ZmFAR1 and ZmABCG26 Regulated by microRNA Are Essential for Lipid Metabolism in Maize Anther

Abstract

The function and regulation of lipid metabolic genes are essential for plant male reproduction. However, expression regulation of lipid metabolic genic male sterility (GMS) genes by noncoding RNAs is largely unclear. Here, we systematically predicted the microRNA regulators of 34 maize white brown complex members in ATP-binding cassette transporter G subfamily (WBC/ABCG) genes using transcriptome analysis. Results indicate that the ZmABCG26 transcript was predicted to be targeted by zma-miR164h-5p, and their expression levels were negatively correlated in maize B73 and Oh43 genetic backgrounds based on both transcriptome data and qRT-PCR experiments. CRISPR/Cas9-induced gene mutagenesis was performed on ZmABCG26 and another lipid metabolic gene, ZmFAR1. DNA sequencing, phenotypic, and cytological observations demonstrated that both ZmABCG26 and ZmFAR1 are GMS genes in maize. Notably, ZmABCG26 proteins are localized in the endoplasmic reticulum (ER), chloroplast/plastid, and plasma membrane. Furthermore, ZmFAR1 shows catalytic activities to three CoA substrates in vitro with the activity order of C12:0-CoA > C16:0-CoA > C18:0-CoA, and its four key amino acid sites were critical to its catalytic activities. Lipidomics analysis revealed decreased cutin amounts and increased wax contents in anthers of both zmabcg26 and zmfar1 GMS mutants. A more detailed analysis exhibited differential changes in 54 monomer contents between wild type and mutants, as well as between zmabcg26 and zmfar1. These findings will promote a deeper understanding of miRNA-regulated lipid metabolic genes and the functional diversity of lipid metabolic genes, contributing to lipid biosynthesis in maize anthers. Additionally, cosegregating molecular markers for ZmABCG26 and ZmFAR1 were developed to facilitate the breeding of male sterile lines.

Keywords: anther and pollen development; genic male sterility; lipid metabolism; maize (Zea mays); microRNA; noncoding RNA.

Figure 1

Posttranscriptional regulation of ZmABCG26 by zma-miR164h-5p during maize anther development: (A) the phylogenetic tree of maize ABCG family genes and their predicted miRNA regulators during anther development; (B) transcriptome analyses of expression patterns of predicted miRNA regulators of maize WBC/ABCG family genes in anthers of WT inbred lines (W23, B73, and Oh43) and GMS mutant lines (ocl4, mac1, ms23). The names of predicted miRNA regulators having negatively correlated expression levels with their ABCG target genes (Pearson correlation test) were marked in blue; (C) the sequence alignment of zma-miR64h-5p and its target site on the ZmABCG26 transcript. *, base mismatch; (D,E) the negatively correlated expression patterns between ZmABCG26 and zma-miR164h-5p in B73 (D1,E1) and Oh43 (D2,E2) inbred lines at anther developmental stages 9 and 10 revealed by transcriptome analyses (D) and at stages 9 to 11 confirmed by qRT-PCR analyses (E). r, Pearson correlation coefficient.

Figure 2

Phylogenetic tree of WBC/ABCG family members, phenotypic characterization, sequencing analysis, and cosegregating molecular marker development of CRISPR/Cas9-induced ZmABCG26 male-sterile mutants: (A) phylogenetic tree of the WBC/ABCG family members in Arabidopsis (At), rice (Os), and maize (Zm). A neighbor-joining tree showed the evolutionary relationships of WBC/ABCG members in the three species; (B) physical map of pCas9-ZmABCG26 construct carrying two gRNAs and the information of target sites in ZmABCG26; (C) gene structure of ZmABCG26 and mutation analysis of three knock-out lines (ZmABCG26-Cas9-1, -2, and -3) created by the CRISPR/Cas9 technology; (D) DNA sequencing of targeted mutation sites or fragments among WT and the three knockout lines; (E) comparison of tassels, anthers, and pollen grains stained with I₂-KI among WT and the three knockout lines; (F) SEM analysis of the whole anthers, pollen grains, anther outer and inner surfaces in WT and one abcg26 mutant at anther developmental stage 13; (G) development of cosegregating molecular markers for genotyping ZmABCG26-Cas9-2 mutation in F₂ plants.

Figure 3

Spatiotemporal expression analysis of ZmABCG26 gene and subcellular localization of ZmABCG26 protein: (A) spatiotemporal expression analysis of ZmABCG26 by qRT-PCR assay with ZmUbi2 (A1) and ZmCynase (A2) as the internal control, respectively. Error bars indicate SD. Each reaction was performed in three biological replicates with three technical repeats; (B) confocal images showing the subcellular localization of ZmABCG26 in maize protoplasts and tobacco leaf cells. The constructs of ZmABCG26-GFP, 35S-GFP, ER-mCherry (mCherry-HDEL), and CFP-AtROP10 were expressed in maize protoplasts (B1) or tobacco leaf cells (B2). The chlorophyll autofluorescence was used as a plastid marker. The ZmABCG26-GFP was cotransformed with the ER-mCherry as an ER marker in maize protoplasts and tobacco leaf cells, and also cotransformed with the CFP-AtROP10 as a plasma membrane marker in tobacco leaf cells. The 35S-GFP vector was used as a negative control. Scale bars, 8 µm (B1) and 40 µm (B2).

Figure 4

Phenotypic characterization, sequencing analysis, and cosegregating molecular marker development of CRISPR/Cas9-induced ZmFAR1 male-sterile mutants: (A) physical map of pCas9-ZmFAR1 construct carrying two gRNAs and the information of target sites in ZmFAR1; (B) gene structure of ZmFAR1 and mutation analysis of three knock-out lines (ZmFAR1-Cas9-1, -2 and -3) created by the CRISPR/Cas9 technology; (C) DNA sequencing of targeted mutation sites or fragments among WT and the three knockout lines; (D) comparison of tassels, anthers, and pollen grains stained with I₂-KI among WT and the three knockout lines; (E) SEM analysis of whole anthers, pollen grains, anther outer and inner surfaces in WT, and one zmfar1 mutant at anther developmental stage 13; (F) development of cosegregating molecular marker for genotyping ZmFAR1-Cas9-3 mutation in F₂ plants.

Figure 5

Spatiotemporal expression analysis of ZmFAR1 and subcellular localization and enzyme activity analysis of ZmFAR1 protein: (A) spatiotemporal expression analysis of ZmFAR1 by qRT-PCR assay with ZmUbi2 (A1) and ZmCynase (A2) as the internal controls, respectively. Error bars indicate SD. Each reaction was performed in three biological replicates with three technical repeats: (B) confocal images showing the subcellular localization of ZmFAR1 in maize protoplasts. The constructs of ZmFAR1-GFP and 35S-GFP were expressed in maize protoplasts. The chlorophyll autofluorescence was used as a plastid marker. The 35S-GFP vector was used as a negative control; (C) ZmFAR1 possessed catalytic activities to three fatty acyl–CoA substrates in vitro: (C1) alignment of ZmFAR1 and its putative orthologs in plants based on the CLUSTALW program. The conserved NAD(P)H binding Rossmann-fold domain (GGTGFLA) and active site motif (YVFTK) in the FAR family were highlighted with the red rectangular borders. The red pentagrams indicated the key amino acids in the GGTGFLA motif for NAD(P)H binding and in the predicted active site motif; (C2) prokaryotic expression and Western blotting analysis of the MBP-tagged protein ZmFAR1 (ZmFAR1-MBP) purified from E. coli. Line 1, SD–SPAGE; line 2, Western blotting; (C3,C4) ZmFAR1-MBP showed different catalytic activities to three substrates C12:0-, C16:0-, and C18:0-CoAs in vitro (C3) with the activity order of C12:0-CoA > C16:0-CoA > C18:0-CoA (C4). Zm, Zea mays; Bd, Brachypodium distachyon, and Si, Setaria italica. * and ** indicated the significant levels of 5% and 1% (Student’s t test, n = 3), respectively; (D) four amino acid mutations (G101A, G104A, Y327F, and K331I) of ZmFAR1-MBP showed a significant reduction of enzymatic activities against all three substrates C12:0-, C16:0-, and C18:0-CoAs in vitro, compared to the WT ZmFAR1-MBP except for G104A against C18:0-CoA. Notes: ** and *** indicated the significant levels of 1% and 1‰ (Student’s t test, n = 3), respectively.

Figure 6

Comparison of genes ZmFAR1 and ZmABCG26 for lipid metabolism of maize anthers by lipidomics analysis: (A) the amount of anther cutin in WT and mutants zmfar1 and zmabcg26 at stage 13: (A1) the total amount of anther cutin per unit surface area; (A2) the changes of 21 specific cutin monomers in zmfar1 and zmabcg26 anthers, compared to those in the WT anther at stage 13; (B) the amount of anther wax in WT and mutants zmfar1 and zmabcg26 at stage 13: (B1) the total amount of wax per unit surface area; (B2) the change range of 20 specific wax constituents in zmfar1 and zmabcg26 anthers at stage 13, compared to those in WT; (C) the number of internal lipid constituents in WT and mutants zmfar1 and zmabcg26 at stage 13: (C1) the total amount of internal lipid constituents per unit dry weight; (C2) the change range of 13 specific internal lipid constituents in zmfar1 and zmabcg26 anthers at stage 13, compared to those in WT. In (A1,B1,C1), ** and *** indicate the significant levels of 1% and 1‰ (Student’s t test, n = 3), respectively. In (A2,B2,C2), the red or blue lines represent the increase or decrease of the cutin, wax, and internal lipid monomers in mutant anthers, respectively. The thickness of the lines represents the change magnitude of the increase or decrease.