Macromolecular intelligence in microorganisms

Biochemistry and molecular biology have been focusing on the structural, ca talytic, and regulatory properties of individual macromolecules from the pe rspective of clarifying the mechanisms of metabolism and gene expression. C omplete genomes of 'primitive' living organisms seem to be substantially la rger than necessary for metabolism and gene expression alone. This is in li ne with the findings of silent phenotypes for supposedly important genes, a pparent redundancy of functions, and variegated networks of signal transduc tion and transcription factors. Here we propose that evolutionary optimizat ion has been much more intensive than to lead to the bare minima necessary for autonomous life. Much more complex organisms prevail. Much of this comp lexity arises in the nonlinear interactions between cellular macromolecules and in subtle differences between paralogs (isoenzymes). The complexity ca n only be understood when analyzed quantitatively, for which quantitative e xperimentation is needed in living systems that are as simple and manipulat able as possible, yet complex in the above sense. We illustrate this for th e glutamine synthetase cascade in Escherichia coli. By reviewing recent mol ecular findings, we show that this cascade is much more complex than necess ary for simple regulation of ammonia assimilation. Simulations suggest that the function of this complexity may lie in quasi-intelligent behavior, inc luding conditioning and learning.