VEGF autoregulates its proliferative and migratory ERK1/2 and p38 cascades
by enhancing the expression of DUSP1 and DUSP5 phosphatases
in endothelial cells
Sofia Bellou,1
Mark A. Hink,2
Eleni Bagli,1
Ekaterini Panopoulou,3
Philippe I. H. Bastiaens,2
Carol Murphy,3
and Theodore Fotsis1,3
1
Laboratory of Biological Chemistry, Medical School, University of Ioannina, Ioannina, Greece; 2
Department of Systemic
Cell Biology, Max-Planck-Institute for Molecular Physiology, Dortmund, Germany; and 3
Biomedical Research
Insitute/Foundation for Research & Technology-Hellas (BRI/FORTH), University Campus, Ioannina, Greece
Submitted 30 January 2009; accepted in final form 9 September 2009
Bellou S, Hink MA, Bagli E, Panopoulou E, Bastiaens PI,
Murphy C, Fotsis T. VEGF autoregulates its proliferative and
migratory ERK1/2 and p38 cascades by enhancing the expression of
DUSP1 and DUSP5 phosphatases in endothelial cells. Am J Physiol
Cell Physiol 297: C1477–C1489, 2009. First published September 9,
2009; doi:10.1152/ajpcell.00058.2009.—Vascular endothelial growth
factor (VEGF) is a key angiogenic factor that regulates proliferation
and migration of endothelial cells via phosphorylation of extracellular
signal-regulated kinase-1/2 (ERK1/2) and p38, respectively. Here, we
demonstrate that VEGF strongly induces the transcription of two
dual-specificity phosphatase (DUSP) genes DUSP1 and DUSP5 in
endothelial cells. Using fluorescence microscopy, fluorescence lifetime
imaging (FLIM), and fluorescence cross-correlation spectroscopy
(FCCS), we found that DUSP1/mitogen-activated protein kinases
phosphatase-1 (MKP-1) was localized in both the nucleus and
cytoplasm of endothelial cells, where it existed in complex with p38
(effective dissociation constant, KD
eff
, values of 294 and 197 nM,
respectively), whereas DUSP5 was localized in the nucleus of endothelial
cells in complex with ERK1/2 (KD
eff
345 nM). VEGF administration
affected differentially the KD
eff
values of the DUSP1/p38 and
DUSP5/ERK1/2 complexes. Gain-of-function and lack-of-function
approaches revealed that DUSP1/MKP-1 dephosphorylates primarily
VEGF-phosphorylated p38, thereby inhibiting endothelial cell migration,
whereas DUSP5 dephosphorylates VEGF-phosphorylated ERK1/2
inhibiting proliferation of endothelial cells. Moreover, DUSP5 exhibited
considerable nuclear anchoring activity on ERK1/2 in the nucleus,
thereby diminishing ERK1/2 export to the cytoplasm decreasing
its further availability for activation.
vascular endothelial growth factors; extracellular signal regulated
kinase-1/2; p38; dual-specificity phosphatase 1/mitogen-activated
protein kinase phosphatase-1; dual-specificity phosphatase-5
VASCULAR ENDOTHELIAL GROWTH FACTORS (VEGFs) are the most
important regulators of vessel morphogenesis. Not only do
they participate in the regulation of both vasculogenesis and
angiogenesis, but they also are among the most important
molecules involved in the pathogenesis of angiogenic diseases
such as diabetic retinopathy and cancer (6). VEGFs are secreted
dimeric glycoproteins of ϳ40 kDa, and in mammals the
VEGF family consists of five members: VEFG-A, B, C, D and
placental growth factor (PLGF) (6). Moreover, alternative
splicing of several of the VEGF family members gives rise to
isoforms with different biological activities. The VEGF ligands
bind in an overlapping pattern to three receptor tyrosine kinases
known as VEGF receptor-1, -2, and -3 (VEGFR1–3), as
well as to coreceptors such as heparan sulfate proteoglycans
and neuropilins (22). VEGFR2 promotes migration, proliferation,
and differentiation of endothelial cells (ECs) being critical
for the regulation of angiogenic sprouting (1). Indeed, during
sprouting, tip cells need to acquire an invasive and motile
phenotype, whereas ECs in the stalk exhibit increased proliferation.
In humans, upon VEGF-A binding, phosphorylation of
VEGFR2 on Tyr-1175 leads to recruitment of PLC-␥, which in
turn, via activation of PKC, phosphorylates mitogen-activated
protein kinase (MAPK)/extracellular signal-regulated kinase-
1/2 (ERK1/2) leading to proliferation of ECs bypassing the
classic Ras-Raf-MEK-MAPK pathway (30), although other
reports implicate also the latter (21). Phosphorylation of Tyr-
1214 of VEGFR2 activates stress-activated protein kinas-2/p38
(14) leading to VEGF-induced actin reorganization and migration
of ECs via phosphorylation of heat-shock protein-27
(HSP27) (27) and LIM-kinase 1 (LIMK1) (13). Thus VEGF
regulates EC proliferation and migration using MAPK-dependent
pathways.
As MAPKs are activated by phosphorylation on threonine
and tyrosine residues within a conserved signature sequence
TxY by a MAPK kinase (MKK or MEK) (25), dual-specificity
(Thr/Tyr) MAPK phosphatases (DUSPs/MKPs) are the largest
group of phosphatases dedicated to the regulation of MAPK
signaling in vertebrates (5). All DUSPs/MKPs contain a
COOH-terminal catalytic HCX5R domain (31) and a NH2terminal
noncatalytic domain containing two regions of sequence
similarity to the catalytic domain of the Cdc25 phosphatase
(11). The NH2-terminal MKP domain also contains the
kinase interaction motif (KIM), a cluster of basic amino acid
residues followed by an LXL site or hydrophobic amino acids
(31). The KIM domain exhibits high-affinity interactions with
an acid-rich region termed the common docking domain (CD)
in the COOH termini of all three MAPKs (31). Differences in
the amino acid sequences of the CD and KIM domains finetune
docking specificities between MAPKs and MKPs. In
addition, MAPK binding to a MKP is often, but not always,
accompanied by enzymatic activation of the COOH-terminal
catalytic domain of the phosphatase thus ensuring specificity
of action (23). Recent studies suggest essential roles for the
DUSPs/MKPs in development, immune system function,
metabolic homeostasis, and regulation of cellular stress
responses (5).
Address for reprint requests and other correspondence: T. Fotsis, Biomedical
Research Institute/Foundation for Research & Technology-Hellas (BRI/
FORTH), Univ. Campus, 45110 Ioannina, Greece (e-mail: thfotsis@uoi.gr).
Am J Physiol Cell Physiol 297: C1477–C1489, 2009.
First published September 9, 2009; doi:10.1152/ajpcell.00058.2009.
0363-6143/09 $8.00 Copyright © 2009 the American Physiological Societyhttp://www.ajpcell.org C1477
byguestonApril29,2013http://ajpcell.physiology.org/Downloadedfrom
In the present study, using DNA microarrays and qRT-PCR,
we demonstrate that, among other genes, DUSP1 and DUSP5
mRNAs are dramatically induced following VEGF administration
in ECs peaking at 30 and 60 min, respectively. Using
fluorescence microscopy, fluorescence lifetime imaging (FLIM),
and fluorescence cross-correlation spectroscopy (FCCS), we
have investigated the localization of DUSP1/MKP-1 and
DUSP5 in ECs and characterized the interaction with their
substrates p38 and ERK1/2. Finally, using gain-of-function and
lack-of-function approaches, we have elucidated the functional
role of DUSP1/MKP-1 and DUSP5 induction by VEGF in the
regulation of VEGF-induced activation of ERK1/2 and p38 and
the consequences on proliferation and migration of ECs,
thereof.
MATERIALS AND METHODS
Cell culture. Human umbilical vein ECs (HUVECs) were isolated
and cultured as previously described (3, 7). HUVECs were pooled
from at least five donors, and passage 2 to 5 was used for all
experiments. For imaging experiments, cells were cultured at 37°C
and 5% CO2 in M199 imaging medium. Human kidney embryonal
(293) cells were cultured in RPMI 1640 medium containing 10% FBS.
Bovine brain capillary ECs were cultured in DMEM medium containing
10% newborn calf serum. EC growth supplement (ECGS) was
isolated from a bovine brain as previously described (16) and was
endotoxin-free (QCL-1000, Bio-Whittaker). All cells were routinely
tested for mycoplasma. An antibody against von Willebrand factor, an
endothelial marker, was used for ensuring cell homogeneity.
Chemicals and antibodies. MEK1 and p38 inhibitor, PD98059 and
SB203580, respectively, were purchased from Sigma (Sigma, St.
Louis, MO) and dissolved in DMSO. Phosphatase inhibitors okadaic
acid and sodium orthovanadate were also from Sigma. Human
VEGF165 was purchased from ImmunoTools (Friesoythe, Germany).
The polyclonal antibodies anti-DUSP1 and anti-p38 were from Santa
Cruz Biotechnology (Santa Cruz, CA). The rabbit polyclonal antiphospho-p38
(anti-P-p38), anti-ERK1/2, and anti-phospho-ERK1/2
(anti-P-ERK1/2) were obtained from Cell Signaling (Cell Signaling
Technology, Beverley, MA). The anti-BrdU and anti-vinculin antibodies
were from Sigma, and the mouse monoclonal anti-␤-catenin
was obtained from Signal Transduction Laboratories (BD Biosciences,
Palo Alto, CA). The 9E10 (anti-Myc tag) monoclonal
antibody was purified from the corresponding hybridoma using standard
techniques. All secondary antibodies were purchased from Dianova
(Hamburg, Germany).
Constructs. Myc-tagged DUSP1wt (DUSP1wt-myc) and Myctagged
catalytically inactive DUSP1 (DUSP1/CS-myc) were kindly
provided by Steve Keyse (Cancer Research UK, Biomedical Research
Centre, Dundee, UK). Myc-tagged DUSP5wt (DUSP5wt-myc) and
Myc-tagged catalytically inactive DUSP5 (DUSP5/CS-myc) were
kindly provided by Jack Dixon (University of California, La Jolla,
CA). The pRSET-B-mCherry vector was a gift from Roger Tsien
(University of California, San Diego, CA) and was cloned into
pcDNA3 vector by Panayiotis Kouklis (University of Ioannina, Ioannina,
Greece). HA-ERK2 and HA-p38 were kindly provided by J.
Silvio Gutkind (National Institutes of Health, Bethesda, MD). All
expression vectors were from Invitrogen (Carlsbad, CA). The constructs
DUSP1-eCFP, DUSP1-eGFP, eCFP-DUSP5, eGFP-DUSP5,
p38-eYFP, p38-mCherry, eYFP-ERK2, ERK2-mCherry, eYFP-p38eCFP,
and eGFP-p38-mCherry were generated by PCR, sequenced by
MWG (Ebersberg, Germany) and purified using Endo-Free kits from
Qiagen (Qiagen) to avoid toxicity from LPS on HUVECs.
Construction of recombinant adenoviruses. DUSP1wt-myc and
DUSP1/CS-myc (AdDUSP1wt or AdDUSP1/CS) were cloned into
the EcoRV site of the expression vector pAD-CMV. DUSP5wt-myc
(AdDUSP5wt) and DUSP5/CS-myc (AdDUSP5/CS) were cloned into
the KpnI-XbaI sites of pAD-CMV. Construction of recombinant
adenoviruses followed as described (www.coloncancer.org/adeasy.
htm). The adenovirus expressing ␤-galactosidase (AdLacZ) was
kindly provided by A. Moustakas (Ludwig Institute for Cancer Research,
Uppsala, Sweden). All vectors for the generation of adenoviruses
were kindly provided by Bert Vogelstein (The John Hopkins
Medical Institutions, Baltimore, MD).
Transfection of HUVECs. HUVECs were transfected 24 h after
trypsinization with Metafectene-proreagent (Biontex, Biontex Laboratories,
Martinsried/Planegg, Germany) at a DNA-to-lipid ratio 1:4.
The complex was prepared in plain M199 medium and was incubated
for 20 min at room temperature. Transfection complex was added to
the cells in serum-reduced M199 medium (5% FBS) with no antibiotics,
heparin, nor ECGS. Transfection medium was replaced with full
HUVEC medium after 4 h.
siRNA transfection. For small interfering RNA (siRNA) experiments,
siRNA for human DUSP1 (siDUSP1: 5Ј-GCACATTCGGGACCAATATTT-3Ј),
DUSP5 (siDUSP5: GGCCTTCGATTACATCAAG),
and the control siRNA (Silencer Negative Control 5) were
provided by Ambion. HUVECs were seeded at 50% confluence and
siRNAs were transfected at a final concentration of 20 nM for 72 h
with Lipofectamine RNAiMAX reagent (Invitrogen). Scrambled
siRNA was included in each transfection as a control.
Microarray analysis. HUVECs, p6, pooled from 30 individuals
(VEC Technology), was plated on 75-cm2
dishes and cultured in
M199 medium supplemented with 20% FCS, endothelial cell growth
supplement (Sigma), heparin (Sigma), and penicillin-streptomycin.
Cells were cultured until they reached 85% confluence and were
serum starved for 6 h in M199 medium supplemented only with 5%
FCS, heparin, and penicillin-streptomycin. Subsequently cells were
induced with VEGF (30 ng/ml) for 3, 6, 12, 17, and 22 h. Total RNA
was then isolated using RNeasy Kit (Qiagen). The concentration and
the quality of the RNA were determined, and the samples were sent
for microarray analysis to VIB facility (Belgium).
Microarray results were confirmed by quantitative reverse transcription-PCR
(qRT-PCR) using The LightCycler 2.0 Instrument
(Roche Diagnostics, Mannheim, Germany) and QuantiTect SYBR
Green RT-PCR Kit (Qiagen). For qRT-PCR experiments, total RNA
isolated at the time points indicated above was used along with RNA
samples isolated after 0.5 and 1 h of VEGF induction.
Indirect immunofluorescence. HUVECs were grown on 11-mm
coverslips at 3 ϫ 104
cells per well and were treated as indicated in
RESULTS, and indirect immunofluorescence followed as previously
described (24). Samples were viewed with a Leica TCS-SP scanning
laser confocal microscope equipped with an argon laser (for excitation
at 488), solid-state 561 laser line, and helium-neon laser (for excitation
at 633). Objectives used were Leica ϫ100 PL APO 1.4 NA and
ϫ63 HCX PL APO 1.3 NA. Images were acquired using Leica
software, and files were subsequently compiled in ImageJ and Adobe
Photoshop.
Nuclear and cytoplasmic extracts. Nuclear and cytoplasmic extracts
were prepared from control and VEGF-treated cells. Briefly,
serum-starved HUVECs from 60-mm diameter dishes were induced
with VEGF for various time points. The cells were then washed with
PBS and 500 ␮l of ice-cold hypotonic buffer were added [20 mM
Tris⅐HCl, pH 7.5, 10 mM NaCl, 1 mM DTT, 0.5% Nonidet-40, 0.2
mM EDTA, 5 mM MgCl2, 10% glycerol, 2 mM sodium vanadate, 50
nM okadaic acid, 20 mM ␤-glycerol phosphate, 50 mM sodium
fluoride and protease inhibitors (Roche Diagnostics)]. The cells were
allowed to swell on ice for 5 min, scraped, and centrifuged at 500 rpm
for 5 min. The supernatant was collected as cytoplasmic fraction, and
the pellet (nuclei) was resuspended in 500 ␮l hypertonic buffer
(hypotonic buffer plus 500 mM NaCl). Equal volumes of these
extracts were loaded on SDS-polyacrylamide gels for Western blotting
with antibodies against P-ERK1/2 or ERK1/2. N-myc and Rab5
were used as nuclear and cytoplasmic markers, respectively.
C1478 VEGF ENHANCES EXPRESSION OF DUSP1 AND DUSP5
AJP-Cell Physiol • VOL 297 • DECEMBER 2009 • www.ajpcell.org
byguestonApril29,2013http://ajpcell.physiology.org/Downloadedfrom
Cell proliferation. HUVECs were seeded in 24-well plates onto
11-mm collagen-coated glass coverslips at 3 ϫ 104
cells per well, and
after overnight incubation the HUVECS were infected with adenoviruses
AdLacZ, AdDUSP1wt, AdDUSP5wt, or the catalytically inactive
mutants of either AdDUSP1/CS or AdDUSP5/CS at a multiplicity
of infection (MOI) of 200. After 24 h, cells were placed into reduced
HUVEC-medium containing 5% FBS lacking ECGS for 16 h. The
cells were then stimulated with 50 ng/ml VEGF and incubated for
24 h. For the final 6 h, cells were treated with 100 ␮M bromodeoxyuridine
(BrdU, Sigma). Cells were then fixed and processed for
indirect immunofluorescence as previously described (24).
Cell migration assay. Twenty four hours postinfection (AdLacZ
and AdDUSP1wt at MOI 200 or AdLacZ and AdDUSP5wt at MOI
300), confluent HUVEC monolayers were wounded with a sterile
plastic pipette tip, cultured in M199 medium supplemented with 5%
FCS, and induced with VEGF (10 ng/ml). Cells were placed in a
37°C, 5% CO2 chamber and monitored using a Leica DM IBRE
microscope equipped with a HRD060-NIK CCD-camera (Diagnostic
Instruments, Sterling Heights, MI) and Metamorph software. Frames
were taken every 10 min for 16 h. Results were expressed as number
of cells per centimeter of wound.
Fluorescence cross-correlation spectroscopy. FCCS was performed
using a Zeiss LSM510-Meta-Confocor 3 (Carl Zeiss Jena)
following the procedure as described previously (17). A water immersion
C-Apochromat ϫ40 objective lens (NA 1.2) was used. To
establish FCCS, a positive control expressing a fusion of eGFP-p38mCherry
was used. Average cross-correlation amplitude corresponding
to 67% complex was obtained. The failure to detect 100%
cross-correlation arose from imperfect overlap of the two confocal
detection volumes at the respective excitation wavelengths. Where
VEGF induction is indicated, cells were serum-starved for 2 h in plain
M199 before growth factor stimulation. The raw data files were
correlated and analyzed using the FCS Dataprocessor 1.5 software
(SSTC, Belarussia).
Fluorescence lifetime imaging in living cells. FLIM measurements
of enhanced cyan fluorescent protein (eCFP) were performed in living
HUVECs as described previously (17). Briefly, eCFP was excited
with a 440-nm diode (Sepia II, Picoquant, Germany). Emission light
passed a 440/530 dichroic mirror and was guided via a multimode
fiber into the detection box containing a HQ480DF40 emission filter
and a single avalanche photon detector (MPD, Italy). A water immersion
ϫ60 UPlanSApo objective lens (numerical aperature 1.2) was
used. Images of 256 ϫ 256 pixels were acquired during 4 min
corresponding to approximately 1 million detected photons. Images of
the donor fluorescence were processed using the SymPhoTime software
package (v4.2, Picoquant). The images were analyzed on a
pixel-to-pixel basis using a two-exponential fitting model including
background. Both fluorescence lifetimes, corresponding to the populations
of non- and interacting eCFP proteins, were retrieved from
analysis of the summed photon count histograms of control and
sample and fixed during analysis.
RESULTS
VEGF enhances transcription of DUSP1 and DUSP5 genes
in ECs. To identify genes that are regulated by VEGF in ECs,
we have carried out DNA microarray analysis on mRNA
collected from HUVECs at several time points post-VEGF
induction. From ϳ22,000 genes examined, VEGF increased
the expression of 116 genes, whereas it decreased the expression
of 30 genes. DUSP1 and DUSP5 were among the genes
that exhibited a statistically significant increase in their expression
at several time points post-VEGF induction. Though most
of the other DUSP genes (DUSP 2, 4, 7, 8, 9, 10, and 16) were
included in the DNA chip, none of them exhibited statistically
significant regulation by VEGF. Validation by qRT-PCR revealed
that the mRNA levels of DUSP1 and DUSP5 were
induced almost 20- and 10-fold at 30 and 60 min post-VEGF
induction, respectively (Fig. 1A), compatible with the notion
that these DUSPs are immediate early genes (8). qRT-PCR
also revealed that VEGF did not exhibit any regulatory effects
on the expression of the DUSP6 gene, the only DUSP missing
from the DNA chip, or the DUSP4 and 13 genes that were also
slightly upregulated at 3 h in the microarray analysis (Fig.
1B). These results suggest that VEGF increases the expression
of DUSP1 and DUSP5 genes, at 30 and 60 min,
respectively, in ECs.
VEGF increased significantly the level of DUSP1/MKP-1
protein after 30 min of induction in HUVECs (Supplemental
Fig. 1A). The increase in DUSP1/MKP-1 protein was, however,
much weaker compared with the strong (20-fold) VEGFinduced
upregulation of the DUSP1 gene. Because binding of
P-ERK1/2 to DUSP1/MKP-1 causes proteolytic degradation of
the phosphatase (15), we wondered whether VEGF-induced
P-ERK1/2 and P-p38 might exert such regulation. Indeed, in
cycloheximide-treated HUVECs, VEGF reduced considerably
the half-lifetime of DUSP1/MKP-1 from 4.5 to 3 h (supplemental
Fig. 1B). Unfortunately, because of the lack of a
reliable antibody against DUSP5, a similar experiment with
endogenous DUSP5 could not be performed.
DUSP1 is localized in the nucleus, cytoplasm, and on the
cell membrane of ECs. Indirect immunofluorescence revealed
that endogenous DUSP1/MKP-1 was present in both the nucleus
and cytoplasm of HUVECs (Fig. 1C), although DUSP1/
MKP-1 is considered to be an inducible nuclear protein (10).
Furthermore, in 50% of the cells, DUSP1/MKP-1 was found on
the plasma membrane colocalizing with ␤-catenin (used as
marker of the plasma membrane) (Fig. 1C, bar graph). Indeed,
DUSP1/MKP-1 was present in both the nuclear and cytoplasmic
fractions of HUVECs (Fig. 1D). In addition, overexpressed
enhanced green fluorescent protein (eGFP)-labeled
DUSP1/MKP-1 was localized in both the nucleus and cytoplasm
of ECs (Fig. 1E) confirming the results with the endogenous
protein (Fig. 1C), whereas eGFP-DUSP5 exhibited
strictly nuclear localization even at high expression levels (Fig.
1E). In agreement, siRNA knockdown of DUSP1 (supplemental
Fig. 2) leads to decreased cytoplasmic and nuclear staining
of the phosphatase (Fig. 1F). Therefore, we conclude that
DUSP1/MKP-1 is not only localized in the nucleus but is also
present in the cytoplasm and on the plasma membrane of ECs.
DUSP1 and DUSP5 interact mainly with p38 and ERK2,
respectively, and this interaction is differentially affected by
VEGF treatment in the nucleus of HUVECs. Determining the
substrate specificity of DUSPs has been under extensive investigation
during the last years. However, this effort has been
proven challenging since regulation of MAPK by DUSPs may
be cell type and stimulus specific (8). In addition, results from
in vitro assays do not always reflect physiological roles in vivo.
To reach a better understanding of these enzyme-substrate
interactions, we estimated the strength of DUSP1-p38 and
DUSP5-p38 interactions using FCCS in living HUVECs (2).
Indeed, with the use of DUSP1-eGFP and p38-mCherry constructs,
the effective dissociation constant (KD
eff
), as defined
previously (17), of the DUSP1-p38 interaction was estimated
in the cytoplasm of HUVECs (Fig. 2A), being essentially
identical in full serum (291 Ϯ 62 nM) or 2 h of serum
starvation (294 Ϯ 23 nM). The corresponding values for the
C1479VEGF ENHANCES EXPRESSION OF DUSP1 AND DUSP5
AJP-Cell Physiol • VOL 297 • DECEMBER 2009 • www.ajpcell.org
byguestonApril29,2013http://ajpcell.physiology.org/Downloadedfrom
Fig. 1. Effect of vascular endothelial growth factor (VEGF) on dual-specificity phosphorylation (DUSP) DUSP1 and DUSP5 genes and DUSP1/mitogen-activated
protein kinase phosphorylation-1 (MKP-1) localization in endothelial cells. A and B: total RNA was isolated after the indicated time points of VEGF-induction of human
umbilical vein endothelial cells (HUVECs) and expression of DUSP1-5 (A) and DUSP3-6-14 (B) was examined by qRT-PCR. Numbers present means Ϯ SD of four
independent experiments. C: For indirect immunofluorescence, anti-DUSP1/MKP-1 and anti-␤-catenin were used as primary antibodies against endogenous proteins and
FITC anti-rabbit and TRITC anti-mouse IgG as secondary antibodies. The average number of cells in which endogenous DUSP1/MKP-1 was nuclear (N), cytoplasmic
(C), and on cell membrane (M) Ϯ SD of three independent experiments is presented. Bar, 10 ␮m. Arrows indicate plasma membrane localization of DUSP1.
D: serum-starved HUVECs were stimulated with VEGF (50 ng/ml) for various time points. Cells were then subjected to cell fractionation. N-myc and Rab5 were used
as nuclear or cytoplasmic markers, respectively. A representative experiment from four independent experiments is shown. E: HUVECs were transfected with either
DUSP1-GFP or DUSP5-GFP. After 24 h, cells were fixed, mounted, and visualized by fluorescence microscopy. Lysates from transfected cells were used for immunoblotting
with anti-green fluorescent protein (GFP) as primary antibody. Bar, 10 ␮m. F: HUVECs transfected with scrambled small interfering RNA (siRNA) or siDUSP1 for 72 h were
fixed, mounted, and visualized by fluorescence microscopy. Quantitation of intensity in the cytoplasm (left) and in the nucleus (right) of more that 100 cells was carried out using
ImageJ on the raw data. Graphs present means Ϯ SD of intensity. The experiment shown is a representative one from three independent experiments.
C1480 VEGF ENHANCES EXPRESSION OF DUSP1 AND DUSP5
AJP-Cell Physiol • VOL 297 • DECEMBER 2009 • www.ajpcell.org
byguestonApril29,2013http://ajpcell.physiology.org/Downloadedfrom
nuclear DUSP1-p38 interaction were 209 Ϯ 27 in full medium
and 197 Ϯ 66 after 2 h of serum starvation (Fig. 2A). The KD
eff
of the DUSP5-eGFP-p38mCherry interaction was 1,314 Ϯ 149
nM in full medium suggesting that DUSP5 and p38 interact
very weakly, if not at all. Interestingly, VEGF treatment of
HUVECs considerably decreased the affinity of DUSP1-p38
interaction (increased the KD
eff
value from 197 Ϯ 66 to 1,505 Ϯ
256 nM) in the nucleus (Fig. 2B and supplemental Fig. 3A),
whereas it had no effect on the DUSP1-p38 interaction in the
cytoplasm (Fig. 2C and supplemental Fig. 3B).
Similar evaluation, using ERK2-mCherry, DUSP5-eGFP,
and DUSP1-eGFP constructs, revealed that the KD
eff
of DUSP5ERK2
interaction in the nucleus of HUVECs was 441 Ϯ 135
and 345 Ϯ 37 nM in full medium or 2 h after serum starvation,
respectively (Fig. 2D), whereas that of DUSP1-ERK2 in full
medium was weaker (603 Ϯ 141 nM). VEGF induction affected
the affinity of DUSP5-ERK2 interaction in the nucleus
of HUVECs in a bimodal manner. It initially increased the KD
eff
to 816 Ϯ 310 nM at 35 min after VEGF induction, thereafter
decreasing it to 150 Ϯ 38 nM at 130 min after VEGF induction
(Fig. 2E and supplemental Fig. 3C).
Next, we have used FLIM in living ECs to confirm the above
findings. In these experiments, if the eCFP-tagged phosphatase
interacted with the relevant enhanced yellow fluorescent protein
(eYFP)-tagged MAPK leading to fluorescence resonance
energy transfer (FRET), this would result in a decrease in the
donor fluorescence lifetime (34). Indeed, the fluorescence lifetime
of eCFP-DUSP1 was decreased statistically significantly from
2.53 Ϯ 0.04 ns (in presence of the eYFP vector) to 2.28 Ϯ 0.07
ns, when HUVECs were transfected with p38-eYFP (Fig. 3B); the
construct eCFP-p38-eYFP was used as a positive control (Fig.
3A). In agreement with the nuclear and cytoplasmic localization
of DUSP1, the interaction with p38 occurred in both
compartments. Similarly, the fluorescence lifetime of DUSP5eCFP
was decreased from 2.53 Ϯ 0.02 to 2.22 Ϯ 0.07 ns in the
presence of ERK2-eCFP and it was confined only to the
nucleus (Fig. 3C).
When taken together these data suggest that, in living HUVECs,
DUSP1 interacts strongly with p38, the interaction with ERK2
being weaker. DUSP5 interacts strongly only with ERK2. VEGF
administration affects the DUSP1-p38 and DUSP5-ERK2 interactions
differently in the nucleus of HUVECs, strongly suggesting
that these complexes have a different fate after VEGF induction.
The bimodal change in the KD
eff
of the DUSP5-ERK2 complex is
compatible with the anchoring effect of DUSP5 on ERK1/2 in the
nucleus of HUVECs (see below).
DUSP1 regulates VEGF-induced phosphorylation of p38 in
ECs. VEGF activation of p38 is one of the main pathways that
stimulate EC migration, a key partial step of angiogenic sprouting
(13, 14, 27). Indeed, in HUVECs, VEGF-induced p38
phosphorylation peaked at 5 min, fully returning to background
levels within 40 min (Fig. 4, A and D), whereas phosphorylaFig.
2. VEGF affects differently the DUSP1/
p38 and DUSP5/ERK1/2 interactions in the
nucleus of endothelial cells. A: HUVECs were
cultured in collagen-coated chambers until
they reached 50% confluence before transfection
with DUSP1-eGFP and mCherry vector or
DUSP1-eGFP and p38-mCherry. Fluorescence
cross-correlation spectroscopy (FCCS) measurements
were taken in 20–50 cells either in
full medium or after 2 h serum starvation.
B and C: FCCS measurements were taken after
VEGF induction (50 ng/ml) and the effective
dissociation constant (KD
eff
) values of the
DUSP1/p38 interaction at various time points
were measured in the nucleus (B) and in cytoplasm
(C) of HUVECs. D: HUVECs were
transfected with DUSP5-eGFP and mCherry
vector or DUSP5-eGFP and ERK2-mCherry.
FCCS measurements were taken in 20–50
cells in either full medium or after 2 h serum
starvation. E: FCCS measurements were taken
after VEGF induction (50 ng/ml) and the KD
eff
values of the DUSP5/ERK2 interaction at various
time points were measured in the nucleus
of HUVECs.
C1481VEGF ENHANCES EXPRESSION OF DUSP1 AND DUSP5
AJP-Cell Physiol • VOL 297 • DECEMBER 2009 • www.ajpcell.org
byguestonApril29,2013http://ajpcell.physiology.org/Downloadedfrom
tion of MEK3/6 also peaked at 5 min but retained a higher than
background level of phosphorylation throughout the experiment
(240 min) (Fig. 4A). In nonstimulated cells, p38 was
found in the nucleus, cytoplasm, and plasma membrane (colocalization
with ␤-catenin) (Fig. 4B). VEGF induction increased
the nuclear localization of phosphorylated p38 (P-p38) from
10% to 70%, (Fig. 4C,a,b and bar graph), however, in 30% of
the cells (Fig. 4C, bar graph) P-p38 exhibited a membrane-like
localization (Fig. 4C,b), which upon costaining with ␤-catenin
was clearly within the cytoplasm (Fig. 4C,c). Further characterization
revealed that P-p38 colocalized with labeled phalloidin
(Fig. 4C, bottom) suggesting localization on the actin
cytoskeleton, where it could influence actin reorganization and
migration of ECs (13).
Next, we investigated whether DUSP1/MKP-1 could influence
the kinetics of VEGF-induced phosphorylation of p38.
Indeed, overexpression of DUSP1/MKP-1 led to complete
inhibition of VEGF-induced phosphorylation of p38, whereas
overexpression of a dominant negative (catalytically inactive)
form of DUSP1 (DUSP1/CS) resulted in a dramatic increase of
both the duration and magnitude of VEGF-induced phosphorylation
of p38 (Fig. 4D). The wild-type and dominant-negative
forms of DUSP5 had no effect on VEGF-induced activation of
p38 (data not shown). The result is compatible with our FCCS
data and the known selectivity of DUSP5 toward ERK1/2 (19)
suggesting that DUSP1/MKP-1 may be the predominant DUSP
that feedbacks on VEGF-induced phosphorylation of p38.
Fig. 4. MKP-1/DUSP1 regulates VEGF-induced phosphorylation of p38 in endothelial cells. A: HUVECs were serum starved in M199 medium for 2 h. VEGF
(50 ng/ml) was added and cell lysates were used for immunoblotting. Representative experiment from four independent experiments is shown. Values represent
the fold induction at each time point compared with the noninduced sample. B: HUVECs were cultured in full medium. Indirect immunofluorescence was
performed with primary antibodies against endogenous p38 and ␤-catenin. Bottom: average number of cells in which endogenous p38 was nuclear (N),
cytoplasmic (C), and cell membrane associated (M) Ϯ SD of three independent experiments. HUVEC lysate was used for Western assay to ensure antibody
specificity. Bar, 10 ␮m. C: serum-starved HUVECs were stimulated with VEGF (50 ng/ml) for 10 min. Indirect immunofluorescence was carried out using
primary antibodies against endogenous P-p38 and ␤-catenin (top) or staining with labeled-phalloidin (bottom). Graph presents average number of cells in which
endogenous P-p38 was nuclear or cytoplasmic before or after VEGF-induction Ϯ SD of three independent experiments. Lysates from stimulated HUVECs were
used for ensuring antibody specificity. See RESULTS for details about a–c. Bar, 10 ␮m. D: HUVECs were infected with AdLacZ, AdDUSP1wt, or AdDUSP1/CS
for 24 h. They were then serum starved for 2 h and induced with VEGF (50 ng/ml) for various time points. Immunoblotting was performed with antibodies against
endogenous P-p38 or p38. Graph shows normalized intensities in arbitrary units. Bottom: infection efficiency with AdLacZ adenovirus and expression levels of
AdDUSP1wt and AdDUSP1/CS. One representative experiment, from three performed, is presented.
Fig. 3. DUSP1 and DUSP5 interact with p38 and ERK2, respectively, in living endothelial cells. HUVECs were cultured in collagen-coated chambers until they
reached 50% confluence. Then they were transfected with enhanced cyan fluorescent protein (eCFP)-p38-enhanced yellow fluorescent protein (eYFP) fusion
protein (A) or DUSP1-eCFP and eYFP vector or DUSP1-eCFP and p38-eYFP (B), or DUSP5-eCFP and eYFP vector or DUSP5-eCFP and ERK2-eYFP (C).
After 24 h, fluorescence lifetime imaging (FLIM) experiments were performed according to MATERIALS AND METHODS. The eCFP intensity images are shown,
the pseudo-colored FLIM and the plots displaying the fluorescence lifetime distribution averaged over several cells. Bar, 10 ␮m. Values present means Ϯ SD
of fluorescence lifetime. The experiment shown is a representative one from three independent experiments. *P Ͻ 0.0001.
C1482 VEGF ENHANCES EXPRESSION OF DUSP1 AND DUSP5
AJP-Cell Physiol • VOL 297 • DECEMBER 2009 • www.ajpcell.org
byguestonApril29,2013http://ajpcell.physiology.org/Downloadedfrom
DUSP5 and DUSP1 regulate VEGF-induced phosphorylation
of ERK1/2 in ECs. VEGF stimulates EC proliferation via
phosphorylation of ERK1/2 (28, 30). In HUVECs, VEGF
caused phosphorylation of ERK1/2 peaking at 10 min and
returning to background levels within 40 min (Fig. 5A). On the
other hand, VEGF-induced phosphorylation of MEK1/2, the
upstream activator of ERK1/2, reached a maximum at 5 min,
retaining phosphorylation levels above background for up to
4 h (Fig. 5A). Interestingly, upon VEGF activation, only a
small proportion of P-ERK1/2 was found in the nucleus, the
majority being localized at focal adhesions costaining with
vinculin (Fig. 5B). Use of PD98059 (an inhibitor of MEK1)
C1483VEGF ENHANCES EXPRESSION OF DUSP1 AND DUSP5
AJP-Cell Physiol • VOL 297 • DECEMBER 2009 • www.ajpcell.org
byguestonApril29,2013http://ajpcell.physiology.org/Downloadedfrom
Fig. 5. Kinetics and localization of extracellular signal-regulated kinase 1/2 (ERK1/2) phosphorylation upon VEGF induction in HUVECs. A: serum-starved
HUVECs were stimulated with VEGF (50 ng/ml) for various time points. Immunoblotting was carried out with antibodies against endogenous P-MEK1/2,
MEK1/2, P-ERK1/2, and ERK1/2. Values represent the fold induction at each time point compared with the noninduced sample. Representative experiment from
three independent experiments is shown. B: serum-starved HUVECs were stimulated with VEGF (50 ng/ml) for 5 or 10 min. Indirect immunofluorescence
followed using antibodies against endogenous P-ERK1/2 and vinculin. Bar, 10 ␮m. C: HUVECs were infected with AdLacZ or AdDUSP5wt or AdDUSP5/CS
for 24 h. Serum-starved cells were then induced with VEGF (50 ng/ml) for various time points. Immunoblotting followed using antibodies against endogenous
P-ERK1/2 and tubulin. For intensity normalization between different gels, we used a common sample as a control (Cnt). Bottom: infection efficiency with
AdLacZ and expression levels of DUSP5wt and DUSP5/CS. The experiment shown is a representative one from five independent experiments. D: HUVECs were
infected with AdLacZ, AdDUSP1wt, or AdDUSP1/CS for 24 h. Serum-starved cells were induced with VEGF (50 ng/ml) for various time points. Immunoblotting
followed using antibodies against endogenous P-ERK1/2 and tubulin. Bottom: infection efficiency with AdLacZ and expression levels of AdDUSP1wt and
AdDUSP1/CS. The experiment shown is a representative one from three independent experiments.
C1484 VEGF ENHANCES EXPRESSION OF DUSP1 AND DUSP5
AJP-Cell Physiol • VOL 297 • DECEMBER 2009 • www.ajpcell.org
byguestonApril29,2013http://ajpcell.physiology.org/Downloadedfrom
effectively reduced VEGF-induced P-ERK1/2 activation and
staining, whereas employment of phosphatase inhibitors (sodium
orthovanadate and okadaic acid) had exactly the opposite
effect (supplemental Fig. 4, A and B), verifying the specificity
of the antibody against P-ERK1/2. Cell fractionation fully
confirmed that VEGF did not exhibit any significant nuclear
localization of P-ERK1/2 (data not shown). Thus we conclude
that VEGF-induced P-ERK1/2 is found predominantly at focal
adhesions exhibiting minor nuclear staining.
Sustained activation of MEK1/2 accompanied by transient
phosphorylation of ERK1/2 suggested that dephosphorylation
of the latter may be a key factor that restricts the magnitude and
duration of VEGF-induced ERK1/2 activation. Overexpression
of DUSP5 led to a slight decrease in the duration of ERK1/2
phosphorylation without affecting its intensity (Fig. 5C),
whereas overexpression of DUSP1 resulted in efficient dephosphorylation
of P-ERK1/2 (Fig. 5D). The result is compatible
with DUSP5 being exclusively localized in the nucleus (Fig.
1E), whereas DUSP1 is localized also in the cytoplasm being
able to dephosphorylate the bulk of VEGF-induced P-ERK1/2
at focal adhesions (Fig. 5B). However, the strong effect of
DUSP1 on P-ERK1/2 is probably the result of overexpression,
and it is unlikely that at endogenous levels DUSP1 will have
such an impact. Indeed, overexpressing the dominant-negative
mutant DUSP1/CS (Fig. 5D) or siRNA silencing of DUSP1
(supplemental Fig. 5A) did not lead to increased P-ERK1/2, the
latter being actually decreased indicating that DUSP1 is not the
only phosphatase responsible for dephosphorylation of VEGFinduced
P-ERK1/2. On the other hand, overexpressing the
dominant-negative mutant DUSP5/CS (Fig. 5C) or siRNA
silencing of DUSP5 (supplemental Fig. 5B) did increase the
intensity of VEGF-induced phosphorylation of ERK1/2 albeit
only at early time points. These results indicate that DUSP5
contributes to the overall dephosphorylation of VEGF-induced
P-ERK1/2.
DUSP5 dephosphorylates and anchors ERK1/2 in the nucleus
of ECs. How could the nuclear DUSP5 contribute in
P-ERK1/2 dephosphorylation, when the latter exhibited minimal
nuclear localization being detected predominantly on focal
adhesions following VEGF induction? One explanation is that
nuclear DUSP5 concentrations (pre- and post-VEGF induction),
as well as other DUSPs, efficiently dephosphorylate
P-ERK1/2 that is translocated to the nucleus, thereby allowing
only a minimal increase of VEGF-induced nuclear P-ERK1/2
levels (Fig. 5B). Indeed, P-ERK1/2 could be detected in the
nucleus of DUSP5/CS-expressing cells (Fig. 6A, j–l), because
the catalytically inactive DUSP5/CS was unable to dephosphorylate
it. Actually, more than 95% of the HUVECs infected
with the adenovirus expressing DUSP5/CS exhibited strong
nuclear staining of P-ERK1/2 after VEGF induction compared
with only 9–10% of the noninduced cells (Fig. 6n). Moreover,
DUSP5 has been reported to function as a nuclear anchor for
ERK2 in mammalian cells (19) sequestering ERK1/2 in the
nucleus, thereby decreasing recycling of ERK1/2 to the cytoFig.
6. Effect of DUSP5 on VEGF-induced phosphorylation and localization of ERK1/2. HUVECs were infected with AdLacZ, AdDUSP5wt, or AdDUSP5/CS
for 24 h. Serum-starved cells were then induced by VEGF (50 ng/ml) for 10 min. Subsequently, indirect immunofluorescence was carried out using either
anti-P-ERK1/2 or anti-ERK1/2, or anti-myc as primary antibodies. See RESULTS for details about a–l. Bottom left: average number of the cells with nuclear
staining of ERK1/2 above threshold Ϯ SD of three independent experiments. Bottom right: mean percentage of cells with nuclear staining of P-ERK1/2 above
threshold Ϯ SD of three independent experiments. Immunoblotting using HUVEC-lysates and antibodies against P-ERK1/2 and ERK1/2 ensured antibody
specificity. Bar, 10 ␮m.
C1485VEGF ENHANCES EXPRESSION OF DUSP1 AND DUSP5
AJP-Cell Physiol • VOL 297 • DECEMBER 2009 • www.ajpcell.org
byguestonApril29,2013http://ajpcell.physiology.org/Downloadedfrom
plasm for further activation and in this way functioning as an
efficient mechanism for signal termination (33) (see also DISCUSSION).
Indeed, we have observed that infection of HUVECs
with adenoviruses expressing either DUSP5 or DUSP5/CS
resulted in a dramatic increase in the intensity of nuclear
ERK1/2 staining in more than 80% of the infected cells (Fig. 6,
a–f, m). Therefore, DUSP5-dependent ERK1/2 nuclear anchoring
could have contributed to the considerable increase of
VEGF-induced P-ERK1/2 nuclear staining following overexpression
of DUSP5/CS (Fig. 6, j–l, n), possibly decreasing the
cytoplasmic pool of ERK1/2 available for activation. Taken
together, these results suggest that DUSP5 appears to participate
in the rapid dephosphorylation of P-ERK1/2 imported to
the nucleus exhibiting also a strong nuclear anchoring activity
on ERK1/2.
Overexpression of DUSP1 and DUSP5 inhibits VEGF-induced
migration and proliferation of ECs. VEGF-induced
activation of p38 and ERK1/2s has been linked to the regulation
of migration and proliferation of ECs, respectively (22).
Indeed, VEGF induced, in a statistically significant manner,
migration of HUVECs both using the wounding (Fig. 7A, top)
and transwell (data not shown) assays, whereas SB203580, a
specific inhibitor of p38, totally abolished VEGF-induced migration
of ECs in the wounding assay (data not shown).
Infection of HUVECs with the adenovirus expressing DUSP1/
MKP-1 resulted in a dramatic inhibition of basal and VEGFinduced
EC migration compared with control cells (Fig. 7A).
The result is compatible with the decrease in VEGF-induced
p38 activation, when DUSP1/MKP-1 was overexpressed in
ECs (Fig. 4D). On the contrary, overexpression of DUSP5 did
not have any appreciable effect on VEGF-induced migration of
HUVECs in wound healing assays (Fig. 7A). Thus DUSP1/
MKP-1 is critical for dephosphorylating VEGF-induced P-p38
levels, thereby controlling VEGF-induced migration of ECs.
Next, we investigated the effect of DUSP1/MKP-1 and
DUSP5 overexpression on VEGF-induced EC proliferation. As
VEGF induces EC proliferation via activation of the ERK1/2s
(30), employment of PD98059, an inhibitor of MEK1, dramatically
decreased VEGF-induced BrdU incorporation in HUVECs
(data not shown). Infection of adenoviruses expressing either
DUSP1 or DUSP5 abolished almost completely VEGF-induced
BrdU incorporation in HUVECs compared with the
effect of the control adenovirus infection (Fig. 7, B, left and
middle). However, overexpression of DUSP5/CS resulted in
Fig. 7. Overexpression of DUSP1 and DUSP5 inhibits VEGF-induced migration and proliferation of endothelial cells. A: HUVECs were infected either with
AdLacZ or AdDUSP1wt or AdDUSP5wt for 24 h. Subsequently, cell migration assay was performed according to MATERIALS AND METHODS. The graph shows
images taken from noninduced cells at time points 0 and 340 min and from induced cells at the same time points, representing the number of the cells per
centimeter of wound. Infection efficiency for AdLacZ, AdDUSP1wt, and AdDUSP5-expression levels are indicated below. The experiment shown is a
representative one from three independent experiments. *P ϭ 0.0038, **P Ͻ 0.0001. Bar, 50 ␮m. B: HUVECs were infected with AdLacZ, AdDUSP1wt,
AdDUSP1/CS (left, *P ϭ 0.0006) or with AdLacZ, AdDUSP5wt, AdDUSP5/CS (middle, **P ϭ 0.0002) for 24 h or transfected with scrambled small interfering
RNA (siRNA) or siDUSP1 or siDUSP5 for 72 h (right, *P ϭ 0.0012, **P ϭ 0.0048). Proliferation assay and indirect immunofluorescence was performed as
indicated in MATERIALS AND METHODS. Graphs indicate percentage of BrdU-incorporated infected cells Ϯ SD derived from four independent experiments in the
case of the left and right and six independent experiments in the case of the middle.
C1486 VEGF ENHANCES EXPRESSION OF DUSP1 AND DUSP5
AJP-Cell Physiol • VOL 297 • DECEMBER 2009 • www.ajpcell.org
byguestonApril29,2013http://ajpcell.physiology.org/Downloadedfrom
overstimulation of VEGF-induced EC proliferation, whereas
DUSP1/CS did not exhibit such an effect (Fig. 7B, left and
middle). Essentially similar results were obtained using
siRNA-knockdown of DUSP1 and DUSP5 (Fig. 7B, right).
The result is remarkably similar to the effect of DUSP1/CS and
DUSP5/CS on VEGF-induced P-ERK1/2 levels (Fig. 5, C and
D), strongly suggesting that DUSP5 is important in regulating
EC proliferation. Thus DUSP5, by efficiently controlling dephosphorylation
and anchoring of VEGF-induced P-ERK1/2 in
the nucleus, plays a key role in controlling EC proliferation.
DISCUSSION
Using microarray analysis and qRT-PCR validation, we
have shown that VEGF dramatically activates transcription of
the DUSP1 and DUSP5 genes. These DUSPs specifically
dephosphorylate ERK1/2 and p38, pathways that are activated
by VEGF in ECs and regulate key angiogenic responses such
as migration and proliferation. Indeed, it has been recently
shown that VEGF induces DUSP1/MKP-1 pointing also to the
importance of DUSP1/MKP-1 in regulation of VEGF-stimulated
EC migration (12). The present study adds DUSP5 to the
VEGF-induced DUSPs in ECs. It appears, therefore, that
VEGF induction of DUSP1/MKP-1 and DUSP5 might feedback
on phosphorylated ERK1/2 and p38, thereby regulating
the intensity and duration of these cascades in ECs. JNK is not
activated by VEGF in HUVECs (27), a result that we have also
observed in our HUVEC cultures (data not shown).
VEGF induces EC migration by activating p38 (13, 14, 27).
Indeed, overexpression of DUSP1/MKP-1 in HUVECs completely
abolished the VEGF-induced p38 phosphorylation and
caused a dramatic inhibition of migration. Interestingly, the
catalytically inactive form of DUSP1/MKP-1 did not lead to
overstimulation of EC migration but rather exhibited no effect
on migration of ECs (data not shown). Perhaps cells cannot be
induced to migrate faster than a certain limit because of the
need of actin polymerization and focal adhesion rearrangements.
In the report of Kinney et al. (12), siRNA silencing of
DUSP1/MKP-1 even decreased migration of ECs, overexpression
of DUSP1/MKP-1 not being employed. It seems, therefore,
that lack-of-function approaches of DUSP1/MKP-1 do
not elicit the opposite effect of overexpression of this phosphatase,
at least as far as migration of ECs is concerned. More
work is apparently required on this issue. On the other hand,
overexpression of DUSP5 did not have any effect on VEGFinduced
phosphorylation of p38 or EC migration. These results
strongly suggest that DUSP1/MKP-1 is the key phosphatase
regulating the intensity and duration of VEGF-induced p38
phosphorylation and is in agreement with previous reports
showing selectivity of DUSP1/MKP-1 toward P-p38 (26).
Interestingly, DUSP1 gene transcription depends on p38 activation
(20), a statement that is valid also in our cell system
(supplemental Fig. 6).
In the present study, we conclusively show (using cell
fractionation, immunohistochemistry, expression of eGFPDUSP1,
FLIM, and FCCS analysis) that DUSP1/MKP-1 is
localized in both the cytoplasm and the nucleus of ECs, though
DUSP1/MKP-1 is classified as an inducible nuclear DUSP (9).
We additionally show that the cytoplasmic DUSP1/MKP-1 is
in complex with unphosphorylated p38 with KD
eff
values of
ϳ200 nM, as suggested by our FLIM and FCCS experiments.
As p38-complexed DUSP1/MKP-1 is catalytically active (29),
the pre-VEGF cytoplasmic concentration levels of DUSP-1/
MKP-1 may function as a threshold for VEGF-induced p38
phosphorylation requiring sufficient MKK3/6 activation by
VEGF to overcome the dephosphorylating activity of DUSP1/
MKP-1. Eventually, VEGF generates sufficient quantities of
P-p38 that enters the nucleus of ECs, where it binds and
activates its nuclear substrates, such as the transcription factor
MEF2c and MK2a (18). Such binding is reflected in a strong
increase of the KD
eff
(decrease in affinity) of the p38/DUSP1
interaction in FCCS, because the influxed P-p38-mCherry to
the nucleus does not exhibit comobility with DUSP1-eGFP, as
it is now bound to transcription factors or MK2a. Whereas
transcription factors regulate transcription, the P-p38/P-MK2a
heterodimer exits the nucleus (32) and phosphorylates cytoplasmic
substrates such as HSP27 and LIMK1, thereby mediating
VEGF-induced migration of ECs (27). As the levels of
DUSP1/MKP-1 are increasing in the nucleus following VEGF
induction, P-p38 will interact more frequently with DUSP1/
MKP-1 eventually leading to dephosphorylation of P-p38,
thereby ceasing activation of downstream transcription factors
and p38-dependent transcriptional regulation. Also, formation
of P-p38/P-MK2a complex will stop, thereby affecting migration
of ECs. Moreover, the increasing levels of DUSP1/MKP-1
in the cytoplasm, following VEGF administration, will probably
dephosphorylate cytoplasmic P-p38 ceasing activation of
MK2a and its downstream substrates HSP27 and LIMK1, key
mediators of VEGF-induced EC migration (13, 27). Thus both
the nuclear and cytoplasmic DUSP1/MKP-1 could contribute
to the regulation of VEGF-induced p-38 MAPK cascades
(summarized in supplemental Fig. 7A), at least in ECs.
VEGF induces proliferation of ECs via activation of ERK1/2
(22). Upon phosphorylation, P-ERK1/2 enters the nucleus,
where it is rapidly dephosphorylated (4, 33), Indeed, following
VEGF induction, we have detected surprisingly low levels of
P-ERK1/2 in the nucleus of ECs. It has been previously shown
that the phosphatases responsible for nuclear inactivation of
P-ERK1/2 are newly synthesized, show tyrosine or dual specificity,
and interact with ERK1/2 via a KIM domain (33). On
this basis, it has been suggested that the only phosphatases
fulfilling these criteria are DUSP1/MKP-1 and DUSP4/MKP-2
(33). Here we show that VEGF induces DUSP1/MKP-1 and
DUSP5, adding also the latter to the list of ERK1/2 inactivating
phosphatases that fulfill the above criteria. The interaction
between DUSP5 and ERK2 is significantly stronger than that
between MKP-1/DUSP1 and ERK2 (19), also confirmed by
our FCCS data in living ECs, suggesting that DUSP5 can
contribute to the inactivation of ERK1/2. Indeed, inhibition of
DUSP1/MKP-1 and DUSP5, by overexpressing the dominantnegative
forms DUSP1/CS and DUSP5/CS (confirmed also by
siRNA silencing) led to VEGF-induced proliferation that was
ϳ80 and 170% that of control, respectively, suggesting that
DUSP5 contributes considerably to the dephosphorylation of
P-ERK1/2 and the regulation of proliferation thereof. Interestingly,
in the adipose tissue and skeletal muscles of mkp-1Ϫ/Ϫ
mice both p38 and ERK1/2 were overactivated, whereas in the
liver only overactivation of p38 was observed, ERK1/2 activity
being slightly, but consistently reduced (35). Thus the relative
involvement of DUSP1/MKP-1 in the dephosphorylation of
ERK1/2 appears to be tissue or cell specific.
C1487VEGF ENHANCES EXPRESSION OF DUSP1 AND DUSP5
AJP-Cell Physiol • VOL 297 • DECEMBER 2009 • www.ajpcell.org
byguestonApril29,2013http://ajpcell.physiology.org/Downloadedfrom
The importance of DUSP5 in regulation of VEGF-induced
ERK1/2-dependent signaling is further intensified by the fact
that DUSP5 is a nuclear anchor for ERK2 (19). Sequestration
of ERK1/2 in a nuclear “anchoring and inactivating center,”
away from the cytoplasmic MEK “activating center,” has been
suggested to function as an efficient mechanism for signal
termination (33). VEGF induction of DUSP5 appears to generate
in ECs such an “inactivating and sequestering center.”
We have shown here that overexpression of DUSP5 and
DUSP5/CS sequesters ERK1/2 and P-ERK1/2 in the nucleus of
ECs, respectively. The dephosphorylating and sequestering
functions of DUSP5 on P-ERK1/2 is illustrated in the determination
of the KD
eff
of the DUSP5-ERK2 complex by FCCS.
After VEGF administration, the KD
eff
of the nuclear DUSP5ERK2
interaction increases as a result of P-ERK2-mCherrry
influx, which interacts with target transcription factors exhibiting
low comobility with DUSP5-eGFP. As VEGF increases
the expression of DUSP5 in the nucleus, DUSP5 dephosphorylates
and sequesters P-ERK1/2, thereby restoring the initial
KD
eff
value of the DUSP5-eGFP/ERK2-mCherry association.
Thus VEGF-induced upregulation of DUSP5 appears to contribute
by dephosphorylating P-ERK1/2 and sequestrating the
dephosphorylated ERK1/2 in the nucleus of ECs, thereby
terminating ERK1/2-dependent transcriptional regulation and
proliferation, thereof (summarized in supplemental Fig. 7B).
In conclusion, VEGF strongly induces the transcription of
two DUSP genes, DUSP1 and DUSP5, in ECs. In distinction to
the current assumption that DUSP1/MKP-1 is an inducible
nuclear protein, DUSP1/MKP-1 is localized in both the nucleus
and cytoplasm of ECs, whereas DUSP5 is a strictly
nuclear protein. Our data suggest that DUSP1/MKP-1 dephosphorylates
primarily VEGF-phosphorylated p38, thereby inhibiting
EC migration, whereas DUSP5 and DUSP1/MKP-1
dephosphorylate VEGF-phosphorylated ERK1/2 inhibiting
proliferation of ECs. Moreover, DUSP5 exhibits considerable
nuclear anchoring activity on ERK1/2, thereby diminishing
ERK1/2 export to the cytoplasm decreasing its further availability
for activation. Thus induction of these phosphatases by
VEGF serves as an autoregulatory circuit that controls activation
of ERK1/2 and p38 cascades by the same growth factor,
thereby regulating angiogenic responses of ECs, such a proliferation
and migration.
ACKNOWLEDGMENTS
We thank the confocal laser microscope facility of the University of
Ioannina for the use of the Leica TCS-SP scanning confocal microscope.
GRANTS
Part of this research was cofunded by the European Union in the framework of
the program “Heraklitos” of the “Operational Program for Education and Initial
Vocational Training” of the 3rd Community Support Framework of the Hellenic
Ministry of Education, funded by 25% from national sources and by 75% from the
European Social Fund (ESF). This study/research project received also financial
support by the European Commission under the 6th Framework Programme
(Contract No: LSHM-CT-2005-018725, PULMOTENSION).
DISCLOSURES
This publication reflects only the author(s)’s views and the European
Community is in no way liable for any use that may be made of the information
contained therein. No conflicts of interest are declared by the author(s).
REFERENCES
1. Adams RH, Alitalo K. Molecular regulation of angiogenesis and lymphangiogenesis.
Nat Rev Mol Cell Biol 8: 464–478, 2007.
2. Bacia K, Kim SA, Schwille P. Fluorescence cross-correlation spectroscopy
in living cells. Nature Methods 3: 83–89, 2006.
3. Bagli E, Stefaniotou M, Morbidelli L, Ziche M, Psillas K, Murphy C,
Fotsis T. Luteolin inhibits vascular endothelial growth factor-induced angiogenesis;
inhibition of endothelial cell survival and proliferation by targeting
phosphatidylinositol 3Ј-kinase activity. Cancer Res 64: 7936–7946, 2004.
4. Costa M, Marchi M, Cardarelli F, Roy A, Beltram F, Maffei L, Ratto
GM. Dynamic regulation of ERK2 nuclear translocation and mobility in
living cells. J Cell Sci 119: 4952–4963, 2006.
5. Dickinson RJ, Keyse SM. Diverse physiological functions for dualspecificity
MAP kinase phosphatases. J Cell Sci 119: 4607–4615, 2006.
6. Ferrara N, Gerber HP, LeCouter J. The biology of VEGF and its
receptors. Nature Med 9: 669–676, 2003.
7. Jaffe EA, Nachman RL, Becker CG, Minick CR. Culture of human
endothelial cells derived from umbilical veins. Identification by morphologic
and immunologic criteria. J Clin Invest 52: 2745–2756, 1973.
8. Jeffrey KL, Camps M, Rommel C, Mackay CR. Targeting dualspecificity
phosphatases: manipulating MAP kinase signalling and immune
responses. Nat Rev Drug Discov 6: 391–403, 2007.
9. Keyse SM. Dual-specificity MAP kinase phosphatases (MKPs) and cancer.
Cancer Met Rev 27: 253–261, 2008.
10. Keyse SM. Protein phosphatases and the regulation of mitogen-activated
protein kinase signalling. Current Opin Cell Biol 12: 186–192, 2000.
11. Keyse SM, Ginsburg M. Amino acid sequence similarity between
CL100, a dual-specificity MAP kinase phosphatase and cdc25. Trends
Biochem Sci 18: 377–378, 1993.
12. Kinney CM, Chandrasekharan UM, Mavrakis L, Dicorleto PE. VEGF
and thrombin induce MKP-1 through distinct signaling pathways: role for
MKP-1 in endothelial cell migration. Am J Physiol Cell Physiol 294:
C241–C250, 2008.
13. Kobayashi M, Nishita M, Mishima T, Ohashi K, Mizuno K. MAPKAPK-2mediated
LIM-kinase activation is critical for VEGF-induced actin remodeling
and cell migration. EMBO J 25: 713–726, 2006.
14. Lamalice L, Houle F, Jourdan G, Huot J. Phosphorylation of tyrosine
1214 on VEGFR2 is required for VEGF-induced activation of Cdc42
upstream of SAPK2/p38. Oncogene 23: 434–445, 2004.
15. Lin YW, Yang JL. Cooperation of ERK and SCFSkp2 for MKP-1
destruction provides a positive-feedback regulation of proliferating signaling.
J Biol Chem 281: 915–926, 2006.
16. Maciag T, Cerundolo J, Ilsley S, Kelley PR, Forand R. An endothelial
cell growth factor from bovine hypothalamus: identification and partial
characterization. Proc Natl Acad Sci USA 76: 5674–5678, 1979.
17. Maeder CI, Hink MA, Kinkhabwala A, Mayr R, Bastiaens PI, Knop
M. Spatial regulation of Fus3 MAP kinase activity through a reactiondiffusion
mechanism in yeast pheromone signalling. Nat Cell Biol 9:
1319–1326, 2007.
18. Maiti D, Xu Z, Duh EJ. Vascular endothelial growth factor induces
MEF2C and MEF2-dependent activity in endothelial cells. Invest Ophthalmol
Visual Sci 49: 3640–3648, 2008.
19. Mandl M, Slack DN, Keyse SM. Specific inactivation and nuclear anchoring
of extracellular signal-regulated kinase 2 by the inducible dual-specificity
protein phosphatase DUSP5. Mol Cell Biol 25: 1830–1845, 2005.
20. McMullen ME, Bryant PW, Glembotski CC, Vincent PA, Pumiglia
KM. Activation of p38 has opposing effects on the proliferation and
migration of endothelial cells. J Biol Chem 280: 20995–21003, 2005.
21. Meadows KN, Bryant P, Pumiglia K. Vascular endothelial growth factor
induction of the angiogenic phenotype requires Ras activation. J Biol
Chem 276: 49289–49298, 2001.
22. Olsson AK, Dimberg A, Kreuger J, and Claesson-Welsh L. VEGF
receptor signalling in control of vascular function. Nat Rev Mol Cell Biol
7: 359–371, 2006.
23. Owens DM, Keyse SM. Differential regulation of MAP kinase signalling
by dual-specificity protein phosphatases. Oncogene 26: 3203–3213, 2007.
24. Panopoulou E, Murphy C, Rasmussen H, Bagli E, Rofstad EK, Fotsis T.
Activin A suppresses neuroblastoma xenograft tumor growth via antimitotic
and antiangiogenic mechanisms. Cancer Res 65: 1877–1886, 2005.
25. Pearson G, Robinson F, Beers Gibson T, Xu BE, Karandikar M, Berman
K, Cobb MH. Mitogen-activated protein (MAP) kinase pathways: regulation
and physiological functions. Endocr Rev 22: 153–183, 2001.
26. Pratt PF, Bokemeyer D, Foschi M, Sorokin A, Dunn MJ. Alterations in
subcellular localization of p38 MAPK potentiates endothelin-stimulated
COX-2 expression in glomerular mesangial cells. J Biol Chem 278:
51928–51936, 2003.
C1488 VEGF ENHANCES EXPRESSION OF DUSP1 AND DUSP5
AJP-Cell Physiol • VOL 297 • DECEMBER 2009 • www.ajpcell.org
byguestonApril29,2013http://ajpcell.physiology.org/Downloadedfrom
27. Rousseau S, Houle F, Landry J, Huot J. p38 MAP kinase activation by
vascular endothelial growth factor mediates actin reorganization and cell
migration in human endothelial cells. Oncogene 15: 2169–2177, 1997.
28. Sakurai Y, Ohgimoto K, Kataoka Y, Yoshida N, Shibuya M. Essential
role of Flk-1 (VEGF receptor 2) tyrosine residue 1173 in vasculogenesis
in mice. Proc Natl Acad Sci USA 102: 1076–1081, 2005.
29. Slack DN, Seternes OM, Gabrielsen M, Keyse SM. Distinct binding
determinants for ERK2/p38alpha and JNK map kinases mediate catalytic
activation and substrate selectivity of map kinase phosphatase-1. J Biol
Chem 276: 16491–16500, 2001.
30. Takahashi T, Yamaguchi S, Chida K, Shibuya M. A single autophosphorylation
site on KDR/Flk-1 is essential for VEGF-A-dependent activation
of PLC-gamma and DNA synthesis in vascular endothelial cells.
EMBO J 20: 2768–2778, 2001.
31. Tanoue T, Nishida E. Molecular recognitions in the MAP kinase cascades.
Cell Signal 15: 455–462, 2003.
32. ter Haar E, Prabhakar P, Liu X, Lepre C. Crystal structure of the p38
alpha-MAPKAP kinase 2 heterodimer. J Biol Chem 282: 9733–9739,
2007.
33. Volmat V, Camps M, Arkinstall S, Pouyssegur J, Lenormand P. The
nucleus, a site for signal termination by sequestration and inactivation of
p42/p44 MAP kinases. J Cell Sci 114: 3433–3443, 2001.
34. Wouters FS, Verveer PJ, Bastiaens PI. Imaging biochemistry inside
cells. Trends Cell Biol 11: 203–211, 2001.
35. Wu JJ, Roth RJ, Anderson EJ, Hong EG, Lee MK, Choi CS, Neufer
PD, Shulman GI, Kim JK, Bennett AM. Mice lacking MAP kinase
phosphatase-1 have enhanced MAP kinase activity and resistance to
diet-induced obesity. Cell Metab 4: 61–73, 2006.
C1489VEGF ENHANCES EXPRESSION OF DUSP1 AND DUSP5
AJP-Cell Physiol • VOL 297 • DECEMBER 2009 • www.ajpcell.org
byguestonApril29,2013http://ajpcell.physiology.org/Downloadedfrom