Selective membrane permeability to a relatively small erythropoiesis stimulatory factor (ESF) was demonstrated ( 1-3). A mixture of a relatively large ESF and an ESF-generating factor (EGF) was separated by selective membrane filtration (4). This report describes the isolation of four erythropoiesis regulatory factors (ERF) , including an erythropoiesis inhibitory factor (EIF) , by combining an electrofractionation technique with selective membrane permeability.
Materials and Methods.A urine concentrate of ESF, fraction I1 + I11 from a patient with paroxysmal nocturnal hemoglobinuria, was the starting material (5). The posthypoxic-polycythemic mouse was used as an assay animal (6). Results were calculated as the mean and standard error of the mean of the ESF activity of five or more mice and reported as percentage 59Fe incorporated during 48 hr and as International Units of standard B (7) ; assays with high activities were repeated at levels commensurate with linear dose-response data. Nitrogen was determined by a micro-Kjeldahl method.Fraction I1 + I11 was placed in the center section of an electrofractionation apparatus ( Fig. 1) designed and constructed for the purpose of isolating ERF. Since the buffers became progressively acidic at the anode and alkaline at the cathode during operation, the buffers chosen were 0.13 M phosphate and pH 7.9 at the anode and 5.0 at the cathode.The entire system was 0.7 M NaCl. The solutions in the anode and cathode buffer compartments were replaced every 3.5 days to 1 Supported in part by Grants FR-0061, FR-5365 and HE-12958 from the National Institutes of H 4 t h .reestablish the initial pH gradient between the compartments; during the period of 3.5 days the pH of the solution in the center section decreased from 7.4 to 7.0. Although the experiments were done at 3 t 2' and high ionic strength, enough phenol was added to the buffers to make the entire system 0.1% phenol as recommended by Lowy and Keighley (8). The chlorine and hydrogen gases evolved during electrolysis were aspirated. The solutions in the sections nearest the buffer compartments were stirred continuously. The fractions in the first anode and cathode sections were collected and replaced with fresh buffer after every 2 wk of electrophoresis through 8 wk. During this period the pH of the solution in the center section decreased to 5.9.During preliminary experiments various combinations of membranes2 were tried. Ultimately an appropriate selection of membranes was made as follows: The buffer compartments (capacity 1,500 ml each) were separated from the fraction collecting sections (capacity 250 ml each) by membranes with a molecular weight cutoff at 10,000; the first anionic fraction collecting section was separated from the second section by a membrane with a molecular weight cutoff at 20,000; the first anionic fraction collecting section was separated from the center section by a membrane with a cutoff at 30,000; the center section (capacity 500 ml) was separated from the first cationic fraction collecting sectio...