Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2016 Oct 11;113(41):11615-11620.
doi: 10.1073/pnas.1605256113. Epub 2016 Sep 26.

RIM-binding Protein 2 Regulates Release Probability by Fine-Tuning Calcium Channel Localization at Murine Hippocampal Synapses

Affiliations
Free PMC article

RIM-binding Protein 2 Regulates Release Probability by Fine-Tuning Calcium Channel Localization at Murine Hippocampal Synapses

M Katharina Grauel et al. Proc Natl Acad Sci U S A. .
Free PMC article

Abstract

The tight spatial coupling of synaptic vesicles and voltage-gated Ca2+ channels (CaVs) ensures efficient action potential-triggered neurotransmitter release from presynaptic active zones (AZs). Rab-interacting molecule-binding proteins (RIM-BPs) interact with Ca2+ channels and via RIM with other components of the release machinery. Although human RIM-BPs have been implicated in autism spectrum disorders, little is known about the role of mammalian RIM-BPs in synaptic transmission. We investigated RIM-BP2-deficient murine hippocampal neurons in cultures and slices. Short-term facilitation is significantly enhanced in both model systems. Detailed analysis in culture revealed a reduction in initial release probability, which presumably underlies the increased short-term facilitation. Superresolution microscopy revealed an impairment in CaV2.1 clustering at AZs, which likely alters Ca2+ nanodomains at release sites and thereby affects release probability. Additional deletion of RIM-BP1 does not exacerbate the phenotype, indicating that RIM-BP2 is the dominating RIM-BP isoform at these synapses.

Keywords: RIM-BP2; active zone structure; calcium channel coupling; release probability; short-term plasticity.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.
Fig. 1.
RIM-BP2 localization at the hippocampal AZ, and effect of RIM-BP2 deletion on synaptic transmission in autaptic hippocampal neurons. (A, iiii) Spatial organization of RIM-BP2 in relation to Bassoon (Bsn) and MUNC13-1 at the AZ of CA3-CA1 synapses in WT mouse brain cryosections imaged by gSTED and analyzed using the mean k-nearest neighbor distance between clusters. (Scale bar: 500 nm.) (B) EPSCs evoked by 2-ms somatic depolarization in RIM-BP2 WT (black) and KO (red) autaptic neurons. Amplitudes (ampl.) were normalized (norm.) to WT mean of the same culture. (C) C.V.s of 24 EPSC amplitudes recorded during a period of 2 min. (D) Synaptic responses to application of hypertonic sucrose (500 mM) solution probing the RRP. (E) PVR of the same cells as in B and D. (F) Spontaneous release and averages of miniature EPSCs (mEPSCs) from the same cells. (G) mEPSC amplitudes and frequencies (freq.). (H) Normalized traces of two EPSCs at an ISI of 25 ms. (I) PPR calculated for the indicated ISIs. Normalized amplitudes of the same cells as in I in response to 5 APs triggered at 50 Hz (J) or 50 APs at 10 Hz (K). The last 10 EPSCs of the 10-Hz train are larger in KO neurons compared with WT. The numbers of neurons and independent cultures analyzed are shown within the bars. Data are expressed as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001.
Fig. S1.
Fig. S1.
RIM-BP expression and interaction with RIM and Munc13-1. (A) Relative (rel.) expression levels of RIM-BP1–3 in cultured murine hippocampal neurons over time. The mRNA levels are normalized to Rpl4. Data are expressed as mean ± SEM, and tested against RIM-BP2 expression levels (pDIV23,RIM-BP1 = 0.001, pDIV10, RIM-BP3 = 0.001, pDIV23,RIM-BP3 < 0.001, Tukey’s multiple comparisons test; n = 4). *P < 0.05; **P < 0.01; ***P < 0.001. (B) Confocal (Left) and corresponding gSTED (Right) images of RIM-BP2 localization to Bassoon (Bsn) and MUNC13-1 at CA3-CA1 hippocampal synapses on a WT mouse brain cryosection. (Scale bar: 500 nm.) (C, iv and D, iv) Estimation of the effective lateral point spread function of the gSTED microscope. The gSTED images of a 40-nm fluorescent bead with excitation and emission maxima at 505 and 515 nm (C, i) or 660 and 680 (D, i), respectively, are shown. (Scale bar: 500 nm.) Three-dimensional surface plot of the local intensity profile (C, ii and D, ii) approximated with a 2D Gaussian fit (C, iii and D, iii). (C, iv and D, iv) Deviation of the 2D Gaussian fit from the intensity profile. (C, v and D, v) Intensity profile and 2D Gaussian fit values along the x axis for constant y (126.25 nm) (Left) and the y axis for constant x (126.25 nm) (Right). In both channels, the smallest bead had a FWHM of 54 nm (C, v) or 45 nm (D, v), indicating a lateral resolution of ∼50 nm. (E) Representative Western blot showing input, IgG control, and immunoprecipitation of RIM-BP2 probed with antibodies against different synaptic proteins. RIM-BP2 forms a complex with the AZ proteins RIM1/2 and MUNC13-1, but not with the endocytic protein Dynamin1 (Dyn1) or the cytomatrix protein GIT.
Fig. S2.
Fig. S2.
Generation and characterization of RIM-BP2 KO mice. (A) Schematic representation of the RIM-BP2 targeting strategy showing the endogenous and recombined locus, as well as the RIM-BP2 locus after Flp-mediated and Cre-mediated excision. The RIM-BP2 locus after Cre-mediated excision corresponds to the constitutive RIM-BP2 KO. (B) Mendelian distribution of RIM-BP2–deficient progeny [total number of litters analyzed = 34; total number of animals: WT, 77; heterozygous (HET), 151; KO, 57]. (C) RIM-BP2 expression in the CA1 area of the hippocampus of WT and KO mice. Note the loss of RIM-BP2 immunoreactivity in the KO (WT, n = 4; KO, n = 2). (Scale bar: 100 μm.) (D, Left) RIM-BP2 protein expression in WT (n = 6) and KO (n = 6) mouse crude P2 membrane preparations probed with an RIM-BP2–specific antibody recognizing amino acids 589–869 of rat RIM-BP2, upstream of the deletion site. (D, Right) Expression levels of the AZ proteins RIM1/2, MUNC13-1, and Erc1b/2 and synaptic vesicle proteins Synaptophysin1 (Syp1) and Synapsin1 (Syn1) in the WT and KO mouse are not affected. HSC70 and β-actin were used as loading controls. (E) Quantification of signals from the Western blot analysis in D showing complete loss of RIM-BP2 levels (n = 6; **P < 0.01, Mann–Whitney U test), but no significant alterations in the levels of other synaptic proteins analyzed (n = 6; P > 0.05, ERC1b/2/Syp1, Mann–Whitney U test; P > 0.05, MUNC13-1/RIMs/Syn1, unpaired Student t test). Data are expressed as median (25th–75th percentiles). Circles indicate outliers, and triangles indicate extremes. (F) RIM-BP2 protein expression in WT (n = 6) and KO (n = 6) mouse crude P2 membrane preparations probed with RIM-BP2 antibody recognizing the last 20 amino acids of mouse RIM-BP2 C terminus. β-Actin was used as a loading control. (G) Representative electron microscopic images of WT and RIM-BP2 KO synapses. (Scale bars: 100 nm.) (H) Bar graph of total length of the AZ (Left) and number of docked vesicles (Right). (I) Distribution of synaptic vesicles relative to the AZ. Data are expressed as mean ± SEM.
Fig. S3.
Fig. S3.
PR, molecular priming and RRP replenishment in RIM-BP2 KO neurons. (AC) Progressive block of the NMDA receptor-mediated component of the EPSC by MK-801. (A) Example traces normalized to the AMPA component of the EPSC. The first, 10th, and 20th EPSCs in the presence of MK-801 are labeled. (B) Normalized NMDA receptor-mediated EPSCs during application of 5 μM MK-801 and double exponential fits. The rate of MK-801 block is reduced in RIM-BP2 KO neurons. (C) Same data as in B, but the x axis was expanded by 18% for RIM-BP2 KO. The alignment of the manipulated curves indicates a reduction in PR in RIM-BP2 KO neurons by ∼18%. (D) Representative traces of synaptic responses induced by 250 mM or 500 mM sucrose solution in autaptic neurons (WT, black; KO, red). (E) Fraction of RRP released by 250 mM sucrose application. (F) Peak release rates in 500 mM sucrose. (G) Spontaneous (spont.) release rates, calculated as the ratio of mEPSC frequency and number of vesicles in the RRP. (H) Representative traces of EPSCs before and after RRP depletion by sucrose application (blue vertical dashed line indicates time point of sucrose application). Responses are normalized to the average EPSC before RRP depletion. Stimulation artifacts were blanked for better visibility. (I) Summary graph of average recovery of EPSC amplitudes after depletion of the RRP by hypertonic sucrose application (WT, n = 35; KO, n = 39). (J) PPRs for different ISIs and in different [Ca2+]exts. (K) EPSC amplitudes and C.V.s of EPSC amplitudes in different [Ca2+]exts. The numbers of neurons analyzed, and independent experiments are shown within the bars. Data are expressed as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001.
Fig. 2.
Fig. 2.
RIM-BP2 deletion alters STP in acute hippocampal slices. (A) Input/output curves relating the amplitude of the presynaptic fiber volley to the fEPSP amplitude in the stratum radiatum of area CA1 of acute hippocampal slices at different stimulus intensities (0.05 mV–0.2 mV: WT, n = 21; KO, n = 19; 0.3 mV: WT, n = 18; KO, n = 14; 0.4 mV: WT, n = 16; KO, n = 13). (B) Normalized fEPSPs in response to a paired pulse with 50-ms ISI. (C) PPRs of fEPSPs at different ISIs. (D) fEPSPs elicited by a 14-Hz stimulation train, with stimulation artifacts blanked for better visibility. (E) mEPSC amplitudes and frequencies recorded from CA1 pyramidal neurons. The numbers of slices and independent animals are indicated. Data are expressed as mean ± SEM. ***P < 0.001.
Fig. S4.
Fig. S4.
STP of fEPSPs in the stratum radiatum of RIM-BP1/2 DKO mice. (A) Schematic representation of the Bzrap1 (RIM-BP1) targeting strategy showing the endogenous and recombined locus, as well as the Bzrap1 locus after Flp-mediated and Cre-mediated excision. The Bzrap1 locus after Cre-mediated excision corresponds to the constitutive Bzrap1 KO lacking exons 23–25, which leads to a frameshift and a premature stop codon in exon 27. If the resulting RNA were translated, the resulting truncated 174-kDa protein would lack the second and third SH3 domains, and therefore would likely not be functional. (B) Southern blot analysis for 3′ homologous recombination (Rec) in ES cells performed by genOway. The genomic DNA of the tested ES cell clones was compared with WT DNA (C57BL/6). The digested DNA samples were blotted on nylon membrane and hybridized with an external 3′ probe (LA-E-A probe) hybridizing downstream of the targeting vector homology sequence (genOway). (C) Southern blot analysis for 5′ homologous recombination in ES cells performed by genOway. The genomic DNA of tested ES cell clones was compared with WT DNA (C57BL/6). The digested DNA samples were blotted on nylon membrane and hybridized with the Neo probe detecting the EcoRV fragment to screen for 5′ homologous recombination events (genOway). (D) Schematic representation of different real-time PCR (RT-PCR) assays for RIM-BP1 and RT-PCR results of cDNA obtained from RIM-BP1 WT, heterozygous (Het), and homozygous (Ho) KO mouse brain. Expression relative to Rpl4 was normalized (norm.) to WT. The RT-PCR assay 21 confirms that exons 23–25 are deleted in the RIM-BP1 KO. Assays spanning exons before and after the deleted region show a 50% reduction of mRNA levels in the KO. The remaining signals are likely due to the presence of frameshift mRNA that has not been completely degraded. This mRNA would not yield a functional protein. (E) Genotyping results obtained using the three primers indicated in A and genomic DNA from tails of WT, Het, and RIM-BP2 Ho KO littermates. The WT band is larger than the KO band, as indicated. In Het animals, both bands are present. (F) Western blot of P2 fractions of RIM-BP1 WT (n = 5) and KO (n = 5) mice using the commercially available N-terminal antibody against RIM-BP1 (Synaptic Systems). In RIM-BP1 KO samples, the upper band (arrow) is missing. This band likely corresponds to the RIM-BP1 isoforms with a molecular mass of ∼200 kDa. However, the RIM-BP1 antibody recognizes several bands at molecular mass that do not correspond to any of the isoforms described so far, indicating unspecific binding. (G) Input/output curves relating the fEPSP amplitude to the amplitude of the presynaptic fiber volley in stratum radiatum of the CA1 region in acute hippocampal slices (0.05 mV: WT, n = 20; KO, n = 21; 0.1 mV: WT, n = 26; KO, n = 26; 0.15 mV: WT, n = 24; KO, n = 25; 0.2 mV: WT, n = 19; KO, n = 26; 0.25 mV: WT, n = 11; KO, n = 18; 0.3 mV: WT, n = 9; KO, n = 19). (H) Summary graph of PPRs of fEPSPs in the stratum radiatum in response to paired stimulation with indicated ISIs (nWT = 28, nKO = 34). (I) Summary graph of fEPSPs in response to a 14-Hz stimulation train (nWT = 10, nKO = 17). Data are expressed as mean ± SEM. n.s., not significant.
Fig. 3.
Fig. 3.
RIM-BP2 deletion alters Ca2+ sensitivity of release. (A) Schematic representation of the genetically encoded Ca2+ indicator synGCamp6f at the membrane of a synaptic vesicle. Heat-colored images of a WT dendrite (B) and quantification of synGCamp6f fluorescence change (ΔF/F) during stimulation with two APs at 20 Hz and 50 APs at 10 Hz (C). (D) Mean EPSCs of WT (black) and RIM-BP2 KO (red) autaptic neurons in different [Ca2+]exts. The [Ca2+]exts of test solutions are indicated (all 1 mM Mg2+). A control solution (2 mM Ca2+/4 mM Mg2+) was applied in between. (E) EPSC amplitudes at different [Ca2+]exts normalized to alternating control responses (Left) and to 10 mM [Ca2+]ext (Right). Hill functions were fitted to the data. (F) Normalized EPSCs in response to a paired stimulus and mean PPRs at indicated [Ca2+]exts. (G) PVR of autaptic hippocampal neurons in 0.5 mM and 4 mM [Ca2+]exts. The numbers of neurons and independent cultures analyzed are shown within the bars. Data are expressed as mean ± SEM. *P < 0.05. n.s., not significant.
Fig. 4.
Fig. 4.
RIM-BP2 loss results in defective CaV2.1 clustering at the AZ. (A) CaV2.1 and Bsn clusters imaged in situ at CA3-CA1 hippocampal synapses of RIM-BP2 WT and KO mice by dual-channel gSTED (Bottom) compared with confocal acquisition (Top). (B) Schematic representation of two kinds of cluster analysis. (B, i) Bsn clusters within indicated sampling distances (e.g., 50 nm, 75 nm) to a given CaV2.1 cluster were quantified and averaged on thousands of CaV2.1 clusters per image. (B, ii) Mean distance between the k-nearest neighbor Bsn cluster and CaV2.1 cluster (k = 1, k = 2, k = 3, k = 4, k = 5). (C) Bsn cluster numbers at short distances from CaV2.1 clusters (WT, n = 5; KO, n = 6). (DF) Mean k distance between Bsn and CaV2.1 clusters (D) and between neighboring Bsn clusters (E) in RIM-BP2 KO mice. (G) Triple-channel gSTED images of CaV2.1, RIM1, and Homer1 at CA3-CA1 hippocampal synapses (WT, n = 9; KO, n = 9). (H) RIM1 clusters found in proximity to CaV2.1 channels. (I) Mean k distance of RIM1 clusters to CaV2.1 clusters. Homer1-RIM1 clustering (J) and Homer1-RIM1 mean k distance (K). Homer1 clusters close to CaV2.1 channels (L), and mean k distances between Homer1 clusters and CaV2.1 (M) are shown. (N) CaV2.1 spatial organization relative to RIM1 and Homer1 at excitatory hippocampal synapses. (Scale bars: A and G, 500 nm.) Distances between clusters are represented in nanometers. Values are expressed as mean ± SEM. *P < 0.05; **P < 0.01.
Fig. S5.
Fig. S5.
RIM-BP2 does not significantly alter the number of Cav2.1 clusters. Cluster analysis performed on gSTED images of CA3-CA1 hippocampal synapses in situ shows that RIM-BP2 deletion does not significantly alter either the number of Bsn and Cav2.1 clusters (A) or their ratio (B) (WT, n = 5; KO, n = 6). (C) Cav2.1, RIM1, and Homer1 distribution at CA3-CA1 hippocampal synapses imaged by conventional confocal (Left) and gSTED (Right) microscopy in brain cryosections of RIM-BP2 WT (Upper) and KO (Lower) mice. (Scale bar: 500 nm.) (D) Quantification of the number of clusters imaged by gSTED (WT, n = 9; KO, n = 9). Also, this independent experiment showed that the number of Cav2.1 clusters does not significantly change in RIM-BP2 KO mice. (E) Similarly, RIM-BP2 deletion does not significantly alter either the number of Homer1 and RIM1 clusters or their ratio to Cav2.1 channels. However, as shown in D and E, we observed a higher variability in the total number of RIM1 clusters, and therefore in the ratio of RIM1/Cav2.1 clusters at RIM-BP2 KO synapses.
Fig. S6.
Fig. S6.
EGTA-AM reduces PVR significantly more in RIM-BP2 KO than in WT. (A) Representative traces of average EPSCs of WT and RIM-BP2 KO neurons preincubated with 25 μM EGTA-AM or DMSO control. (B) Normalized EPSC amplitudes of neurons preincubated with 25 μM EGTA-AM or DMSO control. EPSCs are reduced in EGTA-AM–treated neurons, but the effect is not significant. (C) PVR of the same neurons as in A, normalized to control. EGTA-AM has a significantly stronger effect on the PVR of RIM-BP2 KO neurons than on WT, supporting larger coupling distances. Data are expressed as mean ± SEM. *P < 0.05; ***P < 0.001.

Similar articles

See all similar articles

Cited by 25 articles

See all "Cited by" articles

Publication types

MeSH terms

Feedback