Brain Stains

Short of sticking electrodes directly into an organism's brain, scientists looking to image neural signaling in living systems have few options.

By | November 7, 2005


© Scott Camazine/Science Photo Library

Short of sticking electrodes directly into an organism's brain, scientists looking to image neural signaling in living systems have few options. Until recently, most researchers made do with voltage- or calcium-sensitive organic dyes.

The use of green fluorescent protein (GFP) by scientists in Roger Tsien's lab at the University of California, San Diego, in the early 1990s marked a turning point for the field. By fusing GFP to a gene whose product is sensitive to changes in neural activity – for instance, fluctuations in calcium ion concentration or membrane voltage – scientists could target the probes to specific populations of neurons and read these changes optically.

"This is such a wonderful technique to be able to use. It's such a wonderful fusion of traditional observational anatomy and electrophysiology and molecular reductionist understanding of how the brain works," says neuroscientist Alison Barth of Carnegie Mellon University in Pittsburgh. Scientists have now developed about a dozen such probes and are using them to study processes ranging from olfaction in flies to changes in neuronal activity in mouse-seizure models.


At present, only a few genetically encoded probes can be used to monitor neural activity in living systems via fluorescence microscopy. "The two classes of probes that have been demonstrated to work, and that report different and somewhat complementary aspects of neuronal physiology, are synapto-pHluorin and a variety of calcium probes," says synapto-pHluorin inventor Gero Miesenböck of Yale University.

Synapto-pHluorin, Miesenböck explains, tracks presynaptic transmission, while calcium indicators report calcium influx into the cell. "When a neuron is electrically active, it fires action potentials, or it receives synaptic input. This electrical activity is generally coupled to the influx of calcium ions into the cell interior. And the calcium probes [show] that influx," he says.

Synapto-pHluorin contains a pH-sensitive GFP variant that targets synaptic vesicles' acidic interior. When the vesicle fuses, its lumen is exposed to the cell exterior, which has a neutral pH. "That change in pH leads to a change in the protonation of GFP, which causes an increase in fluorescence of the GFP," says Dierk Reiff of the Max Planck Institute for Neurobiology in Martinsried, Germany.

Though these probes have most often been used in Drosophila because of that organism's genetic pliability, researchers are beginning to use them in higher eukaryotes, too. Recently, Peter Mombaerts and colleagues at Rockefeller University targeted synapto-pHluorin to sensory neurons in intact mouse olfactory bulb.1 Another transgenic probe called fosGFP, developed by Barth and colleagues at Carnegie Mellon University, has been used to study long-term changes in neural activity in living mice. (Synapto-pHluorin, in contrast, reports events on the second-to-minute time scale.)

But while synapto-pHluorin can be used to probe information transfer between cells, it can report only presynaptic events; post-synaptic signaling is best studied with calcium sensors.


The first genetically encoded calcium indicators, called camgaroos and cameleons, use the calcium-binding protein calmodulin and were developed in the Tsien lab. In camgaroo probes, a circularly permutated GFP is fused to a calcium-binding protein; the circular permutation destabilizes the protein so that when it binds calcium, the probe's fluorescence changes. In cameleon probes, a calcium-binding protein is flanked by two different GFPs; when calcium binds, the protein folds, causing the GFPs to exchange energy via fluorescence resonance energy transfer.

The main advantage of such protein-based sensors is that they can be selectively expressed in specific cells. Yet, these probes have a few shortcomings, notably low dynamic range and a tendency for calmodulin to interact with native cellular elements, creating background. Oliver Griesbeck at the Max-Planck Institute for Neurobiology, has developed an alternative sensor that uses troponin C (TnC) as the calcium-binding moiety.2 The TnC-based probe, says Reiff, "improves the time scale on which these indicators actually report changes in intracellular calcium. ... The rise and the decay of the signal gets much faster than the so-far published and used indicators."

Another advantage of protein-based sensors is the ability to target them to specific subcellular locations, says Ehud Isacoff of UC, Berkeley. Isacoff and coworkers have developed systems that target sensors to a specific calcium source of interest. This has two advantages, Isacoff explains: first, it improves signal-to-noise; second, it allows researchers to probe the individual contributions of different sources for the same messenger. "Calcium has a number of different ways of entering the cytoplasm ... and each of these contributes in different ways to signaling inside a neuron. It would be lovely to be able to detect them one by one," he says.

Isacoff and colleagues recently developed a cameleon derivative called synapcam,3 which localizes to the postsynaptic density, reporting calcium influx selectively through glutamate receptors. Whereas cameleon could not resolve calcium signals in the cytoplasm, Isacoff explains, with synapcam "the signal-to-background is so good, that you're able to detect the synaptic transmission during single presynaptic action potentials and to resolve this at the level of single synaptic connections," he says.


Though improvements have made them faster, calcium-based sensors and synapto-pHluorin are still relatively slow. "Neurons generally operate on time scales that are faster than what most of these optical probes can currently detect. I think that's a particularly pressing need, to get things that can report fast electrical activity," says Miesenböck.

<p>Selected Genetically Encoded Neural Probes</p>

The obvious choice: genetically encoded probes that can directly sense changes in membrane voltage potential. Thomas Knöpfel of the RIKEN Brain Science Institute, Japan, calls genetically encoded voltage sensors the "Holy Grail of neuronal circuit imaging." But such probes have been more difficult to develop than their calcium-sensing counterparts.

So far three prototypes have been developed – FlaSh, SPARC, and VSFP – all of which are autonomous probes, meaning they can be transferred as DNA into a cell without a cofactor. Each links GFP to an ion channel, effectively coupling voltage-dependent molecular motions to fluorescence output. "It's a way of linking changes in the membrane potential of a cell to fluorescent intensity," says Vincent Pieribone of the John B. Pierce Laboratory at Yale University.

Yet none of the voltage-sensitive fluorescent probes so far developed works well in intact mammalian neurons, because they are difficult to target to the membrane. "Most of the voltage sensor is actually present in the endoplasmic reticulum and the Golgi [apparatus], which doesn't see changes in voltage. So the problem becomes that you have this huge background fluorescence that's not sensitive to voltage," says Pieribone.

To get around this problem, Isacoff, Knöpfel, and Pieribone (whose groups invented FlaSh, SPARC, and VSFP, respectively) are collaborating with Lawrence Cohen of Yale and Thomas Hughes of Montana State University, Bozeman, who specialize in neuronal imaging and genetics, to develop better probes.

Karl Deisseroth and colleagues at Stanford University recently reported another technical advance. Deisseroth's group coupled a light-sensitive cation channel isolated from green algae to a GFP variant and expressed the construct in cultured rat neurons.4 The team then pulsed the cells with light to trigger voltage spikes that lasted for milliseconds.

Miesenböck and coworkers have similarly used optically gated ion channels to target and control specific sets of neurons in moving flies with millisecond resolution.5 Knöpfel says this approach allows researchers to "write" information into neural systems, whereas indicators only "read." "With conventional imaging, we can observe the system. But if it is just observed, we're not necessarily learning too much about it," he says. "We need to interact with the system, and that requires reading and writing."

Popular Now

  1. Major German Universities Cancel Elsevier Contracts
  2. Grass Routes
    Features Grass Routes

    Researchers are discovering a suite of new locations and functions of endocannabinoid receptors that play roles in sickness and in health.

  3. Studies Retracted After UCLA Investigation
  4. Trump Nominates Sam Clovis to Lead USDA Research