In the present study we investigated the different roles of Mps1 in mitosis of human cells through the use of chemical genetics. We report here two clonal cell lines expressing analog-sensitive versions of Mps1 in which endogenous Mps1 was stably removed by doxycycline-induced shRNA expression. Use of these cell lines circumvents experimental limitation in reproductivity and penetrance associated with transient RNAi reconstitutions 
and allows temporal control of Mps1 activity.
Chemical genetic inhibition of Mps1 was previously used to uncover novel aspects of Mps1 in S. cerevisiae
. A yeast strain carrying the analog-sensitive M516G mutation (similar to the M602G mutation in human Mps1, included in the present study) allowed analysis of Mps1 functions after the moment of spindle pole body duplication, a critical function of Mps1 in budding yeast 
. Using this strain (mps1-as1
), it was verified that Mps1 inhibition prevents mitotic checkpoint activity and causes cell death 
. In addition, Mps1-inhibited yeast cells were found to display spindle morphology defects and chromosome mispositioning 
, possibly as a consequence of errors in chromosome biorientation 
In agreement with these and other reports 
, chemical genetic inhibition of Mps1 in our human cell lines verified that Mps1 kinase activity is critical for chromosome biorientation. Although our data using Mps1 RNAi supported the hypothesis that Mps1 facilitated error-correction by promoting Aurora B activity at inner centromeres 
, we now show that error-correction is also prevented when Mps1 kinase activity is inhibited only during the error-correction process in cells undergoing spindle bipolarisation after monastrol wash-out. Additionally, through the use of phosphate-affinity gel electrophoresis we show that phosphorylation of endogenous Borealin is under the control of Mps1. Borealin phosphorylation enhances Aurora B activity 
yet inhibition of Mps1 did not affect pS10-H3 levels under conditions that showed an increase in Borealin mobility. pS10-H3 was only found reduced when cells entered mitosis without Mps1 activity. These results suggest that partial reduction in Aurora B activity caused by Mps1 inhibition is difficult to detect once pS10-H3 is fully established, possibly because of low turnover of the phosphate group on S10-H3. In support of this, inactivation of Aurora B by a 2 hour exposure of our HCT-derived cells to ZM447439 only caused a moderate reduction of pS10-H3 levels. These data indicate that pS10-H3 serves as a good readout for Mps1-dependent regulation of Aurora B activity when pS10-H3 is not fully established at mitotic entry. Nevertheless, since Mps1 inhibition during mitosis affects Borealin phosphorylation and error-correction, it is likely that acute effects on Aurora B activity by Mps1 inhibition can be visualised with antibodies to the relevant error-correction substrates.
The function of Mps1 kinase activity in the regulation of the mitotic checkpoint has previously been established by various labs including ours, and the present data fully agree with this 
. As expected 
, Mps1 inhibition in our cell lines caused mislocalization of Mad1 and Mad2. This disagrees with the study by Tighe et al., in which chemical genetic inhibition of Mps1 in cells transiently transfected with Mps1 shRNA plasmids had little effect on Mad1 recruitment 
. This discrepancy may be due to differences in penetrance of Mps1 inhibition. Whereas inhibition of Mps1 with 1NM-PP1 in the Tighe et al. study left mitotic progression unaffected in a large proportion of cells, mitosis was disturbed in the vast majority (UTR-Mps1as
) or all (HCT-Mps1as
) cells in our cell lines ( and ). This difference in penetrance may be due to inefficient RNAi by transient shRNA transfection or potency of 1NM-PP1 in the Tighe et al., study. In our hands, 1NM-PP1 was significantly less capable of inhibiting Mps1 than 23dMB-PP1 or 3MB-PP1 (Fig. S2
). Besides Mad1/Mad2, our data further indicate that Mps1 is essential for recruitment of Cdc20 as well as Bub1 to unattached kinetochores, in agreement with some 
but not other 
previous reports using Mps1 RNAi or Mps1 protein depletion.
Mps1 inactivation affects the quality of chromosome segregation in unperturbed mitosis 
. Using time-lapse microscopy we show that inactivation of Mps1 accelerates mitosis. A similar phenotype was reported for cells depleted of Mad2 or BubR1 
. The regulation of mitotic timing by Mad2 and BubR1 is kinetochore-independent, since Mad1 depletion, while disabling the mitotic checkpoint, did not affect normal mitotic timing 
. It is therefore unlikely that lack of recruitment of Mad1 and Mad2 to kinetochores in Mps1-inhibited cells contributes to accelerated mitosis. An explanation for the timing phenotype, proposed by Meraldi et al., and in agreement with experimental data 
, is the possibility that sufficient MCC exists in interphase to constrain APC/C activity long enough to allow full chromosome biorientation in the majority of cells. In this model, only presence of persistent unattached kinetochores requires additional MCC formation to further prolong mitosis. Given the timing phenotype of Mps1-inhibited cells, it is therefore possible that Mps1, in addition to influencing Mad1-dependent MCC formation, affects stability or inhibitory potential of pre-assembled MCC. Further studies are needed to examine this hypothesis.
Due to its central role in mitosis, Mps1 is an attractive target for anti-cancer therapies. A recent study from our lab proposed that elevating the frequency of chromosome segregation errors by partial Mps1 reduction can be of clinical relevance, especially with other treatments that affect fidelity of chromosome segregation 
. We show here for the first time that specific inhibition of Mps1 enzymatic activity efficiently kills human cells. Moreover, a short pulse of inhibition followed by recovery caused lethality in a significant fraction of cells. As such a pulse of Mps1 inhibition is more relevant to potential future clinical situations than potent, long-term kinase inhibition, these data illustrate that specific and penetrant inhibition of Mps1 activity might indeed be a promising anti-cancer strategy.