Ral GTPase promotes asymmetric Notch activation in the Drosophila eye in response to Frizzled/PCP signaling by repressing ligand-independent receptor activation

Abstract

Ral is a small Ras-like GTPase that regulates membrane trafficking and signaling. Here, we show that in response to planar cell polarity (PCP) signals, Ral modulates asymmetric Notch signaling in the Drosophila eye. Specification of the initially equivalent R3/R4 photoreceptor precursor cells in each developing ommatidium occurs in response to a gradient of Frizzled (Fz) signaling. The cell with the most Fz signal (R3) activates the Notch receptor in the adjacent cell (R4) via the ligand Delta, resulting in R3/R4 cell determination and their asymmetric positions within the ommatidium. Two mechanisms have been proposed for ensuring that the cell with the most Fz activation sends the Delta signal: Fz-dependent transcriptional upregulation in R3 of genes that promote Delta signaling, and direct blockage of Notch receptor activation in R3 by localization of an activated Fz/Disheveled protein complex to the side of the plasma membrane adjacent to R4. Here, we discover a distinct mechanism for biasing the direction of Notch signaling that depends on Ral. Using genetic experiments in vivo, we show that, in direct response to Fz signaling, Ral transcription is upregulated in R3, and Ral represses ligand-independent activation of Notch in R3. Thus, prevention of ligand-independent Notch activation is not simply a constitutive process, but is a target for regulation by Ral during cell fate specification and pattern formation.

Fig. 1.

Ral mutant eye phenotype. (A) An apical tangential section through a wild-type adult eye is shown at the top. The orange line marks the equator. The diagram beneath indicates the facet orientations. (B) (Top) Sections of wild-type and Ral^EE1/Y eyes. The diagrams beneath indicate facet orientations and mutant phenotypes. Symbols are defined in C. (C) Quantification of classes of Ral^EE1/Y ommatidia. Reverse orientation facets were not scored because in the context of all of the aberrant ommatidia, it was often difficult to locate the equator. (D) Diagram of five R-cell pre-clusters in a third instar larval eye disc and rotation with respect to the equator. A, anterior; P, posterior; mf, morphogenetic furrow, moving in the direction of arrow. (E,F) Third instar larval eye discs expressing GFP in R2/5 and R3/4, and immunolabeled with anti-Elav (R-cell nuclei) and phalloidin (actin). The genotypes are ro-gfp (wild-type) and Ral^EE1/Y; ro-gfp. The dotted line is the equator, and the morphogenetic furrow is leftward. Numbers indicate R2/R5 and R1/R6. Asterisks are ommatidia in which one of the R1/R6 pair are absent. The lines indicate misrotated ommatidia. Scale bar: in A, 20 μm for A,B; in F, 10 μm for E,F.

Fig. 2.

Genetic interactions between Ral and Notch pathway mutants. (A) External adult eyes of the genotypes indicated are shown. Ral⁻ heterozygous eyes are like wild type (smooth, left), lqf^FDD9 eyes are rough and although this is not obvious in the photographs, eyes with both mutations are more like wild type than like lqf^FDD9. The graph shows quantitative analysis of the wild-type and mutant ommatidia observed in tangential adult eye sections. For each genotype, the data were obtained from five eyes and 600-900 ommatidia. The symmetric ommatidia were scorable only in ommatidia with normal numbers of R cells. ***P<0.0001, *P<0.02; unpaired t-test. Data are mean + s.e.m. (B) External adult eyes of the indicated genotypes are shown. Ral⁻ hemizygous eyes (left) are rough. They are made smoother by heterozygous loss-of-function of lqf, and rougher by heterozygous loss-of-function of Dl or neur. The suppressive effect of lqf is quantified on the right. Ral^EE1/Y data are from Fig. 1 and Ral^EE1/Y; lqf^AG/lqf⁺ data were obtained from tangential sections of 732 ommatidia in four eyes. ***P<0.0005, **P<0.003; unpaired t-test. Data are mean + s.e.m.

Fig. 3.

Interactions between Ral, N⁺-GV3 and N^Sev11-GV3. (A) Diagrams of the protein products of the hs-N⁺-GV3 and hs-N^Sev11-GV3 transgenes (Struhl and Adachi 1998). ecd, extracellular domain; icd, intracellular domain; TM, transmembrane domain; pm, plasma membrane; GV, Gal4/VP16; Sev11, truncated extracellular and TM domains of Sevenless receptor; arrow, cleavage site that generates N^icd-GV3. (B) External eyes of the genotypes indicated at the top are shown. The flies were either not heat-shocked, or else heat-shocked as third instar larvae for 1 or 2 hours at 37^oC to express the hs-N⁺-GV3 or hs-N^Sev11-GV3 transgene. Neither N⁺-GV3 nor N^Sev11-GV3 causes eye roughness, but each enhances the roughness of Ral^EE1/Y eyes. (C) Dissected third instar larvae with gut extruded from cuticle, photographed to visualize GFP fluorescence. Ral-driven GFP expression (part a) is visible in the salivary glands (sg) and anterior midgut (amg), and mainly further posterior in the middle midgut (mmg) in parts d and f. GFP expression was quantified on protein blots (see Fig. S3 in the supplementary material).

Fig. 4.

Analysis of adult ommatidia where pre-R3/pre-R4 are mosaic for Ral⁺ or Ral⁻. (A) An apical tangential section through a w⁻ Ral⁻/w⁻ Ral⁻ clone in a w⁻ Ral⁻ / w⁺ Ral⁺ eye. The genotype is: w Ral^EE1 FRT19A/FRT19A; eyFLP/+. Circled ommatidia have mosaic R3/R4 cells; orange are EQ+/PO− (all asymmetrical) and green are EQ−/PO+ (the starred one is symmetrical). Scale bar: 20 μm. (B) Pooled results from 13 different eye clones are shown. EQ, equatorial cell (pre-R3); PO, polar cell (pre-R4).

Fig. 5.

Analysis of developing larval ommatidia where pre-R3/pre-R4 are mosaic for Ral⁺ or Ral⁻. (A) A Ral^EE1/gfp⁺ eye disc with Ral^EE1 homozygous (gfp⁻) clones is shown. The disc expresses m∂-lacZ and is immunolabeled with anti-β-gal (Notch activation) and anti-Elav (R-cell nuclei). The genotype is: Ral^EE1 FRT19A/ubi-gfp FRT19A; ey-gal4, UAS-flp/+; m∂-lacZ/+. The broken line is the equator, the furrow is on the left. The numbers are R3/R4 cells. Circled ommatidia are enlarged in B. Scale bar: 10 μm. (B) Enlargements of a Ral⁻/Ral⁻ (E/P) ommatidium, and two kinds of mosaics are shown. E, equatorial cell; P, polar cell. Scale bar: 5 μm. (C) Fractions of normally constructed (wild-type) and mutant (symmetrical or reversed) ommatidia in each of the four genotypic classes indicated. The scale at the top is in percent. E, equatorial cell; P, polar cell; + is Ral⁺; − is Ral⁻. (D) Raw data for graph in C. EQ, equatorial cell; PO, polar cell.

Fig. 6.

Ral expression in eye discs. (A) An eye disc is shown on the left that expresses m∂-lacZ and ngfp under the control of a Ral enhancer trap, and immunolabeled with anti-β-gal and phalloidin. The genotype is: Ral^PG69/+; m∂-lacZ/UAS-ngfp. Enlargements of the three different classes of ommatidia indicated are on the right. Scale bars: in the large panel and in B, 20 μm; in the small panels, 5 μm. Arrow indicates morphogenetic furrow. (B) A Ral^PG89/Ral⁺ eye disc containing a clone of Ral^PG89/Ral^PG89 cells (outlined at top), immunolabeled with anti-Ral. The genotype is Ral^PG89 FRT19A/ubi-gfp FRT19A; ey-gal4, UAS-flp/+. (C) Wild-type eye discs and eye discs that overexpress Ral and ngfp in R2/R5 and R3/R4 under the control of ro-gal4 are shown. The genotype is: ro-gal4/+; UAS-Ral/UAS-ngfp. The discs are immunolabeled with anti-Ral and phalloidin. Scale bar: 50 μm in a,b; 5 μm in a′,a″,b′,b″. Arrows indicate morphogenetic furrow. (D) A z-section of a developing ommatidium. A, apical membrane; b, basal membrane; c, cone cell; numbers are R-cells. The horizontal line represents the depth of the xy images on the right. A Ral⁺ (wild-type) eye disc that expresses m∂-lacZ and an eye disc that also overexpresses Ral under control of a Ral enhancer trap (Ral^PG69; UAS-Ral) are shown. The genotypes are: m∂-lacZ/+ (wild type) and Ral^PG69/+; UAS-Ral/m∂-lacZ. Each is immunolabeled with anti-β-gal, anti-Ral and phalloidin. The numbers are R3 and R4. We counted the number of R3/R4 pairs in which there were more Ral⁺ puncta in R3 (R3>R4), where the numbers were similar (R3~R4), and where there were more in R4 (R4>R3) in wild-type and Ral^PG69; UAS-Ral eye discs. In five wild-type discs: R3>R4 (211/290), R3~R4 (31/290), R4>R3 (48/290). In six Ral^PG69; UAS-Ral discs: R3>R4 (161/219), R3~R4 (21/219), R4>R3 (37/219). Scale bar: 5 μm.

Fig. 7.

Control of Ral expression by Fz and Notch. (A) Eye discs are shown containing fz⁻ gfp⁻ clones (white outlines). The discs express nuclear β-gal under Ral control and were immunolabeled with anti-β-gal and anti-Svp (labels R3 and R4). The genotype is: Ral^PG69/eyFLP; UAS-nlacZ/+; fz^P21 fz2^C1 FRT2A/ubi-gfp FRT2A. Arrows indicate the furrow. The disc in a,a′ contains several clones. In b,b′, a single clone from a different disc is shown enlarged. R3 and R4 are indicated. The asterisk indicates a fz⁺ R4 (β-gal⁺) of a mosaic pair where the R3 is fz⁻ (β-gal⁻). (B) An eye disc that expresses sev-fz, m∂-lacZ and ngfp under Ral control, immunolabeled with anti-β-gal and phalloidin, is shown. The genotype is: Ral^PG69/+; sev-fz/+; m∂-lacZ/UAS-ngfp. Enlargements of the three circled ommatidia are on the right. The furrow is on the right. E, equatorial cell; P, polar cell. In a and b, Ral expression is P>E, and in c, it is E~P. (C) Quantification of the three types of Ral expression patterns in ommatidia in sev-fz or sev-N^act discs. ***P<0.001, **P<0.01, *P<0.04. Data are mean + s.e.m. (D) Quantification of the four types of m∂-lacZ expression pattern within each of the three types of Ral expression pattern in the sev-fz discs in C. (E) An eye disc is shown that expresses sev-N^act, m∂-lacZ and ngfp under Ral control, immunolabeled with anti-β-gal and phalloidin. Circled ommatidia representing each of the three classes of Ral expression pattern are enlarged on the right. The furrow is on the left. The genotype is: Ral^PG69/+; sev-N^act/+; m∂-lacZ/UAS-ngfp. (F) A table showing quantification of dominant genetic interactions between Ral mutants, and also null alleles of Notch and Delta, with sev-fz, and Ral mutants with sev-N^act. Data were obtained from sections of adult eyes. omm., number of ommatidia; nd, not determined; *, reversed ommatidia could not be scored because the eye field was too disorganized. Scale bar: 40 μm in A, parts a,a′; 20 μm in A, parts b,b′, and in B,E; 10 μm in the enlargements in B,E.

Fig. 8.

Ral overexpression in pre-R4 has only a subtle effect on R3/R4 determination. (A,A′) An eye disc is shown containing a gfp⁺ cell clone (outlined in A′) that overexpresses Ral. The eye disc also expresses m∂-lacZ and is immunolabeled with anti-Svp and anti-β-gal. The arrows indicate Ral-overexpressing (gfp⁺) R4s (β-gal+), the R3s of which (β-gal⁻) do not overexpress Ral (gfp⁻). The genotype is hs-flp, tub-gal4, UAS-gfp/+; UAS-Ral^wt/m∂-lacZ; FRT82B/FRT82B tub-gal80. Larvae (2nd and 3rd instar) were heat-shocked for 1 hour at 37°C. (B) An analysis of R3/R4 determination in pairs mosaic for wild-type and Ral^CA-overexpressing cells. Pooled results from six mosaic eye discs are shown. EQ, equatorial cell; PO, polar cell. The genotype is hs-flp; Act5C>stoP>gal4, UAS-gfp/+; UAS-Ral^CA/m∂-lacZ. Scale bar: 10 μm.

Fig. 9.

Ral controls R3/R4 cell fate in fz⁻ cells. (A-B″) Two separate gfp⁺ fz⁻ clones (outlined) in an eye disc are shown in A-A″ and B-B″. The eye disc also expresses m∂-lacZ and is immunolabeled with anti-β-gal and anti-Svp. The genotype is hs-flp tub-gal4, UAS-gfp/+; m∂-lacZ/+; fz^P21 fz2^C1 FRT2A/tub-gal80 FRT2A. (C-D″) Two separate gfp⁺ fz⁻ clones that overexpress Ral (outlined) in eye discs are shown in C-C″ and D-D″. The eye disc also expresses m∂-lacZ, and is immunolabeled with anti-β-gal and anti-Svp. The genotype is hs-flp tub-gal4, UAS-gfp/+; m∂-lacZ/UAS-Ral^wt; fz^P21 fz2^C1 FRT2A/tub-gal80 FRT2A. Scale bar: 10 μm. To generate clones, larvae (2nd and 3rd instar) were heat-shocked for 1 hour at 37°C.

Fig. 10.

Model for Ral function in R3/R4 cell fate decision. Fz activation in the equatorial cell results in asymmetric Notch activation through two proposed mechanisms: (1) promotion of Delta signaling through transcriptional upregulation of Delta and neur, and their repression in the polar cell; and (2) direct repression of the Notch receptor through relocalization of a Fz/Dsh complex to the side of the equatorial cell plasma membrane adjacent to the polar cell. We have presented evidence for a distinct Ral-dependent mechanism in the equatorial cell. Ral transcription is upregulated in response to Fz activation, and Ral activity represses ligand-independent Notch activation. Notch activation in R4 does not repress Ral transcription in the polar cell.