Team:Paris/Analysis/Design1

=Generating a FIFO sequential order of expression= A clock that would only indicate one hour, would not be very useful. The same goes for our BacteriO'clock. It is necessary that we add several distinct events in each cycle. We want to use three genes activated successively.

The easiest way to control a sequential expression of genes : the LIFO order


The temporal order of expression that is the easiest to set up is the Last In First Out temporal order. The genetic network that generates this behavior is the Single Input Module (SIM) (Network 1). A transcription factor X activates the expression of different genes Zi. When X concentration increases, the threshold of activation Ki of the different genes Z are reached in a precise order. The expression will stop in the opposite order when X is not expressed anymore. To cut a long story short, the first gene to be switched on will be the last one to be OFF.

This order poses a real problem to us : several states of the system are redundant (Table 1). We can not accept a clock that would display twice the same hour !

The rules to generate a FIFO temporal order
In the contrary, in a FIFO order, there is no confusion possible between different states (Table 1) : they are all distincts. To generate this complex behavior, there are two major possibilities.

The simplest way to generate a FIFO : a simple cascade of genes


The first one, the simplest, which is also the most unsophisticated, is a simple cascade of genes (Network 2). When the first gene Z1 is switched on, it activates the second gene Z2, then the third gene Z3 is turned on. When gene Z1 is turned off, Z2 is switched off first, then Z3...

This network is really simple to implement but presents several problems that makes it incompatible with our project. First of all, the cascade is not resistant to mutations : when one gene is mutated, the cascade is definitely broken. Secondly, the time scale if far too long. Each activation step would take approximately one hour. One full cycle would last almost six hours. Last but not least, the cascade is probably not very resistant to intracellular noise, it is not robust enough.

The more sophisticated way to generate a FIFO behavior : a particular Multiple Output Feed-Forward Loop
The other way to implement a FIFO is based on a logic structure called Feed Forward Loop.

More precisely, the genetic network is called Multiple Output FFL (Network 3). Each gene Zi is regulated by both X and Y, as in a typical FFL. X activates Zi when its concentration reaches the threshold KX/Zi. It is the same for Y and its respective thresholds.

All the MO FFL do not behave according to the FIFO order, three conditions are necessary and sufficient. KX/Z1 < KX/Z2 < KX/Z3 KY/Z1 > KY/Z2 > KY/Z3 This network has several advantages compared to the cascade. It is more resistant to mutations : if gene Z1 is mutated, genes Z2 and Z3 still behave as they should do. This is particularly interesting if we need more than three genes. The MOFFL is more robust concerning intracellular noisy fluctuations of concentration. The temporal expression of genes is more tunable : the expression of gene Z2 can be fully activated even if Z1 is not concentrated enough.
 * The input function for each gene Zi must be a OR or a SUM gate.
 * To have the order of expression Z1, Z2 and then Z3 at the switching ON, the thresholds must be ordered like this :
 * To have the FIFO order, Z1 must be the first gene to be switched OFF when X is no more expressed. When X is not expresses anymore, its concentration decreases. The threshold KX/Z1 will be reached after the thresholds KX/Z2 and KX/Z3. As Y is still expressed when X decreases, its own concentration will decrease with a delay. To have a FIFO, Y concentration must reach KY/Z1 before KX/Z2 and KX/Z3. The necessary order must be :

Implementation of our FIFO
Now that the necessaries and sufficient conditions have been depicted, we can start to describe our FIFO. Rather than designing it from scratch we decided to look in nature where we found and adopted the flagella system of construction in Escherichia coli which then inspired us further.


 * For X we chose flhDC and its natural promoter.
 * For Y, we chose fliA and its natural promoter.
 * For Z1, we chose fliL promoter and ECFP-LVA to report the expression. We chose ECFP because this fluorophore is not very fluorescent. As fliL promoter is the most strongly activated promoter, it compensate its weakness.
 * For Z2, we chose flgA promoter and YFP-LVA to report the expression.
 * For Z3, we chose flhB promoter and mRFP-LVA to report the expression. mRFP is the fluorophore that has the longest time of maturation. When put last in the FIFO, it helps to separate the different states.

The LVA tails are very important to reduce the fluorophores half lives from 20 hours minimum to 30 minutes. Those three fluorophores are compatible together, regarding to their absorption and emission wavelength.