CSHL Imaging Neurons Meeting – Session I

23 03 2007

Wow! A very busy start to the conference on Imaging Neurons and Neural Activity at Cold Spring Harbor Labs. I arrived at 6:30pm Thursday. Since then, I have seen 30 talks with copious note-taking, seen too many posters, given 1 talk, seen Ohio State win, and had lots of informal science talk over some beers. Not a lot of sleep though! There is literally no time to refine my notes, but to be timely with my posts I will be posting raw notes that will be updated and refined over the course of the next week or two.

Session I – Novel photoactivation and tagging methods

Optical Probes

Louis J. DeFelice, Vanderbilt – Neuronal transporters for monoamines analyzed with fluorescent substrates and fluorescently labeled transporters.
The goal – take a small molecule orally and get spatially specific transmitter release. DeFelice demonstrated that MPP+, IDT307, and other similar drugs are taken up by specific monoamine (dopamine, norepinephrine) transporters. This causes depolarizing currents that may be able to induce neuronal excitation and synaptic release in specific monoamine neuron type. However, they still need to demonstrate release, improve specificity and reduce toxicity.

Don Arnold, USC – Using intrabodies generated by phage display to study subcellular trafficking of Kv4.2
Overexpression of proteins (ex. PSD-95:GFP) can cause artifacts in protein localization and function. To gain protein localization information without this potential confound, Arnold has developed the interbody method to label endogenous protein with GFPs. He uses phage display to iteratively screen for single chain Fv antibody fragments that specifically bind selected proteins (ex. T1 domain of the Kv4.2 voltage-gated potassium channel). He then genetically fuses the ScFv gene with GFP and transfects cells. In neurons, these interbodies show labeling that is much more punctate than exogenous Kv4.2-GFP transfection, and thus closer to the real channel distribution. To further enhance specificity, he ubiquitinated the intrabody. Unbound intrabody is rapidly degraded, but binding to the Kv4.2 prevents degradation for unknown reasons. He then finds that in excitatory neurons there is a decreasing gradation of staining intensity outward from the cell body. I suspect this may reflect production, diffusion and distribution of the intrabody rather than the actual channel itself, but Arnold points out that this pattern is not seen in inhibitory neurons.

Andrew Hires, UCSD (that’s me) – Measuring glutamate spillover and uptake with GluSnFRs
Genetically encoded sensors of glutamate concentration have yet to find quantitative applications in neurons due to poor response amplitude in physiological buffers or when expressed on the neuronal cell surface. GluSnFR is made by bracketing CFP and Citrine with a glutamate periplasmic binding protein and then tethering it to the cell surface by fusion to a truncated PDGF receptor. Truncation of 8AA of the N-terminus and 5AA of the C-terminus of the PBP increases the maximum ratio change from 7% to 44% in phyisiological buffers. Rational mutations of the binding pocket give optimized the affinity to 2.5uMkD.
We performed field stimulation of GluSnFR expressing hippocampal cultures on astrocytes. Glutamate release is calcium dependent. The uptake inhibitor TBOA enhances peak [glu] and duration. Single AP stimulations are resolved with spatial averaging. Multiple AP stimulations resolved at bouton level. Spillover from 1 action potential field stimulation peaks near 700nM at around 10ms.
Trains induce steady state spillover >500nM within 4 AP, sufficient to activate NR2B NMDARs. Spillover modulated by stimulation frequency 60% with active uptake, only 10% with uptake blocked. Thus, non-connected neighboring neurons or astros may detect firing rate from spillover glutamate and induce a measured amount of, homeostatic regulation, heterosynaptic LTD or vasoregulation. Differential synaptic independence based on firing patterns.

Optical Stimulation of Neurons

Karl Deisseroth, Stanford – Multimodal fast optical interrogation and control of neurons.
Karl crushed it again. He mostly focused on the results of his article coming out in Nature next week on in vivo optical control of neurons with channelrhodopsin-2 and halorhodopsin. Halorhodopsin is the same yellow light activated chloride channel that Karl’s former postdoc published online in PLoS One this week. We previously covered that technology story here, but Karl’s lab has pushed it further. He showed a pretty demonstration by dual expressing ChR2 and halorhodopsin in motor control neurons of the worm c. elegans. He made the worm expand or contract dependent on the color of light that was flashed. Using lentivirus delivery of ChR2 into the rat motor cortex, they drove whisker movement by piping blue light into the brain with an elegant fiberoptic mount. He also demonstrated a fiber-optic cannula for deep brain optical stimulation. In brain slice, there is the potential for three-color simultaneous calcium imaging with Fura-2, optical stimulation with ChR2 and optical silencing with halorhodopsin. He is interested in using optical stimulation of genetically targeted neurons in humans to treat psychiatric disease. The technology sounds great for potential treatment in humans, except for the lentiviral gene delivery part…

Stefan Herlitze, Case Western – Vertebrate rhodopsin and channelrhodopsin 2 for control of intracellular signaling and physiological response on ion channels, neurons and neuronal circuits
Herlitze expresses vertebrate rhodopsin in the vicinity of presynaptic Ca channels or postsynaptic GIRK potassium channels. Light activation of the rhodopsin induces a G-protein cascade that modulates the function of these channels. Continuous illumination reduces presynaptic Ca-flux while increasing the paired pulse facilitation. Brief 5ms light during 2AP stim also reduces first EPSC while dramatically enhancing the second. On the postsynaptic side, light induces hyperpolarization by modulation of GIRKs. In vivo chick embryo turning off light can synchronize independent rhythmic neuronal networks. Presynaptic optical stimulation with ChR2 does not directly trigger transmitter release. Simulation of motorneurons expressing ChR2 induces muscle contractions.

Dan Huber, Janelia Farm – Channelrhodopsin-2-assisted microstimulation of layer 2/3 barrel cortex neurons detected by freely moving mice

How many 2/3 pyramidal neurons need to be activated to evoke reliable behavior?

Target ChR2 – in utero electroporation of CAGGS-ChR2-GFP. Expresses specifically in 2/3 pyramidals. 530-2364 neurons (mouse dependent) expressing by serial immunohistochemistry. 327-1412 in the window area.

Characterize the light – 1ms of max 7mW/mm^2 max light. Stim with 1ms pulses.

Characterize response – 100% neurons respond up to 20Hz. 50% up to 50Hz, faster than natural spike trains. Very sharp threshold for spiking/light intensity. The threshold is significantly different between cells, can use to reduce total # of cells firing in graded way.

The task – Train rat in a two choice task to find the water. Train rat to perceive 5AP optical train at 20Hz and go right or left for water dependent on stimulus presence. 1AP 65%, 2AP 80%, 5AP 85% correct. Perfomace is enhanced at low AP if you systemically reduce # of APs in stim rather than randomly interleave. Reduce light intensity to 10% of max to stim small subset and still get 65% correct for 5AP.

Working on cleaner fiber optic stimulator (a la Deisseroth), more complex simulation/discrimination tasks with whiskers.

Questions – why barrel cortex? Can you train it to perceive stimuli in arbitrary cortical region?
A: Because we want to do whisker perception experiments in the future. Probably.
Does minimal stimulation activate neighbors through polysynaptic connections?
A: Field recording sees network activation, don’t know if its 2/3s or just downstream.

Richard Kramer, Berkeley – Teathered small-molecule photoswitchable ion channels.
2 parts
Shaker potassium channel blocker teathered to azobenzene. Azobenzene has trans-cis isomerization hinge motion upon 380nm illumination.  Can be driven back by 500nm light. Bound to mouth of constitutively active mutant of K channel.  Several second on/off switching of hyperpolarization with light.  GYGV->GYGQ mutant converts to non-selective ion-channel.  Light on this mutant induces spiking.  [ChR2 appears to have significant response rate and expression advantages over SPARKs, however, both require transfection of exogenous channels.  The second part of the talk focuses on light regulation of endogenous channels]

Photoswitchable Affinity Label (PAL)
Attach a reactive epoxide or acrylamide reactive group to MAQ.  Apply to channels, binds non-specifically, but near mouth due to MAQ affinity for pore.  Hits variety of K channels.  Light turns off firing in hippocampal neurons treated with PAL.  Collaboration with Bill Kristan to modulate firing of HNl HNr neurons in leech [transgenic techniques are undeveloped in the leech.]  Inject PAL into eye in vivo take retina out.  380nm light turns off inhibitory amacrine cells and induces ganglion cells to spike.

Two-Color, Bi-Directional Optical Voltage Control of Genetically-Targeted Neurons

8 03 2007

Looks like Ed Boyden may have another hit on his hands. He and Xue Han reported at the recent Cosyne 2007 meeting a new method of controlling neural circuits. You may remember his work on Channelrhodopsin-2, which gave us millisecond optical control of neuronal spiking. Now they have reported a complementary technique with millisecond optical control of neuronal silencing. From the abstract :

We report that cultured hippocampal pyramidal neurons expressing Halo-GFP under the CaMKII promoter experienced strong hyperpolarizations (> -20 mV) upon exposure to brief pulses of moderate-intensity yellow light (~565 nm). In the absence of light, Halo-expressing neurons were physiologically indistinguishable from wild-type neurons.

But wait! There’s more :

We also demonstrated that in individual neurons expressing both yellow-light driven Halo and the blue-light driven cation channel ChR2, neural inhibition and excitation could be efficiently and independently controlled at the millisecond timescale, by interleaving brief pulses of yellow and blue light.

Hey Ed! Leave some neuroengineering projects for the rest of us please. 🙂