Maize Golden2-like transcription factors boost rice chloroplast development, photosynthesis, and grain yield.

Chloroplasts are the sites for photosynthesis, and two Golden2-like factors act as transcriptional activators of chloroplast development in rice (Oryza sativa L.) and maize (Zea mays L.). Rice OsGLK1 and OsGLK2 are orthologous to maize ZmGLK1 (ZmG1) and ZmGLK2 (ZmG2), respectively. However, while rice OsGLK1 and OsGLK2 act redundantly to regulate chloroplast development in mesophyll cells, maize ZmG1 and ZmG2 are functionally specialized and expressed in different cell-specific manners. To boost rice chloroplast development and photosynthesis, we generated transgenic rice plants overexpressing ZmG1 and ZmG2, individually or simultaneously, with constitutive promoters (pZmUbi::ZmG1 and p35S::ZmG2) or maize promoters (pZmG1::ZmG1, pZmG2::ZmG2, and pZmG1::ZmG1/pZmG2::ZmG2). Both ZmG1 and ZmG2 genes were highly expressed in transgenic rice leaves. Moreover, ZmG1 and ZmG2 showed coordinated expression in pZmG1::ZmG1/pZmG2::ZmG2 plants. All Golden2-like (GLK) transgenic plants had higher chlorophyll and protein contents, Rubisco activities and photosynthetic rates per unit leaf area in flag leaves. However, the highest grain yields occurred when maize promoters were used; pZmG1::ZmG1, pZmG2::ZmG2, and pZmG1::ZmG1/pZmG2::ZmG2 transgenic plants showed increases in grain yield by 51%, 47%, and 70%, respectively. In contrast, the pZmUbi::ZmG1 plant produced smaller seeds without yield increases. Transcriptome analysis indicated that maize GLKs act as master regulators promoting the expression of both photosynthesis-related and stress-responsive regulatory genes in both rice shoot and root. Thus, by promoting these important functions under the control of their own promoters, maize GLK1 and GLK2 genes together dramatically improved rice photosynthetic performance and productivity. A similar approach can potentially improve the productivity of many other crops.

Introduction

Chloroplasts are the sites for photosynthesis and also for biosynthesis of phytohormones, reactive oxygen species, and antioxidants that regulate plant growth, development, and stress tolerance (Jarvis and López-Juez, 2013) and as mediators of plant biotic interactions (Fernandez and Burch-Smith, 2019). Golden2-like (GLK) proteins are transcription factors (TFs) that regulate chloroplast development in plants (Rossini et al., 2001; Fitter et al., 2002; Nakamura et al., 2009; Powell et al., 2012; Nguyen et al., 2014). In addition, GLK factors are also involved in biotic (Jhadeswar et al., 2014; Han et al., 2016) and abiotic (Ahmad et al., 2019) stress responses, such as drought tolerance, disease defense and wounding, and regulation of plant metabolism (Lupi et al., 2019).

C3 plants, such as rice (Oryza sativa L.) assimilate atmospheric CO2 through mesophyll (M) cell chloroplasts while their bundle sheath (BS) cells have few chloroplasts and mainly serve a transport function. In contrast, in C4 plants both M and BS cells are rich in chloroplasts (Edwards and Walker, 1983; Sowjanya et al., 2019). Rice OsGLK1 and OsGLK2 are orthologous to maize (Zea mays L.) ZmGLK1 (ZmG1) and ZmGLK2 (ZmG2), respectively (Rossini et al., 2001). However, while OsGLK1 and OsGLK2 act redundantly to regulate chloroplast development in rice M cells, maize chloroplast ZmG1 and ZmG2 are functionally specialized and preferentially expressed in M and BS cells, respectively (Li et al., 2010; Langdale, 2011; Chang et al., 2012; Liu et al., 2013). It has been shown that rice and maize GLK1 proteins are related to each other with a highly conserved structure. In contrast, compared with OsGLK2, ZmGLK2 lacks an intron and some exons in the proline-rich region, resulting in a lower amino acid similarity between ZmGLK2 and OsGLK2 than that between ZmGLK1 and OsGLK1 (Hall et al., 1998; Rossini et al., 2001). Thus, it is possible that maize and rice GLK2 genes have become differentiated in their role in chloroplast development (Chen et al., 2016). Compared to ZmGLK1, ZmGLK2 and its promoter may have acquired stronger and/or new regulatory functions (e.g. cell- and tissue-specific expression) to confer a superior photosynthetic performance in C4 plants.

The regulation of chloroplast biogenesis requires coordinated synthesis of photosynthetic proteins with chlorophyll (Chl) and other photosynthetic components (Zubo et al., 2018). During Chl synthesis, NADPH:protochlorophyllide oxidoreductase converts protochlorophyllide into chlorophyllide, which is ultimately converted to Chl in developing leaves (Kwon et al., 2017). The Lhca and Lhcb (light-harvesting Chl a/b-binding) genes encode the light-harvesting Chl a/b-binding proteins LHCI and LHCII, which capture light energy and transfer it to Chl in the core reaction centers of photosystem I (PSI) and photosystem II (PSII; Kobayashi et al., 2013). GLKs act as transcriptional activators and promote the coordinated expression of nuclear photosynthesis-related genes that are associated with Chl biosynthesis and light harvesting in rice (Nakamura et al., 2009), Arabidopsis (Arabidopsis thaliana L.; Waters et al., 2009), tomato (Solanum lycopersicum L.; Nguyen et al., 2014), and peanut (Arachis hypogaea L.; Liu et al., 2018). For example, both OsGLK1 and OsGLK2 bind to the promoters of Chl biosynthetic OsPORB and OsCAO1 and light-harvesting Lhca and Lhcb genes (Sakuraba et al., 2017) and transactivate their expression. In peanut, AhGLK1 also augments the expression of the Chl biosynthetic gene AhPORA by binding to its promoter (Liu et al., 2018).

GLK genes have been overexpressed in several plants using constitutive or tissue-specific promoters to study their effects on plant growth and grain yield. However, constitutive expression of either ZmG1 or ZmG2 gene driven by the maize ubiquitin promoter (pUbi) in rice has been found to reduce seed yield in the greenhouse (Wang et al., 2017). In contrast, the overexpression of Arabidopsis GLK1 gene (AT2G20570) with its silique promoter (PAt1G56100) enhanced seed weight (11%) in Arabidopsis. Also, transgenic Arabidopsis overexpressing AtGLK1 under the control of its leaf-specific promoter enhanced leaf photosynthesis and seed yield (25%; Zhu et al., 2018). A recent study reported that transforming rice (spp. japonica cv. Kitaake) with either of the two maize GLK genes, especially the maize GLK2 gene, controlled by the maize ubiquitin promoter, increases rice yield in the field by 30%–40% (Li et al., 2020). The GLK transgenic rice plants also showed increased accumulation of antioxidative pigments, enhanced photoprotection, and higher photosynthetic rates. Thus, these results with Arabidopsis and rice suggest genetic manipulation of GLK genes with a proper promoter may enhance chloroplast development, Chl accumulation, leaf photosynthesis, and crop productivity in some crops.

Most genes exhibit cell-, tissue-, or development-specific expression under the control of their own promoters. GLKs’ own promoters, either from C3 or C4 plants, have not been tested for their effects, including the expression pattern of GLK genes, photosynthesis, and growth in transgenic plants. In contrast to Li et al. (2020), in which the strong constitutive maize ubiquitin promoter was used to drive maize GLK gene expression in rice, the major objective of this study was to induce ectopic expression of the two maize GLK genes, individually and in combination, driven by the maize promoters, or under the control of constitutive promoters, for comparing chloroplast development, Chl accumulation, photosynthetic property, growth, and grain yield in rice.

Results

Establishing five transgenic rice plants (O. sativa L.) expressing maize GLK genes (Z. mays L.)

Five GLK transgenic rice lines overexpressing different maize GLK gene constructs (Supplemental Figure S1) were produced using Agrobacterium-mediated method (Yeh et al., 2015). Specific primers (Supplemental Table S1) were designed to detect these transgenes. The self-pollinated T1 plants derived from four to five independent lines for each transformation with pZmUbi::ZmG1 or p35S::ZmG2 were screened for gene insertion copy number by seedlings resistant to hygromycin during germination (Supplemental Table S2); and gene expression was confirmed by reverse transcription quantitative polymerase chain reaction (RT-qPCR). Homozygous T2/T3 plants showing no further segregation in seedling hygromycin resistance were subsequently obtained. The homozygous plants from each transformation all exhibited dark green leaf/floret phenotypes (Supplemental Table S2) and consistent seed weights/plant (Supplemental Figure S2A). The expression levels of ZmG1 in pZmUbi::ZmG1 and ZmG2 in p35S::ZmG2 representative homozygous lines were assayed by RT-qPCR again (Supplemental Figure S2B) and pZmUbi::ZmG1 (#8-1) and p35S::ZmG2 (#34-3) lines having a high level of expression were selected for further characterization.

Similarly, the self-pollinated T1 plants derived from two independent lines for each transformation with the pZmG1::ZmG1, pZmG2::ZmG2, or pZmG1::ZmG1/pZmG2::ZmG2 construct were analyzed for insertion gene copy number by seedlings resistant to hygromycin (Supplemental Table S3) and total grain weight (Supplemental Figure S2C). Furthermore, thermal asymmetric interlaced polymerase chain reaction (TAIL-PCR) was used to determine the transgene insertion site on rice chromosomes, which facilitated the screening of homozygous plants from the T1 plants for each independent line by PCR genotyping (Supplemental Figure S2D). Notably, for the T1 pZmG2::ZmG2 plants only the heterozygous plants developed into maturity while the homozygous plants showed retarded growth and development and died after 1–2 months of cultivation. The homozygous/heterozygous plants from the independent lines for each of the three transformations all exhibited a dark green leaf/floret phenotype (Supplemental Table S3) and had consistent seed weight/plant (Supplemental Figure S2E). Thus, one representative line from each of these three transformations, #8-8, #2-4, and #3-8, was selected for further characterization. Also, genomic PCR analysis demonstrated the presence of either ZmG1 and/or ZmG2 genes in the five representative GLK transgenic plants (Supplemental Figure S2F). Southern blot analysis demonstrated that one copy or two copies of the maize GLK gene(s) were inserted into a transgenic rice line (Supplemental Figure S2G). This result is consistent with the predicted transgene copy numbers for these transgenic lines, as determined by the segregation pattern to hygromycin (Supplemental Tables S2 and S3).

Phenotype characterization of transgenic rice plants

After germination on 1/2 Murashige and Skoog medium for 8 d, seedlings of pZmUbi::ZmG1 and pZmG1::ZmG1 lines exhibited a retarded growth phenotype, compared to the wild-type (WT; Figure  1A). The seedlings of p35S::ZmG2 and heterozygous and homozygous seedlings of pZmG2::ZmG2 also grew slower and were much smaller. As homozygous pZmG2::ZmG2 plants showed retarded growth and died after 2 months of cultivation heterozygous plants were used for subsequent analyses. All plants were cultivated in 5-L pots (one plant/pot) between May and September 2020 in the greenhouse and the mature plants are shown in Figure  1B. TAIL-PCR analysis revealed one insertion location of the T-DNA into the OsFRD-3-like gene in the pZmG2::ZmG2 transgenic rice plants (Supplemental Figure S3A). RT-qPCR analysis found no substantial difference in OsFRD3 transcript levels in all lines (Supplemental Figure S3B). Thus, the dwarf phenotype of pZmG2::ZmG2 transgenic seedlings was likely caused by overexpression of ZmG2 per se. For other transgenic rice lines, the maize GLK genes were inserted into different chromosomes without affecting known functions (Supplemental Table S3).

Figure 1

Phenotypes of the WT and transgenic rice plants ectopically overexpressing maize GLK genes under the control of constitutive promoters (pZmUbi::ZmG1 and p35S::ZmG2) or maize promoters (pZmG1::ZmG1, pZmG2::ZmG2 and pZmG1::ZmG1/pZmG2::ZmG2). A, Eight-day-old seedlings. Seeds from the homozygous pZmUbi::ZmG1, p35S::ZmG2, pZmG1::ZmG1, and pZmG1::ZmG1/pZmG2::ZmG2 plants and from the heterozygous pZmG2::ZmG2 plant were germinated on 1.5% agar containing 1/2 MS. pZmG2::ZmG2 seedlings included both heterozygous (Hetero.) and homozygous (Homo.) plants. Note the seedlings of homozygous pZmG2::ZmG2 transgenic plants died after 2 months of cultivation. B, Four-month-old mature plants. Note the heterozygous pZmG2::ZmG2 plants showed delayed flowering.

Expression profiles of GLK genes in maize, rice, and transgenic rice plants

The expression profiles of GLK genes in different tissues of maize, including young leaves (YLs, from 9-d-old plants), mature leaves (MLs, from 2-month-old plants), stalk (S), tassel (T), and root (R), were analyzed by RT-qPCR. The data suggest that the promoters of maize GLK genes conferred differential tissue-specific expression, with the highest expression level found in young and MLs (Figure  2A). ZmG1 and ZmG2 were expressed at low levels in maize stem and tassel. The ectopic expression of maize GLK genes and its effect on expression of endogenous rice GLK genes were also examined in newly ML, stem (S), floret (F), and root (R) tissues of all GLK transgenic rice lines. Both ZmG1 and ZmG2 were highly expressed in the leaves of all five GLK transgenic rice plants. In contrast, constitutive maize ubiquitin (pUbi) and 35S promoters drove high levels of maize gene expression in all rice tissues studied (Figure  2B). ZmG1 and ZmG2 were also highly expressed in both shoot and root tissues in pUbi::ZmG1 and p35S::ZmG2 transgenic plants, respectively. Notably, the expression of ZmG1 was promoted whereas the expression of ZmG2 was suppressed in all studied tissues of pZmG1::ZmG1/pZmG2::ZmG2 transgenic plants, suggesting coordinated expression of the two genes, an unexpected observation. As expected, both endogenous OsGLK1 and OsGLK2 were highly expressed in WT rice leaves. However, the ectopic expression of maize GLK genes, individually or in combination, suppressed the expression of rice endogenous GLK genes in all studied tissues of all GLK transgenic plants, except for OsGLK1 in the roots of pZmUbi::ZmG1 and pZmG2::ZmG2 transgenic plants (Figure  2C). Immunoblot analysis revealed that, relative to both maize and rice, pZmUbi::ZmG1 plants accumulated a large amount of GLK protein in the leaves (Figure  2D), consistent with its high level of gene expression as determined by RT-qPCR (Figure  2B). The other four transgenic plants also accumulated a higher GLK protein amount than maize and rice.

Figure 2

GLK gene expression and immunoblot of GLK protein in different tissues of maize and WT and five GLK transgenic rice plants. The heterozygous pZmG2::ZmG2 plants and the homozygous plants of other four transgenic plants were used in the study. A, Expression profiles of ZmG1 and ZmG2 in different maize tissues. Total RNA for RT-qPCR was isolated from YL of 9-d-old seedlings, MLs of 2-month-old plants, stems (S), tassels (T), and roots (R). B–C, Expression of ZmG1 and ZmG2 (B) and expression of OsGLK1 and OsGLK2 (C) in different tissues of WT and transgenic rice plants. Total RNA was isolated from MLs, stems (S), florets (F), and roots (R) of WT and five GLK transgenic rice plants. The expression level of 17S rRNA was used as an internal control for normalization. The expression levels of ZmG1 or ZmG2 in young maize leaves (A and B) and the expression levels of OsGLK1 or OsGLK2 in WT leaves (C) were used as the references for estimating their relative expression levels in different tissues. Values in Figure  2, A–C are means ± sd of three replicates. D, Immunoblot of GLK protein in the YLs of 9-d-old maize and WT and GLK transgenic rice plants using antibodies raised against the consensus peptide of both rice and maize GLK proteins. Immunoblot of Rubisco small subunit is included as an internal control for loading. Maize leaf protein was included as a positive control for GLK. M: molecular weight markers.

Shoot and root transcriptomes of transgenic rice plants

To assess the influences of the maize gene promoters on the expression of ZmGLK and endogenous rice genes, the transcriptomes of 4-d-old shoots and roots of homozygous pZmG1::ZmG1, both homozygous and heterozygous pZmG2::ZmG2, and homozygous pZmG1::ZmG1/pZmG2::ZmG2 transgenic plants were analyzed. The data revealed that maize GLK genes were expressed at much higher levels in shoots than in roots (Figure  3A). Also, ZmG1 and ZmG2 individually and in combination reduced endogenous OsGLK1 transcripts in YLs and roots in these three GLK transgenic plants. Compared to the WT, OsPORA, which is involved in Chl biosynthesis (Sakuraba et al., 2017), was upregulated in the shoots of all three plants. In addition, OsGNC, a key regulator of chloroplast biogenesis (Bi et al., 2005), was upregulated in the shoots of pZmG1::ZmG1/pZmG2::ZmG2 transgenic plants (Figure  3B). Photosynthesis-related genes, including PSII: OsLHCB2.4 (Chl a/b binding protein), OsLHCB3 (Chl a/b binding protein), and OsLHCB6 (Chl a/b binding protein) and PSI: OsPSAE-2 (PSI subunit PsaE-2), OsPSAF (PSI subunit PsaF), OsLHCA2 (Chl a/b binding protein), and OsLHCA4 (LHC I type IV Chl binding protein) were also substantially upregulated in the roots of pZmG1::ZmG1/pZmG2::ZmG2 plants (Figure  3C).

Figure 3

Transcriptome analysis of DEGs in 4-d-old shoots and roots of the WT, pZmG1::ZmG1 (homozygous), pZmG2::ZmG2 (homozygous and heterozygous), and pZmG1::ZmG1/pZmG2::ZmG2 (homozygous) and transgenic seedlings. A–C, Expression of ZmG1, ZmG2, OsGLK1, and OsGLK2 (A) and expression of OsPORA, OsGNC, OsLHCB2.4, OsLHCB3, OsLHCB6, OsPSAE-2, OsPSAF, OsLHCA2, and OsLHCA4 (B and C) in the WT, and pZmG1::ZmG1, pZmG2::ZmG2, and pZmG1::ZmG1/pZmG2::ZmG2 transgenic seedlings. RPKM value was used to define expression level. Shoot and root tissues were pooled from 10 seedlings of 4-d-old WT and transgenic plants (n = 10) for transcriptome analysis.

Functional classification of differentially expressed genes

Differentially expressed genes (DEGs) in pZmG1::ZmG1, pZmG2::ZmG2, and pZmG1::ZmG1/pZmG2::ZmG2 transgenic plants were identified in both shoot and root transcriptomes, relative to those of the WT. In shoot, the numbers of DEGs in these three transgenic plants were 723, 1,693, and 922, respectively (Supplemental Figure S4A). In root, the corresponding numbers were 628, 2,706, and 2,703 (Supplemental Figure S4C). Clearly, ZmG2 tends to transactivate more DEGs. Venn diagrams show the numbers of uniquely and commonly upregulated and downregulated DEGs in shoots (Supplemental Figure S4A) and roots (Supplemental Figure S4C) of these three transgenic plants. Gene ontology (GO) term (biological process) enrichment analysis showed differently upregulated and downregulated DEGs in pZmG1::ZmG1, pZmG2::ZmG2, and pZmG1::ZmG1/pZmG2::ZmG2 shoots (Supplemental Figure S4B) and roots (Supplemental Figure S4D). Differentially expressed TF genes in shoots and roots are shown in Supplemental Figure S5, A and B. Maize GLK genes apparently also modulate several rice TFs, including the major TF families of AP2/ERF (e.g. DREB, ERF, CBF), basic helix–loop–helix), heat shock factors (HSF), myeloblastosis (MYB), NAC (NAM, ATAF and CUC), no apical meristem, WRKY (W-box WRKYGQK-containing factors), basic leucine zipper (bZIP; Supplemental Table S4), which are involved in plant stress and defense responses (Ng et al., 2018). Therefore, the common biological functions enriched in biotic and abiotic stresses and defense responses in rice are promoted by the maize GLK genes.

GO analysis of DEGs

For upregulated DEGs, pZmG2::ZmG2 shoots were substantially enriched in GO terms of responses to biotic/abiotic stimuli and chitin catabolic process (Supplemental Figure S6A), whereas pZmG1::ZmG1/pZmG2::ZmG2 shoots were mainly enriched in lipid transport GO terms (Supplemental Figure S6B). The DEGs of pZmG1::ZmG1 roots were significantly enriched in transcription regulation-related GO terms (Supplemental Figure S6C), whereas those of pZmG2::ZmG2 roots were enriched in disease defense response GO terms, including oxylipin biosynthetic process (methyl jasmonate biosynthesis pathway), chitin catabolic process, and phytoalexin metabolic process (Supplemental Figure S6D). This suggests that both maize genes promote chloroplast biogenesis and development in rice but ZmG1 acts more as a transcriptional regulator, whereas ZmG2 acts more as a defense regulator. Root transcriptome analysis of pZmG1::ZmG1/pZmG2::ZmG2 plants shows enrichment in transcription regulation-related GO terms (also found in pZmG1::ZmG1 roots), chitin catabolic/phytoalexin biosynthetic process (also found in pZmG2::ZmG2 roots), and activation of photosynthesis-related pathways, gibberellin metabolic process, isoprenoid biosynthetic process, xylan catabolic process, and protein ubiquitylation (Supplemental Figure S6E). Taken together, these data suggest that the two maize GLK genes transcriptionally promote rice photosynthesis, metabolism, and defense response genes in a complementary manner.

For downregulated DEGs, pZmG2::ZmG2 roots were substantially enriched in ion transport, amine biosynthetic process, sulfur compound biosynthetic process, and aspartate family amino acid biosynthetic process, whereas pZmG1::ZmG1/pZmG2::ZmG2 roots were mainly enriched in amine biosynthesis and ion transport. This implies that ZmG1 may partially offset the repressive functions by ZmG2 when they are co-expressed in rice (Supplemental Figure S6, F and G). Highly downregulated genes and thus their functions in the homozygous pZmG2::ZmG2 plants may have contributed to their death after germination. This highlights the importance of expressing transgenes at proper levels in transgenic plants.

Chloroplast development, Chl content, Rubisco activity, and photosynthesis

Transmission electron microscopy analysis of newly mature flag leaves revealed that the expression of maize GLK genes substantially induce chloroplast development in both M and BS cells in all five transgenic plants (Figure  4, A and B). The images also reveal that all of these transgenic plants have larger chloroplasts. Wang et al. (2017) also observed increased chloroplast development in BS and mestome sheath cells in transgenic rice leaves overexpressing maize GLK genes driven by the maize ubiquitin promoter. Similar to the WT, the M and BS chloroplast structures in all GLK transgenic plants had normal grana structures and thylakoid membranes (Figure  4, C and D), while BS chloroplast numbers increased in all GLK transgenic plants (Figure  4E). Also, all five GLK transgenic rice lines showed a significant enhancement in stacked thylakoid development per chloroplast area (Figure  4F). The thylakoid membranes of vascular plants are differentiated into stacked granum and unstacked stromal regions. Since PSII (photosystem II) and LHCII (Light harvesting complex II) reside mainly in the stacked thylakoid membranes and PSI and the ATPase are predominantly located in the stromal thylakoids (Andersson and Anderson, 1980), this represents a functional enhancement in photochemical activities in the GLK transgenic rice plants. Chl contents in flag leaves and florets of all five GLK transgenic plants increased respectively by 18%–39% and 10%–127% relative to the WT (Figure  5, A and B) and, as a consequence, Rubisco activity and protein content increased by 27%–78% and 21%–28%, respectively (Figure  5C). The flag leaf photosynthetic rates, stomatal conductance, mesophyll conductance, and light use efficiency (LUE) of all five GLK transgenic lines increased by 5%–11%, 10%–37%, 6%–14%, and 4%-16%, respectively, with the highest increases being observed invariably in pZmG1::ZmG1/pZmG2::ZmG2 transgenic plants (Figure  6, A–D). Thus, all photosynthetic functions in the five GLK transgenic lines are significantly enhanced, especially in the transgenic plant that simultaneously overexpresses both maize GLK genes with their own promoters. Light-saturated photosynthetic rates (Figure  6A) are highly correlated with stomatal conductance (R2 = 0.70, Figure  6B) and mesophyll conductance (R2 = 0.51, Figure  6C) among WT and the five GLK transgenic lines, suggesting that stronger chloroplast development and function in guard cells may promote stomatal conductance to CO2 diffusion and increased Rubisco activity facilitates mesophyll conductance. Moreover, light use efficiencies (Figure  6D) are also highly correlated (R2 = 0.80) with Chl contents (Figure  5A), consistent with the notion that higher Chl contents facilitate plant light harvesting, especially under low light conditions. Consistently, light-saturated photosynthetic rates (Figure  6A) are also correlated (R2 = 0.34) with Chl contents (Figure  5A). Taken together, these data show the benefits of promoting chloroplast development on leaf photosynthetic physiology.

Figure 4

Transmission electron micrographs of flag leaf cross-sections of WT and five GLK transgenic rice plants. A, B, Chloroplasts in M (A) and BS (B) cells of the WT and five GLK transgenic rice plants, with arrows pointing to representative chloroplasts. C and D, Enlarged micrographs of chloroplasts in M (C) and BS (D) cells of the WT and five GLK transgenic rice plants. Bars = 0.5 μm. P, plastoglobuli; CW, cell wall. E, Averaged chloroplast numbers in each BS cell. The number of chloroplasts in each BS cell was determined from 15 BS cells in cross sections. F, Averaged thylakoid area per M chloroplast area. Thylakoid area per chloroplast area was determined from 6 to 8 M cells. Statistical tests were based on the Wilcoxon–Mann–Whitney U test for nonparametric comparison of two groups. The asterisk marked on a box indicates statistical significance in comparison with the WT (*P < 0.05; **P < 0.01). Boxplots showing the median (horizontal line) and interquartile range (boxes). Whiskers extend from the ends of the box to the smallest and largest data values.

Figure 5

Leaf Chl contents, Rubisco contents and activities and total soluble protein contents in the WT and five GLK transgenic rice plants. A–C, Chl contents in the flag leaves (A) and florets (B) and Rubisco contents (analyzed by immunoblot) and activities (assayed by enzymatic reaction) and total soluble protein contents (determined by protein assay) in the flag leaves of 4-month old plants (C). Statistical tests were based on the pairwise t test. Asterisks indicate statistical significance levels in comparison with the WT (*P < 0.05; **P < 0.01; ***P < 0.001). Data = means ± sd, n = 4 (each sample was obtained from leaf sections of three different plants).

Figure 6

Light-saturated photosynthetic rates, stomatal conductance (gs), mesophyll conductance (gm), and maximum LUE in the newly mature flag leaves of 4-month-old WT and five GLK transgenic rice plants. Light-saturated photosynthetic rates, gs and gm were measured at 2,000 μmol·m−2·s−1 PPFD, 30°C, and 400 μL·L−1 CO2 and maximum LUE was measured from the initial linear slope of the light response curve. Plants were grown in the greenhouse between May and September. Statistical tests were based on the Wilcoxon–Mann–Whitney U test for nonparametric comparison of two groups. The asterisk marked on a box indicates statistical significance in comparison with the WT (A–D: *P < 0.05; **P < 0.01, n = 5). Boxplots showing the median (horizontal line) and interquartile range (boxes). Whiskers extend from the ends of the box to the smallest and largest data values.

Growth and productivity traits

Compared to the WT (100%), total shoot biomass (109%–180%), total panicle number (137%–224%), and total grain weight per plant (102%–170%) all increased in each of the five GLK transgenic plants, mainly due to the production of extra tillers (Figure  7, A–C). Notably, pZmUbi::ZmG1 plants produced more small tillers and showed much delayed leaf senescence. Except for the heterozygous pZmG2::ZmG2 transgenic plants, the other plants produced smaller seeds and thus lower 1,000-grain weights (Figure  7D). Most significantly, pZmG1::ZmG1/pZmG2::ZmG2 transgenic plants expressing both maize GLK genes under the control of their own promoters showed the highest increases in growth (80%) and grain yield (70%), with a slightly lower 1,000-grain weight (95%) than that of the WT. The biomass and grain yield of pZmG1::ZmG1 and pZmG2::ZmG2 plants also increased by 35%–39% and 47%–51%, respectively. In contrast, pUbi::ZmG1 transgenic plants with the strong ubiquitin promoter showed no significant increase in either photosynthetic efficiency, growth or grain yield, although they accumulated more Chl in both leaves and florets than pZmG1::ZmG1/pZmG2::ZmG2 transgenic plants. Also, the seeds of pUbi::ZmG1 transgenic plants showed a significantly lower germination rate (85%) than the WT (Figure  7E), which may be related to its smaller seeds (Figure  7D). Thus, in terms of photosynthetic efficiency, growth, and grain yield, transforming both maize GLK genes is better than transforming with one gene alone and using GLK’s own promoters is better than using either the ubiquitin or the 35S promoter. Moreover, high level expression of maize GLK genes with the strong constitutive ubiquitin promoter may lead to negative impacts on seed development.

Figure 7

Growth and yield traits of the WT and five GLK transgenic rice plants. A, Total aboveground biomass/plant. B, Total panicle number/plant. C, Total grain weight/plant. D, 1,000-grain weight/plant. E, Rate of seed germination in water. Plants were grown in the greenhouse between May and September. Statistical tests were based on the Wilcoxon–Mann–Whitney U test for nonparametric comparison of two groups. The asterisk marked on a box indicates statistical significance in comparison with the WT (A–D: *P < 0.05; **P < 0.01; ***P < 0.001, n = 9; E: *P < 0.05; **P < 0.01; ***P < 0.001, n = 4 with each containing 200 seeds). Boxplots showing the median (horizontal line) and interquartile range (boxes). Whiskers extend from the ends of the box to the smallest and largest data values.

Discussion

More than 90% of crop biomass is generated by photosynthesis, which occurs in chloroplasts. Various GLK genes have been overexpressed in Arabidopsis (A. thaliana L.; Waters et al., 2008), rice (O. sativa L.; Nakamura et al., 2009; Wang et al., 2017; Li et al., 2020), and tomato (S. lycopersicum L.; Powell et al., 2012) to evaluate their potential to boost chloroplast development, Chl biosynthesis, photosynthesis and growth (Waters et al., 2008; Nakamura et al., 2009; Powell et al., 2012; Wang et al., 2017; Li et al., 2020) and nonleaf organs (e.g. calli and fruit; Nakamura et al., 2009; Nguyen et al., 2014). However, most of these studies employed constitutive promoters, such as the strong maize ubiquitin or 35S promoters, to drive GLK gene expression in host plants. As anticipated, the effects of overexpressing GLK genes on growth and yield of the transgenic plants are variable, depending on the promoters used because expression of most genes are cell-, tissue-, and development-dependent. The overexpression of transgenes under the control of constitutive promoters, such as the ubiquitin gene promoter, at high levels may lead to negative impacts on plant growth and development (e.g. Wang et al., 2017). Li et al (2020) showed that transgenic rice plants overexpressing either ZmG1 or ZmG2 under the control of maize ubiquitin promoter exhibit a higher photoprotection, due to increased accumulation of the antioxidant pigments xanthophyll and lutein and synthesis of D1 protein to form a repair cycle that restores PSII structure. Specifically, restoration of degraded D1 protein will maintain high rates of electron transport and photosynthesis under high light conditions (Gururani et al., 2015). In this study we used ZmG1 and ZmG2 promoters, besides constitutive promoters, to express ZmG1 and ZmG2 individually or simultaneously in rice. We found that the promoters significantly influence the expression levels of tissue-specific ZmGLK genes and subsequently their vegetative and reproductive growth in rice. The increases in expression of ZmGLKs in turn affected phenotypes associated with chloroplast development (Figure  4) and Chl accumulation in leaves and florets in all of the five GLK transgenic rice plants studied (Figure  5, A and B). Most obviously, we found that high expression levels achieved by the strong constitutive promoters, e.g. maize ubiquitin promoter (Figure  2B), did not translate into higher growth and grain yield in rice (Figure  7, A–D), presumably due to the negative impacts of high level expression of these maize GLK genes (Figure  2B) on rice growth and seed development.

GLK genes regulate chloroplast development and co-regulate the expression of nuclear photosynthetic and Chl biosynthesis in diverse plant species (Nakamura et al., 2009; Waters et al., 2009; Nguyen et al., 2014). The overexpression of OsGLK1 in rice calli promotes nuclear-encoded photosynthetic PSI and PSII genes in response to light, accompanied by a high photosynthetic activity in rice calli (Nakamura et al., 2009). Consistently, expression of photosynthesis-related genes was also induced in roots of pZmG1::ZmG1/pZmG2::ZmG2 transgenic seedling (Figure  3C; Supplemental Figure S4D). In addition, overexpression of either or both of ZmG1 and ZmG2 increased Chl biosynthesis and OsPORA expression level in rice shoots (Figure  3B). More importantly, OsGNC, a major regulator of chloroplast development (Bi et al., 2005), was only upregulated in pZmG1::ZmG1/pZmG2::ZmG2 transgenic shoots (Figure  3B). Since photosynthesis-related genes and Chl biosynthesis genes reside in the nucleus, these genes might be coregulated for efficient photosynthetic development (Waters et al., 2009). Notably, elevated expressions of OsPORA, OsGNC and other photosynthesis-related genes (e.g. nuclear-encoded RbcS and chloroplast-encoded RbcL) in the pZmG1::ZmG1/pZmG2::ZmG2 transgenic rice (Figure  3) led to higher photosynthetic rates (Figure  6A), growth, and grain yields than other GLK transgenic rice plants (Figure  7, A–C).

In previous studies on chloroplast development, Wang et al. (2017) showed ZmUbipro::ZmG2 transgenic rice exhibits elevated Chl content and Rubisco activity in leaves. Nakamura et al., (2009) reported OsGLK1-overexpressing rice exhibits normal stacked thylakoid grana in the chloroplast of leaf sheath and enhances chloroplast development in vascular BS cells of leaves, as well as the occurrence of Chl in roots. Overexpressing GLK in Arabidopsis also induces chloroplast biogenesis in roots (Kobayashi et al., 2013), although potential function in photochemistry in roots is not known. Consistently, in the current study, the overexpression of either one or both ZmG1 and ZmG2 genes with constitutive or maize-own promoters in rice displayed high transcript levels in leaves, florets and roots (Figure  2B), leading to increased leaf and floret chloroplast development (e.g. increased stacked grana, particularly in GLK overexpression lines with their own promoters) in both leaves and roots (Figure  4), higher Chl content, protein content, Rubisco activities (Figure  5), photosynthetic activities (Figure  6), and growth and grain yield (Figure  7). The beneficial effects conferred by overexpressing the maize GLK genes are most significant when both genes are coexpressed in a coordinated manner under the control of their own promoters. In contrast, high-level expression of ZmG1 by the strong constitutive ubiquitin promoter reduces seed size and inhibit seed germination (Figure  7, D and E).

Study with Arabidopsis suggests that increased Chl accumulation in leaves maximizes light absorption, photosynthesis and plant productivity (Gotoh et al., 2018). In rice, a higher seed yield was also related to higher leaf photosynthesis (Oiestad et al., 2019). In this study, enhanced leaf photosynthetic capacity (Figure  6A) and LUE (Figure  6D) were observed in the five GLK transgenic plants by increases in chloroplast development (Figure  4), Chl content (Figure  5, A and B), and Rubsico activity (Figure  5C) in the flag leaves, leading to increased carbohydrates in shoot biomass and panicle number, a key component of grain yield (Khush, 2013). Most significantly, pZmG1::ZmG1/pZmG2::ZmG2 transgenic rice plants simultaneously overexpressing the two maize GLK genes driven by their own promoters produced more biomass (80%) and grain yield (70%) than the WT, mainly due to the development of more tillers (Figure  7, A–C). With own promoters, the biomass and grain yield of pZmG1::ZmG1 and pZmG2::ZmG2 plants also increased by 35%–39% and 47%–51%, respectively.

Besides chloroplast development, GLK genes also play important roles in abiotic and biotic tolerance. AtGLKs confer enhanced biotic resistance to Cucumber mosaic virus (Han et al., 2016). Moreover, Arabidopsis GLK1 and GLK2 control the expression of multiple ABA-responsive genes (Ahmad et al., 2019). ABA in turn acts as a signaling mediator for regulating the adaptive response of plants to drought stress (Long et al., 2019). Consistently, our transcriptome data suggest that the common biological functions enriched in transgenic rice overexpressing maize GLKs are in biotic and abiotic stresses and defense responses (Supplemental Figures S4 and S6). Overexpressing the maize GLK genes in rice also promoted the expression of other stress-response rice TFs families including WRKY, MYB, HSF, bZIP, NAC, and AP (Supplemental Table S4), which are essential for plant stress response by binding to specific cis-acting elements of functional genes. Overexpression of these TF genes usually increases the adaptability of plants to drought and salt stresses (Ma et al., 2019) and confers enhanced disease resistance (Jimmy and Babu, 2019). Mitogen-activated protein kinase cascade activating chitin-responsive WRKY53 has been shown to regulate plant defense responses (Wan et al., 2008). Similarly, upregulated OsWRKY53 may subsequently induce defense- and phytoalexin biosynthesis-related genes to produce phytoalexin in roots of pZmG2::ZmG2 and pZmG1::ZmG1/pZmG2::ZmG2 transgenic rice for enhanced fungus defense (Supplemental Figure S6, D and E). Oxylipins, which are critical regulators in defense responses to pathogens (Genva et al., 2019), are enriched in roots of pZmG2::ZmG2 transgenic rice. The MYB family is also responsible for primary and secondary metabolisms, such as phenylpropanoids biosynthesis, and acts as a key regulator of secondary cell wall formation including xylan (Liu et al., 2015). Similarly, upregulated MYB TFs in pZmG1::ZmG1/pZmG2::ZmG2 transgenic roots may lead to enhanced phenylalanine and xylan metabolism, cell wall development and disease resistance. Overall, transgenic rice expressing maize GLK genes provides insights for improving stress tolerance and disease resistance to mitigate the negative impacts of biotic and abiotic stresses on rice yields. Further studies are needed to evaluate the performance of these GLK transgenic rice plants with constitutive versus maize promoters in the field, including performance under stress conditions. As both Li et al. (2020) and this study used japonica rice to demonstrate the benefits of overexpressing maize GLK genes on rice photosynthesis, growth and productivity, ultimate studies on indica rice will be important since it represents about 80% of world rice production (Wei et al., 2013).

Taken together, our results clearly demonstrate that simultaneous overexpression of both maize GLK genes under the control of their own promoters in rice confers the most significant beneficial effects on the development of chloroplast and the expression of photosynthesis-related and stress-response genes in a complementary manner. This in turn boosts rice photosynthesis and increases growth and grain yield. The same approach can be extended to improve the yields of other crops, with the potential of creating a second “Green Revolution”.

Materials and methods

Promoter and gene cloning

Genomic DNA was extracted from young maize (Z. mays L.) leaves (White Crystal, a glutinous maize cultivar) using the cetyltrimethyl ammonium bromide (CTAB) method. PCR was used to clone the maize GLK1 (ZmG1) and G2 (ZmG2) promoters and 5′-UTR regions based on the sequences of ZmG1 (GRMZM2G026833) and ZmG2 (GRMZM2G087804) in the Ensembl plants database (http://plants.ensembl.org/index.html). Total RNA was extracted from the second leaves of 9-d-old maize seedlings showing peak expression of both GLK genes and used for cDNAs synthesis (Liu et al., 2013). ZmG1 (GRMZM2G026833) and ZmG2 (GRMZM2G087804) cDNAs were cloned by PCR using gene specific primer pairs (Supplemental Table S1). The sequences of the promoters and full-length cDNAs of ZmG1 and ZmG2 were confirmed by Sanger sequencing.

Construction of transformation vectors

Five constructs were made to express ZmG1 and ZmG2 genes in rice (O. sativa L.) under the control of maize’ own or constitutive promoters. The maize promoters and full-length cDNAs of the two genes were cloned by PCR and subcloned into the transformation vector pCAMBIA or Geteway (Invitrogen). The PCR fragments with compatible restriction sites were generated using specific primers (Supplemental Table S1). The ZmG1 promoter and cDNA fragments were treated with XbaI/BamHI and BglII for both 5′-/3′-ends, respectively, and the ZmG2 promoter and cDNA fragments were treated with HindIII/BamHI and BglII for both 5′- and 3′-ends, respectively. The ZmG1 promoter (2,134 bp) fused to its cDNA (1,428 bp) and ZmG2 promoter (1,942 bp) fused to its cDNA (1,386 bp) were separately ligated into an intermediate vector that comprised a NOS terminator to obtain ZmG1p::ZmG1 and ZmG2p::ZmG2 fragments. All fragments in the intermediate vector were confirmed by restriction enzyme digestion analysis. Afterwards, the ZmG1p::ZmG1 fragment cut with EcoRI and the ZmG2p::ZmG2 fragment cut with HindIII/EcoRI from the intermediate vector were subsequently ligated into the EcoRI and HindIII/EcoRI sites of pCAMBIA 1,300 to generate the pZmG1::ZmG1 and pZmG2::ZmG2 constructs, respectively. In addition, the ZmG1p::ZmG1 fragment cut with EcoRI was cloned into the same restriction site of the ZmG2p::ZmG2 vector to generate the pZmG1::ZmG1/pZmG2::ZmG2 construct. To express maize GLK1 and GLK2 in rice under the control of a constitutive promoter, the Gateway entry clone TOPO PCR8 containing ZmG1 or ZmG2 cDNAs was, respectively, transferred into the Gateway donor vector pCAMBIA1302 under the control of the maize ubiquitin or pH2GW7 under the control of 35S promoter to generate pZmUbi::ZmG1 or p35S::ZmG2 construct. The pZmUbi::ZmG1, p35S::ZmG2, pZmG1::ZmG1, pZmG2::ZmG2, and pZmG1::ZmG1/pZmG2::ZmG2 constructs containing the hptII gene (HygR; Supplemental Figure S1) for selection were separately transfected into the Agrobacterium tumefaciens EHA105 strain via electroporation.

Rice transformation

Rice (O. sativa L. sp. japonica cv. TNG67) callus induction, co-cultivation with Agrobacterium, hygromycin selection of transformed callus and plantlet regeneration followed the procedures of Yeh et al. (2015). All positive transgenic seedlings were transplanted into soil and cultivated in the greenhouse for molecular, physiological and anatomical analyses.

Determination of transgene insertion location by TAIL-PCR

The conditions for TAIL-PCR were as described previously (Liu et al., 2005; Møller et al., 2009). Genomic DNAs of T1 transgenic plants were used as templates for three successive runs of PCR using a short arbitrary degenerate primer and three specific nested primers (SP0, SP1, and SP2; Supplemental Table S1) to amplify T-DNA flanking genomic DNA regions. Tail-PCR products were purified and sequenced. Comparison of the PCR product sequences to the O. sativa Japonica Group DNA sequence in Ensemble Plants (https://plants.ensembl.org/Multi/Tools/Blast?db=core) by BLAST pinpoints the site of transgene insertion on a rice chromosome. The sequences identified were then used to evaluate the zygosity of transgenic plants.

Screening of homologous transgenic plants by genotyping PCR

Genotyping primers were designed from the 3′- and 5′-end regions flanking the transgene insertion site as determined by TAIL-PCR. A pair of genome-specific primers (GT-F, GT-R) was designed based on the chromosome DNA sequence across the putative insertion site. In addition, transgene-specific SP2 and genome-specific primer GT-R were used to screen transgene insertion in transgenic plants. DNA was extracted from leaf tissues of T1 transgenic plants using Phire Plant Direct PCR Master Mix (Thermo Scientific) following the manual. PCR amplification was conducted with two primer sets (genome-specific GT-F and GT-R primers, and transgene-specific SP2 and genomic GT-R primers; Supplemental Table S1) in a PCR reaction. WT plant is expected to give rise to only one PCR product from the genome-specific primer pair due to the absence of transgenes. Heterozygous transgenic plants gave rise to two PCR products from amplifying the nondisrupted strand of genome sequence by the genome-specific primer pair and from the presence of one strand of transgene by SP2 and GT-R primers. In contrast, as both DNA strands were disrupted by the transgene insertion at a specific site, only one PCR product from the SP2 and GT-R primers was expected from homozygous transgenic plants. After screening, genomic DNA was isolated from YLs of positive homozygous plants for further confirmation of maize GLK genes by PCR.

DNA extraction and PCR

To detect maize GLK genes in transgenic rice plants, specific primer pairs were used in PCR analysis (Supplemental Table S1). Genomic DNA was isolated from YLs of WT and representative transgenic rice plants by the CTAB method. PCR amplification of maize GLK genes was performed in a thermal cycler under the conditions: 94°C/5 min, 30 cycles of 94°C/30 s, 58°C/30 s, and 72°C/30 s, and a final extension at 72°C/5 min.

Southern blot analysis

Preparation of genomic DNA and Southern gel blot analysis were as described previously (Yeh et al., 2015). Genomic DNA isolated from selected transgenic plants that had been confirmed for the presence of maize GLK genes by genomic PCR was digested with HindIII at 37°C for 16–18 h, electrophoresed on 1% (w/v) agarose gel, and transferred to a nylon membrane for probe hybridization. The random primed labeling probe was labeled with digoxigenin-dUTP as described in the DIG High Prime DNA Labeling and Detection Starter Kit II (Roche).

RNA extraction and RT-qPCR

Total RNA was isolated from leaves of the WT and transgenic plants using TRIZOL reagent (Invitrogen) and purified by acid phenol–chloroform extraction. cDNA synthesis and RT-qPCR were performed as described previously (Yeh et al., 2015). Gene-specific primers for amplification of target genes and 17S gene (accession number X00755) as an internal control for normalized expression values were listed in Supplemental Table S1.

Transcriptome analysis

Total RNA was isolated from the shoots and roots of 4-d-old seedlings of WT, homozygous pZmG1::ZmG1, homozygous and heterozygous pZmG2::ZmG2 and homozygous pZmG1::ZmG1/pZmG2::ZmG2 transgenic plants using TRIZOL reagent (Invitrogen; Liu et al., 2013). Sample preparation and transcriptome analysis followed the methods of Liu et al (2013). The sequencing reads were processed and mapped to the rice genome (IRGSP-1.0) using Tophat (version 2.0.10). Each read was aligned, allowing at most 10 hits. The expression level (RPKM, reads per kilobase exon per million reads mapped) of each gene was estimated, using Cufflinks (version 2.1.1). Genes with RPKM ≥1 in at least one sample were considered “expressed” and selected for further analysis. To compare the expression levels of the selected genes across the shoot and root of WT, homozygous pZmG1::ZmG1, homozygous and heterozygous pZmG2::ZmG2, and homozygous pZmG1::ZmG1/pZmG2::ZmG2 transgenic plants, the upper quartile normalization procedure was adopted (Bullard et al., 2010). The nonparametric method of Tarazona et al (2011) was employed to identify DEGs between two samples and the q value (differentially expression probability) in the method was set to be 0.7 (Bullard et al., 2010). The functional enrichment analysis was conducted with the background set of all expressed genes in this study. Fisher’s exact test with false discovery rate (FDR) <0.05 was applied with functional annotations from MapMan (https://mapman.gabipd.org). For the functional classification of DEGs, the GO analysis was carried out by AgriGO V2 software with Fisher’s exact test with FDR ≤0.05 to get the GO annotations based on biological process (28472432).

To classify the biological functions of DEGs, GO analysis was carried out using the AgriGO V2 software based on biological processes. The direct acyclic graph (DAG) drawer is a visualization tool to illustrate the significant GO terms. The DAG, based on the nature of the GO structure, indicates the inter-relationships between terms. To investigate the effect of maize GLK genes in each transgenic rice, the up- and downregulated DEGs from root and shoot were analyzed using the FDR method ≤0.05.

Protein purification and immunoblot analysis

Total soluble protein was extracted and purified by fractionation in water-saturated phenol from young seedling leaves for GLK detection and newly mature flag leaves for Rubisco detection by immunoblot analyses (Dai et al., 1994; Ku et al., 1999). Antibodies against maize Rubisco small subunit were prepared, as described by Ku et al. (1999). In addition, a partial 681-bp sequence of OsGLK1, consensus to the maize GLK genes (Supplemental Figure S7), was cloned into a pET21b vector to produce recombinant protein for generation of anti-GLK antibodies (Yao-Hong Biotechnology Inc., Taiwan). The anti-GLK antibodies were further purified with the maize GLK1 recombinant protein by affinity chromatography.

Chl measurement and Rubisco assay

Leaf segments from flag leaves or florets were collected and extracted in 96% ethanol. The Chl content was calculated from the absorbances at 665 and 649 nm (Wintermans and de Mots, 1965) and expressed on a leaf area (leaf) or fresh weight (floret) basis. Rubisco activities, based on Chl content or leaf area, were assayed as described previously (Pyankov et al., 2000).

Photosynthesis measurement

The mid-sections of newly mature flag leaves on the major tillers of 4-month-old rice plants grown in the greenhouse during flowering in the summer were used for CO2 and H2O exchange measurements using a CIRAS2 portable photosynthesis system (PP system, Amesbury, MA). Measurements were conducted from 7:00 a.m. to 12:30 p.m. on sunny days and the conditions were 2,000 μmol m−2 s−1 photon flux density, 30°C leaf temperature, 70% relative humidity, and 415 ppm CO2 (Ku et al., 1999). The steady-state photosynthetic rate (Pn), stomatal conductance (gs), mesophyll conductance (gm), intercellular CO2 concentration (Ci), and transpiration rate (E) were recorded.

Transmission electron microscopy

Newly mature flag leaf samples (1 mm × 2 mm) were fixed in 1% glutaraldehyde in 0.1-M phosphate-citrate buffer, pH 7.2 at 4°C overnight. After three 20-min buffer rinses, the samples were postfixed in 1% OsO4 in the same buffer for 2 h at room temperature and rinsed in three 20-min changes of buffer. Samples were dehydrated in an acetone series, embedded in Spurr’s resin, and sectioned with a Lecia Reichert Ultracut S or Lecia EM UC7 ultramicrotome. The ultra-thin sections (70–90 nm) were stained with 5% uranyl acetate/50% methanol and 0.4% lead citrate/0.1-N NaOH. A FEI G2 Tecnai Spirit Twin transmission electron microscope at 80 kV was used for viewing and the images were recorded using a Gatan Orius CCD camera.

Agronomic traits studies

WT and five GLK transgenic rice plants were cultivated in 5-L pots (one plant/pot) in the greenhouse between May and September under natural sunlight conditions. Plants were watered daily and fertilized weekly. Upon maturation, total aboveground biomass, total panicle number, 1,000-grain weight, and total grain weight per plant were analyzed after drying at 43°C/2 d for seeds and 60°C/2 d for shoots.

Accession numbers

Sequence data of ZmG1 (GRMZM2G026833), ZmG2 (GRMZM2G087804), OsGLK1 (LOC4340977), and OsGLK2 (LOC4326363) can be found in the GenBank/EMBL data libraries under accession numbers.

Supplemental data

The following materials are available in the online version of this article.

Specific primers used in this study.

Screening of homozygous T2 and T3 plants from pZmUbi::ZmG1 and p35S::ZmG2 transgenic rice lines, based on the segregation pattern of hygromycin resistance, gene expression level analysis by RT-qPCR, and pigment and seed size phenotypes.

Screening of homozygous T1 plants of pZmG1::ZmG1, pZmG2::ZmG2, and pZmG1::ZmG1/pZmG2::ZmG2 transgenic rice lines, based on the segregation pattern of hygromycin resistance, transgene insertion site determination using TAIL-PCR and genotyping PCR analyses.

Differentially expressed TF genes in shoots and roots of 4-d-old pZmG1::ZmG1, pZmG2::ZmG2, and pZmG1::ZmG1/pZmG2::ZmG2 transgenic rice seedlings.

Constructs used for Agrobacterium-mediated rice transformation.

Screening of homozygous plants from five GLK transgenic rice lines by various molecular and physiological analyses for further study.

The insertion site of T-DNA flanking ZmG2 gene in the intron 6 of OsFRD3-like gene on chromosome 2 and expression of OsFRD3 in leaves of the WT, pZmG1::ZmG1, pZmG2::ZmG2, and pZmG1::ZmG1/pZmG2::ZmG2 transgenic rice plants.

DEGs in shoots and roots of 4-d-old pZmG1::ZmG1, pZmG2::ZmG2, and pZmG1::ZmG1/pZmG2::ZmG2 transgenic rice seedlings.

Differentially expressed TF genes in shoots and roots of 4-d-old pZmG1::ZmG1, pZmG2::ZmG2, and pZmG1::ZmG1/pZmG2::ZmG2 transgenic rice seedlings.

Biological process-related GO terms in shoots and roots of 4-d-old pZmG1::ZmG1, pZmG2::ZmG2, and pZmG1::ZmG1/pZmG2::ZmG2 transgenic rice seedlings.

Protein sequences of rice and maize GLKs.

Funding

This work was supported by Academia Sinica, Taiwan Grants (AS-106-TP-L14, AS-106-TP-L14-1, and AS-106-TP-L14-3).

Conflict of interest statement. The authors declare no conflict of interest.

Click here for additional data file.