| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Basel Institute for Immunology, Basel, Switzerland
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. In addition, EBV
binds to human CR2, accounting for its cellular tropism for B cells and
epithelia (reviewed in Refs. 1 and 2). CR2 plays diverse roles in the immune system. Monovalent binding of B cell CR2 is inhibitory, whereas multivalent binding has costimulatory growth factor activity (3, 4). The coaggregation of CR2 with surface Ig (sIg) results in synergistic activation signals being transduced to the cell cytoplasm through the associated CD19 molecule (5). Since residues in the cytoplasmic tail of CR2 become phosphorylated after some stimuli (6, 7), it is possible that CR2 also has intrinsic signaling activity, although the functional consequences remain unknown. These different signaling modalities are biologically significant, since they provide a mechanism whereby CR2 could help regulate the progression of B cells through the cell cycle and assist in the Ag-induced activation of B cells for those Ags complexed to C3d.
CR2 on FDCs binds C3d-complexed Ags and immune complexes, thereby immobilizing and concentrating them in germinal centers. This is believed to assist in the generation and persistence of efficient germinal center reactions and to enhance the selection of high-affinity sIg mutants (2). Ags are retained on FDCs in native form and are not internalized, whereas CR2 on B cells is endocytosed (8). Although human B cells and FDCs are known to express alternatively spliced transcripts of CR2 (9, 10), it is not known whether this provides a structural basis for the different functional properties of CR2. In addition, some anti-CR2 Abs precipitate a 200-kDa protein from human epithelial cell lysates; however, the relationship of this molecule to human B cell CR2 (140 kDa), and its function, also remain unresolved (11).
Here we identify a novel isoform of CR2 that has ubiquitin attached
covalently to the cytoplasmic domain. Ubiquitinated (190 kDa) and
conventional nonubiquitinated (150 kDa) CR2 isoforms are expressed on
the surface of sheep B cells as noncovalently associated dimers. The
external domains of the two isoforms are distinctive in some way since
one mAb selectively binds the ubiquitinated isoform. The basis for
these unusual higher-order topographical features appears to lie in the
primary structure of sheep CR2, which has in its transmembrane region a
rare amino acid (aa) motif that has recently been shown to mediate
noncovalent dimerization of -helices. Furthermore, the cytoplasmic
region of sheep CR2 contains two conserved lysine residues that provide
potential sites for the attachment of ubiquitin. The primary structures
of sheep ubiquitin and C3d ligand were also determined, and these
proteins are highly conserved.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Healthy White Alpine sheep were obtained from Versuchsbetrieb
Sennweid (Olsberg, Switzerland). In young lambs, the ileal Peyer’s
patch (IPP) is the major B cell-generating organ and contains 
99% B
cells, most of which are immature (12). IPP B cells were isolated using
established procedures (13). Blood was collected by venipuncture using
EDTA as an anticoagulant, and lymphocytes were recovered by
centrifugation on discontinuous Percoll gradients (14). Recirculating
lymphocytes were collected by cannulating efferent lymph ducts (15).
Because B cells comprise 20 to 40% of the lymphocytes in sheep blood
and lymph, and no cell populations in blood or lymph other than B cells
expressed CR2 (Ref 16. and our unpublished observations), there was no
other possible source of CR2 during immunoprecipitation, and further
cell purification was unnecessary. Experimental protocols were approved
by the Kantonalveterinäramt Basel-Stadt.
Monoclonal Abs
A total of 8 new mAbs were produced in two separate fusions. In the first, BALB/c mice were immunized with a suspension of cell membranes isolated from the IPP of a 3-mo-old lamb. A piece of ileum was removed at necropsy, opened, and washed clean with PBS followed by a rinse in absolute ethanol and then with PBS again. The surface mucosa was scraped away with a glass slide. Then, the submucosal tissue, including the IPP follicles, was removed by deep scraping. A total of 12.5 g of this material was suspended in 15 ml of 0.25 M sucrose in 6 mM HEPES buffer and centrifuged at 100 x g for 5 min. The supernatant was removed and diluted with an equal volume of the sucrose/HEPES buffer and then centrifuged at 1,000 x g for 20 min. The supernatant was again removed and centrifuged at 12,000 x g for 30 min to form a pellet weighing 0.2 g. The pellet was resuspended in 5 ml serum-free DMEM and confirmed by light microscopy to be a cell-free mixture of membranous material. Mice were immunized by injection i.p., boosted 9 days later i.v., and after 3 days the immune splenocytes were fused to SP2/0 myeloma cells using standard techniques.
When it became clear that four mAbs from the above fusion (Fusion Du2)
recognized sheep CR2 and had unusual properties, we immunized mice
several times with a mixture of affinity-purified gp150 and gp190, the
two isoforms of sheep CR2, and produced more hybridomas using the
methods described above. This second fusion (Fusion Du14) produced
another four mAbs specific for sheep CR2 (Table I
).
|
Immature B cells isolated from the IPP and mature lymphocytes from blood or efferent lymph were surface iodinated by the lactoperoxidase method, lysed with either 0.5 or 2% Nonidet P-40, 1% digitonin, or 10 mM CHAPS (3-[(3-cholamidopropyl) dimethyl-ammonio]-1-propanesulfonate) and target Ags captured by either solid-phase immunoisolation (17) or protein G-Sepharose (18). Two variations were introduced to this procedure. To identify mAb specificities in situ, cells were incubated with mAbs for 30 min following surface iodination, washed thoroughly, lysed, and reacted promptly with protein G-Sepharose without further addition of mAbs. In other experiments, surface molecules were chemically cross-linked before lysis using either noncleavable (Bis(sulfosuccinimidyl) suberate; Pierce, Rockford, IL; no. 21580) or thiol-cleavable (Dithiobis(sulfosuccinimidylpropionate); Pierce; no. 21578) cross-linkers following procedures recommended by the manufacturer. To remove N-linked carbohydrates, some precipitates were digested in endoglycosidase F (Boehringer Mannheim (Schweiz), Rotkreuz, Switzerland; no. 878 740) before electrophoresis, following the manufacturer’s directions. Proteins were separated by SDS-PAGE in 7.5% gels run under reducing or nonreducing conditions (18) and visualized by autoradiography.
NH2-terminal sequencing
Sheep CR2 was purified from lysates of IPP B cells or efferent lymph lymphocytes (0.5% Nonidet P-40) using mAbs 2-74, 2-87, and 2-128 coupled separately to NHS-HiTrap columns (Pharmacia, Uppsala, Sweden). Protein was eluted with glycine-HCl, pH 2.5, and fractions containing CR2 were pooled and precipitated using TCA. Proteins were separated by SDS-PAGE and electroblotted to a polyvinylidene difluoride membrane (19), and the NH2 terminus was sequenced using an Applied Biosystems (Foster City, CA) 494 Protein Sequencer. Proteins were purified and sequenced at least once using each of the three mAb columns. In total, gp150 and gp190 were independently sequenced four and two times, respectively. On one occasion, cysteine residues were alkylated with 4-vinylpyridine before SDS-PAGE.
Immunoblotting
Aliquots of CR2 proteins immunopurified as above were separated in 7.5% SDS gels in a MiniPROTEAN II electrophoresis cell (Bio-Rad, Hercules, CA) run at 60 mA for 1 h. Proteins were then transferred to hydrated nitrocellulose membranes in methanol buffer using a mini Trans-blot cell (Bio-Rad) run at 100 V for 1 to 1.5h. Immunochemical detection was done as described (20) using antisera specific for sheep CR2 and ubiquitin. To produce anti-sheep CR2 antiserum, rabbits were immunized s.c. with the NH2-terminal peptide IFCDPPPSIKNGRSGYHS synthesized on MAP matrix (21) mixed with Freund’s complete adjuvant and boosted six times using peptide in incomplete adjuvant. Immune, but not preimmune, serum reacted specifically with purified sheep CR2 proteins. Antiserum specific for conjugated ubiquitin was purchased (Dako Diagnostics, Zug, Switzerland; no. Z 0458). Similar results were obtained using two other antisera specific for conjugated ubiquitin (Sigma, St. Louis, MO; no. U-5379, and another kindly provided by Dr A. L. Haas, Medical College of Wisconsin, Milwaukee, WI).
Peptide mapping
Peptide mapping was done as described (22) using endoproteinase Glu-C from Staphylococcus aureus V8 (Boehringer Mannheim; no. 791 156) at 100 µg/ml and 200 µg/ml.
cDNA cloning of CR2
To generate initial clones, a sense oligonucleotide primer (5' TGTGACCCTCCTCCGTCTATCAAAAATGGCCGG) was designed by comparing the NH2-terminal aa sequences of sheep and human CR2 and introducing parsimonious nucleotide changes to the corresponding human DNA sequence. An antisense CR2 primer (5'-CCTCTCTCTGGCCCAGGGTAAC) was chosen from a stretch of the extracellular domain of human CR2 conserved between human and mouse. A commercially available RT-PCR kit (GeneAmp; Perkin-Elmer, Branchburg, NJ) was then used according to the manufacturer’s directions to amplify from IPP B cell RNA a fragment of sheep CR2 covering about half of the external domain. Amplified cDNA was cloned into the pCR vector (Invitrogen, San Diego, CA) and sequenced using a T7 Sequenase version 2.0 kit (United States Biochemical, Cleveland, OH). A similar strategy of using conserved priming sites allowed overlapping fragments that extended into the cytoplasmic domain of sheep CR2 to be cloned. The primers used in this case were (sense 5'-GTCTTTGTAAAGAAATCACCTGCC) and (antisense 5'- AGAATATACTTCTCGTGCTTCTAAATGAA). Once authentic sheep CR2 cDNA sequences were available, a 3' and 5' RACE kit (Boehringer Mannheim) was used to extend the termini, and a full-length sequence was assembled from the overlapping cDNA clones. The mature 4.0-kb transcript covered by these clones included a 5'UT region, a full-length coding sequence, and a 3' UT region ending at a poly(A) tail.
cDNA cloning of ubiquitin
Oligo(dG)-tailed cDNA was prepared from RNA isolated from IPP B cells. An antisense primer for ubiquitin (5' ACAGGTTCAGCTATTACTGA) was designed from the 3' UT region of bovine ubiquitin (23). Ubiquitin cDNA was then amplified by anchored PCR, using the above primer in combination with a sense primer containing oligo(dC). Amplified cDNAs were cloned into pCR vector (Invitrogen) and sequenced.
pCR C3d sequencing
Sheep C3d was sequenced as both protein and cDNA. Sheep C3 was isolated from EDTA-plasma using described procedures (24) except that a Superdex 200 pg HR 16/50 size exclusion column was used as a final step in purification. C3dg was generated by incubating C3 with 2% (w/w) elastase for 5 h at 37°C and purified from C3c and C3a by size exclusion chromatography on a Superdex 200 column equilibrated in PBS. Purified protein (33 kDa) was blotted onto a polyvinylidene difluoride membrane, and the NH2 terminus was sequenced as described above. Internal peptide fragments were produced by trypsin digestion of C3dg, purified, and sequenced.
To clone cDNA encoding the sheep C3d fragment, the nucleotide sequences of human, mouse, rabbit, and Xenopus C3 were aligned and two degenerate primers (sense 5' GAGACCARAATYMTCCTGCAAGGG and antisense 5' CTTGGTCTCTTCYGAYCGCAGGAG, where R = A/G, Y = C/T, M = A/C) were designed to anneal adjacent to the C3dg elastase cleavage sites. cDNA was amplified from sheep liver RNA, cloned into the pCR2 vector (Invitrogen), and sequenced.
Flow cytometry
For two-color FACS analysis, suspensions of lymphocytes were reacted with the primary Abs and then sequentially with conjugated class-specific secondary Abs, e.g., phycoerythrin-conjugated goat anti-mouse IgG1 or anti-mouse IgG2b and FITC-conjugated goat anti-mouse IgM (Southern Biotechnology Associates, Birmingham, AL). Relative fluorescence in the FL1 and FL2 channels was analyzed on a FACScan flow cytometer (Becton Dickinson, Franklin Lakes, NJ).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Eight new mAbs reacting mainly with B cells and FDCs specifically
immunoprecipitated either one or two glycoproteins from lysates of
surface-iodinated sheep B cells, gp150 and gp190 (Table I
). The mAb
2-128 was selective for gp190 while the remaining seven mAbs
precipitated both glycoproteins (Fig. 1,
A–C). Both target molecules contained 30 kDa of
N-linked carbohydrates when expressed on each of three B
lymphocyte populations: immature cells in the IPP, mature cells in
blood, and recirculating cells in efferent lymph (Fig. 1, A
and B). In all samples, the signal strength for gp190
was consistently lower than for gp150 and lysates made from immature
IPP B cells yielded a lower signal strength for gp190 than did lysates
of mature B cells (Fig. 1B). Similar results were
obtained using three different detergents during cell lysis: Nonidet
P-40, digitonin, and CHAPS (data not shown).
|
D). Sequencing of gp190
unexpectedly yielded two equimolar NH2-terminal sequences,
one identical to that of gp150, and another homologous to ubiquitin
(Fig. 1D). These initial results identified gp150 as
the sheep homologue of human and mouse CR2 and pointed to the
possibility that, as for a number of other membrane proteins, sheep CR2
was a natural ubiquitination substrate. Ubiquitin is normally attached
to proteins by isopeptide covalent bonding between a glycine residue in
the COOH-terminus of ubiquitin and lysine residues in the substrate.
This leaves two accessible NH2 termini, accounting for the
two peptide sequences in equimolar ratios. Primary structure of sheep CR2
To better characterize sheep CR2, six overlapping cDNA clones that collectively encoded a full-length CR2 transcript were generated by RT-PCR of sheep B cell RNA. Initially, clones were amplified using nucleotide primers derived from regions of CR2 that are conserved between human and mouse. Once sheep sequences were known, additional overlapping cDNA clones were generated using 5' and 3' RACE and a full-length sequence covering a mature 4.0-kb transcript was assembled.
The NH2-terminal aa of sheep CR2 predicted from cDNA
matched all residues assigned previously by Edman degradation as being
common to gp150 and gp190, confirming their identity. The mature sheep
CR2 transcript has a conserved structure, with 69 and 66% overall aa
identity to the human (25) and mouse (27) proteins, respectively, and
matched the "short" 15 short consensus repeat (SCR) version of
human CR2 (28), which lacks SCR11 (Fig. 2). Three separate cDNA clones generated
by PCR extended through this region of the sheep CR2 transcript and
none of them contained additional sequences corresponding to SCR11
(not shown). The external domains of sheep, human, and mouse CR2 have a
relatively high and uniform level of aa identity (65-80%),
except for SCR1 and SCR12 (41–54%) (Fig. 3A). Eight potential
N-linked glycosylation sites are conserved between the
sheep, human, and mouse proteins. Additional asparagine residues in
each sequence make for a total of 13, 11, and 16 potential
N-linked glycosylation sites in sheep, human, and mouse CR2,
respectively (Fig. 3B).
|
|
The transmembrane region of CR2 is significantly less well
conserved between sheep, humans, and mice than are other regions of the
molecule (29–52% aa identity) (Fig. 3A). A more
detailed examination of protein sequences (Fig. 4A) revealed that the
transmembrane domain of sheep CR2 contained a stretch of sequence
(LIxxxxxGVxxxxxxGVxT) that was remarkably similar to a rare 7-aa motif
(LIxxGVxxGVxxT) identified first in the glycophorin receptor and shown
subsequently to drive the dimerization of transmembrane -helices
(Fig. 4). As far as we are aware, sheep CR2 is only the second native
protein shown to contain this type of motif. It is not strictly
conserved in the human and mouse transmembrane domains, although
individual or paired aa comprising the motif are repeated with a
spacing of 3 to 5 aa, and certain aspects occur in reverse orientation
in the mouse transmembrane domain (Fig. 4B).
|
A). A number of aa residues
that could have functional significance are conserved between them.
Thus, two conserved lysine residues provide potential sites for
ubiquitination and four conserved tyrosine residues, as well as three
conserved threonine/serine residues, offer potential phosphorylation
sites (Fig. 4A). Ubiquitinated CR2 isoform
NH2-terminal sequencing had indicated early on
that the gp190 form of sheep CR2 was specifically ubiquitinated. This
was confirmed by affinity purifying the two target proteins and
analyzing their Western blot reactivity with polyclonal Abs recognizing
an NH2-terminal peptide of sheep CR2 or ubiquitin. Both
gp150 and gp190 reacted with the anti-CR2 antiserum but only gp190
contained ubiquitin (Fig. 5A). Similar results
were obtained using three antisera specific for ubiquitin. None of
these antisera reacted with the surface of sheep B cells when tested by
FACS analysis (data not shown). Taken together, the results presented
to date established unambiguously that gp150 and gp190 were different
isoforms of sheep CR2, one with ubiquitin attached covalently in the
cytoplasmic region (CR2ub 190 kDa), and another that was
nonubiquitinated (CR2no 150 kDa).
|
We then sought to explain why mAb 2-128 distinguished between
these two isoforms, while the other seven mAbs identified both. One
possibility was that in addition to selective ubiquitination, the two
isoforms arose from differentially spliced transcripts such that the
primary structure of their external domains differed and mAb 2-128
identified a unique sequence-dependent epitope of
CR2ub. However, the digestion of surface-labeled
CR2no and CR2ub with endoproteinase did not
reveal any significant difference between the peptide composition of
the external domains of these two proteins (Fig. 5B).
Surface dimerization of sheep CR2
Attempts to explain the reactivities of the mAbs had also to
account for another consistent finding. When the anti-CR2 mAbs were
used in all possible combinations to stain sheep B cells for two-color
analysis, the FACS profiles always resolved as diagonal lines,
suggesting a fixed stoichiometric relationship between Ab binding sites
(Fig. 6A). There were
two possible ways to explain these binding properties: 1) the mAbs
2-54, 2-74, and 2-87 had binding epitopes on both CR2no and
CR2ub, whereas the 2-128 epitope occurred exclusively on
CR2 ub; or 2) CR2no and CR2ub were expressed on
the cell surface as a closely associated complex and, while the 2-128
epitope occurred only on CR2ub, the other mAbs recognized
epitopes that bridged the molecular complex. In view of the primary
sequence of the transmembrane domain of sheep CR2, the possibility of
dimerization between sheep CR2 isoforms seemed plausible, and we tested
whether this occurred and assessed whether it influenced the reactivity
of the mAbs.
|
B). This confirmed
that the mAbs bound specifically to their respective target molecules
while these were in their native state in the B cell membrane.
Next, surface molecules were first iodinated and then cross-linked
chemically before cell lysis and immunoprecipitation. Two different
cross-linking agents were used, noncleavable and thiol cleavable, with
spacer arm lengths of 11.4 and 12 Å, respectively. When using the
noncleavable linker, mAbs 2-74 and 2-128 both precipitated a single
molecular species of comparable mass. In the case of mAb 2-74, the two
native CR2 isoforms were depleted quantitatively and only a trace
amount of noncross-linked CR2no remained (Fig. 6C, lanes 1 and 2).
When a thiol-cleavable linker was used, the single target complex
immunoprecipitated by each mAb redissociated in reducing gels into two
comparable glycoproteins, although their electrophoretic mobilities
were slightly faster than the native CR2 isoforms, perhaps reflecting
some residual intramolecular cross-linking (Fig. 6C).
The effective cross-linking of two glycoproteins into a single
molecular species, and its recognition by both mAbs 2-74 and 2-128,
implied that the physical distance between the two CR2 isoforms when
expressed in the B cell membrane was less than the spacer arm lengths
of the cross-linking agents. However, the strength of signal produced
when mAb 2-128 was used in this way was very low (Fig. 6C). This problematic result was obtained
consistently in several experiments, and suggested that the chemical
treatment partly destroyed the 2-128 epitope.
As a final approach, iodinated B cells were first labeled with the mAbs 2-128 or 2-54, treated with the thiol-cleavable cross-linker, and then lysed. This allowed the mAbs to first bind to their target epitopes. The molecular complex would then be stabilized in its native configuration. The Ags recognized specifically by mAb 2-74 were then immunoprecipitated from an aliquot of each lysate. In parallel experiments, the target complexes recognized by mAbs 2-128 or 2-54 were immunoprecipitated from other aliquots of each lysate (preclearing) and the precleared samples were then also subjected to a final immunoprecipitation using mAb 2-74. It was hoped that this experimental sequence would help clarify the relationships between the epitopes expressed on each CR2 isoform and indicate whether or not CR2 was expressed as a surface dimer.
The mAb 2-74 precipitated both CR2 isoforms after first labeling cells
with either 2-128 or 2-54, with a relative excess of CR2no,
confirming that these Abs labeled separate, noncompetitive CR2 epitopes
(Fig. 6D, lane 1). Significantly, in
comparison to the equivocal results of earlier experiments (Fig. 6C), mAb 2-128 now immunoprecipitated copious amounts
of both CR2 isoforms and the ratio of the signal strengths of the two
proteins approximated 1:1, indicating that these were expressed on the
cell surface as a noncovalently linked dimer (Fig. 6D,
lane 2, left panel).
The mAb 2-54 also immunoprecipitated both isoforms but with a relative
excess of CR2no (Fig. 6D, lane
2, right panel). After two more
immunoprecipitation cycles using either 2-128 or 2-54, respectively,
when very little additional target complex was removed (Fig. 6D, lanes 3 and 4),
the targets identified by mAb 2-74 had been quantitatively depleted by
mAb 2-54. In contrast, substantial levels of CR2no, but not
of CR2ub, remained after preclearing with mAb 2-128 (Fig. 6D, lane 5). Moreover, the
signal intensity produced by the residual CR2no
approximated the combined intensities of the CR2no and
CR2ub that were precipitated by mAb 2-128 (Fig. 6D, compare lanes 2 and 5,
left panel).
Collectively, the results of the experiments outlined above lead to the following conclusions. 1) The mAbs 2-54, 2-74, 2-87, and 2-128 recognize distinct external epitopes on a common CR2 molecular complex. The binding epitopes of the first three mAbs occur on both CR2no and CR2ub whereas the epitope of mAb 2-128 occurs only on CR2ub. 2) CR2 is expressed on the surface of sheep B cells as closely associated noncovalently bound CR2no-CR2ub heterodimers. However, there is no evidence that the binding properties of the mAbs are influenced by, or dependent upon, dimerization. 3) Other molecular forms comprised of CR2no unassociated with CR2ub also exist. The presence of the dimerization motif in the transmembrane region raises the possibility that these could be expressed as CR2no-CR2no homodimers, but this is not shown formally, and the existence of either monomers or higher multimers cannot be excluded. 4) The cell populations examined expressed both CR2 heterodimers and perhaps CR2 homodimers at more or less comparable levels, although the ratio of expression of these two molecular forms on individual B cells is not defined.
B cell polyubiquitin
Because many of the above findings about CR2 structure were
unexpected, and have not been reported to date in other species, we
wondered whether there was also something unusual about the structure
of sheep ubiquitin or C3d, since the first protein was implicated in
posttranslational changes to sheep CR2 and the second serves as its
specific ligand. A polyubiquitin transcript encoding four tandem
repeats of ubiquitin was amplified and cloned from B cell RNA using
conserved PCR primers. Not unexpectedly, since ubiquitin is among the
most conserved of proteins, the predicted aa sequences of sheep, human,
and cow ubiquitin differed at only 1 aa residue (not shown). Also,
similar to the experience with CR2, the NH2-terminal
cDNA sequence of ubiquitin matched that derived earlier by peptide
sequencing of gp190/CR2ub (Fig. 1D).
Primary structure of the sheep C3d ligand
Sheep C3dg was purified from blood plasma and the
NH2 terminus, a number of internal peptide fragments
were sequenced, and a cDNA clone encoding this fragment was amplified
from liver RNA, again using a strategy based on conserved PCR priming
sites (Fig. 7). The aa sequence of sheep
C3d is highly conserved and has 81 and 82% identity to the human (29)
and mouse (30) sequences, respectively. Functionally important motifs
such as the thioester site and a proposed CR2 binding site (31) are
conserved. NH2-terminal sequencing of elastase-generated
sheep C3dg confirmed that this enzyme cleavage site is also
functionally conserved (Fig. 7).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The selectivity of mAb 2-128 for CR2ub implies that the external topography of the two CR2 isoforms differs in some respect, but the structural basis for this remains unresolved. Both isoforms contained comparable amounts of N-linked carbohydrates, their peptide cores were similar, and we could not detect differential association with other cell surface molecules by using nondisruptive detergents during cell lysis. Perhaps the most likely possibility is that alternative splicing leads to protein differences not detected readily by peptide mapping, due to the tandemly repetitive nature of the SCR domains. Although there was no evidence for alternative splicing of CR2 transcripts among the cDNA clones sequenced, a more intensive search is needed to resolve this issue. The only unambiguous distinction between the two isoforms to this point is the covalent attachment of ubiquitin to the cytoplasmic domain of one and its absence from the other.
CR2 can therefore be added to a growing list of glycoprotein receptors that serve as substrates for ubiquitination while being anchored in plasma membranes (reviewed in Refs. 32–34). The mechanisms by which cell surface receptors become ubiquitinated, and the functional consequences, remain poorly understood, although these features clearly differ from those pertaining to cytosolic proteins, which become targeted for degradation in proteosomes after selective ubiquitination. In many cases surface receptors only become ubiquitinated after ligand binding, and in some instances this then signals internalization and traffic of the receptor to the endosomal compartment. However, other modalities may exist, including down-regulation of cell surface receptors through proteosome-mediated degradation of their cytosolic domains (reviewed in 34 . The indications that CR2ub is a topographically distinct receptor isoform that is expressed as a stable surface phenotype on sheep B cells are unusual properties for ubiquitinated proteins and suggest that ubiquitin could play more divergent roles in regulating the function of cell surface receptors.
Two-color FACS analysis and immunoprecipitation experiments provided
convincing evidence that the CR2ub isoform is expressed in
the B cell membrane as a CR2ub-CR2no
heterodimer. The 7-aa motif (LIxxxxxGVxxxxxxGVxT) in the transmembrane
region of sheep CR2 may provide the structural basis for this because a
similar motif in the glycophorin receptor specifically promotes
noncovalent dimerization of hydrophobic transmembrane -helices (35).
Although the linear spacing between the critical residues differs
between the two receptors (a one-turn -helical spacing in the
glycophorin receptor compared with a two-turn spacing for sheep CR2),
the secondary structural alignment is similar and places the key aa
residues at a helix-helix interface. Recent studies showed that these
7-aa residues contributed essentially all of the atoms that
participated in intermonomer van der Waals interactions during
dimerization (36). This does not exclude the possibility that other
parts of CR2 could also mediate interactions between the two isoforms
or between CR2 and other molecules. The aa comprising the motif are
present in human and mouse CR2, with a seemingly appropriate spacing,
although their precise order in the membrane does not match the motif
in either sheep CR2 or the glycophorin receptor. What this might mean
in a functional sense remains unknown.
Human CR2 is phosphorylated after appropriate surface stimulation of the receptor (6, 7), and studies in a cell-free system detected both phosphoserine and phosphotyrosine residues (37). Three serine/threonine residues occur at conserved positions in the cytoplasmic domains of sheep, human, and mouse CR2, and there are additional serine residues in each species at nonconserved sites. However, the sequence context of these residues does not fit closely to known substrate specificities of serine/threonine-kinases (38). In contrast, one of the four conserved tyrosine residues occurs in a motif (HLExxxD/EI/VY) that conforms very closely to known protein tyrosine-kinase substrates in that a potentially basic aa lies 7 residues to the NH2-terminal side of tyrosine, and, relative to the position of the tyrosine residue, two acidic residues occur among the intervening aa, particularly at position -2, and isoleucine or valine occur at position -1 (39, 40).
C3d binds to the two NH2-terminal SCR domains of human CR2
(41), and our analysis adds to other evidence showing that this ligand
has been extensively conserved in evolution. Judging from the sequences
now available, the primary structures of SCR1, SCR12, and the
transmembrane domain of CR2 have been significantly less well
conserved. The diversity in SCR1 suggests that during evolution, this
receptor has been subjected to selective pressures apart from C3d
binding and that these could differ between species. In this regard, it
was puzzling to us that irrespective of the cross-linking or detergent
conditions used during immunoprecipitation experiments, we never
obtained evidence for a CD19-like (95 kDa) molecule associated with
CR2. However, functional cross-linking experiments showed that both
sheep CR2 isoforms transmit activation signals to the cell cytoplasm
and that these synergize with the sIgM signaling pathway (our
unpublished observations). In human and mouse B cells, this effect is
mediated by CD19 (reviewed in 42 and this result therefore
implies the associated presence of CD19 or another signaling molecule
on sheep B cells. Unfortunately, no available Abs specifically
recognize sheep CD19, so it is difficult to properly resolve this
point.
In view of the diverse roles that CR2 plays in the immune system, there
are several ways in which posttranslational changes to the structure of
this receptor might affect its function. First, the dimerization of CR2
may be linked to the binding of C3d or other ligands such as IFN-
(43), CD23 (44), or alternate forms of C3 (4). Changes to the
configuration of CR2 during binding of one ligand may modulate the
affinity of the receptor for another ligand(s) that is bound
subsequently as part of a regulatory cascade. It remains conceivable
that dimerization of some ligands could also play a role, since a
conserved alternate form of C3 may form homodimers (4).
Secondly, as for some other receptors, helix association within the cell membrane might modulate transmembrane signaling of CR2, modify its association with other membrane-spanning molecules, or sort the subcellular localization of molecules interacting with the cytoplasmic domain of CR2 (see 35 . The net effect could be to adjust a built-in set-level that changes the threshold of stimulation needed via sIg to activate a B cell (42). A third possibility is that the two CR2 isoforms could be differentially internalized and then be targeted to distinctive cytosolic compartments. In trying to assess plausible functional implications, it may be pertinent to recall that the affinity of association between the sheep CR2 isoforms seems low, since the two component proteins were only immunoprecipitated jointly by mAbs with epitopes on both or by mAb 2-128 after chemical cross-linking. The wider spacing of the key residues in the dimerization motif may contribute to this, and it may facilitate the rapid membrane sorting of CR2 into higher-order structures.
We recently showed that the coexpression of CR2 and L-selectin by peripheral sheep B cells correlates with their recirculation behavior. A CR2+ L-selectin+ subset of B cells recirculated efficiently through lymph nodes and lymphatic vessels, whereas another subset that had down-regulated both of these markers migrated preferentially to the spleen (16). The expression of CR2ub-CR2no heterodimers on B cells within solid lymphoid tissues is more variable than the relatively uniform pattern seen on migratory cells in lymph or blood and subsets expressing either decreased or enhanced levels occur in some of them (Ref. 16 and our unpublished observations). A synthesis of the results of structural and in vivo studies suggests that selective ubiquitination of one CR2 isoform and changes in receptor topography and expression are likely to occur at checkpoints during B cell development and that coordinated changes in the expression of other molecules then regulates the tissue localization and migratory behavior of peripheral B cells.
Certain aspects of the higher-order structure of CR2 are conserved in mammalian phylogeny. The mAb 2-54, for example, specifically stains B cells in cows, goats, pigs, and humans while the others cross-react in some of these species. Also, some mAbs identifying CR2 in other animals cross-react in sheep (our unpublished observations). In addition to prototype C3d receptors, anti-CR2 Abs have identified 190-kDa proteins previously in lysates of mouse B cells (26, 27) and in transfected cell lines expressing human CR1 and CR2 (45). The larger protein on mouse B cells has been identified as an alternatively spliced product of the Cr2 gene that contains additional NH2-terminal sequences (46). It was proposed that the 190-kDa protein on human cells could be a cross-reacting glycosylation variant of CR1, although this was not confirmed (45). Three anti-CR2 mAbs precipitated two proteins (145 kDa and 180 kDa) from lysates of bovine B cells while a fourth mAb was specific for the 180-kDa molecule, although this protein also remains unidentified (47).
In summary, these experiments have identified CR2 as a natural substrate for ubiquitination and shown that ubiquitinated and conventional isoforms of CR2 are stably expressed as noncovalently associated heterodimers in the plasma membrane of sheep B cells. The attachment of ubiquitin to the cytoplasmic domain of one isoform of sheep CR2 correlates with the presence of specific external epitopes, implying that this isoform has a distinctive external topography, although the molecular basis for this effect remains unresolved. Our observations raise a question as to whether or not similar phenomena occur in B cells of other species and provide a basis from which the functional consequences of ubiquitination of CR2 can be examined.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 cDNA sequences reported in this paper are available from the GenBank data base under the following accession numbers: sheep complement receptor type 2, AF038131; sheep complement component C3dg, AF038130; sheep polyubiquitin, AF038129.
3 Address correspondence and reprint requests to Dr. W. R. Hein, AgResearch, Wallaceville Animal Research Centre, P.O. Box 40063, Upper Hutt, New Zealand. E-mail address:
4 Current address: Dr. Thor Landsverk, Department of Morphology, Genetics and Aquatic Biology, Norwegian College of Veterinary Medicine, Box 8146 Dep., 0033 Oslo 1, Norway.
5 Abbreviations used in this paper: CR2, complement receptor type 2 (also called CD21); CR2no, nonubiquitinated isoform of CR2; CR2ub, ubiquitinated isoform of CR2; FDC, follicular dendritic cell; IPP, ileal Peyer’s patches; SCR, short consensus repeat; sIg, surface Ig; aa, amino acid(s).
Received for publication January 8, 1998. Accepted for publication March 4, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
chain. J. Biol. Chem. 259:13857.-helices. Nat. Struct. Biol. 1:157.[Medline]
This article has been cited by other articles:
|
|
 |
|
  D. Liu, J.-Y. Zhu, and Z.-X. Niu Molecular Structure and Expression of Anthropic, Ovine, and Murine Forms of Complement Receptor Type 2 Clin. Vaccine Immunol., June 1, 2008; 15(6): 901 - 910. [Full Text] [PDF]
|
|
|
|
 |
|
 R. Heggebo, C. McL. Press, G. Gunnes, L. Gonzalez, and M. Jeffrey Distribution and accumulation of PrP in gut-associated and peripheral lymphoid tissue of scrapie-affected Suffolk sheep J. Gen. Virol., February 1, 2002; 83(2): 479 - 489. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Aleksandersen, K.-I. Lie, B. Gjerde, and T. Landsverk Lymphocyte Depletion in Ileal Peyer's Patch Follicles in Lambs Infected with Eimeria ovinoidalis Clin. Vaccine Immunol., January 1, 2002; 9(1): 83 - 91. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. J. Young, W. L. Marston, and L. Dudler Subset-Specific Regulation of the Lymphatic Exit of Recirculating Lymphocytes In Vivo J. Immunol., September 15, 2000; 165(6): 3168 - 3174. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Strous and R Govers The ubiquitin-proteasome system and endocytosis J. Cell Sci., January 5, 1999; 112(10): 1417 - 1423. [Abstract] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
