Oncogene (1997) 14, 2793 ± 2801
 1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00
Transforming activity and mitosis-related expression of the AKT2
oncogene: evidence suggesting a link between cell cycle regulation and
oncogenesis
Jin Quan Cheng, Deborah A Altomare, Matias A Klein, Wen-Ching Lee, Gary D Kruh,
Natalie A Lissy and Joseph R Testa
Department of Medical Oncology, Fox Chase Cancer Center, 7701 Burholme Avenue, Philadelphia, Pennsylvania 19111, USA
The AKT2 oncogene encodes a protein-serine/threonine
kinase containing a pleckstrin homology domain characteristic of many signaling proteins. Recently, it was
shown that AKT2 kinase activity can be induced by
platelet-derived growth factor through phosphatidylinositol-3-OH kinase, suggesting that AKT2 may be an
important signal mediator that contributes to the control
of cell proliferation. We previously reported ampli®cation and overexpression of AKT2 in human cancers. To
investigate the transforming activity of AKT2, we used a
retrovirus-based construct to express AKT2 in NIH3T3
cells. Overexpression of AKT2 was found to transform
NIH3T3 cells, as determined by growth in soft agar and
tumor formation in nude mice. The oncogenic activity of
AKT2 was diminished by truncation of a 70-amino acid
proline-rich region at the carboxyl-terminus. To facilitate the characterization of AKT2, we generated
monoclonal and polyclonal antibodies against this
protein. AKT2 was localized to the cytoplasm by cell
fractionation experiments, immunocytochemistry, and
immuno¯uorescence. Protein levels were more abundant
in mitotic cells than in interphase cells. Western blot
analysis of synchronized pancreatic cancer cells demonstrated that the expression level of AKT2 protein in
mitotic cells is three to ®vefold higher than in their
interphase counterparts. A time-course study of phytohemagglutinin-stimulated lymphocytes revealed that
AKT2 mRNA and AKT2 protein levels are highest
48 ± 72 h after addition of mitogen, when cells are
actively dividing. These data suggest that AKT2 could
play a signi®cant role in cell cycle progression and that
the oncogenic activity of overexpressed AKT2 may be
mediated by aberrant regulation of cellular proliferation.
Keywords: AKT2; transformation; cell cycle; serine/
threonine kinase
Introduction
AKT1 and AKT2 are human homologs of the viral
oncogene v-akt. (Staal, 1987). We previously isolated
AKT1 and AKT2 cDNAs and demonstrated that AKT1
is the true human counterpart of v-akt and that AKT2
encodes a closely related protein-serine/threonine
kinase (Cheng et al., 1992). Homologous sequences in
the mouse, designated Akt and Akt2, have also been
Correspondence: JR Testa
Received 22 November 1996; revised 28 February 1997; accepted
28 February 1997
cloned and shown to be highly conserved in
mammalian evolution (Bellacosa et al., 1993; Altomare et al., 1995). AKT1/Akt and AKT2/Akt2 have
also been designated protein kinase B-a and -b or
RAC-PKa and RAC-PKb, respectively (Jones et al.,
1991; Co€er and Woodgett, 1991; Burgering and
Co€er, 1995; Kohn et al., 1995; Cross et al., 1995).
The AKT1 and AKT2 kinases have signi®cant
sequence and structural homology to members of the
protein kinase C (PKC) and cyclic adenosine monophosphate-dependent protein kinase families (Bellacosa
et al., 1991; Cheng et al., 1992). However, unlike PKC
family proteins, AKT1 and AKT2 possess a Src
Homology 2-like (SH2-like) domain at the aminoterminus (Bellacosa et al., 1991; Cheng et al., 1992).
This region is now known to be contained within a
pleckstrin-homology (PH) domain, which has been
identi®ed in a wide range of proteins involved in
cellular signaling and cytoskeletal organization (Cohen
et al., 1995).
AKT2 has been implicated in several types of human
malignancy. In particular, ampli®cation and overexpression of AKT2 has been found in 10 ± 20% of
ovarian carcinomas (Cheng et al., 1992; Bellacosa et
al., 1995) and pancreatic carcinomas (Cheng et al.,
1996; Miwa et al., 1996). Moreover, antisense AKT2
can signi®cantly inhibit the invasiveness and tumorigenicity of pancreatic cancer cells overexpressing this
gene (Cheng et al., 1996). However, alterations of
AKT1 have not been consistently observed in any
human malignancy. In fact, ampli®cation of AKT1 has
been reported in only a single gastric carcinoma (Staal,
1987). In a previous study, Akt was found to be nononcogenic in nude mice when overexpressed in the
nontumorigenic rat T cell lymphoma cell line 5675
(Ahmed et al., 1993).
Recently, AKT1/Akt was identi®ed as a direct
downstream target of phosphatidylinositol-3-OH (PI3) kinase. Akt can be rapidly activated by a variety of
growth factors, such as platelet-derived growth factor
(PDGF), epidermal growth factor, basic ®broblast
growth factor, and insulin through PI-3 kinase
(Franke et al., 1995; Burgering and Co€er, 1995;
Kohn et al., 1995; Cross et al., 1995; Andjelkovic et al.,
1996), and AKT2 kinase activity also appears to be
induced by PDGF (Franke et al., 1995). These ®ndings
suggest that AKT1 and AKT2 may be important signal
mediators that contribute to the control of cell
proliferation and malignant transformation (Downward, 1995).
Although AKT2 has been proposed to contribute
signi®cantly to the pathogenesis of certain human
Preimmune serum
Anti-AKT2+GST-AKT2
cancers (Cheng et al., 1992, 1996; Bellacosa et al., 1995;
Miwa et al., 1996), experimental evidence supporting
this contention has not been documented to date. In
this communication, we describe the results of
experiments designed to characterize the growthpromoting potential of AKT2. We show that overexpression of human AKT2 transforms NIH3T3 cells.
Similar experiments with a truncated AKT2 cDNA
reveal that the carboxyl-terminus of AKT2 is required
for transforming activity. In addition, we also present
data demonstrating that AKT2 is abundant in mitotic
and actively dividing cells, suggesting that AKT2 could
play an important role in cell cycle progression and
malignant transformation.
Anti-CRK
2794
Anti-AKT2
Transforming activity and mitosis-related expression of AKT2
JQ Cheng et al
KD
— 68
p56AKT2 —
p40CRKI —
— 43
p28CRKII —
— 29
Results
Generation of speci®c anti-AKT2 antibodies
To characterize the cellular function of AKT2, a series
of monoclonal and polyclonal anti-AKT2 antibodies
were produced with bacterial glutathione-S-transferase
(GST) fusion proteins containing either the SH2-like
region (aa 1 ± 78), helix domain (aa 79 ± 148), or PH
domain (aa 1 ± 106) of AKT2. For monoclonal
antibodies, 23 hybridomas were isolated that reacted
with AKT2 protein produced in bacteria. Six of these
were examined for speci®city on Western blots
containing puri®ed Akt/AKT1 proteins, four of which
reacted speci®cally with AKT2, i.e. they reacted
strongly with puri®ed AKT2 but not at detectable
levels with puri®ed Akt/AKT1. Three polyclonal
antibodies against AKT2 were generated, each of
which cross-reacted with Akt/AKT1 to a varying
degree. The four monoclonal antibodies and a single
polyclonal antibody that showed minimal crossreactivity with Akt/AKT1 were used for subsequent
studies.
The polyclonal and monoclonal antibodies were ®rst
used for Western blot analysis of protein lysates from
18 pancreatic carcinoma cell lines. Two of these lines
(PANC1 and ASPC1) have ampli®cation and overexpression of AKT2 and show high levels of expression
of the 56 kD AKT2 protein, whereas AKT2 was barely
detectable in the remaining pancreatic cancer cell lines
tested (Cheng et al., 1996). Subsequently, we determined that these anti-AKT2 antibodies can be used
successfully in immunoprecipitation studies. PANC1
cells were metabolically labeled with [35S]methionine,
lysed, and immunopecipitated with anti-AKT2 monoclonal and polyclonal antibodies. Endogenous AKT2
can be immunoprecipitated with either polyclonal or
monoclonal anti-AKT2 antibody, is competed by GSTAKT2 fusion protein, and is not seen in the
immunoprecipitates obtained with preimmune serum
or anti-CRK antibody (Figure 1).
NIH3T3 ®broblasts overexpressing AKT2 are
morphologically transformed, grow in medium with low
serum, form colonies in soft agar, and are tumorigenic in
nude mice
To determine the role of AKT2 in malignant
transformation, full-length and C-terminal truncated
(70 aa) AKT2 (AKT2-Dtail) were cloned into the
Figure 1 Immunoprecipitation of lysates from [35S]methionine
labeled PANC1 cells with anti-AKT2 monoclonal antibody.
AKT2 protein was precipitated only with anti-AKT2 antibody
(®rst lane) and not with anti-CRK antibody (second lane) or
pereimmune serum (fourth lane). GST-AKT2 fusion protein
competed with endogenous AKT2 (third lane)
LXSN retroviral vector and transfected into mouse
NIH3T3 cells. LXSN vector alone and antisense AKT2
and antisense AKT2-Dtail constructs were also
introduced into NIH3T3 cells as negative controls. In
addition, v-abl was transfected into NIH3T3 cells as a
positive control.
In medium containing 10% calf serum, neither fulllength nor truncated AKT2 transfectants formed foci.
After G418 selection, pooled transfected cells and
clonal cell derivatives expressing AKT2 were generated
to determine the e€ects of such expression on cell
morphology and growth. The pooled cells and four
clonal cell lines were selected for further analysis. The
morphology of the LXSN-AKT2 cells di€ered from
that of parental NIH3T3 (Figure 2a,b), and NIH3T3
cells transfected with LXSN alone, LXSN-AKT2-Dtail,
antisense AKT2, or antisense AKT2-Dtail. The LXSNAKT2-transfected cells were more refractile, rounded,
and larger than the negative control cells.
The growth properties of LXSN-AKT2 cells and
LXSN-AKT2-Dtail cells were evaluated in medium
containing a low concentration of serum (Persons et
al., 1988; Chauhan et al., 1993). The LXSN-AKT2transfected cells, including pooled cells and each of the
four clonal cell lines, grew well in medium with 0.1% calf
serum during a 4 week observation period. In this
medium, LXSN-AKT2-transfected cells formed colonies
that were similar to those formed by v-abl transformed
cells (Figure 2d,e), whereas parental cells and LXSNAKT2-Dtail and LXSN transfected cells did not (Table
1). These results suggest that AKT2-transformed cells
have a reduced serum dependence for growth.
Transforming activity and mitosis-related expression of AKT2
JQ Cheng et al
2795
Figure 2 Cell morphology of parental NIH3T3 cells (a) and NIH3T3 cells transfected with AKT2 (b). Note that LXSN-AKT2
transfected-cells are more refractile, rounded, and larger than parental NIH3T3 cells. Cells transfected with AKT2 form colonies in
soft agar (c). Moreover, LXSN-AKT2 transfected-cells, like v-abl transformed cells (d), grow readily in medium with low serum (e)
Colony formation in soft agar was assayed to
investigate anchorage-independent growth. Colonies
of LXSN-AKT2-transfected cells appeared microscopically after 10 days (Figure 2c) and became visible to
the naked eye after 21 days of incubation. Colonies
were not observed in soft agar cultures of parental
NIH3T3 cells or of cells transfected with LXSN or
LXSN-AKT2-Dtail (Table 1).
To investigate the tumorigenic activity of AKT2,
pooled transfectants and clonal cell lines were injected
subcutaneously into nude and scid mice. All mice
inoculated with LXSN-AKT2-transfected cells formed
tumors within 10 ± 28 days following injection. Tumors
developed sooner in mice given a higher inoculum
(66106 cells/mouse) than those inoculated with 26106
cells. All tumors reached a diameter of 10 ± 30 mm
within 20 ± 35 days. Although parental NIH3T3 cells
and LXSN- and LXSN-AKT2-Dtail-transfected cells
each produced small tumors (55 mm in diameter) in
2 ± 3 of 12 mice/cell line, all of these tumors were
detected after 35 days, and no tumor of 510 mm was
observed during the 42-day observation period (Table
1). Western blot analysis revealed expression of AKT2
protein only in LXSN-AKT2-transfected cells and
tumors (Figure 3; Table 1) and in LXSN-AKT2-Dtail
transfected cells (Table 1). The transfection and
tumorigenicity experiments were repeated on a second
occasion using scid mice, and very similar results were
obtained.
Subcellular localization of AKT2
The subcellular localization of AKT2 was examined in
a normal human ®broblast cell line, AKT2-transfected
NIH3T3 cells, and three tumor cell lines (PANC1,
ASPC1, Ovcar 3) previously shown to exhibit
Transforming activity and mitosis-related expression of AKT2
JQ Cheng et al
2796
Table 1 Tumorigenicity of AKT2 and control transfectant NIH3T3 cell lines
AKT2
protein
level
Growth in
low serum
(28 days)
Soft-agar
growth
LXSN-AKT2 pool
+
+
+
LXSN-AKT2 1
+
+
+
LXSN-AKT2 2
+
+
++
LXSN-AKT2 4
+
+
++
LXSN-AKT2 3
++
+
+++
+a
7
7
LXSN-AKT2-Dtail 1
a
+
7
7
LXSN-AKT2-Dtail 2
+a
7
7
LXSN-AKT2-Dtail 3
+a
7
7
LXSN pool
7
7
7
LXSN 1
7
7
7
LXSN 2
7
7
7
7
7
7
7
Cells
LXSN-AKT2-Dtail pool
LXSN 3
7
Parental NIH3T3
a
No. cells
injected
in mice
Tumors
/injection
26106
66106
26106
66106
26106
66106
26106
66106
26106
66106
26106
66106
26106
66106
26106
66106
26106
66106
26106
66106
26106
66106
26106
66106
26106
66106
26106
66106
6/6
6/6
6/6
6/6
6/6
6/6
6/6
6/6
6/6
6/6
1/6
2/6
0/6
1/6
1/6
1/6
1/6
2/6
0/6
1/6
1/6
2/6
0/6
0/6
1/6
1/6
0/6
1/6
Latency
(days)
20
15
19
14
21
16
21
15
18
10
± 28
± 21
± 28
± 24
± 28
± 22
± 28
± 21
± 25
± 19
36
35 ± 42
Tumors410 mm
in diameter by 42 days
6/6
6/6
6/6
6/6
6/6
6/6
6/6
6/6
6/6
6/6
40
36
36
42
36
36
40
36
36
36
40
Truncated AKT2 protein
1
2
3
4
5
6
7
8
p56AKT2
Figure 3 AKT2 expression in transfected cells and tumor
derivatives. Immunoblot of cell lysates was reacted with a
monoclonal anti-AKT2 antibody. High levels of AKT2 expression are observed in NIH3T3 cells transfected with AKT2 (lane 2)
and in independent tumors derived from these cells (lanes 5 ± 8).
No AKT2 expression was detected in parental NIH3T3 cells (lane
1) or in NIH3T3 cells transfected with the LXSN vector alone
(lanes 3 and 4)
ampli®cation and overexpression of AKT2 (Cheng et
al., 1992, 1996). Immunocytochemical staining and
immuno¯uorescence with anti-AKT2 polyclonal and
monoclonal antibodies revealed that AKT2 protein
localizes to the cytoplasm in normal and tumor cells
(Figure 4) as well as in AKT2-transfected NIH3T3 cells
(data not shown). Cell fractionation studies also
showed that AKT2 protein was only present in the
cytosolic fraction (data not shown).
Expression of AKT2 in mitotic and interphase cells
Interestingly, the immunocytochemistry and immuno¯uorescence studies revealed more abundant AKT2
protein within cells in various stages of mitosis
(including prophase, metaphase, anaphase and telophase) than in interphase cells (Figure 4). Confocal
microscopy and quantitative image analysis revealed
that in both tumor and normal cells the average
intensity of total cellular immuno¯uorescence was two
to threefold higher in mitotic cells than in interphase
cells, suggesting that the amount of AKT2 protein
¯uctuates in a cell cycle-dependent manner and is
increased in mitosis.
To con®rm the cytological evidence suggesting that
AKT2 protein is more abundant in mitotic cells, we
prepared mitotic and interphase cells from PANC1
cells treated with nocodazole, which arrests cells in
mitosis. Equal amounts of protein from mitotic and
interphase cells were subjected to Western blot
analysis. The level of AKT2 protein was three to
®vefold higher in mitotic cells than in interphase cells,
whereas the level of AKT1 did not di€er in these two
cell populations (Figure 5).
Additional evidence that AKT2 is associated with
mitotic activity was obtained by a lymphocyte timecourse study. Phytohemagglutinin (PHA)-stimulated
peripheral blood lymphocytes from two healthy
donors were collected at various time points
following addition of mitogen. Northern and
Western blot analyses revealed that the levels of
AKT2 mRNA and AKT2 protein are highest at 48 ±
72 h (Figure 6a), when PHA-stimulated lymphocytes
are actively dividing (Richman, 1980). In contrast, no
change in the level of AKT1 protein was detected in
these cells (Figure 6b).
To determine whether AKT2 activity is elevated in
mitotic cells, we performed an in vitro kinase assay
with equal amounts of protein from mitotic and
interphase PANC1 cells. This experiment revealed
that AKT2 kinase activity is approximately ®vefold
higher in mitotic cells than in their interphase
counterparts (Figure 5), which parallels the elevated
AKT2 protein level observed in mitotic cells.
Transforming activity and mitosis-related expression of AKT2
JQ Cheng et al
2797
Figure 4 Subcellular localization of AKT2 by immunocytochemistry (a, b) and immuno¯uorescence (c ± e) with anti-AKT2
antibodies. For immuno¯uorescence studies, goat anti-rabbit or anti-mouse IgG-FITC conjugate (green) was used as the secondary
antibody; the cells were counterstained with propidium iodide to delineate the nuclei (red). PANC1 cells (a, b, d, e); normal human
®broblasts (c). AKT2 is localized predominantly in the cytoplasm. Note that in both normal and tumor cells there is intense staining
of mitotic cells, a few of which are indicated by arrowheads
➝
MBP
phosphorylation
➝
AKT2
expression
➝
AKT1
expression
Figure 5 AKT2 kinase activity and expression in mitotic (left
side in each panel) versus interphase (right side) PANC1 cells.
Top, AKT2 kinase activity. Middle, immunoblot of cell lysates
reacted with a monoclonal anti-AKT2 antibody. Bottom, the
same immunoblot reacted with a polyclonal anti-AKT1 antibody
Discussion
In this report, we demonstrate that overexpression of
AKT2 in NIH3T3 cells results in a transformed
phenotype, as judged by several criteria, including
proliferation in medium containing minimal serum,
anchorage-independent growth, and tumor formation
in nude mice. The oncogenic activity of overexpressed
AKT2 can be abolished by truncation of 70 aa from
the C-terminus, suggesting that this region may contain
an important functional sequence involved in AKT2
activity and, therefore, potentially malignant transformation. Sequence analysis revealed that this region is
proline rich (6/70 aa) and that a SH3-binding motif (XP-X-X-P-X) (Pawson, 1995; Alexandropoulos et al.,
1995) resides in this region, suggesting that AKT2 may
interact with SH3 domain-containing proteins that may
mediate AKT2 signaling. The proline-rich region of
AKT2 is similar to a binding site for the SH3 domain
of the adapter protein CRK (Alexandropoulos et al.,
1995). However, AKT2 was not immunoprecipitated
with an anti-CRK antibody (Figure 1).
Cancer is considered to be a disorder of the cell cycle
(Pines, 1995). A number of cell cycle regulators have
been identi®ed, some of which are implicated as
products of either oncogenes or tumor suppressor
genes. For example, p53, Rb, p16, and the cyclins D1,
D2 and D3 all are involved in the G1/S transition. The
oncoproteins Mos and Ras can arrest Xenopus
embryonic cell cleavage in mitosis and maintain high
levels of M-phase-promoting factor (Sagata et al.,
Transforming activity and mitosis-related expression of AKT2
JQ Cheng et al
2798
7d
4d
3d
2d
0
3d
2d
4h
1h
0.5h
3d
2d
AKT2
1d
b
2
1d
4h
1h
0.5h
0
1
1d
a
p56AKT2
28S —
p56AKT1
18S —
Figure 6 Northern (a) and Western (b) blot analyses demonstrating the temporal expression of AKT2 in mitogen-stimulated
lymphocytes from two healthy donors, 1 and 2. Cultures were harvested at the indicated times after addition of PHA. (a) Top,
hybridization of Northern blot to radiolableled AKT2 probe. Bottom, ethidium bromide-stained gel showing equivalent amounts of
RNA in each lane. (b) Top, immunoblot of cell lysates reacted with anti-AKT2 antibody. Bottom, same blot reacted with an antiAKT1 polyclonal antibody (Santa Cruz Biotechnology), shown for comparison. Highest levels of AKT2 transcripts and AKT2
protein are observed at 2-day and 3-day time points
1989; Daar et al., 1991). Moreover, a number of other
oncoproteins such as Src, Raf, Ras, erbB1 and erbB2
mediate mitogenic signals and can regulate cell cycle
progression (Muller et al., 1993). Recent studies have
shown that AKT1/Akt can be rapidly activated by
various mitogenic growth factors through PI-3 kinase
(Franke et al., 1995; Burgering and Co€er, 1995; Kohn
et al., 1995; Cross et al., 1995; Andjelkovic et al., 1996)
and AKT2 can also be induced by PDGF (Franke et
al., 1995). These data suggest that AKT1/Akt and
AKT2 may contribute to the control of cell proliferation. In this report, we demonstrate that AKT2 is
elevated in mitotic cells, indicating that this protein
¯uctuates in a cell cycle-dependent manner. Moreover,
AKT2 mRNA and AKT2 protein levels are elevated in
actively dividing lymphocytes, whereas AKT1 levels are
unchanged, suggesting that AKT2 could play a role in
cell cycle progression. Thus, overexpression of AKT2
could contribute to aberrant cell growth associated
with transformation in vitro and tumor formation or
progression in vivo.
In vitro studies by Ahmed et al. (1993) demonstrated
that Akt is not tumorigenic when overexpressed in the
nontumorigenic rat T cell lymphoma cell line 5675,
whereas v-akt-expressing 5675 cells are highly oncogenic when inoculated in nude mice. Since v-akt arose
by way of an in-frame fusion of Gag and Akt, the
oncogenic di€erence between v-akt and Akt may be
due to di€erences in their subcellular localization, i.e.
Akt is restricted to the cytoplasm, whereas v-akt is
distributed almost equally among the plasma membrane, cytoplasm, and nucleus. This association of vakt with the plasma membrane and the nucleus may
permit its interaction with target molecules mediating
oncogenic signals (Ahmed et al., 1993). Interestingly,
AKT2 is tumorigenic when overexpressed in NIH3T3
cells even though it remains localized in the cytoplasm
(data not shown), suggesting that the mechanism by
which AKT2 transforms cells di€ers from that of v-akt.
In addition, AKT2 has been shown to be ampli®ed and
overexpressed in 10 ± 20% of ovarian and pancreatic
carcinomas as well as in a case of non-Hodgkin's
lymphoma (Cheng et al., 1992, 1996; Bellacosa et al.,
1995; Miwa et al., 1996; Arranz et al., 1996). However,
alterations of AKT1 have been observed in only a
single gastric tumor to date (Staal, 1987). The apparent
di€erences between Akt/AKT1 and AKT2 with respect
to their oncogenic activity and involvement in human
malignancy suggest that the signal transduction pathways of these two closely related proteins may di€er,
particularly with regard to their downstream targets or
substrates.
Finally, we have generated several speci®c antiAKT2 monoclonal antibodies. We have tested several
commercially available anti-AKT2 antibodies and have
found that each reacts with both AKT1 and AKT2.
The use of such antibodies can pose problems in
interpreting certain experimental data, particularly
since AKT1 and AKT2 proteins have the same
mobility on SDS ± PAGE. Thus, our monoclonal
antibodies will be useful for future studies of AKT2
function.
Materials and methods
Expression constructs and DNA transfections
The vector (LXSN) utilized in the generation of the
expression constructs contains a promoter from the 5'
long terminal repeat of the Moloney murine leukemia virus
and a neomycin resistance gene suitable for the selection of
transfectants. The AKT2 construct expresses the entire
AKT2 open reading frame (aa 1 to 481). AKT2-Dtail
expresses a truncated AKT2 protein (aa 1 to 411) with 70
aa deleted from the carboxy terminus. As controls,
antisense AKT2 and antisense AKT2-Dtail constructs were
also generated.
Transforming activity and mitosis-related expression of AKT2
JQ Cheng et al
Transfections were carried out in NIH3T3 ®broblasts.
Lipofectin Reagent (Gibco BRL) was used to transfect cells
with 10 mg of each DNA construct as described previously
(Cheng et al., 1996). Forty-eight h after transfection, G418
was added to the medium (Dulbecco's modi®ed Eagle
medium [DMEM] supplemented with 10% calf serum) at a
concentration of 400 mg/ml. Individual colonies were isolated
and expanded for further analysis.
Preparation of mitotic PANC1 cells
PANC1 cells were grown as monolayers at 378C in DMEM
containing 10% fetal calf serum. To isolate mitotic cells,
nocodazole
(methyl[5-(2-thienylcarbonyl)-1H-benzimidazol-2-yl] carbamate; Sigma) was added to the culture
medium at a ®nal concentration of 0.4 mg/ml. After
incubating for 14 h in the presence of nocodazole,
approximately 50% of the cells were rounded and readily
detached following gentle tapping of the ¯asks and
washing with medium. The ¯oating cells were harvested
by centrifugation and washed in cold phosphate bu€ered
saline (PBS) prior to lysis. An aliquot of the cell suspension
was ®xed in methanol, dropped onto a clean glass slide,
and allowed to air-dry; microscope analysis con®rmed that
more than 90% of these cells were in mitosis. Interphase
cells that remained ®rmly attached to the surface of the
¯asks were washed with cold PBS and then lysed with NP40 bu€er (50 mM Tris-HCl [pH 7.4], 150 mM NaCl,
20 mM EDTA, 0.5% NP-40, 1 mM phenylmethylsulfonyl
¯uoride [PMSF], 2 mg/ml aprotinin, 2 mg/ml leupeptin,
1 mM benzamidine, and 10 mg/ml trypsin inhibitor).
Stimulation of lymphocyte mitotic activity
Peripheral blood mononuclear cells (PBMC) were isolated
from two healthy donors by Ficoll (Pharmacia LKB)
density gradient centrifugation. Subsequent to incubation
on gelatin-coated ¯asks, the PBMC composition was
determined by cell-sorter analysis to be 89% lymphocytes
and 11% monocytes. The PBMC population was resuspended in RPMI 1640 medium supplemented with 15%
fetal calf serum and 2 m M L-glutamine. Cells were plated
in 6-well culture dishes at a density of 26106 cells/ml in the
presence of 0.2% PHA (Gibco BRL) to stimulate
lymphocyte mitogenic activity. PBMCs were collected at
various times (0.5 h to 72 h for RNA; 1 day to 7 days for
protein) after addition of PHA for analysis of gene
expression.
GST fusion protein and generation of anti-AKT2 antibodies
The SH2-like region (aa 1 ± 78), helix domain (aa 79 ± 148),
and PH domain (aa 1 ± 106) of AKT2 were each introduced
into the prokaryotic expression vector pGEX-3X (Pharmacia). The resulting constructs direct the synthesis of
GST-SH2-like, GST-helix, and GST-PH fusion proteins.
Expression and puri®cation of the GST fusion proteins
were essentially as described by Smith and Corcoran
(1994). Brie¯y, logarithmically growing cultures of E. coli
JM83 transformed with the pGEX-3X recombinant were
incubated with 0.1 mM isopropyl-b-D-thiogalactopyranoside at 378C for 6 h. The cells were then pelleted,
resuspended in ice-cold PBS containing 2 mg/ml aprotinin,
1 mg/ml leupeptin, and 1 mM PMSF, and then sonicated on
ice. Debris was removed by centrifugation, and the
supernatant was then applied to a glutathione sepharose
4B column (Pharmacia). GST fusion protein was eluted
from the column and separated by SDS-polyacrylamide gel
electrophoresis and puri®ed with an Electro-Eluter (BioRAD).
Polyclonal anti-AKT2 antibodies were raised in New
Zeealand White rabbits. Approximately 300 mg of GST
fusion protein was used to immunize each rabbit every 2
weeks; rabbits were bled 10 days after each booster injection.
The antibodies were then anity puri®ed. Monoclonal
antibodies were generated by standard methods (McKenzie
et al., 1989). Female CB6 F1 mice were immunized by
intraperitoneal injection on three successive occasions with
20 mg of GST fusion protein. Hybridomas were produced as
described (McKenzie et al., 1989). Test bleeds and
hybridomas were screened for anti-AKT2 reactivity by
ELISA, using puri®ed GST-SH2-like, GST-helix or GSTPH coated on plates at a concentration of 500 ng/ml.
Antibody-antigen complexes were detected by horseradish
peroxidase-conjugated goat anti-mouse IgG (Boehringer
Mannheim).
Metabolic labeling of cells, immunoprecipitation, and immunoblotting
For metabolic labeling with [35S]methionine, subcon¯uent
(50 ± 70%) PANC1 cells were washed twice with prewarmed labeling media (methionine- and cysteine-free
DMEM [ICN] supplemented with 10% dialyzed fetal
bovine serum). After a 30-min incubation with labeling
media, [35S]methionine (Trans 35S-label, ICN) was added to
the medium to a ®nal concentration of *200 mCi/ml and
then incubated for an additional 6 h prior to lysis. All
immunoprecipitation steps were carried out at 48C. Cells
from a 75 cm2 ¯ask were washed twice with cold PBS and
scraped into NP-40 lysis bu€er and then lysed by rotating
for 20 min. Nuclei were removed by centrifugation at
14 000 g for 5 min, and lysates were precleared by mixing
with 20 ml of a 50% suspension of protein A-protein G
(2:1) agarose (Gibco BRL) for 20 min. Following the
removal of the beads by centrifugation, the lysates were
incubated with antibody in the presence of 30 ml of protein
A-protein G agarose slurry for 6 h. The resulting
immunoprecipitates were washed three times in lysis
bu€er and once in 50 mM Tris-HCl (pH 8.0) at room
temperature, resuspended in SDS sample bu€er, and
separated on SDS-polyacrylamide gels. The gels were
®xed with 10% glacial acetic acid and 30% methanol for
30 min, enhanced by impregnating with autoradiography
enhancers (DuPont) for 30 min, and precipitated in water
for 15 min. Enhanced gels were dried and exposed to X-ray
®lm.
Immunoblotting was carried out as described previously
(Cheng et al., 1996). Brie¯y, equivalent amounts of protein
from each cell line were separated on a SDS-polyacrylamide
gel and transferred onto Hybond-C Super membrane
(Amersham) using a Trans-Blot Semi-Dry Transfer Cell
(Bio-Rad). Following overnight incubation in 5% dry milk
in PBS at 48C, the ®lters were exposed for 1 h at room
temperature to anti-AKT2 antibody in blocking solution (1%
dry milk), followed by washing. Detection of AKT2 protein
was carried out with the ECL Western Blotting Analysis
System (Amersham).
In vitro kinase assay
Immunoprecipitation was performed for 4 h using antiAKT2 antibody. Immunopecipitates were washed three
times with lysis bu€er, once with water, and once with
kinase bu€er (20 mM HEPES-NaOH, 10 mM MgCl2,
10 mM MnCl2, [pH 7.4]). Kinase assays were carried out
in 50 ml kinase bu€er containing 10 mCi [g-32P]ATP, 2 mM
cold ATP, and 25 mg myelin basic protein or 25 mg histone
H2B. After incubation at 308C for 30 min, samples were
analysed by 12% SDS ± PAGE and then autoradiographed.
For measurement of protein kinase activity, the phosphorylated substrate bands were excised from the dried gel,
and radioactivity was counted by scintillation spectrophotometry.
2799
Transforming activity and mitosis-related expression of AKT2
JQ Cheng et al
2800
Immunocytochemical staining, immuno¯uorescence, and cell
fractionation
Immunocytochemical staining was performed with an ABC
kit according to the manufacturer's recommendations
(Vector). Immuno¯uorescence was carried out as described
by Willingham (1990). Cells were grown overnight on tissue
culture chamber slides. After ®xation with 3% paraformaldehyde at room temperature for 10 min, the cells were
washed in PBS containing 0.1% saponin (Sigma). Nonspeci®c binding was blocked by treatment with 5% serum,
1% BSA, and 0.1% saponin for 1 h at 378C. Anti-AKT2
antibodies were used at a 1:100 dilution. The cells were
washed thoroughly in PBS containing 0.1% saponin. Goat
anti-rabbit or anti-mouse IgG-FITC conjugate was used as
the secondary antibody, followed by extensive washing in
PBS. The cells were then counterstained with propidium
iodide (Oncor) and observed with a Zeiss Axiphot
¯uorescence microscope. Images were captured on a cooled
CCD camera (Photometrics) connected to a Macintosh
computer work station. Images of propidium iodide staining
and FITC signals were merged using Oncor Image software,
version 1.6. In addition, ¯uorescence corresponding to antiAKT2 binding sites was quanti®ed in single cells (approximately 20 mitoses and 20 interphases from normal
®broblasts and a similar number from PANC1 cells) using
a Bio-Rad MRC 600 confocal laser photometer. FITC was
excited with a 25 mW krypton/argon laser at 10%
transmission, and lines at 488 nm and 568 nm were used
for FITC and propidium iodide, respectively. The fluorescence emission was recorded at wavelengths greater than
515 nm. The image scans were obtained with pixels of
0.3 mm2 using 4095 relative levels of ¯uorescence with a
606oil/NA 1.3 objective and an optical thickness of 0.7 mm.
Cell fractionation studies were carried out as described by
Smith et al. (1993). cells were centrifuged at 200 g and
washed three times with PBS containing 0.5 mM CaCl2. The
pellet was resuspended in lysis bu€er (5 mM Tris-HCl
[pH 7.8], 2 mM MgCl2. The nuclei and unlysed cells were
pelleted by centrifugation at 500 g. The pellet was washed
once with lysis bu€er and saved as the `nuclear fraction', The
supernatant was then centrifuged at 1000 000 g for 1 h; the
resulting pellet was designated the `membrane fraction', and
the supernatant was designated the `cytosolic fraction.'
Soft agar assay and tumorigenicity in nude mice
For soft agar assays, three scalar concentrations (16103,
16104, 16105) of the pooled AKT2-transfected cells were
suspended in 2 ml of 0.35% (wt/vol) agar containing
DMEM/20% fetal calf serum, overlaid onto a 0.5% agar
solution in 35 mm plates. The cultures were fed once a
week for 4 weeks.
Tumor formation in 6-week old female nude/nude mice
and scid mice was tested by subcutaneous inoculation and
assessed for 6 weeks. Parental NIH3T3 cells and cells
transfected with LXSN, LXSN-AKT2 or LXSN-AKT2Dtail were each injected into 6 animals at 26106 cells/mouse
and another six animals at 66106 cells/mouse. Tumor
measurements were made with linear calipers in two
orthogonal directions by the same observer. To establish
cell lines derived from individual tumors that formed in mice,
portions of each tumor were chopped into ®ne pieces and
treated with collagenase type I (400 units/ml) for 2 h. The cell
suspension was centrifuged, washed twice with DMEM
medium without serum, and cultured at 378C in DMEM
medium containing 10% calf serum.
Acknowledgements
We thank Drs T Mysliwiec, K Alpaugh and A Godwin for
assistance with the transfection experiments, lymphocyte
mitogenesis studies and Northern analyses, respectively,
and AM Helt for carrying out the confocal microscopy and
cell imaging studies. We also thank Dr PN Tsichlis for
kindly providing puri®ed Akt/KT1 protein used for the
antibody speci®city tests and Dr A Bellacosa for review of
the manuscript. This work was supported by American
Cancer Society grant #EDT-6, National Cancer Institute
Grant CA-06927, and by an appropriation from the
Commonwealth of Pennsylvania. DAA was supported by
National Institutes of Health Postdoctoral Training Grant
CA 09035 awarded to Fox Chase Cancer Center. JQC and
W-CL are recipients of Special Fellow and Fellow Awards
from the Leukemia Society of America, respectively.
References
Ahmed NN, Franke TF, Bellacosa A, Datta K, GonzalezPortal M-E, Taguchi T, Testa JR and Tsichlis PN. (1993).
Oncogene, 8, 1957 ± 1963.
Altomare DA, Guo K, Cheng JQ, Sonoda G, Walsh K and
Testa JR. (1995). Oncogene, 11, 1055 ± 1060.
Alexandropoulos K, Cheng G and Baltimore D. (1995).
Proc. Natl. Acad. Sci. USA, 92, 3110 ± 3114.
Andjelkovic M, Jakubowicz T, Cron P, Ming X-F, Han J-W
and Hemmings BA. (1996). Proc. Natl. Acad. Sci. USA,
93, 5699 ± 5704.
Arranz E, Robledo M, Martinez B, Gallego J, Roman A,
Rivas C and Benitez J. (1996). Cancer Genet. Cytogenet.,
87, 1 ± 3.
Bellacosa A, Testa JR, Staal SP and Tsichlis PN. (1991).
Science, 254, 274 ± 277.
Bellacosa A, Franke TF, Gonzalez-Portal ME, Datta K,
Taguchi T, Gardner J, Cheng JQ, Testa JR and Tsichlis
PN. (1993). Oncogene, 8, 745 ± 754.
Bellacosa A, Feo D, Godwin AK, Bell DW, Cheng JQ,
Altomare D, Wan M, Dubeau L, Scambia G, Masciullo V,
Ferrandina G, Panici PB, Mancuso S and Testa JR.
(1995). Int. J. Cancer, 64, 280 ± 285.
Burgering BMTh and Co€er PJ. (1995). Nature, 376, 599 ±
602.
Chauhan AK, Li Y-S and Deuel TF. (1993). Proc. Natl.
Acad. Sci. USA, 90, 679 ± 682.
Cheng JQ, Godwin AK, Bellacosa A, Taguchi T, Franke TF,
Hamilton TC, Tsichlis PN and Testa JR. (1992). Proc.
Natl. Acad. Sci. USA, 89, 9267 ± 9271.
Cheng JQ, Ruggeri B, Klein WM, Sonoda G, Altomare DA,
Watson DK and Testa JR. (1996). Proc. Natl. Acad. Sci.
USA, 93, 3636 ± 3641.
Co€er PJ and Woodgett JR. (1991). Eur. J. Biochem., 210,
475 ± 481.
Cohen GB, Ren R and Baltimore D. (1995). Cell, 80, 237 ±
248.
Cross DAE, Alessl DR, Cohen P, Andjelkovich M and
Hemmings BR. (1995). Nature, 378, 785 ± 789.
Daar I, Nebreda AR, Yew N, Sass P, Paules R, Santos E,
Wigler M and Vande Woude GF. (1991). Science, 253,
74 ± 76.
Downward J. (1995). Nature, 376, 553 ± 554.
Franke TF, Yang S-I, Chan TO, Datta K, Kazlauskas A,
Morrison DK, Kaplan DR and Tsichlis PN. (1995). Cell,
81, 727 ± 736.
Jones PF, Jakubowicz T, Pitossi FJ, Maurer F and
Hemmings BA. (1991). Proc. Natl. Acad. Sci. USA, 88,
4171 ± 4175.
Transforming activity and mitosis-related expression of AKT2
JQ Cheng et al
Kohn AD, Kovacina KS and Roth RA. (1995). EMBO J.,
14, 4288 ± 4295.
McKenzie SJ, Marks PJ, Lam T, Morgan J, Panicali DL,
Trimpe KL and Carney WP. (1989). Oncogene, 4, 543 ±
548.
Miwa W, Yasuda J, Murakami Y, Yashima K, Sugano K,
Sekine T, Kono A, Egawa S, Yamaguchi K, Hayashizaki
Y and Sekiya T. (1996). Biochem. Biophys. Res. Commun.,
225, 968 ± 974.
Muller R, Mumberg D and Lucibello FC. (1993). Biochem.
Biophys. Acta, 1155, 151 ± 179.
Pawson T. (1995). Nature, 373, 573 ± 580.
Persons AD, Wilkison WO, Bell RM and Finn OJ. (1988).
Cell, 52, 447 ± 458.
Pines J. (1995). Cancer Biol., 6, 63 ± 72.
Richman DP. (1980). J. Cell Biol., 85, 459 ± 465.
Sagata N, Watanabe N, Vande Woude GF and Ikawa Y.
(1989). Nature, 342, 512 ± 518.
Smith DB and Corcoran LM. (1994). Current Protocols in
Molecular Biology. Ausubel FM, Brent R, Kingston RE,
Moore DD, Seidman JG, Smith JA, Struhl K (eds). John
Wiley & Sons: New York, pp. 16.7.1 ± 16.7.7.
Smith KJ, Johnson KA, Bryan TM, Hill DE, Markowitz S,
Willson JKV, Paraskeva C, Petersen GM, Hamilton SR,
Vogelstein B and Kinzler KW. (1993). Proc. Natl. Acad.
Sci. USA, 90, 2846 ± 2850.
Staal SP. (1987). Proc. Natl. Acad. Sci. USA, 84, 5034 ± 5037.
Willingham MC. (1990). BRL Focus, 12, 62 ± 67.
2801