Protein Underlies Brain’s Response To Activity

MEF2 Roles in Reshaping Synapse May Open Approach To Neurodegenerative and Psychiatric Disorders

The world would seem a chaotic place without our pattern-seeking brains. Yet in its early stages, the brain is an unruly mesh of neurons. The environment, and our actions in it, help to impose order on the welter of connections, coaxing synapses here, cutting back there, not just early on, but all through our lives. Experience helps shape the brain, but how that happens—how synapses are remodeled in response to activity—is one of neurobiology’s biggest mysteries. Though axons and dendrites can be easily spotted waxing and waning under the microscope, the molecular middlemen working inside the cell to shape the neuron’s sinewy processes have been much more elusive.

Photo by Graham Ramsay

Above from left, Brice Gaudillière, Azad Bonni, and Aryaman Shalizi found that the repressor form of MEF2 promotes synaptic development in the rat cerebellum.

Steve Flavell (below left), Tae-Kyung Kim, Michael Greenberg (not shown), and colleagues found that synaptic differentiation in the rat hippocampus was inhibited by the activator form of MEF2.

Photo by Graham Ramsay

Two independent teams of HMS researchers report that they have found a protein that either pares down or promotes a neuron’s synapses, depending on whether or not the neuron is being activated. Rather than work at the far reaches of the cell, in the axon or dendrite, the protein myocyte enhancer factor 2 (MEF2) resides in the nucleus, where it turns on and off genes that control dendritic remodeling. In fact, the researchers have identified some of MEF2’s targets. In addition, one of the teams has identified how MEF2 switches from one program to the other, that is, from dendrite-promoting to dendrite-pruning. The discoveries are reported in back-to-back papers in the Feb. 17 Science.

The uncovering of the MEF2 pathway and its genetic switch helps fill in a theoretical blank in neurobiology, but what excites the researchers are the potential implications for the clinic. “Changes in the morphology of synapses could turn out to be very important in a whole host of diseases including neurodegenerative as well as psychiatric disorders,” said Azad Bonni who, with Aryaman Shalizi, Brice Gaudillière, and colleagues, authored one of the papers. Steven Flavell and Michael Greenberg, who led the other team, believe that the MEF2 pathway could play a role in autism and other neurodevelopmental diseases.

Activity Up, Synapses Down
In some respects, MEF2 appears to be a garden variety transcription factor. Like many transcription factors, it works by either activating or actively repressing target genes. But the results of that active repression are not exactly predictable. One might expect, for example, that the repressor form of MEF2 would inhibit the formation of dendrites and synapses. Working on a group of neurons in the developing rat cerebellum, Shalizi, an HMS research fellow in pathology, and Gaudillière, an HST medical student, along with HMS associate professor of pathology Bonni and their colleagues, found the MEF2 repressor had the opposite effect—it promoted synaptic differentiation.

In a separate study, Flavell, a graduate student in neurology, Greenberg, an HMS professor of neurology at Children’s Hospital Boston, and their colleagues found the MEF2 activator inhibited the development of synapses in the rat hippocampus, an area of the brain associated with memory and learning. Flavell, as well as the Bonni team, found the activated, or synapse-whittling, form of MEF2 comes on in response to increased neuronal activity.

This apparent paradox, that MEF2 activation leads to the inhibition of synapse formation, makes sense in light of what is known about the nervous system. In memory and learning, as well as development, activity leads to a sculpting of synapses, as well as to their growth. What may be more surprising is the way activity causes MEF2 to switch from repressor to activator.

SUMO-wrestling Proteins
When neurons are activated, calcium flows into the cell, which, in turn, triggers calcineurin. Calcineurin’s main role in the life of a cell is to remove phosphate groups from proteins. Normally, phosphate groups alter the structure of a protein in such a way as to allow it to interact with other proteins, usually resulting in the activation of the protein. What Bonni and his colleagues found is that phosphorylation at a particular spot on MEF2 results in the addition of a small ubiquitin-like modifier (SUMO) that, in turn, transforms MEF2 into a repressor. By removing this phosphate group and the SUMO with it, and by allowing an acetyl group to attach to MEF2 instead, activity-dependent calcineurin turns MEF2 from repressor to activator.

Though sumoylation of transcription factors had been observed in neurons, this is the first time it has been observed to be regulated by activity, a discovery that could have significance outside neurobiology. “We think the activity-dependent sumoylation-to-acetylation switch is going to be important not just in the brain but in general for transcription factor biology,” said Bonni.

MEF2 was first identified in neurons in the 1990s at a time when activity-dependent neuronal transcription factors were few and far between. The most notable of these was cAMP response element–binding (CREB) protein. As it turns out, MEF2 sits next to CREB on many gene promoters. In 1999, Zixu Mao, then an HMS research fellow working with Greenberg, Bonni, and colleagues, showed that MEF2 promotes neuronal survival, but little else was known about the protein. They suspected it might, like CREB, play a role in regulating activity-dependent synaptic remodeling and set out to determine if that is the case. RNA interference would be critical in their quest.

Declawing Brain Cells
Shalizi, Gaudillière, Bonni, and their colleagues began by using RNAi to knock down MEF2 in a particular group of developing cerebellar cells, the granule neurons. Normally, granule neuron dendrites end in a distinctive claw-shaped tip. Cells transfected with the MEF2-silencing RNAi exhibited 60 percent fewer dendritic claws. The researchers transfected the neurons with an RNAi-resistant form of MEF2. The dendritic claws reappeared.

RNAi had knocked down both forms of MEF2—activator and repressor. The researchers needed to know which MEF2 version had rescued the claws. They also wanted to see which MEF2 was being triggered by neuronal activity.

Illustration adapted by Rachel Eastwood from original courtesy of Azad Bonni	Claw retraction. Granule neurons normally end in a distinctive claw-shaped tip (schematic and top photo). Cells in which MEF2 has been knocked down with RNA interference exhibited 60 percent fewer dendritic claws. Bonni and colleagues found that the loss of claws (bottom photo) is due to the loss of the repressor form of MEF2.

As it happens, Bonni had earlier identified an unusual MEF2 pattern. “I look at sequences a lot, it’s something I enjoy,” Bonni said. “I found a motif within MEF2 that looked like a SUMO motif.” Shalizi then found a phosphorylation site five amino acids away. The researchers knew that activity-induced calcium influx leads, through calcineurin, to dephosphorylation. To mimic the effect of activity, the researchers altered the nearby phosphorylation site in such a way that it could not be phosphorylated. MEF2 lost its SUMO and gained an acetyl group, suggesting activity was not triggering sumoylation.

Most sumoylated transcription factors act as repressors, and Bonni and his colleagues found that to be the case with MEF2. To determine whether the sumoylated MEF2 repressor was promoting or inhibiting claw formation, they expressed their non-phosphorylatable, non-sumoylated mutant in cerebellar slices. The number of dendritic claws formed was low. Next, they appended a SUMO group to the mutant MEF2. Sure enough, the number of claws rose. Shalizi, Gaudillière, Bonni, and colleagues concluded that the MEF2 repressor was promoting synapse formation.

Synapse Suppressor
Meanwhile, Flavell and his colleagues were working toward the flip-side realization, namely that the MEF2 activator was suppressing synapse formation. They also began their study by knocking down MEF2, in this case in cultured rat hippocampal cells. The cells formed many more synapses than normal. “I saw it right under the microscope the first time we did the experiment,” Flavell said. Like the Bonni group, they still did not know which form of MEF2 was responsible. To find out, they added a constitutively active RNAi-resistant form of MEF2 to the cultured cells. The number of synapses declined precipitously, suggesting the active form of MEF2 was paring down synapses.

Using a novel inducible system, Flavell and his colleagues found that activated MEF2 could even suppress synapses that had already formed. Still it was not clear whether MEF2 was being activated by neuronal activity. To find out, they blocked calcium influx into the cultured hippocampal cells and looked at the number of synapses. As expected, synapse number declined. Next, they knocked down MEF2 using RNAi. There was no change in synapse number, which is exactly what one would expect if the synapse—whittling MEF2 was only active under conditions of calcium signaling.

Taken together, the findings of the two groups might appear puzzling since they seem to say that MEF2 promotes synapse formation by re-pressing genes and suppresses synapse formation by activating genes. The puzzle resolves itself when one considers the possibility that the genes being turned on and off by MEF2 are acting to inhibit synapse formation. In fact, Flavell and his colleagues have identified two of MEF2’s targets, arc and SynGAP, and they seem to do just that. The arc protein plays a role in internalizing glutamate receptors, which occurs when dendrites are being disassembled. SynGAP works to turn off the synapse-promoting ras gene. Bonni and his colleagues have identified yet a third target, Nur77. There are bound to be others.

The identification of these targets, and more generally the opening up of the MEF2 pathway, could lead to new therapies for a host of diseases in which synapses either fail to form or run rampant. In fact, Greenberg is currently a member of a consortium that is trying to get at the molecular underpinnings of autism. “We think the MEF2 pathway may be central,” he said.

—Misia Landau

top