The CPC [chromosomal passenger complex; INCENP (inner centromere protein), Aurora B kinase, survivin and borealin] is implicated in many mitotic processes. In the present paper we describe how we generated DT40 conditional-knockout cell lines for incenp1 and survivin1 to better understand the role of these CPC subunits in the control of Aurora B kinase activity. These lines enabled us to reassess current knowledge of survivin function and to show that INCENP acts as a rheostat for Aurora B activity.
Accurate chromosome segregation requires that kinetochores of sister chromatids bind microtubules that emanate from opposing spindle poles. During prometaphase, various aberrant kinetochore attachments occur which, if not corrected, can lead to improper chromosome segregation and aneuploidy. To avoid this, cells have a surveillance mechanism, called the spindle checkpoint, which delays anaphase until all sister kinetochores are properly captured and under tension. Kinetochore attachment and error correction are directly controlled by the Aurora B kinase, part of the CPC (chromosomal passenger complex).
In most organisms, the core CPC is composed of Aurora B kinase  and three non-enzymatic subunits, INCENP (inner centromere protein) [2,3], survivin and borealin/Dasra B [4–7], that control targeting, enzymatic activity and stability of Aurora B kinase . The CPC controls many aspects of mitosis ranging from chromosome and spindle structure to the correction of kinetochore–microtubule attachment errors, regulation of mitotic progression and completion of cytokinesis . Knockdown by RNAi (RNA interference) of any member of the complex disrupts mitotic progression [4,7,10–13]. Although INCENP has been shown to be responsible for the initial Aurora B activation through direct binding and a phosphorylation feedback loop [9,14,15], the mechanisms responsible for controlling the activity of this important kinase are still poorly understood.
The ability to shut down/off the expression of a gene by RNAi or knockout is widely used to study protein function in many eukaryotic systems. Whereas RNAi relies upon the destruction of the mRNA, which is continuously produced, conditional knockouts block production of the mRNA at its source. Another advantage of knockout strategies over RNAi is the absence of off-target effects . Several cell types are used to generate knockouts including mouse ES (embryonic stem) cells, human RPE, Nalm-6, HT1080 and chicken DT40 cell lines [17–22].
DT40 cells are chicken B lymphoma with a very high rate of homologous recombination (up to 90% for some loci), greatly facilitating gene targeting. DT40 cells have been used to study many biological processes, including DNA damage pathways , transcriptional regulation , calcium signalling , apoptosis [26–29] and chromatin structure [30–32].
Assessment of survivin function in mitosis and apoptosis
Survivin is a protein involved in mitotic progression and apoptotic regulation that is up-regulated in many human tumours. Survivin's expression peaks during mitosis [8,35,36]. The protein is required for targeting the CPC to centromeres [8,37]. Survivin is considered a member of the IAP (inhibitor of apoptosis protein) family, despite lacking some of the key features shared by other members of the family. The exact role of survivin in mitosis and apoptosis remains unclear.
Our laboratory generated two DT40 conditional knockouts of the survivin gene. We used this system to study the role of survivin both in mitosis and apoptosis and to assess the role of key residues of the proteins in these biological processes.
survivin−/− cells were generated by deleting the entire survivin1 gene (Ensembl number ENSGALG00000008713; Figure 1A). Briefly, after targeting and replacement of the first allele by a resistance marker, the survivin cDNA, regulated by a Tet-off system, was introduced into the heterozygous cells. The second allele was then targeted, generating a conditional survivin-knockout cell line which, upon addition of doxycycline exhibits the null phenotype. Stable transfection of this cell line with various survivin cDNA-bearing mutations of interest allowed us to study their phenotypes in a null background.
Schematic representation of the DT40 gene disruption and ‘conditional promoter’
The role of survivin in mitosis
Using our unique survivin−/− model, we confirmed some previous findings on survivin function and revealed important features of key residues of the protein. For instance, we showed that survivin is essential for both chromosome segregation and cytokinesis, concordant with observations from RNAi studies [12,13]. A mutant previously shown to abolish interaction between survivin and Aurora B (human survivinD70A/D71A, chicken survivinD72A/D73A; henceforth, the amino acid numbering corresponds to chicken proteins)  introduced into the survivin-knockout cells was unable to concentrate normally at centromeres. This induced delocalization of the other passenger proteins in early mitosis, surprisingly without affecting subsequent transfer of the CPC to the spindle midzone. Furthermore, this mutant showed a compromised spindle checkpoint. Unexpectedly, despite the presence of these defects, this mutant could progress through mitosis and sustain cell growth, suggesting that survivin binding to Aurora B is not absolutely required for cell survival . In contrast with RNAi studies, we did not observe any significant increase in defects in early phases of mitosis such as chromosome misalignment or abnormal spindles in the absence of survivin [12,13,39–41]. survivin−/− cells exhibited mitotic arrest at high doses of taxol, whereas at lower doses they could override the spindle checkpoint and progress through mitosis. This override was concomitant with increased apoptosis, suggesting abnormal mitotic exit and subsequent cell death occurring in interphase .
The role of survivin in cell death
Similar to what we observed for mitosis, survivin−/− cells enabled us to confirm some previously described results regarding the role of survivin in apoptosis and to contradict others. For instance we could confirm that some mutations in specific domains of survivin did induce a loss-of-function phenotype, such as point mutations in the BIR domain (survivinC59A or survivinC86A) . On the other hand, loss of survivin did not induce an increased sensitivity to pro-apoptotic stimuli. Furthermore, abolition of survivin binding to Smac (second mitochondrial-derived activator of caspase) (survivinD55A), an antagonist of survivin involved in apoptosis, did not prevent cell growth nor did it display a pro-apoptotic phenotype, contradicting previous reports regarding the protective role of survivin in apoptosis [42–46].
Generation of DT40 survivin−/− cells expressing various mutants allowed us to reassess survivin functions both in mitosis and in apoptosis in complete absence of the endogenous wild-type protein. The null background proved to be essential to study survivin function. Discrepancies observed between our study and previous reports demonstrate the importance of a genetically clean system to dissect the protein functions. Our present study ultimately proved that cell death induced by loss of survivin is linked to cell-cycle defects (primarily a failure in cytokinesis) rather than disregulation of apoptosis.
Functional analysis of INCENP–Aurora B interactions
INCENP is the scaffolding protein of the CPC [2,3]. Its N-terminus is required for centromere targeting of the complex, whereas its C-terminus contains the IN box where Aurora B binds and gets activated. To study in more detail the interactions between INCENP and Aurora B, we generated a conditional DT40 knockout for incenp1 (Ensembl number ENSGALG00000007537), expressing INCENP bearing specific point mutations in the IN box. We chose residues predicted to be important for activation (Trp766) or binding of Aurora B (Phe802) to INCENP .
To generate an incenp1 knockout, we used a promoter hi-jack strategy (Figure 1B) [33,47]. Here, the first allele was placed under the control of a Tet-off system, which was achieved by replacing the endogenous promoter with a minimal CMV (cytomegalovirus)/TetO (Tet operator) promoter, together with stable expression of the tTA/VP16 transactivator under the control of the kif4A promoter. The other incenp1 allele had its ORF (open reading frame) disrupted, potentially allowing expression of only the first 28 amino acids of the protein. In these cells, the addition of doxycycline shuts off expression from the endogenous incenp1 allele (Figure 2). This strategy enabled the conditional expression of the multiple spliced isoforms of INCENP (class I and class II) .
The conditional INCENP knockout model
Regulation of Aurora B activity through INCENP binding
We used the system described above to study how interactions between INCENP and Aurora B affect the activity of the kinase along with CPC function and localization. The two mutants generated had a different binding capacity to Aurora B, leading to different levels of kinase activity. Expression of INCENPW766G in the absence of wild-type INCENP abolished the Aurora B interaction with INCENP, reducing Aurora B activity by 70%. The INCENPF802A mutant bound Aurora B in a normal fashion, but kinase activity was reduced by 50%, contrasting with predictions from previous structural analysis . Expression of either INCENPW766G or INCENPF802A in a null background induced cytokinesis failure and rapid cell death .
Both INCENPW766G and INCENPF802A were properly localized at centromeres in early mitosis, but failed to transfer to the spindle midzone in anaphase. We could thus demonstrate that mitotic entry and sister chromatid separation are independent of CPC formation or Aurora B activity, whereas both are required for proper transfer of the CPC to the spindle midzone and completion of cytokinesis . Despite reports from RNAi experiments showing that Aurora B activity is required for outer kinetochore assembly [49,50], we were able to detect CENP-A, -H, -O, -T, -E and the Hec1–Ndc80 complex at kinetochores of prometaphase incenp1−/− cells, suggesting that kinetochore assembly is not affected by the lack of CPC at centromeres . Moreover, CENP-E and PRC-1 (protein regulation of cytokinesis 1) localization to the spindle midzone of incenp1−/− anaphase cells demonstrated that a spindle midzone was present, despite the absence of CPC .
Expressing INCENP mutants in a null background enabled us to analyse the relationship between Aurora B kinase activity and the spindle checkpoint response induced by taxol. We established that various Aurora B activity levels yielded different spindle checkpoint responses when cells were treated with low doses of taxol. A minimal Aurora B activity (~50%, INCENPF802A) was sufficient to trigger a checkpoint response, whereas lower activity (~30%, INCENPW766G or 15% with Aurora B inhibitor) only induced a weak checkpoint response . This probably reflects the ability of the kinase to correct kinetochore–microtubule misattachments.
We mutated residues of INCENP that were predicted by crystallography studies to affect Aurora B binding or activity. Using incenp1−/− cells, we could assess the in vivo effects of mutating these residues. We were able to show that modulation of Aurora B activity regulates CPC localization and function during mitosis. Ultimately, the use of DT40 incenp1−/− cells expressing various mutants allowed us to strengthen the model in which INCENP acts as a rheostat for Aurora B activity .
The DT40 cell line is a powerful system to study protein function. The ability to perform conditional knockouts of essential multiply spliced cell-cycle-regulated genes makes it an ideal system to screen for and characterize functional domains or residues through the analysis of specific mutants against a null background.
Chromosome Segregation and Aneuploidy: An Independent Meeting held at the Royal College of Surgeons, Edinburgh, U.K., 19–23 June 2010. Organized and Edited by Bill Earnshaw (Wellcome Trust Centre for Cell Biology, Edinburgh, U.K.), Kevin Hardwick (Wellcome Trust Centre for Cell Biology, Edinburgh, U.K.) and Margarete Heck (Queen's Medical Research Institute, Edinburgh, U.K.).
S.R. is funded by ‘Ligue contre le Cancer du Grand-Ouest’ comity (districts: 22, 29, 56 and 79). Work in the Earnshaw Laboratory is supported by the Wellcome Trust, of which he is a Principal Research Fellow.