Proc. Nati. Acad. Sci. USAVol. 73, No. 10, pp. 3534-3538, October 1976Biochemistry

Glycosylation of hemoglobin in vitro: Affinity labeling of hemoglobinby glucose-6-phosphate*

(hemoglobin modification/diabetes mellitus/hemoglobin Aic/2,3-diphosphoglycerate/glycoprotein)

DAVID N. HANEYt AND H. FRANKLIN BUNNtDivision of Hematology, Department of Medicine, Peter Bent Brigham Hospital, Harvard Medical School, Boston, Massachusetts 02115

Communicated by Eugene Braunwald, August 9,1976

ABSTRACT To determine the mechanism for the formationof hemoglobin Alc (Hb Ale) in vivo, we incubated human he-moglobin with glucose and metabolites of glucose. ['4CJGlu-cose-6-phosphate (G6P) reacted readily with deoxyhemoglobin,and formed a covalent linkage. The reaction rate was consid-erably reduced in the presence of carbon monoxide or 2,3-di-phosphoglycerate (2,3-DPG). Purified G6P hemoglobin had alowered oxygen affinity and decreased reactivity with 2,3-DPGcompared to Hb A. G6P behaved as a 2,3-DPG analog andreacted specifically at the NHrterminal amino group of the Pchain. In contrast, the interaction of hemoglobin with glucosewas much slower, and was unaffected by carbon monoxide or2,3-DPG. Neither glucose-i-phosphate, fructose-1-phosphate,fructose-6phosphate, nor fructose-1,6-diphosphate formed areaction product with hemoglobin. G6P behaves as an affinitylabel with thephosphate group forming electrostatic bonds atthe 2,3DPG binding site and the aldehyde group reacting withthe NH2-terminal amino group of the a chain. Thus, G6P he-moglobin may be an intermediate in the conversion of Hb A toHb A1c.

About 5% of hemoglobin in normal human red cells is glycos-ylated. This minor component, designated Hb Alc (1), is in-creased at least 2-fold in patients with diabetes mellitus (2, 3).Hb Alc is identical in structure to the main component (Hb A)except that the NH2-terminus of each # chain is attached to aglucose molecule by a Schiff base or aldimine linkage (4-6). Wehave presented evidence indicating that the aldimine undergoesan Amadori rearrangement to form a more stable ketoaminelinkage (6):

HC=O

HLOHHOCH

#A-NH, + H|H

11COH

CHOH

glucose

HC=N-#AHCOH

IHOCH

_ IHCOH

H OH

CH OHaldimine

(Schiff base)

CH2-NH-#A

HOCHAmador

HCOHIICH2OH

ketoamine

In this scheme, glucose is assumed to be the initial reactant.However, this may not be the case. It is possible that hemoglobincondenses with a glycolytic intermediate rather than withglucose.

Because the glycosylation of hemoglobin is a slow and nearlyirreversible process occurring continuously during the 120-daysurvival of the red cell (7), it is likely to be a nonenzymatic re-

Abbreviations: Hb, hemoglobin; G6P, D-glucose-6-phosphate; 2,3-DPG,2,3-diphosphoglycerate.

* This work was presented in part at the 67th annual Federationmeeting of the American Society of Biological Chemists (June 4-10,1976; San Francisco, Calif.).

t Present Address: Department of Biochemistry and Molecular Biology,Northwestern University, Evanston, Ill.

t To whom requests for reprints should be sent.

action. We have examined the interaction of human hemo-globin with glucose and glucose metabolites in a cell free systemdevoid of enzymes or cofactors. Our results indicate that 1)-glucose-6-phosphate (G6P) readily forms a covalent linkagespecifically at the NH2-terminus of the fl chain, while glucosedoes not.

MATERIALS AND METHODSBlood from normal adults was collected in heparin or EDTA.Hemolysate was prepared by the method of Drabkin (8). HbA was purified at 40 on a carboxymethylcellulose column (CM52, Whatman, Inc.; Clifton, N.J.) (9). Approximately 2 g ofhemoglobin was applied to a 4.5 X 25 cm column, and elutedfor 15 h with 0.01 M phosphate buffer containing a linear pHgradient (pH 6.8-9.0) at a rate of 90 ml/hr.

Kinetics of Hemoglobin Modification. Hemolysate wasplaced in two-dimensionally stretched cellophane tubing anddialyzed exhaustively against 0.1 M Tris-HCl at pH 7.2. Twomilliliters of the stripped (phosphate-free) hemoglobin wasplaced in a gas-tight vial and either deoxygenated by passinghydrated argon across the solution, or saturated with carbonmonoxide (CO). [14C]G6P (New England Nuclear, Boston,Mass.) (50,MCi, 0.95 jimol) was mixed with 35.5 Mtmol of unla-beled G6P in 0.5 ml of distilled water and deoxygenated in asimilar manner. In like manner, 50 uCi (10.5 umol) of D-[14C]glucose was mixed with 26.1 umol of unlabeled 1-glucosein 0.5 ml of distilled water and deoxygenated. The deoxy Hbsolution was then mixed with the deoxygenated glucose or G6Psolution. In a parallel experiment, we tested the interaction ofpurified hemoglobin A with unlabeled glucose-l-phosphate,fructose-l-phosphate, fructose-6-phosphate, fructose-1,6-diphosphate, and glucuronic acid (Sigma Chemical Co.). Inmost experiments, the reaction mixture contained the followinginitial concentrations: 3 mM hemoglobin (a# dimer), 15 mMglucose or G6P, and 0.08 M Tris-HCl at pH 7.2. Some incuba-tion vials also contained 5 mM 2,3-diphosphoglycerate (2,3-DPG) (Sigma Chemical Co.). The reaction solutions were in-cubated at 200 for 5 hr. At selected intervals, small aliquots wereremoved, oxygenated, and gel filtered on Sephadex G-25 in 0.1M Tris-HCl at pH 7.2 which effectively removed both free andelectrostatically bound G6P. Hemoglobin concentration wasdetermined from the absorbance of the cyanmethemoglobinderivative at 540 nm. In experiments involving incubations withradiolabeled compounds, the gel-filtered hemoglobin solutionswere bleached with 5% performic acid, and radioactivity wasdetermined in a liquid scintillation counter (Isocap 300, SearleAnalytical). From the hemoglobin concentration and the spe-cific activity of the standard [14C]glucose and [14C]G6P solu-tions, the stoichiometric amount of hemoglobin modificationwas determined.

Structural Analysis of Modified Hemoglobins. Columnpurified Hb A was incubated with G6P for 5 hr under condi-

3534

Dow

nloa

ded

by g

uest

on

May

5, 2

020

Proc. Natl. Acad. Sci. USA 73 (1976) 3535

QX ~~~~~~~~~~~0

n ~~~~~~~0I .o 0.20E

^ 0.15 / ---*DCOHb + G6P

E.5 / 2,3-DPG0.05-

DeoxyHb +Glucose

G2,3-DPG0 2 4 6

TIME (HOURS)

FIG. 1. The incorporation of [14C]glucose and ['4C]G6P intodeoxy- and carboxyhemoglobin. Glucose incubated with: deoxy-hemoglobin (deoxyHb,i); deoxyhemoglobin and 0.5 mM 2,3-DPG(v); carboxyhemoglobin (COHb, o). G6P incubated with: deoxy-hemoglobin (o); deoxyhemoglobin and 5 mM 2,3-DPG (-); car-boxyhemoglobin (0). Hemoglobin concentration, 3 mM (dimer) in0.08 M Tris-HCl buffer at pH 7.2; temperature 200.

tions described above. After gel filtration, the hemoglobin wasreduced with NaBH4 (6) and then exhaustively dialyzed.The NaBH4 reduced reaction mixture was applied to a Bio-

Rex 70 column (1 X 30 cm) and was developed as describedpreviously (7). A normal hemolysate was chromatographed inparallel with the G6P reaction mixture. Hemoglobin solutionswere also analyzed by isoelectric focusing on polyacrylamidegels (10). Globin was prepared in acid acetone from the modi-

1.0

0.5 -Al Alb

- 2.0-0

fied hemoglobins, and then the chains were separated onCM-cellulose in 8 M urea (11). A tryptic digest was preparedfrom the desalted aminoethylated (3 chain (12). Approximately100 nmol of the tryptic digest was used for a tryptic peptidemap on thin-layer glass plates coated with cellulose (Avicel,Analtech; Newark, Del.). (See Fig. 3 legend and ref. 12 fordetails.) Autoradiography was performed by exposing thethin-layer plate to X-omat film (Kodak RP/R54, Rochester,N.Y.) for 5 days. Amino acid analyses were done by the AAALaboratories, Seattle, Wash., on a Durrum D-500 analyzer.Oxygen Equilibria Studies. The purified G6P modified Hb

(G6P Hb) was concentrated by pressure dialysis. To convert anycyanmethemoglobin into oxyhemoglobin, we then gel-filteredit under anaerobic conditions through a zone of excess sodiumdithionite on G-25 (13) into 0.04 M 2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)-1,3-propanediol at pH 7.2 con-taining 0.1 M NaCl. The oxy Hb was deoxygenated in a to-nometer, having a 1-cm light path, and oxygen equilibriummeasurements were performed at 200 as previously described(14).

RESULTSKinetics of hemoglobin modificationThe interactions of [14C]glucose and [14C]G6P with hemoglobinare shown in Fig. 1. The reaction between deoxyhemoglobinand G6P in the absence of any organic phosphates was ap-proximately 20-fold greater than the corresponding reactionwith glucose. The reaction of deoxyhemoglobin with G6P wasreduced 2.5-fold in the presence of 2,3-DPG, while the reactionbetween deoxyhemoglobin and glucose was unaffected by2,3-DPG. Furthermore, carboxyhemoglobin, which has thesame quaternary structure as oxyhemoglobin, reacted only halfas rapidly with G6P compared to deoxyhemoglobin, while thestate of hemoglobin ligation had no effect on its reaction withglucose. It is likely that the incorporation of G6P into hemo-globin involves covalent bonding rather than strong electrostaticinteractions since only a trivial amount of radioactivity wasrecovered in the gel filtered hemoglobin after very short in-cubation periods (Fig. 1). Gustavsson and de Verdier (23) foundthat, unlike 33-DPG, G6P did not bind electrostatically todeoxyhemoglobin.

-20

0.

A11 -109

FRACTION NUMBERFIG. 2. Bio-Rex 70 elution pattern of normal hemolysate (top panel) and G6P reaction mixture (bottom panel). Deoxyhemoglobin was treated

with [14C]G6P for 5 hr and reduced with NaBH4. The minor components were eluted with developer number 6 (1, 3) (fractions 1-70), and HbA was eluted with 0.3 M phosphate at pH 6.4 (fractions 71-100).

Biochemistry: Haney and Bunn

Dow

nloa

ded

by g

uest

on

May

5, 2

020

3536 Biochemistry: Haney and Bunn

CD

CDQ a3

°

. pI

,BTpl1p-N-G6P Tpl

O0+

... ....

FIG. 3. Top: peptide map of aminoethylated tryptic digest of['4CIG6P ,3 chain, prepared from column purified G6P-hemoglobin(Fig. 2). The spot shown by a broken line indicates the position ofnormal ffTpl missing in this peptide map. The stippled spot indicatesthe dominant radioactive spot which was ninhydrin negative. Elec-trophoresis was run in the horizontal direction at 1500 V for 1.5 hr atpH 4.7 in pyridine:acetic acid:water (1:1:114, vol/vol). Chromatog-raphy was run in the vertical direction in amyl alcohol:isobutanol:1-propanol:pyridine:water (1:1:1:3:3, vol/vol). Bottom: Autoradi-ograph of peptide map showing '4C-labeled peptide.

Analysis of the hemoglobin solutions by gel electrofocusingrevealed no new bands in those treated with glucose, glucose-1-phosphate, fructose-l-phosphate, fructose-6-phosphate, or

fructose-1,6-diphosphate. In contrast, G6P treated hemoglobincontained a well resolved minor component having a muchlower isoelectric point than hemoglobin A. Hemoglobin treatedwith glucuronic acid developed a minor component having an

isoelectric point only slightly lower than that of Hb A.

Localization of structural modificationFig. 2 shows the Bio-Rex elution pattern of hemoglobin incu-bated for 5 hr with G6P and then reduced with sodium bor-ohydride. In this experiment, the hemoglobin had previouslybeen purified on CM-cellulose, as described above to preventcontamination by nonheme proteins and minor Hb compo-

nents. The elution profile showed one very sharp radioactivepeak which constituted 14% of the total hemoglobin. Thiscomponent contained 0.9 mol of G6P per mol of a13 dimer.The chain separation of this purified G6P hemoglobin on CM

cellulose in 8 M urea revealed two major peaks and a minorpeak. All of the radioactivity was located in the major chaincomponent which eluted before the normal chain, thus in-dicating that the modification lowered the isoelectric point(making the protein more negatively charged). This componentcontained 1.0 mol of G6P per mol of chain. Between the

Table 1. Amino acid composition of 4C-labeled peptide*

ExpectedAmino acid Found in j3Tpl

Aspartic acid 0.2 0Glutamic acid 2.0 2Glycine 0.1 0Histidine 1.0 1Leucine 1.1 1Lysine 1.0 1Proline 1.2 1Threonine t 1.1 1Valine 0 1

* Values expressed as mol of residue/mol of peptide.t Corrected for losses during hydrolysis.

elution of the labeled chain and the a chain, a small unlabeledpeak eluted in a position identical to that of normal chain. Thelocalization of radioactivity to the chain was found when ei-ther carboxyhemoglobin or deoxyhemoglobin were reactedwith [14C]G6P.A peptide map of the aminoethylated tryptic digest of the

[14C]G6P chain is shown in Fig. 3. The fingerprint resembledthat of normal chain except that the ,BTpl, the normalNH2-terminal tryptic peptide, was absent. An autoradiogramrevealed only one significant radioactive spot which was nin-hydrin negative, migrated more toward the anode than ,BTpl,and had a lower chromatographic mobility. Amino acid analysisof this labeled peptide revealed a composition identical to thatof ,BTpl except that valine (the NH2-terminal residue of 13Tpl)was not detected (Table 1). It is uncertain whether this peptidecontains a phosphate group. These results indicate that thechain was modified at the NH2-terminal valine. In contrast,hemoglobin incubated with [14C]glucose failed to reveal aspecific binding site. The autoradiograms showed several la-beled peptides in both the a and chains.

Oxygen equilibriaIn the absence of organic phosphates, the oxygen binding curve

of purified G6P-hemoglobin showed somewhat lower oxygenaffinity than that of Hb A (Table 2). The addition of 0.5 mM2,3-DPG had a lesser effect on the oxygen affinity of G6P-Hbthan on that of Hb A. These results are very similar to earlieroxygen equilibria measurements on Hb Al, (14).

DISCUSSIONRecent information on the structure (5, 6) and biosynthesis (7)of Hb Alc has not provided a plausible explanation for the

Table 2. Effect of 2,3-DPG on the oxygen affinity ofG6P hemoglobin*

Ps 0 (mm Hg)PS0 plus

Plus 0.5 (2,3-DPG)/No mM 2,3- P50 (No

Hemoglobin 2,3-DPG DPG 2,3-DPG)

G6P Hb 1.26 1.70 1.35Hb A 1.00 1.95 1.95

* Hemoglobin: 2.5 X 10-2 mM (tetramer). Buffer: pH 7.2, 0.04 M2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)-1,3-propanedioland 0.1 M total chloride, temperature 20°. P50 = the partial pres-sure of 02 at which hemoglobin is half-saturated.

Proc. Natl. Acad. Sci. USA 73 (1976)

Dow

nloa

ded

by g

uest

on

May

5, 2

020

Proc. Natl. Acad. Sci. USA 73 (1976) 3537

His-143

10o

FIG. 4. Diagramatic representation of the site of reaction and probable sites of electrostatic interaction of glucose-6-phosphate and deoxy-hemoglobin (redrawn from Arnone, 1972) (17). The aldehyde group of G6P forms a Schiff base linkage with the NH2-termino group of one ,Bchain. The spatial orientation of the G6P molecule cannot be deduced from these experimental data. The precise site of electrostatic bindingof the phosphate group may differ from that shown in this diagram.

specificity of the modification at the NH2-terminus of the ,Bchain. It is likely that the aldehyde group of the sugar condenseswith the amino group by means of a nucleophilic attack. If so,uncharged amino groups such as at the NH2-terminus wouldbe modified in preference to E amino groups of lysine residues.However, the pK of the NH2-terminal amino group of the achain is at least as low as that of the chain (15). Our resultsindicate that G6P behaves as a 2,3-DPG analog, and thereforereacts specifically at the NH2-terminus of the # chain. Like2,3-DPG, G6P reacts more readily with deoxyhemoglobin thanwith liganded hemoglobin. Furthermore, the reaction of G6Pand deoxyhemoglobin is impaired by the presence of a com-

petitive binder, 2,3-DPG. Unlike 2,3-DPG, G6P forms a co-

valent bond with hemoglobin. It is not surprising that 2,3-DPGhad a lesser effect on the oxygen affinity of G6P hemoglobin,compared to unmodified hemoglobin. It is likely that G6Pserves as an affinity label in a manner similar to pyridoxal-5'-phosphate (16). The residues in deoxyhemoglobin responsiblefor the binding of 2,3-DPG have been established from thex-ray diffraction data of Arnone (17). As shown diagramaticallyin Fig. 4, it is likely that the phosphate group of G6P formselectrostatic bonds at the DPG binding site of one , chain, anddirects the formation of a Schiff base linkage at the NH2-ter-minal amino group of the other ,B chain. Unlike the interactionof hemoglobin with pyridoxal-5'-phosphate (16), the reactionof G6P at the NH2-terminus of the , chain does not require thedeoxy conformation. Like deoxyhemoglobin, carboxyhemo-globin reacted with G6P only at the # chains, albeit at a slowerrate.The mechanism proposed in Fig. 4 shows 1 mol of G6P in-

teracting with 1 mol of hemoglobin tetramer. In contrast, thepurified G6P hemoglobin contains 2 mol of G6P per tetramer.As Guidotti et al. (18) pointed out, it is usually not possible toisolate asymmetrical hemoglobin tetramers by conventionalelectrophoretic or chromatographic methods. During the col-umn purification, the hybrid tetramer, a2(3(3G6P, should disso-ciate into unlike a(3 dimers which would then sort with likedimers, forming two chromatographic peaks: a2/32 and a2#2C6P.The same considerations apply to the chromatographic puri-

fication of hemoglobin reacted with pyridoxal-5'-phosphate(16).

Perhaps G6P hemoglobin is a precursor of Hb Alc. G6P Hbmay be identical to one of the negatively charged minor com-ponents, Hb Aia or Alb. There is little structural informationon these hemoglobins. A very small proportion of normalhuman hemoglobin appears to be covalently linked to pyri-doxal-5'-phosphate (19). We have prepared hybrids of Hbs Alaand Alb with canine hemoglobins which indicate that they aref8 chain modifications. Winterhalter and Glatthaar (20) havealso presented data which indicates that Hb Alb involves analteration in the (3 chain. The fact that these minor componentsare elevated in diabetics (ref. 3, K. Gabbay, unpublished results)and have a pattern of biosynthesis similar to that of Hb Alc (7),indicates that one or both may be glycosylated. We are cur-rently investigating the structure of these two hemoglobins.

If G6P is the preferred intermediate in the glycosylation ofhemoglobin, then G6P Hb must undergo removal of phosphateto form Hb Al. Rose and O'Connell (21) have shown that inred cells G6P is relatively resistant to attack by phosphatases.However, the turnover of G6P in the red cell is several ordersof magnitude faster than that of glycosylated hemoglobin.

These results may be applicable to sickle cell anemia. Thepolymerization of sickle hemoglobin can be significantly re-duced by appropriate covalent modifications at the NH2-ter-minus of the , chain (22). By manipulating phosphorylatedintermediates within sickle erythrocytes, it may be possible toglycosylate a significant proportion of Hb S at the fl-NH2-ter-minus with the aim of inhibiting sickling.We are indebted to Dr. Paul M. Gallop and Dr. Kenneth H. Gabbay

for their helpful comments. This work was supported by Grant AMHC 18223-01 and by the Nehemias Gorin Foundation.

1. Allen, D. W., Schroeder, W. A. & Balog, J. (1958) "Observationson the chromatographic heterogeneity of normal adult and fetalhemoglobin: A study of the effects of crystallization and chro-matography on the heterogeneity and isoleucine content," J. Am.Chem. Soc., 80,1628-1634.

2. Rahbar, S. (1968) "An abnorma: hemoglobin in red cells of dia-betics," Cln. Chim. Acta 22, 296-298.

Biochemistry: Haney and Bunn

Dow

nloa

ded

by g

uest

on

May

5, 2

020

3538 Biochemistry: Haney and Bunn

3. Trivelli, L. A., Ranney, H. M. & Lai, H. T. (1971) "Hemoglobincomponents in patients with diabetes mellitus," N. Engl. J. Med.284,353-57.

4. Holmquist, W. R. & Schroeder, W. A. (1966) "A new N-terminalblocking group involving a Schiff base in hemoglobin Alc,"Biochemistry 5,2489-2503.

5. Bookchin, R. M. & Gallop, P. M. (1968) "Structure of hemoglobinAlc: Nature of the N-terminal , chain blocking group," Biochem.Biophys. Res. Commun. 32,86-93.

6. Bunn, H. F., Haney, D. N., Gabbay, K. H. & Gallop, P. M. (1975)"Further identification of the nature and linkage of the carbo-hydrate in hemoglobin Al,," Biochem. Biophys. Res. Commun.67, 103-109.

7. Bunn, H. F., Haney, D. N., Kamin, S., Gabbay, K. H. & Gallop,P. M. (1976) "The biosynthesis of human hemoglobin Alc: Slowglycosylation of hemoglobin in vivo," J. Clin. Invest. 57,1652-1659.

8. Drabkin, D. L. (1946) "Spectrophotometric Studies, XIV. Thecrystallographic and optical properties of the hemoglobin of manin comparison with those of other species," J. Biol. Chem. 164,703-723.

9. Huisman, T. H. J. & Meyering, C. A. (1960) "Studies on the he-terogeneity of hemoglobin. I. The heterogeneity of differenthuman hemoglobin types in carboxymethyl cellulose and inAmberlite IRC-50 chromatography," Clin. Chim. Acta 5,103-123.

10. Drysdale, J. W., Righetti, P. & Bunn, H. F. (1971) "The separationof human and animal hemoglobins by isoelectric focusing inpolyacrylamide gel," Biochim. Blophys. Acta 229, 42-50.

11. Clegg, J. B., Naughton, M. A. & Weatherall, D. J. (1966) "Ab-normal human hemoglobins. Separation and characterizationof the a and ,B chains by chromatography, and the determinationof two new variants, Hb Chesapeake and Hb J (Bangkok)," J.Mol. Biol. 19,91-108.

12. Jensen, M., Oski, F. A., Nathan, D. G. & Bunn, H. F. (1975)

"Hemoglobin Syracuse (a2#21OH21)Hi'PtO), a new high-affinityvariant detected by special electrophoretic methods," J. Clin.Invest. 55, 469-477.

13. Dixon, H. F. B. & McIntosh, R. (1967) "Reduction of methaem-oglobin in haemoglobin samples using gel filtration for continuousremoval of reaction products," Nature 213, 399-400.

14. Bunn, H. F. & Briehl, R. W. (1970) "The interaction of 2,3-di-phosphoglycerate with various human hemoglobins," J. Clin.Invest. 49, 1088-1095.

15. Garner, M. H., Bogardt, R. A. & Gurd, F. R. N. (1975) "Deter-mination of the pK values for the a-amino groups of human he-moglobin," J. Biol. Chem. 250, 4398-4404.

16. Benesch, R. E., Benesch, R., Renthal, R. D. & Maeda, N. (1972)"Affinity labeling of the phosphate binding site of hemoglobin,"Biochemistry 11, 3576-3582.

17. Arnone, A. (1972) "X-ray diffraction study of binding of 2,3-diphosphoglycerate to human deoxyhaemoglobin," Nature 237,146-149.

18. Guidotti, G., Konigsberg, W. H. & Craig, L. C. (1963) "On thedissociation of normal adult human hemoglobin," Proc. Nat.Acad. SM. USA 50,774-782.

19. Srivastava, S. K., VanLoon, C. & Beutler, E. (1972) "Character-ization of a previously unidentified hemoglobin fraction," Bio-chim. Blophys. Acta 278, 617-621.

20. Winterhalter, K. H. & Glatthaar, B. (1971) "Intermediates ofhemoglobin and their relation to biosynthesis," Ser. Haematol.IV, 84-96.

21. Rose, I. A. & O'Connell, E. L. (1964) "The role of glucose-6-phosphate in the regulation of glucose metabolism in humanerythrocytes," J. Biol. Chem. 239, 12-17.

22. Nigen, A. M., Njikam, N., Lee, C. K. & Manning, J. M. (1974)"Studies on the mechanism of action of cyanate in sickle celldisease," J. Biol. Chem. 249, 6611-6616.

23. Gustavsson, T. & de Verdier, C.-H. (1973) "Binding if somephosphorylated intermediates to oxygenated and deoxygenatedhemoglobin A," Acta Biol. Med. Ger. 30,25-31.

Proc. Natl. Acad. Sci. USA 73 (1976)

Dow

nloa

ded

by g

uest

on

May

5, 2

020