Glycogen supports glycolytic plasticity in neurons

Abstract

Glycogen is the largest energy reserve in the brain, but the specific role of glycogen in supporting neuronal energy metabolism in vivo is not well understood. We established a system in Caenorhabditis elegans to dynamically probe glycolytic states in single cells of living animals via the use of the glycolytic sensor HYlight and determined that neurons can dynamically regulate glycolysis in response to activity or transient hypoxia. We performed an RNAi screen and identified that PYGL-1, an ortholog of the human glycogen phosphorylase, is required in neurons for glycolytic plasticity. We determined that neurons employ at least two mechanisms of glycolytic plasticity: glycogen-dependent glycolytic plasticity (GDGP) and glycogen-independent glycolytic plasticity. We uncover that GDGP is employed under conditions of mitochondrial dysfunction, such as transient hypoxia or in mutants for mitochondrial function. We find that the loss of GDGP impairs glycolytic plasticity and is associated with defects in synaptic vesicle recycling during hypoxia. Together, our study reveals that, in vivo, neurons can directly use glycogen as a fuel source to sustain glycolytic plasticity and synaptic function.

Keywords: C. elegans; glycogen utilization; glycolytic biosensor; glycolytic plasticity; neuronal metabolism.

Fig. 1.

Identification of pygl-1/Glycogen phosphorylase as a regulator of glycolysis in the C. elegans hypodermis. (A) Schematic of the microfluidic device used in this study to induce transient hypoxia via nitrogen gas exposure (N₂) (–43). (B) Diagram depicting the role of PYGL-1, the C. elegans ortholog of glycogen phosphorylase, in linking glycolysis and glycogenolysis. (C) Ratiometric image of HYlight expressed in the hypodermis, showing the emission ratio after 488 nm and 405 nm excitation. The calibration bar indicates ratiometric signal intensity (488ex/405ex). (D–G) Hypodermal HYlight ratios in normoxia for WT worms fed empty vector (D), pfk-1.1 RNAi (E), or pygl-1 RNAi (F), and quantification (G). In (G), each dot is one worm (n = 11); **** is P < 0.0001, calculated using Brown–Forsythe and Welch’s ANOVA followed by Dunnett’s multiple comparisons test. For an explanation of the RNAi screen, please see SI Appendix, Fig. S2. (H–K) Ratiometric images of hypodermal HYlight in normoxia in WT (H), pfk-1.1(ola458) (I), pygl-1(tm5211) (J) mutant animals, and quantification (K) done as in (G) (n = 10, 11, 11). [Scale bar in (C and D) is 100 µm; scale of (E–J) corresponds to scale of (D).]

Fig. 2.

pygl-1/Glycogen phosphorylase sustains glycolysis in neurons during energy stress. (A) Schematic of a worm expressing HYlight in the nervous system using the panneuronal promoter rab-3p. The dashed box represents the imaged regions of the nerve ring (the head ganglia) analyzed in the rest of the figure. (B–D) HYlight ratio (488/405 nm excitation) in the nerve ring of WT (B), pfk-1.1(ola458) null (C), and pygl-1(tm5211) mutants (D) during normoxia. (B^′–D^′) As in (B–D), but after 3 min of transient hypoxia. (E) Quantification of HYlight ratios in the head ganglia under normoxia for WT, pfk-1.1(ola458), and pygl-1(tm5211) animals. Each dot represents one worm; ** and *** indicate P values ≤ 0.01 and 0.001 respectively, calculated using Brown-Forsythe and Welch’s ANOVA followed by Dunnett’s multiple comparisons test. Fifteen animals were analyzed for WT, 12 for pfk-1.1(ola458), and 13 for pygl-1(tm5211). (F) Plot of HYlight ratios over time in normoxia and hypoxia (hypoxia period shaded blue) for WT, pfk-1.1(ola458), pygl-1(ola587), pygl-1(tm5211), and panneuronal rescue of PYGL-1A in pygl-1(tm5211) mutants. Error bars represent the SE of the mean at each time point. Fifteen animals were analyzed for WT, 12 for pfk-1.1(ola458), 13 for pygl-1(tm5211), 11 for pygl-1(ola587), and 15 for the neuronal PYGL-1A rescue in pygl-1(tm5211). (G) Graph showing the change in HYlight ratios from normoxia (taken at −1 min) to hypoxia (3 min into hypoxia) for each genotype, using data from (F). Significance is indicated by *** for P-value ≤ 0.001 and ns for P > 0.05, calculated using paired t test. [Scale bar in (B), 10 μm, applies to panels (B–D^′).]

Fig. 3.

PYGL-1A/Glycogen phosphorylase acts cell-autonomously in neurons. (A) Schematic of the ASER sensory neuron located in the head of the worm, with dashed box around the cell body that was imaged in this study. (B) Plot of HYlight ratios over time for WT, pygl-1(tm5211) mutants, and AIY- or ASER-specific rescue of PYGL-1A in pygl-1(tm5211) mutants. The blue shaded area indicates the period of hypoxia, initiated after 1 min of normoxia. Error bars represent the SE of the mean at each time point. Twenty-three animals were analyzed for WT, 19 for pygl-1(tm5211), 8 for AIY-specific rescue, and 13 for ASER-specific rescue. Note that expressing PYGL-1A in AIY did not rescue FBP levels in ASER, further indicating that PYGL-1A is required cell-autonomously to support glycolytic responses during hypoxic stress. (C–E) Ratiometric images of HYlight in ASER of WT worms (C), pygl-1(tm5211) mutants (D), and ASER-specific PYGL-1A rescue in pygl-1(tm5211) mutants (E). (C^′–E^′) As in (C–E), but after 3 min of transient hypoxia. The calibration bars to the Right indicate ratiometric signal intensity. A range of 2.5 to 4.8 was used for WT, and a range of 0.9 to 4.8 was used for pygl-1(tm5211) mutants and ASER:PYGL-1A rescue to better visualize the observed differences. (F) Schematic of the AIY interneuron with a box outlining the specific synaptic region (Zone 2) used for HYlight measurements. (G) Plot of HYlight ratios over time for WT, pygl-1(tm5211) mutants, and AIY-specific rescue of PYGL-1A in pygl-1(tm5211) mutants at AIY Zone 2. The blue shaded area indicates the hypoxia period following 1 min of normoxia. Error bars represent the SE of the mean at each time point. Eleven animals were analyzed per genotype. (H–J) Ratiometric images of HYlight in AIY Zone 2 of WT animals (H), pygl-1(tm5211) mutants (I), and AIY-specific PYGL-1A rescue in pygl-1(tm5211) mutants (J). (H^′–J^′) As in (H–J), but after 3 min of transient hypoxia. The calibration bar to the Right indicates ratiometric signal intensity. [Scale bar in (C and H) is 5 µm and applies to (C^′–E^′ and H^′–J^′).]

Fig. 4.

pygl-1/Glycogen phosphorylase is required for glycolytic plasticity and synaptic vesicle recycling during mitochondrial impairment. (A) Schematic of the dual pump mixing setup and microfluidic device used to deliver desired salt concentrations to worms (Materials and Methods). (B) HYlight ratios in the ASER soma for WT and pygl-1(tm5211) mutants (n = 13). Animals were stimulated by decreasing the salt concentration from 50 mM (gray shaded region) to 0 mM NaCl. Error bars represent the SE of the mean at each time point. (C) As in (B), but for WT, isp-1(qm150), and isp-1(qm150); pygl-1(tm5211) double mutants (n = 13, 9, and 11 animals respectively). (D) Confocal micrographs showing a Zone 3 neurite during normoxia in WT (Above) and pygl-1(tm5211) mutants (Below). Punctate structures represent vesicle pools in the synaptic region, labeled with the synaptic vesicle-associated protein RAB-3 tagged with mCherry. (D^′) Same animal as in (D), but after 10 min of transient hypoxia. RAB-3 becomes diffusely localized during hypoxia in pygl-1(tm5211) mutants, consistent with a defect in synaptic vesicle recycling (61). (E) Quantification of synaptic enrichment for WT and pygl-1(tm5211) mutants under normoxia and transient hypoxia, as described previously (61) (Materials and Methods). Each dot represents one worm; *** indicates P < 0.001, calculated using the unpaired t test with Welch’s correction. Ten animals were analyzed per genotype. (F) Model summarizing the findings of this study. Neurons employ both glycogen-dependent glycolytic plasticity (GDGP) and glycogen-independent glycolytic plasticity (GIGP) to adapt to energetic demands. GDGP is preferentially utilized under conditions of mitochondrial dysfunction (e.g., transient hypoxia) and is essential for sustaining glycolytic flux and synaptic vesicle recycling. In contrast, GIGP allows neurons to increase glycolytic output in response to stimuli such as neuronal activity, even without glycogen metabolism, if mitochondrial function is intact. In the schematic, red dots represent RAB-3 associated with synaptic vesicles. Under dual impairment of mitochondrial and glycogen metabolism, GDGP cannot be deployed, and RAB-3 becomes diffusely localized in the cytoplasm, indicating disrupted synaptic vesicle recycling.