Nobel Winner's Curse

Bert Sakmann visited Duke last Friday to give a seminar on his current work (streaming video here).  By my count, he is the sixth Nobel prize winner I have seen speak.  Three of them (Roger Tsien, Eric Kandel, and Torsten Wiesel) gave overviews of their important works, appropriate to the large audiences they were speaking before.  The other three (Richard Axel, Bert Sakmann, and Susumu Tonegawa) spoke about the current research in their labs, and largely neglected what made them famous.  From their perspective, like a famous rock band, they must decide between playing the audience's favourite songs, or playing the new songs, which while potentially not as good, are what they care about now.

Bert Sakmann is most famous for co-inventing the patch-clamp technique, which allows one to record the actual voltage of individual neurons in the brain.  The basics of the technique is that you take a microscopically thin glass cylinder, called a pipette, and bring it so close to the cell membrane that the membrane adheres to the pipette.  Ideally, this attachment is so strong that it creates a high electrical resistance, or gigaOhm seal.  While you are attached to a cell like this, you can indirectly measure the voltages of the cell.  However, to truly measure the cell voltage, you need direct access to the fluid inside the cell, which can be obtained by opening a hole in the membrane.  This is traditionally achieved by literally sucking on a tube connected to the pipette.  The idea that the fundamental technique of electrophysiology involves puncturing cells by sucking on a straw still tickles me.

Since inventing the whole cell patch technique, Sakmann, like many other famous scientists, has turned his direction "upwards" to more systems oriented questions.  The specific system he is now studying is the barrel cortex.  The barrel cortex is the part of a rodent's brain responsible for decoding what its whiskers are touching.  Each whisker is has a small portion of the cortex dedicated to it called a "barrel."

People have traditionally thought of cortex as having an hierarchical structure, where information flows into one layer, then proceeds through the six layers of cortex until being output not another area of the brain. This theory has come from extensive work in the visual cortex of rats, cats, and other mammals. Sakmann's group  recorded (using patch clamp!) from all of the layers of barrel cortex while stimulating the whiskers, and found that information reaches all layers of the cortex simultaneously, in contrast to visual cortex.  Previously, people have postulated that perhaps cortical columns (small vertically organized units of cortex covering all six layers) were a general processing unit that had been repeated with variation in different parts of the cortex.  By showing that barrel cortex works differently, systems neuroscientists must now be more careful in how the think about information flow. (I should also say at this point that I am not a systems neuroscientist, and do not have a firm knowledge of cortical processing or how information flows through the layers of different types of cortex.  The novelty I present here is based on my interpretation of the seminar.)

In the other section of his talk, he showed recordings from layer five thick-tufted pyramidal cells in barrel cortex, and a part of the brainstem called POm.  My understanding of this part was more hazy, but I believe he showed there is a cortico-thalamic-cortico feedback loop that acts as a threshold detector for input.  His goal in identifying this feedback loop was to give insight into rat decision making, but to this graduate student studying barrel cortex in rats seems like an indirect way to study decision making.

Given he was a Nobel winner, I came into the talk with high expectations, which could not be met.  Sakmann spoke to a packed house, as all Nobel winner's do.  The faculty in attendance were extremely deferential, which was highly unusual.   To these expectations Sakmann presented interesting, but not groundbreaking work, and if anyone else were presenting it, the crowd would have been 1/10th the size. His speaking style was assured, but uninspiring.  As I mentioned at the start, Nobel winners have two choices: to present what they're famous for, or present what they care about now.  And given the magnitude of what they're famous for, everything else pales in comparison.  Maybe that is the Nobel winner's curse, to set such high standards that you cannot help but disappoint in the future.

Assembly and Stoichiometry of the AMPA Receptor and Transmembrane AMPA Receptor Regulatory Protein Complex (TARP):

Today in lab meeting I presented this recent paper from the Tomita lab regarding TARP binding to AMPA Receptors (AMPAR).  Tomita,  as a post-doc in the Bredt lab, was one of the first people to investigate TARP binding and function in depth.  TARPs, as the name suggests, are auxiliary subunits of AMPAR, and are known to modulate AMPAR diffusion and conductances.  They contain a PDZ-binding domain that can bind PSD-95, and thus indirectly link AMPAR to PSD-95.  Single particle tracking studies have shown this binding can regulate AMPAR mobility in the synpase (Bats and Choquet, 2007).  As for the conductances, recordings from AMPAR in oocytes show that AMPAR desensitization is slowed and reduced in the presence of TARPs (Priel et. al. 2005).

In this paper, the authors used fairly simply immunoblotting techniques to investigate how AMPAR form tetramers, how many TARPs can bind to AMPAR, and how TARP binding modulates AMPAR currents. Their basic method was to create AMPAR variants with different weights, so that when they formed tetramers from the different size monomers, the tetramers ran at different speeds.  Around half of the amino acids in AMPAR are in the extracellular n-terminal domain (NTD), making the weight difference significant.

Their first interesting result used AMPAR lacking their NTD, but weighted differently using GFP.  They found that when they mixed AMPARs lacking their NTD, they could form tetramers containing 1, 2, 3, or 4 of the GFP subunits, showing that the tetramers were not formed by dimer-of-dimers, but from monomers.  In contrast, when they mixed full-length AMPAR with AMPAR lacking their NTD, their gel yielded only three bands, showing that the full length AMPAR formed dimers before being dimerized again into tetramers.  From this they concluded that the NTD of AMPAR is important for the initial formation of dimers.

They next turned to finding how many TARPs can interact with one AMPAR.  By transfecting differing amounts of TARP cRNA, running their gels, and staining for GluR1, they were able to find five different size bands, relating binding of 0-4 TARPs.  When they used the minimal transfection situation, which allowed only one TARP to bind, they found that the binding of one TARP was enough to affect the channel conductance.

The above result, however, was in oocytes, and not actual neurons.  To investigate the TARP binding in neurons, they used cerebellar granule cells in a stargazer mutant line, which lacks the TARP stargazin.  By blotting for GluR2/3 in heterozygous mice, they found only two bands of protein, suggesting that AMPARs form associations with TARPs at a fixed stoichiometry.  This fixed stoichiometry suggests that the TARP levels are either minimal or saturated.  When they stained for stargazin, they found no unbound stargazin, and thus concluded that only one TARP binds to each AMPAR.

The most interesting aspect of this paper, to me, is how they were able to investigate what I consider a biophysical property - dimerization and stoichiometry - using simple biophysical techniques (and probably common ones in biophysics at that).  The finding that the n-terminus is important for the initial dimerization does not seem new to me, as it has previously been shown that the n-terminus alone can form dimers (Leuschner and Hoch, 1999), and I have read reviews that the LIVBP portion of the NTD is the initial site of dimerization (Greger and Ziff, 2007).  However, the finding that one TARP alone is enough to modulate AMPAR function is intriguing.  I am not sure that I believe their finding that one and only one TARP binds to AMPAR in vivo, given how indirect the evidence is, and that they mention disagreement with another paper in their discussion.


Bats, C., Groc, L., & Choquet, D. (2007) The Interaction between Stargazin and PSD-95 Regulates AMPA Receptor Surface Trafficking. Neuron 53, 719-734.

Greger, I. H., Ziff, E. B., & Penn, A. C. (2007) Molecular determinants of AMPA receptor subunit assembly. Trends in Neurosciences 30, 407-416.


Leuschner, W. D. & Hoch, W. (1999) Subtype-specific assembly of alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor subunits is mediated by their n-terminal domains. J Biol Chem 274, 16907-16916.

Priel, A., Kolleker, A., Ayalon, G., Gillor, M., Osten, P., & Stern-Bach, Y. (2005) Stargazin Reduces Desensitization and Slows Deactivation of the AMPA-Type Glutamate Receptors. J Neurosci 25, 2682-2686.

PIP3 controls synaptic function by maintaining AMPA receptor clustering at the postsynaptic membrane:

Welcome to Mike and Rohit's Blog O' Science!  We are grad students at Duke and noticed a dearth of good neuroscience blogs, and people kept telling Rohit he should start a blog, so here we are.  I'll be posting my thoughts on papers and current research in the synaptic plasticity field, while Rohit will keep us up to date on in vivo imaging.

The first paper I'd like to cover is a recent paper from the Esteban lab about PI3K signaling during LTP.  Of all the multitudinous signaling pathways involved in LTP, the PI3K pathway was only recently discovered (Sanna et al 2002; Man, et al 2003).  Those papers disagreed slightly on the role of PI3K - Sanna believed it was only necessary for maintenance, while Man thought it was necessary for induction - but both showed that inhibition of PI3K signaling impaired LTP.

This paper started by showing that inhibition of PI3K signaling via overexpressiong PH-Grp1 decreased the basal AMPAR EPSCs in CA1 neurons. PH domains bind to phosphoinositides, and PH-Grp1 specifically binds to PIP3.  Overexpression of PH-Grp1 would thus sequester PIP3, and stop it from functioning in the cell.  They confirmed their genetic result pharmacologically using LY294002, a PI3K inhibitor.  To show that the current decrease was synapse specific, they bath applied AMPA, and showed that the current was the same, and thus number of surface AMPAR was the same.  So far so good.

Next they tried to induce LTP in neurons overexpressing PH-Grp1, and were unable to do so.  Still cool.

To confirm that the decrease in AMPAR currents was due to a decrease in synaptic but not extrasynaptic AMPAR, they transfected neurons with GFP-GluR2 and PH-Grp1.  Here they found that without Grp1, GFP-GluR2 was expressed in both the spine and dendrite; with Grp1, however, GFP-GluR2 shifted more towards the spine.  I originally thought this contradicted other lab's findings, but in verifying this I found that GFP-GluR2 does not have a strong spine bias, while SEP-GluR2 (surface) is punctate in the spine, presumably due to GluR2-containing endosomes in the dendrite (Kopec et. al. 2006).  Still, this result contradicts their earlier finding that Grp1 decreased synaptic but not extrasynaptic AMPAR number.

Since they saw changes in the subcellular distribution of AMPAR when Grp was expressed, they looked at the expression pattern of PSD-95, an integral anchoring protein at the synapse.  They found that in contrast to GluR2, PSD-95 is punctate in control conditions, and loses its punctility after Grp expression.  This contradicts the previous figure, but is in line with the decrease of synaptic AMPAR currents.  To see if this PSD-95 expression pattern changed AMPAR mobility, they performed FRAP on SEP-GluR2, and found that the mobile fraction increased following Grp expression, presumably due to lack of anchoring.  I am usually quite skeptical of FRAP experiments due to the high variability I have encountered doing them myself, and the large disagreement between labs on time constants and mobile fractions.  The FRAP performed here particularly stands out for its bizarrely low mobile fraction and tau recovery (tau of 5 minutes compared to 1-2 minutes in Ashby et. al., 2006 and Makino et. al., 2009).

To try to reconcile the previous two figures, they performed immuno EM to see where in the PSD the AMPAR were located.  Here they found that the AMPAR redistributed from the PSD proper to the perisynaptic region.  While EM has amazing resolution, I am skeptical that a shift of 80 nm in a subset of proteins can be precisely observed.  This finding further does not really explain how the AMPAR are being anchored near the synapse, since PSD-95 binding to TARPs (and thereby, presumably to AMPAR) is intact .

In the discussion the authors mentioned that PIP3, the PI3K end product, is involved in cell polarity, and is present at the tips of branching neurites. They hypothesized that it was playing a similar role in the synapse, which is also a highly polarized compartment.

In the end I found this a frustrating, if interesting paper.  PI3K signaling is an important, if ignored, player in LTP.  However, the contradictory nature of their results is dissatisfying.  I have a hard time understanding how PIP3 simultaneously bring AMPAR into the spine, while reducing AMPAR currents and reducing PSD-95 puncta.  The most intriguing part of this paper, in fact, is the redistribution of PSD-95, which makes me believe future research in PIP3 signaling should focus there, instead of on AMPAR trafficking.

Ashby, M. C., Maier, S. R., Nishimune, A., & Henley, J. M. (2006) Lateral Diffusion Drives Constitutive Exchange of AMPA Receptors at Dendritic Spines and Is Regulated by Spine Morphology. J Neuroscience 26, 7046-7055.

Kopec, C. D., Li, B., Wei, W., Boehm, J., & Malinow, R. (2006) Glutamate Receptor Exocytosis and Spine Enlargement during Chemically Induced Long-Term Potentiation. J Neuroscience 26, 2000-2009.

Makino, H. & Malinow, R. (2009) AMPA Receptor Incorporation into Synapses during LTP: The Role of Lateral Movement and Exocytosis. Neuron 64, 381-390.

Man, H.-Y., Wang, Q., Lu, W.-Y., Ju, W., Ahmadian, G., Liu, L., D'Souza, S., Wong, T. P., Taghibiglou, C., Lu, J., et al. (2003) Activation of PI3-Kinase Is Required for AMPA Receptor Insertion during LTP of mEPSCs in Cultured Hippocampal Neurons. Neuron 38, 611-624.

Sanna, P. P., Cammalleri, M., Berton, F., Simpson, C., Lutjens, R., Bloom, F. E., & Francesconi, W. (2002) Phosphatidylinositol 3-Kinase Is Required for the Expression But Not for the Induction or the Maintenance of Long-Term Potentiation in the Hippocampal CA1 Region. J Neurosci 22, 3359-3365.