Crystal structure of R120G disease mutant of human αB-crystallin domain dimer shows closure of a groove.

Small heat shock proteins form large cytosolic assemblies from an "α-crystallin domain" (ACD) flanked by sequence extensions. Mutation of a conserved arginine in the ACD of several human small heat shock protein family members causes many common inherited diseases of the lens and neuromuscular system. The mutation R120G in αB-crystallin causes myopathy, cardiomyopathy and cataract. We have solved the X-ray structure of the excised ACD dimer of human αB R120G close to physiological pH and compared it with several recently determined wild-type vertebrate ACD dimer structures. Wild-type excised ACD dimers have a deep groove at the interface floored by a flat extended "bottom sheet." Solid-state NMR studies of large assemblies of full-length αB-crystallin have shown that the groove is blocked in the ACD dimer by curvature of the bottom sheet. The crystal structure of R120G ACD dimer also reveals a closed groove, but here the bottom sheet is flat. Loss of Arg120 results in rearrangement of an extensive array of charged interactions across this interface. His83 and Asp80 on movable arches on either side of the interface close the groove by forming two new salt bridges. The residues involved in this extended set of ionic interactions are conserved in Hsp27, Hsp20, αA- and αB-crystallin sequences. They are not conserved in Hsp22, where mutation of the equivalent of Arg120 causes neuropathy. We speculate that the αB R120G mutation disturbs oligomer dynamics, causing the growth of large soluble oligomers that are toxic to cells by blocking essential processes.

Crystal structures of alpha-crystallin domain dimers of alphaB-crystallin and Hsp20.

Small heat shock proteins (sHsps) are a family of large and dynamic oligomers highly expressed in long-lived cells of muscle, lens and brain. Several family members are upregulated during stress, and some are strongly cytoprotective. Their polydispersity has hindered high-resolution structure analyses, particularly for vertebrate sHsps. Here, crystal structures of excised alpha-crystallin domain from rat Hsp20 and that from human alphaB-crystallin show that they form homodimers with a shared groove at the interface by extending a beta sheet. However, the two dimers differ in the register of their interfaces. The dimers have empty pockets that in large assemblies will likely be filled by hydrophobic sequence motifs from partner chains. In the Hsp20 dimer, the shared groove is partially filled by peptide in polyproline II conformation. Structural homology with other sHsp crystal structures indicates that in full-length chains the groove is likely filled by an N-terminal extension. Inside the groove is a symmetry-related functionally important arginine that is mutated, or its equivalent, in family members in a range of neuromuscular diseases and cataract. Analyses of residues within the groove of the alphaB-crystallin interface show that it has a high density of positive charges. The disease mutant R120G alpha-crystallin domain dimer was found to be more stable at acidic pH, suggesting that the mutation affects the normal dynamics of sHsp assembly. The structures provide a starting point for modelling higher assembly by defining the spatial locations of grooves and pockets in a basic dimeric assembly unit. The structures provide a high-resolution view of a candidate functional state of an sHsp that could bind non-native client proteins or specific components from cytoprotective pathways. The empty pockets and groove provide a starting model for designing drugs to inhibit those sHsps that have a negative effect on cancer treatment.

The polycystin-1 C-type lectin domain binds carbohydrate in a calcium-dependent manner, and interacts with extracellular matrix proteins in vitro.

Mutations in the PKD1 gene are responsible for 85% of cases of autosomal dominant polycystic kidney disease (ADPKD). This gene encodes a large membrane associated glycoprotein, polycystin-1, which is predicted to contain a number of extracellular protein motifs, including a C-type lectin domain between amino acids 403--532. We have cloned and expressed the PKD1 C-type lectin domain, and have demonstrated that it binds carbohydrate matrices in vitro, and that Ca(2+) is required for this interaction. This domain also binds to collagens type I, II and IV in vitro. This binding is greatly enhanced in the presence of Ca(2+) and can be inhibited by soluble carbohydrates such as 2-deoxyglucose and dextran. These results suggest that polycystin-1 may be involved in protein-carbohydrate interactions in vivo. The data presented indicate that there may a direct interaction between the PKD1 gene product and an ubiquitous extracellular matrix (ECM) protein.