The root-knot nematode effector MiMSP32 targets host 12-oxophytodienoate reductase 2 to regulate plant susceptibility

mechanisms targeting the biosynthesis of biologically active jasmonates to avoid adverse host responses are unknown. (cid:1) Using yeast two-hybrid, in planta co-immunoprecipitation, and mutant analyses, we identi-ﬁed 12-oxophytodienoate reductase 2 (OPR2) as an important host target of the stylet-secreted effector MiMSP32 of the root-knot nematode Meloidogyne incognita . MiMSP32 has no informative sequence similarities with other functionally annotated genes but was selected for the discovery of novel effector mechanisms based on evidence of positive, diversifying selection. (cid:1) OPR2 catalyzes the conversion of a derivative of 12-oxophytodienoate to jasmonic acid (JA) and operates parallel to 12-oxophytodienoate reductase 3 (OPR3), which controls the main pathway in the biosynthesis of jasmonates. We show that MiMSP32 targets OPR2 to promote parasitism of M. incognita in host plants independent of OPR3-mediated JA biosynthesis. (cid:1) Artiﬁcially manipulating the conversion of the 12-oxophytodienoate by OPRs increases susceptibility to multiple unrelated plant invaders. Our study is the ﬁrst to shed light on a novel effector mechanism targeting this process to regulate the susceptibility of host plants.


Introduction
The tropical root-knot nematode Meloidogyne incognita ranks high on the list of most invasive plant disease-causing agents world-wide (Bebber et al., 2014).This polyphagous plant parasite feeds on both monocots and dicots, including not only many important food crops, but also the model plant Arabidopsis (Sijmons et al., 1991).After penetrating the root epidermis, infective second-stage juveniles (J2) of M. incognita migrate through the cortex and the root apical meristem to enter the host vascular cylinder, where they induce several permanent feeding cells (Wyss & Grundler, 1992;Williamson & Gleason, 2003;Abad & Williamson, 2010).These so-called giant cells are multinucleate and hypertrophied plant cells, which can reach up to 100 times the size of a normal vascular parenchyma cell (Kyndt et al., 2013).Plant-feeding juveniles of M. incognita selectively extract assimilates from the cytoplasm of giant cells with their protrusible oral stylet.Over a period of several weeks, while intermittently feeding on different giant cells, the juveniles of M. incognita develop through successive molts into mature females, which produce offspring as egg masses on the root surface (Gheysen & Mitchum, 2019).
The biosynthesis of jasmonates is thought to be important for orchestrating damage-induced host responses to root-knot nematodes in plants (Gheysen & Mitchum, 2019).The first phase of jasmonate biosynthesis occurs in plastids, where lipidderived linoleic and linolenic acids are converted into 12oxophytodienoic acid (12-OPDA) by 13-lipoxygenases (LOXs), allene oxide synthases (AOS), and allene oxide cyclases.It has been shown that the mutant tomato spr2, which is defective in the conversion of linoleic acid to linolenic acid (Li et al., 2003), is much more susceptible to M. incognita than wild-type plants (Fan et al., 2015;Wang et al., 2019).
Likewise, disrupting the 13-lipoxygenase genes LOX3 and LOX4 in Arabidopsis alters host susceptibility to Meloidogyne javanica infections (Ozalvo et al., 2014).Further downstream, the Arabidopsis mutant, dde2 mutant lacking an allene oxide synthase required for jasmonate biosynthesis, is more susceptible to Meloidogyne hapla (Gleason et al., 2016).The hebiba rice mutant, which is defective in an allene oxide cyclase (Riemann et al., 2013), also shows increased susceptibility to Meloidogyne graminicola (Nahar et al., 2011).After its transport to peroxisomes, 12-oxophytodienoate reductase 3 (OPR3) catalyzes the conversion of 12-OPDA to jasmonic acid (JA), which is then released into the cytosol.Surprisingly, the Arabidopsis 12-oxophytodienoate reductase 3 mutant opr3 shows no difference in susceptibility to M. hapla as compared to wild-type Arabidopsis plants, suggesting that OPDAbut not JAmight regulate host susceptibility to root-knot nematodes (Gleason et al., 2016).However, more recent studies suggest that 12-OPDA derivatives can also be reduced to JA via an alternative pathway, which takes place in the cytosol, and which is catalyzed by 12-oxophytodienoate reductase 2 (OPR2; Chini et al., 2018).Cytosolic JA can subsequently be converted into biologically active JA-isoleucine (JA-Ile) in a reaction catalyzed by jasmonyl isoleucine synthetase (JAR1).In line with other jasmonate biosynthesis mutants, jar1-1 knockout mutant rice plants display increased susceptibility to M. graminicola (Singh et al., 2020).Altogether, these findings illustrate that interfering in the biosynthesis of jasmonates can enhance the susceptibility of host plants to root-knot nematodes.
Meloidogyne incognita secretes a plethora of effectors to manipulate host development and defenses (Mitchum et al., 2013).Most of these effectors are produced in three pharyngeal glands, which are large single-celled specialized secretory glands that are connected via the pharyngeal lumen to the basal opening of the oral stylet.By salivation through the oral stylet, M. incognita delivers effectors into the apoplast or cytoplasm of host cells, where they interact with host targets to promote parasitism (Hussey, 1989;Mejias et al., 2019).Several studies suggest that rootknot nematodes utilize effectors to target damage-induced jasmonate biosynthesis and/or signaling.For instance, ectopic expression of the M. incognita effectors MiISE5 and MiSE6 in Arabidopsis changes the transcriptional regulation of multiple JA signaling genes (Shi et al., 2018a,b).Similarly, the expression of the M. javanica effector Mj2GO2 in transgenic Arabidopsis affects the regulation of genes associated with JA-mediated plant responses, as well as constitutive levels of JA-Ile (Song et al., 2021).However, at present, the host targets and mode of action of root-knot nematode effectors interfering in the biosynthesis and signaling of jasmonates are unknown.
Currently, the number of candidate nematode effectors generated by whole-genome sequencing exceeds by far the capacity to establish which of the candidates are bona fide effectors (Nguyen et al., 2018;Shukla et al., 2018).Prior work with other plant parasitic nematodes suggests that effector families involved in the activation and suppression of host defense are prone to mutations and recombination (and subsequent diversifying selection; Carpentier et al., 2012;Pokhare et al., 2020).
For this study, we used evidence of positive diversifying selection to prioritize the effector candidate MiMSP32 of M. incognita for the further functional characterization in planta (Baskaran et al., 2017).We investigated the importance of MiMSP32 for the parasitism of M. incognita using host-induced gene silencing and effector overexpression in transgenic tomato plants.To identify the host targets of MiMSP32 in tomato and Arabidopsis, we screened a yeast two-hybrid (Y2H) library of nematode-infected tomato roots and confirmed candidate interactors by in planta co-immunoprecipitation.After assessing the susceptibility of Arabidopsis knockout mutants of the candidate host interactors of MiMPS32, we could conclude that this effector targets 12-oxophytodienoate reductase 2 (OPR2) to regulate host susceptibility.AtOPR2 has been shown to catalyze the conversion of 4,5-didehydro-jasmonate (4,5-ddh-JA), a derivative of 12-OPDA, to JA (Chini et al., 2018).Taken together, our data provide the first evidence of a nematode effector regulating host susceptibility by targeting an enzyme in the biosynthesis of jasmonates in plants.

Evidence of positive, diversifying selection in MiMSPs
To first test which of the 27 previously identified pioneer MiMSP genes (Abad et al., 2008) have homologs in current genome assemblies of M. incognita (Kofoid & White, 1919) and other root-knot nematodes, we used their predicted mRNA sequence to query the Wormbase Parasite cDNA database with the BLASTN algorithm (at parasite.wormbase.org,accessed in May 2020).Hits with a BLASTN score below 100 were disregarded as being false positives.The genome assemblies of the root-knot nematodes referred to in this analysis are available at NCBI under the following accession codes: PRJEB8714 dataset for Meloidogyne arenaria, M. incognita, and M. javanica (Blanc-Mathieu et al., 2017); the PRJNA340324 dataset for M. arenaria, Meloidogyne enterolobii, Meloidogyne floridensis, M. incognita, and M. javanica (Szitenberg et al., 2017); the PRJNA438575 dataset for M. arenaria (Sato et al., 2018); the PRJEB6016 dataset version nMf.1.0from nematodes.orgfor M. floridensis; and the PRJNA29083 dataset for M. hapla (Opperman et al., 2008).
For MiMSPs with three or more hits in the genome assembly of M. incognita PRJEB8714, we generated a multiple sequence alignment of the coding sequences of the paralogous genes using CLUSTALW2.The resulting alignment was used to construct a cluster tree based on the Tamura-Nei genetic distance model.MiMSPs with three or more paralogs in the PRJEB8714 genome assembly of M. incognita were also tested for signatures of positive selection using the CODEML algorithm of PAML 4.7 (phylogenetic analysis by maximum likelihood; Yang, 1997Yang, , 2007;;Yang & Bielawski, 2000) within EASYCODEML v.1.21(Gao et al., 2019).As input files, we used the aligned coding sequences of the MiMSPs paralogs, and an accompanying neighbor-joining tree generated under the Tamura-Nei genetic distance model.EASYCODEML was run under the preset mode for nested models using default settings to compare site models M7 vs M8.Significance was calculated using log-likelihood ratio tests for the full model comparisons with an alpha of 0.05.

Nematode infection assays
Nematode preparation and sterilization Meloidogyne incognita eggs (strain 'Morelos' from INRA, Sophia Antipolis, France) were obtained from infected tomato plants (Solanum lycopersicum L. cv MoneyMaker) as described previously (Warmerdam et al., 2018).In short, roots of infected tomato plants were rinsed in water to remove sand particles 10 wk after inoculation and eggs were extracted by incubation in 0.05% (v/v) bleach for 3 min followed by sieving (Hussey & Barker, 1973).Eggs were disinfected using 0.02% sodium azide (NaN₃) for 20 min and washed thoroughly with tap water.Hatching took place at room temperature in the dark on a 25 μm hatching sieve with 1.5 mg ml −1 gentamycin and 0.05 mg ml −1 nystatin.After 4 d, J2s were collected by separation on a 70% sucrose column and sterilized by incubation for 10 min in 0.002% (v/v) Triton X-100, 0.004% (w/v) NaN₃, and 0.004% (w/v) mercury chloride.After surface sterilization, nematodes were rinsed in sterile tap water three times before use and transferred to a 0.7% (w/v) Gelrite solution (Duchefa Biochemie, Haarlem, the Netherlands).
Infection assays on tomato For in vitro infection assays, tomato seeds (S. lycopersicum L. cv MoneyMaker, or the same cultivar with MiMSP32 -sp -derived plasmids; Supporting Information Methods S1, S2) were first incubated for 3 d in tap water at 4°C in the dark.Thereafter, the seeds were briefly washed with 70% ethanol and 2.5% (v/v) bleach and incubated three times for 10 min in sterile tap water.Batches of seeds were sowed on square plates of 10 cm containing ½ MS20 medium (2.35 g l −1 Murashige & Skoog (MS) with vitamins (Duchefa Biochemie), 20 g l −1 sucrose, pH 6.4, 7.0 g l −1 Gelrite (Duchefa Biochemie)).After incubating for 4 d at 24°C in 16 h : 8 h, light : dark, germinated plants were transferred to fresh square plates containing ½ MS20 medium.The transferred plants were allowed to grow for an additional 6 d under the same conditions, after which they were inoculated with 120 infective M. incognita J2s per plant.Plants were kept horizontally in the dark for 2 d at 24°C, after which they were placed diagonally in 16 h : 8 h, light : dark conditions again.To reduce the direct exposure of the roots to light, the bottom 8 cm of the plates was covered with paper sleeves.Seven days postinoculation (dpi), the number of galls formed in the roots was counted by visually inspecting the roots with a dissection microscope.
Infection assays on Arabidopsis Arabidopsis seeds (Arabidopsis thaliana (L.) Heynh.accession Col-0 (CS60000), or mutants mentioned later) were vapor-sterilized for 4 h in 0.7 M sodium hypochlorite and 1% hydrogen chloride before sowing in plates containing MS20 medium (4.7 g l −1 MS with vitamins (Duchefa Biochemie), 20 g l −1 sucrose, pH 6.4, 0.7% Gelrite (Duchefa Biochemie)).Seeds were stratified for at least 3 d at 4°C.Arabidopsis seeds were stimulated to germinate at 24°C under a 16 h : 8 h, light : dark regime.After 6 d, individual plants were transferred to fresh six-well culture plates and allowed to grow and settle for an additional 7 d, after which they were inoculated with 180 infective J2s of M. incognita per plant.A slightly higher amount of J2s per plant was used to ensure a comparable infection pressure, because of the susceptibility differences in between tomato and Arabidopsis.Nematode-containing 6-well plates were incubated at 24°C under dark conditions for the duration of the bioassay.
Data collection and statistical analysis The number of root tips was counted for each plant at the time of inoculation.Data were collected in at least three independent experiments and pooled for statistical analysis and visualization.The number of nematodes inside the roots was counted at 7 dpi by means of acid fuchsin staining on the whole root system.Therefore, clean roots were incubated for 5 min in 2.5% household bleach followed by rinsing for 10 min in tap water.Next, roots were incubated in fuchsinstaining solution (0.2 M acid fuchsin and 0.8% glacial acetic acid in tap water) for 30 s in a microwave oven at maximum power.Finally, roots were transferred to 40% glycerol and nematodes could be counted visually using a dissection microscope.
Counting data from tomato or Arabidopsis were collected in at least two independent experiments with n ≥ 16 and combined for statistical analysis and visualization.To be able to pool the data of independent experiments, we normalized the counts for batch effects (Fig. S1d).To this end, the actual counts were corrected based on the average of counts on wild-type plants from a single independent experiment against the average of counts on wild-type plants over all experiments using where T norm is the normalized number of nematodes or nematode-induced galls, i is the individual experiment, and T is the averaged number.
The batch-corrected data for nematode numbers, gall numbers, and root tip numbers were analyzed in R v.3.6.1.x64.Extreme outliers in the dataset outside the interquartile range of 1.5 were removed (Vinutha et al., 2018).Normality was checked using qq-plots from the GGPUBR package, and statistical comparison was done using either an ANOVA analysis with post hoc Tukey's HSD or Student's t-tests, depending on the number of treatments or plant lines.Data visualization was done using the GGPLOT2 package (Wickham, 2016).

Yeast two-hybrid screen
A yeast two-hybrid screen was performed as a custom service using full-length MiMSP32 minus the signal peptide as bait and a cDNA library of infected tomato roots (S. lycopersicum L. cv M82) as prey (Hybrigenics Services, Evry, France).This cDNA library was constructed with mRNA from an equal mixture of tomato root tissue infected with Ralstonia solanacearum (4-wkold roots on soil, infected with strain GMI1000 or GMI1694 (hrp−), which were harvested at 12 h postinfection), root tissue infected with M. incognita (3-wk old roots (on soil or in vitro)) infected with M. incognita (galls dissected at 4-5 dpi and 7-8 dpi and whole infected roots 7-8 dpi), and uninfected roots (1-to 3-wk-old plants grown on soil or in vitro), and is available as 'Tomato Roots Infected_RP1' (Hybrigenics Services).Hybrigenics Services performed their optimized ULTImate Y2H ™ technique and provided information to separate artifacts from specific interactions by the global predicted biological score (PBS), which is based on a statistical model (Rain et al., 2001).Alignments of individual clone sequences described with both a forward and a reverse sequence were made in GENEIOUS V.8.1.9(Biomatters, Auckland, New Zealand).

Co-immunoprecipitation assays
Transient expression and total protein extraction were done according to Methods S3.Protein extract was first mixed with 50 μl rabbit-IgG agarose (Sigma-Aldrich) and incubated with 50 μl anti-HA microbeads.Co-immunoprecipitation was done with separation columns from μMACS Epitope Tag Protein Isolation Kit (Miltenyi Biotec, Bergisch Gladbach, Germany) according to manufacturer's instructions.Soluble fractions were analyzed by SDS-PAGE on a 12% Bis-Tris gel (Invitrogen), and proteins were subsequently transferred to a PVDF membrane for Western blotting (Thermo Fisher, Waltham, MA, USA).Protein bands were visualized either with a Roche Anti-HA-Peroxidase high affinity from rat IgG1 (Scientific Laboratory Supplies, Wilford, UK) or with a primary goat anti-MYC antibody (Abcam, Cambridge, UK) and a horseradish peroxidaseconjugated secondary antibody of donkey anti-goat (Jackson ImmunoResearch, Ely, UK).SuperSignal West 1 : 1 Femto-Dura Substrate (Thermo Fisher) was used to detect horseradish peroxidase-conjugated antibodies in the G:BOX Chemi System (Syngene, Bangalore, India).To confirm equal protein loading, membranes were also stained with Coomassie brilliant blue.

Subcellular localization using confocal microscopy
To assess the subcellular localization of MiMSP32 -sp without its signal peptide for secretion, and with and without the putative interacting host protein fragments in plant cells, the proteins were transiently expressed by agroinfiltration in Nicotiana benthamiana epidermal cells (Methods S3).Agroinfiltrated leaves overexpressing either HA4_GFP_MiMSP32 -sp or MiMSP32 -sp _GFP_HA4 in combination with the interacting domain of possible interactors (SID) in MYC4_mCh_SID or SID_mCh_MYC4 configuration, or with free mCherry, were collected for microscopic observations with a Zeiss LSM 510 Confocal Microscope (Carl Zeiss).For imaging of constructs carrying a GFP tag, the 488 nm line of an argon-ion laser was used for excitation and GFP emission was selected through a bandpass filter of 505-530 nm for detection.For constructs carrying a mCherry tag, imaging was done using a 543 nm HeNe laser for excitation and mCherry emission was selected by a 600-650 nm band-pass filter.Chlorophyll emission was detected using a 650 nm long-pass filter.Images were equally enhanced in brightness for publication in print.
To quantify differences in subcellular localization, the ratios in fluorescence intensity between cytoplasm and nucleus were calculated in IMAGEJ.Normality of the data was checked using qq-plots from the GGPUBR package in R v.3.6.1 x64, and a oneway ANOVA with post hoc Tukey's HSD was used to determine statistical differences.

Evidence of positive selection in putative M. incognita effectors
Twelve years after the publication of the first M. incognita genome sequence (Abad et al., 2008), we revisited the status of 27 previously identified 'pioneer genes' isolated from gland-specific cDNA libraries by Huang et al. (2003) using the latest root-knot nematode genome assemblies available (Blanc-Mathieu et al., 2017).Querying the WormBase ParaSite nucleotide sequence database with these 27 genes, hereafter named MiMSPs, yielded significantly matching sequences from one or more rootknot nematode species (Fig. 1a).Surprisingly, for only 19 MiMSPs originally identified in M. incognita, we found at least one significantly matching gene prediction in the latest genome assembly of M. incognita (Meloidogyne_incognita_V3; Fig. 1b).Furthermore, for eight MiMSPs, we found an overlap in the matching gene predictions, because of high similarity in the MiMSP query sequences.For six MiMSPs, we identified three or more paralogs in the genome of M. incognita (Fig. 1b,c), which could be further tested for evidence of positive selection using CODEML algorithm of PAML (phylogenetic analysis by maximum likelihood; Yang, 1997Yang, , 2007;;Yang & Bielawski, 2000).The log-likelihood ratio tests of the nonsynonymous over synonymous mutation rates for three MiMSP (i.e.MiMSP17, MiMSP31, and MiMSP32) significantly favored codon substitution model M8, which does allow for positive selection, over M7, which does not allow for positive selection (Table S2; Fig. 1c).We therefore concluded that these three MiMSPs represent As the estimated ω under model M8 for positive selection for MiMSP31 and MiMSP32 proved to be much larger than for MiMSP17 (Fig. 1c), we initially focused our analyses on these two genes.Prior work has shown that MiMSP31 and MiMSP32 encode putative secretory proteins specifically localized in the dorsal pharyngeal gland of parasitic J2s and later stages in M. incognita (Huang et al., 2003).However, closer inspection revealed that MiMSP31 has a high sequence similarity with 150 residues at the amino terminus of MiMSP32 (Fig. S4), and both are likely splice variants of the same transcript.We therefore chose to focus further on the functional characterization of MiMSP32.An initial protein sequence analyses of MiMSP32 revealed no specific motifs, structures, or domains, confirming its classification as 'pioneering gene'.However, we noted that the three paralogous genes of MiMSP32 in the M. incognita genome, called Minc3s00381g11350, Minc3s00827g17824, and Minc3s00593g14814, are classified as

MiMSP32 functions as a bona fide effector
To investigate whether MiMSP32 promotes parasitism by M. incognita in tomato, we first used host-induced gene silencing with a hairpin construct targeting nucleotides 313-522 of the MiMSP32 transcript (Fig. 2a).At 7 dpi, we observed a significantly lower number of nematode-induced galls for two independently transformed tomato lines (Methods S1, S2, S4; Fig. S1a) overexpressing the pSMD:MiMSP32_hp hairpin construct compared with wild-type tomato plants (Fig. 2b).
Next, we tested whether ectopic overexpression of MiMSP32 alters the susceptibility of tomato plants to M. incognita.We challenged seedlings of two independently transformed tomato lines overexpressing MiMSP32 without its native signal peptide with infective J2 in an in vitro bioassay (Methods S1, S2, S4; Fig. S1b,c).At 7 dpi, we observed a significant increase in the number of galls in roots of the two tomato lines overexpressing MiMSP32 as compared to wild-type tomato plants (Fig. 2c).Based on the plant phenotypes of our gene silencing and overexpression constructs, we concluded that MiMSP32 functions as bona fide effector for M. incognita promoting nematode parasitism in roots of tomato plants.
Next, we analyzed predicted physiochemical properties of the six interacting protein fragments that could result in nonspecific binding in the Y2H screen (Table S6).We also determined the positions of the interacting fragments relative to predicted subcellular localization signals and conserved protein motifs and domains in the matching full-length proteins (Marchler-Bauer et al., 2016).For SlOPR2 (Fig. 3a), the minimal interacting fragment contains nine out of 14 active sites, including substrate binding sites and flavin mononucleotide binding sites.For SlSPT6 (Fig. S5), the interacting fragment contains one of the four nuclear localization sites, while the other three are located outside the interacting fragment.However, we could not find any structural properties in the proteins to help us prioritize the probable interacting fragments for further studies in planta.
We therefore proceeded to independently confirm interactions of MiMSP32 with all six probable host targets by co-expressing affinity-tagged constructs of the selected interaction domains of the interactors with pBIN:MiMSP32 -sp _GFP_HA4 (Methods S5) in N. benthamiana Domin leaves.Co-immunoprecipitation with anti-HA magnetic beads showed that five out of six putative interactor fragments (Methods S6) also specifically bind to MiMSP32 -sp _GFP_HA4 in plant cells (e.g.SlOPR2 in Figs 3b,  S6).In addition, we used the same constructs to observe the induction of a possible shift in subcellular localization of these interacting fragments in planta by co-expression with MiMSP32 -sp (Fig. S7a,b).We reasoned that binding to a probable interactor might shift the subcellular localization of MiMSP32 toward the localization of its interactor, or vice versa.
To test this, we transiently co-expressed GFP-tagged MiMSP32 -sp with the putative interactor domain fused with mCherry in N. benthamiana (e.g.SlOPR2 SID in Fig. 3c).When expressed alone, MiMSP32 has a similar fluorescence intensity in nucleus and cytoplasm.Interestingly, we noticed a significant shift in subcellular localization of MiMSP32 -sp _GFP_HA4 toward the cytoplasm when it was co-expressed with three out of six putative interactor fragments (e.g.SlOPR2 SID in Figs 3d,  S7c).The shift in subcellular localization of SlALA1 SID was statistically not significant, yet visually evident.Taking the data of the co-immunoprecipitation experiments and shifts in the subcellular localization together, we conclude that MiMSP32 interacts in planta with fragments of the tomato proteins phospholipidtransporting ATPase 1-like (SlALA1), leucine aminopeptidase (SlLAPA2), oxophytodienoate reductase 2 (SlOPR2), and tripeptidyl-peptidase 2 (SlTPPII).

The MiMSP32 target AtOPR2 regulates the susceptibility of Arabidopsis to M. incognita
To test whether the Arabidopsis orthologs of the possible host targets of MiMSP32 play a role in plant susceptibility to rootknot nematodes, we performed nematode bioassays using T-DNA knockout mutants of AtALA1 (ala1-1 and ala1-2), AtOPR2 (opr2-1), and AtTPPII (tppII-1).The stress-inducible acidic leucine aminopeptidase 2 (SlLAPA2) is thus far only found in a subset of the Solanaceae (Scranton et al., 2012), and therefore, no T-DNA knockout mutant line of a close ortholog of this gene is available for Arabidopsis.In all four knockout mutants, the T-DNA inserts are in gene exons and proved to be homozygous for the insertion (Figs 4a, S2).In multiple independently repeated experiments, only opr2-1 mutant plants harbored a significantly higher number of nematodes per plant at 7 dpi than wild-type Col-0 plants (Fig. 4b).Taken together, our data of the Arabidopsis mutants suggest that M. incognita could enhance host susceptibility by targeting OPR2, and we therefore focused our study further on OPR2.
We first assessed whether MiMSP32 also physically interacts with full-length Arabidopsis AtOPR2 by transiently coexpressing affinity-tagged constructs encoding AtOPR2 and MiMSP32 without its native signal peptide (MiMSP32 -sp ) in N. benthamiana leaves via agroinfiltration.We used MiMSP32 -sp fused to GFP-and 4xHA-tags on either N-or C-terminus of the protein and AtOPR2 carrying C-terminal mCherry-and 4xMYC-tags.Indeed, MiMSP32 -sp pulled down AtOPR2 in a co-immunoprecipitation assay using anti-HA magnetic beads (Fig. 4c).Our data thus showed that full-length AtOPR2 also interacts with MiMSP32 -sp .

OPR2 directs the root transcriptome toward abiotic and biotic stress responses
To resolve gene networks underlying the enhanced susceptibility of the Arabidopsis opr2-1 mutant to M. incognita, we analyzed differential expression patterns in nematode-infected roots of mutant opr2-1 and wild-type Col-0 plants at 0, 1, 4, and 7 dpi using RNA sequencing.In total, we analyzed 1624 950 228 reads, of which 95.3% mapped to the TAIR10 A. thaliana genome (Lamesch et al., 2011;Table S7; Fig. S8; Methods S7).To identify genes specifically affected by the mutation on separate days after inoculation or in interaction with nematode infection, we used a statistical interaction model for plant genotype and M. incognita infection on 1, 4, and 7 dpi (Fig. 5a).We found 13 genes which were differentially regulated in roots of the opr2-1 mutant lines vs wild-type Col-0 upon infection by M. incognita (Fig. 5b; Table S8).As expected, the most affected gene in the opr2-1 mutant at all three time points in both infected and noninfected roots was AtOPR2 (Fig. 5c).In addition, 12 other genes showed deviating expression patterns in the opr2-1 mutant when compared to the wild-type Col-0 line (Table S9).Six of these differentially regulated genes are associated with the response to either biotic or abiotic stress, and one has shown to be directly linked with JA responses.Importantly, we also analyzed the relative expression of all genes included in KEGG module M00113 representing the JA biosynthesis pathway in Arabidopsis (Fig. S9).Here, we observed a minor downregulation of the OPDA reductase AtOPR1 in nematode-infected roots of the opr2-1 mutant compared with wild-type plants.In conclusion, the impact of OPR2 on the transcriptome of nematode-infected Arabidopsis roots points at alterations in responses to abiotic and biotic stress.

OPR2 and OPR3 differently affect plant susceptibility to M. incognita
AtOPR2 is thought to regulate an AtOPR3-independent pathway in the biosynthesis JA (Fig. 6a; Chini et al., 2018).To test whether the two parallel biosynthesis pathways redundantly affect the susceptibility of Arabidopsis to M. incognita, we performed nematode bioassays using opr2-1, opr3-3, and op2-1/opr3-3 mutant lines.We first observed that both single-mutant lines, opr2-1 and opr3-3, were equally more susceptible to M. incognita than wild-type Arabidopsis plants (Fig. 6b).However, the  double-mutant opr2-1/opr3-3 was significantly more susceptible than both single mutants, which implies that opr2-1 and opr3-3 additively affect the susceptibility of Arabidopsis to M. incognita.
It has been shown that interfering with OPR3 in the biosynthesis of JA can induce enhanced secondary root formation (Li et al., 2019), which may affect plant susceptibility to root-knot nematodes.Indeed, we also observed a significantly increased root branching phenotype for the single-mutant opr3-3 at the time of inoculation, but surprisingly not for the opr2-1 mutant (Fig. 6c).Knocking out OPR2 on top of OPR3 in the doublemutant opr2-1/opr3-3 did not result in a further increase the number of root tips either.This indicates that the parallel pathways independently regulated by OPR2 and OPR3 have different physiological outputs, possibly involving other biologically active compounds.
Targeting of host OPR2 by MiMSP32 may alter 12-OPDA but not JA levels Next, we tested whether targeting of host OPR2 by the effector MiMSP32 alters the conversion of 12-OPDA into JA in hypersusceptible tomato plants.Hereto, we measured the concentrations of 12-OPDA and JA in the two independent tomato lines constitutively overexpressing MiMSP32 without its native signal peptide for secretion (Methods S8).Notably, the baseline level of 12-OPDA in wild-type tomato plants was extremely low (i.e.below 0.4 fmol per mg fresh weight).Nonetheless, we observed an increase in the concentration of 12-OPDA in the individual transgenic tomato lines overexpressing MiMSP32, although this was not sufficiently supported in the ANOVA with post hoc analysis (Fig. 6d).However, by aggregating the data of the two independent lines, we found a significant effect of the MiMSP32 -sp -insertion on the 12-OPDA-concentration in tomato in a t-test (Pvalue = 0.0039).By contrast, the JA(-Ile) concentrations in the individual transgenic tomato lines overexpressing MiMSP32, or the aggregated data of the two lines, were not significantly different from the wild-type tomato plants.We therefore concluded that MiMSP32 may target OPR2 in host plants to alter the conversion of 12-OPDA without significantly affecting JA-(Ile) levels.

Discussion
Recent developments in sequencing technologies have accelerated the identification of candidate effectors in genomes of plant parasitic nematodes, many of which have no informative homology with functionally annotated genes.The speed at which such novel candidate effectors are currently being identified exceeds the capacity to establish whether they function as bona fide nematode effectors in planta.Here, we show that evidence of positive, diversifying selection can be used as an additional criterion to prioritize effector candidates for the discovery of novel molecular mechanisms underlying nematode-plant interactions.We demonstrate that the positively selected effector MiMSP32, previously identified as a dorsal pharyngeal gland-specific pioneer gene (named 19F07 in Huang As host-induced gene silencing of MiMSP32 in M. incognita requires the uptake of dsRNA by feeding, this effector operates most likely in later stages of parasitism after feeding on giant cells has begun.This is in agreement with the temporal expression of MiMSP32 in the dorsal pharyngeal gland of M. incognita, which extends from second-stage parasitic juveniles to later parasitic stages (Huang et al., 2003).Our findings also support the hypothesis that dorsal pharyngeal gland-specific effectors of M. incognita involved in later stages of parasitism are more likely to be divergent as a consequence of positive selection by different host defense systems (Da Rocha et al., 2021).Indeed, we show that MiMSP32 enhances the susceptibility of host plants to M. incognita by targeting an enzyme in the biosynthesis of jasmonates, which play a role in regulating host defenses.However, P la n t l i n e d pi 1  2 4 6 0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6 0 2 4  more research is needed to determine whether the sequence diversity observed in the cluster of paralogous genes to which MiMSP32 belongs mirrors adaptations in host targets, such as OPR2, involved in defense.
Our data from the Y2H screen, co-immunoprecipitation assays, and shifts in subcellular location suggest that MiMSP32 can interact with at least four unrelated host proteins.This could point at possible effector promiscuity, which has been observed for effectors of other plant pathogens.For instance, 32% of the bacterial type III effectors have multiple host targets with a similar molecular function, while 36% have multiple host targets with different molecular functions (Khan et al., 2018).Bacterial type III effectors can target multiple unrelated host proteins located in different subcellular compartments allowing for spatial diversification in effector functions.Alternatively, promiscuous effectors also functionally diversify in time by targeting different host proteins at successive stages of the infection process (Thordal-Christensen et al., 2018).Effector promiscuity can be based on the presence of multiple protein domains, providing different protein-protein interaction surfaces (Thordal-Christensen et al., 2018).For example, the HopF1 effector in Pseudomonas syringae harbors two separately acting subdomains, shaped like the 'head' and the 'stalk' of a mushroom (Singer et al., 2004).Smaller structural motifs within a single domain, such as the WY-domain fold, also facilitate binding of effectors to different host proteins (Franceschetti et al., 2017).Within the protein sequence of MiMSP32, we have not found specific structural features, such as unusual charge distribution or a multiple-domain architecture, that could explain its binding to several unrelated host proteins.Nevertheless, evidence from earlier studies on nematode effectors indicates that effector promiscuity may also be a common theme in nematode-plant interactions.For instance, the effectors MiEFF1 of M. incognita (Truong et al., 2021) and MgMO237 of M. graminicola (Chen et al., 2018) interact with different unrelated host proteins.Likewise, the effector Hs25A01 from the beet cyst nematode Heterodera schachtii interacts with an Arabidopsis F-box-containing protein, a chalcone synthase, and the translation initiation factor eIF-2 b subunit (eIF-2bs; Pogorelko et al., 2016).For the highly polyphagous root-knot nematodes, effector promiscuity could be an efficient strategy enabling parasitism on a wide range of unrelated host plant species.
Our work establishes OPR2 as an important effector target of M. incognita for regulating the susceptibility of host plants.Upon nematode infection, OPR2 directs the root transcriptome toward biotic and abiotic stress responses.Here, we found 12 differentially regulated genes in the opr2-1 mutant when compared to the wildtype Arabidopsis.Of the 12 differentially expressed genes, six have been functionally annotated, two of which can be specifically linked to processes known to occur in M. incognita-infected roots.As first, the differentially regulated MAIL1 gene (At2g25010) is shown to be essential for maintaining correct cell division and differentiation in root apical meristems ( Ühlken et al., 2014).This could be a lead to investigate further as root-knot nematodes transform host cells into giant cells by bypassing cytokinesis in the mitotic cell division cycle (De Almeida & Gheysen, 2013).Furthermore, Arabidopsis mail1 mutants show significant differentiation defects leading to disorganized growth of callus-like structures, which have evident similarities to the tumor-like galls formed around infection sites of root-knot nematodes (Olmo et al., 2020).The second gene standing out in the list of OPR2-regulated transcripts, MED8, encodes a subunit of the Mediator complex involved in regulating JAdependent defense responses to biotic and abiotic stresses (Kidd et al., 2009;Zhang et al., 2012;Li et al., 2018).Further research using Arabidopsis mutants may show whether the genes differentially regulated in nematode-infected roots of the opr2-1 mutant are causally related to host susceptibility, or whether they represent a secondary transcriptional fall-out of the catalytic activity of OPR2.
It has only recently been discovered that OPR2 catalyzes the conversion of 4,5-ddh-JA, a derivative of 12-OPDA, to JA (Chini et al., 2018).Cytosolic OPR2 is positioned parallel to the peroxisomal JA biosynthesis pathway regulated by OPR3, which is thought to control the main route in the conversion of 12-OPDA into JA.Herein, OPR3 reduces 12-OPDA to 3-oxo-2(29 [Z]-pentenyl)-cyclopentane-1-octanoic acid (OPC:8), which is further converted through several ß-oxidation rounds into JA (Mussig et al., 2000;Schaller et al., 2000;Stintzi & Browse, 2000).Based on the mildly enhanced susceptibility to M. incognita of the opr2-1 and opr3-3 single mutants and the further increase in the susceptibility of the double-mutant opr2-1/ opr3-3, we conclude that both OPR2 and OPR3 are important and not operating redundantly in the complex hormonal regulation of host susceptibility to root-knot nematodes.This seems to contradict with earlier observations with the opr3-1 mutant in Arabidopsis, which suggested that OPR3 is not required for host responses to root-knot nematodes (Gleason et al., 2016).However, the opr3-1 mutation is a conditional allele in the background on Arabidopsis ecotype Wassilewskija, while loss-offunction of opr3-3 is unconditional in the background of Arabidopsis ecotype Columbia-0 (Chini et al., 2018).Furthermore, Arabidopsis ecotypes exhibit significant quantitative variation in susceptibility to root-knot nematodes (Warmerdam et al., 2018), which underlines that the genetic context of mutations can have a profound effect on observed plant phenotypes.Gleason et al. (2016) also conducted bioassays using the northern rootknot nematode M. hapla, which is only distantly related to M. incognita and may be less sensitive to OPR3-dependent JAmediated host responses.Notably, M. hapla was the only rootknot nematode species included in our comparative genome analyses that seem to lack an ortholog of MiMSP32.
The effector MiMSP32 may target host OPR2 to regulate the pool of 12-OPDA and downstream derivates, but targeting OPR2 alone is probably not sufficient to alter the accumulation of JA.Earlier work on Arabidopsis has shown that only in the background of the opr3-3 mutant allele, OPR2 significantly affects the conversion of 4,5-didehydro-JA into JA and JA-Ile (Chini et al., 2018).Our data show that OPR2 directs the transcriptome of nematode-infected roots toward abiotic and biotic responses, but this does not have a typical JA profile.Likewise, our tomato lines overexpressing MiMSP32 did not show a significant change in JA levels either, although we observed a slight increase in the level of 12-OPDA.The enhanced susceptibility of the single-mutant opr2-1 and the MiMSP32 overexpressing plants to M. incognita may therefore not depend on JA signaling but could instead rely on the perception of 12-OPDA or other derivates.In line with this, Gleason et al. (2016) found that the Arabidopsis 12-OPDA receptor mutant CYP20-3 is more susceptible to infections by root-knot nematodes, albeit it with a different species.Our data could thus add to the increasing evidence pointing at autonomous signaling functions of 12-OPDA distinct from JA in adaptive responses to biotic and abiotic stress (reviewed in Liu & Park, 2021).However, further research is needed to establish whether indeed MiMSP32 alters 12-OPDA levels in nematode-induced syncytia in an OPR2-dependent manner, and whether MiMSP32-enhanced susceptibility requires OPDA perception and signaling (Park et al., 2013;Liu & Park, 2021).
In conclusion, M. incognita delivers the stylet-secreted effector MiMSP32 into host cell cytoplasm to modulate OPR2-dependent host defenses.Like root-knot nematodes, many wound-inducing pathogens and insects of plants are affected by oxylipin-dependent host defenses (Zhang et al., 2017) and artificially manipulating the conversion of the 12-OPDA by 12-OPDA-reductases has been shown to alter susceptibility to multiple unrelated plant invaders (Scalschi et al., 2015;Gleason et al., 2016;Chini et al., 2018).However, to the best of our knowledge, this study is the first to shed to light on a novel effector mechanism targeting this process to promote parasitism in plants.
to the equipment and technical support.Additionally, we would like to thank Unifarm of Wageningen University & Research for support with plant material and the Laboratory of Plant Breeding for support with plant transformations.Table S1 Primers used in this study.
Table S2 All estimates of parameters for the different models of evolution for the 27 MiMSPs.
Table S3 Summary statistics of the yeast two-hybrid screen results.
Table S4 Yeast two-hybrid screen identified 11 probable candidate interactors of MiMSP32, of which six with a high confidence in the interaction.
Table S5 Overview of the six tomato proteins interacting with MiMSP32 in yeast two-hybrid.
Table S6 Physiochemical protein characteristics for MiMSP32 -sp and the six selected interaction domains of the six putative tomato host targets as predicted by ExPASy.
Table S8 RNA-Seq differentially regulated genes between opr2-1 mutant plants and Col-0 wild-type plants, based on an interaction model for plant line and Meloidogyne incognita infection.
Table S9 Details of the 13 differentially regulated genes between nematode-infected roots of the Arabidopsis opr2-1 mutant and wild-type Col-0 plants as extracted from the TAIR database.
Please note: Wiley is not responsible for the content or functionality of any Supporting Information supplied by the authors.Any queries (other than missing material) should be directed to the New Phytologist Central Office.

Fig. 1
Fig. 1 MiMSP17, MiMSP31, and MiMSP32 are under positive diversifying selection.(a) The number of significantly matching genes within all published assemblies of root-knot nematode genomes in the WormBase ParaSite database queried with 27 MiMSPs using BLASTN.Colors indicate nematode species.(b) The number of significantly matching genes for each specific MiMSP in the Meloidogyne incognita PRJEB8714 genome sequence.MiMSPs with three or more paralogous genes in the M. incognita genome are indicated in purple.Small colored rectangles below the bars indicate identical hits caused by high sequence similarity in MiMSP query.(c) Estimated nonsynonymous/synonymous substitutions rates (omega) for six clusters of MiMSP paralogs.In yellow, MiMSP genes with clusters of paralogs for which the log-likelihood ratio test significantly favored model M8 for positive selection (***, P < 0.0001; ns, not significant).

Fig. 2
Fig. 2 MiMSP32 promotes Meloidogyne incognita parasitism in tomato.(a) The target sequence in MiMSP32 for host-induced gene silencing using hairpin construct MiMSP32_hp.AY142116.1 refers to the GenBank accession number of the MiMSP32 mRNA sequence.(b) Number of galls per plant at 7 d postinoculation (dpi) with M. incognita on two independent tomato lines overexpressing the hairpin construct MiMSP32_hp (MiMSP32_hp04 and MiMSP32_hp08) and wild-type tomato plants (MM).(c) Number of galls per plant on two independent tomato lines overexpressing MiMSP32 coding sequence (MiMSP32_ox01 and MiMSP32_ox06) without the predicted signal peptide for secretion and wild-type tomato plants (MM) at 7 dpi with M.incognita.Data were collected in multiple independent experiments, batch-corrected, and combined for statistical analysis with a one-way ANOVA.Boxplots represent the minimum, first quantile, median, third quantile, and maximum.The diamond shape represents the average value, and the circles represent all individual measurements.Asterisks indicate significant differences between transgenic lines and wild-type tomato (as calculated with a one-way ANOVA and Tukey's HSD): *, P < 0.05; ***, P < 0.001.

Fig. 3
Fig. 3 Protein fragment of the putative host target 12-oxophytodienoate reductase 2 (SlOPR2) interacts with MiMSP32 -sp in planta.(a) Projection of six independent clones (A-46 to A-257) and the selected interaction domain (SID) from a yeast two-hybrid screen using MiMSP32 as bait on the coding sequence of tomato SlOPR2.(b) Co-immunoprecipitation of HA-tagged MiMSP32 -sp and MYC-tagged SlOPR2 SID using anti-HA magnetic beads.Equal loading is visualized by Coomassie brilliant blue (CBB) staining on total protein extracts isolated from agroinfiltrated leaves of Nicotiana benthamiana.The yellow arrow indicates SlOPR2SID-mCh-MYC4 and the white arrow indicates YFP_MYC2.(c) Subcellular localization of MiMSP32 -sp _GFP_HA4 with and without mCherry-tagged SlOPR2 SID (right panel) or free mCherry (left panel) in agroinfiltrated leaves of N. benthamiana.The confocal pictures represent the mCherry channel (cyan) and the combined GFP channel (green) with chloroplasts (purple).'N' indicates the position of the nucleus.Bar, 10 μm.(d) The ratio of GFP expression in cytoplasm over nucleus after agroinfiltration in N. benthamiana of HA-tagged MiMSP32 -sp with SlOPR2SID-mCh-MYC4 (right panel) or mCherry alone (left panel).Statistical analysis was done with a one-way ANOVA and a post hoc Tukey HSD.Boxplots represent the minimum, first quantile, median, third quantile and maximum.The diamond shape represents the average value and the circles represent all individual measurements.The horizontal line inside the box represents median.The whiskers mark the values within the first and third quantile.Asterisks indicate significant differences between mCherry-tagged SlOPR2 and MiMSP32 -sp without any co-expressed construct: ***, P < 0.001; ns, not significant.All leaf samples were taken at 48 h after agroinfiltration of the constructs in N. benthamiana.

Fig. 4
Fig. 4 AtOPR2 is a host target of MiMSP32.(a) Detailed locations of the T-DNA insertions (arrows) in the Arabidopsis mutants ala1-1, ala1-2, tppII-1, and opr2-1.(b) Boxplot of the number of acid fuchsin-stained Meloidogyne incognita juveniles (J2s) in mutant and wild-type Arabidopsis plants at 7 d postinoculation (dpi).Data were collected in at least three independent experiments with n ≥ 16, and all data were combined for statistical analysis with ANOVA and Tukey's HSD.Boxplots represent the minimum, first quantile, median, third quantile, and maximum.The diamond shape represents the average value, and the circles represent all individual measurements.Asterisks indicate significant differences between T-DNA lines and wild-type Col-0 Arabidopsis plants: *, P < 0.05; ns, not significant.(c) Co-immunoprecipitation in Nicotiana benthamiana leaf extracts by pulling down MiMSP32 -sp using anti-HA magnetic beads and anti-MYC detection of AtOPR2_mCh_MYC4.Proteins were extracted from multiple leaves and from multiple plants at 48 h after agroinfiltration.Equal loading is visualized by Coomassie brilliant blue (CBB) staining on total protein extracts of agroinfiltrated leaf areas.

Fig. 5
Fig.5Thirteen genes are differentially regulated in roots of Arabidopsis thaliana opr2-1 plants compared with wild-type plants during Meloidogyne incognita infection.(a) Volcano plot of differentially regulated genes using interaction model for opr2-1 and wild-type plants on 1, 4, and 7 d postinoculation (dpi) with M. incognita juveniles (J2) or mock-inoculated.Significantly differentially regulated genes (false discovery rate, FDR < 0.05) are colored in blue; nonsignificant differences are colored in red.(b) Venn diagram with the differentially regulated genes, organized by significant factors from the interaction model.(c) Relative expression as the ratio of transcripts per kilobase million (tpm_rat) of the 13 differentially regulated genes at 0, 1, 4, or 7 dpi with M. incognita juveniles (J2) or mock-inoculated for the opr2-1 mutant or Col-0 wild-type plants.The line represents the average and the ribbon the AESE values.na, not available.
regulates susceptibility to Meloidogyne incognita independently of OPR3.(a) Conversion of 12oxophytodienoic acid (12-OPDA) into jasmonoyl-isoleucine by AtOPR3 and AtOPR2 as proposed by Chini et al. (2018).ACX, acyl-CoA oxidase; KAT, 3-ketoacyl-CoA thiolase; MFP, the multifunctional protein.(b) Number of acid fuchsin-stained M. incognita juveniles (J2s) in wild-type, opr2-1, opr3-3, or opr2-1/opr3-3 mutant plants at 7 d postinoculation (dpi).Data were collected in at least two independent experiments with a total n ≥ 27 and combined for statistical analysis.(c) Number of root tips per genotype at the day of inoculation.(d) Hormone levels of jasmonic acid (JA) (i.e. total of JA and JA-Ile) and 12-OPDA (the total of the four different isomers, fmol mg −1 FW) in leaves of two independent tomato lines overexpressing MiMSP32 coding sequence (MiMSP32_ox01 and MiMSP32_ox06) without the predicted signal peptide for secretion and wild-type tomato plants (MM).All statistical differences were determined with a one-way ANOVA and a post hoc Tukey's HSD test.Boxplots represent the minimum, first quantile, median, third quantile, and maximum.The diamond shape represents the average value, and the circles represent all individual measurements.Asterisks indicate significant differences or between mutant or overexpression plants and wildtype plants: ***, P < 0.001; ^, P < 0.10; ns, not significant.New Phytologist (2023) 237: 2360-2374 www.newphytologist.comÓ 2022 The Authors New Phytologist Ó 2022 New Phytologist Foundation

Fig. S4
Fig. S4 MiMSP31 and MiMSP32 and three paralogous genes in the genome of Meloidogyne incognita.

Fig. S5
Fig. S5 Position of yeast two-hybrid clone inserts interacting with MiMSP32 relative to the whole gene sequence of the six putative host targets.

Fig
Fig.S8RNA-Seq transcripts affected by time and treatments.

Fig. S9
Fig.S9Expression of jasmonic acid biosynthesis associated genes (KEGG module M00113) in nematode-infected roots of the opr2-1 mutant and wild-type Col-0 plants over time.