Investigating the genetic circuitry of mastermind in Drosophila, a notch signal effector

Abstract

Notch signaling regulates multiple developmental processes and is implicated in various human diseases. Through use of the Notch transcriptional co-activator mastermind, we conducted a screen for Notch signal modifiers using the Exelixis collection of insertional mutations, which affects approximately 50% of the Drosophila genome, recovering 160 genes never before associated with Notch, extending the previous roster of genes that interact functionally with the Notch pathway and mastermind. As the molecular identity for most recovered genes is known, gene ontology (GO) analysis was applied, grouping genes according to functional classifications. We identify novel Notch-associated GO categories, uncover nodes of integration between Notch and other signaling pathways, and unveil groups of modifiers that suggest the existence of Notch-independent mastermind functions, including a conserved ability to regulate Wnt signaling.

Figure 1.—

Screen designs and primary screen validation. (A) Primary screen for insertions that modify the C96-GAL4, UAS-MamN (C96-MamN) wing phenotype. C96-MamN individuals exhibit a fully penetrant, dosage-sensitive, dominant wing phenotype (Helms et al. 1999), similar to those associated with loss-of-function mutations in Notch and wingless (Smoller et al. 1990; Helms et al. 1999). Exelixis P-element or piggyBac males were crossed to C96-MamN females. F₁ cultures exhibiting C96-MamN wing-notching enhancement or suppression were selected and retested: 219 enhancers and 385 suppressors were isolated. Three insertions caused lethality and three additional inserts could not be characterized. (B) Control w¹¹¹⁸ and (C) C96-MamN wings. D–J provide examples of the range of C96-MamN modifier phenotypes (materials and methods). Shown are female wings trans-heterozygous for C96-MamN; interacting loci, stock IDs, and identified genes are listed. Strong, moderate, and weak C96-MamN suppressors [(D) d05358 and CG14709; (E) c06331 and CG8090; and (F) c02035, respectively] and enhancers [(G) c06428 and CG14767; (H) d00059 and klumpfuss (klu); and (I) d07432 and Sin3A, respectively] are shown. (J) Secondary screens for Notch-pathway specific (NIs) and MamN-specific (MSIs) C96-MamN interactors. [y w^a nd³; C96-GAL4], [N^55e11/FM6; C96-GAL4], [y w Ax¹⁶; C96-GAL4], or [w dx¹⁵²; C96-GAL4] females were crossed to males as in A (nd³ illustrates the strategy used for each allele). In a C96-GAL4 background, F₁ cultures exhibiting enhancement or suppression of nd³, N^55e11, Ax¹⁶, and dx¹⁵² wing phenotypes were noted (supplemental Table 1 at http://www.genetics.org/supplemental/).

Figure 2.—

Numbers of known Notch and mastermind genetic interactors. Venn diagrams showing the number of genes associated genetically with Notch and mam. The number of genes known to interact genetically with N (green), mam (yellow), or both (orange) as described in FlyBase (A) and disrupted within the Exelixis collection (B) or (C) identified as C96-MamN modifiers from the Exelixis collection. (C) Of the 87 genes (B), 26 (29.9%) within the collection previously known to be associated genetically with N, mam, or both were isolated as C96-MamN modifiers, representing a statistically significant enrichment for pathway modifiers (Yates' χ² = 28.8, P = 8.14 × 10⁻⁰⁸). Identified C96-MamN interactors were enriched for known genetic interactors of N [24 of 79 genes (30.4%)] (Yates' χ² = 27.3, P = 1.70 × 10⁻⁰⁷) and mam [6 of 15 genes (40.0%)] (Yates' χ² = 9.7, P = 1.90 × 10⁻⁰³). (D) Known Notch and mastermind genetic interactors described by FlyBase and represented within the Exelixis collection. Green, yellow, and orange fonts indicate Notch, mastermind, and overlapping genetic interactors, respectively.

Figure 3.—

Mam-independent wing phenotypes. Wing phenotypes of C96-MamN interactors in a mam⁺ background identified loci that interact with Notch pathway elements. Except for B, all panels show wings prepared from genotypes containing C96-GAL4. (A) w¹¹¹⁸ control strain. (B) A null N allele, N^54l9. (C) d01997, fringe, or CG9119. (D) d08885, fringe. (E) d01047, Beadex. (F) d00627, CG15167. (G) C96-MamN. (H–J) d11666, d10223, d03908, cp309. (K) d09084, CG17390. (L) d04807, lark. (M) d00992, smooth. (N) d04790, CG17836. (O) d03727, CG4612. (P) d03841, kekkon-1. (Q) d04745, cropped. (R) d05894, dorsal. (S) d05415, escargot. (T) d05123, CG7443, or CG9603. (U) d08364, Ecdysone inducible protein 75B. (V) d11052, reaper. (W) d11677, distal antenna. (X) d01629, eyegone. All genotypes are trans-heterozygous for the Exelixis insertion.

Figure 4.—

Validation of the secondary screens. C96-MamN NIs modify wing phenotypes resulting from mutations in Notch pathway genes. Efficacy of secondary genetic assays is illustrated for two NI enhancers [scabrous (sca) and klumpfuss (klu)] and two NI suppressors (CG8090 and CG7370). Notch/sca interactions have been reported (Mlodzik et al. 1990); other interactions define novel Notch-signaling modulators. Wings from genotypes contain the following Notch pathway mutations: (A1–A5) C96-MamN. (B1–B5) y w nd³. (C1–C5) w N^55e11. (D1–D5) y w^a N^Ax16. (E1–E5) w dx¹⁵². Strains in columns are mated with (row 2) sca^d09400, (row 3) klu^d00059, (row 4) CG8090^c06331, and (row 5) CG7370^f06222. All phenotypes are heterozygous for C96-GAL4. Each mutation fails to produce wing phenotypes when combined with C96-GAL4 (not shown). (A1) C96-MamN wing nicking is dominantly enhanced by (A2) sca^d09400 and (A3) klu^d00059 and suppressed by (A4) CG8090^c06331 and (A5) CG7370^f06222 heterozygotes. (B1) nd³ combined with C96-GAL4 causes a highly penetrant wing-nicking phenotype. (B2) sca^d09400 and (B3) klu^d00059 heterozygotes increase the number and severity of nd³ wing notching. (B4) nd³; CG8090^c06331 trans-heterozygous wings were nearly suppressed to wild type while (B5) CG7370^f06222 suppresses nd³ wings to wild type. (C1) Distal wing notching exhibited by N^55e11 heterozygotes increased in (C2) sca^d09400 and (C3) klu^d00059 backgrounds and was dominantly suppressed to wild type at high penetrance by (C4) CG8090^c06331 and (C5) CG7370^f06222. (D1) L4 and L5 longitudinal wing-vein shortening in hemizygous N^Ax16 males. (D2) sca^d09400 suppresses L4, but not L5 phenotypes. (D3) klu^d00059, (D4) CG8090^c06331, and (D5) CG7370^f06222 cause additional L4 and L5 longitudinal wing-vein shortening. (E1) dx¹⁵² hemizygotes display distal wing-vein thickening with occasional wing notching, which is dominantly enhanced by (E2) sca^d09400, (E3) klu^d00059, and (E4) CG8090^c06331. For klu^d00059, the degree and penetrance of wing notching is also more severe. (E5) CG7370^f06222 completely suppresses dx¹⁵² vein defects.

Figure 5.—

Overlapping functional categories of C96-MamN modifiers relative to the FlyBase Notch genetic interactors represented within the Exelixis collection. Venn diagrams show the number of overlapping, statistically significant (Fisher's exact test, P < 0.05) functional categories between (A) the FlyBase Notch genetic interactors represented within the Exelixis collection (ExFBNGInts) and the entire set of NIs. (B) the ExFBNGInts and the novel set of NIs. (C) The ExFBNGInts and the MSIs. (D) The ExFBNGInts and the combined list of functional categories identified in the screen (all C96-MamN modifiers, the entire set of NIs, the novel set of NIs, and the MSIs). In A–D, the ExFBNGInts are shown in green, the overlap in yellow-green, and the other subgroups in yellow.

Figure 6.—

GO analysis. The prevalence of specific GO terms among gene subgroups was compared to their prevalence in the Exelixis collection (materials and methods). Intersubgroup comparisons uncovered novel, statistically significant categories (Fisher's exact test, P < 0.05) unique to each subgroup as identified by DAVID and are summarized for (A) the known FlyBase Notch genetic interactors represented by the Exelixis collection (ExFBNGInts), (B) the 175 NIs, and (C) the 160 novel NIs. (D) Statistically significant functional categories (Fisher's exact test, P < 0.05) uniquely associated with 79 MSIs.

Figure 7.—

RNA-processing genes modify Notch haplo-insufficient wing phenotypes. Several RNA-processing genes were tested for their ability to modify N^54l9 and/or N^55e11 wing phenotypes. Shown at ×10 magnification are the distal portion of the wings where typical Notch haplo-insufficiency wing notching is observed. (A and M) Wild-type Drosophila wings. (B) Notch control N^54l9/+, mutant phenotype present in 20% of wings. In B–L, wings are heterozygous for both N^54l9 and the following alleles: (C) sm¹, (D) sm⁰⁵³³⁸, (E) sm^KG03875, (F) spen^AH393, (G) lark¹, (H) msi¹, (I) msi², (J) orb^dec, (K) orb2^BG02373, and (L) CG17838. (N) Notch control N^55e11/+, mutant phenotype present in 23% of wings. In O–X, wings are heterozygous for both N^55e11 and the following alleles: (O) sm¹, (P) sm⁰⁵³³⁸, (Q) sm^KG03875, (R) spen^XFM911, (S) spen^AH393, (T) lark¹, (U) msi¹, (V) msi², (W) orb2^BG02373, and (X) CG17838.

Figure 8.—

Mastermind enhances activated Armadillo eye phenotypes. Modification of the eye phenotype resulting from mis-expression of activated Arm. All eyes are from females heterozygous for (A–E) GMR-ArmS44Y (Y55) or (F–J) GMR-ArmS56F (F76). Eyes from (A) Y55 and (F) F76 individuals exhibit roughness and are smaller than wild type (not shown). In reducing mam levels, (B and G) mam² and (C and H) mam^sAX8 strongly enhance Y55 and F76 phenotypes while haplo-insufficiency for Notch (D) and (I) N^54l9 and (E and J) N^55e11 exert little, if any, effect. Genotypes are as follows: (A) GMR-ArmS44Y/+, (B) GMR-ArmS44Y/mam², (C) GMR-ArmS44Y/mam^sAX8, (D) N^54l9/+; GMR-ArmS44Y/+, (E) N^55e11/+; GMR-ArmS44Y/+, (F) GMR-ArmS56F/+, (G) GMR-ArmS56F/mam², (H) GMR-ArmS56F/mam^sAX8, (I) N^54l9/+; GMR-ArmS56F/+, and (J) N^55e11/+; GMR-ArmS56F/+. (K) In the absence of the Wnt-3A ligand, transfection of 293T cells with MamL1, dominant-negative MamL1 (DnMamL1), full-length and activated human Notch 1 (Fl-N1 and N1-ICD, respectively) fails to activate the TOP-FLASH reporter (K, untreated). However, MamL1, but not DnMamL1, Fl-N1, or N1-ICD, potentiates Wnt-3A-induced activation of Wnt signaling (∼20 times) as measured by the TOP-FLASH assay (K, Wnt-3A treated). Moreover, this ability of MamL1 to activate Wnt signaling is suppressed by the presence of Fl-N1 and N1-ICD. MamL1, DnMamL1, Fl-N1, and N1-ICD had no effect on the FOP-FLASH luciferase reporter, which contains multiple copies of the mutant form of TCF-binding sites, indicating that the observed effect is specific. Normalized luciferase activities for untreated and Wnt-3A-treated cells are shown.