Most recent pharmacophore modelling programmes define additional steric
constraints features. These are called exclusion volumes (XVols), representing the
steric effect of the binding pocket [46]. These features are required to avoid the
clashes of the molecule with the protein surface while mapping. Feature generation
not only facilitates the molecules to be aligned in an easy and rational way, but also
can be used in scoring. The root mean square deviation (RMSD) between matched
features gives quantitative account of the extent of overlay, which is often used as a
fitness score [40]. Hence, the placement of feature points should be accurate, and
one needs to be careful while deciding whether to consider all possible features or
to choose few of them giving adequate information about the spatial orientation of a
group of molecules. For example, sometimes there are huge number of hydrophobic
features as compared to other features, which may bias the alignment and give a
model with good score, but the model will be useless due to lack of specificity.
4 Evolution of the ‘Pharmacophore’ Concept:
Historical Perspective
Paul Ehrlich first used the concept of pharmacophore in the end of nineteenth
century, when he revealed the selective binding of methylene blue to nerve fibres.
This realization ushered the beginning of pharmacophore concept as ‘a molecular
framework that carries (phoros) the essential features responsible for a drug’s
(pharmacon) biological activity’ [37, 38]. Based on this idea, Ehrlich improved the
chemical structure of several compounds to yield efficacious drugs against syphilis
(under the trade name Salvarsan), trypanosome and spirochete infections [37, 38],
which made him win the Nobel prize in 1908 sharing with Ilya Metchnikoff.
Although Ehrlich’s early definition of pharmacophore is almost unchanged for over
a century, Schueler proposed the first modern definition in his book
‘Chemobiodynamics and Drug Design’ in 1960 [47], where the ‘chemical groups’
were replaced by patterns of ‘abstract features’. Beckett and co-workers [48] proposed the first pharmacophore model of muscarinic agents in 1963 that identified
distance ranges between abstract features, and later in 1967, Kier developed the first
‘computed’ pharmacophore model for muscarinic receptor inhibitor binding pattern
[49–51]. Simple pharmacophores were in application as tools for designing new
drug molecules much before the dawn of a well-defined field like computer-aided
drug design. In the 1940s, preliminary structure-activity relationship models were
computed based on simple two-dimensional model structures utilizing the accessible information of the van der Waals sizes and bond lengths [52]. Eventually, in
the 1960s, three-dimensional models could be built with the convenience of X-ray
and conformational analysis techniques. Medicinal chemists could classify some
common molecular frameworks that attributed to high biological activity more
often as compared to other structures by retrospectively analysing the chemical
structures of the various drugs. Evans et al. [52] named such frameworks as
30
C. Choudhury and G. Narahari Sastry
constraints features. These are called exclusion volumes (XVols), representing the
steric effect of the binding pocket [46]. These features are required to avoid the
clashes of the molecule with the protein surface while mapping. Feature generation
not only facilitates the molecules to be aligned in an easy and rational way, but also
can be used in scoring. The root mean square deviation (RMSD) between matched
features gives quantitative account of the extent of overlay, which is often used as a
fitness score [40]. Hence, the placement of feature points should be accurate, and
one needs to be careful while deciding whether to consider all possible features or
to choose few of them giving adequate information about the spatial orientation of a
group of molecules. For example, sometimes there are huge number of hydrophobic
features as compared to other features, which may bias the alignment and give a
model with good score, but the model will be useless due to lack of specificity.
4 Evolution of the ‘Pharmacophore’ Concept:
Historical Perspective
Paul Ehrlich first used the concept of pharmacophore in the end of nineteenth
century, when he revealed the selective binding of methylene blue to nerve fibres.
This realization ushered the beginning of pharmacophore concept as ‘a molecular
framework that carries (phoros) the essential features responsible for a drug’s
(pharmacon) biological activity’ [37, 38]. Based on this idea, Ehrlich improved the
chemical structure of several compounds to yield efficacious drugs against syphilis
(under the trade name Salvarsan), trypanosome and spirochete infections [37, 38],
which made him win the Nobel prize in 1908 sharing with Ilya Metchnikoff.
Although Ehrlich’s early definition of pharmacophore is almost unchanged for over
a century, Schueler proposed the first modern definition in his book
‘Chemobiodynamics and Drug Design’ in 1960 [47], where the ‘chemical groups’
were replaced by patterns of ‘abstract features’. Beckett and co-workers [48] proposed the first pharmacophore model of muscarinic agents in 1963 that identified
distance ranges between abstract features, and later in 1967, Kier developed the first
‘computed’ pharmacophore model for muscarinic receptor inhibitor binding pattern
[49–51]. Simple pharmacophores were in application as tools for designing new
drug molecules much before the dawn of a well-defined field like computer-aided
drug design. In the 1940s, preliminary structure-activity relationship models were
computed based on simple two-dimensional model structures utilizing the accessible information of the van der Waals sizes and bond lengths [52]. Eventually, in
the 1960s, three-dimensional models could be built with the convenience of X-ray
and conformational analysis techniques. Medicinal chemists could classify some
common molecular frameworks that attributed to high biological activity more
often as compared to other structures by retrospectively analysing the chemical
structures of the various drugs. Evans et al. [52] named such frameworks as
30
C. Choudhury and G. Narahari Sastry