Medicinal and Computational Chemistry Dept., ChemDiv, Inc., 6605 Nancy Ridge Drive, San Diego, CA 92121 USA, Service: +1 877 ChemDiv, Tel: +1 858-794-4860, Fax: +1 858-794-4931, Email: ChemDiv@chemdiv.com
Fragment-based drug discovery (FBDD) has become an efficient methodology toward identification of small-molecule leads [ – ,,], and therefore fragment libraries are of great interest in both industry and academia.
The majority of commercially available fragment libraries are predominantly populated with flat (hetero)aromatic chemotypes [, ]. This can be explained by two factors. Initially, fragment libraries were designed to be well detectable in NMR screening. Since (hetero)aromatic compounds usually exhibit well resolved chemical shifts, they are NMR friendly fragments for hit identification . In addition, a large number of fragment hits have been reported against kinase ATP-binding pocket. Since such fragment hits should mimic the adenine base of ATP, almost all of them can be characterized as flat sp2-rich structures.
On the other hand, nature is three-dimensional and therefore recognizes small molecules in a complementary 3D-fashion, and so drugs are likely to be more selective for their targets if they are three-dimensional too [, ]. Not coincidentally, compounds with diverse and well-developed 3D-shapes have become the most attractive ones on the market of screening compounds for HTS for last several years. Furthermore, Fsp3 parameter has become one of the most important criterion of HTS libraries value since it was introduced in 2009 by Frank Lovering et.al  as a measure of three-dimensionality and therefore complexity for libraries members. According to their findings and our further observations, scaffold/molecule saturation may benefit:
¬ More diversity;
¬ More complexity;
¬ Access to greater chemical space;
¬ Improved phys-chem parameters (logP; PSA; water solubility etc.);
¬ More opportunity to reduce scaffold MW;
¬ More opportunity for further scaffold modification;
¬ Natural product-likeness;
¬ Better affinity to target proteins and greater selectivity;
¬ Easy access to IP-clean field.
Noteworthy, this trend retained almost unnoticed on FBDD field until recently perhaps because complexity of such compounds contradicts main principle of FBDD “from simplest fragments toward complex ligands”. Nevertheless, the authors of numerous recent discussion papers are convinced that chemical space and quality of current fragment libraries will be improved significantly if their 3D-diversity is enriched. In their opinion, this can expand the horizons of FBDD enabling new opportunities in most challenging target classes such as PPI, β-secretase etc [3-6,]. Some research groups have made first practical contribution on this direction [8,]. Furthermore, several UK-based non-profit drug discovery institutions, spanning a range of therapeutic foci, have come together to form the 3D Fragment Consortium (http://www.3DFrag.org) aiming to build a shared fragment library with enhanced three-dimensional characteristics and subsequently evaluate them in a range of fragment screens using a variety of screening methodologies.
Taking into account this trend on FBDD field we at ChemDiv have built our own “Beyond the Flatland” 3D-Fragment Library.
STRATEGY OF LIBRARY MEMBERS SELECTION
The library candidates should meet at least one of the following criteria:
¬ Fsp3 ≥ 0.4, preferably due to higher saturation of (hetero)cycle but not side chains;
¬ One and more chiral center in structures;
¬ Bridged structures;
¬ 1,2-Di (bulky)substituted (hetero)cycles
The following filters have been applied for final library population:
¬ MW ≤ 300;
¬ cLogP ≤ 3.0;
¬ HBD ≤ 5;
¬ HBA ≤ 8;
¬ NRB ≤ 4;
¬ No Med-Chem restrictions (widely used and proprietary ChemDiv substructure filters on reactive functionalities, toxicophores, instability suspicions etc.)
¬ The library consists of more than 4,400 fragments;
¬ Assured chemical diversity (diversity coefficient 0.9);
¬ More than 1100 compounds meet strict Astex Rule of Three criteria  (MW ≤ 300, cLogP ≤ 3, H-bond donors ≤ 3, H-bond acceptors ≤ 3);
¬ More than 2850 library members contain at least one chiral center in the structure;
¬ More than 200 bridged fragmens;
¬ More than 450 spiro-fragments.