We therefore prepared a series of oxalyl derivatives 27C32 from merging the bicyclic ring of fragment 3 with the cyclopentyloxalyl core of 8 with methylene and ethylene linkers according to Scheme 2 .23 Open in a separate window Scheme 2 Reagents and conditions: (a) Montmorillonite in MeCN at RT for 3?d; (b) TFA in DCM at 0?C??RT for 1?h; (c) EDCI, HOPO and DIPEA in DMF at RT for 24?h. The most potent analogue 27 achieved a >1000-fold potency enhancement in the biochemical assays (FP IC50?=?2?M, PPIase IC50?=?1.5?M) as well as in the SPR assay (KD?=?2.8?M) compared to the millimolar potencies of the individual fragments 3 and 8 (Table 3 ). Table 3 CypD SAR of optimized oxalamides derived from merging the fragment hits 3 or 4 4 with fragment hit 8.
2720.2126.96.36.19928440.22nt230.2129>100nt>1500308700.16nt2170.1931>100nt1840.1632>30nt>240 Open in a separate window a,bCypD biochemical FP and PPIase assays. processes such as inflammation and vascular dysfunction, wound healing, innate HIV immunity, hepatitis C infection, host-parasite interactions and tumor biology.7 Cyclophilin D (CypD) is the mitochondrial isoform of the enzyme, and a key regulator of the mitochondrial permeability transition pore. Mitochondrial dysfunction has been implicated in a cascade of cellular processes linked to multiple sclerosis and cardiovascular disease, making CypD a therapeutic drug target.7, 8, 9, 10 The crystal structures of several cyclophilins have been determined and show a common fold consisting of two -helices packing against an eight-stranded anti-parallel P-barrel structure.11 The cyclophilins contain a large active binding groove composed by several highly conserved Xylazine HCl hydrophobic, aromatic and polar residues including the catalytic Arg55 located at the entrance Xylazine HCl of the S1 proline pocket.2, 12 A second S2 pocket has been identified nearby: it is deep and relatively non-specific, with access controlled by a set of gatekeeper residues.2 The cyclic peptide CsA binds via specific interactions involving both S1 and S2 pockets with nanomolar potency to cyclophilins, e.g. to CypD with a PPIase IC50 of 20?nM.13 However, CsA and its semisynthetic analogues such as Debio 025 and NIM811 have unfavorable drug-like properties due to high molecular weight, limited solubility and poor BTLA bioavailability.14, 15 Only few small and non-peptidic CypD inhibitors have been published including urea derivatives such as 2, which were discovered by fragment-based lead discovery (Fig. 1).10, 16, 17 These urea derivatives demonstrated in vitro PPIase inhibitory activity and antiviral activity against hepatitis C virus, human immunodeficiency virus and coronaviruses.16 Protein crystallography of 2 in CypD revealed specific binding of the pyrrolidine ring in the S1 pocket, while the aniline substituent is bound in the S2 pocket (Supporting information).13 Our aim was to identify novel chemical hit matter from HTS and fragment screening approaches to develop CypD inhibitors with drug-like properties for prevention of mitochondrial dysfunction in multiple sclerosis. Open in a separate window Fig. 1 Published CypD inhibitors (1C2). We started our hit identification efforts by high-throughput screening on our corporate compound library with ~650,000 compounds using an FP biochemical assay, which resulted in only a small hit rate of 178 hits with IC50s?10?M. Disappointingly, none of these hits could be confirmed in orthogonal biophysical CypD binding assays using surface plasmon resonance (SPR) and protein-based NMR studies. Due to this outcome, we conducted an additional fragment-based screening campaign using our internal fragment library with 2688 structurally diverse fragments (Supporting information). The fragments were screened by SPR at fixed concentrations of 2?mM using immobilized CypD protein and yielded 168 primary hits. For subsequent hit confirmation, we used CsA at 200?nM for SPR-based competition experiments in compound titration series of 10 concentrations up to 10?mM. The affinity determination by SPR confirmed 58 hits with steady state dissociation constants (KD,ss) in the range of 1 1?mM to >10?mM. The identified fragments represented a large chemical diversity consisting of different aromatic as well as saturated rings as potential proline-mimicking motifs. However, the fragments had only millimolar potencies and overall low ligand efficiencies (LEs 0.1C0.3?kcal/heavy atom) beyond the high LE range of >0.3?kcal/heavy atom considered as optimal starting point for fragment optimization.18, 19 We therefore aimed to determine the binding mode in the CypD binding groove for as many fragments as possible by protein crystallography for structure-guided optimization. We evaluated 52 fragments by co-crystallization and by soaking into apo crystals of the CypD K175I mutant and obtained 6 crystal structures with clearly defined fragment electron densities in the active site at resolutions of 1 Xylazine HCl 1.15C2.0?? (Table 1 and Supporting information).20 The 6 fragments displayed a certain variety of binding modes within the CypD binding groove: 3 and 4 are bound in the gatekeeper S2 pocket, 5C7 are located in the proline S1 pocket and 8 is targeting Xylazine HCl both S1 and S2 pockets (Supporting information). All fragment X-ray structures were superimposed with published CypD structures in complex with CsA and urea derivatives such as 2 to define promising fragment linking and merging strategies for hit optimization. These considerations provided the basis of three hit series followed up by medicinal chemistry to improve potency in the biochemical FP and SPR binding assays. Table 1 Overview of SPR-confirmed hits from fragment screening against human CypD confirmed by X-ray crystallography.