Engineering the P450 Workhorse to Secure Supply Chains and Save Endangered Trees with Maria Astolfi

WEBVTT
00:00:00.100 --> 00:00:06.733
Our whole supply chain for vaccine adjuvants rely on trees that grow in mountains in Chile.
00:00:06.733 --> 00:00:16.870
So basically every year to supply for vaccine manufacturing we need to cut down around 10,000 trees.
00:00:16.870 --> 00:00:19.065
We really do not want to do that.
00:00:19.579 --> 00:00:29.312
Welcome to the Climate Biotech Podcast, where we explore the most important problems at the intersection of climate and biology and, most importantly, how we can solve them.
00:00:29.312 --> 00:00:36.926
I'm Dan Goodwin, a technologist who spent years transitioning from software and neuroscience to a career in climate biotechnology.
00:00:36.926 --> 00:00:45.045
As your host, I will interview our sector's most creative voices, from scientists and entrepreneurs to policymakers and investors.
00:00:45.045 --> 00:00:50.334
Hi everyone, we're thrilled to welcome Maria Estolfi for a discussion on climate biotech.
00:00:54.500 --> 00:00:57.692
She's born and raised in the Amazonian rainforest and she's an engineer developing biotechnologies that protect biodiversity.
00:00:57.692 --> 00:01:06.465
She co-founded the first synthetic biology lab in the Amazon, where her work applied advanced technology to environmental challenges that was supported by MIT and Google.
00:01:06.465 --> 00:01:14.192
Then, after working at Ginkgo for a little bit, she's doing her PhD in the famous J-Bay at UC Berkeley with Professor Jay Kiesling.
00:01:14.192 --> 00:01:36.343
Her focus is working on replacing extractive supply chains with sustainable biomanufacturing, and today we're going to be talking specifically around her work on using P450s, one of the really interesting protein families that, frankly, I think have been overlooked, even though they end up being at the center point of everything from liver toxicology to engineering drugs.
00:01:36.343 --> 00:01:44.688
So, maria, I'm so thrilled to be able to have this conversation, and I'll say that Homeworld is thrilled to have supported you with the Homeworld Garden Grants and we'll get into that.
00:01:44.688 --> 00:01:47.191
But first of all, let's start on who you are as a human.
00:01:47.191 --> 00:01:48.034
Who are you?
00:01:48.034 --> 00:01:48.960
Where did you grow up?
00:01:50.283 --> 00:01:51.966
I grew up in a city called Manaus.
00:01:51.966 --> 00:02:01.947
It's in the heart of the Amazon rainforest in Brazil, and growing up there really inspired me to work in biotechnology.
00:02:01.947 --> 00:02:09.424
So seeing biodiversity and being all of nature has been the driving motivation of my career.
00:02:10.948 --> 00:02:11.288
Awesome.
00:02:11.288 --> 00:02:19.700
And so when you were growing up in the Amazon, did you always know that you'd one day be doing cutting edge biotechnology to reinvent chemical pipelines?
00:02:20.942 --> 00:02:36.412
I actually had a hobby when I was a kid that I liked to take different plants and smash them to pretend that I was doing perfume and I would bottle things and smell different plants.
00:02:36.412 --> 00:02:41.687
So if you think about that, we were doing biotech research.
00:02:45.502 --> 00:02:56.213
But I really like drawing and my family thought I would be an artist because I spent a lot of time painting and I loved cooking too.
00:02:56.213 --> 00:03:01.824
So if you think about it again, it's some sort of transformation of biology.
00:03:01.824 --> 00:03:09.562
So met art or making perfume or cooking something and learning from the plants.
00:03:10.204 --> 00:03:17.693
What I immediately hear from that is being rooted in the doing, because I think some people come from biotech from, like a theoretical perspective.
00:03:17.693 --> 00:03:25.526
Oh, I did physics and I really love these questions, and then these huge collection of non-identical interacting particles was what took me into bio.
00:03:25.526 --> 00:03:32.489
But I found that there's this other trajectory of people who started physically rooted, so grinding up plants and making perfumes, cooking.
00:03:32.489 --> 00:03:34.826
If that's the root, it makes total sense to me.
00:03:34.826 --> 00:03:36.426
But am I applying the right narrative here?
00:03:37.060 --> 00:03:40.807
Oh, yes, totally, and I didn't realize that.
00:03:40.807 --> 00:03:45.659
But I also grew up in a place where people are doing a lot of things with biology.
00:03:45.659 --> 00:04:05.086
There are many examples of folks transforming trees into materials or building not only houses or what we eat, but everything around us was biology-based and you could touch it.
00:04:05.086 --> 00:04:09.890
You could swim in the rivers, touch the trees and transform it.
00:04:09.890 --> 00:04:14.168
Yeah, it really was based on doing, if I think about it.
00:04:15.030 --> 00:04:19.795
I love it, and so I'd love to understand the narrative of what made you leave your hometown.
00:04:19.795 --> 00:04:20.959
What did you study?
00:04:20.959 --> 00:04:23.449
And then how did you end up at the famous Jay Bay?
00:04:24.540 --> 00:04:56.831
So we did a lot of research back in Manaus in Brazil and, as you said, I was one of the founders of the Synthetic Biology Lab and we participated in a competition called iGEM, which is like a synthetic biology competition, where students all around the world gather to build with biology and to learn by doing by engineering interrupt quickly just to say that I would say about 90 of the people who've come on this podcast did I gem right?
00:04:56.870 --> 00:05:03.190
people can't hear me smiling, obviously can't see me smiling when you said that, but I just had yeah, and it's because everyone says it the same way I gem.
00:05:03.190 --> 00:05:04.281
Is this the competition?
00:05:04.281 --> 00:05:05.584
I know what it is.
00:05:05.584 --> 00:05:06.565
Everybody knows iGEM.
00:05:06.565 --> 00:05:08.189
I love that.
00:05:08.189 --> 00:05:10.033
Igem was where you grew up.
00:05:10.033 --> 00:05:11.259
This is your undergrad.
00:05:11.560 --> 00:05:14.543
My first time on iGEM was in 2013.
00:05:14.543 --> 00:05:48.779
So it's been more than 10 years and I was so young and dreaming about biology and when I got there, I met so many ambitious people that were really pushing the boundaries of what was possible with biology, and it was also a moment that I was able to meet some of the leaders of the field and I didn't really consider myself, from where I'm from, that I could go to Boston or work at Ginkgo.
00:05:48.779 --> 00:06:07.473
But then I met Tom Knight there and Tom, which is one of the founders of synthetic biology, was professor at MIT, then created Ginkgo and we just talked and Tom was like you should apply to Ginkgo and I was 18.
00:06:07.473 --> 00:06:29.355
I was so young and he saw so much potential in me and when I finished my undergrad, I was like I'm going to apply to Ginkgo Bioworks and I got in and it was like the first person from where I'm from that got to a position at Ginkgo and it was incredible for me.
00:06:42.180 --> 00:06:51.072
I wanted to leave because I wanted to see what cutting edge biotech looked like, to ultimately come back and bring it back to the Amazon and try to connect both worlds, but I wanted to see what was going on inside the startups or how are they solving big challenges in the real world?
00:06:51.072 --> 00:06:59.230
That got me to Boston and being in that ecosystem and then after a while I was like, okay, I've seen enough of Ginkgo.
00:06:59.230 --> 00:07:10.755
I went to do my own PhD because I wanted to now create my own lines of work and chart my own path, basically, and decided that PhD would be the route to do that.
00:07:11.319 --> 00:07:13.699
Wow, I'm feeling this is a motif of just doing right.
00:07:13.699 --> 00:07:17.089
You wouldn't have met Tom if you weren't trying the iGEM thing in the first place.
00:07:17.089 --> 00:07:25.661
Presumably you had something to talk about, which is the platform for having conversations, and it's just.
00:07:25.661 --> 00:07:36.927
It resonates so much with me because so much of my own career has just been 50 no's and then one yes, and that one yes can be just totally random but then ends up opening all these things that are, frankly, a lot better than things that had rejected me.
00:07:36.927 --> 00:07:45.466
I really love that you had this amazing so that way you went straight from Brazil to Boston where you were working in like the original bio foundry for Ginkgo.
00:07:45.880 --> 00:07:49.531
Yeah, and it's so funny to me because I've never seen snow.
00:07:49.531 --> 00:08:04.173
Of course, I was in awe of the robots and the scale of Ginkgo, but also, as a human, I was seeing this new culture and getting out of my country and talking in a different language.
00:08:04.173 --> 00:08:14.237
Getting out of my country and talking in a different language and it was so formative to me to have this courage to leave and talk and survive the winter in Boston.
00:08:15.521 --> 00:08:20.600
For context for readers, I'm currently in Boston so we can fully make fun of the city for awful winters.
00:08:20.600 --> 00:08:24.992
And so, to move the story forward, you have your formative years, I think, at Ginkgo.
00:08:24.992 --> 00:08:28.028
These really are like very important years, I think, for building intuition.
00:08:28.028 --> 00:08:35.102
And so you did a couple of years between undergrad and grad work right and as a quick little editorial, as kind of like an old person.
00:08:35.102 --> 00:08:41.264
It took me two tries at a PhD to get mine, and the first one I went straight from undergrad to grad and I was like I have no idea what I'm doing, I don't care.
00:08:41.264 --> 00:08:48.623
And then what I noticed is the people who did best did a couple of years of work before and they just were so much more grounded and had such a better idea what they're doing.
00:08:48.623 --> 00:08:52.899
And so I think, another example of you being ahead of the curve on a lot of things.
00:08:52.899 --> 00:08:54.403
So you already knew that magically.
00:08:54.403 --> 00:08:58.541
And so you did two years or so and then you decided to do the PhD.
00:08:59.583 --> 00:09:06.357
It's so interesting because in retrospect, I think about it and I'm like, wow, I was into something.
00:09:06.357 --> 00:09:15.317
Because I spent three, four years at Ginkgo and saw real bottlenecks in the industry.
00:09:15.317 --> 00:09:27.062
And when I applied to my PhD, my statement of purpose was very focused on a challenge that I've seen at Ginkgo and that I wanted to try to solve in a PhD.
00:09:27.062 --> 00:09:38.705
So I was laser focused on what I wanted to do and that happened and right now it seems linear, but it wasn't really.
00:09:38.705 --> 00:09:41.692
It was more like I was so excited to do it.
00:09:41.692 --> 00:09:49.330
I was so excited to go to Ginkgo, and then things unfold and you learn and you find challenges that you really want to solve.
00:09:49.330 --> 00:09:52.855
You fall in love with a specific problem.
00:09:53.820 --> 00:10:04.894
Yeah, and this is a good kind of segue into talking briefly about the Homeworld Garden Grants and how you and I met the way we do funding is we separate problems from solutions right, and the spirit is everyone can talk about problems publicly.
00:10:04.894 --> 00:10:08.952
The solutions do have to be held privately for patent or future downstream.
00:10:08.952 --> 00:10:15.725
And I think it's wonderful just to have conversations about problems and you can tell when somebody shows up and talks about a problem they're really excited about that.
00:10:15.725 --> 00:10:22.501
Hey, this thing is cool and important and you can tell so much about the person just before they even tell you what they do, right, the way they set things up.
00:10:22.542 --> 00:10:24.405
And I always remember our first Zoom call.
00:10:24.405 --> 00:10:31.009
You came to office hours and you're like I'm going to work on P450s, because these are holy and I don't want to describe the problem for you, I'm going to toss it over to you.
00:10:31.009 --> 00:10:35.277
But I remember immediately thinking, wow, this is great, I love the way you showed up.
00:10:35.277 --> 00:10:36.721
You had a very clear sense of purpose.
00:10:36.721 --> 00:10:44.051
And then obviously we were very impressed, not only with what you did with your problem, with the solution, but then also where you took it in just one year.
00:10:44.051 --> 00:10:46.312
So enough preamble for me.
00:10:46.312 --> 00:10:48.255
What is the problem that you've been working on.
00:10:49.301 --> 00:10:59.067
The problem I've been working on is how to scale biome manufacturing and when we engineer microbes in the lab to make molecules at scale.
00:10:59.067 --> 00:11:04.917
There is a specific enzyme that's a common bottleneck in many biosynthetic pathways.
00:11:04.917 --> 00:11:14.192
They're called cytochrome P450s and the challenge is how to make that enzyme better.
00:11:14.192 --> 00:11:21.027
But when we think about the enzyme, I was inspired by them because of the products they make.
00:11:21.027 --> 00:11:48.395
So cytochrome P450s decorate terpenes, and terpenes are these incredible molecules that can be everything from a fragrance from mint or roses these are like volatile terpenes or very complex polymers like rubber, and many drugs that we use right now are also plant-derived terpenes.
00:11:49.100 --> 00:12:04.432
So we are talking about the anti-malarial drug, artemisinin, we are talking about the anti-cancer drug, taxol, and in our lab we work with this molecule called QS21, which is in the formulations of vaccines.
00:12:04.432 --> 00:12:26.471
And when we try to engineer that biocentric pathway in microbes, we see that the main bottleneck of that pathway is the step catalyzed by this enzyme called the P450s, and it's something super striking is you start with a gram per liter and you end up with 50 mg per liter.
00:12:26.471 --> 00:12:42.960
So more than 90% of flux is lost because of this enzyme, and we really need to engineer better P450s If we want to scale the next generation of drugs or materials or ingredients in biotechnology.
00:12:42.960 --> 00:12:45.730
We need to find a way to make them better faster.
00:12:47.321 --> 00:12:49.509
A thousand questions to riff on here, but I'll show you.
00:12:49.509 --> 00:12:54.211
So my PhD was in synthetic biology applied to the brain, so I didn't really ever think about P450s.
00:12:54.211 --> 00:13:05.028
But then, when I started thinking about broader climate biotech, I got really into toxicology and it turns out that P450s are the major workhorse in our livers to handle.
00:13:05.028 --> 00:13:06.123
It's everything.
00:13:06.123 --> 00:13:08.529
Toxicology goes through P450 at some point.
00:13:08.529 --> 00:13:10.024
I love the way you set it up.
00:13:10.024 --> 00:13:11.600
Everyone wants better biomanufacturing.
00:13:11.600 --> 00:13:14.450
There's this big bottleneck of this one protein P450s.
00:13:14.450 --> 00:13:16.408
No one can engineer them.
00:13:16.408 --> 00:13:19.847
Before we get into the solution side, I think it's worth learning.
00:13:19.847 --> 00:13:22.542
Just kind of a quick crash course on P450s.
00:13:22.542 --> 00:13:23.323
Tell me about the protein.
00:13:23.323 --> 00:13:24.605
What kind of a quick crash course on P450s.
00:13:24.605 --> 00:13:25.147
Tell me about the protein.
00:13:25.147 --> 00:13:26.048
What kind of reactions does it normally do?
00:13:26.048 --> 00:13:27.591
And then, why has it been so hard to engineer?
00:13:27.591 --> 00:13:29.052
Can't we just alpha fold it?
00:13:38.080 --> 00:13:41.730
No, there's such complex biological problems and complex engineering problems because they are the plant P450s.
00:13:41.730 --> 00:13:47.006
They're first, they're bound in the membrane, so they're not cytosolic enzymes.
00:13:47.006 --> 00:14:00.605
They also require very complex coordination of redox partners, so they require the orchestration of multiple partner proteins that donor electrons so they can do these.
00:14:00.605 --> 00:14:15.692
Sophisticated activation of inert bonds, which is to say that they target specific carbons with incredible precision, which is very difficult to do in chemical synthesis.
00:14:15.692 --> 00:14:28.822
But they target carbons and they can add or take out oxygens, so they can be oxidase hydrolases, so they are basically adding oxygens.
00:14:28.822 --> 00:14:36.325
So adding functional groups to carbon is specifically and precisely a carbon position in a given molecule.
00:14:37.730 --> 00:14:41.003
And so when you just set it up like that, I immediately think that metals have to be involved in some way.
00:14:41.003 --> 00:14:45.000
Is metal coordination part of these redox partners to be involved?
00:14:45.741 --> 00:14:46.042
in some way?
00:14:46.042 --> 00:14:47.924
Is metal coordination part of these redox partners, or is it separate?
00:14:47.924 --> 00:14:54.815
Yeah, so B450s have a heme core and then electrons are donated to them through their reductases.
00:15:03.980 --> 00:15:04.905
So that's how they can perform oxidations.
00:15:04.905 --> 00:15:08.206
Okay, so we can't just throw that into alpha fold and get a magic result out, because it's more than just a single protein.
00:15:08.206 --> 00:15:10.125
It is a coordination on a membrane.
00:15:10.687 --> 00:15:14.288
Yeah, so there's the membrane, there's this complex interactions.
00:15:14.288 --> 00:15:22.886
Many of the directed evolution protocols also required some sort of selection system or a colorimetric readout.
00:15:22.886 --> 00:15:33.933
We cannot easily right now, because we know so little of plant P450s, we cannot rationally design where to mutate or where to start an engineering campaign.
00:15:34.961 --> 00:15:40.360
And the dumb question I have to ask is that I think about liver P450s and you're talking about plant P450s.
00:15:40.360 --> 00:15:42.047
Are they wildly different?
00:15:42.047 --> 00:15:44.407
I'm just very curious how you think about this.
00:15:44.559 --> 00:16:00.850
Yeah which, when you're talking about the application of P450s in human health, it's actually pretty similar to plant health, because plants they evolve these P450s to start the biosynthesis of protection molecules too.
00:16:00.850 --> 00:16:14.621
So they are making insecticides, so they are defending themselves against threats, and they're also functionalized molecules for, in that case, to survive some sort of toxic trigger.
00:16:14.621 --> 00:16:23.784
But they are remarkably different enzymes, bacterial to human, and plants all have P450s, but they have different structures.
00:16:24.044 --> 00:16:24.405
Got it.
00:16:24.405 --> 00:16:31.470
So problem is there's this bottleneck which is doing a 90% loss of yield and for a very important molecule.
00:16:31.470 --> 00:16:36.544
We can talk about the context of what it's currently doing, ecosystem-wide right, but first, you've got this problem.
00:16:36.544 --> 00:16:37.605
How are you solving it?
00:16:38.750 --> 00:16:38.951
Yeah.
00:16:38.951 --> 00:16:42.522
So P450s are extremely hard to engineer.
00:16:42.522 --> 00:16:51.923
We need better P450s for biomanufacturing and there are two ways in which scientists commonly solve this problem.
00:16:51.923 --> 00:17:07.450
It's either through traditional directed evolution, which require screening selection, or the way that I learned at Ginkgo, where we screen thousands of different mutations and candidates of a given enzyme.
00:17:07.450 --> 00:17:25.038
I wanted to be somewhere in the middle, where we use robotics and machine learning to rapidly screen candidate P for 50s and use AI to guide us on which mutations to make to optimize their activity.
00:17:25.038 --> 00:17:40.236
So we use transformer models that learn the language of thousands of P450s to then dictate where are the positions in the protein that can be hotspots for improvement in their activity.
00:17:40.236 --> 00:17:44.541
So learning the language of P450s through machine learning.
00:17:45.303 --> 00:17:46.044
A thousand questions.
00:17:46.044 --> 00:17:47.026
I absolutely love this.
00:17:47.026 --> 00:17:59.974
So, if I'm imagining your lab, you have a liquid handler of some sort and you have built the assay to see if you are or not catalyzing a specific reaction, and then you're also training your own model on top of that data.
00:18:00.931 --> 00:18:02.376
That's precisely that.
00:18:02.376 --> 00:18:13.673
So we first use machine learning to learn evolutionary data of the P450s, thousands of homologues of P450s and then they're scored in their naturalness.
00:18:13.673 --> 00:18:26.146
So there's a specific score that the transformer model can evaluate the P450s, and then candidates are then built and screened in our robotic system.
00:18:26.146 --> 00:19:02.003
It's a Hamilton Vantage workstation that was designed to build and test engineered microbes and that data is then fed again to the machine learning model that learn what is good or bad mutations, learn what is good or bad mutations, and in one round of doing this iteration of robotics and machine learning, we were able to screen 300 strains and find 3x improvement in enzyme activity, which is remarkable for a first generation in the context of a P450.
00:19:03.071 --> 00:19:04.681
Okay, there's a ton to unpack there.
00:19:04.681 --> 00:19:19.862
Yeah, and I think, just to start at the beginning, this whole the dream that people have been pushing for hey, ai meets bio meets robots is you get that positive feedback loop right and it sounds like you now have at least the first round or two of that actually happening.
00:19:19.981 --> 00:19:20.502
Exactly.
00:19:21.730 --> 00:19:28.451
And so then the second thing that's wild is that you have 300 strains, and so what you're just cloning into E coli, do you have?
00:19:28.451 --> 00:19:30.318
Does the workhorse organism here matter?
00:19:30.894 --> 00:19:38.420
We use production strains, so strains that are making gram per liter of substrate in yeast.
00:19:38.809 --> 00:19:47.618
So this is Saccharomyces cerevisiae strains Cool, and you've made it spin around once and you saw 30% improvement over wild type.
00:19:47.618 --> 00:19:48.381
So are you done?
00:19:48.381 --> 00:19:49.142
3x.
00:19:49.522 --> 00:19:50.104
Sorry, how dare I.
00:19:52.551 --> 00:19:53.234
Let's tug on this.
00:19:53.234 --> 00:19:55.818
You got 3x yield on your first cycle.
00:19:55.818 --> 00:19:58.233
Can you run it again?
00:19:59.196 --> 00:19:59.880
That's the goal.
00:19:59.880 --> 00:20:11.826
We ran Gen 1 and we want to run Gen 2, 3 stack improvements and get to 10, 20x improvement in titers, our target molecule.
00:20:11.826 --> 00:20:27.454
And one thing that motivates me a lot is that some of these P450s are structurally similar, and a question that nobody has answered yet is how generalizable are these machine learning models?
00:20:27.454 --> 00:20:33.270
So can we okay, we'll run thousands of mutations in this one P450.
00:20:33.270 --> 00:20:46.246
Can we then use that data to improve every P450 that there is that comes from a plant that is doing similar activity to the one that we are working on.
00:20:46.246 --> 00:20:53.344
So we are really interested in this idea of generalization and knowledge transfer between enzymes.
00:20:54.534 --> 00:20:56.029
There's a lot to tug on there.
00:20:56.029 --> 00:21:04.957
There's generalization across families and across kind of target reactions, but there's also just making sure people understand the impact of why what you're doing is so important.
00:21:04.957 --> 00:21:09.911
And so it starts with why make this molecule at all right?
00:21:09.911 --> 00:21:13.484
If it's in a plant P450, it's probably in the plant.
00:21:13.484 --> 00:21:15.894
Should we just cut all those plants down and run with them?
00:21:15.894 --> 00:21:17.480
Is that what we're doing already?
00:21:17.480 --> 00:21:21.297
What is your work supporting kind of an impact perspective?
00:21:22.099 --> 00:21:28.161
We are working on this vaccine adjuvant called QS21, and it's in the formulations of vaccines.
00:21:28.161 --> 00:21:40.613
They're administered to millions of people globally and right now our whole supply chain for vaccine adjuvants rely on trees that grow in mountains in Chile.
00:21:40.613 --> 00:21:54.057
So basically every year to supply for vaccine manufacturing, we need to cut down around 10,000 trees, which we really do not want to do.
00:21:54.057 --> 00:21:56.143
That, yeah, no.
00:21:57.550 --> 00:21:59.499
Do you know what this analogy immediately makes me think of?
00:21:59.499 --> 00:22:00.996
It's the poor horseshoe crab.
00:22:00.996 --> 00:22:14.781
If you want to make a molecule and it's going to go into humans, there's currently the assay that you have to run it on horseshoe crab blood because there's this one property the blood it's blue, blah, blah, blah, and I think it's an anticoagulation property or it's a toxin.
00:22:14.781 --> 00:22:20.634
Anyway, I'm saying it's so obscure and it's so lame that we have to farm all these crabs.
00:22:20.634 --> 00:22:23.859
There's the ecological impact that it's crazy that we're wrecking these crabs.
00:22:23.920 --> 00:22:27.044
And there's also just the where's the technological ambition?
00:22:27.044 --> 00:22:32.510
Right, it is such a gross hack that we have to farm all these crabs and cut them up just for this one stupid assay.
00:22:32.510 --> 00:22:35.238
That absolutely could be done with modern tools.
00:22:35.238 --> 00:22:36.520
And it's the same thing.
00:22:36.520 --> 00:22:44.432
Like I never would have known that vaccines rely on a I'm going to say this in a loving way random tree in Chile, but it just happens to have this property.
00:22:44.432 --> 00:22:45.635
So where's the abundance mindset?
00:22:45.635 --> 00:22:46.759
Or where's the ambition that?
00:22:46.759 --> 00:22:50.717
Yeah, great, let's just build that molecule ourselves and save ourselves the trip to Chile.
00:22:51.940 --> 00:22:56.538
Yeah, there's a fact that it causes ecological damage.
00:22:56.538 --> 00:23:17.363
But if we think on a supply chain level too, in a scale of supply chain, we've seen, for example, that taxol the production of the anti-cancer drug taxol led to the trees to be in endangered status, so it almost led to all the taxol trees in the Pacific Northwest to go extinct.
00:23:17.363 --> 00:23:39.884
If we think of other examples like erythromycin, the cost variability so we are talking about potentially leading trees to extinction supply chains being incredibly variable because the yield of the tree, how much you can farm every year, will dictate the cost of a given pharmaceutical.
00:23:39.884 --> 00:23:49.817
And because it's so difficult to go to Chile and get the bark of the tree in mountains, the cost is huge.
00:23:49.817 --> 00:23:55.568
A gram of QS21 costs around $100,000 to $200,000.
00:23:56.431 --> 00:23:59.394
Oh, my gosh A gram of this vaccine, edgman.
00:23:59.394 --> 00:24:12.354
And we want vaccines to be affordable, right, like we want people to get immunization against diseases, that we really want this to reach people's arms as affordable as possible.
00:24:13.638 --> 00:24:21.134
And this is one of the stories that gets me so excited and why I love both the problem and the solution that you're deploying is that this is a case where everybody wins?
00:24:21.134 --> 00:24:24.637
Is that this is a case where everybody wins if we can solve this problem right?
00:24:24.637 --> 00:24:29.782
$100,000 a gram is a lot of market forces to do it better, right?
00:24:29.782 --> 00:24:32.967
But then it's also one of those things where you just have to take a step back.
00:24:32.967 --> 00:24:36.397
Wow, that's great that it's like that expensive.
00:24:36.397 --> 00:24:37.788
And then wait, how come nobody's done this?
00:24:38.131 --> 00:24:49.457
Super complex is, if you see the molecule, this is like the longest ever refactored biosynthetic pathway in a microbe.
00:24:49.457 --> 00:24:51.942
That happened in the history of biotech.
00:24:51.942 --> 00:25:07.416
That's why our lab like this became like a nature paper and the technical challenge, the engineering challenge, to be able to precisely put all these functional groups in a terpene is just monumental.
00:25:07.416 --> 00:25:10.763
So now the challenge, the real challenge.
00:25:10.763 --> 00:25:12.031
We made it.
00:25:12.031 --> 00:25:14.316
We can make droplets of this molecule.
00:25:14.316 --> 00:25:16.760
How do we scale it?
00:25:16.760 --> 00:25:21.417
We can start manufacturing QS21 for global immunization.
00:25:22.319 --> 00:25:24.002
Yeah, can you answer your own question.
00:25:24.002 --> 00:25:29.403
Are you at the spot where you're thinking about scaling, or where's the state of play?
00:25:29.403 --> 00:25:30.796
I think is the first question asked.
00:25:31.298 --> 00:25:40.878
Yeah, the state right now is we are showing a tools development, so technology that could rapidly improve every enzyme of the pathway.
00:25:40.878 --> 00:25:48.996
There are over 30 enzymes working together to make this super complex, plain natural product.
00:25:48.996 --> 00:25:57.535
But if we can rapidly engineer and de-bottleneck them, we can unlock their biosynthesis and bioreactors.
00:25:57.535 --> 00:26:36.955
So where I am right now is building that platform to rapidly optimize titers and we do think a lot about scaling QS21 and how would that look like if it's going to be a startup or maybe a non-profit, because these are immunization when our lab commonly we do proof of concepts in the scale of academic labs and spin out companies to then further reach market I'm gonna go out on a limb and say any person with startup experience is gonna hear the way you set it up and just start cold calling you.
00:26:38.901 --> 00:26:53.451
Maria, you need to start a company tomorrow you're selling something that's so high value and eventually but I'll let the randoms from the internet reach out those thoughts and I'm sure I think you're very sophisticated on both sides of for-profit, non-profit, so I am curious where you're going to go with that.
00:26:53.451 --> 00:26:58.932
I do have a technical question though, which is that you've seen stuff on the bi-manufacturing side of Ginkgo.
00:26:58.932 --> 00:27:19.801
You've now been working on these extremely complex pathways, which to me is super inspiring, thinking about a 30 enzyme cascade, and on cascade we had James Waltz from Cascade Bio talking about their polymer-based enzyme immobilization they use for cell-free, and so I'm so curious if you have a take on, kind of, if you would think about scaling.
00:27:19.801 --> 00:27:24.040
Do you ever is this could you put a 30 enzyme cascade in a cell-free environment?
00:27:24.040 --> 00:27:29.450
Or do you think that when you talk about biomanufacturing, it has to be microbial workhorse based?
00:27:29.650 --> 00:27:37.765
We like microbes because they can turn simple sugars so renewable resources into complex molecules.
00:27:37.765 --> 00:27:46.525
So in this case, we could think of a bioreactor that turns sugarcane into QS21.
00:27:47.310 --> 00:27:48.435
And Brazil's got sugarcane.
00:27:48.810 --> 00:27:52.930
And we got plenty of sugarcane now in South America.
00:27:52.930 --> 00:28:07.999
My dream really for when we are talking about synthetic biology and biomanufacturing, is that these and new supply chains can be deployed in the regions from which these technologies come from.
00:28:07.999 --> 00:28:13.356
So I imagine a factory in Chile, right there.
00:28:14.566 --> 00:28:16.432
Just next to the tree, so it's extra.
00:28:16.432 --> 00:28:20.015
You can still do the trip if you want to, but we'll just give it to you in a tube.
00:28:21.286 --> 00:28:43.239
Of course, for the context of the US and securing supply chain, we wanted to be here, but one thing that I really care about is and it's a tangent from your question but how to return investment, return back what we do through biomanufacturing to actually protect the trees in Chile.
00:28:43.239 --> 00:28:50.509
How do we take our benefits and proceeds of biotechnology towards the stewardship of biodiversity?
00:28:50.509 --> 00:28:57.409
It's a question that I'm really interested in as we deploy these technologies for commercialization.
00:28:58.971 --> 00:29:01.417
I love it, one kind of kernel of thought there is.
00:29:01.417 --> 00:29:29.994
One of the first guests we had on the podcast was Braden Tierney from Two Frontiers Project, and what I love so much about the work that he does is he's just a wild man of scuba diving, deep microbial culturing and sequencing, so he sets up these labs in wild places, and so wouldn't it be cool to combine your view of let's take the coolest thing and synthesize it in a decentralized and safe way for the organisms, but then also do it in just a gold standard microbiology lab that you can put right next to the plant?
00:29:30.656 --> 00:29:30.978
right.
00:29:30.998 --> 00:29:31.419
And mine it.
00:29:31.419 --> 00:29:34.150
That's all right that's a collaboration.
00:29:34.150 --> 00:29:34.792
That needs to happen.
00:29:35.054 --> 00:30:07.719
Yeah, my wildest dream is that I will go back to the Amazon, put a lab in a boat and go through the river collaborating with communities to learn about their plant natural products and create supply chains that are sustainable to manufacture these medicinal molecules in a way that will decrease deforestation and also bring a growing economy for communities in biodiverse regions.
00:30:07.719 --> 00:30:10.327
So you bring everything.
00:30:10.327 --> 00:30:14.298
As you said, everybody wins, the community wins.
00:30:14.298 --> 00:30:23.597
We bring incredible technology forward, use the technology to protect biodiversity and just create awesome products.
00:30:23.597 --> 00:30:25.105
That's it, that's the dream.
00:30:26.105 --> 00:30:38.534
Exactly, and the way I play this back to your work is that you're taking a problem that is, there's an obvious market force that wants your problem to be solved, but in the way of it is such a huge technical challenge that nobody thought it was possible.
00:30:38.534 --> 00:30:46.439
The only people in the world that have even thought about this is the J-Bay, and the only person that is actually tackling this is you, right, which is amazing, right?
00:30:46.439 --> 00:30:53.583
So this is one of those things where actually, like technology, this is where I become like a super optimist, which is that this is amazing right.
00:30:55.384 --> 00:30:59.289
This is something that market forces wants to support and hasn't been able to because the technology isn't there, but maybe now's the time.
00:31:00.045 --> 00:31:21.074
Yeah, we started by saying there's this trajectory of doing and I don't really think that QS21 is such a complex and exciting problem and we can tap on how valuable QS21 is to create technology that will enable other biodiverse products.
00:31:21.074 --> 00:31:27.055
So there's a super high value molecule and we need infrastructure to scale it.
00:31:27.055 --> 00:31:35.619
Once we have that infrastructure, we can de-bottleneck many other natural products with that same pipeline.
00:31:35.619 --> 00:31:53.894
But we need one first target that makes economical sense, that will actually lead to ecological impact, that will lead to global immunization, that has metrics and a cost, that makes sense for biomanufacturing.
00:31:53.894 --> 00:32:00.593
Once we have that infrastructure, we can tackle many other problems, but we need a first target that makes sense.
00:32:00.905 --> 00:32:02.865
Oh, I can't agree with you more.
00:32:02.865 --> 00:32:06.496
Climate biotech right now is just starved for a win.
00:32:06.496 --> 00:32:16.813
We need one clear, big win and we've seen this a lot, especially in climate is that people want to rush so quickly to the platform to solve the everything for everyone.
00:32:16.813 --> 00:32:24.924
And I think there's so much bravery in the person who says I just do this and if we do this, then it matters.
00:32:24.924 --> 00:32:36.087
But right now, if we don't get this adjuvant working, then nothing matters, and I think there's so much strength in that we can say that we will scale thousand medicinal molecules.
00:32:36.949 --> 00:32:53.411
I really believe on regional, one specific target kind of action and hyper focusing on that problem so we can actually solve it and see the impact that we are causing in one supply chain.
00:32:53.411 --> 00:33:03.637
I like let's do it one supply chain because of one tree, because of one molecule and then reproduce that to others.
00:33:03.637 --> 00:33:05.567
But we need one successful example.
00:33:06.150 --> 00:33:06.590
I love it.
00:33:06.590 --> 00:33:14.641
As we wrap up, I just have to kind of draw from my previous career chapter where I was in tech right and software and the common phrase is, oh, we're doing X for Y.
00:33:14.641 --> 00:33:24.932
Right, we're doing Uber, but for pets or something like that, and it would be so cool to be like, oh, we're doing Maria's play Y if we're just missing the X.
00:33:24.932 --> 00:33:33.353
I think, in showing that these biomanufacturing pathways can really work, and so I just can't cheer enough for the work you do and would really encourage anyone to reach out for your work.
00:33:33.353 --> 00:33:36.390
But now we get to wrap with our four little rapid fire questions.
00:33:36.390 --> 00:33:38.635
We ask everyone this and we'll run through them.
00:33:38.635 --> 00:33:47.460
First one is Maria, what is a single book, paper, art piece or just idea that blew your mind and shaped your development as a scientist?
00:33:48.545 --> 00:34:17.280
I actually, if I think of the things that really dictated my career, it was not a book, not an art piece, it was just being with nature, like seeing it, swimming with nature and being so, seeing it swimming with nature and being so in relationship with nature and seeing.
00:34:17.280 --> 00:34:19.085
So let me try to rephrase this.
00:34:19.085 --> 00:34:31.929
What I'm trying to say is that when we were talking about making the perfume and just doing with biology that's what I'm trying to recap that that was really the most influential moment to me.
00:34:33.333 --> 00:34:36.398
I love it Just hands-on touching an amazing plant.
00:34:36.965 --> 00:34:39.648
Yeah, just being hands-on with nature.
00:34:40.307 --> 00:34:44.952
I love it All right, what is the best advice line that a mentor gave you?
00:34:45.632 --> 00:35:21.130
Something that actually changed my career when I was at Ginkgo that a colleague of mine said that and it's not necessarily an advice, but it can be applied as an advice that he came and said you really have something to say and there's this specific place that you work in that it's very uniquely you.
00:35:21.130 --> 00:35:33.677
So what he was saying is that he thought that I was able to combine these two worlds together and make sense of it, and that had something to say about that one specific intersection of knowledge.
00:35:33.677 --> 00:35:44.851
The advice here is to try to find that intersection of different knowledges and disciplines that make sense to your personal story, but also professional and so on.
00:35:44.851 --> 00:35:45.853
That's awesome.
00:35:46.434 --> 00:35:50.170
Yeah, that's actually one of my favorite answers to this question and we can talk more about it offline.
00:35:50.170 --> 00:35:57.251
But I think finding the authentic voice is like a LinkedIn forgettable buzzword that people step over really easily.
00:35:57.251 --> 00:36:00.876
But doing something that's uniquely you is the best possible thing.
00:36:00.876 --> 00:36:02.244
I think that's the magic moment.
00:36:02.806 --> 00:36:07.135
I love how you're good at synthesizing things, dan.
00:36:07.135 --> 00:36:10.086
I remember you're talking about our first meeting.
00:36:10.086 --> 00:36:16.528
I was like, oh, the people are 50s and they do this, this, and that I feel like you just did that so well.
00:36:16.528 --> 00:36:17.791
Thank you.
00:36:18.331 --> 00:36:18.952
Yeah, it's very.
00:36:18.952 --> 00:36:20.775
You gave me good material.
00:36:20.775 --> 00:36:30.458
So now, third question if you had a magic wand to get more attention or resources into just one part of biology, where would it be?
00:36:31.425 --> 00:36:43.121
I love grassroots biotech and meaning hyper regional solutions, community labs, that phrase of like local people solving local problems.
00:36:43.121 --> 00:36:47.677
That's where I really would love to see biology going forward.
00:36:47.677 --> 00:37:03.452
We talked about Chile, so yeah, folks in Chile developing a new supply chain for a problem that they face in their backyard, people in the Sahara Desert, people in the Amazon, in the Himalayas, in the Andes.
00:37:03.452 --> 00:37:11.291
I want to see biology everywhere in hyper-regional, so grassroot biology.
00:37:12.887 --> 00:37:14.833
And shout out to our friends Lab to Land, all right.
00:37:14.833 --> 00:37:21.976
Fourth question what's a skill that you think bioscientists need to invest more time into developing?
00:37:22.918 --> 00:37:26.447
Empathy, so much empathy.
00:37:26.447 --> 00:37:53.054
I think so many times we are focused on the technical parts and but I want us to reflect more on the bigger problem, to talk about the tree and where they grow and where does this knowledge come from, and think on higher level and have more empathy for the whole process in a more holistic view.
00:37:53.996 --> 00:37:56.146
I think that's a fantastic answer, maria.
00:37:56.146 --> 00:37:57.789
It's been so much fun to talk with you.
00:37:57.789 --> 00:38:12.257
I'm leaving this really inspired and intrigued, and I love your provocation of how can we be setting up the kind of in situ kind of remanufacturing, so to speak, of getting the best out of bio while still preserving the bio.
00:38:12.257 --> 00:38:19.507
So for anyone listening that wants to get in contact with you or follow along your journey, would you have any suggestions of where people can find you?
00:38:21.050 --> 00:38:24.036
Yeah, I am on LinkedIn.
00:38:24.036 --> 00:38:26.623
Might be the easiest way to find me.
00:38:26.623 --> 00:38:28.407
My email is also.
00:38:28.407 --> 00:38:29.371
I'm super open.
00:38:29.371 --> 00:38:32.775
I'd love to talk with anybody about this.
00:38:32.775 --> 00:38:40.994
Specifically, reconnecting biotech to biodiversity is where I want to be, and if you want to talk about this, let's hang out.
00:38:42.387 --> 00:38:42.791
Awesome.
00:38:42.791 --> 00:38:45.994
Well, maria Stolfi, thank you so much for taking time to talk with us here.
00:38:45.994 --> 00:38:47.090
Thank you, thank you so much for taking time to talk with us here.
00:38:47.885 --> 00:38:48.065
Thank you.
00:38:49.670 --> 00:38:53.197
Thank you so much for tuning into this episode of the Climate Biotech Podcast.
00:38:53.197 --> 00:39:01.702
We hope this has been educational, inspirational and fun for you as you navigate your own journey and bring the best of biotech into planetary scale solutions.
00:39:01.702 --> 00:39:08.376
We'll be back with another one soon and in the meantime, stay in touch with Homeworld Collective on LinkedIn, twitter or Blue Sky.
00:39:08.376 --> 00:39:10.070
Links are all in the show notes.
00:39:10.070 --> 00:39:16.273
Huge thanks to our producer, dave Clark, and operations lead, paul Himmelstein, for making these episodes happen.
00:39:16.273 --> 00:39:17.476
Catch you on the next one.