Deletion of β-Neurexins in Mice Alters the Distribution of Dense-Core Vesicles in Presynapses of Hippocampal and Cerebellar Neurons

Abstract

Communication between neurons through synapses includes the release of neurotransmitter-containing synaptic vesicles (SVs) and of neuromodulator-containing dense-core vesicles (DCVs). Neurexins (Nrxns), a polymorphic family of cell surface molecules encoded by three genes in vertebrates (Nrxn1-3), have been proposed as essential presynaptic organizers and as candidates for cell type-specific or even synapse-specific regulation of synaptic vesicle exocytosis. However, it remains unknown whether Nrxns also regulate DCVs. Here, we report that at least β-neurexins (β-Nrxns), an extracellularly smaller Nrxn variant, are involved in the distribution of presynaptic DCVs. We found that conditional deletion of all three β-Nrxn isoforms in mice by lentivirus-mediated Cre recombinase expression in primary hippocampal neurons reduces the number of ultrastructurally identified DCVs in presynaptic boutons. Consistently, colabeling against marker proteins revealed a diminished population of chromogranin A- (ChrgA-) positive DCVs in synapses and axons of β-Nrxn-deficient neurons. Moreover, we validated the impaired DCV distribution in cerebellar brain tissue from constitutive β-Nrxn knockout (β-TKO) mice, where DCVs are normally abundant and β-Nrxn isoforms are prominently expressed. Finally, we observed that the ultrastructure and marker proteins of the Golgi apparatus, responsible for packaging neuropeptides into DCVs, seem unchanged. In conclusion, based on the validation from the two deletion strategies in conditional and constitutive KO mice, two neuronal populations from the hippocampus and cerebellum, and two experimental protocols in cultured neurons and in the brain tissue, this study presented morphological evidence that the number of DCVs at synapses is altered in the absence of β-Nrxns. Our results therefore point to an unexpected contribution of β-Nrxns to the organization of neuropeptide and neuromodulator function, in addition to their more established role in synaptic vesicle release.

Keywords: dense-core vesicles; electron microscopy; exocytosis; neuromodulators; neuropeptides; secretion; synapse function.

FIGURE 1

Diminished chromogranin A (ChrgA) levels in β-neurexins- (β-Nrxns-) deficient hippocampal neurons. (A) Primary hippocampal neurons from floxed β-Nrxn knock-in (β-KI) mice were transduced with lentivirus expressing inactive Cre recombinase (Cre^mut) and immunostained at DIV18 with anti-ChrgA (A₁) and anti-neurofilament H (NFH, A₂). Merged image (A₃) shows the presence of the dense-core vesicle (DCV) marker ChrgA in axons. Scale bar: 3 μm. (B,C) Cultured neurons as in panel (A) but colabeled for ChrgA (B₁,C₁) and the dendritic marker MAP2 (B₂,C₂). Successful transduction of neurons with lentivirus is demonstrated by nuclear expression of inactive Cre^mut or active Cre recombinase (Cre) fused to GFP (B₃,C₃). Merged images (B₄,C₄) show the exclusion of ChrgA (arrows) from MAP-positive dendrites (arrowheads) independent of the presence of β-Nrxn. Scale bars: 20 μm. (D) Representative images of ChrgA-positive axons of Cre^mut control (D₁) and Cre transduced conditional knockout (cKO) (D₂) neurons from experiments as in panel (A) that were used for measurements of fluorescence intensity (yellow boxes). Note that the fluorescence intensity appears to be overall reduced in β-Nrxn-deficient axons with no change in the labeling pattern and without apparent ectopic accumulation. Scale bar: 4 μm. (E) Histogram comparing the ChrgA fluorescence intensity over axons from Cre^mut control (black bars) and β-Nrxn-deficient Cre neurons (red bars) as shown in panel (D). Intensities (AU = arbitrary units) were normalized to control values. Data are shown as mean ± SEM, measurements are based on n = 15 axonal regions from the three independent cultures per genotype; significance difference indicated as **p < 0.01, two-tailed unpaired t-test.

FIGURE 2

Reduced number of DCVs in presynaptic terminals of β-Nrxn-deficient hippocampal neurons. (A) Representative transmission electron microscopic images of synaptic boutons from primary hippocampal neurons of floxed β-KI mice at DIV18 that were transduced with lentivirus expressing inactive Cre recombinase (Cre^mut). Presynapses of these cultured control neurons often contain 1 or more DCVs (arrows, A₁–A₄). Boutons without DCV are marked by asterisks (A₁,A₄). SV, synaptic vesicle clusters; arrowheads, postsynaptic densities. Scale bar: 250 nm. (B) Ultrastructure of synaptic boutons as in panel (A) but from β-Nrxn-deficient (Cre) neurons revealing fewer presynapses with DCVs (arrows, B₁,B₂,B₄) and more without DCVs (asterisks, B₂–B₄). Scale bar: 250 nm. (C) Histogram showing the area density of DCVs in synaptic boutons of Cre^mut control (black bars) and β-Nrxn-deficient Cre (red bars) neurons as shown in panels (A,B). Numbers of DCVs were normalized to control values (for actual values, see section “Results”). Data are shown as mean ± SEM; dots indicate individual data points, measurements are based on n = 15 areas from three independent cultures per genotype; significance difference indicated as ***p < 0.0001, two-tailed unpaired t-test. (D) Histogram displaying the percentage of presynaptic boutons without DCV = 0, a DCV = 1, or DCV = 2. on random cross-sections of Cre^mut control (black bars) and β-Nrxn-deficient Cre (red bars) cultures as analyzed in panel (C). Note the increase of synaptic profiles without DCVs upon the deletion of β-Nrxn. Data are shown as mean ± SEM; dots indicate individual data points. The relative distribution is based on n = 162 (DCV = 0), n = 64 (DCV = 1), and n = 32 (DCV = 2) boutons of Cre^mut neurons, and on n = 235 (DCV = 0), n = 33 (DCV = 1), and n = 7 (DCV = 2) boutons of Cre neurons from the three independent cultures per genotype; significance difference indicated as ***p < 0.0001, **p < 0.01, or *p < 0.05, one-way ANOVA with Holm–Sidak’s multiple comparison. (E) Similar histogram to panel (D) showing the average area size of boutons without or with DCVs from Cre^mut control (black bars) and β-Nrxn-deficient Cre (red bars) neurons. Data are shown as mean ± SEM. Measurements are based on n = 60 (DCV = 0 and DCV = 1) or n = 30 (DCV = 2) boutons from the three independent cultures per genotype; significance difference indicated as n.s. = non-significant, two-tailed unpaired t-test.

FIGURE 3

Normal ChrgA-to-Bassoon alignment in synapses of β-Nrxn-deficient neurons. (A,B) Primary hippocampal neurons from floxed β-KI mice were transduced with lentivirus expressing inactive Cre^mut (A₁–A₄) or active Cre recombinase (B₁–B₄) and immunostained at DIV18 with anti-ChrgA (arrows, A₁,B₁) and anti-Bassoon (A₂,B₂). Merged images (A₃,B₃) show the more restricted expression of the DCV marker ChrgA (arrows, green label) in comparison to the ubiquitous punctate distribution of the presynaptic marker Bassoon (red label). Note the juxtaposed localization of these molecules in high magnification images (arrows, A₄,B₄). Scale bars in panels (A₁,B₁) for panels (A₁–A₃,B₁–B₃): 20 μm; scale bars in panels (A₄,B₄): 8 μm. (C) Representative traces of whole-cell patch clamp recordings of pharmacologically isolated miniature excitatory postsynaptic currents (mEPSCs) show a visibly reduced frequency of mini events in β-Nrxn-deficient (Cre, red) neurons compared to controls (Cre^mut, black). (D) Averaged individual mEPSC traces from more than 100 consecutive events reveal similar amplitudes and kinetics in β-Nrxn-deficient (Cre, red trace) neurons compared to controls (Cre^mut, black trace). (E,F) The reduced mEPSC frequency in Cre neurons as shown in panel (C) is reflected by longer inter-event intervals (IEIs) (E), whereas the mEPSC amplitude does not differ between genotypes (F). Data are shown as mean ± SEM, dots indicate individual data points, measurements are based on n = 7 Cre^mut control neurons with 647 mEPSC events (black bar) and n = 7 β-Nrxn-deficient Cre neurons with 575 mEPSCs (red bar) from the 3 independent cultures per genotype; significance difference indicated as ***p < 0.0001 or n.s. = non-significant, two-tailed unpaired t-test.

FIGURE 4

Unchanged Golgi marker expression in β-Nrxn-deficient hippocampal neurons. (A,B) Primary hippocampal neurons from floxed β-KI mice were transduced with Cre^mut (A₁–A₃) or Cre (B₁–B₃) expressing lentivirus and immunostained at DIV18 with anti-syntaxin 6 (Stx6) as a marker protein of the trans-Golgi network (tGN, A₁,B₁). Successful transduction with lentivirus is demonstrated by nuclear expression of inactive Cre^mut or active Cre recombinase (Cre) fused to GFP (A₂,B₂). Merged images (A₃,B₃) show similar distribution of Stx6 in the Golgi apparatus. Arrows, examples of Stx6-postive clusters. Scale bars: 20 μm. (C,D) Cultured neurons as in panels (A,B) but immunostained against the cis-Golgi marker protein GM130 (C₁,D₁) with GFP-Cre autofluorescence indicating successful transduction of labeled neurons (C₂,D₂). Merged images (C₃,D₃) reveal no difference in the absence of β-Nrxn. Arrows, examples of GM130-postive clusters. Scale bars: 20 μm.

FIGURE 5

Reduced DCV numbers in cerebellar parallel fiber terminals of β-Nrxn-deficient mice. (A1,A2) Representative electron microscopic images of type 1 synapses in the cerebellar molecular layer of β-KI control mice, likely corresponding to excitatory parallel fiber terminals of granule cells. Arrows point to DCVs at the periphery of synaptic vesicle clusters. SV = synaptic vesicle clusters; arrowheads, postsynaptic densities. (B1,B2) Similar images as in panel (A1,A2) from β-Nrxn-deficient β-TKO cerebellum. Scale bar, for panels (A,B), 250 nm. (C,D) Histograms showing the number of DCVs (C) and the area density of type 1 excitatory synapses (D) in the molecular cell layer from β-KI control (black bars) and β-Nrxn-deficient β-TKO (red bars) cerebella. Data are normalized to control values and shown as mean ± SEM (for actual values, see section “Results”), dots indicate individual data points. Measurements are based on n = 9 cerebellar regions from three mice per genotype, corresponding to 3,123 μm² of total area investigated; significance difference indicated as ***p = 0.0002 or n.s. = non-significant, two-tailed unpaired t-test. (E–G) Histogram summarizing the average area size (E), average number of SVs (F), and average length of the active zone (G) of presynaptic boutons from β-KI control (black bars) and β-Nrxn-deficient β-TKO (red bars) parallel fiber terminals. Samples as in panels (C,D), data are shown as mean ± SEM; significance difference indicated as n.s. = non-significant, two-tailed unpaired t-test.

FIGURE 6

Normal numbers of cerebellar granule cells (CGCs) in β-Nrxn-deficient mice. (A) Nissl-stained parasagittal sections through the brains of β-KI control (A₁) and β-Nrxn-deficient β-TKO (A₂) mice. Ce = cerebellum; box indicates approximate position of area investigated in panels (B,C). Scale bar: 2 mm. (B) Representative images of 1-μm semithin sections from β-KI control (B₁) and β-Nrxn-deficient β-TKO (B₂) cerebellum stained with toluidine blue dye. GCL, granule cell layer; PCL, Purkinje cell layer; MCL, molecular cell layer. Scale bar: 125 μm. (C) Histogram showing the area density of cells in the GCL of the cerebellum of β-KI control (black bars) and β-Nrxn-deficient β-TKO (red bars) mice. Number of granule cells was counted in a 1 mm² area and β-TKO data normalized to control. Data are shown as mean ± SEM, dots indicate individual data points, measurements are based on n = 6 areas from three animals per genotype; significance difference indicated as n.s. = non-significant, two-tailed unpaired t-test.

FIGURE 7

Normal Golgi morphology in CGCs of β-Nrxn-deficient mice. (A₁,A₂) Representative electron microscopic images of the Golgi apparatus (Go) of CGCs from β-KI control mice. Note visible DCVs (magenta arrows) in the vicinity of Golgi stacks. Nu, nucleus; arrowheads, coated vesicles. (B₁,B₂) Similar images as in panels (A₁,A₂) but from β-Nrxn-deficient β-TKO cerebellum. Labels as in panels (A₁,A₂). Scale bar for all images: 250 nm. (C,D) Histograms showing the number of cisternae (C) and the width of cisternae (D) on random cross-sections of the GCL from β-KI control (black bars) and β-Nrxn-deficient β-TKO (red bars) cerebella. Data are shown as mean ± SEM, dots indicate individual data points. Measurements are based on n = 31 Golgi apparatus from three mice per genotype; significance difference indicated as n.s. = non-significant, two-tailed unpaired t-test.