= Protein subunit
A \b[protein subunit] is an individual <polypeptide> component of a <protein complex>. The number of <protein subunits> counts chain copies, whereas the number of subunit types counts distinct kinds of chain. An assembly $A_2B_2$ has four <protein subunits> of two types. A <protein domain> is a region within a chain and need not be a separate <protein subunit>; the arrangement of several chains belongs to <protein quaternary structure>.
The alpha/beta <tubulin> <heterodimer> illustrates the distinction: it contains two related but different <polypeptides>, not two domains cut out of one chain. The 1998 electron-crystallographic model resolved the alpha/beta architecture at $3.7$ angstrom resolution, using zinc-induced <tubulin> sheets.<Structure of the alpha beta tubulin dimer by E. Nogales and others>[Abstract and PDB entry 1TUB: distinct alpha and beta chains and electron-crystallographic method.]{ref}
Subunit counting can also become an optical experiment. In <total internal reflection fluorescence microscopy>, tagged membrane <proteins> can be observed as isolated fluorescent spots; successive <photobleaching> steps reveal individual fluorescent tags disappearing. <Subunit counting in membrane-bound proteins by Maximilian H. Ulbrich and Ehud Y. Isacoff> used this approach on membrane <proteins> expressed in frog oocytes. For NR1/NR3B receptor assemblies, tagging one component type mainly gave two steps, while tagging both gave four, supporting a two-plus-two arrangement.<Subunit counting in membrane-bound proteins by Maximilian H. Ulbrich and Ehud Y. Isacoff>[Figures 1–2 and results on NR1/NR3B stoichiometry.]{ref}
A count of visible steps need not equal the true number of <protein subunits>: some tags fail to fluoresce, and two tags may bleach too close together to resolve. The experiment therefore combines step counts with a model of labeling and detection, rather than interpreting every two-step trace as proof of a two-chain <protein complex>.<Subunit counting in membrane-bound proteins by Maximilian H. Ulbrich and Ehud Y. Isacoff>[Discussion of fluorescent-tag probability, approximately 80%, and binomial interpretation of bleaching-step distributions.]{ref}
Back to article page
