Steroid hormone pathways coordinate developmental diapause and olfactory remodeling in Pristionchus pacificus

Abstract

Developmental and behavioral plasticity allow animals to prioritize alternative genetic programs during fluctuating environments. Behavioral remodeling may be acute in animals that interact with host organisms, since reproductive adults and the developmentally arrested larvae often have different ethological needs for chemical stimuli. To understand the genes that coordinate the development and host-seeking behavior, we used the entomophilic nematode Pristionchus pacificus to characterize dauer-constitutive mutants (Daf-c) that inappropriately enter developmental diapause to become dauer larvae. We found two Daf-c loci with dauer-constitutive and cuticle exsheathment phenotypes that can be rescued by the feeding of Δ7-dafachronic acid, and that are dependent on the conserved canonical steroid hormone receptor Ppa-DAF-12. Specifically at one locus, deletions in the sole hydroxysteroid dehydrogenase (HSD) in P. pacificus resulted in Daf-c phenotypes. Ppa-hsd-2 is expressed in the canal-associated neurons (CANs) and excretory cells whose homologous cells in Caenorhabditis elegans are not known to be involved in the dauer decision. While in wildtype only dauer larvae are attracted to host odors, hsd-2 mutant adults show enhanced attraction to the host beetle pheromone, along with ectopic activation of a marker for putative olfactory neurons, Ppa-odr-3. Surprisingly, this enhanced odor attraction acts independently of the Δ7-DA/DAF-12 module, suggesting that Ppa-HSD-2 may be responsible for several steroid hormone products involved in coordinating the dauer decision and host-seeking behavior in P. pacificus.

Introduction

Organisms integrate multiple sensory stimuli from their environment to make key developmental decisions and engage in context-appropriate behaviors. In nematodes, when environmental conditions are unfavorable, a portion of the population can stop feeding and suspend development. Of the several larval stages that can undergo developmental arrest (Zaslaver et al. 2011), the dauer larva (DL) in Caenorhabditis elegans, as well as the equivalent infective juvenile (IJ) or infective third-larval (iL3) stages in parasitic nematodes, undergo spectacular physiological, morphological, and neuroanatomical changes (Albert and Riddle 1983). At the gene expression level, a number of the dauer entry changes are reversible, while others become permanent life-history traits that mark dauer passage (Lee et al. 2017; Vidal et al. 2018). Entry into the dauer stage is primarily determined by environmental factors, particularly the depletion of food, elevated temperature, or high levels of ascaroside pheromones in response to overcrowding (Cassada and Russell 1975; Golden and Riddle 1982, 1984a; Jeong et al. 2005; Butcher et al. 2008).

In C. elegans, multiple signaling pathways act in parallel to mediate dauer entry, such as the recognition of pheromones by G-protein-coupled receptors in the chemosensory amphid neurons and signal transduction to the DAF-11/guanylyl cyclase pathway (Malone and Thomas 1994). In addition, favorable conditions lead to signaling in the DAF-7/TGF-ß and DAF-2/insulin-like pathways (Ren et al. 1996; Ogg et al. 1997; Gems et al. 1998). This signaling facilitates the production of Δ4- and Δ7-dafachronic acid from the cholesterol precursor, which binds the highly conserved nuclear hormone receptor DAF-12 and prevent DAF-12 from repressing genes necessary for development to reproductive adulthood (Dumas et al. 2010; Werner et al. 2017). Less favorable conditions result in the production of pheromones that reduce signaling in these pathways, subsequently shifting DAF-12 from its inactive ligand-bound form to its active, ligand-free form that specifies dauer development. Despite being a nonfeeding and stress-resistant stage, the DL are actively motile and receptive to external chemosensory and mechanical cues. Like free-living DL, IJs, and iL3s in parasitic nematode species share the ability to survive harsh conditions but remain receptive to semiochemicals from their environment.

In addition to changes in morphology and physiology, dauer development in C. elegans also involves rewiring of the sensory nervous system, as well as changes in neuropeptide and chemoreceptor expression (Albert and Riddle 1983; Lee et al. 2017; Bhattacharya et al. 2019). However, the mechanisms that regulate this remodeling process in other nematodes are largely uncharacterized, thus limiting the scope of generalizable nematode-wide programs. Comparisons of chemosensory behaviors in reproductive free-living adults vs developmentally arrested DL or iL3s have revealed that responses to some chemosensory cues change dramatically across life stages. For example in C. elegans, adults are repelled by carbon dioxide (CO₂) while DL are attracted to CO₂ (Hallem and Sternberg 2008; Hallem et al. 2011). However, most studies of chemosensory behavior in free-living nematodes have been conducted with young adults, while most studies of chemosensory behaviors in parasitic nematodes have been conducted with IJs or iL3s, with very few direct comparisons of chemosensory behaviors between the reproductive and dauer-equivalent stages within the same species (O’Halloran and Burnell 2003; Rasmann et al. 2012; Turlings et al. 2012; Gang and Hallem 2016; Neto et al. 2016; Bryant and Hallem 2018).

The entomophilic nematode Pristionchus pacificus is a powerful model system for investigating the genetic network that coordinates dauer entry with development and behavior in host-associated species. In the wild, P. pacificus has been found as DL on various species of beetles; upon death of the beetle host, the nematodes exit the dauer state and resume reproductive development to feed on host-associated bacteria (Weller et al. 2010; Mayer et al. 2015, 2017). Thus, the DL of P. pacificus is considered to be the host-seeking stage. P. pacificus DL have a number of unique adaptations not found in C. elegans DL, including an oily coat that promotes aggregation and enhanced survival up to a year (Mayer and Sommer 2011; Penkov et al. 2014). Pristionchus adults have distinctive chemosensory preferences for insect pheromones and plant volatiles that differ significantly from those of Caenorhabditis species, and also differ across Pristionchus species (Hong and Sommer 2006). However, due to the difficulty of obtaining large numbers of DL using the P. pacificus reference strain, the chemosensory profiles of P. pacificus DL have not yet been examined.

In P. pacificus, several dauer formation defective (Daf-d) mutants that fail to form DL under inducing conditions have been isolated by unbiased forward genetics as well as reverse genetics. These studies have identified the genes encoding the nuclear hormone receptor DAF-12 and the forkhead transcription factor DAF-16 as important regulators of dauer formation (Ogawa et al. 2009, 2011). While DAF-12 and DAF-16 functions are conserved between C. elegans and P. pacificus, mutations in other Daf-d orthologs suggest divergence in the factors that act upstream of these proteins. For example, in C. elegans, high-population density induces the secretion of dauer pheromones (Golden and Riddle 1984b; Jeong et al. 2005; Butcher et al. 2008; Schroeder 2015), and mutations in the ascaroside biosynthesis enzyme DAF-22 result in a Daf-d phenotype (Golden and Riddle 1984c). Interestingly, in P. pacificus, a recent daf-22 duplication resulted in subfunctionalization of ascaroside production, but neither mutations in the Ppa-daf-22.1/daf-22.2 paralogs alone or in combination resulted in defects in dauer formation (Markov et al. 2016). These studies suggest a significant divergence in P. pacificus and C. elegans dauer regulation upstream of the conserved DAF-12 and DAF-16 modules.

In contrast to the Daf-d mutants, no Dauer formation constitutive (Daf-c) mutants have been genetically characterized in P. pacificus, despite forwarding genetic screens and the existence of all of the homologs of the genes responsible for the Daf-c phenotype in C. elegans. Putative homologs of the C. elegans DAF-11/guanylyl cyclase, DAF-7/TGF-ß ligand (7 paralogs), and DAF-2/insulin receptor have been identified in the P. pacificus genome (www.wormbase.org), but it remains unknown if any of these homologs are involved in dauer regulation. At the same time, targeted mutations in the P. pacificus heat shock protein Ppa-DAF-21/HSP90 and Ppa-DAF-19/RFX transcription factor affected ciliogenesis but did not result in a Daf phenotype (Sieriebriennikov et al. 2017; Moreno et al. 2018). Because cholesterol is a precursor for the synthesis of steroid hormones needed for both proper ecdysis and dauer formation (Matyash et al. 2004), it remains to be seen if mutations in cholesterol metabolism and steroid signaling are among the yet-undescribed Daf-c mutants in P. pacificus.

To identify the unique features and key factors controlling dauer-specific phenotypes in P. pacificus, we examined the development and behavior of Daf-c mutants. Previously, a genetic screen in P. pacificus yielded four dauer constitutive Daf-c alleles, dfc (dauer-formation constitutive), all of which can be rescued by exogenous Δ7-dafachronic acid but none has been molecularly cloned (Motola et al. 2006; Ogawa et al. 2009). We, therefore, isolated a new Daf-c allele, csu60, and following positional mapping and whole-genome sequencing, found that csu60 is a null allele of the sole hydroxysteroid dehydrogenase (HSD) in P. pacificus, an enzyme responsible for the production of dafachronic acids. For both Daf-c alleles, the DL and adults have olfactory behavior resembling wild-type starvation-induced DL, suggesting that the genes that control dauer entry coordinate host-seeking olfactory behavior.

Materials and methods

Nematode strains

P. pacificus and C. elegans strains were maintained at 20°C on NGM plates seeded with OP50 as described previously (Cinkornpumin et al. 2014). The following P. pacificus strains were used: PS312 (California wild-type, synonymous with RS2333), Ppa-obi-1(tu404) ChrI, Ppa-dfc-1(tu391) ChrI, and csu60 ChrII, Ppa-daf-12(tu389) ChrX. csu60 was isolated as an off-target mutant from a CRISPR/Cas9-mediated mutagenesis against a TGF-ß ortholog (Contig115-snapTAU.31). Each allele was backcrossed three times to PS312 for the Daf-c phenotype before further characterization. Double mutant with obi-1(tu404) was initially selected for the long body phenotype of obi-1 along with the Daf-c phenotype to create rlh177 (tu391; tu404) and rlh192 (csu60; tu404). Each double mutant with obi-1(tu404) was confirmed by sequencing the single nucleotide deletion in tu404. The cloning and construction of the Ppa-odr-3p::rfp (tuEx265), Ppa-odr-7p::rfp(tuEx297), and Ppa-che-1p::rfp (lucEx367) transgenic lines were described elsewhere (Hong et al. 2019). To examine the expression of these neuronal markers in the csu60 background, we crossed RFP-expressing transgenic males into csu60 J4 hermaphrodites, and then established individual transgenic lines from RFP+ F₂ progeny with the Daf-c phenotype.

For F₁ complementation tests, we crossed dfc-1(tu391); Ppa-daf-6p::rfp or csu88; Ppa-che-1p::rfp males to Ppa-hsd-2(csu60) hermaphrodites and looked for the frequency of DL in the RFP+ progeny. We found no DL in 28 RFP+ F₁ progeny for the tu391 x csu60 cross, and 14 out of 17 DL in 28 RFP+ F₁ progeny for the csu88 x csu60 cross (P-values < 0.00001 and 0.2581 respectively when testing for the difference from csu60 dauer formation, 53 DL/79 total, Fisher’s Exact Test). For the epistasis tests, we first crossed Ppa-daf-12(tu389) hermaphrodites with csu60/+ males, then isolated F₂ progeny homozygous for the csu60 deletion by PCR genotyping and for the single nucleotide substitution of Ppa-daf-12(tu389) by Sanger sequencing (Ogawa et al. 2009). The double mutant (csu60; tu389) did not produce DL on well-fed or on starved plates, indicating the Daf-d phenotype of Ppa-daf-12(tu389) is epistatic to the Daf-c phenotype of csu60.

Molecular cloning

To identify the mutations, we first rough mapped the mutation to a chromosomal region using Simple Sequence Length Polymorphic (SSLP) markers on 45 mapping lines generated with the mapping strain RS5278 (Bolivia). Each mapping line was a descendent of a single F₂ DL on nonstarved F1 plates that exited when we transferred it to a fresh plate of OP50. Nematodes were washed individually from each mapping line in approximate equal pellet sizes prior to pooling and genomic DNA extraction using the MasterPure Kit (Epicentre). We sequenced the whole genome of csu60 alone well as pooled gDNA from mapping lines (Doitsidou et al. 2010) (∼73 million 100 nt reads for each genome and bulk mapping samples; BGISEQ-500 platform with DNA Nanoball technology, BGI America). Rough mapping indicated linkage to Chromosome II, with the additional fine mapping of 47 recombinant lines and the csu60 genome sequencing revealed a large 95 kb deletion on Chromosome II (1,370,000–1,465,000) that encompassed 10 open reading frames. One of the predicted genes in this interval, PPA10139 (UMM-S10-1.4-mRNA-1 from El Paco whole genome sequence), encodes a 376 amino acid homolog of the C. elegans HSD-2.

To make the Ppa-hsd-2 cDNA rescue transgene, 2063 bp of the Ppa-hsd-2 promoter was amplified from PS312 genomic DNA using RHL1012 and RHL1013. The full-length Ppa-hsd-2 cDNA sequence was amplified from PS312 mixed stage cDNA using RHL1014 and RHL1015. The cDNA sequence is identical to UMM-S10-1.4-mRNA-1. The Ppa-hsd-2 promoter and cDNA were combined with Ppa-rpl-23 3’ UTR to make pHC12. The mix injected into PS312 adult hermaphrodites contains the following: 2.5 ng/µl pHC12, 4 ng/µl Ppa-egl-20p::rfp, 80 ng/µl csu60 gDNA (all digested with PstI). The transgenic line was then crossed into the csu60 mutant background using PCR to genotype for the deletion (RHL1006 and RHL1007 will amplify 693 bp if there is a deletion; RHL1043 and RHL1044 will amplify 1008 bp if the same locus is wildtype) and scored as a percentage of dauer or nondauers (J3/J4/adult) containing the rescue transgene. The csuEx54 transgene transmission rate was approximately 27%.

TurboRFP was introduced downstream of the 2063 bp Ppa-hsd-2 promoter by Gibson Assembly to make the Ppa-hsd-2p::rfp reporter transgene. The injection mix contained PstI digested 5 ng/µl Ppa-hsd-2p::rfp, 5 ng/µl Ppa-egl-20p::rfp, 80 ng/µl PS312 gDNA.

CRISPR/Cas9 mutagenesis

CRISPR/Cas9 mutagenesis using the roller co-injection marker was used to generate an additional Ppa-hsd-2 allele csu88 (Nakayama et al. 2020). Target crRNA, tracrRNA, and Cas9 nuclease were purchased from IDT technologies (San Diego, CA, USA) and hydrated to 100 µM with IDT duplex buffer. An equal volume (0.65 µl) of each of target crRNA (RHL1109) or the co-CRISPR roller crRNA (RHL1123) was mixed with the tracrRNA. The tubes were incubated at 95°C for 5 minutes, followed by room temperature for 5 minutes. Cas9 protein (0.5 µl of 10 µg/µl) was added to each duplex mix and incubated at 37°C for 10 minutes. These two sgRNA products were combined and diluted to a total volume of 40 µl. F₁ and F₂ progeny were screened for the dominant roller phenotype, which were then further examined for mutations in the Ppa-hsd-2 locus by PCR and sequencing. The roller mutation was then crossed out twice into the PS312 wildtype background before characterization.

Reverse transcription and quantitative real-time PCR

We used two pairs of amplicons with each primer spanning over exon boundaries to determine the relative level of Ppa-odr-3 transcript expression in csu60. cDNA was synthesized using random hexamers (N6) and polyT primers using ∼500 ng total RNA from mix staged worms (Thermo Fisher Maxima, M1681). We performed two RNA extractions and three cDNA synthesis to run 12 technical replicates for each of the primer sets. We ran the quantitative real-time PCR (qPCR) reactions using 1–2 µl cDNA in SYBR Green Master Mix (Biorad) totaling 20 µl in a Biorad CFX96 machine using 60°C as the annealing temperature. Ppa-beta-tubulin and Ppa-cdc-42 (RHO GTPase) were used as housekeeping genes and the relative fold change in expression between wildtype and csu60 was calculated using the delta-delta C_t method (Schuster and Sommer 2012). RT-PCR for Ppa-hsd-2 and Ppa-odr-3 transcripts were performed with and without reverse transcriptase as described above.

Phylogeny

The amino acid sequences of potential HSD homologs were first identified by BLASTX searches on WormBase. The phylogeny trees were built using the entire amino acid sequences by first making an alignment and removal of positions with gap using T-COFFEE, followed by Maximum Likelihood phylogeny by PhyML, and finally tree rendering by TreeDyn (www.phylogeny.fr; Dereeper et al., 2008). Midpoint rooting was used and branch support ≥50% is shown.

Temperature response

For each experiment, 10 gravid hermaphrodites were picked onto freshly seeded 6 cm NGM plates and incubated at 15°C, 20°C, or 25°C. Plates were scored for third-stage larvae (active dauer, incarcerated dauer, J3 larvae) after 5 days for Daf-c or 3 days for wild-type PS312 (because they starve after 5 days under the same condition). To determine how long the ensheathed DL remain incarcerated, we transferred the immobilized DL at 20°C onto new OP50 NGM plates. After one day at 20°C or 25°C, these cultures were scored for the presence of active DL and incarcerated DL. To determine if J3 or J4 Daf-c mutants become incarcerated, we picked the worms onto new OP50 NGM plates at 20°C.

Exogenous Δ7-dafachronic acid

Ten microliter of 75 µM Δ7-dafachronic acid (Gift from Corey Lab and Cayman Chemical) or 100% ethanol vehicle control was mixed into 90 µl OP50 before seeding onto NGM plates. After 2 days, we picked 15–20 csu60 or csu88 gravid hermaphrodites onto each plate, which were allowed to lay eggs for 5 hours before being removed for egg synchronization, or if the hermaphrodites lay fewer than 50 eggs they were instead removed after 24 hours. Plates were incubated at 20°C or 25°C for 4–5 days and then scored for DLs and nondauer stages.

Behavioral assays

Pharyngeal pumping rates were observed through a Leica DM6000 upright microscope with a 40x objective, without any anesthetic. J3 larvae, DLs, and incarcerated DLs were mounted onto M9 buffer and 2% agar pads on microscope slides. We counted at least four 15-seconds intervals per animal, which were then summed to obtain pumps/minute.

Fluorescent red latex beads (Sigma Aldrich, L3280) were used to stage worms and characterize the DLs (Nika et al. 2016). Overnight OP50 culture was pelleted and resuspended in 1/10 the volume with fresh LB broth. Three microliters of fluorescent beads were added for every 1000 µl of concentrated OP50, and 100 µl of this mixture was seeded onto each NGM plate. Plates were used immediately after drying or stored at 4°C for up to a week. Approximately 20–30 worms were picked onto the fluorescent bead/OP50 lawn and scored 40–90 minutes later for uptake of fluorescent beads in the gut. Similarly, 20–30 worms were picked into 1 ml of 1% SDS in M9 and assayed for percentage of moving worms after 10 or 40 minutes To determine the correlation between bead uptake and SDS resistance, dauer, J2d, and J3 larvae from the same culture were placed separately on fluorescent bead/OP50 lawn for 90 minutes and assayed for percentage of moving worms in 1% SDS.

The population chemotaxis assays for P. pacificus were performed on covered 10 cm Petri dishes for ∼16 hours at 22°C as previously described (Cinkornpumin et al. 2014). However, because P. pacificus DL are coated with a long-chain polyunsaturated wax ester (nematoil) that makes them highly hydrophobic (Penkov et al. 2014), we modified the chemotaxis assay by adopting a chemotaxis medium containing a detergent (10 mM MOPS pH 7; 2.5% Tween 20; 1.5% Bacto-agar)(Nuttley et al. 2002). This MOPS/Tween medium promotes both adult and DL dispersal, which when 5% rather than 10% ZTDO was used as an attractant, likely contributed to the slightly higher CI values for PS312 and obi-1(tu404) adults than expected compared to previous studies (Cinkornpumin et al. 2014). DLs were loaded to the center of the plate by placing and removing immediately a 1 cm² agar chunk from a culturing plate containing either wild-type dauers from starved cultures or Daf-c DL from densely populated cultures. We were unable to perform chemotaxis assays on the DL of obi-1(tu404) mutants due to its high DL mortality rate, especially in the presence of ZTDO (Cinkornpumin et al. 2014). Young adult worms from nearly saturated cultures were washed twice with M9 buffer and collected by centrifuging at 2500 rpm for 2 minutes. Approximately 100 worms were loaded onto each assay plate. Assays with obi-1(tu404) DL could not be successfully performed due to their fragility and higher mortality. To anesthetize animals at the odor sources, we placed 1.5 µl of 1 M sodium azide on both sources and used 100% ethanol as the counter attractant. Multiple assay replicates constituted each experimental trial, and multiple trials were conducted for each condition. All odorants were diluted with ethanol to the following concentrations: 5% (Z)-7-tetradece-2-one (previous studies used 10% ZTDO, which resulted in more DL paralysis at loading), 1% methyl myristate, 1% (E)-11-tetradecenyl acetate, 1% (Z)-11-hexadecenal, and 10% ß-caryophyllene. Interestingly, we found sodium azide to be superfluous when using ZTDO as the attractant due to ZTDO’s paralyzing effects on DL. Chemicals were purchased from Fisher Scientific, Sigma-Aldrich, or Bedoukian Research (Danbury, CT, USA).

CO₂ chemotaxis assays

Young adults were washed off plates using M9 buffer and collected in a 65 mm-Syracuse watch glass. Animals were then washed 2X with M9 buffer followed by 1X with sterile water and immediately transferred to the center of 6 cm MOPS/Tween agar plates. Excess water was removed with Whatman paper. DLs were directly transferred to the 6 cm MOPS/Tween agar plates on chunks of agar from starved plates. A CO₂ gradient was generated by delivering specific compositions of gasses through holes in the plate lids as previously described (Carrillo et al. 2013). A certified gas mixture containing 21% O₂, balance N₂ (Airgas) air control was pumped through one hole whereas a certified gas mixture containing 10% CO₂, 21% O₂, balance N₂ (Airgas) was pumped through the other hole. Gasses were pumped through ¼-inch flexible PVC tubing using a syringe pump (PHD2000, Harvard Apparatus) at a flow rate of 0.5 ml/minutes. Adults and DL were allowed to migrate in the CO₂ gradient for 1 hour and 90 minutes, respectively. Scoring regions were defined as opposite halves of the assay plate 1 cm away from the center line. At the end of the assay, animals that migrated to opposite scoring regions were scored according to the formula:

Chemotaxis index (CI) = [(# animals at CO₂ side) – (# animals at air side)]/(Total # animals at both sides).

As a control for directional bias, two assays were always run simultaneously with CO₂ gradients in opposite directions. Assays were discarded if the difference in the CI between the two plates was ≥0.9 or if fewer than 7 animals moved to both scoring regions combined on either assay plate.

Ppa-hsd-2 transgenic rescue of csu60

15–20 gravid csu60 hermaphrodites with or without the Ppa-hsd-2 transgenic rescue construct were picked onto plates with food, and allowed to lay eggs for 24 hours before removal to obtain a roughly synchronized population of embryos. The plates were incubated at 20°C for 5 days, then scored for DLs and nondauer stages, as well as for the expression of the co-injection marker Ppa-egl-20p::rfp as a proxy for the presence of the Ppa-hsd-2 rescue transgene. Interestingly, we found that this method of synchronizing embryos led to a higher percentage of the population remaining as DL, compared to not synchronizing the embryos used in the temperature sensitivity assay.

Statistical analysis

Statistical tests were analyzed using GraphPad Prism 8 (La Jolla, CA, USA) and Microsoft Excel. We used t-tests or ANOVAs with post-tests for parametric data, and Kruskal-Wallis tests with post-tests for nonparametric data.

Data availability

Strains and plasmids are available upon request. Genome and transcript sequences are available on www.wormbase.org and www.pristionchus.org. The Reagent Table contains additional details on materials and databases. The csu60 whole genome-sequence is available in figshare at DOI: 10.6084/m9.figshare.14502177. Supplementary material is available in figshare: https://doi.org/10.25386/genetics.14498556.

Results

Daf-c mutations act upstream of the conserved DAF-12 NHR

We isolated a constitutive dauer-formation mutant at 20°C, csu60. When compared to the only other viable but uncloned Daf-c allele, Ppa-dfc-1(tu391) from a previous study, we found that both alleles are incompletely penetrant, recessive, and nonallelic for Daf-c. In C. elegans under reproductive conditions, the binding of the nuclear hormone receptor DAF-12 to steroid hormones, including Δ7-dafachronic acids (Δ7-DA), restrain L2 larvae from becoming DL, which is a conserved function among diverse nematodes, including P. pacificus (Motola et al. 2006; Ogawa et al. 2009). To determine if the Daf-c phenotype of csu60 or tu391 is due to a deficiency in DA, we tested a synthetic DA for its ability to rescue the Daf-c phenotype. Previous work has shown that the constitutive dauer formation phenotype of dfc-1(tu391) could be rescued by feeding the worms Escherichia coli OP50 containing 7.5 µM Δ7-DA (Ogawa et al. 2009). Synthetic Δ7-DA has been shown to prevent dauer formation in both P. pacificus and the ruminant parasite Strongyloides papillosus (Ogawa et al. 2009). We found that both the dauer-specific exsheathment defect (see below) as well as the constitutive dauer-formation phenotype of tu391 and csu60 were rescued by the availability of Δ7-DA in the food (Figure 1, A and B). These results suggest that both mutants cannot produce sufficient dauer-suppressing Δ7-DA.

Figure 1.

Hormonal and transgenic rescue of Daf-c phenotype. (A, B) The constitutive dauer phenotype in both Daf-c mutant alleles is completely rescued by exogenous Δ7-dafachronic acid but not by the ethanol control. (t-test between mock and DA treated. ****P < 0.0001. Sample size: tu391 control = 435; tu391 treated = 426; csu60 control = 210; csu60 treated = 284). (C) A transgene carrying the Ppa-hsd-2 promoter driving its own cDNA strongly rescues the constitutive dauer and exsheathment defects in csu60 (t-test between transgenic and nontransgenic populations. ****P < 0.0001; Sample size: csu60 = 2246; csu60 with the rescue transgene = 2388). Dauer formation is the percentage of DL in all third-stage larvae or older.

To further confirm this genetic pathway for dauer regulation by epistasis analysis, we constructed csu60; daf-12(tu389) double mutants and found that they resembled the Daf-d phenotype of daf-12(tu389) (Ogawa et al. 2009) (Supplementary Table S1), consistent with the expected position of csu60 acting upstream of the nuclear hormone receptor. Similarly, the dfc-1(tu391); daf-12(tu389) double mutants completely masked the Daf-c phenotypes of tu391, placing tu391 also upstream of daf-12 for the regulation of dauer formation. Hence, daf-12 is epistatic to both csu60 as well as tu391. Curiously, a small percentage (3.7%) of the csu60; daf-12(tu389) double mutants exhibited an ecdysis defect in the molts between the late larval stages (J3-J4; J4-adult), rather than J2-DL, which is a synergistic phenotype not found in either mutant alone (Supplementary Table S1). Both Daf-c mutations thus act through the Δ7-DA/DAF-12 module critical for DL formation.

Molecular cloning of the first P. pacificus Daf-c locus

Using the latest genome assembly of P. pacificus, we performed whole-genome sequencing of csu60 and identified a ∼95 kb deletion on Chromosome II that affected 10 predicted protein-coding genes, 6 of which have transcript support (Werner et al. 2018). Among these candidates, the predicted protein sequence of PPA10139 is orthologous to the 3Δ-hydroxysteroid dehydrogenase/Δ5-Δ4 isomerases (3Δ-HSDs) that are key steroidogenic enzymes in vertebrates (GO:0003854). PPA10139 shares 89% similarity with C. elegans HSD-2, which encodes one of three HSDs in C. elegans involved in the biosynthesis of dafachronic acids (Patel et al. 2008; Dumas et al. 2010). Consistent with its potential role in repressing dauer formation, PPA10139 transcript levels are lower in DL relative to the reproductive developmental stages (Baskaran et al. 2015).

To determine if the Daf-c phenotype of csu60 is due to a deficit in Ppa-hsd-2, we made a construct containing a ∼2 kb promoter of PPA10139 driving its full-length cDNA to rescue the csu60 deletion by transgene complementation. Worms containing the transgenic Ppa-hsd-2p::cDNA extrachromosomal array showed dramatically lower constitutive dauer formation, and lower rates of the dauer-specific exsheathment phenotype, than csu60 siblings without the transgene (Figure 1C).

We then investigated the possible evolutionary history of hsd genes in nematodes. We found only one putative ortholog of HSD-2 in each of the completed genomes of the human parasites B. malayi and Onchocerca volvulus, whereas the C. elegans genome contains three paralogs, suggesting that a single HSD is likely the ancestral state (Figure 2A). Each hsd gene in C. elegans differs by the number of transmembrane domains: hsd-1 encodes for two predicted transmembrane domains, hsd-2 encodes for one transmembrane domain, while hsd-3 does not encode for any predicted transmembrane domain (as predicted by SMART utilizing TMHMM). Although the overall protein sequence of PPA10139 shares the highest similarity to C. elegans HSD-2, the two transmembrane domains predicted for Ppa-HSD-2 argues for more functional similarity to C. elegans HSD-1. Furthermore, while the C. elegans hsd-1 is located on Chromosome I, the C. elegans hsd-2 and hsd-3 appear to be recent gene duplications on Chromosome X, suggesting that the appearance of the P. pacificus hsd-2 locus on Chromosome II (syntenic with C. elegans Chromosome II) preceded the divergence of these two species (Rödelsperger et al. 2017).

Figure 2.

Ppa-hsd-2 phylogeny and the Daf-c locus. (A) Orthology assignment for *Ppa-HSD-2 (PPA10139). Maximum likelihood phylogeny trees were generated by PhyML (phylogeny.fr). Midpoint rooting was used and only branch support ≥50% is shown. Protein sequences for P. pacificus are “Ppa-” (orthology same as Wormbase.org) and for C. elegans “Cel-.” Sequences from additional species are Drosophila melanogaster (Dme), O. volvulus (Ovo), B. malayi (Bma), Trichuris muris (Tme), and Homo sapiens (Hs). Dehydrogenases DHS-16 from C. elegans along with DHS-16 and DHS-20 from P. pacificus are closest to the HSDs and are used as an outgroup. (B) JBrowse snapshot of the Daf-c locus on Chr. II shows csu60 and csu88 share a deletion region that encompass Ppa-hsd-2 (PPA10139), Ppa-F35D2.6 (PPA35084), and Ppa-twk-2 (PPA10144) that have transcript support (Rödelsperger et al. 2017).

To determine if the loss of Ppa-hsd-2 is sufficient to cause the Daf-c phenotype, we generated additional alleles by targeted mutagenesis using CRISPR/Cas9 (Witte et al. 2015; Han et al. 2020; Nakayama et al. 2020). We identified another allele, csu88, which did not have a mutation near the crRNA site (Figure 2B), but did exhibit the Daf-c and dauer exsheathment defect characteristic of csu60 (Supplementary Table S1). Using PCR, we inferred a ∼67 kb deletion in csu88 that extended downstream from the Ppa-hsd-2 second intron and confirmed the absence of Ppa-hsd-2 transcripts by RT-PCR (Supplementary Figure S1). Complementation test shows csu88 could not complement csu60, and like csu60, the Daf-c phenotype of csu88 was also rescued by exogenous Δ7-DA. Thus, a second allele at the Ppa-hsd-2 locus phenocopies csu60, and both transgenic complementation with Ppa-hsd-2 cDNA and Δ7-DA are sufficient to rescue the csu60 Daf-c phenotype.

Ppa-hsd-2 is expressed in the CAN neurons and excretory-secretory system

In vivo measurements in C. elegans indicated that peak Δ7-DA concentration occurs in the L2 prior to the commitment to dauer entry, and that Daf-c mutants have DA levels less than 15% of that in wild-type L2s (Li et al. 2013). To determine the stage-specific expression patterns of Ppa-hsd-2, we examined two independent transgenic reporter strains containing the ∼2-kb promoter of Ppa-hsd-2. We found that the cell type that expressed the transgene most robustly at every developmental stage is the canal-associated neurons (CANs), whose cell body are near the midbody with processes that run along the excretory canal laterally along most of the body (Figure 3; Supplementary Table S2). In C. elegans, CAN neurons are required for survival but little is known about their function (White et al. 1986). Whether they function as neurons is also unclear due to their lack of synaptic connections (White et al. 1986). In contrast, in P. pacificus, the ablation of the CAN neurons resulted in an enhanced response to dauer pheromone (Mayer et al. 2015). All developmental stages also show Ppa-hsd-2p::rfp expression in a few intestinal cells as well as the excretory gland cell, though expression in the excretory canal was detectable only in the dauer stage (Figure 3, A and E–G). We speculate that the CAN neurons P. pacificus could innervate the excretory cells, given that the anterior CAN neuron processes reach the nerve ring (Figure 3A). Thus, the Ppa-hsd-2 expression pattern is most prominent in the excretory-secretory system and in the CAN neurons, which are known to inhibit dauer formation in P. pacificus.

Figure 3.

Ppa-hsd-2 is expressed in multiple cell types. Representative DIC and fluorescent overlay images of (A, B) an adult with Ppa-hsd-2p::rfp expression in the excretory gland cells (*) and in the pair of CAN neurons (arrow). The anterior CAN processes are visible near the pharyngeal isthmus (triangles) (C) The earliest Ppa-hsd-2 expression is in the CAN neurons of J1 larvae prior to hatching. (D) CAN neuron expression in a J2 larva. (E) Intestine expression in a J3 larva. (F) Excretory gland cell and intestine expression in a J4 larva. (G) Ventral max projection of a DL, which is the only developmental stage that shows Ppa-hsd-2 expression in the excretory canal. Images were taken at 250–525 ms exposure. Sample sizes are shown in Supplementary Table S2. Scale bar: 25 µm.

Daf-c mutations enhance adult chemotaxis response

Since there have been no studies on the chemosensory behavior of P. pacificus DL, we sought to take advantage of the high proportion of active Daf-c DL using a modified assay to look for potential behavioral changes between P. pacificus young adults and DL. First, to determine if Daf-c DL behave similarly to wild-type DL, we tested DL of wildtype, tu391, and csu60 with a panel of odorants that are known attractants for P. pacificus adults. We focused on the Oriental beetle pheromone (Z)-7-tetradece-2-one (ZTDO) and the fly aggregation pheromone methyl myristate and found that the DL of wildtype and csu60, and to a lesser extent tu391, exhibited strong attraction to both of these insect host pheromones (Figure 4A). Given this wild-type chemosensory response by the Daf-c DL, we expanded the survey of DL chemosensation to the insect pheromones E-11-tetradecenyl acetate (ETDA) and hexadecenal, as well as the plant defense volatile ß-caryophellene, which are all known attractants for P. pacificus adults, but found that neither Daf-c nor wild-type DL responded to these odorants (Supplementary Figure S2A) (Herrmann et al. 2007; Hong et al. 2008). We also examined the responses of Daf-c DL to a panel of C. elegans attractants: 2,3-butanedione (diacetyl), 2,3-pentanedione, and isoamyl alcohol (Bargmann et al. 1993). We found that like P. pacificus adults, the DL also did not respond to these low molecular weight chemicals (Supplementary Figure S2B). Thus, DL exhibited near wild-type responses to two known insect-associated odors but have a narrower overall odor response profile compared to adults.

Figure 4.

Olfactory responses of P. pacificus DLs. (A) DLs of wild-type and Daf-c mutants show robust attraction to 5% Oriental beetle pheromone z-(7)-tetradecen-2-one (ZTDO) and 1% house fly aggregation pheromone methyl myristate. Two-way ANOVA with Dunnett’s multiple comparisons test to wildtype. Sample sizes: dfc-1 ZTDO = 764, csu60 ZTDO = 324, PS312 ZTDO = 272, dfc-1 myristate = 743, csu60 myristate = 1502, PS312 myristate = 136. For each category, 10–33 assays were performed over at least 2 trials. (B) Ppa-hsd-2 alleles have enhanced adult attraction to the host beetle pheromone ZTDO that is dependent on obi-1. Red asterisks indicate differences to PS312 by genotype using One-way ANOVA with Dunnett’s (adult) or Kruskal-Wallis test (DL); Black asterisks indicate differences between adults and dauer by multiple unpaired t-test using Sidak’s method. Sample sizes: PS312 adults = 272, PS312 DL = 315, obi-1 adults = 219, csu60 adults = 397, csu60 DL = 544, csu60; obi-1 adults = 227, csu60; obi-1 DL = 1352, csu88 adults = 541, csu88 DL = 257. PS312 and csu60 DL are the same data as in 4 A, with extra replicates for csu60 DL. For each category, 13–33 assays were performed over at least 2 trials. (C) PS312 Post-dauer adults show a similar response to ZTDO as the never-dauer adults. Significance is shown relative to the adults. One-way ANOVA with Dunnett’s post-hoc test. Sample sizes: adults = 345, DL = 315, post-DL adults = 255. Adults and DL are same data as 4B, with extra replicates for adults. 12–18 assays were performed over at least 2 trials. (D) The steroid hormone receptor encoded by Ppa-daf-12 is not required for the enhanced attraction to ZTDO in adult csu60. One-way ANOVA with Kruskal-Wallis test. Sample sizes: csu60 = 596, daf-12 = 304, csu60; daf-12 = 1830. 7–16 trials were performed over at least 2 trials. (E) Adults from hsd-2(csu60) cultures grown on exogenous Δ7-dafachronic acid (7DA) exhibited the same enhanced ZTDO attraction as ethanol-treated adults, but the Ppa-hsd-2p::cDNA transgene rescued the enhanced ZTDO odor attraction of csu60 adults to unresponsive (P < 0.0001, two-way ANOVA. Sample sizes: −7DA = 729; +7DA = 1418; no transgene = 727; transgene =358. For each category, 9–16 assay replicates were performed over three trials. Violin plots show median with quartiles.

To better define the genetic pathways that are responsible for the chemosensory differences between the host seeking DL and reproductive adults, we focused on the responses of the Ppa-hsd-2 mutants to a host pheromone, ZTDO (Cinkornpumin et al. 2014). We observed that wild-type PS312 showed the strongest difference between the DL and adults, with an almost threefold stronger attraction in DL than in adults (Figure 4B). We also observed significantly stronger attraction by the adults of both Ppa-hsd-2 alleles compared to wildtype, such that there is no difference between the responses of DL and adult stages of csu60 and csu88 (Figure 4B). Because the passage through the dauer stage may introduce lasting post-dauer changes that alter odorant receptor expression and hence modulate behavior (Sims et al. 2016; Vidal et al. 2018; Pradhan et al. 2019), we also examined the ZTDO response of wild-type post-dauer adults. We found that dauer passage alone did not result in enhanced ZTDO response in adults (Figure 4C). Thus, genes that inhibit dauer development also repress the ZTDO attraction in adults.

To determine how the Ppa-hsd-2 mutants interact with chemosensory mutants mediating the ZTDO response, we examined the chemotaxis behavior of the Ppa-hsd-2(csu60) mutants in the loss-of-function obi-1(tu404) background that is defective for its cGMP-dependent responses to ZTDO (Cinkornpumin et al. 2014). We found that the loss of obi-1 significantly reduced the ZTDO response in csu60 adults as well as DL, though DL still showed more attraction to ZTDO than adults (Figure 4B). Interestingly, while Ppa-daf-12 is epistatic to csu60 in dauer formation, csu60; daf-12(tu389) double mutant adults still exhibited the csu60 enhanced attraction to ZTDO (Figure 4D). Furthermore, whereas exogenous Δ7-DA rescued the Daf-c phenotype but did not rescue the enhanced host pheromone attraction in adults, the Ppa-hsd-2p::cDNA transgene rescued the csu60 enhanced chemotaxis phenotype (Figure 4E). Thus, the regulation of chemotaxis responses can be uncoupled from dauer formation. Taken together, our results suggest that the enhanced pheromone attraction of adult Ppa-hsd-2(csu60) depends on obi-1 but not the Δ7-DA/DAF-12 module.

CO₂ is repulsive to both adults and DL

Since host odor attraction in IJs of entomopathogenic nematodes (EPN) is enhanced by CO₂, we next tested the chemotaxis response of P. pacificus DL to CO₂ (Hallem et al. 2011; Dillman et al. 2012). In C. elegans, well-fed adults are repelled by CO₂, while DLs are attracted to the same concentration (Hallem and Sternberg 2008; Hallem et al. 2011). In P. pacificus well-fed adults, CO₂ also elicits a strong avoidance response which, as in C. elegans, is mediated by a pair of internal gas sensing neurons known as the BAG neurons (Hallem and Sternberg 2008). To determine if CO₂ response valence also differs between P. pacificus DL and adults, we examined the response to CO₂ in Daf-c DL and adults using the modified chemotaxis assay. We found that wild-type P. pacificus DL exhibited a slightly stronger avoidance response than adults, while both Daf-c DL and adults showed equally robust repulsion by CO₂ (Figure 5). Thus, unlike the enhanced response to the beetle host pheromone, wild-type P. pacificus DL did not exhibit a valence change to CO₂. Interestingly, iL3s of the mammalian parasites S. stercoralis, S. ratti, Ancylostoma ceylanicum, and Nippostrongylus brasiliensis are also repelled by CO₂ (Castelletto et al. 2014; Ruiz et al. 2017). It is currently unclear how the CO₂ avoidance response of P. pacificus DL, but not C. elegans DL, factors into the different life strategies of these two species in their respective ecologies.

Figure 5.

P. pacificus is repulsed by CO₂. The response to CO₂ differs slightly between wild-type adults and dauers, but does not differ between adults and DLs in the Daf-c mutants. Significance is shown between adults and DL; Two-way ANOVA with Sidak’s method. Sample sizes: PS312 adults = 363, PS312 DL = 243, dfc-1 adults = 439, dfc-1 DL = 334, csu60 adults = 299, csu60 DL = 155. For each category, 12 assays were performed over three trials. Violin plots show median with quartiles. *P < 0.05; **P < 0.01; ****P < 0.001, ****P < 0.0001.

The Ppa-hsd-2 locus affects amphid neuron identity

Given the finding that P. pacificus HSD-2 is upstream of the putative branching point for dauer formation and olfaction, we sought to investigate possible neuronal type changes associated with the enhanced attraction to ZTDO in csu60 and csu88 adults. Using their connectivity profiles and axon trajectories, recent comparisons of the neuroanatomy between C. elegans and P. pacificus identified homologs of the two pairs of amphid olfactory neurons responsible for odor attraction, AWA, and AWC (Hong et al. 2019). The P. pacificus AWC homolog, AM7, expresses Ppa-odr-7p::rfp (rather than the G-protein-coding odr-3 in C. elegans), whereas the AWA homolog, AM3, expresses Ppa-odr-3p::rfp (rather than the nuclear hormone receptor-coding odr-7 in C. elegans). We found that while Ppa-odr-7p::rfp expression is unaltered in Ppa-hsd-2(csu60) mutants (Supplementary Figure S3, A and B), Ppa-odr-3p::rfp expression is dramatically altered in csu60 in all post-hatching stages (Figure 6). Specifically, in addition to the robust Ppa-odr-3p::rfp expression in AM3(AWA) and the weaker expression in AM4(ASK), 68% of csu60 J4 larvae and adults also showed unexpected expression in a third pair of amphid neurons posterior to the AM3(AWA) (Figure 6; Supplementary Table S3). Ppa-hsd-2(csu88) adults also showed ectopic Ppa-odr-3p::rfp expression in this extra pair of amphid neurons (Figure 6; Supplementary Table S3). The level of Ppa-odr-3p::rfp expression was also stronger in csu60 and csu88 compared to wildtype, although it is not clear if the expression represents a duplication of the AM3(AWA) or the AM4(ASK) cell fate.

Figure 6.

Ectopic Ppa-odr-3 expression in amphid neurons of Ppa-hsd-2 mutants. (A–C) Representative DIC and fluorescent overlay images of J4 and adults. In wildtype, Ppa-odr-3p::rfp expression is found in AM3(AWA) and AM4(ASK) amphid neurons (arrows). In csu60 and csu88 mutants, reporter expression in AM3 and AM4 is stronger than wildtype and is found in an additional pair of amphid neurons posterior to the AM3(*). (D–E) Ppa-odr-3p::rfp expression in wildtype J2 and ectopic expression in csu60 J2. (F, I) Ectopic Ppa-odr-3p::rfp expression in a csu88 DL at different focal planes. (G, H) Ppa-odr-3p::rfp expression in wildtype DL and ectopic expression in csu60 DL. Sample sizes are shown in Supplementary Table S3. Anterior is left and dorsal is up. Scale bar: 25 µm.

We then sought to determine if exogenous Δ7-DA can suppress the ectopic Ppa-odr-3p::rfp expression but found that while the same level was sufficient to suppress the csu60 Daf-c phenotype, Δ7-DA could not suppress the ectopic Ppa-odr-3p::rfp expression (Supplementary Table S3). Finally, to confirm that the increased Ppa-odr-3 promoter activity and ectopic cellular expression correspond to increased overall Ppa-odr-3 mRNA transcripts in csu60, we performed qPCR on mix-stage cultures and found that Ppa-odr-3 transcripts are indeed ∼10-fold higher in csu60 than in wildtype (Supplementary Figure S4).

These findings suggest that HSD activity suppresses a default-on Ppa-odr-3-expressing cell fate prior to hatching in the J1 larvae or during embryogenesis. Because the duplicated Ppa-odr-3p::rfp pair was not found in wild-type DL, which show attraction to ZTDO, the difference in the odr-3p::rfp expression pattern alone cannot not fully explain the stronger ZTDO attraction in DL compared to adults in wildtype, but is consistent with the possibility that the transformed wild-type amphid neuron may be involved in restricting dauer entry.

Pristionchus pacificus Daf-c mutants have a dauer exsheathment phenotype

Akin to the infective L3 stage of parasitic nematodes, we found ∼63% of wild-type P. pacificus DL under laboratory condition are ensheathed (n = 122), in which the sheath is thought to be the apolysed but unecdysed J2 cuticle snugly covering the dauer cuticle. C. elegans DL, in contrast, are only very transiently ensheathed (Cassada and Russell 1975). Strikingly, while the ensheathed wild-type DL are motile, both Daf-c alleles exhibited an unusual cuticle defect in which active larvae are trapped and immobilized inside an old cuticle (Figure 7, A–C). Aside from the loosened cuticle, the “incarcerated” larvae in tu391 and csu60 are indistinguishable from DL produced by the wild-type reference strain PS312 under starvation conditions; they are characterized by radial shrinkage of the body, a darkened intestine, and a buccal cap covering the mouth. In EPN from the genera Steinernema and Heterorhabditis, ensheathment appears to confer increased ability to resist desiccation, while exsheathment leads to increased motility and is the first step of the host infection process (Campbell and Gaugler 1991; Menti et al. 1997). Our results reveal that in P. pacificus, ensheathment is unique to the DL and genetically regulated with dauer entry.

Figure 7.

Incarcerated larvae are temperature-dependent constitutive DLs. Ten young adult hermaphrodites on food were transferred to different temperatures for 5 days, and their progeny were then scored for dauer-equivalent J3 stages. (A) Ensheathed wild-type PS312 DLs. The effete J2 cuticle surrounding the body is visible. PS312 DL does not show the incarcerated phenotype. (B) In an incarcerated tu391 DL, the J2 cuticle is not attached at the head in nearly half of tu391 constitutive DLs. (C) The illustration depicts that the incarceration is transient. (D) Incarcerated DLs (n_tu391 = 22; n_csu60 = 30) exhibit pharyngeal pumping rates similar to active mutant dauer (n_tu391 = 22; n_csu60 = 13) and wild-type P. pacificus and C. elegans dauer (n_Ppa = 24; n_Cel = 18), which are significantly different from that of the J3 and J4 larval stages (PS312 J3 n = 32; tu391 J3 n = 40; csu60 J4 n = 13) (letters denote significant differences at P < 0.001, one-way ANOVA, Tukey’s multiple comparisons post-test). (E) The proportions of tu391 J3 larvae, active DLs, and DLs differ at 15°C (n = 577), 20°C (n = 2091), and 25°C (n = 1622). Wild-type PS312 is 100% J3 at all temperatures. (F) The proportion of csu60 J3 to DLs differ between 15°C (n = 1090) and the higher temperatures 20°C (n = 2846) and 25°C (n = 2516). Incarcerated DLs with exsheathment defect are most abundant at 20°C. Additional details are found in Supplementary Table S1 and Figure S5.

To determine if the incarcerated worms are exclusively in the dauer stage, which does not feed, we measured their pharyngeal pumping frequency (Cassada and Russell 1975). We found that the incarcerated tu391 and csu60 larvae display pumping rates significantly lower than the J3 and J4 stages of wild-type and mutant animals, but indistinguishable from their respective nonincarcerated (active) DL siblings as well as wild-type P. pacificus and C. elegans DL (Figure 7D). We also isolated tu391 and csu60 J3 and J4 larvae onto fresh OP50 to determine if they would develop into incarcerated larvae but did not observe incarcerated worms (n_{J3, tu391} = 60; n_{J4, tu391} = 50; n_{J3, csu60} = 95; n_{J4, csu60} = 100), suggesting that the molting defect in tu391 and csu60 is an exsheathment defect exclusive to the formation of the dauer cuticle. To determine if the incarcerated DL is a permanent or transient phenotype, we picked incarcerated tu391 and csu60 DL onto fresh OP50 plates. We found incarcerated tu391 and csu60 DL became free and active within 24 hours (n_tu391 = 213; n_csu60 = 124). Taken altogether, the incarcerated worm is a transient dauer-specific phenotype.

Because temperature modulates the dauer decision (Swanson and Riddle 1981; Golden and Riddle 1984b), we wondered if the constitutive dauer formation and the dauer-specific exsheathment defects are temperature-dependent. We found that for tu391, lower temperature (15°C) completely suppressed the Daf-c phenotype, while higher temperature (25°C) promoted the highest proportion of DL (92% of third-stage larvae) but dramatically suppressed the dauer incarceration defect (Figure 7E, Supplementary Table S1). Similarly, for csu60, free DL formation was highest at 25°C (66%) compared to 15°C (6%), although the increase at 25°C compared to 20°C (53%) was noticeably less than the effect observed for tu391. Like tu391, the highest proportion of csu60 incarcerated DL occurred at 20°C (6%) (Figure 7F, Supplementary Table S1, Figure S5). Thus, increasing the temperature resulted in higher DL formation in both alleles, while the incarcerated defect was most prevalent at 20°C. It is unclear if higher temperature suppresses the initiation of the exsheathment defect or shortens the duration of the incarceration. These findings indicate that the Daf-c mutants in C. elegans and P. pacificus both promote dauer formation at higher temperatures, but DL formation in csu60 is less sensitive to temperature than in tu391, suggesting the genes responsible for these loci occupy different positions in the P. pacificus dauer regulation pathway.

Daf-c mutants reveal a dauer-committed J2 stage

In addition to the cuticle exsheathment phenotype, we also observed the occurrence of partial DLs with csu60; in contrast, partial DLs were not observed with wild-type or tu391 animals. These incomplete dauers exhibited a dark intestine and lethargy typical of DL, but lacked the radial constriction and buccal cap that also define DL (Figure 8, A and B). After hatching, C. elegans L1 larvae can either enter the reproductive development pathway and become L2 larvae, or they can become pre-dauer L2s larvae when treated with exogenous dauer-inducing pheromones. The pre-dauer L2d larvae can either return to reproductive development or enter the dauer diapause pathway, depending on the concentration of dauer-inducing pheromones (Golden and Riddle 1984a; Ilbay and Ambros 2019). Because P. pacificus completes the first J1-J2 molt prior to hatching and emerges as J2 (Lewis and Hong 2014), we hypothesized that the csu60 incomplete DL belong to a dauer-committed J2 stage, which we designated as J2d. To determine whether the partial DL are pre-DL, or J2 that were unable to complete the morphological transition into full DL, we subjected them to the tests used to define C. elegans DL, namely the ability to resist sodium dodecyl sulfate (SDS) (Cassada and Russell 1975) and lack of latex fluorescent bead uptake due to cessation of pharyngeal pumping (Nika et al. 2016).

Figure 8.

Characterization of csu60 J2d and DLs. (A, B) A csu60 J2d pre-dauer lacks a buccal cap and radial constriction, compared to a csu60 DL. The scale bar is 50 µm. (C) None of the dauer (N = 5, n_PS312 = 127, n_tu391 = 137, n_csu60 = 88) and most of the J3 larvae in (N = 5, n_PS312 = 108, n_tu391 = 82, n_csu60 = 103) took up fluorescent latex beads after 90 minutes, compared to 10% of csu60 J2d (N = 5, n = 85). (D) csu60 worms that do not take up fluorescent beads are resistant to SDS exposure. Only the csu60 J2d (N = 9, n = 227) and DL (N = 6, n = 139) that lack fluorescent bead uptake were exposed to 1% SDS and monitored for movement after 10 and 40 minutes. J2d (N = 8, n = 38) and J3 (N = 6, n = 127) that did ingest fluorescent beads were not resistant to SDS (not shown). Error bars represent s.e.m. (E) Diagram of the P. pacificus alternative developmental pathways that share a pre-hatching molt from J1 to J2, followed by 3 post-hatching molts (m). (top) Dauer development occurs under low food conditions, with reversible commitment to dauer entry in the J2d stage, followed by a dauer-specific second molt (m2’), or resumption of reproductive development (m2). Food triggers dauer exit into post-dauer J3 (J3’). (bottom) Reproductive development occurs in the presence of food.

We found that none of the wild-type, tu391, or csu60 DL ingested fluorescent beads, compared to 84–92% of their J3 counterparts (Figure 8C). A minority of the csu60 J2d (10%) took up the fluorescent beads, likely due to nominal pharyngeal pumping. The lack of bead uptake in the L2d-dauer lethargus was also observed in C. elegans (Nika et al. 2016). Next, we tested whether larvae without beads are also SDS resistant. We found that most of the DL that did not ingest beads survived a subsequent SDS treatment, whereas only 3% of the J2d without beads survived in SDS (Figure 8D, Supplementary Figure S6). This suggests that most J2d needed to undergo the J2-dauer molt that enhanced SDS resistance. Taken together, our results indicate that the prolongation of the pre-dauer J2 stage in csu60 mutants is morphologically, physiologically, and developmentally equivalent to the L2d in C. elegans (Figure 8E).

Discussion

In the wild, Pristionchus species associate with beetle hosts as DLs, which resume reproductive development when the hosts die and microbes become available (Herrmann et al. 2007; Meyer et al. 2017). Consequently, the olfactory preferences of Pristionchus species have very little overlap with those of C. elegans. We show in this study that olfactory profiles are further modulated by a developmental switch that coordinates dauer entry with host odor attraction.

Different steroid hormones mediate dauer formation and host-seeking

In order to effectively compare the olfactory behaviors of adults vs host-seeking DLs, we characterized the first two Daf-c loci in the entomophilic nematode P. pacificus. We found that deletion of the Ppa-hsd-2 gene, the sole HSD in P. pacificus, results in a Daf-c phenotype. This gene is likely responsible for the synthesis of a conserved Δ7-DA steroid ligand, which regulates the conserved nuclear hormone receptor DAF-12. Whereas the Ppa-hsd-2p::cDNA transgene rescued both the Ppa-hsd-2(csu60) Daf-c and enhanced host odor attraction phenotypes, exogenous Δ7-DA rescued only the Daf-c phenotype. One likely possibility is that Ppa-HSD-2 produces at least one other type of steroid hormone that acts through a nuclear hormone receptor different from Ppa-DAF-12, and this unknown DA/NHR module normally restricts the dauer-specific neuronal development from being executed in the adult stage (Figure 9).

Figure 9.

A model for the multiple roles of Ppa-HSD-2 in coordinating dauer entry, molting, and olfaction. From the cholesterol precursor, Ppa-HSD-2 is required for the synthesis of a Δ7-dafachronic acid-like steroid hormone (7DA). Binding of 7DA to the DAF-12 nuclear hormone receptor represses decision for dauer entry. Ppa-HSD-2 may also be required for producing other steroid hormones that inhibit the DAF-12-independent and lipid-binding protein OBI-1-dependent dauer-specific host odor attraction in adults.

While DAs appear to play a conserved role in promoting reproductive growth across nematode species (Motola et al. 2006; Li et al. 2013; Ma et al. 2019a,c), the specificity of HSDs and their biosynthetic products are likely more divergent and operationally dependent on other dehydrogenases (Mahanti et al. 2014), since their paralog count in each genome can range from none (Ascaris and Toxocaris) to one (Pristionchus and Brugia) to three (Caenorhabditis) (Patel et al. 2008; Farris et al. 2019). In C. elegans, genetic dissection of the relative contribution of the steroidogenic enzymes is confounded by the yet undistinguished roles the HSD-2 and HSD-3 paralogs play in dafachronic acid biosynthesis (Farris et al. 2019). In the vertebrate parasites Ascaris suum and Toxocara canis, endogenous Δ4- and Δ7-DA have been detected despite lacking a putative hsd ortholog in their genomes, suggesting biosynthesis of DAs can also occur via other enzymes (Ma et al. 2019b). By comparison, the single copy hsd in P. pacificus may facilitate the structural identification of additional DAs and their cognate NHRs involved in the proper regulation of developmental stage-specific chemosensation.

The CAN neurons may serve as a source of DA hormones

The lack of conservation in the cell types that express hsd homologs between P. pacificus and C. elegans suggests different cell types could function as sources of DA. In C. elegans, hsd genes are expressed in the XXX cells (hsd-1), the intestine and somatic gonad (hsd-2), and the hypodermis (hsd-3) (Patel et al. 2008; Farris et al. 2019). The XXX cells also express other components of the steroid hormones important for dauer regulation, including daf-9 (cytochrome 450) as well as ncr-1 and ncr-2 (cholesterol transporters), and are considered neuroendocrine cells in C. elegans that act as the source of DA along with the hypodermis (Li et al. 2004; Schaedel et al. 2012). By contrast, Ppa-hsd-2 expression begins in the J1 stage in the CAN neurons, with expression in the excretory gland cell and the excretory canals during the DL stage. In C. elegans, the gland cell is not necessary for molting and dauer entry or exit when ablated during the L1 or L2 stages (Singh and Sulston 1978; Nelson and Riddle 1984), although the gland cell of DLs uniquely lack secretory granules observed in nonstarved developmental stages (Nelson et al. 1983). The dauerless gene (dau-1) in P. pacificus is exclusively expressed in the CAN neurons, and both CAN ablation and dau-1 mutation increase dauer formation (Mayer et al. 2015). Thus, both inhibitors of dauer formation, Ppa-hsd-2 and dau-1, are expressed in the CAN neurons. It remains to be determined if the P. pacificus CANs and excretory cells are endocrine tissues that function to produce and maintain proper DA levels.

Steroid hormones direct neuronal cell fate

The ectopic expression of Ppa-odr-3p::rfp in csu60 and csu88 shows that Ppa-HSD-2 is responsible for maintaining proper cell fate in putative olfactory neurons in P. pacificus. However, the addition of 7DA could not rescue this phenotype, suggesting other DAs made by Ppa-HSD-2 may be involved in regulating Ppa-odr-3 expression. The homeotic transformation of the AWB to AWC neurons due to mutations in the LIM-4 transcription factor was accompanied by changes in dendritic morphology and chemotaxis behavior (Sagasti et al. 1999). Based on a similar rationale, the loss of Ppa-hsd-2 could transform AM10(ASI) or AM6(ASG), which are normally involved in inhibiting dauer entry in C. elegans, into AM3(AWA) olfactory neurons. Coincidentally, AWA and ASG are sister cells in the C. elegans cell lineage (www.wormatlas.org; last accessed May 2021). However, we did not observe fewer Ppa-che-1p::rfp expression in the amphid neurons of Ppa-hsd-2(csu60) animals (Supplementary Figure S3, C and D), as would be expected if a pair of the neuron mis-expressing Ppa-odr-3p::rfp was derived from the AM6(ASG) or AM5(ASE) neurons (Hong et al. 2019). Given the dynamic transcriptional profiles associated with the dauer stage in P. pacificus (Sinha et al. 2012), Daf-c mutants have the potential to identify genetic circuits coordinating neuronal fate and dauer regulation.

Dauer exsheathment is coupled with dauer entry

Retention of the previous cuticle, or ensheathment, is a hallmark of the infective larvae of many parasitic nematodes as well as insect larvae. In nematodes, the exsheathment of infective larvae in mammalian parasites and EPN often marks the transition to the parasitic phase of the life cycle. Cues for exsheathment are a combination of intrinsic physiological factors, such as the time since emergence from the host, and extrinsic environmental factors, such as temperature or host cues and thus may not be necessarily coupled to dauer entry (Campbell and Gaugler 1991; Wharton 1991; Patel and Wright 1998). For example, exogenous application of an endogenous Δ7-DA promotes the exsheathment of iL3 larvae (Ma et al. 2019c). Although there are no candidate genes known to control dauer exsheathment in C. elegans– because its dauers do not hold on to their L2 cuticle as long as other species-conserved genes involved in molting may nevertheless modulate the formation of the dauer cuticle. We speculate that the early detachment of the old cuticle in the Daf-c DL leads to loss of contact with the mouth plug, resulting in the incarceration of the DL inside the old J2 cuticle. Because dauer entry involves the formation of a buccal cap, the retention of the J2 cuticle in P. pacificus DL may render Daf-c DL especially prone to the incarcerated phenotype by disrupting the timing or manner of exsheathment.

Although both Caenorhabditis and Pristionchus species co-occur on rotten vegetation as both DL and feeding stages (Félix et al. 2018), only DL from Pristionchus species have been isolated from beetles (Weller et al. 2010; Meyer et al. 2017). The robust chemoattraction to insect-associated pheromones shown by this study supports the model that the DL likely functions as the host-seeking stage of Pristionchus. We expect to have a better understanding of how the decision for dauer entry is genetically coordinated with host-associated remodeling of behavior when the cognate olfactory neurons for the host pheromones are identified in P. pacificus. Comparisons of the genes regulating dauer development in C. elegans and P. pacificus could contribute towards a better understanding of how this ancient nematode phenotypic plasticity contributes to the evolution of host associations.