During my postdoctoral work at AMOLF in Amsterdam, I was introduced to the topic of genetic switches. Genetic switches are gene regulatory networks; i.e. collections of genes which act to switch each other on and off. In particular, the "expression" of any gene, or in other words the rate at which protein is produced corresponding to that gene, depends on the rate at which it is transcribed, or in other words the rate at which DNA polymerase enzyme copies its DNA seuence to make messenger RNA molecules, which can then be used to make protein. The rate of transcription is affected by which proteins are bound to the DNA close to the start of a gene. These regulatory proteins, which are called transcription factors, are of course encoded by other genes. Thus the protein product of one gene can turn on or off the expression of its own, or another, gene.
Genetic switches are a particular class of gene regulatory network that shows bistable behaviour. A simple genetic switch is shown below.
The 'rod' in the figure represents a piece of DNA, possibly part of the bacterial chromosome. The light blue and pink rectangles are the genes encoding the light blue and pink proteins, shown as circles. The coloured arrows indicate that proteins are produced from these genes. The green gene encodes Green Fluorescent Protein (GFP) which can be seen under the microscope and is used to follow what is happening to the bacteria. The DNA also contains 'regulatory sequences', shown in the figure as the dark blue and brown rectangles, which are the sites at which transcription factors can bind and control expression of the genes. In this network, the pink protein binds to the DNA at the dark blue regulatory sequence. This prevents the light blue protein from being produced. Likewise, the light blue protein binds to the brown regulatory sequence, preventing production of the pink protein. GFP production is also prevented if the blue protein binds to the brown sequence. The net result is that this network is bistable: it has two alternative stable steady states, one in which a large amount of pink protein, but little blue protein, is present (here GFP is abundant and the bacteria are green), and the other where the blue protein, but not the pink, is abundant (here GFP is not abundant and the bacteria are not green). Because all biological systems are subject to random fluctuations, the network is expected to flip occasionally between these two stable states. One of the subjects of my work is predicting this flipping rate, using computer simulations, and comparing the results of the simulations to experimental measurements.
A very interesting "real" genetic switch is the lysis-lysogeny switch of the bacteriophage lambda. This switch determines whether the virus lambda will multiply inside its host bacterium, killing the bacterium and releasing many progeny viruses (this is called lysis), or whether it will insert itself into the DNA of the bacterium and remain dormant there for many bacterial generations (this is called lysogeny). Many questions remain to be answered about this switch (which seems to be extremely stable to fluctuations), and computer simulations may be a good way to do this. Marco Morelli, Pieter Rein ten Wolde and I worked together to use stochastic simulations to simulate the flipping of this switch. Our simulations showed that a DNA looping interaction can stabilise the switch against intracellular destabilising factors such as nonspecific binding of transcription factors to genomic DNA.