Characterization of the Human B Cell RAG-associated Gene,
hBRAG, as a B Cell Receptor Signal-enhancing Glycoprotein
Dimer That Associates with Phosphorylated Proteins in Resting B
Cells*
Laurent K.
Verkoczy
§,
Barbara-anne
Guinn¶, and
Neil L.
Berinstein
**
From the Department of Immunology and the
Department of Medicine, University of Toronto, Toronto M4N 3N5,
Ontario, Canada and the ** Toronto-Sunnybrook Regional Cancer Centre and
Sunnybrook Health Sciences Centre, Toronto M4W 3M5, Ontario,
Canada
Received for publication, March 6, 2000, and in revised form, April 2, 2000
 |
ABSTRACT |
Affinity-purified polyclonal antibodies against
the hBRAG (human B cell
RAG-associated gene) protein were
generated to characterize hBRAG at the biochemical level.
Immunoblotting and immunoprecipitation experiments with these antibody
reagents demonstrate that this protein can be expressed in B cells as a
membrane-integrated glycoprotein disulfide-linked dimer. However, both
glycosylated and unglycosylated isoforms of hBRAG are detectable with
these reagents. Additionally, their use in cell surface biotinylation
and flow cytometry reveals subcellular hBRAG pools both at cell surface
and intracellular locations. Co-immunoprecipitation experiments with
hBRAG antisera detected the association of hBRAG with phosphorylated
proteins in resting B cells, including the protein tyrosine kinase
Hck, which is subsequently dephosphorylated upon B cell
receptor (BCR) ligation. Consistent with its cell surface expression
and possible link to BCR signaling, experiments in which 
-hBRAG
antibodies were used to generate early activation signals suggest a
modest but specific element of tyrosine phosphorylation occurring
through a putative hBRAG receptor. Additional experiments also suggest that hBRAG may be involved in positively enhancing BCR
ligation-mediated early activation events. Collectively, these results
are consistent with a function for hBRAG as a B cell surface signaling
receptor molecule. Coupled with the earlier observation that hBRAG
expression correlates with early and late B cell-specific RAG
expression, we submit that hBRAG may mediate regulatory signals key to
B cell development and/or regulation of B cell-specific RAG expression.
| |
INTRODUCTION |
In a previous report (1), we developed a differential
display-based screening approach to identify and characterize genes that are associated (either directly or inversely) with human recombination activating gene 1 (RAG1)1 mRNA
expression in the context of the B cell lineage. One of the
differential display cDNA products isolated using this molecular approach lead to the molecular cloning of a novel gene we named hBRAG (human B cell
RAG-associated gene). In
characterizing this gene, we showed that it encodes a type II
transmembrane (TM) molecule that is conserved across vertebrate species
and, as suggested by transfection analysis, has a potential role in the
positive regulation of RAG1 mRNA expression, possibly in
a B cell-specific context (2). Furthermore, hBRAG mRNA
was found to be co-expressed with human RAG1 in human B cell
lines, and found at highest levels in B cell enriched tissues.
In general, type II TM proteins function as intracellular enzymes
involved in biosynthesis and post-translational modification pathways
(3). However, among those with immunological function, there are only
two examples of type II TM molecules that function as enzymes, the
intracellular invariant chain (reviewed in Ref. 3) and the cell
surface-expressed BP-1 neutral endopeptidase (reviewed in Ref. 4). The
rest, which include the LY49 lectin family and the TNF receptor family,
function exclusively as cell surface signaling molecules. Based on the
above type II immune response proteins, there are at least two distinct
ways that hBRAG could function in the processes of B cell development
and RAG regulation. One way would be indirectly, as an intracellular
enzyme or chaperone protein involved in transport, synthesis, or
post-translational modification of B cell-specific factors. The other
way would be directly, as a B cell signal-transducing molecule.
However, hBRAG has no homology to any known protein, making it
difficult to ascertain clues as to its modus operandi in the
aforementioned functions.
To begin biochemical characterization of this molecule, we have
generated antiserum to hBRAG and have characterized their specificities
in various contexts. In the present study, we extend structural
predictions that hBRAG is a glycosylated, membrane-integral disulfide-linked dimer, which may associate with other phosphorylated proteins in resting, but not BCR cross-linked B cells. We also provide
evidence suggesting that hBRAG functions as a cell surface receptor
capable of transducing signals alone and, potentially, in conjunction
with the BCR. Overall, these data suggest that, like other
lymphocyte-specific type II molecules, hBRAG may function in
developmental and lineage-specific signaling. Furthermore, like other
receptors, including IL-7, CD19, and the BCR, this protein may function
as a regulator of RAG expression in B cells via intracellular
signaling. Finally, our finding that a large intracellular fraction of
hBRAG exists cannot exclude a separate, additional role for hBRAG as an
intracellular enzyme or chaperone.
| |
EXPERIMENTAL PROCEDURES |
Cell Lines, Human Tissues, and Tissue Culture--
The cell line
panel used for peptide competition assays included the following human
cell lines: REH (pro-B), 697 (pre-B), A8-6P (mature B), HeLa
(fibroblast), U937 (pro-monocytic). These have been described in detail
previously (2). All human cell lines (unless otherwise indicated) were
cultured in RPMI 1640 media (Wisent, Quebec, Canada) supplemented with
2 mM L-glutamine, 100 units/ml penicillin, 100 µg/ml streptomycin, and 15% bovine calf serum (Wisent). Mouse cell
lines were supplemented additionally with 50 µM 2-ME. All
cells for all experiments were incubated at 37 °C and 5%
CO2 and harvested in the log growth phase. Tissue protein
medleys of human heart, placenta, and bone marrow were obtained from
CLONTECH (Palo Alto, CA), human thymus tissue was kindly provided by David Hogg (University of Toronto, Toronto), and
peripheral blood lymphocytes were obtained from the Sunnybrook Health
Sciences Center. For extracting purified human mature B cells, juvenile
tonsils were recovered post tonsillectomy. Tonsillar tissue was
homogenized manually and subsequently strained through a 70-µm nylon
cell strainer (Falcon, Franklin Lakes, NJ). The strained cells in the
flow-through were then centrifuged in Ficoll-Paque (Amersham Pharmacia
Biotech) density gradient. Isolated lymphocytes were washed in RPMI
1640, and B cells were isolated by negative selection using
neuraminidase-treated sheep red blood cells (Cedarlane, Hornby,
Ontario) to rosette T lymphocytes as per the manufacturer's protocol.
All B cell-enriched preparations were analyzed by flow cytometry for
percent B lymphocyte purity based on CD19+ expression.
hBRAG Peptide Synthesis, Polyclonal Antibody Production, and
Affinity Purification--
A peptide (UT952) derived from the most
hydrophilic N-terminal region of the cloned hBRAG protein, consisting
of residues 82-98, was synthesized by Synpep Corporation (Dublin, CA).
The peptide was constructed on a Pharmacia Biolynx automated peptide synthesizer and was purified by reversed-phase high pressure liquid chromatography. For immunizations, 2422 was conjugated to keyhole limpet hematocyanin with gluteraldehyde. Two New Zealand White rabbits
were immunized by initial intramuscular injections of 500 µg of the
peptide-keyhole limpet hematocyanin conjugate emulsified in complete
Freund's adjuvant. Rabbits were boosted with three subsequent
subcutaneous injections of 250 µg of the antigen in incomplete
Freund's adjuvant every 2 weeks before cardiac puncture. Rabbits were
bled 1 week after each boost for serum collection. For affinity
purifications of polyclonal antibodies from serum bleeds, UT952 was
conjugated to BSA with gluteraldehyde, and peptide-conjugated affinity
chromatography columns were prepared by coupling the BSA-conjugated
UT952 peptide to a CNBr-activated Sepharose 4B bead matrix (Amersham
Pharmacia Biotech) according to the supplier's instructions. 1.5 ml of
sera/column was loaded on the column, and after extensive washing of
the columns in 1× TBS, -hBRAG antibody fractions were eluted in
elution buffer (Pierce) and stored at 4 °C.
Pulse/Steady State Labelings and Immunoprecipitations--
For
pulse labeling experiments, 1 × 107 A8-6P cells were
washed in 1× PBS and resuspended in a 10-ml volume of
L-methionine-free RPMI medium and incubated with 10 µCi/ml Easytag L-[35S]methionine (NEN Life
Science Products) for 3 h. Alternatively, for steady state
labeling experiments, 1 × 107 A8-6P cells were
resuspended in 10 ml of normal RPMI medium and incubated with 10 µCi/ml Easytag L-[35S]methionine (NEN Life
Science Products) for 24 h. Cell pellets were solubilized in 6 ml
of radioimmune precipitation assay (RIPA) buffer (1% deoxycholate, 1%
Triton X-100, 0.1% SDS, 0.15M NaCl, 1 mM EDTA, and 10 mM Tris-HCl, pH 7.5) supplemented with the contents of a
complete protease inhibitor mixture tablet (Roche Molecular Biochemicals). Nuclei, cytoskeletal components, and unlysed cells were
removed by centrifugation at 4 °C for 10 min at 14,000 × g. For each immunoprecipitation reaction, lysates were
precleared before immunoprecipitation by incubating 1 ml of cell lysate
with 40 µl of a 1:1 suspension of protein A-Sepharose (Amersham
Pharmacia Biotech) for 1 h at 4 °C. Lysates were then incubated
with gentle mixing overnight at 4 °C with pre-immune sera, crude
terminal antisera, affinity-purified rabbit IgG control, or
affinity-purified -hBRAG antibodies (antibody concentrations used
are detailed in the figure legends). 20 µl of a 1:1 suspension of
protein A-Sepharose:1× RIPA buffer was added, and samples were mixed
for an additional 2 h at 4 °C. Resins were washed twice with
1-ml aliquots of the relevant ice-cold immunoprecipitation buffer, and
immunoprecipitates were solubilized in 20 µl of 2× electrophoresis
sample buffer (New England BioLabs Inc., Beverly, MA) supplemented with
1% 2-ME and 0.042 mM dithiothreitol.
In Vitro Transcription/Translation and Co-translational
Processing Studies--
Transcription reactions were carried out using
0.5 µg of linearized pBluescript 8-3 hBRAG cDNA, the T1
transcription mix (MBI Fermentas Inc., Flamborough, Ontario, Canada),
and either T3 or T7 RNA polymerase (Promega) for 1 h at 37 °C.
The DNA template was removed with 2 units of DNase I for 15 min at
37 °C. For subsequent translation reactions, the transcription
reaction sample, an amino acid mixture without methionine, and
[35S]methionine (NEN Life Science Products) at 1 µCi/µl was added to the Flexi reticulocyte lysate translation mix
(Promega), all on ice. After incubation at 30 °C for 1 h, tRNA
was digested with RNase A. Translation reactions were divided into four
aliquots, each with a final total volume of 20 µl. In particular,
translations were performed either alone or in the presence of 1 µl
(2 units) of canine pancreatic microsomal membranes (MBI Fermentas),
with 1 µl of microsomes and the N-glycosylation
competitive inhibitor peptide Ac-Asn-Tyr-Thr-NH2 (NYT) at
40 µM or with 1 µl of microsomes and digestion with 50 units of peptide N-glycosidase F (PNGase F) for 2 h at
37 °C subsequent to translation. Membrane integration of the
translation products was determined by alkaline extraction of samples
(10 µl) with 100 µl of ice-cold 0.1 M
Na2CO3, pH 11.5, followed by recovery of the
stripped microsomes by centrifugation (16,000 × g for
15 min), direct resuspension in 2× electrophoresis sample buffer (New
England BioLabs) supplemented with 1% 2-ME and 0.042 mM
dithiothreitol, SDS-PAGE analysis, and autoradiography. Alternatively,
prior to SDS-PAGE analysis, immunoprecipitations of the above
translation reactions were performed with either pre-immune sera or
affinity-purified -hBRAG antibodies under the same conditions as
described above for steady state and pulse labeling experiments. For
cleavage of N-linked oligosaccharide groups in cellular
proteins, lysates were solubilized in 1× electrophoresis sample buffer
(New England BioLabs), diluted 1:5 in 0.5 M
NaH2PO4 (pH 7.5) buffer supplemented with 1%
Nonidet P-40, and treated with 100 units of PNGase F at 37 °C for
2 h.
Cell Surface Biotinylation Assay--
Cells were washed with
ice-cold borate buffer (10 mM boric acid, 154 mM NaCl, 7.2 mM KCl, 1.8 mM
CaCl2, pH 9.0). EZ-Link NHS SS-Biotin (Pierce) at 0.8 mM in borate buffer for 15 min at 0 °C was used to
biotinylate the surface. The cells were then rinsed in 0.192 M glycine, 25 mM Tris, pH 8.3, solution to
quench any unreacted reagent. The cells were then lysed with RIPA
buffer (1% deoxycholate, 1% Triton X-100, 0.1% SDS, 0.15M NaCl, 1 mM EDTA, and 10 mM Tris-HCl, pH 7.5). An
aliquot of the lysate was saved for Western blotting. ImmunoPure
(Pierce) immobilized streptavidin (100 µl) was added to the lysate
for 1 h at 0 °C to bind the biotinylated proteins. The
supernatant was removed, and an aliquot was saved for Western blotting.
The streptavidin beads were washed three times with RIPA buffer. 1×
sample buffer (containing 5% 2-
-mercaptoethanol (2-ME)) was added
to the beads, and the samples were boiled for 30 s to cleave the
disulfide bond in the biotinylating reagent and release the captured proteins.
Biotinylation of Antibodies and Intracellular/Surface Flow
Cytometry--
-BRAG affinity-purified, terminal bleed antibodies
were biotinylated with D-biotin (Molecular Probes) using
the method described previously (3). For surface staining of BRAG
expression on the K562 transfectants, aliquots of 106 cells
were incubated on ice for 30 min with 10 µg of -BRAG-biotin antibodies. Cells were washed twice with cold PBS and incubated on ice
for 30 min with 2 µg of streptavidin-fluorescein
isothiocyanate(Molecular Probes). Cells were washed twice with 4 ml of
cold PBS and analyzed using a FACScalibur flow cytometer (Becton
Dickinson, San Jose, CA). For intracellular staining, BRAG cells were
incubated for 30 min on ice with 1 µg of -BRAG-biotin antibodies
in 1% BSA, 0.3% saponin (Sigma Aldrich Canada Ltd.)/PBS. Cells were
washed twice in 2 ml of 0.1% saponin, PBS and resuspended with 1 µg
of streptavidin-fluorescein isothiocyanate in 1% BSA, 0.3%
Saponin/PBS. Cells were washed twice in 2 ml of 0.1% saponin/PBS,
resuspended in 1% BSA/PBS, and analyzed immediately on the FACScalibur.
Cross-linking and Co-immunoprecipitation Assays--
2 × 107 A8-6P cell aliquots/time point were resuspended at
5 × 106/ml of serum-free RPMI medium in 6-well plates
and either left unstimulated or cross-linked with 30 µg/ml of a
polyclonal affinity-purified goat anti-human µ F(ab')2
antibody (Tago Inc., Burlingame, CA) at 37 °C for either 30 s,
1 min, 3 min, or 5 min. Cells were then washed in 9 ml of cold 1× PBS
prior to direct solubilization in 1 ml of 1× sample loading buffer
(New England BioLabs) for subsequent SDS-PAGE and immunoblot analysis
of whole cell lysate fractions. Alternatively, 8 × 107 A8-6P cells unstimulated or cross-linked with 30 µg/ml anti-human µ F(ab')2 for 3 min were prepared for
immunoprecipitation with -hBRAG, -Hck, or -Lyn antibodies
prior to SDS-PAGE and immunoblot analysis. Finally, for co-ligation
studies, 2 × 107 A8-6P cells were ligated with
various combinations of -hBRAG, -human µ, or
-hBRAG + -human µ at various concentrations and for various
durations (see Fig. 8 legend for details). For competitive blocking of
hBRAG-specific cross-linking, hBRAG antibodies were pre-incubated with
10 µg/ml UT952 peptide overnight at room temperature. Cells were then
solubilized in 4 ml of ice-cold mild (co-immunoprecipitation) lysis
buffer comprising 150 mM NaCl, 50 mM Tris-HCl,
pH 7.5, 5 mM EDTA, 5 mM EGTA, 5 mM
NaFl, 1 mM phenylmethylsulfonyl fluoride, 1 mM
sodium orthovanadate, and 1% Nonidet P-40 supplemented with the
contents of a complete protease inhibitor mixture tablet (Roche Molecular Biochemicals). Immunoprecipitations were carried out as
described above for pulse and steady state labeling experiments. For
blocking of antibody Fc receptors, an
Fc
III/II receptor blocking antibody directed against
human CD16/CD32 (Metarex Corp.) was preincubated with A8-6P cells at 10 µg/ml for 1 h at 37 °C prior to cell stimulation with
-hBRAG or IgG + -human µ.
SDS-PAGE--
Whole cell lysates were sedimented by
centrifugation at 14,000 × g at 4 °C for 10 min to
remove cellular debris, sheared with a 21.5-gauge needle to eliminate
chromosomal DNA, and boiled to denature proteins. Whole cell lysates,
immunoprecipitates, or in vitro synthesized proteins were
electrophoretically resolved on 10% discontinuous SDS-PAGE
Tris-glycine pre-cast minigels (Novex, San Diego, CA) at 100 V for
2 h in Tris-glycine running buffer (25 mM Tris, 192 mM glycine, 0.1% SDS) using XCell IITM Mini Cell gel
runner (Novex). Protein loads varied in various experiments, but in all
cases either cell equivalents or 1 mg of proteins were used, and gels
were stained with Commassie Blue. Cell lines and tissue samples were
washed in 1× PBS and solubilized in either non-denaturing buffer (2×
New England BioLabs electrophoresis sample buffer alone, for runs in
denaturing, non-reducing conditions or in reducing buffer (2× New
England BioLabs electrophoresis sample buffer supplemented with 1%
2-ME and 0.042 mM dithiothreitol) for runs in denaturing,
reducing conditions. In each case, the Benchmark molecular pre-stained
standard protein ladder (Life Technologies, Inc.) was run in parallel.
Following SDS-PAGE, radiolabeled proteins (from steady state labeling,
pulse labeling, and in vitro translation experiments) were
directly subjected to autoradiography for 24-48-h periods.
Immunoblotting Conditions and Primary/Secondary Antibodies
Used--
Unlabeled proteins were electrophoretically transferred to
polyvinylidene difluoride (PVDF) membranes (Millipore, Mississauga, Ontario, Canada) using the Xcell IITM blot module (Novex). The membranes were blocked by incubation in 0.25% gelatin, 10%
ethanolamine, and 0.1 M Tris-HCl (pH 9.0) for 2 h at
room temperature. The blocked nitrocellulose strips were then incubated
with various dilutions of primary antibodies for 2 h at room
temperature. The primary antibodies used in immunoblotting include:
4G10, a polyclonal anti-phosphotyrosine antibody (-PTyr) kindly
provided by Dr. Brian Druker (Oregon Health Sciences University,
Portland, OR), monoclonal -Lyn and -Fyn antibodies kindly
provided by Dr. Kathy Siminovitch (Samuel Lunenfeld Research Institute,
Toronto, Ontario), SC-72, a polyclonal -human Hck
antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), 6C8, a
monoclonal hamster -human Bcl2 antibody (PharMingen),
386.12, a monoclonal mouse -human CD19 antibody (PharMingen), mouse
-actin (Amersham Pharmacia Biotech), pre-immune sera from the above
described bleeds, crude antisera to the hBRAG peptide UT952, or
affinity-purified -hBRAG antibodies. Concentrations of antibodies
used vary in the different experiments and are detailed in the figure
legends. To determine specific immunoreactivity of hBRAG antibodies,
i.e. detection of peptide-blockable bands, duplicate blots
in many experiments were run whereby primary antibodies were
preadsorbed with free peptide (at various concentrations indicated in
the figure legends) overnight at room temperature. Bound primary
antibodies were then detected by incubating membranes with either
horseradish peroxidase-conjugated goat -rabbit or goat -mouse IgG
(Bio-Rad) at dilutions of 1:10000 and 1:5000, respectively, for 2 h at room temperature. For blotting of immunoprecipitates with
antibodies against the human proteins Hck and hBRAG, horseradish peroxidase-conjugated protein A was used as a second-step antibody in
immunoblotting to decrease background caused by IgG cross-reactivity. The buffer used for washes and incubations was 0.25% gelatin, 0.05%
Nonidet P-40, 0.015 M NaCl, 5 mM EDTA, and 0.05 Tris-HCl (pH 7.5). After thorough washing, Western blots were developed using the Renaissance enhanced chemiluminescence detection system (NEN
Life Science Products), and blots were exposed for 5-30 s using
BioMaxTM MR autoradiography film (Eastman Kodak, Rochester, NY).
| |
RESULTS |
Characterization of hBRAG Antibody Specificity--
To begin
biochemical characterization of the hBRAG protein, affinity-purified
polyclonal antibodies were generated against a hydrophilic N-terminal
epitope of the hBRAG protein (see "Experimental Procedures"). Prior
to using the -hBRAG antibodies in immunoblotting and
immunoprecipitation assays, specific reactivity of the various hBRAG
antisera bleeds raised against the hBRAG peptide and the corresponding
affinity-purified antibodies were tested in Western blotting on
BSA-peptide conjugate membrane strips (data not shown). Peptide
competition assays using whole cell lysates from the endogenous hBRAG+ line A8-6P were then used to test the binding of
crude sera and corresponding affinity purified antibodies from initial
bleeds on Western blots (Fig.
1A). These assays showed that,
relative to pre-immune serum controls, these antibodies detect two
specific, i.e. peptide-blockable antigens: a ~52-kDa major
band and a minor ~63-kDa band). Based on the size difference of the
minor product, we hypothesized that it may represent an inefficiently
glycosylated isoform of the hBRAG protein. Using the affinity-purified
terminal bleed -hBRAG antibodies, the same ~52- and
~65-kDa-specific peptide-blockable bands found in A8-6P were also
observed in the stable expressing K562 8-3 pCEP4-1 hBRAG transfectant
but not in the mock-transfected K562 control (Fig. 1B).
View larger version (37K):
[in this window]
[in a new window]
|
Fig. 1.
Immunochemical characterization of rabbit
polyclonal antiserum. A, Western blot analysis of
rabbit polyclonal antiserum reactivity to membrane strips prepared from
endogenous hBRAG+ A8-6P cell line lysates. Two specific
immunoreactive products (a major ~52-kDa product and a minor
~63-kDa product; denoted by  ) were detected by hBRAG 13th week
bleed rabbit polyclonal antisera (lot 2422) raised against the hBRAG
N-terminal peptide antigen UT952 or by corresponding affinity-purified
-hBRAG antibodies. Products were considered specifically
immunoreactive if not found in pre-immune sera and blockable in peptide
competition assays, i.e. incubating multiple duplicates of
membrane strips with each antibody after overnight pre-incubation with
increasing concentrations of competing peptide antigen UT952. Membrane
strips were incubated with antibody alone ( ) or with increasing
amounts of UT952 (triangles). Lanes are as
follows: lane 1, incubation with pre-immune serum at a 1:100
dilution; lanes 2 and 3, incubation with
pre-immune serum at a dilution of 1:100 and pre-adsorbed with 100 ng
and 10 µg of UT952, respectively; lane 4, incubation with
crude terminal bleed hBRAG rabbit polyclonal antiserum at a 1:100
dilution; lanes 5-8, incubation with crude hBRAG antiserum
at a 1:100 dilution; lane 9, incubation of 1 µg/ml
affinity purified -hBRAG antibodies alone; and lanes
10-13, incubation of 1 µg/ml affinity-purified -hBRAG
antibodies pre-adsorbed with 10 ng, 100 ng, 1 µg, and 10 µg/ml
UT952, respectively. B, Western blot analysis of membrane
strips prepared from either untransfected K562 lysates (lanes
1 and 2) or from stable hBRAG transfectant K562 8-3 pCEP4-4 lysates (lanes 3 and 4). Detection of the
specific immunoreactive ~52- and 63-kDa bands () was by incubation
with 1 µg/ml affinity-purified -hBRAG antibodies either alone
(lanes 1 and 3) or in the presence of
pre-adsorbed 10 µg/ml of UT952 (lanes 2 and 4).
Note the greater relative intensity of the specific 65-kDa product
relative to the 52-kDa product in K562 transfectant compared with that
seen in endogenous lysates (panels A and B and
Fig. 3; see "Results" for details). In both A and
B, membrane strips all contain equal protein loads (40 µg)
and lysates have been run under denaturing, non-reducing SDS-PAGE
conditions.
|
|
Identification of hBRAG as a Membrane-integrated, Glycosylated
Protein--
To determine whether the larger specific band seen in
Western blots represent the glycosylated isoform of the hBRAG product, two approaches were taken. First, a cell-free system consisting of a
reticulocyte lysate supplemented with microsomal membranes was used to
assay the N-glycosylation status of the hBRAG protein in vitro (Fig. 2A).
We previously demonstrated that the in vitro translated
hBRAG protein (without post-translational modifications) is present as
a doublet of ~55 and 60 kDa, the smaller band possibly due to an
alternative downstream translational initiation site in the hBRAG
sequence N terminus (2). A similar doublet in cell-free translations
was also seen in the current cell-free translations (Fig.
2A, lane 3). N-glycosylation at a
single site will produce a uniform increase of about 2.5 kDa in the
molecular mass of the protein corresponding to the co-translational
attachment of the high mannose oligosaccharide, which only occurs in
the microsomal lumen (4). Consistent with the structural prediction of
4 potential N-glycosylation sites in the hBRAG extracellular domain, the in vitro synthesized hBRAG protein was found to
shift to an ~10-kDa larger species in the presence of microsomal
membranes (Fig. 2A, lane 4). To further extend
the above results, -hBRAG antibodies were used to immunoprecipitate
specifically in vitro synthesized hBRAG proteins, either
alone or in the presence of microsomes. It was shown that the hBRAG
affinity-purified antibodies detect the full-length hBRAG protein
doublet expressed after cell-free in vitro translation (Fig.
2A, lane 7) and also recognize the post-translationally modified in vitro translated membrane
fraction protein (Fig. 2A, lane 8).
View larger version (36K):
[in this window]
[in a new window]
|
Fig. 2.
N-glycosylation of
in vitro translated hBRAG cDNA and of endogenous
and transfected hBRAG products in cellular lysates. A,
N-glycosylation of the in vitro translated hBRAG
cDNA and recognition of glycosylated and unglycosylated hBRAG
isoforms by -hBRAG antibodies. Autoradiogram of the
[35S]methionine-labeled cell-free translated hBRAG gene
products either directly resolved by SDS-PAGE analysis (lanes
1-5) or immunoprecipitated with affinity-purified -hBRAG
antibodies from terminal bleed antiserum raised against the N-terminal
hBRAG peptide prior to SDS-PAGE UT952 (2422; lanes 6-9).
The open arrow shows the unglycosylated hBRAG doublet
product, and the closed arrow shows the ~70-kDa
glycosylated product (see "Results" for details of relative amounts
of glycosylated and unglycosylated products in whole and
immunoprecipitated cell-free fractions). Numbers to the
left indicate the migration positions of the protein
standards. Lane 1 represents a negative control in which no
cDNA was added. Translation reactions were carried out without
microsomes (lanes 2 and 6), with microsomes only
(lanes 3 and 7), with microsomes and the
competitive inhibitor NYT (lanes 4 and 8), or
with microsomes only and digestion with PNGase F after translation
(lanes 5 and 9). S, indicates the
linearized pBluescript 8-3 hBRAG that was transcribed in
vitro, using the T3 promoter, to generate a sense RNA;
AS, indicates in vitro transcription of 8-3 hBRAG
with the T7 promoter to generate an antisense RNA. B,
PNGase F treatment of endogenous hBRAG-expressing A8-6P cellular
lysates or transfected/untransfected non-endogenous hBRAG-expressing
K562 cellular lysates. Lysates were resolved on SDS-PAGE under reducing
conditions as described for Fig. 1 and immunoblotted using
affinity-purified N-terminal -hBRAG antibody 2422 at 1:1000
dilutions. The + signs (lanes 2, 4,
and 6) represent samples treated with 5000 units of
PNGase F for 1 h at 37 °C; signs (lanes
1, 3, and 5) denote untreated samples. The
putative glycosylated and unglycosylated isoforms (~65 and 52 kDa,
respectively) are shown as open and closed
arrows, respectively. Note that consistent with the structural
prediction of four N-glycosylation sites, the difference in
size between the glycosylated products in lysates and in cell-free
translations is ~10 kDa.
|
|
Two independent experiments were performed in parallel to confirm that
the shift in mobility was specifically due to
N-glycosylation rather than other possible
post-translational modifications. We used the
N-glycosylation cleavage-specific enzyme,
N-glycosidase F, and in addition used a competitive peptide
inhibitor of N-glycosylation, the acceptor peptide NYT. With
either the addition of NYT or treatment with N-glycosidase
F, the N-glycosylation-mediated shift seen in the presence
of microsomes was not observed (Fig. 2A, lanes 4,
5). Furthermore, in vitro translated products
immunoprecipitated with -hBRAG antibodies also demonstrated a
specific loss of the glycosylated hBRAG isoform from the microsomal
fraction, either with NYT pre-incubation or with
N-glycosidase F post-treatment (Fig. 2A,
lanes 9, 10). Taken together, the results suggest that hBRAG
is a membrane-integrated glycoprotein, as has been predicted based upon
the amino acid sequence.
Based on densitometric analysis, the upper (~60 kDa) band of the
unglycosylated hBRAG cell-free translated product doublet is 2.1- and
2.4-fold more intense than the bottom (~55 kDa) band prior to the
addition of microsomes in whole and immunoprecipitated fractions (Fig.
2A, lanes 2 and 6, respectively). Upon
the addition of microsomes, however, the relative amounts of the upper
and lower bands are roughly equivalent (Fig. 2A, lanes
3 and 7), but the upper band returns to increased
amounts relative to the lower band upon either PNGase F or NYT
treatment (Fig. 2A, lanes 4, 5, 8, and
9), suggesting that the larger species is selectively glycosylated. It is unlikely that this selective glycosylation is a
result of alternative translational initiation codon usage, because the
potential alternative translation initiation site with a Kozak-like
sequence is ~40 amino acids C-terminal to the first site and
therefore also encodes a protein with four N-glycosylation sites in its extracellular region. One possible explanation for this
observation is that hBRAG, at least in vitro, may be capable of inserting itself in the membrane in two orientations, either as a
type II TM protein or, in the inverted orientation, as a type I, and as
such perhaps only in one orientation is it glycosylated. The insertion
event itself may depend on the cytoplasmic region and its relative
length, such that a shorter cytoplasmic tail may allow improper
insertion of the protein as a type I rather than type II TM protein,
and the extracellular domain with N-glycosylation sites is
now the cytoplasmic domain and will not be accessible to
N-glycosylation. In this context, alternative initiation
site usage may therefore indirectly determine glycosylation status.
A second test for hBRAG glycosylated and unglycosylated isoforms was to
assess directly the glycosylation status of either the endogenous hBRAG
or transfected cellular hBRAG protein. This assessment was accomplished
by treating the lysates with PNGase F prior to immunoblot analysis
with N-terminal-specific -hBRAG antibodies (Fig. 2B). As
expected, the 52- and 63-kDa bands were seen with immunoblot analysis
of A8-6P and K562 8-3 pCEP4-4 lysates. A nonspecific band at 50 kDa was
seen in all lanes. After PNGase F treatment, the minor, 63-kDa product
is selectively lost, and relative to the untreated sample, there is
more 52-kDa product in both the endogenously expressing cell line A8-6P
and in the hBRAG transfectant K562 8-3 pCEP4-4 (Fig. 2B,
lanes 2 and 6). The low level of glycosylated
hBRAG expression seen in these lanes in Fig. 1, A and
B (seen only with longer exposures), and with affinity-purified but not crude antisera, is consistent with the relatively low percentage of glycosylated product (~25% as assessed by densitometric analysis of glycosylated/unglycosylated product ratios) seen in immunoprecipitates of cell-free translations (Fig. 2,
lane 7). There are two possible explanations for this
finding: either the actual glycosylation of this protein is not
efficient and there may exist a large pool of unglycosylated hBRAG, or
alternatively, this particular -hBRAG antibody is inefficient at
recognizing a possibly different glycosylated conformation. The latter
possibility is consistent with the similar intensity of glycosylated
product in -hBRAG immunoprecipitations and whole cell lysates (Fig.
2A, compare lanes 3 and 7).
Interestingly, relative to peptide competitions with the endogenously
expressing hBRAG cell line A8-6P, a higher glycosylated:unglycosylated
hBRAG ratio is reproducibly seen in stable hBRAG transfectant K562 8-3 pCEP4-4 lysates, indicating that either the transfected hBRAG is more
efficiently recognized by antibody or is more efficiently synthesized
by cellular machinery. It is also formally possible that the weak
glycosylation form seen in endogenous lysates could be the real
antigen, whereas the more ubiquitous 52-kDa antigen is not the
unglycosylated antigen but a cross-reacting, more highly abundant
protein. However, this latter possibility is not consistent with the
in vitro translation data and the shift in size seen after
PNGase F treatment of A8-6P and K562 8-3 pCEP4-4 lysates.
Nevertheless, the fact that the lower molecular weight hBRAG isoform
appears to be differently post-translationally modified could
potentially mean that it serves a different function and/or does not
have a motif that is present in the longer protein, which alters its
cellular distribution, i.e. intracellular versus
cell surface expression.
Analysis of hBRAG Protein Expression in Lymphoid Cell Lines and
Tissues and Identification of hBRAG Disulfide-linked Multimers--
To
determine whether hBRAG could form higher order structures, lysates
from various cell lines and tissues were prepared and run under
denaturing reducing and non-reducing conditions prior to immunoblotting
with -hBRAG antibodies, either in the presence or absence of
blocking peptide (to assess the specificity of products resolved under
both conditions). Nonspecific cross-reactive bands were seen at 50 and
60 kDa in all cell lines. As expected, under reducing conditions, the
specific 52-kDa major antigen was expressed in B cell lines but not in
the non-lymphoid line HeLa or the myeloid cell line U937 (Fig.
3A). This band was also lost
by the addition of competing peptide (+). Under the particular
conditions of this assay and the relatively high background and short
exposure times, the glycosylated product was not detected.
Nevertheless, under non-reducing conditions, in addition to resolving
the single polypeptide hBRAG antigen, a larger disulfide-linked
specific molecule of ~120 kDa could also be resolved in B cells.
Based on its molecular mass, this molecule is not likely to represent
an unglycosylated hBRAG homodimer but is probably either a glycosylated
hBRAG homodimer or an unglycosylated/glycosylated heterodimer. A
predominant ~80-kDa band was also seen in the pre-B cell line 697 under reducing conditions, but it is likely not a hBRAG heterodimer
because it is not competed out by peptide and more likely represents a
697-specific cross-reacting antigen. In tissues, more nonspecific
cross-reactivity, i.e. non-peptide-blockable bands, was
seen, but both the highly expressed 52-kDa- and the lower-expressed
63-kDa-specific single polypeptide products seen previously under
denaturing conditions in immunoblots of B cell line lysates were seen
in bone marrow and peripheral blood lymphocytes but not in thymus,
heart, or placenta unfractionated tissues. Under non-denaturing
conditions, the same specific products were seen in the same
tissues, but additional, larger dimer and possibly trimer
disulfide-linked molecules of ~120, 150, and 180 kDa (representing either hetero/homo glycosylated hBRAG dimers and/or
unglycosylated/glycosylated hBRAG hetero/homotrimers) were also
seen.
View larger version (89K):
[in this window]
[in a new window]
|
Fig. 3.
Western blot analysis of various cell lines
and tissues with -hBRAG antibodies reveals
hBRAG dimer formation. A, Western blot analysis of
protein lysates prepared from 2 × 107 cell
equivalents of various cell lines fractionated by SDS-PAGE under either
non-reducing (lanes 1-8) or reducing conditions
(lanes 9-18). Cells include the human pro-B, pre-B, and
mature B cell lines REH, 697, and A8-6P, respectively, the human
fibroblast cell line HeLa, and myelocytic human cell line U937.
B, Western blot analysis of 1 mg of protein lysates from
various unfractionated human tissues resolved by SDS-PAGE under either
non-reducing (lanes 1-10) or reducing conditions
(lanes 11-20). In both A and B, after
proteins were fractionated by SDS-PAGE, they were transferred to a PVDF
membrane, and affinity-purified antibodies from terminal bleeds of
crude hBRAG antiserum raised against N-terminal peptide UT952 at a
concentration of 1 µg/ml either alone () or in the presence of
excess (500 µg/ml) competing peptide UT952 (+) were used to stain the
blots, which was followed by detection with a chemiluminescence
detection system. The positions of molecular mass standards (expressed
in kilodaltons) are indicated on the left. The positions of
the specifically immunoreactive bands that are competed out by
N-terminal peptide are indicated by arrows. PBL,
peripheral blood lymphocytes.
|
|
To extend the above immunoblotting results of gels run under
non-denaturing conditions and to compare hBRAG isoforms during biosynthesis (prior to, and after post-translational modification), endogenous hBRAG-expressing A8-6P B cells were either metabolically pulse or steady state-labeled, immunoprecipitated with -hBRAG antibodies, and run under non-reducing conditions (Fig.
4). As in reduced gel immunoblots,
larger bands, likely representing multiple hBRAG higher order isoforms
(potential homo- or heterodimers), were seen in both steady state and
pulse-labeled A8-6P immunoprecipitates (Fig. 4). Although the specific
product sizes are similar between steady state-labeled
immunoprecipitates (Fig. 4B) and whole cell lysate products
in Western blot assays (Fig. 3), they are slightly larger in
immunoprecipitations of pulse-labeled cellular proteins (Fig.
4A). This variation cannot be due to alternative splicing because only one transcript is seen in all of the cell lines assessed (2). However, one possibility is that hBRAG may be a highly unstable
protein that is particularly susceptible to proteolysis, such that the
antisera are detecting various hBRAG degradation isoforms depending on
the conditions of the assay. Overall, the results of these assays are
consistent with the structural prediction of hBRAG being expressed in B
cells at least partly as a disulfide-linked dimer, similar to other
lymphocyte-specific cell surface-expressed type II transmembrane
receptors.
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 4.
Analysis of hBRAG biosynthesis as assessed by
immunoprecipitations of metabolic steady state and
pulse-labeled A8-6P whole cell lysates. A,
autoradiogram of SDS-PAGE fractionated lysates from 2 × 106 A8-6P cells that were pulse-labeled with
[35S]methionine and immunoprecipitated with pre-immune
serum at 1:200 and 1:40 dilutions (lanes 1 and 2,
respectively); 1:1000, 1:200, 1:40, and 1:10 dilutions of crude
N-terminal hBRAG antiserum (lanes 3-6, respectively); or a
1:10 dilution of crude N-terminal hBRAG antiserum pre-incubated with
excess (500 µg/ml) blocking peptide UT952 (lane 7).
B, autoradiogram of 1 × 106
A8-6P cells that were steady state-labeled with
[35S]methionine and immunoprecipitated with pre-immune
serum at a 1:10 dilution (lane 2); crude N-terminal hBRAG
antiserum at a 1:10 dilution without or with excess (500 µg/ml)
blocking peptide UT952 (lanes 3 and 4,
respectively); or affinity-purified -hBRAG antibodies at a 1:100
dilution with or without excess (500 µg/ml) blocking peptide UT952
(lanes 5 and 6, respectively). Lane 1 represents the precleared labeled lysate fraction. For both
A and B, the positions of molecular mass
standards (expressed in kilodaltons) are indicated on the
left, and arrows to the right identify
the position of specifically immunoreactive bands.
|
|
hBRAG Intracellular and Surface Isoforms--
To assess whether
hBRAG exists as a cell surface protein and/or as an intracellular
protein, we performed cell surface biotinylation assays on the above
transfectants and on purified human tonsillar B cells (see
"Experimental Procedures," Fig.
5A). Consistent with the flow
cytometry data, both cell surface-expressed and intracellular hBRAG
fractions were seen in hBRAG transfectants, whole tonsillar fractions,
and purified B cells but not in mock transfectants. In the
intracellular fraction, more unglycosylated than glycosylated product was present in endogenously expressing cells, whereas a
relatively equivalent level of both isoforms in this pool was seen in
the K562 hBRAG transfectants. These results are similar to those seen
in cell-free translations and immunoblots of whole cellular lysates
(Figs. 1-3). Interestingly, both the putative unglycosylated and
glycosylated products (~52 and 63 kDa, respectively) are seen in the
intracellular fraction, but only the larger, glycosylated product is
seen in the surface fractions (Fig. 5A). Furthermore, in all
lysates assayed using this assay, a large percentage of the
post-translationally modified hBRAG protein is retained intracellularly relative to that expressed on the plasma membrane, a finding that is
consistent with the structural prediction of a short hBRAG transmembrane region (2). Finally, these results show that there is
more surface and intracellular product in enriched B cells than in
whole tonsil, a result consistent with preferential expression of this
product in human B cell lines and tissues (2). Overall, these results
suggest that hBRAG is expressed as both a low but detectable
glycosylated population on the cell surface and a larger, intracellular
fraction of glycosylated and unglycosylated isoforms.
View larger version (39K):
[in this window]
[in a new window]
|
Fig. 5.
Analysis of intracellular and cell
surface-expressed hBRAG isoforms. A, Western blot
analysis of intracellular and cell surface hBRAG expression.
Intracellular protein fraction lysates (I) or biotinylated,
cell surface protein supernatants (S) from A8-6P K562 empty
vector controls (lanes 5 and 6, respectively), or
K562 hBRAG transfectants (lanes 7 and 8,
respectively) were assessed by immunoblotting using the 2422 affinity-purified -hBRAG antibodies 2422 at a concentration of 1 µg/ml (see "Experimental Procedures" for details of cell surface
biotinylation assay). The same assay was used to assess cell surface
supernatants or intracellular lysates of unfractionated tonsils
(lanes 9 and 10, respectively) or purified
tonsillar B cells (lanes 11 and 12,
respectively). As controls for intracellular and plasma membrane
fractionation, bcl2 (lanes 1 and 2)
and CD19 protein expression (lanes 3 and 4) from
A8-6P fractions were assessed under the same assay conditions. The
open and closed arrows show the putative
glycosylated and unglycosylated hBRAG products, respectively. The
positions of molecular mass standards (expressed in kilodaltons) are
indicated in the middle. B, flow cytometry of
K562 cells stably transfected with empty vector or with vector
containing hBRAG full-length (8-3) cDNA, using biotinylated
affinity-purified -hBRAG antibodies from terminal bleed antiserum
raised against the N-terminal hBRAG peptide UT952. As negative and
positive controls, antibodies were pre-incubated with specific blocking
or irrelevant peptide controls, respectively.
|
|
To extend the results seen by flow cytometry, flow cytometry of K562
cells stably transfected with hBRAG (8-3) cDNA was performed using
biotinylated versions of the hBRAG antibodies used in Western blotting
and immunoprecipitation assays (Fig. 5B). To control for
specific staining, the same assays were performed in the presence of
blocking peptide. Relative to the K562 mock transfectant, positive low
level surface and intracellular staining were seen in the K562 hBRAG
transfectant. Furthermore, this was specific for hBRAG, as the same
shift was not seen in the presence of peptide blocking or when using an
irrelevant biotinylated primary IgG (Fig. 5B).
hBRAG Is Associated with Phosphorylated Proteins in Unstimulated
Cells That Are Dissociated and/or Dephosphorylated in Response to BCR
Signaling--
To determine whether hBRAG is part of a signaling
cascade (either as a signaling receptor or intermediate), proteins were immunoprecipitated with hBRAG antibodies under mild conditions and
immunoblotted with -phosphotyrosine antibodies. As seen in Fig.
6A, hBRAG was found to be
associated with phosphorylated proteins in co-immunoprecipitations of
unstimulated A8-6P cells (Fig. 6A, lane 11). We
then examined the possibility that hBRAG may also be functionally
associated with BCR signaling by assessing changes in physical
interactions with other phosphorylated proteins upon BCR ligation. To
do this, proteins from OCI LY8-C3P cells that had been cross-linked
with saturating concentrations of polyclonal anti-µ antibody were
immunoprecipitated with -hBRAG antisera under mild conditions and
subsequently immunoblotted with anti-phosphotyrosine antibodies. From
previous experiments, we have observed optimal tyrosine phosphorylation
in whole cell lysates using 10 µg/ml F(ab')2 anti-µ for
3 min (5-7). Surprisingly, -phosphotyrosine immunoblots of
co-immunoprecipitates revealed that a ~55-59-kDa protein doublet
associated with hBRAG in resting cells may in fact either dissociate
from hBRAG or, alternatively, be selectively dephosphorylated in
response to BCR ligation (Fig. 6A). A weaker set of larger
products also seems to be associated with hBRAG, which are also either
dephosphorylated or disassociated upon BCR ligation, and could possibly
represent multimers of the major ~55-59-kDa doublet products (Fig.
6A). This doublet (and the larger, minor products) detected
in unstimulated B cells are clearly phosphorylated hBRAG-associated
antigens, because they can be eliminated specifically with the
competing peptide (Fig. 6A, lane 12).
View larger version (72K):
[in this window]
[in a new window]
|
Fig. 6.
Detection of potential hBRAG-associated
phosphorylated proteins in resting but not BCR-ligated A8-6P lysates
immunoprecipitated with -hBRAG
antibodies. A, anti-phosphotyrosine
(-P Tyr) Western immunoblots showing the
effect of BCR ligation on protein tyrosine phosphorylation in whole
cell lysates or -hBRAG-immunoprecipitated lysates prepared from
hBRAG+ mature B cell line variant A8-6P. Whole cell lysates
alone from 2 × 107 unstimulated A8-6P cells
(0) or from 2 × 107 A8-6P cells
cross-linked with 30 µg/ml F(ab')2 anti-µ for either 30 s, 1 min, 3 min, or 5 min were subjected to SDS-PAGE, transferred to PVDF
membranes, and probed with a polyclonal anti-phosphotyrosine antibody.
Alternatively, lysates prepared from 2 × 107
unstimulated A8-6P cells or A8-6P cells cross-linked with 20 µg/ml
F(ab')2 anti-µ for 3 min were immunoprecipitated (IP)
under mild (co-immunoprecipitating) conditions (i.e. 1%
Triton X-100, 150 mM NaCl, 50 mM Tris-HCl, pH
8, 5 mM EDTA, 5 mM EGTA, 5 mM NaFl,
1 mM sodium orthovanadate, 1 mM BSA, and
protease inhibitors) with either irrelevant affinity-purified rabbit
IgG (rIgG) at 10 µg/ml or N-terminal 2422 -hBRAG
antibodies affinity-purified from terminal bleeds at 10 µg/ml
dilutions (-hBRAG); they were then subjected
to SDS-PAGE, transferred to PVDF membranes, and probed with 1.5 µg/ml
polyclonal 4G10 anti-phosphotyrosine antibody. The position of the
putative major ~53/56-kDa hBRAG co-immunoprecipitating doublet
species that is dephosphorylated upon BCR stimulation is denoted by an
arrow, and the positions of molecular mass standards are
indicated in the middle. B, the same Western
blots stripped and re-immunoblotted -hBRAG antibodies. The same
affinity-purified -hBRAG antibodies used in the immunoprecipitations
were used but at a concentration of 1 µg/ml. For both A
and B, the blots were also probed without lysate () to
control for cross-reactivity of the mouse -rabbit second-step
horseradish peroxidase-conjugated reagent with the heavy chain IgG
band. 200 µg/ml irrelevant rabbit IgG alone added to
immunoprecipitation buffer was also used as a negative control. To test
the specificity of bands under the described co-immunoprecipitating
conditions, 10 µg/ml competing peptide U-T952 was used
(+UT952); for comparison, an unstimulated sample without
peptide blocking was run adjacently (UT952). The position
of the putative major ~52-kDa hBRAG unglycosylated species recognized
by affinity-purified 2422 is denoted by an arrow. The human
IgG heavy chain migrates just below this product. Positions of
molecular mass standards (expressed in kilodaltons) are indicated on
the left.
|
|
Various candidate doublet molecules that are constitutively
phosphorylated in B cells in the ~55-59-kDa molecular mass range include three B cell protein tyrosine kinases (PTKs) known to be
associated with BCR signaling, Fyn, Lyn, and Hck, the latter two
existing as doublet isoforms of 56/59 and 53/56 kDa,
respectively. To assess the possibility that hBRAG associates with any
or a combination of these candidate proteins, antibodies against Lyn, Fyn, and Hck were used in co-immunoprecipitation experiments of the
same A8-6P lysates under the same conditions as for the original hBRAG
co-immunoprecipitations shown in Fig. 6. The parallel experiment, co-immunoprecipitation with hBRAG, was also performed, and
immunoprecipitates were resolved on SDS-PAGE and immunoblotted with
either -phosphotyrosine, -hBRAG and either -Lyn, -Fyn, or
-Hck antibodies (Fig. 7). No
association with Lyn (Fig. 7A) or Fyn (data not shown) was detected, but hBRAG demonstrated low level association with Hck (hemopoietic cell kinase) in both
-Hck immunoprecipitates immunoblotted with -hBRAG and in
-hBRAG immunoprecipitates immunoblotted with -Hck antibodies
(Fig. 7B). Based on the intensity of the associated proteins
detected in -phosphotyrosine immunoblots of -hBRAG immunoprecipitates, Hck is probably not the only hBRAG-associated antigen being dephosphorylated or dissociated upon BCR ligation in
A8-6P B cells (Fig. 7A). Also noteworthy is that the
Hck fraction associated with hBRAG in resting A8-6P cells is
likely dephosphorylated rather than dissociated, because comparable
levels of Hck are seen in -Hck immunoblots of
-hBRAG immunoprecipitates either in the presence or absence of BCR
ligation (Fig. 7B). This assumes that Hck at least partially
corresponds to the species that is selectively dephosphorylated in
BCR-stimulated hBRAG immunoprecipitates. Finally, it is noteworthy that
hBRAG itself is not phosphorylated/dephosphorylated, at least not in
the fraction that associates with Hck in -Hck immunoprecipitates.
This observation, however, does not rule out that hBRAG is
phosphorylated/dephosphorylated as part of the 60-kDa whole cell lysate
fraction. Further experiments are underway to examine this
possibility.
View larger version (60K):
[in this window]
[in a new window]
|
Fig. 7.
Co-immunoprecipitation of A8-6P lysates with
anti-hBRAG antibodies or antibodies against candidate hBRAG-associated,
doublet B cell PTKs Lyn and Hck. Lysates were prepared from 2 × 107 A8-6P cells that were unstimulated A8-6P () or
cross-linked with 20 µg/ml F(ab')2 anti-µ for 3 min
(+). Immunoprecipitation conditions and cell and antibody
concentrations used are as detailed in Fig. 6 legend and under
"Experimental Procedures." Immunoprecipitates were resolved in
duplicate on SDS-PAGE under denaturing conditions, transferred to PVDF
membranes, and probed with either -hBRAG N-terminal
affinity-purified antibodies or antibodies directed against either the
Lyn or Hck PTKs. rIgG, rabit
immunoglobulin G.
|
|
Cross-linking hBRAG Mediates Intracellular Signaling Alone and in
Combination with BCR Ligation--
The above data show that hBRAG may
function as a cell surface receptor with signaling capabilities.
Although a large portion of the glycosylated and unglycosylated forms
of this protein seems to be intracellularly retained, a low level can
also be expressed on the cell surface (Fig. 5). To assess hBRAG
signaling potential in the context of the B cell lineage, the putative
hBRAG cell surface receptor was ligated with -hBRAG antibodies, and
proximal signaling parameters, i.e. phosphorylation patterns
in whole cell lysates, were assessed using an -phosphotyrosine
antibody in immunoblots (Fig.
8A). -hBRAG stimulation at
presumably saturating concentrations (10 µg/ml) generated a low but
detectable tyrosine phosphorylation response relative to unstimulated
cells and to the irrelevant rabbit IgG control but one that was much
lower than that seen following BCR stimulation. It is difficult to
ascertain from these experiments whether the phosphorylation pattern is exactly the same or partially overlapping but distinct from that of BCR
ligation (Fig. 8A). Interestingly, when hBRAG was ligated in
combination with BCR, this increased the intensity and the pattern of
BCR mediated tyrosine phosphorylation, suggesting that it may
positively enhance early activation events through the BCR (Fig.
8A, lanes 8 and 9). However, ligation
of the BCR in combination with the irrelevant IgG control also
increased BCR-mediated phosphorylation to almost the same degree as
-hBRAG + anti-µ ligation (Fig. 8A, lane
10). One possible reason for this effect is that the intact IgGs
used in these studies may nonspecifically alter BCR signaling by
co-cross-linking of FcRII/III receptors. Fc
receptor ligation is generally thought to be involved in negative signaling via activation of their ITIM motifs and association with
shp1 Src homology 2 domains, resulting in the subsequent dephosphorylation of selected downstream signaling intermediates (8).
In the context of B cells, FcRII ligation has been shown to inhibit several consequences of BCR stimulation, including phosphoinositide hydrolysis, intracellular calcium flux, cellular proliferation, and Ig secretion. However, ligation of
FcRII in combination with BCR, i.e. ligation,
with intact anti-µ and its effect on resulting phosphorylation
patterns, although not well documented in human mature B cells, appears
to produce either similar or more intense overall tyrosine
phosphorylation patterns in whole cell lysates of mouse
BCR+ B cells and B cell lines relative to
F(ab')2 anti-µ cross-linking (reviewed in Ref. 8).
View larger version (65K):
[in this window]
[in a new window]
|
Fig. 8.
Analysis of tyrosine phosphorylation profiles
in A8-6P whole cell lysates cross-linked with anti-hBRAG,
anti-µ, or both. A, relative
contribution of hBRAG in protein tyrosine phosphorylation induction in
B cells. Whole cell lysates were prepared from A8-6P endogenous
hBRAG-expressing mature OCI LY8-C3P variants that were unstimulated
(lane 1), stimulated with a goat F(ab')2
polyclonal -human µ antibody fragment alone (lanes 2 and 3), hBRAG antibodies alone (affinity-purified from
terminal bleed antiserum raised against the N-terminal peptide UT952;
lanes 4-6), or co-ligated with both (lanes
7-9). The durations and concentrations of antibody stimulation
are indicated in each lane. To control for nonspecific
effects, 10 µg/ml irrelevant affinity-purified rabbit IgG
(rIgG) was also used for co-ligation with 10 mg/ml anti-µ
for 3 min (lane 10). B, relative effect of
specific hBRAG co-ligation in enhancing BCR-mediated tyrosine
phosphorylation. Whole cell lysates were prepared from A8-6P endogenous
hBRAG-expressing mature OCI LY8-C3P variants that were either left
unstimulated or stimulated for 3 min with either 0.1 or 10 µg/ml
-human µ alone (lanes 2 and 3, respectively), 10 µg/ml of both -human µ and -hBRAG (lanes 4-6), or
the combination of 10 µg/ml -human µ and the irrelevant IgG
control. To test for the relative contribution by Fc
receptor components, A8-6P cells were preincubated with 10 µg/ml
blocking antibodies directed against human IgG Fc receptors
(+Fc block, lanes 5 and 8,
respectively). To test for specificity of the hBRAG-mediated effect,
hBRAG antibodies were pre-incubated with 10 µg/ml competing peptide
UT952 (+UT952, lane 6). In both A and
B, 2 × 107 cell equivalents were
fractionated by SDS-PAGE run under denaturing conditions and
transferred to a PVDF membrane, and phosphotyrosine-containing proteins
were detected using the 4G10 antibody. Blots were also stripped and
re-probed with an -actin antibody to control for loading
differences.
|
|
To test for the contribution of Fc-mediated nonspecific
effects versus specific hBRAG effects in combination with
BCR ligation, these components of hBRAG ligation were blocked by
pre-incubation of -hBRAG antibodies either with saturating
concentrations of UT952 peptide or with saturating concentrations of
blocking antibodies directed against the FcRII/III
receptor in A8-6 cells. The results of these experiments suggest that
increases in tyrosine phosphorylation in A8-6P cells are at least
partially mediated by nonspecific Fc-mediated effects, as a
decrease in phosphorylation was observed with IgGs or hBRAG ligation in
the presence of the Fc-blocking antibody (Fig.
8B, lanes 5 and 8, respectively).
However, a substantial component of the -hBRAG ligation effect also
appears to be specific, as blocking with UT952 has a significant effect
in suppressing the enhanced BCR-mediated phosphorylation pattern. This
reduction of phosphorylation signal mediated by the competing peptide
is more marked than that mediated by Fc block (Fig. 8B,
lane 6). In addition, the intensity of the phosphorylation
pattern mediated by -BCR in combination with -hBRAG is increased
relative to that mediated by -BCR in conjunction with irrelevant IgG
(Fig. 8B, lane 6). The differences in signal
intensity were not due to loading differences, as demonstrated by
re-probing immunoblots with an -actin antibody.
| |
DISCUSSION |
As seen in -hBRAG immunoblots of whole cell lysates and
-hBRAG immunoprecipitations of de novo synthesized
proteins under non-denaturing conditions (Figs. 3 and 4, respectively),
we showed that hBRAG exists as a disulfide-linked, dimeric and possibly trimeric protein. We also showed that hBRAG is glycosylated but that a
large unglycosylated pool also exists in whole cell lysates and
cell-free translations (Figs. 2 and 5). The detection of a highly
glycosylated hBRAG isoform with hBRAG antisera is consistent with
earlier structural predictions of a TM region with four
N-glycosylation sites in the hBRAG extracellular domain.
Using cell surface staining and biotinylation assays (Fig. 5), we have
also shown that hBRAG can be expressed at low but detectable levels as
a cell surface-expressed molecule, a finding consistent with the
structural prediction that hBRAG, which lacks an endoplasmic reticulum
retention signal, should at least partly leak out to the plasma
membrane. Finally, we have demonstrated that hBRAG may function as a
signaling receptor and mediates phosphorylation changes after being
cross-linked alone or together with BCR cross-linking. Structurally
consistent with cell surface-expressed hBRAG also being a signaling
receptor is the fact there are various candidate signaling motifs in
its 81-amino acid cytoplasmic N terminus; these motifs include the consensus cAMP and protein kinase C phosphorylation sites and palmitylation residues (2) shared by various other signaling receptors
and intermediates, such as G protein-linked receptors, G proteins, and
Src familly kinases including Src, Blk, Lyn, Yes, Hck, Fyn, and Lck
(reviewed in Ref. 9).
The structural features of hBRAG has revealed many parallels to other
type II TM proteins preferentially expressed in lymphoid cells and/or
with characterized immunologic function. First, like hBRAG, all of
these molecules exist as glycosylated, disulfide-linked, higher order
structures as homodimers (BP-1, CD69, CD72, LY49, CD94), heterodimers
(CD94/NKG2), or homotrimers (CD23, Fas-L; reviewed in Ref. 10). Second,
with the exception of the invariant chain, all are preferentially
expressed on the cell surface and function as signaling receptors with
either negative signaling potential, i.e. the C-lectin
family, or positive/co-stimulatory potential, i.e. the TNF
family, mostly involved in T-T and T-B cell interactions. On the other
hand, hBRAG has several structural features distinguishing it from
other type II proteins with characterized immunologic roles. First,
hBRAG has no overall or domain-specific sequence homology to any of
these proteins and is likely not a multigene family member, based on
genomic Southern blot analysis (2). Second, with the exception of the
C-lectin receptor family members CD23 and CD72, hBRAG mRNA is
preferentially expressed in B cells and at low or undetectable levels
in T cells (2). Third, hBRAG has a shorter TM than the above-mentioned
molecules, a property of Golgi-associated integral membrane proteins
(reviewed in Ref. 11). The fact that the in vitro translated
hBRAG product in the presence of microsomes is efficiently targeted to
the Golgi membrane and glycosylated (Fig. 2) and the relatively large
fraction of intracellular hBRAG in biotinylation experiments (Fig.
5A) both reinforce this notion. The invariant chain is the
only known type II TM protein with characterized immunologic function
that is predominantly retained intracellularly. Its intracellular
localization, however, can occur via either its retention or
endosomal-targeting signals, neither one of which hBRAG contains in its
predicted structure. Like TNF ligands and scavenger receptors, hBRAG
has many cysteine-rich regions, but these are not similar enough to constitute a TNF or SRCR (scavenger receptor cysteine-rich) motif. hBRAG is also more extensively glycosylated (at least in
vitro) than either TNF or C-lectin receptors (Fig. 2), which have
only one or two N-glycosylation sites/polypeptide monomer
(the only exception to this being CD30L).
As an extension to the structural features presented above, several
functional features of the cell surface-expressed hBRAG fraction
support its role as a signaling receptor in the B cell lineage. First,
-hBRAG co-immunoprecipitation results suggest that it and/or various
phosphorylated molecules associated with it are selectively
dephosphorylated (and/or dissociated) within 1 min post-BCR ligation
(Fig. 6); this suggests that hBRAG either directly phosphorylates these
molecules or associates with molecules that are phosphorylated in
response to BCR cross-linking. Second, -hBRAG treatment produces low
but detectable increases in tyrosine phosphorylation relative to
controls and may also enhance BCR-mediated early activation events
(Fig. 8). Finally, hBRAG may associate with Hck because this Src PTK is
expressed predominantly in myeloid and B cell lineages (12, 13).
Although involvement of Hck in myeloid lineage signaling pathways is
well characterized, its precise role in the BCR signaling pathway is
not well understood. However, evidence for its involvement in this
process comes from several observations. For example, upon BCR
ligation, Hck can be phosphorylated and activated, demonstrating
in vitro kinase activity and phosphorylation of various
in vivo BCR signaling intermediates including p120 (Cbl),
Bcr-abl, and RAS-GTPase-activating protein (reviewed in
Refs. 14 and 15). Additionally, Hck can interact with other BCR or
pre-BCR components/signaling intermediates including Ig- and Ig-
(via its Src homology 2 domains), Bruton's tyrosine kinase, and the
Bcr-abl tyrosine kinase (via its Src Homology 3 domains;
Refs. 16 and 17; reviewed in Ref. 18).
Because hBRAG mRNA expression correlates with B cell-specific
expression (2), we hypothesize that hBRAG may mediate potentially important developmental regulatory signals in a B cell-specific signaling pathway. Additionally, based on stable transfection experiments of the hBRAG cDNA, which resulted in RAG1 mRNA
levels being increased in a B but not a myeloid cell line (2), we further hypothesize that such a pathway is important for B
cell-specific RAG expression. Various signaling molecules have
functions in RAG regulation at various stages of B cell development
(19). In particular, hBRAG may be involved in signaling-mediated
regulation of RAG expression in earlier B cell differentiation,
analogous to pre-BCR, pre-T cell receptor, CD19, and the IL-7 receptor. Some of these receptors, such as CD19, have also been shown to have
roles in "negative" as well as positive signaling functions (20).
CD19 can lower the threshold of BCR ligation when co-ligated in
conjunction with BCR but, conversely, can render a cell refractive to
BCR signaling when pre-ligated alone (21). In this context, it would be
interesting to ascertain whether hBRAG may function in an analogous way
to CD19 in the ability to exhibit biphasic signaling responses and may
be involved in an ITIM-independent negative signaling pathway.
Furthermore, hBRAG, like CD19, may be an example of a molecule that can
act both in altering BCR-mediated early activation events and in
positively altering RAG expression in human B cell development.
Additionally, because hBRAG appears to be enriched in peripheral
tissues, it may potentially deliver a signal that re-induces RAG
expression and secondary rearrangements, much like signaling through
the CD40 co-stimulatory receptor, and the IL-4 or IL-7 receptors in
mature germinal center B cells (review
TOP
ABSTRACT
INTRODUCTION
