Symbiont effectors modulate plant signaling to increase host stress resilience

Abstract

Plant-colonizing mutualistic symbionts confer beneficial effects to their hosts, which often includes increased growth as well as biotic and abiotic stress resilience. How these benefits are activated on a molecular level is mostly unknown. Here, we describe effector candidates of the fungal symbiont Serendipita indica (Si), which modulate plant stress signaling pathways. By analyzing the Si effector interactome, we reveal frequent targeting of stress-related host proteins, which are linked to the identified effector signaling functions. Moreover, functional data indicate that Si effectors modulate abiotic stress response of Arabidopsis, as well as resistance to pathogen infection. Analysis of symbiont effectors might not only uncover previously unreported molecular mechanisms that increase plant fitness but might also be used to identify potential genetic traits for crop improvement under changing climates.

Introduction

As sessile organisms, plants cannot escape environmental fluctuations and need to adapt to stress events. These stresses can be abiotic, including drought, cold, heat, high light, and salinity or biotic stresses caused by pathogens and insects. The molecular responses of plants to these stresses are supported by a highly interconnected hormone signaling network. In response to unfavorable conditions, plants activate extensive hormone signaling for adaptation and defense (Verma et al. 2016; Ku et al. 2018; Hou and Tsuda 2022). Strongly interconnected hormone signaling pathways regulate all aspects of plant life (Vega et al. 2019; Altmann et al. 2020; Emenecker and Strader 2020; Khan et al. 2020; Hou and Tsuda 2022). As a result, increased crop stress resistance is linked to essential plant processes and particularly to growth-related pathways. To alleviate the impact of climate change on global agriculture and the growing number of people falling into food insecurity (Ahmadalipour et al. 2019; Mirón et al. 2023), it will be necessary to identify measures for plants to manage increasing environmental stresses to reduce overall crop losses.

Plants have evolved a complex network of hormones to integrate developmental and environmental cues. These include abscisic acid (ABA) and auxin (AUX), which are canonically associated with growth and development (Emenecker and Strader 2020); and salicylic acid (SA) and jasmonic acid (JA), which are widely investigated in plant immunity (Hou and Tsuda 2022). Due to the complexity of the hormone network, no plant hormone operates in isolation, with many points of crosstalk having evolved within the network to co-regulate different processes and stress adaptation (Ku et al. 2018; Yang et al. 2019; Altmann et al. 2020; García-Andrade et al. 2020).

Beneficial symbionts are part of microbial communities and have been shown to enhance crop fitness under many stress conditions, including cold (Acuña-Rodríguez et al. 2020), heat, salinity, drought, and extreme temperatures (Harman 2011; Latef et al. 2016; Begum et al. 2019; Gupta et al. 2021; Mitra et al. 2021; Boorboori and Zhang 2022). At the same time, beneficial microbes also improve defense mechanisms against pathogens. Induced systemic resistance is a priming process used by growth-promoting bacteria and fungi to enhance host defense against pathogens and herbivores (Pieterse et al. 2014). Beneficial symbioses are typically associated with improved nutrient exchange. Mutualists are supplied with carbon and in return supply their host with essential minerals, with many terrestrial plants depending on beneficial fungi for nutrient supply (Behie and Bidochka 2014; Acuña-Rodríguez et al. 2020). Improved nutrient subsistence, however, cannot explain the diversity of beneficial properties symbiont-colonized host plants acquire (Xu et al. 2018; Yan et al. 2019), suggesting additional mechanisms are employed by symbiotic microbes to promote beneficial effects in host plants.

The symbiont Serendipita indica (Si) stimulates growth and development in addition to increasing resilience to both biotic and abiotic stresses (Baltruschat et al. 2008; Vadassery and Oelmüller 2009). So far, these effects have been found to be associated with the ability of the symbiont to influence plant ROS levels but also the manipulation of host hormone networks (Nath et al. 2016; Xu et al. 2018). Similar to pathogens, symbionts must suppress host immunity (Jacobs et al. 2011) and utilize effectors to manipulate host immune responses in the course of colonization (Jacobs et al. 2011; Wawra et al. 2016; Daneshkhah et al. 2018; Yang et al. 2022). While microbes often target hormone homeostasis (Pozo et al. 2015), it is unclear to what extent effectors are involved and reprogram underlying signaling networks. Comparative interactome studies have revealed that Si effector candidates (SIECs) specifically target the Arabidopsis hormone network and to a broader extent than the effector repertoires of different pathogens (Akum et al. 2015; Shen et al. 2018; Osborne et al. 2023). Further functional characterization of these putative effectors has demonstrated that those targeting hormone pathways can improve plant growth when expressed constitutively, highlighting their potential as tools to help understand beneficial host signaling and symbiosis (Osborne et al. 2023) .

Here, we explore the beneficial properties of SIECs in plants exposed to abiotic and biotic stresses. Functional analyses in protoplasts revealed modulation of abiotic stress and pattern-triggered immunity (PTI) signaling. Integration of this information with previously published SIEC-Arabidopsis protein interactome revealed a diversity of SIECs, which might coordinate stress responses in whole plants by altering target activity. In both directed and unbiased phenotyping assays, we demonstrate that SIECs increase host growth, improve resilience to abiotic stresses, and alter host susceptibility to pathogen infection.

Results

SIECs modulate stress and hormone signaling in protoplasts

We previously identified SIECs involved in hormone signaling and root growth promotion (Osborne et al. 2023). Since hormones are required in major signaling and stress pathways of plants (Ku et al. 2018; Khan et al. 2020), we aimed to investigate the function of effector candidates in the modulation of stress responses. We previously analyzed specific hormone signaling markers (pHORMONE) in Arabidopsis thaliana (Arabidopsis). Here we utilized 35S::SIEC coexpressed with markers containing promotors (pSTRESS) of genes responsive to abiotic stress/heat (DREB2a and HSP81.1) or biotic stress/flg22-induced PTI (PHI1 and FRK1) (Fig. 1a). In the first landmark screen, 106 SIECs were tested against each marker in protoplasts (1696 assays in total) (Fig. 1a): 82 of 106 SIECs showed modulation of 1 or multiple stress markers, and 49 and 40 SIECs altered expression of the abiotic stress markers pDREB2a::LUC and pHSP81.1::LUC, respectively. For PTI, 28 SIECs altered pPHI1::LUC and 26 pFRK1::LUC signaling (Fig. 1b, Table S1). The majority of SIECs affected only 1 or 2 pathways (Fig. 1c); most of these SIECs modulated only abiotic stress markers, 16 altered only PTI responses, and 28 affected both (Fig. 1d). We validated the landmark screen by performing 3 additional biological replicates of the top 10 inducers and repressors of the pDREB2a::LUC and pPHI1::LUC markers in protoplast assays. Of the 20 respective validations performed, we were able to confirm effector function in 60% of cases by statistical verification of mock and treatment response relative to the empty vector control (Fig. 1e, Table S1). While somewhat low due to technical variations between biological repeats, this figure was still encouraging that effectors might modulate respective pathways in plants. Subsequent phenotyping assays revealed stress response alterations by effectors, which correspond to the signaling pathway function in protoplast assays (Fig. 4f–i). Taken together, these protoplast assays indicated that Si utilizes effector proteins not only to alter immunity signaling but also to modulate environmental stress responses.

Figure 1.

SIECs modulate stress and hormone signaling pathways. SIEC function on abiotic stress and immunity marker expression in protoplasts. a) Overview of initial stress marker screen in protoplasts. b) Number of SIECs inducing or repressing abiotic stress (DREB2a and HSP81.1) or PTI marker (PHI1 and FRK1) expression calculated from multiplied ratios of mock and treatment conditions. c) Number of SIECs modulating 1 or multiple marker pathways indicating marker-specific SIEC functions. d) Number of SIEC modulating 1 or both stress signaling pathways indicating pathway-specific SIEC functions. e) Confirmation of SIEC function in modulating pDREB2a::LUC and pPHI1::LUC markers, respectively. f) Summary of SIECs with overlapping stress and hormone signaling modulating functions. g) Number of SIECs with overlapping hormone and stress functions in protoplast pMARKER assays. Top bar chart shows number of SIECs modulating stress markers. Bar charts on the left and right indicate the number of SIECs additionally altering individual or multiple hormone markers (hormone data adapted from Osborne et al. 2023). Number of SIECs affecting different combinations of hormone markers and the individual stress pathways is listed between the different bar charts. Abbreviations: AUX, auxin; ABA, abscisic acid; CK, cytokinin; JA, jasmonic acid; SA, salicylic acid.

Figure 4.

Drought and salt stress phenotyping of SIEC-overexpressing seedlings. Roots of 14-d-old seedlings treated with mannitol, NaCl or control conditions were measured. Root responses are shown relative to GFP-expressing plants (in %). a–d) Effect of SIEC expression on root length (a), on hypocotyl length (b), on lateral root number (c), and average distance between lateral roots (d) in control conditions. e) Representative pictures of plant growth on different stress treatment media. f, g) Effect of mannitol (f) and NaCl (g) on root length of SIEC-overexpressing seedlings relative to GFP-expressing control plants (in %). h, i) Lateral root phenotypes are altered in SIEC-overexpressing plants. Number of lateral roots was counted in mannitol and mock-treated plants (h). Lateral root distances were calculated by dividing lateral root numbers by root length (i). Both graphs show relative root trait changes in response to mannitol treatment compared with GFP-expressing control plants (in %). Positive values indicate increases of root traits as compared with GFP plants. GFP-expressing controls are shown in blue; significant effects are indicated in yellow. *, P < 0.05; **, P < 0.01; ***, P < 0.001; statistical significance was calculated using a 2-tailed, unpaired t test of at least 3 replicates. Sample sizes (n) are indicated above each box.

Given that stress responses are tightly linked to plant hormone regulation, we combined our published hormone data (Osborne et al. 2023) with the present stress marker signaling data in an UpSet plot (Fig. 1f, g). Sixty-six SIECs modulate both hormone and stress signaling in protoplasts (80%, 66 out of 82 SIEC with stress function). In concordance with our previous observation that SIECs are highly specific in the hormone pathways they putatively target, most effectors that modulated stress signaling markers affected only 1 hormone pathway in protoplasts (33 SIECs; Fig. 1f, g). To investigate whether the potential hormone and stress function of SIECs was interconnected, we analyzed the dataset toward possible groupings. SIECs only modulating immunity markers similarly affected all hormone markers (Fig. S1a). SIECs only altering heat stress response also changed AUX, SA, and ABA and to a lesser extent CK and JA regulation in protoplast assays (Fig. S1a). While most SIECs that affected pPHI1 and pFRK1 activity also modulated the JA signaling marker activity (Fig. S1b), the alignment of defense hormone and stress signaling was not always as expected.

For example, 19 SIECs out of 49 which modulated pDREB2a activity additionally targeted multiple hormone pathways, and we observed a similar number of effectors targeting each hormone pathway that altered abiotic stress or immunity marker expression (Fig. S1b). The analyses did not reveal a clear correlation between specific hormone and stress signaling functions of SIECs but rather suggested broader interdependencies with hormone signaling pathways. This might reflect the involvement of hormones in very diverse aspects of plant stress responses (Fahad et al. 2015; Ku et al. 2018; Khan et al. 2020) and the functions of proteins in multiple hormone pathways, of which many have not been described yet (Altmann et al. 2020). While these protoplast assays assigned putative pathway-modulating functions to SIECs, it raised the question of how these proteins might alter these pathways to transduce phenotypic traits in plants.

Symbiont effectors interact with stress-associated plant proteins

We previously employed yeast 2-hybrid analyses to map the SIEC interactome. We identified 33 SIECs interacting with 156 Arabidopsis proteins in the initial yeast 2-hybrid screen (Osborne et al. 2023). To support the biological validity of the SIEC-Arabidopsis interactome, we localized all 33 effector candidates except SIEC54 in N. benthamiana and evaluated how frequently SIECs appeared in the same cellular compartment as their 156 target proteins. Confocal microscopy revealed distinct localizations for SIECs in planta, including the nucleus, cytoplasm, and endoplasmic reticulum (Fig. 2a, Fig. S2).

Figure 2.

S. indica effectors possess distinct subcellular localizations and co-localize with their target proteins in silico. a) Confocal laser scanning microscopy images of SIEC-YFP fusions (yellow) transiently expressed in N. benthamiana epidermal cells. RFP localizing to the nucleus and cytoplasm was co-expressed alongside each effector to denote specific organelles (magenta) b) Network rewiring between an effector (hexagon) and its interacting host proteins (circles) to determine the likelihood of random occurrence in the same cellular compartment. c) Distribution of the simulated rate of co-localization in random interaction networks (black points) vs the observed rate in the Arabidopsis Interactome when considering cellular compartment gene ontologies. d) Distribution of the simulated rate of co-localization in random interaction networks (black points) vs the observed rate between SIECs (according to CLSM data) and their Arabidopsis targets (GO CC data).

Next, we utilized a network biology approach to predict the likelihood that an effector would appear in the same subcellular compartment as its interacting protein. GO cellular compartment data were first mapped to the Arabidopsis Interactome (AIMain). Under the hypothesis that 2 interacting host proteins are likely to appear in the same subcellular compartment, we evaluated every pairwise comparison between interacting proteins and their respective localizations to estimate a colocalization rate of 52% within AIMain. We then randomized the interactions within the network and repeated the analysis, with 10,000 simulated networks. In every iteration, we never observed an equivalent level of colocalization, suggesting our analysis was able to calculate the rate at which proteins within a network colocalize. We repeated this analysis with the SIEC-Arabidopsis interactome data by combining the experimental localization data from N. benthamiana to reveal an estimated colocalization rate of ∼50% (Fig. 2b-d).

Comparative interactomics between SIEC target proteins and those of pathogen effectors previously revealed that symbiont effectors more significantly interact with the host hormone network (Osborne et al. 2023). To further understand this observation, we investigated the structural and sequence homology of pathogen vs symbiont effectors to identify specific domains which might influence this activity. We performed reciprocal protein-BLAST and protein domain analysis on the protein sequences of effector candidates from S. indica, Hyaloperonospora arabidopsidis (Hpa), Golovinomyces orontii (Gor), and Pseudomonas syringae (Psy). Direct statistical comparison of bitscores across all respective alignments revealed no significant differences in homology between effectors of any organism vs any other (Fig. S3a), suggesting that effectors are evolutionarily diverse with highly specific functions in line with their respective host ranges. We nonetheless noted that SIECs overwhelmingly possessed unique domains which were not present in the pathogen set, including alpha/beta hydrolases, metalloproteases, and serine proteases (Fig. S3c, d, Table S3). We also found that effectors of all organisms possessed intrinsically disordered domains (IDDs). This raised the question of whether IDDs facilitate convergence by allowing effectors to interact with multiple host targets. We mapped the interactions between Hpa, Gor, and Si effectors and Arabidopsis and classified these based on the presence of an IDD. However, subsequent analysis of this network suggested IDDs do not facilitate convergence (Fig. S3e). In the context of plant-microbe interactions, this suggests an absence of broad-scale homology between effector repertoires of different microbes, aligning with similar observations that have attributed similarity to evolutionary distance (Mukhtar et al. 2011; Weßling et al. 2014). Interestingly however, PCA of the alignment metadata revealed a symbiont specific shift (Fig. S3b). Whether the distinction in SIEC sequence homology can be attributed to its broader host range or its role as a symbiont is unclear. While we did not identify any clear structural distinctions between symbiont and pathogen effectors, these analyses present a starting point for more sophisticated in silico structural analyses.

To further delineate the function of SIECs, we next focused on the function of the host proteins they interact with. We first curated a database of all Gene Ontologies (GO) for genes which we previously described as “Y2H amenable” across the 8 and 12 K spaces and filtered on terms associated with biotic, abiotic, or general (centering on ROS signaling) stress. These stress annotations were then mapped to the Arabidopsis interactome (AIMain; Arabidopsis Interactome Mapping Consortium 2011; Fig. 3a, b). Corroborating previous reports for genes involved in biotic stress, K-shell decomposition analysis of the annotated AIMain network revealed that stress-associated genes exist more centrally within the network and represent hubs that are integral to network structure and are biologically important (Fig. 3c, Fig. S4d). Next, we quantified the rate at which microbial effectors (Si, Hpa, Psy, Gor) target these stress-associated genes; 30, 42, 35, and 15 stress genes were targeted, representing 28%, 34%, 57%, and 33% of respective host targets. Degree preserved network rewiring (10,000 simulations) suggested that in each case, the interaction between effectors and stress targets was not random and was likely a biologically relevant observation (Fig. S4a). Given that GO annotations had been classified as biotic or abiotic, we next asked how central stress-associated proteins are within AIMain and whether different microbes targeted specific groups. Unsurprisingly, stress targets possess a higher K-shell value vs background annotated genes (Fig. 3c). This was especially true for convergent targets that interact with effectors from multiple microbes, although previous annotation of these proteins is likely skewed from our previous definition of these genes as functioning in biotic stress responses (Weßling et al. 2014). Remarkably, however, while all organisms interacted with proteins in both categories, we noted that SIECs interacted with more abiotic proteins than pathogen effectors (Fig. 3e, Fig. S4b).

Figure 3.

S. indica effectors target host stress-responsive proteins. a) Network graph of the Arabidopsis interactome (AI-1Main). Proteins with GO terms associated with stress responses are annotated in green (abiotic stress), yellow (biotic stress) or blue (general stress). b) Bar plot summarizing the frequency of stress annotated proteins within AI-1Main. c) Violin plot describing the K-shell distribution of stress annotated proteins within AI-1Main (gray) vs proteins targeted by effectors of Si, Hpa, Psy and Gor. d) Network graph describing the SIEC-Arabidopsis interaction network. SIECs (hexagons) are coloured according to their function in stress marker assays (left) or hormone marker assays (right). Stress response function of SIEC8, 33, 58, 69, 82, 96, and 125 were confirmed in protoplast validation or phenotyping screen. Interacting Arabidopsis proteins (circles) are colored according to their known gene ontologies in abiotic (aquamarine), biotic (pale yellow), or general (mauve) stress responses. e) Bar plot summarizing the GO stress annotations of effector targets from Si, Hpa, Psy, and Gor irrespective of search space. f) Violin plot describing the K-shell distribution of stress annotated proteins within AI-1Main based on their functional characterization as abiotic, biotic, both, or general. Individual points represent unique or convergent effector targets from Si, Hpa, Psy and Gor.

This prompted us to map our pSTRESS::LUC data onto the interactome to identify putative functional relationships between effectors that affect stress responses and interact with proteins that contribute to those responses. Since hormone signaling also supports stress responses, we further mapped our previously reported pHORMONE::LUC (abscisic acid, auxin, cytokinin, jasmonic acid, salicylic acid) data onto the interactome. All GO terms for every target where “stress” appeared as a character string in their list of gene ontologies (biological process) and combined this with the functional data identified in the pSTRESS marker screen (Fig. 3d). A total of 21 of the 33 SIECs with mapped interactions targeted 55 proteins with GO annotations in stress responses. These included salt, heat, drought, and osmotic stress tolerance, as well as defense responses and regulation of plant innate immunity (Table S2).

We also observed an overlap between the pSTRESS/pHORMONE regulatory activity of SIECs in protoplasts and the known function of their targets. For example, SIEC145 modulated the abiotic stress marker pHSP81.1::LUC (Table S1) and interacted with host proteins involved in heat and water deprivation responses (Fig. 3d). SIEC15, which influenced defense hormone signaling in protoplasts, was found to interact with proteins associated with immunity; PLASTID TRANSCRIPTION FACTOR 1 (PTF1); TEOSINTE BRANCHED1, CYCLOIDEA AND PCF TRANSCRIPTION FACTOR 13 (TCP13) (Wang et al. 2024), and the cell death/hypersensitive response (HR) regulating protein LDS ONE LIKE 1 (LOL1) (Ascencio-Ibáñez et al. 2008). In some instances, an overlap was observed in both the pSTRESS and pHORMONE screens; SIEC17, for instance, showed immunity marker regulation (Table S1) and affected markers for SA and JA signaling in protoplasts. This effector interacted with BINDING PARTNER OF ACD11 1 (BPA1), BPA-LIKE 1 (BPL1), and BPA-LIKE 3 (BPL3), which are known for their role in plant immunity (Li et al. 2019). Similar observations were made for other SIEC-target combinations, including SIECs 16, 26, 54, 56, 69, 81, 82, 108, 126, and 133, representing 33% of all effectors with mapped interactions.

Taken together, the interaction and functional data revealed symbiont effectors target stress-related host proteins. While the targeting of these proteins appeared not to be unique to symbiont effectors, SIECs targeted more frequently host proteins involved in abiotic stress responses. It suggested that S. indica has evolved SIECs to alter host abiotic stress signaling. Based on these analyses, we investigated the capacity for 6 SIECs to confer abiotic stress tolerance in plants, categorized into 3 groups: SIECs with multiple targets linked to different stress responses (group i), SIECs with a single stress-related target (group ii), and SIECs that modulated pSTRESS activity in protoplasts without known targets (group iii) (Fig. S4e).

Symbiont effectors modulate abiotic stress response

To analyze SIEC-induced phenotypic changes, we constitutively expressed them under the 35S promoter in Arabidopsis seedlings. Expression was confirmed by RT-qPCR using SIEC-specific primers and compared with 35S::GFP-expressing control seedlings (Fig. S5, Table S6). We observed enhanced primary root and hypocotyl length in all 6 and in 4 tested 35S::SIEC lines, respectively (Fig. 4a, Table S5). Additionally, all (except for 35S::SIEC91) seedlings produced more lateral roots than GFP-expressing control plants (Fig. 4c, Table S2). As this might be attributed to faster root development, we calculated the average distance between lateral roots. This distance was reduced in 35S::SIEC58 and 35S::SIEC04 plants (Fig. 4d, Table S5), indicating that both effectors altered lateral root density independent of accelerated root development. Taken together, this analysis revealed that SIECs that regulate stress responses in protoplasts or target abiotic stress–related host proteins also appear to promote growth in Arabidopsis seedlings.

After optimizing experimental conditions (Fig. S6a, b, Table S5), we next monitored root development under salt (NaCl) and osmotic (mannitol) stress. NaCl in plant growth media induces mild salt stress response in Arabidopsis seedlings, causing reduction of root growth by restraining metabolic processes (Hongqiao et al. 2021). Mannitol leads to osmotic stress similar to drought in plants by decreasing the osmotic potential of the media, resulting in decreased primary root length and lateral root development (Lipavsk and Vreugdenhil 1996) (Fig. 4e).

By calculating the relative primary root length reduction between control and stress conditions of 35S::SIEC and 35S::GFP plants, we found that 5 of the 6 effectors improved root length under mannitol-induced osmotic stress (SIECs 40, 58, 69, 91, and 93; Fig. 4f, Table S5). Conversely, the relative rate of lateral root induction was suppressed in 35S::SIEC lines (SIECs 4, 40, 58, 93, 69) compared with control seedlings, although this might be due to the existing increased rate of lateral root formation in these genotypes (Fig. 4h, i). Similarly, primary root growth was less impaired in the same genotypes under salt stress (Fig. 4g, Table S5). Interestingly, SIEC04 did not alter primary root length under drought or salt stress (Fig. 4f, g), despite increasing root length and hypocotyl growth under control conditions (Fig. 4a, b). In sum, these root phenotyping analyses revealed effector candidates altered plant root length under salt and simulated drought stress, thereby confirming our predictions based on our protoplast marker assay and network analysis.

To assess whether the protective effect of SIECs against abiotic stress is related to their interacting proteins, we also analyzed the root responses of respective knockout plants to salt and drought stress. For example, SIEC69 interacts with ZFP5 and TCP9 (Osborne et al. 2023) and promoted primary root length under control and abiotic stress conditions (Fig. 4a, f, g) when expressed in Arabidopsis. Conversely, tcp9 and zfp5 mutants produced shorter roots and also exhibited altered responses to salt treatment (Fig. S7). These results indicate that the observed beneficial growth and stress response effects of SIEC69 expression in Arabidopsis might be related to the function of its targets. Furthermore, this analysis has revealed putative functions for ZFP5 and TCP9 in regulating root growth and abiotic stress responses. Other candidates identified from the interactome include DRB4 (SIEC96 target) and ASIL1 (SIEC107 target). While SIEC107 modulated markers for both pSTRESS and pHORMONE (Fig. 3d), the stress responses of asil1 plants were equivalent to control plants. Nonetheless, root length was reduced in this mutant (Fig. S7). Finally, drb4 knockout resulted in shorter roots and improved root length response to both salt and drought stress (Fig. S7), connecting our observation that SIEC96 enhanced pDREB2a::LUC expression (Table S1). Overall, by analyzing the growth and stress responses of selected effector target knockout plants, we confirmed that the functions of SIECs in planta appear to align with those of their interacting plant proteins.

Symbiont effectors modulate plant pathogen resistance

Given that some effectors were able to modulate transcriptional PTI responses (pFRK1, pPHI1) and defense hormone signaling (pWRKY70, pJAZ10), we next tested whether SIECs altered seedling susceptibility to different phytopathogens. Unlike the abiotic stress screen, which was directed by the network analysis, we took an undirected approach for the biotic stress screen. Available SIEC-expressing Arabidopsis genotypes (95 SIEC expressing Arabidopsis lines were available for pathogen infection assays; 86, 50, and 70 were tested for Phytophthora parasitica, Botrytis cinerea, and Rhizoctonia solani infection, respectively) were screened for altered resistance against 3 pathogens with different life strategies and target tissues: the leaf- and root-infecting oomycete Phytophthora parasitica, the leaf-infecting fungus Botrytis cinerea, and the root-infecting fungus Rhizoctonia solani. While both fungi have a necrotrophic lifestyle, the oomycete is a hemi-biotroph that initially colonizes living cells before switching to necrotrophy (Choquer et al. 2007; Meng et al. 2014; Ajayi-Oyetunde and Bradley 2018).

Ten lines possessed altered resistance phenotypes against P. parasitica, with 5 showing increased (SIEC 30, 33, 47, 96, and 99) and 5 reduced infections (SIEC24, 90, 119, 125, and 130; Fig. 5a, Tables S5 and S8). We identified SIEC08- and SIEC19-expressing plants as resistant to B. cinerea infection, while SIEC21-expressing plants were more susceptible (Fig. 5b, Tables S5 and S8). Only SIEC82 conferred resistance to R. solani, while SIEC46 increased susceptibility (Fig. 5c, Tables S5 and S8). Interestingly, no SIEC altered resistance to more than 1 pathogen, suggesting their functions are highly specific and might be explained by the target tissue (roots, leaves) and trophic strategy of each pathogen. A total 87% (13/15) of SIECs that altered pathogen susceptibility in whole plants also altered pFRK1, pPHI1 (PTI markers), pWRKY70 (SA marker), or pJAZ10 (JA marker) activity in protoplasts, demonstrating the predictive power of this assay (Table S1) (Osborne et al. 2023). Four SIECs also had targets with functions related to stress responses; SIEC08 interacted with the defense-related protein KINESIN 7.4 (KIN7.4) (Shimono et al. 2016), SIEC82 with the defense and SA response associated protein XYLOGLUCAN ENDOTRANSGLUCOSYLASE/HYDROLASE 25 (XTH25) (Li et al. 2004; Ascencio-Ibáñez et al. 2008), SIEC96 with the dsRNA binding DOUBLE-STRANDED-RNA-BINDING PROTEIN 4 (DRB4), potentially involved in viral defense (Qu et al. 2008; Jakubiec et al. 2012), and SIEC125 with the oxidative stress–related protein TRYPTOPHAN BIOSYNTHESIS 2 (TRP2) (Zhao et al. 1998). In addition to the abiotic stress phenotyping, these infection assays revealed that SIECs have diverse functions in stress adaptation in plants.

Figure 5.

Pathogen infection response of SIEC expressing plants. a–c) SIEC-expressing plants were analyzed after infection with pathogens. Only significant results are shown in this figure relative to Col-0 infection for Phytophthora parasitica root infection ratios (a), Botrytis cinerea leaf lesion fraction (lesion area relative to the whole leaf area) (b), and Rhizoctonia solani seedling mortality following root infection (c). Col-0 control plants are indicated in blue; significant effects are indicated in yellow. *, P < 0.05; **, P < 0.01; ***, P < 0.001; statistical significance was evaluated by 2-tailed, unpaired t test of at least 2 biological replicates. Sample sizes (n) are indicated above each box.

Discussion

It has long been understood that phytopathogens utilize secreted proteins to suppress host immunity and alter host signaling and metabolism to support their infection and reproduction (Hogenhout et al. 2009). These interactions are generally deleterious to host fitness and negatively impact host responses to abiotic stresses, growth, and yield. In addition to manipulating host signaling, recent studies revealed that beneficial and pathogenic plant symbionts release effectors to directly manipulate plant-associated microbiota. The antimicrobial activity of these “ecological” effectors can shape the composition and thus the effect of microbial communities on plants (Snelders et al. 2022; Mesny et al. 2026). Interestingly, Si and close relatives can activate beneficial effects in host plants in the absence of any other microbe or microbial community (Waller et al. 2005; Deshmukh et al. 2006). Based on this rationale, we hypothesized that effectors from Si might act to enhance, rather than hinder, host fitness.

Si increases plant growth under regular soil nutrition and improves phosphate (Yadav et al. 2010; Kumar et al. 2011), iron (Verma et al. 2022), and sulfate (Narayan et al. 2021) uptake to its host. In addition to altered nutrient supply, Si enhances salt stress tolerance, associated with an improved antioxidant status (Waller et al. 2005; Baltruschat et al. 2008), growth under heat stress through the modulation of JA and ethylene signaling (Chen et al. 2022), and biotic stress tolerance through induced systemic resistance—for example, by modulated gene expression (Molitor et al. 2011). Stimulation of antioxidative metabolism and/or stress-related genes might also be involved in increased tolerance to other stresses (Gill et al. 2016), including chilling (Jiang et al. 2021), pathogen defense (Stein et al. 2008), and drought (Zhang et al. 2018). A natural hypothesis to these diverse benefits is that increased nutrient uptake potentiates systemic host fitness against diverse environmental factors. However, the priming of adaptive genes or modulations to antioxidative metabolism in Si-colonized plants cannot be attributed solely to improved nutrient acquisition. We therefore hypothesized that an essential factor in plant-microbe symbiosis is the utilization of effector proteins to “fine-tune” host signaling, which confers plant benefits.

Consistent with this, we recently identified SIEC141 to interact with thioredoxin-like protein CDSP32 resulting in the activation (monomerization) of NON-EXPRESSOR OF PR1 (NPR1), a key regulator of systemic immunity (Zhang et al. 2024). Here, we show that SIECs, target, and in some cases alter response to biotic and abiotic stresses. Screening effector functions in Arabidopsis leaf protoplasts revealed that 82/106 SIECs (77%) altered either basal or inducible transcriptional activity of luminescence markers for flg22-mediated PTI or heat stress. By analyzing our previously reported SIEC-Arabidopsis interactome in more detail, we found that SIECs interact with more host proteins associated with abiotic stress compared with pathogen effectors and possess unique domain architecture (eg alpha/beta hydrolases, metalloproteases, serine proteases; Table S3), which may represent a symbiont specific fingerprint. To support our interactome, we localized all SIECs that are in the interactome except SIEC54 (32 SIECs) in N. benthamiana and found that their subcellular localizations reflected the localizations of their target proteins (Fig. 2a, Fig. S2). While this approach shows that SIEC possess distinct subcellular localizations, we recognize that transient expression of effectors in N. benthamiana does not reveal the spatiotemporal dynamics of effector secretion during a bona fide Si-Arabidopsis interaction in the root. Given the technical challenge of genetically engineering Si to observe effector secretion and subcellular localization, this method allowed us to evaluate the likelihood that 2 proteins would at least appear in the same organelle and revealed a higher rate of colocalization between effectors and their targets than occurs in random networks.

Based on our protoplast and interactome data, SIECs appear to exert overlapping hormone and stress activities, suggesting that their functions in stress signaling might be coupled to their hormone-modulating activities (Fig. 1g, Fig. S1). ABA, as a major regulatory hormone in abiotic stress responses, for example, integrates extrinsic signals and redirects plant development to cope with drought, high temperature, and salinity (Hong et al. 2013). These functions are often mediated in crosstalk with other hormones, such as auxin, which regulates lateral root branching (Seo and Park 2009). Of the 6 SIECs tested in planta, 5 improved drought stress response and changed ABA marker activity in protoplasts (Fig. 4f, Fig. S4e). Plants expressing these effectors displayed increased lateral root distance and root length after drought exposure (Fig. 4f–i). SIEC04 altered ABA and AUX pathways in protoplast assays (Fig. S4e) and affected lateral root responses to mannitol. Furthermore, SIEC69 and SIEC58 enhanced root length and reduced lateral root density in response to drought stress compared with control plants (Fig. 4a, f, h–i). Interestingly, expression of SIEC69 modulated AUX and JA, in addition to ABA markers, and interacted with multiple stress-related proteins (Fig. 3d, Fig. S4e), suggesting it may have broad regulatory roles in the Si-Arabidopsis symbiosis. This is further reflected in the interactome, where this protein acts as a major hub with 44 putative host interactions (Osborne et al. 2023). SIEC58 affected ABA and SA pathways in plants (Fig. 3d) and interacted with CHLOROPLAST ALDEHYDE REDUCTASE (CHLADR, AT1G54870; Fig. S4e), which is involved in detoxification of reactive carbonyls (Yamauchi et al. 2011). These results indicate that effectors not only increase drought tolerance by promoting root growth but also affect root architecture (e.g. lateral root formation). We further observed that these SIECs reduced the sensitivity of roots to salt stress-mediated growth arrest (Fig. 4g). Here, we focused on understanding seedling responses to osmotic and salt stress. While it is also important to distinguish between sensitivity and survival in these assays, our phenotyping data allude to potential functions of specific effectors in promoting fitness under abiotic stress. Moreover, our data suggest that SIECs and their respective targets converge onto similar physiological and developmental processes to regulate growth and stress responses (Fig. 3d). That both SIEC107 and ASIL1 affect auxin responses (Gao et al. 2009; Osborne et al. 2023) and are nuclear localized (Fig. S2, Table S9) suggests effects on auxin-dependent transcriptional programs that regulate growth rather than stress responses. The observed beneficial functions of Si effectors might depend on the activities of their respective targets and strengthen our rationale that investigating the activity of symbiont effectors can reveal how existing host pathways are fine-tuned to mediate both mutualism and systemic benefit. Combined with our previous study, this also highlights the suitability of using SIECs to detect previously unknown crosstalk points between stress and hormone signaling networks that are critical for plant stress resilience.

In addition to altering abiotic stress response, Si increases host resistance to pathogens (Waller et al. 2005; Stein et al. 2008; Sun et al. 2014). Several SIECs were shown to modulate host immunity in both protoplasts and whole seedlings. Among the SIECs affecting pathogen resistance, the majority altered root colonization by P. parasitica. This might reflect similarities in the infection strategies and the target tissue (roots) of P. parasitica and Si (Jacobs et al. 2011; Meng et al. 2014). Significant effects for R. solani infection were observed less frequently, possibly due to the small sample size and the harsh conditions of seedling mortality assays, making it difficult to detect moderate changes. Nonetheless, we identified 8 SIECs that decreased and 7 that increased disease susceptibility (Fig. 5a–c, Table S8). While this number is lower than the number of effectors which modulated PTI signaling in protoplasts, we believe this might suggest tissue-specific activity of some effectors in roots versus shoots or a potential limitation to screening fungal effectors for altered responses to flg22-mediated activation of immunity. Interestingly, the effectors that altered susceptibility possessed functions in hormone and/or stress signaling in protoplasts or were found to interact with stress-related Arabidopsis proteins (Fig. 3d, Tables S1 and S2). This included altered expression of markers for SA and JA as major plant defense regulating hormones (Hou and Tsuda 2022), in addition to auxin and cytokinin with key functions in growth and development (Brumos et al. 2018; Li et al. 2021). While the latter hormones are not typically linked to pathogen defense, considerable crosstalk between hormone pathways connect auxin and cytokinin to plant immunity, especially in balancing growth and defense (Naseem et al. 2015). It is therefore unsurprising that SIECs which modulate hormone signaling also affect PTI signaling and pathogen susceptibility.

Both pathogens and symbionts activate host immunity, which must be overcome for successful colonization (Jacobs et al. 2011; Yu et al. 2019; Enebe and Erasmus 2023). Si suppresses host immunity and might use SIECs for this purpose. Consistent with this, PIIN_00029 was previously identified as an Si effector-like small-secreted protein with sequence homology to E3 ligases. It was suggested that this protein increases host colonization by suppressing plant defense and ROS generation (Zhang 2014; Plett and Martin 2015). In addition, Si effector PIIN_08944 was reported to interfere with SA homeostasis to suppress PTI, SA-related defense gene expression and ROS burst (Akum et al. 2015). Si effector Dld1 (PIIN_05872) potentially enhances micronutrient accessibility for the microbe and interferes with ROS homeostasis to aid colonization (Nostadt et al. 2020). While these effectors were not present in our top 106 candidates, they illustrate the necessity for Si to suppress host immunity. This is perhaps counterintuitive given that Si colonization can induce systemic resistance within 19 d of colonization (Molitor et al. 2011), and that some SIECs confer disease resistance (Fig. 5a–c). It is, however, explainable in the context of temporal interaction dynamics. Si releases SIECs in a stage-dependent manner, with early effectors delivered during the onset of interaction to inhibit immunity, and late effectors to enhance resistance upon successful accommodation. Characterizing these secreted proteins will be essential to understand their roles in facilitating either colonization, mutualism, or both; especially during the bona fide Si-host interaction.

Just as pathogen effectors have been used to decode plant immunity, we hypothesize that symbiont effectors can be studied to understand symbiosis and identify beneficial host pathways that promote fitness. Here, we not only reveal SIECs with immune suppressing functions, but also those which modulate plant responses to abiotic stress. They are therefore a valuable resource to establish regulatory principles and mechanisms of plant beneficial traits within complex, often cross-talking plant signaling networks. For some Si effectors we could neither identify a host target nor detect any effects in our plant signaling assays. It will be interesting to further our understanding to what extent Si effectors exert unique or additional antimicrobial activities as recently reported for plant pathogen effectors (Mesny et al. 2026). Given its broad host range, we hypothesize that Si effectors offer a sustainable strategy for identifying breeding targets of beneficial signaling pathways in crop species, particularly given the impact of climate change on agricultural productivity (Lobell and Gourdji 2012; Fischer et al. 2021).

Materials and methods

Experimental model and subject details

All Arabidopsis thaliana (Arabidopsis) mutants and transgenic lines are in the Col-0 accession (wild type, WT). A. thaliana genetic material including 35S::SIEC and t-DNA insertion target knockout lines were described previously (Osborne et al. 2023).

Growth conditions for specific experiments are given below.

Protoplast screening

Protoplast screening was performed as described in Osborne et al. (2023), following the procedure in Lehmann et al. (2020). Four- to 5-wk-old Col-0 Arabidopsis plants were used for mesophyll protoplast generation. Following these protocols 10,000 protoplasts were transformed with 3 µg DNA. Each transformation contained 3 different plasmids: 1 µg pUBQ10::GUS (normalization control), 1 µg pSTRESS::LUC (luminescent marker for stress pathway analysis), and 1 µg p35S::SIEC (SIEC expression). Empty vector expressing Nt-HEMAGGLUTININ instead of SIEC expressing ones were used as control. After transformation and overnight incubation protoplasts were treated according to the transformed stress marker. For abiotic stress markers pDREB2a::LUC or pHSP81.1::LUC transformations, heat stress was applied for 1 h at 37 °C or kept in short-day conditions as basal control. Protoplast transformed with PTI markers pPHI1::LUC or pFRK1::LUC were treated with 0.1 mM flg22 or water as basal control. For the landmark screen for each SIEC 2 individual transformations (replicates) per treatment per pMARKER were analyzed. The validation screen used 3 biological replicates with independent transformations for mock and treatment samples. To indicate SIECs with function in signaling pathways mock and treatment ratios were multiplied to show effects in the same direction and reveal promising signaling modulating candidates. SIECs were considered as modulating marker activity if the calculated ratios showed at least 2-fold increase (>2) or decrease (<0.5) relative to the empty vector control.

Interactome and stress-signaling network analysis

The SIEC-Arabidopsis interactome was previously described (Osborne et al. 2023). To identify host targets associated with stress, we mapped all biological process gene ontologies for all proteins within the Arabidopsis Interactome (AI1Main) (accessed 21.08.25). Ontologies were then subcategorized into their association with biotic, abiotic or general stress responses based on the ontology itself, or its parent GO term. General stress responses comprised of terms associated with oxidative stress, DNA damage repair, the ERAD pathway, and reactive oxygen species metabolism, as these terms—while involved in plant stress responses—cannot be described as exclusively biotic or abiotic. Classification of proteins within the interactome allowed the comparison between effector targets of Si, Hpa, Gor, and Psy. To calculate the random likelihood that any organisms' effector repertoire would target stress proteins at random, and degree preserving network rewiring was performed. In short, given X effectors target Y host proteins, Y proteins were sampled randomly from all available proteins with interactions established across our Y2H platform (Y2H amenable). The number of stress associated proteins was then stored (Z), and this analysis was repeated 10,000 times (N). P-values were calculated based on how many simulated networks had more stress associated proteins than was observed experimentally:

$$P\mathrm{value}=\frac{n(Y<Z)}{N}$$

K-shell analysis was performed using the R package “igraph”. Network K-cores were computed using the function “coreness” to evaluate node centrality within AI-Main. Stress annotations were then mapped to respective K-shells to evaluate centrality of those proteins.

Effector domain and homology analysis

To identify organism specific effector domains, proteins sequences of all effectors from Si, Hpa, Gor, and Psy were analyzed using the EBI InterPro tool (https://www.ebi.ac.uk/interpro/) with default settings. Domains were then compared between effector sets to evaluate occurrence of specific domains between organisms. Statistically significant differences between effectors either with or without intrinsically disordered domains were calculated by Wilcoxon test.

For sequence homology analysis, the NCBI BLASTp tool was used (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Reciprocal BLASTp analysis was performed between effector protein sets, using every combination of query and reference (eg Si as reference vs Hpa as query, and vice versa). A PCA was performed comparing bitscores across all comparisons.

Gene expression analysis by real time-qPCR

A pool of seedling roots was frozen in liquid nitrogen and ground into a fine powder. Total RNA was extracted using TRIzol reagent (Invitrogen). RNA was digested by DNase I (Thermo Fisher Scientific) including RiboLock RNase Inhibitor (Thermo Fisher Scientific) to remove genomic DNA. Following manufactures instructions cDNA syntheses was performed using qScrip cDNA Synthesis kit (Quantabio). For expression analysis in qRT-PCR SYBR Green JumpStar Taq ReadyMix (Sigma) was used following standard protocol. To determine absolute and relative expression of SIECs the 2^−ΔCt and 2^−ΔΔCt methods were used (Livak and Schmittgen 2001). Primers used for this analysis are listed in Table S7.

Stress response assays

Arabidopsis seedlings expressing 35S::SIEC and target knockout t-DNA plants were phenotyped for root growth and lateral roots in control and stress conditions. Stress response was analyzed by comparing mock versus treated root length, lateral root number, and density (%) to GFP expressing or Col-0 control plants. Seeds were initially surface sterilized using NaOCl solution and sown on ½ MS media containing 10 µg mL⁻¹ BASTA for SIEC-expressing plants following stratification for 48 h at 4 °C. Selection on BASTA and germination took place for 7 d under short-day conditions in growth cabinets at 22 °C. Seedlings in similar developmental stages were then transferred on ½ MS media containing 50 mM mannitol, 50 mM NaCl, or no treatment. After another 7 d growth under short-day conditions, photos were taken and root length and lateral root number measured using ImageJ. Stress response was quantified by calculating the percentual change of root length/lateral root number/lateral root distance between mock and treatment conditions. Lateral root distance was calculated by dividing the average root length by lateral root number. These values were compared with the percentual change of 35S::GFP or Col-0 plants by dividing 35S::SIEC or target knockout percentage change by the mean of the respective control percentage change to show the percent difference in treatment response. From these values 100 were subtracted to visualize reduced effects with negative values, while positive values indicate increase. Data visualization was performed using BioRender Graphing (BioRender, Toronto, ON, Canada; https://www.biorender.com, accessed [08 2025]; RRID:SCR_018361).

Effector localization in N. benthamiana epidermal cells

Effector subcellular localization was first predicted using the https://localizer.csiro.au/tool (Sperschneider et al. 2017), which classifies compartment targeting to the cytoplasm, nucleus, chloroplast, or mitochondrion based on primary sequence analysis. Next, Agrobacterium tumefaciens–mediated transient transformation was used to express respective YFP-SIEC fusions in N. benthamiana cells. Two days after infiltration, leaf discs were imaged using Leica SP8 confocal laser microscope with excitation at 514 nm and detection between 520 and 545 nm. To compute the random likelihood that effectors and their targets would appear in the same subcellular compartment, gene ontologies for cellular compartment (accessed 02.05.25) were mapped to all effector targets and proteins within the Arabidopsis interactome. To make our analysis more stringent, we only included host proteins which localized within the cytoplasm, nucleus, or endoplasmic reticulum. The colocalization rate was calculated by considering all localizations between proteins within respective interaction networks. For proteins with multiple predicted localizations, all pairwise compartment comparisons were assessed. For example, if protein X is localized exclusively to the nucleus and protein Y to both the nucleus and cytoplasm, their overlap in the nucleus corresponds to a 50% localization rate. Degree-preserving network rewiring was performed to evaluate the likelihood 2 proteins would appear in the same compartment at random https://doi.org/10.5281/zenodo.7749043. Interactions within networks were randomized, while protein localizations were maintained.

Pathogen infection assays

Phytophthora parasitica

35S::SIEC plants were grown vertically on sterile ½ MS plates containing sucrose in a 13-h-light, 11-h-dark cycle at 23 °C for 10 d. Seedlings were then transferred onto sucrose-free ½ MS plates for P. parasitica infection assays. P. parasitica transformant Pp1121, which stably expresses GFP was used for inoculation (Wang et al. 2011). Fresh P. parasitica mycelial plugs (5 mm in diameter) cut from the edge of a culture grown on carrot agar (5% (v/v) cleared carrot juice agar (CA) medium supplemented with 0.002% (w/v) β-sitosterol and 0.01% (w/v) CaCO₃) in the dark at 23 °C for 3 to 5 d were transferred onto the middle of a root of each individual Arabidopsis seedling. The colonization of the P. parasitica hyphae along plant roots was observed using a fluorescence microscope at 48 hpi. The length of the infected root part and the total root length were measured using ImageJ.

Botrytis cinerea

B. cinerea (strain Bc001) was isolated from tomato and cultured on potato dextrose agar (PDA) in a growth chamber at 16 °C for 15 d. Spores were harvested from mycelia with 3 mL potato dextrose broth (PDB) and subsequently were filtered through 4 layers of gauze to remove hyphae. Spore density was adjusted to 5 × 10⁵ spores per milliliter with PDB. Fully expanded leaves of 4-wk-old Arabidopsis seedlings were detached and infected by dropping 5 μL of the spore suspension. Lesions were measured at 48 hpi.

Rhizoctonia solani

R. solani (strain HBZJ-5X) was cultured on PDA in a growth chamber at 23 °C for 7 d. Ten-day-old Arabidopsis seedlings vertically growth on ½ MS (with sucrose) media were transferred onto sucrose-free ½ MS media for subsequent inoculation. Mycelium plugs (diameter 5 mm) were harvested from R. solani PDA culture and put on the middle of individual Arabidopsis seedlings roots. The inoculated seedlings were vertically grown at 23 ℃ with an 11-/13-h-day/-night photoperiod for 9 d.

All graphs involving biotic and abiotic stress phenotyping were created in https://BioRender.com.

Author contributions

Research concept and design: L.R., R.O., S.L., P.F.-B., W.S., and P.S. SIEC-At interactome mapping and data integration: R.O. Protoplast assays: L.R., R.O., S.L., J.R., and E.K. Generation of SIEC lines and stress phenotyping: L.R., R.O., S.L., E.K., J.R., R.E., and Y.Z. Manuscript writing: L.R., R.O., L.R., S.L., P.F.-B., W.S., and P.S.

Supplementary material

Supplementary material is available at Plant Physiology online.

Funding

This work was funded by research grants from Biotechnological and Biological Research Council (BBSRC) / Engineering and Physical Sciences Research Council Grant (EPSRC) of the United Kingdom (BB/M017982/1) and BASF Plant Science Company GmbH.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.