Although I am not a crystallographer I have some comments on the posting by
Dr. Murphy. For several years I have been involved in molecular modeling of
such 7TM or G-protein coupled receptors (GPCRs).
The fact that we currently don't have any published experimental structure of
GPCRs from the neural system does not necessary mean that it does not exist.
With the enormous commercial potential of such structure for e.g. a drug company,
one can imagine the temptation of not sharing the knowledge with the scientific
community. For the time being we must assume that the closest we can get
is the bovine rhodopsin projection map (Schertler et al. Nature 1993, 362:770).
In contrast to bacterial rhodopsin, bovine rhodopsin couple to G-proteins.
Dr. Murphy says that "bacterial-rhodopsin is composed of a concentric ring of
sevel alpha-helical membrane spanning domains". What do you mean by that? What is
A concentric ring? If you mean concentric ringS, which ringS do you mean? The
alpha-helices can be imagined to form relatively rigid rods penetrating the
membrane. These rods are arranged in an oval kind of arrangement such that there
will be a channel/pore into the membrane.
There exist several good quality x-ray structures of e.g. photosynthetic
reaction centers, having the membrane spanning segments in alpha-helical
conformation. Also several porin crystal structures have been published
in which the membrane spanning domains are beta-sheets.
According to the work of Gunnar von Heijne, the insertion of protein into a
membrane is governed by topogenic signals of the amino acid sequence, and
each segment is inserted into the membrane independent of the others (in
one protein), and finally assembled into the functional structure. On the
other hand if the amino acid sequence tells us that the helices are amphi-
patic, I'd think that such helices would not be very happy to be entirely
surrounded by membrane. In the functional form of these proteins, one face
of the helices is oriented towards the membrane and the other face partly
towards the closest helices and the central pore which is probably filled
with water and ions and..... In addition, many of the membrane spanning
helices of membrane proteins with known and unknow 3D structure shows very
weak amphipatic character.
I really don't think that if we could obtain a crystal structure of each
membrane spanning segment, that it would help much in understanding the
structure and function of such membrane proteins. In my opinion much more
important issues are:
1) Which parts of each membrane spanning segment is facing the central
channel/pore? (i.e. relative rotation around each helix axis)
2) Exactly where is the amino acid sequence is the start and end of the
membrane spanning segments? (i.e. translation of potential membrane spanning
alpha helices through the membrane)
3) Is the overall organization of the membrane spanning segments the same
or similar or different for the whole class of proteins?
A large number of 3D models of such proteins have been created, and have
proved to be quite useful in various ways. Many such models have e.g.
predicted the importance of certain residues for binding of certain
For a discussion of various aspects of structure and modeling of such
receptors I can recommend (Bias ON) the paper Dahl & Edvardsen, Molecular
Modeling of dopamine receptors. In Dopamine Receptors and Transporters.
Pharmacology, structure and function. Ed. H.B. Niznik, Marcel Dekker, Inc.
1994, pp 265-282. I can also recommend the compilation of literature
references available from the GCRDb server on
I'd suggest that this discussion may be continued in
| /| _| |_ _| __ _ _| _ _
|/_| \/ \/ | |\| |_| |_ |_| \/ |_|_ | |_| _\ |_' |\|
Dept. of Pharmacology, IMB |
University of Tromsoe | TelePhone: +47 77 64 53 42
MH, Breivika | TeleFax: +47 77 64 53 10
N-9037 TROMSOE | Email: edvard at fagmed.uit.no
NORWAY | URL: http://atf1.fagmed.uit.no/mgl.html