During brain development, neurons connect with each other using branched projections called dendrites. As a dendrite extends away from its neuron’s cell body to connect with dendrites from other neurons, it may also encounter other dendrites originating from the same cell. Neurons must therefore avoid forming connections between dendrites from the same cell, which would cause redundant connectivity. This self-avoidance process depends on proteins called clustered protocadherins, named after the fact that the genome contains a cluster of 53 (in humans) varieties, also called isoforms, of this type of protein. Neurons express 5-15 isoforms of these proteins, and a neuron’s identity is determined by the combination of clustered protocadherins expressed, meaning there are a huge number of possible neuron identities. When two dendrites meet, the clustered protocadherins encounter each other, but only interact if they are the same isoform. A neuron recognizes that it encountered one of its own dendrites – and thus avoids a self-synapse by signaling dendrite pruning – if and only if all protocadherin isoforms are the same on both dendritic membranes and can thus all interact across the two dendrites. Otherwise a synapse or connection will be allowed between dendrites with mismatched isoforms.
A clustered protocadherin is a type of membrane protein with an extracellular domain containing six “cadherin repeats”, a membrane anchor, and a short cytoplasmic tail. The cadherin repeats are responsible for interactions with other protocadherin molecules both in trans, i.e. between dendrites, and in cis, i.e. on the same dendrite. The trans interaction is specific, meaning it only happens with the same isoform, while the cis interaction is likely promiscuous and helps a dendrite confirm that all of its trans interactions are satisfied.
In our new study just published in eLife in collaboration with Debora Marks’s lab (Harvard Medical School), we used x-ray crystallography to determine the atomic structure of a fragment of the human B3 clustered protocadherin isoform. This fragment is responsible for the trans interactions, and the structure reveals how two B3 molecules form an antiparallel homodimer using four of the six cadherin repeats that make up the extracellular portion of clustered protocadherins (see figure). Gratifyingly, this structure confirms our earlier predictions.
We also used bioinformatics analyses to investigate how isoform self-specificity – formation of homodimers – is determined. One particularly noteworthy observation is that different protein interface regions are responsible for the strength (or affinity) and the specificity of the homodimer interaction (see figure).
The human genome contains additional, non-clustered, protocadherin genes which code for proteins that have the same domain organization as clustered protocadherins and are involved in cell adhesion in the nervous system. With our earlier predictions now confirmed, we asked whether the clustered protocadherin homodimer architecture is conserved in these other protocadherins. Our sequence coevolution analysis leads us to predict that at least some of these non-clustered protocadherins indeed dimerize similarly to the clustered protocadherins.
Now that we understand the homophilic interactions of clustered protocadherins across two dendrites, a next step is to examine the interactions that occur between isoforms on the same dendrite, giving rise to complex cell identities that account for all isoforms expressed in a cell. Another future challenge is to uncover the signaling pathway triggered by matching clustered protocadherin interactions to produce self-avoidance. Lastly, the prediction that some non-clustered protocadherins have the same homodimeric architecture as the clustered protocadherins needs to be experimentally validated.
Figure caption: Our structure reveals protocadherin dimer interactions that use the first four (each a different shade of blue or green, and labeled on one molecule) of six cadherin repeats, with the two protocadherin molecules aligned antiparallel to each other. The structure of the 5th and 6th repeats is still unknown. The sixth repeat is then followed by a transmembrane helix and a cytoplasmic tail. The persistence and diversification of this form of trans interaction through evolution leaves traces such as coevolving residue pairs (yellow dumbbells). Different regions of the interface (labeled) are responsible for affinity and specificity. Promiscuous cis interactions on the same dendrite promote the combinatorial behavior of clustered protocadherins that specifies complex neuron cell identities.
Read more at eLife or download PDF
Authors: Jack M Nicoludis, Bennett E Vogt and Rachelle Gaudet