Team:Tuebingen/Bioinformatics
Bioinformatics
“To do something that you feel in your heart that's great, you need to make a lot of mistakes. Anything that's successful is a series of mistakes.”- Billie Joe Armstrong
Model
Overview
Working with proteins as large and potentially toxic as botolinium neurotoxin C without an experimentally derived structure poses a challenge for wetlab scientists. Any modifications on the protein may impair the structure and function of the protein in unexpected ways. Hence, we employed homology modeling algorithms to first-time derive the completely assembled structure of all botolinium neurotoxin C domains. To assess the deduced structure, and various mutations performed on the protein to detoxify it, we conducted molecular dynamics simulations.
The immune system does not only attack malicious molecules, but also proteins which were not originally part of the body, but got injected as parts of drugs. Therefore, to improve the potential usage of botolinium neurotoxin C as a shuttle mechanism for fusion proteins we developed a deimmunization workflow applicable to proteins.
Structure Elucidation
'Form follows function' is a principle routinely applied in industrial design and modern architecture. However, this concept also commonly applies to proteins.
Botolinium neurotoxin C's structure complete structure with all domains. has so far not been experimentally derived in the laboratory and submitted to publicly accessible database. Detailed knowledge of the structure of botulinum neurotoxin C may help scientists to aid in the general understanding of the protein and to assess the impact of sequence modifications on the protein, especially in the context of safety.
Instead of generating the structure of botolinium neurotoxin C using crystallography, we modeled a theoretical structure using homology modeling.
Since the generated structure has not been derived in the laboratory, we verified the structure's folding capabilities and robustness using molecular dynamics simulations.
Homology Modeling
Homology modeling describes the process of the construction of an atomic-resolution model of a target amino acid sequence using experimental three-dimensional structures of related homologous proteins templates. The identification of suitable already known protein structures which resemble the structure of the query sequence and the following alignment of residues of the query sequence to residues in the template sequences heavily influence the quality of the homology model.
Researchers have shown that protein structures are very conserved amongst homologues, even more so than protein sequences[].
Sequences with less than 20% sequence identity are likely to have different three-dimensional strucutes.[]
The presence of alignment gaps in either solely the target or the template further complicate the modeling process, since they indicate a structural region present in only one of the two structures. Moreover, it has been shown that the quality of the homology model gradually decreases with the sequence identity. A typical homology model has ~1–2 Å root mean square deviation between the matched Cα atoms at 70% sequence identity but only 2–4 $Cα agreement at 25% sequence identity.
Moreover, loop regions, where amino acid sequences of target and template proteins may completely differ, usually contain more errors[][].
The generated sequence alignment is then used for the creation of a structural model of the target protein. Significant structural similarity can usually be derived from the detectable levels of sequence similarity, since protein structure is more conserved than DNA[].
empty
empty
