Conversion of partial restorer Swarna into restorer by transferring fertility restorer Rf gene(s) through marker assisted back cross breeding (MABB) in rice

The major constraints in hybrid rice breeding are availability of limited number of parental lines with specific desirable traits and lower frequency of restorers among elite breeding lines. The popular, high-yielding mega-rice variety Swarna, has been identified to be a partial restorer (as it has only one of major fertility restorer genes, Rf4) and hence cannot be utilized directly in the hybrid rice breeding. To convert the partial restorer to complete restorer, a cross was made between Swarna and a stable restorer KMR3R possessing Rf3 and Rf4 genes and developed BC1F5 and BC2F4 populations by marker-assisted back cross breeding (MABB). The SSR marker DRRM-RF3-10 linked to Rf3 gene located on chromosome 1, clearly distinguished restorers from partial restorers. All the improved lines of Swarna possessing Rf3 and Rf4 genes showed complete fertility restoration in test crosses with higher grain yield heterosis. Few rice hybrids developed by using converted restorers were evaluated in multi location testing under the All India Co-ordinated Rice Improvement Project (AICRIP). The results indicated that new rice hybrids expressed higher heterosis with matching grain quality attributes like Swarna. This study provides significantly novel and relevant restorers to enhance and economize future hybrid rice breeding programs.

The restorers with strong restoration ability have two major genes along with modifier genes and a restorer with semi-restoring ability have either one of the two major genes 7 . Rice fertility restoration is governed by two independent dominant genes and one of the genes appeared to be stronger in action 9 than the other. Bharaj et al. 10 reported that fertility restoration is governed by two dominant genes, the restorers with dominant alleles of the two genes in homozygous or heterozygous condition will be fully fertile and the genotypes having dominant alleles of one of the two genes in homozygous or heterozygous condition but homozygous recessive alleles of the other gene will behave partially sterile or partially fertile and vice-versa.
The major loci restoring the fertility has designated as Rf3 and Rf4 and have been mapped on chromosome 1 and 10 using Zhenshan 97A near-isogenic lines (NILs) mapping population 11,12 . The chromosomal location of Rf genes has been resolved using RFLP (Restriction Fragment Length Polymorphism) markers and it showed that effect of the locus on chromosome 10 is larger than chromosome 1 13 . Mapping of two Rf genes of WA-CMS system on the long and short arm of chromosome 10 using SSLP (Simple Sequence Length Polymorphism) markers was done by Jing et al. 14 . The SSR primer RM258 located on chromosome 10 was found linked to the restorer gene at a distance of 9.5 cM 15 . Fertility restorer gene linked to RM6100 was mapped at a distance of 6-7 cM on chromosome 10 in the restorer lines viz., PRR 78 R, IR 40750 and MTU 9992 16 . The candidate gene based marker, DRRM-RF3-10 associated with Rf3 locus showed maximum selection accuracy in identifying restores in comparison with other reported markers viz., RM10305, RM10318, DRRM-RF3-5 and DRRM-RF3-6 17 .
The process of screening for the trait of fertility restoration involves test crossing with a set of CMS lines and evaluation of hybrids/F 1 for their pollen and spikelet fertility. According to the male fertility of F 1 plants, the test lines can be classified as maintainers, restorers, partial restorers and partial maintainers 9 . Molecular markers linked to the fertility restoration trait are useful for evaluation of large number of germplasm/breeding lines to identify restorers in rice without progeny testing 18 . Based on molecular screening with markers linked to Rf4 and Rf3 genes 19 reports that lines possessing both the Rf genes showed higher fertility than the genotypes containing Rf3 or Rf4 individually. In hybrid rice breeding, elite lines and released varieties from varietal improvement programme are being tested for their fertility restoration ability by crossing with the CMS lines to identify restorers and maintainers. The restorer frequency among indica lines is only 40% 20 . The majority of the popular, mega varieties released in India like Swarna, Samba Mahsuri, and MTU1010 were found to be partial or incomplete restorers and hence cannot be utilized as such to produce experimental hybrids.
The major issues in hybrid rice breeding are lack of parental lines with desirable specific traits and lower frequency of restorers and maintainers among elite breeding lines. One of the major constraints, which limit the spread of hybrid rice area in shallow low lands and coastal areas, is non availability of long duration hybrids, which can mature in 145-150 days or more. The popular mega rice variety, Swarna has been widely adopted by farmers for its high yielding ability and adaption under low input conditions, shallow low lands. It is cultivated in an approximate area of 4.5 M ha 21 , occupying more than 30% of total rice area in eastern India 22 . Swarna possesses good grain and cooking quality with desirable grain type. However, as mentioned earlier, it cannot be utilized in hybrid breeding as it is a partial restorer. Therefore in the present investigation, an attempt was made to convert the partial restorer Swarna to a complete restorer by transferring major fertility restorer gene(s) through marker-assisted back cross breeding (MABB), and demonstrated the complete fertility restoration ability of the improved lines and developed new rice hybrids with attributes like Swarna.

Results
fertility restoration status of parents. To study the fertility restoration status of donor and recurrent parents, F 1 hybrids were produced by crossing Swarna and KMR3R with four CMS lines viz., APMS 6A, IR79156A, IR68897A and IR58025A. The F 1 hybrids were evaluated for their pollen and spikelet fertility during both kharif (wet) and rabi (dry) seasons (Table 1). The average pollen and spikelet fertility percentage of partial restorer Swarna was observed to be less than 80%, whereas KMR3R showed more than 90% fertility restoration when crossed with different A lines (Fig. 1).  Table 3). The chromosomal location map of polymorphic markers was prepared by graphical genotypes (GGT ver. 2.0) and same has been presented in the Fig. 3. In   In the backcross generations the phenotypic traits viz., flowering duration, plant height, number of productive tillers and grain yield/plant were recorded and plants were visually selected for heavy pollen load with longer anther filament during anthesis, strong culm with specific features of Swarna like brown glume colour, stay greenness with desirable medium slender grains and were advanced to next generation. More emphasis was given to phenotypic selection for advancing plants to next generation along with background selection. One superior BC 1 F 1 plant with maximum recurrent parent genome (80%) along with better agro morphological traits (i.e. similar to Swarna) was backcrossed with Swarna to produce BC 2 F 1 seeds. Simultaneously, the selected BC 1 F 1 plant with maximum recurrent parent genome was selfed to produce BC 1 F 2 seeds. By foreground selection in BC 2 F 1 and BC 1 F 2 generation plants carrying Rf3 gene in homozygous and heterozygous conditions were identified (Fig. 4). A total 65 plants (22 BC 2 F 1 , 43 BC 1 F 2 ) possessing superior phenotypes were subjected to background selection analysis. BC 2 F 1 and BC 1 F 2 plants with maximum recurrent parent genome (RPG) were selfed to produce BC 2 F 2 and BC 1 F 3 seeds. The BC 2 F 2 and BC 1 F 3 plants were raised and subjected to foreground selection for the presence of Rf3 gene. Marker-assisted screening resulted in identification of 35 plants to carry Rf3 in homozygous condition in BC 2 F 2 and 40 plants in BC 1 F 3 generation. The identified homozygous plants were selfed to raise BC 2 F 3 and BC 1 F 4 generations. Phenotypically superior plants were identified based on morphological traits at BC 2 F 3 and BC 1 F 4 generations and were selfed to produce the subsequent generation seeds. BC 2 F 4 and BC 1 F 5 lines were evaluated in three rows and plants were subjected to stringent phenotypic selection and agro morphological evaluation.  (Table 4) and frequency distribution of each traits are presented in (Fig. 5).
Plant height of selected improved lines ranged from 79.6 cm to 113.7 cm. Swarna was 83.1 cm tall, whereas KMR3R possessed a plant height 115.7 cm. The improved lines possessing more than 100 cm plant height were preferred for using it as restorers, most of the lines possessed taller plant height as compared to recurrent parent Swarna, which is one of the most desirable trait for an ideal restorer parent for utilizing them in hybrid rice breeding. Number of productive tillers and panicle length in the improved lines ranged from 7 to 14 and 21 cm to 26.4 cm respectively, whereas both donor and recurrent parent had 11 productive tillers and possessed panicle length of 18.5 cm (Swarna) and 23.9 cm (KMR3R). The panicle length of converted restorer lines were significantly longer (>7 cm) than Swarna. With respect to mean grain yield per plant of selected backcross derived lines, ranged from 19.0 to 25.0 gm, which is significantly higher than the recurrent parent, Swarna (15.28 gm). The test weight of Swarna (1000 grain weight) was 15.9 gm with short bold grains, whereas KMR3R possessed 26.5 gm of test weight with long bold grain type. The converted restorer lines test weight ranged from 15.1 to 21.9 gm with the grain type of short slender to short bold. Days to 50% flowering (DFF) of selected backcross derived lines ranged from 91 to 121 days (  www.nature.com/scientificreports www.nature.com/scientificreports/ evaluation of experimental hybrids developed by utilizing converted improved restorer lines. The phenotypically superior lines of BC 2 F 4 and BC 1 F 5 generations were crossed with two CMS lines namely APMS 6A and CRMS 32A to develop experimental rice hybrids. The improved lines possessing both Rf3 and Rf4 along with the maximum recovery of recurrent parent genome (RPG) and possessing plant height of more than 100 cm were crossed with above mentioned CMS lines to produce F 1 hybrids. F 1 hybrids were evaluated for pollen and spikelet fertility percentage and grain yield heterosis along with different duration checks. Based on days to 50% flowering three hybrid groups viz., early (<100 days), medium (101-110 days) and late (>110 days) were constituted for evaluation. (Table 5a-c). early duration hybrids. In early duration hybrids, pollen and spikelet fertility percent ranged from 93.23 to 100% and 85 to 95.3%, respectively. Grain yield per plant ranged from 35 to 42 grams and plant height was measured to be 83.75 to 90.60 cm. The early hybrid US 314 was used as a check hybrid for estimating standard heterosis. It was found that restorer line RP5934-70 showed 40% standard heterosis. Hence this could be a potential restorer for developing early duration rice hybrids with characteristic features of Swarna (Table 5a).

Medium duration hybrids.
In medium duration hybrids, the pollen and spikelet fertility percent ranged from 82 to 100% and 85 to 95.3%, respectively. The single plant yield of hybrids ranged from 32 to 55 grams with the plant height of 83.75 to 90.60 cm. The medium duration hybrid HRI 174 was used as a check for estimating standard heterosis. It was found that improved restorer lines viz., RP5934-73, 78 and 80 lines showed more than 48% and RP5934-71 more than 35% of standard heterosis, while crossing with APMS 6A. The converted line RP 5934-100 showed a standard of heterosis of more than 35%, when it was crossed with CRMS 32A. These five restorers can be utilized for the developing medium duration hybrids with higher heterosis (Table 5b).
Late duration hybrids. Eight late duration hybrids were evaluated for pollen and spikelet fertility, yield and yield related traits. The hybrids derived from improved lines viz., RP5934-81 and 89 showed a standard yield heterosis of more than 40%, when they were crossed with CRMS 32A. These restorers could be potential restorers for developing late duration rice hybrids (Table 5c).
Multi-location evaluation of the newly derived hybrids. Based on the performance of converted improved restorer lines in the research station trial, two high yielding hybrids namely IIRRH-111 and IIRRH-112 were nominated for multi-location evaluation in AICRIP (All India Co-ordinated Rice improvement project)-IHRT trial during kharif 2016. In the case of AICRIP-hybrid rice network system, the experimental hybrids developed by public and private sectors were evaluated in Initial Hybrid Rice Trials (IHRT) in different zones in India. The hybrid entries with a yield superiority of more than 5% and 10% over the best hybrid and varietal checks either on over all mean basis or on zonal mean basis are promoted from IHRT to next stages of testing in Advance Varietal Trial 1 (AVT 1) and AVT 2 23 .
The rice hybrid IIRRH-112 showed a superior performance over the checks (AICRIP progress report, 2016) and registered yield superiority over checks during kharif 2016 multi-location testing. With respect to quality data collected from the AICRIP progress report (2016), this hybrid possessed short bold grains as that of Swarna and preferred intermediate amylose content (26.22%) with 60% high head rice recovery. In AICRIP testing rice hybrids were evaluated for not only yield and also for their grain quality characters. The grain quality characters of hybrids developed by crossing CMS lines with improved restorers are presented in  www.nature.com/scientificreports www.nature.com/scientificreports/ restorers, which showed superior performance at IIRR station trial, were nominated for multi-location testing and are presently under AICRIP evaluation.

Discussion
Hybrid rice technology is likely to play a pivotal role in increasing the rice production. Hybrid rice has been commercialized in countries like China, India, Vietnam, Philippines, Bangladesh, Indonesia, Pakistan, Ecuador, Guinea and the United States of America 24 . In India, as a result of concerted efforts by both public and private sectors, a total 102 hybrids have been released for commercial cultivation in the country. In spite of having great potential to enhance rice production and productivity, area expansion under hybrid rice has not increased significantly in the last few years due to major constraints like non availability of long duration hybrids suitable for shallow lowlands and coastal areas, poor grain and cooking quality, marginal heterosis, higher seed cost, non-involvement of public sectors in hybrid rice seed production 25 . The popular, high yielding, mega rice variety Swarna, with good grain and cooking quality, cannot be utilized as such in hybrid breeding due to their partial restoration of fertility of CMS lines (pollen and spikelet fertility is less than 80%) ( Table 1). The effective restorers are expected to have more than 90% of pollen and spikelet fertility to develop new rice hybrids 26 . While considering the above stated constraints and wider scope for large scale adoption of rice hybrids, particularly the ones suited for rainfed shallow low lands and costal area, the present study was planned and carried out with an objective to develop restorers with good grain and cooking quality in the background of mega Indian rice variety Swarna, so that new hybrids can be developed for the long duration segment with better grain and cooking quality with higher level of heterosis.
Molecular screening for the presence of Rf genes with the reported SSR markers ( Table 2) results indicated that partial restorer Swarna showed the presence of Rf4 gene and the absence of Rf3 gene. The marker DRRM RF3-10 located on chromosome 1, could clearly differentiated restorers from partial restorers like Swarna, BPT 5204 (Samba Mahsuri) and Improved Samba Mahsuri, Kavya and MTU 1010 (Fig. 2) in our study. Hence we report here that SSR marker DRRM RF3-10 located on chromosome 1 is a very useful marker in distinguishing partial  www.nature.com/scientificreports www.nature.com/scientificreports/ restorers from restorer and it may be useful in marker assisted selection (MAS) for the fertility restoration trait of WA-CMS system.
An earlier study 18 , reported that the stable restorer KMR3R possessing both the major fertility restorer genes Rf4 and Rf3 and this was also confirmed in our marker analysis. Hence, KMR3R was utilized as a donor parent for Rf gene(s) particularly Rf3 gene located on chromosome 1. Backcross breeding is the most commonly used method for incorporating any essential genes into a rice cultivar. Backcross breeding along with marker-assisted foreground and background selection contributes immensely to accelerate recurrent parent genome (RPG) recovery 27 .
Marker-assisted backcross breeding which includes MAS for the target gene(s) known as foreground selection and MAS for the recovery of recurrent parent genome known as background selection 28 considered to be an efficient strategy for transferring targeted specific genes to elite lines. MABB is a most effective breeding strategy in rice for improving simple and complex traits. Some of the popular rice varieties released for cultivation in  www.nature.com/scientificreports www.nature.com/scientificreports/ India through MAS/MABB for biotic and abiotic stress tolerance are Pusa Basmati 1 with bacterial blight (BB) resistance genes xa13 and Xa21 involving single back cross generation followed by selfing and pedigree based selection 29 and Improved Samba Mahsuri with three BLB resistance genes viz., xa5, xa13 and Xa21 involving four back crossing 30 . Swarna Sub1 with Sub1 QTL for the submergence tolerance 31  To the best of our knowledge, the present study is the first attempt through MABB approach for converting partial restorer Swarna into complete restorer by transferring Rf3 gene. We could recover Swarna genome in the backcross derived lines of converted restorers of BC 1 F 5 and BC 2 F 4 generations by 80% and 92.3%, respectively. The agro morphological evaluation of converted restorers (RP5934-66 to RP 5934-100) in the background of Swarna exhibited desirable traits for an ideal restorer lines viz., improved plant height, longer panicles, more number of productive tillers with heavy pollen load along with traits specific for Swarna, like brown glume or golden husk colour with stay green type etc (Fig. 6). The restorers should be ideally be taller than that of their CMS line for easy and higher hybrid seed production, as taller restorer parents facilitates easy pollen deposition on the stigma lobes of CMS lines. In our study BC 1 F 5 lines were preferred as restorers than BC 2 F 4 lines because of desirable plant height. BC 2 F 4 lines expressed shorter plant height similar to the recurrent parent Swarna. We were able achieve target trait of complete fertility restoration in BC 1 F 5 generation by restricted marker assisted back cross breeding involving only single back cross followed by selfing and selection. To retain some of the useful traits from donor parent 29 , restricted marker assisted backcross breeding strategy has been used for improving basmati rice variety, Pusa Basmati 1 for bacterial blight resistance.
Longer panicle with more number of grains is also preferred traits of restorer for supplying more pollen grains for hybrid seed production. More number of productive tillers results in higher pollen shedding potential per sq.m. during peak anthesis period. Apart from restoration ability converted restorers were observed to be high yielding than Swarna indicating potentiality of these lines for releasing it as variety. The test crosses evaluation of F 1 hybrids based on pollen and spikelet percentage confirmed that Swarna is a partial restorer and KMR3R is a complete restorer. The partial restorer Swarna possessing only Rf4 gene could restore the fertility of WA-CMS lines with an average of 70% pollen and spikelet fertility. This level of pollen and spikelet fertility is not sufficient warrant deployment of Swarna as an effective restorer in hybrid rice breeding. The effective restorers are expected to have more than 90% of pollen and spikelet fertility. The hybrids produced using converted Swarna breeding lines with Rf3 gene showed 15-20% of more pollen and spikelet fertility than the partial restorer Swarna. This clearly demonstrates that that Rf3 gene may be contributing 15 to 30% for the fertility restoration trait of WA-CMS and Rf4 may be contributing about 70% for the same trait and presence of both the genes can bestow complete fertility restoration 13,[33][34][35][36][37][38] .
An effective restorer, IR64 carrying both Rf4 and Rf3 genes showed 85.4% spikelet fertility and Basmati 370 partial restorer carrying only Rf4 gene showed 59.3% of spikelet fertility explaining that presence of both the Rf genes will enhance pollen and spikelet fertility percentage 19 . The results obtained from the present study confirms that presence of only Rf4 gene in restorers may result in partial restoration of fertility in hybrids and presence of both Rf4 and Rf3 genes will certainly results in complete restoration. The rice varieties having both the restorer www.nature.com/scientificreports www.nature.com/scientificreports/ genes expressing higher fertility restoration has been reported 11,39 . The restorer lines carrying two dominant genes for fertility restoration should be preferred to develop rice hybrids using WA-CMS lines and hence the resultant hybrids are expected to have higher fertility and thereby higher heterosis 40,41 .
As cited above most of the previous reports were based only on molecular screening for the presence of Rf4 and Rf3. Our study is the first instance of using marker-assisted selection for targeted introgression of fertility restorer gene(s) in order to develop improved parental lines in the background of an elite Indian mega-variety of rice and thereby confirming the major and minor role of Rf4 and Rf3. The grain quality characters of the converted (i.e. improved) restorer lines in the genetic background of Swarna were similar to that of the original parent. In a panel test for grain quality traits conducted at ICAR-IIRR, Hyderabad, it was observed that, the Rf3 gene introgressed lines were indistinguishable from Swarna in terms of color of paddy, grain size and shape.   Further, replicated field trials that were carried out at twelve different locations across India under the all India coordinated rice improvement project (AICRIP), showed that yield levels of the improved lines were significantly higher than Swarna and the check lines indicating that there is no apparent yield penalty associated with presence of the Rf3 gene 3 . Considering these points, improved version of restorers developed in the genetic background of Swarna would be of great use in hybrid rice breeding for developing late duration hybrids. Large scale seed production of potential hybrids involving long duration improved restorers with different WA-CMS lines are under pipeline for multi location testing. Some of the improved lines were also demonstrated their potentiality as higher heterotic hybrids in the early and medium duration category (Tables 5a,b). Introgression of fertility restorer genes without recovery of other characters of recurrent parent would have been a futile exercise, if the developed lines do not meet expected grain and cooking quality with higher yields. We feel that this was made possible because of the extensive phenotypic selection and also because of the use of molecular markers for background selection. The selected BC 1 F 5 and BC 2 F 4 lines possessed the desirable traits of restorers like intermediate or tall plant height, early to late duration, heavy pollen load, optimum productive tillers and long panicles (Fig. 6). This was achieved by stringent phenotypic selections in each backcross generations as demonstrated in the study 42 . As indicated earlier, two rice hybrids namely IIRRH-111 and IIRRH-112 developed by using these improved lines have undergone multi location testing under the All India Coordinated Rice Improvement Project-Initial hybrid rice trial during kharif 2016. The rice hybrid IIRRH-112 was found promising with yield advantage of 10% superiority over hybrid and varietal checks in zone III (eastern), whereas Swarna is a very popular rice variety and occupying major area under cultivation.
To the best of our knowledge, this is the first attempt to convert partial restorer to complete restorer by transferring Rf genes through marker-assisted backcross breeding. Thus, in the present study, we have successfully developed restorers with characteristic features of popular mega Indian rice variety Swarna for utilization in hybrid rice breeding, specially for development of late maturity hybrids.

Materials and Methods
The experimental materials consist of Swarna (MTU 7029) a popular mega rice variety, which occupies larger area under cultivation in India, derived from the cross Vasistha x Mahsuri and is a partial restorer and it was utilized as the recurrent parent for improvement of its fertility restoration ability. The donor parent for the fertility restoration trait/Rf gene(s) was KMR3R (Jaya/IR29723-143) a restorer parent of popular rice hybrid KRH-2. KMR3R has been identified to carry both fertility restorer genes Rf 4 and Rf 3 genes 18 . To determine the fertility restoration status of recurrent and donor parent, crosses were made with four CMS lines namely APMS 6A, IR 79156A, IR 68897A and IR 58025A to produce F 1 hybrids during rabi 2011. These F 1 hybrids were evaluated for their pollen and spikelet fertility and grain yield heterosis during two rice crop seasons viz., kharif 2012 (i.e. wet season 2012) and rabi 2013 (i.e. dry season 2013) to confirm their fertility restoration status. pollen fertility. Pollen fertility was measured using anthers collected from the spikelets at one or two days before anthesis. The anthers from each spikelet were smeared in a drop of 1% Iodine-potassium iodide (I 2 -KI) solution 43 , on a glass slide and three randomly selected microscopic fields were counted. Stained, well filled and round pollen grains were counted as fertile, while unstained, shrivelled and empty pollen grains were considered as sterile. Pollen fertility was calculated and expressed in percentage as given below   Table 6. Grain quality characteristics of hybrids developed utilizing improved restorers along with checks. www.nature.com/scientificreports www.nature.com/scientificreports/ Spikelet fertility. The panicles that emerged from the primary tiller were bagged before anthesis to avoid out crossing and the number of filled grains and chaffs in the panicle were counted at the time of maturity. The ratio of filled grains to the total number of spikelets was expressed as spikelet fertility as given below: Spikelets fertility (SPF)% Number of filled spikelets in the panicle Total number of spikelets in the panicle 100 = × Plants were classified into four classes based on spikelet fertility percentage, namely, fertile (more than 85% spikelet fertility), partially fertile (50-85%), partially sterile (21-50%) and completely sterile (0-20%) Molecular screening for the presence of Rf genes. DNA isolation and PCR conditions. The total genomic DNA was isolated from the 20 days old transplanted seedlings according to the procedure reported by 44 . PCR analysis was carried out using 30-50 ng DNA as template, 5 pmoles of each primer, 0.05 mM dNTPs, 10X PCR buffer (TaKaRa Taq ™ DNA Polymerase) and 1 U Taq DNA polymerase in a total volume of 10 µl. Template DNA was initially denatured at 94 °C for 5 min followed by 35 cycles of PCR amplification with the following parameters: a 30 s denaturation at 94 °C, a 30 s annealing at 55 °C and 1 min of primer extension at 72 °C. A final extension was done at 72 °C for 7 min. The amplified products were electrophoretically resolved on a 3% agarose gel containing 0.5 µg/ml of ethidium bromide in 0.5X TBE buffer and visualized under UV and results were documented. The recurrent parent Swarna, donor parent KMR 3 R along with few popular varieties (partial restorers) were screened with 20 reported SSR markers linked to Rf4 & Rf3 genes to identify polymorphic marker between restorer and partial restorers for foreground selection (Table 2) and for background selection microsatellite and InDel markers that are polymorphic between donor and recurrent parents were identified by screening 728 markers distributed throughout the rice genome (Table 3). Using the data from polymorphic SSR markers, a schematic map illustrating the genomic contribution of donor and recurrent parents was prepared using Graphical

Marker-assisted backcross breeding (MABB).
A cross was made between Swarna x KMR3R to improve the fertility restoration trait of the recurrent parent Swarna, during rabi 2011 (i.e. dry season 2011). The backcross breeding procedure followed is presented in the Fig. 7. The F 1 plants of Swarna x KMR3R were screened using DRRM-RF3-10 identified polymorphic SSR markers between donor and recurrent parent. The true F 1 s were then backcrossed with Swarna to generate BC 1 F 1 s, which were confirmed for the presence of fertility restorer gene(s) Rf3 and Rf4 with the help of DRRM-RF3-10 and PPR3 markers (i.e. foreground selection). The plants which were positive for the restorer genes Rf4 and Rf3 were subjected to background selection with a set of 89 identified polymorphic SSR markers to identify a single BC 1 F 1 plant, possessing maximum recovery of the recurrent parent genome (RPG). This plant was selfed to generate BC 1 F 2 s and also backcrossed with Swarna to generate BC 2 F 1 s. The MABB process involving foreground and background selection as explained above was repeated among the BC 2 F 1 plants and the best BC 2 F 1 plant was selfed to generate BC 2 F 2 seeds. The BC 1 F 2 and BC 2 F 2 were analyzed with markers specific for Rf3 and Rf4 to identity homozygous plants. Homozygous BC 1 F 2 and BC 2 F 2 plants were then advanced through pedigree method of selection for further phenotypic evaluation based on duration of flowering (days), plant height (cm), number of tillers, panicle length (cm) and single plant yield (g). Phenotypically superior plants were advanced for further evaluation. evaluation of agro-morphological characters of the backcross derived lines. Thirty days old seedlings at BC 1 F 5 and BC 2 F 4 generations along with the parents were transplanted in the main field with the spacing of 20 × 20 cm and fertilizer dosage of 120:80:60 (N:P:K) kg/ha during kharif 2014. The experimental plots were arranged in alpha lattice design in four blocks with two replications. Standard agronomic practices were followed while raising the rice crop. Data was recorded on randomly selected five plants in each replication for the agronomic traits, viz. flowering duration, plant height (cm), number of productive tillers, panicle length (cm), grain yield per plant, and 1000-grain weight. The plants were visually observed for the following traits viz., heavy pollen load during anthesis, strong culm, non-lodging type and Swarna's specific traits of stay greenness and golden hull colour for selecting superior phenotypic plants as restorers. The data was tabulated and statistically analyzed using standard Microsoft office excel and SPSS package.
Generation of experimental rice hybrids utilizing improved restorers and their evaluation. The phenotypically superior lines from BC 1 F 5 and BC 2 F 4 generations possessing Rf4 and Rf3 genes were crossed with two CMS lines viz., APMS 6A and CRMS 32A to produce F 1 hybrids during kharif 2014 (i.e. wet season 2014). Thirty days old seedlings of experimental hybrids along with checks were transplanted to the field with www.nature.com/scientificreports www.nature.com/scientificreports/ the spacing of 20 × 20 cm in two replications. Data on days to flowering, plant height, productive tillers, panicle length, pollen and spikelet fertility percentage, grain yield and grain yield heterosis were estimated. Seeds of Swarna and selected experimental hybrids were stored for three months after harvesting and grain quality tests were carried out by standard grain quality evaluation protocols (as explained in 30 ) with the rice grains having 12 to 14% moisture content. Superior rice hybrids were identified to produce hybrid seed in larger plots during rabi 2016 (i.e. dry season 2016) and nominated to multi-location evaluation of AICRIP trials during kharif 2016 (i.e. wet season 2016).