On a balmy December day in Jupiter, Florida, chemist Ben Shen led the way into a standalone, shed–sized cold room that sits just outside one of the three buildings at the Scripps Research Institute’s East Coast campus. Six-foot-tall racks hung from metal rails on wheels, each with hundreds of cubby holes filled with clusters of glass ampules containing freeze-dried microbes. Along the back wall of the room were tiny drawers with more samples. And this was just half of the collection. Back inside the closest building, more than 30 freezers held thousands more samples of bacteria and fungi.
In total, Shen now oversees a biobank of 217,352 microbial strains, more than 210,000 of which came to Scripps late last year from the drug company Pfizer, which had established and maintained the historic collection since the early 20th century. Microbes produce compounds that have already yielded or inspired hundreds of drugs, such as penicillin, streptomycin, and tetracycline. By continuing to study the biobank of microbes, researchers can look for new therapeutics, agrochemicals, and other useful compounds.
The challenge is daunting, but the opportunity is immense.—Ben Shen, Scripps Research Institute
Shen launched the Natural Product Library Initiative at Scripps in 2011 and already had a collection of about 6,000 microbes that he was mining for natural products. When he heard Pfizer was seeking to unload its collection, he was immediately interested in the resource. Each microbe is estimated to encode an average of 30 natural products; Pfizer’s collection likely held millions of novel molecules waiting to be discovered, and so Shen and his colleagues put together a proposal to take it on.
“Pfizer recognized that the collection’s value might be increased if it were more readily available to researchers outside of the company, and after a competitive process ultimately chose to license the collection to Scripps—a top-ranked international institution that is committed to preserving microbial biodiversity for use in biomedical research,” Pfizer spokesperson Amy Rose tells The Scientist in an email.
In December 2018, Shen and Scripps Executive Vice President Doug Bingham traveled to the Pfizer facilities in Groton, Connecticut, “to figure out what exactly we’d won,” Bingham says. About half of the samples were freeze dried and refrigerated; the other half were frozen at -140 °C. The library was annotated in numerous and disordered notebooks and databases. “It’s an archaic collection; there’s lot of really small labels with handwritten numbers,” says Bingham. “We had to get a handle on how those worked. . . . Getting all the things down here and then not being able to find the strain you wanted was pointless.”
Bingham and his colleagues also decided they wanted to remove any strains that were pathogenic before the collection was loaded into trucks and hauled to Florida. Bingham estimates that the Scripps team identified some 1,000 strains, such as Brucella bacteria, that were pulled from the collection and destroyed. Meanwhile, Bingham oversaw preparations on the Scripps campus. In May, a standalone cold room was installed, and space was cleared for the nearly two dozen freezers that contained the frozen part of the collection. Then, in two separate trips, trucks hauled the collection more than 1,300 miles down I-95 from Groton to Jupiter.
Shen is already studying some of the microbes in the collection. His group has attempted to culture more than 150 strains of Actinobacteria, having success with more than 95 percent of them in one medium. “By varying media, we should be able to grow them all. . . . We should be able to revive most of the strains in the collection,” Shen tells The Scientist in an email. Last month, Shen secured a five-year NIH grant totaling nearly $4 million to develop tools to look for natural products encoded by Actinobacteria. The collection has more than 62,000 Actinobacteria specimens.
Shen’s is not the only group mining the Pfizer strains for natural products. A little over a year before Pfizer got out of the natural products space, the company struck a deal with Adapsyn Bioscience, which can apply its metabolomics approach to the crude extracts that had been created and stored after growing the bacteria in culture. “We can start just with the crude extracts and use our metabolomics pipeline and say there’s something novel,” says Andy Haigh, president and COO of Adapsyn Bioscience. If something of interest turned up in that analysis, the researchers can explore it further by fermenting the bacteria, isolating the natural product, and subjecting it to various bioassays to probe its activity.
This approach is faster and cheaper than genome sequencing, Haigh says, and it more directly gets to the products of interest. “Genomics [studies] give you a good sense of the potential of the bug, but you really need the metabolomics to go and find the actual compound that bug is producing,” he says. He notes, however, that both metabolomics and genomics are essential to mining natural products. The genome sequence, for example, can provide clues about the compounds’ structure and novelty. “Without both sides of the equation, it can be really difficult to find these things,” Haigh says. (Pfizer’s collaboration with Adapsyn was not affected when it gave the microbial collection to Scripps, Rose says.)
It’s an interesting approach, to leverage microbial diversity, but it’s a complicated and high-risk approach as well.—Sylvain Brisse, Institut Pasteur
While the Scripps collection is now one of the world’s largest microbial biobanks, it is by no means the only one. The search for useful products within similar collections is widespread, says Sylvain Brisse, director of the Biological Resource Center of the Institut Pasteur. The institute maintains one of the world’s oldest microbial collections, which holds more 50,000 strains. “There is a trend to develop projects to use the diversity that we have in these collections to look for natural products,” Brisse says.
Like Shen’s group, Dominique Clermont, head of the Institut Pasteur’s bacterial collection, oversees a program to sequence the genomes of the microbes in the collection and mine those sequences for biosynthesis gene clusters, groups of genes involved in the production of natural products. Muriel Gugger, head of Institut Pasteur’s cyanobacteria collection, is doing the same. In the genomes of nearly 90 cyanobacterial strains, for example, her team identified 452 gene clusters for two particular natural product production pathways, pointing to potential novel products to be found in those strains.
“It’s not enough to have a big collection of microorganisms,” says Brisse. “There is all kinds of other expertise and needs down the line to get to the identification of active products. . . . It’s an interesting approach, to leverage microbial diversity, but it’s a complicated and high-risk approach as well.” Still, the potential payoff is great, Gugger notes. “Researchers have . . . to work together to combine their strengths to find what are the natural products and how they can be used for human health.”
Shen, who is helping to organize a meeting on natural products and drug discovery in the genomics era slated for January 12–16 in San Diego, agrees. “The challenge is daunting, but the opportunity is immense.”
Jef Akst is managing editor of The Scientist. Email her at email@example.com.