The receptor-like kinase ARK controls symbiotic balance across land plants

Significance The arbuscular mycorrhizal (AM) symbiosis is considered one of the key adaptations that enabled colonization of land by plants. Here, we identified evolutionarily conserved genes, molecular pathways, and transcription patterns whose association with AM symbiosis arose before the diversification of land plants. Among these, the ARBUSCULAR RECEPTOR-LIKE KINASE (ARK) plays a crucial role in maintaining mutualism in AM symbiosis. We demonstrate functional conservation of ARK from angiosperms to the liverwort Marchantia paleacea, which confirms that its requirement for regulation of symbiosis sustenance evolved before the most recent common ancestor of all land plants.

The mutualistic arbuscular mycorrhizal (AM) symbiosis arose in land plants more than 450 million years ago and is still widely found in all major land plant lineages.Despite its broad taxonomic distribution, little is known about the molecular components underpinning symbiosis outside of flowering plants.The ARBUSCULAR RECEPTOR-LIKE KINASE (ARK) is required for sustaining AM symbiosis in distantly related angiosperms.Here, we demonstrate that ARK has an equivalent role in symbiosis maintenance in the bryophyte Marchantia paleacea and is part of a broad AM genetic program conserved among land plants.In addition, our comparative transcriptome analysis identified evolutionarily conserved expression patterns for several genes in the core symbiotic program required for presymbiotic signaling, intracellular colonization, and nutrient exchange.This study provides insights into the molecular pathways that consistently associate with AM symbiosis across land plants and identifies an ancestral role for ARK in governing symbiotic balance.
The association with the fungus improves plant mineral nutrition, in particular phosphate, in exchange for organic carbon (15)(16)(17)(18)(19)(20)(21)(22)(23).The primary sites of symbiotic nutrient exchange are arbuscules, highly branched tree-like hyphal structures which develop within plant cells.Each arbuscule is surrounded by a plant-derived periarbuscular membrane (PAM), which creates an extensive surface area for nutrient and signal exchange, thereby contributing to plant and fungal fitness.Plant mutants defective in either the uptake of phosphate or the delivery of organic carbon to the fungus result in shriveled arbuscules and in decreased overall root colonization, suggesting that malnourishment of either symbiont is linked to arbuscule development (24)(25)(26)(27)(28)(29).Fungal fitness however also relies on the serine/threonine ARBUSCULAR RECEPTOR-LIKE KINASE 1 (ARK1), which is necessary for maintaining fungal vigor and thus symbiotic balance post arbuscule development (30)(31)(32).ARK is only found in the genomes of AM-competent plants (21,30) suggesting a conserved role in symbiosis across land plants.
Our existing knowledge of the molecular mechanisms governing AM symbiosis predominantly relies on evidence obtained from a limited number of angiosperm models, while the molecular characterization of bryophytes is at its dawn (13,(33)(34)(35)(36).To address the knowledge gap, we focused on examining the extent to which ARK's regulatory role is conserved among terrestrial plants.We conducted a comprehensive investigation involving phylogenetic analysis, comparative transcriptomics, and gene coelimination analysis, using the liverwort Marchantia paleacea as a bryophyte model.We identified a core cluster of 56 AM symbiosis genes conserved in flowering plants and extant bryophytes.The majority of these genes were transcriptionally up-regulated during symbiosis development in M. paleacea and in the angiosperm Medicago truncatula, suggesting their ancestral recruitment for AM symbiosis prior to the MRCA of land plants.The overlay of gene-trait coelimination and transcriptomic data spotlighted ARK as one of the ancestral AM symbiosis genes, and its regulatory role in controlling symbiotic balance was indeed confirmed to be evolutionarily conserved across land plants.

Time-Resolved Transcriptional Responses of M. paleacea to AM
Symbiosis.To understand the molecular changes associated with the development of AM symbiosis in bryophytes, we characterized the transcriptome of the liverwort M. paleacea at 5, 8, and 11 wk post-inoculation (WPI) with Rhizophagus irregularis.R. irregularis was previously described to colonize the central part of the thallus parenchyma, which will be referred to as the midrib region (22,34).
During the initial phases of symbiosis (5 WPI), the primary fungal structures observed within the inoculated thalli were intracellular hyphae (Fig. 1 A-C).As time progressed, all other fungal structures increased in abundance.For instance, arbuscules increased from an average 20% thallus length colonization at 5 WPI to an average 90% at 11 WPI (Fig. 1C).The distinctive purple pigmentation specific to colonized thallus tissue (22) accumulated proportionally to colonization levels across all time points (Fig. 1C).
Similarly, the number of DE genes increased over time in response to progressing AM fungal colonization.This represented a fourfold increase between 5 and 11 WPI, from 292 to a total of 1,145 DE genes, with a distinctive bias towards up-regulated genes (80.4%) (Fig. 1 D-H).The overrepresentation of up-regulated genes mirrors similar observations in distantly related angiosperm models (37)(38)(39).
To investigate the identity of DE genes, the transcriptome of M. paleacea was functionally annotated.Gene ontology (GO) enrichment analysis identified terms significantly overrepresented (adjusted P-value > 0.05) across all stages of AM symbiosis: "gibberellin biosynthetic process", "gibbrellin-20-oxidase activity", "oxidoreductase activity", "nutrient reservoir activity", "apoplast", "iron ion binding" (Dataset S2).The strongest transcriptional response to symbiosis at 11 WPI was additionally enriched in terms associated with fatty acid and carbohydrate metabolism and with strigolactone biosynthesis (SI Appendix, Fig. S1 and Dataset S2).Overall, GO enrichment analysis highlighted that hormonal signaling pathways and symbiotic nutrient metabolismrelated processes, which are commonly associated with angiosperm symbiosis, are also conserved in bryophyte AM symbiosis.

Identification of Evolutionarily Conserved AM Symbiosis-
Associated Genes through Phylogenetics.We traced the evolutionary history of 104 gene families associated with AM symbiosis in angiosperms to compare expression patterns of symbiosis-regulated genes between M. paleacea and tracheophytes (SI Appendix, Table S4 and Dataset S3).Each gene family was selected based on at least one of three criteria: a) GO enrichment analysis indicated that the biological process they contribute to is induced in mycorrhizal M. paleacea, b) the evolutionary history of the clade suggests that the genes are coeliminated with the loss of AM symbiosis across multiple independent events (21,40,41), c) experimental evidence supports a crucial role for the gene in the regulation of AM symbiosis (42,43).For this analysis, 59 publicly available genomes were selected from all major land plant clades, including all nonflowering plant genomes and a selection of mycorrhizal host and nonhost flowering plant models (SI Appendix, Table S5).
Among the 104 families surveyed, 56 displayed distinct orthologs in M. paleacea (Fig. 2A, SI Appendix, Table S4).The majority (49/56) of the evolutionarily conserved gene families (ECGFs) included a one-to-one ortholog in M. paleacea, while in 16/56 ECGFs a single M. paleacea "proto-ortholog" corresponded to several angiosperm genes, as a result of one or multiple independent duplications within the tracheophyte lineage (SI Appendix, Table S4).These 56 orthogroups span a variety of biological processes required for AM symbiosis: NOPE1 is necessary for presymbiotic rhizosphere signaling, RAD1, AP2A, DELLA1, NSP1, and NSP2 are involved in transcriptional regulation of the symbiosis, and AMT2, PT4, HA1, STR, STR2, and SUT2 are required for nutrient exchange at the arbuscule-interface (Fig. 2A and SI Appendix, Table S4).Notably, the SUT2 gene was absent from the latest version of the M. paleacea genome but was included in the 1KP transcriptome (Dataset S3).Of the remaining 48 families examined, no clear ortholog was identified in M. paleacea or its sister clade Marchantia polymorpha.However, in many cases, orthologs were identified in other bryophytes, suggesting that the gene was lost in the genus Marchantia (SI Appendix, Table S4).
To discern whether ECGFs acquired an AM symbiotic function before the MRCA of land plants, analysis of gene coelimination patterns across multiple independent evolutionary losses was performed.The availability of two liverwort genomes from the same genus with M. polymorpha having lost the ability to engage with AM fungi (36,44), allowed us to extend the study of AM gene coelimination to include bryophytes alongside Angiosperms.We identified 30 gene families only conserved in AM-competent angiosperms (21) and in M. paleacea (Fig. 2B).Among these gene families, fourteen have experienced the loss of the corresponding M. polymorpha gene, indicating that these genes tend to be abandoned when symbiosis is lost in plant species.(Fig. 2B and SI Appendix, Table S4).For the remaining half of the genes conserved for symbiosis in angiosperms, an ortholog was identified in both the host M. paleacea and the nonhost M. polymorpha, suggesting that the pattern of coelimination was not conserved for these families (Fig. 2B).Overall, we found an unprecedented number of AM genes whose conservation dates back before the MRCA of land plants.The recursive coelimination of a subset of these genes with the loss of AM symbiosis suggests that the conserved genes acquired their symbiotic function in ancestral land plants.

Comparative Analysis Reveals Deep Homologies in Tran scriptional Responses to AM Symbiosis from Bryophytes to
Angiosperms.The symbiotic expression profile of the 56 ECGFs was compared between M. paleacea and the angiosperm AM symbiosis model M. truncatula (23).Half (28/56) of the genes conserved in M. paleacea were DE during symbiosis (Fig. 3 and SI Appendix, Table S6).In addition, 89.3% (25/28) of the evolutionarily conserved genes induced in M. paleacea, were also up-regulated in M. truncatula during symbiosis, demonstrating their conserved transcriptional regulation across land plants (Fig. 3).Several genes with a known role in presymbiotic signaling are co-up-regulated in M. paleacea and M. truncatula, notably the N-acetylglucosamine transporter NOPE1 (45) and several members of the strigolactone biosynthesis pathway (Fig. 3).However, the majority of genes coregulated in M. paleacea and M. truncatula have been linked to downstream stages of symbiosis: transcriptional regulation, arbuscule dynamics, and nutrient exchange, including the well-characterized transcription factors Required for Arbuscule Development 1 (RAD1) and Ethylene Response Factor 1 (ERF1) (Fig. 3 and SI Appendix, Table S6).Taken together, the above findings suggest that molecular modules necessary for symbiosis establishment and maintenance in angiosperms are conserved and induced during liverwort AM symbiosis.

pnas.org
Acid Oxidase (KAO) were expressed but not induced (Dataset S1).Thus, the synthesis of Gibberellin-like diterpenoids might take place during bryophyte AM symbiosis.
Despite the overall conservation of transcriptional responses, we identified some differences between M. paleacea and M. truncatula AM gene expression profiles.In angiosperms, the H + -ATPase HA1 is responsible for generating the proton gradient required for phosphate transport across the PAM (49,50).Although HA1 is conserved in bryophytes (Marpal_utg000008g0017601), it is not induced during symbiosis in M. paleacea (SI Appendix, Table S6).Instead, three HA proton pumps more distantly related to angiosperm HA1 were induced in M. paleacea (Marpal_utg000039g0074811, Marpal_utg000045g0082001, and Marpal_utg000126g0168721) (Dataset S3).The conservation of a cellular function required for AM symbiosis might thus not be necessarily mediated by conservation of a specific gene.
By overlaying the comparative transcriptomics dataset with the coelimination analysis across angiosperms and liverworts, we identified a core set of ten AM-specific gene families: ARK, RAD1, ERF1, Replication factor C (RFC), Ammonium Transporter 2 (AMT2), short-chain Dehydrogenase/reductase (DHY), Stunted Arbuscule (STR) and Stunted Arbuscule 2 (STR2), ABC transporter B family member 20a (ABCB20a), and ABC transporter B family member 20b (ABCB20b) (Fig. 3).The transcriptional induction of a representative subset of these gene families during AM symbiosis was validated in an independent experiment and confirmed (SI Appendix, Fig. S3).In summary, the embryophyte-wide transcriptional induction during symbiosis and the coelimination of these genes with the loss of AM symbiosis provide independent lines of evidence for their mycorrhizal role before the MRCA of land plants.

The Receptor-Like Kinase ARK Is Required for Symbiosis
Maintenance in Bryophytes.We next wished to address the hypothesis that the newly identified set of ancestral AM symbiosisassociated genes are functionally required for AM symbiosis throughout land plants.We concentrated on the M. paleacea ortholog of ARK, which is exclusively present in AM symbiosis-competent The core conserved AM symbiosis toolkit evolved before the last common ancestor of land plants; evolutionarily conserved genes encompass the following functional groups: presymbiotic signaling, fungal perception and transcriptional regulation, arbuscule-associated genes, and genes of yet unknown function.(B) A set of symbiotic genes is conserved across all land plants that are able to engage in AM symbiosis but is recursively coeliminated from the genome of land plants that do not engage in AM symbiosis (pink box); a small set of symbiosis genes is coeliminated with symbiosis abandonment across angiosperms but not in the genus Marchantia (light blue box); magenta dot = AM trait, transverse dashes = independent AM trait/gene losses.The phylogenetic trees for each gene family can be found in Dataset S3, and comprehensive information on the results of phylogenetic analysis is available (SI Appendix, Table S4).Only genes with clear orthologs in the genus Marchantia have been included.land plants (Fig. 3A), and generated CRISPR/Cas9-edited Mpa_ark mutants (Fig. 4B and SI Appendix, Fig. S4A).
Nonseed plants share a single copy of ARK, while an ancient genome duplication in spermatophytes gave rise to ARK1 and its closest paralog ARK2 (Fig. 4A) (30).We generated three CRISPR/ Cas9 lines carrying indel mutations in the second exon of MpaARK (Fig. 4B), leading to a premature stop codon ahead of the protein's transmembrane domain (SI Appendix, Fig. S4A).Inoculation with R. irregularis resulted in 30 to 50% of thallus colonization in all Mpa_ark KO alleles, relative to >90% colonization of wild-type M. paleacea after 10WPI (Kruskal-Wallis, P < 0.05) (Fig. 4C).Vesicles occupied 80% of the length of wild type thalli but were only present along 30% thallus length in the three independent ark mutant lines (Fig. 4C) while arbuscule morphology was equivalent to wild type in all colonized Mpa_ark lines (SI Appendix, Fig. S6).The M. paleacea ark phenocopies rice ark1 and ark2, displaying lower colonization levels, especially for vesicles, while maintaining wild-type arbuscule morphology (30,32).These findings indicate that the requirement for ARK to maintain fungal vigor is evolutionarily conserved across angiosperms and bryophytes.
To address functional conservation, we examined whether the M. paleacea ortholog of ARK1/ARK2 could functionally complement the rice ark1 mutant.The predicted protein sequence of MpaARK encodes an extracellular SPARK (similar protein to ARK) domain, followed by an alpha-helix transmembrane domain and an intracellular active kinase domain (Fig. 4D).Monocot ARK genes have lost their extracellular SPARK domain, which is instead retained in the closely related SPARK1 clade (Fig. 4D) (30).As MpaARK encodes a SPARK domain, two variants of the construct were designed, to include 1) the full-length MpaARK or 2) the signal peptide, transmembrane, and kinase domains of MpaARK, mimicking the domain structure of OsARK1 (MpaARK-Kin) (SI Appendix, Fig. S4B).As previously reported (32), colonization levels in rice ark1 were significantly reduced (Fig. 4E), which was particularly pronounced for vesicles with 18.8% in the mutant relative to 58.4% in the wild type (TukeyHSD, P < 0.05) (Fig. 4E).This stark reduction in colonization was fully restored to wild type levels when either wild type OsARK1, the full-length MpaARK, or the kinase-only MpaARK-Kin were introduced in the ark1 mutant background (TukeyHSD, P > 0.05) (Fig. 4E).Both the full-length and the kinase-only MpaARK sequences also complemented the vesicle phenotype of ark1 to the same degree as OsARK1 (TukeyHSD, P > 0.05) (Fig. 4E), suggesting that the molecular function of this kinase is evolutionarily conserved.Arbuscule morphology was akin to wild type across all genotypes and root types observed, as previously reported for rice ark1 (30,32).The combined findings from phylogenetic analysis, transcriptomics, and experimental data presented in this research indicate that the essential role of ARK in regulating symbiotic balance is ancestral to land plants.

Discussion
We identified a core AM symbiosis genetic programme of 56 genes conserved across land plants, from bryophytes to angiosperms.Overlaying comparative transcriptomic data with phylogenetic analysis further corroborated that a large portion of these genes is also induced during liverwort M. paleacea symbiosis, providing evidence for their activity in bryophyte symbiosis.Our findings are supported by earlier phylogenetic studies, which collectively identified 24 of these 56 ECGFs to be conserved in bryophytes (13,36,(51)(52)(53).    biological replicates shown; Kruskal-Wallis test was performed to assess significant differences between genotypes, followed by post hoc pairwise comparison using Wilcoxon Rank Sum Test; independent genotypes are grouped into significance classes based on test statistics: Different letters represent significant difference (Wilcoxon Rank Sum, P-value > 0.05), "ab" is not significantly different from either "a" or "b." (D) Schematic representation of ARK and SPARK protein structures in M. paleacea, M. truncatula, and rice.ARK1 and ARK2 in monocots and KIN6 (ARK2) in M. truncatula have lost their SPARK domain, which is retained in the homologous SPARK1 clade (31).(E) Assessment of the AM symbiosis phenotype of WT, ark1−/−, ark1 OsARK , ark1 MpaARK , and ark1 MpaARK-Kinase rice lines inoculated with R. irregularis, after 9WPI; each bar indicates the average percentage value of the fungal structure, all biological replicates are shown; T = total colonization, A = arbuscules, V = vesicles; An ANOVA test was performed to compare each fungal structure across genotypes, followed by post hoc comparison using a Tukey's honest significant difference (HSD) test.Different letters represent significant difference (Tukey HSD, P-value < 0.05).
Among these, functional conservation in bryophyte symbiosis was recently confirmed for the Calcium-and Calmodulin-dependent protein Kinase CCaMK/DMI3 (13), the transcription factors CYCLOPS/IPD3 (13) and MpaWRI (MtERF1) (35), and the strigolactone biosynthesis genes MAX1, CCD8, and ZAS (34).We found several orthologs of PAM-intrinsic proteins to be conserved in M. paleacea, which pushes back the core genetic program associated with arbuscule development and functioning to the MRCA of land plants.In particular, the conservation of the symbiotic ammonium transporter AMT2 (54) and the symbiotic phosphate transporter PT4 (24,25,55), coupled with evidence that M. paleacea is capable of taking up symbiosis-derived phosphate and nitrogen (22), suggests that the molecular mechanisms for symbiotic nutrient uptake are conserved across land plants.Despite previous studies suggesting that phosphate transporters in flowering plants and liverworts have independent origins (36), our analyses point toward a shared origin, indicating that these transporters are in fact orthologs.
Of the 56 ECGFs in M. paleacea identified here, 14 lost the ortholog in the symbiosis-incompetent M. polymorpha.The early loss of these genes after mutualism abandonment across land plants strongly suggests that their biological function is solely required for AM symbiosis.On the other hand, a second set of 16 genes were found to be symbiosis-specific in angiosperms (21) but conserved across both AM-competent and AM-incompetent bryophytes.The presence of these genes in AM-incompetent bryophytes suggests that they might be required for biological functions that are independent of symbiosis.This is the case for the closest homolog of VAPYRIN (VPY), VPY-like.VPY-like is conserved in mycorrhizal angiosperms (40), however, the gene is required for the regulation of a gametophyte developmental stage in the symbiosis-incompetent bryophyte Physcomitrella patens (56).VPY and other genes within this subgroup might be similarly required in bryophytes for processes other than AM symbiosis, which disappeared in the tracheophyte lineage after the switch from a gametophyte-to a sporophyte-dominant life cycle.Given the recent loss of AM symbiosis in M. polymorpha (70.7 to 5 mya) (57), it is also possible that the genetic sequences retained in this nonhost liverwort are undergoing early stages of pseudogenization, as it was previously reported for CCaMK/ DMI3 (13).
Remarkably, orthologs of ARK are required to sustain AM symbiosis in M. paleacea, rice, and M. truncatula, and ark1 mutants are distinctly characterized by a diminished presence of fungal vesicles in colonized tissue (21,30).Moreover, the ability of MpaARK to fully restore fungal colonization in rice ark1 mutants confirms that its signaling function is conserved across more than 450 million years of plant evolution and that ARK was a component of AM symbiosis in its ancestral form.Therefore, the yet undisclosed mechanism by which ARK ensures fungal fitness during symbiosis establishment, but post arbuscule development, is ancestral to land plants.In addition, full-length MpaARK does not hinder downstream signaling at the rice PAM, since the native MpaARK sequence was able to rescue the mutant phenotype of ark1 to the same degree as its variant missing the SPARK domain.We conclude that loss of the SPARK domain from monocot ARK genes reflects a decreased reliance on ARK for extracellular signaling, a function that may however be provided by other SPARK domain-containing proteins in rice (30).
In summary, this research has revealed an unparalleled degree of conservation across all terrestrial plants of the genetic machinery that underpins AM symbiosis and of its transcriptional regulation.Our data support the view that the common ancestor of embryophytes possessed most of the essential molecular tools required for AM symbiosis as we know it today, as evidenced by the unaltered molecular function of the MpaARK kinase after more than 450 million years of evolution and diversification.The strong natural selection under which ARK has been maintained for this extensive period of time is not only indicative of the central importance of its signaling role but also reflecting the substantial benefit that AM symbiosis has provided to land plants throughout their evolution.

Summary of Plant Material and Growth Conditions for Mycorrhizal
Assays.M. paleacea wild type corresponded to the earlier published accession (13).All rice material arose in the Oryza sativa ssp.japonica cv.Nipponbare background.For AMS colonization, 4-wk-old M. paleacea thalli were transferred to sand pots and inoculated with 4% R. irregularis "crude" inoculum (see extended methods in SI Appendix).Plants were grown at 22 °C, 16 h day/night cycle with a light intensity of 200 μmol m −2 s −1 PAR.M. paleacea plants were watered three times per week with "artificial rainwater solution" (pH 5.8) (4).For AMS colonization of O. sativa, germinated seedlings were transferred into cones containing sand mixed with 300 axenic spores of R. irregularis.Rice plants were fertilized twice weekly with Hoagland Solution containing 25 µM KH 2 PO 4 and grown in 12 h day/night cycles at 28/20 °C and 60% humidity.
The extent of fungal colonization was determined by trypan blue staining.For rice, the previously published protocol was used (32).For M. paleacea, thalli were harvested and incubated overnight in 10% (w/v) potassium hydroxide, washed three times with ddH 2 0, incubated in 0.3 M HCl for 15 min at room temperature, then incubated overnight in staining solution [50% (v/v) lactic acid, 25% (v/v) glycerol, 25% (v/v) ddH 2 0, 0.1% (w/v) trypan blue].For sectioning, stained samples were incubated in destaining solution [50% (v/v) lactic acid, 25% (v/v) glycerol, 25% (v/v) ddH 2 0] for 1 h, then embedded in 3.5% (w/v) agarose gel.Three transverse sections (200 μm) were taken at each of 10 equally distanced positions spanning the length of each thallus, using a Hyrax V50 vibratome (Zeiss, Oberkochen, Germany).Three sections per position were imaged under a Labophot light microscope (Nikon, Minato, Japan) (SI Appendix, Fig. S2).For each biological replicate, fungal structures and cell-wall pigment accumulation were quantified by counting their presence or absence at each of the 10 transverse positions imaged.The thallus colonization ratio of biological replicates was calculated by dividing the total number of observations of each structure (x) over the total number of transverse positions imaged (x/10).The following morphological characteristics were quantified with the above method: cell-wall pigment, intracellular hyphae, arbuscules, vesicles, and spores.
Materials and methods are comprehensively detailed in SI Appendix.
Data, Materials, and Software Availability.Transcriptome data have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GSE245702) (58).

Fig. 1 .
Fig. 1.Transcriptional responses of M. paleacea to R. irregularis colonization.(A) Transversal section of a M. paleacea thallus colonized by R. irregularis; fungal structures in the thallus midrib are stained in ink and recognizable by their dark blue coloration.(Scale bar, 100 μm.) (B) Transversal section of an M. paleacea thallus colonized by R. irregularis and stained with wheat germ agglutinin (WGA)-Alexa Fluor 488; yellow arrow = fungal hypha, white arrow = arbuscule, magenta arrow = fungal vesicle.(Scale bar, 50 μm.)(C) Arbuscular mycorrhiza colonization levels of WT M. paleacea thalli inoculated with R. irregularis at 5, 8, and 11 WPI; P = pigment, IH = intracellular hyphae, A = arbuscules, V = vesicles; each independent replicate was screened for all fungal structures over a total of 10 sections, the overall ratio of positive observation over the total number of observations is displayed on the y axis.Bars represent the average number of observations of each fungal structure across replicates, all biological replicates are shown.(D) Volcano plot of differentially expressed (DE) genes (adjusted P-value < 0.05, Log2 fold change > |1|) between mock and R. irregularis-colonized thalli at 5 WPI.Statistically significant DE genes are displayed in magenta.(E) Volcano plot of DE genes between mock and R. irregularis-colonized thalli at 8 WPI.(F) Volcano plot of DE genes between mock and R. irregularis-colonized thalli at 11 WPI.(G) Venn diagram of up-regulated genes in pairwise comparisons of mock and R. irregularis-colonized thalli.(H) Venn diagram of down-regulated genes in pairwise comparisons of mock and R. irregularis-colonized thalli.

Fig. 2 .
Fig. 2. Ancestral AM symbiosis gene families identified through phylogenetics.(A)The core conserved AM symbiosis toolkit evolved before the last common ancestor of land plants; evolutionarily conserved genes encompass the following functional groups: presymbiotic signaling, fungal perception and transcriptional regulation, arbuscule-associated genes, and genes of yet unknown function.(B) A set of symbiotic genes is conserved across all land plants that are able to engage in AM symbiosis but is recursively coeliminated from the genome of land plants that do not engage in AM symbiosis (pink box); a small set of symbiosis genes is coeliminated with symbiosis abandonment across angiosperms but not in the genus Marchantia (light blue box); magenta dot = AM trait, transverse dashes = independent AM trait/gene losses.The phylogenetic trees for each gene family can be found in Dataset S3, and comprehensive information on the results of phylogenetic analysis is available (SI Appendix, TableS4).Only genes with clear orthologs in the genus Marchantia have been included.

Fig. 3 .
Fig. 3.Comparative analysis reveals conserved transcriptional responses to AM symbiosis.Heatmap of variance-stabilized row-centered counts for AM-associated M. paleacea evolutionarily conserved genes that were DE during AM symbiosis.M. truncatula orthologs DE during AM symbiosis as described previously(23); orthologs significantly DE in both species (L2FC > |1|, adjusted P-value < 0.05) are shown in yellow (striped), orthologs not DE in M. truncatula are shown in blue (striped); "angiosperm AM" = the gene clade is conserved across all Angiosperms (white) or it is recursively lost after symbiosis abandonment in Angiosperms (gray)(21)."bryophyte AM" = The gene is conserved in the host liverwort M. paleacea and in the nonhost M. polymorpha (white) or lost in the nonhost liverwort M. polymorpha (gray).

3 AT1T
G49 730 .1_Arabi dop sis_ thal iana P S M E _ 4 1 1 5 4 _ P s e u d o ts u g a _ m e n z ie s ii LOC_Os 04g3511 4.1_Oryz a_sativa C in k a _ R W R 8 3 3 1 7 .1 _ C in n a m o m u m _ k a n e h ir a Se lal ep _6 57 _s na p_ ge ne _0 .25 _m RN A_ 1_ Se lag ine lla _le pid op hy lla Lc hi 34 60 7_ Li ri od en dr on _c hi ne ns e u ja _ p li c a ta M t_ K IN 6 _ M e d tr 6 g 0 0 7 6 9 0 .1 _ M e d ic a g o _ tr u n ca tu la Th up l.2 93 79 11 4s 00 04 .1 .p_T hu ja _p lic at a L c h i1 2 2 8 5 _ L ir io d e n d ro n _ c h in e n s e P ic a b i_ M A _ 5 7 6 5 4 g 0 0 1 0 _ P ic e a _ a b ie s S o ly c 0 5 g 0 0 5 6 6 0 .1 .1 _ S o la n u m _ ly c o p e rs ic u m GSM UA_ Ach r2P 212 60_ 001 _Mu sa_ acu min ata Ad in el _M C O 55 62 54 4. 1_ Ad ia nt um 2 .p_ S o rg h u m _ b ic o lo r C in ka _R W R 79 93 2. 1_ C in na m om um _k an eh ira e Sel mo_ Sm 131 134 _Se lagi nell a_m oell end orff ii P IL A _ 3 6 9 4 8 _ P in u s _ la m b e rt ia n a T h u p l. 2 u ja _ p li c a ta AAN G00 7939 _Ant hoce ros_ angu stus A A L B A 5 B 1 0 3 7 2 2 4 P 1 _ A b ie s _ A lb a Se lm o_ Sm 12 68 22 _S ela gin ell a_ mo ell en do rff ii c a ta e v m _ 2 7 .mo d e l.A m T r_ v 1 .0_ s c a ff o ld 0 0 0 0 6 .30 7 _ A m b o re ll a _ tr ic h o p o d a AagrB ONN_ evm.m odel.S c2ySw M_11 7.177 2.1_A nthoce ros_a gresti s T h u p l. 2 9 3 7 9 8 5 9 s 0 0 0 2 .1 .p_ T h u ja _ p li c a ta Zeama_ GRMZM 5G8494 71_P01 _Zea_m ays

Fig. 4 .
Fig. 4. The function of ARK in AM symbiosis is evolutionarily in land plants.(A).Phylogenetic tree of the URK-2 RLK subfamily comprising the ARK and the SPARK clades.Bootstrap values of important nodes are shown; Dark blue = Angiosperm, Host; Light blue = Nonflowering plant, Host; Magenta = Angiosperm, Nonhost; Pink = Nonflowering plant, Nonhost.(B)MpaARK gene structure and alignment of independent editing events in three Mpa_ark KO lines; exons are boxed and numbered; the position of the gRNA binding site is highlighted by a magenta arrow and magenta text, mutations are highlighted in teal; TM = alphahelix transmembrane domain.(C) AM symbiosis phenotype of WT and ark mutant M. paleacea thalli at 10 WPI with R. irregularis; P = pigment, IH = intracellular hyphae, A = arbuscules, V = vesicles, S = spores; each bar indicates the average ratio of observations of the respective fungal structure over a total of 10, all biological replicates shown; Kruskal-Wallis test was performed to assess significant differences between genotypes, followed by post hoc pairwise comparison using Wilcoxon Rank Sum Test; independent genotypes are grouped into significance classes based on test statistics: Different letters represent significant difference (Wilcoxon Rank Sum, P-value > 0.05), "ab" is not significantly different from either "a" or "b." (D) Schematic representation of ARK and SPARK protein structures in M. paleacea, M. truncatula, and rice.ARK1 and ARK2 in monocots and KIN6 (ARK2) in M. truncatula have lost their SPARK domain, which is retained in the homologous SPARK1 clade(31).(E) Assessment of the AM symbiosis phenotype of WT, ark1−/−, ark1 OsARK , ark1 MpaARK , and ark1 MpaARK-Kinase rice lines inoculated with R. irregularis, after 9WPI; each bar indicates the average percentage value of the fungal structure, all biological replicates are shown; T = total colonization, A = arbuscules, V = vesicles; An ANOVA test was performed to compare each fungal structure across genotypes, followed by post hoc comparison using a Tukey's honest significant difference (HSD) test.Different letters represent significant difference (Tukey HSD, P-value < 0.05).