Signal Model of Gene Expression

The signal model of gene expression - a gene is understood as emitting a signal to the cell who recieves the signal, desciphers it, and acts accordingly (i.e. expresses the gene or not). This model of gene expression is particularly germane because of the recent success of Hidden Markov Models in predicting gene structure. HMMs model the cell as parsing a given gene, and while the cell is parsing a gene, it passes through a series of hidden states that probabilistically determine the eventual outcome of the parsing event. This process is modeled as a linear reading of the genome starting in the 3' promoter region and progressing 5' until a termination state is reached.

Transcription Initiation - An Analogy

Imagine a hot air ballon that is attempting to land and is in the process of being tethered to the ground, but still has enough hot air to want to rise upward and fly away (or maybe it's just windy). The upward force of the hot air (or the wind) is analagous to the chaotic, stochastic forces in the cell's nucleus that try to push the initiation complex proteins apart and away from the promoter region of a gene. The stakes in the ground are analagous to the sequences in the DNA that the initiation complex proteins recognize and bind. For example, the main stake might be a TATA box to which TFIID tethers the rest of the initiation complex. Every stake that connects the ballon to the ground increases its chances of landing successfully. Enhancers are extra stakes that increase the expression of a gene by recruiting the initiation complex proteins or by increasing the stability of the initiation comlex itself (remember that the DNA loops and folds so that sequences that seem far away might actually be just above or below the initiation complex). Silencers stretch the analogy somewhat, but they might be analagous to large trees in the way of the landing zone blocking a clear landing. (Silencers stretch the analogy because the trees would have to float and have their own tethers as well, because silencers are separate domains that repressing transcription factor proteins recoginze and bind to. So maybe steric interference from another ballon might be more approapriate.)

I used the balloon analogy for several reasons: there could be any number of stakes; the stakes/ropes can be anywhere around the ballon; the ballon left by itself will simply blow away; landing a ballon is a stochastic even.

It is important to understand that this process is not deterministic, it is inherently stocastic. Initiation events should not be viewed as on/off switches, but instead as probabilistic switches. Furthermore, in contrast to computers, where signal degredation is always bad and the result of an error, in the cell the strength of the signal is itself part of the signal. A strong consensus promoter will be more likely to initiate transcription and thus will be more highly expressed. Conversely, a weak consensus promoter will be expressed at a lower level because the initiation complex will be less likely to form. Note: the strength of the consensus of a promoter refers to how many of the important nucleotides read by the basal transcription factors have been conserved. Each conserved nucleotide increases the 'strength' of the signal emitted by the gene to the cell making the promoter stickier to the basal transcription factors.

An Example

As an example let's walk throught the imaginary expression of gene A. Initially for gene A, no enhancing transcription factors are being expressed, and therefore only the basal transcription factors are available to form the initiation complex. Imagine in this situation gene A's expression is at 10%. One way of understanding this number is 9 out of 10 times the initiation complex is formed, or begins to form, it is disbanded before transcription can begin. Enhancers work by stabilizing this process. Imagine an enhancing transcription factor binds an enhancer site near the initiation complex and increases the gene's expression to 40%; the enhancer has acted by increasing the binding strength of the initiation complex and stabilizing it, such that now 4 out of 10 times successful transcription occurs. Multiple enhancers can work together as well. Another enhancer could bind another enhancer site nearby and further increase the stability of the initiation complex bringing the expression level up to 60%.

Another (possibly more biologically realistic) alternative to a second enhancer, is that the original enhancer is more highly expressed. Let's say that at the initial level of expression, the original enhancer for gene A binds the enhancer sequence near the gene 20% of the time, and thus 2 out of 10 times the initiation complex forms the enhancer is around to help stabilize it. It's easy to imagine that if more of the enhancer is floating around in the nucleus, more of it will bind the enhancer sequence near the gene and thus the initiation complex will, on average, be more stable. So if the enhancer's expression is upregulated and twice as much of it is expressed, 4 out of 10 times the initiation complex forms the enhancer will be there to stabilize the complex and thus increase the expression of gene A.

Some Conclusions

We have now developed a fairly good picture of eukaryotic transcription. Here is a summary of some of the main points I have tried to make.