Activity-based profiling of glycoconjugate processing enzymes Witte, M.D.

Activity-based profiling of glycoconjugate processing enzymes

Witte, M.D.

Citation

Witte, M. D. (2009, December 22). Activity-based profiling of glycoconjugate processing enzymes. Retrieved from https://hdl.handle.net/1887/14551

Version: Corrected Publisher’s Version

License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden

Downloaded from: https://hdl.handle.net/1887/14551

Note: To cite this publication please use the final published version (if

applicable).

Introduction

Exo-β-glucosidases remove β-linked glucose from the non-reducing end of glycoconjugates.

These enzymes have been found in a great variety of species and their biological roles are just as diverse. For example, β-glucosidases play a role in the acute defense mechanism in plants. Upon disruption of the plant tissues, glucosidases release toxic compounds by hydrolysis of glycosidic bond of the non-toxic precursor.

In mammals, acidic β-glucosidase (GBA-1), also known as glucocerebrosidase, is a key enzyme in the degradation of glycosphigolipids and malfunctioning of this enzyme is at the basis of the storage disease called Gaucher disease.

Cycophellitol based

glucosidase probes

Synthesis and biological

evaluation

Over the past decade, activity-based protein profiling has been used to study enzymes of many classes and a great variety of activity-based probes (ABPs) has been synthesized for this purpose.

It was reasoned that this strategy could also be useful for the profiling of exo- β-glucosidases. The development of glycosidase ABPs, the enzyme class to which the exo-β- glucosidases belong, is challenging and only a few glycosidase ABPs have been reported to date.

Quinone methide probes were the first ABPs that appeared in literature for glycosidases (Figure 1A). Although purified enzymes have been labeled using these probes, severe cross-labeling was observed in complex samples.

Soon after the appearance of the quinone methide probes, ABPs based on fluorinated glycosides were reported (Figure 1B).

Retaining glycosidases react with these probes similar to how they would react with the natural substrate. Protonation of the glycosidic bond by the acid catalyst is followed by nucleophilic attack of the other carboxylic acid present in the active site forming a glycosyl- enzyme adduct. The fluorine substituent slows down both the formation and hydrolysis of the covalent glycosyl-enzyme adduct. The activated leaving group increases the reaction rate of the first step leading to accumulation of the inhibitor-enzyme adduct (Figure 1C).

Vocadlo and Bertozzi were the first to report such an ABP, namely β-galactoside 1. Both galactosidases and glucosidases of several families have been labeled successfully with this probe. Although relatively stable, the adduct slowly hydrolyzes due to the endo-cyclic oxygen present in 1. Lifetimes ranging from seconds to months have been reported for these complexes and cleavage rates increase at pH > 7, limiting the use of these probes in gel-electrophoresis protocols.

Figure 1. General structure of (A) quinone methide probes and (B) fluorinated glycoside probes. (C) Mechanism of inactivation of glycosidases by fluorinated glycosides.

This chapter describes the synthesis of novel glucosidases ABPs 3-5 based on cyclophellitol

(2) (Figure 2A). This glucopyranose-configured glycosyl epoxide selectively inhibits exo-β-

glucosidases and was originally isolated from Phellinus sp.

The oxirane in 2 is protonated

in the active site by the general acid/base catalyst (Figure 2B). Subsequent ring-opening by

the nucleophilic residue in the active site results in covalent modification of the enzyme.

The formed adduct is highly stable,

making cyclophellitol 2 a suitable lead for glucosidase ABPs. It was reasoned that cyclophellitol 2 could be converted into two-step probes or direct probes by the incorporation of a ligation handle or a fluorescent reporter group respectively. The fluorophore of the direct probes allows straightforward in gel visualization of the labeled proteins, but this the bulky reporter group may interfere on binding of the probe. In two-step probes, the bulky fluorophore is substituted by a small ligation handle.

After inhibition and subsequent denaturation of the enzyme, the formed adduct is visualized by conjugating the ligation handle to a reporter group using, for instance, a Staudinger-Bertozzi ligation

or Cu(I)-catalyzed click reaction

. In this fashion, the steric hinderance is minimized. Stringent selectivity of the conjugation step is required to obtain optimal results with these probes. To evaluate which type of probe is most-suited for the labeling of glucosidases, both two-step probe KY170 (3) and direct probes MDW933 (4) and MDW941 (5) were prepared and their inhibitory/labeling potential was compared head to head.

Figure 2. (A) Structure of cyclophellitol and ABPs based on cyclophellitol. (B) Mechanism of inhibition by cyclophellitol.

Results and Discussion

Azido-cyclophellitol 3 was synthesized as depicted in Scheme 1. Core carbocycle 6 was

prepared in seven steps from

-xylose as described by Madsen and co-workers.

Selective

tosylation of the primary alcohol in 6 using p-toluenesulfonyl chloride in CH

Cl

followed

by nucleophilic substitution of the crude tosylate with sodium azide afforded azido alcohol

chromatography prior to azidation resulted in a dramatic drop in the yield (42 % over 2

steps). Next, attention was focused on the introduction of the oxirane ring. In an initial

attempt, the double bond in 7 was reacted with m-chloroperoxybenzoic acid (mCPBA)

Epoxidation proceeded in a moderate yield (52%) and subsequent removal of the benzyl

Scheme 1. Synthesis of azidocyclophellitol 3.

Reagents and conditions: (a) i) p-TsCl, Et3N, CH2Cl2, 0°C; ii) NaN3, DMF, 60°C, 72%; (b) BCl3, CH2Cl2, -78°C; (c) Ac2O, pyr; (d) CF3COCH3, oxone, NaHCO3, MeCN/H2O, 13: 20%, 14: 49 %; (e) NaOMe, MeOH, 75%; (f) BzCl, pyr, 70%.

protective groups with boron trichloride proved even more problematic. Though this method has been successfully used by Serrano et al. to deprotect stereoisomers of cyclophellitol,

in this case it exclusively yielded the ring-opened epoxide. To circumvent ring-opening, the benzyl groups in 7 were removed under the agency of BCl

prior to epoxidation. The resulting hydroxyls of crude 8 were converted to their corresponding acetyl esters by treatment with acetic anhydride in pyridine. Subsequent epoxidation of the alkene in 9 under the agency of mCPBA was uneventful. Treatment of cyclohexene 9 with

afforded oxirane 10 as a mixture of diastereomers which was inseparable both in its protected form 10 and in its unprotected from 11. It was reasoned that changing the protective group pattern on 9 may result in a seperable mixture. To this end, the hydroxyls in 8 were protected as benzoyl esters using benzoyl chloride in pyridine. Subsequent epoxidation of 12 with TFDO afforded epoxides

sodium methoxide in methanol afforded two-step probe 3 which can be applied as such in biological experiments. Azidocyclophellitol 3 was converted to direct probes 4 and 5 by conjugating either BODIPY-alkyne 15 (green emission) or BODIPY-alkyne 16 (red emission) to it employing the copper catalyzed click reaction.

Green fluorescent probe 4 and red fluorescent probe 5 were obtained in moderate to good yield (Scheme 2).

O OH

OH OH HO

BzO OBz

OBz N3

O

HO OH

OH N3

O

HO OBn

OBn N3

HO OBn

OBn HO

HO OH

OH N3

BzO OBz

OBz N3

O 7 steps

ref. 12

6 7

8

13 14

D-Xylose

KY170 (3)

a

b

c

e

AcO OAc

OAc N3

BzO OBz

OBz N3

RO OR

OR N3

O

d

d

f 9

10 R = Ac 11 R = H

12 e

Scheme 2. Synthesis of direct probes 4 and 5 from two-step probe 3.

Reagents and conditions: (a) 15 or 16, CuSO4 (10 mol%), sodium ascorbate (15 mol%), Tol/tBuOH/H2O, 90°C, 4:

56%, 5: 77%.

The inhibitory potential of probes 3, 4 and 5 against almond β-glucosidase and glucocerebrosidase was evaluated. Apparent IC

values of the probes were determined by preincubating the enzymes with a concentration series of the probe for 30 minutes followed by incubation with the fluorescent substrate, 4-methylumbelliferyl glucoside, and measuring of the fluorescence (Table 1). Kinetic studies were performed to obtain insight in the binding-constants of potent probes. Inhibition of an enzyme by a mechanism based covalent inhibitor can be regarded as a two-step process.

First a non-covalent enzyme- inhibitor complex is formed which then reacts to form a covalent adduct. Formation of the initial complex depends on the concentration of both the enzyme and the inhibitor. The second step, which is often rate-limiting, is proportional to the concentration of complex formed. Inhibition will be pseudo-first order when the conditions for an experiment are set such that inhibitor concentration is much greater than the enzyme concentration. Plotting of the obtained pseudo-first order constant versus the concentration in a Kitz-Wilson plot allowed determination of the equilibrium constant for initial binding to almond β- glucosidase (K

) and the inactivation rate constant (k

).

Glucocerebrosidase was rapidly inactivated with low concentrations of inhibitor and therefore Kitz-Wilson plots could not be used to determine binding-constants for this enzyme.

The IC

values and binding constants of the probes 3-5 were compared with the values reported in literature for cyclophellitol 2 and conduritol B epoxide (CBE) 17 where applicable.

The azido group in 3 resulted in a small (5-fold (IC

) to 10 fold (relative k

/K

values)) decrease in the potency for almond β-glucosidase. KY170 (3) however is still 10 (relative k

/K

values) to 17 fold (IC

) more potent than CBE (17).

The increased potency of 3 in perspective to CBE 17 may be accounted to the additional methylene azide and the fact that commercial available CBE 17 is sold as racemic mixture. Incorporation of a BODIPY reporter group as in MDW933 (4) and MDW941 (5) reduced the potency for

N3

HO OH

OH O

MDW941(5) 15

16 N BN F F MeO

MeO N BN F F

MDW933 (4)

3

almond β-glucosidase dramatically (IC

). Most likely, binding of these direct probes to the pocket-shaped active site of the enzyme

is impaired by the sterically encumbered reporter group.

Where the bulky reporter group is not tolerated by almond β-glucosidase, it does appear to fit in the active site of glucocerebrosidase. In fact, the lipophilic BODIPY in MDW933 (4) and MDW941 (5) actually has a beneficial effect on the inhibitory potency. Its incorporation increased the potency by 100 fold (IC

0.95 nM and 1.56 nm compared to 110 nM for KY170 (3)). Entry of hydrophobic substrates/inhibitors is favored, due to the hydrophobic surface located near the active site of GBA-1

and therefore it is hypothesized that the increase in inhibitory potential is caused by the increased overall hydrophobicity.

Table 1. IC50 values of CBE, cyclophellitol 2, two-step probes 3 and direct probes 4 and 5.

almond -glucosidase Glucocerebrosidase Probe

(IC50 M) Ki (mM) ki (min^-1) ki/Ki (IC50 nM)

CBE (17) 461±165 1.70^b 0.13^b 0.076 9,490±420

Cyclophellitol 5^a 0.34^b 2.38^b 7 ND

KY170 (3) 27±2.8 0.90±0.079 0.66±0.027 0.73 120±36

MDW933 (4) 760±427^d ND ND ND 1.24±0.044

MDW941 (5) >800^d ND ND ND 1.94±0.085

a) The literature value of cyclophellitol is reported. This value cannot directly be compared to the IC50 value determined for KY170 (3) since cyclophellitol was not preincubated with the enzyme prior to determining the activity.^b) Reported literature values.^19c) Standard errors are reported and have been determined with graphpad prism 5.0. ^d) Above 100 μM the probes start to precipitated and therefore the determined IC50-values are only an indication.

Prior to evaluating the labeling profile of probes 3-5, the two-step labeling approach using 3 as an ABP was optimized. The activity of recombinant GBA-1 was fully blocked by incubating the enzyme with 3 (10 μM) for 30 minutes and the formed adduct was modified employing either the Staudinger-Bertozzi ligation or the copper catalyzed click reaction.

These experiments revealed that both ligation methods can be used to visualize the inhibited glycosidase. The click reaction in combination with BODIPY-alkyne 15 gave, despite non-specific labeling of BSA, the strongest signal. To reduce non-specific labeling, the influence of the time, the reductor and the amount of SDS, BODIPY-alkyne and CuSO

on the ligation step was investigated. After exposing the inhibited enzyme for 4-16h to the

click-cocktail, saturation of the fluorescent signal was observed (Figure 3A). Tris(2-

carboxyethyl)phosphine (TCEP), sodium ascorbate and dithiothreitol (DTT) were used to

reduce Cu(II)SO

to catalytically active Cu(I). Of these, DTT gave the best results (Figure

3B). At least 0.5-1 mM CuSO

was required to effectively ligate BODIPY-alkyne 15 to the

azide of the inhibited enzyme (Figure 3C). Also the amount of SDS and alkyne 15 used in

the click buffer appeared of great importance. Both the signal and the non-specific labeling

of BSA increased when 0.05% SDS was used instead of 1% (Figure 3D). By using an

Figure 3. Optimization of the click reaction. (A) Glucocerebrosidase (200 ng) pretreated with 3 was incubated with the click buffer (50 mM NaOAc pH 6.0 (80 μL), 10 μM BODIPY-alkyne 15, 1 mM CuSO4, 0.5 mM DTT, 100 μM TBTA) for the indicated time. (B) Pretreated glucocerebrosidase was labeled with click buffer in the presence of the indicated reductor. (C) Influence of the amount of copper sulfate used during the click reaction. (D) Influence of the amount of SDS.

(E) Influence of the amount of BODIPY-alkyne 15.

equimolar amount of BODIPY-alkyne 15 during the click reaction, non-specific labeling could be minimized (Figure 3E). The following conditions proved to be optimal for ligation of the BODIPY fluorophore: inhibited enzyme was diluted with sodium acetate buffer (30 μL, pH 6.0), a premixed solution of TBTA (10 μL, 1 mM in DMF), CuSO

(1 μL, 0.1 M in H

O), BODIPY-alkyne 15 (0.5 μL, 0.2 mM in MeCN) and DTT (0.5 μL, 0.1 M in H

O) was added and the solution was incubated overnight at room temperature. Even with these optimized conditions background labeling was observed. The intensity of remaining background labeling is comparable both in the presence and absence of probe 3 and therefore it is assumed that this is caused by a non-specific reaction of the fluorophore under the click conditions applied.

Figure 4. (A) Enzyme (left panels: recombinant GBA-1 and right panels: almond β-glucosidase) was incubated with the indicated amount of probe at 37°C for 30 min. Enzyme labeled with KY170 (3) was diluted with acetate buffer (50 mM pH 6, 0.05% SDS) before a mixture TBTA (10 μL, 1 mM in DMF), BODIPY-alkyne 15 (1 equiv.

compared to the probe), 1 μL CuSO4 (0.1 M), 0.5 μL DTT (0.1M)) was added. The reaction was incubated for 16h.

The labeled proteins were resolved by SDS-PAGE and visualized by scanning of the fluorescence. (B) Active GBA- 1 is required for labeling. Lane 1: GBA-1 was treated with probe (3: 10 μM, 4 and 5: 0.2 μM) at 37°C for 30 min.

Lane 2: GBA-1 was pretreated with CBE (2 mM) for 30 min followed by incubating with probe (similar concentrations were used as in the control). Lane 3: GBA-1 was incubated with the probe in the presence of AMP- DNM (2 mM). Lane 4: GBA-1 was heat denatured with 1% SDS prior to labeling with the probe.

Having optimized the two-step labeling conditions, labeling of glucocerebrosidase and almond β-glucosidase by the probes was investigated. Both green fluorescent probe MDW933 (4) and two-step probe KY170 (3) labeled the enzymes in a concentration dependent fashion (Figure 4A). Saturation of the fluorescent signal was observed at 0.5 μM (glucocerebrosidase) and 100 μM (almond β-glucosidase). Red fluorescent probe MDW941 also labeled glucocerebrosidase effectively. Labeling of almond β-glucosidase with this probe was however unsuccessful. To validate that the probes are activity-based, heat- inactivated GBA-1 (Figure 4A) was incubated with the probes 3-5. Denaturation resulted in complete loss of signal, indicating that activity is required for labeling. The active site specificity of 3-5 was evidenced by competition experiments with known inhibitors CBE 17 and AMP-DNM

(Figure 4B). Addition of either the irreversible inhibitor 17 or the potent reversible AMP-DNM abolished labeling by 3-5.

To assess the sensitivity of the two methods, decreasing amounts of enzyme were treated with the probes and the signals were compared. Enzyme labeled with direct probes 4 and 5 was directly subjected to gel-electrophoresis and enzyme labeled with the two-step probe 3 was first conjugated to BODIPY-alkyne 15 or 16. As little as 1 ng purified GBA-1 and 30 ng almond β-glucosidase could be visualized using either of the methods (Figure 5). Generally, the sensitivity of the two-step probe is only three fold less sensitive as the direct probes.

Probe 3 in combination with the two-step strategy proved to be especially suitable for almond β-glucosidase. By using this method, less non-specific labeling of BSA was observed. Encouraged by the sensitivity of the probes, both the two-step probe 3 and direct probes 4 and 5 were applied to label the low abundant glucocerebrosidase in homogenates of cultured RAW-cells. Four different forms of this enzyme with a molecular mass ranging from 58-66 kDa due to glycan differences were visualized using MDW933 (4) and

Figure 5. Sensitivity of the two-step probe 3 compared to the direct probes 4 and 5. Labeling of (A) almond β- glucosidase (β-Glu) and (B) recombinant glucocerebrosidase. (C) Labeling of GBA-1 in RAW-homogenates.

Upper panel: signal observed after fluorescent read-out. Lower panel: coomassie brilliant blue staining of the same gel.

MDW941 (5) (Figure 4C). No significant labeling was observed in homogenates treated with KY170 (3). Coomassie brilliant blue (CBB) staining of the gel revealed a marked drop in proteins in the samples subjected to the click-cocktail. This suggests that the proteins aggregate during the click reaction despite the presence of 0.1% SDS which may explain the lack of signal. Further optimization will be required before the two-step strategy can be applied for the labeling of glucocerebrosidase in complex samples.

Conclusion

This chapter describes the synthesis and biological evaluation of a new class of glucosidase activity-based probes. KY170 (3) was synthesized in four steps from previously described carbocycle 6. Biological evaluation showed that two-step probe 3 is a potent inhibitor of almond β-glucosidase and glucocerebrosidase. By applying the two-step strategy, both enzymes were visualized. Further optimization of the two-step approach is needed to be able to label glucosidases in complex samples. Direct probes MDW933 (4) and MDW941 (5) were obtained in a single step from 3. These probes proved to be less suitable for the labeling of almond β-glucosidase compared to two-step probe 3. The sterically encumbered BODIPY group reduced the potency of the compounds dramatically. More importantly, severe back-ground labeling was observed in labeling experiments. Incorporation of a BODIPY had, however, a beneficial effect on the potency for glucocerebrosidase and allowed direct read-out of this low-abundance enzyme. In cell-lysates, GBA-1 could be labeled selectively by activity-based probes 4 and 5.

Experimental section

All reagents were of commercial grade and used as received unless stated otherwise. Diethyl ether (Et2O), ethyl acetate (EtOAc), light petroleum ether and toluene were obtained from Riedel-de Haën.

Acetonitrile, dichloromethane, dimethylformamide (DMF), methanol (MeOH), pyridine, tetrahydrofuran (THF) were purchased from Biosolve. Dichloromethane was distilled from CaH2 and THF was distilled over LiAlH4 prior to use. All reactions were performed under an inert atmosphere of Argon unless stated otherwise. Solvents used for flash chromatography were of pro analysi quality.

Reactions were monitored by TLC analysis using Merck aluminum sheets precoated with silica gel 60 with detection by UV-absorption (254 nm) and by spraying with a solution of (NH4)6Mo7O24.H2O (25 g/L) and (NH4)4Ce(SO4)4.H2O (10 g/L) in 10% sulfuric acid followed by charring at ~150°C or by spraying with 20% sulfuric in ethanol followed by charring at ~150°C.Column chromatography was performed using either Baker- or Screening Device silica gel in the indicated solvents. ¹H NMR and

13C NMR spectra were recorded on a Bruker DMX-400 (400/100 MHz) or a Bruker AV-400 (400/100 MHz) spectrometer in the given solvent. Chemical shifts are reported as δ-values in ppm relative to the chloroform residual solvent peak or tetramethylsilane (TMS) as internal standard. Coupling constants are given in Hz. All given ¹³C spectra are proton decoupled. Spin multiplicities are given as s, d, dd, ddd, dddd, dt, t, td, q and m. High resolution mass spectra were recorded with a LTQ Orbitrap (Thermo Finnigan). LC/MS analysis was performed on a Jasco HPLC-system (detection simultaneously at 214 nm and 254 nm) equipped with an analytical Alltima C18 column (Alltech, 4.6 mmD × 50 mmL, 3μ particle size) in combination with buffers A: H2O, B: acetonitrile and C: 1% aq.

TFA and coupled to a Perkin Elmer Sciex API 165 mass instrument. Optical rotations were measured on a Propol automatic polarimeter (sodium D line, λ = 589 nm). FT-IR-spectra were recorded on a

Paragon-PE 1000. Throughout this report the atoms in all compounds are numbered according to the figure below.

(2S,3R,4S,5S)-2,3-Bis(benzyloxy)-4-hydroxy-5-azidomethyl-cyclohex-6-ene (7)

To a solution of 6 (1.24 g, 3.65 mmol) in dichloromethane (26 mL) were added p- toluenesulfonylchloride (1.04 g, 5.48 mmol, 1.1 equiv.) and triethylamine (0.90 ml, 6.57 mmol, 1.8 equiv.) at 0°C. The solution was stirred for 5h before being poured in 1M HCl solution. The mixture was extracted with Et2O and the organic layer was dried over MgSO4 before being concentrated in vacuo. The tosylated intermediate was immediately subjected to azidation. To a solution of tosylated intermediate (1.75 g, 3.65 mmol) in DMF (35 mL) was added sodium azide (2.40 g, 36.7 mmol, 10.4 equiv.). The solution was stirred for 24h at 60 °C before being concentrated in vacuo. The crude product was diluted with EtOAc, washed with 1M HCl, saturated aqueous NaHCO3

and brine. The combined organic layers were dried over MgSO4 and concentrated in vacuo.

Purification by silica column chromatography (8% EtOAc/PE→16% EtOAc/PE) afforded 7 (900 mg, 2.64 mmol, 72%) as a solid. ¹H NMR (400 MHz, CDCl3) δ ppm 7.33-7.26 (m, 10H), 5.79 (dt, J = 10.4, 2.4 Hz, 1H), 5.58 (dt, J = 10.4, 2.4 Hz, 1H), 5.02 (d, J = 11.3, 1H), 4.7 (dd, J = 11.2, 5.4 Hz, 2H), 4.65 (d, J = 11.2 Hz, 1H), 4.21-4.19 (m, 1H), 3.61-3.53 (m, 3H), 3.44 (dd, J = 12.0, 6.0 Hz, 1H), 2.83 (s, 1H), 2.48 (br, 1H). ¹³C NMR (100 MHz, CDCl3) δ ppm 138.1, 137.2, 128.7, 128.6, 128.0, 127.9, 127.8, 127.7, 83.5, 80.3, 75.0, 71.6, 52.5, 43.6. FT-IR: νmax (neat)/cm^-1: 2095.9, 1497.1, 1453.9, 1275.9, 1092.6, 1050.4, 1027.7. [α]D20 +137.8° (c=1, CHCl3). LC/MS: Rt 9.35; linear gradient 10→90% B in 13.5 min; ESI/MS:

m/z=383.1 (M+NH4)⁺.HRMS: (M+3H⁺-N2) calcd for C21H26NO3 340.19072 found 340.19080.

(2S,3R,4S,5S)-2,3,4-Benzoate-5-azidomethyl-cyclohex-6-ene (12)

Borontrichloride (21 mL, 21.1 mmol, 10 equiv.) was added to a solution of 7 (777.1 mg, 2.11 mmol) in anhydrous dichloromethane (10 mL) at -78°C. The reaction mixture was stirred at -78°C for 6h before being quenched with MeOH.

The solution was concentrated in vacuo giving the triol intermediate 8, which was immediately used for benzoylation. The crude product was coevaporated several times with anhydrous toluene before being dissolved in pyridine (10 mL). Benzoylchloride (2.6 mL, 21.1 mmol, 10 equiv.) was added at 0°C and the reaction mixture was stirred for 18h at ambient temperature. The mixture was quenched with saturated aqueous NaHCO3, extracted with EtOAc, dried over MgSO4 and concentrated in vacuo.

Purification by silica column chromatography (4% EtOAc/PE→6% EtOAc/PE) afforded 12 (701.8 mg, 1.46 mmol, 70%) as a yellow oil. ¹H NMR (400 MHz, CDCl3) δ ppm 7.99 (d, J = 7.2 Hz, 2H), 7.92 (d, J

= 7.2 Hz, 2H), 7.84 (d, J = 7.2 Hz, 2H), 7.53-7.46 (m, 3H), 7.40 (dt, J = 24.4, 8.0 Hz, 5H), 7.26-7.18 (m, 2H), 6.00-5.93 (m, 3H), 5.86 (d, J = 10.0 Hz, 1H), 5.72 (t, J = 9.2 Hz, 1H), 3.64 (dd, J = 12.4, 4.0 Hz, 1H), 3.46 (dd, J = 12.4, 6.4 Hz, 1H), 2.99-2.97 (m, 1H). 13C NMR (100 MHz, CDCl3) δ ppm 166.0, 165.9, 133.3, 133.2, 133.1, 129.8, 129.7, 129.6, 129.4, 129.0, 128.9, 128.5, 128.4, 128.3, 126.2, 127.0, 72.7, 72.7, 72.6, 70.4, 52.0, 42.5. FT-IR: νmax (neat)/cm^-1: 2100.3, 1718.0, 1601.8, 1585.4, 1492.2, 1314.5, 1250.8, 1178.0, 1108.8, 1031.8, 1025.9, 950.7. [α]D20 +173° (c=1.0, CHCl3). LC/MS: Rt 10.68; linear gradient 10→90% B in 13.5 min; ESI/MS: m/z = 498.2 (M+H)⁺.HRMS: (M+Na⁺) calcd for C28H25NO6

520.14791 found 520.14724.

(2S,3R,4S,5S)-2,3,4-Benzoate-7-azidocyclophellitol (13 and 14)

A solution of 0.4 mM Na2EDTA solution in H2O (3.1 mL) and trifluoroacetone (1.34 mL, 15 mmol, 15 equiv.) were added to 12 (497 mg, 1.0 mmol) in acetonitrile (6.7 mL). A mixture of oxone (3.07 g, 5.0 mmol, 5 equiv.) and NaHCO3 (588.1 mg, 7.0 mmol, 7 equiv.) was added to the solution over a period of 15 min. After stirring at 4°C for 4h, an additional amount of 0.4 mM Na2EDTA in H2O (1.5 mL), trifluoroacetone (0.7 mL, 7.5 mmol, 7.5 equiv.) and a mixture of oxone (1.5 g, 2.5 mmol, 2.5 equiv.) and NaHCO3 (290 mg, 3.5 mmol, 3.5 equiv.) were added to the reaction mixture over a period of 15 min. The reaction mixture was stirred at 4°C for 30 min before being diluted with H2O. After extraction of the water layer with EtOAc, the combined organic layers were dried over MgSO4 and

N3

BzO OBz

OBz N3

HO OBn

OBn

concentrated in vacuo. Purification by silica column chromatography (8% Et2O/PE→18% Et2O/PE) afforded 13 (103.9 mg, 0.20 mmol, 20%) and 14 (253.7 mg, 0.49 mmol, 49%) respectively as white crystals.

13: ¹H NMR (400 MHz, CDCl3) δ ppm 8.03 (d, J = 7.4 Hz, 2H), 7.88 (d, J = 7.6 Hz, 2H), 7.79 (d, J = 7.6 Hz, 2H), 7.56 (t, J = 7.2 Hz, 1H), 7.46-7.36 (m, 5H), 7.32 (t, J = 7.6 Hz, 2H), 7.24 (t, J = 7.2, 2H), 5.84 (t, J = 9.2 Hz, 1H), 5.56 (d, J = 8.8 Hz, 1H), 5.43 (t, J = 10.0 Hz, 1H), 3.67-3.62 (m, 4H), 3.44 (s, 1H), 2.71 (dddd, J = 9.3, 7.6, 1.4 Hz, 1H). ¹³C NMR (100 MHz, CDCl3) δ ppm 165.7, 165.6, 165.4, 133.5, 133.3, 133.1, 129.8, 129.6, 129.5, 128.9, 128.7, 128.6, 128.4, 128.3, 128.1, 72.2, 71.4, 67.8, 54.7, 54.2, 50.5, 40.8. FT-IR: νmax (neat)/cm^-1: 2104.5, 1722.6, 1601.9, 1451.6, 1315.2, 1258.3, 1178.4, 1094.8, 1069.6, 1026.1. [α]D20 + 93.6° (c = 1.0, CHCl3). LC/MS: Rt 10.24; linear gradient 10→90% B in 13.5 min;

ESI/MS: m/z = 514.2 (M+H)⁺. HRMS: (M+H⁺) calcd for C28H23N3O7 514.16088 found 514.16007.

14: ¹H NMR (400 MHz, CDCl3) δ ppm 8.02 (d, J = 7.2 Hz, 2H), 7.89 (d, J = 7.2 Hz, 2H), 7.78 (d, J = 7.2 Hz, 2H), 7.53-7.19 (m, 5H), 5.96 (t, J = 9.6 Hz, 1H), 5.77 (d, J = 8.8 Hz, 1H), 5.55 (t, J = 9.6 Hz, 1H), 3.77-3.74 (m, 2H), 3.64 (dd, J = 12.8, 4.0 Hz, 1H), 3.32 (s, 1H), 2.68 (ddd, J = 9.2, 5.2, 3.8 Hz, 1H). ¹³C NMR (100 MHz, CDCl3) δ ppm 166.0, 165.9, 165.6, 133.4, 133.0, 129.9, 129.7, 129.5, 129.0, 128.9, 128.6, 128.4, 128.3, 128.1, 72.1, 70.0, 69.9, 54.6, 53.8, 50.8, 40.9. FT-IR: νmax (neat)/cm^-1: 2104.6, 1717.8, 1602.1, 1451.8, 1249.4, 1178.1, 1093.3, 1069.0, 1026.0. [α]D20 +52.4° (c=1.0, CHCl3). LC/MS: Rt 10.22;

linear gradient 10→90% B in 13.5 min; ESI/MS: m/z = 514.2 (M+H)⁺.HRMS: (M+H⁺) calcd for C28H23N3O7 514.16088 found 514.16017.

Azidocyclophellitol, KY170 (3)

A catalytic amount of NaOMe was added to a solution of 13 (103.9 mg, 0.20 mmol) in MeOH (1.0 mL) and stirred for 1h at ambient temperature. The reaction mixture was neutralized with Amberlite IR-120 H⁺, filtered and concentrated in vacuo. Purification by silica column chromatrography (6% MeOH/CH2Cl2→8%

MeOH/CH2Cl2) provided 3 (30.0 mg, 0.15 mmol, 75%). ¹H NMR (400 MHz, MeOD) δ ppm 3.84 (dd, J = 8.4, 3.6 Hz, 1H), 3.67 (d, J = 8.0 Hz, 1H), 3.51 (dd, J = 12.0, 8.8 Hz, 1H), 3.36 (d, J = 3.2 Hz, 1H, H1), 3.23 (dd, J = 10.0, 8.4 Hz, 1H), 3.13-3.08 (m, 2H), 2.07 (ddt, J = 9.4, 3.6, 1.6 Hz, 2H). ¹³C NMR (100 MHz, MeOD) δ ppm 78.3, 72.7, 68.6, 57.6, 56.1, 52.4, 43.9. FT-IR: νmax (neat)/cm^-1: 3331.7, 3187.9, 2936.1, 2097.6, 1455.8, 1345.9, 1273.4, 1144.2, 1092.5, 1066.5, 1032.1, 995.1, 926.6. [α]D20+ 174.7° (c=0.6, MeOH). LC/MS: Rt 0.95; linear gradient 10→90% B in 13.5 min; ESI/MS: m/z = 219.2 (M+NH4)⁺.HRMS: (M+3H⁺-N2) calcd for C7H14NO4 176.09173 found 176.09179.

MDW933 (4)

Azido-cyclophellitol 3 (8.51 mg, 42 μmol) and BODIPY- alkyne 15 (13.8 mg, 42 μmol) were dissolved in tert- BuOH/Tol/H2O (1.8 mL, 1/1/1 v/v/v). CuSO4 (100 mM in H2O, 42 μL, 4.2 μmol) and sodium ascorbate (100 mM in H2O, 63 μL, 6.3 μmol) were added. Subsequently, the reaction was heated to 80°C and stirred overnight. The solution was diluted with CH2Cl2, washed with H2O, dried and concentrated. Purification over silica gel column chromatography (CH2Cl2→5%

MeOH/CH2Cl2) gave title compound 4 as an orange powder (56%, 12.49 mg, 23.6 Pmol).¹H NMR (600 MHz, CDCl3) δ ppm 7.40 (s, 1H), 6.01 (s, 2H), 4.68 (d, J = 12.0 Hz, 1H), 4.58 (dd, J = 13.4, 7.5 Hz, 1H), 3.66 (d, J = 5.6 Hz, 1H), 3.40-3.34 (m, 1H), 3.20-3.15 (m, 1H), 3.02 (s, 1H), 2.97 (s, 1H), 2.96- 2.91 (m, 2H), 2.73 (t, J = 6.4, 6.4 Hz, 2H), 2.49-2.46 (s, 6H), 2.45-2.40 (m, 1H), 2.33 (s, 6H), 1.86 (td, J

= 15.0, 7.6, 7.6 Hz, 2H), 1.66-1.58 (m, 2H). ¹³C NMR (150 MHz, CDCl3) δ ppm 153.9, 146.0, 140.3, 131.4, 121.7, 77.2, 77.0, 76.9, 76.7, 71.1, 67.4, 56.0, 54.5, 49.6, 43.0, 31.2, 29.5, 28.1, 25.2, 16.3, 14.4.

LC/MS: Rt 6.83; linear gradient 10→90% B in 13.5 min; ESI/MS: m/z = 530.00 (M+H)⁺.HRMS:

(M+H⁺) calcd for C26H34BF2N5O4 530.27447 found 530.27454.

N HO

OH OH O

N N N

B N F F N3

HO O

OH OH N3

BzO O

OBz OBz

N3

BzO O

OBz OBz

MDW941 (5)

Azido-cyclophellitol 3 (5.46 mg, 27 μmol) and BODIPY-alkyne 16 (13.1 mg, 27 Pmol) were dissolved in tert-BuOH/Tol/H2O (1.5 mL, 1/1/1 v/v/v). CuSO4 (100 mM in H2O, 27 PL, 2.7 μmol) and sodium ascorbate (100 mM in H2O, 41 μL, 4.1 μmol) were added.

Subsequently, the reaction was heated to 80°C and stirred overnight. The solution was diluted with CH2Cl2, washed with H2O, dried and concentrated. Purification over silica gel column chromatography (CH2Cl2→5% MeOH/CH2Cl2) gave title compound 5 as an purple powder (77%, 14.32 mg, 20.8 μmol).¹H NMR (400 MHz, CDCl3) δ ppm 7.80 (d, J = 8.5 Hz, 4H), 7.36 (s, 1H), 7.17 (d, J = 3.4 Hz, 2H), 6.89 (d, J = 8.5 Hz, 4H), 6.54 (d, J = 3.8 Hz, 2H), 5.12-4.83 (m, 1H), 4.80-4.46 (m, 3H), 3.77 (s, 6H), 3.75-3.68 (m, 1H), 3.45-3.34 (m, 1H), 3.26-3.13 (m, 1H), 3.11-3.02 (m, 1H), 3.00-2.94 (m, 1H), 2.90-2.78 (m, 2H), 2.73-2.58 (m, 2H), 2.51-2.36 (m, 1H), 2.08-1.91 (m, 2H), 1.84-1.67 (m, 4H).

13C NMR (150 MHz, CDCl3) δ ppm 160.5, 157.4, 144.6, 136.0, 130.9, 126.8, 125.0, 120.0, 113.7, 71.0, 67.2, 56.0, 55.2, 54.6, 49.6, 42.7, 33.0, 30.3, 29.7, 29.4, 25.0. LC/MS: Rt 8.35; linear gradient 10→90% B in 13.5 min; ESI/MS: m/z = 686.07 (M+H)⁺.HRMS: (M+H⁺) calcd for C36H38BF2N5O4 686.29560 found 686.29559.

Determination of the IC50

Prior to determination of the IC50, the enzymes were dissolved in the appropriate buffer. The buffer system employed for glucocerebrosidase was a McIlvain buffer (50 mM citric acid, 100 mM Na2HPO4, pH 5.2 containing 0.2% sodium taurocholate, 0.1% Triton X-100). For almond β-glucosidase, McIlvain buffer (50 mM citric acid, 100 mM Na2HPO4, pH 5.0) was used. The inhibitor (12.5 μL in DMSO/H2O) was added to the enzyme solution (12.5 μL). The solution was incubated at 37°C for 30 min followed by incubation with 4MU-β-glucoside (100 μL, 7.5 mM in McIlvain) at 37°C for 20 min.

The reaction was quenched by the addition of glycine/NaOH (0.3 M, pH 10.6), after which amount of liberated 4MU was determined with a Perkin Elmer Life Sciences Luminiscence Spectrometer LS-30.

IC50 values were obtained by plotting of the residual fluorescence versus the concentration (GraphPad Prism 5).

Determination of the rate constants

The time-dependent interaction of inhibitor (I) with free β-glucosidase (E) was considered to occur in separate stages (A). A rapid reversible interaction is followed by a slower, irreversible reaction that transforms the reversible enzyme-inhibitor complex [EI] into an irreversible enzyme-inhibitor adduct [EI*]. (A)

The first-order association kinetic model (B) describes the time dependence of changes in concentration of the irreversible enzyme–inhibitor adduct [EI*] in respect to the concentration of the inhibitor [I0] at t0 (Marangoni 2003): (B) [(,*] [(T](1exp^k't) with

Almond β-glucosidase: The equilibrium constant for initial binding (Ki)and the rate-constant (ki) were determined as follows. The enzyme was diluted in the appropriate McIlvain buffer (see above) before it was incubated with varying concentrations of the inhibitor. At different time-intervals an aliquot was withdrawn, added to the 4MU-substrate solution and incubated for 20 min. The reaction was stopped by de addition of glycine/NaOH (0.3 M, pH 10.6). The activity of the enzyme was determined by monitoring the release of 4-metylumbelliferone as was described above for the IC50

N HO

OH OH O

N N N

B N F F MeO

MeO

] [

] ' [

0 0

,

 ,

i i

K k k E + I

Ki

[EI] [EI]*

ki

values. The pseudo-first order rate-constants were established by plotting the logarithm of the residual activity versus the time (Figure 6A). Replotting of the reciprocal rate-constants versus the reciprocal concentration gave the Ki and ki values (Figure 6B).

Figure 6. (A) Determination of the pseudo-first order rate-constants. Almond β-glucosidase was incubated with 0.1 mM (squares), 0.3 mM (triangles), 0.5 mM (reverse triangles), 1 mM (diamonds) and 2 mM (circles) KY170 (3) for the indicated time, after which the residual activity was determined. The logarithm of the residual activity was plotted versus the time. (B) Kitz-Wilson plot of the pseudo-first order rate-constants.

Labeling experiments Two-step probe 3

GBA-1: To the enzyme (100 ng) dissolved in the appropriate McIlvain buffer was added KY170 (3) (1 μL, 100 μM). The reaction mixture was incubated at 37°C for 30 min and subsequently diluted with NaOAc buffer (30 μL, 50 mM pH 6.0, 0.05% SDS). A fresh mixture of TBTA (10 μL, 1 mM in DMF), CuSO4 (1 μL, 0.1 M in H2O), DTT (0.5 μL, 0.1 M in H2O) and BODIPY-alkyne 15 (0.5 μL, 0.2 mM in MeCN) was prepared, added to the enzyme solution and the resulting mixture was incubated overnight at room temperature. The reaction was quenched by the addition of 4u sample buffer (15 μL) and loaded on a 7.5% SDS-PAGE gel. The fluorescence was measured in the wet gel slabs using the CY2 settings (λex 488, λem520) on a Typhoon Variable Mode Imager (Amersham Biosciences).

Almond β-glucosidase: To the enzyme (100 ng) dissolved in the appropriate McIlvain buffer was added KY170 (3) (1 μL, 1 mM). The reaction mixture was incubated at 37°C for 30 min and subsequently diluted with NaOAc buffer (80 μL, 50 mM pH 6.0, 1% SDS). A fresh mixture of TBTA (10 μL, 1 mM in DMF), CuSO4 (1 μL, 0.1 M in H2O), DTT (0.5 μL, 0.1 M in H2O) and BODIPY- alkyne 15 (0.5 μL, 2 mM in MeCN) was prepared, added to the enzyme solution and the resulting mixture was incubated overnight at room temperature. The proteins were precipitated by the addition of ice-cold acetone (1 mL) followed by incubating at -20°C for 20 min and centrifugation (16.000 g, 15 min) at 4°C. The proteins were resolved and analyzed as described above.

Direct probes 4 and 5

GBA-1 or almond β-glucosidase (100 ng) was incubated with the direct probe (1 μL, 10u stock) at 37°C for 30 min and subsequently the reaction was quenched by the addition of 4u sample buffer (4 μL). The proteins were resolved on a 7.5% SDS-PAGE gel. In-gel visualization of protein labeling was directly performed in the wet gel slabs using either the CY2 settings (λex 488, λem520) for 4 or the SYPRO-Ruby settings (λex 532, λem 610) for 5 on a Thyphoon Variable Mode Imager (Amersham Biosciences).

Competition experiments

Protocol for irreversible inhibitors: 9 μL of GBA-1(20 ng) in reaction buffer was preincubated with CBE (1 μL, 2 mM final concentration) at 37°C for 30 min followed by incubation with the probe (final concentrations 3: 10 μM, 4: 0.2 μM and 5: 0.2 μM). The reaction was quenched either by the addition

of 4× SDS-PAGE sample buffer (5 μL) (4 and 5) or by the addition of the click buffer (3). The samples were treated and analyzed by SDS-PAGE as described above for the labeling experiments.

Protocol for reversible inhibitors: 9 μL of GBA-1 (20 ng) in reaction buffer was incubated with AMP-DNM (1 μL, 2 mM final concentration) in combination with 1 μL of the probe (final concentrations 3: 10 μM, 4: 0.2 μM and 5: 0.2 μM). After quenching of the reaction, the samples were handled as described above.

Protocol for heat-inactivation: 1 μL SDS (10%) was added to 9 μL of GBA-1 (20 ng) in reaction buffer. The enzyme was heated at 100°C for 5 min, cooled to rt after which 1 μL of the probe (final concentrations 3: 10 μM, 4: 0.2 μM and 5: 0.2 μM) was added. After incubating the sample at 37°C for 30 min, the samples were handled as described above.

Sensitivity of the probes

Decreasing amounts of protein in McIlvain buffer were incubated with the probes 3-5 (1 μL, 10u stock) for 30 min at 37°C. The samples were either subjected to the click-cocktail described above (3) or quenched by the addition of 4u sample buffer (4 and 5). The proteins were resolved on a 7.5% SDS- PAGE gel after which the fluorescence was measured in the wet gel slabs as described above.

References and footnotes

(1) Morant, A.V.; Jorgensen, K.; Jorgensen, C.; Paquette, S.M.; Sanchez-Perez, R.; Moller, B.L.; Bak, S. Phytochemistry 2008, 69, 1795.

(2) Butters, T.D. Curr. Opin. Chem. Biol. 2007, 11, 412.

(3) Evans, M.J.; Cravatt, B.F. Chem. Rev. 2006, 106, 3279.

(4) (a) Timmer, M.S.; Stocker, B.L.; Seeberger, P.H. Curr. Opin. Chem. Biol. 2007, 11, 59; (b) Stubbs, K.A.: Vocadlo, D.J. Methods Enzymology 2006, 415, 244; (c) Stubbs, K.A.; Vocadlo, D.J. Aust. J.

Chem. 2009, 62, 521.

(5) Tsai, C.-S.; Li, Y.-K.; Lo, L.-C. Org. Lett. 2002, 4, 3607.

(6) Vocadlo, D. J.; Bertozzi, C. R. Angew. Chem. Int. Ed. 2004, 43, 5338.

(7) (a) Stubbs, K.; Scaffidi, A.; Debowski, A.W.; Mark, B.L.; Stick, R.V; Vocadlo, D.J. J. Am. Chem.

Soc. 2008, 130, 327; (b) Hekmat, O.; Florizone, C.; Kim, Y.-W.; Eltis, L.D.; Warren, R.A.J.;

Withers, S.G. ChemBioChem 2007, 8, 2125.

(8) (a) Atsumi, S.; Umezawa, K.; Iinuma, H.; Naganawa, H.; Nakamura, H.; Iitaka, Y.; Takeuchi, T. J.

Antibiot. 1990, 43, 49; (b) Atsumi, S.; Iinuma, H.; Nosaka, C.; Umezawa, K. J. Antibiot. 1990, 43, 1579; (c) Legler, G. Z. Physiol. Chem. 1966, 345, 197; (d) Rempel, B.P.; Withers, S.G. Glycobiology 2008, 18, 570.

(9) Gloster, T.M.; Madsen, R.; Davies, G.J. Org. Biomol. Chem. 2007, 5, 444.

(10) Saxon, E.; Berozzi, C.R. Science 2000, 287, 2007.

(11) Speers, A.E.; Adam, G.C.; Cravatt, B.F. J. Am. Chem. Soc. 2003, 125, 4686.

(12) Hansen, F.G.; Bundgaard, E.; Madsen, R. J. Org. Chem. 2005, 70, 10139.

(13) Serrano, P.; Egido-Gabas, M.; Llebaria, A.; Delgado, A. Tetrahedron: Asymm. 2007, 18, 1971.

(14) Yang, D.; Wong, M.-K.; Yip, Y.-C. J. Org. Chem. 1995, 60, 3887.

(15) Verdoes, M.; Hillaert, U.; Florea, B.I.; Sae-Heng, M.; Risseeuw, M.D.P.; Filippov, D.V.; van der Marel, G.A.; Overkleeft, H.S. Bioorg. Med. Chem. Lett. 2007, 17, 6169.

(16) Marangoni, A.G. Enzyme Kinetics: A Modern Approach, (2003) pp. 70 – 78. John Wiley & Sons, Inc. ISBN: 0-471-15985-9.

(17) Kitz, R.; Wilson, B.I. J. Biol. Chem. 1962, 237, 3245.

(18) Conduritol B epoxide (17) is a broad spectrum glucosidase inhibitor, which has been widely applied to study glucosidases. Legler, G. Z. Physiol. Chem. 1966, 345, 197.

Structure of CBE (17):

HO

HO HOHO

O

17

(19) Withers, S.G.; Umezawa, K. Biochem. Biophys. Res Comm. 1991, 177, 532-537.

(20) (a) Namchuk, M.N.; Withers, S.G. Biochemistry 1995, 34, 16194-16202; (b) Zechel, D.L.; Withers, S.G. Ann. Chem. Rev. 2000, 33, 11.

(21) Premkumar, L.; Sawkar, A.R.; Boldin-Adamsky, S.; Toker, L.; Silman, I.; Kelly, J.W.; Futerman, A.H.; Sussman, J.L. J. Biol. Chem. 2005, 280, 23815.

(22) 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.; Aerts, J.M.F.G. J. Biol. Chem. 1998, 273, 265222.

Structure of AMP-DNM:

HO N HO

HO

OH

O