TPX2 phosphorylation by Aurora A is required for normal spindle length. (A) Asynchronous or mitotic HeLa cell lysates were treated with λ-phosphatase. In the middle panel, Mn²⁺-phos-tag was introduced into the gel to enhance mobility shifts of phosphorylated TPX2. The majority of TPX2 is phosphorylated during mitosis (arrow). (B) Mitotic cell lysates were treated with λ-phosphatase, DMSO, or Aurora A inhibitor MLN8237 (MLN), and asynchronous cell lysate was left untreated (left lane). Nonphosphorylated TPX2 accumulates after Aurora A inhibition (arrow) compared with DMSO treatment. (C) Protein sequences of Human and Xenopus TPX2 were aligned by ClustalW and potential phosphorylation sites by Aurora A are indicated with arrows (conserved in both sequences) or an arrowhead (unique to Xenopus TPX2). Asterisks indicate positions that have identical residues on both sequences, double dots indicate highly conserved residues, and single dots indicate weakly conserved groups. (D and E) In vitro kinase assay using recombinant TPX2 as substrate. Phosphorylation of TPX2 is abolished after double serine-to-alanine mutation, and both S121 and S125 contribute to the phosphorylation. TPX2 is indicated by arrows and Aurora A by asterisks. (F–H) HeLa cells were transfected with GFP-TPX2 or -2A. Note that GFP-2A causes short bipolar spindles. Bar, 10 µm. 400 cells from three independent experiments were measured in G, and error bars indicate SD. Three independent experiments were performed in H: n = 30, 30, and 76. Error bars indicate SEM; ***, P < 0.001. (I and J) Isogenic stable HeLa cell lines were depleted of endogenous TPX2 and induced to express RNAi-resistant GFP-TPX2-WT, -2A, and -2D for 48 h. The distance between bipolar spindle poles was measured. Whereas WT- and 2D-expressing cells assemble spindles with normal length, 2A expression causes shorter bipolar spindles. Three independent experiments were performed in J: n = 126, 145, 39, 39, and 35. Error bars indicate SEM. ***, P < 0.001.