Protein structure in molecular systems biology Dr DAVID SHEEHAN, PROTEOMICS RESEARCH GROUP, DEPT. BIOCHEMISTRY, UNIVERSITY COLLEGE CORK, IRELAND McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 THE ECOTOXICOLOGY TRIANGLE POLLUTANTS TOXICOLOGY BIOTRANSFORMATION BIOCONCENTRATION BIOTA SPECIATION DISTRIBUTION SEASONALITY CLIMATE McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 ENVIRONMENT SENTINEL SPECIES IN ECOTOXICOLOGY • CHOSEN AS REPORTERS OF STATUS • ROBUST • EASILY-RECOGNISED • ABUNDANT • SESSILE RATHER THAN MOBILE • WIDE GEOGRAPHICAL DISTRIBUTION McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 SENTINELS - ADVANTAGES • HISTORY OF USE • MULTI-ORGAN MODEL SYSTEMS • PREDICT FOR TOXICITY IN HUMANS • OFTEN BIOCONCENTRATE POLLUTANTS McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 SENTINELS - DISADVANTAGES • MAY HAVE UNIQUE UPTAKE CHARACTERISTICS • MUCH SIMPLER SYSTEMS THAN MAMMALS • POORLY-REPRESENTED IN SEQUENCE D-BASES McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 MUSSELS – GENUS MYTILUS • • • • • • • FILTER FEEDERS SESSILE ABUNDANT WIDELY-DISTRIBUTED RESILIENT TO POLLUTION HISTORY OF USE “MUSSELWATCH” PROGRAMMES McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 SYSTEMS BIOLOGY PROTEIN-PROTEIN INTERACTIONS PROTEOME METABOLOME DNA mRNA PROTEIN TRANSCRIPTOME GENOME METABOLITES GLYCOPROTEOME MODIFIED PROTEINS PHOSPHOPROTEOME REDOX PROTEOME McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 MYTILUS IS POORLY-REPRESENTED IN D-BASES NUMBER OF ENTRIES (NCBI) MYTILUS MOUSE GENES ~ 30,000 33,000 EST 29,102 926,165 PROTEIN SEQ.S 1,768 139,018 STRUCTURE (PDB) 9 1,262 McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 Beta-1,4-D-Endoglucanase Cel45a From Blue Mussel Mytilus Edulis X-Ray Structure Of Beta-Mannanase From Blue Mussel Mytilus Edulis Crystal Structure Of Phosphoenolpyruvate Mutase – 6 str.s Mediterranean Mussel Defensin Mgd-1 McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 SOME IMPLICATIONS OF POOR D-BASE REPRESENTION • SPOT-MATCHING FOR MS IS DIFFICULT • LACK OF STRUCTURAL KNOWLEDGE THROUGH PROTEIN HIERARCHY • DIFFICULT TO LOCATE PROTEIN MODIFICATIONS (e.g. OXIDATION) • PROTEIN-PROTEIN INTERACTION? McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 OVERVIEW OF THIS PRESENTATION • SEQUENCE-INDEPENDENT PROBES FOR REDOX LESIONS IN PROTEINS • SEQUENCE SIMILARITY IN SPOT MATCHING • USE OF ON-LINE HOMOLOGY MODELS McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 SEQUENCE-INDEPENDENT PROBES FOR REDOX LESIONS IN PROTEINS McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 Oxidative stress GSH CAT SOD GPX H202 ANTIOXIDANT O2 + .O - OH2 .OOR . OH DEFENCE ROS McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 PROTEIN EXPRESSION SIGNATURE APPROACH 3 pH IMAGE ANALYSIS SOFTWARE 10 70 kDa Mr UP-REGULATED 17 kDa NO CHANGE McKim Conference on Predictive Toxicology, Duluth MN, Sept. 25DOWN-REGULATED 27th 2007 PROTEIN OXIDATION COMPLICATES THE PROTEOME • PROTEINS ABSORB ~ 70% OF ROS • CARBONYLATION (R → ALDEHYDE/KETONE) • UBIQUITINATION • NITROSYLATION • CYS OXIDATION …. Dowling & Sheehan D. (2006) Proteomics 6, 55975604. Sheehan D. (2006) Biochemical and Biophysical Research Communications 349, 455-462. McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 4 66 pI 7 1 4 pI 7 2 45 31 21.5 3 4 66 45 31 21.5 C = O + HYDRAZINE - DNP Blot Name (Anti-DNP) Spot s Matc hed Matc h Rate *Menadione Gill 33 33 100% Control Gill 16 13 81% H2O2 Gill 64 28 40% CdCl2 Gill 37 25 67% ANTI-DNP HYDRAZONE - DNP BLOT PROTEIN CARBONYLATION:ANIMALS EXPOSED TO PRO-OXIDANTS MYTILUS EDULIS 2-D CARBONYLS __________________________ McDonagh & Sheehan, 2006 Aquat. Toxicol. 79, 325-333 McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 REDOX LESIONS: UBIQUITINYLATION __________________________ McDonagh & Sheehan, 2006 Aquat. Toxicol. 79, 325-333 Ub ANTI-Ub Ub McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 Ub Ub BLOT REDOX LESIONS: NITROSOTHIOLS AND THE BIOTIN SWITCH ASSAY • NITROSOTHIOLS ARE REDUCED BY ASCORBIC ACID BUT DISULPHIDES ARE NOT – SPONTANEOUSLY HYPERTENSIVE RAT KIDNEY MEDULLA SS SNO ASCORBIC ACID SS STREPTAVIDIN McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 SH SS NEM - BIOTIN S-NEM-BIOTIN REDOX LESIONS: SULPHENIC ACID (SOH) PROTEINS PROTEIN EXTRACT SH SOH SH NEM - BIOTIN ARSENITE SOH NEM -S – S- S-NEM S-NEM -S – S- -S – SDTT SOH S-NEM -SH SH- 2D LC-TANDEM MS ATS SH- IDENTIFICATION -SH 2DE . . . . . IMAGE ANALYSIS AFFINITY SELECTION SOH S-NEM FLOW-THROUGH McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 TOXICOLOGY OF NANOPARTICLES IN E. COLI – GOLD AS ANTIOXIDANT -SH- IODOACETAMIDE FLUORESCEIN PROTEIN MENADIONE GOLD McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 CONTROL REDOX LESIONS: NITROSOTYROSINES SS YNO ANTI -3NITROTYROSINE SS YNO LC-TANDEM MS Tyther et al., 2007 Proteomics – Clinical Applications. IN PRESS IDENTIFIED 22 PROTEINS McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 SEQUENCE SIMILARITY IN SPOT MATCHING McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 Cork harbour PROTEINS 7 days O CH 3 DISSECT GILL O 24 hr +/- 1 mM menadione McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 ON OXIDATION…. -SH + IAF -S-IAF A -SH “DISAPPEAR” PDI hsp gp96 45 kDa hmb calreticulin Protease Serine 1 B -S-S “APPEAR” PDI beta alpha 2 tubulin tubulin 45 kDa gelsolin enolase GDP diss. inhibitor GST Pi RNA binding protein transferrn Control Menadione McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 SPOT IDENTIFICATION TRYPTIC DIGEST MS 1 HPLC PROTEIN SPOT ID. PEPTIDES PEPTIDE MS 2 TANDEM MS (i.e. MS/MS) 1 m/z SCREEN SEQ D-BASES McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 FRAGMENT n m/z AMINO ACID SEQUENCE Spot no. Protein Name Species Function MOWSE 1 Heat shock protein gp96 Strongylocentrotus purpuratus (Sea Squirt) Chaperone 260 2 PDI Conus marmoreus(Cone Snail) Protein folding 149 3 Calreticulin Meloidogyne incognita (cotton root-knot nematode) Binds misfolded proteins 69 4 Heavy metal binding protein M. edulis (Blue mussel) Heavy metal binding 227 5 Heavy metal binding protein M. edulis Heavy metal binding 297 6 Heavy metal binding protein M. edulis Heavy metal binding 319 7* Unamed Tetraodon nigroviridis (pufferfish) 8 No identification 9* Protease serine 1 Mus musculus (Mouse) McDonagh & Sheehan, 2007, Proteomics, 7, 3395-3403 McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 34 Cell control 31 USE OF ON-LINE HOMOLOGY MODELS McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 PROCEDURE FOR HOMOLOGY MODELLING TARGET SEQUENCE TEMPLATE STRUCTURE SCRs IDENTIFY STRUCTURALLY CONSERVED REGIONS LOOPS CREATE DATABASE OF LOOPS FROM ELSEWHERE IN PDB ENERGY MINIMISE HOMOLOGY MODEL McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 ON-LINE HOMOLOGY MODELLING WITH Geno3D COMBET et al., “Geno3D an automated protein modelling web server, BIOINFORMATICS, 2002, 18, 213-214 http://pbil.ibcp.fr/htm/ McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 ON-LINE HOMOLOGY MODELLING WITH Geno3D • START AT PROTEIN SEQUENCE DATABASE • AMINO ACID SEQUENCE IS INPUT FOR ON-LINE HOMOLOGY MODELING PROGRAM e.g. GENO3D. • SELECT TEMPLATE (PSI-BLAST) • RESULTS SENT VIA e-MAIL. McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 CATALASE FROM M. Californianus TEMPLATE: HUMAN ERYTHROCYTE CATALASE PDB 1F4J McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 CATALASE FROM M. Californianus McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 MODEL QUALITY McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 STRUCTURAL ANALYSIS MAIN CHAIN PARAMETERS ARE COMPARED TO A STRUCTURE OF 2Å RESOLUTION McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 ROOT MEAN SQUARE DEVIATION a-Cs TEMPLATE x1 x2 x4 x3 x5 X6 .. RMSD = MODEL McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 x2 n PROTEIN (% Id) TARGET (ACCESSION ) TEMPLATE (PDB ID) RMSD (Å) (% Id) RAMACHANDRAN (% DISALLOWED) MODEL HISTONE H3 AAP94664 1KX5 6.86 (98.5) 4.2 CYT. C OXIDASE AAT98405 1V54 11.01 (58.4) 1.1 TROPOMYOSIN BAA19209 2TMA 23.94 (50.9) 6.1 RMSD IS ONLY LOOSELY RELATED TO % SEQ. ID. McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 RAMACHANDRAN ANALYSIS AMIDE PLANE C PHI PSI N SER ARG McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 LACTOPEROXIDASE: OVERLAY OF MODEL ON TEMPLATE (PyMOL) MODEL TEMPLATE MOVIE!! McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 OTHER MODELS METHIONINE ADENOSYLMETHYL TRANSFERASE LYSOZYME P-GLYCO PROTEIN CYTOCHROME C OXIDASE MDR ASSOC.D PROTEIN SUPEROXIDE DISMUTASE McKim Conference on Predictive PHOSPHOFRUCTOKINASE Toxicology, Duluth MN, Sept. 2527th 2007 HEAVY METAL B.P. P-450 CYP4Y1 TWITCHIN RELEVANCE OF HOMOLOGY MODELS • SIGNAL TRANSDUCTION -> “NETWORKS OF PROTEINS” • GUIDE FOR STRUCTURE-FUNCTION • MODEL SOLVENT ACCESSIBILITY • PROTEIN-PROTEIN INTERACTIONS • CAVEATS!! – MODELS ARE NOT STRUCTURES – SIDE-CHAINS McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 CONCLUSIONS • PROTEIN MODIFICATIONS NEED TO BE MODELLED • MORE SEQUENCE DATA ARE REQUIRED FOR NON-STANDARD SPECIES • THREE-DIMENSIONAL MODELS ARE POSSIBLE WHERE 3D STRUCTURE ARE LACKING McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007 FUNDING/ACKNOWLEDGEMENTS • Prog. Res. Third Level Institutions; • IRCSET; EPA; Health Research Board Collaborators: Dr Tim Veenstra (NCI, NIH); Drs Ian Davidson, Phil Cash (Aberdeen); Prof Maria Bebiano (Faro) • Drs. Vera Dowling, Brian McDonagh, Ray Tyther • Sara Tedesco, Ebenezer Kanagaraj, Hu Wentao McKim Conference on Predictive Toxicology, Duluth MN, Sept. 2527th 2007