Conservation of the PBL-RBOH immune module in land plants

The rapid production of reactive oxygen species (ROS) is a key signaling output in plant immunity. In the angiosperm model species Arabidopsis thaliana (hereafter Arabidopsis), recognition of non- or altered-self elicitor patterns by cell-surface immune receptors activates the receptor-like cytoplasmic kinases (RLCKs) of the AVRPPHB SUSCEPTIBLE 1 (PBS1)-like (PBL) family, particularly BOTRYTIS-INDUCED KINASE1 (BIK1).1,2,3 BIK1/PBLs in turn phosphorylate the nicotinamide adenine dinucleotide phosphate (NADPH) oxidase RESPIRATORY BURST OXIDASE HOMOLOG D (RBOHD) to induce apoplastic ROS production.4,5 PBL and RBOH functions in plant immunity have been extensively characterized in flowering plants. Much less is known about the conservation of pattern-triggered ROS signaling pathways in non-flowering plants. In this study, we show that in the liverwort Marchantia polymorpha (hereafter Marchantia), single members of the RBOH and PBL families, namely MpRBOH1 and MpPBLa, are required for chitin-induced ROS production. MpPBLa directly interacts with and phosphorylates MpRBOH1 at specific, conserved sites within its cytosolic N terminus, and this phosphorylation is essential for chitin-induced MpRBOH1-mediated ROS production. Collectively, our work reveals the functional conservation of the PBL-RBOH module that controls pattern-triggered ROS production in land plants.


RESULTS AND DISCUSSION
To examine the conservation of reactive oxygen species (ROS) production mechanisms during land plant immune signaling, we treated the wild-type Marchantia polymorpha (hereafter Marchantia) Tak-1 with the pathogen-associated molecular patterns (PAMPs) flg22 (the 22-amino-acid peptide epitope of bacterial flagellin) or chitin (a major fungal cell wall component). It was previously shown that flg22 treatment failed to inhibit Marchantia gemmaling growth 6 and, similarly, Marchantia thalli were insensitive to flg22 treatment in terms of ROS production ( Figure 1A), which is in line with the absence of an ortholog of the angiosperm flg22 receptor FLAGELLIN SENSING 2 (FLS2) in the Marchantia genome. 7 In contrast, chitin induced clear apoplastic ROS production in this assay ( Figure 1A), consistent with the presence of lysin-motif domain-containing receptors in the Marchantia genome. 6,7,8 In angiosperm species, PAMP-induced ROS production depends on members of the plant nicotinamide adenine dinucleotide phosphate (NADPH) oxidase RESPIRATORY BURST OXIDASE HOMOLOG (RBOH) family. 9- 14 Notably, pre-treatment with the NADPH oxidase inhibitor diphenyleneiodonium (DPI) significantly reduced chitin-induced ROS production ( Figure 1A), suggesting that the RBOH-mediated ROS production machinery is conserved in Marchantia.
The Marchantia genome encodes two RBOH family members, Mp3g20340 (MpRBOH1) and Mp7g00270 (MpRBOH2). Phylogenetic analysis with RBOHs from different land plant species could not clearly establish an orthologous relationship between RBOHs from angiosperm species and those from Marchantia ( Figure 1B). The expression of both MpRBOH genes was detected in multiple tissues and development stages ( Figure S1A), 15 and their mRNA levels were enhanced in response to chitin treatment, with MpRBOH1 displaying a stronger response-similar to that of the immune marker gene MpWRKY22 6 ( Figure S1B). Notably, an increase of MpRBOH1 transcript abundance was observed during infection with the oomycete pathogen Phytophthora palmivora ( Figure S1A). 15 To characterize the function of MpRBOHs, we generated single knockout mutants for both family members with CRISPR-Cas9-nickase. 16,17 Two different alleles of Mprboh1 and two independent lines representing one allele of Mprboh2 were isolated, and their genotypes were confirmed by sequencing ( Figures S1C and S1D). Although Mprboh2 plants responded normally to chitin treatment (Figure 1C), Mprboh1 alleles were completely incompetent of chitin-triggered ROS production ( Figure 1C). ROS production in Mprboh1 was restored by the expression of Cas9-nickase-insensitive wild-type MpRBOH1 (Figures S1E and S1F). Among the ten-member Arabidopsis thaliana (hereafter Arabidopsis) RBOH family, RBOHD is solely required for pattern-induced ROS production. 13,14 These observations suggest that in both Marchantia and Arabidopsis, a single RBOH protein is required for chitininduced ROS production.
RBOHs are involved in multiple plant processes. 18 In addition to the chitin-induced ROS phenotype, Mprboh1 mutants exhibited defects in rhizoid development ( Figures 1D and S1G). In contrast, Mprboh2 mutants exhibited strong overall defects in thallus growth, but were still able to produce rhizoids ( Figures 1D, S1G, and S1H). Notably, while Arabidopsis rbohd mutants do not show any obvious rosette leaves or root development phenotype, 19 another Arabidopsis RBOH family member, RBOHC, is a positive regulator of root hair growth. 20 Rhizoids are functionally reminiscent of root hairs, 21 and both elongate by tip growth, implying that MpRBOH1 might also be involved in the regulation of tip growth in Marchantia.
A single PBL isoform is required for chitin-triggered ROS production in Marchantia RBOH activation during immunity is controlled by several regulatory mechanisms. 22 In Arabidopsis, BOTRYTIS-INDUCED KI-NASE1 (BIK1)/PBL1-mediated phosphorylation is essential for RBOHD activation and function during pattern-triggered immunity. 4,5 We thus tested whether MpRBOH1 is similarly regulated by any of the Marchantia AVRPPHB SUSCEPTIBLE 1 (PBS1)-like (PBL) family proteins. Three PBL-encoding genes were identified in the Marchantia genome: Mp3g25360 (MpPBLa), Mp3g18020 (MpPBLb), and Mp2g14830 (MpPBLc). 7 Our phylogenetic analysis of PBL homologs among land plants identified three major clades dividing into eight subgroups (Figure 2A). Importantly, none of the three MpPBLs falls into the same subgroup as BIK1 or PBL1. In contrast to land plant RBOHs, MpPBLa and MpPBLb are part of clades that are present in all tested land plant species, wherein a single MpPBL gene is orthologous to its bryophyte and tracheophyte homologs. However, MpPBLc clusters together with a single PBL paralog from Anthoceros agrestis, Selaginella moellendorffii, and Arabidopsis, suggesting that those PBLs are more divergent to other land plant PBLs. Based on the phylogeny, we considered MpPBLa and MpPBLb to be the major PBL isoforms, and MpPBLc to be a divergent PBL isoform in Marchantia. We next interrogated publicly available expression data 15 and found that the two major isoforms share a broad expression pattern across tissues and development stages and in response to various treatments ( Figure S2A). The transcript abundance of MpPBLc is relatively low compared to the other two MpPBLs (except in the antheridium), suggesting a potential role for MpPBLc in sexual reproduction.   Figure 2. MpPBLa positively regulates chitin-triggered ROS production in a kinase activity-dependent manner (A) Phylogenetic analysis of PBL family proteins from land plant species. Full-length protein sequences were aligned by MUSCLE and the unrooted phylogenetic tree was constructed using the maximum likelihood method with a 1,000 bootstrap resampling value. Branch colors represent the sequence origins. Branches that did not pass bootstrap analysis are labeled with green dots. Three major clades are highlighted with colored boxes. Phylogenetic tree with all gene identifiers is provided in Figure S4. To study the function of MpPBLs in chitin-induced ROS production, we tried to generate knockout mutants for the two major isoforms of MpPBLs. Several alleles were obtained for MpPBLa (Figure S2B), but we failed to isolate knockout mutants for MpPBLb despite using multiple sets of guide RNAs. Besides plant immunity, PBLs regulate various developmental and reproductive processes in Arabidopsis. 23 It is thus possible that MpPBLb may have a critical role in the regulation of plant growth and development, and that its knockout mutants are therefore lethal. However, multiple, independent Mppbla lines were not able to produce ROS in response to chitin treatment ( Figure 2B), reminiscent of Mprboh1 mutants. To test whether MpPBLa kinase activity is required for chitin-triggered ROS production, we expressed the wild-type or kinase-dead (MpPBLa*, D228N) version of MpPBLa in Mppbla-1 background. The loss of ROS production of Mppbla-1 in response to chitin treatment was rescued by the expression of wild-type MpPBLa, but not  S2C). Together, these data demonstrate that MpPBLa is required for chitin-induced ROS production in a kinase-activity-dependent manner. Previous studies in angiosperm species revealed only quantitative defects in pattern-triggered ROS production in single, double, or higher-order bik1 or pbl knockout mutants, 1,2,24,25 suggesting extensive genetic redundancy within the PBL family, which is further underscored by the targeting of multiple PBL family members by various pathogen-derived effectors. 2,26 Notably, the size of PBL family is considerably small in Marchantia (3 versus 46 of Arabidopsis). In keeping with the much less redundant MpPBL family, chitin-triggered ROS was completely abolished by loss of a single MpPBL, demonstrating the utility of Marchantia as a genetically tractable study system for plant immune signaling.
Interestingly, Mppbla mutants show defects in thallus development but still produce rhizoids (Figures S2D and S2E). Notably, different PBLs regulate diverse plant growth and developmental processes in Arabidopsis, such as SCHENGEN 1 (SGN1)/PBL15 in Casparian strip formation, 27 or PBL34/35/36 in root quiescent center stem cell maintenance. 28 It will be of interest to dissect the potential regulatory roles of MpPBLa in plant growth and development in future studies.
MpRBOH1 is a bona fide substrate of MpPBLa RBOHD is activated by BIK1 and related PBLs through direct interaction with and phosphorylation of RBOHD N terminus. 4,5,29 Given the shared phenotype of Mppbla and Mprboh1 mutants in chitintriggered ROS response, we hypothesized that MpPBLa similarly activates MpRBOH1 via phosphorylation. In vitro pull-down assays with recombinant maltose binding protein (MBP)tagged MpRBOH1 N terminus (MpRBOH1-Nt) and GST-tagged MpPBLa also showed that MpRBOH1-Nt specifically interacted with MpPBLa ( Figures 3A and S3A). We further performed coimmunoprecipitation (coIP) assays in Marchantia, which demonstrate that MpPBLa-citrine associated with FLAG-MpRBOH1 but not the plasma membrane marker Lti6b in planta ( Figure 3B). We next performed in vitro kinase assays, which confirmed that MpPBLa can trans-phosphorylate MpRBOH1-Nt ( Figure 3C). In Arabidopsis, BIK1 specifically phosphorylates several [S/T]-X-X-L motifs in the N terminus of RBOHD and these motifs are also present in RBOHD orthologs in different plant species. 5 Notably, similar motifs are conserved in the N terminus of MpRBOH1 (Figure S3B). To confirm the specificity of MpRBOH1 phosphorylation by MpPBLa, we performed in vitro kinase assays with MpRBOH1-Nt 4A , an MpRBOH1-Nt variant in which four MpPBLa-mediated phosphosites were mutated to non-phosphorylatable (alanine) residues. MpPBLa-mediated phosphorylation was strongly reduced when all four putative phosphosites in MpRBOH1-Nt were mutated to non-phosphorylatable versions ( Figure 3C), and we subsequently confirmed two out of four by liquid chromatography-tandem mass spectrometry (LC-MS/MS) ( Figure S3C). Together, these results indicate that MpPBLa directly interacts with and phosphorylates MpRBOH1-Nt at specific sites that are conserved among land plant RBOHs.
To test the effect of MpPBLa phosphorylation on MpRBOH1 function, we used a heterologous reconstitution system in human HEK293T cells. 30,31 The expression of MpRBOH1 alone resulted only in minor basal ROS generation that was comparable to that of cells transfected with an empty vector control. When MpRBOH1 was expressed together with the wild type, but not the kinase-dead MpPBLa, a significant increase in ROS production was observed ( Figure 3D), confirming that MpPBLa activates MpRBOH1 in a kinase activity-dependent manner. To test the biological importance of MpRBOH1 phosphorylation by MpPBLa, we expressed in Mprboh1-1 a full-length MpRBOH1 variant harboring non-phosphorylatable (alanine) residue at MpPBLa-mediated phosphosites (MpRBOH1 4A ). In contrast to wild-type MpRBOH1, MpRBOH1 4A failed to restore chitin-triggered ROS production ( Figures 3E and S1F), demonstrating the functional importance of MpPBLa-mediated phosphorylation at these sites. In addition to PBL-mediated phosphorylation, RBOHD activation is regulated via phosphorylation at distinct sites by other kinases, such as Ca 2+ -dependent protein kinases. 22 The observation that a minor ROS production could still be detected in lines expressing MpPBLa-mediated phosphosite mutant version of MpRBOH1 suggests additional MpRBOH1 regulatory mechanism(s) in Marchantia, which will need to be deciphered in future studies.
To further interrogate the conservation of this module between Marchantia and Arabidopsis, we reconstituted the reciprocal PBL-RBOH pathway in human HEK293T cells and measured ROS production indicative of the direct PBL-mediated activation of RBOH enzymatic activity. Both BIK1 and MpPBLa could activate their heterologous substrates (MpRBOH1 and RBOHD, respectively) ( Figures 4A, 4B, and S3D). In contrast, co-expression of PBL13, a previously identified Arabidopsis PBL family member that negatively regulates RBOHD activity, 32 did not induce discernible ROS production compared with expression of RBOHs alone (Figures 4A, 4B, and S3D). We next sought to examine the functional conservation of MpPBLa and BIK1 in planta. We transformed the Arabidopsis bik1 pbl1 mutant with MpPBLa under the control of the native AtBIK1 promoter and observed complementation of bik1pbl1 defective chitin-induced ROS production (Figures 4C and S3E). These results indicate the functional conservation of PBL family proteins in chitin-induced ROS production across land plants.
Aside from apoplastic ROS production, mitogen-activated protein kinase (MAPK) activation is another hallmark of early elicitor-induced immune signaling. 33,34 We thus investigated whether the MpPBLa-MpRBOH1 module is involved in chitininduced MAPK activation in Marchantia. Chitin treatment induces an increase of MAPK activation in the wild-type Tak-1 plants, which was similar in Mppbla and Mprboh1 mutants ( Figure S3F). Multiple PBL family members have also been shown to regulate chitin-induced MAPK activation in Arabidopsis and rice, 24,35,36 and it is possible that chitin-induced MAPK activation and ROS production are regulated by distinct PBL family members in Marchantia. Though Mppbla mutants had unaltered chitininduced MAPK activation, we observed impaired chitin-induced expression of the immune marker gene MpWRKY22 6 in Mppbla ( Figure S3G), suggesting that MpPBLa positively regulates immune gene expression in an MAPK-independent manner.
Our results demonstrate that MpRBOH1 is a bona fide substrate of MpPBLa, which is critical for PAMP-induced ROS production in Marchantia and highlight the striking conservation of this key regulatory step for plant NADPH activation. As Arabidopsis BIK1 controls different cellular immune outputs through direct phosphorylation of diverse substrates, 28,35,[37][38][39] it will be interesting to see to which extent orthologous substrates in Marchantia are similarly regulated during immunity in this evolutionary model system. Future studies will also be needed to reveal the biological functions of MpRBOHs and MpPBLs identified in our study, in immunity and beyond.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

RESOURCE AVAILABILITY
Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Cyril Zipfel (cyril.zipfel@botinst.uzh.ch).

Materials availability
Materials generated in this study are available from the lead contact without restrictions.
Data and code availability d The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium. d This paper does not report original code. d Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.
For ROS production assays, Marchantia gemmae were grown on half-strength Gamborg's B5, 1 % agar with a day/night cycle of 10 h:14 h at 21 C and 16 C.
Arabidopsis mutants and transgenic lines were in Col-0 background. All seeds were surface sterilized, sown on MS media containing 1% (w/v) sucrose, stratified for 2 days in the dark at 4 C, and moved to growth chamber with conditions 16 h day/8 h night at respectively 22 C/18 C and 120 mmol m -2 s -1 . 10-day-old seedlings were transplanted into soil and grown with a day/night cycle of 10 h:14 h at 21 C and 16 C.

Gene identification and phylogenetic analyses
Sequences of Marchantia gene and protein were obtained from www.marchantia.info. Sequences of proteins used for analysis were retrieved from Phytozome, except for Picea abies from https://evorepro.sbs.ntu.edu.sg/. 43 All sequences were aligned with MUSCLE or T-COFFEE. 44,45 The phylogenetic analysis was performed by IQ-TREE or MEGA 46 based on PhyML 47 with a bootstrap analysis of 1000 replicates.

ROS production assays
For assays with Marchantia thalli, 12 thalli discs (approximately 4-mm diameter) from 4-week-old Marchantia plants growing in a KKD Hiros chamber (Clitec) with a day/night cycle of 10 h:14 h at 21 C and 16 C were sampled with a biopsy punch and incubated overnight in sterile water. The water was replaced with the solution containing 0.5 mM L-012, 10 mg/mL horseradish peroxidase (HRP), 0.1 mg/mL chitin (Colloidal Chitin polysaccharide, Elicityl) or 100 nM flg22 (SciLight Biotechnology LLC). DPI treatment was performed as previously described 27 with modifications accordingly. The overnight incubation water was replaced with the 10 mM DPI solution (Sigma) or sterile water, and discs were incubated for 3 h before replacing with the elicitor solution. For the preparation of chitin solution, the desired amount of Colloidal Chitin polysaccharide was weighted, and sterile water was added to achieve the working concertation. The chitin solution was mixed by vortexing before adding L-012 and HRP. Luminescence was captured over 30 min in 20-s intervals with a Photek camera (East Sussex). ROS production assays in Arabidopsis leave discs were conducted as previously described. 5