Low-lying excited states of model proteins: Performances of the CC2 method versus multireference methods

Abstract

A benchmark set of relevant geometries of a model protein, the N-acetylphenylalanylamide, is presented to assess the validity of the approximate second-order coupled cluster (CC2) method in studying low-lying excited states of such bio-relevant systems. The studies comprise investigations of basis-set dependence as well as comparison with two multireference methods, the multistate complete active space 2nd order perturbation theory (MS-CASPT2) and the multireference difference dedicated configuration interaction (DDCI) methods. First of all, the applicability and the accuracy of the quasi-linear multireference difference dedicated configuration interaction method have been demonstrated on bio-relevant systems by comparison with the results obtained by the standard MS-CASPT2. Second, both the nature and excitation energy of the first low-lying excited state obtained at the CC2 level are very close to the Davidson corrected CAS+DDCI ones, the mean absolute deviation on the excitation energy being equal to 0.1 eV with a maximum of less than 0.2 eV. Finally, for the following low-lying excited states, if the nature is always well reproduced at the CC2 level, the differences on excitation energies become more important and can depend on the geometry.

CC2 Benchmark for Models of Phenylalanine Protein Chains: 0–0 Transition Energies and IR Signatures of the ππ* Excited State

“…[5][6][7][8][9][10] In this context, we have developed an original innovative computational strategy in order to document the conformer-selective dynamics of capped peptide building blocks serving as models of proteins. Developed on small capped peptides, [11][12][13] this multi-step multi-level computational strategy allows us both to characterize the PES and the dynamics of their low-lying excited states using, first, nonadiabatic dynamic simulations based on time-dependent density functional theory (NA-TDDFT) to provide hints about the critical motions that drive the deactivation, which is then refined at two better levels of theory: i) the standard approximate coupled cluster singles and doubles method (CC2) [14][15][16][17][18] and ii) a multireference configuration interaction (MRCI) method. [19][20][21] This work is fully in line with these previous works [11][12][13] and focuses on the establishment of benchmark of the CC2 method on a series of capped peptides of increasing size and containing residues of different nature.…”

Neutral Peptides in the Gas Phase: Conformation and Aggregation Issues

“…for excitation energies or shifts between S 0 and S 1 IR spectra, the approximate second-order coupled cluster method CC2 366 is mostly preferred for capped amino acids by comparison with multireference methods. 367 Excited state modelling is still a difficult matter, since the precision achieved on the electronic energies is typically of the order of ~1000 cm -1…”

“…for the low lying excited states, 367,368 which is far behind the experimental requirements for spectroscopic assignment owing to the modest spectral shifts between conformers (typically 100 cm -1 at most). However, CC2 methods has recently proven to provide valuable benchmarks on relatively large species.…”

Excited States Computation of Models of Phenylalanine Protein Chains: TD-DFT and Composite CC2/TD-DFT Protocols

“…Among the single reference methods, the time-dependent density functional theory (TD-DFT) [ 6 ], which describes excited states within response theory, exhibits a favorable cost–performance ratio, but even to only perform a qualitative exploration of the PES, the functional has to be judiciously chosen according to the nature of the excited states considered [ 7 , 8 , 9 , 10 , 11 ]. In this context, we developed an original, innovative computational strategy in order to document the conformer selective dynamics of capped peptide building blocks (including the phenylalanine (Phe) residue), serving as models of proteins [ 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 ]. Gas-phase isolation was investigated first, enabling cross-checking between the experimental data and relevant quantum chemistry methods in order to validate the theoretical approach.…”

“…Gas-phase isolation was investigated first, enabling cross-checking between the experimental data and relevant quantum chemistry methods in order to validate the theoretical approach. The challenge in such calculations is that these systems exhibit specificities that orient the choice of the approach: (i) their size (medium-size systems where the smallest one, a capped peptide with one residue, already contains at least thirty atoms); (ii) their lack of symmetry, (iii) their great flexibility due to the non-covalent interactions that govern their structure; and (iv) their multiple singlet close-low-lying excited states featuring very different natures, locally excited states on one or several peptide bonds (nπ* CO ) or on one or several phenyl rings (ππ*) and even charge transfer (CT) states from the backbone to a phenyl ring (nπ*) [ 17 , 18 , 19 , 20 ]. This justifies the “multiscale” character of our strategy, which consists of, first, non-adiabatic dynamics simulations within the TD-DFT framework to provide hints about the critical motions that drive the deactivation and then refine them at two better levels of theory.…”