LOCKE AND HIRSCH 105 1 



may be dialyzed and further purified by a fractional precipitation procedure before 

 concentration. Huntoon has obtained antipneumococcus immune-substance extracts 

 in this way "approximating in antibody content the best immune sera, and contain- 

 ing so low a serum protein content that 5-cc. amounts sensitize guinea pigs only ir- 

 regularly to subsequent injections of horse serum.'" His best preparations contain 

 860 times less nitrogen per immune unit than the immune serum from which they 

 were derived. Ottenberg and Stenbuck have obtained anti-typhoid immune extracts 

 of a similar, remarkable degree of purity.' 



The improved extraction procedures which are the essence of this later work and 

 which have made possible the preparation of immune-substance extracts in prac- 

 ticable yields are, unfortunately, by no means selective. They produce no notable 

 dissociation of the antigen-immune substance union represented by the original, 

 washed antigen-immune serum adsorption sediment but create, rather, a disintegra- 

 tion and dispersion of the sediment, out of which a portion of the antigen substances 

 can be removed by direct centrifugation. A further portion of the dispersed antigen 

 material is removed by the various iso-electric separation procedures which have been 

 proposed, but a considerable fraction persists in the final extracts. 



However, the antigen-immune substance union may be broken, in the experience 

 of the authors,' by a process of ether extraction, permitting the recovery of immune- 

 substance preparations which are almost entirely free from contamination with dis- 

 persed antigen. The ether extraction causes a shift in the iso-electric range of lipoid- 

 containing antigen from a value identical with that of the specific immune substance 

 to a value sufficiently removed to permit the loosening of the antigen-immune sub- 

 stance union and the precipitation of the antigen residue. The procedure, as adapted 

 to the purification of an immune hemolysin, is as follows: 



Fresh sheep blood is defibrinated, centrifugated, and the cell sediment washed five times 

 with physiological salt solution. Fifteen cc. of the packed, washed cell sediment are equally 

 divided between two centrifuge tubes, and to each portion there is added quickly and with 

 vigorous shaking, 40 cc. of perfectly clear, fresh rabbit anti-sheep serum. After two hours, 

 the cells have laked completely and the stroma have flocculated and settled toward the 

 bottom of the tubes. Centrifugation for forty-five minutes completes the separation; 90-99 

 per cent of the hemolysin originally present has become bound to the stroma sediment. The 

 brilliantly clear, red supernatant liquid is decanted from the sediment and replaced by an 

 equal volume of physiological salt solution. After the stroma is finely suspended in the wash 

 liquor by prolonged shaking, the suspension is allowed to stand for twenty minutes and is 

 then strongly centrifugated for thirty minutes. The washing process is repeated (about six 

 times) until the supernatant liquid has no trace of color and gives no trace of foam when 

 shaken. The stroma obtained is perfectly white and has lost little of the originally bound 

 hemolysin. 



The well-washed, hemolysin-saturated stroma is extracted with ether three times. After 

 the removal of the third ether extract, the ether remaining dissolved in the stroma material 

 is removed by centrifugation in a warm centrifuge. The stroma residue material packs at the 

 bottom of the tube, and the salt solution, which made up the cell volume, may be decanted. 

 The residue is washed twice with physiological salt solution and once with distilled water. 

 Considerable hemolysin is lost to the salt solution, but protein impurities, due to surface 

 adsorption, are thereby almost completely removed. The washed residue is extracted re- 



' See p. 1050, note 6. 



