A recent paper in Nature Neuroscience demonstrates the usefulness of the pH sensitivity of fluorescent proteins.
I have briefly mentioned the importance of pH when I discussed mKeima. Here I will describe the work from Richard Tsien’s lab which utilizes the effect of pH on FP in order to study synaptic activity.
One avenue of communication between synapse is vesicle exocytosis, i.e. the release of neurotransmitter-containing vesicles from the pre-synaptic cell. A surge of calcium ions in the postsynaptic cell follows vesicle exocytosis. Interestingly, when a vesicle is exocytosed, the lumen of the vesicle changes its pH from 5.5 to 7.4.
A pH sensitive GFP-based protein called pHluorin was developed as early as 1998. pHluorin was fused to a vesicle protein called VAMP2. Thus, a pH change from 5.5 to 7.5 increase the emission intensity (i.e.the protein becomes brighter when excited).
Although this is a good tool, the fact that it is GFP based limits the use of dual-color microscopy, particularly with other types of fluorescent sensors since most of them are GFP, CFP or YFP based, which makes spectral distinction difficult. Therefore, the authors set out to develop a RFP based pH sensor. After several rounds of mutagenesis and shuffling of mRFP and mStrawberry, they identified a bright pH sensor which they named pHTomato since its ex/em peaks (550nm/580nm) are similar to those of dTomato. Importantly, increasing the pH has increase the brightness of the protein, without affecting the ex/em peaks (unlike mKeima which I mentioned above). Furthermore, fusion to VAMP2 did not change its characteristics; the pH dependence is reversible (i.e. once pH is acidic again, the intensity decreases).
Since VAMP2 gives them a high background expression on the membrane, the authors decided to fuse pHTomato to a protein called synaptophysin (sy), which is more vesicle specific. sypHTomato was then compared to sypHluorin in the same neuronal cell. The authors show (by graph) that both proteins perform similarly. However, it would be nice to also show an image.
Once the authors establish that pHTomato is a good pH-sensor, they turn to the real exciting work of dual color sensing. As mentioned above, following exocytosis, there is a Ca2+ increase at the post-synapse. Electrical stimuli also cause a Ca2+ surge in the presynaptic cell. The authors then used a Ca2+ sensor called GCaMP3. Briefly, GCaMP3 is a GFP-based protein that was modified such that that the two part of the protein are fused to Calmodulin (CaM) and a peptide called M13. In the presence of Ca2+, CaM binds M13, thus bringing the two parts of the GFP together, which results in fluorescence (ex/em 489/509 – similar to EGFP).
Expressing both sensors in the same neurons allowed them to visualize the spike in Ca2+ (green spike) concurrent with an increase in red signal, that deteriorates slowly, upon stimuli to the neurons. I am not a neuroscientist, so I cannot evaluate their system regards to choice of cells/stimuli, but the fluorescence response following the different stimuli they give seems distinct and impressive. However, it is very difficult to see it in the snapshot images of the cells.
Expressing each sensor in a different cell allowed the authors to distinctly visualized pre and post synaptic cells at the same synapse. What they looked at are pre synaptic sypHTomato-positives cells, in contact with post-synaptic GCaMP3 positive cells. I think that the image is beautiful:
Figure 3: Dual-color imaging of synaptic connection by sypHTomato and GCaMP3. (partial figure)
Now that the technical issues have been address, it was time to ask some meaningful biological questions. The first question which they asked is whether the vesicle content at presynaptic termini (called boutons) of the same cell is the same in all synapses of the same cell, or are there differences based on the target, postsynaptic cell (obviously a pre-synaptic cell can create synapses with multiple post-synaptic cells).
They used stimuli to measure the volume of the readily-releasable vesicles, and ammonium chloride (NH4Cl), a strong base, to measure the total volume of all vesicles in each synapse. They found that synapses that were targeted to the same postsynaptic neuron had reduced variability in total and readily-releasable vesicle volume, compared to the variability of synapses targeted to different cells. This indicates that there is some reverse feedback from the post-synaptic cell to the pre-synapses.
It would have been interesting to see if pre-synaptic boutons from different cells also have less variability if they are targeted to the same post-synaptic cell. However, the authors did not measure that.
They then looked what happens when you stimulate the cells electrically. As ecpected, there is an immediate increase in sypHTomato signal (vesicle exocytosis) with an almost immediate increase in GCaMP3 signal (Ca2+surge) in the post synaptic cell. This Ca2+ surge begins at the synapse but propagates along the dendrites, indicating opening of voltage-gated Ca2+ channels with the propagation of the ation potential. They confirmed that the Ca2+ surge is due to the vesicle release by adding inhibitors for the neurotransmitter receptors on the post-synaptic cell, and detecting little or no Ca2+ increase (The inhibitors did not affect vesicle release). This approach, then, was capable to image neuronal activity at single synapse resolution.
But the authors went one step further, to get achieve an all-optical system.
The experiment described above was performed by electrically stimulating the cells. However, such methods are invasive and affect the entire cell. An alternative strategy to activate cells is by chemical application of agonists or antagonists. However, addition of drugs or neurotransmitters to the cell culture media will affect all cells in the media. Here comes optogenetics – a system that allows optical activation of a single cell, or part of it. I do not want to discuss optogenetics here; this should get a post or two of its own. The basic idea is to use genetically encoded light-responsive ion channels. Once you shine light (at a specific wavelength) on the part of the cell you want to activate, the ion channel opens and you get action potential originating from the part of the cell you shined on. So here comes the cool stuff:
The authors co-expressed channelrhodopsin2 (ChR2) with a pHluorin-tagged vesicular glutamate transporter (vGluTpH). ChR2 gets to the plasma membrane, whereas vGluTpH to the vesicles. When they shined blue light (to activate ChR2 and excite pHluorin) they get an increase in the green signal, which is abolished by a drug that inhibits action potential. Similarly, co-expression of VChR1 (another optogenetic tool) with stpHTomato and shining green light (to activate VChR1 and excite pHTomato) led to an increase in red emission, that was abolished by the drug.
Better yet, co-expressing ChR2 with sypHTomato showed an increase in red signal only when cells were illuminated by blue light (that activates ChR2) and not green light (Which doesn’t). This figure should have been a main figure. I don’t know why they put it in the supplementary.
Fig S6(b) Tests of ChR2 in combination with sypHTomato as proof-of-principle of all-optical yet independent monitoring and stimulation. Left, interrogation of sypHTomato with Green light (546 -566nm) excitation alone without activation of ChR2-driven vesicular turnover. Right, positive control with blue light (457-482 nm), showing robust ChR2-driven sypHTomato transient (right).
And there you have it – an all optical system to study vesicle release in neurons.
So what can we do with this system?
The authors state several exciting possibilities:
- Perform dual color experiments with other green sensors.
- Deciphering pre and post synaptic strength in different scenarios.
- Probing neuronal circuits: combining optogenetic photostimulation of two spectrally distinct channlerhodopsins with two spectrally distinct sensors (pH sensors in this case) will allows us to follow the neural pathway when we activate a distinct neuron with a specific color.
But I could think of other options. For instance, the authors did not discuss at all the growing use of caged molecules. In brief, caged molecules are biologically relevant molecules that are in an inactive state. These molecules can be activated by shining light of a specific waveband. Thus, if you have a caged neurotransmitter and you shine the light at a specific location, synapses can be activated only if the uncaged molecule is at high enough concentration (i.e. where you shined your light).
The authors kind of ignored it, but there seems to be a difference in the sypHTomato intensity, and signal decay, when comparing electrical and optical stimulation. This difference could be biologically meaningful and could be further explored.
All in all, I think it is a very nice paper, describing a new system to study neurons.
Yulong Li & Richard W Tsien (2012) pHTomato, a red, genetically encoded indicator that enables multiplex interrogation of synaptic activity. Nature Neuroscience 15, 1047–1053.
A good post on GCaMP at “brain Windows” blog.
Optogenetics resource center
Info on chemical synapses, from Wikipedia
Li Y, & Tsien RW (2012). pHTomato, a red, genetically encoded indicator that enables multiplex interrogation of synaptic activity. Nature neuroscience, 15 (7), 1047-53 PMID: 22634730