wlp3 albino phenotype at the seedling stage
Under field conditions, the wlp3 mutant exhibited a white stripe phenotype at the two-leaf stage continuing through to the tillering stage (Fig. 1A and B). The white stripes were distributed along the leaf veins of the entire leaf. After the four-leaf stage, the newly emerging wlp3 leaves gradually developed a green color until the WT phenotype was restored. At the heading stage, the wlp3 panicle exhibited the albino phenotype again (Fig. 1C and D). Compared to WT, the panicle was longer, the thousand-grain weight was lower, and the number of grains per panicle was higher, which may be related to the increased ear length and tiller number (Fig. 1E–I). Moreover, the wlp3 leaf color was affected by temperature, but the leaves recovered to a WT-like phenotype at 24°C and the chlorophyll content in wlp3 was almost the same as WT (Fig. 2A, D, and E). However, wlp3 exhibited an obvious albino phenotype at 28 and 33°C and the chlorophyll content was significantly lower compared to WT suggesting that this is a temperature-sensitive mutant (Fig. 2B–C, G–J) .
WLP3 affects chloroplast development
To explore the effect of the wlp3 mutation on chloroplast structure, the ultrastructure of the green and white parts of WT and wlp3 plants was observed by transmission electron microscopy (TEM). In WT, the chloroplast developed normally and the thylakoid structure was ordered (Fig. 3A–C). In the albino part of wlp3, the chloroplast did not reach maturity since the structure of the middle chloroplast was abnormal (Fig. 3D–F). Furthermore, the WT and wlp3 leaves were also observed by laser confocal microscopy. The chloroplast autofluorescence in wlp3 was reduced, suggesting that the wlp3 mutation can cause abnormal chloroplast development.
The seedling stage WT and wlp3 leaf surface structures were observed by scanning electron microscope (SEM). The silicoid papillae around the stomata, the number of stomata, and the stomatal openings were all reduced and the pores were narrowed in wlp3 compared to WT (Fig. 3J–L) .Diaminobenzidine (DAB) and Nitrotetrazolium Blue chloride (NBT) staining demonstrated a large number of precipitated blue and brown particles, respectively, indicating a large accumulation of superoxide anion and superoxidase in wlp3 leaves. Since the accumulation was primarily near the leaf veins, which coincides with the albino parts on wlp3 leaves, the damage to the chloroplast structure in wlp3 may cause the accumulation of peroxidase and superoxide anion (Fig. 3M and N) .
To examine if the WLP3 mutation can cause changes in chlorophyll synthesis rate, WT and wlp3 seeds were cultured for seven days in the dark and then transferred to a light incubator. Leaf chlorophyll content was measured. The chlorophyll synthesis rate was faster at 6 h of light, with no significant difference between the WT and wlp3. WT chlorophyll synthesis rate was higher compared to WT from 24 h. From 36 to 48 h, the chlorophyll synthesis rate of wlp3 significantly increased, reducing the gap between the WT. However, the wlp3 rate was always less than in WT. In conclusion, the WLP3 mutation caused the reduction of chlorophyll synthesis and its accumulation in wlp3 (Supplemental Fig. S1) .
Map-based cloning of WLP3
The albino white spike phenotype was not detected in any F1 plant. Self-crossing of the F1 generation produced F2 progeny with a white-leaf and white-spike phenotype similar to wlp3. Further analyses determined that the ratio of the green to the white-leaf and white-spike phenotypes was close to 3:1. The chi-square test showed that the wlp3 phenotype was controlled by a recessive gene (Supplemental Table 1) .
A single F2 plant with the albino phenotype was used to map the WLP3 gene. The results of the initial mapping showed that WLP3 was located between the molecular markers B3-22 and B3-23 on Chromosome 3 (Fig. 4A). A new indel marker was then developed and WLP3 was ultimately localized within a 56 Kb physical distance between the M4 to M5 indel markers (Fig. 4B). Using the Rice Genome Browser (http://rice.uga.edu/cgi-bin/gbrowse/rice/), 11 open reading frames (ORFs) were identified to be included in this interval (Fig. 4C). After sequencing these ORF-specific primers, it was found that the third exon of the LOC_Os03g61260 gene contained a mutation at position 380 of the coding region where the T base mutated to C (Fig. 4E). This transition mutation resulted in the final translation of the amino acid from isoleucine to threonine (Fig. 4F). Therefore, LOC_Os03g61260 was selected as the candidate gene for WLP3.
To verify if the WLP3 mutation leads to the white leaf and white panicle, Agrobacterium-mediated transformation was used to transform pCAMBIA1300-WLP3, containing a 700 bp promoter upstream of WLP3, into the wlp3 calli. Through phenotypic screening, 19 transgenic plants were successfully transformed and all exhibited a normal phenotype (Fig. 4G). The chlorophyll content of the complementary lines was then examined showing that all returned to the WT level (Fig. 4H). The genetic complementation of WLP3 confirmed that LOC_Os03g61260 was the WLP3 gene. The single-base mutation in this gene resulted in the albino phenotype in rice at the seedling stage returning the green color at the later developmental stages.
WLP3 is expressed in leaves and panicles
WLP3 expression in different parts of the plant was analyzed using GUS staining. WLP3 was almost not expressed in mature leaves and leaf sheaths (Fig. 5C), but most highly expressed in panicles (Fig. 5B), followed by stems (Fig. 5A), and also in coleoptiles (Fig. 5E). The staining of GUS transgenic plants at the seedling stage demonstrated that WLP3 expression was higher in leaves at the seedling stage. Real-time quantitative PCR (qRT-PCR) was used to detect WLP3 expression at different stages and in different tissues. The detection results were consistent with GUS staining, further clarifying the cause of the wlp3 leaves turning from white at the seedling stage to green at the later developmental stages (Fig. 5G). At the tillering stage, the transcriptional level decreases since WLP3 no longer plays a role. WLP3 expression increases in the panicle at the heading stage. Therefore, WLP3 mutation causes albino phenotype in the leaves and panicles of wlp3 at the seedling stage.
Subcellular localization of WLP3
Subcellular localization of WLP3 was investigated using ZH11 cDNA as the template, amplified the target fragment using WLP3-Ecoli1-F and WLP-Kpn1-R, and ligated to PYBA1132-eGFP using infusion. The constructed 1132-WLP3-eGFP and PYBA1132-eGFP empty vectors were transferred into EHA105 cells for transient transformation in tobacco. The fluorescence position of empty vectors was observed at various locations in the cell using laser confocal microscopy (Fig. 5H–K). WLP3 fluorescence coincided with chloroplast autofluorescence, suggesting its expression in chloroplasts (Fig. 5L–O) .
Phylogenetic analysis of WLP3
Homologous sequences were identified using phylogenetics in the genomes of rice, maize, Brachypodium, barley, and other plants (Fig. 6A). The signal peptide site (https://services.healthtech.dtu.dk/services/SignalP-5.0/) predicted a chloroplast signal peptide at amino acids 1–21 (Fig. 6B) and the SMART site (http://smart.embl-heidelberg.de/) predicted the presence of a ribosomal L18p domain from amino acids 51 to 170. L18 (L5e) was a ribosomal protein located in the central protrusion of the large subunit (Supplemental Fig. S2A–B) .
The SWISS-MODEL website was used to predict WLP3 protein structure. 3 α helices and 1 β-sheet on the WLP3 protein were identified and the wlp3 mutation site was located in the α on the spiral (Supplemental Fig. S2C). The results of multiple sequence alignment showed large differences in the first 50 amino acids in the WLP3 sequence, while the latter amino acids had small differences among different species, and the amino acids constituting the domain were more conservative (Supplemental Fig. S2D). These results suggest that WLP3 domains possess important organismal functions and are therefore very conserved during evolution.
The genetic information in the biological genome is first transferred to mRNA via RNA polymerase. The mRNA then is translated by ribosomes to encode the biologically active proteins and perform corresponding functions. GO enrichment and Rice FREND analyses demonstrated that WLP3 was primarily co-expressed with protein translation, protein biosynthesis, organic acid metabolism and isogenic groups (Supplemental Fig. S2E), These results suggest that WLP3 is largely responsible for macromolecule synthesis (i.e., proteins) and plays a central role in the normal growth of rice.
WLP3 mutation affects the expression of genes related to chloroplast development
qRT-PCR was used to examine the relevant gene expression at different temperatures in WT and mutant trifoliate three-leaf stages. The results demonstrated that genes related to chlorophyll synthesis in wlp3 including HEMA1, PORA1, CAO1, and YGL1 (Fig. A7A) and some genes related to ribosomal protein synthesis, such as WGL2, ASL2, and AL2 (Fig. 7B) were significantly down-regulated at 28 ºC. The detection of genes related to chloroplast development revealed that RNA binding protein gene V1, guanyl kinase gene V2, RUBP carboxylase small subunit rbc, PSII light-harvested chlorophyll A/B binding protein Cab1R, and SPP expression was significantly down-regulated. OsRpoTp, the gene encoding NEP, was significantly down-regulated, while rpoA and rpoB, the genes encoding PEP, were significantly up-regulated. In rice, NEP contains only one central subunit encoded by OsRpoTp, while PEP contains four small subunits encoded by rpoA and rpoB genes. The WLP3 mutation resulted in a decreased transcript level of NEP and an increased transcript level of PEP-related genes (Fig. 7C). However, at 23 ºC, the expression of genes related to chloroplast development and chlorophyll synthesis in wlp3 was restored to WT levels (Fig. 7D-F). These results suggest that the recovery of leaf color in wlp3 was caused by the recovery of mRNA expression in the plant under low temperatures.
Current studies have shown that when chloroplast growth and development are hindered, special signal substances to transmit to the nucleus are released by chloroplasts reducing chloroplast protein expression encoded by the nucleus. In this way, cells are adapted to chloroplast development. This is achieved via the nucleocytoplasmic retrograde signaling (RS) pathway. However, when certain genes are mutated, the nucleocytoplasmic retrograde signal is blocked, and the translation level of nuclear-encoded chloroplast proteins is not affected by the signal of chloroplast development. Norflurazon (NF) was therefore used to treat the three-leaf stage to determine if WLP3 mutation may lead to the blockage of nucleocytoplasmic retrograde signaling and if wlp3 may cause genome uncoupling at the molecular level. Fluroxypyr can inhibit nuclear gene expression via the tetrapyrrole biosynthesis pathway. LHCB2 and RBCS2, two key target genes of nucleocytoplasmic retrograde signaling, were analyzed using qRT-PCR with the retrograde signal gene PPI2 as the control. The results demonstrated that LHCB2 and RBCS2 were not expressed in WT after treatment with fluroxypyr, while their expression in wlp3 significantly decreased following treatment. The decline in wlp3 was much smaller compared to WT. The level of change of PPI2 in wlp3 was consistent (Fig. 7G–H), suggesting that nucleocytoplasmic retrograde signaling was blocked in wlp3 displaying a partial genome uncoupling phenotype.
wlp3 is insensitive to ABA and abiotic stress response
To determine if the WLP3 mutation affects the sensitivity of plants to ABA, WT and WLP3 were inoculated on 1/2MS medium. The root and stem lengths were measured after seven days. Both the stems and roots in WT were significantly inhibited on the 2.5 µM medium, with a more pronounced inhibition on the 5 µM medium (Fig. 8A-B). The root and stem lengths of wlp3 on the 2.5 µM medium however were significantly increased compared with the control group but slightly decreased on the 5 µM medium compared with 2.5 µM (Fig. 8C-D). These results suggest that ABA treatment at 2.5 and 5 µM promoted the growth of wlp3, but more at 2.5 µM. Furthermore, the WLP3 mutation reduced plant sensitivity to ABA.
To determine if wlp3 is involved in rice drought and salt stress responses, WT and wlp3 seeds were cultured in a nutrient solution of 20% PEG6000 and 1/2MS medium of 100 mM NaCl. Salt tolerance-related gene expression was analyzed by measuring the root and stem lengths of WT and wlp3 after ten days of growth. The RNA of WT and wlp3 plant was then extracted at this time. The results demonstrated that the root length of wlp3 under salt treatment was significantly lower compared to WT (Supplemental Fig. S3A–D), suggesting that the sensitivity of wlp3 was higher under the 100 mM NaCl stress. The expression levels of drought and salt tolerance-related genes were significantly higher than those of wlp3 (Supplemental Fig. S4A–F), suggesting that wlp3 was more sensitive to drought and salt stress.
WLP3 interacts with other ribosomal proteins
The ribosome is a dense ribosomal protein particle that can dissociate into two subunits, each containing different grouped ribosomal subunits. To verify if WLP3 interacts with other ribosomal subunits, the WLP3 coding region was constructed into PGBKT7 and the coding regions of RPL4, WGL2, ASL2, RPL9, RPL5, RPS6 into PGADT7, using the yeast two-hybrid (Y2H) method. WLP3 and PGADT7 could not grow on the Triple and Quadruple Dropout Supplements, indicating that WLP3 does not have self-activating activity. However, WLP3 and RPL4, RPL5, RPS6, ASL2, and WGL2 could grow on the Quadruple Dropout Supplements and RPL9 on the Triple Dropout Supplements (Fig. 9A). These results suggest that WLP3 and RPL4, RPL5, RPS6, ASL2, and WGL2 exhibit strong interactions, while RPL9 has a weaker interaction.
A bimolecular fluorescence complementation (BiFC) assay was used to verify whether WLP3 interacts with other ribosomal proteins in planta assay. The recombinant plasmids with WLP3-nCFP and other cCFP plasmids were co-injected into tobacco leaves. Using fluorescence confocal microscopy, the tobacco leaves co-injected with WLP3, RPL4, WGL2, ASL2, RPL9, RPL5, and RPS6 all exhibited fluorescence and the interactions occurred both in the nucleus and cytoplasm (Fig. 9B). The construction of the wlp3 single-base substitution onto nCFP demonstrated that wlp3 interaction with other ribosomal subunits was not affected (Supplementary Fig. S5 A) .
In BiFC experiments, the fluorescence produced by the interaction of WLP3 with other ribosomal subunits was not only detected in chloroplasts but also in the cytoplasm and nucleus. Therefore, the interaction of WLP3 with ribosomal subunits was hypothesized to change the localization of WLP3. First, the cDNA of RPL4, WGL2, ASL2, RPL9, RPL5, and RPS6 were constructed into the PYBA-1132-eGFP vector. All subunits localized in the chloroplasts using Agrobacterium-transformed tobacco (Supplementary Fig. S5B). Subsequently, RPL4, ASL2, RPL9, RPL5, and RPS6 cDNA were constructed into the PYBA1138-mcherry vector and transferred into A. tumefaciens and then co-injected into tobacco with PYBA1132-WLP3-eGFP. GFP fluorescence was transferred from the chloroplast to the cytoplasm and nucleus (Fig. 9C). Various ribosomal subunits constructed at the cCFP in BiFC with PYBA1132-WLP3-eGFP were co-injected to eliminate the effect of mcherry protein. The fluorescence was also distributed throughout the whole cell (Supplementary Fig. S5C). These results suggest that the interaction of WLP3 with other ribosomal subunits does indeed affect WLP3 localization.
To identify the key domains with which WLP3 interacts in these proteins, WLP3 was divided into two distinct parts (Fig. 9D) : the N-terminal (1-50aa) and the C-terminal (51-170aa). The truncated fragments containing only part WLP3-N or WLP3-C of the C-terminal domain were cloned and connected to nCFP vectors which were then co-transformed with cCFP-RPL4, WGL2, ASL2, RPL9, and RPL5into EHA105. The results demonstrate that RPL4, WGL2, ASL2, RPL9, RPL5, and RPS6 can only interact with WLP3-C. However, WLP3 could not interact with other truncated domains (Fig. 9E) .