mpD alleles are conferred by mutations in IMP9. (A–D) mpD-1 increases the M row number. SR-SIM images of cells stained with the anti-centrin (red) and DAPI (blue), WT or mpD-1 homozygotes were grown at either standard temperature 30°C or 39°C for 6 h. White arrows mark the ends of M rows. (E) Mapping of mpD-1 in the micronuclear genome using ACCA. Note a small linkage peak on chr3 at ∼20 Mb (red star). (F–F″) A comparison of human IMP9 (IMP9Hs) structure with the AlphaFold model of Tetrahymena IMP9/MpD. (F) The experimental (crystal) structure of H2A-H2B–bound IMP9Hs (PDB: 6N1Z) shown in cartoon representation, with the bound histone H2A-H2B cargo removed to focus on the structure of importin-9 (light purple). IMP9Hs is a superhelical/solenoid protein made up of 20 HEAT repeats, each comprising a pair of alpha helices. Most of the IMP9Hs alpha helices are connected by short loops, but four loops are long and colored differently: h7loop (brown), h8loop (aquamarine), h18–19loop (dark purple), and h19loop (blue). An asterisk labels h19loop because it was not resolved in the crystal structure. Instead, we superposed the h19loop from the AlphaFold model of IMP9Hs (AF-Q96P70) onto the crystal structure (PDB: 61NZ). (F′ and F″) Cartoon representations of the predicted model of Tetrahymena IMP9/MpD by AlphaFold (AF-Q23AR9). The model in F′ is colored by the confidence score for each residue of the predicted AlphaFold model. F″ shows the same AlphaFold model colored according to the HEAT repeat helices (green) and the long loops (colored like the homologous IMP9Hs loops). The predicted h9loop of Tetrahymena IMP9/MpD is longer than in human IMP9 and colored light orange. In general, the AlphaFold model of Tetrahymena IMP9/MpD is very similar to the structure of human IMP9 shown in F. Residues G939 and D1148, which are mutated in mpD-1 (G939R) and mpD-2 (D1148N), are marked with arrows. (G–J) Genetic tests show that the IMP9 gene is MPD. (G–I) Confocal images of Tetrahymena cells with the indicated genotypes (grown at 39°C) and labeled by the anti-centrin antibody (red) and DAPI (blue). Asterisks mark OAs with 4 M rows characteristic of mpD-1. Cells shown in I are mpD-1 mutant homozygotes in which the variant codon R939 was replaced by the reference G codon via homologous DNA recombination (Fig. S5 J, lane 6). Note a reduction in the presence of OAs with 4 M rows. (J) The graph quantifies the extra M row phenotype in strains with the indicated genotypes:WT; mpD-1, mutant homozygote (strain IA305); and mpD-1+IMP9-G939R, an mpD-1 homozygous strain carrying the mpD-1–linked variant (G939R codon substitution) with GFP sequence integrated at the end of the coding region of IMP9. Note that the frequency of OAs with 4 M configuration is higher than in the original mpD-1 homozygote strain, possibly because addition of GFP further decreased the functionality of G939R MpD; mpD-1+IMP9-G939; an mpD-1 homozygote strain in which the variant codon (R939) was replaced by the reference codon (G939) and GFP was integrated at the end of the coding region of IMP9; mock knockdown, a strain carrying an “empty” pMCodel vector, in which the level of IMP9 expression was expected to be unaffected; IMP9 knockdown, a strain transformed with a co-deletion vector expressing an RNA that targets a portion of IMP9 for deletion during macronuclear development. There are 90 copies of IMP9 in the macronucleus, and likely not all copies of IMP9 were disrupted, resulting in a “knockdown” of IMP9 expression. Mean ± SD. N = 3 experiments (100 cells scored per experiment and genotype). The P values were obtained using an unpaired t test with Welch’s correction. A more complete set of data related to panels G–J is included in Fig. S5.