Research Focus

Neurons in the brain communicate through chemical synapses. Depending on the activity of pre- and postsynaptic cells, these communication channels can rapidly and persistently change their strength. These functional adaptations, collectively known as long-term plasticity, involve a large number of intracellular signaling systems. On longer timescales, new synapses are established between previously unconnected cells while other synaptic connections are completely removed. Together, these changes in the efficacy and connectivity of brain circuits are thought to be crucial for information processing and memory storage in the brain.

We develop optogenetic methods to stimulate identified neurons and to optically measure the amplitude of postsynaptic calcium transients in dendritic spines. Two-photon laser scanning microscopy allows us to perform such optophysiological experiments in intact brain tissue with high spatial and temporal resolution. Using genetically encoded probes, we monitor the activity of single synapses over several hundred stimulations and measure parameters such as synaptic potency and the probability of glutamate release. Optical induction of plasticity at individual, identified synapses allows us to investigate the underlying electrical and biochemical processes in great detail. The connectivity of our brain constantly changes in response to sensory experience (Huber et al., 2012). A central aim of our research is to understand the rules and molecular mechanisms that govern our extraordinary ability to learn and to remember.


  • Critical point

    Neuronal silencing with light-gated chloride channels

    Simon Wiegert, Iris Ohmert

    The discovery of Channelrhodopsin, a directly light-gated cation channel, had a tremendous impact on neuroscience. Expression of this channel in a defined population of neurons allows activating these neurons with millisecond precision (Schoenenberger et al., 2011). We adapted this technique for our work, inducing specific activity patterns to investigate the effects on synaptic strength and stability (Berndt et al., 2011; Wiegert and Oertner, 2013).Light-induced inhibition of neuronal activity would also be useful for loss-of-function experiments, but is technically more difficult to achieve. Light-driven ion pumps hyperpolarize neurons as long as they are illuminated, but they need very high light intensities and are not suitable for sustained inhibition. Collaborating with Peter Hegemann at the Humboldt University in Berlin, we developed and characterized the first directly light-gated anion channel that can keep neurons from spiking at very low light levels (Wietek et al. 2014, Wietek et al. 2015).

  • Neuron
    CA1 neuron (cyan) and Schaffer coll. boutons (red, with ChR2)

    Activity-dependent regulation of synaptic lifetime

    Simon Wiegert, Celine Dürst

    While intact synaptic plasticity seems to be essential for our ability to learn and to remember, it is less clear if memories are indeed stored as subtle changes in synaptic strength. Alternatively, the connectivity pattern in certain brain areas could change by removal of existing synapses and formation of novel connections. We know this form of ‘binary’ memory storage from our digital computers; it has the advantage that small changes in the strength of individual connections do not affect memory stability or recall. Using optogenetic stimulation and read-out, we could demonstrate that long-term depression indeed had consequences on synaptic lifetime: Synapses with low release probability were efficiently removed from the circuit several days after long-term depression, pointing to a strong correlation between synaptic strength and synaptic lifetime (Wiegert and Oertner, 2013). We have now accumulated evidence that the reverse is also the case: One-time potentiation of individual synapses increases their stability and survival during the following days. Our findings suggest that the connectivity of an adult brain is the direct consequence of a myriad of decisions made at individual synapses.

  •  synCaMPARI_dendrite
    synCaMPARI dendrite
    synCaMPARI spines

    New tools to label active synapses

    Alberto Perez-Alvarez, Brenna Fearey, Simon Wiegert, Christine Gee

    No microscope is fast enough to perform functional imaging of all hippocampal synapses in real time. The goal of this project is to persistently label all synapses that are active in a short time window so we can map their position in 3D by two-photon laser scanning microscopy. This has to work in living tissue so we can repeat the measurements to study the functional stability of synaptic networks. To do so, we have targeted the calcium-dependent photoswitchable protein CaMPARI (Fosque et al., Science 2015) to dendritic spines. A pulse of UV light converts synaptically anchored CaMPARI from green to red, but only in those spines that were previously active. We use the red-to-green ratio to detect all active spines independent of their size. Memory formation is thought to be based on changes in synaptic transmission. Our new tool will help us create three-dimensional maps of synaptic function before and after learning.

  • Optogenetic investigation of spike-timing dependent plasticity

    Christine Gee, Bas van Bommel

    Classic electrophysiological studies of synaptic plasticity consist of a brief induction protocol, followed by low frequency test pulses to assess the stability of synaptic strength over time. The discovery of spike-timing dependent plasticity revealed the possibility that in fact every single action potential changes synaptic strength in a miniscule way, and that synaptic strength tracks temporal correlations between pre- and postsynaptic spike patterns (Hao and Oertner, 2012). This hypothesis is impossible to test with purely electrophysiological approaches, as measuring subthreshold potentials with an electrode perturbs intracellular signaling and eventually kills the recorded neuron. The recent discovery of Channelrhodopsin variants that are sensitive to red light (Klapoetke et al., 2014) enables us to control the timing of action potentials in individual CA3 and CA1 pyramidal cells independently and non-invasively over several days. Using a combination of optogenetic tools with different spectral properties, we are investigating the sensitivity of individual synapses to subtle temporal correlations in pre- and postsynaptic activity patterns.

  • automatic spine detection
    Automatic spine detection compared to EM reconstruction

    The function of endoplasmic reticulum in homeostatic plasticity

    Alberto Perez-Alvarez, Clemens Blumer, Shuting Yin

    Neurons contain a tubular network known as the endoplasmic reticulum (ER) that stretches throughout the entire cell, including the axon and most of the dendrite. The ER is involved in the delivery of proteins and lipids as well as calcium homeostasis. In pyramidal cells, a small subset of dendritic spines contains stable ER in the form of a spine apparatus, while most spines are briefly sampled by single ER tubules from time to time. We could show that synapses on spines containing stable ER express a specific form of plasticity, mGluR-dependent long-term depression (Holbro et al., 2009). We are now using time-lapse two-photon imaging to investigate movements of dynamic ER in hippocampal neurons. Analysis of large 3D time series on the level of individual synapses is not trivial, and we have developed a machine learning approach to detect dendritic spines automatically and to follow them over time (Blumer et al., 2015). To our surprise, ER movements were correlated with changes in spine volume and could be triggered by strong synaptic stimulation, suggesting that ER selectively visits highly active synapses (see figure). Understanding the physiological role of ER in the visited spines is the goal of this project.

  • Cyclic nucleotide signaling in synaptic plasticity

    Daniel Udwari, Christine Gee

    An important second messenger system in neurons is based on the conversion of ATP into cAMP, a reaction catalyzed by the enzyme adenylyl-cyclase (AC). Stimulation of endogenous AC with pharmacological agents (forskolin) triggers the potentiation of synapses, a protocol known as ‘chemical LTP’. Such pharmacological manipulations affect all cell types in the tissue, making it difficult to interpret the results and distinguish between pre- and postsynaptic signaling. In collaboration with the group of Peter Hegemann, we could show that a photoactivated AC from a marine bacterium (bPAC) can be used to control cAMP levels in individual neurons by light (Stierl et al., 2011). To investigate the effects of cAMP elevation on synaptic plasticity, we expressed bPAC in postsynaptic neurons. To our surprise, even strong and sustained cAMP elevation did not induce long-term plasticity of active synapses. As pharmacological elevation of cAMP in all cells is known to induce synaptic plasticity, we suspect the critical compartment to be in the presynaptic neuron or in non-neuronal cell types. Using cell-type specific expression of bPAC, we will dissect the precise location of cAMP action. Furthermore, we develop new tools to control cAMP and cGMP signaling in subcellular compartments (Scheib et al. 2015).

  • . Vesicle cycling at Schaffer collateral synapses

    Tobias Rose, Iris Ohmert

    Studies in dissociated neuronal culture have suggested that not all vesicles in the presynaptic terminal can be mobilized by electrical stimulation. The pool of vesicles that are never used, even during periods of intense stimulation, has been termed the ‘resting pool’. We set out to revisit the question in a more physiological preparation, organotypic slice cultures of rat hippocampus. In order to calculate the fraction of vesicles released from individual Schaffer collateral boutons, we developed a ratiometric sensor based on pH-sensitive GFP and a dimeric red fluorescent protein. In immature cultures, we could indeed detect a resting pool of vesicles, but this pool completely disappeared after 2-3 weeks in culture (see figure). Mature Schaffer collateral synapses also displayed much faster endocytosis, making these synapses capable of sustained high frequency transmission during place cell firing in vivo (Rose et al., 2013).

    releasable vesicles
    In mature boutons, electrical stimulation releases all vesicles.

    Imaging vesicular glutamate release from individual Schaffer collateral boutons in mature organotypic culture. At rest, transmitter vesicles are acidic and green fluorescence is quenched. After stimulation with 1200 action potentials, green fluorescence increases as vesicles are released. Chemical neutralization of vesicular pH (+ NH4Cl, calibration ratio) does not lead to further increase in green fluorescence, indicating that all vesicles were already used during electrical stimulation.

  • Diffusional isolation of dendritic spines as a mechanism for metaplasticity

    Shuting Yin, Christian Schulze

    Dendritic spines act as miniature bio-reactors, trapping activated enzymes close to their targets for up to a second. It is thought that this biochemical isolation enables rapid modification of individual synapses while minimizing the impact on immediate neighbors (Wiegert and Oertner, 2011). We set out to test this hypothesis, comparing the plasticity of synapses on spines with long and thin necks, which are well isolated from the dendrite, to the plasticity of synapses on short and stubby spines. We have previously shown that the spine neck enables stronger depolarization of active synapses, boosting the influx of calcium ions during synaptic transmission (Grunditz et al., 2008; Holbro et al., 2010). On the other hand, diffusion of fresh glutamate receptors from the dendrite into the spine would favor less isolated synapses for potentiation. Plasticity induction by two-photon uncaging of glutamate in combination with optical calcium measurements enables us to compare the plasticity of isolated vs connected synapses.

  • Modulation of synaptic signaling by TRPM4 channels

    Christine Gee, Brenna Fearey,

    Collaboration with Institute for Neuroimmunology and Multiple Sclerosis TRPM4 channels are calcium-activated cation channels. Under conditions where calcium concentrations in neurons are high, they are further depolarized by activation of these channels, leading to a very dangerous situation. Indeed, pharmacological blockade or knock-out of TRPM4 channels makes neurons resistant to apoptosis-inducing stimulation, raising the question of the physiological function of these channels (Schattling et al., 2012). We use a knock-out mouse model, optogenetic stimulation and pharmacology to investigate the physiological role of TRPM4 channels in synaptic function and plasticity.

Future Perspectives

The focus of the institute will remain on synaptic function and plasticity with the ultimate goal to understand the entire life cycle of a synapse, from formation to removal.

We will continue to develop novel optogenetic tools that allow us to monitor and manipulate intracellular signaling in individual neurons with light.

The power of optogenetics allows us to follow the fate of individual synapses over several weeks and simultaneously interfere with electrical activity and biochemical signaling in a controlled and quantitative fashion.

For high speed imaging in 3D, we have constructed a two-photon microscope with advanced scanning mechanisms which allows us to sample functional signals from a large number of synapses near simultaneously. A new line of research will investigate the consequences of altered synaptic plasticity for network function.

In collaboration with Fabio Morellini, we will study the role of hippocampal synaptic plasticity in rodent foraging and spatial learning.