Small peptide–mediated self-recognition prevents cannibalism in predatory nematodes

See allHide authors and affiliations

Science  05 Apr 2019:
Vol. 364, Issue 6435, pp. 86-89
DOI: 10.1126/science.aav9856

Peptides let nematodes know family

To maximize fitness, organisms need to be able to recognize their own species, especially in the proximity of closely related individuals. Lightfoot et al. identified a hypervariable small peptide in the predatory nematode Pristionchus pacificus that is involved in species recognition to prevent predation of kin. They induced modifications in the carboxyl terminus of the peptide with a CRISPR-Cas9 system, which showed that this region is necessary for self-recognition. This molecular recognition system appears to prevent cannibalism and thus enables the worm to focus on appropriate prey species.

Science, this issue p. 86


Self-recognition is observed abundantly throughout the natural world, regulating diverse biological processes. Although ubiquitous, often little is known of the associated molecular machinery, and so far, organismal self-recognition has never been described in nematodes. We investigated the predatory nematode Pristionchus pacificus and, through interactions with its prey, revealed a self-recognition mechanism acting on the nematode surface, capable of distinguishing self-progeny from closely related strains. We identified the small peptide SELF-1, which is composed of an invariant domain and a hypervariable C terminus, as a key component of self-recognition. Modifications to the hypervariable region, including single–amino acid substitutions, are sufficient to eliminate self-recognition. Thus, the P. pacificus self-recognition system enables this nematode to avoid cannibalism while promoting the killing of competing nematodes.

Self-recognition, the capacity to discriminate between self and foreign tissue, or kin, is observed across the tree of life. It regulates diverse biological processes—including communication between viruses (1), aggregation of simple organisms (2, 3), accurate neuronal organization (4), and the vertebrate immune response (5, 6)—and is thought to have played a role in the evolution of multicellularity (7). However, little is known of the molecular mechanisms that regulate self-recognition, and strikingly, descriptions of self-recognition in nematodes are restricted to mating incompatibility (8) and detecting foreign nucleic acids (9), with no organismal self-recognition mechanism thus far described.

The predatory nematode Pristionchus pacificus feeds on the larvae of other nematodes, enabled by the gain of teethlike denticles (1012). These denticles show phenotypic plasticity with two discrete variants: the stenostomatous (St) morph, which has a single small blunt tooth, and the eurystomatous (Eu) morph, which has two large teeth (13, 14). The two morphologies are linked to distinct feeding behaviors because only Eu animals are active predators, whereas St animals feed on bacteria (11). Predatory interactions in Pristionchus species have previously used Caenorhabditis elegans as prey (11, 12). We investigated their interactions with kin, revealing a self-recognition mechanism that prevents the cannibalism of their own offspring while promoting the killing of competing nematodes.

P. pacificus and many of its relatives are predators, feeding on the larvae of other nematodes, such as C. elegans (fig. S1, A and B). We used corpse assays (fig. S2) (11) to investigate the interactions between four different predatory species and their self-progeny (fig. S1C and table S1). All four predatory nematode species killed C. elegans larvae abundantly, with no killing of self-progeny (Fig. 1A). Although killing efficiency varied between species, predators exhibited no bite location preference along the length of the C. elegans body (fig. S3).

Fig. 1 Predatory nematodes avoid self-killing.

(A) Interspecific corpse assays of four predatory Pristionchus species fed own larvae and C. elegans larvae and a diagram of the corpse assay used. Data are shown as median, first and third quartile, and data range. (B) Interspecific corpse assays with four predatory Pristionchus species. In (A) and (B), five predators are fed prey for 2 hours for each assay. n = 3 to 5 replicates for each assay. (C) Biting assay feeding P. pacificus predators mixed populations of fluorescently labeled P. pacificus (egl-20::RFP) and P. exspectatus. Diagram shows schematic of the biting assay by using a mixed culture of prey, including RFP animals. n = 10 animals.

To test whether Pristionchus species would kill species of nematodes that are more closely related than C. elegans, corpse assays were performed between different combinations of these four species (Fig. 1B). The killing of self-progeny was avoided, whereas all predatory species killed and fed on their close relatives. Thus, predatory nematodes exhibit species-level self-recognition specificity.

Described self-recognition systems have reported diverse mechanisms of producing self-specificity (17). In predatory nematodes, we theorized two potential methods of generating a self-recognition signal. The self-signal could be induced by a pheromone secreted from the nematode into its environment, or it could be generated by means of a surface-bound signaling structure on the worm cuticle. To test these hypotheses, we disguised any potential secreted self-pheromone signal by mixing the larvae of self (P. pacificus) with a nonself rapidly killed species (P. exspectatus). Because the larvae of these nematodes are visually indistinguishable under assay conditions, we generated a fluorescently tagged strain of P. pacificus to mix with the P. exspectatus larvae to enable their identification (Fig. 1C). Biting assays (fig. S2) (11) by using the P. pacificus predators that were fed the mixed larval population revealed immediate biting upon nose contact with the nontagged larvae (Fig. 1C) but not self-larvae (movie S1). This suggests that self-recognition was maintained despite the blended pheromones and that self-recognition is generated by a signal localized on the nematode surface.

Next, we explored self-recognition within a single predatory species, focusing on P. pacificus, which has a library of more than 1000 strains isolated from around the world, many of which are genetically diverse (fig. S4). Because most P. pacificus strains fall into three major phylogenetic clades (15), we selected a representative predator from each and fed this nine strains, including self (fig. S4 and table S2). In all assays, the killing of self-progeny was avoided, but predators killed larvae of all other strains (Fig. 2). Thus, P. pacificus demonstrates intraspecific selectivity, with closely related strains identified as nonself and predated on and only self-progeny spared. Such frequent within-species predation hints at a rapidly evolving mechanism that generates diversity between closely related animals.

Fig. 2 P. pacificus intraspecific predation and identification of a self-recognition locus.

Intraspecific corpse assays by using three strains as predators representing the three major clades of P. pacificus. The color scheme is based on phylogeny in fig. S3. In each assay, 20 predators were fed larvae of nine other strains of P. pacificus, including self, for 24 hours. n = 3 replicates for each assay.

We then performed genetic mapping between two P. pacificus strains, PS312 from California and RSB001 from La Réunion, which engaged in predatory feeding on one another. Both PS312 and RSB001 have genotyped recombinant inbred lines (RILs) (16). These RILs were used as prey for both parental strains to identify potential ligands. These phenotypes are mutually exclusive; RILs killed by one parental strain were never killed by the other, indicating that a single causative locus may be involved in self-recognition (table S3).

A quantitative trait loci (QTL) analysis resulted in a single peak at Contig6 on the left arm of chromosome II (fig. S5A), where the most significantly associated single-nucleotide polymorphism (Fisher’s exact test, P < 10−6) was associated with killing behavior in 80% of cases (fig. S5B). Exploiting genetic markers between parental strains identified a region of 600 kb containing 83 predicted genes. However, despite screening 4000 gametes, no further informative recombination events were detected. This lack of recombination may be due to the large genetic distance between these two strains. We therefore substituted our mapping strain, RSB001, for the more closely related PS1843 strain from the state of Washington, which has been previously used to map genes (17) and between which we observed killing. However, this also did not result in any detectable recombination events.

To bypass the lack of recombination, we adapted a CRIPSR-Cas9 method to generate recombination events (18). This system allowed us to induce specific recombination events at target locations during meiosis in P. pacificus (fig. S6, A and B, and table S4), which could be used to map the self-recognition locus. Subsequently, we identified a small locus that contains 10 predicted genes corresponding to the ligand self-recognition phenotype (fig. S6C). Mutations were created in all candidate genes in the PS312 background by means of CRISPR-Cas9 (fig. S7 and table S5) (19). Only Contig6-snap.268 (PDM60460.2), when mutated, resulted in a self-recognition–defective phenotype. Specifically, Contig6-snap.268 mutant animals were killed by PS312 predators, whereas no mutant killed PS312 prey (Fig. 3A and fig. S8). Thus, Contig6-snap.268 appears to be involved in nematode self-recognition. We therefore designated this gene as self-1 for self-recognition-defective.

Fig. 3 Identification of self-1.

(A) Corpse assays over 24 hours by using 20 wild-type predators (PS312 and PS1843) fed animals carrying mutations in genes within the mapping interval. Mutations in Contig6-snap.272 were sterile and unusable for assays. Mutations in Contig6-snap.268 are self-recognition defective; however, killing frequency is lower than observed between strains, as seen in controls by using PS1843 as predators. (B) Predicted gene structure of self-1 and sgRNA target sites used to induce mutations. (C) Corpse assays over 24 hours by using 20 PS312 predators fed various self-1 alleles carrying mutations in different predicted sORFs. For all corpse assays, three to five replicates were performed.

Evaluation of the cDNA sequence of self-1 (Fig. 3B) revealed a single isoform containing only several short open reading frames (sORFS). A single sORF (self-1a), encoding a peptide of 63 amino acids, coincided with the initial self-1 allele, with additional alleles in this sORF also phenocopying the self-recognition defect (Fig. 3C). Two overlapping potential sORFs were also identified located in the self-1a 3′ untranslated region (3′UTR) designated self-1b and self-1c (Fig. 3B). To investigate their potential functionality, we used CRISPR-Cas9 to target a region common to both self-1b and self-1c and analyzed the self-recognition phenotype in these mutants (Fig. 3C and table S6). No loss of self-recognition was detectable, indicating that only self-1a contributes to self-recognition. We then attempted to identify homologous genes in other nematodes, including C. elegans. A self-1 homolog could only be identified in transcriptome data of other Pristionchus species and Micoletkya japonica (20), indicating that self-1 is a taxon-restricted gene (fig. S9). We hypothesize that self-1 may have evolved de novo as a mechanism of generating self-specificity or alternatively may represent a rapidly evolving gene.

A surprising aspect of the P. pacificus self-recognition system is the intraspecific killing observed between even closely related strains (Fig. 2 and fig. S4). We thus generated RNA sequencing (RNA-seq) data from natural isolates and screened for the presence of self-1. A comparative analysis of SELF-1 from 38 strains (fig. S10 and table S2) exhibited conservation across the majority of the peptide, with diversity concentrated at the C terminus and with variation detected in amino acid sequence, length, and self-1 copy number (Fig. 4, A to C, and fig. S10). Within the conserved region, an N-terminal signal domain was identified along with a pair of Lys-Arg residues, which is indicative of potential cleavage sites such as those observed in other peptides (21). We confirmed the importance of the signal peptide by producing a targeted mutation removing this domain, SELF-1∆signal (Fig. 4, D and E), which also showed self-recognition defects.

Fig. 4 Diversity in the SELF-1 peptide is essential for self-recognition.

(A) Graphical representation of the SELF-1 peptide generated from the alignment of 38 P. pacificus strains. A signal peptide and additional potential cleavage sites are indicated. (B and C) Comparative analysis of variation in (B) P. pacificus SELF-1 hypervariable region length and (C) copy number. (D) Specific mutations induced in self-1. (E) Corpse assays by using 20 predators of PS312, fed self-1 mutants carrying various specific mutations. (F) Corpse assays by using 20 predators of the designated strain (PS312 or PS1843) fed self-1 mutants carrying switched hypervariable domains. (G) self-1::RFP transcriptional reporter. Expression is in all epidermal cells.

Because regions of extreme variability often generate specificity in self-recognition systems (16), we next explored whether the hypervariable C-terminal region is sufficient to generate the self-specificity in P. pacificus. Using targeted mutations, we generated several precise alterations in the hypervariable region (tables S7 and S8). First, we produced mutants in which the whole of the hypervariable region was removed: SELF-1∆hypervariable. Second, a version in which the PS312 hypervariable region was replaced with the orthologous hypervariable sequence found in PS1843: SELF-1PS312>PS1843. Last, we generated two single–amino acid changes in the hypervariable domain, one substituting a serine for isoleucine, SELF-1S59I, and another substitution of isoleucine to leucine, SELF-1I58L (Fig. 4D). The self-recognition defect was recapitulated in all mutant strains generated (Fig. 4E). These findings suggest that the hypervariable region is indispensable for the peptide’s self-signaling function, with a single amino acid change in this domain sufficient to disrupt recognition. Mutants that lack the whole hypervariable region display a greater self-recognition defect than that observed in the self-1 null alleles, which may indicate multiple functional domains within SELF-1.

The SELF-1PS312>PS1843 mutant recapitulated the self-recognition defect observed in the self-1 null mutant when fed to PS312 predators (Fig. 4E). This PS1843 motif at the SELF-1 C terminus was also insufficient to prevent killing by PS1843 predators. The reciprocal SELF-1PS1843>PS312 variant was similarly unable to prevent killing by PS312 predators (Fig. 4F), which suggests that additional components are required for self-recognition. Consistent with these observations, attempts to disrupt predation through the soaking of prey or predators in synthetic hypervariable peptides corresponding to the rival strain or the soaking of prey in SELF-1 antibodies did not disrupt self-recognition (fig. S11, A and B, and table S9). Last, the expression of SELF-1 in C. elegans also failed to prevent their killing (fig. S11, C and D).

Different strains may contain multiple copies of self-1 with diverse C-terminal sequences (Fig. 4C and fig. S10). Whereas only a single copy of self-1 is found in PS312, PS1843 contains two copies within 1 Mb, designated self-1.1 PS1843 and self-1.2 PS1843, respectively. Mutations were generated in both genes, but we only observed a self-recognition defect in self-1.1 PS1843 mutants (fig. S11E). Thus, in PS1843, self-1.1 PS1843 is the functional self-1 copy.

Last, we generated a red fluorescent protein (RFP) transcriptional reporter, which revealed that self-1 is expressed in the epidermal layer at all life stages (Fig. 4G and fig. S12). Attempts to identify the localization of the SELF-1 peptide by using CRISPR-Cas9 to generate three epitope-tagged versions of the SELF-1 protein all recapitulated the self-recognition defects observed in self-1 mutants (fig. S13A). Additionally, using SELF-1–specific antibodies was unsuccessful because the P. pacificus cuticle proved immunoreactive, even with control antibodies (fig. S13B). Translation of SELF-1 was still detectable by means of Western blot in whole-worm extracts of SELF-1::2xFLAG animals (fig. S13C). Thus, the SELF-1 peptide is generated in the epidermal layer, but its final localization is currently unresolved.

Together, our findings suggest that self-1 is a component of a complex mechanism that generates self-recognition. In SELF-1, the polymorphic C terminus appears to be essential for specificity and may provide a recognition code for distinguishing kin from non-kin. Although high degrees of sequence polymorphisms are a common feature of self-recognition systems (16), these polymorphisms often reside within membrane proteins that contain extracellular immunoglobulin-like folds (35, 2224), a feature that is thus far absent in peptide-mediated P. pacificus self-recognition. Thus, nematode predation provides a powerful system for understanding self-recognition and the evolution of mechanisms that might facilitate competition between conspecifics.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S13

Tables S1 to S9

References (2536)

Movie S1

References and Notes

Acknowledgments: We thank M. Okumura, A. Lupas, M. Werner, M. O’Donnell, and A. Streit for discussions and critical reading of the manuscript. We thank W. Roeseler and G. Eberhardt for next-generation sequencing expertise. Funding: This work was funded by the Max Planck Society. Author contributions: J.W.L. and M.W. performed all experiments, with assistance from E.M., V.S., and H.W. Bioinformatic data analysis was performed by C.R. All experiments were designed by J.W.L., M.W., and R.J.S. Competing interests: The authors declare no competing interests. Data and materials availability: The self-1 sequence has been submitted to NCBI GenBank under the accession no. PDM60460.2. RNA-seq data has been deposited at the European Nucleotide Archive under the study accession no. PRJEB27362. All other data are available in the main text or the supplementary materials.

Stay Connected to Science

Navigate This Article