This is the tutorial version of my undergraduate thesis work in medicinal chemistry, made for a biology lab.

1. Quantum Mechanical Treatment
of a Binding Mode in Thrombin:
Electrostatics of Binding
Cain Morano, Dr.Dave Hangauer SUNY Buffalo
In collaboration with Dr.Gerhard Klebe and Christof Gerlach
Phillips University, Marburg, Germany

2. Summary
Project summary
Physical chemistry primer
Computational chemistry tutorial
Process and product
Coherent, progressive structure

3. Project Summary
Unraveling the Thermodynamics of
Molecular Recognition:
Enthalpy/Entropy Compensation
Thrombin & available or novel
inhibitors
Collaboration for ITC, X-ray, and
activity
The “Holy Grail” of Drug Discovery:
Quantification of non-covalent interactions

4. But the question is…
Are we King Arthur’s Court…..or Monty Python?

5. Enthalpy/Entropy Compensation
Non-linear relationship
between enthalpy and entropy
Theory supported by
Dudley Williams’ findings
Degrees of freedom and functional groups affect energetics

6. Enthalpy/Entropy Compensation
0,00 UB10 UB11
N
N O N
-10,00 N O
O
N O
kJ/mol
N
-20,00
-30,00
HN N H2
HN N H2
-40,00
Side-by-side comparison

7. Target Compound
S2
S3
O S1
N
N
H 2N O H
Cl
Merck’s ligand w/ binding regions

8. Novel Binding Mode
Charge-transfer bond
Theory supported by
Auffinger et al.
Novel interaction thought to enhance binding

9. Method Summary
Merck crystal structure 1TA2
MM performed with Sybyl 6.9
QM calculations performed with Gaussian94
Visualized with Molekel and Rastop
Process required multiple tools and skill sets

10. Molecular Modeling
Semi empirical
Many functions & features
Powerful tool in the drug discovery toolbox

11. Molecular Modeling
Advantages
 Fast
 User defined functions and parameters
 Rational drug design
Disadvantages
 ‘Global minima problem’
 Requires user intuition and skills
 Requires crystal structure
Another tool, not a paradigm shift

12. Quantum Mechanical Calculations
Ab initio
Main functions
Powerful theoretical tool

13. Quantum Mechanical Calculations
Advantages
 Accurate theory to support practice
 User defined functions
 Detailed physicochemical view
Disadvantages
 Computationally expensive
 Requires specific skill set
 Electron correlation and solvation
Extremely useful for variety of problems

14. Overall method evaluation
Strengths
 Fast
 Inexpensive
 Offers qualitative data
Weaknesses
 No solvation factor
 No cooperativity
 Only a model
Tools that can guide research

15. Math to Model
B3LYP
 DFT – total e density, not Ψ
-
6-31G*
 Split valence basis set with
polarization function
Builds the model piece-by-piece from atomic to molecular orbital

16. Math to Model
Mathematical constructs that emerge into a model

17. Strategy
Calculate for single residues
Calculate binary, ternary, quaternary interactions
Modify with EDG & EWG
Establish collaboration with CCR at SUNY Buffalo
Iterative and progressive process

18. Streamlining the Process
Ligand truncated to meta-chlorotoluene
Amino acids truncated:
 Asp189 to acetate
 Trp215-Gly216 to N-methyl-acetamide
 Tyr228 to para-hydroxytoluene
Begin with small basis set, low level theory
Simplify then build, consider cost and return

19. Binary Interactions
charge scale… -0.1346 to 0.10548 a.u.
Amide bond from backbone with ligand
Areas of counter-intuitive induced polarization

20. Binary Interactions
-0.17249 to 0.07430 a.u.
-0.1346 to 0.10548 a.u.
Manipulate charge scale for details

21. Binary Interactions
charge scale… -0.17249 to 0.07430 a.u.
Compare to lone molecules

22. Quaternary Interactions
charge scale… -0.33983 to 0.02533
Prominent active site electrostatic interactions in complex

23. Quaternary Interactions
Asp189
Amide
Ligand
Tyr228
Electrostatic cooperativity evident in complex

24. Quaternary Interactions
Amide
Ligand
Asp189
Skewed electron density evident on all faces of ligand

25. Atomic Charges (a.u.)
Ligand – m-chlorotoluene
Ligand alone in 4 complex
(+)
1 C -0.515353 -0.520246
3 C 0.160422 0.221632
4 C -0.179828 -0.223321 (-)
5 C -0.059304 -0.044822 (-)
6 C -0.140007 -0.163481
7 C -0.119940 -0.124885
8 C -0.179843 -0.217493 (+)
12 Cl -0.021362 0.002192
Chlorine loses large average electron density

26. Atomic Charges (a.u.)
Tyrosine 228
Tyr alone in 4° complex
1 C -0.529781 -0.520798 (+)
5 C 0.176656 0.185123
6 C -0.188862 -0.186104
8 C -0.191649 -0.180380
10 C 0.345964 0.359216
11 O -0.645287 -0.648089 (+)
13 C -0.156044 -0.149920
15 C -0.184380 -0.187260
Biggest changes were loss of e- density

27. Atomic Charges (a.u.)
Amide bond of peptide backbone
Amide alone in 4° complex (+)
1 C -0.532613 -0.544136
2 C 0.588998 0.599478
6 O -0.508945 -0.520799
7 N -0.580061 -0.546658
8 C -0.310442 -0.304025
(-)
Amide oxygen gains, Amide nitrogen loses e- density

28. Atomic Charges (a.u.)
Aspartic Acid 189
(-)
Asp alone in 4° complex
1 C -0.502038 -0.507450
2 C 0.520851 0.501807
3 O -0.639533 -0.625789
4 O -0.639927 -0.630447
(+)
Oxygens lose e- density, Carbon gains e- density

29. Side-by-Side Comparison
Charge scale -0.1000 to -0.1800
2,5-dichloro m-chloro m-chloro-o-nitro
Other ligands are being synthesized

30. Conclusions
Data weaknesses
Data strengths
Collaboration and new tools
Enough data to point at useful new methods but needs work

31. Current Projects
CH3 O CD3 O
H 3C D 3C N
N
OH OH
H 3C D 3C
Cl Cl
T rim e th y la m m o n iu m b u tyric a c id C h lo rid e la b e ls
O OH
N
HN H
HN OH
H O N
OH
N H O
O O OH
R
O Leu T rp Tyr
R= O OH
F A -P h e -R N HN
H
N
HN NH OH
N
NH2 H
O
A rg H is
Synthesis of labels and dipeptide substrates

32. Current Projects
HEK3, HEK4 analysis
using Analyst and MASCOT
DIGE analysis using
DeCyder software

33. Done!
P.S. Thanks for not falling asleep.