I ain't affected by the negative. It's the Reflection and we dealin with the positive. It's for the love and cause of that we got a lot to give. It's the Reflection kid. Never gotta look between the lines to get the messages. - Talib Kweli

The latest Erin Schuman paper came out two and a half weeks ago in Nature, and it emphasizes something I've been mulling over lately. In the field of synaptic plasticity I think there is this tendency to focus on things being increased. For instance, there are ~7000 PubMed results for "long-term potentiation" and ~1500 for "long-term depression". Also, local protein synthesis is considered a major driving process for long-term potentiation, but people haven't paid nearly so much attention to protein degradation. This paper plus another recent report in J. Neurosci. are turning attention toward the role of protein degradation in LTP.

Bingol and Schuman tagged one of the subunits of the proteasome (the cell's molecular garbage disposal) with GFP and watched it move around in response to stimulation. Depolarization (activity) led to more signal in dendritic spines. Proteins that need to be degraded are often tagged with a protein called ubiquitin. Ubiquitination is a signal to the proteasome that this protein should be destroyed. The authors found that ubiquitination also increases in response in the spine increased 10 minutes after stimulation. After the initial increase, the ubiquitin signal falls off on a time course that matches the increase in the proteasome, implying that activity drives ubiquitination of some synaptic proteins which are then degraded about 10 minutes later when the proteasome shows up. It's important to note that every activity-dependent change in this paper was also shown to be NMDA receptor-dependent. NMDA itself could drive ubiquitin reporter degradation, and using a perfusion system like that I've mentioned before, they were able to show that the degradation occurs local to the site of stimulation rather than being a cell-wide or dendrite-wide event.

They then do some tricky fluorescence experiments with tagged proteasome subunits that make me feel like I am really missing something. Paging Dan Dright. Can you make sense of this? My understanding of FRAP (fluorescence recovery after photobleaching) is that you quench the fluorescence in a certain area by over-exposing the fluorophore, and then you wait to see how long it takes to get more fluorescence in that area. This gives you an idea of how proteins are moving around in the cell. The authors found less FRAP in bleached dendritic spines after activity. I would simpllstically interpret this as less proteasomes moving into the spine following activity. Instead, they "thus conclude that the immobile fraction of spine proteasomes increases from 35% to 90% upon stimulation, indicating that the proteasome is actively sequestered in spines." I suppose one source of confusion is that they studied proteasome movement on the order of minutes while the FRAP time courses are on the order of seconds. Later, they report (using this and another technique) that the proteasome spine entry rate increases about 1.5-fold while the exit rate rate decrease 6-fold. I'll have to take their word for it until I get smarter. So in this model the proteasome is getting 'trapped' in the synapse by activity.

They found some evidence for how the 'trap' is carried out. A large proportion of the proteasomes in their hands were associated with the actin cytoskeleton, an important determinant of spine structure and function. Activity induced an increase in the actin-associated portion of proteasomes. Disrupting the actin cytoskeleton, reduced the proteasome staining in dendritic spines. It's interesting to me that the actin disruption didn't seem to affect CaMKII. This indicates that two major structural elements in the spine are dissociable (i.e. CaMKII doesn't depend on an intact actin skeleton to maintain its spot in the post-synaptic density).

Along with showing that activity regulates protein degradation, the authors also point out that it is a fairly new understanding of the workings of the proteasome to suggest that the proteasome is delivered to its targets instead of vice versa. Along with the recent buzz concerning P-bodies and RNA degradation, I think the negative side of biological processes may be beginning to catch up to the positive. Some may have simplistically mapped protein synthesis onto LTP and synaptic growth, but this isn't the case. LTP and LTD will eventually both be shown to rely on both protein synthesis and degradation. Analogously, I believe an intact memory will rely on both LTP and LTD within the same dendrite. If you do the yin without the yang you just end up paisley.