We at Birkbeck have just hosted a meeting of the London Structural Biology Club. This is a network of students and researchers in structural biology based in London and the South-East of England. Members get together for a couple of afternoons a year to hear research presentations, and the talks are followed by further informal discussion over refreshments (usually pizza and beer).
Four talks were given at the Birkbeck meeting, with each presenting not only new structural studies but also novel insights into molecular function and mechanisms derived from those structures. First to talk was Carien Dekker from the Institute of Cancer Research in London. She described the protein interaction network - known as the "interactome" by analogy with "genome" and "proteome" - of a eukaryotic cytosolic chaperonin, CCT. Chaperonins are a sub-class of chaperones, the proteins that assist other proteins in forming their stable three-dimensional structures, and they consist of two ring-like structures that associate back-to-back forming a cavity in which their substrate proteins fold. Dekker and her co-workers have used a number of different proteomics techniques, including the insertion of a long internal tag into a loop of the protein, to discover the range of substrates for this chaperonin. Proteins involved in functions as diverse as protein import into the nucleus, protein degradation, and chromatin remodelling. CCT is also necessary for the formation of the septin ring complex, and thence for cytokinesis (the last stage of cell division).
This work was published very recently in the EMBO Journal. Dekker and her colleagues are now working on the structure of CCT, which they hope will reveal more details of the function of this chaperonin: watch this space.
The second talk was given by David Komander, who has just moved from the Institute of Cancer Research to set up his own lab at the prestigious MRC Laboratory of Molecular Biology in Cambridge. He described some intriguing details of the ubiquitin system, through which proteins can be tagged for degradation. Ubiquitin (mentioned briefly in section 7 of the PPS course material) is a small protein (only 76 amino acids) with an alpha+beta fold (see PDB entry 1UBQ). Its C-terminus can be covalently linked to lysine side chains or N termini of other proteins. As ubiquitin itself has seven lysine residues (as well, of course, as an N-terminus) it can polymerise to form short chains. Proteomics has shown that all possible combinations of ubiquitin linkages can exist, but linkages in which the molecules are connected through lysines K48 and K63 are the most common. Poly-ubiquitin tags composed of different linkages have been linked with different functions; for example, binding a K48-linked ubiquitin chain to a protein will tag it for proteasomal degradation, whereas a K63-linked chain will tag a protein for signalling. The structures of these two forms of poly-ubiquitin have been shown to be very different, with K63-linked poly-ubiquitin forming an extended chain and K48-linked poly-ubiquitin a compact fold.
Ubuquitinlyation is a reversible process, and Komander has been studying the enzymes (deubiquitinases) that catalyse the hydrolysis of the peptide bonds between two ubiquitins, or between ubiquitin and another protein. These DUBs are analogous to the phosphatases that remove phosphate groups from protein side chains; their specificity , however, is more complex than that of phosphatases. Earlier this year, Komander and David Barford published the structure of the N-terminal domain of one such protein, A20 (Komander & Barford (2008), Biochem. J. 409, 771-785l; full text available). This is a cysteine protease domain known as the ovarian tumour (OTU) domain. These structural studies suggest both a novel architecture for the protein's catalytic triad and a novel mechanism - reversible oxidation - for the regulation of protein ubiquitinylation.
The "home team" at Birkbeck contributed a talk from Han Renaut, in Professor Gabriel Waksman's group. Waksman's own account of this work, on the structure of bacterial secretion systems, was blogged back in May and will not be described in more detail now.
Lastly, we heard from Erhard Hohenester who described an unpublished structure of SPARC, a protein that binds collagen. About 30% of the dry weight of the human body is composed of fibrils of this structural protein. It has a unique structure, being composed of three strands wound round each other in a triple helix. Every third residue of each strand must be a glycine, and the protein also contains a high percentage of proline. Some proline residues are post-translationally modified with the addition of an -OH group to form hydroxyproline. Besides being the major structural component of animal tissue, collagen binds to and forms complexes with many proteins including integrins and some tyrosine kinases. However, until now the only structure of a complex of collagen with another protein was with integrin (see PDB entry 1DZI).
SPARC, or osteonectin, is secreted by osteoblasts during bone formation, and binds calcium as well as collagen. Its structure as an isolated protein has been known for over ten years; it has two domains, one alpha-helical and the other containing many disulphide bonds (PDB 1BMO). Details of the new crystal structure of the collagen-SPARC complex, solved by Hohenester's group, must wait until the paper is published, but it is possible to say that it binds the hydrophobic sequence GVMGFO (which is a rare sequence in collagens, although one often involved in protein-protein interactions) into a hydrophobic pocket on the PARC molecule. [Note that that "O" is not a mistake; it is the single letter amino acid code for hydroxyproline.]
This London Structural Biology Club meeting was sponsored by Alpha Laboratories Ltd.