What is the SMALLEST DNA BINDING DOMAIN?

L.A. Moran lamoran at gpu.utcc.utoronto.ca
Tue Jan 14 09:49:42 EST 1997


In article <5bcck5$ho9 at nntp3.u.washington.edu>,
Angeline Kantola <kantola at u.washington.edu> wrote:
>In article <mthompson-1001971311280001 at phx-ts2-16.goodnet.com>,
>MThompson <mthompson at asu.edu> wrote:
>>I am searching for a small independently folding (or in presence of a
>>metal ion), DNA binding domain. Would like it to be sequence specific, but
>>would greatly appreciate all suggestions. Be as detailed as possible or
>>please cite a reference, or name that I can locate it in the literature.

This request brings up another issue that might be worth discussing. Should
these DNA binding regions be referred to as "domains" or "motifs"? In the
literature on protein structure the word "domain" usually refers to a
discrete, independently folding region with a globular structure. Such
domains consist of large stretches of primary sequence and they can be
recognized as distinct units in the overall structure of a protein. Domains
are usually connected by short linker sequences.

Domains, as defined by structural biologists, consist of multiple regions
of secondary or supersecondary structures. The supersecondary structures
are often referred to as "motifs". Examples of motifs include helix-loop-
helix regions, Greek keys, and beta-alpha-beta units. Some of these
supersecondary structures form the core of a DNA binding region and that's
why most textbooks refer to them as DNA binding "motifs". Examples are
the helix-turn-helix motif and the zinc finger motif. A counter-example
is the homeobox "domain"; a definition that is not consistent with the
terminology used by those who study protein structure.

It's not clear to me what MThompson is looking for. If it's really a
distinct "domain" then most of those who responded have not been helpful
since they have concentrated on "motifs". However, perhaps MThompson
really wanted an example of such a motif?

I realize that this might seem a bit picky but I think that it is important
to be consistent with terminology and I'm pleased to see that even leading
molecular biology textbooks have adopted the word "motif" to describe these
DNA binding regions of protein. This makes the biochemists happy.  (-:

>Cys2-His2 zinc fingers are on the order of 30 amino acids long. As the
>name suggests, a zinc atom is critical for structural integrity of the
>functional domain. Proteins which bind DNA via C2H2 zinc fingers do so
>sequence specifically.  There's no evidence that single zinc fingers bind
>nucleic acids, though. A minimum of three of these fingers are required to
>bind or, as in the case of yeast ADR1, two fingers and a non-finger N
>termial region.

Zinc fingers and other zinc containing structures are examples of motifs
and not, strictly speaking, domains. They are often part of a large domain
within a protein. Motifs are usually short stretches of primary sequence
on the order of 20-100 amino acid residues. Domains are larger.

>C2H2 zinc fingers are everywhere (there are other kinds of zinc fingers as
>well), so finding information about them should be easy. Mia Schmiedeskamp
>and Rachel Klevit wrote a review of zinc fingers a few years back, which
>is probably a good place to start. 

Some people are very sensitive about the naming of zinc containing structures.
The C2H2 structure is called a zinc finger but the others are not referred
to as "fingers". It is probably useful to maintain this distinction.


Larry Moran






More information about the Proteins mailing list