THE JOURNAL OF BIOLOGICAL CHEMISTRY
© 2003 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 278, No. 33, Issue of August 15, pp. 30889 –30895, 2003
Printed in U.S.A.
Molecular Interaction and Enzymatic Activity of Macrophage
Migration Inhibitory Factor with Immunorelevant Peptides*
Received for publication, March 20, 2003, and in revised form, May 7, 2003
Published, JBC Papers in Press, May 9, 2003, DOI 10.1074/jbc.M302854200
Ilaria Potolicchio‡§, Laura Santambrogio§, and Jack L. Strominger‡储¶
From the ‡Department of Molecular Cell Biology, Harvard University, Cambridge, Massachusetts 02138 and
§Department of Cancer Immunology and AIDS, Dana-Farber Cancer Institute, Boston, Massachusetts 02115
Disulfide reduction is an important step in antigen
processing for HLA class II restricted T cell responses.
Migration inhibitory factor (MIF) is a member of the
thioredoxin family and has been classically defined as a
cytokine. Using enzyme-linked immunosorbent assay
and CD analysis, here we describe the binding to MIF of
two peptides, hepatitis B surface antigen (HBsAg) and
insulin B (InsB) with high affinity for HLA class II allotypes, HLA-DP2 and HLA-DQ8, respectively. At neutral
pH, cysteinylated InsB was a substrate for MIF thiol
reductase activity, as assessed by mass spectroscopy/
electrospray analysis. Finally, a biologically active form
of MIF co-immunopurified with mature forms of HLA
DP2/15, and a peptide derived from the HLA-DP ␤1 helix
could be used for affinity purification of MIF. The possibility that MIF participates in class II antigen presentation and/or as a chaperone is discussed.
The macrophage migration inhibitory factor (MIF)1 is a multifunctional protein involved in several inflammatory disorders
(1, 2), such as inflammatory lung diseases (3), septic shock (4),
chronic colitis (5), and some autoimmune diseases, e.g. rheumatoid arthritis (6). MIF is widely expressed in different cell
types but is produced mostly by activated macrophages and
lymphocytes during inflammatory processes (7). The full extent
of its physiological role is not completely defined, although MIF
has been implicated in several biological activities. First, MIF
can function as a pro-inflammatory protein, as assessed by its
inhibitory effect on dexamethasone-mediated TNF-␣ production (2) and its ability to sustain CD3-mediated T cell proliferation (8) and pro-inflammatory cytokine secretion (9). MIF also
possesses enzymatic activities as a tautomerase/isomerase (10)
and thiol oxidoreductase (11). However, the natural substrates
for MIF catalytic activity are unknown even though its enzymatic activity has been characterized in detail using insulin
and L-DOPA as models (11, 12).
X-ray crystallographic analysis of both human and rat recombinant MIF molecules showed a homotrimeric structure
(13, 14). However, MIF is predominantly expressed as the
* This work was supported by National Institutes of Health Grants
AI-49524 and CA-47554 (to J. L. S.) and AI-48832 (to L. S.). The costs of
publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked “advertisement”
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
¶ To whom correspondence may be addressed. Tel.: 617-495-2733;
Fax: 617-496-8351; E-mail: [email protected] or jlstrom@fas.
harvard.edu.
1
The abbreviations used are: MIF, migration inhibitory factor; rMIF,
recombinant MIF; HA, hemagglutinin; HBsAg, hepatitis B surface antigen; InsB, insulin B; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid; LPS, lipopolysaccharide; TNF, tumor necrosis factor; Ab, antibody; Hsp, heat shock protein; ELISA, enzymelinked immunosorbent assay; PBL, peripheral blood leukocytes.
This paper is available on line at http://www.jbc.org
monomer (44%) and dimer (33%), whereas only a smaller fraction (23%) is assembled to form a trimer (15, 16). Native MIF
presumably possesses all three configurations of its recombinant forms. However, the relationship, if any, between configuration and biological activity is still to be defined.
Antigen-presenting cells are able to internalize extracellular
proteins within the acidic endosomal and lysosomal compartment to generate peptides for HLA class II loading. Antigen
processing consists of protein denaturation and fragmentation.
Several proteases have been described to be part of this system
(17). Proteins that contain disulfide bonds require an additional processing step, the reduction of the cysteine residues to
produce antigenic peptides (18). Notably, modification of cysteine residues has been recently shown to regulate T cell responses to HLA class II-restricted epitopes (19 –21). Within the
endosomal pathway the ␥-interferon-inducible lysosomal thiol
reductase, the activity of which is restricted to acidic pH, has
been found associated with antigen processing using reduction
of cysteinylated peptides as the assay (22).
Antigenic peptides can also be generated extracellularly by
secretion of several proteases from professional antigen-presenting cells or stressed cells, such as virally infected and
tumor cells (23–25). Additionally exogenous antigens or already unfolded proteins can bind directly to surface HLA class
II molecule for T cell presentation (26, 27).
Here we describe the MIF peptide binding capacity and,
more importantly, MIF thiol reductase activity at neutral pH
on disulfide bonds in oxidized peptides. Because MIF is a secreted protein, particularly during inflammation, its thiol reductase activity could be associated with peptide modification
in the extracellular milieu. Moreover, because MIF co-precipitated with HLA-DP2/15 molecules, a further role for MIF in
HLA class II antigen presentation is suggested.
EXPERIMENTAL PROCEDURES
Synthetic Peptides—The following peptides were synthesized by Research Genetics (Waltham, MA): residues ␤ 63–77 (KDILEEERAVPDRMA) of HLA-DP2 molecule; residues 306 –318 (PKYVKQNTLKLAT)
of influenza virus hemagglutinin (HA); residues 14 –33 (VLQAGFFLLTRILTIPQSLD) of hepatitis B surface antigen (HBsAg); residues 85–99
(ENPVVHFFKNIVTPR) of myelin basic protein and residues 9 –23
(SHLVEALYLVCGERG) from the insulin B chain (InsB).
Cloning and Bacterial Expression of Human MIF—MIF was amplified from the cDNA of a human B cell line, NP-1, and cloned in an
isopropyl-1-thio-␤-D-galactopyranoside-inducible expression vector
(PGMT-7). PCR primers were designed as previously described with
NdeI (5⬘-GCA TAT ACA TAT GCC GAT GTT CAT CGT A-3⬘) and
BamHI (5⬘-TTG GAT CCT TAG GCG AAG GTG GAG TT-3⬘). MIFPGMT-7 was expressed in Escherichia coli BL21-(DE3) pLysS⫹ cells
(Invitrogen). One liter of bacterial cell culture was grown at 37 °C until
the optical density at 595 nm reached was 0.7–1 absorbance units, and
then the cell culture was incubated with 1 mM isopropyl-1-thio-␤-Dgalactopyranoside for 2 h at 30 °C. The pellet was resuspended in 5 ml/g
BugBaster protein extraction Reagent (Novagen, Madison, WI) for
20 – 45 min in the presence of DNase and protease inhibitors. Cell
30889
30890
Peptides as Substrate of MIF Function
debris were removed by centrifugation, and after filtration through
0.45-␮m filters, the supernatant was applied to Mono Q column (Amersham Biosciences) and an overnight dialysis at 4 °C in 20 mM Tris, pH
7.5, 20 mM NaCl. Because MIF is not retained by Mono Q resin, the
flow-through was collected, and MIF was purified by gel filtration
chromatography (Superdex 200, Pharmacia).
ELISA and CD Spectrum Analysis—ELISA plates were coated overnight with 50 ␮M each peptide and then blocked for 1 h in 2% bovine
serum albumin, phosphate-buffered saline at room temperature. Two
micrograms of recombinant MIF (rMIF) were added and incubated
overnight with the peptides. After washing, bound rMIF was detected
in DuoSet ELISA (R&D, Minneapolis, MN). For CD spectrum analysis
25 ␮M rMIF was incubated overnight at room temperature with or
without 2.5 ␮M each peptide. The incubation was performed in HSB (20
mM Hepes, 0.85% NaCl). The next day samples were diluted 10⫻ in
water, concentrated with MICRON10 (Millipore, Bedford, MA), and
analyzed on a Jasco CD spectrometer. Secondary structure was
analyzed on line by using the k2d program (www.emblheidelberg.de/
⬃andrade/k2d).
Thiol Reductase Activity of MIF—Insulin B, residues 9 –23, was
oxidized as previously described. Briefly, peptide was incubated with
400 ␮M Cys-Hank’s balanced salt solution for 3 h and then dialyzed for
48 h against 10 mM Hepes, pH 7.0. 0.5 ␮g of oxidized peptide was
incubated for 20 min with increasing amounts of rMIF, 100, 200, and
500 ng in 10 ␮l of 0.4 mM Cys-10 mM Hepes buffer, pH 7.0. Samples
were purified by ZipTip (Millipore) and analyzed by mass spectrometry
electrospray. Mass spectrometry electrospray analysis was carried out
at the Chemistry Department at Harvard University.
Immunoaffinity Columns—HLA-DRB1*0401 was purified from 10 g
of B cell line pellet (PRIESS). HLA-DP2/15 was purified from 20 g of a
heterozygous B cell line (NB-1) transformed and previously typed.
Pellets were resuspended in lysis buffer containing 50 mM Tris-HCl, pH
7.5, 0.1 mM NaCl, complete EDTA-free protease inhibitory mixture
tablets (Roche Applied Science), and either 1% Nonidet P-40 or 0.5%
CHAPS. After centrifugation protein extract was filtered through a
0.2-␮m filter and put through a series of columns: POROS(r)20 ALNMS, POROS(r)20 A, POROS(r)20 AL-LB3.1, POROS(r)20 AL-IVD12,
and POROS(r)20 AL-B7/21 as previously described (35). After a protocol
set up in this laboratory, immunoaffinity columns were washed with
solution I (0.1% Nonidet P-40, 50 mM Tris-HCl, pH 8.0), solution II
(0.5% CHAPS, 500 mM NaCl, 50 mM Tris, pH 8.8), and solution III (0.1%
CHAPS, 2 mM Tris-HCl, pH 8.0). Elution was in 0.1% CHAPS, 50 mM
glycine buffer, pH 11.5. Buffer exchange and protein concentration were
performed on Centricon 10. Small scale protein purification was obtained using cell lysate from 5 g of B cell line. Native MIF was isolated
using the monoclonal antibody anti-human MIF (R&D) conjugated to
N-hydroxysuccinimide-activated resin (Amersham Biosciences). Second-step purification was performed using either B7/21 alone or a mix
of HLA class II monoclonal antibodies, B7/21 and IVD12, conjugated to
N-hydroxysuccinimide-activated column. PRIESS and NB-1 were both
analyzed. Proteins were run on 12.5% SDS gels and analyzed by Coomassie Blue staining or Western blot. HLA class II isotypes were
stained with a general anti-HLA class II rabbit serum, generously
provided by Dr. N. Tanigaki. Human MIF was stained using polyclonal
biotinylated antibody specific for human MIF (R&D).
Analysis of the Inhibitory Effect of MIF on Dexamethasone Activity—
MIF function was tested as previously described. Briefly, human monocytes were isolated from whole blood by Ficoll density gradient centrifugation and plated at 5 ⫻ 106 cells/ml in 24-well plates in RPMI, 10%
human AB serum. The monocytes were purified by adherence and
incubated for 1 h with dexamethasone (10⫺9 M) plus 1–10 ng/ml of rMIF
(R&D) or 1 ng of human MIF co-purified with HLA-DP2/15 from B7/21
column. LPS (Sigma) was added at the concentration of 0.5 ␮g/ml.
Sixteen hours of conditioned medium was collected, and TNF-␣ secretion was quantified by ELISA (DuoSet R&D).
Pulse-Chase and Immunoprecipitation—LB and Bg5 B cells lines
were pulsed for 30 min with 1 ␮Ci/ml [35S]methionine (PerkinElmer
Life Sciences) and then chased for 1 and 4 h in complete Dulbecco’s
modified Eagle’s medium supplemented with a 10-fold excess of cold
methionine and cysteine. Cells were lysed in 1% Nonidet P-40 150 mM
NaCl in 50 mM Tris containing a mixture of protease inhibitors (Roche
Applied Science) and equivalent amounts of radioactive post-nuclear
supernatant used for immunoprecipitation after preclearing with protein A- and G-Sepharose beads (Sigma). The immunoprecipitation was
performed using a mouse anti-human MIF mAb (clone 12302.2) (R&D
system). Samples were resolved by SDS-PAGE boiled and non-boiled
under non-reducing conditions.
Peptide Affinity Column—A HiTrap N-hydroxysuccinimide-activated
FIG. 1. Analysis of peptide binding MIF. rMIF was added to
ELISA plates previously coated with 50 ␮M HLA class II naturally
processed peptides. The amount of bound protein was quantified using
a biotinylated polyclonal antibody specific for human MIF followed by
horseradish peroxidase-conjugated-Streptavidin. Absorbance was
measured at 450 nm. The data shown represent a typical experiment
where each sample is assayed in triplicate. The experiment was repeated three times.
1-ml column (Amersham Biosciences) was conjugated with 5 mg of
solubilized peptide following the protocol provided by the company. Five
grams of a transformed human B cell line were lysed in 25 mM Tris-HCl,
0.5 mM NaCl, 0.5% CHAPS, and complete, EDTA-free, protease inhibitor mixture tablets (Roche Applied Science). The affinity column was
loaded overnight at 4 °C with cell lysate and then extensively washed in
lysis buffer. Elution was completed with 100 mM glycine buffer, pH 3.
Fractions corresponding to the elution peak were concentrated on Micron 10 (Millipore) and analyzed by Coomassie staining of 12.5% SDSPAGE. The whole elution was trypsin-digested and sequenced by matrix-assisted laser desorption ionization mass spectroscopy analysis.
Tandem mass spectroscopy analysis was carried out at the Microchemistry Facility at Harvard University.
RESULTS
MIF Peptide Binding Analysis—Macrophage migration inhibitory factor functions as a tautomerase/isomerase and thiol
reductase on small molecule substrates such as L-DOPA and
insulin (11, 12). Pro-1 and Cys-57–Cys-60 form the enzymatic
catalytic sites. Thioreductase activity has an optimal pH ranging between 7.3 and 9 (11). MIF is a soluble protein mostly
secreted extracellularly. Because a number of different molecules with enzymatic properties have been found involved in
extracellular protein fragmentation and processing (23–25) the
possible interaction between MIF and immunologically relevant peptides was examined. Human rMIF was expressed and
purified from E. coli BL21-(DE3) Lys⫹ cells as previously described (28). Correct protein folding assessed by CD spectrum
analysis (see Fig. 2) was identical to that previously published
(29). Several well known high affinity HLA class II peptides
were selected and analyzed for their ability to interact with
rMIF in an ELISA assay (Fig. 1) (30). The highest amount of
rMIF was detected in plates coated with HBsAg14 –33, a peptide
derived from hepatitis B surface antigen, which showed a high
binding affinity for HLA-DP2 (31). A fragment of insulin B
(InsB9 –23) known to bind HLA-DQ8 (32) also showed a significant binding capacity to rMIF, whereas little or no binding was
observed for both myelin basic protein residues 85–99 and
HA306 –318, peptides, known for their high affinity binding to
HLA-DR2 and HLA-DR1, respectively (33). Although ELISA is
a well established method to determine HLA peptide binding, a
better approach to examine protein-protein interaction is analyzing conformational changes by CD spectrum analysis. In
agreement with what has been previously reported, the rMIF
CD spectrum showed a positive ellipticity at 197 nm and broad
negative ellipticity between 205 and 225 nm (Fig. 2) (29). Under the conditions used, the secondary structure obtained with
the K2d program was 37% ␣ helix, 15% ␤ sheet, and 48%
Peptides as Substrate of MIF Function
FIG. 2. CD spectrum analysis of MIF-peptide complexes Far-UV
CD spectroscopy of rMIF (blue) in comparison with rMIF-HBsAg14 –33
complexes (magenta) (a), rMIF-HA306 –318 complexes (green) (b), and
myelin basic protein rMIF-MBP residues 85–99 complexes (red). c,
peptide CD signal.
random coil. Moreover, changes in the MIF CD spectrum were
observed after an overnight incubation with HBsAg14 –33 peptide (Fig. 2a). Secondary structure analysis by K2d revealed a
consistent increase of ␤ sheets and decrease of random coils, 41
and 31%, respectively. No modifications, either in CD spectrum
or in secondary structure, were observed after an overnight
incubation with HA306 –318 or myelin basic protein residues
85–99 peptides (Fig. 2b). At the concentration used all of the
peptides had only a very weak signal that coincided with the
noise level (Fig. 2c). These observations suggest that rMIF has
a flexible conformation, and major modifications of the secondary structure may occur upon complex formation with specific
peptides.
MIF Thiol Reductase Activity—MIF enzymatic activity includes a thiol reductase function (11). Site-directed mutation
analysis pointed to Cys-57 and Cys-60 as part of the MIF
enzymatic catalytic site, determined in vitro using molecular
models (11). Further analysis was conducted here to investi-
30891
gate whether the MIF thiol reductase activity could function on
oxidized peptides. Among the synthetic peptides tested for
their binding to rMIF (Fig. 1), InsB9 –23, molecular mass
1645.8, possesses a cysteine residue in position 11. Cysteinylation of this residue in Cys-Hank’s balanced salt solution
changed the molecular mass to 1765, as shown by mass spectrometry electrospray analysis (Fig. 3, a– b). Peaks at 823 and
at 883 are double-charged and correspond to m/z of reduced and
oxidized peptide, respectively. After 20 min of incubation with
rMIF at pH 7.2 a significant amount of InsB9 –23 was reduced
(Fig. 3c). The ratio peptide reduced/oxidized was dose-dependent as shown in Fig. 3d. Thus, rMIF can function directly to
reduce cysteinylated forms of InsB 9 –23. As expected, no modifications in the molecular mass of the cysteinylated peptide
occurred without adding rMIF (Fig. 3b). Thus, peptides can be
substrates of MIF thiol reductase, and as expected, this reaction occurs at neutral pH
MIF and HLA Class II Association—A possible association
between MIF and HLA class II molecules was also assessed.
Recently a modified methodology based on high pressure immunoaffinity columns that is suitable for rapid large scale
purification of HLA proteins was described (34, 35). This technique was applied to purify HLA-DR (HLA-DRB1*0401) using
mAb LB3.1 and HLA DP (HLA-DPB1*0201/*1501) using mAb
B7/21 affinity columns. The isolated HLA molecules were analyzed on Western blot using a general HLA class II rabbit
antiserum as the detection antibody (Fig. 4a, left). Interestingly, MIF was found associated with HLA-DP2/15 mainly as
the dimer, as detected by Western blot analysis, but not with
the HLA-DR4 preparation (Fig. 4a, right).
MIF has been characterized by its ability to override the
inhibitory effect of dexamethasone on TNF-␣ production from
LPS-stimulated human peripheral blood mononuclear cells (9).
This assay was, therefore, used to assess the biological activity
of MIF found in association with HLA-DP2/15. First, sandwich
ELISA was performed to detect the MIF concentration in the
HLA-DP preparation (Fig. 4b). The HLA-DR preparation was
also analyzed, and as expected, no MIF was found associated.
As shown in Fig. 4c, MIF that had been co-purified with HLADP2/15 as well as purified recombinant MIF partially antagonized the inhibitory effect of dexamethasone on TNF-␣ secretion to an extent similar to that previously reported (9).
Pulse-chase experiments were performed to investigate
whether MIF could be found associated with HLA DP2/15 at
any step of class II trafficking from the endoplasmic reticulum
to the cell surface. HLA class II heterodimers were chased up to
18 h; however, MIF could not be identified after precipitation of
HLA-DP (data not shown). This result was not surprising since,
besides the invariant chain, which associates with all nascent
class II ␣/␤ heterodimers, all other proteins known to pair with
HLA class II molecules (e.g. HLA-DM, HLA-DO (36), tetraspanin (37, 38), and CD1d (39)) only interact with a small
percentage of the class II ␣/␤ complexes. HLA-DP-associated
MIF was detected by first immunopurifying with the mAb
specific for human MIF (Fig. 5a) followed by a second immunopurification step using mAb B7–21 specific for HLA-DP molecules (Fig. 5b). Interestingly, MIF is associated with peptideloaded forms of HLA-DP, as indicated by the presence of an
SDS stable class II ␣/␤ complex. In these experimental conditions, MIF dimers and monomers were purified using both
anti-DP and anti-MIF column (Figs. 4a, 5a, and 6c) and the
dimeric form of MIF became monomeric under reducing conditions (Fig. 5a). In pulse-chase experiments on two different B
cell lines (LB, Bg5), MIF was primarily observed as a monomer.
However, a faint band corresponding to the dimeric form was
also evident (Fig. 5c).
30892
Peptides as Substrate of MIF Function
FIG. 3. MIF thiol reductase activity. Mass spectrometry electrospray analysis of reduced InsB9 –23, m⫹2 823 (a), oxidized InsB9 –23, m⫹2 882
(b), and oxidized InsB 9 –23 after incubation with 200 ng of rMIF (c). d, schematic representation of the dose-dependent reduction of Cys-InsB9 –23
by purified rMIF. The ratio of relative peaks, m⫹2 reduce peptide and m⫹2 cysteinylated peptide, are indicated.
FIG. 4. Analysis of human MIF interaction with HLA class II molecules. a, Western blot analysis of immunoaffinity-purified HLADR4 and HLADP2/15 molecules and Western blot
analysis of MIF co-precipitated with HLA
class II molecules. MIF monomeric and
dimeric forms are indicated. b, quantification of the amount of MIF found in association with HLA-DP2/15 molecules
(filled diamonds) and HLA-DR4 (filled
circle). Protein standards are indicated as
open squares. c, secretion of TNF-␣ in untreated PBLs (Unt, lane 1) and in LPSstimulated PBLs (lane 2). The inhibitory
effect of dexamethasone (Dex) on TNF-␣
secretion in LPS stimulated PBLs in
shown in lane 3. The antagonist effect of
rMIF (1–10 ng/ml) (lane 4) and human
MIF (hMIF, 1 ng/ml) co-isolated with
HLA-DP2/15 (lane 5) on TNF-␣ secretion
in dexamethasone-LPS treated PBLs are
shown. Secretion of TNF-␣ in PBLs is
shown treated with dexamethasone (lane
6), rMIF (lane 7), or human MIF (lane 8)
without LPS stimulation. NB, non-boiled.
Monomeric MIF possess three cysteines, Cys-56, Cys-59, and
Cys-80. Because substitution of Cys-59 led to a significant
reduction of the dimeric and trimeric forms, it is likely that the
same position is involved in intermolecular disulfide bonds, as
suggested by experiments shown in Figs. 4a, 5a, and 6c.
Whether such an S-S bridge is an in vitro artifact or exists
under physiological condition require further investigation.
MIF Can Be Isolated by Column Affinity Using the Synthetic
15-mer DP ␤ 63–77 Found in HLA-DPB1*02—X-Ray crystallo-
graphic analysis of HLA-DR molecules identified a floppy region around the break point of the ␤1 helix important for its
structural flexibility (40). The same region may be involved in
the HLA-DR4/Hsp70 interaction since a peptide spanning residues ␤ 64 –78 has been used to purify Hsp70 (41). Molecular
modeling suggests the presence of the same bulge for HLA-DP2
molecules (42). To investigate whether MIF interacts with this
acidic stretch of HLA-DPB1, a peptide spanning residues ␤
63–77 of HLA-DP1*02 was conjugated with N-hydroxysuccin-
Peptides as Substrate of MIF Function
30893
FIG. 5. Native MIF isolation and analysis of HLA class II molecule association. a, MIF was immunoaffinity-purified and the eluted
protein was analyzed by Western blot. MIF monomeric and dimeric
forms are both visible. b, eluate from the MIF column was re-precipitated with the B7/21 anti-HLA-DP antibody. Samples were analyzed
under non-reducing conditions with anti-HLA Class II rabbit serum. To
ensure specificity, all the monoclonal and polyclonal Abs used in immuno-affinity purification were tested for cross-reactions. c, pulsechase analysis of human MIF in two different B cell lines.
imide-activated Sepharose matrix (Fig. 6a). A column was
loaded with crude proteins extracted from a human B cell line
(see “Experimental Procedures”), and the eluate was analyzed
by separation on SDS-PAGE. Human MIF was the only substance that could be detected by Coomassie Blue staining (Fig.
6a) or by matrix-assisted laser desorption ionization/time of
flight analysis (Fig. 6b). The sample was also analyzed by
Western blot analysis with MIF antiserum confirming the presence of MIF (Fig. 6c).
DISCUSSION
MIF crystal structure analysis showed a homotrimeric structure shaped as a single ring of 15-Å diameter at the open ends,
supposedly the binding site for small molecules (13). However,
crystallographic analysis and NMR are in discordance with the
observation that MIF in physiological conditions exists primarily as a monomer and dimer. Evidence also suggests that dimer
or monomer mediate MIF function (7), although the relation
between structure and functionality is still ill defined.
MIF tautomerase/isomerase activity relies on Pro-1 that promotes enolization and ketonization (10, 43). Because its natural ligand is unknown, MIF enzymatic properties have been
described in detail using molecular models as substrates (44).
Among those that have been tested are the physiological molecules L-DOPA and insulin (45). Because insulin would not fit
within the channel formed by its trimeric assembly, MIF interaction with substrates such as insulin would probably require
a dimeric or monomeric structure. The present report that HLA
class II peptides can also be MIF substrates is a novel
characteristic.
An additional enzymatic activity reported for MIF is its thiol
FIG. 6. Isolation of native MIF using HLA-DP2 ␤ 63–77) peptide
affinity column. Total cell extract from the NB-1 B cell line was
loaded on a peptide affinity column conjugated with a peptide spanning
amino acids ␤ 63–77 from HLA-DP2. Eluted proteins were analyzed by
Coomassie staining of a SDS-PAGE (a), by tandem mass spectroscopy
sequence analysis (b), by Western blot analysis using anti-MIF monoclonal antibody (c).
reductase function (11). The molecular redox center has the
conserved sequence motif of the thioredoxin family consisting
of CXXC (Cys-57–Ala–Leu–Cys-60) (11). MIF thioreductase activity has been well characterized on small proteins. Here, in
addition to defining MIF peptide binding capacity, its thiol
reductase activity is also demonstrated to function on a peptide
substrate as well, cysteinylated (InsB9 –23). Another enzyme of
the thioredoxin family was recently reported, ␥-interferon-inducible lysosomal thiol reductase. ␥-Interferon-inducible lysosomal thiol reductase is constitutively expressed in professional antigen-presenting cells and catalyzes disulfide bond
reactions at low pH (46). Interestingly, ␥-interferon-inducible
lysosomal thiol reductase specifically associates with HLADRw52 (47)2 and has been shown to be very important for HLA
class II peptide presentation. In fact a cysteine-containing peptide (␬I) can be converted from an “inactive” to an “active” form
in inducing T cell proliferation only in the presence of ␥-interferon-inducible lysosomal thiol reductase even though both
peptide forms are similarly loaded on HLA class II molecules
(20). An understanding of the disulfide catalytic activity of MIF
on peptide substrates is a new challenging aspect of its biolog2
P. Cresswell, personal communication.
30894
Peptides as Substrate of MIF Function
ical function. Notably, the reducing property of MIF occurs at
neutral pH, different from ␥-interferon-inducible lysosomal
thiol reductase that functions at lysosomal pH. Thus, either the
extracellular environment or the cytoplasm are the possible
reaction sites of MIF. Several proteases function extracellularly as cathepsin S and L (17, 48), metalloprotease (MMP-3,
MMP-9, and MMP-12) (49, 50), and membrane-associated ectoenzymes (CD13 and CD26) (51). As a result, antigenic determinants can be generated extracellularly for loading onto HLA
proteins. Thus, the peptide-editing activity of MIF on cysteine
residues is likely part of the extracellular processing machinery. Secretion of thioredoxins has been correlated to the generation of a reducing extracellular environment necessary to
sustain T cell activation and proliferation (52). Conversely to
an oxidative milieu, reducing environments would not alter the
oxidative conditions of antigens after their processing. Thus,
the reducing condition required to activate T lymphocytes could
also preserve the extracellular “active” or “inactive” state of
certain peptides during antigen presentation.
Recently, much interest has surrounded the understanding
of the role of soluble “peptide chaperones” during antigen presentation processes. A number of heat shock protein family
members (Hsp70, Hsp90, gp96, Hsp110, gpr170) have been
shown to interact with a broad range of peptides (53, 54).
Historically, Hsp function has been described for class I HLA
loading where peptides are transported from the cytosol to the
endoplasmic reticulum in complex with Hsps (55). However,
some data suggest an intracellular Hsp involvement in HLA
class II presentation as well (55). For example, Hsp70 interacts
with specific peptide fragments derived from the third hypervariable region of HLA-DRB1*0401 and co-precipitates with
HLA-DR4 molecules (56). Similarly, in this report MIF association with mature forms of HLA-DP2/15 molecules is described. All together, these observations, MIF peptide binding
capacity and its interaction with HLA-DP molecules, support
its possible involvement in HLA class II peptide generation and
exchange and, in this sense, its additional function as HLA
class II chaperone. HLA-DP is the least expressed and least
characterized among HLA class II isotypes (34). Its role in
antigen presentation has been established for viral and bacterial antigens and in alloreactivity (57). One of the strongest
HLA class II genetic associations with immunological disorders
is the presence of a glutamic acid at position HLA-DP ␤69,
which confers susceptibility to metal-induced disorders as
chronic beryllium disease and hard metal diseases (58, 59).
MIF association with other HLA class II polymorphic variants
cannot be excluded at the present time and will be a matter of
further investigations.
Because MIF was purified using a peptide affinity column
loaded with a fragment of HLA-DP ␤1 domain-spanning residues 63–77, the same region may also be involved in HLA-DP/
MIF interaction. This region extends around the breaking
point of the ␤1 helix that is known to be particularly sensitive
to peptide-induced structural modifications (60, 61). Furthermore, the positions ␤46 and ␤72 in HLA-DR are anchor residues for Epstein-Barr virus gp42 protein and critical for virus
entry (62). Although MIF has been originally classified as a
cytokine, neither a receptor nor a high affinity membrane molecule has been identified. Its interaction with HLA class II
molecules might represent its mechanism of entry into the cell
similarly to the pan CD91 receptor, responsible for gp97,
Hsp90, Hsp70, and caloreticulin endocytosis (63).
In conclusion, new aspects of MIF biology have been described in this paper; peptide binding capacity, its function as
thiol reductase on peptide substrates, and importantly, its association with HLA-DP molecules. These observations point to
new functional aspects of this small protein and raise interesting questions about the role of MIF during inflammatory
chronic diseases and autoimmunity.
Acknowledgments—We thank Lih Wen Deng for providing purified
HLA-DR4 molecules, Nobuyuki Tanigaki for a general HLA class II
rabbit anti-serum, and Jonathan Boyson for helpful discussion. A special thanks to Lawrence J. Stern and Peter Cresswell for critical reading of the manuscript.
REFERENCES
1. Bernhagen, J., Calandra, T., and Bucala, R. (1998) J. Mol. Med. 76, 151–161
2. Calandra, T., and Bucala, R. (1997) Crit. Rev. Immunol. 17, 77– 88
3. Donnelly, S. C., Haslett, C., Reid, P. T., Grant, I. S., Wallace, W. A., Metz,
C. N., Bruce, L. J., and Bucala, R. (1997) Nat. Med. 3, 320 –323
4. Calandra, T., Echtenacher, B., Roy, D. L., Pugin, J., Metz, C. N., Hultner, L.,
Heumann, D., Mannel, D., Bucala, R., and Glauser, M. P. (2000) Nat. Med.
6, 164 –170
5. de Jong, Y. P., Abadia-Molina, A. C., Satoskar, A. R., Clarke, K., Rietdijk, S. T.,
Faubion, W. A., Mizoguchi, E., Metz, C. N., Alsahli, M., ten Hove, T.,
Keates, A. C., Lubetsky, J. B., Farrell, R. J., Michetti, P., van Deventer,
S. J., Lolis, E., David, J. R., Bhan, A. K., Terhorst, C., and Sahli, M. A.
(2001) Nat. Immunol. 2, 1061–1066
6. Onodera, S., Kaneda, K., Mizue, Y., Koyama, Y., Fujinaga, M., and Nishihira,
J. (2000) J. Biol. Chem. 275, 444 – 450
7. Lue, H., Kleemann, R., Calandra, T., Roger, T., and Bernhagen, J. (2002)
Microbes Infect. 4, 449 – 460
8. Bacher, M., Metz, C. N., Calandra, T., Mayer, K., Chesney, J., Lohoff, M.,
Gemsa, D., Donnelly, T., and Bucala, R. (1996) Proc. Natl. Acad. Sci.
U. S. A. 93, 7849 –7854
9. Calandra, T., Bernhagen, J., Metz, C. N., Spiegel, L. A., Bacher, M., Donnelly,
T., Cerami, A., and Bucala, R. (1995) Nature. 377, 68 –71
10. Rosengren, E., Bucala, R., Aman, P., Jacobsson, L., Odh, G., Metz, C. N., and
Rorsman, H. (1996) Mol. Med. 2, 143–149
11. Kleemann, R., Kapurniotu, A., Frank, R. W., Gessner, A., Mischke, R., Flieger,
O., Juttner, S., Brunner, H., and Bernhagen, J. (1998) J. Mol. Biol. 280,
85–102
12. Taylor, A. B., Johnson, W. H., Jr., Czerwinski, R. M., Li, H. S., Hackert, M. L.,
and Whitman, C. P. (1999) Biochemistry 38, 7444 –7452
13. Sun, H. W., Bernhagen, J., Bucala, R., and Lolis, E. (1996) Proc. Natl. Acad.
Sci. U. S. A. 93, 5191–5196
14. Suzuki, M., Sugimoto, H., Nakagawa, A., Tanaka, I., Nishihira, J., and Sakai,
M. (1996) Nat. Struct. Biol. 3, 259 –266
15. Bendrat, K., Al-Abed, Y., Callaway, D. J., Peng, T., Calandra, T., Metz, C. N.,
and Bucala, R. (1997) Biochemistry 36, 15356 –15362
16. Mischke, R., Kleemann, R., Brunner, H., and Bernhagen, J. (1998) FEBS Lett.
427, 85–90
17. Villadangos, J. A., Bryant, R. A., Deussing, J., Driessen, C., Lennon-Dumenil,
A. M., Riese, R. J., Roth, W., Saftig, P., Shi, G. P., Chapman, H. A., Peters,
C., and Ploegh, H. L. (1999) Immunol. Rev. 172, 109 –120
18. Collins, D. S., Unanue, E. R., and Harding, C. V. (1991) J. Immunol. 147,
4054 – 4059
19. Haque, M. A., Hawes, J. W., and Blum, J. S. (2001) J. Immunol. 166,
4543– 4551
20. Haque, M. A., Li, P., Jackson, S. K., Zarour, H. M., Hawes, J. W., Phan, U. T.,
Maric, M., Cresswell, P., and Blum, J. S. (2002) J. Exp. Med. 195,
1267–1277
21. Jensen, P. E. (1991) J. Exp. Med. 174, 1121–1130
22. Maric, M., Arunachalam, B., Phan, U. T., Dong, C., Garrett, W. S., Cannon,
K. S., Alfonso, C., Karlsson, L., Flavell, R. A., and Cresswell, P. (2001)
Science 294, 1361–1365
23. Santambrogio, L., Sato, A. K., Carven, G. J., Belyanskaya, S. L., Strominger,
J. L., and Stern, L. J. (1999) Proc. Natl. Acad. Sci. U. S. A. 96, 15056 –15061
24. Monji, T., and Pious, D. (1997) J. Immunol. 158, 3155–3164
25. Fox, B. S., Carbone, F. R., Germain, R. N., Paterson, Y., and Schwartz, R. H.
(1988) Nature 331, 538 –540
26. Accapezzato, D., Nisini, R., Paroli, M., Bruno, G., Bonino, F., Houghton, M.,
and Barnaba, V. (1998) J. Immunol. 160, 5262–5266
27. Vacchino, J. F., and McConnell, H. M. (2001) J. Immunol. 166, 6680 – 6685
28. Suzuki, M., Murata, E., Tanaka, I., Nishihira, J., and Sakai, M. (1994) J. Mol.
Biol. 235, 1141–1143
29. Mischke, R., Gessner, A., Kapurniotu, A., Juttner, S., Kleemann, R., Brunner,
H., and Bernhagen, J. (1997) FEBS Lett. 414, 226 –232
30. Chicz, R. M., Urban, R. G., Lane, W. S., Gorga, J. C., Stern, L. J., Vignali, D. A.,
and Strominger, J. L. (1992) Nature. 358, 764 –768
31. Chicz, R. M., Graziano, D. F., Trucco, M., Strominger, J. L., and Gorga, J. C.
(1997) J. Immunol. 159, 4935– 4942
32. Lee, K. H., Wucherpfennig, K. W., and Wiley, D. C. (2001) Nat. Immunol. 2,
501–507
33. O’Sullivan, D., Sidney, J., Appella, E., Walker, L., Phillips, L., Colon, S. M.,
Miles, C., Chesnut, R. W., and Sette, A. (1990) J. Immunol. 145, 1799 –1808
34. Gorga, J. C., Horejsi, V., Johnson, D. R., Raghupathy, R., and Strominger, J. L.
(1987) J. Biol. Chem. 262, 16087–16094
35. Malik, P., and Strominger, J. L. (2000) J. Immunol. Methods 234, 83– 88
36. Brocke, P., Garbi, N., Momburg, F., and Hammerling, G. J. (2002) Curr. Opin.
Immunol. 14, 22–29
37. Engering, A., and Pieters, J. (2001) Int. Immunol. 13, 127–134
38. Kropshofer, H., Spindeldreher, S., Rohn, T. A., Platania, N., Grygar, C.,
Daniel, N., Wolpl, A., Langen, H., Horejsi, V., and Vogt, A. B. (2002) Nat.
Immunol. 3, 61– 68
39. Kang, S. J., and Cresswell, P. (2002) EMBO J. 21, 1650 –1660
Peptides as Substrate of MIF Function
40. Stern, L. J., Brown, J. H., Jardetzky, T. S., Gorga, J. C., Urban, R. G.,
Strominger, J. L., and Wiley, D. C. (1994) Nature 368, 215–221
41. Auger, I., Escola, J. M., Gorvel, J. P., and Roudier, J. (1996) Nat. Med. 2,
306 –310
42. Potolicchio, I., Festucci, A., Hausler, P., and Sorrentino, R. (1999) Eur. J. Immunol. 29, 2140 –2147
43. Lubetsky, J. B., Swope, M., Dealwis, C., Blake, P., and Lolis, E. (1999) Biochemistry 38, 7346 –7354
44. Dios, A., Mitchell, R. A., Aljabari, B., Lubetsky, J., O’Connor, K., Liao, H.,
Senter, P. D., Manogue, K. R., Lolis, E., Metz, C., Bucala, R., Callaway,
D. J., and Al-Abed, Y. (2002) J. Med. Chem. 45, 2410 –2416
45. Zhang, X., and Bucala, R. (1999) Bioorg. Med. Chem. Lett. 9, 3193–3198
46. Phan, U. T., Lackman, R. L., and Cresswell, P. (2002) Proc. Natl. Acad. Sci.
U. S. A. 99, 12298 –12303
47. Arunachalam, B., Pan, M., and Cresswell, P. (1998) J. Immunol. 160,
5797–5806
48. Fiebiger, E., Maehr, R., Villadangos, J., Weber, E., Erickson, A., Bikoff, E.,
Ploegh, H. L., and Lennon-Dumenil, A. M. (2002) J. Exp. Med. 196,
1263–1270
49. Parks, W. C. (2002) J. Clin. Invest. 110, 613– 614
50. McCawley, L. J., and Matrisian, L. M. (2001) Curr. Opin. Cell Biol. 13,
534 –540
51. Ansorge, S., Schon, E., and Kunz, D. (1991) Biomed. Biochim. Acta 50,
799 – 807
30895
52. Angelini, G., Gardella, S., Ardy, M., Ciriolo, M. R., Filomeni, G., Di Trapani,
G., Clarke, F., Sitia, R., and Rubartelli, A. (2002) Proc. Natl. Acad. Sci.
U. S. A. 99, 1491–1496
53. Srivastava, P. K., Menoret, A., Basu, S., Binder, R. J., and McQuade, K. L.
(1998) Immunity 8, 657– 665
54. Wang, X. Y., Kazim, L., Repasky, E. A., and Subjeck, J. R. (2001) J. Immunol.
166, 490 – 497
55. Li, Z., Menoret, A., and Srivastava, P. (2002) Curr. Opin. Immunol. 14, 45–51
56. Auger, I., Lepecuchel, L., and Roudier, J. (2002) Arthritis Rheum. 46, 929 –933
57. Potolicchio, I., Brookes, P. A., Madrigal, A., Lechler, R. I., and Sorrentino, R.
(1996) Transplantation 62, 1347–1352
58. Richeldi, L., Sorrentino, R., and Saltini, C. (1993) Science 262, 242–244
59. Potolicchio, I., Mosconi, G., Forni, A., Nemery, B., Seghizzi, P., and Sorrentino,
R. (1997) Eur. J. Immunol. 27, 2741–2743
60. Santambrogio, L., Sato, A. K., Fischer, F. R., Dorf, M. E., and Stern, L. J.
(1999) Proc. Natl. Acad. Sci. U. S. A. 96, 15050 –15055
61. Chervonsky, A. V., Medzhitov, R. M., Denzin, L. K., Barlow, A. K., Rudensky,
A. Y., and Janeway, C. A., Jr. (1998) Proc. Natl. Acad. Sci. U. S. A. 95,
10094 –10099
62. Mullen, M. M., Haan, K. M., Longnecker, R., and Jardetzky, T. S. (2002) Mol.
Cell 9, 375–385
63. Basu, S., Binder, R. J., Ramalingam, T., and Srivastava, P. K. (2001) Immunity
14, 303–313