Chemical Genetic Inhibition of Mps1 in Stable Human
Cell Lines Reveals Novel Aspects of Mps1 Function in
Tale Sliedrecht1,2, Chao Zhang3, Kevan M. Shokat3, Geert J. P. L. Kops1,2*
1Department of Physiological Chemistry, Cancer Genomics Centre, University Medical Center (UMC) Utrecht, Utrecht, The Netherlands, 2Netherlands Proteomics Centre,
Utrecht, The Netherlands, 3Department of Cellular and Molecular Pharmacology, University of California San Francisco, San Francisco, California, United States of America
Background: Proper execution of chromosome segregation relies on tight control of attachment of chromosomes to
spindle microtubules. This is monitored by the mitotic checkpoint that allows chromosome segregation only when all
chromosomes are stably attached. Proper functioning of the attachment and checkpoint processes is thus important to
prevent chromosomal instability. Both processes rely on the mitotic kinase Mps1.
Principal Finding: We present here two cell lines in which endogenous Mps1 has been stably replaced with a mutant kinase
(Mps1-as) that is specifically inhibited by bulky PP1 analogs. Mps1 inhibition in these cell lines is highly penetrant and
reversible. Timed inhibition during bipolar spindle assembly shows that Mps1 is critical for attachment error-correction and
confirms its role in Aurora B regulation. We furthermore show that Mps1 has multiple controls over mitotic checkpoint
activity. Mps1 inhibition precludes Mad1 localization to unattached kinetochores but also accelerates mitosis. This
acceleration correlates with absence of detectable mitotic checkpoint complex after Mps1 inhibition. Finally, we show that
short-term inhibition of Mps1 catalytic activity is sufficient to kill cells.
Conclusions/Significance: Mps1 is involved in the regulation of multiple key processes that ensure correct chromosome
segregation and is a promising target for inhibition in anti-cancer strategies. We report here two cell lines that allow specific
and highly penetrant inhibition of Mps1 in a reproducible manner through the use of chemical genetics. Using these cell
lines we confirm previously suggested roles for Mps1 activity in mitosis, present evidence for novel functions and examine
cell viability after short and prolonged Mps1 inhibition. These cell lines present the best cellular model system to date for
investigations into Mps1 biology and the effects of penetrance and duration of Mps1 inhibition on cell viability.
Citation: Sliedrecht T, Zhang C, Shokat KM, Kops GJPL (2010) Chemical Genetic Inhibition of Mps1 in Stable Human Cell Lines Reveals Novel Aspects of Mps1
Function in Mitosis. PLoS ONE 5(4): e10251. doi:10.1371/journal.pone.0010251
Editor: Daniela Cimini, Virginia Tech, United States of America
Received February 12, 2010; Accepted March 26, 2010; Published April 22, 2010
Copyright: ? 2010 Sliedrecht et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by the Netherlands Genomics Initiative of the Netherlands Organization for Scientific Research and ERC-StG-242617 to GJPLK.
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: email@example.com
To maintain a stable genome, cells have evolved a variety of
processes that ensure accurate chromosome segregation. In early
mitosis, kinetochores of sister chromatids attach to microtubules
emanating from opposite spindle poles. Correct end-on attach-
ment of microtubules to kinetochores relies on the error-correction
machinery that destabilizes improper attachments through the
actions of the Aurora B kinase . As long as unattached
kinetochores persist, the onset of anaphase is prevented by a
surveillance mechanism called the mitotic checkpoint that will halt
cell cycle progression until all chromosomes are stably attached to
the mitotic spindle . The mitotic checkpoint will be satisfied
upon stable biorientation of all chromosomes, after which
chromosome segregation is allowed to proceed. Proper execution
of chromosome biorientation and mitotic checkpoint signaling
relies on a set of multifunctional kinases, one of which is the dual
specificity kinase Mps1 .
First discovered to regulate spindle pole body duplication in
budding yeast , Mps1 was subsequently found to additionally
regulate the mitotic checkpoint  and spindle assembly .
Regulation of the mitotic checkpoint by Mps1 is evolutionary
conserved and has been shown in fission yeast, fruit flies, Xenopus
egg extracts and human cells [7–11]. Mps1 exerts this control, at
least in part, through regulating kinetochore localization of several
checkpoint proteins including Mad1 and Mad2 [9,11,12].
Recently, Mps1 was also reported to regulate sister chromatid
biorientation in both budding yeast and humans [12,13]. In
human cells, Mps1 promotes biorientation by regulating Aurora B
activity through phosphorylation of the chromosomal passenger
complex (CPC) member Borealin [12,14].
Due to its central role in mitosis, misregulation of Mps1 kinase
activity results in chromosomal instability (CIN) and subsequent
aneuploidy, a hallmark shared by cells from solid tumors [15,16].
Inefficient activation of Mps1 results in weakened mitotic
checkpoint activity and the persistence of falsely attached
PLoS ONE | www.plosone.org1April 2010 | Volume 5 | Issue 4 | e10251
chromosomes, causing frequent but non-lethal chromosome
segregation errors . Conversely, reduction of Mps1 activity
has recently been shown to sensitize tumor cells but not normal
cells to low doses of taxol by elevating the frequency of
chromosome missegregations to near-lethal levels . Partial
inhibition of Mps1 might therefore be an effective anti-cancer
Although RNAi studies have uncovered several aspects of
human Mps1 biology, the multifunctional character of Mps1 has
prevented detailed and temporally controlled investigations into
the different roles Mps1 might play in mitosis. Inhibition using the
small molecules SP600125 and cincreasin has proved to be useful
[18,19], but cincreasin does not inhibit Mps1 in human cells 
and the non-specific nature of SP600125 makes it an unfavorable
choice to study Mps1. A more controlled approach is the use of
chemical genetics, in which endogenous kinase is replaced by an
engineered protein containing a mutated gatekeeper residue .
These gatekeeper mutants render the kinase specifically sensitive
to inhibition by non-hydrolysable bulky ATP analogs such as
chemically modified variants of the Src inhibitor PP1. This
approach has previously been described for Mps1 in budding yeast
[6,13] and in human cells in combination with transient RNAi
. The use of transient RNAi, however, introduces uncertainties
regarding efficiency of knock down and reproducibility.
We present two cell lines in which endogenous Mps1 has been
stably replaced with gatekeeper mutants. In-depth analysis of these
clonal cell lines showed that addition of bulky PP1 analogs allowed
rapid and reversible inhibition of Mps1 kinase activity in a highly
penetrant and reproducible manner. Using these cell lines we
dissected the different roles of Mps1 in mitosis, further establishing
known functions in checkpoint regulation and error-correction
and providing novel insights in the function of Mps1 catalytic
activity in checkpoint regulation.
Mps1 is selectively inhibitable in two engineered cell
To study the role of Mps1 in different mitotic processes, we
created cell lines in which Mps1 kinase activity could specifically
and reversibly be inhibited. For this, we engineered the ATP-
binding pocket of Mps1 to create a kinase with unique preference
for bulky ATP-like small molecules . As previously shown for
Mps1 in a transient expression system , Mps1M602Asignifies
such an analog sensitive (as) version. In addition to M602A, we
also tested M602G, which similarly enlarges the ATP-binding
pocket of Mps1 and may provide stronger inhibition . To
ensure that mutation of the gatekeeper residue did not affect Mps1
functionality, the ability of both Mps1M602Aand Mps1M602Gto
sustain an active checkpoint in Mps1 RNAi cells was investigated.
Similar to LAP-MpsWT, both LAP-Mps1M602A
Mps1M602Greconstituted checkpoint signaling in response to
nocodazole when endogenous Mps1 was transiently replaced with
these proteins (Fig. S1A).
To create stable cell systems in which Mps1 could be inhibited
in a potent and reproducible manner, we stably expressed LAP-
Mps1M602A(hereafter referred to as Mps1as) or LAP-Mps1M602G
in UTRM10 (U2OS-derived) and HCT-TRM (HCT116-derived)
cell lines in which endogenous Mps1 could be removed by
doxycycline-induced expression of Mps1 shRNA . Continuous
growth in doxycycline and clonal selection resulted in stable, viable
clonal cell lines in which endogenous Mps1 was undetectable and
replaced with Mps1WTor Mps1as(Fig. 1A). No HCT-TRM clones
were obtained and only one such
UTRM10 clone grew out. We suspect that the low activity of
Mps1M602G(see below) significantly decreased the chance of
survival of cells expressing only this form of Mps1. Nevertheless,
one UTRM10 cell line expressing Mps1M602Gwas established,
showing that under certain circumstances, Mps1M602Gcan
support viability. All cell clones (renamed HCT-Mps1WT, HCT-
Mps1as, UTR-Mps1WT, UTR-Mps1asand UTR-Mps1M602G)
have been kept in culture for months, showing that replacement
of Mps1 is stable.
Human Mps1 auto-activates by phosphorylating T676 in its
activation loop. T676 phosphorylation is essential for full kinase
activity [16,22,23], and the phosphorylation status of T676 serves
as a read-out for Mps1 kinase activity . To investigate if
Mps1asactivity could be inhibited by bulky PP1 analogs in our
engineered cell lines, we examined Mps1 T676 phosphorylation in
different conditions. Following tests on the potency of different
PP1 analogs to inhibit Mps1as(Fig. S1B), we settled on 23dMB-
midin-4-amine) as the Mps1asinhibitor of choice. As expected,
Mps1WTand Mps1asdisplayed a mobility shift and phosphory-
lation of T676 upon nocodazole addition in both HCT and UTR
cell types (Fig. 1B). Whereas addition of 23dMB-PP1 did not affect
the mobility or T676 phosphorylation of Mps1WT, 23dMB-PP1
prevented hyperphosphorylation as well as T676 phosphorylation
of Mps1as(Fig. 1B). Full inhibition of Mps1aswas established
within 5 minutes after addition of 23dMB-PP1 (Fig. 1C). Further-
more, T676 phosphorylation could be detected 15 minutes after
removal of 23dMB-PP1, showing that this rapid inhibition was
reversible (Fig. 1D). Interestingly, very little T676 phosphorylation
and weak mobility shift was detected in the UTR-Mps1M602G
clone, indicating this engineered Mps1 kinase has strongly reduced
activity compared to Mps1WTand Mps1as(Fig. S2A). Although
any residual activity and all Mps1 functions could be efficiently
inhibited by treatment of the UTR-Mps1M602Gcells with PP1
inhibitors (Fig. S2 and S3B), we focused our investigations on
studies with the Mps1asclones.
Mps1asinhibition by 23dMB-PP1 disables mitotic
Mps1 kinase activity is needed for the mitotic checkpoint to
delay mitosis when unattached kinetochores persist [6,9,12,18,21].
We next studied the ability of the cell lines to maintain a mitotic
checkpoint response in the presence of 23dMB-PP1. In absence of
23dMB-PP1, all HCT-Mps1WTand HCT-Mps1ascell lines
sustained checkpoint activity in the presence of nocodazole as
scored by mitotic index (Fig. 2A). The ability to maintain a
nocodazole-induced mitotic delay was not lost in Mps1WTcells
upon treatment with 23dMB-PP1, showing 23dMB-PP1 did not
have off-target effects that could disable the mitotic checkpoint. In
contrast to Mps1WT-expressing cells, 23dMB-PP1 completely
abolished mitotic checkpoint activity in Mps1ascells. Similar
results were obtained with the UTR-Mps1M602Gcells (Fig. S2B).
To ensure that Mps1 inhibition decreased mitotic index by
inhibiting mitotic checkpoint activity rather than, for instance,
preventing entry into mitosis, mitotic progression in the presence
of spindle poisons was monitored by live cell differential
interference contrast (DIC) microscopy. As a measure for time
spent in mitosis, the time from nuclear envelope breakdown (NEB)
to the first signs of furrow ingression was scored. Engaging the
mitotic checkpoint in these cells by treatment with the spindle
drugs nocodazole or taxol, or the Eg5 inhibitor STLC caused all
cell lines to delay mitosis for hours (Fig. 2B–F and S3). Delays were
most pronounced in cells with the HCT-TRM background where
,95% of cells maintained the arrest for at least 4 hours (Fig. 2C,
Chemical Genetics of Mps1
PLoS ONE | www.plosone.org2April 2010 | Volume 5 | Issue 4 | e10251
D and S3A, C) and ,55% for at least 10 hours (data not shown).
In stark contrast, addition of 23dMB-PP1 caused all HCT-Mps1as
but not HCT-Mps1WTcells to exit mitosis within 60 minutes
(Fig. 2D and S3A, C). These data show that the ability of cells to
delay mitotic progression absolutely requires Mps1 kinase activity.
Moreover, the finding that all 23dMB-PP1-treated HCT-Mps1as
cells had lost this ability illustrated the clonality of this cell line.
Mps1 kinase activity promotes localization of Mad1,
Mad2, Cdc20 and Bub1 to unattached kinetochores
Proper functioning of the mitotic checkpoint requires the
recruitment of checkpoint proteins to the kinetochore. Kinase
activity of Mps1 is required for the recruitment of Mad2 to
unattached kinetochores, at least partly explaining its necessity for
checkpoint signaling [9,12,21]. There is, however, discrepancy
between reports on the requirement of Mps1 for the recruitment of
Mad1 [11,12,21,24], Bub1 [11,12,24,25], BubR1 [11,12,18,23]
and CENP-E [9,12,21,24,26] to kinetochores.
To examine how Mps1 catalytic activity contributes to
recruitment of these and other checkpoint proteins, we assayed
their kinetochore association after inhibition of Mps1asin our cell
lines. Like the ability to delay mitosis, kinetochore recruitment of
Mad1, Mad2, BubR1, Bub1, CENP-E and Cdc20 was altered
neither in DMSO-treated Mps1WTor Mps1ascells, nor in 23dMB-
PP1-treated Mps12WTcells (Fig. 3 and S4). However, addition of
23dMB-PP1 to Mps1ascells significantly reduced the level of
Mad1, Mad2, Bub1 and Cdc20 at unattached kinetochores, but
only marginally affected BubR1 levels and did not influence
CENP-E levels (Fig. 3). These data show that Mps1 kinase activity
promotes kinetochore-binding of Mad1, Mad2, Bub1 and Cdc20.
Mps1 inhibition strongly diminishes MCC levels in mitotic
The mitotic checkpoint functions through the production of an
inhibitory complex called the mitotic checkpoint complex (MCC).
The MCC targets Cdc20, the main activator of the anaphase
promoting complex or cyclosome (APC/C) in prometaphase,
thereby preventing exit from mitosis . The amplification of
MCC assembly by unattached kinetochores is dependent on
kinetochore localization of Mad1 and Mad2 [27,28]. Since
inhibition of Mps1 resulted in loss of kinetochore-bound Mad1/
Mad2, we set out to investigate if Mps1 inhibition influences the
amount of MCC in nocodazole-treated cells. As expected, Cdc20
immunoprecipitations from cells treated with DMSO contained
the MCC subunits Mad2, BubR1 and Bub3 (Fig. 4A). Although
inhibition of Mps1 did not alter the levels of coprecipitated BubR1
and Bub3, the amount of Cdc20-bound Mad2 in mitotic cells was
reduced to interphase levels after treatment with 23dMB-PP1
Figure 1. Stable expression of LAP-Mps1asrenders selective sensitivity to 23dMB-PP1. A: Mps1 and tubulin immunoblots of lysates of
UTRM10 and HCT-TRM cells in which endogenous Mps1 is removed by doxycycline-induced expression of Mps1 shRNA and replaced with the
indicated LAP-Mps1 mutants. B–D: pT676-Mps1 and Mps1 immunoblots of UTR-Mps1as(B) and HCT-Mps1as(B–D) cell lines treated with nocodazole
and MG132 in combination DMSO/23dMB-PP1 for 1 hour (B), with nocodazole and MG132 for 1 hour in combination with an additional 23dMB-PP1
treatment for the indicated amounts of time (C), or with nocodazole and MG132 in combination with 23dMB-PP1 for 1 hour, followed by PBS wash
after which cells were left for 15 or 30 minutes in media without 23dMB-PP1 (D).
Chemical Genetics of Mps1
PLoS ONE | www.plosone.org3April 2010 | Volume 5 | Issue 4 | e10251
(Fig. 4A). Similarly, Mad2 immunoprecipitations displayed little
bound Cdc20 in cells treated with 23dMB-PP1 when compared to
DMSO-treated cells (Fig. 4B). These data show that Mps1
catalytic activity is needed for formation and/or stability of MCC.
Mps1 inhibition decreases the time spent in mitosis
To get more detailed insight in the functions of Mps1 kinase
activity during mitosis, we examined various aspects of chromo-
some segregation in our cell lines by time-lapse microscopy.
Mps1WTcells treated with DMSO or 23dMB-PP1 progressed
through mitosis normally. In addition, 23dMB-PP1 treatment of
Mps1as-expressing cells caused no overt problems in mitotic entry,
bipolar spindle formation or cytokinesis in neither cell line.
Strikingly, however, 23dMB-PP1-treated Mps1as
through mitosis (Fig. 5A). To more carefully quantify this
phenotype, the time from NEB to the first signs of furrowing
Figure 2. Mps1 activity is essential for the mitotic checkpoint. A: Quantification of flow cytometric analysis of the fraction of mitotic cells
treated with DMSO or 23dMB-PP1 alone or in combination with nocodazole for 14 hours. Graph represents average of three experiments (+/2SD). B–
F: Time-lapse analysis by DIC microscopy of HCT- and UTR-derived cells treated with taxol (HCT-derived cells) or STLC (UTR-derived cells) in
combination with DMSO or 23dMB-PP1. Images in B display morphology of HCT-Mps1ascells imaged at the indicated timepoints after addition of
taxol in combination with DMSO or 23dMB-PP1. Graphs in C and E represent percentages of cells delayed in mitosis for 4 hours or longer.
Percentages are indicated at top of each column. Line graphs in D and F show cumulative percentages of cells that display first signs of furrowing at
indicated times after NEB. Dashed lines indicate half-time to furrowing of inhibited drug-treated Mps1ascells.
Chemical Genetics of Mps1
PLoS ONE | www.plosone.org4April 2010 | Volume 5 | Issue 4 | e10251
was measured. Half of DMSO-treated HCT-Mps1WTand HCT-
cells had proceeded to furrow ingression 26 and
22 minutes after NEB, respectively (Fig. 5B). Strikingly, whereas
addition of 23dMB-PP1 had no effect on the time to furrow
ingression of HCT-Mps1WTcells, half the 23dMB-PP1-treated
HCT-Mps1ascells showed signs of furrow ingression in as little as
10 minutes and no cell took longer than 20 minutes to reach this
furrowing stage of mitosis (Fig. 5A, B). Further evidence for the
high penetrance of Mps1 inhibition in these cells was illustrated by
the fact that exit from mitosis in taxol was within a similar
Figure 3. Mps1 kinase activity promotes localization of Mad1, Mad2, Bub1 and Cdc20 to unattached kinetochores. A, B:
Immunolocalization of Mad1, Mad2, Bub1, BubR1, CENP-E, Cdc20, and centromeres (ACA) in HCT-derived (A) or UTR-derived (B) cells treated with
nocodazole and MG132 for 1 hour in combination with DMSO or 23dMB-PP1. Graphs represent quantifications of kinetochore signal intensities for
the investigated proteins at all kinetochores in a single cell as a ratio of the ACA signal. Data are average of three experiments, 10 cells per experiment
(+/2 SD). Representative images are shown.
Chemical Genetics of Mps1
PLoS ONE | www.plosone.org5 April 2010 | Volume 5 | Issue 4 | e10251