The interaction of iron-conalbumin (anion) complexes with chick embryo red blood ccells.

Abstract Equilibrium binding studies and iron transport rate studies of the interaction of conalbumin and transferrin with the chick embryo red cell have supported the following conclusions: (a) that the chick embryo red cell is comparable to the rabbit and human reticulocyte systems traditionally used for the study of the physiological function of the metal complexes of siderophilins; (b) that there may be two classes of binding sites on the chick embryo red cell membrane for Fe2conalbumin, one for which conalbumin competes and one for which it does not compete; (c) that the red cell membrane accepts iron from differiconalbumin at twice the rate as from monoferriconalbumins, suggesting that simultaneous transport takes place from the two iron-binding sites on the protein with respect to iron acquisition; and (d) that the substitution of oxalate for carbonate as the obligate anion in the Fe2conalbumin(anion)2 complex decreases the rate of iron uptake by the cell because the cell must release the anion from the protein complex before it is able to sequester the bound iron.

From the Department of Biochemistry, University of Vermont College of Medicine, Burliqdon, Vermont 05/,Of SUMMARY Equilibrium binding studies and iron transport rate studies of the interaction of conalbumin and transferrin with the chick embryo red cell have supported the following conclusions: (a) that the chick embryo red cell is comparable to the rabbit and human reticulocyte systems traditionally used for the study of the physiological function of the metal complexes of siderophilins; (b) that there may be two classes of binding sites on the chick embryo red cell membrane for Fezconalbumin, one for which conalbumin competes and one for which it does not compete; (c) that the red cell membrane accepts iron from differiconalbumin at twice the rate as from monoferriconalbumins, suggesting that simultaneous transport takes place from the two iron-binding sites on the protein with respect to iron acquisition; and (d) that the substitution of oxalate for carbonate as the obligate anion in the Fe2conalbumin(anion)z complex decreases the rate of iron uptake by the cell because the cell must release the anion from the protein complex before it is able to sequester the bound iron.
Since the report of Walsh et al. (1) that the anucleate erythroid cell, the reticulocyte, is the circulating cell responsible for the acquisition of transferrin-bound iron required for the synthesis of hemoglobin, several groups of investigators have carried out fundamental studies describing the interaction of iron-transporting proteins with reticulocytes. Jandl et al. (2) studied human reticulocytes from patients with severe anemia and established that iron transport from transferrin is a function of cellular oxidative metabolism and of pH and temperature values within the physiological range, and that iron is not freely available to chelators such as EDTA during the transport process. Schade (3) and Morgan and Laurel1 (4)  ferrin system, and demonstrated that the process of iron transport is saturable with respect to the metalloprotein complex. Morgan and his co-workers (5-8) have refined the rabbit reticulocyte studies to define the kinetics and equilibria of binding of the metalloprotein to the cell and the metabolic requirements for the transport of iron to the cell.
Although there is a considerable stock of physical and chemical data available regarding the transition metal complexes of siderophilins,' particularly transferrin and conalbumin, no consensus has yet been reached with respect to the mechanism of iron binding to the protein or the mechanism of iron release from the metalloprotein complex to the reticulocyte. Arguments have been presented to support the concept that the two metal-binding sites are equal and independent (9, 10) and, on the contrary, that iron binds to the two sites in a cooperative manner (11,12). It now appears that the two iron-binding sites differ as revealed by their spectra (13,14). With respect to the release of iron from the protein to the reticulocyte, Schade (3) suggested cooperative release while Fletcher and Huehns (15,16) proposed that the reticulocyte accepts one iron from iron-transferrin more readily than the other. Morgan and Appleton (8) have presented autoradiographic data which suggest that the reticulocyte takes up the entire metalloprotein complex rather than the iron atoms alone.
The resolution of such conflicting information requires not only an understanding of the internal relationship of the metal-binding sites on the protein, but also delineation of their physiological relationship.
To this end, the chick embryo red cell system used in the present study was developed.
This cell system is convenient and reliable, and data obtained from it are intended to complement data obtained from the reticulocyte systems currently in use.

EXPERIMEWTAL PROCEDURES
All solutions and dilutions were made with glass-distilled water. Chemicals were reagent grade and were used without further purification.
Conalbumin and Transfer% Preparation-Conalbumin from hen's egg white was prepared by the method of Woodworth and Schade (17) with the exception that the protein was chromato-1 We adopt "siderophilin" as the class name for all reversible iron-binding globulins found in biological fluids of animal origin and restrict the specific names to particular proteins, e.g. conalbumin is the siderophilin from egg white, transferrin is the siderophilin from blood plasma. This Ils:ige was suggested by Philip Aisen.
graphed on CM-Sephadex C-50 instead of on CM-cellulose. Chicken transferrin was prepared from hen and cock plasma by a modification of the conalbumin preparation which employed isoelectric focusing (18) to separate transferrin from contaminating conalbumin.
Both proteins were stored in the iron-free state at -20".
The iron complexes of conalbumin and transferrin were prepared by adding the appropriate amount of standard 20 mM fersous perchlorate (G. F. Smith Chemical Co.) to the protein dissolved in buffered saline (0.15 M NaCl, 20 mM NaHCOa).
For iron-transport rate studies, 59FeC13 (New England Nuclear) or 55FeCI, (Amersham-Searle) was diluted with the standard ferrous perchlorate solution to obtain the desired specific activity.
Iron binding was measured from the characteristic absorbance of the iron-conalbumin complex at 465 nm. Spectral data for conalbumin have been summarized (17), and the same values were used for chicken transferrin since it differs from conalbumin only in its carbohydrate content (19). Mixed isotope and mixed metal complexes of conalbumin were prepared as previously described (14). Metalloconalbumin complexes containing anions other than carbonate were prepared by adjusting the conalbumin solution to pH 6 in a Thunberg cuvette, adding the anion solution to a final concentration four times that of the conalbumin, and placing enough metal ion solution in the bulb nearly to saturate the conalbumin.
After evacuation (aspirator) to remove COZ, the metal ion solution was tipped in and mixed, and the pH was raised to approximately 8 by addition of NH8 gas through a rubber septum in the sidearm.
About 0.5 ml of air was also admitted to provide sufficient oxygen to oxidize the ferrous ion to the ferric state in the protein complex.
Spectra were taken, then the complexes were dialyzed versus Con-free water to remove ammonium ion, which is inhibitory to the cells. Spectra taken after dialysis showed the characteristic colors of the complexes to be stable and nondialyzable.
Iron-saturated conalbumin and transferrin were trace-labeled with 1251CI by the method of McFarlane (20) at a calculated level of 1 iodine atom per protein molecule, for the equilibrium binding studies. 12510dide solution was obtained from New England Nuclear. stage of development averaged 10 g.
The shell, covering the air space, and the chorioallantoic membrane were removed, and the major extraembryonic blood vessels were severed. The resulting suspension of chick embryo red cells in allantoic fluid was removed from the egg and diluted with a modified Earle's salts solution (21) containing 0.12 M NaCl, 5 mM KCI, 1 rnM Na&IP04, 1 mM MgS04, 1 mivr CaCb, 6 mM glu-U&k Embryo Red Cell Preparation-Red blood cells were taken from white leghorn embryos after 14 to 16 days incubation at 37" and 80% relative humidity.
Embryo wet weights at this gentle shaking in a 37" water bath for 15 min under a continuously flowing atmosphere of 95% air-5% COS. After equilibration, 5gFe-labeled conalbumin or transferrin was added to give the desired experimental protein concentration. Zero time in the rate determination was taken just after adding and mixing the labeled protein with the cells. Incubation was continued as above.
Sampling for a rate study was carried out at appropriate time intervals by removing an aliquot of the cell suspension and immediately diluting it IO-fold with cold Earle's salts solution.
The cells were centrifuged rapidly and washed twice with additional cold suspending medium.
The radioactivity contained in the washed red cell pellet was determined in a well-type scintillation counter.
Hematocrit and pH determinations were made on each reaction mixture at the conclusion of an experiment.
Metal and Anion Release Studies-Incubation mixtures were made up as described above, but disappearance of radiolabeled metal ions and anions from the conalbumin complex was followed by counting a 50-or 100~~1 aliquot of supernatant fluid in a Beckman LSC 250 liquid scintillation counter, rather than by counting the washed cell pellet.
The 55Fe isotope emits a low energy y-ray which can be efficiently counted by this method.
By using a variable &o-set module in the LSC 250 and sacrificing the most energetic 20% of the 55Fe counts, we were able to limit the interference from 14C or 5gFe to 6 and 11% of their respective counts.
In the anion exchange studies the incubation mixture was placed in a dialysis sac mounted on a rapid dialysis rack (22) and dialyzed against Earle's salts solution containing 20 mM NaHC03.
The pH was maintained by bubbling 5% CO*-95% air or 5% COZ-95 y0 NZ through the medium, and the whole apparatus was housed in a 37" water bath. As the anion was released, it dialyzed into the outside medium.
Disappearance of 55Fe and [14C]oxalate was followed as described above by sampling at appropriate intervals the contents of the dialysis sac, by attaching a syringe to a fine plastic tube inserted into the top of the dialysis sac and reaching to the bottom. case, and 2.5 mg per ml of bovine serum albumin (Fraction V, Sigma Chemical Co.). The cells were centrifuged at 700 x g and 4', then washed twice by resuspending in cold diluent, and recentrifuging.
The resulting red cell pellet was very lightly packed.
A standard curve of hematocytometer counts on serial dilutions of the cells plotted against hematocrits determined on the same dilution series allowed cell counts to be made for individual experiments by simply determining the hematocrit of the reconstituted blood used in the experiment. Hematocrits were measured in capillary tubes.
Metal Acquisition Rate XtudiesOne volume of the washed red cell pellet described above was added to 2 volumes of Earle's salts sohrtion which had been made 20 mM in NaHC03, giving a nominal hematocrit of 20% (1.4 x 109 cells per ml).
The suspension was equilibrated with respect to temperature and pH by equilibrium was determined by layering an aliquot of the reaction mixture onto a convenient volume of cold, isotonic sucrose Equilibrium Binding &u&es-Scatchard equilibrium binding (0.200-ml aliquot, 7 ml of 0.25 M sucrose) in a centrifuge tube. The cells were centrifuged through the sucrose column, leaving studies (23) were carried out in the following manner.
the interstitial fluid containing unbound radioactive protein at Red the surface of the column.
cells and suspending medium with bicarbonate were combined In the absence of red cells, no radioactive protein penetrated the sucrose column during centrifugation.
in the proportions described above.
Protein bound per cell was determined from the radio-

AND DISCUSSIOS
From a Scatchard plot of the binding 12"I-Fezconalbumin to the chick embryo red cell (Fig. I), the average number of binding sites per cell, n, at equilibrium is 2 x lo5 with an apparent binding constant of 1 X lo6 M-l with respect to protein. 13aker and Morgan (6) have reported similar values (3 X 1Oj per cell, 1 x 10" M-I) for the rabbit reticulocyte-transferrin system. Fig. 1 also demonstrates the lack of significant binding of Y-Fezconalbumin to the adult chicken red blood cell. In a similar study, the binding sites for 1251-Feztransferrin were determined to be 1 x lo5 per cell, with an apparent binding constant equal to that for conalbumin.
The value of n has been observed to vary within a factor of 2 from one cell preparation to another. The basis for this variability is not presently understood. Fig. 2 is a Scatchard plot comparing the binding of 12zI-Fezconalbumin to the chick embryo red cell as a function of the anion bound in the metalloprotein complex.
The binding parameters of 1251-Fe2conalbumin(oxalate)z and 129-Fezconalbumin(glycinate)z are similar to those for the 12"1-Fe2conalbumin-(carbonate)2 control. However, rate studies discussed below indicate that bound oxalate is capable of inhibiting the delivery of iron from conalbumin to the chick embryo red blood cell. The effect of competing, iron-free conalbumin on the observed value for the Scatchard parameter z is described by Fig. 3. The average number of protein molecules bound per cell at equilibrium is plotted with respect to the ratio of [conalbumin] to [1251-Fe2conalbumin].
1251-Fe2conalbumin was held constant at 6 pM, and t; was determined for that species in the presence of G to 120 PM conalbumin.
The distinct departure from linearity as values for T; decrease in response to increasing conalbumin concentration suggests that there is a class of binding sites for '251-Fe2conalbumin on the red cell membrane for which the metal-free protein successfully competes, as well as a class of sites for which the metalloprotein has a greater affinity. No Scatchard studies were performed with the iron-free protein because iron bound to conalbumin could not be completely removed subsequent to trace iodination.
Metal Acquisition Rate Studies-The chick embryo red cell system was found to be comparable to reticulocyte cell systems with respect to metabolic control of the metal transport rate. The optimum $1 range was 7.2 to 7.4, a result in agreement with that reported by Jandl et al.
(2) for a human reticulocytetransferrin system. Low temperatures, hypoxic conditions, and inhibitors of oxidative metabolism such as 2,4-dinitrophenol also severely inhibit the rate of iron transport from Fezconalbumin to the chick embryo red cell. Comparable data have been reported for the rabbit reticulocyte-transferrin system (7).
Given optimum metabolic conditions for the transport of iron, the rate may be limited by the relative concentrations of membrane-binding sites and reacting protein.
Since iron transport from iron-siderophilin to the reticulocyte is dependent on the formation of an equilibrium complex, the rate of iron uptake by the cell will be a function of the metalloprotein concentration whenever it is too low to saturate the membrane-binding sites. In practical terms, reticulocytes incubated with metallosiderophilins for extended periods, such that the siderophilin becomes depleted of metal, may show rate limitation owing to depletion of substrate.
Lowered iron-transport rates under these circumstances may be misinterpreted unless it is shown that substrate rate limitation is not occurring. Fig. 4 illustrates a determination of the Fezconalbumin concentration required for iron-transport rates independent of substrate concentration in the chick embryo system. 5gFe-saturated plasma was incubated with two levels of chick embryo red cells so that the hematocrits were 25 and 500/,. The decrease in iron uptake after 90 min is due to depletion of substrate in the mixture containing the greater number of cells. At this point, the Fenconalbumin concentration was approximately 7 pM. In all experiments where metal-transport rates were compared, care was taken to exceed this concentration. Fletcher and Huehns (15,16) have described experiments with a reticulocyte-transferrin system which show a decrease in the rate of iron transport from protein depleted of iron by prior incubation with reticulocytes, as compared with an untreated control.
They interpret their observations as indicating that the young red cell accepts iron more readily from one iron-   binding site on the protein than from the other iron-binding site. Fig. 5 illustrates a duplication of Fletcher and Huehn's general experimental design in the chick embryo red cell system. The iron uptake rate measured over the full 2-hour period is the control system with an initial Fesconalbumin concentration of 15 FM. At successive 30.min intervals, an aliquot of the control mixture was removed and centrifuged to separate the red cells. The supernatant fluid containing iron-depleted con-MINUTES FIG. 6. Rates of 59Fe uptake by chick embryo red cells from the following species: upper set of rates, jgFe;-conalbumin(p1) plus 1 eq of 59Fe ( l ), 59Fezconalbumin(pI) (O), and "gFe2conalbumin (control) (W); lower sel of Yates, jgFeiconalbumin(pI) (0 ), EeFe,-SgFei-conalbumin (A), Cr,59Fe;conalbumin (0 ), Ga,hgFeiconalbumin ( l ), and 59Fe,Gajconalbumin (A). p1 indicates species prepared by isoelectric fccusing.
albumin (82, 73, 64, and 587, iroll-saturated at each successive 30-min sampling period) was resuspended with fresh chick embryo red cells and the rate of iron transport again determined over a 40.min period.
The rates of iron donation from each of the iron-depleted protein samples were the same as the COP trol.
In view of recent spectroscopic data which indicate that the two iron-binding sites of conalbumin differ (14), a straightforward determination of the relative rates of iron release to the red cell from those sites is imperative.
The one-iron conalbumin species, Feiconalbumin,2 and various mixed metal-conalbumin species were assayed for their iron-donating ability, as summarized in Fig. 6. The rates of iron transport into the chick embryo red cell from the untreated control species, "gFezconalbumin, the same species isolated by isoelectric focusing, and "'Feiconalbumin to which 1 eq of "9Fe had been added were similar.
Showing half the transport rates of the doubly-labeled proteins were 5gFeiconalbumin, "6Fe,~sFeiconalbumi~l, Ga.5"Feiconalbumin, Cr,59Feiconalbumin, and "gFe,Gaiconalbumin. These results suggest that the chick embryo red blood cell removes whatever iron it finds on a conalbumin molecule and does not discriminate between the iron bound at the two different sites. Occupancy of the other site by a nonacquired metal ion does not affect the acquisition of the protein-bound iron by the cell.
2 The nomenclature for mixed metal and mixed isotope complexes of conalbumin is that previously described (14). h/letal ions specifically bound to the protein are prefixed to conalbumin with a positional subscript, namely Fe,conalbumin means monoferriconalbumin with the iron bound in the "inner" or first-added site, and Ga,Feiconalbumin means monogallicmonoferriconalbumin with iron bound in the inner or first-added site, and gallium bound in the "outer" or second-added site. The ultimate test of the relative rates of release of the two conalbumin-bound ferric ions to chick embryo cells was made by studying the disappearance of "gFe and 5"Fe from 5gFe055Feiconalbumin, produced by adding 1 eq of "gFe to 55Feiconalbumin isolated by isoelectric focusing. Fig. 7 shows the results of two experiments with freshly isolated and 6-month-old conalbumin preparations. Evidently the cells acquire iron at essentially equal rates from the two differentially labeled sites on diferric conalbumin.
Analysis of the final supernatant solution from these uptake studies by isoelectric focusing revealed that the solution contained approximately equal amounts of conalbumin, Feiconalbumin, and Fezconalbumin, and that the "9Fe and ""Fe labels were equally distributed in both the oneiron and two-iron conalbumin species. Similar results were obtained with control incubation mixtures containing 5gFe,""Feiconalbumin and conalbumin but no cells. These findings suggest that under the incubation conditions, or during the isoelcct,ric focusing procedure, the following exchange process occurs: The experiments performed to date do not enable us to decide at what stage this exchange process occurs, but at least they indicate that it does not require the presence of iron-acquiring cells.
It is clear from the foregoing data that the chick embryo red cell is able to accept simultaneously iron bound to the two sites on conalbumin.
This finding is at variance with the Flctcher-Huehns hypothesis (15,16). Whether this discrepancy lies in a difference between the physiological functioning of conalbumin and transferrin or in a difference between the iron-acquiring mechanisms of chick embryo red blood cells and rabbit reticulocytes cannot at present be decided.
In the iron-uptake experiments described thus far, the natural physiological anion, carbonate, has been involved.
The physiological effect of the anion obligately bound in the iron-conalbumin complex was investigated by studying the rates of iron uptake by chick embryo red cells from a series of Fesconal- bumin(anion)z complexes. These complexes have distinct, characteristic visible spectra (Fig. 8), and are stable to dialysis against Cot-free water.
The rates of iron uptake were similar for complexes in which the bound anion was carbonate, glycolate, glyoxalate, glycine, thioglycolate, or salicylate (Fig. 9), but the oxalato complex consistently gave a rate one-half that of the carbonato complex, as has been reported for the transferrinrabbit reticulocyte system (24).
The question then is whether bound osalate allows the selec-  Fig. 10. Each kinetic exchange curve can be resolved into three straight lines representing, in decreasing order (a) washout of excess, nonbound oxalate; (b) exchange of one oxalate for carbonate in the metallozconalbumin(oxalate)~ complex; and (c) exchange, or nonexchange, of the second oxalate in the complex.
In the absence of cells both metallozconalbumin(oxalate)s species exchanged one oxalate for carbonate and retained one oxalatc.
This result is akin to that found for the Feztransferrin([14C]osalate)2 complex (25). In the presence of cells, under aerobic conditions, the second oxalate was released at the same rate as iron uptake into the cells. The cells also catalyzed release of the second oxalate from the gallium complex, although in separate uptake experiments we found little or no incorporation of gallium into cells incubated with @Ga?conalbumin(carbonate)z.
Under anaerobic conditions, however, the cells failed to catalyze release of the second oxalate from either the iron or gallium complex and did not take up iron. Evidently anion release, in addition to iron uptake, is dependent on oxidative metabolism.
We conclude that the cell prob-ably effects release of the bound anion from the metalloconalbumin(anion) complex prior to release of the metal ion. iz protonation step for anion release is a likely possibility, as oxalate, the one inhibitory anion found so far, has significantly lower pK, values than the other anions investigated.
The anion thus may serve to lock in the bound metal ion, thereby protecting it from competing reactions such as hydrolysis.
Such a view is consistent with the finding that iron-transferrin complexes cannot exist in the absence of bound anion (26).
In summary we conclude that the chick embryo red cell does not discriminate between the two ferric ions bound to saturated iron conalbumin, in spite of the demonstrated physical difference between the binding sites, that the cell effects the release of the obligate anion prior to or simult'aneously with sequestering the bound iron, and that such anion release requires aerobic metabolism by the cell.