Xylose-Configured Cyclophellitols as Selective Inhibitors for Glucocerebrosidase

Abstract

Glucocerebrosidase (GBA), a lysosomal retaining β-d-glucosidase, has recently been shown to hydrolyze β-d-xylosides and to transxylosylate cholesterol. Genetic defects in GBA cause the lysosomal storage disorder Gaucher disease (GD), and also constitute a risk factor for developing Parkinson's disease. GBA and other retaining glycosidases can be selectively visualized by activity-based protein profiling (ABPP) using fluorescent probes composed of a cyclophellitol scaffold having a configuration tailored to the targeted glycosidase family. GBA processes β-d-xylosides in addition to β-d-glucosides, this in contrast to the other two mammalian cellular retaining β-d-glucosidases, GBA2 and GBA3. Here we show that the xylopyranose preference also holds up for covalent inhibitors: xylose-configured cyclophellitol and cyclophellitol aziridines selectively react with GBA over GBA2 and GBA3 in vitro and in vivo, and that the xylose-configured cyclophellitol is more potent and more selective for GBA than the classical GBA inhibitor, conduritol B-epoxide (CBE). Both xylose-configured cyclophellitol and cyclophellitol aziridine cause accumulation of glucosylsphingosine in zebrafish embryo, a characteristic hallmark of GD, and we conclude that these compounds are well suited for creating such chemically induced GD models.

Keywords: Gaucher disease; activity-based probe; conduritol B-epoxide; cyclophellitol; glucocerebrosidase; xylose.

Figure 1

(A) Irreversible inhibition by cyclophellitol and cyclophellitol‐aziridine compounds. (B) Reactivity of GBA, GBA2 and GBA3 β‐d‐glucosidases with epoxide and aziridine‐based ABPs 9 and 10, R=Cy5.

Figure 2

Structures of cyclophellitol epoxide and aziridines subject of the research described in this paper.

Figure 3

Selectivity of compounds visualized by competitive ABPP labeling of β‐d‐glucosidase. Lysates of HEK293T cells expressing human GBA, GBA2 and GBA3 were incubated with compounds 1–5 at indicated concentrations for 30 min, following by cABPP with ABP 10.

Figure 4

(A) ABP labeling of β‐d‐glucosidases. A lysate of HEK293T cells expressing human GBA, GBA2 and GBA3 was incubated with indicated ABPs (6, 9 or 10) for 30 min at pH 6.0. Fluorescently labeled proteins were visualized after SDS‐PAGE. (B) Labeling with ABP 6 of wild type or mutant GBA2 (E527G nucleophile mutation and D667G acid/base mutation) expressed in HEK293T cells.

Figure 5

Inhibitory effect of β‐d‐xylo‐configured cyclophellitol 1 and cyclophellitol aziridine 2 on β‐glucosidases in intact HEK293T cells expressing GBA, GBA2 and GBA3. (A) Representative gel images of cABPP where cells were treated for 24 h with indicated inhibitor. Lysates were then prepared and labeled with fluorescent ABP 10. Fluorescently labeled proteins were visualized after SDS‐PAGE (1 set from n=3 replicates). (B) IC₅₀ curves determined by cABPP labeling results. (C) Apparent IC₅₀ values towards β‐glucosidases in intact HEK293T cells producing GBA, GBA2 and GBA3 were determined by the fluorescence quantification based on cABPP SDS‐PAGE results.

Figure 6

In vivo inhibitory effect of β‐d‐xylo‐configured compounds on β‐glucosidases in zebrafish (Danio rerio) larvae. (A) Larvae were exposed to 5 dpf with the indicated inhibitor 1 or 2 in the medium. Larvae were lysed and incubated with fluorescent ABP 10. Fluorescently labeled proteins were visualized after SDS‐PAGE, only GBA and GBA2 were assessed in zebrafish larvae model. (B) Apparent IC₅₀ values towards β‐glucosidases (GBA and GBA2) were determined by the fluorescence quantification based on cABPP results. (C) In vivo inhibition curves. (D) GlcSph levels in zebrafish larvae were determined as described in experiment section, n=2 replicates.