A MIF-like effector suppresses plant immunity and facilitates nematode parasitism by interacting with plant annexins.

Plant-parasitic nematodes secrete numerous effectors to facilitate parasitism, but detailed functions of nematode effectors and their plant targets remain largely unknown. Here, we characterized four macrophage migration inhibitory factors (MIFs) in Meloidogyne incognita resembling the MIFs secreted by human and animal parasites. Transcriptional data showed MiMIFs are up-regulated in parasitism. Immunolocalization provided evidence that MiMIF proteins are secreted from the nematode hypodermis to the parasite surface, detected in plant tissues and giant cells. In planta MiMIFs RNA interference in Arabidopsis decreased infection and nematode reproduction. Transient expression of MiMIF-2 could suppress Bax- and RBP1/Gpa2-induced cell death. MiMIF-2 ectopic expression led to higher levels of Arabidopsis susceptibility, suppressed immune responses triggered by flg22, and impaired [Ca2+]cyt influx induced by H2O2. The immunoprecipitation of MiMIF-2-interacting proteins, followed by co-immunoprecipitation and bimolecular fluorescence complementation validations, revealed specific interactions between MiMIF-2 and two Arabidopsis annexins, AnnAt1 and AnnAt4, involved in the transport of calcium ions, stress responses, and signal transduction. Suppression of expression or overexpression of these annexins modified nematode infection. Our results provide functional evidence that nematode effectors secreted from hypodermis to the parasite cuticle surface target host proteins and M. incognita uses MiMIFs to promote parasitism by interfering with the annexin-mediated plant immune responses.

Introduction

Root-knot nematodes (RKNs; Meloidogyne spp.) are among the most devastating obligate parasites of plants, causing billions of dollars of agricultural losses worldwide annually (Ibrahim ; Jones ). The apomictic RKN species M. incognita, M. javanica, and M. arenaria can infect thousands of plant species (Teixeira ). Disease development requires the motile RKN pre-parasitic second-stage juveniles (J2s) to penetrate the host root apex and migrate towards the vasculature, where they select root cells and transform them into specialized feeding structures known as ‘giant cells’ (Xue ; Favery ), which serve as the sole source of nutrients for the developing nematode (Caillaud ). Like other plant pathogens, RKNs secrete effector proteins into plant cells, to manipulate host cell processes and facilitate parasitism. Several nematode organs can secrete effectors, including the esophageal gland, amphids, phasmids, and hypodermis (Haegeman ). Only a few effectors have been detected as secreted in planta. Some have been shown to be present in the apoplast in planta (Vieira ), and three RKN effectors have been shown to target the giant cell nuclei (Jaouannet ; Lin ; Chen ). Like the esophageal gland effectors, cuticular secretions may play an important role in the suppression of plant defense mechanisms (Davies and Curtis, 2011). For example, a functional glutathione peroxidase (GpX) present in the hypodermis can metabolize larger hydroperoxide substrates and probably contributes to the ability of the potato cyst nematode, Globodera rostochiensis, to suppress host defenses (Jones ). Secreted peroxiredoxins present on the surface of G. rostochiensis (Robertson ) and in the hypodermis and pseudocoelom of M. incognita (Dubreuil ) can improve nematode development by scavenging reactive oxygen species (ROS). In addition, a fatty acid- and retinol-binding protein that accumulates at the surface of the M. javanica cuticle and is secreted into the surrounding host tissues may play a vital role in promoting nematode parasitism and modulating host defenses (Iberkleid ).

Macrophage migration inhibitory factor (MIF) was the first cytokine shown to possess T cell-derived activities involved in the inhibition of macrophage migration (David, 1966). MIF proteins are secreted from cells in a non-conventional secretion pathway, and the secretion of MIF from anterior pituitary cells in response to lipopolysaccharide stimulation suggests that MIF may be a hormone (Bernhagen ). MIFs have been shown to have a broad range of biological functions, including an ability to modulate innate and adaptive immune responses, act as a pro-inflammatory cytokine and a chemokine, regulate angiogenesis, apoptosis, and fibrosis, and function as an enzyme with tautomerase and thiol-protein oxidoreductase activity (Calandra and Roger, 2003; Morand ; Xu ). In animals, there have been many breakthroughs in MIF functional mechanisms. Intracellular signal transduction is initiated by receptor recognition, in which extracellular MIF binds to cell surface protein CD74, along with the potential accessory protein CD44. Sustained extracellular signal-regulated kinases 1 and 2 (ERK1/2) phosphorylation by the MIF–CD74–CD44 complex activates the mitogen-activated protein kinase (MAPK) pathway, which further affects downstream immune responses (Mitchell ; Leng ; Shi ). Meanwhile, MIF could interact with Jun-activation domain-binding protein 1 (JAB1) in the cytosol, antagonize JAB1-controlled pathways, and modulate cellular redox homeostasis (Kleemann ; Lue ). Furthermore, MIF could interact with two chemokine receptors, CXCR2 and CXCR4, of which CXCR2 also physically interacts with CD74. The complex formed of MIF, CXCR2, and CD74 triggers calcium Ca2+ influx and rapid integrin activation (Bernhagen ). A recent study showed that MIF is involved in inflammation by interacting with a NOD-like receptor (NLRP3) and mediating the interaction between NLRP3 and the intermediate filament protein vimentin (Lang ).

MIF-like proteins from invertebrate species have been shown to be involved in immune escape from hosts. The first nematode MIF homolog was characterized in the filarial parasite Brugia malayi, which secretes a MIF-like protein (BmMIF) that can be detected in somatic extracts and in the excretory–secretory products of the parasite, providing the first demonstration that BmMIF could promote parasite survival by modifying host immune responses (Pastrana ). A MIF-like protein produced by the hookworm Ancylostoma ceylanicum, AceMIF, was shown to bind to the human MIF receptor CD74 and to facilitate infection of larval stages and adult worm survival within the intestine; these findings indicate that AceMIF may act as a virulence factor regulating the host immune response (Cho ). Similarly, the filarial parasite Onchocerca volvulus secretes MIF-like proteins into the outer cellular covering of the adult worm body, the syncytial hypodermis, and the uterine wall (Ajonina-Ekoti ). Recently, a MIF-like protein was reported to be secreted by aphids; this protein acts as a cytokine, modulating host immune responses (Naessens ), in a new mechanism of interaction between parasites and plant hosts. However, it remains largely unknown whether and how MIF-like proteins regulate host immune responses or signal transduction during plant–pathogen interactions. Interestingly, plant MIF-like proteins (MDLs) were recently identified, but no functional data are available to date (Panstruga ).

Plants have evolved an innate immune system to protect them against various types of attack, in which pathogen-associated molecular pattern (PAMP)-triggered immunity (PTI) and effector-triggered immunity (ETI) play essential roles (Jones and Dangl, 2006; Dangl ). The model plant Arabidopsis displays induced immune response-related marker gene expression as part of a response to pathogens (Pieterse ; Jaouannet ), involving local callose deposition at the site of penetration (Ellinger ), and the generation of ROS (Tripathy and Oelmüller, 2012). Many Arabidopsis genes have been implicated in the plant immune system, including those encoding annexins, which act as important regulators of various abiotic and biotic stress responses (Clark ; Laohavisit and Davies, 2011). Interestingly, cyst nematodes were reported to secrete annexins to assist parasitism. The beet cyst nematode effector Hs4F01 is an annexin-like protein that interacts with an Arabidopsis oxidoreductase member of the 2-oxoglutarate (2OG)–Fe(II) oxygenase family (Patel ). Another cyst nematode annexin-like effector is Ha-ANNEXIN secreted by Heterodera avenae, which could suppress host immune responses (Chen ). These results highlight the importance of annexins for plant-nematode interactions.

In this study, we characterized MIF-like genes from M. incognita that are up-regulated upon infection. We showed that MiMIF proteins are secreted into plant tissues and giant cells, thereby promoting parasitism. We chose MiMIF-2 as a representative and found that this protein engaged in physical interactions with two Arabidopsis annexins, which suppressed host immune responses. Based on these results, we suggest that MiMIF-2 may act as a novel effector interacting with plant annexins to manipulate host immune responses and signal transduction, to promote the survival of parasitic stages of M. incognita.

Materials and methods

Nematode culture, plant materials and growth conditions

Meloidogyne incognita strains were reproduced initially from a single egg mass on tomato plants (Solanum lycopersicum var. ‘Baiguo’) in a greenhouse at 25 °C. Infective J2s were collected with an improved Baermann funnel for 48–72 h. Different parasitic life stages of the nematode were collected from roots as previously described (Chen ). Infected tomato roots were collected at different days after inoculation, cut into 1 cm sections and digested in a mixture of pectase and cellulose (Sinopharm Chemical Reagent Beijing Co., Ltd, China) at 28 °C and 160 rpm overnight. Nematodes were separated under a stereomicroscope.

Surface-sterilized Arabidopsis ecotype Columbia (Col-0) was sown on MS medium (Sigma-Aldrich, USA) as previously described (Quentin ). The AnnAt1 and AnnAt4 T-DNA mutant lines (SALK_015426 and SALK_096465) were obtained from the Arabidopsis Biological Resource Center (ABRC, USA). Tobacco (Nicotiana benthamiana) seeds were grown in pots under conditions of 16 h light and 8 h dark at 23 °C with 60–75% relative humidity.

For in vivo nematode infection, Arabidopsis or tomato seedlings were inoculated in soil with 300 J2s per plant. Roots were collected at 35 d post-infection (dpi) and stained using the sodium hypochlorite–acid fuchsine method (Bybd ). Egg masses and galls were counted under a dissecting microscope (Olympus, Japan). Juveniles inside roots were isolated by digestion with cellulase. Each experiment was performed three times independently, and 30 plants were counted each time. Data were analysed using Statgraphics Centurion 17 software (Statgraphics Technologies, USA). One-way ANOVA followed by Duncan’s post hoc test was run to identify significant differences between treatments.

RNA isolation and gene amplification

Meloidogyne incognita mRNA was extracted using a Dynabeads® mRNA DIRECTTM Kit (Invitrogen, USA), and complementary DNA (cDNA) was synthesized using reverse transcriptase SuperScript III (Invitrogen). Arabidopsis total RNA was isolated from seedlings using TRIzol Reagent (Invitrogen) and cDNA was synthesized using M-MLV reverse transcriptase (TaKaRa, Japan). MIF-like genes were amplified from cDNA or gDNA of M. incognita by PCR using specific primers. The PCR products were cloned into the pMD18-T vector (TaKaRa, Japan) and sequenced. All primers used in this study are listed in Supplementary Table S1 at JXB online and were synthesized by TsingKe Biotechnology Co. Ltd, Beijing, China.

Accession numbers and databases used

The predicted nucleotide sequences of MiMIFs were obtained from the National Center for Biotechnology Information (NCBI) EST database (http://www.ncbi.nlm.nih.gov/nucest/), and from Meloidogyne genomic resources (http://www6.inra.fr/meloidogyne_incognita/). Multiple amino acid sequence alignment analyses of MIF-like proteins from B. malayi, Arabidopsis, Homo sapiens, M. hapla, M. incognita, Ascaris suum, A. ceylanicum, O. volvulus, Caenorhabditis elegans, and Plasmodium falciparum were conducted using DNAMAN V6 (Lynnon Biosof, USA). All accession numbers are listed in Supplementary Table S2.

Developmental expression analysis and RT-qPCR analysis

Illumina transcriptomic data obtained during developmental stages (eggs, pre-parasitic J2s, parasitic juveniles, and adult females) were described recently (Blanc-Mathieu ). Real-time quantitative PCR (RT-qPCR) was performed on an ABI Prism 7000 real-time PCR system (Applied Biosystems, USA). Meloidogyne incognita tubulin was used as an internal control. RT-qPCR was conducted using a SYBR Premix Ex Taq (TaKaRa, Japan) under standard conditions. The results were analysed using the method (Livak and Schmittgen, 2001).

Anti-MiMIF-2 polyclonal antibody production and immunolocalization

Recombinant MiMIF-2 (rMiMIF-2, with an N-terminal His tag) was purified using QIAexpress® Ni-NTA Fast Start (Qiagen, Germany). Protein concentration and purity were determined by the BCA Protein Assay Kit (Beijing ComWin Biotech Co., Ltd, China) and by SDS-PAGE. The anti-MiMIF-2 polyclonal serum was purified from immunized rabbits (Beijing Protein Innovation Co., Ltd, Beijing, China). Western blotting was conducted to verify the antibody specificity. The MiMIF-2 mutant (rMiMIF-2-mu; N-terminal Pro1→Gly1) was generated by PCR.

Ultrastructural immunocytochemistry was performed as previously described (Dubreuil ). Freshly hatched J2s were fixed with 4% paraformaldehyde (PAF) or 4% PAF and 0.2% glutaraldehyde (PG) in 0.1 M phosphate buffer (pH 7.4). J2s were dehydrated in an ethanol series and embedded in acrylic resin (LR-White) before sectioning and immunogold labelling. The anti-MiMIF-2 antibody (1:200) and protein A–gold (PAG) of 15 nm (1:50) were used. For controls, sections were processed according to the same procedure excepted that the primary antibody was omitted. Eighty-nanometer ultrathin sections were contrasted with 4% uranyl acetate in water and visualized using a JEM 1400 electron microscope operating at 100 kV equipped with a Morada SIS camera. For immunohistochemistry on sections of tomato galls, we used the protocol previously described (de Almeida Engler ). Tomato galls were fixed in 8% formaldehyde, dehydrated and embedded in butyl-methylmethacrylate. MiMIF-2 antibody and secondary antibody (goat anti-rabbit Alexa Fluor 488 conjugate antibody, Molecular Probes) were diluted 250- and 300-fold in blocking solution, respectively. Nuclei were stained by 4′,6-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich). Images were captured by confocal microscopy (Zeiss LSM880, Germany).

In planta RNAi and generation of transgenic Arabidopsis plants

For RNAi experiments, two MiMIF-2 fragments (1–339 bp) were amplified and cloned in the forward and backward pSAT5 intron, and then inserted into pSuper-RNAi (Dafny-Yelin ). For ectopic expression, the open reading frames (ORFs) of MiMIF-2, AnnAt1 (At1g35720) and AnnAt4 (At2g38750) were amplified and inserted into Super1300-FLAG. The vectors were transformed into Agrobacterium tumefaciens GV3101 and used for transformation of Arabidopsis via the floral dip method (Zhang ). Homozygous T3 plants were confirmed by RT-PCR and western blotting.

Cell-death suppression assay in N. benthamiana, callose staining and defense gene expression analyses

The coding regions of MiMIF-2 and GFP were amplified and cloned into PVX vector pGR107 with a FLAG-tag fused at the N-terminus or a HA-tag fused at the C-terminus and transformed into A. tumefaciens GV3101. Agrobacterium tumefaciens cells carrying BAX or RBP1/Gpa2 were used to trigger cell death in N. benthamiana leaves (Sacco ; Jing ). Agro-infiltrations into 4-week-old N. benthamiana leaves were performed as described in Nguyen .

For callose deposition assay, Arabidopsis leaves from 4-week-old MiMIF-2 lines and wild type (WT, Col-0) were treated with 1 µM flg22 or buffer for 24 h. Samples were stained with aniline blue for 3 h and then examined under a photomicroscope.

Defense suppression test was conducted as described previously (Jaouannet ). To analyse defense gene expression, 10-day-old seedlings of MiMIF-2-expressing line and WT were treated with 1 µM flg22. The expression of the defense marker genes PR1, PR3, FRK1, CYP81, RbohD, and RbohF was determined by RT-qPCR. Arabidopsis ubiquitin carboxyl terminal hydrolase 22 (UBP22) was used as an internal control.

Arabidopsis protein immunoprecipitation (IP) experiments and LC-MS/MS analysis

For the IP experiments, Arabidopsis total protein was extracted from MiMIF-2-FLAG-expressing lines (three different lines) and WT plants. IP was performed using the anti-FLAG M2 affinity gel resin (Sigma-Aldrich). The beads were washed twice with 1 ml of ice-cold phosphate-buffered saline (PBS), followed by the immediate addition of 1 ml protein solution and incubation at 4 °C for 4 h. The resin was thoroughly washed five times with 1 ml of ice-cold PBS, and then the proteins were eluted for analysis. A Q-Exactive nanospray ESI-MS mass spectrometer (Thermo, USA) was used for liquid chromatography–tandem mass spectrometry (LC-MS/MS) at China Agricultural University Functional Genomics Platform. The acquired mass spectrometric data were pre-analysed using Mascot Distiller 2.4 (UK) and then anatomized to search an NCBI non-redundant protein database and the Swiss-prot database.

Subcellular localization, bimolecular fluorescence complementation in planta and aequorin-based [Ca2+]cyt measurement

The open reading frame of MiMIF-2 without stop codon was amplified and cloned into the p35S-eGFP vector with an enhanced green fluorescent protein (eGFP) fused at the C-terminus. For bimolecular fluorescence complementation (BiFC) assay, full-length candidate genes or MiMIF-2 coding sequences were amplified and cloned into the binary vectors pSPYNE and pSPYCE, respectively (Walter ). The constructs were transformed into A. tumefaciens strain EHA105. Four-week-old N. benthamiana leaves were agro-infiltrated and images were captured by laser confocal fluorescence microscopy (Leica SP8, Germany) at an excitation wavelength of 515 nm.

Aequorin/MiMIF-2 transgenic lines were generated from a cross between the aequorin line (Laohavisit ) and MiMIF-2 transgenic lines. The aequorin-annAt1 and aequorin-annAt4 seeds were gifts from Yan Guo’s lab. Ten-day-old aequorin-expressing seedlings were used for measurements of root [Ca2+]cyt according to a previously described method (Laohavisit ). Roots were suspended in aequorin buffer (5 mM KCl, 2 mM CaCl2, and 2 mM Tris/MES, pH 5.7, with 10 µM coelentrazine) for 5 h in the dark (25 °C), washed with osmotically adjusted aequorin buffer without coelentrazine, and resuspended in this buffer (100 µl), and left in the dark for 30 min. The baseline luminescence was recorded for 60 s, followed by the addition of 100 µL of buffer (20 mM H2O2). Luminescence was recorded every second for 120 s. Discharge buffer was then injected (final concentration: 10% (v/v) ethanol, 1 M CaCl2), and the luminescence was recorded for an additional 120 s.

In vivo co-immunoprecipitation assay

For a co-immunoprecipitation (Co-IP) assay, the coding regions of AnnAt1, AnnAt4, or MiMIF-2 were cloned into the PVX vector pGR107 with a FLAG-tag fused at the N-terminus and a HA-tag fused at the C-terminus, respectively, and then transformed into A. tumefaciens GV3101. Nicotiana benthamiana leaves were used for co-expressing MiMIF-2 and annexins using the above-described method. Co-IP was conducted as previously described (Liu ), and SDS-PAGE loading buffer (1×) was added to the beads and heated at 100 °C for 10 min to elute the proteins. The eluted proteins were subsequently subjected to immunoblotting using anti-FLAG (Sigma-Aldrich) and horseradish peroxidase (HRP)-conjugated anti-HA antibodies (Roche, USA), respectively.

Protein extraction and western blotting

Total proteins were extracted from 2 dpi agro-infiltrated N. benthamiana leaves or Arabidopsis seedlings using a protein extraction kit (Beijing ComWin Biotech Co., Ltd, China). For western blot, the primary antibody (anti-FLAG or anti-HA, Sigma-Aldrich) was used at 1: 5000. For immunoblot analysis of MAPK activation, 10-day-old Arabidopsis seedling samples were collected after treatment with 1 µM flg22. A phospho-44/42 MAPK (ERK1/2) polyclonal antibody (Cell Signaling Technology, USA) diluted 1:2000 and an anti-rabbit secondary antibody diluted 1:8000 were used. The protein was detected with the eECL Western Blot Kit (Beijing ComWin Biotech Co., Ltd, China).

Results

Identification of M. incognita MIF-like genes and their expression in different life stages

Our searches of the NCBI EST database and M. incognita genome sequences (Abad ; Blanc-Mathieu ) identified four genes encoding MIF-like proteins: MiMIF-1 (Vermeire ), MiMIF-2, MiMIF-3, and MiMIF-4. These genes are organized in tandem arrays on two scaffolds. MiMIF-1 and MiMIF-2, and MiMIF-3 and MiMIF-4 share 97–98% nucleotide sequence identity, whereas MiMIF-1/MiMIF-2 and MiMIF-3/MiMIF-4 only share about 82% sequence identity (Fig. 1A; Supplementary Fig. S1A). The MiMIF-1 gene has a single base-pair deletion in position 260 (Supplementary Fig. S1A, indicated by black arrow), resulting in the premature termination of translation at bp 285 (Supplementary Fig. S1A, indicated by red frame), leading to a protein truncated by 19 amino acids (aa). Interestingly, MiMIF-3 and MiMIF-1 form an operon and are transcribed together to generate a single mRNA strand (Supplementary Fig. S1B, C). Sequence alignment showed that MiMIF-1 (94 aa), MiMIF-2 (113 aa), MiMIF-3 (112 aa), and MiMIF-4 (112 aa) were 75–99% identical (Fig. 1A; Supplementary Fig. S1D). They all contain three conserved residues (Pro1, Lys32, and Ile64) of the active sites for tautomerase activity identified in human MIF (Fig. 1B, black triangle), but MiMIF sequences lacked the Val106 residue (Fig. 1B, red triangle) for tautomerase activity and the C-X-X-C motif found in the thioredoxin superfamily of thiol-protein oxidoreductases (Fig. 1B).

Fig. 1.

Meloidogyne incognita MIF-like genes are more strongly expressed in parasitic stages. (A) Organization of MIF genes in the M. incognita genome. Exons are represented by arrows, introns by blue lines. The M. incognita scaffolds and the percentage nucleotide/amino acid identity between the ORFs/proteins are indicated. (B) Alignment of the deduced amino-acid sequences of the MIF-like proteins from M. incognita, Brugia malayi, Arabidopsis, C. elegans, Plasmodium falciparum, Onchocerca volvulus, Meloidogyne hapla, Myzus percicae, and humans. Identical amino acids are shown in black, highly similar (>75%) amino acid residues are shown in pink, and similar (>50%) amino-acid residues are shown in blue. Triangles (black and red) indicate the tautomerase active site residues conserved in mammalian MIF-like proteins; black triangles are tautomerase active site residues that exist in MiMIFs; red line, C-X-X-C motif. (C) Developmental expression profile of MiMIFs. The mean gene expression is presented as fragments per kilobase of transcript per million mapped reads (FPKM) values and were calculated from three independent biological samples for each developmental stage: eggs, pre-parasitic second-stage juveniles (pre-J2), and parasitic juveniles and females (par-Js). Data are presented as means ±SEM.

The nucleotide sequences of the four MiMIFs displayed a high level of identity, and MiMIF-3 and MiMIF-1 were found to be cotranscribed (Supplementary Fig. S1A, B), making it difficult to study individual MiMIF expression profiles. However, Illumina transcriptomic data obtained during developmental stages (eggs; pre-parasitic J2s; parasitic juveniles and adult females) (Blanc-Mathieu ) revealed higher levels of MiMIFs expression in parasitic juveniles than in pre-parasitic J2s (Fig. 1C).

MiMIFs are localized in the hypodermis and cuticle, and secreted into plants and accumulated in giant cells

To analyse the tissue localization of MiMIFs and demonstrate that MiMIFs are secreted into host-plant tissues, we produced a polyclonal antibody against rMiMIF-2. On western blots, this polyclonal antibody specifically detected rMiMIF-2 and MiMIFs in total proteins from pre-parasitic J2s at the expected size of ~13 kDa (Supplementary Fig. S2). No signal was detected in total proteins from tomato roots or leaves. Ultrastructural immunocytochemistry analyses with the polyclonal antibody were performed on pre-parasitic J2s. MiMIF proteins were mostly detected in the hypodermis in contact with the cuticle, in the cuticle, in the pseudocoelom surrounding the stylet, and barely in the muscle (Fig. 2A, B; Supplementary Fig. S3). No gold particles (black dots) were observed in control treatments (Fig. 2C). Immunohistochemistry analyses were performed on gall sections from tomato plants 15 d post-infection (dpi) with M. incognita. Fluorescence was observed around the nematode head and cuticle adjacent to the surrounding host cells (Fig. 2D, E; Supplementary Fig. S4). This result indicates secretion into and accumulation of MiMIFs in the intercellular space between the nematode body and the host cells. Moreover, several giant cells showed a clear signal at the cytoplasmic level and in particular near the plasma membrane (Fig. 2G, H). In control experiments with secondary antibody only, no signal was observed on sections of infected roots at similar time points (Fig. 2F, I). Thus, MiMIFs are secreted in planta from M. incognita hypodermis to the parasite surface during parasitism and are able to target the giant cells.

Fig. 2.

Localization of MiMIF proteins in nematodes and in planta. (A, B) Ultrastructural immunocytochemistry of M. incognita MIF proteins. Section of the head and anterior part of a pre-J2 fixed in 4% paraformaldehyde (PAF) or 4% PAF and 0.2% glutaraldehyde (PG) showing MIF protein accumulations in the hypodermis (h) and in the cuticle (c), and also detected in the body wall muscle (bwm) and in the pseudocoelom surrounding the stylet (s). ad, amphid. n=9 indicates the number of similar results. (C) Control sections of J2s fixed in PG or PAF without anti-MiMIF-2 antibody. (D–I) Immunodetection of the MiMIF proteins in tomato gall sections. Tomato roots were fixed into methylacrylate after infection with M. incognita for 15 d. The anti-MiMIF-2 antibody (1:250) was applied on gall sections. Alexa Fluor 488-conjugated secondary antibody was used to detect the fluorescence. (D, E) MiMIF proteins accumulated along the nematode cuticle and adjacent plant tissues surrounding the nematode head. (F) Galls containing parasitic J2 nematodes incubated only with secondary antibody as a negative control, showing no signal. (G) Signal was observed around nematode head and in giant cells near the plasma membrane. (H) Signal was detected around the giant cell and nucleus. (I) Galls containing nematode and giant cells incubated only with secondary antibody as a negative control, showing no signal. Micrographs (E, F, G, H) are overlays of images of the differential interference contrast, Alexa Fluor 488-conjugated secondary antibody and DAPI-stained nuclei. Individual images are listed in Supplementary Fig. S3. N, nematode; *, giant cell. Scale bars 20 μm for images (D–I). White arrows indicate the MiMIF-2 fluorescence.

In planta RNAi of MiMIFs affects nematode parasitism

We assessed the role of MiMIF genes during parasitism by silencing MiMIFs in planta through host-derived RNAi. Four homozygous Ri-T3 transgenic Arabidopsis lines expressing the MiMIF-2 full-length coding sequence (CDS) hairpin double-stranded RNA (dsRNA; 339 bp in pSuper vector; the identical phenotypes of different lines are shown in Supplementary Fig. S5) and two GFP hairpin dsRNA lines were obtained. MiMIFs transcript abundance was assessed in parasitic nematodes extracted at 14 dpi, and was found to be strongly reduced (64% to 80%) in the four Ri-T3 lines compared with control lines (Fig. 3A), demonstrating the efficacy of host-derived RNAi-mediated gene silencing. At 35 dpi, the numbers of galls and nematodes were about 60% lower (P<0.05) in Ri-T3 plants than in plants treated with the GFP control lines (Fig. 3B). Similar results were obtained in three independent trials. These findings suggest that MiMIFs play a key role in root-knot nematode parasitism.

Fig. 3.

The effect of MiMIF-2 silencing and expression in planta on M. incognita parasitism. (A) RT-qPCR assays showed the level of MiMIF expression in M. incognita collected from Arabidopsis lines (GFP RNAi lines and Mimif-2 RNAi lines) treated by host-derived RNAi in planta, 14 d post-inoculation (dpi). The tubulin gene was used as an internal control. (B) The total numbers of galls and nematodes were counted at 35 dpi in the Mimif-2 RNAi Ri-T3 lines (Ri-T3-2, -5, -7, and -8). (C) MiMIF-2 expression in Arabidopsis increases susceptibility to M. incognita. The total numbers of galls and nematodes were counted at 35 dpi in MiMIF-2 ectopically expressing lines (MiMIF-2-OE-3, -4, and -5). Data are presented as means ±SD (n≥30). Asterisks denote significant differences (*P<0.05; ***P<0.001, one-way ANOVA). All experiments were performed three times with similar results, and at least 30 plants were analysed per line. (This figure is available in color at JXB online.)

MiMIF-2 expression in Arabidopsis increases susceptibility to M. incognita

We showed that native MiMIFs were indeed delivered into the cytoplasm of host cells and we then expressed MiMIF-2 ectopically in transformed Arabidopsis lines. Transgenic plants were verified by RT-PCR; the phenotype of three independent homozygous T3 lines expressing MiMIF-2-FLAG (lines OE-T3-3, OE-T3-4, and OE-T3-5) showed no obvious difference compared with WT plant, and the presence of a 13 kDa MiMIF-2-FLAG protein was confirmed by western blotting (Supplementary Figs S5, S6). Infection assays with M. incognita showed that all three transgenic lines were significantly (P<0.05) more susceptible than WT to M. incognita infection (Fig. 3C). The transgenic lines had up to 30% more galls and parasitic nematodes at 35 dpi. Similar results were obtained in three independent biological tests, each with 30 plants.

MiMIF-2 suppresses defense responses in plants

As the expression of MiMIF-2 in Arabidopsis enhanced M. incognita infection, we investigated whether MiMIF-2 may regulate plant immune responses. We first tested whether MiMIF-2 has the ability to suppress programmed cell death (PCD) induced by proapoptotic BAX or Globodera RBP-1/Gpa2 resistance protein. BAX and RBP-1/Gpa2 constructs were agroinfiltrated 24 h after tagged MiMIF-2 or controls into N. benthamiana leaves, respectively. BAX-induced PCD was completely suppressed when the leaves were infiltrated with A. tumefaciens cells carrying FLAG–MiMIF-2 but not with FLAG–GFP (Fig. 4A, B). Moreover, we found that MiMIF-2–HA could suppress cell death triggered by Gpa2/RBP-1 (Fig. 4C, D). Msp40, a M. incognita effector, was used as negative control (Niu ) and did not suppress cell death triggered by Gpa2/RBP-1.

Fig. 4.

MiMIF-2 suppressed BAX- and RBP1/Gpa2-induced programmed cell death (PCD) in N. benthamiana. (A) FLAG-MiMIF-2 suppressed BAX-induced PCD. Nicotiana benthamiana leaves were infiltrated with A. tumefaciens carrying the corresponding constructs, and the cell death photographs were taken at 5 dpi (5 plants and infiltrate 3 leaves of each plant, n=15). (B) For suppressing BAX-induced cell death, the presence of MiMIF-2, GFP, and BAX protein was confirmed by immunoblotting with anti-FLAG or anti-BAX antibodies. (C) MiMIF-2-HA suppressed RBP1/Gpa2-induced PCD. Nicotiana benthamiana leaves were infiltrated with A. tumefaciens carrying the corresponding constructs, and the cell death photographs were taken at 5 dpi. n indicates similar leaf phenotype in each experiment (5 plants and infiltrate 3 leaves of each plant). (D) For suppressing RBP1/Gpa2-triggered cell death, the presence of MiMIF-2-HA, Msp40-HA, and RBP1-HA protein was detected by immunoblotting with anti-HA antibody. The expression of Gpa2 was confirmed by RT-PCR; N. benthamiana translation elongation factor 1α (NbEF1α) was used as internal control. (This figure is available in color at JXB online.)

We next investigated whether MiMIF-2 affected PAMP-triggered immunity (PTI). PTI requires signal transduction from receptors of downstream components via MAPK cascades. We compared the MAPK cascades of WT Arabidopsis and MiMIF-2-OE-T3 plants, by monitoring levels of activated phosphorylated MAPKs, MPK3 and MPK6, after treatment with the bacterial PAMP flg22 (flagellin-derived peptide), with a polyclonal α-phospho-44/42 MAPK (ERK1/2) antibody. The activation of MPK3 and MPK6 by flg22 was significantly weaker in MiMIF-2-OE-T3 seedlings than in WT seedlings (Fig. 5A). We analysed the expression of six defense marker genes (PR1, PR3, FRK1, CYP81, RbohD, and RbohF) after treatment with flg22. As expected, flg22 significantly induced the expression of five defense marker genes (PR1, PR3, FRK1, CYP81, and RbohD) in WT plants, but not RbohF. By contrast, these five marker genes were repressed in MiMIF-2-OE-T3 plants, and this repression was particularly strong for CYP81 (Fig. 5B). Moreover, flg22 induced almost no callose deposition in MiMIF-2-OE plants, contrasting with its effects on WT cotyledons (P<0.001, Fig. 5C, D).

Fig. 5.

MiMIF-2 suppresses the defense triggered by flg22 in Arabidopsis. (A) MiMIF-2 suppresses the pathogen-associated molecular pattern (PAMP)-triggered immunity (PTI) response induced by flg22 in Arabidopsis. Twelve-day-old seedlings of wild-type plants (WT) and plants expressing MiMIF-2 (MiMIF-2-OE-3 and MiMIF-2-OE-4) were treated with 1 µM flg22. MAPK activation was analysed by western blotting with an α-phospho-p44/42 MAPK antibody (Cell Signaling Technology, USA). P, Ponceau S staining of the membrane, demonstrating equal protein loading. (B) WT and MiMIF-2 (MiMIF-2-OE-3 and MiMIF-2-OE-4) seedlings were grown for 10 d on MS agar and treated with 1 µM flg22. Induction of the defense marker genes PR1, PR3, FRK1, CYP81, RbohD, and RbohF in response to flg22 was assessed by RT-qPCR and comparison with plants grown without flg22 treatment. (C, D) Callose deposition in the cotyledons of 10-day-old Arabidopsis seedlings from WT and MiMIF-2-expressing plants (MiMIF-2-OE-3 and MiMIF-2-OE-4) 24 h after mock or 1 μM flg22 treatment. Callose deposition was analysed by staining with aniline blue, images were captured by UV epifluorescence microscopy, and ImageJ was used to determine the number of fluorescent spots. Error bars represent the mean ±SD (n=12), and asterisks denote significant differences (***P<0.001, one-way ANOVA). All these experiments were repeated three times with similar results. (This figure is available in color at JXB online.)

MiMIF-2 interacts with AnnAt1 and AnnAt4 in vivo

We identified the target proteins of MiMIF-2 by performing liquid in planta immunoprecipitation (IP) followed by LC-MS/MS analysis on the three MiMIF-2-OE lines and WT Arabidopsis. We focused on candidate proteins that specifically pulled down in the three MiMIF-2-OE plants but not in the WT control (Supplementary Table S3). We then examined the interactions between these candidates and MiMIF-2 using BiFC assays and validated annexin AnnAt1 (AT1G35720) as a MiMIF-2 target. The co-expression of MiMIF-2–nEYFP and AnnAt1–cEYFP in tobacco leaves reconstituted yellow fluorescent protein (YFP) activity in the cytoplasm of transformed cells (Fig. 6A; Supplementary Fig. S7), whereas no YFP fluorescence was observed in YFP plasmids containing control constructs. These results constitute the first validation of AnnAt1 and MiMIF-2 interaction.

Fig. 6.

MiMIF-2 interacts with two Arabidopsis annexins, AnnAt1 and AnnAt4. (A) Bimolecular fluorescent complementation (BiFC) visualization of the MiMIF-2–annexin interaction. nEYFP–MiMIF-2 and AnnAt1–cEYFP or AnnAt4–cEYFP were co-expressed in tobacco leaves. Controls were performed by co-expressing nEYFP and AnnAt1–cEYFP, nEYFP and AnnAt4–cEYFP, and nEYFP–MiMIF-2 and cEYFP. Images were obtained 36 h after co-expression. YFP, yellow fluorescent protein. Bars, 25 or 50 µm as shown. (B) Co-immunoprecipitation of AnnAt1 or AnnAt4 with MiMIF-2. AnnAt1–FLAG, AnnAt4–FLAG or GFP–FLAG was transiently co-expressed with MiMIF-2–HA in tobacco leaves. The Co-IP experiment was performed with anti-FLAG M2 affinity gel resin (Sigma-Aldrich), and the isolated protein was analysed by immunoblotting with anti-FLAG antibodies to detect AnnAt1, AnnAt4, or GFP, and anti-HA antibodies to detect MiMIF-2. GFP, green fluorescent protein. Three independent experiments were performed with similar results. (This figure is available in color at JXB online.)

Eight annexins have been described in Arabidopsis (Clark ; Cantero ). We investigated whether MiMIF-2 could interact with other annexins that are strongly expressed in roots (AnnAt2, AnnAt3, AnnAt4, AnnAt5, AnnAt8) by using BiFC. We found that only AnnAt4 (AT2G38750) interacted with MiMIF-2 (Fig. 6A; Supplementary Fig. S8). Following agroinfiltration into N. benthamiana leaves, AnnAt–GFP and AnnAt4–GFP localized in the cytoplasm (Supplementary Fig. S9), whereas MiMIF-2 was found in the cytoplasm and the nuclei (Supplementary Fig. S10), consistent with interactions between these proteins occurring in the cytoplasm.

To validate these interactions, Co-IP was performed in N. benthamiana. FLAG–AnnAt1 or FLAG–AnnAt4 and MiMIF-2–HA were co-expressed in plants. As a negative control, tobacco leaves were co-infiltrated with FLAG–GFP and MiMIF-2–HA. An anti-FLAG antibody clearly detected single target bands at 37 and 27 kDa, corresponding to the expected sizes of AnnAt1 and AnnAt4, and eGFP, respectively. Analyses of the immunoprecipitated proteins with HA antibody showed that MiMIF-2 (13 kDa) was specifically pulled down by AnnAt1 and AnnAt4, but not by eGFP (Fig. 6B).

The entire MiMIF-2 structure and the Ca2+-binding sites of annexins are required for interaction

The amino-terminal proline residue of MIF is crucial for its biochemical activities (Calandra and Roger, 2003). Likewise, annexins contain Ca2+-binding sites of functional importance (Gerke and Moss, 2002; Laohavisit and Davies, 2009). We investigated the possible involvement of these sites in protein–protein interactions, by generating a series of mutations in the protein sequences of MiMIF-2, AnnAt1, and AnnAt4 (Fig. 7A). A missense mutant (P2/G, MiMIF-2-mu) and a deletion mutant (MiMIF-2-113aa) were generated in the BiFC and Co-IP vectors. Four missense mutations with amino-acid substitutions affecting the Ca2+-binding sites of AnnAt1 (AnnAt1E68A, AnnAt1D299A, and AnnAt1E68A/D299A) and AnnAt4 (AnnAt4E72A) were also constructed. The BiFC and Co-IP assays clearly showed that the entire structure of MiMIF-2, including, in particular, the amino-terminal proline residue, was required for interaction with AnnAt1 and AnnAt4 (Fig. 7B, D). The Ca2+-binding sites of AnnAt1 and AnnAt4 were also essential for interaction with MiMIF-2 (Fig. 7C, D).

Fig. 7.

The first proline residue of MiMIF-2 and the Ca2+-binding sites of annexins were required for protein interaction. (A) Schematic representation of the intact, mutated, and truncated MiMIF-2 and annexin sequences used for interaction assays in tobacco leaves. The assays were performed three times, with identical results in each case. Symbols: −, no interaction; +, interaction. (B) Co-immunoprecipitation of mutated and truncated MiMIF-2 with annexins. (C) Co-immunoprecipitation of mutated annexins with MiMIF-2. (D) BiFC visualization of the mutated and truncated MiMIF-2–annexin interaction. The corresponding proteins were co-expressed in tobacco leaves. Images were taken 36 h after co-expression. Bars, 10, 25 or 50 µm as shown. GFP, green fluorescent protein; YFP, yellow fluorescent protein. All experiments were repeated three times with similar results. (This figure is available in color at JXB online.)

AnnAt1 and AnnAt4 are involved in Arabidopsis response to M. incognita

We investigated the biological functions of AnnAt1 and AnnAt4 in mediating the response to M. incognita by manipulating the levels of AnnAt1 and AnnAt4 expression via overexpression and knockout approaches in Arabidopsis. Homozygous T3 lines overexpressing AnnAt1 and AnnAt4 under the control of the Super promoter (Supplementary Fig. S6C) were less susceptible to M. incognita than control plants (Fig. 8A), illustrated by a reduced number of galls and nematodes inside roots. In contrast, T-DNA insertional alleles of annAt1 (Salk_015426) or annAt4 (Salk_096465) resulted in the opposite phenotype, a higher number of galls and nematodes (Fig. 8B; Supplementary Fig. S11). These results showed the importance of annexin 1 and annexin 4 for restricting M. incognita infection.

Fig. 8.

AnnAt1 and AnnAt4 are involved in Arabidopsis response to Meloidogyne incognita. (A) Overexpression of AnnAt1 or AnnAt4 in Arabidopsis decreased susceptibility to M. incognita, as indicated by the mean numbers of galls and nematodes at 35 dpi in plant roots. (B) The two annexin mutant lines were more susceptible to M. incognita, as indicated by the mean numbers of nematodes and galls in plant roots. Error bars represent the mean ±SD (n≥30), and asterisks denote significant differences (*P<0.05, one-way ANOVA). All experiments were repeated three times with similar results, and each line counted at least 30 plants. (This figure is available in color at JXB online.)

The increase in [Ca2+]cyt triggered by H2O2 is impaired in MiMIF-2-expressing Arabidopsis

Annexin 1 and annexin 4 have been shown to be involved in Ca2+ signature in Arabidopsis roots (Richards ; Ma ). To identify the roles of MiMIF-2 in Ca2+ influx, MiMIF-2-OE-T3 Arabidopsis and two mutants of AnnAt1 and AnnAt4 with a reporter gene encoding aequorin (Supplementary Fig. S12) were challenged with H2O2. A transient increase in [Ca2+]cyt was observed in all genotypes (WT, MiMIF-2-OE-T3, annAt1, and annAt4). However, there was a significant (P<0.05) difference between the maximum [Ca2+]cyt in aequorin-WT (0.819±0.033 µM) versus aequorin-MiMIF-2 (aequorin-MiMIF-2-OE-3: 0.579±0.055 µM; aequorin-MiMIF-2-OE-4: 0.592±0.011 µM), aequorin-annAt1 (0.568±0.013 µM), and aequorin-annAt4 (0.558±0.015 µM) (mean ±SEM; Fig. 9). Together, these results highlight the key role of MiMIF-2 in modulating plant immune responses and Ca2+ signal transduction.

Fig. 9.

The H2O2-induced increase in [Ca2+]cyt is impaired in MiMIF-2 ectopically expressing Arabidopsis lines, annAt1 mutant line, and annAt4 mutant line. An aequorin gene was expressed in 10-day-old seedlings of WT, MiMIF-2 ectopically expressing Arabidopsis (aequorin-MiMIF-2-OE-3 and aequorin-MiMIF-2-OE-4), annAt1 mutant line, and annAt4 mutant line. Root [Ca2+]cyt was assessed by measuring cytosolic aequorin concentration in response to 20 mM H2O2 treatment. The control response (buffer only, at 0–60 s) was similar in aequorin-WT, aequorin-MiMIF-2-OE-3, aequorin-MiMIF-2-OE-4, aequorin-annAt1, and aequorin-annAt4. H2O2 stimulation was initiated at 60 s. The data shown are mean values ±SEM (n=12). The experiments were performed three times, with similar results in each case. (This figure is available in color at JXB online.)

Discussion

MIFs have been shown to be an important endocrine immune factor in animals (Twu ; Sparkes ), and some animal parasites, including nematodes, ticks and protozoa, have been reported to secrete MIFs as a mechanism of host immune system evasion (Pastrana ; Falcone ). Nevertheless, only a few studies have investigated the role of MIFs in plant parasites. Aphid MIFs were secreted from Myzus percicae saliva and were crucial for aphid survival, fecundity, and feeding on a host plant. The ectopic expression of aphid MIFs in leaf tissues inhibits the major plant immune response (Dubreuil ; Naessens ). Interestingly, MIFs have been identified in many plant-parasitic nematodes, including cyst nematodes and RKNs (Vermeire ). We characterized MIF-like proteins of M. incognita, and explored their functions in plant–nematode interaction.

Secreted effectors play essential roles in nematode–host interactions. Research efforts in recent decades have focused largely on effectors from the esophageal gland cells of plant-parasitic nematodes (Hussey, 1989; Davis ; Hewezi, 2015), and there have been a few reports relating to effectors secreted onto the parasite surface. The majority of known plant nematode effectors are synthesized in specialized secretory cells, the esophageal/salivary glands. So far only a few RKN effectors have been localized inside giant cells, all in nuclei (Favery ; Chen ). In planta immunocytological analyses showed that the giant cell apoplasm is an important compartment of nematode stylet-secreted proteins among which are proteins that are predicted to act intracellularly such as a calreticulin (Vieira ). No mechanism of nematode effector translocation has been described and the minute amount of secreted products might prevent their detection in hypertrophied RKN-induced giant cell cytoplasm. Here we demonstrated that MiMIF-2 secreted from the hypodermis into the plant apoplast finally targets the giant cell cytoplasm (Fig. 2). This result highlights a process by which an apoplastic effector enters into plant cells.

In mammals, MIF has been shown to mediate its biological effects via two signaling mechanisms: the classical receptor-mediated pathway and a non-classical endocytic pathway (Donn and Ray, 2004). The first human MIF receptor to be identified was a cell surface class II major histocompatibility chaperone, CD74 (Leng ). MIF has also been shown to be directly taken up into target cells and to interact with intracellular signaling molecules, including the human Jun activation domain-binding protein JAB-1 (Kleemann ). The characterization of the pathway(s) used by MiMIFs, and/or endogenous plant MDLs, to enter into the plant cells would represent a breakthrough in our knowledge of the mode of action of small proteins in plants.

Host-derived RNAi silencing or ectopic expression of MiMIFs in planta significantly decreased or increased, respectively, the infectiousness of M. incognita, indicating MiMIFs are important for nematode parasitism (Fig. 3). Nematodes feed on viable host cells for nutrition via the stylet; thus, efficient mechanisms must be employed to suppress or evade host defenses at this stage. To achieve this, RKNs deliver several effectors into host cells to suppress immune responses including PCD (Niu ; Chen ; Naalden ; Nguyen ; Shi ). We showed that transient expression of MiMIF-2 could suppress PCD triggered by BAX or RBP1/Gpa2 (Fig. 4). In addition, the use of flg22 to stimulate the defense responses of Arabidopsis (Gómez-Gómez and Boller, 2000; Asai ; Navarro ; Lee and Back, 2016) revealed low levels of MPK3 and MPK6 phosphorylation, suppression of defense marker gene expression and low levels of callose deposition in MiMIF-2-expressing lines (Fig. 5). MPK3 and MPK6 are crucial positive regulators of plant defense responses; they are involved in the biosynthesis of both ethylene (Meng and Zhang, 2013) and jasmonate (Schweighofer and Meskiene, 2008). MAPK signaling in Arabidopsis roots has also been demonstrated to be activate following infection by the cyst nematode Heterodera schachtii (Sidonskaya ). Thus, MiMIF-2 expression may interfere with host signal transduction and immune responses as described in mammalian cells.

In an attempt to understand how MiMIFs interfere with these plant pathways, we used IP-LC-MS to identify proteins that interact with MiMIF-2. BiFC and Co-IP demonstrated two annexins were targeted by MiMIF-2 in Arabidopsis (Fig. 6). MiMIF-2–AnnAt1 and MiMIF-2–AnnAt4 complexes were found to localize within the cytosol. Previous studies have indicated that AnnAt1 not only rescues a mutant E. coli strain lacking OxyR from H2O2 stress (Gidrol ), but also may protect (i) human tumor cells against tumor necrosis factor and (ii) mammalian cells against oxidative stress (Jänicke ; Kush and Sabapathy, 2001). AnnAt1 levels have been shown to increase significantly during osmotic stress and powdery mildew infection (Lee ; Cantero ; Chandran ). The annexin from Brassica juncea (AnnBj1) gave resistance to numerous abiotic and biotic stresses in tobacco (Jami ). There is also evidence to suggest that annexins are involved in plant wound responses through their role in membrane repair (Schapire ; Clark ). Interestingly, cyst nematodes induce expression of ANNEXIN genes in Arabidopsis (Puthoff ) and soybean roots (Klink ) and secrete annexin-like effectors to interfere with plant immune responses. Hs4F01 could complement the function of Arabidopsis annexin 1, indicating a mimicry of plant annexins by nematode annexin. Moreover, Hs4F01 may modulate defense responses through interacting with the 2OG–Fe(II) oxygenase (DMR6) (Patel ). These results suggest dual roles of Hs4F01 during cyst nematode parasitism. However, previous studies on transcript changes of RKN-induced galls showed that the expression of Arabidopsis annexin 1 and annexin 4 were both repressed (Jammes ; Barcala ). Here, we showed that AnnAt1- or AnnAt4-overexpressing lines are more resistant to M. incognita and that, on the contrary, the AnnAt1 or AnnAt4 knockout lines are more sensitive (Fig. 8). This is in accordance with the role of annexins in plant defense. While no annexins have been shown to be secreted by root-knot nematodes, we can propose that MiMIF-2 is involved in protection of nematodes from oxidative stress by fine-tuning the plant annexin function.

Annexins are Ca2+- and phospholipid-binding proteins widespread throughout the animal and plant kingdoms (Gerke and Moss, 2002; Laohavisit and Davies, 2011). Ca2+ is widely seen as a common and simple messenger that acts by coupling extracellular signals to cellular responses and integrating biotic and abiotic information (Dodd ; Newton ; Ranty ). Moreover, the Ca2+ signals act in synergy with ROS and activate MAPK cascades (Gilroy ; Lee and Back, 2016). Annexins regulate abiotic stress-induced calcium signal transduction and mediate biotic stress responses to environmental stimuli (Gorecka ; Vandeputte ; Davies, 2014; Yang ). Calcium signatures also activate various plant enzymes, even altering gene expression, and can, ultimately, modify cells and the calcium signals corresponding to their physiological and biochemical reactions (Tuteja and Mahajan, 2007; Kim ). Peroxide-induced increases in [Ca2+]cyt were recently reported to be aberrant in roots and root epidermal protoplasts from an annAt1-knockout mutant, indicating an essential role for AnnAt1 in the [Ca2+]cyt signature and downstream signaling (Richards ). Moreover, AnnAt1 interacts with AnnAt4, which is modulated by Ca2+ binding (Huh ), and recent studies confirmed AnnAt4 interacted with the calcium sensor SCaBP8, which regulated calcium transients and signatures in Arabidopsis (Ma ). These results inspire the possibility that plant annexins with other proteins may form complexes modulating calcium signaling. It’s possible that MiMIF-2 is involved in the complexes that fine-tune annexin-mediated signal transduction and immune responses. Evidence showed that in rat testicular peritubular cells, MIF signal transduction was based on mobilizing intracellular calcium (Wennemuth ). In human neutrophils, MIF induced calcium influx by acting on two chemokine receptors, CXCR2 and CXCR4 (Bernhagen ). Here, we found that the Ca2+ binding sites of annexins were required for interaction between MiMIF-2 and annexins (Fig. 7). This led us to hypothesize that Ca2+ signal transduction might be impaired in MiMIF-2 ectopically expressing Arabidopsis plants. Consistent with this hypothesis, Arabidopsis expressing MiMIF-2 displayed smaller increases in [Ca2+]cyt following stimulation with H2O2, i.e. a decrease Ca2+ influx (Fig. 9), as previously observed in annAt1 and annAt4 mutants (Richards ; Ma ). Our findings suggest a new model for Ca2+, annexins, and pathogen effectors: the combination of annexins with effectors modulates calcium signaling, interfering with the immune response. Ultimately, our discovery further improves our understanding of effector proteins in M. incognita and identifies new avenues of exploration for studies investigating mechanism by which MIF-like proteins function.

Supplementary data

Supplementary data are available at JXB online.

Fig. S1. Characteristics of MiMIF sequences.

Fig. S2. Western blot analysis detected specificity of MiMIF-2 antibody.

Fig. S3. Ultrastructural immunocytochemistry of M. incognita MIF proteins.

Fig. S4. Localization of MiMIF proteins in tomato gall sections.

Fig. S5. Phenotypes of MiMIF-2 ectopically expressing Arabidopsis and Mimif-2-RNAi lines.

Fig. S6. Verification, by RT-PCR and/or western blotting, of the presence of MiMIF-2 transcripts, MiMIF-2 protein, and annexin proteins in transgenic Arabidopsis.

Fig. S7. MiMIF 2 could not interact with a chloroplasts chaperonin (AT3G13470), heat shock protein 70 (AT5G02490) and heat shock protein 9 (AT5G56010).

Fig. S8. MiMIF-2 could not interact with AnnAt2, AnnAt3, AnnAt5, and AnnAt8.

Fig. S9. Subcellular localization of AnnAt1, AnnAt4, AnnAt1E68A, AnnAt1D299A, AnnAt1E68A/D299A, and AnnAt4E72A.

Fig. S10. Subcellular localization of MiMIF-2, MiMIF-2-mu, and MiMIF-2-113aa.

Fig. S11. Verification of homozygous T-DNA insertion mutants of annAt1 and annAt4.

Fig. S12. RT-PCR verification of gene expression in transgenic Arabidopsis.

Table S1. Primers used in this study.

Table S2. Accession numbers of genes or proteins used in this study.

Table S3. Candidate proteins that interact with MiMIF-2 identified by LC-MS.

Click here for additional data file.