(Received for publication, March 4, 1997, and in revised form, May 21, 1997)
From the Department of Pathology, Duke University Medical Center, Durham, North Carolina 27710
Receptor-recognized forms of
2-macroglobulin (
2M*) bind to two
classes of cellular receptors, a high affinity site comprising approximately 1500 sites/cell and a lower affinity site comprising about 60,000 sites/cell. The latter class has been identified as the
so-called low density lipoprotein receptor-related protein (LRP).
Ligation of receptors distinct from LRP activates cell signaling
pathways. Strong circumstantial evidence suggests that the high
affinity binding sites are responsible for cell signaling induced by
2M*. Using sodium hypochlorite, a powerful oxidant produced by the
H2O2-myeloperoxidase-Cl
system,
we now demonstrate that binding to the high affinity sites correlates
directly with activation of the signaling cascade. Oxidation of
2M* using 200 µM hypochlorite completely
abolishes its binding to LRP without affecting its ability to activate
the macrophage signaling cascade. Scatchard analysis shows binding to a
single class of high affinity sites (Kd
71 ± 12 pM). Surprisingly, oxidation of native
2-macroglobulin (
2M) with 125 µM hypochlorite results in the exposure of its
receptor-binding site to LRP, but the ligand is unable to induce cell
signaling. Scatchard analysis shows binding to a single class of lower
affinity sites (Kd
0.7 ± 0.15 nM). Oxidation of a cloned and expressed carboxyl-terminal
20-kDa fragment of
2M (RBF), which is capable of binding
to both LRP and the signaling receptor, results in no significant
change in its binding Kd, supporting our earlier
finding that the oxidation-sensitive site is predominantly outside of
RBF. Attempts to understand the mechanism responsible for the selective
exposure of LRP-binding sites in oxidized native
2M
suggest that partial protein unfolding may be the most likely mechanism. These studies provide strong evidence that the high affinity
sites (Kd
71 pM) are the
2M* signaling receptor.
2-Macroglobulin
(
2M)1 is a
highly conserved, homotetrameric, 720-kDa glycoprotein found in high
concentration in the plasma (2-4 mg/ml). It has the unique ability to
inhibit all mechanistic classes of proteinases by "entrapping" the
proteinase and thereby sterically blocking the access of high molecular
weight substrates (reviewed in Refs. 1 and 2). Proteinases first cleave
the "bait region" of native
2M exposing the internal
-glutamyl-
-cysteinyl thioester bond. Reaction of the thioester
bond with a free amino lysyl residue on the surface of the proteinase
results in bond rupture and a major conformational change in native
2M. The resulting molecule is much more compact as
evidenced by faster migration on a native acrylamide gel (3), electron
microscopy (4, 5), sedimentation behavior (6), and circular dichroism
(7). Consequently the receptor-recognition site is exposed. Small amine nucleophiles, such as methylamine, can initiate this reaction by
directly attacking the thioester bond generating the conformational change and the exposure of the receptor-recognition site without bait
region cleavage. Receptor-recognized
2M
(
2M*) can rapidly eliminate the "entrapped"
proteinase from the circulation by binding to a cell surface clearance
receptor, the low density lipoprotein receptor-related protein (LRP)
(8, 9).
LRP is a multiligand receptor that binds to a wide variety of unrelated
ligands (reviewed in Ref. 10). Binding of all ligands to LRP can be
effectively competed by receptor-associated protein (RAP), which
co-purifies with LRP. Prior investigation of 2M* binding
to LRP has shown that the binding mechanism involves a cluster of
positively charged residues on
2M* interacting with the
second complement-like repeat on LRP, which contains clusters of
negatively charged residues (11). Analysis of the receptor-binding site
on
2M* using monoclonal antibody (12, 13) and
recombinantly expressed protein (14, 15) demonstrates that the carboxyl terminus of
2M* is involved in receptor binding.
Although LRP is the only 2M* receptor identified to
date, some important cellular regulatory functions ascribed to
2M* suggest that an alternate receptor must exist.
2M*, but not native
2M, suppresses the
production of superoxide anion (16), enhances the release of
prostaglandin E2 (17, 18) and platelet activating factor
(19), and stimulates the proliferation of vascular smooth muscle cells
(20). Moreover, our laboratory has characterized a novel signaling
cascade and found that it does not appear to be LRP-mediated (21-23).
Furthermore, we have identified two classes of
2M*
binding sites on peritoneal macrophages and human trabecular meshwork
cells, both of which demonstrate activation of signaling cascades after
exposure to
2M* (24). The lower affinity binding site is
10 times more abundant than the high affinity binding site and has
clearly been identified as LRP (25, 26). The identity of the signaling
receptor remains elusive; however, using site-directed mutagenesis, we
have found that a lysine residue (1374, human numbering) within a
20-kDa fragment constituting the carboxyl terminus of
2M
(RBF) is important for signaling (26).
Very recently, however, some investigators reported that residue 1374 is involved in binding to LRP as well, raising the possibility that the
2M* signaling receptor may not be a separate receptor (27). Our previous attempts to study
2M* binding to LRP
and the signaling receptor using
cis-dichlorodiamine-platinum(II) (cis-DDP)
modification have shown that cis-DDP modifies a region upstream of the 20-kDa carboxyl-terminal of
2M* and that
this modification results in decreased binding to LRP while having no
effect on cell signaling (25, 28, 29). This observation together with
other immunochemical studies (12, 13, 30) suggest that a region outside
of RBF may be involved in
2M* binding to LRP. Further
characterization of the receptor-binding sites using RBF demonstrated
that this fragment both binds to LRP and retains the ability to induce
2M* signaling cascade (23, 26). Mutational studies have
suggested that LRP and the
2M* signaling receptor are
distinct entities. To date, however, no data have demonstrated a
complete dissociation between
2M* binding to LRP and to
the signaling receptor.
Previous studies have shown that 25 µM sodium
hypochlorite completely abolishes the anti-proteinase activity of
native 2M (31, 32). Its effects on
2M*
receptor-recognition have not been examined. In this study we
demonstrate that hypochlorite oxidation of
2M*
completely destroys its ability to bind to LRP without affecting its
ability to bind to the signaling receptor. This modification occurs
predominantly outside of the carboxyl-terminal 20 kDa, consistent with
our previous finding that the cis-DDP-sensitive site is
upstream of RBF. Surprisingly, we also found that although hypochlorite
oxidation of native
2M results in the selective exposure
of the receptor-recognition site to LRP, the ligand cannot signal,
thereby providing direct evidence for the dissociation of
2M* binding to LRP from binding to the signaling
receptor.
Recombinant RAP construct was a kind gift of Dr.
Joachim Herz (University of Texas, Southwestern, Dallas, TX).
Recombinant RBF construct was prepared as in (14). The production of
both recombinant RAP and RBF is described in detail in Ref. 26. RPMI 1640, fetal bovine serum, Hanks' balanced salt solution (HBSS), and
HBSS without CaCl2, MgCl2, MgSO4
were purchased from Life Technologies, Inc. Bovine serum albumin (BSA),
HEPES, sodium hypochlorite, L-methionine, and EDTA were
purchased from Sigma. Tris base was purchased from Boehringer Mannheim.
PD-10® column was purchased from Pharmacia Biotech Inc. (Uppsala,
Sweden). Carrier-free 125I and
125I-Bolton-Hunter reagent for protein iodination was
obtained from NEN Life Science Products.
1-[2-(5-Carboxyoxazol-1-yl)-6-aminobenzofuran]-5-oxyl-2-(2-amino-5
-methylphenoxy)ethane-N,N,N
,N
-tetraacetic acid acetoxylmethyl ester (Fura-2/AM) was purchased from Molecular Probes (Eugene, OR). BCA protein assay kit, IODO-BEADS®, and 10,000 molecular weight cut-off Slide-A-Lyzer® dialysis cassettes were purchased from Pierce. All other reagents were of the highest quality
commercially available.
Human native 2M was purified
according to a previously published protocol (33). Native
2M was activated with 200 mM methylamine in
a buffer containing 150 mM NaCl, 50 mM Tris, pH
8.0, for 16-18 h at room temperature in the dark. Unreacted
methylamine was removed by dialysis for 48 h with four changes of
buffer containing 150 mM NaCl, 25 mM sodium
phosphate, pH 7.4. Dialyzed
2M* was sterile-filtered using 0.22-µm syringe microfilters from Millipore (Bedford, MA), stored at 4 °C, and used within 2 weeks. Native
2M
and
2M* were iodinated using either IODO-BEADS® or
125I-Bolton-Hunter reagent according to the
manufacturer-specified protocol. Specific activity of
125I-
2M* varied from 1000 to 1500 cpm/ng for
IODO-BEADS® labeling and 500-700 cpm/ng for
125I-Bolton-Hunter labeling method. The molecular mass of
2M* used in these experiments is 720 kDa (34).
BALB/c mice were immunized three times at 4-week intervals with 50 µg of antigen (RBF) and Titermax® adjuvant (Vaxcel, Norcross, GA) (35). Two weeks after the third immunization, retro-orbital blood was collected, and polyclonal sera from the individual mice were screened for antigen activity by ELISA with RBF as the coating antigen. The mouse with the best titer was boosted with 50 µg of antigen. Four days following the final boost, the mouse was euthanized and bled by cardiac puncture.
Oxidation of NativeOxidation of native 2M,
2M*,
and RBF was performed essentially as described previously (32). In
brief, native
2M,
2M*, and RBF were
incubated with sodium hypochlorite (from 1.25 µM to 2 mM) for 15 min at 37 °C in phosphate-buffered saline. At the end of the incubation, 20 mM of
L-methionine was added to the mixture to quench residual
oxidants. The sodium hypochlorite concentration was determined by
measuring the absorption at a wavelength of 292.5 nm using 206 M
1 cm
1 as the extinction
coefficient at pH 7.5 (36). To ensure that oxidation did not result in
a loss of 125I label from proteins, trichloroacetic acid
precipitation of 250 µM hypochlorite-oxidized
125I-
2M was performed. No significant loss
of labeling (i.e. less than 5%, n = 4) was
found.
The spectral differences between
hypochlorite oxidized native 2M,
2M*, and
RBF, and nonoxidized native
2M,
2M*, and
RBF, were analyzed on a DU® 640 spectrophotometer (Beckman
Instruments) as described previously (37) with the following
modifications. Native
2M,
2M*, or RBF
(0.25 mg/ml) was first oxidized according to the published protocol. 1 ml of each sample was then added to the sample cuvette and measured
against nonoxidized native
2M,
2M*, or
RBF (0.25 mg/ml) in the reference cuvette. The absorption difference
from
220 nm to
400 nm was calculated. As controls, the absorptions of L-methionine and methionine
sulfoxide at these wavelengths were found to be negligible.
Nondenaturing,
nonreducing gradient (5-15%) PAGE, or reducing SDS-PAGE (7.5%) were
performed to determine the effects of oxidation on the structure of
2M. To visualize the protein bands, gels were fixed in
acetic acid with Coomassie Brilliant Blue. Hypochlorite oxidation at a
concentration up to 125 µM had no effect on the ability
of
2M to be stained by the dye. Oxidation at greater 250 µM, however, appears to decrease the ability of
2M to be Coomassie-stained.
Proteins to be tested were incubated in 96-well
Immulon plates (Dynatech, Chantily, VA) for 1 h in 0.1 M NaHCO3, pH 9.6, at room temperature.
Following incubation, each well was washed twice in PBS-Tween
(phosphate-buffered saline, 0.05% Tween-20) to remove unbound
proteins. 50 µl of PBS-Tween with 4% BSA were then added to each
well for 0.5 h at room temperature to block nonspecific binding
sites. Following incubation, each well was washed twice with PBS-Tween.
To each well was then added 50 µl of 1:100 dilution of primary
antibody against the protein to be tested and then incubation was
continued for 1 h at 25 °C. Following this incubation, each
well was washed twice with PBS-Tween and then 50 µl of 1:400 dilution
of anti-mouse IgG-horseradish peroxidase-conjugated antibody was added.
After 1-h incubation at 25 °C, the unbound anti-mouse IgG antibody
was removed by washing with PBS-Tween and 50 µl of o-phenylenediamine dihydrochloride substrate solution was
added to each well. The enzyme-substrate reaction was allowed to
proceed for 15 min and then the absorbance at a wavelength of 450 nm
was determined with a microplate reader (Molecular Devices, Menlo Park,
CA). Since oxidation may alter the binding of oxidized
2M to the plate, we compared the quantity of the
oxidized proteins bound to the plate with that of the nonoxidized
protein. We found no statistically significant difference
(n = 3) in the amount of proteins bound between
oxidized and nonoxidized
2M.
Pathogen-free female C57BL/6 mice were obtained from Charles River Laboratories (Raleigh, NC). Peritoneal macrophages were obtained as described previously (38). Thioglycollate-elicited peritoneal macrophages were harvested by peritoneal larvage using 5 ml of 150 mM NaCl, 20 mM HEPES, pH 7.4. The macrophages were pelleted by centrifugation at ~800 × g for 5 min and suspended in RPMI 1640 medium containing 12.5 units/ml penicillin, 6.254 mg/ml streptomycin, and 10% fetal bovine serum. Cell viability was assayed by the trypan blue method (Hausser Scientific, Horsham, PA).
Cell Surface Binding AssayMacrophages were plated in
24-well cell culture plates (Becton Dickinson, Lincoln, Park, NJ) at
1 × 106 cells/well (with a maximum of 7% variation
between wells, n = 8) and incubated for 3 h at
37 °C in a humidified 5% CO2 incubator. These plates
were then cooled to 4 °C, and unbound cells were removed by rinsing
three times with ice-cold HBSS containing 20 mM HEPES, 5%
BSA, pH 7.4 (buffer A). As a control for nonspecific 2M*
binding, some of the wells were rinsed three times with ice-cold HBSS
without CaCl2, MgCl2, MgSO4
containing 20 mM HEPES, 5% BSA, and 5 mM EDTA,
pH 7.4 (buffer B) to assess calcium-independent binding. For RAP
competition studies, radiolabeled ligands (2 nM) were added
to each well in the presence or absence of unlabeled RAP. For oxidized
native
2M and oxidized
2M* competition
binding studies, various concentrations of unlabeled inhibitors were
added to each well followed immediately by the addition of radiolabeled ligands (0.5 nM for 125I-
2M* and
11 nM for 125I-RBF). Cells were then incubated
at 4 °C for 12-16 h. Unbound ligand was removed from the wells, and
the cell monolayer was rinsed twice with ice-cold buffer A or B. Cells
in each well were then solubilized with 0.1 M NaOH, 0.5%
SDS at room temperature for > 5 h before transferring to
polystyrene tubes to be counted in a
-counter (LKB-Wallac,
CliniGamma 1272). Specific binding of
2M* to cells was
determined by subtracting calcium-independent binding, which averaged
less than 10% over four experiments, from total binding. To verify
that calcium-dependent binding represents binding to both
the signaling receptor and LRP, radioligand competition binding
experiments in the presence of 100-fold excess unlabeled ligands were
performed (n = 3). No difference was observed between binding in the presence of buffer B or with 100-fold excess unlabeled ligand.
The measurement of intracellular calcium level, [Ca2+]i, was performed according to a previously published protocol (21). Briefly, macrophages were plated on glass coverslips at a density of 400,000 cells/cm2. The cells were then incubated for 16-18 h in a humidified 5% CO2 incubator at 37 °C. After addition of 4 µM of Fura-2/AM, cells were incubated for another 30 min in the dark at 25 °C. The coverslips containing adherent cells were then washed twice with HBSS without calcium and mounted on a digital imaging microscope. After obtaining a stable base-line level of [Ca2+]i, ligands were added and the level of [Ca2+]i was immediately measured. Typical base-line [Ca2+]i was approximately 100-150 nM. A positive response was typically 2-4-fold increase over base line. All ligands and buffers were tested negative (<0.01 enzyme units/ml) for endotoxin using the Limulus amebocyte lysate assay (Associates of Cape Cod, Woods Hole, MA).
Data AnalysisIn direct binding studies, the
Bmax and Kd were derived from
the x intercept and the slope of the Scatchard plot, respectively. These numbers were also verified using least square analysis based on single-site binding using the Systat® 5.0 computer program. The apparent dissociation constant for the unlabeled ligand,
Ki, was determined using the equation
Ki = IC50/(1 + L/Kd), where L and
Kd are the concentration and the dissociation
constant for the radiolabeled protein, respectively (37). The
concentrations of radiolabeled proteins used in oxidized native
2M and oxidized
2M* cold competition
experiments equal the dissociation constant. This reduces the above
equation to Ki = IC50/2.
To quantitate the
concentration dependence of native 2M,
2M*, and RBF modifications, the spectrophotometric
absorption differences between oxidized and nonoxidized native
2M,
2M*, and RBF from
220 nm to
400 nm were determined. As
shown in Fig. 1, the peak difference
between oxidized and nonoxidized native
2M and
2M* occurred at approximately
242 nm
and, to a lesser extent, at
300 nm. The absorption at
242 nm corresponds to a mixture of species with
chloramine modification of amino terminus and lysine side chain being
the most abundant as determined by trinitrobenzene sulfonate titration
(39, 40). The absorption at
300 nm corresponds to the
formation of dichloramine (41). The hypochlorite oxidation of RBF
resulted in minimal modification (Fig. 1C). This is not
unexpected since our previous attempts to modify RBF using
cis-DDP and hydrogen peroxide, which react at the same
residues, have both failed to affect RBF (42-44). Fig. 1D
shows the concentration dependence of native
2M,
2M*, and RBF oxidation at
242 nm. Both native
2M and
2M* were equally
susceptible to hypochlorite oxidation at all concentrations. Moreover,
this modification plateaued at an oxidant concentration of
approximately 2 mM. On the other hand, RBF oxidation showed
minimal protein modification even at high oxidant concentrations.
Binding Competition of Oxidized
The effect of hypochlorite modification on the
receptor binding properties of 2M* was determined. Two
different sets of experiments were performed. The first set showed that
oxidation significantly decreased the ability of
2M* to
compete for the binding of nonoxidized
2M* (Fig.
2A). This effect appeared to
take place only when
2M* was oxidized at a hypochlorite
concentration of at least 37.5 µM. At 75 µM
hypochlorite, oxidized
2M* was unable to compete for
more than 90% of the binding of nonoxidized
2M*. The
second set of experiments showed that oxidation can abolish the
RAP-sensitive binding of
2M* (Fig. 2B).
Consistent with the results in the first set of experiments, no
significant effect on
2M* binding occurred when the
protein was oxidized by less than 37.5 µM hypochlorite. At concentrations equal to or greater than 37.5 µM,
pronounced inhibition of
2M* binding was observed. The
slight increase in the amount of RAP-insensitive binding when the
protein is modified by high concentration of hypochlorite possibly
reflects an increase in electrostatic interaction between oxidized
2M* and the cell surface.
Binding Competition of Oxidized Native
As a control for the binding of oxidized proteins to
the cell surface, we also determined the ability of oxidized native
2M to compete for the binding of
2M*.
Since native
2M does not bind to either LRP or the
signaling receptor, we expect that oxidized native
2M
also would not bind to either receptor. Fig.
3A shows that oxidation of
native
2M with either 12.5 or 25 µM
hypochlorite resulted in an increased ability of oxidized native
2M to compete for the binding of
2M*,
suggesting that its receptor-recognition site had been exposed.
Oxidation with concentrations of hypochlorite greater than 25 µM resulted in a concentration-dependent
decrease in its ability to compete for
2M* binding (data
not shown). These findings are confirmed in Fig. 3B showing
that native
2M that has been oxidized with hypochlorite
can bind specifically to LRP, since its binding can be competed by
RAP.
Comparison of the Effects of Different Labeling Methods on Oxidized
In Figs. 2 and 3, the
iodinated proteins used in the binding experiments were labeled with
IODO-BEADS®. Since this method can potentially oxidize
2M, we performed additional experiments using
125I-Bolton-Hunter-labeled native
2M and
2M* (125I-BH-
2M,
125I-BH-
2M*, respectively) to verify the
results. Fig. 4A shows that
oxidation of 125I-BH-
2M* at a hypochlorite
concentration greater than 200 µM resulted in the
complete absence of binding to LRP. Fig. 4B shows that
oxidation by as little as 25 µM of hypochlorite can
enhance the binding of 125I-BH-
2M. Maximal
enhancement, however, occur at a hypochlorite concentration of 125 µM. Oxidation appear to effect
125I-BH-
2M similarly compared with
IODO-BEADS®-labeled
2M; however, higher concentrations
of oxidants were necessary to achieve the same results. Given the
greater resistance to oxidation by
125I-BH-
2M as demonstrated by receptor
binding, we compared the susceptibility of Bolton-Hunter or
IODO-BEADS®-labeled
2M to structural damage by
oxidation as a mean of determining which labeling method gave a product
that is more representative of the unlabeled protein. A previous study
has shown that
2M oxidation results in fracturing of the
protein along its dimeric axis (32). Fig. 4C shows the
tetramer to dimer transition of
2M upon oxidation. IODO-BEADS®-labeled
2M appear to be more sensitive to
oxidation compared with unlabeled or Bolton-Hunter labeled
2M confirming the results from receptor binding.
Competition Binding of Oxidized RBF to Macrophages
Since RBF
binds to both LRP and the signaling receptor, we investigated whether
oxidation altered its receptor binding properties. Consistent with the
spectral analyses data, oxidation at a hypochlorite concentration up to
125 µM had no significant effect on the ability of RBF to
compete for the binding of nonoxidized RBF to cell surface receptors
(Fig. 5). Additional calcium signaling
experiments revealed that oxidation had no effect on the ability of RBF
to signal (data not shown).
Intracellular Calcium Signaling Studies of Oxidized Native
Since our cell surface
binding studies showed that oxidation can completely abolish the
RAP-sensitive binding of oxidized 2M* and expose the
receptor-recognition site of native
2M to LRP, we
investigated whether oxidized native
2M and oxidized
2M* also increase intracellular calcium levels by
binding to the
2M* signaling receptor. Fig.
6A shows that oxidation at a concentration of hypochlorite up to 200 µM had no effect
on the ability of
2M* to generate an increase in
[Ca2+]i. The intracellular calcium rose from
base-line levels between 100 and 150 nM to peak levels
between 400 and 500 nM within the first 10 s of cell
exposure to the ligands. Boiled
2M* generated no
increase in [Ca2+]i. To determine whether
oxidized native
2M also bound to the signaling receptor,
we measured the changes in intracellular calcium levels when cells were
treated with native
2M that was oxidized with 50 and 125 µM hypochlorite. These concentrations correspond to the
concentrations that generated the maximal exposure of LRP binding
sites. Fig. 6B shows that no increase in
[Ca2+]i was observed with the addition of either
oxidized or nonoxidized native
2M.
Reduced SDS-PAGE of Trypsin-treated Oxidized Native
To determine the mechanism responsible for the
oxidative exposure of the receptor-recognition site of native
2M for LRP, we performed a SDS-PAGE under reducing
conditions of oxidized native
2M that was treated with
20-fold molar excess of trypsin. We hypothesized that if the exposure
of the receptor-recognition site is associated with the unfolding of
the protein, then oxidized native
2M should be more
susceptible to digestion by trypsin. Fig.
7 shows that as the concentration of
hypochlorite used to modify native
2M increased, fewer
large molecular weight protein bands remained, indicating that the
oxidized protein is digested more efficiently by trypsin. Oxidation at
up to 5 µM hypochlorite appeared to have no effect on the
susceptibility of native
2M toward trypsin. However, at
oxidant concentrations greater than 5 µM, native
2M became more susceptible to tryptic digestion. This is
in agreement with our receptor binding data showing that exposure of
the receptor-recognition site begins only after native
2M has been oxidized by at least an oxidant
concentration of 5 µM. To investigate whether other
mechanisms may account for the oxidative exposure of the
receptor-binding site of native
2M for LRP, we performed
thioester bond titration of oxidized native
2M.
Thioester bond rupture results in exposure of the receptor-recognition
site, analogous to the reaction of methylamine with native
2M. This could complicate our interpretations of these
results. In data not presented, we found no thioester bond rupture by
hypochlorite oxidation. These data are consistent with previous studies
of native
2M oxidation (32).
ELISA of Anti-RBF Antibodies against Oxidized Native
To further probe whether unfolding of the
2M secondary structure may result in exposure of RBF, we
performed an ELISA using polyclonal antisera raised against RBF. The
carboxyl-terminal 20 kDa of
2M is partially exposed in
the native conformation (45), and therefore, it is not unexpected that
it would be partially recognized by a polyclonal antisera to RBF. Upon
exposure of native
2M to proteinase or methylamine, this
region becomes more exposed so that it binds to cell surface receptors
and should, therefore, show greater recognition by antibodies raised
against RBF. Fig. 8 shows that native
2M was partially recognized by the anti-RBF antibodies
whereas RBF and
2M* were both fully recognized. Oxidized native
2M was also fully recognized by anti-RBF
antibodies; however, this occurred only after it was oxidized by at
least 12.5 µM hypochlorite.
Direct Binding of Oxidized Native
Since
2M* has two different Kd and
Bmax values, we performed a
concentration-dependent binding experiment using oxidized
125I-BH-
2M and
125I-BH-
2M* to determine whether the binding
of 125I-oxidized
2M* to the signaling
receptor corresponds with its binding to the high affinity site and
whether the absence of binding by 125I-oxidized native
2M corresponds with its binding to the low affinity
site. Fig. 9A shows the
results from direct binding experiments of 125I-oxidized
2M* to macrophages. The Scatchard analysis (A,
inset) shows that only a single class of high affinity
(Kd ~ 83 pM) binding sites exist with
a Bmax of ~ 5.4 fmol per million cells.
These numbers were verified using Systat® analysis program (Kd
71 ± 12 pM,
Bmax
6.2 ± 2.1 fmol/million cells). Binding to the lower affinity LRP is absent since the addition of
100-fold excess RAP did not alter the Kd or the
Bmax. The binding of 125I-oxidized
native
2M to macrophages is shown in Fig. 9B.
The Scatchard analysis (B, inset) shows a single class of
lower affinity (Kd ~ 0.6 nM) binding
sites with a Bmax of approximately 55 fmol/million cells. Analysis using Systat® shows good agreement with
these numbers (Kd: 0.7 ± 0.15 nM;
Bmax: 57 ± 9 fmol/million cells). This
binding is entirely due to the LRP receptor since the addition of
100-fold excess RAP completely abolished specific ligand binding. The
absence of binding to the high affinity sites by oxidized native
2M together with the signal transduction studies described above provide direct evidence that the high affinity sites
represent the signaling receptor.
In this study we demonstrate that hypochlorite oxidation of native
2M or
2M* can generate exposure of the
receptor binding sites to either LRP or the
2M signaling
receptor, respectively. Oxidation of
2M* by 200 µM hypochlorite completely abolished its binding to LRP
without affecting its ability to bind to the high affinity sites or the
signaling receptor. Oxidation of native
2M by 125 µM hypochlorite resulted in the exposure of the
previously buried receptor-binding site to LRP without exposing the
binding site to the signaling receptor. Oxidation of RBF showed no
decrease in its ability to bind to cell surface receptors, supporting
our earlier work showing that the oxidation-sensitive site in
2M* is outside of the carboxyl-terminal 20-kDa
receptor-binding domain. Studies of the mechanism of oxidative exposure
of the LRP-binding site in native
2M suggested that
protein unfolding may be responsible for this phenomenon. These
experiments provide strong proof for the existence of two distinct
2M* receptors and the presence of two independent
receptor-binding regions on
2M*.
Our earlier studies using cis-DDP modification of
2M* and RBF have shown that the
cis-DDP-sensitive site in
2M* is outside of
the 20-kDa carboxyl terminus and appears identical to the
oxidation-sensitive site (28, 29, 37). Although cis-DDP and
oxidation are capable of modifying similar residues, such modification
caused only a 4-5-fold decrease in the binding affinity of
2M* for LRP. Subsequent studies demonstrate that a
lysine 1370 mutant has decreased binding to LRP, and a lysine 1374 mutant is unable to activate the signaling cascade (26). This has been
the best evidence for the existence of two distinct
2M*
receptors; however, recent work by Nielsen et al. (27)
suggested that lysine 1374 mutants also have decreased binding to
LRP.
To investigate further the identity of the two classes of binding
sites, we searched for ligands that could exclusively bind to either
class of binding sites and tested their abilities to signal.
Hypochlorite is a potent oxidant of native 2M (31, 32).
Treatment of native
2M with 25 µM
hypochlorite resulted in complete destruction of its anti-proteinase
activity. We hypothesized that hypochlorite could also inhibit the
ability of
2M* to bind to cell surface receptors. In
this study, we show that hypochlorite treatment completely eliminated
the RAP-sensitive binding of
2M* to macrophages without
affecting its ability to activate the signal transduction cascade or to
bind to the high affinity cell surface receptors. This confirms and
extends our previous observation that RAP competes for the binding of
2M* to the low affinity sites but is unable to inhibit
the ability of
2M* to signal or to bind to the high
affinity sites (24-26). Since hypochlorite oxidation is also able to
cause the exposure of LRP binding sites in native
2M
without inducing its ability to signal or to bind to the high affinity
sites, our studies provide the best direct evidence to date that the
high affinity sites represent the
2M* signaling
receptors.
The oxidative exposure of the 2M-binding site to LRP but
not to the signaling receptor, is unique in a number of ways. All of
the known naturally occurring
-macroglobulins or recombinantly expressed receptor binding fragments activate the signaling cascade (21, 23, 46). RBF mutant 1374 is the first ligand that does not induce
a signal, yet it still binds to the high affinity site, albeit with
lower affinity. Our hypochlorite oxidized native
2M is
the first ligand produced that is incapable of signaling and binding to
the high affinity sites. The fact that it is still capable of binding
to LRP suggests that the binding site on
2M* for the
signaling receptor is distinct from the LRP binding site. That
hypochlorite oxidation can selectively expose only the LRP binding
sites in native
2M or the signaling receptor binding sites in
2M* demonstrates that the ability of
2M* to bind to its two receptors can be uncoupled.
Efforts are currently being made using oxidized
2M* to
isolate and purify the signaling receptor.
Our investigation of the mechanism that may explain the oxidative
exposure of LRP binding site in native 2M suggests that partial protein unfolding may be responsible. Earlier works by Davies
et al. (47-50) have shown that protein oxidation results in
a partial unfolding of the protein secondary structure, which results
in greater susceptibiliy to intracellular degradation by proteosomes.
Similar finding has been resported by Ossanna et al. (51)
showing that extracellular proteins such as
1-antitrypsin may undergo oxidative inactivation
resulting in partial protein unfolding and greater susceptibility to
proteinase digestion.
It is interesting that the exposure of the LRP binding site is
dependent on the concentration of hypochlorite used to treat 2M and on the labeling method. With IODO-BEADS®
labeling, the amount of hypochlorite needed to generate the exposure of
LRP-binding sites begins with as little as 5 µM and peaks
at 25 µM. This is in marked contrast with
2M that has been labeled with Bolton-Hunter reagent,
which generates LRP binding sites with as little as 25 µM
of hypochlorite but does not peak until 125 µM. At
hypochlorite concentrations greater than 125 µM, oxidized
2M binding to LRP decreased with the concentration of
the oxidant. The results obtained from the two labeling methods raise
important questions regarding the effects of radiolabeling on receptor
binding. Radiolabeling with IODO-BEADS® involves oxidation of
tyrosine residues where as Bolton-Hunter labeling modifies amino
terminus and lysine side chains. It is possible that the Bolton-Hunter
reagent may protect lysine residues from hypochlorite oxidation,
thereby generating a ligand that is more resistant to oxidation. Our
results, however, show that Bolton-Hunter-labeled
2M has
similar susceptibility to oxidation as unlabeled
2M,
whereas IODO-BEADS®-labeled
2M is significantly more
susceptible to oxidation. This suggests that receptor binding studies
with
2M should use the Bolton-Hunter labeling method to
minimize protein oxidation.
The selective exposure of the LRP binding site in oxidized
2M suggests that the two receptor binding regions have
distinct properties. We performed an ELISA using polyclonal antisera
against RBF to determine if unfolding of the oxidized native
2M is associated with an increase in the exposure of
RBF. We found that oxidation of
2M at greater than 12.5 µM hypochlorite results in full recognition of RBF by
polyclonal antibodies. This exposure, however, is not associated with
the ability of the ligand to signal. It is possible that recognition by
the signaling receptor requires a more stringent three-dimensional
conformation in the receptor binding domain of
2M* than
recognition by LRP. This is supported by data showing that residues
important for LRP binding appear to fall within a short consensus
sequence having a predominance of positively charged residues, while
the receptor binding region for the signaling receptor appears to
require participation by residues from an exposed helix and from other
regions of RBF (9, 26, 27). It is also possible that the binding site
to the signaling receptor is exposed by oxidation but quickly
destroyed; however, the fact that binding to the signal receptor is
retained in oxidized
2M* even when the protein is
treated with 200 µM hypochlorite suggests otherwise.
Oxidative inactivation of 2M* receptor binding to LRP
suggests an interesting pathophysiological process that may occur
during inflammation.
2M is ubiquitous in serum and
extracellular fluids (6, 52). During inflammation, neutrophils secrete
hypochlorite and proteinases as a defense mechanism against invading
foreign organisms (40, 51, 53-55). In the presence of oxidants
2M that has reacted with proteinase will lose its
ability to bind to its endocytic receptor (LRP) while retaining its
ability to signal. This may have significant pathophysiological
consequences given that
2M* signaling has been
associated with increased production of prostaglandins and
platelet-activating factor as well as increased mitogenesis in vascular
smooth muscle cells (17-20).
2M that has not reacted
with proteinase will lose its anti-proteinase capacity and the ability
to bind to the signaling receptor. The physiological significance of
these mechanisms is highlighted by the finding that activated
neutrophils can create an environment that contains 124 µM hypochlorite in 2 h (40, 51) and that oxidized
2M* can be isolated from inflammatory lesions in humans
(56). Further investigation of the ability of
2M to
inhibit proteinases, bind to cell surface receptors, and carry
cytokines in the presence of oxidants should provide novel
insights into the biological role of this complex molecule during
inflammation.
We thank Tammy Moser and Drs. George Cianciolo, Hanne Grøn, and Yizhi Liang for their comments on the manuscript and Marie Thomas for her assistance with preparation of the manuscript.