Bluelibra You have determined the structure of a 40kD metalloprotein al 2.2A res
ID: 1893470 • Letter: B
Question
Bluelibra
You have determined the structure of a 40kD metalloprotein al 2.2A resolution. The structure was reported to have R = 0.240, Rfree = 0.280, and all but 2 of tin protein residues lie in the "favored" or "allowed" regions of the Ramachandran plot. There are two copies of the peptide chain in the asymmetric unit of the crystal, and they share an extensive interface. There are 200 water molecules in the crystallographic model in addition to the protein itself. What about those two residues with "disallowed" phi/psi angles? Do you go back and try to re-build the model at those locations? Do you attribute it poor resolution or other experimental artifact? What would you say about these in a paper reporting the structure.Explanation / Answer
Yes. The first step is to go back to the electron density and carefully examine the map to see if there is an alternative conformation that better fits the map.
In the context that pretty much all electron density maps have good bits and bad bits, there are many reasons that may explain a poor (phi,psi) main chain angle set.
Perhaps the structure should be checked as to whether there is a cis peptide bond at that location.
Perhaps this main chain is found in a tight turn that may account for significant strain and distortion on the main chain.
/update --> (Some residues are found in strain conformations such as residues in beta buldge or some residues found in the active site of some enzymes.)
There may be features that can only be seen at higher resolution than 2.2 Å. For example, multiple residue conformations cannot be seen at 2.2 Å. What will occur is an average of these conformations, leading to incorrect positioning of the associated residues.
To limit the data to a given resolution, such as 2.2 Å, inherently introduces noise in the electron density map and thus to mis-interpretation. Beside, very low resolution data such as worst than 30 Å and the direct beam are not included with the coefficients in the Fourier Transform. This incompleteness of the data also introduces noise.
There are also errors in the diffraction data measurements. These errors may lead to spurious elements of density in the map, leading to mis-interpretation. Typically higher redundancy helps attenuate the effect of systematic and random errors. Beside, the x-ray measurements are radiation counts that observe a Poison distribution. Random noise is expected.
---
In a paper, these residues must be reported as they are the results. People recognize unfavorable main chain conformations may happen. If the affected residues are found in a tight turn then I would mention this as well.