Mapping quantitative trait loci associated with callus browning in Dongxiang common wild rice (Oryza rufipogon Griff.)

Genetic transformation of indica rice (Oryza sativa ssp. indica) is limited by callus browning, which results in poor in vitro tissue culturability. Elucidating the genes in common wild rice (Oryza rufipogon Griff.) that control callus browning is fundamental for improving the tissue culturability of indica rice varieties. We used a population of 129 O. rufipogon (Dongxiang common wild rice; DXCWR) introgression lines in the elite cultivar GC2 (Oryza sativa ssp. indica) background and 159 simple sequence repeat (SSR) markers to identify quantitative trait loci (QTLs) associated with callus browning. We evaluated callus browning based on the indices of callus browning rate (CBR), callus browning index (CBI), and standard callus browning index (SCBI). We detected 30 QTLs associated with callus browning across all lines, mapping to chromosomes 1, 2, 3, 4, 8, 9, and 12. These genomic regions were repeatedly associated with differences in CBR, CBI, and SCBI. The alleles from DXCWR showed additive effects in reducing callus browning. We identified new QTLs near the markers RM247 and RM7003 on chromosome 12, indicating that these QTLs are unique to DXCWR. Furthermore, we identified six introgression lines with significantly lower callus browning. These lines will be useful germplasms for genetic transformation and fine-mapping of the culturability trait.


Introduction
Efficient genetic transformation is crucial for gene functional studies, molecular breeding, and the development of genetically modified crops. However, Agrobacterium (Agrobacterium tumefaciens)-mediated transformation of indica rice accessions (Oryza sativa ssp. indica) is substantially hindered by callus browning, which is less prominent in japonica accessions (O. sativa ssp. japonica). Callus browning is a common phenomenon in many plant species, resulting in lower regenerative ability, poor growth, and even death of callus [1,2]. Previous studies have shown that various physiological and biochemical indices [2][3][4][5][6] and phytohormones [2,7,8] affect callus browning. However, the most important factor determining the likelihood of callus browning is the genotype [2,9,10]. Identifying the underlying genes associated with a reduction in callus browning and utilizing superior alleles should therefore Yibo Wang and Xin Yang have contributed equally to this work. 1 3 be a major target for functional genetics analyses, genome editing, and molecular breeding [11].
Several quantitative trait loci (QTLs) associated with rice response to tissue culture have been identified [2,12,[12][13][14][15][16][17][18][19][20]; however, only two of the responsible genes associated with callus culturability have been isolated. One of these genes maps to chromosome 1 and encodes ferredoxin-nitrite reductase (NiR), which controls callus differentiation [15]. The second gene is BROWNING OF CALLUS1 (BOC1), which maps to chromosome 3 and encodes SIMILAR TO RADICAL-INDUCED CELL DEATH ONE (SRO), which is involved in reducing callus browning derived from Yuanjiang common wild rice [2]. Wild rice is an important resource of favorable genes [21][22][23]. Twenty-five QTLs related to tissue culturability have been identified in common wild rice (Oryza rufipogon Griff.) as being responsible for variation in callus browning index and callus regeneration rate, two commonly used indices to quantify the extent of browning [10]. Identifying and characterizing the underlying genes would be a major step towards improving transformation efficiency.
Wild rice is an important favorable gene pool [21][22][23]. The genetic background of the introgression lines is single, with only a few genomic fragments from common wild rice otherwise cultivar genomic background, and is thus a useful resource for genetic analysis and functional identification [24,25]. Compared to conventional cultivar materials, introgression lines derived from common wild rice are more accurate and efficient for mapping and cloning genes. In studies using map-based cloning, we previously identified two important genes derived from Dongxiang common wild rice (DXCWR) from introgression lines generated in the indica GC2 background that were associated with the number of grains produced [26] and cold tolerance at the early seedling stage [27]. GC2 is an elite indica rice variety; however, this genotype is susceptible to callus browning during in vitro culture, making genetic transformation of this germplasm challenging.
In the present study, we investigated the genetic architecture of the callus browning trait using 129 introgression lines harboring distinct genomic fragments from DXCWR in the indica cultivar GC2 background after 30 days of growth on induction medium. We genotyped the introgression lines with 159 simple sequence repeat (SSR) markers, which allowed us to perform QTL analysis using the single-point analysis method, laying a solid foundation for the further fine-mapping and cloning of genes involved in tissue culturability.

Plant materials
A total of 129 O. rufipogon introgression lines (BC 4 F 4 ; named 2DILs), derived from a cross between the O. rufipogon accession DXCWR as the donor and the elite indica variety GC2 (O. sativa) as the recipient, were selected from 265 lines on the basis of the coverage of DXCWR introgression segments along chromosomes. These lines were used to assess callus browning.

Tissue culture procedure
After hulling, ~ 90 mature, healthy, dehusked seeds from each line were surface sterilized in 70% (v/v) ethanol for ~ 2 min, followed by a 15% (w/v) sodium hypochlorite solution for 15 min with shaking 220 rpm. Seeds were then rinsed three or four times with sterile water on an ultraclean workbench. The surface-sterilized seeds were placed on induction medium in three plates (25-30 seeds per plate) and incubated in the dark at 28 °C for 30 days. The callus browning phenotypes of calli in each plate were recorded as described below. For each replicate, all introgression lines were divided into five subgroups, each containing 30 introgression lines and the recurrent parent GC2.

Phenotypic evaluation of callus browning
Callus browning was assessed as described by Zhang et al. [2]. The degree of callus browning was categorized into five levels: (0) less than 1/10 of callus tissue is brown; (1) 1/10 to 1/3 brown; (2) > 1/3 to 2/3 brown; (3) > 2/3 to mostly brown; (4) all brown. To minimize phenotyping errors due to observations made by different individuals or differences arising from variations in the cultivation environment, the phenotypes were standardized using the callus browning index (SCBI). The tendency for callus browning was calculated according to Eqs.

QTL analysis
A total of 159 SSR primer pairs was used to determine the genotypes of each of the 129 introgression lines. The association between phenotype and genotype was investigated using a single-point analysis in the software Map Manager QTXb20 [30]. The statistical threshold for the single-point analysis was P < 0.05.

Statistical analysis
The program GGT2.0 [31] was used to visualize individual genotypes. The phenotyping data were converted to percentages in Microsoft Excel, and were analyzed using SPSS v25.0 software, which included drawing the frequency distributions, performing analyses of variance (ANOVAs), and performing correlation analyses of the three indices between the introgression lines.

Phenotypic variation in callus browning between DXCWR, the 2DIL introgression lines, and the recurrent parent GC2
In our previous study, we constructed a set of introgression lines (named 2DILs) using the wild rice O. rufipogon accession DXCWR as a donor and the elite indica cultivar GC2 (O. sativa) as the recurrent parent. DXCWR is relatively resistant to browning, with a CBI of 0.22 (Table 1; Supplementary Fig. 1) [2]. To identify QTLs involved in regulating callus browning in rice, we screened 129 introgression lines by placing mature seeds on unimproved NB medium (NB basal medium without additives, named NB1) for 30 days (Fig. 1). In parallel, we determined the genotypes of the 129 introgression lines with a set of SSR markers and used the information to represent their genotypes with the program GGT2.0 (Fig. 2). We scored callus browning on a five-level scale relative to GC2: level 0 corresponded to no browning; levels 1 and 2 to medium browning; and levels 3 and 4 to severe browning, similar to that of GC2 ( Fig. 1b-f). The recurrent parent GC2 was more susceptible to callus browning than DXCWR ( Fig. 1a; Table 1). The callus browning rate (CBR), callus browning index (CBI), and standard callus browning index (SCBI) varied among the 2DIL population, ranging from 0-100%, 0-0.93, and − 1-0.40, respectively (Table 1 and Fig. 3), which laid a good foundation for further identification of QTLs.

Variance and correlation analysis among the three callus browning indices
The three indices related to callus browning varied substantially among the introgression lines ( Table 2). An analysis of variance (ANOVA) revealed that this variation in each index is due to the genotype ( Table 2). We determined the extent of correlation among these indices by calculating their associated pairwise correlation coefficients (r), which revealed positive correlations between all indices: CBR and CBI (r = 0.895), CBR and SCBI (r = 0.828), and CBI and SCBI (r = 0.938) ( Table 3). Consequently, we used CBR, CBI, and SCBI separately for QTL analysis.

Screening elite introgression lines with low callus browning
We characterized six introgression lines (2DIL18, 2DIL99, 2DIL101, 2DIL103, 2DIL110, and 2DIL112) exhibiting a significant reduction in callus browning compared to GC2 (Fig. 6). Genotype analysis showed that the six selected introgression lines harbor introgression fragments distributed across different chromosomes: 2DIL110 and 2DIL112 had an introgression fragment near marker RM71 on chromosome 2; 2DIL101 had an introgression fragment near marker RM114 on chromosome 3; 2DIL18 and 2DIL110 had introgression fragments near marker RM335 on chromosome 4; 2DIL99, 2DIL103, and 2DIL110 had introgression fragments near markers RM328, OSR29, and RM189 on chromosome 9; and 2DIL99 had introgression fragments near markers RM7003 on chromosome 12, respectively (Fig. 7). The locations of the introgression fragments contained in these six introgression lines was consistent with those identified in our QTL analysis. These lines can be used not only as genetic transformation receptors for indica rice accessions, but also as parent materials for fine-mapping of callus browning traits.

DXCWR contains a wealth of new gene/allele resources conferring tolerance to abiotic stress
An efficient tissue-culture and transformation system has been established for japonica rice using mature seeds; however, the genetic transformation efficiency of indica rice is lower due to callus browning. Notably, most genetic analyses of callus browning have been based on cultivated rice, with few studies investigating those alleles associated with callus browning that have been lost from common wild rice. In our previous study, we proposed that BOC1, derived from a common wild rice accession (Oryza rufipogon G.), was involved in cell senescence and death caused by oxidative stress [2]. In this study, the wild rice germplasm DXCWR displayed significantly less callus browning than GC2 (Table 1), and callus browning indices in the introgression lines (2DILs) showed a wide range of values (Table 1; Fig. 3), indicating that DXCWR may harbor favorable alleles for reducing callus browning when introduced into cultivated rice.

Accuracy and reproducibility of phenotypic assessments
The identification of tissue culture traits is influenced by various factors, such as the genotype, culture medium, selection of explants, and physiological status of the donor material, as well as interactions between these factors [9,32]. Phenotypic evaluation can also suffer from artificial errors introduced by subjective differences between different individuals and different survey criteria, even when using the same genotypes in the same laboratory and when following a standardized scaling system [19]. In our study, we investigated the callus browning trait under the supervision of a single individual and following the same criteria for all lines to minimize errors. We also conducted three replicates per introgression line, which revealed that our callus browning trait results were stable and reproducible. In a number of previous studies, researchers have also observed variation in callus status and proliferation ability in subculture [13,17,18]; however, in the present study, we investigated callus browning cultured on NB1 medium after 30 days to observe the maximum possible phenotypic variation for callus browning in the introgression lines and reduce the incubation time to decrease workload. The selection not to scutellum-derived calli would have increased our working efficiency and speed up the experiments.    [10]. The QTLs located on the short arm of chromosome 4 and the long arm of chromosome 9 appear to be hotspots for the control of the callus browning phenotype in rice. Furthermore, we identified a QTL with high phenotypic variance on chromosome 12 that had not been reported before, indicating that it may be unique to DXCWR. Thus, the QTLs identified in our study should be fine-mapped and their underlying genes should be functionally analyzed to assess their contribution to decreasing callus browning and their potential for improving genetic transformation efficiency.  . Two-tailed Student's t-tests were used to determine significant differences compared to GC2

Conclusion
We used 129 introgression lines (2DILs) derived from DXCWR (O. rufipogon) in the elite indica accession GC2 background to map quantitative trait loci related to callus browning. We identified four QTLs near the markers RM71 on chromosome 2, RM335 on chromosome 4, and RM328 and RM189 on chromosome 9 using CBR, CBI, and SCBI indices as phenotypes. The locations of these QTLs were consistent with those identified in previous studies for callus browning and accounted for 4-8% of the phenotypic variation for this trait. The QTL near RM247 and RM7003 on chromosome 12 appeared to be a previously unidentified QTL whose DXCWR allele reduced callus browning. Moreover, we identified six elite introgression lines with low callus browning potential. These results provide the necessary materials and genetic resources for genetic transformation and solving the molecular breeding bottleneck of indica rice. Author contributions KZ and YF conceived and designed the experiments. YW analyzed the phenotypic data. YW, XY, and GX performed the experiments. XY, YJ, XL, and JS helped to perform the experiments. XL helped to perform the QTL analysis. KZ analyzed the data and wrote the manuscript. CS and YF supervised the research and edited the manuscript. All authors have read and agreed to the submitted version of the manuscript. Data availability All data generated or analyzed during this study are included in the published article.

Conflict of interest
The authors declare that they have no conflict of interest.

Ethical approval
The current study does not involve any animal or human data.
Informed consent Not applicable for current study.