The mode of action of auxin is based on its non-uniform distribution within tissues and organs. Despite the wide use of several auxin analogues in research and agriculture, little is known about the specificity of different auxin-related transport and signalling processes towards these compounds.
Using seedlings of Arabidopsis thaliana and suspension-cultured cells of Nicotiana tabacum (BY-2), the physiological activity of several auxin analogues was investigated, together with their capacity to induce auxin-dependent gene expression, to inhibit endocytosis and to be transported across the plasma membrane.
This study shows that the specificity criteria for different auxin-related processes vary widely. Notably, the special behaviour of some synthetic auxin analogues suggests that they might be useful tools in investigations of the molecular mechanism of auxin action. Thus, due to their differential stimulatory effects on DR5 expression, indole-3-propionic (IPA) and 2,4,5-trichlorophenoxy acetic (2,4,5-T) acids can serve in studies of TRANSPORT INHIBITOR RESPONSE 1/AUXIN SIGNALLING F-BOX (TIR1/AFB)-mediated auxin signalling, and 5-fluoroindole-3-acetic acid (5-F-IAA) can help to discriminate between transcriptional and non-transcriptional pathways of auxin signalling.
The results demonstrate that the major determinants for the auxin-like physiological potential of a particular compound are very complex and involve its chemical and metabolic stability, its ability to distribute in tissues in a polar manner and its activity towards auxin signalling machinery.
The auxin class of plant growth regulatory compounds controls different growth and developmental events in plants. Charles and Francis Darwin first noted the existence of an endogenous substance with biological activity in plants (Darwin & Darwin, 1980). This substance was characterized and reported as indole-3-acetic acid (IAA) (Kögl et al., 1934; Went & Thimann, 1937) and it has been proved to be the major endogenous auxin present in plants. Other compounds with auxin characteristics, such as indole-3-butyric acid (IBA) (Zimmerman & Wilcoxon, 1935), phenylacetic acid (PAA) (Koepfli et al., 1938) and 4-chloroindole-3-acetic acid (4-Cl-IAA) (Porter & Thimann, 1965), were later identified as endogenous auxins. Chemically more stable synthetic auxins, such as naphthalene-1-acetic acid (NAA) and 2,4-dichlorophenoxyacetic acid (2,4-D), are used as auxins in various biological applications, and some of these auxins and their structural analogues, such as, for example, polychloroaromatic acids or chlorinated picolinic acids, are efficient herbicides (Walsh et al., 2006; Grossmann, 2010; Müller & Appleby, 2010). A number of other substances, mostly various carboxy derivatives of benzene, phenol, naphthalene or indole, show varying degrees of auxin-like activity in different bioassays (Koepfli et al., 1938; Porter & Thimann, 1965; Ferro et al., 2006; Christian et al., 2008).
The coordinated growth and development of plants require the maintenance of optimum concentration gradients of active auxin(s) in specific organs, tissues and/or cell types, where they often act in co-operation with other plant hormones (Depuydt & Hardtke, 2011; Leyser, 2011; Rosquete et al., 2012). The regulated development and maintenance of appropriate auxin concentration gradients in tissues are achieved by local metabolic processes, including biosynthesis, conjugation/deconjugation and degradation (Ljung et al., 2002; Chandler, 2009; Ljung, 2013), and/or by polar transport and intracellular auxin compartmentalization (Grunewald & Friml, 2010). These processes result in the establishment of optimum auxin concentration gradients in specific organs, tissues, cells and their compartments, and, in turn, transduce an auxin signal into appropriate biochemical and physiological responses.
Among plant hormones, active polar cell-to-cell transport is a unique property of auxin. The specificity of transport routes for frequently used auxins has been characterized in many plant systems, including maize coleoptiles (Hertel et al., 1969) and roots (Martin & Pilet, 1986), zucchini hypocotyl segments (Depta & Rubery, 1984), Arabidopsis inflorescence stalks (Parry et al., 2001), Arabidopsis root, hypocotyls and inflorescence (Rashotte et al., 2003), suspension-cultured crown gall cells (Rubery, 1977), soybean root cells (Loper & Spanswick, 1991), suspension-cultured tobacco cells of Nicotiana tabacum L. cv Xanthi XHFD8 (Delbarre et al., 1996), cv Bright Yellow-2 (BY-2) (Petrášek et al., 2006) and cv Virginia Bright Italia (VBI-0) (Petrášek et al., 2002), and zucchini and lupin microsomal vesicles (Sabater & Rubery, 1987). At the molecular level, the movement of auxin across the plasma membrane (PM) is mediated by the AUXIN-RESISTANT1/LIKE AUX1 (AUX1/LAX) auxin influx carriers (Bennett et al., 1996; Swarup et al., 2008) and PIN-FORMED (PIN) auxin efflux carriers (Křeček et al., 2009). Moreover, plant orthologues of the mammalian ATP-binding cassette subfamily B (ABCB)-type transporters also transport auxin (Petrášek & Friml, 2009; Yang & Murphy, 2009). The affinity of components of the auxin transport machinery for major auxins and their specific interactions differ in different cell types (Delbarre et al., 1996; Paciorek et al., 2005), and a detailed study of auxin transport specificity is still lacking.
It has been reported that auxin itself can inhibit endocytosis of some PM-localized proteins, including PINs (Paciorek et al., 2005). Recently, a putative auxin receptor, the AUXIN BINDING PROTEIN 1 (ABP1), has been shown to act as a positive regulator of clathrin-mediated endocytosis of PINs and the binding of auxin restricts this effect (Robert et al., 2010). Even though the narrow specificity of PM-associated ABP1 towards auxin-like compounds was suggested decades ago (Hertel et al., 1972; Löbler & Klämbt, 1985; Zažímalová & Kutáček, 1985), details of the specificity of the auxin-dependent inhibition of endocytosis are not yet known.
In addition to ABP1, there is also a well-characterized TRANSPORT INHIBITOR RESPONSE 1 (TIR1)-dependent auxin signalling pathway (Dharmasiri et al., 2005; Kepinski & Leyser, 2005). It is based on the degradation of Aux/IAA transcriptional repressors, mediated by the SCFTIR1 ubiquitin ligase complex in contact with auxin. The involvement of auxin is concentration dependent and auxin acts as a ‘molecular glue’, enhancing the interaction between TIR1 and Aux/IAA (Tan et al., 2007). Recently, the joint action of TIR1/AFBs and Aux/IAAs has been reported for differential auxin sensing, suggesting that the TIR1/AFBs and Aux/IAAs function as auxin co-receptors (Calderon Villalobos et al., 2012). On the basis of the structure of the promoter part of auxin-responsive genes, the ‘synthetic’ auxin-responsive promoter DR5 was constructed (Ulmasov et al., 1997) and, in fusion with reporter genes, it is used widely to visualize the level of response to auxin.
It is obvious that the final mode of auxin action will depend on many partial auxin-related actions; however, to date there has been no comprehensive study of the auxin specificity of these processes. To address this deficiency and to obtain an insight into the complex auxin-related regulatory mechanism(s) in plant development, we have selected a group of compounds (‘auxin analogues’) that are structurally related to the three major auxins, that is, the native IAA and the synthetic NAA and 2,4-D. In order to determine their auxin-like effectiveness and to investigate the selectivity of processes along the auxin response chain from the whole plant to the organ, tissue and cellular level, we analysed the capacity of these auxin analogues to be transported and to induce DR5-driven gene expression. We also investigated their ability to inhibit endocytosis of PM-related PIN proteins. These assays, together with the characterization of the auxin-like physiological potential of selected auxin analogues, showed the similarity of the structural requirements for auxin-like substances for all these processes; however, they also revealed the special behaviour of some compounds.
Materials and Methods
Auxins and auxin-like compounds used
IAA, indole-3-propionic acid (IPA), IBA, PAA, NAA, naphthalene-2-acetic acid (2-NAA), 2,4-D, 2,4,5-T, 2-(2,4-dichlorophenoxy) propionic acid (2,4-DP) and indole-3-lactic acid (ILA) were all obtained from Sigma-Aldrich Inc. (St Louis, MO, USA). 5-Fluoroindole-3-acetic acid (5-F-IAA), 4-Cl-IAA, 5-chloroindole-3-acetic acid (5-Cl-IAA), 5-bromoindole-3-acetic acid (5-Br-IAA) and indole-3-acetyl alanine (IAAla) were purchased from OlChemIm (Olomouc, Czech Republic). Benzoic acid (BA), indole and naphthalene were obtained from Fluka. 6-Chloroindole-3-acetic acid (6-Cl-IAA) was a generous gift from Professor Kjeld Engvild (Riso National Laboratory, Roskilde, Denmark). 5-Cl-IAA stock solution (50 mM) was prepared in dimethylsulfoxide (DMSO) and all other compounds were dissolved in ethanol. [3H]NAA and [3H]2,4-D (both with a molar (specific) radioactivity of 20 Ci mmol−1) and [3H]IBA (molar (specific) radioactivity of 25 Ci mmol−1) were purchased from American Radiolabeled Chemicals Inc. (St Louis, MO, USA).
Plant materials and growth conditions
Wild-type Arabidopsis thaliana Heynh. seeds and tir1-1 mutant (Ruegger et al., 1998) (both ecotype Columbia) were surface sterilized with 70% (v/v) ethanol and 0.05% (v/v) Triton X-100 for 10 min, shaking occasionally. The liquid from the first wash was replaced with the same volume of 100% ethanol for 5 min. Seeds were dried on sterile Whatman filter paper in a sterile hood. Dried seeds were plated on medium containing 0.6% (w/v) agar, 2.3 g l−1 Murashige and Skoog salts, 0.5 g l−1 Mes (pH 5.8), 1% (w/v) sucrose and 0.1 g l−1 myo-inositol. Seeds were kept in the dark for 24 h at 4°C. Plates were wrapped with surgical tape and placed vertically in racks under a 16 h : 8 h, light : dark photoperiod at room temperature (22°C).
Cells of tobacco (Nicotiana tabacum L., cv Bright Yellow-2) line BY-2 (Nagata et al., 1992) were cultivated as described previously (Petrášek et al., 2006) and subcultured weekly. Arabidopsis thaliana, ecotype Landsberg erecta, suspension-cultured cells (May & Leaver, 1993) were maintained in the same medium as BY-2 cells and subcultured weekly.
Arabidopsis root elongation and lateral root formation assays
Wild-type and tir1 Arabidopsis seeds were germinated and grown for 4 d as described above. Seedlings were transferred to solid medium supplemented with a determined concentration of the compound to be analysed and the position of the root tip was marked. The seedlings were grown for an additional 4 d in continuous light under a yellow filter (Stasinopoulos & Hangarter, 1990) and the root length was measured from the marked point to the tip of the root.
For lateral root numbers, the primary root length was measured as explained above, but not from the marked point, but rather from just below the hypocotyl part to the root tip, and then photographed. Lateral root numbers were determined with the image analysis software ImageJ (http://rsb.info.nih.gov/nih-image) by marking every emerged lateral root, and the number of lateral roots per centimetre of primary length was calculated with respect to the primary root length data.
Determination of BY-2 cell division activity
From 1-wk-old BY-2 suspension culture, 7.5 ml was transferred to 500 ml of fresh auxin-free BY-2 medium. This culture was aliquoted (30 ml each) to 100-ml conical flasks, and the specified compound to be tested was added to each flask from the stock to give the required final concentration. The culture was further incubated in the dark at 27°C on an orbital shaker at 120 rpm for 1 wk. Samples of the culture were collected every day and the cell density (number of cells per 1 ml of cell suspension) was determined by counting cells in at least six aliquots of each sample using a Fuchs–Rosenthal haemocytometer slide.
DR5rev::GFP induction assay
Five to six, 4-d-old DR5rev::GFP seedlings (Benková et al., 2003) were transferred to sterile liquid medium (2.3 g l−1 Murashige and Skoog salts, 0.5 g l−1 Mes (pH 5.8), 1% (w/v) sucrose, 0.1 g l−1 myo-inositol). The required concentration (1 or 5 μM) of experimental compound was applied and the seedlings were incubated for 4 or 24 h wrapped with yellow long-pass filter film with transmittance below 450 nm (Stasinopoulos & Hangarter, 1990) on an orbital shaker (150 rpm). The green fluorescent protein (GFP) signal was observed in roots using a Zeiss LSM510 confocal laser scanning microscope (Carl Zeiss, Jena, Germany) with excitation at 488 nm using an argon laser and a ×20 objective. The quantification of the fluorescence intensity was performed with image analysis software ImageJ (http://rsb.info.nih.gov/ij/) by selecting the root tip area (meristem, transition and elongation zones included).
Analysis of the inhibitory effect of auxin on endocytosis
The experiments were performed as described in Paciorek et al. (2005). Four to 5-d-old wild-type Arabidopsis seedlings were incubated in sterile liquid medium supplemented with the compounds at 25 μM concentration for 30 min, followed by 90 min of treatment with 25 μM brefeldin A (BFA). The immunodetection of PIN1 and PIN2 proteins was performed according to Friml et al. (2003). The anti-PIN1 antibody (1 : 1000) (Benková et al., 2003), anti-PIN2 antibody (1 : 1000) (Abas et al., 2006) and fluorochrome-conjugated secondary antibody anti-rabbit-Cy3 (1 : 600) (Dianova, Hamburg, Germany) were used. Imaging was performed as described previously (Robert et al., 2010). The rate of the inhibitory effect of each compound was measured by counting the number of BFA bodies in individual cells.
Auxin accumulation assay
The accumulation of radiolabelled auxins was measured as described in Petrášek et al. (2006).
IBA metabolic assay
Two-day-old BY-2 cells were equilibrated in the uptake buffer as described for accumulation assays. Experiments were performed in the uptake buffer and under standard cultivation conditions. Cells were incubated with 20 nM [3H]IBA for a period from 30 s to 48 h in darkness. After incubation, the cells were filtered and 200 mg fresh weight cells (after draining the medium off under reduced pressure) were flash frozen in liquid nitrogen and stored at −80°C. The extraction and purification of the auxin metabolites from cells and media were performed as described previously (Dobrev et al., 2005). The analysis was monitored by a Ramona 2000 flow-through radioactivity detector (Raytest GmbH, Straubenhardt, Germany) after online mixing with three volumes (1.8 ml min−1) of liquid scintillation cocktail (Flo-Scint III, Perkin Elmer Life and Analytical Sciences, Shelton, CT, USA), and the results were normalized on the basis of the total accumulated radioactivity.
Differences between control and experimental variants were analysed using one-way ANOVA. When significant differences were detected at a 95% level of confidence, the multi-range Tukey's post-hoc test was applied. All statistical tests were performed using Microsoft Excel and GraphPad software (http://graphpad.com).
Selection of the auxin analogues
The selection of compounds was performed on the basis of their structural similarity to major native (IAA) and synthetic (NAA and 2,4-D) auxins. As all active auxins are weak organic acids and contain aromatic or heterocyclic ring(s), we have selected various derivatives of diverse rings bearing one carboxyl group on a short (carbon–carbon or oxygen–carbon) side-chain. Thus, carboxy derivatives of benzene (BA), indole (IAAs, IPA, IBA, ILA), naphthalene (NAAs) and phenol (PAA and phenoxyacetic acids) were used (Supporting Information Fig. S1). Compounds containing only corresponding heterocyclic and aromatic rings, that is, indole and naphthalene, respectively, and an auxin conjugate (IAAla) were chosen as controls.
Effect of auxin analogues on Arabidopsis primary root growth and lateral root formation
The auxin-like physiological competence of selected compounds was analysed in Arabidopsis thaliana seedlings using auxin assays based on the inhibition of primary root elongation and the induction of lateral root formation.
IAA and most of its halogenated derivatives inhibited primary root elongation efficiently at concentrations between 100 and 500 nM. The most active compound tested was 6-Cl-IAA, which exhibited maximum inhibition at a concentration as low as 10 nM (Fig. 1a). Non-halogenated indolic acids (IPA and IBA) required a much higher concentration (10 μM) for maximum effect. Other indole-based compounds had only slight effects on root elongation (Fig. 1b).
The application of NAA caused maximum inhibition of primary root elongation at 500 nM, whereas its positional analogue, 2-NAA, as well as the weak native auxin PAA, inhibited root elongation completely at a concentration one order of magnitude higher (Fig. 1c,d). By contrast, the dichlorophenoxy derivatives 2,4-D and 2,4-DP were very active, with maximum inhibitory effects at 100 nM, but 2,4,5-T, another structural analogue of this class, was comparatively less effective (Fig. 1d).
As TIR1 seems to be the auxin receptor that is important for many processes in plant development, we also analysed the sensitivity of tir1 knockout mutant seedlings to these compounds (Fig. 2). Most of the tested compounds showed more or less similar effects on the primary root elongation of both wild-type and tir1 seedlings. However, treatments with 2,4-D and its structural analogues, that is, 2,4,5-T and 2,4-DP, resulted in less inhibition in tir1 relative to the wild-type. By contrast, IPA showed the highest inhibition of primary root elongation in tir1 (Fig. 2, inset), and it appears to be the least dependent on TIR1 to express its auxin-like activity.
In lateral root initiation (Fig. 3), 6-Cl-IAA was the compound with the greatest activity (Fig. 3a). 5-Cl-IAA showed its highest effect at 1 μM; however, above 1 μM concentration its action was inhibitory. Similarly, NAA showed an increasing effect up to 5 μM and above this concentration its effect was inhibitory (Fig. 3c). IAA, 4-Cl-IAA and 5-F-IAA increased lateral root number with an increase in concentration of the tested compounds up to 10 μM, the last two compounds promoting more lateral roots than IAA. 2,4-D and 2,4-DP also showed high activity (Fig. 3d), but other compounds showed either very weak or no activity in this assay.
Arabidopsis root growth assays showed that, apart from IAA itself, the most potent auxin-like compounds included most of the IAA derivatives halogenated in positions 4, 5 and 6 of the indole ring, as well as 2,4-D and its structural analogues, and NAA.
Effects of auxin analogues on cell division in suspension-grown tobacco BY-2 cells
Compounds showing very obvious effects on root phenotype in Arabidopsis were tested for their cell division-stimulating activity in the auxin-dependent BY-2 tobacco cell culture. Standard culture medium for BY-2 cells is supplemented with 1 μM 2,4-D. The withdrawal of 2,4-D induces typical auxin starvation symptoms, such as the cessation of cell division, abnormal formation of amyloplasts and continuous abnormal cell elongation (Sakai et al., 2004; Petrášek et al., 2006). Therefore, in this assay, we replaced 1 μM 2,4-D in the culture medium with various concentrations of experimental compounds, and measured the cell density within 1 wk after inoculation into fresh medium (Fig. 4).
Surprisingly, the major native auxin IAA did not show any stimulatory effect on cell division at any of the concentrations used. Another native auxin, 4-Cl-IAA, showed a medium effect, but with strong cell division-promoting activity at 5 μM. The application of 5-Cl-IAA and 5-F-IAA also stimulated cell division with increasing concentrations, and both compounds were highly active at 5 μM. NAA was mildly active, whereas its structural counterpart, 2-NAA, exhibited no cell division-promoting activity at any of the concentrations tested. As expected, 2,4-D was the most active compound, but only at concentrations up to 3 μM; higher concentrations had an inhibitory effect. Interestingly, 2,4,5-T was highly active only at 0.5 μM and all other concentrations appeared to be inhibitory. Strikingly, IBA was very active at 5 μM (Fig. 4). As IBA was much more active in this assay than IAA itself, and as it has been suggested to act as an IAA precursor (Strader et al., 2010), we tracked the metabolic fate of radiolabelled IBA in BY-2 cells, tracing free radiolabelled IAA at different time points from 30 s to 48 h after labelled IBA application. Free labelled IAA was not detected at any of the time points measured (Fig. S2), suggesting that, under our experimental conditions, at least at the beginning of the subculture interval, IBA stimulates cell division independently of IAA.
Altogether, this assay confirmed 2,4-D as a preferential auxin-like compound for the standard growth of BY-2 cells and indicated that other 2,4-D-like compounds (namely 2,4,5-T) are able to promote cell division efficiently at low concentration. It also pointed to an IAA-independent action of IBA in BY-2 cells.
Active transport of auxin analogues across PM in suspension-grown cells
For tobacco cells, auxin influx and efflux carrier affinity has been attributed mostly to the major synthetic auxins, either 2,4-D or NAA, respectively, whereas the native auxin IAA is a good substrate for both classes of carriers in cells of tobacco cv Xanthi XHFD8 (Delbarre et al., 1996) or BY-2 and VBI-0 (Paciorek et al., 2005). This distinct substrate specificity of auxin carriers facilitates the quantitative analysis of the active influx and efflux of different compounds during accumulation assays in tobacco cells: using a simple displacement of [3H]2,4-D and [3H]NAA, respectively, by non-labelled auxin analogues, the relative affinity of a particular type of carrier for such analogues can be measured. For comparison, we also performed competitive auxin accumulation assays on suspension-grown Arabidopsis cells.
Of all the compounds tested, at the influx level, IAA maximally decreased the accumulation of [3H]2,4-D (Figs 5a, S3a). In tobacco cells, this decrease was even higher than that resulting from the application of the auxin influx carrier inhibitor 2-naphthoxy acetic acid (2-NOA, Laňková et al., 2010) or unlabelled 2,4-D. Halogenated derivatives of IAA were also highly competitive to labelled 2,4-D in both tobacco (Fig. 5a) and Arabidopsis (Fig. S4a) cells. NAA did not compete for auxin influx carriers but, notably, its structural isomer 2-NAA decreased the accumulation of labelled 2,4-D significantly in both tobacco (Fig. 5a) and Arabidopsis (Fig. S4a) cells. This finding is consistent with its status as an auxin influx carrier ‘inhibitor’ in tobacco cv Xanthi XHFD8 cells (Delbarre et al., 1996). Surprisingly, IBA increased slightly the accumulation of labelled 2,4-D in tobacco cells, but did not show any effect in Arabidopsis cells (Figs 5a, S4a). Representative shapes of accumulation curves for all experimental compounds are shown in Figs S3(a) and S5(a) for BY-2 and Arabidopsis, respectively.
IAA and all of its halogenated derivatives competed efficiently with labelled NAA for active auxin efflux (thus resulting in increased [3H]NAA accumulation in cells), but the major native auxin IAA showed the least activity. Of all the compounds tested in tobacco cells, 5-Cl-IAA competed the most strongly, followed by 6-Cl-IAA and 5-Br-IAA (Fig. 5b); however, in Arabidopsis cells, 5-Br-IAA was the most effective competitor (Fig. S4b). Non-labelled NAA appeared to be less competitive than some other compounds (such as halogenated IAAs), and 2-NAA was only a slightly less active inhibitor than NAA itself in both tobacco and Arabidopsis cells. A widely used chemical inhibitor of active auxin efflux, 1-naphthylphthalamic acid (NPA), was used as a positive control; it showed the maximum inhibition of auxin efflux, being relatively more effective in Arabidopsis than in BY-2 cells. Representative shapes of the accumulation curves for all compounds tested are shown in Figs S3(b) and S5(b) for BY-2 and Arabidopsis, respectively. These findings reveal distinct specificities of auxin influx and efflux carriers towards some synthetic auxins.
The ability of auxin analogues to induce DR5-driven gene expression
To address the possible activity of the auxin analogues in eliciting an auxin response, we tested their ability to induce the expression of DR5rev::GFP in Arabidopsis roots. Generally, IAA and all of its halogenated derivatives induced DR5rev::GFP expression efficiently, and IAA itself and 6-Cl-IAA were the most active. Similarly, 2,4-D and 2,4-DP showed high levels of activity. NAA induced DR5rev::GFP to a slightly lesser extent and other compounds did not show statistically significant effects (Fig. 6a,b). To test the possibility that some of the inactive compounds may only be active at higher concentrations and/or only after longer treatment, the same assay was repeated using higher concentrations (5 μM) of all the weakly active and inactive compounds and for a longer time of treatment (24 h). Under these conditions, 2-NAA induced a considerable effect and 2,4,5-T exhibited a very high activity (Fig. S6).
Taken together, these results show that those compounds that affected the Arabidopsis root phenotype were also active in inducing DR5rev::GFP expression; however, at higher concentrations and after longer treatment, certain other compounds also induced DR5rev::GFP expression to a considerable extent.
The involvement of auxin analogues in the inhibition of endocytosis of PIN auxin efflux carriers
BFA is widely used to inhibit vesicle trafficking from endosomal compartments to the PM in plant tissues, and treatment with BFA leads to the deposition of some PM proteins in so-called BFA compartments inside cells (Geldner et al., 2001). Thus, BFA can be used to visualize proteins internalized by endocytosis. We followed the inhibitory effect of auxins and auxin-like compounds on the BFA-induced intracellular deposition of PIN1 and PIN2 in Arabidopsis root stele and cortex/epidermis regions, respectively. The remarkable inhibitory effect of some auxins (namely 2,4-D and NAA) on endocytosis has been reported previously (Paciorek et al., 2005).
As shown in Fig. 7, 5-Br-IAA was the most active compound in inhibiting the endocytosis of PINs. Halogenated derivatives of IAA, as well as NAA, 2-NAA, 2,4-D, 2,4,5-T and 2,4-DP, inhibited PIN1 and PIN2 internalization with high efficiency, and certain other compounds, including IAA itself, also exhibited activity (Fig. 7b). However, there was one exception among compounds active in both the inhibition of endocytosis and in other assays, namely 5-F-IAA, which was entirely ineffective here (Fig. 7a,b). These results showed that, with the exception of 5-F-IAA, which induces DR5rev::GFP expression, but does not inhibit endocytosis (cf. Robert et al., 2010), auxins and auxin analogues active in DR5rev::GFP induction also inhibit endocytosis.
The observed inhibitory effect of 2-NAA on endocytosis does not agree with previous reports, in which experiments under slightly different conditions suggested that 2-NAA is not effective compared with NAA (Paciorek et al., 2005). Therefore, we compared the effects of 2-NAA and NAA at different concentrations. Our conditions with 25 μM BFA treatments are generally more sensitive than those used in previous studies. Furthermore, we estimated quantitatively the number of BFA bodies, which further increased the precision of our measurements. Using these techniques, we determined that the inhibitory effect of 2-NAA is indeed weaker than that of NAA, but that, at a concentration higher than 5 μM, both compounds are effective (Fig. S7).
Physiological effects of auxins and their analogues
The introduction of different halogen atoms into various positions on the indole ring of IAA had a positive effect on the auxin-like activity in various bio-assays, for example, in the pea curvature test (Porter & Thimann, 1965), and, in particular, substitution by Cl at the 4 position of the indole ring to give 4-Cl-IAA had a strong stimulatory effect on oat and pea growth relative to IAA itself (Böttger et al., 1978; Katekar & Geisler, 1983). In the wheat coleoptile straight growth bio-assay for auxins, 4-Cl-IAA, 5-Cl-IAA and 6-Cl-IAA were considerably more active than IAA itself, their individual activities corresponding to their ability to displace labelled IAA bound to the PM-enriched microsomal fraction (Zažímalová & Kutáček, 1985). A report on pea fruit growth demonstrated the activity of 4-Cl-IAA and the inactivity of IAA itself (Reinecke et al., 1995).
Similarly, in the present study, the observed effect of halogenation on auxin-like activity was different in different bioassays. In the primary root elongation inhibition assay in Arabidopsis, 6-Cl-IAA was more active than IAA, whereas 5-Cl-IAA and 5-Br-IAA showed much lower activity. In lateral root formation, 5-Cl-IAA was much more active than any other halogenated derivative and IAA itself. A negative effect of halogenation appeared in 2,4,5-T, in which the addition of a further Cl atom to 2,4-D at position 5 of the benzene ring reduced its auxin-like activity considerably. In contrast with this, the addition of a methyl group to the side-chain had little effect on activity in both primary root and lateral root assays, as seen in the case of 2,4-DP. These findings show that halogenation at different ring positions results in distinct activities in various physiological processes. They also highlight the fine differences in the specificity patterns of these processes towards compounds which are not native and so could not have created selection pressure during evolution.
Some compounds showed activity only in some of the assays performed. Thus, IAA and NAA, which were very potent in root assays, were not the most active auxins for the stimulation of BY-2 cell division. Here, 2,4-D and related compounds were the most active, probably reflecting differential mechanisms for NAA and 2,4-D action in BY-2 cells (Campanoni & Nick, 2005). Next, 2-NAA, a positional analogue of NAA, was quite effective in both primary root growth inhibition and the promotion of lateral roots at higher concentrations. This observation is very interesting, because this compound was either considered as a very weak auxin (Katekar, 1979) or as an inactive analogue of the potent synthetic auxin NAA (Kepinski & Leyser, 2005). The compound corresponding to the basic core of the IAA molecule – indole –inhibited slightly primary root elongation at high concentrations, possibly as a result of its likely role as a precursor for the tryptophan-independent IAA biosynthetic pathway (Woodward & Bartel, 2005).
As reported previously (Hayashi et al., 2008), lengthening of the side-chain of an auxin molecule reduces its auxin-like activity, and may explain why IBA and IPA are much weaker auxins than IAA. However, and importantly, IBA was effective in the stimulation division of BY-2 cells, whereas IAA showed almost no activity. There are currently two different views on the auxin-like activity of IBA: one notion considers IBA to be a precursor which is converted to IAA by β-oxidation (Zolman et al., 2008; Strader & Bartel, 2011); the other suggests that IBA acts independently of IAA (Ludwig-Müller, 2000). A recent study using stable isotope labelling of IAA and IBA showed that a considerable amount of IBA is converted to IAA in Arabidopsis hypocotyls (Liu et al., 2012). However, interestingly, the presence of endogenous IBA in Arabidopsis thaliana has been challenged recently (Novák et al., 2012). Our study did not reveal any IAA converted from [3H]IBA in tobacco cells; this could be either because the metabolic machinery in Arabidopsis and tobacco BY-2 cells differs considerably in terms of conversion of IBA to IAA, or because the generated IAA is very rapidly (within a few seconds) converted to IAA metabolites.
Comparison of the structure–activity relationships for root elongation, lateral root formation (both in Arabidopsis seedlings) and cell division (tobacco BY-2 cell line) highlights differences in the structural requirements of these auxin-related physiological processes, thus making the differential phenotypic outcome of the same compound a very important aspect of auxin biology.
Intercellular auxin transport
In view of the results described above, it is not surprising that the substrate specificity of both auxin influx and efflux carriers differs between BY-2 and Arabidopsis cells. An interesting observation in several assays performed was that the activity of the major native auxin IAA was much lower than that of some of its synthetic and even native analogues (typically 4-Cl-IAA). This could be explained by the fast metabolic turnover of IAA. With respect to the fact that IAA is the major endogenous auxin, a well-organized metabolic machinery can be expected to be present, with preferential affinity towards IAA. Fast metabolism of NAA and IAA has been reported in tobacco cells (Nicotiana tabacum L., cv Xanthi XHFD8 and cv BY-2; Delbarre et al., 1994 and Hošek et al., 2012, respectively). As these metabolic changes proceed inside cells, they should not affect the auxin influx assay in which IAA is coming from outside (in its non-metabolized form); indeed, IAA shows a high degree of competition with the influx carrier ‘marker’ 2,4-D. In the auxin efflux assay, cold NAA competed with [3H]NAA much less than did some of the halogenated IAAs, suggesting that endogenous auxin efflux carriers in BY-2 cells possess higher affinity towards some halogenated IAAs relative to NAA. Carrier-mediated but AUX1- and PIN-independent transport of IBA has been reported in Arabidopsis (Rashotte et al., 2003; Poupart et al., 2005; Yang et al., 2006). An interesting observation in this study was that IBA consistently stimulated an increase (of c. 10%) in [3H]2,4-D accumulation in BY-2 cells. This may reflect reports that PDR9 (a member of the ABCG/PDR subfamily of ABC transporters) transports 2,4-D out of cells (Ito & Gray, 2006) and that IBA also utilizes the same system for its efflux (Strader et al., 2008); our data may therefore reflect the competition of cold IBA for the PDR-like-driven export of [3H]2,4-D from cells, thus increasing its accumulation. However, in Arabidopsis cells, we did not observe any increase in [3H]2,4-D intracellular accumulation by IBA. This might be explained either by the different relative affinity of efflux carriers towards IBA and 2,4-D in tobacco and Arabidopsis cells and/or by the fact that the Arabidopsis cell suspension used here was derived from stem explants (May & Leaver, 1993), whereas PDR9 genes are widely expressed only in roots (van den Brule & Smart, 2002; Růžička et al., 2010). A parallel IBA transport machinery has recently been proposed to play a role in auxin homeostasis by actively transporting IBA across the PM as a storage form of IAA (Strader & Bartel, 2009, 2011; Růžička et al., 2010).
Basically, auxin signalling pathways result in either the regulation of gene expression (TIR1-related, Calderon-Villalobos et al., 2010) or in effects which do not involve directly transcription or translation (mediated by ABP1, Badescu & Napier, 2006; Robert et al., 2010). In this study, the activation by auxin analogues of DR5rev::GFP, reflecting the transcriptional control of auxin signalling pathways, is in good agreement with Arabidopsis root phenotyping. This is further supported by visible induction of the DR5-GFP signal by IAAla and IBA only after prolonged treatment and at higher concentrations, suggesting that this signal could originate from the IAA released (Rampey et al., 2004) or converted (Woodward & Bartel, 2005) from them. IBA itself does not seem to be a good substrate for the TIR1-mediated pathway, probably because of its longer side-chain (Hayashi et al., 2008).
The resistance of the tir1 mutant to primary root elongation inhibition by auxins was relatively low and may be caused by the functional redundancy between TIR1 and other members of the AFB family of auxin receptors, as reported previously (Parry et al., 2009). Compared with the wild-type, the mutant was notably resistant only to 2,4-D and its structural analogues, namely 2,4-DP and 2,4,5-T. The effect of 2,4,5-T was rather unexpected because of its weak ability to induce DR5rev::GFP expression. Therefore, the auxin-like activity of 2,4,5-T could be attributed mainly to the interaction with the TIR1 co-receptor itself, and its interaction with other members of the AFB family may be very low. A similar effect has been reported previously for a mutant of the TIR1 homologue AFB5, which is resistant to the picolinate-type auxin picloram, but sensitive to 2,4-D and IAA (Walsh et al., 2006). Moreover, the activity of 2,4,5-T might also reflect fine tuning of the recently identified TIR1/AFB-Aux/IAA co-receptor machinery (Calderon Villalobos et al., 2012). As functional redundancy obviously compromises the accuracy of specificity studies performed in vivo, other methods, such as protein crystallography of individual proteins, as performed for TIR1 (Tan et al., 2007), can provide much better information about structure–activity relationships for particular auxin (co-)receptors.
The inhibition of endocytosis was investigated to obtain an insight into the ‘non-transcriptional’ auxin signalling pathway(s). It revealed the non-canonical behaviour of 5-F-IAA, which was not active in inhibiting endocytosis, even though it was very effective in inducing a DR5-GFP signal (and also effective in certain other assays). A more detailed and focused study on the inhibition of endocytosis of PM proteins has been reported by Robert et al. (2010). Interestingly, the inactivity of 5-F-IAA in the inhibition of the endocytosis of PINs may also contribute to its lower activity (relative to other halogenated IAAs) in competition for active auxin efflux: when endocytosis is not inhibited, there are fewer PIN efflux carriers on the PM for which to compete.
Altogether, this study shows that the auxin-like potential is a very complex property of a compound, determined by its chemical and metabolic stability, ability to distribute in a tissue in a polar manner and its efficacy to activate various auxin signalling pathways. In Fig. 8, the auxin activities of selected compounds are presented at selected concentrations and normalized in relation to relevant controls. This figure shows that, except for the presence of both aromatic (or heterocyclic) ring(s) and a carboxyl group, no general criteria can be set for the specificity patterns of different auxin-related processes. The non-canonical behaviour of some synthetic auxin analogues is worth stressing because of their potential to be used as tools in further studies dissecting the molecular mechanism of auxin action. Thus, IPA and 2,4,5-T may serve to distinguish between TIR1 and other AFB protein-related processes, and 5-F-IAA can help to discriminate between transcriptional and non-transcriptional (ABP1-dependent) pathways of auxin signalling. Last, but not least, this study has demonstrated that structure–activity relationships determined precisely at the level of a particular protein (receptor, carrier) may not correspond fully to the final auxin-like physiological activity of a particular compound.
The authors thank Dr Christian Luschnig (University of Natural Resources and Life Sciences (BOKU), Vienna, Austria) for the anti-PIN2 antibody, Professor Mark Estelle (University of California, San Diego, CA, USA) for tir1-1 mutant seeds and, last but not least, to Dr David Morris for critical reading of the manuscript. We also thank Markéta Pařezová and Jana Stýblová for excellent technical assistance. This work was supported by the Grant Agency of the Czech Republic (P305/11/0797 to E.Z. and 13-40637S to J.F.), the Central European Institute of Technology project CZ.1.05/1.1.00/02.0068 from the European Regional Development Fund and by a European Research Council starting independent research grant ERC-2011-StG-20101109-PSDP (to J.F.).