Synthetic Biology and the Regulation of Bacterial RNA Polymerase

Top

00:00:11.11 Today, I'm going to talk about the bacterium,
00:00:14.05 Escherichia coli, and synthetic biology. I'm going to focus
00:00:19.08 on transcription and its regulation. What I want to try
00:00:24.09 and convince you in the next 20 minutes or so, is that
00:00:27.13 by understanding the mechanism of transcription and
00:00:31.00 its regulation, we can lead on to the development of
00:00:36.06 switches and tools that can be useful in the exploitation
00:00:41.23 of Escherichia coli in synthetic biology.
00:00:44.23 I think it's important to understand that manipulating
00:00:49.17 transcription is a means to an end, but not an end
00:00:54.02 in itself. So let me start off by telling you a little
00:00:57.05 bit about Escherichia coli. A typical Escherichia coli
00:01:01.15 cell has a single chromosome that contains something
00:01:06.00 like 4500 genes, and these genes will be organized
00:01:09.04 into 3000 transcription units. The amazing thing is
00:01:14.03 that there's a single species of RNA polymerase
00:01:17.19 that copies those transcription units into RNA.
00:01:22.09 Now if you bust open a typical Escherichia coli
00:01:25.07 cell, you'll find 4000 or so molecules of RNA polymerase.
00:01:31.00 Now, at first sight, that looks as if there's plenty of RNA
00:01:34.16 polymerase to go around all the transcription units, but actually
00:01:38.19 that's an illusion. Because it turns out that some of the
00:01:42.13 transcripts -- a small number of transcription units receive
00:01:47.18 a lot of RNA polymerase. So what this means is that
00:01:51.16 there are a lot of transcription units that are actually
00:01:54.04 -- that are short of RNA polymerase, in other words,
00:01:56.17 in the cell RNA polymerase is in short supply.
00:02:00.11 So, Escherichia coli is very good at distributing its RNA
00:02:05.12 polymerase between different genes, according to the
00:02:08.19 need. Need being which gene needs to be expressed
00:02:12.00 at a particular instance. And the bottom line is the result
00:02:15.23 is that certain genes get a lot of RNA polymerase, certain
00:02:19.20 don't get very much. And what I want to convince you is
00:02:23.06 that by understanding these rules, that govern the
00:02:27.01 distribution of RNA polymerase, we can then subvert
00:02:29.11 them and create switches that of course can be used
00:02:34.08 to enable synthetic biology in Escherichia coli.
00:02:38.24 So, Escherichia coli is a very good tool or chassis, if you
00:02:44.21 wish, for doing this. Now one other thing about Escherichia
00:02:48.02 coli that makes it a good chassis is that actually, there's no
00:02:51.18 such thing as Escherichia coli, there's no such one thing
00:02:54.19 as Escherichia coli. Actually there are millions, if not
00:02:58.02 millions of millions of different sorts of Escherichia coli
00:03:01.08 swimming around in the world. Because, you see the
00:03:05.20 Escherichia coli that we use in the lab is just one
00:03:09.06 of millions and millions of species. And if you look
00:03:13.04 at the sequence of these species, what you see is an
00:03:16.02 enormous diversion in sequence. Actually, the number in
00:03:19.04 common genes between the different species is actually quite
00:03:22.05 small. So what this is telling us is over billions of years,
00:03:26.19 Escherichia coli has diverged and has picked up genes,
00:03:30.22 it's lost genes, and actually it's very, very good at picking up
00:03:35.04 genes, losing genes, and adapting. So of course, from a
00:03:38.12 synthetic biology point of view, this makes it the perfect
00:03:42.08 chassis, because what we want to work with is something
00:03:45.03 which is capable of receiving genes and losing genes.
00:03:49.11 And is going to be adaptable. Okay, so now we
00:03:53.10 need to just focus on the topic, which is the transcription
00:03:57.24 of DNA into RNA and what I'm going to do is, I'm going to show
00:04:05.17 you some models of this wonderful enzyme that does
00:04:10.05 this job. This enzyme's called RNA polymerase. It's a little
00:04:14.13 molecular machine consisting of different subunits,
00:04:17.11 the two major subunits are called Beta and Beta-primed.
00:04:20.05 So they're colored in this board diagram as blue
00:04:23.11 and this rather gaudy pink color. Basically, these two
00:04:26.17 subunits form what's called a crab claw. There's a gap
00:04:30.22 between them, and basically what happens is the DNA
00:04:32.22 that's being transcribed is threaded through. There's a
00:04:36.11 rather more complicated picture of that here. And
00:04:41.03 basically, what this wonderful motor does is it motors along
00:04:45.18 the DNA, copying DNA sequences into RNA. This form of
00:04:51.01 the enzyme, by the way, is known as the core enzyme,
00:04:53.21 consists of the two big subunits that form the active site of the
00:04:57.03 enzyme that actually does the job. And the two big subunits,
00:05:01.07 B and B', are held together by two alpha subunits, so these
00:05:05.07 are shown here in yellow and orange. And this is a highly
00:05:10.06 conserved structure that is found actually in most living
00:05:17.05 cells. Now, the problem with this structure is that while
00:05:21.07 it's good at making RNA, it is incapable of knowing where
00:05:26.20 to start, and starting transcription. Turns out in bacteria
00:05:32.01 starting transcription is really important, because by starting
00:05:37.02 transcription at specific positions, first of all, it ensures
00:05:42.15 that full-length messages are made, but also it allows
00:05:46.01 for regulation. So, the question is, how does bacteria
00:05:50.24 solve the problem of directing this RNA polymerase to
00:05:55.08 specific starts. And basically, bacteria solved this problem
00:05:58.15 by using a very, very simple tool. And this tool is called
00:06:02.11 the sigma subunit. So there's an extra subunit called sigma,
00:06:07.06 that evolved, whose job is really to guide RNA polymerase
00:06:13.01 to start at specific positions. Now, interestingly bacterial
00:06:17.12 use sigma subunits, other organisms, eukaryotes, for example,
00:06:22.19 use a completely different method for solving the same problem.
00:06:27.02 And of course, what this is telling us is that bacteria
00:06:29.21 are of course, during their evolution branched off
00:06:32.14 from the eukaryotes a long time ago and basically
00:06:36.19 have used sigma subunits to drive regulation. So
00:06:41.17 let's take a look at a sigma subunit. Here's a very
00:06:44.08 very elementary diagram of a typical sigma subunit.
00:06:47.16 Most, not all, sigma subunits carry 4 independently
00:06:52.18 folding domains. And like in all proteins that contain
00:06:56.23 independently folding domains, the independently folding
00:06:59.14 domains each do an individual job. In the case of sigma
00:07:03.18 subunits, the domains recognize different bits of
00:07:08.23 DNA sequence at promoters. I remind you that promoters
00:07:12.20 are the sequences at the beginning of genes that specify
00:07:16.14 where transcripts start. And to get a long story short,
00:07:21.11 the four independently folding domains are referred to as
00:07:25.02 domain 1, 2, 3, and 4. Domain 2 contacts something called
00:07:29.05 the -10 element, that targets promoters. Domain 4 contacts
00:07:33.19 something called the -35 element, that targets promoters.
00:07:37.24 Now let me show you a sketch which hopefully will make this
00:07:41.09 really clear for you. This is a sketch derived from some brilliant
00:07:45.04 structural biology done by the Seth Darst lab about 15
00:07:48.13 years ago. Basically, the Darst lab solved the high resolution
00:07:53.00 structure of a bacterial RNA polymerase molecule carrying
00:07:58.02 a sigma subunit. And what you can see, if you look really
00:08:01.08 hard, is you can see the crab claw, the B and B' subunits
00:08:05.19 are colored light blue and pink. And basically, some of the
00:08:11.06 domains of the sigma subunit show up in their structure.
00:08:13.22 And they're shown up in this gold structure, so we have
00:08:17.24 domain 2, domain 3, and domain 4. Actually in this structure,
00:08:21.21 domain 1 doesn't show up. And basically, this very, very
00:08:27.06 simple diagram that comes from a most amazing piece of work
00:08:30.15 shows how sigma subunits work. Basically, the different
00:08:34.21 domains of sigma are splayed across the surface of the
00:08:39.11 core enzyme and they provide a template for the recognition
00:08:43.12 of DNA. And if you look very closely, at this slide, you can
00:08:46.12 see how the three domains of sigma, shown here, 2, 3, and 4,
00:08:51.00 recognize three individual segments or elements of the
00:08:56.20 promoter. Now, there's a simpler way of looking at this
00:08:59.14 and this is just to show a diagram. So here's a schematic
00:09:05.09 diagram of what you just saw, grossly simplified, but basically
00:09:10.02 showing the sigma subunit in orange here. And the idea is
00:09:14.01 that the three domains of the sigma, 2, 3, and 4, contact
00:09:17.17 three different elements at the promoter. Now there's one
00:09:22.04 other point that I need to make here, and this turns out to be
00:09:26.13 really important in a moment. And this point concerns the
00:09:30.14 alpha subunits. Remember I told you that RNA polymerase
00:09:34.07 contains two alpha subunits, and I told you that the alpha
00:09:38.09 subunits are responsible for the assembly of the B and B'
00:09:42.13 subunits. Turns out that each alpha subunit actually
00:09:46.10 contains two domains, an N-terminal domain, a large
00:09:50.11 N-terminal domain, and a small C-terminal domain.
00:09:54.02 Turns out it's the large N-terminal domain that does the
00:09:57.20 holding together of the B and B'. And the small C-terminal
00:10:02.15 domain, shown here as these two cherry-like things
00:10:05.18 attached to the N-terminal domain by a line, which represents
00:10:09.22 a flexible linker. It turns out these two C-terminal domains
00:10:14.11 fold up into a structure that recognizes yet another
00:10:18.02 element at promoters, and this element is called the
00:10:21.07 UP-element. So all together, when RNA polymerase
00:10:25.11 recognizes a transcription start site, there are four
00:10:28.20 main interactions, three made with different elements
00:10:33.12 by the sigma factor, and one made by the two alpha
00:10:37.11 CTDs. And basically, together these elements drive
00:10:43.11 RNA polymerase to promoters and position RNA polymerase
00:10:48.24 so that it can begin transcription at specific positions.
00:10:54.08 One thing you need to know, just before I move on,
00:10:59.01 is that different combinations of these four elements
00:11:03.08 are found at different promoters. So not all promoters
00:11:07.05 have all four elements. And actually, the efficiency of
00:11:11.04 any particular promoter is determined by the combination
00:11:17.04 of the elements. And rather like you can make up a
00:11:21.16 pound or a dollar, I guess I should say, with various
00:11:26.21 small coins, you can make up a promoter with various
00:11:30.05 combinations of these four elements.
00:11:33.19 Now, a moment ago, I told you that different transcription
00:11:38.15 units receive different amounts of RNA polymerase in
00:11:43.18 Escherichia coli, the question is why. So if we look at
00:11:47.14 the textbook, we'll see that there are three reasons. And
00:11:50.11 they're listed here. First one is what I just told you,
00:11:53.04 the different promoter sequence elements, the -10,
00:11:56.06 the -35, something called the extended -10, which is the region
00:12:00.21 between the -10 and the -35, and the UP elements
00:12:03.18 differ from one promoter to another. So according to
00:12:07.06 the precise sequence, a promoter is going to be able to
00:12:11.01 capture polymerase more or less efficiently. Second factor
00:12:15.04 is the sigma factor. Turns out that many bacteria don't just
00:12:21.15 contain one sigma factor, they contain multiple sigma factors.
00:12:25.14 Turns out that most E. coli strains contain seven sigma factors,
00:12:29.08 a major sigma factor which is called sigma 70, and six other
00:12:32.04 sigma factors. These six other sigma factors come into play
00:12:36.04 in response to certain stresses. And basically what they
00:12:40.11 do is these sigma factors capture enzymes and drive
00:12:43.24 it to promoters specified by domains 2, 3, and 4 of these
00:12:49.18 alternative sigma factors. I'm not going to say anything more
00:12:52.13 about sigma factors now, but you should be able to
00:12:55.11 see how by changing sigma factor, you can actually change
00:12:59.19 promoter specificity. This is a strategy used by many
00:13:04.00 bacteria to alter gene expression in response to external
00:13:09.06 cues. And of course, this is something that synthetic biologists
00:13:12.17 are going to be able to exploit in the future. And particularly
00:13:16.09 it's going to be very well placed to exploit it because
00:13:19.20 we know that sigma factors are made up of independently
00:13:23.24 folded domains. So it's not rocket science to see how you could
00:13:27.10 alter particular domains to alter promoter specificity.
00:13:32.17 Now I'm not going to say anything more about sigma factors,
00:13:34.24 because I want to move on rapidly to the third mechanism
00:13:40.04 that drives the distribution of RNA polymerase between
00:13:43.10 different promoters, which is transcription factors.
00:13:46.04 And E. coli contains somewhere between 250 and 300
00:13:49.15 of these factors, or I should more accurately say most
00:13:53.14 E. coli strains contain this sort of number of transcription
00:13:57.21 factors. And these come in two flavors, activators
00:14:01.07 and repressors. So, I guess most of you will know that
00:14:04.13 repressors function by binding at active promoters, so
00:14:08.23 these are promoters that have good -10, -35, UP-elements.
00:14:12.09 Repressors function by binding to those promoters and
00:14:16.01 shutting down their expression. Activators do the inverse,
00:14:20.08 so activators interact with promoters that are defective
00:14:24.17 in some way, such that the promoter is not receiving
00:14:27.10 enough RNA polymerase. The job of the activator is
00:14:30.22 to reverse that and make sure the promoter receives
00:14:35.03 more RNA polymerase, thereby driving transcription.
00:14:39.13 Interestingly, most of these transcription factors contain,
00:14:43.17 also contain domains, most of them, not all of them,
00:14:47.15 but most of them contain what I call a "business" domain,
00:14:50.03 so that's the domain that actually binds to the promoter
00:14:53.00 and does the business of activation or repression.
00:14:56.07 And then most transcription factors contain another domain
00:15:00.04 which is often referred to as a "regulatory" domain, and that
00:15:02.24 ensures that the transcription factor responds to a particular
00:15:06.18 environmental cue. Now again, it's not rocket science
00:15:09.19 to see that by mixing and matching regulatory domains
00:15:13.17 with business domains, a synthetic biologist can create
00:15:17.01 a whole bunch of different transcription factors that
00:15:22.02 can do desired jobs. Now what I want to focus
00:15:25.23 on now, for the rest of the talk, is activators. Because I
00:15:30.06 want to argue, what I want to explain to you first, is how
00:15:33.01 activators work. And then on the basis of that, I want to explain
00:15:36.19 to you how we can create new promoters that are regulated
00:15:42.20 by different combinations of activators. Now just for
00:15:46.08 completeness, I should say that you could do the same
00:15:48.04 with repressors, but for this talk I'm just going to focus
00:15:52.15 on activators. So number one question is, how do
00:15:56.10 activators work? Well, the start point is what I just told you
00:16:00.09 and that is activators function to recruit RNA polymerase
00:16:04.22 to promoters, where the different promoter elements are
00:16:09.21 insufficient to recruit enough RNA polymerase.
00:16:13.18 And a whole lot of studies done by many labs across the world
00:16:17.02 have actually shown the mechanism of action of activators
00:16:19.15 is quite simple. So most activators are dimers, they bind
00:16:23.08 just upstream of the target promoters, and they contain
00:16:26.19 a little patch shown here as a yellow spot. And this little
00:16:31.15 patch is called an activating region. And what's going to happen
00:16:34.21 is this little patch is going to interact with a part of RNA
00:16:39.15 polymerase directly, recruiting the RNA polymerase to that
00:16:44.02 promoter. Why does it need to do that? Well, because
00:16:47.14 the various promoter elements are insufficient to do that.
00:16:51.03 Now, of course it's easy to use Powerpoint to draw this, so
00:16:55.00 here we are. Here comes the RNA polymerase and this
00:16:57.22 tells us that many activators function by making an interaction
00:17:03.07 with the C-terminal domain of the RNA polymerase, alpha
00:17:07.17 subunit. Now, there's an interesting point that I must just
00:17:11.11 remark here, because polymerase contains two alpha subunits.
00:17:15.03 And most activators contain two identical subunits, each
00:17:20.07 of which would have an interacting region, but it turns out,
00:17:23.04 that actually, in order to recruit RNA polymerase, you only
00:17:26.14 need one interaction. And this has been proved experimentally.
00:17:30.22 Now, another thing about this sort of activation, which we
00:17:36.03 call activation by recruitment, or activation by velcro, because
00:17:40.18 at the end of the day, the activating region is just like a little
00:17:43.06 velcro patch that hooks the RNA polymerase to the
00:17:47.20 promoter, is that this activation is crucially dependent on the
00:17:51.21 location of the activator upstream of the promoter.
00:17:56.03 So, you take a single promoter and you move the activator
00:17:59.07 around, you'll find some locations where it works and
00:18:03.13 some locations where it doesn't work. And interestingly, what
00:18:07.07 you find is that the locations where it works are normally
00:18:11.11 separated by 10-11 base pairs, in other words, one
00:18:16.01 turn of a helix. This is some data, very old data from my
00:18:19.24 lab, in which we took a promoter that was being activated
00:18:23.22 by a single activator and basically what we did was we
00:18:27.04 moved the gray boxes, the activator, to different locations.
00:18:32.04 So these locations are shown on the x-axis, the y-axis shows
00:18:37.01 the activity of the promoter. And what you see very, very clearly
00:18:39.20 is that there are some locations where it works, some where
00:18:42.19 it doesn't, and the distance between the locations where
00:18:45.11 it works, correspond to the turn of the helix. So the idea is
00:18:48.23 that in order for activation to take place, the activator
00:18:52.14 and the polymerase have to be lined up on the same
00:18:56.12 face of the helix. Now, course you could ask a very, very
00:19:00.11 interesting question here. Well, hang on a second, if
00:19:03.00 you move the activator by one base pair, you're twisting
00:19:07.06 the activator around, around the helix, why can't the DNA
00:19:12.15 just twist back or if that linker that joins the C-terminal
00:19:16.13 domain with the N-terminal domain, why can't
00:19:19.18 that join? Because the thing is, that requires an energetic
00:19:23.00 penalty. You have to pay energy to do that, and the fact
00:19:27.12 that these spikes are so sharp is telling you that that
00:19:32.22 energetic penalty is too much to pay. And this brings me to a
00:19:38.02 really interesting experiment proposed just a couple of years ago
00:19:42.11 by a professor of biochemistry at Peking University,
00:19:46.18 Yiping Wang, and he suggested that if you could
00:19:51.24 increase the binding of the activator to the RNA polymerase
00:19:57.04 then maybe these peaks would broaden, maybe some of these
00:20:04.23 locations where there was no activation could become
00:20:10.03 locations where there is activation. So his idea was very,
00:20:14.02 very simple. And together, we did some of these experiments
00:20:17.21 and basically, rather than presenting you with all his data,
00:20:21.00 I've just presented you with some cartoons. So we start
00:20:24.04 at the top, this is sort of the promoter that I spoke about.
00:20:26.23 And I want you to imagine that the activator is misplaced,
00:20:31.23 say by one base pair, such that it doesn't work. What Yiping's
00:20:35.24 student did was introduce an UP-element, so this is shown
00:20:39.14 as the blue rectangle here, just downstream of the activator,
00:20:45.00 in the middle. That, of course, increases the binding
00:20:48.18 of alpha CTD to the DNA and actually this little menage-trois
00:20:54.05 of the DNA, the alpha CTD, and the activator, the three
00:20:58.17 components bind cooperatively together and turns out
00:21:01.20 that when you measure the activity of the promoter, this
00:21:04.11 promoter activity actually goes up. So the conclusion
00:21:08.04 of this is that if you beef up the binding, you actually allow
00:21:14.10 the activator to function at a location where it wouldn't
00:21:19.19 normally function. Now, in a moment, I hope you'll see
00:21:23.14 why this is important. Following that, we had a great
00:21:28.06 idea that complemented the Chinese experiment, and that
00:21:31.13 was hey, rather than putting an UP-element, why not
00:21:34.14 put another activator? So this is shown in the bottom
00:21:37.16 of the -- the bottom sketch here. And the little red
00:21:42.02 square, or little red rectangle, that's another activator,
00:21:45.22 and it turns out you can produce exactly the same effect
00:21:48.21 just by putting in another activator. So, essentially just by
00:21:53.17 working with these simple principles, we created a promoter
00:21:57.06 that is dependent on two activators. Now, why is this
00:22:02.23 important for synthetic biologists? Well, it's important
00:22:06.14 for the following reasons. Of course, it would be easy for
00:22:10.23 a synthetic biologist to take the information I just told you
00:22:14.09 and design a promoter that was triggered just by a single
00:22:19.12 activator, a new activator. That would be easy. But
00:22:23.06 it would be much smarter for the promoter to be
00:22:26.24 co-dependent upon two signals, rather than one, or
00:22:31.11 even three signals, because if you could do that you could
00:22:34.19 produce combinatorial regulation. And so this little experiment
00:22:39.22 here actually suggests a great idea, which is that one
00:22:44.05 could exploit the fact that RNA polymerase has two alpha
00:22:47.06 subunits and that activators can function independently
00:22:51.00 binding RNA polymerase to increase the recruitment of
00:22:56.09 RNA polymerase to promoters to create switches,
00:22:59.24 at which expression was actually dependent
00:23:05.16 upon two activators. The question is, has E. coli
00:23:09.21 thought of this already? Metaphorically, of course.
00:23:12.15 And the answer of course, is yes. This is our typical activator
00:23:16.17 doing its stuff by interacting with alpha CTD, it turns out
00:23:19.24 that there are many, many examples of naturally
00:23:23.07 occurring promoters, where a second activator works
00:23:27.00 by interacting with the second subunit of RNA polymerase.
00:23:30.20 Now what I didn't tell you earlier was that there are dozens[,00:23:33.19]if not scores, if not hundreds, of activators that work like this.
00:23:37.16 So, in this example, I've just shown schematically with
00:23:42.02 the yellow and the red, but you -- actually, I think you can see
00:23:46.13 this mechanism could work with pretty much any activator
00:23:50.13 which played this game. And of course what this does is
00:23:54.02 opens the possibility to new combinations. These are just
00:24:00.09 some data to show that it really does work in the lab.
00:24:05.13 So in this experiment here, what we've done is we've
00:24:08.11 taken a promoter with a single activator, anchored one
00:24:13.11 position, we've then moved the position of the second
00:24:16.20 activator on the DNA. So again, it's the same deal.
00:24:19.21 The x-axis denotes the position of the activator that
00:24:24.10 we're moving, the y-axis denotes the activity. What you can
00:24:28.01 see is there are locations where the second activator works
00:24:30.23 and locations where it doesn't work. And again, the
00:24:34.05 phenomenon is face of the helix dependent, in other words,
00:24:38.13 things have to be lined up on the same face of the DNA.
00:24:42.21 Okay right, so this slide shows a few examples taken
00:24:48.11 from the literature and actually on here, there's one example
00:24:52.06 where Ann Hochschild's lab at Harvard Medical School actually
00:24:57.04 took two activators that normally don't talk to each other
00:25:00.04 and showed that they could function synergically together.
00:25:04.01 So just to summarize, the idea is that most activators
00:25:08.07 in E. coli function via mechanism like this similar mechanism
00:25:14.05 and because they function by making contact by different
00:25:18.01 patches on the RNA polymerase, we can mix and match
00:25:22.05 different activators to produce new combinations.
00:25:27.03 But of course, in order for this to work, for the promoter
00:25:30.19 to be co-dependent on both activators present at the same
00:25:35.07 time, what you have to do is you have to stop the promoter
00:25:40.02 being activated just by a single activator. And in order to
00:25:43.24 do that, you play this little trick of misplacing one of
00:25:48.09 the activators. We call this the independent contact model
00:25:52.03 for transcription activation. Just to show you that I'm not
00:25:56.12 making this up, here is some data produced by a member
00:26:01.06 of my lab, Doug Browning. So what Doug has done here is
00:26:04.09 he's compared the activity of a test promoter with either
00:26:09.22 one activator or two activators. Now in both cases, you
00:26:13.15 see when you go from one activator to two activators, the
00:26:16.10 activity increases. That's the y-axis. And what you see is
00:26:21.05 the starting promoter, this is on the left here, the activation
00:26:27.01 is probably something like 2-fold, and that's because
00:26:29.24 the first activator is pretty good at activating. But if you
00:26:33.18 misplace the first activator, so now you moved this
00:26:37.12 slide on the right, now misplaced the first activator
00:26:40.22 the ability of the activator to activate falls right down.
00:26:44.07 And now you get incredible synergy, I think it's 15-fold.
00:26:47.20 15-fold synergy. And you can play this game forever,
00:26:51.15 to tune your promoter to get the output that you want.
00:26:56.13 So we call this activation by independent contacts, and
00:27:00.13 I just stress that one of the reasons it works is because
00:27:04.02 the RNA polymerase has two alpha subunits, and each alpha
00:27:09.04 subunit has a C-terminal domain. And I would just point out
00:27:12.07 these C-terminal domains are highly conserved throughout
00:27:15.23 most bacteria. Interestingly, they're not found in eukaryotes.
00:27:22.24 In eukaryotes, the subunits that encode the -- or fulfill the
00:27:28.03 alpha function, one subunit does have a C-terminal domain.
00:27:32.00 the other doesn't. Okay, so this is activation by independent
00:27:38.09 contacts. Just like to finish off by asking a question
00:27:42.04 which is, is there another way of doing it? And yeah, there is.
00:27:45.18 There is another way of doing it. Let me show you what that
00:27:48.09 way is. So, we're thinking about a promoter that's codependent
00:27:53.20 on two activators, let's call them A and B, okay?
00:27:57.09 So here are A and B, now A and B in this case, don't
00:28:01.24 bind to the DNA. Remember in the previous case, A
00:28:05.13 and B bound independently to the DNA. But in this case,
00:28:08.20 A and B have to interact together before they bind to the
00:28:12.14 DNA. And when they bind, that recruits our RNA polymerase.
00:28:18.13 So basically, this creates the same thing. This creates
00:28:22.11 codependence, you have codependence on A and B,
00:28:26.02 but this time, the codependence is due to the fact that
00:28:29.08 A has to bind to B, and B has to bind to A, before
00:28:32.24 the complex binds to the DNA and interacts with the RNA
00:28:36.19 polymerase. Now, it turns out that in bacteria, or perhaps I
00:28:42.10 should rephrase this, in the bacteria that have been studied
00:28:44.23 so far, because I'll remind you that so far, probably only
00:28:49.06 I think north of 1% of bacteria have been studied. So these
00:28:53.13 ground rules that I'm enunciating that come from E. coli
00:28:56.20 that apply to E. coli, might not apply to some other bacteria.
00:29:00.09 In fact, they probably don't, but we just don't know about it.
00:29:02.17 But in the cases that we've looked at so far, you very,
00:29:05.22 very rarely find this. You find independent binding of A and B,
00:29:10.08 and A and B making independent contacts. And you find
00:29:13.02 a lot of that, but you don't find much of this. And in fact,
00:29:16.01 I believe the last time I looked, I found just three examples
00:29:19.20 amongst thousands of examples. Of course the question is,
00:29:22.17 why is that? And I believe there is a simple explanation.
00:29:26.06 And it comes from the dynamic genomes of bacteria.
00:29:30.11 The fact that bacterial genomes are dynamic and changing
00:29:34.05 all the time. Remember, I told you at the beginning,
00:29:36.13 that there are millions and millions of species of E. coli.
00:29:40.05 Millions and millions of types of E. coli that have widely different
00:29:44.12 genes. Now, the point is, if you're going to fix coregulation
00:29:50.05 by A cooperatively binding to B, A has to have a surface that
00:29:55.10 binds to B, and B has to have a surface that binds to A.
00:29:58.21 So the two have to be committed to each other. And if
00:30:02.09 they're committed to each other, then they're fixed.
00:30:05.08 Okay? Now, of course, if A now wants to go and play with
00:30:10.10 C, it can't. Or D. Or E. Or F. So what I'm trying to get over
00:30:15.06 is that I believe that the independent binding model that
00:30:21.04 I just showed you, which we find very frequently at
00:30:26.19 coregulated promoters in E. coli. I believe that that has been
00:30:31.06 favored in the evolution over the cooperative binding model
00:30:34.17 simply because it allows far more flexibility. It allows far
00:30:40.05 more mix and match. Now interestingly, in eukaryotes,
00:30:45.19 this mechanism of cooperative binding is found very,
00:30:50.15 very commonly to explain codependent activation of genes.
00:30:56.21 And I think there's a basic divergence here in strategy.
00:31:02.02 Anyway, that's a hypothesis that nobody can prove
00:31:05.04 or disprove, but it's an interesting thought. And at least
00:31:08.07 it's an explanation that attempts to explain what we see.
00:31:15.04 And also, it gives synthetic biologists plenty of room
00:31:19.00 for them to play. And on that topic, I think I want to just finish
00:31:24.07 off by stressing this number, 250-300. There really are
00:31:30.08 a lot of transcription factors, and this is illustrated in this
00:31:34.23 wonderful diagram that's actually totally out of date
00:31:39.00 and very, very old, but it makes the point perfectly.
00:31:43.08 This was taken from a review by Julio Collado-Vides,
00:31:46.17 Collado-Vides runs a website in which he tries to collate
00:31:53.22 all the information on E. coli transcription factors and
00:31:58.12 what you see here is that there's a lot of cross-regulation.
00:32:02.13 Most regulators regulate more than one target, most targets
00:32:06.24 are regulated by more than one regulator. There are a lot
00:32:10.12 of regulators and hence, there is a lot of room for synthetic
00:32:15.11 biologists to play. Most of these regulators regulate, not all,
00:32:21.01 I stress there are some exceptions, but most of these
00:32:24.02 regulators regulate by the mechanisms, the simple mechanisms,
00:32:28.02 that I showed you. So my ending point is that if you take this
00:32:33.05 and you take the possibility of mixing and matching different
00:32:37.22 domains, there really is a lot of evolutionary space
00:32:40.09 into which we can move. So I'd like to finish by thanking
00:32:45.00 you for your attention.

This Talk

Speaker: Steve Busby

Audience:

Researcher

Recorded: June 2015

Talk Overview

Dr. Busby gives an overview of the function and regulation of bacterial RNA polymerase. The distribution of RNA polymerase between genes is determined by interactions between RNA polymerase sigma factor, promoter sequences and activating or repressing transcription factors. By engineering additional activators into E. coli and changing where they bind to DNA, Busby and colleagues have been able to alter the regulation of bacterial RNA polymerase on specific transcription units.

Speaker Bio

Steve Busby

Steve Busby is a Professor and Head of the School of Biosciences at the University of Birmingham, where his work focuses on understanding the mechanisms that control gene expression in bacteria. His lab has made fundamental contributions to the understanding of transcription factor activity and promoter recognition by RNA polymerase. Busby is a Fellow of… Continue Reading