TRPV1 expression-dependent initiation and regulation of filopodia.

TRPV1 expression-dependent initiation and regulation of filopodia
C. Goswami and T. Hucho
Signal Transduction in Pain and Mental Retardation, Department of Human Molecular Genetics, Max Planck Institute for Molecular
Genetics, Berlin, Germany
Abstract
Transient receptor potential vanilloid subtype 1 (TRPV1), a
non-selective cation channel, is present endogenously in
dorsal root ganglia (DRG) neurons. It is involved in the rec-
ognition of various pain producing physical and chemical
stimuli. In this work, we demonstrate that expression of
TRPV1 induces neurite-like structures and filopodia and that
the expressed protein is localized at the filopodial tips.
Exogenous expression of TRPV1 induces filopodia both in
DRG neuron-derived F11 cells and in non-neuronal cells, such
as HeLa and human embryonic kidney (HEK) cells. We find
that some of the TRPV1 expression-induced filopodia contain
microtubules and microtubule-associated components, and
establish cell-to-cell extensions. Using live cell microscopy,
we demonstrate that the filopodia are responsive to TRPV1-
specific ligands. But both, initiation and subsequent cell-to-cell
extension formation, is independent of TRPV1 channel activ-
ity. The N-terminal intracellular domain of TRPV1 is sufficient
for filopodial structure initiation while the C-terminal cyto-
plasmic domain is involved in the stabilization of microtubules
within these structures. In addition, exogenous expression of
TRPV1 results in altered cellular distribution and in enhanced
endogenous expression of non-conventional myosin motors,
namely myosin IIA and myosin IIIA. These data indicate a
novel role of TRPV1 in the regulation of cellular morphology
and cellular contact formation.
Keywords: cytoskeletal rearrangement, filopodia, micro-
tubules, myosin, neurite, TRPV.
J. Neurochem. (2007) 10.1111/j.1471-4159.2007.04846.x
Neurons are distinct from most other cells by their extreme
morphology. Axonal and dendritic structures show particular
cytoskeletal composition, morphology, and function.
Through these structures neurons are able to establish
complex functional networks by connecting a large number
of cells. The precise formation of cell-to-cell connections
depends on many factors in a multi-step process, namely:
‘protrusion,’ i.e., initiation of small membranous structures
(initiation of filopodia), ‘engorgement,’ i.e., advancement of
microtubules within the protrusion thereby facilitating trans-
port of membranous organelles (Burmeister et al. 1991), and
finally ‘consolidation,’ i.e., disassembly of actin and shrink-
ing of membranes around microtubules [reviewed in Dent
and Gertler (2003); Jontes and Smith (2000), Song and Poo
(2001)]. After extension toward neighboring cells, contact
formation including the establishment of synapses is initi-
ated.
Filopodial structures extending from axonal growth cones,
from dendrites and from the cell body are similar in many
aspects. All are formed by rearrangement of the actin
cytoskeleton and are involved in ‘chemotaxis’ leading to the
formation of cellular contacts [reviewed in Faix and Rottner
(2006); Small and Resch (2005); Mitchison and Cramer
(1996); Small et al. (2002); Wood and Martin (2002); da
Silva and Dotti (2002)]. Although several cytoskeletal and
regulatory proteins are known to play a role in filopodial
initiation and regulation, a large number of filopodial
proteins remain unidentified.
The complex process of cell–cell contact formation by
developing filopodia requires the involvement of structural
proteins and regulatory factors including intracellular Ca2+. It
has been shown that filopodial Ca2+-transients and Ca2+-
spikes determine the nature and function of these structures
(Goldberg and Grabham 1999; Gomez and Spitzer 2000;
Conklin et al. 2005). Though the participation of a variety of
Ca2+-binding proteins, and Ca2+ channels in the regulation of
filopodial dynamics has been established, the identity of
individual calcium channels and their contribution in this
complex signaling event remains to be elucidated (Gomez
et al. 2001; Robles et al. 2003; Henley and Poo 2004;
Lohmann et al. 2005; Gomez and Zheng 2006).
Previously we showed that TRPV1, a Ca2+-permeable
non-selective cation channel, interacts with tubulin via its
C-terminal cytoplasmic domain (Goswami et al. 2004). We
demonstrated that the C-terminal cytoplasmic domain of
Received April 19, 2007; revised manuscript received June 25, 2007;
accepted June 27, 2007.
Address correspondence and reprint requests to C. Goswami, Signal
Transduction in Pain and Mental Retardation, Department of Human
Molecular Genetics, Max Planck Institute for Molecular Genetics,
Ihnestraße. 73, 14195, Berlin, Germany.
E-mail: goswami@molgen.mpg.de
Abbreviations used: DRG, dorsal root ganglia; GFP, green flourescent
protein; HEK, human embryonic kidney; PBS, phosphate-buffered
saline; PFA, paraformaldehyde; TRPV1, Transient receptor potential
vanilloid subtype 1.
Journal of Neurochemistry, 2007 doi:10.1111/j.1471-4159.2007.04846.x
� 2007 The Authors
Journal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2007) 10.1111/j.1471-4159.2007.04846.x 1

TRPV1 contains two novel tubulin-binding motifs, each of
which can bind tubulin independently (Goswami et al.
2007a). The interaction with the C-terminal domain of
TRPV1 stabilizes microtubuli under certain conditions
(Goswami et al. 2004, 2006). In contrast, activation of
TRPV1 results in disassembly of microtubules (Goswami
et al. 2006). TRPV1 is localized in neurites and growth
cones including axonal filopodia and regulates the mor-
phology and motility of growth cones (Goswami et al.
2007b). Activation of TRPV1 results in rapid retraction of
growth cones as well as in varicosity formation along axons
(Goswami et al. 2007b). We now investigated if TRPV1
plays a role in the process of neurite and filopodia
formation.
Materials and methods
Reagents and antibodies
Resiniferatoxin (RTX), 5¢-iodoresiniferatoxin (5¢-IRTX), were
purchased from Sigma–Aldrich (Deisenhofen, Germany). Mouse
monoclonal antibody against beta tubulin class III (clone
SDL.3D10), mouse monoclonal antibody against a-tubulin (clone
DM1A), antibody against acetylated tubulin (clone 6-11B-1) and
antibody against neurofilament 160 (clone NN18) were purchased
from Sigma (Deisenhofen, Germany). Rat monoclonal antibody
against tyrosinated tubulin (clone YL1/2) was purchased from
AbCam ltd. (Cambridge, UK). Rabbit polyclonal antibody against
detyrosinated tubulin was purchased from Chemicon. Rabbit
polyclonal antibody against the N-terminal sequence of TRPV1
was purchased from Affinity Bioreagents (Golden, CO, USA). The
goat polyclonal antibody against the C-terminal sequence of
TRPV1 was purchased from Santa Cruz Biotechnology (Santa
Cruz, CA, USA). Alexa-594-labeled phalloidin, alexa-594-labeled
anti-rat secondary IgG antibody, alexa-594-labeled anti-mouse
secondary IgG antibodies were purchased from Molecular Probe
(Invitrogen, Karlsruhe, Germany). Cy2-labeled-anti-goat and Cy2-
labeled anti-rabbit IgG antibodies were purchased from Dia nova
(Hamburg, Germany). Affinity purified rabbit polyclonal antibody
against MyoIIIA (5525) and rabbit polyclonal antibodies against
myosin IIA (Sigma) are kindly provided by Dr Andrea Dose
(Berkeley, USA).
Constructs
For heterologous expression in mammalian cells, the full-length rat
TRPV1 cDNA sub-cloned in a pcDNA3.1 vector was used (Jahnel
et al. 2001; Goswami et al. 2004, 2006, 2007b). For expression of
the C-terminally green fluorescent protein-fused with TRPV1
(TRPV1-GFP), a cDNA fragment encoding rat TRPV1 was
amplified by PCR, using 5¢-ATGGAACAACGGGCTAGCTT-3¢
and 5¢-TCTCCCCTGGGACCATGGAA-3¢ primers and sub-cloned
into a pCDNA3.1/CT-GFP-TOPO vector (Invitrogen) (Jahnel 2005).
For preparing the TRPV1-DCt construct, the N-terminal sequence
including the transmembrane sequences of TRPV1 was amplified by
using 5¢-ATGGAACAACGGGCTAGCTT-3¢ and 5¢-GTGCAATC-
TTGTTGACGGTC-3¢ primers, sub-cloned into the pCDNA3.1
vector (Jahnel 2005). For preparing the TRPV1-Nt construct, the
N-terminal sequence of TRPV1 was amplified by using 5¢-GCG-
CGAATTCATGGAACAACGGGCTAG-3¢ and 5¢-GCGCTCTA-
GATTACTTGACAAATCTG-3¢ primers, sub-cloned into the
pCDNA3.1 vector (Jahnel 2005). For preparing the TRPV1-DNt
construct, the transmembrane region with the C-terminal sequences
of TRPV1 was amplified by using 5¢-ATGCTCCTACAGGACAA-
GTGGGAC-3¢ and 5¢ CTTTCTCCCCTGGGACCATGG-3¢ prim-
ers, sub-cloned into the pCDNA3.1 vector (Jahnel 2005).
Cell culture and transfection
F11 cells were cultured in Ham’s F12 medium (Invitrogen)
supplemented with 20% fetal calf serum (Invitrogen). HeLa and
HEK cells were grown in Dulbecco’s modified Eagle’s medium
supplemented with 10% fetal calf serum. Lipofectamine (Invitro-
gen) was used for transient transfection.
TRPV1 activation and live cell imaging
For visualizing the effect of TRPV1 activation, F11 cells were
seeded on glass cover slips and TRPV1-GFP was transiently
expressed. Two days after transfection, F11 cells with complete
medium were imaged at room temperature (25�C). In experiments
aiming to block the basal level of TRPV1 activity, its antagonist 5¢-
iodoresiniferatoxin (5¢-IRTX, 1 lm) was used. For experiments
aiming to activate TRPV1, resiniferatoxin (RTX, 100 nmol/L) was
added.
Immunocytochemistry
For immunocytochemical analysis, cells were fixed with 2%
paraformaldehyde (PFA). To avoid morphological changes of the
dynamic filopodial structures due to phosphate-buffered saline
(PBS) wash, in some experiments, an equal volume of 4% PFAwas
carefully added to the medium without disturbing the cell culture.
The cells were fixed with PFA for 5 min, and subsequently
permeabilized with 0.1% Triton X-100 in PBS (5 min). Subse-
quently, the cells were quenched with 2% glycine in PBS and the
cells were blocked either with 5% normal goat serum or bovine
serum albumin. The primary antibodies were used at the following
dilutions: rabbit polyclonal anti-TRPV1 (1 : 1000), goat polyclonal
anti-TRPV1 (1 : 1000), rat monoclonal anti-tyrosinated tubulin
(1 : 1000), mouse monoclonal anti-acetylated tubulin (1 : 1000),
mouse monoclonal anti-alpha tubulin (1 : 1000), rabbit polyclonal
anti-detyrosinated tubulin (1 : 1000), mouse monoclonal anti-
neurofilament (1 : 1000), rabbit polyclonal anti myosin IIa
(1 : 1000) and rabbit polyclonal anti myosin IIIa (1 : 1000). All
primary antibodies were incubated for 1 h at 25�C in PBST buffer
(PBS supplemented with 0.1% Tween-20). All images were
taken on a confocal laser-scanning microscope (Zeiss Axiovert
100 mol/L) with a 63·-objective and analyzed with the Zeiss LSM
image examiner software.
Estimation of neurite lengths
For quantification of neurite lengths as well as other neurite
properties, cells were fixed with 2% PFA 48 h after transfection. The
cells were subsequently immunostained for TRPV1 and tubulin.
Randomly selected fields containing both, transfected and non-
transfected cells, were imaged. Each cell was scored for the presence
or absence of neurites. Neurite-lengths of each cell were measured
manually by using the Zeiss LSM image examiner software.
2 C. Goswami and T. Hucho
Journal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2007) 10.1111/j.1471-4159.2007.04846.x
� 2007 The Authors

In absence of any neurite, the value was considered as zero. Values
were plotted in ascending order and the graphs were plotted using
Microsoft exel and Origin 7.1 software.
Results
Exogenous-expression of TRPV1 induces changes in
cellular morphology
To study TRPV1, we chose F11 cells. These cells are derived
from rat DRG-neurons by fusion with mouse neuroblastoma
cells (Platika et al. 1985). Due to their neuronal origin, F11
cells resemble DRG neurons in many aspects (Platika et al.
1985). For this reason, we refer to long neurite-like structures
of these cells as neurites. TRPV1 is expressed endogenously
in dorsal root ganglia (DRG) neurons, but not in F11 cells.
Therefore, these cells can be used as a model system to study
TRPV1 function.
Previously we reported that TRPV1 localizes to the plasma
membrane when expressed in F11 cells (Goswami et al.
2004, 2006, 2007b). We now report that these cells often
become elongated after transient expression of TRPV1
(Fig. 1a). In addition, though F11 cells are heterogeneous
in morphology, under our culture conditions exogenous
expression of TRPV1 induces extensive neurite growth in the
majority of transfected cells. We quantified the TRPV1-
induced neuritogenesis by counting neurites per cell and by
measuring the neurite lengths. On average there are eight
times more neurites in transfected cells as compared to non-
transfected cells. And, transfected cells not only show a clear
increase in the number of neurites, but also in their average
length (Fig. 1). By immunofluorescence analysis we ob-
served that TRPV1-expressing cells often show many short
and few long neurites (Figs 1 and 2). The various neurite
sizes are reminiscent of the various stages of neurite
‘budding’, ‘formation’ and ‘polarization’. Furthermore, we
observed that TRPV1 is localized in these neurites.
We observed that due to exogenous expression of TRPV1,
many small protrusions emerge from the neurite shafts and
from the cell bodies. These are variable in shape and length. In
the literature, such protrusions have been termed ‘spikes’,
‘microspikes’, ‘protrusions’, ‘filopodia’, ‘retraction fibres’,
‘veil’, ‘dendritic spine’ etc. Often they are considered to be
distinct entities. However, recent report suggests that all these
different forms have a similar actin-based structure and that
Fig. 1 Exogenous expression of TRPV1 promotes neurite extension.
(a) Immunofluorescence confocal images (superimposed on the
phase contrast image) of F11 cells expressing TRPV1 are shown.
Cells were immunostained for TRPV1 (green), tubulin (red), and DNA
(blue). TRPV1 expressing cells (indicated by white arrow) become
much elongated and reveal distinct morphological changes. The table
at the right side depicts different parameters of neurite-outgrowth in
transfected and non-transfected cells. (b) Graphical representation of
neurite-lengths per cell. Total neurite length (i) and maximum neurite
length (ii) of transfected cells (n = 114) and non-transfected cells
(n = 155) were calculated and plotted in ascending order. Neurite
lengths (in lm, Y-axis) are indicated by blue triangles (transfected
cells) and green asterisks (non-transfected cells). Rank-order of the
cells are indicated along X-axis. Cells with no neurites are considered
as zero neurite length.
TRPV1 expression-induced filopodia 3
� 2007 The Authors
Journal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2007) 10.1111/j.1471-4159.2007.04846.x

these structures are interconvertible (Svitkina et al. 2003).
Therefore, in this study we term them ‘filopodia’ (see Fig. 2).
We observed TRPV1 to be localized in these structures. We
focused on TRPV1-containing filopodia and characterized
their molecular architecture, initiation and function.
TRPV1 is accumulated at the tip of filopodia
While some of the filopodial structures have pointed tips,
some show a distinct bulbous ‘head’ on a thin ‘neck’. Often
we observed that TRPV1 is enriched mostly in the ‘head’
structure, i.e., at the filopodial tip (Fig. 2). We probed these
structures with alexa 594-labeled phalloidin to visualize the
status of polymerized F-actin. We observed the presence of
actin all along the filopodial shaft, but their tips often lack
significant amounts of phalloidin staining, indicating that
polymerized F-actin is not present in large quantities
(Fig. 2b). Filopodial structures with a bulbous head and
with less polymerized actin matches well with descriptions of
‘retracting filopodial structure’ (Medalia et al. 2007).
To confirm that the presence of TRPV1 at the tip of
filopodia is not a ‘fixation artifact’, we performed live cell
imaging using TRPV1-GFP. When expressed transiently in
F11 cells, we observed several filopodia to develop from cell
bodies (Fig. 2c-i.) and also from the much longer neurites
(Fig. 2c-ii). Like in PFA-fixed cells, TRPV1-GFP was
localized also in live cells at these filopodial structures.
And, indeed, TRPV1-GFP was also enriched at their tips,
suggesting that TRPV1 may be part of the filopodial tip
complex (Fig. 2c-iii).
Some, but not all TRPV1-containing filopodia contain
microtubules
TRPV1 expressing cells reveal increased neurite as well as
extensive filopodia formation. Potentially, some of these
filopodia may mature into neurites. To address this, we
characterized their cytoskeleton constituents by immunoflu-
orescence microscopy. Staining with alexa-594-labeled phal-
loidin we confirmed that the majority contains F-actin though
at low level (data not shown). We stained the TRPV1
expressing cells for other major cytoskeleton component,
namely tubulin. While filopodial structures generally contain
actin filaments but no microtubules, the neurite stalk in
(a) (b)
(c)
Fig. 2 Exogenous expression of TRPV1 induces filopodial structures
leading to a change in cellular morphology. (a) Shown is a confocal
immunofluorescence image of an F11 cell transiently expressing
TRPV1 (green). The immunofluorescence image is superimposed on
the phase contrast image. Scale bar: 20 lm. (b) The enlarged con-
focal immunofluorescence images depicting detail morphology of the
filopodia of F11 cells expressing TRPV1 are shown. Cells were stained
for TRPV1 (green) and phalloidin (red). Fluorescence images were
superimposed on the phase contrast images (left side). Intensities of
TRPV1 and phalloidin are represented in false rainbow color (highest
intensity in red and the lowest intensity in blue). A close arrow indi-
cates an intense immunoreactivity of TRPV1 at the filopodial tip. In
comparison, the presence of F-actin at the tip is significantly lower.
Scale bar: 5 lm. (c) The confocal fluorescence live images of F11
cells expressing TRPV1-GFP are shown. Several filopodia originating
from the cell body (i) and from the neurites (ii) are shown. The phase
contrast and fluorescence images are superimposed (i and ii). Inten-
sities of the TRPV1-GFP are represented in false rainbow color (iii).
Intense localization of TRPV1-GFP is observed at the tip (indicated by
white arrows). Scale bar: 20 lm (for i) and 10 lm (for ii and iii).
4 C. Goswami and T. Hucho
Journal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2007) 10.1111/j.1471-4159.2007.04846.x
� 2007 The Authors

contrast is known to be composed of microtubule bundles.
Indeed, we observed that a small number of these structures
contained tubulin immunoreactivity (Fig. 3). Microtubules
are present also at the very end of the filopodial structures
including the TRPV1-enriched ‘head’ region (Fig. 3 and 4).
Different post-translational modifications of tubulin relate
to the stability, dynamic, and age of the microtubules
(Gundersen et al. 1984; Kreis 1987; Wehland and Weber
1987; MacRae 1997; Idriss 2000; Dent and Gertler 2003;
Westermann and Weber 2003). Using modification-specific
antibodies we probed for the presence of post-translationally
modified tubulins in filopodia. Indeed, many contained post-
translationally modified tubulin. We observed the presence of
tyrosinated tubulin (a marker for dynamic microtubules),
detyrosinated tubulin (a marker for stable microtubules), and
beta tubulin subtype III (a neuron-specific tubulin), albeit at a
low level (data not shown). Furthermore, a significant
amount of acetylated tubulin and neuron-specific MAP2a/b
(marker for stable microtubules) was detected (Fig. 4). This
proves the presence of both dynamic and stable microtubules
within TRPV1-induced filopodia. In addition, some of these
structures contain significant amounts of neurofilament (see
below).
TRPV1 expression-induced filopodia are dynamic in
nature and are involved in cellular contact formation
To better understand the nature of TRPV1-induced filopodia,
we expressed TRPV1-GFP and performed live cell imaging.
We observed that the filopodia are dynamic and motile even
in the absence of TRPV1 activation by specific agonists.
They change their length and orientation with time, but can
also remain stable for some time. Next we tested if TRPV1
containing filopodia respond to resiniferatoxin (RTX), a
TRPV1-specific agonist. We observed that the activation of
TRPV1 by RTX causes rapid changes in filopodial dynamics
in the majority of cells. These immediate changes include
increased filopodial bending and/or buckling as well as
disappearance of filopodia soon after addition of RTX.
Nevertheless, also some new filopodia appeared at the same
or at other sites (Fig. 5).
We observed that the neighboring cells influence the
distribution and number of small TRPV1 containing filopo-
dia. They are often generated and extended unevenly, mostly
in an orientation perpendicular to the main neuritic shaft,
especially in the vicinity of a neighboring cell and then being
strikingly orientated towards this cell. Unlike classical
filopodial structures, these few cell-to-cell extension-forming
TRPV1-containing filopodia showed tubulin immunoreac-
tivity (Fig. 6a). Some of them contained neurofilaments, too
(Fig. 6b). Thus, the presence of microtubules, different
(a)
(b)
(c)
(d)
Fig. 3 Some but not all TRPV1 expression-induced filopodia contain
microtubules. Shown are the immunofluorescence confocal images of
F11 cells (a) expressing TRPV1 (T) and an enlarged area of the same
cell (b–d). Cells were immunostained for TRPV1 (green) and tubulin
(red). Confocal immunofluorescence images are superimposed on the
phase contrast images. The presence of extensive neurites and filo-
podia in TRPV1 expressing cell (T) differ from the non-transfected cell
(NT). An even more enlarged area (indicated by white dashed line in d)
reveals further details. Some but not all TRPV1 expression-induced
filopodia contain microtubules (white arrow). Filopodia with or without
microtubules are indicated by close and open arrow, respectively.
Scale bar: 20 lm (for a), 10 lm (b–d) and 2 lm (for enlarged areas at
right).
TRPV1 expression-induced filopodia 5
� 2007 The Authors
Journal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2007) 10.1111/j.1471-4159.2007.04846.x