Rafael Gozalbes / Computational chemistry in the field of human health

1

ProtoQSAR SL

Vivero de Empresas “Creix”

Paseo de la Pechina 15

46008-Valencia, Spain

[email protected]

2

The company

Creation: September 2012. Two PhD, > 15 years of activity in this field, both at

the academic and industrial sides.

Objective: application of computational methods to the prediction of physico-

chemical and/or biological properties of chemicals.

Main work area: "Drug Discovery".

Other areas: nutraceuticals, pesticides, nanomaterials, REACH legislation.

Dissemination: > 70 peer reviewed papers in scientific journals, 7 chapters in

specialized books, > 40 presentations in national/international scientific

conferences.

3

In silico success in DD

Norfloxacin, the first fluoroquinolone antibiotic (Koga et al., J. Med. Chem., 1980), QSAR

models developed to predict the best substituents.

Losartan, the first non-peptide oral angiotensin II antagonist (Duncia et al., J. Med. Chem.,

1990), alignment with a solution structure of angiotensin II.

5-HT1B/1D agonist Zolmitriptan, indicated for migraine (Glen et al., J. Med. Chem. 1995), based

on a pharmacophore approach & ADME optimised by QSAR.

Dorzolamide, a carbonic anhydrase inhibitor used to treat glaucoma (Greer et al., J. Med.

Chem., 1994), ab initio calculations suggested the best conformation.

Antivirals such as Zanamivir (first neuraminidase inhibitor for tratment of influenza) (M.

von Itzstein et al., Nature, 1993) or Amprenavir (HIV-1 protease inhibitor) (Kim et al., J. Amer. Chem.

Soc., 1995).

Others: H2-receptor antagonist Cimetidine, Probenecid, many of the atypical

antipsychotics, selective COX-2 inhibitors NSAIDs, selective serotonin reuptake

inhibitors (antidepressants), etc.

4

Services

Chemoinformatics:

Calculating molecular descriptors, filtering chemicals by applying

standardized rules, similarity and/or chemical diversity analysis, etc.

QSAR/QSPR.

Molecular modeling:

Homology modeling.

Pharmacophore modeling.

Docking.

Molecular dynamics.

5

Virtual Screening

Target

Docking “De novo” design known

Pharmacophores

QSAR

CC & HTS unknown

known unknown

Ligand(s) Available structure:

Virtual screening

6

Virtual Screening

Docking

Ligand Protein

Docking

Protein Library

7

Computational chemistry advantages

Demonstrated good predictivity when selecting/optimizing new chemical

entities.

Great saving of time, resources and money.

Easy and rapid applicability of the models to new structures or molecular

collections.

Limitations on animal testing (3Rs), according to European legislation.

8

QSAR/QSPR

We are particularly featured for our expertise in the development of

mathematical predictive models of structure-activity relationships (QSAR) or

structure-property relationships (QSPR).

These computational models are of great interest when you have

activity/property data for a number of compounds and the structure of the

therapeutic target is unknown.

Our company is possibly the only one in Spain providing currently a

specific and specialized service in the QSAR field.

9

QSAR

Compounds database

Training set

Validation set

Exp.

activity

Descriptors calculation

ACTIVITY = ∑ ci DESCRIPTORi

n

i = 0

Molecular structures

Structures ID MW Num_carbons Num_heteroatoms Carbon-hetero_ratio Halide_count a_acc diameter a_nH Solubility (mg/L)

O(CC(O)CNC(C)C)c1ccc(NC(=O)CCC)cc1C(=O)CAcebutolol 336,43 18 6 3,00 0 5 15 28 590,50

Clc1cccc(Cl)c1Nc1ccccc1CC(OCC(O)=O)=OAceclofenac 354,18 16 7 2,29 2 3 12 13 223,10

Oc1ccc(NC(=O)C)cc1Acetaminophen 151,16 8 3 2,67 0 2 7 9 0,90

OC(=O)\C=C\c1nc(ccc1)/C(=C/CN1CCCC1)/c1ccc(cc1)CAcrivastine 348,44 22 4 5,50 0 4 12 24 57,20

O=C1NC(=Nc2n(cnc12)COCCO)NAcyclovir 225,20 8 8 1,00 0 5 9 11 6,15

O=C1N(N=NN1CC)CCN1CCC(N(C(=O)CC)c2ccccc2)(CC1)COCAlfentanil 416,52 21 9 2,33 0 6 15 32 4008,80

O(CC(O)CNC(C)C)c1ccccc1CC=CAlprenolol 249,35 15 3 5,00 0 3 11 23 420,00

O=C1N(N(C)C(C)=C1N(C)C)c1ccccc1Aminopyrine 231,29 13 4 3,25 0 1 8 17 444,20

S1SCC(NC(=O)C(NC(=O)C(NC(=O)C(NC(=O)C(NC(=O)C(N)C1)Cc1ccc(O)cc1)Cc1ccccc1)CCC(=O)N)CC(=O)N)C(=O)N1CCCC1C(=O)NC(CCCNC(N)=N)C(=O)NCC(=O)NArgipressin 1084,23 46 29 1,59 0 14 26 65 685,30

S1[C@H]2N([C@@H](C(O)=O)C1(C)C)C(=O)[C@H]2NC(=O)Cc1ccccc1Benzylpenicillin 334,39 16 7 2,29 0 4 12 18 66,50

Clc1cc(ccc1)C(=O)C(NC(C)(C)C)CBupropion 239,74 13 3 4,33 1 2 8 18 38,20

S1[C@H]2N(C(C(O)=O)=C(C1)CSc1nn[nH]c1)C(=O)[C@H]2NC(=O)C(N)c1ccc(O)cc1Cefatrizine 462,50 18 13 1,38 0 9 17 18 6,12

s1cc(nc1N)/C(=N/OC)/C(=O)N[C@H]1[C@H]2SCC(CSC3=NC(=O)C(O)=NN3C)=C(N2C1=O)C(O)=OCeftriaxone 554,58 18 18 1,00 0 10 18 18 2315,40

S1[C@H]2N(C(C(O)=O)=C(C1)COC(=O)N)C(=O)[C@H]2NC(=O)\C(=N/OC)\c1occc1Cefuroxime 424,39 16 13 1,23 0 6 14 16 2232,40

S1[C@H]2N(C(C(O)=O)=C(C1)C)C(=O)[C@H]2NC(=O)[C@H](N)c1ccccc1Cephalexin 347,39 16 8 2,00 0 5 12 17 657,10

Structures ID MW Num_carbons Num_heteroatoms Carbon-hetero_ratio Halide_count a_acc diameter a_nH Solubility (mg/L)

ClC(Cl)C(=O)N[C@@H]([C@H](O)c1ccc([N+](=O)[O-])cc1)COChloramphenicol 323,13 11 9 1,22 2 3 11 12 ???

Clc1cc2N(c3c(Sc2cc1)cccc3)CCCN(C)CChlorpromazine 318,86 17 4 4,25 1 1 9 19 ???

Fc1cc2c(N(C=C(C(O)=O)C2=O)C2CC2)cc1N1CCNCC1Ciprofloxacin 331,34 17 7 2,43 1 4 11 18 ???

Clc1cccc(Cl)c1NC=1NCCN=1Clonidine 230,09 9 5 1,80 2 1 7 9 ???

O=C1CC[C@@]2([C@@H]3[C@H]([C@@H]4CC[C@](O)(C(=O)CO)[C@]4(CC3=O)C)CCC2=C1)CCortisone 360,44 21 5 4,20 0 5 12 28 ???

o1ncc2C[C@@]3([C@@H]4[C@H]([C@@H]5CC[C@@](O)(C#C)[C@]5(CC4)C)CCC3=Cc12)CDanazol 337,46 22 3 7,33 0 2 11 27 ???

N(CCCN1c2c(CCc3c1cccc3)cccc2)CDesipramine 266,38 18 2 9,00 0 1 9 22 ???

Virtual screening

10

Available QSAR models

Physico-chemical parameters: cLogP, PSA, water solubility…

ADME properties: oral bioavailability, BBB penetration, Caco-2 permeability,

binding to plasma proteins, cytochromes inhibition, distribution volume,

biological half-time…

Toxicological properties: acute oral toxicity, skin irritation, mutagenicity,

carcinogenicity…

“Global models”: GPCRs, CNS model.

11

Customers & collaborators

Clientes y/o colaboradores.

Mantenemos colaboraciones con grupos de trabajo muy diferentes, tanto por el tipo de

entidad (grupos académicos, entidades privadas sin ánimo de lucro, empresas) como por su

objeto (investigación en química médica y/o combinatoria, en predicción de propiedades

ADME-T, desarrollo de nuevos insecticidas, de nuevos materiales de embalaje, etc.).

Algunos de nuestros clientes son los siguientes:

We collaborate with very different kind of groups, either concerning their entity

(academic groups, private foundations, companies…) or their objective

(medicinal or combinatorial chemistry research, ADME-T prediction, new

insecticides development, new materials for chemical industries, etc.)

12

How to work with us

Application of our QSAR/QSPR models for prediction of properties to

compounds submitted by our customers.

Hiring of our services in molecular modeling and chemoinformatics (medicinal

chemistry projects, generation of targeted librarires, etc).

Collaboration in research projects.

13

Perspectives

Functional food: project with a biotech to search for new peptides preventing

metabolic syndrome.

Agro: collaboration with a private scientific association & INRA to search for red

palm weevil repellents.

miRNAs: collaboration with a BioDonostia group to analyze the possible role

of miRNAs in multiple sclerosis.

NanoQSAR: project with a technological institute to develop innovative models

predicting ecotoxicity of nanoparticles.

14

Thank you for your

attention!

[email protected]