TRANSCRIPT: Meeting 2, Session 2

Jim Wagner:

Good morning. Ladies and gentlemen, Professor James Collins joins us from Boston University, where he is the William F. Warren Distinguished University Professor and Professor of Biomedical Engineering. He co-directs the Center for Biodynamics, is also a Howard Hughes Medical Investigator. He has been the recipient of numerous high-profile awards, including one of the MacArthur Genius Awards. A pioneering researcher in synthetic biology. For those of you that have looked at the literature at all, you know that. And he has used it as a model to design and construct synthetic gene networks. We’re very happy to have you with us today, Professor Collins. Thank you for being here.

 

And Ron Weiss, Associate Professor at MIT’s Department of Electrical Engineering and Computer Science and its Department of Biomedical Engineering. His research focuses primarily on synthetic biology, where he programs cell behavior by conducting and modeling biomedical and cellular computing systems. Dr. Weiss is also a recipient of numerous awards for work in synthetic biology. He was named MIT’s Technology Review magazine’s Top 100 Young Innovators, an exclusive group. Welcome to you as well, Dr. Weiss. We are excited to learn more about both of your work. And let me toss the ball first to Professor Collins, who got up and left the table.

 

[AUDIENCE LAUGHTER]

 

James Collins:

Thank you. I very much appreciate the opportunity to address the commission. What I want to do over the next 15 minutes is to give you a feel for emerging developments in synthetic biology and some of the applications that are also being developed. The field really began to take shape in the mid ’90s. At that time the human genome project was coming up to full speed, and the molecular cell biology and genomics communities were recognizing that as the parts list for the different organisms was being created as a result of the sequencing efforts, one of the next challenges was to figure out how those parts worked together in networks and pathways in living cells. Of note, back to the two very nice talks we had earlier, is that in the genetic engineering revolution of the ’70s, ’80s and early ’90s, there were few, if any, engineers that took part.

 

The genomics community turned to the physicists and engineers in the mid ’90s, recognizing that these group groups could handle complexity quite well. And the challenge there was, could you begin thinking about how they work together inside cells? And we answered that call—myself and others, including Ron Weiss, who will speak next. But what we found was that the data available to do that—that is, to reverse engineer natural networks—were not yet available. Microarrays, that is, technology that allows you to measure simultaneously, say, thousands of genes’ activity inside a cell, had just arrived. It was very expensive, and we had very few publicly available data sets. So I and others ran as fast as we could away from that problem because there were not the data available. So, what they were asking us to do was, could you as an engineer, for example, figure out how a radio was wired up from the outside by taking it apart, which is something that many engineers do.

 

With that challenge in mind, we sat back and said, okay, instead of taking apart a radio what we engineers often do as tinkerers is put together circuits that make up a radio. And it was this thinking about forward engineering—biological circuits—that helped launch synthetic biology. So what we realized was, you could come up with schematics just as an electrical engineer would for a circuit, do some mathematical modeling, and then find the biological parts that matched your circuit, piece them together using enzymes from the genetic engineering revolution in a cut-and-paste fashion, into constructs or plasmas, plasmas being rings of DNA, that you could then get into a cell and see if it functioned in the way that you thought it should function.

 

Tim Gardner, who was a Ph.D. student of mine at the time in the late ’90s, joined me in this effort, and we sat back and said, “Okay, what could be a circuit that would be suitably simple enough that we could analyze it and build it but suitably complicated enough that it would be worthy to build and analyze?” And what we arrived at was this idea of a genetic toggle switch. This was inspired by digital toggle switches or flip-flops or RIS latches, which is a very simple form of a switch that’s at the heart of all PCs and broad computer-based system. And this switch, what you have on the electronic side is that from a single stem you can flip it stably between two states: on/off or 0/1 in a binary fashion. Tim and I sat back and said, okay, how can we do this biologically? And what we came up with is the schematic shown on this slide, where what you have are two genes, repressor one and repressor two. You set them up so that promoters, or on-switches, are constitutive. That is, they always want to be on. You then further design the system so that the protein produced by repressor 1 wants to bind to the on-switch for repressor 2, shutting it off, and the protein produced by repressor 2 wants to bind to the on switch of repressor 1, shutting it off. So you have a system where in two genes each want to be on but each is trying to shut the other off, so it’s a mutually inhibitory network. In principle, you can design this system so it wants to exist in one of two stable states: either state A where gene A is on and gene B is off, or state B where gene B is on and gene A is off. In principle you can flip it between those states by transiently delivering a chemical and environmental stimulus that will temporarily inactivate the currently active gene.

 

Say, for example, you’re in state A, gene A is on and gene B is off. You deliver a chemical, for example, that will now inactivate that gene or destabilize its protein. This now allows gene B, which had been kept off by gene A, to come on, produce its protein, and once it’s at a high enough concentration to shut back off gene A you can remove your stimulus and you’ve flipped the system from State A (gene A is on and gene B is off) to State B (gene B is on and gene A is off).

 

We took this notion, went to where bioengineers often go: we did mathematical modeling, which is quite easy. You don’t have to have a bioethics commission to do computer modeling. We then convinced ourselves that this could work and arrived at certain conditions that could help guide the design. We created DNA plasmids, taking different components, genes and promoters, to work in e. Coli, a bacterium. And then we got this going in bacteria with a number of different versions where you could start the system in the off-state, flip it on with a chemical. The system could stay on indefinitely.

 

We looked out, for example, four days, which is a very long time for e. Coli, which divide every twenty minutes. You could flip this system from the on state to the off state with either another chemical or an environmental stimulus like a shift in heat. We showed this could have a very sharp switching threshold, which becomes important for a number of applications, some of which I’ll address in the next few slides. Wherein as a system would start in the off state, at some very tight range of concentration of our chemical, your switch begins to flip from the off state to the on state and then stays on.

 

So what we did here in one of the first demonstrations of synthetic biology was take inspiration from electrical engineering and build the biological equivalent of cellular memory, where now you have a small circuit that can be generalized to other organisms that can endow your cell with cellular memory. It opened up a whole range of applications, some of which were in biocomputing. There was ten years ago when this came out and still is interest in using programmed bacteria and other organisms to do different computing aspects. I don’t think anytime soon we’ll have a P.C. with “e. Coli Inside” on the label for many reasons, which we can discuss in open questions.

 

But if you think about computing instead, about programming cells with certain functions, this has been really one of the dominant applications of synthetic biology in the last decade. And we took that charge on the heels of these earlier efforts. We began thinking about could you take the engineered gene circuits and other modules from synthetic biology and interface it in a natural organism with fascinating signaling properties and detection properties and interesting phonotypical output properties that in some cases have evolved over billions of years. And one of the challenges we took on was the idea of creating wholesale biosensors. Could we, for example, program bacteria to sense things in the environment, be it a chemical, a pathogen, a heavy metal.

 

What we did to make it easy was just to design e. Coli to be sensors of DNA damaging agents. What we did is, we took our toggle switch and coupled it to the SOS pathway in e. Coli. This is a pathway that senses when there’s been DNA damage and used it so that we set the system up so that it starts in the off-state, and can it now detect the DNA damaging agent? What we used it to do was look at mitomycin C, which is a DNA damaging chemical, as well as UV radiation. The system is highly stable in the off state. If you want to have a good biosensor out in the environment or in your home, you want it to have a low false positive rate. That is, you don’t want it to say it’s detected anthrax unless there’s anthrax there.

 

Less critical but nonetheless still as important, you don’t want your smoke alarm at home going off randomly in the face of no smoke or fire. You only want it to go off when there’s fire. So you want to have a stable off and this system had a very stable off. We looked out 48 hours. The cells in the off state stayed off, but then as you delivered tiny, minuscule amounts of this DNA-damaging chemical, boom, you could begin actually detecting cells, and it increased its response as the amount of damage increased.

 

We might hear it from Ron Weiss, but there are many other applications in this regard where there are natural promoters, on-switches in bacteria, that sense things like heavy metals, be it lead. There are now engineered promoters and receptors from the synthetic biology field that are detecting other chemicals and substances, including TNT. What the toggle and other switches afford you is a form of memory. That is, now you can actually have a detection event, go back and inquire did you actually have a detection event without having to watch the scheme at all once. In this case, we were using fluorescent proteins, which is kind of the favorite output for synthetic biology, but we also engineered the system so that the output would be a biofilm. A biofilm, which I will return to in a couple of slides, is basically a community of bacteria that are encapsulated in a polysaccharide matrix and attached to a surface. So the plaque on your teeth, the gunk in the sink, the stuff on the side of a ship are forms of biofilm.

 

In this case, we had it so that the bacteria would form biofilm when you detect and event, which now could give you an easily visible detection if, for example, you want to sense are there mines in a certain field or has there been a detection of a pathogen. You can now fly over or walk by and see, ah, the purple biofilm has been formed; we need to get in there.

 

There are many other switches, components, networks that have been designed in synthetic biology. In our lab a few years ago with Farren Isaacs and Dan Dwyer, we developed an RNA switch, which is shown here schematically, that quite briefly sets up where we can design a sequence to put in front of an MRNA so that it binds and blocks the ribosome binding site, which is where a ribosome will come in, dock, and produce a protein. With this, you can basically shut off protein production inside a bacteria.

 

We then as second part of the switch designed a short transactivating component that would turn on a non-coding RNA that would interact with the scissure press sequence, pull it off, expose the ribosome, and now allow the protein to be expressed.

 

This switch has a number of interesting properties, in particular that it has low leakage and very fast response times. Where we used it was an interesting application in part to address growing concerns in synthetic biology. The public and others are worried that labs like mine and Ron’s and others may be engineering organisms that could get out in the environment and do damage. They’re worried that as we put things in the environment to sense, will they run amok? Will we not be able to control them? And we realized, well, why don’t we think about using synthetic biology to program them to handle such concerns. And what we came up with were different designs to design synthetic gene networks that could count. So the idea was that If you are going to put a sensor in the environment, have it so that after five cell divisions or five days the cell would count these events and then commit cellular hari kari. And so with Ari Friedland and Tim Lu we came up with two such designs, one of which had the RNA switches basically in linear cascades that create daisy chain cascades so that after the first detection even, be it either that one day has passed or a cell cycle protein has been detected and thus one cell cycle has been undergone, it will now flip a switch, priming the next guy down the line to be ready for the next detection event. And we, in linear fashion, showed you could have these so you could have your bacteria count to two or count to three. In this case it’s only on a output of GFP, or fluorescent protein.

 

Tim Lu, who has just joined the same department as Ron Weiss at MIT as faculty, came up with a separate design, where Tim, instead of using RNA switches, used recombinates. These are enzymes that can chop out a bit of DNA, flip it, and reinsert it. And so we set now up a cascade of these such that the downstream gene was off if the sequence was a certain orientation. Then as you detect an event, it would flip it, flip on this gene, which now primed the next scheme. This guy actually also displayed very nice behavior at counting out to three, operating on the order of hours instead of minutes, as Ari’s RNA switch.

 

Very recently, we went the next step and that is got the cellular hari-kari component going. So, with Jared Callura, on work we just published a month ago, we used his RNA switch to control two different proteins, expressed two different proteins, that when expressed together will cause a cell to lyse, explode and basically kill itself. What this means is that you can now use this with these switches that can be controlled independently inside the same cell, to respond to two different stimuli. So that now if you have this in the environment sensing schemes, you could spray your field with two chemicals to have the switches now commit hari-kari, or you could have it that if you have these sensors inside your body—and Ron is going to talk about some applications that are more clinically oriented—you can now after some period of time have these also be flipped on to basically self-destruct, as we see here.

 

Interestingly, when we published this we have been contacted by a number of biotech and bioenergy companies who are interested in this technology primarily to combat corporate espionage. So I consult to a number of bioenergy companies, and one of their big concerns is that either a corporate competitor and/or an adversarial nation will sneak into their lab, with a little cup dip into their vat and steel their secret recipe of an engineered microbe that can convert sunlight or some other biomass into a fuel. And they’re thinking about how you can protect against that if they steal the secret sauce and run out of the lab to have their organism self-destruct should it leave their bioraft. And we’re talking about how this and later things might be modified.

 

A growing area of focus in synthetic biology is infectious disease. We are facing a major public health crisis from the increasing rise of resistant bacterial strains in this country and around the globe, which is coupled with a decreasing supply of effective antibiotics. We’re in a major crisis that hopefully our country and others will increasingly address. We got intrigued by this going back a few years ago and came up with, for example, two different ways to use synthetic biology to address biofilms.

 

So I mentioned biofilms as an interesting output for one of our devices. But also biofilms are a major problem from an infectious disease standpoint. Bacteria in biofilms are about a thousand times more resistant than they are free-swimming to antibiotics. If you go for an implant, the big risk now, be it for artificial hip, an artificial knee, a pacemaker for your heart or for your brain, is no longer the risk associated with general surgery. Your major risk now is that you are going to get an infection as a result of bacteria growing on your implant.

 

What we did with Tim Lu and others was re-engineer bacteriophage, which are viruses that only infect bacteria, to go in and express enzymes to break up the biofilm as well as then to allow the bacteriophage to go in and explode the cells. We simply went in and used them to go in and produce enzymes that would similarly break up the biofilm. Tim went further and re-engineered the phage to go in and express proteins that together with antibiotics would make the antibiotics much more effective. We are now working with the U.S. Army and the Walter Reed Medical Institute to use the synthetic biology approach to develop phage to help out U.S. soldiers returning from Iraq and Afghanistan. It turn out that the soldiers coming back have infected skin wounds that are infected with resistant bacteria, three different strains they have identified, that are not responding to the antibiotics we have, and we’re now working with them to see if we can actually use the engineered phage to go after them.

 

Lastly—and Ron is going to pick up on this topic—most of the efforts to date in synthetic biology have focused on microbes, bacteria and yeast. Increasingly, synthetic biology is moving into the mammalian world via rats, mice, or humans. We took up this challenge just a few years ago with Tara Deans, when we developed a very tight mammalian switch that could basically keep genes effectively off and then allow to you flip them on from basically that off to very high expression levels, and we did this using repressive proteins as well as RNAi. What Tara was able to achieve was 99.99% repression, which for those who do work in molecular biology of mammalian cells know that it is an incredible achievement that she did achieve. You can get a lot of things off in biology. It’s often harder if you get it cut off, can you get it back on? And Tara was able to get it back on. She could flip these back on an off with a cycle. And just as you can with probably the lights in this room from a dimmer switch, she could go from off to very low levels of expression to very high levels of expression by tuning her switch. This switch was modular, as are many components in synthetic biology, or you could use it to control any gene of interest. And you could use any promoter to drive it, including tissue-specific promoters.

 

The last example I will give is kind of a machismo demo that Tara came up with of the tightness of this switch. What she used was this mammalian switch to control DTA, the expression of DTA, which produces diphtheria toxin, a single molecule of which is known to kill a cell. Tara created a stable cell line, grew these up for many weeks with the cell in the off state. They were healthy and they were fine. She then selectively flipped on the switching cells and killed off the cells, with opens up whole hosts of applications, primarily in a functional genomics and animal model, where you can now imagine having tissue-specific promoters controlling the expression of diphtheria so you have your animal grow up, and now you can kill off certain neurons or tissue types to mimic a disease or to mimic an aging process.

 

There are a whole host of applications. Probably the fastest growing areas on the mammalian side from stem cell, which Ron will address in few minutes, as well as cell therapy. I mentioned things such as bioenergy and biosensors. That space is shifting quickly towards new and novel materials. I think you will see the field, which has focused the last five years on bioenergy, move to using microbes to produce specialty chemicals, because from an economic standpoint it makes much better sense for these companies than it does for energy.

 

And then finally, functional genomics has already taken off with mammalian switches and others in an area which I hope you do hear from, because I know David Relman is coming tomorrow, is on the microbial. Synthetic biology is intrigued now about re-engineering the microbes in our gut. We have ten to a hundred times more bacterial cells making up our bodies than we do mammalian cells. That doesn’t mean that we’re primarily bacterial. It only constitutes about three or four percent of our mass. But increasingly we are recognizing the important role that those bacteria play in our metabolism and health, and synthetic biologists, including our lab, are now thinking about you can re-engineer some of these bacteria with our circuits to express proteins that might be lacking, leading to allergies and/or metabolic conditions and how those could be remediated.

 

[AUDIENCE APPLAUSE]

 

Jim Wagner:

Thank you so much.

 

Ron Weiss:

So it’s a pleasure to be here. Thank you for inviting me to participate in this panel over here. These are exciting times for synthetic biology. Maybe ten, twelve years ago, as Jim mentioned, we really were just starting out to understand what we can engineer inside living cells. We focused quite a bit on trying to turn on fluorescent proteins, so turning on a green fluorescent protein and turning on a red fluorescent protein. And it’s been exciting to be involved in the field, where now we have gone from being able to demonstrate these basic capabilities, where we can control gene expression inside a cell to a situation where we can actually start thinking about applications. Start thinking about applications where we believe synthetic biology is going to make a real difference, where synthetic biology is all of a sudden going to allow you to do things that you couldn’t do otherwise. So, today what I’ll try to give you some flavor of that very briefly.

 

So before we get started on that, let me just skip over this. We think in synthetic biology this is some kind of a work flow that we’d like to be able to embody or engineer, this notion that we’ll sit in front of a computer and have some kind of a high-level application or behavior in mind. Being able to sit in front of that computer and being able to design an application, design some kind of a circuit that can help us address that particular need, and maybe go back and forth with some kind of computational mechanism until we have a validation, until we have an optimized mechanism that allows us to really address that particular need.

 

Once we’re satisfied with these kinds of computer models and computer simulations, we actually have to go ahead and explore this. The idea would be to get some DNA synthesized that is based on the genetic circuits we implemented in silica over here and get that actually physically synthesized, implemented. A few days later, order up a piece of DNA and put that inside the cells and actually try to understand experimentally whether the high-level design that we have over here can actually meet the requirements that we set forth. Obviously, this is not a one-way street over here. As is typical with any engineering mechanism that you’re using, we will have to go back and forth in multiple stages of this work flow until we get things just right. And this is what we’d like to be able to do.

 

Probably in the last meeting, you heard about this notion of a hierarchy in synthetic biology. This is something that we often like to use as a mechanism to understand how to approach the problems. So the notion of genetic engineering has clearly around since the ’70s, and we feel that in synthetic biology what we’re doing is we’re sort of adding value to this genetic engineering process.

 

Initially in this notion of being able to build more complex models that might allow us to engineer fluorescent proteins or engineer the expression of useful proteins, to me this is really—the system integration is the exciting aspect of synthetic biology. This is the thing that synthetic biology brings to the table, this notion of being able to take these modules that have been characterized, have been verified, and then understand how to bring these modules together into systems that actually reliably perform the desired functions. And this is where the challenge is.

 

Once we’re able to do that, we can actually start demonstrating applications. I’ll talk about that today. Start demonstrating applications that benefit from the fact that not only can we put one, two genes together, but maybe we can put fifty genes together. That is to me the biggest challenge in synthetic biology. How do you take these genes from multiple sources, put them together into a functional system that actually performs a particular task reliably? Once we’re able to do that, we’re going to be able to express behaviors that are just not possible without the genetic engineering perspective. What I will argue is that that opens up the possibilities for, for example, environmental applications and biomedical applications that are just not achievable otherwise.

 

What kinds of things are we talking about? Well, synthetic biology, we should not forget, can actually help us understand biology in general. This is certainly something that many of us focus in the field. So building things from the ground up helps us actually appreciate what mother nature has been able to do and also study that in a different way.

 

A lot of emphasis in synthetic biology in these applications that might have immediate effect and optimize drug synthesis. Jim has mentioned this. This is actually something that’s going on right now. For example, the Artemisinin project by Jay Keasling, allowing us to develop an antimilarial drug in an efficient and inexpensive fashion. That’s taking place right now in companies.

 

Looking forward, environmental applications, I think, are going to have a real impact. Biosensing applications where it’s toxins or pathogens in the environment. Environmental remediation is obviously a timely issue. Can we engineer microbes that can clean up various spills?

 

Energy production, as Jim mentioned, is now a very important aspect of the synthetic biology, trying to understand how to make this in a cheap fashion.

 

What I will focus today on are more biomedical related applications, and Jim has mentioned this notion of trying to fight infectious diseases. So I will have a single slide of that, and then I will touch upon various other applications, where I think this notion of being able to program cells is going to allow us to achieve things that are just not possible otherwise. And I will talk about cancer, tissue engineering, and diabetes very quickly.

 

And so the notion of infectious diseases, we have a project right now that are calling program pathogen sense and destroy, and this notion that we need to find new ways to battle these superbugs, these antibiotic-resistant bugs. What we’d like to do and what we are doing right now, we’ve been able to demonstrate, is engineer cells—other cells. Right now, bacterial cells, but we’re also working on mammalian cells as well. For example, potentially T-cells that would be able to detect these pathogens a at a very early stage, perhaps before they have started making the toxins, perhaps before they started to create the biofilms, using things like signal amplifiers that can detect molecules that are secreted by these pathogens, and then secrete a response to that. So, for example, certain proteins that are highly specific to the pathogen that exists that basically are able to kill the cell in a very efficient fashion and even monitor whether that particular therapy is working and, if it’s not working, switch to a different mechanism, switch to a different protein.

 

In our case, it’s something called bacteriocin that can actually target a different aspect of this pathogen. And I think that this approach is really going to allow us to create these highly sensitive, quick-acting, and effective mechanisms to target the superbugs or antibiotic resistant ones.

 

And we’d like to be able to engineer—we’re doing that right now—are what we are calling sentinel killers that can respond not only to one type of pathogen but be able to detect a variety of different pathogens by listening in to different channels, communication channels, if you will, that these pathogens are using to make their own decisions. And then when you see bacteria that might be related but actually are non-harmful, make sure that these engineered cells that we created actually have—are not creating any proteins or anything like that, and maintain their own population so they’re not cause anything problems by themselves. Jim has mentioned this notion of a kill switch. So we want to make sure that this bacterial population remains the at appropriate levels.

 

Another area that we have work on right now is cancer therapy. One of the biggest problems, perhaps the biggest problem right now with cancer is the notion of specificity. If you want to cure cancer it’s actually easy. All you have to do is kill the patient, and you can take care of all of the cells. We’d like to be able to create therapeutic agents that are highly specific, much more so than chemo agent, much more so than other kinds of agents that are being used right now. For example, the agents that are being studied right now, one class of them tries to understand what’s on the cell surface. So look at a particular marker on the cell surface and make a decision about whether the cell is a cancerous cell or not.

 

Often that’s not enough, and that’s why you don’t see that being used everywhere right now. The decision by looking at a single cell surface is not enough. So that’s why we are now investigating this other option, where rather than looking at a single marker that exists outside the cell or perhaps even a single marker that exists inside the cell, embody computational and programming capabilities into this therapeutic agent so that the therapeutic agent can go inside the cell and integrate multiple pieces of information.

 

I’ll show an example right now in the next slide, where six markers that are therapeutic agents can integrate together. And the point is, the six is much better than one. The six markers that we’re using right now—or more markers in the future, ten, twenty in the future—can make a highly precise decision about whether the cell is in fact a cancerous cell or not and then carry out the therapy, meaning killing the cell very specifically if the answer is yes to this. And there are other applications that can be used from this computational capability of a therapeutic agent.

 

So, what is it we’re doing right now? HeLa cells??probably most of you know about the HeLa cells. There’s a recent Times article about that. So this is a HeLa classifier, so this can say, okay, these didn’t come out. These are going to contain a bunch of symbols that don’t make sense. I guess the question mark here is "appropriate?" So you shouldn’t be able to necessarily read the logic function, but it’s actually quite simple. This is what the logic function encodes. It encodes this notion that this marker should be high. The combination of these two markers should be high and these three markers should be low. That’s it. That’s basically the explanation here.

 

Once we can detect this profile inside the cell, we argue that this is going to give us almost a hundred percent specificity. All right? Just by the mere combination of these six inputs into what we are calling an end gate. In theory we can extend this much further. We don’t know where the limits are. And so we’ve been able to actually embed this on the abstract level. The idea is that you take the classifier circuit. It goes inside all cells, and if it is inside a HeLA cell it’s going to have a match, and the match is going to result in the expression of an apoptotic protein. Jim showed you one of those. We’re using a different one called Bax. And if there’s no match, then nothing happens and the therapeutic agent goes away. The hope is that this is near a hundred percent. Very close. I’m not going to guarantee a hundred percent, but we don’t need a hundred percent. If we get 99.9 percent we’ll be celebrating up and down the streets.

 

So that’s a circuit and these are very quickly just experimental results. I just want to point out these are fluorescent results. Look at this. This is fluorescence in HeLA. You can’t see here fluorescence in two other cell types, and it’s good that you can’t see this fluorescence in two other cell types. We’ve recently extended this to four additional cell types and it seems to work quite well. And not only that, we now have also results where we have not just fluorescence as an indication of tumor versus non-tumor but also killing. The cell killing is actually also very specific. That’s why we are excited about the notion of being able to embed computation into these therapeutic agents.

 

Another area where we think synthetic biology is very important—we might as well raise a hot topic right now—is stem cell research. We believe the combination of synthetic biology—let’s bring together two controversial issues. I guess they negate each other and that means we should do both of them, right?

 

Stem cell research has obviously has been looking at—one application area is tissue engineering. We believe that by being able to embed synthetic circuits into these stem cells, not only can we make it safer but we can also add tremendous functionality and perhaps have it be able to do things that are not possible otherwise. So for example, severed spinal cords might be something you want to be able to address, right? It’s not clear that we can take these embryonic stem cells and embed them—and even provide appropriate matrixes and cues—embed them into a site of injury and have it be able to read the cells, be able to proliferate, and be able to regenerate the tissue as appropriate. Clearly stem cells can engineer or can create spinal cords in an embryo, but the question is can they actually do this in an adult patient. Right? And we believe— we strongly believe that in order to combat things like hostile environments where these injuries are taking place or various cues that exists in the very complex environments, we will need to genetically engineer stem cells in order for them to be able to recreate the appropriate tissues using things like engineered cell-cell communication, the toggle switches that Jim has mentioned to you, interfaces to differentiation pathways that we very precisely control.

 

So this, we believe, is a very important application area for synthetic biology. I will mention that if we can do this with non-embryonic stem cells, if we take fibroblasts and be able to take the fibroblasts and de-differentiate them using IPS technologies and use that in these injuries, fine. I’ll be very happy with that. It’s not clear that we can do that, and that’s why we’re spending a lot of resources on embryonic stem cell research, but maybe that’s the case.

 

And so we’ve been able—this isn’t bacteria. We have been able to engineer cells to communicate with each other and form a variety of different patterns, and hopefully some day we will be able to engineer patterns that are as complex maybe as an organ, an actual organ in the body. This is looking long-term.

 

I’m not going to show you all the results that we already have in stem cells, but we can engineer stem cells to become neurons, muscle cells, and a variety of other cells. So this is possible right now.

 

I’ll mention one more example. This other example has to do with diabetes. The 20-second introduction is that in diabetic patients in Type 1 diabetes, what happens is the autoimmune response slowly kills the insulin producing beta cells. Can we engineer a system that can actually address this problem? I’m not going to tell you everything about this. You’re not going to become experts in diabetes or this particular process, but the basic idea is that we would like to be able to put some kind of—maybe adult stem cells, embryonic stem cells into the pancreas, have them grow, have them understand how many beta cells are needed, and then regulate the production of beta cells as necessary so they can maintain the population of beta cells despite the attacks by the immune system. So if we can do that, that would be great. And that’s the algorithm over here. This should give you a headache right now. This is how we’re approaching this. This is a program, this is our bio-program. This is what we would like to be able to implement. This is a formal mechanism to describe behavior that we can then analyze, we can model, and so on. Please don’t try to read this.

 

This is another representation. This is a genetic circuit representation. We’ll take a program, we’ll encode it using something that’s going to be perhaps even more painful to look at, and then model that and so on. And we have been able to take steps along this process. We’ve have been able to engineer stem cells to at least produce insulin. So these are results where the engineered stem cells are now these beta-like cells that in fact are able to produce insulin. We can assay that with a variety of different methods, and we have been able to create many other mechanisms in these stem cells for communication, information processing, and so on.

 

So where do we see things headed in terms of applications? Near-term applications have to probably do more with microbial biochemical synthesis. Mentioned Artemisinin, other pharmaceuticals. Bioenergy production, I think, is something we’re going to—we’re seeing more and more now, has a near-term application. Environmental remediation, toxin sensing, explosives sensing are also things that we believe are going to take place in the next few years. The biomedical applications have more of a long-term impact, we believe. It is absolutely critical that we understand these right now. We experiment with these in the lab, so that within—I’m not going to X number of years; I’m not going to exactly give you the number X?? I don’t know what it is??that we will be able to actually implement these things in a clinic so synthetic biology actually has the ability to affect the human condition.

 

I will stop there. Thank you.

 

[AUDIENCE APPLAUSE]

 

Jim Wagner:

Ron and Jim, thank you both for presentations that not only get us excited about the future potential but show in a different way than we have seen before so the methodology around engineering design for that purpose. I think, Steve, are you going to—no, I think—or you can yield to Raju, who is ready to go.

 

Stephen Hauser:

Raju, do you want to start? And then I will follow you.

 

Jim Wagner:

And then I want to get out of the audience right away. And our second question, if audience members that have a question will get to a microphone, we’ll be sure to call on you.

 

Raju Kucherlapati:

Thank you very much for your presentations. I wanted to understand a little bit from your perspective as to what the boundaries of synthetic biology are. Jim, you know, you talked about your work for many years now about basically taking known biological modules and putting them together and trying to create the switches, and you have talked about that in synthetic biology. Ron gave a definition of basically being able to synthesize the nucleic acids and using the nucleic acids to construct something in synthetic biology.

 

How do both of you think as to what is the boundaries—is there a discrete boundary for synthetic biology, or is this part of a continuum of molecular biology that we have been practicing for the last thirty years?

 

Ron Weiss:

I would actually say that it’s more of a continuum. I think that trying to draw sharp boundaries between synthetic biology and genetic engineering and systems biology and other mechanisms is perhaps useful in the press or something like that. But from what we do in the field itself, what we do in the lab, is we integrate all of the mechanisms together to approach specific problems to make the engineering of biology easier. I think that it’s important to understand the emphasis of synthetic biology from different perspectives, and things that, for example, Jim and I work on and various other people work on often has to do with the emphasis of the system integration, and this is one of the things that I mentioned. And this notion that we’re trying to find rigorous ways by which we can take these genetic parts, we can assemble them into modules, and then we can take these modules and assemble them into larger systems that perform prescribed functions. So I think it’s the methodology that we’re approaching, the desire to create an engineering discipline, that has not existed before that perhaps differentiates synthetic biology from its related activities.

 

James Collins:

I tend to agree with Ron. I just reviewed a paper last night where your first meeting was quoted where Dr. Wagner asked Drew Endy to define synthetic biology, that your commission heard five definitions. And Drew responded, "I don’t think you have them all." I don’t think it’s a big problem for our field. I think it is for systems biology.

 

On the synthetic side, I think it’s generally the notion of taking engineering approaches at the molecular cell level. The distinction with systems is two-fold. One is it tends to be more bottom-up than top-down, and two is that while on systems you’re generally looking to uncover the natural structure. In synthetic biology there always is some element of artificiality, so that you’re either modifying a gene, a protein, a receptor, or constructing circuits or pathways that on their own don’t exist in nature or have now been introduced into a different organism.

 

Raju Kucherlapati:

That’s fair.

 

Jim Wagner:

Steve?

 

Stephen Hauser:

I might ask Ron but also James. Perhaps focusing on the clinical applications, as you think of the first forays outside of the laboratory and into the clinical environment, once one has the requisite cellular and preclinical in vivo animal safety data, aren’t there special considerations, at least over likely short-term applications, that would not be covered by current regulatory controls that have been well established for human research and things that this commission should consider as synthetic biology moves to deployment in all sorts of applications?

 

Ron Weiss:

I think it depends on your perspective, again, of synthetic biology and whether it acts as a continuum. Genetic engineering has clearly been around since the ’70s. There’s been numerous panels to discuss genetic engineering. I think that synthetic biology provides this rational approach to being able to integrate multiple, for example, genetic parts and genetic engineering capabilities into more reliable systems. So from that perspective, I think many of the existing frameworks to understand clinical applications—the clinical applications in some sense is easier. The existing frameworks provide fairly good starting points to understanding them. What are we trying to do in synthetic biology? We are trying to enhance those basic capabilities. We are trying to make them more specific, perhaps trying to make them safer as well by integrating additional mechanisms. And so in some sense it’s not a fundamental change in our basic ability to manipulate the nucleic acids and so on, but it’s more about encompassing these things in a framework that allows more safety, that allows the ability to understand more of what is going on inside the cell, more of what’s going on with respect to the multiple biomarkers, more trying to understand what systems biology, for example, has to offer. So if systems biology is able to offer us some hypotheses about what’s going on in a complex disease. What synthetic biology can do is then take these hypotheses and be able to use them in a clinical, therapeutic application.

 

James Collins:

I think on the clinical end, the current regulations are suitable. I think that for the most part, the great, great majority of our efforts in synthetic biology are also well covered by regulations that were put in place for early efforts in genetic engineering. I thought about this a fair amount to see where could things become problematic, and I think it’s important for the commission to appreciate that we are still taking baby steps in this field, and that while Ron and I can go up and present these interesting talks with results, they often take three, four, five years to accomplish something as simple as getting a two-gene network to work.

 

What I think it’s interesting to consider is, as we engineer new properties into organisms, could you have the possibility where you move from a BL-1 organism to a BL-2 or a BL-1 to a BL-3, and the investigators were not prepared for that shift. So one which doesn’t address a health concern but maybe environmental concern would be, say, if you are going to go after engineering an organism to degrade plastic, which would be a great environmental find. Imagine somebody actually hit upon something quite readily via either large-scale mutations or direct engineering, rational engineering, and now you have something that degrades plastic very quickly. Well, if you are in a BL-1 lab, if you have this it might end up on your lapel and you take it home, and when you wake up you look out and your car is no longer in the driveway. What’s left are the few steel parts are still in modern cars because the organism grew and ate everything in your car, which would be an untoward effect that was unanticipated.

 

[AUDIENCE LAUGHTER]

 

Amy Gutmann:

The only movie that I can remember my life walking out on because I was terrified when I was a really little child is a movie called The Blob. This is the biodegradable blob. Yes. Okay. Dan?

 

Daniel Sulmasy:

Thanks for two presentations which I think really did give us some exciting insights into the biological applications and potentially even clinical applications. I have some questions I would maybe would ask at the break about some of the science, but here a couple of potential worries. Where this, I think, differs from classical engineering is the biological systems are dynamic, right? And I guess there’s a wonder about whether you have seen yet anywhere in the field that some of these systems have evolved past the engineered changes you have made. Have the systems evolved so that you build in an apoptotic mechanism, and just as bacteria are smart enough to become resistant to antibiotics they become resistant to your change.

 

And second, you know, the control of gene expression is all of what this is about. But of course most carcinogens that we know actually affect gene expression. That’s the effect promoters and repressors. And so I wonder in the mammalian models, for instance, if there’s been any tumorigenesis, if anybody has done it past a cell, put it into an organism, and seen any of that. Again, thinking of these out not to the Blob but to some serious kinds of questions, what kinds of precautions we could build in in thinking about these possibilities.

 

James Collins:

I will touch upon your first point. We have seen things evolve in the lab, and it usually goes toward non-functionality. So we have not seen natural evolution towards something getting better or behaving more like we’d like it to.

 

Stephen Hauser:

Did you say usually or always?

 

James Collins:

Always. Always. So I wish I had a usually. But it has always evolved to its non-functionality. So, for example, even in the kill switch the notion would be that these organisms want to live, and so they’ll evolve to shut off the switch. Now, the concern goes—related to that would be that one way they may evolve is, to get rid of it is sending it out. They’re going to drop it into the environment. There, I think, is where precautions need to be made as you begin to think about engineered organisms out into the environment, which are already there. But as we ramp up our schemes, the precaution is to make sure that you’re not sharing material between these organisms and putting in appropriate schemes for the stability for the constructs as well as potentially reduce mutagenesis effects.

 

Amy Gutmann:

Can we go back, though, to Steve’s question on the basis of your answer to Dan’s? Are there regulations in place that would guard against the bringing out into the environment something that’s only now in the lab that might have detrimental effects?

 

James Collins:

I believe there are existing regulations.

 

Amy Gutmann:

I mean, if you don’t know, it’s acceptable to say I don’t know. I just wonder if you do know.

 

Ron Weiss:

I have been involved in one of these NIH panels also. I do believe that there are existing regulations for the kinds of work that you can do. An important debate in the panels has been, how does synthetic biology actually change everything? So, for example, when you look at these regulations, if you’re dealing with some particular pathogen or you’re trying to use genes from a particular pathogen, that can be construed as BL-2, BL-2-plus, BL-3, and so on. What we have been struggling with is, how does synthetic biology actually change this? It’s not clear. And so for example, if you abide by the existing regulations and you make sure never to use—never to incorporate that toxin gene, what else can your synthetic circuit do that would raise it from what—for example, take a whole bunch of BL-1 parts. Can you construct a clever mechanism by which it all of a sudden it becomes a BL-2 or BL-3? I haven’t been able to come up with—a design like that might exist.

 

Amy Gutmann:

Jim, did you want to—

 

James Collins:

My understanding is that the EPA does have current regulations in place that are suitable for what we’re doing. I think we’ll hear from Dr. Brenner tomorrow where—having heard Dr. Brenner speak before, I think he’ll speak to really we’re undergoing an evolution, not a revolution, in terms of this—the capabilities that we have as a result of the things that Ron, myself, and our colleagues are doing.

 

Amy Gutmann:

We have a letter from that EPA states just that. I think we’ll want to delve a little bit more deeply into not only whether the regulations are there, which I think the answer is yes, but whether the oversight on the basis of the regulations are there.

 

Ron Weiss:

Let me quickly come to a point with respect to the notion of evolution. You have to look at the particular organisms as well. So certainly in the lab we have seen bacteria evolve rather quickly, and we’re now exploring as part of our synthetic biology projects mechanisms to integrate fault tolerance. Just in the same way that you want to build fault tolerance into computing systems, not only would you want to have one kill switch but you actually want to have multiple kill switches and bring the safety to an acceptable level. When you go to mammalian applications, the mutation rates are much, much slower. So if we can build systems for clinical applications where the cells are not modified for 20, 30 years, we’re pretty happy. So we—there may be different issues there.

 

Jim Wagner:

Last quick question, Nita.

 

Nita Farahany:

That’s actually quite helpful because I wanted to build just on this question about—so bacteria is, what, about 10 to the negative 8 is the mutation rate per base pair per generation? So If that’s right, if you have multiple ways of doing it, then the likelihood that you’re going to have a single bacteria even in a novel environment that has changes overcome its kill switches or toggle switches is probably pretty low. So in the systems that you are presenting and that you are also speaking of, is that where we’re going, is that there are multiple redundant systems in order to prevent a mutation from being able to be on uncontained blob in the environment?

 

Ron Weiss:

So we have a project right now where we are actually looking at whether they have mutated. It can be done either inside the cell by improving the error correction of DNA replication but also in communities of cells, where if a single cell starts deviating from its neighbors, then you can have the cell execute its own kill switch, multiple kill switches, and so on, so you can detect the errors very quickly. So now you may be able to go to 10 to the minus 8 to 10 to the minus 12, 10 to the minus 15, and the combinatorial effect of those is perhaps multiplicative, and I think it can get you to perhaps acceptable levels.

 

Nita Farahany:

And would you recommend that that’s—given the low error rate, do you think—because EPA regulations don’t go to this now. Do you think it’s necessary before there’s an environmental release of something from synthetic biology to ensure there are multiple systems to prevent mutations from being able to become the Blob?

 

Ron Weiss:

A definite maybe. I think that has to be studied in a particular context, how that relates to existing approaches for releasing organisms and what the scenarios are for these enhanced organisms and whether they will tend to become a problem as opposed to just becoming non-functional, which is perhaps the more likely scenario.

 

Jim Wagner:

Gentlemen, thank you, from a personal and professional perspective. I take a certain comfort in this notion of rational engineering approach to these things, although, of course, discovery is another avenue that takes us to advancement as well. Thank you for spending your time with us today. We appreciate it. Thank you.

 

Ron Weiss:

Pleasure. Thank you.

 

[AUDIENCE APPLAUSE]

 

Jim Wagner:

Yes, I think although the second session got everything done in its time envelope, we never—let’s take a five-minute break and reconvene.