Identification of the Molecular Mechanisms Contributing to Polarized Trafficking in Osteoblasts

Identification of the Molecular Mechanisms Contributing to Polarized Trafficking in Osteoblasts

Experimental Cell Research 282, 24 –34 (2003) doi:10.1006/excr.2002.5668 Identification of the Molecular Mechanisms Contributing to Polarized Traffic...

739KB Sizes 1 Downloads 10 Views

Experimental Cell Research 282, 24 –34 (2003) doi:10.1006/excr.2002.5668

Identification of the Molecular Mechanisms Contributing to Polarized Trafficking in Osteoblasts Cecilia M Preˆle,* ,1 Michael A. Horton,* Paul Caterina,† and Gudrun Stenbeck* ,2 *Bone and Mineral Centre, Royal Free and University College Medical School, The Rayne Institute, London WC1E 6JJ, United Kingdom; and †Anatomical Pathology, The Western Australian Centre for Pathology and Medical Research, Perth, Western Australia, Australia

gaged and free plasma membrane in order to correctly direct vesicular flow. In osteoblasts, matrix proteins are secreted away from neighboring capillaries toward the existing bone surface. ECM proteins are extensively modified in compartments of the secretory pathway, the ER and Golgi apparatus, prior to secretion [1]. Signals, generated by subsequent cell– cell and cell–ECM adhesion, converge to modulate the intracellular trafficking machinery, which is responsible for the establishment of polarized secretion [2]. The current model for intracellular trafficking is that vesicular intermediates bud from the membrane of a starting compartment (i.e., the transGolgi network (TGN)) and subsequently fuse with the membrane of a different, more distal, compartment (i.e., the plasma membrane) [3]. Specificity in the fusion events is envisaged to be mediated by membrane proteins (SNAREs) that are specific for either the vesicle (v-SNARE) or the target membrane (t-SNARE) and by members of the Rab family of small GTP-binding proteins [4]. The interaction of a v-SNARE with its cognate t-SNARE brings the two membranes into close proximity allowing fusion to take place [5]. To date, a large number of v- and t-SNARE pairs have been identified, all localizing to specific intracellular compartments. The best characterized interaction is the one occurring at the presynaptic nerve terminal, which involves the v-SNAREs VAMP1/2, and the t-SNAREs SNAP25 and syntaxin 1. Although expression of these proteins was initially believed to be restricted to the nervous system, a number of studies have demonstrated their expression in other tissues [6]. Most recently, SNAP25 and VAMP expression has been linked to glutamate secretion in osteoblasts [7]. In addition to SNAREs, a number of cytoplasmic proteins are crucial for membrane fusion. Of these accessory proteins, NSF (N-ethylmaleimide sensitive fusion protein) [8] and ␣-SNAP (soluble NSF attachment protein) [9], have been identified in several cell types and form part of the conserved core of proteins required for all membrane fusion events. Additionally, vesicular trafficking to the sites of cell-to-cell contact on the plasma membrane has recently been shown to be dependent on the

The directionality of matrix deposition in vivo is governed by the ability of a cell to direct vesicular flow to a specific target site. Osteoblastic cells direct newly synthesized bone matrix proteins toward the bone surface. In this study, we dissect the molecular mechanisms underlying the polarized trafficking of matrix protein in osteoblasts. We demonstrate using TEM, immunocytochemistry, and cDNA analysis, the ability of osteoblastic cells in culture to form tight junctionlike structures and report the expression of the tight junction associated proteins occludin and claudins 1–3 in these cells. We identify intercellular contact sites and the leading edge of migratory osteoblasts as major target sites of vesicular trafficking in osteoblasts. Proteins required for this process, rsec6, NSF, VAMP1, and syntaxin 4, as well as the bone matrix protein, osteopontin, localize to these sites. We demonstrate that osteoblasts in vivo possess VAMP1 and, furthermore, report the expression of two VAMP1 splice variants in these cells. In addition, osteoblasts express the NSF attachment protein ␣-SNAP and the t-SNARE SNAP23. Thus, cell-to-cell contact sites and the leading edge of migratory osteoblasts contain a unique complement of proteins required for SNARE mediated membrane fusion. © 2003 Elsevier Science (USA) Key Words: osteoblast; SNARE; leading edge; secretion; trafficking; junction.

INTRODUCTION

Matrix deposition is an important step in the formation of all solid tissues and organs. The interaction of cells with the extracellular matrix (ECM) provides the cells with cues critical for the determination of cell fate: proliferation, differentiation, and apoptosis. In vivo matrix deposition is a highly directional process, which requires the cell to distinguish between contact-en1 Present address: Department of Craniofacial Development, Guy’s and St. Thomas’ and Kings College Hospitals, Kings College London, London SW1 9RT, United Kingdom. 2 To whom correspondence and reprint requests should be addressed. Fax: ⫹44 20 7679 6219. E-mail: [email protected]

0014-4827/03 $35.00 © 2003 Elsevier Science (USA) All rights reserved.

24

POLARIZED TRAFFICKING IN OSTEOBLASTS

rsec6/rsec8 complex [2]. This protein complex undergoes relocation from the cytosol to sites of cell– cell contact upon the formation of adherens junctions. Interruption of this process blocks the delivery of basolateral plasma membrane proteins and therefore the maintenance of polarity in epithelial cells [10]. These studies highlight the tight interplay between vesicular trafficking and intercellular junctions. In osteoblasts, a number of morphological studies have demonstrated the formation of intercellular junctions [11, 12]. Gap-junction formation [13] and cadherin expression have been shown to be crucial for osteoblastic differentiation [14] and are developmentally regulated. The polarized nature of matrix-secreting osteoblasts is evident in transmission electron micrographs and was recently confirmed by studies using differentially targeted enveloped viruses, influenza virus, and vesicular stomatitis virus [15]. However, the mechanism responsible for the maintenance of polarization in osteoblasts has not been elucidated. In epithelial cells, spatial cues are provided by cadherin mediated cell-to-cell contacts and integrin mediated cell– matrix contacts, with the different plasma membrane domains being separated by tight junction membrane proteins, the claudins, and occludin [16]. However, in neuronal and osteoclast cells, discrete plasma membrane domains can be detected even without the apparent presence of these “fence” proteins [17]. We used the rat osteosarcoma cell line, ROS 17/2.8, and primary osteoblastic cells isolated from Day 21 embryonic rat calvaria (RCOB) as model systems of targeted matrix secretion in osteoblasts. Characterization of ROS 17/2.8, an osteoblast-like clonal cell line derived from a rat osteosarcoma [18] has revealed that these cells express a number of mature osteoblast phenotype markers and are phenotypically stable [18, 19], thus these cells are an ideal control for the osteoblastic phenotype. In contrast to the availability and phenotypic stability of ROS 17/2.8, isolated RCOB cells are initially a heterogeneous cell population. However, incubation of RCOB cells in the presence of dexamethasone and ascorbic acid for 21 days results in the selection of osteoblastic cells and their progression along the osteoblast lineage. During the culture period, three distinct phases of maturation can be distinguished, proliferation (during the first 5 days of culture), differentiation, and nodule formation (Days 6 to 14) and nodule mineralization (Days 15 to 22) [20]. In mineralizing cultures, cells of the monolayer are confluent with cubodial morphology, whereas cells that are part of the nodule form multilayers [11]. Here, we characterized junction formation between osteoblasts. We established that ROS 17/2.8 and primary osteoblasts express, in addition to occludin, members of the claudin protein family, which are required for tight junction formation. We also investigated the

25

localization of the rsec6/rsec8 complex and SNARE mediated trafficking in these cells and observe a striking polarization of vesicular trafficking toward the leading edge of motile osteoblasts. MATERIALS AND METHODS Cells and cell culture. ␣-Minimal essential medium (␣-MEM) or Dulbecco’s modified essential medium (DMEM) containing Glutamax, heat-inactivated fetal calf serum (FCS), 100 U/ml penicillin/ 100 ␮g/ml streptomycin solution (PS), and 50 mg/ml gentamicin solution were from Gibco (Paisley, UK), ␤-glycerophosphate (␤-GP) was from Fluka Biochemica (UK), dexamethasone (Dex) and L-ascorbate (AA) were from Sigma (UK). NGF differentiated PC12 cells were kindly prepared and supplied by J. Herreros (Cancer Research UK, London). Highly purified rat synaptic vesicles were prepared and donated by S.L. Osborne (Cancer Research UK, London). Antibodies for immunocytochemistry and Western blotting. Antiosteocalcin antibody was a kind gift from K. Bouillion (Belgium). Anti-␣-SNAP and anti-NSF (6E6) antibodies were kindly supplied by J.E. Rothman (Memorial Sloan-Kettering Cancer Center, New York, NY) and anti-VAMP1 antibodies were donated by G. Schiavo (Cancer Research UK, London). Anti-osteopontin antibodies (MPIIIB10) were from Developmental Studies Hybridoma Bank at the University of Iowa (USA), antibodies against VAMP1, SNAP23, and syntaxin 4 were from Synaptic Systems GmbH (Go¨ ttingen, Germany). Anti-pan-cadherin antibodies were from Sigma (UK), anti-ZO-1 antibodies were from Zymed (UK) and anti-rsec6 antibodies were from StressGen Biotechnologies (Canada). Rhodamine-conjugated phalloidin and Alexa 488 conjugated secondary antibodies were obtained from Molecular Probes (Leiden, The Netherlands). Goat anti-mouse and goat anti-rabbit horseradish peroxidase (HRP) conjugated secondary antibodies were from BioRad (UK) and PBS antifade solution was obtained from Citifluor (Cambridge, UK). Cell isolation and culture. The rat osteosarcoma cell line, ROS 17/2.8, was cultured in DMEM (Glutamax) supplemented with 5% heat inactivated FCS and 1% PS. Rat primary osteoblast-like cells (RCOBs) were obtained by sequential collagenase digestion of Day 21 embryonic rat (Wistar) calvaria [21]. Briefly, rats were sacrificed by cervical dislocation and calvaria were dissected free of connective tissue and serially digested in 3 mg/ml collagenase solution. Cell populations obtained from five sequential digestions were cultured overnight in ␣-MEM, 15% FCS, 1% PS, and 25 ␮g/ml gentamicin (standard medium). Following cell attachment, cell fractions II through to V were trypsinized, pooled, and plated directly onto serum-coated glass coverslips for immunocytochemistry. Cells were maintained in culture for 24 h in standard medium conditions and subsequently for a further 21 days in standard media supplemented with 50 ␮g/ml AA, 10 ⫺8 M Dex and 10 mM ␤-GP to stimulate nodule formation and mineralization in vitro. The osteoblastic phenotype of these cells was verified using alkaline phosphatase histochemical staining. Transmission electron microscopy. ROS 17/2.8 cells were fixed as a continuous monolayer in the tissue culture flask with 2.5% glutaraldehyde in 0.05 M cacodylate buffer, pH 7.4, overnight at room temperature following removal of the culture medium. The adherent cells were dislodged with a rubber policeman and formed into a cell pellet [22]. The pellet was postfixed with 1% osmium tetroxide (OsO 4) for 1 h, dehydrated in increasing grades of alcohol, and embedded in araldite (Fluka AG, Switzerland). Ultrathin sections (100 –150 nm) were cut on a LKB microtome (Amersham Biosciences, Sweden), contrast enhanced with saturated aqueous uranyl acetate for 30 min, washed thoroughly in double distilled water, and then contrast enhanced with Reynold’s lead citrate. The sections were

26

ˆ LE ET AL. PRE

examined with a Phillips 410 transmission electron microscope at an accelerating voltage of 80 kV. Immunocytochemistry and confocal laser scanning microscopy (CLSM). ROS 17/2.8 and RCOBs cells were fixed in 4% paraformaldehyde and simultaneously permeabilized and blocked in saponin permeabilization buffer (0.5% BSA, 0.1% saponin, 0.02% sodium azide in phosphate buffered saline (PBS)) prior to staining. Immunostaining was performed for 2 h at 4°C as described [23]. Following sequential washing steps, cells were incubated with Alexa 488-conjugated secondary antibodies. Simultaneous staining for F-actin was performed by adding 1 U/ml of rhodamine-conjugated phalloidin. Fluorescent antibody distribution was monitored with a Leica TCS NT confocal laser scanning microscope (Heidelberg, Germany) using standard filter settings. Images were collected for Alexa 488- and TRITC- fluorochromes using BF 530/30 nm and BP 600/30 nm filters, respectively. Images were collected in the X–Y (horizontal) plane and displayed as single or dual fluorescence images. The thickness of the optical section was calculated with the Leica TCS NT software and set to 0.486 ␮m. For some images the X–Y sections were added together with the extended focus option of the TCS NT software to allow visualization of all parts of the cells. Western blot analysis. Cells were dislodged with a rubber policeman and harvested by centrifugation. Cells were resuspended in SDS loading buffer giving a final cell concentration of ⬃1 ⫻ 10 5 cells per lane. Proteins were resolved by SDS-PAGE and subsequently transferred onto polyvinylidene fluoride (PVDF) membrane. After transfer, the PVDF membrane was stained with Ponceau S (0.1% Ponceau S in 5% acetic acid) to reveal protein loading. Following an initial blocking step (5% dried skimmed milk, 0.02% sodium azide in PBS), membranes were probed with primary antibody for 2 h at room temperature. Postwashing in PBS/Tween, (0.05%), membranes were incubated with HRP-conjugated secondary antibody and bound peroxidase activity was detected using the enhanced chemiluminescence (ECL) detection system (Amersham Biosciences, UK). RT-PCR analysis. Reverse transcription PCR (RT-PCR) was performed on total RNA isolated from cultured ROS 17/2.8 or RCOB cells using TRIzol reagent (Invitrogen, UK). RT-PCR, using Superscript II Reverse transcriptase (Invitrogen, UK) was performed using both random hexamer (RH) and oligo dT primers (dT) (Amersham Biosciences, UK) on 5 ␮g of total RNA. PCR was performed (35 cycles) using 0.25 ␮M of either the gene-specific primers: 5⬘-␤Actin-TGTATGCCTCTGGTCGTACCAC; 3⬘-␤-Actin-ACAGGTACTTGCGCTCAGGAG; 5⬘-occludin-CACTATGAAACCGACTACACGAC; 3⬘-occludin-CCGTCTGTCATAGTCTCCCAC; or of the degenerated primers: 5⬘-claudin1/2-GYTGGKGMCARCATYGTGAC; 3⬘-claudin1/ 2-CAGGARMAGSAAAGKAKGAC; 5⬘-claudin3-GCTGGCTGTGCACCATCGTGC; 3⬘-claudin 3-GCCACGATGGTGATCTTGGCCT. VAMP1 3⬘ RACE reaction. RACE reaction (rapid amplification of cDNA ends) was used to amplify partial cDNA ends between a known point and the 3⬘ end of the VAMP1 cDNA transcript. Total RNA isolated from osteoblastic cell cultures was reverse transcribed using Superscript II reverse transcriptase and a hybrid oligo dT primer (R T) containing 17 dTs followed by a 35 base oligonucleotide sequence (CCAGTGAGCAGAGTGACGAGGACTCGAGCTCAAGCTTTTTTTTTTTTTTTTT) containing the anchor primers R0 (CCAGTGAGCAGAGTGACG) and R1 (GAGGACTCGAGCTCAAGC). Subsequent amplifications were performed in accordance with the classical 3⬘ RACE reaction method described by [24], using PCR and a combination of gene specific and anchor primers, 1st round: 5⬘VAMP1 and R0, 2nd round VAM3 (AGGCAGGAGCGTCAGTGTTTGAGA) and R1. Amplified PCR products were resolved by agarose gel electrophoresis and purified using the QIAquick gel extraction kit (Qiagen, UK). cDNA obtained from gel purified PCR products was ligated into pGEM-T cloning vector (Promega, UK) and subsequently transformed into competent E. coli JM109 cells (Promega, UK). Insert-containing transformants were subcloned, purified, and sequenced.

Sequencing and sequence analysis. Sequencing was performed using the ABI PRISM dye-terminator cycle sequencing ready reaction kit (Perkin Elmer, UK). Approximately 1 ␮g of DNA template was added to 3.2 pmol of sequencing primers (either T7 or Sp6) and amplified (25 cycles). Dye terminators were removed by ethanol precipitation, and the pellet was dried, resuspended, and loaded onto a sequencing gel. Sequence analysis was performed using EditView and Sequence Navigator analysis packages (ABI PRISM, Perkin Elmer, UK). The sequences obtained were compared to existing cDNA sequences in the sequence database using the BLAST search tool (www.ncbi.nlm.nih.gov/BLAST/).

RESULTS

Ultrastructural Analysis of ROS 17/2.8 cell Monolayers Transmission electron microscopy (TEM) revealed closely apposed ROS 17/2.8 cells. Cellular interactions in the form of subplasmalemmal densities were identified between abutting osteoblastic cells (Fig. 1A, arrowhead). Focal cell-to-cell contacts, reminiscent of tight junctions present in other specialized cell types, were identified at contact points between plasma membranes of adjacent osteoblasts. (Fig. 1A, inset, arrowhead). A prerequisite for tight junction formation is the presence of adherens junctions. We confirmed cadherin expression and localization in ROS 17/2.8 cells with a pan-cadherin antibody. In subconfluent ROS 17/2.8 cells, cadherin expression is localized to the plasma membrane (Fig. 1B, a) and in a punctuate manner, along the extended cellular processes (Fig. 1B, a arrow). ZO-1 is a peripheral membrane protein integral to both tight and adherens junction complexes. In confluent RCOB cell monolayers, ZO-1 concentrated to the plasma membrane at sites of cell-to-cell contact (Fig. 1B, b arrow) supporting the data obtained with TEM. Furthermore, Western blot analysis revealed that, in addition to the integral membrane proteins associated with apical junction formation, osteoblastic cells express the adherens junction associated proteins ␣- and ␤-catenin and the cytoskeletal attachment proteins ␣-actinin and vinculin (data not shown) [25, 26]. Subsequent RT-PCR revealed the expression of mRNAs for the tight junction associated membrane proteins; occludin, claudins 1, 2, and 3 in ROS 17/2.8, and both proliferating and mineralizing RCOB cells (Fig. 1C). Osteoblastic cells do not express claudins 4, 5, or 8 (data not shown). Rat liver cDNA samples were used as positive controls. The partial cDNAs identified were identical to existing occludin, claudins 1, 2, and 3 sequences. Interestingly, the amount of claudin 1/2 PCR product for the mineralizing RCOB sample compared to the proliferating RCOB sample was increased, indicating a possible differential expression of claudins 1/2 during osteoblast maturation.

POLARIZED TRAFFICKING IN OSTEOBLASTS

27

FIG. 1. Tight junction-like formation in osteoblastic cells. (A) Transmission electron microscopy (TEM) revealed closely apposed ROS 17/2.8 cells grown in monolayer. Focal areas of cell-to-cell contact were evident between the adjacent cells (arrowhead, OB— osteoblast, PM—plasma membrane). The interaction between the plasma membranes from the two apposing cells is reminiscent to that observed in tight junction-like structures (arrowhead in inset). (B) CLSM localized the apical junction complex associated proteins, cadherin (a) and ZO-1 (b), toward the cell periphery in both ROS17/2.8 and RCOB cells. In subconfluent ROS 17/2.8 cells pan-cadherin immunoreactivity was also detected in a punctate manner along extended cellular processes (arrow in a). In confluent RCOBs, ZO-1 localized to the plasma membrane, at sites of cell-to-cell contact (arrowhead in b). Images shown are representative single sections. (C) RT-PCR revealed the expression of mRNAs for the tight junction associated membrane proteins, occludin, and claudins 1, 2, and 3 in ROS 17/2.8, proliferating (p) and mineralizing (m) RCOB cells. Rat liver cDNA samples were used as positive controls and ␤-actin to access the quality and quantity of the cDNA.

Localization of the rsec6 Subunit of the rsec6/rsec8 Complex in Osteoblastic Cells The correct localization of the rsec6/rsec8 docking complex has been shown to be crucial for epithelial cell polarity. This multisubunit complex is associated with

junctional complexes in MDCK cells and shuttles between the TGN and the plasma membrane in epithelial-like NRK cells [2, 27]. Here we demonstrate a potential role for the rsec6/rsec8 docking complex in vesicular trafficking in osteoblasts (Fig. 2). Using im-

28

ˆ LE ET AL. PRE

FIG. 2. Localization of the rsec6 subunit of the rsec6/sec8 docking complex and of the secreted bone matrix protein osteopontin in osteoblastic cells with immunocytochemistry and CLSM. (A) In subconfluent ROS 17/2.8 cells, rsec6 (a and c, green) was detected in the cytoplasm and at the cell periphery, in particular along the cellular projections and toward the leading edge of the cell (arrows in a and c, green). (B) In RCOBs, rsec6 (a and c, green) immunoreactivity was detected in the cytoplasm in the extended perinuclear region and along cellular protrusions and at their tips (arrows in a and c). (C) Expression of osteopontin (a and c, green) was detected throughout the cytoplasm of RCOB cells, in the biosynthetic organelles of the cell but labeling was more pronounced toward the cell periphery and very strong at cellular protrusions and at the leading edge of the cells (arrows in a and c). F-actin distribution was visualized with rhodamine-phalloidin (A–C, b and c, red). Images shown are representative single sections taken with CLSM.

munocytochemistry and confocal laser scanning microscopy, the rsec6 subunit was localized in both ROS 17/2.8 (Fig. 2A, a and c) and RCOB cells (Fig. 2B a and

c). In subconfluent ROS17/2.8 cells rsec6 has a cytoplasmic distribution but is also detected along cellular projections and toward the leading edge of the cell (Fig.

POLARIZED TRAFFICKING IN OSTEOBLASTS

29

FIG. 3. Osteoblastic cells express NSF, ␣-SNAP, and the plasma membrane t-SNAREs SNAP23 and syntaxin 4. (A) Western blot analysis of extracts of ROS 17/2.8 and proliferating (p) and mineralizing (m). PC12 cells were used as positive control for NSF and ␣-SNAP expression. (B) Western blot analysis of cell extracts from ROS 17/2.8, crude, proliferating (p) and mineralizing (m) RCOB cells is shown. Membranes were also probed with anti-GAPDH antibodies to assess protein loading. (C) The intracellular localization of syntaxin 4 (a and c, green) was determined in proliferating RCOBs using immunocytochemistry and CLSM. Two prevalent pools of syntaxin 4 were detected, one located at

30

ˆ LE ET AL. PRE

2A, a and c, arrows). Consistently, rsec6 immunoreactivity was detected in the cytoplasm of proliferating RCOB cells, predominantly in the extended perinuclear region, though rsec6 was also detected along cellular protrusions and at their tips (Fig. 2B, a and c, arrows). As with rsec6, immunoreactivity for the secreted bone matrix protein, osteopontin, was detected throughout the cytoplasm of RCOB cells, in the biosynthetic organelles of the cells but labeling was more pronounced toward the cell periphery and very strong at cellular protrusions (Fig. 2C, a and c, arrows). Factin distribution was used to define the peripheral boundaries of the cells and was visualized using rhodamine conjugated phalloidin (Fig. 2A–C, b and c, overlay). SNARE Expression in Osteoblasts NSF and ␣-SNAP are ubiquitously expressed proteins that are essential for membrane fusion [8, 9]. Western blot analysis of total cell extracts of ROS 17/2.8, proliferating and mineralizing RCOB cells demonstrated NSF and ␣-SNAP expression in osteoblastic cells (Fig. 3A). NSF localization to the membrane depends on its binding to SNARE complexes [3]. To identify the SNARE proteins involved in NSF localization in osteoblasts we used an antibody screening approach. Plasma membrane localized t-SNARE proteins are syntaxins 1, 2, 3, and 4, and SNAP25 and SNAP23 [28]. In contrast to previous reports, we detected SNAP23 but not SNAP25 expression in osteoblastic cells [7]; this discrepancy may reflect the level of species variation between the cell populations studied. Expression levels of SNAP23 did not alter during nodule formation and mineralization in the primary cultures (Fig. 3B). In other cell types, SNAP23 has been shown to interact with syntaxins 2, 4, 6, and 11 [29 –32]. Osteoblasts have recently been shown to express syntaxin 4 and syntaxin 6, where syntaxin 6 was shown to localize to the perinuclear region and syntaxin 4 was observed at intercellular contact sites [7]. Syntaxin 4 expression in osteoblasts was confirmed in both ROS 17/2.8 cells and in both proliferating and mineralizing RCOB cells (Fig.

3B). Furthermore, in proliferating RCOB cells, we observed two prevalent pools of syntaxin 4, one localizing to the perinuclear region of the Golgi apparatus (Fig. 3C, a and c, arrowheads) and one at the plasma membrane, which predominantly localized toward the leading edge of the osteoblasts (Fig. 3C, a and c, green, arrows). The additional low levels of cytoplasmic staining may represent vesicular carriers shuttling between the TGN and the plasma membrane. Taken together, our data support that the leading edge of osteoblasts is the predominant site for membrane fusion and matrix deposition in these cells, a process likely to be mediated by specific SNARE pairing. VAMP1 Expression in Osteoblasts v-SNAREs that localize to vesicles carrying cargo to the plasma membrane are VAMP1, 2, 5, and 7 [28]. Initially thought to be brain specific, VAMP1 and 2 have been shown to be expressed, though to varying degrees, in a wide variety of cell types, including osteoblasts [7, 23]. Western blot analysis with an antibody directed against the N-terminal portion of VAMP1, showed that ROS 17/2.8 cells express this protein. However, the apparent molecular weight of the protein is 2 kDa larger (Fig. 4A, arrowhead) than that of the VAMP1 detected on purified rat synaptic vesicles (SVs) (Fig. 4A, arrow). PCR and sequence analysis of cDNA derived from ROS 17/2.8 and RCOB cells with primers specific for rat brain VAMP1 revealed that the N-terminal 81% of the transcript expressed in osteoblasts is identical to neuronal VAMP1A (data not shown). Amplification of the extreme C-terminal portion of both osteoblastic and rat brain VAMP1 transcripts, using 3⬘ RACE reaction, led to the identification of two VAMP1 splice variants in osteoblastic cells, rVAMP1B (previously identified in human endothelial cells [33]) and VAMP1OB. The sequences obtained were translated into protein sequence (Fig. 4B). Assuming a similar gene structure of rat VAMP1 to human VAMP1, both rVAMP1B and VAMP1-OB appear to be products of alternative splicing between exons IV and V (Fig. 4B, arrow) and

the perinuclear region (arrowheads in a and c) and one at the plasma membrane. The plasma membrane pool of syntaxin 4 predominantly localized toward the leading edge of the osteoblasts (arrow in a and c). F-actin distribution, detected using rhodamine-conjugated phalloidin, was used to define the peripheral boundaries of the cells (b and c, red). Areas of colocalization of syntaxin 4 with F-actin are shown in the overlay in yellow. Images shown are merged images of 15 ⫻ 0.5 ␮m sections taken with CSLM as described in Materials and Methods. FIG. 4. VAMP1 expression in osteoblasts. (A) Western blot analysis revealed ROS 17/2.8 cells to express an isoform of the v-SNARE VAMP1 with an apparent molecular weight 2 kDa larger in size (arrowhead) than the protein expressed on purified rat synaptic vesicles (SVs) (arrow). (B) PCR amplification of the C-terminal portion of osteoblastic and rat brain VAMP1 cDNAs using 3⬘RACE reaction led to the identification of two VAMP1 splice variants in osteoblastic cells, rVAMP1B [33] and the novel VAMP1-OB (GenBank accession number AF498262). These splice variants are products of alternative splicing between exons IV and V and result in altered C-terminal residues (bold). (C) VAMP1 expression was examined in proliferating RCOBs with immunocytochemistry and CLSM. VAMP1 immunoreactivity (a and c, green) was detected on vesicle-like structures in the perinuclear region and also toward the cell periphery of proliferating RCOB cells (a and c, green, arrow head and arrow, respectively). F-actin, detected using rhodamine-conjugated phalloidin, was used to define cell boundaries (b and c, red). Images shown are representative single sections taken with CLSM.

POLARIZED TRAFFICKING IN OSTEOBLASTS

consequently, contain altered C-terminal residues. VAMP1-OB contains the novel C-terminal residues RQD. A polyclonal antibody directed against the N-terminal 33 amino acids of VAMP1 revealed a vesicle-like distribution of VAMP1 in the perinuclear region and also toward the cell periphery of proliferating RCOB cells (Fig. 4C, a and c, green, arrow head and arrow, respectively); this indicates that, as observed with rsec 6 and syntaxin 4 localization, there are at least two pools of VAMP1 in these cells. F-actin was visualized using rhodamine-conjugated phalloidin (Fig. 4C, b and c, red) and defined the cell boundaries. VAMP1 Localization in Vivo Serial frozen sections (7 ␮m) of long bones derived from 3-day-old rat pups were stained immunohistochemically using anti-VAMP1 or anti-alkaline phosphatase antibodies. Alkaline phosphatase expression was used to identify mature osteoblastic cells (Fig. 5A). Low magnification identified alkaline phosphatase positive cells within the trabecular bone (TB) and within hypertrophic chondrocytes (HC) at the growth plate. Higher magnification demonstrated alkaline phosphatase expression within plump cuboidal cells lining the trabecular bone surface (data not shown). Similarly, VAMP1 localizes to the trabecular bone (TB) (Fig. 5B). VAMP1 expression is low in the surrounding skin (S) and hypertrophic chondrocytes (HC). Higher magnification identified VAMP1 positive cells to be distributed along the trabecular bone surface, representative of osteoblasts lining the bone surface (Fig. 5C, arrowheads). These findings were confirmed in serial sections of embryonic Day 21 rat calvaria (data not shown). DISCUSSION

Intercellular junctions are important signal transduction sites and play a crucial role in the establishment of cell polarity and polarized membrane trafficking [2]. Osteoblastic cells have been shown to form a variety of intercellular junctions, which are important for osteoblast differentiation [13, 14]. We identified areas of tight junction-like contacts between cellular processes of subconfluent osteoblastic cells, indicating that junction formation occurs at initial cell-to-cell contact sites. In addition, the adherens junction protein cadherin was detected at cellular extensions in subconfluent osteoblastic cells, identifying them as initiation sites of junction formation. In confluent osteoblastic cells, the contact sites are distributed along the plasma membrane forming discrete points of cell-to-cell contact. In keeping with this, we detected mRNA expression

31

of a subset of transmembrane proteins implicated in tight junction formation, claudins 1, 2, and 3 and occludin, in osteoblastic cell monolayers. Even though the data are not quantitative, the marked increase in claudin 1/2 PCR product for the mineralizing RCOB sample compared to the differentiating RCOB sample suggests that claudins 1/2 are differentially expressed during osteoblast maturation. Tight junctions are formed by the interaction of individual claudin family members, which have different effects on barrier function. Recently it has been shown that heterologous expression of claudin 2 in MDCK cells, which normally express only claudins 1 and 4, reduces transepithelial resistance [34]. Although further experiments are required to assess the combined effect of claudins 1–3 expression on tight junction barrier function in osteoblasts, our data expand previous findings characterizing junction formation in osteoblastic and odontoblastic monolayers [12, 25, 26]. Moreover, our results confirm the polarized nature of osteoblastic cells [15] and indicate that the molecular machinery utilized by osteoblastic and epithelial cells to achieve and maintain cell polarity is similar. Vesicular trafficking to the sites of cell-to-cell contact is important in establishing cell polarization. In nonpolarized cells, proteins are sorted in the Golgi complex but are delivered to random sites on the plasma membrane [17]. Upon establishment of cell-to-cell contacts, this process becomes directional. Directional vesicular trafficking requires spatial cues for the correct targeting and docking of the vesicles. The recently identified rsec6/8 complex has been shown to regulate the delivery of exocytotic vesicles to these sites [2], thereby identifying points of cell-to-cell contact as potential exocytosis sites. Expression of rsec8 has recently been described in osteoblastic cells, although its specific role in vesicular transport within these cells was not determined [7]. Here, we report the localization of the rsec6 subunit within both ROS 17/2.8 and RCOB cells. The rsec6 subunit is predominantly localized to the cytoplasm in the extended perinuclear region but can also be detected at cellular protusions and at their tips. TGN and plasma membrane pools of rsec6/8 complexes have recently been shown to be required for transport of cargo proteins from the perinuclear region of the cell to the periphery [27]. Interestingly, some proteins, such as transferrin receptor, integrins, and certain viruses are targeted to the leading edge of migratory cells where exocytosis takes place [35]. In keeping with this, we not only demonstrated localization of the rsec6 subunit but also of osteopontin to the cellular projections of osteoblastic cells. This provides evidence that, in addition to cell adhesion molecules, ECM proteins are also targeted to this area of the plasma membrane. To establish that the leading edge of migrating osteoblasts serves as an exocytosis site, we investigated the

32

ˆ LE ET AL. PRE

FIG. 5. Osteoblasts express VAMP1 in vivo. Frozen serial sections (7 ␮m) of 3-day-old rat long bone were stained immunohistochemically using anti-alkaline phosphatase (ALP) or anti-VAMP1 antibodies. (A) ALP expression was used to detect cells with a mature osteoblastic phenotype. Low magnification revealed ALP expressing cells to be located toward the trabecular bone and within hypertrophic chondrocytes (TB and HC, respectively). (B) VAMP1 expression localized toward the area of trabecular bone (TB) and was low in the surrounding skin (S) and hypertrophic chondrocytes (HC). (C) Higher magnification identified VAMP1 positive cells to be distributed along the trabecular bone surface (arrowheads), representative of osteoblasts lining the bone surface.

distribution of proteins necessary for vesicular trafficking. Western blot analysis revealed the expression of NSF, ␣-SNAP, VAMP1, SNAP23, and syntaxin 4 in osteoblastic cells. SNAREs and also NSF (data not shown) can be detected at the leading edge of migratory osteoblasts. Thus, the full complement of proteins required for SNARE-mediated membrane fusion was localized specifically to cell– cell contact sites and the leading edge of migratory cells. Although previously reported in human osteoblastic cells [7], we were un-

able to identify SNAP25 expression in rat derived osteoblast cell populations; this discrepancy may reflect the level of species variation between the cell populations studied. The v-SNARE VAMP is involved in the regulation of exocytosis in neuronal and other cell types [23]. In human osteoblastic cell lines, VAMP has been implicated in the regulation of glutamate exocytosis [7]. To date, the VAMP protein family contains nine members, two of which are closely related isoforms, VAMP1 and

33

POLARIZED TRAFFICKING IN OSTEOBLASTS

VAMP2 [36]. These two isoforms present an overall high degree of sequence homology with the exception of the N-terminus. Additionally, several splice variants of both VAMP1 and 2 have recently been identified [37, 38]. This alternative splicing alters the C-terminal residues, which in the case of human VAMP1B, influences its subcellular localization [33]. VAMP1 expression in cultured rat osteoblastic cells was detected with a VAMP1 specific antibody [23] and expression in vivo was confirmed by immunohistochemistry on sections of rat long bone and calvaria. Through sequence analysis, we identified the presence of two VAMP1 splice variants in rat osteoblastic cells: the previously reported rVAMP1B [39] and a novel splice variant VAMP1-OB, which is related to human VAMP1B [33]. However, we were not able to detect neuronal VAMP1A. To address the influence of the different C termini in localizing the isoforms, experiments expressing exogenous VAMP1A in osteoblastic cells could be used to compare the intracellular distribution of the different VAMP1 isoforms. The variation in primary sequence does not explain the molecular weight difference between osteoblastic and neuronal VAMP1 observed by Western blot. VAMP proteins have been shown to contain posttranslational modifications such as phosphorylation [40] and palmitoylation [41, 42], and the apparent difference in electrophoretic mobility may reflect one such modification. Further experimental work will be required to resolve the role of such posttranslational modifications on VAMP function in bone vs other tissues. During bone formation, some of the matrix-secreting osteoblasts become embedded in the newly formed matrix and terminally differentiate into osteocytes [43]. It is still unclear which signals mediate this change in phenotype and why only a certain proportion of osteoblasts becomes embedded. Due to the directionality of ECM deposition and its relevance in cellular behavior, it is reasonable to propose that an important factor for osteoblast– osteocyte transition is the loss of directionality in protein secretion, which would lead to matrix deposition around the cell. Regulated expression and/or modification of plasma membrane SNARE proteins together with a modified distribution of the rsec6/8 complex could form the basis of this transition, providing a novel link between vesicular trafficking and differentiation in osteoblasts. We thank S.E. Mutsaers, University of Western Australia, Perth, for help with sample preparation for EM, J.E. Aubin, University of Toronto, Canada, for introducing C.M.P to primary osteoblastic cell cultures, and the Molecular Neuropathobiology Laboratory, Cancer Research UK, for help with the sequencing and helpful discussions. We thank J. Price for critical reading of the manuscript and members of the Horton Laboratory for helpful discussions. This work was supported by the Oliver Bird Fund and the Arthritis Research Campaign (G.S.) and the Wellcome Trust (M.A.H). C.M.P. was a Wellcome Trust Prize Student.

REFERENCES 1.

Robey, P. G., Fedarko, N. S., Hefferan, T. E., Bianco, P., Vetter, U. K., Grzesik, W., Friedenstein, A., Van der Pluijm, G., Mintz, K. P., Young, M. F. et al. (1993). Structure and molecular regulation of bone matrix proteins. J. Bone Miner. Res. 8, S483– S487.

2.

Yeaman, C., Grindstaff, K. K., and Nelson, W. J. (1999). New perspectives on mechanisms involved in generating epithelial cell polarity. Physiol. Rev. 79, 73–98.

3.

Rothman, J. E. (1994). Mechanisms of intracellular protein transport. Nature 372, 55– 63.

4.

Chavrier, P., and Goud, B. (1999). The role of ARF and Rab GTPases in membrane transport. Curr. Opin. Cell Biol. 11, 466 – 475.

5.

Jahn, R., and Sudhof, T. C. (1999). Membrane fusion and exocytosis. Annu. Rev. Biochem. 68, 863–911.

6.

Lin, R. C., and Scheller, R. H. (2000). Mechanisms of synaptic vesicle exocytosis. Annu. Rev. Cell Dev. Biol. 16, 19 – 49.

7.

Bhangu, P. S., Genever, P. G., Spencer, G. J., Grewal, T. S., and Skerry, T. M. (2001). Evidence for targeted vesicular glutamate exocytosis in osteoblasts. Bone 29, 16 –23.

8.

Whiteheart, S. W., Schraw, T., and Matveeva, E. A. (2001). N-ethylmaleimide sensitive factor (NSF) structure and function. Int. Rev. Cytol. 207, 71–112.

9.

Stenbeck, G. (1998). Soluble NSF-attachment proteins. Int. J. Biochem. Cell Biol. 30, 573–577.

10.

Grindstaff, K. K., Yeaman, C., Anandasabapathy, N., Hsu, S. C., Rodriguez-Boulan, E., Scheller, R. H., and Nelson, W. J. (1998). Sec6/8 complex is recruited to cell– cell contacts and specifies transport vesicle delivery to the basal–lateral membrane in epithelial cells. Cell 93, 731–740.

11.

Bhargava, U., Bar-Lev, M., Bellows, C. G., and Aubin, J. E. (1988). Ultrastructural analysis of bone nodules formed in vitro by isolated fetal rat calvaria cells. Bone 9, 155–163.

12.

Palumbo, C., Palazzini, S., and Marotti, G. (1990). Morphological study of intercellular junctions during osteocyte differentiation. Bone 11, 401– 406.

13.

Lecanda, F., Warlow, P. M., Sheikh, S., Furlan, F., Steinberg, T. H., and Civitelli, R. (2000). Connexin43 deficiency causes delayed ossification, craniofacial abnormalities, and osteoblast dysfunction. J. Cell Biol. 151, 931–944.

14.

Cheng, S. L., Shin, C. S., Towler, D. A., and Civitelli, R. (2000). A dominant negative cadherin inhibits osteoblast differentiation. J. Bone Miner. Res. 15, 2362–2370.

15.

Ilvesaro, J., Metsikko, K., Vaananen, K., and Tuukkanen, J. (1999). Polarity of osteoblasts and osteoblast-like UMR-108 cells. J. Bone Miner. Res. 14, 1338 –1344.

16.

Tsukita, S., and Furuse, M. (1999). Occludin and claudins in tight-junction strands: Leading or supporting players? Trends Cell Biol. 9, 268 –273.

17.

Keller, P., and Simons, K. (1997). Post-Golgi biosynthetic trafficking. J. Cell Sci. 110, 3001–3009.

18.

Majeska, R. J., Rodan, S. B., and Rodan, G. A. (1980). Parathyroid hormone-responsive clonal cell lines from rat osteosarcoma. Endocrinology 107, 1494 –1503.

19.

Rodan, G. A., and Noda, M. (1991). Gene expression in osteoblastic cells. Crit. Rev. Eukaryot. Gene Expr. 1, 85–98.

20.

Aubin, J. E., Heersche, J. N., Merrilees, M. J., and Sodek, J. (1982). Isolation of bone cell clones with differences in growth, hormone responses, and extracellular matrix production. J. Cell Biol. 92, 452– 461.

ˆ LE ET AL. PRE

34 21.

22.

23.

24.

25.

26.

27.

28. 29.

30.

31.

Bellows, C. G., Aubin, J. E., Heersche, J. N., and Antosz, M. E. (1986). Mineralized bone nodules formed in vitro from enzymatically released rat calvaria cell populations. Calcified Tissue Int. 38, 143–154. Lazzaro, A. V. (1983). Technical note: Improved preparation of fine needle aspiration biopsies for transmission electron microscopy. Pathology 15, 399 – 402. Rossetto, O., Gorza, L., Schiavo, G., Schiavo, N., Scheller, R. H., and Montecucco, C. (1996). VAMP/synaptobrevin isoforms 1 and 2 are widely and differentially expressed in nonneuronal tissues. J. Cell Biol. 132, 167–179. Frohman, M. A., Dush, M. K., and Martin, G. R. (1988). Rapid production of full-length cDNAs from rare transcripts: Amplification using a single gene-specific oligonucleotide primer. Proc. Natl. Acad. Sci. USA 85, 8998 –9002. Luegmayr, E., Glantschnig, H., Varga, F., and Klaushofer, K. (2000). The organization of adherens junctions in mouse osteoblast-like cells (MC3T3-E1) and their modulation by triiodothyronine and 1,25- dihydroxyvitamin D3. Histochem. Cell Biol. 113, 467– 478. Lundgren, T., Nilsson, M., Ritchie, H. H., and Linde, A. (2001). Junctional proteins and Ca 2⫹ transport in the rat odontoblastlike cell line MRPC-1. Calcified Tissue Int. 68, 192–201. Yeaman, C., Grindstaff, K. K., Wright, J. R., and Nelson, W. J. (2001). Sec6/8 complexes on trans-Golgi network and plasma membrane regulate late stages of exocytosis in mammalian cells. J. Cell. Biol. 155, 593– 604. Chen, Y. A., and Scheller, R. H. (2001). SNARE-mediated membrane fusion. Nat. Rev. Mol. Cell Biol. 2, 98 –106. Valdez, A. C., Cabaniols, J. P., Brown, M. J., and Roche, P. A. (1999). Syntaxin 11 is associated with SNAP-23 on late endosomes and the trans-Golgi network. J. Cell Sci. 112, 845– 854. Quinones, B., Riento, K., Olkkonen, V. M., Hardy, S., and Bennett, M. K. (1999). Syntaxin 2 splice variants exhibit differential expression patterns, biochemical properties and subcellular localizations. J. Cell Sci. 112, 4291– 4304. St-Denis, J. F., Cabaniols, J. P., Cushman, S. W., and Roche, P. A. (1999). SNAP-23 participates in SNARE complex assembly in rat adipose cells. Biochem. J. 338, 709 –715.

Received April 9, 2002 Revised version received August 28, 2002 Published online November 11, 2002

32.

Wendler, F., and Tooze, S. (2001). Syntaxin 6: The promiscuous behaviour of a SNARE protein. Traffic 2, 606 – 611.

33.

Isenmann, S., Khew-Goodall, Y., Gamble, J., Vadas, M., and Wattenberg, B. W. (1998). A splice-isoform of vesicle-associated membrane protein-1 (VAMP-1) contains a mitochondrial targeting signal. Mol. Biol. Cell 9, 1649 –1660.

34.

Tsukita, S., Furuse, M., and Itoh, M. (2001). Multifunctional strands in tight junctions. Nat. Rev. Mol. Cell Biol. 2, 285–293.

35.

Bretscher, M. S. (1996). Moving membrane up to the front of migrating cells. Cell 85, 465– 467.

36.

Bock, J. B., Matern, H. T., Peden, A. A., and Scheller, R. H. (2001). A genomic perspective on membrane compartment organization. Nature 409, 839 – 841.

37.

Berglund, L., Hoffmann, H. J., Dahl, R., and Petersen, T. E. (1999). VAMP-1 has a highly variable C-terminus generated by alternative splicing. Biochem. Biophys. Res. Commun. 264, 777–780.

38.

Mandic, R., and Lowe, A. W. (1999). Characterization of an alternatively spliced isoform of rat vesicle associated membrane protein-2 (VAMP-2). FEBS Lett. 451, 209 –213.

39.

Mandic, R., Trimble, W. S., and Lowe, A. W. (1997). Tissuespecific alternative RNA splicing of rat vesicle-associated membrane protein-1 (VAMP-1). Gene 199, 173–179.

40.

Nielander, H. B., Onofri, F., Valtorta, F., Schiavo, G., Montecucco, C., Greengard, P., and Benfenati, F. (1995). Phosphorylation of VAMP/synaptobrevin in synaptic vesicles by endogenous protein kinases. J. Neurochem. 65, 1712–1720.

41.

Couve, A., Protopopov, V., and Gerst, J. E. (1995). Yeast synaptobrevin homologs are modified posttranslationally by the addition of palmitate. Proc. Natl. Acad. Sci. USA 92, 5987– 5991.

42.

Veit, M., Becher, A., and Ahnert-Hilger, G. (2000). Synaptobrevin 2 is palmitoylated in synaptic vesicles prepared from adult, but not from embryonic brain. Mol. Cell Neurosci. 15, 408 – 416.

43.

Aarden, E. M., Burger, E. H., and Nijweide, P. J. (1994). Function of osteocytes in bone. J. Cell Biochem. 55, 287–299.