A dominant-negative form of Arabidopsis AP-3 β-adaptin improves intracellular pH homeostasis


For correspondence (e-mail rserrano@ibmcp.upv.es).


Intracellular pH (pHi) is a crucial parameter in cellular physiology but its mechanisms of homeostasis are only partially understood. To uncover novel roles and participants of the pHi regulatory system, we have screened an Arabidopsis mutant collection for resistance of seed germination to intracellular acidification induced by weak organic acids (acetic, propionic, sorbic). The phenotypes of one identified mutant, weak acid-tolerant 1-1D (wat1-1D) are due to the expression of a truncated form of AP-3 β-adaptin (encoded by the PAT2 gene) that behaves as a as dominant-negative. During acetic acid treatment the root epidermal cells of the mutant maintain a higher pHi and a more depolarized plasma membrane electrical potential than wild-type cells. Additional phenotypes of wat1-1D roots include increased rates of acetate efflux, K+ uptake and H+ efflux, the latter reflecting the in vivo activity of the plasma membrane H+-ATPase. The in vitro activity of the enzyme was not increased but, as the H+-ATPase is electrogenic, the increased ion permeability would allow a higher rate of H+ efflux. The AP-3 adaptor complex is involved in traffic from Golgi to vacuoles but its function in plants is not much known. The phenotypes of the wat1-1D mutant can be explained if loss of function of the AP-3 β-adaptin causes activation of channels or transporters for organic anions (acetate) and for K+ at the plasma membrane, perhaps through miss-localization of tonoplast proteins. This suggests a role of this adaptin in trafficking of ion channels or transporters to the tonoplast.


The regulation of intracellular pH (pHi) is a fundamental activity of living cells because this parameter plays important permissive and regulatory roles in many cellular activities. Most protein machineries (enzymes, motors, transporters, vesicle traffic, ribosomes, spliceosomes, assembly proteins, regulators, etc.) function within a narrow range of pH and, on the other hand, the pH dependence of proteins paves the way for a regulatory role of pHi (Felle, 2001; Parks et al., 2011; Hueso et al., 2012; and references therein).

The physiological mechanisms of pHi homeostasis have been investigated in bacteria (Krulwich et al., 2011), yeast (Mira et al., 2010; Hueso et al., 2012), mammalian cells (Parks et al., 2011) and plants (Raven, 1985; Davies, 1986; Gout et al., 1992; Iuchi et al., 2007; Bissoli et al., 2012). Intracellular acidification may be generated by H+ uptake from the external medium, as during heat shock (Weitzel et al., 1987) and other stresses including acid media, or by metabolic generation of organic acids during fermentations, especially under hypoxic conditions, a process suggested to have driven the evolution of proton pumps in primitive cells (Raven and Smith, 1976). The general picture is that H+ efflux systems are H+-ATPases of plasma and vacuolar membranes that constitute the ‘biophysical pH-stat’ (Felle, 2001) while carboxylation and decarboxylation reactions (H+-generating and H+-consuming, respectively) define the ‘biochemical pH-stat’ (Davies, 1986). Together with cytosolic buffer systems (bicarbonate, phosphate and the imidazol group of histidine; Hochachka and Somero, 2002), these are supposed to be the basic players of pHi homeostasis.

The relative importance and the regulation of the different systems involved in the homeostasis of pHi in plant cells has started to be investigated at the molecular and genomic levels in Arabidopsis thaliana. Mutants with gain (Young et al., 1998) and loss (Haruta et al., 2010) of function of plasma membrane H+-ATPase have improved and reduced growth, respectively, at low external pH. A loss-of-function approach has identified the zinc finger transcription factor STOP1 as a positive regulator of many genes involved in tolerance to external acid pH and to aluminium, two traits that are genetically linked (Iuchi et al., 2007; Sawaki et al., 2009). The regulated genes encode membrane transporters and metabolic enzymes such as a malate transporter (ALMT1), an activator of AKT1 K+ channel (protein kinase CIPK23; Xu et al., 2006), GABA shunt enzymes (GDH1, GDH2, GAD1, GABA-T) and malic enzyme (ME1, ME2). Reverse genetics will be needed to test the relevance of these genes to pHi homeostasis.

In these previous works intracellular acidification was induced by lowering external pH. As external acidification solubilizes aluminium and this cation has specific toxic effects counteracted by secretion of organic anions, some of these cellular responses may be unrelated to intracellular acidification. A more direct method to induce intracellular acidification is the addition of weak organic acids, that diffuse in an undissociated form into the cells and release H+ upon dissociation in the cytosol (Krebs et al., 1983). Genes induced by this form of intracellular acidification (Bissoli et al., 2012) are different from those induced by acidification of the external medium (Lager et al., 2010), pointing to significant differences between these two pH stresses. Over-expression of prolyl cis–trans isomerase ROF2, whose expression is induced by intracellular acidification generated by acetic acid, improves germination in the presence of this acid, activates K+ uptake, decreases membrane potential and activates the efflux of protons mediated by the plasma membrane H+-ATPase (Bissoli et al., 2012).

One approach to identify novel components of pHi homeostasis consist in the identification of genes that upon gain or loss of function improve tolerance to intracellular acidification generated by weak organic acids (Mira et al., 2010; Hueso et al., 2012). We have screened a collection of ‘activation-tagging’ Arabidopsis thaliana mutants (Weigel et al., 2000) for improved resistance to acetic acid during seed germination and report here that a truncated form of AP-3 β-adaptin (encoded by the PAT2 gene; Feraru et al., 2010) behaviors as dominant-negative to improve pHi homeostasis. This β-adaptin is a component of the AP-3 adaptor complex involved in a pathway (the so-called ‘ALP’ pathway) of direct vesicle traffic from the Golgi to the vacuolar compartment (Boehm and Bonifacino, 2002; Feraru et al., 2010; Wolfenstetter et al., 2012). The acetic acid tolerance of this mutant is based on increased rates of K+ uptake and acetate efflux, which by providing electrical balance increase H+ efflux mediated by the electrogenic plasma membrane H+-ATPase. These results suggest a role of this adaptin in trafficking of ion channels or transporters to the tonoplast.

Results and Discussion

Isolation of a mutant altered in pH homeostasis

We have developed an assay to identify mutants with improved pHi homeostasis based on resistance to intracellular acidification generated by acetic acid during germination and seedling establishment. It was interesting to know if Arabidopsis mutants probably altered in pHi homeostasis score in our acetic acid-germination assay. We have tested, mutants with loss of function in the major plasma membrane K+ channel of roots (akt1-1; Hirsch et al., 1998), in the tonoplast dicarboxylate (malate) transporter (tdt; Hurth et al., 2005), in the vacuolar H+-ATPase (vha-a2-1 vha-a3-1 double mutant; Krebs et al., 2010) and in the vacuolar H+-Na+/K+ antiporter (nhx2-1; Barragán et al., 2012). The tdt mutant is more sensitive to low external pH (a condition that lowers pHi) than wild type (Hurth et al., 2005). No data on sensitivity to low external pH existed for the other Arabidopsis mutants but the yeast mutants on the homologous genes were defective in pHi homeostasis. The growth of the yeast null mutants in the plasma membrane K+ uptake system Trk1 Trk2 (Mulet et al., 1999), in the vacuolar H+-ATPase (Nelson and Nelson, 1990) and in the vacuolar H+-Na+/K+ antiporter Nxh1 (Brett et al., 2005) is sensitive to low external pH and, in the case of nhx1Δ mutant, the measured pHi was lower than in wild type (Brett et al., 2005). As indicated in Figure 1, all four Arabidopsis mutants scored as more sensitive to acetic acid than wild type. In addition we have also tested transgenic plants with gain of function by over-expression of AKT1 (Xu et al., 2006) or of the tonoplast H+-PPase (AVP1; Gaxiola et al., 2001) and both scored as more tolerant to acetic acid than control plants (Figure 1). Together with the previous work with ROF2 (Bissoli et al., 2012), these results validate the acetic acid-germination assay as a useful tool to screen for mutants in pHi homeostasis.

Figure 1.

Germination in acetic acid plates of Arabidopsis mutants in plasma membrane and vacuolar ion transporters. The percentage of green cotyledons after 6 days in plates with 3.5 mm acetic acid was recorded for wild type (wt) and for the different mutants and relative values (percent green cotyledon in mutant divided by percent green cotyledons in wild type) are given. Wild-type values in different experiments were in the range of 40–60% of planted seeds. In the absence of acetic acid 95–100% of planted seeds had green cotyledons in all cases. OE corresponds to over-expression of the genes. In the akt1-1 and OE AKT1 mutants the K+ concentration in the medium was reduced to 0.1 mm. Error bars correspond to the standard error of three independent experiments.

We have screened 30 000 lines of an ‘activation-tagging’ mutant seed collection (about 400 000 seeds) looking for resistance to acetic acid during germination and seedling establishment. Five confirmed mutants were analyzed for the presence of T-DNA by Southern blot analysis. One contained no T-DNA, two had multiple (three or four) insertions and two had a unique T-DNA insertion (Figure S1a). The last two were found to correspond to the same mutation and this twice-isolated mutant was selected for further study.

Homozygous mutant plants were crossed with Columbia wild-type and F2 seeds germinated in acetic acid with a 3:1 ratio (resistant:sensitive), as expected if the mutation were monogenic and dominant. Actual data were 143:57 (χ2 = 1.42; = 0.95). F2 acetic acid-tolerant plants were selected and scored for the presence of the T-DNA by PCR analysis. Within 33 F2 plants analyzed, all of them contained the T-DNA, indicating that the insertion was linked to the observed phenotype (Figure S1b). This mutant was named wat1-1D (from weak acid-tolerant, gene 1, allele 1, Dominant), according to the recommendations of Meinke and Koornneef (1997).

In the presence of different weak acids (acetic, propionic, sorbic) wat1-1D is able to germinate and open its cotyledons faster than wild type (Figure 2). This situation indicates that the resistance of the mutant is not specific for the acetic acid utilized in the screening, but that wat1-1D is more tolerant to the decrease in cytosolic pH caused by different weak acids. Under the conditions of our experiments (medium pH buffered at pH 5.5, acetic acid concentrations 3–5 mm) the organic acid does not induce death of the germinating seeds and it just delays germination and the appearance of green cotyledons by about 5 days with respect to medium without acid.

Figure 2.

The wat1-1D mutant is more resistant to weak organic acids during germination and seedling establishment. (a) Representative seedlings of Arabidopsis wild type (wt) and wat1-1D mutant 5 days after sowing in Murashige and Skoog (MS) medium (pH 5.5) supplemented with 3.5 mm acetic acid, 1 mm propionic acid or 0.55 mm sorbic acid. Scale bar corresponds to 2 mm. (b) Statistical data from three independent experiments such as the one shown in (a). Percentage of seedlings with green, expanded cotyledons and standard error (error bars) are given. **P < 0.01 by Student's t-test.

We have measured pHi of epidermal root cells in 15-day-old seedlings (Figure 3) and found that in normal medium both wat1-1D and wild-type seedlings have similar values (7.22 and 7.28, respectively) but after treatment with acetic acid for 5 min pHi drops much less in the mutant than in wild type (6.94 and 6.54, respectively). Therefore, the screening using acetic acid during the first developmental stages of germinating seeds has led to the isolation of a ‘bona fide’ mutant with improved capability for pHi homeostasis.

Figure 3.

The wat1-1D mutant maintains a higher pHi during acetic acid treatment. (a) Records of cytosolic pH of a root epidermal cell of wild type and wat1-1D mutant (15-day-old plants) during an acetic acid perfusion experiment. Arrows indicate the addition of 1 mm of the acid to the medium and the removal of the acid by wash. (b) Values correspond to 5 min after addition of acetic acid or the equivalent time when no addition was made. Mean values (±standard error) of cytosolic pH from four different perfusion experiments such as the one shown in (a).

In normal conditions wat1-1D has no obvious morphological phenotypes, although a more detailed analysis indicated a subtle growth defect (Figure S2).

wat1-1D is a dominant-negative mutant that over-express a truncated AP-3 β-adaptin encoded by the PAT2 gene

We performed plasmid rescue to localize the T-DNA insertion in the genome of the wat1-1D mutant and, as shown in Figure 4(a), it was inserted in an intron of the At3g55480 gene. The 35S transcriptional enhancers found in the T-DNA were 4 kb away from the promoter of the adjacent gene, At3g55470, and northern blot analysis showed that this gene was over-expressed in the mutant (Figure 4b). To test if this over-expression conferred weak acid resistance, we analyzed the phenotype of knock-out mutants and transgenic plants over-expressing At3g55470 cDNA. To our disappointment, the over-expression of At3g55470 in transgenic plants did not confer acetic acid resistance and loss-of-function mutants did not show sensitivity to the acid (Figure S3).

Figure 4.

Position of the T-DNA insertion in the wat1-1D mutant and expression analysis of adjacent genes. (a) Location of the T-DNA in the 7th intron or 8th intron of the At3g55480 gene (splice variants 1 and 2 respectively), as determined by plasmid rescue. The distance of the 35S enhancers to the adjacent promoters is shown. This gene has been named PAT2 (Feraru et al., 2010). (b) Northern blot analysis of At3g55470 (left) and At3g55480 (right) mRNAs from wild type (wt) and wat1-1D mutant. Lower panels correspond to ethidium bromide staining of major ribosomal RNAs (rRNA) as loading controls. In the case of At3g55480, the regions of hybridization for probe 1 and 2 are represented in (a). There are two bands in the wild type because the abundance of the 3.5 kb major rRNA splits the hybridization of the probe at the same position of the blot (the At3g55480 has also 3.5 kb).

A more detailed expression analysis of the T-DNA adjacent genes indicated that the wat1-1D mutant expressed a truncated form of the At3g55480 gene consisting of the first 1.9 kb (Figure 4b). Interestingly, transgenic plants that over-express this truncated form of the gene (Figure S4) reproduced the acetic acid phenotype of the wat1-1D mutant (Figure 5a,b). The expression of a truncated gene can cause gain of function if an inhibitory domain is eliminated or loss of function if the resulting protein is not functional. To check if wat1-1D phenotypes were due to either loss or gain of function of At3g55480, we obtained transgenic plants over-expressing this gene in antisense. The acetic acid phenotype of these transgenic plants was very similar to that of the wat1-1D mutant (Figure 5c,d), indicating that it is the loss of function of the At3g55480 gene that is responsible of the phenotype and that wat1-1D is a dominant-negative mutant. In Figure 5 it is shown a secondary phenotype of the wat1-1D mutant in the germination assay: sensitivity to lithium. This phenotype is also reproduced by expression of either a truncated form or an antisense of the At3g55480 gene (Figure 5a,c).

Figure 5.

Phenotypes of the wat1-1D mutant are due to the loss of function of the At3g55480 gene. (a) Percentage of seedlings with green, developed cotyledons in the wild type, wat1-1D, and two transgenic plants over-expressing the truncated form of the At3g55480 gene (T3(1)trunc and T3(2)trunc) after 5 or 6 days in Murashige and Skoog (MS) medium supplemented with 3.5 mm acetic acid or 23 mm LiCl as indicated. (b) Aspect of seedlings of wild type (wt), wat1-1D, and three transgenic plants over-expressing the truncated At3g55480 gene, after 5 days in MS medium (pH 5.5) supplemented with 3.5 mm acetic acid. (c) Percentage of seedlings with green, developed cotyledons in the wild type, wat1-1D, and two transgenic plants expressing the At3g55480 gene in antisense (T3(1)anti and T3(2)anti) after 5 or 6 days in MS medium supplemented with 3.5 mm acetic acid or 23 mm LiCl. (d) Seedlings of wild type (wt), wat1-1D, and three transgenic plants over-expressing the antisense of At3g55480 gene after 5 days in MS medium (pH 5.5) supplemented with 3.5 mm acetic acid. Scale bar corresponds to 1 cm. Errors bars represent standard error of three independent experiments. *P < 0.05 and **P < 0.01 by Student's t-test.

Recently, an Arabidopsis loss-of-function mutant in At3g55480 was characterized and the gene named PAT2 (Feraru et al., 2010). The authors describe that pat2 loss-of-function mutant has no apparent phenotype in normal conditions with the exception of a defect in shoot gravitropic response. However, in medium that lacked sucrose, the percentage of arrested plants is increased in the mutant compared with the wild type. Although molecular mechanisms are unknown, these phenotypes are reminiscent of mutants affected in vacuolar function, trafficking to vacuoles and post-Golgi membrane trafficking (Kato et al., 2002; Silady et al., 2008). To further demonstrate that the phenotypes of wat1-1D are due to the loss of function of PAT2, we analyzed if wat1-1D shares these features with pat2 mutant. As expected, the establishment of non-arrested seedlings in medium without sucrose is decreased in wat1-1D (Figure S5a,b). Furthermore, as indicated in Figure S5(c,d), the shoot gravitropic response of 90°C-stimulated seedlings is reduced in wat1-1D, as was found for pat2. Taken together, these experiments demonstrate that wat1-1D presents a loss of function of the PAT2 gene and that this mutation confers tolerance to weak organic acids and improves pHi homeostasis.

PAT2 encodes the beta subunit of the adaptor protein complex 3 (β-adaptin 3) (Boehm and Bonifacino, 2002). These complexes are heterotetramers involved in vesicle formation in different endomembrane compartments. The adaptor protein complex 3 (AP-3 complex) is conserved from yeast to animal and plant cells. In yeast this complex participates in protein trafficking from the Golgi to the vacuolar membrane through the ALP (alkaline phosphatase) pathway (Cowles et al., 1997; Piper et al., 1997) and loss of function of different subunits can produce the accumulation of vacuolar alkaline phosphatase in the cytosol (Cowles et al., 1997; Stepp et al., 1997) or the mis-localization of the Yck3 tonoplast protein to the plasma membrane (Sun et al., 2004). In animal cells this complex is involved in protein transport to lysosomes and lysosome-related organelles, and mutations in the human AP-3 complex cause a genetic disorder named Hermansky Syndrome type 2 (Ohno, 2006), in which several lysosomal membrane proteins such as LAMP I, LAMP II and CD63 are delocalized to the plasma membrane (Dell'Angelica et al., 1999). There is growing evidence that in Arabidopsis the AP-3 complex is involved in protein trafficking to the lytic or storage vacuole (Sohn et al., 2007; Niihama et al., 2009) and it has recently been reported that the β-adaptin 3 plays an important role in vacuolar biogenesis and in the transition from storage to lytic vacuoles (Feraru et al., 2010). In yeast lithium sensitivity is typical of mutations in components of ESCRT (Endosomal Sorting Complex Required for Transport) and it can be explained by altered vesicle traffic of some ion transporters for this toxic cation (Bowers et al., 2004). The same mechanism may explain lithium sensitivity in the wat1-1D mutant.

The dominant-negative character of the wat1-1D mutation can result from competition between wild-type and truncated β-adaptin 3 for integration into the AP-3 complex, with over-expression of the latter displacing wild-type adaptin and giving rise to a non-functional complex. Conversely, the improved pHi homeostasis of the wat1-1D mutant in the presence of acetic acid could result from an increased capability for proton extrusion resulting from loss of function of the AP-3 complex. This possibility was addressed in the following experiments.

Proton extrusion by roots in the presence of acetic acid is more efficient in wat1-1D

In order to investigate the mechanism utilized by the wat1-1D mutant to maintain a higher pHi in the presence of acetic acid, we first tested if acetic acid uptake was reduced in the mutant. In yeast, acetic acid uptake occurs through the plasma membrane porin Fps1. This channel is negatively regulated by Hog1 MAP kinase and mutations in these systems result in acetic acid tolerance (Mollapour and Piper, 2007). As indicated in Figure S6, in the case of the Arabidopsis wat1-1D mutant the uptake of acetic acid at short time periods (up to 20 min) is similar to that of wild-type plants. Therefore, the mechanism of tolerance of wat1-1D to acetic acid is not based on reduced uptake of this weak organic acid.

We have also discarded the implication of the ‘biochemical pH-stat’ (Davies, 1986) in the phenotype of the mutant, as the levels of malate and the initial rate of malate degradation of wild-type and wat1-1D seedlings in the presence of acetic acid are similar (Figure S7). As indicated in this figure, the degradation of malate during intracellular acid stress stopped when only 25% of the initial malate was consumed, suggesting that in our experimental system malate decarboxylation offers a limited protection against intracellular acidification.

We next analyzed the activity of the plasma membrane H+-ATPase (PMA), an important component of the ‘biophysical pH-stat’ that could be activated in the wat1-1D mutant. Contrary to our expectations, measurements of the in vitro activity of the enzyme (ATP hydrolysis by purified membranes from roots) indicated that the mutant has slightly lower activity (16% less) than wild type (Figure 6a). It is also shown in the figure that growing wild-type plants in the presence of acetic acid increased the activity of the enzyme by 15%. This is in accordance with the report that intracellular acidification triggered by weak organic acids increases the phosphorylation of the tobacco PMA2 penultimate threonine, activating the pump (Bobik et al., 2010). Acetic acid had no effect on PMA activity in the wat1-1D mutant probably because intracellular acidification is much less than in wild type (Figure 3).

Figure 6.

Activity of plasma membrane H+-ATPase in wild type and wat1-1D mutant. (a) ATP hydrolysis by purified plasma membranes (in vitro activity) of wild type (wt) and wat1-1D. Membranes were purified from roots of plants grown with normal nutrient solution or supplemented with 3.5 mm acetic acid. Mean values and standard errors of six experiments are given. The differences between wild type and wat1-1D were significant by Student's t-test (P < 0.01). (b) Acidification of the external medium by wild type (triangles) and wat1-1D mutant (squares) plants in normal conditions (open symbols) or in the presence of 1 mm acetic acid (closed symbols). The mean of four experiments is shown. Standard errors of the points were in the range 0.4–1.1 nmoles H+ × min−1 × mg fw−1 (fw = fresh weight). The differences between wild type and wat1-1D at the three time points in the presence of acetic acid were significant by Student's t-test (P < 0.05).

The activity of PMA in vitro only shows changes in the activity of the enzyme due to covalent modifications such as phosphorylation. The in vivo activity of PMA, estimated by proton extrusion from roots, is also modulated by physicochemical parameters such as intracellular pH or membrane electrical potential (Serrano, 1985; Michelet and Boutry, 1995). As shown in Figure 6(b), in medium lacking acetic acid root external acidification in the wat1-1D mutant is similar to that of the wild type. However, the addition of acetic acid triggers in both genotypes an increase in proton pumping that is higher in the case of the mutant (Figure 6b). These results indicate that wat1-1D is able to maintain a higher cytosolic pH in the presence of acetic acid probably because it extrudes protons more efficiently than the wild type under these conditions.

The increased proton extrusion of wat1-1D is due to membrane depolarization and increased K+ uptake and acetate efflux

Considering that in the presence of acetic acid proton extrusion is increased in the wat1-1D mutant, it was expected that wat1-1D would have a higher PMA activity and therefore a higher plasma membrane electrical potential under these conditions. Measurements of this parameter in root epidermal cells indicated that, in the absence of acetic acid the membrane potential was very similar in wild-type and wat1-1D cells (Figure 7a,b). Acetic acid hyperpolarizes wild-type cells, as expected if intracellular acidification activates PMA by increasing phosphorylation of the regulatory domain (Bobik et al., 2010) and by taking pHi to values closer to the optimum of the enzyme (Serrano, 1985). However, the electrical potential of wat1-1D cells was almost unaffected by the acid (Figure 7a,b), probably because this mutant exhibits less intracellular acidification in the presence of acetic acid (Figure 3). The final result is that, in the presence of acetic acid wat1-1D cells are depolarized with respect to wild type, a situation opposite to our expectations.

Figure 7.

Membrane potential and potassium uptake in wat1-1D mutant. (a) wat1-1D plasma membrane is not hyperpolarized in the presence of acetic acid. The graph shows the recordings of plasma membrane electrical potential (Em) of a root epidermal cell in a perfusion experiment. Arrows indicate the addition of 1 mm of acetic acid to the medium. (b) Mean values (±standard error) of plasma membrane electrical potential from four different experiments such as the one shown in (a). (c) Potassium uptake was measured by the initial rate of Rb+ uptake. Thirteen-day-old plants of the wild type (triangles) and the mutant (squares) grown in liquid culture were transferred to K+-free medium supplemented with 10 mm RbCl (open symbols) or to the same medium supplemented with 10 mm RbCl and 3.5 mm acetic acid (closed symbols). Samples were taken at the indicated times for intracellular Rb+ measurement as indicated in Experimental Procedures. Values are the means of three independent experiments, with standard errors of the points in the range of 0.6–1.3 nmoles Rb+ per mg dry weight (DW). The differences between wild type and wat1-1D at the three time points in the presence of acetic acid were significant by Student's t-test with P < 0.05. (d) Potassium content of the wat1-1D mutant and the wild type (wt) in roots and shoots. Plants were grown in hydroponic culture for 6 weeks and the K+ content was determined as described in 'wat1-1D is a dominant-negative mutant that over-express a truncated AP-3 β-adaptin encoded by the PAT2 gene'. Results show the mean and standard error of three experiments. *P < 0.05 and **P < 0.01 by Student's t-test.

The steady-state membrane potential of cells (negative inside) is determined by the relative activities of the ‘pump’ (the H+ pumping PMA; inhibited by increasing potential) and the ‘leak’ (mostly determined by K+ uptake and anion efflux; activated by increasing potential; Bissoli et al., 2012; Goossens et al., 2000). A plausible mechanism for the capability of the wat1-1D mutant to maintain a higher pHi in the presence of acetic acid is that it has an increased ion leak with respect to wild type. This factor would explain the observed depolarization of the plasma membrane (Figure 7a,b) and the increase in proton extrusion (Figure 6b) despite of no increase in the in vitro activity of the H+-ATPase (Figure 6a).

We have investigated if the rate of K+ uptake was increased in the wat1-1D mutant because this aspect is one of the most important components of the ‘ion leak’ in plant cells (Thibaud et al., 1986; Bissoli et al., 2012). As shown in Figure 7(c), in normal conditions K+ uptake, estimated from the rate of Rb+ influx, is similar in the wild type and the mutant. This finding is consistent with previous results indicating that, in the absence of acetic acid, pHi (Figure 3), proton efflux (Figure 6b) and membrane potential (Figure 7b) are similar in epidermal root cells of both genotypes. In the presence of acetic acid K+ uptake is increased in wild-type and mutant plants but the initial rate of uptake under these conditions was significantly higher in the wat1-1D mutant (1.8 versus 1.3 μmoles × g DW−1 × min−1; Figure 7c). This activation of K+ transport is supported by the increased content of K+ in both roots and shoots of wat1-1D with respect to wild type (Figure 7d), a difference that could be observed in the absence of acetic acid. Therefore, the decrease in membrane potential of the mutant is probably caused, at least in part, by an increase in K+ uptake.

In yeast the efflux of acetate mediated by the Pdr12 ABC-ATPase is very important for tolerance to acetic acid (Holyoak et al., 1999) and, therefore, we have investigated acetate efflux in wat1-1D and wild-type plants. As indicated in Figure 8, acetate efflux at early times is similar in wild type and the mutant but at longer times (>30 min) it stops in the wild type while continuing in the mutant. Therefore, acetate extrusion is more efficient in the wat1-1D mutant and it could also contribute to its plasma membrane depolarization and enhanced proton efflux. The reduced capability for acetate efflux in wild type may reflect compartmentalization of the anion at the vacuole.

Figure 8.

Acetate efflux is increased in the wat1-1D mutant. Seedlings were incubated for 90 min in medium with 3.5 mm 14C-acetate and at this stage the content of acetate was the same in the wild type and in the wat1-1D mutant [4.8 ± 0.4 nmoles per mg fresh weight (FW)]. Seedlings were washed and transferred to Murashige and Skoog (MS) medium without acetic acid and incubated for the indicated periods of time to determine efflux of radioactivity to the medium. All radioactivity present in the medium was in the form of acetate as determined by high-pressure liquid chromatography (HPLC) (see Figure S8). The mean and standard error of four experiments are shown.

Although our genetic analysis of pHi homeostasis in Arabidopsis is still in an initial stage, the present results and those of our previous report (Bissoli et al., 2012) indicate that the limiting factor for proton extrusion during intracellular acidification may be the activity of secondary ion channels and transporters. By providing electrical balance during H+ pumping, these systems prevent plasma membrane hyper-polarization and inhibition of the H+-ATPase.


The AP-3 adaptor complex is involved in trafficking from Golgi to vacuoles. It has been observed in yeast (Sun et al., 2004) and animal cells (Dell'Angelica et al., 1999) that loss of function of different subunits of this complex can cause mis-localization of tonoplast proteins to the plasma membrane. Therefore, a plausible mechanism to explain the increased rates of K+ uptake and acetate efflux in the wat1-1D mutant could be that loss of function of AP-3 β-adaptin results in mis-localization of some tonoplast ion channels or transporters to the plasma membrane. This adaptin in plants is only known to be required for the biogenesis of lytic and storage vacuoles (Sohn et al., 2007; Niihama et al., 2009; Feraru et al., 2010) and for the delivery to the vacuole of the sugar transporter SUC4 (Wolfenstetter et al., 2012). Our results suggest a role for Arabidopsis AP-3 β-adaptin in trafficking of ion channels or transporters to the tonoplast. The effect on pHi homeostasis is a mutant phenotype, not a physiological function of the adaptin.

The nature of the channels and/or transporters activated in the wat1-1D mutant can only be speculated. In the case of K+ transport, we have discarded mis-localization of the tonoplast TPK1 channel (Gobert et al., 2007) to the plasma membrane (Figure S9a). Plus, concerning acetate efflux, we have also discarded mis-localization of the tonoplast malate channel (Kovermann et al., 2007; Figure S9b). Given the large number of vacuolar K+ and anion channels and transporters (Ward et al., 2009), it seems that the best approach to identify possible mis-localization of some of these compounds would be a proteomic study with isolated vacuolar and plasma membranes. In addition, it cannot be discarded that the protein(s) miss-localized in the wat1-1D mutant are regulatory components instead of channels or transporters.

Experimental Procedures

Plant growth conditions and stress treatments

Growth of Arabidopsis thaliana (ecotype Columbia) in greenhouse and in in vitro culture (solid or liquid medium) was as described (Alejandro et al., 2007; Bissoli et al., 2012). The composition of the medium was 0.4% Murashige and Skoog (MS) salts, 1% sucrose, 10 mm MES (2-(N-morpholino)ethanesulfonic acid) buffer adjusted to pH 5.5 with Tris base. The acetic acid was adjusted also to pH 5.5 with Tris. Hydroponic culture of Arabidopsis thaliana plants for large-scale root production was performed as described by Tocquin et al. (2003).

Isolation of the wat1-1D mutant

The T-DNA ‘activation-tagging’ seed collection donated to the Arabidopsis Stock Center (NASC) by W. Sheible and C. Sommerville was used for screening mutants tolerant to intracellular acid stress. These mutant lines were obtained by transformation of Columbia plants with the pSKI15 vector (Weigel et al., 2000) and were distributed as a T4 generation. Approximately 30 000 different lines (half of the collection), corresponding to about 400 000 seeds, were screened at high density (9 cm diameter Petri plates containing about 2000 seeds per plate) on MS medium with 7 mm acetic acid. After 6 days, mutants with fully expanded green cotyledons were selected, transferred to soil and grown to maturity. The next generation was analyzed again in a secondary screening at low seed density (100–150 seeds per Petri plate) with 3.5 mm acetic acid. Of the 170 putative mutants isolated in the first screening, five exhibited fully expanded green cotyledons on plates with acetic acid and the one with strongest phenotype was selected for further study.

Genetic characterization, co-segregation analysis and plasmid rescue

wat1-1D mutant was crossed with the wild type (Columbia ecotype) by transferring pollen to the stigmas of emasculated flowers. F2 seeds were scored for germination in 3.5 mm acetic acid to determine if the mutation was dominant or recessive. Thirty-four resistant individuals were selected from the segregating F2 generation, and DNA was extracted (see below) individually to check, by PCR, the correlation between T-DNA presence and resistance to acetic acid. The primer sequences used for the amplification of the CaMV 35S enhancer were as follows: forward (35S_F) 5′-CAACATGGTGGAGCACGACA-3′ and reverse (35S_R) 5′-GGGATCTAGATATCACATCA-3′. Plasmid rescue was performed as described before (Alejandro et al., 2007). Plasmids rescued were sequenced with the following primer: 5′LB (5′-AGATTTCCGAATTAGAATAA-3′).

Construction of Arabidopsis transgenic plants

The construct for over-expression of the At3g55470 gene under the control of the CaMV 35S promoter was made as follows: the cDNA that contained the complete open reading frame of that gene was obtained from the Genomics Science Center, RIKEN (Tsukuba, Japan). The cDNA was subcloned into the NotI site of pBS-SK+ (Stratagene, La Jolla, CA, USA), yielding pBS-470. A fragment that included cDNA from At3g55470 was subcloned into pBI121 (Jefferson et al., 1987) by replacing the GUS coding region between the XbaI and SacI sites.

The construct for over-expressing a truncated form of WAT (At3g55480) ‘in sense’ or ‘antisense’ was made as follows: the truncated gene was amplified from Arabidopsis genomic DNA using the following primers: 480-trunc forward 5′-ATATCTAGAATGTGGCGTCACAGTCATCG-3′ and 480-trunc reverse 5′-TATGAGCTCTCAGCCCAAGCTATATAC-3′. The amplified fragment was cloned into the PCR8/GW/TOPO vector following the instructions of the PCR8/GW/TOPO TA Cloning Kit (Invitrogen, Carlsbad, CA, USA). After determining the orientation of the insert, two LR-reactions were performed to transfer, by homolog recombination, the truncated coding region of WAT (‘in sense’ or ‘antisense’) into the binary destination vector pEarleyGate100 (Earley et al., 2006).

The obtained constructs were introduced into Agrobacterium tumefaciens strain pGV2260 by electroporation. A. thaliana wild-type plants were transformed by flower infiltration (Bechtold et al., 1993; Clough and Bent, 1998). T1 transgenic plants were screened on MS agar medium that contained 50 μg L−1 kanamycin (pBI121) or 50 μm glufosinate (pEarleyGate100), and homozygous lines were selected from the T3 generation.

Measurement of membrane potential and cytosolic pH

Seedlings were grown vertically in MS medium for 10–15 days. For the measurements roots were mounted in a Plexiglas chamber (1.1 ml volume) with a constant perfusion of the assay medium (10 ml min−1). The simplified medium used is based on the one described by Felle (1987) and contains: CaCl2 0.1 mm, KCl 0.1 mm, NaCl 0.1 mm, and 12 mm MES-Bis Tris propane (pH 5.5). When required, 1 mm acetic acid was added to this medium. For the measurements double-barreled micro-electrodes are inserted into epidermal cells. These micro-electrodes were made according to the protocol described by Fernández et al. (1999) and one of their bars contained a H+-specific sensor ETH1907. The signal corresponding to this bar was calibrated against different MES-Bis Tris propane buffer solutions (pHs 5.3, 6.3 and 8.3), in the presence of 96 mm KCl.

Acetate uptake and efflux

For acetate uptake, 5-day-old seedlings were transferred to liquid medium and grown there for 7 days. Seedlings were then incubated for different time periods (0, 7, 15, 30 min) in liquid MS medium supplemented with 3.5 mm 14C-acetate (Perkin Elmer, Waltham, MA, USA), specific radioactivity: 1.78 dpm nmol−1. After incubation, seedlings were washed twice with cold water, and frozen with liquid nitrogen. A wash control was performed introducing the seedlings into the medium with 14C-acetate and washing them immediately. For acetate extraction, 500 μl of water were added to the previously ground samples and they were incubated at 95°C for 15 min. Plant debris was removed by centrifugation and the supernatant was used for the measurement. The measurement was performed by mixing 400 μl of the plant extract with 1 ml of the liquid scintillation cocktail. Once the mix was stabilized, radioactivity was measured using a liquid scintillation counter (Wallac 1410; GE Healthcare, Little Chalfont, UK).

For acetate efflux, seedlings prepared as above were incubated for 90 min in medium supplemented with 3.5 mm 14C-acetate. After the acetic load the content of acetate was the same in the wild type and in the wat1-1D mutant (4.8 ± 0.4 nmoles mg fresh weight). Seedlings were washed three times with cold water, transferred to MS medium without acetic acid and incubated for the indicated periods of time. The radioactivity present in the medium was measured as described above. A wash control was performed introducing the washed seedlings in the medium without acetate just for 2 min in ice and this was subtracted to all the values. The radioactivity in the medium was analyzed by high-pressure liquid chromotography (HPLC) after 1 h of efflux. The external liquid was concentrated and injected into a Waters 515 HPLC system (Waters, Milford, MA, USA) with a Radioflow detector (LB509 EG&G BERTHOLD, Bad Wildbad, Germany). The mobile phase was 5 mm H2SO4 with a constant flux of 0.3 ml min−1. The column used was ICSep IC-ION 300 Column (ICE-99-9850) from Transgenomic (Omaha, NE, USA).

Determination of malate

Malate extraction and measurement was performed as indicated in Hurth et al. (2005), with minor modifications. Briefly, 1 ml of water was added to 100 mg of ground tissue, and the samples were heated for 15 min at 95°C in an Eppendorf thermo-incubator. After centrifugation (5 min, 9000 g), the supernatant was used for malate determination with an enzymatic assay described by Passonneau and Lowry (1993).

Plasma membrane isolation

Roots (2 g) of Arabidopsis plants grown in hydroponic culture for 6 weeks were used for plasma membrane isolation. Samples were ground in 0.4 ml of extraction buffer containing: 250 mm Tris–HCl pH 8, 0.5 m KCl, 25 mm ethylene diamine tetraacetic acid (EDTA), 5 mm dithiothreitol (DTT), 30% w/v sucrose and a protease inhibitor cocktail (Roche, one tablet per 5 ml of buffer). Then 4 ml GTED 20 (10 mm Tris–HCl pH 7.6, 1 mm EDTA, 1 mM DTT and 20% v/v glycerol) was added to the homogenized tissue. The plant debris was removed by filtration and the supernatant was centrifuged at 4°C for 5 min at 3400 rev/min. The supernatant was centrifuged again at 4°C for 1 h at 25 000 g. The pellet was resuspended in 500 μl GTED 20 and the resulting solution was added to the top of a discontinuous sucrose density gradient. The gradient was formed by two layers: STET 33 and STET 41, containing 10 mm Tris–HCl pH 7.6, 1 mm EDTA, 1 mm DTT and the corresponding amount of sucrose, 33% (w/w) and 41% (w/w) respectively. The samples were centrifugated for 2 h at 120 000 g, in a Beckman ultracentrifuge (SW 60Ti rotor). After this centrifugation, the protein fraction located in the interface between the two sucrose layers (1 ml) was diluted with 4 ml of cold water and centrifugated for 1 h at 25 000 g. Finally the pellet was resuspended in 100 μl of GTED 20, using a Dounce homogenizer and the amount of proteins was quantified according to the Bradford method (Bradford, 1976).

In vitro measurement of plasma membrane H+-ATPase activity

H+-ATPase activity measurement is based in the determination of the inorganic phosphate released by the hydrolysis of ATP, as described in Serrano (1988). In this assay 65 μl of H+-ATPase buffer (50 mm MES–Tris pH 6.5, 5 mm MgSO4, 50 mm KNO3, 5 mm sodium azide, 0.2 mm ammonium molibdate and 2 mm ATP) was added to 15–20 μg of protein extract in a 96-well multi-well plate. The plate was incubated at 30°C for 30 min, and the reaction was finished adding 130 μl of phosphate reagent (2% (v/v) H2SO4, 0.5% (w/v) ammonium molibdate, 0.5% (w/v) sodium dodecyl sulfate (SDS) and 0.1% ascorbic acid. After 10 min of incubation, absorbance at 690 nm was measured in a microplate absorbance reader (Bio-Rad, Berkeley, CA, USA). For each sample, a control containing 0.1 mm vanadate (H+-ATPase inhibitor) was performed and the absorbance of the control is subtracted from the absorbance of the respective sample.

External medium acidification assay

Seeds were sown and grown in MS medium for 6 days when approximately 100 seedlings per sample were transferred to water and maintained overnight in the dark to perform a starvation treatment. Next day, seedlings were transferred to 10 ml MS medium without buffer and without sucrose (control) or the same medium supplemented with 1 mm acetic acid. The pH of the medium was adjusted to 5.7 with KOH. Proton pumping was started by the addition of 1% (w/v) sucrose and the pH of the medium was recorded every minute for 45 min using a pH meter GLP22 (Crison, Barcelona, Spain), with electrode 52.08. After 40 min 5 μl of HCl 10 mm were added to the medium to make a calibration. With these data, the increase in proton concentration per fresh weight of the root was calculated for different time periods.

Measurement of Rb+ uptake and K+ content

These methods were described previously (Alejandro et al., 2007; Bissoli et al., 2012).


This work was supported by grants BFU2008-00604 from the ‘Ministerio de Ciencia e Innovación’ (Madrid, Spain) and from PROMETEO/2010/038 of the ‘Consellería de Educación’ (Valencia, Spain). We thank Frans J. Maathuis (York, United Kingdom) for the pA7-TPK1-GFP plasmid, Enrico Martinoia (Zürich, Switzerland) for the pART7-ALMT9-GFP plasmid and for the tdt mutant, Karin Schumacher (Heidelberg, Germany) for the vha-a2 vha-a3 mutant, José M. Pardo (Sevilla, Spain) for the nhx2-1 mutant, NASC for the akt1-1 mutant, Roberto A. Gaxiola (Tempe, Arizona) for the AVP1 over-expression line and Wei-Hua Wu (Beijing, China) for the AKT1 over-expression line. R. Niñoles was supported by a JAE-preDOC contract (‘Consejo Superior de Investigaciones Científicas’, Madrid, Spain).