Visualization of clustered protocadherin neuronal self-recognition complexes.

Neurite self-recognition and avoidance are fundamental properties of all nervous systems1. These processes facilitate dendritic arborization2,3, prevent formation of autapses4 and allow free interaction among non-self neurons1,2,4,5. Avoidance among self neurites is mediated by stochastic cell-surface expression of combinations of about 60 isoforms of α-, ß- and γ-clustered protocadherin that provide mammalian neurons with single-cell identities1,2,4-13. Avoidance is observed between neurons that express identical protocadherin repertoires2,5, and single-isoform differences are sufficient to prevent self-recognition10. Protocadherins form isoform-promiscuous cis dimers and isoform-specific homophilic trans dimers10,14-20. Although these interactions have previously been characterized in isolation15,17-20, structures of full-length protocadherin ectodomains have not been determined, and how these two interfaces engage in self-recognition between neuronal surfaces remains unknown. Here we determine the molecular arrangement of full-length clustered protocadherin ectodomains in single-isoform self-recognition complexes, using X-ray crystallography and cryo-electron tomography. We determine the crystal structure of the clustered protocadherin γB4 ectodomain, which reveals a zipper-like lattice that is formed by alternating cis and trans interactions. Using cryo-electron tomography, we show that clustered protocadherin γB6 ectodomains tethered to liposomes spontaneously assemble into linear arrays at membrane contact sites, in a configuration that is consistent with the assembly observed in the crystal structure. These linear assemblies pack against each other as parallel arrays to form larger two-dimensional structures between membranes. Our results suggest that the formation of ordered linear assemblies by clustered protocadherins represents the initial self-recognition step in neuronal avoidance, and thus provide support for the isoform-mismatch chain-termination model of protocadherin-mediated self-recognition, which depends on these linear chains11.

Protocadherin cis-dimer architecture and recognition unit diversity.

Clustered protocadherins (Pcdhs) mediate numerous neural patterning functions, including neuronal self-recognition and non-self-discrimination to direct self-avoidance among vertebrate neurons. Individual neurons stochastically express a subset of Pcdh isoforms, which assemble to form a stochastic repertoire of cis-dimers. We describe the structure of a PcdhγB7 cis-homodimer, which includes the membrane-proximal extracellular cadherin domains EC5 and EC6. The structure is asymmetric with one molecule contributing interface surface from both EC5 and EC6, and the other only from EC6. Structural and sequence analyses suggest that all Pcdh isoforms will dimerize through this interface. Site-directed mutants at this interface interfere with both Pcdh cis-dimerization and cell surface transport. The structure explains the known restrictions of cis-interactions of some Pcdh isoforms, including α-Pcdhs, which cannot form homodimers. The asymmetry of the interface approximately doubles the size of the recognition repertoire, and restrictions on cis-interactions among Pcdh isoforms define the limits of the Pcdh recognition unit repertoire.

Discovery of an O-mannosylation pathway selectively serving cadherins and protocadherins.

The cadherin (cdh) superfamily of adhesion molecules carry O-linked mannose (O-Man) glycans at highly conserved sites localized to specific ß-strands of their extracellular cdh (EC) domains. These O-Man glycans do not appear to be elongated like O-Man glycans found on α-dystroglycan (α-DG), and we recently demonstrated that initiation of cdh/protocadherin (pcdh) O-Man glycosylation is not dependent on the evolutionary conserved POMT1/POMT2 enzymes that initiate O-Man glycosylation on α-DG. Here, we used a CRISPR/Cas9 genetic dissection strategy combined with sensitive and quantitative O-Man glycoproteomics to identify a homologous family of four putative protein O-mannosyltransferases encoded by the TMTC1-4 genes, which were found to be imperative for cdh and pcdh O-Man glycosylation. KO of all four TMTC genes in HEK293 cells resulted in specific loss of cdh and pcdh O-Man glycosylation, whereas combined KO of TMTC1 and TMTC3 resulted in selective loss of O-Man glycans on specific ß-strands of EC domains, suggesting that each isoenzyme serves a different function. In addition, O-Man glycosylation of IPT/TIG domains of plexins and hepatocyte growth factor receptor was not affected in TMTC KO cells, suggesting the existence of yet another O-Man glycosylation machinery. Our study demonstrates that regulation of O-mannosylation in higher eukaryotes is more complex than envisioned, and the discovery of the functions of TMTCs provide insight into cobblestone lissencephaly caused by deficiency in TMTC3.

A two-site flexible clamp mechanism for RET-GDNF-GFRα1 assembly reveals both conformational adaptation and strict geometric spacing.

RET receptor tyrosine kinase plays vital developmental and neuroprotective roles in metazoans. GDNF family ligands (GFLs) when bound to cognate GFRα co-receptors recognize and activate RET stimulating its cytoplasmic kinase function. The principles for RET ligand-co-receptor recognition are incompletely understood. Here, we report a crystal structure of the cadherin-like module (CLD1-4) from zebrafish RET revealing interdomain flexibility between CLD2 and CLD3. Comparison with a cryo-electron microscopy structure of a ligand-engaged zebrafish RETECD-GDNF-GFRα1a complex indicates conformational changes within a clade-specific CLD3 loop adjacent to the co-receptor. Our observations indicate that RET is a molecular clamp with a flexible calcium-dependent arm that adapts to different GFRα co-receptors, while its rigid arm recognizes a GFL dimer to align both membrane-proximal cysteine-rich domains. We also visualize linear arrays of RETECD-GDNF-GFRα1a suggesting that a conserved contact stabilizes higher-order species. Our study reveals that ligand-co-receptor recognition by RET involves both receptor plasticity and strict spacing of receptor dimers by GFL ligands.

Molecular basis of sidekick-mediated cell-cell adhesion and specificity.

Sidekick (Sdk) 1 and 2 are related immunoglobulin superfamily cell adhesion proteins required for appropriate synaptic connections between specific subtypes of retinal neurons. Sdks mediate cell-cell adhesion with homophilic specificity that underlies their neuronal targeting function. Here we report crystal structures of Sdk1 and Sdk2 ectodomain regions, revealing similar homodimers mediated by the four N-terminal immunoglobulin domains (Ig1-4), arranged in a horseshoe conformation. These Ig1-4 horseshoes interact in a novel back-to-back orientation in both homodimers through Ig1:Ig2, Ig1:Ig1 and Ig3:Ig4 interactions. Structure-guided mutagenesis results show that this canonical dimer is required for both Sdk-mediated cell aggregation (via trans interactions) and Sdk clustering in isolated cells (via cis interactions). Sdk1/Sdk2 recognition specificity is encoded across Ig1-4, with Ig1-2 conferring the majority of binding affinity and differential specificity. We suggest that competition between cis and trans interactions provides a novel mechanism to sharpen the specificity of cell-cell interactions.

RET recognition of GDNF-GFRα1 ligand by a composite binding site promotes membrane-proximal self-association.

The RET receptor tyrosine kinase is essential to vertebrate development and implicated in multiple human diseases. RET binds a cell surface bipartite ligand comprising a GDNF family ligand and a GFRα coreceptor, resulting in RET transmembrane signaling. We present a hybrid structural model, derived from electron microscopy (EM) and low-angle X-ray scattering (SAXS) data, of the RET extracellular domain (RET(ECD)), GDNF, and GFRα1 ternary complex, defining the basis for ligand recognition. RET(ECD) envelopes the dimeric ligand complex through a composite binding site comprising four discrete contact sites. The GFRα1-mediated contacts are crucial, particularly close to the invariant RET calcium-binding site, whereas few direct contacts are made by GDNF, explaining how distinct ligand/coreceptor pairs are accommodated. The RET(ECD) cysteine-rich domain (CRD) contacts both ligand components and makes homotypic membrane-proximal interactions occluding three different antibody epitopes. Coupling of these CRD-mediated interactions suggests models for ligand-induced RET activation and ligand-independent oncogenic deregulation.