The cassava (Manihot esculenta) variety (“Huanan5”) was chosen as the wild type during this experiment. Cassava seedings were grown on half-strength Murashige and Skoog (MS) medium at 26°C with a 16 h light and 8 h dark cycle and 70% relative humidity conditions in a growth chamber. The tobacco (Nicotiana benthamiana) plants were also grown in a greenhouse under cycles of 16 h light and 8 h dark at 25°C. The rice (Oryza sativa) seedlings were grown in controlled conditions of 16 h light (30°C) and 8 h dark (28°C) photoperiods with ~70% relative humidity. For the overexpression experiment, the coding sequences of the MeSLAH4 gene were amplified and inserted into pCAMBIA2300-35S-eGFPvector. These constructs were subsequently transferred into an Agrobacterium strain (EHA105) for rice transformation. For the hydroponic culture, the plants were grown in a nutrient solution containing 1.5 mm NH₄NO₃, 0.3 mm NaH₂PO₄, 0.3 mm K₂SO₄, 1.0 mm CaCl₂, 1.6 mm MgSO₄, 0.5 mm Na₂SiO₃, 20 μm Fe-EDTA, 18.9 μm H₃BO₃, 9.5 μm MnCl₂, 0.1 μm CuSO₄, 0.2 μm ZnSO₄, and 0.39 mm Na₂MoO₄, but supplied with different N concentrations, termed as normal nitrogen NN (0.5 mm NH₄NO₃) and low nitrogen LN (0.01 mm NH₄NO₃), pH 5.5. The nutrient solution was changed every 3 days.

The protein sequences of SLAH genes (gene name; locus identifiers) of Arabidopsis, including AtSLAH1 (AT1G12480), AtSLAH1 (AT1G62280), AtSLAH2 (AT4G27970), AtSLAH3 (AT5G24030) and AtSLAH4 (AT1G62262), were downloaded from the TAIR database¹ (Swarbreck et al., 2007). The SLAH protein sequences of genes in rice, such as Os01g0623200 (LOC_Os01g43460), Os01g0385400 (LOC_Os01g28840), Os05g0219900 (LOC_Os05g13320), Os07g0181100 (LOC_Os07g08350), Os01g0226600 (LOC_Os01g12680), Os04g0574700 (LOC_Os04g48530), Os01g0247700 (LOC_Os01g14520), Os05g0269200 (LOC_Os05g18670), and Os05g0584900 (LOC_Os05g50770) were downloaded from the RAP_DB database.² The SLAH protein sequence of M. esculenta (Cassava), including MANES_05G153100 (SLAH4), MANES_11G124900 (SLAC1 homolog 1), MANES_14G020300 (SLAC1 homolog 3), MANES_06G154500 (SLAC1 homolog 3), MANES_06G154600 (SLAC1 homolog 3), and MANES_S089100, along with their homologues in other species such as Zea mays (Corn), Triticum aestivum (Wheat), B. napus (Rapeseed), Selaginella moellendorffii (Spikemoss) etc., were downloaded from the Phytozome database³ (Nan et al., 2021). The phylogenetic trees were constructed based on the SLAC/SLAH protein sequences by IQ-TREE using the Maximum Likelihood (ML) method with 1,000 replicates of bootstrap alignments. RNA-seq data and differential gene expression information were obtained from a published database.⁴

The chromosomal localization information of the SLAC/SLAH genes was obtained from sequences of the cassava genome, and the MG2C⁵ was used to draw the chromosomal distribution of MeSLAH genes.

The structures of the SLAH genes were analyzed using the Gene Structure Display Server (GSDS 2.0)⁶ by aligning the cDNA sequences with their corresponding genomic DNA sequences. Conserved motifs of the SLAH proteins were identified using the online Multiple Expectation Maximization for Motif Elicitation (MEME⁷; Bailey et al., 2006). All obtained SLAH protein sequences were analyzed against the Pfam database to verify the presence of SLAC1 domains (Supplementary Figures 2, 3). The SLAC1 domain was also detected by the SMART program (SMART).⁸ Protein sequences lacking the SLAC1 domain or having E-values of more than 1 e-6 were removed.

Rice protoplasts were isolated for transient transformation of the MeSLAH4 gene (Zhang et al., 2011; Burman et al., 2020). The open reading frame (ORF) of the MeSLAH4 gene was amplified by PCR (95°C for 5 min, then 35 cycles of 95°C for 30 s, 58°C for 30 s, and 72°C for 70 s, with a final extension at 72°C for 5 min). The PCR products (Supplementary Figure 4) were cloned into the 35S-eGFP vector in between the XhoI and HindIII sites and under the control of the cauliflower mosaic virus 35S promoter. The competent cells of Escherichia coli (DH5α) and Agrobacterium (LBA4404) were used for the transformation of recombinants. Listed primers (Supplementary Table 2) for gene cloning and vector construction and plasma membrane marker (35 s-ZmCDPK7-MCHERRY and 35S-HY5-Mcherry; Zhao et al., 2021) were used in this study. The MeSLAH4-eGFP and 35 s-ZmCDPK7-MCHERRY fusion constructs were transiently expressed in rice protoplasts using a polyethylene glycol calcium-mediated method. An empty 35S-GFP vector was served as a negative control. Transfected protoplasts were observed after 16 h of incubation by a confocal laser scanning microscope (Olympus FV3000, Tokyo, Japan).

The morphological, physiological, and agronomic traits of each transgenic rice line (OE) and wild type (WT) was measured at the 2-week-seedling stage and the maturity stage. The morphological characters, including seedlings height (cm) and root length (cm), were measured using a ruler or meter stick. The grains were lined up lengthwise and widthwise along a ruler to measure grain length (mm) and grain breadth (mm), respectively, and the measurements were confirmed using an MRS-9600TFU2L (MICROTEK) grain observation instrument. The shoot weight (g. plant⁻¹ FW), root weight (g. plant⁻¹ FW), and grain yield of a single spikelet (g) were measured using a weighing balance. The root fork numbers (number of branches), root tip numbers, and grain numbers of a single spikelet were calculated with the eye and confirmed by capturing an image with an Epson Expression 10000XL (Epson, Japan) and counting with winRHIZO software (Li et al., 2016a). The root average diagram (mm), root surface area (cm²), root volume (cm³), and grain diameter (mm) measurements were performed using a microscope (MVX10, Olympus), and the winRHIZO scanner-based image analysis system (Regent Instruments, Montreal, QC, Canada; Sun et al., 2014). Total grain protein content (%), and grain moisture content (%) were detected using an XDS Near-Infrared Rapid Content Analyzer (Foss® Analytical, Hilleroed, Denmark; Li et al., 2014).

The chlorophyll was extracted from 0.15 g of fresh leaves at the booting stage using 95.0% ethanol. Briefly, leaves were cut into 3 mm pieces and immersed in 95.0% ethanol for 24 h in the dark at 26°C. The absorbance of the extract was measured using a spectrophotometer at A665 and A649. The chlorophyll-a content (mg. g⁻¹ FW), chlorophyll-b content (mg. g⁻¹ FW), and total chlorophyll content (mg. g⁻¹ FW) contents were determined by the method reported by Arnon (1949). A total of 10 individuals of each transgenic line and wild type plants were assayed (Li et al., 2016b).

The activities of catalase (CAT), peroxidase (POD), and superoxide dismutase (SOD) were measured by employing 0.5 g of seedlings in 5 ml of extraction buffer containing 0.05 M phosphate buffer (Li et al., 2018). The CAT activity was determined spectrophotometrically based on the decrease in absorbance of H₂O₂ (extinction coefficient of 43.6 M⁻¹ cm⁻¹) at 240 nm for 1 min (Aebi, 1984). The POD was measured as the absorbance at 470 nm. The SOD activity was assayed by measuring the ability of the enzyme extract to inhibit the photochemical reduction of nitroblue tetrazolium (NBT; Yoshimura et al., 2000). Phenotypic measurements of the positive transgenic plants were undertaken using three independent lines at least (Li et al., 2014).

Total RNA was isolated from different tissues (leaves, stems, and roots) of cassava and rice using an RNA kit (TRNzol universal reagent, TIANGEN Biotech, Beijing, China) according to the manufacturer’s instructions. The RNA was then reverse transcribed into cDNA using the oligo (dT) primers and ImProm-II reverse transcriptase (Promega). The specific primers for the MeSLAH4 genes along with the housekeeping Actin genes were designed using the Primer Premier 5.0 software (Supplementary Table 2). The real-time PCR reactions were conducted with 2 μl of diluted cDNA, 200 nM of each primer, 2× SYBER GREEN Master Mix (Green I Master Mix, Roche) in a final volume of 20 μl of double sterile water. The thermal cycle conditions were pre-incubation at 95°C for 5 min, then 40 cycles of 95°C for 3 s, 60°C for 10 s, and 72°C for 30 s, with a final extension at 72°C for 3 min in the Light Cycler 480 (Roche, United States). The gene expression levels were calculated with the 2^−ΔΔCt method (Livak and Schmittgen, 2001), and the qRT-PCR assays were performed with three biological and three technical replicates.

Fresh samples (whole plant, grain, and glume) of WT or transgenic lines were harvested at the rice mature stage (n = 4) and heated at 105°C for 30 min. The samples were then dried for 3 days at 75°C. Dry weights were recorded as biomass values. Total N accumulation was assessed using the Kjeldahl method in the different plant samples by multiplying the N concentration with the corresponding biomass.

The data from the experiments were analyzed by one-way ANOVA, Duncan’s multiple range test, and Tukey’s test at p < 0.05 to determine the statistically significant differences among different treatments. All the statistical evaluations were performed using SPSS version 20.0 statistical software (SPSS Inc., Chicago, IL, United States) and MS-Office 2019 software.

In the present study, a close phylogenetic relationship of entire SLAH genes in cassava with previously reported SLAH genes in another species was detected (Figure 1A) by analyzing with a Hidden Markov Model (HMM) profile search along with a conserved model (SLAC1, PF03595) for the SLAH proteins. The phylogenetic tree displayed seven clades (I to VII), and a very close relationship among the six SLAH genes in the cassava genome, five SLAH genes in Arabidopsis, and nine SLAH genes in rice were identified. However, two SLAC1 gene homologues (Os05g0269200 and Os01g0247700) in rice and two SLAC1 homologues (AT1G62280, SLAH1; and AT1G62262, SLAH4) in Arabidopsis were detected in the same clade (Clade II) as the MeSLAH4 (MANES05G153100) gene. Moreover, the MeSLAH4 protein demonstrated about 49.40% similarity with the Os05g0269200 amino acid sequence (Supplementary Table 1; Supplementary Figure 1). Besides, the phylogenetic tree using IQ-TREE of SLAH protein also revealed an evolutionary relationship with other species, including Chondrus crispus (Carrageen Irish moss), Physcornitrella patens (Bryophyta), Marchantia polymorpha (Liverwort), Amborella trichopoda, Z. mays (Corn), T. aestivum (Wheat), B. napus (Rapeseed), and Selaginella moellendorffii (Spikemoss).

RNA-seq data which are available on the database represents diverse expression pattern of SLAH genes in cassava, and the SLAH genes are expressed differentially in different tissues of cassava under different nitrogen conditions (Figure 1B). In particular, the MeSLAH4 (MANES_05G153100) gene is highly expressed in the root under free nitrate concentration (FN, 0 mmol/l, around 1.2-fold) as well as at high nitrate concentration (HN, 10 mm/l, approximately 1.08-fold). Another MeSLAH1 homologue 3 (MANES_06G154600) is also expressed in the roots of cassava but a little bit lower (FN, about 1.08-fold, and HN, roughly 0.96-fold) than the MeSLAH4 gene.

The relative expression pattern of the MeSLAH4 gene in different tissues of cassava revealed variations in expression levels at various concentrations of nitrate levels (Figure 1C). The transcript accumulation patterns that were analyzed in roots, stems, and leaves indicated that the MeSLAH4 gene was mainly expressed in the root. A significantly higher expression (about 370-fold) was observed in the root under a lower concentration of nitrate (0.2 mmol/l) treatment (Figure 1C). Thus, the MeSLAH4 gene exhibited significantly higher relative gene expression levels at lower nitrate concentrations in the root, indicating a potential role in enhancing nitrogen use in plants.

As the root is the principal organ for nutrient uptake in plants, the up-regulation of the MeSLAH4 gene could correlate the relationships among different nitrate concentrations with plant growth and development.

The SLAH genes are distributed on four chromosomes in cassava (chromosomes 5, 6, 11, and chromosome 14). In cassava, one SLAH gene, which has been identified as MeSLAH4 genes, was present on chromosome 5, two SLAH genes were located on chromosomes 6, one gene on chromosome 11, and one gene on chromosome 14 (Figure 2A). The SLAH genes in Arabidopsis are detected on three chromosomes (three genes on chromosome 1, one gene on chromosome 4, and one gene on chromosome 5). In rice, four SLAH genes were found on chromosome 1, one gene on chromosome 4, two genes on chromosome 5, and one gene on chromosome 7 (Figure 2A).

Analysis of the upstream promoter region of SLAH genes represented transcriptional regulation mechanisms. About 2 kb upstream of the initiation codon of SLAH genes of Arabidopsis, cassava, and rice were obtained and submitted to the Plant CARE database for investigating cis-regulatory elements. A total of 9 different cis-elements associated with light responsiveness, stress responsiveness, phytohormone responsiveness and growth regulation have been identified in upstream regions of SLAH genes (Figure 2B).

A linear line has been constructed to present regulatory elements in each corresponding gene (Figure 2B). These results indicate that complex regulatory networks may be implicated in the transcriptional regulation of SLAH genes in different plants. Cis-regulatory elements, CAAT-box was commonly shared by all SLAH genes. G-box elements responding to light existed in the 2-kb upstream region of SLAH genes. Most SLAH genes contain ABRE elements (ABA responsive), but they are absent in the cassava MeSLAH4 gene which suggested that this gene might not involve in regulation and physiological responses of various processes, including stomatal closure, seed and bud dormancy. Moreover, SLAH genes harbored drought responsive cis-elements (DRE) that were not present in the MeSLAH4 gene. The MeSLAH4 gene contains several copies of the CAAT-box and a copy of the G-box, indicating a higher transcription rate with sufficient quantities of suitable binding sites for several transcription factors as well as a highly conserved sequence for evolutionary process and epigenetic regulation.

The transiently expressed fusion protein driven by the 35S promoter through protoplast transformation of the MeSLAH4 gene represented a clear subcellular localization in rice (Figure 3). The green fluorescent signal of eGFP, which represented a negative control, was observed in the cytoplasm. However, the signal of the MeSLAH4-eGFP fusion protein, which coincides with the red fluorescent signal of plasma membrane-localized protein, was detected in the plasma membrane and in the nucleus. The protein localization was further confirmed by the protoplast transformation, which indicated that MeSLAH4 proteins are localized in the plasma membrane and in the nucleus. The microscopic visualization exhibited that the green fluorescence was distributed throughout the whole cell when the control (empty) vector was used. The green fluorescence was exclusively detected on the plasma membrane and nucleus by confocal microscopy when the vectors contained MeSLAH4 (Figure 3). These results indicate that the MeSLAH4 gene may be involved in other functions in the plants.

In the current experiment, the plant phenotype exhibited higher overall growth in OE lines under both nitrated concentrations (0.5 mm NH₄NO₃, NN, and 0.01 mm NH₄NO₃, LN) compared to the WT (Figure 4A). Plant height was significantly different (p < 0.01) under both nitrated concentrations (NN and LN) in both WT and OE lines, but they demonstrated higher plant height at LN compared to NN (Figure 4B). The shoot and root weight exhibited non-significant differences in both WT and OE lines under both nitrate concentrations (LN and NN; Figures 4C,D). However, shoot weight was higher (0.75 g. plant⁻¹ FW) in OE lines at LN compared to both lines (WT and OE) in NN condition. Conversely, the root weight was higher in the WT (4.8 g. plant⁻¹ FW) under NN than under LN. These results point out that the overexpression of the MeSLAH4 gene enhances aboveground biomass (plant height and shoot weight) but decreases the lower ground parts (root weights) under low nitrate conditions.

The chlorophyll-a content was non-significantly different at LN but significantly different (p < 0.05) under NN while it was higher at NN condition (Figure 4E). Conversely, the chlorophyll-b content was significantly different (p < 0.05) under LN but a non-significant difference was observed under NN, while the OE lines showed higher chlorophyll-b content compared with WT and OE lines under LN conditions (Figure 4F). The OE lines showed higher total chlorophyll content under LN conditions, which also demonstrated significant differences (p < 0.05) compared to the WT (Figure 4G). Higher chlorophyll content in OE lines under LN indicates the increasing nitrogen use efficiency and higher conversion of photosynthesis by the plants under low nitrate concentration.

The CAT activity was significantly different under the LN (p < 0.01) and NN (p < 0.001) conditions, but both lines (WT and OE) demonstrated higher activity under the NN condition, where WT had more activity than OE lines (Figure 4H). The POD activity of the OE line was higher than wild type (WT) under both (LN and NN) conditions, but it was highly significant (p < 0.0001) under the NN condition (Figure 4I). The SOD activity was complicated because the WT demonstrated higher SOD activity (650 U.g⁻¹min⁻¹) under the LN condition while the OE lines exhibited higher activity (700 U.g⁻¹min⁻¹) under the NN condition (Figure 4J). Lower CAT and SOD activity of OE lines in LN conditions indicates higher photosynthetic and stress-responsive activities, while the activity of POD in OE lines under both (LN and NN) conditions demonstrates a higher ability to scavenge hydrogen peroxide under prolonged nitrated conditions.

In the field trial, the grain length (Figures 5A,C) and breadth (Figures 5B,D) of transgenic rice were increased significantly relative to the wild type, and the highest increment was observed in the 35S:MeSLAH4 OE-4 line. The relative gene expression of this line (35S,MeSLAH4 OE-4) was higher (about 80-fold, Figure 5E) and the panicle morphology (Figure 5F) was better than the wild type, hence the OE-4 lines have been selected for all other experiments. The grain numbers of a single spike (around 60) in the OE lines have been increased (Figure 5G). These results indicate that the OE lines have up-regulated MeSLAH4 gene expression, which has facilitated higher assimilation rates and more storage in the grain compared to the wild types.

As shown in Figure 5, all other traits of the OE line, including grain protein content (12%), grain moisture content (14%), grain diameter (4.8 mm), and grain yield of a single spike (1.48 g), were higher compared to the WT. Thus, the results of this experiment demonstrated that overexpression of the MeSLAH4 gene increases protein content with higher grain expansion and yield in rice.

The root system indices (phenotype of roots) demonstrated changes in terms of size and shapes in the wild type (WT) and overexpression (OE) lines under different nitrate conditions (Figure 6A). The root fork numbers (number of branches) and root tip numbers were increased (580 and 490, respectively) under the LN condition, and the increment was higher in the OE line compared to the WT line (Figures 6B,C). These results indicate that a lower nitrate concentration facilitated the formation of a higher root number. The root average diagram of WT was increased in NN, but it was increased higher in OE under the LN condition compared to WT, as well as both (WT and OE) under the NN condition (Figure 6D). Similar types of expansions were observed in the OE line for root surface area (22.0 cm², Figure 6E) and root volume (22.0 cm³, Figure 6F). However, root length was higher in the OE line (160 cm) under NN conditions, and it demonstrated non-significant changes compared to the WT (Figure 6G). Enlargement and expansion of root average diagram, root surface area, and root volume indicate that limited nitrate concentration does not inhibit root growth but rather allows it to optimize for absorption of more nutrient resources.

The nitrate accumulation in the transgenic whole plants (35S:MeSLAH4) was higher under both the low nitrate (LN) and normal nitrate (NN) conditions compared to their wild type. However, there was a significant difference in the nitrate accumulation in whole plants at the low nitrate concentration (Figure 7A). During organ or tissue-specific nitrate accumulation analysis, rice grain exhibited significantly higher nitrate accumulation in the transgenic lines compared to the wild type lines (Figure 7B). Conversely, transgenic rice lines demonstrated lower nitrate accumulation in the glume (Figure 7C).

These results showed that overexpression of the MeSLAH4 gene led to more nitrate accumulation by whole plants, more storage in the grain, but lower translocations to the glume in rice.