ENCODE: finished pilot phase (Francesca Camilli, update

1
Projects and collaborations

• ENCODE finished pilot phase (Francesca Camilli,
  update)
• Tress, M., Martelli, P.L., Frankish, A., Reeves,
  G., Wesselink J.J., Yeats, C., Olason, P.I.,
  Albrecht, M., Hegyi H., Giorgetti, A., Raimondo,
  D., Lagarde, J., Laskowski, R., Lopez, G.,
  Sadowski, M.I., Watson, J., Fariselli, P., Rossi,
  I., Nagy, A., Kai, W., Stoerling, Z., Orsini, M.,
  Assenov, Y., Blakenburg, H., Huthmacher, C.,
  Ramirez, F., Schlicker, A., Denoued, F., Jones,
  P., Kerrien, S., Orchard, S., Birney, E., Brunak,
  S., Casadio, R., Guigo, R., Harrow, J.,
  Hermjakob, H., Jones, D.T., Lengauer, T., Orengo,
  C., Patthy, L., Thornton, J., Tramontano, A.,
  Valencia, A.  
• The implications of alternative splicing in the
  ENCODE protein complement . Proc. Natl. Acad.
  Sci., In press
• Server e web service for automatic protein
  modelling Splicing variants analysis. At CRS4 in
  collaboration with Mateo Floris e Massimiliano
  Orsini
• GCSF and beta-interferon collaboration with
  BioKer and Maria Valentini (crs4) MD
  simulations and protein-protein docking. Next
  Final analysis of the bioinformatics part, and
  results comparison with the experimental
  counterpart.
• ITCH collaboration with group of Prof. Gerry
  Melino. Modelling, MD simulations, Normal Modes
  analysis, phosphorylation sites analysis. Next
  Experimental counterpart, new analysis of the
  interface and design of new experiments.
• Ion channels collaboration with C. Michelletti
  and P. Carloni groups at SISSA. Loop modelling,
  normal modes analysis. Next Analysis of how loop
  conformations influence the channels.
• Cancer project Cargo interface, modelling,
  mutations mapping. Next check of automatic
  models and mapping of mutations (annotation on
  PMDB).
• BABP protein collaboration with Henriette
  Molinaris group. Bioinformatic searches.
• Guariento M. , Raimondo D., Assfalg M., S.
  Zanzoni S, Esente P. , Ragona L. ,Tramontano A.
  and Molinari H .
• Identification and functional characterization
  of the bile acid transport proteins in
  non-mammalian ileum and mammalian liver. Proteins
  2007, in press
• Glycodelin proteins collaboration with Henriette
  Molinaris group. Modelling, protein-protein
  interaction. Next Dimer interface analysis and
  design of experiments.

2
Cagliari 31-05-2007Project ITCH/p73

• Alejandro Giorgetti
• Domenico Raimondo
• Anna Tramontano

3
p73

• Structural and functional homologue of p53, able
  to transactivate the promoters of genes involved
  in apoptosis and cell cycle regulation.
• p53 is the most frequently mutated and intensely
  studied tumor suppressor gene.
• After DNA damage or proto-oncogene activation,
  p53 is stabilized by ubiquitylation and exerts
  its anti-tumorigenic activity by inducing cell
  cycle arrest or apoptosis.
• p73 p53 Important difference in the
  ubiquitylation process different E3

4

• Ubiquitin
• Proteins are usually tagged for selective
  destruction in proteolytic complexes called
  proteasomes by covalent attachment of ubiquitin,
  a small, compact, highly conserved protein. 

However, some proteins may be degraded by
proteasomes without ubiquitination. An isopeptide
bond links the terminal carboxyl of ubiquitin to
the e-amino group of a lysine residue of a
"condemned" protein.
5

• .Three enzymes are involved, designated E1, E2
  E3.
• Initially the terminal carboxyl group of
  ubiquitin is joined in a thioester bond to a
  cysteine residue on Ubiquitin-Activating Enzyme
  (E1). This is the ATP-dependent step.
• The ubiquitin is then transferred to a sulfhydryl
  group on a Ubiquitin-Conjugating Enzyme (E2).

6

• A Ubiquitin-Protein Ligase (E3) then promotes
  transfer of ubiquitin from E2 to the e-amino
  group of a Lys residue of a protein recognized by
  that E3, forming an isopeptide bond. 
• There are many distinct Ubiquitin Ligases with
  differing substrate specificity. 
• One E3 is responsible for the N-end rule.
• Some are specific for particular proteins.

7

• More ubiquitins are added to form a chain of
  ubiquitins.
• The terminal carboxyl of each ubiquitin is linked
  to the e-amino group of a lysine residue (Lys29
  or Lys48) of the adjacent ubiquitin. 
• A chain of 4 or more ubiquitins targets proteins
  for degradation in proteasomes. (Attachment of a
  single ubiquitin to a protein has other
  regulatory effects.)

8

• Ubiquitin Ligases (E3) mostly consist of two
  families
• Some Ubiquitin Ligases have a HECT domain
  containing a conserved Cys residue that
  participates in transfer of activated ubiquitin
  from E2 to a target protein.
• Some Ubiquitin Ligases contain a RING finger
  domain in which Cys His residues are ligands to
  2 Zn ions.

9
Model of Ubiquitin Transfer and Ubiquitin Chain
Elongation by HECT Domain Ubiquitin Ligases
HECT
UbcH7
10
Evolution of protein structure families
90
Drug design?
70
Biochemistry?
identical
X-ray cristallography MR
50
30
Molecular Biology?
10
identical
Chothia Lesk (1986)
11
Comparative Modeling
Known Structures (templates)
Template(s) selection
Target sequence
Sequence Alignment
Structure Evaluation
gthTEII MSSPQAPEDGQGCGDRGDPPGDLRSVLVTTV LNLEPLDEDLF
RGRHYWVPAKRLFGGQIVGQ ALVAAAKSVSEDVHVHSLHCYFVRAGDPK
LP
Structure Modeling
Final Structural Models
12
Comparative Modeling
Known Structures (templates)
Template(s) selection
Target sequence

• Protein Data Bank PDB http//www.pdb.org
• Banca Dati dei templati
• Separare in singole catene
• Controllare la qualità delle strutture

Sequence Alignment
Structure Evaluation
Structure Modeling
Final Structural Models
13
Comparative Modeling
Known Structures (templates)
Template(s) selection
Target sequence

• Similarità di sequenza / Fold recognition
• Analisi della struttura (risoluzione, metodo
  sperimentale
• Ci sono altri atomi e/o composti? Sono legati?

Sequence Alignment
Structure Evaluation
Structure Modeling
Final Structural Models
14
Comparative Modeling
Known Structures (templates)
Template(s) selection
Target sequence

• Fondamentale per la modellizzazione per omologia.
• Allineamento globale
• Un piccolo errore nellallineamento può essere
  fatale per il modello.
• Ricordatevi gli allineamenti a coppie
  sussurrano, quelli multipli parlano ad alta voce.
• Sappiamo qualcosaltro? Ci sono sperimenti?

Sequence Alignment
Structure Evaluation
Structure Modeling
Final Structural Models
15
Comparative Modeling
Known Structures (templates)
Template(s) selection
Target sequence
Sequence Alignment
Structure Evaluation

• Assemblaggio di frammenti (Template based
  fragment
• Assembly - SwissMod).
• Minimizzazione della deviazione dai vincoli
  spaziali (Satisfaction of Spatial Restraints
  MODELLER )

Structure Modeling
Final Structural Models
16
Comparative Modeling
Known Structures (templates)
Template(s) selection
Target sequence

• Errori nella selezione dei templati
• Cicli iterativi di allineamento,
  modellizzazione e valutazione.

Sequence Alignment
Structure Evaluation
Structure Modeling
Final Structural Models
17
X-Ray
Orazio Romeo (master Sardinia)
Ubiquitin ligases (E3) act together with the
ubiquitin activating enzyme (E1) and the
ubiquitin conjugating enzyme (E2) to catalyze
protein ubiquitylation
Used template1nd7 80 ID
18
E2 UbcH7 (X-ray)
19
Analysis of the interaction surface (molmol) and
3.5 ns MD simulations (NAMD)
20
Binding Interface
21
Normal modes analysis
22
Normal modes analysis Beta-gm program
23
Putative hinge regions
24
C-lobe

• Two hinge regions found with beta-gm (gaussian
  model)
• C-lobe
• N-lobe small subdomain

N-lobe
25
Beta-GM program

• Normal modes of the complex analysis Vibrational
  modes at low frequencies.
• Normal modes from a MD simulation tens of
  nanosecods.
• Beta-gm program implements a coarse-grained model
  to describe the dynamics of the protein
  (ß-Gaussian network model).
• Provides a reliable (by comparison against full
  atom MD simulations) description of concerted
  large-scale rearrangements in proteins.
• The concerted motions are calculated within the
  quasi-harmonic approximation of the free energy,
  F, around a protein's native state.
• A displacement from the native state dRdr1,
  dr2,...drn (ri being the displacement of Ca atom
  i) is associated with the change in free energy
• ?F (½)dRF dR
• Where F is an interaction matrix constructed from
  the knowledge of contacting Ca and Cß centroids
  in the native state.
• The large-scale motions of the system correspond
  to the eigenvectors of F having the smallest
  nonzero eigenvalues.

26
I. Template based fragment assembly
d) Minimizzazione della energia

• Il processo di modeling produrrà contatti
  ravvicinati fra atomi, e lunghezze di legame
  sfavorevoli.
• ? Riuscire ad avere le geometrie giuste
• Minimizzazione della energia troppo estensiva,
  può allontanarci dalla vera struttura.
• SwissModel utilizza GROMOS 96 force field

27
Eelectrostatic . The electrostatic energy is
evaluated by using the Restrained
Electrostatic Potential (RESP) partial charges.
These charges have the properties of
accurately reproduce the electrostatic potential
multipoles outside the molecule, and they
were calculated in the following way. Ab initio
quantum chemical calculations are performed on
small molecules and the electrostatic potential
j V are calculated on M grid points outside the
molecule.
28
II. Modeling by Satisfaction of Spatial restraints

• Derivate per omologia Ottenute dal
  allineamento.
• Stereochimiche Set di parametri di CHARMM
  parameter - MacKerell et al., 1998 ).
• Energie di Van der Waals e Coulomb dal campo di
  forza CHARMM.
• Esterne Vincoli di distanze esterne.

• Trovare la struttura più probabile a
• partire da un allineamento
• Utilizza probability density functions.
• Minimizza deviazioni dai vincoli.

Comparative protein modeling by satisfaction
of spatial restraints. A. Šali and T.L. Blundell.
J. Mol. Biol. 234, 779-815
29
Cancer Project

• Domenico Raimondo
• Alejandro Giorgetti

30
Sjoblom et al.The consensus coding sequences of
human breast and colorectal cancers. Science.
2006

• From this screening our initial set of sequences
  consisted of 189 (CAN genes) 122 breast genes
  and 69 colorectal ones (two genes overlap), for a
  total of 535 peptides.
• Steps for initial analysis
• Blast search on PDB (In red Genes for that had
  not been strongly suspected to be involved in
  cancer).
• Blast search on BIND database.
• Semiautomatic modelling (hhpred- toolkit and
  visual inspection).
• Submission to PMDB.
• Next Annotation of the mutations directly on
  PMDB.
• Widget for the Cargo web server.

31
Encode
Bioinformatica delle proteine 'Proteine al lavoro'

• Modelli sottomessi su PMDB 30 (20 a 97 ID
  seq. e copertura totale).
• Sequenze con struttura risolta 25 (sempre meno
  di 50 aa mancanti).
• Mappaggio domini-esoni. Solo 3 strutture trovano
  corrispondenza a meno di 5 aa.
• Modelli di varianti di splicing 70
• 47 sequenze e le sue varianti si splicing sono
  state analizzate.16 non hanno splicing
  alternativo, 11 hanno splicing show alternativo
  nelle regioni non codificanti, e 20 hanno
  varianti di splicing (in generale 2-3).

32

• Copertura PARZIALE (41 trascritti)
• Meno di 50 (30) aa (al 5')
• 11 con informazione di struttura
• 4 hanno varianti di splicing
• 2 sequenze identiche
• 1 esone interno mancante (non
  sembra una struttura possibile)
• 1 esone interno mancante (fuori
  copertura X-ray)
• Più di 50 aa (150- 300) mancanti
• 8 con informazione di struttura
• 6 hanno varianti di splicing
• 2 sequenze identiche
• 4 esoni interni mancanti (fuori
  copertura X-ray)
• Copertura TOTALE (39 trascritti)
• 22 hanno varianti di splicing (2 a 4)
• 5 sequenze identiche ()
• 2 esoni alternativi al 5'
• 7 esoni alternativi interni

33
Encode
Bioinformatica delle proteine 'Proteine al lavoro'
AC004039.4 - 001 -002
AC069356.1 - 001 -002
34
Encode Risultati
35
Discussione

• Unlike most evolutionarily related sequences the
  splice isoforms in this set are sequence
  identical except for single deletions or
  insertions Many of these are relatively large.
• Cambiamenti al C-terminal e al N-terminale
  tendono ad essere swaps. Cambiamenti interni
  delezioni.
• In 73 (22) casi le strutture PDB avranno delle
  modifiche dovute alle inserzioni o delezione
  nelle varianti di splicing.
• In 49 (19) casi ci deve essere un grosso
  refolding
• In 24 (3) casi gli effetti nella struttura
  dovranno essere piccoli.
• 994 sequenze hanno un dominio PFAM.
• 42.5 (423 sequenze) hanno un dominio PFAM che è
  diviso in due.
• 53 isoforme hanno due domini interrotti broken
  domains
• 3 sequenze 3 domini sono state splited have
  been split.