The First Case of a Complete Deficiency

of DiphosphoglycerateMutase in Human Erythrocytes

RAYMONDERoSA, MARIE-ODETTEPREHU, YVEs BEUZARD, and JEAN ROSA, Un1itde Rechterches stir les Anieinies, In.stitut Natiotnal de la Sante et de la RecherchleMWdicale U. 91, HMpital Hetnri Monidor, 94010 Creteil, Fratnce

A B S T R AC T An inherited and complete deficiencyof diphosphoglycerate mutase was discovered in theerythrocytes of a 42-yr-old man of French origin whoseblood hemoglobin concentration was 19.0 g/dl. Uponphysical examination he was normal with the exceptionof a ruddy cyanosis. The morphology of his erythro-cytes was also normal and there was no evidence ofhemolysis. The erythrocyte 2,3-diphosphoglyceratelevel was below 3% of normal values and, as a con-sequence, the affinity of the cells for oxygen was in-creased. Diphosphoglycerate mutase activity was un-detectable in erythrocytes as was that of diphospho-glycerate phosphatase. The activities of all the othererythrocyte enzymes that were tested were normal ex-cept for monophosphoglycerate mutase which was di-minished to 50% of the normal value. The levels ofreduced glutathione, ATP, fructose 1,6-diphosphate,and of triose phosphates were elevated, whereas thoseof glucose 6-phosphate and fructose 6-phosphate weredecreased. This report sheds new light on the role ofdiphosphoglycerate mutase in the metabolism of eryth-rocytes.

INTRODUCTION

In human erythrocytes diphosphoglycerate mutase(DPGM)l (EC 2.7.5.4.) catalyses the formation of 2,3-

Receivedfor publication 17 April 1978 atnd in revisedform2 June 1978.

1 Abbreviations used in this paper: ACD, acid-citrate-dex-trose; 1,3-DPG, 1,3-diphosphoglycerate; 2,3-DPG, 2,3-di-phosphoglycerate; DPGM,diphosphoglycerate mutase; DPGP,diphosphoglycerate phosphatase; FDP, fructose 1,6-diphos-phate; F6P, fructose 6-phosphate; GAPDH, glyceraldehyde3-phosphate dehydrogenase; G6P, glucose 6-phosphate;MPGM,monophosphoglycerate mutase; ODC, oxygen dis-sociation curve; PE, phospho-enzyme; P50, partial pressure ofoxygen at which hemoglobin is half saturated with 02; TP,triose phosphates.

diphosphoglycerate (2,3-DPG), an important effectorof the dissociation of oxygen from hemoglobin; erythro-cytes also contain, 2,3-diphosphoglycerate phospha-tase (DPGP) (EC 3.1.3.13). These two enzyme ac-tivities have been separately studied by several authors(1-5). In previous investigations we have provided evi-dence that DPGMand DPGPactivities are carried bya single molecule (6) which in addition contains amonophosphoglycerate mutase activity (MPGM) (EC2.7.5.3.) (7, 8). The trifunctional enzyme has been ob-tained in a purified form by the use of chromatographictechniques (9, 10). This enzyme presents a single elec-trophoretic band (6) and represents one of the threemolecular forms of erythrocyte MPGM(8). Species dif-ferences of erythrocyte DPGMhave been described inthe erythrocytes of mammals, such as the cat, and alsoin ruminants whose cells contain very low levels of2,3-DPG (11). Thus the activity of this enzyme in vivois probably an important determinant of 2,3-DPG levelswithin the erythrocytes.

Until now, only partial deficiencies of erythrocyteDPGMhave been described in three unrelated fam-ilies. In one of them (12) the defect was tolerated with-out any clinical manifestation. In contrast, the two othercases displayed a chronic hemolysis (13, 14). In thepresent report we describe the first case of a completedeficiency of erythrocyte DPGM. Clinical and bio-chemical studies have been performed on the propos-itus, who is homozygous for the defect, and on histwo heterozygous children. These studies have thusallowed us the unique opportunity of evaluating therole of DPGMin the complex metabolic interrelation-ships within erythrocytes.

METHODS

MaterialsWith the exception of some reagents, whose source is speci-

fied in the text, all substrates and commercial enzymes used

J. Clin. Invest.©3 The American Society for Clinical Investigation, Inc., 0021-9738/7811101-907 $1.00 907Volume 62 November 1978 907-915

were purchased from Boehringer Mannheim Biochemicals,Indianapolis, Ind. Buffer salts were obtained from MerckChemical Div., Merck & Co., Inc., Rahway, N. J. Cellogelstrips were purchased from Medical Products, ChemetronCorp., Milan, Italy, whereas 2-phosphoglycolic acid, a-cellu-lose, and Sigma Cell type 50 were from Sigma Chemical Co.,St. Louis, Mo.

ExperimentsGeneral. Routine hematologic examinations, determined

by standard methods (15) and oxygen dissociation curves(ODC), were performed on freshly drawn, heparinized, ve-nous blood specimens. Resting cardiac output was deter-mined with technetium 99 (16). ODCof whole blood cells(2-Iul samples) were recorded at 37°C, pH 7.40, and at a Pco2of 40 mmHg with the Hem-O-Scan oxygen analyzer (Ameri-can Instrument Co., Inc., Silver Spring, Md.). The ODCof hemolysates were determined on samples stripped of or-ganic phosphates on a Dintzis column (17); ODCwere meas-ured according to the discontinuous spectrophotometricmethod of Benesch et al. (18). Measurements of 02, Pco2, andpH were made at 37°C with a BSM3Radiometer gas analyzer(Radiometer Co., Copenhagen, Denmark).

Enzyme assaysBlood collected in the presence of heparin was passed

through a mixture of a-cellulose and Sigma Cell type 50. Thewashed erythrocytes were treated as recommended by theInternational Committee for Standardization in Haematology(19) and enzyme assays were performed on the hemolysate.

Assay of DPGMactivity. The assay was performed accord-ing to Schroter and Kalinowski (20). The assay mixture (1 ml)contained 40 ,umol triethanolamine-HCl buffer (pH 7.5), 1,umol NAD, 7 ,mol fructose 1,6-diphosphate, 7 j,mol KH2PO4,2,umol 3-phosphoglycerate, 1 Ualdolase, 1 U triose phosphateisomerase, and 2.5 U of glyceraldehyde 3-phosphate dehydro-genase (GAPDH). After =5 min preincubation at 37°C, 0.05ml of a 1/15 hemolysate was added. Incubation was subse-quenitly carried out for 10 min and, when necessary, for periodsup to 30 min. DPGMactivity was usually evaluated by de-termination of the NADHformed in the course of the GAPDHreaction and in a few instances by measurement of the 2,3-DPGsynthesized in the DPGMreaction. For this purpose,at the start and after 30 min incubation, 1 ml of the mixturewas removed from the cuvette and deproteinized in a boilingbath for 10 min. After centrifugation, 0.5 ml of the supernatewas taken for determination of 2,3-DPG as described below.

DPGPassay. The assay was performed according to thefollowing principle which has previously been used for itselectrophoretic revelation (21), i.e., in the presence of 2-phos-phoglycolic acid, a strong activator of DPGP, the 3-phospho-glycerate liberated from 2,3-DPG is measured by a reactioncoupled to phosphoglycerate kinase and GAPDH. The assaymixture, in a volume of 1 ml, contained 50 ,umol triethanola-mine-HCl buffer (pH 7.5), 10 ,umol MgCl2, 3 ,umol Na2ATP,0.2 ,imol NADH, 0.8 ,umol 2,3-DPG, 3.3 UGAPDH,2 U phos-phoglycerate kinase and 1 ,umol 2-phosphoglycolic acidas the activator of DPGP. To the incubation medium, 0.05ml of a 1/15 hemolysate was added and incubation carriedout at 37°C for 10 min and, when necessary, up to 30 min.The transformation of NADHto NADat 334 nm was propor-tional to DPGPactivity.

MPGMassay. The assay was performed according to a sim-ilar principle to that employed for DPGP, i.e., the 3-phospho-glycerate formed from 2-phosphoglycerate was measured byreactions coupled to phosphoglycerate kinase and GAPDH.

The incubation medium was the same as that used for de-termination of DPGPactivity except for the following: (a)the 0.8 ,umol of 2,3-DPG was replaced by 0.08 ,umol 2,3-DPGtogether with 0.8 ,umol 2-phosphoglycerate; (b) 2-phospho-glycolic acid was omitted; (c) 10 ,ul of a 1/15 hemolysate wasused and the incubation was carried out at 25°C.

All other glycolytic enzymes, glutathione reductase, ade-nylate kinase, adenosine deaminase, and acetylcholinester-ase, in addition to reduced glutathione, were assayed accord-ing to Beutler et al. (19, 22). Pyrimidine 5'-nucleotidase ac-tivity was measured according to Valentine et al. (23) andinorganic phosphate according to Fiske and Subbarow (24).

Extracts for measurement of 2,3-DPG, ATP,ADP, AMP, and glycolytic intermediatesFresh blood was collected in heparin and deproteinized

extracts were prepared according to Oelshlegel et al. (25) byadding 1 vol of 1 N perchloric acid to 1 vol of blood. Theextracts were then centrifuged, the supernates neutralized toa pH -7.5, (with a solution of 0.5 Mtriethanolamine and 2 MK2CO3), and the precipitated perchlorate removed.

2,3-DPG assay. 2,3-DPG was assayed enzymatically ac-cording to Rose and Liebowitz (26), i.e., by determinationof the 3-phosphoglycerate liberated by the DPGPactivity ofrabbit muscle phosphoglycerate mutase after its stimulationby 2-phosphoglycolic acid. The 3-phosphoglycerate producedwas assayed by coupling the reaction to enolase, pyruvatekinase, and lactate dehydrogenase, and measuring the oxida-tion of NADHin the latter reaction. The incubation mixture(1 ml) contained Tris-HCl buffer, pH 7.5 (16 ,Lmol), KCI (3Jumol), MgCl2 (4 ,umol), ADP (1 ,umol), NADH(0.1 ,umol),2-phosphoglycolic acid (1 ,umol), enolase (4 U), pyruvate ki-nase (4 U), lactate dehydrogenase (5.5 U), and 20 jul of eachneutralized and deproteinized extract. The reaction was ini-tiated by the addition of phosphoglycerate mutase (12.5 U).The blank contained the complete assay mixture with theexception of 2-phosphoglycolic acid which was omitted.

ATP, ADP, AMP, and glycolytic intermediates from erythro-cytes were measured according to Oelshlegel (25) in the sameextracts as those employed for assay of 2,3-DPG.

Incubation of erythrocytes with inosine,inorganic phosphate, and pyruvateBlood stored in sterile flasks containing acid-citrate-dex-

trose (ACD) was supplemented with 10 mMinosine, 4 mMdisodium phosphate, and 4 mMpyruvate according to Oskiet al. (27) and incubated at 370C for 3 h. At the start of incuba-tion and after 2 h, 2-ml portions of the blood were removedand deproteinized. The perchloric acid-treated samples werethen processed, neutralized as described above, and subse-quently assayed for 2,3-DPG, ATP, fructose 1,6-diphosphate(FDP), triose phosphates (TP), glucose 6-phosphate (G6P),and fructose 6-phosphate (F6P).

Concentrated preparation of the enzymesBecause of the undetectable activity of DPGMand DPGP

in the hemolysate from the propositus, a more concentratedsolution was prepared to detect any trace amounts of theseenzymes that might be present. For this purpose hemolysateswere passed through a carboxy-methyl Sephadex column(Pharmacia Fine Chemicals, Piscataway, N. J.) equilibratedwith a 10-mM, sodium-phosphate buffer at pH 6.5. The firstprotein fractions eluted free of hemoglobin usually con-

908 R. Rosa, M-O. Prehu, Y. Beuzard, and J. Rosa

TABLE IHematological Data of the Family

Packed Mean Mean Meancell corpuscular corpuscular corpuscular

Individuals Age Hb Erythrocytes volume volume hemoglobin concentration Reticulocytes

yr gidl xlO12/liter % fl pg g/dl xlO9liter

C. M. (propositus) 42 17.4 5.32 54 94 32 33.5 53.10V. M. (son) 5 14.6 4.91 43.6 89 29.5 33.3 69.10L. M. (daughter) 4 14.3 5.14 43 83 27.0 32.9 95.10

tained the entire DPGMand DPGPactivities. These fractionswere pooled, concentrated by ultrafiltration, and then sub-mitted to electrophoresis.

Electrophoretic studyElectrophoresis of DPGM, DPGP, and MPGMwere car-

ried out according to methods previously described (8) onCellogel strips in a 0.075-M Tris-EDTA-citric acid buffer (pH8.0) at 220 V for 3 h. The samples applied on the strips were5 ,LI of 1/15 hemolysate in the case of MPGM,and 5-10 ,ulof the concentrated preparation of DPGMand DPGP. Afterelectrophoresis, each enzyme activity was revealed by pour-ing 4 ml of a mixture containing 1% agarose, in the samemedium as that used for the enzyme assay, over the respectivestrip. The fluorescent band of DPGMactivity and the darkbands of DPGPand MPGMactivities were detected withultraviolet light (-340 nm) after 5-30 min incubation at 25°C.

RESULTS

Case report. C. M., born in Paris in 1936 of Frenchextraction, was found on routine testing in 1974 to havea hemoglobin level of 19.0 g/dl blood. He was thentreated by phlebotomy every 2 mo for 2 yr and wasreferred in April 1977 to the Henri Mondor Hospital.Except for ruddy cyanosis, he was normal upon physi-cal examination. His spleen was not enlarged and norenal mass was apparent. Extensive cardiac and pul-monary investigations revealed no abnormality and his

z

= 50.

C,,

0 60P 02 mmHg

FIGURE 1 Oxygen dissociation curves of 2 ,lI of whole bloodfreshly drawn from propositus ( ) and control (--- -). Meas-urements were performed with a Hem-O-Scan; 2 ,u of sam-ples; temperature 37°C; CO2, 5%; pH 7.4.

resting cardiac output was normal. Arterial blood gasvalues were normal (Po2, 92 mmHg; Pco2, 45 mmHg; oxygen saturation, 96.5%). The absence of hemoly-sis was shown by the normal levels of haptoglobin andbilirubin. The hematological findings are detailed inTable I. In brief, they show an hematocrit of 54%, nor-mal erythrocyte morphology, and normal leukocyte andplatelet counts; the reticulocyte count was 1.2% (un-corrected). Total blood volume, erythrocyte volume,and plasma volume were 88.0 ml, 44.0 ml, and 41.0ml/kg body weight, respectively, thus revealing an ab-solute erythrocytosis. Analysis of the propositus' hemo-globin by the sensitive isofocusing method (28) did notshow the presence of an abnormal hemoglobin. Thelevel of hemoglobin A2 (2.3%) was normal. The oxygenaffinity curve of his stripped hemolysate was normal.In contrast, the oxygen affinity of his whole erythro-cytes, as determined on the Hem-O-Scan, was increased(Fig. 1), the partial pressure of oxygen at which hemo-globin is half saturated with 02 (P50) being 17.3 mmHg at pH 7.4 and Pco2 of 5% (normal values: 26.5±t1mmHg). The contrast between the normal oxygenaffinity of his stripped hemolysate and the highoxygen affinity of his intact erythrocytes was explainedby the level of erythrocyte 2,3-DPG which was-3%of the normal mean (Table I1).

Family study. The two parents of the proposituscould not be studied. His 5-yr-old son and his 4-yr-olddaughter were examined. In the absence of clinical

TABLE IIFamily Study of the Correlation between

2,3-DPG and P50 Values

2,3-DPG P"

Amol/glHb mmHg

Propositus 0.4 17.3Son 9.2 19.5Daughter 10.6 22.5Normal adult values 15±2 26.5±+1

2,3-DPG concentrations are expressed as micromoles per gramof hemoglobin. The P55 is expressed in millimeters ofmercury at which hemoglobin is half saturated with 02-

Homozygous Diphosphoglycerate Mutase Deficiency 909

abnormalities, their hemoglobin and hematocrit (TableI) were in the upper range of normal for this age group.This can be related to their erythrocyte 2,3-DPG levels,which were found to be decreased, relative to the nor-mal values. The erythrocyte P50 values were in accord-ance with the intracellular 2,3-DPG levels (Table II).

Erythrocyte enzymatic study. The very low levelof 2,3-DPG in the propositus erythrocytes led us tostudy the three enzymes which involve this compoundin their reactions, i.e., DPGM, DPGP, and MPGM.The results are summarized in Table III.

The DPGMactivity of the propositus erythrocyteswas undetectable under the usual conditions even afterrecording NADHformation for 30 min. To detect avery low DPGMactivity, a more sensitive techniquewas employed which involved measurement of the 2,3-DPG synthesized. We did not, however, detect anyadditional formation of 2,3-DPG even after 30 min ofincubation. It could be concluded that, at least underthese conditions, DPGMactivity was undetectable. Inthe propositus' two children, erythrocyte DPGMac-tivity was half the normal value.

DPGPactivity was also undetectable in the proposi-tus' erythrocytes and was half of the normal valuein the two children.

MPGMactivity was also about half of normal in thepropositus' erythrocytes and close to the normal in histwo children.

Activities of other enzymes. Enzymes of the Emb-den-Meyerhof pathway, glucose-6-phosphate dehydro-genase, 6-phosphogluconate dehydrogenase, glutathi-one reductase, adenosine deaminase, adenylate kinase,pyrimidine 5'-nucleotidase, and acetylcholinesterasewere within normal limits in both the propositus' andin his children's erythrocytes.

Study of some other erythrocyte compounds. Thelevels of the glycolytic intermediates, ATP, ADP, andAMPin the propositus' erythrocytes are shown in Fig.2. The very high levels of TP and FDP contrast withthe small amounts of G6P and F6P. The amount of

TABLE IIIFamily Study of DPGM,DPGP, and MPGM

DPGM DPGP MPGM

Ulg Hb/min

Propositus 0 0 27Son 2.53 1.17 36Daughter 2.45 1.15 39Normal adult values 5.72+0.68 2.64+0.45 46+8

The activities of DPGMare expressed as micromoles per gramof hemoglobin per minute of NADHformed in the reaction,whereas those of DPGPand MPGMare micromoles of NADHoxidized per gram of hemoglobin per minute during the re-actions.

:z

lx

zC.,

lx

G1-6P F6P TP PEP LACTG6P FDP 3PGA PYR

2 PGA

ATP AMPADP Pi

FIGuRE 2 Crossover plot of the concentration of glycolyticintermediates in the erythrocytes from the propositus (-0 -)and from his son (-- - -). The glycolytic compounds weredetermined after deproteinization in neutralized acid extractson fresh heparinized blood. Results are expressed for eachintermediate as a percentage of the normal mean calculatedfrom 10 controls. PEP denotes phospho-enol-pyruvate; Pidenotes inorganic phosphates; PYR represents pyruvate;LACT denotes lactate.

ATP is higher than normal, whereas those of ADPandof AMP are lower. The inorganic phosphate levelwas normal but the concentration of reduced gluta-thione was 70 mg per 100 ml of erythrocytes, a valuehigher than normal (50± 11 mg).

Effect of storage. Storage in ACDof normal eryth-rocytes induces a rapid decrease in 2,3-DPG contentand significant alteration in the levels of several phos-phorous compounds (Fig. 3). In the erythrocytes of thepropositus after 2 wk storage, the concentrations ofthese compounds were strikingly modified and had be-come very close to that of the control. More precisely,the values of FDP and TP which were markedly in-

1400

a 1300

C,,

(~300

z" ,

lx 100

0.

G6P F6P FDP TP 2,3-DPG ATP

FIGURE 3 The concentrations of glycolytic intermediates inheparinized fresh blood (-0 -) and in ACD-collected bloodfrom the propositus (- - 0 - -), after 2 wk storage, are comparedto those in ACDcollected blood from a control (- - O - -), afterthe same period of storage. Results are expressed as a per-centage of the concentrations in normal fresh blood.

910 R. Rosa, M-O. Prehu, Y'. Beuzard, and J. Rosa

creased in his fresh blood fell close to those of thecontrol, whereas the concentration of G6Prose in a sim-ilar manner to that in the control.

Incubation with inosine, inorganic phosphate, andpyruvate. The results of incubation of the propositus'erythrocytes with inosine, inorganic phosphate, andpyruvate are compared to those of a control in TableIV and Fig. 4. A slight increase in 2,3-DPG concen-tration was detected in the propositus' erythrocytes,while its level was markedly elevated in the control,to as much as three times the normal value. In addi-tion, all the glycolytic intermediates were increasedin concentration in both the propositus' and in the con-trol's erythrocytes, but their relative levels differed.In the control's erythrocytes, G6P showed the greatestelevation of all the ester phosphates, whereas in thepropositus' cells, FDP was the most increased.

Electrophoretic study. Although the activities oferythrocytes DPGMand DPGPwere undetectable inthe propositus and low in his children, we attemptedto study the electrophoretic pattern of these enzymes.Despite the use of highly concentrated preparations,we could not detect any band of DPGMactivity inthe sample from the propositus. On the contrary, oneband of enzymatic activity appeared in his children'serythrocytes and this band migrated as did the controlsample. Similarly DPGPactivity did not appear in thesample from the propositus but was revealed in hischildren's erythrocytes as one band that moved as thecontrol moved and to the same position as the bandof DPGMactivity. Electrophoresis of MPGMactivitydid not necessitate the use of a concentrated prepara-tion as did DPGMbecause MPGMactivity was suf-ficiently high in the propositus' erythrocytes to permitits identification in crude hemolysates. The pattern ob-tained from the propositus was different from that ofhis children and from the control (Fig. 5). Three bandsof MPGMactivity appeared in the children's hemoly-sates, showing a pattern identical to that of the control,whereas only two bands were detected in the hemoly-sate from the propositus. These bands did not showany modification of charge when compared with the

TABLE IV2,3-DPG Level in Erythrocytes of Control and Propositus

Incubated with Inosine, Pyruvate, andInorganic Phosphate*

Erythrocytes

OTime 2 h

,umol/ml

Propositus 0.12 0.55Control 1.88 6.91

* See conditions in legend of Fig. 4.

1000

2 800

C

600a0

cr

zui 400z0C)

0200

O.)0. f

G6P F6P FDP TP ATP 3PG

FIGURE 4 The levels of glycolytic intermediates, in ACDblood stored for 2 wk, after incubation in 10 mMinosine,4 mMNa2 PO4, and 4 mMpyruvate at 37°C for 2 h. Depro-teinized extracts were made both before and after 2 h incuba-tion. The values of the intermediates after 2 h incubation areexpressed as percentages of their respective concentrations at0 time. - O- Denotes control erythrocytes and - - 0 - - repre-sents propositus erythrocytes.

corresponding bands from the control hemolysate. Itshould be observed that the band that was absent inthe propositus corresponded to that which has beendemonstrated to carry DPGMand DPGPactivities (8).

DISCUSSION

This paper reports the first case of a homozygous de-ficiency of erythrocyte DPGMand the associated lowlevel of 2,3-DPG. This diminished 2,3-DPG value re-sulted in an increased erythrocyte oxygen affinity anda moderate compensatory erythrocytosis. The defect

N P SFIGuRE 5 Electrophoretic patterns of MPGMactivity fromthe hemolysates of a control (N), of the propositus (P), and ofhis son (S). Electrophoresis was performed on cellulose ace-tate strips with a Tris-EDTA citric buffer at pH 8.0. MPGMspecific staining and detection with ultraviolet light at -340nmrevealed the MPGMactivity as dark bands on a fluorescentbackground.

Homozygous Diphosphoglycerate Mutase Deficiency

IV-,

911

was well tolerated by the propositus who exhibitedno sign of hypoxemia, displayed normal activity, andspent several holidays at high altitude without diffi-culty. Such a situation closely resembles that presentedby subjects heterozygous for a hemoglobin with highoxygen affinity in whomthe disorder is mostly benign.There was no evidence for hemolysis in our propositusnor in his children in contrast with the cases reportedby Schroter (13) and Travis et al. (14) who have de-scribed two unrelated families having low DPGMac-tivities associated with hemolysis. In the case of Traviset al., an associated disorder responsible for the de-scribed microspherocytosis could cause the hemolysis.

The crossover plot of the levels of glycolytic sub-strates in the propositus' erythrocytes (Fig. 2) docu-ments the effect of the absence of 2,3-DPG on erythro-cyte glycolysis. This plot is characterized by high levelsof TP, FDP, and phosphoglycerates which contrastwith low levels of both G6Pand F6P. Similar modifica-tions, though less pronounced, have been found in thechildren's erythrocytes, and also in an unrelatedheterozygous propositus (12). Again these results con-trast with those obtained by Travis et al. (14) who didnot find any modification in the glycolytic substratesof their patient's erythrocytes. In our propositus, themarkedly increased amounts of TP and FDPare prob-ably the consequence of an accumulation of 1,3-phos-phoglycerate (1,3-DPG). Unfortunately a direct deter-mination of this phosphorus compound could not beperformed on account of its lability in the extracts. Suchan elevation is, however, highly probable because un-der normal conditions more than half of the 1,3-DPGpasses through the Rapoport shunt (29), which isblocked in the propositus' erythrocytes. An additionalmechanism could also explain the high levels of FDPin the propositus' erythrocytes and the low levels ofG6P and F6P. Thus, these modifications could arisefrom a stimulation of phosphofructokinase which couldbe repressed in normal erythrocytes by 2,3-DPG (30).Such a stimulation of phosphofructokinase would alsocontribute to the high levels of TP and of 1,3-DPG.

The total quantity of 1,3-DPG would then becomeavailable to phosphoglycerate kinase and be directedtowards lactate formation, thereby increasing both thelevel of the intermediary products and the rate of ATPformation. Thus, the high level of ATP in the proposi-tus' erythrocytes probably is the consequence of thestimulation of phosphoglycerate kinase as a result ofthe increase of available 1,3-DPG. In addition, the veryhigh concentration of FDP, a well-known activator ofpyruvate kinase, should contribute to this elevationof ATP.

The low concentration of G6P in the propositus'erythrocytes is not consistent with the inhibitory ef-fect of 2,3-DPG on hexokinase found by some authors(31). This apparently contradictory finding is probably

because of an increased utilization of G6P which isproduced in excess by stimulated hexokinase andpasses through the hexose monophosphate shunt; thispathway should itself be stimulated by the low levelof 2,3-DPG (32). Such a stimulation could also contrib-ute to the large amounts of TP, which are a convergencepoint of the pentose and Embden-Meyerhof pathways.

Another problem concerns the increase in glucose1,6-diphosphate level that is related to the decreasein G6Pconcentration. This apparently paradoxical find-ing could be explained by a stimulation of phospho-glucomutase; the phosphorylated form of this enzymereacts with G6P to give a glucose 1,6-diphosphate-en-zyme complex that is subsequently transformed intoglucose 1-phosphate and phospho-enzyme (PE) ac-cording to the two-step mechanism: (a) G6P+ PE i±glucose 1,6-di-PE, (b) Glucose 1,6-di-PE ± glucose 1-P+ PE. Thus, under normal conditions, glucose 1,6-di-phosphate is not released during the course of the nor-mal reaction. As demonstrated by Alpers (33), phos-phoglucomutase may be phosphorylated by 1,3-DPGwhen the latter is present at a high concentration inthe medium. Under these conditions, it competes withglucose 1,6-diphosphate which is released from the en-zyme complex by the following sequence: (c) Glucose1,6-di-PE + 1,3-DPG ± glucose 1,6-di-P + PE + 3-phosphoglyceric acid. Thus, the PEcan again react withG6P, the level of the latter decreasing in favor of theformation of glucose 1,6-diphosphate. Such an eventu-ality could be considered in the propositus' case be-cause a high concentration of 1,3-DPG could resultfrom the DPGMdefect.

It is of some interest to compare the levels of gly-colytic intermediates in the erythrocytes from the pro-positus with those from some animal species that ex-hibit a low level of 2,3-DPG in their erythrocytes (34).The concentrations of these compounds in the erythro-cytes of these animal species are quite different fromthose found in the propositus. These data apparentlyindicate that the role played by 2,3-DPG in regulatingthe metabolite flow inside the erythrocytes is not ofpredominant importance. In all probability the ob-served differences are a result of enzymatic ratios thatvary from one species to another. Some other differentdata can be observed in human erythrocytes that havebeen stored in ACD solution and that exhibit a fallin their 2,3-DPG concentration (34). FDPand TP wereslightly increased in this condition. Moreover, theerythrocytes of the propositus, when stored under thesame conditions, showed a fall in the level of all theirglycolytic compounds which were originally elevated,i.e., FDP, TP, and 3-phosphoglyceric acid, leading toan increase in lactate. On the contrary, when the eryth-rocytes from the propositus and from the control wereincubated in a medium containing inosine, inorganicphosphate, and pyruvate, the resulting increases in

912 R. Rosa, M-O. Prehu, Y. Beuzard, and J. Rosa

FDP, TP, and 3-phosphoglyceric acid were more ele-vated in the propositus' than in the control cells,whereas the level of G6P was more increased in thelatter. Thus, the erythrocytes of the propositus appearto represent unique material in which to evaluate thecomplex metabolic interrelationships within erythro-cytes.

Small but measurable amounts of 2,3-DPG are pres-ent in the propositus' erythrocytes. These results mustbe correlated with the absence of any detectableDPGMactivity. Perhaps this discrepancy results fromthe fact that our technique for estimation of DPGMis unable to detect the minute amounts of DPGMac-tivity which are nevertheless sufficient to produce thelow quantities of 2,3-DPG present in the propositus'erythrocytes. An alternative explanation arises from theobservation of Laforet et al. (35), according to whichthe low quantities of 2,3-DPG could be synthesizedby the MPGMin erythrocytes. It should be recalledthat these results could not be reproduced by Sheibleyand Hass (36) with purified enzyme.

Electrophoretic patterns have shown that bands ofDPGMand DPGPwere absent from the propositus'erythrocytes, confirming previous reports that providedevidence that DPGMcontained DPGPactivity (6, 9,10). The absence of any detectable DPGPactivity con-firms the results of Hass and Miller (37) who concludedthat only one DPGPexisted in erythrocytes. Thesedata are not in accordance with the observations ofSheibley and Hass (36) on the purified enzyme, butthe techniques used by the latter to determine DPGPdiffered from our own and thus could explain thesedivergent findings. Such a DPGMdeficiency shouldbe a result either of an absence of synthesis of thismolecule, or of the existence of an unstable molecule,or of the presence of an inactive enzyme. Detectionof eventual cross-reactive material should be under-taken to try to solve this problem.

The data on MPGMvalues in the erythrocytes ofour family are of interest. They involve MPGMvaluesclose to normal and a normal electrophoretic patternin the children's erythrocytes, which contrasts with adiminution of 50% in the MPGMactivity of the pro-positus' cells and with a disappearance of the mostanodic band on the MPGMelectrophoretic pattern.Its absence is consistent with our previous demonstra-tion (8) that this band corresponded to a MPGMac-tivity carried by the DPGM-DPGPmolecules. Never-theless, because the part of MPGMactivity carried byDPGMrepresents only 5%of the total MPGMactivityof the erythrocytes, the DPGMdeficiency cannot ex-plain the absence of 50% of MPGMactivity in thepropositus' cells. Also intriguing are the MPGMvaluesclose to normal found in the children's erythrocytes.To solve this problem, one can formulate the hypothe-sis that either a lack of synthesis or a modification of

100

C-)

75.

50.

251

0 1 2 3INCUBATION TIME (hours)

FIGuRE 6 Heat stability of MPGMas a function of time incrude 1/15 hemolysates from the propositus (- 0-) and fromcontrol (- - A - -) and after addition of 5 mM2,3-DPG in hemo-lysates from the propositus (- -) and from control (-- A --).Incubation was performed at 50°C. The noncentrifuged hemo-lysates contained 0.7 mMB-mercaptoethanol and 1 mMEDTA. Results are expressed as a percentage of the MPGMactivity of each nonincubated hemolysate.

a peptide chain common to DPGMand MPGMhasoccurred. The structures of these two enzymes havebeen partially studied and their molecules have beenfound to be constituted of two chains of the same mo-lecular weight (10, 36, 38). However, the molecularweights of the two chains of MPGMdiffered from thoseof the DPGMchains (39). Under these conditions, thehypothesis of a common chain seems an unlikely ex-planation.

A more attractive hypothesis involves a possible pro-tective effect of 2,3-DPG on MPGM.After considera-tion of the results obtained by Omennand Hermodson(40) in brain and by Borders and Wilson (41) in eryth-rocytes, we have studied the effect of 2,3-DPG onMPGMthermostability in hemolysates from the pro-positus and from a control (Fig. 6). In the absence of2,3-DPG in the hemolysate, the MPGMof the pro-positus lost about 70% of its activity after 2 h at 50°C,whereas that of the control lost only 15%. In the pres-ence of 2,3-DPG, no loss of MPGMactivity could bedetected either in the propositus or in control. Thesepreliminary results lead us to suppose that the partialMPGMdeficiency observed in the propositus' erythro-cytes is because of the low level of 2,3-DPG. Thus,MPGMwould be normally synthesized in the erythro-blasts of the propositus, but as a result of the DPGMde-ficiency resulting in insufficient concentration of 2,3-DPGthis enzyme would be unstable and would par-tially lose its activity during erythrocyte maturation.Because the children's erythrocytes contain apprecia-

Homozygous Diphosphoglycerate Mutase Deficiency

00"ll<

913

ble amounts of 2,3-DPG, substantial denaturation ofMPGMshould not occur.

ACKNOWLEDGMENTS

The authors wish to acknowledge Dr. R. E. Benesch and Dr.R. Benesch for their advice and for stimulating discussionand Dr. John Chapman for careful assistance in reviewingthe manuscript.

This work was supported in part by l'Institut National dela Sante et de la Recherche Medicale, grant 76-5-133, by l'Uni-versite de Paris Val de Mamne and by la Fondation pour laRecherche Medicale Frangaise.

REFERENCES

1. Rapoport, S., and J. Luebering. 1950. The formation of2,3-diphosphoglycerate in rabbit erythrocytes: the exist-ence of a diphosphoglycerate mutase. J. Biol. Chem. 183:507-516.

2. Rapoport, S., and J. Luebering. 1951. Glycerate 2,3 di-phosphatase.J. Biol. Chem. 189: 683-694.

3. Rose, Z. B. 1968. The purification and properties of di-phosphoglycerate mutase from human erythrocytes. J.Biol. Chem. 243: 4810-4820.

4. Harkness, D. R., and S. Roth. 1969. Purification and prop-erties of 2,3-diphosphoglyceric acid phosphatase from hu-man erythrocytes. Biochem. Biophys. Res. Commun. 34:849-856.

5. Rose, Z. B., and J. Liebowitz. 1970. 2,3-diphosphoglycer-ate phosphatase from human erythrocytes. J. Biol. Chem.245: 3232-3241.

6. Rosa, R., J. Gaillardon, and J. Rosa. 1973. Diphosphogly-cerate mutase and 2,3-diphosphoglycerate phosphataseactivities of red cells: comparative electrophoretic study.Biochem. Biophys. Res. Comm. 51: 536-542.

7. Rosa, R., I. Audit, and J. Rosa. 1975. Presence of a diphos-phoglycerate phosphatase activity in the multiple molec-ular forms of 3-phosphoglycerate mutase. In Isozymes.C. L. Markert, editor. Academic Press, Inc., NewYork. I:695-706.

8. Rosa, R., I. Audit, and J. Rosa. 1975. Evidence for threeenzymatic activities in one electrophoretic band of 3-phosphoglycerate mutase from red cells. Biochimie.(Paris) 57: 1059-1063.

9. Sasaki, R., K. Ikura, E. Sugimoto, and H. Chiba. 1975.Purification of Bis-phosphoglyceromutase, 2,3-Bis-phos-phoglycerate phosphatase and phosphoglyceromutase.Europ. J. Biochem. 50: 581-593.

10. Kappel, W. K., and L. F. Hass. 1976. The isolation andpartial characterization of diphosphoglycerate mutasefrom human erythrocytes. Biochemistry. 15: 290-295.

11. Harkness, D. R., J. Ponce, and V. Grayson. 1969. A com-parative study on the phosphoglyceric acid cycle in mam-malian erythrocytes. Comp. Biochem. Physiol. 28: 129-132.

12. Labie, D., J. P. Leroux, A. Najman, and C. Reyrolle.1970. Familial diphosphoglycerate mutase deficiency. In-fluence of the oxygen affinity curves of hemoglobin.FEBS (Fed. Eur. liochem. Soc.) Lett. 9: 37-40.

13. Schroter, W. 1965. Kongenitale Nichtspharocytiire Ana-mie bei 2,3-Diphosphoglyceratmutase-Mangel der Eryth-rocyten im fruhen Sauglingsalter. Klin. Wochenschr. 43:1147-1152.

14. Travis, S. F., J. Martinez, and J. Garvin. 1976. Study ofa kindred 'with red cell 2,3-diphosphoglycerate mutase

deficiency and compensated hemolysis. Blood. 48: 993,104. (Abstr.)

15. Dacie, J. V., and S. M. Lewis. 1975. Practical haematology.Churchill Livingstone. London. 5th edition.

16. Bardy, A., H. Fouye, R. Gobin, G. de Tovar, C. Panneciere,and M. Hegesippe. 1975. Technetium 99 m labeling bymeans of stannous pyrophosphate: application to bleomy-cin and red blood cells. J. Nucl. Med. 16: 435-437.

17. Nozaki, Y., and C. Tanford. 1967. Examination of titrationbehavior. In Methods in Enzymology. G. H. W. Hirs,editor. Academic Press, Inc., NewYork and London. 11:715-734.

18. Benesch, R. G., G. Macduff, and R. E. Benesch. 1965.Determination of oxygen equilibria with a versatile newtonometer. Anal. Biochem. 11: 81-87.

19. Beutler, E., K. G. Blume, J. C. Kaplan, G. W. Lohr, B.Ramot, and W. N. Valentine. 1977. International commit-tee for standardization in haematology: recommendedmethods for red cell enzyme analysis. Br. J. Haematol.35: 331-340.

20. Schroter, W., and W. Kalinowsky. 1969. Erythrocyte 2,3-diphosphoglycerate mutase: an optical test in hemoly-sates. Chim. Clin. Acta. 25: 283-285.

21. Rosa, R., J. Gaillardon, and J. Rosa. 1973. Characteriza-tion of 2,3-diphosphoglycerate phosphatase activity: elec-trophoretic study. Biochim. Biophys. Acta. 293: 285-289.

22. Beutler, E. 1971. Red Cell Metabolism. A manual Bio-chemical Methods. Grune & Stratton, Inc. NewYork andLondon.

23. Valentine, W. N., K. Fink, D. E. Paglia, S. R. Harris, andW. S. Adams. 1974. Hereditary hemolytic anemia withhuman erythrocyte pyrimidine 5'-nucleotidase. J. Clin.Invest. 54: 866-879.

24. Fiske, Ch., and V. Y. Subbarow. 1925. The colorimetricdetermination of phosphorus.J. Biol. Chem. 66: 375-400.

25. Oelshlegel, F. J., G. J. Brewer, J. A. Penner, and E. B.Schoomaker. 1972. Enzymatic mechanisms of red celladaptation to anemia. In Hemoglobin and red cell struc-ture and function. G. J. Brewer, editor. Plenum Press,Inc., New York and London. 377-396.

26. Rose, Z. B., and J. Liebowitz. 1970. Direct determina-tion of 2,3-diphosphoglycerate. Anal. Biochem. 35: 177-180.

27. Oski, F. A., S. G. Travis, L. D. Miller, M. D. Papadopou-lous, and E. Cannon. 1971. The in vitro restoration ofred cell 2,3-diphosphoglycerate levels in banked blood.Blood. 37: 52-58.

28. Bunn, H. F. 1973. The use of gel electrofocusing in thestudy of human hemoglobins. Ann. N. Y. Acad. Sci. 209:345-353.

29. Keitt, A. 1966. Pyruvate kinase deficiency and related dis-orders of red cell glycolysis. Am. J. Med. 41: 762-785.

30. Beutler, E., and E. Guinto. 1973. The effect of 2,3-DPGon red cell phosphofructokinase. FEBS (Fed. Eur. Bio-chem. Soc.) Lett. 37: 21-22.

31. Ponce, J., S. Roth, and D. R. Harkness. 1971. Kinetic stud-ies of the inhibition of glycolytic kinases of human eryth-rocytes by 2,3-DPG. Biochim. Biophys. Acta. 250: 63-74.

32. Dische, Z. 1964. The pentose phosphate metabolism inred cells. In The Red Blood Cell. C. Bishop and D. M.Surgenor, editors. Academic Press, Inc., NewYork. 189-209.

33. Alpers, J. B. 1968. Studies on phosphomutases. J. Biol.Chem. 243: 1698-1704.

34. Bartlett, G. R. 1970. Pattern of phosphate compounds inred blood cells of man and animals. In Red Cell Metabo-

914 R. Rosa, M-O. Prehu, Y. Beuzard, and J. Rosa

lism and Function. G. J. Brewer, editor. Plenum Press,New York and London. 6: 245-256.

35. Laforet, M. T., J. B. Butterfield, and J. B. Alpers. 1974.The biosynthesis of 2,3-diphosphoglycerate by mono-phosphoglycerate mutase from muscle and erythrocytes.Arch. Biochem. Biophys. 165: 179-187.

36. Sheibley, R. H., and L. F. Hass. 1976. Isolation and par-tial characterization of monophosphoglycerate mutasefrom human erythrocytes.J. Biol. Chem. 251: 6699-6704.

37. Hass, L. F., and K. B. Miller. 1975. A reassessment ofthe phosphoglycerate bypass enzymes in human erythro-cytes. Biochem. Biophys. Res. Comm. 66: 970-979.

38. Rose, E. B., and R. G. Whalen. 1973. The phosphoryla-

tion of diphosphoglycerate mutase. J. Biol. Cheml. 248:1513-1519.

39. Hass, L. F., R. H. Sheibley, W. K. Kappel, and K. B. Miller.1976. A comparison of some of the physical and chemicalproperties of the phosphoglycerate mutases from humanerythrocytes. Biochem. Biophys. Res. Comnm1. 73: 976-983.

40. Omenn, G. S., and M. A. Hermodson. 1975. Humanphos-phoglycerate mutase: isozyme marker for muscle differen-tiation and for neoplasia. In Isozymes. C. L. Markert, edi-tor. Academic Press, Inc., New York. III: 1006-1018.

41. Borders, C. L., and B. A. Wilson. 1976. Phosphoglyceratemutase has essential arginyl residues. 73: 978-984.

Homozygous Diphosphoglycerate Mutase Deficiency 915