Solanum americanum genome-assisted discovery of immune receptors that detect potato late blight pathogen effectors

Abstract

Potato (Solanum tuberosum) and tomato (Solanum lycopersicon) crops suffer severe losses to late blight caused by the oomycete pathogen Phytophthora infestans. Solanum americanum, a relative of potato and tomato, is globally distributed and most accessions are highly blight resistant. We generated high-quality reference genomes of four S. americanum accessions, resequenced 52 accessions, and defined a pan-NLRome of S. americanum immune receptor genes. We further screened for variation in recognition of 315P. infestans RXLR effectors in 52 S. americanum accessions. Using these genomic and phenotypic data, we cloned three NLR-encoding genes, Rpi-amr4, R02860 and R04373, that recognize cognate P. infestans RXLR effectors PITG_22825 (AVRamr4), PITG_02860 and PITG_04373. These genomic resources and methodologies will support efforts to engineer potatoes with durable late blight resistance and can be applied to diseases of other crops.

Fig. 1. Genome evolution of Solanum americanum.

a, Phylogenetic relationship of S. americanum and neighboring species. The red number indicates the bootstrap of each node. The black number denotes the estimated divergence time (million years ago). The scale bar represents the number of amino acid substitutions per site. b, Genome synteny of S. americanum, potato and eggplant. Ribbons between chromosomes show syntenic regions. Large chromosome rearrangements (>1 Mb in size) are marked in orange. Source data

Fig. 2. Pan-NLRome of S. americanum.

a, The NB-ARC domains of S. americanum SP1102 were predicted by NLR-annotator and used to generate a maximum-likelihood tree using IQ-TREE with the JTT + F + R9 model. Known NLR proteins from Solanaceae species were included (highlighted in red). The NLRs are classified into different groups based on a previous report. The RNL, TNL and NRC superclade are shown. CED-4 from Caenorhabditis elegans was used as the outgroup. The expression profile is shown by a heatmap (white to red) based on the cDNA RenSeq data of SP1102. The P/A polymorphism of NLRs from the three other S. americanum genomes and SMRT RenSeq assemblies of 16 additional accessions are shown by the heatmap (white to blue). The accession order from top to bottom is SP3409, SP3408, SP3406, SP3400, SP3399, SP3370, SP2360, SP2308, SP2307, SP2300, SP2298, SP2272, SP1123, SP1101, SP1034, SP1032, SP2275, SP2273 and SP2271. The absent NLRs are shown by black blocks. The scale bar represents the number of amino acid substitutions per site. b, The physical map of NLR genes in the SP1102 genome. CNLs are shown by yellow blocks, TNLs are shown by red blocks and RNLs are shown by pink blocks. Some functionally characterized NLR clusters are noted on this map. Some previously reported NLR clusters (NRC1, NRG1, R3, ADR1, NRC3, R1, RB, Rpi-vnt1, NRC2, Rpi-chc1, Rpi-amr1, NRC6, NRC4b, NRC5, NRC4a, Rpi-amr3, Gpa2) are shown in the physical map. (c) The proportion of NLR singletons and NLRs in clusters. d, Number of manually curated NLR genes (red circle), and the number of NLR genes predicted by NLR-annotator (yellow circle). Manual curation of NLR genes from SP2275 was not performed. e, Phylogeny of the NRC helper NLR family. The NRC homologs from potato, tomato and N. benthamiana are marked in orange, red and green, respectively. The NRC proteins from S. americanum are in black. The number indicates the bootstrap of each node. The scale bar represents the number of amino acid substitutions per site. f, The log₁₀ transformed TPM values for NLR genes are classified into five groups, and some homologs of known R genes are noted. The NLR IDs are shown in Supplementary Table 2.

Fig. 3. ETI landscape of S. americanum and P. infestans.

A total of 315 RXLR effectors were transiently expressed in 52 S. americanum accessions. The HR index (2, strong HR; 1, weak HR; 0, no HR, NA, not available) was used for the heatmap. These effectors were screened on N. benthamiana, and their recognitions are included in this heatmap. Empty PVX vector was used as negative control, and co-expression of Rpi-amr3–HisFlag and AVRamr3–GFP was used as positive control. The S. americanum accessions were ordered based on the phylogenetic tree; SP3400 is not included in this tree. The scale bar represents the number of amino acid substitutions per site. The S. americanum accessions were classified into four groups (gray or yellow shading). The four reference accessions are marked by red arrows. The effectors were ordered based on the total HR index. For each effector, the numbers of responsive S. americanum accessions is visualized by a bar chart on the top of the heatmap. For each S. americanum accession, the numbers of recognized effectors is visualized by a bar chart on the right of the heatmap. Some RXLR effectors previously characterized or mentioned in this study are indicated by gray arrows.

Fig. 4. Identification and characterization of Rpi-amr4 that recognizes PITG_22825.

a, PITG_22825 is an RXLR effector. 35S::PITG_22825 triggers cell death on S. americanum SP2271 and SP1102 leaves, but not SP2298 leaves. b, Manhattan plot of the GWAS of PITG_22825 recognition. The SNPs associated with PITG_22825 recognition is located in an NLR singleton sp1102chr01_nlr39; (red arrow). c, HR assay of candidate genes. Rpi-amr4-1102 and Rpi-amr4-2271 were expressed alone or co-expressed with either 35S::PITG_22825 or 35S::AVRamr3–GFP constructs in N. benthamiana leaves. Rpi-amr4-2271 is auto-active in N. benthamiana, but when co-expressed with PITG_22825, the HR was stronger. Rpi-amr4-1102 specifically recognizes PITG_22825. Rpi-amr3 was used as control. OD₆₀₀ = 0.5. Four leaves from two plants were used for each experiment and three biological replicates were performed with the same results. HF, HisFlag. d, Phylogeny of Rpi-amr4 homologs in different S. americanum accessions. Rpi-amr3 was used as an outgroup. PITG_22825-mediated HR is shown by red (HR) or blue (no HR) circles. Percent identity of the amino acid sequence relative to Rpi-amr4-1102 is shown. a,c,d–f, The scale bars represent the number of amino acid substitutions per site. e, Rpi-amr4-knockout lines lose the capacity for PITG_22825 recognition. Two sgRNAs (black arrows) were designed on Rpi-amr4-2271. The genotype of the two knockout lines is shown. Both lines failed to recognize PITG_22825, but HR could be complemented when co-expressing PITG_22825 with Rpi-amr4-1102. Wild-type (WT) SP2271 plants were used as control plants. Rpi-amr3 and AVRamr3 were used as positive control. OD₆₀₀ = 0.5. f, DLA with 35S::Rpi-amr4-1102. 35S::Rpi-amr4-1102 (green), Rpi-amr3 (positive control, red) and Rpi-amr3a (a non-functional Rpi-amr3 paralog, negative control, blue) were transiently expressed in N. benthamiana, OD₆₀₀ = 0.5. Zoospores from P. infestans strain T30-4 were used to inoculate the leaves 1 day post-infiltration (dpi). Lesion sizes were measured at 6 dpi. Four biological replicates were performed, and all data points (74 data points per treatment) were visualized as a box-and-whisker plot. Center line, median; box limits, upper and lower quartiles. The whiskers (top and bottom) comprise values within 1.5 times the interquartile range (IQR). The outliers are indicated by black dots. Statistical differences were analyzed by one-way ANOVA with Tukey’s HSD test (P < 0.001) and were indicated by the lower-case letters. Representative leaves are shown. Source data

Fig. 5. Identification of R02860 and R04373 that recognize the RXLR effectors PITG_02860 and PITG_04373.

a, PITG_02860 is an RXLR effector from P. infestans. An illustration and predicted structure are shown. b, An F₂ population was generated from a cross between SP2271 (responds to PITG_02860, R) and SP2272 (no response to PITG_02860, NR). The R bulk (104 progenies) and NR bulk (34 progenies) were used for BSA-RenSeq. c, A total of 218 linked SNPs (red dots) on the top of chromosome 4 of SP2271 were identified. The gray bar represents the chromosome. The physical positions (in Mb) are shown by number. Five molecular markers (S30, S42, S35, S36 and S31) were used for the map-based cloning. The number of recombination events per total tested gametes is shown. d, HR assay of the candidate PITG_02860 receptor. The candidate genes were expressed alone, or co-expressed with PVX::PITG_02860 in N. tabacum leaves. Rpi-amr3 and AVRamr3 were used as controls. NLR16 turned out to be the PITG_02860 receptor (R02860 hereafter). OD₆₀₀ = 0.5. Four-week-old N. tabacum plants were used, and the photos were taken at 4 dpi. Three biological replicates were performed with the same results. e, PITG_04373 is an RXLR effector from P. infestans. An illustration and predicted structure are shown. f, Both backcross (BC₁) and F₂ populations were generated from SP2271 and SP2300. The BC₁ population of 192 responsive plants and 182 non-responsive progenies were bulked for BSA-RenSeq. The F₂ populations were used for fine mapping. g, Informative SNPs (red dots) on the top of chromosome 10 of SP2271 were identified. Five molecular markers (S11, S13, S5, S7 and S16) were used for fine mapping. The number of recombination gametes per total tested gametes is shown. Nine genes from SP2300 SMRT-RenSeq assemblies that mapped to the mapping interval of the SP2271 genome were regarded as candidate genes. All candidates belong to the Rpi-chc1 family except C444. h, HR assay of the PITG_04373 receptor candidates. The candidate genes were expressed alone or co-expressed with 35S::PITG_04373 in N. benthamiana leaves, Rpi-amr3 and AVRamr3 were used as controls. C168 turned out to be the PITG_04373 receptor (R04373 hereafter). OD₆₀₀ = 0.5. Four-week-old N. benthamiana plants were used and photos were taken at 4 dpi. Three biological replicates were performed with the same results.