Chemical Reactions

1. Chemical Reactions

2. Overview • Reactions in organic chemistry, review • Problems in reaction chemistry • Chemoinformatics methods • Applications of reaction chemoinformatics to reaction chemistry problems • Review questions 2

3. Overview • Reactions in organic chemistry, review • Problems in reaction chemistry • Chemoinformatics methods • Applications of reaction chemoinformatics to reaction chemistry problems • Review questions • Chemicals – Reactions Relationship • Reaction Specification • Reaction Mechanisms • Reactivity Principles • Reaction Favorability • Reaction Classification 3

4. + HBr Chemicals – Rxns Relationship • Chemical Space • Chemicals are points in the space • Reactions are “vectors” describing how to reach new points from existing ones • Reactant Chemicals  Product Chemicals • Transformation that forms and breaks bonds • Rearrangement of electron configuration 4

5. C2H5O- Na+ + HBr C2H5OH 70o C Reaction Specification • Simplest reaction specification is a chemical equation indicating starting reactants and resultant products • For practical use and reproducibility, additional information is required: • Catalyst or other reagents • Reaction conditions (temperature, solvent, etc.) • Yield %, etc. 5

6. Reaction Mechanisms • Reactions are fundamentally rearrangements of electron configurations • Mechanisms describe the specific flow of electrons, the transient intermediates, and the final products 6

7. + Mechanistic Principles • Curved arrow diagrams • Depict flow of electrons, NOT atoms • Source must be electrons (bond, lone pair, radical) • Targets should be atoms / nuclei 7

8. Reactivity Principles • Broadly speaking, reactions are the transfer of electrons from • Electron-dense groups (nucleophiles) to • Electron-deficient ones (electrophiles) 8

9. p n s s Reactivity Principles • Molecular orbitals • Distinct spaces around atoms that electrons reside in (high electron probability density) • Up to 2 electrons per orbital • Relative order of reactivity: • radicals (1e) > • n-orbital: Lone pairs > • p-orbital: Double / triple bonds > • s-orbital: Single bonds 9

10. Reaction Favorability • Thermodynamics • Eventually reactions will proceed to thermodynamic equilibrium, maintaining a steady state ratio of products : reactants • Keq: Equilibrium constant defining the stable ratio of products : reactants for a reaction under standard conditions (1 atmosphere, room temperature) • Larger value of Keq thus indicates greater favorability for a reaction • Given competing products, Keq can indicate major ones 10

11. Reaction Favorability • Gibbs Free Energy • Keq is a function of DGo (and temperature) • DG: Difference between product and reactant (Gibbs) free energy • Negative DG is thus favorable • State function, measuring thermodynamic stability • DGo: DG under standard conditions Keq = e-DGo/RT DGo = -RT ln Keq R = Universal gas constant T = Absolute temperature 11

12. Reaction Favorability • Enthalpy and Entropy contributors • G = H – TS • H: Enthalpy, primarily determined by strength of bonds broken and formed in a reaction • S: Entropy, measuring “randomness” of a system, with greater randomness being favorable • For most reactions, DS is small (esp. when Dn = 0), thus • Unless at very high temperatures, DH dominates TDS, thus • Calculating DH provides a good estimate for DG 12

13. Reaction Favorability Scoring • Thermodynamics • DG = DH – TDS (Enthalpy & Entropy contribute) • Hess’ Law simplification • DHreaction = S(BDEbroken) – S(BDEformed) • BDE: Bond Dissociation Energy • Standard lookup values (kcal/mol) C-C : 83 C=O : 178 C=C : 146 C≡N : 213 C=N : 147 etc. • Kinetics • Much less data available 83 + 178 + 83 + 213 557 13 http://www.cem.msu.edu/~reusch/OrgPage/bndenrgy.htm

14. Reaction Favorability • Reaction Kinetics • Thermodynamics: How “far” a reaction will proceed • Kinetics: How “fast” a reaction will proceed 2 H2 + O2 2 H2O • Highly favorable DG, but without a catalyst or flame, reaction proceeds so slowly as to essentially not occur • Measured by rate constants, but much less data exists • Based on relative stability of transition states… 14

15. + Reaction Favorability • Given infinite time, all reactions will reach thermodynamic equilibrium, but • Intervening, unstable intermediates in the pathway impose an activation energy (Ea) barrier • Given limited time and input energy, a reactions may only achieve kinetic equilibrium, settling into an energy local minimum between large Ea barriers Relative Energy Ea DG Reaction Coordinate • DGo = 1.4 kcal / mol ~ 10x Keq • Ea < 22 kcal / mol ~ Room temperature reaction 15

16. Overview • Reactions in organic chemistry, review • Reaction Classification • Specific chemicals • Compatible functional groups • Reactant counts • Bond rearrangement patterns • Functional classification • Mechanism based More General More Informative 16

17. + + + + Reaction Classification/Organization • Specific chemicals • acetic acid + methanamine  N-methylacetamide • Compatible functional groups • carboxylic acid + primary amine  amide + water 17

18. + + + HBr + Reaction Classification • Reactant counts • Substitution Dn = 0 • Addition Dn < 0 • Elimination Dn > 0 18

19. A B C D A B C D A B C D Reaction Classification • Bond rearrangement patterns • 4 atom bond swap covers ~50% of organic reactions 19

20. A F B E C D A F B E C D + Reaction Classification • Bond rearrangement patterns • 6 atom cyclic rearrangement covers ~25% 20

21. H2SO4 Heat + HNO3 + H2O + + Na2Cr2O7 H2SO4 Reaction Classification • Functional classification • Acid-catalyzed, Electrophilic • Base-catalyzed, Nucleophilic • Oxidation- Reduction • Free-radical • Etc. 21

22. Reaction Classification • Mechanism-based • Sn1 • Sn2 • E1 • E2 • etc. • Most informative classification patterns, but • Reaction mechanisms often unknown • Mechanisms cannot be directly observed, can only be proposed and supported with exp. evidence 22

23. Overview • Reactions in organic chemistry, review • Problems in reaction chemistry • Reactions in organic chemistry, review • Problems in reaction chemistry • Chemoinformatics methods • Applications of reaction chemoinformatics to reaction chemistry problems • Review questions • Storing / retrieving reaction information • Combinatorial chemistry / virtual chemical space • Reaction prediction / discovery • Chemical Synthesis • Reaction planning • Synthesis design (retrosynthesis) 23

24. Storing / Retrieving Reactions • DB: Record and classify all reactions, including: • Reactants and products • Reaction conditions, catalysts, solvents, etc. • Literature references, lab notes, etc. • Search: Ability to query for information on all reactions that • Use an epoxide reactant • Produce an aromatic ring • Follow the Sn2 reaction mechanism • Use copper as a catalyst • Can be run at room temperature in aqueous solution 24

25. Combi Chem + Virtual Space • Combinatorial Chemistry • Given a collection of “building block” chemicals, combine them with reactions to produce a diverse set of new products • Virtual Chemical Space • Systems like ChemDB catalog all chemicals available for purchase from different vendors • “RChemDB” would store or allow on-the-fly searching of all chemicals indirectly (but easily) available by applying reactions to directly available chemicals 25

26. Reaction Prediction / Discovery • Given a mixture of reactants and reaction conditions, predict the major products NaOMe D ? + 26

27. Knowledge vs. Principle-based • Knowledge-based • If a reaction database was available, predicting the course of a reaction could just be a matter of finding it (or an analog) in the database • Knowledge-based limitations • Requires construction of the database of many different known reaction profiles to achieve any degree of generalization • DB driven approach would be unlikely to discern competing cases. For example, • carboxylic acid + amine  amide • carboxylic acid + alcohol  ester • carboxylic acid + amino-alcohol  ? 27

28. Knowledge vs. Principle-based • Principle-based • Predict or derive reactions based on general principles of reactivity • Much more flexible and powerful • Entails the ability to discover new reaction profiles that may not be in known in any DB • Principle-based limitations • Complex reactivity can be very difficult to predict • Confounding factors of solvent effects, catalysts, etc. 28

29. Chemical Synthesis • Series of reactions from starting reactants to form a pathway to the final product H2 Pd CaCO3 Quinolone HBr 29

30. ? ? ? Reaction Planning • Derive synthesis pathway given • Starting reactant • Target product • Available reagents / reactions 30

31. ? ? ? Retrosynthesis • Derive synthesis pathway given • Starting reactant pool • Target product • Available reagents / reactions Chemical Vendor Catalog 31

32. Overview • Reactions in organic chemistry, review • Problems in reaction chemistry • Chemoinformatics methods • Applications of reaction chemoinformatics to reaction chemistry problems • Review questions • Problems in reaction chemistry • Chemoinformatics methods • SMILES Extensions • Reaction SMILES • SMARTS • SMIRKS • Quantum Mechanics 32

33. + HBr SMILES Extensions • Reaction SMILES • Reaction equation denoted with delimiters • “.” separates distinct molecules • “>>” separates reactants from products CCC(Br)(C)C>>CC=C(C)C.Br 33

34. C2H5O- Na+ + HBr C2H5OH 70o C SMILES Extensions • Reaction SMILES • Catalyst, solvent or other chemicals may be added between the “>>” delimiters • No natural space to specify non-molecular info such as temperature, yield %, etc. CCC(Br)(C)C>CC[O-].[Na+].CCO>CC=C(C)C.Br 34

35. SMILES Extensions • SMARTS • “Regular expressions” for molecules • SMILES are SMARTS strings, but • SMARTS strings can describe more general matching criteria, such as • Atom types • Bond types • Logical operators (and, or, not) 35 http://www.daylight.com/dayhtml_tutorials/languages/smarts/

36. SMILES Extensions 36 http://www.daylight.com/dayhtml_tutorials/languages/smarts/ for complete rule list

37. SMILES Extensions 37

38. SMILES Extensions • SMIRKS • Reaction profile describing reactants and how to transform them into respective products • Combination of • Reaction SMILES • SMARTS • Atom Mapping • Generally must be manually specified. Limited work done to automatically derive reaction profile from specific examples http://www.daylight.com/dayhtml_tutorials/languages/smirks/ 38

39. 1 O 1 O 8 4 5 10 7,8 3 H NH-R2 + + H2O 2 2 9 3 7 R1 OH 9 4 5 10 R1 NH-R2 SMILES Extensions • Atom Mapping • Necessary to map reactant to product atoms • Proper transform requires balanced stoichiometry • Hydrogens generally must be explicitly specified Carboxylic acid + [O:1]=[C:2]([*:9])[O:3][H:7]. Primary amine  [H:8][N:4]([*:10])[H:5]>> Amide + [O:1]=[C:2]([*:9])[N:4]([*:10])[H:5]. Water [H:7][O:3][H:8] 39

40. + + SMILES Extensions • Atom mapping implies mechanism • Two feasible mechanisms for reaction below • Ambiguity without at least atom mapping • Atom mapping still lacks a complete mechanistic description analogous to “curved arrow” diagram 40

41. Quantum Mechanics • Capable of accurate predictions for • Chemical reactivity • Chemical stability  Reaction favorability • Requires significant computational power, unfeasible for large scale processing 41

42. Overview • Reactions in organic chemistry, review • Problems in reaction chemistry • Chemoinformatics methods • Applications of reaction chemoinformatics to organic chemistry problems • Review questions • Chemoinformatics methods • Applications of reaction chemoinformatics to organic chemistry problems • Reaction databases (storing / retrieving info) • Combinatorial chemistry / virtual chemical space • Reaction prediction / discovery • Synthesis design (retrosynthesis) 42

43. Reaction Databases • Storage • Specific reactions can be recorded with reaction SMILES • More general mechanistic reaction profiles can be stored with SMIRKS • Retrieval • Search by reactant or product is same as usual chemical structure search • Search by bonds that change focuses on reaction centers to find similar classes 43

44. Reaction Databases • Most repositories with thousands of records, some may have millions - CASREACT - Beilstein - ChemInform RX - ChemReact • Generally poor consistency and completion of • Balanced reaction stoichiometry • Atom mapping / mechanistic description • Reaction conditions, etc. • Not publicly available or difficult to access 44

45. Reaction Prediction / Discovery • Algorithm features needed • Hypothesis generating scheme • Thermodynamic scoring system • Kinetic scoring system • Known reactions database NaOMe D ? + 45

46. Reaction Prediction Approximation • Find electron donors (nucleophile) and electron acceptors (electrophile) using rules and rank them • Compute all possible intermediates • Rank by Enthalpy (+Enthropy) • Recurse • Stopping rule (drop in delta G) 46

47. +7 -17.5 Reaction Prediction Example 0 47 Blue: HOMOs / Nucleophiles Red: LUMOs / Electrophiles

48. 0 +300 +315 +415 +300 -25 -30 Reaction Prediction Example 48 Blue: HOMOs / Nucleophiles Red: LUMOs / Electrophiles

49. Dead End Starting Material Dead End Starting Material Retro-Synthesis Tree • Apply retro reactions towards available starting reactants 49

50. Existing Approaches • Retrosynthetic • Interactive: LHASA, SECS • Non-Interactive: SYNCHEM • Forward: SST, CHIRON • Formal: IGOR, WODCA, SYNGEN • Reaction Prediction: CAMEO, EROS 50 Todd, M. H. (2004). "Computer-Aided Organic Synthesis." Chemical Society Reviews(34): 247-266.