Related Experiment Videos
Predicting protein function from structure: unique structural features of proteases.
E W Stawiski1, A E Baucom, S C Lohr
1Graduate Program in Molecular, Cellular, and Developmental Biology, Department of Biology, University of California, Santa Cruz, CA 95064, USA.
Summary
Proteases, enzymes that break down proteins, exhibit unique structural traits like tighter packing and more loops than other proteins. These features, identified by a neural network, aid in predicting protease function and classifying new protein structures.
Area of Science:
- Structural biology
- Biochemistry
- Bioinformatics
Background:
- Proteases are crucial enzymes involved in various biological processes.
- Understanding protease structure is key to their function and inhibition.
- Previous studies have not fully elucidated the unique structural characteristics of proteases compared to other proteins.
Purpose of the Study:
- To identify and characterize distinct structural features of proteases.
- To investigate the relationship between protease structure and function.
- To develop a predictive model for identifying proteases based on structural parameters.
Main Methods:
- Comparative structural analysis of proteases and non-protease proteins.
- Calculation of structural parameters including surface area, radius of gyration, and C(alpha) density.
- Development and training of a neural network model using these structural parameters.
- Validation of the model's accuracy in predicting protease function and identifying novel proteases.
Main Results:
- Proteases exhibit significantly smaller surface areas, smaller radii of gyration, and higher C(alpha) densities compared to other proteins.
- Proteases possess a higher proportion of loops and a lower proportion of alpha-helices.
- The trained neural network achieved over 86% accuracy in predicting protease function.
- The model successfully identified proteases with previously unrepresented folds.
Conclusions:
- Proteases are characterized by a distinct, tightly packed structure with a higher loop content, potentially evolved to prevent self-degradation (autolysis).
- Structural parameters can be effectively used to predict protease function and identify new protease families.
- This approach holds significant potential for classifying proteins generated by structural genomics initiatives.