The New York Times, September 11, 2014
“Microorganisms are the best chemists on the planet,” declared Michael A. Fischbach, a chemist himself at the University of California, San Francisco.
For evidence, Dr. Fischbach points to the many lifesaving drugs that microorganisms produce. In 1928, for example, Alexander Fleming discovered that mold wafting into his lab produced a bacteria-killing chemical that he dubbed penicillin.
Later generations of scientists found drugmaking microorganisms in more exotic locales.
In 1951, a missionary in Borneo named William Bouw shipped a box of jungle dirt to Edmund C. Kornfield, a chemist at Eli Lilly. In that soil, Dr. Kornfield discovered a species of bacteria that made a potent antibiotic, later named vancomycin.
Scientists today are still searching jungles, oceans and other corners of the world for microorganisms that make medicines. But in a new study published Thursday in the journal Cell, Dr. Fischbach and his colleagues suggest that we should also be looking inward.
Analyzing the bacteria that live in our bodies, the scientists identified genes for making over 3,000 previously unknown molecules that may prove to be useful drugs.
“Nobody had thought to look that close to home,” said Dr. Fischbach.
Finding these small molecules — known as natural products — has traditionally been a slow affair. Microbes typically make natural products in exquisitely tiny amounts, and they don’t rely on a single gene to do so. Instead, microbes need dozens of different proteins made by different genes to craft a natural product.
Dr. Fischbach and his colleagues set out five years ago to speed up the search. They wrote a software program that learns how to recognize the genes for natural products.
Those genes tend to sit together in a cluster in a microbe’s DNA, and they are very similar to one another. By shuffling them into different combinations, microbes can produce a staggering range of molecules.
To train the software, Dr. Fischbach and his colleagues introduced it to 732 gene clusters that are already known to make natural products. As the software examined cluster after cluster, it came to recognize distinctive patterns. Eventually the program got so good that it could accurately pinpoint new gene clusters in DNA sequences it had never encountered before.
The scientists wondered what would happen if they turned their well-educated computer loose on the microbes that live in our bodies.
They provided it with a vast genetic library created in an ongoing study called the Human Microbiome Project. The project scientists have collected microbial DNA from five different body sites on 242 healthy volunteers. From that genetic material, they were able to sequence the entire genomes of 2,340 different microbial species, most of which were new to science.
Searching those genomes, the computer spotted more than 14,000 gene clusters for natural products. Dr. Fischbach and his colleagues tossed out the gene clusters that were present in only a few people. They were left with 3,118 common ones.
Their study suggests that the human microbiome is a rich source of previously unknown natural products.
“That wasn’t where I expected to find interesting drug-producing genes,” said Dr. Fischbach. “I was really taken aback.”
To show the potential medical value of these genes, Dr. Fischbach and his colleagues picked out a single cluster to study more closely. It belongs a species of bacteria called Lactobacillus gasseri. They reared huge numbers of the bacteria in the laboratory in order to isolate a speck of one its products, which they dubbed lactocillin.
They found that its structure is similar to a recently discovered antibiotic called LFF571, which the drug company Novartis is now testing in clinical trials. When Dr. Fischbach and his colleagues exposed several species of bacteria to lactobacillin, the microbes died, suggesting that it might also be a good antibiotic.
The idea that our own bacteria are making potent antibiotics may seem strange. If the microbiome is churning out poison, how does it avoid killing itself?
Dr. Fischbach suspects that bacteria only use antibiotics sparingly against their competition. “You don’t wipe the slate clean of bacteria around you,” he said. “This could be something that a hundred thousand microbes use to guard the border of their colony.”
Shaun Lee, a microbiologist at University of Notre Dame who was not involved in the study, said that the fierce competition going on inside our bodies makes it a good place to look for antibiotics. “The human body is the Manhattan of microbial living — a great place to live with plenty of resources,” he said. “But real estate is at a premium.”
A lot of the natural products made by the microbiome may not be antibiotics. Previous studies have shown that some act as signals between microbes. Some even let microbes influence their human hosts.
Since these molecules carry out many jobs for microbes, it may be possible to turn them into drugs besides antibiotics. Consider statins, the drugs commonly prescribed to lower cholesterol. In 1972, the biochemist Akira Endo discovered the first statin in a mold that infects rice.
Of course, the history of research on natural products is also littered with failed molecules that never reached the market because they turned out to be ineffective or even dangerous. While Dr. Fischbach acknowledged that many of the microbiome’s natural products may also fail, he suggested that they might be particularly good molecules to turn into drugs.
For one thing, they’re made by microbes that have adapted to living inside of us for millions of years. Some of them may make natural products that latch precisely onto our own cells in medically useful ways.
The microbiome’s natural products may also be more likely to be safe. After all, microbes pump these molecules into our bodies every day without any apparent harm.
“This might be a privileged set of molecules,” said Dr. Fischbach.
Copyright 2014 The New York Times Company. Reprinted with permission.