Cytoplasmic ankyrin repeats of transient receptorpotential A1 (TRPA1) dictate sensitivity to thermaland chemical stimuliJulio F. Cordero-Moralesa,1, Elena O. Grachevaa,1, and David Juliusa,b,2

Departments of aPhysiology and bCellular and Molecular Pharmacology, University of California, San Francisco, CA 94158-2517

Contributed by David Julius, August 30, 2011 (sent for review August 4, 2011)

Transient receptor potential (TRP) channels are polymodal signaldetectors that respond to a wide array of physical and chemicalstimuli, making them important components of sensory systems inboth vertebrate and invertebrate organisms. Mammalian TRPA1channels are activated by chemically reactive irritants, whereassnake and Drosophila TRPA1 orthologs are preferentially activatedby heat. By comparing human and rattlesnake TRPA1 channels,we have identified two portable heat-sensitive modules withinthe ankyrin repeat-rich aminoterminal cytoplasmic domain of thesnake ortholog. Chimeric channel studies further demonstrate thatsensitivity to chemical stimuli and modulation by intracellular cal-cium also localize to the N-terminal ankyrin repeat-rich domain,identifying this region as an integrator of diverse physiologicalsignals that regulate sensory neuron excitability. These findingsprovide a framework for understanding how restricted changesin TRPA1 sequence account for evolution of physiologically diversechannels, also identifying portable modules that specify thermo-sensitivity.

chemosensation | pain | thermosensation | calcium modulation |somatosensation

Primary afferent (somatosensory) neurons detect a range ofphysical and chemical stimuli, including temperature, pres-

sure, and noxious irritants (1). The transient receptor potential(TRP) channel family has been shown to play a predominantrole in these processes, particularly in regard to thermosensi-tivity and chemosensitivity (2–5). TRPA1, otherwise known asthe “wasabi receptor,” plays a key role in somatosensation inevolutionarily diverse phyla, including vertebrate and inver-tebrate species. Mammalian TRPA1 is expressed by primaryafferent sensory neurons of the pain pathway, where it functionsas a sensor of environmental and endogenous chemical irritants,such as allyl isothiocyanate (AITC), acrolein, and 4-hydroxy-nonena, and contributes to cellular mechanisms underlying in-flammatory pain (6–9).TRPA1 channels show species-specific functional variation to

suit their physiological roles. For example, snakes are uniqueamong vertebrates in that their TRPA1 channels are heat-sen-sitive, which some species (rattlesnakes, boas, and pythons) haveexploited to detect infrared radiation (10). Similarly, insectTRPA1 channels are heat-sensitive and contribute to thermalavoidance behaviors (11–14). In these cases, however, thermo-sensitivity comes at the expense of chemosensitivity, such thatAITC and other chemical irritants still activate these channelsbut with reduced potency compared with mammalian orthologs(10). Despite clear physiological differences between snake andmammalian TRPA1 channels, they share significant amino acididentity (56%), providing a unique opportunity to exploit se-quence comparisons and domain swaps to pinpoint structuralelements associated with stimulus detection and/or gating. De-lineating elements that contribute to TRPA1 function will pro-vide insights into the evolutionary process whereby structuralchanges lead to adaptive variations in physiological function.

Here, we ask whether discrete structural motifs specify physicaland chemical stimulus detection.TRPA1 is a nonselective cation channel consisting of four

subunits, each of which contains six transmembrane segmentsflanked by cytoplasmic N and C termini (15, 16). Ankyrin repeats(ARs) are ∼33-residue protein motifs consisting of two α-helicesconnected by a β-turn. ARs occur in tandem arrangements toform elongated domains [AR domains (ARDs)] critical for manyphysiological processes (17, 18). Although many TRP channelfamily members contain in the range of 3–6 ARs within the N-terminal region, TRPA1 is distinguished by having an unusuallylarge number of such repeats (16–17 ARs) (19). Despite theirprevalence among TRP channel subtypes, little is known aboutfunctional roles for ARs. In the case of the capsaicin receptor,TRPV1, another heat-sensitive vertebrate TRP channel (20),crystallographic and physiological studies suggest a role for ARin the binding of cytoplasmic ATP and calmodulin to regulatechannel desensitization (17, 21). Molecular dynamic simulationsof the extended AR of TRPA1 have led to the hypothesis thatthis region of the channel functions as a molecular spring, per-haps tethering the channel to the cytoskeletal matrix as part ofa mechanical gating mechanism (18). Moreover, mutagenesisstudies suggest that cytoplasmic calcium interacts with residueswithin the AR of TRPA1 to regulate channel function; however,the mechanism remains controversial, with some groups favor-ing a direct binding of calcium to an EF hand-like motif (22, 23),and others suggesting an indirect calmodulin-independent mech-anism of an unknown nature (24). Aside from these examples,relatively little is known about the contribution of ARs to TRPchannel function.Nevertheless, the cytoplasmic N terminus is believed to play an

important role in gating because many TRPA1 agonists arestrong electrophiles that activate the channel through reversiblecovalent modification of cysteine and lysine residues within thisdomain (25, 26). Specifically, these highly conserved residues arelocated in the N-terminal region that connects the AR to the firstputative transmembrane segment. How modification of theseresidues promotes channel opening has not been determined,but their location suggests that conformational changes affectingthe N terminus are key to TRPA1 gating mechanisms. In addi-tion to direct activation by chemical irritants, TRPA1 can func-tion as a “receptor-operated” channel that is activated by ATP,bradykinin, or other proalgesic agents that stimulate phospholi-pase C signaling pathways to promote release of calcium from

Author contributions: J.F.C.-M., E.O.G., and D.J. designed research; J.F.C.-M. and E.O.G.performed research; J.F.C.-M., E.O.G., and D.J. analyzed data; and J.F.C.-M., E.O.G., andD.J. wrote the paper.

The authors declare no conflict of interest.1J.F.C.-M. and E.O.G. contributed equally to this work.2To whom correspondence should be addressed. E-mail: david.julius@ucsf.edu.

See Author Summary on page 18595.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1114124108/-/DCSupplemental.

E1184–E1191 | PNAS | November 15, 2011 | vol. 108 | no. 46 www.pnas.org/cgi/doi/10.1073/pnas.1114124108

Dow

nloa

ded

by g

uest

on

Janu

ary

22, 2

021

intracellular stores, with consequent activation of TRPA1 (6, 7).Therefore, understanding calcium-dependent modulation ofTRPA1 is relevant to elucidating mechanisms underlying painhypersensitivity and for designing novel therapeutic agents thattarget this channel.Another major question in this field concerns the structural

basis of thermosensitivity. Various TRP channel regions havebeen proposed to specify thermosensitivity or thermal activationthresholds; these include C-terminal cytoplasmic, outer pore,and turret domains. For example, swapping C-terminal domainsbetween rat TRPV1 and TRPM8 channels has been reported toswitch heat and cold sensitivity (27, 28). Deletions within theTRPV1 C terminus have been shown to alter thermal threshold(29), and substitutions or point mutations within the turret orouter pore domain of TRPV1 have been shown to alter heatsensitivity (30, 31). Recently, the linker region of TRPV1 con-necting the pore-forming transmembrane core to the AR-con-taining N terminus has been suggested to specify heat sensitivity ofthe channel (32). However, a consensus view has yet to emergeregarding the mechanism(s) and region(s) that determine TRPchannel thermosensitivity, and whether such domains are trulymodality-specific regulators of channel function. Further insightsto these issues could be provided by a system in which faithfultransference of thermosensitivity to a temperature-insensitive chan-nel is achieved.The analysis of chimeric channels represents a classic ap-

proach to addressing such questions in cases where pharmaco-logical or physiological properties can be transplanted amongclosely related family members. Unfortunately, TRP channelsfrom different families show relatively weak protein sequenceconservation; thus, most chimeras, including those formedamong closely related homologs, are often nonfunctional and oflimited value. However, sensory receptors are subject to rapidevolutionary adaptation that drives species-specific differences inchannel properties. Thus, chimeras generated among highlyconserved species orthologs more often render channels that arewell-behaved and can be analyzed to identify functionally rele-vant protein domains. Indeed, we have successfully exploited thisapproach to probe structure-function relationships of sensoryTRP channels. For example, the analysis of avian-mammalian oramphibian-mammalian TRPV1 chimeras led to the identifica-tion of amino acids that specify sensitivity to capsaicin or peptidetoxins, respectively (33, 34). A similar approach revealed siteswithin the cold receptor, TRPM8 (35), that determine sensitivityto cooling compounds (36). Although this method has greatlyfacilitated the analysis of chemical sensitivity, it has met withlimited success in delineating regions that account for thermalsensitivity because TRPV1 and TRPM8 from different speciesretain temperature sensitivity as a highly conserved function,save for differences in precise temperature activation threshold(34, 37). Thus, cross-species chimeric channels cannot be used totransfer heat or cold sensitivity cleanly to a thermoinsensitiveortholog. Snake and mammalian TRPA1 channels provide suchan opportunity; thus, we have exploited this functional variationto delineate regions specifying thermal and chemical sensitivity.Here, we show that the aminoterminal cytoplasmic region of

TRPA1 is a portable unit that specifies stimulus detection, in-cluding sensitivity to environmental stimuli (heat and chemicalirritants) as well as cellular messengers (cytoplasmic calcium).Our data support a bimodal arrangement within the N-terminalARs in which thermal and chemical sensitivity are conferred bytwo independent zones, one of which includes determinants forcalcium sensitivity. These findings suggest that the TRPA1 Nterminus functions as a detector and integrator of physiologicalstimuli that regulate channel gating and excitation of somato-sensory nerve fibers.

ResultsThermal Response of Snake TRPA1 Is Specified by a Portable N-Terminal Domain. In initial experiments, we generated chimerasbetween two closely related snake species, rattlesnake (Crotalusatrox) and rat snake (Elaphe obsoleta lindheimeri), whose TRPA1channels share 81% identity but exhibit substantially differentthermal activation thresholds (28 °C and 38 °C, respectively) andsensitivity to electrophilic agonists (low and high, respectively).We found that exchanging the whole aminoterminal cytoplasmicdomain (chimeras 1 and 2, ∼725–726 amino acids) was sufficientto switch both temperature activation threshold and relativechemical sensitivity precisely between these two species orthologs(Fig. 1). This result suggests that the N-terminal cytoplasmicdomain specifies both thermal and chemical response propertiesof the channel. To determine whether this is more generallyapplicable to vertebrate TRPA1 orthologs, we extended this ap-proach to the analysis of chimeric snake-mammalian channels.Human and rattlesnake TRPA1 channels show even greater

functional disparity in that the former is completely insensitive toheat; thus, we asked whether this difference can also be mappedto a discrete domain. Remarkably, a human TRPA1 chimeracontaining a portion of the rattlesnake N terminus (chimera 3, 1–423 amino acids containing the first 10 ARs) gained substantialheat sensitivity (Fig. 2 A and B and Fig. S1A) while retainingrobust AITC sensitivity characteristic of the human channel (seebelow). At the same time, heat-evoked currents produced by thehuman-rattlesnake chimera 3 were attenuated by HC 030031,a selective antagonist that also abolishes responses to chemicalagonists (38), or by the broad-spectrum pore-blocker Ruthenium

A B

C D

E F

Fig. 1. N-terminal domain of snake TRPA1 determines thermal and chem-ical sensitivity. (A) Schematic representation of chimeras between rattle-snake and rat snake TRPA1 channels. (B) Heat-evoked response profiles ofrattlesnake and rat snake chimeras expressed in oocytes. Data representmean ± SD. I/I, the current at each temperature/the current at 45 °C. (C–F)Current-voltage relationships determined by two-electrode voltage clamprecording from oocytes expressing rattlesnake, rat snake, and chimeraschallenged with heat (42 °C, red trace) or AITC (1 mM, orange trace).

Cordero-Morales et al. PNAS | November 15, 2011 | vol. 108 | no. 46 | E1185

PHYS

IOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Janu

ary

22, 2

021

Red, demonstrating that the observed responses were mediatedthrough heterologously expressed chimeric channels (Fig. 2Cand Fig. S1B, respectively). These results further support thehypothesis that the N-terminal cytoplasmic domain specifies heatsensitivity that is both necessary and sufficient to confer ther-mosensitivity to a related, albeit heat-insensitive, ortholog. Fur-thermore, elimination of key electrophile modification sites (3cysteines and 1 lysine) within the N-terminal linker region ofchimera 3 abolished responses to AITC but not to heat (Fig. 2D),suggesting that activation by thermal and chemical stimulirequires different structural elements within this N-terminal re-gion. Taken together, these data demonstrate that thermal andchemical sensitivities are specified by a portable N-terminal AR-rich domains of the channel.Mammalian TRPA1 channels can also be activated by in-

tracellular calcium (8, 22–24), and we therefore asked whetherheat-evoked responses of the human-rattlesnake chimera 3 werestill observed when oocytes were loaded with the calcium che-lator 1,2-bis(o aminophenoxy) ethane-N,N,N′,N′-tetraacetic acid(BAPTA) and perfused with calcium-free medium. Chelation ofcytoplasmic calcium did not eliminate heat-evoked currents (Fig.S1C), consistent with intrinsic sensitivity of the chimera to heatrather than an indirect action via intracellular calcium. Impor-tantly, exchange of equivalent regions between two heat-in-sensitive TRPA1 orthologs (human and zebrafish) (10) did notproduce a heat-sensitive channel (chimera 4, Fig. S1D), showingthat thermal sensitivity is not simply a default state associatedwith a chimeric TRPA1 protein.To identify a minimal portable module that confers thermal

sensitivity, we constructed an extensive series of human-rattle-snake TRPA1 chimeras (Fig. S2) and measured heat-evokedcurrents in Xenopus oocytes expressing these channels. In thisway, we identified two adjacent regions from the rattlesnakechannel (chimera 16, AR3–8 and chimera 29, AR10–15), eachof which was sufficient to confer heat sensitivity to the human

ortholog (Fig. 3 A–C and Fig. S3). Thus, rattlesnake TRPA1contains two independent and portable heat-sensitive regions,each consisting of six ARs. Importantly, transfer of other ARs(e.g., AR1–6, AR4–9, or AR6–11, corresponding to chimeras 8,18, and 19, respectively) did not confer heat sensitivity, demon-strating a requirement for a specific ARs rather than just a stretchof minimum size. Moreover, although chimeras containing AR1–10, AR3–8, or AR10–15 (chimeras 3, 16, and 29, respectively)showed robust heat-evoked currents, their average temperaturecoefficients of activation (Q10s) were different (3.9 ± 0.7, 4.1 ±0.8, and 10 ± 1.6, respectively) (Fig. 3C and Fig. S1A). Thus,although each module is sufficient to confer heat sensitivity,both likely contribute to the overall thermal response propertiesof native rattlesnake TRPA1 channel (Q10 = 13.9) (10).Drosophila TRPA1 (dTRPA1) is also a heat-sensitive channel;

thus, we asked whether transfer of the cognate ARs to humanTRPA1 would similarly confer temperature sensitivity. In gen-eral, human-fly TRPA1 chimeras exhibited high basal activity,likely reflecting the low sequence identity (<30%) between thesetwo orthologs. Nonetheless, we found that substitution of AR10–15 of the human channel with that of the fly ortholog produceda channel (chimera 40) in which robust heat-evoked responseswere clearly observed over baseline current (Fig. 3 A and D).Interestingly, this chimera exhibited reduced AITC sensitivityresembling that of WT dTRPA1. However, transfer of a regioncontaining AR1–8 produced channels (chimera 41) with theconverse phenotype, namely, meager heat sensitivity and robustAITC sensitivity compared with chimera 40 or WT dTRPA1.These results demonstrate portability of a heat-sensitive module(AR10–15) that is functionally conserved between flies and snakes.Interestingly, this region represents the most highly conservedarea (41% identity) when comparing the N-terminal AR-richdomain of fly and snake TRPA1.

AR-Containing Region of the N Terminus Modulates ChemicalSensitivity. Rattlesnake and rat snake channels exhibit markedlydifferent sensitivities to AITC (low and high, respectively). In-deed, swapping their whole N-terminal domains (chimeras 1 and2) switched not only heat but AITC sensitivity (Fig. 1 C–F), fo-cusing our attention on this region of the channel as a majordeterminant of chemical responsiveness. The cysteines and lysinecritical for AITC activation are conserved in all vertebrateTRPA1 channels (10); thus, other as yet unidentified regionsmust account for the difference in chemical sensitivity. Sub-stantial differences in chemical sensitivity are also observed whencomparing human and rattlesnake channels. Thus, whereas hu-man TRPA1 responds to AITC with a half-maximum effectiveconcentration (EC50) of 62 μM, the rattlesnake ortholog is sub-stantially less sensitive, with an EC50 of >2 mM. Additionally,chemical agonists activate the rattlesnake channel to ≤10% ofmaximal heat-evoked responses (Fig. 4 C and G and Fig. S4A),further illustrating its preferential sensitivity to heat.We next used our series of human-rattlesnake chimeras to

dissect this N-terminal locus further. Indeed, we found that re-placement of the first 10 ARs in the human channel with those ofrattlesnake (chimera 3) retained substantial chemical sensitivity(Fig. 4 A, D, and G) resembling that of the parental humanortholog. The inverse chimera 32 (in which the first 10 ARs fromrattlesnake are replaced with the first 10 human ARs) alsoexhibited AITC sensitivity, as reflected by the large rightwardshift of the dose–response profile (Fig. 4G). Moreover, the rateof AITC activation for chimera 32 was substantially slowercompared with human or chimera 3 channels (Fig. S4B). Finally,we examined the importance of the linker segment connectingthe ARD to the transmembrane core, which contains conservedcysteine and lysine modification sites. When this region of therattlesnake channel was replaced with that from the humanchannel (chimera 33), we observed no improvement in AITC

A B

C D

Fig. 2. Portability of the heat-sensitive domain. (A) Schematic representa-tion of a chimera between rattlesnake and human TRPA1 channels. (B) Heat-evoked response profile of human-rattlesnake chimera 3 expressed inoocytes. A single heat-evoked response trace from the heat-insensitive hu-man TRPA1 is shown (gray) for comparison. Data represent mean ± SD. I/I,the current at each temperature/the current at 42 °C. (C) Heat-evoked cur-rents of chimera 3 were inhibited with the selective TRPA1 antagonist HC030031 (20 μM). (D) Chimera 3 lacking electrophile modification sites (3C-A/K-Q) is not sensitive to AITC (1 mM, orange trace) but retains heat (42 °C, redtrace) sensitivity. The 3C-A/K-Q construct corresponds to a quadruple humanTRPA1 mutant in which four sites (C621A, C641A, C665A, and K710Q) es-sential for AITC activation have been altered to nonreactive amino acids.

E1186 | www.pnas.org/cgi/doi/10.1073/pnas.1114124108 Cordero-Morales et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

22, 2

021

sensitivity (Fig. 4 A and F and Fig. S4A). We reached a similarconclusion by examining analogous chimeras between rattle-snake and rat snake channels (Fig. S4C). Taken together, thesefindings suggest that the AR-containing region of the N terminus(distinct from the conserved downstream cysteines or lysine)tunes chemical sensitivity. Furthermore, although chemical sen-sitivity can be conferred by distinct ARs, those adjacent to thelinker region (AR11–16) appear to be most critical in settingAITC sensitivity.

Calcium-Dependent Desensitization Is Specified by AR11 and Adja-cent ARs. Several studies have shown that calcium potentiatesTRPA1 activity, followed by a process of calcium-dependentdesensitization or inactivation (8, 22–24). Despite the potentialrelevance of these phenomena to nociceptor excitability andinflammatory pain (7, 16), mechanisms underlying calcium-de-

pendent TRPA1 modulation remain controversial, with somegroups favoring direct binding of calcium to an EF hand-like motifwithin the N-terminal domain (22, 23) and others advocating acalmodulin- or EF hand-independent process of unknown nature(24). Interestingly, we found that rattlesnake TRPA1 channelsshow calcium-dependent potentiation (52 ± 6%, n = 3; Fig. S5A),as it was shown for human TRPA1 (8), whereas only the humanchannel exhibits robust calcium-dependent desensitization (with atime constant of 9 s; Fig. 5 A–C), consistent with the notion thatcalcium-mediated potentiation and desensitization represent in-dependent modulatory processes (24). These differential proper-ties of human and snake channels enabled us to focus in on aregion required for calcium-dependent desensitization.Specifically, we found that human-rattlesnake chimeras 23 and

38 containing rattlesnake AR7–11 or AR11–14, respectively, did

A

B

C

Temperature (°C)

0

0.2

0.4

0.6

0.8

1

12 15 18 21 24 27 30 33 36 39

Chimera 16 Chimera 29

Temperature (°C)

Q10 = 11.2

Q10 = 1.2

Q10 = 4.7

1,000 / T (K-1)3.20 3.30 3.40 3.50

C

urre

nt (μA

)

10

1

Q10 = 1.7

1,000 / T (K-1)3.10 3.20 3.30 3.40 3.50

10

1

0.1

Cur

rent

(μA

)

I / I

(38°

C)

0

0.2

0.4

0.6

0.8

1

21 24 27 30 36 39 42 4533

I / I

(45°

C)

D

Human

Rattlesnake

ND

13.9

Chimera 16

Chimera 29

(Q10 )

11.2

4.7

Drosophila

Chimera 40

Chimera 41

ND

ND

ND

Chimera 40 Chimera 41 Drosophila

00.2

0.4

0.6

0.8

1

0 40 80 120 160 200

Time (s)

AITC Heat

00.2

0.4

0.6

0.8

1

0 40 80 120 160

Time (s)

AITC Heat

I / I

(max

)

0 40 80 120 160 200

Time (s)

0

0.2

0.4

0.6

0.8

1AITC Heat

Fig. 3. Rattlesnake andDrosophila TRPA1 channels contain portable heat-sensitive domains. (A) Chimeras generated within ARs of human (yellow), rattlesnake(blue), or Drosophila (red) TRPA1 channels. All chimeras shown responded to a saturating concentration (1 mM) of AITC. ND, Q10 not determined. (B) Thermalresponse profiles of chimera 16 and chimera 29 show the presence of two heat-sensitive modules in the rattlesnake TRPA1, each consisting of six specific ARs. I/I(38 °C), the current at each temperature/the current at 38 °C; I/I (45 °C), the current at each temperature/the current at 45 °C. (C) Representative Arrhenius plotsshow thermal thresholds and Q10 values for baseline and evoked responses for chimera 16 (4.7) and chimera 29 (11.2). Data represent mean ± SD. (D) Com-parison of heat- and AITC-evoked response profiles of WT Drosophila TRPA1, chimera 40, and chimera 41 highlights the presence of one heat-sensitive modulein the Drosophila channel corresponding to a region containing AR10–15. I/I (max), the current at each temperature/the current reached at the maximaltemperature. Chimera 40 contains residues 400–612 from Drosophila TRPA1, and chimera 41 contains residues 1–372 (holding potential = −80 mV, n ≥ 6).

Cordero-Morales et al. PNAS | November 15, 2011 | vol. 108 | no. 46 | E1187

PHYS

IOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Janu

ary

22, 2

021

not undergo calcium-dependent desensitization (Fig. 5 D and E).A more restricted chimera (construct 37) containing just rattle-snake AR11 exhibited clear calcium-dependent desensitization,albeit with a reduced rate compared with the WT human channel(τ = 45 ± 12 s vs. 9.0 ± 1.7 s, respectively) (Fig. 5F). Moreover,chimera 23 resembled WT rattlesnake TRPA1 in regard to cal-cium-dependent desensitization (Fig. 5D), consistent with pre-vious mutagenesis data (24) suggesting that a postulated EFhand-like motif located in AR12 is not, by itself, necessary orsufficient to mediate this process. Taken together, these and otherresults (Fig. S5 B–E) suggest that calcium-dependent desen-sitization is specified by an AR cluster centered around AR11.To determine whether calcium binds directly to negatively

charged residues present in AR7–11 and/or AR11–14 of humanTRPA1 (but absent from the rattlesnake channel), we converted

all such aspartate residues to alanine (8 in all), expressed thesemutants in oocytes, and measured their extent of calcium-de-pendent desensitization. We found that these mutations, whenintroduced singly or together, had minimal effect on AITC-evoked activation or desensitization (Fig. S6), ruling out a simplemodel in which calcium binds directly to negative charges dif-ferentially present in this region of the human channel. Irre-spective of the calcium-binding site, our data identify AR7–11and/or AR11–14 of human TRPA1 as an important determinantof calcium regulation.

DiscussionWe show that the N-terminal domain of TRPA1 is a portableunit that specifies stimulus detection, including sensitivity toheat, chemical irritants, and cytoplasmic calcium. Other domains

A

B

D

F G

E

C

Fig. 4. Contribution of the ARs to chemical sensitivity. (A) Diagram depicting chimeras generated between human and rattlesnake TRPA1. (B–F) Repre-sentative current-voltage relationships determined by two-electrode voltage clamp recording from oocytes expressing human and rattlesnake chimerasfollowing exposure to AITC (1 mM, orange trace) or heat (42 °C, red trace). (G) AITC dose–response curves for human, rattlesnake, chimera 3, and chimera 32TRPA1 channels expressed in oocytes (n = 5 for each data point). Points were obtained by measuring response to a 1-min pulse of test dose at +80 mV andnormalized to the maximum response elicited by 6 mM AITC. The dose–response curve was fit to the data. EC50s for human (62 μM), rattlesnake (>2 mM),chimera 3 (36 μM), and chimera 32 (>700 μM) are shown.

E1188 | www.pnas.org/cgi/doi/10.1073/pnas.1114124108 Cordero-Morales et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

22, 2

021

of thermosensitive TRP channels (e.g., C terminus, pore region)have been implicated in stimulus detection or threshold modu-lation (27, 29–31, 39, 40), but our findings with TRPA1 repre-sent, to our knowledge, a previously undescribed functionaldemonstration that temperature sensitivity can be faithfully andprecisely transferred to a temperature-insensitive but otherwisefunctional ortholog. Mammalian TRPA1 has been suggested tofunction as a cold-activated channel (41, 42); thus, one couldargue that we have interchanged properties between two ther-mally sensitive channel orthologs, albeit in opposite directions(cold-sensitive vs. heat-sensitive). However, under our experi-mental conditions, oocytes expressing human TRPA1 showed nocold-evoked currents with cooling down to 10–12 °C, consistentwith recent reports that cold produces slight or no activation oftransfected cells or sensory neurons expressing TRPA1 (43, 44).Aminoterminal ARs represent a structural hallmark of many

members of the extended family of TRP channels, but theirfunctional roles remain largely unknown. Recent studies of thecapsaicin receptor, TRPV1, suggest that the N-terminal ARDconstitutes a binding site for cytoplasmic regulatory factors, suchas ATP and calmodulin (17, 45), consistent with our findings that

the TRPA1 ARD contributes to channel modulation by envi-ronmental and endogenous agents.

TRPA1 Amino Terminus as a Modular Polymodal Signal Detector.Several TRP channels act as polymodal signal detectors (5, 46),but TRPA1 is unique in that different species use this channelfor distinct physiological purposes, generally favoring one mo-dality (thermal or chemical) over the other (10, 12). This mayreflect evolutionary pressure to reduce background noise fromenvironmental irritants in organisms in which the channel is usedprimarily for thermoreception, and vice versa. Organization ofthe N-terminal AR region into functionally redundant and in-dependent modular domains (Fig. 6) may facilitate evolutionarytuning of thermal or chemical sensitivity while retaining a highdegree of sequence similarity across species. For example, an-cient and modern infrared-sensing snakes (boas/pythons and pitvipers, respectively) likely gained heat sensitivity independentlythrough changes to one or both modules of non–infrared-sensingancestors. Furthermore, the bimodal nature of the TRPA1 aminoterminus AR may account for modality competition. Thus, al-though it is possible to construct channels that exhibit both heatand chemical sensitivity, this may come at some expense to the

Human

Rattlesnake9.0

ND

Chimera 23 ND

Ca2+ Desensitizationτ (s)

Chimera 3

A

Human (+Ca2+)

20 s

6

4

2

-2

8

0

Human (-Ca2+)

3

2

1

-1

0

BC

urre

nt (

µA)

Rattlesnake (+Ca2+)

4

2

0

-4

6

-2

10

8

C

Cur

rent

(µA

)

+ 80 mV

- 80 mV

D

E

Chimera 39 Chimera 38 ND

Chimera 37 45

Chimera 38 (+Ca2+)

AITC

Heat

Chimera 37 (+Ca2+)

6

4

2

-2

0

20 s

20 s30 s

20 s

Cur

rent

(µA

)

11.4

31.8

F

12

8

4

-4

01 min

Chimera 23 (+Ca2+)

Cur

rent

(µA

)C

urre

nt (

µA)

Cur

rent

(µA

)

AITC

AITC

AITC AITC

4

2

0

-4

6

-2

10

8

Fig. 5. Calcium-dependent desensitization is specified by AR11 and adjacent ARs. (A) Diagram depicting chimeras generated between human and rattle-snake TRPA1. The right column denotes desensitization to 2 mM Ca2+. ND, no Ca2+-evoked desensitization. (B) Representative traces showing peak AITC-evoked currents with (Left) or without (Right) extracellular Ca2+ (2 mM) from oocytes expressing human TRPA1. (C) Representative trace showing peak heat-evoked current in the presence of extracellular Ca2+ (2 mM) from an oocyte expressing rattlesnake TRPA1, recorded as above. (D–F) Representative traces ofchimera 23 (AR7–11), chimera 38 (AR11–14), and chimera 37 (AR11), respectively. Representative traces of peak AITC or heat-evoked currents were extractedfrom voltage-clamp ramp protocols at ±80 mV. Chimera 37 contains residues 416–457 from rattlesnake, chimera 38 contains residues 416–563, and chimera 39contains residues 460–563.

Cordero-Morales et al. PNAS | November 15, 2011 | vol. 108 | no. 46 | E1189

PHYS

IOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Janu

ary

22, 2

021

detection of one or both stimuli, as suggested by a decreased Q10for human-rattlesnake chimeras (Fig. 3C) or properties of nativerat snake TRPA1, which exhibits low heat and moderate AITCsensitivity (10). Indeed, this inverse relationship between ther-mosensitivity and chemosensitivity also pertains to invertebrate(insect) orthologs (10, 12), suggesting an ancient arrangement inTRPA1 structures that has been maintained over a long evolu-tionary period and in the face of extensive drift in primary aminoacid sequence (Drosophila and human TRPA1 share <30% se-quence identity). Unlike TRPA1, TRPV1 is heat-sensitive in allvertebrate species thus far examined. As such, thermosensationappears to be its core physiological function, whereas chemo-sensitivity plays primarily a modulatory role whereby detection ofendogenous inflammatory agents enhances the channel’s sensi-tivity to heat to produce thermal hyperalgesia (1). Interestingly,and in contrast to TRPA1, these factors target numerous siteson the channel, including the transmembrane core, outer poreregion, and C terminus, illustrating two different strategies fortuning TRP channel stimulus sensitivity.

Structural Framework for Bimodality in Vertebrate TRPA1. Our ex-tensive chimeric analysis indicates that there are two spatiallydistinct AR modules (AR3–8 and AR10–15) that are each ca-pable of conferring heat or robust chemical sensitivity in snakeand mammalian TRPA1, respectively. Biochemical analyses ofcanonical AR proteins provide a molecular framework for un-derstanding how large stretches of AR can exhibit distinct func-tional properties. For example, a 12-AR region from the ankyrinRprotein has been shown to consist of two modules having distinctthermodynamic properties, such that one module exhibits re-duced stability compared with the other (47). This arrangementprovides directionality and cooperativity in the kinetics of proteinunfolding when this AR model peptide is challenged with cha-otropic agents; the aminoterminal ARs unravel first as an in-termediate to subsequent unfolding of more distal C-terminalARs. This is reminiscent of our finding that the primary module(AR10–15) predominates in regard to specifying stimulus sensi-tivity, as illustrated by relative Q10 for heat or EC50 for chemicalirritants (Figs. 3C and 4G). This suggests a model in which twoindependent ARmodules establish a combinatorial code in whichthe one closest to the transmembrane core establishes stimulussensitivity (thermal or chemical), whereas the one closest to the Nterminus serves as an enhancer (Fig. 6). In the case of flies, ourinitial chimeric analysis suggests amore rudimentary arrangementin which AR10–15 represents the only module conferring bothheat and AITC sensitivity. Later in evolution, snakes may haveacquired the enhancer module to augment thermosensitivity.With regard to the actual gating mechanism, thermal or chemicalstimuli could induce conformational changes in the ARD thatdirectly gate the channel, or they could alter interaction withanother cellular factor, consequently triggering channel opening.

Distinguishing between these possibilities will require biophysicalstudies of TRPA1 function in fully defined reconstituted systems.Interestingly, our chimeric analysis also suggests that the pre-

dominant locus of calcium-dependent desensitization (AR11)overlaps substantially with the primary AR module. This is con-sistent with our findings that robust chemical sensitivity and cal-cium-dependent desensitization are functionally linked in thehuman TRPA1 channel, whereas neither heat- nor AITC-evokedresponses show appreciable desensitization in cells expressingsnake channels. Such an arrangement is perhaps ideal for evo-lutionary coadaptation of traits. The capsaicin receptor, TRPV1,contains six ARs within its cytoplasmic N terminus, equivalent insize to a single module as defined by our analysis in TRPA1 (17).The implication is that an AR consisting of six repeats representsa minimal functional unit in heat-sensitive TRP channels. Doesthis also apply to other thermosensitive TRP channel familymembers? A recent study (32) of chimeras between TRPV1 andthe high-threshold heat-sensitive channel TRPV2 concluded thatheat sensitivity is specified by the linker region connecting theARD to the transmembrane core. Moreover, the two ARs (AR5and AR6) most adjacent to this linker region were found to en-hance heat sensitivity. Our observations diverge with this recentstudy, since our data suggest that in the case of TRPA1, the linkerregion does not specify heat sensitivity per se, but rather theARD predominate in this regard, similar to AR5 and 6 of TRV1.Taken together, our findings show that the TRPA1 cytoplas-

mic N terminus serves as a key integration site for multiplephysiological signals that activate or modulate the channel. Ourfunction-based studies identify regions and potential domainsthat specify sensitivity to thermal and chemical stimuli, but adeep mechanistic understanding of how these modules promoteTRPA1 gating will require further biophysical and structuralanalysis of the N-terminal domain.

Materials and MethodsMutagenesis and Chimera Constructs. Point mutations and chimeras of human,rattlesnake, rat snake, and zebrafish TRPA1 were generated using standardoverlapping PCR or site-directed mutagenesis-based methods (catalog no.250519; Stratagene). In all cases, conserved amino acid sequences were usedfor chimeric junctions. For oocyte expression, genes were subcloned into thecombined mammalian/oocyte expression vector pMO. All constructs wereverified by DNA sequencing.

Oocyte Electrophysiology. cRNA transcripts were synthesized and injected intoXenopus laevis oocytes using Ambion mMessage machine kits according tothe manufacturer’s protocol. Surgically extracted oocytes (Nasco) were cul-tured and analyzed 1–3 d postinjection using a two-electrode voltage clampas previously described by Jordt et al. (8). Briefly, oocytes were injected with0.05–5 ng per cell of RNA and whole-cell currents were measured after24–72 h using a Geneclamp 500 amplifier (Axon Instruments, Inc.). Micro-electrodes were pulled from borosilicate glass capillary tubes to obtainresistances of 0.5 MΩ. Bath solution contained 10 mM Hepes, 120 mM NaCl,2 mM KCl, 0.2 mM EGTA, 2 mM MgCl2, and 1 mM CaCl2 to a final pH of 7.4.For experiments in the absence of Ca2+, we used the same bath solutioncontaining 5 mM EGTA. For BAPTA/acetoxymethyl ester (AM) experiments,currents were induced by increasing the temperature of bath solution aftera 3-h incubation of oocytes with 10–50 μM BAPTA/AM (Sigma). BAPTA/AMpermeates into the cytoplasm and, after hydrolysis, chelates Ca2+ ions. Datawere analyzed using pCLAMP10 software (Molecular Devices).

Determination of Thermal Threshold. Response at each temperature wasnormalized to the response at the maximal applied temperature (holdingpotential = −80 mV, n ≥ 8). All data are shown as mean ± SD. Temperaturethresholds represent the point of intersection between linear fits to baselineand the steepest component of the Arrhenius profile. Values are derivedfrom averages of individual curves (n ≥ 10). Arrhenius curves were obtainedby plotting the current on a log-scale against the reciprocal of the absolutetemperature. Q10 was used to characterize temperature dependence of ioniccurrents as calculated by the following equation:

Cysteines

CN

2 3 4

9 10 11 12 13 14 15 16

1 5 6 7 8

Primary module

Enhancer module

Fig. 6. Schematic representation of the TRPA1 N-terminal region. ARs in redand light red specify heat sensitivity for the primary and enhancer modules,respectively; stars denote location of cysteine residues required for activa-tion by electrophiles.

E1190 | www.pnas.org/cgi/doi/10.1073/pnas.1114124108 Cordero-Morales et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

22, 2

021

Q10 ¼�R2

R1

�10=ðT2 − T1Þ

where R2 is the current at the higher temperature, T2, and R1 is the current atthe lower temperature, T1 (48).

Determination of Time Constant. Desensitization time constants werefitted to a single exponential function from peaks of AITC-evoked currents

as calculated using Origin 7.1 (OriginLab) with the following equation: y =A1·exp(−x/τ1) + ψ0.

ACKNOWLEDGMENTS. We thank A. Priel for advice and assistance withelectrophysiology and A. Chesler, C. Williams, and other members of thelaboratory for insightful discussions. This work was supported by a Pathwayto Independence Fellowship from the University of California, San FranciscoCardiovascular Research Institute (to E.O.G.) and by grants from the NationalInstitutes of Health (to D.J.).

1. Basbaum AI, Bautista DM, Scherrer G, Julius D (2009) Cellular and molecularmechanisms of pain. Cell 139:267–284.

2. Caterina MJ, Julius D (2001) The vanilloid receptor: A molecular gateway to the painpathway. Annu Rev Neurosci 24:487–517.

3. Jordt SE, McKemy DD, Julius D (2003) Lessons from peppers and peppermint: Themolecular logic of thermosensation. Curr Opin Neurobiol 13:487–492.

4. Peier AM, et al. (2002) A TRP channel that senses cold stimuli and menthol. Cell 108:705–715.

5. Ramsey IS, Delling M, Clapham DE (2006) An introduction to TRP channels. Annu RevPhysiol 68:619–647.

6. Bandell M, et al. (2004) Noxious cold ion channel TRPA1 is activated by pungentcompounds and bradykinin. Neuron 41:849–857.

7. Bautista DM, et al. (2006) TRPA1 mediates the inflammatory actions of environmentalirritants and proalgesic agents. Cell 124:1269–1282.

8. Jordt SE, et al. (2004) Mustard oils and cannabinoids excite sensory nerve fibresthrough the TRP channel ANKTM1. Nature 427:260–265.

9. Trevisani M, et al. (2007) 4-Hydroxynonenal, an endogenous aldehyde, causes painand neurogenic inflammation through activation of the irritant receptor TRPA1. ProcNatl Acad Sci USA 104:13519–13524.

10. Gracheva EO, et al. (2010) Molecular basis of infrared detection by snakes. Nature464:1006–1011.

11. Hamada FN, et al. (2008) An internal thermal sensor controlling temperaturepreference in Drosophila. Nature 454:217–220.

12. Kang K, et al. (2010) Analysis of Drosophila TRPA1 reveals an ancient origin forhuman chemical nociception. Nature 464:597–600.

13. Kohno K, Sokabe T, Tominaga M, Kadowaki T (2010) Honey bee thermal/chemicalsensor, AmHsTRPA, reveals neofunctionalization and loss of transient receptorpotential channel genes. J Neurosci 30:12219–12229.

14. Rosenzweig M, et al. (2005) The Drosophila ortholog of vertebrate TRPA1 regulatesthermotaxis. Genes Dev 19:419–424.

15. Latorre R, Zaelzer C, Brauchi S (2009) Structure-functional intimacies of transientreceptor potential channels. Q Rev Biophys 42:201–246.

16. Nilius B, Prenen J, Owsianik G (2011) Irritating channels: the case of TRPA1. J Physiol589:1543–1549.

17. Lishko PV, Procko E, Jin X, Phelps CB, Gaudet R (2007) The ankyrin repeats of TRPV1bind multiple ligands and modulate channel sensitivity. Neuron 54:905–918.

18. Sotomayor M, Corey DP, Schulten K (2005) In search of the hair-cell gating springelastic properties of ankyrin and cadherin repeats. Structure 13:669–682.

19. Gaudet R (2008) A primer on ankyrin repeat function in TRP channels and beyond.Mol Biosyst 4:372–379.

20. Caterina MJ, et al. (1997) The capsaicin receptor: A heat-activated ion channel in thepain pathway. Nature 389:816–824.

21. Phelps CB, Wang RR, Choo SS, Gaudet R (2010) Differential regulation of TRPV1,TRPV3, and TRPV4 sensitivity through a conserved binding site on the ankyrin repeatdomain. J Biol Chem 285:731–740.

22. Doerner JF, Gisselmann G, Hatt H, Wetzel CH (2007) Transient receptor potentialchannel A1 is directly gated by calcium ions. J Biol Chem 282:13180–13189.

23. Zurborg S, Yurgionas B, Jira JA, Caspani O, Heppenstall PA (2007) Direct activation ofthe ion channel TRPA1 by Ca2+. Nat Neurosci 10:277–279.

24. Wang YY, Chang RB, Waters HN, McKemy DD, Liman ER (2008) The nociceptor ionchannel TRPA1 is potentiated and inactivated by permeating calcium ions. J BiolChem 283:32691–32703.

25. Hinman A, Chuang HH, Bautista DM, Julius D (2006) TRP channel activation byreversible covalent modification. Proc Natl Acad Sci USA 103:19564–19568.

26. Macpherson LJ, et al. (2007) Noxious compounds activate TRPA1 ion channels throughcovalent modification of cysteines. Nature 445:541–545.

27. Brauchi S, Orta G, Salazar M, Rosenmann E, Latorre R (2006) A hot-sensing coldreceptor: C-terminal domain determines thermosensation in transient receptorpotential channels. J Neurosci 26:4835–4840.

28. Latorre R, Brauchi S, Orta G, Zaelzer C, Vargas G (2007) ThermoTRP channels asmodular proteins with allosteric gating. Cell Calcium 42:427–438.

29. Prescott ED, Julius D (2003) A modular PIP2 binding site as a determinant of capsaicinreceptor sensitivity. Science 300:1284–1288.

30. Grandl J, et al. (2010) Temperature-induced opening of TRPV1 ion channel is

stabilized by the pore domain. Nat Neurosci 13:708–714.31. Yang F, Cui Y, Wang K, Zheng J (2010) Thermosensitive TRP channel pore turret is part

of the temperature activation pathway. Proc Natl Acad Sci USA 107:7083–7088.32. Yao J, Liu B, Qin F (2011) Modular thermal sensors in temperature-gated transient

receptor potential (TRP) channels. Proc Natl Acad Sci USA 108:11109–11114.33. Bohlen CJ, et al. (2010) A bivalent tarantula toxin activates the capsaicin receptor,

TRPV1, by targeting the outer pore domain. Cell 141:834–845.34. Jordt SE, Julius D (2002) Molecular basis for species-specific sensitivity to “hot” chili

peppers. Cell 108:421–430.35. McKemy DD, Neuhausser WM, Julius D (2002) Identification of a cold receptor reveals

a general role for TRP channels in thermosensation. Nature 416:52–58.36. Chuang HH, Neuhausser WM, Julius D (2004) The super-cooling agent icilin reveals

a mechanism of coincidence detection by a temperature-sensitive TRP channel.Neuron 43:859–869.

37. Myers BR, Sigal YM, Julius D (2009) Evolution of thermal response properties in a cold-activated TRP channel. PLoS ONE 4:e5741.

38. McNamara CR, et al. (2007) TRPA1 mediates formalin-induced pain. Proc Natl Acad SciUSA 104:13525–13530.

39. Grandl J, et al. (2008) Pore region of TRPV3 ion channel is specifically required forheat activation. Nat Neurosci 11:1007–1013.

40. Jordt SE, Tominaga M, Julius D (2000) Acid potentiation of the capsaicin receptor

determined by a key extracellular site. Proc Natl Acad Sci USA 97:8134–8139.41. Karashima Y, et al. (2009) TRPA1 acts as a cold sensor in vitro and in vivo. Proc Natl

Acad Sci USA 106:1273–1278.42. Story GM, et al. (2003) ANKTM1, a TRP-like channel expressed in nociceptive neurons,

is activated by cold temperatures. Cell 112:819–829.43. del Camino D, et al. (2010) TRPA1 contributes to cold hypersensitivity. J Neurosci 30:

15165–15174.44. Knowlton WM, Bifolck-Fisher A, Bautista DM, McKemy DD (2010) TRPM8, but not

TRPA1, is required for neural and behavioral responses to acute noxious coldtemperatures and cold-mimetics in vivo. Pain 150:340–350.

45. Phelps CB, Procko E, Lishko PV, Wang RR, Gaudet R (2007) Insights into the roles ofconserved and divergent residues in the ankyrin repeats of TRPV ion channels.

Channels (Austin) 1:148–151.46. Tominaga M, et al. (1998) The cloned capsaicin receptor integrates multiple pain-

producing stimuli. Neuron 21:531–543.47. Werbeck ND, Rowling PJ, Chellamuthu VR, Itzhaki LS (2008) Shifting transition states

in the unfolding of a large ankyrin repeat protein. Proc Natl Acad Sci USA 105:

9982–9987.48. DeCoursey TE, Cherny VV (1998) Temperature dependence of voltage-gated H+

currents in human neutrophils, rat alveolar epithelial cells, and mammalian

phagocytes. J Gen Physiol 112:503–522.

Cordero-Morales et al. PNAS | November 15, 2011 | vol. 108 | no. 46 | E1191

PHYS

IOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Janu

ary

22, 2

021