density Functional Theory 2.1.5. Mixed QM/MM Methods 2.2. Solvation Models 2.3. Methods for Nonadiabatic Dynamical Calculations 2.4. Comparison with the Experimental Data 3. Isolated Bases 3.1. Uracil and Thymine 3.1.1. Franck−Condon Region in the Gas Phase 3.1.2. Franck−Condon Region in the Condensed Phase 3.1.3. Static Characterization of the Electronic Potential Energy Surfaces: Gas Phase and Solvent 3.1.4. Dynamical Studies 3.1.5. Triplet States © 2016 American Chemical Society
3.1.6. General Picture of Ura and Thy Photophysics According to QM Studies 3.2. Cytosine and Its Derivatives 3.2.1. Franck−Condon Region in the Gas Phase and in the Condensed Phase 3.2.2. Static Characterization of the Relevant Electronic Potential Energy Surfaces in the Gas Phase and in the Condensed Phase 3.2.3. Dynamical Studies 3.2.4. Cyt Triplet States 3.2.5. General Picture of Cyt Photophysics According to QM Studies 3.3. Adenine and Its Derivatives 3.3.1. Franck−Condon Region in the Gas Phase and in the Condensed Phase 3.3.2. Static Characterization of the Electronic Potential Energy Surfaces and Dynamical Studies: Gas Phase 3.3.3. Static Characterization of the Electronic Potential Energy Surfaces and Dynamical Studies: Condensed Phase 3.3.4. General Picture of Ade Photophysics According to QM Studies 3.4. Guanine and Its Derivatives
3541 3542 3542 3543 3543 3543 3544 3545 3546 3547 3548 3549 3549 3549 3551
3551 3553 3554
3555 3555 3555
3556 3558 3559 3559 3559 3559
3561
3563 3563 3564
Received: July 30, 2015 Published: March 1, 2016 3540
DOI: 10.1021/acs.chemrev.5b00444 Chem. Rev. 2016, 116, 3540−3593
Chemical Reviews 3.4.1. Franck−Condon Region in the Gas and in the Condensed Phase 3.4.2. Static Characterization of the Electronic Potential Energy Surfaces and Dynamical Studies: Gas Phase 3.4.3. Static Characterization of the Electronic Potential Energy Surfaces and Dynamical Studies: Condensed Phase 3.4.4. General Picture of Gua Photophysics According to QM Studies 4. Isolated Nucleobases in a DNA-like Environment 5. Oligonucleotides: Base Stacking 5.1. Pyrimidine−Pyrimidine Stacks 5.1.1. Franck−Condon Region in the Gas Phase and in Solution 5.1.2. Characterization of the Electronic Potential Energy Surfaces: Photophysical and Photochemical Pathways 5.1.3. Dynamical Studies 5.1.4. General Picture of Pyrimidine/Pyrimidine Excited State Dynamics According to QM Studies 5.2. Purine−Purine and Purine−Pyrimidine Stacks 5.2.1. Franck−Condon Region in the Gas Phase and in Solution 5.2.2. Characterization of the Electronic Potential Energy Surfaces: Photophysical Pathways 5.2.3. Dynamical Studies 5.2.4. General Picture of Purine/Purine Excited State Dynamics According to QM Studies 6. Oligonucleotides: Hydrogen Bond Pairing 6.1. GC Pairs 6.1.1. Photophysics in the Gas Phase 6.1.2. Photophysics in the Condensed Phase 6.1.3. Dynamical Studies 6.2. AT Pairs 6.3. General Picture of GC and AT Excited State Dynamics According to QM Studies 7. Oligonucleotides: Double Strands 8. Conclusions and Outlook Author Information Corresponding Authors Notes Biographies Acknowledgments Abbreviations References
Review
3564
3565
3566 3567 3567 3568 3568 3568 Figure 1. Schematic description (and atom labeling) of the five nucleobases: (a) uracil (R = H, Ura), uridine (R = ribose, dU), uridine-phosphate (R = ribose-monophosphate, UMP); (b) thymine (R = H, Thy), (deoxy)thymidine (R = (deoxy)ribose, dT), thymidinephosphate (R = (deoxy)ribose-monophosphate, TMP); (c) cytosine (R = H, Cyt), (deoxy)cytidine (R = (deoxy)ribose, dC), (deoxy)cytidine-phosphate (R = (deoxy)ribose-monophosphate, CMP); (d) (9,H)adenine (R = H, Ade), (deoxy)adenosine (R = (deoxy)ribose, dA), (deoxy)adenosine-phosphate (R = (deoxy)ribose-monophosphate, AMP); (e) (9,H)guanine (R = H, Gua), (deoxy)guanosine (R = (deoxy)ribose, dG), (deoxy)guanosine-phosphate (R = (deoxy)ribosemonophosphate, GMP).
3569 3570
3570 3571 3571
3573 3574
radiation spectrum. Considering that most of the prebiotic chemistry probably occurred under intense UV irradiation,11 nucleobase photostability, providing a “natural” protection toward radiation-induced damage, has likely contributed to the selection of these molecules as building blocks of the NAs.4−10 Besides their biological relevance, NAs have also shown promising properties for applications in materials science,12−14 thanks to their remarkable thermodynamic stability, selfassembling capability, and peculiar electronic structure. As a consequence, the study of the photoactivated processes in NAs has attracted remarkable scientific interest, as witnessed by the large number of papers appearing in the literature in the past two decades, as reviewed in refs 4−10. Experimentally the field has been revolutionized by the development and application of time-resolved (TR) pump− probe spectroscopies which have provided fundamental insights into the static and dynamical behaviors of the excited states of NAs and nucleobases.4−10 These studies have unveiled the intrinsic complexity of DNA photophysics. In single nucleobases, a multiexponential behavior is usually observed and, although the ground state (GS) is recovered within a few picoseconds (often on the subpicosecond time scale), longer time components have been detected as well.4,15 For single and double strands the photoinduced dynamics are even more complex, and while a significant part of the excited state (ES) population still decays within ∼1 ps, time components as long as hundreds of picoseconds or nanoseconds have been recorded.15,16 Additionally, both for isolated bases and for single/double strands the ES dynamics often depends on the excitation wavelength, suggesting the existence of competitive decay channels.4,15,16 On these grounds it is not surprising that the study of photoexcited NAs has received important benefits from the advances in the development of quantum mechanical (QM)
3574 3575 3575 3575 3576 3576 3577 3577 3577 3579 3581 3581 3581 3581 3581 3581 3582
1. INTRODUCTION Ultraviolet (UV) light absorption of nucleic acids (NAs) can trigger a cascade of dangerous photochemical events, potentially leading to the damage of the genetic code and to cell apoptosis.1−3 They are largely prevented by the existence of very fast nonradiative decays able to effectively dissipate the harmful electronic excitation into heat.4−10 Pyrimidinic (thymine, Thy; uracil, Ura; cytosine, Cyt) and purinic (adenine, Ade; guanine, Gua) nucleobases, the NA building blocks (sketched in Figure 1), are responsible for NA absorption in the 200−320 nm region, whose red wing intersects the solar 3541
DOI: 10.1021/acs.chemrev.5b00444 Chem. Rev. 2016, 116, 3540−3593
Chemical Reviews
Review
their properties determine the absorption spectra. As a second step, we will describe the PES of the most relevant electronic states, their stationary points, and their crossing regions, including the so-called conical intersections (CoIs). The results of ES dynamics will then be presented describing how, together with the analysis of the excited state PES, they allow assignment and interpretation of the TR spectra. At the end of the section devoted to each base, we shall summarize the main results in a conclusive subsection, trying to highlight what are the most relevant open issues. We shall try to provide a picture as objective as possible of the different hypotheses existing in the literature, because it is clear that different mechanisms are in principle consistent with a given set of experimental data and that any detailed discussion would require a deep analysis of the experimental results themselves. A similar scheme will be followed also in the remaining sections devoted to single- and double-stranded oligomers. In section 4 we will start reviewing the studies dealing with the ES of a single base, treated at the QM level, embedded in a strand treated at the molecular mechanics (MM) level. Although these approaches can capture DNA environmental effects on the ES dynamics of the base, they cannot describe delocalized electronic states. We will then move to discuss (in section 5) the QM studies of fragments of NA single strands (pyrimidine/pyrimidine, purine/purine, and purine/pyrimidine steps), providing insights on the effect of stacking on ES properties and, later, the hydrogen bonded Gua−Cyt and Ade−Thy WC pairs (in section 6 ). Finally, in section 7, we will cover the few QM computational studies reported up to now on larger double-stranded fragments. Section 8 will be devoted to some conclusive remarks and to a brief description of our view on the near-future perspectives in the field. Due to the large number of articles covered by our review, it will not be possible, in several cases, to provide an exhaustive discussion of the results reported in every single paper. On the other hand, several review papers on single topics (for example, isolated bases in the gas phase or in solution, modified nuclebases, single and double strands, and photodimerization processes) have recently appeared, e.g. chapters in refs 9 and 10, and we shall refer the interested reader to them whenever necessary.
methods for the study of excited electronic states. A panoply of methods are nowadays available to describe, with fairly good accuracy, the ES behavior of compounds containing up to 100−150 atoms, both in the gas phase and in solution, making feasible the theoretical study of systems with dimensions approaching those of DNA fragments studied in TR experiments. At the same time, continuous advances in methods and computational power allow nowadays going beyond a pure static description of the potential energy surfaces (PESs), tackling a direct simulation of the photoinduced nonadiabatic dynamics in nucleobases and small DNA fragments. Therefore, in analogy with what happens in several other research fields, computational approaches are nowadays a fundamental tool in the study of photoactivated dynamics in DNA.9,10,17−21 Looking at this research field from a different perspective, it is also clear that the huge number of available experimental and computational data makes the study of DNA a privileged playground for methodology developments. Nucleobases and DNA fragments have become de facto a fundamental benchmark for new QM approaches. Moreover, since DNA is also a challenging multichromophoric system, it also has become an excellent test case for computational methods designed to investigate the broad family of photoexcited supramolecular systems. In this scenario, the increasing number of studies, in some cases with contradictory results, makes it more difficult for experimentalists and the general readership to extract the most relevant information, to get a simple physical picture of the main decay pathways, and to have a clear idea of the issues that can be considered assessed and of those that are still a matter of debate. The aim of this review is to provide a concise yet complete picture of the main computational studies devoted to the study of DNA photophysics and photochemistry in the past decade. The papers published before 2004 are indeed exhaustively reported in the excellent review of Kohler et al.,4 covering both experimental and computational studies. Since the present review is focused on the photoactivated processes in DNA, we shall, necessarily, overlook some interesting issues such as electrochemically induced charge transfer and charge migration, and dynamical processes, mainly involving ionized species, occurring at high excitation energies (
