Proc. Nat. Acad. Sci. USAVol. 78, No. 1, pp. 77-81, January 1976Biochemistry

Modifications of purified glucose-6-phosphate dehydrogenase andother enzymes by a factor of low molecular weight abundant in someleukemic cells

(molecular aging/post-translational modification/leukemia)

AXEL KAHN*, PIERRE BoIVIN, HENRIETTE RUBINSON, DOMINIQUE COTTREAU, JOELLE MARIE, ANDJEAN-CLAUDE DREYFUSCentre de Recherches sur les Enzymopathies, INSERM U24 et ERA 573 du CNRS, H'pital Beaujon, 92110 Clichy, France; and Institut de PathologieMol&ulaire, INSERM U 129, 24 rue du Faubourg St. Jacques, 75014 Paris, France

Communicated by Irving M. London, October 17,1975

ABSTRACT Highly purified platelet glucose;--phosphatedehydrogenase (G6PD; D-glucose-6phosphate:NADP+ 1-oxi-doreductase, EC 1.1.1.49) can be modified in its isoelectricpoint and its molecular specific activity by extracts of someleukemic granulocytes. The "G6PD modifying factors" arerelatively small molecules (molecular weight slightly under5000), thermostable, dialyzable, and ultrafilterable. Thesemolecules are destroyed by various endo- and exopeptidasesand by serine enzymes present in crude extracts of leuko-cytes and commercial preparations of ribonuclease. The al-terations of platelet G6PD due to the "G6PD modifying fac-tors" are stable and not reversible by dialysis or further chro-matography.The leukemic extracts which are able to modify G6PD also

can modify the electrophoretic mobility and (or) the enzy-matic activity of purified leukocyte pyruvate kinase, 6-phos-phogluconate dehydrogenase, and glucosephosphate isomer-ase.The chemical nature of such modifications and their rela-

tionships with post-translational modifications which occurin leukemic or normal cells are discussed.

Once synthesized, human glucose-6-phosphate dehydroge-nase (G6PD; D-glucose-6-phosphate:NADP+ 1-oxidoreduc-tase, EC 1.1.1.49) undergoes modifications of its molecularspecific activity, isoelectric point, and kinetic properties (1,2). In erythrocytes, these modifications depend on the aver-age age of the G6PD molecules and therefore on the age ofthe erythrocytes, since these cells no longer synthesize pro-teins (1-3). However, the age of the enzyme molecules is notthe only factor involved in these post-translational modifica-tions; these alterations are also the result of an active processrequiring tissue factors, whose concentration varies from onecell type to another.We showed previously that some leukemic granulocytic

cells have modified G6PD forms (4) (Fig. 1). The leukemicextracts, when incubated with normal partially purifiedG6PD, led to modifications of this enzyme identical to thoseof the enzyme in the leukemic cell itself (4). These resultssuggested that the modified forms of G6PD in leukemic cellswere due to the high concentration of "G6PD modifyingfactors."The purpose of this work was to characterize the bio-

chemical nature of the "G6PD modifying factors," to eluci-date the type of interaction between these factors and theenzyme, and, finally, to look for an effect of these factors onenzymes other than G6PD.

MATERIAL AND METHODSMaterialThe substrates of the enzymatic reactions were provided byBoehringer-Mannheim and by Sigma; the reagents used forthe buffers came from Merck and Calbiochem. The ion ex-changers (DEAE-Sephadex A-50, CM-Sephadex C-50, andSephadex G-25) were furnished by Pharmacia. The columneluates were scanned at 280 nm in a Beckman DBG spectro-photometer. Acrylamide and bisacrylamide came fromEastman Kodak, the Ampholines from LKB, starch for gelelectrophoresis from Connaught. The electrofocusing exper-iments on acrylamide-Ampholine columns were performedin a Bio-Rad 150 gel electrophoresis cell. The ultrafiltrationswere performed through ultra thimbles UH 100 (Schleicherand Schfill laboratory). The enzymatic reactions were mea-sured in a Zeiss PM QII spectrophotometer connected to aServogor recorder.MethodsThe enzymatic assays of G6PD, glucose phosphate isomer-ase, 6-phosphogluconate dehydrogenase, and pyruvate ki-nase were performed according to the methods summarizedby Beutler (5). The results were expressed in internationalunits at 30'/mg of proteins. Immunological and kinetic

pH 101ib_ SEIIIHb

a a, , O

C -_

a no di cforms1

pH3.54 3 5 7 9 1pHp5 1 2 3 4 5 6 7 8 '9 ilO

FIG. 1. Electrofocusing in polyacrylamide-Ampboline gel ofG6PD of various leukemic cells. 1-Control polymorphonuclears;2-control platelets; 3-granulocytes from a patient with myelofi-brosis; 4-leukocytes (85% myeloblasts) from a patient with acutetransformation of myelofibrosis; 5 and 6-platelets and granulo-cytes from a patient with chronic granulocytic leukemia; 7-blasts(99%) from a patient with chronic granulocytic leukemia in blasticcrisis; 8 and 9-myeloblasts (more than 90%) from two patientswith acute myeloid leukemia; 10-erythroblast-rich cellular frac-tion from a patient with erythroleukemia. The extracts studiedwere obtained from patients whose G6PD is shown in samples 5and 9. Hemoglobin deprived of any G6PD activity was added to allthe extracts as a marker.

77

Abbreviation: G6PD, glucose-6-pbosphate dehydrogenase.* Charge de recherche a l'INSERM. Present address: HOpital Beau-jon, 92110 Clichy, France.

4

Dow

nloa

ded

by g

uest

on

Aug

ust 1

0, 2

021

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

studies were performed according to previously reportedmethods (2). RNase activity was assayed according to Anfin-sen et al. (6).

Purification of Human Platelet G6PD. Human bloodplatelets suspended in citrate-dextrose were obtained, fromthe Centre National de Transfusion Sanguine-Paris. The re-maining erythrocytes and leukocytes were eliminated byone centrifugation for 20 min at 300 X g at 40 and by twocentrifugations for 2 min at 600 X g. The platelets werelysed by digitonin at saturation, and G6PD was then highlypurified according to the method of Kahn and Dreyfus (7).About 0.7 mg of G6PD was obtained, with specific activityof 173 international units (IU)/mg of proteins. The overallyield was 56%. Platelet G6PD was homogeneous when test-ed by acrylamide-sodium dodecyl sulfate electrophoresisand immunodiffusion.

Electrofocusing on 7.5% acrylamide-2% Ampholine ofpH range 3.5-10 and starch gel electrophoresis in Tris-bo-rate buffer, pH 8, were described in previous papers; thegels were specifically stained for G6PD activity (2, 8).

Studies of the In Vitro Modifications of Platelet G6PDby Leukemic "G6PD Modifying Factors." Ten microlitersof purified G6PD dilution (total activity: 10 X 10-3 IU)were mixed with the factors studied in 50 Al of 50 mM sodi-um phosphate buffer, pH 6.4, containing 2 mg/ml of bovinealbumin, 0.2 mM NADP+, 1 mM dithiothreitol, 1 mMEDTA, 10 mM e-aminocaproic acid, and 1 mM diisopropyl-fluorophosphate; then, the preparation was protected fromair by a coat of paraffin oil and incubated for 18 or 42 hr at370. A 50 ul sample of the incubated mixture was depositedon the acrylamide-Ampholine columns for electrofocusing.In some experiments the influence of pH on the alterationsof platelet G6PD by the leukemic "G6PD modifying fac-tors" was studied in Tris chloride buffer pH 8 and pH 9.

Studies on Some Physicochemical Properties of theLeukemic "G6PD Modifying Factors." The leukemic ex-tracts were prepared from the blood of two patients: onewith chronic granulocytic leukemia and the other with acutemyelocytic leukemia. G6PD from the leukemic granulocytesof these patients showed very abnormal electrofocusing pat-terns (4), with active bands of lower isoelectric point (Fig.1). The crude granulocyte extracts were prepared in 5 mMsodium phosphate buffer, pH 6.4, then either heated for 1 hrat 600 or boiled for 5 min, and ultrafiltered. The ultrafiltratewas concentrated by lyophilization, and then deposited onthe top of a Sephadex G-25 column (80 X 2 cm) previouslyequilibrated with 5 mM phosphate buffer, pH 6.4. The flowrate was 10 ml/hr and fractions of 4 ml were collected. Eachfraction was individually Iyophilized and tested against pu-rified platelet G6PD.

Attempts were made to inactivate the "G6PD modifyingfactors" by proteases, phosphoesterases, nucleases, or neu-raminidase.The boiled and ultrafiltered leukemic extracts were incu-

bated for 18 hr at 370 with the following reagents: 1 mg/mlof papain in 50 mM Tris chloride buffer, pH 8, containing 1mg/ml of cysteine and 2 mM EDTA; 1 mg/ml of bromelainin 20 mM sodium phosphate buffer, pH 7, containing 1mg/ml of cysteine and 2 mM EDTA; 1 mg/ml of elastase in50 mM Tris chloride buffer, pH 8.5; 1 mg/ml of trypsin in20 mM phosphate buffer, pH 7, containing 10 mM CaCl2; 1mg/ml of chymotrypsin A in 50 mM Tris chloride buffer,pH 8, containing 50 mM CaCl2; 1 mg/ml of pepsin in 50mM HCl-KCl buffer, pH 2.2; 1 mg/ml of carboxypeptidaseA in 50 mM Tris chloride buffer, pH 8, containing 1 M NaCl

and 2 mM diisopropylfluorophosphate; 0.5 mg/ml of car-boxypeptidase B in 50 mM Tris chloride buffer, pH 7.6, con-taining 100 mM NaCl and 2 mM diisopropylfluorophos-phate; 1 mg/ml of carboxypeptidase C in 50 mM citratebuffer, pH 5.3, containing 2 mM diisopropylfluorophos-phate; 0.5 mg/ml of ribonuclease from bovine pancreas in100 mM acetate buffer, pH 5.6; 0.025 mg/ml of ribonucle-ase Ti from Aspergillus oryzae in 50 mM Tris chloridebuffer, pH 7.5, containing 20 mM EDTA; 1 mg/ml of de-oxyribonuclease from bovine pancreas in 50 mM phosphatebuffer, pH 6.4, containing 10 mM MgCl2; 1 mg/ml of alka-line phosphatase in 100 mM NaOH-glycine buffer, pH 10;2.5 mg/ml of acid phosphatase in 200 mM acetate buffer,pH 5.6; 1 mg/ml of neuraminidase from Vibrio cholerae in200 mM citrate buffer, pH 5.4. After the incubation, the pHwas adjusted to pH 6.4 with diluted HC1 or NaOH, the pro-tein reagents were eliminated by ultra-filtration, and the ul-trafiltrates were then incubated with purified platelet G6PDas described above.An attempt was made to inactivate the leukemic G6PD

modifying factors by 2 mM iodoacetic acid at pH 4.6 and io-doacetamide at pH 8. After incubation overnight with thesereagents, pH was adjusted to 6.4 and the reagents in excesswere neutralized with 10 mM dithiothreitol.

Studies on the Influence of the "G6PD Modifying Fac-tors" on Enzymes Other Than G6PD. Human leukocytepyruvate kinase (A. Kahn, J. Marie, and P. Boivin in prepa-ration) glucose-6-phosphate isomerase (0. Bertrand, A.Kahn, and P. Boivin in preparation) and 6-phosphogluco-nate dehydrogenase (9) were highly purified according tomethods reported elsewhere. These purified enzymes wereincubated at pH 6.4 with boiled and ultrafiltered leukemicgranulocyte extracts in the same way as for G6PD. After in-cubation the residual enzyme activities were assayed and themodifications of electrophoretic mobility were examined byelectrofocusing in plates of 3.75% acrylamide and 1.4% Am-pholines of pH range 9-6 for pyruvate kinase, by electrofo-cusing in columns of 5% acrylamide-1% Ampholines of pHrange 9-11 for glucose phosphate isomerase, and by starchgel electrophoresis for 6-phosphogluconate dehydrogenase.

RESULTS

Highly purified platelet G6PD showed upon electrofocusingone predominant active band (band a), with a minor moreanodic band, (band b) (Fig. 2). When incubated with eitherthe boiled and ultrafiltered leukemic leukocyte extracts orthe extracts heated for 1 hr at 600 with 1 mM diisopropyl-fluorophosphate, this pattern was changed into more anodicforms (Fig. 2) and simultaneously the enzymatic activity de-creased 40-60% (Table 1).By contrast the extracts were left without any influence

on platelet G6PD when heated for 1 hr at 60° without di-isopropylfluorophosphate. Similarly a previous dialysis abol-ished the capacity of the extracts to modify platelet G6PD(Table 1).

Fig. 2 shows the influence of pH on the modifications ofplatelet G6PD by the boiled and ultrafiltered extract: theenzyme was modified at pH 6.4, hardly changed at pH 8,and unmodified at pH 9.

After filtration of the boiled, ultrafiltered and lyophilizedextracts on a Sephadex G-25 column, the fraction able tomodify platelet G6PD was found just after the void volumeof the column (Fig. 3).

78 Biochemistry: Kahn et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 1

0, 2

021

Proc. Nat. Acad. Sci. USA 73 (1976) 79

pHlot

ASn a}a ~~b

__ amuum&--

anodicforms

pH3.5+ 1 2 3 4 5FIG. 2. Influence of pH on the modifications of purified plate-

let G6PD by leukemic extracts. The leukemic extracts were boiledand ultrafiltered. 1-Purified platelet G6PD, nonincubated; 2-incubation under the usual conditions (see Methods); 3-incuba-tion in 50 mM Tris chloride buffeP, pH 8; 4-incubation in 50 mMTris chloride buffer, pH 9; 5-incubation at pH 6.4 without G6PDmodifying factor.

The "G6PD modifying factors" were not destroyed bytrypsin, chymotrypsin A, elastase, alkaline and acid phos-phatases, neuraminidase, carboxypeptidase A, and DNase;by contrast, these factors seemed to be sensitive to pepsin,papain, bromelain, carboxypeptidases B and C, and ribonu-clease preparations. Diisopropylfluorophosphate and boilingfor 15 min at pH 2.2, however, abolished the effects of theribonuclease preparations on the G6PD modifying factors,while the RNase activity was not affected by these treat-ments. (See Fig. 4.) Iodoacetamide and iodoacetic acid didnot inactivate these factors.The more negative charge of modified C6PD led to an

abnormal starch gel electrophoretic mobility (Fig. 5). How-ever, the differences between modified and unmodifiedplatelet G6PD were less obvious in starch gel electrophoresisthan in electrofocusing. The kinetic properties of modifiedplatelet G6PD were identical to those previously reported

r.

E-

C4

Toum

0 60 120 240 mlEFFLUENT

FIG. 3. Gel filtration on Sephadex G-25 column of leukemic ex-

tracts boiled, ultrafiltered, and lyophilized. (...) Absorbance at280 nm; (--- - -) absorbance at 620 nm, detection of dextran blue asmarker of the void volume; (+ + +) capacity of the fractions tomodify purified G6PD. The capacity of the fractions for modifyingG6PD was estimated semiquantitatively by the minimal volumeable to cause the appearance of modified G6PD forms from puri-fied platelet G6PD.

pH 3.5 1 2 3 4 5 6 7 8FIG. 4. Inactivation of the "G6PD modifying factors" in vari-

ous conditions. Before incubation with purified platelet G6PD, theleukemic extracts were treated by: 1, buffer alone; 2, papain; 3,elastase; 4, trypsin; 5, chymotrypsin A; 6, RNase from bovine pan-creas; 7, RNase Ti; 8, RNase from bovine pancreas in the presenceof 2 mM diisopropylfluorophosphate.

for a crude extract of leukocyte G6PD incubated in vitro,and for the hyperanodic G6PD forms isolated by electrofo-cusing (2). The main characteristics of the modified enzymewere a high Michaelis constant for glucose-6-phosphate(about 100 MM). As previously reported (2), molecular spe-

cific activity was decreased to 40-60% of normal.Table 2 shows the influence of the boiled and ultrafiltered

leukemic extracts on other enzymes: glucose phosphateisomerase was drastically inactivated; leukocyte pyruvate ki-nase was stabilized and its isoelectric point was decreased;6-phosphogluconate dehydrogenase showed a more anodicmigration.

DISCUSSIONPlatelet G6PD was purified and used to test the capacity ofleukemic factors to modify G6PD because normal plateletshave been shown to be very poor in G6PD modifying factors(2).The G6PD modifying factors were extracted from leuke-

mic granulocytes in which G6PD itself had enzymatic formswith decreased isoelectric point. This finding seems to indi-cate that these factors are responsible, at least in part, for theabnormal electrofocusing pattern found in freshly extracted

Table 1. Modification of highly purified platelet G6PDby leukemic extracts after various treatments

Inactivation ofCapacity for modi-i G6PD activity,fying pI of G6PD S

Crude extract + 50-60Dialyzed extract 0 0Ultrafiltrate + 50-60Boiled extract + 50-,60Extract heated for

1 hr at 600 with-out DFP 0 0

Extract heated for1 hr at 60° with1 mM DFP + 50-60

In the experiments with crude or dialyzed extracts, G6PD fromthe leukemic leucocytes themselves was mixed with purified plate-let G6PD added to the extract. DFP = diisopropylfluorophosphate.

Mt(.0 co 0a

=L =

pH 10t

anodicforms

Biochemistry: Kahn et al.

i...,Lopli. 7; F-I

Dow

nloa

ded

by g

uest

on

Aug

ust 1

0, 2

021

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

4 co

FIG. 5. Influence of the G6PD modifying factors on the starchgel electrophoresis pattern of purified platelet G6PD. From left toright, alternatively, purified G6PD incubated for 42 hr at 370 withbuffer alone and with boiled and ultrafiltered leukemic extracts(G6PD-MF); 12% starch gel in Tris-borate buffer, pH 8.

leukocytes. There were, however, differences between thein vivo and in vitro modifications of G6PD: in this lattercase there were no intermediate forms between the initialband a and the hyperanodic bands, while in dvo molecularaging led to the transition: band a -- band b -a band c -0

hyperanodic forms (2, 8). Either the mechanism of theabove transition is different from the transformation intohyperanodic forms, or the effects of the same G6PD modi-fying factors are different in vivo and in vitro.The post-translational modifications of G6PD do not seem

to be qualitatively different in either normal or leukemicleukocytes but only much faster in the latter cells than in theformer ones. Consequently it is probable that the leukemic"G6PD modifying factors" are nearly, if not entirely, identi-cal to the physiological factors, and that they exist in a muchhigher concentration.

These factors are molecules of low molecular weight,since they are dialyzable and ultrafilterable. The gel filtra-tion experiment on a Sephadex G-25 column showed thattheir molecular weight was slightly below the limit of exclu-

Table 2. Influence of boiled and ultrafiltered leukemicextracts on enzymes other than G6PD

Enzyme Anodization Inactivation

6-Phosphogluconatedehydrogenase + 0

Glucose phosphateisomerase 0 +

Pyruvate kinase M2 fromleukocytes + Stabilization

All the enzymes were highly purified from human leukocytes.Glucosephosphate isomerase and pyruvate kinase were electro-focused on acrylamide-Ampholine gel according to originalmethods (J. Marie, A. Kahn, and P. Boivin, in preparation).6-Phosphogluconate dehydrogenase starch gel electrophoresis wasperformed according to Blake et al., in phosphate-citrate buffer,pH 7 (14).

sion of the gel, i.e., below 5000.It seemed contradictory that the G6PD modifying factors

were not destroyed by boiling but by heating of the leuko-cyte extracts for 1 hr at 60°.The fact that diisopropylfluorophosphate protected the

modifying factors against heating at 600 probably explainedthis apparent contradiction: at 600 the "modifying factors"would be not destroyed by heat, but by a serine enzyme in-hibited by diisopropylfluorophosphate. The G6PD modi-fying factors were insensitive to the action of various pro-teases with narrow specificity, of neuraminidase, monophos-phoesterases, deoxyribonuclease, and SH or imidazole re-agents. By contrast they were inactivated by pepsin, brome-lain, papain, carboxypeptidases B and C, and ribonucleasepreparations. However, the action of ribonucleases was abol-ished by diisopropylfluorophosphate, although they are notserine enzymes (10); moreover, boiling for 15 min at pH 2.2which did not inactivate pancreatic RNase, in agreementwith Kunitz (11), abolished the capacity of the commercialpreparation to destroy the "G6PD modifying factors." Thesedata indicated that the "modifying factors" were not de-stroyed by ribonucleases themselves, but by a serine enzymecontained in the ribonuclease preparations.Thus the G6PD modifying factors are small thermostable

molecules, destroyed by serine enzyme present in the leuko-cyte extracts and in the ribonuclease preparations, and de-stroyed also by various endo- and exopeptidases. These mol-ecules seem to be, at least in part, oligopeptides, the terminalresidue of which was partially elucidated by the action ofthe various carboxypeptidases: the "G-6PD modifying fac-tors" have a COOH-terminal residue with a free carboxylgroup, since carboxypeptidases B and C were active (12, 13);this COOH-terminal residue seems to be either L-lysine, L-arginine, or L-ornithine, specifically removed by carboxy-peptidase B (13) and slowly or not removed by carboxypep-tidase A (13).Up to now the interactions between G6PD and modifying

factors remain unknown. Obviously a nonenzymatic phe-nomenon is involved. Two hypotheses can be made: the firstone is that all or part of the modifying factors bind to specialgroups of the enzyme. The ionization state of these groupswould be involved in this binding. The second hypothesis isthat the "modifying factors" are nonenzymatic catalysts ofan alteration of the G6PD molecule itself. In any. case, themodifications of G6PD were stable and not reversible by theusual methods.

It is interesting to note that leukemic granulocyte extracts,boiled and ultrafiltered, were able to modify the electriccharge and (or) the enzymatic activity of various other puri-fied enzymes. Therefore, these alterations of enzymes byfactors abundant in some leukemic cells may be a commonphenomenon. The chemical nature of this phenomenon andthe consequences of such enzyme alterations on the functionof the leukemic cell remain to be determined. The possiblerelations of the leukemic factors to the physiological factorsinvolved in molecular aging deserves to be explored.

This work was supported by a grant of Institut National de laSante et de la Recherche Medicale (ATP no. 14-75-37 andC.R.L.75.4.164.2) and of the University Paris VII.

1. Fornaini, G., Leoncini, G., Segni, P., Calabria, G. A. & Dacha,M. (1969) Eur. J. Biochem. 7,214-222.

2. Kahn, A., Boivin, P., Vibert, M., Cottreau, D. & Dreyfus, J. C.(1974) Biochimie 56, 1395-1407.

80 Biochemistry: Kahn et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 1

0, 2

021

Biochemistry: Kahn et al.

3. Marks, P. A. & Bishop, J. C. (1957) J. Clin. Invest. 36,913.4. Kahn, A., Cottreau, D., Bernard, J. F. & Boivin, P. (1975)

Biomedecine, in press.5. Beutler, E. (1971) Red Cell Metabolism, a Manual of Bio-

chemical Methods (Grune and Stratton, New York and Lon-don).

6. Anfinsen, C. R., Redfield, R. R., Choate, W. L., Page, J. &Caroll, W. R. (1954) J. Biol. Chem. 207, 201-211.

7. Kahn, A. & Dreyfus, J. C. (1974) Biochim. Blophys. Acta 334,257-265.

8. Kahn, A., Hakim, J., Cottreau, D. & Boivin, P. (1975) Clin.Chim. Acta 59,183-190.

Proc. Nat. Acad. Sci. USA 73 (1976) 81

9. Kahn, A., Milani, A., Marie, J., Cottreau, D. & Boivin, P.(1975) Biochimie 57,325-35.

10. McDonald, M. R. (1955) in Methods in Enzymology, eds. Co-lowick, S. P. & Kaplan, N. O., (Academic Press, New York),Vol 2, pp. 427-433.

11. Kunitz, M. (1940) J. Gen. Physiol. 24,15-32.12. Zuber, H. (1968) Hoppe Seyler's Z. Physiol. Chem. 349,

1337-1348.13. Neurath, H. (1960) in The Enzymes, eds. Boyer, P. D., Lardy,

H. & Myrback, K. (Academic Press, New York and London),Vol 4, pp. 11-6.

14. Blake, N. M., Saha, N., McDermid, E. M., Kirk, R. L. &Crane, G. G. (1974) Humangenetik 21,347-354.

Dow

nloa

ded

by g

uest

on

Aug

ust 1

0, 2

021