Kazuhiko Kurosawa was running out of variables. For eight months he had made hundreds of cultures of Rhodococcus fascians - manipulating pH, temperature, salt concentration, media type, oxygen levels, even degree of agitation - each time attempting to provoke the bacteria into transcribing a set of genes he knew lay dormant in its genome. But the soil-dwelling bacteria remained recalcitrant.
Anthony Sinskey's lab at Massachusetts Institute of Technology, which Kurosawa joined in 2003, first became interested in Rhodococcus while in collaboration with Merck. The company hoped to use the bacteria to more efficiently make a precursor for one of its HIV protease inhibitors, Crixivan. While analyzing the Rhodococcus genome, the MIT researchers were surprised to find close to 100 genes for nonribosomal peptides, and 30 clusters of genes for polyketides, two major classes of antibiotics. But unlike most other soil-dwelling bacteria, Rhodococcus was not known to produce an antibiotic. Finding the genes "was like Christmas morning," says Philip Lessard, one of the investigators. Unfortunately, he adds, there appeared to be "some assembly required."
However, Sinskey's funding was wrapped up in the lab's main projects. So in 2004 Kurosawa embarked on an "under-the-table study" to provoke one species of the bacteria, Rhodococcus fascians, into using those genes, he recounts in an E-mail. (Despite the need for new antibiotics, Lessard says funding for new antibiotics is often hard to come by because of the availability of cheap and still-effective front line antibiotics like penicillin and ampicillin.) Lessard recalls every roller in the lab constantly drumming with Kurosawa's cultures. After the two researchers ran out of variables to test, they took a hint from nature and decided to introduce R. fascians to an environmental hazard it had yet to confront in the lab - other bacteria.
In the wild, R. fascians lives in soil populated by thousands of microbes. The researchers planned a "kindergarten" version of the scenario, Lessard says, pitting R. fascians against just one other bacterium - a strain of Streptomyces padanus, an aggressive antibiotic producer that Kurosawa isolated from the soil in a flowerpot by his lab bench.
Initially, however, they overshot the mark. "Streptomyces really did kill everything," says Lessard. But one day Kurosawa approached Lessard holding a plate on which the Streptomyces had been wiped out. Lessard, who now works for Agrivida, an agricultural biotech (and retains no patents on the antibiotic), was skeptical - the plate was probably contaminated, he reasoned. But Kurosawa didn't think so. By sequencing the genes encoding ribosomal RNA of the bacteria left on the plate, he proved it was the original strain of R. fascians. Then, Kurosawa used high performance liquid chromatography (HPLC) to analyze the secreted substance, and became "full of confidence" that the bacteria were producing an antibiotic, he recalls. The eager biologist reported his results at a lab meeting in April of 2005. "Prof. Sinskey said, 'No one believes that. Kazu, you should isolate the antibiotic and elucidate the molecule.'"
It took a year of "painfully tedious" work to purify the secreted antibiotic, recalls Kurosawa, still based at MIT (where this author studied journalism). He relied on the other labs to do NMR structural studies of the molecule in their spare time. The molecule turned out to be an aminoglycoside (J Am Chem Soc, 130:1126-7, 2008). When tested for antimicrobial activity, it proved particularly effective against Heliobacter pylori, a bacteria that causes stomach ulcers in humans.
But there was a mystery. Pulse field gel electrophoresis revealed a strange chunk of DNA in the R. fascians genome, about 150 genes that did not exist in the parent strain. Analysis of a tiny extract showed it belonged to Streptomyces. And "in every case we've looked," adds Lessard, "the loss of the ability to make the antibiotic correlates with the loss of this DNA element." In the end, R. fascians may never have actually used the antibiotic-coding genes that prompted the research in the first place. But do the Streptomyces genes encode the antibiotic, or simply provoke R. fascians into making it? The project "was a success followed by ten unknowns," says Lessard.