Genomic organization of rngo and structure of the Rngo protein. (a) The transcription start of rngo is marked by a black flag, and the stop is marked with a red flag. The position of the P element insertion representing a null allele of rngo is indicated. (b) Predicted structure of the Rngo protein. (c) Alignment of the RVP domain of Rngo with the corresponding region of its yeast (Sc, Saccharomyces cerevisiae), mouse (Mm, Mus musculus), and human (Hs, Homo sapiens) homologues. Residues identical to the consensus sequence are boxed in black. Dm, Drosophila.

Subcellular localization of Rngo during oogenesis. (a) Rngo (red) is expressed in somatic cells and germline cells of egg chambers. It localizes to the cytoplasm and to the nuclei. The strongest staining can be seen in the nucleus of the oocyte (arrows), marked by the expression of Orb (green). (b–d) Costaining of Rngo with lamin (b, green), ubiquitin (c, green), and with the proteasome (d, green). (e) Germline cells homozygously mutant for rngo^GE3956 (asterisk) completely lack Rngo staining. Clones are marked by loss of GFP expression (green). (f) L3 larval protein extracts were analyzed by Western blotting. In larvae possessing the wild-type rngo gene (right lanes), a band of 60 kD corresponding to FL Rngo is detectable, which is absent in the extract of rngo^GE3956 mutant animals (left lane). α-Tubulin was used as a loading control. In a–e, DNA was stained with DAPI (turquoise, only shown in the merged images). Egg chambers are at stages 6 and 7. Anterior is to the left. fc, follicle cell; nc, nurse cell; oc, oocyte. Bars: (a–d) 10 µm; (e) 20 µm.

Rngo binds to ubiquitin and to the proteasome and forms homodimers. (a) A schematic diagram of Rngo fusion proteins used in this study. Numbers corresponding to amino acid residues indicate domain boundaries and the extent of the introduced deletions. (b) The UBA domain of Rngo is required for binding to ubiquitin. Ubiquitin-conjugated agarose beads were used to pull down the MBP-Rngo fusion proteins indicated above each lane. MBP was used as a negative control. (c) Direct binding of Rngo to the proteasomal subunit Rpn10. FL MBP-Rpn10 was pulled down by GST–FL Rngo but not by GST alone. (d) Rngo binds to ubiquitin (Ub) and Rpn10 in vivo. Extracts of wild-type ovaries (left) and ovaries expressing Rngo-GFP (right) were used for co-IP experiments with the indicated antibodies. The anti-MBP antibody was used as a negative control. (e) Rngo forms homodimers via its RVP domain. S2 cells were cotransfected with FL Rngo-GFP and either FL or mutated versions of Rngo tagged with the HA epitope. Cell extracts were used for IP with an anti-GFP antibody. IP with an anti-GST antibody was used as a negative control. Untransfected cells and cells transfected only with the respective HA-tagged version of Rngo were used as additional negative controls. White lines indicate that intervening lanes have been spliced out.

Rngo is required for ring canal growth and plasma membrane integrity in the female germline. (a–c) A control germline clone (glc, a) and rngo^GE3956 germline clones (b and c) stained for F-actin (red) and Hts-RC (blue). Note the abnormally large ring canal (b, arrows) and the fusion of at least two nurse cells with the oocyte (c). Insets to the right of a and c show higher magnifications of the oocyte nucleus (red arrows) and an adjacent nurse cell nucleus in the boxed regions. Note the elongated shape of the oocyte nucleus in c. (d and e) A control germline clone (d) and rngo^GE3956 germline clone (e) stained for Filamin (red) and Kel (blue). Note the rudimentary ring canals (arrowheads) and the diffuse cytosolic localization of Kel in e. (f and g) A control germline clone (f) and rngo^GE3956 germline clone (g) stained for PY epitopes (blue). Rudimentary ring canals in g are marked by arrowheads. DNA was stained with DAPI (turquoise). Germline clones are marked by an absence of GFP (green) expression. Note that for imaging of Hts-RC, Kel, Filamin, and PY in rngo^GE3956 germline clones, a much higher gain of the photomultiplier was used than in the respective wild-type controls. Egg chambers are at stages 7 and 8. Anterior is to the left. (h and i) Quantification of outer ring canal diameter of ring canals connecting nurse cells (h) and a nurse cell with the oocyte (i). (j) Quantification of the fusion between nurse cells and the oocyte. In h, n = 40 for stages 3–7, and n = 19 for stage 8. In i, n = 16 for stages 3–7, and n = 6 for stage 8. The statistical significance of the differences was calculated with the Student’s t test; ***, P < 0.01. In j, n = 15 for all stages. Error bars represent SDs from three independent experiments. Bars: 20 µm; (insets) 10 µm.

The levels of ring canal–associated proteins are reduced in rngo^GE3956 germline clones. (a) Western blot of extracts from wild-type ovaries (left) and ovaries with rngo^GE3956 germline clones (glc, right). α-Tubulin was used as a loading control. (b and c) A control germline clone lacking GFP shows the same level of Hts-RC (blue) as the adjacent wild-type egg chamber (b), whereas Hts-RC levels are strongly reduced in an rngo^GE3956 germline clone egg chamber (c). (d–g) The same is true for PY epitopes (d and e, blue), Filamin (f and g, red), and Kel (f and g, blue). Germline clones are marked by the absence of GFP (green). Images are projections of z stacks. Egg chambers are at stages 5–7. Anterior is to the left. Bars, 20 µm.