MAPK3 at the Autism-Linked Human 16p11.2 Locus Influences Precise Synaptic Target Selection at Drosophila Larval Neuromuscular Junctions

Abstract

Proper synaptic function in neural circuits requires precise pairings between correct pre- and post-synaptic partners. Errors in this process may underlie development of neuropsychiatric disorders, such as autism spectrum disorder (ASD). Development of ASD can be influenced by genetic factors, including copy number variations (CNVs). In this study, we focused on a CNV occurring at the 16p11.2 locus in the human genome and investigated potential defects in synaptic connectivity caused by reduced activities of genes located in this region at Drosophila larval neuromuscular junctions, a well-established model synapse with stereotypic synaptic structures. A mutation of rolled, a Drosophila homolog of human mitogen-activated protein kinase 3 (MAPK3) at the 16p11.2 locus, caused ectopic innervation of axonal branches and their abnormal defasciculation. The specificity of these phenotypes was confirmed by expression of wild-type rolled in the mutant background. Albeit to a lesser extent, we also observed ectopic innervation patterns in mutants defective in Cdk2, Gα_q, and Gp93, all of which were expected to interact with Rolled MAPK3. A further genetic analysis in double heterozygous combinations revealed a synergistic interaction between rolled and Gp93. In addition, results from RT-qPCR analyses indicated consistently reduced rolled mRNA levels in Cdk2, Gα_q, and Gp93 mutants. Taken together, these data suggest a central role of MAPK3 in regulating the precise targeting of presynaptic axons to proper postsynaptic targets, a critical step that may be altered significantly in ASD.

Keywords: 16p11.2; Drosophila; MAPK3; autism; copy number variations.

Fig. 1. Aberrant innervation patterns of type III axons and premature axonal defasciculation in rolled MAPK3 mutants

(A) A mutation of rolled (or rl) cause morphologic changes at Drosophila larval NMJs, including ectopic targeting of type III axons to muscle 13 (arrow) and premature defasciculation of axon bundles at the boundary of muscles 13 and 12 (arrowheads), both of which are rarely detected in WT animals. Scale bar, 20 μm. (B, C) The frequency of ectopic type III axons (B) and premature axonal defasciculation (C) is shown for WT and rl larvae. The number of NMJs examined: WT, 102 and rl, 79. ^***, P < 0.001, Fisher’s exact test between WT and rl.

Fig. 2. Phenotypic rescue of axon targeting and defasciculation errors by expression of wild-type rolled in presynaptic motor neurons and postsynaptic muscles

(A) The rl mutant phenotypes, including ectopic targeting of type III axons to muscle 13 (arrow) and premature axonal defasciculation (arrowhead), are rescued by neuron- and muscle-specific expression of WT Rolled (rl⁺). The elav^C155- and mef2-GAL4 lines were used for neuron- and muscle-specific expression of a transgenic rl⁺ construct under the influence of UAS element, respectively. Scale bar, 20 μm. (B, C) The fraction of ectopic type III axons (B) and premature axonal defasciculation (C) is shown for each genotype. The fraction of mutant phenotypes observed in animals containing only a transgenic rl⁺ construct in the mutant background (rl, UAS-rl⁺) serves as a control, along with WT. The data from WT in Fig. 1 are duplicated for comparisons. The number of NMJs examined for each genotype: WT, 102; rl, UAS-rl⁺, (no GAL4) 33; elav^C155-GAL4 → UAS-rl⁺, 57; mef2-GAL4 → UAS-rl⁺, 55. ^***, P < 0.001 for Fisher’s exact test between WT and the genotype indicated. ⁺⁺, P < 0.01 and, ⁺, P < 0.05 for Fisher’s exact test between rl, UAS-rl⁺ (no GAL4) and the genotype indicated.

Fig. 3. Aberrant innervation patterns of type III axons observed in mutants defective in MAPK3-associated proteins

(A) Ectopic targeting of type III axons to muscle 13 (arrows), a phenotype observed in rl mutants, is also detected in mutants defective in Cdk2, Gα_q and Gp93. Scale bar, 20 μm. (B) The fraction of aberrant type III innervation patterns is shown for each genotype. The data from WT and rl groups in Fig. 1B are duplicated for comparisons. The number of NMJs examined: WT, 102; rl, 79; Cdk2, 35; Gα_q, 35; Gp93, 40. ^***, P < 0.001 and ^*P < 0.05 for Fisher’s exact test between WT and the genotype indicated.

Fig. 4. Genetic interactions between rolled and genes encoding MAPK3-associated proteins

(A) Double heterozygous larvae for both rl and one of the associated genes among Cdk2, Gα_q and Gp93 display frequently mistargeted type III axons in muscle 13 (arrows). Scale bar, 20 μm. (B) The fraction of aberrant type III innervation patterns is shown for single and double heterozygous combinations indicated. Upper (black) and lower (light gray) dashed lines represent the fraction observed in rl homozygous mutants and WT, respectively. The number of NMJs examined: rl^−/+, 31; Cdk2^−/+, 35; rl^−/+; Cdk2^−/+, 20; Gα_q^−/+, 36; rl^−/+, Cdk2^−/+, 24; Gp93^−/+, 34; rl^−/+; Gp93^−/+, 20. ^*, P < 0.05 for Fisher’s exact test between rl single heterozygous (rl^−/+) and corresponding double heterozygous mutants indicated. ⁺⁺⁺, P < 0.001 and ⁺, P < 0.05 for Fisher’s exact test between single and double heterozygous pairs indicated.

Fig. 5. Transcriptional regulation of rolled MAPK3 and fasII in Cdk2, Gα_q and Gp93 mutants.

(A–C) The relative levels of mRNA estimated from qRT-PCR analyses with two independent rl primers (A, B) and a single fasII primer (C) are shown for each genotype. The level of mRNA in WT animals is used as a control to normalize the values measured from mutants. The levels of each mRNA are measured from at least three independent sets of triplicate samples. ^***, P < 0.001, ^**, P < 0.01 and ^*, P < 0.05 for one-way ANOVA between WT and the genotype indicated, followed by Tukey’s post-hoc test for pair-wise comparisons.