What is the Optimal Size of the Quantum Region in Embedding Calculations of Two-Photon Absorption Spectra of Fluorescent Proteins?

We systematically investigate an impact of the size and content of a quantum (QM) region, treated at the density functional theory level, in embedding calculations on one- (OPA) and two-photon absorption (TPA) spectra of the following fluorescent proteins (FPs) models: Aequorea victoria green FP (avGFP) with neutral (avGFP-n) and anionic (avGFP-a) chromophore as well as Citrine FP. We find that amino acid (a.a.) residues as well as water molecules hydrogen-bonded (h-bonded) to the chromophore usually boost both OPA and TPA processes intensity. The presence of hydrophobic a.a. residues in the quantum region also non-negligibly affects both absorption spectra but decreases absorption intensity. We conclude that to reach a quantitative description of OPA and TPA spectra in multiscale modeling of FPs, the quantum region should consist of a chromophore and most of a.a. residues and water molecules in a radius of 0.30-0.35 nm ( ca. 200-230 atoms) when the remaining part of the system is approximated by the electrostatic point-charges. The optimal size of the QM region can be reduced to 80-100 atoms by utilizing a more advanced polarizable embedding model. We also find components of the QM region that are specific to a FP under study. We propose that the F165 a.a. residue is important in tuning the TPA spectrum of avGFP-n but not other investigated FPs. In the case of Citrine, Y203 and M69 a.a. residues must definitely be part of the QM subsystem. Furthermore, we find that long-range electrostatic interactions between the QM region and the rest of the protein cannot be neglected even for the most extensive QM regions ( ca. 350 atoms).