|
|
| Line 1,706: |
Line 1,706: |
| | </p> | | </p> |
| | <h2 id="pfont">Since oscillations are a phenomena that require observation at a small scale (level of very few cells or even single cells), we designed microfluidic chambers in order to load our cells and observe oscillations. | | <h2 id="pfont">Since oscillations are a phenomena that require observation at a small scale (level of very few cells or even single cells), we designed microfluidic chambers in order to load our cells and observe oscillations. |
| − | <br><br>
| |
| − | We used standard soft lithography techniques to generate microfluidic channels. In brief, SU8 photoresist was spin coated on a silicon wafer to the height of 50μm. The desired pattern was generated using maskless lithography. A silicone elastomer was added with its curing agent in 10:1 volume ratio and poured over the micro-mold. After 4 hours of incubation at 65oC, PDMS was peeled off the silicon wafer and inlet and outlet holes were punched. The surface of PDMS and a cover slip were modified using a plasma cleaner and microfluidic channels were created by bonding the two together.
| |
| | <br> | | <br> |
| | + | We used standard soft lithography techniques to generate microfluidic channels. In brief, SU8 photoresist was spin coated on a silicon wafer to the height of 50μm. The desired pattern was generated using maskless lithography. A silicone elastomer was added with its curing agent in 10:1 volume ratio and poured over the micro-mold. After 4 hours of incubation at 65oC, PDMS was peeled off the silicon wafer and inlet and outlet holes were punched. The surface of PDMS and a cover slip were modified using a plasma cleaner and microfluidic channels were created by bonding the two together. |
| | + | <br><br> |
| | <img src = "https://static.igem.org/mediawiki/2017/4/45/T--IIT_Delhi--picture3.png" style='border:3px solid #000000'><br> | | <img src = "https://static.igem.org/mediawiki/2017/4/45/T--IIT_Delhi--picture3.png" style='border:3px solid #000000'><br> |
| | <br> | | <br> |
| − | Note, that here, we have not written the reaction where mRNA is being converted to protein, since mRNA is not actually being consumed there or being produced. 1 molecule of mRNA simply produces 1 molecule of protein (assumption). <br>
| + | Cell culture grown overnight containing LB media was loaded directly into the channels, or was diluted 1:50 and flowed through the main channel. Air bubbles were introduced at the T-junction. Hence droplets of water surrounded by air on either side were created in the channel. Cells trapped in these droplets were studied under a fluorescence microscope using a 40x objective. <br> |
| | | | |
| − | Further, it has to be noted that the [DNA] and [mRNA] terms appear in the equation since in writing the model, we assume that mass action kinetics are valid, ie, the rate of the reaction is equal to the rate constant times the concentration of the reactant, raised to a power equal to the number of molecules of the reactant. <br>
| + | Two types of fluorescence microscopes were used for our studies. The first had a mercury lamp as the light source, with a black and white camera. We were able to observe fluorescence in cells that were constitutively expressing GFP, containing the reporter under the PhlF repressible promoter. The system was on a high copy, and PhlF was not being produced, thereby rendering the promoter to be constitutively ON. <br> |
| − | | + | However, for the purpose of our further work, we used the Etaluma Lumascope S40 fluorescence microscope, which had an LED light source, with the appropriate excitation, emission and dichroic filters for observing our GFP levels. Loaded in the channel, our cells showed fluorescence, and a sample image is shown below – <br><br> |
| − | Now, we know that the DNA concentration remains constant and does not change over time. Therefore, the [DNA] term can be included in the constant itself, to give <br><br>
| + | |
| | <img src = "https://static.igem.org/mediawiki/2017/3/3c/T--IIT_Delhi--picture4.png" style='border:3px solid #000000'><br> | | <img src = "https://static.igem.org/mediawiki/2017/3/3c/T--IIT_Delhi--picture4.png" style='border:3px solid #000000'><br> |
| − | <br>
| |
| − | Now, the dynamics of the protein can be similarly written as <br><br>
| |
| − | <img src = "https://static.igem.org/mediawiki/2017/7/77/T--IIT_Delhi--picture5.png" style='border:3px solid #000000'><br><br>
| |
| − |
| |
| − | And that is it! We’ve just written down our first model, for a gene being expressed from a constitutive promoter. Now that we have our model, we can simulate these and find out the dynamics. <br>
| |
| − |
| |
| − | Simulation basically means solving the differential equations to get the variation of the component (mRNA, protein) with time. This can be done by hand for the equations above. However, as models get more complex, implicit equations appear, which are much more difficult to solve by hand. Thus, it is essential to get the hang of modelling software such as MATLAB or R, which solve differential equations and simulate the model for a specified period of time. <br>
| |
| − |
| |
| − | Thus, we write down the model on MATLAB here, and simulate it for a time period of 200 time units. The values of the constants used for alpha, gamma etc and the MATLAB code for the same can be found on the github library link given below. The plot obtained is as follows - <br><br>
| |
| − | <img src = "https://static.igem.org/mediawiki/2017/8/80/T--IIT_Delhi--picture6.jpeg" style='border:3px solid #000000' width="90%"><br><br>
| |
| − |
| |
| − | Changing the parameters for production and degradation rates can give different kinds of graphs, and can be explored by simply changing the values of alpha, gamma, K etc in the model and simulating the same. However, as we can see here, the mRNA and protein levels both rise to a certain fixed value. This is known as the steady state value.
| |
| − | <br>
| |
| − | However, we can make a further simplification in this model. Generally, the mRNA dynamics are faster than the protein dynamics. This means that mRNA levels approach their steady state value faster than proteins do. Therefore, we can say make the assumption and further simplification that before the protein dynamics start to come into play, the date of change of mRNA is zero. This is known as the “quasi steady state assumption”.<br><br>
| |
| − |
| |
| − | Therefore at steady state,
| |
| − | <br><br>
| |
| − | <img src = "https://static.igem.org/mediawiki/2017/b/b6/T--IIT_Delhi--picture7.png" style='border:3px solid #000000'><br><br>
| |
| − |
| |
| − | Thus, <br><br>
| |
| − | <img src = "https://static.igem.org/mediawiki/2017/f/fc/T--IIT_Delhi--picture8.png" style='border:3px solid #000000' solid #000000'><br><br>
| |
| − | Now, we can replace the value of [mRNA] in equation (2) with the value given above, to get - <br><br>
| |
| − | <img src = "https://static.igem.org/mediawiki/2017/e/ea/T--IIT_Delhi--picture9.png" style='border:3px solid #000000'><br><br>
| |
| − |
| |
| − | We can now try to simulate and plot the graph for the protein levels, and compare the time series of the two models - <br><br>
| |
| − | <center><img src = "https://static.igem.org/mediawiki/2017/9/93/T--IIT_Delhi--picture10.png" style='border:3px solid #000000' width="90%"></center><br><br>
| |
| − |
| |
| − | Therefore, we can see that by making the assumption that mRNA is already at steady state at the start of time, the protein levels begin to rise faster than the earlier model. However, the steady state value for protein remains the same. This is because we have only simplified the model by changing the time scale and assuming that at the given time scale, mRNA dynamics are at steady state. We have not changed the steady state per se.
| |
| − |
| |
| | | | |
| | | | |
| Line 1,756: |
Line 1,726: |
| | | | |
| | | | |
| − | <h2 class="h2font">Model for<br> Regulated Gene Expression</h2> | + | <h2 class="h2font">Maintaing<br> Flow Rate</h2> |
| | | | |
| | <p> | | <p> |
| Line 1,767: |
Line 1,737: |
| | <h2 id="pfont"> | | <h2 id="pfont"> |
| | | | |
| − | Regulation of gene expression involves changing the expression of protein or RNA produced by a particular gene. Various mechanisms exist, for doing the same allowing for control at various stages of the expression of the gene. For instance, if the control/regulation is such that it does not allow transcription to happen, it is termed as transcriptional control. Similarly, translational, post translational and several other layers of control exist. <br>
| + | RIn order to visualize the oscillations, we needed to maintain a flow rate in the system. This is because as cells would grow, they would produce toxic compounds inside the chamber, and also reach stationary phase, which would cause a slowdown in protein production, which could kill the oscillations. Flow rate maintenance is something that commonly needs to be done in microfluidic chambers where oscillations are required to be seen, but such sophisticated systems for flow rate maintenance are extremely costly and require specialized equipment to handle. <br> |
| − | | + | |
| − | The simplest and most commonly employed mode of regulation is the transcriptional control by repressor proteins. These are protein molecules that can bind to specific “operator” sites in the promoter region, and stop the promoter to recruit RNA polymerase successfully, thereby inhibiting transcription. Common examples of such systems are LacI, TetR and cI, which can inhibit transcription from the pLac, pTet and pCI promoter respectively. This mode of control is also commonly referred to as repression, and should not be confused with inhibition, which is a separate control mechanism.<br>
| + | |
| − | | + | |
| − | Here, let us try to model what regulated gene expression looks like, by looking at a typical example of transcriptional activation. Consider the following case - <br><br>
| + | |
| − | | + | |
| − | <img src = "https://static.igem.org/mediawiki/2017/6/65/T--IIT_Delhi--picture11.png" style='border:3px solid #000000'><br><br>
| + | |
| − | | + | |
| − | We have a protein X, that can exist in two states, the native (inactive) state X, and an active form X*. The molecule X* can bind to the promoter (say P), and promote transcription of the gene by helping the promoter to recruit RNA polymerase.
| + | |
| − | <br><br>
| + | |
| − | Lets look at another case, where we have transcriptional repression -<br><br>
| + | |
| − | | + | |
| − | <img src = "https://static.igem.org/mediawiki/2017/2/2b/T--IIT_Delhi--picture12.png" width="100%" ><br><br>
| + | |
| − | | + | |
| − | Here we are given 2 proteins, A and B. The protein A is produced and degraded, and is a transcriptional repressor for the gene B. A has its own production and degradation rates, described by alpha and gamma, and B also has its own production and degradation rates, given by beta and gamma respectively. Further, DA and Do represent the two states that the DNA region of the promoter PB can have. DA represents the state where A is bound to the operator, and Do represents the state where A is unbound.
| + | |
| − | <br>
| + | |
| − | The system can be represented by a set of reactions as follows –<br><br>
| + | |
| − | <img src = "https://static.igem.org/mediawiki/2017/9/92/T--IIT_Delhi--picture13.png" style='border:3px solid #000000'><br><br>
| + | |
| − | | + | |
| − | Based on these reactions, we can write the mass action model for the system. This can be represented by the following differential equations –<br><br>
| + | |
| − | <img src = "https://static.igem.org/mediawiki/2017/5/52/T--IIT_Delhi--picture14.png" style='border:3px solid #000000'><br><br>
| + | |
| − | | + | |
| − | Further, we have a 5th equation in the model, which is based on the conservation of DNA. Since all of the DNA of the promoter can either be bound by transcription factor A (DA state) or be unbound (Do state), therefore, the total DNA (DT), at any time, can be represented as –<br><br>
| + | |
| − | <img src = "https://static.igem.org/mediawiki/2017/5/52/T--IIT_Delhi--picture15.png" style='border:1px solid #000000'><br><br>
| + | |
| − | | + | |
| − | Thus, the model of the system, based on mass action kinetics and conservation relations can be represented by<br><br>
| + | |
| − | <img src = "https://static.igem.org/mediawiki/2017/b/b4/T--IIT_Delhi--picture16.png" style='border:3px solid #000000'><br><br>
| + | |
| − | | + | |
| − | Note that in this model, we have taken the rate of change of Do and DA as well, which are DNA molecules. This is because here the DNA concentration also changes because the DNA switches states.
| + | |
| − | <br><br>
| + | |
| − | Now, solving this model and simulating, we get the following results - <br><br>
| + | |
| − | <img src = "https://static.igem.org/mediawiki/2017/e/ec/T--IIT_Delhi--picture17.png" style='border:3px solid #000000'><br><br>
| + | |
| − | <img src = "https://static.igem.org/mediawiki/2017/f/fc/T--IIT_Delhi--picture18.jpg" style='border:3px solid #000000' width="90%"><br><br>
| + | |
| − | | + | |
| − | From the above results, we can see that the time taken for achieving steady state for all the variables (A, Do, DA and B) is more or less similar, and takes about 4-5 hours. This goes against the intuition that the binding and unbinding happens faster, as compared to the production of A and B, which should take a larger time. We see that this does happen, when we run the system of equations for α = 100 nM/hr (results not shown). Thus, the time scale separations become more prominent as the value of α increases (time scale separation was further more prominent for α = 500 nM/hr).
| + | |
| − | <br><br>
| + | |
| − | Further, upon varying the values of k1 and k2 by 100 fold, we see the following –<br><br>
| + | |
| − | <img src = "https://static.igem.org/mediawiki/2017/9/92/T--IIT_Delhi--picture19.png" style='border:3px solid #000000' width="90%"><br><br>
| + | |
| | | | |
| − | In all of these plots, data1, data2 and data3 represent k1 = 0.01, 1 and 100 respectively. We see that on decreasing the value of k1, the effect on the steady state values is not significant. On the other hand, increasing k1 by 100 fold changes the steady state values, and brings down the level of B ultimately produced at steady state. This is because if k1 is high, that means that more A binds to the promoter of B, repressing it. Therefore, a lower steady state level of B is observed.
| + | We designed our own cost effective solution to the problem, by using commonly available material. From the required velocity of the fresh medium that would be needed, the flow rate was calculated. This was in turn used to calculate the pressure difference using the Hagen Poiseuille law. This pressure difference that would be required for the system was achieved by filling media inside a rubber pipe of small diameter (0.4 mm), and placing it up to a height h such that the hydrostatic pressure would be equal to the pressure difference required. <br> |
| | | | |
| | + | The height could be modulated to generate different flow rates and tune them specifically for our requirements. The following videos show the different flow rates that could be achieved – |
| | </h2> | | </h2> |
| | </header> | | </header> |
