Journal of Clinical InvestigationVol. 41, No. 5, 1962

AN INTRACELLULARPROTEIN INTERMEDIATE FOR HEMOGLOBINFORMATION*

By WILLIAM B. GREENOUGH,III, THEODOREPETERS, JR. ANDE. DONNALLTHOMAS

(From The Mary Imogene Bassett Hospital [affiliated with Columbia University],Cooperstoum, N. Y.)

(Submitted for publication November 27, 1961; accepted January 25, 1962)

Iron and amino acids enter young red cells andare incorporated into hemoglobin, but the path-way of iron after entering the erythroblast is notunderstood nor is it clear in what sequence iron,protoporphyrin, and globin join to form a com-plete hemoglobin molecule. Several lines of evi-dence have suggested the presence of an intracel-lular, nonheme iron protein that may be an inter-mediate in the formation of hemoglobin (1-10).This report describes a protein fraction from hu-man and dog marrow cells which appears fromisotopic studies to be such a precursor of hemo-globin. It does not contain heme, and it can bereadily differentiated from siderophilin and fer-ritin.

MATERIALS AND METHODS

General methods. Serum iron and iron-binding ca-pacity were determined by the method of Peters, Giovan-niello, Apt and Ross (11). Quantitation of hemoglobinsamples was done on a Beckman model DU spectro-photometer. Molecular extinction coefficients used were1.2 X 10' at 415 mg for oxy- or carbonmonoxy-hemoglo-bin and 1.08 X 10' at 415 my for cyanmethemoglobin.Other proteins were determined by extinction at 280 m.ior by a turbidometric method with cationic detergent andalkali (12). Heme was isolated by the method ofFischer (13) or by the method of Kassenaar, Morell andLondon (14).

Isotopes and counting. Uniformly labeled L-leucine-Cl'(4.15 me per mmole) was purchased from Merck andCo., Montreal; glycine-2-C1' (2.44 mc per mmole) fromNew England Nuclear Corp.; and Fe'-citrate (approx.500 me per mmole) from Abbott Laboratories.

Radioactive iron was counted in a well scintillationcounter (Baird-Atomic model 810 detector with 1.75 X 2inch crystal, modified to match Packard Autogamma

* This work was supported by Research Grants C2643and H2751 from the United States Public Health Serv-ice. Dr. Greenough was a Public Health Service Re-search Fellow of the National Cancer Institute. Pre-sented in part at the meeting of the American Society forClinical Investigation, Atlantic City, N. J., May, 1961.

spectrometer, series 410A). Efficiency for Fe' wasabout 0.12. C1' counting was performed on a Robinsonwindowless flow-gas counter (15) or on a Picker thin-window flow counter (2975A) with automatic samplechanger. Efficiency was about 20 per cent. C1' sampleswere plated at infinite thinness on platinum discs asliquids and dried under a heat lamp.

Marrow procurement. Marrow was obtained from theribs of hematologically normal human beings undergoingthoracotomy, or from the long bones of dogs 6 to 12months of age, and suspended in Hanks' solution con-taining autologous serum. The marrow was sievedthrough a stainless steel tea strainer and then passedthrough a stainless steel screen with openings 200 Assquare (16). This procedure resulted in a well dis-persed cell suspension suitable for accurate pipetting andcell counts. The marrow was processed and used within1 hour after it was removed from the body.

Incubation of intact marrow cells. Incubations werecarried out at 37.5° C in air with gentle shaking inErlenmeyer flasks with an air: liquid ratio greater than8: 1. Acid-cleaned, siliconed glassware was usedthroughout. The incubation medium was serum inHanks' solution (see legends to figures), buffered topH 7.6, and contained 10,000 U penicillin, 60 Ag strepto-mycin, and 5 mg glucose per ml. At the start of incu-bation, radioactive substrates were added. In the caseof iron the quantities added were always less than thebinding capacity of the serum present. Reactions werestopped by the addition of 1 per cent KCN and 2 percent KFe(CN), with rapid cooling to 00 C.

Preparation of hemoglobin solutions. After incubationthe cells were washed 3 times with 10 vol isotonic salineand centrifuged for 10 minutes at 240 G. The cells werelysed by exposure to 4 vol distilled water for 15 minutesat 40 C with vigorous shaking. The lysate was centri-fuged for 15 minutes at 1,500 G at 40 C, and the super-nate was then centrifuged for 30 minutes at 100,000 G at4' C. The resulting clear red "hemoglobin solution" waskept at 4' C until chromatographed, usually within 24hours.

Hemoglobin solutions stored for long intervals at 40 C,at room temperature, or frozen at -20' C gave alteredchromatograms (17). Dialysis of hemoglobin solutionswas associated with the formation of precipitates, andhence dialysis was not usually performed. In some in-stances brief dialysis was necessary to obtain adsorptionto the resin.

1116

INTRACELLULAR PROTEIN INTERMEDIATE FOR HEMOGLOBIN

Incubation of lysed marrow cells. Marrow cells werewashed twice in Hanks' solution by centrifuging for 10minutes at 375 G. Four vol of distilled water was addedto the cell sediment, and the mixture was shaken at 40 Cfor 15 minutes. Lysis was stopped by the addition of anappropriate volume of a fivefold concentrate of an intra-cellular buffer of pH 7.9. The composition of the dilutedbuffer in milliequivalents per liter was K+, 150; Mg++, 10;Cl-, 130; HCO,-, 30. The lysate was centrifuged for 15minutes at 1,500 G, and aliquots of the clear supernatewere pipetted'into incubation flasks. At the end of incu-bation, reactions were stopped with KCNand K3Fe (CN),as before, particulate matter was removed by centrifugingat 100,000 G for 30 minutes, and the samples were dialyzedovernight against distilled water prior to chromatography.

Chromatography on IRC-5O. Amberlite IRC-50, XE-64, was obtained from Rohm and Haas Co. IRC-50 wasprepared according to Hirs, Moore and Stein (18) andwas titrated from the acid side to near the desired pHwith NaOH. The resin was then washed with a 100 volexcess of distilled water and equilibrated with the startingbuffer until the pH was regulated to within 0.05 pH U ofpH 7.00. Hemoglobin binding varies greatly with verysmall pH changes between pH 6.8 and 7.2, and accurateequilibration of a batch of resin is critical. Columns(1 X 10 cm) were generally packed by gravity 1 to 2hours before use.

Weevolved a linear gradient elution system using twoidentical 500-ml cylindrical bottles connected at the bottomby polyethylene tubing. The starting buffer was 0.0067 Msodium phosphate with 0.0033 M KCN at pH 7.0. Thelimiting buffer was 0.5 M NaCl dissolved in the startingbuffer. Flow rates of 1 to 2 ml per cm2 per hour wereused. The rate of increase in ionic strength was regu-lated by the volumes placed in the reservoirs, as well asby the flow rates.

Initial investigation of rates of incorporation of la-beled amino acids and iron into the various componentsseparated by IRC-50 columns showed no difference ex-cept for the fraction that elutes with the solvent front.This fraction (Fr I), termed V. by Huisman and Prins(19), showed a markedly different kinetic pattern andspecific activity. Accordingly, in many experiments FrI was separated with the starting buffer, and all othercomponents were eluted from the column by changing di-rectly to the limiting buffer.

Recoveries from IRC-50 columns were better than95 per cent, based on measurements of absorbance of 280and 415 miu. Fractions were collected on a volume-ac-tivated fraction collector or, in the case of the two-stagecolumns, by hand.

Chromatography on DEAE-cellulose. This system wasused for the further purification of Fr I. DEAE-cellu-lose (Selectacel, Brown Co., Berlin, N. H.) was alter-nately washed with 0.1 N HCl and 0.1 N NaOH threetimes, ending with an alkaline cycle, and then repeatedlywashed with distilled water to pH 7.0. A linear gradientof increasing ionic strength and decreasing pH was used,starting with 300 ml distilled water. The limiting buffer

was 300 ml 0.01 M acetate at pH 4.0 with 2 M NaCl.Columns of 40-cm length and 1-cm diameter were usedand were packed by gravity.

Immunochemical studies. Antibodies were made infemale rabbits against dog liver ferritin and dog Fr I.The ferritin was made according to Granick (20) andwas recrystallized three times. Fr I was purified byDEAE-cellulose chromatography as described above.Rabbits were injected with 9 mg of Fr I mixed withFreund's complete adjuvant (Difco-Bacto); 12 to 16 in-jection sites in the neck and sacrospinalis muscle groupwere used, each depot being less than 0.5 ml. Thirteenmg of ferritin was used in similar fashion. At 4 weeksthe rabbits were exsanguinated by heart puncture andthe serum harvested. Antibody activity was tested by aprecipitin ring test against dilutions of antigen. Theantiferritin antiserum produced precipitation with 2 ltgof ferritin per ml; the anti-Fr I with 23 img of Fr I perml. Antisiderophilin was a sample of goat antihumantransferrin, kindly provided by Hyland Laboratories.

Concentration of Fr I. This was done to some extentby dialysis against polyvinylpyrrolidone. Considerableloss due to the formation of an insoluble precipitate lim-ited efforts at concentration.

Electrophoresis. Electrophoresis was performed oncellulose acetate (Oxoid) in a Shandon Universal elec-trophoretic apparatus (Consolidated Laboratories, Inc.,Chicago Heights, Ill.) ; 5 X 12 to 18 cm strips were usedin a 0.005 Mveronal buffer, at pH 8.6, with 13 v per cm.After 2 hours the strips were fixed and stained withLissamine green SF for protein, as described by Brack-enridge (21).

RESULTS

Chromatography. Figure 1 ismatogram of a marrow cell lysate.

a typical chro-Fr I, although

FRACTION NUMBER

FIG. 1. IRC-50 CHROMATOGRAMOF NORMALHUMANMARROWLYSATE DEVELOPEDBY A LINEAR GRADIENT AT PH7.0. Details are given in Methods.

1117

W. B. GREENOUGH,III, T. PETERS, JR. AND E. D. THOMAS

ro-ideOwinoinsac-[ol-20to

ionant

heterogeneous, consisted mainly of nonheme pteins, while the subsequent fractions were maup of hemoglobins. After incubation of marryin vitro with Fe59- or with C"4-labeled amiacids, the relative activities of these hemoglobwere similar at all incubation times, while the -

tivity of Fr I was markedly higher. In the flowing discussion the hemoglobins (fractionsto 80) are considered collectively and referredas "Main Hb."

Kinetics. Figure 2 illustrates the incorporaterate of Fe59 into Fr I and Main Hb. A signific,-amount of Fe59 appeared in the Main Hb oi

z i1v4w0I-

E 8,000Nz

NN) 6a,000I-

zm0-) 4,000I-0

2,0001

TIME IN MINUTES

FIG. 2. INCORPORATIONOF FE' BY FR I AND MAIN HBWITH INCREASING TIME OF INCUBATION. Human ribmarrow (96 X 10o nucleated cells per ml) was incubatedwith Fe' (3.6 Ac per ml) in 50 per cent serum in Hanks'solution.

after a delay of 7 to 10 minutes. Thereafter hemo-globin synthesis proceeded at a linear rate for 4to 6 hours. There was no delay in the appearanceof radioactivity in Fr I, and its activity rosesteeply to peak values in 10 to 30 minutes.

Figure 3 shows the incorporation rate ofL-leucine-C14 into Fr I and Main Hb. The in-corporation of leucine-C14 into Fr I and Main Hbshowed the same time curves as that of Fe59.There was no significant difference between humanand dog marrow with respect to the kinetics ofFe59- or C'4-labeled amino acid incorporation.After incubation in vitro there was a wide varia-

TIME IN MINUTES

nly FIG. 3. INCORPORATIONOF C14-LEUCINE INTO FR I ANDMAIN HB. Human rib marrow (75 X 10 nucleated cellsper ml) was incubated with CG4-leucine (0.4 /Ac per ml)in 50 per cent serum in Hanks' solution.

tion from one marrow to another in the degreeof labeling of hemoglobin. This variation wasstill observed after correction for the number ofred cell precursors present and any differences inthe substrate. Different sera have widely differ-ing potency in promoting heme synthesis (16).For this reason a direct comparison between dif-ferent experiments is difficult.

Figure 4 illustrates the ratio of Fe59 incorpora-tion into Main Hb and Fr I at progressive timesin all experiments performed. Despite differentabsolute levels of synthesis, the ratios of Main

30. .~~~~~~~~~~~~~~

DOG... XmU)WdAN

3.01-

20RATIO OFTOTALRADIOACTIVITY

MAIN HB/FR.I

1.0

0

0

a

aIS I

a R

*I,I

I-S a a I a a

100 12020 40 60 80TIME IN MINUTES

FIG. 4. THE RATIO OF FE INCORPORATIONINTO MAINHB ANDFR I AT PROGRESSIVETIMES OF INCUBATION IN ALLEXPERIMENTSPERFORMED.

1118

a

INTRACELLULAR PROTEIN INTERMEDIATE FOR HEMOGLOBIN

z

400

Z FR.I3,000

z2,000. ~~~~~~FR. I

I-

0

i-LOOO MANHB

10 MIN. 4 HRS

FIG. 5. INCORPORATIONOF FE' INTO FR I AND MAINHB. Human rib marrow (133 X 10' nucleated cells perml) was incubated with Fe5" (1.06 ,uc per ml) in 33 percent serum in Hanks' solution. After 10 minutes the cellswere washed, resuspended in the same medium with un-labeled Fe, and reincubated for 4 hours.

Hb to Fr I were similar. In the first 10 minutesthe total Fe59 in Fr I was two to ten times that inthe Main Hb. The activities became equal afterabout 30 minutes, and thereafter there was a pro-gressive increase in that of Main Hb.

Effect of lead. Lead acetate (10-4 M) addedto either dog or human marrow incubationsblocked the incorporation of iron into Main Hb.There was no inhibition of Fe59 incorporation intoFr I. Lead added after peak values for Fr I hadbeen achieved produced no significant change inthe activity of Fr I, although Main Hb synthesiswas diminished.

Reincubation in unlabeled substrate. It seemedlikely on the basis of the kinetic pattern that Fr Icontained an iron compound, possibly an ironprotein, that was an intermediate for hemoglobinformation. The rate of total activity enteringFr I per unit time in the first minutes of reactionwas enough to explain all activity subsequentlyappearing in the hemoglobin fraction. To testthis possibility further the following experimentswere performed (Figures 5 and 6): aliquots of amarrow suspension were incubated for 10 minuteswith Fe59 bound to autologous serum protein orwith leucine-C04. At 10 minutes the incubation

was stopped by lowering the temperature to 40 C.The marrow cells were washed free of externalradioactivity at 40 C and reincubated for 4 hoursin the same concentration of unlabeled iron or un-labeled leucine as was present initially.

There was a fall in total Fe59 activity in Fr Iand a rise in activity in the Main Hb which, al-though less rapid than prior to cell washing, wasdefinite (Figure 5). The stroma, mitochondria,and microsomes also showed a drop in activityafter cell washing. That the Fe59 in Fr I did notfall further may be due to the large amount ofFe59 that became attached to the stroma in theseexperiments. It appeared possible that Fr I re-ceived iron from stroma and yielded it to hemo-globin (8). As long as the stromal iron pool washigh, a large drop in Fe59 in Fr I would not beexpected.

The C14 level of Fr I also fell, while the C14level of Main Hb increased (Figure 6). An activeturnover of proteins that are not on the pathwayof hemoglobin synthesis could account for a greaterdrop in C14 activity of Fr I than appeared in theMain Hb.

Marrow lysate experiments. In order to gainmore direct evidence that the active material inFr I is an intermediate for hemoglobin synthesis,

10 MIN. 4 HRS.RADIOACTIVE LEUCINE "COLD' LEUCINE

FIG. 6. INCORPORATIONOF C"-LEUCINE INTO FR I ANDMAIN HB. Human rib marrow (133 X 10' nucleated cellsper ml) was incubated with C"4-leucine (2.5 Atc per ml) in33 per cent serum in Hanks' solution. After 10 minutesthe cells were washed, resuspended in the same mediumwith unlabeled leucine, and reincubated for 4 hours.

1119

W. B. GREENOUGH,III, T. PETERS, JR. AND E. D. THOMAS

50,0001zidI

z

z

C)

0

SUBSTRATES

M C'4 L-LEUCINEAS FREE AMIOACID

EM C4 FR.! PREPAREDBY INCUBATION OFMARROWCELLS WITH

C14 L-LEUCINE

EXPERIMENTA EXPERIMENTB

FIG. 7. INCORPORATIONOF LABELEDFR I INTO MAIN HBBY A LYSATE OF DOGMARROWCELLS. In each experiment0.6 ,sc (270,000 cpm) C1'-leucine was used as the controlsubstrate. In experiment A, the labeled Fr I contained70,000 cpm; in experiment B, 100,000 cpm. The lysateswere incubated for 30 minutes.

incubation of marrow lysates was carried out.The cell-free lysate system had no ability to in-corporate leucine-C14 into Main Hb in either thepresence or absence of added Fr I. However,when Fr I which had been previously labeled withleucine-C14 was incubated with the lysate, therewas a prompt and striking appearance of C14 inthe Main Hb (Figure 7).

The lysate system was able to incorporate Fe59from added Fe59-citrate into Main Hb (Figure 8).Addition of Fr I enhanced the incorporation ap-

preciably, while addition of serum diminished it.If Fr I previously labeled with Fe59 was used as

the substrate for the lysate system, Fe59 rapidlyappeared in the Main Hb. By addition of carrierhemoglobin with subsequent heme crystallizationit was shown that the nonheme iron in Fr I hadbeen converted to heme in the Main Hb.

I

Z40,000

z

od

z 20,000

0a

z K0,0000

40,500,

_

11,680

5,360

_ 1

CONTROL.5LUV1

11,400

CAMI K)L 1I

FIG. 8. INCORPORATIONOF FE59 INTO MAIN HB BY A

LYSATE OF HUMANMARROWCELLS. In each experiment0.71 ,c (190,000 cpm) of Fet-citrate was added to a 2-mlaliquot of the lysate; 1.4 ml of saline was added to thecontrols, 1.4 ml of serum to the "serum" aliquot, and 1.4ml of a solution of "cold" Fr I (1 mg/ml) to the "Fr I"aliquot. The lysates were incubated for 15 minutes.

Characterization of fraction I

Absence of radioactive heme. Fr I showedmaximal absorption at 280 mju, but did have some

absorption at 415 mfA (Figure 1). When Fr Ilabeled with Fe59, leucine-C14, or glycine-2-C14was mixed with carrier hemoglobin, the isolatedheme contained less than 1 per cent of the activityexpected if all Fr I counts were assumed to be inhemoglobin.

Iron binding. The iron-binding affinity of Fr Iwas tested by dialysis against various buffers. Ta-ble I indicates the binding strength of Fr I com-

pared with that of inorganic iron, siderophilin, and

BLE I

The relative binding affinity of several Fe59-labeled compounds is illustrated by dialysis against various buffers *

Time Acetate Ascorbic Thioglycollicin Distilled Acetate +EDTA acid acid

hours water pH 4.6 pH 4.6 pH 2.7 pH 7.2

Fe59-citrate 2 100 49 22in dist. water 24 89 2 <1

Iron-binding protein 2 2(serum)

Main Hb fraction 2 100 100 93 89(IRC-50) 24 100 100 82 55

Fraction I 2 100 100 90 100(IRC-50) 24 100 100 60 67

* Results are recorded as per cent of original radioactivity remaining in the dialysis bag after the indicated numberof hours.

VM

% Kllf LMIL - irc l CIf%

1120

INTRACELLULAR PROTEIN INTERMEDIATE FOR HEMOGLOBIN

hemoglobin. The iron in Fr I was apparentlybound about as firmly as the iron in heme underthe conditions employed.

Stability. Clear solutions of Fr I became cloudyin a few hours with subsequent slow formation ofa precipitate at 40 C. Room temperature or freez-ing accelerated precipitate formation, and dialysisusually resulted in appreciable loss. In more re-cent experiments it was observed that the additionof a reducing agent, such as 0.005 M cysteine,seemed to stabilize solutions of Fr I. Fr I couldbe heated to 700 C for 15 minutes without pre-cipitation and without alteration of electrophoreticmigration. In this case a dark-brownish protein

CELLULOSE ACETATE ELECTROPHORESIS.06 M BARBITAL BUFFER pH 8.6

+

_ U

FERRITIN FRACTIONI RADIOAUTOGRAPH

FIG. 9. ELECTROPHORESISOF DOG LIVER FERRITIN ANDOF FE'5-LABELED CRUDEFR I FROMAN IRC-50 CHROMATO-GRAM.

precipitate settled out, and the Fe59 radioactivityremained in the supernate.

Distintction from siderophilin. Fr I bound ironmuch more firmly than did siderophilin (Table I).These two substances exhibited different rates ofmigration upon electrophoresis on cellulose acetate,at pH 8.6, with veronal buffer. There was nocross reaction of antisiderophilin antiserum withFr I.

Distinction from ferritin. Fr I and ferritin havein common the property of heat resistance. How-ever, upon electrophoresis the radioactivity incrude Fr I migrated in the opposite direction fromferritin (Figure 9). Crude Fr I placed oppositeantiferritin antiserum produced a faint band inagar diffusion, but this band did not coalesce withbands produced by crystalline liver ferritin (Fig-

FIG. 10. AGAR GEL DIFFUSION PLATE. Center well,antidog liver ferritin; 1, crude Fr I; 2, 4, 6, dog liverferritin; 3, dog hemoglobin; 5, Fr I purified by DEAE-cellulose chromatography.

ure 10). There was no apparent reaction withpurified dog Fr I. Anti-Fr I produced no reac-tion with crystalline liver ferritin.

Purification. A system of DEAE-cellulose col-umn chromatography gave some resolution of thecomponents of dog Fr I. A typical chromatogramis shown in Figure 11. The material in crudeFr I absorbing at 415 mu appeared in the firstpeak. Most of the Fe59 coincided with a subse-quent sharp maximum of absorption at 280 m/.This material had a ratio of D415: D280 of less than1: 1,000 indicating no detectable heme protein.

FIG. 11. CHROMATOGRAMOF CONCENTRATEDDOG FR I ONDEAE-CELLULOSE.

1121

W. B. GREENOUGH,III, T. PETERS, JR. AND E. D. THOMAS

That this purified material is a protein is indicatedby the following facts. It has a single absorptionmaximum at 280 m/u; it is nondialyzable; it isprecipitated by 5 per cent trichloroacetic acid; ithas the staining properties of a protein on elec-trophoresis; it reacted with cationic detergent andalkali in the method of Jacox (12) for proteindetermination.

DISCUSSION

These studies have demonstrated that, whenmarrow is incubated with Fe59, radioactivity ap-pears in the first few minutes not only in stromaand particulate matter but also in a soluble non-heme protein fraction (Fr I). There is a delayof some 5 to 8 minutes before significant activityappears in hemoglobin. Consideration of totalcount flux suggests that Fr I must play a majorrole in iron transport to hemoglobin.

Two pathways for the movement of iron intoyoung red cells have been described: the transferof ferritin from macrophages to young red cells bypinocytosis (4); and the direct transfer of ironfrom the plasma iron-binding protein to the youngred cells (2, 8). Since marrow is composed ofmany cell types, our studies cannot exclude a cell-to-cell transfer. However, radioautographic stud-ies indicate that most of the iron taken up by mar-row cells is in the red cell precursors (22), andmicrospectrophotometric studies show nonhemeiron in these cells (9). Reticulocyte preparationsthat do not contain macrophages are capable ofmaking hemoglobin from Fe59 bound to plasmairon-binding protein (23). Wehave shown thatthe active material in Fr I is neither plasma iron-binding protein nor ferritin. Hence, the weightof evidence indicates that the kinetics describedoccur entirely within young red cells.

Several investigators have suggested an intra-cellular intermediate in the transport of iron fromstroma to hemoglobin (1, 2, 8, 9). The studiespresented here describe an intracellular iron proteinthat was readily differentiated from siderophilinand ferritin, the only known soluble nonheme pro-teins important in iron transport. This solubleprotein appears to be the vehicle by which iron ismoved from fixed stromal pools to hemoglobin.The relationship, if any, of this soluble protein tothe "heme synthetase system" and, especially,

"fraction II" described by Schwartz and col-leagues (24) remains to be elucidated.

The striking parallel in kinetics between ironand amino acid incorporation raises the possibilitythat this protein is a precursor of the globin por-tion as well as of the iron of hemoglobin. Thisconcept is reinforced by the observation of the ap-pearance of C14-leucine in Main Hb when Fr Iprelabeled with C14-leucine was incubated in acell-free marrow lysate. Studies are continuingon the nature of the Fe59 and leucine precursorsubstances.

If the active portion of Fr I is incorporatedin toto into hemoglobin, it is necessary that theleucine-containing material be closely related to,or identical with, part of the globin molecule. Asyet, enough material of sufficient purity is notavailable to make any comment other than thatthere are many properties in common between thetwo. It is interesting to note that a highly labeledprotein closely related to globin present in youngred cells could well contribute an important incre-ment of radioactivity to both in vivo and in vitrostudies of heme versus globin formation. Thisproblem could be avoided by using an IRC-50column to prepare "pure" hemoglobin before theisolation of globin.

A scheme of hemoglobin synthesis which in-cludes a common precursor for the iron and thepeptide chains may be discussed as outlined inFigure 12. Iron from siderophilin, and perhapsferritin, enters the hemoglobin-forming cell, whereit is bound to the insoluble stromal proteins (8).Globin synthesis, from amino acid or peptides, andprotoporphyrin synthesis, via 8-amino levulinicacid and porphobilinogen, occur within the cell.As globin is formed, it, or part of it, combines withthe available iron to form the common precursorwhich, for brevity, might be termed "sideroglobin."Sideroglobin combines with protoporphyrin toform hemoglobin. An alternative pathway, com-bination of protoporphyrin with iron to formheme, has been more frequently considered (24),but it is of interest that an excess of heme inyoung red cells, normal or abnormal, has not beendescribed. On the other hand, an excess of freeprotoporphyrin has been demonstrated in severalabnormal states. As indicated in Figure 12, thismight be expected to occur from a deficiency of

1122

PLASMA ERYTHRON

INTRACELLULAR PROTEIN INTERMEDIATE FOR HEMOGLOBIN

(EXTRACELLULAR) CELL (I NTRACELLULA R)WALL

PROTEIN SIDEROGLOBIN

IRiON

\ POSYNTHSISN OTOPORPHYRIN I

FIG. 12. A SCHEMEFOR HEMOGLOBINFORMATION SUGGESTEDBY OUR EXPERIMENTS. The term sideroglobin

refer t a omintio o irn ita roei, psb.a po. t

FIG.reer12acobntino iroSCHEM FO HEMOGLOBI FORMATIONSUGSE BYrOUR EXEIEtS. Thebi temolesulero

iron, a deficiency of globin synthesis, a failure ofiron to combine with globin to form sideroglobin,or a failure of the final step, the union of sidero-globin with protoporphyrin to form hemoglobin.An excess of free erythrocyte globin has not beenrecognized, but its existence would be expected.Other defects, both inherited and acquired, of thefinal steps of hemoglobin formation should befound and may be better understood in the light ofmechanisms depicted in Figure 12.

SUMMARY

Marrow cells from human and dog were in-cubated for short intervals with radioactive ironor amino acids. The marrow lysate was chro-matographed on an IRC-50 column. A minorfraction (Fr I) became labeled before hemoglobin

ihen marrow was incubated with Fe59 or C14-leucine. Further studies provided additional evi-dence that this fraction contained a precursor tohemoglobin. Heme isolated from the fraction wasunlabeled. Purification of Fr I by DEAE-cel-lulose column chromatography showed that Fe59was bound to a nonheme protein distinct from fer-ritin and siderophilin. These observations suggestthe possibility that in the course of hemoglobinformation iron combines with a protein, possiblya part of the globin molecule, which unites withprotoporphyrin to form heme and the completehemoglobin molecule.

ACKNOWLEDGMENT

The authors wish to express their appreciation to Mrs.Ann Logan and Mrs. Marion Wales for their excellenttechnical assistance.

REFERENCES

1. Thorell, B. Studies on the Formation of CellularSubstances During Blood Cell Formation. London,Henry Kimpton, 1947.

2. 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.

3. Rabinovitz, M., and Olson, M. E. Evidence for aribonucleoprotein intermediate in the synthesis ofglobin by reticulocytes. Exp. Cell Res. 1956, 10,747.

4. Bessis, M. C., and Breton-Gorius, J. Iron particles innormal erythroblasts and normal and pathologicalerythrocytes. J. biophys. biochem. Cytol. 1957, 3,503.

5. Faber, M., and Falbe-Hanseni, I. Nonlhaemi iron inierythrocytes as a precursor for haemoglobin. Na-ture (Lond.) 1959, 184, 1043.

6. Clegg, M. D., and Schroeder, W. A. A chromato-graphic study of the minor components of normaladult human hemoglobin including a comparison ofhemoglobin from normal and phenylketonuric indi-viduals. J. Amer. chem. Soc. 1959, 81, 6065.

7. Schwartz, H. C., Cartwright, G. E., Smith, E. L.,and Wintrobe, M. M. Studies on the biosynthesisof heme from iron and protoporphyrin. Blood1959, 14, 486.

8. Allen, D. W., and Jandl, J. H. Kinetics of intracel-lular iron in rabbit reticulocytes. Blood 1960, 15,71.

1123

W. B. GREENOUGH,III, T. PETERS, JR. AND E. D. THOMAS

9. Sondhaus, C. A., and Thorell, B. Microspectrophoto-metric determination of nonheme iron in maturingerythroblasts and its relationship to the endocel-lular hemoglobin formation. Blood 1960, 16, 1285.

10. Wiggans, D. S., Burr, W. W., Jr., and Rumsfeld,H. W., Jr. The incorporation of leucine into glo-bin in the nucleated erythrocyte. J. biol. Chem.1960, 235, 3198.

11. Peters, T., Giovanniello, T. J., Apt. L., and Ross,J. F. A new method for the determination of se-rum iron-binding capacity I. J. Lab. clin. Med.1956, 48, 274. A simple improved method for thedetermination of serum iron II. Ibid., p. 280.

12. Jacox, R. F. Analysis of the proteins of rat serumby starch electrophoresis and by cationic detergentanalysis. I. Identification of an unusual globu-lin. J. exp. Med. 1959, 110, 341.

13. Fischer, H. Hemin. Organic Synth. 1941, 21, 53.14. Kassenaar, A., Morell, H., and London, I. M. The

incorporation of glycine into globin and the synthe-sis of heme in vitro in duck erythrocytes. J. biol.Chem. 1957, 229, 423.

15. Robinson, C. V. Windowless, flow type, proportionalcounter for counting C14. Science 1950, 112, 198.

16. Thomas, E. D. In vitro studies of erythropoiesis. I.The effect of normal serum on heme synthesis and

oxygen consumption by bone marrow. Blood 1955,10, 600.

17. Hill, R. L. Personal communication.18. Hirs, C. H. W., Moore, S., and Stein, W. H. A

chromatographic investigation of pancreatic ribo-nuclease. J. biol. Chem. 1953, 200, 493.

19. Huisman, T. H. J., and Prins, H. K. Chromato-graphic estimation of four different human hemo-globins. J. Lab. clin. Med. 1955, 46, 255.

20. Granick, S. Ferritin: Its properties and significancefor iron metabolism. Chem. Rev. 1946, 38, 379.

21. Brackenridge, C. J. Optimal staining conditions forthe quantitative analysis of human serum proteinfractions by cellulose acetate electrophoresis.Analyt. Chem. 1960, 32, 1353.

22. Austoni, M. E. Cytoautoradiography Fe' uptake bybone marrow cells. 1. Normal human erythroblastsand young erythrocytes. Proc. Soc. exp. Biol.(N. Y.) 1956, 92, 6.

23. Jandl, J. H., Inman, J. K., Simmons, R. L., andAllen, D. W. Transfer of iron from serum iron-binding protein to human reticulocytes. J. clin. In-vest. 1959, 38, 161.

24. Schwartz, H. C., Goudsmit, R., Hill, R. L., Cart-wright, G. E., and Wintrobe, M. M. The biosyn-thesis of hemoglobin from iron, protoporphyrinand globin. J. clin. Invest. 1961, 40, 188.

124