Libraries for Fragment-Based Drug Discovery

The fragment-based approach to drug discovery (FBDD) has been established as an efficient tool in the search for new drugs. [1], [2] It has been in intensive use since the end of nineties [3] and it already starts delivering compounds which are entering the clinic. [4] In essence, the FBDD differs from the long-used high-throughput screening (HTS) of large, relative high molecular weight compound collections: it identifies simpler, low molecular weight compounds, the "fragments", which bind to the target of interest. The FBDD has several advantages over HTS of large compound libraries. [1], [2] First of all, when following this approach, there is a better chance for the final lead compound to be compliant with the Lipinski’s "Rule of five" [5] to increase the likelihood of having good pharmacokinetic properties. Another important benefit of the FBDD derives from the fact that even a small (few thousands) collection of fragments covers a much greater proportion of all the possible compounds that could exist, termed "chemical space", than large (millions) corporate compound collections for HTS. [6] Finally, the increased chemical diversity allows one to avoid problems caused by existing intellectual property. Because of relatively low affinity of fragments to the biological target, efficient techniques had to be developed to detect their binding. Today, researchers widely employ X-Ray crystallography and NMR screening as well as other techniques for this purpose and the bottleneck of FBDD becomes availability of quality fragments.

Understanding the increasing importance of the FBDD for modern pharmaceutical industry, we set a goal to design a compact library of quality fragments which would represent the whole collection of Enamine compounds and would provide a useful probing tool to identify bindings to any biological target.

The Enamine Fragment Library was designed by application of "Rule of three" filters proposed by Astex Therapeutics [7] and then strict structural filters [8] to a combined dataset of Enamine Screening Compounds and Building Blocks (940,000 and 19,000 respectively at the time of the library preparation). Criteria used in ADME selection are summarized in Table 1. We had identified 6,173 compounds strictly meeting these requirements. Analysis of this set by the variable-length Jarvis-Patrick clustering [9] (Tanimoto coefficient – 0.85, portion of near neighbors overlapping – 0.5) resulted in 830 clusters and 1 819 singletones, which were refined by stringent scientific expertise to yield 1 190 structures of the Fragment Library. An extension library was selected to provide a softer focus and complement design to the main Fragment Library. The compounds in this library may violate the "Rule of three" to the extent mentioned in Table 1 at one of the six selection criteria. This library contains 11 717 compounds distributed between 1 754 clusters and 3 316 singletones. It should be noted, that both databases are composed of the stock compounds, immediately available upon ordering.

Parameter	Fragment Library	Fragment Library Extension Set
Molecular Weight	150 … 300	150 … 350
ClogP	-2 … 3	-2 … 3.5
H-Bond acceptors	0 … 3	0 … 4
H-Bond donors	0 … 3	0 … 4
Rotating Bonds	0 … 3	0 … 4
TPSA	0 … 60	0 … 90

Table 1. The parameters used in design of Fragment Libraries

The graphs (Fig. 1) illustrate the distributions of different parameters over the compounds in the Fragment Library.

Molecular weight	ClogP
Number of rotating bonds	Number of H-bond acceptors
Number of H-bond donors	TPSA

Fig. 1. Parameters of the "rule of three" showed against the number of compounds in the Enamine Fragment Library.

A very important feature of a fragment scaffold is its inherent chemical possibility of further elaboration, allowing "linking" or "growing" the fragments into leads of very high affinity. It is important to note, that the designed Fragment Library is supported by the grand selection of the Building Blocks available from Enamine. Therefore, for the vast majority of fragments in the Enamine Fragment Library there are a number of compounds with the same scaffold in the Enamine Building Blocks and Screening Databases. Availability of these derivatives is a key benefit for the FBDD: after identifying the fragment possessing an affinity to the target of interest, further elaboration of the fragment, or linking fragments are next steps in developing lead compounds. A literature example [10] of this process is given in the Scheme 1. An initial µM fragment hit 1 was elaborated stepwise to the nM lead 7. This elaboration required synthesis of all the compounds which have the same scaffold as the initial hit. Enamine could facilitate the advance by providing the intermediates and functionalized fragments promptly.

Scheme 1.

Enamine provides customers with the analogues of the fragments, which have additional functional groups, "handles" for further elaboration or linking, and even larger additional fragments. For example, Fig. 2 shows a screenshot of an Enamine Fragment Library record for a randomly chosen fragment. The fields "stock compounds with the same fragment(s)" show examples of the analogues, available from stock at Enamine. Total number of the analogues can be seen in the field "number of stock compounds based on the same core cycle(s)".

Fig. 2. Screenshot of a record in the Enamine Fragment Library.

The Fragment Library and complimentary extension can be downloaded here.

[1] Erlanson, D. A.; McDowell, R. S.; O’Brien, T. Fragment-based drug discovery. J. Med. Chem. 2004, 47, 3463-3482.

[2] Rees, D. C.; Congreve, M.; Murray, C. W.; Carr, R. Fragment-based lead discovery. Nat. Rev. Drug Discovery 2004, 3, 660-672.

[3] Shuker, S. B.; Hajduk, P. J.; Meadows, R. P.; Fesik, S. W. Discovering high-affinity ligands for proteins: SAR by NMR. Science 1996, 274, 1531-1534.

[4] Hajduk, P. J.; Greer, J. A decade of fragment-based drug design: strategic advances and lessons learned. Nat. Rev. Drug Discov. 2007, 6, 211-219.

[5] Lipinski, C. A. et al. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv. Drug. Deliv. Rev. 1997, 23, 3-25.

[6] Lipinski, C. A.; Hopkins, A. Navigating chemical space for biology and medicine. Nature, 2004, 432, 855-861.

[7] Congreve, M. et al. A rule of three for fragment-based lead discovery. Drug Discov. Today 2003, 8, 876-877.

[8] Leadlikeness and structural diversity of synthetic screening libraries. Molecular Diversity 2006, 10, No. 3, 377-388.

[9] a) Jarvis, R. A.; Patrick, E. A. Clustering Using a Similarity Measure Based on Shared Nearest Neighbors. IEEE Trans. Comput. 1973, C22, 1025-1034; b) Brown, R. D.; Martin. Y. C. Use of Structure-Activity Data To Compare Structure-Based Clustering Methods and Descriptors for Use in Compound Selection. J. Chem. Inf. Comput. Sci. 1996, 36, 572-584; c) Barnard, J. M.; Downs, G. M. Chemical Fragment Generation and Clustering Software. J. Chem. Inf. Comput. Sci. 1997, 37, 141-142.

[10] Saxty, G.; Woodhead, S. J.; Berdini, V.; Davies, T. G.; Verdonk, M. L.; Wyatt, P. G.; Boyle, R. G.; Barford, D.; Downham, R.; Garrett, M. D.; Carr, R. A. Identification of inhibitors of protein kinase B using fragment-based lead discovery. J. Med. Chem. 2007, 50, 2293-2296.

enamine.net