微生物群落，在自然和实验室 - 面试

My name's Ed DeLong, and I'm a professor at the Massachusetts Institute of Technology. I have a dual appointment actually in civil and environmental Engineering and biological engineering, which may seem like an unusual mix because I have my feet really in two different places. One of the aspects of our research has to do with the environment and specifically with microorganisms, how they influence the environment, how the environment influences them, the, and that's the civil and environmental part, the environmental part of, of our program.

The, the other thing that we study has to do more with biological engineering. What makes microbes tick? How are they put together?

How do the different parts that make up a microbe microbes or microbial communities work together in synergy? So the, the kind of al alliance between these two fields works pretty well. My own specific background is in general microbiology, mixed in a little bit with oceanography.

I, I got my degree in bacteriology at uc Davis, my bachelor's degree. And then I got my PhD at Scripps Institute of Oceanography. My PhD's, actually marine biology, if you can believe that.

And here I am At MIT. How did that happen? The field Of microbiology has been changing really rapidly.

And one of the, the influences, big influences on what we understand about microbes now has been really technological advances is in the area of genomics. Typically, microbes, if you look at them under a microscope, are very hard to identify. There are some different shapes and sizes, but they don't really tell you much about who that organism is or what it's doing.

And so typically, microbiologists, were pretty much relegated to look at microbes in the lab, which has been incredibly powerful. It's kind of an outgrowth of medical microbiology. So you can isolate a microbe from the environment, just like you can isolate a chemical.

And once you have that pure microbe, that pure microbial species you can learn about, its, its phenotype, its characteristics, the different attributes that allows it to function in the environment, so on. There's two problems with that though, with respect to understanding the, the full microbial world. And, and that is microbes in the lab don't always behave the same way that they might behave in the environment.

And the other problem, which is even larger, is that microbial species, for the large part, are really hard to domesticate. The microbial zoos that we have in the lab typically only represent maybe 0.1%of what's out there in nature. So there's a, a vast component of the microbial world that we simply haven't been able to grow in the lab yet, and so we don't understand it.

And yet this represents the, the vast proportion of biomass out in nature. And there's many reasons why that's important. But one of the, the main ones is that microbes are the engines, if you will, of, of the transformations of energy and matter in nature.

So, so fluxes of things like greenhouse gases of energy through ecosystems that, that are, are vast in proportion in terms of some of these gases we're talking about. Gigaton fluxes on an annual basis are really mediated by microbial species in a distributed fashion, although they're very small. It's the power of numbers and the distributed nature of the microbial world throughout our planet moves lots of mass and energy around.

And, and that's one kind of compelling reasons for understanding microbes on their own turf, understanding Them in nature. Well, once Genomics came to the fore, it became easier and easier and cheaper and cheaper to sequence whole microbial genomes. And of course, this started to happen in about the mid nineties.

And once that started to happen, people realized that one could use those same techniques to learn about the properties of genomes, of microbes out in the natural environment without even having to cultivate them. And this kind of is an offshoot of, of earlier studies by really Carl woes and norm pace. And in the mid eighties, norm pace realized that, hey, I can identify microbes without having to grow them.

And the way that was done was to go to any microbial habitat and isolate biomass, and then extract either DDNA or RNA from that and from those nucleic acids. One could, for instance, sequence taxonomically informative molecules like ribosomal RNAs. So you would get a list of different ribosomal RNAs.

They're almost like barcodes for particular taxonomic types. And you could identify the types of microbes in any particular habitat. The limitation of that approach is all you have is an RN gene.

It tells you who the micro is, but it doesn't tell you exactly what they're doing. So now with the application of genomics, you can take that a step further, you can isolate nucleic acids from almost any microbial habitat that you think of, and using a variety of techniques. Isolate either large genome fragments, which are some of the experimental protocols that we're, we're gonna be showing you today.

Or you can blast that DNA into small pieces and do shotgun sequencing. And the easiest way to think about that is that you're getting a parts list of the components of all the microbes that made up that community. And, but that's a parts list that we never had access to before.

So from that parts list, then we can kind of deconvolute what makes up the microbes that are there, what some of their potential functions are. And really what we end up doing is generating hypotheses that we can then go and test using biochemical means. Physiological means going back to the environment and tracking genes in, in space and time.

And this whole area that's evolved, which kind of merges genomics with ecological approaches has been referred to now as metagenomics or environmental genomics. And what one finds today is that these sorts of approaches are going on in, in many different habitats, in soils, in kind of agricultural settings out in the ocean. One of the really, the first largest studies of this type was conducted by Craig Venter out in the Sargasso sea, and also importantly, the human microbiome.

Most people may not realize it, but associated with the human body are more microbial cells than there are human cells. There's about 10 to the 14th microbes and only 10 to the 13th of us. So what I like to say sometimes is, is that God must have had an inordinate fondness for microbes because there's more of them than there is of us.

But these concepts are really important. Think about us. We're never alone.

We always have a microbial community along with us that probably has many different functions. There's evidence now that in fact, with some, some disorders like obesity, the microbial component may have a big influence on, on, on, on basically the human phenotype and how much energy or not we extract from our food sources. So we're a living system composed not just of us, but of the symbiance, if you will, that we carry along with us.

What these new approaches using genomics allow us to do is to, is to begin to dissect those parts of the biological world that we haven't been able to look at before. And therein, therein lies their Power. What's happening in the field Right now is more than just gathering a parts list, but it it's really getting to the notion that there's a real dynamic going on.

Initially when genomes started to be sequenced, of course, a a single genome sequence, really as a singular entity, you have a purified strain of a microbe, it has a genome, and you can determine, say the 2 million bases, base pairs that make up that genome. And it seems like a very fixed and static entity. What we find though, when we go and look at nature, is that there really is rarely such a thing as, as a single strain of a microbe.

When you look into these natural microbial populations, what you find is that there's a theme, a genomic theme around which a species is assembled, if you will. But in any given population, there are many, many, many variants, many variations on that theme, many strains. And what we're starting to see is that genomes are not really static entities, but they, they're ever changing.

They're, they're dynamic. And in fact, they do that in populations. And so you can imagine that as there are subtle variations in the environment that might have to do with nutrients or temperature salinity, that some proportion of those variants in the population can respond to that environmental change a little better.

And there's constant shifting, but they again, are changing too. And so variation is continually supplied by neutral drift and then acted upon by selection. And this gives one a different view, a more dynamic view of what genomes are.

And it's more, instead of like a static picture of a genome, it's more like looking at a motion picture. And one of the powerful things about some of the, the new studies is that we're starting to get the stop frames to put together the motion picture of how genomes evolve, what that means with respect to the phenotype of the organism and its response to the environment and influence on the environment as well. So this, this is quite exciting really because it's a new area for microbiology population genetics, even although one might think would be well developed in microbiology, really isn't because we haven't really had a way to study genetics of natural microbial populations in a very detailed fashion.

And, and that's another component of this sort of research, is that now we can, can delve into that population genetics and population genomics to to watch the motion picture. And, and that's a really quite powerful and, and has many practical implications for, for what we understand about how microbes acquire functions, how functions and properties spread across communities. And one of the surprising things for us is, is that it kind of changes a little bit your notion of how, how phenotypic properties, how how functional properties might, might move amongst Microorganisms in biology, of course The biology of sexual organisms, we typically think about traits being passed on from mother to daughter and so on.

And, and that's how biological properties are, are propagated via that kind of genetic hereditary transmission. The microbial world has a different dynamic though, and we're starting to see that in some of these metagenomic surveys, which have to do, has to do with what's called lateral transfer so that genes or whole arons are passed from very different organisms and by the process of transfer of DNA, and if that DNA is expressed in the recipient organism and has a positive selective value, then it can start to propagate amongst many different organisms. And we've recently seen this using the kinds of techniques that we're, we're looking at in the laboratory, how that, how that works is we're able to isolate very large genome fragments from microbial populations.

And we do that at different points in time and space. In, in this case, what we've been able to do is isolate large portions of genomes along an energy gradient, the gradient of light that penetrates the ocean. And that's a very well-defined gradient in places that we study in the open ocean.

The pho zone is actually quite deep. It extends all the way to 200 meters below the sea, sea surface. And, but the energy of course changes in quality and quantity.

As light gets attenuated, it gets bluer and bluer the deeper you go so that there are only a narrower range of wavelengths available to organisms that live deep. And the energy of light is also less too. So one can start to look at how organisms that use light might adapt and change along that function of the energy gradient, different light quality and quantity.

The interesting thing is you can track particular genes and the sorts of microorganisms that are associated with along these light gradients. And to make a long story short, what we found is that some particular genes involved in, in harvesting of light energy that are called rod dosins, are so easy to express and require. So few genes make the photo system that they've hopped around amongst all different kinds of microbes.

And the reason we can say that is that we've isolated genome fragments from known really vastly different groups as different as bacteria in ArcHa. And we find the same photo system with almost identical properties. So you can track these portions of functional genes as they're moving between organisms In the environment.

It's not just mapping genes on Organisms, but also mapping those genes along the environmental gradient. So the other thing we observe by looking at these large genome fragments that contain these photo systems is that they're quite abundant in the photo zone from depths of about zero meters to 130 meters. But you get down to the base of the photo zone and just below it, those genes disappear.

They're gone even though the same organisms exist deeper. And what that shows us is that in the absence of selection, that is in the absence of light, which allows these genes to function, they're lost amongst all different kinds of organisms as a function of depth. So the reason why that's interesting, well, first of all, it tells us about a particular energy generating mechanism that uses light, where it is in the environment and how it's distributed amongst organisms and the dynamic that moves it around.

But it's also in general interesting, a a new way to think about how to map genes and genomes, not just on organisms, but where those organisms occur in time and space and where those genes occur in time and space and where the intersection is. We're only starting to learn how to do that. That was an example of just one set of genes.

But you could ask the general question, well, how many categories of these cosmopolitan genes that like to move around a lot, are there, we don't know the answer to that question in any given habitat and how common that is, what the mechanisms are for moving those genes around. Is it free DNA in the environment? Is it virus that's propagating the genes from organism to organism?

These are all things that are largely unknown, but that are now approachable using these sorts of techniques of, of mapping genes onto naturally occurring organisms, mapping those organisms onto the environment and understanding their variability. So it's a, it's, it's a really a, a new time for microbiology and we have new access to looking at the motion picture of genomic dynamics out in the environment and, and that that's really put us in a new position to better understand how the microbial world really works Out in nature. One of the challenges Of course, and this is a challenge both for environmental science, and I would say biomedical science too, is, is capturing all the right kinds of data.

And in this case I'm referring to metadata, metadata of course being data about data. If you think about the example I gave in the ocean, there's a number of things you wanna know. You wanna know what the spectral radiance is, how, how, what's the photon flux, what are the wavelengths of light coming through, what was the temperature of that water, the salinity, the pressure, because each of those variables has important consequences with respect to how the organisms evolve.

So besides getting data about the genomes and about the organisms, we need to get fairly rich data about the environment that we can compare to other times, other places, other investigators data. And while environmental scientists have been working with data sets like this for a long time, this merger between bioinformatics and genomics and environmental science is rather new. And databases as they exist right now are not necessarily ready to accommodate such data.

And so there's a little bit of an evolution going on right now, I would say both from the environmental end and since I'm an oceanographer marine biologist, the from the oceanography end trying to crosstalk with folks that are doing genomics and bioinformatics and vice versa. I can give you one example. I'm involved with a an NSF Science and Technology Center called the Center for Microbial Oceanography Research and Education or Seymour.

And at Seymour we're really interested in kind of capturing this whole dynamic, the continuum from genomes to biomes, all these different levels of biological complexity from the genome to the transcriptome, to the proteome, to cells, to populations, to communities, to ecosystems. That's a lot of different sort of information. It's rather ambitious.

And so this project, this NSF Science and Technology Center, Seymour, is, is based in Hawaii and, and, and it's a group of oceanographers, microbiologists, molecular biologists who are trying to get together to figure out how to characterize a particular system in this way. And in this case, the system is largely the upper ocean out in the Central Pacific. An important reason for doing that is that this captures a, a vast kind of biological component that covers our globe, two thirds of our planet in collaboration, Seymour is actually collaborating with another project that's funded by the Moore Foundation called camera.

And Camera is spearheaded by Larry Smar and Craig Venter as a place to capture lots of genomic data that's been collected in the environment and to present it with the synoptic or, or environmental data. So the environmental data that was collected at the same time that the genomic data was. So this is really brand new.

It's just starting and it's gonna require more coordination and, and discussions about data standards and metadata standards. And how do you construct a database that can contain such information? Because right now this is focused on the ocean largely that gives one a focus.

But you can imagine that certainly people studying the human microbiome are gonna need similar kinds of data sets along with clinical data, data about individuals, so on. It's a whole different set of problems, but in many ways it's the same kind of problem capturing the genomic data from natural microbial populations along with all the associated and relevant metadata about where they live, what the conditions Were and so on. The hope is, of course, is that Large relational databases will exist that one can ask fairly interesting questions about particular gene types or categories.

You could think about certain genes that we typically associated with being pathogenic like pill I, are there particular pill I that are always associated with say clinical or pathogenic environments, or are those pill I more widespread and and really just adapted to particular circumstances in, in terms of pathogen formation and so on. There's lots of crosstalk that could potentially happen between these data sets in terms of understanding how the microbial world evolves and how particular adaptations might lead to one or another set of, of kind of environmental relationships or lifestyles. Am I free living?

Am I attached, am I a pathogen? All these things are related. So these relational databases as they grow will probably be really important for both the practical applications as well as the development of theory.

And, and they're not there yet. They Have a long ways to go over the next, Oh, I would say five or 10 years. I think there will be a number of developments, of course in the context of the human genome and, and human medicine and genetics.

The, the thousand dollar genome has been holy grail and technologies are developing quite rapidly, at least for re-sequencing. So that one of the, the big impacts will be kind of the lowering of the bar for just getting genomic informations. By that I mean just the economics.

It's gonna be much cheaper and much faster. And there will probably be wristwatch size devices that you can sequence your own genome on if you want, in the not too far future. That's certainly within the realm of possibility.

When that happens, the game changes a little bit. Right now, a lot of the genomic sequencing happens in very large specialized centers with lots of people. What will happen with genome sequencing, certainly, and it's happening now, is that it'll become much more distributed.

And the interesting thing about that is it poses new challenges to data collection, data standards and data dissemination as the point sources of data generation become more and more distributed. So that's a neat challenge. And and again, it relates to the whole database issue.

Another thing that is starting to happen and will happen even more is that we typically think of, of getting genomic data as a, as a a enterprise where we deposit in a database, it becomes a database and then we move on. DNA sequencing actually has really powerful applications in experimental science. And with some of the new technologies coming to the fore, I think what we'll see is that people will be gathering genomic data more in an experimental mode than just in a data gathering and deposition mode.

And, and one can imagine tracking populations in the environment looking at gene expression, overlaying gene expression patterns on proteomic patterns. All these things are gonna be happening. And, and so we will probably think about genome sequencing more as an experimental tool than as a, a way to to grow databases.

So the these sorts of things are all in the offing and, and, and of course, central to them all will be data capture the infrastructure computational, both in hardware and software to deal with all these data cause the amounts are vast. And, and then on top of that, the kind of development of theory, if you think about biological theory as as, as it applies to ecology in particular, much of it's based on large organisms, large sexually reproducing organisms, and in fact some of the rules of the game and some of the theory doesn't necessarily apply to the microbial world. And so I think for mathematicians and theorists, there's a huge amount of potential in the future for using these data that are now being collected and will be in the future as, as the kind of data models, if you will, for, for developing new theory and predictive theory.

And the end game really ends up being understanding how our planet works, partly because the air that we breathe, the, the fluxes of matter and energy that that really balance the elemental cycles on earth are largely regulated by the collective activities of microbial species. Since we are pushing right now as a species ourselves just as hard on some cycles of energy and matter, take CO2 in the atmosphere for instance, it becomes more and more important to understand the dynamics and details of how the natural balance occurs as we per procured that ourselves and begin to think about ways how we can live a little bit more in balance with the cycles of matter and energy on, on our planet. So I, I believe that some of this kind of foundational, what seems like reduction of science is, is really going to end up evolving through metagenomics and some of these other techniques I talked about and theory into a more holistic science in terms of understanding complex systems, not just at the level of the cell, but at the level of our planet.

And so really right now with some of these techniques we showed you in the lab and, and, and where the field is right now, we're just taking the first baby steps. But I think we can start to see, you know, tiny glimpses of, of where this might head. And it'll, it'll be exciting if we can figure out how to orchestrate it because it requires the skills of, of a vast array of, of technologists, of theorists, mathematicians, engineers, to try and put it all together.

So it's, it's interesting times and it's bound to get more interesting at this point.