Top Curr Chem (Z) (2018) 376:24
1 3
become impracticable when working within an SOS approach, even for the smallest conceivable chromophore aggregates, i.e. the dimeric species. Computational
recipes based on semi-empirical parameterization against our reliable reference
data enabled accurate predictions of the excited-state manifold of dimeric species at a reasonable computational cost, paving the way to the first applications of
2DUV simulation protocols for estimating the effects of inter-chromophore interactions on the nonlinear response (Sect. 3.2).
By embedding multi-configurational/multi-reference computations within hybrid
QM/MM schemes, it was possible to extend the SOS simulation (SOS//QM/MM) of
2D electronic spectra to realistic model systems of proteins and nucleic acids. This
allowed us to demonstrate how 2DUV spectroscopy holds the potential to resolve
inter-chromophore interactions occurring during GS dynamics of nucleic acid and
protein systems. In particular, simulations of 2DUV spectra of the CFYC tetrapeptide and Trp-cage protein model revealed how both electronic (quartic) coupling
shifts and charge-transfer states yield clear spectroscopic fingerprints (mainly associated with ESAs) in the one-color 2DUV–UV and two-color 2DUV–Vis spectra,
respectively. This significant outcome prompted investigations on the potential of
2DUV spectroscopy for tracking the protein folding/unfolding processes of protein
models with high temporal resolution and for the quantitative analysis of nucleobase
stacking in nucleic acid models. In this regard, extensive studies on the unfolding
dynamics of the CFYC tetrapeptide and on the conformational space characterization of the water-solvated ApA dinucleoside monophosphate indicated that the
2DUV technique represented a powerful alternative (or complement) to standard
transient absorption experiments. These results highlight the importance of accurate electronic structure computations in 2DES simulation, to account for appropriate descriptions of the ESAs signals. However, the massive computational costs of
ab initio computations restrict its application to small (still realistic) dimeric systems. Current developments in the field of multi-configurational, density-based and
linear-response techniques [95, 115–122] would certainly provide beneficial tools
for treating large (multi)chromophoric systems, provided they accomplish completeness in the description of excited-state manifolds (e.g. including doubly excited
states) [123]. Efficient QM methods employed in hybrid QM/MM schemes would
also make it possible to account for signal broadening due to solvent rearrangement,
as shown in the ApA case.
Various types of dephasing-induced broadening (due to coupling to nuclear
degrees of freedom and environment, finite excited-state lifetimes) can shape the
2D maps. When neglected, simulated signal line shapes and positions would differ significantly from their experimental counterparts. We moved a step towards
a more realistic description of spectral line shapes through the inclusion of vibrational dynamics in the simulation (Sect. 3.3). The high temporal resolution featured
in 2DUV experiments with ultrashort sub-10-fs pulses allows for the detection of
vibrational features, which manifest in coherent oscillations of the spectral signatures. These features can be computed by combining quantum–classical excited-state
dynamics simulations that account for discrete high-frequency intramolecular vibrational modes, with electronic structure calculations of the manifold of higher-lying
excited states along the dynamics. By accounting for population transfer processes
104
Reprinted from the journal
1 3
become impracticable when working within an SOS approach, even for the smallest conceivable chromophore aggregates, i.e. the dimeric species. Computational
recipes based on semi-empirical parameterization against our reliable reference
data enabled accurate predictions of the excited-state manifold of dimeric species at a reasonable computational cost, paving the way to the first applications of
2DUV simulation protocols for estimating the effects of inter-chromophore interactions on the nonlinear response (Sect. 3.2).
By embedding multi-configurational/multi-reference computations within hybrid
QM/MM schemes, it was possible to extend the SOS simulation (SOS//QM/MM) of
2D electronic spectra to realistic model systems of proteins and nucleic acids. This
allowed us to demonstrate how 2DUV spectroscopy holds the potential to resolve
inter-chromophore interactions occurring during GS dynamics of nucleic acid and
protein systems. In particular, simulations of 2DUV spectra of the CFYC tetrapeptide and Trp-cage protein model revealed how both electronic (quartic) coupling
shifts and charge-transfer states yield clear spectroscopic fingerprints (mainly associated with ESAs) in the one-color 2DUV–UV and two-color 2DUV–Vis spectra,
respectively. This significant outcome prompted investigations on the potential of
2DUV spectroscopy for tracking the protein folding/unfolding processes of protein
models with high temporal resolution and for the quantitative analysis of nucleobase
stacking in nucleic acid models. In this regard, extensive studies on the unfolding
dynamics of the CFYC tetrapeptide and on the conformational space characterization of the water-solvated ApA dinucleoside monophosphate indicated that the
2DUV technique represented a powerful alternative (or complement) to standard
transient absorption experiments. These results highlight the importance of accurate electronic structure computations in 2DES simulation, to account for appropriate descriptions of the ESAs signals. However, the massive computational costs of
ab initio computations restrict its application to small (still realistic) dimeric systems. Current developments in the field of multi-configurational, density-based and
linear-response techniques [95, 115–122] would certainly provide beneficial tools
for treating large (multi)chromophoric systems, provided they accomplish completeness in the description of excited-state manifolds (e.g. including doubly excited
states) [123]. Efficient QM methods employed in hybrid QM/MM schemes would
also make it possible to account for signal broadening due to solvent rearrangement,
as shown in the ApA case.
Various types of dephasing-induced broadening (due to coupling to nuclear
degrees of freedom and environment, finite excited-state lifetimes) can shape the
2D maps. When neglected, simulated signal line shapes and positions would differ significantly from their experimental counterparts. We moved a step towards
a more realistic description of spectral line shapes through the inclusion of vibrational dynamics in the simulation (Sect. 3.3). The high temporal resolution featured
in 2DUV experiments with ultrashort sub-10-fs pulses allows for the detection of
vibrational features, which manifest in coherent oscillations of the spectral signatures. These features can be computed by combining quantum–classical excited-state
dynamics simulations that account for discrete high-frequency intramolecular vibrational modes, with electronic structure calculations of the manifold of higher-lying
excited states along the dynamics. By accounting for population transfer processes
104
Reprinted from the journal