Introduction
Approximately 74 million hectares (ha) of cultivable land in India are affected by soil acidity (pH < 5) (Kumar et al., Reference Kumar, Khan, Singh, Ngachan, Rajkhowa, Kumar and Devi2012). Acid soils comprise more than 81% of the total 26.22 million ha land in the north-eastern region of India. Crops in these soils are subjected to low phosphorus (P) availability, and aluminium and iron toxicities. P is one of the most limiting essential mineral nutrients for plants (Kochian, Reference Kochian2012). In acidic conditions, P is fixed to the soil particles in the organic and inorganic forms, and becomes unavailable for plant uptake. In rice, only one major QTL, Pup1 (phosphorus uptake 1), derived from the traditional aus-type rice variety Kasalath has been identified for conferring tolerance to P deficiency (Wissuwa et al., Reference Wissuwa, Wegner, Ae and Yano2003). With the identification of major genes [e.g. phosphorus-starvation tolerance 1 (PSTOL1), dirigent-like gene (PupK20-2)] underlying this locus, it has become clear that these genes could be used for better performance under poor soil conditions (Gamuyao et al., Reference Gamuyao, Chin, Pariasca-Tanaka, Pesaresi, Catausan, Dalid, Slamet-Loedin, Tecson-Mendoza, Wissuwa and Heuer2012). The Pup1 gene is located in an insertion–deletion (InDel) region that is specific to Kasalath (Heuer et al., Reference Heuer, Lu, Chin, Tanaka, Kanamori, Matsumoto, Leon, Ulat, Ismail, Yano and Wissuwa2009). Using marker-assisted selection and backcrossing, the Pup1 QTL region including the PSTOL1 gene has been transferred into two Asian irrigated varieties (IR64 and IR74) and three Indonesian upland varieties (Chin et al., Reference Chin, Gamuyao, Dalid, Bustamam, Prasetiyono, Moeljopawiro, Wissuwa and Heuer2011). The exact mechanism by which these genes facilitate P uptake is yet to be determined.
Our previous study suggested that Sahbhagi Dhan (LR 23) could be used as a potential donor for P-deficiency tolerance under acidic lowland conditions (Tyagi et al., Reference Tyagi, Rai and Dohling2012). Sahbhagi Dhan has been classified as K type (Kasalath-type) based on markers run across the Pup1 locus. Kasalath cannot be used as a donor for P-deficiency tolerance in our field conditions as its performance is not as good as genotypes growing in these acidic soils. Also, the breeding strategy would be simpler with lesser linkage drag if genotypes previously adapted to acidic soil conditions are used as parents in breeding programmes. With the availability of molecular markers, we can at least partially replace and/or complement phenotypic evaluations in the field; therefore, it is of particular value to identify the donors and characterize them for the genes of interest. Previously, we characterized 60 diverse rice genotypes with respect to allelic status across Pup1 using reported locus-specific primers (Tyagi et al., Reference Tyagi, Rai and Dohling2012), and identified potential donors for acidic soils.
In this study, we report the allelic status of the PSTOL1 and PupK20-2 genes across the previously identified donors for acidic soils, including LR 23 by aligning with Kasalath sequence. It is also well known that the cultivated rice has at least five distinct genetic groups (Garris et al., Reference Garris, Tai, Coburn, Kresovich and McCouch2005), each carrying a set of different allelic combinations. So, there is potential to identify new alleles for these reported genes that may further improve the performance of rice in poor field conditions.
To our knowledge, this is the first study assessing the haplotype status of the locus in genotypes adapted to acidic soil conditions.
Materials and methods
Plant material
In the present study, ten genetically diverse rice genotypes (LR 5, LR 11, LR 18, LR 19, LR 23, LR 26, LR 39, LR 55, RDW 125 and UR 29) (Challam et al., Reference Challam, Kharshing, Yumnam, Rai and Tyagi2013), which are being used as parents in our breeding programme for acidic soils, were selected for a detailed study.
Phenotyping in low-P acidic soils
The field experiment was conducted at College of Post-Graduate Studies (Central Agricultural University; CAU), Umiam. The field selected for evaluation in this study did not receive any P supply for the last 4 years and was maintained as P deficient. The average soil pH was determined by calculating the mean of ten soil pH readings taken while the plants were growing in the field. The mean soil pH value was found to be 5.65, and the P level was 6.25 kg/ha. Three-week-old seedlings were transplanted in the lowland field, with a plant-to-plant spacing of 18 cm and a row-to-row distance of 20 cm. Plants were grown under rainfed conditions; however, as this region of India receives regular rainfall during the rice-growing months, there was no drought stress experienced by the plants. On maturity, plant samples (roots and shoots without grains) were harvested and dried in an oven at 65°C for 5 d. After dry weight determination, the samples were ground using a mortar and pestle, and put in tubes. Di-acid (nitric acid and perchloric acid) at the ratio of 3:1 was added (5–10 ml) to each tube and allowed to be digested. The tissue P concentration in 1 mg of the plant sample was determined colorimetrically by the phosphovanadate method (Gupta, Reference Gupta and Purohit2007). Total P uptake was calculated as the product of dry weight and tissue P concentration.
Designing of markers for allele mining
Genomic sequence of the Kasalath Pup1 region from NCBI (AB458444.1) was targeted for the design of markers for allele mining across PSTOL1 and PupK20-2. Gene-specific markers were designed to give amplicon sizes of about 600–700 bp using software Primer3 (http://biotoolsumassmed.edu/bioapps/primer3_www.cgi). The primers CAU46-1, CAU46-2 and CAUK20-2 were designed so as to obtain overlapping nucleotide regions (Supplementary Table S1, available online; Fig. 1(a) and (b)). In addition, the previously reported markers targeting these two genes were also used (Chin et al., Reference Chin, Lu, Haefele, Gamuyao, Ismail, Wissuwa and Heuer2010; Fig. 1(a) and (b)).
Allele mining across the PSTOL1 (phosphorus-starvation tolerance 1) and PupK20-2 (dirigent-like gene) genes. Gene models for (a) PSTOL1 and (b) PupK20-2 and the position of markers in the gene model. Black boxes in the gene model indicate the conserved domain, and grey boxes indicate exons. The starting and end of the gene is indicated by 
and *, respectively. The grey line at the top indicates the amplicon length. Primers designed for this study are indicated by the full arrows, while the dotted arrows represent the markers designed by Chin et al. (Reference Chin, Lu, Haefele, Gamuyao, Ismail, Wissuwa and Heuer2010). PCR amplification across the (c) PSTOL1 and (d) PupK20-2 genes for acidic soil-adapted rice genotypes. Primer names and amplicon sizes are indicated on the left- and right-hand side of the gel picture, respectively, while the names of the genotype are indicated above the gel picture.
Genomic DNA extraction and PCR analysis
DNA was extracted from the fresh leaves of rice genotypes by the CTAB (cetyl-trimethyl ammonium bromide) method (Murray and Thompson, Reference Murray and Thompson1980). Approximately 50 ng of genomic DNA were used as the template for performing PCR analysis in a total volume of 10 μl [0.6 mM of each primer, 2 μl 5 × Promega PCR buffer, 1.5 μl of 25 mM MgCl2, 0.6 μl of 2 mM dNTPs and 0.4 unit of Taq polymerase (D1806; Sigma-Aldrich, St Louis, MO, USA)]. Briefly, a standard PCR was carried out with the following profile: 5 min at 94°C, 32 cycles of 30 s at 94°C, 45 s at 55°C, 45 s at 72°C, and a final extension of 10 min at 72°C. The PCR was also carried out with the marker reported for the Pup1 locus, i.e. PupK46-1 and PupK20-2, using PCR conditions as mentioned in Chin et al. (Reference Chin, Lu, Haefele, Gamuyao, Ismail, Wissuwa and Heuer2010). The ubiquitin housekeeping gene was used as a positive control. PCR products were size fractionated using 1.5% agarose gel (Ultra Pure Agarose 1000; Sigma), stained with ethidium bromide (0.5 μg/ml), and the image was captured by using the BioRad gel documentation unit.
Sequencing and SNP discovery
All accessions produced single amplicon with gene-specific primers when analysed by agarose gel electrophoresis. The amplicons were directly sequenced from the PCR products using forward and reverse primers. The PCR products were sent for sequencing to Amnion Biosciences Pvt Ltd, Bangalore, India. The sequences obtained were aligned using BioEdit version 7.2.5 (Hall, Reference Hall1999). Kasalath genomic sequence was taken as the reference for alignment, and putative SNPs were identified based on sequence homology.
Homology searches were performed using FASTA, and multiple sequence alignment was done through CLUSTAL W using BioEdit version 7.2.5. Nucleotide sequence data reported herein are available in the DDBJ/EMBL/GenBank databases under the accession numbers KM079160–KM079161.
Estimation of nucleotide variation
Nucleotide diversity (π) was calculated using the following formula: total number of SNPs/total length of aligned sequences.
Results
The indica-type donor ‘Kasalath’ from which the Pup1 locus was first identified (Wissuwa et al., Reference Wissuwa, Yano and Ae1998) was used as a positive control for PCRs. The Kasalath (LR 60) amplicon showed 100% sequence identity with the Kasalath sequence deposited in the genebank with accession number AB458444.
Allele mining across PSTOL1
PSTOL1 is a serine/threonine receptor-like kinase belonging to the LRK10L-2 subfamily, but it lacks the amino-terminal extension. Three primer sets were used for allele mining (Fig. 1(a)). New markers, namely CAUK46-1 and CAUK46-2, were used in addition to the marker PupK46-1 reported by Chin et al. (Reference Chin, Lu, Haefele, Gamuyao, Ismail, Wissuwa and Heuer2010). A common overlapping region of 167 bp was present between the two amplicons.
All the markers were dominant in nature, and amplicons were observed in five genotypes (LR 19, LR 23, LR 39, UR 29 and LR 60; Fig. 1(c)). PCR amplicons were not observed for LR 5, LR 11, LR 18, LR 26 and LR 55. For the genotype RDW 125, amplification was observed only for one set of primers. The sequence alignment revealed that genotypes LR 19 and LR 23 carried the Kasalath-type allele (Fig. 2(a)). UR 29 had 23 nucleotide substitutions within the 975 bp exonic sequence (Fig. 2(b)). Nucleotide substitutions were similar with Oryza glaberrima (Tanaka et al., Reference Tanaka, Chin, Drame, Dalid, Heuer and Wissuwa2014). The protein kinase domain was conserved in all the genotypes sequenced. Although there were variations in the coding DNA sequence (CDS) region, there was no change at the protein level for all the genotypes sequenced. The sequencing across the CDS region revealed two distinct haplotypes. LR 19 and LR 23 showed the Kasalath-type haplotype, whereas UR 29 and LR 39 showed a mixed haplotype having alleles similar to both Kasalath and O. glaberrima (Fig. 2(b)).
Sequence variations across PSTOL1 (phosphorus-starvation tolerance 1). Sequence polymorphism across the (a, b) exonic regions and (c) 3′-UTR (untranslated region). The nucleotide sequence of the PSTOL1 allele from LR 60 (Kasalath) (Oryza sativa ssp. indica) was taken as the reference. Position indicates the SNP position in the exon or UTR of PSTOL1. The dark grey, light grey and white colours indicate Kasalath, Oryza glaberrima and novel allele, respectively. Genotypes are listed on the left. Only polymorphic regions of the sequenced DNA fragments are shown.
Upon sequencing of 177 bp of 3′-UTR, a total of four novel nucleotide variations were observed (Fig. 2(c)). One SNP found at position 1027 was common in all the three genotypes (LR 23, UR 29 and LR 39), while the other three were specific to LR 39 and UR 29. The sequencing of 3′-UTR of the gene revealed that UR 29 and LR 39 had four common nucleotide substitutions (Fig. 2(c)). One novel SNP (1027 bp) was found in the 3′-UTR for both LR 39 and UR 29. Our phenotypic data suggested that both LR 39 and UR 29 were susceptible to low-P conditions under lowland conditions (Table 1). These two genotypes carried similar haplotypes in the 3′-UTR but were different in the CDS region.
Nucleotide diversity (π) for the PSTOL1 (phosphorus-starvation tolerance 1) and PupK20-2 (dirigent-like gene) genes and phosphorus (P) uptake (in mg/plant) for rice genotypes grown under acidic lowland soil conditions
UTR, untranslated region.
a Average P uptake values (in mg/plant) ± confidence interval at the 5% level of significance.
b The terms tolerant (T) and susceptible (S) refer to yield under acidic lowland conditions (data not shown).
c No amplification.
d No variation.
e Kasalath sequence was taken as the reference for diversity calculations.
Allele mining across PupK20-2
The gene model for PupK20-2 is shown in Fig. 1(b). Different rice genotypes including a landrace (LR 39), wild type (LR 5 and RDW 125) and varieties (LR 55, LR 11, UR 29 and LR 60) were used for genotyping (Fig. 3(a) and (c)). Two markers were used for allele mining (Fig. 1(b)). In addition to the marker reported by Chin et al. (Reference Chin, Lu, Haefele, Gamuyao, Ismail, Wissuwa and Heuer2010), one more marker (CAUK20-2) was designed so as to amplify the region that was not covered by the previous marker. There was an overlap of 221 bp between the two sets of primers. Amplification was not observed for LR 18 and LR 26.
Sequence variations across PupK20-2 (dirigent-like gene). Sequence polymorphism across the (a) exonic and (b) intronic regions. The nucleotide sequence of the PupK20-2 allele from LR 60 (Kasalath) (Oryza sativa ssp. indica) was taken as the reference. Position indicates the SNP position in the exons and introns of PupK20-2. The dark grey, light grey and white colours indicate Kasalath, Nipponbare and novel alleles, respectively. Genotypes are listed on the left. Only polymorphic regions of the sequenced DNA fragments are shown.
The genomic length of PupK20-2 was 1189 nucleotides, and it contained three exonic and two intronic regions. The length of three exons was 348, 207 and 36 bp nucleotides, respectively. The 5′-UTR (72 bp) along with the first two exons were fully sequenced, but the third exon was not covered in our study. Sequencing revealed a total of 28 SNPs and Kasalath-specific 3 bp deletion at position 665 bp. The 3 bp deletion was observed in five genotypes (LR 5, LR 19, LR 23, UR 29 and RDW 125), whereas LR 11, LR 39 and LR 55 carried a 3 bp (CAG) insertion. Seven SNPs were found in the exonic regions and 21 in the intronic regions (Fig. 3(a) and (b)). Of the variations observed, four SNPs and a Kasalath-specific 3 bp deletion have been previously reported by Chin et al. (Reference Chin, Gamuyao, Dalid, Bustamam, Prasetiyono, Moeljopawiro, Wissuwa and Heuer2011).
Eleven nucleotide substitutions were specific to UR 29. LR 23 and RDW 125 showed the Kasalath-type haplotype with only one nucleotide substitution at position 363 bp in the intronic region. Four nucleotide substitutions were found to be common between LR 11, LR 39 and LR 55 at positions 518, 655, 620 and 660. Similar nucleotide variations at positions 533, 669 and 671 were obtained for LR 11 and UR 29 (Fig. 3(a) and (b)). A total of 16 transitions and 12 transversions were observed. Three synonymous and one non-synonymous substitution were found. The synonymous substitutions were found at positions 475, 526 and 568. The non-synonymous substitution was found for UR 29 at position 146 (Fig. 3(a)). Although the nucleotide variations occurred in the conserved region of the protein, there was no change in the ‘dirigent’ domain of the protein. Three-base pair insertions were found in LR 11, LR 39 and LR 55 in place of the Kasalath-specific 3 bp deletion (Fig. 3(b)). The complete Kasalath-type allele was found in LR 5 and LR 19.
Nucleotide diversity for the two genes
Nucleotide diversity was calculated for each genotype (Table 1). Only four SNPs found in our study matched with those reported earlier by Chin et al. (Reference Chin, Lu, Haefele, Gamuyao, Ismail, Wissuwa and Heuer2010). The nucleotide variation of UR 29 was higher when compared with other genotypes for both PSTOL1 and PupK20-2. Interestingly, the wild-type genotype LR 5 carried the complete Kasalath-type allele, while the wild-type RDW 125 showed the least nucleotide diversity for PupK20-2 among the genotypes surveyed for polymorphism. A nucleotide diversity of 0.005 was observed in LR 11, while LR 23 showed no nucleotide diversity across the CDS region of PupK20-2.
Haplotypes observed in the acidic soil-adapted rice genotypes
The acidic lowland soil (pH 5.65) used in our study for the field evaluation of rice genotypes was low in available P ( < 22 kg/ha) according to Bray's method.
The P uptake of LR 60 was highest among the nine genotypes. Except for LR 39, LR 11, UR 29 and LR 55, the P uptake of all the rice genotypes used in this study was comparable with Kasalath (Table 1).
Of the ten genotypes sequenced, LR 11, LR 18, LR 19, LR 23 and LR 26 were donors being used in our breeding programme for acidic lowland soils. Both, LR 23 and LR 19 showed the Kasalath-type haplotype for both PSTOL1 and PupK20-2 genic regions. On the other hand, LR 11 lacked the PSTOL1 region and showed a mixed haplotype in the case of PupK20-2. The genotypes LR 26 and LR 18, both having high P uptake, lacked both PSTOL1 and PupK20-2 genes.
Discussion
Implications of allele mining for marker-assisted selection
Allele mining refers to the identification of novel allelic variants associated with a particular phenotype of interest, which can also lead to the identification of the nucleotide sequence changes associated with superior alleles (reviewed in Kumar et al., Reference Kumar, Sakthive, Sundaram, Neeraja, Balachandran, Shobha Rani, Viraktamath and Madhav2010). In our case, we did not sequence a large sample set; however, the aim of the study was to find whether markers reported previously could identify the Kasalath-type haplotype accurately. Also, due to the complex nature of the poor soils found in this part of rice-growing region, one wonders whether the ideal donor could carry something other than the ‘typical’ Kasalath allele. ‘Kasalath’, a landrace, has been previously identified as carrying the desired haplotype across PSTOL1 and PupK20-2 (Wissuwa et al., Reference Wissuwa, Yano and Ae1998; Chin et al., Reference Chin, Gamuyao, Dalid, Bustamam, Prasetiyono, Moeljopawiro, Wissuwa and Heuer2011). However, this genotype cannot be used as a donor parent for low P tolerance due to undesirable traits such as lodging and an average performance under acidic field conditions. Therefore, potential donors adapted to acidic soil conditions carrying the desired haplotypes need to be identified. An earlier survey of 159 diverse rice genotypes conducted by Chin et al. (Reference Chin, Lu, Haefele, Gamuyao, Ismail, Wissuwa and Heuer2010) has suggested that Pup1 is largely absent from lowland and irrigated varieties. We wanted to know the nucleotide variation across PSTOL1 and PupK20-2 in lowland rainfed varieties. More number of lowland rice genotypes was taken for the study as we wanted to search for ideal donors for lowland rainfed conditions. Only one upland genotype, UR 29, was taken as an exception.
It is hypothesized that these two rice genes have a role in changing the root system architecture in low-P conditions, but the exact mechanism is not fully understood (Gamuyao et al., Reference Gamuyao, Chin, Pariasca-Tanaka, Pesaresi, Catausan, Dalid, Slamet-Loedin, Tecson-Mendoza, Wissuwa and Heuer2012). It has been suggested that PSTOL1 might act similar to receptor-like cytoplasmic kinases, PR5K and SNC4, of Arabidopsis (Gamuyao et al., Reference Gamuyao, Chin, Pariasca-Tanaka, Pesaresi, Catausan, Dalid, Slamet-Loedin, Tecson-Mendoza, Wissuwa and Heuer2012). The functional complementation of Arabidopsis thaliana double mutant stn7/stn8 defective in serine/threonine protein kinases showed that recombinant Pstol1 protein restored phosphorylation of the double mutant to almost wild-type levels, confirming that Pstol1 is a functional serine/threonine protein kinase (Gamuyao et al., Reference Gamuyao, Chin, Pariasca-Tanaka, Pesaresi, Catausan, Dalid, Slamet-Loedin, Tecson-Mendoza, Wissuwa and Heuer2012). The dirigent protein family, to which PupK20-2 belongs, has been implicated in response to various abiotic stresses (Chin et al., Reference Chin, Gamuyao, Dalid, Bustamam, Prasetiyono, Moeljopawiro, Wissuwa and Heuer2011). Allele mining across the PSTOL1 gene revealed the conserved protein kinase domain across the genotypes LR 23, LR 19 and LR 39. Although the K46-1 marker revealed no difference between LR 23, LR 19, LR 39 and UR 29 based on amplicon size (Tyagi et al., Reference Tyagi, Rai and Dohling2012; present study), upon sequencing, some nucleotide substitutions similar to O. glaberrima (Tanaka et al., Reference Tanaka, Chin, Drame, Dalid, Heuer and Wissuwa2014) were found in LR 39 and UR 29. Interestingly, the nucleotide variations in LR 39 and UR 29 indicated a haplotype that was a mixture of Kasalath and O. glaberrima haplotypes. UR 29 has been reported to perform better under upland poor soil conditions (Wu et al., Reference Wu, Liao, Hu, Yi, Jin, Ni and He2000). As a result of the mixed haplotypes, it will not be possible to distinguish between UR 29 and LR 23/LR 19/Kasalath alleles just based on the markers reported by Chin et al. (Reference Chin, Lu, Haefele, Gamuyao, Ismail, Wissuwa and Heuer2010) and Tanaka et al. (Reference Tanaka, Chin, Drame, Dalid, Heuer and Wissuwa2014). Moreover, the 3′-UTR of the gene too revealed four novel nucleotide substitutions in UR 29 and LR 39. This implies that markers reported to date will not help in the selection of the Kasalath-type or glaberrima-type allele in the upland field.
Allele mining has been suggested as an important way of expanding disease resistance (Xu et al., Reference Xu, Lv, Shang, Pang, Zhou, Wang, Jiang, Tao, Xu, Li, Zhao, Li, Xu and Zhu2014); however, for P-deficiency tolerance, until we find superior alleles, a large number of mixed haplotypes will only confound the use of a previously reported marker (e.g. K46-2) as a foreground marker.
The other gene with predicted potential is PupK20-2 (Gamuyao et al., Reference Gamuyao, Chin, Pariasca-Tanaka, Pesaresi, Catausan, Dalid, Slamet-Loedin, Tecson-Mendoza, Wissuwa and Heuer2012). Allele mining across this gene led to the identification of a total of 16 transitions and 12 transversions. Apart from four SNPs and one InDel at positions 655, 660, 665 and 671 (Chin et al., Reference Chin, Gamuyao, Dalid, Bustamam, Prasetiyono, Moeljopawiro, Wissuwa and Heuer2011), the rest of the SNPs were novel. The sequencing data revealed that a previously reported marker (K20-2) for PupK20-2 would be unable to distinguish between a good P-uptake genotype (LR 23) and a low-P tolerant genotype (UR 29).
Implications of allele mining for breeding low-P tolerant genotypes adapted to acidic lowland soil conditions
Low P tolerance is a quantitative trait in rice (Kirk et al., Reference Kirk, George, Courtois and Senadhira1998), and there are bound to be more genes involved in tolerance mechanism. Of the four genotypes with the highest P uptake, three have the tolerant Kasalath-type haplotype. Genotypes with less P uptake either do not have the genes or have a large number of SNPs. A large number of SNPs suggest that the region in these genotypes is not under any selection pressure and, therefore, evolves very fast. It has been previously suggested that the Pup1 locus has been selected by farmers in traditional varieties grown under upland conditions (Chin et al., Reference Chin, Lu, Haefele, Gamuyao, Ismail, Wissuwa and Heuer2010). Genotype LR 26 may have a different tolerance mechanism, and this is one of the areas that we are currently pursuing. Mapping populations derived from genotypes such as LR 11, LR 26 and LR 18 can be used to characterize novel genes responsible for P-deficiency tolerance. Biparental populations generated between LR 23 or LR 19 and any of the above three genotypes could be used for mapping novel loci as well as introgressing both PSTOL1 and PupK20-2.
It is important to know the exact status of the candidate gene when using markers lying within the gene so that problems arising due to (1) introgressing alleles that already exist or (2) mixed haplotypes can be avoided. This is more relevant for a trait such as P-deficiency tolerance as PSTOL1 and PupK20-2 are underlying a QTL having a small effect under flooded conditions (Chin et al., Reference Chin, Lu, Haefele, Gamuyao, Ismail, Wissuwa and Heuer2010) and are difficult to phenotype.
The present study provides a way forward for the future breeding programme by providing more insight into the donors being used for acidic lowland soils.
Conclusion
The discovery of novel SNPs across both PSTOL1 and PupK20-2 suggests the existence of novel haplotypes in genotypes adapted to acidic lowland soil conditions. The nucleotide substitutions observed in both the genes do not affect the conserved domain of the protein. Genotypes such as LR 19 and LR 23 with high P uptake carried the Kasalath-type haplotype for PSTOL1, whereas UR 29 and LR 39 showed a mixed haplotype having alleles similar to both Kasalath and O. glaberrima. Therefore, previously reported markers can be used for introgression of PSTOL1 in the case of the former two genotypes, but not for the latter. On the other hand, LR 11 lacked the PSTOL1 region and showed a mixed haplotype in the case of PupK20-2. The genotypes LR 26 and LR 18, both having high P uptake, lacked both PSTOL1 and PupK20-2 genes. Mapping populations derived from genotypes such as LR 11, LR 26 and LR 18 can be used to characterize novel genes responsible for P-deficiency tolerance. These rice genotypes that are now better characterized will lead to the design of an informed breeding programme for acidic soils.
Supplementary material
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/S1479262115000544
Acknowledgements
This work was funded by the National Agricultural Innovative Project (grant no. C30033/415101-036), Indian Council of Agriculture Research, New Delhi, India to W. T., J. S. Y. was supported by a student fellowship from the CAU, Imphal.
