Zyxin links fat signaling to the hippo pathway

Abstract

The Hippo signaling pathway has a conserved role in growth control and is of fundamental importance during both normal development and oncogenesis. Despite rapid progress in recent years, key steps in the pathway remain poorly understood, in part due to the incomplete identification of components. Through a genetic screen, we identified the Drosophila Zyxin family gene, Zyx102 (Zyx), as a component of the Hippo pathway. Zyx positively regulates the Hippo pathway transcriptional co-activator Yorkie, as its loss reduces Yorkie activity and organ growth. Through epistasis tests, we position the requirement for Zyx within the Fat branch of Hippo signaling, downstream of Fat and Dco, and upstream of the Yorkie kinase Warts, and we find that Zyx is required for the influence of Fat on Warts protein levels. Zyx localizes to the sub-apical membrane, with distinctive peaks of accumulation at intercellular vertices. This partially overlaps the membrane localization of the myosin Dachs, which has similar effects on Fat-Hippo signaling. Co-immunoprecipitation experiments show that Zyx can bind to Dachs and that Dachs stimulates binding of Zyx to Warts. We also extend characterization of the Ajuba LIM protein Jub and determine that although Jub and Zyx share C-terminal LIM domains, they regulate Hippo signaling in distinct ways. Our results identify a role for Zyx in the Hippo pathway and suggest a mechanism for the role of Dachs: because Fat regulates the localization of Dachs to the membrane, where it can overlap with Zyx, we propose that the regulated localization of Dachs influences downstream signaling by modulating Zyx-Warts binding. Mammalian Zyxin proteins have been implicated in linking effects of mechanical strain to cell behavior. Our identification of Zyx as a regulator of Hippo signaling thus also raises the possibility that mechanical strain could be linked to the regulation of gene expression and growth through Hippo signaling.

Figure 1. Zyx and Jub influence wing growth.

All panels show wings from male adult flies with nub-Gal4 UAS-dcr2, and (A) no additional transgenes (control), (B) UAS-RNAi-Zyx³²⁰¹⁸, (C) UAS-RNAi-Zyx²¹⁶¹⁰, (D) UAS-RNAi-Zyx²¹⁶⁰ UAS-Zyx:V5, (E) UAS-RNAi-fat, (F) UAS-RNAi-fat UAS-RNAi-Zyx³²⁰¹⁸, (G) UAS-RNAi-fat UAS-RNAi-Jub³⁸⁴⁴², (H) UAS-Zyx:V5, (I) UAS-RNAi-ex, (J) UAS-RNAi-ex UAS-RNAi-Zyx³²⁰¹⁸, (K) UAS-RNAi-ex UAS-RNAi-Jub³⁸⁴⁴², (L) UAS-dco³, (M) UAS-RNAi-Jub¹⁰¹⁹⁹³, (N) UAS-RNAi-Jub³⁸⁴⁴², (O) UAS-RNAi-Zyx³²⁰¹⁸ UAS-RNAi-Jub³⁸⁴⁴², (P) UAS-dco³UAS-RNAi-Zyx³²⁰¹⁸, (Q) UAS-d:V5, and (R) UAS-d:V5 UAS-Zyx:V5. Yellow arrows point to cross-veins. (S) Average sizes for wings of the indicated genotypes, normalized to the average wing size in controls. 9–12 wings were measured per genotypes; error bars show s.e.m. Even modest differences in wing size were statistically significant (e.g., the 9% increase in UAS-Zyx:V5 versus control is significant by pairwise t test, p<0.0005).

Figure 2. Zyx influences Yki activity in wing discs.

(A–D) show third instar wing imaginal discs. In this and subsequent figures, panels marked by prime symbols show individual channels of the stain to the left. Discs in (A,B) are stained for Diap1 (red) and ex-lacZ (green), with posterior cells marked by GFP (blue), and have en-Gal4 UAS-dcr2 UAS-GFP transgenes, and (A) no additional transgenes (control), (B) UAS-RNAi-Zyx³²⁰¹⁸. (C,D) en-Gal4 UAS-RNAi-Zyx³²⁰¹⁸ UAS-dcr2 UAS-GFP, stained for Yki (red/white) and DNA (Hoechst, green/white) with posterior cells marked by GFP (blue) or demarcated by the dashed line. (C) Upper panels show a horizontal section; lower panels show a vertical section. (D) Higher magnification of a portion of the image shown in (C).

Figure 3. Epistatic relationship of Zyx and Jub to wts and fat.

Wing imaginal discs, stained for Diap1 (red) and ex-lacZ (green), with posterior cells marked by GFP (blue), and with en-Gal4 UAS-dcr2 UAS-GFP transgenes, and (A) UAS-RNAi-wts, (B) UAS-RNAi-wts UAS-RNAi-Zyx³²⁰¹⁸, (C) UAS-RNAi-wts UAS-RNAi-Jub³⁸⁴⁴², (D) UAS-RNAi-fat, (E) UAS-RNAi-fat UAS-RNAi-Zyx³²⁰¹⁸, and (F) UAS-RNAi-fat UAS-RNAi-Jub³⁸⁴⁴².

Figure 4. Epistatic relationship of Zyx and Jub to ex, and influence on Yki localization.

Wing imaginal discs, stained for Diap1 (red) and ex-lacZ (green), with posterior cells marked by GFP (blue), and with en-Gal4 UAS-dcr2 UAS-GFP transgenes, and (A,B) UAS-RNAi-ex, (C,D) UAS-RNAi-ex UAS-RNAi-Zyx³²⁰¹⁸, (E,F) UAS-RNAi-ex UAS-RNAi-Jub³⁸⁴⁴². (G–L) show close-ups of portions of discs stained for Yki (red/white) and DNA (Hoechst, green/white) with posterior cells marked by GFP (blue), expressing en-Gal4 UAS-dcr2 UAS-GFP transgenes, and (G) UAS-RNAi-fat, (H) UAS-RNAi-ex, (I) UAS-RNAi-fat UAS-RNAi-Zyx³²⁰¹⁸, (J) UAS-RNAi-ex UAS-RNAi-Zyx³²⁰¹⁸, (K) UAS-RNAi-fat UAS-RNAi-Jub³⁸⁴⁴², and (L) UAS-RNAi-ex UAS-RNAi-Jub³⁸⁴⁴². Panels marked (i) show Yki and DNA, (ii) show Yki, (iii) show DNA, and (iv) show vertical sections, with triple stain at top, Yki in the middle, and DNA at bottom.

Figure 5. Wts Western blots.

(A) Western blot on lysates of third instar wing discs from tub-Gal4 UAS-dcr2 (control), tub-Gal4 UAS-dcr2 UAS-RNAi-Zyx³²⁰¹⁸, tub-Gal4 UAS-dcr2 UAS-RNAi-fat, tub-Gal4 UAS-dcr2 UAS-RNAi-fat UAS-RNAi-Zyx³²⁰¹⁸, and tub-Gal4 UAS-dcr2 UAS-Zyx:V5, probed with anti-Wts and anti-Actin antisera, as indicated. Similar amounts of total protein were loaded in each lane. (B) Quantitation of relative Wts protein levels in wing imaginal disc lysates. Wts and Actin band intensities were measured. To enable comparison across multiple blots, the Wts:Actin ratios were normalized to that detected in the control samples, which was set at 1. The histogram shows the average normalized ratios from five independent blots, error bars indicate s.e.m.

Figure 6. Zyx localization in wing imaginal discs.

All panels show Zyx localization in wing discs, based on UAS-Zyx:V5 (anti-V5, red) or UAS-Ypet:Zyx (red) transgenes. (A,B) Zyx localization versus E-cad (blue) and Ex (green) in an apical horizontal section (A) and vertical sections (B). (C,D) Zyx localization versus Fat (green) in an apical horizontal section (C) and vertical sections (D). (E,F) Close-up of Zyx localization versus Ex (green) in an apical horizontal section (E) and vertical sections (F). (G,H) Close-up of Zyx localization versus Dachs (using Dachs:Citrine, green) in an apical horizontal section (G) and vertical sections (H). (I) Close-up of Zyx localization in a clone. Zyx staining does not exhibit a proximal-distal bias. The stronger staining in the center of the clone presumably reflects the fact that this staining comes from two adjacent cells. (J) Close-up of Dachs localization in a clone. Dachs staining is strong on the distal side (yellow arrows) and weak on the proximal side (white arrows). Proximal-distal orientation is evidenced in these panels by Wg expression (blue) along the dorsal-ventral compartment boundary.

Figure 7. Binding amongst Zyx, Dachs, and Wts.

(A) Schematic of Wts, Dachs, and Zyx proteins, and the constructs used to map interaction domains. LD indicates Lim domain. Binding interactions are summarized to the right; + indicates strong binding, and − indicates weak or no binding. (B–G) show Western blots on co-immunoprecipitation experiments, with upper two blots indicating the relative amount of protein in the lysates used for the experiments and the lower panel indicating the material co-precipitated by the indicated antibody. GFP serves as a negative control. In (B–D) arrow identifies the Zyx-LD:FLAG polypeptide, and other bands in this lane are non-specific background detected by the antibodies. (B) Co-precipitation of V5-tagged Dachs with the FLAG-tagged proteins indicated at top. (C) Co-precipitation of FLAG-tagged Wts with the V5-tagged proteins indicated at top. (D) Co-precipitation of V5-tagged Dachs myosin domain with the FLAG-tagged proteins indicated at top. (E) Co-precipitation of V5-tagged Zyx-LD polypeptide with the FLAG-tagged proteins indicated at top. (F) Co-precipitation of V5-tagged Dachs with the FLAG-tagged proteins indicated at top. (G) Co-precipitation of FLAG-tagged Zyx with the V5-tagged proteins indicated at top. (H) Co-precipitation of V5-tagged Dachs and Zyx with the FLAG-tagged proteins indicated at top, in the presence of increasing amounts of Dachs:V5, as indicated. 1x indicates that equal amounts of pUAS-Zyx:V5 and pUAS-dachs:V5 plasmids were used, and 3x and 6x indicate corresponding increases in amounts of pUAS-dachs:V5 plasmid transfected. Note that in the absence of Dachs, no binding between full-length Zyx and Wts was detected when proteins were precipitated using anti-V5 beads and GFP:V5 was used as a negative control (panel C), but weak binding was detected when proteins were precipitated using anti-FLAG beads and GFP:FLAG was used as a negative control (H).

Figure 8. Models for Zyx function in Fat-Hippo signaling.

(A) Dachs might bridge Zyx and Wts within a trimeric complex; the simplest version of this model (stoichiometric amounts) would predict that in order for Zyx to be co-precipitated with Wts, Dachs and Zyx levels within the complex would have to be equivalent, which was not observed. (B) Dachs might induce a conformational change in Zyx (either directly through binding as shown or by recruiting other factors), exposing the LIM domains and enabling them to bind Wts. (C) illustrates the distinct roles of the LIM-domain proteins Zyx and Jub in Hippo signaling. Zyx influences the levels of Wts protein, presumably by promoting Wts degradation, whereas Jub inhibits Wts activation. The ability of Zyx to interact with Wts is regulated by Dachs, and Dachs in turn is regulated by Fat.