Download In vivo expression of alternatively spliced forms of integrin

Journal of Cell Science 108, 3419-3425 (1995)
Printed in Great Britain © The Company of Biologists Limited 1995
JCS8915
3419
In vivo expression of alternatively spliced forms of integrin-associated
protein (CD47)
Martina I. Reinhold1, Frederik P. Lindberg1,2,4, David Plas1, Stacy Reynolds1, Marion G. Peters1 and
Eric J. Brown1,2,3,*
Departments of 1Medicine, 2Molecular Microbiology, 3Cell Biology and Physiology, and 4Division of Infectious Diseases,
Washington University School of Medicine, 660 S. Euclid Ave., St Louis, MO 63110, USA
*Author for correspondence
SUMMARY
Integrin-associated protein (IAP) is a 50 kDa plasma
membrane protein physically and functionally associated
with β3 integrins in a variety of cells. IAP has an extracellular immunoglobulin domain, five transmembrane
domains and a short intracytoplasmic tail. IAP is recognized by anti-CD47 antibodies and is expressed on cells,
such as erythrocytes and lymphocytes, which do not
express β3 integrins. To learn more about potential
functions of IAP we examined its expression in vivo. Using
the polymerase chain reaction, we detected 4 alternatively
spliced forms of IAP which differ from each other only at
their intracytoplasmic carboxy termini. These alternatively
spliced forms are generated by inclusion or exclusion of
three short exons within 5 kb in the genome and are highly
conserved between mouse and man. There is tissue speci-
ficity of expression of the alternatively spliced forms of IAP
mRNA, with bone marrow-derived cells expressing predominantly one form and neural tissue another. Using
polyclonal antibodies which recognize the alternatively
spliced bone marrow (form 2) and neural (form 4) forms
of IAP, we found that in accord with the mRNA, form 2
protein was expressed in all tissues primarily on bone
marrow-derived cells and endothelia, while form 4 was
highly expressed in the brain and peripheral nervous
system. The evolutionary conservation of IAP isoforms and
their tissue-specific expression suggest an important role
for these intracytoplasmic domains in IAP function.
INTRODUCTION
for ovarian carcinoma cells (Campbell et al., 1992; Lindberg
et al., 1994). IAP is expressed at higher levels than β3 on
many cells. For example, IAP is present on lymphocytes,
which express little αvβ3, and on erythrocytes, which express
no known integrins (Brown et al., 1990). Thus, it is possible
that the functions of IAP extend beyond integrin signal transduction. Alternatively spliced forms of IAP mRNA have been
found in a human cell line (Campbell et al., 1992), but
whether alternative splicing occurs in vivo and if the alternatively spliced mRNAs are translated into different protein
products are unknown. To determine the tissue distribution
of IAP and its alternatively spliced isoforms, we have cloned
the alternatively spliced forms of murine and human IAP and
we have made antibodies which specifically recognize two of
the four alternatively spliced IAP protein products. Using
RT-PCR and these antibodies, we have evaluated the
expression of IAP in murine tissues. IAP is expressed in
essentially all tissues, but the four alternatively spliced forms
of IAP mRNA have distinct tissue distribution. This differential expression of alternatively spliced IAP mRNAs is paralleled by the tissue distribution of the distinct protein
isoforms. While the functions of IAP apart from its role in
integrin β3 signal transduction remain unknown, these data
suggest that there may be tissue-specific functions of
different IAP isoforms.
Integrin-associated protein (IAP) was first identified as a
membrane protein involved in signal transduction from β3
integrins on polymorphonuclear leukocytes (PMN) (Brown et
al., 1990). Antibodies to IAP blocked activation of PMN
phagocytosis, respiratory burst, and chemotaxis induced by
Arg-Gly-Asp-containing proteins and synthetic peptides.
Subsequently, antibodies to IAP have been shown to inhibit
ligand binding to αvβ3 on several cell lines and to block the
increase in [Ca2+]i which occurs upon endothelial cell
adhesion to fibronectin- or vitronectin-coated surfaces,
without disturbing cell adhesion to these surfaces (Schwartz
et al., 1993). These data suggest that IAP is an important
component of the mechanism by which ligation of β3
integrins transmits information to the cell. We have speculated that IAP plays a role in β3 integrin function analagous
to the nonligand binding chains of cytokine receptors,
affecting both ligand affinity and signal transduction
(Lindberg et al., 1993).
Molecular cloning of IAP cDNA revealed that it is a
multiply membrane-spanning protein with an amino-terminal
extracellular sequence consisting of a single IgV-like domain.
CD47 antibodies recognize human IAP, which also is
identical to the OV-3 antigen, initially thought to be specific
Key words: integrin, alternative splicing, immunoglobulin
superfamily, protein expression
3420 M. I. Reinhold and others
MATERIALS AND METHODS
of adult mice. PCR products were sequenced directly or cloned,
sequencing several clones derived from each PCR product.
Cloning of murine and human IAP cDNAs
Standard techniques were used for cloning and other nucleic acid
manipulations. Two potential alternatively spliced forms of human
IAP cDNAs were obtained from screening cDNA libraries from U937
cells and differentiated HL60 cells (Lindberg et al., 1993). To identify
additional cytoplasmic variants of human IAP, PCR primers were synthesized which span the putative splice site. The sense oligonucleotide
(5′GTTGGACTGAGTCTCTGTATTGCGGCGTGT3′) is located 5′
of the splice site, and the antisense oligonucleotide (5′CCCAAGCTTACACTTTCACGTCTTACTACTG3′) is located within the 3′
untranslated region (3′UT). These primers were used to screen clones
from the HL60 library. Among 27 clones, 25 (92%) gave fragments
of a size consistent with form 2 and 2 (8%) with form 4. No clones
corresponding to form 3 were detected. PCR reactions were done in
a final volume of 20 µl containing 50 mM Tris-HCl, pH 9.0, 50 mM
KCl, 0.1% Triton X-100, 0.2 mM each of dATP, dCTP, dGTP and
dTTP, 0.75 µM oligonucleotide primer and 1 U of Taq DNA polymerase (Promega); 1 µl of cDNA was subjected to 30 amplification
cycles under these conditions. Each cycle consisted of 30 seconds at
90°C, 30 seconds at 53°C and 30 seconds at 72°C. PCR products were
run on a 1.5% agarose gel and stained with 1 µg/ml ethidium bromide.
The human form 3 specific expression construct was generated by
PCR using site directed deletion with human form 4 as template. The
carboxy-terminal sequence was shown to be identical to the form 3
sequence previously published (Campbell et al., 1992). Human form
1 corresponds to one of the initially cloned splice variants of IAP
(Lindberg et al., 1993). Cytoplasmic variants of murine IAP clones
from a murine 70Z/3 B cell line were similarly screened by PCR using
the following primers: sense (5′ of splice site) 5′TCCTTGCTTTGGTCGGGCTGTGT3′; antisense (3′ UT region) 5′CTGACTTCCAAGTTACAGTCCGTCAC3′; 1 µl of cDNA was subjected to 30
amplification cycles under above described conditions. Each cycle
consisted of 30 seconds at 90°C, 30 seconds at 65°C and 30 seconds
at 72°C. Of 14 clones, 2 (14%) were of form 1, 8 (57%) were of form
2, and 4 (29%) were of form 3. Form 4 was cloned from brain tissue
A
Intron definition
Genomic mouse liver DNA was prepared from a SV129 mouse: 1 µl
of genomic DNA was subjected to 30 amplification cycles under conditions described above. Each cycle consisted of 30 seconds at 90°C,
30 seconds at 60°C and 4 minutes at 72°C. Introns were detected using
the following primer pairs.
Intron between form 1 and form 2: 5′GCTCTAGCAGAACTACTTGGA3′; 5′CTAGGAGGTTGGATAGT3′;
Intron between form 2 and form 3: 5′CAACCAGAGGACTATCCAACCTC3′; 5′TCGTTAAGGGGTTCCTCTACAGC 3′;
Intron between form 3 and form 4: 5′GCTGTAGAGGAACCCCTTAACG3′; 5′CGTCATTCATCATTCCTTTTGAC 3′;
Intron between form 4 and the 3′UT: 5′GTCAAAAGGAATGATGAATGACG 3′; CTGACTTCCAAGTTACAGTCCGTCAC 3′.
PCR products were run on a 1.5% agarose gel and stained with 1
µg/µl ethidium bromide.
Detection of mRNA
mRNA was prepared from adult mouse tissue and indicated cell lines
using a kit (Micro-Fast Track, Invitrogen). For PCR detection of
mRNA, cDNA was synthesized from poly(A) RNA using a kit (cDNA
Cycle Kit, Invitrogen) according to the instructions of the manufacturer. PCR reactions were done in a final volume of 20 µl containing
50 mM Tris-HCl, pH 9.0, 50 mM KCl, 0.1% Triton X-100, 0.2 mM
each of dATP, dCTP, dGTP and dTTP, 0.375 µM IAP oligonucleotide primer (of which 12.5 nM labeled with [γ-32P]ATP), 0.2 µM
Gαs (GTP binding protein) oligonucleotide primer (of which 12.5 nM
labeled with [γ-32P] ATP), and 2.5 U of Taq DNA polymerase
(Promega). Reverse transcription reactions (2 µl) were subjected to
20 amplification cycles under these conditions. Control experiments
showed that the dose-response of the PCR assay is linear over a range
of 1 to 0.02 µl of cDNA used (Fig. 1). Each cycle consisted of 30
seconds at 90°C, 30 seconds at 60°C, and 1 minute at 72°C. Alternative splice forms of IAP were detected using the following primers:
B
Fig. 1. Quantitative PCR assay for alternatively spliced IAP mRNAs. (A) PCR assays for IAP and for the αs subunit of the heterotrimeric G
protein Gs were performed on reverse transcribed cDNA from J774 cells, using different quantities of starting mRNA (lanes a-f: 1µl, 0.5 µl, 0.2
µl, 0.1 µl, 0.05 µl, 0.02 µl, respectively). The PCR assay, using 32P-labelled oligonucleotide primers, was performed as described in Materials
and Methods. The amplified DNA was quantitated using a Phosphorimager. (B) Under the conditions of the assay, both Gsα and IAP PCR
products increase as a function of input mRNA, and the ratio between the two PCR products is unaffected by the quantity of starting mRNA.
In vivo expression of IAP 3421
5′ of splice site, 5′CATCGTGGTTGTTGGAGCCATC3′; 3′UT
region, 5′ACAGTCCGTCACTTCCCTTCAC3′. This primer pair was
chosen since it eliminated non-specific priming when used with the
Gαs specific primers in the same reaction. The ubiquitous Gαs
message served as an internal control for the efficiency of cDNA
synthesis and PCR amplification and was detected using the following
primers: sense, 5′ATTGAAACCATTGTGGCCGCCATGAGC3′;
antisense, 5′GAAGACACGGCGGATGTTCTCAGTGTC 3′ (Russell
et al., 1991). No product was seen when genomic DNA without
reverse transcription was used. PCR products were run on 4% polyacrylamide gels and then exposed to X ray film.
Preparation of anti-cytoplasmic tail antibodies
To distinguish between different cytoplasmic tail splice variants, we
raised polyclonal rabbit antisera against peptides mimicking the
unique portions of the four cytoplasmic tails. An acetyl-cysteine (Acc) residue was added at the amino terminus of the peptides for
coupling to carrier proteins. Rabbit antisera were generated as previously described (Lindberg et al., 1993). Antisera against peptides
derived from the form 1 and form 3 peptides (Ac-C-QLLGLVYMKFVE and Ac-C-KAVEEPLNE, respectively) failed to recognize IAP
from cells transfected with constructs encoding the different forms of
IAP. The sera were not further characterized. Serum against the form
4 peptide (Ac-C-AFKESKGMMNDE) showed good reactivity
against human recombinant form 4 IAP. The antiserum agaist the
murine form 2 peptide (Ac-C-ASNQRTIQPPRNR) showed weaker
reactivity with both murine and human form 2 recombinant IAPs than
the previously described (Lindberg et al., 1993) serum raised against
the human equivalent peptide (Ac-C-ASNQKTIQPPRNN). Thus, the
sera raised against the form 4 and the human form 2 sequences were
used for further experiments.
In western blots of cells transfected with IAP cDNAs, both the form
2 and form 4 peptide antisera recognized a single band comigrating
with IAP (data not shown). Using 125I-labelled baculovirus produced
recombinant human IAP of forms 1, 2, and 4, the antisera specifically
immunoprecipitated the appropriate form of IAP (Fig. 2) from
solutions in PBS/1% Triton X-100. The antisera precipitated less IAP
than the control mAb 2D3. Presumably, this was due to losses during
200000
Activity Precipitated (cpm)
Antigen
150000
100000
50000
0
AA
IAP-1
IAP-2
IAP-4
A
AA
A
AA
A
AA
A
AA
AA
AA
A
AA AA AA A
Preimmune
anti-2
anti-4
anti-IgV
Antiserum
Fig. 2. Specificity of anti-peptide antibodies. Antibodies against the
IAP form 2 cytoplasmic peptide and the IAP form 4 extension,
prepared as described in Materials and Methods, were used to
immunoprecipitate IAP form 1, form 2, and form 4 proteins made in
Sf9 cells from recombinant baculovirus. The IAP form 2 antibody
immunoprecipitates form 2 IAP but not form 1 or 4; the IAP form 4
antibody immunoprecipitates only form 4.
washing of IAP bound to peptide antibodies of relatively low affinity.
The results show that the antisera were form-specific for immunoprecipitation, even though 11 of the 13 amino acids of the form 2
peptide used are contained within both the form 3 and 4 cytoplasmic
tails (see Fig. 3).
Immunohistochemistry of murine tissue
Immunohistochemical analyses were performed using the ABC
method for staining of cryostat sections. Briefly, cryostat sections of
frozen tissues were fixed in ethanol at 4°C. Sections were preincubated for 30 minutes in PBS (pH 7.2) containing 1% hydrogen
peroxide, then for 15 minutes in PBS containing 0.2% nonfat
powdered milk, 2% BSA and 0.3% Triton X-100. Sections were
incubated with antibodies overnight at 4°C, washed in PBS and then
incubated with 1:100 dilution of biotinylated goat anti-rabbit IgG
(Vector, Burlingame, CA) for 30 minutes. After washing in PBS, the
slides were incubated with streptavidin-peroxidase (Zymed, South
San Francisco, CA) for 10 minutes. After washing in PBS again, the
peroxidase activity was visualized with naphthol (Sigma, St Louis) or
aminoethyl carbazole substrate (AEC kit, Zymed). Binding of the
form-specific peptide antibodies was blocked completely by coincubation of the tissue sections with 50 µg/ml of peptide (not shown).
RESULTS
Alternatively spliced forms of human and murine
IAP mRNA
We found 4 distinct cDNAs for IAP from both human and
murine cDNA libraries (Fig. 3A). The four cDNAs for human
IAP are identical to those reported for the OV-3 antigen
(Campbell et al., 1992). The cDNAs differ from each other
by increasingly long extensions at the 3′ end of the coding
region. Since the carboxy terminus of the protein is intracytoplasmic (Lindberg et al., 1993), each of these cDNAs would
code for protein with a different length intracytoplasmic tail,
which we have for convenience termed forms 1-4 based on
increasing length (Fig. 3C). Each longer cytoplasmic tail
came from the addition of a short peptide sequence to the
extreme carboxy terminus of the protein. We found that each
of these additional peptide sequences was encoded by a short
exon. The three exons encoding the additional extensions of
forms 2-4 were 32 bp, 25 bp, and 36 bp, respectively (Fig.
3A). Each of these exons and the one encoding the form 1
cytoplasmic tail ended in a G in both the mouse and human
IAP gene. The common exon encoding the 3′ UT, which is
identical for each form, began AATAAC in the human gene
and AATAGG in the mouse. For forms 1, 3, and 4, the splice
with the 3′ G of the preceding exon led to a glutamate as the
carboxy-terminal amino acid of the predicted protein,
followed by a stop (TAA or TAG) codon. For form 2, the
frame was shifted and the final amino acids, NN in the human
and NR in the mouse, were encoded in the same exon as the
3′UT sequences prior to a new stop codon. Because the
cDNAs all contained identical 3′UT, we hypothesized that the
different coding regions represented alternative splicing of
the IAP gene. This was proven because these small exons
encoding the cytoplasmic tail peptide sequences were linked
to each other and to the 3′UT exon by PCR on genomic DNA
(Fig. 3B). The intron between the exon encoding the final
transmembrane domain and the 32 bp exon encoding the form
2 cytoplasmic tail is larger than 2.5 kb, so the exact distance
was not mapped by PCR. The distance between the form 2
3422 M. I. Reinhold and others
Fig. 3. Alternatively spliced forms of IAP. (A) The sequences of the human and murine
exons encoding IAP form 1; the form 2, form 3 and form 4 cytoplasmic tail extensions;
and the exon common to all four forms which contains the stop codons and the 3′UT
region are shown. (B) PCR was used as described in Materials and Methods to determine intron size between the exons shown in A. Lane a,
intron between form 1 and 2; lane b, intron between form 2 and 3; lane c, intron between form 3 and 4; lane d, intron between form 4 and 3′
UT. The intron between the common exon encoding form 1 and the exon encoding the form 2 peptide was not amplified by PCR and thus is
likely >2.5 kb. (C) A scheme illustrating the alternative splice possibilities for generation of the known forms of IAP. A common splice donor
from the 3′UT exon can join any of four potential splice acceptor sites.
exon and the next, encoding the form 3 extension is 1.5 kb;
between this exon and that encoding the form 4 extension,
0.5 kb; and between this last coding exon and the 3′UT region
2.5 kb. Thus, the different mRNAs arise from alternative
splicing of mRNA transcribed from a single IAP gene,
leading to 4 distinct cytoplasmic tails arising from the
sequential addition of peptide extensions each encoded by
short exons (Fig. 3C). If the lysine in form 1 is assumed to
be the stop transfer sequence beginning the carboxy-terminal
cytoplasmic tail, IAP form 1 has a 4 amino acid intracytoplasmic extension, encoded on the same exon as the final
transmembrane domain; form 2 a 16 amino acid extension;
form 3 a 23 amino acid extension; and form 4 a 34 amino
acid extension. The 4 intracytoplasmic extensions are highly
conserved between mouse and human. A conservative K to
R change in the form 2 exon peptide between mouse and
human is the only difference besides the N to R change at the
very carboxy terminus of form 2 which is encoded in the
3′UT exon and not reflected in the other forms. This is more
highly conserved than the extracellular IgV-like domain of
IAP (Lindberg et al., 1993). No consensus motifs for kinase
substrates or for protein-protein interactions were found in
any of the forms. No sequence variations were found in other
regions of IAP for either human or murine cDNAs.
Expression of IAP mRNA
To determine the tissue expression of IAP and its alternatively
spliced forms, we performed RT-PCR on murine tissue using
oligonucleotides which were present in all 4 forms of IAP and
which bracketed the region of the mRNA encoding the alternatively spliced cytoplasmic tails. Using this PCR strategy, the
4 forms could be distinguished on the basis of the size of the
PCR product (Fig. 4). PCR conditions were established which
allowed approximate quantitation of the mRNA for the
different IAP isoforms. All tissues tested expressed IAP
mRNA. However, the different forms were expressed at
varying levels in different tissues (Fig. 4). For example,
thymus and spleen expressed predominantly form 2 mRNA,
while brain expressed form 4 to a much greater degree than
any other form.
Primary murine cells, including myoblasts, macrophages,
embryonic stem cells, and microvascular endothelial cells
cultured in vitro expressed predominantly form 2 mRNA
(Table 1). When the endothelial cells were transformed with
In vivo expression of IAP 3423
Fig. 4. Tissue distribution of alternatively spliced
forms of IAP mRNA. mRNA was isolated from
various murine tissues and used for the quantitative
RT-PCR assay, as described in Materials and Methods,
with Gsα as a control in each assay. The sizes of the
four alternatively spliced forms, as determined by
identical PCR reactions on plasmids encoding each of
the four forms, is indicated.
middle T antigen (Dubois et al., 1991), IAP mRNA increased
in abundance and forms 1, 3, and 4 were found in addition to
form 2.
Several human primary cell types were examined for
expression of mRNA isoforms. Like murine thymus and
spleen, human monocytes, monocyte-derived macrophages,
and lymphocytes expressed predominantly form 2 mRNA
(Table 1). Activation of lymphocytes in an autologous mixed
lymphocyte reaction did not alter isoform expression (not
shown). However, in vitro culture of human monocytes for 7
days to allow maturation into macrophages (Bohnsack et al.,
1985) did alter IAP mRNA expression. Not only did mRNA
abundance increase, but all four forms could be found.
However, form 2 remained the predominant mRNA made in
these cells. Human keratinocytes expressed mRNA for IAP
forms 1 and 2. Keratinocytes were the only cell in which significant amounts of form 1 were found. Only form 2 mRNA
was detected in human umbilicial vein endothelial cells. We
also examined several continuous cell lines. The myelomonocytic leukemia cell line HL60 expressed form 2 mRNA, as did
the transitional cell carcinoma T24 and two hepatoma cell
lines. Interestingly, a nasal epithelial carcinoma cell line (2650)
expressed predominantly IAP form 4 mRNA (Table 1).
Immunolocalization of IAP protein
To determine whether the differential expression of IAP
mRNA was found as well in protein, we stained brain and
spleen with IAP form-specific antibodies. Spleen, in which
form 2 mRNA was the most abundant, stained strongly with
the form 2-specific antibody, but did not react with the form 4
antibody (Fig. 5). In contrast, while the anti-form 4 peptide
antibody stained brain cortex strongly, the form 2 antibody
showed minimal staining, consistent with the PCR data on
mRNA expression. The form 2 antibody stained endothelium
in a variety of tissues as well as bone marrow derived cells;
the form 4 antibody showed extensive staining in spinal cord
and peripheral nerves as well as throughout the brain (not
shown). Thus, the staining patterns demonstrate that the tissuespecific expression of alternatively spliced mRNA is reflected
in IAP protein.
Table 1. Expression of alternatively spliced forms of IAP
mRNA
IAP form
Cells/tissue
Form 1
Form 2
Form 3
Form 4
Human
1o keratinocytes
HUVEC
Monocytes
Macrophages
HepG2
Hep3B
HL60
HL60 + retinoic acid
T24 bladder carcinoma
2650 nasal carcinoma
++
+
+
-
++
++
+
++
(+)
+
+
+
++
+
+
+
-
+
+
++
Murine
Myoblasts
Macrophages
ES cells
1o endothelial cells
Transformed endothelial cells
Bladder
Salivary gland
Testis
Brain
Ovary
Intestine
Liver
Thymus
Spleen
+
++
-
++
+
+
++
+++
+++
+
+++
+++
+
++++
++++
+++
+
+
++
+
++
+
+
++++
++
+
+
+
++
+++
+++
-
mRNA level was determined by semi-quantitative PCR, as described in
Materials and Methods. The +-++++ scale covers a range of about 10-fold
difference in mRNA expression, as determined by quantitation of radioactive
bands by Phosphorimager analysis or by densitometry of autoradiograms. (+)
means mRNA detectable only after reamplification of original PCR product.
DISCUSSION
IAP is a plasma membrane molecule involved in β3 integrindependent signal transduction and ligand binding. However, its
precise function, especially on cells such as erythrocytes, lymphocytes and neurons, which do not express β3 integrins,
remains unknown. Although the existence of alternatively
3424 M. I. Reinhold and others
Fig. 5. Staining of cortex and spleen with form-specific
antibodies. Sections of mouse brain cortex and spleen were
stained with anti-form 2 (A and D); anti-form 4 (B and E);
and preimmune rabbit serum (C and F). Many of the cells in
the cortex are stained with the anti-form 4 antibody, but there
is very little staining with the form 2 antibody. In contrast,
spleen stains strongly with the form 2-specific antibody, but
very poorly with the anti-form 4 antibody. ×200.
spliced mRNAs for IAP has been reported, they had been
found only in a single carcinoma cell line (Campbell et al.,
1992) in which any role for IAP was unknown. It was unclear
whether untransformed cells made these various isoforms of
IAP mRNA, whether there were proteins made from these
variant mRNAs, or how their expression might be regulated.
Previously it has been shown that the tissue distribution of IAP
is wide, both by examining RNA from a variety of tissues
(Campbell et al., 1992) and by staining cells with anti-IAP
mAb (Brown et al., 1990). Recently, a systematic examination
of IAP expression in human tissue revealed staining in multiple
tissues including brain (Mawby et al., 1994).
The current work reports a systematic examination of the
expression of IAP and its isoforms in tissue. We have shown
that all 4 forms of IAP mRNA are found in both murine and
human cells, with a very high interspecies conservation with
respect to both peptide sequence of the alternatively spliced
regions and gene structure. All 4 forms are made by untransformed cells. As shown both by RT-PCR and immunohistochemistry, there is cell-type specificity of expression of the
different IAP isoforms. Form 2 and form 4 are the most
abundant IAP splice variants made in vivo. Form 2 is made by
all bone marrow-derived cells, by endothelial cells, and by
fibroblasts. Immunoprecipitation of a variety of bone-marrow
derived cells including erythrocytes, platelets, and lymphocytes with form-specific antibodies followed by extracellular
domain antibodies showed that essentially all the expressed
IAP is form 2 (data not shown). Much of the form 2 message
found in many tissues arises from endothelium, interstitial
fibroblasts, or resident lymphocytes and macrophages. Form 4
is expressed predominantly by neural tissue. Peripheral nerves,
spinal cord, and brain cells including retina all expressed form
4 IAP, the form with the longest cytoplasmic tail. Form 1,
which encodes an IAP without a significant cytoplasmic tail,
is the least abundant message in vivo. Unfortunately, since
there are no unique sequences within form 1, we were unable
to make an antibody specific for this form. Thus, we are unsure
whether the corresponding protein can be found in vivo. In
general, form 3 mRNA is made by many of the tissues which
make form 2, but is much less abundant. Exceptions to this rule
include the liver and the salivary gland. Because we have not
successfully made a form 3-specific antibody, we are unsure
of the cells expressing form 3 in each of these tissues.
However, studies with purified lymphocytes and macrophages
suggest that they can make form 3 IAP mRNA and that these
may well be among the cells which are responsible for the
synthesis of form 3 in tissue.
The preservation of sequence within the four cytoplasmic
tails between mouse and man and the tissue-restricted distribution of the various isoforms suggests that these cytoplasmic
tails play an essential role in IAP function. It is interesting to
speculate that each of the different extensions of the IAP cytoplasmic tail represents a modular unit designed to interact with
unidentified cytoplasmic or membrane proteins. Thus, IAP
might have different functions in different tissues, depending
on the nature of its cytoplasmic tail. It is possible to imagine
that the different tails could interact differently with cytoskeletal elements, integrins, or signal transduction cascades to
mediate different cell-specific functions. As yet, we have no
data as to the function of the IAP cytoplasmic tails. Interaction
of IAP with β3 receptors appears to require only the extracellular Ig domain of IAP (F.P.L. and E.J.B., unpublished). Thus,
it is more likely that the increase in [Ca2+]i which is dependent
on IAP (Schwartz et al., 1993), cell motility dependent on IAP
(Senior et al., 1992; Cooper et al., 1995) or other, as yet
unknown functions of IAP require its cytoplasmic domains.
In vivo expression of IAP 3425
REFERENCES
Bohnsack, J. F., O’Shea, J. J., Takahashi, T. and Brown, E. J. (1985).
Fibronectin-enhanced phagocytosis of an alternative pathway activator by
human culture-derived macrophages is mediated by the C4b/C3b
complement receptor (CR1). J. Immunol. 135, 2680-2686.
Brown, E. J., Hooper, L., Ho, T. and Gresham, H. D. (1990). Integrinassociated protein: a 50-kD plasma membrane antigen physically and
functionally associated with integrins. J. Cell Biol. 111, 2785-2794.
Campbell, I. G., Freemont, P. S., Foulkes, W. and Trowsdale, J. (1992). An
ovarian tumor marker with homology to vaccinia virus contains an IgV-like
region and multiple transmembrane domains. Cancer Res. 52, 5416-5420.
Cooper, D., Lindberg, F. P., Gamble, J. R., Brown, E. J. and Vadas, M. A.
(1995). The transendothelial migration of neutrophils involves integrin
associated protein (CD47). Proc. Nat. Acad. Sci. USA (in press).
Dubois, N. A., Kolpack, L. C., Wang, R., Azizkhan, R. G. and Bautch, V. L.
(1991). Isolation and characterization of an established endothelial cell line
from transgenic mouse hemangiomas. Exp. Cell Res. 196, 302-313.
Lindberg, F. P., Gresham, H. D., Schwarz, E. and Brown, E. J. (1993).
Molecular cloning of integrin-associated protein: an immunoglobulin family
member with multiple membrane spanning domains implicated in alpha-v,
beta-3-dependent ligand binding. J. Cell Biol. 123, 485-496.
Lindberg, F. P., Lublin, D. M., Telen, M. J., Veile, R. A., Miller, Y. E.,
Donis-Keller, H. and Brown, E. J. (1994). Rh-related antigen CD47 is the
signal-transducer integrin associated protein. J. Biol. Chem. 269, 1567-1570.
Mawby, W. J., Holmes, C. H., Anstee, D. J., Spring, F. A., Tanner, M. J. A.
(1994). Isolation and characterization of CD47 glycoprotein: a
multispanning membrane protein which is the same as integrin-associated
protein (IAP) and the ovarian tumor marker OA3. Biochem. J. 304, 525-530.
Russell, D. M., Dembic, Z., Morahan, G., Miller, J. F., Burki, K. and
Nemazee, D. (1991). Peripheral deletion of self-reactive B cells. Nature 354,
308-311.
Schwartz, M. A., Brown, E. J. and Fazeli, B. (1993). A 50 kDa integrinassociated protein is required for integrin-regulated calcium entry in
endothelial cells. J. Biol. Chem. 268, 19931-19934.
Senior, R. M., Gresham, H. D., Griffin, G. L., Brown, E. J. and Chung, A.
E. (1992). Entactin stimulates neutrophil adhesion and chemotaxis through
interactions between its Arg-Gly-Asp (RGD) domain and the leukocyte
response integrin (LRI). J. Clin. Invest. 90, 2251-2257.
(Received 14 March 1995 - Accepted 7 August 1995)