COI1 F-box proteins regulate DELLA protein levels, growth, and photosynthetic efficiency in maize

Abstract

The F-box protein Coronatine Insensitive (COI) is a receptor for the jasmonic acid signaling pathway in plants. To investigate the functions of the 6 maize (Zea mays) COI proteins (COI1a, COI1b, COI1c, COI1d, COI2a, and COI2b), we generated single, double, and quadruple loss-of-function mutants. The pollen of the coi2a coi2b double mutant was inviable. The coi1 quadruple mutant (coi1-4x) exhibited shorter internodes, decreased photosynthesis, leaf discoloration, microelement deficiencies, and accumulation of DWARF8 and/or DWARF9, 2 DELLA family proteins that repress the gibberellic acid (GA) signaling pathway. Coexpression of COI and DELLA in Nicotiana benthamiana showed that the COI proteins trigger proteasome-dependent DELLA degradation. Many genes that are downregulated in the coi1-4x mutant are GA-inducible. In addition, most of the proteins encoded by the downregulated genes are predicted to be bundle sheath- or mesophyll-enriched, including those encoding C4-specific photosynthetic enzymes. Heterologous expression of maize Coi genes in N. benthamiana showed that COI2a is nucleus-localized and interacts with maize jasmonate zinc-finger inflorescence meristem domain (JAZ) proteins, the canonical COI repressor partners. However, maize COI1a and COI1c showed only partial nuclear localization and reduced binding efficiency to the tested JAZ proteins. Together, these results show the divergent functions of the 6 COI proteins in regulating maize growth and defense pathways.

Figure 1.

The maize genome encodes 6 COI proteins. The sequences of 6 maize COI proteins, along with their paralogs from other species, were obtained from Phytozome (https://phytozome-next.jgi.doe.gov/). Two predicted duplication events are marked. Four maize COI1 proteins are marked with green boxes. Two maize COI2 proteins are marked with blue boxes. The species in the Poaceae have 3 COI protein clades, 2 containing COI1 proteins (exemplified by maize COI1a and COI1d and COI1b and COI1c, respectively) and one containing COI2 proteins (exemplified by maize COI2a and COI2b). The protein sequence alignment and species names are presented in Supplementary Data Set 2. A machine-readable (Newick Format) version of the tree is in Supplementary File 1. FASTA sequences of all proteins are in Supplementary File 2. The maximum likelihood, midpoint-rooted tree was produced with IQ-Tree and visualized with MEGA11. Bootstrap values are based on 1,000 replicates. Scale bar indicates substitutions per site.

Figure 2.

Subcellular locations of the maize COI proteins. COI1a, COI1c, and COI2a proteins were fused to EGFP, showing that COI2a is more targeted to the nuclei than the 2 COI1 proteins. Confocal images were taken at 48 h after infiltrating plasmids expressing the COI-EGFP fusion proteins into the N. benthamiana. Leaves were sprayed with water (mock) or 0.02% MeJA 2 h before imaging. White arrows indicate the nuclei, and red arrow represents a cytosolic condensate. Scale bars = 25 μm.

Figure 3.

BIFC between the maize COI proteins fused to the carboxy-terminus of yellow fluorescent protein (cYFP) and JAZ proteins fused to the amino-terminus of yellow fluorescent protein (nYFP). BIFC indicating co-localization is shown as white spots in the images. Magenta represents chlorophyll fluorescence. Maize JAZ proteins have a higher affinity for COI2a than COI1a or COI1c. Leaves were sprayed with 0.02% MeJA 2 h before imaging. Scale bars = 50 μm. These experiments were repeated independently with similar results (Supplementary Fig. S4). JAZ protein names are as in Han and Luthe (2021).

Figure 4.

Fall armyworm (S. exigua) and beet armyworm (S. exigua) caterpillar growth on Coi mutant and wild-type maize. A) Locations of Ds transposon insertions in COI1 genes and Mu transposon insertions in COI2 genes are marked with triangles. Insertions are located at 140, 2682, 344, 2342, −8, and 198 bp from the start codons of COI1a, COI1d, COI1c, COI1b, COI2a, and COI2b, respectively. Black bars represent exons, thin lines are introns, and white bars are noncoding regions of the genes. B) Mass of 10-d-old S. frugiperda caterpillars on wild-type and coi1-4x maize. No significant difference (P > 0.05, t-test). C) Mass of 10-d-old S. exigua caterpillars on wild-type and coi1-4x. *P < 0.05, t-test. D) Mass of 10-d-old S. frugiperda on wild-type and coi2 mutants. No significant difference (NS, P > 0.05), Dunnett's test relative to wild-type-Mu. All data are mean +/− s.e., numbers in bars indicate the number of caterpillars for each treatment. Raw data and statistical calculations are in Supplementary Data Set 18.

Figure 5.

Plants with mutations in 4 maize coi genes (coi1-4x) have impaired growth relative to the corresponding double mutants, coi1a coi1d and coi1b coi1c and wild-type inbred line W22 maize. A) and B) Lengths of the third leaf (dashed arrows in A)) of coi-4x and wild type (WT) 10 d after germination, N = 4 plants, mean +/− s.e., 2-tailed Student's t-test, *P < 0.05, **P < 0.01. Scale bar in A) = 10 cm. C) and D) Plant heights at 60 d after germination. Numbers in bars = number of plants of each genotype, mean +/− s.e., letters indicate significant differences, P < 0.05, ANOVA followed by Tukey's HSD test. Scale bar in C) = 22 cm. E) and F) Internode lengths were compared between 4 genotypes at 60 d post-germination. Positions of stem nodes are marked with white arrows in B). coi1-4x N = 13 plants, WT N = 6 plants, coi1a coi1d and coi1b coi1c N = 5 plants, mean +/− s.e., ***P < 0.01, Dunnett's test relative to WT. Scale bar in E) = 22 cm. Raw data and statistical calculations are in Supplementary Data Set 18.

Figure 6.

Plants with mutations in 4 maize coi genes (coi1-4x) have striped leaves, decreased microelement levels, and reduced photosynthesis. A) The striped-leaf phenotype of the coi1-4x at 20 d post-germination compared to corresponding leaves from the double mutants and WT W22. Scale bar = 5 cm. B) Microelements and C) macroelements in 20-d-old seedlings, N = 7 plants, mean +/− s.e., letters indicate differences (P < 0.05) using Tukey's HSD test. D) Leaf CO₂ assimilation rate (GasEX A) at 400 µmol mol⁻¹ CO₂ and E) the quantum yield of the PSII phytochemistry (øPSII) at 2,000 μmol m⁻² s ⁻¹ of actinic light were measured at 20 d after germination, respectively. Mean +/− s.e., n = 20 plants, 2-tailed Student's t-test, **P < 0.01, ***P < 0.001. Raw data and statistical calculations are in Supplementary Data Set 18.

Figure 7.

Heatmap of gene expression for 13,365 genes that were differentially regulated between 2 or more genotypes or between mock and MeJA induction. Color ranges from blue (minimum) to red (maximum) RPM bp for each gene. The color key represents the normalized log(1 + expression) ranging from blue (indicating low values) to red (indicating high values). RPM data were transformed by log(1 + RPM) prior to clustering. The color gradient in the heatmap reflects the Z scores, with red colors indicating higher-than-average expression levels (Z score > 0) and blue colors indicating lower-than-average expression levels (Z score < 0). Numerical data underlying the heatmap are in Supplementary Data Set 6.

Figure 8.

Genes downregulated in maize plants with mutations in 4 coi genes (coi1-4x) encode 2 main groups of proteins, involved in C₄ photosynthesis and carbohydrate and cell wall metabolism. A) Heatmap showing downregulated genes in coi1-4x relative to other genotypes, with or without MeJA treatment. RPM data were transformed by log(1 + RPM) prior to clustering. The color gradient in the heatmap reflects the Z scores, with red colors indicating higher-than-average expression levels (Z score > 0) and blue colors indicating lower-than-average expression levels (Z score < 0). B) Downregulated genes in coi1-4x were categorized into 5 functional groups. The list of the genes, color-coded by their functional group, as well as ordered based on their position in the heatmap, is presented in Supplementary Data Set 7.

Figure 9.

Genes upregulated in maize plants with mutations in 4 Coi genes (coi1-4x) encode proteins involved in phosphate regulation, lipid and carbohydrate metabolism, hormone regulation, and transcription. A) Heatmap showing upregulated genes in coi1-4x relative to other genotypes, with or without MeJA treatment. The reads are transformed by log(1 + expression) prior to clustering. B) Upregulated genes in coi1-4x were categorized into 6 functional groups. The list of the genes, color-coded by their functional group, as well as ordered based on their position in the heatmap, is presented in Supplementary Data Set 8.

Figure 10.

Immunoblot analyses of maize DELLA proteins. The leaf total proteins from equal surface area (similar weight) leaves (fourth leaf, 20 d after germination) of WT maize inbred line W22 and a line with mutations in 4 Coi genes (coi1-4x) were analyzed by probing with antibodies that react with rice DELLA (SLR1). The membrane was stained with Ponceau S as the loading control.

Figure 11.

Maize DELLA (DWARF9) disappears from the nuclei upon coexpression with the maize COI proteins. Confocal images were taken after transiently expressing genes in N. benthamiana leaves for 48 h. Labels on the left indicate infiltrated genes for each row of images. GFP, green fluorescent protein; EGFP, enhanced green fluorescent protein, RFP, red fluorescent protein. Protease inhibitor (BTZ) was co-infiltrated in the bottom row. Scale bars = 50 μm.

Figure 12.

Expression of the maize COI1 proteins leads to proteasome-dependent degradation of maize DELLA in N. benthamiana. A) Nuclei showing the presence of the COI1a-EGFP and or DELLA-RFP were counted on confocal images from an experiment similar to Fig. 11, with and without protease inhibitor. The bars are means +/− s.e. of N = 48 plants for the mock treatment and N = 71 plants treated with the proteasome inhibitor BTZ. Raw data and statistical calculations are in Supplementary Data Set 18. B) Leaf tissue from the N. benthamiana plants expressing DELLA-RFP alone or with EGFP, COI1a-EGFP, or COI1c-EGFP (similar to those used in Fig. 11), 48 h after infiltration, was used for immunoblot analyses. Membranes were probed with antibodies that react with rice DELLA (SLR1) and antibodies that react with GFP. The first membrane was stained with Ponceau S as the loading control. Nonspecific binding of the SLR1 antibody to an unrelated protein was used as an extra loading control.

Figure 13.

Effect of exogenous MeJA and GA on growth of WT maize inbred line W22 and plants with mutations in 4 maize coi genes (coi1-4x). A) Leaf discoloration symptoms at 30 d post-germination and after 3 wk of JA treatment. Scale bar = 5 cm. B) Plant heights at 23 d post-germination, with mock, MeJA or GA treatments. Scale bar = 10 cm. C) Plant heights at 30 d post-germination, with or without MeJA treatment. WT, wild type. Scale bar = 10 cm. D) Bar chart showing percent change in height relative to the mock-treated controls for the MeJA and GA treatments shown in B) and C). N = 4, mean +/− s.e., 2-tailed Student's t-test, *P < 0.05. Raw data and statistical calculations are in Supplementary Data Set 18.

Figure 14.

Model for differential functions of maize COI proteins in regulating growth and defense. A) Maize COI2 proteins are proposed to have the classical F-box domain protein function of Arabidopsis and tomato COIs. At low JA-Ile concentrations (black arrow pointing down), JAZ proteins prevent activation of JA-Ile-responsive genes by MYC transcription factor (black right-angle arrow with a red X). Higher abundance of JA-Ile (black arrow pointing up) leads to interactions between COI2 and JAZ proteins (arched gray arrow), causing the addition of ubiquitin (U) to JAZ and degradation (blue arrow) by the 16S proteasome (red triangles). The absence of JAZ proteins leads to the activation of transcription by MYC proteins and expression of defense-related genes (black right-angle arrow). B) At low GA levels (black arrow pointing down), DELLA is bound to PIF transcription factors, preventing transcription of GA-responsive genes (black right-angle arrow with a red X). At high GA levels, an as yet unknown maize F-box protein contributes to an E3 ligase complex (dashed blue arrow) that causes the addition of ubiquitin (U) to DELLA and degradation (blue arrow) by the 16S proteasome, leading to activation of GA-responsive genes (black right-angle arrow). Based on the results that are presented, we propose that maize COI1 proteins directly or indirectly cause DELLA degradation (green triangles). In the coi1-4x mutant, there is no COI1, resulting in less DELLA degradation, less growth, and shorter stature of the maize plants.