Drosophila pericentrin requires interaction with calmodulin for its function at centrosomes and neuronal basal bodies but not at sperm basal bodies

Abstract

Pericentrin is a critical centrosomal protein required for organizing pericentriolar material (PCM) in mitosis. Mutations in pericentrin cause the human genetic disorder Majewski/microcephalic osteodysplastic primordial dwarfism type II, making a detailed understanding of its regulation extremely important. Germaine to pericentrin's function in organizing PCM is its ability to localize to the centrosome through the conserved C-terminal PACT domain. Here we use Drosophila pericentrin-like-protein (PLP) to understand how the PACT domain is regulated. We show that the interaction of PLP with calmodulin (CaM) at two highly conserved CaM-binding sites in the PACT domain controls the proper targeting of PLP to the centrosome. Disrupting the PLP-CaM interaction with single point mutations renders PLP inefficient in localizing to centrioles in cultured S2 cells and Drosophila neuroblasts. Although levels of PCM are unaffected, it is highly disorganized. We also demonstrate that basal body formation in the male testes and the production of functional sperm does not rely on the PLP-CaM interaction, whereas production of functional mechanosensory neurons does.

FIGURE 1:

Calmodulin is required for centriole duplication and PLP targeting. (A) Drosophila S2 cells treated with dsRNA against control, CaM, or Sas6 were fixed and stained for PLP (red), Asterless (green), and DNA (blue). Bar, 5 μm. (B) The average number of centrioles per cell was determined for each condition. Mean is indicated on top of each bar, and SE is indicated with red brackets. Both Sas6 and CaM knockdown caused a significant reduction in the number of centrioles/cell compared with controls (analysis of variance [ANOVA] test, ***p < 0.001, three independent experiments, 200 cells counted per condition per experiment). (C) PLP localization strength was determined in a blind experiment and was classified as strong, weak, or no localization (three independent experiments were performed, and at least 200 cells were scored for each). The percentage of “strong” localization is shown above each column. CaM knockdown significantly reduces PLP localization to centrosomes (ANOVA followed by a paired Turkey test, ***p < 0.001, n.s., not significant).

FIGURE 2:

PLP binds CaM through CBD2 within the PACT domain. (A) PLP^PF was divided into five fragments at the indicated amino acid positions. F5 (green) includes the PACT domain, which contains CBD1 and CBD2 (red). Blue blocks indicate regions of predicted coiled-coil. (B) S2 cell transfected with F5-GFP (green) and costained for Asl (red). Insets, enlargements of the indicated centrioles (arrows). Bar, 5 μm. (C) Y2H showing direct interaction of CaM and F5 as indicated by growth on SD –Ade –His –Leu –Trp (QDO) and growth and blue color on SD –Leu –Trp +Aureobasidin A + X-α-Gal (DDOXA). (D) Amino acid sequences of CBD1 and CBD2. The yellow regions indicate the mutated residues. The Y2H column shows a picture and empirically judged strength of the interaction for each. Growth on QDO and growth and color on DDOXA and SD –Ade –His –Leu –Trp + Aureobasidin A + X-α-Gal (QDOXA) plates indicates interaction.

FIGURE 3:

PLP centrosome targeting is disrupted by mutating CBD2. (A) Representative images of S2 cells transfected with GFP-PLP^PF or each of the CBD2 mutants (green). Cells were stained for Asl (red) to identify centrioles. The localization strength of GFP on the centriole was determined empirically to be strong (top), no signal (bottom), or weak (anything between these two levels). Bar, 5 μm. (B) PLP localization strength at the centrosomes was determined for indicated genotypes. At least three independent experiments were performed, and >150 cells were scored. The percentage of “strong” localization is indicated above each column. PLP localization was significantly reduced in PLP^∆CBD2 and PLP^2KR (ANOVA followed by a paired Turkey test, ***p < 0.001) but was normal in PLP^2IQ (n.s., not significant).

FIGURE 4:

The PLP-CaM interaction is required for proper PCM organization in neuroblasts. (A) Metaphase NBs from control (genotype C) and plp⁻ (genotype 1) flies were stained for Asl (green) to detect centrioles and Cnn (red) to detect PCM. Enlargements of the apical (arrow) and basal (arrowhead) centrosomes. (B) NBs from animals expressing the indicated transgenes (GFP, green) in the mutant plp⁻ background. NBs were strained for Asl (red, left) to detect centriole and Cnn (red, right) to detect PCM. All numbers refer to the genotype key (top left corner). Images are maximum intensity projections of the entire NB volume. Bar, 10 μm. (C, D) Circularity and GFP measurements were performed on at least 15 apical and 15 basal centrosomes for the indicated genotypes. Two independent ANOVA tests (blue and green bars above the graph) followed by a paired Turkey test were performed for each of the apical and basal centrosomes. (C) Circularity measurements show a significant centrosome shape change in plp⁻ (genotype 1, ***p < 0.001, ****p < 0.0001) when compared with control (genotype C), but this was rescued by plp^WT (genotype 2, n.s., not significant). Circularity was disrupted in the plp^2KR mutant (genotype 3, ***p < 0.001, **p < 0.01) but not the plp^2IQ mutant (genotype 4, n.s., not significant). (D) Centrosomal GFP signal was normalized to apical plp^WT levels (genotype 2). plp^2KR levels are significantly reduced on the apical and basal centrosomes (genotype 3, ****p < 0.0001). plp^2IQ levels were higher than those of plp^WT on the apical centrosome (genotype 4, ***p < 0.001), which could be attributed to the slightly higher expression of the plp^2IQ transgene. No significant difference was measured on the basal plp^2IQ centrosome.

FIGURE 5:

PLP requires interaction with CaM to localize to neuronal basal bodies. (A) Flies of the indicated genotype were determined to be viable if they fully eclosed and were motile in the vial. One hundred percent of the plp⁻ and plp^2KR individuals were incapacitated and scored “not viable.” (B) Diagram indicating the location of the chordotonal neurons in the second antennal segment. (C) Actin (blue), PLP::GFP expressed from the indicated transgenes (green), and Ana1::tdTomato in chordotonal neurons. Inset, basal bodies of chordotonal neurons. Unlike plp^WT and plp^2IQ, the plp^2KR allele was incapable of localizing to basal bodies, although centrioles still formed and were randomly positioned in the area of the neural cell bodies (dashed region). Bar, 10 μm.

FIGURE 6:

PLP does not require interaction with CaM for its role in spermatogenesis. (A) Cartoon representation of the centriole-to–basal body (BB) transition during spermatogenesis. (B) Testes from third-instar larvae of the indicated genotypes (12 testes from six larvae each). Testes were visually scored to contain abnormal centriole fragments, as clearly seen in the plp⁻ testes. The number of testes in which abnormal centriole fragments were observed over the total number scored is indicated on the left for each genotype. All the transgenes, including plp^2KR, fully rescued the plp⁻ centriole fragmentation. Asterisks indicate the location of the stem cell niche, termed the hub. At this stage, the gross morphology of testes of all genotypes appeared normal. The two right columns are random enlargements from the low-magnification image, intended to highlight the fragmentation phenotype. Bars, 50 μm (left), 15 μm (right enlargement). (C) Enlargements of a single centriole pair from spermatocytes of each of the indicated genotypes. Control and plp⁻ testes were stained for endogenous PLP (green), and the transgenes were imaged using the GFP signal. Sperm motility was determined by eye through a stereomicroscope after disruption of seminal vesicles. Bar, 2 μm.