Mutations in ANKLE2, a ZIKA Virus Target, Disrupt an Asymmetric Cell Division Pathway in Drosophila Neuroblasts to Cause Microcephaly

Abstract

The apical Par complex, which contains atypical protein kinase C (aPKC), Bazooka (Par-3), and Par-6, is required for establishing polarity during asymmetric division of neuroblasts in Drosophila, and its activity depends on L(2)gl. We show that loss of Ankle2, a protein associated with microcephaly in humans and known to interact with Zika protein NS4A, reduces brain volume in flies and impacts the function of the Par complex. Reducing Ankle2 levels disrupts endoplasmic reticulum (ER) and nuclear envelope morphology, releasing the kinase Ballchen-VRK1 into the cytosol. These defects are associated with reduced phosphorylation of aPKC, disruption of Par-complex localization, and spindle alignment defects. Importantly, removal of one copy of ballchen or l(2)gl suppresses Ankle2 mutant phenotypes and restores viability and brain size. Human mutational studies implicate the above-mentioned genes in microcephaly and motor neuron disease. We suggest that NS4A, ANKLE2, VRK1, and LLGL1 define a pathway impinging on asymmetric determinants of neural stem cell division.

Keywords: Ballchen; Bazooka; L(2)gl; MCPH16; Miranda; NS4A; VRK1; aPKC; brain development; congenital infection.

Figure 1:. Mutations in ANKLE2 cause microcephaly in humans and Drosophila.

The structure and transcripts of the Ankle2 gene with insertional mutations or tags are shown in (A). Ankle2^A, noted by *, is an EMS generated L326H mutation, while Ankle2^CRIMIC represents a Crispr-Cas9 mediated MIMiC-like (CRIMIC) insertion in the 4^th coding intron (Lee et al., 2018). ΦC31 mediated cassette exchange replaced the stop-polyA in Ankle2^CRIMIC with an artificial exon cassette consisting of SA-GFP-SD (Venken et al., 2011) to produce Ankle2^IGFP. A C-terminal GFP tagged Ankle2 genomic rescue construct (Ankle2-GFPR) was generated using recombineering (Venken et al., 2008). (B) Lethal stages and rescue by Ankle2-GFPR or (C) complementation tests of Ankle2^A, Ankle2^CRIMIC, and Ankle2^IGFP. (D) Ankle2 gene structure and genomic rescue construct (CH321–85N12; Venken et al., 2009). (E-L) Partial projections of 3^rd instar larval brains stained with Deadpan (Dpn) (purple, neuroblasts) a neuroblast marker (Bier et al., 1992) and Pros (green, GMC, neurons), a neuronal lineage marker (Campbell et al., 1994), to document overall brain structure is shown in (E) wild type (y w FRT19a), (F) y w Ankle2^A FRT19a (G) y w Ankle2^A FRT19a; Ankle2-GFPR, (H) y w Ankle2^A FRT19a; da::hANKLE2^wt, (I) Ankle2^CRIMIC, (J) Ankle2^CRIMIC; Ankle2-GFPR (K) y w Ankle2^A FRT19a; da:: hANKLE2 p.L573V and (L) y w Ankle2^A FRT19a; da:: hANKLE2 p.Q782* animals. The C-terminal GFP-tagged rescue P[acman] clone CH321–85N12, (Ankle2-GFPR, diagramed in A and B) rescues brain morphology, brain size, and lethality in both y w Ankle2^A FRT19a (G) and Ankle2^CRIMIC (J) animals. Quantification of brain size is shown in (M). Box plots hinges represent the 25^th to 75^th percentiles, the central line is the median, and whiskers represent min to max. Note that hANKLE2^wt and hANKLE2 p.L573V rescue brain size and lethality of Ankle2^A mutants, whereas hANKLE2 p.Q782*, hANKLE2 p.A109P, and hANKLE2 p.G201W do not. Here, Ankle2-GFPR is reported as GFPR. (N) Neuroblast specific expression of wild type human ANKLE2 (y w Ankle2^A FRT19a; insc::hANKLE2^wt) partially rescues brain size of Ankle2^A mutants. For quantifications, the total number counted (n) is noted below each graph. See also Figure S1–S3.

Figure 2:. Ankle2 localizes to the ER and is dynamically expressed in the brain.

(A-B) 3^rd instar larval brains from Ankle2-GFPR animals stained for GFP to document Ankle2 expression and localization in (A-B) single slices (arrows point to neuroblasts). (C-E) Live imaging of Ankle2 (D, green) and the ER labeled with Sec61B (E, da-GAL4, UAS-Sec61β-tdTomato, purple) (Summerville et al., 2016) shows strong colocalization. (F) Live Ankle2^A mutant neuroblast display aberrant Sec61β-tdTomato expression. (G-I) Fixed Ankle2-GFPR animals highlighting Ankle2-GFP (green) and another ER marker (Calnexin 99a, purple) (Riedel et al., 2016). (J) Fixed Ankle2 mutant animals (Ankle2^A) display aberrant ER structures (Calnexin 99a, purple). (K-M) Ankle2 (L, green) colocalizes with some portions of the nuclear envelope (Lamin Dm0, purple). (N) Ankle2 mutant animals (Ankle2^A) display disrupted nuclear envelope structure (Lamin Dm0, purple). See also Figure S4 and Movie S1.

Figure 3:. Ankle2 mutations affect asymmetric division, spindle alignment and centrosomes.

Metaphase neuroblasts stained with Baz (green, A-D), Par-6 (green, E-H), aPKC (green, I-L) and Mira (red, I-L) are shown in wild type (A, E, I), Ankle2^A/Y hemizygous (B, F, J), Ankle2^A/Ankle2^CRIMIC transheterozygous (C, G, K) and rescued (D, H, L) animals. Phospho-Histone H3 (pHH3, green, A-C, E-K) was used to identify cell cycle stage. * notes samples where pHH3 was not used. (M-P) quantification of phenotype severity demonstrate that Ankle2 is required for protein localization during asymmetric division in numerous cells. Below each graph, the # of neuroblasts counted for each genotype (n) is noted. (Q-U) Metaphase neuroblasts stained with aPKC (green) and Mira (red) to mark the polarity axis and DNA (white) and CNN (green puncta) to highlight the spindle axis from (Q) wild type (FRT19a) and (R-S) Ankle2 mutant neuroblasts. The angle between the spindle axis and polarity axis is measured and % of metaphase neuroblasts is plotted in 15° in tervals and is shown in (T-U). Below each group, the # of neuroblasts counted for each genotype (n) is noted. See also Movie S2–S5.

Figure 4:. Ankle2 controls Ballchen/VRK1 localization and function.

(A-D) Immunostaining of Ankle2-GFPR (green) and Ball (red) show dynamic localizations during the cell cycle. (B) Ankle2 is localized at the nuclear envelope at prophase. Ball is nuclear (A-B) until nuclear envelope breakdown (C) and then localizes to the cytoplasm through the end of mitosis (D). (E-F) Immunostaining of Ball (red) and Lamin (white) in Ankle2^A/+ heterozygotes and Ankle2^A/Y hemizygous mutants. Note that Lamin is disrupted and Ball becomes mislocalized throughout the cytoplasm during interphase in Ankle2^A/Y hemizygous mutants. Quantification is shown in Figure S3E. (G-H) Confocal projections of immunostaining of VRK1 (green) and DNA (purple) in primary human fibroblasts from (G) parental unaffected (p.L573V/+) and (H) an ANKLE2 compound heterozygous patient (p.L573V/p.Q782*). VRK1 is mislocalized in fibroblasts carrying microcephaly associated ANKLE2 variants (p.L573V/p.Q782* and p.V229G/p.V229G), quantified as nuclear intensity in (I). Arrows in (H) indicate cytoplasmic VRK1 staining, which is minimal in (G) control fibroblasts. (J-L) Partial projections of 3^rd instar larval brains stained with apical marker aPKC (green) and basal marker Miranda (red) in (J) wild type (y,w, FRT19a), (K) Ankle2^A, and (L) Ankle2^A;;ball^e107/+ animals. Note that removal of a single copy of ball rescues the phenotypes of Ankle2^A. (M) Quantification of 3^rd instar larval brain size (as shown in (J-L)). N≥6. One-way ANOVA with multi-comparison post-test. ****p<0.0001, ***p<0.001. Box plots hinges represent the 25^th to 75^th percentiles, a line is at the median, and whiskers represent min to max. (N) Ankle2^A, but not Ankle2^CRIMIC, lethality is rescued with introduction of multiple ball heterozygous mutations. (O-P) Quantification of (O) aPKC or (P) Mira crescent intensity in 3^rd instar metaphase neuroblasts in wild type (y w FRT19a), Ankle2^A and rescued Ankle2^A; ball^e107/+ animals, demonstrating that Ankle2 asymmetric division phenotypes are rescued with ball heterozygosity. Note that wild type and Ankle2^A quantifications were shown in Fig 3. Below each graph, the # of neuroblasts or brains counted for each genotype (n) is noted. See also Figure S3 and S5 and Movie S6.

Figure 5:. Ankle2 affects asymmetric division through aPKC and L(2)gl

(A) Ankle2 and Ball regulate asymmetric division. (B) Western analysis of phosphorylated aPKC in larval brains (mammalian T410 corresponds to T422 in Drosophila) from genomic rescue (Ankle2^A; Ankle2-GFPR), Ankle2^A, and Ankle2^A;; ball^e107/+ animals or total aPKC levels in genomic rescue (Ankle2^A; Ankle2-GFPR) and Ankle2^A mutants. Note that p-aPKC is reduced in Ankle2 mutants but restored with introduction of genomic rescue (GR) or reduction of Ball and is quantified as the ratio of p-aPKC to total aPKC. N=3 replicates. (C) In vivo immunoprecipitation of L(2)gl-GFP using GFP-nAb in L(2)gl^{MI07575-GFSTF} larvae demonstrates L(2)gl can interact with Ball in vivo. (D,F) Partial projections of third instar larval brains stained for Dpn (purple, neuroblasts) and Pros (green, daughter cells and neurons) of (D) Ankle2^A and (E) Ankle2^A/+;l(2)gl^ts3 mutant animals raised at 22°C with brain volume qua ntified in (F). Note that reduction of L(2)gl in an Ankle2^A hemizygous animals rescues brain size defects and lethality at 22°C. Box plots hinges represent the 25^th to 75^th percentiles, a line is at the median, and whiskers represent min to max. Below the graph, the # of brains counted for each genotype (n) is noted. See also Figure S6.

Figure 6:. Zika virus NS4A targets the Ankle2 pathway.

(A) 3^rd instar brains with neuroblast specific (insc-GAL4) NS4A-HA expression stained for HA (green) and DNA (white) shows that NS4A localizes in a pattern similar to Ankle2 (Figure 2). (B-E) Metaphase neuroblasts stained for aPKC (green) and Mira (purple) in brains with neuroblast specific (insc-GAL4) expression of (B) CD8-GFP, (C-D) NS4A, or (E) NS4A in ball^e107 heterozygous animals. aPKC and Mira crescent intensities are quantified in (F-G). (H) Brain volume quantification from animals with ubiquitous expression of CD8-GFP (control), NS4A-HA, or NS4A-HA in ball^e107 heterozygous animals using Act-GAL4 on the 2^nd chromosome. (I) Act-GAL4 (3^rd chromosome) ubiquitous expression of CD8-GFP (control), NS4A-HA, NS4A-HA and human ANKLE2, or NS4A-HA in l(2)gl^ts3 heterozygous animals. Note that human ANKLE2 expression or reduction of Ball or L(2)gl activity rescues NS4A induced brain defects. One-way ANOVA with multi-comparison post-test. ****p<0.0001, **p<0.01, *p<0.05. Box plots hinges represent the 25^th to 75^th percentiles, a line is at the median, and whiskers represent min to max. (J-L) aPKC (green), CNN (green), and Mira (purple) staining of metaphase neuroblasts with insc-GAL4 expression of (J) CD8-GFP, (K) NS4A, and (M) NS4A in ball^e107 heterozygous animals. The angle between the spindle axis and polarity axis is measured and % of total metaphase neuroblasts is noted at 15° intervals. Expression of NS4A causes localizati on defects of aPKC and Mira and spindle orientation defects in metaphase neuroblasts. (M) Zika virus NS4A inhibits the ANKLE2/VRK1 pathway, which regulates asymmetric determinant localization as well as the division axes.