Transcriptome-based strategies for identifying aluminum tolerance genes in popcorn (Zea mays L. var. everta)

Aluminum (Al) toxicity limits crop production worldwide. Although studies have identified genes associated with Al tolerance in crops, a large amount of data remains unexplored using other strategies. Here, we searched for single substitutions and InDels across differentially expressed genes (DEGs), linked DEGs to Al-tolerance QTLs reported in the literature for common maize, and investigated the alternative splicing regulated by Al3+ toxicity. We found 929 substitutions between DEGs in Al-tolerant and 464 in Al-sensitive inbred lines, of which 165 and 80 were non-synonymous, respectively. Only 12 NS variants had deleterious predicted effect on protein function in Al-tolerant and 13 in Al-sensitive. Moreover, 378 DEGs were mapped in Al-QTL regions for the Al-tolerant and 213 for the Al-sensitive. Furthermore, Al stress is primarily regulated at the transcriptional level in popcorn. Important genes identified, such as HDT1, SWEET4a, GSTs, SAD9, PIP2-2, CASP-like 5, and AGP, may benefit molecular assisted popcorn breeding or be useful in biotechnological approaches. These findings offer insights into the mechanisms of Al tolerance in popcorn and provide a ‘hypothesis-free’ strategy for identifying and prioritizing candidate genes that could be used to develop molecular markers or cultivars resilient to acidic soils.

www.nature.com/scientificreports/Despite its agronomic importance, ZmMATE1 7 and ZmALMT1 and ZmALMT2 9,10 have been characterized in common maize, but other genes and mechanisms involved in the tolerance to the phytotoxic Al ion are currently unknown.
Popcorn (Zea mays L. var.everta) has an important role in United States economy and also has a high demand in Brazil 11 .Like traditional maize, popcorn is extensively cultivated through tropics regions in Brazil, thus attracting the attention of breeders to obtain cultivars adapted to Brazilian conditions 12 .Al toxicity affects root development and causes damage and cell disorganization in the apical region in popcorn seedlings, ultimately compromising their growth and nutrient uptake 13 .The tolerance mechanisms in popcorn seedlings are involved with the increasing of sucrose content in the roots and shoots, starch decreasing in roots, and induced secretion of malate and fumarate 14 .
Gene expression is highly influenced by different environmental conditions and can trigger various point mutations across the genome.Plants exhibit diverse transcriptional, physiological, and fitness responses to abiotic stress [15][16][17] , and understanding how these variations affects gene expression as well as plant fitness and adaptation is crucial to providing solutions for crop breeding in the context of climate change.Beyond point mutations, alternative splicing is a dynamic post-transcriptional regulatory mechanism that produces multiple protein variants and regulates many physiological processes essential for plant growth and development, especially in response to stress conditions 18 .In this way, RNA-seq studies allow obtaining a large amount of data and provide new perspectives to uncover transcript dynamics and altered patterns under adverse environment conditions.
Recently, the gene expression profile of two popcorn inbred lines contrasting in Al-tolerance were tracked after 72 h of stress, revealing the mechanisms involved in a long-term Al-exposure 19 .This study used the RNAseq approach and detected genes already known to be related to Al-tolerance, as well as others not previously described in hydroponic or soil experiments.Transcriptomic data analysis can provide valuable information on expressed genes in QTL regions, alternative splicing and SNPs associated with stress tolerance in plants.However, strategies to reduce the list of candidates are necessary when selecting candidates to use in breeding programs.
In this study, we performed comprehensive and extensive analyses to investigate potential candidates for Al-tolerance in popcorn.We assessed single substitutions and InDels across the differentially expressed genes (DEGs), linking them to Al-tolerance QTLs previously reported in the literature for common maize, and investigated the post-transcriptional regulation of popcorn under Al-stress.Our findings provide novel insights into the roles of different molecular processes underlying Al-tolerance in popcorn.They are useful to be explored in genetic engineering tools and/or as biomarkers in popcorn breeding programs.

Al stress triggers polymorphisms in expressed genes
A total of 929 substitutions were identified among 128 DEGs in the Al-tolerant inbred line (11-133), and 464 substitutions were found among 67 DEGs in the Al-sensitive inbred line (11-60) (Supplementary Table 1).In both inbred lines, most of the substitutions were synonymous, and variations were also found in untranslated regions (UTRs) (Fig. 1).
Glutathione transferases are the main cellular detoxification enzymes that act against oxidative damage in plants 22 .The increased expression and activity of GST have been linked to aluminum tolerance in maize 23 .The predicted deleterious effect in GST7 function was by a substitution of the basic Arg with Gly within the conserved N-terminal domain (Table 1).In GSTs, the N-terminal domain forms a thioredoxin-like fold, and changes in this www.nature.com/scientificreports/domain have been correlated with high levels of GST expression in maize 24 .Additionally, a missense substitution with neutral predicted effect was detected in the C-terminal domain in GST7 (Supplementary Table 2).
In addition to GSTs and the total glutathione accumulated in plant tissues under Al 3+ toxicity, phenolic compounds are also regulated in response to this adverse condition, acting as potential defensive compounds.Phenylalanine ammonia-lyase is a key biosynthetic enzyme that produces intermediaries for polyphenolic compounds, and its activity increases in roots of Al-tolerant lettuce when exposed to Al toxicity 25 .In sorghum, the increasing of phytochelatins in roots can be related to Al 3+ chelation 26 .Our findings suggest that this substitution in GST7 and PAL1 may confer an advantage in the detoxification of reactive oxygen species (ROS) and restriction of Al 3+ in root cells of the Al-tolerant inbred line.
A single substitution in the polar residue region of HDT1 resulted in the change of a Pro to a Thr (Table 1).Additionally, five substitutions with neutral predicted effects were found in HDT1 (Supplementary Table 2).Histone acetylation and deacetylation play important roles in gene expression in eukaryotic cells.Maize exposed to different heavy metal stress conditions showed different acetylation levels due to different regulation of histone deacetylases 27 .These substitutions found in HDT1 may be important for the adaptation of popcorn roots to Al toxicity, however, epigenetic regulation in maize under Al-stress is poorly understood and needs further investigation.
In the Al-sensitive inbred line, the down-regulated kinase SnRKβ1 had a deleterious substitution in the AMPK1_CBM domain (Table 2).SnRK1 plays several roles in abiotic stress resistance, and the decrease of its expression is associated with stress sensitivity 28 .This gene also regulates organic acid metabolism and other signaling pathways.SnRK1 knockdown lines of Arabidopsis thaliana under dark conditions altered the content of organic acids, revealing the importance of this gene in metabolic adaptation to stress 29 .Based on this, this amino acid change may cause unfavorable effects on protein function, increasing sensitiveness of 11-60 to Al 3+ .
Although this study focused on non-synonymous variations within protein-coding regions, silent substitutions and variations in intragenic regions among the DEGs may also have benefits for post-transcriptional www.nature.com/scientificreports/mechanisms in response to Al-stress.While synonymous substitutions do not alter the amino acid sequence, they can still affect gene expression, protein folding, and the fitness of the organism 30 .Intragenic substitutions/ InDels within 3′-UTR and 5′-UTR regions can alter gene expression by changing binding sites for transcription factors and miRNAs 20 .A total of 164 variations were found in 3′-UTR and 5′-UTR regions in Al-tolerant inbred line, while 109 in the Al-sensitive, and none of these genes were shared in both inbred lines (Supplementary Table S1).The same transcript of HDT1 that presented a NS variation (Zm00001d012092_T001) also presented two 3′-UTR variations, reinforcing its significance to Al-tolerance in popcorn.Furthermore, important up-regulated genes associated with abiotic stress presented variations in UTR regions.Were detected three variations in the 3′-UTR and 5′-UTR regions in SWEET4a (Zm00001d015905_T001), one in 3′-UTR in a Heavy metal transport/detoxification superfamily protein (Zm00001d027454_T001), one in 5′-UTR in GST12 (Zm00001d027540_T002) and six in both UTR regions for GST6 (Zm00001d027541_T001 and Zm00001d027541_T002). Curiously, GST6 was detected as a central hub in in silico protein-protein network interacting with several Al tolerance related genes 19 .
These finds shed light for the importance of substitutions in UTR regions as an adaptative force to the plant deal with Al stress, enhancing the expression and impacting in the mRNA stability and translation efficiency of Al tolerance related genes.
Linking RNA-seq data with QTLs previously described reduces the number of potential candidates for Al-tolerance A total of 378 DEGs were identified within previous Al-QTL regions in 11-133 (Supplementary Table 4), and 213 genes in 11-60 (Supplementary Table 5).Gene ontology enrichment analysis showed that the majority of these DEGs were categorized under oxidoreductase activity and oxidation-reduction process, moreover, terms related to stress response and defense were enriched only in the Al-tolerant inbred line (Fig. 2).Out of this total, 56 DEGs were shared between both inbred lines (Fig. 3).
Al-related genes found in common maize, like ZmMATE1, ZmALMT1 and ZmALMT2, were not differentially expressed in popcorn.This may be attributed to the timeframe of our study, which aimed to investigate the longterm molecular response to Al exposure, and to the genetic background of these popcorn inbred lines with the common maize genotypes tested in previous studies.
In 11-133, HDT1 and Stearoyl-acyl-carrier-protein desaturase 9 (SAD9, Zm00001d012221) were mapped in QTL7 5 , and both showed changes with predicted deleterious effect on protein function, and the first transcript also presented variations in 3′-UTR region (Table 3, Supplementary Tables 1 and 4).SAD is a key enzyme in determining the global content of unsaturated fatty acids (UFA) 31 .Its expression decreased, and the predicted deleterious effect may lead to the reduction in the UFA content of lipids cells, changing lipid fluidity.Moreover, two Δ 12 -fatty-acid desaturases (Zm00001d023768 and Zm00001d023769) were down-regulated in 11-133 but up-regulated in 11-60 (Fig. 3, Supplementary Table 4), and mapped in QTL2.A high degree of fatty acid desaturation allows for high capability to stabilize the membrane fluidity, alleviating membrane damage from Al stress and limiting the entry of Al 3+ into roots 32 .
Two aquaporins PIP2-2 (Zm00001d005410 and Zm00001d014285) were detected in QTL3, QTL14, and QTL17 5,7,8 , and an ABC transporter (Zm00001d024600) in QTL2 4 for the Al-tolerant inbred line (Supplementary Table 4).Aquaporins and ABC transporters play an essential role in the vacuolar sequestration of Al 3+ , promoting Al tolerance in plants 33 .Aquaporins PIP1-1 and PIP2 were recently identified in the maintenance of hydration in Citrus limonia L. plants when exposed to Al stress 34 .On the same way, ABC transporters facilitate vacuolar sequestration of Al, playing an important role in the tolerance mechanism 35 .Both are involved in diverse processes in plant growth and development under abiotic stress 36 .
The SWEET3b (Zm00001d023673) and SWEET4a (Zm00001d015905) were identified in QTL2 4 and QTL14 7 in 11-133 (Supplementary Table 4).In addition, SWEET4a presented variations in UTR regions (Supplementary Table 1).To deal with environmental stress conditions, plants need to maintain a strict regulation in the storage and transport of vacuolar sugar 37 .The SWEET genes play an important role in tolerance to osmotic stress, cold, high salinity, and drought 38 .The role of SWEET transporters is poorly understood in plants exposed to Al 3+ toxicity, but these genes may be related to the Al tolerant phenotype in popcorn.
We identified DEGs involved in cell wall modification in 11-133 inbred line that were mapped across several Al tolerance QTL (Supplementary Table 4).These genes are mainly involved in primary cell wall and xyloglucan metabolism. 39,40previously reported some of the same genes in Arabidopsis that are involved in Al tolerance. 41etected a xyloglucan endotransglycosylase protein 8 precursor in the Al tolerance QTL Alm1.These findings suggest that the process of structural organization of cell wall is crucial for root growth and development under Al 3+ toxicity conditions in the Al-tolerant inbred line.
CASP-like 5 (Zm00001d010038) and Arabinogalactan peptide 20 (AGP, Zm00001d010434) were both mapped in QTL6 5 and found to be up-regulated in 11-133 but down-regulated in 11-60.CASP protein is involved in stress resistance and nutrient uptake 42 , and in Tamba black soybean under Al-stress, the expression of CASP protein is associated with root growth 43 .On the other hand, AGP plays multiple roles in plant growth and development, and under abiotic stress, it is involved in the thickening of cell wall by oxidative crosslinking mediating the stress signaling response 44 .The presence of these genes within an Al tolerance QTL may be responsive for the root and plant growth in acid soils.
In addition to reducing the number of candidates useful for developing molecular markers for marker-assisted selection or genomic selection for Al-tolerance, the identification of DEGs in QTL regions can provide insights into the molecular mechanisms underlying the traits controlled by these regions.

Response to Al stress is primarily regulated at the transcriptional level
For the Al-tolerant inbred line, we observed a total of 25,026 RI; 15,357 A3SS; 11,782 A5SS; 10,786 SE; and 626 MXE events, respectively.A single gene (Zm00001d002141), which encodes a UBP1-associated protein 2A, underwent differential SE events, with three exons whose inclusions increased upon treatment.A similar frequency of AS events was observed for the Al-sensitive inbred line, which had 25,039 RI; 15,345 A3SS; 11,791 A5SS; 10,743 SE; and 618 MXE events, respectively.Three genes (Zm00001d047479, Zm00001d036136, Zm00001d044494) underwent differential SE events in the Al-sensitive inbred line.These genes encode a superoxide dismutase (SOD), CYSTM domain-containing protein, and pectin acetylesterase (PAE), respectively.
The small number of genes undergoing differential AS (DAS) reveals that post-transcriptional regulation does not play a significant role in popcorn response under long-term Al exposure.In 11-60, the three genes that underwent differential SE events are involved in abiotic stress response.The SOD is already known to play a role in the detoxification mechanism in maize under Al stress 45 .In Arabidopsis, CYSTM is involved in several developmental processes under abiotic stresses conditions 46 .Silencing PAE and annexin in hairy roots of Medicago truncatula increased the sensitivity to Al stress, suggesting that these gene may be involved in Al resistance response 47 .In rice, a significant change of AS profile was observed in the tolerant cultivar under Al stress 48 .Also in rice, the response to cadmium stress is highly controlled at the post-transcriptional level, as several differentials AS events were detected 49 .Similarly, rice response to alkalinity stress also triggered many differentials AS events 50 .
While a substantial number of DAS events were not associated with Al-tolerance in popcorn, we propose that post-transcriptional regulation plays a role in establishing timely patterns of downstream gene expression in response to stress.The interplay between transcriptional and post-transcriptional regulation, including AS, is highly complex and context-dependent, for these reasons we believe that post-transcriptional regulation may act during the growth and development of popcorn plants under Al-stress.Therefore, conducting additional analyses to explore the involvement of AS during a time course exposure of popcorn seedlings to Al may enhance our understanding of the complexity response to Al stress.
In this study, we have explored the transcriptomic data beyond gene expression and demonstrated the potential that this tool can offer to researchers in reducing targets and saving time for further analysis.This strategy has allowed us to identify several genes with a wide range of functions correlated with Al stress response, such transporters, enzymes involved in cell wall modification, fatty acid biosynthesis, and genes involved in stress response and plant development.Although searching for DEGs in Al tolerance QTL regions can significantly www.nature.com/scientificreports/reduce the number of candidates to be explored, the genetic background differences between the genotypes used in QTL studies and our popcorn inbred lines may allow new regions that control Al-tolerance on unmapped chromosomes.Furthermore, as the Al stress response is majorly regulated at the transcriptional level, the genetic variations found in this study can be validated and functionally characterized to understand their roles in Al stress tolerance.This information can be used to develop molecular markers for Al tolerance, which can be utilized in breeding programs to develop cultivars that are resilient to the challenges posed by changing environmental conditions.

Plant material and selection of differentially expressed genes
The first step was the selection of the genes from our previous work 19 .The transcriptome was generated from seven years old seedlings of two contrasting inbred popcorn lines developed by the Popcorn Breeding Program of the Universidade Federal de Viçosa (Viçosa, Brazil): 11-133 (Al-resistant) and 11-60 (Al-sensitive).These genotypes were selected based on a previous study 13 to screen inbred popcorn lines with different Al sensitivity.Seedlings with uniform growth were picked randomly and transferred to a nutritive solution 51,52 with constant aeration to acclimate for 24 h.Then, the treatment group was subjected to aluminum stress with 540 µM of AlCl 3 (160 μM Al 3+ ) for 72 h.To assess only the Al effect on seedlings growth, the pH was adjusted to 4.5 for both control and treatment conditions and the seedlings maintained in a growth chamber at 25 °C with a 12/12 h light/dark cycle.RNA from roots was extracted and sequenced follow by gene expression analysis as described by Pinto et al. (2021) 19 .

In silico mapping
To verify the localization of the DEGs within the previously described Al tolerance QTLs [4][5][6][7][8] , the positions of each delimiting QTL marker identified in different segregating populations in common maize (Zea mays) were examined using the MaizeGDB database (Table 3).To identify the DEGs of interest identified in our transcriptomic analysis inside the QTLs, we performed a simple comparison between the positions of the genes and the markers on the chromosome, as described in Mattiello et al. 41 .

Polymorphism discovery and protein sequence variation prediction
The maize reference genome (B73.RefGen_v4) was downloaded from Phytozome database (https:// phyto zome.jgi.doe.gov).The reads were mapped to the reference genome using BWA-MEM algorithm of BWA version 0.7.17 (http:// bio-bwa.sourc eforge.net/) 53 .A flag was added to identify the respective popcorn sample in each mapping file.The mapping files were processed using SortSam, MarkDuplicates and BuildBamIndex tools of Picard version 2.18.27 (https:// github.com/ broad insti tute/ picard/).Variants were called using FreeBayes version 1.2.0 (https:// github.com/ ekg/ freeb ayes) 54 with a minimum mapping quality of 20, minimum base quality of 20, and minimum coverage of 20 reads at every position in the reference genome.After variant calling, SNPs were filtered using vcftools version 0.16.15 (https:// vcfto ols.github.io/ index.html) and annotated using Variant Effect Predictor 55 available on Ensembl Plants web server (http:// plants.ensem bl.org/ Tools/ VEP).The translated protein sequences were the analyzed using the PROVEAN (Protein Variation Effect Analyzer) software tool (http:// prove an. jcvi.org/ index.php) to predict whether an amino acid substitution or InDel has an impact on the biological function of a protein.A PROVEAN score less than or equal to -2.5 was considered to have a significant biological impact on the protein function.

Figure 1 .
Figure 1.Polymorphisms among differentially expressed genes in Al-contrasting inbred lines under Al-stress.

Figure 3 .
Figure 3. Heatmap of the unweighted pair group method (UPGMA) clustering of co-regulated genes between Al-tolerant and Al-sensitive inbred lines mapped within Al-related QTL regions in maize.

Table 1 .
Predicted deleterious variations in differentially expressed genes in the 11-133 Al-tolerant inbred line under Al stress.

Table 2 .
Predicted deleterious variations in differentially expressed genes in the 11-60 Al-sensitive inbred line under Al stress.

Table 3 .
Al-tolerance QTLs identified in previous work using different genotypes of common maize (Zea mays).