Researchers have used 'directed evolution'—a series of laboratory techniques for modifying DNA that mimics natural evolution—to predict genetic mutations that lead to antibiotic resistance.

Raymond Stevens, of the Scripps Research Institute in La Jolla, California, and colleagues tested directed evolution on the TEM-1 b-lactamase enzyme, which destroys antibiotic drugs like penicillins, cephalosporins, clavams, cephamycins, and carbapenems.

Using three different evolution strategies, Stevens' team generated 11 mutant forms of b-lactamase, of which 8 have already been found in nature. The team also generated an enzyme with three mutations—E104K, M182T, G238S. These same three mutations are present in the TEM-52 clinical isolate of the E. coli bacterium. TEM-52 E. coli is 500 times more resistant to the antibiotic cefotaxime than the wild-type form of the bacteria.

After determining the crystal structure of the TEM-52 b-lactamase, Stevens' team examined the position of the three mutations and found that the E104K and G238S created a wider, more stable site to capture and destroy the cefotaxime. The third mutation, M182T, which lies in the hinge region linking the two halves of the enzyme, seems to stabilize the newly altered, wider binding site.

Drugs designed to interfere with the hinge region could destabilize the whole enzyme and prevent it from functioning. Thus, the hinge region may provide a target for a new class of antibiotics, write the authors in the current issue of Nature Structural Biology.

Stevens' team also identified three mutations that if combined with TEM-52 would increase antibiotic resistance 32,000-fold. "This ensemble of six mutations has not yet been isolated in nature, but an alarming observation is that individual mutations at five of the six positions have been observed in different clinical isolates," the authors write.

Only mutations at 13 positions out of about 290 amino acids in the b-lactamase enzyme seem to confer antibiotic resistance.