Information processing in the brain depends on synaptic transmission and plasticity. Such synaptic functions require specialized organization of neurotransmitter receptors, which are concentrated within membrane-bound protein condensates called postsynaptic density (PSD). The PSD formation through processes such as phase separation has been studied in vitro. However, how individual molecules are organized to form the PSD condensate in situ remains largely elusive, limiting our understanding of molecular mechanisms underlying synaptic formation and functions. Here, we have developed reference-free classification with uniformly over-sampled sub-tomograms to analyze cryo-electron tomograms of cultured hippocampal neurons, enabling us to identify type-A γ-aminobutyric acid receptor (GABAAR) in inhibitory synapses and determine its in situ structure at 19 Å resolution. GABAARs are organized hierarchically: First, the GABAAR super-complexes have a fixed 11-nm inter-receptor distance but variable relative angles; Second, the super-complexes connect to form semi-ordered 2-dimensional multi-receptor networks with reduced Voronoi entropy; Finally, the receptor networks in turn are organized into a mesophasic assembly with a sharp phase boundary. This assembly co-localizes with condensates of postsynaptic scaffolding proteins and putative presynaptic vesicle release sites. Such hierarchical self-organization from ordered super-complexes to mesophasic assembly allows synapses to achieve a “Goldilocks” state with a delicate balance between stability and flexibility, enabling both reliability and plasticity in information processing.

Authors: Yun-Tao Liu^1,3*, Chang-Lu Tao^1,*, Xiaokang Zhang^2,*, Lei Qi¹, Rong Sun¹, Pak-Ming Lau¹, Z. Hong Zhou^3,4,#, Guo-Qiang Bi^1,5,#

¹Center for Integrative Imaging, Hefei National Laboratory for Physical Sciences at the Microscale, and School of Life Sciences, University of Science and Technology of China, Hefei, Anhui, China;  ²School of Medicine and Center of Cryo-Electron Microscopy, Zhejiang University, Hangzhou, Zhejiang, China; ³California NanoSystems Institute, University of California, Los Angeles, CA, USA;  ⁴Department of Microbiology, Immunology and Molecular Genetics, University of California, Los Angeles, CA, USA; ⁵CAS Center for Excellence in Brain Science and Intelligence Technology, University of Science and Technology of China, Hefei, Anhui, China

*These authors contributed equally to this work. #Correspondence should be addressed to: Guo-Qiang Bi (gqbi@ustc.edu.cn) and Z. Hong Zhou (hong.zhou@ucla.edu)