This Week in The Journal

Teresa Esch

Arginine Methylation Helps Motor Neurons Survive

Zoltan Simandi, Krisztian Pajer, Katalin Karolyi, Tatiana Sieler, Lu-Lin Jiang, et al.

Methylation of protein arginine residues is a common post-translational modification that is mediated by a small family of protein arginine methyl transferases (PRMTs). PRMTs have important roles in cellular proliferation and differentiation, and many of their protein targets are involved in transcription, RNA processing, translation, or DNA repair. Dysregulation of these proteins contributes to some forms of cancer and neurodegenerative disease (Blanc and Richard, 2017 Mol Cell 65:8).

Arginine dimethylation (red) is higher in motor neurons (green) than in other spinal cord cells, in part as a result of PRMT8 activity. See Simandi et al. for details.

PRMT8 is unique among PRMTs in that its expression is restricted to the nervous system and it can associate with the plasma membrane. PRMT8 mediates asymmetric arginine dimethylation (ADMA) and is involved in neural fate specification and differentiation. Simandi et al. report that although overall levels of ADMA were lower in mature spinal cord than in proliferating embryonic stem cells, levels remained relatively high in spinal motor neurons, where expression of PRMT8 was enriched. Consistent with previous work, mice lacking PRMT8 showed motor deficits. These deficits worsened with age and were accompanied by loss of motor axons. At 12 months of age, the structure of neuromuscular junctions was abnormal in PRMT8-deficient mice, and both the number of motor units on muscle fibers and the number of motor axons in the ventral horn were reduced in mutants relative to controls. Arginine methylation was also reduced in the spinal cord of 12-month-old PRMT8-deficient mice, whereas levels of the aging-associated pigment lipofuscin and markers of DNA double-stranded breaks were increased. Some of the effects of PRMT8 deletion appeared to be mediated by the transcription factor CREB1. CREB1 levels were reduced in mutant spinal cord and expression of several CREB1 targets was altered. Notably, overexpressing CREB1 increased stress tolerance and survival in PRMT8-deficient motor neurons.

These results suggest that arginine methylation mediated by PRMT8 is important for the survival of motor neurons, partly because it facilitates DNA repair. Additional experiments suggested that arginine dimethylation mediated by PRMT8 or another PRMT helps make motor neurons resilient to oxidative and endoplasmic-reticulum stress. Promoting the function of PRMT8 might therefore enhance neuron survival in neurodegenerative diseases and might even slow the effects of aging.

Spaced Training Strengthens Value Learning

G. Elliott Wimmer, Jamie K. Li, Krzysztof J. Gorgolewski, and Russell A. Poldrack

(see pages 7649–7666)

Animals quickly learn the best places to find food, then regularly return to those locations. To investigate the neural mechanisms underlying this ability, researchers train laboratory animals to associate specific sensory stimuli with food rewards by repeatedly pairing cues and rewards over several trial sessions. Such studies have revealed much about how neutral stimuli acquire incentive value. Functional magnetic resonance imaging (fMRI) studies have suggested that similar mechanisms underlie value acquisition in humans. In humans, however, associations between neutral stimuli and rewards are typically learned in a single session. This might be problematic, because studies of category and motor learning have shown that training over several occasions (spaced training) improves memory and might involve different mechanisms than learning in a single session (massed training). Therefore, Wimmer et al. asked whether spaced training also enhances memory for learned value associations and whether the training protocol used influences brain activity patterns evoked during recall.

Participants learned to associate pictures of scenes with monetary gains or losses through multiple pairings presented within a single session or spaced across 2 weeks. The final session of spaced training occurred on the same day as massed training (involving different scenes), after which fMRI scans were acquired. Notably, activity patterns elicited by reward- and loss-associated stimuli were more distinct if values had been learned during spaced training than if they were learned during massed training. Clusters exhibiting better value discrimination after spaced training were located in dorsolateral and ventromedial prefrontal cortex, orbitofrontal cortex, and medial temporal lobe.

On the final training day, memory for reward- and loss-associated stimuli was high regardless of the training protocol. Notably, however, greater working memory capacity led to greater recall only for items learned during massed training. Memory for value was tested again after 3 weeks, and at that time, memory for values learned during spaced training was better than that for values learned during massed training.

These results indicate that value learning is enhanced by spaced training, likely because it facilitates separation of neural representations in the medial temporal lobe and frontal cortex. Given that learning outside the laboratory normally occurs over days, months, or years, studies on human reward learning should consider using spaced training to better mimic natural learning processes.