MicroRNAs in plant-parasitic nematodes: what are they and why should we care?

Since first being discovered in C. elegans in 1993 (Lee et al., 1993), miRNAs have now been described in many nematode species, including free-living, for example, Pristionchus pacificus, C. briggsae, C. remanei (de Wit et al., 2009), and parasitic species, for example. Haemonchus contortus (Marks et al., 2019), among many others. However, the characterization of miRNAs in PPN species is still in its early stages and is limited to a few species.

The first characterization of miRNA in PPN was in the pinewood nematode, Bursaphelenchus xylophilus (Huang et al., 2010). Although constrained by the methodologies available at the time, this study successfully identified and experimentally validated 57 miRNAs, shedding light on their evolutionary conservation across diverse eukaryotic lineages. Eighteen of these miRNAs were conserved across several animal groups, including C. elegans, Drosophila melanogaster, and humans, and a substantial proportion (38 out of 57) were shared between B. xylophilus and C. elegans, highlighting a high degree of conservation within nematodes. Another study on B. xylophilus expanded its known miRNAome (the complete set of miRNAs encoded in its genome) by identifying several novel miRNAs (Ding et al., 2015). The authors also correlated the expression of two miRNAs, predicted to interact with perfect seed region complementarity, with a glycoside hydrolase involved in plant cell wall degradation. They observed a mutually exclusive expression pattern, suggesting that this key infection step may be regulated by these miRNAs.

However, the most studied species of PPN for miRNAs is Meloidogyne incognita, with the first paper in 2015, which combined sequencing of small RNA-enriched libraries with predictive computational approaches (Wang et al., 2015). Through this study, the authors identified 102 mature miRNA sequences encoded within 71 distinct genes, of which 27 were homologous to known sequences in other nematodes. Several of these MIR genes were shown to be grouped into genomic clusters, an organization that has also been observed in several other organisms to date. A second study on M. incognita further expanded these numbers to 289 miRNAs, of which 35 were specific to this species (Zhang et al., 2016b). In both studies, significant differences in the expression profiles of these miRNAs were observed, with only a few being highly expressed. All of these highly expressed miRNAs were conserved across species and are likely essential in regulating nematode development. Additionally, significant differences in MIR gene expression between developmental stages were observed, and prediction of mRNA interactions revealed how these miRNAs could modulate specific metabolic pathways during the development of M. incognita (Subramanian et al., 2016; Kaur et al., 2017; Liu et al., 2019). Further emphasizing this importance, several miRNAs were found to be stage-specific, suggesting critical roles in developmental or infection processes. The comparison of miRNA expression with that of mRNAs predicted to be their targets in different stages has revealed probable regulatory networks for important functions (Mani et al., 2021).

Most recently, the soybean cyst nematode H. glycines became the latest PPN species to have its miRNAome characterized. In addition to identifying miRNA genes and quantifying their expression across all developmental stages, Ste-Croix et al. (2023) also compared the miRNAome of H. glycines to those of other PPN species. Among the newly identified miRNAs, they found that 75% were shared with H. schachtii (same genus), fewer than 10% with Globodera rostochiensis and G. pallida (same subfamily), and none with Meloidogyne species (same family). Based on estimated divergence times among these groups, this corresponds to a miRNA birth rate comparable to that of other animals, but substantially higher than in C. elegans. Predicting the mRNA targets within the nematode revealed that, on average, each nematode gene was predicted to interact with three different miRNAs, while each miRNA was predicted to target hundreds of genes. Although these numbers varied greatly between individual miRNAs and are likely overestimated due to the nature of computational predictions, they illustrate the extensive regulatory potential of miRNA-mediated gene expression. Furthermore, analysis of miRNA-effector interactions revealed that effectors were primarily predicted to interact with species-specific miRNAs, suggesting recent evolutionary origins and implication in species-specific traits (e.g., host range). Interestingly, a few unique miRNAs were predicted to interact with several effectors and may serve as general modulators of their expression, similar to what was recently observed with transcription factors (Pellegrin et al., 2025).

The involvement of miRNAs in regulating specific effectors has been directly demonstrated in M. graminicola. Using a miRNA target prediction algorithm (see Section “miRNA Discovery and mRNA Target Prediction in PPN”), followed by validation through an in vitro dual-luciferase reporter assay, it was shown that mgr-miR-9 modulates the expression of the protein disulfide isomerase MgPDI (Tian et al., 2023), while mgr-miR-228 regulates the expression of the transthyretin-like protein MgTTL1 (Tian et al., 2025). In both cases, repression of these miRNAs during infection resulted in upregulation of their respective effector genes. Furthermore, the application of synthetic miRNAs led to downregulation of the target effectors, and subsequent bioassays revealed a significant reduction in the reproductive capacity of M. graminicola. These findings support the idea that miRNAs may play a critical role in the fine-tuned and dynamic regulation of gene expression throughout the different stages of infection.

Moreover, miRNA activity may potentially extend to the regulation of host plant genes as well. Exosomes (microvesicles involved in the transport of small molecules) were also isolated from H. glycines and were shown to contain miRNAs, some of which show strong homology to host genes (Ste-Croix et al., 2023). While these predictions remain exploratory due to the fundamental differences in gene regulation between plants and nematodes, they raise a compelling hypothesis about cross-kingdom regulatory interactions in PPN.

RNA, in all its forms, is commonly found in the environment but is typically degraded rapidly. The discovery of stable plant-derived miRNAs circulating in the serum of humans and animals was unexpected (Zhang et al., 2012). These miRNAs, originating from food intake, were not only capable of being absorbed but also exhibited resistance to degradation by RNases and digestive enzymes. Even more surprisingly, plant miRNAs were shown to suppress gene expression in the animal host, mimicking the regulatory function of endogenous eukaryotic miRNAs. Subsequently, it was demonstrated that parasites, such as the gastrointestinal nematode Heligmosomoides polygyrus, can also secrete miRNAs that are detectable in host tissues and modulate the immune response (Buck et al., 2014). Purified exosomes containing nematode-derived miRNAs, and in some cases even the nematode AGO protein, were sufficient to suppress host immunity upon injection, in the absence of the nematode itself. In contrast, the mechanisms of miRNA action in plants differ substantially from those in animals, raising the question of whether eukaryotic parasites that infect plants might similarly manipulate host gene expression through the release of miRNAs.

One of the first examples of cross-kingdom short RNA (sRNA) interactions in a plant pathosystem was found in the causative agent of the gray mold disease, Botrytis cinerea. The fungus has been shown to suppress the immune response in tomato by binding to the plant’s AGO proteins and specifically inhibiting the expression of MAP kinases involved in this defense pathway (Weiberg et al., 2013). Prior to this report, most of the known virulence factors or effectors were of protein origin. However, this study confirmed that sRNAs could also function as effector molecules and may represent key elements of fungal pathogenicity. To reach its targets within the plant cells, the fungus was shown to employ EVs containing sRNAs, which were secreted and subsequently internalized by plant cells via clathrin-mediated endocytosis (He et al., 2023). Interestingly, the reverse was also shown to be true with a tomato-derived miRNA inhibiting the virulence of B. cinerea (Meng et al., 2020). Indeed, bioassays demonstrated that, following infection, the expression of the tomato miR396a-5p was upregulated, while the expression of confirmed fungal virulence genes was subsequently repressed (Wu et al., 2023). Moreover, exogenous application of this miRNA was sufficient to suppress fungal virulence. A similar mechanism has also been observed in the interaction between Verticillium dahliae with cotton, as well as within other pathosystems (Zhang et al., 2016a; Wang et al., 2017). While evidence and review papers on cross-kingdom interactions are accumulating (Liang et al., 2013; Wang et al., 2017, 2018; Zhou et al., 2017; Gualtieri et al., 2020; Chowdhury et al., 2024), there are still few examples in PPNs.

Comparative analysis of H. glycines miRNA sequences with those of its host plant predicted 1,542 potential interactions involving 1,281 unique plant genes (Ste-Croix et al., 2023). When focusing only on miRNAs also detected in nematode-derived exosomes, thus more likely to reach the cytoplasm of plant cells, 540 interactions involving 482 genes remained. Several of these target genes were directly involved in the plant’s defense response against the nematode, supporting the hypothesis that H. glycines may secrete miRNAs as effector molecules to manipulate host gene expression to its advantage. This remains a speculative hypothesis, as multiple barriers must be overcome for these miRNAs to be functional in the host. Specifically, the miRNAs must be expressed at the appropriate time, packaged into exosomes (or secreted via other mechanisms), delivered into plant cells, transported to the cytoplasm, and be compatible with the plant’s AGO proteins and RISC complex. Nonetheless, the possibility of cross-kingdom gene regulation via nematode-secreted miRNAs represents a compelling area for further investigation. Interestingly, in the same study, nine miRNAs were predicted to target both nematode effector genes and host genes located near known soybean resistance loci. These miRNAs were specific either to H. glycines (n = 4) or to the Heterodera genus (n = 5), suggesting they may have evolved relatively recently in association with the obligate parasitism and host specialization of these nematodes.

A similar study was conducted in the tomato-root-knot nematode pathosystem. Based on sequence complementarity, bioinformatic predictions identified 523 putative cross-kingdom interactions, involving 105 M. incognita miRNAs predicted to target 469 tomato genes (Leonetti et al., 2024). In contrast, thousands of interactions were predicted in the opposite direction, with tomato-derived miRNAs potentially targeting M. incognita genes. Functional enrichment analysis of the highest-confidence interactions revealed that the tomato genes targeted by nematode miRNAs are primarily involved in plant development and stress responses. Conversely, plant miRNAs were predicted to influence a wide range of nematode processes, including mobility, host recognition, feeding behavior, and overall development. As noted earlier, these predictions remain speculative. However, in this case, the authors went a step further by experimentally validating a subset of the predicted interactions. M. incognita second-stage juveniles (J2s) were soaked in synthetic tomato miRNAs predicted to have cross-kingdom targets. This treatment led to a significant reduction in the expression of two target genes (Minc11367 and Minc00111). In planta bioassays further demonstrated a marked reduction in root swelling and gall formation in nematodes pre-treated with two specific tomato miRNAs (sly-miR156a and sly-miR169f), compared to untreated nematodes (Leonetti et al., 2024).

These studies indicate that it is possible for plants and their nematode pathogens to exchange miRNAs, often transported via EVs, and that these sRNA effectors play roles in both defense and virulence (Cai et al., 2021). This opens the door to novel control strategies that could exploit this molecular cross-talk for targeted pest and pathogen management.

Identifying miRNAs in non-model organisms requires integrated approaches that combine computational prediction with experimental validation. To support this, a range of tools and databases have been developed. In this section, we review the key resources currently available to researchers investigating these questions in PPNs, as summarized in Table 2.

The conserved nature of miRNA families across diverse organisms makes miRNA sequence databases powerful resources for identifying putative miRNAs in PPNs. By exploiting sequence homology with known miRNAs, such as those cataloged in miRBase, which includes over 38,000 unique miRNAs from various species, researchers have successfully identified candidate miRNAs in species like H. glycines and M. incognita (Subramanian et al., 2016; Lian et al., 2019). Importantly, the conserved nature of miRNA extends beyond sequence similarity; the regulatory functions of many miRNA families are also preserved across taxa, strengthening the rationale for cross-species prediction. Complementing this approach, databases such as TargetWormScan and miRTarBase provide extensive collections of computationally predicted and experimentally validated miRNA targets, offering valuable insights into the potential roles and functional relevance of these candidate miRNAs (Table 2).

Although database-driven discovery of miRNAs in PPNs provides a strong starting point, it is likely to miss species-specific miRNA families, an issue highlighted in H. glycines, with nearly half of its repertoire predicted as genus-specific (Ste-Croix et al., 2023). As a result, current best practices for miRNA studies recommend a combined approach: using database-guided predictions alongside de novo discovery methods based on sRNA deep-sequencing data (Yang and Li, 2011; Yang et al., 2021). Tools such as miRDeep2, miRAnalyser, and sRNABench (Table 2) can predict and quantify miRNAs using only sRNA sequencing data and homology-based training. However, their full potential is realized when genomic information, such as complete genomes or contig assemblies, is available alongside the sequencing data. The inclusion of genomic context allows for more accurate identification of precursor structures, improves the reliability of miRNA annotation, and enhances the prediction of novel miRNAs.

Nevertheless, while predicting miRNAs is an essential step, assigning them biological context and function is typically of greater interest. To this end, numerous tools have been developed for both model and non-model organisms to predict the targets of individual miRNA candidates. However, due to mechanistic differences in the machinery between plants and animals, selecting the appropriate prediction tool is crucial. For instance, tools like miRanda, an algorithm based on mammalian target recognition rules, can predict miRNA targets in nematodes but perform poorly in plant systems. It is also important to note that not all animal-trained algorithms perform equally well. For example, miTAR, which is trained on human data, performs worse in nematodes than miRanda, an algorithm partially trained on interactions identified in C. elegans. Similar considerations also apply to plant-specific predictors, which are typically optimized for plant-specific miRNA-mRNA interaction rules. Although Table 2 provides a useful starting point, more focused reviews on miRNA target prediction tools should be consulted when selecting an appropriate predictor (Carroll et al., 2014).

It is also worth noting the recent shift in PPN miRNA research toward machine learning-based and sequencing-free approaches. For instance, Ajila et al. (2023a) applied a retrained model based on the SMIRP framework, leveraging qualitative structural features, to identify species-specific miRNAs in H. glycines, predicting 3,342 pre-miRNA candidates. In the same study, machine learning was further employed to generate a species-specific model capable of predicting the in vivo targets of these miRNAs. Similar machine-learning strategies have also been implemented in plants and were recently reviewed by Jayasundara et al. (2021). While still in their early stages and prone to overestimating miRNA diversity, these approaches show great promise for expanding miRNA research across a broad range of species, requiring little more than a reference genome.