Improvement of Stable Restorer Lines for Blast Resistance through Functional Marker in Rice (Oryza sativa L.)

Two popular stable restorer lines, CB 87 R and CB 174 R, were improved for blast resistance through marker-assisted back-cross breeding (MABB). The hybrid rice development program in South India extensively depends on these two restorer lines. However, these restorer lines are highly susceptible to blast disease. To improve the restorer lines for resistance against blasts, we introgressed the broad-spectrum dominant gene Pi54 into these elite restorer lines through two independent crosses. Foreground selection for Pi54 was done by using gene-specific functional marker, Pi54 MAS, at each back-cross generation. Back-crossing was continued until BC3 and background analysis with seventy polymorphic SSRs covering all the twelve chromosomes to recover the maximum recurrent parent genome was done. At BC3F2, closely linked gene-specific/SSR markers, DRRM-RF3-10, DRCG-RF4-8, and RM 6100, were used for the identification of fertility restoration genes, Rf3 and Rf4, along with target gene (Pi54), respectively, in the segregating population. Subsequently, at BC3F3, plants, homozygous for the Pi54 and fertility restorer genes (Rf3 and Rf4), were evaluated for blast disease resistance under uniform blast nursery (UBN) and pollen fertility status. Stringent phenotypic selection resulted in the identification of nine near-isogenic lines in CB 87 R × B 95 and thirteen in CB 174 R × B 95 as the promising restorer lines possessing blast disease resistance along with restoration ability. The improved lines also showed significant improvement in agronomic traits compared to the recurrent parents. The improved restorer lines developed through the present study are now being utilized in our hybrid development program.


Introduction
The major goal of rice production is to fulfill the demand for the growing population and improve food security. Among various genetic approaches available today, hybrid rice technology is the most promising and accepted strategy for improving the rice productivity [1]. However, the hybrid rice production system suffers from various biotic stresses including fungal and bacterial diseases [1,2]. Among which, fungal blast (Magnaporthe oryzae), caused by an ascomycete fungus, is one of the major biotic diseases of rice contributing to yield loss up to 10-30% globally [3][4][5][6][7]. In India the disease, under severe and favorable conditions, can cause yield losses ranging from 74 to 100% [8][9][10].

Plant Materials
Two stable restorer lines, CB 87 R and CB 174 R, which are predominantly employed in the hybrid rice program in South India, were used as recurrent parents. Utilizing these lines, two rice hybrids, CO RH 3 (CO 2A/CB 87 R) and CO RH 4 (CO 23 A/CB 174 R), were recently released from this center for farmers' cultivation. These hybrids are mostly preferred by the farmers of South India, because of their high seed setting and high yield (6600 and 7348 kg/ha, respectively), but they are highly susceptible to blast infection. The line B 95 (B 95-1 × Tetep) was used as the donor parent for Pi54 in two independent back-cross breeding steps, and two restorer lines, CB 87 R and CB 174 R, as the recurrent parents. In addition, Tetep as positive check and IR 24 as negative check were used in the screening program. The schematic steps for the introgression of blast resistance genes into restorer lines are presented in Figure 1.  Schematic steps for introgression of blast resistance genes into restorer lines through marker-assisted selection Progenies homozygous for blasts (Pi54) and fertility restoration (Rf3, Rf4) were evaluated for agronomical traits and agro-morphological traits.

Marker-Assisted Selection for Blasts and Fertility Restoration Genes
Marker-assisted back-cross breeding program was adopted to transfer Pi54 into two stable restorer lines, CB 87 R and CB 174 R. The fungal blast susceptible restorer lines, CB 87 R and CB 174 R, were crossed with the donor resistant line, B 95. Back-crossing was carried out up to BC3 generation. The PCR-based functional marker, Pi54 MAS [20], was used to identify the heterozygous nature of hybrids at BC1F1 and subsequent back-cross generations (BC2 and BC3) ( Table 1). One hundred and fifty-six SSR (simple sequence repeat) markers (Supplementary Table S1) were used to recover the recurrent parent background in each back-cross generation. A single BC3F1 plant with high recurrent genome in both the crosses was self-pollinated to produce BC3F2. The segregating population was screened with Pi54 MAS to identify homozygous plants. In addition, to verify fertility restoration genes, Rf3 and Rf4, linked/gene-specific markers were used (Table 1).

Figure 1.
Schematic steps for introgression of blast resistance genes into restorer lines through marker-assisted selection Progenies homozygous for blasts (Pi54) and fertility restoration (Rf3, Rf4) were evaluated for agronomical traits and agro-morphological traits.

Marker-Assisted Selection for Blasts and Fertility Restoration Genes
Marker-assisted back-cross breeding program was adopted to transfer Pi54 into two stable restorer lines, CB 87 R and CB 174 R. The fungal blast susceptible restorer lines, CB 87 R and CB 174 R, were crossed with the donor resistant line, B 95. Back-crossing was carried out up to BC 3 generation. The PCR-based functional marker, Pi54 MAS [20], was used to identify the heterozygous nature of hybrids at BC 1 F 1 and subsequent back-cross generations (BC 2 and BC 3 ) (Table 1). One hundred and fifty-six SSR (simple sequence repeat) markers (Supplementary Table S1) were used to recover the recurrent parent background in each back-cross generation. A single BC 3 F 1 plant with high recurrent genome in both the crosses was self-pollinated to produce BC 3 F 2 . The segregating population was screened with Pi54 MAS to identify homozygous plants. In addition, to verify fertility restoration genes, Rf3 and Rf4, linked/gene-specific markers were used (Table 1).

PCR Amplification
DNAs were extracted from the plants belonging to the parents and back-cross generations employing the simplified mini-scale method [23]. PCR reactions (15 µL) containing 3 µL of genomic DNA, 1× assay buffer, 200 µL of dNTPs, 2 µM MgCl 2 , 0.2 µM each primer (forward and reverse), and 1 unit of Taq polymerase (Bangalore Genei) were amplified. The temperature cycle was programmed as 95 • C for 2 min, 94 • C for 45 s, 52 • C for 1 min, 72 • C for 1.30 min for 35 cycles, and an additional temperature of 72 • C for 10 min for extension and 4 • C for cooling. The amplified products were electrophoresed in a 3% agarose gel and stained with ethidium bromide. Clearly resolved, unambiguous bands were scored visually for their presence with each primer.

Screening for Blast Resistance, Pollen Fertility, and Evaluation of Agronomic Characters in Improved Lines
The back-cross-derived lines (BC 3 F 3 ) along with parents and IR 24 (susceptible check) and Tetep (resistant check) were screened for blast resistance under a uniform blast nursery (UBN) using a virulent local isolate IS (KUL)-6 of M. oryzae [24]. For UBN screening, each BC 3 F 3 line was sown in a single row 50 long and 10 cm apart, and after every 20 test entries and border rows of all the sides of bed the susceptible check IR 24 was planted. This helps to spread the inoculum. Relative humidity is maintained with water sprinklers. The beds were covered during the night to maintain a high humidity until disease development. Additionally, to create severe incidence across the bed, seven-day-old fungal culture maintained as described by Srinivas Prasad et al. [25] was used for artificial inoculation. The adjusted spore concentration of 10 5 spores/mL [2] was sprayed over the entire UBN plot until the entire plant surface of each individual plant become wet. On the fifteenth day after inoculation, the lines were scored for their resistance reaction as per IRRI (International Rice Research Institute) standard evaluation system using the 0-9 disease severity scale (IRRI 1996).
Pollen fertility was estimated from BC 3 F 3 lines by using 1% potassium iodide solution (Supplementary Figure S1). Plants with round and deeply stained pollen grains were classified as fertile. To evaluate the important agronomic traits, twenty-five-day-old seedlings of improved lines along with parents and checks were raised in experimental plot at the Agricultural College and Research Institute, Tamil Nadu Agricultural University, Madurai, India. The experimental plot was arranged in a randomized block design with three replications following standard agronomic practices [26,27]. Each improved line was raised in three rows, each row containing 20 plants. Ten plants in each replication were monitored for the observation of major yield contributing traits. Agronomic traits including days for 50% flowering, plant height (cm), number of productive tillers per plant, panicle length (cm), number of filled grains, 100-grain weight (g), and grain yield per plant (g) were recorded.

Statistical Analysis
The segregation pattern of Pi54 gene was calculated using the chi-square test (χ 2 ) with the formulae χ 2 = (O−E) 2 /E, where O represents the observed frequency and E represents the expected frequency. The SAS (Statistical Analysis System) package (SAS Institute Inc., Cary, NC, USA) was used for the analysis of the significance of variation between parents and improved restorer lines. The agglomerative clustering among the selected pyramided lines and parents was calculated based on seven quantitative traits using "R" software (R, 2013).

Results
The F 1 plants in two crosses were confirmed for their heterozygous nature using the functional marker, Pi54 MAS. Heterozygous plants were back-crossed to respective recurrent parents (CB 87 R and CB 174 R) to develop BC 1 F 1 . In BC 1 F 1 generation, a total of 21/183 (Imp. CB 87 R) and 17/194 (Imp. CB 174 R) plants showed heterozygosity for the marker, Pi54 MAS, which was linked to the blast resistance gene. The positive BC 1 F 1 plants were subjected to background selection using 70 polymorphic markers (https://archive.gramene.org/markers/microsat/ssr.html) to identify the plants with maximum recurrent parent genome (R PG R). The heterozygous plants with a maximum parental genome of 73.6% (CB 87 R) and 74.9% (CB 174 R) were selected and back-crossed to generate BC 2 F 1 plants. Back-crosses were continued up to BC 3 to recover the maximum recurrent genome. Positive BC 3 F 1 plants with the maximum recurrent genome of 91.1% in CB 87 R and 90.3% were self-pollinated to produce BC 3 F 2 plants. A total of 41 homozygous plants in CB 87 R and 46 in CB 174 R for the Pi54 MAS marker were screened for fertility restoring genes (Rf3 and Rf4) using gene specific/linked markers; DRRM-RF3-5 for Rf3 and DRCG-RF4-8 & RM 6100 for Rf4. Nine (CB 87 R) and 13 (CB 174 R) plants showing the co-segregating banding pattern in BC 3 F 3 ( Figure 2) were subjected to the pollen fertility study and agronomic traits. The number of plants scored for both blasts and fertility restoration using gene-specific markers in different back-crosses generation is provided in Supplementary Table S2. The segregation pattern of blast resistance in BC 3 F 2 generation plants was tested using χ 2 statistics. In the CB 87 R × B 95 cross, the observed segregating frequency was 41:106:63, while in the other cross, CB 174 R × B 95, it was about 46:127:87. Collectively, in both the crosses, the χ 2 value was non-significant and hence the observed genotypic ratio was nearly equal to the expected genotypic ratio ( Table 2).  Figure 2) were subjected to the pollen fertility study and agronomic traits. The number of plants scored for both blasts and fertility restoration using gene-specific markers in different back-crosses generation is provided in Supplementary Table S2. The segregation pattern of blast resistance in BC3F2 generation plants was tested using χ 2 statistics. In the CB 87 R × B 95 cross, the observed segregating frequency was 41:106:63, while in the other cross, CB 174 R × B 95, it was about 46:127:87. Collectively, in both the crosses, the χ 2 value was non-significant and hence the observed genotypic ratio was nearly equal to the expected genotypic ratio ( Table 2).

Assessment of Blast Resistance in the Pi54 Gene-Introgressed Lines
Twenty-two three-gene homozygous back-cross-derived lines were included for blast screening in the cross combination, along with recurrent parents, donor, and positive (Tetep) and negative check (IR 24). The disease susceptibility was evaluated for blast pathogen under controlled artificial greenhouse environment. The donor parent, B 95, and positive check, Tetep, harboring the Pi54 gene, showed a high level of resistance to rice blasts ( Figure 3) and scored as "1", whereas, CB 87 R and CB174 R showed susceptibility to rice blasts and were scored as "9". The twenty-two advanced three-gene (Pi54 + Rf3 + Rf4) homozygous-introgressed lines viz. Imp. CB 87 R (9 lines) and Imp. CB 174 R (13 lines) were selected from both of the parental crosses. The developed lines with the blast resistance gene exhibited no lesions on leaves and received a representative score of "0-2". The improved lines with two-gene combinations (Pi54 + Rf3 & Pi54 + Rf4) were excluded from our analysis. The blast resistance scoring and pollen fertility status for the improved lines and parents is presented in Table 3. The donor parent, B 95 (B 95-1 × Tetep) with the Pi54 gene, was found to be resistant to the virulent isolate of the pathogen (IS (KUL)-6) which is more prevalent in South India. At the time of the disease evaluation, each of the improved lines exhibited no symptoms and thus scored as "1". Roundish-shaped, small gray spots, slightly enlarged (1-2 mm in diameter) with a brown margin, present on the lower leaves were scored as "2". High pollen fertility was observed in all the improved lines (Supplementary Figure S1) Genes 2020, 11, x FOR PEER REVIEW 7 of 14

Assessment of Blast Resistance in the Pi54 Gene-Introgressed Lines
Twenty-two three-gene homozygous back-cross-derived lines were included for blast screening in the cross combination, along with recurrent parents, donor, and positive (Tetep) and negative check (IR 24). The disease susceptibility was evaluated for blast pathogen under controlled artificial greenhouse environment. The donor parent, B 95, and positive check, Tetep, harboring the Pi54 gene, showed a high level of resistance to rice blasts ( Figure 3) and scored as "1", whereas, CB 87 R and CB174 R showed susceptibility to rice blasts and were scored as "9". The twenty-two advanced three-gene (Pi54 + Rf3 + Rf4) homozygous-introgressed lines viz. Imp. CB 87 R (9 lines) and Imp. CB 174 R (13 lines) were selected from both of the parental crosses. The developed lines with the blast resistance gene exhibited no lesions on leaves and received a representative score of "0-2". The improved lines with two-gene combinations (Pi54 + Rf3 & Pi54 + Rf4) were excluded from our analysis. The blast resistance scoring and pollen fertility status for the improved lines and parents is presented in Table 3. The donor parent, B 95 (B 95-1 × Tetep) with the Pi54 gene, was found to be resistant to the virulent isolate of the pathogen (IS (KUL)-6) which is more prevalent in South India. At the time of the disease evaluation, each of the improved lines exhibited no symptoms and thus scored as "1". Roundish-shaped, small gray spots, slightly enlarged (1-2 mm in diameter) with a brown margin, present on the lower leaves were scored as "2". High pollen fertility was observed in all the improved lines (Supplementary Figure S1)

Agronomic Traits in Improved Restorer Lines
The twenty-two three-positive BC 3 F 3 plants, along with the parents of both crosses, were screened for agronomic performance parameters. The three-gene pyramided lines 5-3-7-6 and 5-3-8-10, although phenotypically similar to recurrent parents, performed better for most of the agronomic traits studied (  Figure S2). Interestingly, some of the improved lines had earlier maturity (8 days) compared to CB 87 R and CB 174 R. Five out of 22 improved lines possessed increased productive tillers, number of filled grains, and hundred grain weight as compared to recurrent parents. None of the improved lines displayed significant variation for plant height, panicle length, and grain yield compared to the unimproved restorer lines.

Cluster Analysis
Agglomerative clustering with Euclidean values divided the improved restorer lines and parents into four clusters (cophenetic correlation coefficient = 0.716). Cluster one consisted of three parents and two improved lines, while cluster two had the donor parent alone. Cluster three had six improved lines and cluster four had fourteen improved lines (Supplementary Figure S3). The highest yielding improved lines were placed in cluster one and found very similar to both recurrent parents.

Discussion
The present study was carried out with the objective to introgress a blast resistant gene, Pi54, into the background of popular restorer lines, CB 87 R and CB 174 R, through marker-assisted back-cross breeding coupled with screening for improved agronomic traits. Improvement of these lines for blast disease resistance facilitates the development of new generation of hybrids with durable resistance against blast disease in addition to bacterial blight disease. Since the 1960s, more than 100 resistance genes or loci for blast disease have been identified [28,29]. Among them, the majority of the genes have been induced by pathogen infection, while a few of the genes express constitutively. In addition, the expression of many R genes induce resistance against leaf blast at the seedling stage, while only a few R genes confer resistance to panicle blast [30][31][32][33]. Such complex interactions of R genes are an important challenge in blast resistance breeding to achieve broad-spectrum and durable resistance. Earlier reports on pyramiding of Pi genes against blast disease in different background combinations (indica, japonica) indicated the expression of varying levels of resistance. This means that different gene combinations produce different interaction effects, and some show no resistance comprising the simple accumulation of the resistance spectra of the target R genes [34][35][36][37]. However, the blast-resistant gene, Pi54, which was identified in a highly resistant genotype, Tetep, was confirmed to have broad-spectrum resistance against predominant races found in India [2,13]. The Pi54, from the Piz locus, also shows broad-spectrum resistance to both leaf and panicle blasts [12]. Wu et al. [37] also reported that pyramiding of Pi54 along with Pi1, Pi33, and Pigm exhibited high level of resistance to leaf and panicle blasts. Similarly, it was also proved that the R genes of the alleles from the Piz locus exhibit excellent resistance after combining them with certain independently distributed R genes, such as Pi56 and Pish [38]. To date, many successful marker-assisted introgression using the Pi54 gene have been made in the various backgrounds of rice varieties and hybrids all over the world [2,36,[39][40][41]. Keeping in view with this, the present study aimed to introgress the Pi54 gene from Tetep into stable restorer lines CB 87 R and CB 174 R, but susceptible to blast disease through MABB. To the best of our knowledge, ours is the third such report wherein functional marker Pi54 MAS has been utilized for foreground selection. Earlier, Vijay Kumar et al. [1] successfully introgressed the Pi54 gene from a highly blast-resistant genotype, i.e., Tetep, into an elite rice cultivar Samba Mahsuri (BPT 5204), a high-yielding rice variety with good cooking quality, but susceptible to blast disease, through MABB. Subsequently, a Samba Mahsuri variety, JGL1798, was introgressed with the Pi54 gene to resist against blast disease along with BB resistance through functional marker Pi54 MAS [42]. Therefore, pyramiding the alleles of Pi54 along with BB-harboring stable restorer lines would greatly accelerate the hybrid rice program for durable resistance against important biotic stresses of rice.
For foreground selection, the functional marker, Pi54 MAS specific to the Pi54 gene, was used to identify the plants carrying heterozygous alleles for the targeted resistant gene at each back-cross generation. The Pi54 MAS is an InDel targeting a 144-bp insertion/deletion polymorphism between the resistant and susceptible allelic sequences [21]. In addition, its co-dominant nature perfectly co-segregates the blast resistance gene in a 1:2:1 ratio in the segregating mapping population. Earlier, studies also proved that the Pi54 MAS has an accurate genotyping ability (>98%) and predicts the allelic status efficiently in many rice cultivars, and it is routinely used in the many MABB programs [12,21,43] globally. For background analysis, 156 SSR markers distributed across the rice genome were employed to analyze the recovery of recurrent parent genome (RPG) in both improved lines. The stringent phenotypic and background selection from BC 1 , BC 2 , and BC 3 populations recovered 90.3 to 91.1% recurrent parent genomes at BC 3 F 3 generation of both populations. While improving the restorer line for blast resistance, it is absolutely necessary that the back-cross-derived lines should also possess the fertility-restorer genes, Rf3 and Rf4. Closely linked gene-specific markers, DRRM-RF3-10, DRCG-RF4-8, and RM6100, were used for the identification of fertility restoration genes Rf3 and Rf4 along with target gene (Pi54) in the segregating population.
In this study, we used a higher number of parental polymorphic markers with a better coverage per chromosome for genetic background selection. This has certainly resulted in limiting the linkage drag to the regions closer to the target genes [44]. Through stringent phenotypic and background selection at each back-cross generation, we have identified intercrossed lines with increased productive tillers, grain filling, and 100-grain weight than the recurrent parent. This indicates that utilization of the improved blast-resistant restorer lines in the hybridization program would be of great advantage currently and in the future.
Our research group has previously released two rice hybrids, CO RH3 and CO RH4, utilizing the restorers, CB 87 R and CB 174 R [45]. These hybrids played a significant role in improving the rice production from 2010 onwards at both the regional and national level. On the contrary, there were continuous complaints from the farmers about susceptibility of the above hybrids to the blast disease. Earlier, we improved three popular cytoplasmic male sterile lines, CO 2A, CO 23A, and CO 23B, for bacterial blight resistance through functional markers and developed stable hybrids with high yield and BB resistance [19]. Our improved lines, especially 5-3-7-6 and 5-3-8-10 with the Pi54 gene, and similar phenotypic characters of restorer lines have brought back the full genetic potential and yield of two hybrids, CO RH3 and CO RH4. In addition, the improved restorer lines and already improved cytoplasmic male sterile lines will be useful in the development of new blast and bacterial blight-resistant hybrids in the near future.
In conclusion, MABB breeding techniques are much effective than conventional breeding, which helps in the quicker and maximum recovery of genetic backgrounds along with target genes with minimum linkage drag. We successfully introgressed a single dominant blast resistance gene, Pi54, in the susceptible restorer rice lines viz. CB 87 R and CB 174 R, along with the two fertility-restorer genes Rf3 and Rf4. The introgression of Pi54 remarkably increased the blast resistance level and maintained the yield levels in blast endemic regions. The breakdown of resistance is a natural phenomenon and the only alternative is to introgress the various allelic forms of Pi54 in a single genetic background. We expect that the improved lines will provide blast resistance at least for a foreseeable period in Southern India. Meanwhile, our Pi54 allelic pyramided restorer lines will be ready for long-term use in hybrid rice production.