Among my favorite papers are those from the journals of microscopy and ultrastructure research way back when the TEM was first available on a commercial scale. Those guys had great eyes, nothing short of eagle vision. The limitations on figures, and words in an article were not so stringent and the micrographs and wordy descriptions abound. This included descriptions of the pretty obvious cell structures and organelles… yep, the desmosome.  The potential for examining cells was wide open, and insects, plants and fossils were favorites. The tools were primitive, yes, but the dedication was impressive. I am not too far removed from that group….I would have loved todays tools in last century’s electron microscopy boom. All that “looking” requires time, and in the warp factor 9 world of sequencing and mining….it falls behind. There is, however, no substitute for its value.

So, here is another article from the last century (LOL) showing totally different types of desmosomes (septate and comb desmosomes) the latter of which has layering with a central dense line – a little reminiscent of mammalian desmosomes (in hydra), and also desmosomes (found in silk worm), which actually have some similarities (but big differences) to mammalian desmosomes. Reference is here.
These are amazing structures, however, none of the images shown in this publication have any “look” of a connection with mitochondria as an energy and Ca+ supply, nor do they look like they have connections with the thin intermediate filaments.  In fact next to the dense plaque there is a lucent area, then there are the microtubules parallel to the plasmalemma.

One thing that these desmosomes in silkworms have is a central dense intercellular line that was not seen in insect desmosomes in the moth. (vague suggestions of a presence or absence of a central dense line (which in mammalian desmosomes is the interaction of the EC1 domains of two mirrored cadherin molecules that are part of two adjacent cells).

This article calls the picture lower right a hydra septate desmosome. I have to read more of their literature to be sure, certainly they mention the inner part of the trilaminar plasmalemma as being separated more from the outer leafelet, and the outer leafelet being justaposed by dense proteins… (in this case probably analagous to the outer dense plaque in mammalian desmosomes (i.e. the plakophilin and plakoglobin?), and I can see that there is slightly more rigidity to that portion of the plasmalemma.  Just for my own edification, i made a mash-up of some portions of one of their figures with measurements made using their own micron bar markers (red bars=50nm). Top image is a “normal” non-desmosomal membrane with membrane to membrane thickness of 18.5nm (brown bar and text) and an intercellular space between the two cells about 13nm (green bar and text). The two images below are their desmosomes. Left lower image looks like a sandwich of desmosomes (three periods=curley bracket) and red bar=50nm which appears to have three iterations (they show pictures of many more than three iterations). The micrograph (from which I have cropped just the desmosome) on the right is single, has two plasmalemma (one from justaposed cells) and a central dense line (see my line drawn in as an overlay). There are differences in the size of the intercellular space of the desmosome and the non-desmosomal plasmalemma above. Pink line is thicker than the non desmosomal juxtaposed cell membranes, extracellular space from outer leaflet of each cell is pretty similar to top image and the central dense line runs about 6 nm in.