ZmCesA10-12 function in SCW biosynthesis
In Arabidopsis, AtCesA4/7/8 function in SCW biosynthesis. In rice, OsCesA4/7/9 functions in SCW biosynthesis. There are 12 CesA genes in maize (Appenzeller et al., 2004). Functions of those CesAs in PCW and SCW of maize were uncharacterized. In order to categorize 12 maize ZmCesAs, a phylogenetic tree with amino acid sequences of 12 maize ZmCesAs, AtCesA4/7/8, and OsCesA4/7/9, was constructed. ZmCesA10 diverged close to AtCesA4 and OsCesA7 (Fig. 1A). ZmCesA11 diverged close to AtCesA8 and OsCesA4 (Fig. 1A). ZmCesA12 diverged close to AtCesA7 and OsCesA9 (Fig. 1A). Thus, CesA10-12 function in maize SCW cellulose biosynthesis was proposed.
A further amino acid sequence alignment with ZmCesA10-12, OsCesA4/7/9, and AtCesA4/7/8 was conducted to check conserved motifs or domains in ZmCesA proteins. As a result, a conserved domain, Cellulose_synt, was found in ZmCesA10-12 (Fig. 1B), indicating a conserved sequence among CesA protein in both monocotyledonous and dicotyledonous plants.
ZmCesA10-12 localized in cell plasma
Recombined expression vectors with ZmCesA10-12, integrated with GFP, were constructed, and cotransformed into maize protoplasts, as well as a cell plasma (PM) marker tagged with RFP, to visualize their subcellular localization under fluorescence confocal microscopy. ZmCesA10-12 tagged GFP signals merged with RFP signals (Fig. 2B-D), indicating ZmCesA10-12 localize in cell plasma, which was consistent with their putative function in SCW cellulose biosynthesis.
ZmCesA10 - 12 co-expressed in maize tissues
Specific qPCR primers for 12 ZmCesA amplifications (Table S1) were designed to check their expression models in different maize tissues and organs. Wild type B73 at the flowering stage was sampled. RNA was extracted from maize roots, stems, leaves, silks, tassels, embryos, and endosperm. RNA samples were reverse transcripted into cDNA. The expression level of 12 ZmCesA genes was analyzed (Figure S1). In the root, a relatively high expression level was found in ZmCesA2, ZmCesA8, and ZmCesA11 (Figure S1A). In silk, the highest expression level was found in ZmCesA2 (Figure S1B). In culm, a relatively high expression level was found in ZmCesA2, ZmCesA10, and ZmCesA11 (Figure S1C). In embryos, a relatively high expression level was found in ZmCesA2, ZmCesA3, and ZmCesA7 (Figure S1D). In endosperm, the highest expression level was detected in ZmCesA5 (Figure S1E). In tassel, the highest expression level was detected in ZmCesA2 (Figure S1F). In the leaf, the highest expression level was detected in ZmCesA2 (Figure S1G). It was noteworthy that ZmCesA10-12 expressed highly in root and culm (Figure S1).
Expression of ZmCesA1-12 was further checked in the meristematic zone, elongation zone, and maturation zone of the root (Fig. 3A). Expressions of ZmCesA1, 3, 4, and 9 increased from the meristematic zone to the elongation zone, while decreasing from the elongation zone to the maturation zone (Fig. 3A). ZmCesA2, 5, and 8 had a comparable expression level in elongation and maturation zones (Fig. 3A). Expression of ZmCesA6 and 7 increased significantly from the meristematic zone to the elongation zone, and finally maturation zone (Fig. 3A). A 10-fold increment in expression was found in the maturation zones of ZmCesA10, 11, and 12 when compared with that in elongation zones (Fig. 3A). In summary from this part, a similar expression tendency of ZmcesA10-12 was found, indicating their functional relevance.
Expression vectors with GUS driven by native promoters of ZmCesA10-12 were constructed and later transformed into maize KN5585 to visualize their expressed organs and expression abundances (Fig. 3B-H). Transgenic leaves and roots were visualized after GUS staining. ZmCesA10-12 is expressed in the elongation and maturation zones of the root, not in root tips or meristematic zones of the root (Fig. 3B-H). Expression of GUS was also detected in veins and border areas of leaves (Fig. 3B-H).
Pairwise protein-protein interactions among ZmCesA10-12
CSC rosette contains 18–24 CesA proteins. A CSC rosette has at least 3 kinds of CesA, thus protein-protein interactions among ZmCesA10-12 were verified via Y2H and BiFC. No auto-activations were found in ZmCesA10-12 (Fig. 4A). ZmCesA10 can form dimer since their self-interactions, ZmCesA10 had protein-protein interactions with ZmCesA11-12 (Fig. 4C). ZmCesA11 can form dimer since their self-interactions, ZmCesA11 had protein-protein interactions with ZmCesA10 and ZmCesA12 (Fig. 4D). ZmCesA12 can’t form dimer since no interaction was found between PGBKT7-ZmCesA12 and pGADT7-ZmCesA12, ZmCesA12 had protein-protein interactions with ZmCesA11 and ZmCesA12 (Fig. 4E).
BiFc was applied to further verify the interactions among ZmCesA10-12. ZmCesA10 can form dimers since their self-interactions (Fig. 5A). pUC-SPYCE-ZmCesA10 interacted with pUC-SPYNE-ZmCesA11-12 (Fig. 5A). ZmCesA11 can form dimers since their self-interactions (Fig. 5A). pUC-SPYCE-ZmCesA10 interacted with pUC-SPYNE-ZmCesA10 and pUC-SPYNE-ZmCesA12 (Fig. 5B). ZmCesA12 can’t form dimer since no interaction between pUC-SPYCE-ZmCesA12 and pUC-SPYNE-ZmCesA12 (Fig. 5C). pUC-SPYCE-ZmCesA12 interacted with pUC-SPYNE-ZmCesA10-11 (Fig. 5A).
In summary from this part, we concluded that there were protein-protein interactions among ZmCesA10-12, ZmCesA10-ZmCesA11 can form dimers while ZmCesA12 can’t.
Mutants for ZmCesA10-12 were screened
Mutants for ZmCesA10-12, either having stop-gained or non-synonymous mutation, were selected from an EMS-induced mutant library of maize B73 (Lu et al., 2008). A three-generations of crossings and self-crossings between mutants and B73 were made respectively. Finally, four mutants, namedcesa10-1, cesa11-1, cesa11-2, and cesa12-1, which have homozygous mutations in ZmCesA10-12 respectively, were harvested. The schematic structures of cesa10-1, cesa11-1, cesa11-2, and cesa12-1, was displayed in Fig. 6A-C.
Mutant cesa10-1 had a mutation from C to T at 2,933 bases of the genomic sequence of ZmCesA10, leading to an amino acid change from Arg to Thr (Fig. 6A). Mutant cesa11-1 had a mutation from C to T at 3,573 bases of the genomic sequence of ZmCesA11, leading to an amino acid change from Pro to Ser (Fig. 6B). Mutant cesa11-2 had a mutation from G to A at 5,169 bases of the genomic sequence of ZmCesA11, leading to an early termination of protein translation (Fig. 6B). Mutant cesa12-1 had a mutation from G to A at 3,780 bases of the genomic sequence of ZmCesA12, leading to an early termination of protein translation (Fig. 6C).
Harvest-related traits of ZmCesA10-12 mutants
The plant height of B73 was 193.25 cm. Mutant cesa10-1 had a comparative plant height with that of B73 (Fig. 6E, L). The plant height was reduced by approximately 2.5% in the cesa11-1 mutant (**p < 0.01, Fig. 6F, L). The plant height of cesa11-2 significantly reduced, compared with that of B73 (Fig. 6G). The cesa11-2 plant became lighter blue, compared with the dark blue visualized in B73 (Fig. 6G). Mutant cesa11-2 died in its early stage, therefore cesa11-1 was selected for later analysis. The plant height was reduced by approximately 1.9% in the cesa12-1 mutant (*p < 0.05, Fig. 6H, L).
The mean ear length of B73 was 143.07 mm, it reduced significantly by 11.6% and 12.9% in cesa10-1 and cesa12-1 respectively (*p < 0.05, Table S2). There were no significant changes in ear width (Table S2). The mean number of kernel rows of B73 was 15.37, it increased by 14.3% in cesa11-1 (*p < 0.05, Table S2).
The grain morphology of cesa10-1, cesa11-1, and cesa12-1 was compared. The grain length and width of B73 were 11.95 mm and 8.59 mm respectively. Grain length changed without significance, while grain width decreased by 13.8% in cesa10-1, compared with that of B73 (***p < 0.001, Fig. 6M-N). Hundred-grain weight of B73 was 28.83 g, it decreased by 6.7% in cesa10-1, compared with that of B73 (**p < 0.01, Fig. 6O). Grain width decreased significantly by 23.7% in cesa11-1 (***p < 0.001, Fig. 6L, N). The grain length and hundred-grain weight of cesa11-1 had no significant difference from that of B73 (Fig. 6M). Compared with that of B73, the grain length of cesa12-1 reduced significantly by 6.0% (***p < 0.001, Fig. 6M), the grain width of cesa12-1 decreased significantly by 15.1% (***p < 0.001, Fig. 6N), the hundred-grain weight of cesa12-1 decreased significantly by 6.3% (**p < 0.01, Fig. 6O).
The increased culm brittleness of mutants was characterized
Mutants of ZmCesA10-12 displayed increased culm brittleness or decreased culm strength. Mutant cesa10-1 was easier to be break, it had an increased culm and leaf vein brittleness, compared with that of B73 (Fig. 7A, D). Culm strengths were reduced by approximately 39% in cesa10-1 (***p < 0.001, Fig. 7G). Mutant cesa11-1 culms and veins could be easily broken, showing smooth breakpoints (Fig. 7B, E, Figure S2). Culm strength was reduced by approximately 70% in cesa11-1 (***p < 0.001, Fig. 7G). Mutant cesa11-2 culms and veins could be easily broken, showing smooth breakpoints (Figure S2). The brittleness of the cesa12-1 stem was significantly increased compared with that of B73 (Fig. 7C, F). Mutant cesa12-1 culms and veins could be easily broken (Fig. 7C, F). Culm strength was reduced by approximately 7% in cesa12-1 (***p < 0.001, Fig. 7G).
Hybridizations were made between B73 and mutants of cesa10-1, cesa11-1, and cesa12-1. F1 progenies were further self-crossed to generate F2 generations. The population sizes of cesa10-1, cesa11-1, and cesa12-1 F2 generation were 198, 173, and 198187 respectively. The segregation ratios of the mutated progenies and the non-mutated progenies were approximately 1:3 (Degree of freedom = 1, χ2 = 1.13, 1.85, and 2.71), indicating the mutated phenotype, increased brittleness, is the result of a mutation of a single locus.
To further confirm the casual locus of the above mutants to be ZmCesA10-12 respectively. F2 progeny individuals showing increased brittleness phenotypes were selected for the following exon-capture sequencing. The overall mapping result was given in Dataset S1. The SNPs (either G > A or C > T) with an SNP index equal to 1 were selected as candidate mutations. As a result, an SNP of ZmCesA10 (The same SNP shown in Fig. 6A) was included in the candidate SNPs of cesa10-1. An SNP of ZmCesA11 (The non-synonymous SNP of cesa11-1 shown in Fig. 6B) was included in the candidate SNPs of cesa11-1. An SNP from ZmCesA12 (The same SNP shown in Fig. 6C) was included in the candidate SNPs of cesa12-1. In summary from the above results, we concluded that mutations in ZmCesA10-12 were responsible for the increased brittleness phenotype.
Reduced cell-wall thickness and cellulose content in mutants
The Cell walls of culm epidermal cells in cesa10-1, cesa11-1, cesa11-2, and cesa12-1 became thinner, compared with that in B73. Cell walls of culm epidermis cells and vascular bundle of cesa10-1, cesa11-1, cesa11-2, and cesa12-1 were sampled and imaged via SEM. Cell wall thickness was recorded and compared via ImageJ. Cell walls of culm epidermal cells and vascular bundle in cesa10-1, cesa11-1, cesa11-2, and cesa12-1 became thinner, compared with that in B73 (Fig. 8A-H). The cell wall thickness of culm epidermal cells and culm vascular bundle decreased significantly (***p < 0.001) in cesa10-1, cesa11-1, cesa11-2, and cesa12-1 (Fig. 8I-J).
Changes in cell wall morphology and cell wall components would result in changes in plant brittleness. The cellulose content of mutants and B73 was measured. Cell walls of mutants and B73 were observed via SEM imaging. Reduced cellulose contents were found in cesa10-1 (**p < 0.01), cesa11-1 (***p < 0.001), cesa11-2 (***p < 0.001), and cesa12-1 (**p < 0.01) (Fig. 8K). Cellulose content was reduced by approximately 30% in cesa10-1, 39% in cesa11-1, and 20% in cesa12-1 (Fig. 8K). Mutations of ZmCesA genes caused a significant reduction of cellulose biosynthesis, which was as expected.
Based on these results, we proposed that mutations of ZmCesA 10–12, which was found functioning in SCW biosynthesis, resulted in a decreased cellulose content, a changed cell wall morphology, and a decreased cell wall thickness.