RULES FOR THE EVOLUTION OF GENE CIRCUITRYM.A. SAVAGEAUDepartment of Microbiology & Immunology, The University of Michigan MedicalSchool, Ann Arbor, Michigan 48109-0629 USACells possess the genes required for growth and function in a variety of contexts. In anygiven context there is a corresponding pattern of gene expression in which some genes areOFF and others ON. The ability of cells to switch genes ON and OFF in a coordinate fashionto produce the required patterns of expression is the fundamental basis for complexprocesses like normal development and pathogenesis. The molecular study of generegulation has revealed a plethora of mechanisms and circuitry that have evolved to performwhat appears to be the same switching function. To some this implies the absence of rules.However, simple rules capable of relating molecular design to the natural environment havebegun to emerge through the analysis of elementary gene circuits. Two of these rules arereviewed in this paper. These simple rules have the ability to unify understanding acrossseveral different levels of biological organization -- molecular, physiological,developmental, ecological.1. IntroductionRegulation of gene expression and its systemic manifestations are subjects ofintense study. As a result of this effort we shall soon have identified all of thegenes and proteins for a number of simpler organisms. Despite this enormousprogress we are still at a loss to understand the integrated behavior of the organism.Our understanding is still fragmentary and descriptive. We are unable to predictchanges in the organism’s behavior when it is placed in a novel environment orwhen a change is made in one of its genes. Little is known about the forces thatlead to the selection or maintenance of a specific mechanism for the regulation of agiven set of genes in a particular organism. Is this process random, or is itgoverned by rules? The answer to this question is important. It will help us tounderstand the evolution of gene regulation; it also will help us to develop judiciousmethods of redirecting normal expression for biotechnological purposes or ofcorrecting pathological expression for therapeutic purposes.Our goal is to understand the integrated structure and function oforganizationally complex systems in relation to their underlying moleculardeterminants. Moreover, we are particularly interested in identifying the rule-likeproperties of these systems that would allow for some algorithmic compression intheir representation, and not simply a compilation of all the molecular details.In pursuit of this goal we have developed a canonical nonlinear formalism thathas desirable properties for the representation and analysis of organizationallycomplex systems (1). This formalism has been used to characterize alternativemodes of gene control and various forms of coupling among elementary genecircuits. The results allow us to identify a set of rules, or design principles, thatgovern the natural selection of gene circuits. Here we shall review the relevantbiological background and then present results from our analysis of gene circuitry.2. Biological BackgroundThe common metaphor of the genome as a blueprint for construction of theorganism masks the difficult task of relating structure and function of the intactorganism to its underlying genetic determinants (2). The behavior of an intactbiological system can seldom be related directly to its underlying moleculardeterminants. There are several different levels of hierarchical organization that arerelevant. For our present purposes it will be sufficient to consider four differentlevels -- genome sequence, transcriptional unit, elementary gene circuit,environmental context.2.1. The DNA sequence constitutes the genomeThe recent sequencing of the complete genome for a number of simpler organisms,and the projected completion of the sequence for the human genome by the year2005, illustrate the power of modern molecular biology to resolve complex systemsinto their simplest elements. The four bases -- A, T, G, and C -- are strungtogether in sequences that are mind-numbing in their simplicity; yet, thesesequences provide the potential for incredible complexity. Whether it be theversatile metabolism of free-living microbes that can adapt to nearly anyenvironment, or the sophisticated structures of multicellular organisms that can beseen in near endless variety, the physical basis for this complexity is the context-dependent expression of the organism’s genome.2.2. Information is encoded in transcriptional unitsThe mapping from DNA level to organismal level requires a deeper understanding ofhow information is encoded in the genome. DNA sequences are organized intofunctional units that consist of structural genes flanked by a start sequence at whichtranscription begins and a termination sequence at which it ends. In addition, thereare a number of regulatory sites capable of binding specific transcription factors thatinteract with the transcription machinery to modulate the rate of transcriptioninitiation or termination (Fig. 1).M1M2G1G2TPR1R2Figure 1. Unit of transcription. Structural genes (G1 and G2) are bounded by apromoter sequence (P) and a terminator sequence (T), and preceded by upstreammodulator sites (M1 and M2) that bind regulators (R1 and R2) capable of alteringtranscription initiation. The solid arrow represents the mRNA transcript and thescalloped lines indicate the protein products encoded by genes G1 and G2.2.3. Expression is organized into elementary gene circuitsTranscription of DNA is but one step in a cascade of information flow thatconstitutes the expression of a gene (Fig. 2). Each stage of such a cascade is apotential site at which expression can be regulated in a context-dependent fashion.The context is provided by the life cycle of the organism, and the interlockingmechanisms of gene regulation interpret that context.RegulatormRNA-R mRNA-EEnzymeMetaboliteFigure 2. Cascade of information flow from DNA to RNA to protein to metabolite.The processes of synthesis and degradation are represented by horizontal arrows,whereas the catalytic and regulatory influences are represented by vertical arrows.An effector circuit is shown on right and a regulator circuit is shown on the left.2.4. Physiology and ecology are reflected in the organism’s life cycleThe life cycle of some organisms is largely programmed development from egg toembryo to