The plasma membrane H+-ATPase gene family in Solanum tuberosum L. Role of PHA1 in tuberization

Genetic evidence is provided for the role of potato PM H+-ATPases in tuberization, demonstrating that the isoform PHA1 promotes stolon elongation and tuber growth, and is involved in sugar–starch metabolism.


Introduction
Plasma membrane (PM) H + -ATPases are integral membrane proteins that pump protons out of the cell, generating an electrochemical gradient of protons across the plasmalemma. As a result, PM H + -ATPases provide the driving force for the transport of ions and metabolites through channels and transporters (Palmgren, 2001). This process is required for several physiological responses, such as cell expansion, phloem loading/unloading, stress adaptation, and plant growth and development (Michelet and Boutry, 1995;Palmgren, 2001;Gaxiola et al., 2007;Schumacher and Krebs, 2010). PM H + -ATPases have been described in many plant species, such as Arabidopsis thaliana (Gaxiola et al., 2007), Nicotiana plumbaginifolia , Oryza sativa (Baxter et al., 2003), Solanum lycopersicum (Kalampanayil and Wimmers, 2001), Zea mays (Santi et al., 2003), and Cucumis sativus (Wdowikowska and Klobus, 2016).
The plant PM H + -ATPases are regulated at different levels. With respect to post-translational regulation, the mechanism involves the autoinhibitory action of the C-terminal domain that can be displaced by phosphorylation of the penultimate residue, a threonine, and the subsequent binding of 14-3-3 proteins, resulting in pump activation (Palmgren et al., 1991;Svennelid et al., 1999;Maudoux et al., 2000). More recently, it has been described that phosphorylation of other residues can also affect PM H + -ATPase activity Piette et al., 2011), suggesting that these proton pumps are controlled by a complex regulatory mechanism. Other factors can regulate PM H + -ATPase activity. A novel interaction partner of PM H + -ATPase named PPI (proton pump interactor) was identified in A. thaliana (PPI1; Morandini et al., 2002) and Solanum tuberosum (StPPI1; Muñiz García et al., 2011); this protein increases the activity of the proton pump in vitro once the C-terminus has been displaced.
Several studies have indicated that regulation of PM H + -ATPase also occurs at the genetic level. PM H + -ATPases are encoded by five distinct gene subfamilies. The members of subfamilies I and II show ubiquitous expression profiles, while the expression of genes belonging to subfamilies III, IV, and V is restricted to specific tissues (Arango et al., 2003;Gaxiola et al., 2007). PM H + -ATPase transcript levels are differentially regulated by environmental and hormonal signals, such as salt stress (Kalampanayil and Wimmers, 2001;Janicka-Russak and Kłobus, 2007;Sahu and Shaw, 2009), dehydration (Surowy and Boyer, 1991), low temperature (Ahn et al., 1999), and exogenous application of hormones (Frías et al., 1996;Janicka-Russak and Kłobus, 2007).
In potato (Solanum tuberosum cv. Désirée), two genes encoding PM H + -ATPases (PHA1 and PHA2) were isolated long ago by Harms et al. (1994). However, their physiological functions have not yet been elucidated. The economic importance of potato plants resides in the capacity to produce tubers. Potato tubers develop from stolons, which are underground stems. All aspects of tuber development are coordinated by a complex interaction of phytohormones and environmental signals (Sarkar, 2008). Gibberellic acid (GA) is a central negative regulator of tuberization that stimulates stolon elongation and prevents tuber induction (Xu et al., 1998). Abscisic acid (ABA) and auxin also regulate tuberization, acting as promoting factors (Xu et al., 1998;Roumeliotis et al., 2012;Kolachevskaya et al., 2015).
At early stages of tuber formation, stolons display a cessation of elongation and the initiation of subapical radial growth, which is followed by a deposition of starch and storage proteins . During the transition from stolon to tuber, a switch from apoplastic to symplastic sucrose phloem unloading occurs (Viola et al., 2001). Phloem loading of sucrose in source leaves requires the activity of sucrose-H + symporters, and apoplastic sucrose unloading in sink tissues involves sucrose-H + and hexose-H + symporters (Lemoine et al., 2013). These sugar transporters depend on the proton motive force generated by PM H + -ATPases; therefore, these proton pumps might play an important role in stolon and tuber development.
This study presents a comprehensive overview of the PM H + -ATPase gene family (PHA) in potato, including the identification of PHA genes and expression profiles to explore which PHA genes are possibly involved in tuberization. Based on the results of the expression analysis, PHA1 was chosen for further characterization by the development of transgenic plants.

Abiotic stress treatments
The first two fully expanded leaves detached from plants grown in soil, in a greenhouse, for 4 weeks were used for salt stress and drought treatments. Prior to stress treatment, leaves were placed in individual containers with water in a growth chamber at 22 °C, under a 16 h light photoperiod, for 48 h, to allow the wound response components induced by excision to be restored to basal levels. Only the cut end of the petioles was immersed in the solution. For salt stress, leaves were treated with 250 mM NaCl. For drought treatment, water was removed from the containers and leaves were kept in air. For cold treatment, in vitro grown potato plants were exposed to 4 °C, while control plants remained at 22 °C, under a 16 h light photoperiod.

Stolon growth conditions and in vitro tuberization
Tuberization can be studied in vitro, reproducing the process occurring in vivo with the advantages of generating tubers in a controlled environment (Xu et al., 1998;Roumeliotis et al., 2012;Muñiz García et al., 2016). Single-node cuttings obtained from potato plants do not induce tubers when cultured in darkness on standard propagation media (MS medium plus 2% sucrose); however, increasing sucrose concentration (8%) increases the frequency of tuberization.
Shoot apices derived from in vitro plants were cultured on solid MS media containing 20 g l -1 sucrose for 2 weeks, in a growth chamber at 22 °C, under a 16 h light photoperiod, prior to harvesting single-node explants. Nodal segments were grown on MS medium containing 20 g l -1 or 80 g l -1 sucrose (non-inducing and tuberinducing conditions, respectively) solidified with 0.7% (w/v) agar, in a growth chamber in darkness at 19 °C. In some experiments, the media was supplemented with hormones (5 μM GA 3 , 25 μM 3-indoleacetic acid, or 5 μM ABA; purchased from Sigma, St Louis, MO, USA) or the inhibitors of H + -ATPase (1 mM sodium orthovanadate or 50 μM erythrosine B; purchased from Sigma). The single-node cuttings formed etiolated shoots/stolons that, under tuber-inducing conditions, developed tubers. For tuber-inducing conditions, the process was studied for 9-10 weeks. After this time of culture, most of the tubers were fully developed, and the rate of tuber growth decreased significantly.
All the experiments were performed at least three times independently, using 10-20 stolons per condition. A tuberizing stolon was defined as a stolon presenting visible subapical swelling or tubers.
Stolons presenting more than one subapical swelling/tuber were considered as one tuberizing stolon. A branched stolon was defined as a stolon presenting at least one branch ≥5 mm. The total stolon length was determined as the sum of the primary and secondary stolon lengths. Tuber weight was calculated as the sum of the weight of all the tubers divided by number of tubers obtained. If the stolon presented more than one tuber, only the fully developed tubers (minor diameter ≥4 mm) were used for weight determination, since, in most cases, the additional tubers were very small or undeveloped.

Tuberization in soil-grown plants
Yield determination was carried out on plants transferred to soil ex vitro and cultivated in a growth chamber at 19 °C, under a 12 h light photoperiod, for 10 weeks, in 0.5 liter pots with commercial soil mixture (Grow Mix Multipro, Terrafertil Argentina). Stolon length determination was carried out on plants obtained from seed tubers, grown in 0.8 liter pots with commercial soil mixture, in greenhouse, for 4 weeks.

Semi-quantitative reverse transcription-PCR (RT-PCR)
Semi-quantitative RT-PCR to determine PHA gene expression was performed as described in País et al. (2010) using the primers and reaction conditions shown in Supplementary Table S1 at JXB online. RT-PCR bands were quantified relative to the elongation factor 1-α gene (EF1-α) using ImageJ software (http://rsb.info.nih. gov/ij/).
Cloning of the PHA1 gene from Solanum tuberosum cv. Spunta The PHA1 coding sequence was obtained by RT-PCR from RNA isolated from potato (S. tuberosum cv. Spunta) flower bud using the primers PHA1 5'UTR Fw, GGAAGAGAGGAAATTGAGAAAGATG; and PHA1 3'UTR Rv, GCCGATAATGAATGCTGTTATAG. The amplified product was cloned into the pCR-Blunt II-TOPO vector (Thermo Fisher Scientific, Waltham, MA, USA), for sequencing.

Yeast complementation assay
The complementation assay was carried out using the Saccharomyces cerevisiae strain YAK2 (MATα,leu2Δ1,trp1Δ63, lacking the genomic copies of PMA1 and PMA2, the two endogenous H + -ATPase genes. Survival is possible by the expression of the PMA1 gene under the GAL1 promoter on a URA3-bearing centromeric plasmid (de Kerchove d'Exaerde et al., 1995). The YEplac181 plasmid (bearing the 2μ origin of replication and the LEU2 marker) containing the promoter region of the yeast PMA1 gene was used to express the different plant PM H + -ATPase genes. Colony PCR was performed to detect the presence of the corresponding plasmids in the YAK2 transformants. The transformed YAK2 strains were grown at 30 °C in a synthetic medium containing 0.7% yeast nitrogen base without amino acids, supplemented with all of the amino acids except those used for selection (His, Leu, Ura and Trp), and 2% glucose (MGlu-His, Leu, Ura, Trp) or 2% galactose/1% raffinose (MGal-His, Leu, Ura, Trp). Solid media contained 2% agar. The media were supplemented with 20 mM K 2 HPO 4 and buffered to a pH of 6.5. In addition, the YAK2 transformants were grown on MGlu-His, Leu, Trp medium at pH 6.5, containing 0.1% 5-fluoroorotic acid (5-FOA) to counter-select the plasmid bearing the yeast PM H + -ATPase gene.

Generation of PHA1-OE transgenic potato plants
The PHA1 sequence was subcloned into the pBI121 binary vector downstream of the Cauliflower mosaic virus 35S promoter (35S CaMV). Transformation of potato discs was performed using Agrobacterium tumefaciens strain LBA4404 as described in Muñiz García et al. (2014). Regenerated plants carrying no plasmid but obtained from the same explants and by the same regeneration method were used as controls (wild type). All plants were obtained from the same tuber; therefore, regeneration controls and transgenic lines have the same genetic background. Three independent transgenic lines (L1, L2, and L3) were selected and characterized in this study.
PCR analysis was performed to confirm the presence of the transgene using genomic DNA as template and the primers PHA1 5' UTR forward (GGAAGAGAGGAAATTGAGAAAGATG) and tNOS reverse (TGATAATCATCGCAAGACCG) to amplify the transgene but not the endogenous PHA1 gene. PCR was also performed using primers to detect the nptII gene: nptII Fw, ATGATTGAAGAAGATGGATTG; and nptII Rv, GAAGAACTCGTCAAGAAGGCG. The 18S rRNA gene was used as a positive control utilizing the primers 18S forward, GGGCATTCGTATTTCATAGTCAGAG, and 18S reverse, CGGTTCTTGATTAATGAAAACATCCT.
PM H + -ATPase activity PM H + -ATPase activity was determined in purified membranes isolated by a two-step aqueous two-phase partitioning system as described in Olivari et al. (2000) from leaves of potato plants grown in soil for 60 d or stolons cultured in vitro on MS medium plus 8% sucrose for 3 weeks. The PM H + -ATPase activity was determined in 5 mM ATP, 5 mM MgCl 2 , and 10 mM PIPES pH 7.3, with the addition of 5 mM NaN 3 , 0.1 mM sodium molybdate, and 100 mM KNO 3 , in the absence or presence of 100 μM Na 3 VO 4 . The assay was performed in the presence of 0.01 mg ml -1 lysophosphatidylcholine (Papini and De Michelis, 1997). The assay was carried out using the plasma membrane samples (2.5 μg of protein) in a final reaction volume of 150 µl for 30 min at 30 °C. Released Pi was measured using the malachite green method (Hohenwallner and Wimmer, 1973). The PM H + -ATPase activity was determined as the difference between the activity measured in the absence and presence of vanadate.

Observation of stolon epidermal cell imprints
A thick layer of clear nail polish was brushed onto the epidermis of each stolon. Once dried, the nail polish was peeled off, placed on a clear glass, and observed using a light microscope (Olympus BX41, Olympus Optical Co. Ltd, Tokyo, Japan). Photomicrographs of the imprints were obtained at ×200 magnification in the microscope.

Starch content
The samples (50 mg) were ground to a fine powder in liquid nitrogen and extracted three times with 10 ml of 80% ethanol, by boiling the samples in a 90 °C water bath for 20 min to dissolve sugars, dextrins, and tannins, and centrifuged at 1500 g for 5 min. The residues were dried, resuspended in 15 ml of 1 M HCl, and incubated at 99 °C for 45 min for starch hydrolysis. The contents were filtered through Whatman No. 40 filter paper. The filtrates were adjusted to pH 7.0 with NaOH and analyzed for glucose by the Somogyi-Nelson method (Nelson, 1994). Starch content was estimated by multiplying the glucose content by the glucose equivalent of 0.9. Results were expressed as g starch 100 g -1 FW.

Cell wall invertase activity
The pellet mix procedure was used to assess the activities of cell wall invertase (Albertson et al., 2001). About 50 mg of stolon tissue was homogenized in ice-cold extraction buffer at a 1:5 (w/v) ratio containing 50 mM HEPES (pH 7.5), 10 mM MgCl 2 , 1 mM EDTA, 1 mM EGTA, 1 mM benzamidine, 1 mM phenylmethylsulfonyl fluoride (PMSF), 2 μg ml -1 trypsin inhibitor, 5 mM aminocaproic acid, 5 mM DTT, 0.1% (v/v) Triton X-100, and 2% (w/v) polyvinylpyrrolidone (PVP). The extract was centrifuged at 12 000 g for 10 min at 4 °C; the pellet was washed three times with extraction buffer without PVP and used for activity assay after a final resuspension in a 1:4 (w/v) ratio with the same buffer. The activities were assayed in a final volume of 1 ml, containing 125 mM Na-acetate buffer (pH 4.5), 100 mM sucrose, and 80 μl of the pellet mix. The reactions were incubated at 37 °C for 30 min. Reactions were stopped in boiling water for 1 min, centrifuged at 12 000 g for 5 min, and glucose was quantified in the supernatant according to the Nelson-Somogyi method (Nelson, 1994).

Statistical analysis
Statistical analysis was carried out using the Student's t-test. A P-value <0.05 was considered statistically significantly.

Identification of PHA genes in potato
A search in the Potato Genome Sequencing Consortium database, (http://solanaceae.plantbiology.msu.edu/pgsc_download. shtml) revealed the existence of seven sequences presenting a high degree of homology with PM H + -ATPases in the S. tuberosum Phureja genome, designated PHA1-PHA7 (Supplementary Fig. S1; Supplementary Table S2). Analysis of the protein primary structure revealed that the potato PM H + -ATPases present the characteristic features of PM proton pumps: the C-terminal autoinhibitory domain, 10 transmembrane segments, the small cytosolic loop, between the second and third transmembrane segments, and the large cytosolic loop, between the fourth and fifth transmembrane segments ( Supplementary Fig. S1). PHA6 is truncated at its C-terminus and presents a unique region of 88 amino acids, absent in the other members of the PHA family. PHA3 also presents a 34 amino acid sequence which is missing in the other PHAs. Scanning the PM H + -ATPase protein sequences for motif analysis confirmed the presence of typical motifs for plant P-ATPases: TGES, crucial for ATPase activity, DKTGTLT, in which the aspartate is reversibly phosphorylated during catalysis, KGAP, essential for ATP binding, and the GDGVNDA motif, involved the hydrolysis of the acylphosphate intermediate (Thever and Saier, 2009). The last three motifs are not completely conserved in PHA6. The 14-3-3-binding site, located at the extreme C-terminus, is absent in PHA5 and PHA6 (Supplementary Fig. S1).
Multiple alignment of the protein sequences of A. thaliana (AHA), N. plumbaginifolia (PMA), O. sativa (OSA), and S. tuberosum Phureja (PHA) PM H + -ATPases was carried out to generate a phylogenetic tree (Fig. 1). The potato proteins were grouped into four of the five well-established subfamilies (Arango et al., 2003), except for PHA6, which does not fall into any of the subfamilies defined, and was excluded from the phylogenetic analysis. PHA1-PHA5 and PHA7 were clustered into the subfamilies I, II, IV, and V, while no members of PHAs were represented in subfamily III.

Expression profile of PHA genes
The expression profile of PHA genes in different organs was determined ( Fig. 2A; Supplementary Fig. S2). PHA1, PHA2, and PHA3 were expressed in all organs, while expression of PHA4 and PHA5 was restricted to flower buds and flowers. PHA6 was expressed predominantly in flower buds and roots. PHA7 transcripts were not detected in any of the organs analyzed (not shown).
The mRNA levels of PHA genes were determined in stolons cultured in vitro. PHA1, PHA2, and PHA3 were expressed in stolons cultured under tuber-inducing conditions, while PHA4, PHA5, PHA6, and PHA7 transcripts were not detected ( Supplementary Fig. S3). PHA1 and PHA2 were expressed at higher levels in early stages of tuber organogenesis than in the tuber growth stages, while PHA3 displayed an opposite expression pattern (Fig. 2B). The sucrose-H + symporter StSUT1 showed an expression pattern similar to that of PHA1 and PHA2 (Fig. 2B). GA increased PHA1 and PHA2 expression in stolons, while no significant changes were observed for PHA3 (Fig. 2C); no significant changes in the expression of PHA genes were detected in response to auxin or ABA (Fig. 2C). PHA4, PHA5, PHA6, and PHA7 transcripts were not detected at any stage of tuberization, with or without hormone treatment (not shown).
The expression of PHA genes in response to abiotic stress was determined in leaves. Salt stress increased PHA1, PHA2, and PHA3 mRNA levels, while drought only up-regulated the PHA3 gene ( Supplementary Fig. S4). Cold stress increased the expression of PHA2 (Supplementary Fig. S4). PHA4, PHA5, PHA6, and PHA7 transcripts were not detected in leaves in the absence or presence of stress (not shown).

Involvement of PM H + -ATPase in tuberization
In a first attempt to elucidate the role of PHAs in tuberization, the inhibitors of H + -ATPase, sodium orthovanadate (vanadate) and erythrosine B (Kanczewska et al., 2005;Hayashi et al., 2014;Yuan et al., 2017), were used in tuberization experiments in vitro. Dose-response experiments were carried out to determine the appropriate concentration of inhibitors, which were 1 mM and 50 μM for vanadate and erythrosine B, respectively (not shown). The length of inhibitor-treated stolons was drastically reduced (Fig. 3A, C). Both inhibitors increased the percentage of tuberizing stolons cultured under tuber-inducing conditions (MS medium plus 8% sucrose) (Fig. 3B, C). The tubers obtained from stolons treated with vanadate or erythrosine B were significantly smaller than those obtained from untreated stolons (Fig. 3D). When stolons were cultured under non-inducing conditions (MS medium plus 2% sucrose), inhibition of PM H + -ATPases by vanadate or erythrosine B also inhibited stolon growth and promoted tuber initiation ( Supplementary Fig. S5).
Cloning of the PHA1 gene from S. tuberosum cv. Spunta PHA1 was selected for further analysis. The full-length coding sequence of PHA1 was cloned from S. tuberosum cv. Spunta.
The nucleotide sequence appears in the GenBank database under the accession number KX827766. This sequence is almost identical to the S. tuberosum Phureja PHA1 sequence (PGSC0003DMP400055772; Potato Genome Sequencing Consortium database), except for one nucleotide difference within the coding region, with no amino acid change. Comparing the sequence of PHA1 from S. tuberosum cv. Spunta with the sequence of PHA1 from S. tuberosum cv. Désirée (Harms et al., 1994; GenBank accession number: X76536.1), 15 nucleotide differences were found within the coding region, one of which causes an amino acid change (R170A, in Spunta).
To confirm that the PHA1 gene cloned encodes a protein with PM H + -ATPase activity, a functional complementation assay was carried out. The S. tuberosum cv. Spunta PHA1 coding sequence was subcloned in YEplac181 under the control of the constitutive promoter of the yeast PMA1 gene. This plasmid was introduced into the yeast strain YAK2, which is deleted from its own two PM H + -ATPase genes (PMA1 and PMA2), and is able to survive on galactose medium with a centromeric plasmid carrying the yeast PMA1 controlled by the GAL1 promoter (de Kerchove d' Exaerde et al., 1995). Three independent transformants (PHA1-1, PHA1-2, and PHA1-3) obtained in a galactose medium were able to sustain the growth of YAK2 yeast cells when shifted to a glucose medium (Fig. 4A), demonstrating that the PHA1 gene cloned from S. tuberosum cv. Spunta has ATPase and proton pumping activity. The same result was obtained after eliminating the plasmid bearing the yeast PM H + -ATPase from the YAK2-PHA1-3 strain by 5-FOA (Fig. 4B).

Overexpression of PHA1 promotes stolon growth and branching, and increases tuber weight
Transgenic plants expressing the PHA1 gene under the control of the 35S CaMV promoter (PHA1-OE) were developed. Three independent PHA1-OE transgenic lines (L1-L3) were selected for detailed characterization. The presence of the transgene was confirmed by PCR amplification (Fig. 5A). RT-qPCR analysis showed that the mRNA abundance of PHA1 was higher in transgenic lines than in wild-type plants (Fig. 5B). Accordingly, transgenic lines exhibited a higher PM H + -ATPase activity in leaves and stolons (Fig. 5C).
After 2 weeks of culture under tuber-inducing conditions, PHA1-OE stolons were longer than wild-type stolons (Fig. 6A, B). Accordingly, transgenic stolons had longer epidermal cells (Fig. 6C, D). No branching was observed after 2 weeks (Fig. 6B). After 3 weeks of culture, the percentage of branched stolons was significantly higher in PHA1-OE stolons than in wild-type stolons (Fig. 6E, H). The primary stolon length of L1 and L3 was higher, with no statistically significant difference between L2 and wild-type stolons (Fig. 6F), although the total stolon length was increased in all three transgenic lines, due to the increased number of branches (Fig. 6G). PHA1-OE plants grown in soil presented longer stolons than wild-type plants (Fig. 7A, B), confirming the long-stolon phenotype observed in vitro.
The percentage of tuberizing stolons was not significantly affected in PHA1-OE lines (Fig. 8A); however, after 10 weeks of culture, ~26% of PHA1-OE stolons presented more than one tuber (2-5 tubers per stolon), while only 7% of wild-type stolons developed more than one tuber (Fig 8B, C). In most cases, the additional tubers were very small or undeveloped. The tubers developed from PHA1-OE stolons were larger and presented higher starch content than those obtained from wild-type stolons (Fig. 8D-F).
The tuberization capacity of PHA1-OE plants grown in soil was determined. There were no significant differences in the number of tubers obtained per plant between PHA1-OE and the wild type (Fig. 9A). The average weight of tubers obtained from PHA1-OE plants was higher than that of wildtype tubers (Fig. 9B). The tuber yield per plant was higher in transgenic lines with respect to wild-type plants, although this difference was statistically significant only for L2 and L3 (Fig. 9C). It is important to note that the tuber yield was determined in plants transferred to soil ex vitro, which are much smaller than potato plants obtained from seed tubers, and consequently have significantly lower yields.

Overexpression of PHA1 alters cell wall invertase activity, starch content, and StSUT1 expression in stolons
Cell wall invertase activity, starch content, and StSUT1 expression were determined in stolons during the elongation stage. After 2 weeks of culture, PHA1-OE stolons showed higher cell wall invertase activity (Fig. 10A) and accumulated significantly less starch than wild-type stolons, showing very low values of starch content (Fig. 10B). Treatment with H + -ATPase inhibitors inhibited stolon elongation ( Supplementary Fig. S6), decreased cell wall invertase activity (Fig. 10A), and increased starch accumulation (Fig. 10B) in PHA1-OE stolons, abolishing the phenotypic differences between transgenic and wild-type stolons. PHA-OE stolons  an OD 600nm =1.7) of the following YAK2 yeast strains were spotted onto solid media containing glucose (MGlu-His, Leu, Trp) or galactose (MGal-His, Leu, Trp), buffered at pH 6.5: Control -, YAK2 transformed with empty YEplac181; Control +, YAK2 transformed with YEplac181-E14D, that expresses the constitutively active form of the PM H + -ATPase PMA2 Q42932 from N. plumbaginifolia under the yeast PMA1 promoter; PHA1∆N, YAK2 transformed with YEplac181-PHA1∆N that expresses a truncated, inactive form of PHA1, lacking the first 553 nucleotides of the N-terminus (devoid of the TGES motif) under the yeast PMA1 promoter; PHA1-1/2/3, YAK2 transformed with YEplac181-PHA1, that expresses PHA1 under the yeast PMA1 promoter (three different transformed yeast clones were used). (B) Serial dilutions (starting from an OD 600nm =1.7) of the following YAK2 yeast strains spotted onto solid media containing glucose (MGlu-His, Leu, Ura, Trp): Control -, PHA1-PHA3, and PHA1-PHA3 after 5-FOA treatment to eliminate the plasmid bearing the yeast PM H + -ATPase gene. Yeast strains were grown at 30 °C for 48 h.
showed higher expression levels of the sucrose-H + symporter gene StSUT1 than wild-type stolons (Fig. 10C).

Overexpression of PHA1 promotes stem growth and branching
Plant growth parameters were determined in 2-week-old plants generated in vitro from internode cuttings. PHA1-OE plants showed higher stem and first internode lengths, and higher number of leaves than wild-type plants (Fig.11A-D). No significant differences were observed in the number of primary roots or total root length, between transgenic and wild-type plants ( Supplementary Fig. S7). Inhibition of PM H + -ATPase activity by vanadate or erythrosine B resulted in shorter plants ( Supplementary Fig. S8). No branching was observed in 2-week-old plants (Fig. 11D). After 4 weeks of growth in vitro, the percentage of plants presenting branches was significantly higher in PHA1-OE plants than in wild-type plants (Fig. 11E,  F); a similar result was obtained using plants generated from shoot apexes (Supplementary Fig. S9). PHA1-OE plants grown in soil were ~20% taller than wild-type plants (Fig. 11G, H).

The potato PM H + -ATPase family
An in silico screening in the S. tuberosum Phureja genome database retrieved seven sequences, designated PHA1-PHA7, with a high degree of homology with PM H + -ATPases ( Fig. 1 Supplementary Fig. S1), including two genes (PHA1 and PHA2) identified earlier in S. tuberosum cv. Désirée (Harms et al., 1994).
Only PHA1, PHA2, and PHA3 are expressed in stolons ( Supplementary Fig. S3), and their expression levels change during the tuberization process (Fig. 2B), suggesting a role for these PM H + -ATPases in the development of stolons and tubers. Inhibition of PM H + -ATPase activity by vanadate or erythrosine B inhibits stolon elongation, promotes tuber induction, and impairs tuber growth ( Fig. 3; Supplementary  Fig. S5), confirming that PM H + -ATPases are involved in tuberization acting at different stages of the process.

PHA1 promotes stolon elongation
PHA1 was selected for functional analysis by overexpression in potato plants. PHA1-OE stolons elongate faster and present longer epidermal cell length than wild-type stolons after 2 weeks of culture (Fig. 6A-D), indicating that PHA1 activity contributes to the stolon elongation that occurs prior to tuber initiation by promoting cell elongation. In agreement with this, inhibition of PM H + -ATPase activity with vanadate or erythrosine B results in shorter stolon length ( Fig. 3A; Supplementary Fig. S5A). These results are consistent with the role of PM H + -ATPases in cell elongation and expansion described for other tissues such as hypocotyls and roots (Haruta and Sussman, 2012;Inoue et al., 2016), and Nicotiana tabacum BY-2 cells (Niczyj et al., 2016). Acidification of the apoplasm by PM H + -ATPases leads to the wall-loosening process and cell elongation (reviewed in Hager, 2003). GA promotes stolon elongation by stimulating both cell division and cell elongation (Loy, 1977;Xu et al., 1998). PHA1 expression is induced in stolons in response to GA (Fig. 2C), thus it is possible to speculate that PHA1 mediates the GA-induced stolon elongation. However, more studies are needed to establish the mechanistic link between GA and PHA1.

PHA1 regulates sucrose-starch metabolism in stolons
During the stolon elongation phase, apoplastic sucrose unloading predominates (Viola et al., 2001). Although the exact mechanism occurring during this process remains unknown, sucrose can be unloaded from the phloem into the apoplast by the sucrose-H + symporter StSUT1 in its inverse transport mode (Kühn et al., 2003;Carpaneto et al., 2005), or possibly by recently discovered facilitators of the SWEET family (Chen et al., 2012). The apoplastic sucrose can be converted to hexoses by a cell wall invertase, and taken up by the parenchyma cells through hexose-H + symporters, not yet characterized in potato. Sucrose uptake from the aploplast might be mediated by sucrose-H + symporters, possibly different from StSUT1, since it was reported that StSUT1 is localized in sieve elements but not in storage parenchyma in sink tubers (Kühn et al., 2003) (Fig. 12). Apoplastic unloading correlates with high acid invertase activity (Ross et al., 1994;Appeldoorn et al., 1997;Viola et al., 2001), and is associated with rapidly growing vegetative sink tissues (Ehness and Roitsch, 1997;Herbers and Sonnewald, 1998). During the transition from stolon to tuber, a switch from apoplastic to symplastic unloading occurs, accompanied by a switch from an invertase-sucrolytic to a sucrose synthase (SuSy)-sucrolytic pathway, leading to the starch accumulation phase and tuber growth (Viola et al., 2001) (Fig. 12).
During the elongation phase, PHA1-OE stolons grow faster (Fig. 6A), and present higher levels of cell wall invertase activity (Fig. 10A) and significantly lower starch content than wild-type stolons; moreover, starch is almost undetectable in transgenic stolons (Fig. 10B). These metabolic differences are accompanied by higher expression levels of StSUT1 in PHA1-OE stolons (Fig. 10C). In agreement with this, the PHA1-OE stolon phenotype can be reversed by the application of PM H + -ATPase inhibitors, that results in shorter length, decreased cell wall invertase activity, and increased starch content (Fig. 10A, B; Supplementary Fig.  S6). Based on these data, it may be hypothesized that, besides promoting stolon growth by mediating cell elongation, PHA1 regulates the sucrose-starch metabolism in elongating stolons (Fig. 12). PHA1 positively regulates cell wall invertase activity (by an unknown mechanism), and its proton pumping activity can sustain the secondary transport of hexose-H + and sucrose-H + symporters, promoting the uptake of hexoses and sucrose by parenchyma cells; moreover, PHA1 increases the expression of StSUT1 in stolons, promoting sucrose unloading into the apoplast (Fig. 12). By doing so, PHA1 favors apoplastic unloading, sucrolysis by invertase and stolon elongation, and prevents starch accumulation, which is associated with symplastic unloading. Supporting this hypothesis, PHA1 and StSUT1 expression is high in wild-type stolons during the elongation phase, and decreases during tuber growth (Fig. 2B). In this context, PHA1 seems to play a key role in the molecular mechanism that determines the switch from the apoplastic unloading/invertase-sucrolytic pathway to the symplastic unloading/SuSy-sucrolytic pathway.
PHA1-OE stolons show higher expression levels of StSUT1 (Fig. 10C), suggesting that this PM H + -ATPase may be part of the signaling pathway that leads to the activation of StSUT1 transcription. It is possible that PHA1 induces the expression of StSUT1 in companion cells, and the StSUT1 mRNA is transported via plasmodesmata to the sieve elements as previously described (Kühn et al., 1997) (Fig. 12).

Role of PHA1 in tuber development
PHA1-OE tubers are larger and present higher starch content than wild-type tubers (Figs 8D-F, 9B). StSUT1 is essential for sucrose phloem loading (Kühn et al., 1996); therefore, it is possible to speculate that PHA1 might enhance sucrose phloem loading by energizing this sucrose-H + symporter (Fig. 12), promoting sucrose translocation to sink tubers and starch accumulation. The mechanism of sucrose phloem loading in whole plants is different from the process occurring in stolons cultured in vitro; however, in both cases  sucrose loading requires a proton symport mechanism energized by PM H + -ATPases. In plants, sucrose is loaded into the phloem in source leaves (Lemoine et al., 2013), while in stolons cultured in vitro, sucrose is obtained from the culture medium and loaded into the phloem as either hexoses (after hydrolysis by invertase) or sucrose (De Riek et al. 1997). PHA1-OE stolons show an increased branching phenotype (Fig. 6E, H), which results in a higher percentage of stolons presenting more than one tuber, although the additional tubers are undeveloped (Fig. 8B, C). Stolon branching increases the number of potential tuber sites; however, developing more than one tuber per stolon leads to a competition for resources that negatively affects tuber size (Celis-Gamboa et al., 2003). Interestingly, the fully developed tubers obtained from PHA1-OE stolons are larger than those from wild-type stolons (Fig. 8D, E), indicating that PHA1 positively regulates tuber growth. This result agrees with the observation that inhibition of PM H + -ATPase activity results in smaller average tuber weight (Fig. 3D). There are no significant differences in the percentage of tuberizing stolons between PHA1-OE and wild-type plants (Fig. 8A); however, treatment with vanadate or erythrosine B enhances tuber induction ( Fig. 3B; Supplementary Fig. S5B); this effect might be due to the inhibition of PHA2 or PHA3, which are also expressed in stolons, or to a non-specific action.
As observed in vitro, PHA1-OE plants grown in soil also develop larger tubers and show an increased tuber yield with respect to wild-type plants (Fig. 9B, C), suggesting that the PHA1 gene might be a potential tool to increase potato crop yield.

Branching phenotype of PHA1-OE plants
Stems of PHA1-OE plants grown in vitro present a highly branched phenotype, similar to PHA1-OE stolons (Fig. 11E, F; Supplementary Fig. S9). It has recently been shown that lateral bud growth depends on the amount of sucrose translocated to those buds, with sugar distribution, not auxin, being the initial regulator of apical dominance (Mason et al., 2014). It is now clear that sugar supply to axillary buds is not only essential to trigger outgrowth, but is also required to release bud dormancy (Barbier et al., 2015). Sucrose transporters and PM H + -ATPases have been implicated in this process; in rose bush, the onset of bud outgrowth correlates with increased sugar translocation to axillary buds and the up-regulation of the sucrose-H + symporter gene RhSUC2 (Girault et al., 2010;Henry et al., 2011); the dormancy break of Prunus persica buds is associated with sucrose uptake and PM H + -ATPase activity (Aue et al., 1999;Maurel et al., 2004). The branched phenotype of PHA1-OE plants could be due to an increased sucrose unloading towards the lateral buds via sucrose-H + symporters energized by PHA1; however, more studies are required to determine the role of PHA1 in axillary bud outgrowth.

PHA1 promotes plant growth
Like PHA1-OE stolons, the stems of PHA1-OE plants grown in vitro elongate faster, and show longer internodes and more leaves than wild-type plants (Fig. 11A-D). Likewise, PHA1-OE plants grown in soil are taller than wild-type plants (Fig. 11G, H). Thus, PHA1 mediates growth in the aerial part of the plant as well as in underground stolons. Accordingly, inhibition of PM H + -ATPase activity leads to a reduction in stem length (Supplementary Fig. S8). There is much evidence for the role of PM H + -ATPase in plant growth based on studies using proton pump inhibitors; however, strong genetic evidence has so far been sparse. It was reported that overexpression of the unmodified PM H + -ATPase gene PMA4 in tobacco has no effect on plant growth; however, the constitutive expression of a constitutively activated form of PMA4 results in abnormal leaf inclination and twisted stems, suggesting alterations in cell expansion (Gévaudant et al., 2007). Another study has shown that the Arabidopsis aha2 mutant completes its life cycle without any observable growth alteration, and the reduced-growth phenotype only becomes apparent under stress conditions that reduce the transmembrane electrical gradient and/or external proton chemical gradient (Haruta and Sussman, 2012). The present study provides strong genetic evidence for the role of PHA1 as a driver of growth in potato plants, showing a clear enhanced-growth phenotype of PHA1-OE plants.

Supplementary data
Supplementary data are available at JXB online. Table S1. List of primers used for semi-quantitative RT-PCR. Table S2. Chromosome localization, coding region length, predicted protein length, and molecular weight of the PHA isoforms. Fig. S1. Alignment of the protein sequences of PHA isoforms. Fig. S2. Quantification of RT-PCR bands of Fig. 2A. Fig. S3. RT-PCR analysis of PHA genes in stolons. Fig. S4. Expression of PHA genes in response to abiotic stress.