Plasma membrane H⁺-ATPase genes were retrieved from genomic databases for A. thaliana (The Arabidopsis Information Resource, TAIR) and P. trichocarpa (Phytozome v. 12.1). Amino acid sequences were aligned using ClustalW. Evolutionary distances were computed using the Jones-Taylor-Thornton (JTT) matrix-based method with the complete-deletion option (Jones et al., 1992). Phylogenetic trees were constructed by the neighbour-joining (NJ; Saitou and Nei, 1987) and maximum-likelihood (ML) methods. Bootstrap values were calculated with 1,000 replications using the NJ (Felsenstein, 1985) and ML methods in MEGA7 software (Kumar et al., 2016).

Tissue-specific gene expression patterns of 13 Populus PM H⁺-ATPase genes were examined by re-analysing the RNA sequencing data (Shi et al., 2017). We normalised the raw count data set obtained by RNA sequencing (GSE81077) for xylem, phloem, leaf, shoot, and root with trimmed mean M-values using edgeR v. 3.18.1 (Robinson et al., 2010) in R software v. 3.3.2 (R Core Team, 2018; Hori et al., 2020).

Populus tremula×Populus tremuloides (WT clone T89) seedlings were cultured in 0.8% (w/v) agar box containing 0.5× Murashige and Skoog medium (pH 5.7) at 25°C under a cycle of 16-h white light (50μmolm⁻² s⁻¹)/8-h dark. Cultured hybrid aspens were transferred to the soil mix (3:1 fertilised peat moss: vermiculite, v/v) and grown in two different conditions. One is a greenhouse, and the other is an indoor plant growth room. Greenhouse temperatures were maintained at 21.5±8°C, and natural light was supplemented with metal halide lamps (KI Holdings, Yokohama, Japan) to achieve an 18-h daylength (PAR≥200μmolm⁻² s⁻¹)/6-h dark. The indoor plant growth was maintained at 20°C, and plants were grown under a cycle of 16-h white light (80μmolm⁻² s⁻¹)/8-h dark. Plants were watered and fertilised once weekly with 2,000-fold diluted Hyponex 6-10-5 solution (HYPONeX Japan, Osaka, Japan) for all conditions.

For AtGC1::GUS and CaMV35S::AHA2 constructs, AtGC1 and AHA2 were cloned into the pCR8/GW/TOPO vector (Thermo Fisher Scientific, Waltham, MA, United States) and transferred to the pGWB433 and pGWB402 vectors via the Gateway LR reaction (Nakagawa et al., 2007). Construction of AtGC1::AHA2 was described previously (Wang et al., 2014). The binary vectors (pGWB433-AtGC1::GUS, pGWB402-CaMV35S::AHA2, and pPZP211-AtGC1::AHA2) were transformed into Agrobacterium tumefaciens strain GV3101 (pMP90). Transgenic hybrid aspens were generated using the vectors, essentially as described by Eriksson et al. (2000).

Samples were thoroughly rinsed in distilled water and placed in cold 90% acetone on ice for 5min. Acetone was removed, and GUS staining solution was added for 20min, followed by incubation overnight at 37°C. GUS staining solution consisted of 10mM Na₂EDTA, 50mM phosphate buffer (pH 7.0), 1mM K₄Fe(CN)₆, 1mM K₃Fe(CN)₆, 0.5mg/ml X-Gluc (5-bromo-4-chloro-3-indolyl β-D-glucuronide), and 0.1% Triton X-100. Stained samples were soaked in 70% (v/v) ethanol to remove chlorophyll.

Total RNA was extracted from epidermal fragments using the NucleoSpin RNA Plant kit (TaKaRa Bio, Shiga, Japan). Epidermal fragments from whole leaves were isolated from 10-week-old plants as described previously (Hayashi et al., 2011). First-strand cDNAs were synthesised from total RNA using the QuantiTect Reverse Transcription Kit (Qiagen, Hilden, Germany). Reverse transcription PCR was performed using 2μl of cDNA template with Ex Taq PCR Mix (TaKaRa Bio) and specific primers. Primers were shown in Supplementary Table 1. Master mix of PCR reaction was prepared for each gene of interest, and 20μl of reaction mix, including the cDNA template, was pipetted into each tube. The conditions were 1cycle at 95°C for 2min; 24 (UBQ), 28 (AHA2 and Pt×tHA2) cycles of 95°C for 15s, 57°C for 30s, and 72°C for 30s; and a final incubation at 72°C for 2min. Amplified cDNA was detected by electrophoresis.

Immunohistochemical detection was performed as described previously (Hayashi et al., 2011). Polyclonal antibodies against the conserved catalytic domain of the plasma membrane H⁺-ATPase of Arabidopsis (AHA2) were raised in rabbits. The AHA2 DNA fragment was amplified from first-strand Arabidopsis cDNA with PCR using the specific primers 5ʹ-GCCGGATCCATGGATGTCCTGTGCAGTGAC-3ʹ and 5ʹ-GCCGGATCCTCAAGCACCACGAGCAGC-3ʹ. The resulting amplified DNA fragment of 967–1,845bp of AHA2, which contains BamHI sites at both ends, was cloned into the BamHI site of the pET30a vector (Merck, Darmstadt, Germany). The purified proteins from E. coli (BL21) were used as an antigen. The antiserum was used for immunoblots in Arabidopsis (1:1,000 dilution; Hayashi et al., 2010). PM H⁺-ATPase was detected in guard cells using epidermis isolated from hybrid aspen leaf. The amount of PM H⁺-ATPase was estimated using antiserum against the catalytic domain of AHA2. Fluorescence intensity was quantified according to Ando and Kinoshita (2018).

Gas exchange measurements were performed as described previously (Wang et al., 2014) using the LI-6400 system (Li Cor Biosciences, Lincoln, NE, United States), and parameters were calculated with the software supplied by the manufacturer. White light (1,000μmol·m⁻²·s⁻¹) was provided by a fibre optic illuminator with a halogen projector lamp (15V/150W; Moritex, Saitama, Japan) as a light source and a MHAB-150W; power supply (Moritex). Light was attenuated by a series of optical crown glass metallic neutral density filters (Newport Japan, Hakuto, Japan). The molar flow rate of air entering the leaf chamber, leaf temperature, and relative humidity was maintained at 500μmol·s⁻¹, 24°C, and 40–50% (Pa/Pa), respectively. After the initial 10min of dark adaptation, the plants were exposed to white light (1,000μmol·m⁻²·s⁻¹) for 30min.

Plant height was measured weekly from 21days after potting in a greenhouse. Once trees had reached 20cm in height, the stem diameters were measured weekly at 10cm above the soil. The elongation growth rate of plants was evaluated by a curve-fitting procedure (Buchwald, 2007; Edwards et al., 2018). The radial growth rate was calculated by fitting to a linear function. These procedures were conducted in KaleidaGraph v. 4.1 (Synergy Software, Reading, PA, United States). Leaf number and size (leaves 16–25) were measured when sampling leaves. Leaves were imaged using a scanner (Perfection V700 Photo; Epson, Nagano, Japan) at 600dpi, and leaf size was evaluated by ImageJ 1.51 software.¹ Leaves, stems, and roots were collected from each plant and weighed to calculate the fresh weight. Following 3days of drying at 60°C, the leaves, stems, and roots were weighed again to determine the dry weight (DW). The index of stem volume (volumetric index) was calculated as (diameter ÷ 2)²×height×π, from the final diameter (cm) and height (cm) of an individual tree. A 1cm length of stem segment was sampled from 2cm above the soil to determine wood density. Xylem tissues were obtained by peeling off the bark and were then filled with ultrapure water. The weight increase by increased water volume (V) was measured by a balance at 20°C. The xylem samples were dried in an oven at 105°C for 72h, and DW was measured using a balance. The wood density was calculated by the formula: Wood density=DW ÷ V.

Statistical significance was evaluated by Student’s t test followed by the multiple testing correction procedure of Benjamini and Hochberg (1995), performed using Excel (Microsoft Corp., Redmond, WA, United States).

The P. trichocarpa genome has 13 PM H⁺-ATPase (HA) homologs with high amino acid similarity to A. thaliana HA2 (AHA2; Figure 1; Supplementary Table 2). We designated the P. trichocarpa isoforms PotriHA1–PotriHA13. The Populus isoforms have a highly conserved characteristic sequence, GDGVNDAPALKKA, in the catalytic domain of the P-type ATPase (Axelsen and Palmgren, 1998), supporting our proposal that these isoforms are functional homologs in Populus. The C-terminal region of PM H⁺-ATPases is important for catalytic regulation (Palmgren, 2001; Haruta et al., 2015; Falhof et al., 2016; Inoue and Kinoshita, 2017). All Populus isoforms conserve regions I and II, which are important for autoinhibition, in the C-terminal region (Axelsen et al., 1999), and Thr as a penultimate residue, which is important for its activation via phosphorylation (Figure 1A). Several phosphorylation sites in the C-terminal domain (Thr-881, Ser-899, and Ser-931), in addition to Thr as a penultimate residue, are also highly conserved in Populus PM H⁺-ATPases.

Phylogenetic analysis using full-length amino acid sequences showed that PotriHAs were classified into classes I–V (Figure 1B), as defined in Arango et al. (2003). PotriHA10 and PotriHA11 formed a clade with AHA4 and AHA11 in class I. PotriHA1, PotriHA2, PotriHA3, and PotriHA4 formed a clade with AHA1, AHA2, AHA3, and AHA5 in class II. PotriHA13 formed a clade with AHA10 in class III. PotriHA5, PotriHA6, PotriHA7, PotriHA8, and PotriHA9 formed a clade with AHA6, AHA8, and AHA9 in class IV. PotriHA12 formed a clade with AHA7 in class V. PM H⁺-ATPases in class II, including A. thaliana AHA1 and AHA2 and rice OSA7, have a major role in plants (Haruta et al., 2010; Toda et al., 2016). In P. trhichocarpa, PotriHA2, PotriHA3, and PotriHA4 (class II) showed higher expression than the other class genes in xylem, phloem, leaf, shoot, and root (Figure 1C), suggesting that those isoforms are major PM H⁺-ATPases in Populus species.

We first attempted to introduce CaMV35S::AHA2 into hybrid aspen to ectopically overexpress AHA2. However, no transgenic hybrid aspens were generated, even using 109 stem segments for Agrobacterium-mediated transformation. Therefore, we used the guard cell-specific promoter in A. thaliana, AtGC1, to express AHA2 in hybrid aspen. To investigate AtGC1 activity in hybrid aspen, we generated AtGC1::GUS transgenic plants and examined their GUS activity. As shown in Figure 2A, GUS staining was observed in guard cells of the leaf epidermis of AtGC1::GUS transgenic plants, similar to AtGC1::GUS-expressing A. thaliana (Yang et al., 2008). We next transformed the AtGC1::AHA2 construct into hybrid aspen to overexpress AHA2 in guard cells and generated at least three independent transgenic events (#3, #5, and #9; Figure 2B). In the transgenic plants, AHA2 was expressed in leaf epidermis, as was Pt×tHA2, a major PM H⁺-ATPase in Populus. To examine whether the introduction of AtGC1::AHA2 elevated the PM H⁺-ATPase protein level in guard cells, immunohistochemical analysis using an anti-AHA2 antibody was conducted in the leaf epidermis of transgenic and WT plants. Immunofluorescence was brighter in guard cells of AtGC1::AHA2 transgenic plants than WT plants (Figures 2C,D). Fluorescence intensity relative to the WT showed that the protein level of PM H⁺-ATPase was enhanced in guard cells of transgenic plants (70% for #3, 75% for #5, and 150% for #9), indicating that AtGC1::AHA2 transgenic plants over-accumulated PM H⁺-ATPase in guard cells. The density of stomata in AtGC1::AHA2 transgenic plants was comparable to that in WT plants (196 stomata mm⁻² for WT, 203 for #3, 217 for #5, and 203 for #9; Supplementary Table 3). Therefore, the introduction of AtGC1::AHA2 increased its protein levels in guard cells without affecting stomatal development in hybrid aspen, similar to A. thaliana expressing AtGC1::AHA2 (Wang et al., 2014).

To investigate photosynthetic activity in AtGC1::AHA2 transgenic plants, stomatal conductance and the photosynthetic rate (CO₂ assimilation rate) were measured in intact leaves of transgenic and WT plants grown in an indoor-growth room for 82–94days. The AtGC1::AHA2 transgenic plants showed higher stomatal conductance in the dark compared to the WT (0.07 for WT, 0.22 for #3, 0.15 for #5, and 0.16 for #9mol·m⁻²·s⁻¹. In the WT, white light at 1,000μmol·m⁻²·s⁻¹ increased stomatal conductance. Similarly, light illumination increased stomatal conductance in AtGC1::AHA2 transgenic plants. Stomatal conductance was saturated within 10min of the start of light illumination in the transgenic and WT plants. The average stomatal conductance in the transgenic plants was approximately 3-fold higher than in the WT (Figure 3A). Under identical conditions, photosynthetic rates were saturated 20min after the start of light illumination in WT and AtGC1::AHA2 transgenic plants. The photosynthetic rate was 45% higher in the transgenic compared to the WT plants (Figure 3B). Although stomatal aperture is used to estimate stomatal conductance and photosynthetic activity, determining the average stomatal aperture is more problematic in hybrid aspen compared to A. thaliana, because stomatal size varies in the abaxial epidermis of the former (Supplementary Figure 1; Figure 2A). Taken together, our results indicate that the introduction of AHA2 protein to guard cells increased stomatal conductance and the photosynthetic rate in hybrid aspen.

Enhancement of photosynthetic activity in AtGC1::AHA2 transgenic plants was expected to promote growth and biomass production. When AtGC1::AHA2 transgenic plants and WT plants were grown in a greenhouse for 63days, the transgenic plants showed 14–21% greater height compared to the WT (Figure 4). The elongation rates were 11–15% higher in transgenic than WT plants. The stem diameter and radial growth rates were similar between the transgenic and WT plants. The leaf number per plant was increased 7–16% in the transgenic compared to WT plants, although the area of mature leaves was decreased 14–28%. For biomass production, the volumetric index of stem-trunk biomass was enhanced 14–23% in the transgenic compared to WT plants. Furthermore, the DW of leaves, stems, and roots was non-significantly increased in the transgenic compared to WT plants. However, the stem wood density of the transgenic plants (0.37±0.02g·cm⁻³ for #3, 0.36±0.03g·cm⁻³ for #5, and 0.38±0.01g·cm⁻³ for #9) was similar to that of WT plants (0.37±0.01g·cm⁻³). The increment of tree height in the transgenic plants was observed in the indoor-growth room (Supplementary Figure 2), indicating that the enhancement of growth phenology was stable under different light intensities. Together, our results indicate that AtGC1::AHA2 transgenic hybrid aspens had a higher stem elongation rate and greater biomass production than the WT, likely due to the enhanced stomatal opening and photosynthetic rate.