That would be a partial solution to the alpha-L-fucose problem, except
for the fact that currently, the monomer library description for
alpha-L-fucose is a beta! So, unless one is extremely vigilant, they
will get beta fucoses, which I believe do not exist in N-linked sugars.
Currently, another solution is to modify the monomer library to have the
correct alpha fucose (which is a must), and then define a bond and a few
angle restraints (and possibly torsion), in phenix. Another one is to
ask phenix to use a beta linkage, which imposes an alpha linkage on
L-fucose! Apparently, two wrongs do make a right.
The real problem is that every crystallographer I asked tells me that
they use these linkages, without giving much of a thought. This is not
surprising. Figuring out if an anomer is alpha or beta required me to
spend a couple of hours digging up my organic chem notes, and reading
several sections from an IUPAC publication (So, I would appreciate
further confirmation of this).
The real solution is to have another set of linkages for L sugars, in
refmac and phenix. This will not work for every possible sugar, but
should for the standard/common ones.
Engin
On 1/18/10 2:03 PM, Ethan Merritt wrote:
On Monday 18 January 2010, Engin Ozkan wrote:
Hi everybody,
I have a question regarding glycosidic bonds. This relates to refmac,
phenix and cns, so I thought the best forum to pose this was here.
We have these very nifty link descriptions, such as BETA1-4, ALPHA1-6,
etc. that come with refmac and phenix. These essentially describe a
chiral center and torsion angles around position C1. However, the way a
glycosidic linkage is defined as alpha or beta does not solely depend on
the C1 chiral center (see below if interested, and see if I am right).
An ALPHA1-3 link in refmac or phenix works for a alpha-D-mannose, but it
forces my alpha-L-fucoses to become beta, no matter what I do.
I do not know what options are available in phenix, but in the refmac
dictionaries you have the option of specifying the chiral volume of the
anomeric carbon as "both" rather than as "negativ" or "postive".
This will preserve the chirality of the model as input.
So long as you provide the correct anomer in your input model,
you should be OK. Caveat: I have used this successfully for to handle
anomers of GLU, GAL, NAG, and SIA, but I can't rule out a bug of some sort
that hits other sugars.
Ethan
And by
the way, secreted and membrane proteins you make in insect cells will
have ALPHA1-3 and ALPHA1-6 linkages to fucose, and in mammals, ALPHA1-6
to fucose; this should be a very common occurrence (and problem).
This seems to be a fucose-specific problem, since it is the only
standard sugar residue, that's an L sugar, so the defined links will
result in incorrect stereoisomers.
I would be very happy if anyone can check the logic here, and please
correct me (it has been 12 years since I learnt and soon forgot what an
anomer is!).
Best,
Engin
P.S. Alpha or beta: How a sugar is designated as an alpha or beta
"anomer" is actually complicated, and requires one to draw a Fischer
projection. IUPAC says:
Relative stereodescriptors used in carbohydrate nomenclature to describe
the configuration at the anomeric carbon by relating it to the anomeric
reference atom. For simple cases the anomeric reference atom is the same
as the configurational reference atom. Thus in ?-d-glucopyranose the
reference atom is C-5 and the OH at C-1 is on the same side as the OH at
C-5 in the Fischer projection.
Simply checking wikipedia:anomer can show that an alpha or beta anomer
can have opposite chiral centers depending on the identity of the sugar.
Also, the current versions of FUC-a-L in the monomer libraries of refmac
and phenix seem to be beta indeed. Garib Murshudov knows about this. See
http://www.flickr.com/photos/46544...@n03/4274285327/
HIC-UP has it right: http://xray.bmc.uu.se/hicup/FUC/
--
Engin Özkan
Post-doctoral Scholar
Laboratory of K. Christopher Garcia
Howard Hughes Medical Institute
Dept of Molecular and Cellular Physiology
279 Campus Drive, Beckman Center B173
Stanford School of Medicine
Stanford, CA 94305
ph: (650)-498-7111
