BioNetGen: a system for modeling the dynamics of protein-protein interactions

1. BioNetGen: a system for modeling the dynamics of protein-protein interactions Bill Hlavacek Theoretical Biology and Biophysics Group Los Alamos National Laboratory

2. Outline • The biochemistry of cell signaling and combinatorial complexity • The conventional approach to modeling • The rule-based approach to modeling • Tools

3. Multiplicity of sites and binding partners gives rise to combinatorial complexity Epidermal growth factor receptor (EGFR)

4. Multiplicity of sites and binding partners gives rise to combinatorial complexity Epidermal growth factor receptor (EGFR) 9 sites  29=512 phosphorylation states

5. Multiplicity of sites and binding partners gives rise to combinatorial complexity Epidermal growth factor receptor (EGFR) 9 sites  29=512 phosphorylation states Each site has ≥ 1 binding partner  more than 39=19,683 total states

6. Multiplicity of sites and binding partners gives rise to combinatorial complexity Epidermal growth factor receptor (EGFR) 9 sites  29=512 phosphorylation states Each site has ≥ 1 binding partner  more than 39=19,683 total states EGFR must form dimers to become active  more than 1.9108 states

7. Signaling proteins typically contain multiple phosphorylation sites > 50% are phosphorylated at 2 or more sites Source: Phospho.ELM database v. 3.0 (http://phospho.elm.eu.org)

8. Outline • The biochemistry of cell signaling and combinatorial complexity • The conventional approach to modeling • The rule-based approach to modeling • Tools

9. Early events in EGFR signaling EGF = epidermal growth factor EGFR = epidermal growth factor receptor EGF 1. EGF binds EGFR ecto EGFR

10. Early events in EGFR signaling EGF 1. EGF binds EGFR dimerization 2. EGFR dimerizes EGFR

11. Early events in EGFR signaling EGF 1. EGF binds EGFR 2. EGFR dimerizes Y1092 P P 3. EGFR transphosphorylates itself Y1172 P P EGFR

12. Early events in EGFR signaling Grb2 pathway EGF 1. EGF binds EGFR 2. EGFR dimerizes Y1092 P P 3. EGFR transphosphorylates itself Grb2 P P SH2 4. Grb2 binds phospho-EGFR EGFR

13. Early events in EGFR signaling Grb2 pathway EGF 1. EGF binds EGFR Grb2 2. EGFR dimerizes Y1092 P P 3. EGFR transphosphorylates itself Sos P P SH3 4. Grb2 binds phospho-EGFR EGFR 5. Sos binds Grb2 (Activation Path 1)

14. Early events in EGFR signaling Shc pathway EGF 1. EGF binds EGFR 2. EGFR dimerizes P P 3. EGFR transphosphorylates itself Y1172 P P 4. Shc binds phospho-EGFR EGFR Shc PTB

15. Early events in EGFR signaling Shc pathway EGF 1. EGF binds EGFR 2. EGFR dimerizes P P 3. EGFR transphosphorylates itself Y1172 P P P 4. Shc binds phospho-EGFR EGFR Shc Y317 5. EGFR transphosphorylates Shc

16. Early events in EGFR signaling Shc pathway EGF 1. EGF binds EGFR 2. EGFR dimerizes P P Shc 3. EGFR transphosphorylates itself Y1172 P P P 4. Shc binds phospho-EGFR EGFR Grb2 SH2 5. EGFR transphosphorylates Shc 6. Grb2 binds phospho-Shc

17. Early events in EGFR signaling Shc pathway EGF 1. EGF binds EGFR 2. EGFR dimerizes P P Sos Grb2 Shc 3. EGFR transphosphorylates itself Y1172 P P P 4. Shc binds phospho-EGFR EGFR SH3 5. EGFR transphosphorylates Shc 6. Grb2 binds phospho-Shc 7. Sos binds Grb2 (Activation Path 2)

18. A reaction-scheme diagram Species: One for every possible modification state of every complex Reactions: One for every transition among species This scheme can be translated to obtain a set of ODEs, one for each species

19. Combinatorial complexity of early events Monomeric species 2 states 4 states 48 species 6 states EGFR

20. Combinatorial complexity of early events Monomeric species 2 states 4 states 48 species 6 states EGFR Dimeric species EGF N(N+1)/2 = 300 species 24 states

21. 5 proteins 18 species34 reactions A conventional model for EGFR signaling The Kholodenko model* *J. Biol. Chem. 274, 30169 (1999)

22. Assumptions made to limit combinatorial complexity Phosphorylation inhibits dimer breakup Bottleneck for dimers No modified monomers P P P

23. Assumptions made to limit combinatorial complexity Adaptor binding is competitive No dimers with more than one site modified P P P P P

24. Assumptions made to limit combinatorial complexity Phosphorylation inhibits dimer breakup Adaptor binding is competitive Experimental evidence contradicts both assumptions.

25. Outline • The biochemistry of cell signaling and combinatorial complexity • The conventional approach to modeling • The rule-based approach to modeling • Tools

26. Rule-based Approach: A Graph Grammar for Biochemical Systems Faeder et al., Proc. ACM Symp. Appl. Computing (2005) Blinov et al., Proc. BioCONCUR (2005) http://cellsignaling.lanl.gov/bionetgen

27. Proteins in a model are introduced with molecule templates Molecule templates Grb2 Sos Shc EGFR EGF L1 CR1 SH3 SH2 PTB Y317 Y1092 Y1172 Nodes represent components of proteins or Components may have attributes: P

28. Complexes are connected instances of molecule templates An EGFR dimer P P P P P Edges represent bonds between components Bonds may be internal or external

29. Patterns select sets of chemical species with common features Pattern that selects EGFR phosphorylated at Y1092. selects P P P P P P Y1092 P P P P P , , , , … EGFR twice inverse indicates any bonding state suppressed components don’t affect match

30. Reaction rules, composed of patterns, generalize reactions EGF binds EGFR k+1 L1 EGF CR1 + k-1 EGFR Patterns select reactants and specify graph transformation - Addition of bond between EGF and EGFR

31. Rule-based version of the Kholodenko model • 5 molecule types • 23 reaction rules • No new rate parameters (!) 18 species 34reactions 356 species 3749 reactions Blinov et al. Biosystems83, 136 (2006).

32. Dimerization rule eliminates previous assumption restricting breakup of receptors EGFR dimerizes (600 reactions) EGF k+2 dimerization + k-2 EGFR Dimers form and break up independent of phosphorylation of cytoplasmic domains

33. Summary of the rule-based approach • Rigor • Level of detail (sites) consistent with biological knowledge • Physical basis for parameters • Unbiased simplifications • Compositionality • Models can be easily extended by adding new molecules, components, and rules • Reusability of parts • Clarity • Easy to visualize, because model elements are graphs • Easy to store and modify in electronic form (graph data structures) • Rule-based models have about the same number of parameters as conventional models

34. Outline • The biochemistry of cell signaling and combinatorial complexity • The conventional approach to modeling • The rule-based approach to modeling • Tools

35. BioNetGen2: Software for graphical rule-based modeling Graphical interface for composing rules Text-based language Simulation engine

36. BNGL: A textual language for graphical rules L(r) + R(l,d) <-> L(r!1).R(l!1,d) kp1, km1

37. BNGL: A textual language for graphical rules reactant patterns product pattern rate law(s) L(r) + R(l,d) <-> L(r!1).R(l!1,d) kp1, km1 molecule components (unbound) a bond

38. Graphical Interface to BioNetGen Greatly simplifies construction, visualization, and simulation of complex models

39. Conclusions • Biological systems exhibit combinatorial complexity • Conventional approaches to modeling selectively ignore this problem • Rule-based modeling provides a more general and rigorous approach • Parameter requirements can be modest • Models are easy to modify, extend, and exchange • General-purpose software (BioNetGen) is available • Based on formal visual language (portable) • Facilitates precise, intuitive communication about protein interactions • Implements multiple simulation methods • Applications of modeling approach in collaboration with experimentalists

40. Grand challenge: A comprehensive rule-based model EGFR Pathway Map Oda et al., Mol. Syst. Biol. (2005).

41. Michael L. Blinov James R. Faeder G. Matthew Fricke William S. Hlavacek Funding NIH DOE