Drosophila aPKC is required for mitotic spindle orientation during symmetric division of epithelial cells

The correct orientation of the mitotic spindle is crucial for epithelial morphogenesis and the maintenance of tissue integrity (Baena-Lopez et al., 2005; Jaffe et al., 2008; Segalen and Bellaiche, 2009). Spindle orientation relies on the interaction of astral microtubules with the cell cortex and the generation of pulling forces on the attached microtubules. Cortical cues, either intrinsic or extrinsic, may determine spindle orientation through the regulation of cell polarity and cytoskeleton reorganisation (Knoblich, 2008; Siller and Doe, 2009; Yamashita and Fuller, 2008).

The molecular mechanisms driving mitotic spindle orientation have been extensively studied in the context of asymmetric cell division. Drosophila embryonic neuroblasts orient the mitotic spindle along the apicobasal polarity axis in order to generate two daughter cells with different sizes and cell fates (Betschinger and Knoblich, 2004; Knoblich, 2008; Siller and Doe, 2009; Yamashita and Fuller, 2008). The Par3-Par6-aPKC apical complex interacts with Inscuteable (Insc), and together they are important for spindle apicobasal orientation through the induction of apical crescents of Gαi and Pins (Raps – FlyBase) (Schaefer et al., 2001; Schaefer et al., 2000; Schober et al., 1999; Wodarz et al., 1999; Yu et al., 2003; Yu et al., 2000). Pins-like proteins and Gα subunits are key components of a highly conserved molecular machinery that links the cell cortex to the astral microtubules of the mitotic spindle (Siller and Doe, 2009).

Dividing epithelial cells mostly orient the spindle along the plane of the epithelium (planar orientation) for mitosis to produce two identical daughter cells (Jaffe et al., 2008; Segalen and Bellaiche, 2009). Although the molecular machinery that regulates mitotic spindle orientation during asymmetric cell division is well known, the identity of the cortical cues that regulate spindle orientation in symmetrically dividing epithelial cells is less understood. Recent studies in mammalian tissue culture epithelial cells suggest that, similarly to asymmetric cell divisions, the aPKC-Par6 apical complex and its positive regulator Cdc42 (Hutterer et al., 2004) are also important for planar orientation of the mitotic spindle of dividing epithelial cells (Durgan et al., 2011; Hao et al., 2010; Jaffe et al., 2008; Qin et al., 2010). Phosphorylation of LGN (the mammalian homologue of Pins; also known as Gpsm2) by apical aPKC results in the inhibition of its binding to the apically anchored Gαi and its exclusion from the apical cortical region of the dividing cell (Hao et al., 2010; Konno et al., 2008; Zheng et al., 2010). Yet, in chicken neuroepithelial cells, aPKC is not required in vivo for spindle planar orientation during symmetric cell division (Peyre et al., 2011). Since Gαi and Pins/LGN are still required for spindle planar orientation (Peyre et al., 2011) it was proposed that, whereas the molecular machinery that connects the astral microtubules to the cell cortex is conserved, the polarization cues responsible for planar orientation vary between different epithelial cell types and/or developmental processes (Peyre et al., 2011).

In order to investigate whether Drosophila aPKC is required for spindle planar orientation, we took advantage of a temperature-sensitive allele of aPKC (apkc^ts) to modulate in vivo aPKC activity. From our work we conclude that, similar to what has been reported in mammalian tissue culture cells, Drosophila aPKC is required for spindle planar orientation and apical exclusion of Pins during symmetric division of epithelial cells. Our observations suggest that the spindle cortical cues are conserved between Drosophila and mammalian cells, and we provide the first in vivo evidence for a role of aPKC in spindle planar orientation.

Flies were raised using standard techniques. The apkc^ts allele was isolated from a previously reported maternal screen (Pimenta-Marques et al., 2008). apkc^ts zygotic mutant third instar larvae (L3) were obtained by crossing the apkc^ts/CyO Actin-GFP stock with both the Df(2R)l4/CyO Actin-GFP and apkc^k06403/CyO Actin-GFP stocks. apkc^ts/Df(2R)l4 or apkc^ts/apkc^k06403 larvae were selected by the absence of GFP, whereas GFP-positive larvae were used as controls. Additional control heterozygous larvae apkc^ts/+, apkc^k06403/+ and Df(2R)l4/+ were also used as controls for the apoptosis and spindle orientation experiments. After 24 hours egg laying at 25°C, the adults were transferred into a new vial, and the F1 progeny was shifted to 18°C, 25°C or 30°C until wing imaginal disc dissection as L3. The combination of temperatures used depended on the experiment and is indicated in each case.

For the genetic interaction between aPKC and pins, a stock was created carrying the apkc^ts mutation and a strong hypomorphic allele of pins, pins¹⁹³ (Parmentier et al., 2000) (apkc^ts/CyO Actin-GFP; pins¹⁹³/TM6B), which was crossed with the apkc^k06403/CyO Actin-GFP stock. After 24 hours egg laying at 25°C, the adults were transferred into a new vial and the F1 progeny was shifted to 28°C or 30°C. We then analysed the F1 progeny (L3 imaginal discs and adult flies) for dominant genetic interactions.

Maternal mutant embryos were obtained at 25°C from hemizygous females, apkc^ts/Df(2R)l4, selected by the absence of Cy dominant marker and crossed with wild-type males.

To generate homozygous clones of apkc^k06403 or apkc^ts that were negative for nuclear GFP label (nGFPminus), female y, w, hsFLP; FRT42B nGFP/CyO hshid flies were crossed with w; FRT42B, apkc^k06403/CyO hshid or w; FRT42B, apkc^ts/CyO hshid males. The offspring were heat shocked for 1 hour at 37°C at both 24 and 48 hours after a 24-hour egg collection, corresponding to the first and second larval instar. Wing discs were dissected from crawling L3 larvae, which had clones induced at first or at second instar.

Complementation group 5 contained a lethal allele and a maternal sterile allele (Pimenta-Marques et al., 2008). Both alleles were mapped using the Bloomington 2R deficiency kit. Deficiencies were crossed with both alleles and F1 progeny scored for zygotic lethality and maternal sterility. The following seven deficiencies failed to complement both alleles: Df(2R)en-A, Df(2R)en-B, Df(2R)E3363, Df(2R)Exel6060, Df(2R)ix[87i3], Df(2R)ED2219 and Df(2R)ED2155; both mutant alleles mapped to the cytological interval 47E3-47F5. By a candidate gene approach we concluded that both alleles failed to complement a known lethal P-element of aPKC (apkc^k06403).

To molecularly characterise these two potential new alleles of aPKC, genomic DNA was isolated from adult heterozygous flies. PCR and DNA sequencing of the aPKC gene locus was carried out using standard techniques. As a control for DNA polymorphisms, we used a mutant from a different complementation group isolated in the same screen. In this manuscript, we only report the molecular characterisation of the maternal sterile allele of aPKC (apkc^ts). Although the apkc^ts allele originally isolated in our screen had a second-site lethal mutation not associated with the aPKC gene locus and not uncovered by the Df(2R)l4 deficiency (data not shown), an apkc^ts allele without this lethal mutation showed an identical temperature-sensitive viability and larval wing disc apoptosis (data not shown).

Third instar wing imaginal disc fixation and stainings were performed using standard procedures (Lee and Treisman, 2001). For Tubulin staining the discs were fixed in PBS containing 10% fresh formaldehyde (Sigma) and 1 mM EGTA at room temperature for 20 minutes. Wing discs were mounted in Vectashield (Vector Laboratories). For maternal phenotypic analysis, 0-6 hour embryos were fixed and stained using standard procedures (Pimenta-Marques et al., 2008) except for F-actin staining, where the vitelline membrane was removed manually. The oogenesis phenotypic analysis was performed with tissue dissected from 2- or 3-day-old females, where mutant clones were induced by heat shock at first and second instar larval stages, and fixed in PBS containing 4% formaldehyde for 20 minutes. Embryos and ovaries were mounted in Fluorescent Mounting Medium (DakoCytomation).

Antibodies used were: rabbit anti-aPKC at 1:2000 (Santa Cruz, sc-216); mouse anti-aPKC at 1:100 (Santa Cruz, sc-17781); rabbit anti-cleaved caspase 3 at 1:500 (Cell Signaling, 9661S); mouse anti-Armadillo N2 7A1 at 1:50 [Developmental Studies Hybridoma Bank (DSHB)]; mouse anti-Neurotactin clone BP106 at 1:133 (DSHB); rat anti-DE-Cadherin at 1:20 (DCAD2, DSHB); rabbit anti-Par3 at 1:1000 (Wodarz et al., 1999); rabbit anti-pS980 Par3 at 1:50 (Morais-de-Sa et al., 2010); rabbit anti-Par6 at 1:1000 (Petronczki and Knoblich, 2001); mouse anti-pTyr at 1:1000 (Cell Signaling, 9411); rabbit anti-Lgl at 1:100 (Betschinger et al., 2003); mouse anti-Dlg clone 4F3 at 1:250 (DSHB); mouse anti-Tubulin at 1:500 (Sigma, T6199); rabbit anti-Pins at 1:1000 (Yu et al., 2000); and anti-phospho-Myosin light chain at 1:500 (Cell Signaling, 3671). For F-actin staining, a 5-minute incubation with phalloidin-Rhodamine at 1:200 (Sigma; stock concentration 1 mg/ml) was employed. For DNA staining, we used SYTOX Green (Invitrogen) at 1:5000 with 5 μg/ml RNase A in PBT (PBS+0.1% Tween-20) for 30 minutes at room temperature. Cy3- or Cy5-conjugated secondary antibodies were used at 1:1000 (Jackson ImmunoResearch, West Grove, PA, USA) and anti-rabbit Alexa Fluor 488 at 1:1000 (Molecular Probes).

Fluorescence images were obtained on a Leica TCS NT or a Zeiss LSM 510 confocal microscope. ImageJ (NIH) was used to perform measurements on transverse sections. To measure apoptosis levels the areas positive for cleaved caspase 3 were calculated (using the Freehand selection from ImageJ) and compared with a total inner ring (pouch plus hinge) average area of all discs. All measurements were performed using z-projections of larval wing discs.

Drosophila RE60936 full-length aPKC cDNA was cloned into pDONR221 (Gateway System, Invitrogen). Site-directed mutagenesis to insert the apkc^ts point mutation was performed using forward primer 5′-GTCAGCCATCCCTTCCTTAAGAATATGGATT-3′ and reverse primer 5′-AATCCATATTCTTAAGGAAGGGATGGCTGAC-3′. The wild-type and apkc^ts open reading frames were fused to a UASp promoter and a 6× N-terminal Myc tag using the Gateway system (Invitrogen). Both constructs were used to generate transgenic flies (BestGene).

Third instar wing imaginal discs were dissected and collected. Thirty wing discs were lysed in NB buffer [50 mM Tris-Cl pH 7.5, 150 mM NaCl, 2 mM EDTA, 0.1% NP40, 1 mM DTT, 10 mM NaF and Complete Protease Inhibitor Cocktail (Roche)], protein levels were quantified (BioRad protein assay), SDS-PAGE sample buffer was added and samples were heated for 5 minutes at 100°C. Embryos at 0-6 hours after egg laying (AEL) were collected and dechorionated with 50% commercial bleach solution. Each embryonic protein sample was collected by lysing ten embryos with a needle in SDS-PAGE sample buffer and heating for 5 minutes at 100°C. Protein samples were run on 8% SDS-PAGE gels and proteins were then transferred into Hybond-ECL membranes (Amersham). Membranes were blocked overnight at 4°C in 5% non-fat milk in PBT, then primary antibodies were added to membranes and incubated overnight at 4°C. Following washes with PBT, secondary antibodies were added and incubated for 2 hours at room temperature. Protein detection was performed using ECL solution for 1 minute and Hyperfilm ECL (Amersham). Primary antibodies used were rabbit anti-aPKC at 1:2000 (Santa Cruz, sc-216) and mouse anti-alpha-Tubulin Dm1A at 1:20,000 (Sigma). Secondary detection was performed with rabbit and mouse HRP-conjugated antibodies at 1:4000 (Jackson ImmunoResearch).

Embryos laid by females expressing wild-type or apkc^ts Myc-tagged aPKC (aPKC^wt and aPKC^ts) under the control of the Actin5C-Gal4 ubiquitous driver were collected from 0 to 6 hours at 25°C. Embryos were lysed in NB buffer. Extracts were centrifuged three times at 14,000 rpm (20,817 g) for 3 minutes. Owing to a consistent difference in the expression levels of Myc-tagged aPKC^wt and aPKC^ts (supplementary material Fig. S2B), different initial amounts of embryonic total protein were used to obtain comparable amounts of immunoprecipitated aPKC. As a negative control for each kinase assay, an equivalent amount of total protein from wild-type embryos (without Myc-tagged aPKC) was used for immunoprecipitation (supplementary material Fig. S2A). Protein extracts were incubated with mouse anti-Myc antibody at 1:200 (9E10, Covance) for 2 hours at 4°C. Sepharose protein G beads (Sigma) were added and incubated for 2 hours at 4°C. Beads were washed three times in NB buffer and twice with kinase buffer (250 mM HEPES pH 7.4, 0.2 mM EDTA, 1% glycerol, 150 mM NaCl, 10 mM MgCl₂). The kinase assay was performed by adding to the washed beads 20 μl kinase buffer supplemented with 1 μg MBP:Baz (amino acids 829-1168) [kindly provided by Daniel St Johnston (Morais-de-Sa et al., 2010)] and ATP mix [75 μM ATP and 20 μM γ³²P-ATP (10 Ci/mmol)] with incubation at 25°C or 30°C for 30 minutes. Samples were then heated at 100°C for 5 minutes in SDS-PAGE sample buffer. Quantification of the total levels of immunoprecipitated Myc-tagged aPKC was performed by standard western blot. To analyse aPKC kinase activity, samples were run on an SDS-PAGE gel, which was subsequently dried and analysed using a Storm 860 phosphorimager (General Electric) for quantification. The amount of MBP:Baz phosphorylation was monitored by the levels of ³²P incorporation, which were divided by the total amount of aPKC in order to estimate the in vitro kinase activity of aPKC. Negative controls comprised wild-type embryos immunoprecipitated with anti-Myc antibody (supplementary material Fig. S2A) or transgenic embryos carrying a Myc-tagged aPKC but where a mock immunoprecipitation was performed without adding anti-Myc antibody (Fig. 1A). Any non-specific ³²P incorporation observed in each negative control was subtracted from the phosphorylation detected in the respective aPKC kinase assay (Fig. 1A′,B).

Unpaired t-test was performed using Prism 5.00 for Windows (GraphPad Software, San Diego, CA, USA).

Previously, we isolated a collection of maternal mutants defective for early embryonic development (Pimenta-Marques et al., 2008). From this collection we identified a new allele of atypical protein kinase C (apkc^ts) that contained a point mutation in a highly conserved phenylalanine within the protein kinase domain (from a phenylalanine to a leucine at position 532; F532L). By comparison with the human aPKC crystal structure (Messerschmidt et al., 2005) the mutated amino acid was mapped to a hydrophobic pocket within the protein kinase domain (data not shown). Consistently, mutant Myc-tagged aPKC^ts immunoprecipitated from embryonic protein extracts showed a significant decrease in its in vitro kinase activity when compared with wild-type Myc-tagged aPKC^wt (Fig. 1A,A′).

Drosophila aPKC is an essential gene, and hemizygotes between apkc^k06403 (a strong hypomorphic allele) and a deletion that uncovers the aPKC gene locus [Df(2R)l4] did not eclode from pupal cases (data not shown) (Rolls et al., 2003). In contrast to apkc^k06403, hemizygous mutants between apkc^ts and DF(2R)l4 or transheterozygotes between apkc^ts and apkc^k06403 showed temperature-sensitive viability. Larvae hemizygous between apkc^ts and the Df(2R)l4 deletion and transheterozygous between apkc^ts and apkc^k06403 were viable at 25°C (the permissive temperature), without any decrease of viability (supplementary material Fig. S1A; data not shown) and adult flies were morphologically normal. By contrast, at 30°C (the restrictive temperature) most apkc^ts/Df(2R)l4 hemizygous and apkc^ts/apkc^k06403 transheterozygous larvae failed to eclode from the pupal cases, with occasional escapers flies (supplementary material Fig. S1A). At 27-28°C (semi-permissive temperature), hemizygous and transheterozygous mutant viability was highly variable (data not shown), but most adult flies showed significant abdominal midline dorsal closure defects (see below). Nevertheless, although apkc^ts showed temperature-sensitive phenotypes, we failed to detect a decrease of aPKC^ts in vitro kinase activity at the restrictive temperature when compared with the permissive temperature (Fig. 1B; for original kinase assay gel see supplementary material Fig. S2A).

In order to investigate the role of aPKC in symmetric division we took advantage of the highly proliferative larval wing disc epithelium. aPKC is essential for epithelial apicobasal polarity, the correct formation and maintenance of adherens junctions (AJs) and, consequently, for epithelial integrity (Goldstein and Macara, 2007; Knust and Bossinger, 2002; Suzuki and Ohno, 2006). Wing disc clones mutant for apkc^k06403 showed loss of epithelial integrity, cell extrusion and induction of apoptosis (Georgiou et al., 2008; Rolls et al., 2003). Consistently, we failed to detect clones mutant for apkc^k06403 (nGFP negative), in contrast to twin-spot wild-type clones (2× nGFP), if induced during first instar (supplementary material Fig. S7A,B). Moreover, the mutant clones were positive for apoptosis if induced during second instar (supplementary material Fig. S7C-F). Since the adult viability of apkc^ts mutants was temperature sensitive, we investigated whether there was any loss of epithelial integrity at permissive and restrictive temperatures.

apkc^ts wing discs [apkc^ts/Df(2R)l4 hemizygous or apkc^ts/apkc^k06403 transheterozygous] showed a temperature-sensitive induction of apoptosis. Whereas at the permissive temperatures (18°C and 25°C) there were low levels of cell extrusion and apoptosis (Fig. 2D,E, quantification in 2G), at the restrictive temperature (30°C) there were significantly higher levels of cell extrusion and apoptosis, and the discs were smaller than those observed in the control heterozygous larvae (Fig. 2C,F, quantification in 2G; data not shown). We failed to detect any temperature-sensitive induction of apoptosis in the control heterozygous wing discs (Fig. 2A,B, quantification in 2G). It should be noticed that the entire wing disc was mutant for apkc^ts and we could not discern any particular pattern of apoptosis within the disc pouch (data not shown). Large clones mutant for apkc^ts induced during first instar also showed temperature-sensitive apoptosis in L3 wing discs (supplementary material Fig. S7M-O,Y-AA). Since the wing disc epithelium mutant for apkc^ts showed temperature-sensitive apoptosis, we concluded that it was possible to modulate the penetrance of the larval wing disc phenotypes at permissive and restrictive temperatures.

Epithelial cells mutant for apkc^k06403 show a loss of apicobasal polarity, with an abnormal cortical distribution of AJ components and basolateral markers (Georgiou et al., 2008; Rolls et al., 2003). Since our aim was to investigate the role of aPKC in the planar orientation of the mitotic spindle, we first examined the epithelium architecture in apkc^ts wing discs. In order to minimize experimental artefacts, wing discs mutant for apkc^ts [GFP negative; apkc^ts/Df(2R)l4] and control wing discs [GFP positive; apkc^ts/CyO Actin-GFP or Df(2R)l4/CyO Actin-GFP] were all pooled together for immunostaining. At restrictive temperature (30°C), the apkc^ts epithelium was apparently normal, with normal localisation of DE-Cadherin (DE-Cad; Shotgun – FlyBase), Armadillo (Fig. 3A,B,D,E), aPKC (Fig. 3C,F) and cortical F-actin (Fig. 3I,J). We also failed to detect at restrictive temperature a significant reduction of aPKC protein levels in apkc^ts wing discs (Fig. 2H). Moreover, the basolateral proteins Lgl [L(2)gl – FlyBase] and Dlg were efficiently excluded from the apical domain of the epithelium (Fig. 3G-J). Nevertheless, we still observed a significant amount of cell extrusion and apoptosis (cleaved caspase 3) in the basal region of apkc^ts wing discs at the restrictive temperature (Fig. 3J).

The aPKC apical complex includes atypical protein kinase C and two PDZ-containing proteins: Par3 [Bazooka (Baz) – FlyBase] and Par6 (Par-6 – FlyBase). Par3 interacts with the aPKC-Par6 complex and is required for its apical localisation and for AJ maturation (Bilder et al., 2003; Harris and Peifer, 2005; Izumi et al., 1998; Joberty et al., 2000). In embryos, Par3 is localised slightly more basally than the aPKC-Par6 complex (Harris and Peifer, 2005). It was recently shown that aPKC-dependent phosphorylation of Par3 is required for the correct localisation of Par3 at the apical/lateral border, its exclusion from the extreme apical domain, and to regulate the interaction with the apical protein Stardust (Krahn et al., 2010; Morais-de-Sa et al., 2010). Consistent with the observation that the epithelial architecture of apkc^ts wing discs was normal at restrictive temperature, we detected normal apical localisation of Par3 (Fig. 4A-C,E-G) and aPKC-dependent phosphorylation of Par3 (Fig. 4D,H). We conclude that, although there was a significant amount of cell extrusion and apoptosis at the restrictive temperature, the apkc^ts epithelial cells maintained an apparently normal architecture.

Dividing epithelial cells mostly orient the spindle along the plane of the epithelium (planar orientation). Mitotic spindle misorientation can cause a rotation of the cleavage plane of the dividing cell and an asymmetric distribution of cell junctional components between daughter cells, which may lead to epithelial cell extrusion and apoptosis. Since wing discs mutant for apkc^ts showed at restrictive temperature a significant amount of apoptosis in an otherwise normal epithelium, we investigated whether mitotic spindle orientation was affected in dividing epithelial cells. Consistent with our hypothesis, we observed a temperature-sensitive increase in spindle misorientation in epithelial cells mutant for apkc^ts (Fig. 5).

In control heterozygous wing discs [apkc^ts/+ and Df(2R)l4/+] the great majority of dividing cells showed (as expected) the mitotic spindle oriented along the plane of the epithelium (Fig. 5A-C, quantification in 5J; detailed statistical analysis in supplementary material Fig. S3). By contrast, in wing discs mutant for apkc^ts [apkc^ts/Df(2R)l4 and apkc^ts/apkc^k06403] there was a significant decrease in the proportion of mitotic cells with the correct planar orientation of the spindle at permissive (Student’s t-test, P=0.002) and restrictive (P=0.0001) temperatures (Fig. 5D-F,H,I, quantification in 5J; detailed statistical analysis in supplementary material Fig. S3). Importantly, and consistent with a temperature-sensitive increase in apoptosis, wing discs mutant for apkc^ts showed a significant decrease in the proportion of mitotic cells with planar orientation of the spindle at restrictive temperature (30°C) when compared with the permissive temperature (25°C) (P=0.0001) (supplementary material Fig. S3). In control wing discs there was no significant difference in spindle planar orientation between permissive and restrictive temperatures (P=0.682) (Fig. 5J; supplementary material Fig. S3). We conclude that aPKC is required for the planar orientation of the mitotic spindle during the symmetric division of epithelial cells.

In mammalian tissue culture epithelial cells, the Pins homologue LGN has to be correctly excluded from the apical cortex of the dividing cells to ensure the planar orientation of the mitotic spindle (Konno et al., 2008; Zheng et al., 2010). Phosphorylation of LGN/Pins by apical aPKC inhibits its binding to apically anchored Gαi and thereby results in its exclusion from the apical cortical region of the dividing cell (Hao et al., 2010). Since Drosophila wing discs mutant for apkc^ts showed defects in spindle orientation, we investigated Pins localisation during mitosis.

In control wing discs, Pins cortical localisation in dividing epithelial cells (large rounded cells in Fig. 6A) was mainly basolateral when compared with DE-Cad (Fig. 6D-D″′,E-E″′). At restrictive temperature, mitotic cells (large rounded cells in Fig. 6F) from apkc^ts discs showed a clear mislocalisation of Pins towards the epithelial apical domain as compared with DE-Cad localisation (Fig. 6G-H,I-I″′,J-J″′). The proportion of mitotic cells with apical mislocalisation of Pins in control heterozygous wing discs was zero [n=516 dividing cells] and in apkc^ts/Df(2R)l4 hemizygous wing discs was 12.75±7.97% (n=664) (for a breakdown of the results from four replicates see supplementary material Fig. S4).

If the apical mislocalisation of Pins was related to the apkc^ts wing disc phenotype, we anticipated that a reduction in pins expression was likely to partially suppress it. Consistently, we observed that at semi-permissive temperature, but not at restrictive temperature, a strong hypomorphic allele of pins [pins¹⁹³ (Parmentier et al., 2000)] behaved as a dominant suppressor of apkc^ts wing disc apoptosis (Fig. 7A). At 28°C, apkc^ts; pins/+ (apkc^ts/apkc^k06403; pins¹⁹³/+) wing discs showed a significant reduction (Student’s t-test, P=0.0001) of apoptosis in the pouch when compared with apkc^ts (apkc^ts/apkc^k06403) wing discs (Fig. 7A, left). Nevertheless, such a reduction was not observed at the restrictive temperature, where the level of apoptosis was significantly higher in both mutant genotypes (Fig. 7A, right). Control discs (heterozygous: apkc^ts/CyO Actin-GFP or apkc^k06403/CyO Actin-GFP) did not show apoptosis at either temperature.

At the semi-permissive temperature the viability of apkc^ts mutants is variable (data not shown), but most adult flies show dramatic abdominal midline dorsal closure defects (Fig. 7B). Abdominal dorsal closure defects are frequently associated with abnormalities related to histoblast proliferation and morphogenesis (Ninov et al., 2007). Interestingly, at the semi-permissive temperature pins¹⁹³ also behaved as a dominant suppressor, with a significant reduction in the percentage of adult flies with strong abdominal midline dorsal closure defects (from 78% in apkc^ts to 38% in apkc^ts; pins/+) (Fig. 7B). At the restrictive temperature, both mutant genotypes showed identical abdominal midline dorsal closure defects (Fig. 7B). Control flies did not show any abdominal defects at either temperature.

Zygotic mutants of apkc^ts were viable and morphologically normal at the permissive temperature (supplementary material Fig. S1; data not shown), but we detected low levels of cell extrusion and apoptosis in larval wing discs (Fig. 2E,G). We hypothesized that a sublethal reduction in aPKC activity could lead to tissue overgrowth and tumour development if there were inhibition of apoptosis and compensatory cell proliferation. Consistently, at the permissive temperature inhibition of apoptosis with p35 (baculoviral caspase inhibitor protein) induced the development of large tumour-like tissues in the basal region of the larval wing discs mutant for apkc^ts [Scalloped-Gal4/+; apkc^ts/Df(2R)l4; UAS-p35/+] (Fig. 2J; supplementary material Fig. S5E,F), without affecting the apical localisation of aPKC and DE-Cad (supplementary material Fig. S5G,H). Apoptosis inhibition was not associated with the development of tumour-like tissues in the control heterozygous discs (Fig. 2I; supplementary material Fig. S5A-D).

We conclude that although viable and morphologically normal at the permissive temperature, zygotic mutants of apkc^ts were nevertheless extremely sensitive to apoptosis inhibition. Since inhibition of apoptosis failed to suppress the epithelial cell extrusion phenotype observed in apkc^ts wing discs, this suggested that apoptosis was not the cause of epithelial delamination.

Although viable and morphologically normal at permissive temperature, Drosophila females mutant for apkc^ts [hemizygous between apkc^ts and the Df(2R)l4 deletion] were nevertheless sterile, with no larvae hatching from laid eggs even when crossed with wild-type males (n=2428 eggs). Maternal mutant embryos for apkc^ts (hereafter referred to as apkc^ts mutant embryos) failed to complete germ-band extension (GBE), with a complete loss of epithelial integrity (supplementary material Fig. S6A-D). Similar results were obtained for maternal mutant embryos obtained from germline clones for apkc^ts and for maternal mutant embryos laid by females transheterozygous between apkc^ts and apkc^k06403 (data not shown). When embryos mutant for apkc^ts started GBE, there was a rapid delocalisation of the apically localised aPKC into cortically localised aggregates (supplementary material Fig. S6E-G). aPKC-positive aggregates were highly enriched for Par3 (supplementary material Fig. S6H-J), Par6 (supplementary material Fig. S6K,L), Armadillo and F-actin (supplementary material Fig. S6M-T). These morphogenetic defects are similar to those previously observed with a strong allele of aPKC [apkc^k06403 (Harris and Peifer, 2007)] (data not shown).

Yet, we detected some differences between apkc^ts and apkc^k06403. Embryos mutant for apkc^k06403 (maternal mutants) showed a strong reduction in aPKC protein levels (Rolls et al., 2003), but we failed to detect such a change by western blot using apkc^ts embryo protein extracts (supplementary material Fig. S6X). Moreover, and similar to what was observed in larval wing discs (Fig. 3C,F; supplementary material Fig. S7G-L,S-X,AE-AJ), during blastoderm cellularisation and before the onset of GBE, we detected normal levels of apically localised aPKC in apkc^ts (but not apkc^k06403) embryos (supplementary material Fig. S6U-W), which suggested that apkc^ts embryos were able to apically localise aPKC but failed to correctly coordinate epithelial morphogenesis during GBE. Since the maternal phenotypes of apkc^ts were not temperature sensitive and were reminiscent of previously described maternal phenotypes of a strong hypomorphic allele of aPKC (Harris and Peifer, 2007), we concluded [in agreement with a previous suggestion (Kim et al., 2009)] that GBE was likely to require particularly high levels of aPKC activity and that epithelial tissues had differential requirements for aPKC activity during development.

At the permissive temperature (25°C) Drosophila females mutant for apkc^ts laid a significant number of eggs of normal size and shape and with correctly developed dorsal appendages (data not shown). These eggs were fertilised and the embryos developed normally until the onset of gastrulation and GBE (supplementary material Fig. S6; data not shown). Nevertheless, since at the permissive temperature females mutant for apkc^ts showed a reduction in egg laying compared with control heterozygous females (data not shown), we characterised oogenesis in these mutants. Somatic follicular epithelial cells mutant for apkc^ts showed temperature-sensitive cyst encapsulation defects after the induction of mutant clones (supplementary material Fig. S8A-D). Interestingly, similar cyst encapsulation defects were observed at the permissive temperature (25°C) in females mutant for apkc^ts [hemizygous between apkc^ts and the Df(2R)l4 deletion or transheterozygous between apkc^ts and apkc^k06403] (supplementary material Fig. S8E-H), which suggested that the apkc^ts mutant phenotypes could also be modulated by allele copy number. The cyst encapsulation defects could be detected immediately after the mesenchymal-to-epithelial transition in region 3 of the germarium (supplementary material Fig. S8E), but the follicular epithelium that formed showed normal AJs (supplementary material Fig. S8E,G,H; data not shown), apicobasal polarity (supplementary material Fig. S8C,E; data not shown) and actin-myosin cytoskeleton (supplementary material Fig. S8F; data not shown).

Our study provides the first in vivo evidence for the requirement of aPKC in determining the planar orientation of the mitotic spindle during the symmetric division of epithelial cells. Furthermore, our work shows that, similar to observations in mammalian tissue culture cells, Drosophila aPKC is also required for the apical exclusion of Pins during mitosis. Altogether, our work suggests that the cortical cues necessary for spindle planar orientation are conserved between Drosophila and mammalian cells, and that they are likely to be similar to those known to be important for spindle apicobasal orientation during asymmetric cell division.

In this work we characterised a novel hypomorphic allele of aPKC (apkc^ts). This allele contains a point mutation at a highly conserved phenylalanine within the protein kinase domain and shows reduced levels of in vitro kinase activity without any significant reduction of aPKC protein levels or any detectable change in its apical localisation. At restrictive temperature, larval wing discs mutant for apkc^ts showed significant levels of cell extrusion and apoptosis, yet their epithelial architecture was normal. Since loss of aPKC activity was commonly associated with a complete collapse of apicobasal polarity, abnormal AJs, apical constriction and cell extrusion, we hypothesized that the apkc^ts allele specifically failed to complement a function of aPKC that was not directly related to apicobasal polarity and/or the formation and maintenance of junctional components. Previously, it was suggested that different epithelial tissues can have differential requirements for aPKC activity during Drosophila development (Kim et al., 2009). Our data expanded this hypothesis and suggested that, even within the same epithelial cell, aPKC was likely to have distinct thresholds of activity to correctly regulate different cellular processes.

Despite the fact that zygotic mutants of apkc^ts showed significant temperature-sensitive phenotypes, we failed to detect a decrease in aPKC^ts in vitro kinase activity at the restrictive temperature when compared with the permissive temperature. We hypothesized that the observed phenotypes were possibly due to an overhaul property of the epithelial cell whereby, at restrictive temperature, there were higher requirements for aPKC kinase activity. Alternatively, it is also possible that in vivo the aPKC^ts mutant protein behaved as a temperature-sensitive kinase and our in vitro kinase assay failed to detect such behaviour.

Consistent with the mammalian tissue culture work (Durgan et al., 2011; Hao et al., 2010; Qin et al., 2010), our analysis of apkc^tsDrosophila wing discs revealed a requirement of aPKC for spindle orientation during symmetric mitosis. Complete randomization of spindle positioning would potentially lead to spindle misorientation in 50% of epithelial cell divisions. In apkc^ts wing discs, there was a significant decrease in the proportion of dividing epithelial cells with the correct planar orientation of the spindle, which suggested an increase in the randomization of spindle orientation. Although it is unclear when a departure from planarity should be considered functionally significant, an increase in the randomization of spindle orientation is likely to be associated with an increase in the probability of cell extrusion and apoptosis, as misorientation of the mitotic spindle along the apicobasal axis could potentially cause rotation of the mother cell cleavage plane, a reduction of the apical and sub-apical domains of the basally localised daughter cell, and its subsequent extrusion and apoptosis. We hypothesized that spindle misorientation was likely to be one of the main causes for the cell extrusion and apoptosis phenotypes observed in apkc^ts wing discs at restrictive temperature. Consistently, both the apoptosis and spindle orientation defects observed in the apkc^ts wing discs were similarly temperature sensitive.

The highly conserved Pins/LGN protein is important for spindle orientation as it links the cell cortex of the dividing cell with the astral microtubules of the mitotic spindle. In mammalian epithelial tissue culture cells, LGN needs to be correctly excluded from the apical cortex of the dividing cells to ensure planar orientation of the spindle (Durgan et al., 2011; Hao et al., 2010; Jaffe et al., 2008; Konno et al., 2008; Qin et al., 2010; Zheng et al., 2010). It has been proposed that phosphorylation of LGN/Pins by apical aPKC inhibits its binding to apically anchored Gαi, resulting in its exclusion from the apical cortical region of the dividing cell (Hao et al., 2010). Consistent with this model, we observed that at restrictive temperature larval wing discs mutant for apkc^ts showed significant mislocalisation of Pins in a subset of mitotic cells, which suggested that Drosophila aPKC is also required in vivo to exclude Pins from the apical domain of dividing epithelial cells.

Further supporting the hypothesis that apical mislocalisation of Pins is one of the main causes of the phenotypes observed in apkc^ts mutants, we observed that a strong hypomorphic allele of pins (pins¹⁹³) could behave as a dominant suppressor of apkc^ts wing disc apoptosis at the semi-permissive temperature. Moreover, and suggesting that the aPKC-dependent regulation of Pins is not restricted to the wing disc epithelia, we also observed that pins¹⁹³ behaved as a dominant suppressor of the abdomen dorsal closure midline defects observed in adult flies mutant for apkc^ts.

Altogether, our data suggest that aPKC is an important in vivo regulator of spindle orientation during the symmetric division of epithelial cells, with the aPKC-dependent phosphorylation of Pins resulting in its exclusion from the apical domain of dividing cells, a role that is most likely conserved between Drosophila and mammalian cells. Yet, and similarly to chicken neuroepithelial cells (Peyre et al., 2011), it is nevertheless possible that other Drosophila epithelial tissues use distinct (or possibly redundant) polarization cues [e.g. integrin signalling (Fernandez-Minan et al., 2007)] to orient the spindle during symmetric mitosis.

Our observations indicate that identical aPKC-dependent cortical cues are likely to be used to orient the mitotic spindle during symmetric and asymmetric mitosis. Similarly, Pins is required for both types of cell division, which suggests that modulation of the aPKC-dependent apical exclusion of Pins is likely to play a key role in the means by which similar cortical cues can be differentially interpreted in epithelial cells (spindle planar orientation) and neuroblasts (spindle apicobasal orientation).

We thank Alekos Athanasiadis for help with the structural analysis of the apkc^ts point mutation; our colleagues Buzz Baum, Moises Mallo, Florence Janody, Sérgio Simões and Rita Sousa-Nunes for discussion and suggestions that greatly improved the manuscript; and Richard Hampson for manuscript editing.