Demonstration of a Highly Ordered Pattern upon a Mitochondrial Surface

Synaptic bodies (spherules) of the rod cells of the cat retina commonly contain several conspicuous mitochondria. These have been observed frequently to aggregate in an unusually intimate way, in pairs, or in small clusters of as many as seven (Fig. 2). An exceptional feature of this association is that the apposing surface membranes of adjacent mitochondria come to lie closely together with a strictly uniform spacing of approximately 125 Angstrom units between them. In such areas of contact the outer apposed osmiophilic membranes assume a markedly beaded appearance having great regularity (Figs. 1 and 3). The periodicity illustrated by these structured membranes is 160 Angstrom units. This latter measurement has been made with considerable accuracy, for it has been possible to count over I00 consecutive periods, and to compare their average length with a replica of a diffraction grating micrographed under conditions as nearly identical as possible. Poorly defined, osmiophilic cross-linkages are also sometimes visible from one membrane to the other across the uniform gap (Fig. 3). Thus, the two apposing membranes demonstrate a repeating lattice structure of constant dimensions. Where adjacent membranes diverge from one another there is an abrupt change to smoothly contoured surface lines without evidence of periodic structure (Fig. 1). Thus the free surfaces of these mitochondria are in no way marked by unusual features. Also, the inner osmiophilic lines of the mitochondrial envelopes, as well as their reflections over all the surfaces of the cristae, are without a visible organizational pattern. The gap between inner and outer lines is not of particularly regular width, although of course distortions may have occurred during preparative techniques. Clearly, this space lacks the uniformly structured order of the gap between adjacent mitochondria. Thus, these spaces are fundamentally dissimilar in physical character. Much of our present detailed knowledge of cytomembranes derives from our current understanding of the structure of nerve myelin and its relationship to the Schwann cell surface. Thus, Robertson (5) has introduced the concept of the "uni t membrane" and has studied in detail how their surfaces as found in plasma membranes can associate with neighboring membranes, at times to form truly fused or combined structures which he terms "compound membranes." The association between adjacent mitochondrial membranes described here, however, is not so intimate as in an "internal compound membrane" for the gap between the osmiophilic lines can be seen to be a direct extension of the general cytoplasm (Fig. 1). I t is only very rarely that a visibly o rdered cross-linkage between adjacent plasma membranes has been described. But, in desmosome-like structures, particularly in invertebrates, patterns somewhat comparable to that described here occasionally have been noted. The most elaborate of these has been found by Wood in Hydra (9) where a repeating linear component was demonstrated. In the present study sufficiently oblique sections were obtained through the apposed membranes of adjacent mitochondria to reveal linear patterns had they been present. None was found, however, so it is concluded that here we are dealing with an arrangement different from that in the Hydra desmosome. I t is perhaps more comparable with the periodic pattern observed in desmosome-like structures found in the axon sheaths of earthworms by H a m a (2). Whether the latter is truly homologous with the presently described periodic structure is a moot question. I t is evident that it has only been possible to visualize a marked periodicity of cytomembranes under very exceptional circumstances. Synaptic bodies of mammal ian retinal receptor cells have been examined in detail in three previous studies. Sj6strand (7) has made a three dimensional reconstruction of the guinea pig synapse, based upon an exhaustive study of serial sections. The published illustrations indicate the presence of only a few small mitochondria in this area. Ladman (3) studied the synaptic bulbs of rats, and commented that there is regularly but one large mitochondrion in each spherule of this