|Home | About | Journals | Submit | Contact Us | Français|
The stereoselective synthesis of novel photoreactive γ-secretase inhibitors 2 and 3 has been achieved. Key steps of the strategy involve preparation of α-N-Boc-epoxide 8 and formation of lactone 14 in a practical and stereo-controlled fashion. Compounds 2 and 3 are potent γ-secretase inhibitors and directly interact with presenilin-1, a catalytic subunit of γ-secretase.
γ-Secretase cleaves the amyloid precursor protein (APP) to generate β-amyloid (Aβ) peptides, which are believed to play a causative role in the pathogenesis of Alzheimer Disease (AD).1 γ-Secretase is an aspartyl protease composed of at least four proteins, including presenilin, nicastrin, APH and Pen2.2 Genetic and biochemical studies have indicated that presenilin is the catalytic core of γ-secretase3–5 and as such, familial mutations of presenilin have been associated with early on-set of AD6 through alteration of the specificity of γ-secretase. Furthermore, γ-secretase represents a novel class of protease that hydrolyzes the scissile bond within the transmembrane domain of substrate.7, 8
L-685,458 (1) (Figure 1), a potent γ-secretase inhibitor9 that contains a hydroxyethylene isostere, can be modified into a photoreactive compound by replacing an unsubstituted phenyl with a benzophenone (BP). These substitutions at P2, P1′ and P3′ have been synthesized and utilized to study γ-secretase.4, 10 However, synthesis of a photoreactive dipeptide isostere at the P1 position has not yet been achieved. In the current study, we describe the stereo-controlled synthesis of two new analogs of L-685,458 (2 and 3, Chart 1) with BPA (benzophenone alanine) at the P1 position and demonstrate that they directly interact with presenilin, the catalytic subunit of γ-secretase. Moreover, this novel BPA-Phe isostere could be useful as a functional unit to synthesize active site directed inhibitors for profiling aspartyl proteases.
The synthesis of 2 and 3 started with the preparation of epoxide 8 using a modified Barrish-Polniaszek’s method11 (Scheme 1). Methylation of Boc-p-Bz-Phe-OH (4) with TMSCHN2 in methanol12 provided methyl ester 5.
However, an attempt that followed the same synthetic route for preparation of Phe-BPA isostere10 to protect benzophenone 5 as a dioxolane using ethylene glycol, p-TsOH and benzene at reflux for 2 days failed to generate any product. Thus, we changed our strategy by reducing ketone to an alcohol. We intended to find conditions that allow for the stereo- and regioselective reduction the ketone of benzophenone. Initially, we treated 5 with NaBH4 at 0 °C13 with favorable stereoselectivity (85:15 dr) and 70% yield, but this condition also led to the formation of a small amount of reduced methyl ester. However, when we performed the same reaction at −60 °C, we obtained the stereoselective product (85:15 dr) in 86% yield without reducing the methyl ester. Silylation of the resulting secondary alcohol produced 6, which led to the generation of a chiral center at the benzylic carbon. The configuration of 6 is assigned by X-ray crystallographic analysis of intermediate 12 (Scheme 2) as described later in Figure 2. Treatment of methyl ester 7 with excess LDA/CH2ICl provided an α-chloroketone, which was reduced with NaBH4 to give chlorohydrin 7 (9:1 dr) in favor of the desired stereoisomer as demonstrated by X-Ray analysis. The two diastereoisomers of 7 were separated by column chromatography (65% yield of the desired compound, based on recovered starting material 6). Cyclization of chlorohydrin 7 produced epoxide 8 in 95% yield.14
Treatment of epoxide 8 with the sodium salt of diethyl malonate directly provided lactone 9 as a mixture of stereoisomers (Scheme 2).15 Hydrolysis of 9 with aqueous LiOH, followed by decarboxylation gave lactone 10 in 60% yield. Aldol condensation of 10 with benzaldehyde followed by dehydration with acetic anhydride-triethylamine at 120 °C gave the α,β-unsaturated lactone 11 in 80% yield.16 Hydrogenation of 11 with 10% Pd/C (1 atm, 6 h) provided lactone 12 as the sole product. The assignment of three chiral centers, as indicated in Scheme 2, was confirmed by the X-ray crystallographic analysis of 12 (Figure 2). Removal of the silyl group in lactone 12 with n-Bu4NF (TBAF) led to epimerization at the α-lactone position, perhaps due to the basicity of the TBAF reagent. However, we were able to find that treatment of 12 with pyridine/HF overnight successfully removed the silyl protecting group to give 13 without any epimerization.17 Oxidation of the benzylic alcohol with MnO218 gave benzophenone 14 in 76% yield.19 Hydrolysis of lactone 14 with LiOH and silylation of the resulting hydroxy acid produced 15 in 79% yield.
Esterification of Leu-Phe-OH with TMSCl in MeOH,20 followed by coupling of the resulting amine with acid 15, and deprotection of the resulting silyl ether with TBAF, produced the desired compound 2 in reasonable yield (Scheme 3).21 In order to facilitate the purification of the labeled proteins or fragments thereof, biotinylated compound 3 was prepared (Scheme 3). Mild saponification of the methyl ester in 2 led to the corresponding carboxylic acid, which was coupled with 5-(biotinamido)pentylamine in the presence of EDC and HOBt and resulted in compound 3.
We next examined the biological activities of 2 and 3. First, we determined their inhibitory potency against γ-secretase using an in vitro assay.22 The IC50 values of 2 and 3 are 0.7 nM and 0.6 nM, respectively (Fig. 3A), which is similar to the parent compound, L-685,458 (1). These findings have demonstrated that incorporating BPA into the P1 position and attaching a biotin tag at the C-terminus do not affect their potency for inhibition of γ-secretase. Second, we tested whether 3 was capable of photo-crosslinking to γ-secretase. HeLa cell membranes were incubated with 3 at a final concentration of 10 nM in the absence and the presence of 2 μM of L-685,458 for 2.5 h. Then samples were irradiated with UV light (> 350 nm) and the labeled proteins were solubilized and isolated with streptavidin beads.4
The biotinylated proteins were eluted and analyzed by Western blotting with antibodies against presenilin-1 (PS-1). Inhibitor 3 directly photolabels PS-1 (Fig. 3C). Moreover, an excess of L-685,458 is able to block photoinsertion of this probe into presenilin-1. Taken together, these results have demonstrated that compounds 2 and 3 are potent γ-secretase inhibitors that can specifically label the catalytic core of γ-secretase. Therefore, compounds 2 and 3 should be valuable probes for mapping the active site of γ-secretase.
This work was supported by NIH grant AG026660 (YML), the Alzheimer’s Association (Zenith Fellows Award to YML), Mr. William H. Goodwin and Mrs. Alice Goodwin and the Commonwealth Foundation for Cancer Research, the Experimental Therapeutics Center of MSKCC, and the William Randolph Hearst Fund in Experimental Therapeutics. CCS is supported by an NIH NRSA predoctoral fellowship 5F31NS053218-02. We thank Dr. George Sukenick, Ms. Sylvi Rusli (NMR Core Facility, Sloan-Kettering Institute) for mass spectral analyses and Dr. Louis Todaro (Hunter College New York) for X-ray structure analyses.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.