Regulation of Fasciclin II and synaptic terminal development by the splicing factor beag

Abstract

Pre-mRNA alternative splicing is an important mechanism for the generation of synaptic protein diversity, but few factors governing this process have been identified. From a screen for Drosophila mutants with aberrant synaptic development, we identified beag, a mutant with fewer synaptic boutons and decreased neurotransmitter release. Beag encodes a spliceosomal protein similar to splicing factors in humans and Caenorhabditis elegans. We find that both beag mutants and mutants of an interacting gene dsmu1 have changes in the synaptic levels of specific splice isoforms of Fasciclin II (FasII), the Drosophila ortholog of neural cell adhesion molecule. We show that restoration of one splice isoform of FasII can rescue synaptic morphology in beag mutants while expression of other isoforms cannot. We further demonstrate that this FasII isoform has unique functions in synaptic development independent of transsynaptic adhesion. beag and dsmu1 mutants demonstrate an essential role for these previously uncharacterized splicing factors in the regulation of synapse development and function.

Figure 1.

beag mutants have decreased NMJ synaptic bouton number and increased synaptic bouton area. A, B, Representative images of NMJ synaptic terminals at muscle 4 of segment A3 of third instar wild-type and beag^1/1 mutant larvae stained with anti-CSP (green) to label the presynapse and anti-HRP (red) to label the neuronal membrane. Scale bar, 20 μm. beag mutants have a reduction in synaptic bouton number (C), an increase in average synaptic bouton area (D), and an increase in total presynaptic area (E). Expression of a genomic Beag transgene rescues beag mutant bouton number (C), beag mutant bouton area (D), and total presynaptic area (E). C, Quantification of bouton number normalized to muscle surface area. D, Quantification of average bouton area. E, Quantification of total presynaptic terminal area normalized to muscle surface area. Live imaging of developing NMJ synaptic terminals labeled with CD8-GFP-SH in wild-type (F) and beag¹ mutant (G) larvae at the second larval instar and after 48 h at the third larval instar. A new distal bouton addition is indicated by the arrow. Scale bar, 8 μm. H, I, Flag-tagged genomic Beag localizes to the nucleus (H) and colocalizes with Elav in neurons (I, Beag green, Elav red). Scale bar, 10 μm. Error bars indicate SEM. **p < 0.01, ***p < 0.001, significance calculated versus wild-type control except where indicated.

Figure 2.

beag and dsmu-1 mutants have similar NMJ synaptic morphology phenotypes. dsmu1^pBacf03090/Df mutants (B) have decreased synaptic bouton number, no change in total presynaptic area, and increased synaptic bouton area compared with wild type (A). These phenotypes are similar to those seen in beag^EP3260/Df mutants (C). beag^EP/Df dsmu1^pBac/Df double mutants have similar phenotypes to beag and dsmu1 single mutants (D). beag^1/+, dsmu1^Df/+ trans-heterozygotes have a similar decrease in bouton number as either heterozygote alone (E). A–D, Representative images are NMJ synaptic terminals at muscle 4 of segment A3 stained with anti-CSP (green) to label the presynapse and anti-HRP (red) to label the neuronal membrane. Scale bar: (in A) A–D, 20 μm. E, Quantification of bouton number normalized to muscle surface area. F, Quantification of average bouton area. G, Quantification of total presynaptic area normalized to muscle surface area. H, I, UAS-RFP-tagged Dsmu1 (H) expressed in a subset of motor neurons with OK319-Gal4 colocalizes with the neural nuclear protein Elav in the ventral nerve cord and is also present in the cytoplasm (I, Dsmu1 red, Elav green). Scale bar: (in H) H, I, 10 μm. Error bars indicate SEM. *p < 0.05, **p < 0.01, ***p < 0.001, significance calculated versus wild-type control except where indicated.

Figure 3.

Beag and Dsmu1 are required in neurons for normal NMJ morphology. Expression of transgenic Beag in motor neurons with OK6-Gal4 (C) or all neural cells with C155-Gal4 (D) in beag¹/Df (B) mutants rescues synaptic bouton number to wild-type levels (A). Expression of transgenic Dsmu1 in motor neurons with OK6-Gal4 (H) or in neural cells with C155-Gal4 (F) rescues the decrease in synaptic bouton number in dsmu1^pBacf03090/Df mutants (E) to wild-type levels. Neuronal expression of Dsmu1 with C155-Gal4 in beag¹/Df mutants (H) does not rescue synaptic bouton number, while neural expression of Beag with C155-Gal4 in dsmu1^pBacf03090/Df mutants (G) partially rescues bouton number. Quantification of synaptic bouton numbers divided by muscle surface area (H). Representative images are NMJ synaptic terminals at muscle 4 of segment A3 stained with anti-CSP (green) to label the presynapse and anti-HRP (red) to label the neuronal membrane. Error bars indicate SEM. *p < 0.05, **p < 0.01, ***p < 0.001, significance calculated versus wild-type control except where indicated. Scale bar, 20 μm.

Figure 4.

beag mutants have decreased neurotransmitter release. A, B, Representative traces recorded from muscle 6 of segment A3 in wild-type and beag^1/EP3260 larvae. beag^1/EP3260 mutants have decreased EJP amplitude (C) but normal mEJP amplitude (D) and frequency (E). This results in a decrease in quantal content in beag mutants (F). Expression of transgenic Beag in motor neurons with OK6-Gal4 in Beag^1/EP3260 mutants restores EJP amplitude (C) and quantal content (F). dsmu1^pBacf03090/Df mutants also have decreased EJP amplitude (C), similar to beag mutants. The number of active zones labeled by anti-Bruchpilot is unchanged in beag¹/Df mutants (H) compared to wild type (G). Quantification of Bruchpilot punctae per NMJ terminal (I). Error bars indicate SEM. **p < 0.01, ***p < 0.001, significance calculated versus wild-type control.

Figure 5.

FasII levels are altered in an isoform-specific manner in beag mutants. A, FasII has four splice isoforms, A-PEST+, A-PEST−, B, and C, that all include the first seven exons (black) but differ in the inclusion (red) or exclusion (gray) of exons at the 3′ end of the gene. The monoclonal anti-FasII total antibody (34B3) recognizes an epitope in the extracellular domain of all four isoforms. The monoclonal anti-FasII-A antibody (1D4) recognizes an epitope in the intracellular domain of FasII-A-PEST+ and FasII-A-PEST−. Positions of the primers used to amplify both FasII-A isoforms, the C isoform, or total FasII for qRT-PCR are indicated by arrows. B, C, NMJs at muscle 4 of segment A3 in third instar larvae were stained with the monoclonal antibody 1D4 or the monoclonal antibody 34B3. The level of 1D4 staining in beag¹/Df mutant NMJs is reduced compared with wild type (B). In contrast, there is no difference between wild-type and beag NMJs when stained with 34B3 (C). D, Quantification of relative staining intensity with 1D4 and 34B3 at wild-type, beag, and dsmu1 mutant synapses. Expression of a genomic Beag transgene in beag¹/Df mutants rescues synaptic FasII-A staining intensity to wild-type levels. dsmu1 mutants also have decreased synaptic 1D4 staining. E, qRT-PCR measurement of the relative abundance of both FasII-A isoforms, FasII-C, and total FasII. FasII-A mRNA is decreased in beag mutants, while FasII-C mRNA is slightly increased. No change was observed in total FasII mRNA abundance in beag mutants. F, Measurement of FasII-A-PEST+ and FasII-A-PEST− mRNA levels by semiquantitative multiplex PCR in wild-type and beag brains demonstrates a decrease in the ratio of FasII-A-PEST+ to FasII-A-PEST− in beag mutants. G, Normalized ratio of FasII-A-PEST+ versus FasII-A-PEST− in wild-type and beag mutants. Error bars indicate SEM. *p < 0.05, ***p < 0.001, significance calculated versus wild-type control.

Figure 6.

fasII and beag genetically interact. Neither beag^1/+ heterozygotes nor fasII^e76/+ heterozygotes have a significant change in bouton number compared with wild-type (A,B,C,H); however, beag^1/+, fasII^e76/+ trans-heterozygotes have fewer boutons than controls (D, H). Removal of one copy (fasII^e76/+) or two copies (fasII^e76/e76) of fasII does not significantly enhance beag^1/Df mutants and is similar to fasII^e76/e76 (E–H). These results demonstrate that fasII and beag are components of a common genetic pathway in the regulation of NMJ development. H, Quantification of bouton number normalized to muscle surface area. Representative images NMJ synaptic terminals at muscle 4 of segment A3 stained with anti- CSP (green) to label the presynapse and anti-HRP (red) to label the neuronal membrane. Error bars indicate SEM. ***p < 0.001, significance calculated versus wild-type control. Scale bar, 20 μm.

Figure 7.

FasII-A-PEST+ and FasII-A-PEST− are necessary and sufficient for normal NMJ growth. Ubiquitous (Da-Gal4) RNAi-induced knockdown of all FasII isoforms with UAS-FasIIRNAi-Total (B) or only FasII-A-PEST+ and FasII-A-PEST− with UAS-FasIIRNAi-A (C) causes a decrease in synaptic bouton number compared with wild type (A). The decrease in synaptic bouton number due to RNAi inhibition of transmembrane FasII isoforms is not altered by coexpression of FasII-C (D), but is rescued by coexpression of RNAi resistant forms of FasII-A-PEST+ (E) or FasII-A-PEST− (F). Similarly, the decrease in bouton number in fasII^e76/Df mutants (G) is not rescued by expression of FasII-C in motor neurons (OK6-Gal4) (H), while expression of FasII-A-PEST+ or FasII-A-PEST− fully restores the synaptic bouton number of these mutants to wild-type levels (I, J). K, Quantification of synaptic bouton number normalized to muscle surface area. Representative images of NMJ synaptic terminals at muscle 4 of segment A3 stained with anti-CSP (green) to label the presynapse and anti-HRP (red) to label the neuronal membrane. Error bars indicate SEM. *p < 0.05, ***p < 0.001, significance calculated versus wild-type control. Scale bar, 20 μm.

Figure 8.

Overexpression of FasII-A-PEST+ but not the other FasII isoforms rescues beag NMJ morphology. Expression of FasII-A-PEST+ or FasII-C in motor neurons with OK6-Gal4 in a wild-type background (A) does not affect synaptic bouton number, while expression of FasII-A-PEST− causes an increase in synaptic bouton number (G). Expression of FasII-C (C) or FasII-A-PEST− (D) in motor neurons (OK6-Gal4) of beag¹/Df mutants (B) does not rescue synaptic bouton number, while expression of FasII-A-PEST− in motor neurons restores normal synaptic bouton number in beag¹/Df mutants (E). Expression of a fusion protein in which the extracellular and transmembrane domains of CD8 are fused to the intracellular domain of FasII-A-PEST+ (CD8-FasII-A-PEST+Intra) in motor neurons with OK6-Gal4 has no effect on synaptic bouton number in a wild-type background (G); however, expression in beag¹/Df mutants fully restores synaptic bouton number (F). Expression of FasII-A-PEST+AAE in motor neurons with OK6-Gal4 in a wild-type background causes an increase in synaptic bouton number and expression in beag mutants rescues synaptic bouton number (G). Quantification of bouton number normalized to muscle surface area (G). H, Motor neuron expression of FasII-A-PEST+ with OK6-Gal4 also rescues the average bouton area of beag¹/Df mutants. Representative images of NMJ synaptic terminals at muscle 4 of segment A3 stained with anti-CSP (green) to label the presynapse and anti-HRP (red) to label the neuronal membrane. Error bars indicate SEM. **p < 0.01, ***p < 0.001, significance calculated versus wild-type controls. Scale bar, 20 μm.