Oncogene (2008) 27, 4122–4127
& 2008 Nature Publishing Group All rights reserved 0950-9232/08 $30.00
www.nature.com/onc
SHORT COMMUNICATION
Targeted disruption of Aurora A causes abnormal mitotic spindle assembly,
chromosome misalignment and embryonic lethality
K Sasai1,4, JM Parant2,4,5, ME Brandt1, J Carter1, HP Adams2, SA Stass3, AM Killary2,
H Katayama1 and S Sen1
1
Department of Molecular Pathology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA; 2Department of
Molecular Genetics, The University of Texas MD Anderson Cancer Center, Houston, TX, USA and 3Department of Pathology,
The University of Maryland Medical School, Baltimore, MD, USA
Aurora A (also known as STK15/BTAK in humans),
a putative oncoprotein naturally overexpressed in many
human cancers, is a member of the conserved Aurora
protein serine/threonine kinase family that is implicated in
the regulation of G2–M phases of the cell cycle. In vitro
studies utilizing antibody microinjection, siRNA silencing
and small molecule inhibitors have indicated that Aurora
A functions in early as well as late stages of mitosis.
However, due to limitations in specificity of the techniques, exact functional roles of the kinase remain to be
clearly elucidated. In order to identify the physiological
functions in vivo, we have generated Aurora A null
mouse embryos, which show severe defects at 3.5 d.p.c.
(days post-coitus) morula/blastocyst stage and lethality
before 8.5 d.p.c. Null embryos at 3.5 d.p.c. reveal growth
retardation with cells in mitotic disarray manifesting
disorganized spindle, misaligned and lagging chromosomes as well as micronucleated cells. These findings
provide the first unequivocal genetic evidence for an
essential physiological role of Aurora A in normal mitotic
spindle assembly, chromosome alignment segregation and
maintenance of viability in mammalian embryos.
Oncogene (2008) 27, 4122–4127; doi:10.1038/onc.2008.47;
published online 17 March 2008
Keywords: mouse Aurora kinase A (Aurka) gene; mitotic
spindle; chromosome segregation; Cre-LoxP recombination; morula/blastocyst
Aurora A, a member of the conserved serine/threonine
protein kinase family represented by the prototypic Ipl1
kinase in yeast (Carmena and Earnshaw, 2003), has been
identified as a mitosis regulatory protein and a putative
oncoprotein that is naturally overexpressed in many human
Correspondence: Dr S Sen, Department of Molecular Pathology,
The University of Texas MD Anderson Cancer Center, 7435 Fannin,
Houston, TX 77054, USA.
E-mail: [email protected]
4
These authors contributed equally to this work
5
Current address: Department of Oncological Sciences, Huntsman
Cancer Institute, Salt Lake City, UT 84112, USA.
Received 4 January 2008; revised 28 January 2008; accepted 4 February
2008; published online 17 March 2008
cancers (Katayama et al., 2003). Aurora A expression peaks
during G2–M phase of the cell cycle localizing predominantly on the centrosomes and proximal mitotic spindle.
In vitro studies have implicated Aurora A in the
regulation of centrosome maturation separation, mitotic
entry, bipolar spindle assembly, chromosome alignment
and mitotic exit (Berdnik and Knoblich, 2002; Giet
et al., 2002; Conte et al., 2003). Identification of
microtubule-associated proteins as substrates of Aurora
A (Yu et al., 2005; Koffa et al., 2006) as well as their
localization on k fibers in prometaphase human cells
(Sillje et al., 2006) indicates that Aurora A forms
complex with mitotic spindle assembly factors in the
organization of the spindle microtubules. Additionally,
while silencing of Aurora A has been reported to cause
delay or block in mitotic entry (Marumoto et al., 2002;
Hirota et al., 2003; Du and Hannon, 2004; Satinover
et al., 2006), accelerated initiation of mitosis in presence
of active Aurora A has also been observed (Ma et al.,
2003; Liu and Ruderman, 2006). These effects may, in
part, be explained by the reported role of Aurora A in
facilitating expression and activation of critical mitosis
regulators cyclin B and Cdc25B (Mendez et al., 2000;
Tay et al., 2000; Dutertre et al., 2004). However,
involvement of Aurora A in regulating specific cellular
phenotypes still remain uncertain in view of the inherent
limitations in assessing the contributions of the endogenous proteins in ex vivo experiments and also due
to occasional contradictory results published on the
consequence of ablating Aurora A function in cells
grown in vitro (Girdler et al., 2006; Hoar et al., 2007).
In order to elucidate the physiological functions
of Aurora A in vivo, we have carried out targeted
disruption of the mouse Aurora kinase A (Aurka) gene
through Cre-loxP-mediated recombination in mice. The
results reveal that Aurora A is essential for proper
mitotic spindle assembly, chromosome alignment segregation and viability in mammalian embryos.
Results and discussion
Targeted disruption of the Aurka gene was achieved as
outlined in Figure 1a. Ten percent of the recombinant
Aurora A knockout mouse
K Sasai et al
4123
Figure 1 Targeted disruption of the mouse Aurka gene, genotyping and protein expression analyses of genotyped mice tissues.
(a) Targeting strategy for generating Aurora A floxed and null allele (). A CN allele was engineered into embryonic stem (ES) cells by
homologous recombination with a targeting construct consisting of the floxed ATG start codon containing exon 2 and the neomycin
resistance gene (Neo) flanked by the Frt sites. The wild-type allele ( þ ), the targeting vector and the recombinant conditional neo allele
(CN), conditional allele (c) and the null allele are shown. Exons are indicated by black boxes. The selection cassettes containing Neo
and TK are represented by white boxes, small gray boxes indicate the positions of the Southern blot probes. Neo was excised through
FLP-mediated recombination between two FRT sites (circles) to obtain the C allele. The null allele was obtained by Cre-mediated
recombination between two loxP sites (triangles). The arrows show primers (a, b, c, d and e) used for genotyping (Supplementary
Table 1). Restriction enzyme sites are KpnI, K; XbaI, Xb; SalI, S; PstI, P; BamHI, B. (b) For Southern blot analyses, mouse tail DNA
was digested with KpnI and XbaI, probed with 50 probe (top) and 30 probe (bottom). Genotypes are indicated on the top of the blot.
(c) Genotyping PCR of mouse tail DNA used in (b). CN PCR (top), þ /C PCR (middle) and Excision (Exc) PCR (bottom) were
performed with primer sets b/c, b/d and a/e, respectively as described in the Supplementary Table. Each allele was identified based on
the presence of at least two different sized products in the three PCR analyses except for the null allele, which was detected either by the
794 bp product in the ‘Exc PCR’ or by the absence of any product in the ‘ þ /C PCR’ and the ‘CN PCR’. (d) Western blotting of mouse
testis protein from the same mice used in (b and c) probed with anti-Aurora A antibody and anti-HSP90 antibody. Liver and NIH3T3
cell lysate were used as negative and positive controls respectively.
clones, selected for neomycin resistance in G418 and
against TK in gancyclovir to eliminate random insertions, were properly targeted giving rise to the conditional neomycin (CN) allele based on the Southern blot
hybridization results with the 50 and 30 probes. KpnI and
XbaI digestion of DNA from these clones revealed the
correct sized genomic fragments of 2 kb with the 50
probe and 5.4 kb with the 30 probe in addition to the
8 kb fragments from the wild-type allele ( þ ). Microinjection of two targeted cell lines resulted in germ-line
chimeras. The heterozygotes with the CN allele, derived
from the chimeras, were used to generate the conditional
(C) allele through Flp-mediated recombination. The CN
and the C mice were crossed with ZP3-Cre mice to
remove the floxed exon 2, generating the null allele ().
Homozygous and heterozygous mice with the wild-type,
the CN, the C and the null alleles were verified by
Southern blotting of the respective genomic DNA with
the 50 and 30 probes (Figure 1b) and also by PCR
analyses (Figure 1c) with five primers, ‘a’ through ‘e’,
utilized in three different combinatorial pairs (referred
to as ‘CN PCR’, ‘ þ /C PCR’ and the ‘Exc PCR’) as
mentioned in the Supplementary Table 1. Functional
status of the mutant alleles was determined by western
blot analyses performed with testis tissues lysate from
mice with different genotypes since Aurora A protein is
clearly detectable in this tissue. The band intensities
suggested that the CN and the C alleles were expressing
comparable amounts of the protein like the wild-type
allele and the null allele was silenced (Figure 1d).
Aurka þ / intercrosses did not result in any homozygous nulls among live born offspring and 8.5 d.p.c.
(days post-coitus) embryos indicating early embryonic
lethality of this genotype. Among 61 live born mice and
40 embryos at 8.5 d.p.c., wild-type genotype was
detected in 20 and 14, while the numbers of heterozygote
were 41 and 26, respectively. Thus in contrast to the
expected ratio of 1:2:1 for the wild-type homozygote:
heterozygote and homozygote nulls, the observed ratio
of 1:2:0 suggested embryonic lethality of the null
embryos before 8.5 d.p.c. To determine the timing of
embryonic lethality, 40 embryos at 3.5 d.p.c. (morula/
blastocyst) were genotyped. This analysis revealed the
presence of 8 Aurka/ in addition to 12 wild-type and
Oncogene
Aurora A knockout mouse
K Sasai et al
4124
20 heterozygotes (Figures 2a and b). The results,
therefore, demonstrated that Aurka/ embryos were
viable at 3.5 d.p.c. in utero but failed to grow further and
implant.
The 3.5 d.p.c. embryos from heterozygous crosses
were then analyzed for ex vivo growth potential by
culturing in vitro. The embryos were photographed
immediately (3.5 d.p.c.) and up to 3 days of in vitro
culture (E 4.5–E 6.5) followed by genotyping. Total of
four wild-type homozygous, nine heterozygous and five
null homozygous embryos were analyzed. The 3.5 d.p.c.
null mutants were invariably smaller with fewer cells and
cells of irregular size in the inner cell mass compared
with the wild-type and the heterozygous embryos. After
1 day culture (E 4.5), the wild-type and the heterozygous
embryos showed similar growth but the null embryos
remained noticeably smaller with some revealing degeneration of the inner cell mass (Figure 2c). The embryos
genotyped at the end of the second (E 5.5) and third
(E 6.5) days of culture did not reveal any null mutants
with only the wild-type homozygote or heterozygote
detected (data not shown). The results indicated that the
null mutants degenerated sometime between days 4.5
and 6.5 of in vitro culture.
The 3.5 d.p.c. embryos derived from Aurka þ / intercrosses were also analyzed for total cellular content,
nuclear morphology, M phase distribution of cells and
mitotic structures. Due to the small size and fragility of
these embryos, both immunofluorescence staining and
Figure 2 Observed and expected numbers of genotyped offspring
from Aurka þ / intercrosses and in vitro growth characteristic of day
3.5 embryos. (a) Observed and expected numbers of genotyped
progeny at the age of weaning (>21 days), postimplantation
(8.5 d.p.c.) and at the morula/blastocyst stage (3.5 d.p.c.) from
Aurka þ / intercrosses. (b) Genotypes were ascertained based on the
presence of the wild-type and the null allele-specific 322 and the
794 bp fragments generated by the ‘ þ /C PCR’ (trop) and the ‘Exc
PCR’ (bottom). (c) Ex vivo morphology of day 3.5 embryos (E 3.5)
on isolation and following 1 day of in vitro culture (E 4.5) with
different Aurora A genotypes confirmed by PCR.
Oncogene
PCR genotyping was difficult and in most instances only
one of the two methods could be successfully performed.
About a quarter of the embryos that were Aurora Anegative and distinctly smaller in size consisted of fewer
cells with signs of degeneration, as in the case of the null
embryos utilized for ex vivo growth study. Examination
of seven 3.5 d.p.c. Aurora A-negative and ten positive
embryos revealed clear evidence of growth impairment
in the cells of the null embryos. The positive embryos
consisted of an average of 40 cells but the negative
embryos had about 18 cells, with some appearing more
like morulae rather than blastocysts. The average
numbers of mitotic cells and micronuclei were about 5and 10-fold higher, respectively, in the negative embryos
compared with the positive embryos (Figure 3a). The
mitotic cells of the negative embryos displayed disorganized spindle assembly and majority (B69%) of these
cells were in prometaphase, with condensed chromosomes not congressed or aligned on the metaphase plate
(Figure 3b), while about 27 and 4% were in metaphase
and anaphase, respectively (Figure 3c). The null cells
also manifested misaligned and lagging chromosomes at
metaphase and anaphase stages (Figure 3d). The Aurora
A-positive embryos revealed a more even distribution of
cells at different phases of mitosis with about 33% in
prometaphase, 44% in metaphase and 22% in anaphase
displaying predominantly normal alignment and segregation of chromosomes. Presence of an average of 18
cells in the null 3.5 d.p.c. morulae/blastocysts indicates
that these embryos are able to undergo at least four
mitotic cell divisions despite lacking a functional Aurora
A allele. It is plausible that initial cell divisions in the
null embryos are facilitated by the presence of a low
amount of maternal protein since the Unigene cluster
database reveals existence of murine Aurka transcript
(Mm 249363) in oocyte and zygote. Additionally, the
presence of a relatively elevated number of cells in
mitosis in the null embryos indicate that mitosis may be
initiated with a delay rather than inhibited in absence of
Aurora A as has been reported in case of Xenopus early
embryonic cell cycles investigated with egg extracts
depleted of endogenous Aurora A (Liu and Ruderman,
2006; Satinover et al., 2006).
Aberrant spindle assembly is expected due to impaired formation of k fibers requiring active Aurora A
complex with spindle assembly factors, such as TPX2
and HURP, among others (Gruss et al., 2001; Koffa
et al., 2006; Sillje et al., 2006; Tulu et al., 2006). It has
been hypothesized that microtubules can be organized at
multiple, often transient, structures that can catalyze
g-tubulin-dependent microtubule nucleation from their
minus ends, plus ends or sides (Luders and Stearns,
2007) with the polar spindle microtubules ‘collecting’
preassembled chromosome-generated minispindles into
a single bipolar spindle within the cells (Rieder, 2005).
These proposed mechanisms imply that microtubules
nucleated on chromosomes and k fibers mediate capture
and stabilization of the sister kinetochores on the
bipolar spindle and raise the interesting scenario that
absence of Aurora A in the null embryos may be
interfering with the kinetochore capture process giving
Aurora A knockout mouse
K Sasai et al
4125
Figure 3 Fluorescence microscopy analyses of 3.5 d.p.c. embryos from Aurka þ / intercrosses. Embryos were fixed and stained
for detection of Aurora A and mitotic spindle morphology by immunofluorescence with anti-Aurora A and anti-a-tubulin antibodies,
while chromosomes were visualized by 4,6-diamidino-2-phenylindole (DAPI) staining. (a) Bar graph representation of the average
numbers of total cells, mitotic cells and micronucleated cells in the fluorescence stained Aurora A-positive and -negative embryos.
(b) Representative image of an Aurora A-negative embryo with cells in prometaphase displaying abnormal spindle morphology.
Microtubules stained with anti-a-tubulin antibodies (green) and DNA counterstained with DAPI (blue). (c) Representative images of
Aurora A-positive and -negative embryos stained for Aurora A (red), a-tubulin (green) and DNA (blue). Arrows indicate micronuclei.
I and II indicate misaligned and lagging chromosomes shown in the enlarged image in (d).
rise to the misaligned and lagging chromosomes. The
possibility appears conceivable in view of our recent
observation that Aurora A is a key regulator of
kinetochore-associated microtubules formation process
(Katayama et al., submitted).
Severe mitotic defects and early embryonic lethality in
Aurora A null mutant are similar to those seen in case of
null mutants for other known proteins regulating
chromosome segregation and cytokinesis, such as
CenpC (Kalitsis et al., 1998), Incenp (Cutts et al.,
Oncogene
Aurora A knockout mouse
K Sasai et al
4126
1999), CenpA (Howman et al., 2000) and Survivin (Uren
et al., 2000), as well as those involved in the activation of
the spindle assembly checkpoint network like, Mad2
(Dobles et al., 2000), BubR1 (Wang et al., 2004), Mad1
(Iwanaga et al., 2007) and Bub1 (Perera et al., 2007).
While defect in kinetochore-microtubule attachments is
expected to disrupt normal mitotic progression, apparent delay or arrest of mitosis and also chromosome
segregation anomalies in the Aurora A null cells indicate
that loss of Aurora A function influences the spindle
assembly checkpoint network. Induction of similar
chromosome segregation and cytokinesis abnormalities
in Aurora A overexpressing cells also tend to favor this
notion. Although elevated expression of Aurora A has
been reported to override the spindle assembly checkpoint (Anand et al., 2003), specific molecular targets of
Aurora A in the checkpoint network remain unknown.
Aurora A null allele containing cells will provide an ideal
model system for these investigations.
Abbreviations
C, conditional; CN, conditional neo.
Acknowledgements
This study was supported by grants from the National
Institutes of Health (RO1 CA89716) and Institutional
Research Grant award from the University of Texas MD
Anderson Cancer Center to SS. We thank Dr Richard
Behringer for critical reading of the paper and Dr Jan
Parker-Thornburg for help in this study. Institutional Biospecimen Extraction Facility, DNA Analysis Facility are supported by the Cancer Center Support Grant CA16672.
References
Anand S, Penrhyn-Lowe S, Venkitaraman AR. (2003). AURORA-A
amplification overrides the mitotic spindle assembly checkpoint,
inducing resistance to Taxol. Cancer Cell 3: 51–62.
Berdnik D, Knoblich JA. (2002). Drosophila Aurora-A is required for
centrosome maturation and actin-dependent asymmetric protein
localization during mitosis. Curr Biol 12: 640–647.
Carmena M, Earnshaw WC. (2003). The cellular geography of aurora
kinases. Nat Rev Mol Cell Biol 4: 842–854.
Conte N, Delaval B, Ginestier C, Ferrand A, Isnardon D, Larroque C
et al. (2003). TACC1-chTOG-Aurora A protein complex in breast
cancer. Oncogene 22: 8102–8116.
Cutts SM, Fowler KJ, Kile BT, Hii LL, O’Dowd RA, Hudson DF
et al. (1999). Defective chromosome segregation, microtubule
bundling and nuclear bridging in inner centromere protein gene
(Incenp)-disrupted mice. Hum Mol Genet 8: 1145–1155.
Dobles M, Liberal V, Scott ML, Benezra R, Sorger PK. (2000).
Chromosome missegregation and apoptosis in mice lacking the
mitotic checkpoint protein Mad2. Cell 101: 635–645.
Du J, Hannon GJ. (2004). Suppression of p160ROCK bypasses cell
cycle arrest after Aurora-A/STK15 depletion. Proc Natl Acad Sci
USA 101: 8975–8980.
Dutertre S, Cazales M, Quaranta M, Froment C, Trabut V, Dozier C
et al. (2004). Phosphorylation of CDC25B by Aurora-A at the
centrosome contributes to the G2–M transition. J Cell Sci 117:
2523–2531.
Giet R, McLean D, Descamps S, Lee MJ, Raff JW, Prigent C et al.
(2002). Drosophila Aurora A kinase is required to localize D-TACC
to centrosomes and to regulate astral microtubules. J Cell Biol 156:
437–451.
Girdler F, Gascoigne KE, Eyers PAH, Hartmuth S, Crafter C, Foote
KM et al. (2006). Validating Aurora B as an anti-cancer drug target.
J Cell Sci 119: 3664–3675.
Gruss OJ, Carazo-Salas RE, Schatz CA, Guarguaglini G, Kast J,
Wilm M et al. (2001). Ran induces spindle assembly by reversing
the inhibitory effect of importin alpha on TPX2 activity. Cell 104:
83–93.
Hirota T, Kunitoku N, Sasayama T, Marumoto T, Zhang D, Nitta M
et al. (2003). Aurora-A and an interacting activator, the LIM protein
Ajuba, are required for mitotic commitment in human cells. Cell
114: 585–598.
Hoar K, Chakravarty A, Rabino C, Wysong D, Bowman D, Roy N
et al. (2007). MLN8054, a small-molecule inhibitor of Aurora A,
causes spindle pole and chromosome congression defects leading to
aneuploidy. Mol Biol Cell 27: 4513–4525.
Howman EV, Fowler KJ, Newson AJ, Redward S, MacDonald AC,
Kalitsis P et al. (2000). Early disruption of centromeric chromatin
Oncogene
organization in centromere protein A (Cenpa) null mice. Proc Natl
Acad Sci USA 97: 1148–1153.
Iwanaga Y, Chi YH, Miyazato A, Sheleg S, Haller K, Peloponese JMJ
et al. (2007). Heterozygous deletion of mitotic arrest-deficient
protein 1 (MAD1) increases the incidence of tumors in mice. Cancer
Res 67: 160–166.
Kalitsis P, Fowler KJ, Earle E, Hill J, Choo KH. (1998). Targeted
disruption of mouse centromere protein C gene leads to mitotic
disarray and early embryo death. Proc Natl Acad Sci USA 95:
1136–1141.
Katayama H, Brinkley WR, Sen S. (2003). The Aurora kinases: role in
cell transformation and tumorigenesis. Cancer Metastasis Rev 22:
451–464.
Koffa MD, Casanova CM, Santarella R, Kocher T, Wilm M, Mattaj
IW. (2006). HURP is part of a Ran-dependent complex involved in
spindle formation. Curr Biol 16: 743–754.
Liu Q, Ruderman JV. (2006). Aurora A, mitotic entry, and spindle
bipolarity. Proc Natl Acad Sci USA 103: 5811–5816.
Luders J, Stearns T. (2007). Microtubule-organizing centres: a reevaluation. Nat Rev Mol Cell Biol 8: 161–167.
Ma C, Cummings C, Liu XJ. (2003). Biphasic activation of Aurora-A
kinase during the meiosis I- meiosis II transition in Xenopus oocytes.
Mol Cell Biol 23: 1703–1716.
Marumoto T, Hirota T, Morisaki T, Kunitoku N, Zhang D, Ichikawa
Y et al. (2002). Roles of aurora-A kinase in mitotic entry and G2
checkpoint in mammalian cells. Genes Cells 7: 1173–1182.
Mendez R, Hake LE, Andresson T, Littlepage LE, Ruderman JV,
Richter JD. (2000). Phosphorylation of CPE binding factor by Eg2
regulates translation of c-mos mRNA. Nature 404: 302–307.
Perera D, Tilston V, Hopwood JA, Barchi M, Boot-Handford RP,
Taylor SS. (2007). Bub1 maintains centromeric cohesion by
activation of the spindle checkpoint. Dev Cell 13: 566–579.
Rieder CL. (2005). Kinetochore fiber formation in animal somatic
cells: dueling mechanisms come to a draw. Chromosoma
114: 310–318.
Satinover DL, Brautigan DL, Stukenberg PT. (2006). Aurora-A
kinase and inhibitor-2 regulate the cyclin threshold for mitotic
entry in Xenopus early embryonic cell cycles. Cell Cycle 5:
2268–2274.
Sillje HH, Nagel S, Korner R, Nigg EA. (2006). HURP is a
Ran-importin beta-regulated protein that stabilizes kinetochore
microtubules in the vicinity of chromosomes. Curr Biol 16:
731–742.
Tay J, Hodgman R, Richter JD. (2000). The control of cyclin B1
mRNA translation during mouse oocyte maturation. Dev Biol 221:
1–9.
Aurora A knockout mouse
K Sasai et al
4127
Tulu US, Fagerstrom C, Ferenz NP, Wadsworth P. (2006). Molecular
requirements for kinetochore-associated microtubule formation in
mammalian cells. Curr Biol 16: 536–541.
Uren AG, Wong L, Pakusch M, Fowler KJ, Burrows FJ, Vaux DL
et al. (2000). Survivin and the inner centromere protein INCENP
show similar cell-cycle localization and gene knockout phenotype.
Curr Biol 10: 1319–1328.
Wang Q, Liu T, Fang Y, Xie S, Huang X, Mahmood R et al. (2004).
BUBR1 deficiency results in abnormal megakaryopoiesis. Blood
103: 1278–1285.
Yu CT, Hsu JM, Lee YC, Tsou AP, Chou CK, Huang CY. (2005).
Phosphorylation and stabilization of HURP by Aurora-A: implication of HURP as a transforming target of Aurora-A. Mol Cell Biol
25: 5789–5800.
Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)
Oncogene