Roles of microbiota in autoimmunity in Arabidopsis leaves

Abstract

Over the past three decades, researchers have isolated plant mutants that show constitutively activated defence responses in the absence of pathogen infection. These mutants are called autoimmune mutants and are typically dwarf and/or bearing chlorotic/necrotic lesions. Here, from a genetic screen for Arabidopsis genes involved in maintaining a normal leaf microbiota, we identified TIP GROWTH DEFECTIVE 1 (TIP1), which encodes an S-acyltransferase, as a key player in guarding leaves against abnormal microbiota level and composition under high-humidity conditions. The tip1 mutant has several characteristic phenotypes of classical autoimmune mutants, including a dwarf stature, showing lesions, and having a high basal level of defence gene expression. Gnotobiotic experiments revealed that the autoimmune phenotypes of the tip1 mutant are largely dependent on the presence of microbiota as axenic tip1 plants have markedly reduced autoimmune phenotypes. We found that the microbiota dependency of autoimmune phenotypes is shared by several 'lesion mimic'-type autoimmune mutants in Arabidopsis. It is worth noting that autoimmune phenotypes caused by mutations in two Nucleotide-Binding, Leucine-Rich Repeat (NLR) genes do not require the presence of microbiota and can even be partially alleviated by microbiota. Our results therefore suggest the existence of at least two classes of autoimmunity (microbiota-dependent versus microbiota-independent) in plants. The observed interplay between autoimmunity and microbiota in the lesion mimic class of autoimmunity is reminiscent of the interactions between autoimmunity and dysbiosis in the animal kingdom. These parallels highlight the intricate relationship between host immunity and microbial communities across various biological systems.

Fig. 1. The appearance and leaf microbiota phenotypes of the grm1 mutant.

a, Top panel: 4-week-old soil-grown Col-0, bbc and grm1 plants under ambient humidity (~50% RH) for 5 days (basal condition control). Bottom panel: 4-week-old soil-grown plants shifted to high humidity (~95% RH) for 5 days. Images were taken on day 5 of the humidity treatments. Scale bar, 2 cm. b, Population sizes of endophytic leaf microbiota after 5 days of indicated humidity conditions. Ambient humidity (~50% RH; basal condition control) and high humidity (~95% RH). Each column represents bacterial titres as log-transformed colony-forming units per gram of fresh weight (c.f.u. gFW⁻¹). Data are displayed as mean ± s.e.m. (n = 4 biological replicates; each biological replicate contains 1–3 leaves from one plant). Different letters represent significant differences (P < 0.05, two-way analysis of variance (ANOVA) with Tukey’s honestly significant difference (HSD) test). Exact P values for all comparisons are shown in the Source data. Experiment was independently performed twice with similar results. c, Relative abundance of endophytic leaf bacteria at the phylum level and at the class level for Proteobacteria. d, Shannon indexes of endophytic leaf bacteria based on 16S rRNA gene amplicon sequence profiling of indicated genotypes. n = 11 (Col-0), n = 15 (bbc) and n = 20 (grm1) biological replicates for analysis of leaf endophytic bacterial microbiota. The centre lines of the box plot represent means, the box edges are the 75th and 25th percentiles, whiskers extend to 10–90 percentiles, and dots are outliers; statistical analysis was performed using one-way ANOVA with Tukey’s HSD test. Source data

Fig. 2. Characterization of tip1 single mutant plants.

a, A schematic diagram showing various mutant alleles in the TIP1 gene. tip1^W171STOP is the allele isolated from this study which contains a G-to-A mutation at the splicing junction that is expected to cause a premature STOP codon at amino acid residue Trp¹⁷¹ in the ankyrin-repeat domain. SALK_020996 and SALK_052842 are T-DNA insertion alleles obtained from the Arabidopsis Biological Resource Center. b, Images of 4-week-old, soil-grown Col-0, bbc, grm1 and tip1 single mutant plants. Scale bar, 2 cm. c, Population sizes of endophytic leaf microbiota after 5 days under humidity conditions indicated. Ambient humidity (~50% RH; basal condition control) and high humidity (~95% RH). Results represent the mean values ± s.e.m. (n = 4 biological replicates; each biological replicate contains 1–3 leaves from one plant). Different letters represent a significant difference (P < 0.05, two-way ANOVA with Tukey’s HSD test). Exact P values for all comparisons are shown in the Source data. Experiment was independently performed twice with similar results. d,e, Normalized expression levels of immune-responsive PR1 (d) and FRK1 (e) genes in 4-week-old, soil-grown Col-0, bbc, grm1 and tip1 plants under ambient humidity (~50% RH). PP2AA3 expression was used for normalization. PR1 and FRK1 normalized gene expression levels (2^−ΔCt) were calculated by taking the difference between the cycle threshold (Ct) value of the PR1 or FRK1 gene and the Ct value of the internal control gene PP2AA3. This difference is used in the 2^−ΔCt formula to determine the relative expression levels of PR1 and FRK1, normalized to the internal control. Results represent the mean values ± s.e.m. of four biological replicates. Each biological replicate is a pool of three plants. Different letters represent a significant difference (P < 0.05, one-way ANOVA with Tukey’s HSD test). Exact P values for all comparisons are shown in the Source data. Experiment was independently performed twice with similar results. Source data

Fig. 3. The autoimmune phenotypes of tip1 and snc1 mutants.

a,b, Normalized expression levels of PR1 (a) and FRK1 (b) genes in 4-week-old, soil-grown Col-0, tip1 and snc1 plants under ambient humidity (~50% RH). PP2AA3 was used for normalization. Results represent the mean values ± s.e.m. of three biological replicates. Each biological replicate is a pool of three plants. Statistical analysis was performed using one-way ANOVA with Tukey’s HSD test. Experiment was independently performed twice with similar results. c,d, Total bacterial populations in Col-0, tip1 and snc1 leaves 3 days after Pst DC3000 (c) or 5 days after Pst ΔhrcC (d) infiltration. Humidity was kept at ~95% throughout the duration of the disease assay. Each column represents bacterial titres as log-transformed colony-forming units per cm² (c.f.u. cm⁻²) and is the mean of six biological replicates; each biological replicate contains leaf discs from infiltrated leaves from one plant; total of six plants were infiltrated. Error bars indicate s.e.m. Statistical analysis was performed using one-way ANOVA with Tukey’s HSD test. Experiment was independently performed twice with similar results. Source data

Fig. 4. The distinct autoimmune phenotypes of tip1 and snc1 mutants.

a, Top panel: 4-week-old, soil-grown Col-0, tip1 and snc1 plants under ambient humidity (~50% RH) for 5 days (basal condition control). Bottom panel: 4-week-old, soil-grown plants shifted to high humidity (~95% RH) for 5 days. Images were taken on day 5 of the treatments. Scale bar, 2 cm. b, Population sizes of endophytic leaf microbiota after 5 days of plant growth under humidity conditions indicated. Ambient humidity (~50% RH; basal condition control) and high humidity (~95% RH). Results represent the mean values ± s.e.m. of four biological replicates; each biological replicate contains one to two leaves from one plant. Different letters represent a significant difference (P < 0.05, two-way ANOVA with Tukey’s HSD test). Exact P values for all comparisons are shown in the Source data. Experiment was independently performed twice with similar results. c,d, Shannon indexes (c) and relative abundance (d) of endophytic bacterial microbiota at the phylum level and at class level for Proteobacteria of in Col-0, tip1 and snc1 leaves based on 16S rRNA gene amplicon sequence profiling. n = 11 (Col-0), n = 20 (tip1) and n = 17 (snc1) biological replicates for analysis of leaf endophytic bacterial microbiota. The centre lines of the box plot represent means, the box edges are the 75th and 25th percentiles, whiskers extend to 10–90 percentiles, and dots are outliers; one-way ANOVA with Tukey’s HSD test. e,f, Normalized expression levels of PR1 (e) and FRK1 (f) genes in 2.5-week-old plate-grown Col-0, tip1 and snc1 plants. PP2AA3 expression was used for normalization. Results represent the mean values ± s.e.m. of four biological replicates. Each biological replicate is a pool of two seedlings. Statistical analysis by one-way ANOVA with Tukey’s HSD test. Experiment was independently performed twice with similar results. Source data

Fig. 5. The appearance and leaf microbiota phenotypes of Arabidopsis autoimmune mutants.

a, Images of 4-week-old, soil-grown Arabidopsis autoimmune mutants exposed to high humidity (~95% RH) for 5 days. Top panel: Col-0, tip1 and three previously identified ‘lesion-mimic’ autoimmune mutants; bottom panel: Col-0, snc1 and three previously identified autoimmune mutants that showed no visible lesions. Scale bar, 2 cm. b, Population sizes of endophytic leaf microbiota after 5 days of plant growth under high-humidity condition (~95% RH) in tip1 and three previously identified ‘lesion-mimic’ autoimmune mutants. c, Population sizes of endophytic leaf microbiota after 5 days of plant growth under high humidity in snc1 and three previously identified autoimmune mutants with no visible lesions. Results represent the mean values ± s.e.m. of four biological replicates; each biological replicate contains one to three leaves from one plant. Different letters represent a significant difference (P < 0.05, one-way ANOVA with Tukey’s HSD test). Exact P values for all comparisons are shown in the Source data. Experiment was independently performed twice with similar results. Source data

Fig. 6. Microbiota dependency for autoimmunity in Arabidopsis autoimmune mutants.

a, Five-week-old Col-0, tip1 and three previously identified lesion-mimic autoimmune mutants grown in GnotoPots under holoxenic (top panel) or axenic (bottom panel) conditions. Scale bar, 2 cm. Zoomed-in images (white squares) on leaf lesions are shown in Supplementary Figure 2a. b, Five-week-old Col-0, snc1 and three previously identified autoimmune mutants that showed no visible lesions were grown in GnotoPots under holoxenic (top panel) or axenic (bottom panel) conditions. Scale bar, 2 cm. c,d, PR1 expression in tip1 and three previously identified lesion-mimic autoimmune mutants (c) and snc1 and three autoimmune mutants (d) grown in GnotoPots for 5 weeks under axenic (left; with diagonal stripe pattern) or holoxenic (right) conditions. Results represent the mean values ± s.e.m. of four biological replicates. Each biological replicate is a pool of two plants. Different letters represent a significant difference. Statistical analysis was performed using one-way ANOVA with Fisher’s least significant difference (LSD) test. Exact P values for all comparisons are shown in the Source data. Experiment was independently performed twice with similar results. Source data

Fig. 7. Microbiota dependency for autoimmunity in A. thaliana natural accessions.

a, Images of 5-week-old, potting soil-grown A. thaliana accessions (Col-0, Est-1 and C24) exposed to high humidity (~95% RH) for 7 days. Scale bar, 2 cm. b, Population sizes of leaf endophytic microbiota after 7 days of plant growth under high-humidity condition (~95% RH). Results represent the mean values ± s.e.m. of six plants. Statistical analysis was performed using one-way ANOVA with Tukey’s HSD test. Experiment was independently performed twice with similar results. c, Five-week-old Arabidopsis accessions grown using GnotoPots under holoxenic (top panel) or axenic (bottom panel) conditions. Scale bar, 2 cm. Zoomed-in image (white square) on Est-1 leaf lesions is shown in Supplementary Figure 2b. d, PR1 expression in Arabidopsis accessions grown in GnotoPots under axenic (left; with diagonal stripe pattern) or holoxenic (right) conditions. Results represent the mean values ± s.e.m. of three biological replicates. Each biological replicate is a pool of two plants. Statistical analysis was performed using one-way ANOVA with Fisher’s LSD test. Experiment was independently performed twice with similar results. Source data

Extended Data Fig. 1. Mutants isolated from the bbc genetic screen.

Plant images of 4-week-old soil grown grm mutants; scale bar indicates 2 cm. All grm mutant phenotypes were inherited as recessive traits with exception of grm2. grm2 phenotypes were inherited as semi-dominant trait; based on the 1:2:1 segregation ratio, the grm2 mutant image panel contains bbc-like (left), heterozygous (middle), and homozygous (right) plants.

Extended Data Fig. 2. Identifying and confirming the causative mutation in the grm1 mutant.

a, grm1 genomic mapping. Red line represents allele frequency. Blue and purple lines denote 95% and 99% confidence intervals, respectively. b, RT-PCR products using primers flanking the grm1 mutation locus. Genomic (top panel) and complementary (bottom panel) DNA from indicated genotypes were used as templates. The retained intron is expected to lead to a premature STOP codon. c, Images of four-week-old, potting soil-grown Col-0, bbc, grm1 and two independent grm1 complementation lines under ambient humidity ( ~ 50% RH basal control condition; upper panel) or high humidity ( ~ 95% RH; bottom panel) for five days. Scale bar equals 2 cm. d, Transgene complementation of the grm1 mutant. Population sizes of leaf endophytic microbiota after five days of plant growth under ambient humidity ( ~ 50% RH, basal control condition) or high humidity ( ~ 95% RH). Results represent the mean values ± SEM of four plants. Statistical analysis was performed using two-way ANOVA with Tukey’s HSD test. Exact p-values for all comparisons are shown in the Source Data. Experiment was independently performed twice with similar results. Source data Source data

Extended Data Fig. 3. Appearance of tip1 mutant alleles.

a, A schematic diagram illustrating various mutant alleles of TIP1 protein. tip1^W171STOP is the allele isolated from this study which contains a G to A mutation at the splicing junction that is expected to cause a pre-mature STOP codon at amino acid residue Trp¹⁷¹ in the ankyrin-repeat domain. SALK_020996 and SALK_052842 have T-DNA insertions that would affect the catalytic DHHC cysteine-rich domain (DHHC-CRD). b, Images of four-week-old, potting soil-grown Col-0, bbc, grm1 and various tip1 single mutant plants under ambient humidity ( ~ 50% RH, basal control condition; upper panel) or high humidity ( ~ 95% RH) for five days. Scale bar equals 2 cm.

Extended Data Fig. 4. TIP1 and SNC1 gene expression is responsive to PTI elicitor flg22.

Expression of TIP1 (a), SNC1 (b), and FRK1 (c) in wild-type Col-0 plants after flg22 treatment. PP2AA3 expression was used for normalization. Plants were grown under axenic (Ax) or holoxenic (Holo) conditions in GnotoPots. Open circles (∘) indicate the basal, uninduced state whereas closed circles (•) indicate the expression 90 min after 250 nM flg22 induction. Results represent the mean values ± SEM of four biological replicates. Statistical analysis was done by the two-way ANOVA with Fisher’s LSD test. Experiment was independently performed twice times with similar results. Source data

Extended Data Fig. 5. ROS burst dynamics induced by PTI elicitors.

ROS burst dynamics induced by 100 nM of flg22 (a), elf18 (b) and AtPep1 (c) in 4-week-old soil grown plants. Results represent the mean values ± SEM (n = 8 plants). Bar plots show cumulative ROS production over 60 min (right). Results represent the mean values ± SEM of eight biological replicates. Statistical analysis was done by one-way ANOVA with Tukey’s HSD test. Experiment was independently performed twice times with similar results. Source data

Extended Data Fig. 6. Effects of different soil types on tip1 and snc1 plants.

a, soil appearance of Arabidopsis mix, North Carolina (NC) soil, Michigan State University (MSU) soil, and Michigan Benton Harbor (MI-BH) soil. b, Images of four-week-old plants growing on pots prepared using indicated soil types (left) and plant images after five days high humidity condition ( ~ 95% RH; right). c, Population sizes of endophytic leaf microbiota after five days high humidity condition ( ~ 95% RH). Results represent the mean values ± SEM (n = 4 biological replicates; each biological replicate contains leaves from one plant). Different letters represent a significant difference (p < 0.05, two-way ANOVA with Tukey’s HSD test). Exact p-values for all comparisons are shown in the Source Data. Experiment was independently performed twice with similar results. Source data

Extended Data Fig. 7. Effects of bacterial bulk culture (BC) collected from dysbiotic tip1 plants on wild type plants.

a, Col-0 leaf images on Day 0 and Day 6 after infiltrated with BC^Col-0 or BC^tip1. b, Population sizes of endophytic leaf microbiota immediate (Day 0; left) and six days (Day 6; right) after BC infiltrations. Results represent the mean values ± SEM (n = 4 biological replicates; each biological replicate contains 1-2 leaves from one plant). Statistical analysis was done by two-way ANOVA with Fisher’s LSD test. Experiment was independently performed twice with similar results. Source data

Extended Data Fig. 8. Endophytic bacterial microbiota profiling in tip1 and tip1-like autoimmune mutants.

a, PCoA of weighted UniFrac distances obtained from 16 S rRNA gene sequence profiles of endophytic bacterial microbiota in Col-0, tip1 and tip1-like autoimmune mutants under high humidity for 5 days. Pairwise PERMANOVA results are provided in the Source Data. b, The number of ASVs indicating endophytic bacterial microbiota alpha diversity in the indicated genotypes. n = 17 (Col-0) and n = 16 (for tip1 and tip1-like mutant plants). The center lines of the box plot represent means, the box edges are the 75^th and 25^th percentiles, whiskers extend to 10-90 percentiles, and dots are outliers. Statistical analysis was done by one-way ANOVA with Dunnett’s test. c, Relative abundance of bacterial populations at the phylum level. Members of Proteobacteria phylum are further separated into class. Source data

Extended Data Fig. 9. Transcriptomic analysis of Col-0, tip1 and snc1 plants grown in GnotoPots.

a, PCA analysis of genes expressed in Col-0, tip1 and snc1 plants grown in GnotoPots under axenic (Ax), heat-killed “MSU” microbial community (HK) or live “MSU” microbial community (holoxenic; Holo) conditions. Zoomed-in image contains PCA analysis of Col-0 under all three conditions and tip1 grown in GnotoPots under Ax and HK conditions. b, Heat map of DEGs significantly changed between Col-0 and tip1 and/or between Col-0 and snc1 plants grown in GnotoPots under “MSU” microbial community holoxenic condition. Clusters 1 and 2 are enriched for defence-related Gene Ontology (GO) terms, whereas cluster 3 is enriched for growth and photosynthesis-related GO terms. See Supplementary Data 4 for the complete list of genes and their normalized counts. c, Selected defence-related gene expression in Col-0, tip1 and snc1 plants grown in GnotoPots under “MSU” microbial community holoxenic condition. See Supplementary Table 5 for the gene IDs.