Small-molecule crystal structures, data

Parkin A, Collins A, Gilmore CJ, Wilson CC (2008) Using small molecule crystal structure data to obtain information about sulfonamide conformation. Acta Crystallogr B 64 66-77... 

The Cambridge Crystallographic Data Centre produces the Cambridge Structural Database (http //www.ccdc.cam.ac.uk/products/csd/), which covers organic and metal-organic small-molecule crystal structures. 

As observed for the N- H O interactions between main-chain and side-chain groups, the N-ft Ow angles are almost linear, 156(15)°, if the all-a-helix proteins are omitted from the sampling, see Ihble 19.6d. The distribution of these angles is 140° to 180° for 90% of the data, and consistent also with small molecule crystal structures [75, 382, 475]. The spread of hydrogen-bond distances is broad, and wider for C=0 than for N-H, probably because the C=0 oxygen is more readily accessible and multiple C=0- -HOw interactions are frequently observed. 

If distances between main-chain C=0 and N-H groups and hydrogen-bonded water molecules are compared (Thble 23.5 and Fig. 23.5 a), it is apparent that the distribution is narrower about the N-H groups. The reason is that the C=0 groups are more exposed and, in most cases, accessible to more than one or two hydrogen-bond donors. Where this results in overcrowding the C=0 0W distances may be beyond the 3.5 A cut-off limit used in the analysis [596]. In the major parts of both C=0 Ow and N-H ow distribution, the peaks in Fig. 23.5a are at 2.89 A (2.94 A in Thble 23.5a) and 2.97 A respectively, as expected from the data for the hydrogen bonds in small molecule crystal structures. 

Ofc/ values for a small-molecule crystal structure are listed for E values greater than 1.5. Okl data were chosen in order to simplify the description of the method. In practice all E(hkl) greater than 1.5 are used. The crystal is orthorhombic, unit-cell dimensions a — 13.52, h = 26.58, c = 7.26 A, space group P2i2i2i. 

The conformer generator, ROTATE, developed by Schwab [45], can be considered as an extension to CORINA. It is based on a similar set of rules and data as the conformational analysis package, MIMUMBA. The TA library, which reflects the conformational preferences of acyclical parts in small molecule crystal structures, is implemented in ROTATE to explore the conformational space of the open-chain portions of a molecule. A description of the derivation of the TA library from the Cambridge Structural Database, the basis for its application to generate conformations likely to be of biological relevance, and its usage is given in Sec. 5.2.2. 

Fig. 13.8. Stereo-diagram of the preferred spatial location of hydrogen-bond donors about N in a pyridine ring, determined by mapping the composite crystal-field environment data taken from small molecule crystal structures containing this fragment. The data were expanded according to the C2v-symmetry of the pyridine ring...

T2), and 00CCH2-C-(T3), data taken from small-molecule crystal structures... 

Fig. 13.21. Histograms of dihedral angles for fragments which are representative for the nine flexible bonds in the inhibitor methotrexate data taken from small molecule crystal structures (for data with trigonal planar and pyramidal geometry at N have been considered). For each fragment the local symmetry has been taken into account. Some of the angles exhibit clear conformational preferences, others show a broad distribution. The torsion angles for the conformation of methotrexate, adopted at the active site of DHFR (3DFR, 4DFR) from two species E. coli, 1. casei) and in small molecule crystal structures (DOJZADOl, QUIASPIO, FABVUZ (two independent molecules)) are indicated (lines to the right, protein data boldfaced)...

Geometrical aspects of the nucleophilic approach have been derived from structure correlation [107, 108] on small molecule crystal structures. The addition of amine to carbonyl has been shown to proceed via an approach of the nucleophile from above and behind the C = 0 group, in a local mirror plane of the R2C=0 fragment [109]. A more recent study is based on structural data of 31 N... R2C = 0 fragments [110]. In Figure 13.24, these are superimposed by matching the three atoms directly bound to the central carbon, and the observed positions of N are indicated (-h). These results were compared with a computer simulation of a model reaction in which the attack of NH3 on H2C=0 is assisted by a water molecule [110]. 

The low resolution of a protein structure determination implies limited accuracy. Therefore any discussion of bond-length variations should be complemented with data from small molecule crystal structures. Tris-di(thiocarbamate)lron(III) complexes show a high-spin/low-spin equilibrium in the solid state [141, 142]. This process involves a systematic change in Fe-S bond length and affects the atomic displacement parameters [141]. In the low-spin configuration there are no an-... 

Small-molecule crystal structures are identified by their Cambridge Structural Database (CSD) refcodes as described in Chapter 3. Protein crystal structures are identified by their Protein Data Bank four-symbol refcodes. Short format references for both types of refcodes are given in Appendix B. 

As molecular dynamics simulations imply considerable demands on computation time and expertise, a collection of empirical and implicit methods for investigating water-induced effects on ligand binding have been developed. SuperStar, for example, uses information derived from experimental data to identify interaction sites in proteins [85, 86]. It generates propensity maps representing the distribution of one functional group around another found in small-molecule crystal structures. Success rates reported for the prediction of water positions around proteins are promising, in particular for those water positions that are well defined in the X-ray structure [86]. 

By the early twenty-first century, the Cambridge Structural Database contained precise information for more than one-quarter miUion small-molecule crystal structures, and most of the new data had arrived electronically. 

---