| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Department of Pathology, New York University School of Medicine and Kaplan Cancer Center, New York, NY 10016;
Department of Immunology and Medical Genetics, University of Toronto, Toronto, Ontario, Canada;
Boston University Medical Center, Boston, MA 02118;
§
Department of Immunology, University of Toronto, Toronto, Ontario, Canada;
¶
Department of Pathology, McMaster University, Hamilton, Ontario, Canada; and
||
Division of Immunology and Allergy, University Hospital, Lausanne, Switzerland
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2 domain of dimeric IgA is not
involved in high-affinity binding to the T560 pIgR. Inasmuch as this
mouse B cell pIgR binds IgM better than IgA, it is similar to human
pIgR and differs from rat, mouse, and rabbit epithelial cell pIgRs that
bind IgA but not IgM. Possible explanations for this difference are
discussed. All clones of T560 contain some cells that spontaneously
secrete both IgG2a and IgA, but all of the IgA recoverable from the
medium and from cell lysates is monomeric; it cannot be converted to
secretory IgA by T560 cells. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
R) may participate in the
regulation of GALT B cell behavior, but this is difficult to study
because few lymphocytes express them (2, 3, 4, 5, 6, 7, 8). Several years
ago we discovered a GALT-derived murine B lymphoma, T560, that
expressed an IgA receptor (9). Thinking that it might
regulate GALT B cell behavior, we studied its properties only to find
that it differed from the classical FcR in that its binding of IgA
immune complex (IC) was inhibitable not only by IgA but also by IgM,
implying that it was, in reality, an IgA/IgM receptor. IgA/IgM
receptors were already well known as poly-Ig receptors (pIgR),
responsible for transporting both dimeric and polymeric IgA (pIgA) and
IgM (collectively referred to as pIgs) through secretory epithelial
cells into secretions (10, 11, 12, 13). pIgR are not thought to be
importantly expressed on lymphoid cells, but they have been detected on
a T cell hybridoma that also expresses FcR (14), and
the ability of Abs to secretory component (SC), the portion of pIgR
bound to IgA in secretory IgA (S-IgA), to block binding of IgA to
FcR on GALT lymphoid cells of mice has been attributed to
cross-reactivity between the FcR and the pIgR (15).
In this paper, we show that the T560 IgA/IgM receptor is pIgR based on
the following evidence: its binding of IgM is inhibited by pIgA but not
by S-IgA and is J chain dependent as is binding of pIg by human
(16, 17) and mouse (18) pIgRs; it binds IgM
through a different structure from the distinct IgM receptor (FcµR)
of mouse T cells (19); it is precipitable with either IgA
or IgM; its Mr (116 kDa) is consistent
with its being pIgR; and it is recognized by Abs to mouse pIgR on
immunoblots. Furthermore, complete mRNA for pIgR is contained in T560
cells. Studies of the binding properties of mutant mouse IgAs also
presented in this paper indicate that the C2 domain probably does
not mediate interaction between dimeric IgA and the mouse pIgR. Because
a small proportion of T560 cells coexpress IgA with IgG2a (20, 21) and the cell supernatant contains low concentrations of IgA,
we considered the possibility that binding and uptake of autochthonous
IgA might enable T560 to produce S-IgA. However, all of the IgA in both
lysates and secretions of T560 cells proved to be monomeric; none of it
had a high molecular mass consistent with that of S-IgA. T560 evidently
does not convert endogenous IgA into S-IgA.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The T560.2.F7 and CH12.LX B lymphoma and the J774 macrophage cell lines were maintained as previously described (9, 22).
Antibodies
All goat Abs and Ab conjugates were obtained from Southern
Biotechnology Associates (Birmingham, AL). Rat mAbs to the following
four mouse Cµ domains, R33-24, rat IgG2a anti-Cµ1; b7/6, rat
IgG1 anti-Cµ2; IM41, rat IgG1 anti-Cµ3; and C2-23, rat
IgG2a anti-Cµ4 (19), were kindly provided by Dr.
Richard Lynch (Department of Pathology, University of Iowa College of
Medicine, Iowa City, IA). Rabbit anti-recombinant mouse pIgR
reactive with denatured pIgR and SC in Western blots has been described
previously (23). The mouse anti-rat 
-chain mAb
RG7.9.1 (24) was purified from culture supernatants on
protein A beads.
Detection of IgA and IgM binding by flow cytometry
Binding of IgA and IgM for detection by flow cytometry was done
as described in Ref.9 except that binding was at room
temperature rather than 37°C. Various concentrations of IgA, IgM, or
IgA-2,4,6 trinitrophenyl (TNP)-BSA IC (see below) or, for inhibition
assays, mixtures of IgA IC and IgM, IgM and IgA, or IgM and
anti-IgM mAb were used. Maximal binding of IgA or IgM was observed
on T560 cells passed 24 h previously at 2 x
105 cells/ml, incubated at 37°C for 2 h in
fresh medium, then washed three times with Dulbecco’s PBS (DPBS) at
room temperature followed by buffer appropriate to the assay. Binding
of mouse IgM was detected with FITC-goat anti-mouse µ-chain or,
when only MOPC-104E IgM was used, with FITC-goat anti-mouse 
light chain; that of rat IgM was detected with FITC-goat anti-rat
IgM, that of mouse IgA was detected with FITC-goat anti-mouse
-chain, and that of human IgA was detected with FITC-goat
anti-human -chain.
For inhibition studies, the percentage of inhibition was determined from the mean fluorescence (MF) in the various samples using the following formula: % inhibition = 100 (1 - (MF in inhibited sample - MF in low control)/(MF in unihibited sample - MF in low control)).
Immunoglobulins
Human IgA.
Polymeric myeloma IgA (IgA polymer 1) and monomeric myeloma IgA
from a single individual (Steele) were purified in the laboratory of
one of us (B.J.U.). Polymeric myeloma IgA (IgA polymer 2) and
dimeric myeloma IgA from two different individuals, McElhancy and
Latimer, respectively, as well as normal human S-IgA, were kindly given
to us by Dr. J. Mestecky (University of Alabama, Birmingham,
AL).
Mouse IgA.
BALB/c mouse myeloma IgAs, TEPC-15 (T15; IgA) and MOPC-315 (M315;
IgA) were prepared as previously described (9, 21) or
purchased from Sigma-Aldrich. Mutant M315 IgA myeloma proteins were
prepared by cassette mutagenesis, changing Cys301
to Ala, Tyr302 to Phe, and
Cys311 to Ala, and substituting the flexible
human IgA1 hinge or the very short human IgA2 hinge for the mouse hinge
region. Briefly, the pSV2-VH 315 plasmid was
constructed by inserting an XbaI-NaeI fragment of
a cloned VH 315 segment together with an
enhancer-containing Nae-EcoRI fragment of a
JH3-JH4 probe into the XbaI and EcoRI sites of
the multiple cloning site in pSV2NeoM (25). The wild-type
(wt) pSV2-VH-C plasmid was created by
inserting a 1.5-kb XhoI-SalI fragment containing
the BALB/c germline C gene derived from the C30 phage clone
(26) into the corresponding site of
psV2-VH 315. pUC-C30 was made by excising the
1.5 kb XhoI-SalI wt C30 DNA fragment from
pSV2-VH-C and cloning it into a modified pUC18
with no EcoRI and one XhoI site. IgA mutant
constructs were generated from pUC-C30. The region encompassing
residues 297–314 of wt C30 is flanked by EcoRI and
EcoR47III restriction sites, which facilitated insertion of
synthetic oligonucleotides containing the desired mutations. The
mutants were sequenced to ensure a correct reading frame, and the
mutant C was excised from pUC with XhoI-SalI
and inserted into the pSV2-VH 315 plasmid. The wt
and mutant IgA constructs were electroporated into cells of the
light-chain-expressing recipient myeloma line M315.26. Mutant and wt
IgAs in the supernatants of the transfected cells were used to coat
TNP-ox RBC (TNP-ORBC) for rosette assays (9). ELISA and
agglutination titers were used to adjust the IgA concentrations to
approximate those in our standard preparation of M315 IgA myeloma
protein.
Mouse IgM.
MOPC-104E (IgM) and TEPC-183 (IgM) were obtained from
Sigma-Aldrich. TNP-specific normal and mutant IgMs were prepared from
the supernatants of transfected cell lines expressing a TNP-specific
-chain together with a normal or mutant TNP-specific µ-chain as
previously described (27, 28, 29). SP-6 IgM has a wt µ-chain
and consists mainly of polymers with J chain;
J-/IgM-S414 and
J+/IgM-S414 both have µ-chains with Ser
substituted for Cys414, but
J-/IgM-S414 consists mostly of monomers with
some pentamers and tetramers lacking J chain, whereas
J+/IgM-S414 (28) consists of
polymers with J chain. J-/IgM-S575 has
µ-chains with Ser substituted for Cys575 and
consists mainly of monomers with some polymers lacking J chain. These
IgMs were isolated from the supernatants by binding to DNP-Sepharose
followed by elution with 8 mM DNP-glycine. The eluates were dialyzed
against PBS to remove the hapten and concentrated by ultrafiltration. A
total of 200 µl of the concentrated SP-6,
J-/IgM-S414, and
J-/IgM-S575 protein solutions were
ultracentrifuged on sucrose density gradients, essentially as described
(27). Fractions (0.5 ml) collected from the top down were
serially diluted and examined by ELISA, and the concentrations of IgM
were plotted (Fig. 1
). Pools of
"monomer" (fractions 4–8) and "polymer" (fractions 12–20)
were concentrated by ultrafiltration and dialyzed against DPBS to
remove sucrose, and their IgM concentrations were determined in ELISA.
Insufficient SP-6 monomer was obtained for experiments so only the
polymer was used. The J+/IgM-S414 was
concentrated but not subjected to sucrose gradient fractionation. Mouse
IgM was also isolated from supernatants of WEHI-231, which lacks J
chain and makes mainly hexameric IgM, and from two WEHI lines that were
transfected with the J chain gene as previously described
(30). Of these two cell lines, one, 3-H1, remains J chain
negative; the other, 3-C10, is J chain positive. The IgMs were isolated
by binding to Con A beads in the presence of 1 mM
CaCl2, MgCl2, and
MnCl2, and, after washing the beads, eluting the
IgM with 0.2 M methyl--D-mannopyranoside
(Sigma-Aldrich). The eluates were concentrated by ultrafiltration,
dialyzed against DPBS, and the IgM concentrations were determined by
ELISA (see below) and adjusted to working concentrations by dilution
in DPBS.
|
Mouse IgG2a (UPC-10), IgG2b (MOPC-141; Sigma-Aldrich), and
rat IgM (Rockland, Gilbertsville, PA) were taken to have the
concentrations specified by the manufacturer. Normal Sprague Dawley
rat-IgG was prepared by binding to and elution from protein G beads
(Pierce, Rockford, IL). Its concentration was determined from its
OD280.
Other proteins
Mouse whey was kindly given to us by Dr. Michael Lamm (Department of Pathology, Case Western Reserve University, Cleveland, OH).
Phosphatidylinositol-specific phospholipase C (PI-PLC) treatment of T560.2.F7 cells
T560.2.F7 cells were incubated at 37°C with 50 µ/ml PI-PLC from Bacillus cereus (Boehringer Mannheim, Indianapolis, IN) and from Bacillus thuringiensis (ICN Pharmaceuticals, Costa Mesa, CA) in medium or in medium alone for 45 min in the presence or absence of 0.2 ng/ml calphostin C (Calbiochem-Novabiochem, La Jolla, CA), washed three times in medium containing FCS, and resuspended in DPBS for rosette assay as previously described (9).
ELISA
ELISA were performed as described in Ref. 21 using
plates sensitized with goat anti-mouse µ-chain or anti-mouse
or -human -chain in PBS (pH 7.0). The concentrations of the various
mouse IgM and IgA and human pIgA preparations were determined by
interpolation on standard curves. Because polymeric IgM and IgA titrate
in ELISAs with different slopes from their monomeric counterparts, both
the samples and the standards, MOPC-104E for mouse IgM, M315 for mouse
IgA (both from Sigma-Aldrich), were reduced to monomers and alkylated
before titration. Briefly, 0.1-ml volumes of IgM or IgA in PBS
containing 1 mg/ml BSA were mixed with 0.3 ml 0.1 M Tris-HCl buffer (pH
8.6). DTT (0.1 ml, 0.03 M in the same buffer) was added, and the
mixtures were incubated for 1 h at 37°C. After reduction,
iodoacetamide (0.1 ml, 0.07 M in buffer) was added, and the mixtures
were incubated for 20 min at 37°C. Sodium acetate buffer (0.4 ml, 0.1
M) was then added to bring the pH to 
7.0 and the volume to 1.0 ml.
Serial dilutions of the standards and samples were applied to the
sensitized ELISA wells. The human pIgAs had titration curves similar to
that of the standard (human colostral IgA from Sigma-Aldrich) and so
were not reduced, but the monomeric human IgA preparation had a
titration curve of very different slope so its concentration was
calculated simply from its OD. The concentration of rat IgM (Rockland)
was taken as that specified by the manufacturer.
To check that the ability of the rat anti-mouse Cµ domain mAbs to bind to mouse IgM was intact, ELISA wells coated with 10 µg/ml of MOPC-104E IgM were blocked with BSA, and serial dilutions of the mAbs or of normal rat IgG were added to the wells and incubated. After washing, rat IgG bound to the wells was detected with HRP-goat anti-rat IgG Ab. All of the mAbs bound strongly to the IgM-coated wells and binding was not diminished at all by serial dilution down to 1.5 µg/ml. At this concentration, binding of normal rat IgG was at background levels.
Preparation of IgA IC in Ab excess
To assay the ability of IgM to inhibit binding of mouse IgA to the IgA/IgM receptor by flow cytometry, the IgA signal was increased by using IgA IC formed in Ab excess. Dilutions of TNP-BSA (27 TNP groups/mol) were incubated with M315 IgA myeloma protein, and the mixtures were examined by ELISA to establish an amount of TNP-BSA that would barely inhibit binding of 10 µg/ml M315 IgA to plates precoated with TNP-BSA. Half of this concentration was then used together with 10 times the concentration of M315 IgA to prepare IgA IC in such vast Ab excess that no TNP groups could possibly be available to mediate binding of IC to the surface Ig of T560 cells. In practice, equal volumes of BSA-DPBS containing 12.8 µg TNP-BSA/ml and 100 µg M315 IgA/ml were mixed and incubated for 1 h at room temperature. The IC was then diluted 1:2 with diluent or the IgM inhibitor, yielding a final concentration of 25 µg/ml IgA in the mixture.
Periodate-treatment of IgA and IgM
Rat IgM and M315 IgA at 1 mg/ml were treated with an equal volume of 40 mM sodium metaperiodate in 0.6 M sodium acetate buffer (pH 4.0) or with the acetate buffer alone for 2 h at 4°C in the dark (31), blocked with 1% glycine at pH 9.6, and then dialyzed vs DPBS.
Precipitation of 125I-labeled IgA/IgM receptor from the surface of T560.2.F7 cells
To 4 x 107 washed T560.2.F7 or
CH12.LX cells resuspended in 0.4 ml DPBS were added 20 µl DPBS
containing 4 U lactoperoxidase (Sigma-Aldrich) followed by 1.5–2 mCi
carrier-free 125I as NaI (ICN Biomedicals,
Irvine, CA). Ten microliters 0.05% hydrogen peroxide in
H2O were then added five times over the next 30
min, and the mixture was allowed to stand 10 min after the last
addition. The volume was then brought to 10 ml with DPBS, and the cells
were spun down and washed four times with DPBS. Aliquots of 2 x
107 125I-labeled cells in 0.2 ml DPBS were lysed
with 1 ml DPBS-based lysis buffer (pH 6.0) containing 0.5% Nonidet
P-40, 0.01% soybean trypsin inhibitor, 10 µg/ml leupeptin, 1 µg/ml
pepstatin A, 2 µg/ml chymostatin, 2 µg/ml antipain, 100 µg/ml
PMSF, 10 mM benzamidine hydrochloride, 50 mM 
-amino-caproic acid, 20
mM iodoacetamide, and 0.02% sodium azide (32) at 4°C
for 1 h. After centrifugation, the supernatants were transferred
to clean tubes and supplemented with 1/10 volume of 3.5 M NaCl
containing 100 µg/ml PMSF and 3 mg/ml protease-free BSA. The
supernatants were precleared by rotating them at 4°C overnight with
mixtures of 30 µl each of packed plain protein G beads, protein G
beads coated with goat anti-mouse IgM (for CH12 cells), IgA, and
IgG and protein G beads coated with the eventual Ab to be used to bind
IgM or IgA to the beads to precipitate the receptor (i.e., goat
anti-mouse light chain for MOPC-104E or M315, and anti-rat
-chain mAb RG7.9.1 (24) for rat IgM). These beads were
removed by centrifugation, and a second preclearance was performed by
transferring the supernatants into fresh tubes containing 30 µl of
packed MOPC-21 (mouse IgG1)-coated beads, rotating the mixtures at
4°C for 2 h. Equal volumes of the supernatants were then placed
into fresh microfuge tubes containing 30 µl washed protein G beads
precoated with either monoclonal anti-rat -chain followed by rat
IgM or control, BSA-containing buffer, or goat anti-mouse light
chain Ab followed by either M315 dimeric IgA, MOPC-104E IgM, or
control, BSA-containing buffer. The lysates were rotated with the beads
overnight at 4°C. The beads were recovered by centrifugation and
washed twice with 1 ml high-salt lysis buffer, five times with 0.5 ml
of this same buffer, and once with 1 ml low-salt lysis buffer. The
protein was solubilized from the beads by boiling with 50 µl fresh
Laemmli sample buffer (33) with or without 0.7 M 2-ME and
electrophoresed on SDS-PAGE. In some experiments, precipitation was
done in ELISA plates instead of on beads (32). For
preclearance, the wells were coated with goat anti-mouse light
chain or TNP-BSA, washed with PBS, and blocked with BSA. The lysates
were precleared seven times by serial transfer to and incubation at
4°C in such coated wells, then placed in wells coated with goat
anti-mouse light chain, followed by either MOPC-104E IgM,
control BSA-containing buffer, or with TNP-BSA followed by M315 dimeric
IgA or BSA-containing buffer. After adsorption to the coated wells, the
lysates were removed, and the wells were washed seven times with PBS.
Bound proteins were removed from the wells with Laemmli sample buffer
with or without 0.7 M 2-ME and boiled for 5 min before separation on
SDS-PAGE.
Immunoblotting
For immunoblotting, lysates or 10-fold concentrated supernatants
of unlabeled cells were precleared with protein G beads precoated with
goat anti-mouse 
2a chain (T560 cells) or goat anti-mouse
µ-chain Ab (CH12 cells), and the pIgR or IgA present were
precipitated onto protein G beads coated as above with RG7.9.1
anti-rat -chain rat IgM IC, with goat anti-human light
chain human IgA polymer (1) complex, with the Abs alone
with BSA (controls), or with goat anti--chain Ab. For the pIgR,
10% SDS-PAGE was used, and the proteins were electrophoretically
transferred to nitrocellulose membranes (Bio-Rad, Hercules, CA). For
S-IgA, 5% SDS-PAGE was used, and the proteins were transferred to
-probe (Bio-Rad). The membranes were blocked with 10% BSA in PBS
containing 0.05% Tween 20 and 0.02% sodium azide, washed in PBS-Tween
20 without azide, and stained with either rabbit anti-denatured
recombinant pIgR (23) or normal rabbit serum followed by
HRP-conjugated goat anti-rabbit IgG heavy and light chains
(human/mouse adsorbed) or HRP-conjugated goat anti-mouse -chain,
diluted in 5% normal goat serum/PBS-Tween 20. The bands were developed
with ECL Western blotting detection reagents (Amersham, Arlington
Heights, IL).
PCR
PCR were performed using pIgR primers with cDNA templates
prepared from T560.2.F7 cells, from mouse and rat liver, and from
control J774 macrophages and CH12.LX B lymphoma cells, and the PCR
products were TA-cloned and sequenced as previously described
(21). Sequential, overlapping PCRs were done using one
primer designed from the sequence contained within the preceding
amplified fragment and a second primer designed from the mouse liver
pIgR sequence (34) 5' or 3' of the existing fragment.
Combination of the sequences of all these fragments allowed
determination of the complete sequence of the T560 pIgR except for the
5' and 3' ends, which were given by the mouse liver primers. The names
and locations of the primer sequences are given in Table I.
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Rosette assays were previously used to show that T560 cells bind
both mouse IgA- and IgM-coated erythrocytes and that these two Igs are
mutually inhibitory (9). Similar binding and inhibitory
phenomena are here demonstrated by flow cytometry with both mouse and
human IgA and mouse and rat IgM. Binding of polymeric M315 mouse IgA
myeloma protein to T560.2.F7 cells in the presence or absence of
TNP-BSA is shown in Fig. 2A.
IgA binds to T560.2.F7 cells, and the presence of TNP-BSA in Ab excess
markedly enhances the signal from FITC anti--chain Ab,
presumably because there is more IgA to take it up in the bound
complex. Not all of this IgA is necessarily in contact with receptors
on the cell surface, but multipoint attachment to the cell surface
receptors probably increases the avidity of binding of complexes over
that of pIgA alone. No binding to CH12.LX cells, which lack IgA/IgM
receptors, is seen (Fig. 2B).
|
.
More than 50% inhibition of binding of IgA complex containing 25
µg/ml IgA is achieved with as little as 9 µg/ml of rat
IgM.
|
100 µg/ml of IgA polymer (1) and dimer, but
350 µg/ml of IgA polymer (2) and >1000 µg/ml IgA
monomer (Fig. 4). Because the inhibition
curves for polymer and monomer have the same slope, it is possible that
the inhibition by monomer is due to minor contamination by polymer.
|
Binding of IgA and IgM to the IgA/IgM receptor is not due to recognition of a shared carbohydrate
To examine the possibility that IgA and IgM bind to the IgA/IgM
receptor through a shared carbohydrate, the effect of periodate
oxidation of the carbohydrate on binding of rat IgM and M315 IgA was
tested. The data (Table II) indicate that
treatment with periodate has little or no effect on the ability of IgM
to bind to the cells (in the first experiment it reduced it, in the
second it did not) but, in three experiments, it increased the ability
of IgA to do so. It is not clear why it has this effect on IgA but not
IgM. One possibility is that cleavage of sugar rings in the
carbohydrates on IgA renders the IgA susceptible to aggregation,
another is that it relieves a partial blockade of binding of IgA to the
pIgR normally mediated by native carbohydrates. Whatever the
explanation, there being no major reduction in binding of either
protein, it seems unlikely that interaction with the T560 receptor is
mediated through a carbohydrate shared between IgM and IgA.
|
Marked inhibition of binding of murine IgM to the T cell IgM
receptor was shown to occur when either of two different rat mAbs (IM41
and 2911) to the Cµ3 domain were complexed with the IgM at ratios
higher than 5:1 (Ab to IgM) before addition to the cells
(19). Borderline inhibition was seen with one of two mAbs
to the Cµ2 domain and no inhibition was seen with anti-Cµ1 or
with either of two mAbs to the Cµ4 domain. We tested the ability of
some of these same Abs to inhibit binding of MOPC-104E to the T560
IgA/IgM receptor at Ab:IgM ratios ranging from 8:1 down to 1:1. There
was no inhibition of binding with any of the mAbs to any of domains
Cµ1 through Cµ4, most notably not with IM41, which reacts with
Cµ3, indicating that the epitopes bound by the mAbs are not involved
in binding to the T560 IgA/IgM receptor (data not shown). Marked
potentiation of the signal provided by FITC-goat anti-mouse
light chain actually occurred with Abs to Cµ2 and Cµ3 but not with
Abs to Cµ1 and Cµ4, suggesting that the former promoted the
build-up of MOPC-104E-Ab IC on the cell surface.
Role of J chain in the binding of IgM to the IgA/IgM receptor
The best described IgA/IgM receptor to date is the pIgR, normally
expressed on secretory epithelial cells. Expression of the structure
recognized by this receptor is dependent on the presence of J chain in
the IgA or IgM polymer (16, 17, 18). To examine a possible
relationship between the T560 IgA/IgM receptor and the pIgR, binding of
IgMs containing or lacking J chain was investigated. TNP-specific SP-6
mutant and wt IgMs were affinity purified from the supernatants of
transfected cells, separated into monomers and polymers on sucrose
gradients (J-/IgM-S414,
J-/IgM-S575, and wt SP-6) or left unseparated,
(J+/IgM-S414), and their ability to bind to
T560.2.F7 cells was assessed by flow cytometry (Fig. 5, A and B).
|
The role of J chain in binding to the T560.2.F7 IgA/IgM receptor was
further examined by comparing the binding of WEHI-231 IgM with normal
µ-chains but lacking J chain (30) with that of IgM from
two transfected cell lines: one, 3-C10, expressing J chain; the other,
3-H1, not expressing J chain. (Fig. 6).
The results were clear cut. 3-C10 IgM-containing J chain, bound well to
T560.2.F7 cells; WEHI-231 IgM and 3-HI IgM, both lacking J chain,
failed to bind. Because WEHI-231 IgM is substantially hexameric, the
results support the conclusion that it is not sufficient for IgM to be
polymeric; it must contain J chain to bind to the receptor. The results
are consistent with the idea that the T560 IgA/IgM receptor is a form
of pIgR, known to require J chain for binding of polymeric Ig.
|
Radioiodinated IgA/IgM receptor was precipitated with pIg from
T560 cells under several sets of conditions. Both rat and mouse IgM
precipitated it more readily than pIgA. No similar molecule was
precipitated from CH12 cells. In our earliest studies (20)
using the protease inhibitors described in Ref. 14 ,
triethanolamine, iodoacetamide, chymostatin, leupeptin, pepstatin,
antipain, and PMSF, we precipitated small amounts of a molecule similar
in size (36 kDa) to that of the IgA-binding 38-kDa molecule
precipitated from murine T cells (35). However, when we
added soybean trypsin inhibitor, aprotinin, diisopropylfluorophosphate,
and sodium azide to the inhibitors mentioned above (32),
we brought down a band of much higher
Mr (116 kDa; Fig. 7), and the 36-kDa band disappeared. It
seems likely that the 36-kDa band represented an IgM-binding product of
proteolysis derived from the 116-kDa band. The apparent
Mr of the 116-kDa band increased to
about 120 kDa on 7.5% polyacrylamide gels but was unaltered under
reducing vs nonreducing conditions (not shown). The band was also
precipitated by IgA or IgM in PBS in the presence of EDTA indicating
that its binding is not dependent on divalent cation (not shown).
|
The receptor precipitated from Nonidet P-40 lysates of T560 cells
with either rat IgM or human pIgA was detectable as a band of 116
kDa on immunoblots with rabbit Ab to denatured mouse pIgR followed by
HRP-goat Ab to rabbit IgG (Fig. 8). No
such band was precipitated from lysates of CH12 cells, which lack the
IgA/IgM receptor. Control lanes show both free SC (90–95 kDa) and
covalently bound SC in the unreduced S-IgA molecule contained in whey
and of SC alone in reduced whey.
|
The binding characteristics of the IgA/IgM receptor, especially
its requirement for J chain, its high molecular mass, and its ability
to bind anti-murine pIgR Ab suggest that it is pIgR. The
possibility that it might represent some new B cell form of the pIgR,
perhaps with GPI linkage to the cell membrane, as suggested by the
sensitivity of the T560 IgA receptor to PI-PLC (9, 36),
was next explored by RT-PCR. First we demonstrated (Fig. 9) that T560 cells contain message for
the pIgR using primers encompassing 80% of domain I and a small
portion of domain II of the mouse liver pIgR sequence. Identically
sized fragments (337 bp) were amplified from rat and mouse liver and
from both of two separately prepared T560.2.F7 mRNAs but not from mRNA
of either CH12.LX, the B cell line used as control above, or J774, a
receptor-negative macrophage line (9), and not from a
control containing no template. The fragments were TA-cloned and
sequenced; fragments amplified from rat mRNA contained rat sequence
while those amplified from mouse liver and from T560 contained mouse
sequence.
|
. T560 cells clearly contain mRNA
coding for a protein identical with epithelial cell pIgR
throughout its external, transmembrane, and intracytoplasmic regions
except for a single T to C base change at position 476 that results in
a Val
Ala change in domain II. This change, which could well be
allelic, was verified by sequencing the product of an independent
RT-PCR. Two C to T changes, one conservative in domain 4, the other in
the 3' untranslated region, have not been so verified. The sequence
provided no grounds for supposing that the form of pIgR on T560 was
different from that on epithelial cells.
|
We previously noted (9) that treatment of T560 cells
with PI-PLC destroyed their ability to bind IgA-coated erythrocytes,
suggesting that the IgA receptor might be GPI-linked to the cell
membrane. We also found that activation of protein kinase C (PKC) by
PMA caused down-regulation of IgA receptor activity. Subsequently, we
discovered (36) that the effect of PI-PLC was partially
reversed by staurosporine, a protein kinase inhibitor. We now show that
it is completely reversed by calphostin C, which is PKC-specific. As
shown in Table III, treating T560 cells
with PI-PLC reduces their ability to form IgA rosettes by 85%.
Addition of calphostin C diminishes the reduction due to PI-PLC to only
3%. These results suggest that PI-PLC causes loss of IgA receptor
activity (by inference pIgR receptor activity) from T560 cells
indirectly, by activating PKC. PKC activation may follow cleavage by
PI-PLC of a bona fide GPI-linked molecule from the cell surface or
(much less likely) by PI-PLC itself crossing the cell membrane and
cleaving phosphatidylinositol bisphosphate. Whatever the mechanism of
PI-PLC-induced activation of PKC, it is clear that the T560 pIgR is not
GPI-linked to the cell membrane but has conventional type I
transmembrane and cytoplasmic regions consistent with the sequence data
presented above.
|
2 domain or hinge regions on
binding of IgA to the pIgR
Motifs in the C2 domain have been postulated to mediate binding
between pIgA and the pIgR. To explore the roles of C2 domain Cys
residues in the binding of IgA to the pIgR, rosette assays were
performed comparing normal and mutant IgA proteins. The residues
targeted were Cys301, its adjacent
Tyr302, and Cys311. Both
Cys301 and Cys311 are
responsible for inter--chain bonding within the IgA monomer.
In addition, Cys311 of one -chain in an IgA
dimer undergoes a disulfide exchange reaction and binds to a highly
conserved Cys in pIgR domain 5 (Cys467 in human
pIgR) during the formation of S-IgA. At similar concentrations and
agglutination titers, there was no significant difference between
mutant and nonmutant IgAs in their ability to mediate rosette formation
(Table IV). Discrepancies between the
concentrations and agglutination titers probably reflect the ratios of
monomer to polymer in the different supernatants. Because either of two
Cys and a Tyr in the C2 domain can undergo change without impairing
the ability of the dimeric IgA molecule to bind to the pIgR, the C2
domain is probably not directly involved in high-affinity binding to
the pIgR. It also makes no difference to rosette formation whether the
extended, flexible human IgA1 or the very short human IgA2 hinge is
substituted for the mouse hinge region.
|
Because T560 cells can produce IgA and contain mRNA for J chain as
determined by RT-PCR (not shown), we explored the possibility that T560
might make dimeric IgA that could bind to the pIgR and then be
internalized, processed in endosomes, and released as S-IgA. In
immunoblots (Fig. 11), the lanes
containing the supernatants and lysates of T560.2.F7 lanes contain
-chain bands with an Mr of 100
kDa, consistent with -chain dimers (the light chains of mouse IgA
are not covalently bound to heavy chains). No tetrameric or higher
molecular mass -chain bands, such as are seen in the lanes
containing T15 IgA, M315 IgA, and S-IgA (in mouse whey), are
detectable.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
R which may (38) or may not
(39) need the C2 of human IgA1 to be
N-glycosylated for binding. Both polymeric and dimeric human
IgA inhibit binding of IgM to the T560 IgA/IgM receptor, but human
S-IgA does not, suggesting that SC already present in the S-IgA
prevents its interaction with the receptor. Binding of IgM to the T560
IgA/IgM receptor is J chain dependent as are binding of IgM and pIgA to
the epithelial pIgR (16, 17, 18). Binding of IgM to the T560
IgA/IgM receptor does not involve the Cµ3 epitope used in binding to
the mouse T cell FcµR (19). IgA/IgM receptor
precipitated from the T560 cell surface has an appropriately high
Mr (116 kDa), much higher than that of
FcR on either murine T cells (38 kDa; Refs. 14 and
35), human myeloid cells (30 kDa; Ref. 32),
or human B cells (58 kDa; Ref. 40), or of the FcµR on
human B cells (58–60 kDa; Refs. 41 and 42)
or T cells (60 kDa; Ref. 42), and can be detected on
immunoblots with rabbit Ab to murine pIgR. mRNA-encoding pIgR is
expressed within T560 cells, and its sequence indicates that the
molecule has external and transmembrane regions as well as a
cytoplasmic tail identical with that of the molecule expressed on
epithelial cells (34). Of course, because the sequence was
determined from cloned RT-PCR-amplified products, it is possible that
mRNA for a distinct, hypothetical B cell GPI-linked pIgR was also
present, but not detected. This hypothesis is rendered unlikely by the
following two findings: first, the loss of IgA receptor activity from
the cell surface following PI-PLC treatment was shown to be due not to
cleavage of a GPI-linked pIgR from the cell surface but rather to
down-regulation of cell surface pIgR activity following activation of
PKC; second, only one receptor band was detectable on immunoblots. The ease with which wt rat or mouse IgM, as compared with dimeric or pIgA, binds to the pIgR on T560 cells is striking. Only 9 µg/ml rat IgM are needed for 50% inhibition of binding of MOPC-315 IgA IC containing 25 µg IgA/ml to T560 cells, whereas 100 µg/ml of polymeric or dimeric human IgA are needed for 50% inhibition of binding of IgM at 40 µg/ml. The affinity of the T560 pIgR for IgM vs IgA is difficult to compare with that measured in different laboratories because it depends on what species is under consideration and whether the pIgR is free (as SC in solution) or cell bound. For instance, human IgM has a Ka 2–12.5 times that of dimeric IgA in binding to human SC in solution (43, 44) but binds to human pIgR on the surface of transfected Madin-Darby canine kidney epithelial cells with much lower and more nearly equal affinity, though the amount of IgM bound per cell is 3-fold higher (45). The affinity of binding also depends on the source of the pIg; for example, one of us (R.B.C.) has found that while mouse pIgA, like human pIgA and IgM, binds to human SC in an ELISA, mouse IgM does not, implying that differences between human and mouse IgM affect their interaction with human SC; in addition, human pIgA has a lower affinity for human than for rabbit pIgR (46).
Although human pIgR and the T560 mouse pIgR bind both pIgA and IgM, rabbit (44) and rat (47) hepatocyte pIgR bind only pIgA well and do not translocate IgM into bile. Because mouse liver similarly translocates pIgA but not IgM into bile (48, 49), it is generally assumed that mouse hepatocyte pIgR resembles rat and rabbit pIgR and binds IgM poorly or not at all. If this is true, then the difference between the mouse hepatocyte and the T560 pIgR that makes the latter behave more like human pIgR must be explained. Given that the amino acid sequences of the mouse hepatocyte and T560 B cell pIgRs are the same except for the Val to Ala change in domain 2, the difference most likely reflects differential folding or glycosylation of the pIgR, probably the latter. It is easy to imagine that a bulky carbohydrate on hepatocyte-derived pIgR could interfere with IgM but not with IgA binding. Furthermore, it has already been shown that deglycosylation of human SC allows it to inhibit binding of biotinylated native SC to pIgA with 10 times greater efficiency than native SC itself (50), suggesting that some of the carbohydrate moieties on human pIgR may actually impede binding even of pIgA.
Initial, high-affinity, noncovalent binding of pIgA or IgM to the
epithelial cell pIgR involves interaction with its first domain
(50, 51, 52, 53, 54). This first domain of the pIgR contains three
Ig-related complementarity determining region (CDR)1-, CDR2-, and
CDR3-like loops (see Fig. 10
) that may form a binding surface, much as
they would in a conventional Ig (52). The CDR1-like loop
of human pIgR is essential for binding human pIgA (50, 53, 54), but a distinct binding site containing the CDR2-like loop
is critical for binding IgM (55). The amino acid sequences
of mouse and rat pIgR CDR-2 loops are identical but have Asn in place
of the human Glu and rabbit Thr at position 53. However, the amino acid
sequence following this Asn is incompatible with its being a potential
glycosylation site that could be variably glycosylated in mouse
hepatocytes vs B cells (34). If differential glycosylation
is, as suggested above, involved in rendering the pIgR of mouse B cells
more capable of binding IgM than mouse hepatocytes, it must be at some
other site.
Blocking experiments using Fab of Ab to discrete epitopes in human
C2 and C3 suggest that both of these domains are involved in the
interaction between human dimeric IgA and rabbit SC (56),
but recent studies with a panel of IgA1/IgG1 constant region "domain
swap" mutants implicate a peptide in C3 rather than C2 in
binding of human IgA to human pIgR (57). A tentative model
depicting the interaction of SC with dimeric IgA aligns pIgR domain 1
with C2 of one IgA subunit and pIgR domain 5 with C2 of the
second subunit (57). It has been suggested
(50) that initial binding of pIgA to pIgR domain 1 may be
followed by less avid interactions between the remaining pIgR domains
and pIgA that result in conformational changes and closer alignment of
pIgR with pIgA. Late in transcytosis of pIgA-pIgR complexes, a covalent
bond is formed between a highly conserved Cys residue in pIgR domain 5
and Cys311 in the C2 domain in one of the pIgA
heavy chains (58, 59). Obviously, for this to happen these
two domains must somehow be brought close together. That they have an
affinity for one another that contributes to the overall affinity of
binding is not supported, at least for the mouse system, by our data;
it made no significant difference to IgA rosette formation whether
either of the two C2 Cys involved in interchain disulfide bonding
were mutated to Ala. Of particular relevance, mutation of
Cys311, which is actually used in disulfide
bridging to SC, had no effect. This does not, of course, argue against
the idea that a linear epitope of C2 independent of interchain S-S
bonding might be involved in initial binding of dimeric IgA to pIgR
domain 1.
The requirement for J chain for binding of IgM to the pIgR has been used in the present work as a criterion identifying our T560 IgA/IgM receptor as pIgR: two different IgMs proved capable of binding to T560 cells only when they contained J chain. However, it is still not known what function J chain performs in binding to pIgR, and the requirement for it may not be absolute (16, 17, 18) because some pIgAs lacking J chain do bind to pIgR (60, 61). J chain may merely hold the IgA and IgM subunits with their loops in a configuration necessary for their interaction with the pIgR or may itself interact with pIgR. Isolated J chain dimers bind only marginally to SC (16), but Abs to certain J chain epitopes can block binding of intact pIg to SC (62) and prevent pIgR-mediated biliary and epithelial transport of human pIgA (63), suggesting that J chain may participate in binding to pIgR. Only one J chain is present in any given IgA polymer or IgM pentamer (64). Differences in the sizes of the IgA polymers or ratios of IgM pentamer with J chain to hexamer without J chain in the proteins we used might account for the differences in binding observed in our experiments with human polymeric vs dimeric IgA and with MOPC-104E vs TEPC-183 vs the separated polymeric fraction of SP-6 IgM.
The ability of murine IgG2a (UPC-10) to inhibit binding of both pIgA and IgM to the murine pIgR is unexplained. T560 exhibits low-level direct binding of a second mouse IgG2a myeloma protein (UN2S1) in rosette assays (9) so the inhibitory effect may not be restricted to the UPC-10 IgG2a myeloma protein. Inhibition might depend on sharing of one or more binding structures between IgM, IgA, J chain or pIgR, and murine IgG2a. Alternatively, it might depend on some interaction between IgG2a and another molecule on the T560 cell surface that down-regulates the pIgR by, for example, activating PKC (9). Optimal expression of the pIgR on T560 cells is seen when the cells have been cultured in fresh medium before being washed and used in the assay. This thorough washing procedure may allow the pIgR on T560 cells to recover from down-regulation or inhibition by the IgG2a the cells themselves secrete. It was originally instituted because we thought that IgA secreted by T560 (20) might block the IgA/IgM receptor. However, this is unlikely to be the case because all of the IgA contained in both T560 cell lysates and supernatants (no more than 10 ng IgA/ml of supernatant at the end of a 3-day culture) is monomeric. Perhaps T560 cells do not make J chain due to some unknown defect in J chain mRNA (detectable in T560 by RT-PCR). None of the T560 IgA is converted to S-IgA, which is consistent with the fact that monomeric IgA does not bind to the pIgR. We conclude that endogenously produced IgA is not capable of autocrine stimulation of T560 cells through the pIgR.
Activation of PKC down-regulates T560 IgA binding activity (9, 35), which we now know represents pIgR activity. The mechanism of down-regulation is not clear. We previously showed (9) that, after down-regulation by PMA, receptor activity recovered slowly (over several hours) even in the presence of cycloheximide, suggesting that the receptor had not been shed or degraded and that a recycling mechanism might be involved. In Madin-Darby canine kidney cells, activation of PKC by PMA causes transcytosis and apical recycling of transfected pIgR but does not stimulate endocytosis at the basolateral surface; i.e., in these polarized cells, pIgR is transcytosed from the basolateral surface to the apical recycling compartment and then delivered to the apical surface more rapidly in the presence than in the absence of PMA (65). T560 B cells are nonpolar, and nothing is known about the membrane compartment with which their pIgR is associated. Perhaps activation of PKC in T560 cells promotes endocytosis of pIgR rather than an increase in its surface expression.
That T560, a B lymphoma, expresses pIgR at all is a novelty. It raises
the question whether pIgR expression occurs naturally on a small cohort
of normal B cells, of which T560 is a transformed representative, or
whether it is due to some aspect of the transformation process. Support
for the notion that pIgR may be expressed on normal or activated
lymphocytes comes from work showing that the splenocytes of mice
carrying an IgA-secreting myeloma express IgA receptors whose ability
to bind IgA is blocked by rabbit anti-rat SC (15). No
other evidence indicating expression of pIgR by normal lymphoid cells
has appeared, although numerous immunohistochemical studies have shown
binding of anti-SC Abs to human intestinal epithelial cells in
health and disease (65). If binding of anti-SC Abs to
mucosal lymphocytes were common, it should surely have been recorded.
Perhaps expression of the pIgR on the T560 lymphoma initially occurred
as a response to cytokines available during malignant transformation.
T560 is highly activated (22) with heavily mutated Ig
heavy and light chains (21) and originated in the GALT of
an F1 hybrid mouse that had been injected with
parental splenocytes. Although signs of an ongoing graft-vs-host
reaction were not apparent at the time of sacrifice, a graft-vs-host
reaction had probably been initiated and then abrogated. Graft-vs-host
reactions involve release of many cytokines that might up-regulate pIgR
on an activated B cell (64) that was subsequently
transformed. In several human systems, synergy between IFN- and IL-4
in up-regulating pIgR has been reported (66). Whether
these same cytokines would function in the same way with respect to
murine B cells is unknown, but it may be significant that T560 secretes
IL-4 (22). An alternative hypothesis, suggested to us by
Dr. Randy Goldblum (University of Texas Medical Branch, Galveston, TX),
is that the pIgR gene, normally localized to chromosome 1 in both human
(67, 68) and mouse (69), has been switched on
in T560 as a consequence of translocation events that put it under the
control of the promoter of some other constitutively activated gene. We
have not investigated the chromosomes of T560 and do not know the
genetic mechanism of its transformation.
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Julia M. Phillips-Quagliata, Department of Pathology, New York University School of Medicine, MSB 536, 550 First Avenue, New York, NY 10016.
3 Abbreviations used in this paper: GALT, gut-associated lymphoid tissue; FcR, IgA receptor; FcµR, IgM receptor; IC, immune complex; MF, mean fluorescence; pIgR, poly-Ig receptor; pIg, polymeric Ig; pIgA, polymeric IgA; PI-PLC, phosphatidylinositol-specific phospholipase C; SC, secretory component; S-IgA, secretory IgA; TNP, 2,4,6 trinitrophenyl; TNP-ORBC, TNP-ox RBC; wt, wild type; PKC, protein kinase C; CDR, complementarity determining region; RFC, rosette-forming cells; DPBS, Dulbecco’s PBS.
Received for publication February 7, 2000. Accepted for publication June 15, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
) functions of murine B cells. J. Immunol. 144:4571.[Abstract]
cells in vivo and in vitro by purified IgA. J. Immunol. 127:1560.[Abstract]
on murine lymphocytes by IgA in vitro. J. Immunol. 128:888.[Medline]
R+ T cells: role of FcR+ cells for the induction of FcR and formation of IgA-binding factor in Con A-activated cells. J. Immunol. 133:65.[Abstract]
R, ß-galactosyltransferase and anti-secretory component reactive proteins (ASCP). Scand. J. Immunol. 35:469.[Medline]
-chain monotypic and allotypic determinants: reagents for use with rat monoclonal antibodies. Hybridoma 1:257.[Medline]
receptor in humans. J. Exp. Med. 171:597.R protein and development of an anti-FcR antisera. J. Immunol. 145:1745.[Abstract]
receptor interactions. J. Biol. Chem. 273:2260. domains. J. Immunol. 151:1346.[Abstract]
domains of human IgA are involved in in vitro interaction between secretory component and dimeric IgA. Immunol. Lett. 13:221.[Medline]
3 domain motif directs polymeric Ig receptor-mediated secretion. J. Exp. Med. 189:747.This article has been cited by other articles:
|
|
 |
|
  B. Rajan, T. Ramalingam, and T. V. Rajan Critical Role for IgM in Host Protection in Experimental Filarial Infection J. Immunol., August 1, 2005; 175(3): 1827 - 1833. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Pilette, Y. Ouadrhiri, V. Godding, J-P. Vaerman, and Y. Sibille Lung mucosal immunity: immunoglobulin-A revisited Eur. Respir. J., September 1, 2001; 18(3): 571 - 588. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
