| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Convertase Enzyme from Human Arthritis-Affected Cartilage: Isolation of cDNA by Differential Display, Expression of the Active Enzyme, and Regulation of TNF-1
*
Department of Rheumatology and Medicine, and
Department of Orthopedic Surgery, Hospital for Joint Diseases, New York, NY;
Skirball Institute of Biomolecular Medicine, Departments of
§
Medicine and
¶
Pathology, and
||
Kaplan Cancer Research Center, New York University Medical Center, New York, NY 10016.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
convertase enzyme (TACE). TACE shows the presence of an unknown
prodomain, a cysteine switch, a catalytic domain, a zinc binding
region, a disintegrin region, an EGF-like domain, a transmembrane
domain, and a unique cytoplasmic region. A TACE construct harboring the
signal + prodomain + catalytic region (TACE-SPC
DETCy),
expressed in baculovirus could cleave preferentially (
12-fold) the
TNF-specific peptide over the matrix metalloproteases peptide in vitro.
This recombinant protein also cleaved the natural substrate
GST-ProTNF- to TNF- (17 kDa) in vitro. The mRNA for TACE, which
is broadly distributed and differentially expressed in a variety of
human tissues, is up-regulated in arthritis-affected cartilage, but not
normal cartilage. OA-affected cartilage also expressed TNF- mRNA
that was not detected in normal cartilage. The OA-affected cartilage
(in explant assays) spontaneously released TNF- and IL-8 in ex vivo
conditions. Addition of TNF-R fused to IgG Fc fragment (TNF-R:Fc)
in the presence or absence of soluble IL-1R (with which it acted
additively) significantly attenuated the spontaneous/autocrine release
of articular IL-8 in this assay. These experiments demonstrate a
functional paracrine/autocrine role of TNF- in OA-affected cartilage
that may depend, in part, on up-regulated levels of chondrocyte-derived
TACE. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, a pleiotropic
cytokine,
produces a broad spectrum of injurious effects, which makes it an
important target for therapeutic intervention. TNF- is involved in
the pathophysiology of arthritis, AIDS, cancer, autoimmune diseases
(immune complex diseases), lung fibrosis, multiple sclerosis, skin
delayed-type hypersensitivity reactions, and bacterial and parasitic
infections (1, 2, 3, 4, 5, 6).
The gene for human TNF- encodes a prohormone that is inserted into
the cell membrane as a polypeptide with a molecular mass of 26 kDa (5).
This membrane-bound form of TNF- is bioactive as assayed by cell
cytotoxicity and has been implicated in the paracrine activities of
TNF- in various tissues (6). In response to LPS and other stimuli,
the 26-kDa form of proTNF- is proteolytically cleaved (by a
metalloprotease referred to as TNF- convertase) into a soluble
17-kDa polypeptide (6). TNF- binds its cognate receptors (p55 and
p75) as a bioactive trimer and triggers complex intracellular signaling
pathways (6, 7).
Studies in animal models of arthritis indicate that TNF- may be a
pivotal cytokine involved in these disease processes. Injection of
anti-TNF- Abs (8) or soluble TNF-R (9) has proven to be
highly effective in reducing clinical score, paw swelling, and
histologic severity of the disease. Extension of these studies in
patients with rheumatoid arthritis in phase II and III clinical trials
has yielded very encouraging results. These early clinical results
indicate that neutralizing the effects of TNF- may have profound
effects on clinical symptoms and signs (10, 11).
The matrix metalloproteases (MMPs)4 have long been implicated in cartilage degradation in both rheumatoid arthritis (RA) and osteoarthritis (OA), although their precise role in disease progression remains to be determined (12). The two main families of MMPs believed to be responsible for cartilage degradation are collagenases and proteoglycanases (12).
The synthesis of TNF is known to be regulated at transcriptional,
translational, and posttranslational levels (6). Three independent
groups have demonstrated that broad-spectrum inhibitors of MMPs can
specifically inhibit the release of membrane proTNF- (but not
IL-1ß or IL-6) from various cell surfaces, including RA synovial cell
cultures (13, 14, 15, 16, 17). This inhibitor of proTNF- processing could
protect mice against a lethal dose of endotoxin administered to them
(14). The "TNF- convertase" activity was isolated using these
inhibitors as ligands by affinity purification, which resulted in
identification of an 80-kDa protein with the capacity to
cleave the Gln-Ala-
-Val-Arg sequence of proTNF- (15, 16).
In the present study we report the following: a) the full-length cDNA
sequence of a snake venom-like protease from human arthritis-affected
cartilage (cSVP) is identical to TACE cDNA (15, 16); b) the functional
expression of a truncated recombinant TACE showing TNF- convertase
activity; c) the up-regulation of both TACE and proTNF- mRNA in OA-
and RA-affected, but not in normal cartilage; and d) the intraarticular
TNF- expressed in OA-affected cartilage is sufficient to modulate
the autocrine cartilage-derived IL-8 production. We conclude from these
data that TNF-, which is induced in arthritis-affected cartilage,
may be influenced by the up-regulated TNF- convertase to promote
cartilage catabolic activities.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Cartilage slices were taken from the knees of patients with the diagnosis of advanced OA or RA who were undergoing knee replacement surgery, and from nonarthritic knees. The OA/RA patients were free of nonsteroidal anti-inflammatory drugs for at least 2 wk before surgery. Nonarthritic knee cartilage was obtained within 12 h from accident victims who were undergoing knee amputation. Some of the samples were provided by the National Disease Research Interchange (Philadelphia, PA).
OA cartilage organ culture
Organ culture was conducted as previously described (18).
Briefly, knee articular cartilage, from patients with OA or RA
undergoing knee replacement surgery, was obtained and cut into 3-mm
discs; four to six discs (100 mg) were placed in organ culture in 2
ml of medium for 24 to 72 h, with or without modulators. The
medium was analyzed for nitrite (18), PGE2 by RIA
(Sigma Kit, Sigma, St. Louis, MO), and IL-8 and TNF- by ELISA (R&D
Systems, Minneapolis, MN).
RNA extraction from articular cartilage
Total RNA from OA/RA-affected cartilage and normal cartilage was extracted by guanidium thiocyanate-phenol/chloroform extraction method (19), with minor modifications. The cartilage was milled into fine powder and extracted with 4 M guanidium thiocyanate, 25 mM sodium citrate, 0.5% sodium dodecyl sarcosine, and 0.1 M 2-ME. The aqueous phase was layered onto cesium trifluoroacetate gradient for ultracentrifugation (24,000 rpm/24 h) and the RNA was extracted. The RNA obtained with this method is pure enough for RT-PCR analysis.
PCR amplification of OA/RA-affected and normal cartilage RNA
Reverse transcription from 1 µg of total RNA from OA, RA, and normal cartilage was conducted according to manufacturer’s recommendation at 42°C for 1 h with Superscript II and oligo(dT) primers (Life Technologies, Gaithersburg, MD). The first strand cDNA was used for PCR with the following degenerate primers: A) forward primers as a mixture of two cysteine switch domain primers, (i) classical MMP’s cysteine switch degenerate 5'-CCN (A/C)GN TG(C/T) GGN GTN CCN, where N = A, G, C or, T, and (ii) snake venom protease cysteine switch degenerate 5'-CCN AA(A/G) ATG TG(C/T) GGN GTN, where N = A, G, C, or T; and B) reverse primer, a degenerate zinc binding domain, 5'-(G/A)TG NCC (G/A)A(G/A) (C/T)TC (G/A)TG NG(C/T). An amount equal to 100 pmol of each primer was used (forward, a 50-pmol mixture of two cysteine switch domain degenerate primers and 100 pmol of zinc binding domain primers). Three PCR reactions for each RNA species (OA/RA-affected and normal) were conducted in a Perkin-Elmer/Cetus DNA Thermal Cycler (Perkin-Elmer/Cetus, Norwalk, CT) for 40 cycles of denaturation (94°C, 1 min), annealing (40°C/45°C/50°C, 1 min), and extension (72°C, 2 min). The PCR products were excised from low melting agarose gel and purified using Promega PCR Cleaning Kit and ligated in pGEM-T vector (Promega, Madison, WI). The collagenase+/stromelysin+ clones were excluded by colony hybridization as described below. DNA isolated from the rest of the colonies were analyzed. Computer analysis of DNA and protein sequences was performed using the GCG software package of the University of Wisconsin (Madison, WI), Genetics Computer Group (20).
Southern hybridization
The amplified RT-PCR fragments (on the filter) were probed with stromelysin (341-bp to 1477-bp fragment) (21) and collagenase (396-bp to 883-bp fragment) (22). Clone A (cSVP) (which represents 768 bp to 1433 bp of the sequence deposited to GenBank (accession number U92649)) and the 500-bp fragment(s) were also used as probes. All fragments were labeled using[ 32P]dATP (Amersham Life Sciences, Arlington Heights, IL) and Random Primers DNA labeling system (Life Technologies). The membranes were hybridized with 32P-labeled probes overnight at 65°C. After hybridization, the membranes were washed at high stringency twice for 10 min at 65°C. The blots were exposed to Kodak x-ray films (Kodak, Rochester, NY) at -70°C. All methods were conducted according to Sambrook, et al. (23).
Colony hybridization
White colonies, including one positive control each for
collagenase and stromelysin, were taken on grids on 150-mm plates
(400 colonies per plate) and incubated overnight at 37°C. Colonies
were transferred onto nitrocellulose membrane. The filters were then
fixed with a UV Crosslinker (Stratagene, La Jolla, CA). Treatment of
membranes, prehybridization, hybridization, and washing conditions used
were according to published protocols (22). Membranes were exposed to
an autoradiographic film at -70°C. The
collagenase+/stromelysin+ clones were
excluded from the grids before sequencing the inserts of some of the
other clones.
Cloning of full-length cSVP (TACE) by 3' and 5' rapid amplification of cDNA ends (RACE)
The 3' and 5' extensions of cSVP were cloned using the 3'RACE
and 5'RACE methods according to manufacturer’s instructions (Life
Technologies). Briefly, total RNA was extracted from a pool of three
OA-affected cartilage samples. For 3'RACE, 1 µg of this pooled RNA
was synthesized into first strand cDNA using SuperScript Reverse
Transcriptase and adaptor primer 5'-GGC CAC GCG TCG ACT AGT ACT TTT TTT
TTT TTT TTT T (Life Technologies). This first strand cDNA was used to
amplify the 3' end of cSVP; a sense gene-specific primer called primer
MMP.A (5'-TAG TGG TTT GAC GAG CC) and an abridged universal
amplification primer, AUAP (5'-GGC CAC GCG TCG ACT AGT AC; Life
Technologies), were used for 30 cycles in a Perkin-Elmer thermal cycler
(94°C for 1 min, 55°C for 1 min, and 72°C for 2 min). The PCR
products were purified using Promega PCR Purification Kit and
quantitated. Ten nanograms of PCR product was taken and reamplified,
using a gene-specific sense-nested primer called MMP.B (5'-AAG GAA GCT
GAC CTG GTT) and AUAP for 20 cycles at the same PCR conditions
mentioned earlier. PCR products were electrophoresed on 1% agarose
gel, and 1.6- to 2.0-kb bands were excised, purified by PCR Cleaning
Kit (Promega), ligated into pGEM-T (Promega), and transformed into XL-1
Blue MRF'-competent Escherichia coli cells (Stratagene). DNA
was prepared from isolated white colonies and digested with
PstI restriction enzyme (unique site identified in cSVP
Clone 8). The products were run on 1% agarose gel, blotted on
nitrocellulose, and probed with 32P-labeled cSVP Clone 8
(SalI-SphI, 900 bp) fragment. Four
positive clones were picked and DNA were sequenced on an ABI 373
sequencer (Perkin-Elmer).
For 5' RACE, the first strand was synthesized using a gene-specific primer, MMP.C primer (5'-GCG CGA AGC TTA CTC TTC TCC TTC ATC CAC), and SuperScript reverse transcriptase (Life Technologies). The first strand was dC-tailed at the 5' end and amplified using MMP.C (shown above) and anchor primer (5'-CUA CUA CUA CUA GGC CAC GCG TCG ACT AGT ACG GGI IGG GII GGG IIG (Life Technologies), using the above mentioned conditions. The PCR products were run on 1% agarose gel, and bands of 1.0 to 1.6 kb were excised and ligated in pGEM-T and transformed in XL-1 Blue MRF'. White colonies were selected, and DNA was prepared and digested with NcoI restriction enzyme (another unique site identified in cSVP Clone 8) and blotted as mentioned above. Four positive clones were selected, and DNA was prepared and sequenced as above.
Northern blot analysis
The filters for the human multiple Northern (MTN) blot were purchased from Clontech (Palo Alto, CA). These blots contained 2 µg of poly(A)+ RNA per lane. Equal loading of the RNA was confirmed by probing representative filters with ß-actin cDNA probe. The presence of a single 2.0-kb band (for ß-actin) in all lanes was confirmed, except heart and skeletal muscle, which have two isoforms of ß-actin: 2.0 kb and 1.8 kb. These filters were probed with TACE catalytic region (spanning the cysteine switch and zinc binding region) as recommended by the manufacturer. Northern blot analysis was performed using the same probe but with different established cell lines, such as THP-1 (American Type Culture Collection, Rockville, MD (ATCC) 202-TIB), human chondrosarcoma (ATCC 94-HTB), Jurkat (ATCC 152), HEK 293 (ATCC 1573), and C20A4 (human immortalized chondrocyte, a kind gift from Dr. Mary Goldring (24). Total RNA was isolated using the TRI REAGENT kit (Molecular Research Center, Cincinnati, OH).
Construction and expression of recombinant truncated TACE in baculovirus
The TACE cDNA containing the signal region, the prodomain, and
the catalytic domain (TACE-SPCDETCy), which represents amino acids
1–474, was constructed using specific PCR primers and the 5' RACE No.
1-pGEMT clone as template. The following PCR primers were used: sense
primer Met 1, 5'-CGG GAT CCA CC ATG AGG CAG TCT CTC CTA TTC 3' (to
create a BamHI site 5' to ATG); and antisense primer S474,
5'- GCT CTA GAT TAG CTG CGT TCT TGA AAA CAC TC 3' (to create stop
+ XbaI site at serine 474). A typical PCR reaction contained
25 pmols of each primer, 100 ng of template DNA, 5 mM of each dNTP, 2 U
of Vent DNA polymerase (NEB, Beverly, MA), and water to a final volume
of 50 µl. PCR was performed at 94°C for 1 min, 60°C for 1 min,
and 72°C for 2 min, for 20 cycles. The amplified DNA was digested
with BamHI and XbaI and subcloned into pFast
Bac-1 transfer vector (Life Technologies) to generate pFast
Bac/TACE-SPCDETCy. All other recombinant techniques to generate the
recombinant virus and the recombinant protein were conducted as
described (25). The recombinant protein was generated in serum-free
insect cell culture medium: (Express Five SFM) by infecting the host
BTI-TN-5B1–4 (high five) insect cells. Infected cells were harvested
72 h post infection, and the recombinant enzyme recovered from the
medium was used in the assays.
Enzyme assay for recombinant TACE-SPCDETCy with the
TNF--specific and MMP-specific peptides
The TNF--specific peptide substrate (7-methoxy coumarin-4yb)
Acetyl (MCA)-PLAQ
AV-(3-[2,4-dinitrophany)]-2,3-diamino-propianyl-RSSSR-NH2)
and MMP-specific peptide MCA-PLGL-DPA-A R-NH2 were diluted
to a final concentration of 10 µM in a buffer containing 50 mM
Tricine pH 7.5, 100 mM NaCl, 10 mM CaCl2, and 1 mM
ZnCl2. The enzyme reaction contains 150 µl of a
recombinant baculovirus supernatant plus the diluted peptide in 300
µl final volume. This was incubated for 1 h on orbital shaker at
27°C. Reactions were quenched by adding 30 µl EDTA (500 mM), and
plates were read at 320 nM excitation.
Enzyme assay for recombinant TACE-SPCDETCy with proTNF- as a
substrate
Assay of baculovirus-expressed recombinant protein was performed
with [3H]GST-proTNF- as a substrate prepared by
in vitro translation using a rabbit reticulocyte system (Promega). The
acrylamide gel assays were essentially conducted as described by
Gearing, et al. (13). GST-proTNF- was first cloned in pET-3b
expression vector (Promega). The reaction contained
GST-proTNF--pET3b (0.5 µg/µl) 2 µl; RNAsin 1 µl, amino acid
mix (-cys) reaction buffer 2 µl, T7 RNA polymerase 1 µl,
[3H]leucine (5 µci/µl) 12.5 µl, TNT rabbit
reticulolysate 25 µl, and water to a final volume of 50 µl. The
reaction mixture was incubated at 30°C for 90 min followed by
addition of 800 µl of buffer containing 50 mM Tricine, 150 mM NaCl,
0.1% N-octyl glucoside, and 1 mM ß-mercaptoethanol. The
substrate was purified and concentrated using Micro Con 50 (Amicon),
diluted to 1 ml, and stored at 20°C in 100-µl aliquots. Twenty
microliters of baculovirus-expressed TACE-SPCDETCy was mixed with 5
µl of [3H]proTNF- substrate in a 4x buffer mixture
containing 50 mM Tricine, pH 8, 150 mM NaCl, 0.25% Tween-20 (6.25
µl), 10 mM CaCl2, 15% glycerol, and 2.75 µl of a
mixture of protease inhibitors. The reaction was incubated at 37°C
for 30 to 90 min. After incubation, the reaction was quenched by adding
an equal volume of 4x reducing SDS sample buffer. Samples were run on
14% acrylamide gels and stained with Coomassie blue.
RT-PCR analysis of TACE and TNF-
RT-PCR analysis of TACE and TNF- was conducted according to
the manufacturer’s instructions (Life Technologies). The primers used
for TACE were (sense) 5'-CGG GAT CCA TGT TAA AAG TGG ATA ATG AAG and
(antisense) 5'-GCG CGA AGC TTA CTC TTC TCC TTC ATC CAC, which generated
an 900-bp fragment. The primers for proTNF- mRNA amplification
were (sense) 5'-GGA ATT CAT ATG AGC ACT GAA AGC and (antisense) 5'-CGG
CGC CTC GAG TCA CAG GGC AAT GAT, which generated an 700-bp fragment.
The primers to amplify the ß-actin were (sense) 5'-GTG GGG CGC CCC
AGG CAC CA and (antisense) 5'-CTC CTT AAT GTC ACG CAC GAT TTC, which
generated an 500-bp fragment.
Sequence analysis
The sequence search for the cDNAs was performed with the BLAST program, and the modeling was conducted by using the ICM program as recently described (26, 27).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To identify new members of the metalloprotease family produced by human OA- or RA-affected cartilage, two degenerate oligonucleotides were made. One was from the conserved zinc binding region and the other was from the conserved cysteine switch domain from two families of proteases, the matrixins and the snake venom protease, as described in Materials and Methods. The degenerate oligos were used to amplify the first strand cDNA by RT-PCR from RNA obtained directly (without releasing the cells) from normal, OA- and RA-affected cartilage. Using this method, we have identified short half-life (<4 h) mRNAs that are induced in OA-affected cartilage in vivo. These include mRNA for COX-2 (18) and IL-1ß (28).
The annealing temperatures used (for OA, RA, and normal cartilage RNA)
for the RT-PCR (in 3 separate experiments), using equal amount of RNA,
were 40°C, 45°C, and 50°C. The PCR fragments amplified at three
different temperatures for each tissue (i.e., OA, RA, and normal
cartilage) were pooled and analyzed on agarose gels as shown in Figure 1
A. The data showed
amplification of PCR fragments ranging from 200 bp to 1.2 kb. The PCR
products were analyzed for the presence of MMP-1 and MMP-3 by Southern
blot analysis as shown in Figure 1, B and C. As
expected, MMP-1 and MMP-3 fragments (matrixins) could be detected in
the 300- to 400-bp region, which, upon cloning and sequencing, showed
sequence characteristics similar to MMP-1 and -3, respectively (data
not shown). These experiments demonstrated that MMPs could be isolated
using this technique under conditions used in this study.
|
A) that appeared at equal intensity in
normal, OA-affected, and RA-affected cartilage RT-PCR samples after
staining with ethidium bromide (Fig. 1A). This
"500-bp band" has been identified as fibronectin, based on its cDNA
sequence homology (data not shown). This fibronectin cDNA was excised
from the RA lane, purified, radiolabeled, and used as a probe in the
Southern blot hybridization. The filters previously used for probing
MMP-1 and -3 were again utilized (Fig. 1D). The
fibronectin signal was found to be at similar intensity in all 3 lanes.
The data therefore show both identical loading of PCR (on gels) and
equivalent amplification of mRNA (by RT-PCR) of the fibronectin band in
normal, OA-affected, and RA-affected cartilage.
The normal, OA-affected, and RA-affected cartilage RT-PCR bands were
grouped into four regions (A, 200- to 400-bp region; B, 500-bp region;
C, 500- to 800-bp region; and D, above 800-bp region) and cloned in
pGEM-T vector. Colonies were grown on grids and screened as described
in Materials and Methods. Among these, one clone (from
region C of OA-affected cartilage) designated as clone A (668 bp)
showed the presence of a partial zinc binding region (HELGH) at the end
of the 3' region of the insert. On subsequent screening, another clone
was isolated, which was designated as Clone 8 (which also represented
sequences of Clone A). Clone 8 had a complete functional zinc binding
region (HELGHN) and a cysteine switch-like motif (PKVCGY) similar
to that seen in snake venom proteases from Jararhagin (PKMCGV,
GenBank Accession No. 68251) (29). The distance between the
PKVCGY and HELGHN was 690 bp, further excluding it from the
matrixins family of MMPs. To confirm the presence of this
"PCR-amplified" cDNA in the arthritis-affected cartilage, the
original blots (shown in Fig. 1, B and C)
that had been probed with MMP-1, MMP-3, and fibronectin cDNA were
stripped again of their previously used probes and reprobed with Clone
A (Fig. 1E). Similar to MMP-1 and -3, this unknown
cDNA (in the RT-PCR mixture), which hybridizes to Clone A, was also
found to be enhanced in the arthritis-affected cartilage, but the
positive signal was seen at 700-bp region, distinct from the bands
observed with the collagenase- and stromelysin-probed fragments.
Sequence analysis of Clone 8
The sequence analysis of Clone 8 was performed by examining the
characteristic sequence patterns of the PROSITE database. The partial
cDNA sequence of Clone 8 showed only 40% to 50% similarity and
20% to 30% identity to snake venom proteases. The 3-D ribbon
structure of Clone 8 was compared with that of the catalytic region of
atrolysin and adamalysin (30, 31) as the x-ray crystallography
structures of these proteins are solved (pdb code: liag). The peptidase
domain of Clone 8 aligns with only 27% amino acid sequence identity
with the liag sequences. One of the striking features of this alignment
(between Clone 8 and atrolysin/adamalysin) was that all three
histidines important for the catalytic function of atrolysin/adamalysin
were conserved in Clone 8 (32). Furthermore, the key hydrophobic side
chains supporting 3-D folds are also conserved in these three proteins
in the catalytic region. Although the overall sequence identity was not
very high, the alignment score (20.4 for liag) was found to be several
sigmas above the noise level (26). Therefore we concluded (based on the
similarity of the PROSITE signature) that there was a structural and
functional similarity between the Clone 8 peptidase region and that
present in adamalysin (EC:3.4.24.46)/atrolysin (C/D), which belong to
the EC:3.4.24.42 M12B snake venom protease family (32) (Fig. 2A). This partial cDNA
sequence showed >99% homology to the recently described TNF-
convertase (15, 16). These experiments indicated that Clone 8 may be a
fragment of the TNF- convertase, which seems to be up-regulated in
arthritis-affected cartilage.
|
The full-length cDNA was cloned from an OA-affected cartilage mRNA
using both 5' and 3' RACE protocols as described in Materials and
Methods. There was >99% homology in amino acid sequence between
cSVP and TACE, thus indicating that cSVP from OA-affected cartilage was
TACE (15, 16). The sequence analysis for the full-length TACE was
further performed with the new method for sensitive protein fold
recognition (27). Characteristic sequence patterns of the PROSITE
database were searched. The 3-D ribbon structure based on the amino
acid sequence and domain structure for TACE is shown in Figure 2, B and C. The full-length TACE showed the presence
of a signal peptide (aa 1 to 17), a potential prodomain (aa 18 to 216),
a cysteine switch-like region (aa 181 to 188), a catalytic region (aa
217 to 474), a zinc binding region (aa 405 to 417), a disintegrin
cysteine-rich domain (aa 480 to 559), an EGF-like region (aa 571 to
602). a transmembrane domain (aa 672 to 694), and a cytoplasmic tail
(aa 694 to 824).
The prodomain does not show any significant homology to any known
protein. The endopeptidase region, as described above, was similar to
that seen with the snake venom proteases of family M12B (33). The
disintegrin/cysteine-rich region of TACE (29) shows homology to the
TNF-R/nerve growth factor receptor family, which also has a similar
cysteine-rich signature (34, 35). TNF-R has been reported to form
dimers in crystal structures without the ligand, and, therefore, the
potential of TACE to form a dimer cannot be ruled out (35). The TACE
also has an EGF-like domain, a transmembrane domain, and a cytoplasmic
region. The cytoplasmic region, like the prodomain, does not show any
significant homology to any known protein in the public database. The
cytoplasmic region has a putative tyrosine phosphorylation site
(KKLDKQYESL). Substrates of tyrosine protein kinases are generally
characterized by a lysine or an arginine, seven residues to the
N-terminal side of phosphorylated tyrosine. Two lysine residues at
positions 5 and 6 and an aspartic acid residue at position 3 to
N-terminal of the tyrosine (Y) are found in TACE, which is a classical
motif for tyrosine phosphorylation (36). Furthermore, the proline rich
cytoplasmic region of TACE has at least two potential Src-homology 3
domains: PAPQTPGR and PAPVIPSA.
Northern blot analysis of TACE
We examined the distribution of TACE mRNA (using the catalytic
site of TACE as the probe) in various tissues using a commercial
Northern blot (ClonTech) as a source of Poly(A)+ RNA.
We found a 5.0-kb mRNA that was differentially distributed in various
tissues. The salient features of the blot were the high expression in
testis, placenta, lung, and kidney and low expression in ovary, colon,
liver, and brain (Fig. 3A). Northern blot
analysis of 30 µg of total RNA obtained from the various cell lines,
including two chondrocyte-derived cell lines as described in Figure 3B, showed the presence of a dominant 5.0-kb TACE-1
mRNA and a minor 4.0-kb TACE-2 mRNA. Thirty micrograms of total RNA
(pooled from five to ten samples of cartilage) obtained from OA- or
RA-affected cartilage and normal cartilage independently was also
examined by Northern blot analysis. A positive detectable signal could
not be seen in the cartilage-derived mRNA, although TACE mRNA could be
seen in the adjacent lane with equivalent amounts of THP-1 mRNA (data
not shown). This is not surprising, since we have observed that the
poly(A)+ content of human cartilage RNA is approximately
0.1% of the total RNA (data not shown). However, TACE mRNA in
cartilage could be detected by RT-PCR as described below.
|
The TACE-SPCDETCy containing a prodomain (with a putative
signal sequence) and the catalytic region containing a cysteine
switch-like motif and zinc binding domain were cloned in a baculovirus
expression system. This protein was truncated just before the start of
disintegrin domain. The recombinant virus made in Sf9 cells were
further amplified in BTI-TN-5B1–4 cells using the serum-free medium.
The recombinant protein was Western blotted (Fig. 4) using the polyclonal Abs raised
against the catalytic region of TACE (expressed in E. coli;
data not shown) in rabbits. The Western blot shows an 55-kDa
band that corresponds to the unprocessed recombinant protein containing
the prodomain and the catalytic region. This blot also shows an
32-kDa band that seemed to be the processed form of TACE. The
processing of TACE seems to occur at the furin-like cleavage site
(RVKRR) at residue 211 in the prodomain region. Furin and related
enzymes of subtilin-like serine proteases such as PC1/PC3, PC2, PC4,
PC5, PC7, and PACE 4 have been described to utilize the consensus
R-X-K/R-R site for processing of diverse proteins, especially
proenzymes such as factor IX (37), pro-Von Willebrand factor (38),
proalbumin (39), complement pro-C3 (39), and proTGF-ß1 (40). These
results show a) that baculovirus insect cells (Sf9 cells) contain
enzymes that have furin-like activity, and b) that the baculovirus
insect cell cultures can process and secrete the truncated form of the
recombinant TACE. Since TACE contain a cysteine switch-like motif
PKVCGY characteristic of snake venom-like proteases and a zinc binding
domain characteristic of metzincins (41), it is tempting to propose
that the TACE could be activated by cysteine switch mechanism as it was
determined for enzymes of matrixins family such as collagenase 1 and
stromelysin 1 (42).
|
The baculovirus expressed TACE (clone TACE-SPCDETCy) that
harbors the prodomain, and the catalytic site was tested for its
specificity in a fluorogenic assay, using an MCA-labeled peptide
containing the amino acid sequence around the native site, with
proTNF- (MCA-PLAQAV-DPA-RSSSR-NH2) and that of
MMPs. (MCA-PLGL-DPA-AR-NH2) as the substrate. Table I shows the Km and
Vmax of the enzyme in the presence of the two substrates.
The ratio V Ala-Val/V Gly-Leu was 11.94 indicating 12-fold
preferential cleavage of the TNF- peptide over the MMP peptide by
the truncated recombinant enzyme.
|
in vitro.
3H-labeled GST-proTNF- substrate was prepared by in
vitro translation using rabbit reticulocyte system. The substrate was
incubated with the recombinant enzyme for 30 to 90 min, and the
cleavage product (17 kDa TNF-) was examined by SDS-PAGE analysis as
shown in Figure 5. These experiments
demonstrate that the recombinant enzyme cleaved the native proTNF-
to TNF-.
|
in arthritis-affected
cartilage
Our recent observations indicate that mRNA for COX-2 and IL-1ß
are up-regulated in OA-affected cartilage but not in normal cartilage
(19, 28). In view of the up-regulation of TACE mRNA in our initial
screen in arthritis-affected cartilage, as seen in Figure 1, total RNA
was directly isolated from OA- and RA-affected cartilage immediately
after surgery. Normal cartilage was used as control for RT-PCR analysis
using specific primers for TACE. Figure 6
showed up-regulation of TACE mRNA in five arthritis-affected cartilage
samples tested. There was no detectable signal observed in the two
normal cartilage samples tested, even at 35 cycles. It should
be noted that a faint positive TACE signal (after staining the gel with
ethidium bromide) in normal cartilage could be obtained after 40 cycles
of PCR amplification (data not shown) or by probing the
RT-PCR-amplified products with a 32P-labeled TACE
probe by Southern blot analysis, as shown in Figure 1. This may be due
to the constitutive expression of TACE, as reported by Black et al.
(16) and this paper (Fig. 3, A and B).
|
mRNA in
OA-affected and normal cartilage using the same RNA preparations used
for amplifying TACE (Fig. 6). RNA from all the arthritis-affected
patient material showed the presence of TNF- mRNA expression that
was not detectable in the two normal cartilage samples tested. These
experiments indicate that there is an up-regulation of both TACE and
TNF- mRNA in arthritis-affected cartilage.
We have recently observed that OA-affected cartilage explants, when
incubated in ex vivo conditions in F-12 medium (in the absence of any
exogenous supplements including FBS), spontaneously released nitric
oxide, PGE2, and IL-1ß in quantities sufficient to cause
cartilage damage, and these mediators can be down-regulated by sIL-1R
but not by TNF-R:Fc (19, 28). Therefore, we also examined the
spontaneous release of TNF- from cartilage explants obtained from
twelve different patients utilizing the same experimental system. We
could detect TNF- (12.4 ± 11.0 pg/ml per 100 mg OA cartilage,
at 72 h, n = 12) in the serum-free medium when
assayed by ELISA.
We also assayed the spontaneous release of IL-8, which is known to be
stimulated by TNF- in chondrocytes (43), and found it to be present
in significant quantities (5–6 ng/ml) together with nitric oxide
and PGE2 (19, 28). Interestingly, although TNF- was
detected in low concentrations by ELISA in ex vivo conditions,
treatment of explants with TNF-R:Fc significantly inhibited the
spontaneous secretion of IL-8 by OA-affected cartilage (Fig. 7). sIL-1R only marginally inhibited the
spontaneous release of IL-8. However, sIL-1R acted additively with
TNF-R:Fc to markedly inhibit IL-8 release. These experiments
indicate that the OA-affected cartilage spontaneously produces low
levels of TNF- that are sufficient to promote IL-8 production. In
these experiments, TNF-R:Fc + IL-1R could also inhibit the
spontaneous release of nitric oxide and PGE2 production by
more than 50%, as previously reported (28). Thus, chondrocyte-derived
TNF-, like IL-1ß (28), acts in a paracrine/autocrine fashion to
modulate cartilage function.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Moss et al. (15) have shown that the TACE construct harboring signal
region + prodomain + catalytic region + disintegrin
region expressed in baculovirus is functional. Our studies show that a
TACE construct (TACE-SPCDETCy) harboring a signal sequence +
prodomain + catalytic site (devoid of disintegrin domain) and
expressed in baculovirus is sufficient for enzyme activity and
specificity for the TNF- peptide over the MMP peptide. Furthermore,
the truncated recombinant TACE could also cleave the natural proTNF-
substrate to 17-kDa TNF-. Therefore, the C-terminal region of TACE,
unlike the matrixins (where the hemopexin/vitronectin repeats are
implicated in triple helix recognition) (46), has a disintegrin-like
domain that does not seem to have any significant role in the activity
or specificity of the enzyme in in vitro assays. Furthermore, human
TNF- convertase could not cleave mouse proTNF- in vitro (47).
Thus, it seems that human TACE is specific for human proTNF-.
Furthermore, Lunn et al. (48) have shown that ADAM-10 isolated from
bovine spleen also has TNF- convertase activity. This is not
surprising, since several isolated MMPs have also been demonstrated to
cleave proTNF- in vitro (13, 49). It should be noted that T cells
derived from TACE-/- (homozygous mouse) showed
>80% decrease in TNF- release and subsequent enhancement of cell
surface proTNF- accumulation, indicating that TACE is the key enzyme
in proTNF- processing in cell-based assays (16). It is quite
possible that cysteine-rich region may be involved in dimer formation,
based on its sequence homology with TNF-R/NGF-R (35) and/or may act
as an integrin. The disintegrin-like domain-containing proteases
derived from snake venom are responsible for inhibiting platelet
aggregation by interacting with IIbß3 on
platelets (50), and recent studies have also shown that a binding site
in the disintegrin domain of guinea pig fertilin (PH-30) is required
for sperm-egg fusion (51, 52). These observations further emphasize the
importance of disintegrin domain in cell-cell interactions
(51).
One difference between the soluble snake venom-like proteases and their mammalian homologues is the presence of a cytoplasmic region and the transmembrane region (29, 52). The mammalian homologues are membrane bound and may show entirely different function. This may include the inside-out/outside-in signaling of the molecules, in view of a potential tyrosine phosphorylation site and Src-homology 3 binding motifs in the cytoplasmic region of TACE.
Our experiments also suggest that at least two forms of TACE (1 and 2)
may exist based on the molecular weights of the two mRNAs. The
significance of the shorter form is not clear. This is the first
observation that indicates that TACE mRNA is up-regulated in a
pathophysiologic condition where TNF- plays a significant role (4).
Although the role of TNF- in RA is well-documented, it has not been
implicated in the pathophysiology of OA, in view of the comparatively
low levels seen in the synovial fluids of OA patients (53). IL-ß and
TNF- have been shown to independently induce IL-8 in human
chondrocytes (43). Our studies demonstrate that IL-8 produced by
chondrocytes of OA cartilage explants can be inhibited by TNF-R:Fc.
Furthermore, soluble IL-1R, together with TNF-R:Fc, demonstrates an
additive inhibitory effect on the IL-8 production that also highlights
the importance of TNF- in the pathophysiology of OA, not previously
appreciated. These experiments demonstrate that the endogenous TNF-
produced by OA-affected cartilage is functionally active and is present
in quantities sufficient to act in an autocrine or paracrine fashion to
modulate other inflammatory/chemotactic mediators, such as IL-8.
Although OA has been described as a "noninflammatory disease," (54)
we have recently shown superinduction of inflammatory mediators, such
as nitric oxide (18), PGE2 (19), IL-8, IL-6, IL-1 (28), and
MCP-1 (unpublished data) in OA-affected cartilage in ex vivo
conditions. The present studies highlight the role of another
proinflammatory cytokine, TNF-, and its regulator, TACE, along with
other inflammatory mediators described above in the pathophysiology
of OA.
In summary, using a differential display screen between normal and
arthritis-affected cartilage, we have cloned the cDNA for a snake
venom-like protease that has >99% amino acid sequence homology to the
recently described human TNF- convertase (15, 16). We provide
evidence that functional TNF- is secreted in an active form by
OA-affected cartilage, consistent with an up-regulated TNF-
convertase mRNA. These data thus identify two inducible molecules,
TNF- and its regulator, TNF- convertase, that are produced by
OA-affected cartilage and that could serve as future targets for
pharmacologic intervention.
|
Acknowledgments |
|---|
receptors and Dr. M.
Golding for providing the C-20/A4 cells. |
Footnotes |
|---|
2 The first two authors have contributed equally to the paper.
3 Address correspondence and reprint requests to Dr. Ashok R. Amin, Department of Rheumatology, Room 1600, Hospital for Joint Diseases, 301 E. 17th St., New York, NY 10003.
4 Abbreviations used in this paper: MMP, matrix metalloprotease; cSVP, cartilage-derived snake venom-like protease; OA, osteoarthritis; RA, rheumatoid arthritis; sIL-1R, soluble IL-1R; GST-proTNF-, glutathione-S-transferase-proTNF-; TNF-R:Fc, TNF- receptor fused to IgG Fc fragment; TACE, TNF- convertase enzyme; EGF, epidermal growth facor; RACE, rapid amplification of cDNA ends.
Received for publication October 1, 1997. Accepted for publication January 6, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
antibody therapy and its mechanism of action. J. Inflamm. 47:90.
): the good, the bad and potentially very effective. Immunol. Cell Biol. 74:434.[Medline]
precursor by metalloproteinases. Nature 370:555.[Medline]
. Nature 385:733.[Medline]
from cells. Nature 385:729.[Medline]
processing in RA synovial membrane cell cultures.
J. Clin. Invest. 97:2833.
converting activity. Eur. J. Immunol. 27:1275.[Medline]
convertase. FEBS Lett. 400:333.[Medline]
and Notch. Cell 90:589.[Medline]
This article has been cited by other articles:
|
|
 |
|
  Y. Wang, A. H. Herrera, Y. Li, K. K. Belani, and B. Walcheck Regulation of Mature ADAM17 by Redox Agents for L-Selectin Shedding J. Immunol., February 15, 2009; 182(4): 2449 - 2457. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. B. M. Reddy, K. V. Ramana, S. Srivastava, A. Bhatnagar, and S. K. Srivastava Aldose Reductase Regulates High Glucose-Induced Ectodomain Shedding of Tumor Necrosis Factor (TNF)-{alpha} via Protein Kinase C-{delta} and TNF-{alpha} Converting Enzyme in Vascular Smooth Muscle Cells Endocrinology, January 1, 2009; 150(1): 63 - 74. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Charbonneau, K. Harper, F. Grondin, M. Pelmus, P. P. McDonald, and C. M. Dubois Hypoxia-inducible Factor Mediates Hypoxic and Tumor Necrosis Factor {alpha}-induced Increases in Tumor Necrosis Factor-{alpha} Converting Enzyme/ADAM17 Expression by Synovial Cells J. Biol. Chem., November 16, 2007; 282(46): 33714 - 33724. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T Bohgaki, Y Amasaki, N Nishimura, M Bohgaki, Y Yamashita, M Nishio, K-i Sawada, S Jodo, T Atsumi, and T Koike Up regulated expression of tumour necrosis factor {alpha} converting enzyme in peripheral monocytes of patients with early systemic sclerosis Ann Rheum Dis, August 1, 2005; 64(8): 1165 - 1173. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. Li and H. Fan Loss of Ectodomain Shedding Due to Mutations in the Metalloprotease and Cysteine-rich/Disintegrin Domains of the Tumor Necrosis Factor-{alpha} Converting Enzyme (TACE) J. Biol. Chem., June 25, 2004; 279(26): 27365 - 27375. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. V. Bax, A. J. Messent, J. Tart, M. van Hoang, J. Kott, R. A. Maciewicz, and M. J. Humphries Integrin {alpha}5{beta}1 and ADAM-17 Interact in Vitro and Co-localize in Migrating HeLa Cells J. Biol. Chem., May 21, 2004; 279(21): 22377 - 22386. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. F. Seals and S. A. Courtneidge The ADAMs family of metalloproteases: multidomain proteins with multiple functions Genes & Dev., January 1, 2003; 17(1): 7 - 30. [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Chen, K. Diener, I. R. Patel, J. K. Kawooya, G. A. Martin, P. Yamdagni, X. Zhang, A. Sandrasagra, S. Sahasrabudhe, and S. J. Busch Phage Display of Functional Human TNF-{alpha} Converting Enzyme Catalytic Domain: A Rapid Method for the Production of Stabilized Proteolytic Proteins for Assay Development and High-Throughput Screening J Biomol Screen, October 1, 2002; 7(5): 433 - 440. [Abstract] [PDF]
|
|
|
|
 |
|
 J Brynskov, P Foegh, G Pedersen, C Ellervik, T Kirkegaard, A Bingham, and T Saermark Tumour necrosis factor {alpha} converting enzyme (TACE) activity in the colonic mucosa of patients with inflammatory bowel disease Gut, July 1, 2002; 51(1): 37 - 43. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. G. Attur, M. N. Dave, R. M. Clancy, I. R. Patel, S. B. Abramson, and A. R. Amin Functional Genomic Analysis in Arthritis-Affected Cartilage: Yin-Yang Regulation of Inflammatory Mediators by {alpha}5{beta}1 and {alpha}V{beta}3 Integrins J. Immunol., March 1, 2000; 164(5): 2684 - 2691. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Wang, A. R. Johnson, Q.-Z. Ye, and R. D. Dyer Catalytic Activities and Substrate Specificity of the Human Membrane Type 4 Matrix Metalloproteinase Catalytic Domain J. Biol. Chem., November 12, 1999; 274(46): 33043 - 33049. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. A. Solomon, N. Pesti, G. Wu, and R. C. Newton Cutting Edge: A Dominant Negative Form of TNF-{alpha} Converting Enzyme Inhibits ProTNF and TNFRII Secretion J. Immunol., October 15, 1999; 163(8): 4105 - 4108. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. M. Roberts, P. H. Tani, L. C. Bridges, Z. Laszik, and R. D. Bowditch MDC-L, a Novel Metalloprotease Disintegrin Cysteine-rich Protein Family Member Expressed by Human Lymphocytes J. Biol. Chem., October 8, 1999; 274(41): 29251 - 29259. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. J. Garton, P. J. Gough, C. P. Blobel, G. Murphy, D. R. Greaves, P. J. Dempsey, and E. W. Raines Tumor Necrosis Factor-alpha -converting Enzyme (ADAM17) Mediates the Cleavage and Shedding of Fractalkine (CX3CL1) J. Biol. Chem., October 5, 2001; 276(41): 37993 - 38001. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
