These effects of MARK may help shape the synapse as a whole. A second paper, briefly mentioned below, shows that PSD-95 can reach across the synapse to regulate presynaptic function, as well. That work, from Yasunori Hayashi and colleagues at MIT in Cambridge, Massachusetts, appears today in an early online release in Nature Neuroscience.

It has been MARK’s effects on tau that have most interested AD researchers (see review by Drewes, 2004). When MARK phosphorylates tau, tau comes off of microtubules, leaving them destabilized. At the same time, phosphorylation by MARK primes tau for hyperphosphorylation by additional kinases including GSK3 and Cdk5, and this in turn promotes aggregation of tau into neurofibrillary tangles. Early on, MARK showed up localized with neurofibrillary tangles in AD brain tissue (Chin et al., 2000), and in current work published this past Wednesday, Lu’s lab begins to lay out how PAR-1 itself is regulated (Wang et al., 2007).

Previous work from Lu’s lab showed that overexpression of the Drosophila MARK homolog PAR-1 leads to tau-dependent neurodegeneration in flies (see ARF related news story). This led to the speculation that increases in MARK activity could promote the formation of toxic tau species in Alzheimer disease. But PAR-1 overexpression caused a stronger neurodegeneration phenotype than tau itself, suggesting to the researchers that PAR-1 might have additional substrates. When they went looking for those, first author Yali Zhang found PAR-1 was enriched at the postsynaptic region of the neuromuscular junction synapses in Drosophila larvae. Either knockout or overexpression of PAR-1 caused defects in synapse formation—PAR-1 mutants had more synapses on smaller boutons, while PAR-1 overexpressers had fewer boutons and an immature synapse structure.

In PAR-1 overexpressing flies, Dlg/PSD-95 was no longer targeted to postsynaptic structures, but instead was dispersed throughout the muscle cells. Both in-vitro and in-vivo experiments showed that PAR-1 phosphorylated Dlg at a serine residue in a region known to be involved in synaptic targeting. By expressing either a nonphosphorylatable or a phosphomimetic mutant, the investigators demonstrated that the phosphorylation prevented the recruitment of Dlg to synapses in vivo.

Additional experiments supported the idea that Dlg is the main target by which PAR-1 overexpression disrupts synapse development. But loss of PAR-1 also caused synaptic troubles. Ultrastructural analysis revealed exaggerated growth of postsynaptic reticular structures. Consistent with these changes, either overexpression or underexpression of PAR-1 affected synaptic transmission, with PAR-1 excess (or Dlg knockout) reducing transmission and PAR-1 overactivity increasing it.

The results show a previously unrecognized role for MARK in the physiology of a neuron. Not only does the kinase regulate microtubule dynamics via tau phosphorylation, but it also influences the formation of synapses via PSD-95. This work was done in larval neuromuscular junctions, and it will be important to see if the same holds true in central synapses, and how MARK might affect mature synapses in higher organisms. An open question for researchers to tackle is whether the activation of MARK and phosphorylation of PSD-95 contribute to the early synapse loss observed in AD. Certainly, PSD-95 is showing up in AD-related research with increasing frequency. For one, PSD-95 is known to be important for anchoring AMPA receptors in the postsynaptic membrane, a process thought to be affected by the amyloid-β protein.

There are two sides to every synapse, and PAR-1/MARK overexpression affects presynaptic function in flies, too. This should not be too surprising, given that the structure and function of pre- and postsynaptic regions are tightly linked. That’s where Hayashi’s paper ties in. It shows that PSD-95 itself takes part in that coordination. In today’s online issue of Nature Neuroscience, first author Kensuke Futai and colleagues show that postsynaptic PSD-95 links up with the transmembrane protein neuroligin to reach across the synapse and modulate presynaptic neurotransmitter release. The activity of PSD-95 can thus alter presynaptic short-term plasticity, observed here in hippocampal neurons in slice cultures. This effect provides a mechanism for dynamic retrograde signaling across the synapse, a process that helps coordinate both the structure and function of the two-sided communication unit.

Both papers bring home the point that even modest changes in PSD-95 function, for example, under the influence of deregulated MARK, could spell big trouble for synapses.—Pat McCaffrey.

References:
Zhang Y, Guo H, Kwan H, Wang J, Kosek J, Bingwei L. PAR-1 kinase phosphorylates Dlg and regulates its postsynaptic targeting at the Drosophila neuromuscular junction. Neuron. 2007 Jan 18; 53:201-215. Abstract

Futai K, Kim MJ, Hashikawa T, Scheiffele P, Sheng M, Hayashi Y. Retrograde modulation of presynaptic release probability through signaling mediated by PSD-95-neuroligin. Nature Neuroscience. 2007 Jan 21; Advanced online publication. Abstract