5.2.1 Transcription Factors

1. Purpose

Transcription factor proteins regulate the rate of transcription of genes. They do not actually turn them on or off like gene control proteins, but they can essentially shut them down by totally suppressing the genes. However, this affect will only last as long as the factor is still around in enough quantity to effectively suppress the genes. Each gene has a list of enhancer sites that controls which factors can bind to that gene, and what affect they have on its transcription. If the binding ID of a factor matches the binding ID of an enhancer on a gene then that factor regulates the gene. But the next question is "How does it regulate the gene?"

Transcription Factor Properties

Protein Type: All proteins have this property. It is used to determine the type of protein to load. Binding ID: All proteins have this property. For TF's it matches up with enhancer's binding ID on genes to determine which genes that it can regulate. Degrade Rate: All proteins have this property. This determines the rate at which this protein is degraded in the cell.

Table 1. These are the different properties of a transcription factor protein that are defined in the digital genes.

Enhancer Site Properties

Binding ID: This value matches up to the BindingID of the transcription factor protein. Graph Type: This determines which type of graph to use for the expression function of this site. The expression function determines how the protein will affect the regulation of genes with matching Binding ID's A: This is one parameter of the expression function and its affect depends on the graph type. B: This is one parameter of the expression function and its affect depends on the graph type. C: This is one parameter of the expression function and its affect depends on the graph type. D: This is one parameter of the expression function and its affect depends on the graph type.

Table 2. These are the different properties of an enhancer site that are defined in the digital genes.

2. Enhancer Sites

TF's have the basic structure shown in table 1. Like all proteins they have a type, binding ID and a degrade rate. The transcription factor proteins are really simple and it is in the enhancer sites where the real work occurs. It is the site that determines the amount of that gene's protein that will be expressed based on the active quantity of the TF present in the cell. It does this by using an expression function. One of the main goals of this system is to be able to evolve the chromosomes in order to obtain some of them that will develop into neural networks that have desired behaviors. But for evolution to ever be able to work then the functional aspects of the proteins and genes have to be encoded somehow within the structure of the proteins and genes themselves. In nature it is the sequence of nucleotides that determines the folding pattern of the subunits, and the ultimate shape of the protein itself that defines the functionality of the protein. The shape of the protein determines which enhancer regions it will bind to on the gene, and what affect it will have once bound. The protein performs its function because its shape dictates its behavior. But digital proteins and genes don't actually have a shape, and to attempt to simulate a shape would be highly wasteful of processor time and memory. Digital proteins are really just mathematical abstractions, and it is with mathematical processes that they must affect the system. It is not necessary to simulate the shape of the protein or enhancer, just its affects. This is the purpose of the expression function. The basic types of graphs used in this system, their associated equations, and examples of each are shown in table 3. It is this expression function that determines the effect that the protein has on the system. In this particular instance the active quantity of the bound transcription factors that are present in the cell are fed into the expression function of the enhancer to get the quantity of protein to be expressed. The quantity expressed for all enhancers are added together for each gene to determine the actual amount expressed. So one enhancer could be expressing 500 units of the protein, and another enhancer on the same gene could be inhibiting it by 600 units. In this instance no protein would be produced because the inhibited quantity is greater than the expressed quantity. Since the effect of the TF is determined by the sites, and not the protein itself, then the same quantity of active TF could produce one effect for gene A, and a completely different effect for gene B.

Expression Function Equations

GraphType Equation Linear ExpressedQty = (B/A)*ActiveQty + C Bell ExpressedQty = B * e^(-C * (ActiveQty-A)^2) + D Sigmoid ExpressedQty = B/(1+e^(C*(A-ActiveQty))) + D

Table 3.These are the possible graph types for the expression function and their associated equations.

3. Quadrants Transcription Example

Quadrants Chromosome

Figure 1. This figure shows the genes involved in producing the quadrants transcription example.

Quadrants Results

Video 1. This video demonstrates the creation of quadrants using transcription factors.

The quadrants example demonstrates the classic "French flag problem." How does a developing organism use a concentration gradient to split its cells into discrete sections? This can be accomplished by having enhancer regions that regulate the transcription of the gene differently for different concentrations of the same transcription factor. Figure 2 shows an example of this. The curve shows a series of cells that has a gradient of a certain transcription factor. Gene 1 has an enhancer that will up-regulate its expression only if the level of the TF is above the red line. Whereas gene 2 will only up-regulate its expression if the level of TF is below the blue line. Gene 3 would be expressed basally and would be down-regulated by the transcription factors produced by genes 1 and 2. This leads to groups of cells that are each expressing a different set of transcription factors. This could then lead to the flipping on or off of various genes, and then these groups of cells are no longer the same. They are following paths to different fates.

Figure 1 shows the genes involved in producing this example. They work as just described. There is one set for dorsal-ventral segmentation, and another set for anterior-posterior segmentation. There is also a gene that is only expressed in the center of the pattern. The cells start out with two concentration gradients. One vertical gradient of the transcription factor VG (Vertical Gradient) and another horizontal gradient of the transcription factor HG (Horizontal Gradient). The G_Top gene is expressed only for low levels of the transcription factor VG, and the G_Bottom gene is expressed only for high levels of that same transcription factor. Each of these genes produce TF's that inhibit the G_HMiddle gene. That gene normally gets expressed at a basal level in every cell, but when either of the other two TF's are present it is inhibited. This is what forms the horizontal stripes. The vertical stripes are formed the same way using the genes G_Left, G_Right, and G_VMiddle. Finally, G_Middle is only expressed when both HMiddle and VMiddle TF's are present in sufficient quantity. Since this is only found in the center of the graph where the two middle stripes intersect, then this is the only place where that transcription factor is expressed.

Using Gradients

Figure 2. This shows how concentration gradients can be used to divide a collection of cells into different sections. Each gene has enhancers that only transcribe the gene if the quantity of active transcription factor in the cell is above or below a threshold amount. In this figure the green line represents the concentration gradient of a transcription factor in series of cells. Only those with a concentration above the red line will transcribe gene 1. Similarly, only those cells with a concentration of the TF below the blue line will transcribe gene 2. Gene 3 is basally transcribed and is inhibited by genes 1 and 2. There are other ways to work this as well. Gene 3 could stay the same, and gene 2 could only be transcribed for quantities above the blue line. Then you would need gene 1 to inhibit transcription of gene 2 in order to have distinct segments.

4. Diamond Transcription Example

Diamond Chromosome

Figure 3. This figure shows the genes involved in producing the diamond transcription example.

Diamond Results

Video 2. This video demonstrates the creation of a diamond using transcription factors.

It is also possible to create more complex shapes, even using very simple genes. The diamond example produces a diamond shape group of cells that are expressing the TF using only one gene. Like the example above we start with two concentration gradients for both the vertical and horizontal directions. The G_Diamond gene has two enhancers. Each of these enhancers use a bell curve for their expression function. The curve is offset so that only the top of bell is above zero. So within a narrow range of TF levels it will express the gene, and outside of that range it will inhibit expression. One of the enhancers binds to the vertical gradient protein VG and the other binds to the horizontal gradient protein HG. These two enhancers interact so that only when both of these bell curves are positive will the gene actually be expressed. Otherwise the one enhancer inhibits the other one. This produces the diamond shape of TF expression.

5. Doughnut Transcription Example

Doughnut Chromosome

Figure 4. This figure shows the genes involved in producing the doughnut transcription example.

Doughnut Results

Video 3. This video demonstrates the creation of a doughnut using transcription factors.

This demonstration builds on the diamond example to produce an even more complex shape. It creates a doughnut pattern. It starts out just like the diamond example using two bell curve enhancers. They have been slightly modified to make a wider diamond shape. The diamond TF then binds to the enhancer of the G_Doughnut gene. This gene also has two enhancers, but with a twist. Both enhancers bind to the same transcription factor, diamond. The difference is that they each have different expression functions. The first enhancer uses a sigmoid curve to up-regulate expression for all cells with an active diamond quantity greater than 1000. The second enhancer also uses a sigmoid curve to inhibit expression for all cells with an active diamond quantity greater than 2000. This is what causes the doughnut shape. Since the diamond is made with two bell curve enhancers the closer you get to the center point the larger the quantity of TF. So the center of the diamond has higher levels of diamond protein than the surrounding cells, and the two enhancers work together to only express G_Doughnut in those surrounding cells. Pretty Neat!

6. Transcription Factor Overview

Transcription factors and enhancer sites are some of the most important pieces of the developmental simulator. They are key links that allows the expressed proteins to regulate the future expression of other genes. This allows for the creation of positive and negative feedback systems that are a critical part of any dynamic, adaptive system. Each TF can bind to an enhancer on different genes or the same gene. It is how these enhancers for a gene interact that determines the expression rate of that gene. The expression function allows for the encoding of the functionality of the site into a highly compact mathematical relationship that is an integral part of the chromosome structure. This is critical if evolution is to be able to act on the proteins and genes to produce systems that demonstrate desired outcomes. All of these things come together to make a highly flexible system that is also compact and extremely fast during processing.