The hidden treasure in your data: Phasing with unexpected weak anomalous scatterers from routine datasets

Abstract

AbstractSingle wavelength anomalous dispersion (SAD) utilizing anomalous signal from native sulfurs, or other atoms with Z≤20, generally requires highly redundant data collected with relatively long wavelength X-rays. Here we present experiences with two proteins, where anomalous signal from serendipitously acquired surface bound calcium atoms with data redundancy as low as 10 was utilized to drive de novo structure determination. In both the cases, the calcium atoms were acquired from the crystallization solution, and data collection strategy was not optimized for exploiting the anomalous signal from these scatterers. The X-ray data were collected at 0.98Å in one case and at 1.74Å in the other (wavelength optimized for sulfur but anomalous signal from Ca was exploited for structure solution). Similarly, using a test case, we show that the data collected at ~1.0Å wavelength, where the f" value for sulfur is 0.28e, is sufficient for structure determination using intrinsic sulfur atoms from a strongly diffracting crystal. Interestingly, it was also observed that SHELXD was capable of generating a substructure solution from high-exposure data with a completeness of 70% for low-resolution reflections extending to 3.5Å, with relatively low redundancy (i.e., 6.0 fold). Considering the fact that many crystallization conditions contain anomalous scatterers such as Cl, Ca, Mn etc., it appears that the data from well diffracting crystals should be processed, with anomalous pairs kept separate, so as to detect any fortuitous signal that might be present, which could prove useful in either solving the structure de novo or in accurately assigning surface bound atoms.