Skip to main content

Development of novel tools and techniques for the study of the structure and dynamics of GABAergic inhibitory synapses

Final Report Summary - GABAAR (Development of novel tools and techniques for the study of the structure and dynamics of GABAergic inhibitory synapses.)

The overall objectives of the proposed project were to investigate the structure and protein dynamics of GABAergic inhibitory synapses through the development of high-resolution, single molecule imaging techniques that can be applied to cellular as well as integrated systems.

Aim 1. Determine the organization and receptor dynamics of GABAergic inhibitory synapses using photoactivated localization microscopy and three-dimensional imaging techniques. In order to study the gamma-aminobutyric acid type A receptor (GABAAR) and the scaffold protein gephyrin at super-resolution, a number of constructs were made. GABAAR subunits were tagged with the photoconvertible fluorescent protein Dendra 2 while gephyrin was tagged with the mEos2 photoconvertible fluorescent protein. In order to achieve a high and stable expression of these proteins in neurons, lentiviruses of the constructs were prepared and used for neuronal infection. All constructs were successfully expressed in both hippocampal and spinal cord cultured neurons.
GABAAR subunit constructs were tested for two-dimensional photoactivated localization microscopy (PALM) and single particle tracking PALM (sptPALM) and have been shown to exhibit expected properties. The proteins were successfully expressed at the neuronal membrane, exhibited high colocalization with gephyrin and were photoconvertible. All of these results indicate that the subunit constructs are oligomerizing into functional receptors and are being expressed at the surface with correct localization at synapses. sptPALM experiments have also been performed and initial diffusion results were similar to those previously obtained with quantum dots.
Emphasis was also placed on the gephyrin scaffold protein. In collaboration with colleagues of the Triller laboratory, a mEos2-tagged gephyrin construct was used to determine the ultra structure of gephyrin clusters at synapses and to quantify the number of gephyrin molecules within these clusters. It was found that on average, gephyrin molecules clustered with a density of 5000-10000 molecules per μm2. By combining PALM with stochastic optical reconstruction microscopy (STORM), the ultra structure of gephyrin was related to the synaptic distribution of glycine receptors (GlyRs). A high correlation between the scaffold and receptor proteins was observed and a 1:1 stoichiometry between gephyrin molecules and receptor binding sites was determined. Interestingly, competition between glycine and GABAA receptors for these binding sites was also observed.
Visualization of synaptic mEos2-gephyrin clusters at super-resolution revealed the existence of sub-domains with varying protein densities, some of which appeared to change their position with respect to each other. We developed an analytical approach based on the specific localization of individual proteins within a cluster and their movement over time. Our method takes advantage of the highly detailed and precise information obtained from super-resolution imaging to further investigate the internal dynamics and rearrangements within synaptic gephyrin clusters in live neurons. This novel approach was used to describe changes in the morphing of synaptic gephyrin clusters and its tuning through the disruption of cytoskeletal elements and synaptic activity. It was also used to determine the stability of clusters and the level of interaction between proteins, information which was previously not available. We showed that even though gephyrin clusters maintain an apparent stable structure over time, they do in fact exhibit significant internal dynamics that can be altered pharmacologically. These results confirm an important concept in neuroscience: although there is a level of stability, small rearrangements provide the needed plasticity for synaptic activity. Importantly, the described method of analysis has immediate application in the neuroscience field as well as other fields of science for the study of dynamic, clustered structures.
In all, we have established the use of PALM and sptPALM for structural, quantitative, and dynamic studies. These initial studies aimed to obtain a clear picture of the synaptic structure and organization of GABAARs and gephyrin molecules independent from one another.

Aim 2. Employ STORM to image the organization and dynamics of GABAARs genetically encoding unnatural amino acids and labeled with small organic fluorophores via click chemistry. The orthogonality of the aminoacyl-tRNA synthetase / suppressor tRNA (aaRs/tRNA) pair and the proper incorporation of the UAA were successfully demonstrated in COS and HEK cells using a model membrane receptor consisting of a pHluorin-tagged transmembrane domain. The pHluorin tag, which is located at the C-terminus of the transmembrane domain, was used to screen the efficiency of the system, since read-through of the amber codon by the aminoacyl tRNA synthetase should result in the expression of this fluorescent protein. Amber mutations were successfully generated via point mutations of amino acids located in the extracellular portion of the protein. In order to minimally disrupt the native protein structure, the UAA incorporated was a modified phenylalanine: p-azido-L-phenylalanine (AzF). We were able to show that in the presence of the UAA, the fluorescence of the model receptor containing the amber mutation was comparable to that of the wild- type receptor, suggesting the successful incorporation of the UAA. Whereas in the absence of the UAA, the full-length model receptor was not expressed, confirming the specificity of the synthetase for the UAA. Although at lower efficiency, mutagenesis and the successful insertion of the unnatural amino acid was also achieved in neurons.
The next step consisted in establishing the proper conditions to perform click chemistry in HEK cells, Xenopus laevis oocytes, and/or neurons. This process requires the covalent binding of organic fluorophores to the AzF UAA in the protein. For this purpose, we made use of the commercially available dibenzocyclooctyne (DIBO)-AlexaFluor647.
Insertion of AzF at a specific location within the receptors facilitated testing of functional properties via electrophysiology, while the use of fluorescent reporters allowed live cell imaging of receptors. In this process, we found that the use of AzF and (DIBO)-AlexaFluor647 exhibited non-specific fluorescence and slow reaction kinetics. Therefore, we further investigated the incorporation of UAAs and found that trans-cyclooctene-lysine (TCO) has increased kinetics compared to AzF. Implementation of TCO, along with the use of tetrazine dyes, resulted in significant progress in the incorporation and specific labeling of not only the model pHluorin-tagged transmembrane domain, but also functional neuronal receptors. This was all achieved in HEK cells and is now being tested in neurons.
We have continued our collaboration with chemists with the aim of developing a set of organic fluorophores that exhibits faster kinetics and minimum non-specific staining and can be used in combination with the TCO UAA for regular fluorescence and super-resolution microscopy.
In all, we have achieved insertion of an unnatural amino acid in COS-7 cells, HEK cells, neurons, and Xenopus laevis oocytes and have optimized click chemistry conditions in the latter system. Importantly, these UAAs were encoded in both model proteins and functional neuronal receptors. These studies will provide a way to tag innumerable types of proteins in a minimally disruptive fashion and will allow the study of protein structure, conformation, and dynamics using electrophysiology and fluorescence imaging.