Mystery density

From: Martin Montgomery

Posting on behalf of Katia Baranova (see below).  Any suggestions gratefully received!

Hello,

we have a puzzling density and hope someone has an idea of what it may be.

The pictures of it are here:

http://www.mrc-mbu.cam.ac.uk/sites/default/files/images/mystery/top.png
http://www.mrc-mbu.cam.ac.uk/sites/default/files/images/mystery/front.png
http://www.mrc-mbu.cam.ac.uk/sites/default/files/images/mystery/side.png

The space group is I422, resolution 1.6A

The trial contained Tris, DTT, EDTA, sodium, potassium, phosphate, and chloride.

The centre of our unknown feature is about 3.8A from the Arg and it is also on the four fold symmetry axis.

It is + 28 sigma in the center and  about + 9.5 sigma in each of the "tentacles".

Thank you very much!

Katia.

On 30 Nov 2011, at 17:30, Katia Baranova wrote:

Hello,

Well, it doesn't work.
Can you forward it please?

thanks a lot,

K

----------
From: Mark J van Raaij


although it is hard to see in 3 2D views, some of the tentacles appear connected to the protein, could they result from some of the long Lys or Arg side-chains not following the 4-fold symmetry? i.e. alternative conformations?
The density on the 4-fold may be ions, solvent, some metal ion? "Noise appears to collect on symmetry axes" (for want of a better way to put it), so the sigma-levels there are not very reliable. Sometimes one sees quite strong negative peaks on symmetry axes that can not be explained, so I would expect spurious positive peaks to also occur. Of course we normally just put a water in it if it can make reasonable hydrogen bonds.

Mark J van Raaij

----------
From: Carlos Kikuti

It's bang on the Y axis.... I'd just ignore it (treat it as just noise) and work on the rest of the model. There seems to be some trouble in other parts of it. (Had a similar case but at 3.5 A data).  I wonder what is the gap between Rfree and Rwork.

And I'm curious for what more experienced people will say about it.

Carlos


----------
From: Adrian Goldman


We had a similar case in our Eib structure (2xzr), with large peaks on a 3-fold.  We initially thought it was noise, but Andrei Lupas convinced us to look more carefully, and it turned out to be chloride ion.

You could do an wavelength scan at a beamline looking for anomalous signal at the longest wavelength you can get to.  If it's just noise, it shouldn't be particularly visible.  If it is chloride or phosphate, you should at least have a peak there.

Adrian

----------
From: Eleanor Dodson

Well - first thing to do is an DANO map and see if it is a metal ion. The relative height of the DANO peaks for S and anything else can give you some clues.

Eleanor

----------
From: Roger Rowlett

I think a metal ion is unlikely given that the nearest ligand(s) are Arg residues. I would lay money on an anion. The central feature looks very much like a bicarbonate ion, and is very similar to the bicarbonate ion bound in a generic anion binding site near 2 Arg residues on a NCS 2-fold axis in beta-carbonic anhydrase. The rest might be waters. If this sits on a crystallographic symmetry axis, things can look a little funky, but you do seem to have a triangular planar feature in a special position. For a similar binding interaction with bicarbonate, see PDB 3E3I, 3E3F, and Rowlett, et al. (2010) Biochemistry 49, 3640-3647, "Evidence for a Bicarbonate 'Escort' Site in H. influenzae beta-Carbonic Anhydrase."

Cheers
,

----------
From: Ed Pozharski

The center bit looks very much like phosphate - see what happens if one
is placed there

It may be educational to look at the maps in P1, just to see how much
trouble is caused by "noise accumulation on the symmetry axis"

--
Edwin Pozharski, PhD, Assistant Professor