 |
INTRODUCTION |
Nerve growth factor
(NGF),1 a representative of
the neurotrophin family of growth factors, plays a critical role in the
differentiation, proliferation, survival, and cell death of neurons and
neural tumors (2-5). Neurotrophins act on two types of cell surface receptors: high affinity tyrosine kinase Trk receptors (including TrkA,
TrkB, and TrkC) and the low affinity neurotrophin receptor p75NTR. Trk receptors play a crucial role in neuronal
survival and differentiation, whereas p75NTR facilitates
ligand binding to the Trk receptors and is a regulator of cell death
(3). TrkA is a receptor for NGF and neurotrophin-6 (NT-6); TrkB is a
receptor for BDNF and NT4/5; TrkC is the primary receptor for NT-3 (6).
No interactions between NGF and TrkB or BDNF/NT-4/5 and TrkA are
observed. The p75NTR binds all known neurotrophins with
similar equilibrium binding kinetics (7).
Through Trk receptors, neurotrophins stimulate a variety of cellular
responses including survival, differentiation, proliferation, and
apoptosis, in different cell types, by pathways that involve the
activation of Ras, Rap1, Raf, MAPK/extracellular signal-regulated kinase, phospholipase C-
1 (PLC-1), Src, CHK, and
phosphatidylinositol 3-kinase (PI 3-kinase). Although PLC-1 and CHK
directly bind to the Trk receptors, the other pathways are activated
through signaling adapters such as Shc (8, 9), Grb2 (10), Gab1 (11),
rAPS, SH2-B (12), and FRS2 (10, 13). In the intracellular domain (ICD)
of the Trk receptors, there are five tyrosine sites that are
phosphorylated by ligand binding, i.e. Tyr499,
Tyr679, Tyr683, Tyr684, and
Tyr794 in rat TrkA (14, 15). Trk receptors phosphorylate
and activate signaling proteins through the recruitment of proteins to
these tyrosine residues in an active receptor complex. As shown in Fig. 2A, FRS2 and ShcA directly bind to rat TrkA at
Tyr499 (13) whereas the site of PLC-1 and CHK binding to
TrkA is Tyr794 (10). Interestingly, Grb2, in addition to
being indirectly recruited to Trk via multiple adapters such as Shc and
FRS2, also directly binds to the activation loop tyrosine residues,
Tyr683 and Tyr684, on TrkA (10). Receptors
mutated at individual phosphotyrosine, Tyr(P), sites required for
coupling Trk to specific signal transduction pathways have been a
useful tool to correlate changes in biological responses to specific
changes in Trk receptor protein: protein interactions. For example, it
has recently been reported that mutation of Tyr515 on TrkB,
the shared Tyr(P) residue involved in recruiting both Shc and FRS2 to
the Trk receptors (13), is essential for sympathetic neuronal survival
and local axon growth (16).
In an effort to isolate new signaling molecules utilized by the
neurotrophic Trk receptors, we have performed several yeast two-hybrid
screens using the ICD of TrkA, TrkB, and TrkC as baits. As a result, we
isolated a cDNA encoding human ShcB (also called Sck or Sli), a
member of the Shc adaptor family including ShcA, ShcB, and ShcC (also
called N-Shc). ShcB is highly expressed in both the peripheral and
central nervous systems (8, 9, 17). This is of particular relevance
because ShcA is weakly expressed in sympathetic neurons, and not at all
in sensory neurons in the peripheral nervous system (1, 18). By
comparison, ShcC is highly expressed in the central nervous system as
well as in the sympathetic and neural crest-derived, dorsal root
ganglia sensory neurons in the peripheral nervous system (1, 18).
Interestingly, ShcA is expressed maximally in neuronal stem cells in
the brain but is lost in post-mitotic neurons. In contrast, ShcC is not present in neural stem cells but is increasingly expressed during post-natal development (9), suggesting that ShcA and ShcC may support
different roles in the proliferation and differentiation of neurons.
There are three isoforms of ShcA (p46, p52, and p66), one isoform of
ShcB (p68), and two isoforms of ShcC (p55 and p69) generated through
differential splicing or translation initiation codons (17). All Shc
isoforms contain a C-terminal Src homology 2 (SH2) domain, a central
proline-rich region (CH1), and an N-terminal phosphotyrosine binding
(PTB) domain (19). The N-terminal extension of the longer isoforms
(also contained within the single ShcB gene) encodes a second
proline-rich collagen homology region (CH2). ShcA, ShcB, and ShcC share
high sequence identity in their PTB and SH2 domains (over 70%
identical in amino acid sequence) but in the CH1 and CH2 regions, the
sequence identity is less than 35% (1, 17, 18, 20). Binding of the
ShcA PTB domain to target sites is affected by residues N-terminal to
the Tyr(P) residue, whereas SH2 domain binding is modulated by amino
acids C-terminal to the Tyr(P) residue (8). These results indicate that
the two domains utilize different recognition sequences to regulate
protein-protein interactions and suggest that the Shc proteins bridge
together different tyrosine-phosphorylated intracellular proteins.
Moreover, phosphorylated ShcA activates the Ras/MAPK pathway through
Tyr(P) sites in the CH1 region and recruitment of Grb2 (21), suggesting
that each region of Shc has different biological roles.
In this study, we defined the molecular basis and specificity of
interaction between ShcB/ShcC and the Trk receptors. We identified that
ShcB and ShcC bind to TrkA at Tyr499 in a
Tyr(P)-dependent manner. In a similar way, ShcB and
ShcC bind to rat TrkB at Tyr515. The PTB domain of ShcB,
but not the SH2 domain, mediates binding to the Trk receptors. More
importantly, ShcB/ShcC expression in PC12-, TrkB-, or TrkC-expressing
cells supports NGF/BDNF/NT-3-dependent phosphorylation of
ShcB/ShcC, resulting in Grb2 binding and enhanced MAPK activation.
Furthermore, both ShcB and ShcC are activated in mouse cortical neurons
upon stimulation with BDNF and NT3, indicating that they are endogenous
targets of the Trk receptors in primary neurons. Interestingly,
additional tyrosine-phosphorylated proteins were identified in
immunoprecipitation reactions with anti-ShcB/ShcC antibodies in primary
cortical neurons, suggesting that these molecules are additional
downstream targets of neurotrophin-activated ShcB/ShcC in primary neurons.
| |
EXPERIMENTAL PROCEDURES |
Materials and Antibodies--
Mouse NGF (2.5 S) was purchased
from Harlan Bioproducts for Science. Recombinant BDNF and NT-3 were
gifts from A. A. Welcher (Amgen Inc). The anti-hemagglutinin (HA)
antibody, 3F10, was from Roche Molecular Biochemicals. Anti-c-Myc
antibodies were produced by a hybridoma, 9E10 (22), grown as ascites
tumors in 4-6-week-old female BALB/c mice. Rabbit antibodies to the
C-terminal 14 residues of TrkA (J203) were prepared using standard
techniques (13). Monoclonal anti-Trk antibodies (MCTrks) were purchased
from Santa Cruz Biotechnology. Rabbit antiserum to the CH1 domain of
ShcB (amino acids 310-477) (namely anti-ShcB GP) was the generous gift of Tony Pawson (Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, Canada) (8). Independently, we also made a rabbit
antiserum to a peptide of 15 amino acid residues (449-463) (8) from
ShcB (namely anti-ShcB) (Genemed Synthesis Inc., San Francisco, CA),
which was affinity-purified by standard procedures using the ShcB
peptide coupled to NHS-activated Sepharose (Amersham Biosciences). The
rabbit antibodies to Grb2, N-Shc (ShcC), and Fyn were purchased from
Santa Cruz Biotechnology. The mouse monoclonal antibody to Src (Mb327)
(23) was the gift of Joan S. Brugge (Department of Cell Biology,
Harvard Medical School, Boston, MA). The rabbit antibodies against
normal and activated MAPK were purchased from New England Biolabs and
Promega, respectively. The horseradish peroxidase-coupled
anti-phosphotyrosine antibody (RC20) was from Transduction Laboratories.
Cells and Transfections--
COS-7 African green monkey
kidney cells (24) were cultured in Dulbecco's modified Eagle's medium
(DMEM) (Invitrogen) supplemented with 10% fetal calf serum. For
transfection, a total of 4 × 105 cells were seeded
into 60-mm dishes. The next day, cells were re-fed and transfected with
2 µg each of indicated recombinant DNA constructs with LipofectAMINE
2000 (Invitrogen). After 24-48 h of transfection, untreated cells and
cells treated with neurotrophin (NGF, BDNF, or NT-3; 100 ng/ml, 5 min)
were lysed in Nonidet P-40 lysis buffer (1% Nonidet P-40, 137 mM NaCl, 10% glycerol, 2 mM EDTA (pH 8.0), 20 mM Tris (pH 8.0)) containing 1 mM sodium
orthovanadate, 1 mM phenylmethylsulfonyl fluoride, plus 2 µg/ml leupeptin and 10 µg/ml aprotinin.
PC12 rat adrenal pheochromocytoma cells (24) and the derived cell
lines, including nnr5 (14, 25), nnr5-TrkA (14, 25), nnr5-TrkB (25), and
nnr5-TrkC, were maintained in DMEM containing 5% fetal calf serum and
5% horse serum. nnr5 cells are nonresponsive to neurotrophin
stimulation. nnr5-TrkA and nnr5-TrkB cells stably express
~50-100-fold higher levels of TrkA and TrkB than PC12 cells,
respectively (10, 13). nnr5-TrkC cells were made by pcDNA-TrkC
(rat) transfection into nnr5 cells and selecting for stable expression.
These cells express ~10-30-fold higher levels of TrkC than levels of
TrkA on PC12 cells (data not shown). For transfection, cells were
seeded at the concentration of 6 × 105 cells/60-mm
dish and, the next day, cells were transfected as described above. Cell
lysates were prepared as described above.
High-five insect cells were cultured in Grace's complete medium
(Invitrogen) containing 10% fetal bovine serum (FBS).
Recombinant baculovirus expressing wild type TrkA, with a HA N-terminal
tag, was made as described previously (10, 13). Cells were infected with the recombinant baculoviruses (10, 26). After stimulation with 100 ng/ml NGF for 10 min, cells were lysed in Nonidet P-40 lysis buffer as
described above. Clarified lysates were assayed for total protein
(Bio-Rad Dc protein assay kit). The levels of receptor
expression were determined by Western blotting with the anti-HA
antibody, 3F10.
Expression Constructs--
The wild type rat TrkA receptor
cDNA was used as the template to make the following constructs. The
mutants of TrkA, including S3 (a deletion of
493IMENP497), S8 (Y499F), S9 (Y794F), and
S11 (kinase-inactive mutation, K547A), have been previously described
(10, 13, 14). Similar mutants in rat TrkB (accession number NM_012731)
were created using a similar strategy at the following sites: BS3 (a
deletion of 499VIENP513), BS8 (Y515F), BS9
(Y816F), and BS11 (kinase-inactive mutation, K573A). All wild type and
mutant receptors contain an N-terminal HA epitope and were cloned into
the mammalian expression vector, pCMX (10). All constructs were
generated by standard techniques and were confirmed by DNA sequencing
(Robarts Research Institute).
Wild type TrkA was also cloned into the baculoviral expression vectors,
pBacPAK8 (10), as described before (10, 13). The ICD of wild type and
mutant TrkA receptors (S3, S8, S9, and as described above) fused to the
yeast GAL4 activation domain in pAS2.1 (CLONTECH)
have been previously described (27). For yeast two-hybrid screening,
constructs containing the ICD region of rat wild type TrkB (amino acid
residues 458-821) and TrkC (amino acid residues 458-825) fused to the
yeast GAL4 activation domain were created in pAS2.1. The ICD of TrkB
was cut from pGEX-TrkB/ICD (10) with EcoRI and
SalI, and the excised 1.1-kb fragment was ligated into the
respective sites of pAS2.1. The ICD fragment of TrkC was cut from
pCMX-TrkC (10) with HindIII and XbaI and cloned
into the same sites of pSE280 (Invitrogen). As one SalI site
is in the polylinker of pSE280, a SalI fragment containing the ICD region of TrkC was ligated into the SalI site of
pAS2.1.
Full-length human ShcB cDNA without the N-terminal CH2 region
(His111 to Pro573, equivalent to the mouse ShcB
sequence in Ref. 8) was cut or amplified by PCR from pACT2 and ligated
into the pEBG mammalian expression vector (making GST fusion proteins
in mammalian cells; Ref. 10), pGEX4T2 (making GST fusion proteins in
bacteria; Amersham Biosciences), and pRK5, a Myc-tagged mammalian
vector (28). The mouse cDNA for the ShcC p55 isoform was amplified
from adult mouse brain RNA by PCR with primers corresponding to sense
(5'-GTCGACGAGTGCCACCAGGAAGAGCCG-3') and antisense
(5'-GGATCCTCAGGGTTTCCTCTCCACTGGTT-3') sequence reported in the
GenBankTM (accession number NM_009167). PCR products
were cut out with SalI and BamHI and then ligated
into appropriate sites of pRK5 generating an N-terminal fusion with
Myc. The ShcB and ShcC sequences were confirmed by DNA sequencing
(Robarts Research Institute).
The PTB and SH2 domains of ShcB were amplified by PCR using pACT2-ShcB
as a template. The PCR primers for the PTB domain were sense
(5'-GGAATTCTGGGGCCCGGGGTCTCCTA-3') and antisense
(5'-CCGCTCGAGCTCCTCGTCCCCCCAGGC-3'). The primers for the SH2 domain
were sense (5'-GGAATTCTGGTACCACGGCCGGATGA-3') and antisense
(5'-CCGCTCGAGCCGTGAGACCACGCCA-3'). Fusion vectors for GST-ShcB PTB and
GST-ShcB SH2 were constructed by cloning each restriction fragment into
appropriate sites of pGEX-4T (Amersham Biosciences). The sequence was
confirmed by DNA sequencing (Robarts Research Institute). The GST
fusion proteins for GST-ShcB full-length, GST-ShcB PTB, and GST-ShcB
SH2 were obtained by
isopropyl-1-thio-
-D-galactopyranoside induction of
Escherichia coli BL-21 cells harboring the corresponding constructs and purified from bacterial lysates with
glutathione-Sepharose according to the instructions of the manufacturer
(Amersham Biosciences).
Co-immunoprecipitation, SDS-PAGE, and Western Blot
Analysis--
For ShcB binding assays, COS cells were co-transfected
with GST-fused ShcB in pEBG-3 and HA-tagged wild type and mutant Trk receptors in pCMX. Cells were stimulated with NGF or BDNF at 100 ng/ml
for 5 min. Cell lysates, prepared in Nonidet P-40 buffer, were
precipitated by glutathione-Sepharose beads. GST-bound proteins were
resolved on a SDS-polyacrylamide gel and transferred to Immobilon-P. Western blotting was performed with anti-HA 3F10 antibody (1:3000 dilution) as described before (13). For ShcC binding assays, COS cells
were co-transfected with Myc-tagged p55 ShcC in the pRK5 vector and
HA-tagged wild type and mutant Trk receptors in pCMX. Cells were
stimulated as described above. Cell lysates, prepared in Nonidet P-40
buffer, were precipitated by 2 µg of anti-Myc antibody, 9E10. The
bound proteins were resolved on a SDS-polyacrylamide gel, and Western
blotting was performed with anti-HA 3F10 antibody.
To define the region of ShcB that mediates binding to TrkA, equal
amounts (~5 µg) of purified GST fusion proteins (including GST,
GST-ShcB full-length, GST-ShcB PTB, GST-ShcB SH2, and GST-ShcA PTB)
were mixed with 300 µg of insect cell lysates containing baculovirus
expressed wild type TrkA (10). After incubation overnight at 4 °C,
beads were washed twice with Nonidet P-40 lysis buffer and once with
cold phosphate-buffered saline. The bound proteins precipitated by
glutathione-Sepharose beads were resolved on SDS-PAGE followed by
Western blotting with 3F10 antibody (1:3000). Purified GST and GST-ShcA
PTB served as the negative and positive controls, respectively.
For tyrosine phosphorylation assays, the transfected COS cells were
stimulated with neurotrophin (NGF, BDNF, or NT-3 at 100 ng/ml for 5 min) and lysed in Nonidet P-40 lysis buffer containing 1 mM
sodium orthovanadate. Equal amounts of protein (200 µg) were immunoprecipitated with 9E10 anti-c-Myc antibodies (2 µg). After washing, the bound proteins were resolved on a 10% SDS-polyacrylamide gel and transferred to an Immobilon-P membrane. Proteins were visualized by an anti-phosphotyrosine antibody, RC20 (1:2000).
For MAPK assays, Myc-ShcB and Myc-ShcC were overexpressed in PC12 and
nnr5-TrkC cells. After 5-7 days of expression, cells were washed
extensively by serum-free medium and then stimulated with NGF (50 ng/ml, 2 min) and NT-3 (50 ng/ml, 3 min), respectively. Cell lysates
were prepared, and 50 µg of proteins were analyzed by 10% SDS-PAGE.
MAPK activation was detected by using an anti-active MAPK antibody
(1:5,000) (Promega) and horseradish peroxidase anti-rabbit antibody
(1:10,000). The membrane was stripped as described below and then
re-blotted with the anti-normal MAPK antibody (1:10,000) (Cell
Signaling Technology). Blots were analyzed by densitometric scanning
(Bio-Rad, model GS-700) and were normalized for differences in protein
loading per lane. Data shown in Fig. 8 (C and D)
represent the mean values ± S.D. of three to five experiments
performed independently. The enhancement of MAPK in ShcB/ShcC
overexpression versus vector alone was compared using
Student's t test. The results are shown as
p < 0.05.
When indicated, membranes were stripped at 45 °C for 15 min in 62.5 mM Tris-HCl (pH 6.8) containing 1% SDS, 100 mM
-mercaptoethanol and reprobed with the indicated antibodies.
Yeast Two-hybrid System--
The yeast strain PJ69-4A
(MAT
, trp1-901, leu2-3, 112, ura3-52, his3-200, gal4
gal80, LYS2::GAL1-HIS3, GAL2-ADE2,
met2::GAL7-lacZ) was a gift from Philip James
(University of Wisconsin, Madison, WI) (29). All yeast media
were from CLONTECH. 3-Aminotriazole (3-AT) was from ICN.
For yeast two-hybrid screening, the ICDs of wild type TrkB and TrkC,
fused to the GAL4 DNA-binding domain in the pAS2 vector (Trp marker),
were used as bait plasmids. Transformed PJ69-4A bait strains were then
co-transformed with a human mammary gland Matchmaker activation domain
containing cDNA library (Leu Marker) (CLONTECH). Transformed yeast cultures were plated
onto sc medium minus Leu, Trp, and His containing 20 mM
3-AT (ICN). After 5 days of culture at 30 °C, positive interactors
were picked to grid plates containing sc medium minus histidine,
adenine, leucine, and tryptophan (HALT) to eliminate false positives.
To drop out the binding domain plasmid, yeast clones were grown in
media minus leucine for at least 3 days. The yeast DNA was then
isolated by standard procedure and transformed into E. coli
HB101 strain by electroporation using standard procedure (10). To
further select for activation domain-containing plasmids, colonies were
allowed to grow on M9 media (15) lacking leucine. E. coli
DNA was then prepared for DNA sequencing (Robarts Sequencing Facility).
A total of 2 × 106 clones were screened.
For binding assays in yeast, the ICDs of wild type and mutant TrkA (S3,
S8, S9, and S11), fused to the GAL4 DNA-binding domain in pAS2, were
used as baits. Transformed PJ69-4A bait strains were then
co-transformed with ShcB in the pACT2 vector
(CLONTECH). Cultures were plated on media minus Leu
and Trp (LT), and growing colonies were cultured overnight in liquid
LT. To select for colonies expressing the HIS and
ADE reporters, where ShcB interacts with TrkA, cultures were
then grown on HALT with 40 mM 3-AT. The pACT2 vector alone
served as the negative control. For positive controls, pACT2-ShcA and
Grb2 constructs were used because they interact well with the Trk
receptors (10).
Cortical Neurons--
Cortical neurons were isolated from
embryonic day 16 CD1 mice using a standard procedure (30) with minor
modifications. Briefly, embryo forebrains were dissected in Hanks'
balanced salt solution (HBSS) (Invitrogen). Pieces of cerebral cortices
were collected and mechanically triturated in 2 ml of HBSS plus 10% FBS by passing through wide and heat-narrowed Pasteur pipettes 15 times. Dissociated cells were passed through a low binding 70 µM cell strainer (Falcon) and then washed in 10 ml of
HBSS plus 10% FBS. Live cells were counted by trypan blue exclusion assay in a hemocytometer. A total of 3 × 106 cells in
media containing 54% neurobasal medium, 36% DMEM/F-12, and 10% FBS
were plated onto 60-mm dishes precoated with 100 µg/ml poly-D-lysine (Sigma). Cells were allowed to attach for
1 h. Cells were extensively washed with HBSS and then cultured in
media containing neurobasal medium (60%) plus DMEM/F-12 (40%), 15 mM Hepes (pH 7.4) supplemented with glucose (6 g/liter),
insulin (10 mg/liter), apotransferrin (20 mg/liter), putrescine (61 µM), progesterone (20 µM), and sodium
selenite (30 nM) (all from Sigma), for 4 days at 37 °C.
The medium was changed every 2 days. One hour prior to cell harvest,
endogenous neurotrophins were removed by washing twice with medium.
After stimulation with medium alone or 100 ng/ml NGF, BDNF, or NT-3 (5 min), cells were washed with cold phosphate-buffered saline containing
1 mM sodium orthovanadate and then lysed in ice-cold
Nonidet P-40 lysis buffer containing 1 mM sodium
orthovanadate as described above. Lysates containing 400 µg of
protein were immunoprecipitated with anti-ShcB GP antibody (1:100) (8)
and anti-N-Shc antibody (1:100) (Santa Cruz Biotechnology). The
immunoprecipitated proteins were visualized with the anti-Tyr(P) antibody RC20. The membranes were stripped at 45 °C for 15 min in
stripping buffer as described above and then re-blotted with anti-ShcB
(1:2000), anti-N-Shc (1:500), and anti-Src (0.1 µg/ml). To test the
specificity of ShcB-Src interaction in cortical neurons, lysates (400 µg) were also immunoprecipitated with purified normal rabbit IgG (2 µg/ml), anti-ShcB GP (1:100), preimmune serum (1:100), anti-ShcB
(1:100), and anti-Src (2 µg/ml). After resolving by SDS-PAGE,
proteins were visualized with the anti-Src antibody (1 µg/ml). To
test the phosphorylation of Src and Fyn in cortical neurons, lysates
containing 400 µg of protein were immunoprecipitated with anti-Src (2 µg/ml) and anti-Fyn antibody (1:100) (Santa Cruz Biotechnology). The
immunoprecipitated proteins were visualized with the anti-Tyr(P)
antibody RC20.
To test ShcB/ShcC-Trk interaction in cortical neurons, lysates (1 mg)
were immunoprecipitated with the anti-ShcB (2 µg/ml) (purified) or
anti-Trk (J203; 1:250) antibodies. The precipitated proteins were
resolved in SDS-PAGE and visualized with the anti-Trk (MCTrk) antibody
(1:200; Santa Cruz Biotechnology) or anti-N-Shc (1:100; Santa Cruz
Biotechnology). To detect the Trk activation by neurotrophin
stimulations, the same lysates (200 µg) were immunoprecipitated with
anti-Trk antibody J203 (1:250) and then blotted with the anti-Tyr(P)
antibody RC20 (1:2000).
| |
RESULTS |
Isolation and Phosphorylation of the ShcB Adapter by the Trk
Receptors--
To identify additional binding partners of the
neurotrophin Trk receptors, we have performed several yeast two-hybrid
screens using the ICDs of the Trk receptors as baits and commercially available cDNA libraries. In this report, we characterize the isolation of a clone bearing the ShcB cDNA from a human mammary library (CLONTECH) in the pACT2 vector that showed
strong interaction with the ICD of TrkC (data not shown). The human
ShcB cDNA without the CH2 region (His111 to
Pro573, equivalent to the mouse ShcB sequence in Ref. 8)
was excised or amplified by PCR from the pACT2 cDNA plasmid,
subcloned, and used in the following experiments. Although the
isolation of this clone was not unexpected, because ShcA is well known
to interact with the Trk receptors, it is highly relevant because the
ShcB and ShcC adapters are more highly expressed than ShcA in the
nervous system (1, 18). Thus, this provided an opportunity to
investigate whether there are differences in the interaction and/or
activation of the different Shc adapters by the Trk receptors,
particularly in light of a previous report that ShcB did not interact
strongly with or become strongly activated by either TrkA or TrkB
(1).
Because Trk activation results in tyrosine phosphorylation of target
proteins, we first determined whether ShcB was tyrosine-phosphorylated by TrkA, TrkB, and TrkC stimulation. Myc-tagged ShcB and HA-tagged Trk
constructs were co-transfected into COS cells, and lysates were
immunoprecipitated with the anti-Myc antibody, 9E10, followed by
Western blotting with an anti-phosphotyrosine antibody, RC20. As shown
in Fig. 1, ShcB was phosphorylated by
TrkA, TrkB, and TrkC (Fig. 1A, upper
panel). Moreover, Western blotting with anti-HA antibodies,
3F10, demonstrates that all three Trks co-immunoprecipitate with ShcB
to similar levels (Fig. 1A, middle
panel). The doublet Trk bands in Fig. 1 (A and
B), and other figures described below, indicate different
glycosylation states of the receptors (10, 26). The lower
panel indicates that ShcB was expressed to similar levels in
each transfection (Fig. 1A, bottom
panel). These results suggest that ShcB can be
phosphorylated by TrkA, TrkB, or TrkC and, more importantly, that ShcB
can interact with all three Trk receptors in transfected COS cells.
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 1.
ShcB is phosphorylated by Trk
activation. A, COS cells expressing Myc-ShcB and HA-Trk
or Myc-ShcB alone were stimulated with neurotrophin (NGF for TrkA, BDNF
for TrkB, and NT-3 for TrkC, 100 ng/ml) for 5 min and cell lysates were
prepared. Immunoprecipitated (IP) proteins with 9E10
anti-Myc antibody were immunoblotted with anti-Tyr(P) (RC20), followed
by stripping and re-blotting with anti-Myc antibody. For analysis of
Trk expression, immunoprecipitated lysates were also blotted with the
anti-HA antibody (3F10). Arrows indicate the position of
ShcB and the Trks. B, Myc-ShcB and wild type (wt)
or kinase-inactive (S11) HA-TrkA were co-transfected in COS
cells. After stimulation with NGF or no treatment, cell lysates were
prepared. Immunoprecipitated proteins with the anti-ShcB antibody were
immunoblotted with anti-Tyr(P) (RC20), and immunoprecipitated proteins
with anti-Trk antibody were immunoblotted with anti-HA. The position of
ShcB and TrkA is indicated. C, GST-ShcB and HA-Trk were
co-expressed in COS cells and cell lysates were prepared after
stimulation with neurotrophins. GST-ShcB-bound proteins precipitated
(Ppt) with glutathione-Sepharose were immunoblotted with
anti-HA 3F10 antibody, followed by stripping and re-blotting with the
anti-GST antibody. Cell lysates were also immunoblotted with the
anti-HA 3F10 antibody. The position of ShcB and the Trks is indicated
by arrows.
|
|
To further confirm the specificity of ShcB phosphorylation by Trk
activation, we co-expressed ShcB with wild type and kinase-inactive TrkA in COS cells. As expected, co-transfection with kinase-inactive TrkA completely diminished phosphorylation of ShcB (Fig. 1B,
lane 4). ShcB could be phosphorylated in COS cells
constitutively when co-expressed with wild type TrkA (Fig.
1B, lane 2); however, additional NGF stimulation
did significantly enhance the phosphorylation of ShcB (Fig.
1B, lane 3). These results further demonstrated that ShcB specifically interacts with kinase-active TrkA in COS cells.
To further define the ability of ShcB to bind to the Trk receptors, we
employed GST fusion techniques. ShcB was fused with GST in a mammalian
expression vector, pEBG3. GST-ShcB and HA-Trks were then co-expressed
in COS cells. After stimulation with neurotrophins, cells were lysed
and GST-ShcB was precipitated with glutathione-Sepharose. Co-precipitated Trk receptors were visualized with anti-HA antibody 3F10. As shown in Fig. 1C, no visible band was seen when
GST-ShcB was expressed alone (top panel,
lane 1), whereas bands corresponding to TrkA, TrkB, and TrkC
were detected in co-transfected cell lysates. Stripping and re-blotting
with anti-GST antibodies indicate similar levels of GST-ShcB expression
in all lanes (Fig. 1C, middle panel). Similarly, Western blot analysis with anti-HA antibodies showed similar
expression levels of the Trk receptors in whole cell lysates (Fig.
1C, lower panel). These results,
similar to that shown in Fig. 1A, indicate that ShcB binds
to TrkA, TrkB, or TrkC in co-transfected COS cells.
ShcB and ShcC Bind to TrkA at Tyr499--
Five
tyrosine residues in the ICD of TrkA are phosphorylated and activated
upon neurotrophin stimulation, i.e. Tyr499,
Tyr679, Tyr683, Tyr684, and
Tyr794 in rat TrkA (Fig.
2A) (10, 13). ShcA and FRS2
bind to TrkA at Tyr499, whereas PLC-1 and CHK bind to
TrkA at Tyr794 (10, 13, 31-33). To explore the site of
ShcB-TrkA interaction, we employed several mutant Trk constructs,
specifically, S8 (Y499F) and S9 (Y794F), which have been shown to
affect ShcA/FRS2 and PLC-1/CHK binding, respectively (10, 13, 16).
GST-ShcB and HA-tagged TrkA mutants were co-transfected into COS cells, and GST-precipitated proteins were assayed with anti-HA antibodies. As
shown in Fig. 2B, GST-ShcB bound HA-tagged wild type TrkA
and the S9 mutant, respectively (Fig. 2B, top
panel, lanes 2 and 5). The S8 mutation
significantly reduced GST-ShcB binding to TrkA (Fig. 2B,
top panel, lane 4). Similarly, the S3
mutation, which is a deletion of 493IMENP497 in
rat TrkA, and has previously been shown to affect ShcA and FRS2 binding
to TrkA (13), also showed a similar reduction in binding ability of
ShcB (Fig. 2B, top panel, lane
3). More importantly, the kinase-inactive mutation (S11, K547A)
completely blocked the binding of ShcB to TrkA (Fig. 2B,
top panel, lane 7). A combination mutation of S8 and S9 also significantly reduced the binding of ShcB to
TrkA (Fig. 2B, top panel, lane
6), which confirmed that Tyr499 is the binding site of
ShcB to TrkA. Stripping and re-probing with anti-GST antibodies
detected similar levels of GST-ShcB expression in all lanes (Fig.
2B, middle panel). Similarly, Western
blotting with anti-HA antibodies showed similar expression levels of
the TrkA receptors in whole lysates of each co-transfection (Fig. 2B, lower panel). Taken together,
these results indicate that ShcB can bind to TrkA at Tyr499
in a phosphotyrosine-dependent manner.
View larger version (47K):
[in this window]
[in a new window]
|
Fig. 2.
Effects of rat TrkA mutations on ShcB
binding. A, schematic of rat TrkA mutants including S3
(493IMENP497), S8 (Y499F), S9 (Y794F), and
S11 (K547A) in the ICD. Shc and FRS2 bind Tyr499, and
PLC-1 and CHK bind Tyr794. ATP binding site in rat Trk
is Lys547. Single letters are used to
indicate amino acids. B, COS cells expressing GST-ShcB and
HA-Trk or GST-ShcB alone were stimulated with NGF (100 ng/ml) for 5 min
and cell lysates were prepared. GST-ShcB-bound proteins were pelleted
with glutathione-Sepharose, followed by immunoblotting with the
anti-HA 3F10 antibody. Membranes were stripped and then re-probed
with the anti-GST antibody as indicated. For analysis of Trk
expression, cell lysates (50 µg) were blotted with the anti-HA 3F10
antibody. Arrows indicate the position of ShcB and the Trks.
wt, wild type; Ppt, precipitated.
|
|
As mentioned above, the Shc family contains three members,
i.e. ShcA, ShcB, and ShcC, that share a high degree of
identity at the amino acid level in the PTB and SH2 domains (8, 17). To
determine whether the ShcC adapter also binds to TrkA at the same site,
Tyr499, we co-transfected COS cells with a Myc-tagged ShcC
and HA-tagged wild type and mutant TrkA receptors. Lysates prepared
24 h after transfection were used for immunoprecipitation with the
anti-Myc antibody, 9E10. Co-immunoprecipitated proteins were identified by Western blotting with the anti-HA antibody, 3F10. As shown in Fig.
3, ShcC precipitated wild type TrkA
strongly and the S9 mutation (Y794F) did not reduce binding, whereas
the S8 (Y499F) and S3 (493IMENP497)
mutations significantly reduced, and the S11 (K547A) mutation completely blocked, ShcC binding to TrkA. Thus, the binding pattern of
ShcC to TrkA is the same as observed for ShcA and ShcB.
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 3.
ShcC binds to rat TrkA at
Tyr499. COS cells co-expressing Myc-ShcC with HA-TrkA
or Myc-ShcC alone were stimulated with NGF (100 ng/ml) for 5 min.
Lysates were immunoprecipitated (IP) with the anti-Myc
antibody (9E10) and then immunoblotted with the anti-HA (3F10)
antibody. Cell lysates were immunoblotted with the anti-HA antibody.
The position of the Trks is indicated by arrows.
wt, wild type.
|
|
ShcB and ShcC Bind to TrkB at Tyr515--
The ICD
region is highly conserved among TrkA, TrkB, and TrkC receptors (34,
35). Thus, we hypothesized that the interaction between ShcB and TrkB
would involve the homologous residue, Tyr515, in rat TrkB,
which is also involved in ShcA binding (16). To clearly define this, we
made similar mutations in the rat TrkB receptor, namely BS3
(499VIENP513), BS8 (Y515F), BS9 (Y816F), and
BS11 (kinase-inactive mutation, K573A). GST-ShcB was co-transfected
with HA-tagged wild type or mutant TrkB receptors in COS cells. Cell
lysates were precipitated with glutathione-Sepharose and then were
analyzed by Western blotting with anti-HA antibodies. As expected, the
Y816F (BS9) mutation did not affect ShcB binding to TrkB (Fig.
4A, compare lane 2 with lane 5), whereas the kinase-inactive mutation, K573A
(BS11), completely blocked the interaction of ShcB with TrkB (Fig.
4A, lane 6). Similarly, Y515F (BS8) significantly
reduced the binding of ShcB to TrkB (Fig. 4A, lane
4). As described for TrkA, a deletion mutation of
499VIENP513 also reduced ShcB binding to TrkB
(Fig. 4A, lane 3). These results suggest that the
homologous residue, Tyr515, in rat TrkB is involved in the
interaction between ShcB and TrkB.
View larger version (48K):
[in this window]
[in a new window]
|
Fig. 4.
ShcB and ShcC bind to rat TrkB at
Tyr515. A, COS cells co-expressing GST-ShcB
with HA-TrkB (wild type and TrkB mutants) or GST-ShcB alone were
stimulated with BDNF (100 ng/ml) for 5 min. Lysates were precipitated
with glutathione-Sepharose and Western blots performed with the anti-HA
antibody. The blot was stripped and re-probed with the anti-GST
antibody. Cell lysates were immunoblotted with the anti-HA antibody.
Arrows indicate the position of ShcB and the TrkB.
B, COS cells co-transfected with Myc-ShcC and HA-TrkB (wild
type and TrkB mutants) or Myc-ShcC alone were stimulated with BDNF (100 ng/ml) for 5 min. Lysates were immunoprecipitated (IP)with
the anti-Myc 9E10 antibody and Western blots performed with the anti-HA
antibody. Cell lysates were also probed with the anti-HA antibody.
Arrows indicate the position of ShcB and the TrkB. TrkB
mutations including BS3 (a deletion of
499VIENP513), BS8 (Y515F), BS9 (Y816F), and
BS11 (kinase-inactive mutation, K573A) were used. wt, wild
type; Ppt, precipitated.
|
|
Using the same approach, we determined the site of ShcC binding to
TrkB. Accordingly, Myc-tagged ShcC and HA-tagged TrkB were co-transfected into COS cells. Cell lysates prepared 24 h after transfection were immunoprecipitated with anti-Myc antibody 9E10 and
then assayed by Western blotting with anti-HA antibody 3F10. Similar to
results described above, the Y816F (BS9) mutation did not affect ShcC
binding to TrkB (Fig. 4B, compare lane 2 with lane 5), whereas the kinase-inactive mutation, K573A (BS11),
completely blocked the interaction of ShcC with TrkB (Fig.
4B, lane 6). Moreover, both Y515F (BS8) and a
deletion mutation of 499VIENP513 (BS3)
significantly reduced ShcC binding to TrkB (Fig. 4B,
lanes 3 and 4). Thus, ShcB and ShcC both bind
TrkB receptor at Tyr515 in a
phosphotyrosine-dependent manner.
ShcB Interacts with TrkA in Yeast--
To further confirm the
in vivo binding ability of ShcB to TrkA, we used the yeast
two-hybrid system. ShcB fused with the GAL4 activation domain in the
pACT2 vector (containing the Trp marker) and ICD of wild type and
mutant TrkA receptor fused to the GAL4 DNA-binding domain in the pAS2
vector (containing the Leu marker) were co-transformed into the yeast
strain PJ69-4A (29). For controls, ShcA and Grb2, which have been
shown to bind both S8 and S9 mutants of TrkA (10), were also included.
Cultures were plated on LT. Yeast without any vector did not grow,
whereas those containing both pACT2 and pAS2 grew effectively (Fig.
5A). Colonies growing in LT
were then plated on HALT plates containing 3-AT (40 mM)
(Fig. 5B). Cultures containing an empty pACT2 vector died. In contrast, colonies co-transformed with ShcB and wild type TrkA grew
well in HALT, indicating that ShcB-TrkA interact well in yeast. The S9
mutation did not affect the interaction between ShcB and TrkA. As
expected, both the S3 and S8 mutations significantly reduced ShcB
interaction with TrkA, and the S11 mutation failed to interact with
ShcB. These results confirm that ShcB binds to TrkA at
Tyr499 in a phosphotyrosine-dependent manner
in vivo in yeast.
View larger version (45K):
[in this window]
[in a new window]
|
Fig. 5.
TrkA mutations affect ShcB binding in
yeast. The ICDs of wild type and mutant (S3
(493IMENP497), S8 (Y499F), S9 (Y794F), and
S11 (K547A)) TrkA receptors, in pAS2, were transfected or
co-transfected with ShcB, in pACT2, into yeast strain PJ694A. For
controls, ShcA and Grb2 were also included. Cells were plated on LT
(A) and sc medium minus Leu, Trp, His, and Ade plus 40 mM 3-AT (HALT) (B). WT,
wild type; Non, not cotransfected with pAct2.
|
|
The PTB Domain of ShcB Mediates Binding to TrkA--
As mentioned
above, it has been reported that the PTB domain of ShcA and ShcC
mediates binding to TrkA (18, 36, 37). To determine whether the same
region of ShcB also binds the Trk receptors, we made GST fusion
proteins for the PTB and SH2 domains of ShcB. High levels of TrkA were
expressed by baculovirus in Hi-five insect cells (see Fig.
6, middle panel).
Consistent with previous reports, GST fusion proteins containing the
ShcA PTB domain effectively bound TrkA (Fig. 6, top
panel, lane 5). As expected, a GST fusion protein
containing full-length ShcB also bound TrkA (Fig. 6, top
panel, lane 2). Analysis of the PTB or SH2 fusion
proteins indicated that the GST fusion protein containing the PTB
domain of ShcB clearly bound TrkA (Fig. 6, top
panel, lane 3) to a similar level as compared
with full-length ShcB (Fig. 6, top panel,
lane 2). In contrast, GST-ShcB-SH2, like GST protein alone
(Fig. 6, top panel, lane 1), did not
show binding to TrkA (Fig. 6, top panel,
lane 4). Coomassie staining showed that similar amounts of
GST alone and GST fusion proteins were used (Fig. 6, lower
panel). Thus, the PTB domain, but not the SH2 domain, of ShcB binds to TrkA.
View larger version (34K):
[in this window]
[in a new window]
|
Fig. 6.
In vitro binding of ShcB domains
to TrkA. Bacterially expressed GST or GST-ShcB (full-length, PTB
domain, and SH2 domain) fusion proteins (5 µg) were assayed for
precipitation (Ppt) of insect cell lysates (300 µg)
containing baculovirus expressed HA-tagged wild type TrkA. Bound
proteins were precipitated with glutathione-Sepharose followed by
blotting with the anti-HA antibody. Arrows indicate the
positions of TrkA (gp110, gp140). To test protein loading, cell lysates
of insect cells containing baculovirus expressed HA-tagged wild type
TrkA was detected with the anti-HA 3F10 antibody (middle
panel). The same value of GST and GST fusion proteins was
applied to another gel and then stained with Coomassie Brilliant Blue
to test the amount of GST proteins loaded (bottom
panel).
|
|
Activated ShcB and ShcC Bind to Grb2--
Because activated ShcA
interacts with the SH2 domain of Grb2, via YYNS and YVNT motifs in the
CH1 region (38), we assayed ShcB-Grb2 interactions by
co-immunoprecipitation assays. For these studies, we used an nnr5 cell
line, which does not express endogenous Trk receptors, that was
generated to stably express high levels of TrkB, termed nnr5-TrkB (13,
25). Accordingly, ShcB was transfected into nnr5-TrkB cells to assay
BDNF-dependent ShcB/Grb2 interactions. Untransfected nnr5
cells were used as a negative control. BDNF-stimulated or unstimulated
cell lysates were immunoprecipitated with anti-Grb2 antibodies followed
by Western blotting with the anti-ShcB antibody. As shown in Fig.
7A, transfected ShcB in
nnr5-TrkB cells clearly co-immunoprecipitates with the anti-Grb2
antibody in BDNF-stimulated but not in unstimulated cells. A very weak band was observed in BDNF-stimulated nnr5-TrkB cells, suggesting that
ShcB is endogenously expressed in nnr5-TrkB cells. Consistent with
previous results, our data confirmed that only activated ShcB interacts
with Grb2.
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 7.
Activated ShcB and ShcC interact with
Grb2. nnr5-TrkB cells (A and B) and
nnr5-TrkC cells (C and D) were transfected with
or without ShcB (A and C) and ShcC (B
and D). After 24 h of transfection, cells were
stimulated by BDNF (A and B) and NT-3
(C and D), respectively, at a concentration of
100 ng/ml for 5 min. nnr5 cells were used in A as the
negative control. Cell lysates were prepared and immunoprecipitated
(IP) with the anti-Grb2 antibody. Blots were assayed with
the anti-ShcB (A and C) and anti-Myc
(B and D) antibodies followed by stripping and
re-probing with the anti-Grb2 antibody. Arrows indicate the
positions of ShcB or ShcC and Grb2.
|
|
ShcC also possesses conserved YYNS and YVNT motifs in the CH1
region (8). To test the ability of activated ShcC binding to Grb2,
Myc-ShcC was also transfected into nnr5-TrkB cells and the
BDNF-stimulated cell lysates were immunoprecipitated with the anti-Grb2
antibodies followed by Western blotting with the 9E10 antibody. The
results showed that only activated ShcC binds to Grb2 (Fig.
7B). As nnr5-TrkB cells have already expressed high levels
of TrkB, co-transfection of TrkB did not enhance the ShcC-Grb2 interaction (Fig. 7B, lane 3).
To test TrkC activated ShcB/Grb2 and ShcC/Grb2 interactions, we
employed nnr5-TrkC cells. ShcB or ShcC was transfected alone or
co-transfected with TrkC into the cells. After 24 h of
transfection, cells were stimulated with 100 ng/ml NT3 for 10 min. NT-3
stimulation can activate ShcB-Grb2 interaction in nnr5-TrkC cells (Fig.
7C). As TrkC expression in these cells is quite low,
co-transfection of TrkC was used to enhance phosphorylation of ShcB and
increase interaction of ShcB with Grb2 (Fig. 7C, lane
3). Similarly, ShcC-Grb2 interaction was observed by NT-3
stimulation in nnr5-TrkC cells (Fig. 7D).
Taken together, these data demonstrate that both ShcB and ShcC can be
activated by NT-3 through TrkC and that activated ShcB and ShcC both
interact with Grb2. This supports that all three Shc proteins can
activate downstream signaling (i.e. Grb2 activation) in
response to stimulation by the neurotrophin Trk receptors.
Overexpression of ShcB and ShcC Activates MAPK in PC12
Cells--
It is well known that Trk-Shc interactions are an
implicated route to activate Ras-dependent MAPK in response
to neurotrophin stimulation. As nnr5-TrkB cells express 100 times
higher levels of Trk receptors than PC12 cells (13), we could not
detect differences in the level of neurotrophin-dependent
MAPK activation between cells left untransfected and/or transfected
with ShcB (data not shown). Thus, we assayed changes in
NGF-dependent MAPK activation in PC12 cells transfected
with ShcB and ShcC. Cells were stimulated with 50 ng/ml NGF for 2 min
to activate TrkA signaling. As expected, activated MAPK was detected in
NGF-stimulated PC12 cells (Fig. 8A, lanes 2 and
5). Importantly, overexpression of ShcB could enhance the
level of neurotrophin-dependent MAPK activation (Fig. 8A, compare lane 2 with lane 3),
suggesting that ShcB stimulation results in the activation of MAPK.
Similarly, overexpression of ShcC could also enhance
NGF-dependent MAPK activation in PC12 cells (Fig.
8A, compare lane 6 with lane 5).
View larger version (33K):
[in this window]
[in a new window]
|
Fig. 8.
ShcB and ShcC overexpression enhances MAPK
activation. PC12 cells transiently overexpressing ShcB and ShcC
(A) were stimulated with NGF (50 ng/ml) for 2 min. nnr5-TrkC
cells were transfected with or without ShcB and ShcC and then
stimulated with 50 ng/ml NT-3 for 3 min (B). Cell lysates
were immunoblotted with the anti-active MAPK antibody. After stripping,
membranes were re-probed with the anti-normal MAPK antibody. Data shown
in C and D represent the mean values ± S.D.
of three to five experiments performed independently. The enhancement
of MAPK in ShcB/ShcC overexpression versus vector alone was
compared using Student's t test. *, p < 0.05.
|
|
As described above (Fig. 7), ShcB and ShcC also bind Grb2 following
NT-3 stimulation in nnr5-TrkC cells. Thus, we similarly assayed
NT-3-dependent MAPK activation following ShcB and ShcC overexpression. As expected, we found that NT-3 stimulation activates MAPK at low levels in untransfected nnr5-TrkC cells (Fig.
8B, lanes 2 and 4). However, both ShcB
and ShcC transfection enhanced MAPK activation in response to NT-3
stimulation (Fig. 8B, compare lane 3 with
lane 2 and lane 5 with lane 4). The
results further confirm that ShcB and ShcC interact well with TrkC and
enhance NT-3-dependent MAPK activation.
To more precisely determine whether ShcB and ShcC overexpression can
enhance MAPK activation, we repeated the above experiments three to
five times. As shown in Fig. 8C, NGF stimulates MAPK activation to levels ~1.5 times higher than unstimulated cells. By
comparison, overexpression of ShcB and ShcC in PC12 cells significantly increased these levels to ~2 and 2.5 times higher than cells
transfected with vector alone (p < 0.05) (Fig.
8C). Similarly, ShcB and ShcC overexpression in nnr5-TrkC
cells increased levels of MAPK activation to levels ~2 and 2.5 times
higher than vector alone (p < 0.01) (Fig.
8D).
ShcB and ShcC Are Activated in Primary Cortical Neurons--
To
assay neurotrophin-dependent activation of ShcB and ShcC in
primary neurons, we assayed tyrosine phosphorylation of ShcB and ShcC
in primary mouse cortical neurons in response to BDNF and NT-3
stimulation. Primary rodent cortical neurons (E16), which express high
levels of TrkB and TrkC receptors, were cultured for 4 days in the
absence of growth factors. Prior to lysis, cortical neurons were washed
for 1 h and then stimulated with medium alone, NGF, BDNF, or NT-3
at a concentration of 100 ng/ml for 5 min. As shown in Fig.
9 (top panel), both ShcB and
ShcC are phosphorylated in cortical neurons by BDNF and NT-3
stimulation (indicated by the solid arrows). As
expected, endogenous ShcB is larger than the positive control
containing HA-ShcB in transfected COS cells. This reflects the absence
of the N-terminal CH2 region in our ShcB cDNA construct (Fig. 9,
top and second panels, lanes
6 and 12). NGF stimulation did not activate ShcB and
ShcC, as the TrkA receptor is not expressed in cortical neurons (36,
39). Stripping and re-probing the blots with either anti-ShcB GP (8) or
anti-N-Shc (ShcC) antibodies confirmed the identity of the ShcB and
ShcC bands (Fig. 9, second and third
panels, respectively). Importantly, these results suggest
that both ShcB and ShcC are involved in the signal transduction
pathways of neurotrophin-activated Trk receptors in cortical
neurons.
View larger version (51K):
[in this window]
[in a new window]
|
Fig. 9.
ShcB and ShcC are activated in cortical
neurons. A, mouse cortical neurons isolated from day 16 embryos were stimulated with medium, NGF, BDNF, or NT-3 (100 ng/ml) for
5 min. Lysates (400 µg) were immunoprecipitated (IP) with
the anti-ShcB GP and anti-N-Shc (ShcC), and Western blots were assayed
with the anti-Tyr(P) antibody (top panel). Cell
lysates of COS cells un-transfected and transfected with HA-ShcB (CH2
minus) with TrkA (NGF-stimulated) were used as the negative and
positive control, respectively. The positions of endogenous ShcB and
ShcC and COS cell-transfected HA-ShcB are indicated by the
solid arrows. The positions of the unknown
co-immunoprecipitating proteins, pp60 and pp75, are indicated by the
stippled arrows. The blot was stripped and
re-probed first with anti-ShcB (second panel) and
then anti-N-Shc (third panel) antibodies. Note
the specificity of the antibodies for ShcB and ShcC, respectively.
Another half of co-immunoprecipitated products with the anti-ShcB and
anti-N-Shc antibodies were also assayed with the anti-Src antibodies
(fourth panel). B, mouse cortical
neurons isolated and stimulated as described in A. Lysates
(400 µg) were immunoprecipitated with the monoclonal anti-Src and the
rabbit anti-Fyn antibodies. Western blots were assayed with the
anti-Tyr(P) antibody RC20. NGF-stimulated whole cell COS lysates were
also included as a control. The arrow indicates the position
of the tyrosine-phosphorylated c-Src and Fyn proteins.
|
|
Interestingly, we also identified two additional
tyrosine-phosphorylated targets, ~60 and 75 kDa, that were
co-immunoprecipitated with the anti-ShcB GP antibody in
neurotrophin-stimulated cortical neurons (Fig. 9, stippled
arrows). The smaller protein, pp60, was also identified in
anti-ShcC immunoprecipitates. These targets were specific to cortical
tissues as they were not detected in ShcB and Trk co-transfected COS
cell lysates (data not shown) despite significant
Trk-dependent tyrosine phosphorylation of ShcB (Fig.
1A). The identity of these proteins is presently unknown. The possibility that the 60-kDa protein represents a member of the Src
family of soluble tyrosine kinases was considered, particularly because
the Fyn kinase has been shown to associate with activated TrkB in
primary rat cortical neurons (40). Interestingly, we found that ShcB
and ShcC did co-immunoprecipitate with c-Src, albeit constitutively, in
cortical cell lysates (Fig. 9, fourth panel). To
test the specificity of the Src-Shc interactions, we tested
immunoprecipitations with a second anti-ShcB antibody that we generated
and affinity-purified against a peptide specific for ShcB (449-463).
Both anti-ShcB antibodies specifically co-immunoprecipitated Src in
comparison to either normal rabbit IgGs or preimmune sera (data not
shown). These results indicate that ShcB constitutively interacts with
the Src proteins. Probably, however, because Src and Fyn are
constitutively tyrosine-phosphorylated in mouse cortical cultures, and
neither BDNF nor NT-3 stimulation increased their phosphorylation (Fig.
9B), it is likely that the tyrosine-phosphorylated protein
detected in our ShcB/ShcC immunoprecipitations represents another
protein independent of the Src family of kinases.
To demonstrate ShcB/ShcC-Trk interaction in cortical neurons,
additional co-immunoprecipitations using E16 lysates were performed. First, we determined the activation of cortical neurons following neurotrophin stimulation. Lysates were immunoprecipitated with the
anti-Trk antibody J203, and the precipitated proteins were assayed with
the anti-Tyr(P) antibody RC20. As expected, Trk was activated in
cortical neurons stimulated with BDNF and NT-3 but not in nonstimulated
and NGF-stimulated cultures (Fig. 10,
top panel). Using the same lysates (1 mg),
proteins were immunoprecipitated with the anti-ShcB antibody and
Western blots were analyzed with a monoclonal anti-Trk (MCTrk)
antibody. As expected, ShcB-Trk interaction could be detected only in
lysates from cortical neuron stimulated with BDNF and NT-3, whereas
such interaction was not observed in nonstimulated and NGF-stimulated
cortical neurons (Fig. 10, second panel). To
analyze ShcC-Trk interactions in cortical neurons, proteins were
immunoprecipitated by the rabbit anti-Trk antibody J203 (27) and
Western blots assayed with anti-ShcC antibodies. As shown in Fig. 10
(third panel), ShcC and Trk co-immunoprecipitate in lysates prepared from cortical neurons stimulated with BDNF and NT3
but not from nonstimulated and NGF-stimulated lysates. The Trks are
expressed at similar levels in each lane of unstimulated and stimulated
cortical neurons (Fig. 10, bottom panel). These results further indicate that ShcB/ShcC-Trk interactions form in mouse
cortical neurons in a neurotrophin- and
phosphotyrosine-dependent manner.
View larger version (52K):
[in this window]
[in a new window]
|
Fig. 10.
ShcB/ShcC-Trk interaction in cortical
neurons. To detect Trk activation by neurotrophin stimulations,
cortical neuron lysates (200 µg) were immunoprecipitated
(IP) with anti-Trk antibody J203 and then blotted with the
anti-Tyr(P) antibody RC20 (1:2000). The same lysates (1000 µg) were
immunoprecipitated with the anti-ShcB (second
panel) or anti-Trk (J203; bottom two
panels) antibodies. Western blots were assayed with the
anti-Trk (MCTrk) (second and fourth
panels) or anti-N-Shc (third panel)
antibodies. NGF-stimulated COS whole lysates were also included as the
control. The arrows indicate the position of Trk and
ShcC.
|
|
| |
DISCUSSION |
In this study, we addressed the activation and the
site of receptor binding of the neuronal Shc adapter proteins, ShcB and ShcC, to the neurotrophin Trk receptors, TrkA, TrkB, and TrkC. Herein,
we demonstrate that: (i) ShcB and ShcC are tyrosine-phosphorylated by
and interact with TrkA, TrkB, and TrkC in COS cells to similar levels;
(ii) ShcB and ShcC bind to TrkA at Tyr499 and TrkB at
Tyr515 in a Tyr(P)-dependent manner in both
transfected COS cells and in yeast; (iii) the PTB domain, but not the
SH2 domain, of ShcB mediates binding to the Trk receptors; (iv)
neurotrophin-activated ShcB and ShcC bind Grb2, suggesting that active
ShcB/ShcC can stimulate Ras-dependent MAPK activation; (v)
overexpression of ShcB/ShcC in PC12 and TrkC-expressing nnr5 cells
enhances neurotrophin-dependent activation of MAPK; and
(vi) endogenous ShcB and ShcC in primary mouse cortical neurons are
phosphorylated and interact with the Trks upon stimulation with BDNF
and NT-3, confirming that ShcB and ShcC are neurotrophin-activated
signaling proteins in primary neurons.
ShcA, ShcB, and ShcC Share One Site to Bind to the Trk
Receptors--
Three Shc proteins, ShcA, ShcB, and ShcC, share a
similar structure and contain CH2, PTB, CH1, and SH2 domains (8, 17). In this study, we defined that ShcB and ShcC bind to the same site on
TrkA, Tyr499, and TrkB, Tyr515, consistent with
previous reports of ShcA binding to TrkA at Tyr499. Thus,
all three Shc proteins use the same site to bind to the Trk receptors.
This conservation in structure and binding properties may be very
important for ShcA, ShcB, and ShcC to be activated by the Trk
receptors, as there are obvious differences in the temporal expression
of ShcA and ShcC at different stages of cortical neuron development in
the brain (1, 9, 18, 20). Thus, activation of one Shc adapter could be
replaced by the activation of another in different tissues or at
different stages during development. Our results provide a molecular
structural basis by which Trk binding of the Shc proteins can be
exchanged. In this respect, ShcA is expressed embryonically but not in
the adult brain, whereas ShcC expression is lower in the embryonic
brain and increases post-natally (9).
We have shown that the PTB domain of ShcB, as previously shown for ShcA
(41), mediates binding to TrkA suggesting that this domain also binds
TrkB and TrkC. The PTB domains of the Shc proteins are very conserved
in different species and different Shc proteins (8, 17). The amino acid
sequence of the mouse ShcB PTB domain is 86% identical to that of
human ShcB and 94% identical to rat ShcB (1, 8, 17). The PTB domains
of ShcA, ShcB, and ShcC are also very highly conserved (over 70% amino
acid identical to each other) (8, 17). This would provide a good
explanation for how the Shc proteins bind the same site on the Trk
receptors. Although a previous report suggested that ShcB is poorly
activated by the Trk receptors in transfected 3T3 cells (1), we find that stimulated TrkA, TrkB, and TrkC receptors can bind, phosphorylate, and activate ShcB either in transfected COS, PC12, and nnr5 cells or in
primary neurons. The discrepancy between our data and the previous
report using a longer human ShcB cDNA (1) is not entirely clear,
but may represent differences in cell types, levels of expression,
and/or the length of the N-terminal sequence of our respective clones.
The mouse ShcB cDNA used in this study (lacking the N-terminal CH2
region) contains amino acid sequence between His111 and
Pro573, identical to the sequence reported in the
full-length mouse ShcB sequence (8, 17). In comparison, the human ShcB
clone used by Nakamura et al. (1) contains an additional 39 N-terminal residues, which are significantly different from the
comparable sequence in the 5' end of mouse ShcB (8, 17). Although it is
not clear if these differences reflect significant N-terminal divergence between mouse and man, it is possible that the difference in
N-terminal length and/or sequence with our mouse ShcB clone may account
for the differences in our respective abilities to detect Trk interaction.
ShcB and ShcC Signaling Is Involved in MAPK Activation--
ShcB
activation by the three Trk receptors is a very important observation,
as all three Shc proteins are thought to serve different roles in
neural development (9, 16, 19). In this respect, ShcA is present in
neural progenitor cells and is thought to facilitate their
proliferation, whereas ShcC is primarily expressed in post-mitotic
neurons and is thought to be important to their survival/differentiation (17). For ShcB, its role has been best described so far in mice carrying a null mutation in which a specific loss of peptidergic and non-peptidergic sensory neurons, as well as
TrkA-positive fibers, in the dorsal root ganglion have been observed
(8). Moreover, although ShcC
/ mice show no gross
developmental abnormalities, mice lacking both ShcB and ShcC show a
significant loss of neurons in the superior cervical ganglia (SCG) that
is not observed in mice lacking either ShcB or ShcC alone (8). This
suggests that ShcB and ShcC serve redundant roles in SCG development
and/or survival. Interestingly, endogenous ShcA expression in the SCG
of ShcB/ShcC/ mice cannot support the
survival and/or differentiation of these neurons (8), consistent with a
model that the Shc adapters serve different roles in developing
neurons, probably mediated through unique protein-protein interactions.
Although it is becoming increasingly evident that the Shc proteins
likely serve some specificity in intracellular interactions, it is also
clear that ShcA and ShcC, as well as ShcB (this study), can bind to the
adapter protein Grb2 in response to tyrosine kinase activation.
Phosphorylation of two Tyr-X-Asn motifs in the CH1 domain of
the Shc adapters serves as the binding sites to recruit the SH2 domain
of Grb2 to the receptor complexes (38). Grb2, as a multifunctional
signaling adapter, can associate through its SH3 domains with multiple
intracellular targets including the Ras guanine nucleotide exchange
factor Sos1 and Gab-1 (10, 11, 42-44). Active Shc·Grb2·Sos
complexes can, in turn, stimulate MAPK and PI 3-kinase via Ras, whereas
Grb2·Gab-1 complexes facilitate the activation of PI 3-kinase. Both
MAPK and PI 3-kinase activation are associated with
neurotrophin-dependent cell survival and neurite outgrowth
(42-44). Our observation of enhanced
neurotrophin-dependent activation of MAPK in ShcB- and
ShcC-overexpressing PC12 and nnr5-TrkC cells is consistent with a role
of ShcB and ShcC in supporting cell survival and neuritogenesis in
response to the neurotrophins.
ShcB and ShcC Activation in Primary Cortical Neurons--
Null
mutations in the ShcB and ShcC loci have shown an important role for
the ShcB/C adapters in the development and/or survival of the sensory
neurons in the dorsal root ganglia and the sympathetic neurons in the
SCG (8). However, because both ShcB and ShcC are highly expressed in
the brain (1, 18), it is likely that both serve additional roles in
neurotrophin signaling in the central nervous system as well. To this
end, we assayed primary mouse cortical neurons, which express TrkB and
TrkC receptors (45, 46), for neurotrophin-dependent
tyrosine phosphorylation of ShcB and ShcC. Importantly, our data
clearly shows that BDNF and NT-3 stimulate the tyrosine phosphorylation
of ShcB and ShcC in primary mouse cortical neurons. Moreover, we have
shown co-immunoprecipitation between endogenous ShcB/ShcC and the Trk
receptors in neurotrophin-stimulated post-mitotic mouse cortical
neurons, suggesting a role for the ShcB and ShcC adapters in neuronal
signaling and development in the brain. Although ShcC is exclusively
expressed in the brain, in particular, in mature neurons (1, 9), ShcB
is expressed much earlier during the period of cortical neurogenesis,
which occurs between E11 and E17 in the mouse. Specifically, Sakai
et al. (8) have detected ShcB expression as early as E12 in
the brain, which remains stable until at least 12 weeks post-natally. Although no gross developmental losses were reported in the brains of
ShcB/ mice (8), it is likely that ShcB activation
serves a role in cortical neuron development and/or function that
remains to be understood.
Preliminary efforts to assay ShcB activation in cortical neurons have
indicated that two proteins, pp60 and pp75, are tyrosine-phosphorylated and co-immunoprecipitate with BDNF- and NT-3-activated ShcB. The smaller protein is also detected in ShcC co-immunoprecipitates. Although we cannot detect the same bands in ShcB/C and Trk
co-transfected COS cells, we cannot yet conclude that these two targets
are specific to neural tissues because the ShcB cDNA used in our
transfection studies lacked the N-terminal CH2 domain. Thus, these two
targets may be specific to neurons or may be CH2 domain-binding
proteins. The identification of the ShcB/ShcC-interacting proteins,
pp60 and pp75, will be necessary to examine the role of these proteins in neurotrophin-dependent signaling in primary neurons.
and
TOP
ABSTRACT
INTRODUCTION
To whom correspondence should be addressed: Laboratory of
Neural Signaling, John P. Robarts Research Inst., 100 Perth Dr., London, Ontario N6A 5K8, Canada. Tel.: 519-663-5777 (ext. 34304); Fax: 519-663-3789; E-mail: smeakin@robarts.ca.
