Animal behavior is strongly context-dependent and behavioral performance is often modulated by internal state. In particular, different social contexts can alter anxiety levels and modulate social behavior. The vertebrate-specific neuropeptide parathyroid hormone 2 (_pth2_) is directly regulated by the presence or absence of conspecifics in zebrafish. As its cognate receptor, the parathyroid hormone 2 receptor (_pth2r_), is widely expressed across the brain, we tested fish lacking the functional Pth2 peptide in several anxiety-related and social paradigms. Rodents lacking PTH2 display increased anxiety-related behavior. Here we show that the propensity to react to sudden stimuli with an escape response is increased in _pth2-/-_ zebrafish, consistent with elevated anxiety. While overall social preference for conspecifics is maintained in _pth2-/-_ fish until the early juvenile stage, we found that both social preference and shoaling are altered later in development. The data presented suggest that the neuropeptide Pth2 modulates several conserved behavioral features, and may thus enable the animal to react appropriately in different social contexts.
Different aspects of animal behavior, such as aggression (Tulogdi et al., 2014; Zelikowsky et al., 2018; Agrawal et al., 2020), anxiety (Meyer et al., 2017; Shams et al., 2017), or interaction with conspecifics (Shams et al., 2018; Groneberg et al., 2020), are strongly modulated by social context. In several cases, neuropeptides have been shown to regulate this behavioral plasticity (Zelikowsky et al., 2018; Agrawal et al., 2020; Gemmer et al., 2021). Recently, we found that the expression of the neuropeptide _pth2_ (initially described as tuberoinfundibular peptide of 39 residues or _tip39_ (Usdin, 1997)) is quantitatively regulated by the density of conspecifics in zebrafish (Anneser et al., 2020). In rodents, _pth2_ is involved in the regulation of maternal behavior (Coutellier et al., 2011), pain processing (Dimitrov et al., 2013), oxytocinergic signaling (Cservenák et al., 2017), and fear learning (Coutellier and Usdin, 2011). In zebrafish, the cognate receptor of this neuropeptide, _pth2r_, is expressed in approximately 10 % of all neurons, suggesting a potential broad-scale influence (Anneser et al., 2020). It has not been described, however, whether the presence or absence of _pth2_ alters behavior or other biological features in teleosts.
To identify potential roles of _pth2_, we obtained a line (_pth2 sa23129_) carrying a premature stop codon which terminates translation at amino acid 55 (out of 157) (Kettleborough et al., 2013). In the experiments described below, we tested how the absence of Pth2 influences anxiety-related and social behaviors in larval and late juvenile zebrafish.
To investigate the potential roles of Pth2, we in-crossed heterozygous mutants and analyzed the viability of homozygous _pth2-/-_ fish by observing them for several weeks. We validated the complete absence of Pth2 both by sequencing and by immunostaining (Figure 1). The loss of Pth2 did not lead to differences in either the survival rate or body length (Figure S1). This observation is consistent with data from rodents, in which the deletion of PTH2 did not induce obvious external phenotypic effects except for decreased fertility (Usdin et al., 2008). However, in contrast to rodents, matings between _pth2-/-_ zebrafish led to viable offspring.
Figure S1. Survival rate and growth are not affected by _pth2._ Relates to Figure 1.
(A) Kaplan-Meier curve shows the survival rate of _pth2-/-_ and wildtype fish over the first two weeks of development. Shaded areas indicate the .95 confidence interval. For both groups, n = 80. We analyzed the curves with a log-rank test and found no significant difference (p χ2=1.15= 0.28), supported by a difference in mean restricted survival times between groups of 0.025 days. (B) Box plots highlight the average body length at 2 weeks post fertilization. For both groups, n = 20. Distributions were compared using a one-sided, unpaired t-test (p t=0.7= 0.48).
Figure 1. Validation of _pth2_ sa23129 fish lacking a functional peptide.
(A) The _pth2_ gene consists of three exons; the coding sequence (CDS) is contained in exon 2 and 3. In _pth2_ sa23129 fish, a T/A point mutation induces a premature stop codon within the propeptide sequence, thus preventing the successful translation and cleavage of the Pth2 peptide. (B) Homozygous mutants were validated by sequencing and the premature ochre stop codon was identified at base pair 165 out of 474 of the CDS. (C) Transcript levels of _pth2_ were strongly decreased in _pth2-/-_ animals, as demonstrated by fluorescent in-situ hybridization, also some remaining signal was detected (arrowheads). Abolished Pth2 translation was validated using a custom antibody recognizing a 27-mer epitope immediately preceding the peptide sequence (Anneser et al., 2020). For both wildtype and _pth2_-/-, 6 larvae at 6 dpf were imaged. Scale bar: 10 µm. See also Figure S1.
In rodents, administration of PTH2 has anxiolytic effects (LaBuda, Dobolyi and Usdin, 2004) and animals lacking PTH2 show increased anxiety-related behavior (Fegley et al., 2008). A common assay to investigate anxiety-related behavior in zebrafish is the startle response (Burgess and Granato, 2008; Reider and Connaughton, 2015; Tomasi et al., 2020), in which fish are presented with a stimulus and react with stereotyped escape responses, so-called C-starts (Burgess and Granato, 2007). Previous work showed that higher rearing densities in zebrafish decreased the likelihood of a startle response (Burgess and Granato, 2008). As _pth2_ levels are strongly influenced by the number of conspecifics (Anneser et al., 2020), we tested _pth2-/-_ and _pth2+/+_ fish at 5 days post fertilization (dpf) in a startle paradigm and measured when and how often they responded to a vibrational cue with an escape (Figure S2A and Figure 2A). We observed a typical bimodal distribution characterizing the startle response with fish either exhibiting a short-or long-latency response (Burgess and Granato, 2007) (SLC or LLC, Figure 2B) mediated by Mauthner-cells (Burgess and Granato, 2007) or a prepontine cell population, respectively (Marquart et al., 2019). When we compared the time-to-response after stimulus onset between _pth2-/-_ and _pth2+/+_ animals, we found SLCs occurred moderately but significantly earlier in mutants (Figure 2C), suggesting that _pth2-/-_ fish respond more quickly to potential threats (Troconis et al., 2017). In general, _pth2-/-_ fish showed a higher overall startle response rate (see Figure 2D). We analyzed how the proportion of SLCs, LLCs, and non-responding animals per trial were related to the genotype of the fish with an ANOVA and found significant interactions between these ratios, indicating that for both SLCs and LLCs, _pth2-/-_ fish showed higher response rates (Figure 2E). Furthermore, in a logistic regression model relating response ratios and genotype we found the increased overall response rate of mutant fish to be a significant predictor of the _pth2-/-_ genotype (Figure S2B-D).
Figure S2. In-depth analysis of startle response. Relates to Figure 2.
(A) Experimental scheme for the startle experiment. Both the bass shaker for startle cue delivery and the high-speed camera were triggered by an Arduino board at the same time to ensure alignment of videos and cue. (B) Graph shows the bootstrapped accuracy of the logistic regression model fit to the data. We trained the model 1,000 times sampling 70 % of the full dataset as training set and testing the model on the remaining 30 % as test set. The vertical green line indicates chance level of correct assignment of genotype. (C) Confusion matrix of the full model, highlighting the number of true and false predictions. (D) The fraction of SLCs, LLCs, and failures to respond were projected onto the space of the first two principal components to facilitate visualization. Wildtype are plotted as green dots, pth2-/- in magenta. The background color indicates the decision boundary computed by the logistic regression model.
