A common wild rice-derived BOC1 allele reduces callus browning in indica rice transformation

Abstract

Callus browning, a common trait derived from the indica rice cultivar (Oryza sativa L.), is a challenge to transformation regeneration. Here, we report the map-based cloning of BROWNING OF CALLUS1 (BOC1) using a population derived from crossing Teqing, an elite indica subspecies exhibiting callus browning, and Yuanjiang, a common wild rice accession (Oryza rufipogon Griff.) that is less susceptible to callus browning. We show that BOC1 encodes a SIMILAR TO RADICAL-INDUCED CELL DEATH ONE (SRO) protein. Callus browning can be reduced by appropriate upregulation of BOC1, which consequently improves the genetic transformation efficiency. The presence of a Tourist-like miniature inverted-repeat transposable element (Tourist MITE) specific to wild rice in the promoter of BOC1 increases the expression of BOC1 in callus. BOC1 may decrease cell senescence and death caused by oxidative stress. Our study provides a gene target for improving tissue culturability and genetic transformation.

Introduction

In contrast to japonica rice (Oryza sativa ssp. japonica), Agrobacterium tumefaciens-mediated transformation of indica rice (Oryza sativa ssp. indica) is hindered by callus browning. Callus browning is a common feature of many plant species that presents problems for in vitro culture, a process required for transgenic breeding, resulting in decreased regenerative ability, poor growth, and even death1. The use of antioxidants2,3,4, adsorbing agents5, low salt concentrations6, and growth regulators7 may lessen the effects of callus browning to a certain degree, but because some species do not tolerate these treatments, there is no universal solution to this problem.

The physiological and biochemical mechanisms of callus browning have been extensively studied. For instance, reducing the activity of polyphenol oxidase (PPO), which catalyzes the oxidation of phenols to quinones that in turn polymerize to form brown pigments8, can mitigate the browning of callus9. Browning calli have lower antioxidant enzyme activity than non-browning calli and oxidative stress is associated with callus browning10,11. Endogenous12 and exogenous13 ethylene (ET) cause browning in plant cell cultures. In addition, reduced nitrite reductase activity leads to callus browning14.

Different periods of in vitro rice tissue culture are controlled by many complex genes, respectively15. Several quantitative trait loci (QTLs) for culturability traits in rice have been identified16,17,18,19,20,21,22,23,24. Using positional cloning, a major QTL gene encoding ferredoxin-nitrite reductase (NiR) that determines regeneration ability in rice was identified19. QTLs involved in callus browning have been identified on all 12 rice chromosomes17,21,22,23,24. However, to our knowledge, there have been no reports on the isolation of genes related to callus browning in any plant species. The molecular mechanisms underlying physiological processes inducing browning in plant cell culture still remain unclear.

Genetic analyses of rice culturability have mainly been performed on cultivated rice, and few studies have employed common wild rice (Oryza rufipogon Griff.). In this study, we isolate and characterize a gene responsible for mitigating callus browning in rice, named BROWNING OF CALLUS1 (BOC1). The BOC1 allele is specific to common wild rice. BOC1 is an allele of a rice homolog of SRO (SIMILAR TO RCD ONE), termed OsSRO1c. The appropriate upregulation of BOC1 significantly reduces callus browning during callus proliferation. BOC1 is upregulated in common wild rice callus due to the presence of the Tourist MITE in the promoter region of BOC1. Additionally, BOC1 improves the genetic transformation efficiency and frequency of Hygromycin (Hyg)-resistant calli following gene transfer. BOC1 may regulate oxidative stress and programmed cell death (PCD), which may repress the callus browning. Our findings not only elucidate the mechanisms of callus browning during tissue culture, but also provide important insights into establishing a suitable genetic transformation system for the molecular breeding of rice.

Results

Phenotypic analysis of introgression line YIL25

To identify the gene responsible for callus browning in rice, we constructed a set of introgression lines using a Yuanjiang common wild rice (YJCWR, O. rufipogon) accession that is relatively resistant to browning (Fig. 1b) as a donor and an elite indica cultivar Teqing (O. sativa) that is more susceptible to callus browning (Fig. 1a) as the recipient. We screened the introgression line YIL25 by inoculating mature seeds on unimproved NB (UINB) medium (NB basal medium without additives) for 7 days, and transferred the resulting scutellum-derived calli to UINB medium for 3 weeks of subculture (Fig. 1c). The callus browning rate (CBR) and callus browning index (CBI) were lower in YIL25 than Teqing (Fig. 1j, k). YIL25 contained three YJCWR chromosomal segments on the short arm of chromosomes 2, 3, and 5 (Supplementary Fig. 1). We, therefore, chose YIL25 as a reduced-browning line to cross with Teqing. The F1 plants showed similar phenotypes of CBR and CBI to YIL25 (Supplementary Fig. 2). These results indicated that the reduced-browning trait is controlled by a dominant gene. The histological examination indicated that the cell arrangement of YIL25 calli was more even and the cells were smaller and denser compared to Teqing (Fig. 1d, e). Observations using scanning electron microscopy (SEM) revealed that YIL25 calli consisted of many globular nodules (Fig. 1h) and the cells of YIL25 were turgid (Fig. 1i), whereas Teqing calli had fewer globular nodules (Fig. 1f) and the cells of Teqing were flaccid (Fig. 1g).

Fig. 1: Phenotypic analysis of calli subcultured for 21 days.
figure1

ac Callus browning phenotypes observed in the calli after subculture for 21 days including a indica variety Teqing, b Yuanjiang wild rice (Oryza rufipogon, YJCWR), and c the introgression line YIL25. Scale bars, 2 cm. d, e Paraffin sections of d Teqing and e YIL25. Scale bars, 100 μm. fi Scanning electron microscopy of Teqing and YIL25 calli. f Teqing calli. Thirty calli each from Teqing and YIL25 were collected for analysis of paraffin sections and scanning electron microscopy. Scale bars, 300 μm. g Enlarged from (f). Scale bars, 40 μm. h YIL25 calli. Scale bars, 300 μm. i Enlarged from (h). Scale bars, 40 μm. j, k Comparison of the CBR and CBI values of various calli. Values in (k, l) are means ± SE (n = 3 biologically independent samples). Two-tailed Student’s t-tests were performed to determine significant differences. Source data underlying Figs. j, k are provided as a Source Data file.

Map-based cloning of the callus browning trait gene

To clone the gene for the callus browning trait, we constructed a F2 population of 198 segregating plants derived from a cross between YIL25 and Teqing. We used the CBI as a phenotypic index as it reflects the frequency and degree of callus browning. First, we measured the CBI on the seed-derived callus of the 198 individual plants. We then performed QTL analysis using simple sequence repeat (SSR) markers. The callus browning trait QTL, qCBT3, which is located between SSR markers RM3131 and RM3766 on the short arm of chromosome 3, had the largest effect, explaining 14% of the phenotypic variance associated with decreased CBI (Fig. 2a; Supplementary Table 1).

Fig. 2: Map-based cloning of the callus browning trait gene.
figure2

a qCBT3 was initially localized to a region between markers RM3131 and RM3766 on the short arm of chromosome 3 (n = 198 plants of F2 populations derived from a cross between YIL25 and Teqing). Scale bar, 500 kb. b Positional cloning narrowed the qCBT3 locus to an 18.6 kb region between markers SN11 and SN13 (n = 6377 individuals used for high-resolution mapping). R1 through R7 are recombinants. The black, white, and gray regions indicate homozygous for the YIL25 genome, homozygous for the Teqing genome, and the interval in the chromosome where crossover took place, respectively. Scale bar, 10 kb. LOC_Os03g12820 is the candidate gene BROWNING OF CALLUS1 (BOC1). Data are means (n = 3 biologically independent samples), with error bars showing standard error. c Structure of the candidate gene BOC1. White box, white box with arrow, black boxes, and lines between boxes indicate the 5′-UTR, 3′-UTR, exons, and introns, respectively. d Mutation sites of the BOC1 promoter in YIL25 and Teqing. The ellipsis indicates deletion of bases. Scale bar, 50 bp. Source data underlying Figs. a, b are provided as a Source Data file.

For high-resolution mapping, we selected F2 individuals that were heterozygous at the qCBT3 locus (RM3131-RM3766), containing one chromosomal segment from Teqing and one from YJCWR by self-crossing to construct a segregating population to eliminate gene interference on introgression segments of chromosomes 2 and 5 (Fig. 2b). We identified plants with recombination events between markers RM3131 and RM3766 by genotyping the 6377 individuals in the population and evaluated the CBI of these recombinant individuals (Fig. 2b). We then developed single nucleotide polymorphism (SNP) markers and delimited qCBT3 to an 18.6 kb region between SNP markers SN11 and SN13. This region contained only one predicted gene (LOC_Os03g12820) in the Nipponbare reference genome (The Rice Genome Annotation Project Database, Fig. 2c). Sequence analysis of the coding region in this interval revealed no sequence variation between Teqing and YIL25. Next, we compared the promoter sequences upstream of the translation start site between Teqing and YIL25 and identified one 337 bp deletion, one 1 bp deletion, and three SNPs in Teqing (Fig. 2d; Supplementary Fig. 3). Therefore, we chose LOC_Os03g12820 as the candidate gene for BROWNING OF CALLUS1 (BOC1).

Functional analysis of BOC1

To elucidate the function of LOC_Os03g12820, we constructed an RNA interference construct targeting LOC_Os03g12820 and transformed it into introgression line YIL25 (designated as pRi-BOC1-YIL25). Eighteen transgenic lines were obtained. Reverse transcription-quantitative PCR (RT-qPCR) analysis showed that LOC_Os03g12820 was expressed at lower levels in the pRi-BOC1-YIL25 positive transgenic lines than in the negative (non-transformed) plants (Fig. 3e). In addition, RNAi suppression of LOC_Os03g12820 in the pRi-BOC1-YIL25 positive transgenic lines to the level of Teqing caused serious callus browning compared to the non-transformed controls (Fig. 3a–e).

Fig. 3: Functional analysis of BOC1.
figure3

ad Phenotypes of calli from pRi-BOC1-YIL25 transgenic plants (n = 18). Negative represents segregating plants from the transformation process but without the transgene. Scale bars, 1 cm. e Comparison of the relative expression levels and CBI between a negative transgenic line and 10 RNAi lines (n = 3 biological replicates). fi Phenotypes of calli from pCPL-BOC1-Teqing transgenic plants (n = 15). Scale bars, 1 cm. j Comparison of the relative expression levels and CBI in a negative transgenic line and 10 complementation lines (n = 3 biological replicates). kn Phenotypes of calli from pOE-BOC1-Teqing transgenic plants (n = 22). o Comparison of the relative expression levels and CBI in a negative transgenic line and 10 overexpression lines (n = 3 biological replicates). The black line indicates the trend of the relative expression levels, and the gray line indicates the trend of the CBI. Scale bars, 1 cm. Data are means ± SE, two-tailed Student’s t-tests. Source data underlying Figs. e, j, o are provided as a Source Data file.

Furthermore, we introduced a complementation vector containing the promoter, open reading frame and 3′-UTR of LOC_Os03g12820 from YIL25 into Teqing (designated as pCPL-BOC1-Teqing) and obtained 15 positive transgenic plants. RT-qPCR analysis showed that the relative expression levels of LOC_Os03g12820 were significantly elevated in the pCPL-BOC1-Teqing positive plants compared to the negative plants (Teqing) (Fig. 3j). The CBI was reduced in these positive lines when the expression level of LOC_Os03g12820 was elevated at least 1.8 fold compared with the negative plants (Fig. 3f–j). This finding indicates that LOC_Os03g12820 mitigates the callus browning phenotype. These results suggest that LOC_Oso3g12820 is BOC1 regulates callus browning in rice.

Next, we generated transgenic Teqing plants overexpressing the LOC_Os03g12820 coding sequence of YIL25 (as shown in Supplementary Fig. 3, the rose red and red underline represent the nucleotide sequence of the overexpression construct, designated as pOE-BOC1-Teqing) and obtained 22 positive transgenic lines. RT-qPCR indicated that BOC1 was strongly upregulated in the pOE-BOC1-Teqing positive transgenic lines compared to the negative plants (Fig. 3o). Nevertheless, when BOC1 expression levels were 8.8-times higher in the tested pOE-BOC1-Teqing positive lines compared to the negative plants, callus browning was more severe (Fig. 3k–o). This result indicates that only the appropriate expression of BOC1 reduces callus browning.

BOC1 encodes an SRO protein of 463 amino acids, which includes a putative N-terminal RNA Recognition Motif (RRM) domain, a poly (ADP-ribose) polymerase (PARP)-like ADP-ribose transferase catalytic domain, and a C-terminal RST (RCD1, [RADICAL-INDUCED CELL DEATH1]—SRO [SIMILAR TO RCD ONE]—TAF4 [TBP-ASSOCIATED FACTOR4] (RST) domain (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi; Supplementary Fig. 4a). To evaluate the conservation of BOC1 in other plant species, we performed protein sequence alignment using the full-length protein sequence of BOC1 as a query in BLAST analysis and selected 24 putative homologs (sharing > 80% sequence identity with BOC1 at the nucleotide level) to construct a phylogenetic tree. Phylogenetic analysis indicated that BOC1 is closely related to other monocots, such as Oryza brachyantha, Brachypodium distachyon, Triticum urartu, and Aegilops tauschii (Supplementary Fig. 4b).

Transcriptional characterization of BOC1

To elucidate the expression profiles of BOC1, we analyzed the expression pattern of BOC1 in different organs by RT-qPCR. The expression level of BOC1 was highest in YIL25 calli after 21 days of subculture, followed by the tiller base, roots, culms, and leaves, and it was lowest in young panicles (Supplementary Fig. 5). BOC1 expression increased gradually in Teqing and YIL25 throughout tissue culture. BOC1 expression did not significantly differ between Teqing and YIL25 during the early stage of tissue culture, but significant differences were observed during the later stage; BOC1 expression was notably higher in YIL25 than in Teqing in callus after 21 days of subculture (Fig. 4a). We further examined the expression of BOC1 in calli after 21 days of subculture by RNA in situ hybridization and observed that the hybridization signals were stronger in YIL25 than in Teqing, confirming the notion that BOC1 is expressed at higher levels in YIL25 callus (Fig. 4b).

Fig. 4: Transcriptional characterization of BOC1.
figure4

a Expression analysis of BOC1 in calli at different stages of subculture. Data are means, with bars showing SE, n = 3 biological replicates. Two-tailed Student’s t-tests were performed to determine significant differences. b BOC1 expression patterns in Teqing and YIL25 calli subcultured for 21 days, as revealed by mRNA in situ hybridization. Thirty calli each from Teqing and YIL25 were collected for mRNA in situ hybridization. Scale bars, 100 μm. c Transient expression assays of the effects of InDels and SNPs in the BOC1 promoter. Top, structure of the pGreenII 0800-LUC vector. Left, constructs with site-directed mutagenesis of the BOC1 promoter. Luc, the target promoter controlling the firefly luciferase reporter gene. The Renilla (Ren) luciferase reporter gene was used as an internal control. M1 to M5 indicate the constructs with site-directed mutations at the 337 bp InDel, three SNPs, and 1 bp InDel. Right, expression levels of Luc relative to Ren. Data are presented as mean ± SE. Horizontal bars show the SE for all constructs (n = 10 biological replicates, two-tailed Student’s t-tests). d Subcellular localization of BOC1 protein in rice protoplast cells. OsMADS15 is a nuclear localization marker. Scale bars, 20 μm. P-values of (a, c) were calculated with two-tailed Student’s t-tests. The subcellular localization assays were performed with three biologically independent experiments. Source data underlying Figs. a, c are provided as a Source Data file.

Sequencing analysis revealed one 337 bp deletion, one 1 bp deletion, and three SNPs in the promoter region of BOC1 between Teqing and YIL25 (Fig. 4c). To determine whether these deletions and SNPs in the BOC1 promoter region directly contribute to its expression, we generated various constructs by inserting different combinations of mutant promoter fragments of YIL25 into binary vector pGreenII 0800-LUC. We introduced these constructs into rice protoplasts produced from callus that had been cultured for 21 days and examined the expression of the luciferase reporter gene (LUC). As shown in Fig. 4c, no significant differences in activity were observed between the YIL25 promoter fragment and mutated promoter fragments containing three SNPs (constructs M2–M4) and one 1 bp deletion (construct M5). However, the 337 bp deletion (construct M1) significantly decreased the promoter activity of BOC1 to a level similar to that of Teqing (Fig. 4c). These results suggest that the 337 bp deletion is responsible for the reduced expression level of BOC1 in Teqing calli. Using the RepeatMasker Web Server (http://www.repeatmasker.org/cgi-bin/WEBRepeatMasker), we examined the rice transposable element database using the 337 bp sequence as a query and determined that it is the Tourist-like miniature inverted-repeat transposable element (Tourist MITE), which belongs to the PIF/Harbinger superfamily. In addition, using PLACE (Plant Cis-acting Regulatory DNA Elements; http://www.dna.affrc.go.jp/htdocs/PLACE/), we identified three DOFCOREZM (AAAG, DOF TFBS)25, one MYB1AT (TAACCA, MYB TFBS)26, one MYCCONSENSUSAT (CAGATG, MYC TFBS)26,27,28, and one ACGTATERD1 (ACGT, ACGT TFBS)28 in this sequence; these elements are associated with plant stress responses (Supplementary Fig. 6). Subcellular localization analysis indicated that the BOC1-GFP fusion protein was specifically targeted to the nucleus in rice protoplasts (Fig. 4d).

BOC1 influences genetic transformation efficiency

To determine if BOC1 enhances the transformation efficiency of rice due to decreased callus browning, we transformed the binary vector pCAMBIA1300 into YIL25 (YIL25 1300), Teqing (Teqing 1300), and Teqing 1300 as a control. The pCAMBIA1300-BOC1 complementation construct from YIL25 was transformed into Teqing (Teqing pCPL) by Agrobacterium tumefaciens-mediated transformation and cultured on UINB medium for hygromycin (Hyg) selection. The selection frequency of Hyg-resistant calli was more than 2.7-fold higher in YIL25 1300 than in Teqing 1300 (Fig. 5a, b, d), and the transformation efficiency was 2.5-fold higher in YIL25 1300 than in the Teqing 1300 control (Fig. 5e), indicating that YIL25 containing BOC1 is an excellent receptor material for genetic transformation. Meanwhile, the selection frequency of Hyg-resistant calli was 2.1-fold better (Fig. 5a, c, d) and the transformation efficiency was 1.9-fold better in Teqing pCPL compared to the Teqing 1300 control (Fig. 5e). These results indicate that BOC1 improves the selection frequency of Hyg-resistant calli and the genetic transformation efficiency of rice by reducing callus browning during Agrobacterium-mediated transformation.

Fig. 5: BOC1 influences genetic transformation efficiency.
figure5

ac Transformed calli after three rounds of selection on Hyg-resistant UINB medium. Scale bars, 1 cm. d Comparison of the selection frequency of Hyg-resistant callus. Each dot represents the selective frequency of Hyg-resistant callus of 10 independent experiments. (e) Comparison of the transformation efficiency. Each diamond represents the transformation efficiency of 10 independent experiments. Teqing 1300, Teqing calli were infected with Agrobacterium harboring the binary pCAMBIA1300 empty vector as a control. YIL25 1300, YIL25 calli transformed with the binary vector pCAMBIA1300. Teqing pCPL, Teqing calli infected with Agrobacterium harboring the pCAMBIA1300-BOC1 construct. The hollow box in each column in (d, e) represents the means of the selective frequency of Hyg-resistant callus and the transformation efficiency of 10 replicates, respectively. Box edges represent the 0.25 and 0.75 quantiles, with the median values shown by red, blue and yellow lines, respectively. Whiskers extend to data no more than 1.5 times the interquartile range, and the remaining data are indicated by dots. P-values of (d, e) were determined by two-tailed Student’s t-tests. Source data are provided as a Source Data file.

Physiological indices related to oxidative stress

To further explore the molecular mechanisms of callus browning, we measured various physiological and biochemical indices in calli at 21 days of subculture. Compared to Teqing, the H2O2 contents were lower in both YIL25 and pCPL calli and higher in pOE (pOE-BOC1-Teqing) calli (Fig. 6a; Supplementary Fig. 7a). In addition, we used Noninvasive Microtest Technology (NMT) on rice callus to detect intercellular H2O2 fluxes. There was a significant efflux of intracellular H2O2 in YIL25 and pCPL calli, whereas Teqing calli exhibited a significant influx of extracellular H2O2, suggesting that H2O2 was expelled from YIL25 and pCPL cells to alleviate celluar damage (Fig. 6b). MDA (malondialdehyde) is a biomarker of lipid peroxidation and an indicator of reactive oxygen species (ROS) production. Compared to Teqing, MDA contents were notably lower in the YIL25 and pCPL lines but higher contents in the pOE lines (Fig. 6c; Supplementary Fig. 7b), reflecting less damage from H2O2 in YIL25 and pCPL plants. Moreover, catalase (CAT), glutathione reductase (GR), and glutathione S-transferase (GST) activity and antioxidant glutathione (GSH) contents were higher in YIL25 and pCPL vs Teqing callus (Fig. 6d–g). However, GST activity and GSH contents were lower in pOE lines than in Teqing callus although CAT and GR activity did not significantly differ among lines (Supplementary Fig. 7c-f).

Fig. 6: Physiological and biochemical oxidative stress indices of calli subcultured for 21 days.
figure6

a H2O2 contents in the two parents and two positive complementation transgenic plants. b H2O2 flux (n = 4). c MDA contents. df Assay of antioxidant enzyme activities. CAT, catalase. GR, glutathione reductase. GST, glutathione S-transferase. g Antioxidant GSH contents. hj Comparison of phytohormone contents between Teqing and YIL25. SA, salicylic acid; ET, ethylene; ACC, 1-aminocyclopropane-1-carboxylic acid, a direct precursor of ET. All values in (a) and (c) to (j) are presented as means ± SE (n = 3 biologically independent samples). P-values were analyzed by two-tailed Student’s t-tests. Source data are provided as a Source Data file.

Finally, the levels of the plant hormones salicylic acid (SA) and ethylene (ET), as well as the direct ET precursor, 1-aminocyclopropane-1-carboxylic acid (ACC), were significantly lower in YIL25 compared to Teqing (Fig. 6h–j). Collectively, these findings suggest that BOC1 may inhibit PCD caused by oxidative stress, thereby reducing callus browning.

RNA-seq analysis of BOC1

To analyze the molecular function of BOC1, we performed RNA-seq experiments using callus that had been subcultured for 21 days from Teqing and YIL25, Teqing and the complementation line pCPL-BOC1-Teqing (pCPL), and also Teqing and the overexpression line pOE-BOC1-Teqing (pOE). We identified 700 differentially expressed genes (DEGs), including 410 DEGs that were upregulated in YIL25 versus Teqing callus and pCPL versus Teqing callus and downregulated in pOE versus Teqing callus (UUD, Fig. 7a, Supplementary Data 1). The 290 remaining DEGs were downregulated in YIL25 versus Teqing callus and pCPL versus Teqing callus and upregulated in pOE versus Teqing callus (DDU, fold change ≥ 2, FDR < 0.001) (Fig. 7b, Supplementary Data 2).

Fig. 7: Differentially expressed genes in YIL25, the complementation line, overexpression line, and Teqing revealed by RNA-seq analysis.
figure7

a Hierarchical clustering of 410 differentially expressed genes (DEGs) upregulated in YIL25 compared to Teqing, DEGs upregulated in the complementation line (pCPL) compared to Teqing, and DEGs downregulated in the overexpression line (pOE) compared to Teqing (UUD). b Hierarchical clustering of 290 DEGs downregulated in YIL25 compared to Teqing, DEGs downregulated in pCPL compared to Teqing, and DEGs upregulated in pOE compared to Teqing (DDU). c Hierarchical clustering of DEGs associated with stress response and cell death. The numbers in the boxes represent the number of genes per cluster.

Among the UUD genes, 98 genes were involved in stress responses and cell death (Fig. 7c). Notably, these included 26 pentatricopeptide repeat protein genes, whose dysfunction leads to ROS accumulation;29,30 these genes were upregulated in YIL25 and pCPL but downregulated in pOE (Fig. 7c; Supplementary Fig. 8a). GST conjugates with GSH to scavenge ROS in order to protect cellular components from damage. In the current study, 3 GST genes present in the UUD DEG set and absent in the DDU DEG set (Fig. 7c; Supplementary Fig. 8c). In addition, compared to Teqing, GST activity and GSH contents were higher in YIL25 and pCPL callus but lower in pOE callus (Fig. 6f, g; Supplementary Fig. 7d,f). Among the UUD genes were four aldehyde dehydrogenase genes; aldehyde dehydrogenase enhances tolerance to various abiotic stresses by scavenging cytotoxic aldehydes (Supplementary Fig. 8d), and MDA contents were higher in of Teqing and pOE callus compared to YIL25 and pCPL callus (Fig. 6c; Supplementary Fig. 7b). LOC_Os01g67980, which negatively regulates the senescence process, encodes senescence-associated protein OsSAG12-1. RNAi transgenic lines declined the expression of OsSAG12-1 and developed early senescence31 (Fig. 7c; Supplementary Fig. 8f). In addition, the UUD genes included transcription factor genes associated with stress and defense responses, which might function in maintaining ROS homeostasis (Fig. 7c; Supplementary Fig. 8b, g).

Twelve DDU genes were associated with oxygenase and oxidase; cytochrome P450 has monooxygenase activity as a source of ROS32 (Supplementary Fig. 8h). The DDU group also included genes encoding aspartic proteinases, peptidases, and nuclease (involved in cell degradation), as well as genes encoding polygalacturonase, endoglucanase, and glycosyl hydrolases (involved in cell wall degradation) (Fig. 7c; Supplementary Fig. 8i). Emerging evidence suggests that aspartic proteases are involved in plant development and cell death33. Finally, five DDU DEGs encode positive regulators of the senescence process (Fig. 7c). These DEGs include LOC_Os01g53260, encoding WRKY transcription factor OsWRKY23, which positively regulates the senescence process34, and LOC_Os12g36850, encoding a pathogenesis-related bet v I family protein similar to probenazole-inducible protein PBZ1, which induces cell death in rice, tobacco, and Arabidopsis via its RNase activity inside the cell35 (Supplementary Fig. 8j). When we treated callus that was subcultured for 7 days with 0.5% H2O2 for 18 h, PBZ (LOC_Os12g36850) was upregulated compared to the untreated control (Supplementary Fig. 8k).

Taken together, these findings suggest that the appropriate expression of BOC1 may be beneficial for rice cells to be in reducing state, prevented various stress, and then inhibited cell death.

BOC1 promoter sequence variation affects callus browning

To investigate the relationships between mutations in the BOC1 locus and callus browning, we sequenced a ~4.6 kb genomic fragment covering the entire BOC1 gene (2740 bp), the 1452 bp 5′-flanking region, and the 380 bp 3′-flanking region, from 50 O. rufipogon accessions (Supplementary Data 3) and 74 Asian rice cultivars (28 japonica and 46 indica cultivars, Supplementary Data 4). Nucleotide alignment showed that only 19 O. rufipogon accessions contained the Tourist MITE (Supplementary Data 3); however, the element at this locus was absent in all Asian rice cultivars examined (Supplementary Data 4). Therefore, we performed an association test of callus browning after 21 d of subculture vs. sequence variations in the O. rufipogon accessions and Asian rice cultivars. The association test of the common wild rice revealed the strongest association signals at the Tourist MITE and the four SNPs (Fig. 8a). Furthermore, we investigated the CBI of the 50 accessions of common wild rice and determined that this value was lower in the 19 accessions with the Tourist MITE than in the 31 accessions without this insertion (Fig. 8b). Moreover, the BOC1 gene was expressed at higher levels in the 19 accessions with the Tourist MITE than in the 31 accessions without the Tourist MITE (Fig. 8c; Supplementary Data 3). In addition, we obtained previously reported the genome resequencing data for 446 Oryza rufipogon accessions36 and recovered the reads that mapped to the Tourist MITE flanking sequence in these accessions. We detected this site in 226 accessions (48 accessions with this MITE insertion and 178 accessions without this insertion). Forty-one of the 48 accessions with this insertion were in group III (closer to japonica) and 7 were in group II (the intermedia group between indica and japonica) (Supplementary Fig. 9a). By contrast, the 178 accessions without this MITE insertion were distributed among all the three groups including 65 in group I (closer to indica), 49 in group II, and 64 in group III (Supplementary Fig. 9a).

Fig. 8: Association test and expression levels of loci with natural variations in BOC1.
figure8

a Association testing of callus browning in 50 O. rufipogon accessions vs. 16 variant sites in the 4.6 kb genomic region. Black dots represent 16 variations. b Comparison of CBI in calli subcultured for 21 d in 31 accessions with the 337 bp deletion vs. 19 accessions with the 337 bp insertion. c Comparison of BOC1 expression in calli subcultured for 21 d in lines harboring the 337 bp deletion vs. the 337 bp insertion (n = 3 biological replicates). d Association testing of callus browning in 74 Asian rice cultivars with 36 variant sites in the 4.6 kb genomic region. e Comparison of CBI in calli subcultured for 21 d between 59 accessions with the T deletion and 15 accessions with the T insertion. f Comparison of BOC1 expression in callus subcultured for 21 d from rice varieties harboring the T deletion vs. the T insertion (n = 3 biological replicates). Box edges of (b, c) and (e, f) represent the 0.25 quantile and 0.75 quantile with the median values shown by red and blue lines. Whiskers extend to data no more than 1.5 times the interquartile range, and remaining data are indicated by dots. P-values of (b, c) and (e, f) were calculated by two-tailed Student’s t-tests. Source data underlying Figs. b, c, e, f are provided as a Source Data file.

Nucleotide alignment of the 74 Asian rice cultivars revealed 9 Indels and 27 SNPs (Fig. 8d; Supplementary Data 4). To ascertain whether BOC1 was responsible for callus browning in Asian rice cultivars, firstly, an association analysis was carried out between callus browning and sequence variations of 74 Asian rice cultivars. The result indicated that the Indel8 (−433, T indel) in the promoter of BOC1 was the strongest signals (Fig. 8d). However, this site was no polymorphic in the 50 accessions of O. rufipogon (Supplementary Data 3, 4). We further found that the CBI in 15 varieties with T insertion in the promoter of BOC1 was lower than that in 59 varieties without this insertion (Fig. 8e; Supplementary Data 4). We also analyzed the expression levels in these varieties and the result showed that the expression levels of BOC1 was higher in the 15 varieties with T insertion in the promoter of BOC1 than that in 59 varieties without this insertion (Fig. 8f; Supplementary Data 4). The above results suggest that the sequence variation in the BOC1 promoter influence callus browning by affecting the expression level of BOC1.

Finally, we analyzed the fixation index (FST, the level of population differentiation), on chromosome 3 in 1083 cultivated indica and japonica rice varieties36. The FST level in BOC1 was 0.21, indicating that no significant differences in BOC1 exist between the indica and japonica subspecies (Supplementary Fig. 9b). Therefore, we propose that the effect of BOC1 on callus browning is not related to the differences between the indica and japonica subspecies.

Discussion

Here, we demonstrated that the proper expression of BOC1 decreases callus browning and increases the transformation efficiency of rice. BOC1 encodes an SRO (SIMILAR TO RCD ONE) protein that is allelic to OsSRO1c, a regulator of oxidative stress responses37. RCD1 plays an important role in regulating cell death by controlling oxidative stress responses in Arabidopsis responses38,39,40,41,42,43. OsSRO1c is the major stress-responsive gene in the SRO family. Both ossro1c-1 mutants and amiRNA (artificial microRNA)-OsSRO1c plants are hypersensitive to drought stress at the seedling stage. In addition, OsSRO1c-overexpressing rice plants showed more severe damage under drought stress than wild-type plants, whereas plants overexpressing SNAC1, encoding a NAC transcription factor that directly regulates OsSRO1c expression, showed increased drought resistance29. In the current study, RNA interference and BOC1 overexpression plants showed a callus-browning phenotype similar to or more severe than that of Teqing, whereas lines harboring the complementation vector exhibited reduced callus browning, indicating that the appropriate upregulation BOC1 reduces callus browning.

Transposable elements, among the most variable components of the genome, can replicate and integrate into positions, which affects the expression levels of adjacent genes, thereby resulting in genetic diversity44. For example, the terminal-repeat retrotransposon in miniature (TRIM) elements insertion in the promoter of MS2 allele in the male-sterile ms2 mutant resulted in an anther-specific expression pattern in wheat45. In maize, an 82 bp miniature inverted-repeat transposable element (MITE) into the promoter of ZmNAC111 represses its expression; this MITE is significantly associated with natural variation46. In the current study, the presence of Tourist MITE in the BOC1 promoter significantly increased BOC1 expression in rice callus.

Callus cultured in vitro is subjected to complex growth conditions compared to cells in plants. Callus cells are exposed to the oxygen in air, which tends to result in oxidative stress. Oxidative browning is a prevalent problem in plant tissue culture that can cause cell death, resulting in brown tissues47. A moderate increase in the levels of the ROS H2O2 helps protect plant cells from subsequent exposure to more severe abiotic or oxidative stress48. However, the excessive accumulation of H2O2 might be involved in regulating PCD38,39,40,49,50,51. Moreover, high levels of H2O2 induce the accumulation of SA, resulting in severe oxidative damage39,52; in turn, SA pretreatment mediates the accumulation of H2O253. Endogenous ET in in vitro culture may inhibit the synthesis of protective enzymes, such as CAT, for scavenging H2O2, which is involved in senescence, browning, and necrosis of callus54. In the current study, YIL25 callus contained decreased levels of H2O2 and the phytohormones ET and SA, as well as increased levels of CAT, GR, and GST activity compared to Teqing. The RNA-seq data analysis revealed that BOC1 influences the expression of genes associated with ROS production and scavenging, stress response and cell death. In recently report, RCD1 coordinates chloroplast and mitochondrial functions through interaction with transcription factors. RCD1 participates in regulation of mitochondrial respiration and chloroplast ROS processing and may contribute to the survival of plant under a changing environment55. However, as shown in Supplementary Fig. 4b, RCD1 and OsSRO1c have different evolutionary histories which may have contrasting functions in the regulation of ROS related processes. In the current study, we use the calli subcultured for 21 d in the dark and is not necessary for photosynthesis by origin of carbon from medium. Therefore, the interaction and phenotype of BOC1 may be different from seedling stage, which will be an interesting research topic.

Common wild rice, the wild progenitor of cultivated rice, has an important gene pool that includes some alleles no longer present in cultivated rice56. The insertion of the 337 bp Tourist MITE in the BOC1 promoter increases BOC1 expression. This insertion is only detected in certain accessions of wild rice, but not in cultivated rice. Therefore, wild rice could be used to provide alleles for improving the tissue culture characteristics and genetic transformation efficiency of cultivated rice. Moreover, the genetic manipulation of the BOC1 promoter in cultivated rice might also improve tissue culture ability. Perhaps the BOC1 allele of wild rice and even similar alleles in other crops of the Gramineae family could be used to create recipient varieties more amenable to genetic transformation. The use of BOC1 in in vitro rice callus propagation might help improve genetic transformation efficiency and selection to facilitate the molecular breeding of rice.

Methods

Plant materials

The wild rice introgression line YIL25, with reduced levels of callus browning, was derived from a cross between an O. rufipogon accession (YJCWR) as the donor and the elite indica variety Teqing (O. sativa) as the recipient57. The mapping population was developed from a cross between introgression line YIL25 and Teqing. The 50 common wild rice accessions and 74 O. sativa varieties (28 japonica and 46 indica varieties) used in this study are listed in Supplementary Data 3 and Supplementary Data 4.

Tissue culture procedure

Mature, healthy, dehusked seeds were sterilized by immersion in 70% ethanol for ~2 min, followed by 15% sodium hypochlorite solution for 15 min with shaking, and rinsed three or four times with sterile water on an ultraclean workbench. Unimproved NB (UINB) medium (NB medium containing N6 macronutrient components58, B5 micronutrient components, and organic components59, supplemented with 2 mg/L of 2,4-D and 30 g/L of sucrose) was used for callus induction and subculture. The pH of the medium was adjusted to 5.8 with 1 N KOH, and 3 g/L of Phytagel was added before autoclaving at 121 °C for 20 min. Approximately 90 mature seeds per line, evenly distributed among three Petri dishes, were incubated in induction medium for 7 days at 28 °C in the dark. The induced calli were transferred to subculture (UINB) medium and incubated in the dark at 28 °C for 3 weeks. The traits of the calli in each Petri dish were recorded.

Phenotype observations

The tendency for callus browning was categorized into four levels: (0) less than 1/10 of the callus tissue was brown (recorded as no browning); (1) 1/10 ~ 1/3 of the callus tissue was brown, considered light browning; (2) 1/3 ~ 2/3 of the callus tissue was brown, considered medium browning; (3) 2/3 ~ 1 of the callus tissue was brown, considered deep browning; and (4) the callus was completely brown. The tendency for callus browning, selection frequency of Hyg-resistant calli and transformation efficiency were calculated according to the following Eqs. (1, 2, 3, and 4):

$${\mathrm{Callus}}\;{\mathrm{browning}}\;{\mathrm{rate}}\;\left( {{\mathrm{CBR}}} \right) =\ \left( {{\mathrm{number}}\;{\mathrm{of}}\;{\mathrm{calli}}\;{\mathrm{showing}}\;{\mathrm{browning}}}\right./\\ \left.{{\mathrm{number}}\;{\mathrm{of}}\;{\mathrm{transferred}}\;{\mathrm{calli}}} \right) \times 100\%$$
(1)
$${\mathrm{Callus}}\;{\mathrm{browning}}\;{\mathrm{index}}\;\left( {{\mathrm{CBI}}} \right) =\ \left( {\Sigma \;{\mathrm{number}}\;{\mathrm{of}}\;{\mathrm{calli}}\;{\mathrm{at}}\;{\mathrm{each}}\;{\mathrm{browning}}\;{\mathrm{level}} \times {\mathrm{browning}}\;{\mathrm{level}}} \right)/ \\ \left ( {{\mathrm{number}}\;{\mathrm{of}}\;{\mathrm{transferred}}\;{\mathrm{calli}} \times {\mathrm{highest}}\;{\mathrm{browning}}\;{\mathrm{level}}} \right) \times 100\%$$
(2)
$${\mathrm{Selection}}\;{\mathrm{frequency}}\;{\mathrm{of}}\;Hyg \hbox{-} {\mathrm{resistant}}\;{\mathrm{calli}} =\ \left( {{\mathrm{number}}\;{\mathrm{of}}\;Hyg {\hbox{-}} {\mathrm{resistant}}\;{\mathrm{calli}}}\right./ \\ \left. {{\mathrm{number}}\;{\mathrm{of}}\;{\mathrm{infected}}\;{\mathrm{calli}}} \right) \times 100\%$$
(3)
$${\mathrm{Transformation}}\;{\mathrm{efficiency}} =\ \left( {{\mathrm{number}}\;{\mathrm{of}}\;{\mathrm{independent}}\;{\mathrm{transgenic}}\;{\mathrm{events}}\;{\mathrm{regenerating}}\;{\mathrm{plants}}}\right./\\ \left. {{\mathrm{number}}\;{\mathrm{of}}\;{\mathrm{transferred}}\;{\mathrm{calli}}} \right) \times 100\%$$
(4)

Scanning electron microscopy

To prepare histological sections, calli that had been subcultured for 21 days were fixed in 2.5% glutaraldehyde-phosphate buffer saline fixative solution (pH 7.2) for more than 2 h and postfixed in the same buffer containing 1% OsO4. After being dehydrated through an ethanol series and dried using a carbon dioxide critical-point dryer, the calli were cleaned with ethanol and dried at 45 °C. The dry calli were gold plated and photographed under a Hitachi S3400 scanning electron microscope (Japan).

Primers

The primers used in this study are listed in Supplementary Data 5.

Genetic confirmation

The RNA interference construct was generated by inserting a hairpin sequence with a 350 bp and a 346 bp cDNA inverted repeat fragment of BOC1 into the pTCK303 vector driven by the maize Ubiquitin promoter and transformed into YIL25 (pRi-BOC1-YIL25). A 4572 bp genomic fragment from YIL25 harboring the entire BOC1 gene sequence with the 1452 bp 5′-flanking region and the 380 bp 3′-flanking region was amplified and inserted into binary vector pCAMBIA1300 (http://www.cambia.org) to form the complementary construct (pCPL). The overexpression vector (pOE) contained the 1392 bp YIL25-BOC1-ORF, which was amplified and inserted into the pCAMBIA1301 vector (http://www.cambia.org) driven by the Ubiquitin promoter. Both the pCPL and pOE constructs were transferred into Teqing and designated as pCPL-BOC1-Teqing and pOE-BOC1-Teqing, respectively. Relevant primer sequences are listed in Supplementary Data 5.

RT-qPCR

Total RNA was extracted from various samples using TRIzol reagent (Life Technologies) and purified using an RNeasy Mini Kit (Qiagen) following the manufacturer’s instructions. First-strand cDNA was synthesized using Oligo(dT)15 primers (TaKaRa) and SuperScript III Reverse Transcriptase (Invitrogen) with 3 μg of total RNA. The qPCR was performed using a CFX96 Real-Time System (Bio-Rad), with each reaction containing 5 ng of first-strand cDNAs, 4 μM of gene-specific primers, and 5 μL of real-time PCR SYBR MIX (iQ™ SYBR® Green Supermix, Bio-Rad) under the following conditions: initial denaturation at 95 °C for 3 min followed by 40 cycles of 95 °C for 30 s, 58 °C for 30 s, and 72 °C for 30 s. The rice housekeeping gene Actin (LOC_Os03g50885) was used as an internal control to normalize the gene expression data using the relative quantification method (2–ΔΔCT)60. Each set of experiments was repeated three times. The primers used for RT-qPCR and cDNA amplification are listed in Supplementary Data 5.

mRNA in situ hybridization

mRNA in situ hybridization was performed with the method of Javelle et al.61 with minor modifications. YIL25 and Teqing calli that had been subcultured for 21 days were fixed in 3.7% FAA solution at 4 °C overnight, dehydrated, and embedded in paraffin (Paraplast Plus, Fisher Scientific). The calli were sliced into 10-μm sections with a microtome (Leica RM2145). A 344 bp fragment of BOC1 cDNA was amplified and used as the template to generate sense and antisense RNA probes, which were labeled using a DIG RNA labeling Kit (Roche) according to the manufacturer’s instructions. The hybridized slides were observed under a microscope (Leica DMR). Relevant primer sequences are listed in Supplementary Data 5.

Luciferase assays

The promoter sequences upstream of the BOC1 translation start site of YIL25 and Teqing were amplified using specific primers designed with site-directed mutations at the InDels and SNPs. The fragments were inserted into the pGreenII 0800-LUC vector to drive the expression of the firefly luciferase (LUC) gene. The effector plasmid DNA was transformed into rice protoplasts prepared from rice calli that had been subcultured for 21 days in the dark using the polyethylene glycol-mediated method62. The enzyme solution contained the following: 1.5% (w/v) “Onozuka” R-10 cellulose (Yakult Pharmaceutical Industry), 0.75% (w/v) macerozyme R-10 (Yakult Pharmaceutical Industry), 0.5 M D-mannitol, 2 mM 2-(N-morpholino) ethanesulfonic acid (MES; Sigma-Aldrich, cat. no. M3671), and 0.1 × W5 medium, pH 5.7. The protoplasts were lysed with Passive Lysis Buffer (PLB, Promega) and assayed using the Dual-Luciferase Reporter Assay System (Promega). Relevant primer sequences are listed in Supplementary Data 5.

Subcellular localization

The entire coding sequence of BOC1 was cloned into a GFP vector under the control of the 35S promoter to generate p35S::BOC1-GFP. The RFP coding sequence was fused to the C terminus of OsMADS15 to generate OsMADS15-RFP and was used as a nuclear marker. The two constructs were transformed into protoplasts and incubated at 28 °C in the dark for 16 h. Fluorescence was examined under a confocal laser scanning microscope.

Agrobacterium-mediated genetic transformation

The binary plasmid vectors described above were transferred into Agrobacterium tumefaciens strain EHA105. Sterilized mature seeds were inoculated on UINB medium and cultured at 28 °C in the dark. After 7 days of incubating, the calli were collected and used for transformation. After 3 days of cocultivation with Agrobacterium at 25 °C in constant darkness on NB-Acetosyringone (AS) medium (30 g/L sucrose, 10 g/L glucose, 0.3 g/L casamino acids, NB medium salt mixture [Duchefa Biochemie], 0.5 mg/L glutamine, and 2.5 mg/L 2,4-D, 20 mg/L AS, pH 5.3), the calli were washed two or three times with sterile water and rinsed once with sterile water containing 200 mg/L Timentin (Caisson Labs, USA) and 200 mg/L cephalosporin (Wako Pure Chemical Industries). The calli were cultured on UINB medium containing 50 mg/L hygromycin B (Hyg, Merck), 200 mg/L Timentin, and 200 mg/L cephalosporin for three rounds of 15 days each at 28 °C in the dark. For regeneration, vigorously growing calli were transferred to MS regeneration medium (30 g/L sucrose, 30 g/L sorbitol, 2 g/L casamino acids, MS Medium Salt Mixture [Duchefa Biochemie], 1.1 g/L MES, 2 mg/L 1-naphthalene acetic acid, and 1 mg/L kinetin, pH 5.8) and cultured for 15 to 30 days at 28 °C (16 h light/8 h dark). The means in the genetic transformation efficiency test represent data from 10 groups of independent experiments, with ~80 to ~120 transformed calli per group.

Quantitative determination of H2O2 contents and H2O2 flux

H2O2 concentrations were measured in calli subcultured for 21 days63. The calli (100 mg) were homogenized in 1 ml 0.1% (w/v) TCA in an ice bath. The homogenate was centrifuged at 12,000 × g for 15 min, and 1 ml of the supernatant was added to 1 ml 100 mM potassium phosphate buffer (pH 7.0) and 2 ml 1 M KI. The absorbance of the supernatant was read at 390 nm based on the fresh weight of the callus. H2O2 flux was measured using the Non-invasive Micro-test Technology (NMT) System (NMT100-SIM-XY, Younger USA Sci. & Tech. Corp., Amherst, MA, USA) at the Xuyue Beijing NMT Service Center (Xuyue Beijing Sci. and Tech. Co., Ltd., Beijing, China). Callus that had been subcultured for 21 d was fixed onto the bottom of a Petri using a filter paper strip and resin block. The callus in the Petri dish was incubated in measuring solution (0.1 mM KCl, 0.1 mM CaCl2, 0.1 mM MgCl2, 0.5 mM NaCl, 0.3 mM MES, 0.2 mM Na2SO4, pH 6.5) for ~20 min. The measuring solution was discarded and replaced with 5~10 ml of fresh measuring solution. H2O2 flux microsensor was placed a ~10 μm from the calli, and H2O2 flux at each site was measured for 10 min; each measurement repeated four times. Flux data were obtained using imFluxes V2.0 software (Younger USA LLC, Amherst, MA 01002, USA): a positive value represents efflux and a negative value represents influx. The H2O2 flux was calculated based on Fick′s law of diffusion (5):

$$J\left( {{\mathrm{pmol}}/{\mathrm{cm}}^2/s} \right) = - D_o \times \left( {dc/dx} \right)$$
(5)

where J is the analyte H2O2 flux (unit: pmol·cm−2·s−1), dc is its concentration gradient, dx is the distance between the two points (30 μm), and D0 is its diffusion constant (unit: cm−2·s−1).

Measurement of lipid peroxidation

MDA was measured using the thiobarbituric acid (TBA) assay64. The calli (100 mg) were homogenized in 1 ml 0.1% (w/v) TCA in an ice bath. After centrifuging the homogenate at 12,000 × g for 10 min, 0.5 ml of the supernatant was combined with 0.5 ml 0.6% thiobarbituric acid (TBA) and incubated at 95 °C for 15 min. The samples were cooled immediately in an ice bath and the absorbance of the supernatant measured at 450, 532, and 600 nm.

Assay of GSH content

GSH content was assayed by Shanghai Sanshu Biotechnology Co., Ltd (Shanghai, China). The reaction mixture contained plant extract, 0.1 M NaH2PO4 (pH 7.7), and DTNB (5,5-dithiobis-2-nitrobenzoic acid). Absorbance was recorded at 412 nm65.

Enzyme activity assays

The enzyme activity assays were carried out by Shanghai Sanshu Biotechnology Co., Ltd (Shanghai, China). Catalase (CAT) activity was measured based on the consumption of H2O2 and the decrease in absorbance at 240 nm66. Glutathione reductase (GR) was assayed based on the decrease in A340 due to nicotinamide adenine dinucleotide phosphate (NADPH) oxidation. The reaction mixture of GR contained 1.0 ml of 0.1 M potassium phosphate buffer (pH 8.0) including 0.5 mM EDTA, 0.5 mM MgCl2,1 mM NADPH, 10 mM oxidized glutathione (GSSG), and 0.1 ml enzyme extract67. The glutathione S-transferase (GST) was assayed using the substrate 1-chloro 2,4-dinitrobenzene (CDNB), and the activity was detected spectrophotometrically based on the change in absorbance at 340 nm68.

Endogenous hormone level measurements

The calli were ground into a powder in liquid nitrogen in a mortar and each sample (0.05 g) were transferred to a 1.5 ml screw-cap tube. After adding 50 μl of internal standards and 500 μl of extraction solvent (2-propanol: H2O: HCl = 2:1: 0.002) to each tube, the sample was incubated on a shaker at 900 × g for 30 min at 4 °C. After adding 1 ml trichloromethane, the sample was shaken for 30 min and centrifuged at 14,000 × g for 5 min at 4 °C to form two phases. A 1.2 ml sample of solvent from the lower phase was dried in a nitrogen evaporator, dissolved in 0.1 ml methanol and endogenous SA levels quantified by high-performance liquid chromatography-mass spectrometry (HPLC-MS/MS)69.

Ethylene was analyzed according to a published method70 with minor modifications. The calli were placed into 100-mL vials that were promptly sealed with a rubber plug and incubated at 28 °C for 48 h in the dark. Ethylene content in 2 ml of the headspace gas was analyzed with a gas chromatograph (GC-17A, Shimadzu, Japan).

RNA-seq analysis

Total RNA was isolated from Teqing and YIL25, pCPL and Teqing, and pOE and Teqing calli (subcultured for 21 days) with three biological replicates. Paired-end libraries were constructed and sequenced on the Illumina HiSeq 2500 platform at the Novogene Company (China). The RNA-seq reads were mapped to the reference genome (Os-Nipponbare-Refrence-IRGSP-1.0, MSU7) using TopHat2 with default parameters, and the FPKM (fragments per kilobase of exon per million mapped reads) of each gene and the DEGs (fold change ≥ 2, FDR < 0.001) between YIL25 and Teqing, pCPL and Teqing, and pOE and Teqing were calculated71.

Sequencing and data analysis

The ~4.6 kb genomic fragment covering the entire BOC1 gene (2740 bp), the 1,452 bp 5′-flanking region, and the 380 bp 3′-flanking region were amplified using five PCR primer pairs and sequenced using the Sanger sequencing approach. A previously published dataset in 1,083 diverse rice accessions was used to calculate the fixation index (FST) across chromosome 3 between indica and japonica varieties using VCFtools 0.1.15, with a 100-Kb window size36,72.

Data analysis

A two-tailed Student’s t-test was performed using SPSS version 16 (SPSS Inc., Chicago, IL, USA). QTL analysis was performed using MapManager QTXb 2073. Amino acid sequences homologous to BOC1 were downloaded from the National Center for Biotechnology Information website (http://blast.ncbi.nlm.nih.gov/). Phylogenetic trees were constructed using the neighbor-joining algorithm in MEGA version 7 with 1000 bootstrap replications74. The association tests were analyzed by R v3.6.1.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

Data supporting the findings of this work are available within the paper and its Supplementary Information files. A reporting summary for this Article is available as a Supplementary Information file. The datasets generated and analyzed during the current study are available from the corresponding author upon request. RNA-seq data for callus that had been subcultured for 21 days from Teqing and YIL25, Teqing and the complementation line pCPL-BOC1-Teqing, and Teqing and the overexpression line pOE-BOC1-Teqing was deposited in the National Center for Biotechnology Information (NCBI) under GEO accession number GSE142094. The source data underlying Figs. 1j, 1k, 2a, 2b, 3e, 3j, 3o, 4a, 4c, 5, 6, 8b, 8c, 8e, and 8f, as well as Supplementary Figs. 2, 5, 7, 8k, and 9 are provided as a Source Data file.

References

  1. 1.

    He, Y. et al. Changes in morphology and biochemical indices in browning callus derived from Jatropha curcas hypocotyls. Plant Cell Tiss. Org. 98, 11–17 (2009).

    CAS  Article  Google Scholar 

  2. 2.

    Tóth, K., Haapala, T. & Hohtola, A. Alleviation of browning in oak explants by chemical pretreatments. Biol. Plant. 36, 511–517 (1994).

    Article  Google Scholar 

  3. 3.

    Krishna, H. et al. Mango explant browning: effect of ontogenic age, mycorrhization and pre-treatments. Sci. Hortic. 118, 132–138 (2008).

    Article  Google Scholar 

  4. 4.

    Uchendu, E. E., Paliyath, G., Brown, D. C. W. & Saxena, P. K. In vitro propagation of North American ginseng (Panax quinquefolius L.). Vitr. Cell Dev.-Plant. 47, 710–718 (2011).

    CAS  Article  Google Scholar 

  5. 5.

    Thomas, T. D. The role of activated charcoal in plant tissue culture. Biotechnol. Adv. 26, 618–631 (2008).

    CAS  PubMed  Article  Google Scholar 

  6. 6.

    Han, M. L., Gleave, A. P. & Wang, T. C. Efficient transformation of Actinidia arguta by reducing the strength of basal salts in the medium to alleviate callus browning. Plant Biotechnol. Rep. 4, 129–138 (2010).

    Article  Google Scholar 

  7. 7.

    Dong, Y. S. et al. Mechanisms and effective control of physiological browning phenomena in plant cell cultures. Physiol. Plant. 156, 13–28 (2016).

    CAS  PubMed  Article  Google Scholar 

  8. 8.

    Mayer, A. M. & Harel, E. Polyphenol oxidases in plant. Phytochemistry 18, 193–215 (1979).

    CAS  Article  Google Scholar 

  9. 9.

    Murata, M. et al. A transgenic apple callus showing reduced polyphenol oxidase activity and lower browning potential. Biosci. Biotechnol. Biochem. 65, 383–388 (2001).

    CAS  PubMed  Article  Google Scholar 

  10. 10.

    Tang, W. & Newton, R. J. Increase of polyphenol oxidase and decrease of polyamines correlate with tissue browning in Virginia pine (Pinus virginiana Mill.). Plant Sci. 167, 621–628 (2004).

    CAS  Article  Google Scholar 

  11. 11.

    Sarmadi, M. et al. The effects of salicylic acid and glucose on biochemical traits and taxane production in a Taxus baccata callus culture. Plant Physiol. Bioch. 132, 271–280 (2018).

    CAS  Article  Google Scholar 

  12. 12.

    Adkins, S. W., Shiraishi, T. & Mccomb, J. A. Rice callus physiology-identification of volatile emissions and their effects on culture growth. Physiol. Plant. 78, 526–531 (1990).

    CAS  Article  Google Scholar 

  13. 13.

    Kobayashi, Y., Fukui, H. & Tabata, M. Effect of carbon dioxide and ethylene on berberine production and cell browning in Thalictrum minus cell cultures. Plant Cell Rep. 9, 496–499 (1991).

    CAS  PubMed  Article  Google Scholar 

  14. 14.

    Ogawa, T., Fukuoka, H., Yano, H. & Ohkawa, Y. Relationships between nitrite reductase activity and genotype-dependent callus growth in rice cell cultures. Plant Cell Rep. 18, 576–581 (1999).

    CAS  Article  Google Scholar 

  15. 15.

    Abe, T. & Futsufara, Y. Genotypic variability for callus formation and plant regeneration in rice (Oryza sativa L.). Theor. Appl. Genet. 72, 3–10 (1986).

    CAS  PubMed  Article  Google Scholar 

  16. 16.

    Taguchi-Shiobara, F. et al. Mapping quantitative trait loci associated with regeneration ability of seed callus in rice, Oryza sativa L. Theor. Appl. Genet. 95, 823–833 (1997).

    Article  Google Scholar 

  17. 17.

    Taguchi-Shiobara, F. et al. Mapping QTLs that control the performance of rice tissue culture and evaluation of derived near-isogenic lines. Theor. Appl. Genet. 112, 968–976 (2006).

    CAS  PubMed  Article  Google Scholar 

  18. 18.

    Takeuchi, Y., Abe, T. & Sasahara, T. RFLP mapping of QTLs influencing shoot regeneration from mature seed-derived calli in rice. Crop Sci. 40, 245–247 (2000).

    CAS  Article  Google Scholar 

  19. 19.

    Nishimura, A. et al. Isolation of a rice regeneration quantitative trait loci gene and its application to transformation systems. Proc. Natl Acad. Sci. USA 102, 11940–11944 (2005).

    ADS  CAS  PubMed  Article  Google Scholar 

  20. 20.

    Ozawa, K. & Kawahigashi, H. Positional cloning of the nitrite reductase gene associated with good growth and regeneration ability of calli and establishment of a new selection system for Agrobacterium-mediated transformation in rice (Oryza sativa L.). Plant Sci. 170, 384–393 (2006).

    CAS  Article  Google Scholar 

  21. 21.

    Li, Z. et al. A single genetic locus in chromosome 1 controls conditional browning during the induction of calli from mature seeds of Oryza sativa ssp indica. Plant Cell Tiss. Org. 89, 237–245 (2007).

    CAS  Article  Google Scholar 

  22. 22.

    Zhao, L. N. et al. Identification of quantitative trait loci controlling rice mature seed culturability using chromosomal segment substitution lines. Plant Cell Rep. 28, 247–256 (2009).

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  23. 23.

    Li, S. J. et al. Identification of QTLs associated with tissue culture response through sequencing-based genotyping of RILs derived from 93-11 x Nipponbare in rice (Oryza sativa). Plant Cell Rep. 32, 103–116 (2013).

    ADS  PubMed  Article  CAS  PubMed Central  Google Scholar 

  24. 24.

    Zhang, K. et al. Quantitative trait loci for mature embryo culturability traits from Yuanjiang common wild rice (Oryza rufipogon Griff.). Indian J. Genet. Pl. Br. 76, 167–172 (2016).

    CAS  Article  Google Scholar 

  25. 25.

    Yanagisawa, S. & Schmidt, R. J. Diversity and similarity among recognition sequences of Dof transcription factors. Plant J. 17, 209–214 (1999).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  26. 26.

    Abe, H. et al. Arabidopsis AtMYC2 (bHLH) and AtMYB2 (MYB) function as transcriptional activators in abscisic acid signaling. Plant Cell 15, 63–78 (2003).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  27. 27.

    Chinnusamy, V. et al. ICE1: a regulator of cold-induced transcriptome and freezing tolerance in Arabidopsis. Gene. Dev. 17, 1043–1054 (2003).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  28. 28.

    Simpson, S. D. et al. Two different novel cis-acting elements of erd1, a clpA homologous Arabidopsis gene function in induction by dehydration stress and dark-induced senescence. Plant J. 33, 259–270 (2003).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  29. 29.

    Zhu, J. K. Abiotic stress signaling and responses in plants. Cell 167, 313–324 (2016).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  30. 30.

    Liu, Y. et al. ABA overly-sensitive 5 (ABO5), encoding a pentatricopeptide repeat protein required for cis-splicing of mitochondrial nad2 intron 3, is involved in the abscisic acid response in Arabidopsis. Plant J. 63, 749–765 (2010).

    CAS  PubMed  Article  Google Scholar 

  31. 31.

    Singh, S. et al. Down-regulation of OsSAG12-1 results in enhanced senescence and pathogen-induced cell death in transgenic rice plants. J. Biosci. 38, 583–592 (2013).

    CAS  PubMed  Article  Google Scholar 

  32. 32.

    Zangar, R. C., Davydov, D. R. & Verma, S. Mechanisms that regulate production of reactive oxygen species by cytochrome P450. Toxicol. Appl. Pharm. 199, 316–331 (2004).

    CAS  Article  Google Scholar 

  33. 33.

    Niu, N. N. et al. EAT1 promotes tapetal cell death by regulating aspartic proteases during male reproductive development in rice. Nat. Commun. 4, X (2013).

    Article  CAS  Google Scholar 

  34. 34.

    Jing, S. J., Zhou, X., Song, Y. & Yu, D. Q. Heterologous expression of OsWRKY23 gene enhances pathogen defense and dark-induced leaf senescence in Arabidopsis. Plant Growth Regul. 58, 181–190 (2009).

    CAS  Article  Google Scholar 

  35. 35.

    Kim, S. G. et al. The RNase activity of rice probenazole-induced protein1 (PBZ1) plays a key role in cell death in plants. Mol. Cells 31, 25–31 (2011).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  36. 36.

    Huang, X. H. et al. A map of rice genome variation reveals the origin of cultivated rice. Nature 490, 497–501 (2012).

    ADS  CAS  PubMed  Article  Google Scholar 

  37. 37.

    You, J. et al. The SNAC1-targeted gene OsSRO1c modulates stomatal closure and oxidative stress tolerance by regulating hydrogen peroxide in rice. J. Exp. Bot. 64, 569–583 (2013).

    CAS  PubMed  Article  Google Scholar 

  38. 38.

    Overmyer, K. et al. Ozone-sensitive Arabidopsis rcd1 mutant reveals opposite roles for ethylene and jasmonate signaling pathways in regulating superoxide-dependent cell death. Plant Cell 12, 1849–1862 (2000).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  39. 39.

    Overmyer, K. et al. Ozone-induced programmed cell death in the Arabidopsis radical-induced cell death1 mutant. Plant Physiol. 137, 1092–1104 (2005).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  40. 40.

    Ahlfors, R. et al. Arabidopsis radical-induced cell death1 belongs to the WWE protein-protein interaction domain protein family and modulates abscisic acid, ethylene, and methyl jasmonate responses. Plant Cell 16, 1925–1937 (2004).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  41. 41.

    Ahlfors, R., Brosché, M., Kollist, H. & Kangasjärvi, J. Nitric oxide modulates ozone-induced cell death, hormone biosynthesis and gene expression in Arabidopsis thaliana. Plant J. 58, 1–16 (2009).

    CAS  PubMed  Article  Google Scholar 

  42. 42.

    Katiyar-Agarwal, S. et al. The plasma membrane Na+/H+ antiporter SOS1 interacts with RCD1 and functions in oxidative stress tolerance inArabidopsis. Proc. Natl Acad. Sci. USA 103, 18816–18821 (2006).

    ADS  CAS  PubMed  Article  Google Scholar 

  43. 43.

    Hiltscher, H. et al. The radical induced cell death protein 1 (RCD1) supports transcriptional activation of genes for chloroplast antioxidant enzymes. Front. Plant Sci. 5, 475 (2014).

    PubMed  PubMed Central  Article  Google Scholar 

  44. 44.

    Niu, X. M. et al. Transposable elements drive rapid phenotypic variation in Capsella rubella. Proc. Natl Acad. Sci. USA 116, 6908–6913 (2019).

    CAS  PubMed  Article  Google Scholar 

  45. 45.

    Xia, C. et al. A TRIM insertion in the promoter of Ms2 causes male sterility in wheat. Nat. Commun. 8, 15407 (2017).

    ADS  CAS  PubMed  PubMed Central  Article  Google Scholar 

  46. 46.

    Mao, H. D. et al. A transposable element in a NAC gene is associated with drought tolerance in maize seedlings. Nat. Commun. 6, 8326 (2015).

    ADS  CAS  PubMed  PubMed Central  Article  Google Scholar 

  47. 47.

    Jones, A. M. P. & Saxena, P. K. Inhibition of phenylpropanoid biosynthesis in Artemisia annua L.: a novel approach to reduce oxidative browning in plant tissue culture. PLoS One 8, e76802 (2013).

    ADS  CAS  PubMed  PubMed Central  Article  Google Scholar 

  48. 48.

    Gechev, T. S. et al. Reactive oxygen species as signals that modulate plant stress responses and programmed cell death. Bioessays 28, 1091–1101 (2006).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  49. 49.

    Overmyer, K., Brosché, M. & Kangasjärvi, J. Reactive oxygen species and hormonal control of cell death. Trends Plant Sci. 8, 335–342 (2003).

    CAS  PubMed  Article  Google Scholar 

  50. 50.

    Kangasjärvi, S. et al. Photosynthesis, photorespiration, and light signalling in defence responses. J. Exp. Bot. 63, 1619–1636 (2012).

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  51. 51.

    Petrov, V., Hille, J., Mueller-Roeber, B. & Gechev, T. S. ROS-mediated abiotic stress-induced programmed cell death in plants. Front. Plant Sci. 6, 69 (2015).

    PubMed  PubMed Central  Article  Google Scholar 

  52. 52.

    Pan, D. J. et al. Interactions between salicylic acid and antioxidant enzymes tilting the balance of H2O2 from photorespiration in non-target crops under halosulfuron-methyl stress. Pestic. Biochem. Phys. 143, 214–223 (2017).

    CAS  Article  Google Scholar 

  53. 53.

    Chao, Y. Y. et al. Salicylic acid-mediated hydrogen peroxide accumulation and protection against Cd toxicity in rice leaves. Plant Soil. 329, 327–337 (2010).

    CAS  Article  Google Scholar 

  54. 54.

    Housti, F., Coupe, M. & Dauzac, J. Effect of ethylene on enzymatic activities involved in the browning of Hevea brasiliensis callus. Physiol. Plant. 86, 445–450 (1992).

    CAS  Article  Google Scholar 

  55. 55.

    Shapiguzov, A. et al. Arabidopsis RCD1 coordinates chloroplast and mitochondrial functions through interaction with ANAC transcription factors. eLife 8, e43284 (2019).

    PubMed  PubMed Central  Article  Google Scholar 

  56. 56.

    Sun, C. Q. et al. Comparison of the genetic diversity of common wild rice (Oryza rufipogon Griff.) and cultivated rice (O. sativa L.) using RFLP markers. Theor. Appl. Genet. 102, 157–162 (2001).

    CAS  Article  Google Scholar 

  57. 57.

    Tan, L. B. et al. Development of Oryza rufipogon and O. sativa introgression lines and assessment for yield related quantitative trait loci. J. Integr. Plant Biol. 49, 871–884 (2007).

    CAS  Article  Google Scholar 

  58. 58.

    Chu, C. C. et al. Establishment of an efficient medium for anther culture of rice through comparative experiments on nitrogen sources. Sci. China Chem. 18, 659–668 (1975).

    Google Scholar 

  59. 59.

    Gamborg, O. L., Miller, R. A. & Ojima, K. Nutrient requirements of suspension cultures of soybean root cells. Exp. Cell Res. 50, 151–158 (1968).

    CAS  PubMed  Article  Google Scholar 

  60. 60.

    Livak, K. J. & Schmittgen, T. D. Analysis of relative gene expression data using real-time quantitative PCR and the 2-CT method. Methods 25, 402–408 (2001).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  61. 61.

    Javelle, M., Marco, C. F. & Timmermans, M. In situ hybridization for the precise localization of transcripts in plants. J. Vis. Exp. 57, e3328 (2011).

    Google Scholar 

  62. 62.

    Shen, J. et al. Isolation, culture, and transient transformation of plant protoplasts. Curr. Protoc. Cell Biol. 63, 1–17 (2014).

    Article  Google Scholar 

  63. 63.

    Velikova, V., Yordanov, I. & Edreva, A. Oxidative stress and some antioxidant systems in acid rain-treated bean plants: protective role of exogenous polyamines. Plant Sci. 151, 59–66 (2000).

    CAS  Article  Google Scholar 

  64. 64.

    Peever, T. L. & Higgins, V. J. Electrolyte leakage, lipoxygenase, and lipid-peroxidation induced in tomato leaf tissue by specific and nonspecific elicitors from Cladosporium fulvum. Plant Physiol. 90, 867–875 (1989).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  65. 65.

    Griffith, O. W. Determination of glutathione and glutathione disulfide using glutathione-reductase and 2-vinylpyridine. Anal. Biochem. 106, 207–212 (1980).

    CAS  PubMed  Article  Google Scholar 

  66. 66.

    Jiménez, A., Hernández, J. A., DelRio, L. A. & Sevilla, F. Evidence for the presence of the ascorbate-glutathione cycle in mitochondria and peroxisomes of pea leaves. Plant Physiol. 114, 275–284 (1997).

    PubMed  PubMed Central  Article  Google Scholar 

  67. 67.

    Edwards, E. A., Rawsthorne, S. & Mullineaux, P. M. Subcellular distribution of multiple forms of glutathione reduetase in leaves of pea (Pisum sativum L.). Planta 180, 278–284 (1990).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  68. 68.

    Habig, W. H. & Jakoby, W. B. Assay for differentiation of glutathione S-transferases. Method Enzymol. 77, 398–405 (1981).

    CAS  Article  Google Scholar 

  69. 69.

    Pan, X., Welti, R. & Wang, X. Quantitative analysis of major plant hormones in crude plant extracts by high-performance liquid chromatography-mass spectrometry. Nat. Protoc. 5, 986–992 (2010).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  70. 70.

    Roustan, J. P., Latche, A. & Fallot, J. Inhibition of ethylene production and stimulation of carrot somatic embryogenesis by salicylic acid. Biol. Plant. 32, 273–276 (1990).

    CAS  Article  Google Scholar 

  71. 71.

    Trapnell, C. et al. Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat. Biotechnol. 28, 511–515 (2010).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  72. 72.

    Danecek, P. et al. The variant call format and VCFtools. Bioinformatics 27, 2156–2158 (2011).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  73. 73.

    Manly, K. F., Cudmore, R. J. & Meer, J. M. Map Manager QTX, cross-platform software for genetic mapping. Mamm. Genome 12, 930–932 (2001).

    CAS  PubMed  Article  Google Scholar 

  74. 74.

    Kumar, S., Stecher, G. & Tamura, K. MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol. Biol. Evol. 33, 1870–1874 (2016).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

Download references

Acknowledgements

We thank the International Rice Research Institute, Chinese Rice Research Institute, Institute of Crop Sciences of Chinese Academy of Agricultural Sciences, Guangxi Academy of Agricultural Sciences, and Guangdong Academy of Agricultural Sciences for providing the wild and cultivated rice germplasm. This project was supported by the National Natural Science Foundation of China (grant number 31571626) and the National Transgenic Major Project of China (grant number 2016ZX08001-004).

Author information

Affiliations

Authors

Contributions

Y.F. conceived and designed the research. K.Z. performed the experiments, analyzed data, and wrote the paper. J.S., M.X., Z.Z. and J.H. assisted with experiments. X.Z. and X.M helped analyze the data. C.L and Y.L. helped collect the O. rufipogon accessions. L.T., Z.Z., H.C., F.L., H.S., P.G., W.Z. and C.S. provided technical assistance and revised the paper. Y.F. performed data analysis and revised the paper.

Corresponding author

Correspondence to Yongcai Fu.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Nature Communications thanks Kirk Overmyer, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Source data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Zhang, K., Su, J., Xu, M. et al. A common wild rice-derived BOC1 allele reduces callus browning in indica rice transformation. Nat Commun 11, 443 (2020). https://doi.org/10.1038/s41467-019-14265-0

Download citation

Further reading

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.