Researchers have discovered that commensal bacteria of the human microbiome synthesize thousands of potentially drug-like small molecules, including at least one class, the thiopeptides, which is in clinical trials.

The work appeared in the Sept. 11, 2014, advance online edition of Cell.

To date, most studies of the microbiome have been focused on taking a census of sorts, documenting what commensals are present in the human body, first author Mohamed Donia told BioWorld Today. Donia, who is at the University of California at San Francisco, and his colleagues were interested "not just in which bugs are there . . . but in what they are doing." In particular, they decided to glean answers to that question by looking at what sorts of small molecules commensal microbes were producing.

In a video interview posted to Cell's website with the paper, senior author Michael Fischbach said those molecules, collectively, are "the language of interaction between the microbiota and its host," as well as between different bacterial species of the microbiome, and the pathogenic invaders that they sometimes have to contend with.

In their work, Donia explained, the authors used bioinformatics methods to identify so-called biosynthetic gene clusters – groups of several genes that cooperate to make a molecule. The team cast a wide net in these studies, looking at the genomes of about 2,400 human-associated bacteria – "every single genome that we could get our hands on," Donia said.

The method succeeded in identifying 44,000 gene clusters in those genomes, far beyond what Donia and his colleagues had expected.

The team next compared their clusters to metagenomic data from the human microbiome project to understand how many of them were present in the healthy population.

Comparing their data to metagenome samples from 752 healthy U.S. residents, they found slightly more than 3,000 of their clusters in at least one sample, though the prevalence varied widely – some clusters were present in 90 percent of the samples.

Finally, the team looked in greater depth at some of the most common clusters – and discovered that some of them made antibiotics that could garner the respect of a medicinal chemist. Perhaps the most striking was the widespread presence of thiopeptides.

Thiopeptides are a novel class of antibiotics that have been isolated from both the soil and marine organisms; Novartis AG's LFF-571, which is in phase II trials, is a thiopeptide.

In their work, Donia and his colleagues solved the structure of one thiopeptide, which they named lactocillin. That particular molecule, he said, had properties that drug developers would love to give their molecules, if only they could figure out how – it is active against pathogenic bacteria, but other vaginal commensals were by and large resistant to it.

"That's the problem with all the antibiotics we take – they destroy the commensals as well," Donia pointed out. But lactocillin, which is made by the vaginal commensal Lactobacillus gasseri, "was very potent against pathogens, but almost inactive against other vaginal commensals."

The team plans to look at biological functions of other clusters that look like they might have practical relevance, and Donia noted that other researchers could do so as well, as the team published its dataset with the paper. "There's plenty there," Donia said, "and it's a resource for the scientific community."

Fischbach said that collectively, the molecules made by biosynthetic gene clusters on commensal bacteria make up "a pool of molecules that could be quite informative to screen through, to view as leads compounds." And "since you already find them on and inside of humans, perhaps they have a higher likelihood of having an interesting target and maybe a lower likelihood of having an unforeseen toxicity, the kind of problem that could derail a drug development program."

Fischbach noted that "for a very long time, people in the natural products community have gone . . . to the corners of the earth to find exotic microorganisms that will make new molecules that the world has never seen before." But in doing so, they "had been ignoring one of the most interesting ecological sites that's closest to home, which is the human microbiota."

"You don't need to look far," Donia added. "It's just right there."