The new technique may enable researchers to quickly find drug candidates that activate cellular receptors in desired ways.
For the study, researchers from The Scripps Research Institute (TSRI) in US used the technique to target a receptor linked to type 2 diabetes.
The GLP-1 receptor, as it is known, is expressed by insulin-producing 'beta cells' in the pancreas.
Several drugs that activate this receptor - drugs called GLP-1 receptor agonists - are already approved for treating type 2 diabetes.
The GLP-1 receptor belongs to a large class of receptors known as G protein-coupled receptors (GPCRs).
The researchers generated a library of candidate molecules - based on a known GLP-1 receptor agonist, Exendin-4, a small protein (peptide) originally found in the venom of Gila monster lizards; a synthetic version of this protein is now used as a type 2 diabetes medication.
They created about one million new peptides by randomly varying one end of Exendin-4 - the end that normally activates the G protein and beta arrestin pathways.
Using the autocrine selection system, the researchers rapidly screened these variant peptides and eventually isolated one, P5, that potently and selectively activated the GLP-1 receptor's G-protein pathway.
An initial test in healthy mice showed that P5 worked well at boosting glucose tolerance - at about one-hundredth the dose of Exendin-4 needed for the same effect.
The researchers found that while the peptide doesn't make mice fatter or heavier, it triggers the growth of new fat cells.
Exendin-4 induces a feeling of satiety, causing mice (and people) to modestly lower food intake and thus lose weight.
But the researchers found that P5 lacks this mechanism and appears to have no effect on appetite or weight.
