Root angle is controlled by EGT1 in cereal crops employing an antigravitropic mechanism

Significance The growth angle roots adopt are critical for capturing soil resources, such as nutrients and water. Despite its agronomic importance, few regulatory genes have been identified in crops. Here we identify the root angle regulatory gene ENHANCED GRAVITROPISM 1 (EGT1) in barley. Strikingly, mutants lacking EGT1 exhibit a steeper angle in every root class. EGT1 appears to function as a component of an antigravitropic offset mechanism that regulates tissue stiffness, which impacts final root growth angle. EGT1 is a hot spot for selection as natural allelic variation within a conserved Tubby domain that is linked with steeper root angle. Analogous EGT1-dependent regulation of root angle in wheat demonstrates broad significance of EGT1 for trait improvement in cereal crops.


Shoot and leaf growth angle measurements
For shoot growth angle measurements, plants of TM194 and Morex were grown in blue papers for 7 DAG with a day temperature of 21°C (16 h) and a night temperature of 18°C (8 h). Leaf growth angles were measured using the angle tool in FIJI. N=8 plants per genotype were used. For leaf growth angle measurements, plants of TM194 and Morex were grown in the greenhouse, in a peat and vermiculite growing medium (Vigorplant Irish and Baltic peat-based professional mix) in 15 ´ 15 ´ 30 cm polyethylene pots with a day temperature of 22°C (16 h) and a night temperature of 18°C (8 h). Greenhouse lighting was a mix of natural light supplemented with artificial light by 400watt high-pressure sodium lamps (Sylvania SHP-TS 400W Grolux). Leaf growth angles were measured for the first three leaves of each plant, including the flag leaf, at flowering time (Zadoks growth stage 6). A goniometer was used to measure the angle between the proximal region of the adaxial surface of the blade and the stem.

HVEGT1 structure modelling and Hvegt1 mutant allele mapping
The protein sequence obtained by translating Transcript 3 (427 aa) from HvEGT1 (HORVU6Hr1G068970) entry was used to construct a homology model using the Phyre2 (1) server. A homology modelling approach was chosen over de novo structure prediction from first principles as the gene of interest was inferred to have F-box and Tubby-Like domains, which were confirmed by the protein domain analysis using EBI Interproscan tool. Tubby-Like domain was alone used in the structure prediction algorithm. WGS and haplotype analysis identified mis-sense amino acid substitutions (TM194 and Haplotype II and IV, respectively) were mapped on the predicted structure. Splice acceptor mutation (TM3580) was also visualised with respect to organised F-box and Tubby-like domains. Protein sequence was further studied for its conservation to function prediction across plant species using ConSurf algorithm.

Wheat EGT1 mutant identification
Durum wheat (Triticum turgidum) Tdegt1 mutants were identified from a TILLING population developed in tetraploid cv Kronos (2). Two selected lines (Kronos2551 and Kronos3926) carrying premature termination codons in TRITD6Bv1G159700, the TdEGT1 homoeologous gene on the B genome (TdEGT1_wtA/mutB), were both crossed with the line Kronos2708, carrying a splice donor mutation in TRITD6Av1G172130, the TdEGT1 homoeologous gene on the A genome (TdEGT1_mutA/wtB). F1 plants obtained from both crosses were self-pollinated. Progenies of selected wild-type, single and double mutant F2 individuals derived from the two independent initial crosses (TdEGT1_mutA/mutB) were grown in semi-hydroponic system and analysed for seminal root angle analysis as mentioned above.

Phylogenetic analysis of Tubby-like F-Box Protein Sequences in selected monocots
HvEGT1 was used as a seed gene to select orthologous genes (>40% identity) from key monocot species such as barley (Hordeum Vulgare), wheat (Triticum turgidum), rice (Oryza Sativa spp. Japonica), maize (Zea Mays B73) and brachypodium (Brachypodium distachyon) using interactive phylogenetic module of Monocots Plaza 4.5 (3). Protein sequences were aligned using MUSCLE and tree was constructed using FastTree algorithm. Generated Newick file was imported into iTOL to create an unrooted tree.

Lugol's staining assay
To visualise statoliths in root tips of Morex and TM194 mutant, 1 day pregerminated seedlings were grown in paper rolls in 21 ºC, 16/8 daylight photoperiod growth conditions for 5 days. 1 cm root tips were then embedded in 10% low melting point agarose and sliced using vibratome (7000 smz-2, Campden Instruments, UK) set as 50 Hz frequency, 1 mm amplitude and 40 µm section. Sections were stained using Lugol's iodine solution (VWR chemicals) for 3 minutes and then visualised using LEICA DM 550B light microscope.

RTqPCR analysis during NAA and NPA treatments
Morex and egt1 mutant TM194 seeds were sterilized with 70% ethanol for 5 mins then 15% bleach for 5 min and washed 3-5 times with distilled water. Sterilised seeds were sown directly on ½ Hoagland's No. 2 Basal salt (Sigma, H2395), 1% agar plates and plates were kept at 4ºC for 5 days to improve germination rate. Plates were then transferred to growth room with 16/8h photoperiod and temperature of 22°/18°C. 3-day old plants (post germination) were then transferred to plates containing ½ Hoagland's solution, 1% agar, 0.1% DMSO, plus either 10nM NAA or 1µM NPA. Root tips (5mm form tip) from > 3 individual plants (i.e., ~15 plants) were pooled at 0h and 8h post transfer and flash frozen in liquid nitrogen. RNA was extracted using the Monarch® Total RNA Miniprep Kit (NEB, T2010S) as per protocol and cDNA prepared using Thermo Scientific Revertair frist strand cDNA synthesis kit. Quantitative RT-PCR (qRT-PCR) analysis was carried out with SYBRgreen (Meridian bioscience, Sensimix SYBR Hi-ROX Kit) using qTower 384G machine (Analytikjena). HvAlpha-Tub (HORVU1Hr1G081280.1) and HvGADPH (HORVU6Hr1G054520) were used as internal control and auxin responsive genes HvIAA36 (HORVU0Hr1G021630.1), HvIAA22 and HvIAA20 primers used from Shi et al. (4) for primers see Supplementary Table 1. Three independent biological repeats with four technical replicates were used. Data was analysed using delta Ct method and statistical analysis carried out using Student's T-test. Each treated sample per genotype was normalised by respective DMSO sample.

ROS detection assay
TM194 mutant and Morex seeds were surface sterilised using 20% (v/v) bleach for 4 minutes and were then rinsed five times with de-ionised water. Washed seeds were then germinated on a filter paper saturated with de-ionised water in a petri dish kept at 21 ºC for 48 hours. Seedlings with uniform growth were placed on a germination paper, rolled into paper rolls and grown vertically at 21 ºC for 4 days. CM-H2DCFDA (Sigma-Aldrich) dissolved into dimethyl sulfoxide (DMSO, VWR Life Science) was used to visualize the localization of ROS in Morex and TM194 mutant root tips. 20 µM CM-H2DCFDA was prepared in 50 μM potassium chloride buffer (50mM KCl, 10mM MES, pH 6.0) on the day of the experiment. Root samples were taken 1 cm from the tip and were treated with 1 ml of CM-H2DCFDA for 15 minutes under vacuum. After treatment, samples were washed thoroughly with potassium chloride buffer four times. Samples were then placed on a glass slide with 50% glycerol as mounting agent and visualized with the Zeiss Leica DM5000 fluorescent microscope. CM-H 2 DCFDA could be deacetylated by cellular esterase and then subject to oxidisation by ROS to 2',7'-dichlorofluorescein (DCF), which is highly fluorescent and could be detected under excitation and emission spectra of 492-495 nm and 517-527 mm, respectively. To minimize any variation in processing and imaging samples, all roots per seedling were stained, mounted on one glass slide and imaged together. Gain was adjusted for each slide at the saturation limit of the root showing maximum glow and then set for all the roots on the same slide. To identify any spatial differences in ROS accumulation in each root, we took multiple high magnification fluorescent images along the longitudinal axis of root and stitched them into one complete image. This stitched image was then quantified in five different developmental zones: 4 equal length zones between root tip and first visible root hair and the last one as root hair differentiation zone. Mean fluorescent value for each zone was calculated in FIJI. Two biological replicates were performed with 4 seedlings per replicate and 4-5 seminal root tips per seedling were analysed. Statistical analysis was performed using Welch's t-test in "RStudio". *, **, *** indicate significant P-value < 0.05, 0.01, 0.001 (n roots=16-20, n plants=4, n experiments=2).      (8). Magenta, yellow, cyan and white color indicate alpha helices, beta strands, turns and random coils, respectively. G395 indicates the mis-sense substitution position of TM194 mutant on the predicted domain. G395 sits in a highly positively charged cavity, probably stabilized by an adjacent negatively charged C-terminal site. G395E substitution leads to a small, neutral amino acid changing to a larger, negatively charged residue that is highly likely to destabilise this region of the protein, likely affecting its function. b, Amino acid charge distribution in a. c, Domain organisation and mapping of TM3580 and TM194 mutations and haplotypes II and IV identified from the WHEALBI (9) (12) highlighting residues 391 to 400 (red dotted rectangle). G395E and haplotype II (Fig. 4d-e) lie within this highly conserved region. Colored letters indicate, 'e' = exposed and 'b' = buried residues according to the neuralnetwork algorithm; 'f' = predicted functional residue (highly conserved and exposed) and s = predicted structural residue (highly conserved and buried).

Fig. S9. Phylogenetic analysis of Tubby-like F-Box Protein sequences in barley, wheat, rice, maize and brachypodium.
HvEGT1 was used as a seed gene to select orthologous genes (>40% identity) from key monocot species such as barley (Hordeum Vulgare), wheat (Triticum turgidum), rice (Oryza Sativa spp. Japonica), maize (Zea Mays B73) and brachypodium (Brachypodium distachyon) using interactive phylogenetic module of Monocots Plaza 4.5 (3). Protein sequences were aligned using MUSCLE (13) and tree was constructed using FastTree (14) algorithm. Generated Newick file was imported into iTOL (15) to create an unrooted tree. Light green color highlights the distinct clade formed by HvEGT1 with 2 orthologs each from wheat and maize and 1 ortholog each from rice and brachypodium.   Bar graph showing RTqPCR analysis of HvEGT1, HvIAA36, HvIAA22 and HvIAA20 expression in TM194 mutant and Morex measured under control, DMSO control,10nM NAA and 1uM NPA treatments. Here, control represents experimental conditions where plants were grown in 1% Agar plates supplemented with ½ Hoagland's solution and grown in 16/8h photoperiod and a temperature 22ºC/18ºC for 3 days prior transferring to treatment on 1% Agar plates supplemented with ½ Hoagland's solution, 0.1% DMSO and either 10nM NAA or 1µ M NPA. 5mm seminal root tips from >3 plants (i.e., ~15 root tips) were pooled for RNA extraction and 3 independent biological replicates were performed. Data shows mean +/-standard error fold change. T-test was performed to assess the statistical difference. * indicates P-value < 0.05 and 1.5 >=Fold change<= -1.5 between treated vs DMSO samples for each genotype.      showing the stiffness values of the stele and cortical tissues of the root. The pooled data from each area was analysed using a non-parametric Wilcoxon test for significant differences between sample type and area (p < 0.001).

Fig. S20: Protein-protein interaction database analysis suggests that EGT1 regulate cell elongation and cellulose synthesis related proteins.
The protein-protein interactions (experimental, text mining, database curated and co-expression) of HVEGT1 were studied using String Database (https://string-db.org) to identify five high-confidence interactions (4 co-expression -black colored edges and 1 database curated -green colored edge). MLOC_14747.2 (HORVU2Hr1G032710) and MLOC_61785.1 (HORVU2Hr1G063820) are orthologous to Arabidopsis thaliana LONGIFOLIA3, one of the 4 proteins that control polar cell elongation by regulating cell wall modifying enzymes encoded by a multi-gene family xyloglucan endotransglucosylase/hydrolase (19). Quadruple mutant of LNG family members have shown to have reduced cell elongation and reduced organ sizes. MLOC_76406.1 (HORVU7Hr1G061070) is orthologous to Arabidopsis thaliana SHOU4, a plasma-membrane-localised proteins that negatively regulate cellulose synthesis by inhibiting the exocytosis of CESAs (cellulose synthases) (20). SHOU4 mutant cells in inflorescence stem are shown to be smaller and with thinner cell walls. MLOC_5095.4 (HORVU5Hr1G035610) is orthologous to KIPK (KCBP-Interacting Protein Kinase) speculated to be involved in regulation of cell expansion through transport of cell wall material (21). MLOC_17061.4 (HORVU7Hr1G097840) is orthologous to Arabidopsis thaliana a plant-specific GATA-type transcription factor family protein. Bar graphs represent a, measured expression (FPKM values) of EGT2 in Morex and egt1 mutant alleles (TM194 and TM3580) and b, measured expression (FPM values) of EGT1 in wild type and Hvegt2 mutant in different root growth zones (root cap, meristem and elongation zone) (22). These graphs suggest that expression of EGT1 in egt2 mutant largely remain unchanged and vice versa.