*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Near as I can tell, synchrotron radiation is the only way to get
site-specific metal identification information. It's not overkill, it
is essential.
What you do is collect data sets at the Zn absorption edge and just
below (lower photon energy) the edge. The phased difference Fourier of
these two data sets will light up all Zn positions (and ONLY the Zn
positions). You can also do a phased anomalous difference Fourier with
the data set at the Zn edge, but this is not a conclusive test. This
will light up Zn positions, but it will also light up any other
anomalous scatterers, such as Ni. The magnitude of the phased anomalous
difference Fourier peaks for these metals will be similar. However, the
low-energy control data set will have almost no anomalous contribution
from Zn and about the same contribution from Ni. So, if you display
both maps, your Zn sites will be the only ones missing in the low energy
anomalous difference Fourier. The negative control here is important.
Using single-wavelength anomalous data alone, it is not generally
possible to tell the difference between metals. A low-occupancy metal
and a high-occupancy sulphur (or something) in the solvent can have the
same anomalous difference signal. You can use chemistry to try and help
the interpretation, but usually if one is interested in identifying a
bound metal the chemistry is the unkown and not the control.
So, anyway, I highly recommend the two-wavelength "edge contrast"
experiment for deducing the exact location of a particular metal
species. Every element has a unique set of sharp x-ray absorption
edges, so you can get spectacular contrast for anything with an
accessible edge.
You can also use x-ray florescence to identify which metals are present
in your sample. At a PX beamline with an energy-resolving fluorescence
detector, this only takes a few seconds. What you do is dial up a high
photon energy (something higher than the highest edge of the metals you
think you have) and gather the emission spectrum of your sample. Every
metal has a unique x-ray emission spectrum, so you can look up the
photon energies of the fluorescence emission peaks here:
http://xdb.lbl.gov/Section1/Sec_1-2.html
to figure out which metals you have got in your sample. For example, Zn
fluoresces at 8616eV, 8639eV and 9572 eV, but Ni flouresces at 7461 eV,
7480 eV and 8268 eV. A modern energy-resolving detector has error bars
around 200 eV, so you can easily tell the difference between these spectra.
Not every PX beamline has an energy-resolving fluorescence detector (at
ALS, only 12.3.1 and 4.2.2 have them), so you should check with your
friendly neighborhood beamline scientist if you plan to do this
experiment. However, since it is such a short experiment (maybe a grand
total of 10 minutes to mount/scan/dismount) you might be able to ask the
user or beamline scientist for a quick favor.
The other advantage of x-ray fluorescence is that you do not,
neccesarily, need a PX beamline to do it. There are deadicated EXAFS
beamlines at most synchrotrons. They generally look at rocks, but they
might not mind taking a few minutes of their time to tell you what
metals are in your tiny little sample.
You can also use protons to generate x-ray fluorescence:
http://biop.ox.ac.uk/www/lj2001/garman/garman_03.html
with the added advantage that this group has carefully calibrated their
system so that they can give you QUANTITATIVE information about your
metal composition. For example, they can tell you the ratio of Zn to S
pretty precisely. Since the S concentration is generally known from
your sequence (provided you don't have SO4 or DTT around), this can give
you the ratio of Zn to protein monomers exactly. I understand Elspeth
is interested in mail-in samples for her proton machine, so you can
contact her about that if you are interested.
-James Holton
MAD Scientist
Wendy Gordon wrote:
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Hello-
Thanks for entertaining my non-ccp4 question...
I am refining a crystal structure at ~2 Angstroms resolution in which I
find a large, unexpected electron density that is potentially a Zinc
ion because it is coordinated by 2 Histidines, 1 Glutamic acid, and
possibly a water molecule.
This site appears as a strong peak in anomalous difference Patterson
maps (the data was collected at SelMet wavelength 0.979 Angstroms and I
believe that Zn's absorption edge occurs around 1.2 Angstroms). The
problem is, I didn't add any zinc in any purification or
crystallization conditions. I DID affinity purify the protein with
nickel beads, so potentially it could be a Nickel ion. I should also
say that during refinement in refmac where my Zn occupancy is held at
1, that I obtain a negative peak in the 2Fo-Fc in this position, but if
I leave the site unoccupied- I get a huge positive peak- so I either
have the wrong species defined or my occupancy is not 100%- right?
Is there any way short of biochemical means (ITC, mutation, etc.) to
figure out what species is occupying this electron density? I have
thought of atomic absorption- has anyone tried it to determine the
metal species in a protein? Does it seem possible that I could have a
Zn ion in my protein crystal where the Zn could only come from our
standard DI water supply?
Thanks so much!
Wendy Ryan Gordon
