Cover Page
The handle
http://hdl.handle.net/1887/87273
holds various files of this Leiden University
dissertation.
Author: Lahav, D.
Fluorescence Polarization Activity-Based Protein
Profiling on Retaining Glycosidases
PROEFSCHRIFT ter verkrijging van
de graad van Doctor aan de Universiteit Leiden, op gezag van Rector Magnificus prof. mr. C. J. J. M. Stolker,
volgens het besluit van het College voor Promoties te verdedigen op woensdag 8 april 2020
klokke 16.15 uur
door
Daniël Lahav
Promotor: Prof. dr. H. S. Overkleeft
Copromotor: Prof. dr. J. M. F. G. Aerts
Promotiecommissie: Prof. dr. G. A. van der Marel Prof. dr. M. van der Stelt Overige commissieleden: Dr. J. D. C. Codée
Prof. dr. S. Py
Dr. G. Schulzenbacher Prof. dr. J. Brouwer Prof. dr. H. Ovaa
ISBN: 978-94-028-1979-3
Printed by: Ipskamp Printing
Chapter 1 7 General Introduction
Chapter 2 17
A Fluorescence Polarization Activity-Based Protein Profiling
Assay in the Discovery of Potent, Selective Inhibitors for
Human Non-Lysosomal Glucosylceramidase
Chapter 3 49
FluoPol ABPP on Human Lysosomal β-Glucosylceramidase
Identifies Inhibitors from the LOPAC Library
Chapter 4 73
Chapter 5 95
Competitive Fluorescence Polarization Activity Based Protein
Profiling allows the Discovery of New Golgi α-Mannosidase
Inhibitors
Chapter 6 117
Summary and Future Prospects
Samenvatting 123
List of Publications 125
Curriculum Vitae 126
7
1
General Introduction
Glycosidases are important enzymes in the turnover of polysaccharides and glycoconjugates, and are involved in a range of human pathologies including genetic disorders such as Gaucher and Pompe disease, but also in various cancers.1,2 The discovery of potent and selective glycosidase inhibitors for fundamental glycobiology studies and as leads for drug discovery requires access to suitable (glycomimetic) compound libraries, as well as easily applicable assays and screening formats. The research described in this Thesis was aimed to explore how covalent and irreversible glycosidase inhibitors can be applied in the screening of focused compound libraries on various glycosidases using fluorescence polarization activity-based protein profiling (FluoPol-ABPP).
1.1 Detection methods within activity-based protein profiling
Activity-based protein profiling (ABPP) is a powerful method for determining the number of active copies of specific enzymes in a given biological sample and under physiological conditions.3 Protein expression levels do not always correlate to the amount of active enzyme molecules, which might be regulated by post-translational modifications, protein-protein interactions or endogenous small-molecule based inhibition. Measuring mRNA or protein levels could therefore lead to misinterpretation of true enzyme activity levels and maybe even to misinterpreting the physiological role of an enzyme (or enzyme family).
8
enzyme-probe complex.6 The type of detection depends on the reporter group.
Probes containing a fluorophore can be detected after separation of the protein content of the treated proteome via gel-electrophoresis, as in SDS-PAGE, using fluorescent imaging (Figure 1A). Another commonly used procedure is to detect and identify ABP-modified enzymes by mass spectrometry (MS)-based proteomics. Protein that have reacted with biotinylated probes are enriched via streptavidin pull-down and then digested with trypsin, producing peptide fragments derived from the ABP labelled protein(s). These fragments are subsequently identified with MS/MS sequencing (Figure 1B). Combination-ABPs containing a fluorophore and a biotin exist as well, and bioorthogonal chemistry can be employed where direct labelling falls short.7
Figure 1. Schematic representation of a typical ABPP experiment where labelled enzymes are either
directly visualized with SDS-PAGE followed by fluorescent imaging (A) or analysed by mass spectrometry (B). When using FluoPol as readout (C), high polarization signals are obtained when a high fluorescent signal parallel to the plane of excitation is retrieved. Generally this occurs when the apparent size of the fluorophores is large and thus when molecular mobility is relatively low. Free probe (bottom situation) will lead to a depolarized signal due to increased molecular tumbling and scattering of fluorescent light.
9 (in part) formation of probe-enzyme complex, with no (or diminished) in-gel fluorescence as the result. In such a competitive ABPP experiment the probe binding, reflected by the intensity of in-gel fluorescence, is inversely proportional to the degree of competition that occurs, and thus inhibitory potency of the competing compound. Gel-ABPP experiments are time-consuming and labour intensive processes when examining large numbers of compounds. Fluorescence polarization (FluoPol) as readout to monitor probe binding to the target enzyme allows the development of high-throughput screening compatible assays within the ABPP methodology, allowing fast identification of inhibitors or active site competitors.
In FluoPol, excitation of the fluorophore occurs with linear polarized light. The orientation of emitted light and thereby the degree of polarized light parallel to the plane of excitation depends on the movement the excited fluorophore has undergone, which in turn depends on the size of the fluorophore. Large fluorescent moieties, such as those consisting of a fluorescent ABP covalently and irreversibly bound to a target enzyme, move and rotate slower compared to the (much smaller) unreacted and unbound fluorescent ABP. In competitive FluoPol ABPP (inclusion of putative inhibitors prior to exposure of the enzyme of interest to the fluorescent ABP) the degree of inhibition is inversely proportional to the FluoPol-signal. As FluoPol can be performed in microtiter plates, it is possible to screen large compound libraries efficiently for the identification of competitors that reduce or enhance enzyme labelling.
1.2 Targeting enzyme families via activity-based protein profiling
10
of the cysteine thiol following a sequence of events also employed by the cysteine-containing cathepsins shown in entry 2.10 These studies show that specific classes of enzymes can be targeted using specific electrophiles. Introducing electrophilic traps within substrate mimetics, such as biotinylated acyl phosphates (entry 5) of ATP or ADP described by Patricelli et al.,11 demonstrates that this strategy also holds true for kinases. The work described in this Thesis builds on the successful design of electrophilic traps for yet another class of enzymes: the retaining glycosidases.
Figure 2. Structure and working mechanism of serine hydrolase probes (1), cathepsin probes (2 and 3), probes for ubiquitin specific proteases (4) and kinases (5).
1.3 Development of cyclophellitol-based ABPs for retaining glycosidases
β-11 glucosidase inhibition, and that inhibition proceeded through the mechanism as depicted in Figure 3A.13 Hansen et al.14 described in 2005 an efficient synthesis of large quantities of this natural product, starting from D-xylose. Though not the first synthesis of this natural product, the authors were able to produce useful quantities of the natural product following a route that appeared attractive for adaptation towards structural, configurational and functional cyclophellitol analogues.15 The synthetic material obtained in the Madsen studies was used to reveal that retaining β-glucosidase from Thermotoga maritima reacts in a mechanism-based, covalent and irreversible manner with cyclophellitol (6) to form a covalent adduct, with the reacted inhibitor adopting a 4C1 chair conformation, as
witnessed by crystal studies on the enzyme-inhibitor adduct (figure 3A).16
This binding mode, indeed this mode of reaction of cyclophellitol with retaining β-glucosidases can be understood by perusing the molecular pathway through which retaining β-glucosidases process their substrates. Hydrolysis of interglycosidic linkages by retaining glycosidases mostly occurs via a two-step, double displacement mechanism involving a covalent glycosyl-enzyme intermediate as was proposed by Koshland in 1953.17,18 For retaining β-glucosidases, in the first step, (the glycosylation step), the aglycon is protonated through the action of an acid-base residue (either a glutamic acid or aspartic acid depending on the nature of the retaining β-glucosidase) and expelled leaving an oxocarbenium-ion like intermediate. In this transition state, the oxocarbenium ion can adopt a 4H3
half-chair conformation, (see figure 3B) that is then trapped by the catalytic nucleophile (aspartate or glutamate) to form the covalent adduct in a 4C1 chair conformation. In
the second displacement reaction, the deglycosylation step, hydrolysis of the formed covalent intermediate also occurs via an oxocarbenium ion intermediate in
4
H3 conformation. The covalent glycosyl-enzyme intermediate formed in the
12
Figure 3. (A) Chemical structures in of (+)-cyclophellitol (6). Also the mechanism-based inhibition is
depicted. (B) Mechanism of hydrolysis of β-glucosidic linkages by retaining β-glucosidases.
13 following formation of a covalent enzyme-glycoside intermediate and in many cases are bona fide irreversible inhibitors – thus allowing in-depth structural studies. The initial binding and reaction rate is however also slow, much slower than that of cyclophellitol-based compounds. Thus, and though activity-based probes based on 2-deoxy-2-fluoroglycosides and related compounds have been put forward by the groups of Vocadlo and Bertozzi,22 these studies have not been followed up much and this may be because deoxy-fluoroglycosides are – as evident from the above – often relatively poor retaining glycosidase inhibitors.24 In contrast, ABPs based on cyclitol epoxides and especially cyclitol aziridines have been developed in recent years that target a rather wide range of retaining glycosidases. These include GH1 β-glucosidases25, GH27 α-galactosidases26 and GH29 α-fucosidases27, retaining glycosidases that can be addressed by cyclophellitol-aziridine ABPs featuring configuration and substitution pattern
emulating that of the underlying substrate glycosides.28
Figure 4. Chemical structures of cyclophellitol (7) and cyclophellitol aziridine (8)-based probes with
two typical reporter groups (Tags) depicted in the middle. On the right two deoxy-fluoro-glycosides, 9 and 10, based on which activity-based glycosidase probes can be developed as well.
1.4 Aim and outline of this thesis
The research described in this thesis focuses on the development of FluoPol-ABPP assays for the discovery of competitive inhibitors for a number of retaining glycosidases involved in human disease. In Chapter 2 the development of FluoPol-ABPP for the identification of inhibitors of the human cytosolic retaining β-glucosidase, GBA2 is described. These studies, apart from yielding known and new competitive GBA2 inhibitors, also paved the way for the development of related assays directed at other retaining glycosidases. Chapter 3 describes the identification, using a FluoPol-ABPP assay format, of GBA1 inhibitors from the Library of Pharmaceutically Active Compounds (LOPAC1280, Sigma). Several of these
14
inhibitors that might be used as pharmacological chaperones (stabilizing endogenous lysosomal retaining α-glucosidase) for the treatment of Pompe disease. Chapter 5 describes the development of FluoPol-ABPP on Golgi α-mannosidase II (GMII) from drosophila, a homologue of human GMII, which is a pharmacological target for chemotherapeutic agents. In each chapter a screen of an in-house (Leiden Institute of Chemistry, Leiden University) amassed iminosugar library, containing 358 entries as shown in appendix A, is performed with the optimal FluoPol-ABPP conditions. Chapter 6 gives a summary of the results described in this Thesis and discusses some future prospects based on the presented results.
1.5 References
(1) Wennekes, T., van den Berg, R. J. B. H. N., Boot, R. G., van der Marel, G. A., Overkleeft, H. S., and Aerts, J. M. F. G. (2009) Glycosphingolipids - nature, function, and pharmacological modulation. Angew. Chem. - Int. Ed. 48, 8848–8869.
(2) Ernst, B., and Magnani, J. L. (2009) From carbohydrate leads to glycomimetic drugs. Nat. Rev. Drug Discov. 8, 661–677.
(3) Evans, M. J., and Cravatt, B. F. (2006) Mechanism-based profiling of enzyme families. Chem. Rev. 106, 3279–3301.
(4) Willems, L. I., Overkleeft, H. S., and van Kasteren, S. I. (2014) Current developments in activity-based protein profiling. Bioconjug. Chem. 25, 1181–1191. (5) Zweerink, S., Kallnik, V., Ninck, S., Nickel, S., Verheyen, J., Blum, M., Wagner, A.,
Feldmann, I., Sickmann, A., Albers, S.-V., Bräsen, C., Kaschani, F., Siebers, B., and Kaiser, M. (2017) Activity-based protein profiling as a robust method for enzyme identification and screening in extremophilic Archaea. Nat. Commun. 8, 15352. (6) Serim, S., Haedke, U., and Verhelst, S. H. L. (2012) Activity-based probes for the
study of proteases: recent advances and developments. ChemMedChem 7, 1146– 1159.
(7) Willems, L. I., van der Linden, W. A., Li, N., Li, K.-Y., Liu, N., Hoogendoorn, S., van der Marel, G. A., Florea, B. I., and Overkleeft, H. S. (2011) Bioorthogonal chemistry: applications in activity-based protein profiling. Acc. Chem. Res. 44, 718–729. (8) Liu, Y., Patricelli, M. P., and Cravatt, B. F. (1999) Activity-based protein profiling:
The serine hydrolases. Proc. Natl. Acad. Sci. USA 96, 14694–14699.
15
electrophilic substrate analogs. Chem. Biol. 7, 27–38.
(10) Borodovsky, A., Kessler, B. M., Casagrande, R., Overkleeft, H. S., Wilkinson, K. D., and Ploegh, H. L. (2001) A novel active site-directed probe specific for deubiquitylating enzymes reveals proteasome association of USP14. EMBO J. 20, 5187–5196.
(11) Patricelli, M. P., Szardenings, A. K., Liyanage, M., Nomanbhoy, T. K., Wu, M., Weissig, H., Aban, A., Chun, D., Tanner, S., and Kozarich, J. W. (2007) Functional interrogation of the kinome using nucleotide acyl phosphates. Biochemistry 46, 350–358.
(12) Atsumi, S., Umezawa, K., Iinuma, H., Naganawa, H., Nakamura, H., Iitaka, Y., and Takeuchi, T. (1990) Production, isolation and structure determination of a novel beta-glucosidase inhibitor, cyclophellitol, from Phellinus sp. J. Antibiot. (Tokyo). 43, 49–53.
(13) Withers, S. G., and Umezawa, K. (1991) Cyclophellitol: a naturally occurring mechanism-based inactivator of beta-glucosidases. Biochem. Biophys. Res.
Commun. 177, 532–537.
(14) Hansen, F. G., Bundgaard, E., and Madsen, R. (2005) A short synthesis of (+)-cyclophellitol. J. Org. Chem. 70, 10139–10142.
(15) Jiang, J., Artola, M., Beenakker, T. J. M., Schröder, S. P., Petracca, R., de Boer, C., Aerts, J. M. F. G., van der Marel, G. A., Codée, J. D. C., and Overkleeft, H. S. (2016) The synthesis of cyclophellitol-aziridine and its configurational and functional isomers. Eur. J. Org. Chem. 2016, 3671–3678.
(16) Gloster, T. M., Madsen, R., and Davies, G. J. (2007) Structural basis for cyclophellitol inhibition of a β-glucosidase. Org. Biomol. Chem. 5, 444–446.
(17) Koshland, D. E. (1953) Stereochemistry and the mechanism of enzymatic reactions.
Biol. Rev. 28, 416–436.
(18) Davies, G., and Henrissat, B. (1995) Structures and mechanisms of glycosyl hydrolases. Structure 3, 853–859.
(19) Witte, M. D., Kallemeijn, W. W., Aten, J., Li, K.-Y., Strijland, A., Donker-Koopman, W. E., van den Nieuwendijk, A. M. C. H., Bleijlevens, B., Kramer, G., Florea, B. I., Hooibrink, B., Hollak, C. E. M., Ottenhoff, R., Boot, R. G., van der Marel, G. a, Overkleeft, H. S., and Aerts, J. M. F. G. (2010) Ultrasensitive in situ visualization of active glucocerebrosidase molecules. Nat. Chem. Biol. 6, 907–913.
16
(21) Withers, S. G., Rupitz, K., and Street, I. P. (1988) 2-Deoxy-2-fluoro-D-glycosyl fluorides. A new class of specific mechanism-based glycosidase inhibitors. J. Biol.
Chem. 263, 7929–7932.
(22) Vocadlo, D. J., and Bertozzi, C. R. (2004) A strategy for functional proteomic analysis of glycosidase activity from cell lysates. Angew. Chem. Int. Ed. 43, 5338– 5342.
(23) Willems, L. I., Beenakker, T. J. M., Murray, B., Gagestein, B., van den Elst, H., van Rijssel, E. R., Codée, J. D. C., Kallemeijn, W. W., Aerts, J. M. F. G., van der Marel, G. A., and Overkleeft, H. S. (2014) Synthesis of α- and β-galactopyranose-configured isomers of cyclophellitol and cyclophellitol aziridine. Eur. J. Org. Chem. 2014, 6044– 6056.
(24) Walvoort, M. T. C., Kallemeijn, W. W., Willems, L. I., Witte, M. D., Aerts, J. M. F. G., Marel, G. A. van der, Codée, J. D. C., and Overkleeft, H. S. (2012) Tuning the leaving group in 2-deoxy-2-fluoroglucoside results in improved activity-based retaining β-glucosidase probes. Chem. Commun. 48, 10386.
(25) Li, K.-Y., Jiang, J., Witte, M. D., Kallemeijn, W. W., van den Elst, H., Wong, C.-S., Chander, S. D., Hoogendoorn, S., Beenakker, T. J. M., Codée, J. D. C., Aerts, J. M. F. G., van der Marel, G. A., and Overkleeft, H. S. (2014) Synthesis of cyclophellitol, cyclophellitol aziridine, and their tagged derivatives. Eur. J. Org. Chem. 2014, 6030– 6043.
(26) Willems, L. I., Beenakker, T. J. M., Murray, B., Scheij, S., Kallemeijn, W. W., Boot, R. G., Verhoek, M., Donker-Koopman, W. E., Ferraz, M. J., van Rijssel, E. R., Florea, B. I., Codée, J. D. C., van der Marel, G. A., Aerts, J. M. F. G., and Overkleeft, H. S. (2014) Potent and selective activity-based probes for GH27 human retaining α-galactosidases. J. Am. Chem. Soc. 136, 11622–11625.
(27) Jiang, J., Kallemeijn, W. W., Wright, D. W., van den Nieuwendijk, A. M. C. H., Rohde, V. C., Folch, E. C., van den Elst, H., Florea, B. I., Scheij, S., Donker-Koopman, W. E., Verhoek, M., Li, N., Schürmann, M., Mink, D., Boot, R. G., Codée, J. D. C., van der Marel, G. A., Davies, G. J., Aerts, J. M. F. G., and Overkleeft, H. S. (2015) In vitro and in vivo comparative and competitive activity-based protein profiling of GH29 α- L -fucosidases. Chem. Sci. 6, 2782–2789.
17
2
A Fluorescence Polarization Activity-Based Protein Profiling Assay in the
Discovery of Potent, Selective Inhibitors for Human Non-Lysosomal
Glucosylceramidase
2.1 Introduction
Human non-lysosomal glucosylceramidase (GBA2) hydrolyses glucosylceramide (GlcCer) into glucose and ceramide in the cytosol of human cells.1-3 GBA2 is located
at the cytosolic leaflet of the endoplasmic reticulum (ER), Golgi apparatus and endosomes. The substrate, GlcCer, is synthesized at the cytosolic leaflet of the Golgi apparatus prior to translocation to the lumen of the organelle for elongation to complex glycosphingolipids.2,4 GBA2 activity is complementary to lysosomal acid glucosylceramidase (GBA1), which processes GlcCer in lysosomes.5 Genetic mutations in GBA1 are at the basis of the lysosomal storage disorder called Gaucher disease (GD). GD macrophages are rich in GlcCer levels and these lipid-laden macrophages are termed Gaucher cells.5 Other tissues in GD patients, however, appear unaffected in GlcCer levels. It is conceivable that GBA2 compensates for reduced GBA1 activity in these tissues.5
18
detrimental effects.11 One method that was used for the reduction of
neuropathology in NPC mice was via pharmacological inhibition of GBA2 using N-alkyl-deoxynojirimycin (N-alkyl-DNM) derivatives, such as N-butyl-DNM or N-(5)-adamantane-1-yl-methoxy-pentyl-DNM (respectively Zavesca (1) and AMP-DNM (2), Figure 1).12 These N-alkyl-DNM derivatives are reported as potent GBA2 inhibitors, with AMP-DNM (2) as the most potent of the two (Ki = 3 nM). However,
most potent GBA2 inhibitors have significant off-target activity, not only towards GBA1, but also the enzyme responsible for the biosynthesis of GlcCer: glucosylceramide synthase (GCS). Generally, the activity on GBA1 is significantly reduced when changing the configuration of the polyhydroxylated piperidine from
D-gluco- into L-ido-DNM. For example, the potency of L-ido-biphenyl-DNM (3) on GBA1 is significantly reduced. Unfortunately, in the case of 3, the potency on GCS is also significantly increased, showing that there is no simple rule or trend to apply on the design of DNM analogues and realize GBA2 selectivity.
Figure 1. Chemical structures of reported DNM analogues: Zavesca (1), AMP-DNM (2) and
L-ido-biphenyl-DNM (3).
19 2.2 Results and Discussion
GBA2 is part of the glycoside hydrolase family 116 (GH116) according to the Carbohydrate Active Enzyme (CAZY) database.14 The enzyme makes use of a Koshland two-step double displacement mechanism to process GlcCer, as discussed in the general introduction. Cyclophellitol-based probes15 have been designed taking into account the covalent intermediate that is formed in the first step of substrate processing: the glycosylation step. As depicted in Figure 2, cyclophellitol aziridines, with the aziridine nitrogen bearing a reporter group (a fluorophore, or biotin), are able to potently and irreversibly inhibit GBA2. Here, glutamic acid (Glu) 527 acts as the catalytic nucleophile and aspartic acid (Asp) 677 as the catalytic acid/base.16 However, these cyclophellitol aziridines are also able to target GBA1 and GBA3, as they are in the class of broad-spectrum retaining -glucosidases.15 Probes grafted with a fluorophore can be used in gel-based ABPP, where both selectivity and potency of inhibitors per individual target can be assessed.17 As discussed in the general introduction (Chapter 1), fluorescent probes can also be used to monitor binding events in time using fluorescence polarization (FluoPol) as readout. This chapter describes how specific targeting of GBA2 is realized and the resulting FluoPol assay on this enzyme is used to identify GBA2 specific inhibitors.
Figure 2. Mechanism of irreversible reaction of a cyclophellitol-based aziridine with GBA2.
2.2.1 Synthesis of a FluoPol compatible ABP
20
tritylation of the primary alcohol, benzylation of the secondary alcohols and subsequent detritylation. The primary alcohol in the resulting product 5 was converted to an iodide via an Appel reaction to give compound 6. Zinc-mediated reductive fragmentation of 6 afforded aldehyde 7. Subsequently indium-mediated allyl addition using bromo-crotonate in the presence of lanthanide triflate yielded diastereomerically pure 8. Cyclohexene 9 was obtained via ring-closing metathesis on diene 8 using Grubbs 2nd generation catalyst. The ester functionality in 9 was then reduced using DIBAL-H, affording diol 10. Sodium borohydride and water were added in order to avoid aldehyde contamination. The primary alcohol in 10 was selectively converted into the corresponding trichloroacetimidate, after which iodocyclisation led to the formation of iodide 11. Acid-mediated hydrolysis of the imidate in 11 followed by exposure to alkaline conditions (sodium bicarbonate) led to nucleophilic displacement of the iodide in a stereospecific fashion, giving benzyl-protected aziridine 12. Debenzylation using Birch conditions (lithium in ammonia) afforded aziridine 13. This aziridine was alkylated using iodo-pentyne and sodium bicarbonate and the fluorophore, 5’-TAMRA-PEG3-azide 34, was subsequently
21
Scheme 1: Synthetic route towards β-glucosidase aziridine probe (14)
Reagents and conditionsa) AcCl, MeOH, 98% (b) i) TrtCl, DMAP, Et3N, DMF ii) BnBr, NaH, TBAI, DMF,
4°C iii) p-TsOH, DCM/MeOH (9:1), 38% over three steps (c) I2, PPh3, imidazole, THF, reflux, 91% (d) Zn,
THF/H2O (9:1), sonication, 40°C, 76% (e) ethyl 4-bromocrotonate, In, La(OTf)3, H2O, 50% (f) Grubbs 2nd
generation, DCM, 91% (g) i) DIBAL-H, -20°C to ambient temperature ii) NaBH4, -20°C to ambient
temperature, 90% (h) i) CCl3CN, DBU, DCM, 0°C ii) I2, NaHCO3, H2O, 65% (i) i) AcOH/dioxane/H2O
(8:1:1) ii) NaHCO3, MeOH, 90% (j) Li, NH3(liq.), -60°C k) i) 5-iodopentyne, NaHCO3, DMF ii) 34, sodium
ascorbate, CuSO4, H2O, 6% over last three steps.
2.2.2 Compound 14 is a broad spectrum retaining -glucosidase probe
The synthesized probe was evaluated for its labelling capability by comparing 14 with the labelling pattern on tissues of an established broad-spectrum probe, ABP 1515, which contains Bodipy-FL as fluorophore. Lysates from mouse brain were treated with 1 µM probe at pH 5.0 for 30 minutes at 37°C. After incubation the proteins were denatured using Laemmli Buffer, containing an excess of sodium dodecylsulfate (SDS), and the resulting mixture was brought to 100°C for a short period of time. The protein content was then resolved by gel-electrophoresis and, as shown in Figure 3, fluorescence scanning of the wet gel slabs (ex = 488 and em
= 520 for Bodipy green; ex = 530 nm and em = 580 nm for TAMRA), revealed bands
22
respectively GBA2, GBA1 and GBA3. Labeling with ABPs 14 and 15 was performed as well on lysates from HEK293 cells overexpressing GBA2. Cell extracts were incubated with 14 at pH 7, and addition of the ABP was preceded by exposure of the protein mixture to Zavesca (1), AMP-DNM (2) or no inhibitor. Fluorescence scanning of the wet gel slab revealed a major band corresponding to the molecular weight of GBA2 for those samples treated with either 14 or 15 and in which no competitive inhibitor was used (Figure 3B, lanes 1, 4). Importantly, little to no endogenous GBA1 and GBA3 are seen, even though these enzymes are obviously present as well (quantification of the signals reveals that over 85% of the labeled protein correspond to GBA2). Labeling could be prevented by pretreatment with either of the two known competitive GBA2 inhibitors 1 (lanes 2, 4) or 2 (lanes 3, 6) at 100 and 10 µM final concentration, respectively. This experiment shows that relative levels of active GBA2 in the overexpressing cells are such that the preparation can be used to screen for GBA2 inhibitors, and as well that GBA2 labeling is abolished when competitive inhibitors are present.
Figure 3. Resulting image of gel-ABPP experiments on lysates of mouse brains (A) and HEK293 cells
23 2.2.3 Optimization of FluoPol-ABPP with 14
Initial experiments aimed to establish whether ABP 14 can be used in a FluoPol-ABPP assay format to screen for GBA2 inhibitors were carried out in 96 well plates, with 0.5 mg/mL protein and a final volume (Vfinal) of 75 L. Fluorescence intensities
were measured at ex = 530 nm and em = 580 nm. The pH optimum of GBA2 is
reported to be at pH 5.8.2 Assessment in a range from pH 4 to pH 8 (the range at which GBA2 possesses measurable activity), revealed that maximal return of polarized light occurs at pH 7 (Figure 4A). The optimal final probe concentration in the assay was found to be 25-50 nM (Figure 4B). Optimal probe concentration depends on the amount of active GBA2 present in the samples. In this setting a suitable window in FluoPol-signal ( ≈ 180 mPol) was achieved between positive and negative controls, respectively samples containing an excess of AMP-DNM 2 and DMSO only, representing samples in which labelling of GBA2 is blocked completely versus samples representing 100% labelling. The IC50 values of a set of
24
Figure 4. Optimization and validation of FluoPol-ABPP for GBA2. (A) Effect of pH. (B) Effect of probe
concentration. (C) Structures of additional established inhibitors. (D) Inhibition by established inhibitors (1, 2, 16, 17 and 18). Error bars represent standard error of the mean.
2.2.4 Screen of an iminosugar based library on GBA2
The FluoPol assay was further miniaturized into a 384-well plate format (Vfinal = 15
25
Figure 5. (A) FluoPol-ABPP screen of an iminosugar library (358 entries) using to identify potential
GBA2 inhibitors. Competition was performed at 100 nM. (B) Screen where the compounds of the library are categorized based on its sugar configuration (gluco, ido, galacto and others are classified as alternative configurations), deoxygenated variants and pyrrolizidines. (C) Chemical structures of the two identified GBA2 inhibitors, 19 and 20.
26
GCS are shown in Table 1. Neopentyloxypentyl-deoxynojirimycin 19 and its L-ido-congener 20, depicted in Figure 5C, were identified as nanomolar potent GBA2 inhibitors amongst known GBA2 inhibitors. As shown in Table 1, compounds 19 and 20 are also relatively selective GBA2 inhibitors.
Table 1: Potencies (in µM) and selectivity ratios of D-gluco- and L-ido iminosugars on GCS, GBA1 and
28 99 20 0.34 0.01 2000 34 336 0.2 1.5 0.003 67 500 102 1.9 100 0.0058 328 17241 339 0.05 1.25 0.003 17 417 103 4 1.5 0.007 571 214 341 0.075 0.3 0.002 38 150 104 2 50 0.01 200 5000 343 0.05 0.4 0.002 25 200 109 0.43 11.97 0.0015 287 7980 345 0.06 300 0.003 20 100000 111 0.39 0.84 0.001 390 840 347 0.07 400 0.003 23 133333 113 0.2 0.1 0.002 100 50 348 0.33 2.55 0.00004 7500 57955 118 0.025 0.2 0.00012 208 1667 349 0.06 0.3 0.001 60 300 120 0.025 0.1 0.00008 313 1250 350 1.25 0.13 0.01 125 13 122 0.08 10 0.002 40 5000 351 0.025 15 0.003 8 5000 127 0.5 1.5 0.004 125 375 352 0.07 20 0.002 35 10000 129 0.07 60 0.006 12 10000 353 0.008 20 0.003 3 6667 131 35.2 0.5 0.005 7040 100 354 0.03 10 0.001 30 10000 133 14.35 12 0.008 1794 1500 355 0.04 10 0.002 20 5000 139 0.75 0.3 0.001 750 300 356 0.021 100 0.00025 84 401606 141 0.3 0.1 0.002 150 50 357 0.015 12.5 0.002 8 6250 158 50 0.07 0.034 1471 2 358 0.02 500 0.002 10 250000
29
Figure 6. Some examples of the 2nd generation library; 21-25, the design of which was based on inhibitors 19 and 20.22
The isobutyl derivatives 21 and 22 appeared to be more potent GBA2 inhibitors compared to inhibitors 19 and 20 and are also less potent inhibitors of GCS. L-ido configured compounds 20 and 22 are over 100,000-fold selective for GBA2 over GBA1. This selectivity window is larger than that of previously reported GBA2 inhibitors, which are at most a 10,000-fold selective with respect to GBA1.23 Compounds 19, 23 and 24 are about 1000-fold more active on GBA2 compared to their GCS inhibitory activity, while previously reported compounds are at best 150-fold more selective.24 Substitution of the neopentyl moiety for tetrahydrofuranylmethyl groups, as in 24 and 25, had a detrimental effect on inhibitory potency towards GBA1, GBA2 and GCS, and no significant effect on selectivity. In conclusion, neopentyl- and isobutyl derivatives 19-22 are showing remarkable GBA2 selectivity, allowing perhaps specific targeting of GBA2 in situ (in living cells) and even in vivo.
Table 2: Potencies & selectivity ratios of deoxynojirimycin derivatives on GCS, GBA1, GBA2 and GBA3
GCSa GBA1b GBA2b GBA3b GCS/GBA2c GBA1/GBA2c GBA3/GBA2c
1 50 675 0.326 >1000 153 2071 >3067 2 0.15 0.28 0.0035 27.6 43 80 7886 3 0.008 25.3 0.0027 27.8 3 9370 10296 8 0.15 12.4 0.0083 66.1 18 1494 7964 19 5.05 3.8 0.0051 27.6 990 745 5412 20 2.77 120 0.0077 119 360 15584 15455 21 3.87 10.8 0.0075 128 516 1440 17067 22 0.729 1223 0.0095 177 77 128737 18632 23 6 7.3 0.0059 116 1017 1237 19661 24 53 92 0.0536 180 989 1716 3358 25 >50 122 0.1771 145 >282 689 819 a
Inhibition value for in situ assay is given as IC50 (µM) b Inhibition value for in vitro assay is given as K
30
2.2.5 In situ evaluation of neopentyl- & isobutyl-DNMs
A competitive ABPP experiment on live cells was performed in order to evaluate effective GBA2 inhibition in situ. To this end, cells overexpressing either GBA2 or GBA3, both of which are containing endogenous GBA1, were treated with compounds 19 - 22 at various final concentrations for 1 hour. After that the cells were lysed via snap-freeze and then treated with ABP 14 in order to determine the residual glucosidase activity. As can be seen from the images of the SDS-PAGE gels (Figure 7), all compounds are cell permeable and selectively block GBA2 over GBA1 and GBA3 at the concentrations tested.
Figure 7. Competitive ABPP experiment of HEK293T cells overexpressing GBA2 or GBA3 (and
expressing endogenous GBA1) treated with compounds 19 - 22 at various final concentrations prior to cell lysis and ABPP profiling of remaining enzyme activity.
2.3 Conclusion
32
2.4 Experimental section
Chemicals, materials and methods
All solvents and reagents were obtained commercially and used as received unless stated otherwise. Dichloromethane (DCM), dimethylformamide (DMF), tetrahydrofuran (THF) and methanol (MeOH) were dried over molecular sieves (4Å/3Å) for at least 12 hours before use. Moisture sensitive reactions were performed under argon atmosphere and carried out in oven-dried glassware. Reactions were monitored by TLC analysis using sheets with pre-coated silica with detection by UV-absorption (254 nm) wherever applicable and by spraying with 20% H2SO4 in MeOH, an aqueous solution containing KMnO4 (5 g/L) and K2CO3 (95
g/L) or a solution of ninhydrin (6 g/L) in AcOH:MeOH (1:9, v/v) followed by charring at ≈ 200℃. Flash column chromatography was performed on silica gel (40-63 µm). For LC/MS analysis a HPLC-system (detection simultaneously at 213 nm, 254 nm and evaporative light detection) equipped with an analytical C-18 column (4.6 mmD × 250 mmL, 5 µ particle size) in combination with buffers A: H2O, B:
acetonitrile, C: 1.0% aqueous trifluoroacetic acid and coupled with an electrospray interface (ESI) was used. For RP-HPLC purifications, an automated HPLC system equipped with a semi-preperative S2 C18 column (5 μm C18, 10Å, 150 × 21.2 mm) was used. The applied buffers were A: H2O + trifluoroacetic acid (1%) and B: MeCN. HPLC-MS purification was performed on an Agilent Technologies 1200 series automated HPLC system with a Quadropole MS 6130, equipped with a semi-preparative Gemini C18 column (Phenomex, 250 × 10, 5 μm) using buffers A: H2O +
K2CO3 (1%) and B: MeCN. Compounds are characterized by 1H NMR-, 13C NMR-,
33 Scheme 2: Synthesis of 5’-TAMRA-PEG3-azide 34
Reagents and conditions: (a) MsCl, Et3N, THF, 45% (b) NaN3, DMF under reflux, 95% (c) PPh3, 5% HCl
(aq), 77% (d) cat. H2SO4, AcOH under reflux (e) BOP.PF6, DIPEA, DMSO, 2% over two steps
((oxybis(ethane-2,1-diyl))bis(oxy))bis(ethane-2,1-diyl) dimethanesulfonate (27) A mixture of triethylamine (11.2 g, 110 mmol, 2.2 eq.) in THF (10 mL) was added dropwisely at 0°C into a mixture containing tetra-ethylene glycol (26, 9.7 g, 50 mmol) and mesylchloride (12.6 g, 110 mmol, 2.2 eq.) dissolved in dry THF (50 mL). After 30 minutes the cooling bath was removed and the reaction was stirred for 4 more hours at ambient temperature. THF was removed under reduced pressure. Subsequently a mixture H2O (100 mL), aqueous
HCl (100 mL, 1 M) and DCM (200 mL) was poured into the residue. The organic layer was washed with saturated bicarbonate (3 × 100 mL), dried with MgSO4 and
34
afforded 27 (7.943 g, 22.62 mmol, 45%) as brown oil. 1H NMR (300 MHz, CDCl 3) δ
4.27 - 4.14 (m, 4H), 3.66 - 3.57 (m, 4H), 3.50 (s, 8H), 2.93 (s, 6H). 13C NMR (75 MHz, CDCl3) δ 70.1, 69.3, 68.5, 37.1. HRMS: found 351.0773 [M+H]+, calculated for
[C10H22O9S2+H]+ 351.0778.
1-azido-2-(2-(2-(2-azidoethoxy)ethoxy)ethoxy)ethane (28)
A mixture of 27 (7.1 g, 20 mmol) with sodium azide (5.3 g, 82 mmol, 4 eq.) in absolute ethanol (40 mL) and DMF (10 mL) was refluxed overnight. This mixture was poured into a mixture of H2O/DCM (200 mL, 1:1, v/v). The organic layer was washed subsequently with H2O
(3 × 100 mL) and brine (3 × 100 mL). The organic layer was dried over MgSO4 and
filtered. Yellow oil (4.7 g, 19 mmol, 95%) was obtained after removing the solvents under reduced pressure. RF = 0.85 (MeOH:DCM, 1:9, v/v). 1H NMR (300 MHz, CDCl3)
δ 3.62 - 3.54 (m, 12H), 3.36 - 3.23 (m, 4H). 13C NMR (75 MHz, CDCl3) δ 70.6, 70.0,
50.6. HRMS: found 245.1358 [M+H]+, calculated for [C8H16O3N6+H]+ 245.1357.
1-azido-2-(2-(2-(2-azidoethoxy)ethoxy)ethoxy)ethane (29)
Triphenylphosphine(0.8 g, 3 mmol, 0.9 eq.) dissolved in ether (15 mL) was added into a solution of 28 (0.8 g, 3.3 mmol) in 5% aqueous HCl (10 mL). Addition was performed in 30 minutes at room temperature and the reaction was stirred for an additional 2.5 hours. Phases were separated using a separation funnel and the aqueous layer was washed using DCM (3 × 25 mL). The aqueous layer was adjusted to pH 10 using KOH pellets. Product was extracted with DCM (4 × 50 mL). Combined organic layers were dried over MgSO4 and filtered. After removal of the organic solvents under reduced
pressure yellow oil (4.7 g, 19 mmol, 77%) was afforded. RF = 0.2 (MeOH:DCM, 1:9,
v/v). 1H NMR (300 MHz, CDCl3) δ 3.73 - 3.49 (m, 10H), 3.45 (t, J = 5.2 Hz, 2H), 3.36 -
3.28 (m, 2H), 2.80 (t, J = 5.1 Hz, 2H), 1.44 (s, 2H).13C NMR (75 MHz, D2O) δ 73.5,
35
5-carboxy-2-(6-(dimethylamino)-3-(dimethyliminio)-3H-xanthen-9-yl)benzoate(32&33)
Dimethylaminophenol (30, 6.9 g, 50 mmol) and trimellitic anhydride (31, 4.8 g, 25 mmol) were dissolved in AcOH (400 mL). After adding a catalytic amounts of concentrated H2SO4 (ca. 0.5 mL) the mixture was refluxed
overnight. Reaction mixture was concentrated under reduced pressure and pre-purified over column chromatography (DCM 50% MeOH/DCM) to isolate 2.7 g of a mixture containing desired of regio-isomers.
2-(6-(dimethylamino)-3-(dimethyliminio)-3H-xanthen-9-yl)-5-((2-(2-(2-(2-(4-(3-((1R,2S,3S,4R,5R,6R)-2,3,4-trihydroxy-5-(hydroxymethyl)-7-azabicyclo[4.1.0]
heptan-7-yl)propyl)-1H-1,2,3-triazol-1-yl)ethoxy)-ethoxy)ethoxy)-ethyl)carbamoyl)benzoate (34)
BOP.PF6 (74 mg, 0.168 mmol) and DIPEA (54 µL, 0.31 mmol) were added into a mixture containing isomers 32 and 33 (60 mg) and linker 29 (30 mg, 0.14 mmol) dissolved in DMSO (2 mL). The reaction was stirred for 24 hours at ambient temperature. Desired stereoisomeric product 34 (7.41 mg, 11.7 µmol, 2% estimated yield over two steps) was isolated using HPLC purification. 1H NMR (600 MHz, MeOD) δ 8.79 (d, J = 1.7 Hz, 1H), 8.31 - 8.25 (m, 1H), 7.54 (d, J = 7.9 Hz, 1H), 7.15 (d, J = 9.5 Hz, 2H), 7.07 (dd, J = 9.5, 2.5 Hz, 2H), 7.00 (d, J = 2.4 Hz, 2H), 3.77 - 3.60 (m, 14H), 3.35 - 3.33 (m, 2H), 3.32 (s, 12H). 13C NMR (150 MHz, MeOD) δ 168.6, 167.7, 161.0, 159.4, 159.3, 138.4, 138.0, 133.2, 132.7, 132.2, 132.2, 131.7, 115.9, 115.0, 97.8, 72.0 - 70.8, 52.0, 49.9, 41.5, 41.2. HRMS: found 631.2877 [M+H]+, calculated for [C33H38O7N6+H]+ 631.2875. (2R,3R,4R)-methyl-2,3-D-xylofuranoside(4)
A catalytic amount of acetyl chloride (6.00 ml) was added to an ice cooled solution of D-xylose (30.0 g, 200 mmol) in MeOH (300 ml) and stirred at room temperature for 5h. The mixture was quenched with NaHCO3
36
4.99 (d, J = 4.5 Hz, 1H), 4.87 (s, 0.8H), 4.43 (d, J = 5.0 Hz, 0.8H), 4.36 – 4.30 (m, 1H), 4.25 – 4.17 (m, 2.5H), 4.10 (t, J = 4.7 Hz, 1H), 3.95 – 3.91 (m, 2H), 3.90 (dd, J = 4.8, 2.2 Hz, 1.5H), 3.49 (s, 3H), 3.44 (s, 2.4H), 3.16 – 2.12 (br.s., 5.6H). 13C NMR (101 MHz, CDCl3) δ 108.8, 101.7, 82.1, 80.9, 79.1, 78.0, 77.8 – 77.3, 62.2, 61.9, 55.8,
55.7. HRMS: found 165.0756 [M+H]+, calculated for [C6H12O5+H]+ 165.0757.
Methyl-2,3-Di-O-benzyl-5-iodo-D-xylofuranoside (5)
Methyl-xylofuranoside 4 (17.88 g, 108.9 mmol) was dissolved DMF (400 ml) followed by addition of trityl chloride (36.4 g, 131 mmol, 1.2 eq.), DMAP (0.67 g, 5.4 mmol, 0.05 eq.) and Et3N (30.0 ml, 218 mmol,
2 eq.). The reaction mixture was quenched with saturated NaHCO3 after 18h.
Extraction proceeded with Et2O (3x 250 ml) and the combined ether layers were
washed with saturated NaHCO3 and brine. The organic layer was dried with MgSO4,
filtered and concentrated in vacuo. This crude product was dissolved in DMF (100 ml) and slowly added to a solution of NaH (17.4 g 60% in mineral oil, 436 mmol, 4 eq.) in DMF (200 mL) at -20°C. Subsequently benzyl bromide (31.0 ml, 261 mmol, 2.5 eq.) and TBAI (0.402 g, 1.08 mmol, 0.016 eq.) were added to the cooled mixture, which was thereafter stirred at room temperature until full conversion was shown on TLC. The mixture was quenched with MeOH (20 ml) at 0°C and extracted with Et2O (3x 200 ml). Ether layers were washed with H2O (250 ml) and
brine (2x 100ml). The organic layer was dried with MgSO4, filtered and
concentrated in vacuo. The residue was pre-purified via flash column chromatography (pentane 10% EtOAc/pentane) resulting in yellow oil. Rf=0.25
(EtOAc/pentane, 1:9, v/v). The obtained oil was dissolved in MeOH/DCM (250 ml, 3:1, v/v) and the pH of this mixture was adjusted to 2-3 by addition of p-TsOH at 0°C, where an immediate change in color (from yellow into orange) was observed. The reaction mixture was neutralized using Et3N after full conversion of the starting
material was shown on TLC. Product was extracted using DCM (3x 150ml) and the combined DCM layers were washed with saturated NaHCO3 (100 ml) and brine (2x
100ml). The organic layer was dried with MgSO4 and filtered. The filtrate was
concentrated in vacuo and the desired product was purified by column chromatography (10% EtOAc/pentane 50% EtOAc/pentane) which afforded 5 as yellow oil (14.33 g, 41.61 mmol, 38%) and as an anomeric mixture (ratio :; 0.9:1). Rf=0.5 (EtOAc/pentane, 1:1, v/v). 1H NMR (400 MHz, CDCl3) δ 7.43 – 7.22 (m, 19H),
37 82.8, 82.2, 80.6, 76.3, 72.9 – 72.2, 62.24, 62.2, 55.6, 55.1 HRMS: found 345.1700 [M+H]+, calculated for [C20H24O5+H]+ 345.1730.
Methyl-2,3-Di-O-benzyl-5-iodo-D-xylofuranoside (6)
Compound 5 (13.8 g, 40.3 mmol) was dissolved in anhydrous THF (100 ml). Triphenylphosphine (16.05 g, 61.19 mmol, 1.5 eq.) and imidazole (5.72 g, 84.0 mmol, 2 eq.) were added to the resulting yellow mixture and heated to 75°C. A solution of iodine (15.91 g, 62.68 mmol, 1.5 eq.) in anhydrous THF (100 ml) was added to the mixture, which resulted in a dark brown mixture. This mixture was quenched after full conversion was shown on TLC with 10% potassium thiosulphate solution (100ml). The salt was filtered off and the filtrate was concentrated in vacuo. Purification was performed via flash column chromatography (10% EtOAc in pentane) and resulted in yellow oil (16.70 g, 36.75 mmol, 91%). Anomeric mixture (ratio :; 0.75:1). Rf=0.9 (EtOAc/pentane, 1:1, v/v). 1
H NMR (400 MHz, CDCl3) δ 7.49 – 7.01 (m, 17.5H), 4.90 (s, 1H), 4.81 (d, J = 4.1 Hz,
0.75H), 4.61 – 4.32 (m, 8.75H), 4.21 – 4.10 (m, 0.75H), 4.03 – 3.93 (m, 2.75H), 3.41 – 3.25 (m, 7.8H), 3.16 (dd, J = 10.1, 7.5 Hz, 0.75H). 13C NMR (101 MHz, CDCl3) δ
137.2, 128.2 – 127.3, 107.9, 100.5, 86.3, 83.5, 81.8, 81.5, 81.4, 77.3, 72.5 – 71.7, 55.7, 55.1 , 4.7, 3.1. HRMS found 455.0706 [M+H]+, calculated for [C20H23O4I+H]+
455.0714.
(2R,3S)-2,3-bis(benzyloxy)pent-4-enal (7)
Zinc dust (25.1 g, 384 mmol, 19 eq.) was activated by stirring in 2.5M HCl solution (250 ml) for 20 min at ambient temperature. Subsequently the mixture was filtered and washed with H2O, MeOH and Et2O. The zinc was fully dried
under high vacuum at elevated temperature. Compound 6 (9.35 g, 20.6 mmol) was dissolved in THF/H2O (250 ml, 9:1, v/v) and sonicated for 1 h at 40°C under argon
flow, after which the activated zinc was added. After 3h of sonication the mixture was filtered over a pad of celite, which was rinsed with EtOAc. After removal of all solvents in vacuo, the mixture was diluted in Et2O (200 ml) and washed with H2O
(2x 100 ml). The organic layer was dried with MgSO4, filtered and concentrated in vacuo. Purification of the desired product was performed by column chromatography (pentane 25% Et2O/pentane) which resulted in pale yellow oil
(4.621 g, 15.59 mmol, 76%). Rf=0.3 (Et2O:pentane, 1:4, v/v). 1H NMR (400 MHz,
CDCl3) δ 9.66 (d, J = 1.5 Hz, 1H), 7.49 – 7.12 (m, 10H), 5.93 (ddd, J = 17.2, 10.5, 7.6
38
13C NMR (101 MHz, CDCl
3) δ 202.5, 137.6, 137.1, 133.9, 128.5-127.8, 119.9, 85.2,
79.9, 73.5, 70.7.
(2S,3R,4S,5S)-2,3-Bis(benzyloxy)-4-hydroxy-6-vinylhept-5-ethylenoate (8)
Compound 4 (4.30 g, 14.5 mmol) dissolved in degassed milli-Q H2O
(125 ml) was added La(OTf)3 (17.611 g, 30.047 mmol, 2.1 eq.),
indium (3.458 g, 30.12 mmol, 2.1 eq.) and ethyl-4-bromocrotonate (8.00 ml 75% pure, 43.6 mmol, 3 eq.). After 4 days, the mixture was filtered over a pad of celite, which was rinsed with Et2O. The collected mixture was
diluted with Et2O (250 ml) and washed with H2O (3x 50 ml). The organic layer was
dried with MgSO4, filtered and concentrated in vacuo. Purification and separation
of the obtained diastereomers was performed via flash column chromatography (pentane 15% EtOAc/pentane) and resulted in clear oil (3.227 g, 7.861 mmol, 54%). Rf=0.4 (EtOAc:pentane, 9:1, v/v). 1H NMR (400 MHz, CDCl3) δ 7.42 – 7.30 (m, 10H), 5.92 – 5.69 (m, 2H), 5.51 – 5.37 (m, 2H), 5.21 (d, J = 10.0 Hz, 1H), 5.06 (d, J = 11.3 Hz, 2H), 4.65 (t, J = 12.7 Hz, 2H), 4.45 (d, J = 11.5 Hz, 1H), 4.25 (d, J = 7.7 Hz, 1H), 4.15 (q, 6.7 Hz, 2H), 4.01 (s, 1H), 3.59 (d, J = 7.3 Hz, 1H), 3.34 (t, J = 9.1 Hz, 1H), 2.77 (d, J = 9.2 Hz, 1H), 1.27 (t, J = 6.9 Hz, 3H). 13C NMR (101 MHz, CDCl3) δ 172.6, 138.6, 138.5, 135.1, 133.0, 128.5-127.7, 120.1, 120.0, 83.1, 79.5, 74.7, 72.3, 70.9, 60.9, 55.3, 14.3. HRMS found 411.2162 [M+H]+, calculated for [C25H30O5+H]+
411.2166.
(2S,3R,4S,5S)-ethyl-4,5-bis(benzyloxy)-6-hydroxycyclohex-2-enecarboxylate (9) Compound 8 (2.907 g, 7.08 mmol) was dissolved in DCM (408 ml, 0.025 M) and Grubbs2nd catalyst (0.157 g, 0.185 mmol, 2.61 mol%) was added. The mixture was stirred at room temperature for 3 days in the dark after which it was concentrated in vacuo. Purification by column chromatography (10% EtOAc/pentane 13% EtOAc/pentane) resulted in brown oil (2.46 g, 6.43 mmol, 91%). Rf=0.52 (EtOAc:pentane, 1:3, v/v). 1H NMR (400 MHz,
CDCl3) δ 7.38 – 7.27 (m, 10H), 5.80 (dt, J = 10.2, 2.1 Hz, 1H), 5.67 (dt, J = 10.2, 2.1
Hz, 1H), 4.96 (d, J = 11.3 Hz, 1H), 4.79 (d, J = 11.4 Hz, 1H), 4.74 – 4.62 (m, 2H), 4.25 – 4.08 (m, 4H), 3.65 (dd, J = 9.8, 7.5 Hz, 1H), 3.25 (ddd, J = 8.7, 5.7, 2.9 Hz, 1H), 2.96 (s, 1H), 1.25 (s, J = 7.2 Hz, 5H). 13C NMR (101 MHz, CDCl3) δ 172.1, 138.5, 138.1,
39 (2S,3R,4S,5S)-2,3-Bis(benzyloxy)-4-4hydroxy-5-hydroxymethyl-cyclohex-6-ene (10) Dried compound 9 (2.04 g, 5.33 mmol) dissolved in anhydrous THF (300 ml). The mixture was cooled down to -20°C and DIBAL-H (25% in toluene) (35 ml, 53 mmol, 10 eq.) was added. The mixture was stirred 0.5h at 0°C followed by 24h at ambient temperature. Subsequently, the mixture was quenched with EtOAc (10 ml) at 0°C, followed by addition of H2O (25
ml) and NaBH4 (1.41 g, 37.3 mmol, 7 eq.). After overnight stirring at room
temperature, the reaction mixture was concentrated in vacuo. The mixture was diluted with EtOAc (50 ml) and washed with 1 M HCl (10 ml). The organic layer was dried with MgSO4, filtered and concentrated in vacuo. Purification by column
chromatography (10% EtOAc/pentane 45% EtOAc/pentane) resulted in white crystalline product (1.53 g, 4.49 mmol, 84%). Rf=0.5 (EtOAc:pentane, 1:1, v/v). 1H
NMR (400 MHz, CDCl3) δ 7.36 – 7.29 (m, 10H), 5.76 (dt, J = 10.2, 2.0 Hz, 1H), 5.50
(dt, J = 10.2, 2.0 Hz, 1H), 5.01 (d, J = 11.3 Hz, 1H), 4.77 – 4.58 (m, 3H), 4.23 – 4.15 (m, 1H), 3.81 – 3.57 (m, 4H), 2.87 (s, 2H), 2.47 (s, 1H). 13C NMR (101 MHz, CDCl3) δ
138.6, 138.2, 128.8-127.9, 127.6, 127.6, 83.5, 80.4, 75.0, 72.7, 71.6, 65.3, 45.4. HRMS found: 341.1750 [M+H]+, calculated for [C21H24O4+H]+ 341.1747.
(4aR,5R,6S,7R,8S,8aR)-6,7-bis(benzyloxy)-8-iodo-2-(trichloromethyl)-4a,5,6,7,8,8a-hexahydro-4H-benzo[d][1,3]oxazin-5-ol (11)
Dry diol 10 (1.21 g, 3.55 mmol) dissolved in DCM(100 ml). The solution was cooled to 0°C., and trichloroacetonitrile (356 μl, 3.74 mmol) and 1,8-diazobicyclo[5.4.0]undec-7-ene (26 μl, 0.2 mmol) was added. After two hours stirring, TLC analysis revealed complete conversion of the starting material. To the solution was added H2O (15 ml), sodium
hydrogencarbonate (2.99 g, 35.6 mmol, 10 eq.) and iodine (2.73 g, 10.7 mmol, 3 eq.). The resulting mixture was stirred overnight, quenched with 20% potassium thiosulfate solution(12.5 ml) and diluted with DCM. The layers were separated, after which the organic layer was dried with MgSO4 and concentrated in vacuo.
Purification via flash column chromatography (pentane→10% EtOAc/pentane) afforded 11 (1.42 g, 2.33 mmol, 65%) as colorless oil. Rf=0.6 (EtOAc:pentane, 1:9,
v/v). 1H NMR (400 MHz, CDCl3) δ 7.49 – 7.31 (m, 10H), 5.04 (d, J = 11.2 Hz, 1H), 4.92
– 4.81 (m, 2H), 4.77 (d, J = 11.3 Hz, 1H), 4.68 (d, J = 11.2 Hz, 1H), 4.59 (d, J = 11.3 Hz, 1H), 4.33 (dd, J = 11.1, 2.9 Hz, 1H), 4.15 – 4.08 (m, 1H), 3.79 (t, J = 9.2 Hz, 1H), 3.48 (t, J = 10.0 Hz, 1H), 2.78 – 2.62 (m, 2H), 2.52 (s, 1H). 13C NMR (101 MHz, CDCl3) δ
40
(1R,2R,3R,4S,5S,6R)-4,5-bis(benzyloxy)-2-(hydroxymethyl)-7-azabicyclo[4.1.0] heptan-3-ol (12)
Imidate 11 (0.107 g, 0.175 mmol) was dissolved in a 7.0 ml solution of acetic acid, dioxane and water (8:1:1, v/v/v) and this mixture was stirred for three days at room temperature [11] = 0.025M. After LC/MS and TLC showed full conversion of the starting material the mixture solvents were evaporated under reduced pressure. Toluene (3 ml) was used in order to azeotropically evaporate the water. The concentrated residu was redissolved in methanol, to which sodium hydrogencarbonate (0.30 g, 3.5 mmol, 20 eq.) was added and stirred for one day at elevated temperature (ca. 50°C). After LC/MS analysis showed full conversion towards the product, the reaction mixture was filtered over a small path of celite. Purification via flash column chromatography (DCM10% MeOH/DCM) afforded compound 12 (56 mg, 0.16 mmol, 90%) as a white solid. Rf=0.375 (MeOH:DCM, 1:9, v/v). 1H NMR (400 MHz,
CDCl3) δ 7.40 – 7.27 (m, 10H), 4.96 (d, J = 11.3 Hz, 1H, 4.78 (d, J = 11.4 Hz, 1H), 4.66
(dd, J = 11.4, 3.0 Hz, 2H), 3.94 (ddd, J = 15.3, 10.7, 5.3 Hz, 2H), 3.75 (d, J = 8.1 Hz, 1H), 3.53 (t, J = 9.9 Hz, 1H), 3.37 (dd, J = 10.1, 8.1 Hz, 1H), 3.17 – 2.51 (br.s, 2H), 2.43 (dd, J = 6.0, 3.2 Hz, 1H), 2.27 (d, J = 6.1 Hz, 1H), 2.10 – 2.01 (m, 1H), 1.25 (s, 1H). 13C NMR (101 MHz, CDCl3) δ 138.6, 137.9, 128.7-127.8, 84.4, 81.4, 74.9, 72.4,
68.5, 64.8, 42.6, 33.1, 31.6. HRMS found: 356.1855 [M+H]+, calculated for [C21H24O4+H]+ 356.1856.
(1R,2S,3S,4R,5R,6R)-5-(hydroxymethyl)-7-azabicyclo[4.1.0]heptane-2,3,4-triol (13) Ammonia (10 ml) was condensed at −60°C. Lithium (50 mg) was added and the mixture was stirred until the lithium was completely dissolved. To this solution was added a solution of aziridine 12 (45 mg, 0.13 mmol) in dry THF (7 ml). The mixture was stirred for 2h at −60°C and subsequently quenched with MeOH/H2O (5.0 ml, 8:2, v/v) The solution
was allowed to get to room temperature while stirring so all ammonia could evolve. Next, the solution was concentrated in vacuo, redissolved in H2O (10 ml)
and neutralized with Amberlite-H+. Product bound to the resin was eluted with 1M NH4OH in MeOH (25 ml) solution. Solvents were evaporated under reduced
41 = 5.5 Hz, 1H), 1.89 (d, J = 8.7 Hz, 1H). 13C NMR (101 MHz, MeOD) δ 79.5, 74.4, 69.7, 64.0, 45.7, 36.6, 34.2. 2-(6-(dimethylamino)-3-(dimethyliminio)-3H-xanthen-9-yl)-5-((2-(2-(2-(2-(4-(3- ((1R,2S,3S,4R,5R,6R)-2,3,4-trihydroxy-5-(hydroxymethyl)-7-azabicyclo[4.1.0]hept-an-7-yl)propyl)-1H-1,2,3-triazol-1-yl)ethoxy)ethoxy)ethoxy)ethyl)carbamoyl) benzoate (ABP14) Sodium bicarbonate (4.8 mg, 57 μmol) and iodo-pentyne (2.9 mg, 15 μmol) were added to a mixture of crude aziridine 13 (5 mg) dissolved in dry DMF (0.5 mL). The reaction was stirred for 24h at ca. 60°C. The volatiles were removed under reduced pressure, the residue was dissolved in water (10 mL) and washed with DCM (3 × 10 mL) and EtOAc (3 × 10 mL). The aqueous layer was concentrated under reduced pressure. The compound present in the residue containing was redissolved in H2O (0.5 mL). Sodium ascorbate and CuSO4 (5 µL of 1
M solution) were added into the aqueous mixture. After the mixture turned yellow-green TAMRA 34 (1.25 mg, 1.98 µmol dissolved in 1.5 mL H2O) was added. The
reaction was completed after 3 days according to LC/MS analysis. Solvents were evaporated and product ABP 14 (1.457 mg, 1.670 µmol, 6% yield over three steps) was isolated as a purple solid after HPLC-purification. RF = 0.1 (MeOH:DCM, 1:3,
42
Cell culture
Human Embryonic Kidney 293T (HEK293T) cells (Sigma) were cultured in DMEM high glucose (Sigma) supplemented with 10% NBS and 100 units/mL penicillin/streptomycin (Gibco) at 37°C and 5% CO2. RAW 264.7 (American Type
culture collection) were cultured in RPMI (Sigma) supplemented with 10% FCS, 1 mM Glutamax and 100 units/mL penicillin/streptomycin (Gibco) at 37°C and 5% CO2.
Transient and stable overexpression of human GBA2 and GBA3
Primers used for transient GBA2 expressing HEK293T were designed based on NCBI reference sequence NM_020944.2. The full-length coding sequences were cloned into pcDNA3.1/Myc-His (Invitrogen, Life Technologies, Carlsbad, CA. Sub-confluent HEK293T cells were transfected with empty pcDNA3.1 or GBA2pcDNA3.1 (Plasmid:PEI ratio 1:3). Media was refreshed 24 hours later and cells collected 72 hours after transfection in PBS buffer. Cells were centrifuged at 1000 rpm for 10 minutes, after which the supernatant was removed. Cell pellets were snap-frozen with liquid nitrogen and stored at -80°C.
43 Preparation of lysates
Cell pellets were suspended in lysis buffer (20 mM Hepes, 2 mM DTT, 0.25 M sucrose, 1 mM MgCl2, 2.5 U/ml benzonase, pH 7.0). The homogenate was
incubated on ice for 30 minutes after lysis and homogenization using SilentCrusher (Heidolph). Ultracentrifugation was performed at 32.000 rpm for 30 minutes at 4°C. Supernatant fractions were aliquoted in appropriate volumes, after the total protein concentration was determined via a Bradford assay, using BSA (Sigma) for standards and BioRad Quickstart Bradford Reagents. Aliquots were snap-freezed with liquid nitrogen and stored at -80°C. For the lysates of mouse brain tissue the lysis was performed in a similar lysis buffer as described above. Tissue was homogenized using sonication.
Gel-Activity Based Protein Profiling experiments
Prepared lysate was diluted in assay buffer (20 mM Hepes, 2 mM DTT at pH 7.0) until appropriate final protein concentration was reached. Samples from HEK cells were incubated with 500 nM probe at 37°C for 1 hour (Vfinal = 20 µL). Lysates from
mouse tissue were treated with 1 µM of probe in McIlvaine buffer (pH 5.0). Protein content was denatured using Laemli Buffer (4x) at 100°C for 3 minutes. Reactions were resolved by 12.5% SDS-PAGE electrophoresis and wet slabs were scanned for fluorescence (Molecular Imager Gel Doc XR, Biorad). For the in situ competitive ABPP experiments stable GBA2 and GBA3 overexpressing HEK293T cells were incubated for 1 hour at 37°C with various concentrations of the selected inhibitors. Cells were washed 2 times with PBS and then harvested by scraping in 25 mM potassium phosphate buffer (pH 6.5) containing 0.1% Triton X-100 (v/v). Cells were lysed via snap-freezing in liquid nitrogen. Protein concentration was determined via BCA assay (Thermo Fischer). Labelling of the glucosidases was conducted using 500 nM ABP 6 for 30 minutes at 37°C on 20 µg protein. Protein content was denatured using Laemli Buffer (5x) at 100°C for 5 minutes. Reactions were resolved by 10% SDS-PAGE electrophoresis and wet slabs were scanned as described above.
Optimization of the FluoPol-ABPP assay
44
contained Bovine Gamma-Globulin (Sigma) and Chaps (Sigma), respectively 0.5 mg/ml and 1 mg/ml, were carried out in 96-wells plates (flat-black bottom, Greiner). FluoPol-signals were monitored on an Infinite M1000Pro (Tecan) using λex
= 530 nm and λem = 580 nm. Mock containing lysates were used as reference
samples, samples without probe as blanks to correct for background polarization and GBA2 containing lysates without inhibitor as controls. All samples were corrected for background polarization and the residual enzyme activity was calculated based on the polarization signal from the controls. Polarization signals were plotted against time or inhibitor concentration and processed in GraphPad Prism 6.0. IC50 values were calculated via non-linear regression using mentioned
software (N=2, n=3).
FluoPol-ABPP screen of the iminosugar library
The screen, using the optimized conditions as described above, was conducted in a 384-well black-bottom plate (Greiner) with reaction volumes set at 15 µL. Concentration of the iminosugars during the screen was 100 nM. The FluoPol signal was measured on a Clariostar (BMG Labtech). Resulting polarization signals were processed as described above. Residual enzyme activities were plotted against the corresponding iminosugar.
Analysis of selected compounds as inhibitors of enzymatic activity of GCS, GBA1, GBA2 and GBA3
All IC50 values were determined in triplicate and the inhibitors tested were
pre-incubated for 30 minutes at 37°C. Ki values were determined in duplicate using a
range from 0.05 till 5 mM of 4MU-β-D-glucopyranoside in appropriate buffer containing inhibitor. Incubation time and temperature was performed as described below, but without the pre-incubation step. Observed fluorescence was curve-fitted against inhibitor or substrate concentrations using GraphPad Prism 6.0 in order to obtain IC50 or Ki values.
GBA1: Pure recombinant human enzyme (Cerezyme from Genzyme) was used. Activity was measured with with 3.7 mM 4-methylumbeliferone (4MU)-β-D-glucopyranoside (Sigma) in 150 mM McIlvaine buffer pH 5.2 supplemented with 0.2% sodium taurocholate (w/v), 0.1% Triton X-100 (v/v), 0.1% bovine serum albumin (BSA) (w/v) for 30 min. The reaction was stopped with excess 1M NaOH-Glycine (pH 10.3), liberated 4MU fluorescence was measured with a fluorimeter LS55 (Perkin Elmer) using λEx 366 nm and λEm 445 nm.
45 conduritol β epoxide CBE from Sigma)31 were used. Activity was measured with 3.7
mM 4MU-β-D-glucopyranoside in 150 mM McIlvaine pH 5.8, 0.1% BSA (w/v) for 1 h. Reactions were stopped as described above for GBA1.
GBA3: For GBA3 measurements, cellular homogenates of a stable HEK293T over-expressing GBA3 cell line also pre-incubated for 30 min with an inhibitor of GBA1 (1 mM conduritol β epoxide CBE from Sigma) were used. Reactions were conducted in 100 mM hepes pH 7.0, 0.1% BSA (w/v) for 1 h.
GCS: IC50 values for GCS were determined in situ with
6-[N-methyl-N-(7-nitrobenz-2-oxa-1,3-diazol-4-ylaminododecanoyl]sphingosine (NBD-ceramide) as substrate. RAW 264.7 cells were grown to confluence in 6-well plates and pre-incubated for 1h with an inhibitor of GBA1 activity (300 µM CBE), followed by 1h incubation at 37°C with 1 µM C6-NBD-ceramide and in the presence of a range of inhibitor concentrations. The cells were washed 3x with PBS and harvested by scraping. After lipid extraction (described by Bligh and Dyer),30 the C6-NBD lipids were separated and detected by High Performance Liquid Chromatography (λEx 470 nm
and λEm 530 nm). IC50 values were determined in duplicate from the titration curves
of observed formed C6-NBD-glucosylceramide and data was curve-fitted via GraphPad Prism 6.0.
2.5 References
(1) Yildiz, Y., Matern, H., Thompson, B., Allegood, J. C., Warren, R. L., Ramirez, D. M. O., Hammer, R. E., Hamra, F. K., Matern, S., and Russell, D. W. (2006) Mutation of β-glucosidase 2 causes glycolipid storage disease and impaired male fertility. J.
Clin. Invest. 116, 2985–2994.
(2) van Weely, S., Brandsma, M., Strijland, A., Tager, J. M., and Aerts, J. M. F. G. (1993) Demonstration of the existence of a second, non-lysosomal glucocerebrosidase that is not deficient in Gaucher disease. Biochim. Biophys. Acta - Mol. Basis Dis.
1181, 55–62.
(3) Boot, R. G., Verhoek, M., Donker-Koopman, W., Strijland, A., van Marle, J., Overkleeft, H. S., Wennekes, T., and Aerts, J. M. F. G. (2007) Identification of the Non-lysosomal Glucosylceramidase as β-Glucosidase 2. J. Biol. Chem. 282, 1305– 1312.
(4) Körschen, H. G., Yildiz, Y., Raju, D. N., Schonauer, S., Bönigk, W., Jansen, V., Kremmer, E., Kaupp, U. B., and Wachten, D. (2013) The non-lysosomal β-glucosidase GBA2 is a non-integral membrane-associated protein at the endoplasmic reticulum (ER) and golgi. J. Biol. Chem. 288, 3381–3393.
46
Wisse, P., Boot, R. G., Willems, L. I., Overkleeft, H. S., and Aerts, J. M. (2014) Gaucher disease and Fabry disease: New markers and insights in pathophysiology for two distinct glycosphingolipidoses. Biochim. Biophys. Acta - Mol. Cell Biol. Lipids
1841, 811–825.
(6) Hammer, M. B., Eleuch-Fayache, G., Schottlaender, L. V., Nehdi, H., Gibbs, J. R., Arepalli, S. K., Chong, S. B., Hernandez, D. G., Sailer, A., Liu, G., Mistry, P. K., Cai, H., Shrader, G., Sassi, C., Bouhlal, Y., Houlden, H., Hentati, F., Amouri, R., and Singleton, A. B. (2013) Mutations in GBA2 cause autosomal-recessive cerebellar ataxia with spasticity. Am. J. Hum. Genet. 92, 245–251.
(7) Martin, E., Schüle, R., Smets, K., Rastetter, A., Boukhris, A., Loureiro, J. L., Gonzalez, M. A., Mundwiller, E., Deconinck, T., Wessner, M., Jornea, L., Oteyza, A. C., Durr, A., Martin, J.-J., Schöls, L., Mhiri, C., Lamari, F., Züchner, S., De Jonghe, P., Kabashi, E., Brice, A., and Stevanin, G. (2013) Loss of function of glucocerebrosidase GBA2 is responsible for motor neuron defects in hereditary spastic paraplegia. Am. J. Hum.
Genet. 92, 238–244.
(8) Sultana, S., Truong, N. Y., Vieira, D. B., Wigger, J. G. D., Forrester, A. M., Veinotte, C. J., Berman, J. N., and van der Spoel, A. C. (2016) Characterization of the zebrafish homolog of β-glucosidase 2: a target of the drug miglustat. Zebrafish 13, 177–187.
(9) Marques, A. R. A., Aten, J., Ottenhoff, R., van Roomen, C. P. A. A., Herrera Moro, D., Claessen, N., Vinueza Veloz, M. F., Zhou, K., Lin, Z., Mirzaian, M., Boot, R. G., De Zeeuw, C. I., Overkleeft, H. S., Yildiz, Y., and Aerts, J. M. F. G. (2015) Reducing GBA2 activity ameliorates neuropathology in Niemann-Pick type C mice. PLoS One (Dardis, A., Ed.) 10, e0135889.
(10) Mistry, P. K., Liu, J., Sun, L., Chuang, W.-L., Yuen, T., Yang, R., Lu, P., Zhang, K., Li, J., Keutzer, J., Stachnik, A., Mennone, A., Boyer, J. L., Jain, D., Brady, R. O., New, M. I., and Zaidi, M. (2014) Glucocerebrosidase 2 gene deletion rescues type 1 Gaucher disease. Proc. Natl. Acad. Sci. USA 111, 4934–4939.
(11) Aerts, J. M., Hollak, C., Boot, R., and Groener, A. (2003) Biochemistry of glycosphingolipid storage disorders: implications for therapeutic intervention.
Philos. Trans. R. Soc. London. Ser. B Biol. Sci. (Dwek, R. A., Butters, T. D., Platt, F.
M., and Cox, T. M., Eds.) 358, 905–914.
(12) Overkleeft, H. S., Renkema, G. H., Neele, J., Vianello, P., Hung, I. O., Strijland, A., van der Burg, A. M., Koomen, G.-J., Pandit, U. K., and Aerts, J. M. F. G. (1998) Generation of specific deoxynojirimycin-type inhibitors of the non-lysosomal glucosylceramidase. J. Biol. Chem. 273, 26522–26527.
(13) Wennekes, T., van den Berg, R. J. B. H. N., Boot, R. G., van der Marel, G. A., Overkleeft, H. S., and Aerts, J. M. F. G. (2009) Glycosphingolipids - nature, function, and pharmacological modulation. Angew. Chem. Int. Ed. 48, 8848–8869.
