Missense variants in DPYSL5 cause a neurodevelopmental disorder with corpus callosum agenesis and cerebellar abnormalities

Abstract

The collapsin response mediator protein (CRMP) family proteins are intracellular mediators of neurotrophic factors regulating neurite structure/spine formation and are essential for dendrite patterning and directional axonal pathfinding during brain developmental processes. Among this family, CRMP5/DPYSL5 plays a significant role in neuronal migration, axonal guidance, dendrite outgrowth, and synapse formation by interacting with microtubules. Here, we report the identification of missense mutations in DPYSL5 in nine individuals with brain malformations, including corpus callosum agenesis and/or posterior fossa abnormalities, associated with variable degrees of intellectual disability. A recurrent de novo p.Glu41Lys variant was found in eight unrelated patients, and a p.Gly47Arg variant was identified in one individual from the first family reported with Ritscher-Schinzel syndrome. Functional analyses of the two missense mutations revealed impaired dendritic outgrowth processes in young developing hippocampal primary neuronal cultures. We further demonstrated that these mutations, both located in the same loop on the surface of DPYSL5 monomers and oligomers, reduced the interaction of DPYSL5 with neuronal cytoskeleton-associated proteins MAP2 and βIII-tubulin. Our findings collectively indicate that the p.Glu41Lys and p.Gly47Arg variants impair DPYSL5 function on dendritic outgrowth regulation by preventing the formation of the ternary complex with MAP2 and βIII-tubulin, ultimately leading to abnormal brain development. This study adds DPYSL5 to the list of genes implicated in brain malformation and in neurodevelopmental disorders.

Keywords: DPYSL5; brain malformation; corpus callosum agenesis; de novo missense variants; dendrite branching; neurodevelopmental disorder; primary neuronal cultures.

Figure 1

Neuro-anatomical and molecular analyses of DPYSL5 mutations (A) Midsagittal brain MRIs of three individuals carrying a de novo p.Glu41Lys heterozygous variant of DPYSL5 revealing complete agenesis of corpus callosum (ACC) and mild cerebellar hypoplasia (individual 1), total ACC with normal cerebellum size but with mild mega cisterna magna (individual 2), and total ACC and mild hypoplasia of the inferior cerebellar vermis (individual 3). (B) Schematic representation of DPYSL5 transcript and protein and localization of the two variants p.Glu41Lys (E41K) and p.Gly47Arg (G47R). (C) Homo-tetramer model of human DPYSL5. The four subunits are shown in ribbon representation and colored differently (green, blue, brown, and red). In one chain (blue), the location of the mutation sites is labeled and indicated by gray spheres. (D) Close-up view of the mutation site. Mutated residues are shown in stick representation. The crystal structure of human DPYSL5 (4b91) is shown in blue. The modeled structures of E41K and G47R are shown in gold and coral, respectively.

Figure 2

Functional analyses of the impact of DPYSL5 variants on neuronal development (A) Confocal microscopy images of neurons at 6 days of in vitro culture (D.I.V.) transfected at 2 D.I.V. with pEGFP (GFP), pEGFP-DPYSL5 (WT), pEGFP-DPYSL5 p.Glu41Lys (E41K), and pEGFP-DPYSL5 p.Gly47Arg (G47R) and labeled with anti-Tau (axon) and anti-MAP2 (dendrites) antibodies. Scale bar, 50 μm. (B) Graphical representations of the axon and dendrite length measures in dot plots, each plot corresponding to one neuron for which the axonal length and the total dendritic length were measured. A one-way ANOVA parametric test with multiple comparisons (Tukey’s post hoc) was performed and found to be non-significant for the axon length (p = 0.2272; F = 1.488; R-squared = 0.07148) and significant for total dendritic length (p = 0.0003; F = 7.377; R-squared = 0.2762). Three independent transfections including 15 neurons “pEGFP” (solid green circle), 15 neurons “EGFP-DPYSL5 WT” (pink diamond), 16 neurons “GFP-DPYSL5 E41K” (purple triangle), and 16 neurons “EGFP-DPYSL5 G47R” (red triangle). ^∗p = 0.0201 (WT versus G47R) or p = 0.0121 (WT versus E41K); ^∗∗∗p = 0.0001 (GFP versus WT). Data are shown as mean ± SEM.

Figure 3

The p.Glu41Lys and p.Gly47Arg variants impair the physical interaction of DPYSL5 with the neuronal cytoskeleton-associated proteins MAP2 and tubulin βIII (A) Representative immunoblots showing GFP-DPYSL5 or GFP immunoprecipitation and co-immunoprecipitation of DPYSL5-associated proteins DPYSL2, endogenous DPYSL5 (indicated with an arrow), βIII-tubulin, and MAP2 from CAD cells transfected with the indicated DPYSL5 mutants or GFP as a negative control. (B) Bar graph with scatterplots showing relative binding of the indicated proteins to GFP-DPYSL5. Both E41K and G47R mutations decreased binding to βIII-tubulin and MAP2. DPYSL5-E41K showed decreased binding to endogenous DPYSL5 and DPYSL5-G47R had increased binding to DPYSL2 compared with wild type (WT). ^∗p < 0.05, Kruskal-Wallis test with Dunn’s post hoc. Data are shown as mean ± SEM (n = 4).