点击显示 收起
【关键词】 Modified
Serotonin 5-HT2B receptors are often coexpressed with 5-HT1B receptors, and cross-talk between the two receptors has been reported in various cell types. However, many mechanistic details underlying 5-HT1B and 5-HT2B receptor cross-talk have not been elucidated. We hypothesized that 5-HT2B and 5-HT1B receptors each affect the others' signaling by modulating the others' trafficking. We thus examined the agonist stimulated internalization kinetics of fluorescent protein-tagged 5-HT2B and 5-HT1B receptors when expressed alone and upon coexpression in LMTK– murine fibroblasts. Time-lapse confocal microscopy and whole-cell radioligand binding analyses revealed that, when expressed alone, 5-HT2B and 5-HT1B receptors displayed distinct half-lives. Upon coexpression, serotonin-induced internalization of 5-HT2B receptors was accelerated 5-fold and was insensitive to a 5-HT2B receptor antagonist. In this context, 5-HT2B receptors did internalize in response to a 5-HT1B receptor agonist. In contrast, co-expression did not render 5-HT1B receptor internalization sensitive to a 5-HT2B receptor agonist. The altered internalization kinetics of both receptors upon coexpression was probably not due to direct interaction because only low levels of colocalization were observed. Antibody knockdown experiments revealed that internalization of 5-HT1B receptors (expressed alone) was entirely clathrin-independent and Caveolin1-dependent, whereas that of 5-HT2B receptors (expressed alone) was Caveolin1-independent and clathrin-dependent. Upon coexpression, serotonin-induced 5-HT2B receptor internalization became partially Caveolin1-dependent, and serotonin-induced 5-HT1B receptor internalization became entirely Caveolin1-independent in a protein kinase C-dependent fashion. In conclusion, these data demonstrate that coexpression of 5-HT1B and 5-HT2B receptors influences the internalization pathways and kinetics of both receptors.
Serotonin (5-hydroxytryptamine, 5-HT) is a potent vasoactive molecule and also a major neurotransmitter in both the central and peripheral nervous systems (Hoyer et al., 2002). All 5-HT receptors, except 5-HT3, belong to the rhodopsin-like G protein-coupled receptor (GPCR) superfamily. The 5-HT1B, 5-HT2B, and 5-HT2A receptors were found in endothelial and smooth muscle cells from several human and mouse arteries at mRNA, protein and functional levels (Ullmer et al., 1995; Watts et al., 1996). In addition, various human meningeal tissues have been found to coexpress 5-HT1B and 5-HT2B receptor mRNA (Schmuck et al., 1996). Thus, given their coexpression and their role in regulating smooth muscle contractility (Banes and Watts, 2003), 5-HT1B and 5-HT2B receptors have been implicated in the pathogenesis of migraine headaches (Kalkman, 1994; Schaerlinger et al., 2003; Poissonnet et al., 2004), an idea supported by the efficacy of 5-HT1B receptor agonist-based antimigraine therapy [sumatriptan succinate (Imitrex), methysergide maleate (Sansert)]. The antimigraine efficacy of 5-HT2B receptor antagonists is less clear but suspected (Schmuck et al., 1996), although the underlying mechanisms remain to be identified.
A general feature of GPCRs is the existence of complex intracellular regulatory mechanisms that modulate receptor responsiveness. Receptor desensitization and down-regulation are well described for various individual GPCR subtypes and are important homeostatic mechanisms. Homologous desensitization involves the desensitization of a particular receptor subtype upon activation of that receptor subtype. In contrast, heterologous desensitization involves the desensitization of receptor subtype(s) upon stimulation of a different receptor subtype. In view of the increasing number of reports of GPCRs participating in complexes via dimerization or scaffolding proteins, cross-talk between receptor subtypes may represent an important additional regulatory mechanism at modulating sensitivity and/or signal transduction. Agonist-induced internalization of GPCRs uses a pathway determined by a kinase that phosphorylates the receptor: for the 1-adrenergic receptor, protein kinase A (PKA)-mediated phosphorylation directs internalization via caveolae, whereas GPCR kinase (GRK)-mediated phosphorylation directs internalization through clathrin-coated pits (Rapacciuolo et al., 2003). The recruitment, activation, and scaffolding of cytoplasmic signaling complexes occurs via two multifunctional adaptor and transducer molecules, -arrestin-1 and -2 (Lefkowitz and Shenoy, 2005).
Among the described internalization mechanisms, 5-HT-induced receptor desensitization has been reported for receptors closely related to 5-HT2B receptors (i.e., 5-HT2A and 5-HT2C receptors). Desensitization of 5-HT2A receptor was shown to involve receptor internalization through caveolin1 (Cav1), a scaffolding protein enriched in caveolae, in a number of cell lines expressing exogenous 5-HT2A receptors, and rat brain synaptic membrane preparations (Bhatnagar et al., 2004). There is also evidence for functional interactions among 5-HT2A receptors and other plasma membrane microdomain proteins: 5-HT-induced 5-HT2A receptor desensitization can also involve receptor internalization through a clathrin- and dynamin-dependent process (Hanley and Hensler, 2002). Internalization and desensitization of 5-HT2A receptors in some cell types is -arrestin-independent (Gray et al., 2003). A direct interaction between PSD-95 and the 5-HT2A receptor at a type I PSD-95, Dlg, ZO-1 (PDZ)-binding domain at the C terminus regulates the receptor's signal transduction and trafficking (Xia et al., 2003). For the 5-HT2C receptor, constitutively active edited isoform is spontaneously internalized in an agonist-independent manner via the activity of a GRK/-arrestin (Marion et al., 2004).
Receptor oligomerization is a pivotal aspect of the structure and function of GPCRs that has also been shown to have implications for receptor trafficking, signaling, and pharmacology (George et al., 2002). Serotonin 5-HT2C receptors were shown to exist as constitutive homodimers on the plasma membrane of living cells using a confocal-based fluorescent resonance energy transfer (FRET) method (Herrick-Davis et al., 2004). Inactive 5-HT2C receptors can inhibit wild-type 5-HT2C receptor function by forming nonfunctional heterodimers expressed on the plasma membrane (Herrick-Davis et al., 2005). The 5-HT1B and 5-HT1D receptor subtypes that share a high amino acid sequence identity have also been shown to exist as monomers and homodimers when expressed alone and as monomers and heterodimers when coexpressed (Xie et al., 1999). Heterodimerization between 5-HT1 and 5-HT2 receptors, and its functional consequences have yet to be investigated.
The mechanistic details of 5-HT1B receptors internalization have not yet been determined. Despite the coexpression of 5-HT1B and 5-HT2B receptors in various tissues including endothelial and smooth muscle cells (Ullmer et al., 1995), and given the inhibitory effect of 5-HT2B receptors on 5-HT1B receptor signaling (Tournois et al., 1998), physical interaction between the two receptors seems plausible. The internalization of the 5-HT2B receptor is faster than that of 5-HT2A and 5-HT2C receptors (Porter et al., 2001; Schaerlinger et al., 2003; Deraet et al., 2005), although the mechanism underlying this distinction has not been uncovered. In this work, we investigated potential 5-HT1B/2B receptor interactions by examining the colocalization and internalization kinetics of 5-HT1B and 5-HT2B receptors expressed alone or together. Using cyan and yellow fluorescent protein (CFP and YFP) tagged receptors and confocal microscopy, we observed agonist-induced receptor endocytosis in real time. We also performed whole-cell radioligand binding studies as an additional means of measuring receptor internalization. Our results indicate that the stimulation of 5-HT1B receptors affects the internalization dynamics of 5-HT2B receptors and vice versa, with the effect of 5-HT1B receptors on 5-HT2B receptors being more pronounced than the effect of the latter receptor on the former. Furthermore, we used an antibody knockdown strategy to ascertain which pathways each receptor used for internalization. Our findings reveal that coexpression of 5-HT1B and 5-HT2B receptors affects both the kinetics of receptor internalization and the internalization pathway employed compared with either receptor expressed alone.
Reagents. RS127445 was kindly provided by Roche (Mannheim, Germany); CP93129, BW723C86, H-89, Gö 6850-Bisindolylmaleimide I, Gö 6976, and all other chemicals were reagent grade, purchased from usual commercial sources. The radioactive compounds 1-(2,5-dimethoxy-4-[125I]iodophenyl)-2-aminopropane hydrochloride ([125I]DOI; 81.4 TBq/mmol) and [125I]serotonin-5-O-carboxymethyl-glycil-iodo-tyrosamine ([125I]GTI; 81.4 TBq/mmol) were purchased from PerkinElmer Life and Analytical Sciences. Antibodies that were already specificity-tested were used: rabbit antisera against rodent Cav1 (H-97), clathrin (H-300), GRK2,3 (H-222), and GRK5,6 (C-20) and goat antiserum against rodent -arrestin-2 (D-18) and PKC (C-15) (Santa Cruz Biotechnology; Santa Cruz, CA).
Mutagenesis and 5-HT2B Receptor Constructs. The fluorescent-tagged fusion proteins of the mouse 5-HT2B and 5-HT1B receptors were generated by PCR-based subcloning. The receptor coding regions were subcloned into pECFP, pEYFP, pPA-GFP (photo-activable GFP) vectors with the XFP fused to the N terminus of the receptors. The entire coding sequence of all constructs was verified by automated DNA sequencing.
Cell Culture. 5-HT2B and 5-HT1B receptor cDNAs were stably transfected into nontransformed murine fibroblast (LMTK–) cells, which are devoid of endogenous 5-HT receptors because they do not exhibit a concentration-dependent rise in second messengers after 5-HT stimulation (Manivet et al., 2000). LMTK– cell lines were routinely cultured in Dulbecco's modified Eagle's medium (Invitrogen, Carlsbad, CA) containing 10% heat-inactivated fetal bovine serum and 40 µg/ml gentamicin. The stably expressing mouse 5-HT1B and 5-HT2B cell lines were generated by calcium phosphate transfection followed by selection with either G-418 or hygromycin and clonal isolation. Stable receptor-expressing lines were subcultured in serum free medium for at least 24 h before experiments. Stable cell lines with different combinations of receptor expression insured the absence of individual cell based effects.
Radioligand Binding Experiments. Radioligand binding experiments were performed using [125I]DOI or [125I]GTI either on intact cells or on membranes from stably transfected cells as previously detailed (Loric et al., 1995).
Cell Permeabilization. The cells were washed twice with phosphate-buffered saline containing 0.1% bovine serum albumin and exposed to 1 hemolytic unit of alveolysin/106 cells at 22°C under agitation as described previously (Manivet et al., 2000).
Internalization Measurements by Confocal Microscopy. For confocal microscopic studies of living cells, stable or transiently transfected cell lines were plated 36 h before the analysis on 35-mm glass-bottomed dishes (MatTek, Ashland, MA) and grown in a 5.4% CO2 incubator. Twelve hours before the experiment, cells were washed in Dulbecco's modified Eagle's medium without serum and maintained in this medium until the experiment. Confocal analysis were performed at 22°C unless otherwise mentioned in the text to extend the kinetic and 37°C was used to confirm the observed phenomenon at 22°C. Cells were visualized using a confocal microscope (SP2AOBS; Leica, Wetzlar, Germany) with laser excitation lines of 458 nm and 514 nm for CFP- and YFP-tagged receptors and transmitted light. Images of CFP and YFP emission were recorded simultaneously with the transmitted light images. The emission recording channels and the intensity of the excitation lasers were carefully chosen using single tag control linear unmixing technique (Leica) to avoid bleed-through. Images (in xyzt mode) were recorded sequentially every 5 min in two different emission intervals (462–500 nm for CFP and 520–600 nm for YFP) with two different excitation wavelengths (458 and 514 nm). To ensure consistency among z-plane images during time-lapse study, we took at least five z-plane images and manually selected only one plane for the time-lapse analysis. To visualize receptor internalization after agonist treatment, time-lapse series were taken every 5 min over a 30-min period. To calculate the internalization kinetics of the receptors, we selected 10 regions of interest (ROIs) on the plasma membrane per cell and followed the relative intensity changes of these ROIs by time, including at least three to five individual cells per experiment. Data represent more than four independent experiments. The kinetic curves were corrected for bleaching by using the intensity of the whole cell as a normalization factor. PA-GFP was activated with a 405 nm diode laser line pulse lasting 10 to 15 sec and visualized between 495 and 515 nm using excitation at 488-nm.
FRET Measured by Confocal Microscopy. For the colocalization of 5-HT2B and 5-HT1B receptors in living cells, sensitized emission fluorescent resonance energy transfer (FRET) was used. Two additional channels [the FRET channel (excitation, 458 nm; emission, 520–600 nm) and a control channel (excitation, 514 nm; emission, 462–500 nm)] were recorded along with the CFP and YFP channels as described above. The bleed-through of CFP and the direct excitation of YFP by the 458 nm laser light were subtracted from the FRET channel signal. To estimate these artifacts, we used cells transfected with either CFP- or YFP-tagged receptors. Calculation of corrected FRET was carried out on a pixel-by-pixel basis for the entire image. Indeed, positive FRET signal could be obtained using tagged proteins known to interact in similar experimental set-ups (data not shown).
Colocalization Calculation. We considered two proteins to be colocalized if the observed signals of the two corresponding labels were nonzero at the same pixel. The quantitative estimate of colocalization is given as the colocalization coefficients (Manders et al., 1992). To quantify the colocalized fraction of each receptor pair, a threshold value for each channel was estimated and subtracted. Bleed-through in each of the two detecting channels was subtracted using the linear unmixing method (Leica).
Data Analysis. Binding data were analyzed using the iterative nonlinear regression model (Prism 2.0; GraphPad Software, San Diego, CA). This allowed the calculation of dissociation constants (KD) and the number of sites (Bmax). All values represent the average of independent experiments ± S.E.M. (n = number of experiments as indicated in the text). Comparisons between groups were performed using Student's unpaired t test or analysis of variance and a Fischer test. Significance was set at p < 0.05.
Radioligand saturation binding assays were performed on membranes from stable, clonal cell lines expressing either 5-HT1B or 5-HT2B receptors, bearing various N-terminal fluorescent protein tags as well as on cells coexpressing both receptors. The clones chosen for subsequent experiments were selected so as to approximate physiological receptor expression (Bmax of approximately 100 fmol/mg of protein) (Tables 1 and 2). In control experiments with nontagged receptors, we verified that none of the N-terminal fluorescent tags affected radioligand affinity (Tables 1 and 2). In addition, whole-cell radioligand binding experiments on clones expressing either fluorescent protein-tagged or nontagged receptors yielded a Bmax approximately 60% of that measured on membrane preparations, suggesting that the fluorescent tag did not perturb membrane targeting (Tables 1 and 2).
TABLE 1 Pharmacological properties of native or GFP-tagged receptors
Saturation binding assays were performed using [125I]DOI (5-HT2B receptor) or [125I]GTI (5-HT1B receptor) on membrane fraction of cell homogenate (Membranes) or intact cells (Whole Cells) of the different stable clonal cell lines expressing the 5-HT1B receptor and the 5-HT2B receptor alone with or without tag. Binding data were analyzed using the iterative non-linear fitting software GraphPad Prism 2.0 to calculate dissociation constants (KD) and maximum number of sites (Bmax). Insertion of the various GFPs at the N-terminal part of the receptor did not affect receptor affinity. Intact cell Bmax represents only about 60% of that of binding on membranes. Values are means ± S.E. of three independent determinations in triplicate.
TABLE 2 Pharmacological properties of native or GFP-tagged coexpressed receptors
Saturation binding assays were performed using [125I]DOI (5-HT2B receptor) or [125I]GTI (5-HT1B receptor) on membrane fraction of cell homogenate (Membranes) or intact cells (Whole Cells) of the different stable clonal cell lines expressing the 5-HT1B receptor and the 5-HT2B receptor with or without tag. Binding data were analyzed using the iterative non-linear fitting software GraphPad Prism 2.0 to calculate dissociation constants (KD) and maximum number of sites (Bmax). Insertion of the various GFPs at the N-terminal part of the receptor did not affect receptor affinity. Intact cell Bmax represents only about 60% of that of binding on membranes. Co-expression of the receptor subtypes did not alter the receptor affinity for GTI or DOI (Table 1). Values are means ± S.E. of three independent determinations in triplicate.
Kinetics of 5-HT-Induced Internalization of 5-HT2B and 5-HT1B Receptors. To establish a quantitative method to measure internalization kinetics of the tagged receptors, fluorescence intensity at the membrane was compared with that in the cytoplasm in cells treated with 100 nM 5-HT for 0 to 30 min. The time-dependent fluorescence intensity changes at the plasma membrane and in the cytoplasm both fit a one-phase exponential function (decrease for the plasma membrane, increase for the cytoplasm) and had similar half-lives (5-HT2B receptor cytoplasm t1/2 = 23.0 ± 3.0 min; membrane t1/2 = 22.2 ± 3.3 min; 5-HT1B receptor cytoplasm t1/2 = 9.8 ± 2.2 min, membrane t1/2 = 10.7 ± 1.4 min). Because the signal-to-noise ratio was higher for the measurement of decreases in plasma membrane fluorescence, this was chosen to measure endocytosis kinetics at the plasma membrane (Fig. 1, A and B). To further confirm that we were measuring receptor endocytosis, we performed experiments with photoactivatable (PA) GFP-tagged 5-HT1B receptors and compared the rate of decrease in fluorescence intensity of the illuminated membrane region with that determined for plasma membrane YFP-tagged 5-HT1B receptors. As before, for both GFP- and YFP-tagged 5-HT1B receptors, the internalization kinetics fitted a one-phase exponential decay and were identical (5-HT1B-PA-GFP t1/2 = 9.8 ± 0.7 min; 5-HT1B-YFP t1/2 = 10.1 ± 1.2 min), further validating our image-based analysis of receptor internalization kinetics.
Fig. 1. 5-HT stimulus on 5-HT2B and 5-HT1B receptor, dose dependence. Confocal studies in living cells were performed on various LMTK–-transfected cell lines plated on glass-bottomed dishes in Dulbecco's modified Eagle's medium without serum. Quantitative internalization kinetics of the GFP-tagged receptors was assessed by measuring the fluorescence intensity disappearance at the membrane compared with the intensity increase in the cytoplasm of approximately 10 ROI per cell and followed their relative intensity changes. Both signals could be fitted with a single exponential function over time, giving the same half-time. A, series of single confocal plane images taken from living cells expressing 5-HT1B receptor-GFP by time lapse video were used to evaluate the internalization kinetics after stimulation by 100 nM 5-HT (right). Distribution of 5-HT1B receptors expressed at 0 (0) and 50 (50) min of 5-HT stimulation. B, series of single confocal plane images was taken from living cells expressing 5-HT2B receptor-GFP by time lapse video. Internalization kinetics after stimulation by 100 nM 5-HT stimulation (right). Distribution of 5-HT2B receptors expressed at 0 and 50 min of stimulation. These images are representative of more than three cells observed in each of at least four independent experiments. Scale bars, 2 µM. C, similar studies by confocal laser microscopy showed that 5-HT1B and 5-HT2B receptors internalized with a single exponential kinetics after stimulation with 100 nM, 1 µM, or 10 µM 5-HT, and their half-lives were concentration-dependent. D, study by confocal laser microscopy showed that when 5-HT2B receptors were coexpressed with 5-HT1B receptors (5-HT2B1B), the 5-HT2B receptor internalized with a single exponential kinetics after stimulation with 100 nM 5-HT, but its half-life became five times faster. E, when 5-HT1B receptors were coexpressed with 5-HT2B receptors (5-HT1B2B light gray bar), the 5-HT1B receptor internalized after stimulation with 100 nM 5-HT with a similar half-life to 5-HT1B receptor alone (black bar). F, independent radioligand binding experiments using [125I]DOI (5-HT2B receptor) or [125I]GTI (5-HT1B receptor) on intact living cells confirmed that [compared with 5-HT2B receptor expressed alone (dark gray bar)] when the 5-HT2B receptor was coexpressed with 5-HT1B receptors, stimulation with 100 nM 5-HT decreased the 5-HT2B receptor (5-HT2B1B white bar) half-life but not that of 5-HT1B receptors (5-HT1B2B light gray bar). Graphs display the mean± S.E.M. of at least three independent experiments.
Similar confocal laser microscopy studies with various concentrations of 5-HT revealed a concentration dependence on internalization kinetics. At 100 nM, 5-HT-induced internalization of 5-HT1B receptors was twice as fast as that of 5-HT2B receptors; at higher concentrations of 5-HT, internalization of both receptors occurred at similar rate. Furthermore, the 5-HT-induced 5-HT2B receptor internalization rate increased twice as fast as that of 5-HT1B receptors as a function of 5-HT concentration (Fig. 1C).
Having established internalization kinetics for both receptors expressed alone, we set out to ascertain whether coexpression altered 5-HT-induced 5-HT1B and 5-HT2B receptor internalization rates. Coexpression with the 5-HT1B receptor resulted in a 5-fold increase in the rate of 5-HT-induced 5-HT2B receptor internalization (2B1B)(t1/2 = 4.0 ± 1.5 min versus 23.0 ± 3.0) (Fig. 1, D and E). On the other hand, coexpression with 5-HT2B receptors had no effect on 5-HT-induced 5-HT1B receptor internalization rate (1B2B)(t1/2 = 9.0 ± 1.0 min versus 9.8 ± 2.2) (Fig. 1, D and E). Kinetic whole-cell radioligand binding experiments corroborated our microscopy data, revealing the asymmetric effect of receptor coexpression on 5-HT-induced 5-HT1B and 5-HT2B receptor internalization rate (Fig. 1, E and F).
Agonist-Dependent 5-HT2B Receptor Internalization. To determine whether the effect of 5-HT1B receptors on 5-HT2B receptor internalization kinetics involved activation of 5-HT2B receptors, we stimulated cells coexpressing both receptors in the absence and presence of the highly selective 5-HT2B receptor antagonist RS127445 (RS). In cells expressing 5-HT2B receptor alone, 100 nM RS completely blocked 5-HT-induced 5-HT2B receptor internalization. It was striking that RS had no effect on 5-HT-induced 5-HT2B receptor internalization in cells coexpressing 5-HT1B receptors (Table 3). RS did not significantly affect 5-HT-induced 5-HT1B receptor internalization irrespective of 5-HT2B receptor coexpression.
TABLE 3 Ligand dependence of 5-HT1B receptor and 5-HT2B receptor internalization
Quantitative internalization kinetics of the GFP-tagged receptors was assessed by measuring the fluorescence intensity disappearance at the membrane for more than 10 ROIs per cell and following the relative intensity changes of these ROI by time. In 5-HT2B receptors or 5-HT1B receptor transfected cells (single), this signal could be fitted with a single exponential function over time. Applying the same protocol on 5-HT2B receptors in the presence of 5-HT1B receptors transfected cells (double), a change in internalization rate could be observed. Values are means ± S.E. of more than three cells analyzed in at least four independent experiments.
To investigate whether modulation of internalization kinetics was agonist-dependent, we stimulated with the preferential 5-HT2B receptor agonist BW723C86 (BW). Treatment with 50 nM BW induced no 5-HT1B receptor internalization but did stimulate 5-HT2B receptor internalization (t1/2 = 11.0 ± 1.5 min) in noncoexpressing cells (Fig. 2A; Table 3). Coexpression of 5-HT1B and 5-HT2B receptors did not significantly alter the effect of BW on the internalization of either receptor: the internalization kinetics of 5-HT2B receptors changed only slightly in the presence of 5-HT1B receptor (t1/2 = 7.2 ± 1.1 min) (Fig. 2B; Table 3), whereas no internalization of 5-HT1B receptors was observed (Fig. 2C)
Fig. 2. Ligand dependence of 5-HT1B receptor and 5-HT2B receptor internalization. Quantitative internalization kinetics of the GFP-tagged receptors was assessed by measuring the fluorescence intensity disappearance at the membrane of series of single confocal plane images taken from living transfected LMTK– cells by time lapse video. A–C, the fluorescence intensity disappearance has been fitted with a single exponential function over time and used to evaluate the internalization kinetics after stimulation by the preferential 5-HT2B receptor agonist BW723C86 (BW) at the final concentration of 50 nM that produced no effect on the internalization of 5-HT1B receptor alone. D–F, the fluorescence intensity disappearance has been fitted with a single exponential function over time and used to evaluate the internalization kinetics after stimulation by the selective 5-HT1B receptor agonist CP93129 (CP) at the final concentration of 75 nM that produced no effect on the internalization of 5-HT2B receptor alone. The x-axes display 0 to 30 min recording for fast internalization (A–C and E) and 0 to 60 min for slow internalization (D and F). Deduced half-life values are reported in Table 3 with statistics.
Agonist-Dependent 5-HT1B Receptor Internalization. To further investigate putative reciprocal interactions between 5-HT1B and 5-HT2B receptors, we studied the agonist-induced internalization kinetics of the 5-HT1B receptor. The selective agonist CP93129 (CP; 75 nM), induced 5-HT1B receptor internalization (t1/2 = 14.1 ± 0.7) but did not affect 5-HT2B receptor distribution in noncoexpressing cells (Fig. 2D). We observed similar CP-induced internalization kinetics for both receptors (t1/2 = 6.9 ± 0.9 min for 5-HT2B receptors; t1/2 = 4.9 ± 1.3 min for 5-HT1B receptors) upon coexpression (Fig. 2E; Table 3). It is noteworthy that coexpression with 5-HT2B receptors caused a 3-fold increase in CP-induced internalization of 5-HT1B receptors (t1/2 = 4.9 ± 1.3 min versus t1/2 = 14.1 ± 0.7 min) (Fig. 2F; Table 3). These results reveal that, whereas coexpression with 5-HT2B receptor does not affect BW-induced 5-HT1B receptor internalization, the presence of 5-HT2B receptor does modulate CP-induced 5-HT1B receptor internalization.
Temperature Dependence of 5-HT1B and 5-HT2B Receptor Internalization. To verify that agonist-induced receptor internalization was due to an energy-dependent endocytic process, receptor internalization assays were performed at various temperatures. As expected, the kinetics of 5-HT-induced receptor internalization were faster at higher temperatures for both receptors, irrespective of coexpression. In addition, the temperature dependence of agonist (5-HT or subtype selective)-induced receptor internalization rate appeared linear for 5-HT1B receptors and biphasic for 5-HT2B receptor when expressed alone (Fig. 3, A and B).
Fig. 3. Temperature dependence of the of 5-HT1B receptor and 5-HT2B receptor internalization. The receptors half-life (t1/2) were evaluated using transfected LMTK– cells expressing GFP-tagged receptors by measuring the fluorescence intensity disappearance at the membrane after fitting with a single exponential function over time. Upon 5-HT stimulation, the t1/2 values in single receptor transfected cells was evaluated at different temperatures (22°C to 25°C and 37°C) for 5-HT1B receptors (A), for 5-HT2B receptors (B), or alone or coexpressed, after stimulation by CP for 5-HT1B receptors (C–E), or for 5-HT2B receptor (D–E) after BW stimulation. NS, no significant statistical difference. *, P < 0.05, more than three cells were analyzed in at least four independent experiments.
Coexpression of both receptors did not seem to affect the effect of temperature on CP-induced 5-HT1B receptor internalization (Fig. 3C). On the other hand, coexpression with 5-HT1B receptors rendered the temperature dependence of BW-induced 5-HT2B receptor internalization more linear (Fig. 3D). Furthermore, the temperature dependence of CP-induced 5-HT2B receptor internalization rate was identical to that observed for 5-HT1B receptors (Fig. 3E).
Microscopic Analysis Revealed No Receptor Colocalization. The apparent effect of receptor coexpression on agonist-induced 5-HT2B and 5-HT1B receptor internalization led us to hypothesized receptor heterodimerization. Therefore, we performed confocal microscopy on cells coexpressing CFP-5-HT1B and YFP-5-HT2B receptors. Cellular distribution analysis of the two receptors revealed approximately 20% colocalization in the plasma membrane before stimulation (Fig. 4). After 30 min of agonist stimulation, cytoplasmic colocalization of 5-HT2B and 5-HT1B receptors was still approximately 20%, irrespective of agonist. Using enhanced emission to measure FRET, we observed nearly no FRET signal at 5-HT1B/5-HT2B receptor colocalization points (Fig. 4). Thus, our image-based analysis of colocalization was not consistent with agonist-induced 5-HT1B/5-HT2B receptor complex formation.
Fig. 4. Single-cell fluorescence measurements for quantitative analysis. The cellular distribution of CFP-tagged 5-HT2B receptors (2B, red) co-expressed with YFP-5-HT1B receptors (1B green) was video-recorded and is displayed on a single confocal plane before (0 min) and after (30 min) stimulation with different agonists. The corresponding colocalizing points (1B2B, yellow) and deduced FRET values are shown (right panels, white). A, distribution of 5-HT2B receptor expressed at 0 and 30 min of CP stimulation. B, distribution of 5-HT2B receptor expressed at 0 and 30 min of 5-HT stimulation. C, distribution of 5-HT2B receptor expressed at 0 and 30 min of BW stimulation. These images are representative of more than 3 cells observed in each of at least four independent experiments. Scale bars, 2 µM.
Serotonin 5-HT2B and 5-HT1B Receptor Internalization Occurs via Distinct Pathways. In the absence of evidence supporting 5-HT1B/5-HT2B heterodimerization, we investigated the internalization pathways used by the receptors. We performed whole-cell binding studies on cells permeated with antibodies against proteins known to be involved in GPCR internalization. Cell surface receptor expression (Bmax) was measured as a function of agonist exposure time to measure internalization half-life. It is noteworthy that the alveolysin and antibody treatments did not affect agonist and radioligand affinities (data not shown).
We observed that the internalization of 5-HT1B receptors expressed alone was independent of GRK5,6 (data not shown) and clathrin but totally dependent on Cav1 and GRK2,3 when stimulated by 5-HT or CP (Fig. 5, A and B). The internalization of 5-HT2B receptors expressed alone was independent of GRK5,6 (data not shown) and Cav1 but completely dependent on clathrin and -arrestin-2 when stimulated by 5-HT or BW (Fig. 5, A–C). These results established that 5-HT1B and 5-HT2B receptors, when expressed alone, used distinct internalization pathways in an identical cell background.
Fig. 5. Whole-cell binding and antibody knockdown for quantitative analysis. The amount of membrane receptor sites in transfected LMTK– cells was assessed by radioligand binding experiments using [125I]DOI (5-HT2B receptor) or [125I]GTI (5-HT1B receptor) on intact living cells 0, 5, and 10 min after agonist stimulation (KD and Bmax). With unmodified KD (not illustrated), the time-dependent changes in Bmax allowed the approximation of the receptor half-life in response to various agonists (left). Exposure of alveolysin-permeabilized cells to antibodies against caveolin-1, clathrin, -arrestin-2, or GRK2,3 modifies the receptor half-life (expressed as -fold change over unexposed cells, right), upon agonist stimulation 5-HT (A), CP (B), and BW (C). Black bars, 5-HT1B receptors; light gray bars, 5-HT1B receptors coexpressed with 5-HT2B receptors (5-HT1B2B); dark gray bars, 5-HT2B receptors; white bars, 5-HT2B receptors coexpressed with 5-HT1B receptors (5-HT2B1B). Graphs display the mean ± S.E.M. of at least three independent experiments.
We next applied the antibody knockdown strategy to cells coexpressing 5-HT1B and 5-HT2B receptors. When stimulated by 5-HT or CP, but not BW, the internalization of 5-HT2B receptors became partially sensitive to Cav1 antibodies. Furthermore, the internalization of 5-HT2B receptors was entirely sensitive to Arrestin-2 antibodies when stimulated with 5-HT in the absence or presence of 5-HT1B receptors, and with CP in the presence of 5-HT1B receptors (Fig. 5, A and B). However, 5-HT2B receptor internalization was only partially inhibited by anti--Arrestin-2 when coexpressed with 5-HT1B receptors and stimulated with BW (Fig. 5C). This result further suggested that 5-HT1B receptors could affect the internalization pathway of 5-HT2B receptors. Finally, the agonist-induced internalization of 5-HT2B receptors was partially dependent on GRK2,3 when coexpressed with 5-HT1B receptors and stimulated with 5-HT, CP, or BW (Fig. 5). Thus, the effect of 5-HT1B receptor coexpression on 5-HT2B receptors caused a fraction of 5-HT2B receptors to internalize via a Cav1- and GRK2,3-dependent pathway.
With respect to 5-HT1B receptors, coexpression with 5-HT2B receptors caused 5-HT-induced 5-HT1B receptor internalization to become totally independent of Cav1 and GRK2,3. Furthermore, upon coexpression with 5-HT2B receptors, 5-HT-induced 5-HT1B receptor internalization was still independent of clathrin and Arrestin-2. In contrast, the Cav1/GRK2,3 dependence of CP-induced 5-HT1B receptor internalization was not affected by coexpression with 5-HT2B receptors. Thus, the effect of 5-HT2B receptor coexpression on 5-HT1B receptor internalization is to alter the 5-HT-induced internalization pathway from a fully Cav1-dependent pathway to one fully independent of both Cav1 and clathrin (Fig. 5).
Serotonin-Induced Stimulation of PKC by 5-HT2B Receptors Regulates the Pathway of 5-HT1B Receptor Internalization. To investigate the non–Cav1-dependent/non–clathrin-dependent internalization pathway used by 5-HT-stimulated 5-HT1B receptors coexpressed with 5-HT2B receptors, we tested the effect of various protein kinase inhibitors. The wide protein kinase inhibitor staurosporine (5 µM), which inhibits PKC, PKA, and protein kinase G, blocked 5-HT-induced 5-HT1B receptor internalization when coexpressed with 5-HT2B receptors (Fig. 6). H89 (5 µM), which inhibits PKA, protein kinase G, and PKCµ, but not other PKCs (Davies et al., 2000), had no effect on 5-HT1B or 5-HT2B receptor internalization. To further refine these results, we used PKC isotype-selective inhibitors to identify the PKC isozyme involved. We found that Gö 6850-Bisindolylmaleimide I (100 nM), a PKC inhibitor with high selectivity for PKC, -I, -II, -, -, and - isozymes, completely prevented 5-HT-induced 5-HT2B receptor internalization or that of 5-HT1B receptors in the presence of 5-HT2B receptors. This blocking effect was not observed with Gö 6976 (100 nM), which selectively inhibits the Ca2+-dependent PKC and -I. Finally, the blocking effect of Gö 6850-Bisindolylmaleimide I was completely reproduced using PKC antibody knockdown (Fig. 6). These data indicate that the 5-HT stimulation of 5-HT2B receptors triggers 5-HT1B receptor internalization via a pathway that requires 5-HT2B receptor dependent PKC activation.
Fig. 6. Whole-cell binding and pharmacological treatment for quantitative analysis. The amount of membrane receptor sites in transfected LMTK– cells was assessed by radioligand binding experiments using [125I]DOI (5-HT2B receptor) or [125I]GTI (5-HT1B receptor) on intact living cells 0, 5, and 10 min after agonist stimulation (KD and Bmax). With unmodified KD (not illustrated), the time-dependent changes in Bmax allowed to approximate the receptor half-life in response to various protein kinase blockers: 5-HT (A), 5-HT + staurosporine (5 µM) (B), 5-HT + H89 (5 µM) (C), 5-HT + Gö 6976 (100 nM) (D), and 5-HT+Gö 6850-Bisindolylmaleimide I (100 nM) (E). Exposure of alveolysin-permeabilized cells to antibodies against PKC modifies the receptor half-life, upon 5-HT stimulation (F). Black bars, 5-HT1B receptors; light gray bars, 5-HT1B receptors coexpressed with 5-HT2B receptors (5-HT1B2B); dark gray bars, 5-HT2B receptors; white bars, 5-HT2B receptors coexpressed with 5-HT1B receptors (5-HT2B1B). Graphs display the mean ± S.E.M. of three independent experiments.
Given 1) the wide in vivo coexpression of 5-HT1B and 5-HT2B receptors (Kellermann et al., 1996; Ishida et al., 1998, 1999; Stefulj et al., 2000; Banes and Watts, 2003; Nicholson et al., 2003), 2) the established inhibitory effect of 5-HT2B on 5-HT1B receptor signaling (Tournois et al., 1998), 3) the clinical utility of 5-HT1B receptor agonists, and 4) the putative efficacy of 5-HT2B receptor antagonists in treating migraines, knowledge about regulatory mechanisms between the two receptors is of high interest for understanding current and designing novel pharmaceuticals for therapeutic treatments. The main finding of this article is the asymmetric, agonist-dependent cross-regulation of 5-HT1B and 5-HT2B receptor internalization. The evidence for this cross-regulation is that the 5-HT1B receptor agonist CP causes 5-HT2B receptor to internalize only if 5-HT1B receptor is present, whereas the 5-HT2B receptor agonist BW does not similarly affect 5-HT1B receptors. However, CP-induced internalization of 5-HT1B receptor is faster when 5-HT2B receptors are present, demonstrating an effect of 5-HT2B receptors on 5-HT1B receptor internalization. Furthermore, coexpression with 5-HT2B receptors causes 5-HT1B receptors to adopt a Cav1- and clathrin-independent but PKC-dependent 5-HT-induced internalization, whereas a portion of 5-HT2B receptors assumes a Cav1-dependent 5-HT-induced internalization pathway.
The present results demonstrate that individually, 5-HT1B and 5-HT2B receptors expressed in nontransformed mouse fibroblast LMTK– cells use typically described agonist-dependent internalization pathways. The differences in the kinetics and temperature dependence of internalization strongly support the notion that these two receptors—when expressed alone—use different endocytic pathways, each agonist leading to specific output. The antibody knock-down experiments validate these findings and demonstrate that 5-HT1B receptors expressed alone internalize via a Cav1- and GRK2,3-dependent pathway, whereas 5-HT2B receptors expressed alone internalize via a clathrin-, -arrestin-2-, and PKC-dependent pathway (Fig. 7A).
Fig. 7. Model of interactions between 5-HT2B and 5-HT1B receptors in their internalization pathways. A, when expressed as a single receptor in cells devoid of endogenous 5-HT receptors expression and upon 5-HT stimulation, 5-HT1B receptor internalizes via a Cav1- and GRK2,3-dependent pathway, whereas 5-HT2B receptors internalize via a clathrin and arrestin-2-dependent pathway. B, when coexpressed, the two receptors respond differently to selective agonists. A selective 5-HT2B receptor agonist, BW, does not trigger any internalization of coexpressed 5-HT1B receptor and acts as 5-HT on the 5-HT2B receptor internalization (not illustrated). By contrast, stimulation of 5-HT1B receptor by CP, a selective 5-HT1B receptor agonist, makes the 5-HT1B receptor internalizing via Cav1 and GRK2,3 as 5-HT. Although inefficient on cells expressing 5-HT2B receptor alone, CP triggers also the internalization of coexpressed 5-HT2B receptor via Cav1, GRK2,3, and -arrestin-2 imposing for these two receptors the same kinetic for internalization. GRK2,3 becomes a partner of this trans-internalization of 5-HT2B receptors. C, the costimulation of both receptors by 5-HT makes 5-HT2B receptors to internalize via its classic clathrin but also via 5-HT1B triggered Cav1-dependent pathways (large arrow), which imposes in return 5-HT1B receptors to adopt new internalization pathways independent of both clathrin and Cav1 but that becomes dependent on 5-HT2B receptor-induced PKC (small arrow).
The coexpression of these two receptors influences their respective internalization kinetics according to the agonist used for stimulation, although in the absence of apparent colocalization (apparent lack of colocalization and of FRET). It is noteworthy that a slight acceleration of BW-induced internalization of 5-HT2B receptors can be observed when expressed with 5-HT1B receptors. However, a large acceleration of CP-induced 5-HT1B receptor internalization is triggered by the presence of 5-HT2B receptors. A marked modification of the thermodynamic profile of 5-HT2B receptor internalization is also observed in the presence of 5-HT1B receptors: the linear temperature dependence observed for internalization of 5-HT1B and 5-HT1B/5-HT2B receptors, but not for 5-HT2B receptors expressed alone, strongly supports the notion that coexpression of 5-HT1B receptors with 5-HT2B receptors imposes a alternate internalization pathway on 5-HT2B receptors. Stimulation with 5-HT in the presence of RS shows the existence of multiple pathways for 5-HT2B receptor internalization: whereas the 5-HT-induced internalization of 5-HT2B receptors expressed alone is blocked by RS, coexpression with 5-HT1B receptors renders 5-HT2B receptor internalization sensitive to CP and insensitive to RS. The antibody knock-down experiments further validate that activation of 5-HT1B receptors with CP in coexpressing cells results in a switch of 5-HT2B receptor internalization from a totally clathrin-, -arrestin-2-, and PKC-dependent pathway to a partially Cav1- and GRK2,3-dependent pathway (Fig. 7B).
Our experimental results implicate GRK2,3 in mediating the cross-regulation of 5-HT2B receptor internalization by 5-HT1B receptors. When 5-HT2B receptors are coexpressed with 5-HT1B receptors, CP activates 5-HT1B receptors that in turn activate GRK2,3, which leads to agonist-independent, Cav1-dependent internalization of 5-HT2B receptors. One likely possibility is that GRK2,3 phosphorylates 5-HT2B receptors in a way that enables them to use a Cav1-dependent internalization pathway (Fig. 7B).
Activation of 5-HT2B receptors by 5-HT was shown to lead to PKC activation (Cox and Cohen, 1995; Launay et al., 2006). Our data also support the notion that PKC is responsible for the effect of 5-HT2B receptors on 5-HT-induced 5-HT1B receptor internalization (Fig. 7C). Using a combination of pharmacological inhibitors and antibody knockdown, we have identified PKC as the isotype necessary for the 5-HT2B receptor-dependent 5-HT1B receptor internalization. PKC stimulation could phosphorylate—either directly or indirectly—the 5-HT1B receptor, rendering it unable to internalize via a Cav1-dependent pathway. One inconsistency between our experimental data and the proposed model is that activation of 5-HT2B receptors by BW does not affect the Cav1-dependent 5-HT1B receptor internalization. One possible explanation for this discrepancy is that the partial 5-HT2B receptor agonist BW poorly activates PKC. Alternatively, or additionally, both 5-HT-induced 5-HT1B receptor GRK2,3 activation and 5-HT-induced 5-HT2B receptor PKC activation are required for 5-HT1B receptors to internalize via the Cav1- and clathrin-independent pathway (Fig. 7C).
This cross-talk, which affects receptor internalization mechanics, is likely to explain the previously observed Gi uncoupling of 5-HT1B receptors by 5-HT2B receptors. Activation of the 5-HT2B/2C receptor has been shown to inhibit the 5-HT1B receptor function in two independent studies:
Using 5-HT2C and 5-HT2A receptors, stably transfected Chinese hamster ovary cells, it has been reported that activation of 5-HT2C receptors abolishes the endogenous 5-HT1B receptor-mediated inhibition of forskolin-stimulated cAMP accumulation. In contrast, activation of 5-HT2A receptors does not alter the 5-HT1B response and 5-HT2C receptor-mediated inhibition of 5-HT1B receptor function was blocked when 5-HT2A receptors were activated simultaneously (Berg et al., 1996).
Using the teratocarcinoma-derived cell line 1C11 that expresses 5-HT1B and 5-HT2B and then 5-HT2A receptors endogenously and sequentially, Tournois et al. (1998) showed that at day 2 of differentiation, when 5-HT1B and 5-HT2B receptor expression is induced, 5-HT2B receptors exert a dominant-negative regulation of the Gi-coupled 5-HT1B receptor; at day 4, when functional 5-HT2A receptors begin to be expressed, 5-HT2A receptor activation prevents the negative regulation exerted by 5-HT2B receptor on 5-HT1B receptor function. Therefore, it is plausible that 5-HT2B and 5-HT2C receptors should share a common intracellular internalization and signaling pathway by which to control 5-HT1B function, whereas 5-HT2A receptors use alternate pathway(s).
This newly described internalization route fits with other evidence supporting independent intracellular trafficking of 5-HT1B and 5-HT2B receptors (i.e., internalization of 5-HT2B receptors upon activation of 5-HT1B receptors despite the apparent lack of agonist-induced colocalization). Upon coexpression, stimulating one or the other receptor or both generates different cellular responses; this fact is supported by completely independent techniques (i.e., confocal microcopy image analysis and antibody knock-down coupled with whole-cell radioligand binding studies). This work provides the first evidence that one receptor may adopt different internalization pathways within the same cells upon the presence and stimulation of another receptor. Identified interactions regulate receptor internalization and could explain the observed coordination between 5-HT1B and 5-HT2B receptor internalization. Our work suggests that indirect events in trans are mediating the 5-HT1B/5-HT2B receptor cross-regulation that affects their cellular distribution during the endocytic process. Given the wide clinical use of 5-HT1B receptor agonists in the treatment of migraines, and the suspected prophylactic effect of 5-HT2B receptor antagonists, these newly identified functional interactions may be involved in therapeutic effects of these compounds. The phenomenon may also be relevant to the design of novel antimigraine therapies.
Acknowledgements
We thank P. Hickel for excellent technical assistance.
ABBREVIATIONS: 5-HT, 5-hydroxytryptamine; GPCR, G protein-coupled receptor; GRK, GPCR kinase; PKA, protein kinase A; PDZ, postsynaptic density 95/disc-large/zona occludens (PSD-95, Dlg, ZO-1); CFP, cyan fluorescent protein; YFP, yellow fluorescent protein; GFP, green fluorescent protein; RS, RS127445; RS-127445, 2-amino-4-(4-fluoronaphth-1-yl)-6-isopropylpyrimidine; CP, CP93129; CP93129, 1,4-dihydro-3-(1,2,5,6-tetrahydropyrid-4-yl)pyrrolo{3,2-b}pyridin-5-one dihydrochloride; BW, BW723C86; BW723C86, 1-{5-(2-thienylmethoxy)-1H-3-indolyl}propan-2-amine hydrochloride; H-89, N-{2-((p-bromocinnamyl)amino)ethyl}-5-isoquinolinesulfonamide dihydrochloride; Gö 6850-Bisindolylmaleimide I, 2-{1-(3-dimethylaminopropyl)-1H-indol-3-yl}-3-(1H-indol-3-yl)-maleimide; Gö 6976, 12-(2-cyanoethyl)-6,7,12,13-tetrahydro-13-methyl-5-oxo-5H-indolo(2,3-a)pyrrolo(3,4-c)-carbazole; [125I]DOI, [125I]1-(2,5-dimethoxy-4-iodophenyl)-2-aminopropane hydrochloride; [125I]GTI, [125I]serotonin-5-O-carboxymethyl-glycil-iodo-tyrosamine; PKC, protein kinase C; ROI, regions of interest; FRET, fluorescent resonance energy transfer; PA, photoactivatable; Cav-1, caveolin-1.
【参考文献】
Banes AK and Watts SW (2003) Arterial expression of 5-HT2B and 5-HT1B receptors during development of DOCA-salt hypertension. BMC Pharmacol 3: 12.
Berg KA, Maayani S, and Clarke WP (1996) 5-hydroxytryptamine2C receptor activation inhibits 5-hydroxytryptamine1B-like receptor function via arachidonic acid metabolism. Mol Pharmacol 50: 1017–1023.
Bhatnagar A, Sheffler DJ, Kroeze WK, Compton-Toth B, and Roth BL (2004) Caveolin-1 interacts with 5-HT2A serotonin receptors and profoundly modulates the signaling of selected Gq-coupled protein receptors. J Biol Chem 279: 34614–34623.[Abstract/Free Full Text]
Cox DA and Cohen ML (1995) 5-Hydroxytryptamine 2B receptor signaling in rat stomach fundus: role of voltage-dependent calcium channels, intracellular calcium release and protein kinase C. J Pharmacol Exp Ther 272: 143–150.[Abstract/Free Full Text]
Davies SP, Reddy H, Caivano M, and Cohen P (2000) Specificity and mechanism of action of some commonly used protein kinase inhibitors. Biochem J 351: 95–105.
Deraet M, Manivet P, Janoshazi A, Callebert J, Guenther S, Drouet L, Launay JM, and Maroteaux L (2005) The natural mutation encoding a C terminus-truncated 5-hydroxytryptamine2b receptor is a gain of proliferative functions. Mol Pharmacol 67: 983–991.[Abstract/Free Full Text]
George SR, O'Dowd BF, and Lee SP (2002) G-protein-coupled receptor oligomerization and its potential for drug discovery. Nat Rev Drug Discov 1: 808–820.
Gray JA, Bhatnagar A, Gurevich VV, and Roth BL (2003) The interaction of a constitutively active arrestin with the arrestin-insensitive 5-HT2A receptor induces agonist-independent internalization. Mol Pharmacol 63: 961–972.[Abstract/Free Full Text]
Hanley NR and Hensler JG (2002) Mechanisms of ligand-induced desensitization of the 5-hydroxytryptamine2A receptor. J Pharmacol Exp Ther 300: 468–477.[Abstract/Free Full Text]
Herrick-Davis K, Grinde E, Harrigan TJ, and Mazurkiewicz JE (2005) Inhibition of serotonin 5-hydroxytryptamine2c receptor function through heterodimerization: receptor dimers bind two molecules of ligand and one G-protein. J Biol Chem 280: 40144–40151.[Abstract/Free Full Text]
Herrick-Davis K, Grinde E, and Mazurkiewicz JE (2004) Biochemical and biophysical characterization of serotonin 5-HT2C receptor homodimers on the plasma membrane of living cells. Biochemistry 43: 13963–13971.
Hoyer D, Hannon JP, and Martin GR (2002) Molecular, pharmacological and functional diversity of 5-HT receptors. Pharmacol Biochem Behav 71: 533–554.
Ishida T, Hirata K, Sakoda T, Kawashima S, Akita H, and Yokoyama M (1999) Identification of mRNA for 5-HT1 and 5-HT2 receptor subtypes in human coronary arteries. Cardiovasc Res 41: 267–274.[Abstract/Free Full Text]
Ishida T, Kawashima S, Hirata K, and Yokoyama M (1998) Nitric oxide is produced via 5-HT1B and 5-HT2B receptor activation in human coronary artery endothelial cells. Kobe J Med Sci 44: 51–63.
Kalkman HO (1994) Is migraine prophylactic activity caused by 5-HT2B or 5-HT2C receptor blockade? Life Sci 54: 641–644.
Kellermann O, Loric S, Maroteaux L, and Launay J-M (1996) Sequential onset of three 5-HT receptors during the 5-hydroxytryptaminergic differentiation of the murine 1C11 cell line. Br J Pharmacol 118: 1161–1170.
Launay JM, Schneider B, Loric S, Da Prada M, and Kellermann O (2006) Serotonin transport and serotonin transporter-mediated antidepressant recognition are controlled by 5-HT2B receptor signaling in serotonergic neuronal cells. FASEB J 20: 1843–1854.[Abstract/Free Full Text]
Lefkowitz RJ and Shenoy SK (2005) Transduction of receptor signals by beta-arrestins. Science (Wash DC) 308: 512–517.[Abstract/Free Full Text]
Loric S, Maroteaux L, Kellermann O, and Launay J-M (1995) Functional serotonin-2B receptors are expressed by a teratocarcinoma-derived cell line during serotoninergic differentiation. Mol Pharmacol 47: 458–466.
Manders EM, Stap J, Brakenhoff GJ, van Driel R, and Aten JA (1992) Dynamics of three-dimensional replication patterns during the S-phase, analysed by double labelling of DNA and confocal microscopy. J Cell Sci 103: 857–862.
Manivet P, Mouillet-Richard S, Callebert J, Nebigil CG, Maroteaux L, Hosoda S, Kellermann O, and Launay J-M (2000) PDZ-dependent activation of nitric-oxide synthases by the serotonin 2B receptor. J Biol Chem 275: 9324–9331.[Abstract/Free Full Text]
Marion S, Weiner DM, and Caron MG (2004) RNA editing induces variation in desensitization and trafficking of 5-hydroxytryptamine 2c receptor isoforms. J Biol Chem 279: 2945–2954.[Abstract/Free Full Text]
Nicholson R, Small J, Dixon AK, Spanswick D, and Lee K (2003) Serotonin receptor mRNA expression in rat dorsal root ganglion neurons. Neurosci Lett 337: 119–122.
Poissonnet G, Parmentier JG, Boutin JA, and Goldstein S (2004) The emergence of selective 5-HT2B antagonists structures, activities and potential therapeutic applications. Mini Rev Med Chem 4: 325–330.
Porter RH, Malcolm CS, Allen NH, Lamb H, Revell DF, and Sheardown MJ (2001) Agonist-induced functional desensitization of recombinant human 5-HT2 receptors expressed in CHO-K1 cells. Biochem Pharmacol 62: 431–438.
Rapacciuolo A, Suvarna S, Barki-Harrington L, Luttrell LM, Cong M, Lefkowitz RJ, and Rockman HA (2003) Protein kinase A and G protein-coupled receptor kinase phosphorylation mediates -1 adrenergic receptor endocytosis through different pathways. J Biol Chem 278: 35403–35411.[Abstract/Free Full Text]
Schaerlinger B, Hickel P, Etienne N, Guesnier L, and Maroteaux L (2003) Agonist actions of dihydroergotamine at 5-HT2B and 5-HT2C receptors and their possible relevance to anti-migraine efficacy. Br J Pharmacol 140: 277–284.
Schmuck K, Ullmer C, Kalkman H, Probst A, and Lübbert H (1996) Activation of meningeal 5-HT2B receptors: an early step in the generation of migraine headache? Eur J Neurosci 8: 959–967.
Stefulj J, Jernej B, Cicin-Sain L, Rinner I, and Schauenstein K (2000) mRNA expression of serotonin receptors in cells of the immune tissues of the rat. Brain Behav Immun 14: 219–224.
Tournois C, Mutel V, Manivet P, Launay JM, and Kellermann O (1998) Cross-talk between 5-hydroxytryptamine receptors in a serotonergic cell line. Involvement of arachidonic acid metabolism. J Biol Chem 273: 17498–17503.[Abstract/Free Full Text]
Ullmer C, Schmuck K, Kalkman HO, and Lübbert H (1995) Expression of serotonin receptor mRNA in blood vessels. FEBS Lett 370: 215–221.
Watts SW, Baez M, and Webb RC (1996) The 5-hydroxytryptamine2B receptor and 5-HT receptor signal transduction in mesenteric arteries from deoxycorticosterone acetate-salt hypertensive rats. J Pharmacol Exp Ther 277: 1103–1113.[Abstract/Free Full Text]
Xia Z, Gray JA, Compton-Toth BA, and Roth BL (2003) A direct interaction of PSD-95 with 5-HT2A serotonin receptors regulates signal transduction and receptor trafficking. J Biol Chem 278: 21901–21908.[Abstract/Free Full Text]
Xie Z, Lee SP, O'Dowd BF, and George SR (1999) Serotonin 5-HT1B and 5-HT1D receptors form homodimers when expressed alone and heterodimers when co-expressed. FEBS Lett 456: 63–67.
作者单位：Institut National de la Santé et de la Recherche Médicale, U839, Paris, France (V.S., L.M.); Université Pierre et Marie Curie-Paris 6, Institut du Fer à Moulin, Unité Mixte de Recherche S0839, Paris, France (V.S., L.M.); Centre National de la Recherche Scientifique UMR7104, Illkirch, France (A.J., M