HheG from Ilumatobacter coccineus is a halohydrin dehalogenase with synthetically useful activity in the ring opening of cyclic epoxides with various small anionic nucleophiles. This enzyme provides access to chiral β-substituted alcohols that serve as building blocks in the pharmaceutical industry. Wild-type HheG suffers from low thermostability, which poses a significant drawback for potential applications. In an attempt to thermostabilize HheG by protein engineering, several single mutants at position 123 were identified which displayed up to 14 °C increased apparent melting temperatures and up to three-fold higher activity. Aromatic amino acids at position 123 resulted even in a slightly higher enantioselectivity. Crystal structures of variants T123W and T123G revealed a flexible loop opposite to amino acid 123. In variant T123G, this loop adopted two different positions resulting in an open or partially closed active site. Classical molecular dynamics simulations confirmed a high mobility of this loop. Moreover, in variant T123G this loop adopted a position much closer to residue 123 resulting in denser packing and increased buried surface area. Our results indicate an important role for position 123 in HheG and give first structural and mechanistic insight into the thermostabilizing effect of mutations T123W and T123G.

The transfer of electrons over long distances in complex molecular systems is a phenomenon of significance in both biochemistry and technology. In recent years, we have been developing efficient models to study ET in complex systems, including DNA as a prominent example. Ab initio and model approaches have been combined in an "on-the-fly" calculation of ET parameters, which can be used to propagate nuclear and electronic degrees of freedom simultaneously. These previous efforts have aimed at deriving an efficient nonadiabatic quantum mechanical-molecular mechanical (QM/MM) simulation scheme for ET, making nanosecond simulations of ET in realistic systems possible. This, however, is still insufficient for the treatment of large donor-bridge-acceptor systems, like the ET in DNA, overcoming long adenine bridges. Therefore, we have constructed a theoretical model in a bottom-up manner. All quantum-chemical as well as force-field calculations are substituted by theoretical models of the involved phenomena on a molecular level, including polarization and relaxation of the molecular environment, which are often omitted in other recently developed theoretical models of ET. A nonadiabatic simulation scheme is employed, and no assumptions regarding the ET mechanism are needed. Thus, the predictive power of the simulations is preserved, while pushing the limits of the accessible time scales beyond microseconds. This model-based simulation scheme is applied to ET in various DNA species. Good agreement with the "full" atomistic nonadiabatic QM/MM scheme is observed for the archetypal DNA ET systems, the polyA sequence, as well as the sequences GTGGG, containing adenines as bridge sites. Furthermore, ET in larger, more complex DNA sequences is simulated, and the results are discussed.

The conductivity of DNA in molecular junctions is often probed experimentally under dry conditions, but it is unclear how much of the solvent remains attached to the DNA and how this impacts its structure, electronic states, and conductivity. Classical MD simulations show that DNA is unstable if the solvent is removed completely, while a micro-hydrated system with few water molecules shows similar charge transport properties as fully solvated DNA does. This surprising effect is analyzed in detail by mapping the density functional theory-based electronic structure to a tight-binding Hamiltonian, allowing for an estimate of conductivity of various DNA sequences with snapshot-averaged Landauer's approach. The characteristics of DNA charge transport turn out to be determined by the nearest hydration shell(s), and the removal of bulk solvent has little effect on the transport.

There is no theoretical limit in using molecular networks to harvest diffusive sun photons on large areas and funnel them onto much smaller areas of highly efficient but also precious energy-converting materials. The most effective concept reported so far is based on a pool of randomly oriented, light-harvesting donor molecules that funnel all excitation quanta by ultrafast energy transfer to individual light-redirecting acceptor molecules oriented parallel to the energy converters. However, the best practical light-harvesting system could only be discovered by empirical screening of molecules that either align or not within stretched polymers and the maximum absorption wavelength of the empirical system was far away from the solar maximum. No molecular property was known explaining why certain molecules would align very effectively whereas similar molecules did not. Here, we first explore what molecular properties are responsible for a molecule to be aligned. We found a parameter derived directly from the molecular structure with a high predictive power for the alignability. In addition, we found a set of ultrafast funneling molecules that harvest three times more energy in the solar’s spectrum peak for GaInP photovoltaics. A detailed study on the ultrafast dipole moment reorientation dynamics demonstrates that refocusing of the diffusive light is based on ∼15-ps initial dipole moment depolarization followed by ∼50-ps repolarization into desired directions. This provides a detailed understanding of the molecular depolarization/repolarization processes responsible for refocusing diffusively scattered photons without violating the second law of thermodynamics.

Quantum-chemical fragmentation methods offer an efficient approach for the treatment of large proteins, in particular if local target quantities such as protein–ligand interaction energies, enzymatic reaction energies, or spectroscopic properties of embedded chromophores are sought. However, the accuracy that is achievable for such local target quantities intricately depends on how the protein is partitioned into smaller fragments. While the commonly employed naı̈ve approach of using fragments with a fixed size is widely used, it can result in large and unpredictable errors when varying the fragment size. Here, we present a systematic partitioning scheme that aims at minimizing the fragmentation error of a local target quantity for a given maximum fragment size. To this end, we construct a weighted graph representation of the protein, in which the amino acids constitute the nodes. These nodes are connected by edges weighted with an estimate for the fragmentation error that is expected when cutting this edge. This allows us to employ graph partitioning algorithms provided by computer science to determine near-optimal partitions of the protein. We apply this scheme to a test set of six proteins representing various prototypical applications of quantum-chemical fragmentation methods using a simplified molecular fractionation with conjugate caps (MFCC) approach with hydrogen caps. We show that our graph-based scheme consistently improves upon the naı̈ve approach.