Temperature dependence of Uiso constraints in riding hydrogen treatments

"Anisotropic parametrisation of the thermal displacements of hydrogen atoms in single-crystal X-ray structure refinement is not possible with independent atom model (IAM) scattering factors. This is due to the weak scattering contribution of hydrogen atoms. Only when aspherical scattering factors are used can carefully measured Bragg data provide such information. For conventional structure determinations parameters of ""riding"" hydrogen atoms are frequently constrained to values of their ""parent"" heavy atom. Usually values of 1.2 and 1.5 times X-U_eq are assigned to H-U_iso in these cases. Such constraints yield reasonable structural models for room-temperature data. However, todays small molecule X-Ray diffraction experiments are usually carried out at significantly lower temperatures. To further study the temperature dependence of ADPs we have evaluated several data sets of N-Acetyl-L-4-Hydroxyproline Monohydrate at temperatures ranging from 9 K to 250 K. Methods compared were HAR [1], Invariom refinement [2], time-of-flight Neutron diffraction and the TLS+ONIOM approach [3]. In the TLS+ONIOM approach non-hydrogen ADPs from Invariom refinement provided ADPs for the TLS-fit. Hydrogen atoms in all methods were grouped and analyzed according to their Invariom name. We reach a good agreement of the temperature dependence of H-U_iso/X-U_eq. At very low temperatures the ratio H-U_iso/X-U_eq can be as high as 4, e.g. for Hydrogen attached to a sp3 carbon atom with three non-Hydrogen atom neighbors. Since all methods consistently show that the H-U_iso/X-U_eq ratio is temperature dependent, this effect should be taken into account in conventional structure determinations."

SHELXT: Integrating space group determination and structure solution

SHELXT is intended for robust routine solution of small molecule crystal structures. It makes the simple but powerful assumption that the structure consists of resolved atoms, but unlike classical direct methods it is not required that the atoms are 'equal'. This enables it to succeed with poor or incomplete data but makes it unsuitable for structures that are twinned, modulated or contain severe (e.g. 'whole molecule') disorder. SHELXT is a dual-space program that starts with a Patterson minimum superposition and iteratively applies the random omit procedure (also used in SHELXD) with data expanded to space group P1, but does not use phase probability relations or solvent flipping. In the SHELX system it will probably obsolete SHELXS but not SHELXD, which is better for large equal-atom and twinned structures. SHELXT reads any legal SHELX format .ins and (HKLF3 or 4) .hkl files. It extracts the Laue group and tries to find space groups in this Laue group and origin shifts to fit the phases from the best P1 solution, and makes an approximate assignment of element types using the elements specified on the SFAC instruction (and maybe a couple more). This is followed by an isotropic refinement and an attempt to assign the absolute structure if the space group is non-centrosymmetric. It is hoped to release SHELXT as part of the SHELX system (http://shelx.uni-ac.gwdg.de/SHELX/index.php) in time for the 2014 Montreal IUCr Meeting.

Check your metal - not every density blob is a water molecule

Metals play vital roles in both the mechanism and architecture of biological macromolecules, and are the most frequently encountered ligands (i.e. non-solvent heterogeneous chemical atoms) in the determination of macromolecular crystal structures. However, metal coordinating environments in protein structures are not always easy to check in routine validation procedures, resulting in an abundance of misidentified and/or suboptimally modeled metal ions in the Protein Data Bank (PDB). We present a solution to identify these problems in three distinct yet related aspects: (1) coordination chemistry; (2) agreement of experimental B-factors and occupancy; and (3) the composition and motif of the metal binding environment. Due to additional strain introduced by macromolecular backbones, the patterns of coordination of metal binding sites in metal-containing macromolecules are more complex and diverse than those found in inorganic or organometallic chemistry. These complications make a comprehensive library of "permitted" coordination chemistry in protein structures less feasible, and the usage of global parameters such as the bond valence method more practical, in the determination and validation of metal binding environments. Although they are relatively infrequent, there are also cases where the experimental B-factor or occupancy of a metal ion suggests careful examination. We have developed a web-based tool called CheckMyMetal [1](http://csgid.org/csgid/metal_sites/) for the quick validation of metal binding sites. Moreover, the acquired knowledge of the composition and spatial arrangement (motif) of the coordinating atoms around the metal ion may also help in the modeling of metal binding sites in macromolecular structures. All of the studies described herein were performed using the NEIGHBORHOOD SQL database [2], which connects information about all modeled non-solvent heterogeneous chemical motifs in PDB structure by vectors describing all contacts to neighboring residues and atoms. NEIGHBORHOOD has broad applications for the validation and data mining of ligand binding environments in the PDB.