The billions of micro-organisms that naturally occur in soil have played an enormous part in the development of modern medicine, predominantly in the years following the Second World War, when soil-based actinomycetes were used as the basis for a host of antibiotics, including erythromycin, tetracycline and streptomycin. They also represent a major source, either directly or as a biological inspiration, of vital drugs such as anti-cancer agents, immunosuppressants and HDAC inhibitors.
"Plants, animals – natural products in general have been a tremendous source [of clinically useful molecules], but bacteria most of all, and if you want to ask where are there more bacteria, it’s the soil," says Dr Sean Brady, head of Rockefeller University’s Laboratory of Genetically Encoded Small Molecules. But for all the medical discoveries made so far on the back of soil chemistries, for experts like Brady, this progress is little more than a drop in the ocean of clinical potential.
Brady’s lab has been seeking to survey diverse soil samples to build a better idea of the best geographic regions to find the most promising sources of such molecules, as well as developing the technologies necessary to clone DNA from soil samples and activate the genes to produce new molecules in a laboratory setting. On top of this drug discovery effort, the lab is also working with citizen scientists and other institutions to help build up a map of new molecule source sites around the world. With soil-based molecules already yielding new antibiotics that could prove effective against resistant bacteria, we spoke to Brady to get a better picture of the immense clinical potential yet to be tapped in the dirt beneath our feet.
Chris Lo: What remaining potential does soil have as a source of medically useful molecules?
SB: It’s incredible, what we’ve learned over the past ten years. There are two main things. One is that we haven’t cultured the vast majority of microbes. Depending on whose data you believe, it’s somewhere between 99.9% and maybe 99% [that are uncultured]. Maybe, if you want to be very conservative, we’ve cultured 10% of the bacteria, although I think most people would disagree with that.
But even among the bacteria that we do grow, what we realise now is we were leaving behind most of the molecules they could make. We didn’t know how to turn them on; we didn’t know how to turn on the genes that make those molecules. So either we haven’t brought the bugs in, or for the bugs we’ve already brought into the lab, we’ve missed most of the molecules that they make. That suggests that everything we’ve found is just the tip of the iceberg of what’s possible out there.
CL: With all of these potentially useful uncultured soil bacteria, why haven’t we seen a flood of new molecule discoveries?
This year’s Annual Drug Discovery for Neurodegeneration Conference set out to encourage innovation.
SB: There’s a whole slew of things. One is we still can’t culture most of the bugs. That’s what our work is – developing technologies that allow you new ways of getting at those molecules. We don’t grow bugs; we just extract DNA from the environment, and that gives all the DNA from all the bugs. And then you have to develop a whole new set of technologies for problems like how do you screen that DNA to find genes that make molecules? How do you turn those genes on in bugs they weren’t originally in? Because you put them in foreign bugs now and make molecules.
It’s this slow realisation that happened maybe 20 years ago now that there were a lot of bugs out there that we hadn’t seen, and then this realisation that maybe we should rethink how we find molecules, if there are all these bugs. And then we’ve had to develop a set of technologies that would allow us new ways of discovering those molecules. If it’s not just as simple as growing a bug and putting it in a flask and saying ‘Oh look, I’ve found a new antibiotic’; you’ve got to find a whole new set of technologies.
What I would argue is major advances have occurred, like sequencing, which is a very important advance. We are also getting better at heterologous expression. So those tools are now coming together into a unified solution for the problem. We’re still not there yet – we don’t have a perfect unified solution. But one can now see molecules coming out, and at least imagine sometime in the future when large numbers of molecules will come out of those efforts.
CL: Could you describe the work that you’re doing at Rockefeller’s Laboratory of Genetically Encoded Small Molecules in the area of soil and drug discovery?
SB: We have lots of projects, but they have two main themes. One is just to survey environments by sequencing, and ask, ‘Where are environments that might be rich in new chemistries? What are the trends we see?’ The take-home lesson is we really haven’t seen most of what’s out there. We can tell that by sequencing, and it can guide us to sites where we can do the other part of the group, which is, ‘How do we clone this DNA, put it into a bug in the lab, and get it to make these new molecules?’
So there’s really two parts of the group that are related to this. One is to survey environments to ask what’s out there, and can we get some intelligent map that would tell us where to find molecules in the future? And then once we’ve found those sites that we think are rich in new chemistries, can we clone the DNA, can we turn that DNA on, and begin to get the molecules that those genes encode for?
CL: Are there specific drug development areas that could be particularly affected by new molecule discoveries in soil?
SB: I think it’s anyone’s guess. What we do know is what we can do, and we have tools now to take lead structures, or things that exist in the clinic, and ask whether we can do better than them. So if they’re molecules that we can play off of, we now realise that we’ve only scratched the surface of those families. So can you take daptomycin [a lipopeptide antibiotic] and do better? I would argue that unquestionably we can do that. Even [compounds] that were found in the 70s and left behind because they weren’t quite perfect; now we can go back and maybe revitalise those.
What we’re still trying to develop as a field is the de novo technology. How do you scale this so you can find the thousands and thousands of new compounds you need to sort through and find one brand new one that’s really, really good? I think what the field now needs to step forward with is how to scale that so you can say, ‘I never knew anything about this family of compounds, I want to find a brand new family of compounds and see if I can do even better with that family.’ So the de novo discovery is still something that I think is challenging, and that’s great. It’s a science problem to solve.
CL: Do you think microbes from soil have potential as agents to overcome drug-resistant bacteria?
SB: The point is the chemistry pool is extraordinarily large, and we’re likely to find things that do any of a number of things. Like, particularly, kill resistant bugs. If you think about it, we think of clinical resistance as something unique to humans, right? ‘Oh, we have a resistant bug in the hospital.’ Well, most of those bugs, those resistant genes, exist out in the environment. So bugs are fighting that exact same problem. If we can go out and figure out all the antibiotics, well, very likely nature’s come across some that get around those problems that doctors are facing.
I think that’s what we’re hopefully going to find as we go forward and the entire field scales; the discovery process is that lots of bugs maybe have already addressed this resistance issue. And if we can find those antibiotics, maybe we can solve some of these issues we have with resistance.
CL: Have you identified particular regions as high-potential areas for drug discovery?
SB: It turns out our data would suggest that arid regions, dry regions, the deserts of the American South West – we collaborate with lots of groups now, and one group in Brazil has found some dry lands there that look really high in innate diversity. They have lots of bacteria that make lots of new molecules. So the trend seems to be arid environments.
It’s been nearly a year since the formation of the UK’s ‘war cabinet’ to tackle antimicrobial resistance.
And then you have unique environments where for whatever reason, we don’t know why, it’s rich in molecule X, Y or Z of some clinical relevance. A hot spring in New Mexico we found was rich in some specific proteasome inhibitors. Who knows why, but that’s what it is.
There are some consistent trends, but there’s this idea that you just don’t know what you’re going to see and you have to scale the map. There’s an analogy to Google – it really became tremendously valuable when Google mapped the world. You didn’t even know what was going to come out of that; whether it would inform archaeology or city planning. The same thing’s going to be true about biochemistry. As we map the world, hopefully some day, for its biochemistry, you’re going to find hot spots for X, Y or Z, or spots that you didn’t even know had lots of chemistry in them, which we wouldn’t have predicted.
CL: The dataset you’re building up also involves a citizen science component through the Drugs from Dirt project – how did that come about?
SB: Well, it’s expensive to fly around and collect samples. There are lots of things that are citizen efforts. So we figured an easy thing for people to do was collect soil, and that people might be interested.
It became somewhat overwhelming, so most of our effort now is within the United States, and we collaborate with other people who are doing things in other countries. We just got such a response that we’re mostly the United States, trying to get a really nice map of the US, and we’ve got lots of people donating samples for that. The global citizen science will come out of a series of collaborations we’re developing with academics in other regions of the world.[For the citizen science mapping project] we’re not actually collecting enough soil to harvest any molecules; we don’t want to step on people’s biodiversity. Each country owns their own biodiversity; they should be able to control it. So the mapping project is really about a basic science question – can we inform people what’s out there?
CL: In the next 10 to 20 years, how much progress would you hope to have seen in this mapping effort?
SB: I hope that we have a map. I would say in ten or 20 years it’s not going to be like a Google map, but a map where one can go in and say, ‘This is where I go to find this, I go and I have tools where I can specifically pull out that cluster, turn it on and get that molecule.’ That’s what you’d like to be able to do – have this thing where it’s all computationally done. Even if it’s new and I’m looking for sequences that don’t look like anything I’ve seen before, I know exactly where to go, how to pull them out and how to turn on the molecules. Or, say I just discovered a brand-new erythromycin, how do I go and find a hundred derivatives of that? And we’d have all the tools where that’s essentially automated. So it’s an automated process of harvesting, in a directed fashion, chemistry from the environment.
That may be 30 years from now; it may be ten years from now. The field is really growing quite quickly. We’ll see.