|
Originally published In Press as doi:10.1074/jbc.M002594200 on August 4, 2000
J. Biol. Chem., Vol. 275, Issue 42, 32628-32634, October 20, 2000
Changes of Matrilin Forms during Endochondral Ossification
MOLECULAR BASIS OF OLIGOMERIC ASSEMBLY*
Yue
Zhang and
Qian
Chen §¶
From the Musculoskeletal Research Laboratory,
Departments of Orthopaedics and Rehabilitation and § Cell
and Molecular Physiology, the Pennsylvania State University College of
Medicine, Hershey, Pennsylvania 17033
Received for publication, March 27, 2000, and in revised form, July 31, 2000
 |
ABSTRACT |
To understand the molecular properties of
matrilin-3, a newly discovered member of the novel extracellular matrix
protein family, we cloned a MAT-3 cDNA from developing
chicken sterna. Real time quantitative reverse-transcription polymerase
chain reaction indicates that MAT-3 mRNA is mainly expressed in the proliferation zone of a growth plate. It is also expressed in the
maturation zone, overlapping with that of the mature
chondrocyte-abundant matrilin-1 mRNA. This suggests that matrilin-3
may self-assemble in the proliferation zone, in addition to its
co-assembly with matrilin-1 during endochondral ossification.
Transfection of a MAT-3 cDNA into COS-7 cells shows that MAT-3
predominantly forms a homotetramer but also a trimer and a dimer.
Co-transfection of both MAT-3 and MAT-1 cDNAs results in three
major matrilins as follows: (MAT-1)3,
(MAT-3)4, and (MAT-1)2(MAT-3)2.
Thus matrilin-3 may assemble into both homotypic and heterotypic
oligomers. Our analysis shows that the assembly of MAT-3 does not
depend on the number of epidermal growth factor repeats within the
molecule, but the presence of Cys412 and Cys414
within the coiled-coil domain, which form covalent disulfide linkage
responsible for both homo-oligomerization of MAT-3 and hetero-oligomerization of MAT-3 and MAT-1. Our data suggest that the
varying synthetic levels of matrilins in different zones of a growth
plate may result in a change of matrilin oligomeric forms during
endochondral ossification.
 |
INTRODUCTION |
In a cartilaginous growth plate, extracellular matrix
(ECM)1 molecules mediate
cell-matrix and matrix-matrix interactions, thereby providing tissue
integrity and a matrix permissible for chondrocyte differentiation and
subsequent ossification. Some members of matrilins, a novel ECM protein
family, have been shown to be expressed specifically in developing
cartilage rudiments but not in adult articular cartilage (1, 2). The
prototype of the matrilin family, cartilage matrix
protein/matrilin-1, has been shown to interact with both collagens (3)
and aggrecans (4). Thus, matrilins may play an important role in the
assembly of the ECM networks (5).
The matrilin family consists at least of four members (2). Whereas
matrilin-2 and -4 are mainly expressed in non-cartilaginous tissues
such as bone and lung, matrilin-1 is expressed specifically in the
pre-hypertrophic mature zone of a growth plate (6). Matrilin-3, a novel
member of the matrilin family, has also been found to express
exclusively in developing cartilage (7). However, it is not known
whether it is expressed by chondrocytes during a specific developmental
stage, similar to matrilin-1. The identification of the co-assembly
product, a heterotetramer (MAT-1)2(MAT-3)2, from growth cartilage (1) suggests that the synthesis of the two
matrilins may overlap. However, it is not clear whether their expression patterns are identical. Our study will determine the expression pattern of MAT-3 mRNA in a growth plate and compare it
with that of MAT-1.
All the members of matrilin family contain von Willebrand factor
A domains, EGF-like domains, and a heptad repeat coiled-coil domain at
the carboxyl terminus, which is responsible for the oligomerization of
the molecule (8, 9). The structural differences among the members of
matrilins are in two aspects. First, while matrilin-1, -2, and -4 contain two A domains (A1 and A2) separated by EGF-like domains,
matrilin-3 lacks the A2 domain. We have shown that the absence of the
A2 domain, which is adjacent to the coiled-coil domain, may modulate
the oligomerization of matrilins, whereas the A1 domain is not involved
in the oligomer formation process (5). For example, the deletion of the
A2 domain from matrilin-1 converts a trimeric form, the major
oligomeric form of matrilin-1, into a mixture of trimers, dimers, and
monomers (5). In addition, the deletion of A2 domain from matrilin-1
abolishes its ability to form collagen-independent filaments (5). Thus,
the A2 domain plays an important role in matrilin assembly. Matrilin-3
lacks the A2 domain (7). Its oligomeric forms from self-assembly remain
unknown. It is also not clear whether there are other hetero-oligomeric forms, besides (MAT-1)2(MAT-3)2, that result
from co-assembly of matrilin-1 and -3. Our study will determine the
homo- and hetero-oligomeric forms of matrilin-3.
Second, while matrilin-1 contains only one EGF-like domain, matrilin-3
contains four EGF-like domains. Matrilin-2 and -4 also contain multiple
EGF repeats (2). It is not known whether the variation of the number of
EGF repeats will affect the assembly of matrilins. We will test the
role of the EGF repeats in matrilin-3 assembly by examining the
oligomeric formation of the matrilin-3 molecules containing 1-4 EGF repeats.
Among all the matrilins, the position and number of cysteines in every
corresponding domain is conserved (2). Each A domain has two cysteines,
one at the NH2 terminus and one at the COOH terminus. Each
EGF-like domain contains six cysteines. There are two additional
cysteines at the NH2 terminus of the coiled-coil domain. It
has been shown that such two cysteines in matrilin-1 (Cys455 and Cys457) are responsible for forming
intermolecular disulfide bonds linking MAT-1 subunits together (9).
Matrilin-3 is composed of one A domain, four EGF repeats, and a
coiled-coil domain. Thus it contains a total 28 cysteines. We
hypothesize that the two cysteines in the coiled-coil domain of
matrilin-3 (Cys412 and Cys414) are responsible
to form intermolecular disulfide bonds to link covalently matrilin
subunits together, similar to the roles of those corresponding
cysteines in matrilin-1. We will test this hypothesis in this study.
To achieve all of these aims, we cloned a MAT-3 cDNA from chick
sternal cartilage. To examine the oligomerization of matrilin-3, MAT-3
peptides that consisted of both the EGF repeats and the coiled-coil
domain were expressed by COS cells, similar to our approach to
characterizing the oligomer formation of matrilin-1 (9). Furthermore,
MAT-3 cDNAs were co-transfected with MAT-1 cDNAs to
characterize the co-assembly of these two classes of matrilins. The
function of the EGF repeats and the cysteine residues within the
coiled-coil domain in the assembly of matrilin-3 oligomers were also determined.
 |
MATERIALS AND METHODS |
Cloning and Construction of Matrilin-3 cDNAs--
Matrilin-3
cDNA was cloned by RT-PCR from total RNA isolated from chick
sternal cartilage using primers 2 and 4 (Fig.
1). Total RNA was isolated from 17-day
embryonic chick sterna using RNeasy kit (Qiagen). RT-PCR of matrilin-3
mRNA was performed using Titan one-tube RT-PCR system (Roche
Molecular Biochemicals) according to the manufacturer's instruction.
In brief, RNA (500 ng), dNTP (0.2 mM/each), dithiothreitol
(5 mM), RNase inhibitor (5 units), primers (0.4 µM/each), reaction buffer (1×), and enzyme mix (1 µl)
were added in one tube, and the volume was adjusted to 50 µl. The
reverse transcriptions were performed at 50 °C for 30 min and then
heated at 94 °C for 2 min. Two-step PCRs were used in the same tube
under the following conditions: 94 °C 30 s, 50 °C 30 s,
and 68 °C 1.5 min for 10 cycles, and then the annealing temperature
was raised to 55 °C for another 20 cycles. The nucleotide sequence
of a 654-base pair MAT-3 cDNA was determined by DNA sequencing and
was found to be identical to that from GenBankTM (accession
number AJ000055), except the difference of seven nucleotides as
follows: T755C, C920T, A961G, C1028T, T1073C, C1124T, and
T1172C. Among these, only A961G causes an amino acid change from
Lys313 to Arg313. A matrilin-3 cDNA
fragment encoding 217 amino acids (Ala235 to
Val452) was linked, by overlapping PCR, with a signal
peptide cDNA sequence from matrilin-1 to ensure its secretion from
cells (Fig. 1). This cDNA (MAT-3/4EGF) was cloned into an
expression vector pcDNA3.1/V5-His (Invitrogen, Carlsbad, CA). Mat-3
EGF deletions and cysteine (Cys412 and Cys414)
point mutations were made by overlapping PCR and cloned to pcDNA3.1 in a similar fashion (Fig. 1). In addition, a wild-type matrilin-1 cDNA, a mini-MAT-1 cDNA, a MAT-1 cysteine (Cys455
and Cys457) mutant cDNA, and a cDNA of the NC1
domain of type X collagen from previous studies (9, 10) were also
cloned into pCDNA3.1/V5-His vectors. The sequence of all the
inserts was confirmed by DNA sequencing.

View larger version (52K):
[in this window]
[in a new window]
|
Fig. 1.
Construct production and primer set. The
relative locations of the primers used to produce various MAT-3
constructs are shown below the schematic model of matrilin-3
in A. Mutated cysteines are indicated by *. SP,
signal peptide; A, von Willebrand factor A domain;
C-C, coiled-coil domain; Tags, V5 and His tags.
B shows the sequence of all primers used in 5' to 3'
orientation. The primers are numbered as in A. C1-28 indicates the positions of 28 cysteines in MAT-3.
Among these, C25-26 corresponds to Cys382 and
Cys384 of MAT-3, and C27-28 correspond to
Cys412 and Cys414 in MAT-3. The codons of these
cysteines were mutated to those of serines by PCR with the
oligonucleotide primers 5-8 described in
B.
|
|
Real Time Quantitative RT-PCR--
The transcripts of the genes
of matrilin-1, -3, and type X collagen in different zones of a growth
plate were quantified with quantitative real time RT-PCR (PerkinElmer
Life Sciences), as described previously (11). Briefly, the tibiotarsal
growth plate from 15-day chick embryo was cut into three zones,
proliferation, maturation, and hypertrophy, under a dissection
microscope, according to the method published previously (6). Total RNA
was isolated from pooled cartilage pieces from the same zone by RNeasy
Kit (Qiagen). Real time quantitative RT-PCR was performed by with a
PerkinElmer Life Sciences ABI PRISM 7700 sequence detection system.
RT-PCR was performed with the AmpliTAQ Gold polymerase (PerkinElmer
Life Sciences ABI) with 20 ng of total RNA for each reaction.
Real time RT-PCR was performed with specific primers and probes
corresponding to different genes. For each mRNA detection, a
fluorogenic probe and two primers for PCR (forward and reverse) were
synthesized (PerkinElmer Life Sciences ABI). The internal oligonucleotide probe was labeled with a fluorescent dye
carboxyfluorescein (FAM) at the 5' end and
N,N,N',N'-tetramdethyl-6-carboxyrhodamine (TAMRA)
at the 3' end. The probe hybridized with the cDNA regions amplified
by PCR. When both dyes were present in an intact probe, TAMRA acted as
a quencher for FAM by absorbing at the FAM emission spectra. When the
internally hybridized probe was degraded by the 5'-exonuclease activity
of Taq polymerase during the course of PCR, these two dyes
were separated in solution, resulting in a subsequent increase in the
level of fluorescence in the reaction mixture. Thus, the amount of
fluorescence released during each amplification cycle was proportional
to the amount of specific PCR products generated in that cycle.
For chick MAT-3 mRNA detection, the forward and reverse primers
were 5'-515CAAGAGCCACAACAAGCAGTCT534-3', and
5'-581GTGACTGAATCCTGCAATGCAG563-3'. The
internal probe was
5'-(FAM)-537TGGGAAGACCTGCAGTGCTTGCAGT561-(TAMRA)-3'.
For chick MAT-1 mRNA, the forward and reverse primers were
5'-515AGGGCTTCACGCTGAACAAT534-3', and
5'-581AGGGCAGATCCTGACCCAC563-3' respectively.
The internal probe was
5'-(FAM)-537TGGGAAGACCTGCAGTGCTTGCAGT561-(TAMRA)-3'.
They were designed according to GenBankTM accession number
M14792.1 (12). The forward and reverse primers for 1 type X collagen
were 5'-2055AGTGCTGTCATTGATCTCATGGA2077-3' and
5'-2137TCAGAGGAATAGAGACCAT TG GATT2113-3',
respectively. The internal probe of ColX cDNA was 5'-(FAM)- 2084TCAAGTGTGGCTCCAGATGCCAAA2107-(TAMRA)-3'.
They were designed according to GenBankTM accession
number M13496.1 (13). The 18 S RNA was amplified at the same time and
used as an internal control. The cycle threshold (Ct) values
for 18 S RNA and that of samples were measured and calculated by
computer software (PerkinElmer Life Sciences ABI). Relative transcript
levels were calculated as x = 2  Ct, in which  Ct = E C, and E = Ctexp Ct18 S;
C = Ctctl Ct18 S.
The mRNA copy numbers of each gene were calculated after
quantitative real time RT-PCR was performed along with the
amplification of cRNA with known copy numbers, as described previously
(14). cRNA was synthesized from the correspondent cDNA cloned in
pCDNA3.1/V5-His vectors with T7 in vitro transcription
kit (Ambion, Inc. Austin, TX).
Transfection of Matrilin cDNAs--
cDNA constructs of
matrilin-3 and -1 were transfected into COS-7 cells either separately
or simultaneously using LipofectAMINE (Life Technologies, Inc.)
according to manufacturer's instruction. Briefly, COS-7 cell were
trypsinized and counted. Each 60-mm plate was seeded with 6 × 105 cells, and were allowed to attach overnight and reach
70% confluence in Dulbecco's modified Eagle's medium supplied with
10% fetal bovine serum (Life Technologies, Inc.). The following day,
the cells were rinsed with Dulbecco's modified Eagle's medium and subjected to a DNA/LipofectAMINE (Life Technologies, Inc.) mix for
5-24 h. Five µg cDNA were used for single transfection, and 4 µg/each cDNA were used for co-transfection, respectively. The DNA/LipofectAMINE mixture was aspirated and replaced with 3 ml of
Dulbecco's modified Eagle's medium supplied with 1% fetal bovine serum. Transfection efficiency was determined by histochemical staining
of a control transfection of a -galactosidase cDNA under identical conditions, using -galactosidase staining kit
(Invitrogen). Seventy two hours after transfection, the media of
transfected cells were collected for Western blotting.
SDS-Polyacrylamide Gel Electrophoresis and Western
Blot--
Western blot analysis was performed from collected
conditioned medium of transfected cultures. For non-reducing condition, collected medium was mixed with standard 2× SDS gel-loading buffer (15). For reducing conditions, the loading buffer contains 5% -mercaptoethanol and 0.05 M dithiothreitol. Samples were
boiled for 10 min before loading onto 10% SDS-polyacrylamide gel
electrophoresis gels or 4-20% gradient gels as indicated. After
electrophoresis, proteins were transferred onto
Immobilon-polyvinylidene difluoride membrane (Millipore Corp., Bedford,
MA) in 25 mM Tris, 192 mM glycine, 15%
methanol. The membranes were blocked in 2% bovine serum albumin
fraction V (Sigma) in phosphate-buffered saline for 30 min and then
probed with antibodies. The primary antibodies used were a monoclonal
antibody against the V5 tag (diluted 1:5000) (Invitrogen), and D2, a
polyclonal antibody, specifically recognizes MAT-1 (16). Horseradish
peroxidase-conjugated goat anti-mouse or goat anti-rabbit IgG (H+L)
(Bio-Rad), diluted 1:3,000, was used as a secondary antibody.
Visualization of immunoreactive proteins was achieved using the ECL
Western blotting detection reagents (Amersham Pharmacia Biotech) and
exposing the membrane to Kodak X-Omat AR film. Molecular weights of the
immunoreactive proteins were determined against two different sets of
protein marker ladders.
 |
RESULTS |
Real Time Quantitative RT-PCR--
To examine the gene expression
pattern of MAT-3, real time quantitative RT-PCR was performed with
mRNA isolated from the growth plate cartilage of three development
stages as follows: proliferating, mature, and hypertrophic. Consistent
with previously published results obtained by in situ
hybridization (6), real time RT-PCR (Table
I) showed that MAT-1 mRNA was mainly
expressed in the maturation zone, and relatively little in the
proliferation or hypertrophic zone, and that type X collagen mRNA
was mainly expressed in the hypertrophic zone. This indicates that real
time RT-PCR can be used to quantify accurately gene expression during
development. Different from the mRNA distribution patterns of type
X collagen and MAT-1, MAT-3 mRNA was predominantly expressed in the
proliferation zone of a growth plate (Table I). It was also expressed
in the maturation zone but very low in the hypertrophic zone. Thus,
matrilin-3 is expressed at a much higher level than MAT-1 in the
proliferation zone, whereas MAT-1 is expressed at a much higher level
than MAT-3 in the maturation zone (Table I). These data suggest that
matrilin-3 may self-assemble in the proliferation zone in addition to
its co-assembly with matrilin-1.
View this table:
[in this window]
[in a new window]
|
Table I
The mRNA copy numbers of Mat-3, Mat-1, and 1(X) in three zones
of tibiotarsal growth plate cartilage from 15-day embryonic chicks
(copies/ng RNA, normalized to 18 S RNA, n = 3)
|
|
Cloning and Construction of Matrilin cDNAs--
To
characterize the assembly of MAT-3, a MAT-3 cDNA was cloned by
RT-PCR from mRNA isolated from embryonic chicken sternal cartilage
(see "Material and Methods"). Since it has been established that
the oligomeric assembly of matrilins is determined by the COOH-terminal
coiled-coil domain (8) and modulated by the domain adjacent the
coiled-coil (5), we constructed a MAT-3 cDNA that contains the
coiled-coil domain and the neighboring four EGF repeats (Fig.
2, MAT-3/4EGF). A V5 tag and a
His6 tag were attached to the COOH terminus of the molecule
for identification of the recombinant protein with antibodies against
these tags. A signal peptide sequence from MAT-1 was attached to the
NH2 terminus of the molecule to ensure the secretion of the
recombinant protein.

View larger version (21K):
[in this window]
[in a new window]
|
Fig. 2.
Constructs of matrilin-3 and matrilin-1
cDNAs. Full-length MAT-3 consists of a signal peptide, A1
domain, four EGF repeats, and a coiled-coil domain. Full-length MAT-1
consists of a signal peptide, two A domains (A1 and A2) separated by an
EGF domain, and a coiled-coil. All the cysteines in MAT-3 and MAT-1 are
numbered. All constructs derived from the full-length MAT-3 and MAT-1
are indicated.
|
|
Two groups of MAT-3 cDNA mutants were made. To determine whether
the multiple EGF domains of MAT-3 were involved in its assembly, three
EGF deletion mutants were created from deleting the EGF domain one at a
time from WT-MAT-3 (Fig. 2, MAT-3/3EGF, 2EGF, and
1EGF). To determine whether the 27th and 28th cysteine
residues (Cys412 and Cys414), located at the
beginning of the heptad repeat coiled-coil, were involved in covalently
linking MAT-3 subunits, these two cysteines were mutated into serines
(Fig. 2, MAT-3/C27-28). As control, the 25th and 26th
cysteines (Cys382 and Cys384) located in the
fourth EGF repeat of MAT-3 were mutated (Fig. 2,
MAT-3/C25-26).
To examine the co-assembly of MAT-3 with MAT-1, three MAT-1 cDNA
constructs were made. They were wild-type MAT-1 (Fig. 2, WT-MAT-1), mini-Mat-1 that contained the coiled-coil domain
and the adjacent A2 domain, and MAT-1/C11-12, in which the
two cysteines that were responsible for covalently linking MAT-1
subunits (9) were mutated. All the constructs carried V5 tags at their
COOH termini. Thus antibodies against the V5 tags would recognize
both recombinant MAT-1, and MAT-3, whereas an antibody against the A2
domain of MAT-1 (D2) would only recognize MAT-1 (16).
Transfection of a Mat-3 cDNA--
To characterize the
molecular forms of self-assembled matrilin-3, MAT-3/4EGF cDNA was
transfected into COS-7 cells. The secreted MAT-3 protein products were
collected from the medium of transfected cells and analyzed on a
Western blot. Under non-reducing conditions, there were three MAT-3
products as follows: a predominant band at 136 kDa (Fig.
3), consistent with the calculated
molecular weight of a MAT-3 tetramer (Table
II); a weak band at 102 kDa (Fig. 3),
consistent with that of a MAT-3 trimer (Table II); and a very weak band
at 68 kDa (Fig. 3), consistent with that of a MAT-3 dimer (Table II).
In addition, higher order multimeric forms of MAT-3 could also be seen
from the Western blot (Fig. 3). Upon reduction, all of these bands were
reduced to 34 kDa (Fig. 3, Reducing), the predicted
molecular mass of a MAT-3 monomer (Table II). This indicated
that matrilin-3 could self-assemble into a tetramer, a trimer, a dimer,
and other multimeric forms. The predominant form of the matrilin-3
oligomer was a homotetramer. Furthermore, the oligomers were linked
covalently by disulfide bonds.

View larger version (22K):
[in this window]
[in a new window]
|
Fig. 3.
Matrilin-3 forms a series of
homo-oligomers. Conditioned medium of COS cells transfected with
MAT-3/4EGF was separated on a 10% gel, blotted to a membrane, and
incubated with antiserum against the V5 tag. Bound antibodies were
detected with a peroxidase-coupled secondary antibody and a
chemiluminescence detection kit. Lanes 1-3 represents three
independent experiments with transfection efficiencies from low
(lane 1), medium (lane 2), to high (lane
3). All yield similar results. Reducing conditions are indicated
and molecular mass is shown on the right.
|
|
Transfection of Mat-3 EGF Deletion Mutants--
To determine
whether EGF repeats were involved in the oligomeric formation of MAT-3,
a series of MAT-3 EGF deletion mutants were transfected. Under
non-reducing conditions, transfection of the MAT-3/1EGF resulted in a
major product of 52 kDa (Fig. 4,
Non-reducing, lane 1), MAT-3/2EGF, a major product of 80 kDa (Fig. 4, Non-reducing, lane 2), MAT-3/3EGF, a major product
of 108 kDa (Fig. 4, Non-reducing, lane 3), and MAT-3/4EGF, a
major product of 136 kDa (Fig. 4, Non-reducing, lane 4).
These were identical to the calculated molecular weights of MAT-3
tetramers (Table II). In addition, minor products of MAT-3 trimers and
dimers were also seen from MAT-3 EGF deletion mutants (Fig. 4,
Non-reducing, lanes 1-3), similar to the MAT-3/4EGF
(lane 4). Upon reduction, all MAT-3 EGF deletion mutants
were reduced to their monomeric forms (Fig. 4, Reducing),
whose molecular weights were consistent with the calculated ones (Table
II). Thus, the number of EGF repeats did not affect the oligomeric
formation of MAT-3, whose major form was a tetramer.

View larger version (21K):
[in this window]
[in a new window]
|
Fig. 4.
EGF domain deletion mutants behave as
MAT-3/4EGF. Conditioned medium of cells expressing MAT-3/1EGF
(lane 1), MAT-3/2EGF (lane 2), MAT-3/3EGF
(lane 3), and MAT-3/4EGF (lane 4) were run under
non-reducing and reducing conditions as indicated, on a 4-10%
gradient gel and analyzed by Western blotting as described in the
legend of Fig. 2. Molecular mass is shown on the left. BSA,
bovine serum albumin.
|
|
Co-transfection of Mat-3 and Mat-1 cDNAs--
To identify
matrilin-3 and matrilin-1 co-assembly products, MAT-3/4EGF was
co-transfected with a matrilin-1 cDNA (WT-MAT-1). Co-transfection
resulted in three products (Fig.
5A, Non-reducing, V5, lane 1),
which were from co-assembly of MAT-1 and MAT-3 monomers (Fig. 5A,
Reducing, V5, lane 1). Among the three co-transfection products,
the first product of 200 kDa was also a product from MAT-1 single
transfection (Fig. 5A, Non-reducing, lane 3, V5 and D2). This product was established previously to be a MAT-1
trimer (16). The third product of 136 kDa was also a product from MAT-3 single transfection (Fig. 5, Non-reducing, V5, lane 2),
which was established previously in this study as a MAT-3 tetramer
(Figs. 3 and 4). One co-assembly product of 178 kDa was seen (Fig.
5A, Non-reducing, V5, lane 1), which was absent from single
transfections of either MAT-1 or MAT-3 alone (Fig. 5A,
Non-reducing, V5, lanes 2 and 3). The apparent
molecular weight of this co-assembly product was identical to the
predicted molecular weight of (MAT-1)2(MAT-3)2 (Table II). Consistent with these data, the D2 antibody, which only
recognized MAT-1 (16), detected two products containing MAT-1 during
co-assembly as follows: an upper band MAT-1 homotrimer and a lower band
(MAT-1)2(MAT-3)2 (Fig. 5A, Non-reducing,
D2, lane 1).

View larger version (37K):
[in this window]
[in a new window]
|
Fig. 5.
Co-assembly of MAT-3/4EGF with WT-MAT-1
(A), and with mini-MAT-1 (B).
A, conditioned medium of cells co-transfected with both
MAT-3/4EGF and WT-MAT-1 (lane 1), transfected with
MAT-3/4EGF only (lane 2), or transfected with WT-MAT-1 only
(lane 3) were separated under non-reducing and reducing
conditions as indicated, on a 4-10% gradient gel and analyzed by
Western blotting. Both analyses with antiserum against the tag
(V5) and with antiserum against MAT-1 (D2) are
shown. B, conditioned medium of cells co-transfected with
both WT-MAT-3 and mini-MAT-1 were separated under non-reducing and
reducing conditions as indicated and analyzed by Western blotting as
described (A). Note the 136-kDa (MAT-3)4 band was shielded
by the 128-kDa (mini-MAT-1)2(MAT-3)2. This may also artificially
increased the intensity of the (mini-MAT-1)2(MAT-3)2 band.
BSA, bovine serum albumin.
|
|
To verify further the co-assembly products, the MAT-3/4EGF was
co-transfected with a Mini-MAT-1, which contained only the A2 domain
and the coiled-coil of MAT-1. A previous study has shown that the
mini-MAT-1 behaves exactly like wild-type MAT-1 during oligomer
formation (9). Co-transfection resulted in two products that contained
MAT-1 (Fig. 6B, Non-reducing,
V5 and D2) as follows: a mini-MAT-1 homotrimer (90 kDa)
and a mini-MAT-1 and MAT-3 hetero-oligomer (128kDa), whose molecular
weight was identical to that of
(mini-MAT-1)2(MAT-3)2 (Table
III). Thus, MAT-3 assembled with MAT-1 in
the molecular form of (MAT-1)2(MAT-3)2.

View larger version (52K):
[in this window]
[in a new window]
|
Fig. 6.
Cys412 and Cys414 of
MAT-3 are necessary for interchain disulfide bond formation linking
matrilin-3 subunits (A) and linking matrilin-3 with
matrilin-1 through Cys455 and Cys457 of MAT-1
(B). A, conditioned medium of cells
transfected with MAT-3/C25,26 (lane 1) and with MAT-3/C27,28
(lane 2) were separated under non-reducing and reducing
conditions as indicated on an 8% gel and analyzed with Western blot
with antiserum against the tag V5. Cross-reactions of bovine serum
albumin (BSA) in reduced samples are indicated.
B, conditioned medium of cells transfected with WT-MAT-1
only (lane 1), MAT-3/4EGF only (lane 2),
MAT-3/4EGF and MAT-1/C11,12 (lane 3), WT-MAT-1 and
MAT-3/C27,28 (lane 4), and MAT-1/C11,12 and MAT-3/C27,28
(lane 5) were separated under non-reducing and reducing
conditions as indicated on an 8% gel and analyzed with Western blot
with antiserum against the tag V5. Cross-reaction of bovine serum
albumin (BSA) in reduced samples is indicated.
|
|
View this table:
[in this window]
[in a new window]
|
Table III
Amino acid residues at the a and d positions of the heptad repeat of
the coiled-coil domain of chick matrilin-1 and -3
According to Harbury et al. (18), the heptad repeat with
-branched residues ( ) at the a position and Leu (L) at the d
position ( ,L) favors dimer formation, ( , ) and (L, L) favor
trimer formation, and (L, ) favors tetramer formation. The number of
heptad repeats is indicated.
|
|
Site-directed Mutagenesis of Cys412 and
Cys414 of Mat-3--
To determine whether
Cys412 and Cys414 (the 27th and 28th cysteines)
were the sites that covalently link the matrilin-3 subunits, we
performed transfection of the MAT-3 mutant (Mat-3/C27,28), in which the
two cysteines were mutated. Mutation of these two residues completely
abolished the covalent bonds that linked the subunits (Fig. 6A,
Non-reducing, lane 2). In contrast, mutation of the two cysteine
residues in the fourth EGF repeat of MAT-3 (Mat-3/C25,26) did not
affect the tetramer formation of MAT-3 (Fig. 6A, Non-reducing,
lane 1). Thus, Cys412 and Cys414 were
responsible for covalently linking MAT-3 subunits.
Co-transfection of Cys Mutants of Mat-1 and Mat-3--
We
hypothesized that Cys412 and Cys414 (the 27th
and 28th cysteines) of matrilin-3 also formed disulfide bonds with
Cys455 and Cys457 (the 11th and 12th cysteines)
of matrilin-1, which have been shown before to link covalently MAT-1
subunits (9). Thus disulfide bonds among these four residues form
covalent linkages between MAT-1 and MAT-3. To test this hypothesis, a
series of co-transfections were performed between cysteine mutants of
MAT-1 and MAT-3. First, when MAT-1/C11,12 was co-transfected with
MAT-3/4EGF (Fig. 6B, Non-reducing, lane 3), no
hetero-oligomers between MAT-1 and MAT-3 were formed, although
homo-oligomers of MAT-3 (tetramer, trimer, and to a less extent, dimer)
still formed (compare Non-reducing lane 3 to
lane 2). Thus, Cys455 and Cys457 of
MAT-1 were responsible for linking MAT-1 to MAT-3. Second, when
MAT-3/C27, 28 was co-transfected with WT-MAT-1, it failed to form any
hetero-oligomers (Fig. 6B, Non-reducing, lane 4), although
MAT-1 homo-oligomers (trimer and to a less extent dimer) were formed
(compare Non-reducing lane 4 to lane
1). Thus, Cys412 and Cys414 of MAT-3 were
responsible for covalently linking MAT-3 subunits to MAT-1. When
MAT-1/C11,12 was co-transfected with MAT-3/C27,28 (Fig. 6B,
Non-reducing, lane 5), neither the homo-oligomers of MAT-1 and
MAT-3 nor the hetero-oligomers between MAT-1 and MAT-3 (see Fig.
5A) were covalently linked. Thus, disulfide bonds among the
two cysteines (Cys455 and Cys457) of MAT-1 and
the two cysteines (Cys412 and Cys414) of MAT-3
were responsible for covalently linking the subunits of
hetero-oligomers of MAT-1 and MAT-3 together.
 |
DISCUSSION |
ECM in growth plate cartilage is sequentially modified by
differentiating chondrocytes and matured during endochondral
ossification. The spontaneous assembly of the ECM oligomeric molecules
during this maturation process depends on at least two factors as
follows: (a) gene expression patterns of chondrocytes, and
(b) oligomeric properties of the molecule. In this study,
the expression and oligomerization of matrilin-3 are characterized. We
have shown that, although both matrilin-1 and -3 are expressed
specifically in growth plate cartilage, the expression pattern of MAT-3
is different from that of MAT-1. Mat-3 mRNA is expressed
predominantly in the proliferation zone, where much less MAT-1 mRNA
is expressed. In the maturation zone, much more MAT-1 mRNA is
expressed than MAT-3 mRNA. Because the chondrocytes from the
proliferation zone precede those from the maturation zone temporally
during development, our data suggest that matrilin-3 may self-assemble
in the proliferation zone, in addition to its co-assembly with
matrilin-1.
The self-assembly of matrilin-3 was studied by transfection of a MAT-3
cDNA in COS-7 cells that do not synthesize endogenous MAT-3.
Matrilin-3 forms a tetramer predominantly, but also a trimer and a
dimer in small quantities. Thus, unlike matrilin-1 and -2, which form
mainly trimers (9, 17), matrilin-3 forms tetramers predominantly.
Analysis of the amino acid sequence of the coiled-coil domain of MAT-3
(Table III) suggests that, unlike the MAT-1 sequence that favors a
triple helix, the MAT-3 sequence strongly favors the formation of a
tetrameric helix. For example, the presence of -branched residues at
both "a" and "d" positions favors a triple helix, and the
presence of residues at a positions also disfavors a tetrameric
helix (18). Mat-1 contains residues at both positions, and MAT-3
does not contain any residues at a positions (Table III). Thus,
MAT-1 favors trimer formation and MAT-3 favors tetramer formation. By
using this rule to analyze all the members of the matrilin family, it
can be predicted that the major oligomeric forms of matrilin-1, -2, and
-4 are trimers and that of matrilin-3 is a tetramer. Although it has
been shown that matrilin-1 and -2 form trimers (9, 17), and matrilin-3 forms tetramer, the oligomeric form of matrilin-4 remains to be determined.
The second interesting aspect of our finding is that the formation of
MAT-3 oligomers is not exclusively tetrameric. Matrilin-3 also forms
trimers, dimers, and even other multimeric forms in small quantities.
Examination of the MAT-3 sequence (Table III, Matrilin-3) reveals that
there is indeed one pair of Leu residues at the a and d positions,
favoring trimer formation (18), although the majority of the a residues
are Leu, and the d residues are -branched, thus favoring tetramer
formation (18). Therefore, this sequence may dictate that the majority
of the MAT-3 oligomers are tetramers and the minority are trimers.
Similar mixed sequence information, which contains both a majority
component and a minority component, exists for all the matrilins. For
example, although the majority of the MAT-1 sequence ( -branched
residues at both a and d positions) favors trimer formation, the
minority of the sequence ( residues at the a position and Leu at the
d position) may make the formation of a dimer possible as well (Table
III, Matrilin-1). Indeed, a minority of matrilin-1 exists as dimers and
other multimeric forms (1). Consistent with this hypothesis, it has
been reported that, although the majority of matrilin-2 exists as
trimers (19), tetramers and dimers exist as well (20). Thus, each
matrilin may exist in a major oligomeric form and minor oligomeric
forms as well. The heterogeneity of the matrilin oligomeric forms may
contribute to the complexity of the extracellular matrix network.
The co-assembly of matrilin-3 and -1 was studied by co-transfection of
their cDNAs. The major co-assembly product is
(MAT-1)2(MAT-3)2, identical to the one
identified from growth plate cartilage in vivo (1). From our
data, it can be concluded that (MAT-1)2(MAT-3)2 is the major, if not the only, co-assembly product between matrilin-1 and -3. It is interesting to note that, although MAT-1 favors trimer
formation and MAT-3 favors tetramer formation, the hetero-oligomeric form of these two adopt a tetrameric configuration. Thus, the tetrameric configuration of MAT-3 acts in a dominant fashion during the
co-assembly process.
Members of the matrilin family contain different numbers of the
EGF-like domains. Mat-1 contains 1; MAT-2 contains 10, and both MAT-3
and -4 contain 4. Our analysis has shown that the variation of the EGF
repeats has little effect on the oligomeric properties of MAT-3, even
though the EGF repeats abut the coiled-coil domain in the molecule.
This suggests that the EGF-like domains in a matrilin do not influence
the oligomeric formation process. In contrast, the pair of cysteines at
the NH2 terminus of the coiled-coil domain is conserved
among all the matrilins. Our analysis has shown that this pair of
cysteines in MAT-1 and MAT-3 is necessary and sufficient for the
formation of intermolecular disulfide bonds connecting matrilin
subunits. This includes the formation of both homo- and
hetero-oligomeric matrilins.
In summary, we have determined the gene expression pattern of
matrilin-3 in a growth plate, and we identified several novel forms of
matrilin-3. These data may have implications for understanding the
matrix maturation process during development. In the proliferation zone
of a growth plate, chondrocytes are dividing and very close to each
other, thus there is very little matrix space between neighboring
cells. In contrast, mature chondrocytes synthesize and deposit a large
amount of matrix, thereby creating a large interstitial matrix region.
It has been shown that the mature cartilage-specific matrilin-1 could
form a filamentous network to connect collagen fibrils and aggrecans
(5). Matrilin-3, which lacks the A2 domain, may not form filaments by
itself (5). Our data suggest that there is a change of the molecular
forms of matrilins in a growth plate, from the proliferation zone to the maturation zone. In the proliferation zone, matrilin-3 is the major
matrilin form, whereas matrilin-1 is the major form of matrilins
in the maturation zone. This synthetic change may convert matrilin
forms from the ones that are incapable of forming long range matrix
networks into those that can, thereby contributing to the matrix
maturation process during endochondral ossification.
 |
ACKNOWLEDGEMENT |
We thank Deb Grove for assistance in
performing real time RT-PCR.
 |
FOOTNOTES |
*
This work was supported by National Institutes of Health
Grants AG14399 and AG00811 (to Q. C.) and by the Arthritis Foundation (to Y. Z. and Q. C).The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement" in
accordance with 18 U.S.C. Section
1734 solely to indicate this fact.
¶
To whom correspondence should be addressed. Tel.:
717-531-4835; Fax: 717-531-7583; E-mail:
qchen@ortho.hmc.psu.edu.
Published, JBC Papers in Press, August 4, 2000, DOI 10.1074/jbc.M002594200
 |
ABBREVIATIONS |
The abbreviations used are:
ECM, extracellular
matrix;
MAT, matrilin;
RT-PCR, reverse-transcription polymerase chain
reaction;
EGF, epidermal growth factor;
WT, wild type;
TAMRA, N,N,N',N'-tetramdethyl-6-carboxyrhodamine;
FAM, 6-carboxyfluorescein.
 |
REFERENCES |
| 1.
|
Wu, J. J.,
and Eyre, D. R.
(1998)
J. Biol. Chem.
273,
17433-17438
|
| 2.
|
Deak, F.,
Wagener, R.,
Kiss, I.,
and Paulsson, M.
(1999)
Matrix Biol.
18,
55-64
|
| 3.
|
Winterbottom, N.,
Tondravi, M. M.,
Harrington, T. L.,
Klier, F. G.,
Vertel, B. M.,
and Goetinck, P. F.
(1992)
Dev. Dyn.
193,
266-276
|
| 4.
|
Paulsson, M.,
and Heinegard, D.
(1979)
Biochem. J.
183,
539-545
|
| 5.
|
Chen, Q.,
Zhang, Y.,
Johnson, D. M.,
and Goetinck, P. F.
(1999)
Mol. Biol. Cell.
10,
2149-2162
|
| 6.
|
Chen, Q.,
Johnson, D. M.,
Haudenschild, D. R.,
and Goetinck, P. F.
(1995)
Dev. Biol.
172,
293-306
|
| 7.
|
Wagener, R.,
Kobbe, B.,
and Paulsson, M.
(1997)
FEBS Lett.
413,
129-134
|
| 8.
|
Beck, K.,
Gambee, J. E.,
Bohan, C. A.,
and Baechinger, H. P.
(1996)
J. Mol. Biol.
256,
909-923
|
| 9.
|
Haudenschild, D. R.,
Tondravi, M. M.,
Hofer, U.,
Chen, Q.,
and Goetinck, P. F.
(1995)
J. Biol. Chem.
270,
23150-23154
|
| 10.
|
Zhang, Y.,
and Chen, Q. A.
(1999)
J. Biol. Chem.
274,
22409-22413
|
| 11.
|
Wu, Q.,
and Chen, Q.
(2000)
Exp. Cell Res.
256,
383-391
|
| 12.
|
Kiss, I.,
Deak, F.,
Holloway, R. G., Jr.,
Delius, H.,
Mebust, K. A.,
Frimberger, E.,
Argraves, W. S.,
Tsonis, P. A.,
Winterbottom, N.,
and Goetinck, P. F.
(1989)
J. Biol. Chem.
264,
8126-8134
|
| 13.
|
Ninomiya, Y.,
Gordon, M.,
van der Rest, M.,
Schmid, T.,
Linsenmayer, T.,
and Olsen, B. R.
(1986)
J. Biol. Chem.
261,
5041-5050
|
| 14.
|
Suryanarayana, K.,
Wiltrout, T. A.,
Vasquez, G. M.,
Hirsch, V. M.,
and Lifson, J. D.
(1998)
AIDS Res. Hum. Retroviruses
14,
183-189
|
| 15.
|
Sambrook, J.,
Fritsch, E. F.,
and Maniatis, T.
(1989)
Molecular Cloning: A Laboratory Manual
, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY
|
| 16.
|
Chen, Q.,
Johnson, D. M.,
Haudenschild, D. R.,
Tondravi, M. M.,
and Goetinck, P. F.
(1995)
Mol. Biol. Cell
6,
1743-1753
|
| 17.
|
Deak, F.,
Piecha, D.,
Bachrati, C.,
Paulsson, M.,
and Kiss, I.
(1997)
J. Biol. Chem.
272,
9268-9274
|
| 18.
|
Harbury, P. B.,
Zhang, T.,
Kim, P. S.,
and Alber, T.
(1993)
Science
262,
1401-1407
|
| 19.
|
Pan, O. H.,
and Beck, K.
(1998)
J. Biol. Chem.
273,
14205-14209
|
| 20.
|
Piecha, D.,
Muratoglu, S.,
Morgelin, M.,
Hauser, N.,
Studer, D.,
Kiss, I.,
Paulsson, M.,
and Deak, F.
(1999)
J. Biol. Chem.
274,
13353-13361
|
Copyright © 2000 by The American Society for Biochemistry and Molecular Biology, Inc.

CiteULike Complore Connotea Del.icio.us Digg Reddit Technorati What's this?
This article has been cited by other articles:

|
 |

|
 |
 
L. van der Weyden, L. Wei, J. Luo, X. Yang, D. E. Birk, D. J. Adams, A. Bradley, and Q. Chen
Functional Knockout of the Matrilin-3 Gene Causes Premature Chondrocyte Maturation to Hypertrophy and Increases Bone Mineral Density and Osteoarthritis
Am. J. Pathol.,
August 1, 2006;
169(2):
515 - 527.
[Abstract]
[Full Text]
[PDF]
|
 |
|

|
 |

|
 |
 
C Otten, R Wagener, M Paulsson, and F Zaucke
Matrilin-3 mutations that cause chondrodysplasias interfere with protein trafficking while a mutation associated with hand osteoarthritis does not
J. Med. Genet.,
October 1, 2005;
42(10):
774 - 779.
[Abstract]
[Full Text]
[PDF]
|
 |
|

|
 |

|
 |
 
Y. S. Kanwar, J. Wada, S. Lin, F. R. Danesh, S. S. Chugh, Q. Yang, T. Banerjee, and J. W. Lomasney
Update of extracellular matrix, its receptors, and cell adhesion molecules in mammalian nephrogenesis
Am J Physiol Renal Physiol,
February 1, 2004;
286(2):
F202 - F215.
[Abstract]
[Full Text]
[PDF]
|
 |
|

|
 |

|
 |
 
G C Jackson, F S Barker, E Jakkula, M Czarny-Ratajczak, O Makitie, W G Cole, M J Wright, S F Smithson, M Suri, P Rogala, et al.
Missense mutations in the {beta} strands of the single A-domain of matrilin-3 result in multiple epiphyseal dysplasia
J. Med. Genet.,
January 1, 2004;
41(1):
52 - 59.
[Full Text]
[PDF]
|
 |
|

|
 |

|
 |
 
K. Feng, X.-H. Zhou, T. Oohashi, M. Morgelin, A. Lustig, S. Hirakawa, Y. Ninomiya, J. Engel, U. Rauch, and R. Fassler
All Four Members of the Ten-m/Odz Family of Transmembrane Proteins Form Dimers
J. Biol. Chem.,
July 12, 2002;
277(29):
26128 - 26135.
[Abstract]
[Full Text]
[PDF]
|
 |
|

|
 |

|
 |
 
S. Frank, T. Schulthess, R. Landwehr, A. Lustig, T. Mini, P. Jeno, J. Engel, and R. A. Kammerer
Characterization of the Matrilin Coiled-coil Domains Reveals Seven Novel Isoforms
J. Biol. Chem.,
May 17, 2002;
277(21):
19071 - 19079.
[Abstract]
[Full Text]
[PDF]
|
 |
|

|
 |

|
 |
 
A. R. Klatt, D. P. Nitsche, B. Kobbe, M. Macht, M. Paulsson, and R. Wagener
Molecular Structure, Processing, and Tissue Distribution of Matrilin-4
J. Biol. Chem.,
May 11, 2001;
276(20):
17267 - 17275.
[Abstract]
[Full Text]
[PDF]
|
 |
|

|
 |

|
 |
 
Q.-q. Wu, Y. Zhang, and Q. Chen
Indian hedgehog Is an Essential Component of Mechanotransduction Complex to Stimulate Chondrocyte Proliferation
J. Biol. Chem.,
September 14, 2001;
276(38):
35290 - 35296.
[Abstract]
[Full Text]
[PDF]
|
 |
|
|