Portal de Periódicos da CAPES

Pitfalls in auxin pharmacology

2020; Wiley; Volume: 227; Issue: 2 Linguagem: Inglês

10.1111/nph.16491

1469-8137

Julian Dindas, Dirk Becker, M. Rob G. Roelfsema, Sönke Scherzer, Malcolm J. Bennett, Rainer Hedrich,

Plant Reproductive Biology

Auxin represents a key plant hormone that functions to shape plant development and growth (Leyser, 2018). The polar transport (PAT) of indole-3-acetic acid (IAA), the major form of auxin in higher plants, is central to the physiological output of this plant hormone. IAA is a weak acid (pKa = 4.58) and according to a prevailing acidic apoplastic pH (pHext ~ 5.5), the equilibrium for de-protonated (IAA−) and protonated (IAAH) molecules in the cell wall can be calculated to be c. 83% and 17%, respectively (Zazimalova et al., 2010). Even before the molecular identification of auxin transport proteins, it was generally accepted that transport of IAA− was carrier mediated, while IAAH would be able to cross membranes by simple diffusion, due to its lipophilic nature (Gutknecht & Walter, 1980; Delbarre et al., 1996). Once in the cytoplasm, at neutral pH (~ 7–7.5), IAA becomes de-protonated and almost entirely exists in its membrane-impermeable (anionic, IAA−) form. Consistently, the chemiosmotic polar diffusion model predicted that auxin transport is efflux carrier-mediated and PAT requires their asymmetric localization (Rubery & Sheldrake, 1974; Raven, 1975; Goldsmith, 1977). Today, more than 40 years later, it is well established that, at the molecular level PAT is accomplished through the concerted action of auxin influx and efflux carriers together with their dynamic asymmetric localization (Swarup & Bhosale, 2019; Zwiewka et al., 2019). While auxin influx carriers belong to the family of H+-driven AUXIN1/LIKE-AUXIN1 (AUX/LAX) transporters (Swarup & Bhosale, 2019), auxin efflux is accomplished by PIN-FORMED (PIN) and members of the P-GLYCOPROTEIN/ATP-BINDING CASSETTE B4 (PGP/ABCB) family (Geisler et al., 2017; Sauer & Kleine-Vehn, 2019). Local differences in auxin concentration, established via PAT, are translated into physiological responses via transcriptional changes of auxin responsive genes (Weijers & Wagner, 2016). Nuclear auxin perception by F-box proteins of the TRANSPORT INHIBITOR RESPONSE1/AUXIN SIGNALING F-BOX (TIR1/AFB) family governs the expression of auxin responsive genes through the degradation of Aux/IAA class transcriptional repressors (Leyser, 2018). Recently, novel electrophysiological and imaging approaches have been applied to uncover the earliest cellular responses to auxin. Recordings of auxin-induced changes in the apoplastic pH, cytosolic Ca2+ level and the plasma membrane potential of Arabidopsis root cells revealed a novel plasma membrane-delimited auxin signalling pathway that is composed of the major auxin uptake facilitator AUX1, the Ca2+ channel CNGC14 and auxin-receptor F-box proteins (Shih et al., 2015; Barbez et al., 2017; Dindas et al., 2018), which exists in parallel to the well-established nuclear auxin signalling machinery (Kubes & Napier, 2019). The role of auxin receptors and transporters has not only been studied employing available mutants, but pharmacological approaches have also taken advantage of synthetic IAA analogues, such as 1-naphthaleneacetic acid (1-NAA), and PAT inhibitors like triiodobenzoic acid (TIBA), 1-naphthylphthalamic acid (NPA) or more recently with pinstatic acid, as well as with the response inhibitor auxinole (Teale & Palme, 2018; Oochi et al., 2019). Lately, the palette of modifiers of auxin responses was enlarged with an engineered auxin-TIR1 pair, in which convex IAA (cvxIAA) can act through a specifically designed concave TIR1 (cvvTIR1) (Uchida et al., 2018; Yamada et al., 2018). Although commonly used to study auxin physiology, our knowledge about the molecular transport routes of synthetic auxin analogues and their modes of action has remained rudimentary. In one of the few studies that has addressed auxin pharmacology, Yang et al. (2006) expressed the AUX1 transporter in Xenopus oocytes and demonstrated that 1-NAA was unable to compete with 3H-IAA uptake. The authors concluded that, unlike IAA, 1-NAA was not a substrate for AUX1 and suggested that it probably entered oocytes based on its lipophilic nature. The authors further observed that TIBA was unable to block the auxin influx activity of AUX1 in oocytes and therefore concluded that TIBA represents a specific inhibitor for auxin efflux (Geldner et al., 2001). Using microelectrode-based techniques, with Arabidopsis thaliana bulging root hairs as an in vivo model system, we present experiments which provide evidence that (1) 1-NAA uptake depends on a proton-coupled transporter independent of AUX1, whilst (2) TIBA strongly disrupts IAA and 1-NAA uptake, in addition to its accepted role as an auxin efflux inhibitor. Seeds of Arabidopsis thaliana were sterilized for 5 min in 6% NaOCl + 0.05% Triton-X 100, followed by several washing steps with deionized water. Single seeds were placed on the surface of small Petri-dishes (Ø 35 mm), filled with 1 ml of growth medium (0.12% Murashige & Skoog basal salt mixture incl. MES, Duchefa; 0.5% sucrose; 1% agarose, pH 5.8 with Tris). The Petri-dishes were placed vertically in a growth chamber (KBWF 720, Binder, Tuttlingen, Germany, or KKD Hiross, Clitec, Küssnacht, Switzerland) with a day : night cycle of 12 h : 12 h, 21°C and 16°C and a photon flux density of 120 µmol photons m−2 s−1. Seedlings were grown for 3 to 5 d. Sterile grown seedlings were submerged in bath solution (0.1 mM KCl, 1 mM CaCl2, 5 mM MES/BTP pH 5.5), at least 20 min before the measurements. Membrane potentials of bulging root hairs were recorded using single-barrelled microelectrodes that were fabricated from borosilicate glass capillaries (outer diameter 1 mm, wall thickness 0.21 mm, Hilgenberg GmbH, Malsfeld, Germany). The electrodes were pulled either on a horizontal laser puller (P2000; Sutter Instruments, Novato, CA, USA), or a filament puller (DMZ Universal Puller; Zeitz Instruments, Martinsried, Germany). Back pressure-operated application pipettes were made from pulled capillaries, of which the tip was broken back to a diameter of 20 to 40 µm. These pipettes were filled with bath solution containing the indicated concentration of auxin and a back pressure of 2 bar was applied for 1 s, with a Picospritzer II microinjection system (General Valve, Cleveland, OH, USA). Alternatively, the measuring chamber was perfused with auxin containing bath solutions. The bath solution was connected to the ground with a reference electrode using a glass capillary that was backfilled with 300 mM KCl and sealed with an agarose plug (2% agarose in 300 mM KCl). Microelectrodes were backfilled with 300 mM KCl and connected to either a HS180 head stage connected to a VF-102 amplifier (Bio-Logic, Seyssinet-Pariset, France) or an EPC10 USB patch-clamp head stage and amplifier (Heka, Ludwigshafen, Germany). Data were sampled at 0.1 kHz, either using the Pulse software (version 8.74; Heka) with an LIH-1600 interface (Heka), or the WinWCP software (University of Strathclyde, UK) with an NI USB 6259 interface (National Instruments, Austin, TX, USA). In experiments with the EPC10 amplifier, the Patch Master Software (v.2x90.3; Heka) was used in combination with an LIH 8 + 8 interface (Heka). The acquired data was analysed off-line using Excel (Microsoft Corp., Redmond, WA, USA) and Origin Pro 9 software (OriginLab, Northampton, MA, USA). Ion fluxes at the surface of root epidermal cells were monitored using ion-selective scanning microelectrodes. The electrodes were fabricated from borosilicate glass capillaries without filament (outer diameter 1.0 mm, Science Products GmbH, Hofheim, Germany) on a vertical puller (Narishige Scientific Instrument Lab, Tokyo, Japan). The capillaries were baked overnight at 220°C and subsequently silanized with N,N-dimethyltrimethylsilylamine (Sigma-Aldrich, Hamburg, Germany) for 1 h. H+-selective electrodes were backfilled with 40 mM KH2PO4/15 mM NaCl and tip filled with the hydrogen ionophore I cocktail A (Sigma-Aldrich). Electrodes were connected via Ag/AgCl half-cells to head stages of a microelectrode amplifier (IPA-2; Applicable Electronics, New Haven, CT, USA). The ion-selective electrodes were scanning at 10 s intervals over a travelling distance of 100 µm perpendicular to the root surface, using a micro-stepping motor driver (US Digital, Vancouver, WA, USA). The electrode was placed at c. 5 µm distance from the root surface at the early differentiation zone, in between root hairs, using a SM-17 Micromanipulator (Narishige Scientific Instrument Lab) and an upright microscope (Axioskop; Carl Zeiss AG, Oberkochen, Germany). Before the experiments, the seedlings were accustomed to the bath solution (100 μM KCl, 100 μM CaCl2, 100 μM MES/BTP pH 5.5) for at least 20 min. Root cells were stimulated with IAA or 1-NAA, at a final concentration of 10 μM. Raw data were acquired with a NI USB 6259 interface (National Instruments), controlled by custom-made, Labview-based, software ‘Ion Flux Monitor’. Raw voltage data were converted offline into ion flux values, using the Labview-based program Ion Flux Analyser, Excel and Origin Pro 9 software. Live cell imaging was performed on a Zeiss Axiokop 2FS microscope, with an Achroplan 40×/0.80w objective (Zeiss). The microscope was equipped with a CARV2 confocal imaging unit (Crest Optics, Rome, Italy). Image acquisition was carried out using a charge multiplying CCD camera (QuantEM 512SC; Photometrics, Tucson, AZ, USA), controlled by Visiview software (Visitron, Puchheim, Germany). For excitation, light from from a mercury lamp (LQ HXP 120; Leistungselektronik, Jena, Germany) was guided through a bandpass filter (562/40 nm; Semrock, Rochester, NY, USA) and reflected onto the sample by a 590 nm dichroic mirror (Zeiss). The emission signal was filtered with a bandpass filter at 628 nm (628/40 nm; Semrock). The analysis of imaging data was carried out with the ImageJ software (imagej.nih.gov/ij/). The ability of auxin to diffuse across the plasma membrane was studied with oocytes of the African clawfrog Xenopus laevis. The Julius-von-Sachs Institute (Würzburg, Germany) has permission for keeping Xenopus frogs, which is registered at the government of Lower Franconia (ref. no. 55.2-2532-2-1035). Stage V and VI oocytes, obtained from Xenopus laevis frogs, were treated with 0.14 mg ml−1 collagenase I in Ca2+-free ND96 buffer (10 mM HEPES pH 7.4, 96 mM NaCl, 2 mM KCl, 1 mM MgCl2) for 90 min. Oocytes were subsequently washed with Ca2+-free ND96 buffer and stored at 16°C in ND96 solution (10 mM HEPES pH 7.4, 96 mM NaCl, 2 mM KCl, 1 mM MgCl2, 1 mM CaCl2) containing 50 mg l−1 gentamycin. For transient expression of a plasma membrane targeted version of pHluorin (Shen et al., 2013), 25 ng of cRNA was injected into selected oocytes, which were incubated for 2 d at 16°C in ND96 solution with gentamycin. Two-electrode voltage-clamp experiments were conducted with pHluorin expressing oocytes that were perfused with control solution (220 mM Sorbitol, 1 mM KCl, 1 mM MgCl2, 1 mM CaCl2, 1 mM LaCl3, 10 mM MES/Tris pH 5.5), which was supplemented with 10 µM auxins (IAA or 1-NAA), or 10 mM sodium acetate (NaOAC) as indicated. The oocyte membrane was clamped to a holding potential of −100 mV or a test potential of −160 mV. Alternatively, the oocyte was measured in the current-clamp mode, using a current that provoked a membrane potential of c. −120 mV. Time-resolved measurements of cytosolic pH changes were performed with oocytes expressing pHluorin, which were excited at 405 and 470 nm and the emission signal filtered at 510–550 nm was collected with a CCD camera. The fluorescence ratio of raw signals obtained with 470/405 nm excitation light, was used as a measure for changes in in cytosolic pH. To test significances two-tailed Student's t-tests were applied. All data points marked as significantly different from wild-type or control experiments were so with P < 0.05. The synthetic auxin analogue 1-NAA represents a universally used substitute for the natural IAA in experimental auxin physiology. 1-NAA is capable of rescuing the gravitropic phenotype of aux1 auxin influx mutants (Marchant et al., 1999). In accordance with its acid pKa (~ 4.24) and the chemiosmotic hypothesis, this observation suggested that 1-NAA represents a membrane permeable molecule. Recently we published the first in vivo characterization of AUX1 (Dindas et al., 2018). Therein, we report that loss of the AUX1 transporter does not affect the depolarization of root hairs induced by 1-NAA. By contrast, the IAA dependent membrane potential changes were virtually abolished in aux1 knock-out plants. This finding prompted us to test the characteristics of 1-NAA uptake in further detail. We found that application of 1-NAA to bulging root hairs triggered a concentration-dependent depolarization of the root hair plasma membrane potential (EC50 = 0.9 ± 0.5 µM). In line with our previous results, no significant difference was found with respect to the 1-NAA-induced depolarization between wild-type and the aux1 loss-of-function mutant wav5-33 (EC50 = 1.1 ± 0.9 µM; Fig. 1a and inset). Because loss of AUX1 in the wav5-33 mutant does not affect the concentration-dependence of the 1-NAA-dependent depolarization, we conclude that 1-NAA uptake is most likely mediated by a transport mechanism other than AUX1. A shift from pHext 5.5 to pHext 7.5 impaired the 1-NAA induced depolarization (Fig. 1b) suggesting that similar to IAA, the uptake of 1-NAA is co-transported with H+. This is further supported by the observation that 1-NAA application resulted in a reversal of root cell H+-fluxes from an outward to an inward direction, as revealed with H+-selective scanning microelectrodes (Fig. 1c). To exclude the option that it is the lipophilic nature of 1-NAA that is producing the depolarization or the observed changes in proton fluxes, we employed the established Xenopus oocyte system – previously used to study IAA/NAA transport (Yang et al., 2006; Ranocha et al., 2013; Fastner et al., 2017; Skokan et al., 2019) and lacking endogenous high-affinity auxin carriers. Using experimental conditions comparable to in vivo measurements (a membrane potential of −160 mV and pHext 5.5), neither application of 10 µM IAA nor 1-NAA produced detectable currents across the oocyte plasma membrane (Fig. 2a,b). By contrast, upon superfusion with 10 mM NaOAc, frequently used to produce cytosolic acidification (Roos & Boron, 1981; Lacombe et al., 2000), pronounced inward currents were recorded. Simultaneous pHcyt recordings corroborated these findings: application of either IAA or 1-NAA did not affect the near-plasma-membrane cytosolic pH, while NaOAc in the bath medium resulted in instantaneous cytosolic acidification. Likewise, neither IAA nor 1-NAA generated membrane potential depolarizations, as revealed by membrane potential recordings in Xenopus oocytes (Fig. 2c). Together, these results thus suggest that in root hairs, 1-NAA is taken up via a H+-coupled transporter different from AUX1. Given that IAA and NAA transport are dependent on the proton motive force (pmf), we next tested the effects of auxin transport inhibitors NPA and TIBA on auxin-induced fast root hair plasma membrane depolarization. We observed that root hair cells pre-treated for 20 min with a 20 µM concentration of the auxin efflux inhibitor NPA, responded to 10 µM auxin application with the typical rapid membrane potential change (Fig. 3a). In contrast to NPA, pre-treatment with 20 µM TIBA severely shifted the root hair resting plasma membrane potential to a more positive value of −109 mV (n = 5, SE = 3). In addition, TIBA strongly suppressed IAA-induced fast depolarization (Fig. 3a). The fact that TIBA affects the resting potential of the root hair in the absence of IAA suggests that this inhibitor impairs plasma membrane energization by the H+-ATPase. To test this assumption, we analysed TIBA induced changes of the root hair pH profile. In line with the activity of an active H+-ATPase, experiments with scanning H+-selective microelectrodes revealed that the zone of bulging root hairs is characterized by a net proton efflux of c. 100 nmol m−2 s−1 (Fig. 3b). Upon TIBA application however, proton efflux collapsed and finally reversed into an average H+ influx of c. 10 nmol m−2 s−1 (Fig. 3b). Under control conditions and in the absence of TIBA, 1-NAA and IAA induced inward directed H+ fluxes (see Figs 1c and 3c, black bars). In line with the hypothesis that uptake of IAA and 1-NAA depends on the H+-gradient across the plasma membrane, TIBA strongly inhibits H+ influx induced by both auxins (Fig. 3c, red bars) and thus carrier mediated uptake. Fast auxin signalling involves elevation of cytosolic Ca2+ levels through an AUX1- and SCFTIR1/AFB-dependent activation of the plasma membrane localized Ca2+ channel CNGC14 (Shih et al., 2015; Dindas et al., 2018; Retzer et al., 2018). Because of the reduced auxin uptake capacity in TIBA-treated bulging root hair cells, we considered it likely that TIBA also affected auxin-induced Ca2+ signalling. To test this hypothesis, we combined membrane potential recordings with R-GECO1 based Ca2+-live cell imaging, we found that TIBA not only suppressed IAA transport associated depolarization, but in addition, suppressed IAA-induced Ca2+ signals in root hair cells (Fig. 3d,e). Taken together the results of membrane potential and H+-flux measurements strongly suggest that the impact of this classical auxin inhibitor on IAA transport and signalling is indirect. Recent publications have provided increasing evidence, that apart from the well-known canonical auxin signalling pathway controlling the transcriptional responses of auxin responsive genes, SCFTIR1/AFB auxin receptors also control nontranscriptional aspects required for fast auxin signalling (Dindas et al., 2018; Fendrych et al., 2018; Kubes & Napier, 2019). Fast auxin signalling depends on a nonnuclear fraction of SCFTIR1/AFB auxin receptors as well as the activity of the IAA uptake carrier AUX1 (Yang et al., 2006; Dindas et al., 2018). Mutants impaired in AUX1 function or targeting confer selective resistance to various auxins (Dharmasiri et al., 2006). The observation that their agravitropic phenotype can be rescued by 1-NAA, but not IAA or 2,4-dichlorophenoxyacetic acid (2,4-D) (Marchant et al., 1999; Yamamoto & Yamamoto, 1999; Hoyerova et al., 2018) is consistent with the chemiosmotic theory predicting that diffusion of the membrane-permeable auxin parallels carrier mediated transport (Rubery & Sheldrake, 1974). Using recently established electrophysiology measurements on bulging root hairs (Dindas et al., 2018), we show that like IAA, 1-NAA transport is coupled to membrane depolarization mediated by proton influx and 1-NAA symport. 1-NAA influx appears to involve a secondary active transporter that is energized by the plasma membrane H+-ATPase generated proton gradient. Since the transport kinetics of 1-NAA are unchanged in the AUX1 loss-of-function mutant wav5-33, the 1-NAA carrier is likely to be different in molecular nature from AUX1. Our results concur with recent pharmacophore studies which also show that certain auxin herbicides are not transported by AUX1 (Hoyerova et al., 2018). Our results further suggest that 1-NAA uptake is not mediated by simple plasma membrane diffusion. This hypothesis is supported by the observation that neither IAA nor 1-NAA evoke current, membrane potential depolarization or cytosolic acidification in Xenopus oocytes, which lack bona fide auxin carriers (Yang et al., 2006). On the basis of tracer flux experiments with tobacco cells, previous studies have calculated a permeability coefficient for the nondissociated form of 1-NAA (NAAH) of 12 × 10−6 m s−1 (Delbarre et al., 1994, 1996). At our experimental conditions and a pKa = 4.24 for 1-NAA, the NAAH concentration is 0.55 µM and thus the flux of NAAH across a membrane is c. 6.6 nmol m−2 s−1 (see the Materials and Methods section). The measured uptake rate for H+ was c. 100 nmol m−2 s−1 (see Fig. 1c), which is 15-times higher than the expected diffusion of 1-NAA across the membrane of a plant cell. Given that the prevalent anionic form of 1-NAA (1-NAA−) represents the substrate for this carrier, like IAA the stoichiometry of carrier mediated auxin transport is at least 2 H+ for 1 IAA−/1-NAA−. Hence, the transport of 1-NAA by a carrier will exceed diffusion by a factor of approximately eight. Even though it is likely that IAA− and 1-NAA− are taken up together with two H+, the stoichiometry of AUX1 and the unidentified 1-NAA transporter still need to be established. In the past, transport studies with radioactive auxin sources have shed light on the uptake rates of plant cells (Rubery & Sheldrake, 1974; Raven, 1975; Rubery & Skoog, 1980; Gimmler, 1981; Delbarre et al., 1994, 1996). Moreover, this approach has been used to study the uptake or efflux activity of various transporters that were expressed in Xenopus leavis oocytes (Yang et al., 2006; Zourelidou et al., 2014). While the use of radioactive substrates for transport can provide valuable information, the resolution in time and space is often limited. The recently developed fluorescent IAA biosensor, AuxSen (Herud-Sikimic et al., 2020), is likely to open up new possibilities to simultaneously measure changes in the cytoplasmic IAA concentration and H+-uptake rates in intact plant roots. We therefore expect that this sensor will help to unravel the link between auxin-induced H+ fluxes and auxin uptake in detail. While the molecular nature of the proposed 1-NAA carrier is currently unknown, other members of the Arabidopsis AUX1/LAX family or H+-coupled amino acid transporters represent possible candidates for catalysing 1-NAA transport. However, none of the other family members are expressed in root hair cells (Peret et al., 2012). Alternatively, Arabidopsis NRT1.1, a member of the NRT1/PTR FAMILY (NPF) mediates nitrate as well as auxin uptake (Krouk et al., 2010). We propose that 1-NAA is taken up into roots by a plasma membrane transporter that could be identified by screening aux1 knock-out mutants for mutations that prevents complementation of the agravitropic phenotype by 1-NAA. Our attempts to probe the molecular mechanism of auxin uptake by pharmacological means, revealed that TIBA, a well-known inhibitor of polar auxin transport, must be treated with caution. We observed that in the presence of TIBA the membrane potential generated by plasma membrane H+-pumps collapses. By destroying the plasma membrane proton-motive force, TIBA will severely impede AUX1-mediated IAA transport, as well as H+-coupled 1-NAA uptake as well as downstream signalling. The molecular mechanism of TIBA action is scant, but in line with our observations, Geldner et al. (2001) demonstrated, that TIBA not only impaired trafficking of the auxin efflux carrier PIN1 but also that of PM H+-ATPases. The growing use of pharmacological tools in plant research, including auxin analogues or synthetic substrates for engineered, orthogonal auxin-TIR1 receptor pairs (Uchida et al., 2018), correspondingly requires a detailed understanding of their transport routes, to avoid off-target effects as well as interpretation about their molecular modes of action. This work was supported by the German Research Foundation (DFG, FOR964 to RH and DB). All authors declare no competing financial or other interests. JD, RH, DB, MB and MRGR designed the study and wrote the manuscript. JD, DB, SS and MRGR performed the experiments. JD, SS and MRGR analysed the data.