Summary

Graphical abstract

Open in a separate window

Introduction

Proper function of the animal nervous system depends on secretion of chemical neurotransmitters such as acetylcholine (ACh) from the presynaptic terminal into the synaptic cleft, and binding of these transmitters to their respective receptors in the postsynaptic compartment. The molecular mechanisms behind this dynamic process and how it is regulated are still not fully understood1; however, synaptic efficacy is often controlled by neuromodulators.2^,3 Neurotransmitters are stored in small organelles termed synaptic vesicles (SVs) from which they are released by the neuron through exocytosis. Fusion of SVs with the plasma membrane at the active zone of the synapse is triggered by depolarization and the resulting increase in Ca²⁺ concentration. SV exocytosis is promoted in response to intracellular signaling via cyclic adenosine monophosphate (cAMP) and protein kinase A (PKA) signaling, mediated by diverse synaptic PKA targets that induce SV mobilization and SV priming, such that increases in Ca²⁺ concentration have a higher chance of triggering SV fusion events.4 Some known PKA targets are synapsin,5 tomosyn,6 Rim1,7 ryanodine receptor,8 cysteine string protein,9 snapin,10 complexin,11 and SNAP-25.12 In general, it is beneficial if synaptic transmission can be modulated, to adapt to the demands of the current situation, or to alter synaptic efficacy in the long term. On the presynaptic side, modulation of transmission can be mediated by altering the response of the presynaptic machinery to changes in the Ca²⁺ concentration, thus altering the rate at which SVs are mobilized from the reserve pool13 or varying the (quantal) content of neurotransmitter per SV.14 Furthermore, neuromodulators can affect presynaptic transmission, e.g., by altering the number of SVs being produced.3 Postsynaptic modes of regulating signal efficacy have also been described, for example, changes in the amount of postsynaptic transmitter receptors and their interaction with scaffolding proteins, membrane resistance, or other modes of excitability, like upregulation of voltage-gated ion channels.15^,16^,17

Although synaptic architecture varies between cell types and species, the ultrastructure of synapses and the composition of SVs and the “pools” into which they are organized are conserved between invertebrates and vertebrates including zebrafish (Danio rerio).18 Structures at neuromuscular junctions (NMJs), the interface between cholinergic motor neurons and skeletal muscle fibers, have been observed in zebrafish larvae using ultrastructural imaging techniques. As in other organisms, they comprise SVs, dense core vesicles (DCVs) containing neuropeptides, and clathrin-coated vesicles.19 Like other peptide hormones, neuropeptides, which are stored in, and released from, DCVs, originate from longer precursor proteins, which are packaged into vesicles in the trans-Golgi network, that are subsequently acidified. Within the DCV, neuropeptide precursor proteins are converted into small bioactive peptides by pro-protein convertase cleavage, and many, but not all, bioactive peptides are C-terminal trimmed by carboxypeptidase E (CPE), before being released at synaptic terminals.20^,21^,22 In endocrine cells, CPE appears as two isoforms. Full-length CPE associates with membranes in a pH-dependent manner, while a C-terminally truncated, yet enzymatically more active, soluble form is found in the vesicular lumen.23^,24^,25 Although the enzymatic function of CPE during maturation of most neuropeptides is widely accepted, further roles of CPE during sorting of proteins into the regulated secretory pathway have been suggested and loss of CPE was proposed to lead to missorting of peptides into the constitutive pathway.26^,27^,28

In zebrafish, remarkably diverse functions have been associated with neuropeptides. For example, the expression levels of parathyroid hormone 2 (pth2) were shown to mirror the presence of conspecifics and also the density of the swarm.29 Furthermore, galanin (galn) expression in the brain is essential for the generation of color patterns by self-organization of pigment cells,30 while tachykinins (tac) are broadly expressed in the central nervous system and have a supposed role as neuromodulators.31^,32 In mammals, tachykinins perform diverse roles, as neuromodulators, or as regulators of pain, stress, sensory processing, inflammation, etc.33 The human Tac1 precursor gives rise to substance P, neurokinin A, and neuropeptides κ and γ. For zebrafish, the consequences of a broad deficit of neuropeptides, as expected for the cpe knockout, have not been described as yet.

The development of optogenetic tools has revolutionized the field of neuroscience by enabling researchers to manipulate specific neurons or neuronal populations. Today, optogenetic methods are standard in many model organisms, and thus optogenetic tools to influence neuronal activity have also been established in zebrafish.34^,35 Microbial rhodopsins for de- and hyperpolarization were used to study the function of sensory neurons36 or central pattern generators37 as early as day one of embryonic development, allowing to achieve calibrated light-induced currents and behavior in different sets of neurons, among them motor neurons.34 In addition to light-activatable ion channels and pumps, other optogenetic tools have emerged in recent years. One class of tools for optogenetic manipulation are photoactivated adenylyl cyclases (PACs) such as bPAC, mediating light-dependent synthesis of the second messenger cyclic AMP (cAMP) for precise temporal and spatial control and acute tuning of intracellular cAMP concentrations.38^,39^,40^,41 bPAC was shown to be enzymatically active in early zebrafish embryos42 and has been applied to analyze the role of cAMP signaling in axonal regeneration or to induce swimming behavior by activating hindbrain reticulospinal V2a neurons.43

We have previously analyzed the effects of cAMP generation in presynaptic terminals of cholinergic motor neurons in the nematode Caenorhabditis elegans. This uncovered an as yet unknown role of neuropeptides in the regulation of synaptic transmission and plasticity,14 which, apart from increasing the rate of SV release, also affected the filling state of SVs with ACh. Thus, synaptic transmission can be controlled by two regimes: depolarization and Ca²⁺ concentration control the acute fusion of SVs, while cAMP and peptidergic signaling alter the transmitter content per SV. This way, in addition to network activity of central pattern generators, motor neurons can integrate pathways of neuromodulation in different systemic states.

Expression of bPAC and the consequences of increased synaptic cAMP levels have not been described in zebrafish motor neurons to our knowledge. In this study we explored whether the dual control of neuronal transmission, as described in C. elegans,14 is conserved also in vertebrates, specifically in zebrafish. We expressed bPAC in spinal motor neurons and photostimulated larvae at different developmental stages. This evoked exaggerated, but coordinated, locomotion, as well as increased SV fusion rates and miniature excitatory postsynaptic currents (mEPSCs). We generated deletion alleles of the main neuropeptide-processing enzyme, cpe, as well as of the most abundantly expressed neuropeptide in cholinergic motor neurons, tac1. Both mutants showed enhanced behavioral effects in response to bPAC stimulation. Likely, this occurred in response to presynaptic defects, like smaller mEPSC amplitude and lower SV release rates, as a consequence of postsynaptic compensation. The latter involves increased expression of nAChRs in the neuromuscular endplates. Thus, mechanisms of cholinergic neuromodulation by neuropeptides appear to exist also in zebrafish, even though the details seem to differ between vertebrates and invertebrates.

Results

Activation of bPAC in cholinergic motor neurons enhances locomotion activity

In order to analyze the effect of ChR2 and bPAC stimulation in embryonic and larval zebrafish motor neurons, we generated transgenic lines expressing the respective optogenetic tools. The mnx1 promotor44^,45 has been described to be specifically active in primary and secondary motorneurons (PMNs and SMNs, respectively). We took two approaches: (1) we used the mnx1 promotor fragment to directly express a fusion protein consisting of bPAC and EGFP (Tg[mnx1En:bPAC-egfp]); (2) we harnessed the Tg[mnx1:Gal4] driver line as a means to express ChR2 and bPAC from the Tg[UAS:ChR2-P2A-dTomato] and Tg[UAS:bPAC-V2A-mCherry]42 transgenes, respectively. EGFP and dTomato fluorescence were clearly visible in PMN cell bodies and axons as early as 24 h post fertilization (hpf) (Figure 1A, left panels). Corresponding expression of the ChR2 and bPAC transgenes in PMNs and SMNs could be observed at 4 days post fertilization (dpf) (Figure 1A, right panels).

Open in a separate window

Figure 1

Expression and stimulation of ChR2 and bPAC in zebrafish motor neurons alter locomotion behavior

(A) Confocal images of Tg[mnx1:Gal4]/Tg[UAS:ChR2-P2A-dTomato] and Tg[mnx1En:bPAC-egfp] transgenic embryos (24 hpf) and of Tg[mnx1:Gal4]/Tg[UAS:ChR2-P2A-dTomato] and Tg[mnx1:Gal4]/Tg[UAS:bPAC-V2A-mCherry] double transgenic larvae (4 dpf). Asterisks mark motor neuron cell bodies, arrowheads motor axons.

(B) Spontaneous (dark) and evoked (light) tail movements (STMs/ETMs) in 24 hpf embryos.

(C) Head-tail angles in immobilized larvae under dark and blue-light conditions. The maximum head-tail angles were determined for wild-type (WT), bPAC (Tg[mnx1: Gal4]/Tg[UAS: bPAC-V2A-mCherry]), and ChR2 (Tg[mnx1:Gal4]/Tg[UAS:ChR2-P2A-dTomato] transgenic larvae (4 dpf).

(D) Bending angles of freely swimming larvae under dark and photostimulated conditions. The bending angles were determined for animals as in (C).

(E) Swimming speed was measured for individuals 4 dpf before, during, and after blue light stimulation.

(F) Average swimming speed with and without blue light stimulation.

(G) Distance traveled during the photostimulation phase. In (B), median (cross), 25/75 quartiles, and min to max are shown; in (C) and (E), means ± SEM; in (D) and (F and G), median and 25/75 quartiles (thick and thin black lines), min to max are shown. in (B), Kruskal-Wallis test with Dunn post hoc test.; in (C–F), two-way ANOVA; in (G), one-way ANOVA, each with Tukey multiple comparisons of means. Statistical significance given as p < 0.0001, p < 0.001, p < 0.01, p < 0.05.

See also Figure S1.

To test whether cholinergic motor neurons can be activated and neuromuscular signaling increased by elevated cAMP levels, behavioral assays of locomotion phenotypes were used. bPAC enzymatically generates cAMP upon activation by blue light of approximately 441 nm.40 cAMP signaling, through PKA and its presynaptic targets, is expected to evoke increased neuronal activity, also in zebrafish motor neurons. Similarly, ChR2 depolarizes neurons during stimulation with blue light46 and is thus expected to evoke motor activity.35^,47^,48 Blue light stimulus frequency and intensity were optimized for the application of ChR2 and bPAC in accordance with previous work by others.34 Motor activity in zebrafish embryos starts around 17 hpf with a side-to-side movement of the tail, a characteristic behavior called spontaneous tail coiling.49 Since this behavior is mediated by motor neurons and interneurons in the spinal cord and depends on slow muscles,50 we reasoned that it could be a possible early measure for optogenetically evoked neuromuscular activity as well. Wild-type and transgenic zebrafish embryos at 24 hpf showed a similar baseline coiling rate under dark conditions (Figures 1B and S1A). Blue light stimulation of wild-type animals did not evoke significant changes in behavior; however, for both, ChR2 or bPAC, blue light induced a significantly increased coiling rate in transgenic embryos, as compared to the respective condition without stimulus (Figures 1B and S1A; Videos S1, S2, and S3).

Video S1: Light induces no twitching in wild-type embryos, related to Figure 1:

Wild-type embryos, 24 hpf, in chorion. Light stimulation begins after 30s

Click here to view.^{(10M, mp4)}

Video S2: Photoactivated ChR2 induces twitching in wild-type embryos, related to Figure 1:

Embryos expressing ChR2 in motor neurons, 24 hpf, in chorion. Light stimulation begins after 30s

Click here to view.^{(10M, mp4)}

Video S3: Photoactivated bPAC induces twitching in wild-type embryos, related to Figure 1:

Embryos expressing bPAC in motor neurons, 24 hpf, in chorion. Light stimulation begins after 30s

Click here to view.^{(6.7M, mp4)}

To analyze if activation of bPAC and, consequently, the increase of cAMP levels or the stimulation of ChR2 in motor neurons also trigger any locomotion changes in older zebrafish larvae, we performed further assays. Observing induced locomotion in partially immobilized larvae is an established way to determine if a certain input triggers, e.g., an escape response or augmented swimming attempts.35^,51 To this end, we measured the maximum deviation of the tail tip from the body axis, namely the angle between head and tail, in 4 dpf head-mounted animals, in the dark, and following blue light stimulation. Whole-field illumination and respective activation of bPAC and ChR2 in all motor neurons led to large amplitude tail bends (Figure 1C; Videos S4, S5, and S6), while wild-type larvae did not exhibit any change in behavior and no escape responses. This indicates that bPAC as well as ChR2 can specifically activate MNs. We did not observe any seizure- or paralysis-like behavior in this assay. This indicates that MN activity was not induced by massive depolarization but, rather, that their membrane potential was elevated, such that intrinsic circuit activity evoked behavior more readily, yet in a coordinated way, to trigger muscle contraction. Based on the deeper bending angles, contractions were more pronounced for ChR2 stimulation.

Video S4: Light induces no tail beating in wild-type larvae, related to Figure 1:

Head-mounted wild-type larva, 4 dpf. Light stimulation begins after 10s, as indicated.

Click here to view.^{(80K, mp4)}

Video S5: Photoactivated ChR2 induces tail beating in larvae, related to Figure 1:

Head-mounted larva expressing ChR2 in motor neurons, 4 dpf. Light stimulation begins after 10s, as indicated.

Click here to view.^{(101K, mp4)}

Video S6: Photoactivated bPAC induces tail beating in larvae, related to Figure 1:

Head-mounted larva expressing bPAC in motor neurons, 4 dpf. Light stimulation begins after 10s, as indicated.

Click here to view.^{(94K, mp4)}

Strong tail beating as evoked by MN stimulation via bPAC or ChR2 in immobilized larvae is expected to lead to an increased swimming speed in freely behaving animals. In contrast to the coiling behavior described earlier, burst swimming in older larvae is executed mainly by fast muscles upon coordinated input from MNs.50 We established a method to record swimming behavior and quantify swimming parameters like the body bending angle and speed of larval zebrafish over time. Individual larvae (4 dpf) were placed in an agarose arena and subjected to an illumination protocol as described in STAR Methods. Wild-type larvae exhibit a mild escape response upon illumination with bright blue light, as described previously.52 However, a significantly stronger response was observed in bPAC and ChR2 transgenic larvae (Tg[mnx1:Gal4]/Tg[UAS:bPAC-V2A-mCherry], Tg[mnx1:Gal4]/Tg[UAS:ChR2-P2A-dTomato]) (Video S7). This became evident as increases in bending angles (Figure 1D), swimming speed during the blue light stimulation phase (Figures 1E, 1F, and S1B–S1D), and the distance which larvae cover during light stimulation (Figure 1G). Particularly bPAC-expressing animals showed a significant increase in locomotion behavior as compared to non-transgenic wild-type animals. Yet, some bPAC transgenic larvae did not respond to the blue light stimulus (Figures 1F and 1G). Since we did not observe this for ChR2 transgenics (Figures 1F and 1G), we speculated that this might be due to dark activity of bPAC,39 and adaptation of cAMP pathways.

Video S7: Swimming is induced by photoactivated bPAC and enhanced in cpe and tac1 mutants, related to Figures 1 and 3:

4 dpf larvae in swimming arena, five 20 s videos of individuals of the following genotypes, in this sequence: wild type, animal expressing ChR2 in motor neurons, animal expressing bPAC in motor neurons, cpe^−/− animal expressing bPAC in motor neurons, tac1^−/− animal expressing bPAC in motor neurons. Light stimulation begins after 5 s into each video, as indicated.

Click here to view.^{(4.2M, mp4)}

Our findings in behavioral assays demonstrate that ChR2-induced depolarization of cholinergic motor neurons causes rapid behavioral changes (e.g., tail beating and swimming) which originate from enhanced activity of slow or fast muscles, respectively.50 bPAC stimulation evokes similar effects. These findings are in line with the hypothesis that an increased cAMP level in cholinergic cells might lead to increased rates of SV fusion and exocytosis of the neurotransmitter ACh.

Neuropeptide signaling genes are expressed in larval motor neurons

We found that optogenetic cAMP generation induced locomotion behavior. Similar observations were made in motor neurons of other organisms. These could be linked to neuromodulatory activity, specifically signaling via neuropeptides, which in C. elegans was found to regulate the filling state of cholinergic SVs.14 Previous studies revealed that cholinergic neurons in zebrafish express neuropeptides53 and larval motor neurons contain DCVs, as visualized by electron microscopy.19 However, it is not firmly established which roles these neuropeptides have. Given the cAMP effects on locomotion, we wondered if some of these neuropeptides might be involved in modulation of signal transmission at NMJs in zebrafish. Therefore, we analyzed whether key components of the neuropeptide machinery are expressed in motor neurons. Similar to other vertebrates, zebrafish neuropeptides are derived from longer precursor proteins, which undergo enzymatic post-translational processing in DCVs.21 Publicly available single-cell RNA sequencing data54 and RNA in situ hybridization showed that cpe, which is involved in biosynthesis of most (but not all) neuropeptides and peptide hormones,20^,55^,56 is expressed in spinal cord neurons and motor neurons of 24 hpf embryos and 4 dpf zebrafish larvae, respectively (Figures 2A and S2). By mass spectrometry analysis of whole brain lysates, 62 neuropeptides originating from 34 different peptide precursors were identified in zebrafish,57 and expression of several neuropeptides, likely in the central nervous system, was validated. Furthermore, expression profiling of zebrafish neurons at 4 dpf by single-cell RNA sequencing54 identified a particular cluster of mRNA profiles (cluster #143) containing prominent marker genes for motor neurons like mnx1 and isl1 (Figure S2). The same cell cluster expresses the neuropeptide precursor tachykinin 1 (tac1) at high level and with high significance in comparison to the entire cell atlas (Figure S2). Likewise, 24 hpf embryos exhibited strong expression of tac1 mRNA in the brain and in a region of the spinal cord anatomically known to contain motor neurons (Figure 2A). These data suggest that certain neuropeptides are expressed and processed into their active form in, and are likely also released from, motor neurons during developmental stages relevant for our analysis. Hence, in the following experiments, we focused on the candidate genes cpe and tac1.

Open in a separate window

Figure 2

Key components of the neuropeptide signaling pathway are expressed in motor neurons

(A) Bright-field images of in situ hybridization stainings for cpe and tac1 mRNAs, respectively. Scale bars, 100 μm.

(B) Schematic representation of cpe and tac1 gene loci illustrating the CRISPR-Cas9 knockout strategy. Asterisks indicate sgRNA target sites. TSS, transcriptional start site; ATG, start codon; TC, termination codon; PTC, premature termination codon.

(C) qPCR showing cpe mRNA levels relative to gapdh.

(D) qPCR showing tac1 mRNA levels relative to gapdh.

(E) Percentage of wild-type, heterozygous, and homozygous genotypes within a population of 4 dpf cpe or tac1 mutant larvae. In (C) and (D), means ± SEM are shown; Student’s t test. Statistical significance given as p < 0.001.

See also Figure S2.

Neuropeptides modulate signal transmission at zebrafish NMJs

We subjected cpe^−/− and tac1^−/− animals to a swimming assay before and during bPAC stimulation, as described earlier. Basal locomotion speed during swimming before blue light stimulation was similar between wild-type animals as well as bPAC transgenic wild-type and homozygous cpe^−/− or tac1^−/− siblings (Figures S3A and S3B; Video S7). When photostimulated, we found that homozygous cpe mutants swam significantly faster as compared to their wild-type and heterozygous siblings. We could not observe significant differences in swimming behavior of cpe^+/+ and cpe^+/− siblings under bPAC stimulation (Figures 3A, S3C, and S3D). A similar effect was evident in tac1 knockouts, though less pronounced (Figure 3B). These findings could indicate that neuropeptides have a negative contribution to neuromuscular signaling, and thus, in their absence, higher locomotion speed increases are observed. Alternatively, neuropeptide signaling may positively influence synaptic transmission, and the observed increase of the behavioral response could be due to postsynaptic homeostatic scaling, compensating for a presynaptic deficit. tac1 could encode one of the neuropeptides responsible for the observed phenotypes but is not solely responsible for the effects as observed in the cpe knockout animals.

Open in a separate window

Figure 3

Enhanced locomotion behavior during bPAC stimulation in cpe and tac1 knockouts

(A) cpe homozygous knockouts and wild-type siblings (4 dpf) in the Tg[mnx1:Gal4]/Tg[UAS:bPAC-V2A-mCherry] double transgenic background were subjected to a swimming assay and swimming speed measured during blue light stimulation as indicated. Values are normalized to the first 2.5 s before illumination. Two-way ANOVA with Tukey multiple comparisons of means.

(B) Swimming speed of tac1 homozygous knockouts and respective wild-type siblings (4 dpf) carrying the Tg[mnx1:Gal4]/Tg[UAS:bPAC-V2A-mCherry] transgenes before, during, and after bPAC activation. Values are normalized to the first 2.5 s before illumination. Two-way ANOVA with Tukey multiple comparisons of means. Data are displayed as means ± SEM. Statistical significance given as p < 0.001, p < 0.01, p < 0.05.

See also Figure S3.

Patch-clamp recording from muscle cells shows altered transmission in cpe^−/− and tac1^−/− mutants following bPAC photostimulation of MNs

To further characterize the mutants, and to explore the possible cause for their altered locomotion, we turned to electrophysiology. Homozygous mutant animals were chosen for further analysis after genotyping on day 2 of development. We recorded from the superficial (“slow”) muscles,50 which are innervated by the secondary motor neurons60 of dissected larvae at 4 dpf, in voltage-clamp mode, before, during, and after 30 s blue light stimulation of cholinergic neurons expressing bPAC (Figure 4A). During these experiments we found that a certain fraction of wild-type animals did not show any effect during light stimulation in patch-clamp recordings, even though expressing bPAC protein and responding to stimulation by a swimming response (judged by visual inspection; here we could not quantify this due to time constraints; note that we observed non-responders also in behavioral experiments; Figures 1F and 1G). This might be due to the damage of the innervating MNs during the preparation. Since the number of animals which can be evaluated by electrophysiology is very limited, we excluded the individuals described earlier from statistical analysis. We analyzed the rate of mEPSCs, which are evident as single current spikes of different amplitudes and are believed to reflect single SV fusion events.61 The basal rates of mEPSCs were smaller in wild-type than in bPAC-expressing animals (Figure 4B), indicating an effect of bPAC expression, likely due to the known dark activity of bPAC, leading to elevated basal cAMP levels.40 Also, cpe^−/− animals expressing bPAC showed a lower basal mEPSC frequency compared to wild-type and tac1^−/− animals, though not reaching significance. When analyzed over the entire period of the experiment, normalized mEPSC frequency showed a significant increase during the blue light exposure in all bPAC-expressing strains, but not in wild-type non-transgenic animals (Figures 4B and 4C). This increase was significantly smaller in cpe^−/− animals, indicating a role of neuropeptides in regulating the rate of SV release or, possibly, mobilization. This neuropeptide, however, seems not to be encoded by tac1, as there was no difference observed in these mutants.

Open in a separate window

Figure 4

Post-synaptic recording from the NMJ following bPAC photostimulation demonstrates altered signaling in cpe and tac1 knockout animals

(A) Representative recording of mEPSCs before, during, and after bPAC photostimulation (blue bar) in a wild-type animal expressing bPAC in cholinergic motor neurons.

(B) mEPSC frequencies were analyzed in different time bins, as indicated, in wild-type animals, as well as in bPAC-expressing animals, either wild-type siblings or cpe^−/− or tac1^−/− knockout animals.

(C) Normalized mEPSC rates analyzed in 5 s time bins, as indicated, in the respective animals as in (B). Blue shade indicates time of blue light exposure. Data are displayed as means ± SEM.

(D) Number of mEPSC events of a given amplitude interval, as indicated, observed over a 30 s interval before photostimulation (upper panel) or during the photostimulation (lower panel), for animals expressing bPAC, in wild-type, or cpe^−/− or tac1^−/− knockout animals.

(E) mEPSC amplitudes were analyzed for each strain in 30 s intervals as indicated, before, during, and after photostimulation.

(F) All mEPSC amplitudes across all intervals were analyzed for bPAC-expressing animals (wild type, or cpe^−/− or tac1^−/− knockouts, as indicated). In (B) and (D–F), median and 25/75 quartiles (thick and thin black lines), min to max are shown. Two-way ANOVA, Tukey test. In (B) and (E), red significance labels analyses using Fisher Test. Statistical significance given as p < 0.01, p < 0.05; n.s., non-significant.

Next, we analyzed the mEPSC amplitudes (Figures 4D–4F). Amplitudes exhibited a wide range between few dozen pA and up to 700 pA, in rare cases even >1 nA (Figure 4F), as observed previously.62^,63^,64 This is difficult to reconcile, assuming single SV fusion events are causing these high-amplitude mEPSCs. Remarkably high amplitudes could for example be due to different characteristics of release sites that are more or less precisely located opposite of clusters of postsynaptic nAChRs,60 which mediate the currents underlying mEPSCs; but also simultaneous release of multiple SVs from one site may have to be considered, as well as electrical coupling in the slow muscle ensemble.62 When we analyzed the mEPSC amplitudes before and during the photostimulation in the different genotypes, a similar distribution of frequencies and amplitudes became apparent (Figure 4D). Before stimulation, cpe^−/− and tac^−/− animals showed a trend to smaller mEPSC amplitudes, which was significant during photostimulation, while wild-type animals showed significantly more of the large mEPSC events (>80 pA) compared to cpe^−/− mutants. When analyzing mean amplitudes in 30 s intervals before, during, and after the light stimulus, there were significantly smaller amplitude events following the light pulse in wild-type and tac1^−/−, but not cpe^−/−, animals (Figure 4E). tac1^−/− animals, however, showed a significantly more uniform distribution of mEPSC amplitudes throughout, ranging from 20 to 100 pA (Figure 4F).

In sum, the cpe^−/− mutants, which are generally affected for neuropeptide signaling, showed fewer mEPSC events during photostimulation, more of the small and less of the large amplitude mEPSC events, while tac1^−/− mutants had generally more uniform mEPSCs of moderate size. These findings may indicate that several neuropeptides with different modulatory effects could be involved in NMJ signaling. One of them may be inhibitory and lead to the reduced amplitudes after photostimulation, while the other could be promoting larger amplitude events; this latter peptide may be tac1, as these larger amplitudes were clearly absent in tac1^−/− mutants.

cpe^−/− and tac^−/− mutants express more postsynaptic nAChRs, possibly to compensate for altered neuronal ACh signaling

In behavioral experiments, we observed an increased locomotion of cpe and tac1 mutants in response to bPAC photostimulation, while these mutants were defective for presynaptic release of neurotransmitter. This could be explained by the absence of certain inhibitory neuropeptides in these mutants. Alternatively, there may be some homeostatic mechanism effective in the postsynaptic compartment that compensates for the reduced presynaptic cholinergic signal. One possible scenario of how this could be achieved involves reorganization or increased expression of nAChRs on the surface of muscle cells. To probe this hypothesis, we stained nAChRs using fluorescently labeled α-bungarotoxin in 4 dpf larvae, i.e., in the same developmental stage the swimming assays were performed. α-bungarotoxin binds specifically to the α subunit of mature nAChRs; thus the resulting fluorescence signal reflects the amount and distribution of receptors on the muscle cell surface. Commonly, intense staining in large, connected clusters at the boundary regions between myotomes can be observed, representing NMJs of slow muscle (Figures 5A and 5B), as well as nAChRs in small discrete spots, likely representing NMJs innervating fast muscle cells.65^,66^,67 We found that in cpe and tac1 mutants the mean number of small (rectangle in Figure 5B) and large clusters (dashed rectangle in Figure 5B) were not significantly different compared to wild-type animals (Figures S4A and S4B). The size of the small clusters was significantly larger in cpe^−/− mutants, while the size of the large clusters was not affected by loss of cpe or tac1 (Figures S4C and S4D). However, the fluorescence intensity of the small clusters was significantly increased in the cpe and tac1 mutants (Figure 5C) and strongly increased in large clusters for the cpe knockout (Figure 5C). This might indicate that diminished cholinergic output from motor neurons, due to a loss of neuropeptides, is postsynaptically compensated by localizing more nAChRs to the muscle surface.

Open in a separate window

Figure 5

Abundance of nAChRs on muscle cells is altered in neuropeptide mutants

(A) Representative α-bungarotoxin staining at 4 dpf (upper panel). nAChR clusters in green, bPAC-expressing motor neurons in red. Diagram of skeletal muscle cells and clusters of nAChRs on the cell surface (lower panel). Large receptor clusters are assembled on somite boundaries; small receptor clusters are distributed in between.

(B) α-bungarotoxin staining at 4 dpf on wild-type, cpe^−/−, and tac1^−/− animals. The rectangle in the upper panel represents an area containing small receptor clusters on fast muscle; the dashed rectangle marks larger clusters at the end of slow muscle cells.

(C) Quantification of fluorescence intensity of small and large receptor clusters in the respective zebrafish strains. Median and 25/75 quartiles (thick and thin black lines), min to max are shown. Scale bars in (B): 100 μm. One-way ANOVA with Tukey multiple comparisons of means. Statistical significance given as p < 0.001; n.s., non-significant.

See also Figure S4.

Discussion

The function of the NMJ is to excite muscle ensembles in a coordinated manner, enabling movement of body parts, or for general locomotion. Different types of locomotion, i.e., slow, moderate swimming, or corrections of ongoing drifting movement, vs. rapid and vigorous swimming as part of an escape response, require different amounts of neurotransmitter release. Thus, motor neurons need to be able to secrete ACh in different quantities, which can be affected by different regulatory mechanisms, that are listed in the following. (1) Increasing depolarization causes increased action potential frequency, thus leading to fusion and mobilization of more SVs. (2) The level of SV mobilization and priming can affect overall transmitter release, as more SVs will fuse upon arrival of an action potential. (3) As observed in some species, the amount of released transmitter can also be mediated by the amount of ACh loaded into individual SVs. In the latter case, more release of ACh can be achieved from the same number of SVs, and this can be regulated in a very short time, using already existing SVs, and not requiring de novo SV formation at the synaptic endosome. In C. elegans, acute loading of SVs with ACh was observed in response to a rise in synaptic cAMP levels, using optogenetic stimulation,14 and this required the release of neuropeptides from cholinergic motor neurons, thus likely acting in an autocrine fashion to influence the loading of SVs with ACh. In a natural setting, it is conceivable that exogenous signals, or intrinsic states of the animal associated with neuromodulatory signals in the motor nervous system (e.g., alertness and escape responses), may cause the increase in cAMP levels, to quickly upregulate neurotransmitter release in addition to the increase in firing rate. Also, more ACh release could be mediated even if the number of SVs becomes limiting, e.g., during periods of prolonged high activity. Our findings indicate that a similar mechanism may be present in zebrafish.

Neuropeptide signaling in zebrafish embryonal motor neurons was thus far mainly linked to developmental aspects of these cells. By electron microscopy, motor neuron terminals were shown to contain DCVs.19 Genes required for neuropeptide biogenesis, like cpe, as well as neuropeptide-encoding genes, have been identified in single-cell RNA sequencing analyses of embryonal cells, clustering along with markers of motor neurons.54 We showed that two of these genes, cpe and tac1, are expressed in the spinal cord motor neurons, generated knockouts, and analyzed behavioral and electrophysiological phenotypes, at basal level, and in response to optogenetic stimulation of the MNs using bPAC.

Upon photostimulation, wild-type animals showed a robust behavioral response. In both mutants, tac1, and particularly, cpe, even higher swimming speed was observed, indicating not a reduction, but rather a gain-of-function of MNs. Alternatively, (postsynaptic) compensation could affect larger increases in swimming speed following bPAC-induced neurotransmitter release. CPE is a central component of the neuropeptide processing machinery in mice,55 and the observed lethality of juvenile cpe^−/− zebrafish, though we do not know the exact cause of death, emphasizes this previous finding. Nevertheless, there has been substantial debate about the precise function of CPE in the secretory pathway, including a role as sorting receptor.20^,28 In that context, our result that locomotion behavior of cpe wild-type and heterozygous siblings is very similar points to a rather enzymatic, not rate-limiting, role of CPE. This is consistent with recent findings showing that cpe heterozygous mice, with reduced cpe expression and enzymatic activity, do not have a behavioral phenotype and exhibit wild-type-like quantities of mature neuropeptides.68

Our electrophysiological recordings showed that bPAC activation induced an increased rate of mEPSCs. Even though the effect of bPAC stimulation on mEPSC frequency was pronounced, we did observe a large variation in mEPSC amplitudes in slow muscle cells, impeding conclusions as to whether the SV filling state may be altered (increased in response to cAMP signaling in the wild type, lower due to the absence of neuropeptides). Highly variable mEPSC amplitudes have been observed previously in slow muscle, downstream of secondary MNs.62^,63^,64 By bungarotoxin staining and subsequent histological analysis, we observed an increase of large nAChR clusters on slow muscles, in absence of CPE, but only minor effects as in electrophysiological measurements. This may exactly be the outcome of postsynaptic homeostatic upscaling: Lower ACh release due to less SV content is counterbalanced by more postsynaptic nAChRs, in sum leading to similar postsynaptic currents. In the mutants, higher frequency of release (upon bPAC stimulation), meets increased nAChR density, which may lead to increased swimming. The fact that we did not see higher amplitudes in response to bPAC stimulation despite increased behavior may be explained by the involvement of fast muscles, from which we could not measure.

One possible effect of neuropeptide signaling was indicated when we analyzed the mEPSCs in a time-dependent manner. Both wild-type and tac1 mutants showed smaller mEPSC amplitudes following bPAC photostimulation, which was not observed for cpe mutants. While we do not know if some adaptation or fatigue occurred that could explain these reduced amplitudes, the fact that they do not appear in a mutant (cpe) indicates that a neuropeptidergic regulatory mechanism may be involved. A recent study analyzing the prohormone processing in pancreatic β-cells of βCpeKO mice revealed that lack of CPE leads to a decrease of mature peptides, but also an unexpectedly high level of correctly processed (canonical) CPE targets and even upregulation of some incompletely processed proteoforms in cpe knockout cells.69 These findings suggest removal of basic residues by an alternative or compensatory mechanism possibly involving carboxypeptidase D (CPD) and an indirect impact of loss of CPE and its targets on some, otherwise unrelated, peptides. It is unclear how exactly the loss of CPE activity affects the neuropeptidome of zebrafish motor neurons; yet, it can be anticipated that it results in complex changes in the neuropeptide pathway and a reduction in levels of a substantial fraction of neuropeptides. Since the motor neurons express several neuropeptides, there may be a balance of excitatory and inhibitory species that is shifted toward the inhibitory side in the wild type, and not in place in the cpe mutant (i.e., no alteration in signaling occurs); in tac1 mutants, the putative inhibitory signaling is still present, suggesting that tac1 is not encoding this signal.

In sum, we showed that neuropeptidergic acute signaling occurs at the NMJ of larval zebrafish or that neuropeptides shape NMJ signaling such that acute stimulation via cAMP causes different postsynaptic effects and exaggerated behavior. Whether cAMP triggers neuropeptide release acutely will have to be determined in the future, possibly using fluorescent neuropeptide release reporters to image release events in vivo and in real time. However, such reporters have not yet been developed for zebrafish.

STARMethods

Acknowledgments

Notes

Published: August 6, 2024

Footnotes

References