| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 278, Issue 35, 33239-33247, August 29, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Department of Neurobiology, Yale University School of Medicine, New Haven, Connecticut 06510
Received for publication, April 3, 2003 , and in revised form, June 4, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Like all of the lecticans, the structure of BEHAB/brevican includes an N-terminal HA-binding domain, a middle chondroitin sulfate attachment region and a C-terminal selectin-like domain consisting of an epidermal growth factor-like, a C-type lectin, and a complement regulatory protein-like domains (10). Apart from HA, bound by the N-terminal domain of BEHAB/brevican, two other binding partners have been described. BEHAB/brevican binds the ECM glycoprotein tenascin-R and a subset of membrane sulfated glycolipids via its C-type lectin domain, both through calcium-dependent mechanisms (11, 12).
BEHAB/brevican mRNA expression increases over the course of rat brain development, reaching a plateau in adulthood (13). The roles proposed for BEHAB/brevican in normal developing brain include regulation of cell adhesion and neurite outgrowth and a role in synaptic plasticity (14, 15). Our studies, demonstrating a spike in BEHAB/brevican expression early in development in the ventricular zone coincident with gliogenesis (16) and an increase in BEHAB/brevican expression during reactive gliosis after a stab injury (17), have suggested a role for BEHAB/brevican in glial cell proliferation and motility. Consistent with these results, BEHAB/brevican expression is also dramatically up-regulated in surgical samples of glioma, notoriously invasive primary tumors of the central nervous system, as well as in a rodent glioma model (18). We have further shown that BEHAB/brevican up-regulation and its subsequent proteolytic processing contribute to the invasive phenotype of glioma (19, 20). An understanding of the molecular interactions and mechanisms through which these many functions are mediated is still incomplete.
One of the difficulties in characterizing the functions of BEHAB/brevican
is the molecular complexity of this protein. Two isoforms of BEHAB/brevican
are generated by alternative splicing, a full-length, secreted protein and a
C-terminally truncated splice variant
(21). The latter lacks the
entire C-terminal domain, which is replaced by an attachment sequence for a
GPI anchor. The middle region of BEHAB/brevican contains a conserved cleavage
site that is proteolyzed by members of the ADAMTS (disintegrin and
metalloproteinase with thrombospondin motifs) family of proteases
(22,
23). The full-length, secreted
BEHAB/brevican protein (
145–150 kDa), as well as its 90- and 50-kDa
cleavage products, are detected in the adult rat brain
(19,
24). In addition,
glycosylation variants of BEHAB/brevican have led to its characterization as a
"part time proteoglycan," because glycoforms are found both with
and without chondroitin sulfate GAG chains
(9,
24). It is likely that the
different isoforms of BEHAB/brevican created by alternative splicing,
proteolytic cleavage, or differential glycosylation may interact differently
with the cell surface and with other ECM components. To investigate the
functional role of BEHAB/brevican, we first need to gain a more complete
understanding of its molecular heterogeneity.
Here, we have identified a novel isoform of BEHAB/brevican, B/b130, with a developmental expression pattern that differs from all previously described BEHAB/brevican isoforms. Also unlike the previously described BEHAB/brevican isoforms, B/b130 associates with the cell membrane by a calcium-independent mechanism. Our studies indicate that B/b130 is generated by a post-translational process that gives rise to an underglycosylated or unglycosylated form of BEHAB/brevican. Importantly, B/b130 is the major BEHAB/brevican isoform upregulated in an experimental model of invasive glioma. Our work suggests that in addition to alternative splicing and cleavage, differentially glycosylated isoforms of BEHAB/brevican may play unique roles in central nervous system development and glioma.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
For further subcellular fractionation from adult rat forebrain samples, the postnuclear supernatant, obtained as indicated above, was centrifuged according to established protocols (25), yielding a mitochondrial/synaptosomal pellet (P2), a light microsomal pellet (P3), and a final soluble fraction. To separate mitochondrial and synaptosomal membranes, the P2 pellet was resolved in a discontinuous sucrose gradient as previously described (26).
To prepare the subcellular fractions for protein electrophoresis, aliquots of membrane and soluble fractions were equilibrated at a final total protein concentrations of 1–2 mg/ml in 40 mM Tris-HCl, 40 mM sodium acetate, pH 8 (CH buffer), containing 10 mM EDTA and treated with 0.25 unit/ml of protease-free chondroitinase ABC from Proteus vulgaris (EC 4.2.2.4 [EC] ; Seikagaku) for 8 h at 37 °C. Chondroitinase reaction was stopped by boiling the samples in the presence of 1x gel loading buffer.
Release of BEHAB/Brevican Isoforms from Brain Membranes—To characterize the association of different BEHAB/brevican isoforms with the cell membrane, total membranes (1–2 mg total protein/ml) obtained from rat forebrain were resuspended in 50 mM Tris-HCl buffer, pH 7.4, in the presence or absence of 10 mM EDTA or 0.2% Triton X-100, for 1 h at 4 °C. Alternatively, the membranes were resuspended in 100 mM sodium carbonate at pH 11.3 for 30 min at 4 °C. After incubation, the membranes were centrifuged at 20,800 x g for 20 min. Released BEHAB/brevican was recovered in the supernatant, and the membranes containing retained BEHAB/brevican were washed twice with 50 mM Tris-HCl buffer and resuspended in the same initial volume. All of the samples were finally equilibrated with CH buffer and treated with chondroitinase ABC prior to protein electrophoresis. For immunoprecipitation studies, the membranes were extracted for 1 h at 4 °C in 50 mM Tris-HCl, pH 7.4, containing 0.6% w/v CHAPS and further processed according to standard protocols.
PI-PLC Treatment and Detergent Extraction with Triton X-114 —To reveal the GPI-anchored isoform of BEHAB/brevican, rat brain microsomal membranes were resuspended in CH buffer with 0.25 unit/ml chondroitinase ABC and 1 unit/ml of phosphatidylinositol-specific phospholipase C (PI-PLC) from Bacillus cereus (EC 3.1.4.10 [EC] ; Sigma) for 8 h at 37 °C. The samples were subsequently separated by SDS-PAGE. To determine the electrophoretic mobility of GPI-anchored BEHAB/brevican before releasing its GPI anchor, we performed a protein extraction on chondroitinased rat brain membranes using the detergent Triton X-114 (Sigma). This detergent allows the separation of water-soluble membrane-associated proteins from hydrophobic or GPI-bound proteins, according to a previously established procedure (27). Briefly, the microsomal membranes were resuspended in CH buffer at 2 mg of total protein/ml and detergent-extracted for 60 min on ice by adding precondensed 10% v/v Triton X-114 to a final concentration of 2% v/v. Solubilized proteins were recovered in the supernatant after centrifugation of the extract at 20,800 x g for 20 min in a refrigerated centrifuge. This supernatant was warmed at 37 °C, producing an aqueous phase and a detergent-containing phase. Both phases were then isolated by centrifugation at 2000 x g for 10 min at room temperature and exhaustively washed to avoid cross-contamination. Extracted proteins in the detergent phase were precipitated with acidic ethanol at –20 °C, washed with cold acetone, and resolubilized in a small volume of CH buffer in the presence of 0.6% w/v CHAPS. Samples from the aqueous and detergent phases were then treated with PI-PLC for 8 h at 37 °C before SDS-PAGE.
Cell Cultures and Transfections—The rat CNS-1 glioma cell line (generously provided by Dr. W. Hickey, Darthmouth-Hitchcock Medical Center, Lebanon, NH) was grown at 5% CO2 in RPMI 1640 medium supplemented with 10% fetal calf serum (Hyclone, Logan UT), 50 µg/ml penicillin, and 50 µg/ml streptomycin (Invitrogen) (28). The mouse oligodendrocyte precursor Oli-neu line (generously provided by Dr. J. Trotter, Department of Neurobiology, University of Heidelberg, Heidelberg, Germany) was grown on poly-D-lysine-precoated plates (Beckton-Dickinson) in Sato's medium (29) supplemented with 25 µg/ml geneticin (Invitrogen).
CNS-1 cells were transfected as previously described (19) with a pCDNA3.1 vector (Invitrogen) containing the full-length rat BEHAB/brevican cDNA (nucleotides 1–2863; generously provided by Dr. Yu Yamaguchi, Burnham Institute) (30). Control cells were transfected with a pCDNA3.1 vector containing a cDNA insert encoding green fluorescent protein (generously provided by Dr. Tom Hughes, Yale University). Stable transfectants were selected in 1 mg/ml geneticin (Invitrogen). Oli-neu cells were transiently transfected with the same rat BEHAB/brevican cDNA. In addition, the full-length rat BEHAB/brevican cDNA was subcloned in a pCDNA3.1/V5-His6 vector to produce V5/His6-tagged BEHAB/brevican, which was also transiently transfected in Oli-neu cells. In all cases, the expression of the desired transgene was confirmed by Northern blot analysis (19), and the presence of BEHAB/brevican protein was confirmed by Western blot analysis, as described below.
Preparation of Cell Membranes and Immunocytochemistry—The Cells were routinely collected 24–48 h post-transfection and homogenized in 25 mM phosphate buffer, pH 7.4, containing a protease inhibitor mixture (Complete, EDTA-free) and 2 units/ml RNase-free DNase I (Roche Applied Science). The total membranes were obtained by centrifugation at 20,800 x g for 30 min and prepared for protein electrophoresis.
For live immunocytochemical staining of transfected Oli-neu cells, the cultures were grown on glass coverslips in 24-well plates for 24–48 h before transfection with the V5/His6-tagged full-length BEHAB/brevican cDNA. Unfixed, unpermeabilized cultures were rinsed in Dulbecco's modified Eagle's medium (Invitrogen) without serum and then incubated with a monoclonal anti-V5 antibody (Invitrogen) at 4 °C for 30 min. The cultures were again rinsed in Dulbecco's modified Eagle's medium, fixed for 20 min in 4% paraformaldehyde, pH 7.4, rinsed, and then incubated for 60 min with Alexa-conjugated goat anti-mouse IgG1 secondary antibody (Molecular Probes). The cultures were finally rinsed in phosphate-buffered saline, briefly counterstained with propidium iodide (0.2 µg/ml), and prepared for fluorescence microscopy.
Intracranial Grafts—Intracranial grafts of stably transfected CNS-1 cells were performed as described previously (18). Briefly, the cells were harvested at 80% confluence, washed in 1x phosphate-buffered saline and resuspended in injection buffer (1x phosphate-buffered saline supplemented with 1 µg/ml MgCl2, 1 µg/ml CaCl2, and 0.1% w/v D(+)-glucose) at a concentration of 5 x 104 cells/µl. Cell suspension (3 µl) was injected stereotaxically into the thalamus of 45-day-old female Lewis rats over a 5-min period. The animals were returned to their cages and monitored for signs of compromised neurological functions during a 2-week period. After 2 weeks, the rats were terminally anesthetized and decapitated, and the brains were quickly dissected, frozen on dry ice, and stored at –70 °C for further processing. The brains were grossly sectioned, and the samples were obtained from the visualized tumors as well as from the equivalent regions of the contralateral side of the brain, where tumors were not detected. Soluble and particulate fractions of normal rat brain and rat brain gliomas were prepared as described above.
Glycosidase Treatments—To remove O-linked oligosaccharides present in BEHAB/brevican isoforms, chondroitinased samples were equilibrated in 10 mM Tris-HCl, 10 mM sodium acetate, 100 mM NaCl, pH 7, and treated with 20 milliunits/ml O-glycosidase from Diplococcus pneumoniae (EC 3.2.1.97 [EC] ; Roche Applied Science) and 100 milliunits/ml sialidase from Arthrobacter ureafaciens (EC 3.2.1.18 [EC] ; Roche Applied Science). Similarly, N-linked oligosaccharides were removed by incubation with 100 units/ml glycopeptidase F from Chryseobacterium meningosepticum (EC 3.5.1.52 [EC] ; Sigma). In all cases, the samples were incubated with the enzymes for 8 h at 37 °C in the presence of protease inhibitors. Enzyme digestions were stopped by boiling the samples in 1x gel loading buffer.
Western Blot Analysis—The samples were electrophoresed on reducing 7% SDS-polyacrylamide gels, and the proteins were electrophoretically transferred to nitrocellulose. The blots were incubated with an affinity-purified rabbit polyclonal antibody (B6) produced against a synthetic peptide corresponding to the GAG attachment region (amino acids 506–529) of rat BEHAB/brevican. Alternatively, BEHAB/brevican was detected by employing affinity-purified rabbit polyclonal antibodies produced against synthetic peptides corresponding to the amino acids 60–73 of rat BEHAB/brevican (antibody B5) and the amino acids 859–879 of human BEHAB/brevican (antibody BCRP). The 50-kDa cleavage product of BEHAB/brevican was detected with an antibody (B50) directed against the neoepitope originated by the proteolytic processing of the full-length protein. The antibodies B6, B5, and B50 have been previously described for the detection or rat BEHAB/brevican in rat brain samples (22). Antibodies B6 and B5 do not detect mouse BEHAB/brevican. Accordingly, to detect BEHAB/brevican produced endogenously by Oli-neu cells, a mouse-derived cell line, we employed a monoclonal anti-brevican antibody (Beckton-Dickinson) that detects mouse BEHAB/brevican. V5/His6-tagged recombinant full-length BEHAB/brevican was detected with a monoclonal antibody anti-V5 epitope (Invitrogen). In all cases, alkaline phosphatase-conjugated secondary antibodies were employed. Immunoreactive bands were visualized with nitro blue tetrazolium and 5-bromo-4-chloro-3-indoyl phosphate.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
150 kDa (B/b150), as reported previously
(22). The B/b150
isoform was detected predominantly in the soluble fraction early in
development but later was found in both soluble and particulate fractions.
|
In parallel with the increase in expression of B/b150 over the course of development, the major cleavage products of BEHAB/brevican also increased (Fig. 1). These cleavage products are generated by cleavage of B/b150 at the Glu395-Ser396 site by aggrecanase-1/ADAMTS-4, a protease of the ADAMTS family, producing a 50-kDa N-terminal fragment and a 90-kDa C-terminal fragment (22). B/b150 and its cleavage products are first detected in the soluble fraction but over the course of development localize increasingly to the particulate fraction of rat brain homogenates.
The immunochemical analysis carried out here also revealed a new,
130-kDa BEHAB/brevican band (B/b130) in the particulate
fraction. B/b130 was most highly expressed between P3 and P14 and
was expressed at lower levels in the adult.
B/b130 Is Distinct from the GPI-linked BEHAB/Brevican Isoform—Given the particulate localization of B/b130, we asked whether this band corresponded to the previously reported, GPI-linked isoform of BEHAB/brevican (21). Initially, we observed that both B/b150 and B/b130 were released by nonionic detergents such as Triton X-100 (Fig. 2A) and zwitterionic detergents such as CHAPS (not shown) from the particulate fraction. Indeed, both isoforms could be released by detergent concentrations as low as 0.2% w/v Tx-100, indicating a particular sensitivity to the disruption of their hydrophobic interactions with the membranes. This observation suggested that B/b130 might not be a GPI-linked protein, because many GPI-linked proteins sit in plasma membrane rafts and tend to have low solubility in Triton X-100 (31). However, previous work reported that the BEHAB/brevican GPI-linked isoform was sensitive to Triton X-100 extraction (21). On the other hand, brief treatment with alkaline sodium carbonate, which releases membrane-associated but not GPI-linked or integral membrane proteins, also released both B/b150 and B/b130 from the particulate fraction (Fig. 2A). Together, these results provide evidence that both B/b150 and B/b130 are likely to be peripherally associated with the cell membrane.
|
To test further whether B/b130 might represent the GPI-linked
isoform of BEHAB/brevican, we treated membranes from adult rat brain with
PI-PLC, an enzyme that can release GPI-linked proteins. PI-PLC treatment
generated a 120-kDa band, which was clearly differentiable from
B/b150 and B/b130
(Fig. 2B) in straight
7% acrylamide gels. This 120-kDa band was entirely released by PI-PLC and was
not detected in the resulting particulate (membrane) fraction
(Fig. 2C). Treatment
of the soluble fraction of the brain homogenate with PI-PLC did not give rise
to a 120-kDa band (not shown), indicating that the GPI-linked isoform that
gives rise to the 120-kDa band specifically localized to the particulate
fraction, as would be predicted for a GPI-linked protein.
BEHAB/brevican has been described as a part time proteoglycan, because the protein is detected with and without GAG addition (9). It is therefore important to note that detection of the 120-kDa band released by PI-PLC required prior treatment with chondroitinase ABC, suggesting that the majority of the GPI-linked isoform is invested with chondroitin sulfate proteoglycans (data not shown).
In addition to our results obtained with alkaline sodium carbonate and Triton X-100, we observed that although PI-PLC treatment led to the evolution of the 120-kDa band, the immunoreactivity of B/b130 was unaffected. These results suggested that the 120-kDa band was not derived from B/b130 and thus that B/b130 was not a GPI-linked isoform of BEHAB/brevican. Therefore, we were left with the question of what molecular mass the GPI-anchored isoform of BEHAB/brevican runs prior to removal of its lipid anchor, because we observed no obvious depletion of any of the immunoreactive bands following PI-PLC treatment. Answering this question required the isolation of the GPI-anchored isoform from all other BEHAB/brevican forms, prior to the removal of its lipid anchor. To that end, we extracted the membrane fraction with the detergent Triton X-114.
Triton X-114 forms a homogeneous solution with water at 4 °C but
partitions into an aqueous phase and a detergent phase at 37 °C. Previous
work has determined that only proteins with significant hydrophobic domains,
including integral membrane and GPI-linked proteins, partition into the
detergent phase, whereas membrane-associated proteins partition into the
aqueous phase (27).
Accordingly, we expected that all isoforms of BEHAB/brevican should partition
to the aqueous phase, leaving only the GPI-anchored isoform in the detergent
phase. Triton X-114 extraction resulted in only a single, 150-kDa band in
the final detergent phase (Fig.
2D). When proteins precipitated from this phase were
treated with PI-PLC, the 150-kDa band disappeared and was replaced by a
120-kDa band, the position previously observed for GPI-BEHAB/brevican
without its anchor. Further, PI-PLC treatment of the proteins in the
water-soluble phase, which included B/b130, did not lead to any
loss of the 150-kDa band and did not lead to the appearance of a 120-kDa
band. These results confirmed that B/b130 is not the glypiated
isoform of BEHAB/brevican. Furthermore, our results indicate that prior to
removal of its lipid anchor, GPI-BEHAB/brevican runs at a molecular mass
similar to that of the predominant secreted isoform, B/b150, thus
making simultaneous detection of the two forms difficult without prior
extraction.
B/b130 Is Not a Cleavage Product of B/b150—The results described above demonstrated that B/b130 is distinct from the GPI-linked isoform of BEHAB/brevican. Another possible source of B/b130 is that it is a cleavage product of the full-length, B/b150 protein. To investigate this possibility, we asked whether N- and C-terminal epitopes of full-length secreted BEHAB/brevican could be found on B/b130 using antibodies directed against epitopes located within 10 kDa of the N-(B5) and C-(BCRP) termini of the full-length protein (Fig. 3A). Because the BCRP antibody recognizes a subset of proteins carrying the CRP motif, including BEHAB/brevican, to detect specifically BEHAB/brevican with the B5 and BCRP antibodies, we immunoprecipitated BEHAB/brevican from a rat brain membrane preparation with the B6 antibody prior to immunodetection. The immunoprecipitated proteins were probed with the immunoprecipitating B6 antibody as well as with the B5 and BCRP antibodies. If B/b130 represents a terminally truncated cleavage product of B/b150, one of these antibodies would not detect it. However, both B/b150 and B/b130 were detected by all three of the anti-BEHAB/brevican antibodies (Fig. 3B), suggesting that B/b130 was not a cleavage product of B/b150. Furthermore, detection of B/b130 by the C-terminal BCRP antibody provided direct evidence that B/b130 was not the GPI-linked isoform, because the CRP domain is absent from the GPI-linked splice variant.
|
These experiments showed that B/b130 was neither the GPI-linked isoform, nor was it a cleavage product of the full-length transcript. To further investigate the molecular source of B/b130, we transfected the rat glioma cell line, CNS-1, with a BEHAB/brevican cDNA encoding the full-length rat BEHAB/brevican protein. Transfected CNS-1 cells produced both B/b150 and B/b130 isoforms, and both were detected with all three anti-BEHAB/brevican antibodies. Moreover, the distribution of B/b150 and B/b130 closely paralleled that observed in rat brain, with B/b150 found predominantly in the culture medium and B/b130 exclusively membrane-associated (Fig. 3C).
We also studied BEHAB/brevican expression in the mouse oligodendrocyte precursor cell line, Oli-neu, a mouse-derived cell line that endogenously expresses both full-length and GPI-anchored BEHAB/brevican mRNAs (13). Because the B6 antibody does not recognize mouse BEHAB/brevican, we used a monoclonal anti-brevican antibody that does detect the mouse protein. Using this antibody no BEHAB/brevican bands were detected in the media from Oli-neu cells, but a reactive band was detected in the membrane fraction of the cells that corresponds to the B/b130 isoform (Fig. 3D). Because the endogenous BEHAB/brevican is expressed at relatively low levels in parental Oli-neu cells, we overexpressed BEHAB/brevican in Oli-neu cells by transiently transfecting them with an expression vector containing the cDNA encoding the full-length rat BEHAB/brevican. In contrast to transfected CNS-1 cells that expressed both B/b150 and B/b130, we detected only the B/b130 isoform in transfected Oli-neu cells and again found this form exclusively localized to the membrane fraction (Fig. 3E). To confirm these results, Oli-neu cells were transfected with a full-length rat BEHAB/brevican cDNA with a C-terminal V5/His6 epitope. Here again, Oli-neu cells only expressed the tagged B/b130 isoform, detected both with BEHAB/brevican antibodies (not shown) and with anti-V5 antibodies (Fig. 3E), providing further evidence that the B/b130 membrane-associated isoform represents a full-length BEHAB/brevican protein.
B/b150 and B/b130 Arise from Differential Glycosylation of a Single Core Protein—Because B/b130 was derived from the same mRNA as B/b150 and was not a result of proteolytic processing, we hypothesized that a different post-translation modification must be responsible for the size difference between these two isoforms. BEHAB/brevican carries N-linked sugars, chondroitin sulfate GAG chains, and additional O-linked sugars. We asked whether differences in glycosylation could account for the size difference between B/b130 and B/b150. Unlike B/b150, the B/b130 isoform appeared to lack chondroitin sulfate chains. Chondroitinase ABC treatment neither shifted the apparent molecular mass of the band nor enhanced (or decreased) its immunoreactivity. Further, a combination of enzymes that remove N- and O-linked sugars shifted the B/b150 band toward the position of B/b130 (Fig. 4). In contrast, the electrophoretic mobility of B/b130 was not affected by treatment with glycosidases, indicating that this isoform lacked most of the N- and O-linked sugars present on B/b150. These results provide evidence that B/b130 is a distinct, unglycosylated or underglycosylated isoform of BEHAB/brevican.
|
B/b130 Associates with Brain Membranes by a Calcium-independent Mechanism—We next defined conditions in which B/b130 and the membrane-associated component of B/b150 could be released from the membrane preparation. Previous work demonstrated that BEHAB/brevican is bound to tenascin-R and a subset of sulfated glycolipids by calcium-dependent mechanisms (11, 12).
The calcium dependence of the association of BEHAB/brevican with the membrane was investigated using the divalent cation chelator, EDTA. Consistent with previous reports, B/b150 was partially released by EDTA, confirming a calcium-dependent association with the membrane. In marked contrast, B/b130 was not released by EDTA, indicating association with the membrane by a calcium-independent mechanism. Further support for calcium-independent binding of B/b130 to the cell membrane was obtained from V5/His6-B/b130 transfected Olineu (Fig. 5B) and CNS-1 cells (not shown). B/b130 association with the cell membrane in these two cell lines was also EDTA-insensitive, mimicking the behavior of rat brain B/b130.
|
B/b130 Is Enriched in the Microsomal Membrane Fraction and Is Present on the Cell Surface—B/b150 was detected in most of the major subcellular fractions of rat brain, whereas B/b130 was enriched in the light microsomal fraction (Fig. 6), consistent with an association with the cell membrane. B/b130 was also detected in the heavy mitochondrial/synaptosomal fraction, but in a further subfractionation in a discontinuous sucrose gradient, B/b130 was restricted to the synaptosomal subfraction, providing further evidence that this isoform likely localizes to the plasma membrane.
|
Because the synaptosomal and microsomal fractions contain ER membranes, the detection of B/b130 in microsomes together with its lack of oligosaccharides raised the possibility that B/b130 might be a precursor form of secreted BEHAB/brevican, somehow retained in the intracellular secretory pathway because of the lack of post-translational glycosylation. We examined the subcellular localization of B/b130 in Oli-neu cells to determine whether B/b130 was transported to the cell surface. As described above, BEHAB/brevican-transfected Oli-neu cells expressed only the B/b130 isoform (Fig. 3, D and E), thereby providing a model to study its cellular localization in the absence of the full-length, glycosylated BEHAB/brevican isoform. To investigate whether BEHAB/brevican was exported to the cell surface in transfected Oli-neu cells, we employed a live staining protocol in which the antibody is not internalized into the cell and therefore detects only proteins on the extracellular surface. Live cell staining of V5/His6-B/b130 transfected Oli-neu cells showed that this isoform was indeed detected on the extracellular surface (Fig. 7). These results demonstrate that, despite its lack of glycosylation, B/b130 is transported through the secretory pathway to the cell surface.
|
B/b130 Is the Major Isoform of BEHAB/Brevican Up-regulated in a Rat Model of Invasive Glioma—We reported previously that BEHAB/brevican mRNA is highly up-regulated in human glioma as well as in rat models of invasive glioma (18). We next asked whether B/b150 and B/b130 were differentially regulated in a rat model of invasive glioma. We examined BEHAB/brevican expression in tumors produced in adult rat brain following stereotaxic placement of CNS-1 cells transfected with BEHAB/brevican cDNA as described previously (18).
We compared the expression of BEHAB/brevican isoforms in experimental rat brain glioma to control rat brain tissue, obtained from the opposite, untreated hemisphere (Fig. 8). The expression of BEHAB/brevican isoforms in the control side was identical to that described above for normal rat brain (Fig. 1). In the soluble fraction of glioma samples B/b150 expression appeared to be reduced compared with control, although these tumors contain much more contaminating blood proteins than normal tissue, making an accurate assessment of soluble proteins difficult because of spurious total protein readings. However, in the particulate fraction of glioma samples B/b130 was dramatically up-regulated compared with control, and B/b150 was down-regulated so that it was almost undetectable compared with control.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
145–150-kDa isoform that can undergo proteolytic
cleavage to generate 50- and 90-kDa products
(22,
23). Alternative mRNA splicing
produces a second isoform that lacks a portion of the C-terminal region and
instead carries a GPI-anchor
(21). Additionally,
BEHAB/brevican has been described as a part time proteoglycan, because it can
either carry or lack chondroitin sulfate GAGs. The new isoform we report here,
BEHAB/brevican 130-kDa (B/b130), differs from the previously
reported isoforms. Unlike B/b150, which maintains peak expression
levels into adulthood, B/b130 expression peaks during development,
coincident with cellular motility and developmental plasticity. Although
previously reported BEHAB/brevican isoforms associate with the cell surface
through calcium-dependent mechanisms, B/b130 associates with the
cell membrane by a calcium-independent mechanism, but it is distinct from the
GPI-linked BEHAB/brevican isoform. Furthermore, although other isoforms of
BEHAB/brevican may or may not carry chondroitin sulfate glycosaminoglycans,
B/b130 appears to lack all detectable carbohydrate
modifications. The developmental onset of B/b130 expression is consistent with our previous reports, in which BEHAB/brevican mRNA expression was first detected during the period of gliogenesis and glial cell motility but after the peak period of neurogenesis (16). By studying protein expression here, we now have been able to differentiate between the expression of the different isoforms and show that although B/b150 expression increases over the course of rat brain development and plateaus at a high level through adulthood, B/b130 expression increases during early postnatal development and then decreases after postnatal day 14 to relatively low levels in the adult. This expression profile suggests that B/b130 may play a role during development, when the extracellular environment of the central nervous system is more permissive to cell movement and the dynamic rearrangement of neural circuitry.
In contrast to previously described BEHAB/brevican isoforms, which associate with the cell surface either through a GPI anchor or via calcium-dependent mechanisms, B/b130 binding to the cell membrane is calcium-independent, in cell lines as well as in preparations from rat brain. Live staining of BEHAB/brevican-transfected Oli-neu cells (Fig. 7) clearly demonstrates that B/b130 is localized to the extracellular surface. All previously described ligands of BEHAB/brevican bind to the C-terminal C-type lectin domain in a calcium-dependent manner (11, 12). Our observations indicate that the association of B/b130 with the cell membrane utilizes a different mechanism, possibly involving as yet unidentified ligands. Further, the distinct binding and structure of B/b130 suggest that the function of B/b130 is distinct from that of the other isoforms.
We have shown here that B/b130 is a membrane-associated protein,
and we present several lines of evidence that B/b130 is distinct
from the GPI-anchored isoform. First, B/b130 can be released from
membranes by treatments that typically release only membrane-associated and
not GPI-anchored proteins, such as alkaline sodium carbonate. Second, when
membranes are partitioned in Triton X-114, B/b130 partitions
exclusively to the aqueous phase, whereas a 150-kDa, PI-PLC-sensitive band
corresponding to the GPI-anchored isoform of BEHAB/brevican partitions to the
Triton X-114 phase. In addition, in the presence of its GPI anchor, the
glypiated splice variant of BEHAB/brevican migrates in SDS-PAGE at a position
indistinguishable from B/b150, thus being masked from independent
detection until it has been separated from the other BEHAB/brevican isoforms.
Third, when CNS-1 cells, which do not produce any endogenous BEHAB/brevican in
culture, are transfected with a cDNA encoding the full-length BEHAB/brevican
isoform, they produce both B/b150 and B/b130, indicating
that B/b130 is a product of the full-length mRNA, and not of the
truncated, GPI-encoding mRNA. Fourth, B/b130 is detected by an
antibody directed against the C-terminal CRP domain, a domain absent from the
GPI-linked isoform. The experiments reported here, resolving the GPI-anchored
isoform from B/b130, indicate that the 130-kDa BEHAB/brevican
isoform previously described in association with rat neural membranes
(32) is actually the
B/b130 glycovariant we have now identified.
A distinguishing feature of B/b130 is that its electrophoretic
mobility is not affected by treatment with chondroitinase ABC or glycosidases
that remove most N- and O-linked oligosaccharides. These results
indicate that B/b130 is an underglycosylated or perhaps an
unglycosylated form of BEHAB/brevican lacking most or all of the sugars
typically present on other isoforms. Treatment with a combination of
glycosidases shifts the molecular mass of B/b150 to 130 kDa,
so that the difference in molecular mass between B/b150 and
B/b130 may well be entirely accounted for by differential
glycosylation. Along similar lines, we recently reported that regulated,
differential glycosylation produces a large number of different products of
the aggrecan gene in the brain
(33), suggesting that
differential glycosylation may be a key mechanism for generating functional
and structural diversity in the mammalian brain.
The regulation of BEHAB/brevican glycosylation appears not to be tied to cell type. In culture, Oli-neu cells endogenously express BEHAB/brevican but produce almost exclusively the B/b130 isoform. The lack of detection of the glycosylated isoforms of BEHAB/brevican is not due to an intrinsic glycosylation defect of Oli-neu cells, because glycosylated products were detected when they were transfected with cDNAs encoding the 50-kDa N-terminal or 90-kDa C-terminal fragments.2 Glycosylation of these peptides appeared to be essentially "normal" because they were modified by N-linked and O-linked sugars in a manner similar to expected patterns from rat brain. This suggests that B/b130 is not a consequence of defective glycosylation but that it arises from a specific post-translational pathway.
CNS-1 cells transfected with BEHAB/brevican cDNA produce both B/b150 and B/b130 isoforms, indicating that one cell type can produce fully glycosylated and underglycosylated variants. Glycosylated polypeptides can be expressed as a mixture of glycoforms, because of intrinsic heterogeneity in the glycosylation process, such as limited activities of the glycosylating enzymes, differential polypeptide folding, and different transit times through the biosynthetic pathway (34), and so homogeneously glycosylated proteins are rare, whereas oligosaccharide microheterogeneity is common. However, the consistent and specific differences between B/b150 and B/b130 imply the existence of dramatically different mechanisms for the post-translational processing and glycosylation of these isoforms.
We previously reported that BEHAB/brevican mRNA levels are dramatically up-regulated in human glioma and in rat models of glioma and that this up-regulation contributes to the invasiveness of brain tumor cells both in vivo and in vitro (35). Here, we show that the up-regulation of BEHAB/brevican mRNA in glioma does not translate into equal up-regulation of all BEHAB/brevican isoforms but instead into the dramatic, differential up-regulation of B/b130. Taken together with the developmental regulation of B/b130, the 130-kDa glycovariant of BEHAB/brevican may be uniquely involved in glial cell motility and glioma cell invasion.
The up-regulation of an underglycosylated variant of BEHAB/brevican in glioma is consistent with other reports of the expression of unusual glycosylation patterns on the cell surface of cancer cells. Reduced or truncated N- and O-glycosylation of membrane proteins has been observed in several cancers (36, 37). Because oligosaccharides may be involved in the adhesive interactions between cells and the ECM, a lack of normal glycosylation could promote the invasive ability of cancer cells by disrupting their normal association with the surrounding ECM (38). This has been proposed for the adhesion molecule CD44 (39), the principal cell surface receptor for HA in several tissues, including the central nervous system (40). CD44 in neuroblastoma cells lacks several N-linked oligosaccharides involved in binding to HA (41), thus favoring an infiltrative phenotype, because they are less adhesive to the ECM. Because BEHAB/brevican also binds HA, the up-regulation of underglycosylated B/b130 might, similarly, promote invasiveness.
Here, we have described a novel BEHAB/brevican isoform, B/b130, an underglycosylated isoform of BEHAB/brevican. B/b130 localizes exclusively to the membrane fraction of rat brain and binds to the cell membrane by a previously undescribed calcium-independent mechanism, which is distinct from the calcium-dependent binding of other BEHAB/brevican isoforms. Perhaps of greatest significance, B/b130 is dramatically up-regulated in a rat model of invasive glioma and may contribute to the invasive ability of glioma cells. Further understanding of the molecular interactions of BEHAB/brevican isoforms in glioma will advance our knowledge of the function of this ECM component and may help identify new potential therapeutic targets for primary brain tumors.
|
FOOTNOTES |
|---|
To whom correspondence should be addressed: Dept. of Neurobiology, Yale
University School of Medicine, 333 Cedar St., SHM C405, New Haven, CT 06510.
Tel.: 203-785-5941; Fax: 203-785-5263; E-mail:
russell.matthews{at}yale.edu.
1 The abbreviations used are: ECM, extracellular matrix; CHAPS,
3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid; CRP, complement
regulatory protein; GAG, glycosaminoglycan; GPI, glycosylphosphatidylinositol;
HA, hyaluronic acid; PI-PLC, phosphatidylinositol-specific phopholipase C;
En, embryonic day n; Pn, postnatal day
n.
2 M. S. Viapiano and R. T. Matthews, unpublished data.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
CiteULike 
