38 PER-AKE ALBERTSSON 



an increase in the limiting viscosity number [12, 13] and in the laevorota- 

 tion [14] at low pH values indicating a change in the structure of the 

 protein. It has also been reported [ 1 5] that there is an increase in the number 

 of hydrophobic groups on the protein surface at low pH values which 

 could explain why the protein at these pH values favours the top phase 

 containing more methyl groups than the bottom phase. If this explanation 

 is correct it means that partition provides a method for the determination 

 of the " hydrophobicity " of proteins and particles. 



It was mentioned above that the partition in the dextran methylcel- 

 lulose system is only influenced to a small degree by the electrolyte com- 

 position. In contrast, the dextran-polyethylene glycol system shows many 



Theoretical curve 

 Enzyme activity 



tiAn 



20 



40 60 80 

 Tube number 



Fig. 

 dextran 



5. Countercurrent distribution of partly purified ceruloplasmin in a 

 -polyethylene glycol system. From refs. [3] and [18]. 



Striking salt eifects [3, 16]. Thus, a protein or particle may be transferred 

 almost entirely from one phase to the other by changing the salt content. 

 As may be seen in Fig. 4, where the partition coefficients of a number of 

 proteins in a dextran-polyethylene glycol system are recorded, phycoery- 

 thrin is in the top phase {K= ^) with o-oi m phosphate buffer pH = 6-8. 

 If NaCl is added up to o • i m, the protein is transferred to the bottom 

 phase {K=o-2)', further additions of NaCl increase the K value so that 

 the proteins again partitions in favour of the top phase. When the con- 

 centration of NaCl is increased from i to 5 m, all proteins studied transfer 

 to the top phase. Further information on the partition of proteins in the 



