Negative regulation of active zone assembly by a newly identified SR protein kinase

Abstract

Presynaptic, electron-dense, cytoplasmic protrusions such as the T-bar (Drosophila) or ribbon (vertebrates) are believed to facilitate vesicle movement to the active zone (AZ) of synapses throughout the nervous system. The molecular composition of these structures including the T-bar and ribbon are largely unknown, as are the mechanisms that specify their synapse-specific assembly and distribution. In a large-scale, forward genetic screen, we have identified a mutation termed air traffic controller (atc) that causes T-bar-like protein aggregates to form abnormally in motoneuron axons. This mutation disrupts a gene that encodes for a serine-arginine protein kinase (SRPK79D). This mutant phenotype is specific to SRPK79D and is not secondary to impaired kinesin-dependent axonal transport. The srpk79D gene is neuronally expressed, and transgenic rescue experiments are consistent with SRPK79D kinase activity being necessary in neurons. The SRPK79D protein colocalizes with the T-bar-associated protein Bruchpilot (Brp) in both the axon and synapse. We propose that SRPK79D is a novel T-bar-associated protein kinase that represses T-bar assembly in peripheral axons, and that SRPK79D-dependent repression must be relieved to facilitate site-specific AZ assembly. Consistent with this model, overexpression of SRPK79D disrupts AZ-specific Brp organization and significantly impairs presynaptic neurotransmitter release. These data identify a novel AZ-associated protein kinase and reveal a new mechanism of negative regulation involved in AZ assembly. This mechanism could contribute to the speed and specificity with which AZs are assembled throughout the nervous system.

Figure 1. Brp accumulates in srpk79D mutant nerves.

(A) The srpk79D gene region is shown, including the srpk79D^atc transposon insertion site and deleted regions in mutants used for genetic analyses (black bars). Gene loci for srpk79D and the adjacent Csp gene are shown in blue. (B–J) Immunofluorescence images of larval nerves demonstrating large Brp accumulations in srpk79D loss-of-function mutants. Each image shows a section of a single larval nerve, photographed at the same relative position approximately 100 µm from where the nerve exits the CNS. Each nerve contains approximately 85 total axons, including approximately 35 motor axons. Images are shown at two exposures. Images in (B, E, and H) and (C, F, and I) were taken at an exposure in which small, infrequent anti-Brp puncta could be resolved in wild-type nerves. This resulted in overexposure of the Brp puncta in srpk79D mutant nerves. Images (D, G, and J) are identical to (B, E, and H, and C, F, and I) except taken at a lower exposure such that no puncta are found in wild type, and the puncta intensities are not saturated in the srpk79D mutant. (K) Total Brp fluorescence integrated over the nerve area is dramatically increased with srpk79D disruption, whereas loss of Csp does not increase nerve Brp levels. Each bar graph represents data collected from a total of 30 nerves from 12 different larvae. (L) Cumulative frequency plots of individual Brp punctum fluorescence intensities are shifted toward larger values with srpk79D loss of function (gray and blue lines representing srpk79D^atc and srpk79D^atc/Df, respectively, are shifted to the far right whereas other genotypes are clustered to the left). Each curve represents data collected from a total of 30 nerves from 12 different larvae. Sample size for wild type, srpk79D^atc/+, srpk79D^atc, srpk79D^atc/Df, and srpk79D^atc/Csp^X1 = 479, 1,715, 3,387, 3,718, and 2,714, respectively. Significance is indicated according to the following: * = p<0.05, ** = p<0.01, *** = p<0.001, and ns = not significant; Student t-test. Scale bar indicates 10 µm. Error bars indicate ±SEM. au = arbitrary units; Brp = anti-Bruchpilot; HRP = anti-horseradish peroxidase.

Figure 2. Synaptic Brp deficit in srpk79D mutants.

(A–D) Immunofluorescence images of wild-type and srpk79D^atc mutant NMJ reveal a synaptic Brp deficit in srpk79D^atc mutants. NMJs are stained with anti-HRP (red) to label the presynaptic membrane and anti-Brp (green). Larvae in wild type and srpk79D^atc were stained in the same reaction tube and imaged identically. (E) The total synaptic Brp fluorescence is decreased in srpk79D^atc mutants. Each bar graph represents data collected from a total of 30 synapses from nine different larvae. (F) Cumulative frequency plots of synaptic Brp puncta fluorescence intensities are shifted toward smaller values with srpk79D loss of function. Each curve represents data collected from 30 synapses from nine different larvae. n for wild type, srpk79D^atc, and srpk79D^atc/Df = 6,554, 4,713, and 5,952, respectively. au = arbitrary units; HRP = anti-horseradish peroxidase. (G and H) Disruption of srpk79D affects neither synaptic Brp puncta number (G) nor synaptic bouton number (H). Each bar graph in (G) and (H) represents data collected from a total of 38 synapses taken from 15 different larvae. Scale bar indicates 10 µm. Significance is indicated according to the following: *** = p<0.001 and ns = not significant; Student t-test. Error bars indicate ±SEM.

Figure 3. SRPK79D functions in neurons to prevent Brp accumulation.

(A and B) Whole-mount in situ hybridizations demonstrate that srpk79D is widely expressed but is enriched in the CNS. Anterior is to the left. (C and D) Immunofluorescence images of a control nerve (C) and a nerve expressing srpk79D-specific double-stranded RNA (dsRNA) (UAS-srpk79D^RNAi) in neurons (D). Expression of dsRNA causes accumulation of Brp puncta. (E and F) Total Brp fluorescence is increased and cumulative frequency plots are shifted toward larger values when srpk79D^RNAi is expressed in neurons, but not when it is expressed in glia using the glia-specific Repo-GAL4 driver. Each bar graph and curve in (E) and (F) represents data collected from a total of 30 nerves from nine different larvae. In (F), n for C155/+, C155/+;;UAS-srpk79DRNAi/+, Repo/+, and Repo/UAS-srpk79D^RNAi = 1,034, 2,451, 893, and 862, respectively. (G–J) Expression of a Venus-tagged srpk79D transgene (UAS-v-srpk79D-rd*) rescues Brp accumulations in srpk79D mutant nerves. (G and H) Nerves are stained with anti-Brp and imaged identically. Brp accumulations are present in the srpk79D mutant (G), and these accumulations are less intense following rescue of the srpk79D mutant with the UAS-srpk79D transgene (H). (I and J) Quantification of Brp fluorescence intensity (I) and puncta intensities (J), comparing control (C155/+), srpk79D mutant (C155/+;srpk79D^atc), and rescue animals (C155/+; UAS-v-srpk79D-rd*(28)/+;srpk79D^atc). Each bar graph and curve in (I) and (J) represent data collected from a total of 36 nerves from nine different larvae. In (J), n for C155/+, C155/+;;srpk79D^atc, and C155/+;UAS-v-srpk79D-rd*(28)/+; srpk79D^atc = 2,452, 2,696, and 3,645, respectively. Scale bar = 10 µm. Significance is indicated according to the following: *** = p<0.001 and ns = not significant; Student t-test. Error bars indicate ±SEM. au = arbitrary units.

Figure 4. Axonal transport is intact in srpk79D mutants.

(A–L) Immunofluorescence images of wild-type and srpk79D mutant nerves demonstrating the distribution of Bruchpilot (Brp; [A and B]), the synaptic vesicle proteins Synaptotagmin 1 (Syt; [C and D]) and Cysteine String Protein (CSP; [E and F]), mitochondria (mitoGFP; [G and H]), the peri-AZ protein Dap160/Intersectin (Dap160; [I and J]), and the AZ protein Liprin-alpha (Lip-α; [K and L]). Image exposures were selected such that small puncta could be visualized in the wild-type controls, corresponding to a “high exposure” in Figure 1. Wild type and srpk79D mutants were stained in the same reaction tube and imaged identically for each antibody. (M and N) Reducing kinesin heavy chain (Khc8/+) or immaculate connections (imac170/+) in the srpk79D mutant background does not enhance the srpk79D mutant phenotype. Each bar graph in (M) and curve in (N) represents data collected from a total of 30 nerves from nine different larvae. In (N), n = 1,040, 2,928, 2,560, and 3,240 for wild type, srpk79Datc, Khc8/+; srpk79D^atc, and imac170/+; srpk79D^atc, respectively. Scale bar indicates 10 µm. Significance is indicated according to the following: *** = p<0.001 and ns = not significant; Student t-test. Error bars indicate ±SEM. au = arbitrary units.

Figure 5. Comparison of synaptic and axonal Brp assemblies in wild-type and srpk79D mutant animals.

(A–D) Immunofluorescence images of wild-type and srpk79D mutant nerves and synapses. Insets in (C) and (D) show the shape of the nerve terminal arborization at lower magnification based upon costaining with anti-Hrp. (E) Cumulative frequency plots of individual Brp puncta fluorescence intensities. Each curve represents data collected from a total of 18 nerves or synapses from nine larvae. Arrowheads on the x-axis in (E) indicate the average maximum punctum fluorescence intensity for each genotype (indicated by arrowhead color). Scale bars indicate 10 µm. Significance is indicated according to the following: *** = p<0.001; Student t-test. Error bars indicate ±SEM.

Figure 6. T-bar superassemblies are found in srpk79D mutant axons.

(A and B) Electron micrographs of srpk79D mutant motor axons showing large, electron-dense structures that are never found in wild-type axons. (C) Magnified image of region boxed in (B) highlighting an accumulation that is particularly reminiscent of a superassembly of T-bars. (D) A wild-type synaptic T-bar at the same magnification for comparison to the image in (C). Number of animals sectioned: wild type = 5; mutant = 9. Number of sections analyzed: wild type = 150; srpk79D^ATC = 325. All panels are sized relative to the scale bar shown in (A), as follows: 160 nm for (A and B); 80 nm for (C and D).

Figure 7. Brp accumulation in srpk79D mutants is not due to increased Brp expression.

(A and B) Immunofluorescence images of wild-/type and srpk79D mutant nerves. (C) Similar Brp accumulations appear when GFP-Brp is overexpressed in motor neurons using the motoneuron-specific GAL4 driver OK371-GAL4. (D) Brp accumulations persist in srpk79D mutants when one copy of brp is deleted by placing a heterozygous brp⁶⁹/+ mutation in the homozygous srpk79D mutant background. (E and F) Immunofluorescence images of wild-type and srpk79D mutant muscle 4 NMJ stained with anti-Brp. (G) Synaptic Brp is increased following GFP-Brp overexpression. (H) Removal of one copy of brp (brp⁶⁹/+) in a homozygous srpk79D mutant background causes a further decrease in synaptic Brp levels compared to srpk79D mutants alone, but does not cause altered distribution of Brp immunoreactivity. (I) Quantifications of total Brp fluorescence for each indicated genotype normalized to wild type (wt). Each bar graph represents data collected from a total of 29 synapses from 14 different animals. (J) Axonal Brp fluorescence is increased in srpk79D^atc/+ heterozygotes (OK371/+; srpk79D^atc/+) and when Brp is overexpressed in motor neurons (OK371/+; UAS-g-brp). An additive effect is seen when these two perturbations are combined (OK371/+; UAS-brp/srpk79D^atc). In all cases, total synaptic anti-Brp fluorescence is significantly less than that seen in srpk79D mutants (srpk79D^atc). Each bar graph represents data collected from a total of 32 synapses from eight different larvae. Scale bars indicate 10 µm. Significance is indicated according to the following: *** = p<0.001 and ns = not significant; Student t-test. Error bars indicate ±SEM.

Figure 8. SRPK79D kinase activity is necessary, and the SRPK79D N-terminus directs targeting.

(A) A schematic of two previously cloned srpk79D transcripts (srpk79D-rb and srpk79D-rd) 62. srpk79D-rd results from read-through of predicted splice sites in exons 5 and 8. We generated the transcript conforming to the predicted splicing pattern of exons 5 and 8 is also shown (srpk79D-rd*). We have since confirmed the existence of this transcript by reverse-transcriptase PCR. (B) A schematic of yeast Sky1p and human SRPK1 protein domain structures as well as domain structures for the predicted srpk79D protein products (SRPK79D-RB, SRPK79D-RD, and SRPK79D-RD*). Alternative exon usage results in a unique SRPK79D-RB N-terminus (light gray), whereas SRPK79D-RD and SRPK79D-RD* use the same N-terminus (dark gray). The splicing pattern employed in SRPK79D-RD leads to a truncated kinase domain relative to SRPK79D-RB* and SRPK79D-RD. SRPK79D-RD*KD contains a missense mutation at position 376 (K376M) that targets the predicted ATP binding site. Numbers indicate the percent amino acid identity/similarity of the SRPK79D kinase domain to that of human SRPK1. D.m. = Drosophila melanogaster; HRP = anti-horseradish peroxidase; H.s. = Homo sapiens; S.c. = Saccharomyces cerevisiae. (C–F and I–J) Immunofluorescence images of synapses and nerves, respectively, demonstrate near-perfect colocalization between Brp and Venus-SRPK79D-RD* in both the NMJ (C–F) and in the axons (G–J). (K) Table indicating Brp colocalization and ability to rescue axon Brp accumulation phenotype of the srpk79D gene products indicated in (A) and (B). Scale bars indicate 10 µm. Error bars indicate ±SEM. See also Supplemental Discussion (Text S1) relevant to this annotation.

Figure 9. SRPK79D overexpression disrupts synaptic Brp and impairs synapse function.

(A–D) Representative muscle 4 NMJ and individual bouton from control (C155/+) and SRPK79D-RD*-overexpressing larvae demonstrating diffuse Brp staining and reduced total Brp fluorescence. Image offset and gain in (A) and (B) are identical. (C) Example of type-1b boutons from control animals demonstrating typical punctate anti-Brp staining (green) and anti-HRP staining (red) to elucidate the nerve terminal membrane. (D) Examples of diffuse Brp staining observed at type-1b boutons within the NMJ of an SRPK79D overexpressing animal. Image offset and gain in (C) and (D) are identical. (E) SRPK79D-RD* overexpression causes a decrease in total synaptic Brp fluorescence. (F) Overexpression of SRPK79D-RD* causes a highly significant decrease in EPSP amplitude. There is a trend toward an increase in the average amplitudes of spontaneous miniature events comparing SRPK79D-RD*-overexpressing animals to wild type (p = 0.06); representative mEPSPs are shown. (G) Quantification of average EPSP amplitude and quantal content (QC) in SRPK79D-RD*-overexpressing larvae show a greater than 50% decreases in both measures relative to control. C155/+ and C155/+;UAS-v-srpk79D-rd*(F)/+ bar graphs represent data collected from a total of 15 synapses from six different larvae and 12 synapses from five different larvae, respectively. Scale bar in (A) indicates 10 µm, and in (C) indicates 5 µm. Significance is indicated according to the following: *** = p<0.001; Student t-test. Error bars indicate ±SEM. HR = anti-horseradish peroxidase.