Simulation of Two-Dimensional Infrared Spectroscopy of Peptides Using Localized Normal Modes

Artificial Intelligence-based Amide-II Infrared Spectroscopy Simulation for Monitoring Protein Hydrogen Bonding Dynamics

The structurally sensitive amide II infrared (IR) bands of proteins provide valuable information about the hydrogen bonding of protein secondary structures, which is crucial for understanding protein dynamics and associated functions. However, deciphering protein structures from experimental amide II spectra relies on time-consuming quantum chemical calculations on tens of thousands of representative configurations in solvent water. Currently, the accurate simulation of amide II spectra for whole proteins remains a challenge. Here, we present a machine learning (ML)-based protocol designed to efficiently simulate the amide II IR spectra of various proteins with an accuracy comparable to experimental results. This protocol stands out as a cost-effective and efficient alternative for studying protein dynamics, including the identification of secondary structures and monitoring the dynamics of protein hydrogen bonding under different pH conditions and during protein folding process. Our method provides a valuable tool in the field of protein research, focusing on the study of dynamic properties of proteins, especially those related to hydrogen bonding, using amide II IR spectroscopy.

Quantum-chemical calculation of two-dimensional infrared spectra using localized-mode VSCF/VCI

Computational protocols for the simulation of two-dimensional infrared (2D IR) spectroscopy usually rely on vibrational exciton models which require an empirical parameterization. Here, we present an efficient quantum-chemical protocol for predicting static 2D IR spectra that does not require any empirical parameters. For the calculation of anharmonic vibrational energy levels and transition dipole moments, we employ the localized-mode vibrational self-consistent field (L-VSCF)/vibrational configuration interaction (L-VCI) approach previously established for (linear) anharmonic theoretical vibrational spectroscopy [P. T. Panek and C. R. Jacob, ChemPhysChem 15, 3365-3377 (2014)]. We demonstrate that with an efficient expansion of the potential energy surface using anharmonic one-mode potentials and harmonic two-mode potentials, 2D IR spectra of metal carbonyl complexes and dipeptides can be predicted reliably. We further show how the close connection between L-VCI and vibrational exciton models can be exploited to extract the parameters of such models from those calculations. This provides a novel route to the fully quantum-chemical parameterization of vibrational exciton models for predicting 2D IR spectra.

Static Electron Correlation in Anharmonic Molecular Vibrations: A Hybrid TAO-DFT Study

Hybrid thermally-assisted-occupation density functional theory is used to examine the effects of static electron correlation on the prediction of a benchmark set of experimentally observed molecular vibrational frequencies. The B3LYP and B97-1 thermally-assisted-occupation measure of static electron correlation is important for describing the vibrations of many of the molecules that make up several popular test sets of experimental data. Shifts are seen for known multireference systems and for many molecules containing atoms from the second row of the periodic table of elements. Several molecules only show significant shifts in select vibrational modes, and significant improvements are seen for the prediction of hydrogen stretching frequencies throughout the test set.

Monitoring early-stage β-amyloid dimer aggregation by histidine site-specific two-dimensional infrared spectroscopy in a simulation study

Monitoring early-stage β-amyloid (Aβ) dimerization is a formidable challenge for understanding neurological diseases. We compared β-sheet formation and histidine site-specific two-dimensional infrared (2D IR) spectroscopic signatures of Aβ dimers with different histidine states (δ; N^δ1-H, ε; N^ε2-H, or π; both protonated). Molecular dynamics (MD) simulations revealed that β-sheet formation is favored for the δδδ:δδδ and πππ:πππ tautomeric isomers showing strong couplings and frequent contacts between the central hydrophobic core and C-terminus compared with the εεε:εεε isomer. Characteristic blue-shifts in the 2D IR central bands were observed upon monomer-dimer transformation. The εεε:εεε dimer exhibited larger frequency shifts than δδδ:δδδ and πππ:πππ implying that the red-shift may have a correlation with N^δ1-H(δ) protonation. Our results support the tautomerization/protonation hypothesis that attributes Aβ misfolding to histidine tautomers as a possible primary initiator for Aβ aggregation and facilitates the application of histidine site-specific 2D IR spectroscopy for studying early-stage Aβ self-assembly.

Reduced Two-Electron Interactions in Anharmonic Molecular Vibrational Calculations Involving Localized Normal Coordinates

Spatially localized vibrational normal mode coordinates are shown to reduce the importance of calculating the full set of two-electron terms in the molecular electronic Schrödinger equation. Electron correlation and dispersion interactions become less significant in (E,E)-1,3,5,7-octatetraene vibrational self-consistent field calculations when displacing remote atoms along multiple coordinates. Electron correlation interactions between spatially remote modes are also found to be less important compared to their corresponding uncorrelated interaction terms. Attenuation of the Coulomb operator indicates that the two-electron terms between remote electrons become less important for accurately describing the strongly contributing mode-coupling terms between sets of localized vibrational modes.

Computing the frequency fluctuation dynamics of highly coupled vibrational transitions using neural networks

The description of frequency fluctuations for highly coupled vibrational transitions has been a challenging problem in physical chemistry. In particular, the complexity of their vibrational Hamiltonian does not allow us to directly derive the time evolution of vibrational frequencies for these systems. In this paper, we present a new approach to this problem by exploiting the artificial neural network to describe the vibrational frequencies without relying on the deconstruction of the vibrational Hamiltonian. To this end, we first explored the use of the methodology to predict the frequency fluctuations of the amide I mode of N-methylacetamide in water. The results show good performance compared with the previous experimental and theoretical results. In the second part, the neural network approach is used to investigate the frequency fluctuations of the highly coupled carbonyl stretch modes for the organic carbonates in the solvation shell of the lithium ion. In this case, the frequency fluctuation predicted by the neural networks shows a good agreement with the experimental results, which suggests that this model can be used to describe the dynamics of the frequency in highly coupled transitions.

A Machine Learning Protocol for Predicting Protein Infrared Spectra

Infrared (IR) absorption provides important chemical fingerprints of biomolecules. Protein secondary structure determination from IR spectra is tedious since its theoretical interpretation requires repeated expensive quantum-mechanical calculations in a fluctuating environment. Herein we present a novel machine learning protocol that uses a few key structural descriptors to rapidly predict amide I IR spectra of various proteins and agrees well with experiment. Its transferability enabled us to distinguish protein secondary structures, probe atomic structure variations with temperature, and monitor protein folding. This approach offers a cost-effective tool to model the relationship between protein spectra and their biological/chemical properties.

Vibrational Spectroscopic Map, Vibrational Spectroscopy, and Intermolecular Interaction

Vibrational spectroscopy is an essential tool in chemical analyses, biological assays, and studies of functional materials. Over the past decade, various coherent nonlinear vibrational spectroscopic techniques have been developed and enabled researchers to study time-correlations of the fluctuating frequencies that are directly related to solute-solvent dynamics, dynamical changes in molecular conformations and local electrostatic environments, chemical and biochemical reactions, protein structural dynamics and functions, characteristic processes of functional materials, and so on. In order to gain incisive and quantitative information on the local electrostatic environment, molecular conformation, protein structure and interprotein contacts, ligand binding kinetics, and electric and optical properties of functional materials, a variety of vibrational probes have been developed and site-specifically incorporated into molecular, biological, and material systems for time-resolved vibrational spectroscopic investigation. However, still, an all-encompassing theory that describes the vibrational solvatochromism, electrochromism, and dynamic fluctuation of vibrational frequencies has not been completely established mainly due to the intrinsic complexity of intermolecular interactions in condensed phases. In particular, the amount of data obtained from the linear and nonlinear vibrational spectroscopic experiments has been rapidly increasing, but the lack of a quantitative method to interpret these measurements has been one major obstacle in broadening the applications of these methods. Among various theoretical models, one of the most successful approaches is a semiempirical model generally referred to as the vibrational spectroscopic map that is based on a rigorous theory of intermolecular interactions. Recently, genetic algorithm, neural network, and machine learning approaches have been applied to the development of vibrational solvatochromism theory. In this review, we provide comprehensive descriptions of the theoretical foundation and various examples showing its extraordinary successes in the interpretations of experimental observations. In addition, a brief introduction to a newly created repository Web site (http://frequencymap.org) for vibrational spectroscopic maps is presented. We anticipate that a combination of the vibrational frequency map approach and state-of-the-art multidimensional vibrational spectroscopy will be one of the most fruitful ways to study the structure and dynamics of chemical, biological, and functional molecular systems in the future.

Modeling Carbon Dioxide Vibrational Frequencies in Ionic Liquids: IV. Temperature Dependence

In previous papers in the series, the vibrational spectroscopy of CO₂ in ionic liquids (ILs) was investigated at ambient conditions. Here, we extend these studies to understand the temperature dependence of the structure, dynamics, and thermodynamics of CO₂ in the 1-butyl-3-methylimidazolium hexafluorophosphate, [bmim][PF₆], IL. Using spectroscopic mapping techniques, the infrared absorption spectrum of the CO₂ asymmetric stretch mode is simulated at a number of temperatures, and the results are found to be consistent with similar experimental studies. Structural correlation functions are used to reveal the thermodynamics of complete CO₂ solvent cage breakdown. The enthalpy and entropy of activation for solvent cage reorganization are found to be 6.9 and 7.6 (kcal/mol)/K, respectively, and these values are similar to the those for spectral, orientational, and translational diffusion. Caging times for CO₂ are calculated, and it is shown that the short time dynamics of CO₂ are unaffected by temperature, even though the long-time dynamics are highly sensitive to temperature.

Reduced Basis Set Dependence in Anharmonic Frequency Calculations Involving Localized Coordinates

Localized normal coordinates are known to be effective in speeding up anharmonic frequency calculations by reducing the complexity of the nuclear Hamiltonian and wave function. Displacing atoms in localized coordinates can also cause relatively small changes in the electronic structure, which can be exploited for further computational efficiency improvements during ab initio electronic structure calculations of the potential energy surface by reducing the electronic basis set dependence. Three different schemes for reducing the basis set dependence have been investigated in this work. These include combining localized coordinate schemes with general mixed basis sets, distance based force-field reductions, and using coordinate specific basis sets. The importance of accurately describing electronic interactions is found to diminish both for multicoordinate terms involving the displacement of remote atoms and when describing the interactions between more remote atoms within specific coordinates.

Intermediate vibrational coordinate localization with harmonic coupling constraints

Optimized normal coordinates can significantly improve the speed and accuracy of vibrational frequency calculations. However, over-localization can occur when using unconstrained spatial localization techniques. The unintuitive mixtures of stretching and bending coordinates that result can make interpreting spectra more difficult and also cause artificial increases in mode-coupling during anharmonic calculations. Combining spatial localization with a constraint on the coupling between modes can be used to generate coordinates with properties in-between the normal and fully localized schemes. These modes preserve the diagonal nature of the mass-weighted Hessian matrix to within a specified tolerance and are found to prevent contamination between the stretching and bending vibrations of the molecules studied without a priori classification of the different types of vibration present. Relaxing the constraint can also be used to identify which normal modes form specific groups of localized modes. The new coordinates are found to center on more spatially delocalized functional groups than their fully localized counterparts and can be used to tune the degree of vibrational correlation energy during anharmonic calculations.

Effective representation of amide III, II, I, and A modes on local vibrational modes: Analysis of ab initio quantum calculation results

The Hamiltonian matrix for the first excited vibrational states of a protein can be effectively represented by local vibrational modes constituting amide III, II, I, and A modes to simulate various vibrational spectra. Methods for obtaining the Hamiltonian matrix from ab initio quantum calculation results are discussed, where the methods consist of three steps: selection of local vibrational mode coordinates, calculation of a reduced Hessian matrix, and extraction of the Hamiltonian matrix from the Hessian matrix. We introduce several methods for each step. The methods were assessed based on the density functional theory calculation results of 24 oligopeptides with four different peptide lengths and six different secondary structures. The completeness of a Hamiltonian matrix represented in the reduced local mode space is improved by adopting a specific atom group for each amide mode and reducing the effect of ignored local modes. The calculation results are also compared to previous models using C=O stretching vibration and transition dipole couplings. We found that local electric transition dipole moments of the amide modes are mainly bound on the local peptide planes. Their direction and magnitude are well conserved except amide A modes, which show large variation. Contrary to amide I modes, the vibrational coupling constants of amide III, II, and A modes obtained by analysis of a dipeptide are not transferable to oligopeptides with the same secondary conformation because coupling constants are affected by the surrounding atomic environment.

Anharmonic Theoretical Vibrational Spectroscopy of Polypeptides

Because of the size of polypeptides and proteins, the quantum-chemical prediction of their vibrational spectra presents an exceptionally challenging task. Here, we address one of these challenges, namely, the inclusion of anharmonicities. By performing the expansion of the potential energy surface in localized-mode coordinates instead of the normal-mode coordinates, it becomes possible to calculate anharmonic vibrational spectra of polypeptides efficiently and reliably. We apply this approach to calculate the infrared, Raman, and Raman optical activity spectra of helical alanine polypeptides consisting of up to 20 amino acids. We find that while anharmonicities do not alter the band shapes, simple scaling procedures cannot account for the different shifts found for the individual bands. This closes an important gap in theoretical vibrational spectroscopy by making it possible to quantify the anharmonic contributions and opens the door to a first-principles calculation of multidimensional vibrational spectra.

Assessing Spectral Simulation Protocols for the Amide I Band of Proteins

We present a benchmark study of spectral simulation protocols for the amide I band of proteins. The amide I band is widely used in infrared spectroscopy of proteins due to the large signal intensity, high sensitivity to hydrogen bonding, and secondary structural motifs. This band has, thus, proven valuable in many studies of protein structure-function relationships. We benchmark spectral simulation protocols using two common force fields in combination with several electrostatic mappings and coupling models. The results are validated against experimental linear absorption and two-dimensional infrared spectroscopy for three well-studied proteins. We find two-dimensional infrared spectroscopy to be much more sensitive to the simulation protocol than linear absorption and report on the best simulation protocols. The findings demonstrate that there is still room for ideas to improve the existing models for the amide I band of proteins.