e

Experimental Cell Research 269, 13–22 (2001)doi:10.1006/excr.2001.5268, available online at http://www.idealibrary.com on

Evidence for the Involvement of Annexin 6 in the Trafficking betweenthe Endocytic Compartment and Lysosomes

Monica Pons,*,1 Thomas Grewal,†,1 Eulalia Rius,* Tino Schnitgerhans,† Stefan Jackle,† and Carlos Enrich*,2

*Departament de Biologia Celzlular, Institut d’ Investigacions Biomediques August Pi Sunyer, Facultat Medicina,Universitat de Barcelona, Barcelona 08036, Spain; and †Medizinische Kernklinik und Poliklinik,

Universitats-Krankenhaus Eppendorf, D-20246 Hamburg, Germany

p7fwSsetrsbcRdsteet[fe

Annexins are a family of calcium-dependent phos-pholipid-binding proteins, which have been impli-cated in a variety of biological processes includingmembrane trafficking. The annexin 6/lgp120 prelyso-somal compartment of NRK cells was loaded with low-density lipoprotein (LDL) and then its transport fromthis endocytic compartment and its degradation inlysosomes were studied. NRK cells were microinjectedwith the mutated annexin 6 (anx61–175), to assess thepossible involvement of annexin 6 in the transport ofLDL from the prelysosomal compartment. The resultsindicated that microinjection of mutated annexin 6, inNRK cells, showed the accumulation of LDL in largerendocytic structures, denoting retention of LDL in theprelysosomal compartment. To confirm the involve-ment of annexin 6 in the trafficking and the degrada-tion of LDL we used CHO cells transfected with mu-tated annexin 61–175. Thus, in agreement with NRK cellsthe results obtained in CHO cells demonstrated a sig-nificant inhibition of LDL degradation in CHO cellsexpressing the mutated form of annexin 6 compared tocontrols overexpressing wild-type annexin 6. There-fore, we conclude that annexin 6 is involved in thetrafficking events leading to LDL degradation. © 2001

Academic Press

Key Words: annexin 6; prelysosomes; LDL; spectrin;ndocytosis; endosomes; lysosomes; NRK; CHO cells.

INTRODUCTION

The membrane traffic pathway from early to lateendosomes has been studied in detail [1, 2]. In contrast,the events that govern the trafficking from the “lateendocytic compartment” to the lysosome are still poorlycharacterized. Although most trafficking into the lyso-somes comes through the endocytic compartment [3], a

1 These authors made equal contribution.2 To whom correspondence and reprint requests should be ad-

dressed at Departament de Biologia Celzlular, Facultat de Medicina,Universitat de Barcelona, Casanova 143, 08036-Barcelona, Spain.

Fax: 34 93 4021907. E-mail: enrich@medicina.ub.es.

13

subset of different intermediate entities, endocytic car-rier vesicles (ECVs), hybrid organelles, late endo-somes, or prelysosomes, has been described morpholog-ically and biochemically ([1, 4–7] and the referencestherein).

Among the molecules that might be involved in thesesteps of membrane trafficking are the low-molecular-weight GTPases, the Rabs [8], and the membrane fu-sion proteins of the vesicle-associated membrane pro-tein (VAMP), syntaxin and synaptosomal-associatedprotein of the 25-kDa (SNAP-25) families, and the sol-uble N-ethylmaleimide-sensitive factor attachment

rotein (SNAP) receptors (SNAREs) [9]. Whereas Raband 9 [10, 11] are associated with late endosomes,

ew SNAREs have been shown to specifically functionithin the endosomal pathway. Althought severalNAREs, including cellubrevin, VAMP8/endobrevin,yntaxin 13, and syntaxin 7, have been localized inndosomal membranes, their precise location, interac-ions, and function remain to be clarified (see a recenteview on polarized cells by Mostov et al. [12]); wehowed that cellubrevin/VAMP-3 was present in theasolateral “early” endocytic compartment of hepato-ytes and was involved in the transcytosis of pIgA [13].ecently, VAMP-7 (TI-VAMP) [14] was shown to me-iate the vesicular transport from endosomes to lyso-omes [9] and syntaxin 7 and VAMP 8 were reported inhe late endocytic compartment and shown to be nec-ssary for the fusion to lysosomes [15]. However, thencounter (kiss and run) and/or the final fusion be-ween two compartments (endosomes and lysosomes)16, 17] require not only proteins of the docking andusion machinery (Rabs and SNAREs) but a variety offfectors, adapters, GTP-binding proteins, Ca21/cal-

modulin, and tethering proteins [18]. We proposed thatannexin 6 acts as a tether between membranes at thelast stage of the endocytic route [7].

In this study we will highlight the importance of thesubcellular location of annexin 6, and its specific localinteractions with spectrin, as a baseline to understandthe function of annexin 6 in the “late endocytic com-

partment.” However, the ability of annexin 6 (like an-

0014-4827/01 $35.00Copyright © 2001 by Academic Press

All rights of reproduction in any form reserved.

14 PONS ET AL.

nexin II) to change its location according to signalswhich involve intracellular Ca21 mobilization shouldalso be considered [19, 20].

In recent years, the involvement of annexins inmembrane traffic has emerged as one of their predom-inant functions [21, 22]. Annexin 6 was first reportedat the plasma membrane of rat hepatocytes [23, 24],erythrocytes [25], or mammary tissue [26] but alsoassociated with the actin cytoskeleton in fibroblasticcell lines [27], in phagosomes of J774 macrophages [28]or in structures of the endocytic compartment [29–31].By biochemical means, we demonstrated that annexin6 was a major component of rat liver endosomes [32,

FIG. 1. Annexin 6 colocalizes with spectrin in the perinuclear regilocation in NRK cells shown by double-labeling immunocytochemperinuclear region of cells (c, merge); when cells were labeled with a mto spectrin at the cell surface (arrowheads) there was intense labcolocalized with annexin 6 (f, merge); finally, the labeling of annexinstaining in the same region of the cells no significant colocalization

FIG. 2. Interaction of GST–annexin 6 and spectrin. GST–annexiin subcellular fractions of NRK cells and in an endosomal fractionannexin 6 and lgp120 of the Percoll gradients [7], or a crude endosomafor 2 h and then bound proteins were collected and analyzed by Westno binding to spectrin.

33] and we have shown the prominent intracellular

location of annexin 6 in the perinuclear, prelysosomalcompartment of NRK and WIF-B cells [7]. At theplasma membrane (depending on Ca21 mobilization)annexin 6 may be involved in receptor-mediated endo-cytosis and perhaps in the remodeling of the spectrincytoskeleton at the cell surface during endocytosis,thus facilitating the release of clathrin-coated vesiclesfrom the plasma membrane. The binding of annexin 6to spectrin (and to actin), mediates the action of calpainI, which, in turn, may cleave the spectrin–actin cy-toskeleton allowing endocytosis [34]. The fact that cal-pain I was located in clathrin-coated vesicles [35] to-gether with the finding that annexin 6 binds to spectrin

of NRK cells. Immunofluorescence analysis of annexin 6 and spectriny. Anti-annexin 6 (a, d, g) and anti-lgp120 (b) colocalize in theoclonal anti-spectrin antibody (e) it can be observed that in addition

ng at different intracellular locations (arrows) and some of theseg) and the Golgi marker (anti-Golgi 58k) (h) was compared. Despites observed (i, merge). Bar, 10 mm.was used to study the interactions of annexin 6 with other proteinsm rat liver. A prelysosomal fraction, corresponding to the peak ofaction from rat liver was incubated with GST–annexin 6–Sepharoseblotting with anti-spectrin antibody. Control with GST alone shows

onistron

eli6 (wan 6frol fr

ern

[36] and is required for budding in vitro [37] indicates

t6m

nin

15ANNEXIN 6 MEDIATES THE TRAFFICKING TO LYSOSOMES

that a reorganization of the spectrin cytoskeletonmight be necessary to allow the budding at the plasmamembrane.

Interestingly, spectrin is also present in the mem-branes of Golgi [38], in a variety of poorly characterizedintracellular vesicles [39, 40] and in macropinosomes ofNIH 3T3 cells [41]. Supporting the role of annexin 6 atthe plasma membrane, we showed the involvement ofannexin 6 in the endocytosis and trafficking of low-density lipoprotein (LDL) [42]. Here we demonstratethat the mechanism proposed by Kamal and co-work-ers (1998) [34] for annexin 6 at the plasma membrane

FIG. 3. Uptake and transport of LDL in microinjected NRK cellwith the pcDNA of annexin 6 wild type (a) or the pcDNA of the mutahen incubated with DiI-LDL for 1 h, chased overnight, and visualize(green nucleus) show no differences in the pattern of DiI-LDL inticroinjected cells. White arrow (a, b) shows a noninjected cell for cFIG. 4. Colocalization of annexin 6 and internalized DiI-LDL in

with anti-annexin 6 (a, d) after 2 h DiI-LDL internalization in contr(d, e, f) Arrows indicate enlargement of annexin 6 structures contai

can also operate in structures of the late endocytic

compartment, which is the priority location of annexin6 in NRK cells [7]. Our results indicate that, in NRKand CHO cells, annexin 6 is involved in the traffickingevents occurring at the endosomal compartment, andtherefore may modulate the exit of LDL from the pre-lysosomal compartment and eventually its degradationin lysosomes.

MATERIALS AND METHODS

Cell culture and antibodies. NRK and CHO cells were used inthis work. NRK cells were grown in DMEM and CHO cells in Ham’s

epresentative fields of NRK cells co-microinjected into the nucleusform of annexin 6 (anx61–175) (b) and dextran–FITC. Cells were andith confocal microscopy. (a) Cells overexpressing wild-type annexin

alization. (b) DiI-LDL accumulates in large endocytic structures ofparison. Bar, 10 mm.LN-treated cells. Confocal microscopy analysis of NRK cells labeledells (DMSO) (b) or cells treated with ALLN (e). (c, f) Colocalization.g LDL. Bar, 10 mm.

s. Rtedd wernomALol c

F-12 supplemented with 10% fetal calf serum (FCS), L-glutamine (2

5(oSfoHtw

6il(wfss

wbfiabpamTI

w

a

8s

p

d

das

ccwpctt(Tflwp

[rss

oGpiwC

pDpt5dtf

tt

16 PONS ET AL.

mM), penicillin (100 U/ml), and streptomycin (100 mg/ml) at 37°C,% CO2. Fluorescence-conjugated antibodies were from JacksonWest Grove, PA) or Boehringer Mannheim (Barcelona, Spain); per-xidase-labeled antibodies were from Bio-Rad Laboratories (Madrid,pain). N-Acetyl-leucyl-leucyl-norleucinal (ALLN) was purchased

rom Calbiochem–Novabiochem Corp. (Darmstadt, Germany). Allther chemicals were from Sigma Chemical Co. (Madrid, Spain).uman lipoprotein-deficient serum (H-LPDS) was prepared by ul-

racentrifugation of human plasma [43]. Fluorescent LDL (DiI- LDL)as described by Pitas et al. (1985) [44]. LDL was radiolabeled with

125I by the iodine monochloride method [45].The rabbit polyclonal antibody (pAb) to membrane-bound annexinand an anti-sheep polyclonal antibody to annexin 6 were prepared

n our laboratory and affinity-purified [29, 42]. The mouse mAb togp120 (GM10) was provided by Dr. K. Siddle and Dr. J. HuttonUniv. of Cambridge, UK) [46]. Anti-Golgi 58k monoclonal antibodyas from Sigma. The mouse monoclonal anti-spectrin antibody was

rom Sigma; in some experiments we also used a polyclonal anti-pectrin generously provided by Dr. James Nelson (Stanford Univer-ity).Indirect immunofluorescence on fixed cells. NRK or CHO cellsere fixed for 10 min in 3.7% paraformaldehyde in 0.1 M phosphateuffer prior to permeabilization with 0.1% saponin in 0.5% BSA/PBSor 15 min. Then cells were incubated for 1 h with primary antibod-es, rinsed with PBS, and incubated for 45 min with secondaryntibodies conjugated to FITC or TRITC. Dilutions of primary anti-odies in 0.1% saponin/0.5% BSA/PBS were as follows: affinity-urified pAb to annexin 6 (1:10); monoclonal anti-Golgi58k (1:200);scites to lgp120 (1:300). After final washes in PBS, cells wereounted in Mowiol. Confocal images were collected using a LeicaCS NT equipped with a 63X Leitz Plan-Apo objective (NA 1.4).mages represent approx. 1.0-mm optical sections.

Preparation of annexin 6 constructs. The expression vector of therat annexin 6 deletion mutant pcDNA1–175 was constructed using a542-bp HdIII-Asp1 annexin 6 cDNA fragment containing codingsequences from position 11 to position 1527, which lacks six of theeight repeats at the carboxy terminal. The DNA fragment was iso-lated, gel-purified, and cloned into the eukaryotic expression vectorpcDNA3.11 (Invitrogen). For the construction of the GST gene fu-sion vector GST-anx6 full-length 2.0-kb rat annexin 6 cDNA wascloned into pGEX-KG (Pharmacia). DNA sequence analysis using aDye Terminator Cycle Sequencing reaction kit (Perkin Elmer) to-gether with T7, SP6 (Promega), and internal annexin 6 primersconfirmed the full-length and wild-type rat annexin 6 cDNA se-quence in pcDNAanx61–175 and GST-anx6. All cloning procedures

ere performed according to standard protocols [47].Uptake of 125I-LDL and degradation assay. To increase the deg-

radation of 125I-LDL, 1 3 106 NRK cells were incubated for 24 h withLPDS and preincubated with serum-free medium (DMEM, 1% BSA)for 30 min at 37°C. Cells were chilled on ice and 125I-LDL was addedt 106 cpm/ml (in triplicate). After 2 h of LDL internalization cells

were treated with ALLN (500 mM) (Calpain I inhibitor) for 2, 4, 6, orh; 100-fold excess of unlabeled LDL was added to determine non-

pecific uptake of 125I-LDL. At each time point 100 ml of media wasremoved and 3 M ice-cold trichloroacetic acid (TCA) was added to afinal concentration of 10%. The sample was vortexed vigorously andincubated for 10 min on ice. After addition of 250 ml of 0.7 M AgNO3

to remove free iodine [48] the sample was vortexed again and thencentrifuged and the amount of radioactivity in the remaining super-natant was determined.

In some experiments CHO cells were used; 48 h prior to incubationwith radioactive LDL, 1 3 105 cells per well were plated on 12-well

lates. Cells were transfected after 24 h and incubated with 125I-LDL(4–5 mg/ml; in triplicate) the following day. Measurement of degra-

ation was carried out after 6 h incubation at 37°C of cells with125I-LDL as described [42]. After 6 h the media was removed and

100-ml aliquots were analyzed as described above. To determine E

internalization of 125I-LDL cells were incubed from 5 to 60 min asescribed in Grewal et al. [42]; 50-fold excess of unlabeled LDL wasdded to every third sample prior the incubation to determine non-pecific uptake of 125I-LDL. At each time point cells were washed

three times with ice-cold PBS and subsequently with PBS/heparin(20 units/ml) to remove surface-bound LDL and to determine theamount of internalized LDL.

Isolation of subcellular fractions from NRK, CHO cells, and endo-somal fractions from rat liver. The separation of subcellular frac-tions of NRK cells was accomplished using self-generating linearPercoll gradients [49]. Briefly, cells grown in 6 3 150 cm2 tissueulture flasks were washed in ice-cold PBS, scraped, and pelleted byentrifugation in a table centrifuge (750 rpm, 5 min). Cells were thenashed in homogenization buffer (HB: 10 mM Tris/250 mM sucrose,H 7.4), pelleted (2500 rpm, 15 min), resuspended in 2 ml HBontaining protease inhibitors, and homogenized by 10–15 passageshrough a 22-gauge needle. After centrifugation (2500 rpm, 15 min)he postnuclear supernatant was collected, layered over 7 ml of 27%v/v) Percoll in HB, and centrifuged (21,000 rpm, 90 min, 4°C) in a 50i rotor (Beckman Instruments). Aliquots of 500 ml were collected

rom the fractions that corresponded to the peak of annexin 6 andgp120, as described previously [7]. In some experiments CHO cellsere also used and subcellular fractions were prepared as describedreviously [42].Finally, the isolation of rat liver endocytic fractions was performed

50, 51, 29, 13]. Morphological and biochemical characterization ofat liver endocytic fractions was as described [52]. The crude endo-omal fraction was used for the pull-down experiments shown in thistudy.Pull-downs with purified proteins. Recombinant annexin 6 was

btained as glutathione-S-transferase (GST) fusion protein [53].ST-anx6 was expressed in BL21 pLysE Escherichia coli strain andurified by gluthatione–Sepharose chromatography. GST-anx6 wasncubated 90 min with glutathione–Sepharose at 4°C in PBS. Afterashing with 50 mM Tris, 150 mM NaCl, 1% Triton X-100, 0.1 mMaCl2, Sepharose was then incubated, with the two subcellular frac-

tions described above, for 2 h at 4°C and collected by centrifugation.Proteins bound to the column were collected and analyzed by elec-trophoresis and Western blot.

Microinjection of cDNAs to NRK cells. NRK fibroblasts wereseeded onto glass coverslips 24–48 h before injection. Expressionvectors encoding the mutated annexin 6 (anx61–175) or wild-type an-nexin 6 (anx6wt) were mixed with FITC–dextran (150,000 mw,Sigma) to act as an injection marker (final concentration 100 ng/ml

lasmid, 1% dextran in distilled water). Cells were transferred intoMEM (Hepes modification, Sigma) containing 5% FCS medium justrior to injection. Nuclear injections were carried out with the Au-omated Injection System (AIS) (Zeiss), equipped with an Eppendorf246 transjector. Injected cells were placed into fresh DMEM me-ium containing 5% FCS and incubated at 37°C. Cells were allowedo recover for 1 h at 37°C and then incubated with the ligand (pulseor 60 min of DiI-LDL, 30 mg/ml) and chased for 13 h. After the chase,

cells were fixed with 4% PFA, mounted with Mowiol, and observedwith the confocal microscope. Overexpression of wild-type and mu-tant annexin 6 was visualized in cells fixed after microinjection byimmunofluorescence using an anti-sheep polyclonal anti-annexin 6antibody [42].

Transfections. The DNA used for transfections of CHO cells waspurified with the plasmid purification kit from Qiagen; 2–3 3 105

cells were transfected with 1 mg DNA and the FUGENE 6 transfec-ion reagent (Boehringer Mannheim) according to the instructions ofhe manufacturer. Control cells were transfected with pCMVSPORT-

bGal (GIBCO-BRL). For cotransfections of wild-type or mutant an-nexin 6 together with LDL-R (pCMVhLDLR kindly provided by Dr.F. Schniders, Univ. Hamburg) 0.5 mg of each plasmid was used.

xpression of b-galactosidase in 10–25% of cells was confirmed by

l

sce(

aco

17ANNEXIN 6 MEDIATES THE TRAFFICKING TO LYSOSOMES

X-Gal staining and coexpression of plasmids in transfected cells wasconfirmed by double immunofluorescence (data not shown).

RESULTS

Annexin 6 and Spectrin in the PrelysosomalCompartment of NRK Cells

In NRK cells the structures with the highest concen-trations of annexin 6 were mostly positive for lgp120(Figs. 1a–c), and devoid of M6P-R, a late endosomalmarker [7]. Furthermore, it was observed that annexin6 also colocalizes, to some extent, with spectrin in theperinuclear region of NRK cells (Figs. 1d–f). However,the spectrin observed in the perinuclear location,which colocalizes with annexin 6, does not correspondto the Golgi spectrin, since annexin 6 did not overlapwith the Golgi structures, labeled with anti-Golgi 58kantibody (Figs. 1g–i). This indicated that the annexin6–prelysosomal compartment also contains spectrin.

The interaction of annexin 6 with spectrin was con-firmed, in vitro, by pull-down experiments with GST–annexin 6 using a prelysosome/late endosomal fractionfrom NRK cells, and also in rat liver endosomes. Figure2 shows the Western blotting, with anti-spectrin poly-clonal antibody, after the pull-down with GST–annexin6 (GST-anx6wt) or GST, as control; spectrin was posi-tively recognized in both subcellular fractions.

Annexin 6 Is Involved in the Trafficking to Lysosomes

In this study two different approaches that interferewith the removal of spectrin from membranes wereconsidered: overexpression of anx61–175 by microinjec-tion of the pcDNA-anx61–175 into the nucleus of NRKcells (NRK cells showed very low efficiency of transfec-tion) and incubation of cells with ALLN, a calpaininhibitor. Overexpression of anx61–175 was shown toinhibit the LDL internalization, by clathrin-coatedpits, and inhibit the loss of spectrin in isolated mem-branes in vitro [34].

Cells were first microinjected with the pcDNAanx6wt

vector or cDNA corresponding to the mutated anx61–175,allowed to recover for 1 h, and incubated with DiI-LDLfor 1 h and chased overnight. This treatment providesthe accumulation of DiI-LDL into the annexin 6 prely-sosomal/late endocytic compartment, before the ex-pression of microinjected cDNAs (see also Figs. 4a and4b). From this prelysosomal compartment we then in-vestigated whether LDL was able to reach the lyso-somes for degradation. The distribution of DiI-LDL incells microinjected with anx6wt showed no differencefrom noninjected cells (vesicular structures of 200–400nm average size in the Golgi/lysosome region) (Fig. 3a);however, those cells microinjected with the mutatedanx61–175 showed vesicles with DiI-LDL, which were

arger (approx. 500–700 nm) (Fig. 3b) (70% of anx61–175

microinjected cells showed the enlarged LDL contain-ing endosomes). Immunocytochemical analysis of an-nexin 6 expressing the truncated form (anx61–175)howed a diffuse cytoplasmic staining, similar to thoseells overexpressing the wild type (not shown). Thefficiency of microinjection was 85% for the wild typeanx6 wt) and 71% for the mutated form (anx61–175

mutant).In the second approach, the annexin 6 compartment

was loaded with DiI-LDL for 2 h, in NRK cells inducedto express LDL receptors by replacing the fetal calfserum with human lipoprotein deficient serum as pre-viously described [43]. Then, cells were chased for 2 hin the presence of DMSO or ALLN. Figures 4a and 4bshow the colocalization of internalized LDL and an-nexin 6 and also the effect of ALLN on the endocyticcompartment; an enlargement of endocytic structurescan be also observed at the perinuclear region aftertreatment with the calpain inhibitor (a similar effectwas observed in cells treated with chloroquine) [7]. Theenlargement of LDL-containing vesicles, after ALLNtreatment, may be caused by an accumulation of LDLin prelysosomes, with a concomitant inhibition of deg-radation in the lysosomes.

To find out whether this enlargement of endocyticstructures was attributable to an inhibition of LDLdegradation, we repeated the experiment, detailed inFig. 4, but now using 125I-LDL. Medium was collectedafter 2, 4, 6, and 8 h of DMSO/ALLN treatment and themount of free 125I in the supernatant after TCA pre-ipitation was measured. Figure 5 shows the inhibitionf 125I-LDL degradation, at all time points, in cells

treated with ALLN compared with control samples(DMSO). Since it has been described that ALLN mayalso inhibit cathepsins B and L, we tested the effect ofcathepsin inhibitor I using the same conditions. Nomorphological changes in the pattern of labeling of theendocytic compartment were observed (not shown).

Annexin 6 Is Also Involved in the Trafficking andDegradation of LDL in CHO Cells

Since NRK cells showed low efficiency in transfectionand the analysis of the expression and location of mu-tated annexin 6 (annexin 61–175), in microinjected NRKcells was difficult, we used CHO cells shown to havevery low levels of endogenous annexin 6 [42]. Tran-siently transfected CHO cells with wild-type annexin 6or the mutated annexin 61–175 were characterized. Cellswere transiently cotransfected with LDL-R and an-nexin 6 wild type or the mutated anx61–175 and pre-pared for immunofluorescence, after the incubation ofDiI-LDL for 1 h. Figure 6 shows the pattern of intra-cellular wild-type annexin 6 (a) distributed throughoutthe cell but concentrated in structures underneath theplasma membrane (arrows); on the other hand, the

mutated annexin 61–175 (c) showed a more diffuse intra-

A

Aspi

p

18 PONS ET AL.

cellular distribution, being more abundant in the pe-rinuclear region (arrows). In both, internalized DiI-LDL was concentrated in endocytic structures at theperinuclear region (b, d).

In order to assess whether the mutated annexin61–175 was bound to membranes, cells were lysed andthen lysates centrifuged. Pellets and supernatants

FIG. 5. 125I-LDL degradation assay in NRK cells treated withLLN. NRK cells were internalized with 125I-LDL for 2 h and then

chased for 2, 4, 6, or 8 h in the presence of DMSO (gray bars) orLLN (black bars) and the degradation of LDL was compared. Aignificant inhibition of degradation can be observed in those sam-les treated with ALLN. Results are from three independent exper-ments (in triplicate) and values did not vary more than 10%.

FIG. 6. Immunofluorescence analysis of annexin 6 in CHO cecotransfected with LDL-R plus annexin 6 wild type (a) or LDL-R ancotransfected cells were incubated with DiI-LDL for 1 h (in red); wi

referential location of wild-type or mutated annexin 6 in transfected C

were analyzed by Western blotting with anti-annexin 6antibody. Figure 7 shows that whereas the wild-typeannexin 6 was almost 100% membrane-bound, signifi-cant amounts of annexin 61–175 (approx. 50%) was de-tected in the cytosolic fraction.

In some experiments CHO cells were homogenizedand the postnuclear supernatant fraction (PNS) wassubjected to cellular fractionation on sucrose gradients,as characterized in detail in Grewal et al. [42]. Frac-tions from the gradient were separated by SDS–PAGEand then analyzed for the presence of spectrin andannexin 6. Spectrin was detected in the fractions (3–5of the sucrose gradient) corresponding to endosomalmarkers and annexin 6 (not shown).

Finally, we used CHO cells transiently cotransfectedwith annexin 6, wild-type or mutated form, and theLDL-R to analyze the uptake and the degradation of125I-LDL. In Fig. 8 it is shown that there was a signif-icant decreased degradation of 125I-LDL (32%) in thecells expressing the mutated annexin 6 (LDL-R plusanx61–175), compared with LDL-R plus anx61–175. Data ofuptake also showed that, in agreement with Kamal etal. [34], there was a slight decrease on the endocytosisof 125I-LDL in cells transfected with the mutatedanx61–175 compared to anx6wt.

. Confocal immunofluorescence analysis of CHO cells transientlyutated annexin 61–175 (c) (in green), using anti-annexin 6 antibody;

type CHO cells (b) and CHO-anx61–175 cells (d). Arrows indicate the

llsd mld-

HO cells. Bar, 5 mm.

st

dw

aL

19ANNEXIN 6 MEDIATES THE TRAFFICKING TO LYSOSOMES

DISCUSSION

In a recent article by Anderson and co-workers afascinating new mechanism for the regulation of recep-tor-mediated endocytosis was envisaged [34]. It isbased on the interaction of spectrin with annexin 6,and the removal of spectrin, from the plasma mem-brane, mediated by the calcium-dependent cysteineprotease calpain. Upon annexin 6 binding to spectrin,calpain I cleaves spectrin and “opens” the actin cy-toskeleton facilitating the endocytosis. It is assumedthat annexin 6 was located at the clathrin-coated pitson the plasma membrane.

By means of an antibody prepared by purification ofannexin 6 from isolated rat liver endosomes [32] wedemonstrated a precise intracellular staining of endo-cytic structures, by confocal and electron microscopy,in hepatocytes, WIF-B, and NRK cells [29, 13, 7], inagreement with its proposed role in intracellular traf-ficking [21, 31, 39, 42]. In NRK cells annexin 6 wasalmost exclusively found associated with the prelyso-somal compartment [7], This intracellular locationdoes not rule out its presence at other sites such as theplasma membrane, and most probably reflects the var-ious antibodies used, which may recognize differentisoforms or conformations of annexin 6, or differentcells analyzed. Besides, annexins I,II, IV, V, and VIrelocate in response to rises in intracellular calcium;particularly, annexin 6 was shown to move from theperinuclear region to a more homogeneous distributionon the plasma membrane, in human fibroblasts [30,54], or to the sarcolemma to serve as a link of thecytoskeleton, in smooth muscle cells [20]. In addition,we now have evidence that annexin 6 may undergo

FIG. 7. Western blot analysis of annexin 6 distribution in tran-iently transfected CHO cells. CHO cells were transfected with wild-ype annexin 6 or the deletion mutant annexin 61–175 as indicated;

24 h after transfection cells were lysed and centrifuged for 15 min at13,000 rpm. Pellets (membrane) and supernatants (cytosol) werecleared for 1 h (100,000 rpm); 20 mg of protein was loaded and theexpression of annexin 6 was analyzed by Western blotting using anaffinity-purified polyclonal anti-annexin 6 antibody.

ligand-induced translocation [42]. Thus, considering

the priority intracellular location of annexin 6, we in-vestigated whether a similar mechanism postulated forannexin 6 at the plasma membrane also occurs be-tween the endocytic compartment and lysosomes.

In this study we show the presence of spectrin in theannexin 6 intracellular compartment and demonstratethat both proteins interact. The annexin 6 compart-ment of NRK cells was loaded with LDL and then thetransport from this prelysosomal compartment and thedegradation of LDL were analyzed. NRK cells were

FIG. 8. Uptake and degradation of 125I-LDL in transient trans-fected CHO cells. CHO cells (3 3 105) transfected with b-galactosi-

ase (b-gal) or LDL-R, or cotransfected with LDL-R 1 anx61–175 orith LDL-R 1 anx6wt were chilled on ice and 125I-LDL was added at

4°C for 30 min. Unbound LDL was removed and cells were incubatedfor 5–60 min at 37°C. After cell lysis with 0.1 M NaOH the specificradioactivity of solubilized cells was determined (a). Endocytosis of125I-LDL was calculated as -fold increase compared to b-galactosi-dase-transfected cells (1.0 at 5 min) and represents the mean of oneof three representative and independent experiments with triplicatesamples. For degradation (b) cells were preincubated as describedabove and 125I-LDL was added for 6 h at 37°C. A reduction of 32% ofLDL degradation in cells transfected with anx61–175 compared with tothe anx6wt was observed. This is statistically significant for fiveindependent experiments with triplicate samples (6 standard devi-tion). A Student t test was performed: P 5 0.14 for LDL-R andDL-R1 anx61–175; P . 0.02 for LDL-R and LDL-R1 anx6 (*); P .

0.0004 for LDL-R and b-gal (***).

20 PONS ET AL.

microinjected with anx6wt or mutated anx61–175

pcDNAs, to assess the possible involvement of annexin6–spectrin interactions in the trafficking of LDL to thelysosomal compartment. The microinjection of mutatedannexin 6 showed an accumulation of LDL in largerendocytic structures which resemble those observedafter ALLN (calpain I inhibitor) treatment. Alterna-tively in CHO cells, which showed very low levels ofendogenous annexin 6, the LDL degradation in tran-sient transfected annexin 61–175 cells was significantlydecreased, compared with that in cells expressing thewild-type annexin 6.

It has been described that calpains are involved inthe formation of coated vesicles and/or vesicle fusion toendosomes [35]. These calcium-dependent proteasesassociate with coated vesicles and modulate the inter-actions of membranes and cytoskeleton (calpains de-grade membrane-lining proteins that connect mem-brane proteins and cytoskeletal proteins) with aconcomitant effect on the disorganization of the lipidasymmetry of membranes, which is a necessary stepfor membrane fusion. The location of both calpain andannexin 6 to membranes is Ca21-dependent, indicatingthat calpain might be involved in the budding throughthe modulation of annexin 6; also, it was shown thatvesicle-bound annexin 6/calpain complexes were in-volved in the fusion of intracellular membranes [35]. Atleast in NRK cells, at the steady state annexin 6 seemsassociated with spectrin. In the prelysosomal compart-ment, however, changes of intracellular Ca21, whichmay be induced by ligand transport [42] or from signaltransduction upstream events, may recruit calpain I tocleave spectrin, allowing the fusion with lysosomes.

FIG. 9. Proposed model for annexin 6–spectrin–calpain interactiof LDL: (1) LDL binding to cell surface receptors (clathrin-coated pitsto calmodulin (CaM); (3.2) Ca21-induced translocation of residualendosomes; or (3.3) Ca21-induced increased binding affinity of annexenhances the sensitivity of spectrin to calpain [59–61]; (6) Removal( ) Membranes with spectrin/annexin 6 cytoskeleton.

From the results of this study we postulate the model

shown in Fig. 9 based in two experimental evidences:(1) the rise of intracellular Ca21 triggered by LDL [55–57] and (2) the reorganization of annexin 6 (relocate)according to Ca21 mobilization in the cell [20, 54]. Weconclude that, although annexin 6 might be dispens-able (see work in knockout mice by Hawkins et al. [58])for adult structure or development and/or even forthose cellular aspects in which annexin 6 has beeninvolved, in cells expressing annexin 6 it seems that itcan be involved in intracellular trafficking from theendocytic compartment to the lysosomes.

This work was supported by Grants PM96-0083 and PM99-0166from Ministerio de Educacion y Cultura (Spain), Acciones IntegradasHA97-0007, and Generalitat de Catalunya BE2000. We thank toAnna Bosch, Serveis Cientıfic i Tecnics de la Universitat de Barce-lona, for help with confocal microscopy.

REFERENCES

1. Gruenberg, J., and Maxfield, F. R. (1995). Membrane-transportin the endocytic pathway. Curr. Opin. Cell Biol. 7, 552–563.

2. Gu, F., and Gruenberg, J. (1999). Biogenesis of transport inter-mediates in the endocytic pathway. FEBS Lett. 452, 61–66.

3. Mellman, I. (1996). Endocytosis and molecular sorting. Annu.Rev. Cell Dev. Biol. 12, 575–625.

4. Kornfeld, S., and Mellman, I. (1989). The biogenesis of lyso-somes. Annu. Rev. Cell Biol. 5, 483–525.

5. Geuze, H. J., Stoorvogel, W., Strous, G. J., Slot, J. W., Bleeke-molen, J. E., and Mellman, I. (1988). Sorting of mannose 6-phos-phate receptors and lysosomal membrane proteins in endocyticvesicles. J. Cell Biol. 107, 2491–2501.

6. Jahraus, A., Storrie, B., Griffiths, G., and Desjardins, M. (1994).Evidence for retrograde traffic between terminal lysosomes andthe prelysosomal/late compartment. J. Cell Sci. 107, 145–157.

7. Pons, M., Ihrke, G., Koch, S., Biermer, M., Pol, A., Grewal, T.,

in the late endocytic compartment being involved in the degradation2) Ca21 mobilization (increase of intracellular Ca21); (3.1) Ca21 bindstosolic annexin 6 and/or formation of “active” annexin 6 in late6 to spectrin; (4) CaM/Ca21 binds to spectrin; (5) CaM/Ca21 bindingspectrin/annexin 6 and fusion or allow exit (budding) to lysosomes.

ons); (cy

inof

Jackle, S., and Enrich, C. (2000). Late endocytic compartments

1

1

1

1

1

1

1

1

2

23

4

4

21ANNEXIN 6 MEDIATES THE TRAFFICKING TO LYSOSOMES

are major sites of annexin VI localization in NRK fibroblastsand polarized WIF-B hepatoma cells. Exp. Cell Res. 257, 33–47.

8. Novick, P., and Zerial, M. (1997). The diversity of Rab proteinsin vesicle transport. Curr. Opin. Cell Biol. 9, 496–504.

9. Advani, R. J., Yang, B., Prekeris, R., Lee, K. C., Klumperman,J., and Scheller, R. H. (1999). VAMP-7 Mediates vesiculartransport from endosomes to lysosomes. J. Cell Biol. 146, 765–775.

0. Chavrier, P., Parton, R. G., Hauri, H. P., Simons, K., and Zerial,M. (1990). Localization of the low molecular weight GTP bind-ing proteins to exocytic and endocytic compartments. Cell 62,317–329.

1. Lombardi, D., Soldati, T., Riederer, M. A., Goda, Y., Zerial, M.,and Pfeffer, S. R. (1993). Rab9 functions in transport betweenlate endosomes and the trans Golgi network. EMBO J. 12,677–682.

2. Mostov, K. E., Verges, M., and Altschuler, Y. (2000). Membranetraffic in polarized epithelial cells. Curr. Opin. Cell Biol. 12,483–490.

3. Calvo, M., Pol, A., Lu, A., Ortega, D., Pons, M., Blasi, J., andEnrich, C. (2000). Cellubrevin is present in the basolateralendocytic compartment of hepatocytes and follows the transcy-totic pathway after IgA internalization. J. Biol. Chem. 275,7910–7917.

4. Galli, T., Zahraoui, A., Vaidyanathan, V. V., Raposo, G., Tian,J. M., Karin, M., Niemann, H., and Louvard, D. (1998). A noveltetanus neurotoxin-insensitive vesicle-associated membraneprotein in SNARE complexes of the apical plasma membrane ofepithelial cells. Mol. Biol. Cell 9, 1437–1448.

5. Mullock, B. M., Smith, C. W., Ihrke, G., Bright, N. A., Lindsay,M., Parkinson, E. J., Brooks, D. A., Parton, R. G., James, D. E.,Luzio, J. P., and Piper, R. C. (2000). Syntaxin 7 is localized tolate endosome compartments, associates with Vamp 8, and isrequired for late endosome–lysosome fusion. Mol. Biol. Cell 11,3137–3153.

6. Storrie, B., and Desjardins, M. (1996). The biogenesis of lyso-somes: Is it a kiss and run, continous fusion and fission process?BioEssays 18, 895–903.

17. Luzio, J. P., Rous, B. A., Bright, N. A., Pryor, P. R., Mullock,B. M., and Piper, R. C. (2000). Lysosome–endosome fusion andlysosome biogenesis. J. Cell Sci. 113, 1515–1524.

18. Pryor, P. R., Mullock, B. M., Bright, N. A., Gray, S. R., andLuzio, J. P. (2000). The role of intraorganellar Ca21 in lateendosome–lysosome heperotypic fusion and in the reformationof lysosomes from hybrid organelles. J. Cell Biol. 149, 1053–1062.

9. Babiychuk, E. B., Palstra, R.-J., T. S., Schaller, J., Kampfer, U.,and Draeger, A. (1999). Annexin VI participates in the forma-tion of a reversible, membrane–cytoskeleton complex in smoothmuscle cells. J. Biol. Chem. 274, 35191–35195.

0. Babiychuk, E. B., and Draeger, A. (2000). Annexins in cellmembrane dynamics: Ca21-regulated association of lipid do-mains. J. Cell Biol. 150, 1113–1123.

1. Gerke, V., and Moss, S. E. (1997). Annexins and membranedynamics. Biochim. Biophys. Acta 1357, 129–154.

22. Gruenberg, J., and Emans, N. (1993). Annexins in membranetraffic. Trends Cell Biol. 3, 224–227.

23. Tagoe, C. E., Boustead, C. M., Higgins, S. J., and Walker, J. H.(1994). Characterization and immunolocalization of rat liverannexin VI. Biochim. Biophys. Acta 1192, 272–280.

24. Weinman, J. S., Feinberg, J. M., Rainteau, D. P., DellaGaspera, B., and Weinman, S. J. (1994). Annexins in rat en-terocyte and hepatocyte: An immunogold electron-microscope

study. Cell Tissue Res. 278, 389–397.

25. Bandorowicz, J., Pikula, S., and Sobota, A. (1992). Annexins IV(p32) and VI (p68) interact with erythrocyte membrane in acalcium-dependent manner. Biochim. Biophys. Acta 1105, 201–206.

26. Lavialle, F., Rainteau, D., Massey-Harroche, D., and Metz, F.(2000). Establishment of plasma membrane polarity in mam-mary epithelial cells correlates with changes in prolactin traf-ficking and in annexin VI recruitment to membranes. Biochim.Biophys Acta 1464, 83–94.

27. Hosoya, H., Kobayashi, R., Tsukita, S., and Matsumura, F.(1992). Ca21-regulated actin and phospholipid binding protein(68kD-protein) from bovine liver: Identification as a homologuefor annexin VI and intracellular localization. Cell Motil. Cy-toskel. 22, 200–210.

28. Desjardins, M., Celis, J. E., van Meer, G., Dieplinger, H., Jahr-aus, A., Griffiths, G., and Huber, L. A. (1994). Molecular char-acterization of phagosomes. J. Biol. Chem. 269, 32194–32200.

29. Ortega, D., Pol., A., Biermer, M., Jackle, S., and Enrich, C.(1998). Annexin VI defines an apical endocytic compartment inrat liver hepatocytes. J. Cell Sci. 111, 261–269.

30. Seemann, J., Weber, K., Osborn, M., Parton, R. G., and Gerke,V. (1996). The association of annexin I with early endosomes isregulated by Ca21 and requires an intact N-terminal domain.Mol. Biol. Cell 7, 1359–1374.

31. Massey-Harroche, D., Mayran, N., and Maroux, S. (1998). Po-larized localizations of annexins I, II, VI and XIII in epithelialcells of intestinal, hepatic and pancreatic tissues. J. Cell Sci.111, 3007–3015.

32. Jackle, S., Beisegel, U., Rinninger, F., Buck, F., Grigoleit, A.,Block, A., Groger, L., Greten, H., and Windler, E. (1994). An-nexin VI, a marker protein of hepatocytic endosomes. J. Biol.Chem. 269, 1026–1032.

33. Pol, A., Ortega, D., and Enrich, C. (1997). Identification ofcytoskeleton-associated proteins in isolated rat liver endo-somes. Biochem. J. 327, 741–746.

34. Kamal, A., Ying, Y., and Anderson, R. G. W. (1998). AnnexinVI-mediated loss of spectrin during coated pit budding is cou-pled to delivery of LDL to lysosomes. J. Cell Biol. 142, 937–947.

35. Sato, K., Saito, Y., and Kawashima, S. (1995). Identificationand characterization of membrane-bound calpains in clathrin-coated vesicles from bovine brain. Eur. J. Biochem. 230, 25–31.

36. Watanabe, T., Inui, M., Chen, B. Y., Iga, M., and Sobue, K.(1994). Annexin VI-binding proteins in brain. Interaction ofannexin VI with a membrane skeletal protein, calspectin (brainspectrin or fodrin). J. Biol. Chem. 269, 17656–17662.

37. Lin, H. C., Sudorf, T. C., and Anderson, R. G. W. (1992). An-nexin VI is required for budding of clathrin-coated pits. Cell 70,283–291.

38. Beck, K. A., Buchanan, J. A., Malhotra, V., and Nelson, W. J.(1994). Golgi spectrin: Identification of an erythroid b-spectrinhomolog associated with the Golgi complex. J. Cell Biol. 127,707–723.

9. De Matteis, M. A., and Morrow, J. S. (1998). The role of ankyrinand spectrin in membrane transport and domain formation.Curr. Opin. Cell Biol. 10, 542–549.

0. Stankewich, M. C., Tse, W. T., Peters, L. L., Ch’ng, Y., John,K. M., Stabach, P. R., Devarajan, P., Morrow, J. S., and Lux,S. E. (1998). A wide expressed bIII spectrin associated withGolgi and cytoplasmic vesicles. Proc. Natl. Acad. Sci. USA 95,14158–14163.

1. Xu, J., Ziemnicka, D., Merz, G. S., and Kotula, L. (2000). Hu-man spectrin Src homology 3 domain binding protein 1 regu-lates macropinocytosis in NIH 3T3 cells. J. Cell Sci. 113, 3805–

3814.

4

RRP

22 PONS ET AL.

42. Grewal, T., Heeren, J., Mewawala, D., Schnitgerhans, T.,Salomon, G., Enrich, C., Beisiegel, U., and Jackle, S. (2000).Annexin VI stimulates endocytosis and is involved in the traf-ficking of low density lipoprotein to the prelysosomal compart-ment. J. Biol. Chem. 275, 33806–33813.

3. Goldstein, J. L., Basu, S. P., and Brown, M. S. (1983). Receptor-mediated endocytosis of low-density lipoprotein in culturedcells. Methods Enzymol. 98, 241–260.

44. Pitas, R. E., Boyles, J., Mahley, R. W., and Bissell, D. M. (1985).Uptake of chemically modified lipoproteins in vivo is mediatedby specific endothelial cells. J. Cell Biol. 100, 103–117.

45. McFarlane, A. S. (1958). Efficient trace labeling of proteins withiodine. Nature 182, 53–57.

46. Grimaldi, K. A., Hutton, J. C., and Siddle, K. (1987). Productionand characterization of monoclonal antibodies to insulin secre-tory granule membranes. Biochem. J. 245, 557–566.

47. Sambrook, J., Fritsche, E. F., and Maniatis, T. (1989). “Molec-ular Cloning. A Laboratory Manual,” 2nd ed., Cold Spring Har-bor Laboratory, Cold Spring Harbor, NY.

48. Kim, M. J., Dawes, J., and Jessup, W. (1994). Transendothelialtransport of modified low-density lipoproteins. Atherosclerosis108, 5–17.

49. Green, S. A., Zimmer, K. P., Griffiths, G., and Mellman, I.(1987). Kinetics of intracellular transport and sorting of lyso-somal membrane and plasma membrane proteins. J. Cell Biol.105, 1227–1240.

50. Belcher, J. D., Hamilton, R. L., Brady, S. E., Hornick, C. A.,Jackle, S., Schneider, W. J., and Havel, R. J. (1987). Isolationand characterization of three endosomal fractions from the liverof estradiol-treated rats. Proc. Natl. Acad. Sci. USA 84, 6785–6789.

51. Jackle, S., Runquist, E. A., Brady, S., Hamilton, R. L., andHavel, R. J. (1991). Isolation and characterization of threeendosomal fractions from the liver of normal rats after lipopro-

tein loading. J. Lipid Res. 32, 485–498.

52. Pol, A., and Enrich, C. (1997). Membrane transport in rat liverendocytic pathways: Preparation, biochemical properties andfunctional roles of hepatic endosomes. Electrophoresis 18,2548–2557.

53. Kaelin, W. G., Jr., Pallas, D. C., DeCaprio, J. A., Kaye, F. J., andLivingston, D. M. (1991). Identification of cellular proteins thatcan interact specifically with the T/E1A-binding region of theretinoblastoma gene product. Cell 64, 521–532.

54. Barwise, J. L., and Walker, J. H. (1996). Annexins II, IV, V andVI relocate in response to rises in intracellular calcium inhuman foreskin fibroblasts. J. Cell Sci. 109, 247–255.

55. Mollers, C., Drobnik, W., Resink, T., and Schmitz, G. (1995).High-density lipoprotein and low-density lipoprotein-mediatedsignal transduction in cultured human skin fibroblasts. Cell.Signal. 7, 695–707.

56. Allen, S., Khan, S., Al-Mohama, F., Batten, P., and Yacoub, B.(1998). Native low density lipoprotein-induced calcium tran-sients trigger VCAM-1 and E-selectin expression in culturedhuman vascular endothelial cells. J. Clin. Invest. 101, 1064–1075.

57. Cockerill, G. W., and Reed, S. (1999). High-density lipoprotein:Multipotent effects on cells of the vasculature. Int. Rev. Cytol.188, 257–297.

58. Hawkins, T. E., Roes, J., Rees, D., Monkhouse, J., and Moss,S. E. (1999). Immunological development and cardiovascularfunction are normal in annexin VI null mutant mice. Mol. CellBiol. 19, 8028–8032.

59. Harris, A., Croall, D. E., and Morrow, J. S. (1988). The calm-odulin-binding site in a-fodrin is near the calcium-dependentprotease-I cleavage site. J. Biol. Chem. 263, 15754–15761.

60. Harris, A., Croall, D. E., and Morrow, J. S. (1989). Calmodulinregulates fodrin susceptibility to cleavage by calcium-depen-dent protease-I. J. Biol. Chem. 264, 17401–17408.

61. Harris, A. S., and Morrow, J. S. (1990). Calmodulin and cal-cium-dependent protease I coordinately regulate the interaction

of fodrin with actin. Proc. Natl. Acad. Sci. USA 87, 3009–3013.

eceived February 16, 2001evised version received May 2, 2001ublished online July 31, 2001