Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Mol Cell Neurosci. Author manuscript; available in PMC 2009 September 1.
Published in final edited form as:
PMCID: PMC2606928

A Bi-directional Carboxypeptidase E - driven transport mechanism controls BDNF vesicle homeostasis in hippocampal neurons


Anterograde transport of brain-derived neurotrophic factor (BDNF) vesicles from the soma to neurite terminals is necessary for activity-dependent secretion of BDNF to mediate synaptic plasticity, memory and learning, and retrograde BDNF transport back to the soma for recycling. In our study, overexpression of the cytoplasmic tail of the carboxypeptidase E (CPE) found in BDNF vesicles significantly reduced localization of BDNF in neurites of hippocampal neurons. Live-cell imaging showed that the velocity and distance of movement of fluorescent protein-tagged CPE- or BDNF-containing vesicles were reduced in both directions. In pull-down assays, the CPE tail interacted with dynactin along with kinesin-2 and kinesin-3, and cytoplasmic dynein. Competition assays using a CPE tail peptide verified specific interaction between the CPE tail and dynactin. Thus, the CPE cytoplasmic tail binds dynactin that recruits kinesins or dynein for driving bi-directional transport of BDNF vesicle to maintain vesicle homeostasis and secretion in hippocampal neurons.

Keywords: Carboxypeptidase E, cytoplasmic tail, BDNF vesicle transport, dynactin


Brain-derived neurotrophic factor (BDNF), a neutrophin, is primarily synthesized in the soma, packaged into regulated secretory pathway vesicles, transported anterogradely to neurite terminals, and secreted in an activity-dependent manner (Altar and DiStefano, 1998; Goodman et al., 1996; Haubensak et al., 1998; Mowla et al., 1999). The anterograde transport of BDNF vesicles along processes to axonal/dendritic terminals is found to be essential for release to regulate synaptic plasticity, memory and learning (Egan et al., 2003; Lu, 2003; Poo, 2001; Thoenen, 2000). Moreover, a recent study (Shakiryanova et al., 2006) indicates that peptidergic vesicles, if not trapped in the nerve terminal for release in an activity-dependent manner, are retrogradely transported back to the cell body, presumably for degradation, or reuse, of excess vesicles transported to the synapse. However, the mechanism of anterograde and retrograde BDNF vesicle transport is poorly understood.

Studies have indicated that anterograde BDNF vesicle transport is microtubule-based (Gauthier et al., 2004). Various microtubule motors are involved in peptidergic vesicle transport. These include the microtubule plus end-directed motor, kinesin, which is the major conveyer for anterograde transport toward axonal termini, and cytoplasmic dynein, a minus end-directed motor, responsible for retrograde transport to the cell body (Goldstein and Yang, 2000). Retrograde transport of nerve growth factor has been shown to be mediated by binding to transmembrane neurotrophin receptors which associate with components of the dynein motors (Yano et al., 2001). Cytoplasmic dynein requires interaction with the multiprotein complex, dynactin, for its proper function (Gill et al., 1991). However, recent studies indicate that dynactin is involved not only in retrograde, but also anterograde transport on microtubules (Deacon et al., 2003; Dell, 2003; Ross et al., 2006). Furthermore, dynactin is also found to facilitate anterograde transport of BDNF in neuronal cells via its interaction with cytosolic huntingtin-associated protein-1 (HAP1) and Huntingtin (Htt) (Gauthier et al., 2004), but how the vesicles interact with Htt, HAP1 or dynactin is unknown. It is likely that the cytoplasmic tail of a vesicular transmembrane protein is required for direct anchoring of BDNF vesicles to components of the microtubule transport system to mediate bi-directional movement. Precedence for such a mechanism has been reported for the post-Golgi transport of amyloid precursor protein containing vesicles from the cell body to the terminals of neurons (Inomata et al., 2003; Taru et al., 2002).

Carboxypeptidase E (CPE) is a proneuropeptide/prohormone processing enzyme (Fricker and Snyder, 1982; Hook and Loh, 1984) and is associated with BDNF vesicles in hippocampal and cortical neurons (Lou et al., 2005). The luminal domain of the membrane form of CPE acts as a sorting receptor for targeting BDNF to the regulated secretory pathway vesicles in hippocampal and cortical neurons (Lou et al. 2005). Membrane CPE can also exist in vesicles as a transmembrane protein with a cytoplasmic tail of ~10 amino acids at the C-terminus (Arnaoutova et al., 2003; Dhanvantari et al., 2002; Wu et al., 2004). The CPE cytoplasmic tail was found to interact with an activated form of the cytoplasmic small GTPase, Arf6, to recycle CPE and a CPE binding protein, eosinophil cationic protein, from the plasma membrane to the TGN in stimulated neuroendocrine cells (Arnaoutova et al., 2003). Deletion of the CPE cytoplasmic tail, or point mutations (S472A and E473A) at the CPE tail resulted in loss of interaction with Arf6, and elimination of the recycling of these proteins from the plasma membrane. More recently, the CPE cytoplasmic tail was found to be necessary for anterograde transport of pro-opiomelanocortin (POMC)/adrenocorticotropin (ACTH) vesicles to the secretion site for exocytosis in anterior pituitary cells (Park et al., 2008).

In this study, we have investigated the role of the CPE cytoplasmic tail in the mechanism of bi-directional transport of BDNF vesicle in hippocampal neurons. We show that the cytoplasmic tail of CPE controls transport of BDNF/CPE-containing vesicles to the secretion sites in hippocampal neurons, by recruiting dynactin and kinesin-2/kinesin-3 onto the vesicles. Additionally, the CPE tail interacts with dynactin/dynein to mediate retrograde transport of vesicles from the neurite terminus back to the cell body.


Overexpression of CPE cytoplasmic tail reduces localization of BDNF in neurites

Proper sorting of BDNF from the TGN to the regulated secretory pathway vesicles requires interaction with a CPE luminal domain (Lou et al., 2005). Here, we investigated the role of the CPE tail in post-Golgi trafficking of BDNF vesicles to neurites by overexpressing GFP-CPEC10 in E18 srat hippocampal neurons. Endogenous BDNF vesicles were visualized by immunocytochemistry using anti-BDNF antibody (N-20). Overexpressing GFP-CPEC10 in the cytoplasm is expected to compete with endogenous CPE tails for adaptor protein(s) linked to transport system. Fig. 1A shows that BDNF localization in the neurites was diminished in neurons expressing GFP-CPEC10 compared to GFP, suggesting that the cytoplasmic tail of CPE is involved in the transport of BDNF vesicles into the neurites. A similar negative effect was obtained when GFP-CPEC25 that contains extra 15 amino acids upstsream of CPEC10 was overexpressed in hippocampal neurons (Fig.s 1A). The negative effect of CPEC10 was independent of the GFP tag itself since RFP- CPEC10 gave the same results (Fig. 1B). In contrast, post-Golgi constitutive transport of nerve growth factor {NGF} into neurites (Altar and DiStefano, 1998; Lou et al., 2005) was not affected by overexpression of GFP-CPEC10 (Fig. 1C), indicating that the CPEC10 effect is specific to BDNF vesicle transport to the regulated secretory pathway.

Fig. 1
(A) Distribution of immunoreactive BDNF (red) in hippocampal neurons expressing GFP, GFP-CPEC10, or GFP-CPEC25 (green). Arrows indicate endogenous BDNF vesicles. Scale bar: 5 µm (inset-1µm). (B) Distribution of immunoreactive BDNF (green) ...

In order to quantify the extent of inhibition of localization of BDNF along the neurites of hippocampal neurons by different CPE tail constructs, the ratio of mean intensity of endogenous BDNF in the neurite versus cell body was calculated from 10 pairs of neurons transfected with GFP, GFP-CPE10, GFP-CPEC25, RFP, or RFP-CPE10. Neurons expressing GFP-CPEC10, -CPEC25, or RFP-CPEC10 showed a significant decrease in the ratio of BDNF in the neurites versus cell body {GFP-CPEC10 (~58 %), GFP-CPEC25 (~64 %), and RFP-CPEC10 (~56 %)} compared to control {GFP or RFP} (Fig. 2A), indicating that the C-terminal ten amino acids (CPEC10) of CPE is sufficient to confer the negative effect. Conversely, the CPE tail did not affect the level of NGF in the neurites (Fig. 2B), indicating that the CPE tail does not influence neurite-localization of NGF that is transported via the constitutive secretory pathway. Since GFP-CPEC10 emitted a weaker signal than GFP-CPEC25, the latter was used to increase sample numbers to obtain statistical significance. The average immunostaining intensity of endogenous BDNF in the neurites was compared between 44 pairs of neurons transfected with GFP or GFP-CPEC25.Overexpression of GFP had no effect on the level of BDNF in neurites while GFP-CPEC25 caused an ~40% reduction (p<0.001) in neurite-localization of BDNF (Fig. 2C).

Fig. 2
(A) Bar graph showing the ratio of the mean immunostaining intensity of BDNF in the neurites relative to the cell body, in neurons overexpressing GFP alone, GFP-CPEC10, GFP-CPEC25, RFP alone, or RFP-CPEC10. The intensity of BDNF immunostaining from 10 ...

Overexpression of the CPE cytoplasmic tail inhibits bidirectional BDNF vesicle movement

To determine if the cytoplasmic tail of CPE also affects BDNF vesicle transport, we overexpressed cytoplasmic RFP-CPEC10 in neurons (Fig. 3). In live cell imaging, the movement of BDNF-GFP-containing vesicles in neurites of neurons expressing either RFP-CPEC10 or RFP alone was monitored (Fig. 3, supplemental movie S1 & S2). In control RFP-expressing cells, we observed movement of BDNF-GFP vesicles in both anterograde and retrograde directions (Fig. 3A). The linear increase in the distance each vesicle moved over time strongly indicates that the BDNF-GFP vesicles are carried by microtubule-based motors along microtubules. Overexpression of RFP-CPEC10 reduced the number of BDNF-GFP vesicles that moved (Fig. 3B), as well as the total distances moved by ~50% in both the anterograde and retrograde direction (Fig. 3B & Table 1). The average velocities of both anterograde and retrograde movements in neurons expressing RFP-CPEC10 was decreased by ~15% and by ~30%, respectively (Fig. 3B & Table 1). BDNF vesicle pools showing processive movement in the anterograde, as well as retrograde direction disappeared in neurons expressing RFP-CPEC10 (Fig. 3B–C & Table 1). These results demonstrate that the CPE cytoplasmic tail represses the anterograde movement of BDNF-GFP vesicles to the neurite terminals and the retrograde movement to the cell body.

Fig. 3
(A) Graphs showing time-dependent BDNF-GFP vesicle movement for anterograde (plus y-axis) and retrograde (minus y-axis) directions. The displacement of each BDNF vesicle on the trajectory routes based on x- and y-coordinates was calculated. (B) Graphs ...
Table 1
Hippocampal neurons were transfected with BDNF-GFP and either RFP or RFP-CPEC10, or with CPE-RFP and either GFP or GFP-CPEC25. Real-time images of movement of BDNF-GFP/CPE-RFP-containing vesicles in live cells were taken at an interval of 1.97 sec. The ...

Overexpression of CPEC25 interferes with real-time trafficking of CPE-RFP-containing secretory vesicles in live hippocampal neurons

Since the BDNF-GFP signal is weak, thereby limiting the length of time it was possible to monitor BDNF vesicle movement, we carried out studies using RFP-tagged full length CPE, which gave a stronger signal and allowed longer term monitoring, to further support the kinetics of BDNF vesicle movement. First we showed, as expected, colocalization of endogenous BDNF with transfected full-length CPE that was tagged with RFP at its C-terminus (CPE-RFP). Fig. 4A (insets) shows that CPE-RFP was co-localized with endogenous BDNF in a large number of the vesicles in the cell body and in the neurites. Overexpression of GFP-CPEC25, reduced the localization of CPE-RFP in the neurites as reflected by the ~48% decrease in the number of neurons showing CPE-RFP along the neurites (Fig. 4B & C), similar to BDNF (Fig. 2).

Fig. 4
(A) Confocal images of hippocampal neurons expressing CPE-RFP (red) and immunostained with mouse anti-BDNF antibody followed by anti-mouse-Alexa488 antibody (green). Note the co-localization of BDNF and CPE-GFP in punctate vesicles in the cell body and ...

To verify that the observed decrease in neurite-localization of CPE-RFP containing vesicles is due to inhibition of post-Golgi CPE vesicle trafficking we carried out analysis of real-time images of CPE-RFP-containing vesicle movement in live neurons expressing either GFP alone, or GFP-CPEC25. Live cell images of CPE-RFP-containing vesicle movements in the neurites of neurons expressing either GFP alone or GFP-CPEC25 were analyzed with respect to the average velocity and total distance of the movements. CPE-RFP vesicles in the live neurons expressing GFP alone showed normal bidirectional movement along the neurites, as determined by trajectory routes analyzed in a time-based manner for each vesicle (Fig 5A, supplemental movie S3), compared to neurons expressing GFP-CPEC25 (Fig 5A, supplemental movie S4). The overall number of moving vesicles in each neurite was decreased in neurons overexpressing GFP-CPEC25 (Fig. 5B & C). Moreover, the fast-moving (>0.45 µm/sec) pools of CPE-RFP vesicles traveling in the anterograde and retrograde direction specifically disappeared in the GFP-CPEC25 expressing neurons (Fig. 5B & C and supplemental movie S4). Here, we were able to analyze the processivity of CPE-RFP vesicles for a longer distance and duration than with BDNF-GFP vesicles. The processivity and velocity of both anterograde and retrograde movements were significantly decreased by overexpression of GFP-CPEC25 (Fig. 5B & C and Table 1). These results confirm those of BDNF-GFP containing vesicles showing that the C-terminal tail of CPE plays a mechanistic role in the post-Golgi anterograde transport of BDNF/CPE containing vesicles to the neurite terminals for release and in the retrograde transport back to the cell body.

Fig. 5
(A) Graphs showing time-dependent CPE-RFP vesicle movement for anterograde (plus y-axis) and retrograde (minus y-axis) directions. The displacement of each CPE vesicle on the trajectory routes based on x- and y-coordinates was calculated (B) Graphs showing ...

The CPE cytoplasmic tail interacts with dynactin that is connected to kinesins and dynein

To identify motor or motor-related proteins that interact directly or indirectly with the cytoplasmic tail of CPE to couple the vesicles to the microtubule-based motor system, co-precipitation assays using recombinant GST-tagged CPEC10 and mouse brain cytosol were carried out. GST-CPEC10 specifically pulled down dynactin (p150 subunit) and the anterograde motors, KIF3A (kinesin-2) and KIF1A (kinesin-3), and dynein (DIC), a retrograde motor (Fig. 6A). Kinesin-1 probed by two different anti-KHC antibodies (goat and rabbit SUK4) was not detected. HAP1 was also found in the GST-CPEC10 co-precipitate, but huntingtin (Htt) was not detected in the pulldown. Silver staining showed that there was minimum background binding (Fig. 6A). Next, in order to determine what percentage of those CPE tail-interacting proteins among the total proteins in the cytosol were pulled down by GST-CPEC10, we compared the levels of cytosolic proteins that bound to GST-CPEC10 versus those that remained in the cytosol after GST-CPEC10 pulldown using densitometry (Fig. 6B). We found that ~51% of dynactin, ~42% of kinesin-2, ~51% of kinesin-3, ~42% of dynein, and ~30% of HAP1 were among the cytosolic population bound to GST-CPEC10, while little kinesin-1 (KHC) or Htt were bound.

Fig. 6
(A) Immunoblots and silver staining showing proteins pulled down by GSTCPEC10 {G-CPEC10} or GST alone {G} from mouse brain cytosol. {MW: p150 = 150 kD, DIC = 74 kD, KIF3A = 80–85 kD, KIF1A: 200 kD, KHC = 120 kD, HAP1 = 75–85 kD, Htt = ...

To complement the GST pulldown experiment, co-immunoprecipitation was performed using a rabbit antibody against p150, a core subunit of dynactin. Immunoblotting of total cell extracts using rabbit anti-p150 antibody showed specific recognition of p150 by the antibody (Fig. 7A). As shown in the Coomassie stained protein gel (Fig7B), equivalent amounts of IgG heavy (55kD) and light (27kD) chains and only minimal amount of other proteins were precipitated by either the control IgG or p150 antibodies. However, the Coomassie staining did show a clear p150 band precipitated by the p150 antibody, indicating a successful pulldown of dynactin. Dynactin was immunoprecipitated along with endogenous CPE, KIF3A, KIF1A, dynein, and HAP1, but not KHC (Fig. 7B). In order to examine whether endogenous CPE binds to dynactin via its cytoplasmic tail, we conducted a peptide-based competition assay using excess CPEC10 and scrambled peptides (Fig. 7C). Excess CPEC10, but not scrambled peptide, displaced the endogenous CPE but not the motor proteins from dynactin (Fig. 7C), indicating that KIF3A, KIF1A, and dynein interact with dynactin and not the CPE tail. This result indicates that the C-terminal cytoplasmic tail of CPE (CPEC10) is responsible for the interaction between CPE and dynactin. The specificity of the interaction was further demonstrated by the ability of I125-labeled CPEC10 to bind dynactin in a saturable and displaceable manner (Fig. 7D).

Fig. 7
(A) Immunoblot of mouse brain cytosol showing specific immunostaining of the dynactin band with anti-p150 (B) Immunoblot showing proteins co-immunoprecipitated with dynactin (p150) (IP: immunoprecipitation, IgG: rabbit IgG, p150: anti-p150 IgG). Coomassie ...


The neurotrophic factor, BDNF, is sorted into regulated secretory pathway vesicles at the TGN using CPE as a sorting receptor, (Lou et al., 2005), as well as sortilin (Chen et al., 2005) in hippocampal and cortical neurons. The newly formed BDNF vesicles are then transported to the neurite terminals and undergo exocytosis upon stimulation to release BDNF (Thoenen, 2000). In this study, we have investigated the role of the CPE cytoplasmic tail in the post-Golgi movement of BDNF vesicles in hippocampal neurons. Our strategy is cell biological: to overexpress the CPE cytoplasmic tail (CPEC10) in the cytoplasm of neurons and assess its effect on the distribution and movement of vesicles containing BDNF/CPE in live neurons. We showed that overexpression of CPEC10 resulted in a significant reduction of BDNF in the neurites at steady state (Fig. 1), as well as a reduction in the velocity and distance of BDNF/CPE vesicle movement in hippocampal neurons (Fig. 3). These findings demonstrate that the CPE cytoplasmic tail functions to anchor BDNF vesicles to the microtubule-based transport system. The range (0.1–0.8 µm/s) of the average velocities of the BDNF/CPE vesicle movement in hippocampal neurons was similar to that of BDNF vesicle movement in the mouse cortical neurons reported in previous studies by Adachi et al. (2005). Since we did not have available markers for the axons and dendrites of live neurons, we were not able to distinguish between axonal versus dendritic movements of BDNF vesicles; however we observed that overexpression of CPEC10 decreased BDNF/CPE vesicle movements in most of the neurites to < 0.45µm/s. Given that Adachi et al. (2005) showed that the fast-moving vesicles (~0.73 µm/s) were found mainly in the axons while the slow-moving vesicles (~0.47 µm/s) in the dendrites, it would appear that overexpression of the CPE cytoplasmic tail affect both the fast-moving BDNF/CPE vesicles in the axons as well as the slow-moving vesicles in the dendrites.

Knowledge of the microtubule-based motor protein required for anterograde transport of CPE-containing vesicles first came from genetic studies in C. elegans. Mutation of the unc-104 protein, an ortholog of KIF1 (kinesin 3), caused a defect in anterograde movement of egl-21 (CPE) vesicles (Jacob and Kaplan, 2003). To investigate whether KIF1A and what other motor proteins might be involved in BDNF/CPE vesicle transport, we carried out co-precipitation, co-immunoprecipitation, competition and CPEC10 peptide binding assays with mouse brain cytosol. These assays suggested that the CPEC10 domain is where dynactin binds, either directly or indirectly. We propose that dynactin may recruit anterograde (kinesin-3 {KIF1A}, kinesin-2 {KIF3A}) and retrograde (dynein) microtubule motors onto the vesicular CPEC10 tail on the basis of the projected orientation of dynactin and the motor proteins (Schroer, 2005). HAP1 that was known to facilitate dynactin-based BDNF transport (Gauthier et al., 2004) was also found to associate with the CPE-dynactin-motor complex. However, Htt which was previously shown to bind HAP1 (Gauthier et al., 2004) was not associated with the CPE-dynactin complex. This may perhaps be because it was loosely associated and was easily removed by our biochemical procedure, or there was no association with the CPE-dynactin complex. The exact role of HAP1 in the CPE-dynactin complex in anterograde transport of BDNF vesicles is not clear since that study did not identify any kinesin motor proteins associated with HAP1 or Htt (Gauthier et al., 2004). Dynactin recruits both the kinesins and dynein, and depending on which is recruited at the time, these interactions would result in either anterograde or retrograde transport of BDNF vesicles respectively. Previous studies have also shown that dynactin can bind kinesin-2 and dynein (Berezuk and Schroer, 2007), and kinesin-3 (Park et al., 2008) in endocrine cells. In this study, we demonstrate that in neurons, dynactin also acts an adaptor for the kinesin motor protein, KIF1A, which was known to be involved in microtubule-based transport of regulated secretory vesicles, (Klopfenstein and Vale, 2004).

Dynein interacts with dynactin to mediate retrograde transport, since this microtubule-based motor is associated with minus-end movement. A recent live cell imaging study showed that excess neuropeptide-containing vesicles are transported from the site of biogenesis in the soma to the synaptic terminus, but if not trapped at the synapse for release during stimulation, the unused vesicles are transported back to the cell body in a retrograde fashion, presumably for degradation or reuse (Shakiryanova et al., 2006). This phenomenon may also be true for BDNF vesicles since the primary site of BDNF synthesis is at the cell body, unlike classical neurotransmitter vesicles which are recycled and replenished at the synapse (Sudhof, 2004). Excess BDNF vesicles transported to the neurite terminal that are not captured by stimulation would then be expected to return to the cell body for degradation, or perhaps even reuse, to maintain homeostasis. While retrograde transport of BDNF from the postsynaptic terminal has been reported (Bhattacharyya et al., 2002; Curtis et al., 1995; DiStefano et al., 1992; Reynolds et al., 2000; Yano and Chao, 2004), it is unlikely that the BDNF-GFP vesicles we observed moving in the retrograde direction in hippocampal neurons represent vesicles that was released and endocytosed, since we were studying unstimulated neurons. Besides, the BDNF-GFP signal in such putative vesicles, if generated, would be far too weak to be detected. Dynein’s interaction with dynactin in the CPE tail-associated complex can easily serve a role in retrograde transport of excess BDNF vesicles in non-stimulated cells, by switching the microtubule association of kinesins to dynein. Indeed, we showed that overexpression of CPEC10 also reduced retrograde movement of BDNF vesicles, although the factors implementing the switch of the interaction of dynactin with kinesins to dynein to reverse direction of BDNF vesicles remain to be explored. However, we also considered the possibility that the observed decrease in the retrograde BDNF vesicle movement could in part be a consequence of a reduction in the number of BDNF vesicles that reach the neurite terminals since the anterograde movement of BDNF vesicles are inhibited by overexpression of CPEC10. Nevertheless, a pool of BDNF vesicles does reach the neurite terminals even in the presence of CPEC10 (Fig. 1), and hence it is expected that this pool would return to the cell body.

The anterograde transport of BDNF vesicles from the soma along the neurites to the synapses for activity dependent-secretion is necessary for mediating various forms of synaptic plasticity, including long-term potentiation (LTP), learning and memory (Egan et al., 2003; Katz and Shatz, 1996; Lu, 2003; Poo, 2001; Thoenen, 2000). Indeed, a CPE knockout mouse showed a lack of regulated secretion of BDNF (Lou et al., 2005), absence of LTP in hippocampal neurons and deficit in learning and memory (Woronowicz et al., manuscript submitted), consistent with the multiple cell biological functions of CPE in BDNF sorting and BDNF vesicle transport. Interestingly, constitutive secretion of proBDNF caused by lack of CPE and sorting to the regulated secretory pathway was unimpaired in these CPE knockout mice, indicating that constitutive secretion is CPE–independent (Lou et al., 2005).

In conclusion, we have uncovered a mechanism requiring the interaction of the CPE cytoplasmic tail with dynactin which in turn recruits kinesin 2 (KIF3A) and kinesin 3 (KIF1A) motor proteins for anterograde transport, or dynein for retrograde transport of BDNF/CPE vesicles in hippocampal neurons (Fig. 8). Such a bi-directional transport mechanism is critical not only for activity-dependent secretion of BDNF required for synaptic plasticity, learning and memory (Egan et al., 2003; Thoenen, 2000), but also for the maintenance of BDNF vesicle homeostasis in the neuron.

Fig. 8
A model showing (A) that the cytoplasmic tail of transmembrane carboxypeptidase E (CPE) in BDNF/CPE vesicles recruits dynactin which then associates with and confers processivity to KIF1A (kinesin-3) and KIF3A (kinesin-2). KIF1A and KIF3A, plus-end microtubule-based ...

Experimental methods

DNA constructs

A 5’-EcoRI-XhoI-3’ digest of PCR product of CPE tail fragments (CPEC10) amplified from full-length CPE cDNA were subcloned into pGEX4T-2 (GST-CPEC10; Amersham Bioscience, San Diego, CA). A 5’-XhoI-PstI-3’ digest of the PCR product of CPEC10 or CPEC25 was cloned to pEGFP-C1 or pDsRed2 (GFP-CPEC10, GFP-CPEC25, or RFP-CPEC10; BD Bioscience, San Diego, CA). Full-length CPE cDNA was excised from pEGFP-N1 and inserted into pDsRed-Express-N1 (BD Bioscience Clontech Co.) to generate CPE-RFP. BDNF-GFP was gift from Dr. Bai Lu (NICHD).


The rabbit antibodies to kinesin heavy chain (SUK4), p150Glued, BDNF (N-20), huntingtin (htt), HAP1, NGF, the goat antibody to kinesin heavy chain, and the mouse antibody to dynein (74.1) were from Santa Crutz Biotechnology (Santa Crutz, CA). The specificity of the N-20 antibody for BDNF was verified by showing that it stained the hippocampal neurons that contain BDNF, but not pheochromocytoma PC12 cells and anterior pituitary AtT20 cells, both of which do not have endogenous BDNF. (supplemental Fig. S3). The mouse antibodies to CPE, KIF3A, or KIF1A were from Transduction (San Jose, CA). Mouse Ab to ACTH was from Abcam (Cambridge, MA). The rabbit antibody to the C terminus of CPE (Cool et al., 1997) was generated in our laboratory.

Cell culture and immunocytochemistry

E18 rat primary hippocampal neurons (Gene Therapy Systems, San Diego, CA) were completely dissociated by pipetting, and then subjected to transfection by electroporation. Cells (5 × 106) were centrifuged at 800 × g (10 min) and then resuspended in 10 µg of DNA and 100 µl of rat neuron nucleofector solution for electroporation (program: C-13; Amaxa Inc., Gaithersburg, MD). Electroporated cells were seeded onto poly-D-lysine-coated coverslips in DMEM medium plus 10% FBS and incubated for 2 days, and in neurobasal/B27 medium for an additional 2 days prior to experiments. Transfection efficiency of <10% was usually obtained.

AtT20 cells or PC12 cells were grown in DMEM medium supplemented with 10 % FBS or 10 % FBS plus 5 % horse serum, respectively. 2 × 104 cells were seeded on 15 mm2 round coverslips and grown for 18–20 h in DMEM medium before processing for immunocytochemistry.

For immunocytochemistry, cells on the coverslip in a 35mm culture dish were rinsed with PBS, fixed in 3.5% formaldehyde in PBS for 30 min, and permeabilized in 0.1% Triton X-100 in PBS for 30 min. Cells were then blocked in TTBS (TBS, 0.1 % Tween-20, and 2% BSA), incubated for 30 min in primary antibodies [anti-BDNF antibody (N-20) at a dilution of 1:300 or anti-ACTH antibody at 1:250], washed in TTBS (3 × 5 min), and incubated in Alexa Green (488 nm)/Red (568 nm) secondary antibody at a dilution of 1:1000 for 15 min. Samples were washed again and mounted on slides in GEL/MOUNT (Biomeda, Foster city, CA).


The intensity of BDNF immunostaining in the neurites was quantified using a Zeiss Axiovert 200 M inverted microscope (Carl Zeiss, Thornwood, NY) equipped with 63X Zeiss plan-apochromat oil, 1.4 NA, DIC and 100x Zeiss alpha plan fluor oil, 1.45 NA, DIC objectives. Images were acquired by a Meta detector for spectral imaging (Carl Zeiss) and digitized using ‘LSM 510 Meta’ software version 3.5 (Carl Zeiss) and quantification performed using ‘Metamorph’ software (Molecular Devices, Downingtown, PA).

For time-lapse imaging, hippocampal neurons expressing CPE-RFP and GFP/GFP-CPEC10 or BDNF-GFP and RFP/RFP-CPEC10 were maintained in phenol red-free DMEM plus 10% FBS in a temperature (37°C)-controlled Delta T chamber (Bioptechs Inc., Butler, PA) and imaged using the Zeiss Axiovert 200 M inverted microscope and ‘LSM 510 Meta’ software at 1.97-second exposure/shot for 200 shots. Image pixels were converted to µm using ‘LSM 510 Meta’ software and the displacement of each BDNF vesicle on the trajectory routes based on x- and y-coordinates was calculated. The velocity of vesicle movement was calculated from each displacement per 1.97 sec while the total distance was the summation of all of the displacements each vesicle moved. Vesicles moving in either anterograde or retrograde direction with respect to the total number of time intervals (1.97 sec) each vesicle moved were sorted separately. Statistical analysis was carried out using the Student t test (two tailed, non-paired) and the mean values ± SEM are reported for the average of total distances and mean velocities that the vesicles moved.

Co-precipitation and co-immunoprecipitation

0.4mg of bacterially expressed GST or GST tagged CPEC10 was purified on 400 µl of glutathione (GSH) beads (Amesham Bioscience). For preparation of mouse brain cytosol, two mouse brains (~0.8 g wet weight) were washed extensively in PMEE buffer (pH 7.0, 35 mM KOH, 35 mM PIPES, 5 mM MgSO4, 1 mM EGTA, 1%BSA, and 0.5 mM EDTA) containing 1 mM AEBSF and 1x protease inhibitor cocktail, minced to small pieces by a razor, and then homogenized in 4mL PMEE using a 16G needle and a 27G needle. After centrifugation at 14,500 × g (10min), the supernatant from mouse brain was centrifuged at 192,000 × g (30 min) at 4°C to obtain a high-speed cytosol fraction (0.1 g/ml protein). Three 800 µl aliquots of cytosol were pre-cleared by 100 µl of the GSH beads. Each aliquot of precleared cell cytosol was mixed with GST or GST-CPEC10 bound to GSH beads and incubated at 4°C for 18 hr on a rotating platform. The beads were washed six times with 0.5 ml of PMEE buffer, incubated in 200 µl of 10 mM GSH for elution, and centrifuged, (1,000 × g, 5 min). The eluate was boiled in SDS loading buffer and loaded onto NuPAGE gel (Invitrogen, Carlsbad, CA) for Western blotting.

For co-immunoprecipitation studies, mouse brain cytosol was prepared as described above for co-precipitation experiments. 800 µl of the mouse brain cytosol fraction (0.1g/ml protein) was split to two sets of 400 µl. After pre-clearing by 100 µl of protein-A agarose beads (Sigma, St. Louis, MO), the cytosol was mixed with either 10 µg of rabbit IgG (control) or rabbit antibody against dynactin (p150) and incubated for 18 hr at 4°C. Then 100 µl of protein A agarose beads were added to the tubes and incubated for 8 hr. After incubation, the beads were washed seven times with PMEE buffer, boiled in 100 µl of SDS loading buffer and processed for immunoblotting. For competition assays in the co-immunoprecipitation study, all procedures were carried out as described above; except that 10µg/ml of custom synthesized (99.9% pure) CPEC10 peptide or a scrambled peptide (YTNEFLMWMKS) (Phoenix Pharmaceutical Co., Burlingame, CA) was added to an equal volume of cytosol before co-immunoprecipitation with dynactin (p150) antibodies.

In order to examine the specificity of binding of the cytoplasmic tail of CPE (CPEC10) to the dynactin co-immunoprecipitate, we carried out in vitro binding studies using I125-radiolabeled CPEC10 peptide and dynactin co-immunoprecipitates prepared as described above. One mL of 0, 2, 4, 6, or 8nM of I125-CPEC10 was added to 50µL of dynactin coimmunoprecipitate-bound protein A beads (~50µg). After incubation for 18 hr at 4°C, the beads were separated from the supernatant and the radioactivity in each of the bead and supernatant samples were counted in a γ-counter. The total amount of I125-CPEC10 in the final binding reaction was calculated from each set of beads and supernatant. The binding of an I125-labeled CPEC10 scrambled peptide, described above, was used as the background binding. The amount of I125-CPEC10 bound to dynactin immunoprecipitate was calculated from the CPM on the beads minus the background binding by the scrambled peptide. To verify the specificity of binding of I125-CPEC10, the binding assay was done in the presence of excess cold CPEC10 (0.79µM) to displace the I125-CPEC10 binding.

Supplementary Material






We thank Dr. Bai Lu (NICHD, NIH), Dr. Joel Rosenbaum (Yale Univ), Dr. Trina Schroer (Johns Hopkins Univ.), and Dr. Jennifer Lippincott-Schwartz (NICHD, NIH) for their suggestions and critical reading of the manuscript. We also thank Drs. Hong Lou, Guhan Nagappan and Hisatsugu Koshimizu, all from NICHD for technical assistance and helpful discussions. We thank Dr. Vincent Schram and Chip Dye in the NICHD Microscopy Imaging Core for their technical support. This research was supported by the Intramural Research Program of the NICHD, NIH.


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


  • Altar CA, DiStefano PS. Neurotrophin trafficking by anterograde transport. Trends Neurosci. 1998;21:433–437. [PubMed]
  • Adachi N, Kohara K, Tsumoto T. Difference in trafficking of brain-derived neurotrophic factor between axons and dendrites of cortical neurons, revealed by live-cell imaging. BMC Neurosci. 2005;21:42. [PMC free article] [PubMed]
  • Arnaoutova I, Jackson CL, Al-Awar OS, Donaldson JG, Loh YP. Recycling of Raft-associated prohormone sorting receptor carboxypeptidase E requires interaction with ARF6. Mol. Biol. Cell. 2003;14:4448–4457. [PMC free article] [PubMed]
  • Berezuk MA, Schroer TA. Dynactin enhances the processivity of kinesin-2. Traffic. 2007;8:124–129. [PubMed]
  • Bhattacharyya A, Watson FL, Pomeroy SL, Zhang YZ, Stiles CD, Segal RA. High-resolution imaging demonstrates dynein-based vesicular transport of activated Trk receptors. J. Neurobiol. 2002;51:302–312. [PubMed]
  • Chen ZY, Ieraci A, Teng H, Dall H, Meng CX, Herrera DG, Nykjaer A, Hempstead BL, Lee FS. Sortilin controls intracellular sorting of brain-derived neurotrophic factor to the regulated secretory pathway. J. Neurosci. 2005;25:6156–6166. [PMC free article] [PubMed]
  • Cool DR, Normant E, Shen F, Chen HC, Pannell L, Zhang Y, Loh YP. Carboxypeptidase E is a regulated secretory pathway sorting receptor: genetic obliteration leads to endocrine disorders in Cpe(fat) mice. Cell. 1997;88:73–83. [PubMed]
  • Curtis R, Adryan KM, Stark JL, Park JS, Compton DL, Weskamp G, Huber LJ, Chao MV, Jaenisch R, Lee KF, et al. Differential role of the low affinity neurotrophin receptor (p75) in retrograde axonal transport of the neurotrophins. Neuron. 1995;14:1201–1211. [PubMed]
  • Deacon SW, Serpinskaya AS, Vaughan PS, Lopez Fanarraga M, Vernos I, Vaughan KT, Gelfand VI. Dynactin is required for bidirectional organelle transport. J. Cell Biol. 2003;160:297–301. [PMC free article] [PubMed]
  • Dell KR. Dynactin polices two-way organelle traffic. J. Cell Biol. 2003;160:291–293. [PMC free article] [PubMed]
  • Dhanvantari S, Arnaoutova I, Snell CR, Steinbach PJ, Hammond K, Caputo GA, London E, Loh YP. Carboxypeptidase E, a prohormone sorting receptor, is anchored to secretory granules via a C-terminal transmembrane insertion. Biochemistry. 2002;41:52–60. [PubMed]
  • DiStefano PS, Friedman B, Radziejewski C, Alexander C, Boland P, Schick CM, Lindsay RM, Wiegand SJ. The neurotrophins BDNF, NT-3, and NGF display distinct patterns of retrograde axonal transport in peripheral and central neurons. Neuron. 1992;8:983–993. [PubMed]
  • Egan MF, Kojima M, Callicott JH, Goldberg TE, Kolachana BS, Bertolino A, Zaitsev E, Gold B, Goldman D, Dean M, Lu B, Weinberger DR. The BDNF val66met polymorphism affects activity-dependent secretion of BDNF and human memory and hippocampal function. Cell. 2003;112:257–269. [PubMed]
  • Fricker LD, Snyder SH. Enkephalin convertase: purification and characterization of a specific enkephalin-synthesizing carboxypeptidase localized to adrenal chromaffin granules. Proc. Natl. Acad. Sci. U S A. 1982;79:3886–3890. [PubMed]
  • Gauthier LR, Charrin BC, Borrell-Pages M, Dompierre JP, Rangone H, Cordelieres FP, De Mey J, MacDonald ME, Lessmann V, Humbert S, Saudou F. Huntingtin controls neurotrophic support and survival of neurons by enhancing BDNF vesicular transport along microtubules. Cell. 2004;118:127–138. [PubMed]
  • Gill SR, Schroer TA, Szilak I, Steuer ER, Sheetz MP, Cleveland DW. Dynactin, a conserved, ubiquitously expressed component of an activator of vesicle motility mediated by cytoplasmic dynein. J. Cell Biol. 1991;115:1639–1650. [PMC free article] [PubMed]
  • Goldstein LS, Yang Z. Microtubule-based transport systems in neurons: the roles of kinesins and dyneins. Annu. Rev. Neurosci. 2000;23:39–71. [PubMed]
  • Goodman LJ, Valverde J, Lim F, Geschwind MD, Federoff HJ, Geller AI, Hefti F. Regulated release and polarized localization of brain-derived neurotrophic factor in hippocampal neurons. Mol. Cell. Neurosci. 1996;7:222–238. [PubMed]
  • Haubensak W, Narz F, Heumann R, Lessmann V. BDNF-GFP containing secretory granules are localized in the vicinity of synaptic junctions of cultured cortical neurons. J. Cell Sci. 1998;111(Pt 11):1483–1493. [PubMed]
  • Hook VY, Loh YP. Carboxypeptidase B-like converting enzyme activity in secretory granules of rat pituitary. Proc. Natl. Acad. Sci. U S A. 1984;81:2776–2780. [PubMed]
  • Inomata H, Nakamura Y, Hayakawa A, Takata H, Suzuki T, Miyazawa K, Kitamura N. A scaffold protein JIP-1b enhances amyloid precursor protein phosphorylation by JNK and its association with kinesin light chain 1. J. Biol. Chem. 2003;278:22946–22955. [PubMed]
  • Jacob TC, Kaplan JM. The EGL-21 carboxypeptidase E facilitates acetylcholine release at Caenorhabditis elegans neuromuscular junctions. J. Neurosci. 2003;23:2122–2130. [PubMed]
  • Katz LC, Shatz CJ. Synaptic activity and the construction of cortical circuits. Science. 1996;274:1133–1138. [PubMed]
  • Klopfenstein DR, Vale RD. The lipid binding pleckstrin homology domain in UNC-104 kinesin is necessary for synaptic vesicle transport in Caenorhabditis elegans. Mol. Biol. Cell. 2004;15:3729–3739. [PMC free article] [PubMed]
  • Lou H, Kim SK, Zaitsev E, Snell CR, Lu B, Loh YP. Sorting and activity-dependent secretion of BDNF require interaction of a specific motif with the sorting receptor carboxypeptidase E. Neuron. 2005;45:245–255. [PubMed]
  • Lu B. BDNF and activity-dependent synaptic modulation. Learn. Mem. 2003;10:86–98. [PubMed]
  • Mowla SJ, Pareek S, Farhadi HF, Petrecca K, Fawcett JP, Seidah NG, Morris SJ, Sossin WS, Murphy RA. Differential sorting of nerve growth factor and brain-derived neurotrophic factor in hippocampal neurons. J. Neurosci. 1999;19:2069–2080. [PubMed]
  • Park JJ, Cawley NX, Loh YP. Carboxypeptidase E Cytoplasmic Tail-Driven Vesicle Transport Is Key for Activity-Dependent Secretion of Peptide Hormones. Mol. Endocrinol. 2008 {Epub ahead} [PubMed]
  • Poo MM. Neurotrophins as synaptic modulators. Nat. Rev. Neurosci. 2001;2:24–32. [PubMed]
  • Reynolds AJ, Bartlett SE, Hendry IA. Molecular mechanisms regulating the retrograde axonal transport of neurotrophins. Brain Res. Brain Res. Rev. 2000;33:169–178. [PubMed]
  • Ross JL, Wallace K, Shuman H, Goldman YE, Holzbaur EL. Processive bidirectional motion of dynein-dynactin complexes in vitro. Nat. Cell Biol. 2006;8:562–570. [PubMed]
  • Schroer TA. Dynactin. Annu. Rev. Cell. Dev. Biol. 2004;20:759–779. [PubMed]
  • Shakiryanova D, Tully A, Levitan ES. Activity-dependent synaptic capture of transiting peptidergic vesicles. Nat. Neurosci. 2006;9:896–900. [PubMed]
  • Sudhof TC. The synaptic vesicle cycle. Annu. Rev. Neurosci. 2004;27:509–547. [PubMed]
  • Taru H, Iijima K, Hase M, Kirino Y, Yagi Y, Suzuki T. Interaction of Alzheimer's beta -amyloid precursor family proteins with scaffold proteins of the JNK signaling cascade. J. Biol. Chem. 2002;277:20070–20078. [PubMed]
  • Thoenen H. Neurotrophins and activity-dependent plasticity. Prog. Brain Res. 2000;128:183–191. [PubMed]
  • Wu CM, Chang HT, Chang MD. Membrane-bound carboxypeptidase E facilitates the entry of eosinophil cationic protein into neuroendocrine cells. Biochem. J. 2004;382:841–848. [PubMed]
  • Yano H, Chao MV. Mechanisms of neurotrophin receptor vesicular transport. J. Neurobiol. 2004;58:244–257. [PubMed]
  • Yano H, Lee FS, Kong H, Chuang J, Arevalo J, Perez P, Sung C, Chao MV. Association of Trk neurotrophin receptors with components of the cytoplasmic dynein motor. J. Neurosci. 2001;21:RC125. [PubMed]