126 



M. S. C. BIRBECK AND E. H. MERCER 



mentation in fresh medium. We have found that an 

 improved separation may be obtained more quickly 

 by using layering or gradient methods in a swing-out 

 head centrifuge. When the medium contains high 

 molecular weight components, which are required 

 to improve the morphological appearance of the 

 fractions, density gradients may be obtained without 

 increasing the osmotic strength of the medium. 



Fixation of cell fractions by the standard buffered 

 osmium tetroxide solution (7) is satisfactory; no 

 improvement has been found either by the addition 

 of sucrose or high molecular weight components to 

 the fixative, or by using a solution of osmium tetr- 

 oxide in the suspending medium as a fixative. A 

 small volume of a resuspended fraction is added to a 

 larger volume of fixative; the fixed suspension is 

 then spun to produce a pellet. The pellet, which 

 should be less than 0.5 mm thick, is treated as a 

 lump of tissue. It is dehydrated and embedded in 

 methacrylate as in the conventional method. The 

 pellet is oriented on the microtome so that a single 

 section contains all the strata produced by sedimen- 

 tation; in this way a proper estimate of the distribu- 

 tion of material in the pellet may be obtained. 



A promising variant of these methods has been 

 devised by us in which no attempt is made to isolate 

 separate fractions. The tissue is homogenised in a 

 medium, and the suspension is "layered" over a 

 medium of higher density in a centrifuge tube. 

 Using a swing-out head, the suspension issedimented 

 to form a pellet about 0.1 mm thick. This will con- 

 tain "strata" of all components of the suspension 

 separated according to their sedimentation rate. The 

 whole pellet may then be fixed, sectioned and the 

 composition of the various layers determined electron 

 microscopically. As an example of its use, consider 

 the problem of the uptake of a radio-active amino 

 acid in a tissue. The compound is injected, after a 

 certain time tissue is taken from the animal, ho- 

 mogenised and centrifuged by this method. The block 

 is sectioned; a thin section is examined in the micro- 

 scope and a radio autograph is made from a con- 

 tiguous thick section. By comparing these, the par- 

 ticular cell fraction in which the radioactive isotope 

 has been taken up may be found. The method may 

 also be used for the intracellular localisation of other 

 constituents which are still detectable by histo- 

 chemical methods after fixation. 



Methods for nuclei. — Nuclei, separated by many 

 of various standard methods, have been examined 

 (3). None of these methods gives nuclei of 

 comparable appearance to those in sections of fixed 

 tissue; many methods give nuclei, which appear 

 clean in the light microscope, but are badly con- 

 taminated by cytoplasmiccomponents when examined 

 in the electron microscope. Nuclei prepared by the 

 method of Philpot and Stanier (8) were found to be 

 particularly interesting. This method, which only 

 gives small yields, consists of centrifugation in a 

 sucrose glycerol medium with the addition of glycero- 



Fig. 1. Electron micrograph of a single nucleus in a fraction 

 prepared by the method of Philpot and Stanier. The nucleolus 

 may be seen, but the nuclear membrane is absent. 

 Inset. A high power micrograph of the nucleoplasm, showing 

 filaments about 100 A in diameter. 



Fig. 2. Electron micrograph of a few mitochondria prepared 

 in the dextran medium. The internal and external membranes 

 may be seen. 



Inset. A micrograph at the same magnification of a particle 

 in a light mitochondrial fraction, which might be a lysosome. 



phosphate These nuclei (fig. 1) from rat liver are 

 relatively clean of contaminating cytoplasm; this 

 may be partially the result of the loss of the nuclear 

 membrane, to which the cytoplasmic debris tends to 

 adhere. There are. however, numerous small fila- 

 ments (ca. 100 A diam.) dispersed throughout the 

 nuclei (fig. 1, inset). These filaments which signifi- 

 cantly have the same diameter as nucleo-protein 

 filaments, may be an artifact due to the method of 

 separation; however, it is also possible that they also 

 exist in sections of whole tissue, but are rendered 

 invisible by an interstitial substance, which is lost 

 during separation, possibly as a consequence of the 

 removal of the nuclear membranes. 



Methods for niitochondria. — The conventional me- 

 thod for the separation of mitochondria uses 0.25 M 

 sucrose as a suspending medium (5). Mitochondria 

 prepared by this method and by improved methods 

 have been examined by several workers (1, 2, 

 6, 9). Mitochondria prepared in 0.25 M sucrose 

 are swollen, empty and usually contaminated 

 with microsomal debris. We have described an 

 improved complex medium containing 0.25 M 

 raffinose, 6 "o dextran and other additives. Mi- 

 tochondria prepared in this medium (fig. 2) are 

 not swollen and retain their internal double mem- 

 branes and other contents. When the centrifugation 

 conditions are carefully controlled it is possible to 

 obtain cleaner mitochondria using this medium. This 

 may be judged both by the lack of microsomal 

 material in the electron micrographs and also bio- 

 chemically by their low R.N. A. assay. Although the 

 higher density and viscosity of this medium would 

 appear to make high centrifugation speeds necessary, 

 this is not so for mitochondria as the absence of 



