gqsar for gpcr studies

17
© VLife Sciences Technologies Pvt. Ltd. All rights reserved 1 Fragment Based-GQSAR for GPCR Studies Presenter: Kundan B. Ingale Application Scientist kundani@vlifesciences. com

Upload: vlife-sciences-tech-pvt-ltd

Post on 16-May-2015

759 views

Category:

Technology


0 download

DESCRIPTION

GQSAR is a breakthrough patent pending methodology that significantly enhances the use of QSAR as an approach for new molecule design. As a predictive tool for activity, this method is significantly superior to conventional 3D and 2D QSAR. Here we explain application of GQSAR for optimizing GPCR compounds in non congeneric series.

TRANSCRIPT

Page 1: GQSAR for GPCR Studies

© VLife Sciences Technologies Pvt. Ltd. All rights reserved 1

Fragment Based-GQSAR for GPCR Studies

Presenter:Kundan B. IngaleApplication [email protected]

Page 2: GQSAR for GPCR Studies

© VLife Sciences Technologies Pvt. Ltd. All rights reserved 2

• Mediate biological signalling in health/disease

• Commercially validated - 40% of top 100 drugs

• < 2% of proteins in PDB

• Difficult to crystallize or too big for NMR

• Other Issues

• Non-alpha helices

• Loops may contain other secondary structures and domains

• Bias towards TM proteins that are easy to crystallize

• Energetics of TM proteins not completely understood (polar interactions or van der Waals interactions play role in role in helix-helix interaction)

GPCR…

Page 3: GQSAR for GPCR Studies

© VLife Sciences Technologies Pvt. Ltd. All rights reserved 3

Challenges in Multi-target ligand design for GPCRs

• Design of Multi Targeted ligands:

• Advantages over single drug for single target

• Ligands that simultaneously bind to 5HT1A and 5HTT have shown good promise in treatment of major depression

• Can structure based method be used ?

• Are Ligand based methods useful ?

• Shape Based Comparisons ?

• Pharmacophore based ?

• QSAR ?

• What Next ..?

Page 4: GQSAR for GPCR Studies

© VLife Sciences Technologies Pvt. Ltd. All rights reserved 4© VLife Sciences Technologies Pvt. Ltd. All rights reserved 4

Key elements of GQSAR

Where is GQSAR useful

•Lead optimization by using site specific clues from GQSAR model •Scaffold hopping by choosing

• groups/fragments satisfying descriptor

• ranges of actives in the dataset•Novel library generation along with predicted activity of ligands

• Alignment independent fragment based QSAR modeling

• Conformer independent method• GQSAR models generation for both

congeneric and non-congeneric data• Provides site specific clues• Patent pending method

Group QSAR: For lead optimization

Publication references

• QSAR Combi Science 2009, 28:36–51• J Mol Graph Mod 2010;28:683-694

Fig: GQSAR Workflow

Page 5: GQSAR for GPCR Studies

10/10/11

Dataset for GQSAR

• BindingDB database: 162 molecules (5HT1A receptor and 5HTT Inhibitory activities)

• Biological activity: Binding affinity data (Ki nM)

ClassNumber of molecules

Activity (Ki nM) Min Max

C1 (Piperidine) 69

HT1A 0.91 3200.00HTT 0.24 9006.00

C2(Piperazine) 56

HT1A 0.12 475.00HTT 1.30 3900.00

C3(non ring N atom) 26

HT1A 2.00 1470.00HTT 0.50 4700.00

C4(1,2,3,6-

tetrahydropyridine) 8

HT1A 10.90 92.60

HTT 19.80 387.00C5

(azabicyclo[3.2.1]oct-3-ene) 3HT1A 127.70 357.00HTT 8.50 33.00

Page 6: GQSAR for GPCR Studies

NO

OH

NH

CH3

S

10/10/11

Representative Molecules

NO

O

O NH

F

NO

O

OCH3

N

N

O

NH

CH3O

Piperazine (C2)

Non Ring Nitrogen (C3) 1,2,3,6-tetrahydropyridine (C4)

azabicyclo[3.2.1]oct-3-ene (C5)

Piperidine (C1)

NH

NH

O

NH

O

F

Page 7: GQSAR for GPCR Studies

Fragmentation Pattern

Fragment R1 (aromatic region): aromatic ring connected with the core of the molecule i.e. fragment R2

Fragment R2 (anchor region): substituent present in the center of the molecule

Fragment R3 (flexible region): substituent connected to other end of the fragment R2

Page 8: GQSAR for GPCR Studies

8

Relationship between 5HT1A and 5HTT inhibition

Design and optimize molecules for multi-target activity

r2 = 0.051

Fig: Scatter Plot of pKi_5HT1A Vs pKi_5HTT

Page 9: GQSAR for GPCR Studies

Data Processing and Model building

• Biological Activity: negative logarithm of binding affinity i.e. pKi (nM)

• Descriptors: 2D group based descriptors and their squared terms

• Training set: 93 molecules ( From scaffold C2-C5)

• Test set : 69 molecules (From scaffold : C1)

• GQSAR enables identification of common set of descriptors influencing the binding of ligands to both the targets

• GQSAR model: (without Fragment Interaction Descriptors)

• 62% (r2 = 0.620) of variation in the 5HT1A activity

• 49 % (r2 = 0.490) of variation in the 5HTT activity

Page 10: GQSAR for GPCR Studies

10/10/11

GQSAR Model

• Fragment Interaction Descriptors:

• Example : R1_slogp*R2_smr: Product of log P of fragment R1 and molar refraction of fragment R2.

• GQSAR model: (with Fragment Interaction Descriptors)

• 71% (r2 = 0.710) of variation in the 5HT1A activity • 83% (r2 = 0.830) of variation in the 5HTT activity

Fig: Contribution Plot for descriptors in GQSAR equation

Page 11: GQSAR for GPCR Studies

Model Representation

Page 12: GQSAR for GPCR Studies

Model Validation

• Test Set : 69 molecules (chemical class not present in the training set).

• Model Applicability Domain Check: 50 molecules out of 69

• Prediction Correctness: molecules predicted within ±1 log units

• 5HT1A: 46 (92%),

• 5HTT: 40 (80%)

• Prediction accuracy: > 80% with a new scaffold

Page 13: GQSAR for GPCR Studies

Aromatic region (R1) Descriptors: R1-4pathClusterCount (6.33, 1.8)*

Anchor region (R2) Descriptors: R2-PSAExclPandS (2.62, 2.92)*

Flexible region (R3) Descriptors:

R3-4pathClusterCount^2 (-1.47, -1.11)*

R3-SssNHE-index^2 (-4.39, -3.73)*

10/10/11

Model Interpretation

* Figures in bracket indicate contribution of descriptor towards 5HTT and 5HT1A respectively

↑ Branched substitution

↑ Polar surface

↓ Branched Substitution↓ H-don N atom attached to 2 heavy atoms pKi (5HTT): -0.97;

pKi(5HT1A) : 0.04

N

N

O

NH

CH3O

Page 14: GQSAR for GPCR Studies

N

CH3

OOCH3

10/10/1110/10/11

• Aromatic (R1) and Flexible (R3) regions interaction descriptors:

• R3-4pathClusterCount*R1-T==2 (-19.63, -8.22)

• R3-smr*R1-T==6 (-6.83, -14.75)

Model Interpretation (Interaction Descriptors)

* Figures in bracket indicate contribution of descriptor towards 5HTT and 5HT1A respectively

↓sp2 atoms separated by 2 bonds

↓ Branched substitution

↓sp2 atoms separated by 6 bonds

↓ Molar refractivity

pKi (5HTT): -0.59; pKi(5HT1A) : -0.3

Page 15: GQSAR for GPCR Studies

10/10/11

Summary and Conclusion

• GQSAR method can be successfully applied to non-congeneric series of molecules

• With GQSAR, one can identify common set of descriptors that influence the multi-targeted activities of ligands

• GQSAR method provides site specific clues for Lead optimization

• GQSAR method can be effectively used to design Multi Targeted ligands

10/10/11

Page 16: GQSAR for GPCR Studies

References

• GQSAR is patented by VLife Sciences Technologies Pvt. Ltd.

• References:

• "Group Based QSAR (G-QSAR) : Mitigating Interpretation Challenges in QSAR”, QSAR & Combinatorial Science, 28(1),36–51(2009)

• "A Comprehensive Structure-Activity Analysis of Protein Kinase B-alpha (Akt1) Inhibitors“, Journal of Molecular Graphics and Modelling, (2010) doi: 10.1016/j.jmgm.2010.01.007

16

For more information : [email protected] For more information : [email protected]

Page 17: GQSAR for GPCR Studies

VLife Sciences Technologies Pvt Ltd101-102 Pride Purple Coronet, Baner Road, Pune 411 045 (MS) India

Tel / Fax : +91 20 2729 1590/1 Email : [email protected] www.VLifeSciences.com

Copyright © 2005 VLife Sciences Technologies Pvt. Ltd. All Rights Reserved.

VLife Sciences, VLife Logo ,and all other VLife product names and slogans are trademarks or registered trademarks of VLife Sciences Technologies Pvt. Ltd.17