| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Institute of Cell Biology and Neurobiology, Center for Anatomy, Charité–Universitätsmedizin Berlin, Berlin, Germany; and
Institute of Clinical Chemistry and
Institute of Cell Biochemistry and Clinical Neurobiology, University Hospital Hamburg-Eppendorf, Hamburg, Germany
1Correspondence: Institute of Cell Biology and Neurobiology, Center for Anatomy, Charité–Universitätsmedizin Berlin, Charitéplatz 1, D-10117 Berlin, Germany. E-mail: stefan.schumacher{at}charite.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: neuronal differentiation • dendritic tree elaboration • spine morphogenesis • Akt signaling • mass spectrometry
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3
4)
. Although a great number of proteins have been identified as regulators of dendritic morphogenesis and spine maturation (5
6
7)
, much remains to be elucidated about the involved molecular mechanisms. We recently published that the neural epidermal growth factor (EGF) family member chicken acidic leucine-rich EGF-like domain-containing brain protein/neuroglycan C (CALEB/NGC) is a critical mediator in the formation of dendritic arbors and spines in vitro and in vivo (8)
. CALEB/NGC induces dendritic branching of primary neurons in culture and in the intact mouse cortex. Additionally, it increases the density and complexity of dendritic spines. Although the extracellular EGF-like domain of CALEB/NGC is necessary and sufficient for both the induction of dendritic branching and spine formation, the intracellular signaling pathways necessary for the respective processes differ. While CALEB/NGC-mediated dendritic branching is dependent on the activity of the phosphatidylinositide 3-kinase (PI3K) -Akt-mammalian target of rapamycin (mTOR) signaling pathway, spine formation induced by CALEB/NGC is independent of active PI3K (8)
. To date, investigations of the proteins known to interact with CALEB/NGC have not shed light on this divergence. For instance, binding of the extracellular matrix proteins tenascin-C and tenascin-R to CALEB/NGC appears to be dispensable because both proteins interact with the acidic peptide segment and not with the EGF-like domain (9
, 10)
. The only established intracellular interaction partner of CALEB/NGC, the Golgi-associated PDZ domain protein PIST, binds to the juxtamembrane cytoplasmic peptide segment of CALEB/NGC (11)
. However, preliminary data from our laboratory do not support the notion that PIST has a role in regulating CALEB/NGC-induced dendritic branching or spine formation.
To more precisely define the molecular mechanisms required for dendritic tree elaboration and spine morphogenesis induced by CALEB/NGC, we extended the investigation of this issue by performing a yeast two-hybrid screen for further intracellular proteins that may bind to CALEB/NGC. We identified B56β, a regulatory subunit of protein phosphatase 2A (PP2A), as an interaction partner of CALEB/NGC. PP2A is a heterotrimeric protein serine/threonine phosphatase consisting of a 36 kDa catalytic C subunit, a 65 kDa structural A subunit, and a variable regulatory B subunit (12)
. The B subunits determine the subcellular localization and substrate specificity of the enzyme. They are subdivided into three major unrelated families: PR55 (B), B56 (B'/PR61), and PR72 (B''). Of these, the B56 family is the most diverse family, comprising five different genes: B56
, β, 
, 
, and 
(13
14
15)
. PP2A regulators of the B56 family are differentially expressed with high expression of B56β in the brain (13)
. Major functions attributed to B56 family members are inhibition of Wnt signaling by regulating β-catenin degradation (16
17
18)
, modulation of apoptosis (19
, 20)
, and control of extracellular-related kinase (ERK) and Akt activity (21
, 22)
.
Here, we report on the interaction between CALEB/NGC and B56β discovered in a yeast two-hybrid screen. This interaction was confirmed by coimmunoprecipitations, blot-overlay assays, and colocalization studies in primary hippocampal neurons. Results obtained by affinity chromatography in combination with mass spectrometry demonstrated that, in addition to B56β, the structural A subunit and the catalytic C subunit could be recruited to CALEB/NGC via B56β. Dendritic phosphorylation of Akt stimulated by CALEB/NGC was fully inhibited by B56β. Although B56β interfered with CALEB/NGC-induced stimulation of dendritic branching, which is dependent on active Akt, it did not inhibit CALEB/NGC-mediated spine formation, which we found to be independent of Akt. Taken together, we describe a role of B56β in the regulation of dendritic branching. Our results contribute to a more detailed understanding of signaling specificity, which is necessary for independent regulation of dendritic branching and spine formation by CALEB/NGC in consecutive developmental events.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
14
15)
. For mapping the interaction between CALEB/NGC and B56β, different CALEB/NGC constructs were cloned into the vectors pAS2–1 and pACT2, respectively. Primers for the B56β construct were as follows: 5'-CACCCATGGAGACGAAGCTGCCCCCTGC-3' and 5'-CACCTCGAGCTAGCTCTGACCCCCACTGGC-3'. Primers for CALEB/NGC constructs were used as described previously (11)
.
Blot-overlay assay and affinity chromatography
Glutathione-S-transferase (GST) and a fusion protein composed of GST and the total cytoplasmic part of the longer isoform of human CALEB/NGC (hCALEBb) were bacterially expressed in BL21-AITM Escherichia coli (Invitrogen, Karlsruhe, Germany) and purified with a glutathione Sepharose affinity matrix (Amersham Biosciences, Piscataway, NJ, USA) according to the manufacturer’s instructions. Purified recombinant proteins were prepared for SDS-PAGE, transferred to a nitrocellulose membrane, and incubated with detergent extracts of HEK293 cells. HEK293 cells expressing full-length human B56β, NH2-terminally tagged with the myc epitope, were homogenized 2 days after transfection in extraction buffer (Tris-buffered saline, pH 7.4, containing 1.2% Triton X-100, protease inhibitor cocktail, 10 mM NaF, 10 mM Na-pyrophosphate, 1 mM Na-orthovanadate, and 2 mM EDTA). Cellular debris was sedimented by centrifugation, and the supernatant was used for the blot overlay. Bound myc-B56β was detected with an antibody to the myc-epitope (Mab 9E10 anti-myc, Roche, Munich, Germany) and an alkaline phosphatase-conjugated secondary antibody. For the peptide competition assay, the incubation of the nitrocellulose membrane with detergent extract of transfected HEK293 cells was performed in the presence of either 2 µM rCAL1 or rMas1 peptide (see below).
For affinity chromatography, PBS extracts from adult brain (extraction buffer containing PBS supplemented with protease inhibitors) were incubated with bacterially expressed GST-hCALEBb-cyt fusion protein or GST immobilized to cyanogen bromide-activated Sepharose 4 Fast Flow (Amersham Biosciences). After being washed extensively with PBS, proteins were eluted either with high-salt buffer (Tris-buffered saline containing 2 M sodium chloride) or with low-pH buffer (0.1 M glycine, pH 2.5). Elutions were done consecutively. Elution fractions were separated with SDS-PAGE and analyzed by Western blot using anti-B56β antibody [PP2A-B56β (C-19), Santa Cruz Biotechnology, Heidelberg, Germany] or stained with colloidal Coomassie blue (Roth, Karlsruhe, Germany). Bands of interest were isolated, digested with trypsin, and analyzed by mass spectrometry using electrospray ionization tandem mass spectrometry. Mass spectroscopic analysis was performed with a hybrid tandem mass spectrometer (QTOF II; Micromass, Manchester, UK) equipped with a nanoelectrospray ion source. Samples were purified by binding to C18 reverse phase material in a pipette tip (ZipTip; Millipore, Bedford, MA, USA) and eluted with 1 µl of 50% methanol, 5% formic acid into a gold-coated borosilicate vial for nanoelectrospray measurements.
Proteins, peptides, and antibodies
Polyclonal antibodies to CALEB/NGC were directed to a bacterially produced fusion protein composed of maltose-binding protein (MBP) and the extracellular part of mouse CALEB/NGC (MBP-mCALEBextra). This recombinant protein was purified using amylose affinity resin (New England Biolabs, Frankfurt, Germany) according to the manufacturer’s instructions. The antibodies were produced in rabbits and affinity purified using recombinant MBP-mCALEB-extra protein. Polyclonal antibodies to B56β were directed to an internal peptide segment of B56β (VDGFSRRSLRRARP). They were produced in rabbits and affinity purified using immobilized peptide. The rCAL1 peptide used for precipitating recombinant HA-B56β from extracts of transfected HEK293 cells was described previously (11)
. The control peptide rMas1 had the amino acid sequence NTVSIETVV derived from the Mas1 protein (30)
.
Immunoprecipitations, SDS-PAGE, and immunoblots
HEK293 cells were transfected with FuGene 6 (Roche) according to the manufacturer's instructions. Two days after transfection, cells were lysed and extracted as described above. The cell lysates were incubated with anti-HA agarose (Sigma-Aldrich, St. Louis, MO, USA) for 4 h with constant agitation. The agarose beads were washed several times with extraction buffer and prepared for SDS-PAGE. For precipitation experiments with the immobilized peptides described above, detergent extracts of transfected HEK293 cells were incubated with the peptides immobilized to N-hydroxysuccinimide-activated Sepharose (Amersham Biosciences, Piscataway, NJ, USA) for 4 h at 4°C with constant agitation. The Sepharose beads were washed several times with extraction buffer and prepared for SDS-PAGE. After an SDS-PAGE was performed with 8 or 10% acrylamide gels under reducing conditions, the subsequent Western blot was analyzed with either a monoclonal antibody to the HA epitope (Mab HA-7 anti-HA, Sigma-Aldrich) or a monoclonal antibody to the FLAG epitope (Mab M2 anti-FLAG, Sigma-Aldrich) and an alkaline phosphatase-conjugated secondary antibody.
Plasma membrane preparations from P16 rat brains were performed as described previously (23)
. For immunoprecipitations starting from solubilized membrane fractions of P16 rat brain, the lysates were incubated with either affinity-purified anti-B56β polyclonal antibodies or anti-FLAG polyclonal antibodies (Sigma-Aldrich) for 4 h at 4°C. After further incubation with EZview red protein A affinity gel (Sigma-Aldrich) 2 h at 4°C, the agarose beads were washed five times with solubilization buffer (Tris-buffered saline, pH 7.4, containing 1.2% Triton X-100 and supplemented with protease inhibitor cocktail, 10 mM NaF, 10 mM Na-pyrophosphate, 1 mM Na-orthovanadate, and 2 mM EDTA) and then prepared for SDS-PAGE. The subsequent Western blot was analyzed with a monoclonal antibody to CALEB/NGC (BD Biosciences, Heidelberg, Germany) and the polyclonal antibody to B56β mentioned above.
Cell culture and transfection
HEK293 cells were maintained as described previously (11)
. With slight modifications, hippocampal neurons were prepared and cultivated from embryonic day 18–19 Wistar rat pups as described previously (31
, 32)
. Hippocampal cells were transfected with Effectene (Qiagen, Hilden, Germany) according to the manufacturer's instructions with slight modifications. As indicated, Akt III inhibitor (25 µm; Sigma-Aldrich) was added to the cultures 3 h after transfection.
The construct mCALEBb represents full-length mouse CALEB/NGCb with an amino-terminal 3x FLAG tag and a carboxyl-terminal myc epitope, cloned into the expression vector p3xFLAG-Myc-CMV-25 (Sigma-Aldrich). Alternatively, full-length mouse CALEB/NGCb was cloned into the expression vector p3xFLAG-CMV-13 (Sigma-Aldrich). Construct 388 is similar to mCALEBb but lacks the part of the sequence starting from amino acids NKFRTPSE to the carboxyl-terminus. Construct 400 is similar to mCALEBb but lacks the part of sequence starting from amino acids VRKFCDTP to the carboxyl terminus. Coding sequences for B56β and PR55 were cloned into either expression vector pCMV-myc or pCMV-HA (both Clontech-Takara Bio Europe).
Immunocytochemistry
Primary hippocampal neurons were stained for CALEB/NGC with a monoclonal antibody to neuroglycan C (Mab anti-NGC, BD Biosciences) and for B56β with the affinity-purified polyclonal antibody to B56β described above. After fixation, cells were blocked in PBS supplemented with 5% fetal calf serum and permeabilized in the same solution containing 0.2% Triton X-100. The incubations with the primary antibodies were performed for 2 h at room temperature in blocking solution. After being washed with PBS containing 0.1% Triton X-100, the cells were incubated with the secondary antibodies in blocking solution for 1 h at room temperature. As secondary antibodies, we used Cy3-conjugated goat anti-mouse (Dianova, Hamburg, Germany) and Alexa488-conjugated goat anti-rabbit (Molecular Probes, Karlsruhe, Germany) according to the manufacturers’ instructions. Cells were mounted in glycerol gelatin (Sigma-Aldrich) and imaged with a fluorescence microscope (BX50; Olympus, Hamburg, Germany) equipped with a Cool SNAP ES digital camera (Roper Scientific, Ottobrunn, Germany). For fluorescence imaging, the filters U-MWIG, U-MNIBA, and U-MWU2 (Olympus) were used. For close-up views, an oil immersion objective (PLAN APO x60, 1.4 NA) was used.
Image acquisition and data analysis
Hippocampal neurons in culture were transfected and fixed at the indicated time points. Enhanced green fluorescent protein (EGFP) -transfected cells were stained with either a polyclonal antibody to GFP (ab6556, Abcam, Cambridge, UK) or a monoclonal anti-GFP antibody (Mab anti-GFP, clone LGB-1, Biozol, Eching, Germany), and cultures transfected with CALEB/NGC were stained with the monoclonal antibody to the FLAG epitope described above. When double labeling was performed with B56β or PR55, cells were stained with either a monoclonal antibody to the FLAG epitope or a polyclonal antibody to the myc epitope (mCALEBb) and polyclonal or monoclonal antibodies to the myc or HA epitopes (B56β and PR55). Images were captured on a Leica TCS SP2 scanning confocal microscope with 40x HCX PL APO, 1.25 NA, or 63x HCX PL APO, 1.4 NA oil immersion objectives. For dendritic tree and spine analysis, z-stacks were collected at 0.7 and 0.2 µm intervals. Dendritic trees of all healthy transfected neurons with more than five dendritic end tips were analyzed. All end tips of dendrites longer than 8 µm were counted. We used immunohistochemistry against GFP to clearly stain all dendritic branches, because in some cases autofluorescence of GFP is too weak to detect all dendritic branches. Axons could be excluded because of their length, their caliber, and their branching pattern. When we examined the morphology of dendritic spines, spine length was measured from the tip of the protrusion to the point where it meets the dendritic shaft. All protrusions between 0.25 and 6.0 µm in length were regarded as spines, provided that they had a head or head-like structure with a diameter of at least 0.2 µm.
When quantifying the intensity of the phospho-Akt (Ser-473) immunofluorescence signals in dendrites of neurons expressing EGFP or coexpressing EGFP and mCALEBb, EGFP and B56β, or mCALEBb and B56β, we used the same acquisition settings for each set of control and experimental samples. All of the imaging was done in the same session for each experiment. Data acquisition for these experiments was done with wide-field epifluorescence using the equipment described above and Metamorph image analysis software (Universal Imaging Corporation, Ypsilanti, MI, USA). The anti-phospho-Akt monoclonal antibody 587F11 (Cell Signaling Technology, Danvers, MA, USA) was used for these experiments. Control experiments to examine total Akt expression were done with an anti-Akt antibody (Cell Signaling Technology).
Acquisition and data analysis for all experiments were performed by investigators blind to the experimental conditions. Measured data were exported to Excel software (Microsoft, Redmond, WA, USA). All results are means ± SE. Comparison of data and calculation of P values were carried out using 2-tailed Student's t test.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
14
15)
. To narrow down the binding site within the cytoplasmic part of CALEB/NGC, we used B56β as bait and analyzed various CALEB/NGC-derived constructs as prey in the yeast two-hybrid system (Fig. 1
). We found that the following isoform parts of CALEB/NGC bound to B56β: the intracellular part of the shorter isoform of human CALEB/NGC (hCALEBa), which lacks the peptide segment C; the shorter isoform of chicken CALEB/NGC (CALEBa), which only contains the peptide segments A and B of the cytoplasmic part of CALEB/NGC; and the longer isoform of chicken CALEB/NGC (CALEBb), including a D' peptide segment slightly different from the D segment of human CALEB/NGC (9)
. In contrast, the CALEB/NGC-derived construct CC1, which only comprises intracellular peptide segments C and D, did not bind to B56β. These results indicate that the binding site for B56β is located within the peptide segments A and B of the cytoplasmic part of CALEB/NGC, which are both highly conserved in the amino acid sequence of all species examined to date.
|
B56β binds to CALEB/NGC in the blot-overlay assay
To confirm the interaction between CALEB/NGC and B56β found in the yeast two-hybrid system, we first transferred purified recombinant fusion protein composed of GST and the intracellular part of the longer isoform of human CALEB/NGC (GST-hCALEBb-cyt; Fig. 2
A, lane 1) to a nitrocellulose membrane after separation in SDS-PAGE (Fig. 2)
. Purified GST was used as a control (Fig. 2A
, lane 2). This membrane was then incubated with detergent extract of HEK293 cells expressing recombinant myc-tagged B56β (myc-B56β; Fig. 2A
, lane 5). After being washed, the membrane was incubated with an antibody to the myc epitope. Bound myc-B56β was then detected with an alkaline phosphatase-conjugated secondary antibody and a color reaction. Our findings showed that myc-B56β bound to GST-hCALEBb-cyt fusion protein (Fig. 2A
, lane 3) but not to GST (Fig. 2A
, lane 4). The results obtained with the yeast two-hybrid system suggested that the peptide segments A and/or B of the intracellular part of CALEB/NGC bind to B56β (Fig. 1)
. Further mapping analysis confirmed that B56β interacts with intracellular peptide segment A of CALEB/NGC (Fig. 3
C, D). We used the peptide rCAL1 derived from the amino acid sequence of juxtamembrane cytoplasmic peptide segment A of CALEB/NGC (FFAKKLYLLKTENTKLRKT; see Materials and Methods) to interfere with binding of myc-tagged B56β to immobilized GST-hCALEBb-cyt in the blot-overlay assay. We found that the binding of B56β to GST-hCALEBb-cyt is inhibited in the presence of rCAL1 (Fig. 2B
, lane 3 in right panel, arrowhead) but not in the presence of the control peptide rMas1, the amino acid sequence of which (NTVSIETVV) is derived from the COOH terminus of the rat Mas1 proto-oncogene (Fig. 2B
, lane 1 in right panel). Equal amounts of fusion proteins were analyzed, as seen in the Coomassie-stained gel (Fig. 2B
, left panel).
|
|
CALEB/NGC associates with B56β in HEK293 cells
To further analyze the interaction between CALEB/NGC and B56β, we cotransfected HEK293 cells with expression plasmids encoding tagged versions of CALEB/NGC and B56β or PR55, respectively. We prepared detergent extracts of these transfected cells and performed immunoprecipitations (IPs) directed to HA-tagged B56β. B56β was precipitated (Fig. 3B
, bottom), and, in addition, FLAG-tagged CALEB/NGC was coprecipitated (Fig. 3B
, middle). CALEB/NGC was not coprecipitated with HA-tagged PR55, which is another member of the B family of regulatory subunits of PP2A, different from B56β, although HA-tagged PR55 could be precipitated from transfected HEK293 cells with anti-HA antibodies (Fig. 3B
, middle and bottom).
The results of mapping analysis with the yeast two-hybrid system suggested that the intracellular peptide segments A and/or B of CALEB/NGC mediate binding to B56β. To characterize the interaction site more precisely, we immunoprecipitated HA-tagged B56β from HEK293 cells, which had been cotransfected with B56β and either the CALEB/NGC-derived construct 400 or 388, both of which are deletion constructs of CALEB/NGC (Fig. 3A
). While construct 388 contains the full extracellular part of CALEB/NGC and the transmembrane region but only the intracellular peptide segment A, construct 400 additionally includes intracellular peptide segment B (8)
. Both FLAG-tagged constructs 400 and 388 were detected in these immunoprecipitates (Fig. 3C
). This result suggests that the binding site for B56β is located in the juxtamembrane peptide segment A of CALEB/NGC. To confirm this further, we used the peptide rCAL1 immobilized to Sepharose to perform affinity purification from detergent extracts of HEK293 cells expressing HA-B56β. The rMas1 peptide was used as a control. HA-B56β could be isolated from detergent extracts with the rCAL1 but not the rMas1 peptide (Fig. 3D
). Taken together, B56β binds to the juxtamembrane cytoplasmic peptide segment A of CALEB/NGC.
B56β interacts with CALEB/NGC in vivo and colocalizes with CALEB/NGC in dendrites of primary hippocampal neurons
To find out whether CALEB/NGC and B56β interact in neural tissue, we performed IP from membrane preparations of rat brain. We chose a membrane preparation as starting material because CALEB/NGC but not B56β is a transmembrane protein. If B56β bound to CALEB/NGC in vivo, we would be able to coprecipitate both proteins from a solubilized membrane fraction. With an affinity-purified polyclonal antibody to B56β (IP B56β), we could precipitate B56β from the solubilized membrane fraction (Fig. 3E
, bottom). Polyclonal antibodies to an unrelated epitope (the FLAG epitope) did not precipitate B56β (IP FLAG). CALEB/NGC was coprecipitated with B56β in the immunoprecipitation with anti-B56β but not anti-FLAG antibodies (Fig. 3E
, top). Interestingly, the proteoglycan variant of CALEB/NGC (CALEB/NGC-PG-) could not be coprecipitated with B56β (as was also the case when the coimmunoprecipitations were performed from transfected HEK293 cells; Fig. 3A, B
). The reason for this is not clear. Together, these results show that B56β and CALEB/NGC interact in vivo.
Our initial intent in performing the yeast two-hybrid screen was to identify proteins that bind to CALEB/NGC and are involved in signaling pathways important for dendritic tree elaboration and spine formation. As a first step in investigating whether B56β could play a role in these processes, we compared the endogenous expression of B56β in primary hippocampal neurons with the expression pattern of CALEB/NGC. In addition to the cell body, B56β is expressed in discrete spots in main dendrites (Fig. 3F
, middle panel). In these spots, which are often found at dendritic branch points, B56β was found to strongly colocalize with endogenous CALEB/NGC (Fig. 3F
, arrows). Although B56β is not a transmembrane protein, it has a membrane-associated localization in the main dendrites of hippocampal neurons (Fig. 3F
, bottom), which could be due to an interaction with the transmembrane protein CALEB/NGC. Thus, on the basis of its endogenous expression, B56β could be involved in the regulation of signaling pathways necessary for dendrite differentiation induced by CALEB/NGC.
PP2A core composed of structural A subunit and catalytic C subunit is recruited to CALEB/NGC via B56β
B56β, shown to interact with CALEB/NGC, is a regulatory subunit of PP2A. Does B56β bind to CALEB/NGC independent of the PP2A core composed of structural A and catalytic C subunits, or does it act by recruiting the PP2A core to CALEB/NGC? To answer this question, we performed affinity chromatography with the bacterially expressed GST-hCALEBb-cyt fusion protein, which we also used in the blot-overlay assay (Fig. 2)
, and PBS extracts from adult rat brain. Affinity chromatography with GST protein served as a control. After the affinity matrix was incubated with these brain extracts, it was washed extensively with PBS. Two types of elutions were subsequently conducted: a gentle elution with high-salt buffer, followed by a rigorous elution with low-pH buffer (see Materials and Methods). The elution fractions were separated with SDS-PAGE and then either analyzed by Western blot with anti-B56β antibodies (Fig. 4
A, low-pH elution) or stained with Coomassie blue for identification of eluted proteins by mass spectrometry (Fig. 4B
, high-salt elution). B56β was detected in the rigorous low-pH elution fraction from the GST-hCALEBb-cyt but not GST affinity matrix (Fig. 4A
). Two bands present in the high-salt elution fraction of the GST-hCALEBb-cyt but not GST affinity matrix were identified by mass spectrometry. Of these, one band, in the range of 65 kDa, was identified as the structural A subunit of PP2A, while the other, in the range of 36 kDa, was the catalytic C subunit of PP2A (Fig. 4B
). Thus, not only B56β but also the PP2A core can interact with CALEB/NGC. Binding of the PP2A core to CALEB/NGC is likely to be indirect, by means of B56β, as the A and C subunits but not B56β were eluted under the more gentle conditions of high-salt buffer. In addition, we did not detect any interaction between the structural A and the catalytic C subunits of PP2A with CALEB/NGC in the yeast two-hybrid system (data not shown). In summary, the whole PP2A trimer was recruited to CALEB/NGC in our study.
|
B56β inhibits CALEB/NGC-mediated dendritic branching
B56β was shown to colocalize with CALEB/NGC in dendrites of primary hippocampal neurons in culture (Fig. 3F
). Data from our laboratory have demonstrated that CALEB/NGC-induced dendritic branching is dependent on active Akt (8)
. It has also been published that B56β-containing PP2A dephosphorylates and thereby inactivates Akt (22)
. Thus, we investigated whether B56β can interfere with CALEB/NGC-stimulated dendritic branching. Here, we cotransfected hippocampal neurons in culture at 12 days in vitro (DIV12) with CALEB/NGC (isoform mCALEBb) and B56β-expressing plasmids and determined the number of dendritic end tips (TNDET) 2 days later (DIV12+2). Neurons coexpressing mCALEBb and EGFP, EGFP and B56β, mCALEBb and PR55, EGFP and PR55, or expressing EGFP alone served as controls (Fig. 5
A). Hippocampal neurons coexpressing mCALEBb and B56β had a reduced dendritic tree complexity compared with neurons coexpressing mCALEBb and EGFP (Fig. 5A
). This was due to a reduced number of dendritic end tips (Fig. 5B
). TNDET was significant smaller in neurons coexpressing mCALEBb and B56β compared with neurons coexpressing mCALEBb and EGFP, but it was not significantly different compared with neurons coexpressing EGFP and B56β. As a control of specificity for the negative effect of B56β on CALEB/NGC-induced dendritic branching, we coexpressed PR55 together with EGFP or mCALEBb in hippocampal neurons and assessed TNDET. We found that PR55 did not inhibit CALEB/NGC-induced dendritic branching (means of TNDET were 28.6±1.3 for EGFP; 34.1±1.6 for mCALEBb + EGFP; 26.4±1.5 for EGFP+B56β; 26.0±1.1 for mCALEBb+B56β; 29.2±1.4 for EGFP+PR55; and 36.4±1.8 for mCALEBb±PR55; Fig. 5B
). These results indicate that B56β specifically inhibits CALEB/NGC-mediated dendritic branching.
|
CALEB/NGC-induced phosphorylation of dendritic Akt is inhibited by B56β
We recently showed that dendritic branching stimulated by CALEB/NGC is dependent on active Akt (8)
. However, it was not clear whether CALEB/NGC stimulates phosphorylation (Ser-473) and thereby contributes to activation of Akt. We examined this by analyzing phosphorylated Akt in hippocampal neurons overexpressing EGFP, mCALEBb and EGFP, EGFP and B56β, or mCALEBb and B56β. We stained transfected neurons with anti-phospho-Akt mouse monoclonal antibody and quantified the relative fluorescence intensity of these signals in dendrites (not cell bodies). The specificity of this phospho-Akt antibody in cultured hippocampal neurons is documented (see Supplemental Fig. S1). We found that neurons overexpressing mCALEBb and EGFP had a significant higher amount of phosphorylated Akt localized in dendrites than control neurons overexpressing only EGFP (Fig. 6
A). The total amount of dendritic Akt was not significantly increased (data not shown). We further found that coexpression of B56β with mCALEBb completely inhibited the CALEB/NGC-stimulated phosphorylation of dendritic Akt (normalized means are 100±5.9 for EGFP; 133±0.5 for EGFP+mCALEBb; 102±4.3 for EGFP+B56β; and 104±4.8 for mCALEBb and B56β; Fig. 6B
).
|
Currently, it is unclear whether B56β inhibits CALEB/NGC-induced phosphorylation of Akt by directing PP2A to dephosphorylate Akt or by directly inhibiting CALEB/NGC or both. CALEB/NGC contains several potential phosphorylation sites (Ser and Thr) within the intracellular peptide segments A and B that might function as acceptor sites for phosphate. Peptide segments A and B have been shown to be important for CALEB/NGC-mediated dendritic branching (8)
. We systematically produced constructs with point mutations of these amino acids. In a next step, we analyzed whether these CALEB/NGC-derived constructs with point mutations could stimulate dendritic branching. Although not all of these constructs could be expressed in hippocampal neurons, we found one threonine (Thr 452) within peptide segment A as important for CALEB/NGC-induced dendritic branching. Although expressed and localized to dendrites similar to mCALEBb, the CALEB/NGC-derived construct with the point mutation T452A (simulation of dephosphorylated threonine) did not stimulate dendritic branching as did mCALEBb (see Supplemental Fig. S2). In contrast, a construct with the mutation T452E stimulated dendritic branching similar to mCALEBb (see Supplemental Fig. S2). These results suggest that dephosphorylation of threonine 452 may inhibit CALEB/NGC-induced dendritic branching. Taken together, our results show that CALEB/NGC can stimulate phosphorylation of Akt in dendrites of hippocampal neurons and that B56β can inhibit the phosphorylation of Akt that is induced by CALEB/NGC.
CALEB/NGC-induced spine formation is independent of active Akt
We recently showed that dendritic branching stimulated by CALEB/NGC is dependent on an active PI3K-Akt-mTOR signaling pathway. In contrast, CALEB/NGC-mediated spine formation is not dependent on active PI3K (8)
. Here, we investigated whether active Akt is necessary for spine formation induced by CALEB/NGC. We coexpressed EGFP and mCALEBb with or without Akt inhibitor III in hippocampal neurons in culture at DIV12. EGFP-expressing neurons with or without Akt inhibitor III served as controls. We chose Akt inhibitor III because it has been shown to be the most potent inhibitor of CALEB/NGC-stimulated dendritic branching (8)
, while being at the same time more biocompatible than the Akt inhibitors I and II. The neurons were fixed at DIV12 + 4 and then analyzed. As parameters for spine formation, we measured the length (in µm), the number, and the width (in µm) of dendritic spines. In addition, we measured the number of spine branch points. We found that Akt inhibitor III neither inhibited CALEB/NGC-mediated increase in the length (means are 1.1±0.02 for EGFP; 1.3±0.02 for EGFP+mCALEBb; 1.0±0.02 for EGFP+25 µM Akt inhibitor III; and 1.2±0.02 for EGFP+mCALEBb+25 µM Akt inhibitor III; Fig. 7
A, B; also see Supplemental Fig. S3) nor in the number of dendritic spines (means per 10 µm dendritic length are 4.5±0.15 for EGFP; 5.6±0.18 for EGFP+mCALEBb; 5.1±0.15 for EGFP+25 µM Akt inhibitor III; and 5.8±0.15 for EGFP+mCALEBb+25 µM Akt inhibitor III; Fig. 7A, B
; also see Supplemental Fig. S3). It also did not interfere with CALEB/NGC-stimulated spine branching (means of branch points per 10 µm dendritic length are 1.1±0.1 for EGFP; 2.4±0.1 for EGFP+mCALEBb; 1.4±0.1 for EGFP+25 µM Akt inhibitor III; and 2.1±0.15 for EGFP+mCALEBb+25 µM Akt inhibitor III; Fig. 7A, B
; also see Supplemental Fig. S3). The mean width of spine heads was not significantly affected by CALEB/NGC. Inhibition of Akt by Akt inhibitor III led to a general slight increase in the spine width in mCALEBb-overexpressing neurons as well as in control neurons overexpressing EGFP (means are 0.51±0.005 for EGFP, 0.54±0.006 for EGFP + mCALEBb, 0.56±0.006 for EGFP+25 µM Akt inhibitor III, and 0.60±0.007 for EGFP + mCALEBb+25 µM Akt inhibitor III; Fig. 7A, B
, and see Supplemental Fig. S3). Thus, the CALEB/NGC-mediated impact on spine formation is independent of active Akt.
|
B56β does not inhibit CALEB/NGC-mediated spine formation
Due to the fact that CALEB/NGC-stimulated spine formation is not dependent on active Akt, it was conceivable that B56β, which inhibits CALEB/NGC-mediated dendritic branching, does not interfere with spine formation driven by CALEB/NGC. To prove this assumption, we measured the number, length, width, and number of branch points of spines of DIV12 + 4 hippocampal neurons in culture. These neurons coexpressed either mCALEBb and B56β, mCALEBb and EGFP, EGFP and B56β, or expressed EGFP alone. To take into account that mCALEBb is a transmembrane and EGFP a cytosolic protein, we measured the spine length and width of neurons cotransfected with EGFP and mCALEBb on the red channel (mCALEBb staining) and on the green channel (GFP staining). There is a difference between these means of 
0.12 µm. This difference has to be considered when comparing spine length and width of neurons coexpressing EGFP and B56β with neurons coexpressing mCALEBb and B56β. Our results showed that neurons coexpressing mCALEBb and B56β had significantly more spines, longer spines, and more branched spines than neurons coexpressing EGFP and B56β, or expressing EGFP alone [means for spine length are 1.17±0.02 for EGFP; 1.35±0.02 for EGFP+mCALEBb (GFP staining); 1.48±0.02 for EGFP+mCALEBb (mCALEBb staining); 1.13±0.02 for EGFP+B56β; and 1.53±0.02 for mCALEBb+B56β; means for spine number (per 10 µm dendritic length) are 4.2±0.13 for EGFP; 5.2±0.17 for EGFP+mCALEBb; 4.0±0.15 for EGFP+B56β; and 4.9±0.24 for mCALEBb + B56β; means for spine branch points (per 10 µm dendritic length) are 1.0 ± 0.1 for EGFP; 2.2±0.1 for EGFP+mCALEBb; 1.3±0.1 for EGFP+B56β; and 2.4±0.1 for mCALEBb+B56β; Fig. 8
A, B; also see Supplemental Fig. S4]. The mean width of spine heads was not significantly affected by CALEB/NGC when compared with the EGFP control, neither without nor with coexpressed B56β [means are 0.58±0.005 for EGFP, 0.56±0.005 for EGFP + mCALEBb (GFP staining), 0.68±0.005 for EGFP + mCALEBb (mCALEBb staining), 0.50 ± 0.005 for EGFP + B56β, and 0.62 ± 0.005 for mCALEBb + B56β; Fig. 8A, B
, and see Supplemental Fig. S4]. In summary, although B56β completely inhibits CALEB/NGC-stimulated dendritic branching, it has no effect on CALEB/NGC-mediated spine formation.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 24)
. Second, electrophysiological analysis of CALEB/NGC-deficient mice showed disturbances in maintaining normal release probability at early developmental stages (25)
. The latter finding suggests that CALEB/NGC may be implicated in the development of the presynapse. Third, CALEB/NGC was demonstrated to increase dendritic tree complexity and promote spine formation in hippocampal neurons in culture and in the intact mouse cortex (8)
. The signaling pathways within neurons, which are important for these different biological functions of CALEB/NGC, are only partially understood. For instance, an active PI3K-Akt-mTOR signaling pathway was found to be essential for dendritic branching increased by CALEB/NGC (8
, 26
27
28)
. However, active PI3K was shown to be dispensable for spine formation induced by CALEB/NGC (8)
. This is noteworthy, because both stimulation of dendritic branching and spine formation require the same extracellular domain of CALEB/NGC. Recently, we demonstrated that the extracellular EGF-like domain of CALEB/NGC is necessary and sufficient to drive dendritic branching as well as spine formation (8)
. Thus, one could speculate that the EGF-like domain is able to transmit different signals into the cell dependent on the developmental context, but how are these signals transduced within a neuron to ensure the biological outcome and which proteins are involved in the signaling pathways driven by the EGF-like domain of CALEB/NGC? By means of the yeast two-hybrid system, we identified a novel interaction partner of CALEB/NGC: B56β, one of the regulatory B subunits of PP2A. The interaction between CALEB/NGC and B56β was confirmed by several biochemical and immunocytochemical techniques in vitro and in vivo. Of particular interest are our findings that CALEB/NGC can recruit the whole PP2A trimer, that it can be coprecipitated with B56β from a membrane preparation of rat brain tissue, and that B56β and CALEB/NGC colocalize in dendrites but not in spines of hippocampal neurons in culture. B56β not only colocalizes with CALEB/NGC in dendrites but interferes with the function of CALEB/NGC in dendrites by inhibiting CALEB/NGC-induced dendritic branching. Why is endogenously expressed B56β not able to inhibit CALEB/NGC-stimulated dendritic branching when CALEB/NGC is overexpressed in neurons? The most likely interpretation would be that the amount of endogenously expressed B56β is not sufficient to compensate for the increased CALEB/NGC levels. In accordance with this interpretation are our findings that endogenously expressed B56β is not sufficient to inhibit CALEB/NGC-induced Akt phosphorylation (see below). Overexpression of B56β in EGFP-transfected neurons leads to a small but permanent decrease in dendritic branching. This may be due to an enhanced inhibition of endogenously expressed CALEB/NGC or to an interference with dendritic tree complexity through mechanisms independent of CALEB/NGC.
B56β was previously described as a negative regulator of Wnt/β-catenin signaling. This fact could have an effect on the regulation of dendritic branching, since β-catenin was highlighted to be critical for dendritic morphogenesis (29)
. At first glance, our finding that B56β inhibits CALEB/NGC-mediated dendritic branching fits well with a signaling pathway that includes β-catenin, as B56β interferes with β-catenin signaling. However, it was shown that β-catenin is required for activity-dependent dendritic branching. In contrast, the CALEB/NGC-mediated increase in dendritic tree complexity is not dependent on electrical activity (8)
. In addition to its role in regulation of Wnt/β-catenin signaling, B56β was recently identified as the regulatory subunit of PP2A, which targets this enzyme to Akt. The holoenzyme can then function as an Akt phosphatase (22)
. Our data presented here indicate that CALEB/NGC is able to recruit the whole PP2A enzyme complex via B56β. In this way, B56β could interfere with a signaling pathway including active Akt that is necessary for dendritic branching stimulated by CALEB/NGC. To corroborate this hypothesis, we examined whether CALEB/NGC stimulates the phosphorylation of Akt on serine-473 and thus contributes to the activation of Akt. Our results clearly show that enhanced expression of CALEB/NGC in hippocampal neurons leads to an increased phosphorylation of dendritic Akt, while coexpression of B56β completely inhibits the phosphorylation of Akt induced by CALEB/NGC. The conclusion of these results is that B56β could really interfere with a signaling pathway driven by CALEB/NGC and leading to phophorylated Akt. However, so far the interference of B56β with CALEB/NGC-induced dendritic branching and stimulation of Akt phosphorylation was only shown in the case of overexpression of CALEB/NGC and B56β. Future studies have to be conducted to interfere with the binding of endogenously expressed B56β to CALEB/NGC. Then, the functional outcome in terms of dendritic branching and Akt phosphorylation could be analyzed without overexpressing CALEB/NGC or B56β. Furthermore, it will be important to examine whether and, if so, how the interaction between B56β and CALEB/NGC is regulated. We previously showed that the extracellular EGF-like domain of CALEB/NGC is necessary and sufficient to induce dendritic branching (8)
. This could happen either by triggering an intracellular signaling cascade that is directly stimulatory for dendritic branching or by abrogating an intracellular inhibition of CALEB/NGC-induced dendritic branching. It will be interesting to see whether a binding of an (so far unknown) extracellular ligand of the EGF-like domain could regulate the intracellular interaction between B56β and CALEB/NGC.
Up to now, it has not been clear whether B56β or B56β-dependent PP2A interfere with Akt phosphorylation by directly inhibiting CALEB/NGC or by dephosphorylating Akt, or both. However, we show here that the point mutation T452A but not T452E of the amino acid threonine-452 of CALEB/NGC abolishes the potency of this protein to induce dendritic branching. The amino acid threonine-452 is located within the intracellular peptide segment A of CALEB/NGC. B56β binds to this peptide segment and is able to recruit the whole PP2A trimer to CALEB/NGC. Although we have no definitive proof for this hypothesis so far, it is tempting to speculate that B56β-dependent PP2A may dephosphorylate the amino acid threonine-452 of CALEB/NGC and thereby contributes to the inhibition of CALEB/NGC-induced dendritic branching. The observation that overexpressed B56β does not significantly reduce dendritic Akt phosphorylation in the absence of overexpressed CALEB/NGC compared with the control is in line with the interpretation that B56β likely inhibits dendritic branching by a direct effect on CALEB/NGC.
On the other hand, we demonstrate here that CALEB/NGC stimulates spine formation by increasing spine number, spine length, and spine branching independent of active Akt. Neurons expressing EGFP or coexpressing EGFP and CALEB/NGC in the presence of Akt inhibitor III develop more spines, which additionally are shorter, have more branch points and slightly larger spine heads than neurons expressing the same constructs in the absence of this Akt inhibitor. This may be due to the fact that other, Akt-dependent signaling pathways could also contribute to spine formation. However, expression of CALEB/NGC results in a significant increase in spine number, spine length, and spine branching compared with EGFP expression in the absence or presence of Akt inhibitor III. Overexpression of CALEB/NGC leads to an increase of Akt phosphorylation not only in dendrites but also in spines. It will be interesting in the future to examine whether CALEB/NGC is not only involved in shaping the morphology of spines in an Akt-independent manner but also in modulating the functional status of spines and synapses dependent on phosphorylated Akt.
In accordance with these results, we found that B56β did not interfere with spine formation stimulated by CALEB/NGC. Furthermore, B56β colocalizes with CALEB/NGC in dendrites but there was no evidence for a colocalization in spines. Even during overexpression, B56β did not localize to dendritic spines in a significant manner.
An important next step in the elucidation of CALEB/NGC-dependent signaling to mediate dendritic tree complexity and spine formation will be to figure out whether the interaction of B56β with CALEB/NGC and the subsequent inhibition of Akt phosphorylation are regulated by the EGF-like domain of CALEB/NGC. To initiate these experiments, it will be important to identify the ligand of this EGF-like domain, which is currently not known.
In summary, we have shown that B56β is a novel interaction partner of CALEB/NGC, which is involved in regulating CALEB/NGC-mediated dendritic branching but not spine formation.
|
ACKNOWLEDGMENTS |
|---|
Received for publication August 11, 2007. Accepted for publication February 21, 2008.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
