RETRACTED ARTICLE: BrRNE cleaves RNA in chloroplasts, regulating retrograde signals in Brassica rapa L. ssp. pekinensis

KEY MESSAGE
Brassica rapa RNE participates in the processing of polycistronic precursor transcripts into mature monocistronic mRNAs in plastids, thereby sending strong retrograde signals. Leaf color is one of the most important agronomic traits for Chinese cabbage. Not only is it closely linked to photosynthesis, thereby affecting plant growth, but it also influences consumer preference in the marketplace. A pale-green mutant rne was produced by EMS mutagenesis of Chinese cabbage inbred line A03. Chlorophyll content, photosynthetic rate, actual quantum efficiency (φPSII), and maximum quantum efficiency (Fv/Fm) of photosystem II (PSII) were all reduced in rne plants. Genetic analysis indicated that the pale-green trait was controlled by a pair of recessive alleles. Using mixed pool sequencing of F2 individuals derived from an rne × wild-type cross, we identified the essential gene Brassica rapa RNase E (BrRNE), which is responsible for chloroplast development. BrRNE cleaves polycistronic RNA in Chinese cabbage A03 plastids, but rne plants are defective in RNA processing and show reduced translation levels of the seven plastid genes, BrpsaB, BrpsaA, BrpsbA, BrpsbD, BrpsbB, BrpetA, and Brycf4. Abnormal RNA processing in the plastids sends retrograde signals that markedly regulate the expression of nuclear genes, upregulating genes that participate in ribosome and DNA replication pathways and repressing photosynthesis-associated nuclear genes (PhANGs). Our study reveals a new regulatory mechanism by which plastid RNA cleavage influences plastid development and leaf color, sending retrograde signals that affect the expression of nuclear genes in Brassica.


Introduction
Photosynthesis is essential for plant growth and development, and chlorophyll (Chl) is the main pigment that absorbs light energy and drives electron transport in the photosynthetic reaction centers of higher plants (Tanaka and Tanaka 2006). Defects in Chl biosynthesis, degradation, or other related pathways often result in leaf color mutants. These mutants are widely distributed in nature and produce a variety of phenotypes, including albino, virescent, chlorina, xanthas, maculate, striped, and dark green (Jung et al. 2003;Manjaya 2009;Singh and Ikehashi 1981). Leaf color mutants are ideal genetic materials with which to study the molecular mechanisms of plant photosynthesis and chloroplast development.
Leaf color mutants are induced by multiple genetic and environmental factors, among which genetic change plays a decisive role. Previous studies have reported that the main molecular mechanisms of chl-related mutations are: (1) mutations in genes of the chl biosynthesis pathway such as CAO (Tanaka 1998), CHLH (Jung et al. 2003), DVR (Nagata et al. 2005), PORB, and PORC (Frick et al. 2003;Beale 2005); (2) mutations in heme metabolism genes that influence chromophore biosynthesis (Parks and Quail 1991;Terry and Kendrick 1999;Muramoto et al. 1999);and (3) mutations in genes related to chloroplast differentiation and development, including ALBINO3 (Sundberg et al. 1997), PGP1 (Babiychuk et al. 2003), and RNE (Mudd et al. 2008;Walter et al. 2010;Schein et al. 2008). Significant progress has been made in the study of leaf color in model plants such as Arabidopsis thaliana, rice, and maize, but there have been fewer such studies on vegetable crops. The development of next-generation sequencing technology, combined with classical genetic methods, has enabled the discovery of leaf color-related genes in vegetables. For example, six genes (LG1_162414, GST, CAD, MYB113, bHLH42, and ANS) have been found to regulate the diversity of leaf color in lettuce (Zhang et al. 2017a, b).
Chinese cabbage (Brassica rapa L.ssp. pekinensis) originated in central China; it is an important vegetable crop that is most widely grown in Asia. B. rapa is among the most closely related species to the model plant Arabidopsis thaliana, in which research on leaf color development has laid a foundation for its study in Brassicas. However, the Chinese cabbage genome has undergone triplication, and some genes are present in multiple copies, making the regulation of leaf color development more complex. To date, a variety of leaf color mutants have been obtained in Brassica crops such as cauliflower (Chiu et al. 2010), oilseed rape (Zhao et al. 2001), pakchoi , and Chinese cabbage . The orange inner leaf of Chinese cabbage is controlled by a single recessive gene (or) that causes abnormal accumulation of carotene (Feng et al. 2012). BrCRTISO, a carotenoid isomerase specifically required for carotenoid biosynthesis, was identified as a candidate gene for the control of orange inner leaf (Zhang et al. 2013. The purple-leaf mutation of Chinese cabbage is regulated by a pair of dominant alleles, BrPur  and BrMYB2 (He et al. 2020). In non-heading Chinese cabbage, the evergreen leaf mutant phenotype is controlled by a pair of recessive nuclear alleles, Brnye1  and BrNYM1 (Wang et al. 2020). Overall, despite the identification of some leaf color-related genes in Chinese cabbage, little is known about their regulatory mechanisms.
Here, we identified a pale-green Chinese cabbage mutant in an EMS-mutagenized population. It retained a pale-green leaf color and exhibited reduced chlorophyll content relative to the wild-type during the seedling stage. We discovered that the mutation responsible for the palegreen trait was a stop-gain base pair change in chromosome A07. Compared with the wild-type, the mutant had an RNA processing defect in the chloroplast and exhibited repressed plastid gene translation. These results revealed that the function of BrRNE was to regulate plastid RNA cleavage and translation. Our findings may aid in the development of new tools for the genetic improvement of horticultural crops.

Plant materials
A mutant library of Chinese cabbage was generated by treating the seeds of inbred line A03 with ethyl methanesulfonate (EMS) (Lu et al. 2016), and the pale-green mutants (rne and rne2) were isolated from the M 6 generation. A03 × rne F 1 and F 2 populations were developed and used for the genetic analysis of mutant traits. The plants were grown in green house at Hebei Agricultural University in Baoding (115.47 E,38.87 N), China, in 2016 and 2018. In August 2016, 60 M 6 plants of A03 and rne were grown in the same green house at Hebei Agricultural University. At the rosette stage (40 days after sowing), the soft portion of the second leaf from the interior, excluding the leaf petiole, was sampled ( Fig. 1c). All leaf samples were snap frozen in liquid nitrogen and stored at − 80 °C for RNA extraction or immunoblotting.

Chlorophyll and carotenoid contents, photosynthetic rate, φPSII and Fv/Fm
At the rosette stage, chlorophyll and carotenoids contents of rosette leaves were measured according to Lichtenthaler et al. (1983). Photosynthetic rate of rosette leaves was determined using the LI-COR LI-6400XT portable photosynthesis system (LI-COR Biosciences, Lincoln, NE, USA). φPSII and Fv/Fm were measured using the PAM-2500 portable chlorophyll fluorescence apparatus (Walz, Germany). All Fig. 1 Phenotypic characterization and morphology of the palegreen rne mutant. a-d The phenotype of wild-type and rne plants at the seedling stage (a) and the rosette stage (b-d). Bars = 2 cm. e-h Chlorophylls and carotenoids contents (e), photosynthetic rate (f), actual quantum efficiency (φPSII) (g), and maximum quantum efficiency (Fv/Fm) (h) of wild-type and rne leaves at the rosette stage. Chl a, chlorophyll a; Chl b, chlorophyll b; Car, carotenoids. Data represent mean ± s.d. of three independent biological replicates (from different seedling leaves). **P-value < 0.01 (Student's t-test). ***P-value < 0.001 (Student's t-test). i-n transmission electron microscopy images of cells from wild-type (i-k) and rne (l-n) leaves at the rosette stage. Magnification is indicated on the bottom of each image ◂ R E T R A C T E D A R T I C L E measurements were conducted with three independent biological replicates (from different seedling leaves).

Transmission Electron Microscopy (TEM)
For TEM, the middle portion of the rosette leaves at the rosette stage was cut into 1-mm 2 fragments and fixed in 2.5% glutaraldehyde in 0.1 M sodium phosphate buffer (PBS, pH 7.3) for at least 4 h at 4 °C. The tissue was post-fixed with 1% osmium tetroxide for 2 h after extensive washing in PBS at room temperature. After dehydration in a graded ethanol series, the tissue was infiltrated using a Spurr Low Viscosity Embedding Kit (Sigma-Aldrich). Ultrathin sections (70-90 nm) were cut with a diamond knife on a Leica EM UC7 Ultramicrotome (Leica Microsystems) and examined using a Hitachi H7600 transmission electron microscope at 75-100 kV.

RNA extraction and qPCR analysis
Total RNA was extracted using TRIzol reagent (Invitrogen). RNA was reverse transcribed into cDNA using the Prime-Script RT Reagent Kit with gDNA Eraser (Takara) and used for qRT-PCR. All qRT-PCR experiments were performed in three biological replicates and three technical replicates using SYBR Green Master Mix (Vazyme) and the CFX Connect Real-Time PCR System (BioRad). The procedure for qRT-PCR was as follows: 10 min at 95 ℃, followed by 40 cycles of 10 s at 95 ℃, 10 s at 57 ℃, and 10 s at 72 ℃. Then, a melting curve was generated by gradually increasing the temperature to 95 ℃ to ensure the specificity of PCR amplification. Relative gene expression levels were calculated with the 2 −ΔCT method. Specific primer sequences are summarized in Table S2.

Subcellular localization in Nicotiana benthamiana
The coding sequences of BrRNE and BrmRNE were amplified by PCR from the cDNA using the KOD-Plus-Neo enzyme (TOYOBO). Purified PCR products were cloned separately into the N-terminus of eGFP between the NcoI and SpeI restriction sites under the control of the CaMV 35S promoter using the In-Fusion HD Cloning Kit (Takara), and Sanger-sequencing was used to validate the resulting constructs. The validated expression vectors were transformed into the Agrobacterium tumefaciens strain GV3101. Five-week-old fully expanded N. benthamiana leaves were infiltrated with 35S::BrRNE-GFP, 35S::BrmRNE-GFP, and 35S::GFP (an empty vector control) that had been diluted to an OD 600 of 0.6-0.8 with a solution of 10 mM MES (pH 5.6), 150 μM acetosyringone, and 10 mM MgCl 2 . Two days after infiltration, small leaf sections were examined under a Leica TCS SP5 confocal microscope. A 488 nm argon laser and a PMT detector with emission bandwidth set to 500-530 nm were used to monitor GFP fluorescence. Specific primer sequences are presented in Table S3.

Northern blotting
Total RNA was extracted from wild-type and rne plants using TRIzol reagent (Invitrogen). For northern blotting, RNA samples (20 μg total RNA) were electrophoresed in formaldehyde-containing 1% agarose gels and blotted onto nylon membranes (Millipore). Hybridization probes were generated by direct polynucleotide synthesis and labeled with digoxin at their 5′ ends. Hybridizations were performed at 42℃ in a hybridization buffer (Roche). After anti-digoxigenin AP-conjugate (Roche) incubation, CSPD chemiluminescent substrate (Roche) was added, and images were obtained with a Tanon 5200 imaging system.

Transcriptome sequencing and data analysis
Total RNA was extracted from wild-type and rne plants using TRIzol reagent (Invitrogen). Transcriptome sequencing was performed as previously described (Zhao et al. 2015). P-values were adjusted using the Benjamini and Hochberg (1995) method, and genes with an adjusted P-value < 0.05 and |log 2 (FoldChange)|> 1 were considered to be differentially expressed. GO enrichment analysis of the differentially expressed genes was performed using the GOseq R package (Young et al. 2010). GO terms with a corrected P-value < 0.05 were considered to be significantly enriched. The KOBAS software was used to test the statistical enrichment of differentially expressed genes in KEGG pathways (Mao et al. 2005).

MutMap and kompetitive allele-specific PCR (KASP)
MutMap genome sequencing was performed as previously described (Abe et al. 2012). The primers used for PCR sequencing of SNPs in genes are listed in Table S1. In brief, a pool containing equal amounts of DNA from 30 pale-green rne × wild-type F 2 plants was re-sequenced (30 × coverage). Delta SNP index = 1 was used to filter SNP loci in the mutant genes. SNP markers were used for KASP genotyping. The co-segregating SNP marker was selected by analysis of both the allelic site and the leaf color trait.

Phenotypic characterization and morphology of the pale-green rne mutant
Pale-green rne plants were clearly smaller than wild-type plants and displayed a chlorotic phenotype (Fig. 1a-d). Visible chlorosis appeared initially in the cotyledons of the mutant plants ( Figure S1). As the plants grew, all the true leaves became chlorotic as they emerged. When grown under photoautotrophic conditions in the soil, rne plants grew somewhat slowly (Fig. 1a-d) but survived well, flowering and producing seeds, albeit in limited quantities.
We next characterized the physiological basis of the rne mutation in detail. Chlorophyll and carotenoid contents, photosynthetic rate, and the efficiency of photosynthetic electron transport were determined by spectroscopy, infrared gas analysis, and chlorophyll fluorescence measurements. As suggested by their pale-green phenotype, rne plants had a reduced chlorophyll content (Fig. 1e). However, the ratio of chlorophyll a to chlorophyll b and the carotenoid content did not differ between rne and wild-type plants (Fig. 1e). Photosynthetic rate was also significantly reduced in rne plants (Fig. 1f), as were the actual quantum efficiency (φPSII) and the maximum quantum efficiency of photosystem II (PSII; Fv/Fm), which are standard measures of PSII integrity ( Fig. 1g-h). These results suggested that a proportion of the PSII reaction centers in the rne mutant may have been damaged.
To confirm the defect in chloroplast development indicated by the pale-green leaves of rne plants, the true leaves of wild-type and rne plants were observed by transmission electron microscopy (TEM). TEM analysis revealed that chloroplasts in the leaves of wild-type seedlings contained a well-formed thylakoid system of stromal and granal thylakoids, but thylakoid formation was severely altered in the mutant. In the chloroplasts of rne mutant leaves, both the stromal and granal thylakoids took the form of unusual vesicle-like structures. The plastids themselves were smaller and had fewer thylakoids and shorter granal stacks than

R E T R A C T E D A R T I C L E
wild-type chloroplasts ( Fig. 1i-n). Nonetheless, the chloroplasts of both rne and wild-type plants accumulated starch granules normally ( Fig. 1j-k, m-n), suggesting that the photosynthetic capacity of the rne plants was not strongly affected.

Inheritance of the mutant trait
Genetic analysis showed that the phenotypic traits of all F 1 plants were similar to those of the wild-type. Analysis of the F 2 populations in 2016 and 2018 showed that the proportions of wild-type and mutant plants were 3.2:1 and 2.9:1, respectively, conforming to a 3:1 ratio (χ 2 = 0.132 and 0.062, respectively) (Table 1). Therefore, we initially speculated that the pale-green trait was controlled by a pair of recessive alleles.

Identification of candidate mutant genes
DNA samples were collected from 30 recessive mutant F 2 progeny (rne DNA pool) and 30 wild-type plants (wild-type DNA pool) and bulked sequencing with substantial genomic coverage (30 ×), respectively. Comparison of sequence reads from the wild-type and rne DNA pools, identified a single unique genomic region that harbored a cluster of SNPs with SNP index of > 0.8 on chromosome A07 that contained 72 SNPs (Fig. 2a). Among the SNPs with SNP index of 1, only four were predicted to cause amino acid changes, including one non-synonymous and three stop-gain substitutions, representing four different mutant genes.
To verify whether the mutant phenotype was caused by a mutation in the candidate gene, sequencing primers were designed near the mutation site and used to amplify and sequence the PCR products. The results indicated that only two of the stop-gain substitutions were real ( Figure S3).
Kompetitive Allele-Specific PCR (KASP) analysis was further performed in 217 F 2 progeny (rne × wild-type). Of the two validated SNP markers (causing stop-gain substitutions), only the SNP marker in BraA07000683 cosegregated with leaf color (Fig. 2c); 55 pale-green F 2 progeny had T:T, and 162 green F 2 progeny had T:C or C:C (108 T:C and 54 C:C). The proportion of green and pale-green plants conformed to a 3:1 ratio (χ 2 = 0.014), and the proportion of T:T, T:C and C:C conformed to a 1:2:1 ratio (χ 2 = 0.021). In addition, transcript levels of BraA07000683 in roots, stems, and leaves indicated that the gene was preferentially expressed in leaves (Fig. 2d). Thus, the gene BraA07000683 was considered to be responsible for the leaf color development phenotype.
BraA07000683 is the homolog of the Arabidopsis RNase E/G gene AtRNE (AT2G04270) and was therefore named BrRNE. B. rapa contains a single RNase E/G gene in its haploid genome. The stop-gain substitution occurred before the RNase_E_G domain (PF10150) of BrRNE, causing a defect in this domain (Fig. 2b, Figure S4).
To confirm that BrRNE was the key gene responsible for the pale-green phenotype, we investigated the DNA sequencing data from the Brassica rapa EMS mutant population (Lu et al. 2016). One SNP was observed in the first base of the 11th intron of BrRNE in a pale-green leaf EMS line (rne2) (Figure S5a-c). The transcript of BrRNE was analyzed further, and the splice donor variant (BrmRNE') of BrRNE was also obvious in rne2 ( Figure S5d), strongly suggesting that BrRNE plays an essential role in the regulation of leaf color morphology in Chinese cabbage.
To investigate putative RNE homologs in higher plants, we searched for RNE protein homologs in 12 angiosperm genomes: eight eudicot species, four monocot species, and a basal angiosperm. Domain analysis revealed that all the angiosperm RNEs contained an RNase_E_G domain, which was highly conserved in flowering plants ( Fig. 2e-g). Phylogenetic analysis showed that BrRNE clustered together with Arabidopsis RNE (AtRNE) (Fig. 2e), suggesting that the functions of BrRNE in Chinese cabbage may be similar to those of AtRNE.

BrRNE cleaves RNA and affects plastid gene expression in the chloroplast
To determine the subcellular localization of the wild-type and mutant BrRNE proteins, we constructed a gene fusion of BrRNE or mutant BrRNE (BrmRNE) with the GFP gene and transformed the chimeric gene into tobacco epidermal cells. The green fluorescence of the reporter protein overlapped with the red fluorescence of chlorophyll, strongly

R E T R A C T E D A R T I C L E
suggesting that BrRNE was localized to the chloroplast. By contrast, BrmRNE was localized to both the chloroplast and the plasma membrane, and free GFP signal was observed throughout the entire cell (Fig. 3a). This result is consistent with previous reports of AtRNE subcellular localization in the chloroplast (Schein et al. 2008;Mudd et al. 2008;Walter et al. 2010). Taken together, our data indicate that BrRNE is localized in plastids, suggesting that its role may be similar to that of AtRNE in plastid development.

R E T R A C T E D A R T I C L E
Endoribonuclease E (RNE) is a member of the RNase E/G ribonuclease family (Mudd et al. 2008). To determine the role of BrRNE in B. rapa chloroplasts, the complete B. rapa plastid genome sequence was downloaded from NCBI (GenBank: MT726210.1) and assembled into a circle graph ( Figure S6) with the OGDRAW tool (http:// ogdraw. mpimpgolm. mpg. de/, Greiner et al. 2019). The B. rapa plastid genome is 153,494 bp in length, contains 132 genes, and includes two large inverted repeats (IRs, IRA, and IRB) separated by large and small single-copy regions (the LSC and SSC) ( Figure S6).
The putative defects in chloroplast RNA processing in rne plants were further investigated by northern blotting in wild-type and rne plants. We analyzed the transcription accumulation of four plastid genes, including BrpsbB (a PSII gene that is part of a large polycistronic transcription unit), Brycf4 (a PSI gene that is also part of a large polycistronic transcription unit), BrrpoA (an RNA polymerase gene), and Brrpl20 (a chloroplast ribosomal gene). Apparent massive overaccumulation of unprocessed polycistronic precursor transcripts was observed in rne plants, suggesting that RNE is involved in processing polycistronic precursor transcripts into mature monocistronic mRNAs in B. rapa chloroplasts (Fig. 3c). In addition, immunoblot analyses using antibodies against BrpsaB (a core protein of PSI multisubunit complex comprised of psaA and psaB), BrpsaA (another core protein of the PSI), BrpsbA (the D1 protein of PSII), BrpsbD (the D2 protein of PSII), BrpsbB (the CP47 subunit of photosystem II), BrpetA (a cytochrome f apoprotein of photosystem I), and Brycf4 (a photosystem I assembly protein) directly confirmed their reduced accumulation (Fig. 3b), suggesting that impaired precursor processing reduced protein translation in plastids.

Abnormal plastid RNA processing causes substantial retrograde signaling to the nucleus
Changes in chloroplast development and/or gene expression can cause massive changes in nuclear gene expression (Chi et al. 2015). We further analyzed our RNA-seq data to investigate global changes in the nuclear gene expression of rne plants. The expression of 4026 nuclear genes was increased and that of 4277 nuclear genes was reduced in rne plants compared with the wild-type (P < 0.05; |log 2 (FoldChange)|> 1) ( Figure S7).
We further categorized the differentially expressed genes using gene ontology (GO) enrichment analysis. Genes upregulated in rne versus wild-type plants were significantly enriched in DNA replication (biological process), ribosome (cellular component), and structural constituent of ribosome (molecular function) (Fig. 4a). The genes upregulated in rne were also enriched in a number of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways, including ribosome, DNA replication, homologous recombination, mismatch repair, and base excision repair (Fig. 4c).
Among the genes downregulated in rne versus wildtype plants, the most highly enriched GO terms were photosynthesis (biological process), thylakoid (cellular component), and tetrapyrrole binding (molecular function) (Fig. 4b). The top six KEGG pathways enriched in downregulated genes were photosynthesis, starch and sucrose metabolism, phorosynthesis-antenna proteins, glyoxylate and dicarboxylate metabolism, plant-pathogen interaction, and carbon fixation in photosynthetic organisms (Fig. 4d). These results suggest that the expression of photosynthesis-associated nuclear genes (PhANGs) is repressed in the rne mutant and that PhANGs with reduced expression in rne primarily functioned in PSI and PSII, light harvesting, carbon fixation, and electron carrier processes (Table S4).

Leaf color mutants are valuable genetic resources for Chinese cabbage breeding
As a group, the Brassicas are one of the most important vegetables worldwide and include Chinese cabbage, pakchoi (Brassica rapa L. ssp. chinensis), cabbage (Brassica oleracea var. capitata), cauliflower (Brassica oleracea var. botrytis), and others. Leaf color is one of the most important agronomic traits for Brassica vegetables; it is closely associated with consumer choice and affects the Fig. 3 B. rapa RNase E (BrRNE) cleaves polycistronic RNA in Chinese cabbage plastids. a Subcellular localization of the BrRNE-GFP and B. rapa mutant RNase E (BrmRNE)-GFP fusion proteins in Nicotiana benthamiana epidermal cells were observed and imaged by confocal microscopy. The GFP fluorescence (GFP), the chlorophyll autofluorescence (Chlorophyll), the bright-field images, and the merged images are shown. The bottom panel shows a negative control using an empty vector with free GFP. Bars = 50 μm. b Accumulation of BrpsaB (a core protein of the PSI multisubunit complex comprised of BrpsaA and BrpsaB), BrpsaA (another core protein of the PSI complex), BrpsbA (the D1 protein of PSII), BrpsbD (the D2 protein of PSII), BrpsbB (the CP47 subunit of PSII), BrpetA (a cytochrome f apoprotein of PSI), and Brycf4 (a photosystem I assembly protein) determined by immunoblot analysis with specific antibodies. Rubisco stained with Ponceau S is used as a control for total chloroplast proteins. c RNA accumulation and processing patterns of the four plastid genes, BrpsbB, Brycf4, BrrpoA and Brrpl20 in Chinese cabbage rne mutants. Northern blots were hybridized to specific probes for the four plastid genes. Sizes of marker bands are given in kb. Precursor transcripts that overaccumulated in rne plants are indicated by asterisks. For quantitative comparison, a dilution series of the wild-type (WT) and rne RNA samples was loaded ◂

R E T R A C T E D A R T I C L E
vegetables' quality and marketability. Although leaf color development is well studied in Arabidopsis, maize and rice, the essential genes and regulatory mechanisms that determine leaf color in Brassica vegetables remain elusive.
There has been recent progress in understanding the color development of Brassica vegetables. The stay-green genes Brnye1 and Brnym1 have been identified and analyzed in pakchoi and Chinese cabbage, respectively (Wang et al. , 2020. The Or gene is known to cause weakly pigmented or unpigmented tissues to turn orange in cauliflower (Li et al. 2001), and the Br-Or gene was mapped and shown to confer orange inner leaf color in Chinese cabbage (Feng et al. 2012). Molecular markers derived from Br-or and BrCRISTO1 cause orange-colored inner leaves in Brassica rapa (Zhang et al. 2013;Lee et al. 2014;Zou et al. 2016), and loss of BrCRISTO1 function confers orange color to the inner leaves of Chinese cabbage (Su et al. 2015;Zhang et al. 2015). In addition, BrPur and BrMYB2 have been mapped and shown to be tightly linked to the purple-leaf phenotype He et al. 2020). Nevertheless, other genes that regulate leaf color remain to be discovered in Chinese cabbage. Although a pale-green EMS mutant was shown to harbor a mutation in the plastid gene Rps4 (Tang et al. 2018), no nuclear genes that control pale-green traits have been reported previously in Brassica. In this study, an EMS Chinese cabbage mutant with pale-green leaves was used to map the gene BrRNE, which is essential for leaf color development ( Fig. 1a-d,  Fig. 2a). In addition, the regulatory mechanism by which BrRNE influences chloroplast development in Chinese cabbage was analyzed.

Plastid RNA processing is controlled by the nuclear gene RNE in Brassica
Chloroplasts evolved from cyanobacterial ancestors by endosymbiosis. They have their own genome and gene expression system, including bacterial-type transcription and translation machineries (Harris et al. 1994;Dyall et al. 2004) that perform intron splicing (Zhang et al. 2017a, b), post-transcriptional cleavage of polycistronic mRNA into monocistronic units (RNA cutting) (Westhoff and Herrmann 1988), tRNA modification in translation , and ribosomal RNA processing (Liu et al. 2015). The steps of each process are coordinated, and defects in these processes affect chloroplast development, leading to distinct albino or pale-green phenotypes. RNase E/G type endoribonucleases have an essential role during plastid RNA cleavage in higher plants (Stoppel and Meurer 2012).
RNE is an endoribonuclease that has been studied primarily in Escherichia coli, where it plays a prominent role in the processing and degradation of RNA (Schein et al. 2008). An RNase E-like protein in higher plants was first characterized in Arabidopsis (Schein et al. 2008;Mudd et al. 2008). Arabidopsis RNE is present in the chloroplast, cleaves RNA similarly to the E. coli enzyme, and is essential for chloroplast development and autotrophic growth (Schein et al. 2008;Mudd et al. 2008). RNE participates in the intercistronic processing of primary transcripts from chloroplast operons, and rne plants show plastid ribosome deficiency because of the disturbed maturation of a transcript that encodes essential ribosomal proteins. Nonetheless, RNE is not essential for the survival of Arabidopsis plants growing in soil (Walter et al. 2010). Although RNE lacks a degradosome homolog in plant plastids compared with that in E. coli (Schein et al. 2008;Mudd et al. 2008;Stoppel et al. 2012), RHON1 directly interacts with RNE in the same highmolecular-weight (HMW) multiprotein complex, binding to single-stranded (ss) RNA to ensure the efficient processing of plastid transcripts by RNE in Arabidopsis (Stoppel et al. 2012). At present, little is known about RNE in other crops. In our study, rne plants grew autotrophically and survived well in soil and MS medium containing sucrose, but they could not survive in MS medium without sucrose under autotrophic conditions ( Fig. 1a-b, Figure S2), This result suggests that the survival of rne plants may depends on carbon source supplementation. BrRNE was localized to the chloroplast (Fig. 3a) and cleaved chloroplast operons into monocistronic units (Fig. 3c). Its mutation reduced plastid protein translation levels (Fig. 3b), affecting chloroplast development and resulting in pale-green leaves in Chinese cabbage ( Fig. 1a-d, l-n).

Aberrant plastid RNA cleavage causes strong retrograde signaling
It is now widely accepted that chloroplasts evolved from free-living prokaryotic organisms that were capable of photosynthesis (McFadden 2001). After serial endosymbiotic events, most genes encoding chloroplast proteins reside in the nucleus (Jarvis 2001). Chloroplast development is regulated by the coordinated expression of both chloroplast and nuclear genes (Pogson and Albrecht 2011). Signaling between the chloroplasts and the nucleus is bidirectional (Jung and Chory 2010). In anterograde regulation, the biogenesis and homeostasis of chloroplasts are controlled by their own genetic system in coordination with the nucleocytosolic system (Jarvis and López-Juez 2013). In retrograde signaling, nuclear gene expression is regulated as a result of signals generated in the plastids (Jung and Chory 2010). In the absence of chloroplast development, the expression of nuclear genes that encode chloroplast proteins such as light harvesting complex proteins is repressed (Jung and Chory

R E T R A C T E D
A R T I C L E 2010; Liu et al. 2020). To date, retrograde signaling caused by impaired processing of plastid operons has not been reported. The size and organization of the B. rapa plastid genome ( Figure S6) were similar to those of the chloroplast chromosomes of most land plants (Palmer 1985;Shinozaki et al. 1986). Our RNA-seq data showed that a large number of nuclear genes in the ribosome and DNA replication pathways were significantly upregulated in the rne mutant (Fig. 4c), whereas the expression of nuclear genes involved in carbon metabolism and photosynthetic pathways was markedly decreased (Fig. 4d). These results indicate that RNA cleavage triggers chloroplast-to-nucleus retrograde signaling.
In this study, we demonstrated that a functional defect of BrRNE leads to impaired RNA processing, thereby affecting the translation of plastid genes in the chloroplasts (Fig. 3b-c). RNA-seq analysis revealed that the expression of nuclear genes was dramatically altered in the rne mutant, with increased expression levels of genes involved in ribosome composition, DNA replication, and mismatch repair pathways. This result suggests suggesting that the nucleus may compensate for the impaired translation of Fig. 4 RNA-seq analysis of nucleus-encoded genes in wild-type and rne plants at the rosette stage. a-b Gene ontology analysis of genes with increased (a) and reduced (b) expression levels in rne com-pared with the wild-type. c-d KEGG pathway analysis of genes with increased (c) and reduced (d) expression levels in rne compared with the wild-type plastid genes and restore plastid RNA cleavage in the rne mutant (Fig. 3c). On the other hand, as reported by Liu et al. (2020), the suppression of PhANGs in rne may facilitate plant survival by conserving energy through reduced plant growth (Fig. 1a-d, Figure S1). Taken together, our results suggest the following working model of the regulatory mechanism that underlies BrRNE function in wild-type and rne plants (Fig. 5). In In the wild-type, BrRNE cleaves polycistronic precursor transcripts into mature monocistronic mRNAs in chloroplasts. In rne, the defective function of BrRNE results in impaired RNA processing. Thus, aberrant chloroplast RNA cleavage leads to an abnormal thylakoid system and a pale-green phenotype. In addition, responding to retrograde signaling, the transcript levels of photosynthesis-associated nuclear genes (PhANGs) in the nucleus are reduced in rne the wild-type, BrRNE cleaves polycistronic precursor transcripts into mature monocistronic mRNAs in chloroplasts. In rne, the defective function of BrRNE results in impaired RNA processing. Aberrant chloroplast RNA cleavage then leads to reduced levels of photosynthetic proteins, resulting in an abnormal thylakoid system and a pale-green phenotype. In addition, responding to retrograde signaling, the transcript levels of PhANGs in the nucleus are reduced in rne, leading to smaller plants.