Roles of microbiota in autoimmunity in Arabidopsis

Over the past three decades, researchers have isolated plant mutants that display constitutively activated defense responses in the absence of pathogen infection. These mutants are called autoimmune mutants and are typically dwarf and/or bearing chlorotic/necrotic lesions. From a genetic screen for Arabidopsis genes involved in maintaining a normal leaf microbiota, we identified TIP GROWTH DEFECTIVE 1 (TIP1), which encodes a 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, displaying lesions, and having a high basal level of defense 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. Interestingly, autoimmune phenotypes caused by mutations in NLR genes do not require the presence of microbiota and can even be partially alleviated by microbiota. Our results therefore suggest the existence of two classes of autoimmunity (microbiota-dependent vs. 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.

4 was more than three orders of magnitude higher (Fig. 1b). In addition to a drastic increase in leaf 118 endophytic microbiota, the relative abundance of leaf microbiota members in the grm1 mutant also 119 shifted overwhelmingly to Proteobacteria (Fig. 1c, ~97% in grm1 leaves compared to ~45-55% in 120 Col-0 and bbc leaves). In the grm1 mutant, amplicon sequence variants (ASVs) in Proteobacteria 121 belong predominantly to the genus Pseudomonas of the class Gammaproteobacteria, while 122 Bacillus and Paenibacillus belonging to Firmicutes became nearly undetectable ( Fig. 1c and   123 Extended Data Table 1). Shannon index that measures richness and evenness of a microbial 124 community composition also decreased in the grm1 mutant, reflecting the overwhelming presence 125 of Proteobacteria (Fig. 1d).

126
Identification of the causal mutation in the grm1 mutant.

127
To identify the causative mutation in the grm1 mutant, grm1 plants were backcrossed with its 128 progenitor, the bbc triple mutant, to generate a segregated F2 population. Analyzing the mapping-129 by-sequencing data from grm1 co-segregates revealed that the mutation was located on

142
To determine if the dysbiotic phenotypes of the grm1 mutant (i.e., the bbc tip1 quadruple mutant) 143 are dependent on the background bbc mutations, we segregated the tip1 mutation from the bbc 144 triple mutations by outcrossing the grm1 mutant with wild-type Col-0 plants and genotyping the 145 resulting F2 population (see Extended Data Table 5 for primes used to genotyping). From The

146
Arabidopsis Biological Resource Center (ABRC) 26 , we also obtained two independent tip1 single 147 mutant alleles (SALK_020996 and SALK_052842) carrying T-DNA insertions in the TIP1 gene 148 (Fig. 2a). All three tip1 single mutants were larger than the original grm1 (bbc tip1) mutant, but 149 still smaller than wild type Col-0 plants (Fig. 2b). Interestingly, the humidity-dependent dysbiosis  The tip1 mutant has features of autoimmune mutants.

156
We noticed that the morphological phenotypes of grm1 and tip1 (i.e., small statures and chlorotic 157 leaves) were reminiscent of typical autoimmune mutants, which have been isolated in the past few 158 decades. One hallmark of autoimmunity is constitutive high basal expression of immune-related 159 marker genes in the absence of pathogen attacks. We therefore analyzed two immune-related was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023. ; https://doi.org/10.1101/2023.03.06.531303 doi: bioRxiv preprint 1 (FRK1). Indeed, both grm1 and tip1 plants have heightened PR1 and FRK1 expression under 162 basal condition (i.e., in the absence of pathogen inoculation) ( Fig. 2d and 2e). TIP1 gene expression 163 itself was found to be induced by flg22, a flagellin-derived peptide that induces many PTI-164 associated genes 27,28 (Extended Data Fig. 5).

166
We were intrigued by the morphological phenotypes and heightened immune-related marker gene 167 expression of the tip1 mutant as they point to a connection between dysbiosis and autoimmunity.  Table 3). In contrast, snc1 plants had a diverse leaf endophytic microbiota 192 composition, similar to that of Col-0 plants ( Fig. 4c and 4d). Finally, when grown in aseptic agar 193 plates, snc1 plants continued to exhibit heightened immune-related marker gene expression in the 194 absence of microbes, whereas PR1 and FRK1 expression in tip1 mutant plants greatly subsided to 195 close to the low levels observed in wild type Col-0 plants ( Fig. 4e and 4f).

196
A broad role of microbiota in "lesion-mimic" autoimmunity in Arabidopsis.

197
The interesting contrast in microbiota-dependency for autoimmune phenotypes between the tip1 was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023. ; https://doi.org/10.1101/2023.03.06.531303 doi: bioRxiv preprint zinc-finger protein in Arabidopsis 37 . The snc1-like autoimmune mutant category consists of (i) the 206 chs3 mutant, which carries a gain-of-function mutation in a TIR-type NLR immune receptor 38 , (ii) 207 dnd1 39 and (iii) dnd2 40 mutants, which carry mutations in two cyclic nucleotide-gated cation 208 channels. Mutant plants in the tip1-like category showed severe leaf lesions ( Fig. 5a; top panel)

209
and harboured high levels of endophytic leaf microbiota under high humidity (Fig. 5b), whereas 210 mutant plants in the snc1-like category had no visible lesions ( Fig. 5a; bottom panel) and carried 211 low levels of endophytic leaf microbiota similar to wild-type Col-0 plants (Fig. 5c).

212
To further characterize a possible microbiota dependency of autoimmune phenotypes in these two 213 categories of mutants, we grew mutant plants in the absence (axenic) or presence (holoxenic) of a 214 natural soil-derived microbiota using GnotoPots, a peat-based gnotobiotic system as recently   Est-1 and between snc1 and C24 prompted us to investigate if microbiota is required for the 249 autoimmune phenotypes in Est-1 and C24. As shown in Figure 7c, under holoxenic condition, Est-250 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023. ; https://doi.org/10.1101/2023.03.06.531303 doi: bioRxiv preprint 7 1 shows lesions on leaves, albeit to a lesser extent compared to Est-1 grown under the conventional 251 potting soil growth condition (Extended Data Fig. 6 and Extended Data Fig. 7b). Interestingly, like 252 tip1, Est-1 plants did not show leaf lesions under the axenic condition. Furthermore, like the tip1 253 mutant, the heightened PR1 expression in Est-1 subsided to a low level when grown in the axenic 254 condition ( Fig. 7d; left panel). Conversely, like the snc1 mutant, C24 plants had elevated PR1 255 expression regardless of growth in the presence or absence of a microbial community (Fig. 7d).

256
Another similarity between the snc1 mutant and C24, which is in contrast to the tip1 mutant, is the 257 alleviation of their stunted growth morphology in the presence of microbiota ( Fig. 6b and 7c).

269
Among putative mutants isolated was grm1, which we characterized in detail.

270
The grm1 mutant contains a missense mutation in the TIP1 gene that encodes a S-acyltransferase.

271
The first mutant allele of TIP1 was isolated in a genetic screen for mutants that had defects in root

280
Prior to this study, however, the connection between TIP1 and leaf microbiota homeostasis was 281 not known. A comprehensive Arabidopsis acylome using multiple tissue types identified close to to identify specific TIP1 substrate(s) that is required for microbiota homeostasis in Arabidopsis 289 leaves.

290
A key finding in this study is that not only the tip mutants are unable to control the proliferation 291 or maintain a normal composition of a leaf microbiota, but they also display dysbiosis-associated 292 tissue damages and autoimmunity in the presence of microbiota (Fig. 4). The microbiota-293 dependent autoimmune phenotypes of the tip1 mutant led us to broadly examine a potential 294 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023. ; https://doi.org/10.1101/2023.03.06.531303 doi: bioRxiv preprint connection between microbiota and previously reported "autoimmune" mutants in Arabidopsis.

295
Based on how they respond to the existence of microbiota, it appears that autoimmune mutants in 296 Arabidopsis can be divided into at least two classes (Extended Data Fig. 8). One class, exemplified 297 by the tip1 mutant, exhibits microbiome-dependent autoimmunity. The autoimmune phenotypes 298 in this class largely disappeared when grown in the axenic conditions. Given that these mutants 299 also have an increased microbiota load, this result suggests that the autoimmune phenotypes in this 300 class of mutants are a consequence of harbouring an overabundant microbial community. The 301 other class of autoimmunity is independent of microbiota and is represented by the snc1 mutant.

302
The autoimmune phenotypes of this class do not require the presence of microbiota. i.e., they have 303 small statures and high PR1 expression regardless of presence or absence of microbial 304 communities. In fact, the presence of microbiota alleviates the stunted growth morphology of snc1 305 and chs3 (Fig. 6b), which is in striking contrast to those of the tip1 class (Fig. 6a).

306
We find it interesting that microbiota-dependent and -independent autoimmunity can also be C24 and Est-1 may be able to survive and reproduce to avoid extinction of the entire population.

312
The microbiota-dependent and -independent expression of autoimmunity in Est-1 vs. C24, as 313 observed in our study (Fig. 7), may reflect different paths by which the two types of autoimmunity 314 have convergently evolved in natural populations under different abiotic and biotic pressures.

315
Future research may uncover other natural accessions that show a continuum range of autoimmune 316 phenotypes in terms of microbiota-dependency, as observed for the dnd1 and dnd2 mutants, which 317 exhibit microbiota-amplification of heighted basal defense gene expression (Fig. 6d).

318
Overall, results from this study begin to illustrate conceptual parallels in microbiota interactions  was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made     Fig. 2a); the G to A mutation in the TIP1 (At5g20350) gene is tightly associated 371 with the grm1 phenotype (Extended Data Fig. 2 and Extended Data Table 2

417
16S rDNA amplicon data processing 418 Raw Illumina data for 16S rDNA amplicon were processing as described previously 20 using 419 QIIME2 version 2022.2 59 . In brief, primer sequences were removed using Cutadapt 60 followed by 420 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made     QuantStudio 3 real-tame PCR system (Applied Biosystems). PP2AA3 was used for normalization.

447
The primer sets used to quantify gene expression in this study are listed in Extended Data Table 5.

496
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023.          was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023.     was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023.

782
a, Images of four-week-old, soil-grown Arabidopsis autoimmune mutants exposed to high 783 humidity (~95% RH) for five days. Top panel, Col-0, tip1 and three previously identified "lesion-    was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted July 4, 2023. ; https://doi.org/10.1101/2023.03.06.531303 doi: bioRxiv preprint