Article Figures & Data
Figures
- Figure 1. Construction of epitope-tagged rim alleles using CRISPR/Cas9.
(A) Overview of the Drosophila melanogaster rim locus on chromosome III and illustration of genomic targeting strategy. #1 and #2 mark positions of the gRNA probes. (B) Location and size of the rim gene fragment excised by CRISPR/Cas9 and replaced by φC31-mediated integration to generate rimrescue-Znf (gray), rimV5-Znf (blue), and rimHA-Znf (magenta) alleles. UTR, untranslated region. (C) Drosophila RIM protein domain positions: Rab3/zinc finger (Znf), PDZ, C2A, PxxP, and C2B domains and position of the N-terminal epitope tag (V5-Znf or HA-Znf) is indicated.
- Figure 2. Intact synaptic transmission at rimHA−Znf neuromuscular junctions.
(A) Miniature excitatory postsynaptic currents (mEPSCs) at WT (black), rimrescue−Znf (gray), rimV5−Znf (blue), and rimHA−Znf (magenta) neuromuscular junctions (NMJs). (B) Mean ± SEM mEPSC amplitude and frequency in wt (n = 15 NMJs in seven animals), rimrescue−Znf (n = 17/8), rimV5−Znf (n = 12/6), and rimHA−Znf (n = 12/5). Scatter plots show individual data points, individual P-values are indicated. (C) Evoked EPSCs (eEPSCs) samples at NMJs of the same four genotypes. (D) eEPSC amplitude and paired-pulse ratios with 30 ms interstimulus interval in wt (n = 10 NMJs in seven animals), rimrescue−Znf (n = 13/7), rimV5−Znf (n = 12/5), and rimHA−Znf (n = 11/5). (E) Mean ± SEM eEPSC amplitude, mEPSC amplitude and quantal content in rimrescue−Znf and rimHA−Znf animals treated with PhTx in DMSO (+, open circles) or DMSO (−, filled circles). rimHA−Znf larvae show undisturbed presynaptic homeostatic potentiation in response to PhTx stimulation (rimrescue−Znf: 14 NMJs in seven animals in DMSO, 13/6 in PhTx; rimHA−Znf: 11/5 in DMSO, 13/5 in PhTx).
- Figure 3. Expression of RIMHA−Znf in the Drosophila central and peripheral nervous systems.
(A) Confocal image of the ventral nerve cord of a male third instar larva stained with α-HRP (blue), BrpNc82 (green), and rabbit monoclonal α-HA antibody to visualize RIMHA−Znf (magenta). (B, C) α-HRP, BrpNc82, and RIMHA−Znf at a neuromuscular junction (NMJ) on abdominal muscles 6/7. White box highlights the enlarged region in (C). (C) Co-localization of BrpNc82 and RIMHA−Znf in presynaptic boutons. (D) Whisker plots with median for Brp puncta per NMJ (normally distributed data, mean indicated by dashed line), NMJ area, and number of boutons per NMJ in rimrescue−Znf (n = 24 NMJs in nine animals, only 23 NMJs for boutons per NMJ) and rimHA−Znf (n = 22/9). Scale bars in (A) 100 μm, in (B) 10 μm, and in (C) 3 μm.
- Figure S1. Expression of RIMV5−Znf at presynaptic terminals.
Staining of α-HRP (blue) and RIMV5−Znf (magenta) shows the expression of RIMV5−Znf at presynaptic type Ib and type Is boutons of a third instar larval neuromuscular junction on abdominal muscles 6/7 in segment A3. Scale bar 3 μm.
- Figure 4. RIMHA−Znf clusters at presynaptic AZs.
(A) Two-channel dSTORM localizations for a rimHA−Znf type Ib bouton. Left: RIMHA−Znf stained with α-HA antibody and Alexa Fluor647 conjugated F(ab’)2 fragments (magenta). Middle: overlay with BrpNc82 labeled with Alexa Fluor 532-conjugated IgGs (green). (C) Asterisk marks enlarged AZ in (C). Right: RIMHA−Znf localizations from left panel with all localizations with Euclidian distance >20 nm to Brp localizations removed. The removed signal is considered noise. Individual RIMHA−Znf subclusters (SCs) were extracted by HDBSCAN and assigned to nearest AZs by color. (B) Averaged H function (gray, mean ± SD) from n = 542 RIMHA−Znf first-level clusters obtained from 18 NMJs and nine animals (maximum of the curve indicates a mean SC radius of 6 nm) and histogram (black) of the mean radius from n = 11,094 RIMHA−Znf SCs (estimated from SC area under the assumption of a circular area, median (25th–75th percentile): 6.4 (3.4–10.9) nm). Dashed black line indicates the prediction for a random Poisson distribution. (A, C) Enlarged plots of the AZ marked in (A). Left: two-channel dSTORM localizations of RIMHA−Znf and BrpNc82. Right: RIMHA−Znf SCs extracted by HDBSCAN in different colors. Colored lines indicate alpha shapes used for area determination. The center of mass (c.o.m.) of the corresponding AZ (x) is indicated. Dashed red line shows the Euclidian distance between the AZ c.o.m and an SC c.o.m., referred to as radial distance. Parameters in HDBSCAN were: minimum cluster size = 100 localizations, minimum samples = 25 localizations for BrpNc82; minimum cluster size = 2 localizations, minimum samples = 2 localizations for RIMHA−Znf; α-value BrpNc82 = 800 nm2, α-value RIMHA−Znf = 300 nm2; exclusion criteria for BrpNc82 clusters were area ≤ 0.03 μm2 and ≥ 0.3 μm2. (D) Distributions of RIMHA−Znf SC area (11,094 SCs from 18 NMJs in nine animals) and the number of RIMHA−Znf SCs per AZ (n = 893 AZs from 18 NMJs in nine animals) in all AZs without selection according to AZ circularity (see the Material and Methods section). Inset in the left panel highlights the range between 0 and 400 nm2 SC area. Solid red line indicates median, dashed red lines, 25th and 75th percentiles. Scale bars in (A) 1 μm, in (C) 100 nm.
- Figure 5. RIMHA−Znf subclusters are recruited to SpCs in acute PHP.
(A) Localizations of a type Ib AZ from a phtx larva. Left: two-channel dSTORM data of RIMHA−Znf (magenta) and BrpNc82 (green). Right: RIMHA−Znf SCs extracted by HDBSCAN in different colors. Colored lines indicate alpha shapes used for area determination. The AZ c.o.m. is indicated (x). (B) Number of localizations per RIMHA−Znf SC, SC area, and SC localization density in ctrl (filled boxes, n = 11,094 SCs from 18 NMJs in nine animals) and phtx (dashed boxes, n = 13,568 SCs from 19 NMJs in 12 animals) shown as box plots (horizontal lines show median, box boundaries show 25th and 75th percentiles, whiskers show 10th and 90th percentiles). (C) Number of RIMHA−Znf SCs per AZ, the total number of RIMHA−Znf localizations per AZ, the entire RIMHA−Znf area per AZ, and radial distance of RIMHA−Znf SCs per AZ in ctrl (n = 893 AZs from 18 NMJs in nine animals) and phtx (n = 963 AZs from 19 NMJs in 12 animals). (D) Representative, circular AZs (circularity ≥ 0.6) from a ctrl and a phtx type Ib bouton. Green lines indicate alpha shapes used for AZ area determination. Left: ctrl AZ from Fig 4C with gray dots indicating individual RIMHA−Znf localizations and black dots indicating SC c.o.m.s that do not belong to a supercluster (SpC). Right: two SpCs (shown in blue and yellow) were extracted by HDBSCAN from the c.o.m.s (red dots) of RIMHA−Znf SCs in the phtx AZ. Black dots represent SC c.o.m.s that are unclustered according to HDBSCAN analysis and gray dots show localizations of the corresponding SCs. Red x indicate c.o.m.s of SpCs. Dashed red line indicates the distance between a SC c.o.m. and the AZ c.o.m. (E) Percentage of SC c.o.m.s that are organized in SpCs per AZ for ctrl (n = 542 AZs from 18 NMJs in nine animals) and phtx (n = 545 AZs from 19 NMJs in 12 animals) shown as histograms. Dashed lines indicate mean, solid lines indicate median, 25th and 75th percentiles. Note that the 25th percentile and median fall together at zero in the left panel. (F) Number of SpCs per AZ. Median values, indicated in green, and statistical comparison refers to AZs with at least one SpC (ctrl: n = 247 AZs from 18 NMJs in nine animals; phtx: n = 290 AZs from 19 NMJs in 12 animals). Scale bars in (A) and (D) 100 nm.
- Figure S2. RIMHA−Znf is not recruited from the AZ vicinity during PHP.
(A) Contour plot of the median number of RIMHA−Znf clusters per dSTORM image (n = 18 NMJs) depending on the HDBSCAN parameters minimum samples and minimum cluster size. (B) Crosshairs indicate the parameter combination (5 and 20, respectively) used in (B). (B) Scatter plot of RIMHA−Znf localizations of a type Ib bouton in ctrl imaged as in Fig 4 but analyzed with a single-channel HDBSCAN algorithm (see the Material and Methods section). Colors indicate individual clusters with unclustered localizations in gray. Boxes in (B, C, D) mark corresponding positions and enlarged regions in (E, F). (B, C) BrpNc82 localizations of the same bouton as in (B) analyzed with a single-channel HDBSCAN algorithm. (B) Clustered and unclustered localizations displayed as in (B). (B, C, D) RIMHA−Znf localizations of the same bouton as in (B, C) analyzed with a second-level HDBSCAN applied only to the clustered (e.g., denoised) localizations from (B), resulting in RIMHA−Znf subclusters (SCs) that are assigned to the nearest BrpNc82 cluster by colors used in (C). (E, F) Scatter plots of a BrpNc82 cluster (gray) and assigned RIMHA−Znf SCs (different colors). Colored lines in (E) indicate alpha shapes for area determination. (F) Extrasynaptic RIMHA−Znf SCs (black), extrasynaptic RIMHA−Znf SCs > 50 and ≤ 400 nm distant from the AZ (e.g., AZ vicinity, blue), and intrasynaptic RIMHA−Znf SCs (magenta) are illustrated in (F). (G) Number of localizations per RIMHA−Znf SC, SC area, and SC localization density in extrasynaptic SCs, extrasynaptic SCs ≤ 400 nm distant from the AZ and intrasynaptic SCs for ctrl (filled bars, n = 8,697, 4,675, and 11,775 SCs, respectively, from 18 NMJs in nine animals) and phtx (dashed bars, n = 10,197, 4,931 and 14,081 SCs, respectively, from 19 NMJs in 12 animals). (H, I) Number of RIMHA−Znf SCs per AZ, RIMHA−Znf localizations per AZ, total RIMHA−Znf area per AZ (in (I)) and radial distance in ctrl (filled bars) and phtx (dashed bars) in extrasynaptic SCs ≤ 400 nm away from the AZ ((H), n = 525 and 532 circular AZs for ctrl and phtx, respectively) and intrasynaptic SCs (I, n = 892 and 961 AZs for ctrl and phtx, respectively). Scale bars in (B, C, D) 1 μm, in (E, F) 100 nm.
- Figure S3. Analysis of RIMHA−Znf superclusters.
(A) Cumulative plots show distance distributions between RIMHA−Znf SC c.o.m.s at circular AZs for ctrl (n = 93,552 distances from 5,876 SCs and 542 AZs) and phtx (n = 108,618 distances from 6,491 SCs and 545 AZs). Inset shows x and y axes from 0 to 100 nm and 0–8%, respectively. (B) Nearest neighbor distance (NND) and second to fifth neighbor distance (ND) of RIMHA−Znf SCs at circular AZs for ctrl (n = 5,876, 5,842, 5,770, 5,558, and 5,328 distances from 542, 525, 501, 448, and 402 AZs for NND and second to fifth ND, respectively) and phtx (n = 6,491, 6,463, 6,409, 6,253, and 6,108 distances from 545, 531, 513, 474, and 445 AZs for NND and second to fifth ND, respectively). (C) Number of SCs per supercluster (SpC). Median values are indicated in green (n = 713 and 823 SpCs for ctrl and phtx, respectively). (D) Distance between c.o.m.s of SCs and SpCs for both groups at AZs with at least one SpC.
- Figure S4. AZ changes in PHP deciphered by super-resolution imaging.
(A, B) Model of a type Ib AZ before ((A), ctrl) and after induction of PHP ((B), phtx) based on data acquired with dSTORM and HDBSCAN analysis. BrpNc82 (green) and RIMHA−Znf (magenta) were imaged as described in Mrestani et al (2021) and the present study and imaging data delivered numerical values for the model. Note the compaction of the entire AZ area and RIMHA−Znf SCs and RIMHA−Znf SpC formation during PHP. Brp SCs and their compaction are not shown for clarity. Scale bar, 100 nm.
Supplementary Materials
Table S6. dSTORM analysis of RIMHA−Znf and Brp.
Table S9. Primer sequences.
