In the past century, the Nobel Prize winner Christian Boehmer Anfinsen had demonstrated clearly that a protein is capable of finding its way back to its “native” 3D structure once it is placed under “denaturing conditions,” in which the structure of the protein is unfolded.
Associate Professor Frans Mulder and coworkers at Aarhus University have developed ODiNPred (Prediction of Order and Disorder by evaluation of NMR data), a software tool developed for the prediction of protein order and disorder. Image Credit: Frans Mulder.
The significant conclusion of Anfinsen’s experiments was that the information that controls the search back to the native state is apparently concealed in the sequence of amino acids.
Thermodynamic considerations subsequently proposed a view in which the folding process is similar to rolling vigorously downhill to the lowest point—that is, to the exclusive native structure.
Most often, these discoveries have been intertwined with the central belief of molecular biology. Thus, the sequence codes for a certain structure, while a gene codes for a sequence of amino acids.
Enter intrinsically disordered proteins
The introduction of rapid and low-cost genome sequencing following the human genome project led to the next breakthrough; at one time, scores of genomes of numerous organisms were sequenced, but now investigators have made a stunning discovery—there were loads of genes that coded for low-complexity proteins.
To put this in simpler terms, such proteins lacked the right amino acids to fold up and experiments proved that they stayed “intrinsically disordered.” The human genome was also found to have over one-third of its genes that code for protein disorder.
How to detect protein disorder?
Disordered proteins are very flexible, and hence they are not conducive to crystallization. Therefore, no data can be acquired from X-ray diffraction on protein crystals—the method that has been so crucial for protein folding. Rather, such proteins should be analyzed in solution, and for this purpose, nuclear magnetic resonance (NMR) spectroscopy is believed to be the most suitable tool.
In this approach, a quantum physical property known as “spin” is quantified in a powerful magnetic field for every atom in the molecule. The accurate precession frequencies of the spins are essentially a function of their surroundings, and it is precisely this frequency that enables scientists to quantitatively determine to which extent each amino acid is disordered or ordered in the protein.
In the latest study published on September 8th, 2020, Dr Rupashree Dass in association with Associate Professor Frans Mulder and Assistant Professor Jakob Toudahl Nielsen employed machine learning along with experimental NMR data for scores of proteins to construct a novel bioinformatics tool which they dubbed ODiNPred.
Through this bioinformatics program, other scientists can make the best possible predictions of which areas of their proteins are likely to be flexible and which are actually rigid. This data is useful for structural studies, and also for interpreting the regulation and biological role of intrinsically disordered proteins.
Dass, R., et al. (2020) ODiNPred: comprehensive prediction of protein order and disorder. Scientific Reports. doi.org/10.1038/s41598-020-71716-1.