 |
INTRODUCTION |
The focus of our laboratory has been the study of the cell and
molecular biology of one group of basement membrane molecules, basement
membrane proteoglycans. These molecules have been traditionally divided
into two families based on the glycosaminoglycan chains they bear;
i.e. heparan sulfate proteoglycans
(HSPG)1 or chondroitin
sulfate proteoglycans (CSPG). The most well recognized and
characterized member of the HSPG group is perlecan, originally isolated
from the Engelbreth-Holm-Swarm tumor (1, 2). Other proteoglycan core
proteins bearing heparan sulfate (HS) chains have been identified and
characterized, using both biochemical and immunological approaches, as
potentially being gene products unique from perlecan (3, 4). However,
with the exceptions of perlecan (5-7) and, most recently, agrin (8),
none have been characterized at the molecular level.
The cell and molecular biology of CSPG present in basement membranes
has not been as thoroughly investigated as their HSPG counterparts.
Previous work by ourselves (9, 10) and from the laboratories of other
investigators (11-13) demonstrated the possibility of the existence of
several unique gene products for which core proteins bear chondroitin
(CS)/dermatan sulfate (DS) glycosaminoglycan (GAG) chains. Little was
known about the core protein structure of this species until the
molecular characterization of one member of this group, bamacan, by Wu
and Couchman (14). However, although the bamacan core protein is
presently thought to be a "full time" CSPG, the core protein of
perlecan also qualifies as inclusion into the basement membrane CSPG
family, since it has been shown to be secreted in certain tissues and
cell lines in culture as a proteoglycan core protein with CS/DS chains
(15-17).
However, there is some evidence that highlight the fact that basement
membranes might, in fact, contain other CSPG proteoglycans for which
core proteins might differ from those reported above. In earlier
reports, using core protein-specific monoclonal antibodies to
BM-CSPG/bamacan, we showed that BM-CSPG/bamacan was a component of the
mesangial matrix and not present in the walls of capillaries within the
glomerulus (10, 18). However, two studies using monoclonal antibodies
directed against chondroitin carbohydrate-specific epitopes have shown
specific immunostaining the wall of the glomerular capillary for the
presence of CS chains (13, 19). Moreover, a recent report by Schittny
et al. (20) demonstrated the presence of a heterodimeric
proteoglycan bearing DS chains in some but not all basement membranes.
These findings imply the possibility that CS/DS-bearing proteoglycans,
other than BM-CSPG/bamacan, might also exist in basement membranes, or
at the very least, within the glomerular capillary basement membrane.
The present report fulfills in part this premise, describing the
molecular characterization of a novel proteoglycan, leprecan
(leucine proline-enriched proteoglycan), the protein core of which bears CS chains
and which can be immunolocalized to basement membranes.
| |
EXPERIMENTAL PROCEDURES |
Methods
Isolation of Proteoglycans--
L-2 cells, a rat parietal yolk
sac tumor cell line (a generous gift from Dr. Ulla Wewer, Institute of
Pathology and Anatomy, University of Copenhagen, Copenhagen, Denmark),
were grown to confluence in roller bottle culture in Dulbecco's
modified Eagle's medium (Mediatech) supplemented with 4% fetal bovine
serum (HyClone Laboratories, Logan, UT), 4 mM glutamine,
and penicillin/streptomycin. The culture medium was changed weekly and
the cultures maintained in 10% CO2 at 37 °C. The total
proteoglycan pool from the conditioned culture medium was isolated
using a combination of ion exchange chromatography and CsCl density
gradient ultracentrifugation as described previously (10). Briefly, 2 liters of conditioned medium were concentrated under positive pressure
at 4 °C in an ultrafiltration cell (100,000 Mr cut-off, Amicon, Beverly, MA) to a final
volume of 250 ml. Guanidine hydrochloride (Sigma) was added to a final
concentration of 4 M, and solid CsCl (Fisher) was added to
give a final starting density of 1.3 g/ml. 4% CHAPS (Sigma), 10 mM Na2EDTA, 1 mM dithioerythritol
(DTE), and the protease inhibitors 0.2 mM PMSF and 5 mM benzamidine were added to the solution to result in the
final concentrations listed above. The solution was aliquoted into
individual ultracentrifuge tubes and subjected to ultracentrifugation
at 40,000 rpm, for 40 h at 10 °C using a 50.2TI rotor (Beckman
Instruments) in a RC-80 ultracentrifuge (Sorvall). Subsequently, each
tube was divided into five fractions, the density of each fraction
measured, and fractions of approximately the same density pooled.
SDS-PAGE analysis of the pooled fractions indicated that the two most
dense fractions 4 (1.35 g/ml, low buoyant density proteoglycan) and 5 (1.42 g/ml, HD-PG) contained proteoglycans. These were subsequently
dialyzed into 8 M urea, 0.15 M NaCl, 10 mM Na2 EDTA, 1 mM DTE, 50 mM Tris-HCl (pH 8.0) with the above protease inhibitors.
Afterward, each fraction was further purified by ion exchange
chromatography, initially loading each at a rate of 1 ml/min onto a
2.5 × 10-cm column of DEAE Trisacryl (IBF Biotechnics,
Villeneuve-la-Garrenne, France) equilibrated in 8 M urea,
0.15 M NaCl, 0.5% Tween 20, 10 mM
Na2EDTA, 1 mM DTE, 50 mM Tris-HCl
(pH 8.0) with protease inhibitors (as above). After loading, the column
was rinsed with 250 ml of the same buffer, then 250 ml of 8 M urea, 0.15 M NaCl, 10 mM
Na2EDTA, 0.5% Tween 20, 1 mM DTE, 50 mM sodium acetate (pH 4.2) with protease inhibitors. Afterward, the column was eluted with a NaCl gradient (0.15-2.0 M in the same base buffer as the latter column rinse
buffer) monitored by an in-line conductivity meter, and the eluent
collected in 2-ml fractions. The elution of proteoglycans in the
gradient was confirmed by SDS-PAGE. A subsequent Western blot
immunoassay using a polyclonal antibody, R44, which recognizes the
carbohydrate "stubs" remaining on proteoglycan core proteins after
prior digestion with chondroitinase ABC (EC 4.2.2.4, Seikagaku America)
(10), was used to confirm the presence of CSPGs within the respective fractions.
Characterization of Proteoglycans--
In order to identify
HSPGs and CSPGs in the pool, 50-µl samples of the HD-PG pool were
digested overnight either with heparitinase/heparinase (EC 4.2.2.8 and
4.2.2.7, respectively) in 100 mM sodium acetate (pH 7.0), 1 mM calcium acetate, 10 units of leupeptin, 1 mM
benzamidine, or with chondroitinase ABC in 50 mM Tris, 30 mM sodium acetate (pH 8.0), 20 mM EDTA, 10 mM n-ethylmaleimide, 0.2 mM PMSF.
Afterward, both untreated samples and digested samples were
characterized by SDS-PAGE using a 4-15% resolving gel. The separated
proteoglycans were subsequently visualized by silver staining after electrophoresis.
Monoclonal and Polyclonal Antibody Production--
An aliquot of
the HD-PG pool was used as an antigen for the production of monoclonal
antibodies as described previously (10). Tissue culture supernatants
from the resultant hybridomas were screened by Western blot immunoassay
with the HD-PG pool. Those clones for which supernatants were positive
in the screening procedure were further subcloned by limiting dilution,
expanded to high density, and injected into pristane-primed BALB/c mice
for the production of ascites fluid. The determination of mAb subclass was done using a commercially available isotyping kit (Amersham Pharmacia Biotech). The monoclonal antibody, Mo27, was subsequently characterized and used during the course of this study. The identity of
the epitope recognized by Mo27 on proteoglycan core proteins present in
the HD-PG preparation is not certain, but the data (see below) indicate
that it is shared by three of the core proteins present within the
preparation (180, 130, and 100 kDa).
For the production of polyclonal antibodies against the 100-kDa
proteoglycan core protein, an aliquot of the HD-PG pool was fractionated using gel filtration chromatography (CL-4B, 0.5 × 120 cm) in denaturing conditions (50 mM Tris, pH 8.0, 0.35 M NaCl, 0.1% SDS, 0.02% sodium azide, 0.2 mM
PMSF, 5 mM NEM, 10 mM EDTA), the elution of
the target proteoglycan from the column identified in the collected
fractions using the monoclonal antibody, Mo27, in a Western blot assay
(data not shown). A polyclonal antiserum (Rb4127) was raised against
the isolated native proteoglycan in rabbits using biweekly immunization
protocol. 500 µg of antigen suspended in an equal volume of Freund's
complete adjuvant (Sigma) was used in the initial immunization,
followed by a second immunization after 2 weeks using a comparable
amount of antigen suspended in Freund's incomplete adjuvant (Sigma).
The resultant polyclonal antibody was a polyvalent antibody,
recognizing the 180-, 130-, and 100-kDa proteoglycan core proteins
present in the HD-PG preparation.
A polyclonal antibody (Rb2096) directed against a fusion protein (see
below) representing the carboxyl half of the molecule was developed
using the same immunization protocol.
cDNA Library Screening--
A cDNA library developed
from the mRNA of rat parietal endoderm cells (oligo(dT)- and
random-primed) in the Uni-ZAP XR vector was purchased from Stratagene
(La Jolla, CA) and screened as an expression library according to the
developer's instructions using the polyclonal antiserum Rb4127. Clones
found positive after a third round of preliminary screening were
subcloned into appropriate host bacteria. To identify putative clones
for further development, all positive clones from the third round of
screening were initially sequenced (21) using both a 
20 universal
primer and a forward primer (5'-ATTAACCTCACTAAAG-3') with BST
polymerase (Bio-Rad). Extension of sequence information was managed
using a primer walking strategy. The obtained sequence information from
all clones was compared for homology to a data base constructed from
known proteoglycan sequences using the data base search and analysis
routine in MacVector (Oxford Molecular Group, Campbell, CA). Contigs
from DNA sequencing were aligned using AssemblyAlign (Oxford Molecular Group).
To obtain a full-length cDNA, an aliquot of the L-2 cDNA
library was amplified and mass excised according to the manufacturer's (Stratagene) directions. This DNA was subsequently used in a solution hybridization screening strategy (Genetrapper cDNA Positive
Selection SystemTM, Life Technologies, Inc.) according to the
developer's instructions. The capture oligonucleotide
(5'-TCCTGGCACTCGTCATCA-3') used in this procedure was obtained from
Oligos Etc, Inc. (Wilsonville, OR). After completion of the cDNA
capture protocol, the captured single-stranded DNA was converted to
double-stranded DNA using a different oligonucleotide
(5'-TCCTTGATACACACGAGGGATTCG-3') as a repair primer, according to the
developer's instructions. The repaired double-stranded DNA were
transformed by electroporation into ElectroMAX DH10B cells (Life
Technologies, Inc.) using a Gene Pulser apparatus (Bio-Rad) under the
following conditions: 0.8 kV in 0.1-cm gap chamber at settings of 200 ohms and 25 microfarads. The cells were then plated on LB/ampicillin
plates (100 µg/ml) and incubated overnight at 37 °C. The resultant
plates contained four dominant colonies that were selected and grown
overnight at 37 °C in 5 ml of Terrific Broth supplemented with 100 µg/ml ampicillin. Plasmids were purified using a Wizard Miniprep DNA system (Promega), and the identity of the isolated clones was confirmed
by DNA sequencing. One clone, SJ3, was found to encode the full length
of the leprecan molecule.
Northern Blotting
Probe Production--
An 850-base pair fragment of clone P2 was
excised from P2 cDNA by digestion with BamHI. For probe
development, 250 ng of the isolated probe DNA was biotinylated using a
NEBlot phototope kit according to the developer's directions (New
England Biolabs Inc., Beverly, MA). The specificity and sensitivity of
the probe was tested by dot blot assay prior to use on in the Northern
blotting protocol.
Isolation of RNA and Northern Blotting--
Total RNA was
isolated from L2 cells using TRIzolTM LS reagent (Life Technologies,
Inc.). Total RNA (18.75 µg/lane) was electrophoresed at 5 V/cm on a
1.0% agarose, 2.2 M formaldehyde gel. Afterward, the gel
was rinsed five times with diethylpyrocarbonate-treated water and the
RNA was transferred to a nylon membrane (NytranTM, Schleicher & Schuell) following the manufacturer's protocol for neutral transfer
using a downward transfer system (TurboblotterTM, Schleicher & Schuell) overnight in 20× SSC. The next day, membrane was rinsed in
5× SSC for 20 min and the lanes cut into individual strips and
cross-linked with UV irradiation (Gene Linker, Bio-Rad) for total
delivery of 125 mJ of energy. One strip was stained with 0.02%
methylene blue, 0.3 M sodium acetate, pH 5.5 to visualize the 28 and 18 S rRNA bands. The adjacent lane was incubated in prehybridization solution (6× SSC, 5× Denhardt's reagent, 0.5% SDS,
100 µg/ml salmon sperm DNA) at 65 °C for 1.0 h. Afterward, the biotinylated P2 probe was added to the prehybridization solution and allowed to hybridize overnight at 65 °C. Afterward, the membrane was rinsed as follows: 2× SSC, 0.1% SDS at room temperature twice for
5 min each; 0.2× SSC, 0.1% SDS at 25 °C twice for 10 min; 0.2×
SSC, 0.1% SDS at 58 °C twice for 15 min; 2× SSC at 25 °C for 10 min. Detection of the biotinylated probe was done using a NEBlot
photodetection kit (New England Biolabs, Inc., Beverly, MA) following
the manufacturer's protocol with minor modifications; the blocking and
rinsing times were increased to 1.0 h and 30 min, respectively.
The membrane was exposed to BioMax MR film (Eastman Kodak Corp.,
Rochester, NY) and subsequently developed in an automatic film
processor. The experiment was repeated using duplicate strips from the
same blot.
Expression of Bacterial Fusion Proteins
The sequence encoded by clone P2, which represents the carboxyl
half of the core protein, was expressed as a fusion protein by
subcloning a KpnI and SacI restriction fragment
of the sequence into a commercially available fusion protein expression
system (QiaexpressTM, Qiagen) that incorporates a 6×His tag 5' to the polylinker region allowing affinity purification of the fusion protein
product on a nickel-nitrilotriacetic acid resin. Competent M15 (pREP4)
cells (Qiagen) were transformed with the appropriate vector according
to the manufacturer's protocol. Expression of fusion proteins in
liquid cultures was induced by
isopropyl-1-thio-
-D-galactopyranoside according to the
developer's instructions. Afterward, the cells were harvested,
centrifuged, and the cell pellet stored at 70 °C until processing.
For purification of fusion proteins, cell pellets were solubilized in 8 M urea, 0.1 M sodium phosphate, 0.01 M Tris-HCl, pH 8.0. Purification of the fusion protein
(P2E) was performed using a low pressure chromatography system
(Econo-SystemTM; Bio-Rad) to load, rinse, and elute the expressed
protein. The protein was loaded at a flow rate of 0.25 ml/min onto a
0.5 × 0.5-cm nickel-nitrilotriacetic acid resin affinity column
equilibrated in 8 M urea, 0.1 M sodium
phosphate, 0.01 M Tris-HCl, pH 8.0, rinsed with 8 M urea, 0.1 M sodium phosphate, 0.01 M Tris-HCl, pH 5.9, and eluted in 1-ml fractions with 8 M urea, 0.1 M sodium phosphate, 0.01 M Tris-HCl, pH 4.5. The fractions containing the P2E
protein, as measured by absorbance at 280 nm, were
subsequently characterized by 10-20% SDS-PAGE and Western immunoblot
assays using polyclonal antibody Rb4127. The fractions for which
proteins were immunoreactive with Rb4127 on immunoblot assay were
pooled, dialyzed extensively against PBS, concentrated to a 2 mg/ml
concentration, and used to immunize a rabbit to develop a polyclonal
antiserum directed against the fusion protein (Rb2096, see above).
Transfection and Expression in Mammalian Cells
Four T-75 flasks were seeded with equal numbers of CHO-K1 cells
(American Type Culture Collection, Rockville, MD) in Ham's F-12 medium
(Mediatech, Inc., Herndon, VA) supplemented with 10% fetal bovine
serum (HyClone Laboratories) incubated with 10% CO2, at
37 °C 24 h prior to transfection. Each flask represented one of
the following treatment groups: 1) control, no transfection; 2) sham
transfection, i.e. the cells exposed to transfection agent only; 3) inappropriate transfection, i.e. cells transfected
with the control plasmid vector pZeoSVLacZ (Invitrogen, San Diego, CA);
4) cells transfected the pZeoSV plasmid bearing the SJ3 cDNA, representing the full-length coding region for the 100-kDa protein. Protocol for transfections followed that provided by the manufacturer (Life Technologies, Inc.) adjusting all volumes and ratios of the
components in accordance to the relative surface area of the flasks.
Conditioned culture medium from each flask was harvested 24 h and
again at 48 h following the start of the transfection, the medium
placed in 50-ml tubes and clarified by centrifugation at 3000 rpm for
10 min, and the cleared supernatants aspirated to new tubes. For
morphology studies, a similar transfection protocol was used for CHO-K1
cells seeded in six-well plates, adjusting all volumes and component
ratios to the relative surface area of the wells.
Immunoprecipitation of CHO-K1 Cell Culture Supernatants
For immunoprecipitation assay, affinity-purified Rb2096
antiserum was biotinylated using the EZ-link sulfo-NHS biotinylation kit (Pierce) according to the manufacturers' protocol. The
biotinylated antibody was subsequently tested for sensitivity in dot
blot assays and also for specificity via Western immunoblot analysis
using the P2E fusion protein.
Cell culture medium (1.75 ml) representing each transfection condition
(normal, sham transfection with LipofectAMINE, lacZ transfection, and
SJ-3 transfection) was precleared in microcentrifuge tubes with 50 µl of protein A plus resin (Pierce Chemical Co,) with gentle
shaking for 2.0 h at room temperature. Afterward, the
supernatants were clarified by centrifugation at 2000 × g for 2.0 min, and 1.70 ml of each aspirated and dispensed
into fresh tubes. 100 µl of affinity-purified polyclonal antibody
Rb2096 (2.2 mg IgG/ml) was added to each tube and the
immunoprecipitation carried out overnight at 4 °C with gentle
agitation. The next morning 50 µl of protein A resin/buffer slurry
was added to each of the tubes and the supernatants incubated an
additional 3.5 h with gentle agitation. The supernatants were
subsequently clarified by centrifugation at 2000 × g
for 2.0 min, the resin pellets removed from the tubes and rinsed with
TTBS (20 mM Tris-HCl, pH 7.5, 150 mM NaCl,
0.5% Tween 20) for a total of seven 2-min rinses, followed by a final
rinse step with 500 µl of sterile water. The supernatant was
aspirated, and 25 µl of chondroitinase buffer and 2 µl of chondroitinase ABC (Seikagaku America) enzyme were added directly to
the suspension. The samples were incubated overnight at 37 °C.
Subsequently, the resin was rinsed twice with 500 µl of sterile water
and the proteins eluted from the resin by incubating at 100 °C with
two 25-µl aliquots of DTE mixture for 5 min each. After each
incubation, the resin was spun at 2000 × g for 2.0 min
and the sample aspirated and placed in a clean tube. The samples were
then separated by SDS-PAGE using a 4-15% gradient resolving gel
(Bio-Rad) and then electrophoretically transferred onto nitrocellulose (Schleicher & Schuell). Afterward, the blot was rinsed in TBS (20 mM Tris-HCl, pH 7.5, 150 mM NaCl) and then
blocked for 1.0 h at room temp. in SuperblockTM blocking buffer
(Pierce). Afterward, the blot was rinsed three times for 2 min each in
TBS and incubated with biotinylated antibody Rb2096 at 1:1000 dilution
in TBS buffer. The blot was then rinsed twice for 5 min each with TTBS,
three times for 10 minutes each in immunoblot wash buffer (0.0125%
Nonidet P-40, 0.0125% SDS, 0.0312% deoxycholic acid in TBS, pH 7.5),
and twice for 5 min each with TTBS. The blot was then incubated for 30 min at room temperature with neutravidin conjugated to horseradish peroxidase (Pierce) at 1:10,000 dilution in TBS. Following incubation, the blot was then rinsed twice for 5 min in TTBS, twice for 10 min in
immunoblot wash buffer, twice for 5 min in TTBS, and five times for 2 min each in TBS. The blot was developed via chemiluminescence using the
Biomax chemiluminescent detection system (Eastman Kodak Corp.).
Immunodetection of KDEL
Immunoblots of electrophoretically transferred 4-15% SDS-PAGE
gels of secreted proteoglycan (as described above) and cell layer
preparations isolated during several stages of the study were performed
as described previously (10). For the determination of the presence of
the KDEL signal in native leprecan, cell layer extracts were prepared
as follows. Confluent monolayers of L-2 cells were rinsed three times
for 10 min each with PBS followed by extraction with 4 M
urea, 250 mM Tris (pH 8.0), 150 mM sodium acetate, 100 mM Na2EDTA, 100 mM
n-ethylmaleimide, 2 mM PMSF. After extraction,
10-µl aliquots of cell extract were diluted with water to 50 µl,
and some aliquots subjected to overnight treatment with chondroitinase
ABC. Both cell layer aliquots and secreted proteoglycan (± chondroitinase ABC digestion) were electrophoretically separated on a
4-15% gradient gel and subsequently transferred to nitrocellulose for
immunoblotting. The resultant blots were probed with either Rb2096 or
with a monoclonal antibody directed against the sequence SEKDEL
(StressGen, Victoria, British Columbia, Canada). Detection of binding
of primary antibodies was done using species- or isotype-specific secondary antibodies (Jackson Immunoresearch) conjugated to alkaline phosphatase, followed by color development with an alkaline phosphatase color development kit (Bio-Rad).
Cell and Tissue Immunohistochemistry
CHO-K1 cells in six-well plates were fixed for 10 min in
Clark's solution (22), rinsed three times in PBS, and the wells blocked for 20 min with 1% bovine serum albumin in PBS. Afterward the
wells were rinsed three times for 10 min each with PBS, and incubated
with the primary antiserum Rb2096 (1/100 dilution in PBS) for 45 min.
The wells were subsequently rinsed three times for 10 min each with PBS
and then incubated with secondary antibody, goat anti-rabbit IgG
conjugated with fluorescein isothiocyanate (1/100 dilution in PBS,
Jackson Immunoresearch, Malvern, PA) for 45 min. The wells were rinsed
three times for 10 min each with PBS, mounting medium applied, and then coverslipped.
Frozen sections from rat tissues were immunostained with polyclonal
antiserum Rb2096 using single label immunohistochemistry protocols
described previously (10, 18).
The resultant plate wells or slides were imaged using a Zeiss
Axiovert-35 inverted microscope equipped with epifluorescent illumination. Images were digitized using a Photometrics SenSys camera
(Photometrics Ltd., Tuscon, AZ) and ported to a Macintosh PowerPC 9500 hosting the image processing software IPLabSpectrum (Scanalytics Corp.,
Fairfax, VA).
| |
RESULTS |
L-2 Cells Produce Multiple Proteoglycans--
The total pool of
HD-PG was isolated using a protocol combining cesium chloride density
gradient ultracentrifugation and ion exchange chromatography under
denaturing conditions. SDS-PAGE of aliquots of that pool indicated the
possibility of the existence of four or more species of proteoglycan
within the pool (data not shown). Treatment of aliquots of that pool
with chondroitinase ABC demonstrated that at least eight core proteins
were resolvable after separation with SDS-PAGE, with the majority of
the protein cores migrating within the range of 100-200 kDa. Digestion
of parallel aliquots of the HD-PG pool with heparitinase/heparinase revealed no clarification of proteoglycan core proteins within the
SDS-PAGE gel (data not shown), suggesting that the majority of the
proteoglycans within the pool were chondroitin/dermatan sulfate
proteoglycans in nature. Western immunoblotting of another aliquot of
the HD-PG pool digested with chondroitinase ABC with the polyclonal
antiserum R44 provided additional evidence that the core proteins
within the pool bore CS/DS GAG chains (Fig. 1A).
View larger version (77K):
[in this window]
[in a new window]
|
Fig. 1.
Panel A is a photograph of a
Western blot immunoassay of chondroitinase ABC-treated total HD-PG
after SDS-PAGE on a 4-15% gel and transfer to nitrocellulose. In this
case the blot was probed with the R44 antiserum, a polyclonal antibody
recognizing remnant carbohydrate "stubs." Within the region of the
blot spanning 200-100 kDa, there are at least six proteoglycan core
proteins that bear CS/DS side chains. Note that there are also several
smaller core proteins (~35-50 kDa) identified by this antibody in
this preparation. Panel B is a Western blot
immunoassay of total HD-PG electrophoresed on a 4-15% gel and
transferred to nitrocellulose. Lanes 1 and
3 represent intact proteoglycan, lanes 2 and 4 represent the HD-PG sample treated with
chondroitinase ABC and with sequential treatment with chondroitinase
ABC then heparitinase/heparinase, respectively. The blot was probed
with Mo27, which recognizes material that runs as a smear at 200 kDa
(lane 1) that, when treated with the enzyme
chondroitinase ABC resolves into several lower molecular mass bands
(180, 130, and 100 kDa). c'abc, chondroitinase ABC;
+hep, c'abc followed by heparitinase/heparinase;
unt, untreated.
|
|
An aliquot of HD-PG pool was used as an antigen to immunize mice for
the development of a panel of monoclonal antibodies. Screening of the
resultant hybridoma tissue culture supernatants using immunoblot assay
yielded one clone, Mo27, for which immunoreactivity against
proteoglycans was characterized by Western immunoblot assay. In those
assays, Mo27 recognized a "ladder" of core proteins 180, 130, and
100 kDa that were resolved after treatment of the HD-PG pool with
chondroitinase ABC (Fig. 1B, lane 2).
This ladder is sometimes present in GAG-lyase-treated preparations
where the GAG chains on the proteoglycan core protein were not entirely cleaved by the lyase. However, prolonged digestion with chondroitinase ABC failed to resolve the three core proteins into a single band (data
not shown). Treatment of another aliquot of the HD-PG pool with
heparinase/heparitinase failed to change the position of the higher
molecular weight protein cores on Western blots. This indicated that
presence of the higher molecular weight core proteins recognized by
Mo27 on the immunoblot was not due to substitution with HS rather than
CS chains. In order to exclude the possibility that the higher
molecular weight core proteins might be hybrid proteoglycans, a third
aliquot of HD-PG pool was treated in a sequential manner with
chondroitinase ABC followed by heparinase/heparitinase. This treatment
protocol resolved the same pattern of core proteins on Western blots in
a manner similar to that seen with treatment by chondroitinase ABC
alone, indicating that L-2 cells export all three core proteins as a
CS-substituted core protein (Fig. 1B, lane
4). The fact that neither alternative digestion protocol was
unable to resolve the core protein ladder into a single band could be
indicative that Mo27 recognizes a shared epitope present on all three
core proteins in the HD-PG pool.
In order to maximize the potential to detect either the entire
proteoglycan core protein or partially expressed protein core in a
cDNA expression library, an aliquot of the HD-PG pool was fractionated over a CL-4B column. Mo27 was used in a Western blot immunoassay to identify those fractions containing the proteoglycan population of interest. Using the contents of those fractions as an
immunogen, the polyclonal antiserum Rb4127 was developed against the
intact native proteoglycans of the HD-PG pool. The resultant antiserum
was characterized in a manner similar to that described above. The
antiserum recognized the same pattern of core proteins recognized by
Mo27 (Fig. 2A,
lanes 1 and 2) and is considered to be
a polyvalent antiserum, the activity of which is directed against
several CSPG core proteins present in the fraction.
View larger version (36K):
[in this window]
[in a new window]
|
Fig. 2.
Panel A is a Western
immunoblot of total HD-PG electrophoresed on a 4-15% gradient gel as
either the intact, native PG (lanes 1 and
3) or after treatment with chondroitinase ABC
(lanes 2 and 4). The blot was probed
with a polyclonal antiserum, Rb4127, that was directed against the
native proteoglycan species in the HD-PG fractions (lanes 1 and 2) or with Rb2096, the polyclonal antiserum
directed against the P2 fusion protein (lanes 3 and
4). Both antisera recognize the 100-kDa core protein;
however, Rb2096 is not immunoreactive with the higher molecular weight
core proteins present in the HD-PG fractions. Abbreviations are defined
in legend for Fig. 1. Fig. 2B is a Western immunoblot of a
10-20% gradient gel over which the P2E fusion protein had been
electrophoresed. The blot was probed with Rb4127, which readily
recognized the fusion protein. Un, no induction of
expression; Ind, induced expression. Fig. 2C is a
Northern blot of total RNA isolated from L-2 cells that was probed with
cDNA probe derived from clone P-2. Adjacent to the blot is a
parallel strip of membrane that was stained with methylene blue to
visualize the 28 and 18 S rRNA. The resultant band detected by the P2
probe migrated to a position within the gel to yield an estimated size
of 3.1 kb for the mRNA encoding the protein.
|
|
cDNA Library Screening--
From the expression screening
protocol using Rb4127 antiserum as a probe, a total of 10 unique clones
were isolated, each clone sequenced for approximately 0.5 kb from each
end, and the cDNA sequence information obtained from each clone was
compared with a proteoglycan cDNA data base to detect potential
homology to sequence information of known proteoglycans. None of the 10 clones isolated had overlap with the other. However, one of these clones from the Rb4127 screening panel, P2, had a stop codon and polyadenylation signal at the 3' end, indicating that the cDNA encoded for the carboxyl terminus of a novel protein. Moreover, proteoglycan data base comparison showed P2 to have homology over a
limited region with several members of the small leucine-rich repeat
family of proteoglycans (SLRP). Translation of the sequence encoded by
the cDNA revealed the presence of at least one optimal acceptor
site for the addition of GAG chains to the core protein of the molecule
(see below). Extension of sequence information was performed using a
solution hybridization protocol, using a biotinylated oligonucleotide
(5'-TCCTGGCACTCGTCATCA-3') to hybridize to complementary
single-stranded cDNA, and the hybridized cDNAs isolated by
magnetic separation. One clone, SJ3, contained the sequence found in
clone P2 in its entirety, and upon DNA sequencing was found to
represent the entire cDNA of the molecule (see below).
Expression of PG Fusion Protein--
In order to further verify
that clone P2 encoded the sequence of a proteoglycan core protein, the
entire length of P2 cDNA was expressed as a fusion protein (P2E)
subsequently purified by affinity chromatography. An aliquot of the P2E
fusion protein was characterized by SDS-PAGE (10-20% gel). Coomassie
staining indicated that the size of the P2E was approximately 25 kDa.
Subsequent Western blot immunoassays of P2E using Rb4127 as a probe
provided partial confirmation that the isolated cDNA encoded the
carboxyl terminus of a proteoglycan core protein (Fig. 2B).
However, probing an identical blot with Mo27 failed to demonstrate
recognition of the fusion protein (data not shown).
In order to provide further evidence that the fusion protein
represented a portion of the proteoglycan core protein, a polyclonal antiserum, Rb2096, directed against the fusion protein, was developed. The resultant antiserum recognized the 100-kDa proteoglycan core protein resolvable after prior digestion with chondroitinase ABC in the
HD-PG fraction (Fig. 2A, lane 4). This
100-kDa protein was also recognized by Mo27 (Fig. 1B,
lanes 2 and 4) and Rb4127 (Fig.
2A, lane 2). However, Rb2096 did not
recognize the 180- and 130-kDa proteoglycan core proteins in the same
preparation as did Mo27 and Rb4127.
Features of Leprecan--
Northern blot analysis of total RNA
extracted from L-2 cells gave an estimated 3.1-kb size for the message
(Fig. 2C). The isolated SJ3 cDNA contained the entire
open reading frame (ORF) for leprecan (Fig.
3) was 2.541 kb in size; the size of the
ORF for leprecan in this cDNA was 2.183 kb in length. The remainder of the sequence consisted of an 0.325-kb untranslated region and a poly
A tail. During library construction, the EcoRI and
XhoI sites in the polylinker region of the UniZap XRTM
vector were used for the directional insertion of cDNAs. Because of
this, we believe that the 5'-untranslated region is truncated, due to
the presence of an EcoRI site 13 bases 5' to the putative
translational start site. The deduced amino acid sequence for the
leprecan core protein is 727 amino acids in length, having an estimated
Mr of 82,384 and an estimated pI of 4.84. The
translational start site (23) for the protein leads into a presumptive
hydrophobic signal peptide corresponding to the first 14 amino
acids of the deduced sequence. A putative signal peptidase cleavage
site (24) was found starting at Ala12. An RGD cell
attachment motif was found at Arg38, suggesting that this
region serves as a binding site for the integrin family of cell
adhesion receptors. Within the deduced sequence, three sites for
potential GAG attachment were found (Ser182,
Ser488, and Ser645). Two (Ser182
and Ser488) of the three sites were flanked on
one side by acidic clusters, indicating that they might serve as a site
of the addition of either CS or HS chains (25). Also within the
sequence were four N-glycosylation sites at
Asn308, Asn450, Asn459,
and Asn532. Fifteen cysteine residues were found in the
sequence (positions 65, 69, 111, 115, 234, 238, 270, 274, 315, 475, 560, 583, 592, 637, and 662), the majority (n = 8) of
which occurred as pairs of cysteines separated by three amino acids
(CXXXC). Because of the extensive disulfide bonding that
occurs within the molecule, in SDS-PAGE the leprecan core protein
migrates as a protein with an apparent mass of 
80 kDa under
non-reducing conditions (data not shown) and at/or slightly greater
than 100 kDa with reduction. An endoplasmic reticulum retrieval signal,
KDEL, was found in proper context as the terminal four amino acids of
the core protein, suggesting the possibility that this molecule might
also be recycled in the ER-Golgi pathway and may have an intracellular
function. (26). The predicted secondary structure of leprecan shows the molecule to consist of a series of short 
helices interrupted at
irregular intervals by sheets and turns (Fig.
4). Sequence homology searches showed
that the amino-terminal region of leprecan has homology (34%) with the
protein CASP (27), a protein associated with the cartilage
extracellular matrix (Fig. 5), and
related proteins (see below). Comparison with cDNA information in
the GenBankTM EST data base shows that this molecule is potentially highly conserved in both human and murine tissues (Fig.
6).
View larger version (131K):
[in this window]
[in a new window]
|
Fig. 3.
cDNA sequence of the leprecan coding
region with the annotated amino acid sequence directly subjacent.
The heavy underscore denotes the extent of the
cDNA present in clone P2, the clone isolated in the first round of
screening an L-2 expression cDNA library with Rb4127. The
translation start site and polyadenylation signal are denoted with
appropriate markers, and asterisk (*) denotes the stop codon
at the end of the ORF, which is immediately preceded by the ER
targeting signal, KDEL. The three putative glycosaminoglycan acceptor
sites (SG) are boxed, and the four
N-glycosylation sites are underscored.
|
|
View larger version (36K):
[in this window]
[in a new window]
|
Fig. 4.
Panel A is a plot of the
predicted secondary structure of leprecan, using the Chou-Fasman
secondary structure algorithm (top), the Robinson-Garnier
secondary structure algorithm (middle), or a combined plot
(bottom) of regions of consensus between the two algorithms.
The consensus strand predicts the leprecan sequence to contain at least
15 small helical regions interupted at intervals by either sheets or turns. Panel B shows a schematic
diagram of leprecan that is aligned to the upper secondary structure
plots shown in A. The three glycosaminoglycan acceptor
sites, the N-glycosylation sites, and the KDEL site
predicted by primary structure analysis all occur outside of the helical regions of leprecan.
|
|
View larger version (146K):
[in this window]
[in a new window]
|
Fig. 5.
Alignment of the deduced amino acid sequence
of leprecan with those of CASP, SC65 (rat), No55 (human), and Unknown
Protein B (human). The presence of the CXXXC clusters
and their respective enumeration are denoted in bold text
above the aligned sequences. The three SG acceptor sites present in
leprecan are also demarcated in a similar manner. The areas within the
dark gray boxes denote areas of
sequence identity; the lighter gray areas denote
areas of conservative amino acid substitution. Leprecan and Unknown
Protein B (U47926) exactly match two out of three SG acceptor sites
(sites 1 and 3); site 2 in the latter is offset from the leprecan site
by two amino acids in the aligned sequence. Both molecules share the
latter two sets of CXXXC clusters.
|
|
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 6.
Alignment of the cDNA sequence of
leprecan with sequence information present in the GenBankTM EST data
base. The accession numbers and direction of the EST clones is
given to the right of each aligned segment. The
box at right of the figure shows in tabular
format the EST clone number, the species the cDNA clone was derived
from, and the percentage of homology at the nucleic acid level each
clone has to leprecan.
|
|
The Leprecan cDNA (SJ3) Is a Full-length Transcript Encoding a
CSPG--
Further studies were done to substantiate the fact that a
viable translation initiation sequence was present in the ORF of SJ3
cDNA and that the protein encoded by the SJ3 cDNA would be exported into the extracellular milieu as a proteoglycan. For these
studies, CHO-K1 cells were transiently transfected with the entire SJ3
cDNA construct that had been subcloned into the pZeo-SV vector. To
use the pZeo-SV vector for protein expression studies, the inserted
cDNA must have a start codon, stop codon, and polyadenylation
signal present within its sequence. In parallel wells, non-transfected
cells or CHO-K1 cells transfected with either pZeoSVLacZ or pZeo-SV
served as controls for the immunoprecipitation experiments. Twenty-four
hours after transfection, the medium from each cell condition was
harvested, the cell layers fixed, and processed for
immunocytochemistry (Fig. 7) using Rb2096
as a probe. Immunofluorescence microscopy showed that no
leprecan-associated immunostaining was present in native CHO cells
(Fig. 7A), sham-transfected CHO cells (data not shown), or
lacZ-transfected cells (Fig. 7B). Leprecan-associated
immunofluorescence was seen in cells transfected with SJ3 cDNA
(Fig. 7C and background), the pattern of immunolocalization for this molecule resembled ER and Golgi staining.
View larger version (127K):
[in this window]
[in a new window]
|
Fig. 7.
Series of micrographs representing the
results of transfection studies in CHO-K1 cells. The micrographs
in the inset represent native CHO-K1 cells (A),
or CHO K-1 cells transfected with either the lacZ gene (B)
or the SJ3 construct (C), which represents the entire coding
region of leprecan. The cultures were immunostained with the polyclonal
antiserum Rb2096. Both native CHO K1 cells (A) and lacZ-transfected
cells (B) were negative for immunostaining, whereas the CHO
K1 cells transfected with the SJ3 construct were positive. The larger
micrograph is a higher magnification of an SJ3-transfected culture, the
cells within that culture showing varying degrees of immunoreactivity
for the SJ3 protein as well as the limited localization of the protein
to structures resembling ER and Golgi in appearance. In both the
transfected cells and in native L-2 cells immunostained for leprecan,
nuclear/nucleolar immunostaining was not noted.
|
|
The secreted proteoglycan was immunoprecipitated from equal volumes of
medium from the cell transfection groups with Rb2096. Because the
intact leprecan core proteoglycan runs as a smear within the upper
confines of the resolving gel, in order to resolve the CS substituted
core proteins, the bound, immunoprecipitated proteoglycan was digested
with chondroitinase ABC, run on a 4-15% SDS-PAGE under reducing
conditions, and electrophoretically transferred to nitrocellulose. The
blot was subsequently probed with biotinylated affinity-purified
antibodies derived from the same antiserum (Fig. 8) and detected with avidin-horseradish
peroxidase. Leprecan could not be immunoprecipitated from conditioned
medium taken from cultures of either sham-transfected cells
(lane 1) or cells transfected with lacZ
(lane 2). Rb2096 was able to immunoprecipitate
leprecan (lane 3) from the medium of CHO-K1 cells
transfected with SJ3 cDNA, as evidenced by the dense, broad band
present on the gel at 100 kDa. Densitometry measurements of areas of
even exposure within the corresponding lanes showed that CHO cells
transfected with the SJ3 insert gave a signal strength 2.42 and 3.70 times greater than that of lacZ and sham transfectants,
respectively.
View larger version (80K):
[in this window]
[in a new window]
|
Fig. 8.
Lumigraph of a chemiluminescence-based
Western blot immunoassay of a chondroitinase ABC-treated material
electrophoresed on a 4-15% SDS-PAGE gel, isolated by
immunoprecipitation from CH0-K1 cell-conditioned medium using
Rb2096. The medium was derived from cells that were either
sham-transfected with LipofectAMINE (lane 1), transfected
with lacZ (lane 2), or transfected with the
entire SJ3 construct (lane 3). The
boxes directly adjacent to the lane demarcate the
extent of the area within the lane that was used for quantification
(see text). The measurements were taken in each lane in a region of
continuous density immediately to the left of each box.
Rb2096 immunoprecipitated material in the conditioned medium from the
cells expressing the SJ3 construct that was sensitive to chondroitinase
ABC treatment that migrated with an apparent size of approximately 100 kDa. This material was not seen in medium from either the native cells
(data not shown), cells sham-transfected, or cells transfected with
lacZ, nor was it resolved in lumigraphs of similar experiments where
the immunoprecipitated proteoglycan was not treated with chondroitinase
ABC prior to electrophoresis (data not shown).
|
|
Leprecan Is Secreted as a Chondroitin Sulfate Proteoglycan by L-2
Cells--
The information from immunoblot assays of secreted
proteoglycans isolated from L-2 conditioned medium and the pattern of
tissue immunostaining (see below) with the Rb2096 polyclonal antiserum provided evidence that the leprecan core protein should be secreted by
cells as an extracellular CSPG. However, the presence of the KDEL motif
at the COOH terminus of the deduced sequence also implied that this
molecule should be retained in the ER-Golgi circuit (26). To determine
if the intracellular and secreted forms of leprecan did contain the
KDEL recognition signal, an immunoblot assay on duplicate membranes was
performed using the polyclonal antiserum Rb2096 and a monoclonal
antibody recognizing the KDEL sequence (Fig.
9). The anti-KDEL monoclonal antibody
recognized a band approximately 100 kDa that was resolved by
SDS-PAGE after prior treatment of a total L-2 PG pool with
chondroitinase ABC. A band that migrated in an identical manner in
SDS-PAGE was also recognized by Rb2096, the polyclonal antiserum
directed against the P2 fusion protein. These results indicate
that the native leprecan core protein secreted by L-2 cells possesses
the KDEL motif at the COOH terminus.
View larger version (47K):
[in this window]
[in a new window]
|
Fig. 9.
Western blot of a 4-15% SDS-PAGE through
which secreted L-2 PG (lanes 1, 2, 5, and
6) and PG extracted from confluent
L-2 cell layers (lanes 3, 4, 7, and
8) were electrophoresed. One half of the blot was
probed with Rb2096 (lanes 1-4), which was
directed against the carboxyl half of leprecan; the other half of the
blot (lanes 5-8) was probed with a monoclonal
antibody directed against the KDEL sequence, the ER targeting signal
present on the carboxyl terminus of leprecan (see "Experimental
Procedures"). The material present in lanes 1,
3, 5, and 7 was intact proteoglycan
(unt) whereas the material run in lanes 2, 4, 6, and 8 was
proteoglycan treated with chondroitinase ABC (c'abc) prior
to electrophoresis. The intact proteoglycan is barely evident on the
resolving gel, showing mainly as a faint smear within lanes 1, 3, 5, and 7. With prior
treatment with chondroitinase ABC, a band migrating to approximately
100 kDa is detected in the secreted PG (lanes 2 and 6) and in the proteoglycan extracted from the cell layer
(lanes 4 and 8) using both antibodies.
This provides evidence that leprecan, which contains a KDEL targeting
signal, is secreted by L-2 cells as a proteoglycan bearing CS
glycosaminoglycan chains.
|
|
Leprecan Is a Basement Membrane Proteoglycan--
A survey of
various tissues including cardiac muscle, skeletal muscle, central
nervous system (cerebral cortex and cerebellum), intestinal tract,
trachea, ear, skin, liver, and kidney was done (Fig.
10). A consistent finding was that, in
every tissue examined, leprecan antibodies immunostained the basement
membranes of the vasculature and smooth muscle associated with each
organ.
View larger version (158K):
[in this window]
[in a new window]
|
Fig. 10.
These micrographs show the patterns of
leprecan immunostaining of frozen sections of the renal glomerulus
(A), renal tubules and peritubular capillaries
(B), hyaline cartilage from trachea
(C), and skeletal muscle (D). In
the glomerulus (A), the glomerular basement membrane
(arrow), Bowman's capsule, and mesangial matrix are stained
with this antiserum. Antiserum also immunostained the basement
membranes of renal tubules and the peritubular capillaries
(arrow). In cartilage (C), leprecan immunostained
the pericellular material surrounding individual chondrocytes
(arrow) in lacuna but not the cartilage matrix. In skeletal
muscle (D), the basement membranes of the microvessels
(arrow) in between the skeletal muscle are stained but the
skeletal muscle basement membranes are negative for leprecan.
|
|
In the kidney, Rb2096 immunostained the glomerular basement membrane,
mesangial matrix, and Bowman's capsule of the nephron (Fig.
10A). In the renal parenchyma leprecan antibodies stained the basement membranes of tubules and blood vessels (Fig.
10B). Because of the homology of leprecan with CASP (27),
sections of trachea and ear were also immunostained (Fig.
10C). Leprecan immunoreactivity was found around the
perimeter of resident chondrocytes in lacunae (arrows)
similar to that reported for perlecan (28) but not in the cartilage
matrix proper as reported for CASP (27). Leprecan weakly immunostained
the basement membranes around smooth muscle in the stomach and small
intestine but did not immunostain the basement membranes of the
intestinal crypts, villi, or the gastric glands. (data not shown). In
liver, leprecan immunostained the basement membranes of large
vasculature as well as small components of the hepatic sinusoidal tree
(data not shown). Consistent with its distribution in the above
tissues, leprecan was found in the vasculature of the brain, cardiac
muscle, and skeletal muscle (Fig. 10D).
| |
DISCUSSION |
Our results show that we have characterized a novel proteoglycan
core protein, the tissue distribution of which in our initial survey
studies appears to be localized to some, but not all, basement membrane
matrices. Although this particular sequence was identified in our early
screening studies by remote homology to fibromodulin and lumican, two
members of the small leucine-rich repeat proteoglycan (SLRP) family,
the deduced amino acid sequence is far removed enough from the SLRP's
to show this protein to be unique in primary structure from all other
known proteoglycan core proteins. Moreover, alignment of the cDNA
sequence of leprecan with those of other proteoglycans resident to the
basement membrane, including perlecan (6, 7), agrin (8), or bamacan
(14), shows that the leprecan core protein is unrelated to the known
proteoglycans, and is a member of a separate family of molecules that
have yet to be characterized in detail (see below).
The deduced amino acid sequence information suggests that this core
protein has the potential to be heavily glycosylated, having a total of
four putative N-glycosylation sites (Asn308,
Asn450, Asn459, and
Asn532) and a total of three SG sequences, which might
serve to function as acceptor sites for the addition of the GAG chains.
Although the immunoblot data demonstrate that this molecule is secreted as a CSPG, the number and size of the attached GAG chains is presently not known. However, two of the three SG motifs are associated with
flanking acidic residues, which are thought to promote the addition of
GAGs (25). The first site, at Ser182, is associated with a
downstream acidic sequence, E186ED. The second site
(Ser488) has the motif SGDG, which is similar to the
glycosaminoglycan site found in the perlecan core protein (7). This
site is flanked on the amino-terminal side by the sequence
D472DEC, which might also permit this particular site to
serve as an acceptor site for a heparan sulfate GAG chain. The third
site at Ser645 might serve as an acceptor site for GAG
chains but is not immediately flanked on either side by clusters of
acidic residues (25), suggesting that this site might serve for the
exclusive attachment for CS chains. Thus, leprecan could possibly be
secreted by cells as a proteoglycan bearing either CS or HS chains, or
possibly a hybrid molecule. In the case of the native core protein
secreted by L-2 cells, immunoblot assays addressed this possibility,
showing the core protein to be susceptible to chondroitinase ABC and
insensitive to treatment with heparanase/heparitinase (data not shown).
Sequential treatment with both GAG degrading enzymes showed no shift in
the electrophoretic mobility of the core protein, indicating that L-2
cells did not secrete leprecan as a hybrid proteoglycan (Fig. 1B). However, recognizing the fact that the perlecan core
protein can exist as either a CS- or HS-substituted proteoglycan (17), one must acknowledge the potential for tissue-specific patterns of GAG
substitution for leprecan.
Of particular interest is that the carboxyl terminus of this
proteoglycan bears the ER retrieval signal, KDEL, implying that this
molecule might participate in some way in ER processes, possibly interacting with other secreted matrix molecules. This observation is
corroborated in part by the pattern of immunolocalization of leprecan
synthesized by transiently transfected CHO K1 cells, showing that the
construct is readily detectable in the ER and Golgi (Fig. 7).
Immunostaining of L2 cells in culture for leprecan also shows that
cells handle the native protein in a similar manner (data not shown).
Immunoblots of SDS-PAGE gels of whole L2 cell homogenates and secreted
proteoglycan from tissue culture supernatant bulk preparations with a
monoclonal antibody directed against the KDEL sequence demonstrates
that the KDEL retrieval signal is retained in both the intracellular
and in the secreted proteoglycan (Fig. 9). KDEL is known to function as
an ER retrieval signal that potentially limits the localization of
proteins bearing this sequence to the ER and Golgi circuit (26).
However, several recent reports describe other KDEL-tagged molecules as
exiting the secretory/sorting pathway, being exported by the cell as
cell surface molecules (29-31) or as a soluble secreted protein (32). Our data show that leprecan follows that trend, for L-2 cells secrete
the native leprecan protein into the cell medium and CHO K1 cells
transiently transfected with the leprecan sequence secrete leprecan as
a CS substituted proteoglycan. This finding implies a mechanism by
which retention in the ER is bypassed to allow the core protein to
escape the ER-Golgi-ER circuit for further processing. Several
mechanisms have been postulated on how ER retention of KDEL proteins
may be defeated including subunit dimerization (33) or
posttranslational cleavage of the KDEL domain (34). Others have
postulated that the KDEL signal is "leaky" (32), delaying transit
of the protein through the secretory pathway but allowing eventual
export of the KDEL containing protein. In the case of leprecan, two out
of the three mechanisms are possible, but have yet to be investigated
in detail. As mentioned above, our studies show that the KDEL targeting
signal is retained on the secreted leprecan core protein, thus
proteolytic cleavage is unlikely. Dimerization or aggregation into
larger order aggregates is a possibility considering the numerous
cysteine residues (n = 15) present in the deduced
sequence. The latter possibility might also be occurring in the case of
leprecan, where the export of leprecan by the cell is deliberately
delayed due to the KDEL signal. Site-directed mutagenesis studies are
currently in progress to test the significance of the loss of the KDEL
signal in the secretion of leprecan.
The deduced amino acid sequence has little homology between the
leprecan sequence when compared to most of the known proteoglycan core
protein sequences by BLAST algorithms. Our data base search of
GenBankTM showed that the amino-terminal region of the proteoglycan leprecan has homology with a recently described glycoprotein present in
the extracellular matrix of cartilage, called CASP (EMBL accession no.
X97607)/Dualin (Swiss Protein data base accession no. Q90830) (27)
(Fig. 5). Although CASP was shown to be highly regulated during
morphogenesis and development, the function of CASP/Dualin in cartilage
is yet unknown. Interestingly, both leprecan and CASP/Dualin have, in
turn, sequence similarities with a nucleolar protein, No55 (35) (Swiss
Protein accession no. Q92791), a synaptonemal protein SC65 (36) (EMBL
accession no. X65454), and with an unknown protein B (37) (GenBankTM
accession no. U47926). Moreover, besides apparent familial similarities
(see below), scanning the GenBankTM EST data base reveals that in part
the human and murine forms of leprecan are highly conserved (Fig.
6).
The leprecan cDNA sequence information indicates that the encoded
core protein is substantially larger in size (estimated Mr 82,384) than CASP, No55, and SC65, primarily
due to the longer carboxyl "tail" of leprecan (Fig. 5). Leprecan
shows areas of similarity with the sequence encoded by U47926,
especially obvious in the carboxyl tail region of both molecules where
the other members of the family do not overlap. In its amino-terminal region, leprecan also differs from three members of this family in the
location of the putative glycosaminoglycan acceptor site. That acceptor
site, Ser182, is not conserved in CASP, No55, or SC65 but
is present in U47926. However, a different SG motif is present at a
site in No55, and SC65 (Fig. 5) far downstream from the putative GAG
acceptor site found in leprecan. The third GAG acceptor site is
conserved between leprecan and U47926 at Ser645, while the
second leprecan acceptor site Ser488 is offset in U47926 by
two amino acids. What is not known at this time is whether these SG
motifs serve as GAG acceptor sites for the other members of the family.
Comparison of the homologous regions of leprecan with the former
molecules shows that seven cysteine residues present in regions of
sequence similarity are highly conserved among the proteins, with the
CXXXC motif being conserved in three of these proteins. The
fourth CXXXC match is shared among leprecan, SC65, No55, and U47926, but is missing in CASP. Conservation of these cysteines is
indicative that disulfide bonding is important for the secondary structure of these molecules. Although the CXXXC motifs in
the homologous regions among the family members is consistent, what does vary among members in the family is the distance between one
CXXXC motif and the next, with leprecan having the greatest distance between CXXXC 1 and 2. The distance between
CXXXC 2 and 3 for all family members with the exception of
leprecan is identical. U47926 has the greatest distance between
CXXXC 3 and 4, followed by leprecan; SC65 and No55 have
identical distances. Because of these minor differences, it is
anticipated that the secondary structure between members of the family
will differ slightly. However, at this point the significance of this
pattern of organization is presently unknown and the extent of actual
intra- versus interpeptide bonding has yet to be
ascertained; those experiments are currently in progress.
Analysis of the secondary structure of leprecan (Fig. 4A)
shows leprecan to consist of 15 short (18-25 amino acids long) helical regions with sheets and turns interspersed between the helical regions. Given this secondary structure, we believe that leprecan must be globular in configuration rather than a long, extended
configuration. This prediction can be collaborated in part by its
differential mobility in SDS-PAGE, based on whether or not the molecule
is reduced prior to electrophoresis (data not shown). Within the
predicted secondary structure of leprecan, the glycosaminoglycan
binding sites, the N-glycosylation sites, and the KDEL
retrieval signal all lie outside the boundary of the helical
regions where they would be readily accessible. Three of the four
CXXXC motifs (CXXXC 1-3) present in leprecan begin at the end of an helical region and terminate immediately outside of the predicted helix. The last CXXXC domain
spans what is predicted as a turn in the secondary structure. It
may be that the CXXXC motif (1-3) serves to lock the
secondary structure of the respective helices into a specific conformation.
Leprecan's secondary structure is unlike the basement membrane CSPG
bamacan (14), the secondary structure of which is conferred by the
presence of two distinct regions of closely spaced helices. Several
of these helices are between 50 and 100 amino acids in length
before they are terminated. Because of the extensive helices
present in bamacan, its secondary structure has been predicted to
consist of two coiled-coil domains separated by a short flexible hinge
region (14). Perlecan, a basement membrane proteoglycan, has small helices (10-20 amino acids) within its sequence. However, these are in
minority compared with the numerous sheets and turns predicted to
occur in perlecan. The predicted secondary structure of agrin, the
other heparan sulfate proteoglycan present in basement membranes,
resembles that of perlecan in that the sheets and turns far
outnumber the small helices present in the predicted secondary structure.
The immunohistochemistry studies in this report show that leprecan is
associated with the vascular tree, in the media/adventitia of vessels.
Leprecan is present in the glomerulus, resident in both the glomerular
basement membrane and mesangial matrix (Fig. 10). Its presence in
the glomerular capillary wall of the glomerulus is unlike what we had
previously described for BM-CSPG/bamacan (10, 18). In those studies we
reported that BM-CSPG/bamacan was present only in the mesangial
matrix of the glomerulus and not present in the glomerular capillary
wall. However, two studies that used monoclonal antibodies directed
against carbohydrate epitopes (13, 19), present on CSPGs after prior
digestion with chondroitinase ABC, were able to demonstrate the
presence of a CSPG in the glomerular capillary wall. Because of the
data in this present report, it is possible that leprecan, not
BM-CSPG/bamacan, was being detected in the glomerular basement
membrane. In some respects, the pattern of distribution for leprecan
parallels that reported for two members of the original panel of
monoclonal antibodies (mAbs 5A3 and 4D5) directed against the
proteoglycan BM-CSPG (10, 18), that is now named bamacan (14, 17). In
the earlier report we showed that mAb 5A3 immunostained the
microvasculature of skeletal muscle and weakly, if at all,
immunostained the basement membranes of skeletal muscle. The pattern
for leprecan immunostaining is similar in that it stains vascular
basement membranes and not basement membranes surrounding skeletal
muscle. mAb 4D5 and Rb2096 also show parallel patterns of
immunostaining in cartilage. These patterns of immunostaining early in
our studies suggested to us that the cDNA that we had isolated was,
in fact, BM-CSPG. However, comparing both cDNA sequences by BLAST
analysis, the deduced amino acid sequences for both molecules by
Clustal W analysis, and the secondary structures by Chou-Fasman
analysis shows leprecan and bamacan to be completely different.
and
TOP
ABSTRACT
INTRODUCTION
