| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Biophys J, August 2002, p. 968-976, Vol. 83, No. 2
and
From the Departments of *Pathology and Laboratory Medicine,
Molecular and Medical Pharmacology, and
Neurobiology, UCLA School of Medicine, University of
California, Los Angeles, California 90095-1732 USA
 |
ABSTRACT |
|---|
 TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Nerve growth factor (NGF) promotes neuronal survival and differentiation by activating TrkA receptors. Similar to other receptor tyrosine kinases, ligand-induced dimerization is thought to be required for TrkA receptor activation. To study this process, we expressed TrkA receptors in Xenopus laevis oocytes and analyzed their response to NGF by using a combination of functional, biochemical, and structural approaches. TrkA receptor protein was detected in the membrane fraction of oocytes injected with TrkA receptor cRNA, but not in uninjected or mock-injected oocytes. Application of NGF to TrkA receptor-expressing oocytes promoted tyrosine phosphorylation and activated an oscillating transmembrane inward current, indicating that the TrkA receptors were functional. Freeze-fracture electron microscopic analysis demonstrated novel transmembrane particles in the P-face (protoplasmic face) of oocytes injected with TrkA cRNA, but not in uninjected or mock injected oocytes. Incubating TrkA cRNA-injected oocytes with the transcriptional inhibitor actinomycin D did not prevent the appearance of these P-face particles or electrophysiological responses to NGF, demonstrating that they did not arise from de novo transcription of an endogenous Xenopus oocyte gene. The appearance of these particles in the plasma membrane correlated with responsiveness to NGF as detected by electrophysiological analysis and receptor phosphorylation, indicating that these novel P-face particles were TrkA receptors. The dimensions of these particles (8.6 × 10 nm) were too large to be accounted for by TrkA monomers, suggesting the formation of TrkA receptor oligomers. Application of NGF did not lead to a discernible change in the size or shape of these TrkA receptor particles during an active response. These results indicate that in Xenopus oocytes, NGF activates signaling via pre-formed TrkA receptor oligomers.
| |
INTRODUCTION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Nerve growth factor (NGF) is required for the
development and maintenance of the nervous system. NGF initiates many
of its biological effects by activating TrkA tyrosine kinase receptors in the plasma membrane (Barbacid, 1995
; Klesse and Parada, 1999; Kaplan
and Miller, 2000). Similar to other receptor tyrosine kinases, NGF is
thought to induce receptor dimerization as a first step for signal
transduction (Meakin and Shooter, 1991; Schlessinger and Ullrich,
1992). This hypothesis is supported by the bivalent nature of NGF that
can bring TrkA receptor monomers into close proximity such that each
monomer phosphorylates multiple tyrosine residues on its partner,
thereby initiating downstream signal transduction (Kaplan and Stephens,
1994; Mcdonald et al., 1991; Meakin and Shooter, 1991; Schlessinger and
Ullrich, 1992). However, several studies have shown that biological
response can be elicited with monovalent ligands or a chimeric
NGF/neurotrophin-4 dimer (Maliartchouk et al., 2000a,b; Treanor
et al., 1995). Therefore, ligand binding per se may not be responsible
for receptor dimerization. In line with these observations, recent
studies of other receptor tyrosine kinases suggest that 1) tyrosine
kinase receptors can form ligand-independent oligomers on the plasma
membrane of cells (Gadella and Jovin, 1995; Wiseman and Petersen,
1999); 2) dimerization/oligomerization of these receptor tyrosine
kinases is necessary but not sufficient to activate signaling; and 3)
binding of ligand induces rotational coupling of these pre-formed
receptor oligomers to initiate receptor transphosphorylation and signal
transduction (Jiang and Hunter, 1999; Burke et al., 1997; Burke and
Stern, 1998; Livnah et al., 1996, 1998, 1999; Remy et al., 1999).
To investigate the oligomeric state (i.e., monomer, dimer, or larger
size oligomer) of TrkA receptors and the possible role of NGF binding
in inducing oligomerization, we expressed TrkA receptors in
Xenopus laevis oocytes and examined their
response to NGF by using a combination of functional, biochemical, and structural approaches. Two properties made oocytes ideally suited for
structure/function studies. First, the plasma membrane of the oocyte
contains a low density of endogenous transmembrane proteins
(200-300/µm2 instead of
>2000/µm2 in most cells), which allows the
identification of the exogenous proteins from the size and shape of the
newly inserted particles induced by cRNA injection. Second, oocytes
synthesize and transport to the plasma membrane a large number of
exogenous channels and transporters as well as many of their
genetically engineered mutants. While in the plasma membrane, the
exogenous proteins are characterized by electrophysiological and
optical methods before the same oocytes are processed for electron
microscopy (Eskandari et al., 1998, 1999, 2000; Zampighi et al., 1995,
1999). These properties of the oocyte expression system provide a
simple yet powerful approach to determine whether functional membrane
proteins are assembled as monomers, dimers, or larger size oligomers in
a heterologous cellular environment.
| |
MATERIALS AND METHODS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
cRNA preparation and oocyte microinjection
The carboxy-terminal FLAG-tagged TrkA receptor was generated by
a PCR strategy to insert two FLAG epitopes in tandem on the carboxy-terminal (Mischel et al., 2001). A PCR reaction was done with
the rat TrkA cDNA as a template, using forward primer 5'CA GGG ACT AGT
GGT CAA GAT GAT TGG A and reverse primer 5' CTT ATC ATC ATC ATC CTT GTA
ATC GCC CAG AAC GTC CAG G to amplify a fragment encoding the last 405 basepairs of rat TrkA receptor coding sequence (minus the stop codon)
including the 5'Spe1 site (part of the coding sequence) and
the first FLAG epitope on the 3' end. This product was then used as a
template for a second PCR reaction using the same forward primer and a
reverse primer encoding the second FLAG tag followed by a stop codon
and Xho1 restriction site 5'CTG CTC GAG CTA CTT ATC ATC ATC
ATC CTT GTA ATC CTT ATC ATC ATC. The amplified product was digested
with Spe1 and Xho1 and subcloned into similarly
digested full-length TrkA in pGEMHE (a modified pGEM vector that
contains 5' and 3' Xenopus 
-globin untranslated sequences
for RNA stability (Liman et al., 1992; Mischel et al., 2001). TrkA was
transcribed using the T7 mMessage Machine from Ambion (Austin, TX).
Transcripts were purified using an RNAeasy kit purchased from QIAGEN
Inc. (Valencia, CA), analyzed on agarose gels, and quantitated by
spectrophotometry (and by densitometric comparison to RNA standards).
Mature Xenopus laevis oocytes (Dumont stage V-VI) were isolated, defolliculated, and injected with 5 ng of TrkA cRNA. Oocytes were maintained at 17-18°C in ND96 (in mM): 96 NaCl, 2 KCl, 1.8 CaCl2, 1 MgCl2, 5 HEPES, pH 7.5, and 1 µg/ml ciprofloxacin, 6 µg/ml ceftazidime, and 1 µg/ml gentamicin.
Immunoblot analysis
For analysis of TrkA expression, 10 oocytes microinjected with TrkA cRNA (and 10 control oocytes) were manually homogenized in an ice-cold solubilization buffer (65 mM Tris-HCl, 50 mM NaCl, 5 mM EDTA, 2% Triton X-100, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 1 mM phenylmethylsulfonyl fluoride, 5 µg/ml aprotinin, 5 µg/ml leupeptin, 1 mM sodium orthovanadate, 10 mM sodium fluoride, pH 7.5). Lysates were incubated at 4°C for 1 h and the soluble phase was partitioned from the insoluble and lipid phases by three rounds of centrifugation at 18370 × g for 5 min at 4°C. For preparation of a membrane fraction, oocytes were manually homogenized in ice-cold homogenization buffer and samples were centrifuged at 735 × g for 5 min at 4°C. The pellet (containing the membrane) was solubilized in 500 µl of a buffer containing 130 mM Tris-HCl, pH 7.5, 150 mM NaCl, 2% Triton X-100 and 10 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 µg/ml aprotinin, 1 µg/ml leupeptin, and 1 mM sodium orthovanadate. Samples were pre-cleared with protein A/G Sepharose beads (Pierce, Rockford, IL), and immunoprecipitated with anti-FLAG M2 antibody (Stratagene, La Jolla, CA). Immunoprecipitates were washed three times with 1 ml of solubilization buffer and protein was recovered by boiling for 5 min in 3× Laemmli sample buffer. SDS-PAGE was performed on 7.5% polyacrylamide gels. After transfer to nitrocellulose membranes (Hybond, Amersham Pharmacia Biotech, Piscataway, NJ), samples were blocked in TBST (20 mM Tris-HCl, 150 mM NaCl, 0.1% Tween 20, pH 8.0) with 5% non-fat dried milk and incubated with the anti-FLAG M2 monoclonal antibody (Stratagene) at a concentration of 10 µg/ml overnight at 4°C. After washing, blots were incubated with horseradish peroxidase (HRP) conjugated goat anti-mouse antibody (Promega, Madison, WI) or HRP conjugated goat anti-rabbit antibody (Promega) at a dilution of 1:5000 for 1-2 h at room temperature. After washing, bands were visualized by incubating with Supersignal (Pierce), followed by exposure to Kodak BIOMAX film (Eastman Kodak Company, Rochester, NY). Experiments were repeated three times using independent oocytes.
For analysis of tyrosine phosphorylation, 100 ng·ml-1 of NGF (4 nM) (Sigma, St. Louis, MO) was added to the media of 10 TrkA-expressing oocytes and 10 control oocytes for 5 min at room temperature. Oocytes were manually homogenized in ice-cold solubilization buffer, incubated for 1 h at 4°C, and centrifuged at 18370 × g for 5 min at 4°C. Samples were resolved electrophoretically for immunoblot analysis as mentioned above, and incubated with anti-phosphotyrosine monoclonal antibody 4G10 (Upstate Biotechnology, Lake Placid, NY). Blots were washed as above and incubated with HRP conjugated goat anti-mouse antibody (Promega), and visualized by using enhanced chemiluminescence (Amersham Pharmacia Biotech). Experiments were repeated three times using independent oocytes.
For analysis of endogenous NGF expression in oocytes, 20 uninjected oocytes were incubated in 60 µl of ND96 for 24 h. Oocytes was manually homogenized in ice-cold solubilization buffer, solubilized and fractionated as described above. Samples were separated electrophoretically for immunoblot analysis and incubated with anti-2.5S-NGF monoclonal antibody (Sigma). For analysis of NGF expression in the oocyte culture medium, 20 µl (one-third of total culture medium) was resolved electrophoretically for immunoblot analysis.
Electrophysiology
Stage V-VI Xenopus laevis oocytes were
isolated, defolliculated, injected with TrkA cRNA (5-20 ng), and
cultured for 2 days in Barth's solution (Umbach et al., 1990). Oocytes
with resting potentials more negative than 
50 mV were voltage-clamped
using the two-electrode voltage clamp method and an Axoclamp-2A
amplifier (Axon Instruments, Foster City, CA). The membrane potential
was held at 60 mV and current measurements were made at 20-22°C
during bath application of NGF (100 ng· ml-1)
in Barth's solution. Data were recorded on a PrimeLine (Soltec, San
Fernando, CA) chart recorder and also stored on Axotape (Axon Instruments) for analysis. The reversal potential of the NGF-evoked current was estimated by stepping the holding potential to depolarizing potentials during an active response (Umbach et al., 1990). Membrane capacitance was determined at the beginning of each experiment by
integrating (using Clampfit, pCLAMP 6.0, Axon Instruments) the
capacitive current elicited by +10 and +20 mV depolarizing steps (30 ms) from the holding membrane potential (Zampighi et al., 1999).
Whole-cell capacitance is reported as the mean calculated from at least
six pulses per cell. Immediately following electrophysiological measurements, oocytes were fixed in 4% glutaraldehyde in 0.2 M sodium
cacodylate (pH 7.4) (Zampighi et al., 1999), and prepared for
freeze-fracture electron microscopy (Eskandari et al., 1998, 1999,
2000; Zampighi et al., 1999).
Freeze-fracture electron microscopy
The freeze-fracture structural data were collected from a total
of eight oocytes (3 TrkA-expressing oocytes treated with NGF, 3 TrkA-expressing oocytes without NGF treatment, and 2 control uninjected
oocytes). For particle density determinations, images of protoplasmic
(P) fracture faces were digitized at a final magnification of
150,000×. Particle densities were determined by counting P-face particles from known areas of the plasma membrane (NIH Image). For each
oocyte, a minimum of 1000 particles was counted from at least 10 µm2 of plasma membrane. For measurement of the
dimensions of the freeze-fracture particles, particles were sampled
from two replicas from two oocytes. For construction of the frequency
histograms, the particle diameter was measured directly from the
negatives using a comparator (Nikon, Model 6c) at a final magnification of 1,000,000×. The diameter measurements were plotted in a frequency histogram with a bin size of 1.0 nm and were fitted to a multiple Gaussian function (Eskandari et al., 1998, 1999, 2000; Zampighi et al.,
1999) (see Fig. 3). All values are reported as mean ± SE.
| |
RESULTS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Expression of functional TrkA receptors in Xenopus oocytes
As a prelude to the morphometric characterization of TrkA
receptors expressed in the plasma membrane of Xenopus
oocytes, we performed experiments to verify that functional TrkA
receptors were expressed in these cells. Prior investigations have
documented, and we have confirmed (data not shown), that
Xenopus oocytes do not express either TrkA or p75NTR
receptors (Mischel et al., 2001; Nebreda et al., 1991; Sehgal et al.,
1988). Injection of Xenopus oocytes with TrkA receptor cRNA
led to the appearance of a 140/110 kDa TrkA immunoreactive doublet,
which was not detected in uninjected (Fig.
1 A) or water injected oocytes
(data not shown) (Carriero et al., 1991; Mischel et al., 2001; Nebreda
et al., 1991). The 140 kDa band, which represents the fully
glycosylated membrane-bound receptor (Barbacid, 1995; Meakin and
Shooter, 1991) was also present in a plasma membrane fraction isolated
from TrkA receptor cRNA-injected oocytes, but not in control oocytes
(Fig. 1 A) (Mischel et al., 2001). These TrkA receptors were
catalytically active, as demonstrated by tyrosine phosphorylation of a
number of proteins, including TrkA receptors, in response to NGF (Fig.
1 B).
|
Further evidence that the 140/110 kDa immunoreactive species of Fig. 1, A and B corresponded to functional TrkA receptors came from the electrophysiological analysis of oocytes expressing TrkA. In the absence of NGF the resting electrophysiological properties of TrkA-expressing oocytes did not differ from those of uninjected or water-injected oocytes (not shown). Bath application of a saturating concentration of NGF (100 ng·ml-1) elicited a slowly developing (10-15 s) oscillating inward current in oocytes injected with TrkA receptor cRNA (Fig. 1 C). This NGF-evoked current was not seen in control oocytes, even after several minutes of exposure to this ligand (Fig. 1 C). Overall, 10 oocytes injected with TrkA receptor cRNA exhibited responses similar to that seen in Fig. 1 C, while uninjected oocytes (n = 5) were uniformly unresponsive to NGF.
Freeze-fracture analysis of the plasma membrane of oocytes
To investigate whether NGF induces oligomerization of TrkA
receptors, we used freeze-fracture electron microscopy to examine TrkA
receptors expressed in the plasma membrane of oocytes. First, we
characterized the plasma membrane particles of control oocytes. The
P-face of control cells contained a low density of particles at
200 ± 10 per µm2 (Fig.
2 A). Consistent with previous
observations (Eskandari et al., 1998, 1999, 2000; Zampighi et al.,
1999), these endogenous P-face particles were circular in cross-section
with a mean diameter of 7.6 ± 0.5 nm (Fig. 2 A).
Independent assessment of exoplasmic (E) face particles revealed a
density of 890 ± 40 per µm2 (Eskandari et
al., 1999). Water-injected oocytes exhibit a similar population of
P-face and E-face particles (data not shown). We then examined the
freeze-fracture particles of oocytes expressing TrkA in the absence of
exposure to NGF. Oocytes expressing the TrkA receptor revealed an
additional population of large particles in the P face of the plasma
membrane (Fig. 2 B and Fig. 3,
A and C). High magnification analysis
(1,000,000×) of a selected subset of these particles
(n = 48) demonstrated a rectangular shape with axes
measuring 10 ± 1.3 nm × 8.6 ± 0.5 nm (Figs. 2
B and 3); a/b = 1.16 ± 0.05. These
particles were randomly oriented in the xy-plane. The
presence of this new population of particles in the plasma membrane led
to an increase in the total density of P-face particles to 340 particles/µm2, an increase of ~140
particles/µm2. No change in the density of
E-face particles was observed. This partitioning of the TrkA receptors
to the P-face of the membrane is consistent with other heterologously
expressed integral membrane proteins (Eskandari et al., 1998, 1999,
2000; Zampighi et al., 1995, 1999). To exclude the possibility that
these particles resulted from the expression of an endogenous membrane
protein initiated by the injection of TrkA receptor cRNA, we cultured
the oocytes in actinomycin D (0.1 mg·ml-1), which inhibits transcription
of endogenous genes (Umbach and Gunderson, 1987). This treatment did
not block the appearance of the novel TrkA-induced particles, or
electrophysiological responses of oocytes to NGF (data not shown).
Therefore, the 10 × 8.6 nm freeze-fracture particles that
appeared after injection of TrkA cRNA arose specifically from
translation of this foreign cRNA. Using whole-cell capacitance measured
in the same cell as an index of the cell surface area (Zampighi et al.,
1999), we estimated that there were 
3.4 × 109 TrkA receptor particles present in the plasma
membrane of the oocytes analyzed.
|
|
Xenopus oocytes do not secrete NGF
NGF mRNA is present in Xenopus oocytes, and an immature
NGF precursor protein has been observed in the cytoplasm of oocytes (Carriero et al., 1991). However, mature NGF protein has not been detected in oocytes or their secreted medium (Carriero et al., 1991;
Wion et al., 1984). Similarly, neurotrophic activity could not be
detected in the conditioned medium of oocytes or ovarian follicles
(Sehgal et al., 1988). We detected a 32 kDa immunoreactive band in the
cytoplasmic fraction of oocytes (Fig. 1 D) consistent with
the NGF precursor, as has previously been demonstrated (Carriero et
al., 1991; Wion et al., 1984). However, the mature form of NGF was not
detected in the cytoplasm or culture medium of oocytes (Fig. 1
D). Similarly, no mature NGF was detected in the culture medium from water-injected or TrkA cRNA-injected oocytes (data not
shown). These findings argue that oocytes do not secrete NGF, consistent with our electrophysiological result in which the large NGF-evoked inward currents (Fig. 1 C) were observed only
after addition of NGF to the oocyte bathing medium. For NGF to have provoked the responses exemplified in Fig. 1 C, there must
have been unliganded TrkA receptors in the plasma membrane. Therefore, these results suggest that the 10 × 8.6 nm TrkA receptor
particles in the plasma membrane (Figs. 2 and 3 A)
correspond to the inactive form of the receptor before NGF application.
NGF does not induce TrkA receptor aggregation
To determine whether NGF induces TrkA receptor aggregation, we studied the size and shape of freeze-fracture particles in oocytes fixed during an active response to the ligand, as determined by the inward transmembrane current (as illustrated by the response in Fig. 1 C). Simple visual inspection showed that dimensions of the P and E freeze-fracture particles remained unchanged upon addition of NGF. This qualitative assessment was confirmed by comparing size frequency histograms of P-face particles in the plasma membrane of NGF-treated and untreated oocytes (Fig. 3, C and D). For the ligand-induced dimerization hypothesis to be correct, the 10 × 8.6 nm TrkA receptor particles would have to form substantially larger (20 × 8.6 nm or 10 × 17.2 nm) freeze-fracture particles. It can be argued that the electrical response elicited from saturating concentration of NGF might have resulted in the dimerization of just a few receptor molecules. However, the predicted 20 × 8.6 nm or 10 × 17.2 nm diameter particles are so large with respect to the size distribution observed in untreated oocytes that it would have been easily detected by simple visual inspection of the large number of replicas examined. Therefore, NGF does not alter the dimensions of the TrkA receptor particles in the plasma membrane of the oocyte during an active response.
| |
DISCUSSION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
It has been proposed that ligand-induced dimerization of
TrkA receptor monomers is a prerequisite for NGF signaling (Jing et
al., 1992; Meakin and Shooter, 1991; Schlessinger and Ullrich, 1992).
NGF, which is bivalent (Mcdonald et al., 1991), is thought to bridge
two TrkA receptor monomers, allowing each monomer to phosphorylate its
partner. The finding that NGF promotes chemical cross-linking of TrkA
receptors lends support to this hypothesis (Jing et al., 1992; Meakin
and Shooter, 1991; Schlessinger and Ullrich, 1992). However, evidence
obtained from the use of monovalent ligands that cannot bridge TrkA
receptors suggests that ligand binding per se may not be responsible
for receptor dimerization (Maliartchouk et al., 2000a; Treanor
et al., 1995). To test this hypothesis, we combined
electrophysiological, biochemical, and freeze-fracture electron
microscopy data to study the effect of NGF treatment on the structure
and function of TrkA receptors expressed in the plasma membrane of oocytes.
Our data indicate that functional TrkA receptors were expressed in the
plasma membrane of oocytes injected with TrkA cRNA. Injection of TrkA
cRNA correlated with the appearance of a 140/110 kDa doublet on
immunoblot analysis. These TrkA receptors were catalytically active, as
the TrkA receptor protein exhibited an NGF-dependent increase in
tyrosine phosphorylation (Clary et al., 1994; Jing et al., 1992; Kaplan
and Stephens, 1994; Kaplan and Miller, 2000; Segal and Greenberg,
1996). Electrophysiological analysis indicated that addition of NGF to
the bathing medium evoked an oscillating inward current with a reversal
potential of 29 mV, consistent with a chloride current. TrkA-mediated
activation of phospholipase C stimulates the release of ionized calcium
from cytoplasmic stores (Kaplan and Stephens, 1994; Klein et al., 1991; Loeb et al., 1994; Mischel et al., 2001; Obermeier et al., 1993; Segal
and Greenberg, 1996), which activates endogenous calcium-dependent chloride channels in oocytes (Callamaras and Parker, 2000; Hartzell, 1996; Weber, 1999). The NGF-evoked current was not seen in control oocytes, indicating that the response to NGF was via the heterologously expressed TrkA receptors.
To visualize the TrkA receptors on the plasma membrane we performed
freeze-fracture electron microscopy. This method splits plasma
membranes to produce complementary P and E fracture faces, where
integral membrane proteins appear as "particles" in the fracture
faces (Branton, 1966; Eskandari et al., 1998, 1999, 2000; Zampighi et
al., 1999). Proteins heterologously expressed in oocytes appear as
distinct P-face particles whose size and shape can be distinguished
from the endogenous integral membrane proteins (Eskandari et al., 1998,
1999, 2000; Zampighi et al., 1999). This technique can resolve proteins
as small as 2.5 nm (Eskandari et al., 1999). In oocytes expressing the
TrkA receptor, a new population of rectangular P-face particles
(10 × 8.6 nm) was observed in the plasma membrane. Several lines
of evidence indicate that these freeze-fracture particles represent the
functional TrkA receptors in the plasma membrane observed by
electrophysiological and biochemical methods. First, they were present
in TrkA-expressing oocytes, but not in mock-injected or uninjected
oocytes. Second, they partitioned only within the P-face of the
membrane, as has been demonstrated for a number of other heterologously
expressed transmembrane proteins (Eskandari et al., 1998, 1999, 2000;
Zampighi et al., 1995, 1999). Third, their presence in the plasma
membrane correlated with responsiveness to NGF, as detected both by
electrophysiological analysis and receptor phosphorylation. Finally,
incubating oocytes with the transcriptional inhibitor actinomycin D did
not prevent the appearance of the P-face particles or
electrophysiological responses to NGF, indicating that they did not
arise from de novo transcription of an endogenous Xenopus
oocyte gene.
Previously, we have studied integral membrane proteins that consist of
a multiple transmembrane domains connected by small loops of
extracellular domain (Eskandari et al., 1998, 1999, 2000). Therefore,
the cross-sectional area of these freeze-fracture particles provided an
index of the number of transmembrane 
helices of the proteins
(1.4 ± 0.03 helix/nm3) (Eskandari et al.,
1998, 1999). In contrast, TrkA receptors have a single transmembrane
domain and a large globular extracellular domain (Holden et al., 1997;
Wiesmann et al., 1999). Because the freeze-fracture technique results
in asymmetric cleavage of the integral membrane protein, it is likely
that the juxtamembrane region of the extracellular domain contributes
to the area of the freeze-fracture particle. The TrkA receptor
particles in the P-face of the plasma membrane measured 8.6 × 10 nm. After correcting for the thickness of the metal replica (2.4 nm on
each side) (Eskandari et al., 1998, 1999, 2000), the TrkA receptor
particles were 6.2 ± 0.5 nm × 7.5 ± 1.3 nm. Because
the crystal structure derived from the NGF binding domain of two TrkA
monomers complexed to NGF is 6.0 × 9.5 nm (Wiesmann et al.,
1999), our data suggest that each TrkA receptor particle consists of a
minimum of two TrkA receptors. Alternatively, if the extracellular
domain of TrkA does not contribute to the freeze-fracture particles,
then each TrkA receptor particle consists of larger aggregates of TrkA receptors. In either case, the data indicate that TrkA receptors are
present on the plasma membrane as pre-formed oligomers before NGF exposure.
To ensure that TrkA receptors were not oligomerized in response to
endogenously secreted NGF, we examined oocytes and their culture media
for evidence of NGF secretion. Uninjected oocytes and oocytes injected
with mRNA isolated from submaxillary glands produce an immature 32 kDa
NGF precursor protein, which is present in the cytoplasm (Carriero et
al., 1991; Wion et al., 1984). However, they do not secrete detectable
mature NGF (Carriero et al., 1991; Wion et al., 1984), prompting the
hypothesis that oocytes lack the machinery to process it (Wion et al.,
1984). We detected the 32 kDa NGF precursor protein in control and TrkA
cRNA-injected oocytes, but we did not detect any NGF secretion (or more
importantly, production of mature NGF). Although we cannot formally
dismiss the possibility that the NGF precursor protein induces
oligomerization of TrkA receptors before its insertion in the plasma
membrane, our data indicate that TrkA oligomerization is not sufficient to activate signaling. Moreover, our data show that the receptors expressed on the plasma membrane were competent to respond to NGF,
suggesting that they were not occupied by NGF precursor protein.
To further investigate our hypothesis that TrkA receptors were
pre-oligomerized, we proceeded to study the effect of NGF application on TrkA receptor aggregation. NGF had no discernible effect on the
dimensions of the TrkA receptor particles compared to the TrkA-expressing controls, which had not been exposed to NGF. If NGF
promoted aggregation of TrkA receptors to activate signaling (Jing et
al., 1992; Meakin and Shooter, 1991; Schlessinger and Ullrich, 1992),
we should have detected a new population of particles of approximately
twice the dimensions of the unliganded receptors. The expected increase
in the particle dimensions is well within the limits of resolution of
the freeze-fracture replicas (2.5 nm) (Eskandari et al., 1999) and
should have been evident by simple visual inspection of the replicas.
Our method of Gaussian fit to study the size distribution of particles
would not be able to distinguish a difference in TrkA-expressing
oocytes treated with or without NGF if <5% of the TrkA receptors were
responding to NGF. However, the extent of TrkA receptor phosphorylation
in response to bath application of NGF (Fig. 1 B) strongly
suggests that a sufficient number of TrkA receptors were activated that would have enabled us to detect any large particles due to NGF-mediated oligomerization. Furthermore, the predicted freeze-fracture particles (20 × 17.2 nm) would be so large with respect to the size
distribution observed in the untreated oocytes that they would have
been easily identified by visual inspection, even at particle densities
below 1 per µm2. Therefore, our results
indicate that NGF did not induce changes in the state of
oligomerization of TrkA receptors in the plasma membrane.
TrkA receptors were expressed at a relatively high density in oocytes
compared to PC12 cells, which are a common model of TrkA signaling
(Connolly et al., 1985; Klesse and Parada, 1999). TrkA density in the
oocyte plasma membrane was ~140 per µm2. If
each particle represents a TrkA dimer, then there are ~280 TrkA
receptor monomers per µm2 in TrkA-expressing
oocytes. In contrast, PC12 cells have a lower density of TrkA
receptors. In PC12 cells, 2100-3000 TrkA binding sites have been
estimated to be in the plasma membrane per cell (Clary et al., 1994;
Hempstead et al., 1991). Because the surface area of undifferentiated
PC12 cells is 290 µm2 (Connolly et al.,
1985), this suggests that there are 7-10 TrkA binding sites per
µm2. Therefore, we estimate that there is up to
a 40-fold difference in the density of TrkA receptors in the plasma
membrane of oocytes relative to PC12 cells. It can be argued that
overexpression of TrkA receptors may contribute to their
oligomerization. Although we detected a low level of phosphorylation of
these receptors in the absence of NGF, TrkA tyrosine phosphorylation
was dramatically increased by NGF treatment. In addition,
TrkA-expressing oocytes did not reveal a large inward current in the
absence of NGF, and only exhibited this current in the presence of NGF.
Therefore, our data suggest that NGF activation of constitutive TrkA
receptor oligomers is physiological, and that TrkA receptor
oligomerization is insufficient for TrkA signaling. This conclusion is
further supported by the observation that in oocytes, oligomerization is an intrinsic property of the proteins that often occurs in the rough
endoplasmic reticulum and is not a function of density of receptors in
the plasma membrane (Eskandari et al., 1998, 1999, 2000; Zampighi et
al., 1999). For example, the Na+/glucose
cotransporter is monomeric over a very wide density range (100-3500/µm2) (Eskandari et al., 1998); the
neuronal glutamate transporter (EAAT3) is pentameric
(100-800/µm2) (Eskandari et al., 2000); and
the gap junction hemichannel connexin50 is hexameric
(100-800/µm2) (Zampighi et al., 1999). These
results suggest that oligomer formation in the absence of ligand is an
intrinsic property of TrkA receptors, and is not sufficient to activate
signaling. Therefore, TrkA oligomers are fully competent to respond to
NGF, and initiate signaling after NGF binding.
The experimental evidence underlying the current model of
ligand-induced TrkA receptor dimerization is derived primarily from chemical cross-linking studies (Jing et al., 1992; Meakin and Shooter,
1991; Schlessinger and Ullrich, 1992). Similar cross-linking studies of
the platelet derived growth factor receptor (PDGF-) and the
epidermal growth factor receptor (EGFR) also suggested that these
receptors exist in the plasma membrane as monomers that dimerize upon
addition of ligand (Cochet et al., 1988; Heldin et al., 1989; Inui et
al., 1993). However, the use of new biophysical methods such as
fluorescence resonance energy transfer (FRET) and image correlation
spectroscopy (ICS) have challenged this model. In contrast to the
chemical cross-linking data, FRET analysis demonstrated that EGF
receptors exist in a pre-oligomerized state in the plasma membrane of
A431 cells (Gadella and Jovin, 1995), and ICS analysis of
non-transformed human fibroblasts demonstrates that PDGF- receptors
are pre-aggregated as tetramers in the absence of ligand (Wiseman and
Petersen., 1999). Our finding that TrkA receptors are pre-oligomerized
is consistent with the these observations. In the future, it will be
important to apply these other biophysical techniques to study TrkA
aggregation in mammalian cells.
The hypothesis that TrkA is a constitutive oligomer in the oocyte can
explain several observations that are difficult to reconcile with the
notion that TrkA receptor dimerization is sufficient for signaling. A
dimer composed of NGF conjugated to NT-4 activates transphosphorylation
of TrkA receptors and initiates biological responses, even though NT-4
(at physiological concentrations) does not bind to TrkA receptors
(Treanor et al., 1995). This strongly suggests the NGF/NT-4 heterodimer
activates pre-formed TrkA receptor dimers. Similarly, small peptide
monovalent ligands, can activate TrkA receptors and promote neurite
outgrowth in PC12 cells (Maliartchouk et al., 2000a) and primary
embryonic dorsal root ganglion (DRG) neurons (Maliartchouk et al.,
2000a), prompting the hypothesis that small molecule
peptidomimetics stabilize TrkA receptor conformations that promote
receptor activation.
Recent studies suggest that TrkA receptors must be in close proximity
to each other in the plasma membrane before ligand binding. TrkA
receptors bearing deletions and point mutations within the first or
second immunoglobulin-like binding domains are constitutively phosphorylated and biologically active in the absence of NGF (Arevalo et al., 2000). The first immunoglobulin-like domain of TrkA appears to
play a key role in regulating this ligand-independent activation (Zaccaro et al., 2001). To explain these findings, Arevalo et al. have
suggested that there is a balance of attractive and repulsive forces
between TrkA receptor monomers in the membrane, which is altered by NGF
binding (or mutations within the ligand binding domain of TrkA)
(Arevalo et al., 2000). Our findings are also not at odds with the
observation of Woo and colleagues, who demonstrated that NGF promotes
dimerization of the soluble extracellular domain of TrkA (Woo et al.,
1998). Rather, they argue for the importance of the transmembrane
domain in promoting ligand-independent TrkA oligomerization. Consistent
with this, most growth factor receptor tyrosine kinases, including TrkA
receptors, contain a conserved transmembrane dimerization motif
(Sternberg and Gullick, 1990) that may enable them to form inactive
dimers in the absence of ligand (Jiang and Hunter, 1999). Our study
suggests that TrkA receptors may be activated in a similar fashion to
other tyrosine kinases receptors, such as the ErbB2/neu and EPO
receptor, for which dimerization is necessary, but not sufficient, to
activate signaling (Burke et al., 1997, 1998; Jiang and Hunter, 1999;
Livnah et al., 1996, 1998, 1999; Remy et al., 1999). Precise rotational coupling of receptor monomers induced by ligand-binding may be required
to initiate transphosphorylation of critical tyrosine residues that are
required for signal transduction (Jiang and Hunter, 1999).
In summary, the unique characteristics of the Xenopus oocyte expression system allowed us to use combined functional, biochemical, and structural methods to study the role of NGF binding in TrkA receptor dimerization. We conclude that, in the absence of NGF, TrkA receptors form oligomers in the plasma membrane of oocytes. Therefore, activation of signal transduction via the TrkA receptors likely involves conformational changes that are induced by NGF binding to pre-formed oligomeric receptor complexes. In the future, it will be important to resolve the nature of the conformational changes that accompany NGF binding, and determine whether constitutive TrkA receptor complexes are present in the plasma membrane of mammalian cells as well.
| |
ACKNOWLEDGMENTS |
|---|
We thank Michael Kreman for his skillful preparation of the freeze-fracture replicas and Dr. Jeffrey Twiss and Dr. Uri Saragovi for critical review of this manuscript.
This work was supported by Grant U01 CA88127 from the National Cancer Institute/National Institutes of Health and Grant K08NS43147-01 from the National Institute of Neurological Disorders and Stroke/National Institutes of Health (to P.S.M.); a Henry E. Singleton Brain Tumor Fellowship and a Stop Cancer Award (to P.S.M.); and a generous donation to the UCLA Comprehensive Brain Tumor Program from the Kevin Riley family. This work was also supported in part by the Harry Allgauer Foundation through The Doris R. Ullmann Fund for Brain Tumor Research Technologies. J.A.U. (NS31934), C.B.G. (MH-59938) and G.A.Z. (EY-04110) are supported by grants from the National Institutes of Health. S.G.S was the recipient of a pre-doctoral fellowship from the Jonsson Cancer Center Foundation.
| |
FOOTNOTES |
|---|
Address reprint requests to Paul S. Mischel, Department of Pathology and Laboratory Medicine, UCLA School of Medicine, 10833 Le Conte Avenue, Los Angeles, CA 90095-1732. Tel.: 310-794-5223; Fax: 310-206-0657; E-mail: pmischel{at}mednet.ucla.edu.
Submitted October 8, 2001, and accepted for publication April 3, 2002.
Sepehr Eskandari's present address is Biological Sciences Department, California State Polytechnic University, Pomona, CA 91768-4032.
| |
REFERENCES |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
) currents in Xenopus oocytes is modulated by voltage.
Am. J. Physiol. Cell Physiol.
278:C667-C675
Biophys J, August 2002, p. 968-976, Vol. 83, No. 2
© 2002 by the Biophysical Society 0006-3495/02/08/968/09 $2.00
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
