Researchers use new approach to predict protein function

July 11, 2007
Enolase Enzyme
Homology modeled complex (in cyan) of the enolase enzyme, BC0371, with its substrate, N-succinyl-L-arginine, matches very closely the experimentally determined structure (in yellow) via X-ray crystallography. Credit: Image courtesy of John A. Gerlt and Nature Chemical Biology

In a paper published online this month in the journal Nature Chemical Biology, researchers report that they have developed a way to determine the function of some of the hundreds of thousands of proteins for which amino acid sequence data are available, but whose structure and function remain unknown.

The research team, led by University of Illinois biochemistry professor John A. Gerlt, is the first to use a computational approach to accurately predict a protein’s function from its amino acid sequence. Their “in silico” (computer-aided) predictions were validated in the laboratory by means of enzyme assays and X-ray crystallography.

The new approach involved searching databases of known proteins for those with amino acid sequences that had the greatest homology to the unknown proteins. The researchers then used the three-dimensional structures of the most closely matched known proteins in their analyses of protein function.

Using the structural data obtained from this homology modeling, the team performed computerized docking experiments to quickly evaluate whether the unknown proteins were likely to bind to any of a vast library of potential target molecules, or substrates. Determining which substrate binds to a given protein is vital to understanding the protein’s function.

“This study describes an integrated approach using experimental techniques, computational techniques and X-ray crystallography for predicting the function of a protein of previously unknown function,” Gerlt said.

These methods will speed the task of identifying the biological roles of some of the hundreds of thousands of proteins whose functions have not yet been discovered.

“Rather than trying to do (laboratory) experiments on 30,000 compounds to determine if they are substrates, with this approach you might do experiments on 10,” Gerlt said.

The study involved a family of proteins within the large and diverse enolase superfamily. Enolases are enzymes that catalyze the breakdown of glucose and related compounds into other molecules as needed for metabolism.

The enzymes within the enolase superfamily utilize similar reaction mechanisms to one another but catalyze different reactions, complicating the task of discovering their function. There are more than 3,000 proteins in the enolase superfamily, and a majority of them have not yet been fully – or accurately – characterized. (The new study also revealed that one family of enolase proteins had been misclassified.)

Gerlt and his colleagues expect that the computational approach they pioneered will help scientists more efficiently tackle the problem of understanding these – and other – unknown proteins.

“There are 4 1/2 million protein sequences in the sequence databanks, and maybe the functions are known for, can be assigned to, half of those with some reliability,” Gerlt said. “That tells you that there is a lot of biology to be discovered.”

Source: University of Illinois at Urbana-Champaign

Explore further: Bacterial protein can help convert stem cells into neurons

Related Stories

Bacterial protein can help convert stem cells into neurons

November 19, 2015

As the recipe book for turning stem cells into other types of cells keeps growing larger, the search for the perfect, therapeutically relevant blend of differentiation factors is revealing some interesting biology. A study ...

Mitochondria on guard of human life

November 18, 2015

A group of researchers from Lomonosov Moscow State University in collaboration with Russian Science Foundation has developed a unique method for the selective study of electron transport chain in living mitochondria by using ...

Researchers develop the first platform for DNA simulations

November 17, 2015

Molecular dynamics is a technique that allows the simulation of DNA movement, including its folding into double, triple or quadruple strands, and even its interaction with proteins and drugs. Molecular dynamics is used to ...

Recommended for you

A common mechanism for human and bird sound production

November 27, 2015

When birds and humans sing it sounds completely different, but now new research reported in the journal Nature Communications shows that the very same physical mechanisms are at play when a bird sings and a human speaks.

Moonlighting molecules: Finding new uses for old enzymes

November 27, 2015

A collaboration between the University of Cambridge and MedImmune, the global biologics research and development arm of AstraZeneca, has led researchers to identify a potentially significant new application for a well-known ...


Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.