A total of 8 phosphorylated residues were found. In agreement with a recent study, one of the identified phospho-residues, T676, is located in the activation -loop of the kinase domain. Another potential T-loop phosphorylation on T686 was identified, but confidence of assignment to this site was quite low. Similar identification of phosphorylation sites on in vitro auto-phosphorylated recombinant GST-Mps1 purified from insect cells revealed that three additional residues were in vitro phosphorylated, and that five of the residues found phosphorylated in vivo, including T686, were the result of autophosphorylation. In agreement with this, T676 and T686 were recently found phosphorylated on bacterially expressed recombinant Mps1. Many protein kinases are activated by phosphorylation on Tloop residues. Since both T676 and T686 were reported recently to contribute to kinase activity, we created Threonine-to-Alanine mutants of these two sites to study their contribution to mitotic PB 203580 checkpoint signaling, attachment-error-correction and chromosome segregation. To this end, endogenous Mps1 was depleted from U2OS cells using shRNA and replaced with either LAP-tagged wild-type Mps1 or the T676A or T686A mutants. Consistent with this, treatment of cells with low concentrations of a specific Aurora B kinase inhibitor did not cause observable inhibition of Aurora B at the level of substrate phosphorylation, chromosome alignment or cytokinesis yet resulted in persistent merotely in a significant fraction of cells. Therefore, it is likely that a small, virtually undetectable, local reduction in Aurora B activity in cells expressing Mps1- T676A underlies the occasional lack of correction of a merotelic attachment in these cells. Cells expressing Mps1-T676A arrested potently when treated with nocodazole but initiated anaphase in the presence of misaligned chromosomes. Thus, although T676 phosphorylation is not essential when the combined signals from all kinetochores is strong, it is needed for amplified checkpoint signaling from one or a few misaligned chromosomes. This supports the hypothesis that weakening of the mitotic checkpoint can cause non-lethal chromosomal instability. The contribution of such checkpoint weakening to CIN in tumor cells was recently questioned by a study in which filming of chromosome segregation in CIN tumor cells did not PI-103 reveal checkpoint defects but suggested merotely as the most common way to aneuploidy in these cells. Clearly in the three cell lines examined in that study, checkpoint signaling strength was sufficient to delay anaphase onset until all chromosomes had bioriented. In various other tumor cells, however, checkpoint strength was found to be diminished, and recent evidence from mouse models indicate that in cases where such weakening occurs, it can contribute to carcinogenesis.