Báo cáo khoa học: WD-repeat-propeller-FYVE protein, ProF, binds VAMP2 and protein kinase Cf pptx

WD-repeat-propeller-FYVE protein, ProF, binds VAMP2 and
protein kinase Cf
Thorsten Fritzius, Alexander D. Frey*, Marc Schweneker

, Daniel Mayer
à
and Karin Moelling
Institute of Medical Virology, University of Zurich, Switzerland
We have recently identified the propeller-FYVE
(domain identified in Fab1p, YOTB, VAC1p, and
EEA1) protein (ProF) as a binding partner for Akt and
protein kinase (PK)Cf [1]. ProF contains seven WD
repeats, which form a b-propeller-like structure, provi-
ding a protein-binding platform [2]. Furthermore, ProF
harbors a FYVE domain that specifically interacts with
phosphatidylinositol-3-phosphate [3] and targets ProF
to internal vesicles. Deletion of the FYVE domain or
inhibition of phosphatidylinositol-3-phosphate forma-
tion by a phosphoinositide-3-kinase inhibitor resulted
in loss of vesicular localization. ProF preferentially
bound to the kinases Akt and PKCf upon hormonal
stimulation of the cells with insulin-like growth factor 1
(IGF-1) [1]. Because of this stimulation-dependent bind-
ing to kinases and due to its vesicular localization and
its broad tissue distribution, we suggested that ProF
plays a role in a number of secretory pathways [1].
In order to better understand the role of ProF in
inducible vesicle trafficking, we searched for substrates
Keywords
protein interaction; protein kinase Cf;
VAMP2; vesicle transport; WD repeats

Correspondence
K. Moelling, Institute of Medical Virology,
University of Zurich, Gloriastrasse 30, Zurich
CH-8006, Switzerland
Fax: +41 44 6344967
Tel: +41 44 6342652
E-mail:
Website: />Present address
*Institute of Microbiology, ETH Zurich,
Switzerland
Gladstone Institute of Virology and
Immunology,San Francisco,CA,USA
àDepartment of Virology, Institute for
Medical Microbiology and Hygiene,
University of Freiburg, Germany
(Received 20 December 2006, accepted 16
January 2007)
doi:10.1111/j.1742-4658.2007.05702.x
We have recently identified a protein, consisting of seven WD repeats, pre-
sumably forming a b-propeller, and a domain identified in Fab1p, YOTB,
VAC1p, and EEA1 (FYVE) domain, ProF. The FYVE domain targets
the protein to vesicular membranes, while the WD repeats allow binding of
the activated kinases Akt and protein kinase (PK)Cf. Here, we describe the
vesicle-associated membrane protein 2 (VAMP2) as interaction partner of
ProF. The interaction is demonstrated with overexpressed and endogenous
proteins in mammalian cells. ProF and VAMP2 partially colocalize on
vesicular structures with PKCf and the proteins form a ternary complex.
VAMP2 can be phosphorylated by activated PKCf in vitro and the pres-
ence of ProF increases the PKCf-dependent phosphorylation of VAMP2
in vitro. ProF is an adaptor protein that brings together a kinase with its

substrate. VAMP2 is known to regulate docking and fusion of vesicles and
to play a role in targeting vesicles to the plasma membrane. The complex
may be involved in vesicle cycling in various secretory pathways.
Abbreviations
EGF-1, epidermal growth factor 1; FYVE, domain identified in Fab1p, YOTB, VAC1p, and EEA1; GLUT4, glucose transporter type 4; GST,
glutathione S-transferase; HA, hemagglutinin; Hrs, hepatocyte growth factor-related tyrosine kinase substrate; IGF-1, insulin-like growth
factor 1; MBP, myelin basic protein; PK, protein kinase; ProF, propeller FYVE protein; P-VAMP2, phosphorylated VAMP2; SNAP,
synaptosomal-associated protein; SNARE, soluble N-ethylmaleimide-sensitive fusion protein-attachment protein receptor; t-SNARE, target-
SNARE; VAMP2, vesicle-associated membrane protein 2; Vps4, vacuolar protein sorting-associating protein 4; v-SNAREs, vesicular-SNARE.
1552 FEBS Journal 274 (2007) 1552–1566 ª 2007 The Authors Journal compilation ª 2007 FEBS
of Akt and PKCf on vesicles. While this work was in
progress, the Akt substrate of 160 kDa has been found
to be located in adipocytes on vesicles containing the
glucose transporter 4 (GLUT4) [4]. The Akt substrate
of 160 kDa affects the trafficking of GLUT4-contain-
ing vesicles to the plasma membrane upon Akt phos-
phorylation [5–7]. Several PKCf substrates on vesicles
have been described previously. The vesicle-associated
membrane protein 2 (VAMP2) may be one of them
[8]. VAMP2 belongs to the vesicular soluble N-ethyl-
maleimide-sensitive fusion protein-attachment protein
receptors (v-SNARE). This protein family comprises
eight members involved in secretory pathways [9].
VAMP2 is widely expressed in a large variety of tis-
sues, such as brain, kidney, adrenal gland, liver and
pancreas [10]. VAMP2 is crucial for stimulus-depend-
ent secretion in various cell-types including insulin-
stimulated GLUT4 translocation in adipocytes and
muscle cells [11], fusion of early and sorting endosomes
[12,13], and synaptic vesicle fusion with the plasma

membrane in neurons [14,15]. The fusion of VAMP2-
containing vesicles with the plasma membrane is medi-
ated by complex formation of the v-SNARE with the
target (t)-SNARE synaptosome-associated protein
(SNAP) and syntaxin [16]. VAMP2 has previously
been reported to be phosphorylated in myotubes over-
expressing PKCf, which correlated with increased
GLUT4 translocation and glucose uptake [8].
Previous analyses have shown that ProF binds spe-
cifically to the atypical PKC isoform PKCf [1]. In the
present study, we show that ProF also interacts with
VAMP2 both in vitro and in vivo. We demonstrate that
all three proteins can form a complex and that ProF
can mediate the binding of PKCf to VAMP2 in a con-
centration-dependent manner. Furthermore, we show
that VAMP2 is directly phosphorylated by activated
PKCf in vitro and that ProF leads to increased phos-
phorylation of VAMP2 by activated PKCf in vitro.
Thus, ProF can integrate the kinase PKCf, and its
substrate VAMP2, which, upon phosphorylation, may
contribute to vesicle cycling in secretory pathways.
Results
VAMP2 is a binding partner of ProF
We have recently identified a protein, consisting of
seven WD repeats, presumably folding into a b-propel-
ler-type structure, and a FYVE domain (Fig. 1A), des-
ignated as ProF. ProF interacted via its WD repeats
with the serine ⁄ threonine kinases Akt and PKCf, and
was located on internal vesicles via its FYVE domain.
These two kinases preferentially bound to ProF after

hormonal stimulation of the cell [1]. Therefore, the
question arose whether ProF can bring together the
kinase with putative kinase substrates.
In order to identify such candidate substrates, we
performed a yeast two-hybrid screen using a human
B-cell-specific embryonic cDNA library and full-length
ProF as bait [17]. Out of this screen, two positive
clones were obtained. One of them was identified as
VAMP2, the other as an as yet not described protein.
VAMP2 is a v-SNARE protein, associated with vesicu-
lar membranes via its C-terminal transmembrane
domain. Its central SNARE domain of 60 amino
acids allows the interaction of VAMP2 with its cog-
nate t-SNARE proteins (Fig. 1A) [9]. The VAMP2
fragment, which interacted with the full-length ProF in
the yeast two-hybrid assay, contained the amino acids
1–111. Because the complete human VAMP2 protein
consists of only 116 amino acids, little information
about the ProF interaction domain of VAMP2 could
be deduced from the yeast two-hybrid screen. All
serine (Ser) residues of VAMP2 are indicated in
Fig. 1(A). Four of them were mutated to alanine
(Ser to Ala) in order to generate a VAMP2 mutant
mt(1–4).
To verify the results obtained by the yeast two-
hybrid screen, we first analyzed the interaction of
VAMP2 with ProF by coimmunoprecipitation of over-
expressed proteins. For this purpose, Myc-tagged
ProF- and Flag-tagged VAMP2-expression constructs
were cotransfected in COS-7 cells. Cell lysates were

treated with an anti-Flag IgG (Fig. 1B, left) or an
anti-Myc IgG (Fig. 1B, right) and the precipitates were
analyzed by immunoblotting for the presence of copre-
cipitating proteins. As can be seen, Myc-tagged ProF
indeed coprecipitated with Flag-tagged VAMP2
(Fig. 1B, left, lane 3) and Flag-tagged VAMP2 copre-
cipitated with Myc-tagged ProF (Fig. 1B, right,
lane 3). VAMP2 did not interact with the hepatocyte
growth factor-regulated tyrosine kinase substrate,
hepatocyte growth factor-regulated tyrosine kinase
substrate (Hrs) (Fig. 1B, left, lane 4), another FYVE-
domain containing protein that is also localized to
intracellular vesicles [18,19], or with the vacuolar pro-
tein sorting-associating protein 4, Vps4 (Fig. 1B, left,
lane 7), a protein involved in intracellular vesicle for-
mation and protein trafficking [20]. ProF interacted
weakly with Hrs (Fig. 1B, right, lane 5), possibly via
heterodimerization of the FYVE domains of ProF and
hepatocyte growth factor-regulated tyrosine kinase
substrate, as ProF can form oligomers via its FYVE
domain [1]. Furthermore, ProF did not interact with
Vps4 (Fig. 1B, right, lane 6), demonstrating the specif-
icity of the interaction of ProF with VAMP2.
T. Fritzius et al. WD-FYVE protein interacts with VAMP2 and PKCf
FEBS Journal 274 (2007) 1552–1566 ª 2007 The Authors Journal compilation ª 2007 FEBS 1553
We further characterized the interaction between
ProF and VAMP2 with deletion mutants of ProF. We
analyzed a Myc-tagged mutant of ProF, lacking the
FYVE domain (ProFDFYVE) and two mutants lack-
ing the FYVE domain and containing only blades 1–3

(ProF 1–3) or blades 4–7 (ProF 4–7) of the
seven-bladed b-propeller (Fig. 1A). Interaction of trun-
cated WD-repeat proteins, containing only one or two
b-propeller blades, with its binding partners has been
shown earlier ([21,22]), indicating that the remaining
blades are still able to fold properly. We coexpressed
these proteins together with Flag-tagged VAMP2 in
Fig. 1. VAMP2, a new interaction partner of ProF. (A) Domain structure of ProF and VAMP2. ProF consists of seven WD repeats (WD1–7),
binding to proteins, and a FYVE domain, binding to phosphatidylinositol-3-phosphate on vesicular membranes (top, left). A model of the
three-dimensional structure of ProF without the FYVE domain is shown, with the seven WD repeats indicated (top, right). VAMP2 is
anchored to vesicular membranes through its C-terminal transmembrane domain. A central SNARE motif is essential for the interaction with
its target SNARE proteins. Serine (S) residues, which are potential PKCf-phosphorylation sites are indicated below as wild-type (wt) and
mutant [mt(1–4)] (bottom). (B) Coimmunoprecipitation assay of overexpressed Flag-tagged VAMP2 (lanes 1, 3–4, and 7) in the presence or
absence of Myc-tagged ProF (lanes 2–3 and 5–6), HA-tagged Hrs (lanes 4–5), and GFP-tagged Vps4 (lanes 6–7) in COS-7 cells. Immunopre-
cipitation was performed with an antibody against the Flag-tag (left) or the Myc-tag (right) and immunoprecipitates were analyzed by immu-
noblotting with antibodies against Flag, Myc, HA and GFP epitopes. Direct lysates (DL) are shown as expression control (bottom). (C) Myc-
ProF wild-type (wt), ProF lacking the FYVE domain (Myc-ProFDFYVE), ProF lacking the FYVE domain and containing only blades 1–3 or 4–7
(Myc-ProF 1–3 or Myc-ProF 4–7) were overexpressed together with Flag-VAMP2 in COS-7 cells. Immunoprecipitation and subsequent immu-
noblot show the interactions.
WD-FYVE protein interacts with VAMP2 and PKCf T. Fritzius et al.
1554 FEBS Journal 274 (2007) 1552–1566 ª 2007 The Authors Journal compilation ª 2007 FEBS
COS-7 cells and tested their interaction by coimmuno-
precipitation assays. As can be seen, all ProF mutants
interacted equally well with VAMP2 (Fig. 1C). In sum-
mary, this result suggested that multiple binding sites
on ProF are involved in the binding of VAMP2.
VAMP2, ProF, and PKCf colocalize on vesicular
structures
Further indications for the interaction of both proteins
were obtained by confocal immunofluorescence analy-

sis showing their subcellular distribution. For that,
Flag-tagged VAMP2 and Myc-tagged ProF were coex-
pressed in COS-7 cells and analyzed by confocal micro-
scopy. As can be seen, a partial colocalization of
VAMP2 (green signal) and ProF (red signal) on vesicu-
lar structures was detectable (Fig. 2A). Colocalization
of the two proteins is indicated by the orange color,
detectable in the merged picture, showing the super-
position of the two signals.
We have previously reported that ProF binds PKCf
[1] and show here the interaction between ProF and
VAMP2. This raised the question whether ProF could
interact with both proteins, VAMP2 and PKCf. First,
this was tested by colocalization analysis using confo-
cal microscopy. Flag-tagged VAMP2, hemagglutinin
(HA)-tagged PKCf and Myc-tagged ProF were transi-
ently expressed either together (Fig. 2B) or alone
(Fig. 2C) in COS-7 cells and analyzed by confocal
microscopy. When PKCf, VAMP2, and ProF were
Fig. 2. Colocalization of VAMP2, ProF, and
PKCf. (A) Flag-VAMP2 and Myc-ProF were
overexpressed in COS-7 cells. Confocal
microscopy analysis with antibodies against
Flag- and Myc-tag revealed areas of colocali-
zation as visualized in yellow on the merged
picture (right). (B) COS-7 cells were transi-
ently transfected with HA-PKCf (green),
Flag-VAMP2 (red), and Myc-ProF (blue) and
analyzed by confocal microscopy using anti-
bodies against HA, Flag, and Myc epitopes.

The lower part shows detail. A partial colo-
calization on cytoplasmic punctuate struc-
tures (white) is observed (merged). (C) As a
control, COS-7 cells were transfected with
either HA-PKCf (green), Flag-VAMP2 (red),
or Myc-ProF (blue) and were analyzed by
confocal microscopy. Flag-VAMP2 (red) and
Myc-ProF (blue), signals are confined to spe-
cific structures, while HA-PKCf (green) and
is distributed throughout the cell. (D) Confo-
cal microscopy analysis of COS-7 cells
cotransfected with HA-PKCf (red) and
Myc-ProF (green) revealing colocalization of
both proteins on punctuate structures (top).
Confocal microscopy analysis of COS-7 cells
cotransfected with HA-PKCf (green) and
Flag-VAMP2 (red) show colocalization of
both proteins on punctuate structures
(bottom). (E) For confocal immunofluores-
cence analysis of 3T3-L1 pre-adipocytes,
cells were serum-starved for 2 h prior to
staining for endogenous PKCf (green),
endogenous VAMP2 (red), and stably
expressed Myc-tagged ProF (magenta). A
partial colocalization on cytoplasmic punctu-
ate structures (white) is observed (merged).
Lower part shows detail.
T. Fritzius et al. WD-FYVE protein interacts with VAMP2 and PKCf
FEBS Journal 274 (2007) 1552–1566 ª 2007 The Authors Journal compilation ª 2007 FEBS 1555
expressed together, a partial colocalization of the three

proteins on intracellular vesicles was detected (Fig. 2B,
right). Colocalization of the three proteins is indicated
by the white color in the merged picture, showing the
superposition of the three signals. When expressed
alone, overexpressed VAMP2 (red) and ProF (blue)
were found to be located on distinct intracellular vesi-
cles, while PKCf (green) was more evenly distributed
in the cytoplasm (Fig. 2C). These data indicate that
VAMP2 and ProF can alter the subcellular localization
of PKCf. In order to find out whether ProF and
VAMP2 alone were also able to target PKCf to vesi-
cles, HA-PKCf and Myc-ProF (Fig. 2D, top) or
HA-PKCf and Flag-VAMP2 (Fig. 2D, bottom) were
transiently coexpressed in COS-7 cells and analyzed by
confocal microscopy. As can be seen, expression of
HA-PKCf together with Myc-ProF (Fig. 2D, top) and
Flag-VAMP2 (Fig. 2D, bottom), led to a partial local-
ization of the kinase on punctuate structures. We fur-
ther substantiated these data by confocal microscopy
studies with 3T3-L1 pre-adipocyte cells, stably expres-
sing Myc-tagged ProF. As can be seen, Myc-ProF,
endogenous VAMP2, and endogenous PKCf colocal-
ized on perinuclear vesicular structures (Fig. 2E).
These results are in agreement with the previous find-
ings that ProF is located on internal vesicles in various
cell lines, e.g. in 3T3-L1 pre-adipocyte cells [1].
VAMP2, ProF, and PKCf form a complex
We have shown earlier that ProF interacts specifically
with the atypical PKC isoform PKCf, but not with
novel PKC isoforms, and binds weakly to the classical

PKC isoform PKCa [1]. To investigate whether there
was also a specific interaction between PKCf and
VAMP2, human embryonic kidney 293T cells were
transiently transfected with constructs expressing Flag-
VAMP2 and VAMP2 was immunoprecipitated with an
antibody against Flag. Overexpressed Flag-VAMP2
coprecipitated endogenous aptyical PKCf ⁄ k but not
the novel isoforms PKCd and PKCe or the classical
isoform PKCa (Fig. 3A), suggesting a specificity of
VAMP2 for atypical PKC isoforms.
Next, we analyzed whether these three proteins
physically interacted by performing a sequential
precipitation procedure. We overexpressed the
epitope-tagged forms of all three proteins in human
embryonic kidney 293T cells. Then we immunoprecip-
itated Myc-ProF and showed coprecipitation of Flag-
VAMP2 and HA-PKCf by western blot analysis of
an aliquot of the immunoprecipitate (Fig. 3B, lane 2).
The precipitated complex was thereafter eluted with a
Myc-peptide and half of the eluate was used for
immunoprecipitation of HA-PKCf. As can be seen,
the coimmunoprecipitation of Myc-ProF and Flag-
VAMP2 was demonstrated by western blotting
(Fig. 3B, lane 3). Furthermore, immunoprecipitation
of Flag-VAMP2 using the other half of the lysate led
to the coimmunoprecipitation of HA-PKCf and Myc-
ProF (Fig. 3B, lane 4). Additionally, reciprocal immu-
noprecipitation was performed to verify the existence
of a complex. Immunoprecipitation of HA-PKCf
allowed the coprecipitation of Flag-VAMP2 and

Myc-ProF, as evidenced by western blotting analysis
(Fig. 3C, lane 2). The precipitated complex was there-
after eluted with a PKCf-peptide and Myc-ProF
(Fig. 3C, lane 3) or Flag-VAMP2 was immunoprecipi-
tated (Fig. 3C, lane 4). A strong coimmunoprecipita-
tion of HA-PKCf and a weak coimmunoprecipitation
of Flag-VAMP2 in the case of Myc-ProF (Fig. 3C,
lane 3) can be demonstrated by western blotting. Fur-
thermore, we found coimmunoprecipitation of HA-
PKCf and Myc-ProF in the case of Flag-VAMP2
(Fig. 3C, lane 4). As only a weak VAMP2 signal was
detected after immunoprecipitation of Myc-ProF, this
could indicate that ProF might bind more strongly to
PKCf than to VAMP2, or that the VAMP2 binding
to the protein complex is more susceptible to the
mechanical disruptions performed during the elution
of the complex. Nevertheless, we have shown that
Myc-ProF forms a complex with both proteins, Flag-
VAMP2 and HA-PKCf.
These findings raised the question how ProF would
affect the interaction between VAMP2 and PKCf.In
order to test this, we expressed Flag-VAMP2 and
HA-PKCf in the absence or presence of increasing
amounts of Myc-ProF in COS-7 cells (Fig. 3D). As
can be seen, in the absence of Myc-ProF, only small
amounts of HA-PKCf were coprecipitated (Fig. 3D,
lane 5). Coexpression of small amounts of Myc-ProF
led to the coprecipitation of large amounts of HA-
PKCf by Flag-VAMP2 (Fig. 3D, lane 6). Further
increasing concentrations of ProF caused the opposite

effect, a decreased coprecipitation of HA-PKCf by
Flag-VAMP2 (Fig. 3D, lanes 7 and 8). This is further
corroborated by a quantification of three individual
experiments, performed by densitometric scanning of
western blots (Fig. 3D, bottom), demonstrating that
low concentrations of ProF lead to a strongly (3.5-
fold) and significantly increased binding of VAMP2 to
PKCf.
In summary, these results indicate that ProF can
regulate the binding of VAMP2 and PKCf in an adap-
tor protein-like fashion. It increases the binding of
PKCf to VAMP2 under optimized conditions, which
is a characteristic of adaptor proteins.
WD-FYVE protein interacts with VAMP2 and PKCf T. Fritzius et al.
1556 FEBS Journal 274 (2007) 1552–1566 ª 2007 The Authors Journal compilation ª 2007 FEBS
Fig. 3. Interactions of VAMP2 with ProF and PKCf . (A) Flag-VAMP2 was transiently expressed in human embryonic kidney 293T cells. The
complex comprising Flag-VAMP2 and endogenous PKC isoforms was subjected to immunoprecipitation using an antibody against Flag. Inter-
action of VAMP2 with the PKC isoforms was analyzed as shown by immunoblot against PKCa, PKCd, PCKe, and PKCf (from left to right,
top) and VAMP2 (bottom). Direct lysates show expression controls. (B) HA-PKCf, Flag-VAMP2, and Myc-ProF were transiently coexpressed
in human embryonic kidney 293T cells. The complex comprising HA-PKCf, Myc-ProF, and Flag-VAMP2 was subjected to immunoprecipita-
tion with an anti-Myc IgG (lane 2). The complex was eluted by addition of an excess of a competing Myc-peptide followed by immunoprecip-
itation using an antibody directed against PKCf (lane 3) or the Flag-epitope (lane 4). Immunoprecipitations of the different steps were
analyzed by immunoblot against the indicated proteins. Samples were loaded onto one gel, separating lines were included later for clarity.
(C) The complex comprising HA-PKCf, Myc-ProF, and Flag-VAMP2 was subjected to immunoprecipitation with an anti-PKCf IgG (lane 2). The
complex was eluted by addition of excess of competing PKCf-peptide followed by immunoprecipitation against the Myc- (lane 3) or the Flag-
epitope (lane 4). Immunoprecipitations of the different steps were analyzed by immunoblot against the indicated proteins. Samples were
loaded onto one gel; separating lines were included for clarity. (D) HA-PKCf, Myc-ProF, and Flag-VAMP2 were transiently overexpressed in
COS-7 cells. The interactions of VAMP2 with ProF alone (lane 4), PKCf alone (lane 5), or PKCf in the presence of increasing amounts
(0.25 lg, 1 lg, and 4 lg) of Myc-ProF (lanes 6–9) were analyzed as shown by immunoprecipitation and subsequent immunoblot (top). DL
(direct lysates) show expression controls (bottom). To the right, a quantification of PKCf binding to VAMP2 in the absence or presence of

ProF is shown. Results were obtained by densitometric scanning of immunoblot bands from three independent experiments. The interaction
of PKCf with VAMP2 was normalized to binding in the absence of ProF ¼ 1. Values represent mean±
SD of three separate experiments
(*P < 0.05, **P < 0.01).
T. Fritzius et al. WD-FYVE protein interacts with VAMP2 and PKCf
FEBS Journal 274 (2007) 1552–1566 ª 2007 The Authors Journal compilation ª 2007 FEBS 1557
Endogenous VAMP2 interacts with ProF and
PKCf
So far we have analyzed overexpressed proteins; we
now need to confirm these results with endogenous
proteins. ProF and VAMP2 have been reported to be
expressed in the brain [1,23]. Therefore, we used mouse
brain lysates to test the interaction between endo-
genous ProF and VAMP2. For this, brain lysates were
treated with an anti-ProF IgG with and without
peptide competition to demonstrate the specificity
of the reaction. The precipitates were analyzed by
western blotting. As can be seen, coimmunoprecipi-
tation of VAMP2 with ProF was detectable, while
the presence of a competing peptide inhibited ProF
precipitation and VAMP2 coprecipitation (Fig. 4A).
To verify the interaction of ProF and VAMP2, we per-
formed a reciprocal immunoprecipitation by treatment
of mouse brain lysates with an anti-VAMP2 IgG
(Fig. 4B, lane 2). As can be seen, coprecipitation of
ProF with VAMP2 was detectable. Peptide competi-
tion during western blotting (Fig. 4B, lane 3) and
immunoprepitation performed with an irrelevant anti-
Myc IgG (Fig. 4B, lane 1) demonstrated the specificity
of the reaction. Thus we also confirmed the interaction

of ProF and VAMP2 for endogenous proteins in brain
tissue.
In order to show the interaction of all three endo-
genous proteins, we immunoprecipitated ProF from
mouse brain lysates and analyzed the precipitates by
western blotting for the presence of ProF, VAMP2,
and PKCf. As can be seen, after precipitation with
Fig. 4. Interactions of endogenous VAMP2 with ProF and PKCf in
brain. (A) Immunoprecipitation of murine brain lysates, performed
with an anti-ProF IgG in the absence or presence of an excess of
competing peptide, used before as an antigen to raise the anti-ProF
IgG, and subsequent immunoblot. (B) Immunoprecipitation of
murine brain lysates, performed with an irrelevant antibody (anti-
Myc, lane 1), and anti-VAMP2 IgG (lane 2 and 3). The subsequent
immunoblot was performed with anti-ProF IgG in the absence
(lane 2) or presence (lane 3) of an excess of competing peptide.
Direct lysates (right) show expression of the endogenous proteins
in mouse brain lysate. (C) Immunoprecipitation of murine brain
lysates with an anti-ProF IgG and subsequent immunoblot (left).
Direct lysates (right) show expression of the endogenous proteins
in mouse brain lysate.
Fig. 5. VAMP2 is phosphorylated by PKCf in vitro. (A) Recombinant GST-VAMP2 wild-type (wt) and VAMP2 serine to alanine mutant [mt(1–
4)] were expressed in bacteria, purified and subjected to an in vitro kinase assay with c-
32
P-ATP in the presence of 3 lg of recombinant act-
ive Akt (lane 2), 100 ng recombinant active PKCf (lanes 3 and 5) and 200 ng recombinant active PKCf (lanes 4 and 6). GST-VAMP2 wt and
mt(1–4) phosphorylation and Akt ⁄ PKCf autophosphorylation were analyzed using a PhosphoImager. Expression of GST-VAMP2 wt and
mt(1–4) was analyzed by immunoblot as indicated. One representative of three independent experiments is shown (top). A quantification of
VAMP2 phosphorylation (P-VAMP2) in the presence of PKCf is shown (bottom, left). Results were obtained by densitometric scanning of
PhosphoImager signal bands from three independent experiments. The PhosphoImager signal of P-VAMP2 was normalized to GST-VAMP2

substrate phosphorylation in the presence of 100 ng recombinant active PKCf ¼ 1. PKC f activity was verified by addition of the PKCf sub-
strate MBP (bottom, right). (B) Flag-VAMP2 was overexpressed in COS-7 cells and subjected to immunoprecipitation with an antibody
against the Flag-epitope. Immunoprecipitations were subjected to in vitro kinase assay with c-
32
P-ATP without (lane 1) or with (lane 2–7)
increasing amounts of active PKCf (10 ng, 50 ng and 200 ng). PKCf inhibitor (100 l
M) was added in lanes 5–7 as indicated. VAMP2 phos-
phorylation and PKCf autophosphorylation were analyzed by PhosphoImager. Immunoprecipitation and immunoblot were performed as indi-
cated. A quantification of VAMP2 phosphorylation in the presence of PKCf is shown at the bottom. Results were obtained by densitometric
scanning of PhosphoImager signal bands from three independent experiments. The PhosphoImager signal of P-VAMP2 was normalized to
Flag-VAMP2 substrate phosphorylation in the presence of 200 ng recombinant active PKCf ¼ 1. (C) COS-7 cells, transiently overexpressing
HA-PKCf, were left either unstimulated or stimulated for 10 min with 100 ngÆ mL
)1
EGF-1, 100 nM insulin or 100 ngÆmL
)1
IGF-1, as indicated
before lysis. Lysates were subjected to immunoprecipitation with an antibody against PKCf. Immunoprecipitations were subjected to in vitro
kinase assay with c-
32
P-ATP in the presence of GST-VAMP2. VAMP2 phosphorylation and PKCf autophosphorylation were analyzed by Phos-
phoImager (top). Immunoprecipitation and immunoblot were performed as indicated. A quantification of VAMP2 phosphorylation in the pres-
ence of PKCf is shown (bottom). The PhosphoImager signal of P-VAMP2 was normalized to GST-VAMP2 substrate phosphorylation in the
presence of unstimulated HA-PKCf ¼ 1. Results were obtained by densitometric scanning of PhosphoImager signal bands from three inde-
pendent experiments (**P < 0.01). PKCf activity was verified by addition of MBP to immunoprecipitated PKCf (right).
WD-FYVE protein interacts with VAMP2 and PKCf T. Fritzius et al.
1558 FEBS Journal 274 (2007) 1552–1566 ª 2007 The Authors Journal compilation ª 2007 FEBS
anti-ProF IgG all three proteins were present in the
immunoprecipitates (Fig. 4C). The specificities of the
three antibodies used in this study were confirmed by
western blotting, which did not lead to any significant

unspecific detection of unrelated proteins (data not
shown). This result also supports the role of ProF as
interaction partner for VAMP2 and PKCf for the
endogenous proteins in brain tissue.
VAMP2 is a substrate of PKCf
Next, we wanted to find out whether VAMP2 is a sub-
strate of activated PKCf. To investigate whether
VAMP2 is directly phosphorylated by active PKCf,we
generated glutathione S-transferase (GST)-tagged
VAMP2 for expression in bacteria. Furthermore, we
generated a mutant of GST-VAMP2, in which several
serine residues were mutated to alanine. Out of the six
serine residues conserved in mouse and rat VAMP2
(Ser2, Ser28, Ser61, Ser75, Ser80, Ser115) (Fig. 1A),
we excluded Ser2 from mutation because of its posi-
tion at the very N-terminus and Ser115, because of its
C-terminal position and its location inside the vesicle,
which seemed to be an unlikely target for phosphoryla-
tion. The four remaining serine residues were mutated
together to alanine [VAMP2 mt(1–4); Fig. 1A]. Three
T. Fritzius et al. WD-FYVE protein interacts with VAMP2 and PKCf
FEBS Journal 274 (2007) 1552–1566 ª 2007 The Authors Journal compilation ª 2007 FEBS 1559
of these sites are located within the SNARE motif
(Ser61, Ser75 and Ser80). The fourth one is located in
the N-terminal sequence (Ser28) and has previously
been reported to represent a PKC phosphorylation site
in vitro [24]. Purified recombinant wild-type GST-
VAMP2 and the mutant GST-VAMP2 mt(1–4) were
subjected to in vitro kinase assay using recombinant
active PKCf, expressed in bacteria, and were subse-

quently analyzed for the presence of
32
P-phosphoryla-
tion (Fig. 5A). We found that wild-type GST-VAMP2
was specifically phosphorylated by PKCf in a concen-
tration-dependent manner (Fig. 5A, lanes 3 and 4), but
not by Akt (lane 2), while phosphorylation of GST-
VAMP2 mt(1–4) was strongly decreased (70%) when
compared with Flag-VAMP2 wild-type (lanes 6 and 7).
The activity of PKCf was verified by addition of the
PKC substrate myelin basic protein (MBP). This result
indicates that VAMP2 is directly phosphorylated by
PKCf.
We further investigated the specificity of the PKCf -
dependent VAMP2 phosphorylation by means of a
PCKf inhibitory peptide. For this, we transiently over-
expressed Flag-VAMP2 in COS-7 cells. After lysis
VAMP2 was immunoprecipitated using an anti-Flag
IgG and the precipitates were subjected to an in vitro
kinase assay using increasing amounts of recombinant
active PKCf in the absence (Fig. 5B, lane 2–4) or pres-
ence (Fig. 5B, lane 5–7) of a PKCf inhibitory peptide.
The precipitates were subjected to SDS ⁄ PAGE and
analyzed for radioactive signal by using a PhosphoI-
mager (Molecular Dynamics, Sunnyvale, CA, USA).
As can be seen, addition of recombinant active PKCf
led to a concentration-dependent substrate phosphory-
lation of the immunoprecipitated Flag-VAMP2. Fur-
thermore, addition of a PKCf inhibitory peptide
decreased PKCf autophosphorylation and abolished

the substrate phosphorylation of VAMP2. These data
confirm that VAMP2 is specifically phosphorylated by
active PKCf in vitro.
So far we have shown that VAMP2 is a substrate of
PKCf. Next, we wanted to find out whether VAMP2
could also be phosphorylated by PKCf that is activa-
ted by hormonal stimulation of the cells. We addressed
this question by transient overexpression of HA-PKCf
in COS-7 cells. These cells were either left unstimulated
or were stimulated with 100 ngÆmL
)1
epidermal growth
factor (EGF)-1, 100 nm insulin, or 100 ngÆmL
)1
IGF-1
as indicated in order to activate PKCf. Ten minutes
later, cells were lysed and HA-PKCf was immunopre-
cipitated using an anti-PKCf IgG. GST-VAMP2 was
added to the immunoprecipitates and subjected to an
in vitro kinase assay. Subsequently the proteins were
separated by SDS ⁄ PAGE and analyzed for radioactive
signals using PhosphoImager. As can be seen, hormo-
nal stimulation of the cells by epidermal growth fac-
tor 1 (EGF-1), insulin and IGF-1 led to strongly
increased substrate phosphorylation of recombinant
VAMP2 (3.5- to 4.5-fold) and to phosphorylation of
the immunoprecipitated PKCf (Fig. 5C). The activity
of PKCf was verified by addition of a MBP. These
results further support the idea that VAMP2 phos-
phorylation depends on activated PKCf.

ProF increases the PKCf -dependent VAMP2
phosphorylation in vitro
In a final experiment, we tested the effect of ProF on
the phosphorylation of VAMP2 by PKC f. In order to
test this, Flag-VAMP2 wild-type (wt) and Flag-
VAMP2 mt(1–4), were transiently expressed with and
without Myc-ProF in COS-7 cells. Flag-VAMP2 was
immunoprecipitated with an anti-Flag IgG. The preci-
pitates were subjected to an in vitro kinase assay using
recombinant active PKCf and subsequently analyzed
for the presence of
32
P-phosphorylation. We found that
phosphorylation of Flag-VAMP2 wt by active PKCf
was slightly (30%) but significantly (P<0.05)
increased in the presence of Myc-ProF (Fig. 6, lane 1
and 2). A strongly decreased in vitro
32
P-phosphoryla-
tion (90%) was found when Flag-VAMP2 mt(1–4)
was used as substrate (Fig. 6, lane 3), proving the spe-
cificity of the substrate phosphorylation. The activity of
PKCf was verified by addition of MBP. In summary,
these data indicate that Myc-ProF increases the in vitro
phosphorylation of Flag-VAMP2 by activated PKCf.
Discussion
We have previously identified ProF as a molecule that
is located on internal vesicles and which preferentially
binds to the activated kinases Akt and PKCf upon
hormonal stimulation of the cells [1]. This raised the

question of putative kinase substrates, which might
also interact with ProF. To address this question, we
performed a yeast two-hybrid screen, which indicated
VAMP2 as binding partner of ProF. We confirmed the
physical interaction of VAMP2 and ProF by coimmu-
noprecipitation of overexpressed and endogenous pro-
teins in both directions. VAMP2 is known to be
anchored via its transmembrane domain to secretory
vesicles in numerous cell lines, where it represents the
v-SNARE protein responsible for mediating fusion of
vesicles. Many vesicle cycling events rely on the inter-
action of v-SNARE and t-SNARE proteins, which
allow docking of vesicles to their target membranes.
SNARE complex formation is thought to bring the
WD-FYVE protein interacts with VAMP2 and PKCf T. Fritzius et al.
1560 FEBS Journal 274 (2007) 1552–1566 ª 2007 The Authors Journal compilation ª 2007 FEBS
opposing membranes close enough for fusion [25].
These SNARE-dependent fusion events include a num-
ber of secretory processes, such as insulin release from
pancreatic b-cells [26–28], synaptic vesicle exocytosis
[29], granule release in hematopoetic cells [30], and
aquaporin- [31], or GLUT4 translocation to the
plasma membrane [11,32]. In general, secretory events
are regulated by a variety of mechanisms including
phosphorylation of SNARE and accessory proteins
[29]. In adipocytes and skeletal muscle cells, VAMP2
has been described to bind to the t-SNARE proteins
syntaxin-4 and SNAP-23, found at the plasma mem-
brane [33,34], whereas in neurons VAMP2 interacts
with syntaxin-1 and SNAP-25 at the plasma membrane

for neurotransmitter release [14,15]. These findings
highlight the importance of VAMP2 in a number of
secretory systems.
In this study, we showed that ProF can act in an
adaptor protein-like fashion to mediate the interaction
between PKCf and VAMP2. ProF, VAMP2, and
PKCf partially colocalized on vesicular structures and
formed a complex. The contribution of additional pro-
teins to the formation of this complex cannot be exclu-
ded at the moment. Furthermore, because ProF is able
to form oligomers [1], it is possible that one ProF mole-
cule is not simultaneously interacting with VAMP2 and
PKCf, but instead that different ProFs may individu-
ally bind to VAMP2 and PKCf. Further studies using
mutants of ProF will investigate this question.
Finally, we hypothesized that ProF may be import-
ant for the phosphorylation of VAMP2. We found
that VAMP2 can be phosphorylated by activated
PKCf in vitro and that the presence of ProF increased
the PKCf- dependent VAMP2 phosphorylation. These
data support and expand earlier studies, which showed
that insulin-stimulated or overexpressed PKCf induced
serine phosphorylation of GLUT4 vesicle-associated
VAMP2 in vivo in rat myotubes, while expression
of dominant-negative PKCf completely abolished
VAMP2 phosphorylation [8]. Furthermore, it has been
shown that PKCf specifically associated with a
GLUT4- and VAMP2-positive cellular compartment,
and that overexpression of PKCf led to GLUT4 trans-
location to the plasma membrane and increased glu-

cose uptake even in the absence of insulin stimulation
[8]. Based on these studies, it is conceivable that the
PKCf-mediated VAMP2 phosphorylation affects the
fusion of vesicle with the plasma membrane. It is
currently unknown if the PKCf-dependent phosphory-
lation of VAMP2 influences the interaction of the
v-SNARE protein with its cognate t-SNAREs or with
accessory proteins. Whether the PKCf-dependent
phosphorylation of VAMP2 decreases or increases, the
interaction between the v-SNARE and t-SNARE pro-
teins and how this phosphorylation might regulate
vesicle cycling should be investigated in future studies.
We specified four serine residues within the VAMP2
molecule as potential phosphorylation sites and
Fig. 6. Phosphorylation of VAMP2 by PKCf is increased in vitro by ProF. Flag-VAMP2 wt and the VAMP2 mutant mt(1–4) were over-
expressed either with or without Myc-ProF in COS-7 cells. Flag–VAMP2 and Flag–VAMP2–Myc-ProF complexes were obtained by immuno-
precipitation with an antibody against the Flag epitope. Immunoprecipitations were phosphorylated by addition of 200 ng active recombinant
PKCf and c-
32
P-ATP. VAMP2 phosphorylation and PKCf autophosphorylation were analyzed by PhosphoImager, immunoprecipitation and
immunoblot were performed as indicated (top, left). Direct lysate shows expression controls (bottom, left). A quantification of PKCf- medi-
ated phosphorylation of VAMP2 in the absence or presence of ProF is shown (top, right). Results were obtained by densitometric scanning
of immunoblot bands from three independent experiments. The PhosphoImager signal of P-VAMP2 was normalized to Flag-VAMP2 sub-
strate phosphorylation in the absence of ProF ¼ 1. Values represent mean±
SD of three separate experiments (*P < 0.05, **P < 0.01). PKCf
activity was verified by addition of MBP (bottom, right).
T. Fritzius et al. WD-FYVE protein interacts with VAMP2 and PKCf
FEBS Journal 274 (2007) 1552–1566 ª 2007 The Authors Journal compilation ª 2007 FEBS 1561
mutated them to alanine. Three of the four mutated
serine residues were found within the highly conserved

SNARE motif of VAMP2. This motif fits to the
SNARE motifs of syntaxin-4 and SNAP-23 and would
allow a twisted, parallel 4-helical bundle [35]. As the
driving forces for the generation of the helical bundle
are mostly hydrophobic interactions, one or several
highly polar phosphorylated serine residues could dis-
turb the formation of the bundle. However, several
reports showed that decreased binding of SNARE pro-
teins to each other could lead to an increased fusion of
vesicles. For example, SNAP-25 is phosphorylated by
PKC at Ser187, which lies within the C-terminal
SNARE motif [36–38]. Activation of PKC by various
agents resulted in phosphorylation of SNAP-25 [37].
This phosphorylation decreased binding of SNAP-25
to syntaxin-1 and increased neurotransmitter release,
possibly by accelerating the SNARE complex dissoci-
ation and thus enhancing the rate of vesicle cycling
[37]. Thus, PKCf-mediated phosphorylation of
VAMP2 in response to stimulation of the cells, as
shown here, may affect the interaction of the
v-SNARE protein VAMP2, thereby influencing vesicle
trafficking. Another possibility would be that the
phosphorylation of VAMP2 by PKCf influences its
interaction with accessory proteins, which could up- or
down-regulate SNARE–SNARE protein interactions.
Our results show that ProF stimulates phosphoryla-
tion of the SNARE protein by PKCf in vitro, possibly
by recruitment of PKCf to VAMP2. This could be a
possible mechanism by which PKCf influences vesicle
trafficking. A recent publication supports our idea of

ProF as a regulator of vesicle trafficking. Knockdown
of the protein using small interfering (si)RNA was
reported to affect vesicle cycling and to inhibit endo-
cytosis in mammalian cells, as well as in the nematode
Caenorhabditis elegans [39]. As the expression profile
of ProF suggested a broad tissue distribution, we
hypothesize that a complex of ProF, VAMP2 and
PKCf may occur in various tissues and may be
involved in several secretory pathways.
Experimental procedures
Antibodies and reagents
Antibodies against Myc-epitope (A14, rabbit polyclonal
and 9E10, mouse monoclonal), against HA (Y-11, rabbit
polyclonal), PKCa (C-20, rabbit polyclonal), PKCd (C-17,
rabbit polyclonal), PKCe (C-15, rabbit polyclonal), PKCf
(C-20, rabbit and goat polyclonal), and PKCf-blocking
peptide were obtained from Santa Cruz Biotechnology
(Santa Cruz, CA, USA). Insulin was from Novo Nordisk
(Bagsvaerd, Denmark), the antibody against the GFP-
epitope (632377, rabbit polyclonal) was derived from BD
Biosciences (San Jose, CA, USA), and the antibody
against the GST-epitope (goat polyclonal) was from
Amersham Pharmacia (Piscataway, NJ, USA). The mouse
monoclonal and the rabbit polyclonal antibodies to
VAMP2 were from Synaptic Systems (Goettingen, Ger-
many) and ABR Affinity Bioreagents (Golden, CO, USA),
respectively. The antibody to Flag-epitope (M2, mouse
monoclonal), EGF-1, and MBP were from Sigma (St
Louis, MO, USA). IGF-1 was from Calbiochem Signal
Transduction (La Jolla, CA, USA). A polyclonal peptide

antibody directed against the 15 C-terminal amino acids
of ProF was raised in rabbits and affinity purified on the
peptide used for immunization [1]. This peptide was also
used for peptide competition in endogenous interaction
analysis during immunoprecipitation or, later, during west-
ern blotting. All secondary antibodies for western blotting
and indirect immunofluorescence staining were from Amer-
sham Pharmacia and Jackson Immuno Research (West
Grove, PA, USA), respectively.
Yeast two-hybrid analysis
A human B-cell-specific cDNA library was obtained from
S. J. Elledge (Baylor College of Medicine, Houston, TX,
USA) [40]. The yeast two-hybrid analysis was performed
essentially as described using full-length ProF as a bait [17].
Recombinant DNA manipulation and plasmid
constructs
Serine to alanine point mutations were inserted into the
coding sequence of Flag-VAMP2 using the Quick Change
Mutagenesis Kit (Stratagene, La Jolla, CA, USA). Plasmid
pEF-Flag-VAMP2 was used as a template. The primers for
mutagenesis were obtained from Microsynth (Balgach,
Switzerland), mutation 1 (Ser28Ala): forward, cca aac ctt
act gct aac agg aga ctg, reverse, cag tct cct gtt agc agt aag
gtt tgg; mutation 2 (Ser61Ala): forward, gac cag aag ttg gcg
gag ctg gat gac, reverse: gtc atc cag ctc cgc caa ctt ctg gtc,
mutation 3 (Ser75Ala): forward, gca ggg gcc gcc cag ttt
gaa, reverse: ttc aaa ctg ggc ggc ccc tgc, mutation 4 (Ser80-
Ala): forward, cag ttt gaa aca gct gca gcc aag ctc, reverse,
gag ctt ggc tgc agc tgt ttc aaa ctg. The inserted mutation is
in bold in the forward primer. Mutants containing four sin-

gle and one mutant harboring all four mutations were con-
structed. All mutations were verified by DNA sequencing.
N-Terminally Myc-tagged human ProF encoding
constructs were described earlier [1]: Myc-ProF, Myc-
ProFDFYVE lacking the FYVE domain for phosphatidyl-
inositol-3-phosphate binding, Myc-ProF 4–7 lacking blades
1–3 and the FYVE domain, and Myc-ProF 1–3 lacking
blades 4–7 and the FYVE domain were used in this study.
WD-FYVE protein interacts with VAMP2 and PKCf T. Fritzius et al.
1562 FEBS Journal 274 (2007) 1552–1566 ª 2007 The Authors Journal compilation ª 2007 FEBS
A HA-tagged PKCf construct was obtained from F. J.
Johannes (Stuttgart, Germany). The plasmid encoding
Flag-VAMP2 has previously been described [41] and was
kindly provided by M. Fukuda (RIKEN, Saitama, Japan).
Cloning and expression of GST-VAMP2
The coding sequence of VAMP2 and VAMP2 mt(1–4) was
excised from a pEF-plasmid containing Flag-tagged
VAMP2 (kindly provided by M. Fukuda, Saitama, Japan)
using BamHI and EcoRI restriction enzyme sites and cloned
into the pGEX-6p-2 vector (Amersham Biosciences) for
expression of GST-tagged VAMP2 and VAMP2 mt(1–4) in
transformed BL21 + (Invitrogen, Carlsbad, CA, USA)
Escherichia coli cells. Colonies were grown overnight in
20 mL of LB medium at 37 °C. LB medium (400 mL) was
added to this preculture and bacteria were grown at 28 °C
until the A
600
reached 0.6. Addition of 0.1 mm isopropyl
b-d-thiogalactoside to the bacterial culture, which was fur-
ther grown at 28 °C for another 6 h, induced protein

expression. After 6 h cells were collected by centrifugation
at 6000 g and 4 °C in a Sorvall RC-5B superspeed centri-
fuge using a Sorvall GSA rotor (Sorvall Instruments Inc.,
Newton, CT, USA). For extraction of recombinant pro-
teins, 10 mL of sucrose, Tris-HC1, and EDTA (STE)-buffer
was added to the bacterial pellets, containing 20 mm
Tris ⁄ HCl, pH 7.5, 300 mm NaCl, 2 mm EDTA, pH 8.0,
1 lgÆmL
)1
aprotinin, 1 lgÆmL
)1
leupeptin, and complete
EDTA-free protease inhibitor tablets (Roche Diagnostics,
Basel, Switzerland). Resuspended cells were lysed by two
freeze–thaw cycles in liquid nitrogen, followed by three cycles
of sonification (Branson sonifier 250) on ice. The bacterial
cell lysate was thereafter incubated with 100 lL dry volume
of glutathione-Sepharose beads (Amersham Biosciences) for
2 h at 4 °C in a spinning wheel. Beads were washed twice
with wash buffer, containing 50 mm Tris ⁄ HCl, pH 7.5,
250 mm NaCl, 1 lgÆmL
)1
aprotinin, 1 lgÆmL
)1
leupeptin,
and complete EDTA-free protease inhibitor tablets, to
remove unbound material. GST-tagged proteins were eluted
from beads by addition of three times 200-lL fractions of
elution buffer, containing 50 mm Tris ⁄ HCl, pH 7.5, 100 mm
NaCl, and 20 mm reduced glutathione by vigorous shaking

for 20 min at 15 °C. Fractions were pooled, dialyzed using
Centricon centrifugal filter devices (Millipore, Bedford, MA,
USA), aliquoted, and stored at )20 °C.
Retroviral transduction and generation of stably
transduced 3T3-L1 fibroblasts
Retroviruses containing the construct pRTP-Myc-ProF or
the empty pRTP vector as control were produced using the
BOSC-23 packaging cell-line as described [42,43]. Early pas-
sage 3T3-L1 fibroblasts were incubated in virus-containing
medium for 48 h. The cells were used for immunofluores-
cence studies.
Immunoprecipitation
For transient expression, the simian kidney-derived cell line
COS-7 (ATCC CRL-1651) and the human embryonic kid-
ney cell line 293T (ATCC CRL-11268) were transfected
with expression vectors encoding the indicated proteins
using Lipofectamine 2000 (Invitrogen) [1]. Cells were lysed
using a lysis buffer containing 100 mm NaCl, 1 mm EDTA,
20 mm Tris ⁄ HCl, and 0.5% NP-40 (NETN) [1] including
1 lgÆmL
)1
aprotinin, 1 lgÆmL
)1
leupeptin, and complete
EDTA-free protease inhibitor tablets (Roche Medicals) and
were cleared by centrifugation for 10 min at 16 000 g and
4 °C in a 5415 R refrigerated benchtop centrifuge from
Ependorf (Hamburg, Germany).
Interactions of endogenous proteins were analyzed in
mouse brain extracts. Proteins were extracted in an extrac-

tion buffer containing 20 mm Tris ⁄ HCl, pH 7.5, 150 mm
NaCl, 1 mm EDTA, 0.2% Nonidet NP-40 and supplemen-
ted with protease inhibitor tablets, by incubating the lysates
at 4 °C for 30 min with vigorous shaking. Glycerol was
added to a final concentration of 10% after clearing of ly-
sates. Protein concentrations were determined by Bradford
assay in a microtiterplate reader at 595 nm using bovine
serum albumin as standard.
Lysates were precleared for 1 h using 30 lL Protein-A-
Sepharose (Amersham Pharmacia Biotech). Cleared lysates
were immunoprecipitated with 1 lg of the appropriate
antibody for at least 3 h at 4 °C and then for 1 h in the
presence of 10 lL of protein G-sepharose (Amersham
Pharmacia Biotech). Immunoprecipitation of endogenous
ProF and VAMP2 was conducted overnight at 4 °C. For
immunoprecipitation of endogenous VAMP2 bringing
down ProF, the anti-VAMP2 IgG (ABR Affinity Biorea-
gents) was covalently coupled to protein G-sepharose in
order to reduce the signal of the IgG antibody heavy chain
as described earlier [1]. In brief, 10 lL of protein G-seph-
arose (Amersham) were incubated with 2 lg of anti-ProF
serum in 500 lL of washing-binding buffer (all buffers were
from Pierce, Rockford, IL, USA) for 30 min at room tem-
perature, followed by washing twice with this buffer and
incubating in 260 lL of cross-linking buffer with 850 lL
of freshly added dimethylpimelidate. Thereafter, protein
G-sepharose was incubated for 1 h at room temperature
and washed twice with cross-linking buffer, incubated for
10 min with blocking buffer at room temperature, washed
twice with blocking buffer, washed three times with elution

buffer, before boiling the samples for 5 min at 95 °C. For
competition studies, 1 lg anti-ProF IgG was preincubated
with 10 lg peptide on ice for 30 min for competition in
immunoblot and immunoprecipitate.
For studying the protein complexes, elution of overex-
pressed Myc-ProF and HA-PKCf was performed by vigor-
ous shaking for 1 h at 24 °C, followed by two rounds of
shaking for 45 min at 4 °C in 150 lL of NETN buffer con-
taining 900 lgÆlL
)1
of C-Myc-peptide (Roche) or PKCf
T. Fritzius et al. WD-FYVE protein interacts with VAMP2 and PKCf
FEBS Journal 274 (2007) 1552–1566 ª 2007 The Authors Journal compilation ª 2007 FEBS 1563
(C-20) blocking peptide (Santa Cruz). After elution with
C-myc, one-half of the eluate was used for immunoprecipi-
tation against HA-PKCf using 2 lg of anti-PKCf (C-20)
IgG (Santa Cruz). The other half of the eluate was used for
immunoprecipitation against Flag-VAMP2 using 2 lgof
anti-Flag (M2) IgG (Roche). After elution with PKCf
blocking peptide, one half of the eluate was used for immu-
noprecipitation against Myc-ProF using 2 lg of anti-Myc
(9E10) IgG. The other half of the eluate was used for
immunoprecipitation against Flag-VAMP2 using 2 lgof
anti-Flag (M2) IgG.
Lysates and the resulting immunoprecipitates were
resolved using commercial 10–20% SDS ⁄ PAGE (Invitro-
gen). The primary and secondary antibodies were used at a
dilution of 1 : 10 000. Western blotting of endogenous
ProF was performed overnight at 4 °C with a 1 : 1000 dilu-
tion of the antibody. All incubation and wash steps were

performed in 1% nonfat dry milk in Tris-buffered saline
containing 0.1% Tween-20. Detection was performed using
a chemiluminescence detection kit (Amersham Pharmacia
Biotech).
Confocal microscopy
COS-7 cells were grown on glass cover slips and were tran-
siently transfected with fugene-6 (Roche Medicals) accord-
ing to the manufacturers’ instructions. Twenty hours after
transfection, cells were washed with NaCl ⁄ P
i
, fixed with
3% paraformaldehyde and permeabilized in NaCl ⁄ P
i
con-
taining 0.25% Triton X-100. 3T3-L1 cells were fixed with
4% paraformaldehyde and permeabilized with 0.5% Triton
X-100 in NaCl ⁄ P
i
. Afterwards, the cover slips were succes-
sively incubated with appropriate primary and secondary
antibodies before they were mounted in Mowiol (Hoechst
Pharmaceuticals, Frankfurt, Germany). The cells were
examined by sequential excitation at 488 nm (fluorescein
isothiocyanate), 568 nm (tetramethylrhodamine isothiocya-
nate) and 633 nm (cyanine-5) using a confocal microscope
(SP2, Leica, Wetzlar, Germany) and a 40 · 1.25 oil objec-
tive (Leica). The images were processed by using photo-
shop (Adobe Systems, San Jose, CA, USA).
In vitro kinase assays
Flag-VAMP2 was transiently expressed in the presence or

absence of Myc-ProF or HA-PKCf in COS-7 cells. Cells
were lysed as described above and lysates were immunopre-
cipitated using an anti-Flag IgG. Immunocomplexes were
washed in a kinase reaction buffer (20 mm Tris ⁄ HCl,
pH 7.5, 20 mm MgCl
2
,1mm dithiothreitol). For the kinase
reaction, the buffer was supplemented with 3 lm protein
kinase A inhibitor, 20 lm ATP, and 10 lCi c-
32
P-ATP with
a specific activity of 3000 mCiÆmol
)1
(Amersham Phar-
macia). The reaction was started with addition of the
indicated amounts of recombinant active PKCf (Upstate
Biotechnology, Lake Placid, NY, USA) to the samples. The
phosphorylation reaction was conducted at 30 °C for
30 min and was stopped by boiling the samples in
SDS ⁄ PAGE loading buffer. For Fig. 5C, the reaction was
started without or with addition of 100 lm of PKC pseudo-
substrate inhibitory peptide II (Calbiochem). In all cases,
the activity of overexpressed or recombinant active PKCf
was verified by addition of 0.5 lg of the PKC substrate
myelin basic protein (Sigma). In vitro kinase experiments
with GST-tagged proteins were performed by addition of
1 lg of GST-VAMP2 and GST-VAMP2 mt(1–4) to 30 lL
of the kinase reaction buffer (20 mm Tris ⁄ HCl, pH 7.5,
20 mm MgCl
2

,1mm dithiothreitol), followed by addition
of 200 ng PKCf. For Fig. 5A, 3 lg of active recombinant
Akt1 (Upstate), or 100 or 200 ng of PKCf were added to
the reaction buffer and the kinase assay was thereafter per-
formed as described above. For the PKCf activity experi-
ments, HA-PKCf was transiently expressed in COS-7 cells,
which were left either unstimulated or stimulated for
10 min with 100 ng ÆmL
)1
EGF-1 (Sigma), 100 nm insulin
(Novo Nordisk) or 100 ngÆmL
)1
IGF-1 (Calbiochem)
before lysis as described above. Lysates were immunopre-
cipitated using an anti-PKCf IgG, immunocomplexes were
washed in kinase reaction buffer and subjected to in vitro
kinase assay as described above.
Samples were resolved on 10–20% SDS ⁄ PAGE. Expres-
sion of Flag-VAMP2 and Myc-ProF was verified by west-
ern blotting and phosphorylation was visualized using a
Storm 840 PhosphoImager.
Statistical analysis
Statistical analysis was performed using students t-test
(*P < 0.05, **P < 0.01).
Acknowledgements
The authors thank Dr M. Fukuda (Saitama, Japan) for
kindly providing us the plasmid encoding Flag-
VAMP2, F. J. Johannes (Stuttgart, Germany) for the
plasmid encoding HA-tagged PKCf, and B. Schaub
and J. Rohrer (Zurich, Switzerland) for the plasmids

encoding HA-Hrs and GFP-Vps4. The authors also
thank the Electron Microscopy Center (EMZ) at the
University of Zurich for providing the facilities for con-
focal microscopy and for expert help. Thanks go to
S. Dettwiler and C. Hagedorn for excellent technical
assistance. E. Haas (Zurich, Switzerland) and G. Burk-
ard (Berne, Switzerland) were very helpful during initial
studies. T. Fritzius: Figs 1,3,4(B), 5 and 6, generation
of GST-VAMP2 and GST-VAMP2 mutants. A. Frey:
Fig. 2,4(A,C), generation of Flag-VAMP2 mutants.
M. Schweneker: generation of Flag-VAMP2 mutants.
D. Mayer: initial yeast two-hybrid screen.
WD-FYVE protein interacts with VAMP2 and PKCf T. Fritzius et al.
1564 FEBS Journal 274 (2007) 1552–1566 ª 2007 The Authors Journal compilation ª 2007 FEBS
References
1 Fritzius T, Burkard G, Haas E, Heinrich J, Schweneker
M, Bosse M, Zimmermann S, Frey AD, Caelers A,
Bachmann AS et al. (2006) A WD-FYVE protein binds
to the kinases Akt and PKCzeta ⁄ lambda. Biochem J
399, 9–20.
2 Sondek J, Bohm A, Lambright DG, Hamm HE &
Sigler PB (1996) Crystal structure of a G-protein beta
gamma dimer at 2.1 A
˚
resolution. Nature 379, 369–374.
3 Stenmark H, Aasland R & Driscoll PC (2002) The
phosphatidylinositol 3-phosphate-binding FYVE finger.
FEBS Lett 513, 77–84.
4 Larance M, Ramm G, Stockli J, van Dam EM, Winata
S, Wasinger V, Simpson F, Graham M, Junutula JR,

Guilhaus M et al. (2005) Characterization of the role of
the Rab GTPase-activating protein AS160 in insulin-
regulated GLUT4 trafficking. J Biol Chem 280, 37803–
37813.
5 Sano H, Kane S, Sano E, Miinea CP, Asara JM, Lane
WS, Garner CW & Lienhard GE (2003) Insulin-stimu-
lated phosphorylation of a Rab GTPase-activating pro-
tein regulates GLUT4 translocation. J Biol Chem 278,
14599–14602.
6 Zeigerer A, McBrayer MK & McGraw TE (2004) Insu-
lin stimulation of GLUT4 exocytosis, but not its inhibi-
tion of endocytosis, is dependent on RabGAP AS160.
Mol Biol Cell 15, 4406–4415.
7 Kane S, Sano H, Liu SCH, Asara JM, Lane WS,
Garner CC & Lienhard GE (2002) A method to identify
serine kinase substrates – Akt phosphorylates a novel
adipocyte protein with a Rab GTPase-activating protein
(GAP) domain. J Biol Chem 277, 22115–22118.
8 Braiman L, Alt A, Kuroki T, Ohba M, Bak A, Tennen-
baum T & Sampson SR (2001) Activation of protein
kinase C zeta induces serine phosphorylation of
VAMP2 in the GLUT4 compartment and increases glu-
cose transport in skeletal muscle. Mol Cell Biol 21,
7852–7861.
9 Gerst JE (1999) SNAREs and SNARE regulators in
membrane fusion and exocytosis. Cell Mol Life Sci 55,
707–734.
10 Rossetto O, Gorza L, Schiavo G, Schiavo N, Scheller
RH & Montecucco C (1996) VAMP ⁄ synaptobrevin iso-
forms 1 and 2 are widely and differentially expressed in

nonneuronal tissues. J Cell Biol 132, 167–179.
11 Grusovin J & Macaulay SL (2003) Snares for GLUT4 –
mechanisms directing vesicular trafficking of GLUT4.
Front Biosci 8, d620–641.
12 McBride HM, Rybin V, Murphy C, Giner A, Teasdale
R & Zerial M (1999) Oligomeric complexes link Rab5
effectors with NSF and drive membrane fusion via
interactions between EEA1 and syntaxin 13. Cell 98,
377–386.
13 Sun W, Yan Q, Vida TA & Bean AJ (2003) Hrs regulates
early endosome fusion by inhibiting formation of an
endosomal SNARE complex. J Cell Biol 162, 125–137.
14 Sudhof TC (2004) The synaptic vesicle cycle. Annu Rev
Neurosci 27, 509–547.
15 Jahn R, Lang T & Sudhof TC (2003) Membrane fusion.
Cell 112, 519–533.
16 Hong W (2005) SNAREs and traffic. Biochim Biophys
Acta 1744, 493–517.
17 Schneider S, Buchert M, Georgiev O, Catimel B,
Halford M, Stacker SA, Baechi T, Moelling K &
Hovens CM (1999) Mutagenesis and selection of PDZ
domains that bind new protein targets. Nat Biotechnol
17, 170–175.
18 Komada M & Soriano P (1999) Hrs, a FYVE finger
protein localized to early endosomes, is implicated in
vesicular traffic and required for ventral folding
morphogenesis. Genes Dev 13, 1475–1485.
19 Komada M, Masaki R, Yamamoto A & Kitamura N
(1997) Hrs, a tyrosine kinase substrate with a conserved
double zinc finger domain, is localized to the cytoplas-

mic surface of early endosomes. J Biol Chem 272,
20538–20544.
20 Scheuring S, Rohricht RA, Schoning-Burkhardt B,
Beyer A, Muller S, Abts HF & Kohrer K (2001) Mam-
malian cells express two VPS4 proteins both of which
are involved in intracellular protein trafficking. J Mol
Biol 312, 469–480.
21 Steele MR, McCahill A, Thompson DS, MacKenzie C,
Isaacs NW, Houslay MD & Bolger GB (2001) Identifi-
cation of a surface on the beta-propeller protein
RACK1 that interacts with the cAMP-specific phospho-
diesterase PDE4D5. Cell Signal 13, 507–513.
22 Lopez-Bergami P, Habelhah H, Bhoumik A, Zhang W,
Wang LH & Ronai Z (2005) RACK1 mediates activa-
tion of JNK by protein kinase C. Mol Cell 19, 309–320
(erratum appears in Mol Cell 19, 578–579).
23 Elferink LA, Trimble WS & Scheller RH (1989) Two
vesicle-associated membrane protein genes are differen-
tially expressed in the rat central nervous system. J Biol
Chem 264, 11061–11064.
24 Nielander HB, Onofri F, Valtorta F, Schiavo G,
Montecucco C, Greengard P & Benfenati F (1995)
Phosphorylation of VAMP ⁄ synaptobrevin in synaptic
vesicles by endogenous protein kinases. J Neurochem 65,
1712–1720.
25 Bonifacino JS & Glick BS (2004) The mechanisms of
vesicle budding and fusion. Cell 116, 153–166.
26 Regazzi R, Wollheim CB, Lang J, Theler JM, Rossetto
O, Montecucco C, Sadoul K, Weller U, Palmer M &
Thorens B (1995) VAMP-2 and cellubrevin are

expressed in pancreatic beta-cells and are essential for
Ca(2+)-but not for GTP gamma S-induced insulin
secretion. EMBO J 14, 2723–2730.
T. Fritzius et al. WD-FYVE protein interacts with VAMP2 and PKCf
FEBS Journal 274 (2007) 1552–1566 ª 2007 The Authors Journal compilation ª 2007 FEBS 1565
27 Sadoul K, Lang J, Montecucco C, Weller U, Regazzi R,
Catsicas S, Wollheim CB & Halban PA (1995) SNAP-
25 is expressed in islets of Langerhans and is involved
in insulin release. J Cell Biol 128, 1019–1028.
28 Wheeler MB, Sheu L, Ghai M, Bouquillon A, Grondin
G, Weller U, Beaudoin AR, Bennett MK, Trimble WS
& Gaisano HY (1996) Characterization of SNARE pro-
tein expression in beta cell lines and pancreatic islets.
Endocrinology 137, 1340–1348.
29 Burgoyne RD & Morgan A (2003) Secretory granule
exocytosis. Physiol Rev 83, 581–632.
30 Logan MR, Odemuyiwa SO & Moqbel R (2003) Under-
standing exocytosis in immune and inflammatory cells:
the molecular basis of mediator secretion. J Allergy Clin
Immunol 111, 923–933.
31 Gouraud S, Laera A, Calamita G, Carmosino M, Pro-
cino G, Rossetto O, Mannucci R, Rosenthal W, Svelto
M & Valenti G (2002) Functional involvement of
VAMP ⁄ synaptobrevin-2 in cAMP-stimulated aquaporin
2 translocation in renal collecting duct cells. J Cell Sci
115, 3667–3674.
32 Nielsen S, Marples D, Birn H, Mohtashami M, Dalby
NO, Trimble M & Knepper M (1995) Expression of
VAMP-2-like protein in kidney collecting duct intracel-
lular vesicles. Colocalization with Aquaporin-2 water

channels. J Clin Invest 96, 1834–1844.
33 Foster LJ, Weir ML, Lim DY, Liu Z, Trimble WS &
Klip A (2000) A functional role for VAP-33 in insulin-
stimulated GLUT4 traffic. Traffic 1, 512–521.
34 Watson RT, Kanzaki M & Pessin JE (2004) Regulated
membrane trafficking of the insulin-responsive glucose
transporter 4 in adipocytes. Endocr Rev 25 , 177–204.
35 Sutton RB, Fasshauer D, Jahn R & Brunger AT (1998)
Crystal structure of a SNARE complex involved in
synaptic exocytosis at 2.4 A
˚
resolution. Nature 395,
347–353.
36 Kataoka M, Kuwahara R, Iwasaki S, Shoji-Kasai Y &
Takahashi M (2000) Nerve growth factor-induced
phosphorylation of SNAP-25 in PC12 cells: a possible
involvement in the regulation of SNAP-25 localization.
J Neurochem 74, 2058–2066.
37 Shimazaki Y, Nishiki T, Omori A, Sekiguchi M,
Kamata Y, Kozaki S & Takahashi M (1996) Phosphor-
ylation of 25-kDa synaptosome-associated protein.
Possible involvement in protein kinase C-mediated
regulation of neurotransmitter release. J Biol Chem 271,
14548–14553.
38 Gonelle-Gispert C, Costa M, Takahashi M, Sadoul K
& Halban P (2002) Phosphorylation of SNAP-25 on ser-
ine-187 is induced by secretagogues in insulin-secreting
cells, but is not correlated with insulin secretion. Bio-
chem J 368, 223–232.
39 Hayakawa A, Leonard D, Murphy S, Hayes S, Soto M,

Fogarty K, Standley C, Bellve K, Lambright D, Mello
C et al. (2006) The WD40 and FYVE domain contain-
ing protein 2 defines a class of early endosomes neces-
sary for endocytosis. Proc Natl Acad Sci USA 103,
11928–11933.
40 Durfee T, Becherer K, Chen PL, Yeh SH, Yang YZ,
Kilburn AE, Lee WH & Elledge SJ (1993) The retino-
blastoma protein associates with the protein phospha-
tase type-1 catalytic subunit. Gene Dev 7, 555–569.
41 Fukuda M (2002) Vesicle-associated membrane protein-
2 ⁄ synaptobrevin binding to synaptotagmin I promotes
O-glycosylation of synaptotagmin I. J Biol Chem 277,
30351–30358.
42 Heinrich J, Bosse M, Eickhoff H, Nietfeld W, Rein-
hardt R, Lehrach H & Moelling K (2000) Induction of
putative tumor-suppressing genes in Rat-1 fibroblasts by
oncogenic Raf-1 as evidenced by robot-assisted complex
hybridization. J Mol Med
78, 380–388.
43 Pear WS, Nolan GP, Scott ML & Baltimore D (1993)
Production of high-titer helper-free retroviruses by
transient transfection. Proc Natl Acad Sci USA 90,
8392–8396.
WD-FYVE protein interacts with VAMP2 and PKCf T. Fritzius et al.
1566 FEBS Journal 274 (2007) 1552–1566 ª 2007 The Authors Journal compilation ª 2007 FEBS