RacGAP50C directs perinuclear gamma-tubulin localization to organize the uniform microtubule array required for Drosophila myotube extension

Abstract

The microtubule (MT) cytoskeleton is reorganized during myogenesis as individual myoblasts fuse into multinucleated myotubes. Although this reorganization has long been observed in cell culture, these findings have not been validated during development, and proteins that regulate this process are largely unknown. We have identified a novel postmitotic function for the cytokinesis proteins RacGAP50C (Tumbleweed) and Pavarotti as essential regulators of MT organization during Drosophila myogenesis. We show that the localization of the MT nucleator gamma-tubulin changes from diffuse cytoplasmic staining in mononucleated myoblasts to discrete cytoplasmic puncta at the nuclear periphery in multinucleated myoblasts, and that this change in localization depends on RacGAP50C. RacGAP50C and gamma-tubulin colocalize at perinuclear sites in myotubes, and in RacGAP50C mutants gamma-tubulin remains dispersed throughout the cytoplasm. Furthermore, we show that the mislocalization of RacGAP50C in pavarotti mutants is sufficient to redistribute gamma-tubulin to the muscle fiber ends. Finally, myotubes in RacGAP50C mutants have MTs with non-uniform polarity, resulting in multiple guidance errors. Taken together, these findings provide strong evidence that the reorganization of the MT network that has been observed in vitro plays an important role in myotube extension and muscle patterning in vivo, and also identify two molecules crucial for this process.

Fig. 1.

Drosophila RacGAP mutants have defects in somatic muscle patterning. Stage 16 embryos stained for Muscle myosin. Muscle patterning is shown at low (A,B) and high (D-I) magnification in wild-type (wt) (A,D,F,H) or RacGAP50C^DH15 (B,E,G,I) embryos. (A) Muscle pattern as observed in the whole embryo. (B) No significant loss of muscle tissue or unfused myoblasts are observed in RacGAP50C^DH15 mutants. (D,E) Arrows in D mark normal attachments of DO1. In RacGAP50C^DH15, DO1 is in the incorrect position (E, arrows). (F,G) Arrowheads in F mark muscle 22 (LT2). Arrowheads in G mark an LT muscle of an abnormal shape. The arrow marks a rounded muscle that fails to migrate. (H,I) Arrowheads in H mark normal attachments for VO muscles. In RacGAP50C^DH15 mutants, these muscles fail to fully extend (I, arrowheads). (C) Schematic of the wild-type muscle pattern of a single abdominal hemisegment, showing the names and numbers for muscles as referred to throughout this work. DA, dorsal acute; DO, dorsal oblique; LT, lateral transverse; VL, ventral lateral; VO, ventral oblique.

Fig. 2.

RacGAP^DH15 mutants can form stable muscle attachments. Wild-type (A,C) or RacGAP^DH15 mutant (B,D) stage 16 Drosophila embryos. To the left is a schematic of the wild-type muscle pattern (magenta) and tendon cells (green). (A,B) Muscle myosin and β-galactosidase staining of embryos with the sr-lacZ reporter labels muscles (magenta) and tendon cells (green). Tendon cells are specified in RacGAP^DH15 mutants (B), but muscles often select the wrong attachment sites. Arrowheads in A mark normal attachments for LT muscles, as compared with arrowheads in B that mark muscles making inappropriate attachments in mutants. Asterisks mark tendon cells to which muscles should be attached. (C) αPS2 integrin staining in the wild-type labels muscle ends at their contact site with tendon cells. (D) RacGAP^DH15 mutants make stable attachments as indicated by αPS2 integrin staining.

Fig. 3.

RacGAP functions cell-autonomously in muscles. Muscle myosin staining in stage 16 Drosophila embryos. Brackets mark LT muscles in two adjacent segments. (A) RacGAP^DH15 homozyogotes have muscle attachment site (MAS) selection and morphology defects. Shown is a LT muscle of abnormal shape, extending past its attachment site (top right bracket). (B) G14-GAL4/+;UAS-tum::myc/+ embryo showing that overexpression of RacGAP in muscles does not cause an abnormal phenotype. (C) RacGAP^DH15;69B-GAL4/UAS-tum::myc embryo showing that muscle defects are not rescued by RacGAP expression in tendon cells. (D) G14-GAL4, RacGAP^DH15/+,RacGAP^DH15;UAS-tum::myc/+ embryo showing that muscle defects are significantly rescued by expressing RacGAP specifically in muscles. LT muscles are now properly positioned.

Fig. 4.

FC specification and myoblast fusion are not affected in RacGAP mutants. (A,B) In wild-type Drosophila embryos at stage 12, Eve is expressed in the founder cell (FC) for muscle 1 (DA1) and two pericardial cells (A). These cells are properly specified in RacGAP^DH15 embryos (B). (C) Runt staining labels FCs for muscles 10 (DO2), 28 (VO3) and 15 (VO4) at stage 12. (D) In RacGAP^DH15 embryos, Runt-positive FCs are properly specified, although some segments contain an extra Runt-positive muscle 10 nucleus (arrow), whereas FCs for muscles 28 and 15 are unaffected (arrowhead). (E,F) Anti-Vestigial labels a subset of FCs at stage 12 in the wild type (E), and this staining is normal in RacGAP^DH15 mutants (F). (G,H) By stage 14, FCs have fused with fusion-competent myoblasts (FCMs), showing an increase in Vestigial-positive nuclei of wild-type dorsal muscles (G). Fusion also occurs in RacGAP^DH15 mutants (H). (I,J) 5053A-GAL4/UAS-lacZ embryo stained for Muscle myosin to label all muscles (magenta) and for β-galactosidase to label nuclei (green) of muscle 12 (VL1). In the wild type (I), muscle 12 spans each hemisegment and contains ∼11 nuclei. In RacGAP^DH15 mutants (J), muscle 12 is specified and contains the correct number of nuclei, but is mispositioned in 6% of hemisegments (n=52) (brackets). Asterisks mark intestinal cells that are also labeled with the 5053-GAL4 driver.

Fig. 5.

Localization of RacGAP and Pavarotti in muscle fibers. Stage 16 Drosophila embryos. The schematic on the left shows interior muscles 6, 7, 12 and 13 (VL1-4), which are exposed by fillet preparation performed for embryos stained for RacGAP, Roundabout and γ-tubulin. (A,B,G) RacGAP (A,G) and Muscle myosin (B) staining shows RacGAP throughout all muscle fibers and in discrete cytoplasmic puncta, in contrast to RacGAP nuclear localization in the epidermis (G, arrowheads mark epidermal cells). (C,C′) Merge of RacGAP (magenta) and Muscle myosin (green) in wild type showing discrete puncta of RacGAP (C′, arrowheads) at the periphery of the nuclei (C′, n). (D-F′) MEF2-GAL4/UAS-GFP::pav embryo stained for RacGAP (D) and GFP (E) and merged (F,F′). Pav is nuclear in all muscle fibers (E-F′ green), and co-staining of RacGAP (magenta) shows an increase in RacGAP concentration (F′, arrowhead) at the nuclear periphery. (H) pav^B200 mutants have a similar muscle phenotype to RacGAP^DH15 (arrowheads mark muscles making inappropriate attachments). (I) RacGAP is mislocalized to the ends of muscle fibers in pav^B200 mutants (arrowheads). (J) In RacGAP^DH15 mutants, Pav (green) still localizes to the nuclei of the muscle fibers (magenta). (K,L) w¹¹¹⁸ (K) and scraps⁸ (L) mutants have normal muscle patterning. (M,M′) α-Spectrin (magenta) and DAPI (green) staining shows that scraps⁸ mutants have cytokinesis defects in the epidermis. The boxed region in M is shown at high magnification in M′, where arrowheads indicate binucleate cells.

Fig. 6.

RacGAP mutants do not have general defects in the actin or microtubule cytoskeleton. (A-H) Phalloidin labeling showing LT (A-B′,E-F′) and ventral (C,D,G,H) muscles. During MAS selection, F-actin concentrates at the ends of myotubes in the wild type (A-C) and in RacGAP^DH15 mutants (E-G). Magnification of LT ends shows multiple filopodia in both wild type (B,B′) and RacGAP^DH15 (F,F′). At stage 16, F-actin continues to accumulate at muscle ends (D). In RacGAP^DH15 mutants, F-actin accumulation occurs even when muscles attach to the wrong site (H, arrowhead) or contain ectopic extensions (H, arrow). (I-L) Anti-β3-tubulin labels microtubules (MTs) in somatic muscles. At stage 13 during myoblast fusion, wild type (I) and RacGAP^DH15 (K) have a similar β3-tubulin pattern. By stage 16, wild-type MTs are organized along the longitudinal axis of each muscle (J). In RacGAP^DH15 mutants, MTs are present, but muscles make inappropriate attachments (L, arrowheads).

Fig. 7.

RacGAP is required for uniform MT polarity. (A-I) G14-GAL4 driving UAS-nod::GFP labels MT minus ends. Brackets mark individual muscles, arrowheads mark muscle ends. As wild-type muscles extend (A,C) and make attachments (D,G), Nod::GFP accumulates towards the center of the fiber (arrows) and is notably absent from muscle ends. In RacGAP^DH15 (B,C′,E,H) and pav^B200 (I) mutants, Nod::GFP is present throughout the entire muscle, which shows a disruption of MT polarity. C and C′ are higher magnifications of stage 14/15 Nod::GFP expression in wild type and RacGAP^DH15 mutants. (F) Co-expression of UAS-tum::myc and UAS-nod::GFP in RacGAP^DH15 mutants results in significant rescue of muscle patterning and restoration of MT polarity, as shown by the absence of Nod::GFP at muscle ends. G and H are higher magnifications of D and E, respectively, showing Nod::GFP at the interior of muscle fibers in the wild type (G) and at the ends of muscles in RacGAP^DH15 (H).

Fig. 8.

RacGAP controls MT polarity through interaction with γ-tubulin. Yellow arrows mark segment borders, where myotubes from adjacent segments make attachments. (A,B) In wild type, γ-tubulin (A) and RacGAP (B) have a similar localization pattern in cytoplasmic puncta that are concentrated towards the center of the myotube. (C,C′) Merge of RacGAP (red) and γ-tubulin (green) shows that γ-tubulin colocalizes with RacGAP puncta at the nuclear periphery. (D-G) Gain was increased for γ-tubulin imaging in D,D′,E,E′,F,F′,G. The segment borders indicated by the yellow arrows in D, E and F are magnified in D′, E′ and F′, respectively. D″, E″ and F″ are schematics of γ-tubulin localization within extending myotubes at a segment border in the wild-type, pav and RacGAP backgrounds, respectively. In wild-type mononucleated myoblasts (G), γ-tubulin (green) is present diffusely throughout the cytoplasm and is absent from the nucleus (magenta, labeled with anti-Mef2). In wild-type myotubes (D,D′), γ-tubulin is found in the cytoplasm (D), with puncta concentrated at the nuclear periphery (D′, arrowheads), but is absent from the muscle ends (yellow arrow). In pav^B200 mutants (E,E′), γ-tubulin is not as highly concentrated around nuclei (E′, n), but is localized to patches at muscle ends (E′, arrowhead). In RacGAP^DH15 mutants (F,F′), γ-tubulin fails to localize to perinuclear sites and instead is diffusely localized throughout the muscle cytoplasm, similar to its pattern in mononucleated myoblasts (G).

Fig. 9.

Model of RacGAP and Pav function during myotube extension in Drosophila. (A) Mononucleate myoblasts have MTs dispersed throughout the cytoplasm and diffuse cytoplasmic γ-tubulin localization. (B) One end of an elongating myotube. In multinucleated myotubes, the MT array must be reorganized in the longitudinal axis of the muscle to allow for elongation and extension. This organization requires MT polarity, characterized by minus ends at the interior of the myotube and plus ends at the periphery, to drive extension. RacGAP, Pavarotti (KLP, kinesin-like protein) and γ-tubulin are required to establish the proper MT array in migrating myotubes. Pav localizes RacGAP to discrete cytoplasmic puncta at the nuclear periphery, and RacGAP localization determines γ-tubulin distribution. RacGAP may transport incoming MTs to the nuclear periphery after myoblast fusion and/or promote the nucleation of new MTs in the appropriate orientation by increasing γ-tubulin at the nuclear periphery, thus establishing the MT array required for myotube extension.