Cytosolic acidification but not auxin at physiological concentration is an activator of MAP kinases in tobacco cells

Summary

In higher plants, MAP kinase cascades are involved in the transduction of numerous stress-related signals but much less is known about the effect of mitogenic signals. We have analysed MAP kinase activation in tobacco cells after treatment by auxin, a growth factor required at physiological concentrations for mitosis in plant cell cultures. From in-gel assay of myelin basic protein kinase and from immunochemical detection of ERK related kinases, we show that the mitogenic effect of auxin, which was confirmed by the specific increase of several mRNAs species, did not rely on MAP kinase activation within the first 2 hours. These data contest previous results which could be due to the activation of MAP kinase by a signal other than auxin. In the second part of this study, we show that the treatment of the cells with high concentrations of various weak lipophilic acids such as auxin, in a non-physiological concentration range, butyric or acetic acid is sufficient to activate transiently a MAP kinase. The data show that MAP kinase activation is the consequence of cytosolic acidification. Moreover, it is not sensitive to the protein kinase inhibitor staurosporine. These results suggest a functional role for cytosolic acidification as a second messenger mediating MAP kinase activation in the response of plant cells to various stresses.

Introduction

The Mitogen Activated Protein Kinase (MAP kinase) cascade is a conserved signalling element in the transduction of numerous extracellular signals controlling growth, differentiation and stress adaptation in animals and fungi ( Guan 1994;Pelech et al. 1993 ;Seger & Krebs 1995). Major targets of this cascade are nucleus-located transcriptional regulators, which makes the MAP kinase cascade an essential link between transmembrane signalling and transcriptional control. MAP kinases are ss ( Hill & Treisman 1995;Posas et al. 1996 ;Post & Brown 1996)erine-threonine kinases which are activated by dual phosphorylation on tyrosine and threonine residues owing to dual specificity kinases, called MAPKKs or MEKs. These latter are activated by phosphorylation through the serine kinases MAPKKKs, which include the two groups MEKKs and Raf. The MAP kinase family represents three related subfamilies of kinases which respond to different extracellular signals ( Davis 1994): the Extracellular Regulated Kinases (ERKs) respond primarily to growth factors; the Stress Activated Protein Kinases (SAPKs/JNKs), and the p38 kinase and HOG1-homologous kinases respond to various environmental and cellular stress signals. In addition, MAP kinases possibly defining new subfamilies have been discovered recently ( Robinson & Cobb 1997). In Saccharomyces cerevisiae, five different MAP kinase cascades have been defined in response to a variety of signals ( Bardwell et al. 1996 ;Herskowitz 1995). A complex pattern of regulation of the MAP kinase pathways has emerged from studies in animal and yeast: depending on cellular context and receptors, one MAP kinase cascade may be involved in different responses related to differentiation or proliferation. Moreover, a given signal may be mediated by different pathways according to cell type ( Blumer & Johnson 1994;Kosako et al. 1996 ).

Recently, structural, genetic and biochemical lines of evidence have demonstrated the conservation of MAP kinase cascades in higher plants. A number of genes encoding homologues of MAP kinases ( Hirt 1997;Mizoguchi et al. 1993 ;Mizoguchi et al. 1997 ), and of MAPKKs and MAPKKKs ( Banno et al. 1993 ;Mizoguchi et al. 1996 ;Mizoguchi et al. 1997 ;Nishihama et al. 1995 ;Shibata et al. 1995 ) have been isolated in plants. Genetic analysis has demonstrated a Raf homologue, CTR1, to be involved in ethylene signalling ( Kieber et al. 1993 ), downstream of a two-component sensor-regulator transmembrane protein ( Chang 1996;Hirt 1997). MAP kinase-like activities were shown to increase rapidly and transiently in plants in response to a number of biotic and abiotic stimuli including abscisic acid ( Knetsch et al. 1996 ), salicylic acid ( Zhang & Klessig 1997), elicitors ( Ligterink et al. 1997 ;Suzuki & Shinshi 1995;Zhang et al. 1998 ), wounding ( Bogre et al. 1997 ;Seo et al. 1995 ;Usami et al. 1995 ), low temperature and drought ( Jonak et al. 1996 ), and hydration ( Wilson et al. 1997 ). A specific feature of plants, as compared with other eukaryotes, is that some MAP kinases, which represent one subfamily ( Zhang & Klessig 1997), are also regulated at the mRNA level ( Bogre et al. 1997 ;Huttly & Phillips 1995;Jonak et al. 1996 ;Ligterink et al. 1997 ;Mizoguchi et al. 1996 ;Seo et al. 1995 ).

Very few studies have been reported on MAP kinase activation in relation to cell division in plants, whereas in animals, ERKs are typically activated in response to proliferative signals. Auxins are pleiotropic plant growth factors which are intimately involved in many aspects of plant development including cell division and cell elongation ( Abel & Theologis 1996;Napier & Venis 1995;Walden & Lubenow 1996). Plasma membrane-located auxin receptors are believed to play a key role in the control of pumps and ionic channels ( Napier & Venis 1995), but limited data are available about auxin signalling in the control of cell division ( Coenen & Lomax 1997). Preliminary evidence in favour of MAP kinase activation by auxin has been obtained recently from tobacco cells which rely on auxin for cell division ( Mizoguchi et al. 1994 ). Given the importance of this result in the elucidation of auxin signalling, we have investigated in depth the effect of auxin on MAP kinase activation. We show that tobacco cells do not respond to auxin by activation of a MAP kinase-related MBP kinase. We also found, in the course of this study, that a MAP kinase pathway can be activated in plant cells merely by cytosolic acidification, which indicate a key role for cytosolic pH in mediating MAP kinase activation in plants.

Results

The response of tobacco cells to auxin

The growth of tobacco cells (line BY-2) was dependent on the presence of the synthetic auxin 2,4-D at 0.9 μm in the culture medium. The subculture of the cells with a low 2,4-D concentration (0.05 μm) reduced the extent of cell divisions, the mitotic index of the culture becoming negligible after 4–5 days because of auxin starvation ( Fig. 1a). When 2,4-D was added at optimal concentration for growth (0.9 μm) to 4-day-old auxin-starved cells, the mitoses resumed after 10 h and a peak of synchronous divisions occurred at 20 h ( Fig. 1b). The re-addition of auxin induced a sustained growth of the cells for 3 days, with a mitotic index similar to the one recorded during a standard growth cycle with 0.9 μm 2,4-D ( Fig. 1a).

Figure 1. The control of cell division by auxin in the tobacco BY-2 cell line.

The cells were cultured with 2,4-D at the concentration of 0.9 μm (•) or 0.05 μm (auxin starvation, ○). At day 4 (arrow), 2,4-D was added to auxin-starved cells at the concentration of 0.9 μm (▪). The data are the mean of two separate experiments in which 1000 DAPI-stained nuclei were counted for each measurement of mitotic index. The standard deviations of mitotic indices were equal to or below 1%.

(a) The mitotic index was measured once a day in the three conditions.

(b) The mitotic index was measured during the first day after auxin addition to auxin-starved cells.

The treatment of the auxin-starved cells by 2,4-D at 0.9 μm led to an increase of the abundance of several mRNA species chosen as markers of auxin response ( Fig. 2). Two of them known to be expressed only in dividing cells, ArcA ( Ishida et al. 1993 ) and a tobacco A-type cyclin CycA3;2 ( Reichheld et al. 1996 ), did accumulate up to 4–6-fold above their basal level 10 h after auxin treatment. They were slightly more abundant at that time than in cells from a normal growth cycle with 0.9 μm 2,4-D, indicating the relative synchronisation of cell divisions ( Fig. 2, lanes ‘Expo’). After 10 h of auxin treatment their level declined and returned to the basal level at 24 h ( Fig. 2b). The relative abundance of two other auxin-responsive mRNAs increased with a kinetics similar to ArcA and CycA3;2 ( Fig. 2). The most dramatic increase in abundance (ninefold) was for a glutathione transferase gene, Nt103–1, a gene not involved in cell division ( van der Zaal et al. 1991 ); whereas there was approximately a fourfold increase for NtIAA4.3, a gene belonging to a large multigene family of auxin-responsive putative transcriptional regulators ( Abel & Theologis 1996). This transcript was also at a higher level in exponentially growing cells.

Figure 2. The control of mRNA abundance by auxin.

Total RNAs were extracted from exponentially growing tobacco cells of 4-day-old cultures with 0.9 μm 2,4-D (lane ‘Expo’ in a and b) and from auxin-starved cells from 4-day-old cultures with 0.05 μm 2,4-D which had been further treated for various times with 2,4-D at 0.9 μm (closed symbols in b) or with a control solution (open symbols in b). Ten μg of total RNAs were analysed by Northern blotting. They were hybridised with random-labelled cDNAs for tobacco ArcA (▴, ▵), CycA3;2 (▪, □), NtIAA4.3 (♦, ◆), Nt103–1 (•, ○) and 25S rRNA.

(a) Autoradiographic signal.

(b) Quantification of mRNA abundance by reference to the hybridisation signal of rRNA. The values were further normalised relative to the signals from auxin-starved cells at time zero.

Auxin activates MBP kinase only at toxic, non-physiological concentrations

MAP kinase activation was assessed by in-gel assay of MBP kinase activity in soluble protein extracts. Nearly no kinase activity was detected in untreated tobacco cells ( Fig. 3a, lane 1). Three activities of 32, 36 and 47 kDa were weakly detected in overexposed gels in the presence as well as in the absence of MBP in the gels, which indicates autophosphorylation (data not shown). The addition of 2,4-D at 0.9 μm to auxin-starved cells was never followed by MBP kinase activation during the 2 h treatment ( Fig. 4a). This result was observed in several experiments, with 5 min intervals during the first hour. Since the lack of MBP kinase activation after auxin treatment opposes a previous result ( Mizoguchi et al. 1994 ), we tested the possibility that this response requires a concentration of 2,4-D higher than 0.9 μm. No activation of MBP kinase was recorded for 2,4-D concentrations up to 30 μm, i.e. effective growth-promoting concentrations ( Fig. 4b). However, a significant activation of a single 47 kDa MBP kinase was observed after 15 min treatment by 2,4-D at a very high concentration, 300 μm ( Fig. 3a lane 2;Fig. 4b). The activated kinase was detected in the presence of MBP in the gel, and, to a much lower extent, in the presence of histone H1. It was not detected in the presence of casein or in the absence of substrate in the gel (data not shown). The concentration of 300 μm of 2,4-D, together with 90 μm which did not activate MBP kinase, was toxic for the cells ( Fig. 4b). Cell death was detected several hours after treatment with 2,4-D concentration of 90 μm and higher, and it affected the whole cell population after 18 h treatment. The auxin analogue 2,3-D, which does not have the growth promoting effect of 2,4-D but shares the same toxicity, displayed the same dose–response of MBP kinase activation as 2,4-D ( Fig. 4b).

Figure 3. Comigration of the MBP kinase activated by high 2,4-D or butyric acid concentrations with a polypeptide related to the catalytically activated form of ERK1/2.

Tobacco cells from 4-day-old cultures grown in the presence of 2,4-D at 0.05 μm were treated with a mock solution (1), 300 μm 2,4-D (2) or 2 m m butyric acid (3). Protein extracts (40 μg), prepared after 15 min treatment for control and 300 μm 2,4-D, and after 6 min treatment for butyric acid, were analysed by SDS–PAGE. The arrowhead shows the 47 kDa molecular mass.

(a) In-gel kinase assay of MBP phosphorylation.

(b) Western blot analysis of polypeptides related to ERK1. The duration of detection of the luminescence signal was less than in (c) and (d).

(c) Western blot analysis of polypeptides related to phospho-MAP kinase. The duration of detection of the luminescence signal was the same than in (d).

(d) Control Western blot analysis with only the secondary antibody antirabbit IgG.

Figure 4. The effect of auxin on MAP kinase activation.

Tobacco cells from 4-day-old cultures grown in the presence of 2,4-D at 0.05 μm were treated with different concentrations of active (2,4-D) or inactive (2,3-D) auxins and soluble protein extracts were prepared at indicated times. Fifty μg protein were analysed by in-gel kinase assay of MBP phosphorylation and by immunodetection of total and catalytically activated MAP kinase with anti ERK1 and anti phospho-MAP kinase antibodies. The mass of the polypeptides shown is 47 kDa.

(a) Kinetic analysis of MBP kinase after treatment by 2,4-D at 0.9 μm.

(b) 2,4-D and 2,3-D dose–response for MBP kinase. The extracts were prepared 15 min after the addition of the auxins to the cultures. The effect of each treatment on cell viability was assessed in separate samples after 24 h by microscopic examination with vital dyes (+ , over 75% of the cells remained alive; -, all the cells were dead).

(c) Kinetic analysis of MBP kinase, ERK1- and phospho-MAP kinase-related polypeptides after treatment of the cells by 2,4-D or 2,3-D at 300 μm.

(d) Effect of a subculture of the cells in a fresh medium in the absence or in the presence of 0.9 μm 2,4-D. Auxin-starved cells were subcultured into a fresh medium by filtration on a 110 μm nylon mesh and resuspension into the fresh medium. MBP kinase was analysed in the extracts prepared 10 min after the subculture. In the same experiment, two control treatments were performed. The negative control consisted in adding a mock solution to the suspension of auxin-starved cells. The positive control consisted in adding 2,4-D at 300 μm to the suspension of auxin-starved cells. For both controls, MBP kinase was analysed in the extracts prepared 15 min after the treatments.

An antibody raised against mammalian ERK1 recognised five major polypeptides of 22.4, 40.5, 44.2, 47 and 55.7 kDa in tobacco cell extracts ( Fig. 3b). The MBP kinase activity detected after treatment with 2,4-D at 300 μm comigrated with the 47 kDa ERK1-related polypeptide ( Fig. 3a–b, lane 2). The abundance of this polypeptide did not vary according to auxin concentration ( Fig. 3b). In contrast, an antibody directed against the catalytically activated form of mammalian ERK1/2 did not specifically detect any polypeptide in untreated tobacco cells. It detected only one 47 kDa polypeptide in the cells treated by 2,4-D at 300 μm, which comigrated exactly with the MBP kinase and the ERK1-related polypeptide ( Fig. 3a–d, lanes 1,2).

MBP kinase activity was the highest 15 min after addition of 2,4-D or 2,3-D at 300 μm and declined afterwards ( Fig. 4c), returning to the initial level of activity 2 h after treatment. The relative abundance of the 47 kDa polypeptide recognised by the antibody against ERK1 did not vary during 2 h of auxin treatment. The antibody against activated ERK1/2 recognised the 47 kDa polypeptide mostly in the extracts made after 15 min auxin treatment, i.e. displaying an active MBP kinase ( Fig. 4c).

When auxin addition was not made by simply adding the hormone, but by subculturing auxin-starved cells into a fresh medium containing 0,9 μm 2,4-D, a 47 kDa MBP kinase was activated after 10 min ( Fig. 4d). However, this activation did not require the presence of auxin in the fresh medium, as a control treatment in which the cells were subcultured in the absence of auxin also led to MBP kinase activation ( Fig. 4d). This experiment demonstrates the importance of appropriate controls when assessing MBP kinase activation in plant cells.

Cytosolic acidification is an activator of MBP kinase

The previously described activation of MBP kinase by high concentrations of both active and inactive forms of auxin could be due to an acidification of cytosol induced by the accumulation of these weak acids in the cells. To evaluate this hypothesis, the cytosolic acidification induced by these two auxin forms was measured from the relative accumulation of [¹⁴C]-benzoic acid in the cells ( Mathieu et al. 1996 ), and MBP kinase activation by two non-auxin lipophilic weak acids was studied. The cytosolic pH of auxin-starved cells was found to be ≈ 7.2, with the method enabling the detection of variations as low as 0.08 pH units. When the cytosolic pH was measured after auxin addition, as a function of time and of 2,4-D concentration, a significant acidification of the cytosol was only observed for 300 μm 2,4-D ( Fig. 5a). It started after 2.5 min and reached 1.5 pH units or more at 15 min. It was observed that the concentration of 2,4-D optimal for cell division, 0.9 μm, led to nearly no change, or only a slight early increase by less than 0.2 pH unit, of the cytosolic pH. A significant cytosolic acidification was also found for the highest concentration of 2,3-D, 300 μm, ( Fig. 5b).

Figure 5. Time dependence and dose effect of auxins and butyric acid treatments on the cytosolic pH.

Different concentrations of 2,4-D, 2,3-D or butyric acid were added at time zero to 4 ml tobacco cell cultures grown for 4 days with 2,4-D at 0.05 μm, and previously equilibrated for two hours with 50 m m Mes-Tris pH 5.2. After various incubation times [¹⁴C]-benzoic acid was added to the cultures at the final concentration of 0.5 μm. The cytosolic pH was calculated from the accumulation ratio of benzoic acid in the cells versus the medium after 2.5 min incubation of the labelled compound as described in Experimental procedures. The cytosolic pH of cells from cultures untreated by auxins or butyric acid was calculated to be ≈ 7.2, with SD = 0.06 determined within each experiment on eight separate samples. The data show the effect of treatments on the variation of the cytosolic pH from this latter value (mean ± SD of 3 independent experiments).

(a), Effect of 2,4-D at 0.9 μm (•), 90 μm (○), 300 μm (▪).

(b) Effect of 2,3-D at the same concentrations as 2,4-D.

(c) Effect of butyric acid at 300 μm (•), 1 m m (○), 2 m m (□), 5 m m (▪).

When the cells were incubated with various concentrations of butyric acid, a decrease of the cytosolic pH was recorded for all the concentrations tested between 0.3 and 5 m m ( Fig. 5c). The decrease in pH was the greatest after 2.5 min, with the extent of acidification ranging from 0.3 pH unit for 0.3 m m butyric acid to 1.6 pH unit for 5 m m butyric acid. This was followed by a recovery phase to a higher cytosolic pH value in all cases except at 5 m m butyric acid ( Fig. 5c). No toxic effect of butyric acid was observed within 24 h for concentrations up to 5 m m.

A MBP kinase was found to be activated after 6 min treatment by the various concentrations of butyric acid ( Fig. 6a). It was hardly detectable for the concentration of 0.3 m m butyric acid, and it reached a maximum for the concentration of 5 m m butyric acid ( Fig. 6a). It comigrated with the 47 kDa MBP kinase activated by 300 μm 2,4-D ( Fig. 3a, lanes 2–3). A time course experiment with 2 m m butyric acid showed that the 47 kDa MBP kinase was strongly activated between 3 and 15 min, with a maximum at 6 min, and disappeared thereafter ( Fig. 6b). The activation of MBP kinase by butyric acid was quantitatively correlated with the presence of a 47 kDa polypeptide immunologically related to the catalytically activated form of a ERK1/2-related polypeptide, whereas the overall abundance of ERK1-related polypeptides did not change ( Fig. 3c, lane 3;Fig. 6b).

Figure 6. Effect of butyric and acetic acids on MAP kinase activation.

Tobacco cells from 4-day-old cultures grown with 2,4-D at 0.05 μm were treated with different concentrations of butyric or acetic acid and soluble protein extracts were prepared at indicated times. Fifty μg protein were analysed by in-gel kinase assay of MBP phosphorylation and by immunodetection with anti ERK1 and anti phospho-MAP kinase antibodies. The mass of the polypeptides shown is 47 kDa.

(a) Dose–response of butyric and acetic acids for MBP kinase. The extracts were prepared 6 min after the addition of the acid to the cultures.

(b) Kinetic analysis of MBP kinase, ERK1 and phospho-MAP kinase-related polypeptides after treatment of the cells by butyric acid at 2 m m.

As butyric acid may exert effects not linked to pH change, we have also analysed the effect of another lipophilic acid widely used for acid-loading, acetic acid, on MBP kinase activation. Treating the cells with raising concentrations of acetic acid induces the activation of a 47 kDa MBP kinase, in a manner similar to butyric acid at the same concentrations ( Fig. 6a).

We have analysed the effect of the protein kinase inhibitor staurosporine on the activation of the MBP kinase by 2,4-D at 300 μm or by butyric acid at 2 m m ( Fig. 7). When staurosporine was added at the concentration of 5 μm, the activation of MBP kinase by butyric acid was modified very little, whereas the effect of 2,4-D was reduced by half. This latter result was only obtained with concentrations of staurosporine above 1 μm.

Figure 7. Effect of staurosporine on MBP kinase activation.

The cells were treated with butyric acid at 2 m m during 6 min or with 2,4-D at 300 μm during 15 min. The protein kinase inhibitor staurosporine (5 μm) was eventually added to the cell suspensions 9 min before butyric acid or at the same time as 2,4-D. Soluble protein extracts were prepared and MBP phosphorylation was analysed by in-gel kinase assay. The mass of the MBP kinase is 47 kDa.

Discussion

Several plant MAP kinases, which share conserved features with mammalian ERKs and yeast MAP kinases, recently demonstrated to be involved in the transduction pathways of many stress-related signals ( Bogre et al. 1997 ;Jonak et al. 1996 ;Knetsch et al. 1996 ;Ligterink et al. 1997 ;Seo et al. 1995 ;Suzuki & Shinshi 1995;Usami et al. 1995 ;Zhang & Klessig 1997;Zhang et al. 1998 ). Curiously, given the typical involvement of MAP kinases in response to proliferative signals in animal cells, few studies have analysed MAP kinase activation by such signals in plants. In this paper, we show that auxin, a growth factor intimately linked to the control of cell division in plants, does not activate an MAP kinase at physiological concentrations which induce mitoses in auxin-starved non-dividing tobacco cells. This result is not the consequence of technical inadequacy or of a constitutive lack of this activity in the tobacco cells since an MBP kinase could be activated in our system by very high, non-physiological, auxin concentrations, by butyric and acetic acids as well as by plasmolysis with 0.8 m sorbitol (data not shown). Thus, our results oppose a previous study with auxin-starved tobacco BY-2 cells, in which a MAPKK and an MBP kinase were stimulated 5 and 10 min, respectively after treatment by 0.9 μm 2,4-D ( Mizoguchi et al. 1994 ).

The lack of auxin-activated MBP kinase may be explained by only a fraction of the cells actually responding to auxin, making the related MBP kinase signal too weak (diluted) to be detected. Several points contest this hypothesis. First, when extracts with a high MBP kinase activity, such as those of cells treated by 300 μm 2,4-D, were diluted to various extent by extracts from untreated cells, a linear relationship betwen MBP kinase activity and the dilution factor was observed (data not shown). From these data, it was estimated that the in-gel kinase assay could detect MBP kinase activation in a heterologous cell population in which only 10% of the entire population contain active MBP kinase. This value was clearly lower than the actual fraction of the cells responding to auxin. Auxin addition triggered a rise of the mitotic index to values equal or higher than those recorded during the exponential growth phase of a standard culture cycle, where most of the cells divide ( Fig. 1). Moreover, an increase in the relative abundance of several auxin-responsive (NtIAA4.3, Nt103–1) and cell-cycle associated (CycA3;2, ArcA) mRNAs was clearly detected by hybridisation experiments ( Fig. 2), including weakly expressed mRNAs such as NtIAA4.3 and CycA3;2. Interestingly, the increase of ArcA was similar to the increase previously described in the same system in response to physiological concentrations of auxin ( Ishida et al. 1993 ). Taken together, these data indicate that a large fraction of the cells responded to auxin, which ruled out any over-dilution of auxin-induced MBP kinase in the total extract.

Recent results clearly illustrate the responsiveness of plant MAP kinases to mechanosensing and osmotic stress ( Bogre et al. 1996 ;Hirt 1997). Unknown parameters related to cell manipulation may activate MAP kinases ( Suzuki & Shinshi 1995). To avoid this phenomenon, we ensured a minimum perturbation of the cell suspensions at the level of gaseous environment, solvent addition, and change in pH, for the addition of all the studied compounds. The previous report of MAPKK and MBP kinase induction by auxin in a similar system ( Mizoguchi et al. 1994 ) actually reflects the activation of a MAP kinase cascade but this was most likely due to another parameter of the method used rather than to auxin itself. We show for example that a 47 kDa MBP kinase may be activated when the cells are subcultured in a fresh medium, irrespective of the presence or not of auxin ( Fig. 4d). It is concluded that the induction of mitosis by physiological concentrations of auxin does not require the activation of a MAP kinase with MBP kinase property. However, the possibility that auxin activates a MAP kinase devoid of MBP kinase activity or unrelated to mammalian ERK1/2 remains open.

In the second part of this study, we have shown that high concentrations of lipophilic weak acids rapidly stimulate an MBP kinase activity with the constitutive hallmarks of a typical MAP kinase ( Figs 3, 4 and 6): (i) it comigrated in SDS–PAGE with a 47 kDa polypeptide which was recognised by a polyclonal antibody raised against animal ERK1 and which was present at a constant amount throughout the experiments; (ii) it was activated concurrently with the catalytically activated form of an ERK1/2-related polypeptide detected by a specific antibody; (iii) it did not depend on calcium, and was much more efficient with MBP as a substrate than with histone H1, casein or autophosphorylation; (iv) it was activated transiently in the cells, reaching its highest value 6–15 min after the treatment. The stimulation of an MAP kinase by 2,4-D at a high concentration of 300 μm was not related to a specific effect of auxin, since 2,3-D, an inactive auxin analogue of 2,4-D, was as efficient in inducing MAP kinase ( Fig. 4b). In addition, it was not related to the toxic effect of high auxin concentrations as toxic effects were observed from the concentration of 90 μm of 2,4-D and 2,3-D, whereas cytosolic acidification and MAP kinase activation were detected only from a concentration of 300 μm (compare Figs 4b and 5a–b). Moreover, MAP kinase could also be induced by non-toxic concentrations of butyric or acetic acids ( Fig. 6), lipophilic acids without specific biological activity in plant cells.

Several lines of evidence show that the activation of an MAP kinase by high concentrations of auxin and other lipophilic acids was mediated by the significant cytosolic acidification which they produce: (i) cytosolic acidification due to the loading of the cytosol by these molecules was detected shortly after the addition of the acids, but prior to MAP kinase activation ( Fig. 5). Acidification and MAP kinase activation were more rapidly detected after a butyric acid treatment in comparison to auxin treatment, probably because of the higher permeability coefficient of butyric acid; (ii) the various compounds were effective at the same concentrations both to activate MAP kinase and to acidify significantly the cytosol; (iii) the concentration of 2,4-D required to activate MAP kinase and to acidify the cytosol, i.e. 300 μm, was lower than the equivalent for butyric or acetic acid, i.e. 5 m m. This was in agreement with the fact that 2,4-D and 2,3-D are stronger acids (pKa = 2.8) than butyric or acetic acids (pKa = 4.8); (iv) at least two different weak acids, acetic and butyric acids, were used to trigger acid loads in the cytosol; these two compounds were able to activate MAP kinases, ruling out an effect of these molecules different than pH change.

These results indicate that cytosolic acidification is a sufficient trigger of MAP kinase activation in the tobacco cells. The method used to calculate the cytosolic pH could not show the maximum extent of the acidification, since each measurement was done every 2.5 min, but the results were similar to those shown by ³¹NMR measurement of cytosolic pH in plant cells during acid load ( Guern et al. 1986 ). Notably the biphasic response of the cytosolic pH after butyric acid loading was indicative of regulatory mechanisms, thus contributing to pH homeostasis. It was estimated from Fig. 5 that MAP kinase was activated when the cytosolic pH was decreased by more than ≈ 0.4 pH units. Although such variations are unlikely to occur frequently in plant cells, they may be physiologically relevant in the context of some stress such as anoxia or elicitor treatment ( Felle 1996;Kurkdjian & Guern 1989;Mathieu et al. 1996 ). The pathway between cytosolic acidification and MAP kinase activation remains unknown. From the effect of butyric acid, no staurosporine-sensitive protein kinase is involved in this pathway ( Fig. 7). 2,4-D is known to enter the cells by combined diffusion and transport ( Delbarre et al. 1996 ). The partial inhibition by staurosporine of MAP kinase activation by 300 μm 2,4-D ( Fig. 7) may be explained by a reduction of 2,4-D influx because of a phosphorylation-dependent step during its transport into the cells.

Five MAP kinases have been cloned in tobacco ( Seo et al. 1995 ;Wilson et al. 1995 ;Wilson et al. 1993 ;Zhang & Klessig 1997), and results from Arabidopsis thalianaMizoguchi et al. 1993 ) suggest that this number is far from being complete. We cannot assess whether one or more tobacco MAP kinases are activated by cytosolic acidification. Several reports have shown that tobacco MAP kinases with the same mass are activated in response to elicitors ( Suzuki & Shinshi 1995;Zhang et al. 1998 ), salicylic acid ( Zhang & Klessig 1997) and wounding ( Seo et al. 1995 ;Usami et al. 1995 ). The role of cytosolic acidification in these responses should therefore be analysed. Henceforth, our results provide a causal link between two reports that fungal elicitors induce both cytosolic acidification by 0.5–0.8 pH units ( Mathieu et al. 1996 ) and MAP kinase activation ( Suzuki & Shinshi 1995;Zhang et al. 1998 ) in tobacco cells. In animal PC12 cells, a direct intracellular acidification by propionic acid is sufficient to activate p42-p44 MAP kinases and to produce thereafter mitogenesis ( Thomas et al. 1996 ). The finding that MAP kinase may be activated in plant cells by cytosolic acidification in the absence of a mitogenic effect of the acidifying treatment points both to similarities and differences between the two kingdoms. The highly regulated cytosolic pH of eukaryotic cells ( Madshus 1988) may be a second messenger preceding the MAP kinase cascades, but the plant mitogenic signal auxin and animal mitogenic signals are probably transduced by different mechanisms.

Experimental procedures

Cell culture and treatment

The tobacco BY-2 cell line (Nicotiana tabacum L. cv Bright Yellow 2) was cultured in the dark at 26°C with 0.9 μm 2,4-dichlorophenoxyacetic acid (2,4-D) as the exogenous auxin required for growth ( Nagata et al. 1992 ). For auxin starvation, stationary phase cells (7 or 8-day old) were subcultured with a 1/60 dilution in a fresh medium containing 0.05 μm 2,4-D for 4 days. To study MAP kinase activation various compounds, 2,4-D, 2,3-dichlorophenoxyacetic acid (2,3-D), butyric or acetic acids were added to 4-day old auxin-starved cells by making sure to minimise the perturbation linked to the treatment: the compounds were solubilized in aqueous solutions of pH between 6 and 7; they were added through the foil capping the flask, while stopping agitation for only 10 sec. In control treatments, a mock solution of similar pH and ion content (KCl) was added. The mitotic index was measured by mixing 1 ml suspension with 10% formaldehyde (v/v), 1% Triton X-100 (v/v) and 1 μg ml^–1 diaminophenylindole. In each sample the mitoses were counted by fluorescence microscopy in a total number of 1000 nuclei.

Extraction of soluble proteins

The cells were separated from the medium on a fritted glass filter, frozen in liquid nitrogen, and stored at –80°C. For protein extraction, liquid nitrogen-ground cells were mixed in one volume of 100 m m KCl, 10 m m MgCl₂, 10 m m dithiothreitol, 2 m m EGTA, 20 m mβ-glycerophosphate, 10 m mp-nitrophenylphosphate, 10 m m NaF, 1 m m orthovanadate, 1 m m phenylmethylsulfonyl fluoride, 1 μm leupeptin, 5 m m ascorbic acid and 20 m m Tris–HCl, pH 7.4. The extract was centrifuged at 200 000 g for 30 min. The supernatant, referred to as the soluble extract, was stored at –80°C. Protein concentration was determined using the Bradford assay ( Bradford 1976) with bovine albumin as a reference.

Measurement of MBP kinase activity

In-gel kinase assays were performed according to Gotoh et al. (1990) and Suzuki & Shinshi (1995). The proteins (50 μg) from soluble extracts were separated by SDS–PAGE in 10% gels polymerized in the presence of 0.5 mg ml^–1 myelin basic protein (MBP, Sigma). After electrophoresis, the gels were washed 2 × 1 h in 20% 2-propanol, 5 m mβ-mercaptoethanol, 50 m m Tris–HCl pH 8.0 and 1 × 1 h in 5 m mβ-mercaptoethanol, 50 m m Tris–HCl pH 8.0. The proteins were denatured 2 × 30 min at room temperature in 6 m guanidine, 5 m mβ-mercaptoethanol, 50 m m Tris–HCl pH 8. They were renatured at 4°C by incubating the gel for 3 × 6 h in 0.05% Tween 20, 5 m mβ-mercaptoethanol, 50 m m Tris–HCl pH 8.0. To reveal kinase activities, the gels were incubated for 30 min in 100 μm EGTA, 20 m m MgCl₂, 2 m m dithiothreitol, 40 m m Hepes-KOH pH 7.5, and then in the same buffer containing 25 μmγ-[³²P]-ATP (7.4 GBq.mmol^–1, ICN). The reaction was stopped by adding 5% trichloracetic acid and 1% sodium pyrophosphate. The gels were washed extensively with this solution before drying and autoradiography. All the experiments have been reproduced at least three times.

Immunodetection of MAP kinase

The proteins from soluble extracts (20–100 μg) were separated by SDS–PAGE (10% gel), and transferred to 0.2 μm nitrocellulose (Sartorius) in 10% methanol, 10 m m 3-[cyclohexylamino]-1-propanesulfonic acid (CAPS), pH 11, at 35 V during 16 h. The blots were probed with polyclonal antibodies raised against the conserved subdomain XI of rat ERK1, which recognises ERK1 and ERK2 (Santa Cruz Biotechnology), or against the phosphotyrosine-containing peptide of human p44 MAP kinase (residues 196–209, New England Biolabs). The latter antibody recognised only ERK1/2 related polypeptides which are catalytically activated at Tyr-204. The immune complexes were detected by a peroxidase-bound antirabbit IgG antibody (Sigma) and revealed by luminescence (ECL, Amersham). All antibodies were used at 1/1000 dilution.

Estimation of cytosolic pH

The cytosolic pH was measured from the accumulation in the cells of a weak lipophilic acid, [¹⁴C]-benzoic acid ( Mathieu et al. 1996 ). Four ml of auxin-starved cells (50–100 mg fresh weight) were equilibrated in an open flask for 1 h with 50 m m Mes-Tris pH 5.2. During various treatments, 2 nmol [¹⁴C]-benzoic acid (2.24 Gbq.mmol^–1, ICN) were added to the suspension. After 2.5 min the cells were collected on GF/A glass fiber filters (Whatman), and weighed. The intracellular radiolabelled molecules were extracted in 300 μl methanol for 30 min and the total intracellular radioactivity was counted by liquid scintillation. The actual amount of intracellular free benzoic acid was measured after fractionation of the cellular extract by thin layer chromatography on a silica plate, with the eluent chloroform:methanol: acetic acid (75:20:5 v/v/v). The radioactive free benzoic acid was located by autoradiography (R_f = 0.82), and measured by liquid scintillation. Estimates of the cytosolic pH were calculated from the ratio of the intracellular versus extracellular concentrations of benzoic acid as described ( Mathieu et al. 1996 ).

mRNA analysis

Total RNAs were isolated from the cells by phenol–SDS extraction ( Tournaire et al. 1996 ), electrophoresed on formaldehyde/agarose gels ( Sambrook et al. 1989 ) and transferred to Biotrans(+) membranes (ICN). Hybridisation with radiolabelled probes was performed under standard high stringency conditions ( Tournaire et al. 1996 ). The probes were complete tobacco cDNAs for the CycA3;2 cyclin ( Reichheld et al. 1996 ) and for the auxin-responsive genes ArcA ( Ishida et al. 1993 ), NtIAA4.3 (A. Dargeviciute, F. Sitbon and C. Perrot-Rechenmann, unpublished) and Nt103–1 gluthathione transferase ( van der Zaal et al. 1991 ). A partial cDNA for tobacco 25 S rRNA (NT7, 191 bp, J.P. Renaudin, unpublished) was also used as a probe to normalise the hybridisation signals by quantification with a PhosphorImager (Molecular Dynamics).

Acknowledgements

This work was supported by the National Institute for Agronomic Research (INRA) and the National Centre for Scientific Research (CNRS), and by a grant from the French Ministry of Education to G.T. We thank Dr C. Perrot-Rechenmann for the gift of the plasmids containing the tobacco cDNAs ArcA and NtIAA4.3. as well as Dr E.J. van der Zaal for the gift of the plasmid containing the tobacco cDNA Nt103–1. We acknowledge the suggestions of Professor J. Guern and the critical reading of the manuscript by Drs J.F. Briat, J.C. Davidian and J. Vidmar. We thank Mrs C. Charlery for revising the English language.