NHGRI Oral History Collection: Interview with Stephen Fodor

Stephen Fodor My name’s Steve Fodor and currently I head a little biotech group in Palo Alto where we’re exploring new technologies. You know, it wasn’t a straight path. And you know, my dad in the early years – I grew up in the late ‘50s, early ‘60s in Seattle. My dad was a physician and so he always, you know, made sure that there were, you know, microscopes and telescopes and chemistry sets and would bring me home blood smears from his office and stuff like that, x-rays. So, we had a house that had a lot of science in it. So, you know, I sort of did that. I actually – when I graduated from high school, I was not going to go into science.

01:05 - It was in the late ‘60s and so I was actually somewhat of an anti-establishment type rebel and decided I was going to go live out in the country. And so, I did not go to college right after high school. I went to Eastern Washington and actually worked on a potato farm and grew potatoes for a year or two with a farmer there. And one day, I was with the farmer and, you know, we went through the whole process of planting, irrigating, harvesting, and sorting all the potatoes, putting them into bags. And a guy in a truck drives up and the farmer sells them to the guy in the truck for 35 cents a bag and he drives them to the store and sells them for 70 cents.

01:55 - And so, I said to the farmer, I said, “Well, why don’t you buy a truck?” And he didn’t want to step on the middleman and all this. And so, you know, I thought to myself right then that, you know, if this is really the way the world is, you know, I should go to college. And so, I was going to be a farmer. That’s what I thought I was going to do. And plus, I was pretty tired of working hard labor. But I went to college and then, you know, started to take some agriculture classes and got totally disillusioned with the courses. They sort of taught by repetition. And so, I was going to drop out. And my dad told me, “Well, don’t drop out.

02:43 - Just take courses that are interesting to you.” And so, I started to take a lot of biology, chemistry, and math courses and got very much interested in science at that point. Well, you know, I came to the end of my undergraduate work and had to declare a major. And so, my first major was in biology actually. And part of that was to complete a senior project, senior thesis-like program.

03:12 - And I met a guy named Arthur Cohen, who actually ran the electron microscopy section at Washington State, which is where I was. And, you know, they had an electron microscopy facility and electron microscopes that were totally accessible to undergraduates. And so, I started to learn how to use those. And it was just fascinating being able to, you know, take a sample that you could see with your eye and then, you know, end up magnifying it 700,000 fold to get down to pretty much the limit of what you could do with electron microscopy, which was in the 10 angstrom or so, you know, 20 angstrom range back then. And so, that was fantastic to me. And then I met – I was looking for my project, some DNA because I wanted to try this technique that was called the Kleinschmidt Technique at the time, which was basically floating DNA on the surface interface between a layer of cytochrome C, I believe, and the meniscus of a bubble and put it again on an electron microscope grid.

04:37 - And I started calling different people and I ended up finding a guy names Keith Dunker, who was in the biochemistry department at WSU. And he gave me some samples of bacteriophage FD or M13, which is the same thing. He gave those to me and said, “Come on back if you get anything.” And so, I did them and I got these gorgeous micrographs and I brought them back to him. And as I walked in and showed it to him, he said, “Do you want a job?’ because apparently, they had been trying for months to get some good photographs. And so, he got me interested in this.

05:19 - I stayed at WSU and got a master’s in biochemistry and biophysics. During that period, I also met a fellow by the name of Paul Stein, who was a visiting scientist there from the east coast. And Paul was sort of an expert in laser resonance Raman spectroscopy. And so, I started to learn from him. This was in the days where the first sort of commercial research lasers were becoming available. And so, you know, he taught me a lot about how to work with lasers, how to do some spectroscopy.

06:02 - And that became the basis for my work when I applied to graduate school. At graduate school, I went to Princeton in Tom Sparrow’s [spelled phonetically] lab. He was an expert in resonance Raman of hemoglobin. But we also at the time – he had been getting some of these new pulse lasers, these yag-lasers [spelled phonetically], at the time. And he wanted to do some new work in the ultra-violet realm.

06:40 - And so, I started to work with actually a guy named Rich Rava, who I became close friends with and worked for many years. But we had some of the first ultra-violet laser light at the time for again, for biomolecular research. And we did some of the earliest work on proteins and nucleic acids by excitation of the chromophores, both nucleic acids and amino acids, back in the early 80s. And so, you know, Tom Sparrow, whose lab we were in, his – that was my first exposure to a very large lab. There were probably a dozen graduate students and probably a dozen post-docs. We had a big facility.

07:36 - But it was there that I really learned, you know, starting to do some innovative things had a big payoff. By doing these first early experiments on nucleic acid proteins, we then put in some grant applications for Tom. And so, that started a whole new funding cycle for Tom’s lab. And you know, these were actually, I thought, pretty easy experiments. Because when you get to do it for the first time, you always break a lot of ground. So, you know, I very much enjoyed it. I always enjoyed working with my hands.

08:10 - I enjoyed experiments and making things work. And so, this was a big lesson for me, you know, as well as moving from a place like WSU to Princeton. The exposure was much, much larger at Princeton. But the exposure to the larger lab, the lessons that if you’re innovative, you develop a new technique or new data, it opens new doors for you. And so, that was a real eye opener for me. So, at Princeton, when I completed there, you know, we had done some very nice work. I was applying for post-docs and I really wanted to go out to a guy named Rich Matthew’s lab at Berkley. And so, I applied there, and Rich said, “Yeah, you can come.” And but, the moment I got there he said, “Well, you have to get your own money now.” And so, I applied for an NIH fellowship at that time.

09:05 - And I was working – at that time, I had already had experience in pulsed lasers, that experience with generating, you know, ultraviolet lasers and light. And Rich had been working on time resolves spectroscopy, particularly of bacterial and plant pigments. And so, I took on a couple projects. One was in bacterial rhodopsin and another halo rhodopsin and another sensory rhodopsin and finally phytochrome, one of the plant pigments, over my post-doc time there. But we – I’d learned a lot more about very, very high sensitivity detection at that point. And you know, and this is a little bit of a prelude to what we’ll get into, but I was working in Raman spectroscopy, which is orders of magnitude less signal that you get, for example, with fluorescence spectroscopy.

10:03 - And so, I’d learned, you know, on probably some of the hardest problems at the time. And in addition, I started to get a good background on photochemistry there. So, you know, it was a combination of being able to work with light, being able to work with biological molecules, understand high sensitivity detection, and also photochemistry. After working on that for a while, my ambition was to go into academics. So, I applied to a few academic spots and got a couple of interviews.

10:43 - Turned out to be number second choice in both of those. And so, I was going to stay for another year at Berkley and then reapply. And it was during that period that I actually got a call from Lubert Stryer, who was at Stanford, and of course Lubert is a very famous scientist that not only wrote the textbook on biochemistry that was used by practically every medical student for years. But, you know, Lubert had done this very nice work in the biochemistry and photochemistry of retinal, of rhodopsin in site, in vision. And in addition, had done some very innovative things with avoid strep dowden system for fluorescent tagging, done some work with the – he was very well known for his work with the forcer energy transfer experiments of measuring distances by energy transfer with fluorescent molecules.

11:44 - And you know, so when Lubert called, I listened. And we chatted, but, you know, he said, “Well, I’m going to take a year leave of absence from Stanford to start this little company called Affymax and I’d like someone with your background to come joint me.” He said, “Look, you know, if you stay for a year, if it doesn’t work out, then, you know, we’ll see about, you know, where is a good spot for you,” sort of thing. And so, I thought, “Well, okay.” And so, I went down to this little company called Affymax. And you know, I have to say my motivations were really to go and do some science.

12:26 - My motivations were not at all to create a new company nor to necessarily even go after what the ambitions of the company were. I was invited by Lubert and you know, the prospect of doing some great science I thought was great. But when I got down there, there was this guy named Alex Zaffaroni, who was a well-known biotechnology entrepreneur in Silicon Valley. He had worked on companies like DNAX, earlier – I’ve forgotten the name of the company – Syntex, that he and Carl Djerassi and others did. And of course, they came up with a lot of steroid compounds and things that were used for birth control and so on.

13:14 - He started a company names ALZA, which did skin patch technology for slow release of things into the bloodstream. And he had started – he started this group called Affymax. And so, his idea was – now, remember, this was late 80s. And at the time, you know, people had now begun to clone genes and be able to get receptors and so on in pure form. And so, his idea of Affymax – and that’s where the name Affymax comes from is from this concept of the affinity matrix.

13:52 - And the idea was that if you had all of the receptors, for example, in pure form and you could clone these and isolate them in pure form, he’d like to see ways were you could assay that, those pure receptors against a whole bunch of chemistry. So, where there’s natural products, synthetic, chemistry, collections, whatever. And so, he called this the affinity matrix because you want to measure the affinity of receptors against chemical diversity. And so, we started to play around with a whole bunch of ideas. And as I mentioned, some of them were things like chemical collections, things like natural products.

14:38 - We were introduced to a whole bunch of really good characters at the point, guys like Josh Lederberg were there thinking about cauldron chemistry, what he used to call cauldron chemistry where you would, you know, cook up a tar like from the tar pits and see what kind of magic compounds were hidden in there. I mean, just some really great fascinating stuff, remembering that this was actually a long – you know, quite a while ago before much of this was very popular. One of the ideas, a guy named Leit Norid [spelled phonetically] said, “Well, gee” -- he was infatuated by semiconductor world – “isn’t there some way that we could use light to direct chemical reactions and put this together?” And I thought that was – Lubert and I both thought that was a fascinating idea. And so, I started thinking about how to do that. And some of the chemists had some interesting ideas about photochemical protecting groups and so on, Michael Parongal [spelled phonetically] was one of them, Dennis Solis [spelled phonetically] was another, Pete Schultz was coming in at the time and had some interesting contributions.

15:45 - And so, you know, I just started to do some really simple experiments. And, you know, how could you do, you know, light-directed synthesis in very precise locations at a microscopic level on a solid surface? When I first went to Affymax, I was somewhat horrified because, you know, it was a startup. I’d been used to now being in pretty well-funded labs. And I went in there and I was given a desk and empty rooms. And I had to call up on the phone and order, you know, pipets and big tips, and chem wads.

16:24 - And I thought, “Oh my god, you know, I’d just committed career suicide. I’ve gone to school for the last, you know, freaking 25 years or something and now, you know, what have I done?” You know, I started to do some really simply experiments. Built a very, very simple apparatus in the beginning which was basically a light source, couple mirrors, and took some – I did some things like printed a checker board on a piece of paper on a laser printer and then took a picture of it and then had the film developed and I would use that as a lithographic mask and shine the light through it and image that down onto surfaces. And started to do some very basic experiment derivatizing the glass, attaching and working with some chemists to attach these photochemical removal groups onto the covalently linked surface molecules and demonstrating that indeed we could excite the allieving [spelled phonetically] groups in different areas on the glass and well-defined areas at different resolutions. And actually, in a way that was not available to you by printing normally, you know, by putting pipets down or doing something physical.

17:45 - We could then generate designs on surfaces or patterns on surfaces and direct chemistry through the use of light on a surface. And so, we started to work on that, started to build some things on the surface. Started, as you might expect, because of working with Lubert and because of my background, began to use fluorescence as a read out. And demonstrated very quickly that we could start to synthesize molecules on a surface. And we used some very simple peptide recognition systems in the beginning.

18:26 - We would, you know, synthesize a little peptide that we can an antibody that could recognize. We could fluorescently tag the antibody and you could show now you could spatially resolve a synthesis compounds. So, I ended up, you know, building then – we had to have a reader. And so, I ended up building a laser through a microscope down to a surface with xy stages. Then started to do raster scanning and so on so you could build up a two-dimensional image of the fluorescence map, if you will, of a surface.

19:03 - And by the way, none of these type of experiment equipment existed back then. You know, there were some epi-fluorescent microscopes that you could look at very, very small fields of view. But nothing where if you wanted to scan, you know, a couple centimeters by a couple centimeters you could do. And so, I built these systems. And then continued to work on making, you know, compounds in different places. We then kind of stumbled upon – well, say you wanted to build a matrix of many, many different types of compounds – because that was the original ambition -- you know, how could you build millions of compounds on the surface? And it became very apparent that well, you can’t really do it by using millions of steps.

19:50 - And so then the question, “Well how do you actually build large sets of molecules?” And so, we came up with some primitive ideas on how to build some overlapping sets. And Lubert actually had some great ideas on how to build a sort of a, you know, a constant set of oligonucleotides or peptides. But it was actually the first experiment that we were doing where, you know, we now, you know, by hand drew out some lithographic masks. Fabian Pease is another guy, by the way. I should bring this in. When we started to put together a lithographic system in order to make the first, what are now called microarrays. We went down to some old junk shops in Silicon Valley and bought some used aligners that were used to make printed circuit boards and so on and brought this up to our little office building in Palo Alto.

20:46 - We had to chop out some weeds and so on to get this into a room. But we had set it up now with an old lithographic aligner that was used in the semiconductor industry. I had hand drawn lithographic mask designs that could be used to do some pattern work to synthesize our sort of first arrays. And we had bought, at that point, an ABI synthesizer that synthesized, you know, one oligonucleotide at a time and taken it apart, but used that as a chemical reagent dispenser and plumed it over to the lithographic instrument with a holder in microfluidics and so on. And so, we had all these pieces put together.

21:36 - And then of course, as I mentioned, we also had the first scanners that were in the other room ready to scan this thing. And so, as we started to do the first photolysis, you know, we were going to lay down a particular building block of an amino acid. And then as I turned the micrometer to go to the next region where I was going to do the photolysis, I realized that, “Well, if I overlapped it on the very next step, not only would I put compound A down and compound B, but I would also build AB.” And so, I went into Lubert. I said, “Wait a minute, you know. There’s something we don’t understand. There’s this whole field of how do we do these overlaps in order to generate lots of chemical compounds.” And so, it turns out this got us on a whole path of really thinking about, you know, how many compounds can you make in how many chemical steps.

22:38 - And so, we had some great discussions about it, and it dawned on us that this was really a very simple binary process. And we had stumbled on to, you know, a formalization of the world of combinatorics, combinatorial chemistry. And so, we formalized this and it turns out that you could make two to the n compounds in n chemical steps, which means, you know, if you want to make a million compounds, no it doesn’t take you a million chemical steps. In fact, you know, two to the tenth is about a thousand. Two to the twentieth is about a million. So, it actually turns out that in about twenty chemical steps, you can actually make a diversity of about a million. And it just grows exponentially. And we would actually formalize this.

23:23 - We put it into matrix algebra and figured out how to create the lithographic masks from the matrix algebra. So, it was very rewarding and a very exciting thing. There’s a combination of things that goes on. So, one is at that time and that place, you know, we actually set out to invent something, I think. Because we said, we want to have something that can do the following. And at that time and at that point, we had a nice bolus of money that was not earmarked specifically for particular things. So, I had a tremendous amount of freedom. You know, I could try this experiment. If it didn’t work, I could try something else. If I wanted to, you know, spend $100,000 to build a nice scanning instrument, I didn’t have to ask anybody for permission. I didn’t have to write a grant. I didn’t have to wait for them to give it to me and to go through the whole review process. And we can talk about that later because that brings in a whole – it’s not that that’s bad.

24:27 - It’s just that if you want to move fast and if you’ve got an idea, you want to be funded in a way that is very, very flexible. And I think that’s it. And I think a lot of it is the funding. A lot of it is just the culture that you set up. You know, this was going to be a company. But its first main purpose was to invent some new ways to look at biology and new ways to screen chemical diversity. And so, our purpose at that point was to not make money. Our purpose at that point was to open a new field and to develop new technology.

25:06 - And I think that’s – it’s academics at the time, that would have been hard. Because I think you could -- and even today, it’s probably kind of tough. There are some places that give you great startup funding and you maybe have a little window to do that. But, you know, it kind of falls into two buckets these days. One is you have a tremendous amount of freedom and not much money. Or you have a big project to undertake and so you’re extremely well-funded. And then people figure out, “Well, how do I leverage that into having a little bit of freedom?” But this was a lot easier back then in many ways because this was a brand-new area that nobody really had any results in. So, you know, as we started to work on this -- one thing I should have said is that it should be obvious right now that we took an approach which was a combinatorial approach, which was a building block approach. And of course, we took our inspiration from nature. Nature creates chemical diversity through the combinatorial assembly of a limited set of building blocks. It does this in the nucleic acids. It does it in sugars.

26:17 - It does it in peptides, amino acids to peptides to proteins and so on. And so, you know, we took this building block approach. And the two obvious building blocks were peptides and nucleic acids. Both of those chemistries had been worked out the solid phase chemistry of both of those important chemistries. As I started – so, for the science paper that came out in 1991, a lot of the recognition because, you know, we had the reagents and so on, was done on peptides.

26:55 - But in that paper, we also worked out the fundamental chemistry and methods of doing nucleic acids. And it was just – it was very limited, but we enabled the chemistry in that paper. And actually, in the abstract of that paper, as you’ll read – and this is what got me very excited – was in the early – in, well this was 1990ish now at this point. The Dramanis [spelled phonetically] and Zircunyakov [spelled phonetically] as well as Andrei Mirzabekov had come up with this idea of sequencing by hybridization. And the idea was, well let’s say I could make a complete set of all the eight-mers.

27:42 - So, that’s 65,436 or something like that, you know, 64k. And if I could make all of those eight-mers, I should be able to add those eight-mers to an unknown strand of DNA and then depending on which one bound, I could tell you what that sequence was. And so, this was also in the context at the time that everyone was talking about doing the human genome. And wouldn’t it be great if, you know, we had new sequencing methods and so on. And of course, there was Sanger sequencing, which was great and extremely accurate and still a gold standard today.

28:17 - But this was an idea of doing something new and very innovative. And so, you know, as we started to develop out our chemistry and realize that we could make these matrices of high-density chemical compounds, it became very obvious that one of the things we really wanted to do was nucleic acids. They also had a certain advantage because there’s only four building blocks. Whereas proteins have 20, okay? And so, your number of reagents for proteins goes way up and the medicinal chemists, of course, came up with all kinds of new building blocks all the time they wanted to add into your repertoire. So, I started to get to this position where Affymax wanted to be a drug discovery company. They wanted to go on.

29:04 - They wanted to explore peptides, natural compounds, natural products, and eventually be a pharmaceutical company. But, when I started to develop this technology, to me, it was exceedingly obvious the right home for this was in nucleic acids because there were only four building blocks. And then when you looked at something like this SBH technology, sequencing by hybridization, they only wanted to make 65,000 of them. Well, that was a piece of cake. So, I knew that was going to be a piece of cake. And so, you know, this was all theoretical, you’ve got to remember.

29:38 - It was a theoretical construct that we should be able to do SBH. And here, I’d been developing this technology and I read these papers and I said, “Well, hell, we can do that. That’s a piece of cake.” And so, I started working on it. Now, that didn’t go over so well with everybody at Affymax. You know, me, I was still – you know, I still appreciated the problems they had. You know, but let me tell you that, you know, I started I think in July of ‘89 and we had the cover of Science about 18 months later. So, this was very fast moving. I mean, it was – the experiments worked.

30:17 - It just went click, click, click, click, click. And then I wanted to turn all my attention to nucleic acids because I just thought that was the future. And you know, everyone remembered there was no human genome. There was – people were still arguing about whether they should do the human genome project in 1990. And so, through this interaction at Affymax, I had come to make friends with a number of different people including, Ron Davis at Stanford, and told him about what we were doing, and Paul Berg who was also at Stanford.

30:54 - Now, Paul was I believe on the early council or advisors or something to recommend the human genome project. And in 1991, after the Science paper, he invited me to come give a talk in front of – you know, there was Jim Watson and some – I think this was in Concord or something. But it was a huge panel of people that wanted to go after, you know – that wanted to sequence the human genome. And so, they were very interested in alternate sequencing technologies. And I’ll never forget this because Paul invited me to go speak in front of these guys.

31:39 - And you have to remember at this point, you know, I got spoiled very quickly at Affymax because they gave me – because this program was so successful, they basically gave me all the – Alex Zaffaroni gave me all the money I wanted. And although it wasn’t fitting into the ambitions and, you know, what everybody wanted to do from a business perspective at Affymax, the science was just so cool that Alex just, you know, protected me and made sure I had the funding I needed to pursue these things. So, when I went in front of the panel for the human genome project, I showed what we have done, and I also showed them what I thought we could do in terms of building arrays of nucleic acids. And I remembered Jim Watson at the time – and you have to remember, this was – it’s very commonplace now. Everybody knows what an array is. Everybody is used to array formats. But back then nobody was because it was all brand new.

32:41 - And so, I had these nice scans and pictures of these fluorescent maps of these compounds. And I was showing this, and Jim Watson was watching this and watching this. And he finally stopped me in the middle of my talk, and he says, “Wait a minute. Wait a minute. Is this real data or are these just pretty pictures?” And you know, I was very indignant at the time. I said, “Just pay attention.” You know, and so, you know, he – if you know Jim, you know he was very crotchety about the whole thing.

33:08 - But here were all these guys that were very serious. So, I was somewhat – I was completely a newbie at this and very naïve but coming at it from the technology side and from the sort of science of technology side and knew just very clearly I had a very, very clear vision that we could accomplish this. And so, you know, we talked about it there. I was just barraged with questions from the group and people wanting to know, you know, “How are you going to solve repetitive elements?” “How are you going to solve the branch structures?” all this sort of stuff. I actually remember Jasper Ryan was there and he asked some very pointed questions, but not from an evil perspective, you know, from a very constructive perspective, but a very hard question. And I basically said, “I don’t know.” I said, “I do not know.

” 34:01 - All I know is that we will make these things and we will test these ideas. And that was my attitude. Uou know there was, complicated politics around the whole genome project and all the people involved. And everybody had their own idea about the directions they should go. But, you know, I was looking at it from a little bit different perspective. Because by then, you know, I’d really realized, you know, for example, you know, we knew that we wanted to make things on the order of 30 bases long because they gave very nice stability and you could actually do single base pair mismatches with something of that size.

34:44 - And so, it was – but 30 long, you know, that’s a number like 40 to the thirtieth. I mean, that is a really, really, really, big number, okay. And you know, two to the sixtieth. I’m not even sure how large that number is, but it’s ridiculously large. But the thing is that combinatorally, you could make all of 30-mers in four times 30 steps or 120 steps, which means you could make any subset of them. And so, I started to think about this more as an information technology that what would really happen is that the genome would be sequenced and then we could copy basically the genome down onto these chips.

35:25 - And then you’d have a format to look at many genomes. And that was the direction that I was going. Well, there were two. There was one, you could make these complete sets of eight-mers, 10-mers and so on and say, could I do sequencing by hybridization. And then you could also make arrays that were designed as basic memory devices against certain pieces of DNA. [end of transcript] .