1 Sequence and Phylogenic Analysis of OsMed16
Owing to its high homology to AtSRF6 (AtMed16), the rice gene LOC_Os10g35560 was previously named OsSRF6 (Wathugala et al. 2011). However, as a subunit of the Mediator complex, LOC_Os10g35560 should be named OsMed16 according to the common unified nomenclature for Mediator subunits (Bourbon et al. 2004). In Wathugala’s studies, OsSFR6 (OsMed16) was predicted to encode a protein of 1170 amino acids. When searching in GenBank (National Center for Biotechnology Information, NCBI) and Rice Genome Annotation Project database, we found the ORF of OsMed16 was 3906 bp in length, and thus encoded a putative protein with 1301 amino acid residues, which is 131-aa longer than OsSRF6 reported by Wathugala et al. (2011). To test this, the full-length ORF of OsMed16 (3906 bp) was amplified from the model japonica rice variety Nipponbare by high-fidelity PCR, and verified by sequencing. Subsequently, the gene structure of OsMed16 was analyzed, which contains 16 exons and 15 introns (Fig. 1a).
To understand the evolutionary relationship of OsMed16, its counterparts were obtained from different plant species, including algae, mosses, ferns, gymnosperms and angiosperms. Then sequence alignment and phylogenetic analysis were performed. On the whole, the phylogenetic tree is organized into two major clades. The Med16 subunits from unicellular algae (CrMed16, VcMed16, GpMed16) were grouped in one clade and shared less than 15% identity with OsMed16 (Fig. 1b); The Med16 subunits from other plant species were grouped in another clade and shared higher identity with OsMed16 (Fig. 1b). Among the sequences retrieved from NCBI database, OsMed16 displays the highest percentage of identity with ObMed16 from Oryza brachyantha (96%), and has 69% identity with AtMed16.
2 OsMed16 mRNA Expression Pattern and Protein Subcellular Localization
Quantitative real-time PCR (qRT-PCR) assays were performed with total RNA isolated from rice root, leaf, stem, leaf sheath and young panicle. The results showed that OsMed16 mRNA was expressed in all the examined organs, which have similar expression level except a little lower in leaf sheath (Fig. 2a). Furthermore, public microarray databases (eFP browser) indicated that OsMed16 was also expressed in inflorescence and seed (Additional file 1 Fig. S1) (Winter et al. 2007). The wide expression pattern of OsMed16 is consistent with its function as a basic transcriptional regulator.
To determine the subcellular localization of OsMed16, a p35S-OsMed16-GFP construct was generated and transiently expressed in rice protoplasts with a red fluorescent protein (RFP) fused to OsGhd7, a nucleus-localized protein (Xue et al. 2008). The p35S-GFP empty vector was used as a control. As a result, the green fluorescence signal in the control was observed in cytoplasm, while OsMed16-GFP fluorescence was present in the nucleus, co-localized with the OsGhd7-RFP protein (Fig. 2b). These results indicated that OsMed16 is localized in the nucleus, which was in agreement with its role as a Mediator subunit.
3 Overexpression of OsMed16 Caused Rice Growth Inhibition and Spontaneous Cell Death
To investigate the function of OsMed16 in planta, it was disrupted using CRISPR/Cas9 genome-editing technology (Additional file 1 Fig. S2a). The osmed16 mutants exhibited a stunted growth phenotype, failed to head, and died prematurely (Additional file 1 Fig. S2b), indicating that disruption of OsMed16 caused rice seedling lethality.
We further employed gain-of-function approach to investigate the roles of OsMed16. OsMed16 overexpression vector driven by CaMV 35S promoter was constructed and transformed into Nipponbare via an Agrobacterium-mediated method. The expression level of OsMed16 in transgenic plants was detected using qRT-PCR assay, and two representative homozygous transgenic lines with high expression level of OsMed16 (named OsMed16-OE) were used for further investigation (Additional file 1 Fig. S3). Unexpectedly, overexpression of OsMed16 also inhibited rice growth. Compared with the wild-type, OsMed16-OE lines had a dwarf phenotype with fewer tillers (Fig. 3d). Another distinct visible phenotype observed was spontaneous cell death in OsMed16-OE lines. At three leaves stage, small necrotic spots first appeared on the leaf sheath of OsMed16-OE seedlings (Fig. 4a), and then were also observed on leaves (spotted leaf, Fig. 4b-c). As plants grew, the brown spots gradually became large irregular lesions (Fig. 4b-c). The cell death was further confirmed by Trypan Blue staining. The OsMed16-OE leaves showed increased staining intensity compared to wild-type leaves (Fig. 4d-f). Accumulation of reactive oxygen species may cause cell damage and even death (Khanna-Chopra. 2012). By DAB (3,3-diaminobenzidine) staining, the over accumulation of H2O2 was observed in the leaves of OsMed16-OE plants (Fig. 4g-i). We also used NBT (nitroblue tetrazolium) staining and observed the increase of superoxide anion in OsMed16-OE plants (Fig. 4j-l). With the increasing of number and size of lesions, the old leaves of OsMed16-OE lines withered prematurely, and the whole plants exhibited early senescence (Fig. 3b).
4 Overexpression of OsMed16 Reduced Rice Grain Yields
In addition to the growth inhibition, plants overexpressing OsMed16 exhibited significant yield reduction. Compared to wild-type plants, grain yield per plant was reduced by 91.8% and 91.3% in the two overexpression lines (Fig. 5a-b). The yield components were further analyzed. The panicle number per plant, panicle length, and 1,000-grain weight of OsMed16-OE plants decreased significantly compared to the wild-type (Fig. 5c-e). Additionally, the seed length and width were also compared between the OsMed16-OE lines and the wild-type. The results showed that the seed length was unchanged (Additional file 1 Fig. S4a-b), but seed width decreased slightly in OsMed16-OE lines (Additional file 1 Fig. S4c-d).
5 Transcriptome Changes in OsMed16-OE Plants
To assess the influence of OsMed16 overexpression on gene expression, OsMed16-OE plants exhibiting necrotic lesions were harvested, and RNA sequencing (RNA-Seq) was performed on wild-type and OsMed16-OE plants. Overall, we obtained 6 transcriptome data sets, each of which contains an average of about 50 million paired-end (PE) reads (Additional file 1 Fig. S5). The raw sequencing reads were first trimmed and mapped to the rice reference genome using HISAT2. More than 96% reads could map to unique loci per sample (Additional file 1 Fig. S5). Differentially expressed genes (DEGs) were determined with stringent criteria: log2 fold change ≥ 1 and P-value (false discovery rate, FDR) ≤ 0.05. Compared with the wild-type, 2402 DEGs were detected in OsMed16-OE plant leaves, of which 1419 were up-regulated (Additional file 2 Table.S1), whereas 983 were down-regulated (Additional file 2 Table.S2). Gene ontology (GO) enrichment analysis indicated that the up-regulated genes in OsMed16-OE plants were involved in multiple biological processes, including heme binding (66), tetrapyrrole binding (66), oxidoreductase activity (57) and iron ion binding (56). Among these DEGs, CYP71Z2 (LOC_Os07g11739) is the rice cytochrome P450 gene, which participates in plant defense by regulating the secondary metabolism of plant phytoalexin (Li et al. 2013;Li et al. 2015). Rice D3 gene (LOC_Os06g06050), a multi-tiller dwarf gene, encodes an F-box protein rich in leucine repeat sequence, which is not only necessary for the signal transduction of strigolactone (SL), but also involved in leaf senescence and cell death (Ishikawa et al. 2005). OsCAld5H1 (LOC_Os10g36848) encodes a ferulic acid 5-hydroxylase, whose biological function is mainly involved in the synthesis of rice lignin, and its expression affects the composition of S/G lignin in the main nutritional tissues of rice without affecting the structure of vascular bundles (Takeda et al. 2017). HAN1 (LOC_Os11g29290) encodes an oxidase which can catalyze the conversion of biologically active Jasmonate-L-isoleucine (JA-Ile) into the inactive 12-hydroxy-jasmony-L-isoleucine (12Oh-Ja-ILE) and regulate JA-mediated low temperature reaction and cold tolerance as a negative regulator of cold tolerance (Mao et al. 2019). Whereas the down-regulated genes mapped to categories including tetrapyrrole binding (40), heme binding (39), and oxidoreductase activity (34) (Fig. 6a, b). Among these genes, OsAPX2 (LOC_Os07g49400) is an ascorbic acid peroxidase gene that plays an important role in the growth and development of rice by clearing reactive oxygen species to protect seedlings from abiotic stress (Zhang et al. 2013). CYP93G2 (LOC_Os06g01250) encodes the flavanone 2-hydroxylase, which is not only a member of the cytochrome P450 gene but also the first enzyme in its biosynthetic pathway (Du et al. 2010). Our results confirmed that the up-regulated genes and the down-regulated genes were indeed associated with multiple biological pathways in rice.
Overexpression of OsMed16 led to spontaneous cell death in rice, which resembled the hypersensitive response (HR) caused by pathogenic infection. This led us to speculate that overexpression of OsMed16 might trigger the expression of defense-related genes. Thus, we examined these genes in the RNA-seq data. Indeed, some defense-related genes, including PR1a and PR1b, were up-regulated in OsMed16-OE compared with the wild-type. To confirm these results, we further performed qRT-PCR to check the expression levels of eight defense-related genes in the OsMed16-OE and wild-type plants. The transcript levels of all these genes were elevated in OsMed16-OE plants (Fig. 7), suggesting that overexpression of OsMed16 did activate the expression of defense-related genes.