Introduction

Maize is cultivated extensively due to its versatility as a forage, feed, and food crop, as well as its capacity to produce great yields (Erenstein et al. 2022). Globally, maize is produced each year more than any other grain with a total global production exceeding 1 billion tons (Ranum et al. 2014). Maize’s significant intraspecific genetic diversity in abiotic stress performance and ease of cross-pollination makes it a crop of choice for crop improvement programs for African vast marginal regions where maize is an important staple. As an annual crop, maize is subject to seasonal abiotic and biotic pressures, with climate change and unsustainable land use presenting the most serious threats to maize production (Masuka et al. 2012). The most significant obstacle to maize production is escalating water scarcity and soil deterioration (Falkenmark 2013), and salinity is a key factor in soil degradation in Africa (Thiam et al. 2021). For example, salinization accounts for 50% of irrigated land and is constantly increasing (Thomas and Middleton, 1993).

Due to the complexities of tackling salinity holistically, current research on maize production should be directed toward in-planta and in-vivo phenotyping, genotyping, and genetic breeding. For instance, so far, studies of critical candidate genes involved in growth and response to saline conditions have been summarized in various forage crops such as sorghum (Sorghum bicolor) (Amombo et al. 2022), turfgrasses (Fan et al. 2020; Amombo et al. 2017), alfalfa (Medicago sativa) (Bhattarai et al. 2020), and white clover (Trifolium repens) (Wang et al. 2009). It is worth noting that most of the reports focus on tolerance/resistance rather than forage quality. Furthermore, the most introduced improved cultivars do not have all the beneficial qualities that make them appropriate for farming in the vast arid regions of Africa and other continents. As a result, contemporary germplasm genetic improvement programs should be exploited to accelerate the introduction of improved cultivars with high forage value as well as resistance to salt and drought.

Sugar is an incredibly critical biomolecule to consider because it plays a significant role in carbohydrate accumulation which is a major component of forage quality (Ruckle et al. 2017). Sugar transport is essential for C translocation between sources and sinks during and after photosynthesis (Lemoine et al. 2013). Sugar transport in plants is controlled by gene families which consist of several proton-coupled sugar/H+ symporters, sugar/H+ antiporters, and uniporters (Bazzone et al. 2018). Among the transporters, sucrose transporters (SUTS) have received considerable interest from maize researchers (Slewinski et al. 2009; Leach et al. 2017). Besides SUTS, Sugars Will Eventually Be Exported Transporters (SWEET) is a novel and underexplored class of genes with highly conserved functions in sugar distribution between photosynthetic leaves and agronomically important sinks with high potential for forage improvement (Anjali et al. 2020). Distinguished by their conserved MtN3/saliva domain, the SWEET gene family has been identified and analyzed at the genome-wide level in several model plant species, including barrel clover (Medicago truncatula) (Hu et al. 2019), rockcress (Arabidopsis thaliana), and black cottonwood (Populus trichocarpa) (Zhang et al. 2021) which has provided a foundation for studying important forage crops like maize.

Understanding the transcriptional regulation and function of this gene in maize through a forward genetic approach, as well as mapping its chromosomal location via reverse genetics provide a rich genetic resource for the targeted development of superior varieties with high tolerance to salinity and drought, as well as high sugar content, which is an important component of livestock production. Therefore, this work looked at the gene architectures, conserved motif compositions, and chromosomal location of SWEET genes in maize. Furthermore, we investigated the tissue-specific expression of ZmSWEETs and their transcriptional regulation and functional analysis in two varieties with contrasting sugar amounts under normal and saline conditions.

Materials and Methods

In Silico ZmSWEET Gene Identification

The most recent versions of maize genome annotations were obtained from the genome assembly (https://maizegdb.org), while the SWEET full-length and homeodomain amino acid sequences were obtained from the Arabidopsis Information Resource (TAIR: http://www.arabidopsis.org/), aligned with multiple sequence alignment program MAFFT v5.3 (Katoh et al. 2019), and then loaded to sequence homologs searching software HMMER v3.0 (Potter et al. 2018) for the construction of Hidden Markov models (HMM). Using an E-value threshold of 1.0, the HMM profiles were used as queries against annotated maize protein databases. For further query, a BLASTP search with an E-value threshold of 0.01 was also conducted using both the full-length and homeodomain amino acid sequences of Arabidopsis SWEET to detect extra putative SWEET genes. The protein sequences obtained using the two approaches described above were combined, and redundant entries were manually removed. Pfam (https://pfam.xfam.org/) and SMART (http://smart.embl.de/) were used to examine the target sequences for the presence of the conserved homeobox domain (Finn et al. 2016). For evolution analysis, a bootstrapped phylogenetic tree was generated using the neighbor-joining technique, while the genetic distance was computed using MEGA v 5.1 software (http://www.megasoftware.net/).

To understand the gene structure, the exon–intron architectures were examined and visualized using the Gene Structure Display Server (GSDS: http://gsds.cbi.pku.edu.cn/); while MEME 4.9.1 (http://meme-suite.org/) was used to find conserved motifs in model SWEET and maize SWEET genes, WebLogo (Crooks et al. 2004) (http://weblogo.berkeley.edu/logo.cgi) was used to visualize them. The following parameters were set: the frequency of motif occurrences was set to zero or one per sequence; the maximum number of motifs was set to eight; the optimal motif width was set to six and one hundred, and the optimal number of sites for each motif was set to two and two hundred.

Chromosome Distribution and Subcellular Localization

The SWEET gene annotation information in the maize genome database was used to evaluate the location of the maize SWEET family members on maize chromosomes. The plant genome duplication database service (http://chibba.agtec.uga.edu/duplication/index/locket) was used to find duplicate gene pairs. The Clustalw algorithm was used to identify the amino acid sequence of the partially repeated SWEET gene (Thompson et al. 1994). The pre-built ngLOC model database (http://ngloc.unmc.edu/) was used to retrieve a web-based interface for predicting subcellular localization. SWEET protein sequences in FASTA format were used to create predictions, and maize species were used as the default. The MLCS (Multi-Localization Confidence Score) (King et al. 2012) was used to determine the prediction level of the top two sites.

Plant Materials and Growth Conditions

Two maize varieties (INRA16595 and Dracma) with contrasting sugar yields were selected (Fig. S1). From our phenotyping experiment, we observed that INRA16595 (a local germplasm-designated variety V1) is a high-sugar-yielding variety sourced from the National Institute for Agricultural Research (INRA) in Morocco, while Dracma (the silage variety designated variety V10) is the farmer-preferred commercial variety sourced from Syngenta with significantly lower sugar content but higher dry mass. We conducted a pot experiment simulating field saline conditions. The seeds were sown in plastic pots filled with commercial soil and were placed under the light with a 14 h photoperiod, a dark/light temperature cycle of 25/30 °C, and relative humidity of 55–65%. The plants were irrigated every other day with 500 mL of deionized water (control), 4 dS/m water (low), and 8 dS/m (medium) salinity for each variety. For the salt treatments, the extra water flowing at the bottom was reirrigated back until there was no further flow through. Sampling for RNA extraction was done at the transition between the tassel stage and the earliest reproductive stage. During this growth stage, transcriptional regulation of flowering and sensitivity to environmental stress is high. Leaves were collected and frozen at − 80°C for RNA extraction. The experiment consisted of six biological replicates.

RNA Extraction, cDNA Synthesis, and RT-qPCR

Young clean shoots were removed and homogenized in liquid nitrogen. With some modifications, the total RNA was isolated using the Canvax Total RNA extraction Kit (Carl Stuart UK Limited, Surrey, United Kingdom). The RNA double bands were confirmed using 1% agarose electrophoresis on the horizontal electrophoresis system (Avantor, Hamilton Street, Allentown, PA, USA) dyed using ethidium bromide and visualized on an ultraviolet gel imager (Mini 6, G: box, Syngene, MD, USA). The HiScript II One-Step qPCR SYBR Green kit (Red Maple Hi-tech Industry Park, Nanjing, PRC) was used for reverse transcription to cDNA and qPCR. The cDNA synthesis and qPCR were prepared as follows: 10 μL 2 × One-Step SYBR green mix, 1 μL One-Step SYBR Green enzyme mix, 0.4 μL 50 × ROX reference dye 1, 0.4 μL gene-specific primer forward (10 μM), 0.4 μL gene-specific primer reverse (10 μM), 2 μL template RNA, and topped up with RNase-free ddH2O to 20 μL. Then the qPCR was operated by AriaMx Real-Time PCR (qPCR) instrument (Agilent, Santa Clara, CA, USA) with melting curves inspection at the end of each reaction. The PCR reaction consisted of the following steps: reverse transcription (1 cycle, 55 °C, 15 min), initial denaturation (1 cycle, 95 °C, 30 s), cycling reaction (40 cycles, 96 °C, 10 s and 60 °C for 30 s), melting curve (1 cycle, 95 °C, 15 s, 60 °C, 60 s, 95 °C, 15 s). Elongation Factor 1 (EF1) gene was used as a reference, each reaction had three duplications. The values of the relative expressions for SWEET family genes were calculated using the 2−∆∆Ct method. Table S1 shows primers used for real-time PCR in this study.

Homologous Overexpression of ZmSWEET7

Immature Dracma maize embryos were prepared following D'Halluin et al. (1992) protocol with significant modifications. Briefly, the seeds were isolated surface sterilized, and treated with 0.3% macerozyme for 3 min at pH 5.6. The embryos were washed and placed into a disposable cuvette containing 200 µL of phosphate buffer saline. In each cuvette, 15 μL of plasmid DNA with ZmSWEET7 insert from variety V1, and the GUS reporter gene were introduced into the enzyme-treated embryo. After 1 h of incubation, the cuvettes were placed in an ice bath for 10 min after which electroporation was performed by discharging one pulse with a field strength of 375 Vlcm from a 900-pF capacitor (BTX Twin Waveform Electroporation Systems, Holliston, MA, USA). Then, embryos were washed and transplanted back into a nutrient medium for further growth. Following germination, the seedlings were transplanted to pots with commercial soil and watered every other day with a specified amount of water. The following treatment was imposed: The mutant and wild types were both treated to 8 dS/m of saline water, while the control was treated with deionized water (EC = 0 dS/m). Excess saline water was trapped beneath the surface and reirrigated until no more water trickled and EC measurements were taken frequently to maintain constant salinity. Plants started to exhibit phenotypic differences after 8 days of treatment when sampling for physiological analysis began. To confirm electroporation success, fresh leaf samples were collected from the pots and cut into quarters. The sections were transferred to 0.5 mL of X-Gluc stain and incubated overnight at 37 °C. The stain was then rinsed in warm 70% ethanol until the chlorophyll color disappeared. The GUS stain was examined using a dissecting microscope (VWR, SN 545036).

Phenotypic Analysis

The phenotypic selection was based on the following criteria: (a) salt sensitivity (chlorophyll a fluorescence), (b) CO2 dynamics (quantum efficiency of CO2 assimilation (phiCO2), intra and extracellular CO2, CO2 conductance, and CO2 assimilation rate), (c) water status (relative water content (RWC), transpiration rate, and water conductivity), and (d) yield (total sugar and dry matter). Photosynthesis measurements were done using the combined induction kinetics and gas exchange measurement using Li-COR 6800 equipment (Li-COR Biosciences, Lincoln, NE, USA). All the measurements took place in the morning after the extra dark adaptation. Measurements were performed on the ear leaves in each pot. Chlorophyll fluorescence measurements were performed both on the control, low, and medium salinity treatments on the same day with minimum time wastage. After the adaptation of leaves to darkness, a light pulse at a flow rate of 500 µ mol/m2/s was applied with the help of a light-emitting diode. The fast fluorescence kinetics (F0 to FM) was recorded to 1 s. For each variety and treatment, at least 6 repetitions were applied. The measured data were analyzed by the JIP test according to (Strasser et al. 1995). CO2 assimilation was measured using the CO2 response curve at various CO2 concentrations i.e., 0, 200, 400, 600, 800, and 1000 µ mol mol−1.

The Deans et al. (2018) approach was used to determine the total soluble sugars. In summary, dried harvested shoot samples were finely milled, and 20 mg of the powder was transferred into a glass test tube blended with 1 ml of 0.1 M H2SO4 and heated for 1 h in a water bath. The samples were chilled in a lukewarm water bath before being centrifuged for 10 min at 15,000 rpm. A 15 µL of the supernatant was transferred to a clean glass test tube and 400 mL of distilled water was added. 400 mL of 5% phenol was added, followed by 2 mL of concentrated H2SO4. The reaction mixture was vortexed and incubated for 30 min at room temperature. A spectrophotometer was used to measure absorbance at 490 nm. The glucose standard curve was used to calculate the values.

For RWC from fully expanded leaves, 1.5 cm wide by 4 cm long portions were cut with scissors from the area between the mid-vein and the edge. Three samples (replications) were collected from each plot, each sample representing a different plant. To avoid physiological changes, sampling proceeded quickly. Each sample was placed in a pre-weighed airtight plastic vial with its basal part to the bottom. The vials were immediately placed in a cold box but not frozen and transported to the laboratory as soon as possible. In the laboratory, the vials were weighed to obtain fresh weight (FW), after which the leaves were immediately hydrated in deionized water to full turgidity for 4 h under normal room light and temperature. After hydration, the leaves were removed from the water and dried using tissue paper to remove any residual surface moisture and immediately weighed to obtain the fully turgid weight (TW). The leaves were oven-dried for 24 h at 80 °C and weighed to obtain the dry weight (DW). The RWC was calculated as RWC (%) = [(FW–DW)/(TW–DW)] × 100, where FW is Sample fresh weight, TW is Sample turgid weight, and DW is Sample dry weight. Harvesting took place 40 days following silking, at silage harvest maturity just before senescence. The fresh plant samples were oven-dried at 60 °C and dry weight measurements were taken.

Data Analysis

All the experimental data for phenotyping and gene expression consisted of six replicates. Means and standard deviation were analyzed using SPSS version 16, while the multivariate and chlorophyll a fluorescence curves were analyzed using Origin lab Pro version 2022b.

Results

Together, the HMM search using SWEET domains as queries, as well as BLAST using AtSWEET and O. sativa OsSWEET sequences as queries revealed a total of 19 ZmSWEETs. The ZmSWEETs were designated by their orthologous genes in A. thaliana. The genes encoded amino acids with sizes ranging from 14,609.5 to 153,183.84 kDa and isoelectric points (pI) ranging from 5.33 to 9.79 which indicated that most of the ZmSWEET proteins were basic proteins. All the ZmSWEET proteins were hydrophobic proteins with a grand average of hydropathicity, (GRAVY) above 0. These results indicated that the basic properties of the proteins encoded by members of the maize ZmSWEET gene family were different (Table 1). Multiple sequence alignment showed that the 7 alpha-helical transmembrane domains (7-TMs) were basically conserved in ZmSWEETs, while ZmSWEET17 lacked the TM1 domain (Fig. S2).

Table 1 The ZmSWEET gene family members in Zea mays
Full size table

Phylogeny and Gene Structure

We performed a phylogenetic study of the discovered genes with other species i.e., A. thaliana; O. sativa; Vitis vinifera, and Litchi chinensis to better understand their probable unique functions. Maize and the other species were classified into four evolutionary clades. Clade 4 was the smallest clade in any of the investigated species with only three ZmSWEET genes. Clade III had the most evolutionary-related genes, with 9 from maize, indicating that SWEET members from other species and maize may be linked. There is a tandem duplication of ZmSWEET4 (SWEET4 and SWEET4C) distributed in clades II and I (Fig. 1).

Fig. 1
figure 1

Phylogenetic tree of SWEET sequences from maize and other plant species. Clades I, II, III, and IV are indicated by purple, pink, yellow, and green sections, respectively. At, A. thaliana, Os, O. sativa; Vv, Vitis vinifera, Lc, Litchi chinensis (Color figure online)

Full size image

The motif organization was investigated to better comprehend the detected maize gene structure. All ZmSWEET genes had 5 to 10 motifs. The number of motifs in SWEET varied among clades. ZmSWEET6, ZmSWEET8, ZmSWEET17, and ZmSWEET18 from clade III, for example, had the most motifs (10), while ZmSWEET10 from clade II had just five. Within the same group, small variations in gene features were also seen. For example, ZmSWEET15 in clade I had 8 motifs, but ZmSWEET13 in the same clade had just 7. Untranslated areas were found in all 19 individuals, the great majority of whom belonged to clade III. The arrangement of motifs in a gene family may reveal information about its functional development (Fig. 2A). Cis-regulatory elements are distinct DNA sequences located upstream of gene coding regions that regulate gene expression by interacting with transcription factors. In this study, we observed that light responsiveness had the most cis-regulatory elements among all the genes, suggesting a sugar transport-light link in maize. The second most prevalent cis-regulatory element is linked to stress-signaling hormones such as MeJA, salicylic acid, abscisic acid, and gibberellins, as well as endosperm and meristem control. Other notable features were seen in metabolism and regulation, circadian control, MYB binding, and anaerobic control (Fig. 2B).

Fig. 2
figure 2

A Conserved motif, and gene structure analysis of ZmSWEET genes. Ten motifs were displayed in different colors. B Predicted cis-elements as analyzed by PlantCARE. The upstream length to the translation starting site can be inferred according to the scale at the bottom (Color figure online)

Full size image

Chromosomal Distribution and Segmental Duplication

All the SWEET genes in maize were found on 9 of the 10 maize chromosomal pairs. The number of genes found on each chromosome ranged from one on chromosomes 2 and 6 to four on chromosomes 7 and 9. Members of comparable gene clades were found on the same chromosomes. Clade 1 ZmSWEET1 and ZmSWEET2 members, for example, were all found on chromosome 1. ZmSWEET3 is the sole maize gene on chromosome 2, whereas its group III orthologous counterparts such as ZmSWEET5, ZmSWEET6, ZmSWEET7, and ZmSWEET4 are all situated on chromosome 3. Chromosome 9 similarly contained only one gene, ZmSWEET17 (Fig. 3A). The BLASTP and MCScanX algorithms grouped ZmSWEET into four duplication events. For example, ZmSWEET5 and ZmSWEET13, ZmSWEET1 and ZmSWEET17, ZmSWEET7 and ZmSWEET16, and ZmSWEET8 and ZmSWEET3 can be produced by fragment duplication. Based on the findings, these ZmSWEET genes were most likely created through gene segmental rather than tandem duplications (Fig. 3B).

Fig. 3
figure 3

Schematic representations of the A chromosomal distribution of the ZmSWEET genes. The chromosome number is indicated on the left of each chromosome and/or scaffold. B Synteny of SWEET genes in maize

Full size image

Predicted and Confirmed Expression in Agronomic Phenotypes

Expression prediction revealed that ZmSWEET16, ZmSWEET7, and ZmSWEET10 were found to have considerably higher expression levels in the tassel inflorescence, whereas ZmSWEET15, ZmSWEET11, and ZmSWEET19 were found to be strongly expressed in the vascular leaf. ZmSWEET3 is the sole highly expressed gene in maize endosperm, indicating a potential function in grain filling; ZmSWEET10 and ZmSWEET11 are also strongly expressed in the pericarp. In the shoot axis internode, ZmSWEET15 and ZmSWEET11 are co-expressed. As a result, we considered these SWEET genes to be candidates for further functional analysis (Fig. 4A). Through qPCR, all 19 ZmSWEETs were detected in the shoots. We, therefore, tested the expression of the genes in shoot tissues of agronomic importance. The leaf sheath has the most upregulated genes. Among the most upregulated genes are ZmSWEET10 and ZmSWEET19 in the leaf sheath, while ZmSWEET5 and ZmSWEET9 are upregulated in the tassel. The highest upregulation is ZmSWEET7 in the shoot internode which is also upregulated in the leaf sheath and whole leaf. Consistent with bioinformatic prediction, most of the genes are downregulated (Fig. 4B).

Fig. 4
figure 4

A Gene expression patterns of ZmSWEET genes in maize phenotype. The color scale represents fragments per kilobase of exon model per million mapped reads values of RNA sequencing, with red indicating the highest transcript abundance and blue indicating the lowest transcript abundance. B Cluster heat map of gene expression via qPCR in phenotypes of agronomic interest. Blue is upregulation and red is downregulation

Full size image

Expression of SWEET Under NaCl Treatment

We compared the expression profile of salt-responsive genes in the shoots, and 11 genes are responsive to salt treatment in both varieties by real-time qPCR and gene-specific primers. In V1, three genes ZmSWEET3, ZmSWEET11, and ZmSWEET1 are upregulated by salinity. All these genes are downregulated by salinity in Dracma. The only upregulated gene by salinity in Dracma is ZmSWEET7 which is also highly upregulated in V1. Also, with more than sixfold change compared to the control, ZmSWEET7 stands out as the most upregulated gene. Interestingly, this gene was also predicted to be of high abundance in the shoot axis internode. Due to its distribution in traits of agronomic interest as well as high expression under salinity, we, therefore, considered ZmSWEET7 our target for further functional analysis under salt stress (Fig. 5).

Fig. 5
figure 5figure 5

Bar plots of gene expression levels under low and medium salinity in two maize varieties. Error bars indicate the standard deviation of means. Bars below 1 signify downregulation, while above 1 is upregulation

Full size image

Phenotypic Variation in Wild and Mutant Samples

The GUS::ZmSWEET7 fusion protein could be visualized as dark blue colorization on the epidermal cells under fluorescent microscopy (Fig. 6A) indicating integration of ZmSWEET7 promoters. The morphology of the seedlings varies after transplanting and salt treatments. The plant height, number of leaves, and leaf length are notable morphological variations. The control has the greatest height and longest leaves, followed by the mutant (Fig. 6B, C). This trend was consistent up to silage maturity.

Fig. 6
figure 6

A Abaxial visualization of GUS::ZmSWEET7 in the leaves of Dracma as depicted by light microscopy with 20-fold magnification using X-GlcA Cyclohexylammonium dye. B Early morphological variation after salt treatment C Morphological variation after salt treatment at the reproductive stage before harvest

Full size image

There is a significant increase in the total soluble sugar content in the mutants (46.78 mg/g DW) compared to the wild (38.66 mg/g DW) which despite being a significant increase is still below the one observed in V1 from the previous phenotyping study. This can be attributed to the fact that ZmSWEET7 is just one of the several genes in V1 that play a role in sugar accumulation. An increase in the sugar content under saline conditions could also be an osmotic stress adaptive mechanism since sugars act as important compatible osmolytes under osmotic stress caused by salinity and drought. It was, therefore, necessary to determine the water status of plants by measuring the RWC. There is an insignificant decline in the RWC of salt-treated wild type compared to the control. Mutants exhibited a slightly significant increase (Fig. 7).

Fig. 7
figure 7

Bar plots of soluble sugar content and relative water content

Full size image

Chlorophyll a Fluorescence and CO2 Assimilation Rate and Efficiency

The results revealed that different treatments resulted in considerably varied slow kinetics of chlorophyll a fluorescence behavior. For instance, the wild type had the lowest F0 of 496.7, whereas the control treatment had a much higher F0 of 890. The F0 was medium in the mutant at 579.8. Generally, after the first F0, all treatments experienced an O-J increase that occurred between 0.00001 and 0.0001 s. To achieve the greatest FM, the J-I-P phase of the fluorescence induction curve rise time spanned from 0.0001 to 0.001 s. The control had the highest FM value of 1948.72, followed by the mutant, and the wild type had the lowest of 996.18. Notably, the J-I-P rise was much delayed in the wild type (Fig. 8A). The mutant exhibited the highest rise in A between 0 and 200 ppm followed by the wild type. However, the wild type reached the plateau at the earliest (Fig. 8B). A similar trend is observed in the dry matter weight where the mutant, despite being lower than the control, exhibits a significantly higher DW than the wild type (Fig. 8C). The mutant and control FV/FM levels do not differ significantly. However, the wild type had much lower values (Fig. 8D), demonstrating that ZmSWEET7 could protect photosystem II in maize.

Fig. 8
figure 8

Plots of A fast chlorophyll a fluorescence transient (OJIP) curve up to 1 s, B CO2 assimilation rate, C dry matter weight, and D maximum quantum yield of photosystem II in wild type and mutants’ maize grown under salt stress

Full size image

Further analysis of the photosynthesis parameters indicated that GUS::ZmSWEET7 mutants experienced a significant increase in the phiCO2 compared to the control and wild type with a small difference between the control and wild type. Salt treatment caused a significant decline in the stomatal conductance to water and CO2 in both wild and mutant types which coincided with the transpiration rate. However, the conductance despite being lower than the control was significantly higher than the wild type. The same pattern is observed in intracellular and extracellular CO2 (Fig. 9).

Fig. 9
figure 9

Measured photosynthesis parameters in control, mutant, and wild-type maize

Full size image

Multivariate analysis reveals a strong correlation between A and phiCO2 in both salt-treated wild and mutants. In the wild type, A is strongly positively correlated with the dry matter, intracellular and extracellular CO2 but weakly correlated with the soluble sugar content. Transpiration was the only positive correlation with sugar, while there is a strong negative correlation between transpiration and FV/FM. The phiCO2, on the other hand, is strongly positively correlated with dry weight (0.65), moderately correlated with extracellular CO2, and there is almost zero correlation with the soluble sugar content. There is also a strong negative correlation between phiCO2 and FV/FM. In the mutants, A is strongly positively correlated with the extracellular CO2 and the FV/FM. However, there is still a very low correlation between A and soluble sugar, while the strongest negative correlation is observed with transpiration rate. The phiCO2 on the other hand displays strong positive correction with the DW and FV/FM and soluble sugar content. There is a negative correlation however between conductance to CO2 and intracellular CO2, while the extracellular CO2 exhibited a slightly positive correlation (Fig. 10). These findings indicate that ZmSWEET7-mediated increase in phiCO2 plays a positive pleiotropic role in C accumulation in the form of sugar or dry matter via increased FV/FM.

Fig. 10
figure 10

Pearson’s correlation plot of combined induction kinetics and gas exchange measurement, sugar, and dry matter yield under salt stress. A Wild and B mutant. Gtw-stomata conductance to water, E-transpiration rate, Ci-intracellular CO2, Ca-extracellular CO2, DW-dry weight, A-CO2 assimilation rate, RWC-relative water content, gtc-stomata conductance to CO2

Full size image

Discussion

Despite maize's economic and agronomic importance, its study as a forage in Africa has often lagged compared to other forages such as Napier grass (Pennisetum purpureum) (Balehegn et al. 2021), particularly in terms of developing genotypes with high abiotic stress tolerance while retaining high nutritive value. As a C4 crop, maize uses sunlight to generate carbon-based macromolecules in its foliar tissues, which is the basis of energy flow through the trophic levels in the ecosystem (Dusenge et al. 2019). Among the photosynthates, sugars are the most prevalent carbon molecules which play an integral role in biomass accumulation (Aluko et al. 2021). The important connection between soluble sugars and forage quality has been well documented by Capstaff et al. (2018). As primary consumers, herbivorous livestock actively manage their sugar intake and assimilation from forages to meet physiological demands, which is critical to their survival and productivity (Sarwar et al. 1992). Therefore, forage quality and soluble sugar content are intricately connected. Perennial ryegrass (Lolium perenne), for example, is a popular forage crop whose excellent digestibility in livestock has been associated with its high soluble sugar content (Ruckle et al. 2017). Besides, sugars are also required for plant response to biotic and abiotic stress (Ciereszko. 2018; O'Hara et al. 2012; Eveland et al. 2012; Lastdrager et al. 2014). Thus, sugar distribution and accumulation patterns are key aspects to consider when selecting excellent fodder for optimum livestock production.

In this study, using a genome-wide approach, we identified 19 ZmSWEET genes in maize. Their AA lengths were varied and distributed across various organelles. The varied structure of ZmSWEETs indicates that they have distinct functions in different biological processes or under different growth settings. Gene structure analysis revealed that the bulk of ZmSWEET genes had 10 motifs. This was higher than those in closely related species like sorghum (Miao et al. 2017). To validate the conserved motif analysis, we conducted a phylogenetic analysis. These genes were classified into four clades (Clades I–IV) based on their phylogenetic evolutionary relationship, which corresponded to the classification of SWEETs in the model species A. thaliana. Our findings revealed that gene members in each clade had a unique conserved motif indicating that they may play a variety of roles in maize. However, there are disparities in the number of subfamilies revealed; for example, in Clade III, 9 members of ZmSWEET were discovered, compared to only one member of AtSWEET and OsSWEET. In addition, the number of duplicated gene pairs differed across clades compared to Arabidopsis. For instance, we observe four duplication events on chromosomes 2–4, 3–8, and 1–9. This suggests that during maize evolution, gene duplication and divergence events happened more in a segmental rather than in a tandem manner compared to A. thaliana and O. sativa.

Due to the intricate connection between SWEET and sugar, we validated the bioinformatics results and tested the expression level of SWEETs in phenotypes of agronomic importance. ZmSWEET7 was highly distributed in almost all tested phenotypes. While the highest expression level is observed in the ZmSWEET19, ZmSWEET19 is an ortholog of AtSWEET2 whose vacuolar transcription regulates sugar release by decreasing the access of glucose sequestered in the vacuole, minimizing carbon export (Chen et al. 2015). In this study, the transcript levels of ZmSWEET19 were highest in the leaf sheath indicating that these genes play an important role in sugar accumulation in an organ-specific manner. We also looked at the expression patterns of SWEET family genes from distinct phylogenic clades in salt stress to determine if functional differentiation occurred. A total of 11 genes were responsive to salt stress. Among them, all the ZmSWEET genes in clades I, II, and IV can respond to salt stress conditions, highlighting the common involvement of these three phylogenic clades in maize salt stress response. Most of the genes are upregulated by salinity in V1 compared to Dracma with ZmSWEET7 having the highest upregulation level. Interestingly, the highly expressed ZmSWEET19 during normal conditions was not strongly salt inducible.

Therefore, we probed the potential role of ZmSWEET7 in salt tolerance by overexpressing it in Dracma through direct immature embryo electroporation. Dracma varieties that overexpress the gene have increased sugar content and RWC compared to the wild type but lower than the control. The RWC represents a plant's water state and is inextricably linked with the osmotic potential (Paulino et al. 2020), and recently, the RWC has been documented to be among the most reliable physiological marker for salt-stressed plants (Soltabayeva et al. 2021). Plants under salt stress, like those under water stress, experience physiological drought because of increased osmotic pressure in the surrounding soil. To overcome this, plants must enhance their osmotic pressure by accumulating osmotically compatible solutes, and sugars are among the osmotically active substances. Interestingly, there is a 20.98% gain (from 38.66 to 46.78 mg/g DW) from the wild type, which despite being a significant increase is still below observed in V1 from the previous phenotyping study. This can be attributed to the fact that ZmSWEET7 is just one of several genes in variety V1 that play a role in sugar accumulation. An increase in the sugar content under saline conditions could therefore also be an osmotic stress adaptive mechanism since sugars act as important compatible osmolytes. Turner (2018) reviewed osmotic adjustment studies on crops for the past 40 years and found that sugar had an essential function in enhancing crops' ability to withstand osmotic stress. Chimenti et al. (2006) reported notable intraspecies variation in osmotic adjustment in maize, which could be utilized to select drought-tolerant lines. The differential accumulation of soluble sugars and their positive association with water content in this study suggest that like in drought, soluble sugar could be utilized as a phenotypic marker for salt tolerance in maize.

Photosynthesis is sensitive to environmental changes and can determine the abiotic tolerance level of a plant as well as biomass buildup, both of which are essential for crop forage yield (Yadav et al. 2020). Chlorophyll fluorescence has been demonstrated to be one of the most sensitive physiological measures when subjected to salt stress, making it a reliable phenotypic marker (Papageorgiou and Govindjee 2004). As a result, we investigated the FV/FM from chlorophyll a fluorescence curve of Dracma variety that overexpresses ZmSWEET7. The findings demonstrated that different treatments resulted in significantly different behavior. Following the initial F0, all treatments saw an O-J rise suggesting that salinity modifications had no effect on this phase. This phase represents the photochemical step of Chl fluorescence induction. Thus, higher F0 values in the mutant relative to the wild type showed a larger physical separation of the PSII reaction center from their corresponding pigment antennae, which has been shown to contribute to better salt tolerance by restricting energy input into the electron transport chain (Srivastava et al. 1997). The J-I-P phase of the fluorescence induction curve rising time was set at 0.0001–0.001 s to produce the highest FM. Notably, the wild-type J-I-P increase was substantially delayed, demonstrating that this phase relates to plastoquinone accumulation, whereas it increased in the mutant. The large increase in I translates to slower electron transit to the PSI acceptors. In mutants, there is also a higher plateau, indicating a bigger number of PSI end acceptors, which are typically linked with alternative electron transfer routes that function as electron sinks.

In-planta CO2 dynamics define the pathway for carbon accumulation which in turn determines plant growth and production (McCarthy et al. 2010). An assimilation rate curve plotted versus intercellular CO2 concentration can reveal numerous insights into the biochemistry of a leaf or plant. For instance, a ZmSWEET7-mediated increase in stomatal conductance, which regulates gas exchange (CO2 and water), can allow maize to increase their CO2 uptake and subsequently enhance photosynthesis, while its negative correlation with Ci and A could be a strategy of increased CO2 and water use efficiency under saline conditions as evidenced by higher intracellular CO2 and assimilation rate and lower transpiration rate, respectively. The photosynthetic quantum yield efficiency is a crucial but seldom measured biophysical quantity which estimates both net carbon intake and net oxygen evolution concurrently (Du et al. 2018). The high correlation between phiCO2 and dry matter accumulation and sugar in mutants indicates that SWEET mediates carbon accumulation via increased efficiency of CO2 fixation per unit of supplied CO2. There have not been many maize studies that explicitly relate PSII with phiCO2. The positive association between phiCO2 and PSII, on the other hand, supports Edwards et al.’s (1993) claim that, across a wide variety of environmental conditions, fluorescence characteristics may be employed to predict accurately and rapidly CO2 assimilation rates in maize.

In conclusion, through a genome-wide analysis, we discovered 19 SWEET genes in maize and analyzed their chemical structure, chromosomal distribution, phylogeny, and promoter regions. To validate the work, we conducted a molecular stress physiology experiment which indicated that the 19 genes are expressed differentially in various phenotypes of agronomic interest, while 11 are salt inducible. Among the salt-inducible genes, ZmSWEET7 is the most upregulated in the high-sugar variety and its homologous overexpression in low-sugar variety enhanced the phiCO2 which is positively associated with DW and soluble sugar accumulation and FV/FM under saline conditions. Although dry matter and total sugar are quantitative traits controlled by multiple genes, this study provides insights into the potential role of ZmSWEET7 in these important carbon dynamics under saline conditions especially via phiCO2. To understand the specific transcriptional role of this gene, complete knock-out studies using genome editing will be needed.