Mechanistic Insights into Engineered Riboswitches

00:00:11.16 Hi, I'm Beatrix Suess and I will talk
00:00:13.19 about engineered riboswitches.
00:00:15.19 RNA is a very interesting and very fascinating
00:00:20.20 molecule due to its internal flexibility
00:00:24.00 it can adapt different conformation or alternative structures.
00:00:30.02 And nature uses this for control of gene expression.
00:00:35.11 And these are the so-called riboswitches.
00:00:38.05 Riboswitches are structured RNA elements which are located
00:00:43.19 in the 5' untranslated region of many bacterial
00:00:48.01 messenger RNAs and they are composed of two domains,
00:00:53.26 the aptamer domain in orange can recognize the ligand
00:01:02.12 with very high affinity and specificity, and we have
00:01:06.07 a second domain, the expression platform,
00:01:08.22 and the expression platform interprets the binding
00:01:12.13 state of the aptamer and leads to the regulation of
00:01:17.29 gene expression. Either by transcription termination
00:01:20.22 or by the control of translation initiation.
00:01:24.25 There are several advantages of riboswitch control.
00:01:29.25 We have a direct interaction between the ligand and
00:01:34.20 the RNA with a very high affinity and specificity.
00:01:38.11 And the RNA here is a sensor but also the regulator,
00:01:42.08 here we do not need further cofactors for regulation.
00:01:46.13 Due to the molecular nature of RNA, you can construct
00:01:50.23 even more complex RNA regulators. So this
00:01:56.29 mechanism or this riboswitch mechanism for
00:02:04.12 to build up engineered riboswitches for the conditional
00:02:08.07 control of gene expression. And another advantage of RNA
00:02:12.25 is that you can build up such binding or such sensing domains,
00:02:20.18 the so-called aptamers, de novo by a process called in vitro
00:02:25.07 selection or SELEX. And this method I want to introduce you
00:02:30.13 here on the next slide. You start with an initial
00:02:36.18 RNA pool, which is very huge library of different RNA
00:02:42.24 species. It's a 10^14 different RNAs. And then you
00:02:48.24 have your target of choice, and you immobilize your target.
00:02:53.10 And then you incubate the RNA with your target,
00:02:58.12 and you will remove all the RNAs which cannot bind
00:03:02.13 to your target, but a few of them have the ability to recognize
00:03:06.19 your target and you can specifically elute them and
00:03:11.02 then transcribe them back into DNA. You can amplify
00:03:18.02 the DNA and with this DNA, you can then in vitro transcribe
00:03:22.11 back into your RNA and then you end up here after the first
00:03:28.07 cycle. And then you can repeat this selection process
00:03:32.25 several times. And if you are lucky, after let's say five,
00:03:37.28 six, seven, ten or fifteen rounds of selection, you will end up
00:03:42.29 with small RNAs which can recognize your ligand of choice.
00:03:48.12 With a very high affinity or specificity. In principle, you can
00:03:54.04 make this procedure against a multitude of targets. So
00:04:03.08 you can find aptamers against small molecules, against
00:04:08.04 peptides, proteins, or even against whole cells.
00:04:12.17 And yeah, finally, you will end up with such small guides
00:04:19.11 here, which in this case recognize the drug tetracycline.
00:04:28.07 Such aptamers are characterized, as I already said,
00:04:33.20 by a very high binding affinity and specificity. And
00:04:38.14 they have also, or some of them undergo structural
00:04:45.18 changes upon ligand binding. And this we wanted to use
00:04:49.00 to build up our synthetic riboswitches. The principle idea
00:04:56.22 is shown here. In the upper part, you see that we have
00:05:01.07 integrated an aptamer into the 5' UTR of a eukaryotic
00:05:06.19 messenger RNA. And without a ligand, the aptamer
00:05:10.24 is only slightly structured and then if you add the ligand,
00:05:18.10 here in this case, the aptamer adapts the defined tertiary
00:05:26.22 structure. And in this conformation, it's able to interfere
00:05:30.29 with cellular processes. Here in this case, it can
00:05:35.16 stop the scanning ribosome and this can't recognize the
00:05:41.25 START codon, and so it leads to the stop of translation.
00:05:46.12 We started our work with an aptamer which can bind
00:05:51.12 tetracycline, as you can see on the left side, this L-shaped
00:05:54.28 RNA structure. It forms a very complex structure here and
00:06:03.16 in green, you see tetracycline, which is deeply buried within
00:06:07.16 the RNA molecules. And nearly all chemical moiety of
00:06:11.29 tetracycline are read out by the RNA, so it's a very
00:06:16.10 tight complex. And this aptamer, we inserted in front
00:06:21.20 of a GFP reporter gene in yeast. And here you can see that
00:06:29.10 with increase in tetracycline in the graph, we obtain a
00:06:34.22 decrease in the GFP fluorescence. The control is slightly
00:06:41.14 tight. If you insert more than one but less than three copies
00:06:46.08 of this aptamer, then you obtain really a tight regulation
00:06:52.07 of gene expression. You can even use this to control
00:06:57.29 essential genes. Here's an example we've shown for
00:07:01.14 NOP14, it's a protein involved in ribosome genesis.
00:07:06.00 And if you bring this under control of our tetracycline
00:07:09.04 aptamer, and if you add tetracycline, you can completely
00:07:13.13 inhibit the growth of tetracycline as shown here.
00:07:17.29 It's a very rapid decrease, so you can't detect the protein
00:07:21.24 anymore after four hours. Beside the control of
00:07:29.05 translation initiation, we also wanted to control splicing.
00:07:33.18 And again, we started in yeast and the idea is
00:07:38.22 schematically shown here. Now we have an intron
00:07:42.10 which is inserted in the GFP reading frame of
00:07:50.18 yeast. And then, within this intron, we inserted our
00:07:56.00 small aptamer close to the 5' splice site, which is shown
00:08:00.16 there as a red box. And the idea is that the 5' splice
00:08:07.25 is accessible without tetracycline, so the spliceosome can
00:08:13.00 bind and then if you add tetracycline, you have a tightening of
00:08:19.08 the complex and the 5' splice site is not accessible anymore.
00:08:26.06 Here you can show some of the constructs, we have
00:08:31.20 inserted the aptamer two nucleotides away from the 5'
00:08:39.16 splice site. This is the left construct, directly behind
00:08:43.16 the splice site. And we also have integrated the 5'
00:08:47.16 splice site into the stem of the aptamer. And what you
00:08:58.16 can see is that as soon as you are very close on the splice
00:09:03.13 site, then you can nicely inhibit gene expression, as you
00:09:08.13 can see on the gray graph. And the best construct we have
00:09:13.22 is if the splice site is completely integrated within the closing
00:09:19.08 stem. And we think that there is a certain breathing of the
00:09:22.23 closing stem and after binding of tetracycline, you
00:09:26.23 inhibit this breathing and then the splice site is not
00:09:32.20 accessible any longer. And here you can see that
00:09:37.16 you indeed on the level of RNA can distinguish
00:09:43.01 between the construct without and plus tetracycline.
00:09:47.22 Our next cellular target besides translation initiation
00:09:56.17 and splicing was mRNA degradation. We wanted to
00:10:03.18 control, here we made use of small RNA molecules,
00:10:07.24 so-called ribozymes. And these ribozymes can cleave
00:10:12.13 itself. So if you integrate such a ribozyme in an mRNA
00:10:17.18 and it cleaves itself, then the RNA is rapidly degraded,
00:10:22.13 as you can see on the left side of this picture.
00:10:28.14 But we wanted to control the cleavage activity by
00:10:32.10 an aptamer shown on the right side, so that we can
00:10:36.18 regulate if the mRNA is degraded or not.
00:10:40.26 Here I want to show you our construct. This is
00:10:46.19 a secondary structure and a tertiary structure on the left
00:10:49.25 of the hammerhead ribozyme. Only if the ribozyme folds
00:10:55.14 into the structure you can see on the left side, then
00:10:59.17 it cleaves itself exactly on the position marked by the red arrow.
00:11:06.00 And then we attached our tetracycline aptamer on the
00:11:12.26 right side. And very important for the activity of
00:11:19.23 the ribozyme, is the loop-loop interaction, which is shown
00:11:25.06 here. Between this and the loop on the other side. And
00:11:30.10 we thought maybe you can disturb this loop-loop interaction
00:11:33.24 by our tetracycline aptamer in a way that the nucleotides,
00:11:41.20 which are encircled, are free and are there for loop-loop
00:11:48.01 interaction. If the closing stem of the aptamer here
00:11:55.08 is somehow breathing and not rigid. And then if we add
00:12:02.12 tetracycline, as shown in the last experiment, there is
00:12:07.04 a tightening of this closing stem and these 3 nucleotides
00:12:11.16 are no longer accessible for this loop-loop interaction
00:12:15.11 hence the ribozyme can't cleave any longer. And so we
00:12:22.04 looked for some closing stems which exactly behaved
00:12:27.29 in this way. Here you can see some constructs which
00:12:31.13 we tested. And indeed there, we found three stem
00:12:37.08 compositions which showed exactly the behavior we
00:12:43.02 wanted to see. Here, without tetracycline, you have
00:12:50.25 this breathing in the stem and this allows the ribozyme
00:12:56.28 to form. And you have the mRNA degradation that means
00:13:02.07 that without tetracycline, as shown here. Without
00:13:08.18 tetracycline you have no gene expression and then if
00:13:14.13 you add tetracycline, you prevent or you prevent the breathing
00:13:19.20 of the closing stem and the loop-loop interaction is
00:13:24.23 disturbed. And this leads to stabilization of the mRNA
00:13:28.25 and to an increase in gene expression.
00:13:32.25 We tested this not only in yeast, but also in
00:13:36.25 eukaryotic cells, in HeLa cells. Without tetracycline,
00:13:41.17 no fluorescence. If you add tetracycline, then
00:13:45.24 you have an increase in reporter activity.
00:13:50.25 To sum of this first part, I hope I could convince you that
00:13:56.07 you indeed can use such aptamers to control gene
00:14:00.25 expression. You can use a tetracycline aptamer, there are
00:14:05.18 some others known in the literature which bind theophylline,
00:14:10.18 neomycin. You can work in different organisms, bacteria,
00:14:14.24 yeast, archaea, human cells. And you can target
00:14:19.18 different cellular targets, so for instance, the translation
00:14:25.16 initiation, you can control splicing, or mRNA stability.
00:14:29.18 However, if you have a look into the literature,
00:14:34.01 you realize that these three aptamers, tetracycline,
00:14:38.18 neomycin, and the other aptamer, these are the only ones
00:14:43.26 which are used to build up riboswitches. And that's why
00:14:50.12 we were interested in the question, what turns an aptamer
00:14:53.22 into a riboswitch? And therefore, we tried to find
00:15:00.02 further of these guys, further of these regulating aptamers.
00:15:03.29 This time, the selection was done with neomycin
00:15:08.12 as a target. After several rounds of selection, we ended
00:15:13.11 up with an enriched pool of aptamers. So this means
00:15:17.14 within this pool, there are several aptamers which can
00:15:20.26 bind theoretically, to neomycin. But then we did not
00:15:26.13 continue with this pool and sequence it, and looked
00:15:29.25 for a single colony. But we used the complete pool
00:15:34.04 and we cloned the pool in front of GFP reporter genes.
00:15:37.29 And then we screened on 96-well plates for aptamers
00:15:43.22 which have exactly the phenotype we looked for. So here
00:15:49.18 we have some, we have found one aptamer that
00:15:55.23 if we have it inserted in front of a GFP reporter gene,
00:16:00.20 it led to a decrease in fluorescence if we add neomycin.
00:16:06.10 The sequence is shown here, however, if you have a look
00:16:12.12 into the enriched aptamer pool here, then you -- sorry,
00:16:20.06 this aptamer behaviors similarly to the tetracycline
00:16:23.02 aptamer and it leads to a nice dose dependent decrease
00:16:26.18 in gene expression if you add neomycin.
00:16:31.11 However, if you sequence this in vitro selected aptamer
00:16:35.20 pool, and this is the red guy here. Then you end up with
00:16:39.23 a completely different aptamer. You have the same stem loop
00:16:44.07 structure, but the sequence is completely different.
00:16:48.02 That's why we were interested in why the one guy
00:16:53.23 is active, the black one, and the red one was not able to
00:16:59.26 for building riboswitches. So we characterized this
00:17:05.23 active aptamer and we could show that the other part
00:17:11.12 is important for the binding of the ligand. However,
00:17:15.11 we could identify here, a small element, we call it
00:17:19.28 asymmetric bulge. So some nucleotides on the left side
00:17:24.02 of the aptamer, which we identified as important
00:17:30.08 for regulation, so this bulge can be between 1-5
00:17:36.23 nucleotides long. It's mostly sequence independent with
00:17:41.29 a small preference for pyrimidines. However, it's important
00:17:45.28 that it is present and it's important that you do not have
00:17:49.28 a nucleotide on the other side of this bulge.
00:17:53.17 At this point, we decided to team up with an NMR
00:17:59.14 group in Frankfurt, the group of Jens Wohnert.
00:18:02.24 And he analyzed the structure of this aptamer
00:18:06.12 for us. And he could see, he could find out that
00:18:11.01 the aptamer is largely unstructured in the absence of
00:18:15.29 neomycin. You only can identify the seven good base pairs,
00:18:22.11 which are shown in the secondary structure. However,
00:18:25.06 if you add in neomycin, you can see in the lower spectrum
00:18:31.00 a lot of new signals. These red dots, and they correspond
00:18:35.04 to several new base pairs. Here we have two U-U base
00:18:42.11 pairs. The lower ones are hydrogen mediated, so it nicely fits
00:18:49.13 into a form helix, and this allows the G-C, G-C base pair
00:18:54.04 to form. And then the lower part of this small aptamer
00:19:03.13 steps on each other, so you have a long rigid
00:19:08.13 helix. And this helix then, is able to interfere with cellular
00:19:14.08 processes. So, if you compare our active aptamer here,
00:19:21.18 then you have this bulge. And this bulge seems to
00:19:27.00 keep open your ground state and this is not the case
00:19:34.02 for the inactive aptamers, as you can see on the left side,
00:19:37.29 these similar aptamers however, in this case, they are
00:19:42.24 non-active in vivo and they also do not show any
00:19:47.17 conformational changes. So, we think, or after our further
00:19:53.29 analysis of this mechanism, we could show that we have
00:19:57.07 in the ligand free state, a kind of dynamic ensemble
00:20:02.10 of a kind of open and closed state in equilibrium.
00:20:07.20 And then the ligand comes and binds the pre-formed
00:20:12.01 closed state and stabilizes like a clamp, which holds
00:20:16.16 the upper and the lower part of the helix together.
00:20:20.12 And this is not the case for the inactive aptamer, because
00:20:25.24 there we already have the major population of this closed
00:20:31.02 conformation, and now the stabilized population.
00:20:35.01 I want to summarize my talk, I hope I could
00:20:40.13 convince you that you can indeed use small molecular
00:20:44.02 binding aptamers to control gene expression. To engineer
00:20:49.22 riboswitches, either by using the aptamers alone or
00:20:53.25 by building up more modular riboswitches. You can
00:20:58.13 control different cellular processes, like translation
00:21:03.18 initiation, splicing, or mRNA stability. And you can build
00:21:09.06 them up for different organisms. And in the second part,
00:21:12.17 I gave you some mechanistic insights into engineered
00:21:17.02 riboswitches. I showed you some experiments
00:21:23.28 for a neomycin aptamer, where large conformational changes
00:21:28.25 are important for the regulatory activity of this aptamer.
00:21:38.02 And very important, the active aptamer was completely
00:21:42.19 underrepresented in the pool after in vitro selection and
00:21:46.10 it was important to add an in vivo screening step
00:21:50.00 to identify this regulating aptamer. Thank you very much.

This Talk

Speaker: Beatrix Süess

Audience:

Researcher

Recorded: June 2015

Talk Overview

Dr. Beatrix Suess explains that riboswitches are highly structured RNA elements that are located in the 5’ untranslated regions (UTR) of many bacterial mRNAs. Riboswitches bind ligands with high specificity. This causes a termination of transcription or the inhibition of translation initiation. Suess reveals how to engineer riboswitches into the 5’ region of an mRNA in order to control its gene expression by a specific ligand, such as tetracyline or neomycin. She discusses the ideal traits that make a good riboswitch.

Speaker Bio

Beatrix Süess

Beatrix Suess is a Professor at the Technical University, Darmstadt. She received her PhD from the University of Erlangen, Germany where she was also a post-doctoral fellow. Suess was also a research fellow at the University of Vienna, Austria and at Yale University, USA. Dr. Suess’s research focuses on the ability of RNA to operate… Continue Reading