| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
NEUROSCIENCE |
1 Institut de Pharmacologie Moléculaire et Cellulaire, CNRS Université de Nice Sophia-Antipolis, UMR-6097, Institut Paul Hamel, 660 route des Lucioles, Sophia Antipolis, 06560 Valbonne, France
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionSupplemental materialReferences |
|---|
(Received 3 August 2005;
accepted after revision 9 November 2005;
first published online 10 November 2005)
Corresponding author M. Lazdunski: Institut de Pharmacologie Moléculaire et Cellulaire, CNRS Université de Nice Sophia-Antipolis, UMR-6097, Institut Paul Hamel, 660 route des Lucioles, Sophia Antipolis, 06560 Valbonne, France. Email: ipmc{at}ipmc.cnrs.fr
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionSupplemental materialReferences |
|---|
ASICs are predominantly expressed in both central and peripheral nervous systems (CNS and PNS, respectively) (Waldmann & Lazdunski, 1998; Krishtal, 2003) suggesting an important function of these channels in signal transduction associated with local extracellular pH variations during normal neuronal activity. Several studies have indicated that they are involved in synaptic plasticity, learning, memory and acquired fear-related behaviour (Bianchi & Driscoll, 2002; Wemmie et al. 2002; Chu et al. 2004). In the PNS, ASICs are expressed in nociceptive neurones (Waldmann & Lazdunski, 1998; Krishtal, 2003) where they are thought to be responsible for the perception of pain that accompanies tissue acidosis, particularly in muscle and cardiac ischaemia (Waldmann & Lazdunski, 1998; Benson et al. 1999; Kress & Zeilhofer, 1999; McCleskey & Gold, 1999; Pan et al. 1999; Immke & McCleskey, 2001; Sutherland et al. 2001; Chen et al. 2002; Immke & McCleskey, 2003), and to mediate cutaneous acid-induced pain (Price et al. 2001; Ugawa et al. 2002; Jones et al. 2004). ASICs have also been proposed to play a role in sour taste perception (Waldmann et al. 1997b; Liu & Simon, 2001; Lin et al. 2002), visual perception (Ettaiche et al. 2004) and in mechanosensation (Price et al. 2000, 2001; Page et al. 2004).
Since the cloning of the first ASIC in 1996 (Waldmann et al. 1996), six different members have now been cloned: ASIC1a (Waldmann et al. 1997b), ASIC1b (Chen et al. 1998), ASIC2a (Waldmann et al. 1996), ASIC2b (Lingueglia et al. 1997), ASIC3 (Waldmann et al. 1997a) and ASIC4 (Grunder et al. 2000); these members are encoded by four genes. ASIC1b and ASIC2b are splice variants of ASIC1a and ASIC2a, respectively. In both variants, the region comprising the intracellular N-terminus, the first transmembrane domain (M1) and the first third of the extracellular loop is exchanged. Psalmotoxin 1 (PcTx1), a tarantula peptide of 40 amino acids, is a specific blocker of ASIC1a but is devoid of effect on its splice variant ASIC1b (Escoubas et al. 2000, 2003). Considering the increasing number of important roles played by ASICs both in normal and pathological conditions, the emergence of specific toxins for ASICs (Escoubas et al. 2000, 2003; Diochot et al. 2004) provides a powerful tool for further understanding their involvement in neuronal excitability and pain coding.
The first purpose of the present study was to produce a recombinant variant toxin named PcTx1YN which could be radiolabelled by iodination (125I-PcTx1YN) to conveniently identify the toxin target; as has been previously and extensively been carried out for other toxin/channel pairs (Rehm & Lazdunski, 1988; Catterall, 2000; Bourinet et al. 2001; Frenal et al. 2004). The second purpose was to use this new radiolabelled tool, combined with an electrophysiological approach, to analyse independently the toxin binding parameters and the inhibitory effect on ASIC currents generated by ASIC1a and chimeras between ASIC1a and ASIC1b or between ASIC1a and ASIC2a. Our results identify structural elements involved in PcTx1 binding, in the ASIC1a extracellular loop.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionSupplemental materialReferences |
|---|
Oocyte preparation and cRNA injection have been previously described (Diochot et al. 1998). Animals were chilled on ice and humanely killed by decapitation after final collection of oocytes from both ovaries. After the first collection, the animal was allowed to recover from anaesthesia and surgery for at least 3 weeks before the second collection. Animal handling and experiments fully conformed with French regulations and were approved by local governmental veterinary services (authorization No. B 06-152-5 delivered by the Ministère de l'Agriculture, Direction des services vétérinaires). cRNA of rat ASIC1a was synthesized with the mCAP RNA capping kit from Stratagene. Oocytes were kept at 19°C in ND96 solution containing (in mM): 96 NaCl, 2 KCl, 1.8 CaCl2, 2 MgCl2 and 5 Hepes (pH 7.4 with NaOH); ND96 solution was supplemented with penicillin (6 µg ml–1) and streptomycin (5 µg ml–1). Currents were recorded 2–4 days after cRNA injection under voltage-clamp (Dagan TEV 200 amplifier, Dagan Corporation, Minneapolis, MN, USA) using two standard glass microelectrodes (1–2.5 M
) filled with a 3 mM KCl solution. Stimulation, data acquisition, and analysis were performed using pCLAMP software (Axon Instruments, Union City, CA, USA). All experiments were performed at 19–21°C in ND96 solution. Changes in extracellular pH were induced by a microperfusion system that allowed local and rapid changes of solutions. Hepes was replaced by Mes to buffer solutions with a pH between 6 and 5. Bovine serum albumin (BSA; 0.1%) was added to extracellular solutions containing PcTx1 to prevent its adsorption.
Patch-clamp recording of ASIC currents
COS-7 cells, seeded at a density of 20 000 cells per 35-mm diameter Petri dish, were cotransfected with pCI-rASIC1a or the same vector containing one of the constructs listed in Fig. 3, and with an expression vector containing the CD8 receptor cDNA (5:1 ratio) using the Diethylaminoethyl-Dextran (DEAE–dextran) method (Baron et al. 2001). Cells were used 1–3 days after transfection. Successfully transfected cells were recognized by their ability to bind CD8-antibody-coated beads (Dynal, Biotech ASA, Oslo, Norway). Currents were recorded by the whole-cell patch-clamp technique (Hamill et al. 1981). Data were sampled at 500 Hz and low-pass filtered at 3 kHz using the pClamp8 software (Axon Instruments). The statistical significance was estimated using Student's t-test. The pipette solution contained (in mM): 140 KCl, 5 NaCl, 2 MgCl2, 5 EGTA, 10 Hepes (pH 7.35). The bath solution contained (in mM): 150 NaCl, 5 KCl, 2 MgCl2, 2 CaCl2, 10 Hepes (pH 7.4). Mes was used instead of Hepes to buffer bath solution pH at levels ranging from 6 to 5. Currents were activated by a pH drop from pH 7.4, by shifting one out of eight outlets of a microperfusion system in front of the cell. Experiments were carried out at room temperature (20–24°C). BSA (0.1%) was added to extracellular solutions containing PcTx1 to prevent its adsorption to tubing and containers (Escoubas et al. 2000, 2003). Unless otherwise mentioned, PcTx1 was applied at the resting pH 7.4, before and between pH drops, and the maximal concentration tested was 100 nM.
|
A PcTx1 variant containing an extra tyrosine residue at the amino-terminal of the PcTx1 sequence (PcTx1YN) was produced recombinantly using the Drosophila S2 cell system as previously described for the production of recombinant PcTx1 (Escoubas et al. 2003). The additional tyrosine residue was introduced by PCR using a standard oligonucleotide strategy. Purification of PcTx1YN was carried out using the same combination of ion-exchange and reversed-phase batch chromatography and semipreparative HPLC as previously described (Escoubas et al. 2003). The molecular mass of the recombinant peptide was measured by MALDI-TOF (Perspective Biosystems Voyager DE-Pro, Applied Biosystems, Framingham, MA, USA)mass spectrometry and was found to be identical to the value calculated from sequence data (4849.24 Da monoisotopic mass). The ability of PcTx1YN to interact with ASIC1a channels was confirmed in Xenopus oocytes injected with ASIC1a cRNA.
MALDI-TOF mass spectrometry
Peptides or HPLC fractions were analysed on an Applied Biosystems Voyager DE-Pro system Applied Biosystems, Framingham, MA, USA) using 
-cyano-4-hydroxycinnamic acid matrix (10 mg ml–1).
Iodination of PcTx1YN
PcTx1YN was iodinated using the chloramine-T method (Hunter & Greenwood, 1962). Briefly, 125I-Na (0.75 nmol) was added to PcTx1YN (0.5 nmol) in 13.5 µl NaH2P04, pH 6.5 (100 mM final concentration). The reaction was initiated by adding three 15 µl aliquots of chloramine T (3 x 4.5 nmol) at 1 min intervals with gentle mixing. After 1 min of incubation at room temperature, the reaction mixture was diluted to 1 ml with 16% acetonitrile (ACN), 0.1% (v/v) trifluoroacetic acid (TFA) and loaded on an analytical Supelco C5 reversed-phase column (250 mm x 4.6 mm) for separation of mono-iodinated 125I-PcTxYN from unlabelled and di-iodinated toxin. The column was equilibrated at 16% ACN (0.1% TFA) then eluted with a linear gradient to 34% ACN in 60 min (0.3% ACN min–1) at 0°C. Flow rate was 1 ml min–1, and the elution was monitored at 280 nm. Fractions of 1 ml were collected in low-protein adsorption test tubes containing 50 µl BSA 1% (w/v), 1 M Hepes (pH 7.0) and 0.2% NaN3. An aliquot of each fraction was counted for 
-radiation. Fractions containing mono-iodinated 125I-PcTxYN (20–100 nM) were stored at 4°C.
Plasmid constructions and mutagenesis
Rat ASIC1a, rat ASIC1b and rat ASIC2a coding sequences (accession numbers U94403, AJ309926 and U53211, respectively), and their chimeras or point mutations were subcloned into the NheI/NotI restriction sites of pCI vector (Promega, Corporation, Madison, WI, USA) for expression in COS or Chinese hamster ovary (CHO) cells. The construction of chimeras (Fig. 3) between ASIC1a and ASIC1b, or between ASIC1a and ASIC2a, was carried out using recombinant PCR strategies (Ho et al. 1989) and was entirely verified by sequencing. Briefly, the different domains were amplified with primers containing the sequence of the desired junctions between ASIC1a and ASIC1b, or between ASIC1a and ASIC2a, and were joined by a second recombinant PCR with primers flanking the entire open reading frame.
Growth and transfection of CHO cells
Chinese hamster ovary (CHO) cells were maintained at 37°C in 5% CO2 Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum, penicillin/streptomycin mix, HT media supplement (Sigma), and proline 0.3 mM. Cells were harvested using trypsin/EDTA and washed once in serum-free OptiMEM before counting. Cells (15 million) were transfected by electroporation using a Biorad (Bio-Rad Life Science, Hercules, CA, USA) Genepulser Xcell (single square-wave pulse, 320 V, 15 ms duration, in a 4 mm Eurogentec cuvette-Eurogentec EGT Group, Selaing, Belgium) with 30 µg of plasmid coding for ASIC subunits (ASIC1a, 1b, 2a, 2b, 3 or 4) or chimeras (ASIC1b/1a, 2a/1a) or point mutants of ASIC1a (Fig. 3). Cells were plated onto 150-mm diameter culture dishes in 25 ml of medium and lysates made 3 days after transfection.
125I-PcTx1YN binding
CHO cells in 150-mm dishes were washed twice with Ca2+- and Mg2+-free phosphate-buffered saline, and lysates were made in 1 ml of 20 mM Hepes (pH 7.0), 0.2 mM EDTA, 1 mM phenylmethylsulphonic fluoride (PMSF), sonicated for 30 s and stored at –20°C. Protein concentrations of whole cell lysates were determined by the Bradford method. Rat brain membranes were prepared as previously described (Lambeau et al. 1989). Male 3-month-old Wistar rats were anaesthetized with isoflurane (2% inhaled concentration) and killed by decapitation. All binding assays were carried out at room temperature (20–24°C). Preliminary kinetic experiments indicated that the plateau for association of 125I-PcTx1YN was reached after 30–40 min and was stable for at least 2 hours. Therefore, in further binding assays, samples were incubated for 45–50 min. Conditions for radioligand binding studies were in a NaCl-containing buffer consisting of: 20 mM Hepes (pH 7.25), 5 mM KCl, 2 mM CaCl2, 2 mM MgCl2, 0.1% BSA and 150 mM NaCl. pH dependence, saturation and competition binding assays were done on both rat brain membranes and lysates of transfected CHO cells. A concentration of 125I-PcTx1YN around 100 pM (50–200 pM depending on specific activity of the ligand batch) was used in pH-dependence (Fig. 2B) and competition (Figs 4A, 5A and 6A) binding assays. Separation of bound and free 125I-PcTx1YN was achieved by filtration on 0.2 µm cellulose acetate filters (Sartorius AG, Goettinger, Germany) soaked in binding buffer (pH 7.25, 150 mM NaCl and 3% BSA), and washing with 10 ml binding buffer (pH 7.0, 150 mM NaCl). Saturation and competition binding data were analysed using the built-in equations in the GraphPad Prism software (v3.02, GraphPad Software Inc., San Diego, USA). Based on the linear Scatchard plot, competition curves were fitted according to a one-site competition equation.
|
|
|
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionSupplemental materialReferences |
|---|
Experiments were performed on ASIC1a currents expressed in COS cells and activated by successive pH drops separated by at least 60 s to allow the full reactivation of the current. As previously described (Escoubas et al. 2000), PcTx1 inhibited ASIC1a current in a concentration-dependent manner (IC50
1 nM; see also Figs 2A (
) and 4C (•)) when applied at the resting pH 7.4, and before and between the successive pH drops (see Fig. 1B for original current traces). The maximal level of inhibition, obtained after several current activations, was similar whether the toxin was applied only at the resting pH 7.4, or before and during the pH drops (not shown). The inhibition of ASIC1a by PcTx1 was the same at holding potentials of –50 and 0 mV (Fig. 1D). In order to determine whether the opening of ASIC1a channels was necessary for PcTx1-induced inhibition, we tested the frequency dependence of the inhibition. Figure 1A shows that the time course of PcTx1-induced inhibition depends on the total duration of the toxin application, independent of the number of pH drops during this application. The maximal inhibition induced by 1 nM PcTx1 (50% inhibition) is reached after a 3-min application. The PcTx1-induced inhibition was not dependent on the pH reached during the pH drop (Fig. 1C), but was reduced when the ASIC1a current was activated from resting pH 8 or 9 (Fig. 1E). Consequently, throughout this work, PcTx1 was applied at pH 7.4 before and between pH drops, and the maximal inhibition was measured after 3 min
|
In order to facilitate iodination, we produced a variant of PcTx1 containing an extra tyrosine residue at the amino-terminal extremity (PcTx1YN) in S2 cells and purified it with yields similar to that of wild-type PcTx1 (Escoubas et al. 2003). PcTx1YN caused concentration-dependent inhibition of ASIC1a currents in Xenopus oocytes with an IC50 of 1.98 ± 0.24 nM (n= 3), similar to the IC50 of 1.17 ± 0.11 nM (n= 3) for native PcTx1 (Fig. 2A). Binding of 125I-PcTx1YN depends on the pH of the buffer solution. This pH dependence is the same for ASIC1a expressed in both CHO cells and rat brain membranes (Fig. 2B). The curve is bell shaped with a pH of half-maximal activation (pH0.5) of 6.3 and 7.45 for the left and right slopes, respectively. Increasing the Ca2+ concentration in the buffer solution inhibits the binding of 125I-PcTx1YN with half-maximal inhibition at 3.2 mM Ca2+, no inhibition at 0.1 mM and full inhibition at 100 mM (not shown).
PcTx1 binds specifically to ASIC1a
Saturation and competition curves were determined on native channels in rat brain membranes as well as on heterologously expressed ASIC1a channels. 125I-PcTx1YN binding to both ASIC1a/CHO lysates and rat brain membranes was specific and saturable with Kd values of 213 ± 35 (n= 3) and 371 ± 48 pM (n= 4), respectively; and Bmax values of 36.9 ± 3.2 fmol mg–1 of CHO lysate protein (n= 3) and 49.5 ± 10.2 fmol mg–1 of brain membrane protein (n= 4) (Fig. 2C and D). The Scatchard plots (insets in Fig. 2C and D) are linear, indicating a single family of binding sites. Under the same binding conditions, competition assays using unlabelled wild-type PcTx1 resulted in IC50 values of 128 ± 7 (n= 5) and 149 ± 25 pM (n= 3) for ASIC1a/CHO lysates and rat brain membranes, respectively (Table 1). The differences in both the Kd and IC50 values for rat brain membranes and for ASIC1a/CHO lysates were not significant (P > 0.05, Student's unpaired t-test). PcTx1 is known to inhibit ASIC1a current (Escoubas et al. 2000), but it was not known whether it was able to bind to other members of the ASIC family without blocking their activity. At concentrations up to 1 nM, 125I-PcTx1YN appears to bind only to ASIC1a (Fig. 2E).
|
At 1 mM, amiloride blocks 100% of ASIC1a (IC50= 10 µM), the anti-inflammatory agent diclofenac blocks 100% of ASIC3 (IC50= 92 µM), and the analgesics ibuprofen or flurbiprofen block 80% of ASIC1a (IC50= 349 µM) (Voilley et al. 2001). At this concentration, amiloride, flurbiprofen or diclofenac did not show any competition with 125I-PcTx1YN binding to rat brain membranes (Fig. 2F). The same result was obtained with ibuprofen (not shown). FMRFamide and related peptides have also been shown to interact with both ASIC1a and ASIC3 currents (Catarsi et al. 2001; Perry et al. 2001; Deval et al. 2003; Xie et al. 2003). These peptides affect channel gating as they can slow down desensitization and increase current amplitude. FMRFamide has an EC50 of 50 µM on dorsal root ganglion H+-gated currents (Deval et al. 2003). FMRFamide (up to 500 µM) showed no significant effect on 125I-PcTx1YN binding to rat brain membranes (Fig. 2G).
Chimeric construction strategy to explore the PcTx1 binding site and its mechanism of inhibition
For this study, the extracellular loop was divided into two main parts separated by the splice junction (Fig. 3A), and each part was divided into three domains. The limits between each domain were chosen in conserved regions to avoid important global structural modifications. These six domains were used in combination to construct the different chimeras (Fig. 3B and C). For example, the introduction of domains 1 and 2 from ASIC1a to ASIC1b gives the chimera named 1b/1a:12 (see chimeras ASIC1b/1a, Fig. 3B). The second part of the loop is identical between ASIC1a and ASIC1b, therefore requiring chimeric constructions between ASIC1a and ASIC2a to explore its contribution to the binding site of PcTx1. Thus, the first part of the loop has been transferred from ASIC1a to ASIC2a either alone, or with all the possible combinations of domains 4, 5 and 6 (see chimeras ASIC2a/1a, Fig. 3C). All the residues or regions of the ENaC/DEG family with known structure or function, and their overlap with the six different domains, are indicated in Figs 3A and 6.
All chimeras expressed in COS cells generated ASIC currents in response to a pH drop. Some functional characteristics are summarized in the right part of Table 1 (pH dependence of activation and inactivation time constant). These data confirmed that pH0.5 is principally determined by the first part of the extracellular loop. The IC50 values for PcTx1YN binding and PcTx1-induced inhibition of currents were compared by their relative variations from their control values (ratio columns in Table 1).
The N-terminal cytoplasmic and the first transmembrane M1 segments of ASIC1a are not involved in the PcTx1 binding site
In the 1b/1a:123 chimera (Fig. 3B), containing the extracellular loop of ASIC1a with the N-terminus cytoplasmic and the transmembrane M1 segments of ASIC1b (different from those of ASIC1a), toxin binding was normal (IC50 of 127 ± 10 versus 128 ± 7 pM for ASIC1a) (Fig. 4A and Table 1) and the current inhibition was only slightly lower (IC50 of 1.62 nMversus 1.17 nM for ASIC1a) (Fig. 4B and C, and Table 1). Therefore, the N-terminus cytoplasmic and the transmembrane M1 segments of ASIC1a are not essential for the binding of PcTx1, while the first part of the extracellular loop (i.e. domains 1, 2 and 3) of ASIC1a appears necessary to build up a full binding site in the ASIC1b scaffold.
The CRDI (or domain 3) of ASIC1a is involved in the PcTx1 binding site
Domains 1, 2 and 3 were individually transferred from ASIC1a to ASIC1b (Fig. 3B and Table 1). Only domain 3, which corresponds to CRDI, is able to create a toxin binding site albeit with an affinity 2.6-fold lower than that of ASIC1a (IC50 of 337 ± 22 versus 128 ± 7 pM for ASIC1a) (Fig. 4A and Table 1). PcTx1YN was able to bind to all chimeras containing domain 3 of ASIC1a (Fig. 4A), but the inhibition of the current induced by 10 nM PcTx1, a concentration nearly fully inhibiting ASIC1a, was only significant for the two chimeras 1b/1a:123 and 1b/1a:13 (Fig. 4B and D). Indeed, the IC50 values of the PcTx1-induced current inhibition were 14.8, 31.6 and 136 nM for 1b/1a:13, 1b/1a:23 and 1b/1a:3, respectively; compared with the control value of 1.17 nM for ASIC1a (Fig. 4C). Furthermore, PcTx1 could neither bind to nor inhibit 1b/1a:12, further supporting the proposed crucial role of domain 3 in the binding of PcTx1.
The post-M1 region (domains 1 and 2) is necessary to mediate the inhibitory effect of PcTx1 and influences PcTx1 binding on CRDI (domain 3)
When domain 3 is transferred together with domain 1 from ASIC1a to ASIC1b, the binding affinity is not significantly improved, as the 1b/1a:13 chimera has an IC50 value of 307 ± 43 pM as compared with an IC50 of 337 ± 22 pM for the 1b/1a:3 chimera. In contrast, the inhibitory effect is strongly increased (a 9-fold decrease in IC50 from 136 nM for 1b/1a:3 to 14.8 nM for 1b/1a:13). These data, and the absence of PcTx1 binding on the 1b/1a:1 chimera (Table 1), indicate that domain 1 is not essential for PcTx1 binding, but is probably very important for PcTx1 to induce current inhibition. When domains 2 and 3 are transferred together (1b/1a:23 chimera), the inhibitory effect is increased 4.3-fold (IC50 values of 136 nM for 1b/1a:3 and 31.6 nM for 1b/1a:23 chimera), correlated with a 3.5-fold increase in binding affinity (IC50 values of 337 ± 22 pM for 1b/1a:3 and 95 ± 7 pM for 1b/1a:23). Even if domain 2 does not constitute the main PcTx1 binding site (as shown by the absence of toxin binding on the 1b/1a:2 chimera (Table 1)), it influences the binding of the toxin to domain 3. This could be related to the variation in size of domain 2 between ASIC1a and ASIC1b (Fig. 3A), and could explain the difference between 1b/1a:3 and 1b/1a:23 chimeras. Several point mutations were made in domain 2 of ASIC1a, supporting its regulatory role in PcTx1 binding. Results are provided as supplemental material (Supplemental Results and Suppl. Figure 1).
The first part of the extracellular loop is necessary but not sufficient for the PcTx1 binding site
In contrast to ASIC1b, the transfer of the first part of the extracellular loop of ASIC1a into ASIC2a (chimera 2a/1a:123) did not confer toxin binding (Table 1), and the current remained insensitive to PcTx1 (Fig. 5B). In contrast, the introduction of the entire extracellular loop (domains 1–6) of ASIC1a into ASIC2a confers binding with an IC50 of 310 ± 14 pM and current inhibition with a slightly higher IC50 of 5.37 nM (chimera 2a/1a:123456; Fig. 5 and Table 1). The fact that the extracellular loop of ASIC1a is sufficient by itself to restore a proper binding site for PcTx1 in ASIC2a allows us to exclude any important contribution of the transmembrane segment M1 or M2 or of the N- and C-terminal cytoplasmic tails to toxin binding. The results obtained with these two chimeras also indicate that, in addition to the first part of the loop, one or several regions located in the second part of the loop is/are necessary to constitute the complete PcTx1 binding site.
|
The transfer of domain 4 (chimera 2a/1a:1234) or 6 (chimera 2a/1a:1236), alone or in combination (chimera 2a/1a:12346), did not restore PcTx1 binding or inhibition, suggesting that they do not constitute the principal target for the toxin (Fig. 5B and Table 1). The absence of domain 6 in chimera 2a/1a:12345 did not affect binding (IC50 value of 458 ± 43 pM for chimera 2a/1a:12345 versus 310 ± 14 pM for chimera 2a/1a:123456), whereas the inhibitory effect is dramatically affected (IC50 value of >400 nM for chimera 2a/1a:12345 versus 5.37 nM for chimera 2a/1a:123456). This supports an important role of domain 6 in the mechanism of PcTx1-induced current inhibition but not in the toxin binding site. Only chimeras that contain domain 5 of ASIC1a (i.e. CRDII), in addition to the first part of the extracellular loop, showed a current affected by PcTx1, suggesting a central role for CRDII (Fig. 5D and Table 1; chimeras 2a/1a:123456, 2a/1a:1235, 2a/1a:12345 and 2a/1a:12356). Amazingly, the current generated by chimeras ASIC2a/1a:1235 and ASIC2a/1a:12356 containing domain 5 were found to be increased by PcTx1 (Fig. 5D), whereas no binding could be measured. The increase in current amplitude was clearly concentration dependent for chimera ASIC2a/1a:1235, with a 21 ± 7% (n= 9) increase induced by 10 nM PcTx1 (Fig. 5B and D), a 39 ± 17% increase (n= 4) induced by 20 nM PcTx1 and up to a 2- to 3-fold increase with 30–100 nM PcTx1. With the chimera ASIC2a/1a:12356, an increase in current amplitude was only observed with 100 nM PcTx1 (Fig. 5D). This suggests an apparent affinity of the toxin for these chimeras in the 10–100 nM range compared with an IC50 of 1 nM for its inhibitory effect on ASIC1a. In both chimeras, the increase in current amplitude was associated with a PcTx1-induced slowing of the inactivation rate, which could explain – at least in part – the increase in current amplitude. The ASIC2a/1a:1235 current activated at pH 5 showed an increase in 
from 0.20 ± 0.02 (n= 14) up to 0.41 ± 0.04 s (n= 8, P < 0.001) in the presence of 10 nM PcTx1, whereas the ASIC2a/1a:12356 current activated at pH 5 showed an increase in from 1.04 ± 0.12 (n= 8) up to 1.93 ± 0.4 s (n= 4, P < 0.05) in the presence of 10 nM PcTx1 and up to 4.17 ± 1.18 s (n= 4, P < 0.005) in the presence of 100 nM PcTx1. In these chimeras, the PcTx1-induced slowing down of inactivation suggests that PcTx1 stabilizes the open state rather than a closed state, as appears to be the case in wild-type ASIC1a (see Discussion and Chen et al. (2005)). This hypothesis could explain the lower affinity that leads to difficulty in identifying significant 125I-PcTx1YN binding to both chimeras (Table 1). The PcTx1-induced increase in ASIC2a/1a:1235 current amplitude was similar at holding potentials of –50 mV and 0 mV, and was still observed when the current was activated from a resting pH of 8 (Suppl. Figure 2). Results obtained from point mutations made in domain 5 of ASIC1a, suggest that D349 could participate directly in the binding site (Supplemental Results and Suppl. Figure 1).
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionSupplemental materialReferences |
|---|
The binding of 125I-PcTx1YN to recombinant and native channels is saturable, specific and of very high affinity. In both cases, 125I-PcTx1YN binds to a single family of sites with similar high affinities, suggesting that ASIC1a is indeed the only specific target of PcTx1 in CNS tissue. FMRFamide peptide, which potentiates ASIC current (Lingueglia et al. 1995), has no effect on the binding of 125I-PcTx1YN to ASIC1a, suggesting that it recognizes a different modulatory site on the channel (Fig. 2G). In contrast, several non-steroidal anti-inflammatory drugs (NSAIDs) have been shown to have a direct inhibitory effect on ASICs (Voilley et al. 2001). Neither ibuprofen nor its analogue flurbiprofen, both of which inhibit ASIC1a, inhibited 125I-PcTx1YN binding. Diclofenac, which inhibits ASIC3 but not ASIC1a, was also without effect (Fig. 2F). Therefore, PcTx1 binds to a unique site on ASIC1a, whose occupancy results in a profound inhibitory effect, thus making it a very attractive target for the development of novel therapeutics.
PcTx1 does not bind directly to the ion pore
Several groups have identified different functional regions of ENaC/DEG/ASICs (Fig. 3A and Fig. 6), strongly suggesting that transmembrane M1 and M2 segments and intracellular pre-M1 and post-M2 regions are involved in the pore structure. Our approach with chimeras has suggested that these regions are not involved in either the PcTx1 binding site or in its mechanism of inhibition of ASIC1a. The fact that amiloride, a known pore blocker of ENaC/DEG/ASICs (Schild et al. 1997), does not inhibit PcTx1 binding also supports this conclusion. Moreover, a physical occlusion of the ion pore is not compatible with the PcTx1-induced stimulatory effect observed with both chimeras 2a/1a:1235 and 2a/1a:12356. Our results are fully supported by a very recently published paper (Chen et al. 2005) showing that PcTx1 acts as a gating modifier on ASIC1a, by shifting the channel from its resting towards its inactivated state through an increase of its apparent affinity for protons, the natural ligands of ASICs. The PcTx1-induced shift of the pH-dependent inactivation curve of ASIC1a towards higher pH values was found to be Ca2+ dependent, increasing extracellular Ca2+ resulting in a decrease of the PcTx1 inhibitory effect (Chen et al. 2005). We also observed that Ca2+ inhibits PcTx1YN binding with a half inhibition at 3.2 mM Ca2+. The reduction in the inhibitory effect of PcTx1 observed when ASIC1a current was activated from a higher resting pH (Fig. 1E) is also in favour of the mechanism proposed by Chen et al. (2005).
Patch-clamp and 125I-PcTx1YN binding data provide complementary information to localize the PcTx1 site
Results obtained using the patch-clamp technique and using 125I-PcTx1YN binding are not always superimposable. Whereas binding experiments were realized on isolated membrane fractions exposed to symmetrical ionic concentrations, at 0 mV, and on presumably closed channels (probably in the inactivated form), patch-clamp experiments were carried out on functional channels activated by repeated pH drops with asymmetrical ionic concentrations. Presumably, IC50 values for 125I-PcTx1YN binding reflect only binding affinities for channels (ASIC1a or chimeras) in the closed state, whereas IC50 values for inhibition reflect the binding of PcTx1 to its receptor followed by a modification of ASIC1a gating (Chen et al. 2005). When only PcTx1 binding is altered in a chimeric construct, changes in IC50 values for binding and inhibition are expected to be correlated. When only the ability of PcTx1 to modify channel gating is altered, a difference between relative IC50 variations of binding and inhibition is expected. This is the case for chimera 1b/1a:23, where the absence of domain 1 from ASIC1a decreases the PcTx1-induced inhibition (27-fold weaker than that of ASIC1a) despite an IC50 of binding similar to ASIC1a. Similar results were also obtained with domain 6. CRDI was identified as a necessary part of the binding site mainly based on 125I-PcTx1YN binding results. The important role of CRDII was proposed mainly based on patch-clamp data. This later conclusion mainly stems from the observed increase in current amplitude (instead of an inhibition) of 2a/1a:1235 and 2a/1a:12356 chimeras.
Role of cysteine-rich domains in the PcTx1 binding site
Our results show that CRDI and CRDII are the principal targets for PcTx1 binding and that the post-M1 region (i.e. domains 1 and 2) and the pre-M2 region (i.e. domain 6) play a crucial role in the ability of PcTx1 to modify channel gating. These results are supported by the fact that the ASIC pore structure has been shown to be regulated by structural elements located in the extracellular loop (Figs 3A and 6). Interestingly, the action of trypsin on the first part of the ASIC1a extracellular loop prevents inhibition by PcTx1 (Poirot et al. 2004). We cannot rule out an involvement of other extracellular regions that are conserved within the ASIC family. However, if conserved regions were to contain key residues for association with the toxin, one would probably expect all ASICs to show at least some binding capacity for PcTx1, as well as some inhibition at high toxin concentrations, which is not the case (Fig. 2E).
The role of domain 4 between CRDI and CRDII
PcTx1 binding is undetectable in chimeras 2a/1a:1235 and 2a/1a:12356 (Table 1) which contain domain 4 of ASIC2a. This could be due to the affinity being too low. The PcTx1-induced increase in the current generated by these chimeras is observed with PcTx1 concentrations from 10 to 100 nM (and 100 nM is probably still not high enough to produce a maximal effect). Conversely, the PcTx1-induced inhibition of ASIC1a is complete at around 10 nM. The number of 125I-PcTx1YN binding sites on ASIC1a is low, in the tens of fmol mg–1 range. Therefore, a decrease in affinity by a factor of 10, or more, probably renders detection impossible. A stimulatory effect of PcTx1 on wild-type ASIC1a has very recently been reported by Chen et al. (2005), showing that, although under unusual experimental conditions, PcTx1 could induce a brief opening of ASIC1a on its way to the inactivated state. Such effects are seen with chimeras 2a/1a:1235 and 2a/1a:12356, and apparently more pronounced. Interestingly, binding as well as inhibitory effects were restored when domain 4 of ASIC1a was reintroduced into PcTx1-stimulated chimeras (Fig. 5D and Table 1). Located between CRDI and CRDII, both involved in PcTx1 binding, domain 4 may determine their correct positioning. A computerized data analysis approach revealed that CRDI and CRDII could be involved in an intramolecular interaction (Tavernarakis & Driscoll, 2000). Moreover, His162 and His339 (Baron et al. 2001), which are essential for the Zn2+-potentiating effect on ASIC2a, are found in CRDI and CRDII, respectively; suggesting that they play an important role in the global rearrangement of the extracellular loop modulating channel gating in the ASIC family. A functional relationship between the post-M1 domain and residues in CRDII has also been recently proposed by Coric et al. (2005). A disulphide bond within the ENaC extracellular loop has been described (Firsov et al. 1999), corresponding to Cys93–Cys193 in ASIC1a (Figs 3A and 6). This putative disulphide bridge between domains 1 and 4 supports a close relationship between proton-induced conformational changes of the post-M1 region and a rearrangement of the CRDI–domain 4–CRDII region where PcTx1 acts.
The production of iodinated PcTx1YN and binding assays, in addition to electrophysiological recordings, has allowed us to identify the regions of ASIC1a that interact with PcTx1. This work provides several elements that help to understand the mechanism of inhibition by PcTx1. The prominent role for ASICs as mediators of acid-induced pain emphasizes the importance and interest of this work in understanding and treating this type of pain. ASICs might also play an important role in pathological situations such as brain ischaemia or epilepsy, which produce significant extracellular acidification (Biagini et al. 2001; Yermolaieva et al. 2004). Moreover, the findings of Xiong et al. (2004), showing that ASIC1a blockers and gene knockout can dramatically reduce the amount of brain injury after ischaemia, suggest that targeting these channels may prove to be an effective therapy for stroke. It has also been suggested that PcTx1 may be helpful both for diagnostic and therapeutic treatments of aggressive malignant gliomas (Bubien et al. 2004). The fact that PcTx1 does not compete with NSAIDs, amiloride or FMRFamide, indicates that the binding site of PcTx1 constitutes a new site of action on the ASIC1a for the development of novel and very specific ASIC blockers, thus leading to the putative development of new neuroprotective agents and analgesics. The binding assay using iodinated PcTx1YN may constitute a powerful tool for the pharmacological screening of compounds active on ASIC1a.
|
Supplemental material |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionSupplemental materialReferences |
|---|
This material can also be found as part of the full-text HTML version available from http://www.blackwell-synergy.com
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionSupplemental materialReferences |
|---|
Baron A, Deval E, Salinas M, Lingueglia E, Voilley N & Lazdunski M (2002a). Protein kinase C stimulates the acid-sensing ion channel ASIC2a via the PDZ domain-containing protein PICK1. J Biol Chem 277, 50463–50468.
Baron A, Schaefer L, Lingueglia E, Champigny G & Lazdunski M (2001). Zn2+ and H+ are coactivators of acid-sensing ion channels. J Biol Chem 276, 35361–35367.
Baron A, Waldmann R & Lazdunski M (2002b). ASIC-like, proton-activated currents in rat hippocampal neurons. J Physiol 539, 485–494.
Bassler EL, Ngo-Anh TJ, Geisler HS, Ruppersberg JP & Grunder S (2001). Molecular and functional characterization of acid-sensing ion channel (ASIC) 1b. J Biol Chem 276, 33782–33787.
Benson CJ, Eckert SP & McCleskey EW (1999). Acid-evoked currents in cardiac sensory neurons: a possible mediator of myocardial ischemic sensation. Circ Res 84, 921–928.
Biagini G, Babinski K, Avoli M, Marcinkiewicz M & Seguela P (2001). Regional and subunit-specific downregulation of acid-sensing ion channels in the pilocarpine model of epilepsy. Neurobiol Dis 8, 45–58.[CrossRef][Medline]
Bianchi L & Driscoll M (2002). Protons at the gate: DEG/ENaC ion channels help us feel and remember. Neuron 34, 337–340.[CrossRef][Medline]
Bourinet E, Stotz SC, Spaetgens RL, Dayanithi G, Lemos J, Nargeot J & Zamponi GW (2001). Interaction of SNX482 with domains III and IV inhibits activation gating of 1E (CaV2.3) calcium channels. Biophys J 81, 79–88.
Bubien JK, Ji HL, Gillespie GY, Fuller CM, Markert JM, Mapstone TB & Benos DJ (2004). Cation selectivity and inhibition of malignant glioma Na+ channels by Psalmotoxin 1. Am J Physiol Cell Physiol 287, C1282–C1291.
Canessa CM, Horisberger JD & Rossier BC (1993). Epithelial sodium channel related to proteins involved in neurodegeneration. Nature 361, 467–470.[CrossRef][Medline]
Canessa CM, Schild L, Buell G, Thorens B, Gautschi I, Horisberger JD & Rossier BC (1994). Amiloride-sensitive epithelial Na+ channel is made of three homologous subunits. Nature 367, 463–467.[CrossRef][Medline]
Catarsi S, Babinski K & Seguela P (2001). Selective modulation of heteromeric ASIC proton-gated channels by neuropeptide FF. Neuropharmacology 41, 592–600.[CrossRef][Medline]
Catterall WA (2000). From ionic currents to molecular mechanisms: the structure and function of voltage-gated sodium channels. Neuron 26, 13–25.[CrossRef][Medline]
Champigny G, Voilley N, Waldmann R & Lazdunski M (1998). Mutations causing neurodegeneration in Caenorhabditis elegans drastically alter the pH sensitivity and inactivation of the mammalian H+-gated Na+ channel MDEG1. J Biol Chem 273, 15418–15422.
Chen CC, England S, Akopian AN & Wood JN (1998). A sensory neuron-specific, proton-gated ion channel. Proc Natl Acad Sci U S A 95, 10240–10245.
Chen X, Kalbacher H & Grunder S (2005). The tarantula toxin psalmotoxin 1 inhibits acid-sensing ion channel (ASIC) 1a by increasing its apparent H+ affinity. J General Physiol 126, 71–79.
Chen CC, Zimmer A, Sun WH, Hall J & Brownstein MJ (2002). A role for ASIC3 in the modulation of high-intensity pain stimuli. Proc Natl Acad Sci U S A 99, 8992–8997.
Chu XP, Wemmie JA, Wang WZ, Zhu XM, Saugstad JA, Price MP, Simon RP & Xiong ZG (2004). Subunit-dependent high-affinity zinc inhibition of acid-sensing ion channels. J Neurosci 24, 8678–8689.
Coric T, Zhang P, Todorovic N & Canessa CM (2003). The extracellular domain determines the kinetics of desensitization in acid-sensitive ion channel 1. J Biol Chem 278, 45240–45247.
Coric T, Zheng D, Gerstein M & Canessa CM (2005). Proton-sensitivity of ASIC1 appeared with the rise of fishes by changes of residues in the region that follows TM1 in the ectodomain of the channel. J Physiol.
Coscoy S, Lingueglia E, Lazdunski M & Barbry P (1998). The Phe-Met-Arg-Phe-amide-activated sodium channel is a tetramer. J Biol Chem 273, 8317–8322.
Cottrell GA, Jeziorski MC & Green KA (2001). Location of a ligand recognition site of FMRFamide-gated Na+ channels. FEBS Lett 489, 71–74.[CrossRef][Medline]
Deval E, Baron A, Lingueglia E, Mazarguil H, Zajac JM & Lazdunski M (2003). Effects of neuropeptide SF and related peptides on acid sensing ion channel 3 and sensory neuron excitability. Neuropharmacology 44, 662–671.[CrossRef][Medline]
Diochot S, Baron A, Rash LD, Deval E, Escoubas P, Scarzello S, Salinas M & Lazdunski M (2004). A new sea anemone peptide, APETx2, inhibits ASIC3, a major acid-sensitive channel in sensory neurons. EMBO J 23, 1516–1525.[CrossRef][Medline]
Diochot S, Schweitz H, Beress L & Lazdunski M (1998). Sea anemone peptides with a specific blocking activity against the fast inactivating potassium channel Kv3.4. J Biol Chem 273, 6744–6749.
Escoubas P, Bernard C, Lambeau G, Lazdunski M & Darbon H (2003). Recombinant production and solution structure of PcTx1, the specific peptide inhibitor of ASIC1a proton-gated cation channels. Protein Sci 12, 1332–1343.
Escoubas P, De Weille JR, Lecoq A, Diochot S, Waldmann R, Champigny G, Moinier D, Menez A & Lazdunski M (2000). Isolation of a tarantula toxin specific for a class of proton-gated Na+ channels. J Biol Chem 275, 25116–25121.
Ettaiche M, Guy N, Hofman P, Lazdunski M & Waldmann R (2004). Acid-sensing ion channel 2 is important for retinal function and protects against light-induced retinal degeneration. J Neurosci 24, 1005–1012.
Firsov D, Gautschi I, Merillat AM, Rossier BC & Schild L (1998). The heterotetrameric architecture of the epithelial sodium channel (ENaC). EMBO J 17, 344–352.[CrossRef][Medline]
Firsov D, Robert-Nicoud M, Gruender S, Schild L & Rossier BC (1999). Mutational analysis of cysteine-rich domains of the epithelium sodium channel (ENaC). Identification of cysteines essential for channel expression at the cell surface. J Biol Chem 274, 2743–2749.
Frenal K, Xu CQ, Wolff N, Wecker K, Gurrola GB, Zhu SY, Chi CW, Possani LD, Tytgat J & Delepierre M (2004). Exploring structural features of the interaction between the scorpion toxinCnErg1 and ERG K+ channels. Proteins 56, 367–375.[CrossRef][Medline]
Grunder S, Geissler HS, Bassler EL & Ruppersberg JP (2000). A new member of acid-sensing ion channels from pituitary gland. Neuroreport 11, 1607–1611.[Medline]
Grunder S, Jaeger NF, Gautschi I, Schild L & Rossier BC (1999). Identification of a highly conserved sequence at the N-terminus of the epithelial Na+ channel alpha subunit involved in gating. Pflugers Arch 438, 709–715.[CrossRef][Medline]
Hamill OP, Marty A, Neher E, Sakmann B & Sigworth FJ (1981). Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflugers Arch 391, 85–100.[CrossRef][Medline]
Ho SN, Hunt HD, Horton RM, Pullen JK & Pease LR (1989). Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77, 51–59.[CrossRef][Medline]
Hong K & Driscoll M (1994). A transmembrane domain of the putative channel subunit MEC-4 influences mechanotransduction and neurodegeneration in C. elegans. Nature 367, 470–473.[CrossRef][Medline]
Huang M & Chalfie M (1994). Gene interactions affecting mechanosensory transduction in Caenorhabditis elegans. Nature 367, 467–470.[CrossRef][Medline]
Hunter WM & Greenwood FC (1962). Preparation of iodine-131 labelled human growth hormone of high specific activity. Nature 194, 495–496.[CrossRef][Medline]
Immke DC & McCleskey EW (2001). Lactate enhances the acid-sensing Na+ channel on ischemia-sensing neurons. Nat Neurosci 4, 869–870.[CrossRef][Medline]
Immke DC & McCleskey EW (2003). Protons open acid-sensing ion channels by catalyzing relief of Ca2+ blockade. Neuron 37, 75–84.
) just before the second transmembrane domain (
) crucial for Ca2+ block of ASIC1a (
) (
) and selectivity filter on the first and second transmembrane domain are also indicated (
) and ASIC1a/CHO cell lysates (•) (n= 2–3). 125I-PcTx1YN concentrations of around 100 pM (50–200 pM depending on specific activity) were used. C, saturation curve of 125I-PcTx1YN binding to ASIC1a/CHO lysate (Kd= 213.7 ± 35.2 pM, Maximum Binding, Bmax= 36.9 ± 3.2 fmol mg–1, n= 3); inset, linear Scatchard plot indicating a single family of binding sites. D, saturation curve of 125I-PcTx1YN binding to rat brain membranes (Kd= 371.6 ± 48.5 pM, Bmax= 49.5 ± 10.2 fmol mg–1, n= 4); inset, linear Scatchard plot indicating a single family of binding sites. E, selectivity of 125I-PcTx1YN (1 nM) binding to lysates of CHO cells expressing different ASIC subtypes (n= 2). In this experiment, the cell lysates came from a different transfection to that in panel C, thus explaining the change in Bmax value. F, other known inhibitors of ASIC1a (amiloride (Amil) and flurbiprofen (Flurb)) and of ASIC3 (diclofenac (Dicl)), up to a concentration of 1 mM, failed to compete with 125I-PcTx1YN (100 pM) binding to rat brain membranes (n= 2). 125I-PcTx1YN binding was displaced by 50 nM PcTx1 (Tx). G, FMRFamide, a positive modulator of ASIC1a and ASIC3, up to a concentration of 500 µM, failed to compete with 125I-PcTx1YN (100 pM) binding to rat brain membranes (n= 2). 125I-PcTx1YN binding was displaced by 50 nM PcTx1 (Tx).
, the first currents recorded after PcTx1 applications for different durations; •, currents recorded after repeated activation by successive pH drops. Results were obtained from 11 different cells. Currents were activated by pH drops from pH 7.4 at a holding potential of –50 mV. Inset, original currents traces showing the inhibition of ASIC1a current by 1 nM PcTx1 (Tx 1) after 60 s application (first recorded current) and 180 s. B, original current traces recorded from the same cell showing the inhibition induced by different durations of 10 nM PcTx1 (Tx 10) application (top and middle panels). The first recorded currents are indicated (1st). The washout of PcTx1 10 nM inhibition is illustrated in the bottom traces. Currents were activated by pH drops from pH 7.4 to pH 6 every 2 min at a holding potential of –50 mV. Variation of control current amplitude is due to a spontaneous run-down of ASIC1a current throughout the experiment. C, PcTx1-induced inhibition of ASIC1a current is independent of the pH reached during the pH drop. ASIC1a currents were activated by pH drops from a resting pH of 7.4 down to the pH value indicated below (test pH). Holding potential, –50 mV. The amplitude of 1 nM PcTx1-inhibited ASIC1a current was measured at steady state, expressed as a percentage of control current and plotted as mean ±S.E.M. (n= 5–8). D, PcTx1-induced inhibition of ASIC1a current is independent of the holding potential. ASIC1a currents were activated by pH drops from a resting pH 7.4 down to pH 5 at the holding potentials indicated below. The amplitude of ASIC1a currents in the presence of 1 nM PcTx1 (filled bars) or 10 nM PcTx1 (shaded bars) was measured at steady state, expressed as a percentage of control current and plotted as mean ±S.E.M. (n= 3–14). E, PcTx1-induced inhibition of ASIC1a current is dependent on the resting pH value. ASIC1a currents were activated by pH drops from the resting pH values indicated below each bar down to pH 5. Holding potential, –50 mV. The amplitude of ASIC1a currents in the presence of 1 nM PcTx1 (filled bars) or 10 nM PcTx1 (shaded bars) or 100 nM PcTx1 (open bar) was measured at steady state, expressed as a percentage of control current and plotted as mean ±S.E.M. (n= 4–14). **P < 0.01 and ***P < 0.005 compared with respective control values.