Protonation state summary
I've gotten a number of usefull responses. As was expected the
response turned to determining the pKa's of the AA's in the protein (I
didn't specifically ask this question because I wanted to see if there
where other answers not anticipated). Two other references from Bill
Welsh's summary from May 23 are also usefull. Rob Setlik recommends:
Bashford et al., (1993) Biochem 32, 8045.
Yang et al., (1993) Protein 15, 252-265.
Both of these are good papers. The Bashford et al. lists an ftp site
ftp.scripps.edu in pub/electrostatics) to obtain the programs and data (Is
this site possible a David Case piece of work, could be :) ). I've been
there and there is an alpha version of a the program MEAD. I haven't
unpacked this yet so I can't tell you much other than it exist.
Thanks for the help.
Untill next time,
Wright State University
Department of Biochemistry
Dayton, OH 45435
e-mail: cletner[ AT ]remcure.bmb.wright.edu
----------------- response summary -----------------------------------
From: "Dr. Walter J. Stevens" <walt[ AT ]ibm1.carb.nist.gov>
Quite a bit of work has been done on ionization
states of amino acid side chains in proteins.
See for example:
M.K. Gilson, Proteins: Structure, Function and Genetics
-15-, 266-282 (1993)
J. Antosiewocz, J.A. McCammon, and M.K. Gilson
J. Molec. Biol. -238-, 415-436 (1994)
and references therein.
This is a good review of the complexity of
the problem you have addressed.
Dr. Walter J. Stevens
Center for Advanced Research in Biotechnology
National Institute of Standards and Technology
and University of Maryland Biotechnology Institute
9600 Gudelsky Drive
Rockville, MD 20850
Phone: (301) 738-6264
FAX : (301) 738-6255
E-MAIL: walt[ AT ]ibm1.carb.nist.gov
From: "Jeffrey L. Nauss" <nauss[ AT ]ucmod2.che.uc.EDU>
I'm running into the same situation myself. The rule of thumb I am
using is that if a protein is accessible to the solvent it is
protonated as appropriate for the pH of the system under study.
Fortunately, I am looking at systems at pH 7.0 or at low pH levels
(3-4). Therefore, I feel I can still go for an all or nothing
protonation for a given type of amino acid (i.e. all Glu will be
protonated at the low pH). And, I do not have to worry about subtle
variations in pKa due to local environments.
As for salt bridges, the residues involved are charged. Residues at
the active site present a different problem. There subtle pKa changes
may have dramatic effects. Frankly, I am still a bit baffled as to
how to handle them, especially one in my system of interest with a
solvent accessible surface of only 1 Angstrom**2.
I hope this rambling will be of some benefit to you. I am interested
in your summary.
* UU UU Jeffrey L. Nauss, PhD *
* UU UU Director, Molecular Modeling Services *
* UU UU Department of Chemistry *
* UU UU CCCCCCC University of Cincinnati *
* UU UU CCCCCCCC Cincinnati, OH 45221-0172 *
* UUUU CC *
* CC Telephone: 513-556-0148 Fax: 513-556-9239 *
* CC *
* CCCCCCCC e-mail: nauss[ AT ]ucmod2.che.uc.edu
* CCCCCCC *
From: "Jim Briggs, Ph.D., U of Houston,
Chem. (713)743-3315" <BRIGGS[ AT ]kitten.chem.uh.edu>
I can't speak to your main question of the importance of having
correct protonation states except that if you get one wrong in a
critical area, you will likely see a large effect.
I did want to respond to your secondary question of assigning
protonation states to AAs in proteins. Using the approach that
you suggest (i.e. the "null" model) should work most of the time.
Assuming that residues that are on the surface are appropriately
ionized depending on the pH and that internal residues are ionized
if in "polar" environments (again, depending on pH). Obviously, you
need to take a look at ALL histidines...
Anyhow, there is an approach that was developed recently to
predict ionization states depending on the pH, ionic strength, and
electrostatic environment of each AA. The approach uses the
electrostatics module of the UHBD (University of Houston Brownian
Dynamics) program. The procedure and results are described in a
recent paper: J. Antosiewicz, A. McCammon, M. Gilson J. Mol. Biol.
1994, 238, 415-436. In most cases, this approach does better than
the "null" model as compared with experimentally determined
ionization states. The UHBD manual can be downloaded for free from
the anonymous ftp site: 188.8.131.52 under: pub/uhbd in DVI and
PostScript forms. The code is available for a nominal fee to
academic institutions and is distributed commercially through
Molecular Simulations, Inc.
I'm sure that a similar approach is available in conjunction
with Barry Honig's DELPHI program...
JBriggs[ AT ]uh.edu
From: Soaring Bear<bear[ AT ]ellington.pharm.Arizona.EDU>
I'm glad you posed that question because I've been
wondering the same thing about drugs which contain amines
which can equilibrate between both protonated & unprotenated
states at physiological pH (as can Histidine). The grunt
way of handling it would be to run the model each way
seperatly but it would seem this is a common enough situation
that modeling software ought to incorporate handling of this
as a feature.
I look foward to seeing the responses you get.
* UU UU SOARING BEAR *
* UU UU Computer Aided Drug Design *
* UU UU A University of Arizona, New Pharmacy 404 *
* UUUU AAA Tucson, AZ 85721 602-626-6133 *
* AA AA where the sun shines *
* AAAAAAA *
* AA AA e-mail:bear[ AT ]ellington.pharm.arizona.edu *
From: Xiaoou Xiang <xiang[ AT ]auriga.rose.brandeis.edu>
The only way to make certain about the protonation state is, of course, by
getting hands wet with test tubes.
The computational way to do that, not always reliable, is to locally minimize
with the side chain under consideration in all protonation states, and then
choose the one in which the side chain has the lowest energy of interaction
with the local environment.
In practice, one often only needs to ascertain the protonation state of
residues in the region of interest in the protein, and assign arbitrary
protonation states to other residues, assuming that the influence of
(de)protonation does not propagate beyond 8-10 anstroems.
Phil X. Xiang
From: Paul Beroza <ppb[ AT ]coeus.ucsd.edu>
People have been worried about this question, which usually
falls under the topic, how do you determine an amino acid's pKa
when it is bound to a protein?
Here are a few references that I have at my fingertips.
(pberoza[ AT ]ucsd.edu)
Russel and Warshel, Journal of Molecular Biology, 185:389 (1985).
"The Energetics of Ionized Groups in Bovine
Pancreatic Trypsin Inhibitor"
Bashford and Karplus, Biochemistry, 29: 10219 (1990),
"pKa's of Ionizable Groups in Proteins: Atomic Detail
from a Continuum Model"
Beroza, et al., Proc. Natl. Acad. Sci., 88:5804 (1991),
"Protonation of Interacting Residues in a Protein by a Monte Carlo
Method: Application to Lysozyme and the Photosynthetic Reaction Center
of Rhodobacter sphaeroides"
Yang, et al., Proteins, 15:252 (1993),
"On the Calculation of pKa's" in Proteins"