Fall 2003 Catalyst

Two Approaches
to Gene Regulation

by John W. Little, PhD

We are interested in the process of gene regulation — how cells turn their genes on and off. We study this process in bacteria and their viruses. These model organisms are easy to study: they grow rapidly, and are good systems for both our favorite approaches, genetics and biochemistry. In addition, they show interesting gene regulatory behaviors, as described below.

We take two diverse approaches to studying gene regulation. One, a standard approach for biochemists, is often called “reductionist”. In this approach, a system is analyzed in ever finer and finer detail, focusing progressively on mechanisms by which the regulation works. The other approach, a relatively new one for us and one that is becoming trendy throughout biology, is often termed “systems behavior”. In this approach, one looks at the behavior of the overall system, seeking to understand and predict how the interactions among its components lead to the overall observed behavior of the system. A familiar example is the operation of a thermostat to maintain a constant temperature in a home.

Mechanisms of gene regulation — the SOS regulatory system

This bacterial regulatory system controls the cellular response to treatments that damage DNA. It is controlled by two proteins — a “repressor” called LexA that turns off a group of about 40 “SOS genes” during normal cell growth, and a second protein called RecA that is activated when the cellular DNA is damaged. Activated RecA inactivates LexA by stimulating a specific proteolytic cleavage of LexA; this reaction cuts LexA into two fragments, which are no longer active as a repressor. This turns on the SOS genes for as long as the RecA remains activated.

Our major contribution to this field has been to figure out how LexA cleavage works. RecA, it turns out, acts indirectly to stimulate a latent self-cleavage activity of LexA. That is, LexA has a built-in self-destruction capability, but this is normally held in check as described below. Our evidence shows that LexA has an active site, like any enzyme, which contains a binding pocket for the substrate and a catalytic center to carry out the chemistry of cleavage at a specific bond. Although we do not yet understand how RecA stimulates LexA self-cleavage, the following structural evidence provides a tempting and testable model.

We recently determined the crystal structures of several forms of LexA, together with colleagues at the University of British Columbia (Vancouver, Canada). These structures support a model to help us understand why LexA cleavage does not occur in an uncontrolled way. We see two forms of the LexA molecule (see Figure 1): A “cleavable” (C) form, in which the cleavage site is positioned in the active site in a way that allows rapid cleavage of the protein to occur, and a “non-cleavable” (NC) form, in which the cleavage site lies distant from the active site. We believe that the latter form normally predominates in the cell, preventing cleavage, and that RecA acts in some way to stabilize the cleavable conformation, perhaps by binding tightly to it. This would greatly speed the rate of cleavage. To help understand how RecA acts, we are currently trying to determine where RecA binds on the LexA molecule, partly in collaboration with Prof. Vicki Wysocki (Department of Chemistry).

Fig. 1: Two forms of LexA. The active site and cleavage site are depicted
in white; the red portion of the molecule moves into the active site in the form on the right.

Systems behavior of phageλ

Complex gene regulatory circuits have many interlocking components, each of which influences the action of several others. These interactions give rise to “systems behavior” such as negative feedback (like a thermostat), positive feedback (like a microphone in front of a speaker), and non-linearity. These factors make the operation of such circuits very difficult to intuit. Phageλ is one of the best-understood regulatory circuits: Most or all of the components are known; their interactions are understood; and the consequences of these interactions are known. Importantly, however, λ has all of the systems behaviors mentioned above, making it difficult to predict the behavior of mutants (which alter components and/or their interactions). λ also has several other “systems” properties. Its genes can be expressed in either of two stable patterns, making a so-called “bistable switch”; it can switch from one stable state to the other (the “genetic switch”); and the genetic switch has “threshold behavior” (Fig. 2), that is, it responds poorly to a low level of stimulus, but at a particular set-point, response abruptly becomes efficient.

Graphic of threshold behavior of the phage lambda genetic switch

Fig. 2: Threshold behavior of the phage lambda genetic switch. At higher stimulus,
the switch becomes efficient (as detected by an increase in the output).

We are studying the λ circuitry in three different ways: First, we are removing one or several of the components that are believed to make the circuitry work properly; second, we are analyzing systems properties, such as threshold behavior, to determine how these operate and how their behavior is determined; and finally, we are carrying out computer simulation of the λ circuit, in collaboration with Prof. Adam Arkin (U.C. Berkeley) with the goal of making a computer model that can predict the properties of mutants in the circuitry.

When we remove several different components of the circuitry that the textbooks say are important, we find that we can compensate for their loss by other changes in the system. These findings suggest that complex systems evolved by first making a simple ground plan, and then elaborating this in a later stage by adding refinements that gave more optimal behavior. In this view, the components we have removed are such refinements. These studies also show that the circuitry is "robust" — it can tolerate quantitative or qualitative changes in the components without destroying its operation. We believe that robustness is also important in the evolution of complex circuitry. We can isolate mutants that alter the set-point for the threshold behavior; these mutations alter one of several components. These studies help show how the set-point is controlled, and support the idea that this set-point has evolved to an optimal value. Finally, our computer simulations, still in progress, have developed a model that has bistable behavior; this behavior is altered by changes that we know experimentally to affect the operation of the circuitry. We hope soon to add the genetic switch to this model.

Biological Sciences West
P.O. Box 210088 ·Tucson, AZ 85721-0088
Tel: (520) 621-9185 FAX (520) 621-9288
Department of Biochemistry and Molecular Biophysics
The University of Arizona

Two Approaches to Gene Regulation

Mechanisms of gene regulation — the SOS regulatory system

Systems behavior of phageλ

Two Approaches
to Gene Regulation