An interesting article today by Bruce Cohen on the use of “next-generation” molecular light generators can be found here. These inorganic atoms, which are typically doped with some other material, are designed to have desirable excitation properties. For example, their use minimizes the amount of deleterious photobleachingin the sample. On the other hand, they are often large, making them difficult to use biologically without altering the system under study. However, as their size decreases, these molecules will likely play a large role in neuroscience in the coming years.
Postsynaptic densities (PSDs) in the forebrain contain ~ 100 – 400 proteins, including receptors (NMDA; AMPA) scaffolding molecules (DLG4, DLG1; HOMER1; SHANK2), kinases, phosphatases, and cytoskeletal components. They are typically 20–30 nm thick and 300 nm long.
In order to build a model of the PSD using DLG4*, Chen et al (here) cultured hippocampal neurons, incubated primary antibodies to DLG4, conjugated the antibody to secondary 1.4-nm gold bodies, and silver enhanced the gold. They then visualized 8 dendritic spines free from ice damage with EM tomography with pixel sizes of 0.48 – 0.75 nm. Finally, they built a model of the PSD in excitatory (glutamate-based) synapses based on imaging each component.
For example, via the following an EM image of 120 nm thick section (A), they were able to create a 1.5 nm thick virtual image without so much overlap (B), and then render the vertical filaments in a 3d reconstruction (C):
The vertical filaments are typically 5 nm in diameter (4.9 ± 0.2 nm, n = 22) and 20 nm long (21 ± 3 nm, range 16–25 nm, n = 38). They are also uniformly spaced. On the basis of the immunolabeling, it appears that these filaments are DLG4 proteins, with their N-termini attached to the membrane.
After doing this for each component of the PSD, they propose the following model, where red is vertical filaments of DLG4, cyan is NMDA-like receptors, blue is AMPA-like receptors, purple is horizontal filaments associated with NMDA / AMPA – like receptors, and white cross-links the vertical filament meshwork, especially near NMDA receptors:
It’s possible that some of the vertical filaments are not DLG4 proteins but instead some other type. However, there are ~ 300 DLG4 proteins and ~ 30 NMDA receptors per PSD (see here) so it is certainly possible that there could be 1 vertical filament for each NMDA receptor with spares left over for AMPA receptors. Regardless, this is a useful mental map for the molecular structure of the PSD, with implications for the mechanisms of AMPA / NMDA receptor-dependent synaptic plasticity.
* DLG4 is AKA PSD-95, but I’m going with the wikipedia definition.
Chen X, et al. 2008 Organization of the core structure of the postsynaptic density. PNAS doi: 10.1073/pnas.0800897105 .
Chen X, et al. 2005 Mass of the postsynaptic density and enumeration of three key molecules. PNAS doi: 10.1073/pnas.0505359102
Tadpoles are a model system for neural development and have well-characterized retinal ganglion cells that guide axons out of the retina. This is promoted by gradients of the guidance cue netrin. There is a poorly named protein called “deleted in colectoral cancer” (DCC) that acts as the netrin-1 receptor.
Manitt et al used DCC primary antibodies, IgG secondary antibodies coupled to 1 nm gold particles, and then serial section TEM to visualize this system. The picture below shows that DCC is present on presynaptic vescicles (arrows in J + K), the surface of presynaptic membranes (L), and on the surface of axonal filopedia (M). It also indicates synapses via triangles, and has scale bars of 200 nm:
This is consistent with their model that DCC receptor proteins mediate netrin signaling in axon pathfinding and synaptogenesis. BDNF is another molecule that promotes both increased synaptic density and axon branching. But they have slightly different mechanisms, showing that different cues can lead to the subtly differences in neural circuitry that defines the developing brain.
Manitt et al, 2009. doi:10.1523/JNEUROSCI.0947-09.2009. Pubmed here.
The role of the guanine nucleotide exchange factor Tiam1 has long been known to stimulate the outgrowth of neurons (see here) by shifting the Rac / Rho actin polymerization equilibrium (see here) towards Rac. Now Fard et al. have used RNA interference in cell culture to show that Tiam1 is specifically associated with NGF/TrkA-dependent neurite elongation. Here is the mechanism they propose, involving tyrosine kinase receptors:
RNA interference is a useful technique for evaluating mechanisms. As opposed to pure gene knock-out studies, it can allow for the evaluation of a more dose-dependent relationship.
Shirazi Fard S, Kele J, Vilar M, Paratcha G, Ledda F (2010) Tiam1 as a Signaling Mediator of Nerve Growth Factor-Dependent Neurite Outgrowth. PLoS ONE 5(3): e9647. doi:10.1371/journal.pone.0009647
Witcher et al recently published a paper looking at the relationship between astroglia and synapse morphology of three epileptic patients and one control patient. Following surgical resection, these patients allowed for the analysis of some of their hippocampal brain tissue. It took the surgeons less than 2 minutes to resect the hippocampus after cutting the main blood supply, and around 5 mins to collect the samples into the cold and oxygenated artifical cerebrospinal fluid. The quick pace produces healthy, physiologically relevant tissue samples. Eventually these hippocampal samples were cut into 40-50 sections ~45 nm thick and visualized with transmission electron microscopy. These are some of the components that they detected:
- Dendritic spines. These were classified on the basis of their morphology into thin, mushroom, stubby, and giant. Also, the researchers were able to determine the amount of extracellular space around them, the appearance of their organelles, and the organization of their microtubules.
- Docked presynaptic vescicles. These are vescicles that have been primed to form “soluble N-ethylmaleimide-sensitive factor attached protein receptor” complexes, and thus are generally immobilized at the release site before exocytosis. Vescicles were counted as docked if the vescicular membrane was adjacent to the plasma membrane of the presynaptic axon at the “active zones” of the synapse that have high postsynaptic density.
- Neurotransmitter type of synapse. On the basis of morphology, the researchers could make educated guesses as to the neurotransmitter type of the synapse. For example, some asymmetric synapses with round vesicles were assumed to be excitatory and gluatamergic. Some symmetric synapses were assumed to be inhibitory and GABAergic.
One of their findings was that the number of docked synaptic vesicles correlated positively with the postsynaptic density of the synapse, r = 0.68 in the tissue with mild neurodegeneration. This suggests a coordination between pre- and post-synaptic composition that has been found in rats and assumed to be the case in humans. They also found more invasive astroglia processes as the severity of the neurodegeneration increased, indicating that reactive astrogliosis may be responsible for some of the epileptic symptoms of the patients.
Witcher MR, et al. 2009 Three-dimensional relationships between perisynaptic astroglia and human hippocampal synapses. Glia doi:10.1002/glia.20946 .
Nofal S, et al. 2007 Primed Vesicles Can Be Distinguished from Docked Vesicles by Analyzing Their Mobility. J Neuro doi:10.1523/JNEUROSCI.4714-06.2007.
Quantum (q) dots have a number of advantages over conventional organic fluorophores, and their application may prove fruitful to constructing input-output models for different neuron types. The dots can be as small as 5-8 nanometers, although their hydrophobic region has been reported to be at least 16 nm. Their small size allows researchers to combine them to individual protein molecules and thus visualize intracellular processes.
Pathak et al (2007) considered the actual binding potential of 605 nm q dots to the most common human antibody, immunoglobulin G, which has two light chains and two heavy chains linked by disulfide bonds–see a picture of the molecule here. It’s important to validate that the quantum dots not only bind to the antibody but that they bind in such a way that the antibody can still bind to its typical ligand and have normal biological function. Since the light chain is the part of the antibody that binds to other proteins, it needs to be oriented outward, and moreover the antibody molecule itself should not be cleaved by the q dot binding.
Of their two techniques, direct conjugation and biotin-streptavidin based conjugation, the latter yielded significantly more q dots bound correctly per antibody molecule. In this latter technique, the researchers coated their q dot with the protein streptavidin and added biotin groups to the immunoglobins, likely at some of the antibody’s primary amine groups (see biotinylation). The noncovalent interaction between streptavidin and biotin has one of the lowest known dissociation constants, ~ 10^-15 mol/L, and thus it leads to a strong interaction between the q dot and the antibody.
Ultimately, the highest ratio they were able to produce was 1.3 +/- 0.35 of antibodies bound per q dot with a 2:1 antibody to q dot molar ratio. Since not all of these bound antibodies will be functional, in part because some will have the light chain of the immoglobulin molecule blocked, that number represents an upper bound on functional antibodies. The researchers note that since there is currently no way to control the binding orientation of immoglobulin molecules, Brownian motion means there is no guaranteed way to ensure functionally bound antibodies. More on q dots and neuro to come–this will be an important tech in the years to come, no doubt.
Pathak S, et al. 2007 Characterization of the Functional Binding Properties of Antibody Conjugated Quantum Dots. doi:10.1021/nl062706i, pdf.
Neuronal processes (axons, dendrites) are highly autonomous from the soma. This allows for synaptic plasticity (i.e., growth or shut down of receptors), alterations in spine morphology, and specific types of navigation towards extracellular guidance cues. Holt and Bullock implicate three major examples of this capability in neurons:
1) Synapse Plasticity. mRNA localization can affect the neuron’s ability to respond to activations with structural changes. For example, consider transcription of the activity-regulated cytoskeletal associated (Arc) gene. In activated hippocampal neurons, Arc mRNA is sent to dendrites that contain recently activated synapses with NMDA receptors and is locally translated into protein, where it probably modulates spine morphogenesis (i.e., causes shape changes in dendrites). It has a critical role in long term but not short term memory, as suggested by selective knock-out mice studies, where researchers replaced the 3′ untranslated region of the gene with a nonlocalizing transcript sequence. There are other examples in which mRNA localization can be necessary for synaptogenesis, the most important form of synapse plasticity. For example, localizing the neuropeptide-encoding sensorin mRNA into synapses is probably necessary for synapse formation in mechanosensory neurons in Aplysia and Helix pomatia.
2) Directionally Responsive Axon Protrusion. In neurodevelopment, changes in growth and directional steering of axons is dependent upon extracellular cues. For example, in growth cones beta-actin mRNA is concentrated near regions of attractive stimulus gradients, indicative of how the cell can transduce extracellular gradients into intracellular asymmetry. Inhibiting local beta-actin mRNA translation blocks attractive cues, but not repulsive ones, turning towards the favored extracellular stimuli. Presumably there are similar mechanisms that target other aspects of axon guiding during neurodevelopment to weave the intricate neuronal networks that underly everything we think, do, and say.
3) Spatially Dependent Gene Expression. mRNA’s translated in particular regions of the neuron may be modified at some amino acid residues selectively depending upon which part of the cell they are in. When the protein travels back to the nucleus following translation, its pattern of phosphorylation can then signal which part of the cell the mRNA was translated in, which can change patterns of gene expression following the same transcription factor entering the nucleus. For example, mRNA encoding the transcription factor cAMP response element–binding protein (CREB), which promotes the survival of certain neurons, can be translated locally in axons in response to neurotrophic nerve growth factor. CREB is localized to the distal axons of neurons (as indicated by Boyden chamber axon isolation and fluorescent in situ hybridization), its mRNA is selectively translated in response to local innervations of nerve growth factor, and the phosphorylation patterns of CREB at sites other than serine-133 will affect its transcriptional effects. Thus the cell can tell whether the pCREB in the nucleus came from distal axons or from the soma, and alter gene expression accordingly.
Casadio A, et al. 2003 Distribution of sensorin immunoreactivity in the central nervous system of Helix pomatia: Functional aspects. 10.1002/jnr.1084.
Li L, et al. 2005 The neuroplasticity-associated arc gene is a direct transcriptional target of early growth response (egr) transcription factors. doi:10.1128/MCB.25.23.
Leung KL, et al. 2006 Asymmetrical bold beta-actin mRNA translation in growth cones mediates attractive turning to netrin-1. doi:10.1038/nn1775.
Cow LJ, et al. 2008 Intra-axonal translation and retrograde trafficking of CREB promotes neuronal survival. doi:10.1038/ncb1677.
Holt CE, Bullock SL. 2009 Subcellular mRNA localization in animal cells and why it matters. doi:10.1126/science.1176488
One assumption of cortical circuitry is that the strength of interactions between neurons and layers should be directly correlated to the amount of overlap between axons and dendrites of pyramidal cells. However, this assumption is cracking under evidence that the cell’s class (as determined by morphology, postsynaptic target, and protein complement) also helps to determine the strength of the synaptic connection between any two neurons. For example, Petreanu et al mapped the dendrite locations of channelrhodopsin-2 expressing pyramidal neurons in the neocortex to determine the axon-dendrite overlap between its various inputs. But when the researchers quantified the strength of inputs in a given column with identical laser powers, one layer of barrel cortex (L5B cells) had 62-fold less input from the posterior medial nucleus than another group (L5A cells), despite the fact that L5B dendrites had more overlap with posterior medial nucleus axons than L5A dendrites. So, the functional connectivity between neurons cannot simply be deduced from the structure and overlap of the axons and dendrites. The actual class and region of the two given neurons has predictive power for synaptic strength as well.
Brown SP, Hestrin S. 2009 Cell-type identity: a key to unlocking the function of neocortical circuits. Current Opinion in Neurobiology 19:415-421. doi:10.1016/j.conb.2009.07.011
Petreanu L, et al. 2009 The subcellular organization of neocortical excitatory connections. Nature 457:1142-1145. doi:10.1038/nature07709.
Chondroitin sulphate proteoglycans (CSPG) are chondroitin sulfate polysaccharides that attach to the hydroxyl groups of serine residues on proteins. The negative charges on the chondroitin sulfate are important for regulating the brain’s perineuronal nets, mostly prominently via the inhibition of the sprouting and growth of axons. One region where it has been shown to regulate plasticity is in the visual cortex, which normally has plasticity during an early critical period of development which is lost in adults. When Pizzorusso et al degraded the CSPG glycosaminoglycan (GAG) chains of live adult rats using chondroitinase-ABC, it shifted the ocular dominance towards the non-deprived eye, an indicator of plasticity. The authors suggest that removing the inhibitory CSPG GAG chains may have restored experience-dependent generation of synaptic connections and/or their rearrangement.
Now Gogolla et al have shown that a similar mechanism of plasticity exists in the amygdala for CS-US conditioned fear response memories. These memories can be extinguished via CS presentations with the US in juvenile (i.e., post-natal day 16) rats, but this is not normally possible in adult rats, which exhibit spontaneous recovery of the CS-US fear response despite extinction. In an attempt to convert the adult phenotype to the juvenile phenotype, the researchers injected chondroitinase ABC into the basolateral amygdala of 3-month old rats in order to destroy their extracellular perineuronal nets. After fear conditioning, these enzyme injected adult rats were identical to controls, as indicated by the levels of freezing in response to the CS. And after extinction training, both condroitinase ABC injected and control rats exhibited the same depression of freezing responses. But when the rats were retested at 7 or 28 days post-extinction, the adult condrotinase ABC injected rats did not exhibit spontaneous recovery of the fear response, thus restoring the juvenile phenotype, while the control rats did exhibit spontaneous recovery. Since juveniles rats also do not have well developed perineuronal nets, it appears that these extracellular nets prevent the unlearning of a conditioned fear response and thus impede plasticity in the amygdala.
The fact that perineuronal nets regulated by CSPG are responsible for mediating the development shift away from plasticity with age in both the visual cortex and the amygdala indicates that it may have a role for information storage changes in neuronal circuits in general. The synapse is the key circuit element of the human brain and the ability to manipulate its plasticity would be a major breakthrough towards neural engineering.
Pizzorusso T et al. 2002 Reactivation of ocular dominance plasticity in the adult visual cortex. Science 298:1248-1251. doi: 10.1126/science.1072699.
Gogolla N et al. 2009 Perineuronal nets protect fear memories from erasure. Science 325:1258 – 1261. doi: 10.1126/science.1174146.
Cool study by Je et al, using genetically engineered specific inhibition of eIF2 – alpha mediated protein synthesis to study the role of de novo protein synthesis in neurons. Late-phase long-term potentiation (L-LTP) in hippocampal slices was shown in 1988 to be dependent upon protein synthesis by general protein inhibition, but it has remained unclear whether the inhibition is due to changes in presynaptic CA3 neurons or postsynaptic CA1 neurons. With their fast-acting virally transduced system, they were able to show that specifically inhibiting protein synthesis in CA1 (i.e., postsynaptic) neurons reduces L-LTP following theta burst stimulation as compared to controls. But, crucially, similar treatment in the CA3 region did not inhibit LTP, strongly suggesting that L-LTP is dependent upon postsynaptic protein synthesis. From what I can tell, this is consistent with our understanding of postsynaptic changes and lends importance to the current search for tagging systems that tell the nucleus to transport newly synthesized proteins to specific dendrites.
Je HS, et al. 2009 Chemically inducible inactivation of protein synthesis in genetically targeted neurons. Journal of Neuroscience 29:6761-6766. doi:10.1523/JNEUROSCI.1280-09.2009.