In a paper published in Nature in 1998, Paul Rainey and Michael Travisano showed that identical populations of the bacterium Pseudomonas fluroescens diversify morphologically when provided with ecological opportunity, but show no divergence when opportunity is restricted. Further, the morphs follow a predictable sequence of evolution maintained by competition. These results provided support for the idea that new designs could evolve rapidly through just the processes of mutation and selection. Nineteen years after the paper was published, I spoke to Paul Rainey about his motivation for doing the work presented in this paper, memories of lab work and the collaboration with Michael Travisano and what we have learnt since about adaptive radiation in Pseudomonas bacteria.
Citation: Rainey, P. B., & Travisano, M. (1998). Adaptive radiation in a heterogeneous environment. Nature, 394(6688), 69-72.
Date of interview: 17 March 2017 (via Skype)
Hari Sridhar: What was your motivation for this particular set of experiments and this paper?
Paul Rainey: It’s is a long story. It started out in 1985 when I was a Master’s student at the University of Canterbury. It wasn’t published till 1998! It started, as actually a lot of things that I’ve done, by curiosity. I had begun my Master’s thesis in New Zealand. The thesis topic was related to the mushroom industry. It was a very applied problem. It concerned the role of what’s known as the casing layer. When you grow mushrooms, the white mushrooms, first, a compost is prepared of, typically, straw and horse manure – it is quite a complex process – to be ready for inoculation with the mycelium of the fungus. The mycelium runs through the compost. It takes about two weeks. If you do nothing else other than let the mushroom mycelium run through the compost, you won’t get any fruit bodies; no mushrooms. To get mushrooms it’s necessary to put on a layer of peat and lime. This layer of peat and lime is known as the casing layer. Once you put that casing layer on, about a week later, you get lots of mushrooms. I was taken on to do a project that looked at alternatives to peat. Peat is a valuable resource, and also one that is often conserved, and so the industry was interested in alternatives. So, I did that. And then, it was also clear that it was necessary to have something more academic or basic, at least, to complement the applied work. So, I’d been reading around, and there was a curious report, I think from 1969, which concerned the reason that the casing layer was necessary to promote fruit body formation. The authors reported an experiment to test the importance of biological activity. So, they sterilized the casing layer and found that the sterilized material did not support formation of mushroom fruit bodies. Okay, so the obvious conclusion was that there is some living material, most likely microbial, that is in the casing layer and required for mushroom fruiting. And these investigators went on and provided evidence that it was bacteria of the genus Pseudomonas, specifically Pseudomonas putida. If you applied this bacterium back to the sterilized casing layer, fruit body formation would be promoted. So, interestingly, here’s a bacterium that is apparently necessary to promote fruit body formation in a mushroom. Actually, I think that was a wrong conclusion, but it set me off on an interesting route. I’m pretty sure that their observation is a consequence of having autoclaved the casing layer. I think that what triggers fruiting is the transition from a nutrient rich environment (compost) to a nutrient poor environment (casing layer). When you autoclave the casing layer, it liberates lots of nutrients, so the mycelium doesn’t experience the necessary environmental change and no fruit bodies are formed. Add bacteria though and they utilize the excess nutrients, converting the casing layer to a nutrient poor environment necessary for fruiting. So, I don’t actually think there was any very specific interaction. Anyway, that was a slight digression.
I’m trying to think…I can’t remember whether someone had developed an in-vitro agar plate based assay for fruiting earlier, or whether that was what I did for the first time. Anyway, I wanted to be able to try and reproduce this phenomenon in a simplified system, basically on an agar plate, where I would havemushroom mycelium growing and then I put the bacteria around the outside and I hoped I would see the bacteria promote fruit body formation. Actually, I did achieve some success on this – and that’s another story – but I found the results were variable. And I recall that there were other studies that perhaps tried to do the same,and they also found variable results. Sometimes, they got fruiting and sometimes they didn’t. So this seemed to be a bit of a problem, and, of course, I was curious as to why there should be this variability. Was it just that I had the wrong strain – because I’d isolated my own bacteria. I was completely naïve about this. Anyway, I’m now getting much closer to what sparked my interest in what turns out to be evolution in a test tube. I was using another bacterium, as kind of a negative control, Pseudomonas tolaasii, which is a mushroom pathogen, on which I subsequently did quite a lot of work on. Now, Pseudomonas tolaasii shows a very clear evolutionary response on agar plates, which you have to take notice of, if you are interested in studying the causes of pathogenecity. When you isolate Pseudomonas tolaasii, the colonies are almost white. They produce a toxin that you can easily assay and which causes a bacterial blotch disease in mushroom. Now, very rapidly on an agar plate, these white colonies sector. What you will see is a little outgrowth from the white colony that is translucent. So, you have a very obvious phenotypic transition between this opaque type and these spontaneous mutants that arise within the white opaque colony and have a selective advantage. They grow more rapidly. So you could see these outgrowths and these outgrowths are mutants that don’t produce a toxin. With this particular organism, this phenomenon of what we were referring to as phenotypic variation, is something you really have to take notice of. It made me wonder, in the context of the variability in fruit body formation, or the capacity of Pseudomonas putida to promote fruit body formation, whether Pseudomonas putida might also do something similar, but was a little more subtle, and perhaps I was working with the wrong variant. And it turned out that Pseudomonas putida also did the same sort of sectoring when you look carefully at the colonies, but indeed, it was less obvious. So at this stage, I became aware of the fact that, actually, a lot of bacteria show sectoring when you put them on an agar plate. It wasn’t at all clear to me what was going on other than you needed to take notice of it. And I guess I already then had a sense that these sectored types, these new types that arose, were actually mutants. If you took the translucent form of Pseudomonas tolaasii and streaked that out, it would breed true. The other thing that was interesting was that you would never see an opaque toxin producing variant arise from the non-toxin producing type. It looked like evolution on the plate. However, whether this was actually evolution on a plate was something that I really had a hard time getting across to other people, even though basic phenomenology is completely consistent with this, i.e. as these different types breed true. And the reason for this being a little difficult was, and I guess also my own thinking was influenced by, the emergence of a body of work about the same time, on a phenomenon called phase variation or epigenic variation in bacterial pathogens. We now know that what I was observing, and goes on to underpin the Rainey & Travisano work, is evolution driven by spontaneous mutations happening on an agar plate. The phenomenon of phase variation and so on that had been studied, and still is studied a lot in pathogens, also involves differences in colony morphology -that’s how it was picked up – on agar plates. However, one thing that very much marked it as different, in terms of basic phenomenology, was the capacity for the switch to go both ways. You could go from type A to type B and back again. So it seemed to be able to go both ways. We know that now, and it was emerging around about that time, that this capacity to switch backward and forward reflected highly mutable loci in bacterial pathogens. There are molecular mechanisms that allow pathogenic bacteria that interact with hosts, particularly the host immune system, to generate very substantial phenotypic variability in those outer membrane components, the polysaccharide proteins and so on, that interact with the host,within a host immune system, with the hope of avoiding recognition. Those are often referred to, rather confusingly, as programmed responses. But you know, what was meant was that there was a genetic machinery in the bacteria that allowed switching backward and forward between many different phenotypic states. So, I was looking at something that had some of the hallmarks of this, with these Pseudomonas strains, but I never saw them switch back, so I always felt it was probably different.
So, where are we…we’ve got Pseudomonas, my interest in fruit body formation in mushroom, and recognition that you’re getting, on an agar plates, what appears to be a fairly rapid evolutionary response to being on a plate environment. By this stage, I’d moved on. I remained at University of Canterbury for a PhD, and the role of these bacteria in fruiting became the topic of my thesis. An important step in terms of making progress was inspired by a chapter in a book in a series called The Bacteria. I think it was Volume 10. It was , I think, in 1986, and the author of the chapter was Jim Shapiro in Chicago. He had also noticed similar phenotypic variation, but in an even more dramatic sense. He was doing some early genetic studies with Pseudomonas, and if I remember rightly, he was introducing Mud-lac transposons into a Pseudomonas putida strain. If X-gal was included in the plate, then the β-galactosidase activity would allow him to see a blue color. What is frequently observed were colonies expression elaborate blue-white patters due to, presumably, movement of the transposon. I don’t know what prompted him to do this, but he was clearly also trying to figure out what was going on. The chapter contained a whole series of photographs of the different colony morphologies that he got. And these arose from experiments in which he would take one aged (sectored) colony from a plate, resuspend it in buffer. Then replate. So, let’s say he started off with Phenotype 1. When he saw the colonies that arose after a few days, let’s say, maybe three different colony morphologies, he would record this, producing long lists of lineages of colonies. Anyway, Jim’s work suggested to me the possibility of doing similar, but allow my colonies a period of growth in unshaken culture. So I did this with a range of different Pseudomonas and found much more variation in colony morphology that ever observed in sectors on agar. But it gave me the idea to use static broth test tubes and do the same. I did this, and I found that after a period of growth in unshaken both, the colonies that appeared on agar plates after dilution-plating were much more variable than I’d seen before. So I knew that if you put them into these static broth environments, you saw even more variation. I was curious; I didn’t have any answers. And so, I remained very interested in this phenomenon of phenotypic variation. I didn’t understand it for what it was. I didn’t understand it, really, in the broader perspective at all, but I remember being curious. So, I finished my PhD. I went to a postdoc in Cambridge where I worked on Pseudomonas tolaasii. I took with me my continued interest in this phenotypic variation. I was working mostly on the genetics and biochemistry of tolaasin toxin production, but I continued to explore aspects of this colony sectoring, but no major advances there other than describing behaviors and so on.
After a couple of years at Cambridge I went on to a second postdoc at the NERC Institute of Virology and Environmental Microbiology in Oxford. It no longer exists! I again worked on Pseudomonas and my primary task was to work with the isolate P. fluorescens SBW25,which has subsequently become the focus of much work in my lab and others. It was isolated at the university farm from the phyllosphere of sugar beet. The lab as a whole was, actually, trying to get some sense of the risk of introducing genetically modified organisms back into the environment. The thinking had been to go out and get lots of Pseudomonas from leaf surfaces or sugar beet,the focal plant, bring them back into the lab, type them all, and then choose one common type. That’s how SBW25 was chosen. My task during this postdoc was to map the genome of this organism. This as well before sequencing was routine. That was a huge project. Anyway, I was doing that, but I also had opportunity to get back to my curiosity about sectoring. Again, I could see exactly the same sort of behavior in P.fluorescensSBW25. And it was there that I began to make some progress, sort of doing obvious experiments, with hindsight. This was the early 1990s. One experiment that I did during that time was to grow SBW25 shaken and static. When I did that, I saw that I got all of this diversity in the static environment, but none in the shaken environment. That was a eureka moment. I guess that made me think much more solidly within an evolutionary framework. Well, it probably took me a while to get to that stage,but I certainly recognized that there was something fundamentally different about the environment, when you shook or didn’t shake. I was also aware of all these different colony morphologies, but I had no idea what they meant. And when I first began to see all of the diversity of different morphologies that came from cultures that had been propagated in static microcosms, I was simply overwhelmed. I had no idea how to think about it. And it’s kind of interesting, you know, you need to give names to things, in order to be able to make sense.But in this situation, where do you draw the boundary? How do you know what’s different to something else? Well, this is where I began to use these terms like ‘Wrinkly spreader’ and ‘Fuzzy spreader’. They seemed to convey certain classes that were distinct. At this stage, I didn’t do the obvious experiment – I’ll tell you what the obvious experiment was in a moment – but what I thought I would try and do was to capture something of the dynamic of the origin of these different mutants. I figured they were mutants because they bred true, but and that was something that was contentious for quite a while. I’ll come back to that. I set up dozens and dozens of microcosms, all founded by the same ancestral type, and every day I harvested cells from the microcosms and scored the frequencies of different types. Actually, I was very concerned about being misled by my categorization of diversity. So rather than scoring things on the basis of colony phenotype, I measured another phenotypic attribute of these types, which was the diameter of the swarm that they produced when I took these colonies and stabbed them individually into semisolid agar. I would harvest my microcosms, make serial dilutions, and the next day I’d have pinprick sized colonies too small to distinguish on the basis of their morphology. I would then laboriously pick a 100 colonies, from every single plate, with a needle and poke one at a time into semisolid agar, incubate them overnight and then I’d measured the diameter of the swarm. And I do that with multiple replicates on multiple days. Anyway, when I began to look at that data, it was absolutely amazing. First of all, you have to remember here that, to do this experiment, I was destructively harvesting the microcosms. So in order to run, say, through 10 days of evolution, I would have a set of replicates for each day that I would harvest.So every replicate is in independent, biological, evolutionary experiment. It was amazing to me that, not only did I have considerable agreement between my replicates, but through time, I could see trends. Things I came to subsequently call Wrinkly Spreaders were detected first on day two, and there’d be more of them on day three, more on day four, maximal at day five, and then they began to decrease. Fuzzy Spreaders would come up later and so on. So it seemed like a community had emerged.And indeed, in effect, that’s exactly what had happened. So once I saw this, and I began to think a bit more clearly, I did the obvious experiment.If I had done it a long time earlier, a lot would have made sense.The obvious experiment was to take these different morphotypes and to place them back in pristine microcosms and that’s when I saw – my goodness me – the wrinkly spreaders grow on the top, and the fuzzy spreaders grow on the bottom, the smooth ones grow in the broth phase: evidence of niche specialization. But that’s when I realized what I was dealing with here was, effectively an adaptive radiation. I was seeing the process by which a single genotype diversified into a range of different types that appeared to be adaptive. Of course, it all made complete sense that when I shook these vials, I imagine that I destroyed the niches and there was minimal diversity.
All of that was done, actually, even before I met Mike Travisano. My meeting with him is an interesting story too, and it is connected to Rich Lenski. As a consequence of a letter that Richard Moxon and I wrote to Science commenting on a paper that Richard Lenski recently published, Science said, oh, do you know that Richard is on sabbatical in Oxford? Well, I’d never met Richard Lenski and this was now an opportunity to meet up.We did and talked about the paper on directed mutation that he had published, and ideas that Richard Moxon and I have been developing on contingency loci that could provide an explanation in part for the directed mutation controversy, and my contribution arose from thinking about evolution as I’d been seeing it in my test tubes. I met with Rich Lenski many times during that time, and he was interested in what I was doing. I told him about these experiments and he was super excited. This was great for me. I was very excited about it. I guess his enthusiasm spurred me on, although it would still be some, years later before the experiments were finalized. Let me just tell you another thing for the record, because it’s kind of interesting. I left NERC and took up a position at the University of Oxford in 1994. So it was probably in 93 when I had begun to understand what was going on; that this was an adaptive radiation. We would have six-monthly lab meetings with the director of the NERC Institute. And on this particular occasion, I was the one who had to give a presentation. I think maybe there were two or three of us each time that did. I presented the progress I’d made on mapping the SBW25 genome, and then talked rather excitedly about what I just relayed to you. And, I would never forget it. The director just looked at me, after I’d finished, and was completely unimpressed. He said, “if you ever do this work again in my institute, you will be fired”. He said it’s purely phenomenological, there’s no value in it and continue it at your peril. I was pretty upset about that. It seemed pretty cruel; it’s not the sort of thing you would do to anybody. I could never imagine doing that to anybody,especially not if they were excited about what they were discovering. I remember going for a walk after that meeting, I walked down to the university’s wonderful natural history museum, and there was a display of Niko Tinbergen’s work, his behavioral imprinting and so on. Of course, he got a Nobel Prize for this work, and I looked at that and said to myself, that’s pure phenomenology! It was certainly around that time, I can’t remember if before or after, when Rich Lenski had also given me extra reason to persevere with this. Also, having been told I couldn’t do something, I stubbornly went out of my way to do it. Those efforts not only underpinned the Rainey &Travisano paper, but also “Evolution of cooperation and conflict in experimental bacterial populations” that appeared in Nature in 2003.
Jumping forward, Rich Lenski had told me about a Gordon conference on microbial population biology, and encouraged me to go. I must have gone in 1995. Let me backtrack. Before I went there, I had my own career to think about. My second postdoc was coming to an end, and I was really wondering if I had a career in science at all. Most of my work was on the plant microbe side, and I had plans to go back to New Zealand and to set up an agricultural business around growing plants to extract essential oils. However, I wrote a proposal to the (then) Agricultural Food Research Council (AFRC) for a five year fellowship to understand how a bacterium like SBW25 was able to survive in the plant rhizosphere. And that was funded. So, I suddenly found myself with five years of money for my own salary, and funds to set up a lab. With that I went to University of Oxford, just across the road from the NERC institute, and to the Department of Plant Sciences, where I then spent a very profitable and happy decade or more. Now, setting up my own lab and having been funded to do a particular project, there really wasn’t any opportunity to take further the work that would become Rainey &Travisano. I guess I began to write that work up, but it was by no means complete. I met Mike Travisano in 95 at the Gordon conference. Mike had done his PhD with Rich Lenski and was at that time working in Japan. With funds I had at Oxford I was able to offer Mike a postdoc position. And so Mike joined the team in, I guess, in 1997 and contributed some additional work that added some very nice polish to the Rainey &Travisano story. In particular, he brought the recognition of the value of performing invasion from rare experiments to explore the question as to whether diversity among different niche specialist types was stably maintained. That was a very important addition to the paper.
That is a long answer to the question of the origins of Rainey and Travisano! It was never motivated as a test of the hypothesis that ecological opportunity was important for adaptive radiation. It was purely curiosity-driven.
HS: That is a fascinating story, and I’m glad you described it in all its detail. One minor follow-up: why did you turn to sugar beet to isolateSBW25?
PR: P. fluorescens SBW25 is nothing special. As I’d gone from PhD to first postdoc to second postdoc, I continued my interest in phenotypic variation and I just switched to use whatever bacterium was the main focus of the lab. What the lab eventually did was to release a genetically marked version of SBW25 into a field setting. And the field that had been chosen was a sugar beet field because there were sugar beet plots at the university farm. These bacteria you find commonly associated, or always associated, with plant leaves and plant roots. So there’s no particular reason for choosing it other than, you know, the lab that I joined was already focused on this bacterium.
HS: Where was this lab located?
PR: The main experiments which I recounted, static versus shaken, the dynamic, a lot of the figures actually, that are in Rainey &Travisano, they were done in 1992-93, in Oxford at the NERC Institute of Virology and Environmental Microbiology. And then when Mike joined me, I was, literally, just across the road in the Department of Plant Sciences, where I set up my lab and was a faculty member there until 2005.
HS: Were there other people who were involved in doing the experiments? In the Acknowledgements, you thank a K. McCallum for technical assistance.
PR: Katrina McCallum was my partner. She would on occasions return with me to the NERC institute in the evenings to help preparation of the media and counting and so on. This would have been in 1992 and 1993.
HS: Apart from her, was it mainly you and Mike Travisano?
PR: Yes, these were projects that were done “on the side”, if you like. So there was no other technical help other than that provided on occasion by Katrina.
HS: Actually, can we go over the other names in the Acknowledgements to get a sense of who they were and how they helped?
HS: You thank J Baker for photography.
PR: John Baker was the photographer at the Department of Plant Sciences who took the images that appeared in Figure 1.
HS: Dieter Ebert
PR: Dieter is someone I still see from time to time .He was a postdoc in zoology. He and I talked about this work, and also work that he was involved in. We even published jointly on a project that was of his development. Dieter was quite excited about my findings and I think, if I recall correctly, he actually he did some early statistical analysis on some preliminary data sets that I didn’t eventually use. He was acknowledged more for his discussions as a friend and colleague.
HS: S Kahn
PR: Sophie Kahn was one of my first PhD students at Oxford. She is a New Zealander. She came to work on the genetics of the wrinkly spreader. She pioneered that work. She is acknowledged as someone in the lab at the time with whom I discussed many ideas.
HS: B Haubold
PR: Bernhard joined the lab a little before Sophie. He was working on the population genetics of fluorescent Pseudomonas on plant leaves. I would talk to him and Sophie often about what I was doing. They were acknowledged as members of the team.
HS: ER Moxon
PR: That’s Richard Moxon, a longtime colleague and very good friend, who I wrote to, probably, in the early 90s.There was a fellowship at Pembroke College, Oxford, the BTP fellowship, which I applied for with a research proposal, all based around phenotypic variation in bacteria. Richard had done pioneering work in this field. He’s now retired but still active. Richard was an infectious disease pediatrician and also ran a molecular microbiology lab, working mostly on Haemophilus. I saw a parallel with what I was looking at, even though in my case I had no idea of the molecular mechanisms. Learning that Richard was in Oxford, I sent him the proposal. It would have been 1992. He wrote back and said: “thank you very much. I find it very interesting. I’m on sabbatical at the moment. I’m back next year, so let’s meet.”. And so we met. And we subsequently went on to publish a number of articles together. And our discussions, I think Richard will tell you, were very influential in developing a more evolutionary perspective on the way that he saw phase variation in Haemophilus. And a delightful collaboration emerged. Richard is somebody that I’ve always discussed these sorts of ideas with.
HS: P. Sniegowski
PR: Ah, right, Paul was someone that I also met at the very first Gordon Research Conference on microbial population biology that I attended. We shared an interest in evolutionary mechanisms of what you might consider “evolvability”. We discussed ideas about the adaptive radiation.
HS: You thank him for comments on the manuscript.
PR: I must have sent the manuscript to Paul. He must have commented, although I do not now recall on precisely what aspects.
HS: I Moore
PR: Ian Moore was a great friend of mine and colleague in the Department of Plant Sciences. He was a plant biologist, but someone I would often drink beer with and talk about many things, including this work.
HS: Do you know what has happened to your labs at the NERC institute and the Department of Plant Sciences after you moved out?
PR: I can tell you about Plant Sciences. The NERC institute no longer exists. The building still does. I think it’s now occupied by University of Oxford, but I don’t know what goes on there. I occupied two labs in the Department the time that I was there. The first I occupied is where Sophie and Bernhard and Mike and I were working. Well, I was actually just back in Oxford, end of last year, and I visited my own lab, but not this very first lab that I was in, but a space that was renovated specifically for my expanding team. After I left the space was occupied by Ian Moore.
HS: Do you continue to work on SBW25 and the mutant morphs you describe in this paper?
PR: Yes. It’s gone in many different directions. But, absolutely. SBW25 has become something of a model organism, used quite widely in many areas of microbiology, evolution and ecology. A lot of the work of my lab and team over the years has been focus on the genetic and biochemical mechanisms of the different mutant types. Its genome was sequenced early on, it’s become a model of for plant-microbe interactions, particularly, through genetic tools developed, but also a model through the genetic and biochemical tools and evolutionary biology, and microbiology more generally. It is still at the heart of much that goes on. I guess a particular recognition that the wrinkly spreader morph is actually a mutant type that gains benefit by being part of a group has led to work that I find extremely exciting, on the origins of multicellularity, ideas about how the germ-soma division of labour might have emerged. So yes, I continue to be focused on SBW25.
HS: Were the names of the three morphs used for the first time in this paper?
PR: Yes. They were names that I just began to use,as I said earlier, as a way of making sense of diversity. As it turned out, wrinkly spreader has particular properties that distinguish it from a fuzzy spreader. And, in fact, there are many other more minor classes as well, that have never been investigated. We’ve got some stuff coming up, actually, on some of these other morphotypes. The diversity is phenomenal. And there’s no evidence from all of the mechanistic studies that we’ve done that this is anything other than just spontaneous mutation and selection. There’s no mutation mechanism that makes this happen at some higher rate or anything like that.
HS: How did you come up with the names for these morphs? Did you consider other names?
PR: It was purely for my own benefit alone, to give them names that meant something.I do, remember the first time, must have been in 97, when I was invited to talk about this work in one of the Gordon research conferences. My use of all of these names resulted in great hilarity in the audience. Maybe it was, in a way, smart. I never thought about it in terms of a marketing thing. But I guess it made people take notice. I did it purely for practical, pragmatic reasons I needed to be able to count and for that I needed to be able to name. I could have called them As and Bs and C, but the names describe how they look on an agar plate.
HS: From what you have said so far, I get the sense that this paper was probably written in bits and pieces…
PR: Yeah, it took me a long, long time to write this. As it is with other pieces of work that are describing something new, when you are talking about it at meetings and conferences, you realise that it is not easily categorised. So, how it’s then presented becomes critical. Another recent paper, a couple of years ago now, on our work on the evolution of multicellularity, again featured these wrinkly spreaders very prominently, took me two years to write. I got very close to submitting it on two occasions and then just tore it up and started again, purely because,in that case even more so, it really is utterly new and you’re trying to get people to see the world in a very different way than they would ordinarily see it. When you do that, the temptation is to try and shove the message down the readers’ throat. But that is not a good strategy. It tends to annoy people. And you often lose your ability to convey what you mean. I think it is better to present things in a kind of mundane way to begin with. Let the narrative and the data tell its own story, so that by the time you get to the discussion, you can bring in the stuff you really want to say. By that time, hopefully, the reader who has persevered, has come to a position where they realize, ah ha, this is different. They are taking me somewhere I’ve not been before. Whereas if you put that to begin with, they just go, “rubbish”! So, I think, it was a similar thing with Rainey &Travisano. I remember visiting my father in New Zealand, well before I met Mike even. He used to have a house with water access only, so I was there for 10 days, working quietly undisturbed, no internet or anything like that, and just really making so little progress. Maybe I wrote a paragraph or something. It took a while. It took a while because it’s not a formulaic piece of work. It was new and the challenge was, how to get across the message. How do you communicate something that’s new? If I’m following something that’s been done before, it’s easy; I just take an established format. But this was not formulaic and it took a lot of careful thinking. I talked about it with a lot of people. As you talk about it, you realize what they don’t understand or are confused about and therefore what you might need to explain more. It definitely took a while.
And it didn’t have a smooth ride through the publication path. In fact, in my office in New Zealand, I still have all of the paperwork surrounding it in the folder. I wanted to make sure I don’t lose that. It was a bit of a roller coaster ride, with ups and definitely downs.I guess about the time that it was nearing completion, I talked about the work at a meeting of the European Society for Evolutionary Biology. It was somewhere in the Netherlands and it must have been, in 1997. And there was a journalist in the audience from Science Magazine, Virginia Morell. The talk itself had created a lot of buzz, and she came up to me and did an interview and published an article in Science on it. I thought this is this is great. Science should thus be the journal to whom we submit the paper. So, we submitted it to Science. I remember submitting it before going to the Gordon Research Conference. This must have been in 1997. Maybe, I talked about it, first, at the Gordon research conference in 1995. Mike and I both went to the GRC in 1997. The paper went to Science before we traveled to the US. We hadn’t had a rejection. Maybe this was three weeks by the time we went to the US, so we thought it must have gone out to review. We came back and still no news and so we’re sure it must have gone to review. But no, it was rejected without sending it to review. It just took them two months. That was disappointing, of course. Given that it hadn’t been reviewed I probably didn’t change it much before submitting it again. The intention was to submit it to Nature, but I realized that maybe we were in breach of protocol here. It could have been perceived that we had solicited attention of our work by Science and that would be cause for Nature to decline to consider the paper. So I remember enquiring with Nature,and they said, no, it’s fine as long as you didn’t solicit attention. Virginia Morell wrote to Nature and confirmed we hadn’t solicited her article. Nature said that’s fine, we would never want to block scientific exchange. We submitted to Nature, it went out for review, and came back with one very strong positive review and a very negative review. The paper was rejected. The letter from Nature said something like, we feel that the negative comments from a significant expert in the field, simply preclude publication. What I remember about this negative review was that it was super angry. It was, as far as I was concerned, a rather misguided review. The reviewer made a number of claims in the critique that were simply false. So I decided to respond to this, to appeal against the decision, I remember spending a long time, no doubt with Mike, going through and rebutting the negative comments. And then, I made a case for it to be sent to a third referee. Nature agreed. I remember coming into work one morning and there was a letter in an envelope from Nature with the news that the paper had been accepted. The third reviewer had received the previous reviews and our rebuttal. He had effectively critiqued the negative referee. I remember one case where the negative referee, in relation to the frequency dependent interactions, had said, the authors are patently wrong. I forget exactly why, but that was one of the things I completely objected to. The third reviewer wrote, it is the second reviewer that is patently wrong, and not the authors. There was not much revision required. I think it went to somebody who really couldn’t grasp what we were saying. From a purely microbiological perspective, it seemed almost magical that we get this radiation driven by what we didn’t know at the time. We had all sorts of strong circumstantial evidence, but we didn’t have the molecular detail. We know now that it’s all driven by spontaneous mutation, hundreds of which had been characterized. But at the time it seemed bizarre to many that you could have this phenomenology going on. It must reflect some kind of unknown evolutionary mechanism or something like this. I suspect the reviewer was probably an evolutionary biologist who really struggled with making sense of what we were offering, particularly as we didn’t have the mechanisms worked out.
HS: After Mike Travisano joined the work, did he also get involved in the writing?
PR: Yes, I’m sure he did. I guess it was so much my baby, if you like. I’d already done a lot of it. But he must have absolutely contributed to the writing.
HS: Do you remember how it was received when it was published? Did it attract a lot of attention?
PR: Back in those days, there was no Twitter or other things like that, so I don’t think one was particularly aware of how a paper was received, until after people began to cite it. I remember being on holiday in Wales actually, when it came out, and finding a copy of Nature at a news agent. Actually, there was some media attention, not on the BBC or anything, but National Geographic did a piece. Thinking back, actually, there were quite a number of articles that appeared in magazines in various places. So, I guess it did, provoke some interest.
HS: Today, the paper has been cited over 800 times. Are you surprised by how much attention as it has received over the years? Do you have a sense of what it mostly gets cited for?
PR: I haven’t looked at the trajectory. I don’t know if it took a while to get cited or whether it began straight away. It was great to see it get that attention. With the benefit of hindsight, perhaps it’s not something completely surprising. That paper captures something quite complex in very simple, overly simplistic, terms, I might say now. It’s a great metaphor, if you like, for a complex problem. It enabled something that was often more theoretical, and even with experimental studies, rather difficult to grasp. It gave some concrete reality to it. I think often, papers that, in simplifying, capture the essence of something, often do end up with a lot of attention, whereas those that provide a huge amount of detail, and are arguably far more scholastic, are often not cited so well. In terms of who cites it: I’m aware that it’s been quite widely cited. For example, in cancer biology, where people, I thought, very nicely, drew attention between what happens in the tumor when it radiates and what we capture on these simple microcosms. Often it’s cited, I think, as an example of imaginative experimental evolution. And perhaps it also ends up getting cited because it’s been cited.
HS: In the paper you say, “The ecological mechanisms maintaining diversity within the spatially heterogeneous environment are complex and not fully understood.” Did this form a focus of research for you subsequently? Do we know more today?
PR: Yes. That’s a long story, and I won’t go into it. But, yes, we have done more, but nowhere near as what could be done. I guess what I said to some of my collaborators in the early days was that, after Rainey &Travisano was published, there were a number of papers that simply took that system and applied it to fairly simple ecological problems. We now had an assay that was powerful. And I used to encourage some of the early members of the team to try and understand the system better, or get further into the ecological mechanisms and so on. And I would do so pointing out that, it took many years to get to this assay, and you’re now in a fortunate time where you can exploit it, but it would also be very valuable to contribute to a greater understanding of what’s going on these little microcosms. We haven’t done as much as I would have liked. However, in recent times, we have, and there’s more stuff coming out, which gets at some of the extraordinary complexities underlying this radiation, including quite a substantial revision to the model that we applied to understand the radiation. It is true that competition for oxygen is the primary deal, which, of course, as I guess I was aware of way back then, was at odds with what the morphotype called Fuzzy Spreader appeared to be doing. As you can see in the paper, it appears to grow on the bottom. But that doesn’t make a lot of sense because the bottom should be anaerobic. In fact, we know it’s anaerobic. Once you get microbial metabolism taking place in the broth, oxygen is rapidly depleted, and these bacteria need oxygen. So I always thought it was odd and actually looked for various explanations for Fuzzy Spreader. The other thing that we show in Rainey & Travisano is that fuzzy spreader can invade against the ancestral smooth type, which suggests some sort of allelopathic factor, which also seemed very odd to me. I remember hunting for a possible phage that might have been liberated under anaerobic conditions that might have killed the ancestral type, and things like this. Anyway, the short of it is that our understanding of fuzzy spreader is quite wrong. That was revealed when Gayle Ferguson, a postdoc in the lab, took a time lapse movie video of fuzzy spreader when put into the microcosm on its own. You can view the movies if you consult her paper in Genetics published in 2013. This revealed that fuzzy spreader, like the wrinkly spreader, colonizes the air-liquid interface, but it’s not very good at it. It forms little rafts that very quickly get too heavy and they then fall to the ground. So, what we saw photographed in that paper is the detritus, if you like, of life at the interface, and this turned out to be interesting from various perspectives. So we have fuzzy spreader, like wrinkly spreader, like smooth, trying to get oxygen. That’s the limiting deal. Wrinkly spreader forms this very substantive mat due to overproduction of a cellulosic polymer; the cells are glued to one another and form this mat. That’s not the deal with fuzzy spreader. Actually what really prompted us to make progress toward understanding the ecology of the fuzzy spreader a bit more was when began to get at the genetics underpinning fuzzy spreader. To my amazement, the locus where a mutation arises that causes fuzzy spreader isn’t a locus that we know is required for expression of the wrinkly spreader phenotype. That seemed to be really odd. Anyway, that mutation that generates fuzzy spreader alters a surface component. The altered component provides the possibility, or facilitates interactions, of an electrostatic nature between the cells. They’re not as strong or robust as those between the wrinkly spreader cells and can’t form a thick mat like the wrinkly spreader does. Nonetheless, it forms these thin rafts of cells. It has a niche in there, it cannot make it in the face of the wrinkly spreader mats, but when the wrinkly spreader mat has collapsed, the fuzzy spreader has opportunity for a while. Understanding the ecological basis, the genetic and biochemical basis, of fuzzy spreader also led to an explanation of why fuzzy spread could invade against the ancestral type. What it was able to do is, unlike the ancestor, really can colonise the meniscus, gain access to oxygen, and out-compete the ancestor through competition for oxygen. And also, the nature of electrostatic interactions means, unlike in the case of wrinkly spreader, which can be taken advantage of by the smooth genotype, the smooth genotype cannot hitchhike in among the fuzzy spreader mat.
So well, some detail there. It’s something I would like to study much more. understanding more fully what’s really going in these microcosms. The more we learn, the more we realize how much more remarkable and complex they are than we once thought. The original model, has been viewed as an example of an adaptive radiation driven by different ecological niches: one growing on the top, one in the middle and one at the bottom, for example. But really that’s not the case. Everything is being driven by competition for oxygen, and the different strategies coexist in a time lag frequency dependent manner, really as a consequence of the fallibility of each strategy due to the effects of gravity. But there’s much, much more to learn.
HS: Towards the end of the paper, you talk about how the approach in this study might provide a framework to look at phenotypic variation in other bacteria. To what extent has that happened?
PR: I’m just thinking what I meant by that. I think what I meant there was not so much that you take this experimental system and you just put another bacteria in it; although, people have certainly done that. I think what I was getting at was, everything that we knew then, and subsequently learned in detail about the genetics and so on, says that what we are seeing here is an instance of evolutionary change that is remarkably rapid and remarkably repeatable, but it’s attributable to nothing other than mutation and selection. What I was saying is, look, you see all of this stuff and it almost looks like it’s “programmed”. Provided you have a large enough population size, you will always get these various dynamics. And therefore it’s tempting, particularly in the microbiological world, to want to attribute that to some sort of evolved regulatory system. There was a prior history of being able to do this sort of stuff, and there is a gene regulatory network that’s specifically there to control this kind of radiation. That’s not the case at all. I think that’s what I was meaning basically, that microbes evolve and adapt by natural selection and that can be very rapid and with surprising effects. When we see it, we shouldn’t always just assume that it is underpinned by some regulatory, programmed mechanism. So, I think in that regard, it has been something that myself and others have sort of banged on about, the need to understand that every time you grow a bacterium in a vial, it’s going to evolve and you can’t stop evolution happening. That was the thrust of that comment, I believe, to heighten the awareness of evolution.
HS: Would you consider this piece of work and this paper as major landmark in your career, in terms of the impact it had on future research? And would you count this as one of your favourite papers?
PR: Yes, for sure. It’s a paper I’m very proud of, really. I’m sure Mike would say the same. This is not just because it is so well-cited, but also, to me, it’s sort of a hallmark of a number of other papers that have been published over the years, which I guess revealed a degree of creativity in the way I approach science that I like. It wasn’t easy to get it published, it wasn’t easy to write it, it was new and different, and it all came out of curiosity. It required some imagination and tenacity to see it through. So yeah, it was the first example of a piece of work like this that I’d seen through to publication, and it’s definitely high impact. Yes, certainly, for me, it was a landmark.
In terms of career, yes, sure. It was my first paper in Nature. Although one wishes this stuff didn’t matter to careers, it certainly is not a bad thing to have one or two under your belt, especially as a young researcher. No doubt about it. Although, in that regard too, I think it was tremendously important to have followed up that paper, within a few years, with some serious genetics and ecology to counter the often prevalent view that what gets into Nature and Science is cute and without substance. Following that through with some painstaking genetics, to show really what’s going on, was also tremendously important, and added a great deal to the impact of Rainey &Travisano.
HS: The adaptive radiation community is quite large and diverse today, in terms of the taxa being studied. Did this paper help you become part of that community of researchers?
PR: Yeah, it’s certainly true that before this paper I was not interacting with that community. My particular interest moved at that time, I guess, from the adaptive radiation itself to other matters. But certainly, in recent years, I have had the pleasure of meeting some of the leading figures in the study of adaptive radiation. And to my delight, they know about this piece of work and the ensuing studies that have come from it. I was particularly delighted to meet, two or three years ago, Peter and Rosemary Grant. Okay. I’d never met them, and they said, they always wanted to meet the person who’s responsible for this. Jonathan Losos is another person I met, who I’d not met before.
HS: Have you ever read this paper after it was published?
PR: Interesting question. I guess I haven’t, from beginning to end. I’m awfully familiar with what’s in it, although, you know, you have drawn attention to details that I forgot with your questions. I certainly think about it. I certainly think about some of those early studies and unsolved issues that still interest me today. When you write a paper, especially a short paper for Nature or Science, you really can only have one message. If you’ve got multiple messages, it’s easy to get bogged down or lose the reader. There’s one thing in Rainey &Travisano that I always wish I had found a way of making more explicit. It’s there in the paper, and I think it’s unique: we observed the de novo evolution of the game of rock-paper-scissors. And I think I deliberately decided not to do that because it would just add too much and it may detract. But the fact that you end up with these non-transitive interactions evolving de novo, and where the three main morphotypes play rock-paper-scissors is quite something. My friend Ben Kerr published a lovely paper on rock-paper- scissors, but it was an ecological paper where it showed that you could get a rock-paper-scissors if you had these particular genotypes with these particular fitnesses. In the work of Rainey and Travisano these interactions evolved right there in the test tube!
HS: What would you say to a student who is about to read this paper today? Would you guide his or her reading in some way? Would you point them to other papers they should read along with this? Would you add any caveats they should keep in mind when reading this paper?
PR: Well, I think the paper can be approached on its own largely because it’s so visceral. I guess what was really lacking from the paper back then was any information on the underlying mutational process. And so, I would want to draw their attention to some of the papers that have described the underlying genetics (There are, for example, a series of papers in Genetics with me as last author) so that there’s no doubt that this really is all driven by spontaneous mutation selection. In terms of caveats, it’s not a caveat really, but rather how science works. I rather like the fact that we were able to show that we had it wrong with regard to the fuzzy spreader. As you investigate things, you may find what you once thought to be true, no longer the case. Some piece of information has caused you to revise a particular position. And I like very much, and get students to read, the paper by Gayle Ferguson and Frederic Bertels in Genetics in 2013, which is titled something like “A revision of the Pseudomonas radiation”. This is a good example of how science works. Again, this is not really a caveat, but something that I would emphasize. My advice to someone reading for the first time, might be to sit back and see how extraordinarily simple this piece of work is, and yet how challenging it might have been to unravel. Sometimes, just looking at the world differently can take us to unexpected places.