Very little is known about the genetic modulation of nictation, a host-finding strategy used by infective juveniles of many pathogenic nematodes, as well as a phoretic strategy of Caenorhabditis elegans dauers. We here use a mass spectrometry-driven approach to find neuropeptides involved in the modulation of this behavior. We identified 126 neuropeptides in infective juveniles of the commercially relevant entomopathogenic nematode Steinernema carpocapsae, 75 of which have orthologs in the model organism C. elegans. To prioritize candidates for causality testing in light of the stage-restricted nictation behavior, we developed a targeted approach to carry out quantitative neuropeptidomics measurements in C. elegans. Our results show that virtually all 164 quantified neuropeptides are more abundant in dauers compared to L3 juveniles. Amongst these, we identify at least the neuropeptide genes flp-7 and flp-11 as novel regulators of nictation.

PRM method development - Synthetic versions of all 427 known and predicted neuropeptides in C. elegans were subjected to a data-dependent acquisition LC-MS/MS workflow. This was used to create a parallel reaction monitoring method for the quantification of neuropeptides. As a proof of principle, the method was tested to four technical replicates of C. elegans neuropeptide extractions. Relative quantification of neuropeptides in L3 vs dauer - The developed PRM method was used to quantify 300 neuropeptides at once in L3 and dauer stage C. elegans worms (four independent replicates each).

For quantitative samples (synchronous C. elegans L3 and dauer juveniles), 500 µL of 1.4 mm ceramic beads (zirconium oxide, Bertin Instruments, France) were added to each pellet together with 1 mL of 90/9/1 methanol/milli-Q water/acetic acid, and then homogenized for 10 cycles of each 15 s at 6800 rpm with 45 s rest, in a Precellys homogenizer with Cryolys cooling unit (kept at -15 °C to -25 °C, Bertin Instruments, France). Of the three frozen aliquots per liquid culture, two were homogenized and then combined to continue with 1 mL biological material per sample. Each sample was topped up to 10 mL with the 90/9/1 solution. Further extraction was performed as of the sonication step as described above. For retention time calibration, iRT standard peptides (Biognosys AG, Switzerland) (Escher et al., 2012) were added as specified by the manufacturer’s protocol. Samples were dried (Thermo Savant SpeedVac) and stored at -80 °C until use (1-3 weeks), when they were dissolved in 45 µL of 2% acetonitrile with 0.1% formic acid prior to analysis. The mixed-stage samples that were used for developing the peptidomics method were done as described for the quantitative method except using a glass homogenizer instead of using the Precellys homogenizer. For the quantitative samples, targeted parallel reaction monitoring (PRM) measurements were performed on a Dionex UltiMate 3000 UHPLC coupled online to a Q Exactive HF-X mass spectrometer (Thermo Scientific). Two separate injections were performed, each time loading 5 µl of sample on a trap column (ReproSil-pur C18-AQ, 5 µm, Dr. Maisch, 20 mm x 75 µm, self-packed) at a flow rate of 5 µL/min in 0.1% FA. After 10 minutes of loading, samples were passed onto the analytical column (ReproSil Gold C18-AQ, 3 µm, Dr. Maisch, 450 mm x 75 µm, self-packed), from which they were eluted at a flow rate of 300 nL/min, using a 50 min linear gradient from 4 to 32% solvent B (0.1% formic acid and 5% DMSO in acetonitrile) in A (0.1% formic acid in 5% DMSO). Analytes eluting from the HPLC undergo electrospray ionization right before introduction into the MS instrument. Full scan MS1 spectra were recorded in the 360 to 1300 m/z range with a resolution of 60,000 at 200 m/z, using an AGC target of 3e6 and maximum injection time of 100 ms. A normalized collision energy of 26 was used for HCD, and targeted MS2 spectra were acquired with a resolution of 15,000 at 200 m/z, an AGC target value of 1e6, a maximum injection time of 22 ms and an isolation window of 0.9 m/z. Depending on the run, the number of targeted precursors was adjusted to not exceed 50 or 70-80 simultaneous targets at a given retention time. For quantitative analysis, first a C. elegans neuropeptide spectral library had to be created. An in-house collection of 427 synthetic neuropeptides, making up the known and predicted C. elegans neuropeptidome was subjected to data-dependent acquisition on a Dionex UltiMate 3000 coupled online to a Q Exactive HF-X mass spectrometer (Thermo Scientific) and analyzed with PEAKS Studio X+ (v.10.5, Bioinformatics Solutions Inc., Canada). For de novo searches, parent mass error was set at 10 ppm, with a fragment mass error of 0.04 Da. Enzyme was set to “none” and the following variable posttranslational modifications were taken into account: oxidation of methionine (+15.99 Da), pyroglutamation of N-terminal glutamic acid (-18.01 Da) or glutamine (-17.03 Da), C-terminal amidation (-0.98 Da) and half of a disulfide bridge on cysteine (-1.01 Da). For PEAKS DB searches, the same error tolerances and posttranslational modifications were applied. Here, the enzyme parameter was again set to “none” and digest mode to “unspecific”. A spectral library was built from these data using Skyline-daily (64-bit, version 21.1.1.223) (Pino et al., 2020). The quality of the spectral library was manually evaluated, taking into account datapoints across the peak, peak shape and dotp value.For quantitative analysis, first a C. elegans neuropeptide spectral library had to be created. An in-house collection of 427 synthetic neuropeptides, making up the known and predicted C. elegans neuropeptidome was subjected to data-dependent acquisition on a Dionex UltiMate 3000 coupled online to a Q Exactive HF-X mass spectrometer (Thermo Scientific) and analyzed with PEAKS Studio X+ using the same settings as described earlier. A spectral library was built from these data using Skyline-daily (64-bit, version 21.1.1.223), and was manually curated to retain only the good-quality fragmentation spectra (based on number of fragmentation peaks and intensity). All PRM data were analyzed using Skyline-daily (64-bit, version 21.1.1.223). For all target neuropeptides, the most intense precursor charge state and the six most intense fragment ions were selected automatically by Skyline from the neuropeptide spectral library. Raw PRM data were imported into Skyline and reviewed: if necessary, integration boundaries were manually adjusted and strongly interfering transitions were removed from the dataset, keeping at least 5 transitions per peptide. Peptide ion peaks were only retained for differential analysis when detected in all replicates of both the L3 and dauer conditions, or when detection was limited to all replicates of only a single condition (reflecting on/off-type differences between conditions). Since absence of a suitable peak to integrate results in missing values (NA), integration boundaries for signals below the detection limit were set around similar retention times as those in samples in which they were successfully detected, effectively integrating the background signal. In some cases, this resulted in a zero value (0) because of a flatline signal. In these cases, any remaining zero values for the area under the curve of the fragment ions were replaced by the minimum value of the entire dataset. All data were exported from Skyline and further analyzed with R (Core R Team, 2019). In cases were multiple ions were observed(Core R Team, 2019). Any remaining zero values for the same neuropeptide sequence (due to different charge state or the presence of a oxidized methionine), the total fragment area of all the respective ions was summed. Subsequently, fragment peak areas of each run were normalized using the median abundance of all MS1 features. After log10 normalization, the peptide precursor abundancies of the individual peptides were compared and significance was determined using a Student’s t-test.