Progress in analyzing proteinprotein interactions

1
Progress in analyzing protein-protein
interactions Chia Jer-ming Prasanna R Kolatkar
LinKui(BNU)
2
Structure guys

• Paaventhan Palasingam
• Jeremiah S Joseph

3
Protein Functional Dbase Protein
Interactions Rosetta Stone Text
Information/Dbase MS/yeast two-hybrid
4
(No Transcript)
5
(No Transcript)
6
Value of Kleisli/K1
Data doesnt have to be re-stored locally in a
specific format Efficient Flexible
7
Protein-Protein Interaction Database (PPiDB v
1.0)                                            
                                                  
           
8

•    Protein-Protein Interaction Queries
•    Query Interactions
• by Species 
• by Protein
• by Text Search engine.

9

• P V PROTEIN KINASE 1 PKGA lt-
  P34101Also called EC 2.7.1.-, FRAGMENT.From
  Dictyostelium discoideum.Interacting domain
  Eukaryotic protein kinase domain.

10

•    Information of Eukaryotic protein kinase
  domain
•    DB links Pfam PubMed Swissprot PDB
  Genbank DIP
•    Interacting with 55 domains
• Actin
• Adenylate and Guanylate cyclase catalytic domain
• Ank repeat
• BTK motif
• C2 domain
• CNH domain
• Cadherin domain
• Cyclic nucleotide-binding domain
• Death domain
• DnaJ domain
• Double-stranded RNA binding motif
• EF hand
• EGF-like domain
• F5/8 type C domain
• FHA domain

11

•    Interacting domains Eukaryotic protein
  kinase domain ltgt Death domain
•    Shared proteins
• death-associated protein kinase 1, also EC
  2.7.1.-, DAP KINASE 1, from Homo sapiens
• probable serine/threonine protein kinase pelle,
  also EC 2.7.1.37, from Drosophila melanogaster
• serine/threonine protein kinase rip, also EC
  2.7.1.-, CELL DEATH PROTEIN RIP, RECEPTOR
  INTERACTING PROTEIN, from Mus musculus
• serine/threonine protein kinase rip, also EC
  2.7.1.-, CELL DEATH PROTEIN RIP, RECEPTOR
  INTERACTING PROTEIN, from Homo sapiens
•    Protein pairs
• Species Bovine - Bos taurus
• activin receptor type i precursor ltgt fasl
  receptor precursor
• activin receptor type ii precursor ltgt fasl
  receptor precursor
• angiopoietin 1 receptor precursor ltgt fasl
  receptor precursor

12

• P V PROTEIN KINASE 1 PKGA lt-
  P34101Also called EC 2.7.1.-, FRAGMENT.From
  Dictyostelium discoideum.Interacting domain
  Eukaryotic protein kinase domain.

13
(No Transcript)
14

•    Interactions
• P34101protein kinase 1
• ltgt Q021581-phosphatidylinositol-4,5-bisphosphate
  phosphodiesterase
• ltgt P05987camp-dependent protein kinase
  regulatory chain
• ltgt P08796contact site a protein precursor
• ltgt P34125myosin heavy chain kinase
• ltgt P42527myosin heavy chain kinase a
• ltgt P90648myosin heavy chain kinase b
• ltgt P22467myosin ia heavy chain
• ltgt P34092myosin ib heavy chain
• ltgt P42522myosin ic heavy chain
• ltgt P34109myosin id heavy chain
• ltgt Q03479myosin ie heavy chain
• ltgt P54695myosin if heavy chain
• ltgt P54696myosin ih heavy chain
• ltgt P08799myosin ii heavy chain, non muscle
• ltgt P54697myosin ij heavy chain
• ltgt P13833myosin regulatory light chain

•    Interactions
• P34101protein kinase 1
• ltgt Q021581-phosphatidylinositol-4,5-bisphosphate
  phosphodiesterase
• ltgt P05987camp-dependent protein kinase
  regulatory chain
• ltgt P08796contact site a protein precursor
• ltgt P34125myosin heavy chain kinase
• ltgt P42527myosin heavy chain kinase a
• ltgt P90648myosin heavy chain kinase b
• ltgt P22467myosin ia heavy chain
• ltgt P34092myosin ib heavy chain
• ltgt P42522myosin ic heavy chain
• ltgt P34109myosin id heavy chain
• ltgt Q03479myosin ie heavy chain
• ltgt P54695myosin if heavy chain
• ltgt P54696myosin ih heavy chain
• ltgt P08799myosin ii heavy chain, non muscle
• ltgt P54697myosin ij heavy chain
• ltgt P13833myosin regulatory light chain

• Interactions
• P34101protein kinase 1
• ltgt Q021581-phosphatidylinositol-4,5-bisphosphate
  phosphodiesterase
• ltgt P05987camp-dependent protein kinase
  regulatory chain
• ltgt P08796contact site a protein precursor
• ltgt P34125myosin heavy chain kinase
• ltgt P42527myosin heavy chain kinase a
• ltgt P90648myosin heavy chain kinase b
• ltgt P22467myosin ia heavy chain
• ltgt P34092myosin ib heavy chain
• ltgt P42522myosin ic heavy chain
• ltgt P34109myosin id heavy chain
• ltgt Q03479myosin ie heavy chain
• ltgt P54695myosin if heavy chain
• ltgt P54696myosin ih heavy chain
• ltgt P08799myosin ii heavy chain, non muscle
• ltgt P54697myosin ij heavy chain
• ltgt P13833myosin regulatory light chain

15
Scoring System
Computational Identification Weak Same keywords
Moderate Experimental evidence from DIP or
elsewhere Good References in literature
relating names of proteins Good More than 1
support Strong
16
PPDB Current Status Putative Filtered 753508
350819
17
Christian von Mering, Roland Krause, Berend
Snel, Michael Cornell, Stephen G. Oliver,
Stanley Fields Peer Bork NATURE VOL 417 23
MAY 2002
18

Evolutionary Relationships Functional
Relationships Putative Drug Targets
19
Current BIND Database Statistics DatabaseRecord
Count Interaction Database 20000 Biomolecu
lar Pathway Database 8 Molecular Complex
Database 851 Organisms represented 12 GI
Database 4651 DI Database 0 Publication
Database 428
20
BIND Interactions RAS-GTP active form of RAS
bound to GTP RAF Y2H Homo sapiens
21
DIP DATABASE STATISTICS Number of
proteins 6963 Number of organisms 113 Number
of interactions 18059
22
Improving Quality

• Integrating high quality structural data
• Separating interaction categories
• Decreasing false positives

23
Using 3-D data

• 3-D data from PDB (Thornton,Ofran and Rost)
• Inter-domain interactions distances and
  comparison between sequence separation
• Also do for inter-protein interactions

24
Better structural analysis

• Careful analysis of the structural data needed
• Transient,Permanent,homo-oligomers,hetero-oligomer
  s
• ML could be highly useful with better
  categorization

25

• Sarah Teichman domains separated by 30 residues
  are the ones that have interaction

26

• Testing the rule
• Need a good data set to test the rule

27
X-ray crystallography

• Structural Information important for detailed and
  mechanistic understanding
• Least populated data
• Highly useful when merged with lots of functional
  information

28
(No Transcript)
29
(No Transcript)
30

• PFAM/PDB
• Single chain with multiple domains including
  complexes
• Only use non-redundant chains
• calculate distances between the domains

31

• Criteria
• 6A
• Number of contacts you can choose
• Between domains
• If xtal contact omit

32

• What did we see with the 1273 chains
• Teichman rule basically obeyed
• Exceptions
• Binding proteins
• SH3 proteins

33
(No Transcript)
34
Applying PPDB

• PPDB can be used to predict a set of interacting
  proteins.
• Intersection with Y2H studies and other methods
• Help direct structural genomics of complexes and
  improve PPDB

35
Thermatoga structural genomics

• Great model to look at a large set of complexes
• Will be useful for looking at interactions in
  other systems
• Can be used to build a database of interacting
  motifs

36
Thermatoga current state

• Crystallized several hundred and scores of
  structures
• Initial Yeast two hybrid data
• Large scale-up facilities

37
Thermatoga Y2H vs PPDB

• 15 overlap
• Improving

38
TFs

• Stem Cell totipotency TFs
• Hep B TFs
• ER TFs

39
Support

• Mass Spec
• MA
• Structure
• Other info

40
Genomic analysis

• Careful genomic data analysis can greatly
  accelerate discovery (i.e regulatory networks)

41

• Pombe kinase
• What are the possible interactions?

42
(No Transcript)
43
(No Transcript)
44
(No Transcript)
45
Which way is right?

• Strong et al Eisenberg Dec 15 NAR Functional
  linkage M. Tuberculosis
• Giot et al.Rothberg Dec 5 Science Drosophila Y2H