Archives of Biochemistry and Biophysics 436 (2005) 323–332

www.elsevier.com/locate/yabbi

0003-9861/$ - see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.abb.2005.01.013

Cathepsin X binds to cell surface heparan sulfate proteoglycans

Fábio D. Nascimento b, Claudia C.A. Rizzi a, Iseli L. Nantes a, Ivica Stefe c, Boris Turk c, Adriana K. Carmona d, Helena B. Nader b, Luiz Juliano d, Ivarne L.S. Tersariol a,¤

a Centro Interdisciplinar de Investigação Bioquímica (CIIB), Universidade de Mogi das Cruzes, São Paulo, Brazilb Departamento de Bioquímica, Universidade Federal de São Paulo, Escola Paulista de Medicina, São Paulo, Brazil

c Department of Biochemistry and Molecular Biology, Joqef Stefan Institute, Ljubjana, Sloveniad Departamento de Biofísica, Universidade Federal de São Paulo, Escola Paulista de Medicina, São Paulo, Brazil

Received 13 December 2004, and in revised form 14 January 2005Available online 2 February 2005

Abstract

Glycosaminoglycans have been shown to be important regulators of activity of several papain-like cathepsins. Binding of glycos-aminoglycans to cathepsins thus directly aVects catalytic activity, stability or the rate of autocatalytic activation of cathepsins. Theinteraction between cathepsin X and heparin has been revealed by aYnity chromatography using heparin–Sepharose. Conforma-tional changes were observed to accompany heparin–cathepsin X interaction by far UV-circular dichroism at both acidic (4.5) andneutral (7.4) pH. These conformational changes promoted a 4-fold increase in the dissociation constant of the enzyme-substrateinteraction and increased 2.6-fold the kcat value also. The interaction between cathepsin X and heparin or heparan sulfate is speciWcsince dermatan sulfate, chondroitin sulfate, and hyaluronic acid had no eVect on the cathepsin X activity. Using Xow cytometrycathepsin X was shown to bind cell surface heparan sulfate proteoglycans in wild-type CHO cells but not in CHO-745 cells, whichare deWcient in glycosaminoglycan synthesis. Moreover, Xuorescently labeled cathepsin X was shown by confocal microscopy to beendocytosed by wild-type CHO cells, but not by CHO-745 cells. These results demonstrate the existence of an endocytosis mecha-nism of cathepsin X by the CHO cells dependent on heparan sulfate proteoglycans present at the cell surface, thus strongly suggest-ing that heparan sulfate proteoglycans can regulate the cellular traYcking and the enzymatic activity of cathepsin X. 2005 Elsevier Inc. All rights reserved.

Keywords: Heparin; Heparin sulfate; Cathepsin X; Lysosomal enzyme; Cysteine proteinase

Heparan sulfate is a ubiquitous glycosaminoglycan ofanimal cells [1]. These classes of compounds are hetero-polysaccharides made up of repeating units of disaccha-rides, an uronic acid residue, either D-glucuronic acid orL-iduronic acid, and D-glucosamine with N- and 6-O-sul-fates and N-acetyl substitutions [2]. Heparan sulfate ispresent at the cell surface and in extracellular matrix inthe form of proteoglycan. Most of cellular heparan sul-fate derives from the syndecans and the glypicans. Thesyndecan family is associated with cell membranes viathe transmembrane core proteins [3,4], whereas the

glypican family is anchored by glycosilyl phosphatidyl-inositol-anchor core proteins [5]. In addition, cells pro-ducing basement membranes secrete heparan sulfateproteoglycans such as perlecan and agrin [6]. Anotherglycosaminoglycan, closely related to heparan sulfate, isheparin. Collectively, all these heparin/heparin sulfate-like glycosaminoglycans play a complex role inregulation of a wide variety of biological processes,including those in the extracellular matrix, such ashemostasis, inXammation, angiogenesis, growth factors,cell adhesion, and many others [7].

Several of these regulatory functions ofglycosaminoglycans have been achieved through theinteraction with proteases. There is considerable

¤ Corresponding author. Fax: +55 11 4798 7102.E-mail address: ivarne@umc.br (I.L.S. Tersariol).

324 F.D. Nascimento et al. / Archives of Biochemistry and Biophysics 436 (2005) 323–332

evidence that heparin-like glycosaminoglycans can mod-ulate the activity of various serine proteinases and theirnatural inhibitors, with thrombin–antithrombin interac-tion as the best characterized example [8–11]. In addi-tion, there is growing evidence that glycosaminoglycansalso interact with cysteine proteases, primarily thosefrom the papain family. The lysosomal cysteine protein-ases are implicated in a variety of pathologies includingthose involving tissue remodeling states, such as inXam-mation [12], parasite infection [13], and tumor metastasis[14]. Heparin and heparin sulfate have thus been shownto bind to papain and cathepsin B, thereby substantiallystabilizing the enzymes at neutral pH [15,16]. Moreover,collagenolytic activity of cathepsin K is critically depen-dent on the interaction with chondroitin sulfate presentin bone matrix [17,18], whereas collagenolytic activitiesof cathepsins L and S are potently inhibited by glycos-aminoglycans [19].

Human cathepsin X is a recently discovered cysteineproteinase belonging to the papain superfamily [20,21],which was identiWed from the EST1 database and wasobtained by PCR ampliWcation from a human ovarycDNA library [20]. The enzyme is primarily a carboxy-monopeptidase and diVers from other related enzymesby the presence of a three-residue insertion, called the“mini-loop,” which is responsible for the exopeptidaseactivity of this enzyme [22–26]. In addition, several spe-ciWc interaction motifs, such as an integrin-binding motifArg-Gly-Asp (RGD), found in proteins involved in celladhesion like Wbronectin [27,28], and a Glu-Cys-Aspsequence, a recognition site for �2�1-integrins present inthe type I collagen [29,30], were identiWed in the cathep-sin X sequence. Recently, it had been showed thatcathepsin X, also termed cathepsin Z-1 in Caenorhabditiselegans, has a critical role in nematode morphogenesisby controlling collagen base membrane matrix extracel-lular assemble process [31].

In view of previous results, heparan sulfate proteogly-cans seem to be important physiological ligands andactivity regulators of cysteine cathepsins at the cell sur-face and basement membrane. Therefore, we decided toinvestigate the interaction between cathepsin X and gly-cosaminoglycans with the aim to further understand thebiological role of this new cysteine proteinase. Here, wehave shown that cathepsin X binds to the heparin aYn-ity column with considerable aYnity and characterizedthe interaction between heparin and cathepsin X by cir-cular dichroism and enzymatic assays. Binding to cellsurface glycosaminoglycans was also conWrmed in a

cellular model, suggesting a role for heparan sulfateglycosaminoglycans in intracellular traYcking ofcathepsin X.

Experimental procedures

Enzymes

Human cathepsin X clone (IMAGp998J24840) inpT7T3D vector was obtained from RZPD GermanResource Center for Genome Research (Berlin, Heidel-berg, Germany). The cDNA sequence of cathepsin Xwas then veriWed using T7 and T3 sequencing primersand automated DNA sequencer Abi Prism310 geneticanalyzer (Perkin-Elmer, USA). The correct cDNAsequence coding for human procathepsin X was clonedinto the pPIC9 expression vector (Invitrogen, USA)using AvrII and NotI sites. Recombinant human cathep-sin X was then expressed in yeast Pichia pastoris andactivated using cathepsin L essentially as described [23],except that in the last puriWcation step cathepsin X wasseparated from cathepsin L by ion exchange chromatog-raphy on SP Sepharose Fast Flow column (Amersham–Pharmacia Biotech, USA) at pH 4.4. BrieXy, the samplewas dialyzed against 50 mM acetate buVer, pH 4.4, con-taining 1 mM EDTA and applied to the column equili-brated with the same buVer. Cathepsin X eluted in thenon-bound fractions, whereas cathepsin L bound to thecolumn and was eluted with 0.8 M NaCl. Active cathep-sin X was pooled, concentrated, and stored at ¡70 °C.For cellular experiments, recombinant human cathepsinX was conjugated with Alexa Fluor 488 using the AlexaFluor 488 Protein labeling kit A-10235 from MolecularProbes (Eugene, USA).

Other materials

We have used a size-deWned (10 kDa) bovine lungheparin (The Upjohn), prepared by size exclusion chro-matography [2,7]; heparan sulfate (16 kDa) from bovinelung was a generous gift from Dr. P. Bianchini (OpocrinResearch Laboratories, Modena, Italy) [2]; hyaluronicacid, dermatan sulfate (12 kDa), and chondroitin sulfate(25 kDa) were purchased from Seikagaku Kogyo(Tokyo, Japan); heparin–Sepharose resin was purchasedfrom Amersham–Pharmacia Biotech (USA); polyclonalimmunospeciWc antibodies against human cathepsin Xwere prepared as described previously [25]; the lys-osomal marker Lyso Tracker Red DND-99 and FITC-labeled goat anti-rabbit secondary antibody werepurchased from Molecular Probes (Eugene, USA);Fluormont-G was purchased from Electron MicroscopySciences (Ft. Washington, PA, USA); and the irrevers-ible inhibitor E-64 and saponin were purchased fromSigma.

1 Abbreviations used: AFU, arbitrary Xuorescent units, EST,expressed sequence tags; Abz-FQK(Dnp)-OH, ortho-aminobenzoyl-L-phenylalanyl-L-glutamyl-L-lysine-�-N-2,4-dinitrophenyl; E-64, trans-epoxysuccinyl-L-leucyl-amido-(4-guanidino)butane; DTT, dithiothreitol;CA030, ethyl ester of epoxysuccinyl-Ile-Pro-OH; CA074, n-propyl-epox-ysuccinyl-Ile-Pro; ADAM, A disintegrin and metalloprotease.

F.D. Nascimento et al. / Archives of Biochemistry and Biophysics 436 (2005) 323–332 325

Substrate of cathepsin X

In the present work, we followed the cathepsin X carb-oxymonopeptidase activity using the intramolecularlyquenched Xuorogenic substrate Abz-FQK(Dnp)-OHcontaining the Dnp group incorporated to the �-NH2 ofa Lys residue. This peptide was synthesized by the solid-phase methodology, using Fmoc-Lys(Dnp)-OH to intro-duce the quencher group [32]. The substrate stocksolution (1 mM) was prepared in water containing 10%DMSO. The amount of DMSO used in the kinetic assaysnever exceeded 1%.

Cell lines and culture conditions

Wild-type Chinese Hamster Ovarian cells (CHO-K1)and their mutants deWcient in all cellular glycosamino-glycans as a consequence of the removal of xylosyl-transferase (CHO-745) were kindly donated by Prof. Dr.JeVrey D. Esko (Glycobiology Research and TrainingCenter, University of California, San Diego La Jolla,USA). CHO cells were cultured in F-12 medium supple-mented with 10% fetal bovine serum, 10 U/ml penicillin,and 10 �g/ml streptomycin. Cells were cultured in ahumidiWed incubator containing 2.5% CO2 at 37 °C.

Heparin–Sepharose aYnity chromatography

Pre-activated cathepsin X (0.3 �M) was applied on aheparin–Sepharose column (2 ml) previously equili-brated with 0.05 M sodium acetate buVer, pH 4.5. A lin-ear NaCl gradient (0–1 M) was used to elute the boundmaterial at a Xow rate of 0.5 ml/min. The collected frac-tions were monitored for cathepsin X enzymatic activityusing Abz-FQK(Dnp)-OH substrate.

Kinetic measurements

The inXuence of glycosaminoglycans on cathepsin Xmonocarboxypeptidase activity was followedspectroXuorometrically, using the Xuorogenic substrateAbz-FQK(Dnp)-OH at 37 °C in 0.05 M sodium acetatebuVer, pH 4.5, containing 0.2 M NaCl, 1 mM EDTA, and2 mM DTT. Fluorescence intensity was monitored in athermostated Hitachi F-2000 spectroXuorimeter withexcitation and emission wavelengths set at 320 and420 nm, respectively. Prior to the assays, the enzyme wasactivated for 5 min at 37 °C in the assay buVer. Themolar concentration of active cathepsin X used duringkinetic measurements was 2 nM, while the concentrationof active enzyme was determined by titration with theirreversible inhibitor E-64 as described previously [33].For the determination of pH activity proWles, thekinetics of Abz-FQK(Dnp)-OH hydrolysis were per-formed in the absence or in the presence of diVerentheparin concentrations at 37 °C in 0.05 M sodium

formate (pH 2.5–3.5), 0.05 M sodium acetate (pH 3.7–5.7), and 0.05 M sodium phosphate (pH 6.0–7.0),containing 200 mM NaCl, 1 mM EDTA, and 2 mMdithiothreitol. The concentrations of Abz-FQK(Dnp)-OH were kept 10-fold below KM value, and the values ofkcat/KM were determined. The kinetic parameters weredetermined by measuring the initial rate of the hydroly-sis at various substrate concentrations in the presence orabsence of diVerent glycosaminoglycan concentrations.The kinetic model depicted in Eq. (1) can describe theeVect of heparin on the hydrolysis of Abz-FQK(Dnp)-OH by cathepsin X

(1)

where S is Abz-FQK(Dnp)-OH, Hep is heparin, KS is thesubstrate dissociation constant, KH is the heparin appar-ent dissociation constant, � is the factor of the perturba-tion of KS in the presence of heparin, and � is the factorof the perturbation of Vmax (kcat) in the presence of hepa-rin.

Circular dichroism spectrometry

Circular dichroism spectra were recorded in a JASCOJ-810 spectropolarimeter equipped with a thermostaticcell holder. Far UV measurements (260–200 nm) ofcathepsin X (1.5 �M) were performed at 25 °C at a scan-ning rate of 50 nm/min in a 0.1 cm cell. The assays wereperformed at pH 4.5, in 0.05 M sodium acetate, and atpH 7.4, in 0.05 M Tris–HCl buVers, containing 0.2 MNaCl and 1 mM EDTA, in the absence or in the presenceof 40 �M heparin. The observed ellipticity was normal-ized to units of degrees cm2/dmol. Baseline recordingswere routinely made in the presence or in the absence ofheparin, and used to correct cathepsin X spectra, whichwere analyzed for the percentage of secondary structuralelements as described previously [34].

Cell surface labeling with Alexa Fluor488-conjugated cathepsin X

CHO-K1 and CHO-745 cells were harvested in PBSsolution containing trypsin or in PBS solutioncontaining 5 mM EDTA and incubated at 37 °C for15 min, followed by a PBS washing step. CHO cellsharvested by trypsin treatment were prior to the last PBSwashing step incubated with a molar excess of SBTI toneutralize trypsin. Recombinant human cathepsin Xconjugated with Alexa Fluor 488 was diluted to 4 �g/mlwith PBS solution containing 0.1% BSA in the presenceor in the absence of 50 �M heparin and incubated at 4 °Cfor 30 min with CHO cells (condition that preventsendocytosis) followed by FACS analysis.

326 F.D. Nascimento et al. / Archives of Biochemistry and Biophysics 436 (2005) 323–332

Flow cytometry analysis

Flow cytometry was conducted on a FACSCaliburXow cytometer with a 15-mW, 488-nm, air-cooled argon-ion laser and analyzed by the CELLQUEST software(Beckton–Dickinson, USA). Interaction of Alexa488-conjugated cathepsin X with CHO cell surface wasdetermined as the percentage of cells emitting a Xuores-cent signal in FL1. The boundary between cells thatstained positive and negative for the Alexa488-conju-gated cathepsin X was determined according to the Xuo-rescence distribution of positively stained relative tounstained samples.

Localization of endogenous cathepsin X in CHO cells

CHO cells were grown on cover glasses, rinsed withPBS, and incubated at 20 °C for 60 min before loadingwith 0.5 �M Lyso Tracker DND-99 for 30 min at 37 °Cin the medium without serum. Thereafter, cells werechased in the complete culture medium at 37 °C for120 min, followed by PBS washing and Wxation with 2%formaldehyde in PBS (v/v) for 30 min at 37 °C. Afterthis step, cells were permeabilized with 0.01% saponin(v/v) for 5 min at 20 °C. Cells were then incubated with1% BSA in PBS solution (w/v) to prevent non-speciWcbinding. Primary polyclonal rabbit anti-human cathep-sin X antibodies were diluted in PBS containing 0.1%BSA and incubated with cells overnight. After washing,cells were incubated for additional 120 min at 37 °CFITC-labeled goat anti-rabbit antibodies as the second-ary antibody [43]. In some experiments, the nuclei of thecells were also labeled by incorporation of 0.2 �M DAPIfor 15 min at 37 °C. Cells were washed with PBS,mounted upside-down on slides with Fluormontreagent, and observed with a Zeiss LSM 510 confocalmicroscope (Germany).

Endocytosis of Alexa Fluor488-conjugated cathepsin X by CHO cells

For binding and endocytosis studies, CHO cells werekept at 20 °C in F-12 medium supplemented with 10%fetal bovine serum, 10 U/ml penicillin, and 10 �g/mlstreptomycin medium before mounting on microscopeslides and in vivo confocal microscopy at 37 °C. CHOcells were grown on cover glasses, rinsed with PBS, andincubated at 20 °C for 60 min in medium without serum.The endocytic compartments of CHO cells were labeledin vivo with 1 nM Lyso Tracker DND-99 for 30 min at37 °C [35]. Thereafter, the cells were washed with PBSand incubated with 4 �g/ml Alexa Fluor488-conjugatedcathepsin X in PBS solution containing 0.1% BSA for30 min at 37 °C. The Xuorescent signal of Lyso Trackerand Alexa Fluor488-conjugated cathepsin X as well asthe phase contrast micrographs of CHO cells were taken

with confocal laser scanning microscope (LSM 510,Zeiss, Jena, Germany).

Results

Cathepsin X binds heparin–Sepharose

The interaction of cathepsin X with heparin wasobserved by heparin–Sepharose aYnity chromatogra-phy. Cathepsin X was eluted from heparin–Sepharosecolumn at high ionic strength of 0.6 M (Fig. 1). Previousaddition of 40 �M heparin or heparan sulfate to cathep-sin X solution abolished speciWcally this interaction.Other sulfated glycosaminoglycans, such as dermatansulfate, chondroitin sulfate, and hyaluronic acid, werenot capable of dislodging this binding. These resultsshow that the binding of cathepsin X to heparin-like gly-cosaminoglycans is speciWc and seems to be governedmainly by electrostatic interactions. These data led us toinvestigate the possible eVect of heparin, as well as ofother glycosaminoglycans, on the enzymatic activity ofcathepsin X.

The inXuence of heparin upon cathepsin X pH activity proWle

The pH dependence of cathepsin X activity towardthe substrate Abz-FQK(Dnp)-OH and the inXuence ofheparin upon cathepsin X pH activity proWle aredepicted in Fig. 2. The cathepsin X pH activity proWleobserved for the hydrolysis of the substrate Abz-FQK(Dnp)-OH showed the maximum activity at pH4.5. HPLC and mass spectrometry analyses showed thatGln-Lys(Dnp) was the only peptide bond of the sub-strate Abz-FQK(Dnp)-OH cleaved by cathepsin X both

Fig. 1. Cathepsin X binds heparin–Sepharose column. Activatedcathepsin X (0.3 �M) was chromatographed from a heparin–Sepharose column (3 ml), equilibrated in 50 mM sodium acetate buVer,pH 4.5. A linear NaCl gradient (0–1 M) was used to elute the bondedmaterial; cathepsin X was eluted from the column at 0.6 M of ionicstrength.

F.D. Nascimento et al. / Archives of Biochemistry and Biophysics 436 (2005) 323–332 327

in the absence and presence of heparin, showing thatcathepsin X processed this substrate as exopeptidase.The substrate Abz-FQK(Dnp)-OH was hydrolyzed bycathepsin X at pH 4.5 with a kcat value of 4.8 § 0.2 s¡1,KS value of 22 § 1 �M and a kcat/KS value of2.2 £ 105 M¡1 s¡1. Also, this substrate was chosen to fol-low only a single carboxymonopeptidase activitybecause the product Abz-FQ-OH was resistant to fur-ther hydrolysis. In addition, Abz-FQK(Dnp)-OHcovers the S2, S1, and subsites, which are the main

substrate binding sites in papain-like cysteine protein-ases [26]. As shown in Fig. 2, a large eVect of heparinwas observed upon the cathepsin X pH activity proWle.Basically, heparin promoted a general decrease in Abz-FQK(Dnp)-OH second-order kcat/KS rate without shift-ing the pH activity proWle, while the pH optimum of theenzyme was not changed in the presence of heparin;suggesting that the pKs values of the prototropic groupsin cathepsin X active site were not dislocated by heparinbinding.

EVect of heparin on the exopeptidase activity of cathepsin X

Fig. 3 shows the eVect of heparin on the carboxymono-peptidase activity of cathepsin X. The data show that thepresence of heparin in cathepsin X kinetic assays resultedin a 2.6-fold increase in the kcat value (�D2.6§0.1), aswell as in a 3.7-fold increase of the KS value (�D3.7§0.4).The eVect of heparin on the carboxymonopeptidase activ-ity of cathepsin X can be described by a hyperbolic,mixed-type inhibition depicted in Eq. (1), which showsthat the eYciency of the system can be altered by changeseither in KS (� parameter) or in Vmax (� parameter). Theresults show that heparin binds cathepsin X with a disso-ciation constant of KH D16§2�M.

We have tested other glycosaminoglycans in anattempt to probe the speciWcity for cathepsin Xinteraction. Table 1 shows that, besides heparin, onlyheparan sulfate was capable of binding to cathepsin X

Fig. 2. The pH activity proWles of cathepsin X. The kinetics of Abz-FQK(Dnp)-OH hydrolysis were performed in the absence (I) or in thepresence (�) of 40 �M heparin concentration, at 37 °C, in 50 mMsodium formate (pH 2.5–3.7), 50 mM sodium acetate (pH 3.7–5.7), andsodium phosphate (pH 6.0–7.0) containing 200 mM NaCl, 1 mMEDTA, and 2 mM dithiothreitol.

Fig. 3. EVect of heparin on Abz-FQK(Dnp)-OH hydrolysis by cathepsin X. The inXuence of heparin concentration upon cathepsin X exopeptidaseactivity was determined spectroXuorometrically as described under Experimental procedures. Heparin promotes a decrease in the aYnity of cathep-sin X for the substrate (� D 3.4 § 0.4) (A) and also increases cathepsin X kcat values (� D 2.6 § 0.1) (B).

Table 1Kinetic parameters for hydrolysis of Abz-FQK(Dnp)-OH by human cathepsin X in the presence of diVerent glycosaminoglycans at pH 4.5

GAG (40 �M) Molecular mass (Da) Sulfation degree (SO4/disaccharide) kcat (s¡1) KM (�M)

Control 4.8 § 0.2 22 § 1Heparin 10,000 2.5 12 § 1 81 § 8Heparan sulfate 16,000 0.8 10 § 1 76 § 6Chondroitin sulfate 25,000 0.9 4.5 § 0.4 24 § 2Dermatan sulfate 12,000 1.1 4.9 § 0.5 24 § 2Dextran sulfate 5,000 4.0 14 § 2 86 § 9

328 F.D. Nascimento et al. / Archives of Biochemistry and Biophysics 436 (2005) 323–332

and altering its kinetic constants of Abz-FQK(Dnp)-OHhydrolysis, while the data show that the presence of hep-arin or heparan sulfate was able to increase the KM andkcat constant values of substrate hydrolysis simulta-neously. Other glycosaminoglycans tested, namely hyal-uronic acid, chondroitin sulfate, and dermatan sulfate,had no eVect on the enzyme activity. On the other hand,high sulfated dextran polysaccharide showed a very sim-ilar eVect to that observed for heparin on cathepsin Xactivity. These results show that besides structural fea-tures of glycosaminoglycans, the electrostatic eVect isimportant to this interaction.

EVect of heparin on cathepsin X structure

The eVect of heparin on cathepsin X conformation wasexamined by CD spectroscopy. Fig. 4 shows that at pH 7.4the �-helix content is decreased, with a proportionalincrease in the �-sheet content, when compared to cathep-sin X spectrum at pH 4.5. In addition, the presence of hep-arin at pH 4.5 promotes a signiWcant decrease (22%) incathepsin X �-helix content, while its �-sheet content isincreased. The addition of heparin at pH 7.4 does not alterthe �-helix content of the enzyme; however, a signiWcantloss of the �-sheet structure added to an increase of theremaining structure content observed (Fig. 4). The addi-tion of 1 M NaCl induced a spectral change consistentwith the dissociation of the cathepsin X-heparin complex;the spectrum obtained at high ionic strength conditions isessentially identical to the spectrum of cathepsin X alonein the presence of 1M NaCl only (Table 2).

Cathepsin X binds cell surface heparan sulfate proteoglycans

Based on our in vitro data showing that heparin-likeglycosaminoglycans interact with cathepsin X, we testedthe ability of Alexa488-conjugated cathepsin X to bindcell surface heparan sulfate proteoglycans. Flow cytome-

try analysis (Fig. 5) showed that recombinant humancathepsin X conjugated with Alexa Fluor 488 was capa-ble of binding to the surface of the wild-type CHO cells.This interaction seems to be quite speciWc since the pre-incubation of 20-fold molar excess of unlabeled cathep-sin X abolished 99% of Alexa488-conjugated cathepsinX binding to the cell surface of CHO wild-type cells(data not shown). In contrast, CHO-745 cells, which arealmost completely deWcient in glycosaminoglycans(>95%), did not interact with Alexa488-conjugatedcathepsin X. Similarly, preincubation of Alexa488-conjugated cathepsin X with 50 �M heparin abolishedmore than 95% of cathepsin X binding to the cell surfaceof CHO wild-type cells. Also, Alexa488-conjugatedcathepsin X did not bind CHO wild-type cells pretreatedwith mild trypsin, showing that Alexa488-conjugatedcathepsin X binds a cell surface receptor. These datastrongly suggest that heparan sulfate proteoglycans cananchor cathepsin X to the cell surface.

Localization of endogenous cathepsin X in CHO cells

To identify the cellular localization of endogenouscathepsin X, immunolabeling experiments with thecathepsin X-speciWc antibodies were performed. As canbe seen in Fig. 6, the green Xuorescence of the FITC-labeled secondary antibody (Figs. 6A and D) and the redXuorescence of the Lyso Tracker marker (Figs. 6B andE) colocalized within the vesicles concentrated in theperinuclear regions of both CHO cell lines, whereas nocolocalization with nuclear compartment was observed

Fig. 4. EVect of heparin on cathepsin X circular dichroism spectra. Cathepsin X circular dichroism spectra were determined in the absence (I-I) orin the presence of 40 �M heparin (�-�) at pH 4.5, and at pH 7.4 in the absence (�-�) or in the presence of 40 �M heparin (�-�). The cathepsin Xspectra were analyzed for the percentage of secondary structural elements (Table inset).

Table 2EVect of salt upon recombinant human cathepsin X secondary struc-ture at pH 4.5

Enzyme condition �-helix �-sheet (%) Remaining

Control 18.5 34.5 47.01.0 M NaCl 10.4 44.8 44.81.0 M NaCl + heparin 40 �M 11.0 44.0 45.0

F.D. Nascimento et al. / Archives of Biochemistry and Biophysics 436 (2005) 323–332 329

(Figs. 6C and F, orange-yellow signals), implying thatcathepsin X was localized in lysosomes. These resultsshow that defect in glycosaminoglycan synthesis presentin CHO-745 cells did not aVect cathepsin X biosynthesis.Also, it some heterogeneity for labeling cathepsin X vesi-cles was observed since three patterns of vesicular label-ing were detected: (1) vesicles staining for both cathepsinX and lysosomes, (2) vesicles staining for only cathepsinX, and (3) vesicles staining for only lysosomes. The pres-ent results suggest that cathepsin X is present in particu-lar subpopulation of lysosomes in CHO cells.

Endocytosis of Alexa Fluor488-conjugated cathepsin X in CHO cells

In another experiment, exogenous Alexa Fluor 488-conjugated cathepsin X was added to CHO cells at 4 °C

for 30 min. Thereafter, cells were chased in the completeculture medium, followed by PBS washing. Following30 min of incubation at 37 °C, Xuorescently labeledcathepsin X (green Xuorescence) was found to be local-ized inside the CHO-K1 wild-type cells (Fig. 7A),whereas no exogenous cathepsin X was observed insidethe CHO-745 cells (Fig. 7D). Simultaneously, endo-somal/lysosomal cellular vesicles were labeled with LysoTracker (Figs. 7B and E, red signals). A clear colocaliza-tion between cathepsin X and Lyso Tracker wasobserved in CHO-K1 wild-type cells (Fig. 7C, orange-yellow signals), suggesting that exogenous cathepsin Xwas endocytosed/internalized into endosomal/lysosomalvesicles. In contrast, exogenous cathepsin X could not bevisualized in the CHO-745 cells (Fig. 7F). Takentogether, these data show that heparan sulfateproteoglycans can anchor cathepsin X at cell surface andthis interaction may be related to the endocytosis ofcathepsin X.

Discussion

Heparin–cathepsin X interaction was initiallyobserved by aYnity chromatography in heparin–Sepharose resin (Fig. 1). Kinetic assay data have shownthat heparin binding inhibits cathepsin X monocarboxy-peptidase activity (Fig. 2). The eVect of heparin on thecarboxymonopeptidase activity of cathepsin X can bedescribed by a hyperbolic this mixed type inhibition(Fig. 3), the presence of heparin promotes an aYnityreduction of the enzyme for the substrate (�D 3.7 § 0.4),and an increase in catalytic activity (�D 2.6 § 0.1). Theinhibition promoted by heparin upon monocarboxypep-tidase activity of cathepsin X strongly suggests thatheparin interacts with cathepsin X at mini-loop regionsand heparin–cathepsin X complex perturbs the substratebinding to the central �-helix, where the enzyme catalyticsite is located. His23 residue, which is present in cathepsinX mini-loop, is essential for the enzyme carboxypepti-dase activity; His23 anchors, by electrostatic interactions,the C-terminal carboxylate group of the substrate. Inaddition, His23 and Tyr27, which are also present at themini-loop, occlude cathepsin X subsite and determinethe enzyme monocarboxypeptidase activity [23,26]. Simi-larly, cathepsin B dipeptidyl carboxypeptidase activity isdetermined by the presence of two histidines, His110 andHis111, which are located at the occluding loop, andinteract with the C-terminal carboxylate group of thesubstrate [36,37]. It was demonstrated that His111 has afundamental role in cathepsin B–heparin interaction,promoting a linear, competitive inhibition of thedipeptidyl carboxypeptidase activity [16]. Similar tocathepsin B, it is quite possible that His23 is inXuenced,or has a direct participation, in heparin–cathepsin Xinteraction.

Fig. 5. Alexa488-conjugated cathepsin X binds cell surface heparansulfate proteoglycans in CHO cells. Recombinant human cathepsin Xconjugated with Alexa Fluor 488 (4 �g/ml) dissolved in PBS solutioncontaining 0.1% BSA was incubated at 4 °C for 60 min with CHO cells,and then Xow cytometry analysis was performed on collected cells.Untreated CHO-K1 wild-type cells (A), untreated CHO-745 cells (B),both not incubated with Alexa488-conjugated cathepsin X; CHO-K1wild-type cells treated with 4 �g/ml Alexa488-conjugated cathepsin X(C), CHO-745 incubated with 4 �g/ml Alexa488-conjugated cathepsinX (D), CHO-K1 wild-type (E), and CHO-745 (F) cells were preincu-bated for 30 min with 50 �M heparin and then incubated with 4 �g/mlAlexa488-conjugated cathepsin X; and CHO-K1 wild-type (G), andCHO-745 (H) cells were preincubated for 15 min with trypsin and thenincubated with 4 �g/ml Alexa488-conjugated cathepsin X.

330 F.D. Nascimento et al. / Archives of Biochemistry and Biophysics 436 (2005) 323–332

It was also observed that this interaction is quite spe-ciWc; only heparin-like glycosaminoglycans were eVectivein binding to cathepsin X. Other glycosaminoglycans,such as chondroitin sulfate and dermatan sulfate, werenot capable of making interactions with the enzyme(Table 1).

Heparin–cathepsin X interaction was also moni-tored by circular dichroism spectroscopy analysis(Fig. 4). It was shown that the alkaline pH-inducedinactivation of human cathepsin B [38] and cathepsinL [39,40] was shown to be correlated with the loss ofconformational stability. The deprotonation of His199

Fig. 6. Localization of endogenous cathepsin X in CHO cells. Confocal Xuorescence micrographs of formaldehyde-Wxed and saponin-permeabilizedCHO-K1 wild-type (A–C) and CHO-745 (D–F) cells. Single channel Xuorescence of CHO cells, immunolabeled with polyclonal immunospeciWc rab-bit anti-human cathepsin X antibodies, revealed with FITC-labeled goat anti-rabbit secondary antibodies (A and D, green) and CHO endosomes/lysosomes revealed with Lyso Tracker (B and E, red). Merged images and CHO nucleus revealed in blue with DAPI (C and F), orange-yellow signalsare indicative of colocalization. N, nuclei. (For interpretation of the references to colours in this Wgure legend, the reader is referred to the web ver-sion of this paper.)

Fig. 7. Endocytosis of Alexa Fluor 488-conjugated cathepsin X in CHO cells. Confocal Xuorescence micrographs of CHO-K1 wild-type (A–C) andCHO-745 (D–F) cells. Single channel Xuorescence of exogenous Alexa Fluor 488-conjugated cathepsin X (4 �g/ml) added to CHO cell cultures(A and D, green) and CHO endosomes/lysosomes marked with Lyso Tracker (B and E, red), merged imagines (C and F, orange-yellow). Arrowsshow in detail the phase contrast image with merged Xuorescence (lower panel inside of C and F pictures), orange-yellow signals are indicative ofcolocalization. N, nuclei. (For interpretation of the references to colours in this Wgure legend, the reader is referred to the web version of this paper.)

F.D. Nascimento et al. / Archives of Biochemistry and Biophysics 436 (2005) 323–332 331

in cathepsin B, catalyzed by OH¡ ions, is considered acrucial event for alkaline pH-induced inactivation ofthis cysteine proteinase [38]. The break of the thio-late–imidazolium ion pair Cys29-His199 inXuences ion-ization and solvent exposure of some chargedresidues located at the interdomain interface, such asAsp40-Arg202 and Arg41-Glu163, resulting in conforma-tional changes that promote destabilization of thecentral �-helix [38–40].

Fig. 4 shows that the secondary structure of cathepsinX is changed at pH 7.4, with loss of �-helix andconcomitant gain of � structures. The addition ofheparin to cathepsin X at pH 7.4 promotes a decrease of� structures as well as a subsequent elevation of remain-ing structures contents, which shows that, in addition toalkaline pH, the presence of heparin promotes a higherdisorganization of the enzyme. This eVect was the oppo-site of that observed for cathepsin B, where heparin pro-motes the stabilization of the secondary structure,preserving its �-helix content at alkaline pH [16]. Hepa-rin increases the molar ellipticity value of cathepsin XCD spectrum at [�]222nm, which suggests that the pres-ence of heparin promotes a decrease in the �-helix con-tent of the enzyme. This result corroborates to thespectrum deconvolution, which shows that heparin, atpH 4.5, is capable of promoting a decrease in the �-helixcontent of cathepsin X (Fig. 4). Such eVect is the oppo-site to that observed for cathepsin B, where the molarellipticity decreased drastically, showing that the pres-ence of heparin promotes an increase of the �-helix con-tent of the enzyme [16].

As expected, the interaction between cathepsin X andheparin was disrupted by the addition of 1.0 M NaCl. Itwas observed that the CD spectrum obtained in thepresence of 1.0 M NaCl is very similar to the spectrumobtained for the cathepsin B alone in the presence of1.0 M NaCl (Table 2). The data show that cathepsin Xinteraction with heparin is mainly governed by ionicinteractions.

The results suggest that the cathepsin X conforma-tion changes promoted by heparin binding corroboratewith the decrease in the catalytic eYciency of theenzyme, altering its aYnity for the substrate and itscatalysis. Taken together, these data suggest that hepa-rin binding perturbs cathepsin X active site at the samemanner.

Although a heparin binding sequence has not beendemonstrated in cathepsin X yet, the sequence formed byArg43-Ile44-Asn45–Ile46-Lys47-Arg48-Lys49 (RINIKRK) isa candidate heparin binding site [41]. The heparin bindingsite proposed for cathepsin X is located at the C-terminalregion of the central �-helix [24]. The presence of this con-sensus sequence in this region supports the hypothesis thatheparin inXuences cathepsin X binding to the substrate,which is in agreement with kinetic and circular dichroismdata presented in this work.

Heparan sulfate proteoglycans play a major role insignaling from the extracellular matrix in cell adhesion[42]. It has been shown that heparan sulfate proteogly-cans are essential as partners of integrins in focal adhe-sion formation [43,44]. Many proteins can bindheparan sulfate proteoglycans and regulate its functionin cell adhesion, e.g., heparanase [45], transglutaminase[46], laminin alpha3 chain [47], and thrombin [48]. Pro-cathepsin X contains RGD and ECD sequences [20,21].The RGD and ECD sequences serve as binding recog-nition sites for cell-surface integrin receptors and arealso found in metalloproteinases belonging to theADAM family, named disintegrins [49]. In addition,our data show that heparan sulfate proteoglycans arerelated to insertion processes of cathepsin X at cellsurface (Fig. 5). As procathepsin X contains RGD andECD sequences, and matures cathepsin X also interactwith heparan sulfate proteoglycans in living cells, it istempting to hypothesize that interaction of cathepsin Xwith heparin-like glycosaminoglycans may be involvedin the control of cell adhesion processes. The presentdata show that heparan sulfate proteoglycans arerelated to insertion processes of cathepsin X at cell sur-face controlling the enzyme cellular traYcking.

Despite no literature data about exact cellular locali-zation of cathepsin X, the present results show thatcathepsin X is an ER-Golgi-lysosome resident protein.Cathepsin X is localized within the lysosomal vesiclesconcentrated in the perinuclear regions of both CHOcell lines (Fig. 6). In addition to the presence of endoge-nous lysosomal cysteine proteases, many cell types canalso secrete them [14,17–19,35]. Secreted lysosomalenzymes can be reinternalized to lysosomes by the 300kDa M6P-receptors, which are also responsible for thenormal traYcking of newly synthesized lysosomal pro-teases [50]. These receptors seem to function normallyin both CHO-K1 and CHO-745 mutant cells, as endog-enous cathepsin X was found within endosomal/lyso-somal compartments in both cell lines (Fig. 6).Therefore, our data strongly suggest that internaliza-tion of cathepsin X proceeds through diVerent mecha-nism, which is probably dependent on heparan sulfateproteoglycans present at cell surface of CHO cells(Fig. 7). On this basis it can be further suggested thatheparan sulfate proteoglycans are involved in cellulartraYcking of this lysosomal enzyme.

Acknowledgments

This work was supported by Fundação de Amparo àPesquisa do Estado de São Paulo Grants 03/08161-1 andFundação de Amparo ao Ensino e Pesquisa Universid-ade de Mogi das Cruzes, SP, Brazil. We thank Prof. Dr.JeVrey D. Esko (Glycobiology Research and TrainingCenter, University of California, San Diego La Jolla,

332 F.D. Nascimento et al. / Archives of Biochemistry and Biophysics 436 (2005) 323–332

USA) for supplying the CHO-K1 wild-type cells and theCHO-745 cells, defective in glycosaminoglycanbiosynthesis.

References

[1] C.P. Dietrich, H.B. Nader, H.A. Straus, Biochem. Biophys. Res.Commun. 111 (1983) 865–871.

[2] C.P. Dietrich, I.L.S. Tersariol, L. Toma, C.T. Moraes, M.A. Porcion-atto, F.W. Oliveira, H.B. Nader, Cell Mol. Biol. 44 (1998) 417–429.

[3] M. Yanagishita, V.C. Hascall, J. Biol. Chem. 267 (1992) 9451–9454.[4] K. Elenius, M. Jalkanen, J. Cell Sci. 107 (1994) 2975–2982.[5] G. David, FASEB J. 7 (1993) 1023–1030.[6] R. Iozzo, I.R. Cohen, S. Grässel, A.D. Murdock, Biochem. J. 302

(1994) 625–629.[7] H.E. Conrad, Heparin Binding Proteins, Academic Press, New

York, 1998.[8] P.G.W. Gettins, P.A. Patston, S.T. Olson, Serpins: Structure, Func-

tion and Biology, R.G. Landes, Austin, TX, 1996.[9] J. ErmolieV, C. Boudier, A. Laine, B. Meyer, J.G. Bieth, J. Biol.

Chem. 269 (1994) 29502–29508.[10] M.A. Fath, X. Wu, R.E. Hileman, R.J. Linhardt, M.A. Kashem,

R.M. Nelson, C.D. Wright, W. Abraham, J. Biol. Chem. 273 (1998)13563–13569.

[11] V. Kainulainen, H. Wang, C. Schick, M. BernWeld, J. Biol. Chem.273 (1998) 11563–11569.

[12] N. Katunuma, Intracellular Proteolysis, Japan ScientiWc SocietiesPress, Tokyo, Japan, 1989.

[13] E. Del Nery, M.A. Juliano, A.P.C.A. Lima, J. Scharfstein, L. Juli-ano, J. Biol. Chem. 272 (1997) 25713–25718.

[14] D. Cavallo-Medved, B.F. Sloane, Curr. Top. Dev. Biol. 54 (2003) 313–341.

[15] P.C. Almeida, I.L. Nantes, C.C.A. Rizzi, W.A.S. Júdice, J.R. Cha-gas, L. Juliano, H.B. Nader, I.L.S. Tersariol, J. Biol. Chem. 274(1999) 30433–30438.

[16] P.C. Almeida, I.L. Nantes, J.R. Chagas, C.C.A. Rizzi, A. Faljoni-Alário, E. Carmona, L. Juliano, H.B. Nader, I.L.S. Tersariol, J.Biol. Chem. 276 (2001) 944–951.

[17] Z. Li, W.S. Hou, D. Bromme, Biochemistry 20 (2000) 529–536.[18] Z. Li, W.S. Hou, C.R. Escalante-Torres, B.D. Gelb, D. Brömme, J.

Biol. Chem. 32 (2002) 28669–28676.[19] Z. Li, Y. Yasuda, W. Li, M. Bogyo, N. Katz, R.E. Gordon, G.B.

Fields, D. Brömme, J. Biol. Chem. 279 (2004) 5470–5479.[20] I. Santamaria, G. Velasco, M. Cazorla, A. Fueyo, E. Campo, C.

López-Otín, Cancer Res. 58 (1998) 1624–1630.[21] D.K. Nägler, R. Ménard, FEBS Lett. 434 (1998) 135–139.[22] J. Sivaraman, D.K. Nägler, R. Zhang, R. Ménard, M. Cygler, J.

Mol. Biol. 295 (2000) 939–951.

[23] D.K. Nägler, R. Zhang, W. Tam, T. Sulea, E.O. Purisima, R.Ménard, Biochemistry 38 (1998) 12648–12654.

[24] G. Gunbar, I. Klemenbib, B. Turk, V. Turk, A. Karaoglanovic-Car-mona, L. Juliano, D. Turk, Structure 8 (2000) 305–313.

[25] I. Klemenbib, A.K. Carmona, M.H.S. Cezari, M.A. Juliano, L. Juli-ano, G. Gunbar, D. Turk, I. Kriqaj, V. Turk, B. Turk, Eur. J. Bio-chem. 267 (2000) 5404–5412.

[26] C. Therrien, P. Lachance, T. Sulea, E.O. Purisima, H. Qi, E.Ziomek, A. Alvarez-Hernandez, W.R. Roush, R. Ménard, Bio-chemistry 40 (2001) 2702–2711.

[27] Y.F. Cheng, R.H. Kramer, J. Cell. Physiol. 139 (1989) 275–286.[28] C.A. Buck, A.F. Horwitz, Annu. Rev. Cell Dev. Biol. 3 (1987) 179–

205.[29] M. De Luca, C.M. Ward, K. Ohmori, R.K. Andrews, M. Berndt,

Biochem. Biophys. Res. Commun. 206 (1995) 570–576.[30] A.S. Kamiguti, C.R. Hay, M. Zuzel, Biochem. J. 320 (1996) 635–641.[31] S. Hashmi, J. Zhang, Y. Oksov, S. Lustigman, J. Biol. Chem. 279

(2004) 6035–6045.[32] Y.R. Hirata, M.H.S. Cezari, C.R. Nakaie, P. Boshcov, A.S. Ito,

M.A. Juliano, L. Juliano, Lett. Pept. Sci. 1 (1994) 299–308.[33] A.J. Barrett, A.A. Kembhavi, M.A. Brown, H. Kirschke, C.G.

Knight, M. Tamai, K. Hanada, Biochem. J. 201 (1982) 189–198.[34] N. Sreerama, R.W. Woody, Anal. Biochem. 209 (1993) 32–44.[35] M. Linke, V. Herzog, K. Brix, J. Cell Sci. 115 (2002) 4877–4889.[36] D. Musil, D. Zucic, D. Turk, R.A. Engh, I. Mayr, R. Huber, T.

Popovic, V. Turk, T. Towatari, N. Katunuma, W. Bode, EMBO J.10 (1991) 2321–2330.

[37] K. Nagler, A.C. Storer, F.C.V. Portaro, E. Carmona, L. Juliano, R.Ménard, Biochemistry 36 (1997) 12608–12615.

[38] B. Turk, I. Dolenc, E. Zerovnik, D. Turk, F. Gubensek, V. Turk,Biochemistry 33 (1994) 14800–14806.

[39] E. Dufour, V. Dive, F. Toma, Biochim. Biophys. Acta 995 (1988) 58–64.[40] B. Turk, I. Dolenc, V. Turk, J.G. Bieth, Biochemistry 32 (1993) 375–380.[41] A.D. Cardin, H.J.R. Weintraub, Atherosclerosis 9 (1989) 21–32.[42] J.R. Couchman, A. Woods, J. Cell Sci. 112 (1999) 3415–3420.[43] F. Echtermeyer, P.C. Baciu, S. Saoncella, Y. Ge, P.F. Goetinck, J.

Cell Sci. 112 (1999) 3433–3441.[44] S. Saoncella, F. Echtenmeyer, F. Dense, J.K. Nowlen, D.F.

Mosher, S.D. Robinson, R.O. Hynes, P.F. Goetinck, Proc. Natl.Acad. Sci. USA 96 (1999) 2805–2810.

[45] O. Goldshmidt, E. Zcharia, M. Cohen, H. Aingorn, I. Cohen, L. Nadav,B.Z. Katz, B. Geiger, Vlodavsky, FASEB J. 17 (2003) 1015–1025.

[46] D.M. Beauvais, A.C. Rapraeger, Exp. Cell Res. 286 (2003) 219–232.[47] T. Hirosaki, Y. Tsubota, Y. Kariya, K. Moriyama, H. Mizushima,

K. Miyazaki, J. Biol. Chem. 277 (2002) 49287–49295.[48] R. Bar-Shavit, Y. Eskohjido, J.W. Fenton II, J.D. Esko, I. Vlodav-

sky, J. Cell Biol. 123 (1993) 1279–1287.[49] R.M. Scarborough, M.A. Naughton, W. Teng, J.W. Rose, D.R.

Phillips, L. Nannizzi, A. Arfsten, A.M. Campbell, I.F. Charo, J.Biol. Chem. 268 (1993) 1066–1073.

[50] K. Von Figura, Curr. Opin. Cell Biol. 3 (1991) 642–646.