*This webpage was produced as an assignment for an undergraduate course at Davidson college. *


Programmable ligand-controlled riboregulators of eukaryotic gene expression

Travis S. Bayer & Christina D Smolke

Nature Biotechnology v.23 n3 March 2005 337-343.

A review by Matt Gemberling:


In this paper, Bayer and Smolke are pursuing the development of trans -acting RNA's to regulate gene expression. They do this by designing riboswitches, which can be turned on and off through allosteric modulation by various ligands. The design of these anti-switched begins with the selection of an aptamer. The aptamer stem of the anti-switch contains a specific ligand binding sequence and allows for allosteric modulation of the construct. The second portion of the anti-switch is the anti-sense stem which contains anti-sense RNA to bind to a specific mRNA, in turn shutting down production of this gene's protein. The authors of this paper use two different Aptamer portions of the anti-switch, one that binds to theophylline and one that binds to tetracycline. These two ligands are easily introduced in vivo and in this case the genes they are specific to GFP and YFP respectively. GFP is an ideal protein because it is easy to calculate the relative gene expression, without performing another operation on the cells. One can use a machine to calculate the relative fluorescence given off by the cells. The authors have used this knowledge to create several different anti-switch constructs, to better understand the design of these anti-switches. They have designed switches that are both "on" and "off" switches, meaning that without ligand present they have designed one switch that turns the production off under these conditions as well as one that does not alter the expression of the gene. They have shown that the Anti-switches are ligand dependent as the concentration of the ligand must reach a critical level before total inhibition will occur. Expression was not prone to gradual declines; rather it showed that there was a threshold at which the amount of ligand becomes sufficient to inhibit the majority of expression. The authors have also demonstrated the ability to use multiple anti-switches in the same cell at the same time. As long as the ligand and the target gene are different, the anti-switch displays a very selective nature. This is very important as anti-switches could one day be used to regulate entire protein cascades.


Description of figures followed by conclusions:

Figure 1a:

Shows the general mechanism by which the s1, anti-switch, will work both by itself and in the presence of theophylline, its ligand. It shows how the switch will be in relation to the GFP mRNA. Without theophylline the anti-switch will not bind to the GFP mRNA allowing GFP production to continue. Once theophylline is added the anti-switch becomes allosterically modulated and the anti-sense arm will bind to GFP mRNA and production will cease.

Figure 1b:

The structure of the anti-switch is shown in the presence of theophylline. The more stable and more energetically favorable form of the anti-switch in the presence of theophylline is when it is bound to the GFP mRNA. This allows them to predict that the majority of the anti-switch in the presence of theophylline will be bound to GFP mRNA stopping its production.

Figure 1c:

This figure is a Graph of the Relative GFP expression versus the concentration of theophylline present. There are four experimental conditions: The aptamer construct + theophylline is a positive control and the GFP levels remain around 1. The s1 + caffeine conditions are also a negative control which should show no decrease in GFP production with and increased concentration of theophylline. The anti-sense construct + theophylline is a positive control and inhibits GFP expression. S1+ theophylline is the experimental condition we are supposed to notice. The levels of GFP decrease slowly until the point of ~0.75 mM theophylline in which inhibition of GFP increase drastically to the levels of the + control.


Shows the affect on GFP in vivo of the addition theophylline to cells that either contain do not contain the anti-switch.   After the addition of theophylline the GFP in the cells containing s1, decreases at a rate similar to the half-life of GFP protein. This suggests that no new protein is being made in these cells and that the previously made protein is degraded as normal.

Figure 1e:

This figure is a gel showing the affinity of the anti-switch to bind to theophylline in different concentrations. The sharp increase between 2 and 10 is due to both the ligand and target binding. This experiment was done to try and show that the riboswitch was binding to only one mRNA. The riboswtiches were radioactively labeled.

Figure 2a:

In this figure four riboswitch constructs are shown: s1, s2, s3, and s4. S1 is the original construct of the anti-switch. S2, has a one base mismatch in the antisense stem, this makes the antisense arm less stable. S3 has a longer antisense arm, which makes the compound more stable. It also has 3 new base pairs in the aptamer stem making the construct even more stable. S4 contains an altered loop sequence on the antisense stem, causing this conformation to be much less stable.

Figure 2b:

Shows the in vivo expression of GFP in the presence of different concentration of theophylline with the different anti-switch constructs. It shows that S4 as expected being the most unstable would inhibit GFP expression the most. S2 showed the second greatest ability to inhibit GFP expression, which shows that it is less stable than s1, but more stable than s4. s2 and s4 greatly inhibit GFP expression at about 0.1 mM theophylline. S3 displays less inhibition than s1 at low concentrations, but inhibits in a similar fashion at about 1 mM theophylline. All constructs however do display inhibition of GFP.

Figure 2c:

This graph shows the in vivo concentration of theophylline required to effectively reduce the GFP expression to approximately 0. The constructs used in this Graph are s1, the original anti-switch, and s5, and s6, two new anti-switch constructs. S5 has a tenfold lower affinity for theophylline; with this lower affinity it takes a higher concentration to result in the same level of GFP inhibition. S6 a destabilized form of s5 displays greater inhibition of GFP between 5 and 10 mM than s6. It is important to notice that the concentration required for total inhibition is tenfold high in both s5 and s6 mirroring the tenfold decrease in affinity for theophylline.

Figure 2d:

In this graph s1 is the original anti-swtich while s7, is a modified version that responds to tetracycline rather than theophylline. This figure shows that as the in vivo concentrations of these ligands increases the same amount of inhibition in the expression of GFP occurs. This shows that the different ligands do not have different levels of inhibition suggesting that this inhibition is concentration not type of ligand dependent.

Figure 3a:

Show the mechanism of s8 a newly designed anti-switch that is an "on" switch in the presence of theophylline but off without its presence.

Figure 3b:

This graph shows the in vivo levels of GFP production in s1 and s8 cells before and after the addition of theophylline. As expected the s8 cells have very low GFP until the addition of theophylline and gain GFP rapidly after this addition. S1 cells act as normal with high levels and the gradual decline is GFP once theophylline is added.

Figure 4a:

Shows two constructs, one a GFP anti-switch, whose ligand is theophylline, and another YFP anti-switch whose ligand is tetracycline. It shows the mechanism by which these bind and affect expression.

Figure 4b:

This graph shows the in vivo production of GFP and YFP under four different treatment conditions. With no ligand present the expression levels of both GFP and YFP are high. With only theophylline present only GFP expression was affected and with only tetracycline present only YFP expression was affected.   However with both ligands present both YFP and GFP expression was strongly inhibited. This shows the specificity of the anti-switches as well as the possibility of using multiple anti-switches at the same time or in a sequence.


The conclusions reached in this paper are simple and straight forward.

1. Riboswitches can be made to control gene expression through a ligand controlled method.

2. These riboswitches are capable of significant gene inhibition.

3. Differences in anti-switch structure and stability relates to the relative gene expression levels.

4. On can contruct both and "on" and "off" riboswitch.

5. Two riboswitches can be used in conjuction to target two different gene transcripts, and can be regulated seperately by two different ligands.


I would first like to say that I really enjoyed this paper. The authors present the material in a very logical and concise manner which makes reading it easy to follow and understand. The figures were clear and provided necessary information for the understanding of this paper. I believe the conclusion the authors arrived at are substantiated by their data. The authors conclude that they are able to engineer, ligand-controlled anti-switches, which based on the data I believe to be true. The authors demonstrate that they can create several different constructs with different stabilities and that these small differences will have an impact on the ability and sensitivity of the anti-switch to regulate gene expression. The most compelling figure for this is construct s5 in which they use an aptamer stem with a tenfold lower affinity for theophylline and they find that it requires a tenfold higher concentration of theophylline to display total inhibition. I believe that this figure goes a long way to proving that they can accurately design and control the anti-switch, essentially controlling gene expression of the target gene. The experiment with two anti-switches that respond to two different ligand and control two different genes also goes a long way to establishing their ability to control one specific gene, with one ligand-controlled anti-switch.

There is one more test I would have liked to have seen them perform. Figure 1d shows the GFP expression in cells with and without the anti-switch, both before and after the addition of theophylline. I would have liked to see an analysis done one either s5 or s6 and on s7 to show the relative amounts of GFP over time. This would eliminate the possibility that the different plasmids used to insert the anti-switch as well as the GFP gene acted differently over time. Especially in anti-switch, s7, it would have been nice to see the difference between cells with or without the anti-switch.

Another question I have is whether or not GFP or YFP is the only gene that is being turned "off" by the anti-switch. They show that the switch is working to regulate the expression of GFP, but is it also affecting other mRNA's as well? Figure 1e suggests that it is not, but I don't know that i'm convinced. I would have liked to see them do another test to confirm or dismiss the possibility that they are binding to other mRNA as well. Another question I have is does it matter if it is binding to other mRNA's if the cells are still functional and alive?

Other than these two suggestions I feel as though the authors have covered their bases, even going as far as to create a switch they could turn from "off" to "on". They have used the appropriate controls in all experiments and have shown the ability to create functional riboswitches that from the data appear to work as expected.

Future Studies :

One area for future research that I believe should be looked into is possible drug discovery, based on the riboswitch. Many known diseases are caused by the malfunction of one or more genes. Using the information in this paper one could envision using these riboswitches to control the gene expression of disease causing genes. Use of this method in vivo does not seem to exhibit any notable problems as all GFP expression levels were taken in vivo . Riboswitches could be designed specifically for each gene of interest and these genes could then be turned "off", when the patient has taken a drug such as tetracycline. Another very important fact is that the use of a ligand like tetracycline is beneficial because tetracycline is already known to be safe for human use. This technique would be most effective if used on genes that do not need to be active for major cell functions.

There are some potential draw backs for this project though. It would be unlikely that one could insert a plasmid containing the coding for the riboswitch into every human cell. However if a specific tissue was the target one could envision the possibility of targeting this organ and getting plasmids into enough cells to make a substantial impact. Another potential drawback is that if this gene of interest is necessary for normal cell function it would not be enough to silence the gene; one would then have to include a new copy of the gene on the plasmid to allow normal protein synthesis.

Another possible approach is that this technique could be used in organisms to determine gene function. The genome of many organisms has been sequenced and many genes are known to exist however there is very little functional data on the majority of these genes. With this technique one could pick a gene of interest and create a riboswitch to turn "off" this gene. With this gene turned off, it could be possible to gain general knowledge as to the function of the gene. If you turn the gene "off" and observe what is happening in the cell, one may be able to ascertain the function of this gene. In organisms which have vast gene expression cascades this could also be used to determine genes required to continue the cascade.

These are just two examples of how a riboswitch could be used in further projects, but this is not even scratching the surface of the possible uses of riboswitches.



Bayer TS, Smolke CD. 2005 March. Programmable ligand-controlled riboregulators of eukaryotic gene expression. Nature Biotech 23(3): 337-343.



Back to Matt's Molecular Page

Molecular Biology Home Page

Back to Students Molecular Pages

© 2005 Department of Biology, Davidson College, Davidson, NC 28036

Any questions or concerns regarding this page should be directed to Matt Gemberling.