Identification of Significant Amino Acids in Multiple Transmembrane Domains of Human Transient Receptor Potential Ankyrin 1 (TRPA1) for Activation by Eudesmol, an Oxygenized Sesquiterpene in Hop Essential Oil

Background: Transient receptor potential ankyrin 1 (TRPA1) is activated by many spicy compounds by unknown mechanisms. Results: The amino acids critical for the activation of TRPA1 by the hop-derived sesquiterpene β-eudesmol were identified. Conclusion: Multiple transmembrane domains are crucial for β-eudesmol-derived TRPA1 activation. Significance: New insight for activation mechanism of TRPA1 is revealed by our study. Transient receptor potential ankyrin 1 (TRPA1) is a calcium-permeable non-selective cation channel that is activated by various noxious or irritant substances in nature, including spicy compounds. Many TRPA1 chemical activators have been reported; however, only limited information is available regarding the amino acid residues that contribute to the activation by non-electrophilic activators, whereas activation mechanisms by electrophilic ligands have been well characterized. We used intracellular Ca2+ measurements and whole-cell patch clamp recordings to show that eudesmol, an oxygenated sesquiterpene present at high concentrations in the essential oil of hop cultivar Hallertau Hersbrucker, could activate human TRPA1. Gradual activation of inward currents with outward rectification by eudesmol was observed in human embryonic kidney-derived 293 cells expressing human TRPA1. This activation was completely blocked by a TRPA1-specific inhibitor, HC03–0031. We identified three critical amino acid residues in human TRPA1 in putative transmembrane domains 3, 4, and 5, namely threonine at 813, tyrosine at 840, and serine at 873, for activation by β-eudesmol in a systematic mutational study. Our results revealed a new TRPA1 activator in hop essential oil and provide a novel insight into mechanisms of human TRPA1 activation by non-electrophilic chemicals.


Transient receptor potential ankyrin 1 (TRPA1) is a calciumpermeable non-selective cation channel that is activated by various noxious or irritant substances in nature, including spicy compounds. Many TRPA1 chemical activators have been reported;
however, only limited information is available regarding the amino acid residues that contribute to the activation by non-electrophilic activators, whereas activation mechanisms by electrophilic ligands have been well characterized. We used intracellular Ca 2؉ measurements and whole-cell patch clamp recordings to show that eudesmol, an oxygenated sesquiterpene present at high concentrations in the essential oil of hop cultivar Hallertau Hersbrucker, could activate human TRPA1. Gradual activation of inward currents with outward rectification by eudesmol was observed in human embryonic kidney-derived 293 cells expressing human TRPA1. This activation was completely blocked by a TRPA1-specific inhibitor, HC03-0031. We identified three critical amino acid residues in human TRPA1 in putative transmembrane domains 3, 4, and 5, namely threonine at 813, tyrosine at 840, and serine at 873, for activation by ␤-eudesmol in a systematic mutational study. Our results revealed a new TRPA1 activator in hop essential oil and provide a novel insight into mechanisms of human TRPA1 activation by non-electrophilic chemicals.
The transient receptor potential (TRP) 2 channel family is known to be involved in detection of various noxious stimuli, such as thermal, chemical, and mechanical stimulus (1). Among them, transient receptor potential ankyrin 1 (TRPA1) channel is a nonselective cation channel mainly expressed in primary sensory neurons (2). It responds to a wide variety of irritant sensory stimuli, including pungent ingredients of various spices and herbal medicines (3), which indicates its involvement in pungent sensations. Many TRPA1 activators have a reactive electrophilic group, allowing covalent binding to cysteine residues in the cytosolic N terminus (4,5). In addition, non-electrophilic TRPA1 activators also exist in nature (6 -8). The activation mechanisms of these compounds are not fully understood, although important amino acids for activation of TRPA1 by non-electrophilic activators have been reported (9,10).
Plants have defense mechanisms such as chemical irritant accumulation to avoid predation and infection. In addition, plant-produced chemical compounds also contribute to human life as natural medicines, fragrances, seasonings, or spices. Essential oils whose major components are volatile terpenoids are used worldwide as flavorings; they have large structural variations, and each compound possesses a characteristic flavor. Hop cones, the immature inflorescences of the female plant of Humulus lupulus L., are widely used in beer production to add flavor and bitterness. Essential oils in hop cones contain various volatile terpenoids that are important components, determining beer flavor. In the production of new flavor patterns in beer, many hop cultivars with differing volatile terpenoid patterns have been developed.
Previous research has focused on the olfactory sensation of hop essential oil (11)(12)(13); however, little is known about their effects on sensory receptors, yet sensory stimulation should be an important factor in beer development. A particular hop cultivar, Hallertau Hersbrucker (HHE), is described as adding a "spicy" character to beer (14). A previous report showed that the oxygenated sesquiterpenoid fraction from hops contributed to its spicy character (15), but mechanisms underlying this have not been well studied. Previous reports have implied that characteristic ingredients of HHE essential oil can modulate sensory receptors such as TRP channels (14,15).
In this report we provide evidence that the human TRPA1 (hTRPA1) channel is activated by eudesmol, a characteristic oxygenated-sesquiterpenoid of HHE. There are three eudesmol structural isomers, namely ␣-eudesmol, ␤-eudesmol, and ␥-eudesmol (Fig. 1A, inset), which do not have a reactive electrophilic group, suggesting they should be classified into a non-electrophilic type activator of TRPA1. Site-directed mutagenesis identified three critical amino acids in transmembrane domains 3, 4, and 5 of hTRPA1 required for activation by ␤-eudesmol. Our results provide further understanding of the activation mechanism of human TRPA1 by non-electrophilic activators.
GC-MS Analysis-Hops were ground to a fine powder using a pestle and mortar, and dichloromethane (50 mg dry weight/ml for 1 h) was used in the extraction procedure. Borneol was used as an internal standard for quantification. Extracts were dehydrated using Na 2 SO 4 , and then 1 l was used for GC-MS. GC-MS was performed using a QP2010 mass spectrometer (Shimadzu, Kyoto, Japan) coupled with a GC-2010 gas chromatograph (Shimadzu) in split mode (1:30). The ion source was operated at 70 eV. The gas chromatograph was equipped with a HP-INNOWAX capillary column (60 m ϫ 0.25 mm, 0.25-m film thickness; Agilent Technologies, CA). The oven temperature was programmed from 40°C (0.18 min hold) to 240°C at a rate of 3°C/min using helium as the carrier gas at 2.07 ml/min. Quantitative data are shown as the means of two independent experiments.
Plasmid Vector-All human TRP channel expression vectors were purchased from OriGene Technologies. Nucleotide substitution was performed using the QuikChange site-directed mutagenesis kit (Agilent Technologies) with slight modification. Oligonucleotides used for nucleotide substitution are listed in Table 1. The coding regions were fully sequenced before they were used.
Calcium Imaging-Human embryonic kidney-derived 293 (HEK293) cells were maintained in Dulbecco's modified Eagle's medium (supplemented with 10% fetal bovine serum, penicillin, streptomycin, and L-glutamine) and transfected with expression vector using Lipofectamine TM LTX (Invitrogen). Cells were then incubated for 24 h in phenol red-free DMEM medium (supplemented with 1% fetal bovine serum) before use in calcium-imaging experiments. Fluo-4 Direct TM Calcium Assay kits (Invitrogen) were used according to the manufacturer's instructions. Compound addition and fluorescence measurement was performed using the FDSS Cell system (Hamamatsu Photonics, Hamamatsu, Japan). The DMSO solution of eudesmol and the buffer used for dilution were kept at 40°C before the experiments to maximize solubility. Dilutions were carried out just before the experiments, and the diluted solution was used immediately in the experiments to avoid deposition of eudesmol. This procedure was also adopted for other cell-based assays.
Whole-cell Patch Clamp Experiment-HEK293 cells were transfected with human TRPA1 or human TRPV3 and 0.1 g of pGreen Lantern-1 using the same method as in the calciumimaging experiment and incubated with Dulbecco's modified Eagle's medium (supplemented with 10% fetal bovine serum, penicillin, streptomycin, and L-glutamine) for 14 -24 h. The bath solutions for the patch clamp experiments contained 140 mM NaCl, 5 mM KCl, 2 mM MgCl 2 , 2 mM CaCl 2 , 10 mM glucose, 10 mM HEPES, pH 7.4 (with NaOH). Calcium-free bath solutions for patch clamp experiments contained 140 mM NaCl, 5 mM KCl, 2 mM MgCl 2 , 10 mM glucose, 5 mM EGTA, 10 mM HEPES, pH 7.4 (with NaOH). The pipette solution contained 140 mM KCl, 5 mM EGTA, and 10 mM HEPES, pH7.4 (with KOH). Data from whole-cell voltage-clamp recordings were sampled at 10 kHz and filtered at 5 kHz for analysis (Axon 200B amplifier with pCLAMP software; Axon Instruments). The cell was voltage-clamped at Ϫ60 mV. The current-voltage relation-

Name
Forward Reverse ship was obtained using 500 ms voltage-ramp pulses from Ϫ100 to ϩ100 mV applied every 5 s. All experiments were performed at room temperature except for eudesmol solution preparation described above. Statistical Analysis-The Mann-Whitney U test, Kruskal-Wallis test followed by the Steel test, one-way ANOVA followed by the Bonferroni test, or one-way ANOVA followed by Dunnett's test were used for statistical evaluation. Significance was assumed if the p value was Ͻ0.05.

Eudesmol Is a Characteristic Oxygenated Sesquiterpenoid Found in the Hop Cones of Cultivar Hallertau Hersbrucker-
The characteristic spicy taste from hop cones of cultivar HHE is reported to be due to oxygenated sesquiterpenoid (15). Quantitative GC-MS data showed that HHE accumulated the highest amount of eudesmol among 44 hop cultivars; 7 other cultivars also contained eudesmol but at a lower level (Fig. 1A). Eudesmol has been reported to be one of the representative oxygenated sesquiterpenoids in hop (16), and there are three structural isomers whose structural formulas are shown in the insets of Fig. 1A. Representative total ion chromatograms and mass spectrums of HHE extract and eudesmol standards are shown in Fig. 1, B and C. ␣-Eudes-mol and ␤-eudesmol accumulated in abundance in HHE in comparison with ␥-eudesmol ( Fig. 1, A and B). Other volatiles, such as sesquiterpene, ␤-caryophyllene, and ␣-humulene, accumulated in almost all the cultivars tested (Fig. 2). These results indicated that eudesmol is one of the characteristic compounds that accumulates in HHE.
Activation of Human TRPA1 and TRPV3 by Eudesmol-To investigate whether eudesmol could activate human sensory channels, we investigated six representative hTRP channels. HEK293 cells transiently expressing each hTRP channel were utilized to investigate changes in intracellular Ca 2ϩ concentrations ([Ca 2ϩ ] i ), which were detected by fluorescence of Fluo-4. hTRPA1 was the most activated channel after eudesmol treatment among the hTRP channels tested. Although apparent activation of hTRPV3 was observed, distinct activation of other hTRP channels, hTRPV1, hTRPV2, hTRPV4, and hTRPM8, were not observed under our experimental conditions (Fig. 3A).
Representative traces of calcium imaging of hTRPA1 are shown in Fig Electrophysiological experiments were carried out to confirm hTRPA1 activation by eudesmol. Representative wholecell patch clamp current traces are shown in Fig. 4A. Each eudesmol isomer (100 M) gradually brought inward currents with an outwardly rectifying current-voltage relationship, a characteristic property of the TRPA1 channel. Densities of the currents activated by ␣-eudesmol, ␤-eudesmol, and ␥-eudesmol at a holding potential of Ϫ60 mV were calculated as 62.4 Ϯ 19.6, 35.2 Ϯ 9.01, and 34.1 Ϯ 17.5 pA/pF, respectively. In Fig. 4A we utilized a high concentration of AITC (100 M) as a positive control. Currents activated by AITC (100 M) after eudesmol treatment were relatively small because of desensitization. The current-voltage (I-V) relationship differed between eudesmol-and AITC-activated currents in that eudesmol-activated currents exhibited clear outward rectification, whereas currents activated by a high concentration of AITC exhibited a more linear I-V relationship. It is well known that TRPA1-mediated currents show different I-V-relationships depending on the channel activation (Fig. 4B) (17). The data also indicate that eudesmol is a relatively weak TRPA1 agonist. Inward currents evoked by eudesmol showed significant differences in densities between hTRPA1-expressing cells and vector control cells (Fig.  4C). No significant differences were observed in the densities of currents activated by isomers of eudesmol in hTRPA1-expressing cells, suggesting the similar abilities of the isomers of eudesmol to activate hTRPA1.
We utilized ␤-eudesmol as the representative in subsequent experiments because almost the same results were obtained for the structural isomers and because ␣-eudesmol and ␥-eudesmol were not commercially available. To verify that a response by ␤-eudesmol was elicited via hTRPA1 channel activation, the effect of an inhibitor was examined. HC03-0031, a TRPA1specific antagonist, completely suppressed both inward and outward currents induced by ␤-eudesmol (Fig. 6, A and B), supporting our findings that ␤-eudesmol activates hTRPA1. It is  well known that Ca 2ϩ ions entering the cells through TRPA1 enhance TRPA1 channel activity; therefore, we performed the experiments in the absence of extracellular Ca 2ϩ . Currents activated by ␤-eudesmol in the absence of extracellular Ca 2ϩ were very small, whereas much bigger currents were elicited by ␤-eudesmol and AITC in the presence of 2 mM extracellular Ca 2ϩ (Fig. 6C). Inward currents showed significant differences in densities between the absence and presence of 2 mM Ca 2ϩ (Fig. 6D). The result indicated an extracellular Ca 2ϩ dependence of the ␤-eudesmol-evoked current response, further supporting the activation of TRPA1 by ␤-eudesmol. In addition, ␤-eudesmol did not affect the inward currents evoked by the maximum dose of AITC (Fig. 6, E-G). These data revealed that hTRPA1 is activated by ␤-eudesmol. Three Amino Acids in Human TRPA1 Transmembrane Domains 3, 4, and 5 Are Important for Activation by ␤-Eudesmol-Next, we attempted to identify the amino acid residues involved in ␤-eudesmol-induced activation of hTRPA1. ␤-Eudesmol is thought to be a non-electrophilic activator of hTRPA1 because it does not have the ␣,␤-unsaturated carbonyl structure characteristic of electrophilic activators like AITC. Previous studies of ligand-gated TRP channels suggested that transmembrane domains contain important amino acids for hydrophobic activators, and these amino acids tend to possess hydroxyl groups in  their side chain that may contribute to hydroxyl bonding with activator molecules (9,18). Because ␤-eudesmol is a hydrophobic compound and has a hydroxyl group, we focused on amino acids possessing hydroxyl groups in the transmembrane domains of hTRPA1. Based on the prediction of transmembrane domains using the SOSUI program, we prepared 21 hTRPA1 mutants in which serine, threonine, or tyrosine in putative transmembrane domains were substituted with alanine. Four mutant channels (Y726A, T734A, S781A, and T869A) were eliminated from subsequent analyses because no activation by AITC was observed. In addition, these four mutant channels did not respond to carvacrol, a non-electrophilic TRPA1 activator. These four mutant channels may have lost channel function or may not have been expressed in the plasma membrane. Fig. 7A shows a schematic diagram of the putative structure of hTRPA1 in which the remaining 17 amino acid positions are indicated by circles. The first screening results used a calcium imaging method in which ␤-eudesmol-evoked responses were normalized to AITC-evoked responses to exclude any unexpected effects for protein expression levels or ion channel functions. Seventeen mutant channels were analyzed, and five mutants (T813A, Y840A, Y849A, S873A, and T874A) showed significant reductions in response to ␤-eudesmol (Fig. 7B). To confirm the importance of these five amino acids, electrophysiological experiments were carried out. All five mutant channels produced significantly smaller whole-cell currents compared with WT (Fig.  7C, upper). To investigate the significance of these mutants in terms of responses to ␤-eudesmol, we normalized the currents by ␤-eudesmol to those by AITC because these five mutants also exhibited significantly small AITC-evoked currents (Fig. 7C, middle). T813A, Y840A, and S873A showed significant reductions in the current ratios, whereas Y849A and T874A did not (Fig. 7C,  lower). These data suggested that Thr-813, Tyr-840, and Ser-873 could be more important amino acids for ␤-eudesmol-evoked hTRPA1 activation. Representative traces of these three mutants are shown in Fig. 7D. These mutant channels almost completely lost the response to ␤-eudesmol, whereas responses to AITC were observed. The data indicated that these three amino acids play a critical role in hTRPA1 activation by ␤-eudesmol.

DISCUSSION
Empirical sensory assessments have evaluated that HHE produces a spicy character in beer (14); our results of hTRPA1 activation by eudesmol could help explain these sensory phenomena. HHE-flavored beer contains eudesmol at about 1 M, which is lower than the effective concentrations activating hTRPA1 revealed in this study. This discrepancy could be explained by potential synergistic effects of other chemicals in beer. There are reports that combinations of TRPA1 activators at lower concentrations could evoke a large synergistic activation (19). Because many TRPA1 activators are found in natural compounds, beer may include other hTRPA1 activators to produce synergistic TRPA1 activation with eudesmol in vivo.
Various stimulations have been found to evoke TRPA1 activation, and in many cases these stimulations have been viewed as noxious. Meanwhile, TRPA1 activators have been found in food ingredients such as horseradish, garlic, cinnamon, and  5-6). The Mann-Whitney U test was adopted. One-way ANOVA followed by the Bonferroni test was used for the statistical evaluation among isomers. Significance was assumed if the p value was Ͻ0.05. *, p Ͻ 0.05.
pepper. These ingredients could produce the desired sensory stimulations and health benefits under appropriate conditions (3). Eudesmol appears to be a weak activator of hTRPA1, indicative that this compound is unlikely to be a useful pharmacological tool; however, identification of a TRPA1 activator in hop cones provides the possibility of favorable control of the sensory stimulation of beer. ␤-Eudesmol activated hTRPA1 under extracellular calciumfree conditions, as shown in Fig. 6C, although activation was reduced compared with 2 mM Ca 2ϩ extracellular conditions. This result together with the fact that gradual current activation was observed upon ␤-eudesmol application (Fig. 4A) suggests that calcium ions entering the cells through TRPA1 activated by ␤-eudesmol enhance the channel activity. Previous studies revealed that intracellular calcium ions directly activate TRPA1 via cytosolic N-terminal EF-hand calcium binding domains (20).
We also found eudesmol produced weak but significant activation of hTRPV3, a warm-sensitive Ca 2ϩ -permeable cation channel. Carvacrol, a major component of essential oil from oregano, also activates both TRPA1 and TRPV3 (21). A previous report showed that carvacrol produced both a pungent and warm sensation on the tongue (21), whereas previous sensory assessments of HHE have not reported a warm sensation. The data that eudesmol-evoked hTRPV3-mediated currents at a holding potential of Ϫ60 mV were negligible, similar to the current in the vector control cells, could imply that eudesmolevoked hTRPV3 activation may be too weak to contribute to the occurrence of a warm sensation. Structural determinants for TRPA1 activation have been studied in other non-electrophilic TRPA1 activators. Oleocanthal, another non-electrophilic TRPA1 activator found in extra virgin olive oil, needs a particular double bond in the molecule for TRPA1 activation (6). Three eudesmol structural isomers possess double bonds in different positions, whereas the ability of these isomers to activate hTRPA1 did not differ. These data indicate that the position of the double bond in eudesmol is not critical for TRPA1 activation and suggests other structure determinants for TRPA1 activation in the eudesmol molecule.
To explore the critical amino acids for hTRPA1 activation by ␤-eudesmol, we screened for amino acids with hydroxyl groups in their side chains within the transmembrane domains because this has been reported in other TRP channels. For example, tyrosine and serine in transmembrane domains 3 and 4 are known to contribute to capsaicin binding in TRPV1 (22). A menthol binding site was identified as tyrosine in transmembrane domain 3 of TRPM8 (18). In addition, a menthol binding site in TRPA1 was reported as β-Eudesmol threonine in transmembrane domain 5 (9). Recently, a tyrosine residue in hTRPA1 transmembrane domain 3 was shown to be essential for inhibition by borneol, a monoterpene alcohol (23). Systematic screening identified three critical amino acids in transmembrane domains 3, 4, and 5 that were required for hTRPA1 activation by ␤-eudesmol. Xiao et al. (9) presented a model of hydrogen bonding between the hydroxyl groups of threonine and menthol in transmembrane domain 5 of TRPA1. Our results may also suggest that hydrogen bonding occurs between ␤-eudesmol and the identified amino acids of hTRPA1, and this might contribute to channel activation. The published model shows a sole menthol binding site in TRPA1, whereas the three amino acids in different transmembrane domains, shown in the current study, may contribute to TRPA1 activation in different ways. To reveal the roles of the identified amino acids in the ␤-eudesmol-evoked activation of hTRPA1, experiments examining the shift of EC 50 values and differences in maximal potentiation would provide valuable information (24). It has been reported that multiple tyrosine residues play distinct roles in ligand-activated channels, such as the GABAc receptor and 5-hydroxytryptamine receptor (25,26). Ligand binding sites in TRP channels were studied using sitedirected mutagenesis, chimeric analysis between different TRP channels, and structural data (27). The three-dimensional structure of TRPA1 was revealed by electron microscopy (28) in which TRPA1 was shown to form tetramer. Multiple cytosolic AITC binding sites in the N terminus were shown to locate at overlapped regions of each monomer, and ligand binding was thought to be affected by either monomer or tetramer interaction to gate the channel open. Our results demonstrated that important identified amino acids were distributed in multiple transmembrane domains, suggesting that multiple binding sites of ␤-eudesmol exist in hTRPA1 as shown by covalent modification by electrophilic activators such as AITC. It is also conceivable that a single well organized binding site consists of multiple transmembrane ␣-helices. Either way, hydroxylated amino acids, threonine, serine, and tyrosine are known to break the ␣-helical symmetry (29). This property may lead to speculation that transmembrane domains 3, 4, and 5 form hinges at identified hydroxylated amino acids. In the case of TRPA1 activation by menthol, hydroxylated amino acids serine and threonine in transmembrane domain 5 are shown to be important, and it is suggested that these amino acids may also act as the hinge (9). Two hydroxylated terpenes, menthol and ␤-eudesmol, might share the TRPA1 activation mechanism to elicit conformational changes through hydroxylated amino acids in transmembrane domains, although the distribution patterns of important amino acids differed. Different regions have been reported to be crucial for activation among non-electrophilic TRPA1 activators. A recent report revealed that amino acids critical for TRPA1 activation by a non-electrophilic TRPA1 activator, methyl anthranilate, were not in the transmembrane domains but in the N-terminal cytosolic region (30). Methyl anthranilate may affect either monomer or tetramer structures by changing the N-terminal environments to gate the channel open without covalent modification. This difference in important regions for activation among non-electrophilic activators implies that different activation mechanisms may exist among non-electrophilic TRPA1 activators. However, it is important to note that our results do not indicate the direct binding of ␤-eudesmol to hTRPA1. It is possible that indirect effects of ␤-eudesmol occur through the modification of the lipid bilayer. Thus, either direct interactions or indirect effects may be detected by these hydroxylated amino acid residues, and this could contribute to conformational changes leading to the channel opening. Our study revealed important amino acids required for activation of hTRPA1 by ␤-eudesmol located in multiple transmembrane domains, as observed in other ligandgated TRP channels (31,32). However, we cannot exclude the involvement of other amino acids. More intensive studies are required for accurate understanding of the TRPA1 activation mechanisms by non-electrophilic activators and the roles of the amino acids identified in this study.