DNA...where is the missing information?

So, you've got your DNA strand, and it's just toddling along like so...

............................

And every so often (in eukaryotes, at any rate), there is a gene.

............:::::::::::::::::::::::::::.............

Now, on it's own, not much happens with this gene. But if it's lucky, there is a region of DNA upstream of it that makes things happen.

.......!!!!!!.....:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::.....

The !!!!! region can do any number of things. A chemical messenger might bind to it and block the :::::: stretch so that the gene cannot be transcribed. Or the !!!!! region might attract a messenger that helps the gene to get transcribed.

Genes usually have multiple !!!!! regions that respond to different messaengers and have different effects on the gene.

Similarly, different genes have similar !!!!! so they all get turned off or on by the same chemical messenger. Many, many genes, for instance, have !!!!! areas (or regulatory elements, I think some folks call them) that respond to corticosteroid hormones (e.g., corticosterone or cortisol, depending on whether you're a man or a dog).





That's one tiny piece of the explanation, anyway. The story during development (when this kind of stuff is really, really important) is decidedly more complicated, but it has been mapped out in pretty good detail in some critters -- fruit flies and sea urchins, for instance.

Patiodog,

Ok, I basically get this explanation, except that it seems to have some major holes which I am trying to understand.

The genes that get expressed are determined entirely by which chemical messengers are binding or interacting with the DNA. But how does it happen that each cell is recieving the proper messengers to activate the right part of DNA for that cell's location?

The explanation that I have heard is that the ratios of chemical messengers present at any cell in the body provide enough information to uniquely identify the position of the cell...basically by using a more complicated form of triangulation.

While this may make sense for creating simple Turing patterns, it does not appear to me that this kind of logic can extend to arbitrary complexity. The signal to noise ratio just seems impossibly low.

Don't have any big anwers for you, but you may not be fully appreciating that the cell's identity has as much or more to do with it's history than with its location. The initial distinctions between cell type during development are very rudimentary ones -- e.g., am I mesendoderm (if that was the right term) or am I ectoderm? Am I mesoderm or am I endoderm? It's not as though all switches are thrown at once, but rather that each switch narrows the range of switches available at the next step of determination.

And these early (and later, for that matter) switches have a probabalistic element to them. We're perhaps most familiar with this process in the immune system, where, for instance, a naive CD4+ T cell might become a Th1 cell or a Th2 cell, depending on its chemical environment upon activation. Bacterial and viral infections will tend to favor a switch to Th1, parasitic infections (or allergy, through a not-understood mechanism) tend to favor a switch to Th2. There always will be a mixture of the two cell types, though. The trick is that Th1 cells produce messengers that favor the development of more Th1 cells and Th2 cells favor the development of more Th2 cells (and each inhibits the other type) -- so an early disparity between numbers of Th1 and Th2 are magnified during subsequent generations.

That's it for now. Largely talking out of my ass and on the fly, so who knows if this is accurately applied or not...

Thinking also about the nervous system (which bears comparison in many respects, including lineage to the immune system)... The complexity and precision of the nervous system is largely the result of winnowing down. Early on in development, everything is hooked up to everything around it, and over time the number of cells and the number of connections between cells actually decreases (in the cases of cerebral synapses, for instance, by about 90%). The system isn't built to conform to a blueprint, but rather it is greatly overbuilt and then circuits that don't perform useful functions are interrupted and useless elements die off. (More positive feedback stuff, like with the Th1 and Th2 cells.)

The degree to which architecture and function in the nervous system is conserved is impressive, but also worthy of note is the degree to which neural anatomy and function varies between individuals. Dentists are well aware that a sizeable proportion of the population has "abnormal" nerve structures in their mouths that make them difficult to numb for basic procedures -- and yet these abnormal folks have no deficits in terms the actual functions of the nerve.

I asked my sister (a pathologist) and she had this to say:



Okay, you both seem to agree that things are determined "early on"...but that doesn't tell me how they are determined

Big question. Depends on what stage of development, what kingdom/phylyum/class/etc. you're talking about, blah blah blah.

Got my animal development textbook out (Gilbert, 7th edition).

Early experiments with what is called "fate mapping" found that if you injected a particular part of the egg of a tunicate (for instance) with a dye and let it develop, particular tissues would end up displaying that dye. For instance, the "vegetal pole" of the egg (the bit with more yolk) would become endoderm (the lining of the GI tract and it's glandular tissues) while the "animal pole" of the egg would become ectoderm (epidermis, most neural tissue).

Early experiments also found that zygotes were anteriorly and posteriorly determined. Cells that arose from the posterior of the zygote near the equator (that is, between the animal and vegetal poles) became mesodermal tissues -- specifically, muscle and mesenchyme (the cells that make up the "pulp," if you will, of organs. Cells that arose from the anterior of the zygote near the equator became notochord (a sort of proto-vertebral column that induces in the development of the spinal cord and, in vertebrates, is the substrate around which the spinal column will be formed). When the developing embryo folds in upon itself (a process called gastrulation) these tissues assume the relationship to each other that we find in the adult tunicate (which is taken to be the protochordate).

Later molecular analysis has found that the poles of the embryo also correspond to concentration gradients of maternally-derived mRNA in the ovum. These mRNA molecules, then, end up in the cells which arise through the process of cleavage from these regions of the ovum, which in turn determines the fate of those cells.

A pdf with some good imagery (at least at first glance it appears to be) can be found here. It glosses some determinant factors in development in various phyla. There also might be some good stuff at NIH's web site here and here.

I really should be doing other (bureaucratic) stuff, so I'm not really taking the time to look it over. My very quick summary is -- "you owe most of it to mom, and a little bit to where the sperm hit the egg."





Of course, the answer to most of the questions that arise from your line of inquiry is "no one knows." It's biology, after all.

One new study indicates that there may be an additional layer (or layers) of encoding overlaid on the basic layer of information that we have recognized for years.