Advances in methods for macromolecular structure solution: ab initio and MR approaches

New methods have been recently developed to improve the structure solution of macromolecules by ab initio (Patterson or Direct Methods) and non ab initio (Molecular Replacement) approaches. Phasing proteins at non-atomic resolution is still a challenge for any ab initio method. The combined use of different algorithms [Patterson deconvolution and superposition techniques, cross-correlation function (C-Map), the VLD (Vive la Difference) approach included in the Direct Space Refinement (DSR) procedure, a new probabilistic formula estimating triplet invariants and capable of exploiting a model electron density maps, the FREE LUCH extrapolation method, a new FOM to identify the correct solution] allow to overcome the lack of experimental information. The new methods have been applied to a large number of protein diffraction data with resolution up to 2.1Å, under the condition that Ca or heavier atoms are in the structure. Results show that solving proteins at limited resolution is a feasible task, achievable even by new Direct Methods algorithms, against the traditional common believe that atomic resolution is a necessary condition for the success of a direct ab initio phasing process. A new procedure (REVAN) , aiming at solving protein structures via Molecular Replacement and density guided optimization algorithms, has been assembled. It combines a variety of programs (REMO09, REFMAC, COOT) and algorithms (Cowtan-EDM, DSR, VLD, FREE LUNCH), and can successfully lead to the structure solution also when the sequence identity between target and model structures is smaller than 0.30 and data resolution up to ~ 3Å. The application to a wide set of test structures (including difficult cases proposed by DiMaio et al. (2011), solved by using MR procedures together with energy guided programs) suggests that REVAN is quite effective even far from atomic resolution and, in combination with EDM techniques and sequence mutation algorithms, it is able to efficiently extend and refine the set of phases, reducing its average error. The final step of the automatic solving process (ab initio or MR approaches) is the application of an Automated Model Building program (i.e. Buccaneer, Nautilus, ARP-wARP or Phenix-Autobuild) in order to recover the correct structure. Results suggest that the quality of the phases at the end of the phasing process is good enough to lead the AMB program to success. These new efficient procedures are implemented in the current version of the software package SIR2014.