Hoch

Publikationen

Hier finden Sie alle wissenschaftlichen Publikationen und Arbeiten bzw. Veröffentlichungen von Dr. Gerald Böhm aus der Zeit von 1988 bis 2007. Die PDF-Dateien stehen zum Download gemäß den jeweiligen Urheberrechten bereit.

Halobacterium volcanii mutants that are resistant to the dihydrofolate reductase inhibitor trimethoprim contain DNA sequence amplifications. This paper describes the cloning and nucleic acid sequencing of the amplified DNA sequence of the H. volcanii mutant WR215. This sequence contains an open reading frame that codes for an amino acid sequence that is homologous to the amino acid sequences of dihydrofolate reductases from different sources. As a result of the gene amplification, the trimethoprim-resistant mutant overproduces dihydrofolate reductase. This enzyme was purified to homogeneity using ammonium sulfate-mediated chromatographies. It is shown that the enzyme comprises 5% of the cell protein. The amino acid sequence of the first 15 amino acids of the enzyme fits the coding sequence of the gene. Preliminary biochemical characterization shows that the enzyme is unstable at salt concentrations lower than 2 M and that its activity increases with increase in the KCl or NaCl concentrations.

The aim of the present work is to suggest that protein folding is a highly complex process which generally cannot be simulated on digital computers. This limitation is not due to the availability of computing ressources or exact force field parameters, as it has been suggested previously; it is obviously impossible to quantify parameter(s) for any deterministic algorithm with sufficient accuracy to describe the dynamics of protein folding. Molecular dynamics simulations on crambin, a small protein with 46 amino acids whose three-dimensional structure is known, suggest the native state to be a ‘fixed attractor’. The results show that any ab initio calculation of protein structure must fail if the folding process of a protein is controlled by a kinetic process, i.e. when the native state is the kinetically accessible minimum on the energy hyperspace but not the thermodynamically possible global minimum. In this case only non-dynamical methods like pattern recognition or database processing (knowledge-based approaches) can provide a reasonable three-dimensional structure. Novel computational methods like genetic algorithms and neural network methods may be more valuable for the design and description of protein structures than the traditional force-field based algorithmic methods used to date. Knowledge-based modelling is therefore the most promising method to date to deduce the structure of an unknown protein from the sequence information solely.

A new method based on neural network theory is presented to analyze and quantify the information content of far UV circular dichroism spectra. Using a backpropagation network model with a single hidden layer between input and output, it was possible to deduce five different secondary structure fractions (helix, parallel and antiparallel β-sheet, β-turn and random coil) with satisfactory correlations between calculated and measured secondary structure data. We demonstrate that for each wavelength interval a specific network is suitable. The remaining discrepancy between the secondary structure data from neural network prediction and crystallography may be attributed to errors in the determination of protein concentration and random noise in the CD signal, as indicated by simulations.

 
 
Powered by Phoca Download
Go to top