Recognizing the Next CRISPR-Level Tech for Biology

Jennifer Doudna won the 2020 Nobel Prize in Chemistry for her co-discovery of CRISPR/Cas9, a versatile genome editing platform. In the decade since its discovery the toolbox of CRISPR technologies has exploded, acting like rocket fuel for curiosity-driven science. It is also increasingly a foundational technology for many biotech companies.

In this conversation, Doudna chats with a16z general partner Vijay Pande. Previously, he was a Professor at Stanford University, where he directed the Biophysics department. During his time there he also founded the Folding@Home Project and Globavir Biosciences. 

Pande and Doudna grapple with questions facing scientists at this inflection point. How do you recognize a discovery that will open further opportunities to engineer biology? What will happen as CRISPR tools mature? What does a biologically engineered future look like, and what responsibility do scientists have to ensure these tools are used responsibly?

Along the way, Doudna touches on what she’s struggling with, what surprised her, and what may never be engineerable.

Note: This interview was originally published as an episode of Bio Eats World. The transcript has been lightly edited for clarity. You can listen to the full episode here.

VIJAY PANDE: There is so much excitement for our ability to engineer biology, and to take what we’ve learned and create new therapies, new things, and synthetic biology. The product and company side is really blossoming. At the same time, if we didn’t have that basic research, we probably wouldn’t be where we are now. Given the arc of what you’ve seen, where you stand on that, how should we be thinking about that balance?

JENNIFER DOUDNA: It’s a pleasure to be here. 

I think you bring up a great point. And that is, how do we get the right balance between fundamental science and engineering or focused applied science? You know, I’ve always done what you would call curiosity-driven science for the most part. And increasingly, I find myself faced with problems or challenges that we’re working on that are right at the edge of that. You sort of ask yourself, do we know enough that this is now an engineering problem, or is there still really important, fundamental work that needs to happen that could be very enabling, but maybe not for a few years?

He was just kind of shocked at the way we do science. His word for it was artisanal. 

VIJAY: Yeah. You know, it’s a tricky question. And I think part of it is also just the timescales. When I think about basic research, I was thinking of the discovery and invention of CRISPR, almost to be akin to that of the transistor, where it’s really only now–50 years later–when you can pack 10 billion, 50 billion transistors on a chip, and you can do these things that are mind-blowing. So you can’t expect to get immediate returns, even 10-year returns out of basic work. 

On the other hand, it is these major discoveries like CRISPR, like the transistor, that really can make these huge shifts. So there naturally has to be a balance. So much of biology is discovery. There’s just so much to learn, so much to discover, compared to, let’s say, in physics, where you can do so much more theoretically and drive it, or even compared to engineering where you can sin principle grind things out more. 

What would the industrialization of bio look like?

VIJAY: I am really curious about the ways we can shift even just the process of discovery from an art to an industrialized process. Can we industrialize discovery? Where are we now with that and where do you think we can go?

JENNIFER: Yeah, it’s a great question. It reminded me of at one point, I had a visitor from Google who came up to the lab at Berkeley. He wanted to have a tour of a working experimental biology lab. And he was just kind of shocked at the way we do science. His word for it was artisanal. He said, “This looks artisanal to me.” And he said, “I think you guys could do a lot to automate your work and this and that.” 

But in the end, it hasn’t been that easy, really, to automate or industrialize the work that we’re doing. Now, certainly, in some ways that’s happened just by the power of computing, and having more programmers and people that think computationally involved in biology has been a huge plus. That’s really had a very positive impact. But there’s something about biology that there are stochastic things that you just can’t yet really predict.

Now, every now and then, something happens that makes me think, “Huh, maybe we’re on the cusp of a real change.” For example, the work that was recently announced about being able to computationally predict protein folds accurately. That really seems like a really interesting advance that could revolutionize that field, right? And so you could imagine that that kind of thing could extend in other directions too. Maybe eventually it becomes much easier to assign function to genes because we’ll have enough predictive information that if you feed that all into the right algorithm, you get a very limited number of possibilities that come out, and that makes your experimental work a lot easier or more robust.

VIJAY: One of the things here is that just the aspects of automation is pretty hardcore. You get like a big robot like a Tecan or something like that. It’s pretty expensive. And that’s only for a specific kind of high throughput workflow. Whereas lots of biology is N equals five or maybe a lot of replicates. But not 5,000 or 5 million. 

I’m curious whether, just like the innovation that we’ve seen in kits over the last 20, 25 years, whether a kit could be both the reagents, and the software to drive a little desktop robot, like Opentrons. That desktop robot maybe is the equivalent of a PC here, in that it can be fast and nimble and do things, and because it comes in the kit, with the reagents and with the software to drive it, then people will build upon kits, kits upon kits, and so on. And you finally get to something that’s useful. 

Because I think maybe the point you’re making is that if you had a big robot, that wouldn’t be faster if you have to do the small end, right? It would probably be more work than pipetting by hand. Do you think that’s getting closer in the right direction?

I thought, how can I really defend this as something that has anything to do with human health?

JENNIFER: I’m trying to think about where the real bottlenecks are just in my own research world. It was really two and one can’t be solved with a robot, at least until we get robots that are thinking on their own, probably, because that’s really at the level of the gut feeling. There’s lots and lots of ideas out there, but only some of them are good. And so, how do you figure out what you’re going to spend time going after. So, there’s still that problem. 

But once you’re onto a good idea, then just getting through the experiments, I think that’s where having nimble, small and not super expensive robots in the lab could be really enabling. I have to say that, you know, we’ve worked with a number of [robots]… And yeah, as you said, it’s typically a big box of a thing that is designed to do one type of task. At least in my experience, they’re often very fussy.

So, you have to spend quite a bit of time just getting the whole thing working with whatever you’re trying to do, and maybe even training a person or hiring a person that’s going to be responsible for running that robot. And then you might run it for some months, and then decide, “Oh, now I wanna change my experiment, do some different thing but now that robot’s no good for that,” right? I think if there were a way to have small robots that were easily adaptable to different tasks, which could do them very accurately… I guess it could be the case that you had individual small, not too expensive robots that were good at a certain type of task, and you have a different robot for different types of tests, that could work. I think that could be really enabling.

VIJAY: Well, and I think this is where the industrialization [applies]. If you’re building a shoe factory, you’re gonna make shoes. And you’ll make maybe slightly different shoes, but you’re not gonna make teddy bears or something like that. Whereas, you have to be super nimble, and you may be doing a radically different experiment the next week, or next day, or something like that. And I think it is that generalizability that we need. But, you know, maybe the most exciting point is this shift. I see so many people making the shift from having done basic curiosity-driven research towards applied.

JENNIFER: That really, in many ways, has underscored a lot of the things that I’ve done over the years in my own lab, beginning with all the way back to when I started my faculty career looking at the structures of ribosomes. You know, that really took us into the field, eventually, of RNA interference and RNA molecules in viruses that are part of the machinery for controlling translation in infected cells. And then from there to CRISPR. 

These were always projects that were, in my lab, framed from the perspective of: how does this work? You know, how does this work from a molecular perspective, whether it’s the actual structures of the underlying molecules or their enzymatic or biochemical behaviors? That’s how we approach CRISPR as well. It was really, for us, in the beginning that this looks like an adaptive immune system in bacteria that is RNA directed in some way. So how does that work? It was a project that very much started with that really fundamental question.

On making the leap from biology to tool

VIJAY: There’s this seemingly big gap between studying an adaptive immune system of bacteria to the ability to engineer genomes, and developing new classes of therapeutics for things that were previously undruggable. How did you start to see the sort of connecting the dots?

JENNIFER: Quite frankly, when we began that work now almost a dozen years ago, I certainly didn’t expect it to go the way it did. In fact, I was a little bit reticent about working on it in the beginning, because I was receiving funding from the NIH and from Howard Hughes Medical Institute. I thought, how can I really defend this as something that has anything to do with human health? And now, as we all know, it has everything to do with human health. It started with those very fundamental questions of how does this immune system work? And then a very specific question about one particular protein, Cas9, that was clearly implicated as a central player in the CRISPR immune systems of some bacteria.

And then it was pretty obvious from those biochemical data that this enzyme, which works as an RNA-guided cleaver of DNA, can be directed to cleave a desired DNA sequence. That concept converged so well with all the other work that was going on in genome editing because people were looking for ways to cut DNA in cells in a way that made a double-stranded break that would induce the cell to repair the DNA by introducing a change in the sequence. So, here we had this cleaver that was programmable, so you could tell it where to go and make a cut. And that just converged beautifully with all of the work on genome engineering using earlier technologies. It’s just that this is a much easier way to do it.

Nature-made for engineering

VIJAY: One of the fun things about things that came out of natural selection is that it seems like [CRISPR systems] were evolved to be evolvable. I think about chaperones and things helping proteins do things. One of the hallmarks of bringing in an engineering mentality or approaches is that you can have iterative improvement. Things can get a little better year over year. And often that improvement is compounding almost like compounding interest, where you could sense that there was a shift from ‘this is the time to be curious’ to ‘this is the time to engineer.’

JENNIFER: Well, one of the things that’s so exciting about CRISPR, from an engineering perspective, is that it’s turned out to be a system that’s highly amenable to modification. I think you make a really good point that nature kind of sets things up that way anyway. We see that in natural CRISPR biology because there’s a large collection of these enzymes that have evolved in different bacteria, and they can look really quite different from each other, and have a range of activities. So, clearly, nature is doing this tweaking and fine-tuning these proteins for their native environment. In my mind, I have this vision of this whole toolbox that’s all built around this RNA-guided mechanism, that adds all kinds of interesting different chemical activities that allow these types of manipulation and genomes.

They all look very interesting. So, we struggle to figure out where we want to focus our efforts and whether it’s worth working on the next CRISPR system versus casting our net in a different direction.

In 2013, there was a cascade of publications that came out that year from different groups showing that you could use Cas9 inhuman cells, you can use it to engineer zebrafish. There were lots of really interesting proof of principle discoveries that were put forward using the CRISPR/Cas9 system that made it clear that this was going to be a transformative tool for doing all kinds of science. Not only fundamental research–the kinds of things that were enabled by being able to probe the function of genes, make knockouts in targeted ways and cells–but frankly, also to use it in a very applied way. Namely to make, for example, corrective mutations in genes that would fix the sickle cell mutation, things like that. 

My mindset was already thinking about, how do we use these? They’re clearly interesting enzymes. They clearly have utility in the research arena. That just kind of expanded infinitely from our original thinking. That was: can we use these to do diagnostics or use them to detect different kinds of viral RNAs, essentially taking advantage of what they do in nature, but do it in an in vitro setting as a research tool? But I think there’s still a lot of runway there.

VIJAY: Yeah, absolutely. 

Recognizing the next engineerable system

VIJAY: I’m curious about how you have a sense for what are going to be the next things that are engineerable in biology. Are there things that you’re excited about? Or are there tips that you would give for people for how they could even identify that?

JENNIFER: Well, that’s tough. It’s one of those things where you’re either looking under the lamppost for things that look like things you already know about, or you’re doing fundamental work, on whatever topic, but you have an eye out towards, you know, ‘if I happen to come across something that looks like it’s gonna be useful or engineerable, I’m gonna pull that aside.’ 

So, Jillian Banfield at Berkeley has  been working on bacterial metagenomes for a long time. That basically just means being able to take the DNA sequences from microbes and stitch them back together, so we know what their entire genome looks like. Then, you learn fundamental biology by doing various kinds of analysis. She was actually one of the very first people to come across CRISPR sequences by doing that kind of thing.

As you can imagine, she’s coming across all sorts of really interesting observations in her work. One of the challenges that we have is that she’s often coming to me and saying, “Hey, I have this really cool observation and, you know, what do you think?” And they all look very interesting. So, we struggle to figure out where we want to focus our efforts and whether it’s worth working on the next CRISPR system versus casting our net in a different direction. To some extent, we try to do both, but I struggle with this. It’s not really very easy to figure out where the next big insight or technology will be coming from.

Sometimes when that happens, people also can get tunnel vision, right? Everybody starts working in one direction. Yet, there might be something very interesting over there that the crowd isn’t focused on but is actually really, really important.

VIJAY: Yeah. Well, I’m curious to test a hypothesis on you and see what you think. You should feel free to completely shoot this down, it would only break my heart, that’s all. One of the really interesting hallmarks about biology is the modularity. You know, from amino acids to proteins, to complexes, large things to cells, organelles, tissues, and organs, and so on, there’s a sort of modularity at many scales. And, you can mess with the amino acid or mess with the protein or you can do things at different scales. That way, not everything has to be redesigned atom by atom. You can redesign parts or so on so modularity is one part. Then you can start taking these building blocks and putting them together in interesting ways, and we’ve obviously seen that in so many different ways. So, have aspects of natural selection really been driving the engineering ability here or can you think of times where they’re in opposition? Because it doesn’t have to be the case.

JENNIFER: Right. No, it doesn’t have to be the case. As you were asking the question, I was thinking back to our shared history with ribosomes. Because, you know, back in the 1980s when people were discovering these catalytic RNAs, there was a tremendous excitement about being able to engineer something not found in nature. I think now, if you look back, it hasn’t been that easy to do a lot of engineering on ribosomes to make them do things differently from what you find in nature. Then if you look naturally, we also find that there aren’t huge numbers of diverse types of ribosomes.

VIJAY: In comparison to enzymes, which have a great diversity.

JENNIFER: Exactly. So, I think that’s one example where your hypothesis holds up. Then, with CRISPR, it’s kind of the opposite in a way in the sense that we see a large number of very diverse forms of CRISPR/Cas proteins in nature. They have the same mechanism, but they work a bit differently. So I think that’s consistent, at least with the idea that we find, in the lab, that nature has also found this to be a very pliable platform for manipulating DNA, or in some cases RNA, in cells.

VIJAY: Yeah. I’m always looking for that moment where we feel like we’ve made that transition. That moment is really important for bringing in collaborators or thinking about pouring in research funding for doing venture funding. How do you know we’ve found that moment? It almost sounds like you have to try a few things. 

I mean, one of the most important catalytic machineries on Earth, the ribosome, is a ribozyme. So, you might have high hopes for it. But it doesn’t have to be. As long as you can read, write, edit, modify, you can start making variants and start trying to do these things. And some things will be engineered when something’s going on. I guess you’ll see whether it catches. We see this in science and in startups where just people start piling in and realizing that there’s really something here.

JENNIFER: Yes. Well, I’ll tell you a little bit. Back when we were starting to work on CRISPR proteins in the mid to late 2000s, we started to get the idea that these could be very useful enzymes for research purposes. So, the first call I ever had with a venture capitalist was a call where I described to him the data we had for these CRISPR/Cas proteins that can bind and cut RNA in a very precise fashion, and how you might be able to use that activity as a way of detecting particular RNA sequences. You know, we spent an hour on the phone talking about, “What’s the killer app for this?” And nothing really gelled. There were ideas but it didn’t really gel and how would you even modify a protein like that to make it more useful? It’s not really clear. So, I kind of came away from that call thinking, “Okay, well, this is probably not yet at a point where it’s going to have that kind of opportunity to expand in a lot of directions.”

And that was very different than with Cas9, right? Because kind of immediately you knew, you didn’t need to ask anyone. It was like, yeah, this is clearly going to be something that’s going to be really useful. Then the question was, just how broadly can you engineer it to do different things? And, like you said, then as people start jumping into a field, and they start to get traction in their own projects, and you see exponential growth. That’s really exciting when you see that happening in science. We’ve seen it also in the area of imaging technologies in the last few years, as well as in cancer immunotherapies, where there’s just so many opportunities and lots of people jumping into it. I’m curious how you think about this, too, with your VC hat on.

Technologies like CRISPR, more often than not, come out of left field in the sense that they come from fundamental curiosity-driven science.

But sometimes when that happens, people also can get tunnel vision, right? Everybody starts working in one direction. Yet, there might be something very interesting over there that the crowd isn’t focused on but is actually really, really important. So, how do you think about that when you see this kind of exponential frenzy in a field and yet you have a sense that maybe we’re missing something?

VIJAY: It’s a really hard question. Like anything, you handle it with a portfolio, right? Whether it’s a portfolio of grad students and postdocs in your lab doing different things, or a portfolio of dollars, or a portfolio of companies, a portfolio of ideas. I think some of the most exciting things are the contrarian ones. But, with that said, it all is whether the data bears out and whether there’s something really there. One of the things that my strongest mentors always enforced upon me is that as PIs or as investors, we have to have some sense of good taste, right? Have some sense of some guess, some gut feeling for where are the interests or just even where our curiosity is, right?

JENNIFER: I couldn’t agree more. There’s something unquantifiable about the gut feeling about a project that is very real.

Choosing your direction

VIJAY: You know, you’ve been a founder or co-founder of many startups now. What sort of lessons have you learned or what advice would you give people that are coming up behind you that want to follow in those footsteps? Especially given all the things that we can do that we couldn’t do even just a few years ago. How does that affect the way you think about company building?

JENNIFER: So, I’m struggling with this right now actually, Vijay, because there’s a number of opportunities that build on some of the work that’s coming out of CRISPR biology and technology that could be ready for a company. Like, one of the challenges with CRISPR is the whole question of delivery. How do you deliver CRISPR molecules into cells, whether it’s in plants, or whether it’s in people? It’s a problem, right? And it’s a problem that hasn’t really been addressed in a comprehensive way. So, is that an engineering problem? Yes. But is it also going to require some fundamental discovery? I think probably the answer is yes. So, you kind of need both. 

So, is that better done in a company or is it better done in academic labs? Again, the answer is probably both. Then, it’s trying to figure out how you parse out a challenge like that and build, let’s say, a company team around it with the right people. Ideally, for something like that, you would do it with the right investors who are acknowledging that, “Yeah, this is not a short-term problem. It’s gonna be solved over a period of time.” Hopefully, you have some shorter-term goals built in there so that, from a company perspective, you can gain traction. But you have to have a team that’s going to be willing to really put in the R&D effort to make some breakthroughs.

Moving forward responsibly

VIJAY: So, thinking about this world, maybe 10, 20 years from now. You think about engineered CRISPR, engineering the rest of biology in so many different ways. We could talk about healthcare, we could talk about energy, and climate change, we could talk about feeding 10 billion people on the planet in a sustainable, healthy way. When I think about a lot of the challenges that are facing the world, they are inherently biological at some level, or could be addressed with the sorts of engineering biology technologies we’re doing. 

I’m curious how you think about the principles for how to handle what we can do, because the flip side is also potentially scary, right? The things that people could do with this great power–and they could want to do the opposite of what we described. I’m curious what you think about the guiding principles for how we should handle this new power.

JENNIFER: Cool. Wow. You threw me a tough one at the end here, Vijay. Well, I do think that part of the solution to that comes from active engagement.  I’m a big proponent of transparency and engagement of scientists, especially academic scientists, with people outside of that academic ivory tower. I think that’s very important. It’s certainly been helpful to me, honestly, over the last few years with CRISPR in thinking about all of the challenges there. And like you said, there’s many scientific opportunities with it, so which ones are going to be most important to focus on? That’s one question. But then also just making sure that the technology is advancing in ways that are productive and not destructive, right? So, for myself, I think that it’s really about engaging as broadly as possible, but also looking for ways to build synergies.

Let’s take the climate change example. It’s probably the big existential threat that we’re facing right now across humanity. Is it appropriate to be addressing that with biological solutions? Absolutely. So, then the question is how to do that. Going back to the CRISPR example, the way that I’m thinking about that is by working with colleagues that are focused on the soil microbiome. What are the ways that you can manipulate soil microbes to enhance carbon capture, but also to enhance the production of food, and deal with issues of a changing climate, from the perspective of the soil and agriculture? So, that’s one area. Now, is that something I work on? It’s not, right? But it’s something where I would love to enable others to do that to convene groups and make people aware of what the opportunities are with this technology that could apply to problems that they’re working on.

VIJAY: Yeah. You know, when I think about this question, I think the North Star for me is trying to do things that we think can be in alignment with existing biology. So, you think about fossil fuels, where you pump all this stuff out of the ground, and then you have all this residual waste, which maybe we’ve turned to plastic, which becomes different types of waste. 

But one of the key principles in biology has been the circular nature to things where the main input is energy coming in from the sun, but the rest moves along, because there’s always going to be unknown unknowns. But if we can stick to that sort of alignment, we have a chance. And what gets me really excited about CRISPR or other bioengineering technologies is that it feels like it’s the best hope for being in alignment with nature because we’re doing it in a hopefully more natural way.

JENNIFER: No, that’s very interesting. And it gets back to this question of, are engineered organisms natural or not? I mean, you’re right. If you’re using engineering to get to organisms that would exist if they had enough time to evolve, then it’s just that you don’t want to wait a million years, right?

VIJAY: That’s exactly right. You’re just sort of shooting it along a little bit, like curling, to keep it going in the right way but nothing extreme.

So just in the last minute or so, CRISPR is an example of a technology that is well-known very broadly in the public. I think people hear lots of different things about it. I’m curious if there’s anything that you wish the public understood better about the science that you’ve done?

JENNIFER: Well, I guess it comes back to where we started, in a way. I think it’s important to understand that technologies like CRISPR, more often than not, come out of left field in the sense that they come from fundamental curiosity-driven science. So, it really is important to support that kind of work, in concert with people that are taking those discoveries and applying them. Something like this doesn’t just get created, right? It has to be uncovered by a more stochastic process of fundamental science.