Maize transcriptome profiling reveals low temperatures affect photosynthesis during the emergence stage

Abstract

Introduction: Earlier sowing is a promising strategy of ensuring sufficiently high maize yields in the face of negative environmental factors caused by climate change. However, it leads to the low temperature exposure of maize plants during emergence, warranting a better understanding of their response and acclimation to suboptimal temperatures.

Materials and methods: To achieve this goal, whole transcriptome sequencing was performed on two maize inbred lines - tolerant/susceptible to low temperatures, at the 5-day-old seedling stage. Sampling was performed after 6h and 24h of treatment (10/8°C). The data was filtered, mapped, and the identified mRNAs, lncRNAs, and circRNAs were quantified. Expression patterns of the RNAs, as well as the interactions between them, were analyzed to reveal the ones important for low-temperature response.

Results and discussion: Genes involved in different steps of photosynthesis were downregulated in both genotypes: psa, psb, lhc, and cab genes important for photosystem I and II functioning, as well as rca, prk, rbcx1 genes necessary for the Calvin cycle. The difference in low-temperature tolerance between genotypes appeared to arise from their ability to mitigate damage caused by photoinhibition: ctpa2, grx, elip, UF3GT genes showed higher expression in the tolerant genotype. Certain identified lncRNAs also targeted these genes, creating an interaction network induced by the treatment (XLOC_016169-rca; XLOC_002167-XLOC_006091-elip2). These findings shed light on the potential mechanisms of low-temperature acclimation during emergence and lay the groundwork for subsequent analyses across diverse maize genotypes and developmental stages. As such, it offers valuable guidance for future research directions in the molecular breeding of low-temperature tolerant maize.

Keywords: climate change; low temperature stress; maize; photosynthesis; whole transcriptome profiling.

Figure 1

The RNA data analysis pipeline. Raw reads, acquired through 150 bp PE sequencing, were passed through quality control and processed to remove reads containing adapter and poly-N sequences and low quality reads (Phred score > 30, N% < 10%). The trimmed reads were then aligned onto the Zea mays B73 NAM 5.0 reference genome. mRNAs could be quantified after mapping, but lncRNAs and circRNAs required additional steps. To identify lncRNAs, the reads were assembled, and filtered (by length, expression level, exon number, and genome position), in addition to assessing the absence of coding potential (CPC2, CPAT, Plek). Additionally, the potential cis- and trans-targets of the identified lncRNAs were predicted. circRNA detection was accomplished through a custom pipeline in Circexplorer2. After all the coding and non-coding RNAs were identified, the quantification, differential expression (DE) analysis, and network construction were performed. Furthermore, functional enrichment (GO, KEGG) analysis was performed for the identified DE mRNAs. Bioinformatics tools used in each step are also shown next to the box containing the step name.

Figure 2

mRNA expression summary. (A) Percentage of explained variances of selected principal components (PCs). (B) PCA graph of individual libraries, explained by the first two PCs (PC1 and PC2). (C) mRNAs differentially expressed between the control and treatment, in the two genotypes (L_S, L_T), at the two time points (6h and 24h). Up-regulated mRNAs are shown in blue, while down-regulated are in orange. (D) Hierarchical cluster analysis, based on the expression patterns expressed in FPKM of 26,023 mRNAs across the eight libraries.

Figure 3

Differentially expressed (DE) genes. (A) Venn diagram showing the common and unique differentially expressed (DE) genes for each comparison: L_S 6h is shown in orange, L_S 24h in yellow, L_T 6h in blue, and L_T 24h in green. (B) Expression profiles of selected DE genes for each comparison (L_S 6h, L_S 24h, L_T 6h, and L_T 24h). Fold change is shown as its log 2 value (log2FC), in the range from -10 (blue), over 0 (white), to 10 (red).

Figure 4

Functional enrichment of differentially expressed mRNAs. (A) Bar chart showing the results of over-representation analysis (ORA) for the Biological Process GO class. (B) Emap plot showing the results of over-representation analysis (ORA) for the Biological Process GO class.

Figure 5

Summary of the lncRNA identification and quantification. (A) Potential lncRNAs identified through the three different approaches: CPC2 (yellow), CPAT (orange), and Plek (blue); and the identified lncRNAs common for different pairs and all three approaches. (B) Classification of identified lncRNAs into four classes: intergenic (orange), intronic (purple), sense (blue), and antisense lncRNAs (yellow). (C) lncRNA FPKM distribution across the libraries (Ls_C_6, Ls_C_24, Ls_T_6, Ls_T_24, Lt_C_6, Lt_C_24, Lt_T_6, Lt_T_24).

Figure 6

Differentially expressed lncRNA. (A) lncRNAs differentially expressed between the control and treatment, in the two genotypes (L_S, L_T), at the two time points (6h and 24h). Up-regulated lncRNAs are shown in blue, while down-regulated are in orange. (B) Unique and common differentially expressed lncRNAs between the two genotypes (L_S, L_T) and time points (6h and 24h). Ls_6h is shown in orange, Ls_24h in yellow, while Lt_6h is presented in blue, and Lt_24h in green.

Figure 7

Identified circRNAs. (A) circRNA FPKM distribution across the samples (Ls_C_6, Ls_C_24, Ls_T_6, Ls_T_24, Lt_C_6, Lt_C_24, Lt_T_6, Lt_T_24) of the selected 135 circRNAs. (B) Length distribution across the eight libraries of the selected 135 circRNAs. (C) Exon number of the selected 135 circRNAs. 83% of circRNAs had one exon and are shown in blue, those with two accounted for 14% and are shown in orange, while the rest of circRNAs were represented with 1% in each category: circRNAs with three exons shown in yellow, those with four in purple and six in light blue. (D) The number of parent genes (y-axis) that generated different numbers of circRNAs (x-axis).

Figure 8

The lncRNA-mRNA coexpression network. Orange nodes represent the lncRNAs, while the blue rectangles represent the target mRNAs. Positive correlation between the lncRNA and mRNA expression is shown with a green arrow, while the negative is shown with a red one. The network showed that the expression of 19 lncRNAs was correlated to the expression patterns of 41 target mRNA.