1. Molecular Interaction Maps:
Mirit I. Aladjem
Laboratory of Molecular Pharmacology, NCI
Molecular Interaction Maps:
Circuit
Diagrams for Bioregulatory Networks
Kind of networks we wish to describe?

2. Cells use bioregulatory networks to assess:
Their environment (growth factors, proximity of other cells in a tissue, biophysical parameters)
Their metabolic status (energy levels, availability of key components such as nucleotides)
Then, the network will aid in making decisions:
Cell cycle progression (replication, mitosis, senescence)
Programmed cell death (apoptosis)
Activation of a gene expression program (e.g. differentiation)
The challenge:
Organize information about complex networks in a concise graphical manner while presenting sufficient detail to describe models for simulation.
NCI interested in how cells regulate their growth so

3. Specific Features of Regulatory Interaction Networks
There are several different types of possible interactions
Interactions may involve distinct intracellular domains
Interactions may affect specific intramolecular domains
Each component of the network may interact with several other components
Interactions may affect the ability to form other interactions
Requirement: all the pertinent interactions involving each component should be traceable on the diagrams
Molecular Interaction Maps depict bioregulatory interactions unambiguously in diagram form using specific lines and nodes.

4. Multi-protein complexesA B C
The complex of (A&B) & C
primary molecular species
Each molecule appears only once per diagram. Interaction outcomes - complexes or modified molecules - are depicted as nodes on the interaction lines.

5. Covalent modification (e.g., protein phosphorylation)
P = PO4, a phosphate group
Inhibitory phosphorylation:
Phosphorylation of A blocks the kinase activity of A.
Inhibition convention B P
Catalysis line convention
color not important

6. obligatory requirement
Activating phosphorylation:
The phosphorylated form of kinase A is the active form,
phosphorylates B P A
Bar added to indicate
obligatory requirement B P

7. Interaction types Reactions Contingencies
Catalysis
Stimulation
required
Inhibition
Binding (non-covalent)
Covalent Modification
(e.g. phosphorylation)
Bond cleavage
(e.g. Phosphatase)
Reactions operate on molecular species; contingencies operate on reactions or on other contingencies. Reaction outcomes (nodes) are treated as molecular species.
Stochiometric
Conversion (A to B) A B
Interaction types are depicted by different arrowheads
Second set of reactions derivatives of the first
Transcription/translation
Degradation
Dimerization A
Transport

8. Interaction relationships Multiple Substrates : Boolean Functions :
A can phosphorylate B and C
a AND b AND c required
a OR b OR c required
x = complex of A and B
y = complex of A and C
z = either x or y
Competition :
interatction relationships - the angle by which interaction lines touch each other
a OR (b AND c) required
*Lines crossing at right angles = no relationships

9. Assembly of a multimolecular complex:
ORC, the origin recognition complex
(involved in cell cycle regulation)
Cyclin A/B
Cdk1 P D N A ( o r i )
These two complexes can bind DNA and play roles in cell cycle regulation P

10. Intramolecular domains (two compartments)
EGFR example for intramolecular binding
Reagent binding
How we depict membranes

11. Dimerization
Signal from the outer membrane conveyed inside

15. transfer to the nucleus

17. An Explicit MIM

18. Translation of the Explicit MIM: a Reaction Table and ODEs Based on Mass Action Laws. . . . .

19. An Explicit MIM Depicts Molecular Interactions With Sufficient Detail Required for Simulation

20. A Heuristic MIM
four types of EGFR and more than one pathway - not assigning order hence Heuristic
Heuristic = “an aid to discovery”

21. Heuristic MIMs aim to organize information about complex networks in a concise manner, as an aid to discovery

22. MIMs depict what molecules “see” - potential interactions between depicted molecular species. Heuristic MIMs can describe complex networks of potential interactions without encountering “combinatorial explosion”.

23. Combinatorial Explosion
48 Receptor-ligand-dimer combinations;
96 phosphorylation reactions involving a single tyrosine residue (Y1068);
144 phosphorylation reactions involving two tyrosines (Y1148/73)
A simple model for SOS recruitment by EGFR enumerates 3749 individual reactions (Blinov et al., 2005)
each receptor a combination of “states” sometimes there is only one state or a group of state that are capable of the next interaction

24. Can biological processes be inferred from the potential interactions described in heuristic MIMs?

25. What does the ORC complex do?
Cyclin A/B
Cdk1 D N A ( o r i )

26. Role of ORC in events leading to DNA synthesis
note the dotted lines not strictly part of the annotation but useful

27. ORC binds DNA during early G1
highlighting only part of the map - again not part of the annotation but illustrative G1

28. The ORC-DNA complex binds adaptor molecules
adapter molecules G1 G1

29. Adaptor molecules recruit a helicase (inactive)
second complex, note that it binds only the second complex of ORC not the first.
S-phase G1
Helicase

30. Kinases Activate the Helicase and Start Replication
Inhibitory
Kinases
Activating
Kinases G1
higher level - control S

31. More Information: Annotated MIMs and e-MIMs
Annotation
links to external pertinent info

35. Inferring Regulatory Pathways from eMIMs
this map does not exist

36. Cells Do Not Replicate DNA During Mitosis
only the kinase partner cyclin B1 is present

37. Cells Inhibit Replication In Response to DNA Damage

38. Cells Inhibit Replication In Response to DNA Damage

39. Heuristic MIMs can help elucidate the control principles underlying biological pathways and contributing to their robustness and sensitivity

40. Coherent Biological Processes Inferred from a Single MIM
PT68
PS516
ATM
Chk2
PS15/20
AcK373 P
PS17
PY394
c-Abl Ub
p53
Sometimes several biological pathways in the network lead to the same outcome
Mention promoter
p300
Mdm2

41. Coherent Biological Processes Inferred from a Single MIM
PY394
PS516
p53
PS15/20
AcK373
c-Abl P
PT68
PS17
ATM
Chk2
c-Abl Ub
p300
Mdm2

42. Coherent Biological Processes Inferred from a Single MIM
PY394
PS516
p53
PS15/20
AcK373
c-Abl P
PT68
PS17
ATM
Chk2
c-Abl Ub
p300
Mdm2

43. Coherent Biological Processes Inferred from a Single MIM
PY394
PS516
p53
PS15/20
AcK373
c-Abl P
PT68
PS17
ATM
Chk2
c-Abl Ub
add to robustness of the pathway
p300
Mdm2

44. Transcriptional Regulatory Complexes Depicted in MIM
transcriptional control a combinatorial process; two partners of myc two kinds of promoters
S. Pasa

45. Depiction of intramolecular interactions in MIMs:
SRC Activation By Receptor Tyrosine Kinases

46. Src has the following domains:
SH3: binds to proline-rich domains
SH2: binds to phosphotyrosine motifs
Pro: proline-rich domain
kinase: tyrosine kinase domain
The SH3 domain binds the Pro domain.

47. The Src tail region can be phosphorylated at Tyr 527.

48. The SH2 domain binds to the tyrosine-phosphorylated tail.

49. The 2 intra-molecular bonds form cooperatively, and fold the Src molecule, hiding the kinase domain and keeping Src in an inactive configuration.

50. Src’s tyrosine kinase domain could phosphorylate various substrates

51. Phosphorylation of Tyr416 is required for the kinase to be active.
Requirement contingency

52. Activated (phosphorylated) EGFR could phosphorylate Tyr416

53. However, access to Tyr416 is blocked by the intra-molecular folding.

54. Through one of its phosphotyrosines, activated EGFR recruits p85.
SRC activation:
Through one of its phosphotyrosines, activated EGFR recruits p85.

55. Pro domain of p85 competes with Pro of Src for binding to SH3 of Src
If binding is to p85, then this inhibition is relieved

56. Phosphotyrosine of EGFR competes with P-Tyr527 of Src for binding to SH2 of Src
Then EGFR can phosphorylate Tyr416
And activate SRC
If binding is to EGFR, then this inhibition is relieved

57. Then a phosphatase can remove Tyr527 . . .
With Tyr527 gone, Src cannot refold and remains active even if it dissociates from the EGFR:p85 complex.
Thus multiple Src’s can be activated by a single active EGFR -- i.e., an amplification step.
Then a phosphatase can remove Tyr
. . . because Tyr527 is no longer blocked by binding to the SH2
Kohn, K.W Chaos 11:84-97

58. A dynamic animated map of Src activation by EGFR
Inactive Src
Active Src

60. MIMs address the challenges of depicting complex networks:
MIMs can depict:
different types of bioregulatory reactions
contingencies affecting such reactions
intra-molecular interactions
Explicit MIMs describe models suitable for simulation
Heuristic MIMs summarize large data sets with different levels of detail
eMIMs provide links to pertinent external information
MIMs are useful to represent the combinatorial complexity of biological networks.
we seek to develop tools that will simplify the production and modifications of MIMs to allow wide adoption of this tool by the community

61. The MIM Team NCI: Kurt W. Kohn Sohyoung Kim Yves Pommier
John Weinstein
David Kane
Margot Sunshine
Hong Cao
NIST:
Geoffrey McFadden
Italian Cancer
Institute:
Sylvio Parodi
Stefania Pasa

62. Kitano et al., 2005; Oda et al., 2005