T H E J O U R N A L O F C E L L B I O L O G Y
© The Rockefeller University Press $30.00
The Journal of Cell Biology, Vol. 179, No. 6, December 17, 2007 1123–1131
To understand the mechanism of membrane fi ssion, we identi-
fi ed and used a compound called Ilimaquinone (IQ), which ve-
siculates the Golgi apparatus via a trimeric G protein subunit βγ
and a serine/threonine kinase called protein kinase D (PKD)–
dependent process (Takizawa et al., 1993; Jamora et al., 1997,
1999). Importantly, PKD is necessary for the biogenesis of TGN
to cell surface transport carriers (Liljedahl et al., 2001; Bard and
Malhotra, 2006). The binding of PKD to TGN requires DAG
(Baron and Malhotra, 2002) and is activated by Golgi-associated
PKCη (Diaz Anel and Malhotra, 2005). PKD activates the lipid
kinase activity of PI4kinase IIIß to generate phosphoinositide
4-phosphate (PI4P) from PI, and regulates the binding of
ceramide transfer protein CERT to PI4P. PI4P is required for TGN-
to-cell surface transport (Walch-Solimena and Novick, 1999;
Audhya et al., 2000; Godi et al., 2004; Hausser et al., 2005,
2006; Fugmann et al., 2007). The evidence for PKD’s role in
the formation of TGN to cell surface transport carriers is though
use of a kinase-dead (KD) form and pharmacological inhibitors.
The best evidence for PKD’s direct involvement in membrane
fi ssion requires that its depletion inhibits protein secretion.
However, the problem is exacerbated by the fact that there are
three isoforms of PKD in the mammalian cells (1, 2, and 3)
(Rykx et al., 2003), and all are involved in the formation of
basolaterally directed transport carriers (Yeaman et al., 2004).
We believe we have now addressed this issue. Our fi ndings
reveal that HeLa cells contain predominantly PKD2 and PKD3,
and virtually no PKD1. PKD2 and PKD3 dimerize at the TGN
and we suggest they activate different substrates. Importantly,
depletion of PKD2 and PKD3 by siRNA inhibits TGN-to-cell
surface transport. Under these conditions, cargo containing
tubules and reticular membranes accumulate at the TGN. In con-
trast, overexpression of an activated PKD causes extensive vesi-
culation of TGN. These results demonstrate convincingly that
PKD is a bona fi de component of membrane fi ssion used to
regulate the number and size of TGN-to-cell surface transport
carriers depending on the physiological (cargo) needs.
Results and discussion
Depletion of PKD2 and PKD3 inhibits TGN-
to-cell surface protein transport
RT-PCR–based analysis revealed that of the three PKD iso-
forms, only PKD2 and PKD3 were expressed in HeLa cells
(Fig. 1 A). These results were confi rmed by quantitative RT-
PCR (qRT-PCR): PKD1-specifi c mRNA is virtually undetect-
able (10- and 12-fold lower) compared with PKD2 and PKD3,
respectively (Fig. 1 B). Specifi c siRNAs were designed to de-
plete PKD2 and PKD3 in HeLa cells. Western blotting with
specifi c antibodies revealed a 70–75% reduction in the level
of PKD2 and PKD3, respectively (Fig. 1, C and E). By com-
parison, the level of β-actin was not affected by PKD-specifi c
siRNAs (Fig. 1 D).
To test the effect of PKD2 and PKD3 depletion on protein
secretion, control cells and depleted HeLa cells were cotransfected
with a plasmid expressing HRP containing the N-terminal sig-
nal sequence (SS) as described previously (Bard et al., 2006)
Dimeric PKD regulates membrane fi ssion to form
transport carriers at the TGN
Carine Bossard,1 Damien Bresson,2 Roman S. Polishchuk,3 and Vivek Malhotra1
1Section of Cell and Developmental Biology, University of California, San Diego, La Jolla, CA 92093
2La Jolla Institute for Allergy and Immunology, Developmental Immunology 3, La Jolla, CA 92037
3Department of Cell Biology and Oncology, Consorzio Mario Negri Sud, Santa Maria Imbaro (CH) 66030, Italy
TGN to cell surface transport carriers. We now provide
defi nitive evidence that PKD has a function in membrane
fi ssion. PKD depletion by siRNA inhibits traffi cking from
rotein kinase D (PKD) is recruited to the trans-Golgi
network (TGN) through interaction with diacylglyc-
erol (DAG) and is required for the biogenesis of
the TGN, whereas expression of a constitutively active PKD
converts TGN into small vesicles. These fi ndings demon-
strate that PKD regulates membrane fi ssion and this activ-
ity is used to control the size of transport carriers, and
to prevent uncontrolled vesiculation of TGN during pro-
Correspondence to Vivek Malhotra: email@example.com
V. Malhotra’s present address is Cell and Developmental Biology program, Centre
de Regulació Genòmica (CRG), Dr. Aiguader 88, 08003 Barcelona, Spain.
Abbreviations used in this paper: CA, constitutively active; KD, kinase dead;
PKD, protein kinase D; PLAP, placental alkaline phosphatase; SS, signal sequence;
JCB • VOLUME 179 • NUMBER 6 • 2007 1124
together with a plasmid expressing placental alkaline phospha-
tase (PLAP), a GPI-anchored protein that contains an apical
sorting signal (Lisanti et al., 1990; Lipardi et al., 2000). The ac-
tivities of HRP and PLAP released into the medium were mea-
sured by chemiluminescence (Bard et al., 2006). HRP secretion
is inhibited by 82% in cells transfected by PKD2 siRNA and
by 80% in cells transfected by PKD3 siRNA compared with
control siRNA-transfected cells (Fig. 2 A). None of the PKD
siRNAs have any effect on PLAP secretion (Fig. 2 B). These
fi ndings strengthen our previous proposal that similarly to po-
larized cells that have distinct apical and basolateral targeting
pathways, nonpolarized HeLa cells sort proteins into apical-
like (PKD-independent) and basolateral-like (PKD-dependent)
pathways (Yeaman et al., 2004). Simultaneous depletion of
PKD2 and PKD3 from HeLa cells did not have a synergistic
effect in inhibiting HRP secretion (unpublished data). Because
siRNA-based depletion is not complete, we suggest the residual
PKD is suffi cient to support traffi cking from the TGN to the
cell surface. To further ascertain the involvement of PKD2 and
PKD3 in traffi cking, their involvement in the transport of vesic-
ular stomatitis virus (VSV)–G protein was monitored. HeLa
cells were fi rst cotransfected with either siRNA specifi c for
PKD2, PKD3, or a control, and a fl uorescent-labeled siRNA to
identify transfected cells. After 30 h, the cells were transfected
with the tsO45 strain of VSV-G–GFP, which has a thermosensi-
tive mutation causing it to unfold in the ER at the nonpermissive
temperature of 39.5°C. Upon shifting cells to the permissive
temperature of 32°C, tsO45VSV-G protein folds and is trans-
ported to the cell surface. The amount of VSV-G at the cell
surface was quantifi ed by immunofl uorescence fl ow cytometry
permitting analysis of more than 5,000 cells transfected with
both siRNA and VSV-G-GFP. In control cells, VSV-G protein
was found at the cell surface within 40 min of shift to 32°C.
In PKD2- and PKD3-depleted cells (either individually or to-
gether), 50% less VSV-G was found at the cell surface (Fig. 2 C).
This is a signifi cant effect on the traffi cking of VSV-G to the cell
surface considering that siRNA-based depletion of PKD is not
complete, and the residual levels (25–30%) would support traf-
fi cking, albeit at a slower rate.
Immunoelectron microscopy was used to monitor effects
on the secretion of ss-HRP (cargo) in cells depleted of PKD2
and PKD3. Compared with control cells (Fig. 2 D), PKD2 and
PKD3 depletion resulted in accumulation of HRP containing
highly fenestrated membranes, and large tubules in the trans
region of the Golgi stacks (Fig. 2, E and F). To quantitate the
number of tubules in the TGN, 20 different Golgi stacks were
visualized. In control cells there were 1 to 3 tubules per stack,
whereas in PKD-depleted cells the number of tubules increased
to 3 to 6 per stack. In sum, PKD depletion results in accumula-
tion of HRP in the TGN and on average a 2.9-fold increase in
the number of HRP-containing tubules per Golgi stack. We suggest
that these tubules and fenestrated HRP-containing membranes
would ordinarily be converted into small transport carriers by
membrane fi ssion. However, depletion of PKD2 and PKD3 in-
hibits events leading to membrane fi ssion, thus accumulating
cargo (in this instance HRP) in tubular elements.
PKD2 and PKD3 dimerize in vitro and
in vivo and transphosphorylate
Why does depletion of either PKD2 or PKD3 affect Golgi-
to-cell surface transport? Why is the other (nondepleted) isoform
not functional under such conditions? Do these forms dimerize
to activate downstream targets? To test this hypothesis, PKD2
and PKD3 were immunoprecipitated separately from HeLa cell
lysates and Western blotted with anti-PKD3 and anti-PKD2,
respectively. Western blotting pure recombinant PKD2 and 3
confi rmed the specifi city of the antibodies to PKD2 and PKD3
(Fig. 3 A). The endogenous PKD2 and PKD3 were found to co-
precipitate (Fig. 3 B). Exogenously expressed GST-PKD2 and
Flag-PKD3 also coprecipitate specifi cally (Fig. 3 C). GST alone
Figure 1. Relative expression of PKD iso-
forms in HeLa cells and their depletion by siRNA.
(A) Analysis of mRNA expression by RT-PCR
shows that PKD2 and PKD3 are the only PKD
isoforms expressed in HeLa cells. RT-PCR reaction
without the reverse transcriptase (RT−) was used
as a negative control and PCR with the corre-
sponding PKD cDNA (C) as a positive control.
(B) Quantitative real-time RT-PCR analysis was
performed on RNA extracted from HeLa cells.
Bars represent the mean (±SD) of the relative
mRNA expression of each PKD isoform com-
pared with the average expression of β-actin.
*, P < 0.01 compared with PKD1. (C) PKD2 and
PKD3 protein levels in HeLa cells transfected
with the indicated siRNA were detected by
immunoblot analysis after immunoprecipitation
from 100 μg of cell lysate using anti-PKD2 and
anti-PKD3 antibodies, respectively. (D) β-Actin
expression in the lysates used for immunopreci-
pitation was monitored as a loading control.
(E) The effect of PKD2 and PKD3 siRNA was
quantifi ed by densitometry and normalized to
the expression of PKD2 and PKD3, respectively,
in cells transfected with control siRNA.
MEMBRANE FISSION BY PKD • BOSSARD ET AL.1125
does not coprecipitate with Flag-PKD3, and Flag-HRP used as
a negative control for Flag-tagged protein does not coprecipitate
with GST-PKD2. It is interesting to note that Flag-PKD3 con-
struct, when expressed, co-migrates with a fragment of a 34-kD
apparent molecular weight. This product is specifi c for PKD3
because it only appears when Flag-PKD3 is expressed. Because
the Flag tag is present at the N-terminal part of PKD3, this 34-kD
fragment is a degradation product of PKD3 containing its cysteine-
rich domain. More interestingly, this degradation product (Flag-
PKD3-Nterm) is also coimmunoprecipitated with GST-PKD2
(Fig. 3 C, top right panel), suggesting that PKD3’s cysteine-rich
domain binds PKD2. Purifi ed, recombinant-tagged PKD2 and 3
incubated in vitro also coprecipitate, which reveals a direct inter-
action between PKD2 molecules, PKD3 molecules, and between
PKD2 and PKD3 (Fig. 3 D).
PKDs cycle between cytosol and the TGN. A kinase-
dead form binds to the TGN but fails to dissociate. Under such
conditions, cargo-fi lled transport carriers form but fail to detach
and grow into large tubules. PKD-KD is found both on the TGN
and the tubules. Flag-PKD2-KD and GST-PKD3-WT were expres-
sed in HeLa cells and the cells stained with anti-PKD2 and anti-
GST antibody. Flag-PKD2-KD was localized to the TGN and
the tubules and more interestingly, so was PKD3-WT. Similarly,
in cells expressing GST-PKD3-KD, Flag-PKD2-WT was found
on the TGN and the tubules (Fig. 3 E). These fi ndings further
strengthen our arguments that PKD2 and PKD3 form a dimer
and reveal for the fi rst time the localization of the wild-type
kinase on the TGN and the TGN-derived tubules, when the corre-
sponding partner is a kinase-inactive form. The fact that these
tubules do not detach indicates that both kinases have to be
functionally active for membrane fi ssion. The binding of PKD
to the Golgi membranes requires DAG (Baron and Malhotra,
2002). However, PKD2 and PKD3 can interact in cells depleted
of DAG (unpublished data). These fi ndings reveal that PKD2
Figure 2. Depletion of PKD2 or PKD3 inhibits secretion of ss-HRP. (A and B) ss-HRP and PLAP cDNAs were cotransfected in HeLa cells depleted or not of PKD2
or PKD3. 20 h after transfection, HRP activity secreted in the medium was measured by chemiluminescence. PLAP activity in cell lysates and secreted into the
medium was measured as described in Materials and methods. Bars represent the mean ± SD of HRP (A) or PLAP (B) activity in the medium normalized by PLAP
activ ity in cell lysates/total protein concentration. (C) Depletion of PKD2 or PKD3 inhibits VSV-G transport. tso45VSV-G-GFP plasmid was transfected in HeLa cells
depleted or not of PKD2 or PKD3. The levels of VSV-G at the cells surface and inside the cells were determined by FACS analysis. The bars represent the relative
ratio of VSV-G at the surface to the total expressed in depleted cells compared with siRNA control transfected cells ± SEM. *, P < 0.01 compared with the control.
(D–F) PKD2- and PKD3-depleted HeLa cells and control cells were transfected with ss-HRP and the organization of the Golgi membranes and the localization of
HRP monitored by electron microscopy. In control cells (D), the arrow indicates a single tubular profi le and arrowheads indicate round profi les (presumably vesi-
cles) in the trans-Golgi region. In contrast, in PKD2- and PKD3-depleted cells (E and F), HRP accumulates in tubular membranes. The arrows indicate multiple tubu-
lar profi les at the TGN. The open arrowheads show pearling tubules. The arrowhead in F is a clathrin-coated vesicle revealing the trans side of the Golgi stack.
JCB • VOLUME 179 • NUMBER 6 • 2007 1126
Figure 3. PKD2 and PKD3 dimerize and transphosphorylate. (A) Specifi city of PKD2 and PKD3 antibodies. Pure recombinant Flag-PKD2 and GST-PKD3
were Western blotted with anti-PKD2 and -PKD3 antibodies. Anti-PKD2 antibody recognizes PKD2 (lane 2) and anti-PKD3 antibody recognize PKD3 (lane 3).
(B) PKD2 and PKD3 interact. PKD2 or PKD3 was immunoprecipitated from HeLa cell lysates with specifi c antibodies and the precipitates blotted with
anti-PKD2 antibody (lanes 1 and 2) or PKD3 antibody (lanes 3 and 4). (C) Exogenously expressed PKD2 and PKD3 interact. GST (lane 1) or GST-PKD2
(lanes 2 and 3) was coexpressed with Flag-HRP (lane 2) or FLAG-PKD3 (lanes 1 and 3) in HeLa cells. The cells were immunoprecipitated with anti-Flag anti-
body (top left) or anti-GST antibody (top right) and Western blotted with anti-GST or anti-Flag antibody, respectively. To verify that each tagged protein was
expressed and immunoprecipitated, the Flag and the GST precipitates were respectively blotted with anti-Flag (bottom left) and anti-GST (bottom right).
(D) PKD2 and PKD3 interact directly. Pure recombinant GST-tagged proteins were incubated in vitro with pure recombinant Flag-tagged proteins. After GST
pull-down, the precipitates were Western blotted with anti-Flag antibody, followed by anti-GST antibody. (E) PKD2 and PKD3 colocalize on PKD-KD tubes.
HeLa cells were cotransfected with Flag-PKD2-KD and GST-PKD3-WT (top) or with Flag-PKD2-WT and GST-PKD3-KD (bottom). The cells were visualized by
fl uorescence microscopy with anti-PKD2 and anti-GST antibody. (F) PKD2 and PKD3 transphosphorylate. Pure recombinant GST-PKD2-WT was incubated
MEMBRANE FISSION BY PKD • BOSSARD ET AL.1127
and PKD3 dimerize in the cytoplasm before DAG-dependent
recruitment to the TGN.
PKD1 in addition to autophosphorylation transphosphory-
lates other PKD1 molecules (Sanchez-Ruiloba et al., 2006).
We therefore tested whether PKD2 and PKD3 share this property.
Flag-PKD2-KD, Flag-PKD3-KD, GST-PKD2-WT, and GST-
PKD3-WT were expressed in 293T cells and immunoprecipitated
by specifi c antibodies. The isolated (soluble) Flag-tagged PKD-
KD (2 or 3) was incubated with or without the GST-tagged
kinases WT (2 or 3) attached to the beads in a kinase buffer.
The phosphorylation status of the kinase-dead proteins by wild-
type kinases was determined by SDS-PAGE followed by auto-
radiography. Our results reveal that PKD2 transphosphorylates
other PKD2 molecules and also PKD3, and similarly PKD3 trans-
phosphorylates PKD2 and other PKD3 molecules (Fig. 3 F).
A constitutively activated PKD converts
TGN into small vesicles
If PKD depletion inhibits membrane fi ssion then its overacti-
vation should cause extensive vesiculation. A constitutively acti-
vated PKD containing a CAAX domain was generated to test
this hypothesis. Proteins containing CAAX domain at their
C-terminal can be prenylated, which confers a greater hydro-
phobicity and membrane anchoring (Choy et al., 1999; Wright and
Philips, 2006). We reasoned that PKD-CAAX upon recruitment
to the TGN through its C1domain (Maeda et al., 2001) will be
inserted into the membrane via prenylation. This will retain acti-
vated PKD on TGN and hyperactivate the fi ssion process. A con-
stitutively activated PKD (PKD-CA) was generated by replacing
Ser744 and Ser 748 with glutamic acid to mimic the phosphory-
lated form, as described by Iglesias et al. (1998). The kinase-dead
form of PKD (PKD-KD) refers to PKD-K618N as described pre-
viously (Liljedahl et al., 2001).
HeLa cells stably expressing a GFP-tagged form of man-
nosidase II (MannII-GFP) (Sutterlin et al., 2005) were trans-
fected with PKD-WT, PKD-CAAX-WT, PKD-CAAX-CA, or
PKD-CAAX-KD. 24 h after transfection the cells were visual-
ized with anti-GST antibodies to detect transfected cells, and
the organization of Golgi apparatus was visualized with MannII-
GFP (early Golgi) or an antibody against TGN46 (late Golgi).
In cells transfected with PKD-CAAX-CA, the Golgi (cis Golgi
as well as TGN) was found fragmented, whereas in cells trans-
fected with PKD-WT or PKD-CAAX-KD the Golgi was un-
affected (Fig. 4 A). In cells transfected with PKD-CAAX-WT, the
Golgi was fragmented when expressed at high levels (Fig. 4 A,
line 2). The recruitment of PKD to TGN requires DAG, which
is inhibited by treatment of cells with fumonisin B1 (FB1) that
lowers the intracellular pool of DAG (Baron and Malhotra,
2002). We found that depletion of DAG by FB1 inhibited the
recruitment of PKD-KD and PKD-CAAX-KD to the Golgi
membranes, and PKD-CAAX-CA mediated fragmentation of
Golgi membranes (Fig. 4 B). Thus, the CAAX motif does not
provide any specifi city to PKD recruitment but simply anchors
it to the TGN in a DAG-dependent manner.
To further ascertain the effects of constitutively activated
PKD on the organization of Golgi membranes, the following
constructs were coexpressed in HeLa cells; sialyltransferase (ST)-
HRP and either PKD-WT, PKD-CAAX-WT, PKD-CAAX-CA,
or PKD-CAAX-KD. The cells were fi xed and thin sections visu-
alized by immunoelectron microscopy with anti-HRP antibody
to visualize the resident enzyme ST of TGN (Fig. 5, A–D, left
panel), or anti-TGN46 antibody, which is a cargo protein of the
TGN-to-cell surface pathway (Fig. 5, E–H, right panel). There
was a defi nitive increase in the number of vesicles in the vicinity
of Golgi stacks in cells expressing either PKD-CAAX-WT or
PKD-CAAX-CA compared with PKD-CAAX-KD or PKD-WT
(Fig. 5). A quantitation of these images revealed a fi ve- to seven-
fold increase in the number of vesicles containing TGN46 in
cells expressing CAAX variants of PKD-WT or the constitu-
tively activated form compared with PKD-WT or PKD-CAAX-KD
(Table I). Interestingly, there was a 20- and 25-fold increase in
the number of ST-containing vesicles in cells expressing PKD-
CAAX-WT and PKD-CAAX-CA, respectively, compared with
PKD-WT or PKD-CAAX-KD. Thus, overexpression of the wild-
type or constitutively activated form of PKD hyperactivates the
fi ssion reaction, and causes extensive vesiculation of the TGN.
We suggest that this reaction fi rst generates the normal cargo
(TGN46 containing) transport vesicles, but the reaction continues
and consumes domains containing Golgi resident enzymes that
are ordinarily excluded from participating in TGN-to-cell surface
Our fi ndings clearly demonstrate the regulation of membrane
fi ssion by PKD. But how does PKD regulate membrane fi ssion?
As mentioned above, PKD is recruited by a DAG-dependent
manner to the TGN and then activated by PKCη. We suggest
that a PKD-dependent increase in local concentration of DAG
separates transport carriers from the TGN by periplasmic fusion.
This coat- and dynamin-independent (fi ssion) reaction is there-
fore fundamentally different from the process of COPI, COPII, and
clathrin-coated vesicle biogenesis.
Materials and methods
Total RNA was extracted from HeLa cells with Nucleospin RNAII (Machery-
Nagel) and reverse-transcribed with Titanium One-Step RT-PCR kit (Clontech
Laboratories, Inc.) according to the manufacturer’s instructions.
Reverse transcription was performed using 1 μg of purifi ed RNA, oligo dT,
and SuperScript III reverse transcriptase (Invitrogen) in a 20-μl reaction.
Amplifi cation of each specifi c transcript was performed using RT2
PCR primer Set (SuperArray). Plasmids containing cDNAs for each specifi c
PKD isoform were used as standards for real-time quantitative PCR amplifi -
cation (Q-PCR). These plasmids were 10-fold serially diluted and used as
alone (lane 1) or mixed with either Flag-PKD2-KD (lane 3) or Flag-PKD3-KD (lane 7) for in vitro kinase assays. Similarly, GST-PKD3-WT was incubated alone
(lane 4) or mixed with Flag-PKD3-KD (lane 6) or Flag-PKD2-KD (lane 8). The lack of kinase activity of Flag-PKD2-KD and Flag-PKD3-KD is shown in lanes
2 and 5, respectively.
JCB • VOLUME 179 • NUMBER 6 • 2007 1128
templates for the Q-PCR to generate standard curves (ranging from 102 to
105 copies/μl). Real-time Q-PCR assays were performed with an Mx4000
Multiplex Quantitative QPCR System (Stratagene). Reactions were per-
formed using 200-nM primers, 1 μl template/25 μl PCR reaction and the
iTaq SYBR Green Supermix with ROX (Bio-Rad). A two-step PCR method
(denaturation at 95°C for 30 s and annealing/extension at 60°C for 1 min)
was used. Each assay included the analysis of the samples in duplicates.
In addition, samples were run at least three times to check for interassay
variability. Melting curve analyses were performed on all PCR reactions to
check for specifi city of the amplifi cation. Real-time qRT-PCR analyses for
β-actin were included as housekeeping genes to normalize the data.
Antibodies in this study included goat anti-GST (GE Healthcare), rabbit
anti-GST (AbCam), sheep anti–human TGN46 (AbD Serotec), rabbit affi nity-
purifi ed anti-PKD3 (Bethyl), rabbit affi nity-purifi ed anti-PKD2 (Bethyl) mono-
clonal anti-β-actin (Sigma-Aldrich), monoclonal anti-Flag (Sigma- Aldrich),
AMCA donkey anti–rabbit (Jackson ImmunoResearch Laboratories), and Texas
Figure 4. Expression of an overactivated form of PKD fragments the Golgi apparatus. (A) HeLa cells stably expressing MannII-GFP were transfected with
GST-tagged PKD WT, PKD-CAAX-WT, PKD-CAAX-CA, or PKD-CAAX-KD. The localization of PKD and the organization of Golgi membranes were monitored
by fl uorescence microscopy with anti-GST and TGN46 antibody, respectively. (B) PKD-CAAX-CA–dependent Golgi fragmentation requires DAG. HeLa
MannII-GFP cells were pretreated with fumonisin1 (FB1) 24 h before transfection with PKD constructs. 24 h after transfection, cells were fi xed and stained
with anti-GST and anti-TGN46 antibodies.
MEMBRANE FISSION BY PKD • BOSSARD ET AL.1129
red donkey anti–sheep (Jackson ImmunoResearch Laboratories). The mono-
clonal antibody 8G5F11, which recognizes the extracellular domain of
VSV-G, was provided by Dr. Douglas Lyles (Wake Forest University School
of Medicine, Winston-Salem, NC).
Cell culture and transfection
293-T cells, HeLa cells, and the cell line stably expressing a GFP-tagged
form of mannosidase II (HeLa MannII-GFP) were grown in complete me-
dium consisting of DME (Cellgro) containing 10% FCS and supplemented
Figure 5. Analysis of HeLa cells transfected with PKD CAAX isoforms by electron microscopy. HeLa cells transfected with PKD-WT, PKD-CAAX-WT, PKD-CAAX-CA,
or PKD-CAAX-KD alone (right) or with ST-HRP (left) were processed for electron microscopy. (A) Gold particles indicating the presence of PKD-WT reveal
its presence at the TGN (where ST-HRP is also visible). Few vesicles lacking ST-HRP were also visible in sections (arrow). (B) Expression of PKD-CAAX-WT
increased the number of ST-HRP–positive vesicles (arrows). (C) PKD-CAAX-CA induced an increase in the number of vesicles with ST-HRP (arrowheads);
however, number of unlabeled vesicles also increases (arrows). Row of vesicles (empty arrows) between ST-positive cisterna and the rest of the stack may
represent cisterna consumed by vesicles, and this also might explain why ST-positive cisterna appears to be peeling off. (D) Cells expressing PKD-CAAX-KD
exhibit regular trans-Golgi cisterna (arrow) labeled with ST-HRP. Some tubular structures (presumably TGN) are also visible at the trans-face of the Golgi
(arrowhead). (E) Cells expressing PKD-WT exhibit regular TGN with both tubular (arrow) and vesicles (arrowhead). PKD is detected at the TGN46-positive
membranes. (F) PKD-CAAX-WT induced an increase in the number of TGN46-positive vesicles (arrows). (G) Expression of PKD-CAAX-CA induced extensive
vesiculation of the Golgi apparatus. Both TGN46-positive (arrows) and TGN46-negative (arrowheads) vesicles increased in numbers under these conditions.
PKD is detected at the TGN46-positive membranes. (H) Long TGN46-positive (arrows) tubular structures were detected in PKD-CAAX-KD–expressing cells
without any obvious increase in number of vesicles.
JCB • VOLUME 179 • NUMBER 6 • 2007 1130
with 0.8 mg/ml of geneticin for HeLa MannII-GFP at 37°C in a 7% CO2
incubator. The cells were transfected with FuGene 6 (Roche) or Lipofectamine
2000 (Invitrogen) following the manufacturer’s recommendations.
The day before transfection, HeLa cells were plated in order to ensure 50%
confl uency on the day of transfection. Knockdown transfections were per-
formed using 80 nM of purifi ed siRNA and Lipofectamine 2000 according
to the manufacturer’s protocol. For the VSV-G transport assay, specifi c tar-
geting siRNA and siGlo Risc free-labeled nonspecifi c siRNA were mixed
(ratio 4:1) with a fi nal concentration of 100 nM. siRNAs controls were from
Dharmacon. The siRNA PKD3 and the siRNA PKD2 are Silencer-validated
siRNAs from Ambion.
ssHRP and PLAP secretion assay
30 h after transfection with siRNA, the cells were cotransfected with the
SS-HRP-Flag and the pSEAP2-basic (Clontech Laboratories, Inc.) plasmid
(carrying PLAP cDNA), using Lipofectamine 2000 (Invitrogen). 30 μl of
extracellular media was harvested 48 h after the initial siRNA transfection.
HRP activity was measured using enhanced chemiluminescence (ECL) as
described previously (Bard et al., 2006). PLAP activity in the medium and
in the cells was measured using the Phospha-Light System (Applied Bio-
systems) following the manufacturer’s protocol. PLAP activity inside the cells
was normalized by total protein concentration and used to normalize both
HRP and PLAP secreted into the medium.
VSV-G transport assay
30 h after transfection with siRNA, the cells were transfected with ts045VSV-
G-GFP construct and cultured at 40°C for 20 h. 100 μg/ml of cycloheximide
was then added before a 2-h incubation at 20°C. After an incubation
at 32°C for 40 min the cells were harvested with a cell dissociation buffer
(Invitrogen) and fi xed with 4% paraformaldehyde. After blocking with PBS
containing 1.5% serum and 0.1% sodium azide, the labeling of surface
VSV-G–GFP was performed for a 30-min incubation at 4°C with the anti-
VSV-G mAb 8G5F11, which is specifi c for the extracellular domain of VSV-G.
After washings with the blocking buffer, the cells were incubated with the
secondary (APC)-labeled anti–mouse IgG antibody (Jackson ImmunoResearch
Laboratories) for 30 min at 4°C. After washing, the cells were analyzed on
a FACScalibur fl ow cytometer (BD Biosciences). The amount of VSV-G pre-
sent at the cell surface (APC positive) of cells transfected with both siRNA
(cy3 positive) and VSV-G (GFP positive) after subtracting the background
was normalized by the GFP intensity.
Transfected 293-T cells or HeLa cells were lysed in 50 mM Tris, pH 7.4,
150 mM NaCl, 1% Triton X-100, and protease inhibitors for 30 min at
4°C. After centrifugation at 13,000 rpm for 10 min, protein concentrations
were measured in the lysates. 100 μg of extracts were incubated with the
primary antibody (1:100) at 4°C and after 2 h, 30 μl of G-Sepharose
beads (GE Healthcare) were added for 1 h. Immobilized proteins were re-
leased by boiling in Laemmli buffer and analyzed by SDS-PAGE. For Flag-
tagged proteins, 100 μg of cell lysates were incubated with anti-Flag M2
affi nity gel and eluted with 3×Flag peptide following the manufacturer’s
In vitro kinase assay
An equal amount of the indicated proteins purifi ed by immunoprecipitation
from transfected 293-T cells was incubated for 10 min at 32°C in a buffer
containing 50 mM Tris-HCl, 30 mM MgCl2, 0,3 mM ATP, 2 mM DTT,
0.2 μM PdBu, and 5 μCi ATPγP32. The reaction was stopped by addition of
6× SDS sample buffer and the samples were processed for SDS-PAGE
In vitro binding assay
After immunoprecipitation with anti-GST antibody from transfected 293-T
cell lysates, the purifi ed GST-tagged proteins bound to the beads were
incubated with equal amount of purifi ed Flag-tagged proteins for 2 h at
4°C in 300 μl of PBS containing 0.1% Triton X-100 and 0.2%BSA. After
extensive washes in the same buffer, the precipitates were eluted in 1×
SDS sample buffer and processed for Western blotting.
PKD-CAAX constructs cloning
PKD-CAAX-CA, PKD-CAAX-WT, and PKD-CAAX-KD were cloned by PCR using
GST-PKD-CA, GST-PKD-WT, and GST-PKD-KD, respectively, as templates as
previously described (Maeda et al., 2001). The CAAX motif was appended
by using a reverse primer containing the nucleotides coding for the CAAX
motif MVLC: 5′-A A A T C T A G A A A G C T T T C A C A T A A C G A G A C A G A G G A T G-
C T G A C A C G C T C A C T G -3′.
24 h after transfection, HeLa cells expressing MannII-GFP grown on cover-
slips were fi xed with 4% formaldehyde in PBS for 10 min, blocked, and
permeabilized with blocking buffer (0.05% Saponin and 0.2% BSA in
PBS) for 20 min. The coverslips were incubated with primary antibodies
diluted in blocking buffer for 2 h, washed, incubated with secondary
antibodies diluted in blocking buffer for 1 h, washed, mounted using Fluor
Save Reagent (Calbiochem), and visualized with a Nikon Eclipse TE2000-U
microscope. Pictures were taken using MetaMorph software and decon-
volved using AutoVisualize+AutoDeblur 9.3 software. The pictures were
then opened in ImageJ v1.37 and Adobe Photoshop.
HeLa cells were depleted of both PKD2 and PKD3 as described above.
The cells were transfected with SS-HRP and processed for immunoelectron
microscopy, and HRP was visualized by staining with DAB and H2O2 as
described previously (Polishchuk et al., 2000). HeLa cells were transfected
with GST-tagged PKD WT, PKD-CAAX-CA, PKD-CAAX-WT, or PKD-CAAX-
KD. Cells were fi xed in the mixture of 4% paraformaldehyde and 0.5%
glutaraldehyde, washed, and labeled with anti-TGN46–specifi c antibody
followed by an antibody conjugated with peroxidase as described previ-
ously (Polishchuk et al., 2000). Then cells were incubated with polyclonal
antibody against GST and subsequently with Nanogold-conjugated Fab
fragment of anti–rabbit IgG. Nanogold particles were enhanced using the
manufacturer’s kit (Nanoprobes). ST-HRP-expressing cells were incubated
directly with DAB and H2O2 (Polishchuk et al., 2000) and then labeled
with anti-GST antibody as described above. Cells were then embedded in
Epon 812 and thin sections visualized in a Tecnai-12 electron microscope
(FEI, Philips). Images were taken using an Ultra View CCD digital camera.
Morphometric analysis of Golgi stacks was performed in 20 cells for each
experimental condition using the ANALYSIS software.
Tubular profi les were defi ned as HRP-positive structures with length
twice or more higher than thickness.
The statistical signifi cance of the difference between means was determined
using the t test. Differences were considered signifi cant at P < 0.01.
This work was supported by National Institutes of Health grants (GM46224
and GM53747), and a senior investigator award from the Sandler’s program
for Asthma research to V. Malhotra. C. Bossard was supported by a Human
Frontier Science Program postdoctoral fellowship. R.S. Polishchuk was sup-
ported by Telethon EM Facility Grant GTF05007 and Telethon Rersearch
Submitted: 27 March 2007
Accepted: 15 November 2007
Audhya, A., M. Foti, and S.D. Emr. 2000. Distinct roles for the yeast phospha-
tidylinositol 4-kinases, Stt4p and Pik1p, in secretion, cell growth, and
organelle membrane dynamics. Mol. Biol. Cell. 11:2673–2689.
Table I. Quantitation of the ST-HRP– and TGN-positive vesicles
PKD-WT PKD-CAAX-WTPKD-CAAX-CA PKD-CAAX-KD
0.30 ± 0.48
1.50 ± 0.50
5.90 ± 1.07
7.22 ± 2.40
7.40 ± 2.57
6.85 ± 2.80
0.45 ± 0.25
1.20 ± 0.90
Bars represent mean ± SD of the number of vesicles and show that they originate from the TGN in cells expressing PKD-CAAX-CA and PKD-CAAX-WT.
MEMBRANE FISSION BY PKD • BOSSARD ET AL.1131
Bard, F., and V. Malhotra. 2006. The formation of TGN-to-plasma-membrane
transport carriers. Annu. Rev. Cell Dev. Biol. 22:439–455.
Bard, F., L. Casano, A. Mallabiabarrena, E. Wallace, K. Saito, H. Kitayama, G.
Guizzunti, Y. Hu, F. Wendler, R. Dasgupta, et al. 2006. Functional ge-
nomics reveals genes involved in protein secretion and Golgi organization.
Baron, C.L., and V. Malhotra. 2002. Role of diacylglycerol in PKD recruitment
to the TGN and protein transport to the plasma membrane. Science.
Choy, E., V.K. Chiu, J. Silletti, M. Feoktistov, T. Morimoto, D. Michaelson, I.E.
Ivanov, and M.R. Philips. 1999. Endomembrane traffi cking of ras: the
CAAX motif targets proteins to the ER and Golgi. Cell. 98:69–80.
Diaz Anel, A.M., and V. Malhotra. 2005. PKCeta is required for beta1gamma2/
beta3gamma2- and PKD-mediated transport to the cell surface and the
organization of the Golgi apparatus. J. Cell Biol. 169:83–91.
Fugmann, T., A. Hausser, P. Schoffl er, S. Schmid, K. Pfi zenmaier, and M.A.
Olayioye. 2007. Regulation of secretory transport by protein kinase
D-mediated phosphorylation of the ceramide transfer protein. J. Cell Biol.
Godi, A., A. Di Campli, A. Konstantakopoulos, G. Di Tullio, D.R. Alessi, G.S.
Kular, T. Daniele, P. Marra, J.M. Lucocq, and M.A. De Matteis. 2004.
FAPPs control Golgi-to-cell-surface membrane traffi c by binding to ARF
and PtdIns(4)P. Nat. Cell Biol. 6:393–404.
Hausser, A., P. Storz, S. Martens, G. Link, A. Toker, and K. Pfi zenmaier. 2005.
Protein kinase D regulates vesicular transport by phosphorylating and
activating phosphatidylinositol-4 kinase IIIbeta at the Golgi complex.
Nat. Cell Biol. 7:880–886.
Hausser, A., G. Link, M. Hoene, C. Russo, O. Selchow, and K. Pfi zenmaier.
2006. Phospho-specifi c binding of 14-3-3 proteins to phosphatidylinositol
4-kinase III beta protects from dephosphorylation and stabilizes lipid
kinase activity. J. Cell Sci. 119:3613–3621.
Iglesias, T., R.T. Waldron, and E. Rozengurt. 1998. Identifi cation of in vivo phos-
phorylation sites required for protein kinase D activation. J. Biol. Chem.
Jamora, C., P.A. Takizawa, R.F. Zaarour, C. Denesvre, D.J. Faulkner, and V.
Malhotra. 1997. Regulation of Golgi structure through heterotrimeric G
proteins. Cell. 91:617–626.
Jamora, C., N. Yamanouye, J. Van Lint, J. Laudenslager, J.R. Vandenheede, D.J.
Faulkner, and V. Malhotra. 1999. Gbetagamma-mediated regulation of
Golgi organization is through the direct activation of protein kinase D.
Liljedahl, M., Y. Maeda, A. Colanzi, I. Ayala, J. Van Lint, and V. Malhotra. 2001.
Protein kinase D regulates the fi ssion of cell surface destined transport
carriers from the trans-Golgi network. Cell. 104:409–420.
Lipardi, C., L. Nitsch, and C. Zurzolo. 2000. Detergent-insoluble GPI-anchored
proteins are apically sorted in fi scher rat thyroid cells, but interference
with cholesterol or sphingolipids differentially affects detergent insolu-
bility and apical sorting. Mol. Biol. Cell. 11:531–542.
Lisanti, M.P., A. Le Bivic, A.R. Saltiel, and E. Rodriguez-Boulan. 1990.
Preferred apical distribution of glycosyl-phosphatidylinositol (GPI) an-
chored proteins: a highly conserved feature of the polarized epithelial cell
phenotype. J. Membr. Biol. 113:155–167.
Maeda, Y., G.V. Beznoussenko, J. Van Lint, A.A. Mironov, and V. Malhotra.
2001. Recruitment of protein kinase D to the trans-Golgi network via the
fi rst cysteine-rich domain. EMBO J. 20:5982–5990.
Polishchuk, R.S., E.V. Polishchuk, P. Marra, S. Alberti, R. Buccione, A. Luini,
and A.A. Mironov. 2000. Correlative light-electron microscopy reveals
the tubular-saccular ultrastructure of carriers operating between Golgi
apparatus and plasma membrane. J. Cell Biol. 148:45–58.
Rykx, A., L. De Kimpe, S. Mikhalap, T. Vantus, T. Seufferlein, J.R. Vandenheede,
and J. Van Lint. 2003. Protein kinase D: a family affair. FEBS Lett.
Sanchez-Ruiloba, L., N. Cabrera-Poch, M. Rodriguez-Martinez, C. Lopez-
Menendez, R.M. Jean-Mairet, A.M. Higuero, and T. Iglesias. 2006.
Protein kinase D intracellular localization and activity control kinase
D-interacting substrate of 220-kDa traffi c through a postsynaptic
density-95/discs large/zonula occludens-1-binding motif. J. Biol.
Sutterlin, C., R. Polishchuk, M. Pecot, and V. Malhotra. 2005. The Golgi-associated
protein GRASP65 regulates spindle dynamics and is essential for cell
division. Mol. Biol. Cell. 16:3211–3222.
Takizawa, P.A., J.K. Yucel, B. Veit, D.J. Faulkner, T. Deerinck, G. Soto, M.
Ellisman, and V. Malhotra. 1993. Complete vesiculation of Golgi mem-
branes and inhibition of protein transport by a novel sea sponge metabo-
lite, ilimaquinone. Cell. 73:1079–1090.
Walch-Solimena, C., and P. Novick. 1999. The yeast phosphatidylinositol-4-OH
kinase pik1 regulates secretion at the Golgi. Nat. Cell Biol. 1:523–525.
Wright, L.P., and M.R. Philips. 2006. Thematic review series: lipid posttrans-
lational modifi cations. CAAX modifi cation and membrane targeting of
Ras. J. Lipid Res. 47:883–891.
Yeaman, C., M.I. Ayala, J.R. Wright, F. Bard, C. Bossard, A. Ang, Y. Maeda,
T. Seufferlein, I. Mellman, W.J. Nelson, and V. Malhotra. 2004. Protein
kinase D regulates basolateral membrane protein exit from trans-Golgi
network. Nat. Cell Biol. 6:106–112.