Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Apr 3;216(4):1143-1161.
doi: 10.1083/jcb.201606086. Epub 2017 Mar 6.

ELKS1 Localizes the Synaptic Vesicle Priming Protein bMunc13-2 to a Specific Subset of Active Zones

Affiliations
Free PMC article

ELKS1 Localizes the Synaptic Vesicle Priming Protein bMunc13-2 to a Specific Subset of Active Zones

Hiroshi Kawabe et al. J Cell Biol. .
Free PMC article

Erratum in

Abstract

Presynaptic active zones (AZs) are unique subcellular structures at neuronal synapses, which contain a network of specific proteins that control synaptic vesicle (SV) tethering, priming, and fusion. Munc13s are core AZ proteins with an essential function in SV priming. In hippocampal neurons, two different Munc13s-Munc13-1 and bMunc13-2-mediate opposite forms of presynaptic short-term plasticity and thus differentially affect neuronal network characteristics. We found that most presynapses of cortical and hippocampal neurons contain only Munc13-1, whereas ∼10% contain both Munc13-1 and bMunc13-2. Whereas the presynaptic recruitment and activation of Munc13-1 depends on Rab3-interacting proteins (RIMs), we demonstrate here that bMunc13-2 is recruited to synapses by the AZ protein ELKS1, but not ELKS2, and that this recruitment determines basal SV priming and short-term plasticity. Thus, synapse-specific interactions of different Munc13 isoforms with ELKS1 or RIMs are key determinants of the molecular and functional heterogeneity of presynaptic AZs.

Figures

Figure 1.
Figure 1.
Differential expression and subcellular localization of Munc13-2 variants in mouse hippocampal neurons. (A) bMunc13-2 and ubMunc13-2 expression levels in mouse hippocampal neurons. (Left) EYFP-fused Munc13-2 isoforms were extracted from cultured hippocampal neurons of Munc13-2EYFP KI mice. (Right) EYFP-fused Munc13-2 isoforms were immunoprecipitated from detergent extracts of hippocampal synaptosomes from Munc13-2EYFP KI mice. Extracted proteins or immunoprecipitates (IP) were separated by SDS-PAGE along with recombinant bMunc13-2–EGFP and ubMunc13-2–EGFP (right two lanes) and analyzed by Western blotting (WB) with anti-GFP (top), anti–bMunc13-2 (middle), and anti–ubMunc13-2 (bottom) antibodies. Two GFP-positive bands appeared in samples from Munc13-2EYFP KI hippocampal neurons (left) and from hippocampal tissue (right), whose molecular weights corresponded to those of recombinant bMunc13-2–EGFP (arrowheads) and ubMunc13-2–EGFP (arrows), respectively. The band with larger molecular weight cross-reacted with the anti–bMunc13-2 antibody, whereas the one with lower molecular weight cross-reacted with the anti–ubMunc13-2 antibody. By densitometric quantification of GFP signals on films using ImageJ, ratios of over 4.5:1 and 6:1 were determined between expression levels of bMunc13-2–EYFP and ubMunc13-2–EYFP in Munc13-2EYFP KI neurons and in hippocampal tissues, respectively. (B and C) Subcellular localization of Munc13-2 in mouse hippocampal neurons. Cultured autaptic hippocampal neurons from Munc13-2EYFP KI mice were immunostained for Munc13-1, EYFP, and Bassoon. (B) The overview image of an autaptic neuron shows that AZs containing Munc13-1 only (red) are much more abundant than AZs containing Munc13-2–EYFP (green). Bar, 50 µm. (C) A high-magnification image of the boxed image section in B shows some AZs containing Munc13-1 and Munc13-2–EYFP (arrowheads) among numerous AZs containing only Munc13-1 (arrows). Only 23% of puncta double positive for Munc13-1 and Bassoon were also enriched for Munc13-2–EYFP (n = 1,942 synapses, n = 12 images). Bars, 10 µm.
Figure 2.
Figure 2.
bMunc13-2 and ELKS1 form a protein complex. (A) Domain structures of bMunc13-2 and ELKS1. The sequences covered by the bait vector pGBDC2-bMunc13-2(1–605) and the bMunc13-2–binding ELKS1 prey clone pGAD-ELKS1(771–976) obtained in the YTH screen are indicated. Coiled-coil motifs were predicted by COILS/PCOILS. CC, coiled coil motif; C1, C1 domain; C2, C2 domain. (B) The ELKS1 binding site in bMunc13-2 maps to the coiled-coil motif in the N terminus. Shown are the domain structures of bMunc13-2 and Munc13-3 and the representations of bait vectors used in YTH binding assays with the ELKS1 prey vector pGAD-ELKS1(771–976). (C) Isoform-specific binding of MBP-bMunc13-2(1–305) to GST-ELKS1 (771–976), but not to GST-ELKS2(751–969), in cosedimentation assays. Proteins that bound to immobilized GST-ELKS1(771–976) were subjected to SDS-PAGE and stained with Coomassie Brilliant blue. Note that MBP-bMunc13-2(1–305) bound to GST-ELKS1(771–976) robustly, whereas binding to GST-ELKS2(751–969) was not detectable. Data are representative of three independent experiments. (D) Stoichiometric interaction of MBP-bMunc13-2(1–305) and GST-ELKS1(771–976). A 3.5× higher amount of MBP-bMunc13-2(1–305), 0.14 mg/ml, was loaded to the GST-ELKS1(771–976) affinity matrix than in the assay shown in C. Absorbance units of Coomassie Brilliant blue–stained MBP–bMunc13-2(1–305) versus GST-ELKS1(771–976) in this experiment were 12,056:11,428. Given that the molecular weights of the two proteins are ∼80 kD and ∼45 kD, MBP-bMunc13-2(1–305) and GST-ELKS1(771–976) are interacting with an apparent 1:1.69 stoichiometry. The image of loaded MBP-bMunc13-2(1–305) (left) is from the same SDS-PAGE gel as the one for stoichiometric interaction (right). Data are representative of two independent experiments.
Figure 3.
Figure 3.
bMunc13-2 is preferentially localized to ELKS1-enriched synapses. (A and B) Localization of endogenous bMunc13-2, ELKS1, and Bassoon (A) and of bMunc13-2, ELKS2, and Bassoon (B) in the WT mouse cortex. Note that bMunc13-2 shows a similar distribution pattern to that of ELKS1 with a restriction to a limited number of synapses (arrowheads in A and B), whereas ELKS2 is accumulated in most Bassoon-positive AZs (B). Bars, 5 µm. For images represented as heat maps, quantitative score scales are presented at the bottom left of the respective image. (C) Quantification of the percentage ELKS1- and ELKS2-positive synapses containing bMunc13-2. Data are means ± SEM and were analyzed by Student’s t test (n = 15 images; ***, P < 0.001).
Figure 4.
Figure 4.
ELKS1 binding of bMunc13-2 regulates synapse function. (A) Mean evoked EPCS amplitudes in Munc13-1/2 DKO neurons expressing bMunc13-2–EYFP or bMunc13-2(Δ145–187)–EYFP. For each cell, EPSC amplitudes of 13–20 pulses at 0.2 Hz were averaged. Cells efficiently rescued with bMunc13-2(Δ145–187)–EYFP (n = 84) showed significantly smaller EPSC amplitudes than cells rescued with bMunc13-2–EYFP (n = 87). (B) Mean RRP sizes in Munc13-1/2 DKO neurons expressing bMunc13-2–EYFP or bMunc13-2(Δ145–187)–EYFP. Cells rescued with bMunc13-2(Δ145–187)–EYFP (n = 37) showed a significantly smaller RRP size than cells rescued with bMunc13-2–EYFP (n = 59). (C) Mean Pvr from Munc13-1/2 DKO neurons expressing bMunc13-2–EYFP or bMunc13-2(Δ145–187)–EYFP. Cells rescued with bMunc13-2(Δ145–187)–EYFP (n = 35) showed a Pvr similar to cells rescued with bMunc13-2–EYFP (n = 59); Student’s t test, P > 0.05. (D) Short-term plasticity in Munc13-1/2 DKO neurons expressing bMunc13-2–EYFP (n = 87) or bMunc13-2(Δ145–187)–EYFP (n = 83). Neurons were initially stimulated at 0.2-Hz stimulation frequency, and then a 10-Hz train was applied for 5 s. Frequency facilitation of the EPSC was measured during the train, and augmentation was measured 1.4 s after the train. Data were normalized to the mean EPSC amplitude of the first nine data points at 0.2-Hz stimulation frequency. Munc13-1/2 DKO cells rescued with bMunc13-2(Δ145–187)–EYFP showed a significantly smaller facilitation and augmentation than cells rescued with bMunc13-2–EYFP. (E) Changes in evoked EPSC amplitudes, RRP sizes, and Pvr at 1.4 s after a 10-Hz stimulation train. Data are normalized to the values measured before the 10-Hz stimulation. Irrespective of the bMunc13-2 variants expressed, all three parameters are increased in the augmented state. The degree of EPSC amplitude and RRP size augmentation was significantly larger in cells expressing bMunc13-2–EYFP as compared with cells expressing bMunc13-2(Δ145–187)–EYFP (n = 83–87 for EPSC amplitude augmentation and n = 22–28 for RRP size increases), whereas the increase in Pvr in the augmented state was similar in the two groups (n = 11–12). (F) Similar absolute Pvr values at 1.4 s after a 10-Hz stimulation train in cells expressing bMunc13-2–EYFP or bMunc13-2(Δ145–187)–EYFP (n = 11–12); Student’s t test, P > 0.05. Note that the ELKS1 binding–deficient construct bMunc13-2(Δ145–187)–EYFP failed to rescue synaptic transmission in ∼40% of Munc13-1/2 DKO neurons analyzed. These neurons were not included in any of the analyses shown. ***, P < 0.01, Student’s t test. Data are means ± SEM.
Figure 5.
Figure 5.
AZ recruitment of bMunc13-2 depends on the presence of ELKS1. (A) Immunohistochemical analyses of somatosensory cortex of control (ELKS1f/f) and ELKS1 conditional KO mice (ELKS1f/f;Emx1Cre) using anti–bMunc13-2 and anti-Bassoon antibodies. The low-magnification image of the control brain shows discrete bMunc13-2–positive signals, whose number was reduced in the ELKS1 conditional KO sample. Arrowheads in the high-magnification images indicate bMunc13-2–positive structures. Bars: (low magnification) 50 µm; (high magnification) 5 µm. For images represented as heat maps, quantitative score scales are presented at the bottom left of the respective image. (B) Quantification of the percentages of bMunc13-2–positive puncta among Bassoon-positive puncta in control and ELKS1 conditional KO mouse cortex. Areas for analyses were selected based on the Bassoon signal (n = 5 images, Student’s t test, ***, P < 0.001). (C and D) Frequency distributions of the intensity of bMunc13-2 signals in control mice (ELKS1f/f) and ELKS1 conditional KO mice (ELKS1f/f;Emx1-Cre) as measured in all bMunc13-2 spots (C) and in bMunc13-2 spots colocalized with Bassoon (D). Note that a small population of bMunc13-2 spots with high bMunc13-2 labeling intensity is present in control, but not ELKS1 conditional KO, tissue. n = 372 for control mice and n = 53 for ELKS1 conditional KO in C; n = 35 for control mice and n = 16 for ELKS1 conditional KO in D. (E) Immunohistochemical analysis of the hilus of the dentate gyrus of control mice (ELKS1f/f) and ELKS1 conditional KO mice (ELKS1f/f;Emx1-Cre) with anti–bMunc13-2 (red) and anti-Bassoon (green) antibodies. The low-magnification images of the control brains (insets) show discrete bMunc13-2–positive signals, whose number was reduced in the ELKS1 conditional KO sample. Arrowheads indicate bMunc13-2–positive structures. Bars: (low -magnification, insets) 50 µm; (high magnification) 5 µm. Data are means ± SEM.
Figure 6.
Figure 6.
Partitioning of bMunc13-2 into the insoluble AZ protein fraction depends on the presence of ELKS1. (A) Western blot (WB) analysis of bMunc13-2 levels in cortex homogenates (total) and in detergent-insoluble and detergent-soluble fractions of cortex synaptosomes from ELKS1f/f and ELKS1f/f;Emx1Cre mice. The two panels at the very bottom show that PSD95 was enriched and RabGDI was depleted in the insoluble fraction. Note that different batches of prestained molecular weight markers were used for the blot sections at the bottom of the top panel (total) so that the molecular weight marker for 250 kD runs at a slightly smaller apparent molecular weight. Further, the soluble bMunc13-2 fraction (third and fifth blot sections from top) typically appears as two bands, both of which are absent in Munc13-2 KO samples. (B–D) Quantification of bMunc13-2 signals in detergent-insoluble (B) and soluble (C) fractions of synaptosomes and in homogenate (D). Band intensities were determined densitometrically on films using ImageJ (soluble and insoluble fractions) or by using Odyssey imaging of blots after incubation with fluorescent secondary antibodies (LICOR). The signals were then normalized to the ones obtained with anti-PSD95 (for B), anti-RabGDI (for C), and antitubulin (for D) antibodies. n = 5 for control and n = 3 for ELKSf/f;Emx1-Cre mice for B and C; n = 4 for control and for ELKSf/f;Emx1-Cre mice for D; Student’s t test, **, P < 0.01 for B, P = 0.756 for C, and P = 0.90 for D. Data are means ± SEM.
Figure 7.
Figure 7.
ELKS binding regulates the recruitment of bMunc13-2 to AZs in cultured hippocampal neurons. (A and B) Munc13-1/2 DKO neurons expressing the indicated bMunc13-2–EGFP fusion constructs. Neurons were costained for GFP, ELKS1, and Bassoon. bMunc13-2–EGFP expressed in Munc13-1/2 DKO neurons formed punctate structures and colocalized with immunostained Bassoon (A), whereas bMunc13-2(Δ145–187)–EGFP puncta showed only partial colocalization with Bassoon (B). Note that ELKS1 showed colocalization with Bassoon in A and B. Bars, 5 µm. (C and D) Quantification of the fractions of EGFP-positive (C) and ELKS1-positive (D) Bassoon spots in neurons expressing bMunc13-2–EGFP or bMunc13-2(Δ145–187)–EGFP. n > 25,000 spots were analyzed for EGFP, ELKS1, or Bassoon. Student’s t test (**, P < 0.01; *, 0.01 < P < 0.05). (E) Tukey box and whisker plot for mean intensities of EGFP and Bassoon signals (Bsn) in Bassoon spots of cultured neurons expressing bMunc13-2–EGFP (green) or bMunc13-2(Δ145–187)–EGFP (red). Boxes represent values from upper and lower quartiles, horizontal lines represent medians, vertical lines cover 99% of all values, and outliers are represented as filled circles. Bassoon puncta intensities show a slight reduction in bMunc13-2(Δ145–187)–EGFP-expressing neurons as compared with Munc13-2–EGFP-expressing neurons (5.441 ± 0.016, n = 16,850 in Munc13-2–EGFP-expressing neurons; 5.391 ± 0.015, n = 18,208 in Munc13-2(Δ145–187)–EGFP-expressing neurons; *, P = 0.0173). EGFP signals show a stronger intensity reduction in bMunc13-2(Δ145–187)–EGFP-expressing neurons than Munc13-2–EGFP-expressing neurons (0.991 ± 0.010, n = 16,850 in Munc13-2–EGFP-expressing neurons; 0.770 ± 0.007, n = 18,208 in Munc13-2(Δ145–187)–EGFP-expressing neurons; ***, P < 0.001). (F) Frequency distributions of EGFP signal intensities in Bassoon spots in bMunc13-2–-EGFP- (green) or bMunc13-2(Δ145–187)–EGFP-expressing (red) neurons. Note that bMunc13-2–EGFP-expressing neurons have a small fraction of Bassoon spots with high EGFP intensity. n = 16,850 for bMunc13-2–EGFP-expressing neurons and n = 18,208 for bMunc13-2(Δ145–187)–EGFP-expressing (red) neurons. (G) Scatterplots showing relationships between mean intensities of EGFP and Bassoon signals in bMunc13-2–EGFP- (green) or bMunc13-2(Δ145–187)–EGFP-expressing (red) neurons. Mean intensities were quantified from Bassoon and EGFP colocalizing spots. n = 1,549 for bMunc13-2–EGFP-expressing neurons and n = 895 for bMunc13-2(Δ145–187)–EGFP-expressing (red) neurons. Data are means ± SEM.
Figure 8.
Figure 8.
ELKS binding anchors bMunc13-2 to AZs in cultured hippocampal neurons. (A) Representative three-channel images of neurons expressing bMunc13-2–EGFP (top two rows) or bMunc13-2(Δ145–187)–EGFP (bottom two rows) together with tdTom-ELKS1. Neurons were fixed and stained with an anti-Bassoon antibody. Note that most of the recombinant bMunc13-2 signals colocalizing with tdTom-ELKS1 were accumulated at Bassoon-positive AZs (arrowheads). Bars, 5 µm. (B) Overview of axonal terminals imaged in the FRAP experiment in C. The region of interest for the FRAP experiment is indicated by the white box. Bar, 5 µm. (C) Images of a representative FRAP experiment with bMunc13-2–EGFP in Munc13-1/2 DKO neurons. The synapse was bleached, and images were obtained at 2-min intervals for the first 10 min after bleaching and subsequently at 5-min intervals. Images are shown at 10-min intervals after the first 10 min. (D) Time course of FRAP at the synapse shown in C. (E) Mean FRAP time course of the indicated bMunc13-2–EGFP constructs at synapses in Munc13-1/2 DKO neurons. The fluorescence recovery of bMunc13-2(Δ145–187)–EGFP was faster as compared with that of bMunc13-2–EGFP. Solid lines are mean traces measured for the two constructs. Dashed lines are results of fitting with two exponential curves. Data for bMunc13-2–EGFP are from 31 synapses (nine cells), and data for bMunc13-2(Δ145–187)–EGFP are from 31 synapses (seven cells). Data are means ± SEM.
Figure 9.
Figure 9.
Turnover of bMunc13-2 at AZs depends on ELKS1 levels. (A) Overview of axonal terminals imaged in the FRAP experiment in B. The regions of interest for the FRAP experiment are indicated by white boxes (a and b). Bar, 5 µm. (B) Representative images of tdTom-ELKS1 (top) and bMunc13-2–EGFP (second panels from top) before bleaching and recovery of EGFP signals along the time course at synapses with high (a) or low (b) tdTom-ELKS1 levels. (C) Mean FRAP time courses of bMunc13-2–EGFP at synapses with high (black, n = 22) or low (red, n = 22) levels of tdTom-ELKS1. Note the more rapid recovery of bMunc13-2–EGFP at synapses with low levels of tdTom-ELKS1 in B and C. Data are means ± SEM.

Similar articles

See all similar articles

Cited by 13 articles

See all "Cited by" articles

References

    1. Andrews-Zwilling Y.S., Kawabe H., Reim K., Varoqueaux F., and Brose N. 2006. Binding to Rab3A-interacting molecule RIM regulates the presynaptic recruitment of Munc13-1 and ubMunc13-2. J. Biol. Chem. 281:19720–19731. 10.1074/jbc.M601421200 - DOI - PubMed
    1. Aravamudan B., and Broadie K. 2003. Synaptic Drosophila UNC-13 is regulated by antagonistic G-protein pathways via a proteasome-dependent degradation mechanism. J. Neurobiol. 54:417–438. 10.1002/neu.10142 - DOI - PubMed
    1. Augustin I., Betz A., Herrmann C., Jo T., and Brose N. 1999a Differential expression of two novel Munc13 proteins in rat brain. Biochem. J. 337:363–371. 10.1042/bj3370363 - DOI - PMC - PubMed
    1. Augustin I., Rosenmund C., Südhof T.C., and Brose N. 1999b Munc13-1 is essential for fusion competence of glutamatergic synaptic vesicles. Nature. 400:457–461. 10.1038/22768 - DOI - PubMed
    1. Banker G.A., and Cowan W.M. 1977. Rat hippocampal neurons in dispersed cell culture. Brain Res. 126:397–42. 10.1016/0006-8993(77)90594-7 - DOI - PubMed

Publication types

MeSH terms

Substances

Associated data

Feedback