Does anyone know, or can anyone point me to articles on how our DNA codes for repeating structures? I’m most interested in the brain and neural structures.
An example of a repeating structure would be the layers of cells in a retina (ganglion then amacrine then horizontal then bipolar and finally cone and rod), this series of cells is repeated many times. Maybe thinking of them as a series of cells is inaccurate and it’s better to think of them as multiple layers of similar cells. Either way there are horizontal and vertical boundaries that contain these types of cells, and there is a whole lot of repetition within those boundaries.
Thoughts regarding this issue:
We don’t have enough genes to code for every neuron, doesn’t this imply some method of compression in DNA?
It seems like a mechanism in DNA for repetition would be an advantage from an evolution standpoint because it would be easier to duplicate a structure that exists and works than it would be to keep re-inventing that cell sequence through random mutation to get where we are with our eye (and I’m sure other structures within the brain).
Well, I’m not so sure about your example of the neural pathways of retinas, those are all distinct cell types that are repeated in a fairly exact manner throughout the millions of neurons that make up the retina, so I guess I’m not entirely sure how it relates to the question of repeating structures, but let me take a stab at the broader question.
Yes, human molecular biology and DNA do have some “compression technology” built in.
To start with, between different cells types, the overwhelming majority of cellular machinery is conserved. The calcium pumps that help run your muscles are the same calcium pumps that run your nerves, the DNA helicase that helps the skin in your big-toe divide and make more skin cells is the exact same DNA helicase that is pumping out new epithelial (lining) cells for your small intestine. It isn’t like your DNA codes for, “okay, now, just on the underside of Jimmy’s spleen let’s put fibroblast #213324/278957” and here’s all the info to make that fibroblast. Your DNA says, “here’s everything in every cell that any cell might need.” Regulation of transcription, translation, and the entire field of developmental biology deals with the how and why certain undiferentiated stem from an 8-cell embryo end up as a particular rod in your retina.
And even these pieces of cellular machinery are overwhelmingly recycled with various degrees of change between species. Hence, you share a lot of your most basic genetic machinery with slugs, yeast, frogs and chimps.
Anyway, the really crazy numbers show up in your immune system. You’ve got roughly 20,000 genes in a human, and your body somehow manages to produce on the order of millions upon millions of different antibodies so that it can have just the right match for some crazy cold virus that no other human on the face of the earth has ever encountered. This is done in part via the magic of alternate splicing.
Basically, genes have different segments, kind of like components of a car, and alternate splicing lets you pick and choose and even rearrange these segments to make up some totally wierd looking things and this combinatorial diversity is how you get millions of different gene products from thousands of genes. Keywords for further reading would include introns, exons, and alternative splicing.
Okay, obviously that’s a huge topic covered in a five minute post, but is that a vaguely useful answer?
The bolded part is really what the question is about. Here are some fake-pseudocode-like-dna-type things that might make it more clear:
Example 1 - Brute Force Method of coding repeating structures
Gene #1-Cell type A
Gene #2-Cell type B connected to previous cell
Gene #3-Cell type C connected to previous cell
Gene #4-Cell type A (start of same structure adjacent to first one)
Gene #5-Cell type B connected to previous cell
Gene #6-Cell type C connected to previous cell
Gene #7-Cell type A
Gene #8-Cell type B connected to previous cell
Example 2 - Compressed/Tokenized version
Gene #1-Codes for Cell type A, B and C all connected in sequence
Gene #2-Copy gene #1 50 times
Example #1 doesn’t seem practical, so it seems something else must be going on, example #2 is just for illustration. So the question is, how does our DNA code for the structure (specific cell types connected in a specific manner) occuring many times?
Oh, okay, I think I have a better idea of what you’re asking now.
In that case, the answer is definitely Example 2.
Again, there’s really nothing that codes for every single cell like in the first example. Rather, the entire field of developmental biology deals with the question of how a small bunch of completely identical cells sitting in an embryo communicate and decide among themselves, “okay, you go that way, I’ll go this way,” and there are plenty of development genes that help to map and plan this out. But it’s more general than saying there are going to be 100 cells in a certain nuclear nucleus. It’s more of a stochastic process. Some population of developing cells is exposed to a certain amount of a cellular signal and that will cause on average a certain portion of those cells to develop into a certain cell type. At least in humans.
In nematodes, a much simpler animal of only a few thousand cells at full size, scientists have been able to describe almost the entire developmental fate of the entire animal, but I still think it’s a mater of a relatively small number of cellular signaling components leading to a much larger number of ultimate cellular fates.
Example 2 is not correct. Threemae however does describe in his next paragraph how stuff really does happen. It is not that you copy the gene 50 times therefore you get 50 repeats. it’s that the DNA codes for chemical signals which then set off a kind of chain reaction of differentiation and chemical signaling.
Let me expound a little bit with what little I know. (1 course in developmental bio!)
Lets say we’re talking about your repeating structure. It would be group of cells A lets off a set of chemical signals. This causes the cells adjacent to them to develop into cell-type B. Cell type B has it’s own set of expressing genes (genes can be expressed differentially, despite all cells having the same DNA). This cell-type B lets off it’s own cocktail of cell signals and causes the cells adjacent them to turn into Cell type C.
As for your other questions I Don’t really know, but i’ll take a stab at it.
1,2,3) Could be a thousand different mechanims. Some cells may code for surface protein X. Which happens to lock into surface protein Y expresed on another cell. Or Cell-type A might expressing chemokine Z which attracts Cell type B.
4,5) Again, it could be a billion reasons. But one of them is basic diffusion. Lets say CellType A which expresses chemical signal X is responsible for the entire retina. Lets place cell type A in the center of the retina. Then chemical signal X would be most concentrated in the area immediately around it and would be less and less concentrated as you move away from it until eventually it reaches a critically low concentration and that defines hte boundary.
Of course it may not be as simple as that. It is possible that different concentrations of chemical signal X may be responsible for different cell types. Lets say you need 100 parts per million to make precurser cell turn into cell type B. But as you move out to 50 part per million it’s no longer concentrated enough to make cell type B and the precurser cells instead become cell type C. and so on and so forth. The Interesting experiment would then be if you took SOme Cell type A(which make chemical signal X) and graft them to the outer boundary of the retina you’ll see that those 50 part per million cells which became celltype C are now exposed to 100 parts per million and will instead become type B.
Theres alot of different kinds of mechanisms at work in the human body. I’m sure humanity has barely scratched the surface of how it all comes together.
Hormones. A cell that is at a certain stage of growth releases chemicals that are picked up by any adjacent cells that are also in the right stage of growth. The two cells then move and grow together along the concentration gradient of the hormone. Only cells of the right type witll have the hormone receptor proteins “switched on”, and thus only cells of the proper type will reposnd. Adjacent cells of different types won’t have those receptors in production.
Quite simply it does’t. That is why helaing dmage to the CNS is almost impossible. The brain only connects to cells that are close by. When the embryo is less than an inch long the brain can connect to any cell in the body by following hormonal signals. As the body grows the brain remains in connnection with the body parts by lengthening those original connections, but the brain can’t sythesis a completely novel connection to a distant body part. It has to follow an established neural pathway to do so.
As far as we can tell it’s a combination of placement, counting replications and hormones.
The retinal cells can only develop from eye tissue, and that is laid down very, very early in development. Other cells in the body simply can not develop into retinal tissue, so the tissue can’t grow throughout the entire body.
As for more precise boundaries, that is where we get into counting replications. I dont; think anyone knows the excat mechanism for the retina, but we might imagine a scenarion like this:
We start with a single cell that is “eye”.
After it has divided 8 times the cell closest to the spinal cord becomes a retinal cell. It knows it’s closest to the spinal cord because it’s recieveing more homrones from the spinal cord than any other retinal cells.
It then divides a further 20 times to produce several hundred retinal cells. Once that number of divisions has passed it switches off and can never replicate again.
Alternatively it may be a case of it divinding once, and one of those cells producing hormone. The other cell line then keeps dividing until the concentration of hormone from the undivided cell drops below threshold X. Then the line stops dividing. If the hormone can only diffuse 2 mm then the retina can never be more than 2 mm thick.
It gets more complicated than that in practice. with other hormones acting as antagonists, agonists and synergists to produce structures that grow in specific directions and so forth, but if you can imagine how to grow a sphere under hormonal influence then you can imagine how to construct any other shape by adding to or subtracting from that sphere.
For more complex shapes the nerves and blood vessels themselves secrete hormones that cause the structure to grow into shape. Basically every developing cell knows how far it is from every other cell by the amount of hormone it recievs from that cell, and that allows very complicated 3D structures to be made.
More hormones. In simplified terms the body plan starts spherical and then due to gravitational effects some cells differentiate, this gives us our “top” and “bottom”, which thereafter are maintained by hormonal dominance rather than gravity in mammals. From that point some of the “top middle” cells differentiate into our spinal cord. The other structures than develop based on how far they are from the top and the middle. So for example the ears develop from where they are high levels of hormones secreted from “top” cells, so they only appear near the head of the embryo. By also knowing how much “bottom” hormone they are recieving they can position themselves a specific distance along the body axis. Similarly by responding to hormones produced by the spinal cord and the skin they can also position themselves a specific distance out from the central line.
In reality ears are far more complex than that, but the basic principle holds. By cells in the midline, top and bottom producing hormones the cells can be located anywhere within the three dimensional shape provided they are produced in pairs.
I know we’re not supposed to clutter the SDMB with content-less posts, but I just wanted to thank the OP for this thread, and especially the posters for their answers. Fascinating, crucial, basic stuff that we should all know about but few of us do.
One big misconception people have is that DNA is like a blueprint. It’s nothing like a blueprint. DNA doesn’t contain a master plan for building a human, with instructions for where to put the hair, what shape the bones should be, or how to construct an eye.
It’s much more analgous to a recipe: “Divide twenty times, then produce eznyme A, and if you are exposed to enzyme B turn into cell type C, but if exposed to enzyme D turn into cell type E. If you are cell type C start producing enzyme F until the concentration of enzyme G reaches a certain level, then switch to producing enzyme H. If you are cell type E stick to any neighbor cell that produces enzyme I and start dividing, but stop dividing if you come in contact with enzyme J, and start producing more enzyme J.” And so on and so on.
I might be nitpicking here, but I wanted to back up and note that DNA doesn’t instruct cells to do anything as described above. Proteins are what do all this stuff. Think of proteins as nature’s own nanotechnology, tiny little machines. DNA is the code that describes how to build them, and RNA actually builds them. Once built, the proteins are largely directed by chemical equilibrium and hormonal messaging.