Sometimes, it helps to overlook the big picture and zoom in on the minutiae.
That’s the strategy an Indian research team is hoping will accelerate the process of assigning functions to the growing number of hypothetical and so-called orphan proteins as new genomes get sequenced.
Besides speeding up function assignment, a peptide-library-based homology search tool (PLHost) developed at India’s Center for Biochemical Technology in New Delhi might also help pinpoint novel drug targets against nasty microbes.
While standard homology-based function assignment tools compare long stretches of peptides to determine the degree of homology between proteins, PLHost looks for tiny peptide fragments, just a few amino acids long, that occur in two or more organisms.
Given the 20 amino acids, the chance of a particular sequence of, say, eight amino acids occurring elsewhere works out to one in more than 25 billion. “We’re assuming that common amino acid sequences of this size turn up in different organisms not by sheer chance, but because they’re associated with specific protein functions,” says CBT scientist Debashish Dash.
He calls the common sequences “functional signatures” — peptide fragments of lengths varying from eight to 12 amino acids — associated with specific functions no matter which genome they’re in. CBT research shows that such fragments make up parts of proteins with unique functions.
“The mere presence of a functional signature in a hypothetical or an orphan protein gives us a clue about what that protein is likely to do,” says CBT director Samir Brahmachari, a molecular biophysicist. Brahmachari says the technique works because from billions of possible combinations, nature appears to have used a relatively small set of peptides for proteins.
The CBT team has used PLHost to identify a set of six common peptides, all part of an exclusively bacterial protein found in Mycobacterium tuberculosis, Helicobacter pylori, and Haemophilus influenzae, among other bacteria.
“This conserved protein is a kind of an Achilles’ heel, a potential target for novel drugs,” says Dash, “particularly because it is not found in eukaryotes.”
PLHost, the CBT researchers concede, won’t work all the time. It won’t be able to assign functions to unique proteins that lack functional signatures. “This is likely to happen when a protein has evolved independently in an organism and has no evolutionary heritage,” says Dash.
— Ganapati Mudur