Dynamic charge densities of amino acids and proteins

This PhD thesis deals with the notion of dynamic electron density and describes the effect of temperature on the electron density distribution by analyzing both static and dynamic densities of crystals. The dynamic electron densities have been successfully computed by inverse Fourier transform of accurately computed structure factors from the structure model by employing the method of fast Fourier transform (FFT). Static and dynamic electron densities corresponding to independent atom models (IAM), structure models based on high-order refinement of the IAM (IAM-HO), invariom (INV) models and multipole (MP) models have been constructed for several molecular crystals. Based on all four structure models, the static and dynamic electron densities have been calculated and compared using the low-temperature (T ≈ 20 K) high-resolution data sets of α-glycine, D,L-serine, L-alanine and L-alanyl-L-tyrosyl-L-alanine (Ala-Tyr-Ala) as well as the protein Crambin (T = 100 K). By using a multi-temperature data set of D,L-serine, the effect of temperature on electron densities have been analyzed. Density values near atomic maxima are found to be much smaller in dynamic than in static electron densities due to the thermal smearing in dynamic densities. The electron densities at bond critical points (BCPs) of covalent bonds obtained from dynamic electron densities possess slightly smaller values in comparison to the static densities. However, rather larger differences have been observed for Laplacians. The discrepancy increases with increasing polarity of the bond and with increasing temperature. Nevertheless, at temperatures below 100 K, topological properties at BCPs of dynamic electron densities provide at least a semi-quantitative estimate of the topological properties of static electron densities. In contrast to covalent bonds, electron densities at BCPs of hydrogen bonds possess slightly larger values in dynamic electron densities compared to static densities. In case of the protein Crambin, it has been found that the ADPs of Crambin at 100 K are larger or equal to ADPs of D, L-serine at 298 K. In corresponding dynamic densities, bonding features have been found to be attenuated due to the masking effects of large ADPs. As a result, the topological properties obtained from dynamic densities of Crambin at 100 K appear to be similar with the topological properties of small molecules at room temperature. Calculations using the maximum entropy method (MEM) have been undertaken with all four types of dynamic model densities, mentioned above, as prior densities. Electron density analysis by the MEM has been performed for all small molecules studied in this thesis. It is shown that MEM density maps show a tendency to converge to a density map that is independent of choice of prior. Apart from the MP model as prior, it has been found that a good characterization of chemical bonds, at least in organic molecules, can be obtained by the MEM using the IAM-HO or the INV dynamic model densities as prior, while the IAM dynamic model density as prior leads to slightly inferior MEM densities. It will become especially important for the intended application to large systems (for example proteins) where the free refinement of MP model is not possible.

Level thesis:degree_level

thesis.doctoral

Grantor dc:publisher

Universität Bayreuth

Year

2013

Author dc:creator

• Siriyara Jagannatha, Prathapa

Contributors dc:contributor

• van Smaalen, Sander

Repository record source_url

https://epub.uni-bayreuth.de/id/eprint/127/

OAI identifier oai:identifier

oai:epub.uni-bayreuth.de:127