In the hybrid rice system, several types of Cytoplasmic Male Sterility (CMS) systems are known, but only a few are used for commercial hybrid seed production (Sankar Deepa et al., 2019). The Wild Abortive (WA), Bao Tai (BT) and Honglian (HL) are commonly used for commercial hybrid seed production. Amongst the three, WA system is the most stable, showing complete pollen sterility and is the most commonly used CMS system (Shinjyo and Omura, 1966). The genetics and inheritance of fertility restoration (Rf genes), particularly in WA-CMS system has been significantly reviewed (Bharaj et al., 1995; Zhang et al., 1997; Govindaraj and Virmani, 1988; Yao et al., 1997; Ali et al., 2003). Although several restorer genes have been reported in various restorer lines, most of the time either or both dominant restorer alleles i.e., Rf-3 and Rf-4 were suspected to be responsible for the fertility restoration in the majority of restorer lines for WA-CMS lines (Tan et al., 1998). The use of CMS system in developing hybrids in crops is possible only when stable CMS lines, their effective maintainer and restorer lines with high pollen load and spikelet fertility are identified (Hasan et al., 2015). The CMS lines introduced from elsewhere may not be well adapted in the target area. Therefore, it becomes imperative to identify maintainers and restorers from well adapted local germplasm for the development of component lines in a hybrid program (Rajkumar and Ibrahim, 2019). Traditionally, maintainers and restorers are identified through evaluating the pollen or spikelet fertility or both of the testcross hybrids (Sutaryo et al., 1989). Traditional method of identifying maintainers and restorers is tedious and time consuming. Revathi et al., (2013) have reported efficiency of 85-92% of tightly linked markers of Rf-3 and Rf-4 genes in identifying maintainers and restorers. The use of molecular markers linked to Rf genes can enhance the selection efficiency, save time and avoid the complications associated with phenotype-based screening (Singh et al., 2025). Identification of locally adapted maintainers and restorers which show complete sterility and consistently high degree of restoration of CMS lines would be of great value in a commercial hybrid breeding program. Thus, the overall objective of the study was to evaluate the efficiency of restorer genes and identify locally adapted maintainers and restorers using molecular markers and testcrossing for development of rice hybrids suitable to Punjab.

In the present study, a set of 79 promising rice genotypes was used for molecular characterization of Rf-3 and Rf-4 genes. Candidate gene-based markers DRRM-RF3-10 for Rf-3 and DRCG-RF4-14 for Rf-4 genes (Balaji et al., 2012) were used. Twenty-one different CMS lines were crossed in various combinations with 79 promising lines to generate 109 test crosses in kharif 2024 (Table 1) in the Rice experimental farm, Punjab Agricultural University (PAU). These test crosses were evaluated in the field along with the checks viz., PR 126, to identify potential maintainers and restorers in kharif 2025. The material was planted in an augmented design comprising of 8 blocks. Observation was recorded on 10 randomly selected plants for various characters such as plant height and days 50% flowering. Pollen fertility and spikelet fertility were recorded to identify maintainers and restorers. Genotypes categorized as Maintainers exhibit pollen fertility between 0-1% and spikelet fertility of less than 5%. Genotypes classified as Partial Maintainers show pollen fertility ranging from 1.1% to 50%, with spikelet fertility between 5% and 20%. In contrast, Partial Restorer genotypes achieve pollen fertility levels from 50.1% to 80% and exhibit spikelet fertility from 20% to 70%. Finally, genotypes categorized as Restorers display pollen fertility greater than 80% and spikelet fertility exceeding 70% (Virmani et al., 1997; Mallikarjuna et al., 2013).

Screening for restorer genes (Rf-3 and Rf-4) using candidate gene-based markers
 
The marker DRRM-RF3-10 amplified 210 bp for Rf3 allele and 155 bp for rf3 allele, likewise DRCG-RF4-14 amplified 800 bp for Rf-4 allele and 900 bp for rf4 (Fig 1 and 2). Out of 79 elite genotypes, 8 (10%) had dominant functional allele for Rf-3(Rf3rf4), 44 (56%) had dominant allele Rf-4 (rf3Rf4), 8 (10%) positive for both the alleles (Rf3Rf4) and 19 (24%) had recessive allele for both genes (rf3rf4). The Rf-4 genes were predominant than Rf-3 in the studied material. The reason for this probably lies in the pedigree relationship. The parents with rf3rf3/Rf4Rf4 genotype have been used for developing most of these lines (Katara et al., 2017). Rf4 gene on an average, showed higher efficiency in restoring male fertility as compared to Rf3 gene. Relatively, genotypes carrying only Rf4 when crossed with different CMS lines restored the fertility fully to partially in more than 70% testcross hybrids, while genotypes carrying only Rf3 gene when crossed with the CMS line could restore fertility fully or partially only in 54% of the testcross hybrids. Interestingly, it was also observed that, absence of Rf3 and Rf4 genes did not affect the restoration ability of some lines such as in PCMS 18A/HHZ 23-DT16-DT1-DT1, PCMS 22A/IR 12L144 and PCMS 23A/ HHZ 4-SAL-12-LI1-LI1 which had spikelet fertility of 87.61%, 82.42% and 77.99% and lines like IR 62829A /NVSR-V-2057, PCMS 27A /HHZ17-DT6-SAL3-DT1and PCMS 31A /HHZ 3-SAL 6-Y1-Y1 acted like maintainer despite the presence of Rf-genes indicating there might be other minor/modifier genes present in the pollinator which restores the fertility (Yadav et al., 1997).




 
Evaluation of the test cross nursery
 
Based on pollen fertility and spikelet fertility, these genotypes were grouped into maintainers, partial maintainers, partial restorers and restorers. Pollen fertility ranged from 0% (completely sterile) to 100% (completely fertile), while spikelet fertility ranged from as low as 0% to as high as 93.33%. The results of pollen and spikelet fertility showed that among the testcrosses, 17 were completely sterile, 37 were completely fertile while rest showed partial fertility. It was also observed that some testcrosses with medium pollen fertility, such as IR 62829A/HHZ 17-DT6-SAL3-DT1, PCMS 12A/HHZ 22-Y3-DT1-Y1, PCMS 12A/HHZ 2-SUB 2-DT1-DT1, PCMS 13A/HHZ 15 SAL 13-Y3, PCMS 23A/IR 11A318 also showed high spikelet fertility. This could be explained as each spikelet may receive many pollens but one fertile pollen would be sufficient to fertilize the spikelet (Joshi et al., 2007). Hence, for identification of restorer lines, spikelet fertility is potential trait than pollen fertility (Babu et al., 2010).
       
An interesting observation was made on fertility restoration behaviour such as genotype HHZ17-DT6-SAL3-DT1, behaved as restorer for PCMS 10 A and IR 62829A giving spikelet fertility of 69.99% and 85.46%, respectively but as maintainer for BOA and PCMS 27A showing complete sterility in testcross hybrid, whereas PR 124 behaved as maintainer for PCMS 27A and PCMS 31A showing complete sterility and as restorer for PCMS 11A giving 76.99% spikelet fertility. The genotypes showed differential fertility behaviour with different CMS lines. The variation in behaviour of fertility restoration may be attributed to varied penetrance and expressivity of restorer genes in the different female backgrounds, as also reported in earlier studies (Wilson, 1968; Virmani et al., 1997). This could also be due to the effect of the environment on the restoration ability of genotypes (Virmani and Edwards, 1983) or it may also be attributed to the presence of modifier genes (Ngangkham et al., 2010). Similar results have been reported in the study conducted by Hasan et al., (2015) and Hariprasanna et al., (2005), wherein the same genotypes behaved differently for different CMS line.
       
Thirty-seven restorers were identified with the highest spikelet fertility of 93.33% recorded for testcross hybrid PCMS 23A /BP 10618F-BB8-18 followed by PCMS 19A /Peeli Pusa 1 with 90.20%, which were comparable with the checks. Other top performing restorers were HHZ 24-DT11-LI1-LI1 and HHZ 22-Y3-DT1-Y1 with 89.23% and 87.99% of spikelet fertility in their respective testcross PCMS 12A /HHZ 24-DT11-LI1-LI1 and PCMS 12A /HHZ 22-Y3-DT1-Y1. The identified restorers can be put into restorer improvement program by crossing with the landraces or other elite restorers and selection in advanced generation, which will lead to locally adapted restorers for hybrid rice production. The potential maintainers identified include IR 75478-282-5-1, IRMT 4402, PAU 5216-10-1-3-1, NVSR-V-2057, PR 121, IRRI 123, HHZ 15 SAL 13-Y3, HHZ17-DT6-SAL3-DT1, HHZ 14-SAL 19-Y1, R-RHZ-R56, 2k3-322-5-1-1-18-1-1-1-2-1-1, 2k3-322-5-1-1-9-1-1-1-1-1-1, HUANGHUAZHAN, PAU 4320-21-1-3-1, PR 124, HHZ 4-SAL12-LI1-LI1, HHZ 21-Y4-Y2-Y1, HHZ 3-SAL 6-Y1-Y1 and HHZ 24-DTI1-LI1-LI. These identified maintainers showed complete pollen sterility to <5% pollen fertility and 0 to <10% spikelet fertility in the respective test cross. These identified maintainers as well be put into conversion backcross breeding programs for its diversification by backcrossing with their respective test cross hybrids to develop new CMS lines. The maintainers and restorers identified can be directly used in strengthening the hybrid breeding programme in PAU for Punjab.

The comprehensive screening of promising genotypes for restorer genes Rf-3 and Rf-4, combined with the evaluation of test cross nurseries, has yielded significant insights into the potential for enhancing hybrid rice production in Punjab. The results indicate a predominance of the Rf-4 gene, which demonstrated higher efficiency in restoring male sterility compared to Rf-3. A total of 37 restorer lines, 17 maintainers, 17 partial maintainers and 18 partial restorers were identified. In conclusion, the findings of this study lay a robust foundation for the ongoing hybrid rice improvement program at Punjab Agricultural University (PAU) by integrating high-performing restorer and maintainer lines.

We gratefully acknowledge the financial support from the Indian Council of Agricultural Research (ICAR) and Punjab Agricultural University for providing all the laboratory and field facilities.
 
Disclaimers
 
The views and conclusions expressed in this article are solely those of the authors and do not necessarily represent the views of their affiliated institutions. The authors are responsible for the accuracy and completeness of the information provided, but do not accept any liability for any direct or indirect losses resulting from the use of this content.

The authors declare that there are no conflicts of interest regarding the publication of this article. No funding or sponsorship influenced the design of the study, data collection, analysis, decision to publish, or preparation of the manuscript.