Glycohydrolases β-hexosaminidase and β-galactosidase are associated with lipid microdomains of Jurkat T-lymphocytes

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Research paper

Glycohydrolases b-hexosaminidase and b-galactosidase are associated with lipidmicrodomains of Jurkat T-lymphocytes

Alessandro Magini a, Alice Polchi a, Brunella Tancini a, Lorena Urbanelli a,Andrej Hasilik b, Carla Emiliani a,*aDepartment of Experimental Medicine and Biochemical Sciences, University of Perugia, Via del Giochetto, 06123 Perugia, Italyb Institut fur Physiologische Chemie, Philipps-University Marburg, 35033 Marburg, Germany

a r t i c l e i n f o

Article history:Received 12 August 2011Accepted 21 September 2011Available online 29 September 2011

Keywords:Acidic glycohydrolasesMembrane microdomainsT-lymphocytes

a b s t r a c t

Growing evidence suggests the presence of active lysosomal enzymes in extra-lysosomal compartments,such as the plasma membrane. Although in the past little attention was paid to glycohydrolases acting oncellular compartments different from lysosomes, there is now increasing interest on plasma membrane-associated glycohydrolases because they should be involved, together with glycosyltransferases, in gly-cosphingolipids oligosaccharide modification processes regulating cell-to-cell and/or celleenvironmentinteractions in both physiological and pathological conditions. Starting from the previous evidence of thepresence of b-hexosaminidase and b-galactosidase at the plasma membrane of cultured fibroblasts, wehere investigated the association of these glycohydrolases with lipid microdomains of JurkatT-lymphocytes. Monosialoganglioside GM3 represents the major glycosphingolipid constituent of T-cellplasma membrane and its amount largely increases after T-cell stimulation. b-hexosaminidase andb-galactosidase cleave specific b-linked terminal residues from a wide range of glycoconjugates and inparticular are involved in the stepwise degradation of GM1 to GM3 ganglioside. Here we demonstratedthat fully processed plasma membrane-associated b-hexosaminidase and b-galactosidase co-distributewith the lipid microdomain markers and co-immunoprecipitate with the signalling protein lck inJurkat T-cell. Furthermore, Jurkat cell stimulation up-regulates the expression and activity of lysosomalb-hexosaminidase and b-galactosidase and increases their targeting to lipid microdomains. The non-random distribution of plasma membrane-associated b-hexosaminidase and b-galactosidase and theirlocalization within lipid microdomains, suggest a role of these enzymes in the local reorganization ofglycosphingolipid-based signalling units.

� 2011 Elsevier Masson SAS. All rights reserved.

1. Introduction

Glycosphingolipids (GSLs) are ubiquitous components of cellplasma membrane where they function as modulators of signaltransduction pathways regulating cell proliferation, survival anddifferentiation [1e3]. The concentration of GSLs is determined bya complex metabolic pathway involving biosynthesis, catabolismand intracellular trafficking. In addition to this, past and recent datasuggest that sphingolipids can be enzymatically remodelled at theplasma membrane. To date, there are several reports indicating thepresence of sialyltransferases and glycohydrolases associated to theextracellular site of cell plasma membrane [4e8]. The presence ofenzymatic activities involved in GSLs modification in extra-lysosomal compartments may be of biological relevance, indi-cating a possible role of these enzymes in signal transductionpathways regulating cell growth, differentiation and death. Despitethe relevance that their biological role may have in normal and/or

Abbreviations: ASB-14, amidosulfobetaine-14; Brij 98, polyoxyethylene (20)oleyl ether; CNS, central nervous system; CT-B, cholera toxin B subunit; EEA1, earlyendosome antigen 1; EZ-link Sulfo-NHS-LC-Biotin, sulfosuccinimidyl-6-(biotinamido)hexanoate; flot-2, flotillin-2; Gal, b-galactosidase; GalNAc, b-D-N-acetylgalactosamine;GlcCer, Glc-ceramide; GlcNAc, b-D-N-acetylglucosamine; GM1, Galb1,3GalNAcb1,4-(NeuAca2,3)-Galb1,4Glc-ceramide; GM2, 3GalNAcb1,4-(NeuAca2,3)-Galb1,4Glc-ceram-ide; GM3, NeuAca2,3Galb1,4Glc-ceramide; GSL, glycosphingolipid; LacCer, Galb1,4Glc-ceramide; Hex, b-hexosaminidase; IL-2, interleukin-2; LAMP-2, lysosomal associatedmembrane protein 2; a-Man, a-mannosidase; b-Man, b-mannosidase; MbCD, methyl-b-cyclodextrin; MUGal, 4-methylumbelliferyl-b-D-galactopyranoside; MUG, 4-methylumbelliferyl-N-acetyl-b-D-glucosaminide; MUGS, 4-methylumbelliferyl-N-acetyl-b-D-glucosaminide-6-sulphate; MUa-Man, 4-methylumbelliferyl-a-D-mannopyranoside; MUb-Man, 4-methylumbelliferyl-b-D-mannopyranoside; Neu3, sia-lidase-3; PHA, phytohemagglutinin; PMA, phorbol 12-myristate 13-acetate; TX-100,Triton X-100; Tween 20, polyethylene glycol sorbitan monolaurate; b-tub, b-tubulin.* Corresponding author. Tel./fax: þ390755857436.

E-mail address: [email protected] (C. Emiliani).

Contents lists available at SciVerse ScienceDirect

Biochimie

journal homepage: www.elsevier .com/locate/b iochi

0300-9084/$ e see front matter � 2011 Elsevier Masson SAS. All rights reserved.doi:10.1016/j.biochi.2011.09.021

Biochimie 94 (2012) 684e694

Author's personal copy

pathological conditions few details are available about membrane-associated glycohydrolases.

Glycohydrolases b-hexosaminidase (Hex, EC 3.2.1.52) andb-galactosidase (Gal, EC 3.2.1.23) are both involved in the stepwisedegradation of GM1 to GM3 ganglioside. Hex is an acidic glycohy-drolase that cleaves terminal N-linked GlcNAc or GalNAc residuesfrom oligosaccharides, glycolipids, glycoproteins and glycosami-noglycans [9]. GM2-gangliosidosis are two GSL lysosomal storagediseases caused by a genetic deficiency of acid Hex which results inthe accumulation of GM2 ganglioside, primarily in the CNS [9].Mencarelli et al. [10] originally demonstrated the presence ofplasma membrane-associated Hex A having the same biochemicalproperties of the lysosomal counterpart, thus indicating its lyso-somal origin. In addition, this form is active towards the naturalsubstrate GM2 ganglioside. Mature Hex was also found associatedto lysosomal membrane [11]. Glycohydrolase Gal catalyzes thehydrolysis of terminal N-linked galactosyl moiety from oligosac-charides and glycosides [12]. Genetic deficiency of acid Gal leads toGM1-gangliosidosis, a GSL lysosomal storage disease which resultsin the accumulation of GM1-ganglioside and its asialo-form (GA1)primarily in the CNS [13]. Also Gal enzyme has been recently re-ported to be associated with plasma membrane of human fibro-blasts [6]. Currently, very little is known about the biochemicalpathways leading from GM1 or GM2 gangliosides accumulation topathogenesis, and there are no effective therapies for the treatmentof these disorders [14].

Plasma membrane GSLs play a key role as signalling moleculesin both physiological and pathological processes regulating cell-to-cell and/or celleenvironment interactions [15]. Changes in thesphingolipid- and, more specifically, in the ganglioside-content atthe plasma membrane are often accompanied by dramatic modi-fications in the cellular properties. Ample evidence has accumu-lated on changes in cell membrane ganglioside pattern in thecourse of malignancy [15e17] as well as during neurodegenerationand inflammation [18e20]. GSLs are not homogeneously distrib-uted throughout the membrane surface. Rather, they concentratetogether with signalling proteins in restricted membrane areascalled GSL signalling domains, or lipid microdomains [21]. Theycontain a minor part of cell proteins, ranging from 1 to 3% of thetotal, but many of them play a crucial role in important pathwayssuch as cell signalling. Lipid microdomains are detergent-resistantmembrane areas that can be recovered exploiting their low densityduring sucrose-density gradient centrifugation [22]. Triton X-100(TX-100) is the most commonly used detergent for lipid micro-domain purification, although recently many detergent withspecific properties have been proposed [23,24].

In T-cells, lipid microdomains function as signalling platformsand the presence of microdomains at the cell surface increases withtheir stimulation [25,26]. Monosialoganglioside GM3 representsthe major glycosphingolipid constituent of T-cell plasmamembrane (72% of total ganglioside-content) [27], and stronglyincreases after T-cell stimulation [28]. The organization andsegregation of different types of lipid microdomains are essentialprocesses during T-cell migration as microdomain GM3 gangliosidepolarises to the leading edge, while GM1 localises to the uropod[29]. Moreover, microdomain GM1 has also been observed to coa-lesce to the immunological synapse during antigenic stimulationand GM3 has been shown to specifically associate with ZAP-70during activation [27,30].

Here we demonstrated that fully processed Hex and Galco-localise with lipid microdomains in Jurkat T-cell. We also shownthat Jurkat cell stimulation up-regulates the expression and activityof lysosomal Hex and Gal and gives rise to an increased activity ofthese enzymes in lipid microdomains. By contrast, a-mannosidase(a-Man, EC 3.2.1.24) and b-mannosidase (b-Man, EC 3.2.1.25), two

glycohydrolases not involved in GSLs catabolism, were not detect-able in lipid microdomains, although their intracellular activitywere comparable to that of Gal.

2. Materials and methods

2.1. Cell culturing and stimulation

JE6e1 Jurkat T-lymphocytes (American Type Culture Collection,Manassas, VA, U.S.A.), were cultured in RPMI 1640 medium supple-mented with 10% (v/v) heat-inactivated bovine foetal serum (FBS),2 mM L-glutamine, 100 units/ml penicillin, 100 mg/ml Streptomycin ina humidified incubator under 5% CO2 at 37 �C. Cell viability wasdetermined by Trypan blue method. Stimulation of Jurkat cells wascarried out by using 1 mg/ml of PHA (SigmaeAldrich) and 50 ng/ml ofPMA (SigmaeAldrich) and incubating for 6, 12 or 24 h at theconcentration of 1 � 106 cell/ml.

2.2. RNA extraction and real-time Q-PCR

RNA was extracted from 1 � 107 cells with PureLinkTM TotalRNA Purification System (Invitrogen). 2 mg of RNA were reverse-transcribed into cDNA using random hexamers and Super-Script� II Reverse Transcriptase according to the manufacturer’sprocedure (Invitrogen). cDNA was used as template for theestimation of Hex b-subunit (HEXB), Hex a-subunit (HEXA), Gal(GLB1) and interleukin-2 (IL-2) genes expression by quantitativePCR (Q-PCR) in a Stratagene Mx3000P Q-PCR machine (AgilentTechnologies). Reactions were performed in triplicate usingBrilliant II SYBR Green Q-PCR Master Mix. Primers weredesigned using Primer-BLAST software (http://www.ncbi.nlm.nih.gov/tools/primer-blast). Sequences used for the amplifica-tion of HEXB, HEXA, GLB1 and IL-2 genes were: 50-TTTGGGAGGAGATGAAGTGG-30 (forward) and 50-AAACCTCCTGCCAGACAATG-30

(reverse), 50-GCATTTGAAGGTACCCCTGA-30 (for) and 50-TCAACTTGTTGCTCCACAGC-30 (rev), 50-GTTATAACAGTGCAGGTTGAAAATGAA-30 (for) and 50-CCCAGATGGTGGCGAAAG-30 (rev), 50-TGCACTTGTCACAAACAGTGCACCT-30 (for) and 50-TCTGTGGCCTTCTTGGGCATGT-30 (rev). b-Actin (ACTB) gene was amplified asendogenous control using the following primers 50-AGAAAATCTGGCACCACACC-30 (for) and 50-GGGGTGTTGAAGGTCTCAAA-30

(rev). Data were analyzed using the DDCt method. DCt wascalculated subtracting the average Ct value of ACTB to theaverage Ct value of each target gene. DDCt, calculated for eachgene, was the difference between the DCt of resting cells and theDCt of stimulated cells. The relative quantity (RQ) was calculatedby 2�DDCt.

2.3. Isolation of lipid microdomains and cholesterol depletion

Lipid microdomains from resting and stimulated Jurkat cellswere isolated as described by Rodger et al. [22], with some modi-fications. Briefly, 1 � 108 cells were resuspended with 0.5 ml of10 mM Tris, 150 mM NaCl, 5 mM EDTA, pH 7.4 (TNE) containingeither 1% (v/v) TX-100, 1% (v/v) Brij 98 (SigmaeAldrich), 1% (v/v)Tween 20 (SigmaeAldrich) and no detergent. Extraction was doneon ice for 30 min, except for Brij 98, which was applied at 37 �C for15 min. 20 ml of protease inhibitor cocktail were added. The cellswere mechanically disrupted by Dounce homogenization (10strokes). The lysate was centrifuged at 1500 g for 5 min at 4 �C toremove nuclei and large cellular debris. The supernatant was thenmixed with 0.5 ml of 85% (w/v) sucrose solution in TNE, transferredinto a polyallomer centrifuge tube, then carefully overlaid with2.75 ml of 35% (w/v) sucrose solution in TNE, and finallywith 1.25 ml of 5% (w/v) sucrose solution in TNE. Thereafter,

A. Magini et al. / Biochimie 94 (2012) 684e694 685

Author's personal copy

discontinuous sucrose-density gradient centrifugation was per-formed at 4 �C for 18 h at 50,000 rpm, using a MLS-50 rotor and anOptima Max ultracentrifuge. Eleven fractions of equal volume(450 ml) were collected from the top to the bottom of the discon-tinuous sucrose-density gradient.

Cholesterol was depleted by incubating the cells (1 �107/ml) inserum-free medium for 30 min at 37 �C in the presence of 15 mMMbCD (SigmaeAldrich) prior to cell lysis.

2.4. Biotinylation and affinity purification of cell surface lipidmicrodomain proteins

Jurkat cells (2 � 108) were washed three times by ice-cold 0.1 Msodium phosphate, 0.15 M NaCl, pH 7.4 (PBS) and incubated, undergentle shaking, for 30 min at room temperature in 20 ml of PBScontaining the EZ-Link Sulfo-NHS-LC-Biotin (Thermo Scientific) atthe concentration of 1 mg/ml. The cells were washed three timeswith PBS containing 100 mM glycine to quench and remove excessbiotin reagent and homogenized in 0.5 ml of TNE containing 1% (v/v) TX-100 and 20 ml of protease inhibitor cocktail. Successively, lipidmicrodomains were isolated as reported above.

In order to eliminate sucrose and concentrate lipidmicrodomainvesicles, flot-2 positive fraction 3 was mixed with four volumes ofTNE containing 1% (v/v) TX-100 in polyallomer centrifuge tube.Samples were ultracentrifuged at 60,000 rpm at 4 �C for 2 h usinga TLA-100.3 rotor and an Optima Max ultracentrifuge. The pelletwas resuspended in 100 ml of PBS containing 1% (v/v) TX-100 andlipid microdomain complexes were disaggregated by incubatingthem for 10 min at 37 �C. Thereafter, four volumes of PBS wereadded and disaggregated lipid microdomain preparation wasloaded in a 0.5 ml Monomeric Avidin Affinity Column (ThermoScientific). The column was washed with five resin-bed volumes ofPBS until all unbound proteins had been washed off and the A280returned to baseline. Biotinylated proteins were then eluted using5 mM D-biotin in PBS.

2.5. Horseradish peroxidase assay

Jurkat cells suspended in RPMI 1640 containing 10% (v/v) FBS ina6-well tissue cultureplate (1�106/ml)were loadedwithhorseradishperoxidase (HRP, SigmaeAldrich) in a concentration of 2 mg/ml. Cellsweremaintainedat 37 �C in ahumidifiedatmosphereof5%CO2 for 2h.Successively, cells were recovered by a low speed centrifugation andwashed with RPMI 1640 medium at room temperature and resus-pended in the fresh medium containing 10% (v/v) FBS. After 16 h ofincubation, cells were stimulated for 24 h as described above and theculture medium was recovered for HRP activity determination using2,20-Azino-bis(3-ethylbenzothiazoline)-6-sulfonic acid (ABTS) assubstrate [31]. Briefly, 4e6 ml of diluted culturemediumwere added to1ml of 0.7mMABTS in 100mMpotassiumphosphate buffer, pH5.0 at25 �C; then 30 ml of fresh prepared 0.3% (v/v) H2O2 were added,samples were mixed by inversion and the increase in absorbance at405 nm was immediately recorded for 3 min. DA405 nm/min wascalculated using the maximum linear rate. The same procedure wasperformed with resting cells as negative control. One unit (U) is theamount of enzyme that oxidizes 1 mmol of substrate/min at pH 5.0at 25 �C.

2.6. Determination of enzyme activities and protein concentration

Total Hex, Hex A, Gal, a-Man and b-Man activities were deter-mined using a final concentration of 2 mM MUG (SigmaeAldrich),2 mM MUGS (Toronto Research Chemicals Inc.), 1 mM MUGal (Sig-maeAldrich), 2 mM MUa-Man (SigmaeAldrich), 2 mM MUb-Man

(SigmaeAldrich) respectively, in 0.1 M citric acid/0.2 M disodiumphosphate buffer, pH 4.5.

The optimum pH values for the Hex and Gal activities weredetermined using MUG, MUGS and MUGal substrates in 0.1 M citricacid/0.2Mdisodiumphosphate buffers for pH values ranging frompH3.5 to 6.5 and 0.2 M phosphate buffer for pH 7.0 and 7.5 values.MichaeliseMenten constantKmwas obtained using LineweavereBurkplot. We obtained Km values for Hex using the following finalconcentrations of the substrates MUG and MUGS: 2, 1, 0.5, 0.25, 0.12,0.06, 0.03, 0.015 and 0.0075 mM dissolved in 0.1 M citric acid/0.2 Mdisodium phosphate buffer, pH 4.5. Moreover, Km value for Gal wasobtained using the following final concentrations of the substrateMUGal: 1, 0.5, 0.25, 0.12, 0.06, 0.03, 0.015, 0.0075 and 0.00375 mMdissolved in 0.1M citric acid/0.2Mdisodiumphosphate buffer, pH 4.5.

Reactions were performed in triplicate in 96-well black multi-plates (Greiner, Frickenhausen, Germany) at 37 �C. At the end of thereaction period, 0.290 ml of 0.4 M glycine/NaOH buffer, pH 10.4were added. Fluorescence of the liberated 4-methylumbelliferonewas measured on a Infinite F200 fluorimeter (Tecan, Mannedorf,Switzerland) at 360 nm excitation, 450 nm emission. One enzy-matic unit (U) is the amount of enzyme that hydrolyses 1 mmol ofsubstrate/min at 37 �C. Enzymatic activity was expressed as enzy-matic units � 10�3 (mU).

Protein concentration was determined by the method of Brad-ford [32] using bovine serum albumin as standard.

2.7. Western blotting and dot blot analysis

Fractions resulting from isolation of both lipid microdomainsand biotinylated lipid microdomain proteins from avidin affinitycolumn were subjected to 10% SDS-PAGE under reducing condi-tions according to Laemmli [33]. Proteins were transferred to PVDFmembrane (Biorad), blocked with 5% (w/v) skim milk in 80 mMNa2HPO4, 20 mM NaH2PO4 and 100 mM NaCl containing 0.1% (v/v)Tween 20, and reacted with primary antibodies for 1 h at thefollowing dilutions: 1:5000 for mouse monoclonal anti-flot-2 (BDBiosciences), 1:200 for mouse monoclonal anti-lck (Santa CruzBiotechnology), 1:1000 for mouse monoclonal anti-EEA1 (ABcam),1:100 for mouse monoclonal anti-LAMP-2 (Santa Cruz Biotech-nology), 1:500 for mouse monoclonal anti-b-tubulin (Santa CruzBiotechnology), 1:500 for a goat antiserum specific for Hexa-subunit, raised to a mixture of synthetic peptides belonging tothe sequence of mature human enzyme [11], 1:500 for mouse anti-Gal (SigmaeAldrich). After being washed, the blots were incubatedwith secondary antibodies anti-mouse IgG (GE Healthcare) or anti-goat IgG HRP-conjugated (SigmaeAldrich), and were developed byECL detection system (GE Healthcare).

Fractions from discontinuous sucrose-density gradient andbiotinylated lipid microdomain proteins from avidin affinitycolumn were spotted on nitrocellulose membrane (Biorad). Bio-tinylated proteins were revealed by using 1:500 streptavidin HRP-conjugated (Thermo Scientific) for 1 h and developed by ECLdetection system. In fractions from discontinuous sucrose-densitygradient, GM1 was revealed by using 1:500 cholera toxin Bsubunit biotin-conjugated (SigmaeAldrich) for 1 h. Successively,the membrane was incubated with 1:1000 streptavidin HRP-conjugated and developed by ECL detection system.

2.8. Immunoisolation of lipid microdomains

Lipid microdomains were prepared and concentrated fromresting or stimulated Jurkat cells as described above. Microdomainpellets from 5 � 107 cells were resuspended in 200 ml TNE/1% (v/v)TX-100 or 200 ml TNE/1% (w/v) ASB-14 (SigmaeAldrich) and incu-bated with 2 mg monoclonal anti-lck at 4 �C for 3 h under constant

A. Magini et al. / Biochimie 94 (2012) 684e694686

Author's personal copy

agitation. These suspensions were incubated with 50 ml Dynabeadsprotein G (Invitrogen) overnight. Controls were conducted withbeads that received no antibody. After magnetic application,supernatants were collected and the beads were washed five timeswith 500 ml of 200 mM NaCl. The bound aggregates were treatedwith 150 ml of TNE/1% (w/v) ASB-14 for 10 min at 37 �C to disag-gregate lipid microdomain complexes. The supernatant and thereleased aggregate fractions were assayed for Hex and Gal activitiesor subjected to immunoblotting with anti-flot-2 as a control.

3. Results

3.1. Jurkat cell lipid microdomains contain Hex and Galglycohydrolases

Jurkat cell lipid microdomains were obtained by flotationexperiments [21] by using different approaches as described in the

“Materials and methods”. Cells were treated with TNE containingeither 1% TX-100, 1% Brij 98, 1% Tween 20 or without detergent, andlow-density floating material was separated from the unfloatingmaterial by using discontinuous sucrose-density gradient centri-fugation. Fractions of the gradient were analyzed for the presenceof the specific microdomain markers GM1 and flotillin-2 (flot-2) byimmunoblotting analysis [34]. As shown in the Fig. 1A and B, in allcases where it was used detergent, no matter what, GM1 and flot-2were highly enriched in the light-density fractions 2e4. In partic-ular, using the TX-100 detergent, GM1 and flot-2 were recoveredmainly in the fraction 3, corresponding to the 5e35% sucroseinterface. Differently, using the detergent-free method, lipidmicrodomain markers were recovered both in light-density frac-tions 3e5 and in high-density fractions 8e11. Protein quantificationassay showed the presence of a distinct peak with a maximum atthe light-density fraction 3 for detergentmethods, whereas floatingproteins obtained using detergent-free method were detected asa broad peak in the fractions 3e5 (Fig. 1C).

Fig. 1. Association of Hex and Gal with lipid microdomains in Jurkat cells. The lysates from Jurkat cells (1 � 108) in TNE buffer containing either 1% TX-100, 1% Brij 98, 1% Tween 20and no detergent were fractionated by a discontinuous sucrose-density gradient consisting of 5, 35, and 42.5% sucrose in TNE buffer. Eleven fractions (450 ml) were collected fromthe top of the gradient. (A) Aliquots of each fraction (1.5 ml) were spotted in an nitrocellulose membrane and the presence of GM1 was revealed by biotinylated CT-B. Representativeblot of three independent experiments is reported. (B) One-fifteenth volume of each fraction was subjected to immunoblotting for flot-2. Representative Western blotting of threeindependent experiments is reported. (C) Gradient distribution of proteins obtained using TX-100, Brij 98, Tween 20 and no detergent (detergent-free) is shown as mg of proteinsrecovered in each fraction. Values are the mean � SD of three independent experiments. (D) Distribution of Total Hex, Hex A and Gal enzymatic activities obtained using TX-100, Brij98, Tween 20 and no detergent (detergent-free) was expressed as total mU (tot. mU) in each fraction. Values are the mean � SD of three independent experiments. LM, Lipidmicrodomains; H, High-density fractions.

A. Magini et al. / Biochimie 94 (2012) 684e694 687

Author's personal copy

Collected fractions were assayed for enzymatic activity of Hexusing the two substrates MUG, which is hydrolysed by both a- andb-subunits forming Hex isoenzymes (Total Hex), and MUGS, whichis hydrolysed only by the a-subunit-containing isoenzyme (Hex A).Gal enzymatic activity was assayed using MUGal substrate. As theacidic glycohydrolase b-galactocerebrosidase (GALC, EC 3.2.1.46)may contribute to the hydrolysis of the substrate MUGal [35], thepresence of Gal was confirmed by Western blotting analysis usingthe specific antibody (data are shown in Fig. 3, panel C insert). Asreported in Fig. 1D, the Total Hex, Hex A and Gal activities havesimilar distribution showing a distinct peak with a maximum at thelight-density fraction 3 and co-distributing with GM1 and flot-2,when detergent methods were used. In a different way, enzymeactivity profiles obtained using the detergent-free method showeda broad Hex and Gal enzymes distribution in fractions 3e5, similarto that seen for protein distribution. As expected, the major portionof the Hex and Gal activities was recovered in the high-densityfractions 10e11 for both detergent and detergent-free methods.

Results show that lipid microdomain extraction methods thatuse detergent give a lower diffusion of the microdomain markersGM1 and flot-2 compared to the detergent-free method. Amongthese, the TX-100method, although allows aminor recovery of Hexand Gal enzyme activities in light-density fraction 3, providesa more stringent method to exclude non-lipid microdomain

proteins contamination, due to the higher solubilizing capacitywith respect to Brij 98 and Tween 20 [36]. It is for this reason thatnext studies of Hex and Gal lipid microdomains-association werecarried out using TX-100 method.

3.2. TX-100 insoluble microdomains characterization

TX-100 insoluble microdomains from Jurkat cells were preparedas above described using discontinuous sucrose-density gradientcentrifugation. Fractions of the gradient were analyzed for thepresence of the lipid microdomain markers GM1 and flot-2, andfurther characterized for the presence of the early endosomemarker EEA1, the lysosomal marker LAMP-2 and the cytosolicmarker b-tubulin. As shown in the Fig. 2A and B, whereas GM1 andflot-2 were highly enriched in the light-density fraction 3, EEA1,LAMP-2 and b-tubulinwere detectable in the high-density fractions8e11, thus demonstrating the successful partition of membranemicrodomains. Under these conditions, about 1.75% of the totalproteins was recovered in lipid microdomain fractions (Fig. 1C).Collected fractions were assayed for Hex and Gal activities (Fig. 1D)and for activity of a-Man and b-Man, which are lysosomal glyco-hydrolases not involved in GSLs catabolism. Results reported inFig. 2C show the absence of both a-Man and b-Man in flot-2 posi-tive fractions, although their total activity in cell extract wascomparable to that of Gal. Microdomain-associated Hex and Galdisplayed pH activity profiles similar to their soluble lysosomalcounterparts, with a maximum at pH 4.5 and 4.0, respectively,and similar residual enzymatic activity at pH 7.0 (Table 1).Furthermore, Km values of microdomain-associated Hex and Galtowards synthetic substrates were similar to those of their solublecounterparts (Table 1).

Fig. 3. Cell surface localization of lipid microdomain-associated Hex and Gal enzymes.Cell surface proteins from Jurkat cells (2 � 108) were in vivo labelled with EZ-LinkSulfo-NHS-LC-Biotin, lipid microdomains were extracted and biotinylated proteinswere purified on an avidin affinity column as described under “Materials andmethods”. (A) One-thirtieth volume of each fraction from sucrose-density gradientwas subjected to immunoblotting for flot-2. Representative Western blotting of threeindependent experiments is reported. (B) Aliquots (1.5 ml) of concentrated and solu-bilized fraction 3 lipid microdomains (LM3), avidin affinity column flow-through (F)and eluate (E) were spotted on nitrocellulose membrane and biotinylated proteinswere revealed by using streptavidin HRP-conjugated (Strep-HRP). (C) Approximately15 mg of proteins from LM3, F and E were subjected to immunoblotting for Hex a-subunit and Gal. Representative Western blotting of three independent experiments isreported. *, 54 kDa; **, 64 kDa.

Fig. 2. Characterization of lipid microdomains purified by using TX-100 method. Thelysate from Jurkat cells (1 � 108) in TNE buffer containing 1% TX-100 was fractionatedby a discontinuous sucrose-density gradient as reported in the legend of Fig. 1. Elevenfractions (450 ml) were collected from the top of the gradient. (A) Aliquots of eachfraction (1.5 ml) were spotted in an nitrocellulose membrane and the presence of GM1was revealed by biotinylated CT-B. Representative blot of five independent experi-ments is reported. (B) One-fifteenth volume of each fraction was subjected to immu-noblotting for flot-2, EEA1, LAMP-2 and b-tubulin. Representative Western blotting offive independent experiments is reported. (C) Distribution of Gal (dotted line), a-Manand b-Man enzymatic activities was expressed as total mU (tot. mU) recovered in eachgradient fraction. Values are the mean � SD of three independent experiments. LM,Lipid microdomains; H, High-density fractions.

A. Magini et al. / Biochimie 94 (2012) 684e694688

Author's personal copy

3.3. Fully processed Hex and Gal glycohydrolases are associatedto the external leaflet of Jurkat plasma membrane

To further characterize Hex and Gal glycohydrolases, lipidmicrodomain proteins exposed to the external leaflet of Jurkat cellplasma membrane were purified by in vivo cell surface bio-tinylation, followed by lipid microdomains extraction. Thereafter,lipid microdomain biotinylated proteins were recovered by avidinaffinity chromatography to obtain highly purified microdomainproteins exposed to the extracellular environment.

TX-100 insoluble microdomains were purified as abovedescribed and flot-2 positive fraction 3 (Fig. 3A) was recovered andultracentrifuged to eliminate sucrose and concentrate lipid micro-domains. To perform the affinity chromatography, lipid micro-domain complexes were opportunely disaggregated by usingTX-100 at 37 �C (LM3). Avidin affinity chromatography was fol-lowed for the presence of biotinylated proteins by dot blot usingHRP-conjugated streptavidin. As shown in Fig. 3B, biotinylatedproteins were almost completely bound to the column and recov-ered in the eluate fraction (E). Immunoblot analysis of

Table 1Optimum pH and MichaeliseMenten (Km) values of microdomain-associated and soluble Hex and Gal. Optimum pH and Km for microdomain-associated and soluble Hex andGal were determined in fraction 3 and fraction 11, respectively.

Optimum pHa pH 7.0 residual activitya (%) Kma (mM)

Total Hex Hex A Gal Total Hex Hex A Gal Total Hex Hex A Gal

Microdomain-associated 4.5 4.5 4.0 14 � 2 11 � 2 16 � 2 0.58 � 0.04 0.40 � 0.05 0.26 � 0.03Soluble 4.5 4.5 4.0 13 � 2 12 � 1 15 � 3 0.49 � 0.06 0.49 � 0.06 0.31 � 0.03

a Values are the mean � SD of five independent experiments.

Fig. 4. Increase of Hex and Gal enzymatic activities in lipid microdomains of stimulated Jurkat cells after perturbation of lipid microdomains by MbCD treatment. Resting and 24 hPHA-stimulated cells (1 � 108), either non-treated or treated with 15 mM MbCD, were solubilized by using TNE buffer containing 1% TX-100. Lysates were fractionated bya discontinuous sucrose-density gradient as reported in the legend of Fig. 2. Eleven fractions (450 ml) were collected from the top of the gradient. (A) Aliquots of each fraction (1.5 ml)were spotted in a nitrocellulose membrane and the presence of GM1 was revealed by biotinylated CT-B. The PHA addition is indicated. Representative blot of five independentexperiments is reported. (B) One-fifteenth volume of each fraction was subjected to Western blot analysis for flot-2 and lck (*, p56lck; **, p60lck) and the addition of PHA and/orMbCD are indicated. Representative Western blotting of five independent experiments is reported. (C) Gradient distribution of Total Hex, Hex A and Gal enzymatic activities forresting, resting MbCD treated, stimulated and stimulated MbCD treated cells were expressed as total mU (tot. mU) in each fraction. Values are the mean � SD of five independentexperiments. Histograms represent the results of gene expression analysis, performed by real-time Q-PCR. Reactions were performed in triplicate, using SYBR green binding todetect amplification. ACTB gene was used as endogenous control. The fold expression of HEXB, HEXA and GLB1 genes in stimulated (þ) with respect to resting (�) Jurkat cells isreported. The values are expressed as Relative Quantity (RQ). The mean � SD of three independent experiments is reported. LM, Lipid microdomains; H, High-density fractions.

A. Magini et al. / Biochimie 94 (2012) 684e694 689

Author's personal copy

chromatographic fractions highlighted the presence of Hex a-subunit and Gal polypeptides in eluted fraction E, so indicating thatbothmicrodomain-associated enzymes are exposed on the externalleaflet of the plasma membrane (Fig. 3C). Furthermore, bandsrevealed a 54 and 64 kDa molecular weight for Hex a-subunit andGal, respectively, corresponding to the fully processed proteins.Moreover, mature Hex and Gal are also present in flow-throughfraction (F) suggesting that a portion of these lipid microdomain-associated enzymes, as expected, is not exposed on the cell surface.

3.4. Stimulation of Jurkat cells causes a recruitment of Hex and Galin membrane microdomains

Jurkat cells were treated with 1 mg/ml PHA for 24 h and cellstimulation was confirmed by the increased production of IL-2[37] as assessed by real-time Q-PCR. To determine whether Hexand Gal increased in lipid microdomains after cell stimulation weisolated these membrane domains from resting and stimulatedJurkat cells by flotation experiments as described above. Fractionsof discontinuous sucrose-density gradient were analyzed for GM1,flot-2 and also for lck [38], which is known to be hyper-phosphorylated and translocated into the plasma membranemicrodomains upon cell stimulation [39]. As shown in Fig. 4A andB, in lipid microdomain fraction 3 Jurkat cell stimulation did notchange appreciably the GM1 content, whereas the amount of thespecific marker flot-2 resulted increased. Apparently, cell stimu-lation enhanced the content of the p56lck band and causeda prominent rise in the p60lck hyperphosphorylated band. Bothevents confirmed that specific lipid microdomain componentswere correctly recruited to the membrane during T-cell stimula-tion. Recovered fractions were assayed for Hex and Gal enzymaticactivities and results, reported in Fig. 4C, show an evident increaseof activity in flot-2 positive fraction 3 of stimulated cells: theincrease was quantifiable in 2.7-fold for Total Hex, 2.9-fold for HexA and 2.5-fold for Gal, with respect to non-stimulated cells. Bycontrast, Total Hex, Hex A and Gal activities in stimulated cellcrude extract were only 1.5-, 1.4- and 1.5-fold higher with respectto resting cells, indicating selective recruitment of these enzymesto the plasma membrane microdomains during cell stimulation.

Real-time Q-PCR experiments revealed that in stimulated Jurkatcells there was an increase of Hex and Gal transcription levels thataccount for theobserved increaseof theenzymaticactivities (Fig. 4C).

a-Man and b-Man enzymatic activities assayed in lipid micro-domain fraction 3 of stimulated cells did not show any noticeableincrease with respect to resting cells, even if their total enzymaticactivity increasedaswellas theotherglycohydrolases(datanotshown).

Since cholesterol is a major component of membrane micro-domains, we also assessed the effect of membrane cholesterolcontent perturbation on Hex and Gal lipid microdomain associa-tion. Tomodify the organization of lipid microdomains we depletedcholesterol by MbCD cell treatment. Resting and stimulated Jurkatcells were incubated with MbCD prior to the TX-100 treatment. Asshowed in Fig. 4B, microdomain perturbation by cholesteroldepletion significantly reduced lipid microdomain association offlot-2 and lck markers, making them more soluble in TX-100, asexpected. Under this condition the activity of Total Hex, Hex A andGal as recovered in fraction 3 was also consistently reduced in bothresting and stimulated cells (Fig. 4C).

3.5. Increase of Hex and Gal enzymatic activities in lipidmicrodomains is time-dependent and proportional to theintensity of stimulation

To investigate whether Hex and Gal activities increased in lipidmicrodomains in a time-dependent manner during stimulation, we

isolated these membrane domains from Jurkat cells stimulatedwith PHA for 6, 12 or 24 h. Data confirmed the previously observeddistribution of Total Hex, Hex A and Gal activities within sucrose-density gradient and highlighted that the activity increase in frac-tion 3 was time-dependent, with the beginning clearly detectableafter 6 h and a continued increase up to 24 h from stimulation(Fig. 5).

Jurkat cell stimulation can be markedly increased by PMAaddition to PHA, an event that is accompanied by IL-2 over-production [37]. PHA/PMA stimulation resulted in about 4-foldIL-2 increase with respect to the PHA stimulation, as monitoredby real-time Q-PCR. In our experiments, the co-stimulationresulted: i) in no appreciable change in the GM1 content(Fig. 6A); ii) in an increased accumulation of the flot-2

Fig. 5. Time-dependent increase of Hex and Gal activities in lipid microdomains ofstimulated Jurkat cells. Resting and stimulated Jurkat cell (1 � 108) lysates wereprepared in TNE buffer containing 1% TX-100 and subjected to discontinuous sucrose-density gradient separation as reported in the legend Fig. 2. Eleven fractions werecollected from the top of the gradient. Total Hex, Hex A and Gal enzymatic activities ofresting and 1 mg/ml PHA-stimulated cells, incubated for 6, 12 or 24 h were assayed inlipid microdomain fraction 3. Histograms show the % of enzymatic activity in stimu-lated cells with respect to resting cells. Values are the mean � SD of three independentexperiments.

A. Magini et al. / Biochimie 94 (2012) 684e694690

Author's personal copy

microdomain marker in fraction 3 (Fig. 6B); iii) in a concomitantincrease of 2.2-, 2.0- and 1.6-fold of Total Hex, Hex A and Galactivities, respectively, compared to the PHA-stimulated cells.Moreover, the increase of Total Hex, Hex A and Gal activities was5.0-, 5.5- and 3.9-fold with respect to resting cells, respectively(Fig. 6C). Again, a-Man and b-Man enzymatic activities assayed inlipid microdomain fraction 3 of PHA/PMA treated cells did notshow any relevant increase with respect to resting cells, even iftheir total enzymatic activity increased as well as the other gly-cohydrolases (data not shown).

3.6. Jurkat cell stimulation induces fusion of lysosomeswith cell surface

Like other lysosomal enzymes, after de novo synthesis, Hex istransported from TGN to lysosomes, where undergoes to the finalproteolytic processing [9]. The presence of fully processed Hexa-subunit on lipid microdomains suggests that the enzyme is firsttargeted to the lysosome and later to the cell surface, probably asresult of fusion events of the lysosomal membrane with the plasmamembrane.

To determine if such lysosome-to-plasma membrane transportpathway was altered by cell stimulation, Jurkat cells were incu-bated for 2 h at 37 �C in culture medium containing 2 mg/ml HRP,rinsed and then chased for 16 h in fresh culture medium. There-after, cells were stimulated for 24 h and the culture medium wasrecovered for HRP activity determination using ABTS as substrate.Results reported in Fig. 7 show that HRP activity released in theculture medium increased approximately 1.7 and 2.0-fold in PHAand PHA/PMA treated cells, respectively, compared to the restingcells (P < 0.01).

3.7. Lck-containing microdomains include Hex and Gal enzymes

To perform the immunoprecipitation of microdomains withthe anti-lck antibody, flot-2 positive fraction 3 was diluted inTNE containing 1% TX-100 and subjected to high-speed centri-fugation. Microdomain pellet was resuspended with 200 ml TNEbuffer containing 1% TX-100 or alternatively it was resuspendedwith the same volume of ASB-14, a zwitterionic amidosulfobe-taine detergent with linear 14-C alkyl chains, which is able to

Fig. 6. Recruitment of Hex and Gal in lipid microdomains of PHA/PMA-stimulatedJurkat cells. Resting and 24 h PHA/PMA-stimulated cells (1 � 108) were lysed byusing TNE buffer containing 1% TX-100. The lysates were fractionated by a discontin-uous sucrose-density gradient as reported in the legend of Fig. 1. Eleven fractions werecollected from the top of the gradient. (A) Aliquots of each fraction (1.5 ml) werespotted in an nitrocellulose membrane and the presence of GM1 was revealed bybiotinylated CT-B. Representative blot of three independent experiments is reported.(B) One-fifteenth volume of each fraction was subjected to Western blot analysis forflot-2. Representative Western blotting of three independent experiments is reported.(C) Gradient distribution of Total Hex, Hex A and Gal enzymatic activities of resting andstimulated cells were expressed as total mU (tot. mU) in each fraction. Graphics (topright) show the enzymatic activity recovery in fractions 8e11 for resting and stimu-lated cells. Values are the mean � SD of three independent experiments. LM, Lipidmicrodomains; H, High-density fractions.

Fig. 7. Cell stimulation induces lysosomal content secretion. Jurkat cells were incu-bated for 2 h at 37 �C in culture medium containing 2 mg/ml HRP, rinsed and thenchased for 16 h in fresh culture medium. Thereafter, cells were stimulated for 24 h asdescribed in “Materials and methods” and the culture medium was recovered for HRPactivity determination using ABTS as substrate. Histograms show the % of HRP enzymeactivity in stimulated cells culture medium with respect to resting cells culturemedium. Values are the mean � SD of five independent experiments. *, P < 0.01(PHA/PMA or PHA-treated cells vs resting cells) according to unpaired two-tailedStudent’s t test.

A. Magini et al. / Biochimie 94 (2012) 684e694 691

Author's personal copy

solubilize the major part of lipid microdomain structureswithout affecting proteineprotein interactions [40].

To determine if lck-decorated microdomains contained Hexand Gal glycohydrolases we used a polyclonal anti-lck antibodyadsorbed to protein G magnetic beads. Immunoprecipitation of

lipid microdomain complexes was validated by immunoblottingusing anti-flot-2 antibody. Results reported in Fig. 8A show thepresence of flot-2 signal in TX-100 immunoprecipitates (IP)whereas minor quantities of flot-2 were detectable in TX-100supernatant (IP-Super). Since ASB-14 completely dissolved lipidmicrodomain complexes, flot-2 signal was present only in ASB-14supernatant and no flot-2 signal was found in the lck immuno-precipitated fraction (data not shown). Immunoblotting alsoconfirmed that cell stimulation leads to an increased flot-2recruitment to lipid microdomains. As reported in Fig. 8B, aftersolubilizing the membranes with TX-100 and immunoprecipita-tion with the anti-lck antibody, about 45% of Total Hex and Galactivities, and 30% of Hex A activity from both resting and stim-ulated cells, were recovered adsorbed to protein G magneticbeads. The remaining activity was recovered in TX-100 super-natant. Moreover, results reported in Fig. 8B show that Hex andGal activities are totally recovered in ASB-14 supernatant. Theseactivities were not detected in the lck immunoprecipitatesfrom extracts that were prepared in the presence of ASB-14.Beads incubated without anti-lck antibody as control showedno microdomain immobilization (data not shown). Datareported above indicate that Hex and Gal reside in lck-decoratedmicrodomain vesicles and demonstrate that lck is not directlycoupled to these glycohydrolases by proteineprotein interac-tions, although it is possible that extraction of the lipid bilayermay disrupt low-affinity proteineprotein interactions.

4. Discussion and conclusions

The present work provides evidence for the presence of thefully processed glycohydrolases Hex and Gal, both involved in thestepwise degradation of GM1 to GM3 ganglioside, within cellsurface TX-100 insoluble lipid microdomains in resting and stim-ulated Jurkat T-lymphocytes. Furthermore, Jurkat cell stimulationup-regulates the expression and activity of Hex and Gal andincreases their targeting to lipid microdomains. Hex and Gal actmainly inside endosomal/lysosomal environment, where theyparticipate to the degradation of a wide range of glycoconjugatedsubstrates [41]. However, recent studies have demonstrated thepresence of glycohydrolases at the plasma membrane, pointingto their potential role in modifying glycosphingolipids duringsignalling and cellular communication [4e8]. In particular, thetranslocation of sialidase Neu1 from lysosomes to plasmamembrane has been observed during T-cell activation [42,43] andmonocyte differentiation [44]. Moreover, it has been showed thatsequential activity of Neu3, Gal and b-glucocerebrosidase at thecell surface level of human fibroblasts lead to the production ofceramide from GM3 [45]. Recently, the presence and the behav-iour of Neu3, Gal and b-glucocerebrosidase activities at the plasmamembrane of rat cerebellar granule cells have been reported [46].In a previous paper [10], we also identified a mature form of Hexassociated to the plasma membrane of human cells, and demon-strated its ability to hydrolyse the natural substrate GM2 in vitro.A similar form was also found to be associated to the lysosomalmembrane [11].

GSLs are components of the external leaflet of the plasmamembrane and are particularly enriched in lipid microdomains.Proteins contained in lipid microdomains play important roles incell signalling [47]. Assuming that lysosomal glycolipid metabo-lizing enzymes may play an active role in in situ remodellingglycolipid composition and pattern of microdomains, we investi-gated in detail the presence of acidic glycohydrolases in lipidmicrodomains, using Jurkat T-lymphocytes as cell model. In T-cells,lipid microdomains play a fundamental role. Interaction of T-cellswith the specific antigen is accompanied by the polarized

Fig. 8. Lipid microdomain-associated Hex and Gal are co-immunoprecipitated by anti-lckantibody. Lipid microdomain (LM) pellets were prepared from 5 � 107 resting (�) and24 h PHA-treated (þ) Jurkat cells as described in the “Materials and methods”. Afterresuspension in TNE/1% TX-100 or TNE/1% ASB-14, lipid microdomain pellets were incu-bated with 2 mg polyclonal anti-lck antibody for 3 h before the addition of 50 ml Dynabeadsprotein G for 3 h. Immunoprecipitates were washed five times and disaggregated bytreatmentwith TNE/1%ASB-14. Supernatants andsolubilizedmicrodomainswereusedbothfor immunoblotting with antibody to flot-2 as lipid microdomain control (A) and to assayHex andGal activities expressed as totalmU (tot.mU) (B). Values are themean� SD of threeindependent experiments.

A. Magini et al. / Biochimie 94 (2012) 684e694692

Author's personal copy

recruitment of lipid microdomains to form immunological synapsethat facilitate intermolecular interaction and the propagation ofsignalling transduction pathways from the site of TCR [26,48]: thisis a very complex event and not all its molecular details are known.Gangliosides such as GM2 and GM3 play an important role ina large variety of cellular functions such as cell differentiation andproliferation and cellecell interaction in both normal and patho-logical conditions [49e51]. In T-lymphocytes, GM3 is the majorganglioside of cell surface, mainly concentrated in lipid micro-domains, also defined “GM3 enriched domains”, and interacts withthe molecules involved in signal transduction mechanisms duringlymphocyte activation [27].

Analysis of Jurkat cell lipid microdomains purified by usingdifferent non-ionic detergents at 4 and 37 �C and a detergent-freemethod provides evidence for the presence of both Hex A andHex B isoenzymes as well as Gal enzymatic activities in flot-2positive fractions. The finding that these enzymes are associatedto lipid microdomains also using different extraction methods hasallowed us to conclude that the presence of Hex and Gal in lipidmicrodomains is not due to artifacts generated by the presence ofdetergents or the use of non-physiological temperatures. Therefore,further characterization analysis was carried out using the deter-gent TX-100 that ensured a higher stringency to exclude non-lipidmicrodomain protein contamination. In this experimental condi-tion the obtained lipid microdomain fraction 3 was not contami-nated by proteins from early endosomes, lysosomes and cytosol asattested by using specific markers. Moreover, the association ofmature Hex a-subunit and Gal to plasma membrane-lipid micro-domains was definitively confirmed by using in vivo cell surfacebiotinylation that permitted to obtain highly purified microdomainproteins exposed to the extracellular environment. After de novosynthesis, Hex and Gal, like the other lysosomal proteins, aredelivered from TGN to lysosomes, where they undergo to the finalproteolytic processing. Therefore the presence of these glycohy-drolases in the mature processed form in cell surface lipid micro-domains implies their lysosomal origin. Secretion experimentindicates the existence of a lysosome-to-plasma membrane trans-port pathway that may mediate the translocation of these lyso-somal enzymes to the cell surface.

Stimulation of Jurkat cells up-regulates the expression andactivity of Hex and Gal and increases their targeting to lipidmicrodomains. However, the recruitment of Hex and Gal in lipidmicrodomains observed after cell stimulation appeared moresignificant when compared to the enhancement of Hex and Galcellular activities in the same conditions. This indicated a specificrecruitment of these enzymes to the plasma membrane duringT-cell stimulation. Lipid microdomains are stabilized by additionof cholesterol, and usually increased protein solubility observedin TX-100 resistant membrane after cholesterol depletion in thepresence of MbCD provides a more stringent test of proteinassociation to lipid microdomains [52]. In our case, microdomainperturbation by cholesterol depletion significantly reduced theHex and Gal activities found in lipid microdomains of resting andstimulated Jurkat cells. All these data support the conclusion thatHex and Gal can be considered a bona fide associated to lipidmicrodomains, excluding an eventual contamination from othercellular compartments, i.e. early/late endosomes and lysosomes.In addition, the specific recruitment to the plasma membrane ofHex and Gal was further characterized: i) the time-coursemonitoring of T-cell stimulation showed that the process ofglycohydrolases recruitment to the plasma membrane is a quiterapid event, as it appears after 6 h and increases up to 24 h; ii) theenhancement of T-cell stimulation by combined PHA/PMAtreatment resulted in a major recruitment of Hex and Gal in flot-2positive fractions, with respect to cells stimulated only with PHA.

The tyrosine kinase lck is essential for T-cell development and T-cell receptor signalling. It is a Src-family kinase that connects theT-cell antigen receptor to Syk kinase ZAP-70. Upon stimulation,lck is massively translocated to the lipid microdomains. Glyco-hydrolases activity recovered in TX-100 immunoprecipitatesobtained using anti-lck antibody provided a further evidence ofthe association of Hex and Gal to the same lipid microdomainvesicles where lck is located, indicating that these two enzymesare recruited in specific lipid microdomains where proteinsinvolved in signalling pathways coalesce to form the immuno-logical synapse. However, the results of ASB-14 immunoprecip-itation experiments excluded that lck is directly bound to the Hexand Gal glycohydrolases. In fact, the use of zwitterionic amido-sulfobetaine ASB-14 detergent to solubilize the isolated lipidmicrodomains interfered with the isolation of Hex and Gal alongwith lck. Therefore a direct interaction between lck and theseglycohydrolases was not likely.

Overall results provide evidence that glycohydrolases Hex andGal, which are deputed to hydrolyse series “a” monosialoganglio-sides, are not randomly distributed on the cell surface but localisein specialized areas of the plasma membrane external site, wherethey may have a role in the local GSLs composition rearrangementthat accompany many important cellular processes.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgements

This work was supported by “Fondazione Cassa di Risparmio diPerugia” (2008.021.375) grant to CE.

References

[1] A.R. Todeschini, S. Hakomori, Functional role of glycosphingolipids andgangliosides in control of cell adhesion, motility, and growth, through gly-cosynaptic microdomains, Biochim. Biophys. Acta 1780 (2008) 421e433.

[2] P.H. Lopez, R.L. Schnaar, Gangliosides in cell recognition and membraneprotein regulation, Curr. Opin. Struct. Biol. 19 (2009) 549e557.

[3] S. Hakomori, Y. Igarashi, Functional role of glycosphingolipids in cell recog-nition and signalling, J. Biochem. 118 (1995) 1091e1103.

[4] P.M. Crespo, V.T. Demichelis, J.L. Daniotti, Neobiosynthesis of glyco-sphingolipids by plasma membrane-associated glycosyltransferases, J. Biol.Chem. 285 (2010) 29179e29190.

[5] E. Monti, M.T. Bassi, N. Papini, M. Riboni, M. Manzoni, B. Venerando, G. Croci,A. Preti, A. Ballabio, G. Tettamanti, G. Borsani, Identification and expression ofNEU3, a novel human sialidase associated to the plasma membrane, Biochem.J. 349 (2000) 343e351.

[6] M. Aureli, A.P. Masilamani, G. Illuzzi, N. Loberto, F. Scandroglio, A. Prinetti,V. Chigorno, S. Sonnino, Activity of plasma membrane beta-galactosidase andbeta-glucosidase, FEBS Lett. 583 (2009) 2469e2473.

[7] S. Sonnino, M. Aureli, N. Loberto, V. Chigorno, A. Prinetti, Fine tuning of cellfunctions through remodeling of glycosphingolipids by plasma membrane-associated glycohydrolases, FEBS Lett. 584 (2010) 1914e1922.

[8] M. Aureli, N. Loberto, V. Chigorno, A. Prinetti, S. Sonnino, Remodeling ofsphingolipids by plasma membrane associated enzymes, Neurochem. Res.(Dec 23, 2010). doi:10.1007/s11064-010-0360-7.

[9] D.J. Mahuran, Biochemical consequences of mutations causing the GM2 gan-gliosidoses, Biochim. Biophys. Acta 1455 (1999) 105e138.

[10] S. Mencarelli, C. Cavalieri, A. Magini, B. Tancini, L. Basso, P. Lemansky, A. Hasilik,Y.T. Li, V. Chigorno, A. Orlacchio, C. Emiliani, S. Sonnino, Identification of plasmamembrane associated mature beta-hexosaminidase A, active towards GM2ganglioside, in human fibroblasts, FEBS Lett. 579 (2005) 5501e5506.

[11] A. Magini, S. Mencarelli, B. Tancini, V. Ciccarone, L. Urbanelli, A. Hasilik,C. Emiliani, Identification and characterization of mature beta-hexosaminidases associated with human placenta lysosomal membrane, Bio-sci. Rep. 28 (2008) 229e237.

[12] S. Okada, J.S. O’Brien, Generalized gangliosidosis: beta-galactosidase defi-ciency, Science 160 (1968) 1002e1004.

[13] Y. Suzuki, H. Sakuraba, A. Oshima, Beta-galactosidase deficiency (beta-galactosidosis): GM1 gangliosidosis and Morquio B disease. in:

A. Magini et al. / Biochimie 94 (2012) 684e694 693

Author's personal copy

C.R. Scriver, A.L. Beaudet, W.S. Sly, D. Valle (Eds.), The Metabolic andMolecular Bases of Inherited Disease. McGraw-Hill Publishing Co., NewYork, 1995, pp. 2787e2823.

[14] L. Urbanelli, A. Magini, A. Polchi, M. Polidoro, C. Emiliani, Recent develop-ments in therapeutic approaches for lysosomal storage diseases, Recent Pat.CNS Drug Discov. 6 (2011) 1e19.

[15] S. Hakomori, Glycosilation defining cancer malignancy: new wine in an oldbottle, Proc. Natl. Acad. Sci. U S A 99 (2002) 10231e10233.

[16] S. Hakomori, Cancer-associated glycosphingolipid antigens: their structure,organization, and function, Acta Anat. (Basel) 161 (1998) 79e90.

[17] S.A. Saddoughi, P. Song, B. Ogretmen, Roles of bioactive sphingolipids incancer biology and therapeutics, Subcell. Biochem. 49 (2008) 413e440.

[18] T. Ariga, M.P. McDonald, R.K. Yu, Role of ganglioside metabolism in thepathogenesis of Alzheimer’s disease, J. Lipid Res. 49 (2008) 1157e1175.

[19] T. Yamashita, Y.P. Wu, R. Sandhoff, N. Werth, H. Mizukami, J.M. Ellis,J.L. Dupree, R. Geyer, K. Sandhoff, R.L. Proia, Interruption of gangliosidesynthesis produces central nervous system degeneration and altered axon-glial interactions, Proc. Natl. Acad. Sci. U S A 102 (2005) 2725e2730.

[20] I. Jou, J.H. Lee, S.Y. Park, H.J. Yoon, E.H. Joe, E.J. Park, Gangliosides triggerinflammatory responses via TLR4 in brain glia, Am. J. Pathol. 168 (2006)1619e1630.

[21] K. Simons, E. Ikonen, Functional rafts in cell membranes, Nature 387 (1997)569e572.

[22] W. Rodgers, J.K. Rose, Exclusion of CD45 inhibits activity of p56lck associatedwith glycolipid-enriched membrane domains, J. Cell. Biol. 135 (1996)1515e1523.

[23] P. Drevot, C. Langlet, X.J. Guo, A.M. Bernard, O. Colard, J.P. Chauvin, R. Lasserre,H.Y. He, TCR signal initiation machinery is pre-assembled and activated ina subset of membrane rafts, EMBO J. 21 (2002) 1899e1908.

[24] B. Brügger, C. Graham, I. Leibrecht, E. Mombelli, A. Jen, F. Wieland, R. Morris,The membrane domains occupied by glycosylphosphatidylinositol-anchoredprion protein and Thy-1 differ in lipid composition, J. Biol. Chem. 279(2004) 7530e7536.

[25] M. Martin, H. Schneider, A. Azouz, C.E. Rudd, Cytotoxic T lymphocyte antigen4 and CD28 modulate cell surface raft expression in their regulation of T cellfunction, J. Exp. Med. 194 (2001) 1675e1681.

[26] C.R. Monks, B.A. Freiberg, H. Kupfer, N. Sciaky, A. Kupfer, Three-dimensionalsegregation of supramolecular activation clusters in T cells, Nature 395 (1998)82e86.

[27] T. Garofalo, L. Lenti, A. Longo, R. Misasi, V. Mattei, G.M. Pontieri, A. Pavan,M. Sorice, Association of GM3 with Zap-70 induced by T cell activation inplasma membrane microdomains: GM3 as a marker of microdomains inhuman lymphocytes, J. Biol. Chem. 277 (2002) 11233e11238.

[28] S. Tani-ichi, K. Maruyama, N. Kondo, M. Nagafuku, K. Kabayama, J. Inokuchi,Y. Shimada, Y. Ohno-Iwashita, H. Yagita, S. Kawano, A. Kosugi, Structure andfunction of lipid rafts in human activated T cells, Int. Immunol. 17 (2005)749e758.

[29] C. Gómez-Moutón, R.A. Lacalle, E. Mira, S. Jiménez-Baranda, D.F. Barber,A.C. Carrera, C. Martínez-A, S. Mañes, Dynamic redistribution of raft domainsas an organizing platform for signaling during cell chemotaxis, J. Cell. Biol. 164(2004) 759e768.

[30] A. Viola, S. Schroeder, Y. Sakakibara, A. Lanzavecchia, T lymphocyte cos-timulation mediated by reorganization of membrane microdomains, Science283 (1999) 680e682.

[31] B. Tancini, A. Magini, L. Latterini, L. Urbanelli, V. Ciccarone, F. Elisei, C. Emiliani,Occurrence of an anomalous endocytic compartment in fibroblasts fromSandhoff disease patients, Mol. Cell. Biochem. 335 (2010) 273e282.

[32] M.M. Bradford, A rapid and sensitive method for the quantitation of micro-gram quantities of protein utilizing the principle of protein-dye binding, Anal.Biochem. 72 (1976) 248e254.

[33] U.K. Laemmli, Cleavage of structural protein during the assembly of the headof bacteriophage T4, Nature 227 (1970) 680e685.

[34] Y. Sasaki, Y. Oshima, R. Koyama, R. Maruyama, H. Akashi, H. Mita, M. Toyota,Y. Shinomura, K. Imai, T. Tokino, Identification of flotillin-2, a major protein onlipid rafts, as a novel target of p53 family members, Mol. Cancer Res. 6 (2008)395e406.

[35] H. Tanaka, K. Suzuki, Substrate specificities of the two genetically distincthuman brain beta-galactosidases, Brain Res. 122 (1977) 325e335.

[36] S. Schuck, M. Honsho, K. Ekroos, A. Shevchenko, K. Simons, Resistance of cellmembranes to different detergents, Proc. Natl. Acad. Sci. U S A 100 (2003)5795e5800.

[37] A. Weiss, J. Imboden, D. Shoback, J. Stobo, Role of T3 surface molecules inhuman T-cell activation: T3-dependent activation results in an increase incytoplasmic free calcium, Proc. Natl. Acad. Sci. U S A 81 (1984) 4169e4173.

[38] R. Tavano, G. Gri, B. Molon, B. Marinari, C.E. Rudd, L. Tuosto, A. Viola, CD28 andlipid rafts coordinate recruitment of Lck to the immunological synapse ofhuman T lymphocytes, J. Immunol. 173 (2004) 5392e5397.

[39] S. Danielian, R. Fagard, A. Alcover, O. Acuto, S. Fischer, The lymphocyte-specific protein tyrosine kinase p56lck is hyperphosphorylated on serine andtyrosine residues within minutes after activation via T cell receptor or CD2,Eur. J. Immunol. 19 (1989) 2183e2189.

[40] M. Chevallet, V. Santoni, A. Poinas, D. Rouquié, A. Fuchs, S. Kieffer,M. Rossignol, J. Lunardi, J. Garin, T. Rabilloud, New zwitterionic detergentsimprove the analysis of membrane proteins by two-dimensional electro-phoresis, Electrophoresis 19 (1998) 1901e1909.

[41] R.A. Gravel, J.T.R. Clarke, M.M. Kaback, D.J. Mahuran, K. Sandhoff, K. Suzuki, in:C.R. Scriver, A.L. Beaudet, W.S. Sly, D.L. Valle (Eds.), The metabolic basis ofinherited disease, McGraw-Hill Publishing Co, New York, 1999, pp.2839e2879.

[42] X.P. Chen, E.Y. Enioutina, R.A. Daynes, The control of IL-4 gene expression inactivated murine T lymphocytes: a novel role for neu-1 sialidase, J. Immunol.158 (1997) 3070e3080.

[43] X.P. Chen, X. Ding, R.A. Daynes, Ganglioside control over IL-4 priming andcytokine production in activated T cells, Cytokine 12 (2000) 972e985.

[44] F. Liang, V. Seyrantepe, K. Landry, R. Ahmad, A. Ahmad, N.M. Stamatos,A.V. Pshezhetsky, Monocyte differentiation up-regulates the expression of thelysosomal sialidase, Neu1, and triggers its targeting to the plasma membranevia major histocompatibility complex class II-positive compartments, J. Biol.Chem. 281 (2006) 27526e27538.

[45] R. Valaperta, V. Chigorno, L. Basso, A. Prinetti, R. Bresciani, A. Preti, T. Miyagi,S. Sonnino, Plasma membrane production of ceramide from ganglioside GM3in human fibroblasto, FASEB J. 20 (2006) 1227e1229.

[46] M. Aureli, N. Loberto, P. Lanteri, V. Chigorno, A. Prinetti, S. Sonnino, Cellsurface sphingolipid glycohydrolases in neuronal differentiation and aging inculture, J. Neurochem. 116 (2011) 891e899.

[47] D. Lingwood, K. Simons, Lipid rafts as a membrane-organizing principle,Science 327 (2010) 46e50.

[48] W.R. Burack, K.H. Lee, A.D. Holdorf, M.L. Dustin, A.S. Shaw, Cutting edge:quantitative imaging of raft accumulation in the immunological synapse,J. Immunol. 169 (2002) 2837e2841.

[49] A.R. Todeschini, J.N. Dos Santos, K. Handa, S.I. Hakomori, Ganglioside GM2-tetraspanin CD82 complex inhibits met and its cross-talk with integrins,providing a basis for control of cell motility through glycosynapse, J. Biol.Chem. 282 (2007) 8123e8133.

[50] M. Sorice, A. Longo, T. Garofalo, V. Mattei, R. Misasi, A. Pavan, Role of GM3-enriched microdomains in signal transduction regulation in T lymphocytes,Glycoconj. J. 20 (2004) 63e70.

[51] P. Wang, P. Wu, J. Zhang, T. Sato, S. Yamagata, T. Yamagata, Positive regulationof tumor necrosis factor-alpha by ganglioside GM3 through Akt in mousemelanoma B16 cells, Biochem. Biophys. Res. Commun. 356 (2007) 438e443.

[52] L. Bini, S. Pacini, S. Liberatori, S. Valensin, M. Pellegrini, R. Raggiaschi, V. Pallini,C.T. Baldari, Extensive temporally regulated reorganization of the lipid raftproteome following T-cell antigen receptor triggering, Biochem. J. 369 (2003)301e309.

A. Magini et al. / Biochimie 94 (2012) 684e694694