DOI:10.1002/cbic.201600561 Full Papers

ASpecific Activity-Based Probe to Monitor Family GH59 , the Deficient in Krabbe Disease AndrØ R. A. Marques+,[a, f] LianneI.Willems+,[b, g] DanielaHerreraMoro,[a] BogdanI.Florea,[b] SaskiaScheij,[a] RoelofOttenhoff,[a] Cindy P. A. A. van Roomen,[a] Marri Verhoek,[c] JessicaK.Nelson,[a] WouterW.Kallemeijn,[c] AnnaBiela-Banas,[d] Olivier R. Martin,[d] M. BegoÇa Cachón-Gonzµlez,[e] Nee Na Kim,[e] Timothy M. Cox,[e] Rolf G. Boot,[c] Herman S. Overkleeft,*[b] and Johannes M. F. G. Aerts*[a, c]

Galactosylceramidase (GALC) is the lysosomal b-galactosidase ring specificity for GALC,equipped with aBODIPY fluorophore responsible for the hydrolysis of galactosylceramide. Inherited at C6 that allows visualization of active enzyme in cellsand tis- deficiency in GALC causes Krabbedisease, adevastating neuro- sues. Detection of residual GALC in patient fibroblasts holds logical disordercharacterizedbyaccumulation of galactosylcer- great promise forlaboratory diagnosis of Krabbedisease.We amide and its deacylated counterpart, the toxic sphingoid base furtherdescribe aprocedure for in situ imaging of active GALC galactosylsphingosine (psychosine). We report the design and in murinebrain by intra-cerebroventricularinfusionofthe ABP. application of afluorescently tagged activity-based probe In conclusion, this GALC-specific ABP should find broad appli- (ABP) for the sensitive and specific labeling of active GALC cations in diagnosis, drug development, and evaluation of molecules from various species. The probe consists of a b-gal- therapyfor Krabbe disease. actopyranose-configured cyclophellitol-epoxide core, confer-

Introduction

Glycoside family 59 (GH59) humangalactosylcerami- the b-anomericconfiguration of the released galactopyrano- dase (GALC, galactocerebrosidase) is an 80 kDa protein respon- side (Scheme 1A). The two carboxylic acid residuesinthe sible for the lysosomalturnover of galactosylceramide andgal- that function as anucleophile and ageneral acid/ actosylsphingosine. Newly synthesized GALC contains at least base have been identified as the glutamic acidresidues E258 four N-linked glycosylation sites, which are responsible for lyso- and E182, respectively.[1,3] somal trafficking through the mannose-6-phosphate receptor.[1] Deficiencies in GALC are at the basis of the autosomal reces- After enteringthe , the enzyme is cleaved into 30 sive lysosomal storage disorder Krabbe disease, also termed and 50 kDa subunits without effect on enzymatic activity.[2] globoid cell leukodystrophy.More than 70 mutationsinthe Crystalstudies indicate that no dissociation of these subunits encoding GALC have been implicated in the develop- occurs.[3] Substrate hydrolysis by GALC occurs through aKosh- ment of this disease.[1] The main pathological consequences land double displacementmechanism with overall retention of are found in the peripheraland central nervous systems. The

[a] Dr.A.R.A. Marques,+ D. HerreraMoro, S. Scheij, R. Ottenhoff, [e] Dr.M.B.Cachón-Gonzµlez, N. N.Kim, Prof. Dr.T.M.Cox C. P. A. A.van Roomen, Dr.J.K.Nelson, Prof. Dr.J.M.F.G.Aerts DepartmentofMedicine,University of Cambridge Department of Biochemistry,Academic MedicalCenter Addenbrooke’s Hospital University of Amsterdam Hills Road, CambridgeCB2 2QQ (UK) Meibergdreef 15, 1105 AZ Amsterdam (The Netherlands) [f] Dr.A.R.A.Marques+ [b] Dr.L.I.Willems,+ Dr.B.I.Florea, Prof. Dr.H.S.Overkleeft Present address:Institute of Biochemistry Department of Bio-organic Synthesis Christian-Albrechts-University of Kiel Leiden Institute of Chemistry, Leiden University Otto-Hahn-Platz9,24098 Kiel (Germany) Einsteeinweg55, 2300 RA Leiden (The Netherlands) [g] Dr.L.I.Willems+ E-mail:[email protected] Present address:Department of Chemistry, Simon FraserUniversity [c] M. Verhoek, Dr.W.W.Kallemeijn, Dr.R.G.Boot, Prof.Dr. J. M. F. G. Aerts 8888 University Drive, BurnabyV5A 1S6, BC (Canada) Department of Biochemistry,LeidenInstitute of Chemistry,Leiden University [+] These authorscontributed equally to this work. Einsteinweg 55, 2300 RA Leiden (The Netherlands) Supportinginformation for this article can be found under: E-mail:[email protected] http://dx.doi.org/10.1002/cbic.201600561. [d] Dr.A.Biela-Banas, Prof. Dr.O.R.Martin Institute of Organic and Analytical Chemistry,UniversitØ D’OrlØans Rue de Chartres, B.P. 6759, 45100OrlØans(France)

ChemBioChem 2017, 18,402 –412 402  2017 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Papers

Scheme1.Galactocerebrosidase and structuresofinhibitors and probes 1–7.A)Mechanism of substrate hydrolysisbyGALC. B) Proposed mechanism of GALC binding by compounds 1–4.C)Structures of novel retaining b-galactosidase inhibitor 1 and ABPs 2–4. b-Glucopyranosyl-configured compound 5 targets glu- cocerebrosidase. ABB166 (6)[31] and galactosylsphingosine (psychosine, 7)are competitive GALC inhibitors. mechanism behind this neuropathologyhas not been fully elu- fibroblasts, which is usually 0–5%ofnormallevels.However, cidated.[1,2] Reduced activity of GALC resultsinreduced catabo- the amount of residual activity is neither directly correlated to lism of galactosphingolipids,including galactosylceramide and the clinical symptomsnor to the course of the disease. Carriers galactosylsphingosine (psychosine). Galactosylceramide is the might have as little as 10–20%ofnormalGALC activity with- main lipid component of myelin, the protective sheath around out being affected. The only availabletreatment for Krabbe neuron axons that is essential for correctfunctioning of the disease involves symptomatic treatment and physical therapy. nervoussystem. Accumulation of the toxic metabolite galacto- Clinicaltrials with hematopoietic stem cell transplantation, sylsphingosine eventually leads to apoptosis of myelin-forming which aim to restoreGALC activity in the central nervous cells and consequentdemyelination and neurodegenera- system and therebypreventfurtherdemyelination, hold prom- tion.[4,5] Infantile Krabbe disease is usually diagnosed before ise in slowing the course of juvenile Krabbedisease when di- one year of age and lethal beforethe age of two years. Early agnosed at an early stage.[8,9] symptomsinclude limb stiffness, developmental delay,and Further investigations of GALC and its involvement in severe irritability.When diagnosed in adolescents or adults, Krabbedisease would benefit from the availability of an activi- other symptoms can be observed, such as seizures,feeding dif- ty-based probe (ABP) that specifically targets this enzyme. Nu- ficulties,slowing of mental and motor development, muscle merous fluorescent b-galactosidaseprobeshave been reported weakness, spasticity,deafness, and blindness. The onset and in literature. However, most of these are reversibleand there- severity of symptoms, as well as the course of the disease in fore cannot be used in, for example, gel-based assays. Exam- adult Krabbe patients, is highly variable, even in patients carry- ples include substrates consisting of a b-galactose moiety with ing the same mutation.[6,7] Diagnosis can be confirmed by aluminescentorfluorogenic tagattached to the aglycon posi- measuring residual GALC activity in leukocytes or cultured skin tion that is released after cleavage by the enzyme,[10–16] as well

ChemBioChem 2017, 18,402 –412 www.chembiochem.org 403  2017 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Papers as fluorescently tagged competitive inhibitors.[17,18] In addition, Results afew ABPs that enable irreversible mechanism-basedlabeling of retaining b- have been reported. These in- Labelingand inhibition of recombinant galactocerebrosi- clude suicide substrates bearing alatent quinone methide pre- dase cursor as the aglycon, to whichafluorescent or fluorogenic tag is attached,[19–22] and 2-fluorogalactoside inhibitors in which First we evaluated the ability of compounds 1–4 to inhibit re- the hydroxy group at C6 is substituted with an azide, which combinant GALC by measuring residual enzyme activity using enablestwo-step labelingbyStaudinger–Bertozzi ligation.[23] the fluorogenicsubstrate 4-methylumbelliferyl b-d-galactopyr- To date, noneofthese probes has been used for the labeling anoside(4-MU b-Gal) after 30 min of pre-incubationwith vary- of human retaining b-galactosidases. ing concentrations of the probes. Plots of residual activity Our work on activity-based retainingglycosidaseprobes against inhibitorconcentration reveal acleardose-dependent used the naturalproduct and retaining b-glucosidaseinhibitor, inhibition of GALC by all probes(Figure 1A). The apparent IC50 cyclophellitol, as starting point.[24] Substituting the epoxidefor values calculated from these curvesare shown in Table1.The an aziridine and grafting areportergroup onto the aziridine nitrogen yielded ABPs that were broadly selectivefor various members within agiven class of retaining glycosidases. Glyco- Table 1. Inhibition of recombinant galactocerebrosidase activity. sidase family selectivity was dictated by the configurationof Compound Apparent ki/Ki Inhibition [%] the cyclophellitol aziridine derivative,anapproachthat was [a] 1 1 IC50 [mm] [mmÀ minÀ ]30min 6h shown valid for GH1-retaining b-,GH79-retaining epoxide 1 0.038 12.06 98 (1 mm)100 (1 mm) a-galactosidases and, most recently, for GH29-retaining a-fuco- azido-epoxide 2 70 n.a. 62 (100 mm)96(100 mm) [25–27] sidases. In our first forays into activity-based glycosidase BODIPY-epoxide 3 2.8 0.014 97 (100 mm)100 (100 mm) profiling,however,westudied cyclophellitol derivatives modi- biotin-epoxide 4 30 n.a. 81 (100 mm)97(100 mm) [28] fied at C6 (glucopyranose numbering) with afluorophore. [a] Inhibitionofrecombinant GALC, as determined by hydrolysis of 4-MU These epoxide probesexhibited much higherselectivity than b-galafter pre-incubation for 30 minor6hwith probes 1–4.Apparent their aziridine analogues. Because of their high selectivity and IC50 values were calculatedfrom residual GALC activity as afunction of potencyfor human lysosomal (GBA)—the probe concentration (Figure 1A). Second-order rate constants were calcu- lated from progress curves over time (Figure S1). n.a.:not assessed. enzyme deficient in Gaucher patients—we now routinelyuse the fluorescent cyclophellitol derivativestomonitor GBA activi- ty in vitro, in situ, and in vivo in healthyand Gaucher models.[28] non-tagged epoxide 1 provedtobeaverypotent inhibitor of

We realized that this probe design might also hold potential GALC with an apparent IC50 value of 38 nm.Substitution of the for the development of aGALC ABP suitable for monitoring hydroxy group at the C6 position with an azide (2), however, this enzyme in the context of Krabbe disease (Scheme 1B), the resultedinadramatic loss of potencywith an almost 2000- more so as GALC and GBA have related glycosylceramide sub- fold increase of the apparent IC50 value (70 mm). We found strates. Here we describe the evaluation of b-galactopyranose- probe 2 to reach full inhibition after aprolonged incubation configured epoxides 1–4 as inhibitors and ABPs for GALC time. Interestingly,the inhibitory potency was partially restored (Scheme 1C). The cyclophellitol core of theseprobes, synthe- by incorporation of aBODIPY dye (3), but not abiotin tag (4), sized as reported,[24,29,30] was designed to mimic the substrate’s at the same position. The second order rate constantsfor in- terminal galactosyl moiety and bind covalently to the target hibition (ki/Ki)ofcompounds 1 and 3 were in agreement with enzyme through nucleophilic attack of the catalytic residue in this observation (see Table 1and Figure S1 in the Supporting the active site on the b-configured electrophilic epoxide Information). Hence, it appears that the hydrophobic fluoro- moiety (Scheme 1B). The non-tagged inhibitor 1 was included phore leads to enhanced bindingofthe probe to its target in our studies as agalactose-configured isomer of the known enzymeand might be better tolerated in the active site of the retaining b-glucosidase inhibitor cyclophellitol.[24,28] For poten- enzymethan asmall polar azide moiety.Asimilar result has tial use in two-step activity based profiling studies, in ABP 2 been found previously forthe inhibition of retaining b-glucosi- the primary hydroxy group is substituted with an azide that dases by analogousC6-modified probes, although the benefi- can be used fortwo-step labeling by copper(I)-catalyzed or cial effect of the BODIPY dyewas much larger in that case.[28] copper-free strain promoted alkyne-azide[2+3] cycloaddition On the contrary,previous findings suggest that neither of the chemistry or Staudinger–Bertozzi ligation. In addition, this a-configured isomers of the C6-functionalized epoxide probes probe could serve as acontrol probe and an inhibitor.ABPs 3 2–4 appears to inhibit retaining a-galactosidases,[26] indicating and 4 wereobtained by functionalization of 2 with aBODIPY that this phenomenon is dependent on the specific active site fluorophore and abiotin tag, respectively. features and substrate tolerance of each individualglycosidase. To demonstrate the irreversibility of GALC inhibition by ABP 3 we preincubated GALC with ABP 3 fordifferent time periods. Next, residual enzymatic activity in the samples was deter- mined and aliquots of the same samples were subjected to gel electrophoresis to quantify ABP-labeled GALC (Figure S2).

ChemBioChem 2017, 18,402 –412 www.chembiochem.org 404  2017 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Papers

Figure 1. Labeling and inhibition of recombinant GALC. A) Inhibition of recombinant GALC activity.Recombinant GALC was treated for 30 min with inhibitors 1–4,after which residual activity was determined from hydrolysis of 4-MU b-gal. B) Labeling with BODIPY-epoxide 3 (0.01–10 mm)for 1h,where indicated enzymewas denatured prior to labeling by boiling at 100 8Cfor 3min in assay buffer(+)orinbuffer with 1% SDS (+SDS). C) Labeling with 0.2 mm 3 after pre-incubation with inhibitors 1 (0.2 mm), 2 (10 mm)or4 (10 mm)for 1h.D)Labeling with 0.1–10 mm of biotin-epoxide 4 for 1h;where indicated enzyme was denatured priortolabeling by boiling at 100 8Cfor 3min. E) Recombinant GALC was labeled with 0.5 mm 3 for 1hin buffers of varyingpH. Samples were an- alyzed by 10%SDS-PAGE with fluorescent readout, followed by CBB staining (B, C, E) or streptavidin western blotting (D). M: protein marker.F)Quantification of gel bands in A(*), as compared to GALC activity on 4-Mu b-Gal (~)atdifferent pH values. G) Selective labeling of retaining b-galactosidases and retaining b-glucosidases with differently configuredepoxide ABPs. Recombinant GALC and (GBA) were labeled with 1 mm of ABP 3 or ABP 5 for 1h,and labeledproteins wereresolved by 7.5%SDS-PAGE with fluorescent readout.

Enzymeactivity was irreversibly lost during the pre-incubation Having shownthat compounds 1–4 were able to irreversibly in time- and concentration-dependentmanner.Loss of activity inhibit GALC, we next assessed the visualizationofthe re- of GALC by pre-incubation with ABP 3 correlated with its fluo- combinant enzyme on gel by using fluorescently labeled ABP rescent labeling. To further demonstrate irreversibility of label- 3. ing, we incubated GALC with ABP 3 for one hour and separat- Exposure to BODIPY-epoxide 3 for 1hresulted in concentra- ed enzymeand small compound by using aspin dialysis car- tion-dependentfluorescent labelingofaband around80kDa, tridge with 7kDa cutoff. No activity wasfound to be recovered corresponding to the molecular weightofnon-dissociated over time after removal of free unbound ABP 3 from the GALC (Figure 1B). The labeling was completely abolished by enzyme (Figure S3). We also incubated enzyme andABP 3 denaturation of the enzyme prior to addition of the ABP,con- briefly (2 min) before performing the same separation:inthis firming the specific binding of the probe to catalytically active case part of the separated enzyme population was still active. enzyme. In addition, asecond fluorescently labeled band of Again, the amountofenzyme activity did not increasefollow- approximately 60 kDa wasvisible at higherprobe concentra- ing the separation from free ABP 3.Analysis of labeled GALC tions. This protein most likely represents serum albumin, acom- revealed no loss of label following separation of pre-labeled ponent of the cell culture medium. Unwanted labeling of albu- enzyme from free ABP 3 (Figure S3). In order to assess the se- min by BODIPY-functionalized probescan be caused by non- lectivity of inhibitors 1–4 forretaining b-galactosidases over specific interactions of the hydrophobic dye with the protein. the related class of retaining a-galactosidases, we determined Evidence for the non-specific nature of these interactions is inhibition of recombinant a-galactosidase A(Fabrazyme) by provided by the fact that denaturation of the protein by boil- using asimilar assay with the fluorogenic substrate 4-methyl- ing in assay buffer prior to labeling with ABP 3 resulted in an umbelliferyl a-d-galactoside.Wedid not detectany inhibition even stronger fluorescentsignal, which could only be eliminat- after 30 min of pre-incubation with up to 100 mm of the ed by addition of the surfactant SDS. probes(data not shown).

ChemBioChem 2017, 18,402 –412 www.chembiochem.org 405  2017 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Papers

Pre-treatment of samples with non-taggedinhibitor 1 led to nerve of wt and heterozygous animalsbut not of Twitcher complete disappearance of the fluorescent labeling of GALC mice (Figure 2A). This band corresponds to the 50 kDa subunit by BODIPY-tagged epoxide 3,whereas the labeling of albumin of GALC that is formed after proteolytic cleavage of the was unaffected (Figure 1C). Pre-incubation with azide- and enzymeinthe .The fact that no labeling by probe 3 biotin-tagged probes 2 and 4 resulted in partial blockingof was detected in wt liver is in line with the low expression the fluorescent labeling of GALC. We also used biotinylated levels of GALC in this organ.[33] To confirm the identity of the ABP 4 to visualize GALC activity directly by streptavidin west- labeled proteins, we performed astreptavidin affinity purifica- ern blotting(Figure 1D). In agreement with the higherIC50 tion after labeling of mousekidney lysates with biotinylated value of this probe as compared to its fluorescently labeled an- ABP 4.Proteins were subjected to LC-MS/MS identification fol- alogue 3,arelatively large amount of probe is required to lowing tryptic digestion. Analysis of those proteins specifically label the enzyme. At the highest concentration tested (10 mm), enriched in samples treated with probe as compared to the asingle biotin-labeled band was clearly visible that corre- no-probecontrol revealed the specific labelingofseveralpep- sponds to catalytically active GALC. tides from the N-terminal part of the active 50 kDa GALC pro- tein by ABP 4 (Table S1). Next, we studied competitionofABP 3 labeling of GALC by pH dependence of galactocerebrosidase labeling two known competitive inhibitors of the enzyme, ABB116[31] As GALC is alysosomalenzyme, itsactivity is highest in aslight- (Scheme 1C,compound 6)and galactosylsphingosine or psy- ly acidic environment with an optimum around pH 4.3.[3] We chosine(Scheme 1C,compound 7). Labeling of GALC in examined the pH dependence of GALC labeling by BODIPY- mouse brain and kidney lysates by probe 3 wasfully competed functionalized ABP 3 by exposing the recombinant enzyme to by both inhibitors at millimolar concentrations;this confirmed the probe in buffers of varyingpH(Figure 1E). The intensity of that the ABP binds to the active site of the enzyme (Fig- fluorescent labeling of GALC is highest at pH 4–5, whereas it is ure 2B). almostcompletely abolished at pH 3and lower or pH 7and BesidesGALC, lysosomes contain another acid b-galactosi- higher.Quantification of the fluorescent gelbands revealed dase involved in degradation of various substrates like ganglio- that the amount of labeling by ABP 3 overlaps perfectly with side GM1, lactosylceramide, glycoproteins, and keratan sulfate- the enzymatic activity,asdetermined by fluorogenicsubstrate derived oligosaccharides.[34–36] Malfunctioning of this b-galacto- hydrolysis (Figure 1F). These resultsindicate that binding of sidase can cause GM1 gangliosidosis and Morquio Bsyn- the probe occurs in an activity-based manner and support the drome.[34,37] To evaluate the enzymespecificity of probe 3 and proposed bindingmechanism depicted in Scheme 1B. epoxide inhibitor 1 within the family of retaining b-galactosi- dases, we investigated their effect on the enzymatic activity of lysosomal acid b-galactosidaseand GALC. To distinguish the Selectivity of galactocerebrosidase labeling two enzymeactivities, we followed the protocol described by

Next, we set out to validate our hypothesis that selective label- Martino et al.,which employs AgNO3 as aselective competitive ing of retaining b-galactosidases and retaining b-glucosidases inhibitor of lysosomal acid b-galactosidase.[38] Pre-incubation of could be achieved with differently configured epoxide-based brain and kidney lysates with 10 mm of probe 3 resulted in ABPs. For this purpose, we used b-galactosidase ABP 3 and the asignificant decrease in GALC activity (ca. 50 %ofuntreated previously reported b-glucosidaseABP 5[28] (Scheme 1C)to values) in thesetissues,whereas b-galactosidaseactivity was label either recombinant GALC or recombinantglucocerebrosi- unaltered (Figure 2C and D). In contrast, incubation of the ly- dase, alysosomal retaining b-glucosidase. The two ABPs were sates with 0.5 mm of inhibitor 1 completely abrogated both functionalized with different BODIPY fluorophores and could enzymeactivities. therefore be visualized by using differentscanner settingsfor The human body contains one additional b-galactosidase in in-gel fluorescent readout. Although GALC was labeled exclu- the small intestine:, also knownaslactase-phlorizin hy- sively by probe 3,glucocerebrosidase was labeled by ABP 5 drolase (LPH).[39,40] This enzyme,which also has b-glucosidase but not by its stereoisomer, 3 (Figure 1G). Theabsence of activity,cleaves lactose into galactose and glucose, and defi- cross-reactivity demonstrated the selectivetargeting of each of ciency in its activity causes lactose intolerance. We transfected these by the appropriately configured ABPs. human embryonic kidney (HEK293) cells with LPH and studied To further demonstrate the specificityofprobe 3 towards labeling of the enzymebyABP 3 and ABP 5 by using concen- GALC and to show its application to label endogenous trationsofthese probesatwhichthey are known to target enzyme, we used the probe to label varioustissue extracts of LPH (Figure S4).[28] Whereas the b-glucosidaseABP 5 resulted in Twitcher (twi/twi)mice.[32] These animalsare anaturally occur- prominentlabeling of LPH,this was negligible for ABP 3,again ring modelofKrabbedisease and lack GALC protein. Wecom- indicating specificity of ABP 3 for binding to GALC and not pared the labeling of tissues of wild-type (wt), Twitcher,and LPH. heterozygous mice with ABP 3 and ABP 5.Asexpected, label- ing by b-glucosidase probe 5 occurred in all tissues of all the In situ and in vivoGALC labeling genotypes (Figure 2A). However, incubation of tissue lysates with GALC probe 3 resulted in fluorescent labeling of asingle Finally,weexamined labeling of GALC by ABP 3 in intact cells band of approximately 50 kDa in the kidney,brain, and sciatic and tissues. We notedearlier that the stereoisomer ABP 5 was

ChemBioChem 2017, 18,402 –412 www.chembiochem.org 406  2017 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Papers

Figure 2. Labeling of tissues of wild-type, GALC-deficient (Twitcher) and heterozygous mice. A) Homogenates of brain,kidney,sciatic nerve and liver of wt (Galc+/+), carrier (Galc+/ ), and Twitcher(Galc / )mice were incubated with ABP 3 (1 mm)orABP 5 (100 nm). Samples were analyzedby7.5 %SDS-PAGE À À À with fluorescentreadout, followed by CBB staining. M: protein marker.B)Competition of ABP 3 labeling ofGALC in mouse kidney and brain homogenates by ABB166(6)and psychosine (7). C) Kidney homogenates were pre-incubated with 10 mm ABP 3 and 0.5 mm compound 1 for 1h,after which the residual acid b-galactosidase and GALC enzyme activity was determined. D) Brain homogenates were pre-incubated with 10 mm ABP 3 and 0.5 mm inhibitor 1 for 1h,after which the residual acid b-galactosidase and GALC enzymeactivity was determined.Data (n= 3per group, mean SD) were analyzed by using one-way ANO Æ VA followed by the Dunnett’s multiple comparison test:*p<0.05;**p< 0.01;***p<0.001.

ChemBioChem 2017, 18,402 –412 www.chembiochem.org 407  2017 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Papers

Figure 3. In situ labeling of GALC in intact cells in cultureand in vivo in mouse brain. A) In situ labeling with ABP 3 of control HEK293 cells (top) and HEK293 cells overexpressing GALC (bottom). Left to right:nuclearDAPI staining,ABP 3 labeling, and overlay of DAPI staining and ABP 3 labeling (all in unfixed cells), and overlay of immuno-detection of lysosomal membranemarker LAMP1 and DAPI staining in fixed cells. Scale bars in third panels:10mm, scale bars in right panels: 20 mm. B) In vivo labeling with ABP 3 of wt mouse cerebellar cortexfollowing i.c.v. administration. Left to right:DAPI staining, ABP 3 labeling, anti- GalC Ab (top) or anti-LAMP1 Ab (bottom), and overlay of ABP 3 and Ab staining. Scale bar:10mm. able to penetrate cells by diffusion andefficiently labels lysoso- Krabbedisease patients andhealthy volunteers simultaneously mal glucocerebrosidase in situ.[28] Cultured HEK293 cells, with with ABP 3 and ABP 5.Labeling was preceded by an affinity and without overexpression of GALC,were exposed for 1hour purification of glycosylated proteins with concanavalin A to 5nm ABP 3,and confocal fluorescencemicroscopy was beads. Aband of approximately 50 kDa, corresponding to used to detect labeled GALC in the unfixedcells. The fluores- human GALC, was labeled by ABP 3 in fibroblasts of the con- cent labeling in control HEK293 cells was low (Figure 3A), in trol subjects and Gaucher disease patients(Figure 4) but was agreement with the virtual absence of endogenous GALC la- virtually absent in samples from individuals diagnosed with beling in HEK cells extracts (FigureS4). Lowexpression levels Krabbedisease. As expected, activeGBA labeled by ABP 5 was of GALC were furtherconfirmed by western blot (not shown). much less prominent in fibroblasts of Gaucher disease patients However,incells overexpressing GALC, aperinuclear vesicular compared to the other cell types (Figure 4). labeling,characteristic of lysosomes, was observed (Figure 3A). The perinuclear vesicular labeling pattern of lysosomes was Discussion and Conclusion confirmed by labeling of fixed cells with the lysosomal marker LAMP1 (Figure 3A,last panel). Prompted by the successful design of b-glucopyranose-config- Next, we examined ABP 3 labeling of endogenous GALC in ured epoxide-based probes that are able to label lysosomal the brain of living mice. For this purpose, we intra-cerebroven- glucocerebrosidase (E.C. 3.2.1.45)inanactivity-based manner tricularlyinfused mice for 2h with 1nm ABP 3.The animals with high selectivity and sensitivity,[28] we examined asimilar were sacrificed, and brain slices were examined by confocal approachfor the related enzymegalactocerebrosidase (E.C. fluorescencemicroscopy. Pronounced perinuclear labeling was 3.2.1.46). The b-galactopyranose-configured epoxide-based detected in cells of the cerebellar cortex,overlapping with probes 1–4 were demonstrated to inhibit recombinant and immuno-histochemical detectionofGALC and lysosomal endogenous rodentGALC covalently and irreversibly.Ofthese marker LAMP1 (Figure 3B). probes, non-tagged epoxide inhibitor 1 was the mostpotent To investigate the potential of ABP 3 for the diagnosis of inhibitor.Substitution of the hydroxy group at C6 with an Krabbedisease, we labeled fibroblasts from Gaucher and azide (2)resulted in loss of potency,whereas installment of

ChemBioChem 2017, 18,402 –412 www.chembiochem.org 408  2017 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Papers

and heterozygousmiceand is absent in the same tissues of Twitcher animals. Importantly,the reported b-galactopyranose-configured ep- oxide-based probescan be applied as diagnostic tools in mon- itoring Krabbedisease by visualizing the levels of residual GALC activity,aswas demonstrated here by the labeling of active enzymeincontrolcompared to Krabbe disease fibro- blasts. The probescould also be used in activity-based protein profiling studies, aiding in the developmentofnovel therapeu- tic strategies by facilitating the screening of potential chaper- ones interacting with the catalytic pocket of GALC. Last but not least, the probes will likely prove of great value to evaluate efficacyofexperimental therapiesinmouse modelsofKrabbe disease.

Figure 4. Labeling of Gaucher and Krabbe patient fibroblastswith ABPs 3 and 5.Fibroblasts from Gaucher and Krabbe disease patients, and healthy volunteerwere lysed and incubated overnight with concanavalin Abeads Experimental Section for enrichment of glycosylated proteins. Bead-bound GALC was labeled with ABP 3(1mm), followed by labeling of GBA with ABP 5 (0.1 mm). Samples Synthesis of b-galactopyranose-configured epoxide-based were resolved by 10%SDS-PAGE with fluorescent readout, followed by CBB probes: b-Galactopyranose-configured cyclophellitol epoxides 1–4 staining. M: protein marker. were synthesized as previously reported.[24] Enzymes: Recombinant murine GALC was expressed in HEK293 [1] aBODIPY tag at the same position (3)partiallyrestored the in- cells, as previously described. The mouse enzyme is 83 %homolo- gous to human GALC.[1] Culture medium containing the secreted hibitory potency.Ahigh-resolution crystal structure of murine [3] recombinant protein was used directly for fluorogenic substrate GALC was recently reported. Soaking of crystallized enzyme assays and labeling assays. Recombinant a-galactosidase A(Fabra- with ABP 3 should yield an explanation for the noted differen- zyme) and recombinant b-glucocerebrosidase (Cerezyme) were ob- ces in affinity of inhibitor 1 and ABPs 2–4. tained from Genzyme. BODIPY-functionalized ABP 3 and biotinylated ABP 4 enable Cells: HEK293 cells (ATCC CRL 1573) were cultured in Dulbecco’s the visualizationofcatalytically active GALC on gel. The epox- Modified Eagle Medium (DMEM) with high glucose (Gibco), supple- 1 ide probeslabel the enzymeinanactivity-based manner.The mented with 10%FBS (Bodinco) and 100 unitsmLÀ penicillin/ selectivity of the labeling by ABP 3 is remarkable.Inparticular, streptomycin (Gibco). HEK293 cells with and without stable overex- BODIPY-epoxide 3 differs only in the configuration of asingle pression of murine GALC were generated as previously described.[1] substituent from the previously reported glucopyranose con- Animals: Twitcher mice (twi/twi), anatural model of Krabbe dis- figured stereoisomer 5,yet the two probes enableselective ease resulting from amutation in the GALC gene, along with wild- targeting of GALC and glucocerebrosidase, respectively.Atlow type littermates, were generated by crossing heterozygous (+/twi) concentrations, ABP 3 targets neither the relatedglycosidase mice in-house. Heterozygous C57BL/6J B6.CE-Galctwi/Jmice (stock a-galactosidase Anor,importantly,any of the other retaining number 000845) were obtained from the Jackson Laboratory (Bar b-galactosidases known to be present in humans, lysosomal Harbor,USA). Mouse pups were genotyped as previously de- scribed.[33] Mice ( 3weeks old) received the rodent AM-II diet acid b-galactosidase (EC 3.2.1.24) and intestinal LPH(E.C. Æ (Arie Blok Diervoeders, Woerden, The Netherlands). Animals were 3.2.1.106). We previously noted the same high degree of selec- housed and experiments were conducted according to protocols b tively for BODIPY-functionalized -glucopyranose-configured approved by the Institutional Animal Welfare Committee of the [28] epoxide 5. At low concentration,this probe selectively la- Academic Medical Centre Amsterdam in the Netherlands. beled lysosomal glucocerebrosidase but not the non-lysosomal glucosylceramidaseGBA2, cytosolic b-glucosidase GBA3, and Twelve week-old wt C57Bl/6 mice underwent intra-cerebroventricu- lar (i.c.v.) administration of ABP 3.The animals were kept in individ- LPH,[28] all other enzymes degrading glucosylceramide. ual cages at constant temperature (23 28C) and a12/12 hlight/ Æ GALC is synthesized as aprecursor of about 80 kDa that is dark cycle. They were exposed to ad libitum food and water processed into 30 and 50 kDa fragments after lysosomal before and after the experimental procedures. uptake. The two fragments do not dissociate but remainlinked Fibroblasts: Fibroblasts were obtained from skin biopsies with in- to each other by disulfide bridges.[1] The cleavage of the al- formed and signed consent from the donors. ready active80kDa precursor does not affect the enzymatic activity.[41] In line with this, we found that ABP 3 labels recom- Expression and labeling of human lactase phlorizin hydrolase binantly produced 80 kDa precursor GALC well. In lysates of (LPH): Primers were designed based on NCBI reference sequence NG_008104.2. The full-length cDNA sequence was cloned into GALC-overexpressing HEK293 cells, we could detect both the pcDNA3.1 in frame with the myc/His vector (Invitrogen). Confluent 80 kDa precursor and the 50 kDa mature subunit with ABP 3 HEK293 cells were transfected with empty pcDNA3.1 vector or the (data not shown). The50kDa GALC subunit, containing the vector with the described insert, in conjunction with FuGENE catalytic nucleophile residue E258, was the major GALC form (Roche), and harvested after 72 hbyscraping into potassium phos- visualized with ABP 3 in kidney, brain, and sciatic nerve of wt phate (KPi)buffer (25 mm,pH6.5) supplemented with 0.1%(v/v)

ChemBioChem 2017, 18,402 –412 www.chembiochem.org 409  2017 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Papers

Triton X-100 and protease inhibitor cocktail (Roche). Avolume error bars indicate standard deviation (SD). Graphpad Prism 5soft- equivalent to 50 mgofprotein was labeled with 1 mm (final concen- ware was used to determine apparent IC50 values. tration) of ABP 3 or ABP 5 for 2hat 378C. For tests of pH dependence, culture medium containing recombi- SDS-PAGE analysis and fluorescence scanning: Protein samples nant GALC was diluted 2:1(v/v)with McIlvaine buffer (10 mL, (recombinant enzyme, cell and tissue homogenate) were dena- pH 4.3) and mixed with substrate mix of various pH values (100 mL, 1 tured by adding 5” Laemmli sample buffer containing 2-mercap- 0.23 mgmLÀ 4-methylumbelliferyl-b-d-galactopyranoside in McIl- 1 vaine (pH 3–8)/H O(1:1, v/v), with 0.2m NaCl and 0.1%BSA). After toethanol ( =4 of sample volume) and boiling for 4min at 1008C. 2 The samples were then run on a7.5%or10% SDS-PAGE gel, and incubation at 378Cfor 30 min, the reaction was quenched, and wet slab gels were scanned for fluorescence by using aTyphoon fluorescence measured as described above. Displayed values repre- Variable Mode Imager (Amersham Biosciences, Piscataway,NJ, USA) sent mean values from six experiments, and error bars indicate SD. l l with ex =488 nm and em =520 nm (band pass:40) for green fluo- Fluorogenic substrate assay for recombinant a-galactosidase A: rescent ABP 5 and lex =532 nm and lem =610 nm (band pass:30) 1 Fabrazyme was diluted to aconcentration of 0.1 ngmLÀ in McIl- for red fluorescent ABP 3.Asaloading control, gels were stained vaine buffer (pH 4.6)/H2O(1:1, v/v)containing 0.1%BSA for stabili- with Coomassie Brilliant Blue (CBB) and destained with milliQ zation of the recombinant protein. Asolution of Fabrazyme (1 ng, water. 1 20 fmol per experiment, 10 mL, 0.1 ngmLÀ )was exposed to the in- Western blotting: Proteins were transferred onto aPVDF mem- dicated concentrations of compounds 1–4 (10 mL2”solution in H2O) for 30 min at 378C, before addition of substrate mix (100 mL, brane by using aBio-Rad Trans-Blot Turbo Transfer System. Mem- 1 branes were blocked with 1% BSA in Tris-buffered saline (TBS) with 1.5 mgmLÀ 4-methylumbelliferyl-a-d-galactopyranoside in McIl- 0.1%Tween-20 (TBST) for 1h at room temperature, hybridized vaine buffer (pH 4.6)/H2O(1:1, v/v)containing 0.1%BSA). After in- with streptavidin-horseradish peroxidase (HRP) for 1h at room cubation at 378Cfor 20 min, the reaction was quenched glycine temperature (1:10000 in blocking buffer) (Molecular Probes, Life (0.3m) adjusted with NaOH to pH 10.6, and fluorescence was mea- Technologies), washed with TBST and TBS, and then visualized by sured by using an LS55 fluorimeter (PerkinElmer) with lex =366 nm using an ECL+ western blotting detection kit (Amersham Biosci- and lem =445 nm. All samples were corrected for background fluo- ences). Protein standards were PageRuler Plus Prestained Protein rescence (sample without enzyme), and residual enzyme activity Ladder (Thermo Scientific) and biotinylated protein marker (Bio- was calculated as compared to acontrol sample incubated in the Rad). same manner but without inhibitors. In vitro labeling assays with recombinant GALC: In atypical ex- Proteomics: One kidney from afive week old wt mouse was ho- periment, culture medium containing recombinant GALC (5 mL) mogenized in 150 mm McIlvaine buffer,pH4.5 supplemented with was diluted with McIlvaine buffer (pH 4.3)/H O(2:1, v/v,14mL) and protease inhibitors. The homogenate (100 mL, containing 3.5 mg 2 exposed to the indicated concentrations of ABPs 3 or 4 (1 mLof protein) was incubated for 2h at 37 8Cwith ABP 4 (100 mL, a20” solution in DMSO) for 1h at 378C. Labeling of denatured 0.18 mm)inMcIlvaine buffer (150 mm,pH4.5). Analysis was per- enzyme was performed by pre-heating the enzyme to 1008Cfor formed as previously reported.[42] Peptides were desalted on stage 3min in buffer with or without 1% SDS (9 mLtotal volume) before tips[43] and analyzed with atrap-elute system on aC reversed- 18 addition of the probes. One half of each sample was resolved on phase nano LC with a45min 10–60%ACN/0.1%formic acid gradi- 10%SDS-PAGE. In-gel visualization of the fluorescent labeling by ent and aThermo LTQ-Orbitrap mass spectrometer by using a probe 3 was performed directly on the wet gel slabs. In the case top 3data-dependent protocol (60000 resolution, m/z range 300– of probe 4,biotinylated proteins were detected by performing 2000, 1000 ms fill time in the Orbitrap, 35 units of CID energy for streptavidin western blotting. MS/MS fragmentation, 120 ms max fill time, AGC 50 e3, and 750- count threshold). Ions of z=2+ and higher were selected to be Competition assays: Culture medium containing recombinant fragmented twice within 10 s, prior to exclusion for 150 s. Peak GALC (5 mL) in McIlvaine buffer (pH 4.3)/H2O(2:1, v/v;14mL) was lists were extracted and searched against the Uniprot mouse first exposed for 1h at 378Ctoeither 1 (0.2 mm), 2 (10 mm), or 4 (decoy) database, with carbamidomethylation of cysteine as afixed (10 mm,all 1 mLof20” solution in DMSO), before labeling with 3 modification and oxidation of methionine as avariable modifica- (0.2 mm, 1 mLof4mm stock in DMSO) as described above. tion, 20 ppm peptide tolerance, trypsin as protease, and with two missed cleavages allowed by using aMascot (matrix science) pH-dependent labeling assay: Culture medium containing re- m search engine. combinant GALC (5 L) was diluted with McIlvaine buffers (pH 3– 8)/H2O(2:1, v/v;14mL) and labeled with ABP 3 (0.5 mm; 1 mLof Fluorogenic substrate assay for recombinant GALC: Culture 10 mm stock in DMSO) for 1h at 37 8C. Gel bands were quantified medium containing recombinant GALC was diluted 2:1(v/v)with by using Image Lab 4.1 (Bio-Rad) software. Displayed values repre- McIlvaine buffer at (10 mL, pH 4.3) and exposed to the indicated sent mean values ( SD) from four independent experiments. Æ concentrations of compounds 1–4 (10 mL2” solution in H O) for 2 In vitro labeling assays with recombinant GALC and glucocere- 30 min at 37 8C, before addition of substrate mix (100 mL, 1 brosidase: Recombinant glucocerebrosidase (Cerezyme) in McIl- 0.23 mgmLÀ 4-methylumbelliferyl-b-d-galactopyranoside in McIl- vaine buffer (pH 5.2)/H O(1:1, v/v;9mL0.22 mm,2.0 pmol, 0.12 mg vaine buffer (pH 4.3)/H O(1:1, v/v)with 0.2m NaCl and 0.1%BSA). 2 2 per experiment) or culture medium containing recombinant GALC After incubation at 37 Cfor 30 min, the reaction was quenched 8 (2.5 mL) in McIlvaine buffer (pH 4.3)/H O(2:1, v/v;6.5 mL) were ex- with glycine (0.3m) adjusted with NaOH to pH 10.6, and fluores- 2 posed to either 1 mm of epoxide 3 or 1 mm of epoxide 5[28] (1 mLof cence was measured by using an LS55 fluorimeter (PerkinElmer) 10 mm stock in DMSO) for 1hat 378C. The reaction mixtures were with l =366 nm and l =445 nm. All samples were corrected for ex em then resolved on 7.5%SDS-PAGE, and in-gel visualization of the background fluorescence (sample without enzyme), and residual fluorescent labeling was performed directly on the wet gel slabs. enzyme activity was calculated as compared to acontrol sample incubated in the same manner but without inhibitors. Displayed In vitro labeling assay with mouse tissue homogenates: Animals 1 values represent mean values from triplicate experiments, and were first anesthetized with adose of Hypnorm (0.315 mgmLÀ

ChemBioChem 2017, 18,402 –412 www.chembiochem.org 410  2017 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Papers

1 phenyl citrate and 10 mgmLÀ fluanisone) and Dormicum fluorescent dyes:donkey anti-rabbit Alexa 488/Alexa 594 (Life Tech- 1 (5 mgmLÀ midazolam). The given dose was 80 mL/10 gbody nologies). The slides were then rinsed three times with TBS, weight. Animals were sacrificed by cervical dislocation. Tissues mounted, and covered with DAPI to be observed in aconfocal were collected, snap frozen in liquid N and stored at 808C. Later, laser scanning microscope (Leica SP5). Images were taken with 2 À homogenates from the frozen material were made in KPi buffer lex =488 nm for Alexa 488 and 561 nm for Alexa 594 and ABP 3.To (25 mm,pH6.5), supplemented with 0.1%(v/v)Triton X-100 and capturing images, a63” objective was used. In the case of single- protease inhibitors, and protein concentration was determined cell images, afurther threefold magnification was used. (BCA kit, Pierce). For labeling experiments, avolume of tissue ho- Fluorescence microscopy: HEK293 cells with and without stable mogenate equivalent to 50 mg(for kidney,brain and liver) or 25 mg overexpression of murine GALC were cultured on glass slides pre- of protein (for sciatic nerve) was completed to 10 mLwith water coated with poly-l-lysine. Cells were incubated with ABP 3 (5 nm) and incubated for 30 min at 378Cwith 1 mm of ABP 3 (10 mL2mm in the medium for 1h.Next, the cells were washed twice with PBS, in McIlvaine buffer 150 mm pH 4.3) or 100 nm of ABP 5 (10 mL and new medium was added. The cells were placed in the incuba- 200 nm in McIlvaine buffer 150 mm pH 5.2, 0.2%(w/v)sodium taur- tor overnight (approximately 16 h) before being washed again and ocholate, 0.1%(v/v)Triton X-100). The samples were then resolved then mounting the glass slides without fixation. For LAMP1 stain- on 7.5%SDS-PAGE and analyzed as described above. For competi- ing, the cells were fixed with 3% paraformaldehyde (v/v)before tion experiments the indicated concentrations of psychosine (7) being washed and blocked in donkey serum. Afterwards, the cells and ABB166 (6)dissolved in water were pre-incubated with the were stained with LAMP1 antibody for 1h,washed and incubated tissue lysate for 30 min on ice before addition of ABP 3 (final con- with the secondary Alexa 594-conjugated antibody (see above). centration 1 mm)and subsequent incubation at 37 8Cfor 15 min. The cells were observed by using aconfocal laser scanning micro- [25] Fluorogenic substrate assay with mouse tissue homogenates: scope (Leica SP5) as described previously. For measurement of b-galactosidase and b-galactocerebrosidase Fibroblast concanavalin Aaffinity purification and ABP labeling: activity,wefollowed the protocol previously described by Martino Cell pellets (from T-75 cm2 flasks) of fibroblasts from patients and [38] et al. with slight alterations. Briefly,avolume of tissue lysate healthy volunteer were lysed in KPi buffer (pH 6.5, 100 mL, 25 mm), equivalent to 7.5 mgofprotein was completed with water (6.25 mL supplemented with 0.1%(v/v)Triton X-100 and protease inhibitors. total volume) and exposed to the indicated concentration of ABP 3 The lysate was incubated overnight at 48Cwith pre-washed conca- or inhibitor 1 (6.25 mLofa2” solution in McIlvaine buffer,pH4.3) navalin Abeads (200 mL, Sigma–Aldrich). The beads were then for 60 min at 37 8C. Afterwards, AgNO3 (12.5 mLof110 mm solution washed with wash buffer according to the manufacturer’s instruc- 1 in water;11mm)and substrate mix (100 mL, 0.23 mgmLÀ 4-methyl- tions. Bead-bound glycoproteins were labeled with ABP 3 (1 mm)in umbelliferyl-b-d-galactopyranoside in McIlvaine pH 4.3) were McIlvaine buffer (pH 4.3) for 1h at 378C, followed by labeling for added, and the mixture was incubated for 45 min (for b-galactosi- 30 min with ABP 5 (0.1 mm)inMcIlvaine buffer (150 mm,pH5.5, dase) or 90 min (for GALC) at 378C. The reaction was then 0.2%(w/v)sodium taurocholate, 0.1%(v/v)Triton X-100). The sam- quenched with glycine (0.3m) adjusted with NaOH to pH 10.6, and ples were boiled in the presence of 5” Laemmli sample buffer and the fluorescence was measured with afluorimeter as described resolved by 10%SDS-PAGE. above. Data (n=3per group) represent mean SD. Statistical anal- Æ ysis was performed by using one-way ANOVA, followed by Dun- nett’stest. Acknowledgements ABP intra-cerebroventricular administration: For this purpose, we followed apreviously described procedure.[44] Briefly,intra-cere- The authors acknowledge financial supportfrom the European broventricular stainless steel guide cannulas were implanted in the Research Council(ERC (AdvancedGrant CHEMBIOSPHING) and lateral ventricle by using the following stereotaxic coordinates: the Netherlands Organization for Scientific Research (NWO-CW). AP 0.3, L+1.0, and V 2.2. After arecovery period of seven days, We thank Kassiani Kytidou for her assistance with the concana- À aneedle connected to atube was introduced in the guide cannula, valin Aaffinity purification. and asolution of ABP 3 (1 nm in PBS) was administered at arate 1 of 0.1 mLminÀ for 10 min. After 2h,the animals were sacrificed by

CO2 euthanasia and transcardially perfused with 250 mL of 0.9%of Keywords: beta-galactosidase · fluorescent probes · saline solution. Brains were isolated and immediately frozen for fur- galactosylceramidase · hydrolase · Krabbe disease ther biochemical and histological analysis. [1] J. E. Deane, S. C. Graham, N. N. Kim, P. E. Stein, R. McNair,M.B.Cachón- Fluorescent and immuno-histochemical analysis of brain sec- Gonzµlez, T. M. Cox,R.J.Read, Proc. Natl. Acad.Sci. USA 2011, 108, tions: Brains were cut in 30 mmslices with acryostat and attached 15169. to SuperFrost slides (Thermo Scientific). All of the following steps [2] S. Nagano,T.Yamada,N.Shinnoh, H. Furuya, T. Taniwaki, J. Kira, Clin. were performed in the dark to protect the fluorescence of the ABP. Chim.Acta 1998, 276,53. Slides were extensively washed in 0.01m TBS to remove non-spe- [3] C. H. Hill, S. C. Graham, R. J. Read, J. E. Deane, Proc. Natl. Acad. Sci. USA cific fluorescence and were covered in DAPI (Vector Lab, Burlin- 2013, 110,20479. game, USA) mounting media. [4] P. Formichi, E. Radi, C. Battisti,A.Pasqui, G. Pompella, P. E. Lazzerini, F. Laghi-Pasini, A. Leonini, A. Di Stefano,A.Federico, J. Cell. Physiol. 2007, For the immuno-histological analysis, brain slices were also exten- 212,737. sively washed in TBS and then incubated overnight in TBS with [5] K. Tanaka, H. Nagara, T. Kobayashi, I. Goto, BrainRes. 1988, 454,340. [6] D. A. Wenger,M.A.Rafi, P. Luzi, J. Datto, E. Costantino-Ceccarini, Mol. 0.5%(v/v)Triton X-100, 0.025%(v/v)gelatin, and rabbit-raised pri- Genet. Metab. 2000, 70,1. mary antibodies against LAMP1 (Millipore) and GALC (12887-1-AP, [7] D. A. Wenger,M.A.Rafi,P.Luzi, Hum.Mutat. 1997, 10,268. Proteintech, Chicago, IL, USA) at aconcentration of 1:1000. After [8] W. Krivit, E. G. Shapiro, C. Peters, J. E. Wagner,G.Cornu, J. Kurtzberg, primary antibody incubation, slides were incubated for 2h with D. A. Wenger,E.H.Kolodny, M. T. Vanier,D.J.Loes, K. Dusenbery, L. A. the appropriate secondary antibodies (1:1000) conjugated with Lockman, N. Engl. J. Med. 1998, 338,1119.

ChemBioChem 2017, 18,402 –412 www.chembiochem.org 411  2017 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Papers

[9] N. Sakai, Brain Dev. 2009, 31,485. J. D. C. CodØe, G. A. van derMarel, G. J. Davies, J. M. F. G. Aerts, H. S. [10] T. S. Wehrman,G.von Degenfeld,P.O.Krutzik, G. P. Nolan, H. M. Blau, Overkleeft, Chem. Sci. 2015, 6,2782. Nat. Methods 2006, 3,295. [28] M. D. Witte,W.W.Kallemeijn, J. Aten,K.-Y.Li, A. Strijland, W. E. Donker- [11] J. Han, M. S. Han, C.-H. Tung, Mol. Biosyst. 2013, 9,3001. Koopman, A. M. C. H. van den Nieuwendijk, B. Bleijlevens, G. Kramer,B.I. [12] M. Kamiya, D. Asanuma, E. Kuranaga,A.Takeishi, M. Sakabe, M. Miura, T. Florea,B.Hooibrink, C. E. M. Hollak, R. Ottenhoff, R. G. Boot, G. A. van Nagano,Y.Urano, J. Am. Chem. Soc. 2011, 133,12960. der Marel, H. S. Overkleeft, J. M. F. G. Aerts, Nat. Chem. Biol. 2010, 6,907. [13] Y. Urano,M.Kamiya, K. Kanda, T. Ueno, K. Hirose,T.Nagano, J. Am. [29] Y. Harrak, C. M. Barra, A. Delgado, A. R. CastaÇo, A. Llebaria, J. Am. Chem. Soc. 2005, 127,4888. Chem. Soc. 2011, 133,12079. [14] C.-H. Tung, Q. Zeng, K. Shah, D.-E. Kim, D. Schellingerhout, R. Weissled- [30] A. Alcaide, A. Trapero, Y. PØrez, A. Llebaria, Org.Biomol. Chem. 2015, 13, er, Cancer Res. 2004, 64,1579. 5690. [15] N.-H. Ho, R. Weissleder,C.-H. Tung, Chembiochem 2007, 8,560. [31] A. Biela-Banas´,F.Oulaı¨di, S. Front, E. Gallienne,K.Ikeda-Obatake, N. [16] B. Rotman,J.A.Zderic, M. Edelstein, Proc. Natl.Acad. Sci. USA 1963, 50, Asano, D. A. Wenger,O.R.Martin, ChemMedChem 2014, 9,2647. 1. [32] K. Suzuki, Brain Pathol. 1995, 5,249. [17] R. F. G. Frçhlich, K. Fantur, R. H. Furneaux, E. Paschke, A. E. Stütz,J. [33] N. Sakai, K. Inui, N. Tatsumi, H. Fukushima, T. Nishigaki, M. Taniike, J. Wicki, S. G. Withers, T. M. Wrodnigg, Bioorg.Med. Chem. Lett. 2011, 21, Nishimoto, H. Tsukamoto, I. Yanagihara,K.Ozono, S. Okada, J. Neuro- 6872. chem. 1996, 66,1118. [18] P. Greimel,H.Häusler,I.Lundt,K.Rupitz, A. E. Stütz,C.A.Tarling, S. G. [34] J. W. Callahan, Biochim. Biophys. ActaMol. Basis Dis. 1999, 1455,85. Withers, T. M. Wrodnigg, Bioorg. Med. Chem. Lett. 2006, 16,2067. [35] A. Zschoche,W.Fürst, G. Schwarzmann, K. Sanhoff, Eur.J.Biochem. [19] K. D. Janda, L. C. Lo, C. H. Lo, M. M. Sim, R. Wang, C. H. Wong, R. A. 1994, 222,83. Lerner, Science 1997, 275,945. [36] H. Tanaka, K. Suzuki, Brain Res. 1977, 122,325. [20] M. Kurogochi, S.-I. Nishimura, Y. C. Lee, J. Biol. Chem. 2004, 279,44704. [37] U. Ohto,K.Usui, T. Ochi, K. Yuki, Y. Satow,T.Shimizu, J. Biol. Chem. [21] D. H. Kwan, H.-M. Chen, K. Ratananikom, S. M. Hancock, Y. Watanabe, 2012, 287,1801. P. T. Kongsaeree,A.L.Samuels, S. G. Withers, Angew. Chem.Int. Ed. 2011, [38] S. Martino, R. Tiribuzi, A. Tortori, D. Conti, I. Visigalli, A. Lattanzi,A.Biffi, 50,300; Angew.Chem. 2011, 123,314. A. Gritti, A. Orlacchio, Clin. Chem. 2009, 55,541. [22] T. Komatsu, K. Kikuchi, H. Takakusa, K. Hanaoka, T. Ueno, M. Kamiya, Y. [39] A. K. Campbell,J.P.Waud, S. B. Matthews, Sci. Prog. 2009, 92,241. Urano,T.Nagano, J. Am. Chem.Soc. 2006, 128,15946. [40] J. Potter,M.W.Ho, H. Bolton, A. J. Furth, D. M. Swallow,B.Griffiths, Bio- [23] D. J. Vocadlo, C. R. Bertozzi, Angew.Chem.Int. Ed. 2004, 43,5338; chem. Genet. 1985, 23,423. Angew.Chem. 2004, 116,5452. [41] W. C. Lee, D. Kang, E. Causevic, A. R. Herdt, E. A. Eckman,C.B.Eckman, [24] L. I. Willems, T. J. M. Beenakker, B. Murray,B.Gagestein, H. van denElst, J. Neurosci. 2010, 30,5489. E. R. van Rijssel, J. D. C. CodØe, W. W.Kallemeijn, J. M. F. G. Aerts, G. A. [42] B. I. Florea, M. Verdoes, N. Li, W. A. van der Linden, P. P. Geurink, H. van van der Marel, H. S. Overkleeft, Eur.J.Org. Chem. 2014,6044. den Elst, T. Hofmann, A. de Ru,P.A.van Veelen, K. Tanaka, K. Sasaki, S. [25] W. W. Kallemeijn, K.-Y.Li, M. D. Witte, A. R. A. Marques, J. Aten, S. Scheij, Murata,H.den Dulk, J. Brouwer,F.A.Ossendorp, A. F. Kisselev,H.S. J. Jiang, L. I. Willems, T. M. Voorn-Brouwer,C.P.A.A.van Roomen, R. Ot- Overkleeft, Chem. Biol. 2010, 17,795. tenhoff, R. G. Boot, H. van denElst, M. T. C. Walvoort, B. I. Florea, J. D. C. [43] J. Rappsilber,M.Mann, Y. Ishihama, Nat. Protoc. 2007, 2,1896. CodØe, G. A. vander Marel, J. M. F. G. Aerts, H. S. Overkleeft, Angew. [44] D. HerreraMoro Chao, W. W. Kallemeijn,A.R.A.Marques, M. Orre,R.Ot- Chem. Int. Ed. 2012, 51,12529; Angew.Chem. 2012, 124,12697. tenhoff, C. van Roomen, E. Foppen, M. C. Renner, M. Moeton, M. van [26] L. I. Willems, T. J. M. Beenakker,B.Murray,S.Scheij, W. W. Kallemeijn, Eijk, R. G. Boot, W. Kamphuis, E. M. Hol, J. Aten, H. S. Overkleeft, A. Kals- R. G. Boot, M. Verhoek, W. E. Donker-Koopman, M. J. Ferraz, E. R. van beek, J. M. F. G. Aerts, PLoS One 2015, 10,e0138107. Rijssel,B.I.Florea, J. D. C. CodØe, G. A. van derMarel, J. M. F. G. Aerts, H. S. Overkleeft, J. Am. Chem. Soc. 2014, 136,11622. Manuscript received:October 24, 2016 [27] J. Jiang, W. W. Kallemeijn, D. W. Wright, A. M. C. H. van den Nieuwendijk, V. C. Rohde, E. C. Folch,H.van den Elst, B. I. Florea,S.Scheij, W. E. Accepted Article published:December 20, 2016 Donker-Koopman, M. Verhoek,N.Li, M. Schürmann, D. Mink, R. G. Boot, Final Article published:January 31, 2017

ChemBioChem 2017, 18,402 –412 www.chembiochem.org 412  2017 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim 本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP 图书馆。 图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具