Team:Edinburgh/project/population
From 2014.igem.org
Samireland (Talk | contribs) |
|||
(14 intermediate revisions not shown) | |||
Line 4: | Line 4: | ||
<div id="topbit" style="margin-bottom: 40px;"> | <div id="topbit" style="margin-bottom: 40px;"> | ||
<table align="center" style="border-spacing: 5px;"><tr><td><a href="https://2014.igem.org/"><img src="https://static.igem.org/mediawiki/2014/archive/3/3f/20140702191450%21Igem.png" width="50px" height="45px"></a></td><td id="navlink"><a href="https://2014.igem.org/Team:Edinburgh">Home</a></td><td id="navlink"><a href="https://2014.igem.org/Team:Edinburgh/team/">Team</a></td><td id="navlink"><a href="https://igem.org/Team.cgi?id=1399">Profile</a></td><td id="navlink"><a href="https://2014.igem.org/Team:Edinburgh/logic/">Background</a></td><td id="navlink"><a href="https://2014.igem.org/Team:Edinburgh/project/">Project</a></td><td id="navlink"><a href="https://2014.igem.org/Team:Edinburgh/HP/">Policy and Practices</a></td><td id="navlink"><a href="https://2014.igem.org/Team:Edinburgh/modelling/">Modelling</td><td id="navlink"><a href="https://2014.igem.org/Team:Edinburgh/log">Daily log</a></td></table></div> | <table align="center" style="border-spacing: 5px;"><tr><td><a href="https://2014.igem.org/"><img src="https://static.igem.org/mediawiki/2014/archive/3/3f/20140702191450%21Igem.png" width="50px" height="45px"></a></td><td id="navlink"><a href="https://2014.igem.org/Team:Edinburgh">Home</a></td><td id="navlink"><a href="https://2014.igem.org/Team:Edinburgh/team/">Team</a></td><td id="navlink"><a href="https://igem.org/Team.cgi?id=1399">Profile</a></td><td id="navlink"><a href="https://2014.igem.org/Team:Edinburgh/logic/">Background</a></td><td id="navlink"><a href="https://2014.igem.org/Team:Edinburgh/project/">Project</a></td><td id="navlink"><a href="https://2014.igem.org/Team:Edinburgh/HP/">Policy and Practices</a></td><td id="navlink"><a href="https://2014.igem.org/Team:Edinburgh/modelling/">Modelling</td><td id="navlink"><a href="https://2014.igem.org/Team:Edinburgh/log">Daily log</a></td></table></div> | ||
- | <div id="tourleft"><a href="https://2014.igem.org/Team:Edinburgh/project/ | + | <div id="tourleft"><a href="https://2014.igem.org/Team:Edinburgh/project/cisgenic">← Cisgenic wires</a></div> |
- | <div id="tourright"><a href="https://2014.igem.org/Team:Edinburgh/ | + | <div id="tourright"><a href="https://2014.igem.org/Team:Edinburgh/zeigler">The Zeigler report→</a></div> |
<div id="igem"> | <div id="igem"> | ||
<div id="header"> | <div id="header"> | ||
Line 14: | Line 14: | ||
<div id="PC"> | <div id="PC"> | ||
<h2>The circuit</h2> | <h2>The circuit</h2> | ||
- | <p>We designed an autoregulatory system for establishing an equilibrium between the populations of three different strains of E. coli, each with a different battery of biodegradation enzymes.</p> | + | <p>We designed an autoregulatory system for establishing an equilibrium between the populations of three different strains of <em>E. coli</em>, each with a different battery of biodegradation enzymes.</p> |
<img src="https://static.igem.org/mediawiki/2014/1/11/Ed14_Pc1.png"> | <img src="https://static.igem.org/mediawiki/2014/1/11/Ed14_Pc1.png"> | ||
Line 31: | Line 31: | ||
<p>Splitting the individual metabolic pathways between the different strains would effectively provide them with a means of communication. This has previously been described as Metabolic Wiring (Silva-Rocha and de Lorenzo, 2013). Each strain would thus ‘sense’ the presence of the other two strains in the media by detecting the intermediary metabolites that each of them release. Let’s call these intermediaries ‘signalling metabolites’.</p> | <p>Splitting the individual metabolic pathways between the different strains would effectively provide them with a means of communication. This has previously been described as Metabolic Wiring (Silva-Rocha and de Lorenzo, 2013). Each strain would thus ‘sense’ the presence of the other two strains in the media by detecting the intermediary metabolites that each of them release. Let’s call these intermediaries ‘signalling metabolites’.</p> | ||
- | + | <h2>The output</h2> | |
<p>Using a synthetic AND gate (Wang, Barahona and Buck, 2013), each strain would integrate the signal emitted by the other two strains. The output of that AND gate was coupled with the suppression of one of a series of growth modulating genes that we cloned. This way, if any one strain is not receiving a signal from one or both other strains (thus not fulfilling the criteria of the AND gate), growth regulating genes would be activated causing it to pause its growth, effectively “waiting” for the other strain(s) to catch up.</p> | <p>Using a synthetic AND gate (Wang, Barahona and Buck, 2013), each strain would integrate the signal emitted by the other two strains. The output of that AND gate was coupled with the suppression of one of a series of growth modulating genes that we cloned. This way, if any one strain is not receiving a signal from one or both other strains (thus not fulfilling the criteria of the AND gate), growth regulating genes would be activated causing it to pause its growth, effectively “waiting” for the other strain(s) to catch up.</p> | ||
Line 38: | Line 38: | ||
- | <p><em>Figure 1. Cartoon illustration of our circuit’s behaviour, each ‘cell’ symbolises a strain’s population. In 1, all strains are growing at the same rate. In 2, population A is lagging in comparison to populations B and C. As a result, populations C and B are consuming the signalling metabolite produced by A at a faster rate than it is being produced (simply because there are more of them). This leads to a decrease in the concentration of the signalling metabolite in the media, triggering the growth regulating genes in populations B and C to become activated (3),inducing them to arrest growth. As a result, in 4, population A can catch up. In 5, population A has successfully caught up with populations B and C, and is now producing enough signalling metabolite to for populations C and B to continue growing.</em></p> | + | <p><em>Figure 1. Cartoon illustration of our circuit’s behaviour, each ‘cell’ symbolises a strain’s population. In 1, all strains are growing at the same rate. In 2, population A is lagging in comparison to populations B and C. As a result, populations C and B are consuming the signalling metabolite produced by A at a faster rate than it is being produced (simply because there are more of them). This leads to a decrease in the concentration of the signalling metabolite in the media, triggering the growth regulating genes in populations B and C to become activated (3), inducing them to arrest growth. As a result, in 4, population A can catch up. In 5, population A has successfully caught up with populations B and C, and is now producing enough signalling metabolite to for populations C and B to continue growing.</em></p> |
Line 50: | Line 50: | ||
<h2>Does it work?</h2> | <h2>Does it work?</h2> | ||
<p>The system was modelled mathematically and proven to produce the desired behaviour (fig.2). We used the mathematical model inform and fine tune the components of our circuit. We understood that in order for the signalling metabolite produced by each strain to represent the relative size of each strain’s population, it had to be degraded at the same rate that it was being produced, in order to avoid its saturation or depletion in the media.</p> | <p>The system was modelled mathematically and proven to produce the desired behaviour (fig.2). We used the mathematical model inform and fine tune the components of our circuit. We understood that in order for the signalling metabolite produced by each strain to represent the relative size of each strain’s population, it had to be degraded at the same rate that it was being produced, in order to avoid its saturation or depletion in the media.</p> | ||
- | <p> Since for every one signalling metabolite there are two strains consuming it (receivers) and only one producing it (sender), the consumption rate by each of the two receivers had to be exactly half of the production rate by the sender, so that when summed together the net production rate is zero. Only then will the concentration of the each signalling metabolite correlate with the size of the population producing it. As part of our project we characterised the activity of multiple degron tags, which would allow us to fine tune the degradation rates of the enzymes and transcription factors relating to the consumption and production of each signalling metabolite, in order to achieve the correct rate ratios (consumption:production 1:2).</p> | + | <p> Since for every one signalling metabolite there are two strains consuming it (receivers) and only one producing it (sender), the consumption rate by each of the two receivers had to be exactly half of the production rate by the sender, so that when summed together the net production rate is zero. Only then will the concentration of the each signalling metabolite correlate with the size of the population producing it. As part of our project we characterised the activity of multiple <a href="https://2014.igem.org/Team:Edinburgh/project/degrons">degron tags</a>, which would allow us to fine tune the degradation rates of the enzymes and transcription factors relating to the consumption and production of each signalling metabolite, in order to achieve the correct rate ratios (consumption:production 1:2).</p> |
- | <p>The resulting model indicated that the | + | <p>The resulting model indicated that the threshold for each population size was dynamic and relative to that the other strains. The concentration of the signalling metabolites in the medium is defined by the ratio of the its consumption and production rates, which in turn are defined by the ratios of the corresponding populations (the more cells, the more degrading/producing enzymes present globally). This meant that each population would adjust its size according to the ratio of its own size and the size of the other two. </p> |
<img src="https://static.igem.org/mediawiki/2014/4/42/Ed14_Simulation.png"> | <img src="https://static.igem.org/mediawiki/2014/4/42/Ed14_Simulation.png"> | ||
- | <p><em>Figure 2: We modelled how the relative populations size of each strain affects the growth rate of the other two. Here, you can see how population 2 (blue) and 3 (green) will enter a transient stationary phase at 50 minutes. This is caused by the fact that population 1 (red), which is has been lagging up to that point, is being outgrown, which results in there not being a high enough concentration of its signalling molecule in the media to allow strain 2 and 3 to grow. For a description of the differential equations used to generate these graphs, please refer to the modelling page.</em></p> | + | <p><em>Figure 2: We modelled how the relative populations size of each strain affects the growth rate of the other two. Here, you can see how population 2 (blue) and 3 (green) will enter a transient stationary phase at 50 minutes. This is caused by the fact that population 1 (red), which is has been lagging up to that point, is being outgrown, which results in there not being a high enough concentration of its signalling molecule in the media to allow strain 2 and 3 to grow. For a description of the differential equations used to generate these graphs, please refer to the <a href="https://2014.igem.org/Team:Edinburgh/modelling/maths">modelling page</a>.</em></p> |
- | <h2>Selection of | + | <h2>Selection of growth regulating genes</h2> |
<p>We performed an extensive literature research in order to find an array of genes that could be used for modulating the growth rate of a bacterial population at real time. The genes had to fulfil the following criteria:</p> | <p>We performed an extensive literature research in order to find an array of genes that could be used for modulating the growth rate of a bacterial population at real time. The genes had to fulfil the following criteria:</p> | ||
Line 67: | Line 67: | ||
<p>The genes we found were RelA, RelA’ (a truncated version of RelA), rsd and rmf, which are described in more detail below. We found that overexpression of these genes was likely to induce the desired behaviour in our cell populations. We successfully cloned these genes in high copy plasmids and had them sequenced. We found that these genes were toxic at such high copy number, therefore we swapped the vector for a lower copy one. We advise future teams to use low copy plasmids for genes detrimental to the cells’ growth, and co-transform with a high copy plasmid harbouring the repressor, in order to achieve minimum leaking.</p> | <p>The genes we found were RelA, RelA’ (a truncated version of RelA), rsd and rmf, which are described in more detail below. We found that overexpression of these genes was likely to induce the desired behaviour in our cell populations. We successfully cloned these genes in high copy plasmids and had them sequenced. We found that these genes were toxic at such high copy number, therefore we swapped the vector for a lower copy one. We advise future teams to use low copy plasmids for genes detrimental to the cells’ growth, and co-transform with a high copy plasmid harbouring the repressor, in order to achieve minimum leaking.</p> | ||
- | <p>We found that another way to approach the same challenge, would be to silence a growth-related target gene, using antisense RNA (asRNA). We picked a gene that encoded a subunit of RNA polymerase that is necessary for exponential growth, sigma factor 70 (rpoD). The protocol we decided to follow (Nakashima et al., 2012), used a system of pair termini for stabilisation. We designed an algorithm that detected the optimal segment of the RNA for silencing, which is described here.</p> | + | <p>We found that another way to approach the same challenge, would be to silence a growth-related target gene, using antisense RNA (asRNA). We picked a gene that encoded a subunit of RNA polymerase that is necessary for exponential growth, sigma factor 70 (rpoD). The protocol we decided to follow (Nakashima et al., 2012), used a system of pair termini for stabilisation. We designed an algorithm that detected the optimal segment of the RNA for silencing, which is described <a href="https://2014.igem.org/Team:Edinburgh/modelling/software">here</a>.</p> |
+ | |||
+ | |||
<h2>Progress</h2> | <h2>Progress</h2> | ||
Line 77: | Line 79: | ||
2] we had to find vectors with lower copy number, in order to avoid overwhelming toxicity to the cell.</p> | 2] we had to find vectors with lower copy number, in order to avoid overwhelming toxicity to the cell.</p> | ||
<p>Just to be sure, we decided to combine the two, and express the growth controlling gene from a low copy plasmid, while at the same time express the repressor from a high copy plasmid.</p> | <p>Just to be sure, we decided to combine the two, and express the growth controlling gene from a low copy plasmid, while at the same time express the repressor from a high copy plasmid.</p> | ||
+ | |||
+ | <h2>Talks with industry</h2> | ||
+ | |||
+ | <p> Finally, we spoke with people working in the field to gain insights on how our project would be applied in the real world. <a href="https://2014.igem.org/Team:Edinburgh/zeigler">Here is a description of the interview</a>.</p> | ||
Latest revision as of 03:59, 18 October 2014
The circuit
We designed an autoregulatory system for establishing an equilibrium between the populations of three different strains of E. coli, each with a different battery of biodegradation enzymes.
![](https://static.igem.org/mediawiki/2014/1/11/Ed14_Pc1.png)
The wiring
Splitting the individual metabolic pathways between the different strains would effectively provide them with a means of communication. This has previously been described as Metabolic Wiring (Silva-Rocha and de Lorenzo, 2013). Each strain would thus ‘sense’ the presence of the other two strains in the media by detecting the intermediary metabolites that each of them release. Let’s call these intermediaries ‘signalling metabolites’.
The output
Using a synthetic AND gate (Wang, Barahona and Buck, 2013), each strain would integrate the signal emitted by the other two strains. The output of that AND gate was coupled with the suppression of one of a series of growth modulating genes that we cloned. This way, if any one strain is not receiving a signal from one or both other strains (thus not fulfilling the criteria of the AND gate), growth regulating genes would be activated causing it to pause its growth, effectively “waiting” for the other strain(s) to catch up.
![](https://static.igem.org/mediawiki/2014/2/20/Ed14_Pc2.png)
Figure 1. Cartoon illustration of our circuit’s behaviour, each ‘cell’ symbolises a strain’s population. In 1, all strains are growing at the same rate. In 2, population A is lagging in comparison to populations B and C. As a result, populations C and B are consuming the signalling metabolite produced by A at a faster rate than it is being produced (simply because there are more of them). This leads to a decrease in the concentration of the signalling metabolite in the media, triggering the growth regulating genes in populations B and C to become activated (3), inducing them to arrest growth. As a result, in 4, population A can catch up. In 5, population A has successfully caught up with populations B and C, and is now producing enough signalling metabolite to for populations C and B to continue growing.
Why Metabolic Wiring?
Our system relies on Metabolic Wiring in order to work, therefore we investigated three possible metabolic pathways that can be used for this purpose. These are described on our Metabolic Wires page. The reasons why we chose Metabolic Wiring instead of quorum sensing for our intercellular genetic circuit are:
- it is orthogonal to native signalling systems, thus only activates the genes it has been designed to.
- it provides a wider variety of potential signalling molecules, with virtually no crosstalk between them.
- unlike quorum sensing, the signal is consumed when received, thus avoiding saturation of the signal in the media which would cause the circuit to false fire. This is called signal quenching, and it is vital for virtually any inter-cellular genetic circuit in liquid culture.
Does it work?
The system was modelled mathematically and proven to produce the desired behaviour (fig.2). We used the mathematical model inform and fine tune the components of our circuit. We understood that in order for the signalling metabolite produced by each strain to represent the relative size of each strain’s population, it had to be degraded at the same rate that it was being produced, in order to avoid its saturation or depletion in the media.
Since for every one signalling metabolite there are two strains consuming it (receivers) and only one producing it (sender), the consumption rate by each of the two receivers had to be exactly half of the production rate by the sender, so that when summed together the net production rate is zero. Only then will the concentration of the each signalling metabolite correlate with the size of the population producing it. As part of our project we characterised the activity of multiple degron tags, which would allow us to fine tune the degradation rates of the enzymes and transcription factors relating to the consumption and production of each signalling metabolite, in order to achieve the correct rate ratios (consumption:production 1:2).
The resulting model indicated that the threshold for each population size was dynamic and relative to that the other strains. The concentration of the signalling metabolites in the medium is defined by the ratio of the its consumption and production rates, which in turn are defined by the ratios of the corresponding populations (the more cells, the more degrading/producing enzymes present globally). This meant that each population would adjust its size according to the ratio of its own size and the size of the other two.
![](https://static.igem.org/mediawiki/2014/4/42/Ed14_Simulation.png)
Figure 2: We modelled how the relative populations size of each strain affects the growth rate of the other two. Here, you can see how population 2 (blue) and 3 (green) will enter a transient stationary phase at 50 minutes. This is caused by the fact that population 1 (red), which is has been lagging up to that point, is being outgrown, which results in there not being a high enough concentration of its signalling molecule in the media to allow strain 2 and 3 to grow. For a description of the differential equations used to generate these graphs, please refer to the modelling page.
Selection of growth regulating genes
We performed an extensive literature research in order to find an array of genes that could be used for modulating the growth rate of a bacterial population at real time. The genes had to fulfil the following criteria:
- It must slow or arrest their growth.
- It must be simple to induce (one gene ideally) and reversible.
- The cells must still be able to produce and secrete signal of their own in this state.
- The cells must be able to stay in this state for long periods of time without dying.
The genes we found were RelA, RelA’ (a truncated version of RelA), rsd and rmf, which are described in more detail below. We found that overexpression of these genes was likely to induce the desired behaviour in our cell populations. We successfully cloned these genes in high copy plasmids and had them sequenced. We found that these genes were toxic at such high copy number, therefore we swapped the vector for a lower copy one. We advise future teams to use low copy plasmids for genes detrimental to the cells’ growth, and co-transform with a high copy plasmid harbouring the repressor, in order to achieve minimum leaking.
We found that another way to approach the same challenge, would be to silence a growth-related target gene, using antisense RNA (asRNA). We picked a gene that encoded a subunit of RNA polymerase that is necessary for exponential growth, sigma factor 70 (rpoD). The protocol we decided to follow (Nakashima et al., 2012), used a system of pair termini for stabilisation. We designed an algorithm that detected the optimal segment of the RNA for silencing, which is described here.
Progress
- PCRed out all the overexpression genes
- cloned them into high copy plasmids pSB1C3, under the control of Plac promoter
- sequences looked right for 2/4, but for both relAs there seemed to be a chunk omitted at around the middle of the sequence. We hypothesize that the reason for that was the toxicity of those genes at high copy, resulted in the cells “recombining out” part of the gene, thus effectively losing it.
This meant two things: 1] we needed to somehow impose a tighter control over the toxic gene, whether by using a tighter promoter or overexpressing the transcription repressor protein. 2] we had to find vectors with lower copy number, in order to avoid overwhelming toxicity to the cell.
Just to be sure, we decided to combine the two, and express the growth controlling gene from a low copy plasmid, while at the same time express the repressor from a high copy plasmid.
Talks with industry
Finally, we spoke with people working in the field to gain insights on how our project would be applied in the real world. Here is a description of the interview.
References
Lynd, L., Van Zyl, W., McBRIDE, J. and Laser, M. (2005). Consolidated bioprocessing of cellulosic biomass: an update. Current opinion in biotechnology, 16(5), pp.577--583.
Nakashima, N., Goh, S., Good, L. and Tamura, T. (2012). Multiple-gene silencing using antisense RNAs in Escherichia coli. Springer, pp.307--319.
Silva-Rocha, R. and de Lorenzo, V. (2013). Engineering multicellular logic in bacteria with metabolic wires. ACS synthetic biology, 3(4), pp.204--209.
Wang, B., Barahona, M. and Buck, M. (2013). A modular cell-based biosensor using engineered genetic logic circuits to detect and integrate multiple environmental signals. Biosensors and Bioelectronics, 40(1), pp.368--376.