Overexpression of β spectrin in a wild-type background produced dominant negative effects in vivo. (A) Overexpression of UAS-α and UAS-β spectrin transgenes in the wing disk. Top, overexpression of the UAS-α-Spec³⁷ transgene in the wing disk driven by MS1096-Gal4 did not affect wing development. Bottom, overexpression of the UAS-β-Spec⁹⁵ transgene blocked normal wing development. (B) Overexpression of the UAS-β-Spec⁹⁵ transgene driven by the neuronal elav-Gal4 caused defects in wing inflation/unfolding. (C) Overexpression of UAS-β-Spec⁹⁵ in the salivary gland. Left, in control salivary glands, β spectrin was localized to the basolateral membranes of epithelial cells. β spectrin was detected with a rabbit anti-β spectrin antibody and a Cy3-labeled secondary antibody. Right, UAS-β-Spec⁹⁵ was expressed in the salivary gland (sg) via basal activity of heat-shock Gal4, which resulted in perturbed epithelial cell polarity. β spectrin was aberrantly associated with the apical membrane domain, and some cells no longer showed cuboidal morphology. β-Spec⁹⁵ was detected with mouse anti-myc antibody and Cy3-labeled secondary antibody. Scale bar, 10 μm.

Neuronal expression of UAS-β-Spec⁹⁵ was sufficient to rescue the lethality of β spectrin mutants. (A) Rescue cross. The neuronal driver elav-Gal4 was recombined onto the X chromosome of a female fly heterozygous for the spectrin loss-of-function mutation β-Spec^em6. Heterozygous elav-Gal4-β-Spec^em6/FM7; +/+ female flies were then crossed to male flies carrying the myc-tagged wild-type transgene (+/Y; UAS-β-Spec^95/ UAS-β-Spec⁹⁵). Successful rescue was scored by the presence of a non-FM7 F₁ male class (which lacks endogenous β spectrin). (B) Adult rescue flies. Flies rescued with the UAS-β-Spec⁹⁵ transgene at 25°C were generally smaller than wild-type siblings (FM7/Y) and had wing unfolding defects similar to wild-type flies that overexpressed UAS-β-Spec⁹⁵ (Table 1). In contrast, flies rescued with UAS-β-Spec⁴⁹³ at any temperature or with UAS-β-Spec⁹⁵ at 22°C did not exhibit wing unfolding or body size defects. (C) Quantitative analysis of rescue. Neuronal expression of either the UAS-β-Spec⁹⁵ or the UAS-β-Spec⁴⁹³ transgenes (both at 25°C) efficiently rescued the lethal β-Spec^em6 mutation at about the same rate as a ubiquitously expressed β spectrin transgene described previously (Dubreuil et al., 2000). Expression of UAS-α-Spec²⁷ (negative control) did not produce the rescue class of male progeny.

Production of modified β spectrin and DAnk2 transgenes. (A) Top, β spectrin is divided into discrete structural domains, including an N-terminal actin-binding domain (ABD), a C-terminal pleckstrin homology (PH) domain, 16 degenerate ∼106-amino acid–long spectrin repeats (ellipses), and one partial repeat near the C-terminus (Bennett and Baines, 2001). Two of the degenerate repeats (14 and 15) have been implicated in ankyrin (ank)-binding activity. Modified transgenes include α13 (ankyrin-binding repeat 15 of β spectrin replaced with repeat 12 of α spectrin; Das et al., 2006) and α8 (ankyrin-binding repeat 14 of β spectrin replaced with repeat 8 of α spectrin). Bottom, conventional ankyrins are divided into three major structural domains: an N-terminal membrane binding domain (MBD) with 24 copies of a 33-amino acid ankyrin repeat, a central spectrin-binding domain (SBD), and a C-terminal domain that diverges widely among ankyrin isoforms (Bennett and Baines, 2001). UAS-DAnk transgenes were engineered with a C-terminal EGFP tag. The DAnk2S isoform produced here contains the N-terminal ankyrin repeat domain and a truncated spectrin-binding domain. The C-terminal domain found in other Dank2 isoforms (and DAnk1) is absent in Dank2S. (B) Western blots demonstrated that the transgenes are stably expressed and are the expected size. Total protein from flies constitutively expressing β-Spec^a8 (lane 1, arrowhead), β-Spec^α13 (lane 2), or no transgene (lane 3) was probed with mouse anti-myc antibody and rat anti-α actinin as a loading control. Total protein from induced (lane 4) or uninduced (lane 5) bacteria carrying a DAnk2S-EGFP insert was probed with anti-EGFP antibody. IPTG induction strongly elevated expression of the expected DAnk2S-EGFP product (arrow). Molecular-weight standards are indicated to the left in kDa (additional 66-kDa standard indicated for different percentage gel on right). (C) Mouse anti-myc antibody staining of the β-Spec^α8 transgene product in the salivary gland of a mutant larva lacking endogenous β spectrin. A goat anti-mouse Texas Red–conjugated secondary antibody was used. The β-Spec^α8 exhibited the same polarized basolateral distribution as wild-type β spectrin (arrowhead). Scale bar, 10 μm.

β spectrin knockdown with dsRNA. β spectrin was detected with a rabbit anti-β spectrin antibody and a Cy3-labeled secondary antibody. (A) Expression of UAS-β spectrin dsRNA in the salivary gland (using heat-shock Gal4 as in Figure 1) reduced β spectrin abundance. Top, β spectrin staining of wild-type salivary gland and fat body cells. Bottom, β spectrin dsRNA expression in the salivary gland reduced spectrin staining to background levels (*). Fat body cells (fb) not expressing the dsRNA construct remained strongly stained for β spectrin. Scale bar, 10 μm. (B) Expression of UAS-β spectrin dsRNA in the fat body with Cg-Gal4. Top, wild-type fat body cells, demonstrating that β spectrin (red) colocalized with the plasma membrane marker UAS-mCD8-GFP (green). Bottom, β spectrin immunofluorescence in the fat body was reduced to background levels by Cg -Gal4–driven expression of UAS-β spectrin dsRNA in the fat body (*). A salivary gland (sg) beside the fat body still expressed β spectrin at wild-type levels. Scale bar, 50 μm.

Lifespan of UAS-β-Spec⁴⁹³-rescued β-Spec^em6 mutants. Age at death was plotted for β-Spec^em6 mutants rescued by the β-Spec⁴⁹³ transgene (blue) compared with FM7 controls (red) all reared at 25°C. Similar results were obtained with the β-Spec⁹⁵ transgene (not shown). Adult progeny from a representative rescue cross were scored daily to determine posteclosion age at time of death. Neuron-specific expression of β-Spec⁴⁹³ was sufficient to allow development of β spectrin mutants to adulthood, but the rescued flies were unusually short-lived.

Targeting of DAnk1-EGFP and Dank2S-EGFP to the plasma membrane in β spectrin mutants. Transgene expression was driven in the salivary gland by heat-shock Gal4 and was detected by EGFP fluorescence and confocal microscopy. (A–D) DAnk1-EGFP. (E–H) DAnk2-EGFP. In wild-type controls (A and E) both DAnk-EGFP isoforms were prominently enriched at the basolateral surface of the salivary gland epithelial cells. In the absence of endogenous β spectrin (β-Spec^em6 mutant) the membrane-labeling pattern was reduced to near background level (B and F). Note the variable level of DAnk-EGFP expression from cell to cell, resulting in strong cytoplasmic signal in some cells. Both spectrin repeat modifications (β-Spec^α13, C; β-Spec^α8, D) had a similar effect on the behavior of DAnk1-EGFP, reducing the membrane-specific signal to near background levels. The β-Spec^α8 repeat modification had a comparable effect on the behavior of DAnk2S-EGFP (H). In contrast, DAnk2S-EGFP recruitment to the plasma membrane was left largely intact in the β-Spec^α13 mutant (G). A nuclear signal (presumably an artifact of the DAnk2S-EGFP fusion) that was barely detectable in wild-type (E) became more conspicuous in mutant cells with high-level expression (F–H, arrowheads) but was never observed with DAnk1-EGFP (A–D).