| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Tumor Biology |
Ludwig Institute for Cancer Research and the Experimental Medicine Unit of the Université Catholique de Louvain, B-1200 Brussels, Belgium [J-B. D., C. U., D. L., J-C. R.]; Institute for Biomedical Research, D-60596 Frankfurt/Main, Germany [B. G.]; and Department of Surgery, University of British Columbia, Jack Bell Research Centre, Vancouver Hospital and Health Sciences Centre, Vancouver, British Columbia, V6H 3Z6 Canada [A. M.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
), we demonstrated
that this factor is an important mediator of IL-9-dependent Ba/F3 cell
growth. Mutation of the STAT binding site of the IL-9 receptor
(tyr116phe) results in an important decrease in STAT activation and
inhibition of proliferation in the presence of IL-9. A small number of
cells escape this inhibition, and IL-9-dependent cell lines could be
derived. The selected cells required activation of STAT5 for growth,
which was blocked by STAT5 expression and enhanced by overexpression
of wild-type STAT5. In contrast to parental cells, Ba/F3-Phe116 cells
growing in the presence of IL-9 further progress to
cytokine-independent tumorigenic clones. These tumorigenic clones
exhibited a strong cytokine-independent activation of JAK1 and STAT5,
which most likely supports their proliferation. Transfection of a
constitutively activated variant of STAT5 promoted the growth of
wild-type Ba/F3 cells in the absence of cytokine. Finally, the
expression of the proto-oncogene pim-1 was
correlated with STAT5 activation and cell growth. Our data suggest that
STAT5 is an important mediator of IL-9-driven proliferation and that
dysregulation of STAT5 activation favors tumorigenesis of lymphoid
cells. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
IL-9 binds to a receptor comprising a specific chain (IL-9R) and the
c chain, which is shared by receptors for IL-2, IL-4, IL-7, IL-9,
and IL-15 (11)
. IL-9 effects are mediated through the
activation of JAK1 and JAK3 tyrosine kinases and STAT transcription
factors, namely STAT1, STAT3, and STAT5 (12–15)
.
Activation of the JAK-STAT pathway by cytokines has been studied
extensively (16)
. On ligand binding to hematopoietic
receptors, JAK kinases are activated and phosphorylate the cytoplasmic
part of the receptor, creating phospho-tyrosine docking sites for STAT
factors. Recruited STATs are subsequently phosphorylated by JAKs,
dissociate from the receptor, and form stable dimers that migrate into
the nucleus, where they bind to promoter sites and regulate the
expression of genes (16)
. A transactivation domain has
been located in the COOH terminus of most STATs. Deletion of this
domain results in a dominant negative phenotype (17)
.
STATs are active players in malignant transformation. Fusion proteins generated by chromosomal translocation, such as BCR-ABL or TEL-JAK2, and viral oncogenes, for instance v-SRC, HBx, and v-Eyk, are able to activate STATs (16, 18) . STAT3 is required for transformation of NIH3T3 cells by v-SRC (19, 20) . A constitutively active STAT3 mutant was shown to transform fibroblasts in vitro (21) . Many human leukemias and lymphomas are associated with constitutive activation of the JAK-STAT pathway (22–24) . In most cases, however, the mechanism that underlies STAT activation in these tumors is unknown.
Apoptosis inhibition by STAT3 and STAT5 is well documented, and it is likely that it accounts for their role in oncogenesis (14, 25) . The effect of STATs on proliferation is more complex. On the one hand, STAT1, STAT3 and STAT5 have been shown to regulate cell cycle inhibitors, such as p21waf1, resulting in cell growth inhibition and differentiation (26, 27) . On the other hand, STAT3 and STAT5 play a role in proliferation induced by IL-6 and IL-3, respectively (17, 28) . Moreover, hematopoietic cells and T-lymphocytes from STAT5-deficient mice exhibit a decreased response to growth stimuli (29) . Two recent studies have suggested that STATs simultaneously regulate genes that stimulate and inhibit cell growth, the phenotypic outcome depending on the intensity and duration of the expression of both types of genes (26, 28) .
We have reported that STAT activation by IL-9 correlates with the induction of proliferation and apoptosis inhibition (12, 14) . Here, we provide direct evidence that STAT5 plays an important role in IL-9-dependent growth and the malignant transformation of lymphoid cells.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Wild-type and mutated human IL-9R cDNAs were inserted into either the
pEFbos/puro plasmid (12)
or pEF/myc/cyto plasmid
(Invitrogen, Carlsbad, CA), which contain a resistance gene to
puromycin or to neomycin, respectively. Ba/F3 cells were transfected by
electroporation (300 V, 1500 µF, 74 
, SEDD device,
Eurogentec, Belgium) with 50 µg of DNA, and selected with puromycin
(3 µg/ml; Sigma, Bornem, Belgium) or with G418 (3 mg/ml; Sigma).
Resistant populations and clones of cells were used. Human IL-9R
expression was similar in transfected cells, as tested by
fluorescence-activated cell sorter analysis with biotinylated
anti-IL-9R antibody AH9R1 (12)
, and
phycoerythrin-conjugated streptavidin (Becton Dickinson, San Jose, CA).
STAT5/VP16/JAK2 was inserted into the pEFbos/puro plasmid and
transfected in Ba/F3 cells as described above (31)
. Cells
were cultured for 24 h in the presence of IL-3, then were washed
and cloned in the absence of cytokine by limiting dilution, as
described by Nosaka et al. (26)
.
Proliferation Assays.
Short-term proliferation was assessed by the hexoseaminidase method.
Briefly, cells were extensively washed and seeded in microtiter plates
(3000 cells/well) in the presence of cytokines. After 3 days, cell
growth was measured by colorimetric determination of hexoseaminidase
levels (30)
. SDs were calculated from triplicate culture.
The frequency of autonomous clones was assessed in microtiter plates in the absence of cytokine (5 x 104 washed cells/well). Plates were checked for proliferating clones for 3 weeks. Clones were subsequently cultured in the same conditions as Ba/F3 cells, but without IL-3.
EMSAs.
EMSA experiments were performed as described previously
(32)
. Briefly, nuclear extracts (4 µl) were incubated
for 5 min with 11 µl of water and 4 µl of buffer [60
mM HEPES (pH 7.6), 50 mM KCl, 2.75
mM EDTA, 17.5% glycerol, 0.5 mg/ml
poly(dI-dC)·poly(dI-dC)] from ICN (Costa Mesa, CA). The labeled
oligonucleotide (105 cpm) that was derived from
the GRR of the FcRI gene promoter (upper strand:
5'-ATGTATTTCCCAGAAA-3'; bottom strand: 5'-CCTTTTCTGGGAAATAC-3') was
then added for 25 min. This probe binds all of the known STAT factors
(12)
. Super-shifts were performed with the following
antibodies: anti-STAT1 (0.75 µg; G16920; Transduction
Laboratories); anti-STAT3 (1 µg; Zymed, San Francisco, CA); or
anti-STAT5 (1 µg; SC835; Santa Cruz, CA). Complexes were
separated in a 5% nondenaturing polyacrylamide gel and visualized by
autoradiography.
Tumorigenicity Assay in SCID Mice.
Cells were washed and resuspended in serum-free medium. Cells
(106) were injected i.p. into female SCID mice (6–9 weeks
old). Mice were kept in sterile cages with sterile water and food
ad libitum. Survival was followed for 3 months.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
), which is
truncated after tyrosine 683 and lacks a transactivation domain
(17)
. In these cells, STAT5 expression is controlled by
a tetracyclin-sensitive tTA regulator (Fig. 1A
). We introduced the IL-9R into this cell line and observed
that STAT5 completely inhibited the activation of STAT5 by IL-9
(Fig. 1B
) and significantly reduced cell growth (Fig. 1C
). Activation of STAT1 and STAT3 was barely affected. To
rule out a nonspecific effect of STAT5, we analyzed cells that
overexpress wild-type STAT5 in addition to STAT5. Overexpression of
wild-type STAT5 compensated for the effects of STAT5 and even
increased proliferation in response to IL-9 (Fig.
1D–F). These experiments provide direct
evidences that STAT5 is an important mediator of IL-9-dependent cell
growth.
|
). We called these selected cells Ba/F3-Phe116/9. In the
presence of high IL-9 concentrations (>100 units/ml), they
proliferated at the same rate as Ba/F3-IL-9R (Fig. 2B
).
However, they were unable to grow in low IL-9 concentrations (1–10
units/ml, corresponding to ± 0.1–1 ng/ml). This
selection was reproduced with cell lines and clones derived from three
independent transfections.
|
). The 44–98 IL-9R domain is involved
in the activation of JAK kinases (12, 13)
. In summary, it
was possible to select Ba/F3 cell clones that expressed a receptor
lacking the STAT-binding site, and nevertheless proliferated in the
presence of IL-9.
Proliferation of Ba/F3-Phe116/9 Cells Depends on Residual STAT5
Activation.
We next tested whether the selection processes affected STAT
activation. Mutation of IL-9R tyrosine 116 into phenylalanine
reduces STAT activation by IL-9 to a very low level (Fig. 3
; Refs. 12, 15 ). A weak STAT5
activation was still detectable, however. After selection for
IL-9-dependent growth, it was increased (Fig. 3, right)
.
Similar results were obtained with three independent polyclonal
Ba/F3-Phe116 cell populations and with three clones. The expression of
IL-9R and STAT5 proteins was not affected by the selection process (not
shown).
|
transfected with Phe116. In these cells,
STAT5 completely blocked the activation of STAT5 and the selection
of IL-9-responsive cells at high IL-9 concentrations for long
observation periods (Fig. 4
). As observed for the wild-type receptor, expression of a large amount
of wild-type STAT5 restored—and even increased—IL-9 responsiveness
(Fig. 4C
). These data indicated that proliferation of
Ba/F3-Phe116/9 cells requires STAT5 activation.
|
). Moreover, no autonomous clones were obtained from Ba/F3-IL-9R cells
cultured in IL-9 for 1 month, which showed that the emergence of
autonomous clones was linked to the selection process, and not to
normal culture with IL-9. Autonomous clones were tumorigenic in
vivo when injected into immunodeficient SCID mice (Fig. 5
) whereas Ba/F3-IL-9R, Ba/F3-Phe116, and Ba/F3-Phe116/9 cell lines were
not. Similar results were obtained in irradiated BALB/c mice (not
shown). These experiments indicate that the selection of
IL-9-responsive Ba/F3-Phe116/9 cells is correlated with the acquisition
of a pretransformed phenotype.
|
|
Constitutive Activation of STAT5 and JAK1 in Tumorigenic Clones.
We next tested whether STAT5 was activated in autonomously growing
Ba/F3 cells. In EMSA experiments, strong constitutive activation of
STAT5 was observed in tumorigenic Ba/F3 clones (Fig. 6A
). A recent report (36)
suggests that
constitutive activation of STAT5 in Ba/F3 cells, induced by
transfecting a constitutively activated STAT5 mutant, is enough to
promote proliferation in the presence of serum. We obtained similar
results with a construct encoding a fusion protein composed of
STAT5, the transactivator domain of VP16 and the kinase domain of
JAK2. This STAT5/VP16/JAK2 protein specifically mimics STAT5 activation
in various cells including Ba/F3 (Fig. 6B
; Refs 14, 31
) and was able to support the growth of the Ba/F3 cells in the
absence of cytokine (Fig. 6C
). Thus, the constitutive
activation of STAT5 observed in autonomous clones is sufficient to
account for the autonomous phenotype of these cells. Incidentally,
neither autonomous clones nor Ba/F3-STAT5/VP16/JAK2 cells were able to
grow in the absence of serum (data not shown).
|
10-6; Ref. 3
). TS1
is an IL-9-dependent cell line derived from T-helper cells grown in the
presence of keyhole limpet hemocyanin antigen and antigen-presenting
cells (30)
. We tested the presence of activated STATs in
four TS1 autonomous clones and observed a constitutive activation of
STATs, including STAT5 (Fig. 6D
). This is in line with our
results in Ba/F3 cells. Thus, in the TS1 system, the constitutive
activation of STATs also correlated with autonomous growth and
tumorigenesis.
STAT activation in transformed cells has been associated with JAK
tyrosine kinase activation (16, 24)
. In autonomous clones,
we found that JAK1 was constitutively phosphorylated on tyrosines (Fig. 7
). JAK2 phosphorylation was not detectable.
|
shows that the Phe116 mutation dramatically affected pim-1
induction by IL-9, as described previously (14)
. In
IL-9-responsive Ba/F3-Phe116/9 cells, we observed significantly
increased effects of IL-9 on pim-1 expression. As expected, pim-1 was
constitutively expressed in tumorigenic clones derived from
Ba/F3-Phe116/9 cells (Fig. 8, right)
. pim-1 was also present
in autonomously growing Ba/F3 cells obtained by transfecting
STAT5/VP16/JAK2 but not in those transformed with an activated mutant
of M-ras (34)
. Thus, the expression of pim-1 correlated
with STAT5 activation and may be a mediator of IL-9-dependent and
autonomous proliferation of Ba/F3 cells.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, which suggests redundancy in IL-9
signaling. Indeed, STAT1 and STAT3 may also play a role in the process,
as suggested by a previous mutagenesis study (14)
. Previous experiments indicated that the mechanism of STAT5 activation by IL-9 involves the docking onto phosphorylated tyrosine 116 of the receptor. Mutation of this single STAT-binding site of the IL-9R results in impaired proliferation and STAT activation (12) . However, a weak STAT5 activation by IL-9 was still detectable in Ba/F3-Phe116/9. Because tyrosine 116 is the only phosphorylated tyrosine of the IL-9R (12) , the residual activation of STAT5 observed in Ba/F3-Phe116 cells must occur independently of IL-9R tyrosine phosphorylation. Fujitani et al. (39) have reported that STAT5, but not STAT1 or STAT3, can be activated directly by JAK1 and JAK3. This is consistent with our observations, because we were able to adapt Ba/F3 cells only when they were transfected with receptors that still activate JAK1 and JAK3 (IC115 or IC98, but not IC44 or shorter; Refs. 12, 13 ). Moreover, the IC98 and IC115 receptors contain only two cytoplasmic tyrosines, which are located in the box1 motif and are not surrounded by amino acids found in consensus STAT5-binding sites (12) . Collectively, these results point out an alternative mechanism of STAT5 activation by IL-9 that is independent of IL-9R phosphorylation. This mechanism appeared to be less efficient than the classical one, but it was required for the selection of Ba/F3-Phe116/9 cells.
The IL-9-adapted Ba/F3-Phe116/9 cell lines showed increased STAT5 responsiveness and were prone to cytokine-independent STAT5 activation leading to autonomously growing clones with oncogenic activity. The selection of pretransformed Ba/F3-Phe116/9 cells from normal Ba/F3-Phe116 cells was blocked by dominant negative STAT5 but was facilitated by overexpression of wild-type STAT5. Taken together, these observations indicate that quantitative or qualitative modifications in STAT5 activation may favor tumorigenesis.
Constitutive activation of STAT5, as we observed in tumorigenic clones derived from both Ba/F3-Phe116/9 and TS1 cells, has also been reported in various human hematopoietic cancer cells (22–24) . In most human tumors, the mechanism of constitutive STAT5 activation remains obscure. In Ba/F3 cells, it was associated with JAK1 phosphorylation. Activation of JAK1 and STAT5 could indicate the presence of an autocrine loop, involving the secretion of IL-9, for instance, because murine IL-9 can bind to the transfected human IL-9R. However, neither IL-9 nor a Ba/F3 growth factor activity was detected in the supernatant of autonomously growing clones, and no IL-9 mRNA was found by reverse transcription-PCR. Alternatively, various oncogenes have also been shown to activate JAK1 and STAT5, including BCR-ABL and HBx (16, 18) . Down-regulation or de-activation of JAK/STAT-inhibitory proteins, such as PIAS, SOCS, or phosphatases may also help to activate STAT5 (16) .
Several STAT5-target genes have been identified that could explain the effect of STAT5 on proliferation. Here, we focused on pim-1 because it is regulated by IL-9, and we showed that its expression correlates with proliferation. Interestingly, transgenic mice expressing IL-9 or pim-1 have similar afflictions: they suffer from an increased lymphoma incidence, particularly when mice are treated with small amounts of mutagenic compound (38) . Hence, pim-1 may be a mediator of IL-9 oncogenic properties, downstream of STAT5. However, other genes are probably involved in the regulation of proliferation by STAT5 and IL-9. A major goal for future studies will be to identify them, taking advantage of the present model.
Our data show that STAT5 and its downstream target, pim-1, are important mediators of cell growth in response to IL-9. Moreover, the mechanism of STAT5 activation may be affected during tumorigenesis, leading to a constitutive activation of this factor.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 This work was supported in part by the Belgian
Federal Service for Scientific, Technical and Cultural Affairs and by
the Actions de Recherche Concertées, Communauté
française de Belgique, Direction de la recherche scientifique.
J-B. D. is a senior research assistant, D. L. is a research
assistant, and J-C. R. is a research associate with the Fonds National
de la Recherche Scientifique, Belgium. 
2 To whom requests for reprints should be
addressed, at: Ludwig Institute for Cancer Research, avenue Hippocrate,
74, B-1200 Brussels, Belgium. Phone: 32-2-764-7465; Fax: 32-2-762-9405;
E-mail: Jean-Baptiste.Demoulin{at}bru.licr.org
3 The abbreviations used are: IL, interleukin;
IL-9R, IL-9 receptor; EMSA, electrophoretic mobility shift assay; STAT,
signal transducer and activator of transcription; JAK, Janus kinase;
GRR, IFN-response region; SCID, severe combined
immunodeficient/deficiency.
Received 12/ 3/99. Accepted 5/16/00.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
chain. Cytokines Cell. Mol. Ther. (Lond.), 4: 243-256, 1998.[Medline]
chain in interleukin-9-mediated cell proliferation and signal transduction. J. Biol. Chem, 272: 21334-21341, 1997.c chain with the interleukin-9 receptor subunit leads to STAT activation and prevention of apoptosis. J. Biol. Chem, 273: 9255-9260, 1998.
cklin E., Groner B. Cytokine receptor-independent, constitutively active variants of STAT5. J. Biol. Chem, 272: 30237-30243, 1997.This article has been cited by other articles:
|
|
 |
|
  T. G. Jones, J. Hallgren, A. Humbles, T. Burwell, F. D. Finkelman, P. Alcaide, K. F. Austen, and M. F. Gurish Antigen-Induced Increases in Pulmonary Mast Cell Progenitor Numbers Depend on IL-9 and CD1d-Restricted NKT Cells J. Immunol., October 15, 2009; 183(8): 5251 - 5260. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Elyaman, E. M. Bradshaw, C. Uyttenhove, V. Dardalhon, A. Awasthi, J. Imitola, E. Bettelli, M. Oukka, J. van Snick, J.-C. Renauld, et al. IL-9 induces differentiation of TH17 cells and enhances function of FoxP3+ natural regulatory T cells PNAS, August 4, 2009; 106(31): 12885 - 12890. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Hornakova, J. Staerk, Y. Royer, E. Flex, M. Tartaglia, S. N. Constantinescu, L. Knoops, and J.-C. Renauld Acute Lymphoblastic Leukemia-associated JAK1 Mutants Activate the Janus Kinase/STAT Pathway via Interleukin-9 Receptor {alpha} Homodimers J. Biol. Chem., March 13, 2009; 284(11): 6773 - 6781. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Nakasato, Y. Shirakura, M. Ooga, M. Iwatsuki, M. Ito, S.-i. Kageyama, S. Sakai, M. Nagata, and F. Aoki Involvement of the STAT5 Signaling Pathway in the Regulation of Mouse Preimplantation Development Biol Reprod, October 1, 2006; 75(4): 508 - 517. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Qiu, R. Lai, Q. Lin, E. Lau, D. M. Thomazy, D. Calame, R. J. Ford, L. W. Kwak, R. A. Kirken, and H. M. Amin Autocrine release of interleukin-9 promotes Jak3-dependent survival of ALK+ anaplastic large-cell lymphoma cells Blood, October 1, 2006; 108(7): 2407 - 2415. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. J. Peltola, K. Paukku, T. L. T. Aho, M. Ruuska, O. Silvennoinen, and P. J. Koskinen Pim-1 kinase inhibits STAT5-dependent transcription via its interactions with SOCS1 and SOCS3 Blood, May 15, 2004; 103(10): 3744 - 3750. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Iavnilovitch, B. Groner, and I. Barash Overexpression and Forced Activation of Stat5 in Mammary Gland of Transgenic Mice Promotes Cellular Proliferation, Enhances Differentiation, and Delays Postlactational Apoptosis Mol. Cancer Res., November 1, 2002; 1(1): 32 - 47. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Buettner, L. B. Mora, and R. Jove Activated STAT Signaling in Human Tumors Provides Novel Molecular Targets for Therapeutic Intervention Clin. Cancer Res., April 1, 2002; 8(4): 945 - 954. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Zhu, L. M. Ramirez, R. L. Lee, N. S. Magnuson, G. A. Bishop, and M. R. Gold CD40 Signaling in B Cells Regulates the Expression of the Pim-1 Kinase Via the NF-{kappa}B Pathway J. Immunol., January 15, 2002; 168(2): 744 - 754. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-B. Demoulin, J. Van Snick, and J.-C. Renauld Interleukin-9 (IL-9) Induces Cell Growth Arrest Associated with Sustained Signal Transducer and Activator of Transcription Activation in Lymphoma Cells Overexpressing the IL-9 Receptor Cell Growth Differ., March 1, 2001; 12(3): 169 - 174. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
) tetracyclin (1 µg/ml). The
number of viable cells was determined every 2 days by trypan blue
exclusion. Cells were diluted in fresh medium before confluence
(106/ml). Growth indices were calculated as the ratio
between the cell number at a given time and the initial cell number.
D, E, F,
Ba/F3-tTA-STAT5
) or 106 cells/ml
(Phe116, 
) with IL-9 (500 units/ml). Cells were
counted every 2 days and diluted before confluence. Growth indices were
calculated as in Fig. 1
) the
selection process. Cells were seeded in microtiter plates in the
presence of various concentrations of IL-9, and the extent of
proliferation was assessed after 3 days by determining hexoseaminidase
activity. Dotted line, the IL-9 concentration that gave
a 50% proliferative response. One significant experiment of three is
shown, with SEs.
), using the pEF/myc/cyto plasmid. Cells were analyzed
(left) as in Fig. 2
