Series: Mathematics Colloquium

Date: Thursday, February 7, 2002

Time: 4:30 - 5:30 PM

Place: 102 McAllister Building

Host: Howie Weiss

Refreshments: 4:00 - 4:30 PM, in 212 McAllister

Speaker: Thomas D. Schneider, National Cancer Institute

Title: Flippers, Flappers and Flip-Flops in DNA Binding

Abstract: 

The mathematics of information theory provides powerful tools for
cracking problems in molecular biology.  For example, the packing of
spheres in high dimensional spaces is directly related to molecular
states.  In this talk I will show how we used information theory to
understand two kinds of genetic control system.

Genetic control is often done in a simple way: a protein binds to a
spot near the start of a gene and physically prevents the gene from
being used (that is, it blocks the DNA from being transcribed into
RNA).  The amount of pattern in the DNA can be measured in bits of
information, and this reveals that the protein shape molds to the
surface of the DNA during evolution.  DNA has major (ie wide) and
minor (ie narrow) grooves and proteins can get more information from
the major groove (2 bits) than they can from the minor groove (1 bit).
This can be seen clearly using information theory.

In the first genetic system, a virus called P1 replicates its DNA in
the intestinal bacterium Escherichia coli by using a control protein
called RepA.  Surprisingly, RepA violates the idea given above.  When
a theory appears to fail it can teach us new things.  In this case
RepA probably is flipping a base out of the DNA (swivelling half of a
DNA ladder rung out of the ladder) in order to start the opening of
DNA for DNA replication.  Strong experimental evidence was found to
support this hypothesis.

In the second genetic system, we discovered that the Fis protein
(which controls all kinds of stuff in E. coli) frequently uses pairs
of sites 7 or 11 base pairs apart on DNA.  Two overlapping Fis sites
separated by 11 base pairs are found in the E. coli origin of
chromosomal replication (the place where replication begins).  We
found that only one of the two overlapping Fis sites is bound by Fis
at a time, so the structure is a molecular flip-flop. Since the two
sites are precisely positioned between two DnaA sites, and these
determine the orientation of the DnaB helicase, we suggest that the
flip-flop directs alternative firing of replication complexes in
opposite directions.  Since they can implement Boolean logic,
molecular flipflops could be used to build molecular computers.  Both
of these discoveries were made only because of the clarity that
information theory provides.