Les Mécaniciens: Salon des Refusés
Biological Information and Some Problems for Regularity Requirements in the Analysis of Mechanisms
Richard M. Burian, Virginia Tech
Preliminary Text for My Presentation; Longer Text Eventually Available: ‘Protein synthesis’ and ‘gene expression’ are a bad labels for a multitude of cellularly regulated steps that deploy specifically different machinery in different cells, sometimes at several removes from the actual processing of DNA to yield products. The steps fall into five groups: (1) some alter whether certain stretches of DNA are accessible for transcription, (2) some are involved in transcribing specific stretches of DNA, with different likelihoods of starting and stopping at different points in different cells, (3) some are involved in processing the resulting mRNAs in various ways in the nucleus (including, in eukaryotes, sometimes editing their nucleotide sequences before transporting them out of the nucleus), (4) some process RNA in yet further ways in the cytoplasm, differently in different cells, and, for those mRNAs that are finally processed by ribosomes to yield polypeptide chains, (5) some are involved in processing those polypeptide chains to make proteins, adding non-nucleotide components (e.g., sulfur molecules, methylations, phosphorylations, etc.) linking polypeptide chains together or deleting parts of them, and establishing specific conformations. Even then the resulting molecules may be altered or destroyed by chaperonins and associated machinery, in effect because many cells contain devices to eliminate conformations of specific polypeptide sequences not encountered early in the history of the relevant cell lineage. These last steps are not trivial: identical polypeptide chains in different cells can yield different proteins! For example a human aldehyde dehydrogenase and a particular human lens crystallin are made from the same exons of a single gene; these are distinct proteins with the same sequence of amino acids. (An excellent general reference on this topic is Piatigorsky J., 2007, Gene Sharing and Evolution: The Diversity of Protein Functions, Cambridge, MA: Harvard University Press).
There are regulatory controls affecting every single step that I have described – and I have described only an outline of the complexities involved. Thus regulation of gene expression can occur at each and every stage of the steps involved in all five phases of protein synthesis (and similar things hold for the processing of RNAs that are not set up to make proteins). A full development of the findings that lie behind this description can be mobilized that gene expression and protein synthesis are not mechanisms, but shorthand levels for a branching congeries of mechanisms that operate, contingently, in ways that no right-minded mechanist would want to count as a mechanism.
All the machinery inside cells is constructed using proteins encoded in DNA. But many of the machines built from those proteins made in different cells (e.g., hemoglobin traps for oxygen) are cell-specific and regulated by cellular contents. Some of the machines already present in those cells affecting transcription, RNA processing, and assembly of proteins deployed in transcribing DNA are cell-specific –e.g., those affecting accessibility of DNA (cf. histone packaging), those affecting the start and stop points of transcription, those affecting RNA editing, some of those affecting RNA processing in the cytoplasm (cf. RNA-induced silencing complexes, charged with particular microRNAs, which detect mRNAs which will then either be cut apart or made less likely to be translated until a releasing signal is received). These devices, in turn, are affected by signals affecting those cells, the availability of energy sources within the cells, and the regulatory molecules that they contain at the relevant times. Thus, the contents of cells depend on the history of gradients of signal molecules within and outside of cells, on cell adhesion molecules, on cell-cell signaling, cell movements, etc. Indeed, the history of a cell and its ancestors within the cell’s lineage – particularly including the effects of organism-external and organism-internal signals that have affected the ancestors as well as the cell itself – are enormously important in determining the specific machines and settings of the machines present in that cell. This explains the fact most kinds of cells cannot be transformed backwards to assume the earlier fates of cells in their lineage except by extreme biotechnological tricks.
Thus, there is no fixed genetic or developmental program in the fertilized egg; the entire contents of the egg are not sufficient to explain the contents of particular cells or the distribution of cells in the resulting organism. The switches that determine cell fate and content are thrown in sequential series, many of which depend on cell-cell or cell-environment interactions. And different kinds of cells contain, literally, different machines that regulate the different products they make. In this way, the organism is constructed “on the fly” without central control, but with local controls that are distributed in regular ways because the early cells, and then their successors, are (re)constructed in ways that draw on the relevant regularities of the (external and internal) environments as well as on the available RNA and DNA molecules – regularities that include gravity, and cyclicities in the light regimes, sonic regimes, ambient pressures, chemical regimes, plus relations to neighboring cells and interactions with the cells and signals of other organisms. The regularities of the environment and of cell contacts as cells move and emit signals are as much a part of the developmental program as the actual contents of the fertilized egg. It follows that although the generalities of protein synthesis are similar from cell to cell, the mechanisms within different cells are different. This, in turn, suggests that genetic and developmental programs – interpreted as programs that can be localized in fertilized eggs – are not sufficient to explain the structures and properties of organisms. But it also yields a critique of straightforward accounts of the regularities of development in terms of mechanisms bringing about development. I conclude from this is that the notion of biological information – the information utilized in constructing the organism, that guarantees, e.g., that quadrupeds will pass through a stage that involves the induction of four limbs even if they grow up without limbs – is much richer than sequence information. It draws on cellular histories as discussed above.
My argument suggests that the regularities that determine what proteins are synthesized in eukaryotes cannot be explained by an account resting only on the machinery present in the fertilized egg. The cellular machines that influence an organism’s development are themselves altered, step-by-step, by interactions among cells and between the organism and the external environment, with the result that different cells synthesize different proteins from the same coding sequences, i.e., that sequence information is processed in different ways than the Watson-Crick theory of sequence information and Shannon and Weaver information theory applied to DNA molecules as signal sources would have us believe. Couched in terms of machinery, this is because the machinery in different cells is structured and set differently so that the readouts achieved from the same initial information are different in different cells. *
*I should mention that Benjamin Jantzen of Carnegie Mellon University (and soon to join the faculty at Virginia Tech) has a paper in progress that argues that (in spite of all these complications?) the existence of the genetic code shows that an information-theoretic analysis of coding relations should work. I will be interested whether his arguments undermine mine, mine undermine his, or, perhaps, whether the complications I have examined can be set aside as secondary and not fundamental to the issues at stake.