Zip is an interacting partner of Notch

In an effort to identify novel components involved in Notch signaling and its regulation, we carried out two independent protein interaction screens, one based on the identification of cellular protein complexes using immunoprecipitation followed by mass spectrometry and other based on yeast two-hybrid system. Both the screens identified Zip as an interacting partner of Notch. For immunoprecipitation, tissue lysate was prepared from hyperplastic wing imaginal discs over-expressing Notch-ICD under vg-GAL4 driver and from wild type wing discs as well that served as the control. Anti-Notch antibody was used to precipitate Notch along with its interactors from both the lysates. The protein bands exclusively present in the lane with over-expressed Notch lysate and absent in the control lane were excised from the SDS-PAGE and were processed for mass spectrometry. One of them was identified as Zip, a 205 kDa protein as a novel interactor of Notch (Fig. S1). Zip was also recovered in a standard yeast two-hybrid screen in which approximately 6 × 106 cDNAs from a random primed Drosophila 0–24 h embryonic library were screened with a cDNA corresponding to amino acids 1765–1895 of Drosophila Notch fused in frame to the LexA DNA binding domain used as a bait [24]. In the same screen, multiple positive clones of Su(H), a well-established binding partner of Notch-ICD, were also identified, which validated our approach. Sequence analysis of three identified positive clones of Zipper (Zip) revealed that the carboxy-terminal part of Zip (amino acids 1692–1972) binds to Notch-ICD. In addition, we have also carried out a large-scale proteomic analysis of Zip based on co-immunoprecipitation of Zip along with its interacting proteins followed by mass spectrometry (MS). Notch was identified as one of the many interacting partners in our MS analysis. Along with Notch, many other cytoskeletal components and endocytic machinery proteins were also identified in the MS analysis (Fig. S7).

Physical interaction of Notch with Zip was further validated by co-immunoprecipitation studies. Protein lysate was prepared from adult head tissues co-expressing full-length Notch (Notch-FL) and GFP-Zip driven by GMR-GAL4. Anti-Notch antibody was used to immunoprecipitate Notch along with its interacting partners in a complex which were fractionated on SDS-PAGE. Co-immunoprecipitated GFP-conjugated Zip protein was detected on a western blot using anti-GFP antibody (Fig. 1A). In addition, we also used anti-Zip antibody that detected both endogenous and over-expressed GFP-Zip that appeared as doublet band on the western blot (Fig. 1A). In the same set of experiments, we also demonstrated the physical interaction of over-expressed Zip with endogenous Notch protein. As in case of over-expressed Notch, Zip was also detected as a physically associating partner of the endogenous Notch with anti-GFP as well as anti-Zip antibody (Fig. 1A). To rule out the presence of Zip in our western blot due to non-specific binding with AG beads, we also performed immunoprecipitation experiment without adding primary antibody. No expression of Zip could be seen in the negative control compared to the presence of a doublet band of endogenous and over-expressed GFP-tagged Zip in the experimental lanes (Fig. S1C). Conversely, we used anti-GFP antibody to immunoprecipitate GFP-tagged Zip along with its binding partners and Notch FL was found to be co-immunoprecipitated with Zip which was detected using anti-Notch antibody on the western blot (Fig. 1A). In the same set-up, we also performed co-immunoprecipitation studies using protein lysate from only Notch FL over-expressed condition. Here, anti-GFP antibody could not immunoprecipitate endogenous Zip (absence of over-expressed GFP-Zip). Hence, we did not observe presence of Notch upon immunoblotting with anti-Notch antibody (Fig. 1A). These results confirmed that Notch is indeed physically associated with Zip.

Fig. 1

Physical interaction between Notch and Zip. A Zip was identified as an interacting partner of Notch. Co-immunoprecipitation was carried out with head tissue lysates over-expressing GFP-Zip and Notch FL. M indicates the marker lane, (+) symbol indicates the presence and (−) shows the absence of the specified reagent. Notch-FL and endogenous Notch immunoprecipitated endogenous and GFP-tagged-Zip that was detected by anti-GFP and anti-Zip antibody on the western blot. Star marks indicate the bands of GFP-tagged Zip and endogenous Zip. No GFP-Zip protein bands were observed in the negative control (Fig. S1C). In the other direction, anti-GFP immunoprecipitated Notch-FL that was detected by anti-Notch antibody on the western blot. Lower blots show the presence of the specified protein bands in the experimental and the control lysates. B–D Zip colocalizes with Notch in the over-expressed condition in the cytoplasmic compartment. Zip and Notch-FL were co-expressed under vg-GAL4. GFP-tagged Zip forms cytoplasmic aggregates upon over-expression and colocalizes with Notch-FL puncta on the cytoplasmic membrane. Anti-Notch antibody was used for the detection of over-expressed full-length Notch. Panel D is the merged image of B, C. The insat image in the panel D represents enlarged view of the image showing extent of colocalization between Notch and Zip. E–G Zip colocalizes with Notch in wild type wing discs. Merged image of panel G shows that Zip colocalizes with Notch on the cytoplasmic membrane. The antibody used for detection of wild type Zip in panel E is anti-Zip and for Notch detection in panel F is anti-Notch. The insat image in the panel G represents enlarged view of the images showing extent of colocalization between these two proteins. Scale bar: 20 μm

To corroborate our physical interaction experiments of Notch and Zip, we performed colocalization experiments through immunostaining using anti-Notch and anti-Zip antibodies. We checked their localization in GFP-tagged Zip and Notch-FL co-expressed larval wing imaginal discs. We observed that Zip colocalized with Notch in over-expressed condition on the cell membrane and in the cytoplasm (Fig. 1B–D, Fig. S1G–G″). As in case of colocalization of Notch and Zip in over-expressed condition, endogenous Zip and Notch also colocalized in the same cytoplasmic compartment in the wild-type wing imaginal disc (Fig. 1E–G). In the over-expressed condition, GFP-tagged Zip was found to form large cytoplasmic aggregates [29,30,31], and these aggregates colocalized with Notch-FL (Fig. 1B, D) indicating that the two proteins colocalize and perhaps functionally regulate developmental processes together.

zip genetically interacts with Notch pathway components

To address the functional implications of physical interaction between Zip and Notch, we studied the genetic interactions between zip mutants and mutants of Notch pathway components in trans-heterozygous combinations. We used two independent amorphic alleles of zip, zip1, and zip2. We checked their genetic interaction with the Notch mutants, the amorphic allele of Notch, N54l9, and the hypomorphic allele of Notch, Nnd3. A trans-heterozygous combination of zip alleles with N54l9 and Nnd3 resulted in an increased number of flies with wing-nicking phenotype (Fig. 2B–C″). It was observed that 28% of N54l9 mutant flies exhibited wing-nicking phenotype which was increased to 60% and 52% (42/80) when these flies were combined heterozygously with zip1 and zip2 alleles, respectively (Fig. 2B‴). Similarly, 28% of Nnd3 mutant flies exhibited wing-nicking phenotype which was increased to 37% and 35% when these flies were combined heterozygously with zip1 and zip2 alleles respectively (Fig. 2C‴). This indicated a further reduction of Notch signaling upon lowering the dose of zip in the same background thus validating a functional relevance between the two genes. The wing vein thickening phenotype of Delta (Dl5f) was also enhanced by reducing the dose of zip (Fig. 2D–D″). The null allele of deltex (dx), a cytoplasmic modulator of Notch resulted in mild wing-vein thickening. However, we noticed an enhancement in wing vein thickening phenotype in dx hemizygous condition with zip alleles. Since vein thickening is a Notch loss-of-function phenotype, the enhancement of it with zip alleles in the background indicates further decrease of Notch signaling upon lowering the dose of Zip (Fig. 2E–E″). The functional interaction predicted to be caused by the colocalization of Notch and Zip in the same cytoplasmic compartment (as shown in Fig. 1B–G) is evident from genetic interaction between zip and Notch mutant alleles (Fig. 2B–C‴). Additionally, we also checked the genetic interaction with C96-GAL4-driven dominant-negative form of C-terminal Mastermind truncation (MamH), that display a fully penetrant wing-nicking phenotype [32, 33]. Reducing the dose of zip in this background elicited enhanced wing notching combined with reduced marginal bristles. It was observed that the approximate number of bristles exhibited by C96-Mam H flies was 40 on the entire posterior margin which was reduced to 19 and 21 when the flies were heterozygously combined with zip1 and zip2 alleles (Fig. 2F‴). This modulation in the wing phenotype could be attributed to a compromised dose of Zip in the cytoplasm that ultimately affects the downstream Notch signaling processes, resulting in an enhanced wing-nicking and loss of marginal bristles in the background of C96-GAL4-driven dominant-negative Mastermind (Fig. 2F–F″). The genetic interaction between zip and Notch pathway components demonstrated that effects caused by decreased Notch signaling were further enhanced when zip mutants were brought in trans-heterozygous combination indicating that Zip positively regulates Notch signaling and its loss results in the reduction of Notch signaling (Fig. 2).

Fig. 2

Genetic interaction of zip with Notch pathway components. Representative wings from different Notch pathway component mutants are shown in first column and in trans-heterozygous condition with zip mutants zip1 and zip2, are shown in second and third column respectively. Wing from wild type is shown in A and from zip alleles are shown in A′–A″. B-C″ Wings from heterozygote N54l9 (B) and hemizygous Nnd−3 (C) shows nicking phenotype which were further increased in number in trans-heterozygous combination with loss-of-function alleles of zip, zip1 (B′, C′), and zip2 (B″, C″). D–E″ Heterozygous wings of DL5f (D) and hemizygous dx (E) showed vein thickening phenotype which was enhanced in trans-heterozygous condition with zip mutants (D′, D″ and E′, E″) respectively. (F–F″) Representative wings of loss-of-function allele of Mastermind driven by C96-GAL4 showed serrated wing margin phenotype, which was further enhanced by combining zip alleles trans-heterozygously. (B‴, C‴ and F‴) Graphs representing the percentage of wings showing nicking phenotype (B‴ and C‴) and approximate number of bristles on the posterior margin of the wing (F‴) in trans-heterozygous combination of zip mutants with the components of Notch signaling. (B‴ and C‴) All experiments were performed in triplicates (n). For consistency a total of 80 flies/vial was observed (B‴) No. of wings with nicking phenotype N54l9/ + : n1 = 22/80, n2 = 20/80 and n3 = 24/80; N54l9/zip1: n1 = 48/80, n2 = 42/80 and n 3 = 52/80; N54l9/zip2: 42/80, n2 = 40/80 and n3 = 46/80. (C‴) No. of wings with nicking phenotype Nnd3: n1 = 22/80, n2 = 22/80 and n3 = 24/80; Nnd3/zip1: n1 = 30/80, n2 = 27/80 and n3 = 32/80; Nnd3/zip2: 28/80, n2 = 20/80 and n3 = 36/80. F‴ The genotype of the flies mentioned on the X-axis of the graph are as follows: C96-GAL4/UAS-Mam H, C96-GAL4/UAS-Mam H + zip1, C96-GAL4/UAS-Mam H + zip2. Unpaired t-test was performed to determine p-value (**p < 0.01, ***p > 0.001). Scale bar: 3 cm

The loss-of-function of zip renders wing phenotypes similar to Notch mutant phenotypes and perturbs the expression pattern of Notch targets, Cut and Deadpan

The integration of Zip with Notch signaling was further verified by loss-of-function studies in which Zip was downregulated using UAS-zip-RNAi and UAS-GFP-zip DN under various wing-specific GAL4 driver lines. vg-GAL4-driven UAS-zip-RNAi displayed slightly bent, crumpled wing phenotype with serration on the lower margin of the wing blade. Some of these wings also exhibited irregular marginal bristles with thickened veins and extra vein material. Similarly, abrogating Zip using UAS-GFP-zip DN in the ventral domain of the wings using the same vg-GAL4 driver resulted in wing-nicking phenotype (Fig. S2A–A′). Down-regulating Zip using UAS-zip-RNAi in the dorsal region of the wing with ap-GAL4 yielded outward-directed erect wings with a crumpling phenotype. These wings also harbored several other phenotypes including irregular marginal bristles, extra vein material, and extra cross-veins. Similarly, eliminating Zip using UAS-GFP-zip DN in the dorsal region of the wing resulted in pupal lethality (Fig. S2B–B″). en-GAL4-driven UAS-zip-RNAi resulted in 100% pupal lethality. However, blocking Zip using Zip-DN in this region with same Gal4 driver resulted in wings harboring various phenotypes similar to Notch loss-of-function that included wing-nicking, mispatterning of wing hair, disruption of vein pattern and the cross-veins, and wing blister (Fig. S2C–C″). C96-GAL4-driven UAS-zip-RNAi resulted in slightly bent wings with fully penetrant serration on the lower margin of the wing blade. These wings also exhibited mild crumpling with ectopic bristles on the anterior portion of the wings. Reducing the activity of Zip using UAS-GFP-zip DN in the wing margin resulted in wing-nicking phenotype (Fig. S2D–D″). Reducing the dose of zip on the anterior/posterior boundary using ptc-GAL4 resulted in pupal lethality with very small number of flies eclosing that harbored wing-nicking on the anterior/posterior region of the wing blade (Fig. S2E–E″). Down-regulating zip at the anterior/posterior boundary using dpp-GAL4 displayed disrupted wing bristles and mild wing-nicking at A/P boundary (Fig. S2F–F″). These phenotypes shown by downregulation of zip using different GAL4 lines are reminiscent to Notch loss-of-function phenotypes, indicating that lowering the dose of Zip results in further reduction of the Notch signaling (Fig. S2G).

The modulation of the Notch signaling caused by down-regulation of Zip was further validated in the wing imaginal discs by examining the expression level of Notch downstream targets, Cut, and Dpn. Cut encodes a homeodomain transcription factor that bears a significant structural and functional similarity with several vertebrate proteins. Notch is required for the activation of Cut in a cell-autonomous manner [34]. Similarly, activated Notch directs the expression of Dpn by binding to the Notch-responsive enhancer present in the regulatory region of Dpn [35]. ptc-GAL4 and en-GAL4-driven UAS-GFP-zip DN in third instar larvae resulted in the perturbation of the expression of Cut and Dpn. RGB Plot profiles display the decreased fluorescence intensity of Cut and Dpn in the regions with compromised Zip compared to the internal control (Fig. 3A1–D1). The expression of Cut and Dpn has been observed to be normal in the wing imaginal discs of wild type flies (Fig. S3A, B). These observations confirmed that Zip not only positively regulates Notch signaling but also play a vital role in the modulation of Notch downstream targets (Fig. 3). Perturbation of Notch targets, Cut, and Dpn upon down-regulating Zip at A/P boundary (using ptc-GAL4) and in posterior domain (using en-GAL4) was observed in all the discs that were examined (total number of wing discs examined = 30). Image J was used for the intensity profiling where integrated density/area of the domain was used for quantification purpose. A total number of 5 discs were used for quantification in each case which was subjected to unpaired t-test to determine the significance of our findings. Mean intensity for internal control of Cut was 29, whereas that of en-GAL4 > UAS-GFP-zip DN was 20.6. Mean intensity for internal control of Dpn was 33.4, whereas that of en-GAL4 > UAS-GFP-zip DN was 28.2. In case of ptc GAL4-driven UAS-GFP-zip DN, mean intensity of internal control for Cut expression was 31.2, whereas that of Zip DN domain was 15.2. For Dpn, mean intensity for internal control was 38, whereas UAS-GFP-zip DN was 22.

Fig. 3

A–D Reducing the dose of Zip results in lowered Notch signaling. Representative wing discs display the expression pattern of Cut and Dpn upon eliminating Zip on the A/P boundary and in the posterior region using ptc-GAL4 and en-GAL4 respectively. Abrogating Zip in the patched and engrailed domain using UAS-GFP-zip DN results in the perturbed expression of Cut (A, C) and Dpn (B, D) in these regions. A′–D′ Merged images showing the expression pattern of Cut and Dpn along with the patched and engrailed domain marked by GFP-tagged Zip DN. A1–D1 RGB plot profiles indicating the intensity of Cut and Dpn in A, B, C and D, respectively. E–J′ Loss of Cut and Dpn expression due to Zip DN on the A/P boundary using dpp GAL4 (E and H respectively) is rescued by over-expressing Notch ICD in the background (G and J respectively). Over-expressing UAS-Notch-ICD with dpp-GAL4 resulted in ectopic expression of Cut and Dpn throughout A/P boundary (F and I respectively). (E′–G′ and H′–J′) Merged images showing the expression of Cut and Dpn in dpp domain with GFP-tagged UAS-zip-DN (E′ and H′), and GFP-tagged UAS-zip-DN + UAS-Notch ICD (G′ and J′). (F′ and I′) Merged images showing the expression of Cut and Dpn along AP boundary with over-expressed Notch ICD. Absence of GFP indicates the absence of GFP-zip DN. (D″–D‴) Graphs representing the intensity profiling of Cut and Dpn upon abrogation of Zip at the A/P boundary (D″) and in the posterior region (D‴) of the wing imaginal discs. Unpaired t-test was performed to determine the p-value (*p < 0.05, ***p < 0.001). D″″ Western blot showing the expression of Cut, Dpn and internal control Tubulin (Tub). Scale bar: 20 μm

To verify whether the diminished expression of Cut and Dpn downstream to Zip-DN is mediated by reduced activity of Notch signaling, we supplied Notch-ICD in the Zip compromised background. Over-expression of UAS-GFP-zip DN using dpp-GAL4 alone leads to Cut and Dpn reduction in the D/V boundary straddling the A/P domain of dpp-GAL4 (Fig. 3E, H), whereas both Cut and Dpn were found to be ectopically increased upon over-expressing Notch ICD in the Dpp domain (Fig. 3F, I). When UAS-Notch ICD was over-expressed in the background with dominant-negative zip, the Notch targets were found to be rescued thus indicating that loss of Cut and Dpn resulting due to perturbed Zip is Notch-mediated (Fig. 3G, J), thus highlighting the role of Zip in modulating the activity of Notch signaling. To validate the loss of Notch signaling targets in Zip DN background, we prepared protein lysates from wing discs of en-GAL4 > Oregon R and en-GAL4 > UAS-GFP-zip DN and checked for the expression of Cut and Dpn using Western blotting. Both Notch downstream targets were found to be reduced in Zip DN condition compared to the control condition (Fig. 3D″″).

The loss-of-function of zip leads to cell surface accumulation of Notch receptor and endocytic component Rab5

Since down-regulation of Zip leads to reduced expression of Notch targets, we wanted to check the status of Notch receptor in Zip compromised background. Down-regulation of Zip in the posterior domain of the wing disc in UAS-GFP-zip DN using en-GAL4 driver, resulted in altered Notch localization. The wild type membranous Notch appeared to be accumulated in Zip compromised domain. These studies indicated that loss of Zip might result in perturbation of Notch receptor processing subsequently leading to down-regulated Notch signaling (Fig. S4A–A″). An accumulation of Notch in Zip compromised domain of the wing discs was observed in 100% of the wing discs that were examined (total number of discs examined = 20). A quantification of the membranous Notch expression was used as a readout for accumulated Notch at the cell surface. About 10 discs were used for the quantification purpose using Image J. Mean intensity of the wild type expression of Notch in the internal control was 22, whereas that in the posterior domain with compromised Zip was 31.5. To verify the accumulation of Notch at the cell surface, we also checked the colocalization of accumulated Notch with membrane marker actin. Actin associates with the plasma membrane and provide mechanical support, determine cell shape, etc. [36]. Here, accumulated Notch in the posterior region with compromised Zip using UAS-zip RNAi with en-GAL4 was observed to colocalize with Phalloidin marked actin (Fig. 4A–B″). Western blotting also showed a higher band intensity of Notch FL in the Zip compromised condition compared to the wild type that served as the control (Fig. 4A‴).

Fig. 4

Loss of Zip leads to accumulation of Notch receptor and Rab5. A–B′″ Zip downregulation with UAS-zip RNAi in the posterior region of the wing disc using en-GAL4 resulted in an accumulated expression of the Notch receptor (A′) and actin (A) on the cell membrane compared to its endogenous expression in the anterior domain (A). Panel A″ represents a merged image of A and A′ where Notch was observed to colocalize with Phal. Panel B–B′ shows the expression of Phal, Notch at a higher resolution. Panel B″ represents merged image of B and B′. (A′″) Western blotting showing the expression of Notch in control and Zip compromised condition. C–C″ Down-regulation of Zip using UAS-zip RNAi with en-GAL4 resulted in accumulated Notch (C) and Rab5 (C′) in the posterior region of the wing disc. Panel C″ is the merged image of C and C′. Scale bar: 20 µm. B‴ Graph representing the accumulation of Notch and Rab 5 upon downregulating Zip in the posterior region. Error bar denotes the error between the mean values of the specified genotypes

Accumulation of the Notch receptor at the cell surface has been correlated with impaired endocytosis of the receptor in the signal receiving cell leading to compromised signaling [37]. Myosin II is implicated in the endocytic process where via force generation, it assists in pulling the clathrin-coated pit along with the receptor toward inside of the cell leading to invagination of the membrane thus initiating the process of early endosome formation [38]. Rab5 belongs to the family of GTPases that regulate trafficking into and between the early endosomes [39]. Hence, to understand the mechanism behind accumulation of Notch receptor in the signal receiving cell upon abrogating Zip, we checked the status of early endosomal marker Rab5 in Zip compromised condition. Down-regulating Zip using UAS-zip RNAi with en-GAL4 revealed an accumulated pattern of Notch at cell surface in the posterior domain of the wing imaginal disc. Similar to Notch accumulation, Rab5 also appeared to be accumulated in the engrailed domain compared to the anterior domain with endogenous Zip. Accumulated Notch was observed to partially colocalize with Rab5 (Fig. 4C–C″). This suggested that perturbation of Zip compromised the formation of early endosome as evident from the accumulated expression pattern of Rab5. Loss of Zip led to loss of pulling force required for the formation of endocytic vesicles indispensable for receptor internalization, leading to accumulated Notch receptor at the cell surface. Our findings suggested that Zip is necessary for Notch receptor internalization in the signal receiving cell leading to activation of the signaling pathway. The accumulation of Notch along with Rab5 was observed in all the wing discs that were examined (total number of wing discs examined = 15). Among them, 5 discs were used for the quantification of the intensity.

To rule out the involvement of the regulatory light chain in the functional role of Zipper in Notch regulation, we also examined the status of Notch in regulatory light chain compromised background. The gene spaghetti squash (sqh) encodes the myosin regulatory light chain in Drosophila. Notch appeared to be unaltered in the wing imaginal discs obtained from null mutant of spaghetti squash (Fig. S4C) indicating that the absence of myosin heavy chain (encoded by zip) is mainly responsible for the accumulation of Notch receptor at the cell surface (Fig. S4).

Over-expression of zip rescues Notch loss-of-function phenotype and this rescue is facilitated by motor domain of Zip

Through epistatic interaction studies, here we wanted to check whether over-expressing Zip in the compromised Notch background can rescue Notch loss-of-function phenotypes. At this end, we over-expressed Zip in larval wing imaginal discs of UAS-GFP-zip individuals and reduced the expression of Notch in the same tissue by dominant-negative Notch using C96-GAL4 driver. Notch, upon being downregulated under C96-GAL4, yielded a highly serrated wing phenotype which was significantly rescued by over-expressing zip in the same background (Fig. 5A–C). This rescue in the wing serration upon upregulating Zip in Notch DN background using C96 GAL4 has been shown via a bar graph (Fig. 5G) where massive nicking of the wing margin has been referred to as “severe serration” (e.g., Fig. 5B) and a lesser nicking of the wing margin has been referred to as “mild serration” (e.g., Fig. 5C). Here, the experiment was performed in two batches (Batch1 and Batch2) where a total number of 80 wings were examined for each case. In batch 1, C96 GAL4-driven UAS-GFP-zip yielded all the wings with no serration, UAS-Notch DN yielded 66 wings with severe serration and 14 wings with mild serration, UAS-GFP-zip + UAS-Notch DN yielded 66 mildly serrated wings (rescued) and 14 severely serrated wings. In batch 2, C96 GAL4-driven UAS-GFP-zip yielded wild type wings with no serration, UAS-Notch DN displayed 67 wings with severe serration and 13 wings with mild serration, UAS-GFP-zip + UAS-Notch DN yielded 60 mildly serrated wings (rescued) and 20 severely serrated wings.

Fig. 5

Over-expression of zip rescues Notch loss-of-function phenotype. A–C C96-GAL4-driven expression of UAS-Notch-DN leads to severe wing-nicking phenotype (B) that is rescued upon expression of zip in the same background (C). Over-expression of zip alone results in wild-type wing (A). Scale bar: 3 cm. G Graph representing the number of wings showing rescue in wing serration on co-expressing zip with Notch-DN using C96-GAL4. The genotype of the flies mentioned on the X-axis of the graph are as follows: C96-GAL4/UAS-GFP-zip, C96-GAL4/UAS-Notch-DN, and C96-GAL4/UAS-GFP-zip + UAS-Notch-DN. D–F′ Representative wing discs showing the expression pattern of Cut. Downregulating Notch in C96-GAL4 region resulted in loss of expression of Cut on the DV boundary (E) which was mildly rescued upon over-expressing zip in the same background (F). Over-expression of zip alone resulted in the expression of Cut similar to wild type (D). D′, F′ Merged images showing the expression pattern of Cut under C96-GAL4-driven GFP-zip and GFP-zip with Notch DN, respectively. Similarly, E′ represents the merged image showing the expression pattern of Cut under C96-GAL4-driven UAS-Notch DN. Absence of GFP in this image denotes the absence of GFP-zip. Scale bar: 20 μm. H–H‴ Zip interacts with Notch via motor domain. H Co-expression of full-length UAS-zip with UAS-Notch-DN using C96-GAL4 lead to rescued wing serration caused by Notch loss of function alone. However, this rescue failed to occur when UAS-Notch-DN was co-expressed with UAS-Myo II Neck Rod having truncated motor domain (H′), UAS-Myo II-Rod (H″) and UAS-Myo II-Rod (delta Nterm58) (H‴) having truncated head and neck domain respectively. Scale bar: 3 cm. I A depiction of full-length UAS-GFP-zip and domain truncation stocks. UAS-GFP-zip has all the domains including head domain, neck domain harboring binding sites for essential and regulatory light chains, and rod domain. UAS-Myo II Neck Rod has truncated head domain. UAS-Myo II Rod has truncated head and neck domains. UAS-Myo II-Rod (delta Nterm58) has truncated head, neck and deletion of 58 amino acids of Rod domain. J Graphical representation of the percentage of wings showing mild and severe wing serration upon co-expression of UAS-Notch-DN with full-length zip and its truncation domains. The genotype of the flies mentioned on the X-axis of the graph is as follows: C96-GAL4/UAS-Notch-DN + UAS-GFP-zip, C96-GAL4/UAS-Notch-DN + UAS-Myo II Neck Rod, C96-GAL4/UAS-Notch-DN + UAS-Myo II Rod, C96-GAL4/UAS-Notch-DN + UAS-Myo II-Rod (delta Nterm58)

This was further validated in wing discs where we wanted to assess the expression of the Notch signaling target Cut upon over-expressing zip in the lowered Notch background. Over-expression of zip alone using C96-GAL4 resulted into the expression of Cut almost like wild type (Fig. 5D). Downregulating Notch signaling using dominant-negative form of Notch under C96-GAL4 resulted in complete loss of Cut staining in the wing discs (Fig. 5E). However, the loss of Cut expression was mildly rescued (Fig. 5F) upon over-expressing zip in the same background indicating an important role of Zip in Notch signaling (Fig. 5D–F).

Subsequently, we wanted to investigate the functional domain of Zip that is responsible for the rescue of the serrated wing phenotype caused by Notch loss-of-function. Several studies suggest that not all the domains of non-muscle myosin II are required for various biological processes [40]. It has been observed that some processes require the contractility based function of non-muscle myosin II while other processes take place normally even if the motor domain is perturbed by amino acid replacements [41,42,43]. Hence, we wanted to characterize the functional domain involved in Notch and Zip interaction that would subsequently lead to a significant rescue of the nicked wing phenotype of Notch dominant-negative individuals. In Drosophila, separate genes encode each subunit of non-muscle myosin II: zip encodes the heavy chain (zip/MyoII), spaghetti squash encodes the regulatory light chain (sqh/RLC) and mlc-c encodes the essential light chain (mlc-c/ELC) [16, 17, 40, 44]. Zip i.e., each individual heavy chain consists of (1) a globular N-terminal motor or head domain (~ 800 amino acids) that contains the ATP and actin-binding sites; (2) a neck domain (~ 50 amino acids) composed of two IQ motifs that bind one ELC and one RLC; (3) a coiled coil or rod domain (~ 1100 amino acids) composed of heptad repeats; and (4) a short C-terminal segment, termed the tailpiece of ~ 34–47 amino acids in length [45]. Truncation alleles with truncated domains of non-muscle myosin II was used for the study of the interaction between Zip domain and Notch. UAS-Myo II-Neck-Rod truncation allele lacks only the motor domain, UAS-Myo II-Rod encompasses the entire rod domain, thus lacking the motor domain and the neck region and UAS-Myo II-Rod (delta Nterm58) lacks only the first 58 amino acids of the rod domain along with motor and neck domains (Fig. 5I) [40]. It was observed that no rescue of the wing-nicking phenotype could occur when UAS-Notch-DN was co-expressed with the truncated allele, UAS-Myo II-Neck-Rod, UAS-Myo II-Rod and UAS-Myo II-Rod (delta Nterm58) (Fig. 5H′–H‴). Our findings indicated that the truncation of motor domain fails to rescue the nicked wing phenotype caused by Notch loss-of-function suggesting that motor domain is indispensable for the function of Zip in regulation of Notch.

zip synergises with Notch

It is apparent from genetic interaction experiments that zip modulates Notch signaling activity. Hence, we wanted to explore the integrative effect of Zip in Notch signaling. To investigate this, we over-expressed Notch-FL and zip together using vg-GAL4 driver. Over-expression of zip alone resulted in wild-type wing phenotype and over-expressed Notch-FL alone resulted in multiple wing phenotypes, such as wing crumpling, wing blisters, ectopic outgrowth, vein disorganization, and ectopic marginal bristles. Over-expression of both zip and Notch-FL together resulted in enhancement of only Notch-FL over-expression induced wing phenotypes (Fig. S5A–C). Massive wing disorganization, such as wing duplication, large wing blisters, severe vein disorganization, etc., was observed upon co-expression of Notch-FL and zip (Fig. S5C).

In order to check the synergistic effect of Notch-FL and Zip on the wing phenotype using other GAL4 driver line, we co-expressed UAS-Notch-FL and UAS-GFP-zip in the posterior wing compartment using en-GAL4 driver. Zip yielded almost wild type wings when over-expressed using en-GAL4. Notch-FL over-expression in the posterior domain resulted in multiple wing phenotypes including fifth vein shortening, extra vein material, wing crumpling and blisters. When Zip was co-expressed with Notch-FL, it culminated in 100% pupal lethality, thus confirming a significant synergistic effect of Notch and Zip (Fig. 6A–C).

Fig. 6

A–I′ Zip synergises with Notch-FL to lead to pupal lethality and increase in the ectopic expression of Cut and Wg on D/V boundary. A–C Zip synergy with Notch lead to pupal lethality. A Over-expression of zip by en-GAL4 resulted in wild type wings whereas Notch-FL over-expression in the posterior domain resulted in disruption of wing morphology with shortening and thickening of fourth and fifth vein, presence of extra vein material, wing crumpling, and loss of cross veins (B). Flies failed to emerge from the pupal case when UAS-Notch–FL was co-expressed with UAS-zip resulting in 100% pupal lethality (C). Scale bar: 3 cm. D–I′ Representative wing discs show the expression patterns of Cut and Dpn. Ectopic expression of Cut (H) and Dpn (I) gets enhanced upon co-expression of zip with Notch (FL) under en-GAL4 in contrast to over-expressed zip that shows wild-type expression of Cut and Dpn (D, E) and over-expressed Notch-FL that shows mild ectopic expression of Cut and Dpn (F, G). D′, H′ are the merged images showing expression pattern of Cut under en-GAL4-driven expression of GFP-Zip and GFP-zip with Notch-FL, respectively. Similarly, E′, I′ are the merged images showing the expression pattern of Dpn using GFP-zip and GFP-zip with Notch-FL, respectively, in the posterior region. F′, G′ panel represent the expression pattern of Cut and Dpn respectively under Notch (FL) driven by en-GAL4. Absence of GFP in this image represents absence of GFP-zip. J, K Graphs representing the intensity profiling of Cut and Dpn in en-GAL4 > UAS-zip, en-GAL4 > UAS-Notch FL and en-GAL4 > UAS-zip + UAS-Notch FL. Scale bar: 20 μm

To further validate the synergistic interaction between Notch and Zip, we analyzed the levels of Notch targets, Cut and Dpn. Over-expressing GFP-tagged zip alone using en-GAL4 driver did not result in any significant increase in the expression of Cut and Dpn. Over-expression of only Notch-FL in the posterior compartment of wing disc resulted in increased Cut and Dpn expression. However, co-expression of zip and Notch-FL under en-GAL4 resulted in an elevated level of Cut and Dpn in the posterior domain of wing imaginal discs (Fig. 6D–I′). Image J was used for the intensity profiling. A total of 23 discs were examined and all of them showed consistent results. 6 discs were used for the quantification. Mean intensity for Cut expression in en-GAL4-driven > UAS-GFP-zip was 19.4, for UAS-Notch FL was 21.2 and for UAS-GFP-zip + UAS-Notch FL was 26. Mean intensity for Dpn expression in en-GAL4-driven > UAS-GFP-zip was 34, for UAS-Notch FL was 44.5 and for UAS-GFP-zip + UAS-Notch FL was 52.17.

At this point, it was interesting to explore the synergistic interaction between Zip and activated Notch. Hence, we explored the co-operative effect of zip with processed Notch (Notch-ICD). To analyze this, we co-expressed both Notch-ICD and zip using C96-GAL4 driver. C96-GAL4-driven zip alone did not show any phenotype, whereas C96-GAL4-driven Notch-ICD resulted in wing margin defects with irregular marginal bristles resulting in mild crumpling of the wing. When Notch-ICD and zip were co-expressed together with C96-GAL4 driver, it resulted in increased irregular marginal bristles and severely crumpled wings (Fig. S5D–F).

In order to determine that GAL4 is not a limiting factor in our experimental set-up and there is a similar level of expression of transgenes downstream to single copy of UAS (UAS-GFP zip/UAS-Notch FL/UAS-Notch ICD) in control case and double copy of UAS (UAS-GFP zip + UAS-Notch FL) in experimental condition, we performed immunoblotting to check the expression level of Zip and Notch in protein lysates from UAS-GFP-Zip, UAS-Notch FL, UAS-Notch FL + UAS-GFP Zip and UAS-Notch ICD + UAS-GFP-Zip driven with GMR-GAL4. A similar level of Zip and Notch expression was observed in the transgenes with single UAS and double UAS copy (Fig. S1D′–F′). The graphs depicting a similar band intensity suggest that irrespective of single or double copy of UAS, the expression level of transgenes is almost same in all the cases and that GAL4 is not a limiting factor here (Fig. S1D″–F″). Similarly, we also analyzed the GFP-Zip intensity between en-GAL > UAS-GFP-zip and en-GAL > UAS-GFP-zip + UAS-NotchFL condition (Fig. 6D′, H′; E′, I′) to rule out the potential GAL4 dilution effect. The mean intensities of GFP-Zip between respective genotypes were observed to be 32.5 in Zip alone condition and 37.5 in Zip co-expression with Notch FL. The difference in the intensities between both the genotypes was calculated to be non-significant (p > 0.05), thus indicating that there has been no case of GAL4 dilution effect in our studies (Fig. S1F″).

Interactivity of Zip and activated Notch results into hyperpla