Concerns traditionally central to poetics (pity and fear; to delight and to teach; truth, beauty; etc.) also matter in other domains of inquiry. This is the first installment of a series of interviews that pursue such “poetic” concerns with practitioners of other domains of inquiry, such as science and philosophy. When they were paired in a recent collaborative project involving scientists and artists, hosted by the Ucross Foundation, H. L. Hix took the opportunity to interview microbiologist Naomi Ward about her recent work, with particular focus on her recent paper disclosing a discovery about the bacterium Gemmata obscuriglobus, just published in the Proceedings of the National Academy of Sciences U.S.A.
H. L. Hix: By way of introduction, how would you clarify or amplify (or simply correct!) this amateur understanding of your article? The most widely-held view about the cellular processes of transcription and translation is that, in bacteria, they are always coupled: that is, they always take place close to one another within the bacterium. But you and your colleagues have shown that in the bacterium Gemmata obscuriglobus, the processes do not always occur in proximity.
Naomi L. Ward: You’ve really captured it. There are a few things I could say for clarification and background, that add some complexity to the story (like you needed more complexity, right?). I should mention that the coupling is not just about location. In bacteria, transcription and translation can be almost simultaneous, so that as the transcript is being copied from the DNA, the ribosome (the cell’s protein factory) is already jumping onto that transcript and starting to make proteins from it. It’s not just that transcription and translation occur together in space; they’re together in time as well. In the more complex cells of eukaryotes (plants, animals, humans), that transcript is made in the nucleus, but then exported out of the nucleus. There’s a journey there, a gap in space and time.
The background, and what makes the story more complex, is that when we started the project, we thought we had a rule and an exception. So the rule, as you’ve mentioned, is that bacterial transcription and translation are always coupled. We proposed that we had, in the cells of Gemmata, an exception to the rule. Because of the way the cell was structured with lots of internal membranes, we thought there was a membrane barrier between transcription and translation. So we started down that track. Technical complications caused the work to take a very long time (four years from beginning the project to submitting the paper). During that time, some papers were published that provided other exceptions to the rule, in model organisms from which we have learned nearly everything we know about the way bacteria work (E. coli and Bacillus subtilis). So there we were, getting ready to submit our paper, having spent four years on what we thought was our unique exception, to then see these other papers coming out, saying that even in simple cells translation was sometimes occurring spatially segregated from transcription. What we maintained, though, is that in Gemmata we still had a layer of complexity that wasn’t there in the simpler cells of these model organisms. We made the argument that it was a unique cellular context. Although it was no longer a unique phenomenon among bacteria, there were challenges and aspects to coupling of transcription and translation in Gemmata that didn’t exist in these other organisms.
HLH: Continuing this thought about introducing this conversation to an audience that hasn’t yet read your paper, let me jump forward to something you touched on in that response, which has to do with the relationship between the bacterium that you’re studying and processes in eukaryotic cells, the cells of plants and animals. Most of us aren’t microbiologists, and when we think about the processes of reproduction, we think of things that are at our scale. We think of the birds and the bees, animals and plants. But bacteria are different in structure from the cells of animals and plants, so could you speak a little more to the distinction between the processes in bacteria and those in eukaryotic cells?
NLW: We’ve tried to be very careful not to extrapolate too far from our findings. Our findings are simply that the two processes, at least some of the time, occur in different places in the Gemmata cell. From there, you can take all kinds of roads that lead you to comparison with the eukaryotic cell, and in fact those roads have been explored before our paper, based on other evidence, other data from Gemmata. One road has been to argue that the eukaryote-like features of Gemmata are homologous with those of eukaryotic cells. In other words, they shared a common ancestor at some point. On that view, the resemblance we see between the Gemmata cell and a eukaryotic cell is not superficial, it actually has an evolutionary origin that includes a common ancestry. That has been countered by the argument that it is just a superficial resemblance, that Gemmata has independently come up with the same solution to a given problem as have eukaryotic cells. In that case, we would call the features that Gemmata shares with eukaryotic cells not homologous but analogous. The other phrase that is used is convergent evolution. I think that either of those explanations is very interesting; I don’t particularly favor one over the other. Our data could actually be used to support either argument.
But one thing that is novel about our work is that, although the decoupling of transcription and translation in Gemmata had been previously proposed — someone had already thought of this before we did — it hadn’t been experimentally tested. That’s where we did something that was different.
HLH: I’m latching onto elements of your response. Your answer is very carefully formulated. Part of the reason for conducting this interview is our shared interest in questions of process, how we ask questions and check answers and so on, so I’m interested in the particular form that your caution takes. In another context, from a person engaged in something other than science, I would expect less reticence about making a sweeping claim. If this were a conversation about political matters and you were a senator, say, I’d expect a readiness to jump to those large sweeping claims. If we were talking about a poem instead of a research paper, I would expect you as author of the poem to be entitled to say what you want about the poem, to infer whatever you want. So what is it about science that creates the kind of caution you have just displayed, and why is that caution valuable? How does it help us learn stuff?
NLW: My response to that would be nested: a general response about scientists in general — I’m now going to make sweeping statements! — and then within that a response that’s particular to the research I do and the research community that I work within. As scientists, we are trained from an early age to not speculate too much from our findings. Of course, it’s natural, when you find an answer to a question, to ask the next question, and sometimes the next question takes the form of a speculation. From any set of results we have, you can think of the questions as runaway horses, and part of the fun of the process is that you see all these horses go off in different directions. But as I said we’re trained to corral them, because they are not yet supported by data.
At one point in our paper (and other people do this too, of course) we say “Here’s an idea, but we have no experimental support for this speculation.” It’s part of our training, especially in a paper that is a primary research article, as this one is, to clearly distinguish between what we do and what we don’t have experimental support for. In a different kind of scientific article, such a review paper or (especially) an opinion piece, those are where you can let the horses go. These kinds of papers may be peer-reviewed, but the tolerances are going to be increased, depending on the format of the article.
That’s the general sweeping statement. Nested within that, the community of people who work with Gemmata and its relatives is, right now, very small, a fraction of the size of the community that would work with a medically important, disease-causing organism such as the E. coli that causes food poisoning or a drug-resistant, flesh-eating Staphylococcus. Gemmata does no harm to man, beast, or plant, or any other thing, as far as we know, and none of its relatives do. Which is good, on one side, but it means that our funding sources to do this are different, smaller than what would be available to work on an organism with biomedical relevance. And that’s probably as it should be.
Gemmata and its relatives are not terribly difficult to work with (to grow, to study, to apply standard methods to), but they are more difficult to work with than the aforementioned E. coli and Staphylococcus. That’s another reason why the community is relatively small. We just had our very first conference on this group of bacteria last year in Heidelberg, Germany, which was wonderful because this small community came together. This is leading to the observation that, when you have a small research community, everyone knows everyone, either through personal contact or through the literature. In order to keep a good scientific discussion going, which is really the whole point, it helps to be not dogmatic, and to take your speculations only as far as your data will support.
The process of writing and revising this paper was interesting because the question I mentioned earlier, about whether Gemmata is a homologue or analogue of the eukaryotic cell, meant that the revision of this paper required an extra level of thought. Not just for the purely pragmatic perspective of getting it published, but to really feel like we were making the right statement for the data that we have. Maybe ten years from now we’ll have more data, different kinds of data that will let us make a stronger statement in support of one or the other of those opposing views, but it seems like the right degree of caution for where we are right now.
HLH: I’m becoming self-conscious now about being incoherent in my questions! I’m supposed to be creating a through-line to the conversation…
NLW: … it’s alright, you have grasshopper mind!
HLH: Yes! You mentioned that Gemmata doesn’t harm humans, that it doesn’t have “biomedical relevance.” OK, but then why might a person be interested in it? If it doesn’t pose an immediate danger to humans, it seems possible for a human being to ignore it. I could go through my whole life, if I’d never met you and read your article, and never even hear the name Gemmata obscuriglobus. Why might a person be interested in this? Why did you choose this bacterium to study?
NLW: I’ll start with the second part, why I chose it for study. That story is a combination of life history and serendipity. I did my undergraduate study in microbiology at the University of Queensland, in Australia, and Gemmata obscuriglobus was isolated and studied by a researcher in the same department. His name is John Fuerst, and he’s really the father of Gemmata biology. He was also one of my lecturers. I was generally aware, as an undergraduate, of this organism and that it was unusual, but then serendipity kicks in. I wanted to get into undergraduate research, and I was very interested in immunology, so I went toward the end of the semester and knocked on the door of the immunology professor. No answer. Somebody stuck their head out of the lab and said, “Oh, Bill’s on sabbatical, so we can’t take anybody into the lab right now,” and then they said, “but the new head of the department, who has just moved here from Germany, is setting up his lab, maybe you should go ask and see whether they want any help.” The new head of the department was Erko Stackebrandt, who was later my doctoral advisor. He had previously worked with Gemmata and its relatives, and he was a leader in using the molecular methods that I’ve talked to you about for identifying and classifying bacteria and studying their ecology. Those things came together: my exposure to Gemmata through John’s work, and then reinforced through my experience as an undergraduate. So this organism has been with me from the age of 19: we go back!
I did my graduate work on this organism and its relatives, took a break from it for a few years, and then came back in the era of whole genome sequencing. People know now that you can sequence the human genome, but of course you can also sequence microbial genomes, and in fact, it was done with microbial genomes first, because they’re smaller and less expensive. I worked in an institute where we did that kind of work (The Institute for Genomic Research, TIGR for short), and Gemmata was one of the organisms that I got funding to work with. My Gemmata research continued after I went to Wyoming, and in fact, it was a cornerstone of my new research program. In hindsight, it was maybe not an entirely wise choice for a new tenure-track faculty member, because as I mentioned these organisms are not quick and straightforward to work with, and the community is small, so there aren’t a lot of resources and shared expertise yet. I made things a little difficult for myself!
That’s how I came to the organism. In terms of broader interest, you’re right. Gemmata doesn’t make us sick, or do anything useful for us. It doesn’t produce ethanol from corn, or help us make bread, or carry out some useful process in the environment — as far as we know. But we haven’t looked enough yet. It has a very, very large genome for a bacterium, so there’s a repertoire in there for lots of different kinds of activities. We just don’t know what they are. So maybe it does do something useful. Maybe it does make somebody sick — but we’re not allowed to do those kinds of experiments!
So the interest and relevance is much more of a basic-science question: it’s a question of the relationship between structure and function, with a connection to the framework of evolutionary biology. How did animals get to be as complex as they did, in comparison to bacteria, and why? And while Gemmata may not (in fact probably doesn’t) have direct answers to those questions, even if it is only a superficial resemblance, it may tell us something about a more complex state and how that might have come to be. Tangentially we might get some ideas: “Well, if that’s how it happened in Gemmata, can we test whether that’s also how it happened in eukaryotes as well?” This has been proposed, mainly by John Fuerst, as being the value of working in Gemmata – the ability to test hypotheses.
I’m not sure if that’s what you’re getting at…
HLH: Absolutely. We share a preoccupation with ways of knowing, and I’m interested in a difference between poetry and science in this way. It seems from the way you use “basic science” that there is an important sense in which a paper in science is a part of a large, dynamic, very multiple inquiry that’s going on. You’re able to assume that a reader of your paper is seeing it in light of lots of other stuff that’s happening, and that seems slightly different from poetry, or at least a difference of emphasis. We typically think of a poem, or the experience of reading a poem, as a more or less self-enclosed thing. Although there are actually a lot of connections being made, to other poems that one has read, to historical questions, contemporary social-political questions, and so on, a lot of that stuff recedes. It’s more or less invisible in the experience of reading. It might seem as if, or feel as if, I’m just looking at this one self-enclosed thing. Everything is more or less self-contained, and this thing can be isolated from the rest of the world. I’m isolated from everything else in the world when I read this thing. But that appears to contrast with a scientific paper, where it’s part of this huge community operating — parallel doesn’t quite catch the sense, but happening all at once — and all the different questions relate to one another, and they temper one another and condition one another and inform one another.
I don’t even know where I’m going with this!
NLW: But you’re right, a scientific paper cannot stand by itself. In a standard research paper format, there’s always an introduction, which provides context and background for the question that’s being asked. There’s a process whereby you say what is known about a particular topic, what we don’t know, and then why the not knowing is a problem that prevents our moving forward. With that you establish that our goal is to find this out. It’s impossible to write a standard research paper without that kind of context. Often then (but not always) the paper will come to a hypothesis: a proposed answer to the question, one that will be rigorous and testable and falsifiable. But there are also papers that are not hypothesis-driven, they are exploratory; a report of a genome sequencing project would be a good example. There are two places in the paper where you can make connections to other people’s work: in the introduction, where you’re setting the stage, and in the discussion and conclusion section, where you take what you’ve found and say this is how our findings support something someone else did or contradict it. That process of extrapolation and speculation — the wild horses — is part of that as well, and there sometimes you can give your wild horses some more credibility if you say, “As so-and-so has suggested (reference x), this may be a mechanism for y.” Even there in the process of extending what we know or think we know, we draw on the community.
I mentioned the other day the idea of lineages. Particularly in European science, it’s common to refer to your scientific father or grandfather (or, increasingly often, your scientific mother or grandmother): it’s a pedigree. People actually use the word “pedigree” when they’re reviewing candidates for a faculty job, for example! “This person has a good pedigree,” they’ll say, like they’re talking about a race horse. So not only are there lateral connections to other scientists that occur through the research questions; there are vertical connections, too, that can affect the way you structure the questions you’re going to ask, and when you’re going to ask them. Pedigree, in a professional sense, also affects your credibility. So in a number of ways your work doesn’t stand by itself. I assume that in poetry there are mentors, but maybe it’s not the same kind of relationship.
Another example in science of the strength of that pedigree is that, as a professor, when you train graduate students and post-doctoral fellows, and they’re ready to go on to the next position, you write them a letter of recommendation — hopefully a strong one! — and that obligation is lifelong. So you write the first letter for the graduate student applying for the postdoctoral fellowship, and you will be called on again to write the next letter, for a faculty position or a job in industry. These letters evolve over time, because you start to incorporate what the person has done since they left your lab. There’s almost a family connection there that is very strong. And the more I think about it, value judgments are made about people, sometimes not just on the content of the letters of recommendation and who they came from, but the speed with which they are received. If a super-famous, respected scientist — very busy, very important — is asked to provide a letter of recommendation for their former student, and it’s there the next day, then — Wow! — not only was this candidate trained by this top person, but this person, this mentor, cares enough about the candidate, values them highly enough, to get right on that letter. That’s where it’s like family, because you do more for a family member than you would for those outside of the circle. So there are lots of connections in science…
HLH: … which also seems related to another difference between science and many other enterprises. Of course, because of the context of our discussion, I’m thinking particularly of poetry. Science seems to be a team or group activity, whereas poetry seems most typically to be a very solitary activity. That’s a bit misleading in relation to poetry, because a lot happens communally: sending one’s poems to friends or colleagues for critique, workshop settings, and so on. But still, in poetry — however you prepare, however you engage in dialogue about it — the core activity of sitting down and writing a poem usually happens when you’re by yourself, and usually as solitary as you can get. If you can seclude yourself in your garret, away from all other people and all other noise, the more solitary the better. But in science we talk about a lab, meaning a place in which many people are working on the same thing, working together as an organized team.
Again I’m not sure of my question! Why might this matter? How does this help in relation to the acquisition of knowledge?
NLW: Science used to be like your description of poetry, but over time that has gone away in most of the experimental sciences. In some of the theoretical sciences, which may be more akin to poetry, people do have papers with only their own name on the author line. But in the experimental sciences, the team nature of it reflects a number of things. One is that, as we have specialized, many of us lack all the tools we need to produce the data that we need to make a convincing story. So if you look back to scientists of the nineteenth century, it was intimidatingly diverse, the range of things they could do: natural history, chemistry, geology. They were Renaissance people, and could do a little of everything. There was less to know, then, in terms of training and background. But now we’re all highly specialized, so there is more interdependence, not just of expertise but of resources, big expensive pieces of equipment, access to field sites for ecologists. If the question you want to ask is in Ecuador, it helps to have a collaborator in Ecuador who can help you with that access and help you work there.
Another part is what I was just describing, our training activities. But what I’ve observed is that the ties between student and advisor seem to be tighter in microbiology and molecular biology than in some other types of research. In my field, when a student writes a paper, in general, we write together. It’s a symbiosis. There are specializations within that symbiosis. The student is usually the one who has done the hands-on work, and is often more familiar with the techniques and their intricacies; the advisor has the bigger picture, maybe more of the history, maybe more flexibility in interpreting results. So part of the collaboration comes from the training that we do, and that may be why there is this family structure in science: you “grow up” with these people as a scientist, and in fact, scientists who trained in the same labs — siblings — stay in touch. When they get older, they invite each other for seminars, because there’s a strong connection that’s forged when you’re a student working alongside other students. Your experiments are failing, it’s the middle of the night — there are strong bonds there! I think that’s part of it, too.
There are also ways in which the roles of a collaborative team in science are encoded in the author line of a paper, and the quite enormous importance, career-wise, of how your code is being read, what it’s telling the reader about your contributions — hopefully accurately! We actually have a codified structure in that author line that represents the nature of the collaboration.
HLH: You mentioned that collaboration might take various forms, one of them access to equipment. This relates to another question we’ve hit on a little bit in prior conversation, about the role of imaging in verifying your hypothesis. You mentioned one of your colleagues and the equipment and process for imaging. Could you speak a little about that? How does this imaging work? Why is it important to verifying your hypothesis? What does it allow to be seen that needed to be seen in order to check out what you thought was happening? What does it show?
NLW: There are a couple of different ways of answering that. One is that our question is essentially a spatial question. We’re asking about spatial relationships. Imaging is not the only way to capture that from a cell. You can, for example, take a cell, break it apart, separate the different parts of the cell, and then take each of those fractions and interrogate them with your question. But it’s more common to answer a spatial question with an image. The other reason is more pragmatic: we relied heavily on imaging in our analysis because we were constrained technically. We cannot yet successfully fractionate the Gemmata cell, so we couldn’t use the approach I just mentioned. Or couldn’t in a convincing way. We were also constrained to certain types of imaging and processing of samples for imaging by the fact that we also cannot use what is now a central tool for asking questions about the bacterial cell, and that is genetic manipulation. A much quicker and more standard way of asking where a protein exists in a cell is to take the gene for that protein and tag it with a label of some kind, then put it back into the cell and see where it is. When the gene gets made into a protein, the protein will have the tag. We can’t do that with Gemmata. Or at least in our hands. There may be others who can, but we can’t.
So we had a relatively narrow range of experimental techniques that we could use. One of the reasons the project took so long to come to fruition is because of reliance on those techniques, because they take longer. Even preparing the tools that we needed to ask the question took a long time.
The other aspect of using imaging data is that seeing an image and the distribution of something within that image is the most intuitive way for us as humans to answer a “where” question. We can very easily look and say, “Oh, yeah, there’s more of that over there than there is over here.” But usually that’s not enough to convince another scientist that such a distribution represents the true distribution in all individuals, all cells that you might possibly look at. It’s a sampling problem. We may have five cells in an image, and those five cells may all support our conclusion, but it’s very easy for someone to look at that and say, “Well, you just found the five cells that supported your story.”
HLH: Right! These are the only five cells in the world that do!
NLW: And they might be, so that’s why it’s important to add a more quantitative approach, to look at a larger sample of cells, and ask how often this distribution occurs, as opposed to some other distribution. That’s a fairly standardized thing across many sciences that use images: the imposition of a more quantitative analysis of that imaging data, on top of the image.
HLH: What you saw by means of this equipment was … what? For a person who hasn’t read your article, what was it that you saw?
NLW: Do you mean literally, looking down a microscope?
HLH: Yes. It’s a spatial question that you’re asking, so you used microscopy in a way that generated images. What were those images images of? What spatial relationship was revealed by them?
NLW: We saw markers — flags, if you will — for transcription and translation often occurring separated from each other in the cell. Not always. Sometimes we saw those flags overlapping (evidence of co-localization), but when we did the quantification that I was just mentioning, we found evidence that there was more separation than there was togetherness. When we looked at those images, we were also able to orient those flags with respect to other features of the cell, particularly the internal membranes that make the Gemmata cell so distinctive. So we were able to say not only do these processes seem to be separate spatially, but here is their relationship to the surrounding landscape. When for example we looked at that merged image with all the colors in there, that allowed us to say, “Oh, here is the red signal and here is the blue signal, and in between there’s a green line and it’s a membrane.” It’s a form of mapping. We’re mapping the processes that we tagged, with respect to other features. Which we can then make wild speculations about!
HLH: That mapping is what I was getting at. You had just used the word “flags,” that you were seeing flags. So I was thinking, by way of a metaphor or analogy, of something like tracking. What the hunter or tracker sees is evidence of thing that he or she is looking for. I’m trying to find the mountain lion that I’m told is living around Ucross, but that is very secretive. I haven’t actually seen the mountain lion yet, but I’ve observed fresh mountain lion droppings, and I observe a freshly-killed deer that was not killed by bullets but by fangs, and I observe mountain lion tracks in the mud, and so on. So I know the mountain lion is there even though I haven’t seen the mountain lion. I’m curious then: is that in any way analogous to what you’ve done with imaging? What you’ve seen are flags that are evidence of these processes, transcription and translation. You haven’t seen these processes occurring separately, but you’ve seen the tracks.
NLW: It’s a good analogy. If you want to take the wild animal metaphor a little further, we could see it as two different animals. Do their ranges overlap? We have these signs that they do or they don’t, or that they sometimes do but not always, or that mostly they do. It’s a good analogy because the imaging is an indirect observation. The other method that I mentioned, in which you would separate the cell and look for a particular protein in a particular part of the cell, that’s much more direct because you can take that fraction and chemically determine exactly what proteins are there, with a high degree of confidence. So that’s a more direct approach. You can also see the relative abundance of one protein in relation to others, and so on. But as I said we can’t do that in our lab.
There are other approaches to biology, where we also look for tracks. So when I mentioned features of the Gemmata cell that make it look somewhat like a eukaryotic cell, there is a particular class of proteins in eukaryotic cells that are responsible for curvature of membranes, which is biologically important. They have a characteristic structure, and one group of researchers found Gemmata proteins that were predicted to have a similar structure to these membrane curvature proteins of eukaryotes, and while that was not completely unique to Gemmata, it was fairly rare. That’s an indirect piece of information that suggests a relationship without actually taking that protein out of a Gemmata cell and putting it in a membrane and making it bend, which would be the direct way of testing that. Often we move from these flags or signposts towards a more direct approach. This often moves science from an exploratory phase to a more hypothesis-driven phase, because a hypothesis needs some kind of background information. Nobody makes a hypothesis in a vacuum. You’ve read something, or you’ve seen something that made you think, “Oh, maybe this is how that works.”
Sometimes getting the direction for a hypothesis requires that exploratory phase. So in the example I just gave about the predicted protein structure, that came from when someone determined the whole genome sequence. (Actually, it was me! Or at least I did part of it. It wasn’t completed, so I can’t honestly say I did it.) Sequencing a genome is about as exploratory as it gets in science. It’s like the natural history of the organism captured in its DNA. You’re going to get a long list of genes and the best estimate of what they do for the organism when they are translated to proteins. Then these researchers ran software on the genome sequence that was looking for this particular structure, and they found it in several predicted proteins. This gave them a short list of Gemmata proteins that may have this membrane curvature function. Even though I said that it hadn’t been directly tested, they took one of those short-listed proteins, tagged it, and found it in parts of the membrane system that are curved. So there is more than just theoretical evidence to support that idea. That’s an example of starting with a very large pool of information that’s generated in an exploratory way with no particular direction in mind, and then you interrogate that: what’s interesting about this very large pool of information? That leads you then to a hypothesis that you can test.
This is one of the challenges of science: which approach is more appropriate at any given time? Often, hypothesis-driven science is prioritized. It’s more focused, and more likely to yield a definitive answer. But it is dependent on the exploratory work that came before, as well as a lot of other hypothesis-driven science. All the connections you were mentioning are drawn together in formulating a new hypothesis. When a researcher submits a proposal for funding, a frequent reviewer comment might be “Too exploratory. This is a fishing expedition.” The phrase “fishing expedition” is used a lot in the review of grant proposals! In many disciplines, it’s seen as a bad thing, a thing you should already have done without being given money for it.
HLH: You’ve led me to another observation that I think is a question. You responses make me think about the interpretive nature of what you’re doing, but I want to be careful not to associate this with the connotation of “interpretation” that makes it loosey-goosey and speculative, guesswork, highly subjective, and so on. I’m thinking about St. Augustine, who in his On Christian Doctrine talks about different kinds of signs. His interest is specifically in biblical interpretation, but he distinguishes kinds of signs, and one of the kinds is what he calls “natural signs,” his example for which is smoke as a natural sign of fire. The thing being interpreted is not a dubious matter, not just a subjective thing: it’s a very strict natural connection that’s being observed. Where there’s smoke, there’s fire. But that’s a very familiar natural sign: we all of us have had perceptual experience of this connection, and there’s broad social consensus about this connection, and so on. We already have all the equipment and the context that we need in order to interpret that natural sign.
But now tell me if this is any way a fair characterization of what you do. Is there any sense in which what you’re doing is creating that context where it hasn’t existed before, or where we haven’t had it before, to interpret a natural sign? We all know how to interpret the natural sign smoke. What you have done with your various experimental techniques is create the context in which we are now able to interpret this natural sign which was there before. Now I can get to the fire, which is the fact that transcription and translation sometimes occur separately. I’ve only actually seen the smoke, but I now have the context that lets me determine what it is a sign of. I’ve never seen transcription and translation occurring in Gemmata obscuriglobus (as I have seen fire), but I now have enough context that the thing I have seen tells me that it occurs separately.
NLW: I never thought about it in that way, but yes I think so. Building that context occurs in a number of ways. One is just the use of multiple approaches: different types of microscopy, different labels, other approaches if you’re able to use them. That’s one way in which we build a context that makes the link believable. Another way we do so is through replication. In other words, imagine that ten times you have seen smoke, but only three times has someone shown you that there is a fire; other times you walk over there and there’s no fire. When you make an observation that supports a link repeatedly, that strengthens the connection. Another way is by having as part of the context experimental methods that have been used to show a similar kind of link but for a parallel question. In our case, we were interested in transcription and translation; somebody else might be interested in where is the activity of an enzyme that is needed to break down sugar? The questions are very different, other than sharing an interest in location, but they likely use fluorescent tags to detect the protein, as we did. So the fact that we were using an established approach (or at least some of our approaches are established) was in our favor. It strengthened the connection.
In our case, where we were technically very constrained, we also (out of desperation!) tried something that hadn’t been done before. Most of the detection that we did was using antibodies. This is familiar to us from the biology of our own bodies: you get sick, your body develops an antibody that recognizes that agent the next time it invades, and protects you. So antibodies are also a very standard tool for locating proteins, and actually capturing proteins and seeing if they are active. We did that, but in our desperate search for a different way to support that antibody work, we also thought about an antibiotic called gentamicin. Gentamicin kills bacteria by binding to the ribosome, the protein factory of the cell. If a cell can’t make proteins, it dies. However, it can also be a flag for us. So if we want to know where protein is being actively made, then we send this antibiotic in, like a guided missile. Yes, it’s going to kill the cell, but because we’ve put a flag on it, we’re going to know exactly where it was when it killed the cell. We didn’t invent the technique of putting the flag on the antibiotic; it had been used before, to look where within the human body the antibiotic goes when it helps make us better. My student Ekaterina found a paper where this flagged antibiotic had been shown to be located in certain cells of the ear, connected to ear infections. But her thought was, well if it can work in an ear, it can probably work for what we want to do. It was not a completely novel approach, but it was a novel application of something that already existed. However, we found out that because it was novel, after we got our reviews back on the first version of the paper, the reviewers wanted some more substantial evidence that this was a valid way to get data to answer this question. We ended up doing more experiments that increased reviewer confidence in the appropriateness of using this tool. So sometimes you have to create new context, and if you do that you have to justify it. If we had had other tools at our disposal, I’m sure we never would have come up with that, because we wouldn’t have needed to. Those are the ways that we strengthen our arguments through providing context.
HLH: I’m conscious of time, so let me pledge that this will be the last question. Keeping in mind our potential reader who will not already be familiar with your work, what in lay terminology are translation and transcription? For me as a poet arriving at your work those terms are very compelling. They’re loaded in terms of language and poetry, but what are they as cellular processes?
NLW: The processes are two different stages of the expression of a gene. (There’s another word, “expression,” taken from nonscientific English.) The bacterium has all of its genes, but at any given time not all of them are being expressed. The same way not every thought in your head gets verbalized.
NLW: A cell can express different genes at different times, depending on its needs and its environment. There’s a whole regulatory framework that determines when the gene is expressed. There are two stages of expression. Transcription comes first. Transcription takes the gene — a unit of information in the DNA, the genome of the organism — and makes a copy. It unwinds the double helix of the DNA so it can get access to the information, and it makes a copy. All of these verbs — transcribe, copy — make a lot of sense because we’re essentially dealing with letters, or chemicals that we represent as letters. The variation in DNA comes from the order of four chemicals that we abbreviate as A, C, G, and T. From a transcription point of view, DNA is just strings of As and Cs and Gs and Ts. In transcription a copy is made, and its purpose is to be a message, so it’s called messenger RNA. RNA is chemically very similar to DNA; they’re both nucleic acids. It differs from DNA in that one of the four letters is substituted by a different letter: Ts become Us, and also differs in the structure of the sugar (ribose) that is a component of both DNA and RNA. As a message, it’s also only one strand, rather than the double helix of the DNA. There are organisms that have double RNAs, but they’re unusual.
That’s transcription: making a message, a transcript from the instructions. Translation takes that message and converts it into a protein. The word translation is used because there you see really a switch in language. In the transcript you have pretty much the same language but with just one letter substituted for another. Translation takes a code that’s embedded in the message; it’s a triplet code, so every three letters means something. The something is a particular amino acid, which are the building blocks of protein. Translation, which occurs in a specialized “factory” in the cell (the ribosome), is the process of decoding the transcript and translating it into a completely different language. I guess another way you could look at it is, whereas a gene and its transcript are maybe dialects of one language, the transcript to the protein is like English to Basque, or Aramaic, or something that looks completely different, and which maybe, way back in evolutionary time, the same way that there are relationships between the romance languages and languages from other places, there’s a tie there. There are linkages: there have to be, or you couldn’t translate. But it looks completely different. So that’s translation, and it ends with the production of a protein.
That’s not the end of the story. A lot more has to happen, in most cases, to that protein before it can do something useful for the cell. It often has to be modified, folded, activated, transported to a particular place in the cell, or even out of the cell. Sometimes the cell has to send the protein out to do a job. There’s also turnover, so proteins have a life and they get degraded, and recycled. But that’s post-translational …
H.L. Hix, author of As Much As, If Not More Than, loves the interview form as a way of thinking together (itself a condition of democracy, justice, philosophy, and other ideals and practices he values), and as one element in a community poetics. A collection of his interviews on The Conversant can be read here.
Naomi Ward is an associate professor in the Department of Molecular Biology at the University of Wyoming. A graduate of the University of Queensland (Australia), she earned her Ph.D. at the University of Warwick (U.K.). Her research interests include microbial community structure and function, interactions between microbes and their environment, and evolution of new function in microbial genomes. The scientific research described in this interview was supported by an award from the U.S. National Science Foundation MCB-0920667.