,81€¦ · j. h. jandl, j. k. inman, r. l. simmons, and d. w. allen iron in this material was...

TRANSFER OFIRON FROMSERUMIRON-BINDING PROTEINTO HUMANRETICULOCYTES*

By JAMESH. JANDL, JOHNK. INMAN, RICHARDL. SIMMONS, ANDDAVID W. ALLEN

(From the Thorndike Memorial Laboratory and Second and Fourth (Harvard) Medical Serv-ices, Boston City Hospital, and the Department of Medicine, Harvard Medical

School, Boston, Mass.)

(Submitted for publication August 7, 1958; accepted September 11, 1958)

The metabolism of iron involves a series ofcompounds which restrict iron to undissociatedforms. Upon absorption by gut mucosal cells,iron is enveloped within ferritin molecules (1, 2).In the presence of reducing agents, ferritin ironthen is released to the surfaces of ferritin mole-cules (3) to become bound by a specialized trans-port protein of the plasma.' This iron-bindingprotein (IBP) is a ,81 globulin by electrophoreticanalysis but resembles albumin in chemical prop-erties (5, 8). Each molecule of IBP from humanplasma or from the plasma of several other mam-malian species studied, in common with the IBPof egg white, "conalbumin," binds two iron atomsto form a colored complex (5, 9-12). Ironbound by these proteins is in the ferric state (5,8) and is tightly chelated in complexes which ap-parently contain three phenolic groups and onebicarbonate ion (7, 11, 12). At a physiologicpH and bicarbonate concentration the iron-IBPcomplex is extremely stable. In some manner,however, iron is transferred from IBP to im-mature red cells for incorporation into heme. Al-though it is well established that immature redcells, either as intact cells or as hemolysates, canutilize ionic iron for heme synthesis (13-20),little is known of the mechanism releasing ironfrom plasma IBP for cellular utilization. It isoften assumed that this mechanism for transfer-

* This investigation was supported in part by a grantfrom the Helen Hay Whitney Foundation, and by a re-search grant, P. H. S. No. RG3507 (C5), from the Na-tional Institutes of Health, Public Health Service.

1Although several metals, including Cu+, Zn+ (4, 5),Co' and Mn"I (6) may combine with this protein, itssole known physiologic function is to transport iron throughplasma. Thus the terms transferrin (4), siderophyllin(7), and iron-binding protein, of which the last will beemployed in this report, are perhaps preferable to lessspecific terms such as metal-binding protein or metal-combining globulin.

ring iron from IBP to immature red cells includestransition through marrow iron storage sites.Recently, indeed, Bessis and Breton-Gorius (21,22) have interpreted their electron-microscopicstudies of immature red cells to indicate thaterythroblasts may obtain iron for hemoglobinsynthesis by removing iron particles from con-tiguous macrophages. The physical disadvan-tages of such a system as compared to a solubletransfer system are great, however, and the ob-servation by Walsh and co-workers (15) thatintact immature red cells (human reticulocytes)are capable of taking up iron bound to plasmain vitro suggests that normally iron is transferreddirectly from IBP to immature red cells.

The interaction between IBP and immaturered cells represents a physiologic transfer systempeculiarly susceptible to study in vitro. Observa-tions characterizing this interaction, partiallydescribed in a preliminary report (23), are pre-sented below.

MATERIALS

Iron-binding protein of a high degree of purity wasprepared as described elsewhere by Inman (24), using asa starting material Fraction IV-7, obtained by standardethanol fractionation procedures and saturated with iron.2After recrystallization six times this material was ana-lyzed by electrophoresis and found to contain 97.6 percent iron-binding protein. Its molecular weight was cal-culated from data obtained by ultracentrifugation and bydiffusion to be 87,000 (6).

Iron-saturated conalbumin, prepared from hen's eggs,was generously supplied by Dr. Robert C. Warner.8 The

2 In this method the pH is not reduced below 5.2. Be-low a pH of 5, this protein is at least partially denatured.Partial denaturation of IBP during its preparation pre-sumably accounts for differences reported (25) in theturnover of iron bound to unaltered plasma IBP as com-pared to iron bound to certain commercial fractions ofIBP.

3Department of Biochemistry, New York UniversityCollege of Medicine, New York, N. Y.

161

J. H. JANDL, J. K. INMAN, R. L. SIMMONS, AND D. W. ALLEN

iron in this material was removed by dialysis against a0.4 per cent solution of the disodium salt of ethylenedi-aminetetraacetic acid (EDTA) 4 in physiologic salineadjusted to pH 5.6, and EDTA was subsequently re-moved by repeated dialysis against iron-free saline buf-fered at pH 7.4. Throughout these experiments Fe'5was derived from dilute acid solutions of radioactive fer-ric chloride to which four parts of stable carrier ironwere usually added. Studies of cobalt utilization by redcells were carried out with CoC°Cl2.5

Whole heparinized blood samples rich in reticulocyteswere generally obtained from patients either with per-nicious anemia during therapeutic response or with hemo-lytic anemias. In certain instances the proportion of retic-ulocytes was increased by centrifuging and withdrawingthe top portion of packed red cells. Finally, saline sus-pensions of red cells were usually prepared by washingthe specimens of heparinized peripheral blood three timesby centrifugation in approximately 20 volumes of iron-free physiologic saline. For the sake of brevity suchreticulocyte-rich red cell suspensions are referred to be-low in many places simply as "reticulocytes."

METHODS

Hematocrit and hemoglobin values of the washed redcell preparations were determined by standard methods(26); reticulocytes were enumerated by the dry method(26) after resuspension of a drop of washed cells in twodrops of plasma.

Complexing of iron and IBP was produced by slowlyadding Fe'Cl8 to solutions of IBP in iron-free physiologicsaline containing 10' M NaHCO,. This mixture wasallowed to react for one hour at room temperature be-fore addition to red cells, although color due to the for-mation of the Fe-IBP complex regularly reached a maxi-mumwithin a few minutes of admixture.

Cobalt and manganese also form stable colored com-plexes with IBP and, as with iron, the trivalent state isrequired for this reaction (6). Unlike iron, which auto-oxidizes readily to the ferric state under physiologicconditions, cobalt requires the presence of an oxidant inorder to remain in the cobaltic form. Thus one-tenthvolume of 0.1 MHO, was added to the reaction mixturecontaining Co'Cl,, NaHCO. and IBP.

In the majority of experiments a 30 1AM concentration ofIBP was employed and equal volumes of 20 or 30 /AMconcentrations of metal were added, resulting in one-thirdand one-half saturation, respectively, of the iron-bindingcapacity of IBP. Actual determinations of the iron-bind-ing capacity of purified IBP, plasma or serum were madeby the method of Rath and Finch (27), and iron con-centrations were measured by the method of Kitzes,Elvehjem and Schuette (28).

Following addition of washed red cells to the Fe"- orCo'-containing test mixtures of the type described above,

4"Sequestrene," three times recrystallized, AlroseChemical Company, Providence, R. I.

5 Abbott Laboratories, North Chicago, Ill.

the mixed suspensions were incubated at 370 C. with agi-tation. At the end of the incubation period the red cellswere washed with physiologic saline five times. Thepacked washed red cells were then diluted with physio-logic saline to a given volume in volumetric flasks, fromwhich 3 ml. samples were withdrawn for determination ofwhole red cell radioactivity in a well-type scintillationcounter. The red cell radioactivity was compared to thatof the standards, and the red cell uptake of metal wasgenerally expressed as micrograms of metal per millilitersof cells.

The fraction of radioactivity not present in the red cellmembranes was estimated in most experiments by centri-fuging 5 ml. of the final cell suspension and hemolyzingthe sedimented cells with 3 ml. of distilled water; 0.7 ml.of 5 per cent NaCl was then added, and the suspension wasmade up to volume in a 5 ml. volumetric flask with 0.85per cent saline. The hemolyzed suspension was then cen-trifuged clear of cell membranes at approximately 1,000 Gfor 15 minutes and the supernatant solution was assayedfor radioactivity. The localization of Fe' in red cellfractions was more precisely defined in certain instancesby differential centrifugation of hemolysates with an ul-tracentrifuge.6 The incorporation of Fe' into heme wasdetermined in certain experiments by measuring the ra-dioactivity of heme extracted from the red cells by themethod of Fischer (29).

In order to estimate simply and simultaneously the dis-tribution of Fe' between hemoglobin, IBP and the red cellmembranes, filter paper electrophoresis and autoradiog-raphy were employed. In this procedure, the washedreticulocytes were suspended in equal volumes of salineand hemolyzed by repeated freezing and thawing. At pH6.35, midway between the isoelectric points of IBP (5.9)and hemoglobin (6.8), these two proteins migrated op-positely, whereas the particulate matter of the cells re-mained precipitated at the origin. The proportion ofradioactivity traveling with the hemoglobin spot was de-termined by planimetric analysis and quantitated by com-parison with autoradiographs of various dilutions of theFe standard. In intact cell systems, good conformity ap-peared between measurement of hemoglobin incorporationof Fe' by this method, and heme incorporation of Fe' bythe method of Fischer. By the autoradiograph method,over 80 per cent of the radioactivity of supernatant solu-tions derived from water hemolysates of Fe'-labeledreticulocytes, prepared as described above, resided in thehemoglobin spot.

RESULTS

The uptake of Fe59Cl, from saline by mature andimmature red cells

Fe59 from a saline solution of Fe59Cl, wasbound to, or adsorbed onto, both mature andimmature red cells (Figure 1). The uptake ofradioactive Fe+-+ by both mature and immature

6 Spinco Model L Preparative Ultracentrifuge.

162

163IRON TRANSFERTO RETICULOCYTES

_-4 IMMATURE RED CELLSO-O ADULT RED CELLS

ALBUMIN

0I

0A

a2 = = r0ow

0 1 2 a 4

IRON-BINDIN16 PROTEIN

T-CLOSULIN

0 1 2 4

NOURS

FIG. 1. THE INFLUENCE OF PROTEIN BINDING ON THE UPTAKE OF IRON BY IMMATUREAND

MATURERED CELLS

Fe'Cl, was added to saline, albumin, iron-binding protein and 'y-globulin solutions. Afterone hour at room temperature these mixtures were incubated with a washed suspension of redcells containing 31.5 per cent reticulocytes from a patient with acquired hemolytic anemia (im-mature red cells) and with a washed suspension of normal adult red cells, only 0.3 per centof which were reticulocytes.

The uptake of Fe' by the washed cells after various periods of times was then measured as

described under "Methods." Note the striking selectivity with which immature red cells tookup Fe' from the iron-binding protein.

red cells was initially very rapid, and graduallydiminished thereafter. Approximately 90 percent of the Fe59 adsorbed by these red cells re-

mained in the membrane fraction after waterhemolysis and centrifugation at 1,000 G; andwhen the uptake of Fe59+++ exceeded about 5 Kg.per ml. red cells, red cell agglutination ensued, a

finding reported in more detail elsewhere (30).Siderocyte granules were not observed in suchiron-coated red cells.

Although the uptake of Fe... from saline bythe reticulocytes of patients with various diseaseswas usually similar to the uptake by mature redcells obtained from these same patients, the Fe++,uptake of reticulocyte-poor suspensions of red

cells from patients with acquired anemia and a

positive Coombs test was consistently smallerthan that of reticulocyte-rich suspensions preparedfrom the same blood specimens (Figure 2) or ofsuspensions of normal mature red cells (Figure1). In pursuance of this finding, the effect ofred cell sensitization with antibodies upon Fe"'uptake was assessed by comparing the uptake ofFe59 from 30 MMFe59C18 in one hour by washednormal red cells and by normal cells which hadbeen sensitized before washing with serum con-

taining incomplete anti-D. The iron uptake ofanti-D sensitized red cells was 2.18 Mug. per ml.red cells as compared to 3.43 Mug. of iron per ml. ofunsensitized red cells.

-i

0

a

ma

0

U.

a.

to

Ua.

1.0'

0*

0.4

0.2

A.

.1.0


The uptake of Fe59 bound to IBP by mature andimmature red cells

Iron bound to IBP (as 30 AM, half-saturatedFe59 IBP) was not taken up by mature red cellsdespite incubation at 370 C. for as long as 24hours. However, with red cell suspensions rich inreticulocytes an appreciable transfer of iron fromIBP to the cells occurred, with a slowly diminish-ing rate of uptake over the course of severalhours of incubation (Figure 1). By contrast,iron bound to human albumin (Cohn FractionV), in a concentration equimolar to that of theIBP studied, was transferred to some extent tomature red cells. Although iron was preferenti-ally taken up from albumin by reticulocyte-richcell suspensions, the uptake by these cells wasless than was the case with IBP (Figure 1).Thus, with iron bound to albumin there was less

(I)-i-iw0aw

0UN.Id

In

U.

FIG. 2. Ti

discrimination between mature and immaturecells than with iron bound to IBP. When ironwas bound to an equimolar concentration of hu-man gamma globulin (Cohn Fraction II-1,2)the iron uptake by adult red cells was still greater(Figure 1) but with no discrimination betweenadult and immature red cells.

In contrast to free iron, iron bound to IBP(Fe59 IBP) was taken up by a suspension ofred cells in direct proportion to its concentra-tion of reticulocytes (Figure 2). By extrapola-tion of data, assuming that iron was taken uponly by reticulocytes, the quantity of iron trans-ferred in two hours from IBP to the red cells ofvarious patients ranged from 0.5 to 4.5 pg. perml. reticulocytes. In general the amount trans-ferred correlated with the apparent morphologicimmaturity of the reticulocytes. Also in contrast

ont-O re IN SALINE0 * Fe9ANO I OP

0 10 go 30 40 50 60 70 60 S0 l0RETICULOCYTES. PER CENT

[E RELATION BETWEENRETIcULOCYTE CONCENTRATIONAND THE UPTAKE OF FEWFROMFE`CL. ANDFROMFE' IBP

Cell suspensions containing various reticulocyte concentrations were obtained by combiningthe top (reticulocyte-rich) and bottom (reticulocyte-poor) fractions of the centrifuged redcells of patients with acquired hemolytic anemia in different proportions. The uptake of Fe'from Fe' IBP is seen to occur in direct proportion to the reticulocyte concentration. As notedin the text, the somewhat greater uptake of Fe' from Fe'Cl, by reticulocytes than by adult cellswas peculiar to patients with acquired hemolytic anemia.

164

.__

IRON TRANSFERTO RETICULOCYTES

32.2 % RETICULOCYTES: Fess I BP

0

0-

z

0.

j04

°0 \ S ~~~~~~~2sRETICULOCYTES: Fe" Cls

bi.3It0 0

I--

\ ~~3.t% RZTICULOCYTE~s fe5s C13

0i

0.1 I 10 100

EDTA, MILLIMOLES /LITER

FIG. 3. THE ELUTA.BMITY FROMREDCELsS OF FeD TAKRENUP IN VITROAs indicated by the solid circles, very little of the Fe' taken up from Fe' IBP

by a reticulocyte-rich suspension of red cells could be eluted by subsequent ex-posure to high concentrations of EDTA, albeit over half of the Fe' was in thecell membranes. Fe'@ taken up from Fe59Cl' by a reticulocyte-poor suspensionof red cells from the same patient was almost entirely eluted by EDTA (solidtriangles). Fe' taken up from Fe5'Cl, by the reticulocyte-rich cell suspensionwas partially eluted, but almost 30 per cent of the Fe'§ resisted elution by highconcentrations of EDTA (empty circles), although less than 12 per cent of theFe'9 had been incorporated into hemoglobin.

to the 10 per cent of Fe59 derived from Fe59Cl3in saline, from 40 to 90 per cent of red cell Fe59derived from Fe59 IBP was recoverable in thesupernatant solution of water hemolysates of reti-culocyte-rich blood, spun at about 1,000 G. Thusfrom 10 to 60 per cent of the Fe59 was associatedwith the reticulocyte membranes. In a repre-

sentative experiment, 49.1 per cent of the Fe59IBP taken up by reticulocytes in vitro was in the

supernatant solution of the hemolysate spun at1,000 G. A small portion (4.5 per cent of thecell uptake) of the Fe59 in the original supernatewas sedimented with the ribonucleic acid-richmicrosome fraction, obtained by centrifuging thesupernate at 105,000 G for 60 minutes. Anothersmall portion of Fe59 (0.9 per cent of the celluptake) was recovered in that fraction obtainedby centrifuging the supernate at 26,360 G for 20

165


minutes and believed to contain the mitochondria.However, most of the nonmembrane Fe59 re-mained in the water-soluble phase of the lysedred cells and was incorporated into heme (seebelow). When a suspension of reticulocytes con-taining Fe59 derived from Fe59 IBP was washedfree of Fe59 IBP and then incubated for up to24 hours in native plasma or in 60 ,MM unsaturatedIBP, there was no appreciable elution of Fe59into the medium, and during this period the pro-portion of Fe59 in the supernate of the hemolyzedcells to that of the whole cells (62 per cent) re-mained unchanged.

Iron taken up from Fe59 IBP by a red cellsuspension containing 32.2 per cent reticulocytes

2.0

-a

a

1U.hi

I~-IL

0IL

was not appreciably eluted by subsequent ex-posure to EDTAin pH 6.8 or pH 7.4 phosphatebuffer, even at concentrations as high as 0.1M EDTA (Figure 3). This was true bothof intact Fe59-labeled reticulocytes, in whichapproximately 55 per cent of the Fe59 was in themembrane (as determined by centrifugation ofhemolysates), and of the Fe59 in cell membraneswashed "free" of hemoglobin and then resus-pended in EDTA. Thus the Fe59 was not un-elutable simply because of incorporation intoheme. On the other hand, all but 28 per cent ofthe iron taken up by the same red cell suspension(containing 32.2 per cent reticulocytes) fromFe59C13 in saline was eluted by EDTA. That

0 10 20 30 40 30 60 70 SO 90 100

I BP SATURATION, PER CENTFIG. 4. THE EFFECT OF THE IRON SATURATION OF IBP UPONTHE IRON UPTAKEOF RETICULO-

CYTE-CONTAINING RED CELL SUSPENSIONSFROMFOUR SUBJECTSNote the tendency for Fe' uptakes to level off at IBP saturations of over 20 to 30 per cent.

If the iron uptake data above is plotted against the reciprocal of the IBP saturation (see Fig-ure 5), straight-line relationships obtain at saturations from 0 to 50 per cent. The symbolsrepresent red cells from subjects with the following hematologic conditions: acquired hemolyticanemia with 38.9 per cent reticulocytes (empty triangles) ; sickle cell anemia with 18.4 per centreticulocytes (solid triangles); pernicious anemia during a therapeutic response with 14.9per cent reticuloyctes (empty circles) ; and normal with 1.2 per cent reticulocytes (solid circles).Except with the acquired hemolytic anemia red cells a heavy "nonspecific" adsorption of iron onred cells became evident when the IBP saturation exceeded 60 to 85 per cent.

166


I1

f .7

4

S .2,

0Dh).4

PI A

| 2

.II

S

0

.1

.. 1 2 3 4 5 6 i 8MOLAR RkTIO IBP/Fe59

FIG. 5. THE COMPETITIVE EFFECT OF IBP ON IRON UPTAKEBY RETICULOCYTESEach point represents the Fe' uptake by equal volumes of reticulocytes incubated with con-

stant amounts of Fe' and increasing concentrations of IBP. Since each molecule of IBP iscapable of binding two atoms of iron, the points represent IBP saturations ranging from 50to 6 per cent. The expression "molar ratio IBP/Fe5" is equivalent to half the reciprocal of thefractional iron saturation of IBP.

taken up by the mature red cells was almostentirely eluted by EDTA (Figure 3) as is trueof other multivalent metals adsorbed onto redcells (30). Thus, the 28 per cent of the radio-activity that remained unelutable despite highconcentrations of EDTA (Figure 3) was presum-ably bound to reticulocytes, but was not incorpor-ated into heme, for 88 per cent of it was attachedto the cell membranes.

In an effort to determine the chemical fractionof reticulocyte membranes to which Fe59 wasbound, a reticulocyte-rich red cell suspension wasincubated with autologus plasma, which had beenbrought to half-saturation with Fe59. Thesecells were then washed five times with physiologicsaline, after which they were hemolyzed in threevolumes of cold water and later suspended in 50volumes of cold acetate buffer for 18 hours.The stroma residue which precipitated from thissuspension was subjected to lipid extraction bythe method of Folch and associates (31). Lessthan 0.1 per cent of the Fe59 taken up by thecells was recovered in the lipid extract of their

membranes, whereas 34.8 per cent remained boundto the stroma residue. This Fe59-containingresidue was partially soluble in 30 per cent so-dium cholate. However, efforts to characterizethe iron-binding component of this cholate-sol-uble extract by paper electrophoresis and bycolumn chromatography were unsuccessful.

The transfer of iron from IBP to reticulocyte-rich suspensions of red cells was greatly affectedby the degree of iron saturation of the IBP. InFigure 4 it is evident that the Fe59 uptake of retic-ulocytes approached a plateau when the ironsaturation of the IBP exceeded 20 to 30 per cent.At iron saturations in excess of 60 to 85 per centthere was usually a sharp increase in Fe59 up-take; this appeared to, be a nonspecific adsorption,for a similar sharp increase in Fe59 uptake alsooccurred when adult red cells were so studied.This "nonspecific" iron uptake at high iron satura-tion was partially diminished by increasing thebicarbonate concentration. Comparable experi-ments in which iron-deficient human plasma wassaturated to various extents with iron revealed

167


that only when the plasma iron-binding capacitywas exceeded did iron "nonspecifically" attachto red cells. Even when the iron-binding capac-ity of plasma was exceeded, however, a majorityof the excess iron was bound by other proteins.As indicated by filter paper electrophoresis andautoradiography, iron added in excess of theplasma iron-binding capacity at an alkaline pHwas largely bound to a protein migrating mid-way between the alpha-1 and alpha-2 globulins;greater excesses of iron also migrated with al-bumin. There was no difference in this respectbetween normal plasma and the plasma of pa-tients with hemochromatosis.

When the IBP was less than about 20 per centsaturated, the iron uptake of reticulocytes dimin-ished despite an amount of iron in the system wellin excess of that utilizable by the reticulocytes.The competition of IBP for iron in the presenceof a given volume of reticulocytes was demon-strated either by maintaining the IBP concen-tration constant and varying the amount of ironadded (Figure 4) or by keeping the iron con-centration constant and varying the amount ofIBP (Figure 5).

When to a given amount of reticulocytes vari-ous quantities of IBP of a constant Fe59 satura-tion (33.3 per cent) were added, the Fe59 uptakeincreased in proportion to the IBP concentra-tion until a concentration of 160 juM was exceeded(Figure 6). Thereafter, further increase in Fe59IBP caused no further uptake.

In an effort to detect whether IBP enteredreticulocytes or became attached to their sur-faces, measurements were made of the IBP con-centration of a preparation before and after ex-posure to a relatively large volume of reticulo-cytes. Despite removal by the reticulocytes ofover 30 per cent of the iron, there was no changein the total IBP concentration.

The uptake by reticulocytes of Fe59 bound tovarious mammalian plasmas and to conalbumin

The uptake by human reticulocytes of ironfrom human plasma labeled with 1 Mug. Fe59 perml. plasma was similar to that from comparableconcentrations and saturations of purified humanIBP. As with human red cells, iron bound to30 juM human IBP was not transferred to maturerabbit or rat red cells (iron uptake 0.03 pg. per

.7J8 .6

a.

0 0

ONE-TIRD SATURATOD 18P, MICROMOLE8/LITER

FIG. 6. THE EFFECT OF IBP CONCENTRATIONON THE UPTAKEOF IRON BY RETICULOCYTESIBP which had been 33.3 per cent saturated with Fe59 was added in increasing concentrations

to a series of test tubes holding equal volumes of a suspension of red cells containing 18.0 percent reticulocytes. Iron uptake increased in direct proportion to the Fe'9 IBP concentration un-til the latter exceeded 160 ,uM (1.39 Gm. per cent). The Fe' uptake at physiologic concen-trations of IBP (approximately 27 ,uM) was about 77 per cent of that at 160 /AM concentrationsof IBP.

168


TABLE I

Utilization of Fer9 by human reticulocytes

Iron Fe59 incorporation Red cell Fe5' Feft incorporation Red cell Fe59preparation Fes uptake into heme (25) in heme into hemoglobin* in hemoglobin

sg./ ml. red cells pg./Gm. heme % pg./Gm. hemoglobin %Fell IBP 0.24 15.6 76.5 0.485 64.6Fell Cl3 2.32 10.5 5.7 0.251 3.5

* Determined by autoradiography as described in the text.

ml. cells per two hours). A suspension of rabbitred cells containing 33.0 per cent reticulocytes,however, took up 0.60 jg. iron per ml. cells in twohours from human IBP, as compared to 1.4Mg. from rabbit plasma; and a suspension of ratcells containing 14.0 per cent reticulocytes tookup 0.17 Mg. iron per ml. red cells in two hoursfrom human IBP, as compared to 0.37 Mg. fromrat plasma. Thus in each instance iron wastransferred from human IBP to heterologous retic-ulocytes, albeit not as readily as from the na-tive plasma.

Fe5' bound to 30 MuMhalf-saturated conalbuminwas not taken up by mature human red cells(0.006 Mg. iron per ml. red cells in two hours).However, a preparation of human red cells con-taining 33.5 per cent reticulocytes took up 0.535Mg. iron per ml. red cells in two hours from 30MMconalbumin, as compared to 0.613 Mg. ironper ml. red cells in two hours from 30 MMhu-man IBP. Thus, in vitro, conalbumin was ap-proximately as efficient an iron transport ve-hicle for human reticulocytes as was human IBP.Conversely, chicken (nucleated) red cells (6.2per cent reticulocytes) took up in two hours halfas much iron (0.52 Mg. per ml. red cells) fromhuman IBP as they did from conalbumin (0.94Mg. per ml. red cells).

The fact that chicken red cells utilized ironfrom conalbumin somewhat more readily thanfrom human IBP suggested the possibility thatthis protein, which normally is concentrated inhens' eggs, may also function as an avian plasmatransport protein. As was intimated by Jensen,Ashenbrucker, Cartwright and Wintrobe (19),the addition of radioactive iron to fresh chickenplasma (which had an iron content of 160 Mg.per cent) caused no colorimetric change. Theadded Fe59 was largely dialyzable thereafter andwas "nonspecifically" adsorbable by both matureand immature human red cells. However, when

native chicken plasma was dialyzed at pH 5.6against EDTAto free it of iron, and then dialyzedfree of EDTA, at a pH of 7.4, this iron-freeplasma was found capable of binding about 200Mug. per cent of radioactive iron, as determinedby dialysis experiments. Filter paper electro-phoresis (0.05 M barbital buffer at pH 8.6) andautoradioautography of chicken plasma so labeledwith Fe59 revealed a specific iron-binding proteinwith the same mobility as that of conalbumin:midway between the /3 and y positions.

Incorporation of Fe59 into heme and into hemo-globin by human reticulocytes

In systems containing only labeled iron, IBPand washed reticulocytes suspended in saline andincubated at 370 C. for two hours, from 20 to50 per cent of the Fe59 taken up by the reticulo-cytes was recoverable in the heme fraction ob-tained by the method of Fischer (29). Cor-recting for heme lost during fractionation, it wascalculated in various experiments that from 30to 80 per cent of the reticulocyte Fe59 was in-corporated into heme. Despite the higher totalcell uptake of Fe5' derived from Fe5"Cl, thanfrom Fe5' IBP, the portion of this Fe5' whichwas recovered with heme was far less, rangingfrom 2 to 12 per cent and varying in direct pro-portion to the percentage of reticulocytes. Pre-sumably this is due to the fact that mature redcells, which adsorb free iron but do not synthesizeheme, constituted a majority of the cells in mostsuspensions studied. In a single experiment theincorporation of Fe5' from Fe5Cl, and from Fe59IBP into both heme, as determined by the Fischermethod, and into hemoglobin, as determined byfilter paper electrophoresis and autoradiography,was compared (Table I). Both methods were inessential agreement. Note that the calculatedabsolute amount of Fe59 incorporated from Fe"QC,

169


into the heme fraction of reticulocytes was onlyabout two-thirds of the amount of Fe59 incorpor-ated into the heme of reticulocytes incubated withFe59 IBP.

Factors affecting the transfer of iron from IBPto reticulocytes

a) Temperature. The effect of temperature oniron uptake by reticulocytes was determined byincubating reticulocyte-rich red cell suspensions

hi

a)

*00S.JSI-Z

and Fe59-containing preparations together at cer-tain temperatures, after these components hadbeen separately equilibrated at these temperaturesfor 15 minutes. After incubation for two hours,the reaction was stopped by the addition of NaCN(see below), and the cells were then washed asusual. The uptake by reticulocytes of Fe59bound to IBP was markedly temperature-de-pendent. As indicated in the upper part of Fig-ure 7, there was very little uptake at 0.50 C.

D.3

.2

Fe59 IBP/\

'.6

.5

.4

Fe59 Cl3

.2.

.I-

i , . A= -nv10 30

TEMPERATURE,DEGREESCENTIGRADE

FIG. 7. THE EFFECT OF THE TEMPERATUREOF INCUBATION ONTHE UP-TAKE BY RETICULOCYTESOF FE' FROMFE" IBP ANDFROMFE"CLs

In each instance the whole cell uptake of Fe" (solid circles) increasedwith increases in temperature up to 370 C. However, Fe" uptake fromFe" IBP (above) was more strikingly suppressed at 0.5' C. and at42.50 C. than that from Fe'Cl, (below). In each study Fe" incorporationinto hemoglobin (empty circles), as determined by autoradiography, wasalso clearly depressed by nonphysiologic temperatures but an approxi-mately constant proportion between hemoglobin Fe' and whole cell Fe"remained.

170

c

.11


Above 140 C. iron uptake increased steadily toa maximum rate at 370 C.; at 42.50 C. the up-

take was sharply inhibited. Based on these datathe heat of activation (A E) was calculated as

12,200 calories per mole of iron. Almost identicaltemperature dependency was observed when Fe59bound by plasma was incubated with reticulocytes.In parallel studies the uptake from Fe59CJ9 insaline both by reticulocyte-rich cell suspensionsand by reticulocyte-poor cell suspensions alsoincreased with rising temperatures, but this effectwas less striking. The ratio of Fe59 uptake fromFe59 IBP at 37° C. over that at 0.50 C. was 12.6;the ratio of Fe59 uptake from Fe59C13 at 370 C.over that at 0.50 C. was 6.2. As seen in Figure7 there was also a fall-off of the uptake of ironfrom Fe59Cl at 42.50 C., but this effect was slightcompared to the sharp reduction when Fe59 IBPwas employed. The incorporation of Fe59 fromeither Fe59C1I or Fe59 IBP into hemoglobin as

determined by autoradiography showed a tem-perature dependence very similar to that of thewhole cell uptake of Fe59 (Figure 7).

b) pH. The effect of pH on iron uptake was

determined both in systems containing Fe59 IBPand in systems with Fe59-labeled plasma. Bothsystems were adjusted to various pH levels withN/3 HCl or N/3 NaOH, and the final experi-mental pH level was determined after addition ofthe washed reticulocytes. A maximum of Fe59uptake ranging from pH 7.2 to pH 7.6 was

found in both systems, with sharp reductions out-side of this range. Thus, there was very littleiron uptake below pH 6.8 and above pH 8.0.

c) Oxygen tension and the state of oxidation.The effect of oxygen tension on iron uptake byreticulocytes in vitro was investigated, employingautologus plasma labeled with Fe59. After re-

moval by centrifugation of the red cells of bloodsamples of patients with pernicious anemia under-going treatment with vitamin B12, Fe59CJ3 was

added to the native, hypoferremic, heparinizedplasma in an amount calculated to half-saturatethe iron-binding protein. Meanwhile, aliquots ofthe reticulocyte-rich red cells in unlabeled plasmawere placed in a series of air-tight separatoryfunnels, each of which was equilibrated three timesat one atmosphere pressure with one of the seriesof gas mixtures which contained various concen-

trations of 02 in N2 with a constant amount

TABLE II

Effect of oxygen tension on the utilization of iron byhuman reticulocytes

SubjectP02 A B C D Average

Fe"9 uptake (jig./ml. red cells)mm. Hg

0 0.139 0.227 0.061 0.068 0.12420 0.08339 0.176 0.234 0.075 0.082 0.14281 0.224 0.243 0.082 0.090 0.160

113 0.212 0.252 0.079 0.089 0.158144 0.210 0.261 0.084 0.091 0.162724 0.210 0.273 0.086 0.095 0.166

Per cent of red cell Fe"9 in "hemoglobin"*0 42.3 81.2 47.6 57.0

2039 32.1 73.8 38.5 48.181 28.2 72.5 38.2 46.3

113 27.6 72.0 36.9 45.5144 29.6 66.3 39.3 45.1724 22.1 63.2 33.9 39.7

* Determined only presumptively in these experimentsas that radioactivity remaining in the supernatant solutionfrom centrifuged reticulocyte hemolysates.

(4.7 + 0.2 per cent) of CO2. Fe59-labeled plasmawas then added to each separatory funnel and themixtures were re-equilibrated three times with thesame gas mixtures and incubated at 370C. forthree hours. As portrayed in the upper part ofTable II, the iron uptake of the whole cells (pre-sumably reticulocytes) was least under anaerobicconditions and increased somewhat with increasing02 tensions. On the other hand, the portion ofred cell Fe59 which was present in the superna-tant solution derived from hemolysates of thesecells was greatest under anaerobic conditions anddiminished with increasing 02 tensions.

Although it was possible to decolorize par-tially the iron-IBP complex with certain reduc-ing agents such as sodium metabisulfite (2 X10-2 M), cysteine (10-3 M) and glutathione (2 X10-3M), these agents did not facilitate Fe59transfer from IBP to reticulocytes. Rather,Fe59 uptake by reticulocytes was somewhat dimin-ished by chemical reduction, and was moderatelyincreased by 5 x 10-2 M H202 and by 5 X 103M KMnO4.

d) Hemolysis. The effect of preliminary hemo-lysis of reticulocytes on their capacity to take upFe59 from Fe59 IBP was determined by incubat-ing Fe59 IBP with three aliquots of a reticulo-cyte-rich suspension of red cells, (a) in saline,

1 71


CONTROL Pb

Fe59 IBP

Fe..

5'9G 3

FIG. 8. THE EFFECT OF LEAD ON THE INCORPORATIONBY RETICULOCYTESOF FE' FROMFE' IBP AND FROMFECL, INTO HEMOGLOBIN

After incubation with Fe', human reticulocytes were washed free of unutilized iron prior to being hemolyzedand subjected to electrophoresis at pH 6.35 in 0.1 M phosphate buffer. The upper (lighter) of each of the fourpairs of strips is the unstained hemoglobin paper electrophoretogram. In each the heavy spot to the left repre-sents hemoglobin migrating toward the anode, and the small spot to the right is at the origin and consists largelyof cell membranes. The lower (darker) of each of the four pairs of strips is the autoradiogram of the paper strip.As compared to incubation without lead (left), reticulocytes incubated with 10-' M PbCl2 (right) showed amarked inhibition of incorporation of Fe' into hemoglobin, whether the Fe' was derived from Fe' IBP or FeWCI.At this concentration of lead, however, the total reticulocyte uptake of Fe' was only slightly inhibited, causing arelative accumulation of Fe' on the reticulocyte membranes. Note in the control study how much more efficientlyFe's was incorporated into hemoglobin when derived from Fe' IBP than when derived from FeCl,.

(b) after hemolysis with three volumes of dis-tilled water, and (c) after hemolysis by rapidfreezing and thawing. After incubation for twohours at 370 C. the intact reticulocytes in (a)were also hemolyzed with distilled water, andthe cell membranes in all tubes were then washedthree times with cold physiologic saline. TheFe59 content of the membranes from the reticulo-cytes hemolyzed in (b) and (c) before incubationwith Fe59 IBP was only 7 per cent of that ofmembranes from reticulocytes in (a) which wereintact during the incubation period.

e) Metabolic accelerators and inhibitors. Theinfluence of several metabolic accelerators andsubstrates on the transfer of Fe59 from bufferedsolutions of Fe59 IBP to human reticulocyteswas evaluated during two hour periods of in-cubation at 37.5° C. Glucose in concentra-tions of 2 x 10-2 M increased Fe59 uptake by

reticulocytes moderately, with values rangingfrom 110 per cent to 150 per cent of controlvalues. Inosine (5 x 10-2 M), glycine (10-1M), succinic acid (10-1 M), methylene blue (20mg. per cent) and citrate (10-1 M), all in thepresence of 2 x 10-2 M glucose, did not increaseFe59 uptake further, either singly or in combi-nation.

The uptake of Fe59 from Fe59 IBP by reticulo-cytes was moderately depressed, but not abolishedby HCO5- concentrations in excess of 10-2 M.At given HCO5- levels, sodium acetazoleamideat concentrations as high as 50 mg. per ml. cellshad no effect on iron uptake. Each of several"metabolic inhibitors" studied (10-2 M sodiumcyanide, 10-2 M sodium fluoride, 10-2 M sodiumazide, 10-3 M sodium arsenate and 10-i3 M iodo-acetic acid) reduced the Fe59 uptake by reticulo-cytes (suspended in Krebs-Ringer phosphate so-

172

.f


lution containing glucose) to less than 20 percent of control values. Marked inhibition ofiron uptake was caused by as little as 104 Mconcentrations of 2,4-dinitrophenol and 10-5 M2,4-dinitrophenylphenol. However, concentra-tions of fluoracetate as high as 2 X 10-2 M hadno inhibitory effect.

The possibility that iron uptake was suppressedindirectly by inhibition of heme synthesis was ex-

amined by observing the effect of lead, whichinterferes with several stages in the synthesis ofheme (32, 33). Although inhibited by 5 X 104

concentrations of Pb++, the whole cell uptake ofFe59 by human reticulocytes during two hoursat 37.5° C. was only slightly diminished (- 20.6per cent) in the presence of 1 X 10-4 M Pb++.However, the incorporation of Fe59 into hemoglo-bin was almost entirely prevented, as is evidentin Figure 8, whether the Fe59 was derived fromFe59 IBP or from Fe59Cl5 in saline. Thus theamount of iron remaining in the cell membranewas increased. This pile-up of iron on the mem-

brane was not the result of an increase in mem-

brane hemoglobin and the increased membraneFe59 in lead-poisoned reticulocytes was not re-

covered in heme extracts of the membrane.The results of a comparable study of the effect

of lead on the uptake of Fe59 from rabbit plasmaby the reticulocytes of bled rabbits are presentedin Table III. In this experiment a reticulocyte-rich suspension of washed rabbit red cells was

incubated for five minutes with 10-4 M PbCl2,after which Fe59-labeled plasma was added andincubated with the cells at 37°C. under oxygen

for one hour before cell fractionation was carriedout. It is evident from Table III that, althoughthe total cell uptake of Fe59 was moderately dimin-ished by this concentration of Pb++, there was a

striking increase of Fe59 in the stromal (mem-

brane) fraction of lead-poisoned cells, and, pro-

portionately, a reduction in Fe59 in the hemoglobin-containing supernate. There was no evidencethat lead caused iron to accumulate in the micro-somal fraction or in the fraction believed to con-

tain the mitochondria.f) Metals and chelating agents. The question

whether, in its transfer from IBP to reticulocytes,iron exists at any time in the free ionized statewas examined by adding Fe59 IBP to each of a

series of tubes containing a reticulocyte-rich sus-

pension of red cells to which various amountsof nonradioactive FeCl. had been added. Therewas no competitive inhibition of Fe59 uptake bythe added ionized iron. The effect on iron up-

take of several other metals known to participatein enzymic systems was also explored. Concen-trations of KCl, MgCl2, and CaCl2 ranging from10-4 to 10-2 M had no effect. CoCl2 concentra-tions of 10-3 M, or greater, moderately suppressediron uptake.

More evidence that iron did not exist in thefree state during transfer from IBP to reticulo-cytes, and that traces of nonferric metals were

not involved in the reaction, was obtained byuse of EDTAwhich competes with IBP for iron,As is shown in Figure 9, EDTAdissolved in pH7.4 phosphate buffer completely blocked the up-

take by red cells of Fe59 from Fe59Cl3 at a molarratio of EDTA/Fe of two. EDTA was equallyeffective in this respect on reticulocytes (Figure9) and on adult red cells (not shown), indicatingthat it did not function as an iron transfer agent.

In contrast to the findings with Fe59C15, the trans-fer of iron from IBP to reticulocytes was not

significantly affected by the presence of EDTAuntil the ratio of EDTA/Fe exceeded 10; andFe59 uptake from Fe59 IBP was not abolished even

at a ratio of EDTA/Fe of 100, despite the fact that

TABLE III

The effect of lead on the uptake and intracellular distribution of iron in rabbit reticulocytes

Fe5 activity (cpmlml. cells)

Sediment from spun hemolysate

Reticulocyte Whole 7,000 G X 15 min. 26,000 G X 20 min. 105,000 G X 60 min. Solublesuspension cells (Cell membrane) (Mitochondria) (Microsomes) supernate

Control 446,250 23,200 800 700 412,200Lead-poisoned 210,200 113,100 1,600 900 76,720

(10-4 MPb++)

173


lI

0.9

-e-Jw

0

a

Ir00

q.

0nLu

I~-

CL

09it

0a.s

0.7

06

05

0.4

0.3

0.2

0.1

0 90 20 30 40 50 60 70

MOLAR RATIO EDTA/Fe59

80 90 100

FIG. 9. THE EFFECT OF EDTAON THE UPTAKEOF FE BY RETICULOCYTESFROMFE'9 IBP AND

FE'CL,With a molar ratio of EDTA/iron of two or more, iron was not taken up from FeWCl, by

reticulocytes. However, a relatively large excess of EDTAwas needed to suppress the trans-fer of iron from Fe' IBP to reticulocytes.

this amount of EDTAcaused over 50 per cent dis-sociation of the Fe59 IBP complex as determinedcalorimetrically.

g) Effect of enzymatic alteration of red cellsurfaces. In order to explore the hypothesisthat the immature red cell membrane includes on

its surface receptor sites which bind iron, theeffects of certain enyzmes which alter the red

cell surface were observed. Reticulocyte-richwashed red cell suspensions were incubated forone hour at 370 C. in pH 7.6 phosphate bufferwith various concentrations of trypsin.7 Thesecells were then washed in phosphate buffer (pH7.4) once, and in physiologic saline twice, andwere then incubated, in parallel studies, with Fe59IBP and with Fe59Cl3 in saline. Pretreatment ofreticulocytes with as little as 0.001 per cent tryp-

7 Trypsin, crystalline (salt-free, lyophilized). MannResearch Laboratories, New York, N. Y.

sin (0.01 mg. per ml.) markedly suppressedtheir ability to take up Fe59 IBP (left portion ofFigure 10). On the other hand, the adsorptionof Fe59CI, by these enzyme-treated cells was un-

affected. As indicated by the interrupted line inFigure 10, the uptake of iron from Fe59 IBPby trypsinized reticulocytes was consistentlygreater (although still markedly depressed) whenglucose (0.015 M) was included in the system

during incubation of cells and Fe59 IBP. Whenreticulocytes were first trypsinized, then incu-bated with glucose for two hours, and finally incu-bated with Fe59 IBP, with two saline washingsof the cells between each incubation, the effect oftrypsinization upon iron uptake was again dimin-ished but not to the extent observed when glucosewas present during the incubation of cells withFe59 IBP. The presence of iron in the mediumduring trypsinization did not affect the subse-

O-O 20 PM Fe9 IN SALINE

*@* tO pM Fens IN 30,pM I SP

Is *

I

I

I

j

174

%p


quent inhibition of Fe59 uptake. Trypsinization ofreticulocytes curtailed the subsequent uptake ofFe59 from Fe59 bound to plasma to the sameextent as that bound to purified IBP. Reticulo-cytes, previously incubated with various concen-trations of trypsin as described above, were re-suspended in plasma containing Fe59 and werethen incubated in Warburg flasks for three hours.During this time oxygen consumption was ob-served, and at the end of incubation the lactatecontent of the supernatant plasma specimens andthe Fe59 activity of the red cell specimens weredetermined. Although Fe59 uptake by the reticulo-cytes was, once again, sharply diminished by pre-treatment with 0.001 per cent, 0.01 per cent, and

-I~~~~~~~~~~~~~~~~~~D

1OO _ ~~~~~Fe59C1390

S.j60

I-

I-

zu \

oWON\IX50 \IL4e \

4 \A -A c \ \

0.1 per cent trypsin, these concentrations of tryp-sin had no effect upon either oxygen consumptionor lactate production. Incubation of reticulo-cytes with 0.1 per cent trypsin, after these cellshad taken up Fe59 from Fe59 IBP, did not causeelution of Fe59 into the medium, even though 50per cent of the Fe59 was in the cell membranesas determined by centrifugation. Chymotrypsin 8treatment also blocked the uptake of Fe59 byreticulocytes, although it was not quite as potentin this respect as trypsin.

The effect of enzymatic destruction of thesialic acid of the reticulocyte surfaces by receptor-

8 Chymotrypsin, crystalline. Mann Research Labora-tories, New York, N. Y.

0.001 0.01 o.0 I 10 o0.o 0.1 I 10 100TRYPSIN, MG/CC RECEPTOR-DESTROYINGENZYME, UNITS/CC

FIG. 10. THE EFFECT OF ENZYMATICALTERATION OF RETICULOCYTE SURFACESON THEIR SUBSEQUENTABILITY TO TAKE UP IRON FROMFE5CL, AND FE IBP

Neither trypsin (left) nor receptor-destroying enzyme (RDE) (right) significantly affected the up-take by reticulocytes of Fe' from Fe5Cl,. Exposure of reticulocytes to extremely low concentrations oftrypsin, however, caused a marked suppression of their uptake of iron from either purified IBP (left) orfrom plasma (not shown). This effect of trypsin was partly reversed by the presence of glucose (inter-rupted line). RDEalso diminished the subsequent uptake of reticulocytes of Fe from Fe IBP. How-ever, the quantity of RDE preparation required to cause inhibition far exceeded that of trypsin.

175


A

2.5

W a 2.0. 0.01 CC Fe5 SERUM

z0

cc1.5.z0z0

* 1.0 00.06 CC Fe59 SERUM

0I

!i 05

0.80 CC FeS9 SERUM

0~~~~~~~~~0 1 2 3 4 5 6

HOURS

FIG. 11. THE UPTAKEOF FE FROMRAT SERUMBY RAT LIVER SLICES AS AFUNCTIONOF TIME

Rat liver slices took up iron from rat serum in direct proportion to the periodof incubation. Only when comparatively small amounts of Fe'9-labeled rat se-rum were present, however, was an actual concentration gradient achieved.Thus when five rat liver slices were incubated with 0.01 ml. of Fe'n serum forsix hours (solid triangles), the concentration of Fe5' in the slices was 3.1 timesthat of the medium. However, when 0.8 ml. of Fe'n serum was present (solidcircles) the concentration of liver slice Fe' was only 0.15 that of the medium.

destroying enzyme (34) was investigated. Lyoph-ilized "Receptor-Destroying Enzyme" (RDE) ,'derived from vibrio cholera filtrates by the methodof Ada and French (35), and quantitated in themanner of Burnet, McCrea and Stone (36), wasdissolved in saline, and various amounts wereadded to a series of tubes containing bufferedsaline (pH 6.6) and 0.002 M CaCl2. Washedsuspensions of reticulocyte-rich red cells were in-cubated with the RDEpreparations, and the cel-lular uptake of Fe59 was then determined as de-scribed above for the experiments with trypsin.Like trypsin, RDEhad no effect upon the abilityof red cells to adsorb Fe.. from Fe59Cl insaline. However, at concentrations of this en-zyme exceeding 1 unit per ml., reticulocytes

9 Behringwerke, Marburg-Lahn, Federated Republic ofGermany.

were impaired in their ability to take up ironfrom Fe59 IBP (Figure 10). Nevertheless, com-pared to trypsin, much greater concentrations ofthe RDE preparation were required to inhibitsuch iron uptake. Thus, 240 mg. of RDEpro-tein per ml. red cells was required to suppress ironuptake to the degree effected by 0.01 mg. oftrypsin per ml. red cells. Preincubation, in ananalogous manner, of reticulocyte-rich red cellsand potassium periodate (KIO3, which alsoinactivates sialic acid and related substances (37),caused a moderate (27 per cent) diminution ofsubsequent iron uptake by the reticulocytes fromFe"5 IBP. However, the concentration of KI04(10-3 M) required to accomplish this inhibitionwas very close to that concentration (2 x 103M) causing hemolysis.

176


Comparisons between the uptake of Fe59 by ratreticulocytes and rat liver slices

The mechanism of uptake of iron from IBPby a nonerythroid, iron-requiring tissue was ex-plored by utilizing rat liver slices. Livers, freshlyobtained from decapitated Lobund rats, were cutwith a Stadie-Riggs microtome into slices weigh-ing 50 to 100 mg. The weighed slices werethen placed in groups of five in iron-free flaskscontaining 2 ml. of Krebs-Ringer-phosphate so-lution to which glucose had been added. Afterequilibration by agitation at 370 C. under 95per cent oxygen and 5 per cent CO2, 1 ml. vol-umes of autologous rat serum which had beenhalf-saturated with Fe59 and diluted with salinewere added to each flask. Appropriately, at vari-ous time intervals, the supernate of each selectedflask was decanted and saved, and its liver sliceswere rinsed 10 times in 50 volumes of saline.The Fe59 activities of each supernate and of the

3.5

3.0

RETI

2.5

2.0.

0

i .

o 1.0/0

.5'

corresponding slices were then determined andexpressed as the ratio of the Fe59 concentrationper gram of liver slice to that per milliliter ofmedium.

As shown in Figure 11, rat liver slices accumu-lated Fe59 at a steady rate during six hours' in-cubation. However, in this system a concentra-tion of Fe59 by liver slices exceeding that of themedium was only manifest after several hourswhen the Fe59 serum was present in relativelysmall amounts (< 0.1 ml. serum, half-saturatedwith Fe59, per flask). Fe59 accumulated by ratliver slices did not leak out if the slices were subse-quently incubated with rat serum half-saturatedwith nonradioactive iron. On the other hand,high concentrations of human IBP (50 juM) re-moved two-thirds of the Fe59 from rat liverslices within three hours, although this concentra-tion of IBP failed to remove Fe59 from the stromaof rat reticulocytes. Preliminary boiling of rat

ICU LOCYTES

LIVER SLICES

0 10 20 30 40 50 60 70 80 90 100

IBP SATURATION, PER CENT

FIG. 12. THE EFFECT OF THE IRON SATURATION OF HUMANIBP ON THE UPTAKEOF FE BYRAT RETICULOCYTESANDBY RAT LIVER SLICES

As with human reticulocytes, the uptake of Fe' from IBP by reticulocytes approached a plateauat an iron-binding protein saturation of 20 or 30 per cent, and a further sharp increase occurredat saturations above 75 per cent. In contrast rat liver slices took up comparatively small amountsof Fetm until high saturations (above 62 per cent) were reached, when Fe' adsorption sharplyincreased.

177


liver slices for five minutes not only failed todiminish the subsequent accumulation of Fe59from plasma by liver slices, but actually aug-mented it moderately. Boiling of rat reticulo-cytes, on the other hand, virtually abolished theircapacity to take up iron.

A comparison between the uptake of iron byreticulocytes and by liver at various iron satura-tions of IBP was made by separately incubatingautologous rat reticulocytes and liver slices, re-spectively, with purified human IBP which hadbeen incubated with various amounts of Fe59Cl.The reticulocytes and liver slices were taken froma rat which had been bled 20 ml. per Kg. bodyweight three and four days before sacrifice. Asdepicted in Figure 12, rat reticulocytes resembledhuman reticulocytes in that the Fe59 uptaketended to level off when the iron saturation of theIBP preparation exceeded 20 to 30 per cent,only to rise further as full saturation was ap-proached. With liver slices, on the other hand,Fe59 uptake increased slightly in a linear rela-tionship with IBP saturation until, again, heavydeposition of Fe59 appeared at high saturations.It is apparent from this figure that at low IBPsaturations the relative capacity of reticulocytes totake up iron was great, as compared to liver slices,whereas at high saturations of IBP the relativeability of liver slices to take up iron increasedsharply.

The uptake of cobalt by mature and immaturered cells

Cobalt, like most multivalent metals, combineswith the membranes of normal washed red cells(30). Unlike iron or chromium, however, co-balt (as saline solutions of Co60C12 or Co60Cl5)was invariably taken up more avidly by reticulo-cytes than by adult cells. The red cells of severalpatients with increased reticulocytes were centri-fuged and samples were removed from the top frac-tions (which contain most of the reticulocytes),the middle fractions, and the bottom fractions(said to contain the oldest cells [38] ). In TableIV, it is apparent that the most immature redcells took up Co++ most extensively, and thatCo++ uptake was roughly inversely proportionalto cell age. Cobalt oxidized to Co... by thepresence of 10-2 M H202 was similarly accumu-

TABLE IV

The uptake of Co++ by mature and immature red cells

Type of Centrifuged Reticulo- Conc. Cobaltred cells fraction cytes CoGOCl2 uptake

pg./Ioo ml.% ApM red cells

Pernicious Top 59.1 30 20.48anemia Middle 1.6 10.87

Bottom 0.8 5.36


Bottom 0.4 4.77


Bottom 0.5 3.87

Acquired Top 39.9 10 13.17hemolytic Middle 26.1 8.73anemia Bottom 17.5 4.60

lated preferentially from a saline solution by im-mature red cells. In suspensions of reticulocytesand of adult red cells, the uptake of Co+++ was 1.47and 1.38 times, respectively, that of Co++, followingthe rule that the more highly charged cations aremore extensively bound by red cell membranes(30).

Co+++ reacts with IBP to form an amber-coloredcompound with a light absorptive maximum at402 m~k (6). This compound is stable as longas the cobalt is maintained in the trivalent stateby a suitable oxidant, and unlike Cuw and Zn",Cot++ is not displaced from IBP by Fe++. WhenCo60 IBP was incubated with adult red cells,little Co60 became attached to the red cell mem-brane until the IBP was more than half-satur-ated. When the cobalt saturation of IBP wasless than 50 per cent, however, the Co60 uptakeof reticulocyte-rich red cell suspensions was from5- to 12-fold as great as that of adult red cells.It is to be noted that this was an even greater de-gree of preferential binding by reticulocytes thanoccurred with ionized cobalt. However, the up-take of Co60 from Co60 IBP by reticulocytes wasconsiderably less than that of Fe59 from compar-able preparations of Fe59 IBP, averaging aboutone-tenth as much.

Colorimetric interference by serum made itdifficult to determine by spectrophotometrywhether or not Co+++ in the presence of H202formed a colored complex with the IBP or iron-deficient serum. However, with several speci-mens of human and of rat sera it was possible

178


to show that CoCl., added serially to plasma in amanner comparable to that employed in the de-termination of the serum iron-binding capacity(27), produced small increments in light absorp-tion at a wavelength of 400 mJu. The erythro-poietic effect of Co IBP in vivo was studied infour rats. Each was injected intravenously twiceweekly for eight weeks with 2.5 ml. volumes ofrat plasma which had been freshly incubated forone hour in the presence of H202 and sufficientCoCl2 (10 pug.) theoretically to saturate fully theplasma IBP. No hematologic change occurred inrats injected with small amounts of cobalt in thisform.

DISCUSSION

In common with most multivalent metals (30),iron in the free cationic form is adsorbed ontothe membranes of either mature or immature redcells. This adsorption is injurious to red cellsand may lead to their agglutination and even tohemolysis (30). Although both mature and im-mature red cells may take up free Fe+++, only theimmature cells have the capacity to incorporatethis membrane-bound iron into heme. The non-specific capacity of red cell membranes to absorbcations such as iron presumably explains an ob-servation by Ellis, Brandt and Thacker (39)that more iron was taken up in vitro by washedthan by unwashed reticulocytes with which se-rum binding of iron would be in competition.

The mature red cells from patients with ac-quired hemolytic anemia whose red cells are ag-glutinable by Coombs serum, as well as red cellssensitized with incomplete anti-D antibodies, takeup less free Fe.. than do normal red cells, sug-gesting that those chemical groups on the redcell membrane reacting with Fe+++ are partiallyblocked by such protein "coatings."

In contrast to free cationic iron, iron boundto human serum IBP was readily utilized byimmature red cells but was completely unavail-able to mature red cells. In this sense the irontransporting protein of plasma is involved in aremarkable mechanism for restraining iron fromreacting with, and perhaps causing injury to, cellsnot requiring iron while making iron available tocells which utilize iron. In comparing several sub-stances with IBP in terms of the selectivity withwhich iron bound to them was utilized by reticulo-

cytes and not by mature red cells, albumin wasfound to be much less selective, and gammaglobu-lin and EDTAdid not function in selective trans-port capacity at all.

The transfer of iron from IBP to reticulocytesis not species specific. Indeed, it is not even classspecific, since human reticulocytes readily utilizediron bound to conalbumin, the iron-binding proteinof egg white, although conalbumin does inhibit ironutilization by bacteria (9). Conversely, a recentreport by Paoletti (40) indicates that avian redcells are able to utilize iron from a preparationof human IBP. These observations indicatethat the structure of the protein bearing the iron-chelating groups of IBP is not critical to thetransfer of iron to immature red cells. On theother hand, the iron-chelating rings of IBP andof conalbumin are believed to be very similar(12).

The transfer of iron from IBP to immature redcells did not approach a maximum until the ironsaturation of the protein reached 20 or 25 percent, and nonspecific "unloading" of iron fromIBP on to mature red cells began to occur atsaturations above 60 per cent. It is thereforeof interest that these are approximately the lowerand upper limits, respectively, of normal IBPsaturation in man. Thus, it is tempting to con-clude that these homeostatic limits are preservedbecause below a saturation of 20 per cent IBPcompetes successfully with the marrow for iron,while above 60 per cent saturation iron is un-loaded onto tissues having comparatively lessaffinity for iron. Experimental support for sucha mechanism in the regulation of serum iron levelsis provided in the comparative study of iron up-take by rat reticulocytes and rat liver slices. Itis apparent in Figure 12 that at high iron satura-tions of IBP there is a sharp increase in the pro-portion of iron deposited in liver as compared toreticulocytes. The inability of liver slices, ascompared to reticulocytes, to take up iron at lowIBP saturations may explain the observation ofBeutler (41) that rat liver cytochrome C levelsare more strikingly reduced during mild or mod-est iron deficiency than are hemoglobin levels.

Since the total IBP concentration of serum in-creases during iron deficiency as well as duringpregnancy (42), with consequent lowering of thepercentage iron saturation, the resultant more

179

L. SIMMONS, AND D. W. ALLEN

active competition by the more unsaturated IBPfor iron (Figure 5) might be considered disad-vantageous to the immature red cells. A similarsituation pertains to serum thyroxine-binding pro-tein, the levels of which increase in myxedemaand in pregnancy (43). Thus, this transportprotein also appears to compete with tissues forthe transported metabolite (44). It is, of course,possible that the levels of these two proteins arehomeostatically adjusted primarily with respectto achieving maximum loading rather than un-loading of iron and thyroxine, respectively.

The reticulocyte membrane acts in a role inter-mediate between the stage of plasma transportand heme synthesis, as suggested by Walsh andhis associates (15), and may be considered thesite of an iron-trapping mechanism. It is clearthat IBP, with a biological half-life of at least12 days (45), is not consumed during the cellularutilization of iron, which normally is half-clearedfrom the plasma in less than two hours. Al-though theoretically iron may be released fromIBP by reduction, a drop in pH, or removal ofHCO8-, it is also clear from the studies withEDTAreported here that, in transfer, iron is notreleased from IBP for any appreciable time ordistance as a free ion but is directly and inti-mately transferred from IBP to the reticulocytemembrane. No evidence was obtained that IBPitself entered or became attached to the surfacesof the reticulocytes, although a momentary phys-ical interaction of this sort would be hard to ex-clude. It is probably more reasonable to assumethat IBP molecules in close proximity to the re-ticulocyte membrane relinquish iron to specificiron-binding receptors on the cell surface. Con-sistent with this explanation is the finding of acompetition between IBP and reticulocytes foriron, in vitro, and the striking inability of reticulo-cytes to take up iron from IBP after these cellshad been exposed to very small concentrations oftrypsin and chymotrypsin. It is not possible toassign any specificity to this action of trypsinexcept to note that extremely low concentrationsof trypsin were effective, and that the respira-tory and glycolytic activities of the reticulocyteswere unaffected by these concentrations. Al-though trypsin readily lyses the neuraminic acidgroup of substances, such as occur prominentlyon red cell surfaces (34, 46), the comparatively

weak effect of "receptor-destroying enzyme" andof periodate in inhibiting iron uptake by reticulo-cytes indicates that the iron receptors of redcells are not identical with those red cell receptorswhich react with many viruses and which con-sist primarily of n-acetylneuraminic (sialic) acid(34, 47, 48). The inability to remove more than5 or 6 per cent of the Fe59 from reticulocytemembranes with high concentrations of IBP orEDTA is presumably attributable to the rapidincorporation of membrane iron into undissociablecomplexes.

It is difficult to say whether immature red cellsdiffer from mature red cells in their ability totrap iron because immature red cells possessunique receptors or because they alone are ableto metabolize membrane iron further, leavingsuch receptors continually unsaturated. Tishkoff,Robscheit-Robbins and Whipple have shown thatthe red cell stroma of dogs contains 12.4 jug. ofiron per ml. red cells and that most of this ishemin iron (49). Human mature red cell stromawas found by Tishkoff normally to contain lessthan 1 ug. of nonhemin iron per ml. red cells, al-though in certain patients with erythroblastemiastromal iron levels were distinctly higher (50).The concentration of iron in the stroma of im-mature red cells at various stages of developmentis not known, but it is known by optical micro-scopy and, recently, by electron microscopy (21,22) that the stroma of immature red cells, includ-ing reticulocytes, normally contains numerousiron-rich granules. Many of these granules havethe aspect of ferritin (1) by electron microscopy(21, 22). It thus seems reasonable to supposethat in its stromal phase iron may exist in partin the form of ferritin, although, using chemicalmethods, we have been unable to demonstratethis. Nevertheless, it is probable that the pro-portion of iron localizing in the reticulocyte mem-branes in the experiments reported here far ex-ceeds that which occurs in vivo. Whereas themechanism for trapping iron remains relativelyintact for several hours in vitro, it is presumedthat the mechanisms for metabolizing the ironfurther partially break down.

Certain bacteria and smut fungi which re-quire iron for the synthesis of cytochrome C andother iron-containing substances produce organiccompounds with a strong affinity for iron when

180 J. H. JANDL, J. K. INMAN, R.


grown in an iron-deficient medium 10 (51). Itis believed that these organic compounds serveas carriers for the transfer of ferric ions fromthe medium into the bacterial cell for use in hemesynthesis (51) and are analogous to the mineral-dissolving organic compounds excreted by cer-tain higher plants capable of growing on rocks(52). Neilands (51) found several compoundswhich, when added to simple media, served torender iron available for bacterial utilization;these included citric acid and fructose-phenylala-nine. The function of the latter in this respectis of interest, since red cell membranes containketose-amino acids (46) and since Borsook andhis associates (53, 54) have reported that ketose-amino acids stimulate the incorporation of aminoacids into the proteins of reticulocytes when addedin the presence of iron. In order to examine thepossibility that ketose-amino acids participate inthe transfer of iron from IBP to immature hu-man red cells, fructose-phenylalanine, fructose-glycine and fructose-glutamine were prepared bythe method of Abrams, Lowy and Borsook (55).The effect of various concentrations of these sub-stances on Fe59 uptake from Fe59 IBP by suspen-sions of mature red cells, of immature red cellsand of trypsinized immature red cells was thenassessed. In no instance was there an accelera-tion of iron uptake or of iron incorporation intohemoglobin by these cells. These negative find-ings fail to support, but do not exclude, the pos-sibility that such compounds, in proper stericarrangement, may function as iron-binding or iron-carrying components of the membranes of im-mature red cells. It is reasonable to suppose thatthe iron-binding substances manufactured by iron-deficient bacteria and fungi are designed merelyto render inorganic iron in the surrounding me-dium soluble and diffusible, and that the evolutionin animals of cell receptors capable of directlyremoving iron from IBP supplanted that mecha-nism. Bacteria, on the other hand, ordinarilyhave no need for such receptors, and indeed do

10 Interestingly, the smut fungus, U. sphaerogena, be-comes pale and releases iron-binding compounds intothe medium when the iron concentration of the mediumfalls to levels similar to those in the sera of iron-deficientpatients (25 ltg. per cent); these compounds disappearwhen the iron concentration is raised to that of normalsera (>100 jsg. per cent) (51).

not possess them, as indicated by their inabilityto utilize conalbumin-bound iron (9).

Since the uptake of Fe59 from Fe59 IBP wasonly moderately diminished by anoxia in vitro,it is improbable that iron transfer is directly de-pendent upon oxidative mechanisms. However,the effect of temperature upon iron uptake by re-ticulocytes, with striking inhibition of uptake attemperatures only a few degrees above or below370 C., is suggestive of an active metabolic mecha-nism.'1 This conclusion is supported by the find-ing that iron uptake was stimulated by the pres-ence of glucose and was readily inhibited by a num-ber of agents which block intermediary carbohy-drate metabolism, including cyanide, azide, fluoride,arsenate, 2,4-dinitrophenol and 2,4-dinitrophenyl-phenol. Immature red cells are well known to bemetabolically active, and reticulocytes have beenshown to possess numerous enzymes, includingvarious elements of the tricarboxylic acid cycle(57), and these are situated at least in part in thecell membrane (57, 58). Since hemolysis wasfound here to block the transfer of Fe59 from Fe59IBP to reticulocytes, it is of interest that Hofmannand Rapoport (59) found that the respiration ofreticulocytes in vitro was considerably reduced byhemolysis, probably by disruption of the tricar-boxylic acid cycle. One may reasonably con-clude that the uptake of iron by immature red cellsis dependent at some stage upon metabolic en-ergy, although this energy may not be directlyrequired for the actual transfer of iron fromIBP. In its dependence upon metabolic energy,iron uptake from IBP by immature red cellsdiffers from that by rat liver slices. Saltman,Fiskin, Bellinger and Alex (60) have reportedthat rat liver slices accumulate iron against aconcentration gradient, and that this accumulationis unaffected by metabolic inhibitors. However,the iron employed in those studies was not boundto IBP, and the findings are more analogous tothose described here with red cells in which un-bound iron was used.

The fact that appropriate concentrations of

"Although calculation of the heat of activation of acomplex and incompletely-understood reaction as is in-volved here is very crude, the value for heat of activationcalculated from the data in Figure 6 is within the rangeencountered in "active metabolic" or "enzymatic" reac-tions (56).

181

1. H. JANDL, J. X. INMAN, R. L. SIMMONS, AND D. W. ALLEN

lead block iron incorporation into the hemoglobinof reticulocytes without blocking the cellular up-take of iron from IBP indicates that the rate oftransfer of iron from IBP is not directly de-pendent upon the rate of heme synthesis. Fur-thermore, these studies indicate that under theseconditions lead-poisoned reticulocytes tend to ac-cumulate nonhemoglobin iron on the cell mem-branes. Direct electron-microscopic evidence ofiron accumulation in the stroma of immature redcells in lead poisoning has recently been reportedby Bessis and Breton-Gorius (61). The findinghere that blocking heme synthesis with leadcaused an accumulation of iron in the reticulo-cyte membrane rather than in the water-solublephase of the cell interior also indicates that,whereas the uptake of iron by the immature redcells may not be dependent upon heme synthesis,the release of iron from the membrane into thecell interior may be so dependent.

The failure of anoxia to stimulate iron uptakeby reticulocytes in vitro (indeed its mild inhibi-tory effect) is in accord with observations ofothers that the maturation of immature red cells(62) and their incorporation of iron into heme(63) are somewhat inhibited, rather than stimu-lated, by anoxia. Thus, in terms of the ability ofimmature red cells to take up iron, to incorpor-ate iron into heme and to mature morphologically,there is no evidence of a direct anoxic stimulus tored cell metabolism. The possibility that therate of uptake of iron from IBP by reticulocytesin vitro is influenced by the presence of a plasmaerythropoietic factor seems theoretically remote,since, in rabbits, this factor appears to act bystimulating cell division rather than the rate ofmaturation or hemoglobin metabolism (64). In-deed, it is characteristic of the active erythro-poietic response to acute anoxia that the increaseof red cells exceeds that of hemoglobin. Never-theless, in an attempt to explore this possibility,the uptake by rabbit reticulocytes of Fe59 boundto rabbit plasma was measured in the presence oflarge volumes of plasma drawn before, and twodays after, strenuous bleeding of the same rab-bits. As determined by the Fe59 uptake andmeasurement of the iron content of the plasmas,there was no difference in the iron uptake byreticulocytes in the presence of "anemic" plasmaas compared to "nonanemic" plasma.

To recapitulate the above, it is suggested thatthe utilization of iron by immature red cells canbe separated into three phases: 1) Iron boundto IBP is transferred by competitive binding tospecific receptors on the cell surface. These re-ceptors may be effaced by certain proteolytic en-zymes. 2) The functioning of these receptors isdependent upon metabolic energy, possibly forsynthesizing receptors or for facilitating the re-lease of iron from the receptors into the cellinterior. This phase may be blocked by variousinhibitors of oxidative metabolism. 3) Iron isthen incorporated into hemoglobin, presumablyunder the control of the microsomes and mito-chondria of the cytoplasmic reticulum.'2 Thisstage may be blocked by low concentrations oflead.

It is of great interest that trivalent cobalt isbound by IBP in a manner comparable to iron.Cobalt, like iron, is capable of forming reversiblecombinations with molecular oxygen, when in thecobaltous state and linked to histidine (66). Itsposition in the porphyrin ring of the vitamin B12molecule is analogous to that of iron in heme.Finally, it has the unique but still obscure capacityto stimulate erythropoiesis. It was found possibleto substitute cobalt for iron in the interaction be-tween IBP and the reticulocyte surface. How-ever, the highly oxidant milieu needed for chela-tion of cobalt by IBP probably does not existphysiologically. Nor was it possible to producepolycythemia in rats repeatedly given cobalt com-bined with rat IBP in vitro. Finally, the factthat free Co++, unlike iron and chromium, is it-self taken up by immature red cells with greateravidity than by adult red cells may be more per-tinent to the mechanism of cobalt polycythemiathan the reactions between cobalt and IBP.

SUMMARY

Iron bound to purified human iron-binding pro-tein (IBP) or to conalbumin, the iron-binding

12Although the nucleated red cells of birds have beenwidely employed in studies of heme synthesis, it has beenclear for some time that the basophilic reticulum of thecytoplasm of immature red cells, not the nucleus, is thesine quo non of hemoglobin formation (65). As shownby Jensen and his associates (19) suspensions of avianred cells synthesize hemoglobin in vitro only in propor-tion to the concentration of reticulum-containing cells.

182


protein of egg white and of avian plasma, isreadily and preferentially transferred to immature,but not to mature, mammalian red cells. Irontransfer is dependent upon the extent of saturationof IBP, and there exists a competition for ironbetween IBP and iron-binding receptors on thereticulocyte membrane. These cellular receptorsare destroyed by certain proteolytic enzymes.Iron appears to be directly transferred to thesereceptors without existing in a free form; andiron once bound to reticulocyte membranes is notexchangeable with free iron and cannot be elutedwith high concentrations of IBP or of ethylene-diaminetetraacetic acid (EDTA).

The transfer of IBP-iron to immature red cells,unlike the adsorption of free iron, is related tometabolic energy-producing processes. Thus, itis accelerated by glucose and oxygen and is dimin-ished or blocked by low concentrations of meta-bolic inhibitors, by nonphysiologic temperaturesand by hemolysis. The uptake of iron is notprimarily linked to heme synthesis. Appropriateconcentrations of Pb++ blocked iron incorporationinto heme by reticulocytes without reducing irontransfer, thus causing a pile-up of iron on the cellmembranes.

Unlike reticulocytes, rat liver slices take up ironin appreciable amounts only at high IBP satura-tions. The transfer of iron from IBP to liverslices is not dependent upon metabolic energy andiron accumulated by liver slices can be elutedwith unsaturated IBP.

Although free iron is usually bound equally bymature and immature red cells, free cobalt ismore avidly bound by the immature cells. Triva-lent cobalt, which forms colored chelates withpurified IBP as does iron, also is bound by theIBP of whole plasma. Cobalt bound to IBPcan, to a limited extent, be transferred prefer-entially to reticulocytes. However, the combina-tion of Co+++ and IBP probably does not occurunder physiologic conditions and is probably notinvolved in the induction of cobalt polycythemia.

REFERENCES

1. Farrant, J. L. An electron microscopic study of fer-ritin. Biochim. biophys. Acta 1954, 13, 569.

2. Granick, S. Iron metabolism. Bull. N. Y. Acad.Med. 1954, 30, 81.

3. Mazur, A., Baez, S., and Shorr, E. The mechanismof iron release from ferritin as related to its bio-logical properties. J. biol. Chem. 1955, 213, 147.

4. Holmberg, C. G., and Laurell, C. B. Investigationsin serum copper. I. Nature of serum copper to theiron-binding protein in human serum. Acta chem.scand. 1947, 1, 944.

5. Surgenor, D. M., Koechlin, B. A., and Strong, L. E.Chemical, clinical and immunological studies onthe products of human plasma fractionation.XXXVII. The metal-combining globulin of humanplasma. J. clin. Invest. 1949, 28, 73.

6. Inman, J. K. Studies on human p1l-metal-combiningprotein. Thesis presented to the Division of Medi-cal Sciences, Harvard University, 1956.

7. Schade, A. L., Reinhart, R. W., and Levy, H. Carbondioxide and oxygen in complex formation with ironand siderophilin, the iron-binding component ofhuman plasma (1, 2, 3). Arch. Biochem. 1949, 20,170.

8. Koechlin, B. A. Preparation and properties of se-rum and plasma proteins. XXVIII. The fi-metal-combining protein of human plasma. J. Amer.chem. Soc. 1952, 74, 2649.

9. Schade, A. L., and Caroline, L. Raw hen egg whiteand the role of iron in growth inhibition of Shigelladysenteriae, Staphylococcus aureus, Escherichiacoli and Saccharomyces cerevisiae. Science 1944,100, 14.

10. Schade, A. L., and Caroline, L. An iron-binding com-ponent in human blood plasma. Science 1946, 104,340.

11. Warner, R. C., and Weber, I. The metal combiningproperties of conalbumin. J. Amer. chem. Soc.1953, 75, 5094.

12. Warner, R. C. The metal chelate compounds of pro-teins. Trans. N. Y. Acad. Sci. 1954, 16, 182.

13. London, I. M., Shemin, D., and Rittenberg, D. Thein vitro synthesis of heme in the human red bloodcell of sickle cell anemia. J. biol. Chem. 1948, 173,797.

14. Shemin, D., London, I. M., and Rittenberg, D. Thein vitro synthesis of heme from glycine by the nu-cleated red blood cell. J. biol. Chem. 1948, 173,799.

15. Walsh, R. J., Thomas, E. D., Chow, S. K., Fluharty,R. G., and Finch, C. A. Iron metabolism: Hemesynthesis in vitro by immature erythrocytes. Sci-ence 1949, 110, 396.

16. London, I. M., Shemin, D., and Rittenberg, D. Syn-thesis of heme in vitro by the immature non-nucleated mammalian erythrocyte. J. biol. Chem.1950, 183, 749.

17. Sharpe, L. M., Krishman, P. S., and Klein, J. R.Uptake by duck erythrocytes of iron added toblood. Arch. Biochem. 1952, 35, 409.

18. London, I. M., and Yamasaki, M. Heme synthesisin non-intact mammalian and avian erythrocytes.Fed. Proc. 1952, 11, 250.

183


19. Jensen, W. N., Ashenbrucker, H., Cartwright, G. E.,and Wintrobe, M. M. The uptake in vitro of ra-dioactive iron by avian erythrocytes. J. Lab. clin.Med. 1953, 42, 833.

20. Goldberg, A., Ashenbrucker, H., Cartwright, G. E.,and Wintrobe, M. M. Studies on the biosynthesisof heme in vitro by avian erythrocytes. Blood 1956,11, 821.

21. Bessis, M. C., and Breton-Gorius, J. Iron particlesin normal erythroblasts and normal and pathologi-cal erythrocytes. J. biophys. biochem. Cytol. 1957,3, 503.

22. Bessis, M., and Breton-Gorius, J. Le cycle du ferdans l'organisme revele par le microscope elec-tronique. Sem. H6p. Path. et Biol. 1957, No. XIX,2173. (Ann. Rech. med. 1957, No. 8, 829.)

23. Jandl, J. H., Inman, J. K., and Simmons, R. L.Transfer of iron and cobalt from serum iron-bind-ing protein to human reticulocytes. Clin. Res.Proc. 1957, 5, 144.

24. Inman, J. K. Preparation of crystalline beta.-metal-combining protein in Proc. 10th Conference on thePlasma Proteins and Cellular Elements of theBlood, Jan. 5-6, 1956, sponsored by the ProteinFoundation, Inc., Cambridge, Mass., p. 46.

25. Bothwell, T. H., Hurtado, A. V., Donohue, D. M.,and Finch, C. A. Erythrokinetics. IV. The plasmairon turnover as a measure of erythropoiesis.Blood 1957, 12, 409.

26. Ham, T. H. A Syllabus of Laboratory Examinationsin Clinical Diagnosis. Cambridge, Mass., HarvardUniversity Press, 1950.

27. Rath, C. E., and Finch, C. A. Chemical, clinical, andimmunological studies on the products of humanplasma fractionation. XXXVIII. Serum iron trans-port. Measurement of iron binding capacity of se-rum in man. J. clin. Invest. 1949, 28, 79.

28. Kitzes, G., Elvehjem, C. A., and Schuette, H. A.The determination of blood plasma iron. J. biol.Chem. 1944, 155, 653.

29. Fischer, H. Hemin in Organic Synthesis, E. C.Horning, Ed. New York, John Wiley and Sons,Inc., 1955, coll. vol. 3, pp. 442-43.

30. Jandl, J. H., and Simmons, R. L. The agglutinationand sensitization of red cells by metallic cations:Interactions between multivalent metals and thered-cell membrane. Brit. J. Haemat. 1957, 3, 19.

31. Folch, J., Ascoli, I., Lees, M., Meath, J. A., and Le-Baron, F. N. Preparation of lipide extracts frombrain tissue. J. biol. Chem. 1951, 191, 833.

32. Watson, C. J. The eryth1rocyte coproporphyrin.Arch. intern. Med. 1950, 86, 797.

33. Eriksen, L. Lead intoxication. I. The effect of leadon the in vitro biosynthesis of heme and free eryth-rocyte porphyrins. Scand. J. clin. Lab. Invest.1955, 7, 80.

34. Klenk, E., and Lempfrid, H. tUber die Natur der Zell-receptoren fur das Influenza-virus. Z. physiol.Chem. 1957, 307, 278.

35. Ada, G. L., and French, E. L. Purification of the re-ceptor destroying enzyme of V. cholerae. Aust. J.Sci. 1950, 13, 82.

36. Burnet, F. M., McCrea, J. F., and Stone, J. D.Modification of human red cells by virus action;receptor gradient for virus action in human redcells. Brit. J. exp. Path. 1946, 27, 228.

37. Burnet, F. M. Principles of Animal Virology. NewYork, Academic Press Inc., 1955.

38. Borun, E. R., Figueroa, W. G., and Perry, S. M.The distribution of Fe' tagged human erythrocytesin centrifuged specimens as a function of cell age.J. clin. Invest. 1957, 36, 676.

39. Ellis, G. H., Brandt, C. S., and Thacker, E. J. Fac-tors influencing the uptake of iron by blood and bybone marrow cells in vitro. Science 1954, 119, 94.

40. Paoletti, C. R6le des 8-globulines plasmatiques dansle transport du fer utilise par les cellules erythro-formatrices. C. R. Acad. Sci. (Paris) 1957, 245,377.

41. Beutler, E. Iron enzymes in iron deficiency. I.Cytochrome C. Amer. J. med. Sci. 1957, 234, 517.

42. Laurell, C. B. Studies on the transportation andmetabolism of iron in the body with special refer-ence to the iron-binding component in humanplasma. Acta physiol. scand. 1947, 14, Suppl. 46, 1.

43. Dowling, J. T., Freinkel, N., and Ingbar, S. H. Thy-roxine-binding by sera of pregnant women, new-born infants, and women with spontaneous abor-tion. J. clin. Invest. 1956, 35, 1263.

44. Freinkel, N., Ingbar, S. H., and Dowling, J. T. Theinfluence of extracellular thyroxine-binding pro-tein upon the accumulation of thyroxine by tissueslices. J. clin. Invest. 1957, 36, 25.

45. Gitlin, D., Janeway, C. A., and Farr, L. E. Studieson the metabolism of plasma proteins in the ne-phrotic syndrome. I. Albumin, 'y-globulin andiron-binding globulin. J. clin. Invest. 1956, 35,44.

46. Yamakawa, T., and Suzuki, S. The chemistry of thelipids of posthemolytic residue or stroma of erythro-cytes. II. On the structure of hemataminic acid.J. Biochem. (Tokyo) 1952, 39, 175.

47. Gottschalk, A. Neuraminic acid; the functional groupof some biologically active mucoproteins. Yale J.Biol. Med. 1956, 28, 525.

48. Howe, C., Rose, H. M., and Schneider, L. Enzymicaction of influenza virus on human erythrocytestroma components. Proc. Soc. exp. Biol. (N. Y.)1957, 96, 88.

49. Tishkoff, G. H., Robscheit-Robbins, F. S., andWhipple, G. H. Red cell stroma in dogs. Varia-tions in the stroma protein and lipid fractions re-lated to experimental conditions. Blood 1953, 8,459.

50. Tishkoff, G. H. Personal communication.51. Neilands, J. B. Some aspects of microbial iron me-

tabolism. Bact. Rev. 1957, 21, 101.52. Hutner, S. H., Provasoli, L., Schatz, H., and Haskins,

C. P. Some- approaches to the study of the role

184


of metals in the metabolism of microorganisms.Proc. Amer. phil. Soc. 1950, 94, 152.

53. Borsook, H., Abrams, A., and Lowy, P. H. Fruc-tose-amino acids in liver: Stimuli of amino acidincorporation in vitro. J. biol. Chem. 1955, 215,111.

54. Kruh, J., and Borsook, H. Hemoglobin synthesis inrabbit reticulocytes in vitro. J. biol. Chem. 1956,220, 905.

55. Abrams, A., Lowy, P. H., and Borsook, H. Prepara-tion of 1-amino-l-deoxy-2-keto hexoses from aldo-hexoses and a-amino acids. I. J. Amer. chem.Soc. 1955, 77, 4794.

56. Glasstone, S. Textbook of Physical Chemistry, 2nded. New York, D. Van Nostrand Co., Inc., 1946.

57. Rapoport, S., and Hofmann, E. C. G. Unter-suchungen fiber den Atmungsstoffwechsel vonReticulocyten. Biochem. Z. 1955, 326, 493.

58. Schweiger, H. G., Rapoport, S., and Sch6lzel, E.Role of non-protein nitrogen in the synthesis ofhaemoglobin in the reticulocyte in vitro. Nature1956, 178, 141.

59. Hofmann, E. C., and Rapoport, S. Der Einfluss derHkmolyse auf die Atmung von Reticulocyten.Biochem. Z. 1955, 326, 499.

60. Saltman, P., Fiskin, R. D., Bellinger, S. B., and Alex,T. The metabolism of iron by rat liver slices. Theeffect of chemical agents. J. biol. Chem. 1956, 220,751.

61. Bessis, M., and Breton-Gorius, J. Etude au micro-scope electronique du sang et des organes hemo-poietiques dans le saturnisme experimental. Sem.Hop. Path. et Biol. 1957, No. IV, 411. (Ann.Rech. med. 1957, No. 2, 151.)

62. Rosin, A., and Rachmilewitz, M. Studies on bonemarrow in vitro: III. The effect of anoxia andhyperoxia on explanted bone marrow. Blood 1948,3, 165.

63. Thomas, E. D. In vitro studies of erythropoiesis.II. The effect of anoxia on heme synthesis. Blood1955, 10, 612.

64. Erslev, A. J. The effect of anemic anoxia on thecellular development of nucleated red cells. Clin.Res. 1958, 6, 187.

65. Reimann, F. Retikulare Substanz und Hamoglobin-bildung. Schweiz. Z. Path. 1943, 6, 35.

66. Hearon, J. Z. The configuration of cobaltodihisti-dine and oxy-bis (cobaltodihistidine). J. nat. Can-cer Inst. 1948, 9, 1.

185

,81€¦ · j. h. jandl, j. k. inman, r. l. simmons, and d. w. allen iron in this material was...

Documents