Seeds of wild type Arabidopsis (Col-0), a transgenic line overexpressing LEA5 (OEX2-5; Mowla et al., 2006) and T-DNA insertion mutant lines (lea5) were used in the following studies. The lea5 T-DNA SALK line (SALK_099663), which has an insertion in the second exon of AT4G02380, was obtained from TAIR. A T-DNA insertion rh22 mutant line with a T-DNA insertion in sixth intron of the At1g59990 gene, was also used in these studies. The seeds of this line were a gift from Professor Masatake Kanai (Kanai et al., 2013). Genotyping was performed as recommended by the T-DNA primer design website http://signal.salk.edu/tdnaprimers.2.html.

Arabidopsis seeds were vapor-sterilized using commercial bleach with concentrated HCL for 2 h, then sown onto Petri dishes containing 0.5x Murashige and Skoog (MS) basal salts, 1% agar, pH 5.7 and sealed with micro pore tape and later stratified for 2 days at 4°C.

The plates were then either transferred to controlled environment chambers at 20°C and grown under a long photoperiod (16 h light and 8 h dark), with a light intensity of 150 μmol m^–2 s^–1 and relative humidity (70%) for 14 days, or germinating seeds were transplanted to soil and grown in controlled environment chambers, under growth conditions, as described above.

Seeds of wild type barley (Hordeum vulgare CV Golden Promise) and transgenic barley lines expressing the Arabidopsis LEA5 sequence (9.1, 10.1, and 11.2) were sterilized using 10% (^v/_V) bleach for 15 min, followed by 3 washes (each 10 min) with 80% of ethanol and 5 washes with 50 ml of sterilized water in a laminar flow cabinet. Plants were grown on compost under controlled environment conditions with a 16 h light/8 h dark photoperiod regime (400 μmol m^–2 s^–1).

For these studies, seeds were sown onto Petri dishes, as described above, except that the plates contained either no added oxidants (Control), or methyl viologen (0.1 μM), or menadione (0.1 mM). Seedlings were grown under a 16 h photoperiod (150 μmol.m^–2. s^–1 irradiance) at 22°C ± 2°C for 21 days prior to analysis.

For these studies, seedlings of wild type barley, transgenic lines expressing the Arabidopsis LEA5 sequence (9.1, 10.1, and 11.2) and empty vector and GUS controls were grown for 7 days on soil under well-watered conditions in controlled environments, as described above. Thereafter, half of the plants were maintained under well-watered conditions and half were deprived of water for 7 days. Photosynthesis measurements were then performed.

Intact chloroplast and mitochondria fractions were isolated from 14-day old plate-grown seedlings according to protocols described by Aronsson and Jarvis (2011) and Murcha and Whelan (2015), respectively. Tape-Arabidopsis Sandwich protocols were used for protoplast isolation and PEG transformation (Wu et al., 2009).

The percentage of intact mitochondria present in isolated fractions was measured via the latency of cytochrome C oxidase (COX) activity in the absence and presence of Triton X-100 using Clark-type oxygen electrode (Neuberger et al., 1982; Shaw et al., 2017).

Respiratory control at the level of Complex 1 was measured as described by Shaw et al. (2017). Firstly, mitochondria were added to a 1 ml liquid phase oxygen electrode chamber at final concentration 100 μg/ml. Pyruvate, thiamine pyrophosphate (TPP) and malate were added to the final concentrations of 2, 3, and 2 mM, respectively. After 2 min 20 μl of 100 mM ADP was added and after a further 2 min the reaction was stopped with the addition of 2 μg/ml oligomycin. Then 0.5 μM carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP) was added.

Respiratory control at the level of Complex II was measured as described by Shaw et al. (2017). Mitochondria were first added to a 1 ml liquid phase oxygen electrode chamber at final concentration 100 μg/ml. Then 10 mM succinate was added to the chamber. After 1 min, 0.5 μM rotenone was added to inhibit complex I activity. Thereafter, 20 μl of 100 mM ADP was added followed 2 μg/ml oligomycin and 0.5 μM FCCP.

Intact organelle fractions were suspended in assay buffer, prior to analysis. Aliquots (60–100 μg protein) were added to buffer containing the mixture of 19 amino acids. Reactions were started by the addition of 225 uCi [35S] methionine (1,000 Ci/mM) and incubated at 25°C with constant rotatory shaking for 60 min. Mitochondrial fractions were incubated in the dark, while chloroplast fractions were incubated at low light conditions (100 μmol m^–2 s^–1). Reactions were stopped by the addition of cold buffer containing non-labeled methionine. Proteins were separated on 14% SDS-PAGE gels. The incorporation of [35S]-Met into protein bands was determined by soaking the gels in ENLIGHTNING™ Rapid Autoradiography Enhancer. After drying, the gels were analyzed using a phosphorimager (Bio Rad) at high resolution for 48 h.

Reverse transcription of 1 μg of RNA aliquots into cDNA was performed using the QuantiTect Reverse Transcription Kit (Qiagen). Thereafter, qPCR was performed using QuantiFast SYBR Green PCR kit (Qiagen) in the presence of 0.5 μM primers in a CFX96 thermocycler (Biorad, Hercules, CA, United States) following the manufacturer’s instructions. The two-step cycling protocol was programmed as follows: incubation at 95°C for 5 min; 40 cycles of amplification comprised of 95°C for 10 s, 60°C for 30 s and 72°C for 30 s. The mean value of three replicates was normalized using actin 11 as internal controls. All amplifications for the cloning purposes were done with Thermo Scientific Phusion High-Fidelity DNA Polymerase according to manufacturer’s instructions. All primer sequences are listed in Supplementary Table 2.

Intact leaf mesophyll protoplasts were prepared from 14-day old Arabidopsis seedlings expressing the LEA5-YFP protein. Images of these protoplasts or 5 days old plate-grown seedlings expressing the LEA5-YFP protein were prepared using a Zeiss LSM700 laser scanning confocal microscope using 20×/0.8 Plan-Apochromat, 40×/1.2 W C-Apochromat or 63 × /1.4 Oil Plan-Apochromat in multi-track channel mode. Three fluorescent proteins were excited with the following excitation wavelengths: GFP with 488 nm, YFP with 514 nm, and RFP with 651 nm. Mitochondria were labeled with red-fluorescent dye, Mitotracker Red CMXRos (Thermo Fisher Scientific) after excitation at 578 nm. Images were collected using LP emission filters 500–710 nm and processed using Zeiss ZEN 2011 (black edition).

A range of constructs (RH22-YFPn and LEA5-YFPc or RH22-YFPc and LEA5-YFPn) were designed to study protein-protein interactions and transiently expressed in Arabidopsis mesophyll protoplasts. Homologous co-transfections using either LEA5-YFPn and LEA5-YFPc or RH22-YFPn and RH22-YFPc were performed as controls. For these studies protoplasts were co-transfected with 10 μg of plasmid mixture the appropriate constructs, as described by Wu et al. (2009). Images of the interactions were prepared using confocal microscopy (Horstman et al., 2014).

Barley immature embryos were extracted and co-cultivated with Agrobacterium cells containing the BRACT214_SAG21-YFP for 3 days in the dark (Comadira, 2015). After callus induction and regeneration of shoots on hygromycin selection, putative transgenic plantlets were transferred to soil and grown in the glasshouse. PCR was performed using hygromycin B and Arabidopsis-YFP primers to confirm the presence of transgene. Thirty-six lines were selected as positive. Homozygous seedlings were selected from T3 generation in 0.5x MS plates supplemented with hygromycin B. Three lines with the closest 3:1 ratio (resistant: sensitive) were subsequently chosen; 9.1; 10.1 and 11.2. For controls, barley plants were transformed in parallel with pBRACT214 empty vector and pBRACT204 vector containing the GUS fragment (Comadira, 2015).

T3 generation plants, control lines and the wild type were grown to maturity in compost in glasshouses at the University of Leeds and the James Hutton Institute in Scotland under a 16 h/8 h day photoperiod regime with supplemental lighting. The number of fertile tillers were counted and the total seed yield quantified.

Arabidopsis cell suspension cultures (PSB-D) were transformed using a co-cultivation method with direct selection in liquid medium, as described previously (Van Leene et al., 2011). TAP tagging of protein complexes was performed using a GS^rhino tag (Van Leene et al., 2015). Aliquots (200 mg) of protein were precipitated. Proteins were then separated on SDS-PAGE as described by Bürckstümmer et al. (2006) and Van Leene et al. (2015). Protein interactors were identified by mass spectrometry using an LTQ Orbitrap Velos mass spectrometer. Proteins with at least two matched high confident peptides were retained for further analysis (Supplementary Table 1). Background proteins were filtered out based on the frequency of occurrence in the co-purified proteins in a large dataset containing 543 TAP experiments using 115 different baits (Van Leene et al., 2015).

Samples (5 μg of total RNA) were denatured in 1 volume of NorthernMax™-Gly Sample Loading Dye for 30 min at 50°C. They were then separated on a 1.2% (w/v) agarose gel, blotted to positively charged nylon membrane by capillary blotting, and then fixed by cross-linking. Hybridization was performed with ³²P-labeled probes in 5 × SSC, 5 × Denhardt’s solution, and 0.5% SDS at 65°C overnight. The membranes were washed twice in 2 × SSC, 0.1% SDS and once in 1 × SSC, 0.1% SDS at 65°C.

Hybridization probes were prepared by PCR amplification of 23S-4.5S rrn region (primer sequences listed in Supplementary Table 2). The fragments were separated on 1.5% agarose gels and extracted using QIAquick Gel extraction kit (Qiagen). They were then labeled with 20 μCi [α-32P]dCTP, 3000 Ci/mmol (PerkinElmer) using a “Random primed” DNA labeling kit (Roche).

Light response curves for photosynthesis were measured using a portable Infrared Gas Analyzer (model LI-6400XT) LI-COR. Measurements were performed at 20°C and a CO2 concentration of (400 μmol mol^–1) in the leaf chamber. The leaves were exposed to each of the following light intensities: [(0, 50, 250, 500, 750, 1,000, 1,250, 1,500, and 1,750 μmol m^–2 s^–1) photosynthetically active radiation (PAR)] allowing the leaves to acclimatize to each irradiance for at least 15 min prior to measurement to allow stabilization of parameters. Measurements were made on 3 plants per line per experiment.

The lea5 mutants and the transgenic plants overexpressing the LEA5 protein (OEX2-5) had a similar vegetative growth phenotype to the wild type (Figures 1A,D). The growth of the wild type seedlings was decreased in the presence of the prooxidants menadione (Figures 1B,D). and paraquat (Figures 1C,D). The prooxidant-related decreases in plant growth were greater in the lea5 mutant seedlings than the wild type (Figures 1B–D). However, decreases in growth caused by the prooxidants were much less pronounced in the OEX2-5 seedlings than the wild type (Figures 1B–D).

Confocal microscopy of Arabidopsis mesophyll protoplasts expressing a LEA5-YFP fusion protein, revealed that the LEA5 protein was localized to the chloroplasts as well as the mitochondria (Figures 2A–D). In contrast, LEA5-YFP protein was only detected in the mitochondria of the mesophyll cells of 5-day old leaves of Arabidopsis seedlings (Figures 2F,G).

A tandem affinity purification (TAP) approach was used to identify protein interaction partners for LEA5. For these studies, a LEA5-YFP construct was constitutively expressed in transgenic Arabidopsis ecotype Landsberg erecta suspension cultures grown either in the dark (PSB-D) or the light (PSB-L). We characterized the proteins that interacted with LEA5 in the absence or presence of hydrogen peroxide (H₂O₂) in order to determine whether oxidative stress altered the repertoire of LEA5 protein binding partners. Several proteins that interacted with the LEA5 protein in the cultures grown in the absence or presence of H₂O₂ were identified (Table 1). The list includes a chloroplast-localized DEA (D/H)-box RNA helicase family protein (RH22; AT1G59990), which is involved in chloroplast ribosome biogenesis and plays a key role in Arabidopsis growth and stress responses (Gu et al., 2014; Liu and Imai, 2018), as well as another DEA(D/H)-box RNA helicase family protein (RH20) and pumilio (PUM) 24, which regulates mRNA degradation and translation repression. Since LEA5 interacted with RH22 in all tagging experiments, we analyzed this interaction further using bimolecular fluorescence complementation (BIFC) assays. This method confirmed the interaction of LEA5 with RH22 specifically in the chloroplasts of Arabidopsis leaf protoplasts (Figure 3).

To explore the functions of LEA5 in organelles, the expression and processing of chloroplast rRNA was measured in wild type Arabidopsis and in mutants lacking RH22 (rh22), lea5 mutants and in plants overexpressing LEA5 (OEX2-5; Figure 4). In contrast to the rh22 mutant, which showed aberrant chloroplast ribosome processing, the lea5 mutant and OEX2-5 plants showed similar levels of chloroplast rRNAs to the wild type (Figure 4).

Finally, organellar translation in chloroplasts and mitochondria was analyzed in wild type, OEX 2-5 and lea5 by in organello translation assays which measures the incorporation [35S] methionine to newly synthesized proteins of isolated organelles (Figure 5). The intensity of labeling for some mitochondrial protein products was seen to be decreased in mitochondria isolated from lea5 and increased in mitochondria isolated from OEX 2-5 relative to the wild type (Figure 5A). Conversely, the intensity of labeling of some translated chloroplast protein products was observed to increase in chloroplasts isolated from lea5 mutants and decreased in chloroplasts isolated from OEX 2-5 relative to the wild type (Figure 5B). Measurement of mitochondrial respiration revealed that mitochondria isolated from the OEX 2-5 exhibited significantly higher rates of respiratory electron transport to the cytochrome C oxidase than mitochondria from wild type controls (Table 2).

LEA5 overexpression has previously been reported to increase the growth and flowering of Arabidopsis plants (Mowla et al., 2006; MohdSalleh et al., 2012). We therefore set out to investigate whether LEA5 overexpression also had a positive effect on the growth and productivity of crop species. Barley was selected for these studies because an efficient transformation system is available for this important crop. The Arabidopsis LEA5 protein in barley was characterized for the growth phenotypes, yield parameters and photosynthesis in a large number of independent transgenic lines. Vegetative and reproductive phenotypes were compared in three overexpression lines (9, 10, and 11) that had high levels of LEA5 expression compared to the wild type barley plants and empty vector and GUS control lines (Figure 6). Total biomass, the number of tillers, flowers, heads numbers, total seed yield per plant and seed yield per tiller were measured in T3 generation plants that had been grown for 6 months under glasshouse conditions (Figure 6 and Table 3). The number of fertile tillers and seeds produced in transformed barley lines (9; 10, and 11) were significantly higher than the controls (Table 3).

We have previously shown that the expression of AtLEA5 was strongly induced by drought, such that the levels of transcripts correlated with the drought-dependent inhibition of photosynthesis (Mowla et al., 2006). Moreover, Arabidopsis lines LEA5 were more susceptible to drought than the wild type, in terms of the drought-induced inhibition of photosynthesis (Mowla et al., 2006). We therefore next explored the responses of the transformed barley lines (9.1; 10.1, and 11.2) to drought by withholding water for 7 days. Photosynthetic CO2 assimilation was monitored throughout the period of drought course using infrared gas analysis. The AtLEA5-overexpressing barley lines exhibited comparable rates of photosynthesis to the wild type under water-replete conditions (Figure 7A), with comparable rates of stomatal conductance (Figure 7C). The light response curves for photosynthesis were similar in all lines grown under well-watered conditions (Figure 7A). Photosynthetic CO₂ assimilation rates were decreased in all lines following exposure to drought (Figure 7B). However, the light-saturated rates of photosynthesis were lower in the overexpression lines than controls after 7 days of water deprivation (Figure 7B). Stomatal conductance was decreased in all lines following exposure to drought (Figure 7D). All lines showed similar changes in stomatal conductance in response to increasing irradiance under these conditions (Figure 7D).

To investigate the subcellular localization of the Arabidopsis LEA5 protein in the leaves and roots of the transgenic barley lines, lines expressing 35S-LEA5-YFP were generated and analyzed. Confocal microscopy images revealed that the LEA5-YFP protein was present exclusively in the mitochondria of the leaves and roots of the transgenic barley lines (Figure 8).