In this protocol, we introduce a method for purifying the dendritic filopodia-rich fraction from the phagocytic cup-like protrusion structure on cultured hippocampal neurons by taking advantage of the specific and strong affinity between a dendritic filopodial adhesion molecule, TLCN, and an extracellular matrix molecule, vitronectin.
Dendritic filopodia are thin and long protrusions based on the actin filament, and they extend and retract as if searching for a target axon. When the dendritic filopodia establish contact with a target axon, they begin maturing into spines, leading to the formation of a synapse. Telencephalin (TLCN) is abundantly localized in dendritic filopodia and is gradually excluded from spines. Overexpression of TLCN in cultured hippocampal neurons induces dendritic filopodia formation. We showed that telencephalin strongly binds to an extracellular matrix molecule, vitronectin. Vitronectin-coated microbeads induced phagocytic cup formation on neuronal dendrites. In the phagocytic cup, TLCN, TLCN-binding proteins such as phosphorylated Ezrin/Radixin/Moesin (phospho-ERM), and F-actin are accumulated, which suggests that components of the phagocytic cup are similar to those of dendritic filopodia. Thus, we developed a method for purifying the phagocytic cup instead of dendritic filopodia. Magnetic polystyrene beads were coated with vitronectin, which is abundantly present in the culture medium of hippocampal neurons and which induces phagocytic cup formation on neuronal dendrites. After 24 h of incubation, the phagocytic cups were mildly solubilized with detergent and collected using a magnet separator. After washing the beads, the binding proteins were eluted and analyzed by silver staining and Western blotting. In the binding fraction, TLCN and actin were abundantly present. In addition, many proteins identified from the fraction were localized to the dendritic filopodia; thus, we named the binding fraction as the dendritic filopodia-rich fraction. This article describes details regarding the purification method for the dendritic filopodia-rich fraction.
Dendritic filopodia are thought to be precursors of spines. Actin filaments in the dendritic filopodia regulate their extension and retraction1,2,3. After contacting with an axon, selected dendritic filopodia begin their maturation into spines, and a synapse is formed4,5. Components of spines have been determined from comprehensive analysis of postsynaptic density fractions6,7, while components of dendritic filopodia remain largely unknown. It has been shown that telencephalin (TLCN), ERM, SynGAP, Ras, PI3 kinase, Akt, mTOR, polo-like kinase 2, CaMKII, syndecan-2, paralemin-1, ARF6, and EphB regulate dendritic filopodia formation5,8,9,10,11, while a method has not been developed for the comprehensive analysis of molecules present in the dendritic filopodia.
TLCN (ICAM-5) is specifically expressed by spiny neurons in the most rostral brain segment, the telencephalon12. TLCN has 9 Ig-like domains in its extracellular region, a transmembrane region, and a cytoplasmic tail13. TLCN binds to vitronectin (VN) and LFA-1 integrin in its extracellular region, to presenilin in its transmembrane region, and to phospho-ERM and α-actinin in its cytoplasmic region5,8,14,15,16. TLCN binds to the actin cytoskeleton through phospho-ERM at the tips of dendritic filopodia and α-actinin in spines and dendritic shafts8,16.
We showed that overexpression of TLCN enhanced dendritic filopodia formation and induced the reversion of spines to filopodia10. The constitutive active form of ezrin bound to the TLCN cytoplasmic region and enhanced dendritic filopodia formation8. Thus, TLCN regulates dendritic filopodia formation through actin-binding proteins. Esselens et al. demonstrated that microbeads induced TLCN accumulation on cultured neurons17. We showed that phagocytic cup structures were formed on neuronal dendrites around VN-coated microbeads in a TLCN-dependent manner15. Constituents of dendritic filopodia are similar to those of the phagocytic cup. It is difficult to collect dendritic filopodia, but it is relatively easier to collect the phagocytic cup using magnetic microbeads. Thus, we developed a method to purify the phagocytic cup instead of dendritic filopodia18. Here, we describe the purification method for the dendritic filopodia-rich fraction.
All methods described here have been approved by the Institutional Animal Care and Use Committee of RIKEN Wako.
1. Culture of Hippocampal Neurons
2. Purification of Dendritic Filopodia-rich Fraction
3. Silver Staining and Western Blot Analysis
In cultured hippocampal neurons, TLCN was abundantly localized to the dendritic filopodia, shaft, and soma and colocalized with F-actin (Figure 1A, B). When polystyrene microbeads were added to cultured hippocampal neurons, the beads were automatically coated with vitronectin (VN) derived from fetal bovine serum (FBS) in the culture medium; they were mainly bound to dendrites, and they induced the formation of phagocytic cups (Figure 1B-E). Phagocytic cups were based on sheet-shaped actin filaments around microbeads on dendrites. TLCN, phospho-ERM, and PI(4,5)P2, which are markers for dendritic filopodia, are highly accumulated around beads (Figure 1D)8,15. Phagocytic cups were only formed on wild-type hippocampal neurons, but not on TLCN-deficient hippocampal neurons (Figure 2A-D). Thus, the phagocytic cup formation was crucially dependent on the presence of TLCN in dendrites.
Constituents of the dendritic filopodia appeared similar to those of phagocytic cups. Next, we purified phagocytic cups instead of dendritic filopodia. The protocol for the purification of phagocytic cups is schematically shown in Figure 3. Magnetic microbeads were added to wild-type cultured hippocampal neurons, which induces the formation of phagocytic cup structures. The phagocytic cup structures were lysed with a weak detergent. The microbeads were collected after the lysis using a magnet separator. After washing of the beads, their binding proteins were boiled and eluted in an SDS sample buffer.
The amount of proteins in the bound and unbound fraction was measured using the BCA protein assay kit. The same amount of proteins in the unbound and bound fractions was separated by SDS-PAGE and stained by silver staining (Figure 4A). The protein band patterns were almost the same for the unbound and bound fractions, but the intensities at 50 and 70 kDa in the bound fraction were lower than those in the unbound fraction. However, the band intensity was not obviously different between the unbound and bound fractions prepared from TLCN-deficient culture hippocampal neurons. To confirm the purification of phagocytic cup structures, we performed Western blotting using anti-TLCN-C, anti-bovine vitronectin, anti-actin, and anti-α-tubulin (Figure 4B). TLCN and VN were mainly detected in the bound fraction. Actin, ezrin, Gαq, PLCβ1, MAP-2, and spectrin were detected in both the bound and unbound fractions. Moesin, PSD-95, α-actinin, and α-tubulin were detected in the unbound fraction.
Figure 1: Localization of TLCN and F-actin in dendritic filopodia and phagocytic cups. (A) Immunofluorescence staining of dendrites of a cultured hippocampal neuron expressing gapVenus with anti-GFP antibody (blue in a merged image), anti-TLCN antibody (red in a merged image), and phalloidin (green in a merged image). TLCN and F-actin are abundantly observed in dendritic filopodia. (B, C) Formation of phagocytic cup structures on neuronal dendrites. Fluorescent microbeads (blue in merged images of B, C) added to cultured hippocampal neurons strongly adhere onto dendrites and induce the accumulation of TLCN (red in merged images of B, C) and F-actin (green in merged images of B, C). (D) A lateral view of a dendritic phagocytic cup reconstructed from confocal images reveals TLCN (red in merged images of D) surrounding the fluorescent bead (blue in merged images of D). (E) A dendritic phagocytic cup induced by a magnetic microbead attached onto a neuronal dendrite and immunostained with anti-VN antibody (blue in merged images of E), anti-TLCN antibody (red in merged images of E), and phalloidin (green in merged images of E). Scale bars = 1 μm in (D); 2 μm in (A), (C), and (E); and 20 μm in (B). This figure has been modified from a previous study18. Please click here to view a larger version of this figure.
Figure 2: TLCN-dependent formation of phagocytic cup-like structure. (A–D) Triple fluorescence images of wild-type (WT; A, C) and TLCN-deficient (KO; C, D) neurons treated with VN-coated fluorescent beads (red in merged images of A, B, C, D) and labeled with anti-TLCN antibody (green in merged images of A, B, C, D) and anti-phospho-ERM antibody (blue in merged images of A, B) or Alexa488-phalloidin (blue in merged images of C, D). Scale bars = 5 μm. This figure has been modified from a previous study15. Please click here to view a larger version of this figure.
Figure 3: A schematic diagram illustrating the purification procedure of the dendritic filopodia-rich fraction. Magnetic microbeads were added onto cultured hippocampal neurons to induce the formation of dendritic phagocytic cups. After 1 day of incubation, the neurons were solubilized with lysis buffer containing 0.01% Triton X-100. The beads were separated from the unbound fraction using a magnetic separator. After washing, the bound proteins were eluted with SDS sample buffer and used as the bound fraction. Red: VN, green: TLCN, other colors: bound proteins. This figure has been modified from a previous study18. Please click here to view a larger version of this figure.
Figure 4: Confirmation of phagocytic cup fraction. (A) Silver staining of proteins in the unbound and bound fractions of the microbeads. The same amount (50 ng) of proteins in the unbound and bound fractions purified from wild-type (WT) and TLCN-deficient (TLCN KO) hippocampal neurons were separated by SDS-PAGE and visualized with silver staining. (B) Western blot analysis of the unbound and bound fractions. The same amount (50 ng) of proteins were separated by SDS-PAGE and subjected to Western blot analysis using anti-TLCN, anti-VN, anti-actin, anti-ezrin, anti-moesin, anti-Gαq, anti-PLCβ1, anti-MAP-2, anti-spectrin, anti-PSD-95, anti-α-actinin, and anti-α-tubulin antibodies. Note that TLCN, VN, actin, ezrin, PLCβ1, MAP-2, and spectrin are observed in the dendritic filopodia-rich fraction. This figure has been modified from a previous study18. Please click here to view a larger version of this figure.
We developed a purification method for the dendritic filopodia-rich fraction using affinity between the cell adhesion molecule TLCN and the extracellular matrix protein vitronectin. Compared to PSD fraction, it could be possible to identify the synaptic proteins acting on the immature synapse from the dendritic filopodia-rich fraction. Thus, the constituents of the dendritic filopodia-rich fraction are different from those of the PSD fraction by 74%. Different from PSD fraction, we used cultured hippocampal neurons to actively form phagocytic cups, and the cells need to be alive. To form the phagocytic cup, we used the interaction between TLCN and vitronectin. TLCN expression is limited in the telencephalon. Thus, we cannot use cultured neurons derived from the cerebellar. However, if we coat the beads with different proteins, this activity-dependent purification method could be applied to cerebellar neurons. For example, when the N-terminal domain of glutamate receptor delta2-coated microbeads are applied to cerebellar granule cells, presynaptic neurexin and cbln1 were identified from the binding proteins. Therefore, this activity-dependent method could be used, if the coating proteins are changed. Since the access of vitronectin-coated microbeads is limited to the brain slice and brain tissue, phagocytic cups are not formed efficiently. Thus, future tasks include developing the purification method of the dendritic filopodia-rich fraction from brain slice and tissue.
In this protocol, low-density culture condition is used for hippocampal neurons, and the growth of glial cells are inhibited by the addition of Ara-C10,20,21. Hippocampal neurons are cultured in MEM prepared by ourselves, but not in neurobasal medium, which is widely used for the culture of hippocampal neurons. Hippocampal neurons often die by 14 DIV when cultured on neurobasal medium, which indicates that the neurobasal medium is not suitable for our low-density culture condition. In addition, an important aspect of low-density culture is coating the dish with poly-L-lysine hydrobromide, which cannot be replaced with poly-L-lysine.
To obtain sufficient amount of proteins from the dendritic filopodia-rich fraction, the numbers of hippocampal neurons and magnetic microbeads are important factors. For immunostaining dendritic filopodia, hippocampal neurons were plated on a 35 mm dish at 5.6 x× 104 cells/dish, while the neurons were plated at 7.0 x 104 cells/dish for purification of the dendritic filopodia-rich fraction. At 14 DIV, almost no hippocampal neurons overlapped at 5.6 x 104 cells/dish for immunostaining, while many of hippocampal neurons partly overlapped with the other neurons at 7.0 x 104 cells/dish for purification of the dendritic filopodia-rich fraction. However, dendritic filopodia were abundantly present at both densities of hippocampal neurons. High-density cultures of hippocampal neurons often mature faster than low-density cultured hippocampal neurons; thus, adjusting the density of hippocampal neurons to low-density culture is indispensable to obtain the dendritic filopodia-rich fraction. Microbeads were added at 1 x 106 microbeads/dish for immunostaining and at 3 x 106 microbeads/dish for purification of the dendritic filopodia-rich fraction. For purification of the fraction, once a sufficient amount of beads was added, the unbound beads were washed out from hippocampal neurons.
In this protocol, microbeads were automatically coated with VN, which is present in the culture medium. However, microbeads can be coated with recombinant VN and added to hippocampal neurons. Because VN is a very sticky protein, VN-coated microbeads easily form aggregates. Thus, VN-coated microbeads are sonicated and pipetted to dissociate from each other just before addition to hippocampal neurons.
We previously showed that VN-coated microbeads induces phosphor-ERM, PI(4,5)P2, and F-actin together with TLCN accumulation15. When the dendritic filopodia-rich fraction was analyzed by Western blotting, phospho-ERM was not detected by anti-phospho-ERM polyclonal antibody and was slightly detected by anti-ezrin and anti-moesin monoclonal antibodies. It appears that the sensitivity of the monoclonal antibodies was higher than the anti-phospho-ERM polyclonal antibody. In addition, ERM proteins were localized to almost all regions of the hippocampal neurons, but phospho-ERM proteins were only localized to the tip of the dendritic filopodia and the phagocytic cup. It is considered that the amount of phospho-ERM binding to TLCN is limited compared to the amount of nonphosphorylated ERM. Thus, detection of phospho-ERM in the dendritic filopodia-rich fraction appears to be difficult.
To detach the microbeads from the cell membrane, we used a weak detergent, 0.01% Triton X-100. TLCN is linked to actin filament through phospho-ERM in the dendritic filopodia and phagocytic cup structures. To purify proteins indirectly linked to TLCN through the actin filament, we used a weak detergent in this protocol. However, the concentration of Triton X-100 could be changed to a higher concentration depending on the purpose of the experiment.
Esselens et al. has shown that the phagocytic uptake of microbeads induced TLCN, PIP2, and F-actin accumulation in cultured hippocampal neurons17. According to our analysis via confocal microscopy after 24 h of incubation with microbeads, the microbeads were not completely taken up into the cytoplasm. TLCN and PIP2, which are localized to the cell membrane, were localized around beads, especially on the bottom of the microbeads. In addition, 319 proteins identified from the dendritic filopodia-rich fraction were analyzed using the KEGG pathway analysis. Phagocytosis and autophagy pathways were not detected, but cytoskeleton organization, exocytosis, actin-filament-based process, and microtubule-based process were significantly enriched in the fraction18.
The molecular mechanism of dendritic filopodia formation remains largely unknown. Analysis of the phagocytic cup structure could help understand the molecular constituents and dynamic functions of dendritic filopodia. It would be interesting to analyze the dendritic filopodia-rich fraction prepared from mouse models of neurodevelopmental and neuropsychiatric disorders.
The authors have nothing to disclose.
We thank Shigeo Okabe and Hitomi Matsuno for the low-density culture of hippocampal neurons, Masayoshi Mishina for TLCN-deficient mice, Sachiko Mitsui and Momoko Shiozaki for technical assistance, and members of the Yoshihara laboratory for helpful discussions. This work was supported by JSPS KAKENHI Grant Nos. JP20700307, JP22700354, and JP24500392 and MEXT KAKENHI Grant Nos. JP23123525 to YF and JP20022046, JP18H04683, and JP18H05146 to YY.
1 M HEPES | Gibco | 15630-080 | |
1.7 ml Low Binding MCT | Sorenson BioScience | 39640T | |
200 mM L-Glutamine | Gibco | 2530149 | |
35-mm plastic cell culture dishes | Corning | 430165 | |
Anti-actin | Sigma-Aldrich | A-5060 | |
Anti-alpha-Actinin | Sigma-Aldrich | A-5044 | |
Anti-alpha-tubulin | Sigma-Aldrich | T-9026 | |
Anti-Ezrin | Sigma-Aldrich | clone3C12, SAB4200806 | |
Anti-Galphaq | Santacruz | sc-393 | |
Anti-MAP2 | Chemicon | clone AP20, MAB3418 | |
Anti-Moesin | Sigma-Aldrich | clone 38/87, M7060 | |
Anti-PLCbeta1 | Santacuz | sc-5291 | |
Anti-PSD95 | MA2 | ABR | |
Anti-Spectrin beta | Chemicon | MAB1622 | |
B27 | Gibco | 0080085SA | |
BCA protein assay kit | Thermo | 23227 | |
Bromophenol blue | Merck | 1.08122.0005 | |
calcium chrolide, hydrous | Wako | 038-19735 | |
Cell scraper | Falcon | 353085 | |
Cell strainer | Falcon | 352350 | |
Choline chloride | Sigma-Aldrich | C7527 | |
Complete EDTA free protease inhibitor cocktail | Roche | 11873580001 | |
Cytosine beta-D-arabinofuranoside | Sigma-Aldrich | C-6645 | |
DNase-I | Sigma-Aldrich | DN-25 | |
D-Pantothenic acid hemicalcium salt | Sigma-Aldrich | P5155 | |
DynaMag-2 Magnet | Thermo | 12321D | |
ECL Prime Western Blotting Detection Reagent | GE | RPN2232 | |
e-PAGEL 5-20% SDS-PAGE gradient gel | ATTO | E-T520L | |
Folic acid | Sigma-Aldrich | F8758 | |
HBSS | Gibco | 14175095 | |
HRP-conjugated anti-rabbit IgG | Jackson ImmunoResearch | 111-035-144 | |
i-Inositol | Sigma-Aldrich | I7508 | |
LAS-1000 mini | Fuji Film | LAS-1000 mini | For detection of luminescence from WB membrane |
Magnetic polystyrene microbeads | Sperotech | PM-20-10 | |
MEM amino acid solution | Gibco | 11130-051 | 30 mM L-Arginine hydrochloride, 5 mM L-Cystine, 10 mM L-Histidine hydrochloride-H2O, 20 mM L-Isoleucine, 20 mM L-Leucine, 19.8 mM L-Lysine hydrochloride, 5.1 mM L-Methionine, 10 mM L-Phenylalanine, 20 mM L-Threonine, 2.5 mM L-Tryptophan, 10 mM L-Tyrosine, and 20 mM L-Valine |
Mini-slab size electrophoresis system | ATTO | AE-6530 | |
Niacinamide | Sigma-Aldrich | N0636 | |
Penicilin / Streptomycin | Gibco | 15070063 | |
PhosSTOP phosphatase inhibitor cocktail | Roche | 4906845001 | |
Poly-L-lysine hydrobromide | Nacali | 28360-14 | |
Pyridoxal HCl | Sigma-Aldrich | P6155 | |
Riboflavin | Sigma-Aldrich | R9504 | |
Silver Stain 2 Kit wako | Wako | 291-5031 | |
Thiamine HCl | Sigma-Aldrich | T1270 | |
Trans-Blot SD Semi-Dry Transfer Cell | Bio-rad | 1703940JA | |
Ultra pure water | MilliQ | For production of ultra pure water |