Identification and Characterization of Crr1a, a Gene for Resistance to Clubroot Disease (Plasmodiophora brassicae Woronin) in Brassica rapa L.

Clubroot disease, caused by the obligate biotrophic protist Plasmodiophora brassicae Woronin, is one of the most economically important diseases of Brassica crops in the world. Although many clubroot resistance (CR) loci have been identified through genetic analysis and QTL mapping, the molecular mechanisms of defense responses against P. brassicae remain unknown. Fine mapping of the Crr1 locus, which was originally identified as a single locus, revealed that it comprises two gene loci, Crr1a and Crr1b. Here we report the map-based cloning and characterization of Crr1a, which confers resistance to clubroot in Brassica rapa. Crr1aG004, cloned from the resistant line G004, encodes a Toll-Interleukin-1 receptor/nucleotide-binding site/leucine-rich repeat (TIR-NB-LRR) protein expressed in the stele and cortex of hypocotyl and roots, where secondary infection of the pathogen occurs, but not in root hairs, where primary infection occurs. Gain-of-function analysis proved that Crr1aG004 alone conferred resistance to isolate Ano-01 in susceptible Arabidopsis and B. rapa. In comparison, the susceptible allele Crr1aA9709 encodes a truncated NB-LRR protein, which lacked more than half of the TIR domain on account of the insertion of a solo-long terminal repeat (LTR) in exon 1 and included several substitutions and insertion-deletions in the LRR domain. This study provides a basis for further molecular analysis of defense mechanisms against P. brassicae and will contribute to the breeding of resistant cultivars of Brassica vegetables by marker-assisted selection. Data deposition The sequence reported in this paper has been deposited in the GenBank database (accession no. AB605024).


Introduction
Clubroot is one of the most serious diseases of cruciferous crops in the world. Plasmodiophora brassicae, the causal agent of this disease, is not a fungus but a member of the phylum Cercozoa in the kingdom Rhizaria [1,2,3]. The first record of clubroot in Japan dates from 1892 in cabbage and 1898 in Chinese cabbage [4], and this disease is now considered a major problem in cabbage and Chinese cabbage production in Japan and Korea [5]. The life cycle of P. brassicae is divided into two phases: a primary phase occurring in the root hairs and a secondary phase occurring in the stele and cortex of the hypocotyl and roots [4]. During the secondary phase, secondary plasmodia induce abnormal tissue proliferation of infected roots, leading to the formation of galls (clubs). These symptoms prevent the uptake of water and nutrients, stunting the infected plants and severely reducing crop yield and quality [6]. The primary phase has been observed in both susceptible and resistant plants. In the secondary phase, the development of the plasmodia is quantitatively reduced or delayed in resistant plants [7,8,9]. Because the resting spores released from decayed clubs can survive for many years in soil, agricultural practices such as liming and crop rotation are insufficient to keep crops healthy. In addition, reducing the use of agrochemicals is preferred for the production of vegetables. Therefore, the breeding of resistant cultivars is one of the most efficient ways to control clubroot.
European fodder turnips (Brassica rapa) were identified as sources of resistance and have been used to transfer their clubroot resistance (CR) genes into Chinese cabbage, oilseed rape, and Brassica oleracea [5,10]. The CR trait of European turnip cultivars such as Siloga, Gelria R, and Debra was considered at first to be controlled by a single dominant gene [11]. Yet after the release of more than 50 CR F 1 cultivars of Chinese cabbage so far, the breakdown of resistance, caused by variation in the pathogenicity of P. brassicae, has been reported only recently [12,13]. Several differential tester sets have been used to study the interaction between P. brassicae and hosts [12,13,14,15,16]. Four pathotypes (groups 1 to 4) were identified in Japanese field isolates through the use of two commercial CR F 1 cultivars of Chinese cabbage [12,13]. But since the number and identity of resistance genes in the tester sets are unknown [10], information on the performance or pathotype specificity of CR genes remains limited.
Genetic analysis and quantitative trait locus (QTL) mapping studies have identified at least 8 CR loci in B. rapa, 22 QTLs in B. oleracea, and 16 QTLs in Brassica napus [10,17]. In B. rapa, Crr1 and Crr2 were identified on chromosomes A08 and A01, respectively [18,19]. These two loci were detected by using two P. brassicae isolates, the mild Ano-01 and the more virulent Wakayama-01. Crr1 was necessary for the resistance to both isolates, but plants having Crr1 alone were susceptible to Wakayama-01. Crr2, which by itself does not show any effect against either isolate, was necessary for resistance to Wakayama-01 in combination with Crr1. Therefore, Crr1 may play a role in a common pathway of resistance, and Crr2 may be a modifier locus for the resistance expressed by Crr1 [19]. Four CR loci-CRa [20,21,22], CRb [23], Crr3 [24,25], and CRk [26]-were identified on chromosome A03. The relationships between common markers and the location of the CR loci suggest that CRa and Crr3 are identical, allelic, or closely linked to CRb and CRk, respectively [10]. CRc [26] was mapped on A02, and a weak QTL, Crr4 [19], on A06. Comparative mapping of CR loci between B. rapa and Arabidopsis revealed that Crr1, Crr2, and CRb are syntenic with the central region of Arabidopsis chromosome 4 [19,25]. Because this region is located within a disease resistance gene cluster, it has been suggested that CR genes are members of these clusters [17,19]. However, although many studies have mapped CR loci in B. rapa, the molecular mechanisms of resistance remain unknown.
In a previous study, we fine-mapped the Crr1 locus and found that it was likely to comprise two gene loci, a major locus for clubroot resistance and another locus with minor effect [3]. We named the former locus Crr1a and the latter Crr1b. Here, we report the map-based cloning and characterization of Crr1a, the major locus for resistance to Ano-01, derived from European fodder turnip 'Siloga'. Crr1a G004 encodes a TIR-NB-LRR disease resistance protein and is expressed in the stele or cortex of hypocotyl and roots, where the secondary infection phase occurs. Transgenic Arabidopsis and susceptible B. rapa harboring Crr1 G004 showed resistance to P. brassicae isolates similar to that of the resistant B. rapa.

Map-based Cloning of Crr1a
We previously fine-mapped the Crr1 locus by analyzing 1920 F 2 plants derived from a cross between clubroot-resistant G004 and susceptible A9709 and found that Crr1 was likely to consist of two gene loci in the region around insertion-deletion (indel) markers BSA2 and BSA7 [3]. We named the major locus for resistance, located near BSA7, Crr1a, and the other locus, with minor effect, around indel marker BSA2, Crr1b (Fig. 1). Here we attempted to delimit the candidate region of the Crr1a locus. A BAC library was screened with BSA7, and three BAC-end markers were developed (Fig. 1). We genotyped these markers in the F 2 population of 3700 plants and found 39 F 2 plants with recombination in the region between BSA7 and BZ2-DraI. A clubroot resistance test using F 3 seeds derived from the F 2 plants revealed that three F 3 populations obtained from F 2 plants in which recombination occurred between markers B355H7 and B359C3 were susceptible to isolate Ano-01 ( Fig. 1). Therefore, Crr1a was estimated to lie within this 8-kb region. We determined the sequence of BAC clone 208F8. 59-RACE (rapid amplification of cDNA ends) and 39-RACE experiments revealed that four open reading frames (ORFs) predicted in this region formed a single gene with similarity to a TIR-NB-LRR-type R (resistance) gene.
To determine whether this candidate gene confers resistance to P. brassicae, we connected a full-length cDNA of Crr1a G004 cloned by RT-PCR with the Lactuca sativa Ubiquitin (LsUbi) promoter and transferred it into clubroot-susceptible Arabidopsis Col-0. The Crr1 locus was effective against Ano-01 but not against the more virulent Wakayama-01 [18]. Therefore, we tested T 2 seeds derived from 11 independent transformants (T 1 ) for resistance to Ano-01 and Wakayama-01. Col-0 plants and LsUbi promoter::GUS lines (UGU_07, 08) were fully susceptible to Ano-01 ( Fig. 2A, B). Nine T 1 lines were highly resistant to Ano-01 (mean disease index [DI] ,1.0) (Fig. 2C) and 2 were intermediate (mean DI .1.0). In contrast, 7 T 1 lines showing resistance to Ano-01 were fully susceptible (DI .2.6) to Wakayama-01 (data not shown). Three T 3 lines derived from resistant T 2 plants homozygous for the Crr1a G004 transgene (UpCc_02, UpCc_04, UpCc_17) were also resistant to Ano-01 and susceptible to Wakayama-01 (Fig. 2D). For further verification, we expressed the full-length cDNA of Crr1a G004 under the control of its own 2.5-kb upstream region; all 11 T 2 lines tested were highly resistant to Ano-01 (Fig. S1).
Previously, using two CR F 1 cultivars of Chinese cabbage-CR Ryutoku and SCR Hiroki-as differential hosts, we classified Japanese field isolates into four pathotypes: group 1, which was virulent on both cultivars; group 2, which was virulent only on CR Ryutoku; group 3, which was virulent only on SCR Hiroki; and group 4, which was not virulent on either [13]. Ano-01 was classified into group 4, while Wakayama-01 was classified between groups 1 and 2 because of an intermediate response by SCR Hiroki. We inoculated two transgenic T 2 lines homozygous for the Crr1a G004 transgene with a representative isolate of each pathotype and found that Crr1a G004 was effective against groups 2 and 4, but not against groups 1 and 3 (Fig. S2). These results were similar to those of B. rapa with a homozygous Crr1 locus. Thus, these results strongly indicate that this candidate gene confers resistance to P. brassicae isolates and is Crr1a.

Crr1a Encodes a TIR-NB-LRR-type Disease Resistance Protein
Comparison of the genomic DNA and cDNA sequences revealed that Crr1a G004 consists of 4 exons and 3 introns. Crr1a G004 encodes a putative 1225-aa protein consisting of an N-terminal Toll-Interleukin-1 receptor (TIR) domain, a nucleotide-binding site (NB) domain, and leucine-rich repeats (LRRs) at the Cterminus of the NB domain (Fig. 3). Conserved motifs characteristic of the NB domain-the P-loop, kinase-2, RNBS-B, GLPL, and MHDV domains-were also identified [27]. In the LRR region, at least 11 imperfect repeats with lengths of 22 to 24 aa were identified. A BLAST search revealed that the predicted Crr1 G004 protein showed similarity to At5g11250 (58% identity), which is a TIR-NB-LRR-type R gene in the TNL-G subgroup [27]. Among known R genes, the predicted Crr1a G004 protein showed moderate similarity to Arabidopsis RPP1s (55% identity), which confer resistance to the biotrophic oomycete Peronospora parasitica (downy mildew) [28].

Structural Differences between Resistant and Susceptible Alleles
We determined the nucleotide sequence between markers B355H7 and B359C3 in the susceptible A9709 ( Fig. 1) and compared it with that in the resistant G004. Crr1a A9709 had three large insertions: a 357-bp insertion 37 bp downstream of the start codon, and 333-and 4982-bp insertions in exon 4 (Fig. 4). The 4982-bp insertion, 157 bp upstream of the termination codon of Crr1a G004 , showed high similarity to a Ty1-copia long terminal repeat (LTR) retrotransposon comprising two identical 171-bp LTRs bordered by 6 bp of target site duplication (TSD) (Fig. S3).
Interestingly, 59-and 39-RACE analysis placed the 59 end of Crr1a A9709 25 bp downstream of the 39 end of the 357-bp insertion (Fig. 4) and the 39 end in the first LTR region of the retrotransposon (Fig. S3). Translation from the first available inframe start codon would yield an NB-LRR protein lacking 80 residues of the TIR domain and 49 residues of the C-terminal domain (Fig. 3). Comparison of the predicted proteins of both alleles showed that the remaining TIR and NB domains were well conserved, but the LRR domain was highly variable (56% identity; Fig. 3). The LRR domain of the predicted Crr1a A9709 protein included 50 substitutions, 2 deletions, and insertions of 2 and 111 residues relative to the predicted Crr1a G004 protein.
We found a highly homologous sequence (Bra020861) in the reference B. rapa genome sequence derived from a clubrootsusceptible line, Chiifu-401 [29], and considered that it is an allele of Crr1a. Comparison of the nucleotide sequences between Crr1a A9709 and Crr1a Chiifu-401 revealed that the sequences between markers B355H7 and B359C3 were almost identical except in the insertion in exon 1. Interestingly, large insertions in exons 1 and 4 were also found in Crr1a Chiifu-401 , both with high similarity to the LTR retrotransposon (Fig. 4, Fig. S3). Although the sequence similarity and domain order suggest that the insertion in exon 1 of Crr1a Chiifu-401 is a Ty1-copia LTR retrotransposon, it is not probably an intact retrotransposon, because it has no start codon and it has an inverted repeat of LTRs at the 59-end. Because the 357-bp insertion in Crr1a A9709 showed high similarity to the LTR region at the 59-end of the insertion in Crr1a Chiifu-401 and TSDs were well conserved between both alleles (Fig. 4), it is likely that the 357-bp insertion is a solo-LTR.

Expression Analysis of Crr1 G004
Semi-quantitative RT-PCR revealed Crr1a G004 transcripts in roots and leaves of the resistant R4-8-1 (see ''Experimental procedures'') and the susceptible A9709, more so in the former (Fig. 5A). We examined the expression of Crr1a G004 in transgenic Arabidopsis carrying a Crr1a G004 promoter::GUS reporter gene Map-based cloning of Crr1a: a genetic map around the Crr1 locus and graphical genotypes of F 3 populations in which recombination occurred between AT27 and BZ2-DraI. The major locus for resistance to Ano-01 was predicted around BSA7, and another locus with minor effect was predicted around BSA2 (Suwabe et al., 2011). We named the former locus Crr1a and the latter Crr1b. Gray bars indicate BAC clones isolated by using BSA7 as a probe. Black bars, homozygous for resistant G004 allele; white bars, homozygous for susceptible A9709 allele. Phenotypes of resistance (R) and susceptibility (S) of F 3 populations to Ano-01 are indicated to the right of the graphical genotype, with the number of F 3 populations of each in parentheses. The 3 F 3 populations in which recombination occurred between B359C3 and B355H7 were susceptible to Ano-01. The position of Crr1a was delimited to an 8-kb region between B355H7 and B359C3. The predicted ORFs of the Crr1a candidate are shown below. doi:10.1371/journal.pone.0054745.g001 construct. In 12-day-old seedlings, the construct was expressed in the stele and cortex of the primary root and hypocotyl, the vascular bundles of the cotyledons and leaves, and at the center of rosettes in a region corresponding to the shoot apical meristem (Fig. 5B-D). No expression was detected in root hairs, the root cap, or the adjacent elongation zone (Fig. 5B-D).
In the course of cloning the full-length cDNAs by RT-PCR, we obtained two clones, each containing a cryptic intron, one in exon 2 and the other in exon 4. This result indicates that Crr1a G004 is alternatively spliced like other TIR-NB-LRR-type R genes [30,31]. RT-PCR with primers flanking introns 2 and 3 produced faint bands in addition to the major band. We cloned and sequenced the PCR products and, in addition to the regular transcript with introns spliced out (RT-a and RT-d in Fig. 5E), obtained a longer transcript with a retained intron 2 (RT-b). In addition, the shorter PCR product obtained with primers flanking introns 2 and 3 represented transcripts in which a cryptic intron within exon 2 or exon 4 was spliced out (RT-c and RT-e, respectively). No PCR products with retained introns 1 and 3 were obtained. Because the retained and cryptic introns gave rise to inframe stop codons or a frame shift, all three alternative transcripts encode truncated TIR-NB proteins.

Characterization of Transgenic B. rapa Plants Expressing Crr1a cDNA
To assess the ability of the Crr1a G004 allele to confer resistance in susceptible B. rapa, we inserted the Crr1a G004 promoter::Crr1a G004 cDNA construct into a susceptible cultivar of B. rapa. T 1 seeds of 15 independent transgenic plants (T 0 ) were tested for resistance to Ano-01. Five of the T 0 lines produced resistant T 1 plants (DI = 0 or 1; Table 1, Fig. S4). Other T 1 plants with the transgene showed severe (DI = 3.0) or moderate swelling (DI = 2.5) on the main roots. Because Crr1a G004 was expressed in both roots and leaves, we analyzed the expression of the transgene in leaves of resistant and susceptible T 1 plants derived from two resistant T 0 lines. Expression of the transgene in the resistant plants was 2 to 3 times that in the susceptible plants (Fig. 6). Similar results were observed in 2 independent tests. These results indicate that Crr1a G004 functions in susceptible B. rapa and dose-dependently confers resistance to Ano-01.

Discussion
The Crr1 locus was originally identified as a single locus for resistance to clubroot isolate Ano-01 [19]. Fine mapping of this locus revealed two gene loci [3], Crr1a, with a major effect, and Crr1b, with a minor effect. We have now cloned and characterized Crr1a G004 . Recent findings show that a pair of NB-LRR genes function together in disease resistance [32]. In contrast, our finding that Crr1a G004 alone could confer resistance to Ano-01 in transgenic Arabidopsis and B. rapa suggests that Crr1b is not required for Crr1a G004 -mediated resistance.
Resistance to the obligate biotroph protist P. brassicae also might be mediated via a gene-for-gene interaction, called effector-triggered immunity [33]. Expression of Crr1a G004 was detected in the stele and cortex of hypocotyl and root, where secondary infection occurs, but not in root hairs, where primary infection occurs (Fig. 5B-D). These results are consistent with previous histological findings of the incompatible interaction [7,8,9,10]. Because primary infection occurs in both resistant and susceptible plants [10], this phase is not likely to be associated with resistance. Tanaka et al. [9] reported that the plasmodia remained immature with a small number of nuclei and did not form resting spores even by 40 days after inoculation of Kukai 70, a CR cultivar of Chinese cabbage. Similar findings were also reported in resistant radish cultivars and accessions of Arabidopsis [7,8,9]. Therefore, Crr1a G004 may inhibit plasmodial development during the secondary infection phase.
Gain-of-function analysis proved that Crr1a G004 alone confers resistance to Ano-01 in susceptible Arabidopsis and B. rapa. Resistant and susceptible T 1 plants of B. rapa segregated, with the level of resistance dependent on the expression level of the transgene (Fig. 6). It is likely that T 1 plants homozygous for the transgene were resistant and plants heterozygous were susceptible. Our unpublished data suggest that Crr1 is incompletely dominant, because genetic analysis of an F 2 population derived from a cross between G004 and A9709 showed that a heterozygous Crr1 locus was insufficient for complete resistance to Ano-01. Our results here suggest that a threshold level of Crr1a expression is required for complete resistance, and the level of expression may explain the incomplete dominance of Crr1. However, we cannot rule out the possibility that a low level of susceptible Crr1a protein acts as a dominant-negative regulator of Crr1a G004 in heterozygous plants.
In the case of the tobacco N gene, a member of the TIR-NB-LRR class of R genes, loss-of-function alleles caused by TIR deletion or point mutations in TIR interfered with the wild-type N function in heterozygous plants [34]. In Arabidopsis, all T 2 plants with the transgene showed resistance, and Crr1a G004 behaved as a dominant resistance gene (Fig. 2). This different behavior may be due to the difference in inoculation methods between Arabidopsis and B. rapa. The level of expression of the transgene in heterozygous plants might be high enough to confer resistance in to lower concentration of resting spores.
The genomic regions of B. rapa adjacent to Crr1, Crr2, and CRb are syntenic with Arabidopsis chromosome 4 [19,23,25]. Therefore, these CR loci may have been derived from the same region of the ancestral genome, and triplicated and dispersed among 3 chromosomes in B. rapa during the evolution of the Brassica genome [5]. The genomic region around Crr3 is syntenic with Arabidopsis chromosome 3, and this locus is thought to have a different origin from Crr1, Crr2, and CRb [25]. Here, however, we found that Crr1a G004 showed the highest similarity to At5g11250 on Arabidopsis chromosome 5. Because two additive QTLs controlling partial resistance to clubroot in Arabidopsis accession Bur-0 have been identified on chromosome 5 and one of them colocalized with several clusters of resistance genes [35], either locus might be a functional ortholog of Crr1a and the ancestor of CR genes. Sequence information on Crr1a G004 will accelerate the cloning of other CR genes identified in Brassica and Arabidopsis, and may clarify the evolution of CR genes. In fact, a BLASTP search of the B. rapa genome sequence revealed a large number of predicted proteins with partial similarity to Crr1a G004 , some near markers linked to other, previously reported CR loci (data not shown).
The susceptible Crr1 A9709 allele had the solo-LTR in exon 1 and appeared to generate a truncated NB-LRR protein lacking more than half of the TIR domain (Fig. 3). The TIR domain of the NB-LRR protein plays an important role in the induction of defense responses or in recognition specificity for the pathogen effector [36,37,38]. Therefore, it is likely that deletion of the TIR domain abolishes the Crr1a A9709 function and results in susceptibility. The large LTR retrotransposon-like insertion in exon 1 was also found in the susceptible Crr1a Chiifu-401 allele, and the insertion site was conserved between Crr1a A9709 and Crr1a Chiifu-401 . These results suggest that both susceptible alleles are derived from the same retrotransposon insertion event, which causes loss of resistance, and that this retrotransposon has a role in the differentiation of CR genes. However, this insertion is not the sole cause of susceptibility alleles in B. rapa, because PCR analysis using primers flanking this insertion revealed that 18 of 24 non-CR cultivars of Chinese cabbage and turnips tested did not have such an insertion (data not shown). Furthermore, since the LRR domain plays an important role in recognition specificity [37], we cannot exclude the possibility that indels or substitutions in the LRR domain result in susceptibility.
Crr1 G004 was alternatively spliced (Fig. 5E). Alternative splicing is a common feature of NB-LRR-type R genes. Although its biological role in the function of R genes is still unknown, transcript variants of tobacco N and Arabidopsis RPS4 are required for complete resistance, and the expression of the alternative transcripts is induced during defense responses [30,31]. Our transgenic plants with the full-length Crr1 cDNA, which did not produce truncated transcripts by alternative splicing, were resistant to Ano-01 (Fig. 2, 6); this result suggests that transcript variants are not necessary for Crr1 G004 function. A similar finding was also reported in relation to the flax rust resistance gene L 6 [39].
An effective measure to increase the durability of resistance is the pyramiding of 3 or more CR loci in a single cultivar by marker-assisted selection [5]. Our results will enable the development of Crr1a-specific or closely linked markers that will improve the efficiency of marker-assisted selection and avoid the introduction of undesirable traits into improved cultivars by linkage drag [40]. Furthermore, knowledge of the pathotype specificity of CR genes is important. Crr1a G004 was effective against pathotypes of groups 2 and 4, but not groups 1 and 3 (Fig. S2). The CR line PL9, with both Crr1 and Crr2, is effective against isolate No. 5 of group 1 [41], and Wakayama-01 distinguished groups 1 and 2 [18], indicating that Crr2 is useful for resistance to the more virulent isolates. Because Crr1 comprises Crr1a and Crr1b, further analysis of the relationships among Crr1a, Crr1b, and Crr2 in resistance to different pathotypes is necessary. Recently, Kato et al. [41] showed that CRb was effective against isolates of group 3. The pyramiding of Crr1a G004 and/or Crr1b, Crr2, and CRb could confer resistance to most isolates currently present in Japan. In this study, we used field isolates, which are regarded as heterogenic; and multi-pathogenic races have been found in a single field [42,43]. Single-spore-derived isolates (SSIs) are considered to be most valuable for the genetic study of resistance and virulence surveys of pathogens [43]. The availability of single CR genes and SSIs of P. brassicae will allow us to differentiate pathogenicity and pathogenhost interactions more precisely.

Plant Materials
A clubroot-resistant (CR) doubled-haploid (DH) line, G004, with Crr1 and Crr2, and a susceptible DH line, A9709, were used as parents for the F 2 population [19]. To develop a CR inbred line with Crr1 and Crr2 in the A9709 background, we selfed one of these CR F 2 plants to generate a CR F 3 plant, and backcrossed that line three times to A9709 to generate BC 3 F 3 plants. During this process, SSR markers linked to Crr1 and Crr2 ) were used to select plants with both loci in each generation. Plants homozygous for the G004 alleles of the Crr1 and Crr2 loci were selected from the self progeny of the BC 3 F 3 plant by SSR genotyping; one of the selected plants was named R4-8-1.

Test for Clubroot Resistance
The P. brassicae field isolates Ano-01, Wakayama-01, and Nos. 5, 7, 9, and 14 were used [12,13,18] to test clubroot resistance in A. thaliana according to Jubault et al. [35] with modifications: T 2 seeds were plated on MS medium containing kanamycin to select plants carrying the transgene. Ten-day-old kanamycin-resistant T 2 plants were transplanted into soil and inoculated by the injection of 2-4 mL of a resting-spore suspension (1.0-1.5610 6 spores/mL) into the soil near the roots. Inoculated plants were grown in a controlled environment under 14-h light/10-h dark at 23/18uC. Resistance responses were evaluated 3 to 4 weeks after inoculation, and the disease index (DI) was scored on a scale of 0 to 3 (Fig. S5). The mean DI of each T 1 line was expressed as the mean of two or three clubroot tests (9 T 2 seedlings per test) on different dates. Clubroot tests of B. rapa were carried out by the insertion method [11,12]. Each of 2 or 3 tests, performed on different dates, used 8 T 1 seedlings derived from an independent transgenic B. rapa (T 0 ). We determined whether the T 1 plants had the transgene or not by PCR analysis before evaluating resistance. Root symptoms were graded as: 0, no symptoms; 1, very slight swelling on main roots; 2, a small gall on main roots; 2.5, moderate swelling on main roots; 3, severe swelling on main roots (Fig. S4).

Map-based Cloning of Crr1
A BAC library constructed from the resistant G004 was screened with markers BSA2, AT27, and BSA7 as anchors, and a BAC contig covering 0.6 cM between BSA2 and BSA7 was assembled [3]. We selected F 2 plants in which recombination occurred in the interval between markers developed from BACend sequences. F 3 plants derived by selfing of the selected F 2 plants were tested for resistance to Ano-01. To sequence the candidate region for Crr1a, we shotgun-sequenced BAC clones, 355H7 and 208F8. Sequences were assembled with Sequencher v. 2 software (Hitachi). ORFs were predicted with Genetyx v. 10 software (Genetyx).

Vector Construction and Transformation
A 3.7-kb coding sequence of Crr1a G004 was amplified from firststrand cDNA synthesized from root poly(A) + RNA. The Lactuca sativa Ubiquitin (LsUbi) promoter::Crr1a G004 and Crr1a G004 promoter::Crr1a G004 constructs were generated from the pU-C198AAUGU plasmid [44], in which the 1.9-kb LsUbi promoter, the GUS gene, and the 0.6-kb LsUbi terminator were inserted in a modified pUC19 vector, pUC198AA [44]. The amplified Crr1a G004 cDNA replaced GUS between the LsUbi promoter and terminator in pUC198AAUGU to generate the LsUbi promoter::Crr1a G004 construct. A 2.5-kb 59-upstream sequence of Crr1a G004 replaced the LsUbi promoter::Crr1a G004 construct to generate the Crr1a G004 promoter::Crr1a G004 construct. The gene cassettes were ligated into the plant binary vector pZK3B [45]. These constructs were transformed into Arabidopsis Col-0 using Agrobacterium tumefaciens strain GV3101 by the floral dip method [46]. For B. rapa transformation, a commercial F 1 cultivar, Osome (Takii Seed Co.), was used as the recipient, and hypocotyl explants were transformed with A. tumefaciens strain GV3101 as described [47]. Primer sequences used for vector construction are listed in Table S1.

GUS Assay Construct and Histochemical Staining
The 2.5-kb 59-upstream sequence of Crr1a G004 replaced the LsUbi promoter of pUC198AAUGU to generate the Crr1a G004 promoter::GUS construct. The resulting construct was introduced into Arabidopsis Col-0, and T 2 plants were used in the reporter gene assay. The GUS assay was carried out according to Ariizumi et al. [48].   Table 1. Clubroot resistance of the transgenic Brassica rapa (T 1 ) carrying Crr1a promoter::Crr1a cDNA construct. column DNase I treatment, and then converted into first-strand cDNA with a SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen). Semi-quantitative RT-PCR for Crr1a G004 was performed using cDNA synthesized from RNA isolated from plants at 14 days after inoculation with Ano-01. To test splicing variation, we performed PCR using primer pairs flanking introns 2 and 3 and cloned the PCR products into the pGEM-T Easy vector (Promega) for sequencing. RACE was performed with a BD SMART RACE cDNA Amplification Kit (TaKaRa Bio); 2 independent RACE experiments were carried out. Primer sequences used for these amplifications are listed in Table S1.

Real-time PCR Analysis
T 1 seedlings derived from 4 T 0 lines and wild-type B. rapa 'Osome' were inoculated with Ano-01, and total RNA was isolated from leaves 4 weeks after inoculation. Transcript levels of the Crr1a G004 transgene were analyzed in 2 independent clubroot tests by real-time PCR using SYBR Premix ExTaq II (TaKaRa Bio). BrACT2 was used as an internal control for normalization [49]. Primer sequences used for these amplifications are listed in Table  S1.  Figure S4 Typical root symptoms of transgenic Brassica rapa inoculated with Ano-01. Resistance responses were evaluated 5 weeks after inoculation. Root symptoms were graded as: 0, no symptoms; 1, very slight swelling on main roots; 2, a small gall on main roots; 2.5, moderate swelling on main roots; 3, severe swelling on main roots. Scale bar indicates 5.0 cm. (TIF) Figure S5 Typical root symptoms of Arabidopsis (Col-0) inoculated with P. brassicae. Resistance responses were evaluated 3 weeks after inoculation. Root symptoms were graded as: 0, no symptoms; 1, very slight swelling on lateral roots; 2, moderate swelling on lateral roots and taproot; 2.5, severe swelling on all roots but no swelling on hypocotyl; 3, severe swelling on all roots and hypocotyl. Scale bar indicates 1.0 cm. (TIF)