Who Needs Genes?

Who Needs Genes?

One might object that the implicit information mentioned here so far is qualitatively different from genetic information simply because only genetic information is stable across generations; it can be inherited. But this is not so, either. Implicit, or "epigenetic" information can indeed be heritable, through a variety of mechanisms. Some forms of epigenetic inheritance are unarguable and unproblematic. Nerve cells beget other nerve cells and kidney cells beget other kidney cells, even though they both have precisely the same set of DNA. Whatever makes a nerve cell into a nerve cell can indeed be passed along to its offspring, and no one thinks that impossible, though the details remain mysterious.

Well before the discovery of the nature of DNA, Max Delbrück wrote in 1948 that biological homeostasis was a possible form of inheritance, though he added that it would work in cooperation with some kind of "definite series of genes."²⁵ According to his idea, the interaction between several different interdependent chemical reactions could produce many different stable states. Each state would be stable enough to propagate forward through generations of organisms.

Some years later, the discovery of gene regulation made possible such a system, and it wasn't many years more before one such system was observed in bacteria, in the lac gene. Part of the lac gene codes for an enzyme that allows lactose through the cell wall, a "permease." In a bath containing a low concentration of lactose, some bacteria will, just by random fluctuation, happen to switch on the lac gene, and make this enzyme. Once the enzyme is made, lactose comes into the cell, disables the repressor, and keeps the gene operating. When the cell divides, both children will continue to let lactose in, and digest it, as will their children and so on. The cells that don't switch on the gene will make no permease, and let in no lactose, and neither will their children.²⁶

This sort of inheritance is not as stable as inheritance through the genes; researchers observe that they can last for hundreds of generations, but tend eventually to decay. But this is inheritance nonetheless, and of an acquired character to boot, albeit one corresponding to a genetic predisposition.

There are also forms of structural inheritance, some of which have been observed in experiments with single-celled Paramecium. It is possible, through micro-surgery, to create physical modifications of Paramecia that are transmitted to the succeeding generations, showing both that acquired traits can be passed along, and that animals with identical genes need not have identical bodies.²⁷ It is not clear how relevant these effects are to multi-cellular organisms, but, to debunk yet another myth, cells do not build themselves from scratch. They divide, producing new cells with some of the same materials from the parent. That is, part of the reason a nerve cell may beget another nerve cell is the complex of genes that happen to be repressed or activated at the moment of division, and another part may simply be that the parent was a nerve cell.²⁸

Another form of epigenetic inheritance is regularly seen in sex-linked traits. Placental mammal genomes contain a variety of genes that are typically "imprinted." That is, their activity seems to depend on whether the gene was inherited from the mother or the father. ²⁹ Much of the development of the placenta and the fetus seem to proceed from genes imprinted with their parental origin.

DNA comes in chromosomes in the nucleus of each cell. To make a chromosome, you coil the DNA up tight, and wrap it around little protein plugs. The combination is called "chromatin." In addition to the kinds of activity recorded on the DNA--repressors and activators and the like--chromatin can apparently receive different kinds of "marks." One well known mark involves additional chemical groups that may be attached to some DNA bases, acting to suppress the activity of some regions of the DNA. The additional chemical group is often a methyl group, a carbon atom surrounded by three hydrogen atoms, and scientists refer to some stretch of DNA as "methylated" if it is disabled in this way. Sex-linked imprinting is thought to work this way.

Through the action of the enzyme "methyltransferase," methylation appears to be generally conserved across replication of DNA, though at lower fidelity than the DNA base pattern itself. That is, methylated patterns on a DNA copy will correspond to similar patterns on the original, providing an easy way for the suppression of a gene in a parent to be communicated to daughter cells. It appears that much of the methylation pattern on a gene is stripped off during the early development of eggs, but apparently not all, and it may be that patterns of methylation are behind the sex-linked imprinting of genes.

So not only is there implicit information available in the context in which DNA exists, but it is uncontroversial that this context can be modified to carry information from one generation to the next. There may, of course, be many more methods awaiting discovery than are outlined here. Genetic inheritance is still, of course, predominantly due to the effects of DNA and its various coding regimes, but there is sophistication enough in the context of the cell to carry a substantial amount of information from one generation to the next. A few specific mechanisms of this sort have now been identified in complex animals, such as mice,³⁰ as well as in plants,³¹ and there is no principled reason why many more shouldn't exist.

One difficulty with the information and the information-carriers outlined here is that their information does not reside in "patterns" instantly comprehensible to some external observer. The information, such as it is, is carried by systems of activity, relative concentrations of chemicals in solution, and the complex shape of biological actors. These forms of information are not readily reducible to the kind of information easily analyzed by classic information theory. That theory was developed for a theoretical framework in which to analyze the sending and receiving of a message. But that is simply not the way a cell works.³²

Erwin Chargaff, one of the pioneers of molecular biology, put it this way, in 1968:

We usually think of the message of DNA as being read by the cell, or by a nucleus. But our knowledge of the process is at once more detailed than that, and more confusing. If we think of a reader as an instance of the reading apparatus--the enzymes, the messenger RNA, the ribosomes--that makes DNA into a protein, then the message encoded in DNA is read by hundreds or thousands of different entities at the same time. These entities appear, do a little reading, interact with other entities engaged in the same kind of work, and disappear. What they certainly don't do is start from the "beginning" and work their way to the "end," as one might read a novel. Or a popular book on genetics.

One group of researchers has proposed that DNA is more aptly compared to data to be operated on by a program of a massively parallel architecture, "embedded in the global geometrical and biochemical structure of the cell." But the authors go on to say that even this comparison is marred by the complexity and interdependence of cell processes, not to mention the impoverishment of currently available parallel processing paradigms.³⁴

Who Needs Genes?