As you suggest, it depends on the contour level. Looking through a list of 17 dicarboxylates that I found problematic, there were 4 (actually two and their ncs-mates) that showed disconnected density for the carboxylate at 1.4 sigma as in the figure I sent. Going up to 2 sigma, four more became disconnected and linear (backbone density is continuous to 3.5 sigma). Going down to 0.6 sigma, disconnected residual density showed up for several more. So I guess there are varying degrees of decarboxylation, and varying extent of retention of the fragment. I have the impression these are mostly on the protein/solvent boundary, which could explain their disorder, but perhaps would also expose them to greater concentration of radicals generated in the solvent channels(?).
Ed
On 05/04/2017 06:25 AM, Andrew Leslie wrote:
> Dear Ed,
>
> I find your electron density quite interesting, because generally (I think, I would be happy to be corrected on this) when de-carboxylation of Asp/Glu occurs due to radiation damage, there is no evidence of what happens to the resulting CO2 group. One interpretation of this is that it diffuses away from the side chain and is effectively totally disordered, so no electron density is seen, but I was surprised that this would always be the case, especially as I would have thought that diffusion would be quite limited at 100K (maybe I’m wrong about that too, but that is supposed to be one reason why radiation damage is less at 100K).
>
> If the residual density is due to partial de-carboxylation, then I would have expected density for the CG-CD bond, which is not present (at your chosen contour level).
>
> Do many of your Glu side chains have the residual density?
>
> Best wishes,
>
> Andrew
>
>
>> On 3 May 2017, at 22:19, Edward A. Berry <[log in to unmask]> wrote:
>>
>>
>>
>> On 05/03/2017 02:46 PM, Gerard Bricogne wrote:
>>> Dear Ed,
>>>
>>> Have you considered the possibility that it could be a water
>>> stepping in to fill the void created by partial decarboxylation of the
>>> glutamate? That could be easily modelled, refined, and tested for its
>>> ability to flatten the difference map.
>>>
>>> Gerard.
>>>
>> Actually some of them do appear decarboxylated. Is that something that can happen? In the crystal, or as radiation damage?
>> However when there is density for the carboxylate (figure), it appears continuous and linear, doesn't break up into spheres at H-bonding distance - almost like the CO2 is still sitting there- but I guess it would get hydrated to bicarbonate. I could use azide. Or maybe waters with some disorder.
>> Thanks,
>> eab
>>
>> Figure- 2mFo-DFc at 1.3 sigma, mFo-DFc at 3 sigma, green CO2 is shown for comparison, not part of the model.
>>
>> <decarbox.gif>
>
>
|