1. Introduction

Clubroot causes severe yield losses of brassica oil and vegetable crops world-wide ^[1]. The organism responsible for this disease is Plasmodiophora brassicae. Although originally referred to as a fungus, P. brassicae is a plasmodiophorid protist ^[2][3]. Plasmodiophorids contain chitin in the cell wall of their resting spores, but they are not related to fungi and unlike fungi or oomycetes do not show filamentous growth. Plasmodiophorids are part of the highly diverse eukaryotic group Rhizaria and belong to the Phytomyxea, a group of Endomyxa (Retaria) ^[4][5][6][7]. The Phytomyxea consist of obligate biotrophic parasites of brown algae, oomycetes (Phagomyxids) and diverse range of plant hosts (Plasmodiophorids) ^[8]. Due to the high agricultural and economically damage ^[9], P. brassicae is the best studied plasmodiophorid, although other species have a high impact on agriculture as well, such as Spongospora subterranea and Polymyxa species. After the initial infection when zoospores encyst on the roots and inject themselves into the host cells, the root cortex is colonized. Plasmodiophorids are rare examples of plant pathogens that reside entirely inside their hosts where they multiply and form new resting spores ^[10][11]. Once re-released into the soil, P. brassicae can render infested fields unsuitable for brassica crop cultivation due to the persistence of resting spores in the soil for up to 20 years ^[12][13][14]. Chemical control for this soil borne disease is not possible at present and cultural practices, such as long crop rotation times, can only limit the soil infestation with P. brassicae. However, long crop rotation times are often not economical feasible especially for the cultivation of oilseed rape ^[13][14]. Therefore, the development of resistant cultivars is considered the most economical and efficient method for clubroot control ^[14]. Breeding for clubroot resistant plants has its own challenges, as it is work and time intensive, and resistance can be broken ^[15][16][17]. A deeper understanding of the molecular interaction between P. brassicae and its hosts, would facilitate the developing of new breeding and management strategies. Due to the truly intracellular lifestyle of P. brassicae, clubroot is a complicated system to study and research of this plant pathogen system lacks somewhat behind other plant–pathogen relationships. Most research has also been made using field isolates (P. brassicae resting spores collected from an infested plant or field) which might be heterogenic, and not with single spore isolates (SSI), a population derived from a host infected with a single spore.

The first P. brassicae genome sequence from the European SSI e3 originally isolated from Brassica rapa was only published in 2015 (e3_2015) ^[18]. At the time, this was also only the third species of Rhizaria with genome information, which is one of the eukaryotic groups with the least molecular data [4]. By now genome drafts for the plasmodiophorids Polymyxa betae ^[19] and Spongospora subterranea ^[20] are also published. A further 48 P. brassicae genomes were assembled ^{[21][22][23][24][25]} and deposit in the NCBI genbank (https://www.ncbi.nlm.nih.gov/genome/browse/#!/eukaryotes/38756/ access on 10-11-2020). The genomes described in [24] were assembled based on the e3_2015 reference genome, whereas the other genomes described in this review were assembled de novo. The most recent assembly combined long- and short-read sequencing and accomplished a nearly complete assembly of the e3 genome (e3_2018). It consists of only 20 contigs (of which 13 are assembled chromosomes from telomere to telomere) with a total size of 25.1 Mb and also includes the complete mitochondrial sequence of 114 kbp in length ^[25]. The high quality of the e3_2018 genome will facilitate the reference-based genome assembly of additional P. brassicae isolates in the future. Assemblies in the NCBI database range between 24.05 and 25.25 Mb. The small size is due to a low presence of repeated sequences (2–5%) and a reduction of intergenic elements in the genome. The number of protein coding genes is around 10,000, but gene models are only published for the e3 genomes. Many of the predicted proteins of P. brassicae do not show high similarities to protein models of other species or do not contain known functional domains, making the prediction of their function difficult ^[25][26].

As the protists cannot be cultivated without host, gene studies using reverse genetic methods (i.e. constructing of gene knock-out mutants) cannot be applied to study gene function. Thus the P. brassicae gene function remains for most cases hypothetical, despite the genome information. However, the genome information and transcription studies gave some insights into the pathogen metabolism (for a review see ^[27]). The clubroot pathogen appears to be dependent from host metabolites as the genome appears to be contain several incomplete metabolic pathways, a characteristic common with other eukaryotic biotrophic plant pathogens ^[28][29]. The missing genes encode proteins involved in sulfur and nitrogen uptake, and arginine, lysine, thiamine, and fatty acid biosynthesis pathways. In addition, only a few carbohydrate active enzymes (CAZymes), involved in the synthesis, metabolism, and transport of carbohydrates, were found. The P. brassicae genome contains genes potentially able to manipulate plant hormone metabolism, such as the auxin-responsive Gretchen Hagen 3, isopentenyl-transferases, a SABATH type methyltransferase and cytokinin oxidase ^[30]. The investigation of proteins associated with lipid droplet organelles or protein families such as the E3 ubiquitin ligases of P. brassicae ^[31] or immunophilins ^[32] also benefitted from the available genome information.

However, despite the presence of the genomes, most transcriptional studies just focus on the host response to an infection by P. brassicae and ignoring the information of the P. brassicae gene expression. Even though the P. brassicae gene expression pattern will be mainly descriptive, it contains important information. Even without functional domains encoded in the proteins, P. brassicae candidate genes can be selected from the transcript information for further studies to better understand how they manipulate the host and gives insights about how the metabolism of the pathogen changes ^{[33][34][35][36]}. Jiang et al. ^[35] did report differential expression of identified effector candidates in Canadian P. brassicae isolates 5I and 5X in resistant and susceptible B. napus hosts. Thus, the regulation of effector genes might lead to a host specific virulence of different isolates and should be investigated in the future. To identify differences in the gene regulation of effector candidates and other genes, analyses of the P. brassicae transcripts in more transcriptomic studies would be very helpful. That information should not be ignored, to better understand the clubroot disease and therefore also the resistance of the hosts. Additionally, it should be considered that the transcriptional host response is different in root tissue that is colonized by the pathogen than in P. brassicae free tissue ^[36] in a cell- and stage-specific manner ^[37], so that transcriptional analyses of whole roots contain diluted information about host response and pathogen gene expression.

2. Genomes and Pathotypes

The genomes sequences enable now comparative analyzes between different P. brassicae isolates. A comparative analyses of P. brassicae isolates is of high interest as P. brassicae exists in different pathotypes or races. The pathotypes are distinguished by the ability to infect different Brassica species or causing more severe disease symptoms and overcome resistance on certain Brassica hosts compared to other hosts. Knowing the pathotype that is present in the soil of a certain field, would be great advantage as farmer could chose to grow a crop variety which is less susceptible to the present P. brassicae isolate and thereby diminishing anticipated crop losses. To date the pathotype is determined by work and time intensive bioassays, which test the grade of infection on a set of hosts thereby identifying the ability of a P. brassicae isolate to infect different plant host genotypes harboring resistance genes. Currently different host sets are used internationally, such as the European Clubroot Differential system (ECD) [38], and pathotyping according to Somé [39] and Williams [40]. Additional adaptations were made using regional economical important hosts to fit local needs, such as the Canadian Clubroot differential set (CCD) with a focus on rapeseed resistance [41]. Other systems were focusing on Chinese cabbage resistance [42–44]. Those different systems make it difficult to compare pathotypes between studies, as the pathotype determined by one system cannot be translated into another system. However, the CCD system assigns P. brassicae isolates based on their Williams classification, along with a letter denoting their virulence pattern on the additional hosts of the CCD set and also includes the differential hosts of Somé ^{[38][39][40][41][42][43][44][45][46][47][48][49][50][51][52][53][54][55][56][57][58][59][60][61][62][63][64][65][66][67][68][69]}.

Other systems focusing on Chinese cabbage resistance Different pathotypes occurring dominant in different regions or areas in the world. The Williams pathotype 3 appears to be dominant in Alberta, Canada [45,46], and Korea [43], whereas Williams pathotype 4 is dominant in China [42]. Using the ECD system dominant pathotypes were also determined in Australia as 16/3/12 and 16/3/31 [47]. In Germany, the P. brassicae isolates with the Somé pathoypes 1 and 3 or ECD pathotypes 16/31/31, 16/14/30, and 16/14/31 were most frequently found [17]. The occurrence of pathotypes is somewhat fluent and new pathotypes become present in fields, and there is variation of virulence inside a pathotype, when tested on additional hosts [15,42–44,46]. In addition, P. brassicae field isolates and even isolates from an individual plant root can consist out of a mixture of pathotypes and genetically different strains [48–50]. Thus, the homogeneity of the pathogen material can only be guaranteed if it has been multiplied from a single spore. Most pathotyping is performed with field isolates and the pathotype should be interpreted with caution. In the field, a less prominent P. brassicae pathotype might be present and become prominent, when a different host is cultivated.

Still, replacing the time-consuming bioassays by a fast and cheap molecular distinction between P. brassicae races, would be a huge advantage. Therefore, it must be known if and which sequence variations correlate with the race characteristics of the different isolates. A standardized pathotype system would therefore be beneficial to compare the molecular data from international isolates with each other. Furthermore, for isolates used in molecular studies a pathotype is often not determined. However, a large number of pathotyped isolates derived from different hosts and geographic origins is needed to identify molecular markers of different P. brassicae pathotypes and isolates.

One obstacle for comparing the P. brassicae genome data with pathotype and other information is that many of the sequenced isolates have been named differently in different publications and again differently in the NCBI database. We summarized the currently available P. brassicae genome assemblies, linked with the information about origin, other assigned isolate names and pathotype (if known) in Table 1. The majority of the sequenced P. brassicae strains were isolated from canola and are of Canadian origin. Indeed, within the 43 genomes published recently, two originated in the USA, five from China, and the other sequences were obtained of Canadian isolates [24]. This study also included a number of SSIs and many of the isolates were pathotyped. The first reported Canadian P. brassicae genomes came from a variation of pathotypes [21]. The two genome assemblies deposited in the public databases derived from SSI of Williams pathotypes P3 (AAFC-SK-Pb3) and P6 (AAFC-SK-Pb6) whereas the first Chinese P. brassicae genome derived from SSI of Williams pathotype 1 (ZJ-1) [23]. Currently there are three genome assemblies of P. brassicae from Europe: the original sequence of the SSI e3 [18] and its updated version (e3_2018) [25] and the sequence of the selection isolate eH [22]. The eH isolate has a pathotype P1 according to Somé, but it is not SSI. However, both the isolate e3 and the isolate eH, originally derived from the same isolate "e" from a stubble turnip [51]. As the three European genome sequences come from P. brassicae isolated from the same clubroot, they do not allow a deeper insight into the genomic variation of European or even German P. brassicae isolates. RFLP analyses show a high genomic variation in European isolates [52,53], but to date genomic data are missing to analyze the variation in more detail. Currently additional European P. brassicae non-“e” sequences are only published from transcriptomes of clubroot infected kohlrabi (Brassica oleracea var. gongylodes) from Austria [36].

Table 1. Summary of available P. brassicae genome data. The country and region of origin, the host the isolate derived from, alternative names, and pathotyping results have been retrieved from available literature [15,18,21–25,45,54–59] and information provided by the authors from [24].

Abbreviations: AB: Alberta, BC: British Columbia: PEI; Prince Edward Island, ON: Ontario, MB: Manitoba, QC: Quebec, CHN: China, USA: United States of America, ND: North Dakota, SK: Saskatoon, GER: Germany, CAN: Canada, BRAP: Brassica rapa; BNAP: B. napus; BOLE: B. oleracea; VEG: vegetable, EDC: European Clubroot Differential; CCD: Canadian Clubroot Differential. The numbers and letters in the columns for the spore classifications are based on those in the original publications. An * indicates single-spore isolates (SSI).

3. Gene Variation and Molecular Markers

Several studies tried to associate gene sequences to certain pathotypes or isolates. Comparisons of single nucleotide polymorphisms (SNPs) reveal differences in the genome assemblies of P. brassicae isolates [21,24]. A phylogeny based on SNPs of Canadian, Chinese and P. brassicae isolates from North Dakota (USA) in comparison to the e3_2015 sequence distinguished 5 different groups of P. brassicae, which however did not cluster according to their pathotypes [24]. Other studies looked at specific genes for their specificity of pathotypes. Polymorphism within the 28S rDNA of P. brassicae were reported which potentially could distinguish P. brassicae pathotypes, but unfortunately, the reported variation in LSU sequence of the rDNA of P. brassicae was due to chimeric PCR products of P. brassicae DNA and other soil inhabiting cercozoan species [60–62]. A set of markers was also reported to distinguish Korean isolates with different virulence patterns on clubroot resistant and susceptible cultivars of Chinese cabbage [63]. Markers were selected through sequence characterized amplified region (SCAR) by comparing the whole genome sequences of P. brassicae isolates from Korea with the genome of the e3_2015. However, while primer sequences were published the authors did not provide information about the sequence of the amplified regions or the genome assemblies used in their study, so the Korean sequences cannot be used in comparative studies.

Molecular markers were reported to distinguish the predominant Williams pathotypes P11, P9, P7 and P4 in China [64–66], as well as for P5 [67] and the new emerged pathotype P5X in Canada [68]. For now, it remains difficult to trust the reported PCR assays in [64–66]. Marker genes for the Williams P4 and P9 were identified using the e3_2015 genome and additional identified genes from transcriptome data, which were not predicted in the e3_2015 [64,65]. Unfortunately, the authors did not report the sequences of their new identified genes. It would be of interest to see if the sequences of the reported genes are present in the available P. brassicae genome assemblies and if those markers are useful for pathotype determination and indeed of P. brassicae origin. The authors reported further that the genes encoding for PBRA_003263, PBRA_003268, and PBRA_000003 can identify Williams pathotype P4. However, in the public available genome data the gene sequences for PBRA_000003 is present in all sequenced P. brassicae isolates without any sequence variation, inclusive Williams P1–P3, P5, P6, and P8 pathotypes. In contrast, PBRA_003263 and PBRA_003268 are missing in the AAFC-SK-Pb3 and AAFC-SK-Pb6 assemblies and PBRA_005772 is additionally not in part of the ZJ-1 assembly, but all are present in a variety of pathotypes. If those genes are only present in certain P. brassicae isolates or if the assemblies are incomplete needs to be tested.

In a similar investigation by the same group PCR assays were reported to differentiate other pathotypes, using sequences of novel genes for which the P. brassicae origin was not confirmed, as well as the e3_2015 sequences for PBRA_007750, PBRA_008439, and PBRA_009348. From those genes PBRA_007750 and PBRA_008439 are partially present in the AAFC-SK-Pb6 assembly and present in all other genomes, albeit with sequence variations (Supplementary Data S1). There are no genome assemblies in NCBI databases with the Williams pathotype 4 or 7, but the primer pairs used to amplify the PBRA_007750 sequence would amplify the markers from Williams pathotypes 1–3, 5, 6, and 8 (Supplementary Data S1), but the reported distinction between Williams pathotypes 4 and 7 might be possible. PBRA_009348 is missing in AAFC-SK-Pb3 and has one nucleotide different in P.b-3 and P.b-17, and the sequence is otherwise identical in the genomes of the NCBI database. PBRA_000303, reported to be specific for pathotype P7 [66], is missing in the AAFC-SK-Pb3 and ZJ-1 genome assembly but present without sequence variation in all other genomes.

It might well be that the reported primers for the P. brassicae genes found in all genome sequences in the databases only amplify the genes in the reported Chinese isolates of a certain pathotype. Some reported primers do not appear to match the genome sequences in the NCBI database without mismatches and the primer sequences for PBRA_003263 published in [65] do not match the PBRA_003263 sequences obtained from the genome assemblies deposited in the NCBI database. The lack of information of the retrieved sequences of the Chinese isolates used in the studies above, do not allow to check sequence variations with other strains and pathotypes. The genes are however present in most if not all sequenced Williams pathotypes (P1–3, P5, P6, and P8) and it is therefore questionable if the reported PCR assays can be used to undoubtedly identify P4, P7, and P9. However as there are currently no genome assemblies of P. brassicae isolates from Williams P4, P7, and P9, so the reported markers might be able to differentiate between those pathotypes.

The CR811 (KJ683723.1) gene was reported to be specific for the Canadian P. brassicae isolates of Williams pathotypes P5 and P5X [67]. However, the according CR811 sequence is not part of any of the published P. brassicae genome assemblies, including genome sequences for Canadian isolates of P5 and P5X; thus, it is either missing in the assemblies or not part of the P. brassicae genome and therefore not a specific marker. The origin of the CR811 gene should however be identified, as it could be that a higher virulence is associated with the presence of other microorganisms, which harbor this gene.

Generally, a PCR assay to distinguish P. brassicae pathotypes does have additional obstacles. Pathotype diversity within single root galls appears to be a common occurrence. In a Canadian study 50 of 79 investigated galls consisted of more than one strain [48]. Therefore, the results using a single-gall or field population for pathotyping or molecular research, especially for the identification of pathotype specific markers, should be treated with caution. In a field and even in a single club several different isolates can be present [48,49]. While one isolate of a certain pathotype might be dominant, the PCR assay can still amplify DNA from the less present pathotypes. Additionally, it should also be shown that DNA derived from clubroots or soil can be amplified with a positive control, especially if a marker is supposed to be absent in certain pathotypes. It is likely that false positive or negative results from the PCRs will occur frequently. One solution might be a multiplex PCR assay. Yang et al [69] used two genes that were able to differentiate two groups of P. brassicae isolates via PCR. While this duplex PCR assay could differentiate between P. brassicae isolates that could break resistance on resistant canola cultivar 45H29 or not, the assay could also not determine pathotypes. The study also showed that field isolates are usually mixed population. In field isolates both specific bands were amplified and showed potential for quantitative analyses of different pathotypes in parallel [69].

This entry is adapted from the peer-reviewed paper 10.3390/pathogens10030259