1 3
Top Curr Chem (Z) (2018) 376:24
signals in the T-stacked conformation, due to bright S 1–2 → CT transitions. It is
worth noting that the appearance of these important ESA signals in the 2D maps
would be completely missed in standard exciton modeling, with SOS//QM/MM
computations of chromophore aggregates here showing their relevance.
In another example, we benchmarked low-cost methods for indole [82]. The first
two (ππ*) excited states (L b and L a ) of indole (see Fig. 5) have essentially opposite character: L b is covalent and apolar, with low oscillator strength, while L a is
ionic, polar and bright. Thus, indole represents a more intricate challenge compared
to benzene and phenol, which possess a single (covalent) state absorbing in the
NUV. Understandably, additional compromises had to be made in the calibration
procedure. Nevertheless, a RAS(4,5|0,0|4,4) level of theory followed by a singlestate RASPT2 energy correction utilizing an imaginary level shift parameter of 0.5
a.u. (Figure 11a) shows reasonable agreement with the reference data obtained with
RAS(10,9|2,12) level of theory (Figs. 5c, 11a) for states below 8 eV, with significantly reduced computational effort. Employing the cost-efficient protocols developed for computing the electronic structures of indole and phenol, we applied the
SOS//QM/MM scheme to the Trp-cage peptide, a common protein model for studying protein folding/unfolding, which contains Trp and Tyr side chains. As for the
Phe- and Tyr-containing CFYC tetrapeptide (Fig. 10), two-color 2DUV–Vis spectroscopy can be used to detect ESA signatures of chromophore–chromophore interactions. We provide an indirect way to detect Tyr and Phe, whose absorption bands
remain hidden under the (more) intense envelope of Trp in linear absorption experiments. A signal-free probing window (between 18,000 and 24,000 cm
−1
) along the
Ω 1 trace of Trp seems to facilitate this undertaking. Specifically, weak CT signals
are resolved in this spectral window in a folded Trp-cage conformation with the two
chromophores at a distance of about 5 Å. This outcome indicates that an experimentally detected enhancement of the ESA signal in Tyr- and Trp-containing peptides,
in the (otherwise) signal-free Vis spectral window between 18,000 and 24,000 cm
−1
,
should be regarded as a clear signature of chromophore–chromophore proximity of
5 Å or below.
From the above examples, we can thus conclude that 2DUV spectroscopy can
be used to distinguish between folded and unfolded structures of protein models
with interacting and non-interacting UV-chromophores, respectively. The question then arises as to whether, given the high temporal (femtosecond) resolution of
2DUV experiments, this technique could also be used to monitor folding/unfolding dynamics of protein models and, more generally, whether it represents a useful
tool for tracking GS conformational dynamics. We have exploited the SOS//QM/
MM computational recipes described above to tackle these questions by simulating the 2DUV spectra along the unfolding dynamics of the CFYC tetrapeptide [54].
Figure 12 shows the structural changes relative to the inter-chromophore distance
(d) and angle (between vectors normal to the aromatic planes, α) observed during a
ratchet-and-pawl biased molecular dynamics (rMD) simulation [84], indicating how,
starting from a closed T-stacked conformation (with d = 4.5–6 Å and α > 60°), the
CFYC peptide unfolds (d > 6.5 Å), passing through twisted offset stacked conformations (with d = 5–6.5 Å and α = 40–60°). Notably, given the restricted flexibility of
the peptide backbone, the two aromatic side chains are never observed in a parallel
89
Reprinted from the journal
Top Curr Chem (Z) (2018) 376:24
signals in the T-stacked conformation, due to bright S 1–2 → CT transitions. It is
worth noting that the appearance of these important ESA signals in the 2D maps
would be completely missed in standard exciton modeling, with SOS//QM/MM
computations of chromophore aggregates here showing their relevance.
In another example, we benchmarked low-cost methods for indole [82]. The first
two (ππ*) excited states (L b and L a ) of indole (see Fig. 5) have essentially opposite character: L b is covalent and apolar, with low oscillator strength, while L a is
ionic, polar and bright. Thus, indole represents a more intricate challenge compared
to benzene and phenol, which possess a single (covalent) state absorbing in the
NUV. Understandably, additional compromises had to be made in the calibration
procedure. Nevertheless, a RAS(4,5|0,0|4,4) level of theory followed by a singlestate RASPT2 energy correction utilizing an imaginary level shift parameter of 0.5
a.u. (Figure 11a) shows reasonable agreement with the reference data obtained with
RAS(10,9|2,12) level of theory (Figs. 5c, 11a) for states below 8 eV, with significantly reduced computational effort. Employing the cost-efficient protocols developed for computing the electronic structures of indole and phenol, we applied the
SOS//QM/MM scheme to the Trp-cage peptide, a common protein model for studying protein folding/unfolding, which contains Trp and Tyr side chains. As for the
Phe- and Tyr-containing CFYC tetrapeptide (Fig. 10), two-color 2DUV–Vis spectroscopy can be used to detect ESA signatures of chromophore–chromophore interactions. We provide an indirect way to detect Tyr and Phe, whose absorption bands
remain hidden under the (more) intense envelope of Trp in linear absorption experiments. A signal-free probing window (between 18,000 and 24,000 cm
−1
) along the
Ω 1 trace of Trp seems to facilitate this undertaking. Specifically, weak CT signals
are resolved in this spectral window in a folded Trp-cage conformation with the two
chromophores at a distance of about 5 Å. This outcome indicates that an experimentally detected enhancement of the ESA signal in Tyr- and Trp-containing peptides,
in the (otherwise) signal-free Vis spectral window between 18,000 and 24,000 cm
−1
,
should be regarded as a clear signature of chromophore–chromophore proximity of
5 Å or below.
From the above examples, we can thus conclude that 2DUV spectroscopy can
be used to distinguish between folded and unfolded structures of protein models
with interacting and non-interacting UV-chromophores, respectively. The question then arises as to whether, given the high temporal (femtosecond) resolution of
2DUV experiments, this technique could also be used to monitor folding/unfolding dynamics of protein models and, more generally, whether it represents a useful
tool for tracking GS conformational dynamics. We have exploited the SOS//QM/
MM computational recipes described above to tackle these questions by simulating the 2DUV spectra along the unfolding dynamics of the CFYC tetrapeptide [54].
Figure 12 shows the structural changes relative to the inter-chromophore distance
(d) and angle (between vectors normal to the aromatic planes, α) observed during a
ratchet-and-pawl biased molecular dynamics (rMD) simulation [84], indicating how,
starting from a closed T-stacked conformation (with d = 4.5–6 Å and α > 60°), the
CFYC peptide unfolds (d > 6.5 Å), passing through twisted offset stacked conformations (with d = 5–6.5 Å and α = 40–60°). Notably, given the restricted flexibility of
the peptide backbone, the two aromatic side chains are never observed in a parallel
89
Reprinted from the journal