Concepts and Case Studies in Chemical Biology (Waldmann/Concepts and Case Studies in Chemical...

191

13

Rational Design of Activity-Based Retaining 𝛃-ExoglucosidaseProbes

Kah-Yee Li, Wouter Kallemeijn, Jianbing Jiang, MartheWalvoort, LianneWillems, Thomas

Beenakker, Hans van den Elst, Gijs van der Marel, Jeroen Codée, Hans Aerts, Bogdan Florea, Rolf

Boot, MartinWitte, and Herman Overkleeft

13.1

Introduction

Activity-based protein profiling (ABPP) is one of the most visible areas of

research in chemical biology where organic chemistry plays an essential role.

Activity-based probes (ABPs) have been developed for numerous serine hydro-

lases, cysteine proteases, and threonine hydrolases (see also Chapter 12), but less

frequently for other enzyme families. This chapter details the successful develop-

ment and application of a number of activity-based retaining β-exoglucosidaseprobes. The design principles of these probes can serve as a blueprint for the

development of ABPs aimed at various retaining glycosidase families, next to

exoglycosidases and also endoglycosidases.

13.2

The Biological Problem

Biochemical and biological research on carbohydrates and glycoconjugates, their

structure, and their function is complicated. Carbohydrates and glycoconjugates

often exist only transiently, are heterogeneous in structure, and their biosynthesis

is only indirectly controlled by the genetic code. The combined action of glycosyl

transferases and glycosidases – enzymes that create and break glycosidic linkages,

respectively – in conjunction with substrate levels determine the nature of the

eventual carbohydrate structures. Therefore, the nature of the pool of carbohy-

drates and glycoconjugates present in a given organism (termed a glycome) can, in

contrast to proteins and nucleic acids, not be extracted from the genetic material

of this organism.

Studies toward the glycome are further complicated by its structural complex-

ity. The chemical space covered by carbohydrate-containing compounds is vast.

Compared to their biopolymer counterparts, nucleic acids and peptides/proteins,

Concepts and Case Studies in Chemical Biology, First Edition. Edited by Herbert Waldmann and Petra Janning.© 2014 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2014 by Wiley-VCH Verlag GmbH & Co. KGaA.

192 13 Rational Design of Activity-Based Retaining β-Exoglucosidase Probes

which are both synthesized from a relatively small set of building blocks,

carbohydrates and glycoconjugates are constructed from a large variety of

monosaccharide building blocks. A limited set of monosaccharides (predomi-

nantly D-glucose, D-mannose, D-galactose, D-glucuronic acid, D-xylose, D-ribose,

L-idose, D-neuraminic acid) is used to construct the glycome in humans, but,

for instance, bacterial glycomes contain up to hundreds of monosaccharides

differing in stereochemistry and functional group pattern. Monosaccharide

building blocks can, and are, interconnected through glycosidic bonds to

various positions of the core of other monosaccharides forming oligomeric

structures, called oligosaccharides and polysaccharides (biomolecules composed

of monosaccharide building blocks exclusively). Hybrid biomolecules composed

of carbohydrates and lipids (glycolipids), carbohydrates, and peptides (glycopep-

tides), as well as glycoconjugates involving other biomolecules, also exist. Nucleic

acids and amino acids are linked through achiral linkages (phosphodiester bonds

and amide bonds, respectively), whereas the glycosidic linkages that make up

oligosaccharides and glycoconjugates involve a chiral (anomeric) carbon center,

increasing the structural complexity even further.

The structural complexity, together with the fact that the glycome is nontem-

plated encoded, limits the use of molecular biology techniques, and therefore

other means of studying the glycome are often employed. One attractive and

often-used strategy is to study the glycome by perturbation, which can be

achieved by manipulating the corresponding glycoprocessing enzymes, the

glycosyl transferases, and glycosidases. A host of natural and synthetic (fluoro-

genic) substrates and inhibitors that act on, predominantly, glycosidases [1] exist.

With these, the activity of a given glycosidase can be monitored (fluorogenic

substrate) or inhibited, yet direct insight in the presence and/or nature of a

glycoprocessing enzyme in a biological sample cannot be established directly and

unambiguously. ABPP does provide the means to do so but requires that suitable

ABPs are available. This chapter discusses how, by rational design, suitable ABPs

for retaining β-exoglucosidases can be designed.

13.3

The Chemical Approach

In designing ABPs for a specific enzyme/class of enzymes, both the nature of the

substrate and the mechanism employed by the enzyme are taken into considera-

tion. Ideally, the enzyme of interest forms a covalently bound enzyme-substrate

intermediate at some point of the catalytic cycle. Analysis of such a covalent inter-

mediate allows the design of a mechanism-based inhibitor, normally a substrate

analog that undergoes part of the catalytic process as if it were a substrate, only to

get stuck at the covalent intermediate stage because this covalent intermediate is

(much)more stable than that of the corresponding enzyme-substrate adduct.This

strategy has met with most success in the design of ABPs for hydrolytic enzymes,

in particular serine hydrolases, cysteine proteases, and threonine proteases

13.3 The Chemical Approach 193

OOR

OO

O

OH

H

O

OO

OR

H

OO

δ+

δ−

δ+

OH

H

HO δ−

δ−

O

O

OHO

OHROH

OOR

OO

O

O

OO

OR

H

OO

δ+

δ−

δ+

O

OO

OO

OOH

OHO

O

O

OO

OH

H

OO

δ+

δ−

δ+

O

H

H

HO

HO

ROH

HOHO

HOHO

HOHO

HO

HOHO

HOHO

HOHO

HO

HOHO

HOHO

HOHO

HOHO

HOHO

HO

HOHO

HOHO

H

H

−O

−O

−O

−O

++

++

++

(a)

(b)

Figure 13.1 Mechanism of (a) retaining and (b) inverting β-glucosidases.

(proteasomes, see also Chapter 12). Glycosidases are a large family of hydrolytic

enzymes that hydrolyze the acetal linkages that characterize oligosaccharides

and glycoconjugates to form hemiacetal linkages [2]. Mechanistic studies on

glycosidases revealed that many, although not all, enzymes from the glycosidase

family develop covalent intermediates during glycosidase action as is exemplified

for the β-exoglucosidases (Figure 13.1). In theory, such glycosidases that do form

a covalent adduct are amenable to active site labeling and therefore ABPP.

Figure 13.1a depicts the classical double displacement mechanism proposed

originally by Koshland as employed by retaining β-exoglucosidases [3]. The

enzyme active site contains two carboxylic acid (aspartate or glutamate) residues:

a general acid/base (carboxylic acid at the onset of enzyme catalysis) situated

above the β-glucose substrate and a nucleophile (carboxylate) situated below.


Upon binding, the substrate adopts a 1S3 skew-boat conformation [4]. This

distorted conformation positions the aglycon in a pseudoaxial position and aligns

the σ* orbital of the acetal linkage to the nucleophilic carboxylate residue, and

minimizes steric hindrance by H3 and H5. After protonation of the aglycon

(exocyclic oxygen of the acetal linkage) and expulsion of the leaving group,

a transient oxocarbenium ion is formed, with concomitant flattening of the

pyranose ring to the 4H3 half-chair. This putative intermediate is trapped by the

carboxylate nucleophile to form the covalently linked glucosyl-enzyme adduct

with inversion of configuration at the anomeric center (alpha product). In the

second step of the catalytic process, the covalent adduct is hydrolyzed via a similar

oxocarbenium ion transition state to produce β-glucose with overall retention

of configuration. Figure 13.1b depicts an alternative mechanism employed

by inverting β-exoglucosidases. Although these enzymes are quite similar to

retaining β-exoglucosidases with respect to their substrate (β-glucosides) and the

composition of their enzyme active site, the overall stereochemical outcome of

the hydrolysis is different: net inversion versus net retention. From a structural

point of view, the two catalytic carboxylates in inverting β-exoglucosidasesare positioned more distal (8-9Å compared to the 4-5Å observed in retaining

β-exoglucosidases) and a water molecule positioned below the scissile acetal

linkage can now be accommodated within the enzyme active site. The substrate

binds in a distorted 2S0 skew-boat and upon protonation of the aglycon, as before,

the developing oxocarbenium ion can now be trapped directly by water, itself

deprotonated by the alpha-carboxylate in the process, to yield alpha-glucose.

From the viewpoint of ABPP, the lack of a covalent intermediate makes the

development of activity-based inverting glycosidase probes rather complicated,

much more so than is the case for retaining glycosidases. Although not the

subject of this chapter, it should be noted that the development of probes to

label enzymes that do not form a covalent enzyme–substrate intermediate

mostly relies on photoreactive groups (such probes are often referred to as

photoactivatable affinity-based probes) and a recent report describes the design

of such a probe based on the competitive inhibitor, deoxynojirimycin equipped

with a photoactivatable aryl azide and a bioorthogonal tag for probing inverting

glycosidase activities [5].

As stated, inhibitors that proceed through the catalytic process, but form a long-

lived covalent intermediate, are good leads for ABP development. With respect to

retaining β-exoglucosidases, two compound classes that meet this requirement

have been studied in detail in the past decades: 2-deoxy-2-fluoroglucosides

(Figure 13.2a) and cyclitol epoxides (Figure 13.2b) [6, 7]. Substitution of the

2-hydroxyl by an electron-withdrawing fluorine, as in compound 1, results in the

formation of an enzyme-glucoside adduct that is comparatively more stable than

that formed from the natural substrate, because the 2-deoxy-2-fluoroglucoside is

comparatively less able to sustain a developing positive charge that accompanies

hydrolysis of the enzyme-substrate adduct. This feature, formation of the

oxocarbenium ion, is also inherent to the first step of the catalytic cycle and the


(a)

(b)

O OH

OO

O

OO

OO

O

OH

OHHO

HOO

OH

OHHO

HO

OF

OO

O

O

OO

OO

O

H

HO

OH

OHO

O−O

Slow

HF

OHO

HOHO

F

F

1

2 3

HOHO

HO

F

HOHO

HOF

HOHO

HOF

HOHO

HOHO

HOHO

HOHO

H

H

−O

−O

Figure 13.2 Overview of mechanism-based retaining β-exoglucosidase inhibitors and their

mode of action. (a) 2-Deoxy-β-1,2-difluoroglucose 1 and (b) cyclophellitol 3.

fluorine residue therefore also decreases the rate of formation of the glucosyl-

enzyme adduct. A good leaving group (here: fluorine) is thus a requirement to

assure that the first step of the catalytic cycle proceeds uneventfully [6]. It should

be noted that 2-deoxy-2-fluoroglycosides were employed by theWithers group [8]

in a seminal paper demonstrating the involvement of covalent enzyme-substrate

adducts in the action of retaining glycosidases and thus in proving the mecha-

nism hypothesized by Koshland (Figure 13.1a) correct. In an alternative design,

replacement of the monosaccharide core by a cyclitol analog equipped with an

pseudoequatorial epoxide produces after enzyme catalysis (protonation of the

epoxide followed by nucleophilic substitution) the ester adduct, comparatively

more stable than the acetal formed as depicted in Figure 13.1a, thereby effectively

inactivating the enzyme. As much as five decades ago, Legler and coworkers

reported on the use of conduritol B epoxide (CBE)2 (Figure 13.2b) for this purpose

[9].Themost effective inhibitor of this class is the natural product, cyclophellitol 3

(Figure 13.2b), amolecule that closely resembles β-glucopyranose in configurationand substitution pattern and as such appeared a highly potent mechanism-based

inhibitor of retaining β-exoglucosidases from different origins [10].

13.3.1

Development of a Human Acid Glucosylceramidase Activity-Based Probe

Human acid glucosylceramidase, or GBA (glucosidase, beta, acid), catalyzes

the hydrolysis of glucosylceramide to glucose and ceramide. As such, it is

responsible for the penultimate step in the turnover of glycosphingolipids,

an important metabolic pathway malfunctioning of which is responsible for


numerous inherited metabolic disorders. Mutations in the gene encoding GBA

can lead to partial malfunctioning of the enzyme, leading to accumulation

of its substrate, glucosylceramide [11, 12]. This is in a nutshell the basis of

the lysosomal storage disorder, Gaucher disease. Two Gaucher therapies are

practiced in the clinic. In enzyme replacement therapy, patients are treated with

recombinant GBA, whereas in substrate reduction therapy glucosylceramide

levels are downregulated through partial inhibition of the enzyme responsible

for glucosylceramide biosynthesis: glucosylceramide synthase [13–15]. A third

potential clinical strategy that received much attention in recent years is called

chemical (or pharmacological) chaperone therapy and this strategy aims to

enhance the activity of mutant GBA through stabilizing molecules [16]. Both for

monitoring GBA levels in healthy and Gaucher patients and for assessment of the

effect of interference in glucosylceramide metabolism, it would be advantageous

to have access to potent and selective activity-based GBA probes.

With the aim of developing such tools, a comparative study was performed

on the merits of the two scaffolds described earlier – 2-deoxy-2-fluoroglucosides

and cyclitol epoxides – as activity-based GBA probes. Figure 13.3 depicts the four

probes that were designed for this purpose: two direct probes and two probes rely-

ing on two-step bioorthogonal ligation (see for bioorthogonal chemistry in con-

junction to ABPP, Chapter 12). GBA is a member of the large family of exoglycosi-

dases, an enzyme class normally rather particular to the nature of their substrates.

At the onset of the studies, it was therefore considered unlikely that attachment

of a bulky group such as a fluorophore or a biotin would be accepted within

the enzyme active site and thus 1,2,6-deoxy-6-azido-1,2-difluoroglucoside 4, its

click-conjugated fluorescent counterpart 5, as well as the corresponding azido-

cyclophellitol and BODIPY-cyclophellitol (boron dipyrromethene difluoride)

derivatives 6 and 7, respectively, were designed [17, 18]. Comparison of the

inhibitory potency of these compounds relative to that of the knownmechanism-

based inhibitors, CBE (2) and adamantane pentyloxy deoxynojirimycin (AMP-

DNM,MZ21, 8) for both almond retaining β-exoglucosidase (ABG, theworkhorse

4

2 8

5 6 7

O O

O O

OO

N

N

NN

B

N

NF

F

N3HO

HOF

F

BODIPYHO

HOF

F

N3HO

HOOH

BODIPYHO

HOOH

HOHO

HOOH

HOHO

HOOH

BODIPY

Figure 13.3 Mechanism-based GBA inhibitors 4–8 for comparative studies.


Table 13.1 Apparent IC50 of 2, 4–8 for almond β-glucosidase and glucocerebrosidase.

Compound Almond 𝛃-glucosidase Glucocerebrosidase

IC50 (𝛍M) IC50 (𝛍M)

2 461 9.49

4 > 10 000 1 665

5 > 1 000 785

6 27 0.120

7 56.5 0.0012

8 — 0.2

retaining β-glucosidase in the field) and GBA yielded a rather surprising result

(Table 13.1). The ABG inhibitory potency of the small set of compounds was as

expected. The cyclitol epoxides outperform the 2-deoxy-2-fluoroglycosides, with

the close glucose mimic cyclophellitol as the most potent inhibitor, and partial

to complete loss of inhibition was observed for the C6-modified compounds.

In contrast, by far the most potent GBA inhibitor proved to be cyclophellitol

derivative 7 equipped with a bulky fluorescent group at C6 [17, 18].

This superior inhibition becomes also evident in a comparative direct and

two-step bioorthogonal ABPP experiment on the four compounds. Figure 13.4a

depicts a general strategy for direct and two-step labeling on cells and cell extracts,

whereas Figure 13.4b gives a representative image of the potency and specificity

of the various ABPs on GBA. Labeling of GBA with BODIPY-cyclophellitol 7 is

very clean both in vitro and in situ, much more so than is the case in two-step

bioorthogonal labeling using copper(I)-catalyzed click reaction conditions start-

ing with azidocyclophellitol 6. No GBA-specific labeling was achieved with either

direct or two-step probes 4 and 5 based on the 2-deoxy-2-fluoroglucoside scaffold

[17]. The latter result is perhaps not so surprising as 2-deoxy-2-fluoroglucosides

are rather poor glucosidase inhibitors, likely because OH-2 of the corresponding

substrates is an important structural feature in binding to the enzyme active

site. Another intrinsic feature of 2-deoxy-2-fluoroglucosides that sets these

apart from cyclitol epoxides is their tempered reactivity as a result of the

electron-withdrawing fluorine at C2 (i.e., the same effect as that in stabilizing

the enzyme-glycoside adduct; see Figure 13.1). To offset this disadvantage,

good anomeric leaving groups (fluoride, dinitrophenyl) are often employed.

Figure 13.4c depicts a few structures that were employed to further look into the

labeling activity of this class of compounds [19]. Anomeric imidate 9 proved by

far the most potent of these series. Moreover, 1,2-difluoroderivative 5 labeled

mutant GBA in which the acid–base residue (Glu235) is mutated for Gln equally

well, whereas imidate 9 proved inactive toward this mutant. Arguably, imidate

9 is therefore a more “true” ABP that truly recruits the glucosidase active site

residues. At the same time, BODIPY-cyclophellitol 7 out-competes imidate 9 by

several orders of magnitude and cyclophellitol therefore appears the superior

scaffold for retaining glycosidase ABP design.


a

bN3

N3

C

a b b + c

SDS-PAGE

analysis

N

N N

In vitro

In situ

4 5 6 7

4 5 6 7

GBA1

GBA1

BODIPY

BODIPY

BODIPY BODIPYHO

HOHO

HO HOHOHO HO

O

OO

S

NF

F B

BODIPY

N NN

NF

OO O

O

F

S

O

OPhOPh

P

10 119

12

F FNPh

CF3

(a) (b)

(c)

Figure 13.4 Direct and two-step bioorthog-

onal labeling of GBA in cells and cell extracts.

(a) General workflow. (b) BODIPY-cyclophelli-

tol 7 is the most effective in vitro and in

situ probe. (c) Tuning of the leaving group

on 2-deoxy-2-fluoroglucosides yields com-

paratively more potent GBA probe.

13.3.2

Cyclophellitol Aziridine Is a Broad-Spectrum Activity-Based Retaining 𝛃-ExoglucosidaseProbe

The mammalian genome contains at least four retaining β-exoglucosidase genes.Next to GBA, these are the nonlysosomal retaining beta glucosidase (GBA2),

another cytosolic glucosidase termed GBA3 and lactase-phlorizin hydrolase – an

−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−→Figure 13.5 Synthesis of broad-spectrum

retaining 𝛽-exoglucoside probe cyclophelli-

tol aziridine 13 and 15. (a) (i) CCl3CN, DBU

(1,8-diazabicyclo[5.4.0]undec-7-ene), CH2Cl2,

0 ∘C, 2 h, (ii) followed by the addition of H2O,

NaHCO3, I2, 18 h, (b) (i) 37% HCl, MeOH, 3.5 h,

(ii) 37% HCl, dioxane, 60 ∘C, 1 h, (iii) NaHCO3,

MeOH, 4 days (over five steps 60%), (c) (i)

Li (s), NH3, THF (tetrahydrofurane), −60 ∘C,30min, (ii) EEDQ (ethoxycarbonyl-ethoxy-

dihydroquinoline), hept-6-ynoic acid, DMF

(N,N-dimethylformamide), 0 ∘C, 1 h, 20%, (d)

bodipy-azide, sodium ascorbate, CuSO4, DMF,

1 h, 45%, (e) EEDQ, 7-azido-octanoic acid, DMF,

0 ∘C, 1 h, and (f ) biotin-Ahx-alkyne, sodium

ascorbate, CuSO4, DMF, 1 h, 17%.


HO

HO

OBn

OBnHO

OBn

OBn

NO

I

HO

OBn

OBn

HO

NH

HO

OH

OH

HO

N

O

N

BN

F

F

14 16

13

HO

OH

OH

HO

N

O

HN S

NHHN

O

H

H

O

NH

O

3

c

d

e

f

a b

OH

OH

HO

OHN

O

N

BN

F

F

N N

N

OH

OH

HO

OHN

O

5N

NN HN N

HO

S

HN NH

O

HH

O

5

15

17

18

CCl3

N3

N3


intestinal, dual-activity glycosidase containing both a β-glucosidase activity

and a β-galactosidase activity [20]. Of these, only GBA and, to a lesser extent,

lactase-PGH are targeted by BODIPY-cyclophellitol 7, presumably because

the bulky substituent at C6 is not accepted by the other enzymes. Indeed, and

as stated before, one would expect exoglycosidases to be rather particular to

the substitution pattern and configuration of the glycomimetic emulating the

corresponding natural substrate. In contrast, exoglycosidases are often much less

selective toward the aglycon – the anomeric leaving group – as is evident both

from the range of natural substrates and artificial substrates (including fluoro-

genic substrates often used for glycosidase kinetics studies). Figure 13.5 depicts

broad-spectrum retaining β-exoglucosidase probe 13, the design of which is

based on the thought that pointing the bulky reporter group toward the direction

normally occupied by the substrate aglycon would result in a mechanism-based

inhibitor accepted by all the enzymes mentioned [20].

A key aspect to ABPP studies is, next to the design of an ABP, obviously

also its synthesis. This is often not an easy task and one complicating factor

is that ABPs are intrinsically reactive. Their reactivity needs to be balanced

such that they are – if only just – stable under physiological conditions, yet

react efficiently with their target enzyme(s). The synthesis and purification

of cyclophellitol aziridine 13 is depicted in Figure 13.5a. Partially protected

cyclohexenol 14 is an advanced intermediate in the synthesis of the natu-

ral product, cyclophellitol, as reported by Madsen and coworkers [21]. This

compound proved an ideal intermediate both in the synthesis of epoxides

6 and 7 and aziridines 13 and 15 (Figure 13.5). Installation of the acetimi-

date at O6 is followed by iodocyclization, acidic hydrolysis of the resulting

iminal and base-induced cyclization to give aziridine 16 in a complete stere-

ospecific manner. Global deprotection and aziridine acylation is followed by

copper(I) catalyzed [2+3] azide-alkyne cycloaddition to biotin-alkyne 17 or

BODIPY-azide 18. Purification of the resulting compounds 13 and 15 has

to be conducted with care, as they are both acid- and base labile. HPLC

(high-performance liquid chromatography) using neutral conditions (sol-

vent H2O/ACN (acetonitrile)) followed by lyophilization afforded the cyclitol

aziridines.

Table 13.2 presents a head-to-head comparison of GBA-specific probe 7 and

aziridine 13 [20]. As expected, aziridine 13 labels all four murine retaining β-exoglucosidases depending on their expression in various tissues. As is the case

with epoxide 7, aziridine 13 is both cell permeable and tissue permeable and both

probes are therefore amenable for in vivo labeling experiments.The question why

GBA, but not the other enzymes, accepts (and, in fact, prefers) a bulky substituent

at C6 remains unanswered. However, both probes react with a considerable num-

ber of bacterial glucosidases, some of which appear to have evolved from endoglu-

cosidases and it might well be that GBA is evolutionary related to these bacterial

enzymes [20].

13.4 Biological Research/Evaluation 201

Table 13.2 Head-to head comparison of epoxide 7 and ariridine 13.

Glucosidases Epoxide 7 Aziridine 13

GBA1 + +GBA2 − +GBA3 − +LPH − +Bacterial + +

13.4

Biological Research/Evaluation

ABPs, in general, find various uses in biology research. They can be used to dis-

cover new enzymatic species and their active site residues (comparative ABPP)

and in the evaluation of the potency and selectivity of putative inhibitors aimed at

one of the enzymes targeted by the ABP (competitive ABPP). Depending on the

bioavailability of the probe, these studies can be conducted in an in vitro, an in

situ, or an in vivo research setting. Such studies are common practice with serine

hydrolase, cysteine protease, and threonine protease probes (see also Chapter 12),

yet only start to emerge in the field of glycobiology – this for the obvious reason

that suitable glycosidase probes were until recently not available. In the following

two sections, two examples of biochemical and biological studies are described

to highlight the potential of activity-based glycosidase probes in chemical biology

research.

13.4.1

In situMonitoring of Active-Site-Directed GBA Chemical/Pharmacological Chaperones

Chemical or pharmacological chaperones form a conceptually new approach to

treat inherited diseases that are characterized by point mutations in a hydrolytic

enzyme that causes its partial dysfunctioning. Gaucher disease is caused by point

mutations in GBA that lead to lower enzyme activity in total, and this lowered

activity appears to be caused by a comparatively lower number of GBA copies

that reach the lysosome, rather than a lower activity of an individual GBA pro-

tein. Indeed, probing tissues from different Gaucher type patients with epoxide

7 (Figure 13.6a) shows GBA labeling in varying intensities, corresponding to the

severity of the disease (or the impact of the nature of the point mutation) [18].

This, while the intrinsic reactivity of the GBA mutants toward probe 7, is largely

invariable and therefore the partial loss in lysosomal activity is thought to rely on

the partial impairment of ER (endoplasmic reticulum) folding of the mutants.

Chemical/pharmacological chaperone strategies aim at correcting this

impaired folding through stabilization of the enzyme in its proper fold, which

can be achieved by inhibition of the enzyme active site.The caveat of this strategy


GBA

Rec

NCI

L444

P

N37

0S

Wild

-type

In vitro In situ

(Isofagomine)

Rela

tive

activity (

%)

Rela

tive

labelin

g (

%)

(Isofagomine)

Culture withdifferent

Scrape cellsand lysisIsofagamine

Epoxide 7

Determine GBA activityusing fluorogenic subtrateassay

(a) (c)

(b)

Figure 13.6 In situ monitoring of GBA

activity with epoxide 7. (a) Wild-type

versus mutant GBA from healthy and

Gaucher tissue. (b) Workflow for in vitro

and in situ GBA activity profiling in the

presence of chemical chaperones. (c) In

vitro and in situ effect of the pharmaco-

logical chaperone, isofagomine, on GBA

activity.

is that in this way, a larger number of enzymes may traverse to the lysosome

but these will be accompanied by their active site inhibitor. Thus, an increase in

activity within lysosomes may not be the actual result. BODIPY-cyclophellitol

7 allows for the first time to probe intracellular GBA activity directly, all or not

in the presence of a chemical chaperone. Isofagomine is the archetypal GBA

chemical chaperone studied broadly in the field, yet almost exclusively in a setting

in which, after cell culture treatment with this compound or derivatives thereof,

the sample is lysed and after which GBA activity is monitored using a fluorogenic

substrate assay (Figure 13.6b). For various reasons (dilution of the chemical

chaperone being the most obvious one), such a research setting may not reflect

the intracellular situation. As is depicted in Figure 13.6c, this appears indeed true.

In vitro measurement reveals a marked increase in activity of mutant (N370S)

GBA activity in the presence of isofagomine, whereas in situ measurement

using BODIPY cyclophellitol 7 as the readout shows a comparatively much less

pronounced activity increase. It should be noted that these assays are rather

complicated and that care has to be taken in the interpretation of the results.

13.5 Conclusions 203

At the same time, this observation should serve as a warning to the field: for

an enzyme active-site-directed chemical chaperone to be effective, it should

bind within the ER, there stabilize the enzyme in its proper fold, and once in the

endo-lysosomal compartments dissociate to become an inactive bystander. It may

not be so easy to reach this result using iminosugars, intrinsically basic by nature

and therefore prone to be trapped in acidic milieu. Interestingly, the Withers

group recently proposed the use of 2-deoxy-2-fluoroglycosides as potentially

useful alternative chemical chaperones based on their mechanism-based binding

followed by slow but sure release to produce 2-deoxy-2-fluoroglucose as such an

inert molecule [22].

13.4.2

Mapping of Human Retaining 𝛃-Glucosidase Active Site Residues

An intrinsic nature of ABPs is their covalent attachment to enzyme active

site nucleophiles. In case the nature of these is unknown, they can, in fact, be

unearthed using ABPs following the workflow as depicted in Figure 13.7a. Such

studies can be executed on recombinant purified enzymes and therefore both

2-deoxy-2-fluoroglucosides and cyclitol epoxides/aziridines can be used for this

purpose. Figure 13.7b–d provides a representative example. In contrast to GBA

and GBA3, the active site acid base and nucleophile of GBA2 were unknown.

Moreover, at least six aspartate/glutamate residues appeared suitable candidate

nucleophiles. Figure 13.7e–g depicts the GBA2 active site as determined using

the flow of experiments as outlined in Figure 13.7a [23].

13.5

Conclusions

In conclusion, rational design has resulted in the development of a panel of active

and selective activity-based retaining β-exoglucosidase probes. Cyclophellitol is anatural product and it is therefore fair to state that, as before (see the epoxomicin-

based proteasome probes in Chapter 12), nature has paved the way for these stud-

ies. The first-generation probes, represented by epoxide 7, provided a rather sur-

prising result: an active and highly selective probe for the Gaucher enzyme, GBA.

Arguably, this design principle – the fluorophore or biotin grafted at C6 – will not

meet with success when applied to other retaining glycosidases. The aziridine-

based scaffold, in contrast, holdsmore promise, and as one can learn fromCazype-

dia [24], there are numerous retaining exoglycosidases that follow the general

Koshland mechanism and that are, in principle, amenable to ABPP using cyclitol

aziridines emulating in configuration and substitution pattern the corresponding

substrate glycosides.

Another intriguing feature of the epoxide and aziridine probes is the highly

potent activity they display. Next to offering a suitable electrophile (epoxides, acy-

laziridine) to the general acid/base, the (putative) half-chair conformation they


Selection of (putative) acid/base, nucleophile residues

Site-directed mutagenesis

In vitro labeling with activity-based probes (ABPs) Azide-mediated activity rescue

Acid/base

β-Epoxide 7

β-Aziridine 13

Nucleophile

No No

NoYes

Acid/base Nucleophile

− Azide

+ Azide

No No

Yes Yes

COOH

Wild-type

Mock

Mock

E527G

Wild-type

Other alterations

D677G

GBA

GBA2

E34

0Q

E34

0G

E23

5G

E23

5G

Moc

KW

ild-typ

e G

BA

E52

7GD

659G

D66

3GE

667G

E67

3GD

677G

Moc

KW

ild-typ

e G

BA

2

β-aziridine 13

β-epoxide 7

α-myc

β-Aziridine 13

α-myc

[N3−]

[N3−]

No acid/base (E235G)

Su

bstr

ate

hyd

roly

sis

Su

bstr

ate

hyd

roly

sis

No nucleophile (E340G)

GBA:

GBA2:

H2N

H2N

E527

D659D663 E667

D673

D677

COOH

E235 E340Acid/base Nucleophile

(a)

(b) (c) (d)

(e) (f) (g)

Figure 13.7 (a–g) Retaining β-exoglucosidase active site mapping using activity-based

probes. (a) General workflow. (b) GBA2 active site as mapped by using probe 7 and 13.

O

RO

H

OO

O

H

OO

OHOHHO

HO

OHOHHO

HO

O

H

OO

OHOH

HOHO

O NR

O+

O−O− O−

Figure 13.8 Cyclitol epoxides and aziridine may feature ideal conformational behavior for

retaining β-exoglucosidases inhibition.

References 205

adopt resembles that of the developing oxocarbenium ion that is the result of

aglycon protonation andmay be at the basis of this activity (Figure 13.8). Not only

are they highly potent, they also appear to react almost instantaneously, further

suggesting that they fit exceedingly well within the active site. All this bodes well

for the future development of ABPs aimed at other retaining glycosidases and per-

haps – given that there are literature speculations on the covalent intermediacy of

some glycosyl transferases as well – for glycosyl transferases.

References

1. Wennekes, T., van den Berg, R.J., 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. Davies, G.J. and Henrissat, B. (1995)

Structures and mechanisms of

glycosyl hydrolases. Structure, 3,

853–859.

3. Koshland, D.E. (1953) Stereochem-

istry and the mechanism of enzymatic

reactions. Biol. Rev. Camb. Philos. Soc.,

28, 416–436.

4. Vocadlo, D.J. and Davies, G.J. (2008)

Mechanistic insights into glycosidase

chemistry. Curr. Opin. Chem. Biol., 12,

539–555.

5. Gandy, M.N., Debowski, A.W., and

Stubbs, K.A. (2011) A general method

for affinity-based proteomic profiling of

exo-α-glycosidases. Chem. Commun., 47,

5037–5039.

6. Withers, S.G., Street, I., Bird, P.,

and Dolphin, D. (1987) 2-deoxy-

2-fluoroglucosides- a novel class of

mechanism-based glucosidase inhibitors.

J. Am. Chem. Soc., 109, 7530–7531.

7. Atsumi, S., Umezawa, K., Iinuma, H.,

Naganawa, H., Nakamura, H.M., Iitaka,

Y., and Takeuchi, T. (1990) Production,

isolation and structure determina-

tion of a novel β-glucosidase inhibitor,

cyclophellitol, from Phellinus sp. J.

Antibiot., 43, 49–53.

8. Vocadlo, D.J., Davies, G.J., Laine, R.,

and Withers, S.G. (2001) Catalysis by

hen egg-white lysozyme proceeds via

a covalent intermediate. Nature, 412,

835–838.

9. Legler, G. (1968) Investigations on

the mechanism of action of glycoside-

splitting enzymes. 3. Labelling of the

active centre of a β-glucosidase from

aspergillus wenti with (14C) conduritol

B epoxide. Hoppe-Seyler’s Z. Physiol.

Chem., 349, 767–774.

10. Atsumi, S., Iinuma, H., Nosaka, C., and

Umezawa, K. (1990) Biological activ-

ities of cyclophellitol. J. Antibiot., 43,

1579–1585.

11. Beutler, E. and Grabowski, G.A. (2001)

in The Metabolic and Molecular Bases

of Inherited Disease, 8th edn (eds

C.R. Scriver, W.S. Sly, and D. Valle),

McGraw-Hill, New York, pp. 3653–3668.

12. Aerts, J.M.F.G., Hollak, C., Boot, R.,

and Groener, A. (2003) Biochemistry

of glycosphingolipid storage disorders:

implications for therapeutic intervention.

Philos. Trans. R. Soc. Lond. B: Biol. Sci.,

368, 905–914.

13. Barton, N.W., Furbish, F.S., Murray,

G.J., Garfield, M., and Brady, R.O.

(1990) Therapeutic response to intra-

venous infusions of glucocerebrosidase

in patient with gaucher disease. Proc.

Natl. Acad. Sci. U.S.A., 87, 1913–1916.

14. Cox, T., Lachmann, R., Hollak, C., Aerts,

J.M.F.G., van Weely, S., Hrebicek, M.,

Platt, F., Butters, T., Dwek, R., Moyses,

C., Gow, I., Elstein, D., and Zimran, A.

(2000) Novel oral treatment of Gaucher’s

disease with N-butyldeoxynojirimycin

(OGT 918) to decrease substrate biosyn-

thesis. Lancet, 355, 1481–1485.

15. McEachern, K.A., Fung, J., Komarnitsky,

S., Siegel, C.S., Chuang, W.L., Hutto,

E., Shayman, J.A., Grabowski, G.A.,

Aerts, J.M., Cheng, S.H., Copeland, D.P.,

and Marshall, J. (2007) A specific and


potent inhibitor of glucosylceramide

synthase for substrate inhibition therapy

of Gaucher disease. Mol. Genet. Metab.,

91, 259–267.

16. Benito, J.M., García Fernández, J.M., and

Ortiz Mellet, C. (2011) Pharmacological

chaperone therapy for Gaucher disease:

a patent review. Expert Opin. Ther. Pat.,

21, 885–903.

17. Witte, M.D., Walvoort, M.T.C., Li, K.-Y.,

Kallemeijn, W.W., Donker-Koopman,

W.E., Boot, R.G., Aerts, J.M.F.G.,

Codee, J.D.C., van der Marel, G.A., and

Overkleeft, H.S. (2011) Activity-based

profiling of retaining b-glucosidases: a

comparative study. ChemBioChem, 12,

1263–1269.

18. 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.

19. Walvoort, M.T.C., Kallemeijn, W.W.,

Willems, L.I., Witte, M.D., Aerts,

J.M.F.G., van der Marel, G.A., Codee,

J.D.C., and Overkleeft, H.S. (2012)

Tuning the leaving group in 2-

deoxy-2-fluoroglucosides results in

improved activity-based retaining b-

glucosidase probes. Chem. Commun., 48,

10386–10388.

20. Kallemeijn, W.W., Li, K.-Y., Witte,

M.D., Marques, A.R.A., Atemn, J.,

Scheij, S., Jiang, J., Willems, L.I.,

Voorn-Brouwer, T.M., van Roomen,

C.P.A.A., Ottenhoff, R., Boot, R.G., van

den Elsts, H., Walvoort, M.T.C., Florea,

B.I., Codee, J.D., van der Marel, G.A.,

Aerts, J.M.F.G., and Overkleeft, H.S.

(2012) Novel activity-based probes for

broad-spectrum profiling of retaining

β-exoglucosidases in situ and in vivo.

Angew. Chem. Int. Ed., 51, 12529–12533.

21. Hansen, F.G., Bundgaard, E., and

Madsen, R. (2005) A short synthesis

of (+) cyclophellitol. J. Org. Chem., 70,

10139–10142.

22. Rempel, B.P., Tropak, M.B., Mahuran,

D.J., and Withers, S.G. (2011) Tailor-

ing the specificity and reactivity of a

mechanism-based inactivator of gluco-

cerebrosidase for potential therapeutic

applications. Angew. Chem. Int. Ed., 50,

10381–10383.

23. Kallemeijn et al., to be published in

detail.

24. CAZypedia (2007) http://www.

cazypedia.org (accessed 29 December

2014).

http://www

Concepts and Case Studies in Chemical Biology (Waldmann/Concepts and Case Studies in Chemical...

Documents

Transcript of Concepts and Case Studies in Chemical Biology (Waldmann/Concepts and Case Studies in Chemical...