Background

Soil-borne plant parasites, members of the genus Meloidogyne (root-knot nematodes, RKNs) are some of the most yield-limiting plant-parasitic nematodes (PPNs). On a global scale, PPNs cause huge economic losses in the agricultural sector with annual losses estimated at $US157 billion [1, 2]. To date, approximately 100 RKN species are documented. The four dominant globally distributed RKN species are M. arenaria, M. javanica, M. incognita, and M. hapla. These highly polyphagous nematodes are capable of infecting nearly all higher vascular plants including economically important crops such as potato, tomato, maize, rice, tobacco, sweet potato, pineapple, and woody perennial plants [1, 3].

Initial infection of any parasitic organism is challenged by the host defense system. However, these parasites have devised ways to dampen the host immune system and mechanisms for their survival [4]. Parasites produce a diverse range of excretory-secretory products (ESPs, also known as secretome) during the infection process to interact with the host proteins. The secretome constitutes functionally distinct molecules including digestive and detoxification enzymes, toxins, proteases, serpins, as well as small RNAs that can imitate host miRNAs [4, 5]. The ESPs are the main actors that mediate successful interactions at the host-parasite interface [6]. The ESPs are essential for parasite survival and are involved in various cellular activities including host invasion, nutrient acquisition, parasite development, evasion, and modulation of the host immune system [5].

As effective management strategies against RKN species diminish and become obsolete, there is a need for the development of new environmentally benign approaches [7, 8]. Identifying so far unexplored nematode ESPs and understanding their mechanism in host cells unfolds the possibility of developing novel control strategies against nematode pests.

In the genus Meloidogyne, many secretory proteins have been identified and studies geared towards deciphering their mechanism in plant parasitism are developing at an increasing pace [9]. Candidate secretory proteins have been identified from proteomes, genomes, and/or transcriptome analyses. For instance, mass spectrometry analysis has been used directly to identify 486 M. incognita secretory proteins [10]. Previously, [11] and [12] mined 37 and 18 putative effectors, respectively from M. incognita cDNA libraries of esophageal glands.

Global transcriptomic analysis of pre-parasitic and parasitic phases of M. graminicola identified putative secretory proteins [13]. Moreover, dual RNA sequencing analysis, a tool that can simultaneously profile both host and pathogen transcriptional changes during the pathogenicity, led to the discovery of M. incognita secretory proteins in a stage-specific manner [14, 15]. Secretome analysis using computational tools identified 1,886 secretory proteins encoded in M. incognita genome [16]. In addition, phylogenomics in conjunction with RNA-seq analysis allowed the discovery of 68 phytoparasite-specific proteins strongly expressed during the endophytic stages of M. incognita [17].

In view of the above, it is evident that previous studies have largely focused on identification of secretory proteins from M. incognita using in silico tools, proteomics, and genome-wide approaches. This could be due to the availability and accessibility of M. incognita genome sequences earlier than other RKN species [18], its highly polyphagous nature, and its classification as the most damaging crop pest worldwide [2]. Now, with the accessibility of genome sequences from other important Meloidogyne spp. including M. arenaria, M. javanica, M. hapla, M. graminicola, M. floridensis, M. luci, M. exigua, and M. enterolobii [19,20,21,22,23,24] genome-wide analysis, and/or comparative genomics are plausible to reveal the common mechanism that drive RKN infections and potentially guide towards identification of new control strategies specifically directed against Meloidogyne spp.

The availability of M. javanica genome sequences allowed us to explore its secretome systematically and comprehensively using bioinformatics approaches. Indeed, computational tools have been used widely in the discovery of putative ESPs in PPNs and animal parasitic nematodes (APNs) [16, 25,26,27]. Here, we identified classically and non-classically secreted proteins, followed by functional annotation and in-depth insights into the predicted M. javanica secretome. The present study not only adds to the secretory proteins repertoire from M. javanica but also provides a resourceful biological database that lays a framework for future functional characterization of these putative M. javanica secretory proteins. In addition, this research provides a foundation for understanding the interaction between RKNs and their host plants.

Results

In silico identification of putative secretory proteins from the M. javanica genome

Nematode secretions released by the infective juveniles are crucial for a successful infection process and participate in various stages of nematode parasitism [28]. Here, we applied various bioinformatics tools to predict M. javanica secreted proteins. Of the 97,208 proteins encoded in the M. javanica genome, 5,258 and 2,369 were identified as classically and non-classically secreted proteins, using SignalP and SecretomeP softwares. The 7,627 secreted proteins identified were without transmembrane domains according to TMHMM and Phobius analysis. TargetP analysis identified 11 proteins as mitochondrial-target proteins, and these were discarded.

Further, PS-SCAN was used to scan endoplasmic reticulum-based proteins while PreGPI detected secretory proteins but linked to the membrane by a GPI anchor. In total, 26 proteins with endoplasmic reticulum retention signal and 132 proteins predicted to be GPI anchored proteins were removed from the data set. Thus, the bioinformatic pipeline predicted a total of 7,458 (7.7%) of the entire proteome as the secretome, and these were used for further analysis (Fig. 1, Additional file 2: Table S1). Overall, 1,269 (17%) predicted secreted proteins (Additional file 2: Table S2) matched to the NCBI dbEST database (https://www.ncbi.nlm.nih.gov/nuccore/?term=Meloidogyne+javanica). This implies that these proteins are expressed during the nematode infection cycle and may be involved in nematode parasitism.

Fig. 1
figure 1

Overview of bioinformatic pipeline to predict putative secretory proteins from the genome of root-knot nematode, Meloidogyne javanica (See the materials and methods section for detailed criteria for secreted proteins prediction)

Full size image

The size distribution analysis of the predicted secreted proteins showed that 50.1% (3,739) of the proteins were between 101 and 300 aa, followed by proteins with a sequence length of below 100 aa accounted for 32.4% (2,413 proteins). The remaining secretome is made up of 906 (12.1%) proteins with sequence lengths between 301 and 500 aa, 344 (4.6%) proteins between 501 and 1000 aa, and 56 (0.2%) proteins with more than 1000 aa (Additional file 1: Figure S1). The large proportion of the predicted secretome constituted small-sized proteins that can be easily released to the extracellular medium of the parasite where they mediate interactions with their host.

We searched for the Mel-DOG promoter motif (Meloidogyne DOrsal Gland; TGCMCTT) identified in the promoter of Meloidogyne incognita dorsal gland encoding effectors [29] within 2 kbupstream of the coding region of genes of the predicted M. javanica secretome. The core motif was detected in the 2 kb upstream of the coding region of 1,608 secretory proteins out of the 7,256 promoter regions retrieved from WormBase database (Additional file 2: Table S3), adding to the criteria for identifying RKN effectors.

Functional annotation of M. javanica secretory proteins

Gene ontology (GO) analysis

Out of the 7,458 secreted proteins, 24.1% (1,797) were assigned to 3,513 GO terms with 1,044 GO terms from biological process category, 1,216 GO terms from molecular function and 1,253 GO terms from cellular component category. A summary of the GO term annotation is shown in Fig. 2, Additional file 2: Table S4.

Fig. 2
figure 2

Gene ontology distribution of secretory proteins from Meloidogyne javanica

Full size image

GO terms such as cellular (GO:0009987; 43.6%), and metabolic process (GO:0008152; 47.2%), biological regulation (GO:0065007; 26%), regulation of biological process (GO:0050789; 23.8%), developmental process (GO:0032502; 13.9%) and response to stimulus (GO:0050896; 9.5%) were the most represented GO terms under biological process category (Fig. 2, Additional file 2: Table S4).

The top represented GO terms at the molecular function category level were catalytic activity (GO:0003824; 35.7%) binding (GO:0005488; 39.3%), and molecular function regulator (GO:0098772; 5.8%) (Fig. 2). The most dominant sub-categories in catalytic activity terms were hydrolase activity (GO:0016787; 20.1%), catalytic activity action on a protein (GO:0140096; 14.4%), and transferase activity (GO:0016740; 8%). In binding term, ion (GO:0043167; 29.4%), heterocyclic compound (GO:1,901,363; 20.4%), organic cyclic compound (GO:0097159;20.4%), protein (GO:0005515; 5%), drug binding (GO:0008144; 5.6%), small molecule (GO:0036094; 6.1%) and carbohydrate derivative binding (GO:0097367; 5.8%) were the most predominant sub-categories (Additional file 1: Figure S3). Lastly, enzyme regulator activity (GO:0030234; 4.9%) was the abundant sub-category in molecular function regulator terms.

At the cellular component category, membrane (GO:0016020; 50.9%), membrane part (48.9%) cell (GO:0005623; 31.6%), cell part (GO:0044464; 31.6%) organelle (GO:0043226; 24.2%), and extracellular region (GO:0005576; 16.5%) were found to be highly represented GO terms within this category (Fig. 2).

Gene ontology enrichment analysis

GO enrichment analysis was performed using the WEGO genomics tool to detect significantly enriched GO terms in the predicted secretome compared to the total proteome at a P-value of 0.05. In the cellular component category, cell surface, extracellular region, extracellular region part, extracellular space, extracellular matrix, transporter complex, and receptor complex were found to be significantly enriched terms (Fig. 3, Additional file 1: Figure S3). Carbohydrate binding, molecular function regulator, catalytic, activity acting on proteins, and enzyme regulator activity were over-represented in the molecular function category. Finally, under biological process category, anatomical structure development, response to biotic stress stimulus, regulation of metabolic process, regulation of molecular function, response to other organisms, biological and cell adhesion, secondary metabolic process, and pigment metabolic process showed significant enrichment in the predicted secretome in comparison to the distribution of the same GO terms in the total proteome (Fig. 3, Additional file 1: Figure S3).

Fig. 3
figure 3

Gene Ontology enrichment analysis of Meloidogyne javanica secretome in comparison to the total proteome. This figure only displays secretome GO terms that are significantly enriched at (P ˂0.05)

Full size image

Identification of peptidases and peptidase inhibitors in M. javanica secretome

Peptidases are hydrolytic enzymes found in parasitic nematodes. These enzymes are important virulence factors that influence nematode-host interactions and facilitate penetration of host tissues, sustenance through host protein digestion, parasite growth and survival, and immune system evasion [30]. Here, 276 (3.7%) secreted proteins were identified as peptidase enzymes representing the five major catalytic classes belonging to 43 families presented in Fig. 4, Additional file 2: Table S5. These include 65 cysteines (C), 104 serine (S), 70 metallo (M), 17 aspartic (A), and 10 threonines (T) peptidases distributed into 43 families. The most abundant peptidases were C1A, S9, S10, S1A, and M12A. Moreover, 248 (3.3%) proteins were annotated as peptidase inhibitors (I) belonging to 18 families in the predicted M. javanica secretome. The most represented families include I2 (Kunitz inhibitors, 50 proteins) I63 (90 proteins), I35 (19 proteins), and I43 (25 proteins) among others (Fig. 4, Additional file 2: Table S4).

Fig. 4
figure 4

Distribution of peptidases and peptidase inhibitors identified in Meloidogyne javanica secretome

Full size image

Analysis of cell wall degrading enzymes (CWDEs)

Cell wall degrading enzymes are a major category of PPNs secretome that mediate compatible interactions [31]. Here, 146 secretory proteins were mapped into 31 families representing five main functional classes of CWDEs including glycoside hydrolases (GH, 90 proteins), glycosyltransferases (GT, 19 proteins), pectate lyases (PL, 36 proteins), and carbohydrate-binding modules (CBM, 1 protein) shown in Additional file 1: Figure S4, Additional file 2: Table S5. The most represented Carbohydrate-Active Enzymes (CAZymes) sub-families include GH5-2 (27 proteins), GH30 (9 proteins), GH28 (6 proteins), GH18 (10 proteins) PL3-3 (21 proteins), and PL3 (13 proteins) (Additional file 1: Figure S4 and Additional file 2: Table S5).

Protein domain analysis

Using InterProScan domain search, 1915 (25.7%) of the predicted secreted proteins were mapped into 412 known protein domains and families. The most abundant InterProScan domains summarized in Table 1, Additional file 2: Table S7 include protein kinase domain, peptidase C1A papain C-terminal (cysteine peptidase family C1, sub-family C1A (papain family, clan CA), C-type lectin-like domain, Transthyretin-like, CAP (cysteine-rich secretory proteins) domain, Pancreatic trypsin inhibitor Kunitz domain, Modulator of levamisole receptor-1, Peptidase S10 serine carboxypeptidase, Pectate lyase PlyH/PlyE-like, Glycoside hydrolase, family 5, and DnaJ domain.

Table 1 Top 20 highly represented protein domains
Full size table

Metabolic pathway analysis

Out of the 7,458 predicted secretory proteins, 672 (9.0%) were mapped into 470 KEGG pathways presented in Additional file 2: Table S8. The most abundant KEGG pathways displayed in Table 2 include lysosome, protein processing in endoplasmic reticulum, autophagy, apoptosis pathways of neurodegeneration-multiple diseases, pathways in cancer, antigen processing, and representation, and calcium signaling pathway. In addition, KEGG functional hierarchies revealed molecular entities linked to membrane trafficking, exosome, peptidases and peptidase inhibitors, chaperones and folding catalyst, transcription factors, protein kinases, chromosome, and associated proteins and spliceosome were found to be highly represented (Additional file 2: Table S9) which are essential for the parasite’s existence inside the host [16].

Table 2 Top 20 highly represented KEGG pathways
Full size table

Identification of Meloidogyne javanica species-specific secretome and discovery of potential targets for functional analysis

The 7,458 predicted secreted proteins of M. javanica were searched for homologues in other closely related Meloidogyne species. From this comparative analysis, the tropical Meloidogyne spp. shared 323 orthologous groups with 1,737 proteins which can be explored further to uncover typical mechanisms RKNs use to undermine host immunity and promote disease processes. A set of 599 clusters with 1,684 proteins was specific to M. javanica secretome (Fig. 5, Additional file 1: Figure S5).

Fig. 5
figure 5

 A Venn diagram depicting the orthologous protein clusters shared in the secretome of four root-knot nematode species

Full size image

A further BLAST analysis of the 1,684 proteins against the NCBI non-redundant proteins database showed that the majority of these proteins had homologues to M. graminicola and M. enterolobii suggesting that RKN-secreted proteins are highly conserved within the genus. Among these, a total of 382 proteins were identified as M. javanica species-specific secreted proteins since there were no homologous proteins in the NCBI non-redundant protein database (Additional file 2: Table S10). According to Argot functional prediction analysis [32], 63 protein sequences out of 382 proteins were annotated with 123 different GO terms (Additional file 2: Table S11). The REVIGO tool was used to obtain of non-redundant GO terms which were classified into functional groups [33]. Major biological processes associated M. javanica species-specific secreted proteins include spliceosomal complex assembly, carbohydrate metabolic process, cell death, defense response, protein ubiquitation, DNA modification, regulation of DNA-templated transcription, microtubule-based process, negative regulation of Wnt signaling pathway, homophilic cell adhesion via plasma membrane adhesion molecules and positive regulation of siRNA production. The major molecular function categories include transcription cis-regulatory region binding, DNA binding, DNA-binding transcription factor activity, actin binding, aspartic-type endopeptidase activity, hydrolase activity, hydrolyzing O-glycosyl compounds, hydrolase activity, acting on glycosyl bonds, lipid binding, N-methyltransferase activity, protein-containing complex binding, and collagen binding. The major cellular component represented include extracellular region, intracellular anatomical structure, spliceosomal complex, nuclear pore, nucleus, cytoskeleton, external enscapulating structure and other organism part (Additional file 2: Table S12).

Of the 382 secretory proteins, 36 secretory proteins were supported at the transcriptional level (EST mapping, Additional file 2: Table S10). In addition, 50 predicted secreted proteins had the Mel-DOG promoter motif (Additional file 2: Table S10) which is associated with effector proteins expressed within the nematode dorsal oesophageal glands [29]. Taken together, these secreted protein sets constitute the best candidates for functional studies, as revealed by our analysis of their likely involvement in M. javanica infection process. The host target repertoire of this effectome can be uncovered by further functional research, which will unvarel the virulence mechanism and host range susceptibility of M. javanica species.

We subjected the 382 protein sequences to a fold recognition analysis using the Phyre2 algorithm to obtain additional functional information for this set of specific secreted proteins. The Phyre2 tool determines the structure and function of novel protein sequences using advanced homology detection algorithms [34, 35]. In this case, if the predicted structure of the query protein is derived with high confidence, the template protein function’s can be tentatively attributed to the query protein. Table 3 shows the proteins that attained confidence scores of 50% which was the minimum cut off value. The protein M. Javanica_Scaff8743g053112 had the highest structural similarity to repeat five residue (rfr) protein whose function is yet to be known.

Table 3 Phyre 2 confident prediction discovered specific in the secretome of M. javanica
Full size table

We performed BLAST analysis to identify M.javanica secreted proteins with homologues in C. elegans. M. javanica proteins homologous to C.elegans RNAi lethal phenotype genes can be potential targets for functional analysis using RNA interference (RNAi) tool and inform designing of control strategies. Meloidogyne javanica protein sequences with homologues in host proteomes were excluded from this analysis considering that they are less likely to be exploited as nematode control targets due to off-targets effects. In total, 5,289 secretory proteins were detected as RKN-specific shown in Additional file 1: Figures S6 were blasted against C. elegans’ lethal phenotype genes. On the premise that homologous genes retain similar functions in other organisms and share other crucial features [36]. In total 227 secretory proteins were similar to C. elegans RNAi lethal phenotype genes including transthyretin-like protein family, peptidases, and peptidase inhibitors (Additional file 2: Table S13).

Discussion

Meloidogyne javanica and M. incognita are both tropical Meloidogyne species with comparable secretomes. Despite being closely related, the two RKN species exhibit differences in their host ranges [37]. It is now understood that phytonematode genes encoding effector proteins are key determinants of host specificity [38, 39]. Consequently, variations in pathogen effector repertoires could potentially influence pathogen host range changes, including host range expansion, including host range expansion, which leads to reproductive isolation and subsequent pathogen speciation [40]. It is therefore important to identify secreted proteins of an individual parasite as these are essential for molecular dialogue between the host and the parasite as well as the parasite’s survival within host tissues [5, 6].

The ever-increasing number of sequenced genomes of PPNs provides an unparalleled opportunity to delve into their secretome, which is critical for their ability to survive within host cells. To this end, we applied various computational approaches to predict and annotate the secretome encoded in the M. javanica genome (Fig. 1). This work represents the first in silico prediction of M. javanica secretome which accounts for 7.7% of the total proteome. The size of M. javanica secretome is comparable to those reported in other PPNs in comparison to their proteome size [16, 27, 41] indicating that ˂10% of the total proteome is secreted.

We examine the composition of RKNs secretomes by comparing the functional annotation of M. javanica secretome, reported in this study with the secretome of M. incognita predicted by [16]. Parallels can be drawn between M. javanica and M. incognita secretome, given that Meloidogyne species have the same lifestyle, as obligate biotrophs [42]. One observation of GO enrichment analysis is that RKN-secreted proteins are mostly present on the cell surface, extracellular region, space, and matrix cell compartments (Fig. 3) that are probably involved in virulence and mediate parasite-host interactions [43, 44].

Molecular entities linked to hydrolase activity, lyase activity, and carbohydrate-binding activity under MF were enriched in the secretome compared to the proteome in the GO enrichment analysis (Fig. 3). These molecular activities are associated with the functions of CWDEs. RKN predicted proteins were mapped into various CWDEs families, namely glycoside hydrolase, glycosyltransferase, polysaccharide lyases, and carbohydrate-binding modules. CWDEs provide an essential catalog of RKN secretomes, which not only facilitate nematode’s penetration and migration through host tissue, but they also help nematodes utilize plant nutrients as a source of food and develop nematode feeding sites [31, 45, 46]. The composition of plant cell walls is important in the interaction between plants and nematodes [47]. As a result, RKNs harbour various families of CWDEs in their secretome. This reflects the strategic ways employed by these parasites to break down and modify the multifaceted plant cell wall structure to invade and exploit host nutrients necessary for a successful infection.

Pectate lyases and glycoside hydrolase 5 represent the most abundant families found in this study and are also among the differentially expressed effector-encoding genes implicated in cell wall modification during M. javanica infection process [48]. The functional characterization of M. graminicola pectate lyase (Mg-PEL1) revealed that it is involved in virulence and facilitates RKN parasitism [49].

Protein domain analysis revealed that M. javanica secretome harbors secretory proteins with distinct protein domains which are largely shared with the closely related M. incognita secretome. The protein kinase domain (IPR000719, also supported by our KEGG analysis) was, for example, one of the most represented domains according to InterProscan analyses (Table 1, Additional file 2: Table S7). Signaling transduction pathways such as protein kinases are critical in facilitating microbes’ transition and adaptation inside a host [50]. Several C-type (IPR001304) proteins were among the top represented domains (Table 1, Additional file 2: Table S7), which have previously been shown to aid nematode parasitism by regulating plant defense responses [51, 52].

Additionally, secreted proteins containing transthyretin-like (IPR001534) and thioredoxin (IPR013766) domains (Table 1, Additional file 2: Table S7) facilitate compatible host-nematode interactions by protecting the nematode against toxic reactive oxygen species molecules generated when the host perceives a parasitic nematode [53, 54]. Altogether these findings eminently show that the various domains represented by RKNs secretome play diverse roles in shaping plant-nematode interactions. Functional annotation analysis indicates that RKN-secreted proteins associated with lysosome, protein processing in endoplasmic reticulum, spliceosome, cAMP signaling pathway, calcium signaling pathway, and pathways in cancer (Table 2, Additional file 2: Table S8) are important for phytonematode survival inside the host [16, 27].

Enzyme regulator function was found enriched in M. javanica secretomes (Fig. 3) which are associated with the regulation of enzymes such as peptidases which are important virulence factors in nematodes [30]. Peptidases and peptidase inhibitors constitute a significant component of PPNs-secreted proteins [27, 55]. In our secretome analysis, 276 proteins were annotated as peptidases mapped into five major catalytic classes distributed across 43 families. Similarly, the various peptidase classes are spread in the secretome of various PPN groups including the genus Meloidogyne [27]. The most abundant peptidase classes in this study were cysteine and serine peptidases (Fig. 4) which were supported by our domain analysis (Table 1). Both catalytic classes play important roles in PPN infection processes such as host tissue invasion, nematode development, reproduction, and pathogenicity. Furthermore, it has been demonstrated that silencing cysteine and serine peptidases in parasitic nematodes is detrimental to nematode fecundity and parasitism [56,57,58,59]. Peptidase inhibitors, on the other hand, are implicated in defense mechanisms that avoid peptidase-mediated defense, manipulate host defense responses, and facilitate nematode feeding [30, 60]. The present study shows that peptidase inhibitors are important for nematode-host interactions where several peptidase inhibitors were identified and distributed across 18 families of known peptidase inhibitor classes.

Although there are parallels between M. javanica and M. incognita secretomes, it was noteworthy that secondary and pigment metabolic processes, developmental processes, and anatomical developmental structure were enriched in M. javanica secretome (Fig. 3) but not in M. incognita secretome [16]. Secondary metabolites and pigment compounds are crucial for parasite survival and pathogenicity [61]. Though precise functions in RKN infection are unknown. However, comparable roles may be hypothesized for RKN parasitism. Enrichment of developmental process and anatomical developmental structure implies that RKN-secreted proteins function in the development and maintenance of giant cells, which provide nutrients to the developing nematode after host colonization [62].

Further comparative analysis with different Meloidogyne species secretomes was carried out to uncover species-specific M. javanica secretome. The species-specific proteins may reveal the parasite-specific molecular mechanism of host adaptation and specialization [40]. Here, 382 M. javanica-specific secreted proteins were discovered that may help to explain nematode lifestyle, virulence, and host specificity and could eventually be applied to design control strategies for M. javanica infestations.

These secreted proteins might be critical in mediating M. javanica parasitic survival, as shown by functional groups identified such as defense and cell death. Remarkably, positive regulation of siRNA production and regulation of miRNA-mediated gene silencing were among the represent GO terms. Current research indicates that pathogen small RNAs play an active role in virulence [63]. However, their regulatory role in PPNs is unknown. Molecular activities associated with these secreted protein sets include hydrolase and aspartic-type endopeptidase activity which are related to typical function of secreted proteins. Many of these predicted sets of secreted proteins (Additional file 2: Table S10 and Table S13) can be exploited for functional characterization and can also be utilized as a panel of effectors in root-knot nematode ‘effectorome’ investigations.

Conclusions

In conclusion, this study sheds light on what constitutes the secretome of M. javanica. To the best of our knowledge, this is the first account of M. javanica secretome being identified and functionally annotated. In silico analysis of secretomes encoded by parasites’ genomes can precede and complement the laborious and complex experimental methodologies applied in deciphering the function of putative excreted-secreted proteins.

Materials and methods

Prediction of excretory-secretory proteins from the M. javanica genome

The proteome deduced from M. javanica genome under BioProject ID PRJEB8714 [20] was retrieved from Wormbase parasite database version WBPS14. Our bioinformatics workflow integrated various computational tools to identify and annotate ESPs from the proteome is summarized in Fig. 1. Firstly, the SignalP v5.0 [64] was applied to predict signal peptides of classically secreted proteins, with parameters set for eukaryote organisms, cleavage site determined by the software. SecretomeP v1.0 with default option for mammalian organisms and cut-off over 0.9 NN- scores was used to predict proteins secreted through unconventional secretory protein pathways accounting for a set of non-classically secreted proteins [65]. Both sets of predicted secretory proteins were scanned with TMHMM (version 2.0) [66] and Phobius [67] for the presence of transmembrane helices using default parameters. All the proteins with a transmembrane region were discarded from this dataset. Mitochondrial targeting proteins predicted by TargetP v1.1 [68] with a probability score over 0.90 were eliminated from the dataset. The resulting list of proteins was subsequently scanned for the presence of endoplasmic reticulum retention signal and glycosylphosphatidylinositol (GPI) anchor signals by PS SCAN [69] (Prosite pattern: PS00014) and PredGPI [70], respectively using default parameters. Finally, all the protein sequences exhibiting ER signals and GP-anchor signals were excluded from the dataset and the remaining proteins were considered secretory proteins for further investigation. To determine the proportion of secretory proteins supported at the transcriptional level, the predicted secretory proteins were matched to M. javanica 7,734 expressed sequence tags (EST) retrieved from the NCBI dbEST database (https://www.ncbi.nlm.nih.gov/genbank/dbest/) using tblastn at a cut-off value of 1e-05 and percentage identity of ≥ 50%.

All the secretory proteins were scanned for the presence of the recently identified core motif Mel-DOG (Meloidogyne DOrsal Gland; TGCMCTT) associated with the expression of dorsal gland effectors [29]. The 2 kb putative promoter regions upstream of the coding region were retrieved from the Wormbase database [71]. FIMO webserver (v5.4.1) [72] was used to search for the core motif occurrence (p-value ˂ 0.0001) in the promoter regions of the predicted secretory proteins.

Functional annotation of secretory proteins

Functional annotation of the secretory proteins comprised assignment of Gene Ontology (GO) terms, GO enrichment analysis, mapping protein domains, and pathway associations using KEGG pathways. The secretory proteins were assigned to Gene Ontology (GO) terms into three functional categories of Biological Process (BP), Molecular Function (MF), and Cellular Component (CC) using the PANNZER2 functional annotation web server [73]. GO visualization and enrichment analysis was performed by WEGO 2.0 web tool https://wego.genomics.cn/view/WEGOID72126237138727 [74]. Significantly enriched GO terms were considered at a P-value ˂ 0.05 using the entire proteome GO terms as the reference group. Additionally, peptidases (proteolytic enzymes) and peptidase inhibitors were identified by scanning against the MEROPS database, v11.0 [75] using BLASTP (e-value ˂1e-5). Secretory proteins coding for carbohydrate-active enzymes (CAZymes) were identified using the dbCAN web server v10.0 HMMER (e-value ˂ 1e-15, coverage ˃ 0.35) [76]. Protein families and domains analyses were mined from the Blast2GO suite (version 6.03) [77].

The KAAS (KEGG [Kyoto Encyclopedia of Gene and Genomes], Automatic Annotation Server, v2.1) [78] was used to map all the secretory proteins to KEGG pathways and KEGG BRITE using BBH (bi-directional best hit) method to assign the orthologs [79]. The eukaryotes and nematode datasets were used as reference datasets in KAAS mapping.

Functional analysis of specific M. javanica secretome

To identify putative orthologs in the secretome of closely related Meloidogyne species including M. incognita, M. enterolobii, and M. floridensis we performed orthologous gene clusters using orthovenn2 with the following parameters; e-value 1e-05 and an inflation value 1.5 [80]. The species-specific predicted secretory proteins were searched for sequence similarities against the NCBI nr database using BLASTP (e-value < 1e-05).

Discovery of M. javanica proteins with potential for functional analysis using RNA interefence (RNAi) tool

The predicted M. javanica secretome was searched for similarity against the host proteomes (Solanum tuberosum, Arabidopsis thaliana, Oryza sativa, and Zea mays) using Orthovenn 2 default parameters; e-value 1e-05 and an inflation value 1.5 [80]. The resulting RKN-specific proteins were screened for the presence of Caenorhabditis elegans lethal RNA interference (RNAi) phenotypes genes retrieved from the WormMine database (WS284) using the tblastn program and filtered using an e-value threshold of 1e-05 and a percentage identity of 50%.