- Email: [email protected]
Mobilization of iron from specifically labeled reticulocyte ghosts
Mobilization of Iron from Specifically Labeled Reticulocyte Ghosts Erwin F. Workman, Jr., and George W. Bates Departmetzt of Biochemistry Texas A&M University
and Biophysics
and the Texas Agricitltural
Experiment
Station
ABSTRACT
SpecificaIly labeled 59Fe ghosts have been prepared by incubation of whole reticulocyte; with ~~Fe3+-transferrin~O32- followed by washing and ghost isolation. The binding of 59Fe by the membrane fraction is quite stable over a wide range of conditions, but Iron mobilization occurs on incubation with chelating agents or cell lysate. The time course of 59Fe mobilization by unlabeled reticulocyte lysate exhibits fme apparently zero-order phases. The rate of iron mobilization is linearly dependent on the concentration of 59Fe ghosts present ln the incubation mixture. In contrast, the relative concentration of lysate appears to exhibit a saturation dependence with regard to membrane iron mobilization. Bathophenanthroline sulfonate follows a multiphasic time course of iron mobilization similar to that found with the lysate. Lysate from mature erythrocytes was found to mobilize iron with kinetics that are identical to reticulocyte lysate. The number and duration of the phases is independent of the mobilizing agent. The role of the membrane fraction in regulating the rate of iron release to cytosol was also investigated by the repetitive incubation of 59Fe ghosts with fresh lysate. The rate of 59Fe mobilization depended on the condition of the ghost with regard to prior SsFe depletion. This publication emphasizes the active role of the membrane fraction in determining the rate at which iron will become available to the cytosoland the possibility that cytosol factors modulate the action of membrane bound components-
INTRODUCTION The molecular mechanisms and biological controls operative in the assimilation, transport, and utilization of iron by the developing red blood cell are topics of considerable interest and significance in the field of iron metabolism. Morgan and co-workers [l, 21 and other groups [3,4] have presented evidence that is consistent with the entry of molecule into the cell. It is assumed that transferrin the Fe3+-transferrin-C032donates iron to cytoplasmic components or subcellular organelles for utilization in hemoglobin and ferritin biosynthesis_ Another hypothesis describes a direct donation Address reprint requests to: George W. Bates, Texas Agricultural Experiment Station, Texas A&M University, College Station, Texas 77843. Journalof InorganicBiochemistry 10, 41-51
0 Elsevier North Holland, Inc., 1979
(1979)
41 0162-0134/79/010041+11$01.75
Erwin F. Workman,
Jr. and George W. Bates
of iron to a plasma membrane associated iron receptor site [S] . Consistent with this view is the work of Fielding and Speyer [6, 73, who have reported the isolation of transferrin and iron-binding components from the plasma membrane of reticulocytes. These investigators suggest that a transferrin-binding molecule, presumably at the surface of the membrane, extracts iron from transfenin and subsequently donates the iron by a series of steps to either the cytosol or a membrane-bound iron-storage component_ We have investigated iron release from the reticulocyte membrane fraction by preparing specifically labeled 5sFe ghosts from whole cells previously incubated with 59Fe3+-transferrinCOa2[8, 9]_ Various media, including hemolysate, ATP, and chelating agents, are able to moclilize iron from the cell stroma. The form of the solubilized iron can be analyzed by fractionation with gel-filtration media. Subsequent to the completion of the work described herein a related study was reported by Blackbum and Morgan IlO] . It focused on the cytoplasmic factors involved in iron mobilization and transport_ The two articles are complementary since they point out that both the membrane fraction and specific cytoplasmic factors play active roles in providing and controlling the flow of iron for hemoglobin synthesis_
MATERIALS
AND METHODS
Transferrin Transferrin was prepared from fresh Pel Freez rabbit serum using ammonium sulfate fractionation and chromatographic techniques [3, 11, 12]_ Transferrin was saturated with iron using published methods [13] _ Iodination was carried out according to the method of Katz [ 14,15]_
Reticulocytes Reticulocytes were obtained from New Zealand white rabbits rendered reticulocytotic by phlebotomy; 40 ml were obtained from the marginal ear veins on alternate days for at least 10 days. Heparinized whole blood was obtained immediately prior to use. The plasma and buffy coat were removed following centrifugation and the cells washed twice in standard buffer (5 mM Tris, 0.15 M NaCl, and 0.01 M glucose, pH = 7.45). Reticulocytes were enumerated after staining with brilliant cresyl blue.
Reagents and isotopes All chemicals were of reagent grade or better and not further purified_ Water was glass distilled and deionized_ Iron-59. 5-l 0 mCi/mg Fe as FeCl a in 0.5 M HCl, and 1z51, carrier free, in 0-I M NaOH were obtained from New England Nuclear. Radioactivities were measured in a Nuclear Chicago gamma well-scintillation counter_
5gFe-LabeIed
ReticuIocyte
Ghosts
Specifically labeled 5gFe ghosts were prepared as described earlier [8], with the exception that the whole cells were allowed to react for 45 min with the 5sFe3+-
43
Membrane Iron Mobilization
transferrin-COa2--. Control experiments with 59Fe-[1251] iodotransferrinC032indicate that less than 15% of the 5gFe of the 5gFe ghosts was associated with transfenin.
UnlabeledReticulocyteLysate Appropriate volumes (see Results) of packed whole cells and 20 mOsM Tris were mixed to achieve lysis_ The mixture was then centrifuged at 43,000 X g for 20 min and the Iysate removed. The lysate was prepared just prior to use and was made 0.15 M in NaCl and 0.01 M in glucose.
Modilization Experiments Figure 1 is a flow diagram depicting the sequence of steps involved in the mobilization studies. Aliquots of 59Fe ghosts prepared from a known volume of packed whole cells and containing a known amount of 5gFe were placed in polycarbonate reaction vessels
and aIIowed to come to 37”. Preincubated reagents or lysate were added to the ghost suspension at time zero. The final volume was 10 ml. At various times, llml aliquots of the incubation mixture were removed and rapidly cooled to 4*. The extent of mobilization was subsequently determined by separating the ghosts and soluble fraction by centrifugation at 43,000 X g for 10 min and analyzing the radioactivity of the supernatant fraction. Resolution of the various iron binding components of the soluble fraction was carried out using Sephadex G-100 as previously described [SJ _ Chromatographic fractions were monitored at 280 run using a Cary 15 spectrophotometer.
FIGURE 1. Row diagram depicting sequence of steps involved in mobilization studies. Experimental details are given in the text
WASHED
CELLS
,
1 1 INCUBATION
1 CHROMATOGRAPH
MIXTURE
xNTS(q
Erwin F. Workman,
44
Jr. and George W. Bates
RESULTS Mobilization
as a Function
of SoIution
Conditions
An important point regarding the feasibility of mobilizaticn studies is the stability of the 59Fe associated with the specifically labeled ghosts. We have examined 59Fe ghosts under a variety of chemical and physical conditions_ Figure 2 illustrates the release of sgFe and [ 1251] iodotransferrin as a function of several variables. In panel A it is seen that Ca*+ and Mg*+. which play a role in membrane integrity, have little effect on the mobilization process. Panel B depicts the effect of the NaCl. At low ionic strengtlrc there is an increase in the amount of iron and [1251] iodotransferrin mobilized during the 45min incubation period_ Some protein release from erythrocyte membranes at ionic strength extremes is known to occur. It appears that those components binding the sgFe are not greatly affected_ In panel C the effect of temperature on iron and [ 12sI] iodotransferrin release is shown. Although release of the [l *sI] iodotransferrin from the membrane is significantly enhanced at higher temperatures. the stability of iron binding is unaffected_ The effect of pH on iron mobilization by buffer is exhibited in Fig_ 3. A marked enhancement of iron release from the ghost is observed at pH extremes_ A less pronounced increase of [ 12511 iodotransferrin release at the pH extremes is noted. We do not know whether the iron release at pH extremes represents release of iron binding membrane component:, or simply the release of iron to the solution_
Kinetics
of Iron Mobilization Figure 4 shows the time course of iron mobilization from the 5gFe ghosts by unlabeled lysate in which 26 points were obtained at times ranging from 15 see to 180 min. There is a rapid initial mobilization_ phase 1. occuring within the first 15 set after addition of lysate_ This is followed by four linear mobilization phases. the last of which is linear at 32-150 min. as shown in the figure inset (Fig. 4). A segmented linear regression model involving an intercept and four slope parameters corresponding to the intervals (0.25, 5)_ (5, I?), (12, 32). and (32, 180) was fitted by multiple regression analysis_ This model accounted for 99.94% of the variation in 59Fe mobilization and involved a residual error rms of 27. Thus the reaction mathematically fits a multiphasic zero-order kinetics model Iron mobilization from the 5gFe ghosts was calculated with reference to the number of whole cells used in the ghost preparation and the specific activity of the 5gFetransferrin used in the labeling- Phase 1 was found to have a minimal rate of 6 X 105 Fe nlin-l cell-l. This rate is twice as fast as the calculated cytoplasmic iron uptake by whole cells, 3 X IO5 Fe min-l cell-l [ 15]_ Phases 3,3,4, and 5 of Fig. 4 exhibited rates of 3.7, 1.9, 0.8, and 0.3 X 10” Fe min-l cell-l. respectively. It would appear that only the phase-l rate is sufficient to be characteristic of turnover rates of iron transport components under physiological conditions.
Mobilization
as a Function
of “SFe-Ghost
Concentration
An experiment was carried out in which a constant amount of lysate was present in each of six incubation mixtures containing variable amounts of 59Fe ghosts_ The ghost component varied over a 40-fold range. Curves similar in shape to that in Fig_ 4 were
Membrane
Iron
45
MobiIization
IO
-.__~. .__ c
A
:NaC<, M IO
C _.o-__--
5 t
0'
__-- -0
-
.
-
________-o--------
;_. o________d ------- -a=.--. lo203040506070 TEMPERATURE
:C
FIGURE 2. Mobilization of 59Fe and [I2511 iodotransferrin during 45-min incubations at 37” in the presence of standard buffer, modified as depicted in the diagraa. Standard buffer consists of 0.15 M NaCl, 5 mM Tris, and 10 mM glucose, pH = 7.45. In panel A the standard buffer was supplemented with various concentrations of Ca2* and Mg 2+_ Panel B depicts the effect of varying the final NaCl concentration. Panel C indicates the effect of temperature. In all panels, closed circles indicate 59Fe and open circles, [1251] -iodotransferrin. The radioactivities of these components per sample were 15,000 cpm and 7000 cpm. respectively.
FIGURE 3. Mobilization of 59Fe and [1251] iodotransferrin during 45min incubation at 37” in standard buffer, varied with regard to pH. Closed circles, 59Fe; open circles, [==I ] -iodotransferr-in. Other conditions were as described in Fig. 2.
PH
Erwin F. Workman, Jr. and George W. Bates
46
I
I lo
20so405060708090
TIME,min FIGURE 4. Time course of mobiIization of 59Fe from 5sFe ghosts by unlabeled reticulocyte Iysate. First, 12 mI of 59Fe-ghost preparation and 18 ml of reticulocyte lysate were brought to 37” and then rapidiy mixed, Iron-59 activity of the incubation mixture was 2450 cpm ml-l. The Iysate was isotonic and at a 5.6 dilution factor relative to whole cell cytoplasm. At timed intervals, I mi of incubation mixture was removed with an Eppendorf pipette and added to 0.5 ml of standard buffer in an ice bath at @. The material was centrifuged at 43,000 X g for 10 min and 1 ml of the supematant removed for counting. As a control, the mobilization of 59Fe by standardbuffer was monitored at various times. The insert illustratesthe reaction at 30-180 min. Kinetic phases were analyzed mathematically as described in the text_
generated, with the exception that since it was technically impractical to obtain as many data points, phase 4 was not resolved_ Figure 5 shows the amount of 5gFe mobilized in phase 1 (open circles) and in the total 60-min incubation (closed circles), as a function of the concentration of 59Fe ghost- Both values are linearly dependent on the amount of 5gFe ghosts present. The rate and extent of all the phases were found to be linearly dependent on 5sFe-ghost concentration_ When log (rate) was plotted as a function of log (59Fe-ghost concentration), a slope between 0.8 and 1 was obtained for the phases, indicating a first-order dependence on 59Fe-ghost concentration_ Mobilization as a Function of Lysate Concentration Figure 6 depicts the time course of mobilization of iron from 5gFe ghosts by lysate at five concentrations_ An effect of the relative lysate concentration on the mobilization process is seen. To examine this effect, we plotted the rate of iron mobilization in each of the three resolved phases as a function of Iysate concentration. The rates of phases 2 and 3 depend in a hyperbolic fashion on the lysate concentration. Factors in the lysate are apparently able to enhance the rate of these phases; however, as the Iysate concentration nears the physiological range the effect reaches a plateau level. Ln
Membrane
47
Iron Mobilization
cpm
5sFe-GHOST,
FIGURE 5. Extent of sgFe mobilized from 5oFe ghosts by unlabeled lysate during phase 1 fopen circle, dashed line) and the entire 6fkmin incubation period (solid circles, solid line) as a function of SsFe-ghost concentration in incubation mixture. Aliquots of a 59Fe-ghost preparation ranging within 0.14 ml were added to sufficient standard buffer to give a volume of 8.0 ml_ These mixtures were brought to temperature, and at to 2.0 ml of preequilibrated isotonic lysate were added, Lysate concentration in the equilibrium mixture was 0.05 relative to whole-cell cytoplasm. Time course of g9Fe mobilization was determined as described in the text. contrast,
phase 5 is independent of lysate concentration, as can be observed directly in
Fig. 6. Repeat incubations The experiments described in Figs. 4-6 suggest a mobilization system in w-hich membrane-associated factors determine the rate at which iron becomes available to the cytosol. The cytosol, in turn, appears to have some role in modifying the activity of the membrane fraction_ To determine whether the multiphasic nature of the mobilizaFIGURE 6. Family of mobilization patterns obtained during incubation of constant amount of g9Fe ghosts with variable concentrations of unlabeled lysate. Curve designations refer to concentration of lysate relative to that of physiological cytoplasm. Incubations were carried out at 370 in a total volume of 10 ml, which included 2 ml of 59Fe-ghost preparation containing 11,000 cpm. Experimental details are presented in other legends and in the text.
‘0
I
1
tb
2b
&
40
I
I
50
60
I
Erwin F. Workman,
48
Oo L
IO
20
xl
40
50
60
0
102030405060010
Jr. and George W. Bates
20
30
40
50
60
TIME. min.
FIGURE 7. Repeat incubation eqxx-iment in which effect of fresh lysate on previously treated 59Fe ghosts is tested. Iron-59 ghosts were incubated one, two, and three times for panels labeled Iysate I, Iysate 2, and Iysate 3, respectiveIy. The 2-hr control and buffer control indicate that there was not significant deterioration of the lysate or the SgFe-ghost fraction over the duration of the experiment. Detailed experimental procedures are described in the Materials and Methods and the Results sections.
tion phenomenon
migiit in some way depend on the saturation
devised a repeat incubation experiment_ Four incubation mixtures were prepared containing incubation mixtures 1, 2, and 3 were added sufficient
of lysate factors,
we
4 ml of radioactive ghosts. To lysate to give a final volume of
10 ml and a lysate concentration of 0.125 relative to physiological cytosol. Incubation mi..ture 4 was taken to 10 ml in standard buffer_ The latter mixture.was sampled at various &es over the next 4 hr. Mixtures 1. 2, and 3 were incubated for 60 min. During the first 60-min period aliquots were taken only from incubation mixture l_ At the end of 60 min, the remainder of incubation mixture 1 was discarded, incubation mixtures 2 and 3 centrifuged, the 59Fe-ghost fraction pelleted and saved, and the amount of mobilized iron determined (see arrows in Fig_ 7). By adding 15 ml of standard buffer at 4O to the incubation mixtures we could stop the mobilization process and effect a degree of washing. The 5gFe-ghost fractions from incubations 2 and 3 were taken up in 3 ml of standard buffer at 37” and allowed to equilibrate to temperature. At ro, 5 ml of Iysate was again added to each of the two preparations. During the next 60-min period, aliquots were only taken from incubation mixture 2. At the end of 60 min, the washing procedure was repeated,after which a kinetic study of mobilization from incubation mixture 3 was performed. Figure 7 shows the results of this experiment. It will be seen that at to incubation mixture 1 shows the characteristic rapid multiphasic mobilization of iron. When fresh lysate is added to the preincubated ghosts of incubation mixture 2, the bulk of the mobilization continues at the slow rate that was observed when the prior incubation was terminated. When the fresh lysate is added--to the 59Fe ghosts of incubation mixture 3, we again see the slow rate of the fifth kinetic phase. It appears that the rates
are governed
by the state of the membrane
and not the state of the lysate.
Membrane Iron Mobilization
49
25
0
0
10
20
30 TIME,
40
50
60
min.
FIGURE 8. Mobilization of 59Fe from 59Fe ghosts by three concentrations of bathophenanthroline sulfonate. Molar concentrations of chelate are as shown in curve de&nations. Bathophv throline sulfonate was dissolved in standard buffer containing, in addition, 1 mM Caz+ and 1 mM Mg2+_ Mobilization experiments were carried out in a fashion identical to that descrii above in which unlabeled lysate was used as the mobilizing agent. The buffer-only control is also depicted.
The question arks as to whether the lysate pool used in the repeat incubations could have deteriorated during the l-2 hr that it was incubated prior to use. On the right-hand section of the graph-is shown a 241 control in which both Iysate and 59Fe ghosts were maintained at 37” for 2 hr prior to utilization in a mobilization experiment_ It wih be seen that the overall mobilization time course of the 2-hr control is very close to that of the results observed with mixture 1. Mobilization by Bathophenanthroline Sulfonate The time course of the mobilization of 5aFe by bathophenanthroline sulfonate was examined in a manner parallel with the lysate-ghost incubation experiments. The media included 1 mM Mg2+ and 1 mM Ca2+. Figure 8 shows the time course of iron mobilization by three concentrations of bathophenanthroline sulfonate as indicated by the figure labels. Again we observe multiphaslc iron mobilization from the specifically labeled 59Fe ghosts. Phase 1 contributes 4.5-9.5% of the total iron. This is a quite narrow range when one considers that the chelate concentration varies over a lOCkfold range. The extents and rates of the following phases are independent of chelate concentration. The parallelism in these curves suggests that chelation is not rate limiting between 10-a M and 10s6 M. It appears that the membrane directs the rate at which iron becomes available for mobilization. Comparison of the mobilization of 5aFe by bathophenanthroline sulfonate with that of iysate reveals that the rates and extents of mobilization are greater in the latter case (with the exception of phase 5). Lysate may influence the rates of the phases rather than their number or duration. The rate of the final’ phase (observedat times >30 min) is invariant under a wide range of lysate and bathophenanthroline sulfonate concentrations, distinguishing it kinetically from the other phases.
Erwin F. Workman, Jr. and George W. Bates
50
Mobilization
by kythrocyte
Lysate
Lysate factors appear to exert some influence on membrane activity_ We thus investigated mobilization by mature red-blood-cell lysate. Rabbits not previously bled were utilized as erythrocyte sources. Heparinized samples contained less than 1% reticulocytes and were shown to be unable to take up iron from 5eFe-transferrinC032-. On the other hand, the reticulocyte preparation was found to be quite active. Reticulocytes were used for preparing 5sFe ghosts in the usual manner. The ghosts were then incubated with either unlabeled reticulocyte lysate or erythrocyte lysate. The tire course of mobilization by both lysate preparations was parallel and similar to the top curve of Fig. 6. The reticulocyte lysate mobilized 52.4% of the iron in 60 min, whereas the erythrocyte lysate mobilized 493% in the same time. The results are in excellent agreement with those of Blackbum and Morgan, who found 50.5% and 47.5%~~respectively [IO] _ We examined the fate of the mobilized iron in both lysate preparations_ A striking parallel was found in the Sephadex G-100 profdes. The lysate of the mature red blood cell incorporated iron into the fractions containing fenitin, hemoglobin (as confirmed by heme extraction), and the 6000-molecular-weight iron-binding component [S] _The time course of the appearance of radioactive heme in both incubation supematants was also examined. Over the first 15-min period, lysate from reticulocytes and mature erytthrocytes mobilized heme at approximately the same rate; however, at longer tunes the appearance of heme in the mature erythrocyte Iysate slowed markedly. After 60 mm the mature erythrocyte had incorporated 62% as much iron into heme as had the reticuiocyte lysate. These results do not necessarily indicate heme synthesis, since Blackbum and Morgan [IO] reported that much preformed heme is present in the membrane fraction.
DISCUSSION The experirrents reported herein point to the stability of the SeFe-ghost preparation, and along with the recent work of Blackbum and Morgan [lo], emphasize the potential of the lysate-ghost incubation system for probing reticulocyte iron metabolism_ The central point of this article is the evidence that the membrane fraction is not a passive, homogeneous iron depot, but an active component that makes iron available for mobilization at regulated rates. The rate at which iron is mobilized is dependent on the state of repletion of the membrane, as shown in Fig. 7. From this and previous studies [8-lo], we can infer that iron transport in the whole reticulocyte depends on both membrane-bound and cytosol factors. It is tempting to speculate that the multiphasic mobilization of iron from the specifically labeled 59Fe ghost is a manifestation of a multicomponent membrane iron trammort and storage system. Consistent with this interpretation is the work of Fielding and Speyer [6,7] , who used a variety of labeling, solubilization, chromatographic, and incubation procedures to obtain evidence for the presence of at least tiree membraneassociated iron-binding components_ An Fes+-transferrin-CO,*receptor site appears to have the role of interacting with the transferrin and perhaps removing the iron. The iron appears to move to a second component, which may have a relationship to ferritin. The subsequent pathway appears to have a branch point allowing iron to be stored in a membrane-associated component or transferred to cytoplasmic iron-transport agents.
51
Membrane Iron Mobilization
These studies also suggest that the membrane-associated components pky a dominant role in determining the rate at which iron becomes available to the cytosol. Corroborating evidence on the importance of the plasma membrane of the developing erythrocyte is provided by other investigators. Tanaka and Brccher [16] have suggested that fenitin plays an important part in iron exchange at the plasma membrane. Evidence was presented for the synthesis of apoferxitin at or near the ceil surface under conditions of iron overload. Denton et al. [ 171 have provided evidence that the cell stroma is a major iron-storage compartment during the erythroblastic stages. Iron release from the stromal fraction coincides with an increase in hemoglobin biosynthesis during the cell maturation process. The reticulocyte contains comparatively little stromal iron and is in what appears to be the final stages of stromal iron depletion. The plasma membrane of the developing erytbrocyte may play an important role in celhdar iron metabolism, not only for obtaining iron from transferrin, but also for storing iron and regulating its availability to the cytoplasm during cellular maturation. Blackbum and Morgan [lo] have pointed to the importance of heme as a major portion of the iron being mobilized from the stroma fraction, which includes a portion of the mitochondria of the cell. The multiphasic nature of mobilization could be accounted for in terms of heme and nonheme iron complexes being solubilized at varying rates. The fact that chelating agents arc not as effective as lysate in mobilizing iron may be a manifestation of their inability to complex heme. At this point it se&s safe to say that reticulocyte iron metabolism must involve an orchestrated interaction of membrane-bound and cytosol fractions, sub&h&u organelles, transfer&, and perhaps other serum factors.
This investigationwas supported by National Institutes of Health Research Cmnr No_ I ROI AM I 7790 from the National Institute of Arthritis. Metabolism. and Digestive D-es and represents partial fut”Iiment of the requirementsfor a doctoral degree by Edwin F. Workman,Jr_
REFERENCES 1. E. H. Morganand T. C. Appleton,Natlue223,1371-1372 (1969). 2. E. H. Morganand E. Baker,Biochim. Biophys Acta l&4,442454 (1969). 3. J. Martinez-MedeUiu and H_M. Schulman,Biochim Biophys Acta 264,272-284 (1972). 4. J. Borova,P. Ponka, and J. Neuwirt, B&him. Biophys- Acta 320.143-156 (1973). 5. 67. 8. 9.
10. 11. 12. 13. 14.
J. H. Jandl, J. K. Irunan. R. L.. Simmons, and D. W. Allen, J. C7k Invesx 38,161-X35 (1959). B. IL Speyer and J. Fielding,Biochim Biophys Actn 332,192-200 (1974). J. Fieldingand B. E. Speyer,B&him. Biosphys. Acta 363,387-396 (1974). E. F. Workmanand G. W Bates,B&hem Biophys Res. Cbmmun. 58,787-794 (1974). E. F. Workmanand G. W. Bates, in R. R. Chrichton,Ed.,&teinsoflron Storageand lkansport in Biochemistry and Medicine. North Holland,Amsterdam.1975, pp. 155-160. G. W. Blackbumand E. H. Morgan, Biochim Biophys Acta 497,728-744 (1977).
E. H. M0rgan.J. PIzysioLLand. 171,2641(1964). H. G. Van Eyk and B. Leijnse,Biochim Biophyr AC&~ 160,126-128 (1968). G. W. Batesand M. R. Schlabach,J. BioL them 248,3228-3232 (1973). J. H. Katz,L C7in.Invest. 40,2143-2152 (1961). 15. E. F. Workman,G. Graham,and G. W. Bates,B&him. Biophys. Acra 399,254-264 (1975). 16. Y. Tanakaand G. Brecher, Blood 37,211-219 (1971). 17.
M. J. Denton, H. T_ Delves, and H_ R_ V_ Amstein.Biochem
13 (1974). Received I 9 Apnl I9 78
Bkphys
Res
Commun.
61, S-
and Biophysics
and the Texas Agricitltural
Experiment
Station
ABSTRACT
SpecificaIly labeled 59Fe ghosts have been prepared by incubation of whole reticulocyte; with ~~Fe3+-transferrin~O32- followed by washing and ghost isolation. The binding of 59Fe by the membrane fraction is quite stable over a wide range of conditions, but Iron mobilization occurs on incubation with chelating agents or cell lysate. The time course of 59Fe mobilization by unlabeled reticulocyte lysate exhibits fme apparently zero-order phases. The rate of iron mobilization is linearly dependent on the concentration of 59Fe ghosts present ln the incubation mixture. In contrast, the relative concentration of lysate appears to exhibit a saturation dependence with regard to membrane iron mobilization. Bathophenanthroline sulfonate follows a multiphasic time course of iron mobilization similar to that found with the lysate. Lysate from mature erythrocytes was found to mobilize iron with kinetics that are identical to reticulocyte lysate. The number and duration of the phases is independent of the mobilizing agent. The role of the membrane fraction in regulating the rate of iron release to cytosol was also investigated by the repetitive incubation of 59Fe ghosts with fresh lysate. The rate of 59Fe mobilization depended on the condition of the ghost with regard to prior SsFe depletion. This publication emphasizes the active role of the membrane fraction in determining the rate at which iron will become available to the cytosoland the possibility that cytosol factors modulate the action of membrane bound components-
INTRODUCTION The molecular mechanisms and biological controls operative in the assimilation, transport, and utilization of iron by the developing red blood cell are topics of considerable interest and significance in the field of iron metabolism. Morgan and co-workers [l, 21 and other groups [3,4] have presented evidence that is consistent with the entry of molecule into the cell. It is assumed that transferrin the Fe3+-transferrin-C032donates iron to cytoplasmic components or subcellular organelles for utilization in hemoglobin and ferritin biosynthesis_ Another hypothesis describes a direct donation Address reprint requests to: George W. Bates, Texas Agricultural Experiment Station, Texas A&M University, College Station, Texas 77843. Journalof InorganicBiochemistry 10, 41-51
0 Elsevier North Holland, Inc., 1979
(1979)
41 0162-0134/79/010041+11$01.75
Erwin F. Workman,
Jr. and George W. Bates
of iron to a plasma membrane associated iron receptor site [S] . Consistent with this view is the work of Fielding and Speyer [6, 73, who have reported the isolation of transferrin and iron-binding components from the plasma membrane of reticulocytes. These investigators suggest that a transferrin-binding molecule, presumably at the surface of the membrane, extracts iron from transfenin and subsequently donates the iron by a series of steps to either the cytosol or a membrane-bound iron-storage component_ We have investigated iron release from the reticulocyte membrane fraction by preparing specifically labeled 5sFe ghosts from whole cells previously incubated with 59Fe3+-transferrinCOa2[8, 9]_ Various media, including hemolysate, ATP, and chelating agents, are able to moclilize iron from the cell stroma. The form of the solubilized iron can be analyzed by fractionation with gel-filtration media. Subsequent to the completion of the work described herein a related study was reported by Blackbum and Morgan IlO] . It focused on the cytoplasmic factors involved in iron mobilization and transport_ The two articles are complementary since they point out that both the membrane fraction and specific cytoplasmic factors play active roles in providing and controlling the flow of iron for hemoglobin synthesis_
MATERIALS
AND METHODS
Transferrin Transferrin was prepared from fresh Pel Freez rabbit serum using ammonium sulfate fractionation and chromatographic techniques [3, 11, 12]_ Transferrin was saturated with iron using published methods [13] _ Iodination was carried out according to the method of Katz [ 14,15]_
Reticulocytes Reticulocytes were obtained from New Zealand white rabbits rendered reticulocytotic by phlebotomy; 40 ml were obtained from the marginal ear veins on alternate days for at least 10 days. Heparinized whole blood was obtained immediately prior to use. The plasma and buffy coat were removed following centrifugation and the cells washed twice in standard buffer (5 mM Tris, 0.15 M NaCl, and 0.01 M glucose, pH = 7.45). Reticulocytes were enumerated after staining with brilliant cresyl blue.
Reagents and isotopes All chemicals were of reagent grade or better and not further purified_ Water was glass distilled and deionized_ Iron-59. 5-l 0 mCi/mg Fe as FeCl a in 0.5 M HCl, and 1z51, carrier free, in 0-I M NaOH were obtained from New England Nuclear. Radioactivities were measured in a Nuclear Chicago gamma well-scintillation counter_
5gFe-LabeIed
ReticuIocyte
Ghosts
Specifically labeled 5gFe ghosts were prepared as described earlier [8], with the exception that the whole cells were allowed to react for 45 min with the 5sFe3+-
43
Membrane Iron Mobilization
transferrin-COa2--. Control experiments with 59Fe-[1251] iodotransferrinC032indicate that less than 15% of the 5gFe of the 5gFe ghosts was associated with transfenin.
UnlabeledReticulocyteLysate Appropriate volumes (see Results) of packed whole cells and 20 mOsM Tris were mixed to achieve lysis_ The mixture was then centrifuged at 43,000 X g for 20 min and the Iysate removed. The lysate was prepared just prior to use and was made 0.15 M in NaCl and 0.01 M in glucose.
Modilization Experiments Figure 1 is a flow diagram depicting the sequence of steps involved in the mobilization studies. Aliquots of 59Fe ghosts prepared from a known volume of packed whole cells and containing a known amount of 5gFe were placed in polycarbonate reaction vessels
and aIIowed to come to 37”. Preincubated reagents or lysate were added to the ghost suspension at time zero. The final volume was 10 ml. At various times, llml aliquots of the incubation mixture were removed and rapidly cooled to 4*. The extent of mobilization was subsequently determined by separating the ghosts and soluble fraction by centrifugation at 43,000 X g for 10 min and analyzing the radioactivity of the supernatant fraction. Resolution of the various iron binding components of the soluble fraction was carried out using Sephadex G-100 as previously described [SJ _ Chromatographic fractions were monitored at 280 run using a Cary 15 spectrophotometer.
FIGURE 1. Row diagram depicting sequence of steps involved in mobilization studies. Experimental details are given in the text
WASHED
CELLS
,
1 1 INCUBATION
1 CHROMATOGRAPH
MIXTURE
xNTS(q
Erwin F. Workman,
44
Jr. and George W. Bates
RESULTS Mobilization
as a Function
of SoIution
Conditions
An important point regarding the feasibility of mobilizaticn studies is the stability of the 59Fe associated with the specifically labeled ghosts. We have examined 59Fe ghosts under a variety of chemical and physical conditions_ Figure 2 illustrates the release of sgFe and [ 1251] iodotransferrin as a function of several variables. In panel A it is seen that Ca*+ and Mg*+. which play a role in membrane integrity, have little effect on the mobilization process. Panel B depicts the effect of the NaCl. At low ionic strengtlrc there is an increase in the amount of iron and [1251] iodotransferrin mobilized during the 45min incubation period_ Some protein release from erythrocyte membranes at ionic strength extremes is known to occur. It appears that those components binding the sgFe are not greatly affected_ In panel C the effect of temperature on iron and [ 12sI] iodotransferrin release is shown. Although release of the [l *sI] iodotransferrin from the membrane is significantly enhanced at higher temperatures. the stability of iron binding is unaffected_ The effect of pH on iron mobilization by buffer is exhibited in Fig_ 3. A marked enhancement of iron release from the ghost is observed at pH extremes_ A less pronounced increase of [ 12511 iodotransferrin release at the pH extremes is noted. We do not know whether the iron release at pH extremes represents release of iron binding membrane component:, or simply the release of iron to the solution_
Kinetics
of Iron Mobilization Figure 4 shows the time course of iron mobilization from the 5gFe ghosts by unlabeled lysate in which 26 points were obtained at times ranging from 15 see to 180 min. There is a rapid initial mobilization_ phase 1. occuring within the first 15 set after addition of lysate_ This is followed by four linear mobilization phases. the last of which is linear at 32-150 min. as shown in the figure inset (Fig. 4). A segmented linear regression model involving an intercept and four slope parameters corresponding to the intervals (0.25, 5)_ (5, I?), (12, 32). and (32, 180) was fitted by multiple regression analysis_ This model accounted for 99.94% of the variation in 59Fe mobilization and involved a residual error rms of 27. Thus the reaction mathematically fits a multiphasic zero-order kinetics model Iron mobilization from the 5gFe ghosts was calculated with reference to the number of whole cells used in the ghost preparation and the specific activity of the 5gFetransferrin used in the labeling- Phase 1 was found to have a minimal rate of 6 X 105 Fe nlin-l cell-l. This rate is twice as fast as the calculated cytoplasmic iron uptake by whole cells, 3 X IO5 Fe min-l cell-l [ 15]_ Phases 3,3,4, and 5 of Fig. 4 exhibited rates of 3.7, 1.9, 0.8, and 0.3 X 10” Fe min-l cell-l. respectively. It would appear that only the phase-l rate is sufficient to be characteristic of turnover rates of iron transport components under physiological conditions.
Mobilization
as a Function
of “SFe-Ghost
Concentration
An experiment was carried out in which a constant amount of lysate was present in each of six incubation mixtures containing variable amounts of 59Fe ghosts_ The ghost component varied over a 40-fold range. Curves similar in shape to that in Fig_ 4 were
Membrane
Iron
45
MobiIization
IO
-.__~. .__ c
A
:NaC<, M IO
C _.o-__--
5 t
0'
__-- -0
-
.
-
________-o--------
;_. o________d ------- -a=.--. lo203040506070 TEMPERATURE
:C
FIGURE 2. Mobilization of 59Fe and [I2511 iodotransferrin during 45-min incubations at 37” in the presence of standard buffer, modified as depicted in the diagraa. Standard buffer consists of 0.15 M NaCl, 5 mM Tris, and 10 mM glucose, pH = 7.45. In panel A the standard buffer was supplemented with various concentrations of Ca2* and Mg 2+_ Panel B depicts the effect of varying the final NaCl concentration. Panel C indicates the effect of temperature. In all panels, closed circles indicate 59Fe and open circles, [1251] -iodotransferrin. The radioactivities of these components per sample were 15,000 cpm and 7000 cpm. respectively.
FIGURE 3. Mobilization of 59Fe and [1251] iodotransferrin during 45min incubation at 37” in standard buffer, varied with regard to pH. Closed circles, 59Fe; open circles, [==I ] -iodotransferr-in. Other conditions were as described in Fig. 2.
PH
Erwin F. Workman, Jr. and George W. Bates
46
I
I lo
20so405060708090
TIME,min FIGURE 4. Time course of mobiIization of 59Fe from 5sFe ghosts by unlabeled reticulocyte Iysate. First, 12 mI of 59Fe-ghost preparation and 18 ml of reticulocyte lysate were brought to 37” and then rapidiy mixed, Iron-59 activity of the incubation mixture was 2450 cpm ml-l. The Iysate was isotonic and at a 5.6 dilution factor relative to whole cell cytoplasm. At timed intervals, I mi of incubation mixture was removed with an Eppendorf pipette and added to 0.5 ml of standard buffer in an ice bath at @. The material was centrifuged at 43,000 X g for 10 min and 1 ml of the supematant removed for counting. As a control, the mobilization of 59Fe by standardbuffer was monitored at various times. The insert illustratesthe reaction at 30-180 min. Kinetic phases were analyzed mathematically as described in the text_
generated, with the exception that since it was technically impractical to obtain as many data points, phase 4 was not resolved_ Figure 5 shows the amount of 5gFe mobilized in phase 1 (open circles) and in the total 60-min incubation (closed circles), as a function of the concentration of 59Fe ghost- Both values are linearly dependent on the amount of 5gFe ghosts present. The rate and extent of all the phases were found to be linearly dependent on 5sFe-ghost concentration_ When log (rate) was plotted as a function of log (59Fe-ghost concentration), a slope between 0.8 and 1 was obtained for the phases, indicating a first-order dependence on 59Fe-ghost concentration_ Mobilization as a Function of Lysate Concentration Figure 6 depicts the time course of mobilization of iron from 5gFe ghosts by lysate at five concentrations_ An effect of the relative lysate concentration on the mobilization process is seen. To examine this effect, we plotted the rate of iron mobilization in each of the three resolved phases as a function of Iysate concentration. The rates of phases 2 and 3 depend in a hyperbolic fashion on the lysate concentration. Factors in the lysate are apparently able to enhance the rate of these phases; however, as the Iysate concentration nears the physiological range the effect reaches a plateau level. Ln
Membrane
47
Iron Mobilization
cpm
5sFe-GHOST,
FIGURE 5. Extent of sgFe mobilized from 5oFe ghosts by unlabeled lysate during phase 1 fopen circle, dashed line) and the entire 6fkmin incubation period (solid circles, solid line) as a function of SsFe-ghost concentration in incubation mixture. Aliquots of a 59Fe-ghost preparation ranging within 0.14 ml were added to sufficient standard buffer to give a volume of 8.0 ml_ These mixtures were brought to temperature, and at to 2.0 ml of preequilibrated isotonic lysate were added, Lysate concentration in the equilibrium mixture was 0.05 relative to whole-cell cytoplasm. Time course of g9Fe mobilization was determined as described in the text. contrast,
phase 5 is independent of lysate concentration, as can be observed directly in
Fig. 6. Repeat incubations The experiments described in Figs. 4-6 suggest a mobilization system in w-hich membrane-associated factors determine the rate at which iron becomes available to the cytosol. The cytosol, in turn, appears to have some role in modifying the activity of the membrane fraction_ To determine whether the multiphasic nature of the mobilizaFIGURE 6. Family of mobilization patterns obtained during incubation of constant amount of g9Fe ghosts with variable concentrations of unlabeled lysate. Curve designations refer to concentration of lysate relative to that of physiological cytoplasm. Incubations were carried out at 370 in a total volume of 10 ml, which included 2 ml of 59Fe-ghost preparation containing 11,000 cpm. Experimental details are presented in other legends and in the text.
‘0
I
1
tb
2b
&
40
I
I
50
60
I
Erwin F. Workman,
48
Oo L
IO
20
xl
40
50
60
0
102030405060010
Jr. and George W. Bates
20
30
40
50
60
TIME. min.
FIGURE 7. Repeat incubation eqxx-iment in which effect of fresh lysate on previously treated 59Fe ghosts is tested. Iron-59 ghosts were incubated one, two, and three times for panels labeled Iysate I, Iysate 2, and Iysate 3, respectiveIy. The 2-hr control and buffer control indicate that there was not significant deterioration of the lysate or the SgFe-ghost fraction over the duration of the experiment. Detailed experimental procedures are described in the Materials and Methods and the Results sections.
tion phenomenon
migiit in some way depend on the saturation
devised a repeat incubation experiment_ Four incubation mixtures were prepared containing incubation mixtures 1, 2, and 3 were added sufficient
of lysate factors,
we
4 ml of radioactive ghosts. To lysate to give a final volume of
10 ml and a lysate concentration of 0.125 relative to physiological cytosol. Incubation mi..ture 4 was taken to 10 ml in standard buffer_ The latter mixture.was sampled at various &es over the next 4 hr. Mixtures 1. 2, and 3 were incubated for 60 min. During the first 60-min period aliquots were taken only from incubation mixture l_ At the end of 60 min, the remainder of incubation mixture 1 was discarded, incubation mixtures 2 and 3 centrifuged, the 59Fe-ghost fraction pelleted and saved, and the amount of mobilized iron determined (see arrows in Fig_ 7). By adding 15 ml of standard buffer at 4O to the incubation mixtures we could stop the mobilization process and effect a degree of washing. The 5gFe-ghost fractions from incubations 2 and 3 were taken up in 3 ml of standard buffer at 37” and allowed to equilibrate to temperature. At ro, 5 ml of Iysate was again added to each of the two preparations. During the next 60-min period, aliquots were only taken from incubation mixture 2. At the end of 60 min, the washing procedure was repeated,after which a kinetic study of mobilization from incubation mixture 3 was performed. Figure 7 shows the results of this experiment. It will be seen that at to incubation mixture 1 shows the characteristic rapid multiphasic mobilization of iron. When fresh lysate is added to the preincubated ghosts of incubation mixture 2, the bulk of the mobilization continues at the slow rate that was observed when the prior incubation was terminated. When the fresh lysate is added--to the 59Fe ghosts of incubation mixture 3, we again see the slow rate of the fifth kinetic phase. It appears that the rates
are governed
by the state of the membrane
and not the state of the lysate.
Membrane Iron Mobilization
49
25
0
0
10
20
30 TIME,
40
50
60
min.
FIGURE 8. Mobilization of 59Fe from 59Fe ghosts by three concentrations of bathophenanthroline sulfonate. Molar concentrations of chelate are as shown in curve de&nations. Bathophv throline sulfonate was dissolved in standard buffer containing, in addition, 1 mM Caz+ and 1 mM Mg2+_ Mobilization experiments were carried out in a fashion identical to that descrii above in which unlabeled lysate was used as the mobilizing agent. The buffer-only control is also depicted.
The question arks as to whether the lysate pool used in the repeat incubations could have deteriorated during the l-2 hr that it was incubated prior to use. On the right-hand section of the graph-is shown a 241 control in which both Iysate and 59Fe ghosts were maintained at 37” for 2 hr prior to utilization in a mobilization experiment_ It wih be seen that the overall mobilization time course of the 2-hr control is very close to that of the results observed with mixture 1. Mobilization by Bathophenanthroline Sulfonate The time course of the mobilization of 5aFe by bathophenanthroline sulfonate was examined in a manner parallel with the lysate-ghost incubation experiments. The media included 1 mM Mg2+ and 1 mM Ca2+. Figure 8 shows the time course of iron mobilization by three concentrations of bathophenanthroline sulfonate as indicated by the figure labels. Again we observe multiphaslc iron mobilization from the specifically labeled 59Fe ghosts. Phase 1 contributes 4.5-9.5% of the total iron. This is a quite narrow range when one considers that the chelate concentration varies over a lOCkfold range. The extents and rates of the following phases are independent of chelate concentration. The parallelism in these curves suggests that chelation is not rate limiting between 10-a M and 10s6 M. It appears that the membrane directs the rate at which iron becomes available for mobilization. Comparison of the mobilization of 5aFe by bathophenanthroline sulfonate with that of iysate reveals that the rates and extents of mobilization are greater in the latter case (with the exception of phase 5). Lysate may influence the rates of the phases rather than their number or duration. The rate of the final’ phase (observedat times >30 min) is invariant under a wide range of lysate and bathophenanthroline sulfonate concentrations, distinguishing it kinetically from the other phases.
Erwin F. Workman, Jr. and George W. Bates
50
Mobilization
by kythrocyte
Lysate
Lysate factors appear to exert some influence on membrane activity_ We thus investigated mobilization by mature red-blood-cell lysate. Rabbits not previously bled were utilized as erythrocyte sources. Heparinized samples contained less than 1% reticulocytes and were shown to be unable to take up iron from 5eFe-transferrinC032-. On the other hand, the reticulocyte preparation was found to be quite active. Reticulocytes were used for preparing 5sFe ghosts in the usual manner. The ghosts were then incubated with either unlabeled reticulocyte lysate or erythrocyte lysate. The tire course of mobilization by both lysate preparations was parallel and similar to the top curve of Fig. 6. The reticulocyte lysate mobilized 52.4% of the iron in 60 min, whereas the erythrocyte lysate mobilized 493% in the same time. The results are in excellent agreement with those of Blackbum and Morgan, who found 50.5% and 47.5%~~respectively [IO] _ We examined the fate of the mobilized iron in both lysate preparations_ A striking parallel was found in the Sephadex G-100 profdes. The lysate of the mature red blood cell incorporated iron into the fractions containing fenitin, hemoglobin (as confirmed by heme extraction), and the 6000-molecular-weight iron-binding component [S] _The time course of the appearance of radioactive heme in both incubation supematants was also examined. Over the first 15-min period, lysate from reticulocytes and mature erytthrocytes mobilized heme at approximately the same rate; however, at longer tunes the appearance of heme in the mature erythrocyte Iysate slowed markedly. After 60 mm the mature erythrocyte had incorporated 62% as much iron into heme as had the reticuiocyte lysate. These results do not necessarily indicate heme synthesis, since Blackbum and Morgan [IO] reported that much preformed heme is present in the membrane fraction.
DISCUSSION The experirrents reported herein point to the stability of the SeFe-ghost preparation, and along with the recent work of Blackbum and Morgan [lo], emphasize the potential of the lysate-ghost incubation system for probing reticulocyte iron metabolism_ The central point of this article is the evidence that the membrane fraction is not a passive, homogeneous iron depot, but an active component that makes iron available for mobilization at regulated rates. The rate at which iron is mobilized is dependent on the state of repletion of the membrane, as shown in Fig. 7. From this and previous studies [8-lo], we can infer that iron transport in the whole reticulocyte depends on both membrane-bound and cytosol factors. It is tempting to speculate that the multiphasic mobilization of iron from the specifically labeled 59Fe ghost is a manifestation of a multicomponent membrane iron trammort and storage system. Consistent with this interpretation is the work of Fielding and Speyer [6,7] , who used a variety of labeling, solubilization, chromatographic, and incubation procedures to obtain evidence for the presence of at least tiree membraneassociated iron-binding components_ An Fes+-transferrin-CO,*receptor site appears to have the role of interacting with the transferrin and perhaps removing the iron. The iron appears to move to a second component, which may have a relationship to ferritin. The subsequent pathway appears to have a branch point allowing iron to be stored in a membrane-associated component or transferred to cytoplasmic iron-transport agents.
51
Membrane Iron Mobilization
These studies also suggest that the membrane-associated components pky a dominant role in determining the rate at which iron becomes available to the cytosol. Corroborating evidence on the importance of the plasma membrane of the developing erythrocyte is provided by other investigators. Tanaka and Brccher [16] have suggested that fenitin plays an important part in iron exchange at the plasma membrane. Evidence was presented for the synthesis of apoferxitin at or near the ceil surface under conditions of iron overload. Denton et al. [ 171 have provided evidence that the cell stroma is a major iron-storage compartment during the erythroblastic stages. Iron release from the stromal fraction coincides with an increase in hemoglobin biosynthesis during the cell maturation process. The reticulocyte contains comparatively little stromal iron and is in what appears to be the final stages of stromal iron depletion. The plasma membrane of the developing erytbrocyte may play an important role in celhdar iron metabolism, not only for obtaining iron from transferrin, but also for storing iron and regulating its availability to the cytoplasm during cellular maturation. Blackbum and Morgan [lo] have pointed to the importance of heme as a major portion of the iron being mobilized from the stroma fraction, which includes a portion of the mitochondria of the cell. The multiphasic nature of mobilization could be accounted for in terms of heme and nonheme iron complexes being solubilized at varying rates. The fact that chelating agents arc not as effective as lysate in mobilizing iron may be a manifestation of their inability to complex heme. At this point it se&s safe to say that reticulocyte iron metabolism must involve an orchestrated interaction of membrane-bound and cytosol fractions, sub&h&u organelles, transfer&, and perhaps other serum factors.
This investigationwas supported by National Institutes of Health Research Cmnr No_ I ROI AM I 7790 from the National Institute of Arthritis. Metabolism. and Digestive D-es and represents partial fut”Iiment of the requirementsfor a doctoral degree by Edwin F. Workman,Jr_
REFERENCES 1. E. H. Morganand T. C. Appleton,Natlue223,1371-1372 (1969). 2. E. H. Morganand E. Baker,Biochim. Biophys Acta l&4,442454 (1969). 3. J. Martinez-MedeUiu and H_M. Schulman,Biochim Biophys Acta 264,272-284 (1972). 4. J. Borova,P. Ponka, and J. Neuwirt, B&him. Biophys- Acta 320.143-156 (1973). 5. 67. 8. 9.
10. 11. 12. 13. 14.
J. H. Jandl, J. K. Irunan. R. L.. Simmons, and D. W. Allen, J. C7k Invesx 38,161-X35 (1959). B. IL Speyer and J. Fielding,Biochim Biophys Actn 332,192-200 (1974). J. Fieldingand B. E. Speyer,B&him. Biosphys. Acta 363,387-396 (1974). E. F. Workmanand G. W Bates,B&hem Biophys Res. Cbmmun. 58,787-794 (1974). E. F. Workmanand G. W. Bates, in R. R. Chrichton,Ed.,&teinsoflron Storageand lkansport in Biochemistry and Medicine. North Holland,Amsterdam.1975, pp. 155-160. G. W. Blackbumand E. H. Morgan, Biochim Biophys Acta 497,728-744 (1977).
E. H. M0rgan.J. PIzysioLLand. 171,2641(1964). H. G. Van Eyk and B. Leijnse,Biochim Biophyr AC&~ 160,126-128 (1968). G. W. Batesand M. R. Schlabach,J. BioL them 248,3228-3232 (1973). J. H. Katz,L C7in.Invest. 40,2143-2152 (1961). 15. E. F. Workman,G. Graham,and G. W. Bates,B&him. Biophys. Acra 399,254-264 (1975). 16. Y. Tanakaand G. Brecher, Blood 37,211-219 (1971). 17.
M. J. Denton, H. T_ Delves, and H_ R_ V_ Amstein.Biochem
13 (1974). Received I 9 Apnl I9 78
Bkphys
Res
Commun.
61, S-