In terms of detecting pseudo translational symmetry, it sounds like both crystals have it, but the relationship between the two complexes is closer to perfect in the 2nd crystal. Looking at that problematic density, it does indeed suggest that the lattice is slightly messed up. For instance, you could have a sort of twinning-ish problem. Say, if your molecules usually sit in a row like such: Mol1 - x Angstroms - Mol2 - 0.9x Angstroms - Mol1 - etc. but some rows of the crystal pack: Mol1 - 0.9x Angstroms - Mol2 - x Angstroms - Mol1 etc, the Mol1s will superimpose but the Mol2s won't. Or could it be that sometimes the protein binds 1bp over from where you wanted it to? Do you have any halogens on the DNA that you could use to make sure the DNA sequence register is always the same (i.e. if you put in 1 bromo-dU, do you get 1 anomalous peak or two)? Since you don't have a problem with the density derived from your first data set, you might get lucky just by cranking through many different crystals of the 2nd complex. I might try micro seeding as well in hopes of somehow tweaking the crystal growth pattern.

As a side point, a Patterson peak that is 18% of the origin may not be huge, but it seems unusually large to me. I suspect there's a broad grey area where you need some additional analysis to decide what is really noteworthy pseudo translational symmetry, such as noticeable patterns of dark vs. light spots. As a "control", I looked up the validation report now posted for one of my lab's old structures that had pseudo translational symmetry clearly visible in the diffraction pattern (and later in the coordinates) (1szp). Xtriage found a 19%-of-origin peak for that, and decided "no signifiant pseudo translation."

Phoebe

++++++++++++++++++++++++++++++++++++++++++

Phoebe A. Rice
Dept. of Biochemistry & Molecular Biology
The University of Chicago

773 834 1723; [log in to unmask]
http://bmb.bsd.uchicago.edu/Faculty_and_Research/

http://www.rsc.org/shop/books/2008/9780854042722.asp

From: CCP4 bulletin board [[log in to unmask]] on behalf of Minyun Zhou [[log in to unmask]]
Sent: Wednesday, July 30, 2014 9:58 AM
To: [log in to unmask]
Subject: [ccp4bb] Pseudo-translation problem

Dear all,

I am trying to determine the structure of a protein-DNA complex. I collected several datasets of the same protein in complex with two dsDNAs. Two DNAs have the same sequence, but one DNA contain a single central mispair, which is cleaved by the enzyme leaving an abasic site, while the other contains a non-cleavable mispair. Two datasets have almost the same space group and cell parameter, however, the later one has been detected with pseudo-translation while the other not.

The details of two datasets are as follows:

Dataset 1. Protein with abasic site containing dsDNA:

C2221: a=106, b=111, c=182.2, α=β=γ=90

Resolution: 2.45 A

Xtriage summary: The largest off-origin peak in the Patterson function is 18.57% of the height of the origin peak. No significant pseudotranslation is detected.

Dataset 2. Protein with non-cleavable lesion containing dsDNA:

C2221: a=106.1, b=111.1, c=183, α=β=γ=90

Resolution: 3.0 A

Xtriage summary: The analysis of the Patterson function reveals a significant off-origin peak that is 60.75 % of the origin peak, indicating pseudo translational symmetry.

I first determined the structure using the dataset 1 by MR. There are two protein-DNA complexes in one asymmetric unit, and two molecules are similar except the conformation of a small loop at the active site. The final model has the R_work/R_free of 19.3%/23.0%. Then I use this refined structure as model to solve the second structure with dataset 2. The final R_work/R_free is 24.3%/28.9%. The DNA density looks okay while the protein part is problematic. After refinement, although the main chain of the protein is fitted well into the density, there is still a lot of unidentified continuous positive density next to the polypeptide chain, especially near the region involved in the crystal packing. I attached a snapshot below, which shows one of these positive densities lies along a helix. Since there is no other polymer in the reservoir solution that can fit, the additional density seems to indicate another slightly shifted conformation of the protein itself. My question is whether this seemingly “dual conformation” is caused by pseudo-translation and indicates the solution I found is incorrect. And how can I eliminate the effect of pseudo-translation to get the correct structure?

Thanks for your help!

Best regards,

Minyun