That has changed with a new report from Barbara Calabrese and Shelly Halpain from the Scripps Research Institute in La Jolla, California. In a paper published this week in Neuron, Calabrese and Halpain show that MARCKS functions in the maintenance and remodeling of dendritic spines in cultured hippocampal neurons. Their work also reveals that protein kinase C-catalyzed phosphorylation of MARCKS leads to dendritic spine remodeling, giving the protein a starring role in synaptic plasticity, learning and memory.

MARCKS knockout mice die around birth, so Calabrese and Halpain turned to cultured neurons to investigate the effects of MARCKS knockdown or overexpression. When the scientists expressed RNAi for MARCKS in 3-week-old cultures of rat hippocampal neurons, they observed a reduction in the density, width, and length of dendritic spines. But overexpression of MARCKS also caused a reduction in spine number, with increased length. From this, the authors concluded that MARCKS dynamically regulates spine stability and morphology, and this regulation is finely tuned.

MARCKS is a multitalented protein that binds to membranes (via a myristoyl moiety) and F-actin (via the effector domain), serving as a bridge between the cell surface and the actin cytoskeleton. When the effector domain gets phosphorylated by protein kinase C, MARCKS falls off the membrane, and its binding to actin is disrupted. To dissect the role of each domain, the researchers overexpressed mutant MARCKS proteins and carefully observed spine morphology and actin dynamics. Mutations in the effector domain that abolished PKC phosphorylation sites, or changed them to aspartic acid to mimic phosphorylation, both resulted in decreases in spine number, but in somewhat different ways. The unphosphorylatable mutant caused an increase in spine length, similar to that seen with overexpression of wild-type MARCKS. The opposite mutation, pseudophosphorylated MARCKS, induced a morphology similar to the knockdown, with reduced width and length. The different effects of the mutants reflected their different subcellular distributions: wild-type and nonphosphorylatable MARCKS were mostly membrane-associated, while the pseudophosphorylated protein was mainly cytosolic.

A look at the cytoskeletal rearrangements in these cells showed that the pseudophosphorylated MARCKS enhanced actin clustering in the spine head and reduced head motility. Decreased motility is often seen after glutamate receptor activation, raising the possibility that MARCKS phosphorylation by PKC could be responsible for real-life regulation of spine plasticity. Indeed, when the researchers treated neurons with the PKC activator phorbol ester, they recapitulated the loss of spines achieved with the pseudophosphorylated mutant, and this loss was prevented by expression of the nonphosphorylatable MARCKS.

It came as a bit of a surprise that loss of spines elicited by MARCKS mutant proteins was not associated with synapse loss. The same number of synapses seemed to be redistributed onto the remaining spines, and the synapses were functionally equivalent to those in normal cells as indicated by unchanged postsynaptic excitatory currents. These results show that even in the midst of MARCKS-influenced spine changes, synaptic function is maintained.

Dendritic spines put the plastic in synaptic plasticity—they proliferate during long-term potentiation and retrench during long-term depression. The identification of MARCKS as a player in this process jibes with another recent report implicating MARCKS in learning and memory in mouse hippocampus (McNamara et al., 2005), and should lead the way to more understanding of its involvement in AD and other neuropathologies.—Pat McCaffrey.

Reference:
Calabrese B, Halpain S. Essential Role for the PKC Target MARCKS in Maintaining Dendritic Spine Morphology. Neuron. 2005 Oct 6;48:77-90. Abstract

Matus A. MARCKS for Maintenance in Dendritic Spines. Neuron. 2005 Oct 6;48:4-5. Abstract