A Computational Study of Optical Properties of Nonaromatic Proteins: The Case of α3C

The mechanisms by which ultraviolet (UV) and visible light interact with nonaromatic molecules remain poorly understood. Traditionally, proteins lacking aromatic residues were believed to be optically silent at wavelengths longer than 250 nm. However, growing experimental evidence over the past decade has challenged this assumption, showing that proteins devoid of aromatic and conjugated groups can absorb light in the near-UV beyond 300 nm and emit in the visible region. Understanding the origins of this phenomenon offers exciting opportunities for designing noninvasive spectroscopic probes to study local interactions in biological systems. Among the emerging cases of nonaromatic optical activity, the synthetic protein α3C stands out as a representative example. Recent experiments demonstrated that α3C exhibits a broad UV-visible absorption band between 250–800 nm [1], attributed to charge-transfer excitations between charged amino acids, and emission in the 310–550 nm range upon excitation at 295 nm [2]. In this work, we investigate the origins of this unconventional absorption and emission in α3C. An unsupervised machine learning approach is used to automatically detect statistically significant structural motifs, that then subjected to QM/MM simulations of the full protein in explicit solvent using the time dependent density-functional tight-binding (TD-DFTB) method. This integrated approach streamlines the identification of statistically significant structural motifs and their direct connection to the observed absorption and emission features. Our simulated absorption spectra calculations reveal unconventional absorption features spanning 250–350 nm, with the arginine–glutamic acid interactions contributing to all transitions beyond 300 nm. However, the simulated absorption tail remains notably shorter than what is observed experimentally. In particular, our calculations do not predict any absorption between 400 and 800 nm. Transitions in this range appear only when environmental interactions are neglected - a simplification we consider physically inaccurate. To investigate the source of this discrepancy between simulated and experimental results, we examined the influence of nuclear quantum effects (NQEs) on the absorption spectra. Incorporating these effects significantly broaden the spectrum and redshift it by approximately 100 nm, extending the calculated absorption to 450 nm. Excited-state dynamics provide additional insight into the emission behavior. Although the arginine-glutamic acid interactions are more prone to low-energy electronic transitions, they contribute minimally to emission. Instead the backbone interactions appear to be the most important for the emission of the protein, when its α-helical structure remains intact. The resulting emission profile considering contributions from all the three types of interactions (arginine–glutamic acid, other side-chain interactions, and backbone H-bond contacts) combined, are in better agreement with the experimentally reported emission spectrum. Nonradiative relaxation to the ground state primarily proceeds through secondary-structure distortions, proton transfer, and arginine deplanarization. In summary, this study elucidates the molecular basis of unconventional nonaromatic fluorescence using α3C as a model system. Hydrogen bonding and charge-transfer interactions, particularly between arginine and glutamic acid residues, drive near-UV absorption, while backbone interactions dominate emission. The results highlight challenges and opportunities from both computational and experimental perspectives. Discrepancies at longer wavelengths highlight the need to incorporate nuclear quantum effects for quantitative accuracy. Experimentally, α3C variants with targeted arginine mutations and minimal peptide analogs could clarify the roles of hydrogen bonding and structural rigidity, guiding the design of new noninvasive fluorescent probes.

The mechanisms by which ultraviolet (UV) and visible light interact with nonaromatic molecules remain poorly understood. Traditionally, proteins lacking aromatic residues were believed to be optically silent at wavelengths longer than 250 nm. However, growing experimental evidence over the past decade has challenged this assumption, showing that proteins devoid of aromatic and conjugated groups can absorb light in the near-UV beyond 300 nm and emit in the visible region. Understanding the origins of this phenomenon offers exciting opportunities for designing noninvasive spectroscopic probes to study local interactions in biological systems. Among the emerging cases of nonaromatic optical activity, the synthetic protein α3C stands out as a representative example. Recent experiments demonstrated that α3C exhibits a broad UV-visible absorption band between 250–800 nm [1], attributed to charge-transfer excitations between charged amino acids, and emission in the 310–550 nm range upon excitation at 295 nm [2]. In this work, we investigate the origins of this unconventional absorption and emission in α3C. An unsupervised machine learning approach is used to automatically detect statistically significant structural motifs, that then subjected to QM/MM simulations of the full protein in explicit solvent using the time dependent density-functional tight-binding (TD-DFTB) method. This integrated approach streamlines the identification of statistically significant structural motifs and their direct connection to the observed absorption and emission features. Our simulated absorption spectra calculations reveal unconventional absorption features spanning 250–350 nm, with the arginine–glutamic acid interactions contributing to all transitions beyond 300 nm. However, the simulated absorption tail remains notably shorter than what is observed experimentally. In particular, our calculations do not predict any absorption between 400 and 800 nm. Transitions in this range appear only when environmental interactions are neglected - a simplification we consider physically inaccurate. To investigate the source of this discrepancy between simulated and experimental results, we examined the influence of nuclear quantum effects (NQEs) on the absorption spectra. Incorporating these effects significantly broaden the spectrum and redshift it by approximately 100 nm, extending the calculated absorption to 450 nm. Excited-state dynamics provide additional insight into the emission behavior. Although the arginine-glutamic acid interactions are more prone to low-energy electronic transitions, they contribute minimally to emission. Instead the backbone interactions appear to be the most important for the emission of the protein, when its α-helical structure remains intact. The resulting emission profile considering contributions from all the three types of interactions (arginine–glutamic acid, other side-chain interactions, and backbone H-bond contacts) combined, are in better agreement with the experimentally reported emission spectrum. Nonradiative relaxation to the ground state primarily proceeds through secondary-structure distortions, proton transfer, and arginine deplanarization. In summary, this study elucidates the molecular basis of unconventional nonaromatic fluorescence using α3C as a model system. Hydrogen bonding and charge-transfer interactions, particularly between arginine and glutamic acid residues, drive near-UV absorption, while backbone interactions dominate emission. The results highlight challenges and opportunities from both computational and experimental perspectives. Discrepancies at longer wavelengths highlight the need to incorporate nuclear quantum effects for quantitative accuracy. Experimentally, α3C variants with targeted arginine mutations and minimal peptide analogs could clarify the roles of hydrogen bonding and structural rigidity, guiding the design of new noninvasive fluorescent probes.