Insertion of Transposable Elements in AVR-Pib of Magnaporthe oryzae Leading to LOSS of the Avirulent Function

Rice blast is a very serious disease caused by Magnaporthe oryzae, which threatens rice production and food supply throughout the world. The avirulence (AVR) genes of rice blast are perceived by the corresponding rice blast resistance (R) genes and prompt specific resistance. A mutation in AVR is a major force for new virulence. Exploring mutations in AVR among M. oryzae isolates from rice production fields could aid assessment of the efficacy and durability of R genes. We studied the probable molecular-evolutionary patterns of AVR-Pib alleles by assaying their DNA-sequence diversification and examining their avirulence to the corresponding Pib resistance gene under natural conditions in the extremely genetically diverse of rice resources of Yunnan, China. PCRs detected results from M. oryzae genomic DNA and revealed that 162 out of 366 isolates collected from Yunnan Province contained AVR-Pib alleles. Among them, 36.1–73.3% isolates from six different rice production areas of Yunnan contained AVR-Pib alleles. Furthermore, 36 (28.6%) out of 126 isolates had a transposable element (TE) insertion in AVR-Pib, which resulted in altered virulence. The TE insertion was identified in isolates from rice rather than from Musa nana Lour. Twelve AVR-Pib haplotypes encoding three novel AVR-Pib variants were identified among the remaining 90 isolates. AVR-Pib alleles evolved to virulent forms from avirulent forms by base substitution and TE insertion of Pot2 and Pot3 in the 5′ untranslated region of AVR-Pib. These findings support the hypothesis that functional AVR-Pib possesses varied sequence structures and can escape surveillance by hosts via multiple variation manners.


Introduction
In the coevolution of plants and pathogens, the latter can adapt to the host and environment, and selection is the major evolutionary force.Up to now, the "arms race" and "trench warfare" hypotheses of coevolution between host resistance (R) genes and pathogen avirulence (AVR) genes have been proposed.In the principal hypothesis of the arms-race, the mutation of R genes and AVR genes is derived by directional selection.Contrarily, it is derived by undirectional selection in the trench-warfare hypothesis [1].
The effectiveness of resistance of Pib has been examined in different rice production provinces in China.Pib exerts a high level of resistance to M. oryzae from Heilongjiang Province, and can be applied as a parent for resistance breeding in Heilongjiang Province [4].Pib is moderately resistant in Fujian Province [5], but Pib exhibits partial resistance in Guangdong, Sichuan and Guizhou Provinces [6,7].Different resistance spectra of Pib were detected in 282 blast isolates collected from Indica rice-and Japonica rice production regions in Yunnan Province [8].Those results showed that the Pib gene exhibits different resistance to the Chinese rice blast from different rice-growing regions.There were 42 out of 54 varieties containing Pib in Chinese elite hybrid rice varieties, two haplotypes of Pib were identified and 9 different haplotypes of AVR-Pib were found among 27 M. oryzae [9].The reactions of differential isolates on 54 rice varieties between the frequency of AVR-Pib haplotypes in the differential isolates showed a good correlation [9].They showed that the Pib gene was widely distributed in rice varieties in China, and the adapt variation of AVR-Pib of M. oryzae occurred.The rice resistance gene of Pib is located on the long arm of chromosome 2 [10,11].The cDNA length of the Pib gene contains 306 bp of 5 ′ untranslated regions (UTRs), 3753 bp of open reading frames (ORFs) (containing 3 exons) and 229 bp of 3 ′ UTRs, and encodes a NBS-LRR protein with 1251 amino acids [12].In China, 16 out of 204 (7.8%) varieties have been detected hoarding the Pib gene in a mini-core collection of Chinese rice germplasm using their functional markers [13].Moreover, 11 varieties were identified with the Pib gene among 58 leading rice cultivars or hybrid rice parents in China using functional DNA markers [14].Pi-b was detected in 33 landraces among 176 landraces (18.8%) from Yunnan Province in China [15].Pib gene homologs (87-bp deletion in exon 1 of Pib, leading to a loss of the resistance function of Pib) were identified in Yunnan Yuanjiang type of common wild rice (Oryza rufipogon Griff) [16].In the Philippines, 32 out of 52 commercial rice varieties have been shown to contain Pib as detected by polymerase chain reaction (PCR) primers specific to Pib [17].They showed that the Pib gene was widely used in rice breeding in China.
Further clarification of the diversification and evolution of the AVR gene is useful for the prediction of the effectiveness and durability of R genes for resistance breeding.Here, we wished to: (i) detect the diversity of nucleotide sequences of AVR-Pib alleles of M. oryzae under field conditions; (ii) determine the avirulence function of AVR-Pib variations to the Pib gene; (iii) reveal the molecular diversification principles of AVR-Pib alleles in M. oryzae in Yunnan Province.Our results provide useful information for rice-blast disease controlling and resistant breeding in China.

Effectiveness of the Pib Gene and Frequency of AVR-Pib Alleles
The efficacy of the Pib gene was examined by pathogenicity assays.A total of 223 of the 366 M. oryzae isolates tested were avirulent to the Pib gene-containing rice monogenic line IRBLb-B (Table 1).The percentage of avirulent isolates to Pib was 60.9%, whereas the remaining 143 isolates were virulent to the Pib gene (Table 1).The percentage of the avirulent isolate was 100, 75.9, 75.0, 57.6, 53.6, and 48.2% in northwestern, central, northeastern, southeastern, southwestern and western Yunnan Province, respectively.Among 366 isolates, 44.3% of isolates with the AVR-Pib allele were amplified by AVR-Pib-specific primers (AVR-Pib F1/AVR-Pib R1) (Table 1; Figure S1), and three genotypes (L1 with 1231 bp, L2 with 3100 bp and L3 with both 1231 bp and 3100 bp) of AVR-Pib alleles in 162 isolates were amplified (Table 1; Figure S1).The highest percentage of amplification of AVR-Pib was 73.3% in the rice blast isolates collected from northwestern Yunnan Province, whereas the lowest percentage was 36.1% from northeastern Yunnan Province (Table 1).The percentage of AVR-Pib was 46.3, 36.1, 73.3, 51.5, 67.9 and 39.0% in central, northeastern, northwestern, southeastern, southwestern and western Yunnan Province, respectively.The percentage of AVR-Pib was 47.0 and 42.4% in Xian/Indica (XI) rice-and Geng/Japonica (GJ) rice production areas in Yunnan.The genotype of L1, L2 and L3 alleles of AVR-Pib was detected in 104, 53 and 5 isolates, with percentages of 28.4%, 14.5% and 1.3%, respectively (Table 1).The genotype of L1, L2 and L3 alleles of AVR-Pib was detected in the XI rice production area, whereas L3 was absent in the GJ rice production area (Table 1).

Virulence Function of AVR-Pib Variations against the Pib Gene
Twelve AVR-Pib haplotypes (H01 to H12) (Table 2), excluding the original AVR-Pib allele (GenBank accession number, KM887844), were detected on the nucleotide sequence assemblies of 90 isolates of L1 alleles containing a 719-bp 5 ′ -region, 225-bp coding DNA sequence (CDS) and 302-bp 3 ′ -region of AVR-Pib (Table 2; Figure S2).Moreover, insertion of Pot2 (at position −275) and Pot3 (at position −240) was identified based on the DNA sequence assemblies of six and 30 isolates (Table 2; Figure 1), respectively, and the amplicon size difference between L1 and L2 (Figure S1).The 12 novel AVR-Pib haplotypes (H01-H12) were identified compared to previous published alleles [3,9].Alignment of DNA sequence assemblies of the AVR-Pib allele from 90 isolates revealed 18 mutation sites, in-cluding six mutant sites in the CDS region which were not in the signal-peptide region (Table 2; Figures S2 and S3).Six mutant sites in the CDS region led to changes in amino acids (Table 3).The CDS sequence assemblies of the AVR-Pib allele among the 126 isolates (including L1, L2 and L3) were predicted to produce four AVR-Pib proteins (Table 3).Among them, amino acid variations were predicted to occur at six positions (Table 3).Amino-acid variations at F54L in H05 and H06, E46V, F53S and F54V in H07 were found; these isolates of the corresponding haplotypes were avirulent on the monogenic line IRBLb-B (with Pib).Meanwhile, the amino acid variations at F47L, I49T and R50G were found in one isolate with H08, which was virulent on the monogenic line IRBLb-B (with Pib) (Table 3) and the amino acid variations at F47, I49 and R50 in H01, H02, H03, H04, H05, H06, H07, H09, H10 and H12; these isolates were avirulent on the monogenic line IRBLb-B (with Pib), whereas the amino acid variations at 47L, 49T and 50G in H08, as well as the isolate, were virulent on IRBLB-b (Table 3).This finding suggested that the amino acids F47, I49 and R50 were crucial for the avirulence function of AVR-Pib.

Distribution of Haplotypes of AVR-Pib of M. oryzae
Among 12 AVR-Pib haplotypes, none were identical to the original AVR-Pib (GenBank accession number, KM887844) (Table 2).Eight haplotypes, as well as the Pot2 and Pot3 insertion, were detected in 50 M. oryzae isolates from western Yunnan Province (Table 4).Five haplotypes, as well as Pot3 and Pot3 reverse-insertion, were identified in 23 M. oryzae isolates from central Yunnan Province.Three haplotypes, as well as Pot2 and Pot3 reverse-   The different haplotypes of H01, H02, H03, H04, H09, H10, H11 and H12 had no change on amino acid sequence (Table 3).Three-dimensional protein structures built by homology modeling (SWISS-MODEL; https://swissmodel.expasy.org/,accessed on 18 February 2019) showed the different protein structures of these four (H1, H5, H7 and H8) AVR-Pib variants (Figure S4).Isolates of H01 (amino acids that were the same as that with a GenBank accession number of KM887844), H02, H03, H04, H05, H06, H07, H09, H10 and H12 haplotypes hold AVR-Pib because these isolates were avirulent to the Pib-containing monogenic line IRBLb-B (Table 3).The isolate of H08 defeated the resistance of Pib because this isolate was virulent to the Pib-containing monogenic line IRBLb-B (Table 3).Furthermore, Pot2 and Pot3 inserted in the 5 ′ UTR of the Pib gene were identified in six isolates and 30 isolates (Table 3; Figure 1), respectively, and these isolates were virulent to the Pib-containing monogenic line IRBLb-B (Table 3).These findings suggested that the insertion of TEs (Pot2 and Pot3) and small segments of the nucleotide in the promoter region, and the nuclear substitution in the ORF region, resulted in a variation of AVR-Pib from avirulence to virulence, and that the diverse mutations of the AVR-Pib allele of M. oryzae were involved.

Distribution of Haplotypes of AVR-Pib of M. oryzae
Among 12 AVR-Pib haplotypes, none were identical to the original AVR-Pib (GenBank accession number, KM887844) (Table 2).Eight haplotypes, as well as the Pot2 and Pot3 insertion, were detected in 50 M. oryzae isolates from western Yunnan Province (Table 4).Five haplotypes, as well as Pot3 and Pot3 reverse-insertion, were identified in 23 M. oryzae isolates from central Yunnan Province.Three haplotypes, as well as Pot2 and Pot3 reverseinsertion, were identified in 15 isolates of M. oryzae from northeastern Yunnan Province.Three haplotypes and Pot2 insertion were detected in 19 isolates from northeastern Yunnan Province.Three haplotypes and Pot3 insertion were identified in six M. oryzae isolates from southeastern Yunnan Province.Three haplotypes were identified in 13 isolates of M. oryzae from northwestern Yunnan Province (Table 4).Eleven and nine haplotypes were detected in GJ rice-and XI rice production areas, and the diversity index (DI) of haplotypes was 0.84 and 0.79 for these areas, respectively.The DI of AVR-Pib was 0.72, 0.71, 0.70, 0.63, 0.59 and 0.54 for southeastern, central, western, southwestern, northeastern and northwestern Yunnan Province, respectively (Table 4).[28].Diversity index = (1 − ∑ n i=1 p i 2 ) (where pi is the frequency of the haplotype i in a population).c XI and GJ indicates Xian/Indica and Geng/Japonic, respectively.
In brief, the DI of AVR-Pib alleles in Yunnan Province was in the order: southeastern > central > western > southwestern > northeastern > northwestern.The DI of AVR-Pib alleles in the GJ-rice production area was higher than that in the XI rice production area.These results indicate that the genetic divergence of AVR-Pib of M. oryzae in each rice-growing region occurred depending on each field's condition.
Eighteen nucleotide variable sites in AVR-Pib alleles were identified (Table 2; Figures S2 and S3).A haplotype network based on sequence variations of 90 isolates of L1 alleles was developed (Figure 2).Four main lineage branches (A to D) of AVR-Pib were divided among 90 field isolates (Figure 2), and a different evolution of AVR-Pib among them was noted.Isolates of B and D lineage branches of AVR-Pib were avirulent to IRBLb-B (with Pib) (Figure 2; Table 3).Isolates of H11 of the A-evolved branch and H08 of the C-evolved branch were virulent to the rice-blast-resistant gene Pib, respectively (Figure 2; Table 3).These data suggested that the A and C branches of AVR-Pib had evolved to virulence from avirulent origins via base substitution and insertion, and evaded the recognition of the riceblast-resistance gene Pib in field isolates.The virulence of H08 and H11 was identified in southeastern and western Yunnan Province (Table 4).Moreover, TE insertion in rice samples in all regions except northwestern Yunnan Province (Table 4) suggested that virulent evolution of AVR-Pib occurred in most rice production areas of Yunnan Province.

Selection Pressure on AVR-Pib in M. oryzae
The natural-selection pressure on AVR-Pib was calculated by Tajimaʹs neutrality test on 126 AVR-Pib CDS sequences: the Tajimaʹs D value was not significantly different from zero (D = −1.61687;0.10 > P > 0.05) (Table S1).This result suggested that AVR-Pib may suffer neutral selection and evolve neutrally in the population of M. oryzae.Furthermore, the results of three positive-selection models kept a higher similarity (Figure S5).The "sliding window" under M8, M8a and M7 models showed values of Ka/Ks (Ka, rate of nonsynonymous substitutions; Ks, rate of synonymous substitutions) across all 74 amino acids (Figure S5).The Ka/Ks value of all sites was >1 under the M8 and M8a model, and the value was 1 under the M7 model for entire residues.These results implied that the sites may have suffered from neutral selection.These findings suggest that the AVR-Pib maybe under a neutral evolution.
Only the L1 genotypes (with the expected size) of AVR-Pib were detected in isolates from E. indica, E. coracana and Triticum aestivum Linn, suggesting that these isolates did not have

Selection Pressure on AVR-Pib in M. oryzae
The natural-selection pressure on AVR-Pib was calculated by Tajima's neutrality test on 126 AVR-Pib CDS sequences: the Tajima's D value was not significantly different from zero (D = −1.61687;0.10 > P > 0.05) (Table S1).This result suggested that AVR-Pib may suffer neutral selection and evolve neutrally in the population of M. oryzae.Furthermore, the results of three positive-selection models kept a higher similarity (Figure S5).The "sliding window" under M8, M8a and M7 models showed values of Ka/Ks (Ka, rate of nonsynonymous substitutions; Ks, rate of synonymous substitutions) across all 74 amino acids (Figure S5).The Ka/Ks value of all sites was >1 under the M8 and M8a model, and the value was 1 under the M7 model for entire residues.These results implied that the sites may have suffered from neutral selection.These findings suggest that the AVR-Pib maybe under a neutral evolution.
Only the L1 genotypes (with the expected size) of AVR-Pib were detected in isolates from E. indica, E. coracana and Triticum aestivum Linn, suggesting that these isolates did not have a TE insertion in the AVR-Pib allele.Three genotypes of L1, L2 (with TE insertion) and L3 (with TE insertion) of AVR-Pib were detected in 18 isolates from the Pib homolog-containing O. rufipogon, and the isolate of YN441 (with the H9 haplotype of AVR-Pib and identical with the original haplotype of KM887844) was virulent to O. rufipogon (Figure 3).These findings showed that the diversification of AVR-Pib of M. oryzae was dependent upon the Pib homolog-containing O. rufipogon, and that the variation in TE insertion in AVR-Pib could be selected and adapted to rice and other Gramineae species.
a TE insertion in the AVR-Pib allele.Three genotypes of L1, L2 (with TE insertion) and (with TE insertion) of AVR-Pib were detected in 18 isolates from the Pib homolog-cont ing O. rufipogon, and the isolate of YN441 (with the H9 haplotype of AVR-Pib and iden with the original haplotype of KM887844) was virulent to O. rufipogon (Figure 3).Th findings showed that the diversification of AVR-Pib of M. oryzae was dependent upon Pib homolog-containing O. rufipogon, and that the variation in TE insertion in AVR could be selected and adapted to rice and other Gramineae species.

Phylogeny of Pib Allele Partial to CDS Regions
Fifty-seven sequences of Pib were obtained from GenBank (Table S3).Eleven of th were from five wild-rice species

Phylogeny of Pib Allele Partial to CDS Regions
Fifty-seven sequences of Pib were obtained from GenBank (Table S3).Eleven of them were from five wild-rice species (seven from O. rufipogon, one from O. meyeriana, one from O. officinalis, one from O. longistaminata and one from O. nivara), and 46 accessions from O. sativa, including the original Pib (GenBank accession number, AB013448.1)(Table S3).These sequences were aligned.A minimum-evolution phylogenetic tree was constructed based on the nucleotide sequences of exon 1 of Pib in 34 accessions and partial regions of exon-3 nucleotide sequences (from 7633 to 8484 of AB013448.1) of Pib in 49 accessions, respectively (Figure 4).

Discussion
We identified 12 new haplotypes, as well as Pot2 and Pot3 insertion in the AVR-Pib DNA sequences among rice blast isolates from different rice-growing areas in Yunnan Province.The many virulent isolates to Pib-containing rice varieties implied that Pib was overcome in these rice-growing regions because of the massive exploitation of Pib in China.
Pib alleles have been used widely and have shown strong resistance to disease in China [14].Complete deletions have occurred in AVR-Pib sequences among field isolates of M. oryzae from various rice-growing regions of Guangdong, Hunan and Liaoning Provinces [18].Moreover, TE insertion has occurred in AVR-Pib in M. oryzae isolates from south and northeast China [18] and the Philippines [17].These data are consistent with our results.The L1 genotype of AVR-Pib identified in rice blast isolates collected from rice fields implied that Pib has been effective in preventing rice blast.Li and colleagues showed that rice cultivars with Pib were resistant to 74.9% of isolates (282 isolates) from Yunnan

Discussion
We identified 12 new haplotypes, as well as Pot2 and Pot3 insertion in the AVR-Pib DNA sequences among rice blast isolates from different rice-growing areas in Yunnan Province.The many virulent isolates to Pib-containing rice varieties implied that Pib was overcome in these rice-growing regions because of the massive exploitation of Pib in China.
Pib alleles have been used widely and have shown strong resistance to disease in China [14].Complete deletions have occurred in AVR-Pib sequences among field isolates of M. oryzae from various rice-growing regions of Guangdong, Hunan and Liaoning Provinces [18].Moreover, TE insertion has occurred in AVR-Pib in M. oryzae isolates from south and northeast China [18] and the Philippines [17].These data are consistent with our results.The L1 genotype of AVR-Pib identified in rice blast isolates collected from rice fields implied that Pib has been effective in preventing rice blast.Li and colleagues showed that rice cultivars with Pib were resistant to 74.9% of isolates (282 isolates) from Yunnan Province [8].The corresponding value was 2.1% in Guangdong Province (146 isolates were tested) [28], and the percentage resistance was <31% in Hunan Province [29].These results show that Pib alleles had poor effects in these rice production areas.The further inspection of variation in AVR-Pib DNA sequences in these isolates could reveal the molecular evolutionary patterns of AVR-Pib and predict the durability and effectiveness of Pib allelemediated resistance under field conditions in rice production regions.
AVR-Pia, AVR-Pii and AVR-Pita1 located on telomere regions tend to be unstable, and effective mutants in these genes were identified [21, 30,31].The retrotransposon (MINE) insertion in the ACE1 gene [23] and Pot3 insertion in AVR-Pita1 [32,33] and AVR-Pizt [22] caused new virulent alleles.TE (Pot2 and Pot3) insertion, complete absence, segmental deletion and a point mutation were found in AVR-Pib alleles, all of which lead to a gain of virulence [18].Three expression patterns were identified among different haplotypes of AVR-Pib [18].Recently, insertion of a Pot3 transposon in AVR-Pib was shown to mediate the loss of function of AVR-Pib in all 248 isolates collected from the Philippines [17].These findings showed that rice blasts can use transposons to suppress the expression of AVR genes to defeat the rice blast resistance gene.The AVR-Pib allele was identified in nearly half of rice blast isolates (44.3%) in Yunnan Province (Table 1).This percentage was higher than that in rice blast isolates in Jilin and Heilongjiang Provinces, but lower than those in Guangdong, Hainan and Liaoning Provinces [18].Meanwhile, 28.6% of isolates contained a TE insertion in AVR-Pib.Among them, 21.4% isolates contained a Pot3-element insertion and 2.4% of isolates contained a reverse Pot3-element insertion, and 4.8% isolates contained a Pot2-element insertion in AVR-Pib of M. oryze from Yunnan Province (Table 4; Figure 1).These insertions resulted in the variation from avirulence to virulence to the corresponding R gene.Several nucleotide variations in AVR-Pib alleles were identified, which led to variations in amino acids and implied that AVR-Pib alleles suffer from strong selection pressure in rice production regions of Yunnan Province.Whereas, mutations that impact gene expression are further acknowledged by transcript level.
We observed no TE insertions in AVR-Pib of isolates from E. indica or E. coracana, and TE insertion in AVR-Pib was selected by the host, data that are consistent with the results of Zhang et al. [18].Pot2 and Pot3 insertions were identified in the XI-rice-growing areas, whereas only Pot3 insertion was identified in GJ-rice-growing areas of Yunnan Province.TE insertion of AVR-Pib was noted in all rice-growing regions except northwestern Yunnan Province, and Pot3 insertion was distributed mainly in western Yunnan Province.These results showed that the virulent AVR-Pib alleles were involved in most of the rice-growing regions of Yunnan Province.Hence, the monitoring of these virulent alleles in field populations is important for employing Pib-containing rice varieties.
Various mutations were identified in CDS regions of AVR-Pib, and 12 AVR-Pib haplotypes were found based on 18 variant nucleotides among 90 isolates of L1 alleles collected from Yunnan Province (Table 2).Six new variant amino acids of the AVR-Pib loci variants were identified in the 90 M. oryze isolates in the present study, and resulted in the identification of four novel haplotypes.A more holonomic network was constructed based on the new variations among different alleles of AVR-Pib.The putative and secreted proteins of AVR-Pib in 126 isolates were identified (Table 3), and they were in accordance with the results of Zhang et al. [18].Nine isolates had variations at the amino acid position F54L; three isolates had variations at the amino acid positions of E46V, Y53S and F54V; one isolate had variations at the amino acid positions of F47L, I49T and R50G (Table 3).These isolates were virulent to the monogenic line IRBLb-B (with Pib), suggesting that these amino acids are crucial for avirulent function.Isolates of H11 with insertion of ATTA in the 5 ′ UTR may change AVR-Pib expression and cause a loss of the avirulent function (Tables 2 and 3; Figure S3).
In the course of interactions and co-evolution between pathogens and plants, the R genes of plants can discern the cognate AVR genes of pathogens and inspire immunity [1].
The genetic variation of the AVR genes of the pathogen is dependent upon the R genes of the host and changeable environmental conditions.The DI of AVR-Pib was higher in GJ rice production areas than that of XI rice production areas (Table 4).Different variations were observed in AVR-Pib between XI rice-and GJ rice production areas (Table 4).These findings imply that the adaptive mutations of AVR-Pib occurred in Yunnan Province under natural conditions, and these results were similar in previous studies [18].
Yunnan Province is abundant in genetic resources of rice.The wild species of O. officinalis, O. meyeriana and O. rufipogon also coexist in this province [34].More than 5000 rice accessions germplasms have been conserved in Yunnan Province.Among them, 12 out of 227 accessions carried the Pib resistance gene screened by the resistance gene identification using different isolates [34].Pib gene homologs were identified in wild rice O. rufipogon from Yuanjiang County [16], and four genotypes (L0 to L3) of AVR-Pib were detected in M. oryzae and O. rufipogon in Yuanjiang County.TE insertion of L2 and L3 genotypes of AVR-Pib was absent in the isolates from D. sanguinalis and M. nana Lour.These results suggest that the adaptive variation of AVR-Pib is involved during interactions and co-evolution between AVR-Pib of M. oryzae and Pib of O. rufipogon.The Tajima's D value of −1.61687 (Table S1) indicates that AVR-Pib loci may suffer from a purifying selection by the corresponding R gene in Yunnan rice production areas.
Massive variations and stepwise mutations in AVR-Pib of rice blast isolates were observed in Yunnan Province (Table 2; Figure 2), which suggests that there is an abundant diversity of rice accessions and M. oryzae isolates in Yunnan Province.Pot2 insertion in AVR-Pib was found in western, southwestern and northeastern Yunnan Province.Pot3reversed insertion was found in central and northeastern Yunnan Province, which was not observed in the previous studies [18].Moreover, Pot3 insertion occurred mainly in western Yunnan Province.Pot3 insertion of AVR-Pib was found in GJ rice and XI rice production areas, whereas Pot2-and Pot3-reversed insertion was found only in the XI rice production area and GJ rice production area, respectively (Table 4).The virulent haplotype of H11 was detected in the XI rice production area and GJ rice production area, and H08 was detected in the XI rice production area.These data showed a high variation of AVR-Pib in different rice-growing regions, which may be due to rice variety and the environment.
The stepwise mutations that result in a loss of avirulence function have been identified in AVRL567 [35] and AVR-Pik [36][37][38][39].Based on the Pib homologs identified by Yang et al. [16], and our result for AVR-Pib in the present study, the potential interactions and coevolution of AVR-Pib alleles in M. oryzae and Pib alleles of rice were constructed (Figure S6).The AVR-Pib homolog L1 (H01) originated from an ancestral M. oryzae gene.The Pib allele (87-bp deletion in exon 1 of Pib) in O. rufipogon could not recognize the L1 alleles of AVR-Pib.Thus, the other Pib allele (gained 87 bp in exon 1 of Pib) in cultivated rice evolved to recognize the L1 alleles (H01) of AVR-Pib, whereas the altered alleles L2 and L3 evolved to virulence from avirulent origins by TE insertion, base substitution (H08) and segment insertion (H11) to avoid recognition by Pib (Table 2; Figure S6).These actions indicated a stepwise evolution of AVR-Pib as well as Pib interaction and co-evolution.Intriguingly, the AVR-Pib alleles H08 and H11 were derived from H01, and could escape recognition by Pib (Table 2; Figure S6), but several extinct or missing haplotypes were not identified in the sample (Figure 2).These findings imply that: (i) the AVR-Pib loci of M. oryzae evolved gradually during the interaction and coevolution between the Pib loci of M. oryzae in field conditions; (ii) the genome organization of the AVR-Pib locus is much more intricate than anticipated.

Blast Isolates, Rice Accessions, Culture and Pathogenicity Identification
The rice blast fungus single spores were obtained from infected leaves or panicles incubated on moist filter paper in a Petri dish at room temperature for 24 h according to Jia et al. [38], and the tests made fulfilled the Kock's postulates for all isolates.The seedlings of the rice monogenic line IRBLb-B (which contains Pib) and a susceptible cultivar Lijiangxintuanheigu (LTH; does not contain Pib) were used for pathogenicity assays (the seeds were acquired originally from Cailin Lei).Four thousand isolates were collected from six rice-growing regions from 1997 to 2012 in Yunnan Province, and a total of 366 isolates are selected from six rice-growing regions as representative isolates.The total of 366 isolates of M. oryzae in the present study were the same in the published paper [38].The storing and culturing methods of the isolates were described in the published paper [38].Disease reactions were referred to as the method of Jia et al. [39].In a few words, when rice seedlings were in the 3-to 4-leaf stage, they were inoculated with a spore suspension (1-5 × 10 5 spores/mL containing 0.05% Tween 20).After innoculation, rice seedlings were put into a plastic bag and sealed securely to keep a high relative humidity of 90-100% at 25 • C for 24 h in the dark.Then, the plants were shifted into a greenhouse for 6 days to develop the disease lesion extension fully.
Disease reactions were monitored externally on the second youngest leaf based on the number and degree of lesions using a 0-to-5 disease scale (Figure S7).The disease scale method was described in the published paper [38].Five seedlings at a time, repeating twice, were arranged in the experiment.In addition, the average value of disease scales was used to discriminate resistance versus susceptibility.The disease reaction of wild rice species O. rufipogon, O. meyeriana and O. officinalis (which were conserved in Yunnan Academy of Agricultural Sciences) was determined using a detached-leaf method, as described by Jia et al. [39].One virulent blast isolate was used for inoculation.Disease reactions were evaluated 5-7 days after inoculation.

DNA Extraction, PCR Amplification and DNA Sequencing
A culturing method of vegetative mycelia of M. oryzae isolates was described in the published paper [38].The genomic DNA of each isolate was extracted from vegetative mycelia by CTAB method [40].The primers AvrPibF1 (5 ′ -GGACAAGGGAGGCAAATCT AAC-3 ′ ) and AvrPibR1 (5 ′ -ATGCCGACAATGCGAGGTAT-3 ′ ) were used to amplify the AVR-Pib allele, as well as for sequencing according to the method of Zhang et al. [18].Each PCR reaction was amplified in a total reaction volume of 50 µL containing the following components: 25 µL of 2× Taq PCR MasterMix (Tiangen Biotech, Beijing, China), 1 µL (10 µM) of each primer, 2 µL of template DNA, and 21 µL of ddH 2 O (provided in the Tiangen kit).PCR procedure was conducted in a C1000 Touch™ thermal cycler (Bio-Rad Laboratories, Hercules, CA, USA) in the following steps: initial denaturation at 94 • C for 3 min, followed by 29 cycles at 94 • C for 45 s, 55 • C for 45 s and 72 • C for 2.5 min, and a final extension at 72 • C for 5 min.Each reaction was repeated twice.The size of the PCR products was valued by a DNA marker (DL2000, Tiangen Biotech).Amplicons were sequenced twice by Life Technologies Biotechnology (Shanghai, China).

Data Analyses
DNA sequences of AVR-Pib were assembled and aligned using DNASTAR v7.1.0(www.dnastar.com/,15 January 2019).DnaSP v5.10.01 [41] was used for the calculation of polymorphic sites (π), the number of DNA haplotypes and the sliding window.TCS1.21 (http://darwin.uvigo.es/,accessed on 1 February 2019) [42] was used for analyses of the haplotype network of AVR-Pib.The DI (haplotype diversity index) was counted in M. oryzae populations following the method of Fontaine et al. [43]: where p i is the frequency of haplotype i in a population.Tajima's neutrality test was conducted using MEGA X (www.megasoftware.net,accessed on 12 February 2019) [28].
The Selection Server program (http://selecton.tau.ac.il, accessed on 19 February 2019) was used for analyses of purifying the selection.The purifying of the selected sites of AVR-Pib was identified by used three models: M8 (positive selection enabled, beta + w ≥ 1), M7 (beta, null model) and M8a (beta + w = 1, null model).Then, the sliding window of purifying the selected sites of AVR-Pib was drawn under the M8, M7 and M8a models by Excel™ (Microsoft, Redmond, WA, USA).MEGA X [28] was used for the construc-tion of phylogenetic trees by the minimum evolution method [44].The SWISS-MODEL (http://swissmodel.expasy.org,accessed on 18 February 2019) with ProMod v3.7.0 was used to build the protein homology model.The significant difference of distribution of AVR-Pib alleles and avirulence isolates of M. oryzae in each region was analyzed by Excel software with CHITEST.

Conclusions
We detected twelve novel haplotypes in the field population by using 90 isolates and a transposable element (TE) insertion in 36 of 126 isolates, constructing a complex network of AVR-Pib alleles and assessing the efficacy of Pib alleles in rice production areas of Yunnan; we also analyzed the adaption of TE insertion of AVR-Pib in the isolates from a different host.Our findings support the hypothesis that functional AVR-Pib possesses varied sequence structures and can escape surveillance by hosts via multiple variation manners.Haplotype H08 and H11 can overcome all detected Pib alleles to date, and Pot insertion can change the avirulent function of AVR-Pib.Despite the H08, H11 haplotypes and TEs insertions have low frequencies; the monitoring of these alleles in field populations is critical because of their high risk for Pib-holding rice varieties.The TE insertion was not detected in the AVR-Pib allele in the isolates from E. indica, E. coracana,and Triticum aestivum Linn, while three genotypes of AVR-Pib were detected in isolates from O. rufipogon.The selected and adapted variation of TE insertion in AVR-Pib is the occurrence of the long-term coevolution between M. oryzae and hosts (rice and other Gramineae species).

Supplementary Materials:
The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ijms242115542/s1.Table S1.Tajima's Neutrality Test of AVR-Pib in M. oryzae.Table S2.Genotype of AVR-Pib in isolates from different hosts.Table S3.Lists accessions of Pib were obtained from GenBank used in this study.Figure S1.Detected alleles of AVR-Pib by PCR assays.The amplicons of AVR-PibF1/R1 were used to distinguish four alleles: L1 and L2 (one fragment of different size), and L3 (two fragments of L1 and L2), L0 (no amplification); The isolates 1 to 8 are YJW-1-1d, YJW-1-1e, YJW-2-1a, YJW-2-1b, YJW-2-1c, YJW-2-1d, YJW-2-1e and D-1-1g, respectively.M: DNA marker DL2000.L1 indicates the AVR-Pib genotype with the expected size (1231 bp), L2 and L3 indicates the AVR-Pib genotype with TE insertion (L2 with 3100 bp, L3 with both of 1231 bp and 3100 bp). Figure S2.Diversification of AVR-Pib in avirulent isolates.Distribution of variation of the AVR-Pib alleles was analyzed using sliding window.X-axis shows the distribution of variation within the full region, including signal peptide and exon of AVR-Pib.Lower pane indicates the corresponding schematic presentation of the signal peptide and exon of AVR-Pib.Window length: 1; Step size: 1. π value corresponds with the level of variation at each site because it is the sum of pair-wise differences divided by the number of pairs within the population.Figure S3.Part of insertion of sequences of AVR-Pib.A, indicates ACTTA, AGTTA, ATTA insert, respectively; B, indicates ACGTTA insert; C, indicates ACA insert; D, indicates C insert. Figure S4.SWISS-MODEL homology modelling, built with PROMOD v. 3.70 and method: X-ray.Amino acid variations at R50G in H1 (same as KM887844, H2, H3, H4, H9, H10 and H11), H5 (same as H6), H7 and H8 protein haplotypes of AVR-Pib, respectively.A: Ramachandran Plots.B: the 50th amino acid of AVR-Pib.A value of 0 to 1 is classified as "resistant", 2 denotes "moderately resistant", and 3 to 5 is classified as "susceptible".

Figure 1 .
Figure 1.Characterization of allelic variation at AVR-Pib.The functional nucleotide polymorphic maps of the six natural alleles in 126 isolates of M. oryzae in Yunnan Province.Dis.indicates disease reaction on monogenic line IRBLb-B (containing Pib), V indicates the isolates were virulent to IRBLb-B.

Figure 1 .
Figure 1.Characterization of allelic variation at AVR-Pib.The functional nucleotide polymorphic maps of the six natural alleles in 126 isolates of M. oryzae in Yunnan Province.Dis.indicates disease reaction on monogenic line IRBLb-B (containing Pib), V indicates the isolates were virulent to IRBLb-B.

Figure 2 .
Figure 2. The haplotype network for the 12 AVR-Pib alleles.Haplotype network analysis was performed using TCS1.21(http://darwin.uvigo.es/,accessed on 12 March 2019.).The original AVR-Pib allele was designated as the H01 haplotype in the network.Each haplotype was separated by mutational events.The node in the network represents an extinct or a missing haplotype not found among the samples.Each haplotype was separated by mutational events.All haplotypes were displayed as circles.The size of the circles corresponds to the haplotype frequency.The KM887844 (GenBank Accession No.) of AVR-Pib was obtained from GenBank.White color indicates avirulent to the Pib gene and yellow color indicates virulent to the Pib gene.A to D, four major haplotypes of AVR-Pib in Yunnan Province of China, are shaded.

Figure 2 .
Figure 2. The haplotype network for the 12 AVR-Pib alleles.Haplotype network analysis was performed using TCS1.21(http://darwin.uvigo.es/,accessed on 12 March 2019.).The original AVR-Pib allele was designated as the H01 haplotype in the network.Each haplotype was separated by mutational events.The node in the network represents an extinct or a missing haplotype not found among the samples.Each haplotype was separated by mutational events.All haplotypes were displayed as circles.The size of the circles corresponds to the haplotype frequency.The KM887844 (GenBank Accession No.) of AVR-Pib was obtained from GenBank.White color indicates avirulent to the Pib gene and yellow color indicates virulent to the Pib gene.A to D, four major haplotypes of AVR-Pib in Yunnan Province of China, are shaded.

Figure 3 .
Figure 3. Disease reaction of the identification isolate of YN441 (with the H9 haplotype of AVR which was identical with the original haplotype of KM887844) on Oryza officinalis, O. rufipogon O. meyeriana.LTH: LijiangxinTuanHeigu.

Figure 3 .
Figure 3. Disease reaction of the identification isolate of YN441 (with the H9 haplotype of AVR-Pib which was identical with the original haplotype of KM887844) on Oryza officinalis, O. rufipogon and O. meyeriana.LTH: LijiangxinTuanHeigu.
Exon 1 of Pib in wild-rice species (O.rufipogon, O. meyeriana, O. officinalis and O. longistaminata) was close to that in O. sativa (Figure 4B).The DQ317978.1 group of the wild rice O. rufipogon shared >90% identity with the nucleotide sequences of the JN564624.1 group of Indica.Two major clades emerged in one part of exon 3 of Pib (Figure 4C).One clade contained two wild-rice species (O.rufipogon and O. nivara) and O. sativa.The EF642422.1 group of the wild-rice species O. nivara shared >90% identity in nucleotide sequences with the EF642423.1 group of O. sativa.The EF642442.1 group of O. rufipogon shared >75% identity of nucleotide sequences with the EF642423.1 group of O. sativa.The other clade contained O. rufipogon and O. sativa.The EF642440.1 group of the wild-rice species O. rufipogon shared >75% identity of nucleotide sequences with the EF642433.1 group of O. sativa.The isolate of YN441 (with the H9 haplotype of AVR-Pib) was virulent to O. rufipogon, O. meyeriana, and O. officinalis (Figure 3).These results suggested that different regions of the Pib gene may have suffered different selection pressures in the host rather than domestication.Int.J. Mol.Sci.2023, 24, x FOR PEER REVIEW 10 of 17 was virulent to O. rufipogon, O. meyeriana, and O. officinalis (Figure 3).These results suggested that different regions of the Pib gene may have suffered different selection pressures in the host rather than domestication.

Figure 4 .
Figure 4. Phylogenetic tree constructed with the nucleotide sequences of Pib gene and different partial CDS regions from wild rice and O. sativa using minimum evolution method of MEGA X.The numbers associated with individual branches indicate confidence levels based on 1000 bootstrap replicates.(A), structure of Pib from AB013448.1 (GenBank ID); (B), the phylogenetic tree constructed based on the nucleotide sequences of exon1 of Pib regions from 34 accessions.(C), the phylogenetic tree constructed based on the nucleotide sequences of partial exon3 (from 7633 to 8484 of AB013448.1) of Pib from 49 accessions.All accessions of Pib were obtained from GenBank.

Figure 4 .
Figure 4. Phylogenetic tree constructed with the nucleotide sequences of Pib gene and different partial CDS regions from wild rice and O. sativa using minimum evolution method of MEGA X.The numbers associated with individual branches indicate confidence levels based on 1000 bootstrap replicates.(A), structure of Pib from AB013448.1 (GenBank ID); (B), the phylogenetic tree constructed based on the nucleotide sequences of exon1 of Pib regions from 34 accessions.(C), the phylogenetic tree constructed based on the nucleotide sequences of partial exon3 (from 7633 to 8484 of AB013448.1) of Pib from 49 accessions.All accessions of Pib were obtained from GenBank.

Figure S5 .
Sliding window of positive-selection sites of the AVR-Pib alleles under M8, M8a and M7 models.The Y-axis indicates the ratio of the rate of nonsynonymous substitution (Ka) to the rate of synonymous substitution (Ks) (Ka/Ks); the X-axis indicates the position of the AVR-Pib amino acids in the site.Figure S6.Possible scenario for M. oryzae AVR-Pib alleles-rice Pib alleles interactions and co-evolution.AVR-Pib homolog L1 genotypes (H01) were derived from an ancestral M. oryzae gene.L1 genotypes are not recognized by Pib in wild rice.In response to this situation, an 87bp insertion in exon1 of Pib in cultivated rice evolved and can recognize L1 genotypes (H01) of AVR-Pib alleles.Then, another AVR-Pib alleles with TE insertion (L2/L3 genotypes), H08 and H11, were derived that cannot be recognized by Pib in cultivated rice.Figure S7.Disease reaction of Magnapothe oryze isolate on rice leaves.Number of 0 to 5 on the label on top of the Figure indicates 0-to-5 disease scale.

Table 1 .
Frequency of AVR-Pib genotypes and avirulent isolates of Magnaporthe oryzae collected from Yunnan, China to IRBLb-B.
a L1 indicates the AVR-Pib genotype with the expected size (1231 bp), L2 and L3 indicates the AVR-Pib genotype with TE insertion (L2 with 3100 bp and L3 with both 1231 bp and 3100 bp).The frequencies in bracket.The superscript of A and B indicates the significant difference at 0.01 level, and * indicates non significant.b Indicates pathogenicity assay on monogenic line IRBLb-B containing Pib. XI and GJ indicates Xian/Indica and Geng/Japonic, respectively.The frequencies in bracket.The superscript of A, B and C indicates the significant difference at 0.01 level; ** indicates significant difference at 0.01 level.

Table 2 .
Haplotypes of AVR-Pib loci in Magnaporthe oryzae field populations of Yunnan, China.
a Indicates the same with KM887844 (GenBank Accession No.).The KM887844 of AVR-Pib was obtained from GenBank.rev: indicates reverse insertion of Pot3 in AVR-Pib.

Table 3 .
Variation of the AVR-Pib loci proteins in M. oryzae populations of Yunnan, China.
a Indicates the same with KM887844.b Indicates pathogenicity assay on the monogenic lines IRBLb-B containing the resistant gene of Pib.R, M and S indicate disease reaction were resistant, moderate resistant and susceptible, respectively; Ex. 24R indicated 24 isolates were avirulent to IRBLb-B; and ?indicates unknown.c rev: indicates reverse insertion of Pot3 in AVR-Pib.

Table 4 .
Haplotype distribution of AVR-Pib in different Yunnan rice-growing regions.
a Number and frequency (in bracket) of isolates of each haplotype.b Diversity index was calculated as the frequency of haplotypes types in the M. oryzae population following Fontaine's method