Real-space quantum-based refinement for cryo-EM: Q|R#3

Abstract

Electron cryo-microscopy (cryo-EM) is rapidly becoming a major competitor to X-ray crystallography, especially for large structures that are difficult or impossible to crystallize. While recent spectacular technological improvements have led to significantly higher resolution three-dimensional reconstructions, the average quality of cryo-EM maps is still at the low-resolution end of the range compared with crystallography. A long-standing challenge for atomic model refinement has been the production of stereochemically meaningful models for this resolution regime. Here, it is demonstrated that including accurate model geometry restraints derived from ab initio quantum-chemical calculations (HF-D3/6-31G) can improve the refinement of an example structure (chain A of PDB entry 3j63). The robustness of the procedure is tested for additional structures with up to 7000 atoms (PDB entry 3a5x and chain C of PDB entry 5fn5) using the less expensive semi-empirical (GFN1-xTB) model. The necessary algorithms enabling real-space quantum refinement have been implemented in the latest version of qr.refine and are described here.

Structural bases for aspartate recognition and polymerization efficiency of cyanobacterial cyanophycin synthetase

“…Each of the four models of G core and M core in the Te CphA1 tetramer was manually built based on the post-processed D 2 cryo-EM map of the apo state using Coot 40 . Model refinement was performed using phenix.real_space_refine with the secondary-structure restraints generated by Phenix.secondary_structure_restraints in the PHENIX program suite 41 , 42 . The built models were fit to the C 2 cryo-EM map of the substrate-bound state with UCSF Chimera 43 .…”

Optimal clustering for quantum refinement of biomolecular structures: Q|R#4

“…Here we continue our series [30,49,50,51] of publications that document the evolution and progress of the Q|R project. In what follows we describe how divide-and-conquer procedure can be optimized to minimize these errors without substantial increase of the computational time, which sets another milestone along the road of the Q|R project.…”

“…The QM energy gradients for all atoms from each of the fragments are computed with only the gradients from each cluster combined to create the total gradient. This procedure has been already successfully applied to improve several molecular models by quantum re nement in connection with experimental data from X-ray crystallography [49,50] and cryo-EM [51]. It has been also used in the QM optimization of protein structure [51].…”

“…This procedure has been already successfully applied to improve several molecular models by quantum re nement in connection with experimental data from X-ray crystallography [49,50] and cryo-EM [51]. It has been also used in the QM optimization of protein structure [51]. The divide-and-conquer procedure can potentially introduce and accumulate errors at the dividing boundaries of the clusters due to, for example, insu cient size of the buffer region.…”

In situ ligand restraints from quantum-mechanical methods

“…For the entire macromolecule, Phenix has the option of using the Amber (Case et al, 2018) and OPLS3e (Roos et al, 2019) molecular-mechanics force fields (Moriarty, Janowski et al, 2020;Zundert et al, 2021). The Q|R package makes it possible to derive model geometry restraints for refinement from ab initio quantum-chemical calculations (Zheng et al, 2017(Zheng et al, , 2020Wang et al, 2020). Another approach is to focus the QM method on the ligand and to use non-QM methods for the rest of the model.…”