Differential effect of neuraminidase-treatment on the surface charge-associated properties of rat reticulocytes and erythrocytes. Studies by partitioning in two-polymer aqueous phases

Rat reticulocytes undergo charge-associated surface changes, detectable by cell partitioning in charged dextran-poly(ethylene glycol) aqueous phase systems, as they become mature erythrocytes. Young reticulocytes have a lower partition coefficient, i.e., quantity of cells in the top phase as a percentage of total cells added, than do mature erythrocytes. Sialic acid is the main charge-bearing group on red blood cells and, in the case of the rat, most of the sialic acid can be removed by treatment of the cells with neuraminidase (Vibrio cholerae). By combining isotopic 59Fe-labeling of reticulocytes with countercurrent distribution of the entire red blood cell population in charged dextran-poly(ethylene glycol) aqueous phases we have now studied the relative effect of neuraminidase-treatment on rat reticulocytes and mature erythrocytes. It was found that neuraminidase-treatment (a) does not eliminate surface differences, detectable by partitioning, between rat reticulocytes and erythrocytes and (b) reduces the partition coefficient of mature erythrocytes to a greater extent than the partition coefficient of reticulocytes indicating a differential effect of this enzyme on the two cell populations.     

Differential effect of neuraminidase-treatment on the surface charge-associated properties of rat reticulocytes and erythrocytes: Studies by partitioning in two-polymer aqueous phases

Rat reticulocytes undergo charge-associated surface changes, detectable by cell partitioning in charged dextran-poly(ethylene glycol) aqueous phase systems, as they become mature erythrocytes. Young reticulocytes have a lower partition coefficient, i.e., quantity of cells in the top phase as a percentage of total cells added, than do mature erythrocytes. Sialic acid is the main charge-bearing group on red blood cells and, in the case of the rat, most of the sialic acid can be removed by treatment of the cells with neuraminidase (Vibrio cholerae). By combining isotopic ⁵⁹Fe-labeling of reticulocytes with countercurrent distribution of the entire red blood cell population in charged dextran-poly(ethylene glycol) aqueous phases we have now studied the relative effect of neuraminidase-treatment on rat reticulocytes and mature erythrocytes. It was found that neuraminidase-treatment (a) does not eliminate surface differences, detectable by partitioning, between rat reticulocytes and erythrocytes and (b) reduces the partition coefficient of mature erythrocytes to a greater extent than the partition coefficient of reticulocytes indicating a differential effect of this enzyme on the two cell populations.     

Detection of surface differences between two closely related cell populations by partitioning isotopically labeled mixed cell populations in two-polymer aqueous phases. I. Human red blood cell subpopulations

The partition behavior of cells in dextran-poly(ethylene glycol) aqueous phases (i.e., the cells' relative affinity for the top or bottom phase or their adsorption at the interface) is greatly dependent on the polymer concentrations and ionic composition and concentration. Appropriate selection of phase system composition permits detection of differences in either charge-associated or lipid-related surface properties. We have now developed a method that can reveal differences by partitioning that fall within experimental error if one were to compare countercurrent distribution (CCD) curves of two closely related cell populations run separately. One cell population is isotopically labeled in vitro (e.g., with 51Cr-chromate) and is mixed with an excess of the unlabeled cell population with which it is to be compared. The mixture is subjected to CCD and the relative specific radio-activities are determined through the distribution. As control we also examine a mixture of labeled cells and unlabeled cells of the same population. The feasibility of this method was established by use of cell mixtures the relative partition coefficients of which were known. The procedure was then used to test for human erythrocyte subpopulations. 51Cr-chromate-labeled human young or old red blood cells were mixed with unfractionated erythrocytes and subjected to CCD in a phase system reflecting charge-associated properties. It was found that older cells had a high, young cells (probably only reticulocytes) a low partition coefficient. Because of the small differences involved these results were not previously obtained. It was further determined, by repartitioning 51Cr-labeled cells from the left or right ends of a CCD of human red blood cells admixed to unlabeled unfractionated erythrocytes, that a subpopulation with higher partition coefficient exists (probably constituting the old red cells). These experiments serve to illustrate (a) that human red blood cells, contrary to a previous report, can be subfractionated by partitioning and (b) the usefulness of this new method in detecting smaller surface differences between closely related cell populations than was heretofore possible by partitioning alone.     

Detection of surface differences between two closely related cell populations by partitioning isotopically labeled mixed cell populations in two-polymer aqueous phases

The partition behavior of cells in dextran-poly(ethylene glycol) aqueous phases (i.e., the cells' relative affinity for the top or bottom phase or their adsorption at the interface) is greatly dependent on the polymer concentrations and ionic composition and concentration. Appropriate selection of phase system composition permits detection of differences in either charge-associated or lipid-related surface properties. We have now developed a method that can reveal differences by partitioning that fall within experimental error if one were to compare countercurrent distribution (CCD) curves of two closely related cell populations run separately. One cell population is isotopically labeled in vitro (e.g., with⁵¹Cr-chromate) and is mixed with an excess of the unlabeled cell population with which it is to be compared. The mixture is subjected to CCD and the relative specific radio-activities are determined through the distribution. As control we also examine a mixture of labeled cells and unlabeled cells of the same population. The feasibility of this method was established by use of cell mixtures the relative partition coefficients of which were known. The procedure was then used to test for human erythrocyte subpopulations⁵¹Cr-chromate-labeled human young or old red blood cells were mixed with unfractionated erythrocytes and subjected to CCD in a phase system reflecting charge-associated properties. It was found that older cells had a high, young cells (probably only reticulocytes) a low partition coefficient. Because of the small differences involved these results were not previously obtained. It was further determined, by repartitioning⁵¹Cr-labeled cells from the left or right ends of a CCD of human red blood cells admixed to unlabeled unfractionated erythrocytes, that a subpopulation with higher partition coefficient exists (probably constituting the old red cells). These experiments serve to illustrate (a) that human red blood cells, contrary to a previous report, can be subfractionated by partitioning and (b) the usefulness of this new method in detecting smaller surface differences between closely related cell populations than was heretofore possible by partitioning alone.     

Detection of surface differences between two closely related cell populations by partitioning isotopically labeled mixed cell populations in two-polymer aqueous phases. II. A correction

The partition behavior of cells in dextran-poly(ethylene glycol) aqueous phases (i.e., the cells' relative affinity for the top or bottom phase or their adsorption at the interface) is greatly dependent on the polymer concentrations and ionic composition and concentration. Appropriate selection of phase system composition permits detection of differences in either charge-associated or lipid-related surface properties. We have now developed a method that can reveal differences by partitioning that fall within experimental error if one were to compare countercurrent distribution (CCD) curves of two closely related cell populations run separately. One cell population is isotopically labeled in vitro (e.g., with 51Cr-chromate) and is mixed with an excess of the unlabeled cell population with which it is to be compared. The mixture is subjected to CCD and the relative specific radio-activities are determined through the distribution. As control we also examine a mixture of labeled cells and unlabeled cells of the same population. The feasibility of this method was established by use of cell mixtures the relative partition coefficients of which were known. The procedure was then used to test for human erythrocyte subpopulations. 51Cr-chromate-labeled human young or old red blood cells were mixed with unfractionated erythrocytes and subjected to CCD in a phase system reflecting charge-associated properties. It was found that older cells had a high, young cells (probably only reticulocytes) a low partition coefficient. Because of the small differences involved these results were not previously obtained.(ABSTRACT TRUNCATED AT 250 WORDS)     

Separation of cell mixtures by immunoaffinity cell partitioning: strategies for low abundance cells

The partitioning of cells in aqueous two-phase systems formed by poly(ethylene glycol) (PEG) and dextran can be changed by incubating the cells with a PEG-modified antibody directed specifically against its surface. We have developed a new approach for immunoaffinity cell partitioning (IACP) in which the antibodies are first reacted with tresylated monomethoxy PEG (TMPEG) in sodium phosphate buffer, pH 7.5, the excess TMPEG is quenched by reaction with bovine serum albumin, and the resulting preparation is used directly for incubation with the cells without any isolation of the monomethoxyPEG (MPEG)-antibody conjugates. We have demonstrated the specificity of this IACP method by showing that MPEG-modified anti-human red blood cell antibody increases the partition of human erythrocytes from the interface to the PEG-rich top phase (up to 100%) but not the partitioning of either neutrophils or HL60 cells. Irrelevant antibodies do not affect the partitioning of red blood cells. The partitioning behaviors of erythrocytes and HL60 cells in mixtures varying from 75 to 10% red blood cells subjected to IACP are similar to those of the pure cell population, i.e., erythrocytes ca. 100% and HL60 cells 3% in top phase. Thus, the population of erythrocytes can be almost completely extracted into the top phase in a single step. The contaminant cells represent only a small percentage (less than 5% in most of the cases) of the cell mixture recovered in top phase. Both cell populations can be completely separated by countercurrent distribution (CCD).(ABSTRACT TRUNCATED AT 250 WORDS)     

Aging of erythrocytes results in altered red cell surface properties in the rat,but not in the human Studies by partitioning in two-polymer aqueous phase systems

Aqueous solutions of dextran and of poly(ethylene glycol) when mixed give rise to two-phase systems useful in separating cells, on the basis of their surface properties, by partitioning. Depending on whether salts with unequal or equal affinity for the two phases are chosen, phases with or without an electrostatic potential difference between the phases are obtained. At appropriate polymer concentrations the former yield cell partition coefficients (i.e., the quantity of cells in the top phase as a percentage of total cells added) based on charge-associated surface properties while the latter reflect membrane lipid-related parameters. With increasing cell age, rat erythrocytes have diminishing partition coefficients in both charged and uncharged phases. Using the elevated aspartate aminotransferase levels of younger red cells as a marker, we have now found that young mature erythrocytes of human do not have the highest partition coefficient in the red cell population as they do in rat. Experiments with isotopically labeled dog red cells yield results similar to those found with human erythrocytes. Furthermore, density-separated young and old red cells from human give overlapping countercurrent distribution curves. Finally, counter-current distribution of human red blood cells followed by pooling of cells from the left and right ends of the distribution and subjection of these cells to a redistribution gives curves that overlap with each other and with the original countercurrent distribution. This indicates that not only are human red cells not subfractionated based on possible age-related surface alterations, but also that they are not subfractionated by partitioning based on any surface parameter.These results are consistent with our previous findings that membrane sialic acid/hemoglobin absorbance is essentially constant through the extraction train after countercurrent distribution of human erythrocytes in a charged phase system; and with the recent reports of others that there is no difference in electrophoretic mobility between human young and old red cells.     

Aging of erythrocytes results in altered red cell surface properties in the rat, but not in the human. Studies by partitioning in two-polymer aqueous phase systems

Aqueous solutions of dextran and of poly(ethylene glycol) when mixed give rise to two-phase systems useful in separating cells, on the basis of their surface properties, by partitioning. Depending on whether salts with unequal or equal affinity for the two phases are chosen, phases with or without an electrostatic potential difference between the phases are obtained. At appropriate polymer concentrations the former yield cell partition coefficients (i.e., the quantity of cells in the top phase as a percentage of total cells added) based on charge-associated surface properties while the latter reflect membrane lipid-related parameters. With increasing cell age, rat erythrocytes have diminishing partition coefficients in both charged and uncharged phases. Using the elevated aspartate aminotransferase levels of younger red cells as a marker, we have not found that young mature erythrocytes of human do not have the highest partition coefficient in the red cell population as they do in rat. Experiments with isotopically labeled dog red cells yield results similar to those found with human erythrocytes. Furthermore, density-separated young and old red cells from human give overlapping countercurrent distribution curves. Finally, countercurrent distribution of human red blood cells followed by pooling of cells from the left and right ends of the distribution and subjection of these cells to a redistribution gives curves that overlap with each other and with the original countercurrent distribution. This indicates that not only are human red cells not subfractionated based on possible age-related surface alterations, but also that they are not subfractionated by partitioning based on any surface parameter. These results are consistent with our previous findings that membrane sialic acid/hemoglobin absorbance is essentially constant through the extraction train after countercurrent distribution of human erythrocytes in a charged phase system; and with the recent reports of others that there is no difference in electrophoretic mobility between human young and old red cells.     

Effect of cell exposure to top or bottom phase prior to cell partitioning in dextran-poly(ethylene glycol) aqueous phase systems: erythrocytes as a model.

Cells exposed to dextran (Dx)-rich bottom phase prior to cell partitioning in Dx-poly(ethylene glycol) (PEG) aqueous two-phase systems have lower partition ratios than cells exposed to PEG-rich top phase. Aspects of this previously observed phenomenon were explored. In the present work charge-sensitive phases made with Dx T500 and PEG 8000 were used exclusively. It was found that: (1) even on countercurrent distribution (CCD) red cells (RBC) loaded in bottom phase have a lower apparent partition ratio, G, than the same cells loaded in top phase; (2) when part of the same cell population is loaded into top phase and part into bottom phase of the same load cavities for CCD, with the cells loaded into top or bottom bearing an isotopic tracer (51Cr), the cells loaded into top phase have a higher G value than the cells loaded into bottom phase; (3) the shift in the CCD curves of human or of rat RBC between cells loaded in top or bottom phase using systems having the same polymer concentration (though different salt compositions) shows no striking difference and is, for the number of experiments run, not statistically significant; (4) when the quantity of cells loaded for CCD is reduced from 10(9) to 10(8), the G value of cells loaded in top phase is reduced slightly while that of cells loaded in bottom phase is diminished more appreciably; (5) increasing polymer concentrations yield larger differences in G values between (rat) RBC loaded in top or bottom phase; (6) when cells exposed to top or bottom phase, respectively, are centrifuged and suspended in bottom or top phase, respectively, their CCD patterns are qualitatively similar to cells exposed to these latter respective phases initially; (7) rat RBC populations containing 59Fe-labeled cells of different but distinct age are fractionated on CCD irrespective of whether loaded in top or bottom phase. An exception are populations containing very young mature labeled cells (e.g., 4-d old) which are resolved when loaded in top phase but not in bottom phase. Thus cell populations exist which can be resolved by CCD when loaded in one of the phases but not when loaded in the other. Glutaraldehyde-fixed rat RBC containing 4-d old labeled cells are fractionated by CCD irrespective of whether loaded in top or bottom phase.     

Choice of an aqueous polymer two-phase system for cell partition

Partition of human erythrocytes in aqueous two-phase polymer systems produced by Ficoll and different molecular weight fractions of dextran and polyethylene glycol and the influence of the ionic composition on the cells' partition in the systems was studied. It is found that the Ficoll-dextran-40 system is characterized by a number of advantages as compared with the common dextran-polyethylene glycol system or the others systems under study. The main advantage of the system appears to be that it is possible to concentrate the red cells in the top phase or in the bottom phase of the system, depending on the system ionic composition. The influence of the nature and the concentration of salt additives on this two-phase system formation is examined.     

Partitioning of pristinamycins in aqueous two-phase systems: A first step toward the development of antibiotic production by extractive fermentation

The partitioning of pristinamycins was studied in dextran and polyethylene glycol (PEG) aqueous two-phases systems. Pristinamycins partitioned preferentially into the PEG-rich top phase. The partition coefficient was independent of molar mass of PEG and dextran and of antibiotic concentration, but, increased exponentially with the tieline length of the system. Partition of pristinamycins was greatly improved when fatty acids esters of PEG were mixed with PEG. In such mixtures, the partition of coefficient increased up to a value of 24, dependent on the carbon chain length of fatty acids and the modified PEG concentrations. Moreover, in such system, the two groups of pristinamycins, I and II, were extracted in accordance with their hydrophobicity. Recovery of pristinanamycins produced by Streptomyces pritinaespiralis in a fermentation broth was achieved with a dextran/PEG system. Cells were confined into the bottom phase and pristinamycins partitioned in the top phase. However, due to binding of the pristinamycins to the cells, the partition coefficient was slightly lower than of pure antibiotics solutions. (c) 1994 John Wiley & Sons, Inc.     

Differential effect of pH on the density and volume of rat reticulocytes and erythrocytes. Relevance to their fractionation by centrifugation.

Rats were injected with 59Fe-ferrous citrate and bled thereafter at different times (16 h to 49 d). This gave rise to red cell populations in which cells corresponding in age to the time elapsed between injection and bleeding were labeled. The anticoagulant used was either acid-citrate-dextrose (ACD) with a pH adjusted to 7.3 or ACD (pH 5.1). Final pH of the collected blood was about 7.2-7.4 in the former case and 6.4-6.7 in the latter. Red cells were then centrifuged (5) and approximately 7-10% of the packed cells from the top and 7-10% from the bottom of the cell column collected. When reticulocytes are the predominant labeled red cell population, as in blood obtained for about 24 h after isotope injection, a fractionation of these cells and mature erythrocytes is in evidence only when blood is collected at the higher pH. Thus, at pH 7.2-7.4 ratios of specific radioactivities of cells in top fraction/cells in an unfractionated sample are about 3, whereas at pH 6.4-6.7, the analogous ratios are 1 or less. These differences in specific activity ratios, as a function of pH at collection, virtually disappear after about 4 d following isotope injection. The lower pH is known to increase the volume and decrease the density of mature red blood cells. The marked effect of pH on cellular fractionation could be correlated with the smaller change in rat reticulocyte density and volume in acid medium. At pH 6.4-6.7, the densities of mature erythrocytes and reticulocytes are so close that their physical separation by centrifugation is not feasible.     

Production of pectinases by Polyporus squamosus in aqueous two-phase system

The ability of Polyporus squamosus to grow and produce pectinases in an aqueous two-phase medium composed of polyethylene glycol and crude dextran is reported. Fungal growth was restricted to the bottom phase leaving the top phase cell free. Amounts of produced biomass and endo and exo-pectinase activities were superior or equal to those obtained in homogeneous medium. The partition coefficient for the endo-pectinase was 1.52 followed by a top phase yield of 70.86%. Although the phase system composition favours partition of a greater part of exo-pectinase activity to the bottom phase (K(exo) was 0.6 and yield in top phase 48.56%) the partitioned activity in the top phase was equal to that produced in homogeneous cultivation.     

Immunoglobulin-positive mononuclear cells in human peripheral blood: detection by mixed anti-globulin and direct anti-globulin-rosetting reactions.

Ig-bearing mononuclear cells were identified in Ficoll-Hypaque preparations of human peripheral blood by using mixed anti-globulin (MAG) and direct anti-globulin rosettes; indicator cells consisted of sheep erythrocytes coated with human F(ab')2 or anti-F(ab')2 antibody, respectively. Of the cell population isolated from 10 normal subjects, a mean of 68% was lymphocytes. However, fewer than 50% of the cells with detectable surface Ig were lymphocytes. On viable cell preparations using chromic chloride-treated sheep erythrocytes (CrCl3SRBC) coated with anti-F(ab')2 antibody, a mean of 20.1% of the lymphocytes formed rosettes, i.e., were B. Up to 6% of peripheral blood lymphocytes formed mixed Ig-rosettes and E-rosettes. On viable lymphocytes using F(ab')2-coated CrCl3SRBC, MAG rosettes were insensitive in detection of B lymphocytes. Formaldehyde treatment of lymphocytes increased the number of B cells detectable to 25.5% of the lymphocyte population. Study of T-enriched and B-enriched populations showed that the observed increase in B cell reactivity was real and not due to MAG-rosetting T cells. A one-stage procedure for T and B lymphocyte separation is described.     

Fixation with even small quantities of glutaraldehyde affects red blood cell surface properties in a cell- and species-dependent manner. Studies by cell partitioning

Partitioning in charge-sensitive dextran-poly(ethylene glycol) aqueous phase systems reveals that fixation with even small concentrations of glutaraldehyde (e.g., 0.1% w/v) changes the surface properties of cells. While fixation with larger concentrations of glutaraldehyde (i.e., 1.85%) increases erythrocyte partition ratios, the effect of lower glutaraldehyde concentrations on the partition ratios appears to be species-specific. The differential effect of glutaraldehyde on rat reticulocytes and erythrocytes indicates that fixation is also cell-dependent. These data, together with the previous report that glutaraldehyde fixation does not change the characteristicrelative partition ratios of rat mature erythrocytes of different cell ages, suggest that the nature and extent of glutaraldehyde alteration of cell surfaces must, in each case, be empirically evaluated.     

Nonspecific human lymphocyte-mediated cytotoxicity induced by immobilized IgG aggregate

Normal human lymphocytes were induced to lyse nonsensitized erythrocytes when concomitantly incubated on immobilized IgG aggregate with various erythrocyte target cells. These included ox, sheep, chicken, and human red blood cells. Only immobilized aggregate would initiate cytolysis. The IgG aggregate was prepared from normal, healthy adult donors and did not possess target cell specificity (e.g., human erythrocyte lysis was initiated by autologous IgG). Normal human lymphocytes could be induced to lyse the red blood cell targets only after a preincubation with adherent mononuclear cells; however, freshly prepared lymphocytes depleted of IgG-FcR^? cells were cytolytic. Cytolysis induced by immobilized IgG-aggregate can be distinguished from NCMC and ADCC by its requirement for immobilized IgG aggregate and the absence of target cell specificity in the IgG-aggregate preparation.     

Differential effect of pH on the density and volume of rat reticulocytes and erythrocytes

Rats were injected with⁵⁹Fe-ferrous citrate and bled thereafter at different times (16 h to 49 d). This gave rise to red cell populations in which cells corresponding in age to the time elapsed between injection and bleeding were labeled. The anticoagulant used was either acid-citrate-dextrose (ACD) with a pH adjusted to 7.3 or ACD (pH 5.1). Final pH of the collected blood was about 7.2–7.4 in the former case and 6.4–6.7 in the latter. Red cells were then centrifuged (5) and approximately 7–10% of the packed cells from the top and 7–10% from the bottom of the cell column collected. When reticulocytes are the predominant labeled red cell population, as in blood obtained for about 24 h after isotope injection, a fractionation of these cells and mature erythrocytes is in evidence only when blood is collected at the higher pH. Thus, at pH 7.2–7.4 ratios of specific radioactivities of cells in top fraction/cells in an unfractionated sample are about 3, whereas at pH 6.4–6.7, the analogous ratios are 1 or less. These differences in specific activity ratios, as a function of pH at collection, virtually disappear after about 4 d following isotope injection. The lower pH is known to increase the volume and decrease the density of mature red blood cells. The marked effect of pH on cellular fractionation could be correlated with the smaller change in rat reticulocyte density and volume in acid medium. At pH 6.4–6.7, the densities of mature erythrocytes and reticulocytes are so close that their physical separation by centrifugation is not feasible.     

Subfractionation of cell populations by partitioning in dextran-poly (ethylene glycol) aqueous phases

Partitioning of cells in dextran-poly(ethylene glycol) aqueous two-phase systems depends on the interaction between the surface properties of the cells and the physical properties of the phases. The latter can be manipulated to a considerable extent by selection of polymer concentrations and ionic composition and concentration. If salts (e.g., phopshate) are used that have an unequal affinity for the two phases, an electrostatic potential difference between the phases results and, at appropriately high polymer concentrations, the partition coefficient of cells is determined predominantly by membrane charge-associated properties. By “balancing” the magnitude of the electrostatic potential difference against that of the interfacial tension (primarily a function of polymer, but also phosphate, concentrations) one can obtain phase systems that give usable partition coefficients for most cell populations (1). In work under way in our laboratory on the effects of different chemical and enzymatic modifications on the relative surface properties of rat red blood cells of different ages, we have now found that certain phase compositions did not resolve such treated cell subpopulations while other phase compositions did. Thus not all charged phase systems in which cell populations as a whole have usable partition coefficients are equally capable of detecting or subfractionating cell subpopulations. It is therefore essential, before drawing conclusions on the nonseqarability of cell subpopulations, to test cell separability in charged phase systems of different compositions if the system initially chosen does not afford a subfractionation.     

Synthesis and application of a poly(ethylene glycol)-antibody affinity ligand for cell separations in aqueous polymer two-phase systems

The possibility of producing biospecific affinity ligands for separating cells in two polymer aqueous phase systems on the basis of cell surface antigens was investigated. Rabbit anti-human erythrocyte IgG was reacted with cyanuric chloride-activated monomethyl poly(ethylene glycol) (PEG) fractions (molecular weights approximately 200, 1900, and 5000) at various molar ratios of PEG to protein lysine groups. The partition coefficient of the protein in a Dextran/PEG two-phase system increased with increasing degree of modification and increasing PEG molecular weight. There was a concomitant loss in ability to agglutinate human erythrocytes. The ability of the modified IgG to bind to a DEAE-cellulose column was almost eliminated by reaction with the PEG 5000, and was decreased to a lesser extent by PEG 1900. This PEG 1900-modified IgG substantially increased the partition of fresh or fixed human erythrocytes into the PEG-rich phase of a suitable phase system, while having no effect on rabbit cell partition. The partition increase could be inhibited by unmodified anti-human red cell IgG but not by nonspecific unmodified human IgG, demonstrating that the ligand effects were specific for the cell type against which the antibody was raised. A mixture of rabbit and human erythrocytes, which ordinarily have very similar partitions in the phase systems used, could be separated on a countercurrent distribution apparatus using the modified IgG. These results demonstrate the feasibility of producing immunologically specific affinity partition ligands for cell separation.