The PHR proteins function in a range of developmental events, including axon termination, axon guidance, and synapse formation. In C. elegans, RPM-1 functions cell autonomously to regulate synapse formation in the GABAergic motor neurons, and functions cell autonomously in the mechanosensory neurons to regulate both axon termination and synapse formation [13, 14, 25, 30]. We now show that RPM-1 also functions cell autonomously in the GABAergic motor neurons to regulate axon termination. Prior to our study, axon termination defects caused by rpm-1 (lf) might have been considered cell-specific defects associated with the mechanosensory neurons and, as such, of lesser interest. Our results here demonstrate that this is not the case, and strengthen the argument that a core function of the PHR proteins is to regulate axon termination.
Our observation that RPM-1 is concentrated in two distinct subcellular compartments, the mature axon tip and the presynaptic terminal, points to a possible explanation for why RPM-1 regulates both axon termination and synapse formation in an individual neuron. RPM-1 may differentially regulate specific local signals or the intensity of core signals at different subcellular locations. This idea is consistent with a prior study, which showed that RPM-1 negatively regulates signaling by UNC5 (UNC-5) and Robo (SAX-3) to control axon termination in the mechanosensory neurons . Thus, RPM-1 localized at the axon tip may regulate local signaling triggered by axon guidance cues. In contrast, RPM-1 signaling at the presynaptic terminal is unlikely to regulate signaling by guidance cues, and therefore presumably has a distinct role in synaptogenesis. It was previously proposed that RPM-1 might coordinate axon extension and termination with synapse formation [3, 16]. Our finding that RPM-1 is localized to distinct subcellular compartments within individual neurons provides cell biological evidence to support this model further, and raises the interesting possibility that different activating or inhibitory signals might converge on RPM-1 in distinct locations.
Because RPM-1 is concentrated at the mature axon tip, it is possible that RPM-1 acts to cap a growing axon and trigger termination of extension. Presumably, localization of RPM-1 at the mature axon cap continues to silence signaling (most likely by guidance cues, as discussed previously) to ensure that the axon termination site is maintained. Transgenic studies in flies have shown that Hiw is enriched at presynaptic boutons, and there is a presynaptic bouton at the tip of fly motor neurons . Thus, it is plausible that Hiw is concentrated at the mature axon tip, although this has not been explicitly examined. To our knowledge, it remains uncertain whether vertebrate PHR proteins are concentrated at the mature axon tip; however, work in mice has shown that Phr1 is localized throughout the axon and excluded from portions of the growth cone in actively growing axons that have not formed synapses . Collectively, these observations raise the interesting possibility that at some point during development, PHR proteins may concentrate in the growth cone to trigger formation of a mature axon cap that is no longer capable of extension. Further support for or against this model is likely to be obtained by addressing a number of questions. Does RPM-1 localize to the axon tip prior to or following termination of axon outgrowth? What is the temporal relationship between RPM-1 localized to the axon tip and RPM-1 localized to the presynaptic terminal? Finally, where does RPM-1 localize during active axon growth prior to termination?
rpm-1 regulates axon termination and axon extension
Our genetic analysis showed that rpm-1 (lf) mutants have defects in axon termination of the GABAergic motor neurons at the anterior tip, the posterior tip, and within the dorsal cord. rpm-1−/−; syd-2−/− double mutants had enhanced axon termination defects (evident by increased numbers of axons showing overgrowth), and also had enhanced defects in axon extension exclusively at the posterior tip of the dorsal cord (evident by increased numbers of neurons with axon undergrowth). Thus, the use of a sensitizing genetic background has allowed us to determine that rpm-1 functions primarily in axon termination and secondarily in axon extension in the motor neuron (VD13) that forms the posterior dorsal cord termination site. To our knowledge, this is the first evidence that, in a single cell, rpm-1 regulates both axon termination and extension of the same process. This provides further support for the model that RPM-1 is a general and key regulator of axon length.
Previous studies with fish cortical neurons and with murine motor neurons reported the surprising and differing result that Phr1 regulates microtubule disassembly and assembly, respectively [5, 38]. Initially, we assumed that this paradox was due to a difference in the type of neuron analyzed. While this is still a potential factor, our finding that rpm-1 regulates axon termination at the anterior and posterior tip of the dorsal cord, but regulates extension exclusively at the posterior, suggests the interesting possibility that the location of a neuron and its environment might also instruct how the PHR proteins regulate axon length. However, given the anatomical differences between different GABAergic motor neurons (for example, the VD1 neuron has unique axon anatomy), it remains possible that intrinsic differences in individual motor neurons dictate whether RPM-1 regulates axon extension or termination.
We have also found that synaptic activity, acting as a secondary player, functions coordinately with RPM-1 to regulate termination of the anterior tip of the dorsal cord. Thus, enhanced termination defects at the anterior of the dorsal cord in rpm-1−/−; syd-2−/− double mutants are likely to reflect enhancer effects associated with loss of synaptic transmission resulting from severely impaired synapse formation in these double mutants. By contrast, synaptic activity does not function coordinately with RPM-1 to regulate termination at the posterior tip of the dorsal cord. As a result, the enhanced defects in posterior termination of rpm-1−/−; syd-2−/− double mutants are likely to reflect loss of a signal other than synaptic transmission, possibly synaptic connectivity. Given that SYD-2 is localized to the active zone of presynaptic terminals and not axon tips [22, 33], it is unlikely that SYD-2 functions at the axon tip to regulate axon termination. Overall, our results demonstrate that axon termination is established coordinately by a core signal from RPM-1, and secondary signals (such as synaptic activity), which are dependent upon the location or the type of neuron in question.
Prior studies showed that two Wnts, LIN-44 and EGL-20, and the canonical β-catenin BAR-1 regulate axon termination of the GABAergic motor neurons at the posterior tip of the dorsal cord  and within the dorsal cord . Wnts acting in the posterior of C. elegans have also been shown to regulate synapse position in the cholinergic DA9 motor neuron , and axon polarization in the PLM mechanosensory neurons [41–43]. Given that both Wnt (lf) mutants, and rpm-1 (lf) mutants have axon termination defects at the posterior tip of the dorsal cord, it is plausible that Wnt and RPM-1 signaling function together to regulate axon termination in GABAergic motor neurons. Consistent with this, we have found that bar-1 functions in the same genetic pathway as rpm-1 to regulate axon termination in the PLM mechanosensory neurons, and synapse formation in the GABAergic motor neurons (Tulgren and Grill, unpublished observation). In the future, we hope to address the question of whether RPM-1 and Wnt signaling converge differentially on BAR-1, thereby providing multiple mechanisms for regulation of the BAR-1 β-catenin.
The role of DLK-1 in axon termination and axon extension varies with location
Previous studies established the role of RPM-1, and PHR proteins in general, as negative regulators of the MAP3K DLK-1 (called Wallenda in flies and Dlk in mammals) [5, 11, 30, 44]. We have found that axon termination defects caused by rpm-1 (lf) are suppressed by dlk-1 (lf) at the anterior, but not the posterior tip of the dorsal cord. However, it is notable that dlk-1 (lf), while unable to suppress enhanced overextension defects in rpm-1−/−; syd-2−/− double mutants, strongly suppressed undergrowth defects in rpm-1−/−; syd-2−/− double mutants. These results demonstrate that RPM-1 regulates axon extension at the posterior tip of the dorsal cord by inhibiting DLK-1, but RPM-1 does not function through DLK-1 to regulate axon termination in this location. Our findings are consistent with prior studies, which showed that RPM-1 and Hiw function only in part through DLK-1 signaling [11, 25, 34]. Further, our results suggest that the anatomical location of a neuron may dictate whether DLK-1 regulates axon termination or axon extension. We propose three possible explanations for the role that location plays in the variable contribution of DLK-1 to axon termination or extension. (1) Extracellular cues may differentially regulate RPM-1 effects on DLK-1. (2) The extracellular environment may shape the relative contribution of DLK-1 signaling or the activation of DLK-1 independent of RPM-1. (3) Intrinsic differences between GABAergic motor neurons in different anatomical locations may affect the contribution of DLK-1 signaling to axon termination and extension.
rpm-1 regulates both axon termination and synapse formation in GABAergic motor neurons
Our analysis indicated that the anterior and posterior dorsal cord termination defects in rpm-1 (lf) mutants probably reflect overextension of the VD1 and VD13 processes, respectively. In the interior of the dorsal cord, rpm-1 (lf) mutants have axon termination defects in the DD5 motor neuron. Previous work showed that synapse formation defects are also observed along the length of the dorsal cord in rpm-1 (lf) mutants and, thus, are occurring in DD5 [13, 30]. The presence of both axon termination defects and synapse formation defects in the DD5 neuron of rpm-1 (lf) mutants is consistent with our observation that RPM-1 is localized to the axon tip and the presynaptic terminal of DD5 (Figure 7B). Thus, RPM-1 regulates both axon termination and synapse formation in a single motor neuron, DD5. Similar logic suggests that the same situation exists in VD1 and VD13.