 |
 |
OCMS Research Summary 1998 - 2001
The OCMS research programme during 1998 - 2001 was divided into five core areas encompassing:
Eight collaborative projects were undertaken, requiring expertise from several groups as well as support from OCMS facilities and staff:
(i) Single Crystal EnzymologyThis has been an exciting and productive area for OCMS. The combination of very high resolution (ca. 1Å) analyses with use of the OCMS single crystal spectrophotometer and other techniques (fluorescence, EXAFS) has led to ground breaking time–resolved studies on IPNS, cytochrome dependent oxygenases and proteases (e.g. elastase) (Wilmouth et al. Nat Prod Reps 17, 527-533 2000; Wilmouth et al. Nat Struct Biol). Rather than reinforcing existing mechanistic proposals, some of the observations have been unexpected, leading to new insights, including rearrangements during catalysis. Exposure of crystals of IPNS complexed to a synthetic substrate analogue to pressurised dioxygen led to a remarkable reaction in which the ferryl intermediate was trapped by the side-chain of the substrate (Burzlaff et al. Nature 401, 721-724, 1999). Time-resolved crystallographically on clavaminic acid synthase allowed observation of an unprecedented rearrangement around the iron centre and led to a general mechanistic proposal for 2-oxoglutarate oxygenases (Zhang et al. submitted). In collaboration with Prof. J. Hajdu (Uppsalla) we have developed the use of single crystal gradient (i.e. crystals in which a chemical gradient exists) methodology, in combination with pH jumps and freezing, to visualise intermediates during serine protease catalysis. We have been able to rationalise how oxy-anion hole stabilisation of a tetrahedral intermediate is compatible with efficient product release (Wright et al. Biochem J 351, 335-340, 2000; Wilmouth et al. Nat Struct Biol, 8, 689-694, 2001). The results have helped us to understand the processes underlying inhibition of hydrolytic enzymes by beta-lactams (Wilmouth et al. Biochemistry 38, 7989-7998, 1999; idem, ibid, 37, 17506-17513, 1998). (ii) Characterisation of Macromolecular Assemblies by ESI-MSOCMS scientists pioneered the use of mass spectrometry to study post-translational modifications, enzyme mechanism and protein folding. Today soft ionisation MS is used routinely for characterisation of single biopolymers, their folding, (Coyle et al. Nat Struct Biol 6, 683-690, 1999), and small molecule/metal complexes, and exciting studies in these areas have continued at OCMS (Last et al. Curr Opin Chem Bio. 3, 564-570, 1999). In the last few years we have developed MS as a technique for studying non-covalent macromolecular complexes (Vis et al. JACS 120, 6427-6428, 1998; Chung et al. Protein Science, 8, 1962-1970, 1999). The work is likely to establish MS as a front-line technique for mapping interactions between biopolymers. Intact ribosomes have been analysed by means of nanoflow electrospray techniques. By lowering the Mg(II) ion concentration in solutions containing ribosomes the particles were found to dissociate into 30S and 50S subunits (Benjamin et al. PNAS 95, 7391-7395, 1998; Rostom et al. PNAS 97, 5185-5190, 2000). Further dissociation into smaller macromolecular complexes and then individual proteins could be induced by subjecting the particles to increasingly energetic gas phase collisions. The ease with which proteins dissociated from the intact species was found to be related to their known interactions in the ribosome particle. Another giant macromolecular assembly characterised by MS was the intact MS2 virus capsid (Tito et al .JACS 122, 3550-3551, 2000). MS has also been used to characterise the structure and mechanism of fibril formation from a range of medicinally important proteins including detailed conformational analyses using hydrogen-deuterium exchange analyses (Tito et al. J Mol Biol 303, 267-278, 2000). MS has also been used to characterise previously unidentified interactions between proteases and naturally occurring protein inhibitors and has been developed as a technique for mining libraries (Freitas et al. J Am Soc Mass Spectr 11, 1023-1026, 2000; Wright et al. Biorg Med Chem Letts 10, 1219-1221, 2000). (iii) Protein Engineering of Oxidases and OxygenasesStructural studies on non-haem oxygenases have made enormous advances in the last three years and OCMS is a world-leader in this area (Schofield and Zhang Curr Opin Struct Biol 9, 722-731, 1999). Following from the structure of isopenicillin N synthase (IPNS), the first from the ubiquitous family of 2-oxoglutarate and dependent oxygenases, four more structures from the family have been determined. The first reported of these, deacetoxycephaosporin C synthase (DAOCS) catalyses formation of the medicinally important cephalosporin antibiotics (Valegard et al. Nature 394, 805-809, 1998; Lloyd et al. J Mol Biol 287, 943-960, 1999). Engineering of DAOCS has lead to variants with selectivity altered towards hydrophobic penicillins that are in development for commercial use (Lee et al. J Mol Biol 308, 937-948, 2001; Lee et al. J Biol Chem 276, 18290-18295 2001). Crystal structures of proline-3-hydroxylase (Zhang et al. submitted), clavaminic acid synthase (Zhang et al. Nat Struct Biol 7, 127-133, 2000) and anthocyanidin synthase (Turnbull et al. Chem Comm 2473-2474, 2000; Wilmouth et al., Structure 10, 93-103, 2002) have led to the identification of variations on a common fold and powerful evidence for divergent evolution within the family. The combination of structural, genomic, and chemical expertise within OCMS has led to the identification of related enzymes in other metabolic pathways, dealkylation of DNA and disease states, including Refsum's Disease, (Mukherji et al., Hum Mol Genet 10, 1971-1982, 2001; Chem Comm (11), 972-973, 2001) hypoxic sensing and tumour vascularisation (Epstein et al. Cell 107, 43-54, 2001). (iv) Substrate Recognition by Kinases and PhosphatasesProtein kinases are among the most common domains in the genomes of eukaryotes, indicating the importance of phosphorylation in cellular processes, especially signalling. Kinases comprise a common catalytic core but each recognises specific substrates and is regulated by different mechanisms. Following our demonstration that phosphorylation on the activation segment is a key component of regulation in kinases (Johnson et al. Cell 85, 149-158, 1996), we have now demonstrated that this segment is essential for substrate recognition. Crystal structure analysis of the cell cycle regulatory kinase, phospho-CDK2/cyclin A with substrate and recruitment peptides have shown the activation segment adopts an unusual conformation that enables it to select the Ser-Pro motif of the substrate and that the phospho-threonine contributes to specificity for a basic group on the substrate (Brown et al. Nature Cell Biology 1, 438-443, 1999). The remote (40 Å) recruitment site plays a definite, but as yet incompletely understood, role in defining substrate specificity. The recent structure of the Kinases Associated Phosphatase (KAP) with phospho-CDK2, the first structure of a phospho-protein/enzyme complex, has demonstrated key elements in control by dephosphorylation and phospho-protein recognition (Song et al. Mol Cell 7, 615-26, 2001). The association with KAP results in significant conformational changes in phospho-CDK2 to allow the phospho-threonine to reach the catalytic site. The major recognition is through a remote docking site on CDK2 that is also utilised by the regulatory protein Cks1. Structures of CDK2 in complex with inhibitors are providing insights for potential anti-cancer drugs. (v) Electron Microscopy of ProteinsFollowing the establishment of electron microscopy in OCMS a number of different proteins have been analysed. Phosphorylase kinase holoenzyme (MW 1.3 x 106) is among the largest and most complex of protein kinases. Negatively stained images of tilted and untilted particles with the random conical tilt method have led to a 3D model at 22 Å resolution of the (abgd)4 complex (Venien-Bryan et al. Structure 10, 33-41, 2002). The 3D reconstruction shows a butterfly-like shape (roughly 260Å by 215Å by 160Å) with two wing-like lobes connected by thin oblique bridges. Two-dimensional crystallisation technique for proteins on a lipid monolayer film has been applied with HupR, a microbial response regulator controlled by phosphorylation. 2D crystals of were formed by exploiting the interaction between the N-terminal polyhistidine-tag of the His6-HupR protein and nickel ions carried by specially synthesised lipid (Ni-NTA-DOGA). A projection electron density map at 9Å resolution of the full-length response regulator HupR revealed the regulatory domain and the C-terminal domain that contains the DNA binding motif (Venien-Bryan et al. J Mol Biol 296, 863-871, 2000). We have extended the technique of crystallization on a lipid monolayer to his-tagged membrane proteins utilising nickel containing fluorolipids. In collaboration with Professor W. Kuhlbrandt (Frankfurt), we have successfully used this technique for crystallization of an H+-ATPase from Arabidopsis Thaliana. Image processing of images taken from frozen hydrated specimen yielded a projection map at 9Å resolution (Lebeau et al. J Mol Biol 308, 639-47, 2001). (vi) Host InvasionPathogens usually invade host cells by binding to cell surfaces or the extracellular matrix. The N-terminal region of fibronectin is known to have a role in bacterial pathogenesis. Both Staphylococci and Streptococci have been shown to bind to fibronectin module pairs from the N-terminal region domain. In a good example of the type of collaboration fostered within OCMS, the Campbell and Dobson groups began to explore the nature of these interactions a few years ago, initially with SmithKline Beecham. At first it was shown that the binding region of the bacterial peptide was unfolded in solution (Penkett et al. Biochemistry 37, 17054-67, 1998). These studies were then extended to show, using NMR spectroscopy and isothermal titration calorimetry, how the bacterial peptides bind to fibronectin Residues in the 4F15F1 module pair of fibronectin have been identified that bind to a fibronectin-binding peptide from Staphylococcus aureus (Penkett et al. Biochemistry 39, 2887-2893, 2000). More recently, the structure of a complex between a fibronectin-binding peptide found in Streptococcus dysgalactiae with the 1F12F1 module pair has been determined (Potts et al. Biochemistry 38, 8304-8312, 1999; Schwarz-Linek et al. FEBS Lett 497, 137-40, 2001). This structural information gives new insight into how bacteria bind to extracellular matrix proteins and has allowed predictions to be made about how several other pathogens bind to other regions of fibronectin of known stucture (Pickford et al. EMBO J 20, 1519-1529, 2001). This knowledge has implications for therapy and is the subject of a patent application. (vii) Folding Intermediates: Experimental Studies using Electron MicroscopyProteins fold in a variety of environments. Many folding experiments
are carried out in free solution following chemical denaturation
of the structure; they are refolding analyses. In vivo however,
there are two major environments where proteins fold; co-translationally
while still bound at their C-terminus to the ribosome, and within
chaperones. In fact, the mechanisms of folding are distinct in these
two cases, with co-translational folding being significantly more
rapid than chaperone-assisted refolding. This is partly at least
due to the sequential folding of domains within a multidomain protein
during its translation. We have chosen to analyze by cryo-EM an
actively translating ribosome in order to gain insights into the
mechanism by which proteins first fold, during their synthesis.
The structures we have obtained have provided a wealth of mechanistic
insight into the structural basis of protein synthesis as a dynamic
process played out through the ribosome. This analysis has been
greatly assisted by being able to fit the crystal structures of
the 50S and 30S subunits published in the past year (Bamford
et al. Curr Opin Struct Biol 11, 107-113, 2001). The fits have
revealed conformational changes in the ribosome itself due to being
in an actively translating state. We have also observed the peptidyl
site tRNA on the ribosome, attached to a nascent polypeptide chain
within the long tunnel that passes from the peptidyltransferase
centre through to the back of the 50S. The nascent protein is largely
visible as a globular mass. Careful analysis of our maps and comparison
with the crystal structure suggest that the ribosome has a chaperoning
function towards nascent proteins and the mechanism by which protein
folding is co-translationally regulated. This is the first study
of a defined actively-translating ribosome and the profound insights
it supplies suggest a number of further experiments, such as analysis
of multi-domain nascent chains to follow folding intermediates down
the tunnel and beyond. (viii) Cell Surface ReceptorsStructural studies on cell surface receptors have been a long term interest linking several of the OCMS groups. Increasingly the focus has been on direct analysis of functional complexes, rather than just the individual proteins. Recent successes have generated structural data on both high affinity interactions, for example the complex of TRAIL/DR5 (Mongkolsapaya et al. Nature Struct Biol 6, 1048-1053, 1999) and low affinity interactions (for example the dimerisation of B7-1, Ikemizu et al. Immunity 12, 51-60, 2000). Since many of the target proteins require eukaryotic expression systems a major commitment of our resources has been needed to set up dedicated tissue culture facilities and gain expertise in the use of these systems. One of the first projects to exploit this investment has been a series of structural studies of the N-terminal domain of sialoadhesin (Sn-d1) in complex with sialic acid based inhibitors. Sialoadhesin is the prototypic member of a family of human cell surface receptors that mediate their function through specific recognition of sialylated glycoconjugates. Our analysis of Sn-d1 in its functional complex with 3’ sialyllactose provided the first insights into the recognition requirements for this receptor family (May et al. Molecular Cell 1, 719-728, 1998) and triggered an interdisciplinary programme of work. The structural data allowed us to identify regions of the sialic acid molecule that could fruitfully be modified to yield ligands with enhanced binding characteristics. The most interesting of these compounds have been fed back into structural studies. To date five inhibitor-Sn-d1 complex structures have been determined (Zaccai and Jones, in preparation) and these are now guiding further drug design. These reagents are already proving invaluable tools to dissect out aspects of function, demonstrating the potential to aid functional studies by yoking together structural and chemical biology. |
© Copyright 2009 University of Oxford