A solid technique

When we think of techniques for determining three-dimensional (3D) protein structure, X-ray crystallography and solution NMR probably first come to mind. But, what do we do when proteins are inaccessible to solution NMR or do not readily crystallize? Could solid-state magic-angle-spinning (MAS) NMR be a solution?

To determine a 3D NMR structure, you need a large number of distance restraints, but for solid-state NMR — which could be applied to the tricky situations above — there has been no easy way to collect these restraints. Oschkinat and co-workers therefore devised a technique that allowed them to determine the structure of the α-spectrin Src-homology 3 (SH3) domain using only solid-state MAS NMR. Restraints between carbon atoms are essential for defining a protein's 3D structure, but so-called dipolar truncation effects make it difficult to measure these restraints in fully labelled 13C samples. The authors therefore made three differently labelled forms of the α-spectrin SH3 domain — one fully labelled sample and two biosynthetically site-directed 13C-enriched samples. Altering the labelling pattern in the latter two samples reduced the problem of dipolar truncation effects and allowed the observation of long-range interactions in the resulting NMR spectra.

The 3D structure of the α-spectrin SH3 domain determined by solid-state MAS NMR is remarkably similar to the X-ray structure, and this work “...paves the way for the structure determination of amyloids, small membrane proteins or receptor–ligand complexes...”.

Strong ties

Pseudomonas aeruginosa can infect most human tissues when the immune system is compromised, and this opportunistic pathogen is particularly dangerous to cystic fibrosis patients. In these patients, cell-surface glycoconjugates are increasingly fucosylated, and this makes them targets for P. aeruginosa, because this bacterium makes large amounts of a lectin called PA-IIL that binds L-fucose with a high affinity. In Nature Structural Biology, Imberty and colleagues now increase our understanding of PA-IIL by reporting its high-resolution crystal structure in complex with fucose. In addition, they used binding and modelling studies to show that antigens of the Lewis a series might be the preferred ligands of PA-IIL.

The crystal structure of PA-IIL is a tetramer, and the four independent monomers each consist of a nine-stranded antiparallel β-sheet. Each monomer binds two Ca2+ ions and one fucose ligand, and the authors found that “...fucose locks onto both calcium ions, a binding mode that is unique among protein–carbohydrate interactions.” This unique interaction explains the high affinity of PA-IIL for fucose — for most C-type animal lectins, which also bind two Ca2+ ions, only one ion is involved in sugar binding.

The strong ties between PA-IIL and fucose indicate that it could be a target for oligosaccharide-based therapeutics, and the structural detail provided by Imberty and co-workers should help in the design of antibacterial-adhesion prophylactics. REFERENCE Castellani, F. et al. Structure of a protein determined by solid-state magic-angle-spinning NMR spectroscopy. Nature 420, 98–102 (2002) Mitchell, E. et al. Structural basis for oligosaccharide-mediated adhesion of Pseudomonas aeruginosa in the lungs of cystic fibrosis patients. Nature Struct. Biol. 9, 918–921 (2002)