First check - look for anom peaks.
If your S atoms are showing up in the Dano map, them you can check
whether this is a compound with S in it. crystallisation and
cro-protectants contain a wealth of small molecules which often bind
Eleanor
9 sigma peaks are rarely due to multiple site waters..
Eleanor
Borhani, David wrote:
Maria, 1.7 A is too long for common diatomic molecules (N2, 1.1 A; O2, 1.2 A;
CO, 1.1 A) or molecules like hydrogen peroxide (~1.48 A), methanol (ditto), etc.
Does the "water" have unusually low temp. factors (i.e., it's really a much
heavier atom, and the peak you see is a Fourier ripple)?
It may be indeed that you have (one) water at two alternate positions, but I
think you need to reset to nothing modeled to be certain of density
interpretaion moving forward.
Thus, try removing the water you have already built, any surrounding waters or
other ligands, and also the protein atoms to which all these are hydrogen
bonded (set occ = 0), refine a few cycles, and then look at the diff. maps.
Dave
David Borhani, Ph.D.
D. E. Shaw Research, LLC
120 West Forty-Fifth Street, 39th Floor
New York, NY 10036
[EMAIL PROTECTED]
212-478-0698
http://www.deshawresearch.com
-----Original Message-----
From: CCP4 bulletin board [mailto:[EMAIL PROTECTED] On
Behalf Of Maria Håkansson
Sent: Friday, September 19, 2008 8:00 AM
To: CCP4BB@JISCMAIL.AC.UK
Subject: Re: [ccp4bb] Unexplained electron density
Hello David and others,
Thanks for yur comments.
I guess it might be as simple as water molecules, present in the
structure but not at the same time.
The density looks like a rod with uneven distribution. Both ends of
the rods (1.7-1.8 Å in between)
make hydrogen bonds to protein or other water molecules - normal
distances (2.3-3.3 Å). Could it be
strong water molecules but with partial occupancy meaning that both
sites are occupied but not at the same time?
I guess refmac automatically refines the molecules that way
although I
have not specified it in my file. So after
refinement as too close water molecules there is no clash just nice
density. However I assume it is appropriate to
specify these water molecules as the same water but with an
alternative conformation in the pdb file.
Best Regards,
Maria
19 Sep 2008 kl. 11:10 skrev David Briggs:
Hi Maria,
Initial questions:
1) What's present in crystallisation/purification buffers?
2) Are any other ligand visible for the 9sigma peak?
3) Does the 9 sigma peak also have a peak in an anomalous
difference
map?
Assuming 1.7A is accurate (and with 1.5A resolution, you'd hope it
would be!) a metal-ion - water interaction is looking less likely.
Take a look at Marjorie Harding's papers and website for
target metal
ion - ligand distances, and the closest you'll see is water:Mg2+, at
2.07.
http://tanna.bch.ed.ac.uk/newtargs_06.html
So one would assume it is a covalently bonded compound.
So back to question 1.
What's in your buffers?
A quick search for bond length tables suggests Carbon-Sulphur (1.8)
and Carbon-Chlorine at 1.7.
hope this helps
David
2008/9/19 Maria Håkansson <[EMAIL PROTECTED]>:
Hello All,
I have a problem with a 9 sigma positive peak 1.7 Å away
from a water
molecule (or what I believe is
a water molecule). There are several similar peaks in my
map though
only one
is as high as 9 sigma.
My first thought was to exclude these too close waters.
However the
R-values
increased by more than 0.5 %. Could it
be carbonmonoxide or oxygen atoms? By the way, It is 1.5 Å
resolution data.
Any suggestions?
Best Regards,
Maria Håkansson
Maria Håkansson, Ph.D.
tel: +46 (0) 76 8585 706
Senior Scientist, Max-lab, Lund University
fax: +46 46 222 47 10
Ole Römers väg 1 (P.O. Box 188)
www.maxlab.lu.se
SE-221 00 Lund, Sweden
[EMAIL PROTECTED]
--
============================
David C. Briggs PhD
Father & Crystallographer
http://drdavidcbriggs.googlepages.com/home
AIM ID: dbassophile
============================
Maria Håkansson, Ph.D.
tel: +46 (0) 76 8585 706
Senior Scientist, SARomics Biostructures AB fax: +46 46 19 12 77
Scheelevägen 22 (P.O. Box 724)
www.saromicsbiostructures.com
SE-220 07 Lund, Sweden
[EMAIL PROTECTED]
