Title: The Core Structure of TMC-95A Is a Promising Lead for Reversible Proteasome Inhibition This work was supported by the SFB 469 of the Ludwig-Maximilians-Universität München and the SPP 1045.
Abstract: Based on information from the X-ray structure of the TMC-95A/proteasome complex, TMC-95A (A) was reduced to the minimum core structure (B). The inhibitory potency of the synthetic analogue confirms that the simplified ring structure may well serve as the lead for further improvement of affinity and selectivity of reversible proteasome inhibitors. The proteasome is an intracellular multicatalytic protease complex which in combination with the ubiquitin pathway plays a central role in major cellular processes, such as antigen presentation, cell proliferation and differentiation, and apoptosis.1 Proteolysis occurs in a barrel-shaped core structure known as 20S proteasome, which consists of four stacked rings arrayed in an α7β7β7α7 mode.2a In eukaryotic proteasome three β-subunits of each β-ring are enzymatically active with an N-terminal threonine residue as the active nucleophile involved in proteolysis2b with three more or less distinct substrate specificities, that is chymotrypsin-like (CL), trypsin-like (TL), and peptidyl-glutamyl-peptide hydrolase (PGPH) activities.3 Because of the physiological role of proteasome in critical intracellular processes, this enzyme represents a promising target for drug development in inflammatory and autoimmune diseases as well as in tumor therapy.4 Correspondingly, great attention has recently been paid to the discovery of potent and selective proteasome inhibitors by structure-based design or natural product screening approaches. Most of the synthetic inhibitors consisting of peptide aldehydes, boronates, and vinylsulfones, as well as the natural products lactacystin and epoxymicins inhibit in a more or less selective manner the proteasome by reaction with the N-terminal threonine residue (for a recent review see ref. 5). A notable exception is the highly selective and competitive proteasome inhibitor TMC-95A, which was isolated from the fermentation broth of Apiospora montagnei Sacc. TC 1093.6 This cyclic peptide metabolite consists of L-tyrosine, L-asparagine, a highly oxidized L-tryptophane, and the (Z)-1-propenylamine and 3-methyloxopentanoic acid moities with a phenyl/oxindole ring junction (structure A in Scheme 1).7 A similar biaryl moiety was so far encountered only in chloropeptin,8 complestatin,9 diazonamide,10 and the kistamicins.11 Because of the pharmacological interest of proteasome inhibition and the distinct phenol/oxindole system the synthesis of TMC-95A has attracted considerable interest.12 Structure of TMC-95A (A), the minimal skeleton (B) for binding to the proteasome as derived from the X-ray structure of the TMC-95A/proteasome complex, and the analogue (C) with propylamide as R1 and the side chain of asparagine as R3 residue. Compound C was synthesized for the validation of the design concept. X-ray structural analysis of the complex of TMC-95A with yeast proteasome clearly confirmed the noncovalent binding of this low molecular weight cyclic peptide, with a network of hydrogen bonds between the β-type extended peptide moiety and the main chain of the protein13 which contributes to the high affinity for all three active sites.6 From this crystal structure the minimum structural elements of TMC-95A for binding to the proteasome were derived (structure B in Scheme 1). To validate our working assumption we selected compound C (Scheme 1) as our first synthetic target; compound C contains a propylamide residue as R1 for interaction with the S1 pocket of the active sites, and the side chain of asparagine, which is present in the natural product, as R3 for occupancy of the S3 pocket. All other modifications at the level of the core structure of TMC-95A were not expected to significantly affect the binding affinity for proteasome. The synthesis of this first TMC-95A analogue was accomplished and its inhibitory potency fully confirmed our working hypothesis. The main difficulty in the synthesis of compound C was expected from cyclization to the constraint ring structure. Based on previous experiences reported for the synthesis of the chloropeptin ring,14 a first synthetic route by acid-catalyzed ring contraction of an 18-membered macrolactame precursor was attempted. This foresees a Suzuki cross-reaction for the preparation of the biaryl system followed by lactamization of the linear precursor. As meanwhile also reported by Estiarte et al.,15 cyclization of the 18-membered precursor failed. Thus, the compound C was synthesized according Scheme 2 and 3 by formation of the “correct” biaryl junction at the level of the linear precursor followed by cyclization. For this synthetic route 3-iodo-L-tyrosine was converted into the suitably protected derivative 2 which was then transformed to the aryl boronate 3 by the Miyauri–Suzuki reaction16 using bis(pinacolato)diboron, [PdCl2(dppf)], and KOAc (Scheme 2). 7-Bromo-L-tryptophane was produced from 7-bromoindole and L-serine by the use of tryptophane synthetase following essentially known protocols17 and then converted into the related Nα-Boc propylamide derivative 6 (Scheme 2). The derivatives 3 and 6 were coupled under standard conditions of the Suzuki cross-reaction18 to generate the key intermediate 7 in 80 % yield after flash chromatography (Scheme 3). Upon C-terminal extension of the peptide chain with the asparagine tert-butyl ester, the linear precursor 8 was oxidized at the indole moiety by standard procedures with DMSO/concentrated HCl19 (40 % yield) prior to peptide cyclization with PyBOP/HOBt/DIEA. After HPLC purification the desired compound C (10) was isolated in about 40 % yield as a homogeneous product (Scheme 3). NMR analysis confirmed the correct ring structure including the S configuration at the C3 atom of the oxindole.20 Apparently only this configuration allows cyclization of the precursor into the highly restricted ring. From distance and dihedral angle constraints derived from 2D NMR spectra a spatial structure of compound C was calculated which is superimposable to that of TMC-95A determined by X-ray analysis of its complex with the proteasome (Figure 1). Synthesis of 3 and 6: a) SOCl2, MeOH; b) Z-OSu; c) K2CO3 (10 equiv), MeI (5 equiv); d) B2O4C12H24 (1.1 equiv), [Pd(dppf)Cl2]⋅CH2Cl2 (5 mol %), KOAc (3 equiv); e) L-serine, tryptophane-synthetase, PLP, 37 °C; f) Boc2O; g) nPrNH2, EDCI, HOBt. Z-OSu=Nα-(benzyloxycarbonyloxy)succinimide, dppf=Ph2PC5H4FeC5H4PPh, PLP=pyridoxal 5′-phosphate, Boc=tert-butoxycarbonyl, EDCI=N′-(3-dimethylaminopropyl)-N-ethylcarbodiimide, HOBt=1-hydroxy-1H-benzotriazole, DME=dimethoxyethane. Synthesis of 10. a) 3 (1.1 equiv), [Pd(dppf)Cl2]⋅CH2Cl2 (5 mol %), K2CO3 (3 equiv), DME/H2O (7:1), 70 °C; b) H-Asn-OtBu, EDCI, HOBt; c) DMSO (20 equiv), AcOH/HCl (4:1); d) PyBOP (4 equiv), HOBt (4 equiv), DIEA (6 equiv). PyBOP=benzotriazol-1-yloxytripyrrolidinophosphonium hexafluorophosphate, DIEA=diisopropylethylamine. Conformation of compound C determined by NMR spectroscopy (20 lowest energy structures; gray) superimposed on the three-dimensional structure of TMC-95A as determined by X-ray analysis of its complex with yeast proteasome (black); for better visualization the N-terminal phenyl group is omitted. In the context of the structural analysis of the TMC-95A/yeast proteasome complex the natural product was found to inhibit the three proteolytic activites of yeast and human proteasome with similar potencies13 as previously reported for human proteasome.6 The inhibitory effects of compound C are reported in Table 1, and compared with those of TMC-95A and the calpain inhibitor I, that is Ac-Leu-Leu-Nle-H. Despite the difficult quantitative comparison of IC50 values obtained with different proteasome preparations and assay conditions, the synthetic TMC-95A analogue was found to inhibit all three proteolytic activities more efficiently than the tripeptide aldehyde. Compared to the natural product, compound C retains almost full inhibition of the TL and PGPH activities, but it is significantly less potent against the CL activity. This may well be attributed to the exchange of the (Z)-1-propenylamide with the more flexible propylamide as P1 residue for occupancy of the specificity subsite S1 of the enzyme. Full retainment of affinity of compound C for the two other active sites of the proteasome confirms that the N-terminal acyl residue as well as the tyrosine hydroxy group are not critically involved in the interaction with the protein counterpart, whilst the biaryl moiety restricts the peptide backbone into the extended β-strand conformation for optimal hydrogen bonding to the active site clefts. The data also confirm that the degree of oxidation of the tryptophane residue can be reduced and that the oxindole group is sufficient for the additional hydrogen bond to the protein backbone to be established. CL activity TL activity PGPH activity compound C 8.0[a] 10.6[a] 7.4[a] 1.9[b] 3.7[b] 1.8[b] TMC-95A6 0.012[b] 1.5[b] 6.7[b] Ac-Leu-Leu-Nle-H 35.4[a] 142.5[a] 88.0[a] In conclusion, the inhibitory potencies of compound C confirm the correctness of our inhibitor design based on minimization of the TMC-95A skeleton, and clearly indicate that by optimization of the R1 and R3 residues both selectivity and affinity for the three active sites may significantly be improved. These results also suggest that not all of the complex structural elements of the natural product are required for inhibition of proteasome, thus markedly facilitating the synthesis of TMC-95A related compounds.