|This article is of interest to the following WikiProjects:|
- 1 Untitled
- 2 Took out overkill
- 3 Mechanisms
Yes, thanks Nrets, that is a more accurate description
OK, Synaptidude, I think all that stuff about alternative explanations is a bit of overkill for this kind of article. You definitely seem to have more of a stake in this than I do! I think that the review article I suggested simply showed that there had been some sort of controversy, not necesarily to push an agenda. As far as spillover is concerned, I am not completely convinced that in some cases, what appears to be a silent synapse may actually be NMDAR detecting glutamate spillover. This is not always the case and it will really depend on the specific anatomy and release probabilities of a given set of synapses. Rather than put that long controversy in the article, perhaps you can move it to the talk page. If you want, please add a ref. to a different review article that compliments the one I put in. Nrets 01:57, 3 August 2005 (UTC)
Took out overkill
Yes, you are right, it was overkill. I took it out, although I had fun writing it. I'll paste it below in case we want to use any of it sometime. I do have a problem with the Voronin review however. Despite being recent, is completely out of date. This field might not even be particularly appropriate for a wikipedia article because it is emerging and moving fast. Most of what Voronin has in the review has now been discredited. Besides, it's less of a review and more of position piece. It is he who is pushing an agenda in that paper, and unfortunately, one that has been soundly disproved by later experimental evidence. So it doesn't really represent where knowledge of the field is. The authors of that paper are like the Japanese soldiers you sometimes hear about who have spent 60 years on an island, unaware that the war is over. They are firing their gun but the battlefield has moved on and they have no bullets. Synaptidude 05:52, 3 August 2005 (UTC)
- Yikes, you must really not like those guys. I sense years of them heckling you at your SFN meeting posters, etc. Feel free to remove the link, BUT please add one which you consider more current or correct. I still think spillover hasn't been completely ruled out in many cases (even if it has in the hippocampus). Cheers. Nrets 12:00, 3 August 2005 (UTC)
- I just don't like bad science. Unfortunately, the field is so young, there really hasn't been a considered review written on it yet, just position papers. Also, I don't think spillover has been ruled out at all, even in the hippocampus. It's just not the mechanism of silent synapses there. It clearly happens in inhibitory synapses, and probably in excitatory ones too, just for some different purpose.Synaptidude 13:26, 3 August 2005 (UTC)
While this article is written with the up-to-date knowledge of the nature of silent synapses in mind, it would not be fair to suggest that there has not been some controversy about what constitutes a silent synapse. The cited review (Voronin, et al., 2004) covers four potential mechanisms that might underlie silent synapses:
- postsynaptic lack of AMPA receptors (deaf synapse)
- presynaptic changes in the diameter of the synaptic vesicle fusion pore (whispering)
- presynaptic very low probability of release (Pr; mute synapse)
- glutamate spillover
A lack of postsynaptic AMPA receptors is now the mechinsim widely accepted by scientists in the field, but the other three were not without their adherents.
Postsynaptic lack of AMPA receptors
This is the currently accepted phenotype of a silent synapse. The synapse does not transmit a signal because the postsynaptic side does not generate an ion current in response to glutamate binding. AMPA receptors in the silent synapse are not in the surface membrane, but instead are stored in recycling endosomes inside the cell. NMDA receptors are found in the surface postsynaptic membrane, but do not allow current flow (unless the postsynaptic cell is depolarized) because they are magnesium blocked. These synapses might more aptly be called "deaf" synapses, but the term "silent" is entrenched.
Whispering synaspes are proposed to be silent based on the combination of two biophysical mechanisms. First, it is a known fact that the affinity of NMDARs to glutamate is much higher than the affinity of AMPARs. Second, it is proposed that silent synapses have only a partial or slowed release of glutamate, compared to non-silent synapses. This diminshed release of the neurotransmitter glutamate is proposed to be caused by a hypothetical release process called "kiss and run", where the synaptic vesicle fuses with the presynaptic membrane, but opens only a small pore to the extracellular space, before being retrieved directly back into the presynaptic membrane. The idea is that this small pore only allows a "leak" of glutamate to encounter the postsynaptic cell, thus activating fewer postsynaptic receptors. This "leaking" of glutamate would result in a lower concentration of glutamate in the synaptic cleft than would a normal synapse's "full release" event. The Whispering Synapse Hypothesis holds that this reduced glutamate concentration interacts effectively only with the NMDA receptors, because of their lower affinity.
Why silent synapses don't whisper
The primary scientific paper supporting the idea of a whispering synapse is Choi et al. (2000). However, this paper suffers from several subtle but fundamental flaws that render it's conclusions moot. Some of these flaws are:
- The "feasibility calculation" in the paper that shows that the concentrations of synaptic cleft glutamate in a "kiss and run" or a "full fusion" mode are feasible, suffers from a crushing error. The calculation assumes that only one glutamate molecule binds to a receptor to cause it to open. It is long established that two glutamate molecules must bind to a receptor before it will open. This enters a 2-power error in their calculated result. The predicted concentrations that one gets when one calculates with a corrected equation are outside of biologically possible limits.
- Measurements of the slope of the rising phase of the EPSC that are critical for the evaluation of the data in this paper are made improperly, so such that an increase in the apparent duration of the EPSC is mismeasured as an increase in slope. Slope is proportional to glutamate concentration in a non-saturated system, but duration is not.
- This paper suffers from the same fundamental procedural error that is decribed in more detail below in the "low Pr synapse" section, for the Gasparinin et al. paper.
Put together, these flaws mean that the data supporting whispering synapses is lacking. Of course, absence of evidence is not evidence of absence. A later paper, Montgomery et al., 2002, showed that in fact, the concentration of glutamate in the cleft did not change before and after awakening of silent synapses, thus disproving the whispering model.
Mute synapses are the term sometimes used to describe synapses that are physically present and intact, but have a very low presynaptic probability of release (Pr). Hypothetically, a low Pr synapse would appear silent because it rarely releases glutamate in response to a presynaptic action potential. Under this hypothesis, these synapses would become non-silent by becoming High Pr synapses. Experimentally, the hypothesis of Low Pr synapses is supported by the paper Gasparini, et al., 2000.
Why silent synapses are not "mute"
The experiments decribed in the Gasparini et al. paper suffers from a subtle but fundamental procedeedural error that renders their interpretation moot. To record synaptic potentials, or lack thereof, from silent synapses, Gasparinin et al. use a technique called "minimal stimulation". Minimal stimulation is a valid technique, if it is performed properly, and is in fact, the method used to first identify silent synapses in the hippocampus (their very first sighting was in Goldfish Mauthner cells). Minimal stimulation is the practice of recording from a single postsynaptic neuron using whole cell recording while stimulating a large population of axons with an extracellular stimulating electrode. The idea is to turn the stimulus down while providing test shocks through the stimulating electrode. In theory, as the stimulus intensity is decreased, the number of axons stimulated goes down. Eventually, one could turn the stimulus down until only one axon is stimulated. In minimal stimulation, one continues turning down the stimulus until it is at a level JUST BELOW where NO MORE AMPAR responses are recorded over many trials. In this situation, you would then be in one of two situations: 1) You have turned the stimulus down so low that you are no longer stimulating any axons (hence the lack of response) or, 2) You have turned the stimulus down so low that you are only stimulating a single axon, but that axon is connected to your postsynaptic cell via a silent synapse.
Experimentally, you tell the difference between these two conditions by depolarizing the postsynaptic cell and then observing if there is an NMDAR-mediated response. If there is, you have a silent synapse, if not, then you don't.
The experimental error committed by Gasparini et al. (and Choi et al. above) was as follows:
Once they had turned the stimulus down to just below where they lost their last AMPA (i.e. non-silent synapse) response, they then turned it back up to the level where they saw AMPA responses. This is, by definintion, a complete violation of the principle of the method. If you have turned the stimulus down to just below the level where normal AMPAR synapses are transmitting, and then you turn it back up, then the synapses you are studying are normal synapses, not silent synapses.
Thus while there are certainly low-Pr synapses in the hippocampus, they are distinct from silent synapses. The Gasparini method was (presumably) inadvertantly designed to record from normal synapses from a single axon rather than silent synapses. Indeed their method in incapable of isolating silent synapses. So while some subpopulation of their "normal synapses" might be low Pr, they are not silent.
Again, absence of evidence is not evidence of absence. Montgomery et al, 2002 showed that silent synapses were not low Pr, by artificially raising Pr but showing no response from silent synapse. Montgomery et al. did not use the minimal stimulation spproach, but rather recorded from individual pre and postsynaptic cells simultaneously. This is a far superior method because there can be no mistaking that a single axon is being activated (and the same one every time), and there is no way to mistake an active and a silent synapse.
One hypothesis that was put forward to explain silent synapses is that they are perfettly normal synapses, containing both AMPA and NMDA receptors, but in some cases their NMDA receptors would detect glutamate released from neighboring synapses that the AMPA receptors could not. Under this idea, local release of glutamate, from the corresponding presynaptic terminal, would activate both receptor types, while "spillover" of glutamate from neighboring synapses would arrive at the local synapse in too low a concentration to activate the higher affinity AMPA receptors.
Why silent synapses are not spillover detectors
When presented with a presynaptic action potential, a presynaptic terminal will oftentimes fail to release glutamate. The rate of these failures derives from the release probability of the synapse. A synapse with a Pr of 1 would release on every action potential, 0.5 on half of all action potentials, 0.25 on a quarter and so on (randomly destributed). Failures are detected by
Choi S, Klingauf J, Tsien RW. Postfusional regulation of cleft glutamate concentration during LTP at 'silent synapses'. Nat Neurosci. 2000 Apr;3(4):330-6.
Gasparini S, Saviane C, Voronin LL, Cherubini E. Silent synapses in the developing hippocampus: lack of functional AMPA receptors or low probability of glutamate release? Proc Natl Acad Sci U S A. 2000 Aug 15;97(17):9741-6.