The mutation in rodZ3 affects RZZ streaming in neuroblasts. (A) Quantification of RZZ streaming in WT, rodZ3, and rodZ3/+ neuroblasts. Cells were classed by the intensity of GFP–Zw10 streaming: high, low, and no streaming. RZZ streaming is not observed in homozygous rodZ3 cells and is substantially reduced in rodZ3/+ cells. (B) Dynein recruitment is reduced but not eliminated in rodZ3 neuroblasts. Images of GFP–DLIC2 at kinetochores in live WT, rodZ3 and rod-null cells treated with colchicine. Dynein is recruited to unattached kinetochores in WT and rodZ3 cells, but not in rod-null cells. Scale bars: 2 µm. (C) Quantification of GFP–DLIC2 at kinetochores, normalized to Spc25–RFP, in colchicine-treated WT (blue, n = 36), rodZ3 (red, n = 34) or rod-null (gray, n = 18) cells. In 90% of rodZ3 cells, GFP–DLIC2 is detectable at kinetochores but its signal is reduced compared to that in WT cells. Horizontal bars represent the mean. **P<0.01 (Student's t-test). (D) RodZ3 dominantly disrupts RZZ streaming. Top, metaphase spindles of neuroblasts expressing two different colors of WT Rod [Cherry (Ch)–Rod+ and GFP–Rod+, first row, or one WT copy and one mutant copy (Ch–Rod+ and GFP–RodZ3, second row] in a rod-null background. In cells expressing only WT Rod, streaming is normal, and particles containing both colors are found on the spindle fibers. When GFP–RodZ3 is co-expressed with Ch–Rod+, the streaming of both is reduced. Scale bars: 2 µm. Bottom, quantification of Rod streaming, using criteria described in A. The fraction of cells with robust streaming is substantially reduced when GFP–RodZ3 is present.