There are four major types of cell–cell junctions in the polarized epithelial cells of vertebrates. Farquhar and Palade (1963) defined three of them.
Electron microscopy supplied the means to look at adhesion structures, but by the early 1960s the field had settled only on desmosomes (Wood, 1959) and something called a terminal bar, which may or may not have been a collection of desmosomes. Other investigators came across structures that they named quintuple-layered cell interconnections, external compound membranes, and nexuses. But there was no study that brought together and organized these various observations.
Farquhar and Palade remedied this situation by defining, based on their very different characteristic appearances and relative locations, three structures. The zonula occludens (closing belt; now known as the tight junction) was nearest the apical surface of the columnar epithelium. It featured closely apposed membranes, and was shown to act as a diffusion barrier. Next was the zonula adherens (adhering belt; now known as the adherens junction) with an intercellular space separating strictly parallel membranes, all framed by dense cytoplasmic matrices. Most basal were the desmosomes (here named the macula adhaerens [adhering spot], a name that did not last). Whereas the first two junctions were continuous and belt-like, the desmosomes were dispersed buttons of adhesion.
Farquhar and Palade (when they were both at Rockefeller University, New York, NY) had originally set out to study how the kidney glomerulus filters the blood to form urine. They induced glomerular damage and saw that filtration of tracers by the basement membrane was compromised. But in response the epithelium formed intercellular connections that were “tightened and increased in depth” (Farquhar and Palade, 1961). That morphological change associated with increased permeability made Marilyn Farquhar take a closer look. To see the tripartite structure of the junction properly, she says, “you need just the right section. It takes a lot of looking. You have to have a reason for looking, and people hadn't had that functional connection before.”
The new work surveyed junctions in 13 different epithelial tissues in exhaustive detail. Farquhar says that she and George Palade were in no hurry to publish a less complete account. “It was not like now where there are multiple groups breathing down each others' necks,” she says. “Nobody was doing work on this topic at this level of detail.” As a result, she says, the 1963 paper “was pretty much accepted from the beginning as the definitive work in the area, because nobody had done anything comparable.”
Farquhar acknowledges that in the 1960s “it was much easier to find new things with the electron microscope. But it was just as demanding as it is now to put that structure into a functional context.” Many people had excellent microscopes and high quality preparations, but “once it got past the looking they were stuck.” Adding the functional data, such as the impermeability of tight junctions to protein tracers, “was really a hallmark of the Rockefeller school at that time,” she says.
The initial study reported on the orderly organization of a columnar epithelium, but a similar hierarchy was found in the less rigidly organized stratified cells of the epidermis (specifically in the basal layer; Farquhar and Palade, 1965). The depth (number of occluding strands) in a tight junction was later found to correlate with the leakiness or tightness of a tight junction (Claude and Goodenough, 1973). Eventually, the molecular components of the junctional complexes were determined: the occludin (Furuse et al., 1993) and claudins (Furuse et al., 1998) of tight junctions; actin-linked cadherins of adherens junctions; and intermediate filament-linked cadherins of desmosomes.