Address correspondence and reprint requests to F. Hucho, Freie Universität Berlin, Institut für Chemie/Biochemie, Thielallee 63, 14195 Berlin, Germany. E-mail: firstname.lastname@example.org
The transmission of pain signalling involves the cytoskeleton, but mechanistically this is poorly understood. We recently demonstrated that the capsaicin receptor TRPV1, a non-selective cation channel expressed by nociceptors that is capable of detecting multiple pain-producing stimuli, directly interacts with the tubulin cytoskeleton. We hypothesized that the tubulin cytoskeleton is a downstream effector of TRPV1 activation. Here we show that activation of TRPV1 results in the rapid disassembly of microtubules, but not of the actin or neurofilament cytoskeletons. TRPV1 activation mainly affects dynamic microtubules that contain tyrosinated tubulins, whereas stable microtubules are apparently unaffected. The C-terminal fragment of TRPV1 exerts a stabilizing effect on microtubules when over-expressed in F11 cells. These findings suggest that TRPV1 activation may contribute to cytoskeleton remodelling and so influence nociception.
TRPV1 forms a homotetramer with six transmembrane sequences per monomer. Both the N- and the C-terminus of TRPV1 are located in the cytoplasm. They are subject to regulatory post-translational modifications (reviewed by Cortright and Szallasi 2004), and supposedly interact with a number of proteins. Several reports suggest that cytoskeletal structures play an important role in pain pathways (Bhave and Gereau 2003; Dina et al. 2003). Microtubule-active reagents, such as vinca alkaloids and taxol, are capable of changing nociceptor morphology and responsiveness (Topp et al. 2000; Alessandri Haber et al. 2004).
We recently reported a direct interaction of the C-terminal domain of TRPV1 with tubulins and hypothesized that the microtubule cytoskeleton is a downstream effector of TRPV1 activation (Goswami et al. 2004). We now aim to further substantiate this hypothesis. We report immunocytochemical and biochemical data indicating that activation of TRPV1 results in selective destabilization of dynamic microtubules, whereas the C-terminal fragment of TRPV1 alone clearly exerts a stabilizing effect on the microtubules. The actin or neurofilament cytoskeleton, however, is not affected by TRPV1 activation.
Materials and methods
Antibodies and reagents
Digitonin was purchased from Calbiochem (San Diego, CA, USA). The TRPV1 agonist resiniferatoxin (RTX), antagonist 5′-iodo-resiniferatoxin (I-RTX) and the microtubule-depolymerizing drug nocodazole were purchased from Sigma (Deisenhofen, Germany). Mouse monoclonal antibodies anti-α-tubulin (clone DM1A), anti-β-tubulin (clone D66), anti-β-tubulin class III (clone SDL.3D10), anti-acetylated tubulin (clone 6-11B1), anti-tyrosinated tubulin (clone TUB1A2), anti-polyglutamylated tubulin (clone B3), anti-γ-tubulin (clone GTU-88) and anti-160-Kd neurofilament (clone NN18) were all purchased from Sigma. The rat monoclonal antibody YL1/2 was purchased from AbCam Ltd (Cambridge, UK). Mouse monoclonal anti-actin antibody (clone JLA20) was purchased from Oncogene (Cambridge, MA, USA). Affinity-purified rabbit polyclonal antibody against de-tyrosinated tubulin (glu tubulin), mouse monoclonal antibody against the 200-kDa neurofilament (clone RT97) and mouse monoclonal antibody against tau (clone Tau-1 PC1C6) were purchased from Chemicon (Chandlers Ford, UK). Rabbit polyclonal anti-N-terminal TRPV1 antibody and the respective blocking peptide (sequence M1EQRASLDSEESESPPQENSC21, corresponding to the first 21 amino acid residues of TRPV1, were from Affinity Bio Reagents (Golden, CO, USA) and from Alexis Biochemicals (San Diego, CA, USA) respectively. Goat polyclonal anti-TRPV1 antibody raised against the C-terminus of TRPV1 was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Alexa-594-labelled phalloidin, alexa-594-labelled anti-rat IgG secondary antibody and alexa-594-labelled anti-mouse IgG secondary antibody were purchased from Molecular Probes (Invitrogen, Karlsruhe, Germany). Cy2-labelled anti-goat and Cy2-labelled anti-rabbit IgG were purchased from ; (Hamburg, Germany).
For mammalian expression, the full-length rat TRPV1 cDNA subcloned in pcDNA3.1 vector was used (Jahnel et al. 2001). For the expression of the C-terminus only, a cDNA fragment representing only the C-terminal cytoplasmic domain (amino acids 681–838) from rat TRPV1 was amplified by PCR using the primers 5′-ATGGGTGAGACCGTCAACAA-3′ and 5′-TTATTTCTCCCCTGGGACCA-3′, and subcloned into the vector pcDNA3.1 (Jahnel 2005).
Generation of a stable TRPV1-expressing F11 cell line
The cDNA encoding TRPV1 was subcloned into pBICD4 (Liu et al. 2000), employing the EcoRI and NotI restriction sites of the vector. F11 cells were transduced with retroviral particles obtained from a triple transfection of human embryonic kidney (HEK)293T cells with plasmids BICD4-TRPV1, pVPack-GP and pVPack-eco (Stratagene, La Jolla, CA, USA). CD4-positive cells were stained using a phycoerythrin-conjugated CD4 antibody (clone EDU-2; Dianova) and isolated by flow cytometry on a Becton Dickinson FACS Vantage cell sorter (Heidelberg, Germany). Expression of TRPV1 in this cell line was confirmed by western blot analysis and immunofluorescence analysis. This cell line is subsequently referred to as TRPV1-F11 cells.
Cell culture and transfection
F11 cells and TRPV1-F11 cells were cultured in Ham's F12 medium (Invitrogen) supplemented with 20% fetal calf serum (Invitrogen). HEK cells were maintained in Dulbecco's modified Eagle's medium with 10% fetal calf serum. Cells were maintained in a humidified atmosphere that contained 5% CO2 at 37°C. For transient transfection, lipofectamine (Invitrogen) was used according to the manufacturer's instructions.
F11, TRPV1-F11 and HEK cells were grown and transfected on glass coverslips. Two days after seeding or transfection, the cells were fixed with either 2% paraformaldehyde at room temperature (25°C) or 80% methanol in phosphate-buffered saline (PBS) at − 20°C for 10 min, permeabilized with 0.4% Triton X-100 in PBS for 5 min, followed by incubation with 100 mm glycine dissolved in PBS for 1 h. The cells were blocked with 5% normal goat serum or bovine serum albumin. After incubating the cells with the primary antibody for 1 h at room temperature, the cells were washed three times with PBS containing 0.1% Tween 20 (PBST) and incubated with secondary antibody diluted in PBST. The coverslips were mounted on to glass slides with fluromount G (Southern biotech, Eching, Germany). All anti-tubulin staining reported here was done with the YL1/2 antibody unless stated otherwise. The mouse monoclonal antibody against β-tubulin was used to study the effect of C-terminal sequence of TRPV1 (TRPV1-Ct) on microtubule stabilization. Alexa-594-labelled phalloidin was used to visualize the actin cytoskeleton. Images were taken on a confocal laser scanning microscope (Axiovert 100M; Zeiss, Berlin, Germany) with a 63 × objective and analysed using Zeiss LSM image examiner software.
TRPV1 activation assay
In order to visualize the effect of TRPV1 activation on the cytoskeleton, F11 cells expressing TRPV1 transiently or TRPV1-F11 cells were grown on glass coverslips for 2 days. Cells were washed gently with Hank's balanced salt solution (HBSS) Invitrogen) at room temperature, incubated with HBSS buffer supplemented with 1 mm CaCl2 and RTX (100 nm) for 1 min, and either fixed immediately or further extracted with membrane permeabilization buffer for 1 min before fixation. Membrane permeabilization buffer contained 50 mm PIPES, pH 6.8, 1 mm EGTA, 0.2 mm MgCl2, 10% glycerol, 50 µg/mL digitonin and completeTM protease inhibitor cocktail (Roche, Indianapolis, IN, USA). Quick extraction of cells in this buffer permeabilized the membrane, but cell morphology essentially retained intact; this method was thus suitable for observation of the stable cytoskeleton (Lieuvin et al. 1994). For blocking the TRPV1, cells were incubated with 1 µm I-RTX for 10 min and activation of TRPV1 by RTX was done in presence of I-RTX.
For biochemical analysis of the activated cells by western blot analysis, TRPV1-F11 cells were scraped from the culture vessels, collected by a brief centrifugation and resuspended in HBSS. An equal volume of suspension was distributed into different tubes. Cells were activated by addition of an equal volume of HBSS supplemented with RTX and CaCl2. In control experiments, an equal volume of HBSS only was added. To examine the cytoskeleton of the cells after TRPV1 activation, an equal volume of 2 × membrane permeabilization buffer was added and mixed gently without homogenization for 1 min, followed by centrifugal separation of supernatants and pellets at 150 gfor 5 min at room temperature.
Depolymerization of microtubules in cell culture
F11 cells expressing TRPV1-Ct after transfection were incubated with 1 µm nocodazole for 15 min at 37°C, washed with HBSS at room temperature, extracted with membrane permeabilization buffer and fixed with Paraformaldehyde (PFA).
Western blot analysis
The amount of protein present in the extracts was determined by the bicinchoninic acid method (Pierce, Rockford, IL, USA). For analysis of the proteins, extracts were resolved by sodium dodecyl sulfate–polyacrylamide gel electrophoresis on 10% gels according to the method of Laemmli (1970). For western blot analysis, proteins were transferred to a nitrocellulose membrane by the semidry method, and membranes were blocked with 5% fat-free dry milk powder suspended in Tris-buffered saline (TBST; 20 mm Tris-HCl pH 7.4, 150 mm NaCl, 0.1% Tween 20). Blocked membranes were incubated with primary antibody in TBST buffer for 1 h, washed three times with TBST buffer, then incubated with secondary antibody in TBST buffer for 1 h. Finally, the membranes were washed with TBST and developed with an enhanced chemiluminescence kit (Amersham Biosciences/GE Healthcare, Freiburg, Germany).
Activation of TRPV1 alters microtubule cytoskeleton morphology
We recently reported that the C-terminal cytoplasmic sequence of TRPV1 interacts with tubulin dimers and polymerized microtubules. It thereby alters some physicochemical properties of microtubules (Goswami et al. 2004). To understand the effect of TRPV1 activation on the tubulin cytoskeleton in a cellular context, we treated TRPV1-transfected cells with the selective agonist RTX, and monitored the integrity of cytoskeletal structures after RTX activation of TRPV1 by indirect immunofluorescence (Fig. 1). We expressed TRPV1 in F11 cells by transient transfection. The bright immunostaining of TRPV1 in some cells allowed us to distinguish TRPV1-expressing cells from the non-transfected cells which showed no immunoreactivity for TRPV1.
TRPV1 localized to the plasma membrane as distinct patches as well as to intracellular membranes, in line with previous reports (Jahnel et al. 2001; Goswami et al. 2004). Both filamentous microtubules and actin fibres were clearly visible in non-activated F11 cells over-expressing TRPV1 (Fig. S1). Accumulation of some tubulin in the TRPV1-enriched patches and significant co-localization in these areas were observed in contrast to findings in non-transfected cells (Fig. S1a). TRPV1-enriched patches were never found to contain actin (Fig. S1b).
To observe the effect of TRPV1 activation, the receptor was activated by RTX. Upon activation, the microtubule cytoskeleton was dispersed within 1 min in TRPV1-expressing F11 cells (Fig. 1). In these cells, the microtubule structure was lost and tubulin immunoreactivity was dispersed all over the cell body. Non-transfected cells, however, showed normal microtubule structures after RTX treatment (Fig. 1). Microtubules in TRPV1-expressing cells had normal fibrous structure even after RTX activation when the cells had been preincubated with I-RTX, an antagonist of TRPV1 (data not shown). In contrast, we observed no significant change in the actin and neurofilament cytoskeletons of TRPV1-expressing F11 cells after the same treatment with RTX (data not shown). When we performed the same experiment in HEK293 cells transfected with TRPV1, we observed a similar effect on microtubules (data not shown). These results indicate that it is the activation, but not the over-expression, of TRPV1 that affects the fine microtubular structures.
Activation of TRPV1 results in increased tubulin solubility owing to disassembly of microtubules
The apparent dispersal of microtubules subsequent to activation of TRPV1 detected by indirect immunofluorescence suggests that the microtubules may have been disassembled into small tubulin oligomers and dimers. To analyse whether activation of TRPV1 renders an increased fraction of tubulin soluble, we extracted cells with digitonin in an isotonic buffer immediately after activation by RTX. Under these conditions, elements of the stable cytoskeleton and associated proteins are expected to remain unaffected while all soluble cytoplasmic proteins are extracted. After extraction, all cells that expressed TRPV1 showed a significantly reduced tubulin content (Fig. 2a). Some residual immunostaining for tubulin, which was observed in the cell body, appeared diffuse (Fig. 2a,iv). In some instances, it could be assigned as fragmented microtubules (Fig. S2). However, a perinuclear anti-tubulin immunoreactive structure was retained even after RTX activation (Fig. 2a; see also Fig. S2). No effect of RTX was observed in non-transfected cells (Fig. 2a). To prove further that the observed disassembly of microtubules indeed results from TRPV1 activation, we incubated the cells with antagonist I-RTX for 10 min before activation with RTX and extraction with digitonin. As expected, I-RTX-treated TRPV1-expressing F11 cells did not show loss of microtubules from the cell body after RTX activation and the microtubules had a normal fibrous pattern (Fig. 2b). In addition, TRPV1-transfected, but unstimulated cells displayed a normal microtubule structure comparable to that of non-transfected cells (Fig. 2c). These results indicate that the loss of microtubule structures in the cell body is rapid and occurs within a minute of TRPV1 activation. After RTX treatment and digitonin extraction, the actin and neurofilament cytoskeleton remained unchanged, and visible all over the TRPV1-expressing cells, comparable to that in control cells (Fig. 2d). This indicates that activation of TRPV1 has no effect on actin and neurofilament cytoskeletons.
TRPV1-mediated microtubule dispersal in stably transfected cells
In order to achieve a homogeneous TRPV1-expressing cell population for biochemical studies, we stably transfected F11 cells with TRPV1 cDNA. We obtained stably transfected F11 cells by a viral transduction method and subsequent fluorescence-activated cell sorting, according to a method described previously (Liu et al. 2000). Expression of TRPV1 in this cell line was confirmed by western blot analysis with anti-TRPV1 antibody (data not shown) and immunofluorescence analysis. Expression of TRPV1 was detected in all cells by indirect immunofluorescence analysis with anti-TRPV1 antibody, and the expression level was much lower than that in transiently transfected cells (data not shown). Low and homogeneous expression of TRPV1 in TRPV1-F11 cells allowed us to study biochemically the effect of TRPV1 activation.
We observed by immunofluorescence analysis that activation of TRPV1 after RTX treatment of TRPV1-F11 cells resulted in the selective disassembly of microtubules but not of the actin cytoskeleton in the majority of cells (data not shown). This effect was similar to that observed in transiently transfected cells. Co-staining of actin and microtubules in TRPV1-F11 cells confirmed that the cells that showed a drastic reduction in microtubules in the cell body retained a normal actin cytoskeleton, even after activation of TRPV1 followed by detergent extraction (Fig. 3).
To confirm the apparent increase in soluble tubulin subsequent to activation of TRPV1 by an independent method, we applied the detergent extraction treatment to the TRPV1-F11 cells, but then separated the soluble and insoluble fractions by centrifugation. We further analysed the soluble and insoluble fractions by sodium dodecyl sulfate–polyacrylamide gel electrophoresis, and determined the abundance of cytoskeletal compounds across the fractions by western blot analysis (see flow chart in Fig. 4a). The amount of tubulin increased significantly in the soluble fraction and decreased in the insoluble fraction after activation of TRPV1 (Fig. 4b). In contrast, only a small increase in actin and almost no change in neurofilament protein were observed in the soluble and insoluble fractions following TRPV1 activation compared with levels in non-stimulated cells. These results suggest that the predominant disassembly of the tubulin cytoskeleton is a downstream effect of TRPV1 activation.
Dynamic microtubule structures are affected by TRPV1 activation
We always observed an area near the perinuclear zone that retained strong tubulin immunoreactivity even after activation of TRPV1 and extraction with detergent. This finding suggested that this structure was made of mainly stable microtubules that are resistant to Ca2+ influx and detergent extraction. One candidate structure for this is the microtubule organizing centre (MTOC), which characteristically contains γ-tubulin (Joshi 1993; Oakley and Akkari 1999). To determine whether the TRPV1 activation-resistant structures contain γ-tubulin, RTX-activated and detergent-extracted TRPV1-F11 cells were co-immunostained with antibodies against tubulin and γ-tubulin. We observed that the centre of this anti-tubulin-immunoreactive region indeed contained γ-tubulin and thus represents the MTOC Fig. S3a). Tubulin staining was retained in some cell extensions even after RTX treatment and digitonin extraction (Fig. S3b), suggesting that the effect of TRPV1 activation on the microtubule cytoskeleton is cell-site specific and that the peripheral dynamic microtubules in the cell body are most affected, whereas microtubules at the MTOC and in neurite-like processes are more stable.
To analyse this effect of TRPV1 activation on microtubule subpopulations in more detail, we applied the fractionation scheme according to the flow chart (Fig. 4a), but probed the insoluble and the soluble fractions with antibodies specific for various post-translationally modified tubulins, as post-translational modifications of tubulin may alter the physicochemical properties of microtubules (MacRae 1997; Westermann and Weber 2003). We also probed the same samples for other microtubule cytoskeleton proteins, namely γ-tubulin (as a component of MTOC), the neurone- specific β-tubulin subtype III, and tau. Under conditions of RTX activation of TRPV1-F11 cells, we observed no change in the distribution of γ-tubulins in comparison to that in non-activated cells, and the majority of the γ-tubulin remained in the insoluble fraction (Fig. 5). This is in line with data from our immunofluorescence studies (Fig. S3a). It suggests that the MTOC, a rigid and stable structure composed mainly of stable microtubules, γ-tubulin and other modified tubulins, remains unaffected upon TRPV1 activation. In contrast, levels of tyrosinated tubulin, a marker for dynamic microtubules (Gundersen et al. 1984; Kreis 1987; Wehland and Weber 1987) increased significantly in the soluble fraction after TRPV1 activation compared with levels under control conditions (Fig. 5). A certain amount of tyrosinated tubulin, however, remained in the insoluble fraction even after TRPV1 activation and detergent extraction. This is in full agreement with our immunostaining results showing that stable microtubules at the MTOC and neurites also contain a considerable amount of tyrosinated tubulin (Figs S3a and b). In contrast to the tyrosinated tubulin, the proportion of de-tyrosinated tubulin (glu tubulin), polyglutamylated tubulin, and acetylated tubulin was not significantly altered in the soluble fraction after TRPV1 activation. Nor did the distribution of neurone-specific β-tubulin subtype III change upon TRPV1 activation. It is important to note that de-tyrosinated tubulin, acetylated tubulin, polyglutamylated tubulin and neurone-specific β-tubulin subtype III are all associated with stable microtubules. After TRPV1 activation, no significant change in the soluble fraction was observed for the neurone-specific tau, which stabilizes microtubules (Weingarten et al. 1975).
In summary, these results suggest that the dynamic microtubules (enriched with tyrosinated tubulin and/or non-modified tubulins) were affected by TRPV1 activation, whereas stable microtubules were not.
We observed that TRPV1, when over-expressed in F11 cells, was localized at the plasma membrane as distinct patches (Fig. S1a). This is in line with our previous report that TRPV1 directly interacts with tubulin dimers and polymers via its C-terminal domain (Goswami et al. 2004). In order to understand whether this tubulin forms part of a stable structure in such TRPV1-enriched patches, F11 cells transiently transfected with TRPV1 were extracted with digitonin or Triton X-100 subsequent to RTX activation of TRPV1, and then subjected to indirect immunofluorescence analysis. Although activation of TRPV1 followed by detergent extraction resulted in the loss of microtubule structures from most of the cell body, TRPV1-enriched patches at the plasma membrane still retained some tubulin immunoreactivity (Fig. 6; data for Triton X-100 extraction is shown), indicating that the tubulin in the TRPV1-enriched patches forms part of stable microtubule structures.
Over-expression of TRPV1-Ct results in the formation of bundled microtubules and stabilizes them
As reported before, we observed that the C-terminus, but not the N-terminus, of TRPV1 provides stability to microtubules in vitro (Goswami et al. 2004). Therefore we next asked whether it is the C-terminal domain of TRPV1 that confers stability to microtubules localized in TRPV1 patches. In order to assess this, we transiently expressed only the C-terminal cytoplasmic domain of TRPV1 in F11 cells and performed immunostaining of the C-terminal fragment as well as of tubulin. TRPV1-Ct immunofluorescence appeared as distinct spots throughout the cytoplasm (Fig. 7). These spots were much bigger than the expected size of vesicles. Interestingly, we observed an uneven distribution and bundling of microtubules in TRPV1-Ct-expressing cells (Fig. 7a). This bundling occurred especially in regions that contained clusters of TRPV1-Ct. In contrast to the transfected cells, non-transfected cells did not show uneven distribution or bundling, and a normal microtubule structure was visible (Fig. 7a). This result supports a stabilizing effect of TRPV1-Ct on the microtubules in vivo. No changes in the actin cytoskeleton of TRPV1-Ct-expressing cells were observed (Fig. S4). We next extracted the cells with detergent before fixing them, and performed a similar immunostaining. We observed that the cluster-like spots that contained TRPV1-Ct were not extractable (Fig. 7b). After detergent extraction, the cells lose soluble tubulin, but retain microtubules. In TRPV1-Ct-expressing cells, unevenly distributed microtubules all over the cell body and bundled microtubules in areas enriched with TRPV1-Ct spots became prominent after detergent extraction. Apart from these spots, some TRPV1-Ct immunoreactivity was also observed along with the microtubules after detergent extraction, especially in areas with enriched microtubule structures (Fig. 7b). In contrast, non-transfected cells showed uniform distribution of microtubules all over the cell body and no bundling of microtubules (Fig. 7b).
In order to further substantiate our conclusion that the C-terminus of TRPV1 provides stability to the microtubules in cultured cells, we incubated cells over-expressing TRPV1-Ct with nocodazole, a microtubule-destabilizing drug. We then extracted the cells with an isotonic buffer containing digitonin. The TRPV1-Ct clusters remained visible even after nocodazole treatment and digitonin extraction (Fig. 8). We observed by immunofluorescence analysis that the amount of tubulin immunoreactivity left after nocodazole treatment and detergent extraction (nocodazole-resistant microtubules) in TRPV1-Ct-expressing cells was significantly higher than that in non-transfected cells (Fig. 8). We also noted a correlation between the amount of nocodazole-resistant microtubules left after detergent extraction and the level of expression of TRPV1-Ct. This was confirmed by a comparative analysis of the intensity of tubulin immunoreactivity of nocodazole-resistant microtubules in TRPV1-Ct-expressing cells and non-expressing cells (Fig. 8). Similar results were obtained when we probed the cells with anti-α-tubulin antibody (data not shown). More often we observed the presence of nocodazole-resistant microtubules all over the cell body, including the distal peripheral areas in the TRPV1-Ct-expressing cells, whereas the distribution of the same nocodazole-resistant microtubules in non-transfected cells was limited only to the presumed MTOC and nearby areas (data not shown). These results accord well with our earlier observation that MBP-TRPV1-Ct, a recombinant Maltose binding protein (MBP) fusion protein with the TRPV1 C-terminus, interacts with polymerized microtubules in vitro and provides stabilization against nocodazole (Goswami et al. 2004).
In summary, our results indicate that TRPV1 activation results in destabilization of dynamic microtubules whereas the C-terminus of the channel has a microtubule-stabilizing effect. Such an interplay between stabilizing and destabilizing effects could provide the basis for the participation of TRPV1 in microtubule cytoskeletal remodelling during pain transmission.
We recently reported a Ca2+-sensitive interaction of TRPV1 with the tubulin cytoskeleton (Goswami et al. 2004), and hypothesized that the tubulin cytoskeleton might be a downstream effector of TRPV1 activation. Extending the results of our previous study, we now provide evidence in a cellular context that the integrity of the tubulin cytoskeleton is indeed affected by TRPV1 activation. Activation of TRPV1 results in a selective depolymerization of microtubules into soluble tubulin. This depolymerization affects primarily unmodified and tyrosinated tubulins, which are known to be part of dynamic microtubules (Gundersen et al. 1984; Kreis 1987; Wehland and Weber 1987). Disassembly of microtubules resulting from activation of TRPV1 is not dependent on the level of TRPV1 expression as stably transfected TRPV1-F11 cells, which express a much lower level of TRPV1, also exhibit the same effect. We found, on the other hand, that some microtubules remain attached to TRPV1 after activation of the ion channel, and provide data to support our conclusion that this is due to a stabilizing effect of the C-terminal portion of TRPV1 on the microtubules that interact with the channel. This is supported by our finding that over-expression of the C-terminal fragment of TRPV1 alone in F11 cells leads to bundling and stabilization of microtubules, which makes them resistant to detergent and nocodazole. The occurrence of destabilization of microtubules by TRPV1 activation and stabilization of microtubules by the C-terminus of TRPV1 indicates that a fine balance exists between TRPV1 and the microtubule cytoskeleton. TRPV1 displays characteristics that make it a good candidate as an effector of remodelling of the microtubule cytoskeleton in pain transmission.
The mechanism of tubulin cytoskeleton disruption upon TRPV1 activation, which was not the subject of this study, still remains unclear. It is known that Ca2+ has a depolymerizing effect on microtubules in vitro (Karr et al. 1980; Job et al. 1981). Depolymerizing effects of Ca2+ on microtubules were demonstrated in vivo and it has been shown that Ca2+ leads to two distinct processes, dynamic destabilization and signal cascade-induced fragmentation of microtubules (Lieuvin et al. 1994). Enzymatic pathways might also be involved as RTX activation does not result in the loss of microtubules when the enzymatic activities of the cells are reduced by lowering the temperature (data not shown). Activation of Ca2+-dependent proteases, for example, may occur upon TRPV1 activation, which triggers proteolysis of structural proteins as a downstream effect (Chard et al. 1995).
Prolonged stimulation of responsive neurones with capsaicin has been known for a long time to induce neuronal cell death (Jancso et al. 1984). In fact, retraction and degeneration of sensory neurones, which may well involve events that affect cytoskeletal integrity, may be the basis for the analgesic effect of topical capsaicin treatment (McMahon et al. 1991). Cytotoxicity due to TRPV1 activation and subsequent deletion of TRPV expressing neurones was also reported for other parts of the TRPV1-expressing nervous system (Karai et al. 2004; Kim et al. 2005).
In the experiments reported here, activation of TRPV1 did not affect the stability of the neurofilament cytoskeleton, but affected the actin cytoskeleton to a certain extent as some actin appeared in the soluble fraction after TRPV1 activation. This increasement in actin in soluble fraction is not surprising as actin and microtubule filaments are interconnected within the cell (Griffith and Pollard 1978, 1982). Immunofluorescence analysis of the TRPV1-activated and detergent-extracted cells revealed that both actin and neurofilament cytoskeletons remained as intact polymers (Fig. 2d). Moreover, the N- and C-terminus of TRPV1 do not interact with purified actin or enriched neurofilaments (data not shown), whereas the C-terminus of TRPV1 interacts with tubulin in a Ca2+-sensitive manner (Goswami et al. 2004). This apparent specificity for the microtubule cytoskeleton is remarkable in the light of the well established role of microtubule-active drugs in neuropathic pain. Vincristine and paclitaxel, two agents used as chemotherapeutics in the treatment of cancer, can produce a painful peripheral neuropathy (Polomano and Bennett 2001; Quasthoff and Hartung 2002). Notably, both vincristine and paclitaxel regulate microtubule stability. It has also been reported that TRPV1 is involved in bone cancer pain (Ghilardi et al. 2005). However, the precise mechanism that links the actions of vincristine and paclitaxel to peripheral neuropathy is currently unclear. On the other hand, there are reports that the integrity of the cytoskeleton plays an important role in the development of persistent pain states in certain experimental paradigms (Dina et al. 2003). Furthermore, TRPV4, another member of the TRPV family of ion channels, has been reported to be essential for the development of pain in a rat model of paclitaxel-induced neuropathy (Alessandri Haber et al. 2004). In summary, all these reports suggest that the effect of TRPV1 on the microtubule cytoskeleton may well be relevant to TRPV1-mediated pain transmission, particularly in physiopathological situations.
The technical assistance of Doris Krück and Jutta Metz is gratefully acknowledged. We thank Dr R. Jahnel for preparing the TRPV1-Ct construct. We acknowledge Oliver Bogen, who provided the enriched neurofilament fraction and engaged in useful discussions. This work was supported by grant no. 01 GG 9818/0 from Molecular Pain Research, by the Deutsche Forschungsgemeinschaft, Sfb 515, and by the Fonds der Chemischen Industrie. MD is supported by the Wellcome Trust.