The Rice Blast Resistance Gene Pid Family has been Strictly Diverged into Indica and Japonica Subspecies

Ruipeng Chai South China Agricultural University Jinyan Wang South China Agricultural University Xing Wang South China Agricultural University Jianqiang Wen South China Agricultural University Xuemei Ye South China Agricultural University Yaling Zhang South China Agricultural University Yongxiang Yao South China Agricultural University Jianfu Zhang Fujian Academy of Agricultural Sciences Yihua Zhang Fujian Academy of Agricultural Sciences Ling Wang South China Agricultural University Qinghua Pan (  panqh@scau.edu.cn ) South China Agricultural University https://orcid.org/0000-0002-0767-4089

. As a result, following the mutation of matching avirulence genes, major gene-based resistances are prone to rapid breakdown. That in turn enable it to create a new genotype of resistance gene to overcome the emerged new race with its new resistance speci city (Zeigler et al. 1995 At least 100 major genes encoding resistance to P. oryzae are known, an increasing number of which have been isolated (Liu and Wang 2016;Zhao et al. 2018). Among the latter are the three genes Pid-2, -3 and − 4, present as a cluster on chromosome 6 (here after the Pid family) (Chen et al. 2006(Chen et al. , 2018Shang et al. 2009). The objective of the present study was to devise a set of reliable FNP markers based on variations in genomic sequences of the Pid family, and to use these to exploit the extent of allelic variation available in rice germplasm. A particular focus was to reveal the genetic basis underlying resistance gene divergence between indica and japonica subspecies.

Marker veri cation
The FNP assays were validated by testing a larger set of control cvs (so-called CKs), namely Digu (DIG), Tetep (TTP), CO39, Zhenshan97 (ZS97), Tadukan (TDK), Nipponbare (NPB), Koshihikari (KSH) and Genotyping and data analysis A smaller set of control cultivars, i.e., DIG, TTP, ZS97, NPB, SN265, and CO39, were involved in each genotyping experiment (Table S2). The functional and nonfunctional haplotypes of each member of the Pid family were rstly determined with a couple of the functional and nonfunctional FNP markers, and new alleles were then determined with a set of allelic FNP markers by testing a regular panel consisting of each 30 representative indica and japonica type cultivars. For con rming genetic divergence of alleles between indica and japonica rice groups (if any), allele mining was then extended to two additional germplasm panels, one consisting of 40 indica type cultivars used as parents in rice breeding programs based in the southern province of Guangdong and the other of 28 japonica type cultivars used -similarly in the north-eastern province of Heilongjiang (Table S2). A χ 2 test was used to determine whether the two genepools had or had not experienced divergence. The test was based on the formula where a and b represent the number of indica type entries scored as, respectively, harbouring or not harbouring a given allele or genotype, while c and d represent the same for the japonica type entries. N denotes the total number of alleles or genotypes detected for each Pid gene or genotype (Fernando et al. 2018;Zhang et al. 2019;Huang et al. 2021). If all alleles derived from a given resistance gene, which was extremely diverged into indica group, then the resistance gene was de ned as indica type one, and that in turn called as japonica type one.

Pid-2 alleles
An alignment of Pid-2 coding sequences (CDSs) of the 15 reference cultivars revealed the presence of seven single nucleotide polymorphisms (Fig. S1). A pair of FNPs, Pid2-F/N C1022T and Pid2-F/N A1383G effectively distinguished between the functional and the non-functional alleles. DIG, TTP, CO39, ZS97 and TDK each carried a functional allele, while NPB, KSH and SN265 carried a non-functional one (Fig. 1).
Pid2-DIG T2058C was informative for Pid2-DIG allele (DIG), and Pid2-ZS A555G for Pid2-ZS allele (TTP, CO39, ZS97, and TDK). When the regular panel was screened, all 30 indica type entries were found to carry a functional Pid-2 allele, but this was the case for only four of the 30 japonica types ( Fig. 2; Table S2). Of the 34 Pid-2 carriers, 14 belonged to Pid2-DIG, and 18 to Pid2-ZS, and two carried a distinct allele (hereafter referred to as Pid2-New). The screen of the additional 40 indica rice panel revealed that of the 39 carrying a functional copy of Pid-2, 32 had the Pid2-ZS allele, six the Pid2-DIG allele and one the Pid2-New allele. None of the additional japonica germplasm panel carried a functional copy of Pid-2 (Fig. S2). A homogeneity test suggested that divergence of Pid-2 was speci c to the indica genepool (Table 1). It was, therefore, de ned as an indica type resistance gene.

Pid-3 alleles
The variation in the Pid-3 CDSs identi ed in the 15 reference cultivars comprised 29 SNPs and one indel (Fig. S3); 18 of the SNPs and the indel were targeted for marker development (data not shown). The Pid3-F/N G2009A and Pid3-F/N C2209T were both effective for distinguishing between functional and nonfunctional alleles: the ve cultivars DIG, TTP, CO39, ZS97 and TDK each carried a functional allele, while NPB, KSH and SN265 each carried a non-functional one ( Fig. 3; Table S2). Three pairs, Pid3-DIG G775A vs Pid3-DIG G2695A , Pid3-TTP C1136T vs Pid3-TTP C1623G , and Pid3-ZS G477A vs Pid3-ZS C525T , were ensured as allele-speci c FNP markers responsible for Pid3-DIG, Pid3-TTP, and Pid3-ZS, respectively (Fig. 3). As was the case for Pid-2, all 30 members of the indica panel carried a functional Pid-3 haplotype, whereas only four of the japonica panel did so ( Fig. 4; Table S2). The distribution of effective alleles was highly uneven: 29 of the Pid-3 carriers harboured the Pid3-ZS allele, three the Pid3-DIG allele, one the Pid3-TTP and one a novel allele (Pid3-New). The distribution was similarly uneven in the additional indica panel, where 28 of the Pid-3 positive entries carried the Pid3-ZS allele, three the Pid3-DIG allele, one the Pid3-TTP allele and one Pid3-New; none of the members of the additional japonica panel carried an effective allele (Fig. S4). A homogeneity test implied that divergence for Pid-3 has only occurred in the indica genepool (Table 1). It was, also, referred to as an indica type resistance gene.

Pid-4 alleles
Pid-4 was by far the most diverse of the three members, with 149 SNPs and six InDels identi ed in the CDSs plus one intron of the 13 reference cultivars (Fig. S5), a sample of these (17 SNPs and two InDels) were targeted for marker development (data not shown). Both Pid4-F/N C1217G and Pid4-F/N A1452G were informative with respect to functionality: ve cultivars, DIG, NPB, KSH, CO39, and SN265, were recognized as carriers of functional alleles, while TDK, TTP and ZS97 harboured non-functional alleles ( Fig. 5; Table  S2). Two pairs, Pid4-DIG A1149T vs Pid4-DIG A1898G , and Pid4-NPB G1362A vs Pid4-NPB C1554A , were con rmed as allele-speci c FNP markers responsible for Pid4-DIG and Pid4-NPB, respectively; and Pid4-SN/CO T1841A coupled with Pid4-SN/CO G2250C responsible for both Pid4-SN and Pid4-CO (Fig. 5). Unlike the situation in both Pid-2 and Pid-3, functional Pid-4 alleles were present in many (28/30) of the japonica type entries, while the frequency of functional alleles was only moderate (12/30) in the indica germplasm ( Fig. 6; Table S2). The distribution of the various alleles was more even than was the case for Pid-2 and Pid-3, with 14 entries carrying the Pid4-SN allele, 11 the Pid4-NPB allele, eight Pid4-New and six the Pid4-CO allele (Fig. 6). Extending the screen to the two additional panels revealed that 28/40 japonica type cultivars harboured a functional allele, while only 7/40 indica type cultivars did so. Of the 35 functional haplotypes, 12 were present in entries carrying the Pid4-CO allele, 11 in those carrying the Pid4-NPB allele, seven in those carrying Pid4-New, ve in those carrying the Pid4-SN allele, while just a single entry carried the Pid4-DIG allele (Fig. S6). A homogeneity test con rmed that signi cant divergence at Pid-4 has occurred in the japonica genepool (Table 1). Thus, it was termed 'japonica type resistance gene'.

Discussion
The FNP markers for the Pid family have been comprehensively integrated In the present study, a deeper allele mining was carried out on the Pid family in the three panels consisting of 70 indica and 58 japonica cultivars, which were selected from various regions across landrace and modern rice eras (Tables 1; S2 , allele mining of each member of the Pid family was, therefore, initiated from haplotype differentiation with haplotype speci c FNPs, which enabled us to identify any new allele in a given cultivar belonging to the functional haplotype. That in turn, allele mining could be stopped when the panel where was not any functional haplotype (Figs. S2, S4). Whereas allele mining was pursued to individual alleles with a series of allele-speci c FNPs, which helped us nding out more certain alleles within the functional haplotypes. Even so, there were still 3, 2, and 14 cultivars in Pid-2, Pid-3, and Pid-4 categories, respectively, which were presumed to carry new types of alleles, compared to the de ned alleles (Table 1). Collectively, the FNP markers for allele mining in the present study were largely improved from those used in the previous researches, as almost of those were focused on individual SNPs and/or InDels, which were unable to nd out a series of new alleles as were the three cases shown The Pid family has been strictly diverged into indica and japonica subspecies Four alleles were detected at Pid-2, of which the three functional ones were almost entirely restricted to indica type cultivars, while the null allele was only present in japonica type germplasm ( Table 1). The distribution of alleles at Pid-3 was very similar: the four functional ones were harboured for the most part by indica type entries and the null allele was common in the japonica genepool ( Table 1). The latter result echoes a prior nding that the alleles of Pid-3 present in japonica type cultivars are pseudogenes (Shang et al. 2009;Lv et al. 2017). In contrast, the distribution of alleles at Pid-4 featured ve functional alleles which were shared evenly among the japonica type entries, with the null allele found only in indica type ones (Table 1). It might be the rst time to nd out and de ne both indica and japonica type resistance genes within individual cultivars through allele mining with a set of comprehensive FNP markers (Table  S2). The data revealed by the FNP screen suggest a plausible genetic basis for the stable and broad blast resistance exhibited by the modern cultivars, Digu, R207, Lu28S, TianfengB, R217, Zhonghua 11, Gumeizao 4, Moliruanzhan, Yuehesimiao and Yuejingsimiao 2, in that all of these cultivars harbour a functional allele at each of the three Pid genes (Table S2). It might be truly expected that integration of both indica and japonica resistance genes into upcoming cultivars would be one of the most promising ways to enlarge their genetic diversities of resistance genic resources thereby withstanding ever-growing pressure from the pathogen across indica and japonica rice areas (2 3 -1=7), shortly as: d2, d3, d4; d2-d3, d2-d4, d3-d4; and d2-d3-d4, irrespective of speci c alleles. However, only three genotypes, d4, d2-d3, and d2-d3-d4, were detected in the three panels consisting of 128 diversi ed rice germplasms (Tables 1, S2). The indication is therefore that rather limited genotypes of the Pid family have been integrated into both indica and japonica rice cultivars in China. Since all the three members have been strictly diverged into the two subspecies across landrace and modern rice eras, d2-d3 was centralized in indica group and d4 in japonica one both reached at overwhelming proportions; and d2-d3-d4 also in indica group but with a rather moderate rate ( Table 1). The genetic structure of the region harbouring the Pid family does not suggest any obvious barrier to local recombination (Fig. S7). That is, there were four types of such barriers in the target genomic region, the key subspecies hybrid sterile gene cluster S5 (Chen et al. 2008), the heading date gene Hd1 (Yano et al. 2000), the photonasty gene Se5 (Izawa et al. 2000), and the centromere of rice chromosome 6 (Zhao et al. 2019), all of which were enough far from the genic positions of the Pid family. In addition, genomic intervals among the three members were also enough long for independently segregation each other in a given genetic cross (Fig.  S7).
The most possible genetic determinants leading to establish such speci c allelic and genotypic structures of the Pid family was, therefore, due to the speci c lineage(s) of the Chinese rice population. That is, the Chinese rice population has, indeed, been derived from rather limited founder parents for an age. By reviewing the pedigrees of Top-10 of several cultivar types including the general cultivars and F 1 hybrid crosses, Liu (2021) pointed out that the speci c lineage, 'Zhenzhuai 11-ZS97', both were recognized as Pid2-ZS_Pid3-ZS, has been central to Chinese indica type rice breeding programs since the 1960s (also see www.ricedata.cn/variety). The speci c lineage perfectly addressed to both questions: why there were two indica type alleles with much higher rates in the respective allelic structures, Pid2-ZS with 71.4% (50/73), and Pid3-ZS with 86.3% (63/73); and why there was not any single gene genotype for d2 or d3, but d2-d3 for being predominant among the three effective genotypes for a long time (Tables 1,  S2). That is, the unique allelic structures of the three members of the Pid family have been mainly constructed by the genotype, d2-d3, exactly, Pid2-ZS_ Pid3-ZS, carried by the lineage in rice breeding programs in China since the 1960s. It was, again, concerned that updating the lineage would be the key to

Conclusions
The study has demonstrated that all the three members of the Pid family have been strictly diverged into indica and japonica subspecies: Pid-2 and Pid-3 were de ned as indica type resistance genes, and Pid-4 as japonica one. Owning to the speci c lineage, 'Zhenzhuai 11-ZS97', rather limited genotypes of the Pid family have been effective in both indica and japonica rice groups, of which Pid2-ZS + Pid3-ZS has been central to the Chinese rice population since the 1960s.

Declarations
Author's contributions

Availability of Data and Materials
The data sets supporting the results of this article are included within the article and its supporting les.
Ethics Approval and Consent to Participate Not applicable.

Consent for Publication
Not applicable.  Table   Due to technical limitations, table 1 is only available Table S2. M, DL-500 size marker.    Table S2. M, DL-500 size marker.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.