Transgressive segregation reveals mechanisms of Arabidopsis immunity to Brassica-infecting races of white rust (Albugo candida)

Significance Most plants resist most plant pathogens. Barley resists wheat-infecting powdery mildew races (and vice versa), and both barley and wheat resist potato late blight. Such “nonhost” resistance could result because the pathogen fails to suppress defense or triggers innate immunity due to failure to evade detection. Albugo candida causes white rust on most Brassicaceae, and we investigated Arabidopsis NHR to Brassica-infecting races. Transgressive segregation for resistance in Arabidopsis recombinant inbred lines revealed genes encoding nucleotide-binding, leucine-rich repeat (NLR) immune receptors. Some of these NLR-encoding genes confer resistance to white rust in Brassica sp. This genetic method thus provides a route to reveal resistance genes for crops, widening the pool from which such genes might be obtained.


Pathogen strains and maintenance
Albugo candida races Ac2V, Ac7V and AcBoT were propagated as follows: zoosporangia from plants inoculated 14 days earlier were suspended in cold water and incubated on ice for 30 min. The spore suspension was then sprayed on plants using a spray gun, and plants were incubated in a cold room (5°C) in the dark over-night. Infected plants were kept under 10 hrs light and 14 hrs dark cycles with a 21°C day and 14°C night temperature. Albugo candida race Ac2V was maintained on Brassica juncea cultivar Burgonde, Ac7V on B. rapa cultivar Just Right and AcBoT on Brassica oleracea DH1012 (1). Pathogen infection assays were carried out using two weeks old Arabidopsis plants and the inoculations were carried out as described above.

DNA Extraction
For both genetic mapping and RenSeq, gDNA was extracted from young leaves using the DNeasy Plant Mini Kit (Qiagen, Hilden, Germany) following the manufacturer's protocol.

Genetic Mapping
To identify linkage with the WRR locus of interest, DNA extracted from fully susceptible F2 plants derived from the crosses between MAGIC.329 and the indicated Arabidopsis accession were used for genotyping with the molecular markers located on different Arabidopsis chromosome arms that show polymorphism between MAGIC.329 and the indicated Arabidopsis parent. Molecular markers and corresponding primer sequences are in Dataset S1.
Once a linkage established, further molecular markers were developed and used to delineate the WRR locus of interest using additional fully susceptible F2 plants.

Cloning of WRR genes
Genomic fragment of WRR4A Col-0 was amplified from Arabidopsis accession Col-0 with Platinum Pfx DNA Polymerase (Life Technologies, Carlsbad, CA, USA) using the specific primers given in Table S3. Following digestion with BamHI (New England Biolabs, Inc., Ipswich, MA, USA) , this genomic fragment was cloned in to BamHI and SmaI (New England Biolabs, Inc., Ipswich, MA, USA) digested pCambia2300 vector. WRR4B alleles as well as its paralog At1g56520 were amplified from Arabidopsis accessions Col-0 and Ws-2 with Platinum Pfx DNA Polymerase (Life Technologies, Carlsbad, CA, USA) as multiple fragments using the primers containing a BsaI recognition site (Table S3). Using the Golden Gate assembly (2), these fragments were assembled into full-length genes and cloned into the binary vector pICH86966. WRR8 Sf-2 and its paralog At5g46260 from Sf-2, WRR9 Hi-0 and its paralogs At1g63730 and At1g63740 from Hi-0 as well as WRR12 Ler-0 allele from MAGIC.329 were PCR amplified with Kapa Hifi Uracil+ (Roche) using primers supplemented with specific 5' and 3' extensions (Table S3) to make them compatible with custom USER (Uracil-Specific Excision Reagent) expression vectors (3). 30 ng of purified PCR product was then hybridized with 30 ng PacI and Nt.BbvCI (New England Biolabs, Inc., Ipswich, MA, USA) digested USER vector pICSLUS0002 in the presence of 1 µL of USER enzyme mix (New England Biolabs, Inc., MA, USA). All constructs were verified by DNA sequencing. Plasmids containing the candidate genes were then transformed into Agrobacterium tumefaciens strain AGL1. We submitted novel WRR gene sequences to NCBI GenBank (4) (accession numbers): Antibiotic resistant shoots were removed from explants and placed in sterilin jars containing rooting media with 300 mg/L Timentin, 25 mg/L kanamycin and 0.5 mg/L IBA.

RNA extraction, cDNA synthesis and RT-PCR
To determine the expression of WRR8 Sf-2 and WRR9 Hi-0 in transgenic Brassica juncea, total RNA was extracted using the TRI-reagent (Sigma-Aldrich, MO, USA) and Directzol RNA Mini-prep kit (Zymo Research, CA, USA), following manufacturer's recommendations.
First-strand cDNA was made with Invitrogen Superscript III Reverse Transcriptase. The obtained cDNA was used as template for RT-PCR with specific primers listed in Table S3.

Extraction of sequences of WRR alleles from SMRT RenSeq assemblies
Eighteen MAGIC parents as well as Ws-2 were sequenced using the SMRT RenSeq technology (11) as part of a collaborative effort.
http://ftp.tuebingen.mpg.de/ebio/alkeller/pan_NLRome/. We also independently sequenced one of the MAGIC parents Can-0 using the SMRT RenSeq technology and assembled using CANU version 1.3 (12). Full sequencing, assembly statistics and associated quality control steps for the eighteen MAGIC parents and Ws-2 will be published in an upcoming submission.
Reads and assemblies for these accessions can be found online at http://ftp.tuebingen.mpg.de/ebio/alkeller/pan_NLRome/. SMRT RenSeq sequence reads for Can-0 were submitted under ENA project PRJEB26457. To discover putative WRR alleles from each accession, the genomic sequence of each respective WRR gene was queried using the blastn command (13) (megablast) against the assembly of SMRT RenSeq reads. The highest ranked hits from these searches were compiled. Hits of less than 95% identity and much shorter than the query sequence were considered non-orthologous loci and are marked in grey in Dataset S2. To obtain predicted protein sequences the assemblies were submitted to the Augustus gene prediction server (14) using the Arabidopsis thaliana species parameters. Blastp was used to query the predicted amino acid sequence encoded by each WRR gene against a combined database of all the predicted proteins in the MAGIC founders and Ws-2.

Phylogeny construction
In order to construct the phylogeny of Arabidopsis thaliana proteins containing TIR and NB domains, TIR-domain containing proteins were identified from the Araport11 Col-0 annotation using NLR-parser (15). Additionally, APAF-1 (Homo sapiens) and L6 (Linum usitatissimum) were chosen as outgroups. Where NB domains were detected, these were extracted and a core region of ~200 amino acids was aligned using MUSCLE (16)