The three grain amaranth species (Amaranthus hypochondriacus cv. “Revancha,” A. cruentus cv. “Dorada,” A. cruentus cv. “Tarasca” and A. caudatus) employed in this study were provided by Dr. Eduardo Espitia Rangel, INIFAP, México, curator of the Mexican amaranth germplasm collection. The plants were germinated in a conditioned growth chamber, as described previously (Délano-Frier et al., 2011). Briefly, seeds were germinated in 60-well germinating trays containing a sterile soil preparation composed of a general soil mixture (three parts Sunshine Mix 3TM [SunGro Horticulture, Bellevue, WA], one part loam, two parts mulch, one part vermiculite [SunGro Hort] and one part perlite [Termolita S.A., Nuevo León, México] and coconut paste [Hummert de México, Morelos, México] in a 1:1 v/v relation). Germination occurred shortly after in a growth chamber maintained at 27 ± 1°C, ≈75% relative humidity (RH) and within a 16: 8 h light (under a photosynthetically active radiation [PAR] of ~300 μmol m⁻² s⁻¹) dark photoperiod. Three weeks after germination, the seedlings were transplanted to 1.3 L plastic pots containing 250 g of a general substrate and were grown in the same conditions as above. They were fertilized once, 1 week after transplant, with a 20:10:20 (N: P: K) + micronutrients nutrient soil drench solution according to the manufacturer's instructions (Peters Professional; Scotts-Sierra Horticultural Products, Marysville, OH, USA). Grain amaranth plants having 10–12 expanded leaves, were employed for all (a)biotic stress treatments, except defoliation (see below).

Arabidopsis thaliana plants, ecotype Columbia, were used for the generation of transgenic plants. A. thaliana seeds were germinated and grown following standard methodologies (Rivero et al., 2014). Briefly, seeds were sterilized by rinsing once with 70% ethanol, for 2 min and then in a 20% hypochlorite solution for 5 min. They were washed thrice with sterile distilled de-ionized (dd) water and kept under-water for 48 h at 4°C. Vernalized seeds were germinated on 100 × 15 mm Petri dishes containing 0.1 × Murashige and Skoog (MS) media (Murashige and Skoog, 1962) prepared with 0.43 g/ L MS salts (Caisson laboratories, North Logan, UT, USA), plus 0.5% sucrose, 100 mg/ L myo-inositol and 1% agar (Research Organics Inc., Cleveland, OH, USA). The dishes were transferred to growth chambers maintained under controlled conditions of light (PAR of 60–80 μmol m⁻² s⁻¹), photoperiod (16 h light/8 h dark) and temperature (22 ± 1°C). Plants were grown for 10 d post germination before transferal to 125 mL plastic pots filled with 60 g of a soil mixture composed of three parts Sunshine Mix 618 3TM (Sun-Gro Horticulture), one part loam, two parts mulch, one part vermiculite (SunGro Horticulture) and one part perlite (Termolita S.A.) and fertilized with a 20:10:20 (N: P: K) + micronutrients solution (Peters Professional). Plants were subsequently cultivated in growth rooms conditioned at 22°C 16 h light/ 8 h dark photo-period.

The gene and mRNA sequences of the grain amaranth Nattering-4-like-1 gene (AhN4L-1) were retrieved from the A. hypochondriacus genome (Phytozome accession number: AHYPO_022599-RA CDS). The intronic and exonic regions of the AhN4L-1 gene were determined directly via the Phytozome platform (Goodstein et al., 2012).

A MatInspector Release professional 8.0.5 program (Genomatix), in addition to the New PLACE database of plant cis-acting regulatory DNA elements (Higo et al., 1999) were used to detect regulatory elements in the 1,600 bp promoter region of the AhN4L-1 gene. The protein encoded by the open reading frame (ORF) of the AhN4L-1 cDNA was deduced with the aid of ExpasyTranslate platform (https://web.expasy.org/translate/). The AhN4L-1 protein sequence was used to perform BLASTp searches against different plant genomes within the NCBI platform. Multiple alignment with amino acid sequences of related proteins in other plant species, based on the NCBI taxonomy, was used to construct the phylogenetic trees, which were generated with MEGA software (version 11.0) using the maximum like-hood method based on the JTT matrix-based model (Jones et al., 1992).

The AhN4L-1 protein was modeled entering the amino acid sequence, in FASTA format, into the Iterative Threading ASSEmbly Refinement or I-TASSER server platform (http://zhang.bioinformatics.ku.edu/I-TASSER), able to accept sequences ranging from 10 to 1,500 aa. The link predicted the most viable 3D structures which were modeled and visualized using the PYMOL software (http://pymol.sourceforge.net) (Roy et al., 2010).

Groups of six grain amaranth plants (i.e., A. hypochondriacus, A. cruentus cv. “Dorada” or cv. “Tarasca” and A. caudatus) were used to perform all stress experiments. These were performed in the same chamber under used for germination and growth. The plants were subjected to acute salt stress by watering the 1.3 L pots for eight consecutive days with 100 mL of a 400 mM NaCl solution. The electrical conductivity of the soil at field capacity at the end of the stress treatment was ≈ 12.1 dS/m. It was measured using a portable HI 98130 conductometer (Hanna instruments Inc., Woonsocket, RI, USA). The conductivity reached by the soil in the salt treatment was within a range known to affect moderately salt-sensitive crops such as alfalfa, corn, millet, sugar cane, tomato, potato, and spinach (http://www.fao.org/docrep/005/y4263e/y4263e0e.htm). Control plants were maintained in the same conditions and were watered with dd water. The water stress treatments were performed using A. cruentus cv. “Dorada” as described previously (González-Rodríguez et al., 2019). Insect herbivory experiments were performed by placing three 3rd instar Trichoplusia ni (Lepidoptera, Noctuidae) larvae per plant. Larvae were starved for 16 h prior to the assays. All plants were covered with a net to restrict larval movement. Larvae were allowed to feed for 8 h and were retired from the plant. Plant tissue sampling of the herbivore-damaged plants was done 6, 12, and 24 h posterior to larval removal. Damaged and undamaged leaves, in addition to stems and roots, were collected. The bacteria infection assays were performed with Pseudomonas syringae pv. syringae 3525 (Pss) and Clavibacter michiganensis subsp. Michiganensis acr42 (Cmm). These pathogens were previously shown to interact with grain amaranth as avirulent and virulent bacterial pathogens, respectively (Casarrubias-Castillo et al., 2014). Briefly, aliquots of Pss and Cmm bacterial cultures were diluted to an OD₆₀₀ of 0.2 (Pss) and 0.3 (Cmm) with sterile 0.05 M phosphate buffer, pH 7.0. They were subsequently used to infect the amaranth plants (16 plants per species) by spraying 400 μL of the respective cultures on the abaxial surface of the leaves closest to the apical meristem. For the Cmm assays, an additional 100 μL were inoculated in the stem, in a section adjacent to these leaves. The plants were incubated at 27 ± 1°C and infected leaves, uninfected leaves, stems and roots of 12 plants were sampled 6, 12, and 24 h after inoculation. Control plants were maintained in the same conditions after treatment with phosphate buffer only. Defoliation treatments were performed using A. cruentus plants cv. “Tarasca” as described before by Vargas-Ortiz et al. (2015) and Cisneros-Hernández et al. (2021). The choice of the “Tarasca” cultivar reflected its previous use as the model plant for ample defoliation-related research in grain amaranth (Castrillón-Arbeláez et al., 2012; Vargas-Ortiz et al., 2013, 2015). Briefly, the plants were subjected to severe defoliation at two different vegetative and reproductive development stages, respectively. Plants at the “vegetative 1” (V1) stage were utilized 4–5 weeks after germination (wag) and had 6-to-8 expanded leaves, whereas V2 plants were used at 6-to-7 wag and had 11–14 expanded leaves. Plants at panicle emergence (PE) were defoliated 9–10 wag, while those at the flowering stage (FL), were damaged 11–12 wag, when they had reached ~50% anthesis. The defoliation treatment was done with a punch-hole borer at a rate of 1/3 loss of foliar tissue area per diem for the first two days. On the third day, the leaf tissues remaining after ca. 66% percent leaf area loss, were cut from the base of the petioles to reach total plant defoliation and were immediately flash frozen in liquid nitrogen and stored for subsequent analysis. Twenty-four h later, leaves, roots, stems and panicles (when applicable) were sampled from intact plants (controls), whereas only roots, stems and panicles were collected from defoliated plants. All plant tissues/organs sampled were stored at −80°C until needed for analysis. Then, they were ground to a fine powder in liquid N₂ with frozen mortar and pestle.

Total RNA was extracted and subsequently purified from 100 to 200 mg of frozen and ground plant tissues as described previously (Palmeros-Suárez et al., 2015). The integrity of the purified RNA was assessed by visualizing the 18S and 28S RNA bands after electrophoretic separation in ethidium bromide-containing denatured formaldehyde/formamide 1% agarose gels. RNA purity and concentration were determined in a NanoDrop 2000 apparatus as instructed (ThermoFisher Scientific; Waltham, MA, USA). cDNA synthesis from 4 μg of total RNA was subsequently realized using 200 units of the reverse transcriptase SuperScript II enzyme as instructed by the manufacturer (Invitrogen, Carlsbad, CA, USA). The cDNA produced was diluted 25-fold with dd water prior to quantitative PCR (qPCR) analysis. All qPCR assays were performed using SYBR Green detection chemistry in a CFX96 Real Time System (Bio-Rad, Hercules, CA, USA). qPCR reactions were prepared using the primers listed in Supplementary Table S1, which were designed using the following programs: Primer3 (http://bioinfo.ut.ee/primer3/), Beacon Designer (http://www.premierbiosoft.com), UNAFold (http://www.idtdna.com/UNAFold) and Oligo Evaluator (http://www.oligoevaluator.com) (Thornton and Basu, 2011). Reaction mixtures, 20 μL total volume, contained 2 μL cDNA solution, 4 μL of each 2 μM oligonucleotide solutions, 8 μL de SYBR Green Jumpstart TAQ Ready Mix (Sigma-Aldrich St. Louis, MO, USA) and 2 μL of sterile dd water. The amplification process was performed within the following parameters: 15 min at 95°C to activate the Taq Polymerase, followed by 40 cycles of denaturation (95°C/15 s) annealing (60°C/1 min) and extension (72°C/30 s) and a final extension step (60°C/1 min). AhN4L-1 expression levels were determined in different tissues of grain amaranth plants subjected to (a)biotic stress. These were reported in relation to the expression levels in representative tissues of control plants, which were set to a value of 1. In transgenic A. thaliana, foliar AhN4L-1 transgene expression levels were calculated relative to trace background levels in WT plants. Relative gene expression was calculated using the 2^−ΔΔCT comparative cycle threshold method (Livak and Schmittgen, 2001). The genes employed as normalizing controls were the AhACT7 and AhβTubulin (for grain amaranth) and actin and EF-1α (for A. thaliana) housekeeping genes. In all cases, plant tissues for qPCR assays were obtained from combined pools of six plants. Each pool was subsequently subjected to three independent sampling procedures prior to analysis. qPCR data are reported as the mean of three repetitions ± SE of one representative experiment. qRT-PCR expression analyses were validated in two independent experiments.

The 1581 base pair (bp) open reading frame of AhN4L-1 was PCR amplified from A. hypochondriacus cDNA using specific primers (Table 1) and purified using a standardized procedure (Protocol kit; Qiagen, Germatown, MD, USA). Plasmids for expression of the AhN4L-1 under the control of the CaMV 35S promoter were assembled using the Gateway TA Cloning technology (Invitrogen). AhN4L-1 coding sequences were amplified as PCR products, having an additional deoxyadenosine at their 3′-ends, using cDNA obtained from RNA extracted from 9-day-old plant leaves from A. hypochondriacus L. cv. Revancha. The complete attB PCR product was obtained using, nested PCR using the primers listed in Table S1. The PCR program was run using the following steps: initial 5 min at 94°C; then, 30 cycles (30 s at 94°C, 30 s at 65°C and 1.5 min at 72°C), and a final elongation step for 10 min at 72°C. Subsequently, the fragments were inserted into the pCR8/GW/TOPO TA entry vector (Invitrogen). After sequencing of the entry clones, the LR reaction was performed for subsequent cloning in the pB7WG2D destination vector (Cambia, Canberra, Australia), under the control of 35S CaMV promoter. In both cases, spectinomycin resistance (at 50 μg/mL) was used for selection of transformed E. coli TOP10 cells. The binary construct was extracted and purified following the manufacturer's instructions (Invitrogen). After confirmation of correct orientation and size of the insert with the BglI, PstI, and XbaI restriction enzymes and subsequent sequencing (Instituto de Biotecnología de la UNAM, Cuernavaca, Morelos, México) it was electroporated into Agrobacterium tumefaciens strain GV2260. After a 48 h incubation period at 28°C on LB media spiked with rifampicin (25 μg/mL) and spectinomycin (100 mg/L), they were used to transform A. thaliana plants by the floral dip method (Clough and Bent, 1998).

Transformed plants were grown to maturity under controlled conditions, as described above, and the seeds were harvested at maturity. One-week old T1 transgenic A. thaliana plants were selected by growth in MS medium containing N-acetyl-L-phosphinothricin (PPT) (20 μg/mL), by monitoring the emission of GFP-related fluorescence under a Zeiss Primo Star fluorescence microscope (Carl-Zeiss GmbH, Oberkochen Germany) and by the PCR amplification of the GFP marker gene using specific primers (Table 1). T2 seeds were collected from individual transformants (T1) and plated again on the selection medium to determine PPT-resistant vs. PPT-sensitive plants segregation ratios. Homozygous T3 transgenic plants produced no PPT-sensitive seedlings from seeds of T2 plants. All further analysis of AhN4L-1 OE lines was performed using T3 plants homozygous for the transgene.

Seeds of WT controls and T3 homozygous A. thaliana transgenic plant lines differing in AhN4L-1 transgene dosage (i.e., L5, L6, and L17; Supplementary Figure S1), were germinated and grown as described previously (Massange-Sanchez et al., 2015; Palmeros-Suárez et al., 2015). Independent sets of OE transgenic and wild-type plants were grown for the different analyses performed.

A 1600 bp DNA fragment covering the 5'flanking region of the A. hypochondriacus AhN4L-1 gene was PCR-amplified from A. hypochondriacus genomic DNA. The primer pairs used in the PCR reaction are listed in Table 1. The PCR DNA amplicon was double-strand sequenced to verify its authenticity. Using the Gateway technology described above, it was inserted first into the pCR8/GW/TOPO TA vector and then into the pFAST-G04 vector harboring a GUS reporter gene. This vector was electroporated into A. tumefaciens GV2260 cells, which were selected by incubation for 48 h and 28°C on LB media supplemented with rifampicin (50 mg/L) and spectinomycin (100 mg/L). Transformed agrobacteria were subsequently used for A. thaliana transformation as described above.

Plasmids for expression of the AhN4L-1 sequence N-terminally fused to GFP under the control of the CaMV 35S promoter were constructed using the Gateway Cloning technology (Invitrogen), as above. Coding sequences were amplified as PCR products via nested PCR with the primers listed in Table S1. After the entry clone was validated by sequencing, it was inserted into the pk7WGF2 destination vector in order to fuse the GFP sequence N-terminally to AhN4L-1 (Karimi et al., 2002). These vectors were electroporated into A. tumefaciens GV2260 cells, which were selected and subsequently used for A. thaliana transformation as described above.

GUS histochemical staining was performed in diverse organs of transgenic A. thaliana plants (line 4, having the highest transgene expression; G. Cabrales-Orona, personal communication) harboring the pAhN4L-1::GUS fusion construct at different development stages (5 and 35 day-old plants). GUS staining buffer was prepared in 5 mM sodium phosphate buffer, pH 7.0, containing 40 mM potassium ferricyanide (K₃Fe(CN)₆), 40 mM potassium ferrocyanide (K₄Fe(CN)₆ and 0.004% Triton X-100. This buffer was stored at −20°C until required. Five μg/ mL of 5-bromo, 4-chloro, 3indolyl-β- D-glucuronide (X-Gluc) was dissolved in DMSO and added to the buffer. The different plant organs were incubated in X-Gluc solution for 16 h at 37°C. The visualization of the blue stained organ sections indicative of GUS activity was possible after repeatedly rinsing with buffer Z (40 mM NaH₂PO₄, 60 mM NaHPO₄, 10 mM KCl, 1 mM MgSO₄) and 70% ethanol, followed by chlorophyll removal with an acetone: methanol (1:3 v/v) solution (Jefferson et al., 1987). The images of blue-colored plant organs and plants at different developmental stages were recorded using a ZEISS Stemi 2000-C stereo microscope (Carl Zeiss, Jena, Germany). GUS-stained organs and plants shown in this study were obtained from a representative experiment, performed with the L4 transgenic plant line, that was repeated thrice with similar results.

Fluorescent images of roots and flowers were obtained from 3- and 35-day old transgenic plants (Line 34; Supplementary Figure S1), respectively, transformed with the N-terminal GFP::AhN4l-1 construct. They were recorded using a confocal laser scanning inverted microscope LSM 800 (Carl Zeiss, Jena, Germany). Propidium iodide (PI) at 0.01 mg/mL was used as a counterstain to highlight root cell morphology. GFP and PI were excited using the 488-nm and 514-nm lines, respectively, of an Argon laser. GFP emission was filtered with a 500–550-nm bandpass filter and PI emission was filtered with a 575-nm long pass filter. Images were captured using a Rebel T3 camera (Canon USA, Inc., Melville, NY, USA).

The effect on vegetative and reproductive growth and seed yield was determined in the three transgenic OE A. thaliana plants lines (L5, L6, and L17; Supplementary Figure S1) having differential expression levels of the AhN4L-1 gene. Germination rate was determined 3, 6, and 9 days after placing 100 seeds in Petri dishes (100 × 15 mm) containing sterile solid MS medium (0.5 × MS salt, 1% sucrose, and 1% agar, at pH 5.8). Root growth was similarly determined in 3, 5, and 7 days-old seedlings grown vertically on MS-containing 100 × 15 mm Petri dishes. Leaf number, leaf area, plant rosette area, inflorescence height and seed yield and weight were also determined in 5–6 weeks-old transgenic A. thaliana plants. The onset of flowering was also recorded. To measure yield, dry seeds were harvested manually from fully matured plants, approximately after 8 weeks of growth. Seed weight was determined as the measure of 1,000 seeds. All experiments were performed in the A. thaliana conditioned growth room, as described above.

The L5, L6, and L17 lines were used to perform the stress experiments, as follows. Twenty 3-weeks old plants of each transgenic line were grown individually in single 125 mL plastic pots containing the A. thaliana soil mixture described above. All plants were kept in the conditioned growth room maintained at constant temperature of 22°C and under a 16 h light/8 h dark photo-period. Salt stress was applied by watering the plants with 20 mL each of a 400 mM NaCl solution on day 1 and with an identical volume of a 100 mM NaCl solution on day 3. The soil salinity at this point (8.2 dS/m), measured according to the NOM-021-RECNAT-2000 environmental norm (DOF, 2002) with a Hanna HI 98130 conductometer (Hanna instruments Inc., Woonsocket, RI, USA) corresponded to moderate to high saline soils (Artiola et al., 2019). In planta determination of photosynthetic parameters (see below) were performed immediately after the application of the second 100 mM NaCl solution, whereas the foliage and roots of all plants, divided into 2 pools of 10 plants each, were collected 3 weeks later. The plant tissues were flash frozen with liquid N₂ and stored at −70°C until needed for analysis. Soluble and insoluble non-structural carbohydrate (NSC) content, as well as the levels of chlorophyll, carotene, total phenols and flavonoids, in addition to the catalase and superoxide dismutase enzymatic activities (see below) were determined in control and salt-stressed plants. The water deficit stress assays were performed according to Massange-Sanchez et al. (2015) and Palmeros-Suárez et al. (2015).

A portable CIRAS-3 equipment (PP Systems, Amesbury, MA, USA) was employed to measure CO₂ assimilation (A) and transpiration (E) rates, respectively, in addition to stomatal conductance (gs) and water use efficiency (WUE). The equipment parameters were set at 380 ppm CO₂, 55% RH and a PAR of 1,000 μmol m⁻² s⁻¹. The Fv/Fm measurement of the photosystem II efficiency in a dark-adapted state was determined in A. thaliana plants kept in the dark for 30 min using a portable chlorophyll fluorometer (Pocket PEA Chlorophyll Fluorimeter; Hansatech Instruments; Norfolk, England).

Total chlorophyll content in A. thaliana leaf tissues was determined after extraction with cold 80% acetone according to Arnon (1949). Briefly, 200 mg of frozen leaf tissue was extracted twice with 800 μL of 80% acetone for 1 h at 4°C and centrifuged. Both extracts were combined to spectroscopically measure total chlorophyll levels in a UV-visible xMark microplate absorbance spectrophotometer (Bio-Rad) by recording the absorbance at 645 (A₆₄₅) and 663 nm (A₆₆₃). An acetone blank solution was used to calibrate the apparatus. Total chlorophyll was determined with the following equation: 8.02 × A₆₆₃ + 20.2 × A₆₄₅. Total carotene content was calculated according to Biehler et al. (2010) with some modifications. Thus, 100 mg of frozen leaf tissue, previously ground to a fine powder in liquid N₂, together with 20 mg of calcium carbonate, was extracted with 400 μL metanol and 100 μL of 30% KOH in methanol for 15 min in the dark. The resulting extract, clarified by centrifugation and mixed with 1.5 mL of a saturated NaCl solution, was extracted twice with 500 μL of an hexane:chloroform (1:1 v/v) solution. The combined extracts were dried overnight by vacuum centrifugation (Maxi-Dry Lyo; Heto-Holten, Denmark) and re-dissolved in 0.5 mL acetone. Total carotenes were determined in a UV-visible spectrophotometer, as above, by recording the absorbance at 450 nm.

Soluble (i.e., glucose [Glc], fructose [Fru] and sucrose [Suc]) and insoluble (i.e., starch) NSC contents in lyophilized ground tissues were determined enzymatically using a coupled assay as described previously (Cisneros-Hernández et al., 2021). Proline was assayed colorimetrically, at 520 nm, via the measurement of the red chromogen produced by its reaction with ninhydrin at low pH (Bates et al., 1973). Total phenols and flavonols were determined according to colorimetric methods described before by Maranz et al. (2003) and Sakanaka et al. (2005), respectively. Catalase (CAT; E.C. 1.11.1.6) and superoxide dismutase (SOD; E.C.1.15.1.1) activities were determined using commercially available kits as described before (Palmeros-Suárez et al., 2015).

These experiments were performed with Pseudomonas syringae pv. tomato DC3000 (Pst3000) as described by Yao et al. (2013). Briefly, bacterial glycerol stocks were removed from freeze-storage at −70°C, were plated on Luria-Bertani (LB) agar and were incubated at 28°C for 48 h. A single colony was used to inoculate 5 mL of liquid LB media which was incubated at 28°C for ≈ 12 h until the bacterial growth reached an OD₆₀₀ of between 0.8 and 1.0. The culture was centrifuged at 5,000 rpm for 10 min at room temperature and the resulting bacterial cell pellet was resuspended in sterile dd water to an OD₆₀₀ of 0.002 which was equivalent to 1 × 10⁶ colony forming units/mL (CFUs/mL). Three fully expanded leaves of WT and transgenic A. thaliana plants (five 20-day old plants per genotype: WT, L5, L6, and L17, respectively) were each inoculated, at their abaxial surface, with 10 μL aliquots of the Pst3000 bacterial suspension. An identical number of control plants were similarly inoculated with sterile dd water. The plants were subsequently incubated for 3 days at 28°C. At this time, photographic register was obtained and leaves were sampled to determine the bacterial titers as CFUs/cm². The latter were obtained as follows: the leaves were surface sterilized by submersion in a 15% ethanol solution for 15 s and rinsed by dipping in sterile dd water for 10 to 15 s. Leaf discs, 0.5 cm², were then obtained from the leaves using a cork borer. Each disc was then placed in a 1.5 mL Eppendorf tube containing 100 μL of sterile dd water, where the leaf tissue was thoroughly macerated. The macerated tissue was diluted with the addition of 900 μL dd water and was vortexed. Serial 10 × dilutions of the extracts were produced by adding 100 μL aliquots in 900 μL of dd water and 10 μL of each were plated on LB agar plates. The plates were sealed with parafilm material and incubated at 28°C for 24 to 30 h until colonies were visible. These were counted in order to calculate the number of UFCs/cm² of leaf tissue. A total of 3 leaf discs per plant per four plants were analyzed to reach the final bacterial counts. The experiment was repeated thrice.

Agar squares containing spores of F. oxysporum f. sp. lycopersici were placed on freshly prepared potato-dextrose agar plates and incubated at 28°C for 7 days to obtain dense fungal growth. Spores collected from these cultures were resuspended in 2 mL of sterile dd water. Spore counts were performed using a hemocytometer, followed by dilution with sterile dd water to reach a final count of 10⁶ spores/mL. Volumes of this suspension (10 mL) were employed to infect the A. thaliana plants by their direct application to the soil as drench solutions. A total of twenty 20-day old plants (5 per genotype: WT, L5, L6, and L17) per assay were employed. The plants were incubated at 28°C for 15–25 days, when the disease symptoms started to be apparent. Photographic register of the plants and roots was obtained 25 days after infection. The fresh weight of the shoot and roots was also determined. Root samples were further processed for DNA extraction which was subsequently used for PCR amplification of Fusarium oxysporum-specific ITS amplicons. Root DNA was extracted using the urea-phenol method described by Shure et al. (1983). Briefly, 100 mg of frozen root tissue was ground in liquid N₂ and mixed with 700 μL of an extraction buffer composed of 7 M urea, 0.35 M NaCl, 50 mM Tris-HCL (pH 8.0), 20 mM EDTA, 0.5% sarcosyl, 0.25 mL of molecular grade phenol and 0.2 mL of a 20% SDS detergent solution. The mixture was homogenized by vortexing at room temperature and extracted with 700 μL of a phenol: chloroform; isoamyl alcohol (50:48:2 v/v/v) solvent mixture. It was centrifuged at 12,000 rpm for 10 min at 4°C. The upper aqueous phase was transferred to a fresh tube where 400 μL of cold isopropanol and 40 μL of a potassium acetate solution were added and mixed by gently inverting the tube. The tube was centrifuged again at 12,000 rpm for 10 min. After discarding the supernatant, the resulting pellet was washed with 1 mL of a 99% ethanol solution and centrifuged again at 12,000 rpm for 3 min. The washed DNA pellet was dissolved in 100 μL of sterile dd water, spiked with 1 μL of ARNase (Invitrogen) and incubated at 37°C for 30 min. The resulting DNA was used as template for a PCR reaction performed using the ITS1 and ITS4 primers listed in Table 1. The PCR amplification program used was the following: denaturing at 94°C/3 min followed by 30 cycles of 30 s/94°C, 30 s/60°C and 1 min/72°C. The PCR products were separated via 1% agarose gel electrophoresis and the resulting bands, visualized with ethidium bromide, were photographed using a photo-documenter (Bio-Rad).

All the experiments were performed using a completely randomized design with 20 plants per genotype. Data generated were analyzed using a one-way analysis of variance (ANOVA) to examine the effect of different stress conditions on the expression levels of the AhN4L-1 gene in grain amaranth plants and of the overexpression of this gene on various physiological variables in A. thaliana plants. The means were compared using the Tukey-Kramer test to identify statistically significant differences between them. The statistical analysis was performed with the aid of R (http://r-project.org/) Rstudio (https://www.rstudio.com) and JMT Pro13 (jmp.com) statistical software.

A gene coding for an uncharacterized A. hypochondriacus Nattterin-4-like1 protein (AhN4L-1) possessing an AAT domain arrangement was identified and characterized. The AhN4L-1 protein contains two successive agglutinin (A) domains linked to a single toxin domain (or T domain), with sequence similarity to aerolysin, the pore-forming toxin produced by Aeromonas sp. It differs from the amaranthin-like AAT proteins recently reported in Caryophillales, including A. hypochondriacus, and several other monocot and dicot plant orders (Dang et al., 2017). Sequence alignment of the mRNA and the genomic contig sequences revealed the 10754 bp gene has three exons of 394, 795, and 392 bp, respectively, that code for a 526 aa protein (Supplementary Figure S2). It has a homolog of 518 aa (AhN4L-2 protein Phytozome accession number: AHYPO_022596-RA CDS; Supplementary Figure S2). The A domains of the AhN4L-1 protein are localized at positions 14-102 and 175-274, respectively, whereas the T domain is present in position 294-423 (Figures 1A,C). Bio-informatic analyses (https://zhanglab.ccmb.med.umich.edu/I-TASSER/output/S553256/) predicted a 3D structure with 2 agglutinin-like domains within the N-terminal half of the protein, and one aerolysin-like domain at the C-terminal end (Figure 1B). A phylogenetic tree (Figure 2) showed that the AhN4L-1 and AhN4L-2 proteins from A. hypochondriacus are grouped together with other uncharacterized N4L proteins present in Beta vulgaris and Chenopodium quinoa. Blast and Blast4OneKp alignments retrieved a maximum resemblance of 43% with the LOC104887222 protein from B. vulgaris. No homologous sequences were detected in A. thaliana. The tree indicated lower homology levels with ATT and Nattering-like proteins identified in other monocot and dicot plants species.

Previous transcriptomic data of AhN4L-1 in A. hypochondriacus plants revealed that this gene was strongly induced by water-deficit and salinity stress and by chewing insect herbivory (Délano-Frier et al., 2011). These results were corroborated by qPCR data obtained from plants of the three grain amaranth species subjected to different stress conditions (Table 1). Thus, its responsiveness to insect herbivory was confirmed by its strong, rapid and stable induction in leaves damaged by chewing insect herbivory in all three species, particularly in A. cruentus. A weaker systemic expression of this gene was also observed in A. cruentus and A. hypochondriacus but was absent in A. caudatus. Bacterial infection with virulent Cmm infection predominantly induced this gene in the stems of the three grain amaranth species 12 h after infection whereas treatment with avirulent Pss induced a rapid (6 h) expression of this gene in most organs of A. hypochondriacus, except roots, and an intermediate to late (12 to 24 h) transcript accumulation in the roots of the three grain amaranths. Salinity led to a predominant expression of this gene in stems and roots, while moderate WDS and subsequent recovery induced its expression in leaves and roots of A. cruentus cv. Dorada plants. Severe defoliation of A. cruentus cv. Tarasca plants also led to the induction of this gene, mostly in leaves and roots and during V1 and V2 vegetative stages and during PE. The expression of this gene was limited to the stem when A. cruentus plants were defoliated at FL.

To investigate the promoter activity of the AhN4L1 gene, a genomic DNA fragment (1.6 kbp) containing the 5′ flanking region of this gene was cloned after PCR amplification. This promoter region was fused to the E. coli uidA reporter gene encoding a GUS enzyme and introduced into A. thaliana through Agrobacterium-mediated gene transfer. Transgenic T3 A. thaliana L4 plants containing the homozygous recombinant transgene were subjected to histochemical staining for GUS activity. An in silico analysis of this promoter region segment unveiled the prominence of five elements (AHBP, DOF, GTBX, L1BX and MADS) involved in development and (a)biotic stress responses. These were also prominent in the promoter region of the AhN4L-2 gene. Interestingly, four cis-regulatory elements associated with abscisic acid (ABA) and ethylene signaling, root and apical meristem regulation and control of storage proteins synthesis, were only found in the promoter region of the AhN4L-1 gene (Supplementary Table S2). The prevalent elements present in the AhN4L-1 promoter were in agreement with the gene expression patterns observed in these transgenic plants (Figure 3). Thus, the detection of GUS staining along the stem in young germinated plantlets (Figure 3A) and in developing root tips (Figures 3C,D) concurred with the regulation of development-related functions by genes rich in AHBX and GT-box elements, as shown in Supplementary Table S2. Likewise, the expression in developing flowers, particularly in the anthers (Figure 3E), pollen (Figure 3G), and segments of the pistil (Figure 3F), coincided with the control of flower development, including flowering structures and floral meristems, that characterizes genes rich in L1- and MADS-boxes (Supplementary Table S2).

To investigate the subcellular localization of the A. hypochondriacus AhN4L-1 protein, N-terminal GFP fusion constructs were generated and expressed in A. thaliana plants. The fluorescence emitted by the cells of the transgenic plants was analyzed using confocal microscopy. The localization of the green fluorescence emitted by A. thaliana transgenic plants able to constitutively express GFP- AhN4L-1 fusion proteins coincided with the GUS-activity detected in growing root tips (Figures 4C,D) and in segments of the pistil, around the style and pollen tube (Figures 4O,P), and in the anthers (Figures 4S,T). A GFP signal was also observed in the stomata of leaves (Figures 4K,L), suggesting a role in the regulation of stomatal movement, and in the root vasculature (Figures 4G,H). The images indicated that these proteins appeared as punctate agglomerates, perhaps in the periphery of protein bodies and/or the cell wall-protoplast interface. Equivalent images obtained from root tips (Figures 4A,B), roots (Figures 4E,F), leaf surface (Figures 4I,J), pistils (Figures 4M,N) and anthers (Figures 4Q,R) of untransformed control plants are shown.

Three A. thaliana transgenic plants lines with differing AhN4L-1 expression levels (high [L17], middle [L5], and low [L6]) were selected for experimentation; see above). Various physiological parameters were affected in these plants. The germination percentage of seeds generated by the three transgenic plant lines, measured at 3, 6, and 9 days after placing on germinating media, was significantly lower than untransformed controls (Figure 5A). Although the number of leaves remained unchanged, they were significantly smaller in 30-day old transgenic OE lines (Figure 5B). The inflorescences were also significantly smaller than control plants (Figure 5C) whereas seed weight (Figure 5D), but not seed yield (Figure 5E), were generally lower in the transformed plants. The advent of flowering was similar in WT and transgenic plants. Root length was also significantly reduced in the transgenic A. thaliana plants (Figure 6). Compared to WT plants, the transgenic plants showed no increased tolerance to WDS or high salinity stress (Supplementary Figure S3 and G. Cabrales-Orona, personal communication). Despite the fact that AhN4L-1 overexpressing plants showed no increased salinity tolerance, they nevertheless maintained their significantly basal higher chlorophyll content plants after 21-days of saline stress. However, carotenes were mostly reduced in response to salt-stress, particularly in the L6 and L17 transgenic lines. The accumulation of proline and Suc in response to salt stress was also less pronounced in these transgenic plant lines, whereas the L17 line showed significantly higher Glc and Fru, but lower starch levels, in response to salt stress. No change in the total phenols and total flavonoids content due to salt stress, which were similar to those in WT plants, was observed in all three transgenic plant lines examined (Supplementary Figure S4). Also, no changes in CAT and SOD activity levels were observed between WT and transgenic plants kept under control and salinity stress conditions (Supplementary Figure S5). On the other hand, all the photosynthetic parameters tested (i.e., A, E, gs and WUE) remained above zero in the L6 and L17 transgenic lines under salt stress conditions, whereas, in WT plants, E and gs were significantly lower and A and WUE became negative (Supplementary Figure S6). Fv/Fm variations in response to stress were similar in WT and AhN4L-1 overexpressing plants (G. Cabrales-Orona, personal communication).

In contrast to the abiotic stress assays, the overexpression of AhN4L-1 in A. thaliana plants conferred resistance to both biotrophic bacterial and hemi-biotrophic fungal pathogens in a dose-dependent manner, since it was mostly observed in the L17 line having the highest transgene expression levels, Thus, counts of Pst3000 CFUs were significantly lower after 3-day post-inoculation periods in this transgenic plant line (Figure 7A), whose plants were also less developmentally affected by bacterial infection (Figure 7B). Likewise, L17 transgenic line plants were significantly less susceptible to F. oxysporum fungal infection. They showed better and healthier vegetative and reproductive development (Figures 8A,B), characterized by less affectation of shoot and root growth (Figures 8C,D). Moreover, the pathogen load was lower, as determined by PCR amplification of the ITS marker gene sequence (Figure 8E).