Arabidopsis thaliana ecotype Columbia [wild type (WT)] was used in this study. The T-DNA insertion line (SALK_080811) was obtained from Arabidopsis Biological Resource Center (ABRC, Ohio State University). The T-DNA insertion site was identified by PCR amplification with the T-DNA left border primer LB3 and the gene-specific primers, Left Primer (LP) and Right Primer (RP). For laboratory work, surface-sterilized seeds were sown on Murashige and Skoog (MS) medium supplemented with 2% sucrose and 0.8% (w/v) agar. Plants were grown at 22°C under 16 h light/8 h dark conditions at 30 μmol photons m^–2s^–1.

For genomic DNA isolation, samples were homogenized in extraction buffer [200 mM Tris-HCl, pH 7.5; 25 mM NaCl; 25 mM EDTA;0.5% (w/v) SDS], then the homogenate was extracted through phenol/chloroform. After centrifugation, DNA was precipitated from the supernatant by adding cold isopropyl alcohol. After washing with 70% (v/v) ethanol, the DNA was rehydrated in distilled water. For total RNA extraction, a commercial RNAiso Plus kit (Takara, Otsu, Japan) was used in accordance with the manufacturer’s instructions. Briefly, 2-week-old plants were homogenized in liquid nitrogen, then the homogenate was lysed in the appropriate amount of RNAiso Plus buffer. The mixture was incubated for 5 min at room temperature, then centrifuged at 12,000 rpm for 5 min at 4°C. The supernatant was collected into a new microtube and treated with chloroform. After centrifugation at 12,000 rpm for 15 min at 4°C, an equal volume of isopropanol was added to the supernatant. The pellet was recovered after centrifugation at 15,000 rpm for 10 min at 4°C and dissolved in distilled RNase-free water. Total RNA samples (2–5 μg) were used as templates for the synthesis of the first-strand cDNA with a PrimeScript™ RT Reagent Kit with gDNA Eraser (Takara, Otsu, Japan), according to the manufacturer’s instructions. RT-PCR reactions were performed with specific primers for RP8. Actin2 gene (At3g18780) was used as an internal positive control. Quantitative PCR was performed using the SYBR Green PCR amplification mixture on a StepOnePlus™ Real-Time PCR System (Applied Biosystems, United States). Specific amplification has been confirmed by melting curve analysis. Three biological repeats were performed independently and each sample was operated in triplicate. For Northern blot analysis, equal amounts of total RNA (10 μg) were transferred to positively charged nylon membranes (Roche, Switzerland) after formamide denaturing agarose gel electrophoresis and was further probed with digoxigenin (DIG)-labeled nucleic acid probes (Roche, Switzerland). The probes were synthesized with a PCR DIG synthesis mix (Roche)¹ with the specific primers listed in Supplementary Table 1. Both hybridization and chemiluminescence detection was carried out, according to the Roche manual.

As for the genomic complementation analysis, the 2.9 kb full-length genomic fragment of At4g37920 was amplified using high-fidelity KOD plus polymerase (TOYOBO, Japan)² with the gene-specific primers listed in Supplementary Table 1, and sub-cloned into a binary vector pCAMBIA1301. The construct was introduced into heterozygotes mediated by Agrobacterium tumefaciens strain using floral dip transformation as described (Clough and Bent, 1998). Transgenic plants were screened on MS medium with 80 mg L^–1 hygromycin B (Roche). The genomic backgrounds of these hygromycin-resistant transformants were further analyzed with the specific primers listed in Supplementary Table 1.

Leaf segments were fixed with 2.5% glutaraldehyde in phosphate buffer (pH 7.2) for 24 h at 4°C and washed three times with the same buffer. The tissues were postfixed overnight in 1% OsO₄ at 4°C. The fixed samples were dehydrated through a series of ethanol solutions, infiltrated with a series of epoxy resin in epoxy propane, and embedded in Epon 812 resin. Ultrathin sections were cut with a diamond knife and mounted onto copper grids. Then, the samples were stained with 2% uranyl acetate for 10 min followed by lead citrate for 2 min and observed with a transmission electron microscope (Phillips CM120).

To construct p35S:RP8-GFP, a 297-bp coding fragment containing the transit peptide was amplified using specific primers and cloned into the XhoI and SpeI sites of the GFP vector. GFP was transiently expressed in protoplasts of N. benthamiana using polyethylene glycol protocol (Lyznik et al., 1991). The GFP fluorescence (green) and chloroplasts autofluorescence (red) of N. benthamiana protoplasts were imaged using a confocal laser scanning microscope (Zeiss LSM500). The filter sets used were BP505-545 (excitation 488 nm; emission 505-545 nm) and LP585 (excitation 488 nm; emission 585 nm) to detect GFP and the chlorophyll autofluorescence.

Total proteins were extracted from 2-week-old plants, according to the method described (Yu et al., 2011). Different samples were quantified by Protein Assay (Bio-Rad, United States). A total of 30 μg of protein was loaded per lane and separated by 12–15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). After electrophoresis, proteins were transferred onto a.45 μm PVDF membrane (Millipore, United States) and incubated with antibodies against chloroplast proteins (Agrisera, Sweden).³ Signals were identified by an ECL plus Immunoblot detection system (GE) following the manufacturer’s instructions.

The coding sequence for the RP8 protein without the N-terminal putative transit peptide was amplified with the KOD plus polymerase (TOYOBO, Japan)² with gene-specific primers listed in Supplementary Table 1 and subcloned into pET51b vector, in frame with a 6 × His tag at C-terminus. The RP8 protein was overexpressed in the Escherichia coli BL21 (DE3) strain. When it was grown to OD₆₀₀ = 0.6, 100 mM isopropyl-β-D-thiogalactopyranoside was added. Then, the E. coli was cultured at 16°C for 12 h to induce the recombinant protein. All protein purification steps were carried out as described in a study by Sun et al. (2019). The 5 mg purified RP8 protein was sent to Shanghai Orizymes Biotech Company, Shanghai, China to make an antibody of RP8.

To identify novel nuclear-encoding factors essential for chloroplast development, we screened pigment-deficient mutants of T-DNA insertional lines from the Arabidopsis Biological Resource Center, ABRC. Among the obtained mutant lines, an albino mutant from SALK_080811 with seedling lethal phenotype was characterized and subsequently named as rp8 (RNA Processing 8). PCR sequencing confirmed that the T-DNA was inserted into the fifth exon of the locus, At4g37920 (Figure 1A) as claimed in the SALK database. Genetic analysis revealed that the progenies from heterozygous segregated approximately at a 3:1 ratio, indicating that the mutant was caused by a single recessive mutation. When cultivated on sucrose-supplemented medium under low light conditions at 30 μmol photons m^–2s^–1, homozygous lines still showed an albino phenotype which only can survive for approximately 3 weeks with six to eight leaves (Figure 1B). Examination of young siliques of the heterozygous line also showed that approximately 25% of developing ovules were white (Supplementary Figure 1). RT-PCR results showed that no transcript of the RP8 gene was detected in the rp8 mutant, while it was present in the wild type (Figure 1C). Further immunoblot analysis showed that the corresponding size of 43 kDa protein was not detected in the rp8 mutant, but clearly present in the wild type (Figure 1D).

To further confirm whether the knockout of the At4g37920 locus is responsible for the albino phenotype, we performed the genetic complementation analysis. We cloned a full-length genomic fragment of the At4g37920 gene, then transformed it into the heterozygous lines through Agrobacterium transformation. We obtained more than seven independent transgenic complementation lines with RP8 genotype but showed the wild-type phenotype (Figure 1B). RT-PCR analysis detected the full-length transcript of the RP8 gene in these complementation lines and a specific 43 kDa protein could also be detected by immunoblot analysis as in wild type (Figures 1C,D). Taken together, these results demonstrated the albino mutant was due to the knockout of the At4g37920 locus which is essential for photoautotrophic growth and plant viability in Arabidopsis.

The genome locus At4g37920 encodes a 427-aa-long protein with a predicted molecular mass of 48.7 kDa. It was annotated as an endoribonuclease E-like protein (RNase E-like) in the TAIR database.⁴ Nevertheless, alignment analysis revealed no significant similarity between the RP8 and RNase E proteins, and conserved domain analysis also showed no known information in the currently available database (see text footnote 5). This analysis indicated that RP8 encoded an unknown protein, and it does not belong to the endoribonuclease E-like family. Sequence searching of the PPDB database identified that one homologous of RP8, RP8-like protein (AT1G36320), is probably located in the chloroplast of Arabidopsis. They share 41% identity and 67% similarity at the amino acid level⁵ (Supplementary Figure 2). No obvious functional domain or motif was predicted in either RP8 or the RP8-like protein by any available bioinformatic tools. The BLASTP searching database in the NCBI website revealed that RP8 and the RP8-like protein are present in both dicotyledons and monocotyledons, such as Camelin sative, Brassica napus, Capsella rubella, Ricinus communis and Oryza, Sorghum bicolor, Zea mays, and Physcomitrium patens. But only RP8-like homologous proteins are found in the species including Selaginella moellendorffii, moss, and a green alga. No homologous protein is detected in bacteria (Figure 2A). These data suggested that RP8 and its homologous proteins are widely present in green plants, which probably originated from green alga.

Both the TargetP⁶ (Emanuelsson et al., 2000) and Predator software predicted that the RP8 protein contains a putative chloroplast trans-peptide at the N-terminal region. To further confirm the subcellular localization of RP8, we constructed the coding sequence of the N-terminal 99-amino acid fused to a green fluorescent protein (GFP) under the control of the cauliflower mosaic virus 35S promoter. The plasmid containing the chimeric gene was transformed into the protoplasts of N. benthamiana through a polyethylene glycol (PEG)-mediated transformation method. Transient expression of the fusion protein was then examined by a confocal laser-scanning microscope. Our results showed that the green fluorescence of the chimeric protein was co-localized with chlorophyll autofluorescence (Figure 2B). By contrast, the fluorescence from free GFP protein was present ubiquitously in the cytoplasm. These data suggested that the N-terminal region functions as a trans-peptide that is able to target the RP8 protein to the chloroplast exclusively. Thus, the RP8 protein is localized in the chloroplast. To further demonstrate its location in the chloroplast, the stroma and thylakoid fractions were separated and immunoblot analysis using antibodies against corresponding marker proteins of stroma (RbcL) and thylakoid (PsbA and PetB) were performed. These results indicated that RP8 is associated with the thylakoid membranes (Figure 2C). Taken together, we concluded that RP8 is a chloroplast-localized protein associated with the thylakoid membranes in Arabidopsis.

We examined the expression of the RP8 gene using the publicly available microarray data and Genevestigator v3 (Zimmermann et al., 2004). The developmental expression analysis revealed that the RP8 gene is highly expressed in seedlings, leaves, and flowers (Supplementary Figure 3). The anatomical expression analysis showed a higher expression of the RP8 gene in juvenile leaves compared with other tissues, suggesting that RP8 plays an important role during the early stage of chloroplast development (Supplementary Figure 3). To investigate the expression pattern of the RP8 gene, total RNA was extracted from various tissues of the wild type, quantitative reverse transcription RT-qPCR analysis was performed. Our results showed that the RP8 gene was expressed in leaves, stems, flowers, siliques, and especially high in 8-day-old seedlings. In contrast, the RP8 transcript in roots was barely detected. This result suggests that the RP8 gene is widely expressed in photosynthetic tissues (Figure 3A). We also used an antibody against RP8 to confirm its accumulation in different plant tissues. As Figure 3B showed, the RP8 protein accumulated highly in the 8-day-old seedlings, which is consistent with the transcript levels. Unexpectedly, the RP8 protein can still accumulate in roots, although transcripts of RP8 were barely detected in roots. Taken together, our data showed that the RP8 gene is strongly expressed in photosynthetic tissues at both transcription and protein levels.

To examine the development status of the chloroplast in the rp8 mutant, we observed the chloroplast from 2-week-old seedlings by transmission electron microscopy (TEM). In the wild type, the chloroplast contains a well-organized stroma and stacked grana thylakoids in both cotyledons and true leaves (Figures 4A–C). In contrast, the thylakoids were nearly absent in the cotyledons of the rp8 mutant, and only a few thylakoid lamellae can be observed from its true leaves (Figures 4D–F). These observations indicated that chloroplast development in the rp8 mutant was seriously arrested at the early stage and that RP8 is essential for thylakoid formation and chloroplast development in Arabidopsis. We also investigated the accumulation of photosynthetic-related proteins in the rp8 mutant. These proteins, including three components of PSII complex (PsbA, PsbD, and OEC33), two components of PSI (PsaB and PsaD), one component of cyb6/f complex (PetA), two ATP synthase subunits (AtpB and AtpF), one subunit of light-harvest complex and chlorophyll a/b-binding protein (Lhcb1), and the large subunit of rubisco (RbcL), were checked using their corresponding antibodies. Our results showed that the amounts of AtpB and RbcL were reduced in the rp8 mutant (Figure 5 and Supplementary Figure 5). PsbA, PsbD, OEC33, PsaB, PsaD, and Lhcb1 are barely detected in the rp8 mutant, although these proteins can be detected in the wild type. These results indicate that components of photosynthetic complexes were seriously reduced in the rp8 mutant. Altogether, these results prove that the accumulation of photosynthetic proteins and the development of chloroplasts were seriously blocked in the rp8 mutant.

Because numerous nucleus-encoded genes regulate chloroplast gene expression (Zhou et al., 2009; Chateigner-Boutin et al., 2011; Gao et al., 2011; Yu et al., 2013; Huang et al., 2018; Wang et al., 2020), we - want to check the putative role of RP8 in chloroplast RNA metabolism by investigating chloroplast transcript profiles in the rp8 mutant by Northern blot. The psaA, psbA, psbB, petB, and rbcL genes were chosen because they are PEP-dependent (class I). The rps11, rpoA, rpoB, and rpoC2 were chosen because they are NEP-dependent (class III). The atpB and clpP were chosen because they are both PEP and NEP-dependent (class II) (Hajdukiewicz et al., 1997). As for those PEP-dependent chloroplast transcripts, their accumulations were seriously reduced in the rp8 mutant compared with those of the wild type (Figure 6A). In contrast, the accumulations of NEP-dependent transcripts (rps11, rpoB, rpoC2) were increased in the rp8 mutant (Figure 6B). As for the atpB operon, 2.6- and 2.0-kb transcripts are transcribed by PEP (σ dependent) and NEP, respectively (Schweer et al., 2006). In the rp8 mutant, the 2.6-kb transcripts observed in the wild-type plant are not transcribed, whereas a novel 4.8-kb transcript transcribed by NEP was detected in the rp8 mutant (Figure 6C). We also checked the transcripts of two nucleus genes Lhcb1 and AtpC1, and no significant alteration was detected between the rp8 mutant and the wild type (Figure 6D). This expression pattern is similar to other PEP-deficient mutants that have been reported earlier (Pfalz et al., 2006; Gao et al., 2011; Yu et al., 2013). Thus, our data showed that the null mutation of RP8 affected the proper expression of chloroplast genes, resulting in reduced amounts of PEP-dependent chloroplast transcripts.

Noticeably, the rpoA gene encoding the alpha subunit of the PEP complex was transcribed by NEP. Here, we found that the number of mature transcripts of rpoA (about 990 nt) was clearly decreased in the rp8 mutant. In contrast, no obvious difference in the processing of the rps11 transcript located in the rpoA polycistron was observed between the wild type and the rp8 mutant. Since the accumulation of mature transcripts of rpoA was decreased in the rp8 mutant, we further checked whether the level of the RpoA protein was affected. Immunoblot analysis showed that the protein level of RpoA in the mutant was dramatically reduced by 50% (Figure 6F). This result suggests that the defective rpoA polycistronic processing affects the accumulation of the RpoA protein. Similarly, we investigated the accumulation of another core protein of the PEP complex, RpoB, in the rp8 mutant. Our data showed that the RpoB protein level was also dramatically reduced by 25% in the rp8 mutant (Figure 6G). Taken together, our data suggested that the RP8 deletion affects the processing of rpoA transcripts, which resulted in the defective accumulation of the PEP core complex and impaired expression of PEP-dependent chloroplast genes at the early growth stage.

Ribonucleic acid processing 8 (RP8) is associated with thylakoids, as translationally active ribosomes are (Zoschke and Barkan, 2015). In the rp8 mutant, accumulation of both alpha- and betta-PEP subunits (RpoA and RpoB) are affected (Figures 6F,G), and photosynthetic membrane proteins are barely detected (Figure 5). We further checked the possibility of RP8 interference with translation. First, we checked the accumulation of chloroplast rRNAs in the rp8 mutant compared with that of the wild type. Northern blot analysis showed that the abundance of the chloroplast rRNA, including 23S, 16S, 4.5S, and 5S rRNA, was severely decreased in the rp8 mutant (Figure 7A). Second, we checked the amount of chloroplast ribosome proteins in the rp8 mutant. Our immunoblot analysis showed that the amounts of chloroplast ribosome proteins, including PRPS1 (uS1c), PRPS5(uS5c), PRPL2 (uL2c), and PRPL4 (uL4c), were substantially reduced in the rp8 mutant (Figure 7B), compared with those in the wild type. These data showed that knockout of RP8 seriously affected chloroplast translational machinery.