Developmental Function of the PHR Protein RPM-1 Is Required for Learning in Caenorhabditis elegans

Abstract

The PAM/Highwire/RPM-1 (PHR) proteins are signaling hubs that function as important regulators of neural development. Loss of function in Caenorhabditis elegans rpm-1 and Drosophila Highwire results in failed axon termination, inappropriate axon targeting, and abnormal synapse formation. Despite broad expression in the nervous system and relatively dramatic defects in synapse formation and axon development, very mild abnormalities in behavior have been found in animals lacking PHR protein function. Therefore, we hypothesized that large defects in behavior might only be detected in scenarios in which evoked, prolonged circuit function is required, or in which behavioral plasticity occurs. Using quantitative approaches in C. elegans, we found that rpm-1 loss-of-function mutants have relatively mild abnormalities in exploratory locomotion, but have large defects in evoked responses to harsh touch and learning associated with tap habituation. We explored the nature of the severe habituation defects in rpm-1 mutants further. To address what part of the habituation circuit was impaired in rpm-1 mutants, we performed rescue analysis with promoters for different neurons. Our findings indicate that RPM-1 function in the mechanosensory neurons affects habituation. Transgenic expression of RPM-1 in adult animals failed to rescue habituation defects, consistent with developmental defects in rpm-1 mutants resulting in impaired habituation. Genetic analysis showed that other regulators of neuronal development that function in the rpm-1 pathway (including glo-4, fsn-1, and dlk-1) also affected habituation. Overall, our findings suggest that developmental defects in rpm-1 mutants manifest most prominently in behaviors that require protracted or plastic circuit function, such as learning.

Keywords: C. elegans; PHR protein; RPM-1; habituation; learning.

Figure 1

Loss of RPM-1 affects exploratory locomotion. (A) Multi-Worm Tracker was used to quantitatively analyze the path of 30 wild-type C. elegans during 5 min of exploratory behavior on an agar plate with no food (left); paths were segmented into forward and backward movement (right). The majority of locomotion is forward, with relatively infrequent backward movement. (B−C) Proportion of time animals spent moving forward (B) and backward (C) during the final minute of observation. Shown are wild-type animals (wt; gray), rpm-1 (lf) mutants (red), and rpm-1 mutants with a transgene that uses the native rpm-1 promoter to express RPM-1 (blue). (D−E) Average speed during the final minute of observation of forward (D) and backward (E) movement for wild-type animals (wt; gray), rpm-1 mutants (red), and rpm-1 mutants with a transgene that uses the native rpm-1 promoter to express RPM-1 (blue). Points represent the mean of 30 animals for a single independent trial. Lines represent the mean of the points for each genotype. For transgenic rescue (blue), each point represents an independently derived transgenic line. **P < 0.01; ***P < 0.001 for Student’s unpaired two-tailed t-test between indicated groups.

Figure 2

RPM-1 is required for the harsh touch response. (A) Schematic of C. elegans harsh touch response. (B) Harsh touch responses of wild-type animals (wt; gray), rpm-1 (lf) mutants (red), and rpm-1 mutants with a transgene that uses the native rpm-1 promoter to express RPM-1 (blue). Response was measured as the number of reverse body bends an animal makes following a harsh touch stimulus to the head. Points represent the mean response of 20 animals on different experimental days. Lines represent the mean of all points within each genotype. For transgenic rescue (blue), each point represents an independently derived transgenic line. ***P < 0.001 for Student’s unpaired two-tailed t-test between indicated groups.

Figure 3

Loss of RPM-1 disrupts habituation, a simple form of learning. (A) Schematic of the tap stimulus and habituation protocol (top), and the reversal response that follows a tap stimulus (bottom). (B−C) Reversal probability after each tap stimulus of wild-type animals (gray) and rpm-1 (lf) mutants (red). Connected points represent the mean ± SEM of the tap response for each stimulus (n = 3−6 experiments; each experiment consists of 50−100 animals). Smooth thick lines indicate the best-fit exponential curves. Habituation level was measured as the asymptote of the curve (plotted inside the cyan column labeled HL with ± SEM). (D) Habituation level (asymptote) of wild-type animals (gray) and rpm-1 mutants at different ages. rpm-1 mutants have a strong habituation defect at all ages tested. (E) Habituation level (asymptote) of wild-type animals (gray), rpm-1 mutants (red), four independent transgenic lines of rpm-1 mutants expressing transgenic co-injection markers (control; magenta), and five independent transgenic lines of rpm-1 mutants in which the native rpm-1 promoter is used to express RPM-1 (blue). All RPM-1−expressing lines were significantly lower than the median control line and 4/5 RPM-1 lines were significantly less than the minimum control line (asterisks). (F) Representative example of a complete tap habituation response profile for one of the transgenic lines summarized in panel E (median transgenic lines are shown). *P < 0.05; **P < 0.01; ***P < 0.001 for Student’s unpaired two-tailed t-test between indicated groups.

Figure 4

RPM-1 affects habituation to tap by functioning in the sensory neurons. (A) Summary of the neural circuit that mediates the tap withdrawal response, and promoters used to drive transgene expression in different parts of the circuit. (B) Habituation level of wild-type animals (gray), rpm-1 (lf) mutants (open red), rpm-1 mutants expressing transgenic coinjection markers (control; red), and five transgenic lines of rpm-1 mutants in which RPM-1 is expressed in different neurons of the tap habituation circuit (various colors). Only expression of RPM-1 in the sensory neurons significantly rescued the mutant defect compared with the paired control line (3/5 lines, asterisks; orange vs. red). (C−F) Representative examples of complete tap habituation response profiles for the genotypes summarized in (B). Connected points represent the mean ± SEM of the tap response for each stimulus (n = 4−6 experiments). Smooth thick lines indicate the best-fit exponential curves. Points in the cyan bar labeled HL indicate the habituation level (asymptote of the curve ± SEM). Expression in the sensory neurons partially rescued the rpm-1 mutant defect (C). (G) Habituation of wild-type animals (gray), rpm-1 mutants (open red), and five transgenic lines of rpm-1 mutants using the sensory neuron promoter to express GFP (green) or RPM-1 (orange). All transgenic lines expressing RPM-1 significantly rescued habituation defects compared with the median control line, and 3/5 RPM-1 lines were significantly rescued compared with the minimum control line (asterisks). Therefore, expression of RPM-1, and not the presence of the sensory neuron promoter in transgenic arrays, was responsible for rescuing habituation defects in rpm-1 mutants. (H) Representative examples of complete tap habituation response profiles for the genotypes in panel G. For all panels, the rpm-1 allele used was ju44 and asterisks indicate transgenic lines that were significantly lower than the indicated control using Student’s unpaired two-tailed t-tests (*P < 0.05; **P < 0.01; ***P < 0.001). Note that different, independently derived transgenic lines were used for B and G, thus collectively 6/10 sensory neuron lines showed significant rescue.

Figure 5

Adult-specific expression of RPM-1 does not rescue habituation defects in rpm-1 mutants. (A) Schematic of the protocol for induction of adult-specific expression and sustained developmental expression of RPM-1 using the hsp-16.2 heat shock promoter. (B) Habituation of wild-type animals (gray), rpm-1 (lf) mutants (red), and rpm-1 mutants expressing GFP (green) or RPM-1 (blue) using the heat shock promoter under the indicated expression conditions. Note that adult-specific expression of RPM-1 did not rescue the habituation defect in rpm-1 mutants, whereas sustained expression of RPM-1 partially rescued the habituation defect in two independently derived transgenic lines compared with the minimum control line (asterisks). (C−D) Representative examples of complete tap habituation response profiles for the genotypes and conditions in panel B. Connected points represent the mean ± SEM of the tap response for each stimulus (n = 4−6 experiments). Smooth thick lines indicate the best-fit exponential curves. Points in the cyan bar labeled HL indicate the habituation level (asymptote of the curve ± SEM). (E) Schematic of the protocol for induction of intense adult-specific expression using the hsp-16.2 heat shock promoter. (F) Habituation of wild-type animals (gray) and rpm-1 mutants expressing RPM-1::GFP using the heat shock promoter under the indicated expression conditions (various colors). Data are represented as in (C) and (D). (G) Images of the head (solid outline) of Phsp-16.2::RPM-1::GFP positive C. elegans 18 hr after heat shock at 33° for 4 hr (top panel) or not heat shocked (bottom panel). Four internal structures are dash-outlined (from left to right): anterior gut (incomplete shape), posterior pharynx (circle), nerve ring (comma-shaped), and anterior pharynx (extended circle). Only heat shocked animals had visible RPM-1::GFP expression in the nerve ring. Stronger RPM-1::GFP fluorescence was present in the gut and pharynx. Low level autofluorescence is present in gut without heat shock (bottom panel). Scale bar is 25 μm. ns = not significant, *P < 0.05, **P < 0.01 for Student’s unpaired two-tailed t-tests between indicated groups.

Figure 6

Molecules that mediate RPM-1 function in neuronal development affect habituation. (A) Known mediators of RPM-1 function in development; blue arrows represent activation and red bars represent inhibition. (B−D) Tap habituation response profiles for the indicated genotypes. Connected points are means ± SEM of tap response (n = 4−6 experiments). Points in cyan bar are habituation levels (HL; asymptotes of fitted curves ± SEM). Note that glo-4 (B) and fsn-1 (C) mutants caused a strong, but partial, habituation defect compared with wild-type and rpm-1 mutants. dlk-1 (lf) strongly suppressed the habituation defect caused by rpm-1 (lf), as indicated by habituation levels in rpm-1; dlk-1 double mutants compared with rpm-1 single mutants (D). *P < 0.05; **P < 0.01; ***P < 0.001 for Student’s unpaired two-tailed t-tests between indicated groups.