Identification and characterization of a putative protein disulfide isomerase (HsPDI) as an alleged effector of Heterodera schachtii

The plant-parasitic nematode Heterodera schachtii is an obligate biotroph that induces syncytial feeding sites in roots of its hosts. Nematodes produce effectors that are secreted into the host and facilitate infection process. Here we identified H. schachtii protein disulphide isomerase (HsPDI) as a putative effector that interferes with the host’s redox status. In situ hybridization showed that HsPdi is specifically localized within esophageal glands of pre-parasitic second stage juveniles (J2). HsPdi is up-regulated in the early parasitic J2s. Silencing of HsPdi by RNA interference in the J2s hampers their development and leads to structural malfunctions in associated feeding sites induced in Arabidopsis roots. Expression of HsPDI in Arabidopsis increases plant’s susceptibility towards H. schachtii. HsPdi expression is up-regulated in the presence of exogenous H2O2, whereas HsPdi silencing results in increased mortality under H2O2 stress. Stable expression of HsPDI in Arabidopsis plants decreases ROS burst induced by flg22. Transiently expressed HsPDI in N. benthamiana leaves is localized in the apoplast. HsPDI plays an important role in the interaction between nematode and plant, probably through inducing local changes in the redox status of infected host tissue. It also contributes to protect the nematode from exogenous H2O2 stress.


Results
Sequence analysis of HsPDI. A partial sequence of the Heterodera glycines esophageal gland cells putative secretory protein hsp3 (AF273730.1) was used to identify a homologue in H. schachtii. BLASTn results revealed 100% similarity of hsp3 (AF273730.1) with one of the contigs in our draft transcriptome. The contig was extracted and further analysed. A detailed sequence analysis showed that the resulted contig contains a signal peptide of 22 amino acid residues and lacks a transmembrane domain. Further domain analysis with NCBI CD search program revealed four conserved thioredoxindomains (a,b,b' ,a') with two catalytic domains (CGHC), as outlined in Fig. 1A. This arrangement of the domains is a classical feature of the protein disulfide isomerase family members 47 . For that reason, the described sequence was annotated as HsPDI.
Alignment results of HsPDI with other PDIs from different organisms showed high conserved active domains (CGHC) in eukaryotic organisms including some protozoans ( Supplementary Fig. S1). However, overall the highest level of sequence similarity among PDIs was within nematodes (identity >70%). In comparison, similarity was less to the other organisms PDIs (Supplementary Table S1).
Scientific RepoRts | 7: 13536 | DOI: 10.1038/s41598-017-13418 -9 HsPdi is expressed in oesophageal glands during early parasitic stages. We used in situ hybridization to localise the expression of the HsPdi in the pre-parasitic juveniles of H. schachtii. The labelled antisense riboprobe of HsPdi gave a clear signal in the sub-ventral gland cell of the pre-parasitic J2s while no signal was observed in the labelled sense (negative control) (Fig. 1B). To further investigate the expression pattern of HsPdi during different developmental stages of H. schachtii, we used the qRT-PCR using cDNA generated from nematode RNA isolated at different pre-parasitic (eggs and freshly hatched J2s) and parasitic developmental stages (J3s, J4s and young females). HsPdi expression level increased during the sedentary stages of nematode development, reaching its maximum in J3 with 15-fold increase compared with unhatched J2 in eggs. In J4s and young females, expression decreased, respectively, to 2.3 and 1.6-fold compared with unhatched J2s in eggs and hatched pre-parasitic J2s (Fig. 1C).

HsPDI is involved in parasitism.
To analyse whether HsPdi plays a role in parasitism, we performed in vitro RNAi targeting HsPdi (see methods for details). Our results showed that RNAi caused a significant decrease in the transcript abundance of HsPdi in J2s of H. schachtii ( Supplementary Fig. S2). Next we infected the roots of Arabidopsis wild-type plants with nematodes soaked in dsRNA targeted against HsPdi or GFP and counted the number of females and males at 12 DAI. We also measured the average sizes of females, associated syncytia and cysts and counted the average number of eggs per cyst. Our analysis showed no significant difference in number of infecting nematodes found on plants infected with J2s treated with dsRNA targeting HsPdi compared to those treated with dsRNA targeting GFP ( Fig. 2A). However, the average size of syncytia, average size of females, average size of cysts and average number of eggs per cyst were reduced significantly in plants infected with juveniles treated with dsRNA against HsPdi compared with GFP ( Fig. 2B-E).
To characterise further these differences, we examined the anatomical and ultrastructural features of syncytia induced by HsPdi or GFP dsRNA-treated juveniles, and performed a detailed time-course analysis via light and transmission electron microscopy. To obtain comparable materials for these analyses, we took sections in the middle region of syncytia. At 5 DAI, syncytia induced by HsPdi dsRNA-treated juveniles were composed of a similar number of elements as syncytia induced by the GFP dsRNA-treated control juveniles ( Fig. 3A and B). Nevertheless, the hypertrophy of a single syncytial element was lower with HsPdi silenced nematodes. The openings were also observed to be narrower than those found in syncytia induced by GFP dsRNA-treated juveniles. At 10 DAI, the differences between syncytia induced by HsPdi and GFP dsRNA-treated juveniles were less obvious ( Fig. 3C and D), but the number and extent of cell wall openings were lower and the regions with confluent cytoplasm were smaller in syncytia induced by HsPdi dsRNA-treated juveniles.
Parallel examinations of the ultrastructure of syncytial elements revealed developmental abnormalities in syncytia induced by HsPdi dsRNA-treated juveniles. They showed differences in the electron density of the cytoplasm, the organisation and composition of endoplasmic reticulum (ER) and vacuole formation. First of all, cytoplasm electron density was lower in syncytia induced by HsPdi dsRNA-treated juveniles at 5 DAI (syncytia associated with sedentary J2) and 10 DAI (syncytia associated with young females) than in syncytia induced by GFP dsRNA-treated juveniles (Fig. 4A,B,C vs E,F). However, this difference was less pronounced when comparing syncytia at 10 DAI (Fig. 4C,D vs. G,H). In addition, large organelle-free regions were present in syncytia induced by HsPdi dsRNA-treated juveniles at 5 DAI (Fig. 4A,B). At the interface to the organelle-containing cytoplasm, small vesicles or dilated cisterns of accumulated ER appeared (Fig. 4A,B). Besides, the organization and composition of the ER differed strongly between syncytia induced by both groups of juveniles. In control syncytia, numerous cisternae of ER were present (Fig. 4F), whereas they were almost absent in syncytia induced by HsPdi dsRNA-treated juveniles at 5 DAI (Fig. 4C). Interestingly the total number of ER cisternae decreased during syncytia development in both groups. At 10 DAI they were still quite numerous in syncytia induced by GFP dsRNA-treated juveniles (Fig. 4G,H), but completely absent in syncytia induced by HsPdi dsRNA-treated juveniles (Fig. 4C,D). The ER system in syncytia induced by HsPdi dsRNA-treated juveniles consisted predominantly of tubular ER that occupied large regions of syncytial cytoplasm (Fig. 4A,C). Tubular ER was present also in syncytia induced by GFP dsRNA-treated juveniles, but never appeared alone in any syncytial element ( Fig. 4E-H). Thirdly, syncytia induced by HsPdi dsRNA-treated juveniles were strongly vacuolated at 10 DAI (Fig. 4C,D). Organelles such as nuclei, plastids or cell walls, displayed no structural changes ( Fig. 4A-H), and their ultrastructure was typical as described for syncytial elements 4,48 . To analyse the function of HsPdi in more detail, we produced transgenic Arabidopsis plants over-expressing HsPDI (35S::HsPdi-GFP). The relative expression of HsPdi in transgenic lines was measured using qRT-PCR ( Supplementary Fig. S3). A detailed phenotypic analysis did not reveal any significant differences in plant growth as indicated by number of lateral roots, length of main root, fresh root weight and fresh shoot weight between HsPDI expressing plants and Col-0 ( Supplementary Fig. S4). Next, we analysed these lines for susceptibility via nematode infection assay and found that at least one of the HsPDI expressing lines was more susceptible to H. schachtii infection as compared with wild-type control (Fig. 5A). However, considering significantly higher average female, cyst, and syncytium sizes, both HsPDI expressing lines appear to be more susceptible than the wild type Col-0 plants (Fig. 5B,C). Furthermore, the average number of eggs were also increased significantly compared with the control (Fig. 5D).

HsPdi expression is triggered by H 2 O 2 and increases H 2 O 2 tolerance. We tested whether treating
J2s with H 2 O 2 influence the survival of nematodes. To test vitality, pre-parasitic J2s were soaked in 5, 10 or 25 mM H 2 O 2 and the percentage of dead juveniles was evaluated after 15, 30, 45, and 60 min. We found that J2s can survive up to 30 min in 10 mM H 2 O 2 without significant increase in mortality rate (Fig. 6A), but prolonged treatment substantially increased mortality of J2s. Next, we used qRT-PCR to analyse the expression of HsPdi in response to H 2 O 2. By analysing the juveniles that were exposed to H 2 O 2 (5 or 10 mM) for 30 min, we found a significant increase in transcript abundance of the HsPdi when compared with water-treated control J2s (Fig. 6B).
We examined the mortality rate of HsPdi or GFP dsRNA-treated J2s after soaking them in 5 mM H 2 O 2 for 30 min. As a control, we incubated dsRNA-treated J2s in water (Fig. 6C). We observed a significantly lower percentage of J2s that survived in H 2 O 2 in case of HsPdi dsRNA-treated nematodes, indicating that HsPdi plays a role in protecting J2 from the impact of the exogenous H 2 O 2 . Following up on these results, we investigated whether HsPDI could modulate the plant endogenous ROS burst. We incubated the plants leaf discs in presence of the bacterial peptide flg22 and total ROS burst was measured in HsPDI expressing transgenic plants and wild type Col-0. We observed a significant decrease in total ROS in both transgenic plant lines expressing the HsPDI to half of the total amount of ROS produced by Col-0 plants (Fig. 6D, Supplementary Fig. S5), while no significant changes were found in the transgenic plant lines expressing Hs-Tyr compared with Col-0 ( Supplementary  Fig. S6).

HsPDI is localised in the plant apoplastic space.
To determine the sub-cellular localization of the HsPDI protein inside the host cells, we transiently expressed HsPDI::GFP in Nicotiana benthamiana epidermal cells with the constitutive cauliflower mosaic virus (CaMV) 35S promoter and assessed its localization by co-infiltrate it with an apoplastic marker under the confocal microscope ( Fig. 7A-D). Our observations clearly showed that GFP signal was localized in the cell periphery and co-localized with the mCherry signal. To further investigate the specificity of subcellular localization, we induced plasmolysis by adding 1 M NaCl to the leaf tissue. Upon dissociation of plasma membrane from the cell wall, the signal was observed in the apoplastic space indicating the localization of HsPDI in outer cell periphery (Fig. 7E-H).

Discussion
Plant endo-parasitic nematodes spend most of their life cycle inside host tissue. To establish parasitism, nematodes have evolved a repertoire of physical and chemical tools including secretions of proteinaceous and non-proteinaceous effectors into the host tissues 8,49 In the current work, we identified from H. schachtii draft transcriptome a novel putative effector gene (HsPdi) encoding a protein with a protein disulfide isomerase domain. We demonstrated its importance in nematode parasitism and ROS detoxification. HsPDI encodes a 22-amino acid, signal peptide at its N-terminal and lacks a transmembrane domain. Moreover, four thioredoxin domains and two active catalytic motifs were also detected (CGHC and CGHC) (Fig. 1A). Being a typical PDI protein, HsPDI contains the main structural building block. It also contains a-type domains containing two cysteines in a CXXC active-site motif with an intervening GH sequence, which is the most common CGHC motif in the PDIs 47 . Presence of a signal peptide and lack of a transmembrane domain supported the role of HsPDI as a putative effector 50,51 . Previous studies have shown that the active motifs are essential for the protein function as mutations in the active site domains result in the loss of protein functions 52 . It was also shown that, as a virulent factor, mutating the active motifs of the PpPDI1 abolished necrosis-inducing activity of the oomycete plant parasitic Phytophthora parasitica, indicating that the cell death-inducing function might be related to the catalytic properties 46 . Here we showed that the active catalytic motifs are highly conserved in the tested sequences from various eukaryotes (Supplementary Fig. S1).
The observation that transcript for HsPdi was localized in oesophageal gland cells supports the hypothesis that the HsPDI protein is secreted into the host tissues to facilitate parasitism 8,49 . Furthermore, our expression analysis found that transcript abundance for HsPdi was increased significantly during early stages of infection reaching its maximum at 5 DAI, which coincides with rapid enlargement of the nematode induced syncytium 4 . This particular expression pattern points towards the importance of HsPDI in the early stages of infection including syncytium formation and maintenance. This hypothesis is further supported by results from infection assays where silencing HsPdi expression via RNAi led to impaired nematode development and ultrastructural and anatomical abnormalities in associated syncytium. Additionally, the described set of ultrastructural features strongly resembles the ultrastructure of syncytia associated with developing male juveniles 48 . This suggests that syncytia induced by HsPdi depleted-nematodes may suffer from a shortage of nutrients, which can lead to smaller females with a lower numbers of eggs. Earlier studies suggest that in H. schachtii (i) sex determination is regulated epigenetically by the composition and amount of nutrients withdrawn from syncytia 53 , (ii) sex differentiation occurs during the sedentary late J2 stage 3 . Therefore we conclude that HsPdi dsRNA-treated juveniles can induce fully functional syncytia that support their development into females during J2 sedentary stage. Afterwards these syncytia start to develop structural abnormalities. Theses abnormalities are similar to ultrastructural features of syncytium senescence typically occurring in degrading syncytia associated male J4 and adult males which had ceased feeding. The impaired function of their syncytia makes associated females developing smaller and producing fewer eggs.
It has been shown previously that ROS oxidize DNA, proteins, and lipids, which causes damage to the cellular organelles and inhibits cell functions 54 . Plant-parasitic nematodes encodes various antioxidant enzymes, such as superoxide dismutase (SOD), catalase, ascorbate, p-phenylenediamine-pyrocatechol (PPD-PC), o-dianisidine, guaiacol isoperoxidases, peroxiredoxins and glutathione peroxidases, which are important for parasitism and could have protective function against ROS 27,31,37,55 . The fact that HsPdi expression was elevated in presence of exogenous H 2 O 2 and that silencing the expression of HsPdi using dsRNA decrease the tolerance of nematodes to 5 mM H 2 O 2 points to the role of HsPdi to protect the nematode. These observations are supported by previous studies where a similar expression pattern was observed for M incognita peroxiredoxins. Silencing peroxiredoxins expression of M incognita impaired the nematode infectivity on tomato and their tolerance to exogenous H 2 O 2 31 . We therefore conclude that HsPdi plays a dual role protecting nematodes against ROS and modulating the apoplastic redox status in the infected host tissue.
Our data for subcellular localization showed that HsPDI (without signal peptide) is localized in apoplast; however, this observation also raises the question of how an effector that is putatively secreted into the cytoplasm of infected tissues is ultimately translocated to the apoplastic space. Although the exact mechanism is currently unknown, it is plausible that host trafficking machinery is manipulated to deliver effectors into the apoplastic space 56,57 .
In plants, apoplastic ROS are actively produced through the action of NADPH oxidases and class III peroxidases, but the biological significance and the mechanism by which these ROS are scavenged during the plantnematode interaction are not well understood. Recently, it was shown that H. schachtii infection of Arabidopsis plants stimulates ROS burst via host NADPH oxidases (RbohD and RbohF). Surprisingly, knocking out RbohD and RbohF restricted nematode development and nurse cell formation and triggered massive cell death upon nematode infection 25 . Combining the previous studies with our result we conclude that although ROS are needed for successful nematode infection, they are harmful for the plants and nematodes when exceeding a certain level. Therefore, nematode not only induce ROS burst, but they also manipulate their levels for optimal infection and less plant damage. HsPDI is proposed as one of the effectors which may serve this purpose. Alternatively, it is possible that Rbohs are guarded by a nucleotide-binding and leucine-rich repeat protein (NLR), leading to a strong immune response in mutants deficient in Rboh genes upon infection 9,58 . In conclusion, our results strongly indicate that HsPDI is a nematode effector that is secreted into host tissues and become a part of the host antioxidant mechanisms as plant ROS scavenger not only during invasion but also during sedentary parasitism. Clarifying further details of the interplay between various ROS-producing and ROS-scavenging systems during plant-nematode interaction will provide exciting information on nematode parasitism.

Methods
Plant growth and nematode culture. Arabidopsis thaliana L. Heyn. ecotype Col-0 plants and two transgenic lines expressing HsPDI were grown aseptically on agar medium supplemented with modified 0.2 Knop's nutrient solutions for 16 h light and 8 h dark at 25 °C as described previously 2 .
Heterodera schachtii Schmidt used in the experiments was reared on white mustard (Sinapis alba L. cv. Albatros) plants which were grown aseptically on agar containing 0.2% Knop medium. Mature cysts were collected in funnels and hatched in 3 mM ZnCl 2 2 . The hatched pre-parasitic J2s were collected and used in the experiments.
Infection assay. Nematode infection assays on Arabidopsis plants either for the nematode RNA interference (RNAi) experiments or on HsPDI-expressing lines were performed as described previously 25 . Briefly, sterilised seeds were placed on 0.2% Knop medium. After 10 days, roots were inoculated with 60-70 J2s per plant. For each experiment, 12 plants per treatment were used. Numbers of adult males and females were counted per plant at 12 days after inoculation (DAI). Furthermore, sizes of females and associated syncytia were measured after 13 DAI, sizes of cysts and numbers of eggs per cyst were examined at 45 DAI using Leica M165C Binocular (Leica Microsystems, Wetzlar, Germany) and Leica Application Suite software. Experiments were repeated three times and analysed using Student's t-test. Sequence analysis. HsPdi (KU948160) was identified by performing a BLASTn of the esophageal gland cells putative secretory hsp3 (AF273730.1) which was isolated from Heterodera glycines 50 against draft transcriptomic data of H. schachtii. The draft transcriptome was generated by using next generation sequencing (Illumina, 100 bp paired end reads) of pre-parasitic J2s. The assembly was done using CLC genomics workbench after trimming the reads for adapter sequence and low quality nucleotides (less than 0.05) and ambiguity nucleotides (more than 2 adjacent ambiguous nucleotides).
The deduced protein was analysed to predict its functional domain(s). The conserved domains search was performed using the National Centre for Biotechnology Information NCBI CD search program (http://www.ncbi. nlm.nih.gov/Structure/cdd/wrpsb.cgi), signal peptide was identified using signalP4 server 59 , and transmembrane domains using TMHMM algorithm 60 .
In situ hybridization. Digoxigenin (DIG)-labelled probes complementary to identified HsPdi fragments were amplified in asymmetric PCR with single sense (negative control) or antisense primer and DIG-labelled deoxynucleoside triphosphates (dNTPs) (Roche) in the reaction mixture (see Supplementary Table S2). The hybridization was performed on the pre-parasitic J2s of H. schachtii at temperature of 47 °C as performed in previous study 61 . The hybridized nematodes were examined using Leica DMI2000 compound microscope.
Developmental expression pattern analysis. Transcription of HsPdi was analysed in different developmental stages (eggs, pre-parasitic J2, parasitic juveniles and females) by quantitative PCR (qRT-PCR) using specific primers (Supplementary Table S2). Around 3,000 eggs and 3,000 pre-parasitic J2s per biological replicate were collected directly from hatching funnels 2 . Around 500-600 nematodes per biological replicate were collected manually from nematode infected roots of A. thaliana at 5, 10, 15 DAI, what corresponds to J3, J4, and young female stages.
Total RNA of the biological replicates was extracted using NucleoSpin RNA kit (MACHEREY-NAGEL) following manufacturer's protocol. Quality and quantity of the extracted RNA samples were tested using the Agilent 2100 Bioanalyzer system (Agilent Technologies). The RNA with RNA integrity number (RIN) value higher than 8.5 was used for first strand cDNA synthesis using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems) and oligo-dT primer. The resulted cDNAs were tested for the expression changes using the Stepone Plus Real-Time PCR System (Applied Biosystems) with 95 °C for 15 s and 60 °C for 30 s (40 cycles) for amplification. Each sample contained 10 μl of Fast SYBR Green qPCR Master Mix (Invitrogen), 9 μl of the specific primer mixture with final concentration 1 μM for each primer, 1 μl of cDNA. The amplified data were analysed by using the Stepone Plus Real-Time PCR software to create Ct values. The data were analysed and relative expression was calculated following Pfaffl 62 . Actin was used as internal control for all experiments (Supplementary Table S2). Three biological replicates from each stage in three technical replicates for each biological replicate were used.
Double stranded RNA (dsRNA) and gene silencing in nematodes. HsPdi dsRNA was synthesized using RiboMAX RNA Large-Scale Production System (Promega) according to the manufacturer's instruction. The forward and reverse primers (Supplementary Table S2) were supplemented with the SP6 and T7 promoter sequences and used for dsRNA synthesis. The GFP template was used for synthesis of a dsRNA construct that was used as a negative control.
About 10,000 freshly hatched J2s of H. schachtii were soaked in 50 µL of dsRNA incubation mixture consisting of 25 µL HsPdi-dsRNA or GFP-dsRNA (2 mg*mL −1 ) in 5 µL of 10x M9 buffer (55 mM KH 2 PO 4 , 21 mM NaCl, 47 mM NH 4 Cl) supplemented with 100 mM spermidine (1.5 µL), 500 mM octopamine (5.0 µL) and 13.5 µL (nematode suspension) for one day. Incubated J2s were washed three times in H 2 O, sterilized in 0.05 M HgCL 2 for 2 minutes and washed again three times with sterile water. Afterwards, the batch of juveniles was divided equally and one part was used for plant infection assay whereas the second part was used to evaluate the level of HsPdi silencing using qRT-PCR as described above.
Ten days old A. thaliana Col-0 roots were inoculated with 60-70 J2s incubated in HsPdi or GFP dsRNA. Numbers of developed females and males were counted after 12 DAI, sizes of females and syncytia associated with females were measured at 13 DAI, whereas sizes of cysts and numbers of eggs per cyst were examined at 45 DAI. All measurements were conducted under a Leica M165C stereo microscope using manufacturer's software. Experiments were repeated three times with 12 plants per treatment in each experiment. The obtained data were merged and analysed using the Student's t-test.
Syncytium anatomy and ultrastructure. Segments of roots containing nematode induced syncytia were dissected from Arabidopsis Col-0 plants inoculated with HsPdi or GFP dsRNA-treated juveniles at 5 and 10 DAI. They were processed for light and transmission electron microscopy as described 63 . Semi-thin sections (3 μm thick) taken on a Leica RM2165 microtome (Leica) were stained with hot 1% (w/v) aqueous solution of Crystal violet (Sigma, St. Louis, MI, USA) for 1 min at 65 °C. They were examined with an Olympus AX70 'Provis' (Olympus, Tokyo, Japan) light microscope equipped with an Olympus DP50 digital camera. Ultra-thin sections (70-80 nm thick) taken on a Leica UCT ultramicrotome (Leica Microsystems) were collected and on formvar-coated (Fluka, Buchs, Switzerland) single-slot copper grids and stained with uranyl acetate (Fluka) and lead citrate (Sigma) 4 . They were examined with an FEI 268D 'Morgagni' transmission electron microscope (FEI, Hillsboro, OR, USA) operating at 80 kV. The images were recorded with an SIS 'Morada' digital camera (Olympus SIS, Münster, Germany) at 10 Mpix resolution. The images were equalized for similar contrast and brightness, resized and cropped using Adobe Photoshop graphic software. stress, J2s were incubated for 30 min in 5 or 10 mM H 2 O 2 , and washed in sterile tap water. J2s incubated in sterile tap water were used as a control. RNA was extracted, cDNA was generated and expression of HsPdi was quantified by qRT-PCR as described above. Three biological replicates were carried out and each one was replicated three times.

Effect of HsPdi depletion on nematode survival under the H 2 O 2 stress. Around 150 freshly hatched
J2s were incubated overnight in solution of HsPdi dsRNA and in GFP dsRNA as described above and then washed in tap water. Afterwards, the J2s were incubated in 5 mM H 2 O 2 or in tap water as a control for 30 min. The numbers of dead J2s were counted and their percentage was calculated and analysed using Student's t-test. Each treatment consisted of 4 replicates and the experiment was repeated 4 times.

Construct generation and production of transgenic HsPDI expressing plants. The ORF of the
HsPdi with no signal peptide was cloned into the binary Gateway over-expression vector pMDC83 to obtain a C-terminal fusion with GFP 64 using the primers listed in Table S2. The construct was transferred to Agrobacterium tumefaciens GV3101::pMP90 strain 65 and transformed into A. thaliana Col-0 using the floral dip method 66 . The transformed plants were selected for hygromycin resistance on modified 0.2 Knop medium and grown for 3 generations to obtain homozygous lines for infection assays. The expression of HsPdi in homozygous lines was confirmed using qRT-PCR as described previously 62 .
The homozygous lines were grown on Murashige and Skoog media (MS) plates for 10 days. The number of lateral roots, main root length, fresh root weight and fresh shoot weight were measured and compared with Col-0. The experiment was repeated 3 times and each experiment consists of 10 plants for each line. After phenotyping, plants were infected with H. schachtii as described above.

Agroinfiltration and subcellular localization of HsPDI. A. tumefaciens transformed with
HsPdi::pMDC83 construct was grown overnight in 50 ml of YEB liquid medium 65 supplemented with 10 mg*mL −1 gentamycin, 50 mg*mL −1 kanamycin and 50 mg*mL −1 rifampicin to an OD600 of 0.8 in an incubator/shaker at 28 °C. Bacteria were pelleted then they were re-suspended in an infiltration buffer 65 . Bacterial suspensions were diluted with the infiltration buffer to get the required OD600 of 1. After incubation for 4 h at RT, the transformed bacteria were injected in the leaves abaxial side of 6 week-old Nicotiana benthamiana plant using 1 mL hypodermic syringe without needle. For co-infiltration of RNA silencing inhibitor P19 and the apoplastic marker, equal volumes of a bacterial suspensions harbouring pBin61::P19 67 , HsPdi::pMDC83 and apoplastic marker constructs were mixed and injected. To perform co-localization experiments in the apoplast, the N-terminal region of a membrane-localized receptor-like-kinase (At4g31250) was amplified and cloned in frame with mCherry under the control of the 35 S CaMV promoter and terminator. The cloning of only N-terminal region assured the delivery of fusion protein into extracellular region. The complete expression cassette was further cloned with restriction enzyme SdaI into the binary vector pGreenII 68 resulting in a pGreen-apo-mCherry marker. The primers sequences are given in Table S2. Infiltrated plants were incubated in the growth chamber (16hrs light, 8hrs dark and at 25 °C) for 6 days. Slides were made from the infiltrated leaves and examined for the presence of fluorescence signal using Zeiss CLSM 710 with and without plasmolysis with 1 M NaCl 2 . ROS measurement in HsPDI over-expressing plants. ROS production in Col-0 and transgenic HsPDI over-expressing plants after treatment with bacterial peptide flg22 (100 nM) was measured on leaf disc samples using luminol-based assay and 96 wells plate luminometer Mithras LB 940 (Berthold Technologies) as described previously 69 . Light emission was measured in relative units over 120 min long incubation period and analysed Scientific RepoRts | 7: 13536 | DOI:10.1038/s41598-017-13418-9 using instrument software. ROS production was checked in transgenic plants expressing the nematode Hs-Tyr 70 and compared with Col-0 as a control. Data was tabulated and analysed statistically using Student's t-test. Four leaf discs were used per line as technical replicates. The experiment was repeated three times. Total ROS production was calculated and presented.