Aurora-A Phosphorylates, Activates, and Relocalizes the Small GTPase RalA
The small GTPase Ras, which transmits extracellular signals to the cell, and the kinase Aurora-A, which promotes proper mitosis, can both be inappropriately activated in human tumors. Here, we show that Aurora-A in conjunction with oncogenic Ras enhances transformed cell growth. Furthermore, such transformation and in some cases also tumorigenesis depend upon S194 of RalA, a known Aurora-A phosphorylation site. Aurora-A promotes not only RalA activation but also translocation from the plasma membrane and activation of the effector protein RalBP1. Taken together, these data suggest that Aurora-A may converge upon oncogenic Ras signaling through RalA.
MOLECULAR AND CELLULAR BIOLOGY, Jan. 2010, p. 508–523 Vol. 30, No. 2
Copyright © 2010, American Society for Microbiology. All Rights Reserved.
Aurora-A Phosphorylates, Activates, and Relocalizes the Small
‡ Donita C. Brady,
‡ David F. Kashatus,
Brooke B. Ancrile,
Channing J. Der,
Adrienne D. Cox,
* and Christopher M. Counter
Department of Pharmacology and Cancer Biology, Department of Radiation Oncology, Duke University Medical Center, Durham,
North Carolina 27710,
and Departments of Pharmacology
and Radiation Oncology,
Lineberger Comprehensive Cancer Center,
University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599
Received 9 June 2008/Returned for modiﬁcation 22 July 2008/Accepted 26 October 2009
The small GTPase Ras, which transmits extracellular signals to the cell, and the kinase Aurora-A, which
promotes proper mitosis, can both be inappropriately activated in human tumors. Here, we show that Aurora-A in
conjunction with oncogenic Ras enhances transformed cell growth. Furthermore, such transformation and in some
cases also tumorigenesis depend upon S194 of RalA, a known Aurora-A phosphorylation site. Aurora-A promotes
not only RalA activation but also translocation from the plasma membrane and activation of the effector protein
RalBP1. Taken together, these data suggest that Aurora-A may converge upon oncogenic Ras signaling through
Ras small GTPases (H-, N-, and K-Ras) function as regu-
lated binary switches, typically at the plasma membrane,
whereby extracellular signal-stimulated cell surface receptors
stimulate guanine nucleotide exchange factors (GEFs) to pro-
mote GDP/GTP exchange to favor the formation of active,
GTP-bound Ras. This, in turn, induces a conformational
change in the effector binding domain in Ras, permitting the
binding and activation of effector proteins, such as Raf pro-
teins, phosphatidylinositol 3-kinase (PI3K), and RalGEF pro-
teins, that mediate Ras signaling (19, 53). One-third of human
cancers harbor point mutations in Ras that render the protein
in a constitutively active GTP-bound state, promoting a host of
cancer cell phenotypes (40).
Aurora-A belongs to a family of three related serine/threo-
nine mitotic kinases critical for many stages of mitosis. Studies
of a number of model systems indicate that Aurora-A phos-
phorylates a growing number of proteins in a spatially and
temporally restricted manner to ensure proper centrosomal
maturation and separation, mitotic entry, mitotic spindle as-
sembly, chromosome alignment and separation, and subse-
quent cytokinesis (18, 42). Overexpression of Aurora-A is seen
in human cancers (22, 31, 32) and can cause growth transfor-
mation of Rat1 and NIH 3T3 rodent ﬁbroblast cell lines (5).
That it can cause tumor formation in the mammary epithelia of
mice only after a long latency (60, 68) and that alone it does
not transform primary rodent cells (1) or induce pancreatic
cancer formation in mice (61) argue that Aurora-A acts in
concert with other changes to promote a transformed state.
Although the mechanism by which Aurora-A promotes on-
cogenesis remains to be understood, emerging evidence sug-
gests that Aurora-A may cooperate with the Ras oncoprotein.
First, activating mutations in KRAS occur in nearly all pancre-
atic cancers (28), and Aurora-A has been found to be overex-
pressed in this tumor type by gene ampliﬁcation (21) or by
elevated levels of mRNA or protein (21, 35). This overexpres-
sion likely fosters tumor growth, as suppression of Aurora-A
expression by interfering RNA or treatment with Aurora-A
inhibitors also impairs pancreatic cancer cell growth (26, 48).
Second, overexpression of Aurora-A enhances Ras-induced
transformation of murine 3T3A31-1 ﬁbroblasts (59). Third,
Aurora kinases physically interact with RasGAP in vitro (23),
and inhibition of Aurora-B binding to RasGAP causes apop-
tosis (49). Fourth, two components of the RalGEF-Ral effector
pathway of Ras, which is known to promote Ras oncogenesis
(38), are substrates for Aurora-A (66). Speciﬁcally, active Ras
binds to RalGEF proteins, a family of guanine nucleotide
exchange factors (GEFs) and activators of the related small
GTPases RalA and RalB. Both the RalGEF protein RalGDS
and RalA were shown to be phosphorylated by Aurora-A.
With regard to the latter, RalA is phosphorylated at S194 in its
C-terminal membrane binding domain, leading to elevated
levels of activated RalA-GTP. Constitutively active RalA
(G23V) also cooperated with ectopically expressed Aurora-A
to promote anchorage-independent growth of MDCK epithe-
lial cells, whereas a RalA mutant that could not be phospho-
rylated by Aurora-A (RalA
) was impaired in this
activity (66). Moreover, overexpression of both Aurora-A and
RalA mRNA is associated with advanced human bladder can-
cer (58). Finally, as indirect evidence for the importance of the
Aurora-A phosphorylation site in RalA, S194 and S183 were
identiﬁed as sites of dephosphorylation by the phosphatase
PP2A. Short hairpin RNA (shRNA) silencing of PP2A expres-
sion increased phosphorylation of S194 and S183 and forma-
* Corresponding author. Mailing address for Adrienne D. Cox:
Lineberger Comprehensive Cancer Center, University of North Caro-
lina at Chapel Hill, Chapel Hill, NC 27599. Phone: (919) 966-7712.
Fax: (919) 966-9673. E-mail: firstname.lastname@example.org. Mailing
address for Christopher M. Counter: Department of Pharmacology
and Cancer Biology, Department of Radiation Oncology, Duke
University Medical Center, Durham, NC 27710. Phone: (919) 684
-9890. Fax: (919) 684-8958. E-mail: email@example.com.
† Supplemental material for this article may be found at http://mcb
‡ These authors contributed equally to this study.
Published ahead of print on 9 November 2009.
tion of RalA-GTP, whereas replacing endogenous RalA with
S194A or S183A mutants resulted in a loss of tumorigenic
growth of human embryonic kidney (HEK) cells expressing
oncogenic H-Ras, hTERT, and the early region of simian virus
40 (SV40) (56). Given these observations, we explored the
molecular connection between Aurora-A and the Ras-Ral-
MATERIALS AND METHODS
Plasmids. pSuper-Retro-Puro plasmids encoding shRNA against RalA, RalB,
RalBP1 (5⬘-GTAGAGAGGACCATGATGT), or a scramble sequence; pBabe-
Neo plasmid encoding shRNA-resistant Myc-tagged RalA; pBabe-Puro plasmid
encoding shRNA-resistant Flag-tagged RalA or Q72L; pBabe-Bleo plasmid en-
coding hemagglutinin (HA)-tagged Rlf-CAAX; pBabe-Puro Ras
rived effector mutants; small t antigen (t-Ag); DsRed-Rab11; pMT3-mycRalBP1;
and pcDNA3-mycSec5 were previously described (16, 25, 36, 45). S194A and
S194D point mutations in RalA were introduced to pBabe-Neo and pBabe-puro
plasmids encoding shRNA-resistant RalA, respectively, by site-directed mu-
tagenesis, and a Myc epitope tag (9E10) or a Flag epitope tag was added to the
N terminus by PCR. Human Aurora-A was PCR ampliﬁed from a cDNA tem-
plate (MGC-1605; ATCC) to add an N-terminal HA-epitope tag, after which
K162R and T288D mutants were generated by site-directed mutagenesis. Re-
sultant wild-type (WT), K162R, and T288D cDNAs were cloned into pBabe-
Bleo or pBabe-Hygro for stable expression and pCGN for transient expression.
shRNA directed against Aurora-A (5⬘–ATGCCCTGTCTTACTGTCA) was
subcloned into pSuper-Retro-Puro-TET. Green ﬂuorescent protein (GFP)-RalA
fusion proteins were generated by subcloning wild-type RalA, RalA
, and RalA
in frame to the C terminus of GFP in pEGFP-C2.
HEK-TtH cells, 293 cells, human pancreatic cancer cell lines and derived cell
lines, inducible Aurora-A shRNA HPAC cells, and tissue samples. HEK-TtH
cells (human embryonic kidney cells stably expressing the SV40 early region and
hTERT) were previously described (24). 293, 293T, AsPC-1, HPAC, HPAF-II,
PANC-1, SW1990, Capan-1, Capan-2, CFPAC-1, and MIA PaCa-2 (38) were
purchased and cultured as suggested by supplier (ATCC), and T3M4 cells were
kindly provided by M. Korc (Dartmouth-Hitchcock Medical Center, Hanover,
NH). Where noted, these cell lines were stably infected with retroviruses encod-
ing the indicated shRNAs or transgenes as previously described (47). Generation
of HPAC cells expressing inducible Aurora-A shRNA was done as described
previously (37) except that a pSuper-Retro-Puro-TET plasmid (37) engineered
to encode Aurora-A shRNA (26) was used and, at the indicated times, the cells
were treated with either 60 mg/ml of doxycycline (Sigma) or a vehicle. Frozen
surgical samples of normal human pancreatic tissues were kindly provided by
A. D. Proia and D. Tyler (DUMC).
Immunoblotting. Whole-cell lysates isolated in Triton X-100 lysis buffer from
cultured cells (serum starved for 48 h for measurement of Ras effector pathways)
or isolated in radioimmunoprecipitation assay (RIPA) buffer from tissue mate-
rial were immunoblotted with the following antibodies, diluted according to the
manufacturers’ recommendations: antiactin, anti-RalBP1 (Santa Cruz), anti-
RalA, anti-RalB (BD Transduction Laboratories), anti-Aurora-A (Cell Signaling
Technology), and antitubulin (Sigma) (to detect endogenous actin, RalBP1,
RalA, RalB, Aurora-A, and tubulin proteins, respectively) and anti-HA (Roche),
anti-Myc 9E10 (Invitrogen), anti-RalA (BD Transduction Laboratories) and
anti-pan-Ras (Santa Cruz) (to detect the ectopic proteins HA-Aurora-A
, HA-Rlf-CAAX, shRNA-resistant
RalA [wild type, S194A, or S194D], and H-Ras or effector domain mutants,
Immunoprecipitation. Whole-cell lysates prepared in Triton X-100 buffer
from 293T cells transiently transfected with plasmids encoding FLAG epitope-
tagged RalA (wild type or S194D) and Myc-RalBP1 or Myc-Sec5 were subjected
to immunoprecipitation with anti-FLAG M2 resin (Sigma) and immunoblotted
with anti-FLAG M2 (Sigma) to detect FLAG-RalA and anti-Myc 9E10 (Invitro-
gen) to detect Myc-RalBP1 or Myc-Sec5. To detect phosphorylated RalA, cells
expressing a vector control or HA-tagged Aurora-A were subjected to immuno-
precipitation with anti-RalA antibody (BD Transduction Laboratories) and
immunoblotted with antiphosphoserine (4A4; Millipore) to detect serine-phos-
phorylated RalA. For detection of endogenous GTP-bound small GTPases,
GTP-RalA, GTP-Cdc42, and GTP-Rac1 were captured for pulldown analyses by
incubating cell lysates with glutathione-agarose-bound recombinant glutathione
S-transferase (GST)–RalBD for RalA or GST-PakBD for Cdc42 and Rac1 and
detected with the anti-RalA, anti-Cdc42, and anti-Rac1 antibodies (BD Trans-
duction Laboratories), respectively, as previously described (4, 65). The total
amount of the corresponding small GTPase in the whole-cell lysate served as a
Soft agar assay. Fifty thousand cells per 35-mm plate with a 2-mm grid were
suspended in soft agar as described previously (17, 25). In the case of comparing
-transformed HEK-TtH cells in the absence or presence of HA-Aurora-
, or HA-Aurora-A
cells were seeded for
analysis. Colonies with ⬎30 cells were scored after 5 weeks. Assays were done in
triplicate and two to three times independently.
Isolation of internal membrane fractions. Internal membrane fractions were
prepared as described previously (8). Brieﬂy, cells were resuspended in hypo-
tonic lysis buffer (10 mM KCl, 1.5 mM MgCl
, 1 mM EDTA, 1 mM EGTA, 250
mM sucrose, 1 mM dithiothreitol [DTT]) and homogenized with a Dounce
homogenizer. Extract was centrifuged at 750 ⫻ g to pellet nuclei, and the
supernatant was further centrifuged at 10,000 ⫻ g to pellet the crude membrane
fraction. The crude membrane fraction was resuspended in MIB (250 mM
sucrose, 40 mM KCl, 40 mM succinate, 40 mM HEPES, 20 mM EGTA, 800 mM
mannitol) and layered over a discontinuous Percoll gradient composed of 42%,
37%, 30%, and 25% Percoll prepared in MIB. Percoll gradients were centrifuged
at 25,000 rpm for 25 min in a Beckman TLS-55 rotor with the brake off. The
endoplasmic reticulum (ER)-rich internal membrane fraction sedimented be-
tween the 25% and 30% bands.
Xenograft tumorigenesis assay. Cells (10
) mixed with Matrigel were injected
subcutaneously into both ﬂanks of SCID/beige mice (Charles River Laboratory),
with four injection sites per cell line, and tumor volumes were determined twice
per week and calculated as (length/2)
⫻ width. These experiments were ap
proved by the Duke University Institutional Animal Care and Use Committee.
Immunoﬂuorescence. The described cells were plated on glass microslides the
previous day, ﬁxed, permeabilized, and incubated with anti-RalA or anti-RalB
(BD Transduction Laboratories), anti-Myc 9E10 (Invitrogen), and/or EEA1 (BD
Transduction Laboratories) primary antibodies, followed by incubation with the
anti-mouse Alexa-488 secondary antibody (Molecular Probes) or Texas Red-
phalloidin to detect endogenous RalA, RalB, actin, or ectopic Myc-tagged pro-
teins as previously described (36). For visualization of transiently expressed GFP
fusion proteins, HEK-TtH cells on glass microslides transfected with GFP-RalA
fusion proteins with or without HA-tagged Aurora-A (WT, K162R, or T288D)
and Myc-tagged RalBP1 or Sec5 were ﬁxed, mounted, and visualized by confocal
microscopy. A Zeiss LSM 410 confocal microscope, an Olympus Fluoview 300
laser scanning confocal imaging system conﬁgured with an IX70 ﬂuorescence
microscope ﬁtted with a PlanApo 60⫻ oil objective, or a Leica DMI6000CS
scanning confocal imaging system ﬁtted with a Leica Plan Apochromat 100⫻/1.4-
to 0.70-numerical-aperture oil objective was used for imaging.
RESULTS AND DISCUSSION
cooperates with the RalGEF pathway to pro
mote Ras-mediated transformation. Both oncogenic Ras mu-
tations (7) and overexpression of Aurora-A occur in human
cancers (22, 31), such as pancreatic cancer (27, 28, 35). In
experimental systems, overexpression of Aurora-A enhanced
Ras- and RalA-induced transformation of a murine cell line
(59) and a canine cell line (66), respectively. Given this func-
tional association between Ras oncogenic signaling and
Aurora-A activation, we tested whether Aurora-A also en-
hances Ras-mediated transformation of human cells. For these
experiments, we utilized HEK-TtH cells (normal HEK cells
immortalized and rendered sensitive to Ras transformation by
ectopic expression of the SV40 early region encoding large T
and small t antigens and of the telomerase catalytic subunit
hTERT) (24, 33). As these cells depend upon oncogenic Ras
for tumorigenesis, and the other genetic changes required for
tumorigenesis are known, Ras-transformed HEK-TtH cells
provide a simpliﬁed, genetically deﬁned, and malleable system
for dissecting the relationship between Aurora-A and onco-
genic Ras in transformation and tumorigenesis of human cells.
HEK-TtH cells were stably infected with a retrovirus either
encoding a constitutively activated T288D mutant (5) of hu-
VOL. 30, 2010 AURORA-A PHOSPHORYLATES AND RELOCALIZES RalA 509
man Aurora-A (Aurora-A
) or carrying no transgene, in
the absence or presence of oncogenic (G12V) H-Ras
). Cell lines were veriﬁed to express the appropriate
transgenes, as assessed by immunoblot analysis (Fig. 1A), and
assayed for growth in soft agar as a measure of transformation.
We found that activated Aurora-A
alone could not pro
mote transformation of these cells in the absence of oncogenic
Ras (Fig. 1B) but that Aurora-A
Ras-mediated transformation twofold (Fig. 1B). These results
indicated that the cooperative activity of Aurora-A and Ras
signaling to promote growth transformation observed previ-
ously in murine ﬁbroblast (59) and canine epithelial (66) cell
lines can be extended to human cells.
To explore at what point Aurora-A may converge upon
oncogenic Ras signaling, we tested whether Aurora-A cooper-
ated with a speciﬁc Ras effector pathway to promote transfor-
mation. We focused our analyses on the three major Ras ef-
fectors involved in oncogenesis (Raf, PI3K, and RalGEF) by
using oncogenic H-Ras effector domain mutants Ras
, and Ras
, which retain preferential ac
tivation of the Raf, RalGEF, and PI3K pathways, respectively
(34, 54, 62, 64). Kinase-active Aurora-A
or an empty
vector as a negative control was therefore expressed in HEK-
TtH cells in conjunction with no transgene or a transgene
encoding each Ras effector domain mutant. Appropriate ex-
pression of Aurora-A
and the Ras
was veriﬁed by immunoblot analysis (Fig. 1A), and the result-
ant six cell lines in addition to vector controls were assayed for
anchorage-independent growth. As previously reported (25),
activation of the RalGEF pathway, but not the PI3K or mito-
gen-activated protein kinase (MAPK) pathway, promoted the
growth of these cells in soft agar. Addition of kinase-active
did not endow anchorage-independent growth
to either vector control cells or cells expressing Ras mutants
FIG. 1. Aurora-A
promotes Ras-induced transformation through the RalGEF pathway. (A and D) Appropriate expression, as detected
by immunoblot analysis, of kinase-active HA-Aurora-A
, kinase-inactive HA-Aurora-A
, constitutively activated HA-Rlf-CAAX, and
ectopic and endogenous Ras (Pan-Ras). Actin serves as a loading control. (B, C, and E) Anchorage-independent growth in soft agar of polyclonal
HEK-TtH cells stably expressing the indicated transgenes, expressed as average numbers of colonies formed ⫾ standard errors of the means (SEM)
for six plates (two independent experiments conducted in triplicate). The same vector (with [⫹] or without [⫺] HA-Aurora-A
) was used as
a control in all of these experiments. Signiﬁcant P values (⬍0.001) are indicated by ⴱⴱ. Tukey’s multiple-comparison test was used to determine
signiﬁcance between cell lines.
510 LIM ET AL. M
activating the Raf or PI3K pathways but did enhance twofold
the transformed growth of cells expressing Ras
1C). Thus, Aurora-A cooperates synergistically with the Ral-
GEF signaling arm of oncogenic Ras in cellular transforma-
To further validate these results, we tested whether Aurora-A
similarly enhanced transformation of cells expressing an acti-
vated variant of the RalGEF protein Rlf/Rgl2 (Rlf-CAAX)
(63) in place of Ras
. Speciﬁcally, HEK-TtH cells
were engineered to coexpress both Rlf-CAAX and Aurora-
proteins, or, as a negative or positive control, an empty
vector or Rlf-CAAX, respectively. Appropriate transgene ex-
pression was veriﬁed by immunoblot analysis (Fig. 1D). As
expected, negative-control vector cells did not grow in soft
agar, whereas the addition of Rlf-CAAX promoted trans-
formed cell growth (Fig. 1E). In agreement with the ability of
to enhance transformation by
, cells expressing both the activated Rlf-CAAX
proteins grew more robustly in soft agar
(Fig. 1E). We next tested whether the kinase activity of
Aurora-A was required to enhance RalGEF-mediated trans-
formation by expressing a kinase-inactive mutant of Aurora-
(5) in Rlf-CAAX-transformed cells (Fig. 1D). Whereas
transformation, the kinase-inactive Aurora-A
hibited growth in soft agar (Fig. 1E). Thus, Aurora-A enhances
Ras-RalGEF signaling to promote cellular transformation.
-induced increase in RalGEF-induced trans
formation of human cells is lost upon mutation of the Aurora-A
phosphorylation site S194 of RalA. RalGEF proteins activate
the RalA and RalB isoforms of Ral GTPases, which are related
both structurally (82% sequence identity) and biochemically.
Despite these similarities, however, the proteins have distinct
roles in oncogenesis. Whereas RalA is required for Ras-medi-
ated anchorage-independent growth (36, 38), RalB can instead
be critical to cell viability, at least in some contexts (14, 15),
particularly when cells are grown in the absence of substrate or
during metastasis (38). The primary sequence differences be-
tween RalA and RalB are concentrated in their C-terminal-
30-residue membrane-targeting sequences in the hypervariable
domain (20), which have been shown to contribute to some of
the functional differences described above (36). Upstream of
the C-terminal CAAX tetrapeptide motif that signals for post-
translational modiﬁcation by a geranylgeranyl isoprenoid re-
quired for membrane association, these divergent sequences
target RalA to the plasma membrane and late endosomes but
target RalB to the plasma membrane only (57). Aurora-A
phosphorylates this C-terminal region at a serine residue,
S194, found in RalA but not in RalB (66), and mutating this
site inhibited the ability of an overexpressed and constitutively
activated RalA mutant (G23V) to cooperate with ectopically
expressed Aurora-A to promote the transformed growth of a
canine cell line (66). Conversely, knockdown of PP2A A␤ in
HEK cells expressing T-Ag, hTERT, and oncogenic Ras re-
sulted in elevated phosphorylation of S183 and S194 of RalA,
and when RalA was also knocked down in these cells, they
were able to form tumors upon restoration with ectopic wild-
type RalA but not with an S194A mutant version (56). Con-
sistent with these observations, we found that endogenous
RalA is phosphorylated at serines in both human 293T and
HEK TtH cells (Fig. 2A), likely owing to expression of t-Ag in
these cells (56), and that ectopically expressed wild-type Au-
rora-A increased the level of this phosphorylation (Fig. 2A).
Similarly, we found that the Aurora-A
hancement of RalGEF-mediated transformation of HEK-TtH
cells requires phosphorylation of RalA at S194. First, in HEK-
TtH cells stably expressing both activated Rlf-CAAX and Au-
proteins, expression of endogenous RalA protein
was stably knocked down by shRNA, as assessed by immuno-
blot analysis (Fig. 2B). Next, the loss of RalA expression was
then complemented by a vector either carrying no transgene
(negative control) or encoding an shRNA-resistant RalA pro-
tein in the wild-type (positive control) or S194A mutant (66)
conﬁguration, as assessed by immunoblot analysis (Fig. 2B). As
FIG. 2. Aurora-A potentiates RalGEF-transformation through
phosphorylation of RalA S194. (A) Appropriate expression, as de-
tected by immunoblot analysis, of wild-type HA-Aurora-A, endoge-
nous RalA, and phosphorylated RalA (detected by immunoprecipita-
tion [IP] of endogenous RalA followed by immunoblot analysis [IB]
with an anti-phospho-serine [␣ p-serine] antibody) in 293T and HEK-
TtH cells expressing the indicated transgenes. Total RalA serves as a
loading control. (B) Appropriate expression, as detected by immuno-
blot analysis, of HA-Aurora-A
, HA-Rlf-CAAX, knockdown of
endogenous RalA, and complementation by ectopic RalA resistant to
RalA shRNA in HEK-TtH cells stably infected with retroviruses car-
rying no transgene (vector [V]) or the indicated transgenes in the
presence of shRNA speciﬁc to RalA or a scrambled version (scram) of
this sequence. Actin serves as a loading control. (C) Anchorage-inde-
pendent growth in soft agar of the aforementioned polyclonal HEK-
TtH cells infected with retroviruses encoding the indicated shRNAs
and carrying transgenes, expressed as average numbers of colonies
formed ⫾ SEM for six plates (two independent experiments conducted
in triplicate). Signiﬁcant P values (⬍0.001) are indicated by ⴱⴱ. Tukey’s
multiple-comparison test was used to determine signiﬁcance between
OL. 30, 2010 AURORA-A PHOSPHORYLATES AND RELOCALIZES RalA 511
expected (36), knockdown of RalA was found to reduce trans-
formed cell growth in soft agar to the level of cells not trans-
fected with activated RalGEF and Aurora-A. This loss of
transformation was rescued, almost to the level of the positive
control scramble control cells, by expression of shRNA-resis-
tant wild-type RalA but not the S194A mutant of RalA (Fig.
2C). Lastly, the Aurora-A
-enhanced transformation of
Rlf-CAAX-expressing cells was not reduced upon knockdown
of RalB (see Fig. S1 in the supplemental material), which lacks
the Aurora-A phosphorylation site (66). Thus, Aurora-A
enhancement of RalGEF-mediated transformation is lost spe-
ciﬁcally upon mutation of the Aurora-A substrate S194 of
RalA S194 is required for transformed growth of human pancre-
atic cancer cell lines. RalA is frequently activated in pancreatic
cancers and is essential for transformed growth of pancreatic
cancer cell lines in vitro and in vivo (38). Aurora-A protein is
similarly overexpressed in this disease (35). Aurora-A and
RalA mRNA overexpression is also associated with advanced
human bladder cancer (58). Indeed, we detected elevated lev-
els of this kinase in a panel of 10 KRAS mutation-positive
pancreatic carcinoma cells, compared to the level for normal
pancreatic cancer tissues, whereas the total protein levels of
RalA did not differ between tumor cells and normal tissue (Fig.
3A). Thus, while RalA is upregulated at the level of protein
activation (36, 38), Aurora-A is upregulated at the level of
protein expression in pancreatic cancer cells. Given this cor-
relation and the requirement of S194 of RalA for transformed
or tumorigenic growth of canine (66), human HEK-TER (56),
and HEK-TtH (Fig. 2C) cells, we tested whether S194 is also
required for tumorigenic growth of human pancreatic cells.
Speciﬁcally, a panel of pancreatic cancer cell lines, including
AsPC-1, HPAC, HPAF-II, Panc-1, SW1990, Capan-1,
CFPac-1, MIA PaCa-2, and T3M4 cells, were stably infected with
a retrovirus encoding RalA shRNA or a scramble control, after
which the RalA knocked-down cell lines were complemented by
infection with retroviruses encoding shRNA-resistant forms of
either wild-type RalA or the RalA
analysis veriﬁed appropriate knockdown of endogenous RalA
and subsequent reexpression of the shRNA-resistant forms of
RalA at levels roughly similar to that of endogenous RalA in
all nine cell lines (Fig. 3B). All 36 cell lines were then assayed
for anchorage-independent growth.
Consistent with the known role of RalA in transformation
(36, 38), knockdown of RalA expression signiﬁcantly impaired
the anchorage-independent growth of all nine parental cell
lines, ranging from 55% to 90% decrease in colony numbers.
This decrease in transformed growth was, in large part, over-
come by ectopic expression of the shRNA-resistant wild-type
RalA protein. However, expression of a protein that was ex-
actly the same except that it harbored the single point mutation
at S194A that renders RalA resistant to Aurora-A phosphory-
lation extinguished the ability of RalA to rescue the loss of
endogenous RalA and restore transformed cell growth in each
of the nine cell lines (Fig. 3C). Based on these data, S194
appears to be broadly required for RalA promotion of anchor-
age-independent proliferation of human pancreatic cancer
RalA S194 is required for tumorigenic growth of some hu-
man pancreatic cancer cell lines. Since RalA is critical for
tumorigenic growth of pancreatic cancer cells in vivo (38), we
next determined whether S194 is also required for tumorigen-
esis of such cells. The CFPac-1, HPAC, and Capan-1 cell lines,
which exhibit differing levels of RalA-GTP (Fig. 4A), were
engineered to stably express scramble control shRNA, RalA
shRNA, or a RalA protein resistant to RalA shRNA in either
the wild-type or the S194A mutant conﬁguration. These lines
were injected subcutaneously into both ﬂanks of immunocom-
promised mice. Compared to what was found for scramble
control cells, loss of RalA prolonged the latency period of
tumorigenesis at least twofold and impeded subsequent tumor
growth in all three pancreatic cancer lines (Fig. 4B and C). In
Capan-1 and HPAC cells, which have prominent RalA-GTP
(Fig. 4A), RalA
was defective in restoring tumor growth
to the same level as wild-type RalA (Fig. 4B and C), whereas
in CFPac-1 cells, with weak RalA-GTP, wild-type and S194A
RalA proteins were equally effective. The ﬁnding that
did not rescue soft agar growth (Fig. 3C) but did
rescue tumor growth of CFPac1 cells treated with RalA
shRNA (Fig. 4B and C) suggests that RalA, but not Aurora-A,
may be required for tumor growth. Whether there is an inverse
correlation between the level of RalA-GTP (Fig. 4A) and the
degree of tumor growth restoration by RalA
type RalA (Fig. 4B and C) or this is simply a coincidence
remains to be determined. If the former is the case, this oc-
currence may reﬂect the possibility that cells with higher basal
levels of RalA-GTP are more dependent for tumorigenicity on
RalA function and are therefore more sensitive to perturba-
tions in RalA modulation. If the latter is the case, the variabil-
ity observed may simply mean that not all oncogenic Ras-
driven cancer cells equally require Aurora-A phosphorylation
of RalA to promote tumorigenesis.
promotes the translocation of endogenous
RalA from the plasma membrane. The C-terminal hypervari-
able domain of RalA, but not RalB, contains the S194 Auro-
ra-A phosphorylation site, and the hypervariable regions of
both proteins are known to account for the partially overlap-
ping but distinct subcellular membrane localization patterns of
the two proteins (36, 57). Further, phosphorylation by protein
kinase C (PKC) of a similarly situated C-terminal serine in
K-Ras4B (S181) can induce translocation of K-Ras4B from the
plasma membrane to internal organelles (6). Given these ob-
servations, we speculated that in addition to its role in activat-
ing RalA (66), Aurora-A may alter the subcellular location of
To address this possibility, we monitored the subcellular
localization of endogenous RalA in both gain (ectopic Aurora-
A)- and loss (Aurora-A shRNA)-of-function situations. In the
gain-of-function approach, the subcellular distribution of en-
dogenous RalA was ﬁrst assessed by immunoﬂuorescence in
HEK-TtH cells. In cells that stably expressed either empty
vector or kinase-inactive Aurora-A
, endogenous RalA
was found abundantly at the plasma membrane as well as
throughout the cytoplasm and at internal membranes. How-
ever, in cells expressing kinase-active Aurora-A
was a clear loss of RalA at the plasma membrane and a con-
comitant increase in the internal pool of protein (Fig. 5A), and
some RalA colocalized with Rab11 and EEA1, markers of
recycling endosomes and early endosomes, respectively (not
shown). This effect of Aurora-A kinase activity on accumula-
512 LIM ET AL. MOL.CELL.BIOL.
tion of RalA on internal membranes was borne out by bio-
chemical fractionation. Speciﬁcally, in the presence of kinase-
, RalA was greatly enriched in the
isolated internal membrane fraction (deﬁned by high levels of
the ER-resident protein calnexin), compared to the amount
seen in that fraction in the presence of kinase-inactive Aurora-
(Fig. 5B). Furthermore, ectopic Aurora-A, whether in
the kinase-active or -inactive mutant form, did not alter the
levels of endogenous RalA (Fig. 5B).
In the loss-of-function analysis, we ﬁrst tested whether re-
ducing the level of endogenous Aurora-A affected the local-
ization of endogenous RalA in the human pancreatic cancer
cell line HPAC (46), chosen as an example of cells that harbor
a mutated KRAS allele (38), exhibit elevated levels of Aurora-A
protein (Fig. 3A), and are sensitive to Aurora-A phosphoryla-
tion of RalA
for transformation (Fig. 3C) and tumorige
nicity (Fig. 4B and C). Because constitutive knockdown of
endogenous Aurora-A expression leads to rapid G
and subsequently to apoptosis (26), HPAC cells were engi-
neered to express a doxycycline-sensitive TET repressor and
Aurora-A shRNA (or the scramble sequence) driven by a
TET-ON operator (37), such that Aurora-A protein levels
FIG. 3. RalA S194 is broadly required for transformed growth of pancreatic cancer cell lines. (A) Detection of Aurora-A and RalA by
immunoblot analysis of the indicated pancreatic cancer cell lines, compared with the results for three normal (N) pancreatic tissue specimens.
(B) Immunoblot analysis of RalA in the indicated nine pancreatic cell lines, each retrovirally infected with a scramble sequence (S) or shRNA
against RalA (RalAi) complemented by an empty vector (V), shRNA-resistant RalA in the wild-type (WT) or S194A mutant (SA) conﬁguration.
Actin serves as a loading control. (C) Photographs illustrating anchorage-independent growth in soft agar of the indicated polyclonal pancreatic
cancer cells stably expressing either a RalA scramble sequence (scram) or RalA shRNA complemented with an empty vector (vector), the
shRNA-resistant wild-type RalA protein (RalA WT), or the S194A mutant RalA protein (RalA
). Shown are average numbers of colonies
formed ⫾ SEM as calculated from triplicate plates. Data are from one representative experiment of two independent assays. Signiﬁcant P values
(⬍0.001) are indicated by ⴱⴱ. Tukey’s multiple-comparison test was used to determine signiﬁcance between cell lines.
OL. 30, 2010 AURORA-A PHOSPHORYLATES AND RELOCALIZES RalA 513
could be inducibly reduced in the presence of doxycycline (Fig.
5C). As expected, in scramble control cells with or without doxy-
cycline in the culture medium, endogenous RalA was detected by
immunoﬂuorescence throughout the cytoplasm, at intracellular
organelles, and at the plasma membrane along cell-cell borders.
On the other hand, cells induced to express Aurora-A shRNA
exhibited a noticeable increase in the concentration of RalA pro-
tein at the plasma membrane and a concomitant loss at internal
membranes (Fig. 5D), which was even more obvious upon bio-
chemical fractionation (Fig. 5E). Thus, by both gain- and loss-of-
function analyses, we demonstrate that, akin to PKC promoting
translocation of K-Ras4B from the plasma membrane (6), Auro-
ra-A kinase activity promotes dissociation of endogenous RalA
from the plasma membrane.
Phosphorylation of S194 by Aurora-A
is required for
redistribution of RalA. To address whether S194 of RalA is
required for redistribution of RalA in the presence of
Aurora-A, we expressed GFP-tagged RalA
in which the
FIG. 4. RalA S194 is required for tumorigenesis of pancreatic cancer cell lines. (A) RalA-GTP levels as detected in the indicated pancreatic
cancer cell lines. HEK-TtH cells expressing empty vector or Ras
serve as a negative or positive control, respectively. Total RalA and tubulin
serve as loading controls. (B) Shown are tumor volumes (mm
) ⫾ standard deviations versus times (days) observed for the indicated cell lines stably
expressing a Ral scramble sequence (䡺), RalA-shRNA (f), or RalA-shRNA complemented by expression of shRNA-resistant wild-type RalA (Œ)
(F) injected into the ﬂanks of immunocompromised mice. (C) Representative subcutaneous ﬂank tumors observed in mice (top) and
resected (bottom) from RalA shRNA-treated CFPac-1, HPAC, or Capan-1 cells transduced with shRNA-resistant wild-type RalA or the
mutant at 42 days (CFPac-1) or 49 days (HPAC and Capan-1) after the cells were injected.
514 LIM ET AL. M
Aurora-A phosphorylation site was mutated in HEK-TtH cells
ectopically expressing either kinase-active Aurora-A
no transgene (vector). Cells expressing kinase-active Aurora-
were smaller, possibly because constitutive activation of
Aurora-A may promote more-rapid cell division (Fig. 6A). We
found by immunoﬂuorescence analysis that RalA
tained in the plasma membrane, whether Aurora-A was acti-
vated or not. In contrast, GFP-RalA
, a putative phospho
mimetic version of RalA, was observed primarily at internal
membranes, consistent with the localization of GFP-RalA in
the presence of kinase-active Aurora-A
(Fig. 6A). These
observations were also supported by biochemical fractionation.
In the presence of kinase-active Aurora-A
, a higher pro
portion of the total wild-type RalA was found in the internal
membrane fraction than in the S194A mutant, after controlling
for different levels of exogenous RalA expression (Fig. 6B). In
support of a speciﬁc effect of Aurora-A on RalA, expression of
did not alter the subcellular lo
calization of GFP-RalB (see Fig. S3 in the supplemental ma-
terial), which is highly similar to RalA but lacks the Aurora-A
phosphorylation site (66), or of K-Ras4B (see Fig. S2 in the
supplemental material), in which phosphorylation of a similar
C-terminal serine residue by PKC displaces GFP-K-Ras from
the plasma membrane (6). Lastly, we demonstrate that the last
20 amino acids of RalA are sufﬁcient for Aurora-A-mediated
relocalization of RalA. Speciﬁcally, a GFP fusion protein with
the most-C-terminal 20 amino acids of RalA, comprising the
hypervariable membrane targeting domain and containing
the Aurora-A phosphorylation site, was relocalized from the
plasma membrane upon expression of kinase-active Aurora-
. Moreover, this relocalization was blocked when S194
was mutated to alanine (Fig. 6C). Thus, relocalization of RalA
in the presence of Aurora-A
depends upon S194 of RalA.
The RalA effector RalBP1 is translocated from the plasma
membrane and activated in the presence of Aurora-A
explore the relationship of RalA with its effectors in the con-
text of Aurora-A-mediated phosphorylation, we determined
whether Aurora-A kinase activity, which alters RalA subcellu-
lar localization, also alters that of a key RalA effector, RalBP1/
RLIP76 (10, 29, 50). We compared the localization of RalBP1
to that of ectopic RalA in HEK-TtH cells expressing kinase-
versus kinase-active Aurora-A
In the presence of kinase-inactive Aurora-A
RalBP1 colocalized with RalA in the cytoplasm, internal struc-
FIG. 5. Aurora-A regulates RalA subcellular localization. (A) Immunoﬂuorescence demonstrating distribution of endogenous RalA by use of
an anti-RalA antibody in HEK-TtH cells stably expressing an empty vector, kinase-inactive HA (epitope-tagged)-Aurora-A
, or kinase-active
. Arrows, plasma membrane localization. Scale bar, 20 m. (B) Immunoblot analysis of endogenous RalA, calnexin, and tubulin
in a whole-cell extract and an internal membrane fraction isolated from HEK-TtH cells expressing HA-Aurora-A
(C) Reduction in endogenous Aurora-A protein, as detected by immunoblot analysis, in HPAC cells expressing doxycycline (dox)-inducible
Aurora-A shRNA treated with dox for 16 h, compared to the level for untreated cells or dox-treated cells expressing a dox-inducible scramble
control shRNA. Tubulin serves as a loading control. (D) Relocalization of endogenous RalA from the cytoplasm to the plasma membrane, as
detected by immunoﬂuorescence, in HPAC cells expressing dox-inducible Aurora-A shRNA treated with dox for 16 h, compared to the level for
untreated cells or dox-treated cells expressing a dox-inducible scramble control shRNA. Arrows, plasma membrane localization. Scale bar, 20 m.
(E) Immunoblot analysis of endogenous RalA, calnexin, and tubulin in a whole-cell extract and an internal membrane fraction isolated from HPAC
cells expressing dox-inducible Aurora-A shRNA treated with dox for 16 h, compared to the level for untreated cells.
OL. 30, 2010 AURORA-A PHOSPHORYLATES AND RELOCALIZES RalA 515
tures, and protrusions of the plasma membrane. In the pres-
ence of kinase-active Aurora-A
, the plasma membrane
pools of both RalA and RalBP1 repartitioned internally (Fig.
7A). These data indicate that Aurora-A coordinately regulates
RalA and RalBP1. Expression of the effector domain mutant
, which perturbs association of RalBP1 with RalA,
did not result in changes in RalBP1 localization in the presence
of kinase-inactive Aurora-A
versus kinase-active Aurora-
(see Fig. S4 in the supplemental material), indicating
that Aurora-A-mediated internalization of RalBP1 is depen-
dent on the association between RalA and RalBP1. Further-
more, ectopic Aurora-A expression, either in the kinase-active
or in the inactive mutant form, did not alter the levels of
endogenous RalBP1 (see Fig. S5 in the supplemental mate-
rial). To determine biochemically if this coordinate relocaliza-
tion enhanced the association of RalBP1 with RalA, vectors
encoding RalA (the wild type or the putative phosphomimetic
S194D mutant) or RalBP1 were cotransfected into 293T cells,
and the amount of RalBP1 coimmunoprecipitating with RalA
was assessed by immunoblot analysis. In the absence of serum,
when the majority of RalA is in the inactive GDP-bound state,
RalBP1 did not readily coimmunoprecipitate with wild-type
RalA (Fig. 7B) but did weakly associate with RalA
dition of serum to activate RalA promoted its association with
FIG. 6. Aurora-A
-mediated internalization of RalA depends upon S194. (A) Distribution of GFP-RalA constructs in HEK-TtH cells
stably expressing vector or HA-Aurora-A
in combination with GFP-RalA or GFP-RalA
, compared to the level for HEK-TtH cells
expressing the phosphomimetic mutant RalA protein GFP-RalA
with vector alone. GFP-RalA localization to both the plasma membrane and
the internal membrane (PM⫹IM), the plasma membrane only (PM), or the internal membrane only (IM) in 50 cells was quantitated in two
independent experiments for each condition. Representative images with the primary location are displayed. Scale bar, 20 m. (B) Immunoblot
analysis of endogenous RalA, calnexin, and tubulin in a whole-cell extract and an internal membrane fraction isolated from HEK-TtH cells
with either wild-type or S194A Flag-RalA. Eightfold more RalA
than wild-type RalA was expressed in
whole-cell extract. Therefore, wild-type Flag-RalA levels in the whole-cell extract and the internal membrane fraction were normalized 8:1 to
S194A Flag-RalA levels. (C) Distribution of the last 20 amino acids of the RalA C terminus fused to GFP (GFP-RalA-C term) or, as indicated,
in HEK-TtH cells stably expressing either an empty vector, HA-Aurora-A
, or HA-Aurora-A
localization to both the plasma membrane and the internal membrane (PM⫹IM), the plasma membrane only (PM), or the internal membrane only
(IM) in 50 cells was quantitated in two independent experiments for each condition. Representative images with the primary location are displayed.
Scale bar, 20 m.
516 LIM ET AL. M
RalBP1, and this interaction was enhanced nearly ﬁvefold in
the case of RalA
, compared to the level for wild-type
RalA (Fig. 7B). However, Aurora-A activity did not affect all
Ral effector interactions equally. Another validated Ral effec-
tor, Sec5, neither redistributed in the presence of kinase-active
nor preferentially bound RalA
S6 in the supplemental material). Compared to what was found
for wild-type RalA, the association between RalA
Sec5 was reduced by about one-third; however, the nature of
this difference is unknown (see Fig. S6 in the supplemental
material). Whether the selective inﬂuence of Aurora-A on
effector association is due to the context of RalA activation,
subcellular localization, or indirect effects, such as competition
with RalB for the same effector, remains to be determined.
Given that compared to RalA, the RalA
if anything, more strongly to RalBP1 than to Sec5, we explored
a possible relationship between RalBP1 and Aurora-A. Be-
cause RalBP1 has been shown to be a GTPase-activating pro-
tein (GAP) for two Rho family GTPases, Rac1 and Cdc42 (10,
29, 50), we investigated the signaling consequences of the ob-
served increased association between RalA and RalBP1. Spe-
ciﬁcally, we examined the levels of endogenous activated GTP-
FIG. 7. Aurora-A promotes cytoplasmic translocation and activation of RalBP1. (A) Distribution of GFP-RalA and Myc-tagged RalBP1
(MycRalBP1), visualized by immunoﬂuorescence using an anti-Myc antibody in HEK-TtH cells stably expressing either a vector control,
, or kinase-active HA-Aurora-A
. Arrows, plasma membrane. Scale bar, 20 m. (B) Aurora-A fosters the
association of RalA with RalBP1. Immunoprecipitation (IP) of the Flag-tagged WT or the phosphomimetic S194D (SD) mutant version of RalA,
followed by immunoblot analysis (IB) for detection of Flag-tagged, immunoprecipitated RalA protein or the presence or absence of coimmuno-
precipitated Myc-tagged RalBP1 in the presence or absence of serum for activation of the RalA protein, as assessed by the level of Flag RalA-GTP.
Total MycRalBP1 and Flag RalA serve as loading controls. (C) Aurora-A decreases Cdc42 and Rac1 activation. Shown are GTP-Cdc42 and
GTP-Rac1 levels detected in HEK-TtH cells in which endogenous RalA was either not activated (vector) or activated by expressing HA-Rlf-CAAX
in the presence, as assessed by immunoblot analysis, of either kinase-active HA-Aurora-A
or kinase-inactive HA-Aurora-A
. Total Cdc42
and Rac1 serve as loading controls. Cdc42-GTP and Rac1-GTP levels are normalized to levels for total Cdc42 and Rac1, expressed as fold
changes ⫾ standard deviations for three independent experiments (P ⫽ 0.029). (D) Knockdown of RalBP1 activates Cdc42 and Rac1. Shown are
GTP-Cdc42 and GTP-Rac1 levels in HEK-TtH cells stably expressing shRNA against either the vector control or RalBP1. Total Cdc42 and Rac1
serve as loading controls. Cdc42-GTP and Rac1-GTP levels are normalized to the levels for total Cdc42 and Rac1, expressed as fold changes ⫾
standard deviations for three independent experiments. (E) Distribution of GFP-RalA or GFP-RalA
and actin organization, visualized by
immunoﬂuorescence using Texas Red-phalloidin in HEK-TtH cells stably expressing either a vector control or kinase-active HA-Aurora-A
Formation of ﬁlopodia and lamellipodia in 50 cells was quantitated in two independent experiments for each condition. Representative images are
displayed. Arrows, ﬁlopodia or lamellipodia. Scale bar, 20 m. (F) Appropriate expression, as detected by immunoblot analysis, of HA-Aurora-
(TD), HA-Rlf-CAAX, or a knockdown of endogenous RalBP1 in HEK-TtH cells stably infected with retroviruses carrying no transgene
(v) or the indicated transgenes in the presence of shRNA speciﬁc to RalBP1 or a scrambled version (scram) of this sequence. Tubulin serves as
a loading control. (G) Anchorage-independent growth in soft agar of the aforementioned polyclonal HEK-TtH cells, infected with retroviruses
encoding the indicated shRNAs and carrying the indicated transgenes, expressed as average numbers of colonies formed ⫾ SEM for six plates (two
independent experiments conducted in triplicate). Signiﬁcant P values (⬍0.001) are indicated by ⴱⴱ. Tukey’s multiple-comparison test was used
to determine signiﬁcance between cell lines.
OL. 30, 2010 AURORA-A PHOSPHORYLATES AND RELOCALIZES RalA 517
bound Cdc42 and Rac1 as readouts of RalBP1 GAP activity
and assessed whether the expression of kinase-inactive Aurora-
or kinase-active Aurora-A
altered RalBP1 GAP
activity. To do this, we ﬁrst activated endogenous RalA in
HEK-TtH cells by expression of Rlf-CAAX and measured the
amounts of activated GTP-bound Cdc42 and Rac1 in the pres-
ence of either kinase-active Aurora-A
. In cells expressing kinase-active Aurora-
, activation of RalA signaling by Rlf-CAAX reduced
the level of GTP-bound Cdc42 by one-half to one-third com-
pared to the level for vector control cells (Fig. 7C). The nature
of the selectivity toward decreased Cdc42-GTP levels com-
pared to Rac1-GTP levels is unknown but, we speculate, could
be due to differences in subcellular localization. As RalBP1
GAP activity is not altered upon binding activated RalA in
vitro (50), recruitment of RalBP1 to speciﬁc subcellular sites
may instead underlie the reduction in GTP-bound Cdc42. Fur-
thermore, shRNA-mediated knockdown of RalBP1, as con-
ﬁrmed by immunoblot analysis, increased the levels of both
activated GTP-bound Rac1 and Cdc42 both in the presence
and in the absence of Aurora-A
(Fig. 7D; see also Fig. S7
in the supplemental material). Such a result is not surprising if
indeed RalBP1 acts as a GAP for Cdc42 and Rac1. These
results are consistent with a model in which Aurora-A-medi-
ated phosphorylation of RalA leads to an enhanced interaction
with RalBP1, perhaps to alter its subcellular localization, re-
sulting in a decrease in GTP-bound Cdc42.
We next explored the cellular consequences of the increased
association between RalA and RalBP1 and the decrease in
GTP-bound Cdc42 promoted by Aurora-A
. Cdc42 acti
vation causes formation of actin microspikes and ﬁlopodia,
whereas Rac activation promotes concentration of actin at the
leading edge of moving cells and the formation of lamellipodia.
Therefore, we assayed for changes in cellular morphology in
HEK-TtH cells transiently expressing GFP-RalA in the pres-
ence of either empty vector or kinase-active Aurora-A
We ﬁrst conﬁrmed that stable expression of kinase-inactive
or kinase-active Aurora-A
did not alter
actin cytoskeleton organization in the absence of exogenous
RalA (see Fig. S8 in the supplemental material). Consistent
with the biochemical decreases in Cdc42- and Rac1-GTP levels
in the presence of kinase-active Aurora-A
⬃76% of HEK-TtH cells expressing GFP-RalA and empty
vector displayed ﬁlopodia and lamellipodia, whereas ⬃30% of
cells expressing both GFP-RalA and kinase-active Aurora-
displayed ﬁlopodia and lamellipodia, indicative of de
creases in both Cdc42 and Rac1 activations (Fig. 7E). In agree-
ment, HEK-TtH cells transiently expressing a phosphomimetic
mutant of RalA, GFP-RalA
, displayed an absence of
ﬁlopodia and lamellipodia (Fig. 7E). Taken together, these
data support the notion that Aurora-A
creases the association between RalA and its effector RalBP1,
leading to enhanced RalBP1 function, as measured by de-
creased Cdc42- and Rac1-GTP levels and decreased Rho fam-
ily GTPase-driven morphology.
These studies demonstrate that Aurora-A has a biological
impact on RalA activation of RalBP1, but it remains to be
determined whether such activation of RalBP1, or even sup-
pression of Cdc42, fosters transformation. On one hand,
RalBP1 contributes positively to transformation. Speciﬁcally,
knockdown of RalBP1 in HEK-TtH cells transformed by Rlf-
CAAX in the absence or presence of Aurora-A
ﬁrmed by immunoblot analysis (Fig. 7F), exhibited reduced
anchorage-independent growth, compared to the level for the
scramble control counterparts (Fig. 7G). The basis of this de-
creased transformation is unknown, but given that RalBP1
knockdown cells can be cultured extensively with no overt
impact on passaging (not shown) and that mice homozygous
for a gene trap mutation of RALBP1 (RIP1 and RLIP76) are
viable (3), the decrease in anchorage-independent growth may
be related to a function RalBP1 plays in transformation, al-
though this remains to be formally tested. Similarly, preventing
RalA association with RalBP1 by introduction of the D49N
effector domain mutation within a constitutively active
background slightly reduced transformation (33).
On the other hand, RalA also contributes to transformation by
pathways aside from RalBP1. Speciﬁcally, knockdown of
RalBP1 (Fig. 7G) did not reduce transformation to the same
extent as knockdown of RalA (Fig. 2C). Similarly, a D49E
mutation, which inhibits binding of Sec5 and Exo84, reduced
the transforming activity of RalA
to a greater extent than
the D49N mutation (33). Since RalBP1 could potentially serve
as a GAP for other untested Rho GTPases (29) and has ad-
ditional activities (2, 30, 67), it also remains to be tested if
Cdc42 is the relevant target for RalBP1 in transformation.
Moreover, activation of Cdc42 has been reported to both pro-
mote and suppress transformation. While RNA interference
(RNAi) suppression of a Cdc42GEF or dominant-negative
Cdc42 enhanced soft agar growth of human colon carcinoma
cells (44), activation of Cdc42 promoted transformation of
rodent ﬁbroblasts (51). Thus, while activation of RalBP1 and
suppression of Cdc42 constitute one plausible mechanism for
promoting transformation, others are certainly possible.
Differential effects of Aurora-A versus Aurora-A
RalA functions. Aurora-A is upregulated, rather than muta-
tionally activated, in human cancers (22, 31); hence, we also
explored the effect on RalA function in cells expressing wild-
type versus constitutively active Aurora-A. It has been dem-
onstrated that expression of either version of Aurora-A re-
sulted in elevated levels of GTP-bound and thus active RalA
and that mutation of S194, but not S183, reduced this effect
(66). Similarly, knockdown of the PP2A phosphatase subunit
a␤ has been found to lead to elevated levels of both RalA S183
and S194 phosphorylation and GTP-bound RalA (56). These
data support the notion that phosphorylation of S194 activates
wild-type RalA. We thus measured the level of RalA-GTP in
human cells expressing Aurora-A versus Aurora-A
ciﬁcally, human 293 cells were transiently transfected with a
vector carrying no transgene (as a negative control), encoding
t-Ag (as a positive control, since t-Ag blocks the ability of
PP2A to dephosphorylate S183 and S194 of RalA, resulting in
elevated phosphorylation at these sites ), encoding Aurora-
, or encoding wild-type Aurora-A. Expression of these
transgenes and of endogenous RalA was conﬁrmed by immu-
noblot analysis, and the levels of activated GTP-bound endog-
enous RalA were assessed by pulldown analyses (Fig. 8A). As
previously reported (56), t-Ag promoted robust activation of
RalA, compared to the level for vector control cells. Similarly,
wild-type Aurora-A was highly effective in activating RalA.
Unexpectedly, expression of Aurora-A
did not cause a
518 LIM ET AL. MOL.CELL.BIOL.
signiﬁcant increase in RalA-GTP, compared to the level for
vector control cells (Fig. 8B). Since Aurora-A
Ral transforming activity (Fig. 2C; see also Fig. 9B, C, and D)
yet does not robustly activate RalA (Fig. 8B), Aurora-A may
foster Ral-mediated transformation by additional mechanisms
independent of RalA-GTP formation.
Given this result, we tested whether phosphorylation and
subcellular localization of RalA were differentially affected by
Aurora-A versus Aurora-A
. 293 cells were therefore sta
bly infected with retroviruses carrying no transgene, encoding
Aurora-A, or encoding Aurora-A
. RalA phosphorylation
levels were measured by immunoprecipitation of endogenous
RalA, followed by immunoblotting with a phospho-speciﬁc
serine antibody, and endogenous RalA-GTP and total RalA
levels were determined by pulldowns and immunoblot analyses
total cellular lysates versus lysates derived from fractionated
internal membranes. As expected (66), RalA serine phosphor-
ylation was elevated upon expression of Aurora-A, and this
was further increased in cells expressing constitutively active
. In contrast to their differential effects on RalA
phosphorylation and GTP loading, both versions of this kinase
resulted in higher levels of total RalA in the internal mem-
FIG. 8. Differential effects of Aurora-A versus Aurora-A
on RalA functions. (A) Appropriate expression, as detected by immunoblot
analysis, of t-Ag, wild-type HA-Aurora-A, kinase-active HA-Aurora-A
, endogenous RalA, and GTP-bound RalA levels (detected by GST
pulldown of endogenous RalA) in 293 cells expressing the indicated transgenes. Total RalA serves as a loading control. (B) Immunoblot analysis
of phosphorylated RalA (detected by immunoprecipitation of endogenous RalA followed by immunoblot analysis with an antiphosphoserine [␣
p-serine] antibody) and RalA GTP-levels (detected by GST pulldown of endogenous RalA) in a whole-cell extract and an internal membrane
fraction isolated from 293 cells expressing the indicated transgenes. (C, D) Distribution of GFP-RalA or GFP-RalA
in HEK-TtH cells stably
expressing vector, HA-Aurora-A
, or HA-Aurora-A
. Arrows, plasma membrane localization. GFP-RalA localization to both the plasma
membrane and the internal membrane (PM⫹IM), the plasma membrane only (PM), or the internal membrane only (IM) in 50 cells was
quantitated in two independent experiments for each condition. Representative images with the primary location are displayed. Arrows, plasma
membrane localization. Scale bar, 20 m. (E) Immunoblot analysis of RalA-GTP levels in HEK-TtH cells stably expressing HA-Rlf-CAAX and
wild-type RalA (WT), RalA
(SA), or RalA
(SD). Tubulin serves as a loading control.
OL. 30, 2010 AURORA-A PHOSPHORYLATES AND RELOCALIZES RalA 519
brane fraction, although a greater proportion of the internal
pool was phosphorylated in cells expressing Aurora-A
(Fig. 8B). This result was independently validated by immuno-
ﬂuorescence; GFP-RalA accumulation at internal membranes
increased in cells expressing either version of Aurora-A (Fig.
8C), and in both cases, mutating S194A in RalA reduced this
translocation (Fig. 8D). The reason for the increase in RalA
phosphorylation with no corresponding increase in GTP load-
ing in the presence of Aurora-A
is unclear. One possibil
ity is that perhaps increased phosphorylation of RalA by Au-
promotes translocation of phosphorylated RalA
away from the plasma membrane, where RalGEF proteins
associate with activated Ras to then reduce RalGEF-mediated
RalA activation. In support of this model, RalA
, a puta
tive phosphomimetic version of RalA, was enriched at internal
membranes, compared to the level for wild-type RalA (Fig.
6A), but was less active (lower GTP levels), when coexpressed
with an activated, plasma membrane-targeted RalGEF pro-
tein, than the wild-type or even the S194A mutant of RalA
(Fig. 8E). Nevertheless, it still remains to be resolved why
robust phosphorylation of RalA by Aurora-A
coincide with elevated RalA-GTP levels.
Wild-type versus activated-Aurora-A-mediated transforma-
tion. Wild-type Aurora-A and Aurora-A
RalA phosphorylation and promoted RalA translocation,
whereas wild-type but not constitutively active Aurora-A
potently increased RalA-GTP levels. Capitalizing on the ability
of these two forms of Aurora-A to differentially alter RalA
functions, we tested whether elevated RalA-GTP or endo-
membrane translocation was associated with Aurora-A-medi-
ated transformation in cells with activated Ral signaling. HEK-
TtH cells in which the Ras-RalGEF pathway was activated at
different levels by expression of oncogenic Ras (Ras
oncogenic Ras mutant that preferentially activates RalGEF
) or an activated RalGEF protein (Rlf-
CAAX), were stably infected with retroviruses either carrying
no transgene or encoding kinase-inactive Aurora-A
negative control), Aurora-A, or Aurora-A
was assessed by immunoblot analysis (Fig. 9A), and cells were
assayed for anchorage-independent growth. As noted previ-
ously (Fig. 1), Aurora-A
, but not negative-control kinase-
, cooperated to various degrees with all
three activators of Ral to promote transformation. Wild-type
Aurora-A enhanced this transformation only marginally better
grounds (⬃1.1-fold) (Fig. 9B and C) but approximately twofold
in the Rlf-CAAX background (Fig. 9D). Given that only Au-
rora-A robustly activates RalA-GTP, whereas both Aurora-A
promote translocation of RalA and growth
transformation, their ability to enhance transformation was
attributed more to endomembrane translocation than to for-
mation of RalA-GTP in these experiments. Conversely, how-
ever, the ability of constitutively activated RalA (RalA
promote anchorage-independent growth of HEK-TtH cells
was not altered upon introduction of either the S194A muta-
tion or the S194D mutation (see Fig. S9 in the supplemental
material). Thus, in the context of constitutively active RalA,
the loss of S194 phosphorylation has no effect on RalA trans-
Summary. Aurora-A normally functions as a mitotic kinase
to ensure proper chromosome separation and cytokinesis (42).
This kinase is frequently overexpressed in various human can-
cers and can become mislocalized to the cytoplasm, where it
FIG. 9. Wild-type versus kinase-active Aurora-A-mediated poten-
tiation of Ras-induced transformation through the RalGEF pathway.
(A) Appropriate expression, as detected by immunoblot analysis, of
empty vector (v), wild-type HA-Aurora-A (WT), kinase-active HA-
(TD), kinase-inactive HA-Aurora-A
stitutively activated HA-Rlf-CAAX, or ectopic and endogenous Ras
(Pan-Ras). Actin serves as a loading control. (B, C, D) Anchorage-
independent growth in soft agar of polyclonal HEK-TtH cells stably
expressing the indicated transgenes, expressed as average numbers of
colonies formed ⫾ SEM for six plates (two independent experiments
conducted in triplicate). The same vector (v) (with [⫹] or without [⫺]
HA-Aurora-A [WT], HA-Aurora-A
[TD], or HA-Aurora-A
[KR]) was used as a control in all of these experiments. Signiﬁcant P
values (⬍0.001) are indicated by ⴱⴱ. Tukey’s multiple-comparison test
was used to determine signiﬁcance between cell lines.
520 LIM ET AL. M
may phosphorylate inappropriate substrates, leading to both
genomic instability and altered signaling, promoting cancerous
development (22, 31, 32). We ﬁnd that activated Aurora-A
cooperates with the RalGEF-Ral effector signaling arm of on-
cogenic Ras to promote transformation in human model cells
(Fig. 1 and 2) and, further, that transformation (Fig. 3) and in
some cases also tumor growth (Fig. 4) of pancreatic cancer
cells characterized by oncogenic Ras mutations depend upon
S194 of RalA, the site phosphorylated by Aurora-A (66). In
contrast to RalA, the nearly identical RalB protein cannot
support anchorage-independent growth transformation or tu-
morigenicity of human cells (36, 38) and is neither phosphor-
ylated (66) nor required for Aurora-A to promote RalGEF
transformation of HEK-TtH cells (see Fig. S1 in the supple-
mental material). Thus, in addition to its effect on chromosome
stability, aberrant overexpression or cell cycle-independent ex-
pression of Aurora-A in cancer may also foster transformation
and tumorigenesis through phosphorylation of RalA, a key
substrate of the oncogenic Ras-RalGEF effector pathway (66).
Indeed, the ﬁnding (9) of Aurora-A overexpression in a cell
cycle-independent fashion and its localization throughout tu-
mor cells, as opposed to its restriction to the nuclei of normal
cells, suggests a nonmitotic role for this kinase when it is
aberrantly expressed. Overexpression of Aurora-A is not, how-
ever, the only mode of fostering the tumorigenic activity of
RalA through phosphorylation. Phosphatase PP2A A␤ also
affects both the phosphorylation status and the tumorigenicity
of RalA in HEK cells (56). Thus, RalA activation through
RalGEF stimulation by oncogenic Ras in cooperation with
phosphorylation by kinase activation or phosphatase inactiva-
tion promotes tumorigenesis.
We also report that phosphorylation of RalA by Aurora-A
leads to internalization of RalA and to elevated RalBP1 GAP
activity (Fig. 5, 6, and 7). The signiﬁcance of differential sub-
cellular localization and effector utilization upon Aurora-A-
mediated phosphorylation of RalA remains to be determined.
We speculate that Aurora-A phosphorylation of RalA at S194
promotes internalization of RalA, and increases in the associ-
ation of RalA with RalBP1 may spatially restrict the activation
of Cdc42 and Rac1, thereby leading to changes in actin dy-
namics. It is not clear if these changes underlie the above-
mentioned requirement for RalA phosphorylation in transfor-
mation and tumorigenesis. Increased RalA translocation upon
expression of kinase-active Aurora-A
is transforming and
depends upon S194 phosphorylation, even in the absence of
robust additional activation of RalA-GTP. In contrast, mutat-
ing the S194 Aurora-A phosphorylation site did not alter the
ability of constitutively activated RalA to transform cells (see
Fig. S9 in the supplemental material). Whether this reﬂects
pleiotropic effects of Aurora-A
or the ability of an onco
genic activated mutation in RalA to overcome the need for
phosphorylation for transformation is unclear. Thus, while cer-
tainly the increased GTP loading of RalA in the presence of
activated Aurora-A promotes transformation, it is unclear if
the same holds true for internalization of RalA and, if so,
whether this reﬂects suppression of Cdc42 activity or another
Although our studies relate only to the situation in cancer
cells, it is possible that regulation of RalA by Aurora-A in
normal cells may also play a role in normal entry and exit from
mitosis, as Aurora-A is required for chromosome separation
and cytokinesis (1, 18, 41, 43, 68), RalA is also required for
proper cytokinesis (11–13), and the RalBP1 homolog cytocen-
trin is involved in proper centrosome duplication and segrega-
tion (52). In normal cells, Aurora-A-mediated phosphorylation
of RalA may promote the activation of the RalA-RalBP1 com-
plex in a spatially restricted manner to promote the switching
off of endocytosis during mitosis to ensure proper mitotic entry
(13, 55). Even so, as both kinase-active and -inactive versions
of Aurora-A can impair mitosis (43), yet only the kinase-active
mutant transforms rodent cell lines (5) and, as we demonstrate
here, human cells in cooperation with oncogenic Ras, the effect
of Aurora-A on cell transformation may not be solely through
chromosome instability. Phosphorylation of RalA either by a
decrease in phosphatase PP2A a␤ expression (56) or by ectopic
expression of Aurora-A (66) has been found to increase the
level of RalA-GTP.
In summary, ﬁrst, we and others (56, 66) report that the
concurrent activations of two seemingly disparate proteins,
Ras and Aurora-A, converge through a common protein,
RalA, to promote tumorigenesis. Aurora kinase inhibitors are
currently under clinical evaluation for cancer treatment (32,
39). Thus, RalA may be a potential target for some of the
antitumor activity of these inhibitors. In turn, RAS mutations
may be a genetic determinant for patient response to these
inhibitors and establish RalA phosphorylation at S194 as an
important biomarker for their antitumor efﬁcacy. Second, we
found that Aurora-A promotes both internalization of RalA in
an S194-dependent fashion and activation of RalBP1, as mea-
sured by reduced Cdc42-GTP levels. Whether this internaliza-
tion of RalA plays a role in transformation or instead reﬂects
another function of the protein remains to be determined.
We thank Mike White for plasmids pMT3-mycRalBP1 and
This work is supported by NIH grants CA94184 and CA126903
(C.M.C.), CA42978 and CA67771 (C.J.D. and A.D.C.), and CA109550
(A.D.C.). C.M.C. is a Leukemia and Lymphoma Scholar, D.F.K. is a
Leukemia and Lymphoma Fellow, and K.-H.L. and B.B.A. were De-
partment of Defense Breast Cancer Research Predoctoral Scholars.
D.C.B. is supported by an NIH T32 Training Fellowship.
1. Anand, S., S. Penrhyn-Lowe, and A. R. Venkitaraman. 2003. AURORA-A
ampliﬁcation overrides the mitotic spindle assembly checkpoint, inducing
resistance to Taxol. Cancer Cell 3:51–62.
2. Awasthi, S., J. Cheng, S. S. Singhal, M. K. Saini, U. Pandya, S. Pikula, J.
Bandorowicz-Pikula, S. V. Singh, P. Zimniak, and Y. C. Awasthi. 2000.
Novel function of human RLIP76: ATP-dependent transport of glutathione
conjugates and doxorubicin. Biochemistry 39:9327–9334.
3. Awasthi, S., S. S. Singhal, S. Yadav, J. Singhal, K. Drake, A. Nadkar, E.
Zajac, D. Wickramarachchi, N. Rowe, A. Yacoub, P. Boor, S. Dwivedi, P.
Dent, W. E. Jarman, B. John, and Y. C. Awasthi. 2005. RLIP76 is a major
determinant of radiation sensitivity. Cancer Res. 65:6022–6028.
4. Bagrodia, S., S. J. Taylor, C. L. Creasy, J. Chernoff, and R. A. Cerione. 1995.
Identiﬁcation of a mouse p21Cdc42/Rac activated kinase. J. Biol. Chem.
5. Bischoff, J. R., L. Anderson, Y. Zhu, K. Mossie, L. Ng, B. Souza, B. Schryver,
P. Flanagan, F. Clairvoyant, C. Ginther, C. S. Chan, M. Novotny, D. J.
Slamon, and G. D. Plowman. 1998. A homologue of Drosophila aurora
kinase is oncogenic and ampliﬁed in human colorectal cancers. EMBO J.
6. Bivona, T. G., S. E. Quatela, B. O. Bodemann, I. M. Ahearn, M. J. Soskis, A.
Mor, J. Miura, H. H. Wiener, L. Wright, S. G. Saba, D. Yim, A. Fein, I. Perez
de Castro, C. Li, C. B. Thompson, A. D. Cox, and M. R. Philips. 2006. PKC
regulates a farnesyl-electrostatic switch on K-Ras that promotes its associa-
VOL. 30, 2010 AURORA-A PHOSPHORYLATES AND RELOCALIZES RalA 521
tion with Bcl-XL on mitochondria and induces apoptosis. Mol. Cell 21:481–
7. Bos, J. L. 1989. ras oncogenes in human cancer: a review. Cancer Res.
8. Bozidis, P., C. D. Williamson, and A. M. Colberg-Poley. 2007. Isolation of
endoplasmic reticulum, mitochondria, and mitochondria-associated mem-
brane fractions from transfected cells and from human cytomegalovirus-
infected primary ﬁbroblasts. Curr. Protoc. Cell Biol. 37:3.27.1–3.27.23.
9. Burum-Auensen, E., P. M. De Angelis, A. R. Schjolberg, K. L. Kravik, M.
Aure, and O. P. Clausen. 2007. Subcellular localization of the spindle pro-
teins Aurora A, Mad2, and BUBR1 assessed by immunohistochemistry.
J. Histochem. Cytochem. 55:477–486.
10. Cantor, S. B., T. Urano, and L. A. Feig. 1995. Identiﬁcation and character-
ization of Ral-binding protein 1, a potential downstream target of Ral
GTPases. Mol. Cell. Biol. 15:4578–4584.
11. Cascone, I., R. Selimoglu, C. Ozdemir, E. Del Nery, C. Yeaman, M. White,
and J. Camonis. 2008. Distinct roles of RalA and RalB in the progression of
cytokinesis are supported by distinct RalGEFs. EMBO J. 27:2375–2387.
12. Cawthon, R. M., P. O’Connell, A. M. Buchberg, D. Viskochil, R. B. Weiss, M.
Culver, J. Stevens, N. A. Jenkins, N. G. Copeland, and R. White. 1990.
Identiﬁcation and characterization of transcripts from the neuroﬁbromatosis
1 region: the sequence and genomic structure of EVI2 and mapping of other
transcripts. Genomics 7:555–565.
13. Chen, X. W., M. Inoue, S. C. Hsu, and A. R. Saltiel. 2006. RalA-exocyst-
dependent recycling endosome trafﬁcking is required for the completion of
cytokinesis. J. Biol. Chem. 281:38609–38616.
14. Chien, Y., S. Kim, R. Bumeister, Y. M. Loo, S. W. Kwon, C. L. Johnson,
M. G. Balakireva, Y. Romeo, L. Kopelovich, M. Gale, Jr., C. Yeaman, J. H.
Camonis, Y. Zhao, and M. A. White. 2006. RalB GTPase-mediated activa-
tion of the IkappaB family kinase TBK1 couples innate immune signaling to
tumor cell survival. Cell 127:157–170.
15. Chien, Y., and M. A. White. 2003. RAL GTPases are linchpin modulators of
human tumour-cell proliferation and survival. EMBO Rep. 4:800–806.
16. Choudhury, A., M. Dominguez, V. Puri, D. K. Sharma, K. Narita, C. L.
Wheatley, D. L. Marks, and R. E. Pagano. 2002. Rab proteins mediate Golgi
transport of caveola-internalized glycosphingolipids and correct lipid traf-
ﬁcking in Niemann-Pick C cells. J. Clin. Invest. 109:1541–1550.
17. Cifone, M. A., and I. J. Fidler. 1980. Correlation of patterns of anchorage-
independent growth with in vivo behavior of cells from a murine ﬁbrosar-
coma. Proc. Natl. Acad. Sci. U. S. A. 77:1039–1043.
18. Cowley, D. O., J. A. Rivera-Perez, M. Schliekelman, Y. J. He, T. G. Oliver, L.
Lu, R. O’Quinn, E. D. Salmon, T. Magnuson, and T. Van Dyke. 2009.
Aurora-A kinase is essential for bipolar spindle formation and early devel-
opment. Mol. Cell. Biol. 29:1059–1071.
19. Downward, J. 2003. Targeting RAS signalling pathways in cancer therapy.
Nat. Rev. Cancer 3:11–22.
20. Feig, L. A. 2003. Ral-GTPases: approaching their 15 minutes of fame. Trends
Cell Biol. 13:419–425.
21. Fukushige, S., F. M. Waldman, M. Kimura, T. Abe, T. Furukawa, M. Su-
namura, M. Kobari, and A. Horii. 1997. Frequent gain of copy number on
the long arm of chromosome 20 in human pancreatic adenocarcinoma.
Genes Chromosomes Cancer 19:161–169.
22. Giet, R., C. Petretti, and C. Prigent. 2005. Aurora kinases, aneuploidy and
cancer, a coincidence or a real link? Trends Cell Biol. 15:241–250.
23. Gigoux, V., S. L’Hoste, F. Raynaud, J. Camonis, and C. Garbay. 2002.
Identiﬁcation of Aurora kinases as RasGAP Src homology 3 domain-binding
proteins. J. Biol. Chem. 277:23742–23746.
24. Hahn, W. C., C. M. Counter, A. S. Lundberg, R. L. Beijersbergen, M. W.
Brooks, and R. A. Weinberg. 1999. Creation of human tumour cells with
deﬁned genetic elements. Nature 400:464–468.
25. Hamad, N. M., J. H. Elconin, A. E. Karnoub, W. Bai, J. N. Rich, R. T.
Abraham, C. J. Der, and C. M. Counter. 2002. Distinct requirements for Ras
oncogenesis in human versus mouse cells. Genes Dev. 16:2045–2057.
26. Hata, T., T. Furukawa, M. Sunamura, S. Egawa, F. Motoi, N. Ohmura, T.
Marumoto, H. Saya, and A. Horii. 2005. RNA interference targeting aurora
kinase A suppresses tumor growth and enhances the taxane chemosensitivity
in human pancreatic cancer cells. Cancer Res. 65:2899–2905.
27. Hezel, A. F., A. C. Kimmelman, B. Z. Stanger, N. Bardeesy, and R. A.
Depinho. 2006. Genetics and biology of pancreatic ductal adenocarcinoma.
Genes Dev. 20:1218–1249.
28. Jones, S., X. Zhang, D. W. Parsons, J. C. Lin, R. J. Leary, P. Angenendt, P.
Mankoo, H. Carter, H. Kamiyama, A. Jimeno, S. M. Hong, B. Fu, M. T. Lin,
E. S. Calhoun, M. Kamiyama, K. Walter, T. Nikolskaya, Y. Nikolsky, J.
Hartigan, D. R. Smith, M. Hidalgo, S. D. Leach, A. P. Klein, E. M. Jaffee, M.
Goggins, A. Maitra, C. Iacobuzio-Donahue, J. R. Eshleman, S. E. Kern,
R. H. Hruban, R. Karchin, N. Papadopoulos, G. Parmigiani, B. Vogelstein,
V. E. Velculescu, and K. W. Kinzler. 2008. Core signaling pathways in human
pancreatic cancers revealed by global genomic analyses. Science 321:1801–
29. Jullien-Flores, V., O. Dorseuil, F. Romero, F. Letourneur, S. Saragosti, R.
Berger, A. Tavitian, G. Gacon, and J. H. Camonis. 1995. Bridging Ral
GTPase to Rho pathways. RLIP76, a Ral effector with CDC42/Rac GTPase-
activating protein activity. J. Biol. Chem. 270:22473–22477.
30. Jullien-Flores, V., Y. Mahe, G. Mirey, C. Leprince, B. Meunier-Bisceuil, A.
Sorkin, and J. H. Camonis. 2000. RLIP76, an effector of the GTPase Ral,
interacts with the AP2 complex: involvement of the Ral pathway in receptor
endocytosis. J. Cell Sci. 113:2837–2844.
31. Katayama, H., W. R. Brinkley, and S. Sen. 2003. The Aurora kinases: role in
cell transformation and tumorigenesis. Cancer Metastasis Rev. 22:451–464.
32. Keen, N., and S. Taylor. 2004. Aurora-kinase inhibitors as anticancer agents.
Nat. Rev. Cancer 4:927–936.
33. Kendall, S. D., S. J. Adam, and C. M. Counter. 2006. Genetically engineered
human cancer models utilizing mammalian transgene expression. Cell Cycle
34. Khosravi-Far, R., M. A. White, J. K. Westwick, P. A. Solski, M. Chrza-
nowska-Wodnicka, L. Van Aelst, M. H. Wigler, and C. J. Der. 1996. Onco-
genic Ras activation of Raf/mitogen-activated protein kinase-independent
pathways is sufﬁcient to cause tumorigenic transformation. Mol. Cell. Biol.
35. Li, D., J. Zhu, P. F. Firozi, J. L. Abbruzzese, D. B. Evans, K. Cleary, H.
Friess, and S. Sen. 2003. Overexpression of oncogenic STK15/BTAK/Aurora
A kinase in human pancreatic cancer. Clin. Cancer Res. 9:991–997.
36. Lim, K. H., A. T. Baines, J. J. Fiordalisi, M. Shipitsin, L. A. Feig, A. D. Cox,
C. J. Der, and C. M. Counter. 2005. Activation of RalA is critical for
Ras-induced tumorigenesis of human cells. Cancer Cell 7:533–745.
37. Lim, K. H., and C. M. Counter. 2005. Reduction in the requirement of
oncogenic Ras signaling to activation of PI3K/AKT pathway during tumor
maintenance. Cancer Cell 8:381–392.
38. Lim, K. H., K. O’Hayer, S. J. Adam, S. D. Kendall, P. M. Campbell, C. J.
Der, and C. M. Counter. 2006. Divergent roles for RalA and RalB in
malignant growth of human pancreatic carcinoma cells. Curr. Biol. 16:2385–
39. Malumbres, M., and M. Barbacid. 2007. Cell cycle kinases in cancer. Curr.
Opin. Genet. Dev. 17:60–65.
40. Malumbres, M., and M. Barbacid. 2003. RAS oncogenes: the ﬁrst 30 years.
Nat. Rev. Cancer 3:459–465.
41. Marumoto, T., S. Honda, T. Hara, M. Nitta, T. Hirota, E. Kohmura, and H.
Saya. 2003. Aurora-A kinase maintains the ﬁdelity of early and late mitotic
events in HeLa cells. J. Biol. Chem. 278:51786–51795.
42. Marumoto, T., D. Zhang, and H. Saya. 2005. Aurora-A—a guardian of poles.
Nat. Rev. Cancer 5:42–50.
43. Meraldi, P., R. Honda, and E. A. Nigg. 2002. Aurora-A overexpression
reveals tetraploidization as a major route to centrosome ampliﬁcation in
p53-/- cells. EMBO J. 21:483–492.
44. Mitin, N., L. Betts, M. E. Yohe, C. J. Der, J. Sondek, and K. L. Rossman.
2007. Release of autoinhibition of ASEF by APC leads to CDC42 activation
and tumor suppression. Nat. Struct. Mol. Biol. 14:814–823.
45. Moskalenko, S., D. O. Henry, C. Rosse, G. Mirey, J. H. Camonis, and M. A.
White. 2002. The exocyst is a Ral effector complex. Nat. Cell Biol. 4:66–72.
46. Norman, J., M. Franz, R. Schiro, S. Nicosia, J. Docs, P. J. Fabri, and W. R.
Gower, Jr. 1994. Functional glucocorticoid receptor modulates pancreatic
carcinoma growth through an autocrine loop. J. Surg. Res. 57:33–38.
47. O’Hayer, K. M., and C. M. Counter. 2006. A genetically deﬁned normal
human somatic cell system to study ras oncogenesis in vivo and in vitro.
Methods Enzymol. 407:637–647.
48. Okada, T., T. Sawada, T. Osawa, M. Adachi, and K. Kubota. 2008. MK615
inhibits pancreatic cancer cell growth by dual inhibition of Aurora A and B
kinases. World J. Gastroenterol. 14:1378–1382.
49. Pamonsinlapatham, P., R. Hadj-Slimane, F. Raynaud, M. Bickle, C. Cor-
neloup, A. Barthelaix, Y. Lepelletier, P. Mercier, M. Schapira, J. Samson,
A. L. Mathieu, N. Hugo, O. Moncorge, I. Mikaelian, S. Dufour, C. Garbay,
and P. Colas. 2008. A RasGAP SH3 peptide aptamer inhibits RasGAP-
Aurora interaction and induces caspase-independent tumor cell death. PLoS
50. Park, S. H., and R. A. Weinberg. 1995. A putative effector of Ral has
homology to Rho/Rac GTPase activating proteins. Oncogene 11:2349–2355.
51. Qiu, R. G., A. Abo, F. McCormick, and M. Symons. 1997. Cdc42 regulates
anchorage-independent growth and is necessary for Ras transformation.
Mol. Cell. Biol. 17:3449–3458.
52. Quaroni, A., and E. C. Paul. 1999. Cytocentrin is a Ral-binding protein
involved in the assembly and function of the mitotic apparatus. J. Cell Sci.
53. Repasky, G. A., E. J. Chenette, and C. J. Der. 2004. Renewing the conspiracy
theory debate: does Raf function alone to mediate Ras oncogenesis? Trends
Cell Biol. 14:639–647.
54. Rodriguez-Viciana, P., P. H. Warne, A. Khwaja, B. M. Marte, D. Pappin, P.
Das, M. D. Waterﬁeld, A. Ridley, and J. Downward. 1997. Role of phospho-
inositide 3-OH kinase in cell transformation and control of the actin cy-
toskeleton by Ras. Cell 89:457–467.
55. Rosse, C., S. L’Hoste, N. Offner, A. Picard, and J. Camonis. 2003. RLIP, an
effector of the Ral GTPases, is a platform for Cdk1 to phosphorylate epsin
during the switch off of endocytosis in mitosis. J. Biol. Chem. 278:30597–
522 LIM ET AL. MOL.CELL.BIOL.
56. Sablina, A. A., W. Chen, J. D. Arroyo, L. Corral, M. Hector, S. E. Bulmer,
J. A. DeCaprio, and W. C. Hahn. 2007. The tumor suppressor PP2A Abeta
regulates the RalA GTPase. Cell 129:969–982.
57. Shipitsin, M., and L. A. Feig. 2004. RalA but not RalB enhances polarized
delivery of membrane proteins to the basolateral surface of epithelial cells.
Mol. Cell. Biol. 24:5746–5756.
58. Smith, S. C., G. Oxford, A. S. Baras, C. Owens, D. Havaleshko, D. L.
Brautigan, M. K. Safo, and D. Theodorescu. 2007. Expression of ral
GTPases, their effectors, and activators in human bladder cancer. Clin.
Cancer Res. 13:3803–3813.
59. Tatsuka, M., S. Sato, S. Kitajima, S. Suto, H. Kawai, M. Miyauchi, I. Ogawa,
M. Maeda, T. Ota, and T. Takata. 2005. Overexpression of Aurora-A po-
tentiates HRAS-mediated oncogenic transformation and is implicated in
oral carcinogenesis. Oncogene 24:1122–1127.
60. Wang, X., Y. X. Zhou, W. Qiao, Y. Tominaga, M. Ouchi, T. Ouchi, and C. X.
Deng. 2006. Overexpression of aurora kinase A in mouse mammary epithe-
lium induces genetic instability preceding mammary tumor formation. On-
61. Warner, S. L., R. M. Munoz, D. J. Bearss, P. Grippo, H. Han, and D. D. Von
Hoff. 2008. Pdx-1-driven overexpression of aurora A kinase induces mild
ductal dysplasia of pancreatic ducts near islets in transgenic mice. Pancreas
62. White, M. A., C. Nicolette, A. Minden, A. Polverino, L. Van Aelst, M. Karin,
and M. H. Wigler. 1995. Multiple Ras functions can contribute to mamma-
lian cell transformation. Cell 80:533–541.
63. Wolthuis, R. M., B. Bauer, L. J. van ’t Veer, A. M. de Vries-Smits, R. H. Cool,
M. Spaargaren, A. Wittinghofer, B. M. Burgering, and J. L. Bos. 1996.
RalGDS-like factor (Rlf) is a novel Ras and Rap 1A-associating protein.
64. Wolthuis, R. M., N. D. de Ruiter, R. H. Cool, and J. L. Bos. 1997. Stimulation
of gene induction and cell growth by the Ras effector Rlf. EMBO J. 16:6748–
65. Wolthuis, R. M., B. Franke, M. van Triest, B. Bauer, R. H. Cool, J. H.
Camonis, J. W. Akkerman, and J. L. Bos. 1998. Activation of the small
GTPase Ral in platelets. Mol. Cell. Biol. 18:2486–2491.
66. Wu, J. C., T. Y. Chen, C. T. Yu, S. J. Tsai, J. M. Hsu, M. J. Tang, C. K. Chou,
W. J. Lin, C. J. Yuan, and C. Y. Huang. 2005. Identiﬁcation of V23RalA-
Ser194 as a critical mediator for Aurora-A-induced cellular motility and
transformation by small pool expression screening. J. Biol. Chem. 280:9013–
67. Yadav, S., S. S. Singhal, J. Singhal, D. Wickramarachchi, E. Knutson, T. B.
Albrecht, Y. C. Awasthi, and S. Awasthi. 2004. Identiﬁcation of membrane-
anchoring domains of RLIP76 using deletion mutant analyses. Biochemistry
68. Zhang, D., T. Hirota, T. Marumoto, M. Shimizu, N. Kunitoku, T. Sasayama,
Y. Arima, L. Feng, M. Suzuki, M. Takeya, and H. Saya. 2004. Cre-loxP-
controlled periodic Aurora-A overexpression induces mitotic abnormalities
and hyperplasia in mammary glands of mouse models. Oncogene 23:8720–
VOL. 30, 2010 AURORA-A PHOSPHORYLATES AND RELOCALIZES RalA 523