Human erythrocyte hexokinase deficiency. Characterization of a mutant enzyme with abnormal regulatory properties.

In the erythrocytes of a patient with hereditary nonspherocytic hemolytic anemia, a homozygous expression of hexokinase deficiency was detected. The mutant enzyme was characterized by normal kinetic parameters with respect to its substrates, glucose and MgATP2-, normal pH optimum, normal heat stability at 40 degrees C, but abnormal behavior with respect to its regulation by glucose-1,6-diphosphate and inorganic phosphate, and an altered electrophoretic pattern. Interpretation of the results revealed the presence of two different hexokinases type I in normal human erythrocytes: one enzyme with a high affinity for glucose-1,6-diphosphate, the inhibition of which is regulated by inorganic phosphate; and another enzyme with a lower affinity for the inhibitor, not regulated by inorganic phosphate. The former enzyme was not detectable in the erythrocytes of the patient, whereas the presence of the latter enzyme could be demonstrated.

Pyruvate kinase and the "high ATP syndrome".

The erythrocytes of a patient with the so-called "high ATP syndrome" were characterized by a high ATP content and low 2,3-diphosphoglycerate level. The pyruvate kinase activity was specifically increased (about twice the normal level). After separation of the erythrocytes according to age by discontinuous Percoll density centrifugation, the pyruvate kinase activity was found to be increased in all Percoll fractions. Pyruvate kinase of the patient's cells was characterized by a decreased K0.5 for the substrate phosphoenolpyruvate and no inhibition by ATP. The Michaelis constant (Km) value for ADP, the nucleotide specificity, the thermostability, pH optimum, and immunological specific activity were normal. It is concluded that the high pyruvate kinase activity is due to a shift in the R(elaxed) in equilibrium T(ight) equilibrium to the R(elaxed) form.

Orotic aciduria in two unrelated patients with inherited deficiencies of purine nucleoside phosphorylase.

The urines of two unrelated children with inherited deficiencies of purine nucleoside phosphorylase have been found to contain significant quantities of orotic acid in addition to the previously reported purine nucleosides. The data are consistent with some cell types of these immunodeficient patients being deplete of pyrophosphoribosylphosphate, a precursor of both purine, and pyrimidine nucleosides. It is suggested that the pyrophosphoribosyl-phosphate-depleted cells may be some component of the thymus-dependent immune system.

Deoxyguanosine triphosphate as a possible toxic metabolite in the immunodeficiency associated with purine nucleoside phosphorylase deficiency.

Purine nucleoside phosphorylase (PNP) deficiency is associated with a severe defect in thymus-derived (T)-lymphocyte function combined with normal bone marrow-derived (B)-lymphocyte function. To investigate the role of this enzyme deficiency in the resulting immune dysfunction, we measured the levels of ribonucleoside and deoxyribonucleoside triphosphates in erythrocytes from two unrelated PNP-deficient, T-lymphocyte-deficient patients. Both PNP-deficient patients have abnormally high levels of deoxyguanosine triphosphate (deoxy-GTP) in their erythrocytes (5 and 8 nmol/ml packed erythrocytes). In contrast, normal controls and adenosine deaminase-deficient, immunodeficient patients do not have detectable amounts of deoxyGTP (<0.5 nmol/ml packed erythrocytes). We propose that deoxyguanosine, a substrate of PNP, is the potentially lymphotoxic metabolite in PNP deficiency. The mechanism of toxicity involves phosphorylation of deoxyguanosine to deoxyGTP, which acts as a potent inhibitor of mammalian ribonucleotide reductase.

Erythrocyte metabolism in purine nucleoside phosphorylase deficiency after enzyme replacement therapy by infusion of erythrocytes.

Purine nucleoside phosphorylase deficiency is associated with a severely defective T-cell immunity. A patient with purine nucleoside phosphorylase deficiency was treated with transfusions of irradiated erythrocytes and plasma. This resulted in a remarkable correction of the metabolic disturbances in the patient. The urinary excretion of inosine, deoxyinosine, guanosine, and deoxyguanosine decreased, whereas uric acid excretion as well as serum uric acid concentration increased. It could be shown that the enzyme activity of the circulating erythrocytes correlated inversely with the urinary excretion of nucleosides and directly with the excretion of uric acid. As a consequence of the therapy, several glycolytic intermediates of the erythrocytes were increased, especially 2,3-diphosphoglycerate. The high 2,3-diphosphoglycerate level caused a shift to the right of the oxygen dissociation curve (P50 = 32.9 mm Hg). The immunological status of the patient showed definite improvement after the enzyme replacement therapy.

Isozymic composition and regulatory properties of phosphofructokinase from well-differentiated and anaplastic medullary thyroid carcinomas of the rat.

Acceleration of glycolysis is, in general, a characteristic of neoplasia. Previous studies have shown that this increase in glycolysis is achieved by quantitative increases in the activities of the key regulatory enzymes, hexokinase, phosphofructokinase (PFK) and/or pyruvate kinase, which are often accompanied by isozymic alterations that facilitate glycolysis. In this study, we investigated the alterations in the activity, isozymic profile, and kinetic-regulatory properties of PFK from the medullary thyroid carcinomas of the rat, which represent a model for the neuroectodermally derived tumors in humans. Contrary to the expected, we found that undifferentiated tumors showed a decrease in the enzyme activity as compared to the highly differentiated tumors. This decrease in PFK activity was accompanied by an increase in the expression of the liver-type isozyme of PFK. The enzymes from the 2 tumor types showed no significant differences in their affinity and cooperativity toward the substrates, fructose 6-phosphate and adenosine triphosphate (ATP). However, the tumor PFKs showed major differences with respect to their behavior toward the allosteric regulators of the enzymes, ATP, citrate, and fructose 2,6-diphosphate; the latter is a recently discovered activator of the enzyme. The enzyme from the undifferentiated tumor was less sensitive to citrate inhibition, which was more readily reversed by cyclic adenosine 3':5'-monophosphate. In addition, it was less sensitive to ATP inhibition at low fructose 6-phosphate concentrations. More importantly, the enzyme from the undifferentiated tumors was more sensitive to the activation by fructose 2,6-diphosphate especially when inhibited by citrate and ATP. The altered regulatory properties of the enzyme from the undifferentiated tumors most probably reflect its altered isozymic composition, i.e., increase in the liver-type isozyme. The preferential expression of the liver-type isozyme by undifferentiated and rapidly replicating cancer cells may be explained in terms of the unique regulatory properties of this isozyme. Although the concentrations of fructose 2,6-diphosphate were comparable in these 2 tumor types, the higher sensitivity of the liver-type PFK to activation by this compound may permit accelerated glycolytic flux observed in undifferentiated tumors, despite a decrease in total PFK activity.

Subunit composition, regulatory properties, and phosphorylation of phosphofructokinase from human gliomas.

In this study, we investigated the alterations in the activity, subunit profile, and kinetic regulatory properties of phosphofructokinase (PFK) from human gliomas compared with those from normal human brain. Gliomas showed a decrease in the enzyme activity as compared to normal brain. This decrease in PFK activity was accompanied by a relative increase in the expression of the liver type subunit of PFK. The enzymes from the tumor and normal brain showed no significant differences in their affinity toward the substrate fructose 6-phosphate. However, tumor and normal brain PFK showed major differences with respect to their behavior towards citrate and fructose 2,6-bisphosphate. The enzyme from the gliomas was less sensitive to citrate inhibition. More importantly, the enzyme from the tumor was more sensitive to the activation by fructose 2,6-bisphosphate. In addition, we found that in gliomas the L-type subunit could be phosphorylated, most probably by a cyclic AMP-independent protein kinase. This phosphorylation could not be detected in normal human brain. It is proposed that the preferential expression of the liver type subunit by undifferentiated cancer cells may be explained in terms of the unique regulatory properties of this isozyme.

Characterization of some glycolytic enzymes from human retina and retinoblastoma.

GLycolytic enzymes were studied from normal human retinas (both fetal and adult) and from retinoblastomas of eight patients and an established retinoblastoma cell line. No significant differences were found between the enzyme activities in the tissues investigated except for hexokinase and pyruvate kinase, which were significantly decreased in the tumor cells. In fetal retina, five different forms of pyruvate kinase could be detected by electrophoresis (K4, K3M, K2M2, KM3, and M4). In adult retina the K4 isozyme is almost absent, while in retinoblastoma the M4 isozyme is hardly present. In the retinoblastoma cell line, the M4 isozyme is completely absent. Alanine inhibition of pyruvate kinase from the retinoblastoma cell line is more inhibited compared to the pyruvate kinase of fetal retina and retinoblastoma and is even more inhibited compared to adult retina. Electrophoresis of aldolase from adult retina revealed the presence of all potential A-C hybrids (A4, A3C, A2C2, AC3, and C4). Fetal retina, however, is characterized by the predominance of the A type. The same patterns were observed in the retinoblastoma cell line and retinoblastoma. However, in other brain tumors, e.g., gliomas of adults, a five-membered A-C hybrid set is found. Electrophoresis of hexokinase from normal fetal and adult retina revealed the predominance of hexokinase type I; retinoblastoma and retinoblastoma cell line are both characterized by the presence of considerable amounts of hexokinase type II. The isozyme shifts in retinoblastoma result in an enzyme pattern identical to that of fetal retina except for the presence of hexokinase type II.

Phosphorylation of pyruvate kinase type K in human gliomas by a cyclic adenosine 5'-monophosphate-independent protein kinase.

In recent years, we reported the isozyme shift of pyruvate kinase from the M- toward the K-type in human neuroectodermal tumors. To investigate whether this shift enables phosphorylation of pyruvate kinase in these tumors, we studied 29 different specimens of human brain tumors for endogenous pyruvate kinase phosphorylation. While in normal human brain no phosphorylation of pyruvate kinase was detected, in all brain tumors pyruvate kinase became phosphorylated. There was no correlation between the extent of the pyruvate kinase phosphorylation and the histological classification and grading or the pyruvate kinase isozyme composition of the tumors. Only pyruvate kinase type K, and not type M, served as a substrate in the phosphorylation reaction. The phosphorylation of pyruvate kinase could be completely inhibited by addition of fructose 1,6-bisphosphate, a positive effector of pyruvate kinase type K; alanine, however, a negative effector, and phospho-enol-pyruvate, a substrate in the pyruvate kinase reaction, had no effect. While pyruvate kinase type L in liver is phosphorylated by a cyclic AMP-dependent protein kinase, the incorporation of phosphate into pyruvate kinase in human brain tumors appeared to be cyclic AMP independent and occurred exclusively on serine residues.

Isozymes of pyruvate kinase from human brain, meningiomas, and malignant gliomas.

Pyruvate kinase isozymes were studied in normal brain tissue (both fetal and adult) and in meningiomas and malignant gliomas. In fetal brain five different forms could be detected by electrophoresis (K4, K3M, K2M2, KM3, and M4). In adult brain the M4-type, K3M hybrid, and K4-type are present; the M isozyme is largely predominant. Alanine inhibition of pyruvate kinase is in agreement with the electrophoretic pattern. Pyruvate kinase from fetal brain and brain of a newborn is more inhibited compared with pyruvate kinase from adult brain. The Lineweaver-Burk plots for pyruvate kinase from fetal brain and brain of the newborn are nonlinear due to the presence of hybrids. Pyruvate kinase from meningiomas and malignant gliomas is strongly inhibited by alanine. Electrophoresis proved the presence of mainly K4 type and the hybrid K3M, which is in agreement with the alanine inhibition. Determination of the Km's for phosphoenolpyruvate supports this conclusion. The determination of the alanine inhibition of pyruvate kinase may be a diagnostic tool in surgery for gliomas.

Tyrosine kinase activity in breast cancer, benign breast disease, and normal breast tissue.

Tyrosine specific protein kinase activity was determined in 70 specimens of the human mammary gland. These included 28 cancers of the breast, 21 benign breast diseases, and 21 normal breast tissues. We measured tyrosine kinase activity in the cytosol fraction and in the membrane fraction of the homogenates. In addition cytosolic aldolase activity was measured. Tyrosine kinase activity was determined using poly(glutamic acid:tyrosine = 4:1) as an artificial substrate. Cancers of the breast exhibited considerable higher tyrosine kinase activities in both cytosol and membrane fractions, compared to benign breast tumors (P less than or equal to 0.001). Benign tumors demonstrated increased activities in cytosol in comparison to normal breast tissues (P less than 0.001). Furthermore, there appears to be a strong association of an enhanced expression of activity of tyrosine kinase in cytosol of primary carcinomas and early systemic relapse. In combination with aldolase activity a nearly complete discrimination is achieved between malignant specimens on one hand and benign and normal tissues on the other.

Characterization of protein tyrosine kinases from human breast cancer: involvement of the c-src oncogene product.

Tyrosine phosphorylation is an important regulatory mechanism in response to the action of growth factors and oncogenes. Since many oncogenes code for tyrosine kinases, increased or altered oncogene expression may be reflected in increased tyrosine kinase activity. In a recent study (Hennipman et al., Cancer Res., 49: 516-521, 1989), we found that the tyrosine kinase activity of the cytosolic and membrane fractions of malignant human breast tissue was significantly higher compared to the benign or the normal breast tissue. Moreover, the increase in the cytosolic fractions was found to be of prognostic value. In the present study we determined the protein tyrosine kinase (PTK) activity of another 72 breast cancer specimens, and it could be shown again that the PTK activity in all 72 of these tumors was elevated compared to normal controls. We characterized these cytosolic PTKs by anion exchange chromatography using fast protein liquid chromatography, and it could be shown that at least two different forms of PTK exist. Using antibodies against a number of known oncogene products, we could determine that at least 70% of the PTK activity in the cytosol originated from the presence of the c-src oncogene product. Both of the PTK activity peaks seen in the fast protein liquid chromatography patterns could be precipitated with the anti-Src antibody. Furthermore, using the MCF-7 breast cancer cell line, it could be shown that the antibody against c-src also precipitated a part of the cytosolic PTK activity. In normal human peripheral lymphocytes, no precipitation of the cytosolic and membrane PTK activity could be achieved using the anti-Src antibody. Inasmuch as the cytosolic PTK activity parallels the malignancy in breast tumors (Hennipman et al., Cancer Res., 49: 516-521, 1989), and the majority of this activity is precipitated by anti-Src antibodies, the c-src protooncogene may play a key role in the manifestation of breast cancer.

Functional changes associated with the sequential transformation of L'4 into L4 pyruvate kinase.

The functional changes, associated with the sequential transformation of L'4 into L4 pyruvate kinase (ATP:pyruvate 2-O-phosphotransferase, EC 2.7.1.40) were studied. L'4 enzyme from human erythrocytes shows strong hysteretic behaviour: the initial rate of the enzyme preincubated with an unsaturating concentration of phosphoenolpyruvate is much higher than of the enzyme preincubated with ADP, at the same phosphoenolpyruvate concentration, although the "final activity" (the activity of the linear part of the reaction progress curve) was the same in both cases. This phenomenon was observed both in the presence and absence of fructose 1,6-diphosphate. High concentrations of both Mg2+free and MgATP2- diminish the difference in initial rate, between the ADP and phosphoenolpyruvate preincubated enzymes: Mg2+free by stabilizing the phosphoenolpyruvate-induced form; ATPMg2- by stabilizing the ADP-induced form. The magnitude of the difference in initial rates of the ADP-or phosphoenolpyruvate-preincubated enzyme is a function of both substrates. L4 pyruvate kinase (either from human liver or trypsin treated L'4 enzyme) does not, or to a very slight extent, show such behaviour. L'2L2 pyruvate kinase shows behaviour intermediate between L'4 and L4 enzymes. A model is proposed to describe the kinetic behaviour of L'4 and L4 enzymes.

Regulation of human erythrocyte hexokinase. The influence of glycolytic intermediates and inorganic phosphate.

Human erythrocyte hexokinase (ATP: D-hexose 6-phosphotransferase, EC 2.7.1.1) was inhibited competitively with respect to MgATP2- by glucose-6-P (Ki - 10.8 muM) and fructose-6-P (Ki = 160 muM). Low concentrations of inorganic phosphate were competitive with respect to glucose-6-P and fructose-6-P, although higher concentrations of Pi were not able to overcome completely the inhibition by the hexose phosphates. The results are consistent with a model in which hexokinase exists in equilibrium either as free or phosphate-associated enzyme, the latter having a reduced but still substantial affinity for hexose phosphate. An alternative explanation could be found in the presence of two different enzymes, one with a high affinity for glucose-6-P being sensitive to regulation by Pi, one with a lower affinity for glucose-6-P being insensitive to Pi. A similar but less pronounced effect of Pi, was found on the inhibition by 2,3-diphosphoglycerate (Ki = 4.0 mM). Pi in the absence of inhibitor was also a competitive inhibitor with respect to MgATP2- (Ki = 20 mM). Furthermore a competitive inhibition with respect to MgATP2- was found by fructose 1,6-diphosphate (Ki = 4.3 mM), glycerate-3-P (Ki = 3.8 mM), glycerate-2-P (Ki = 12.5 mM), MgADP- (Ki = 1.0 mM) and MgAMP (Ki = 1.7 mM).

Kinetics of human erythrocyte hexokinase. Influence of temperature, ATP4- and magnesium ions.

Human erythrocyte hexokinase (EC 2.7.1.1) is inhibited competitively with respect to Mg-ATP2- by uncomplexed Mg2+ (Ki = 16--18 mM) and ATP4- (Ki = 1.6 mM). No real activation by low concentrations of Mg2+ could be detected and no allosteric behaviour was observed under the conditions tested. The temperature dependence of the enzyme was studied in relationship to the presence of Mg2+ or ATP4-. At equal concentrations of Mg2+ and ATP4- a break in the Arrhenius plot was observed at 27.5 degrees C, the higher temperature form of the enzyme having the lower activation energy. This break point in the Arrhenius plot was shifted to 36 degrees C in the presence of 5 mM Mg2+. A straight-line relationship was observed in the presence of 2.5 mM ATP4-. The Km for Mg-ATP2- showed a linear increase at temperatures over about 36 degrees C independent of the presence of Mg2+ or ATP4-. The nature of these phenomena is discussed.

Purification and some properties of human erythrocyte hexokinase.

1. Human erythrocyte hexokinase (ADP:D-hexose 6-phosphotransferase, EC 2.7.1.1) was purified 50 000--100 000-fold with a final specific activity of about 25--50 units/mg protein using gel-filtration, ion-exchange chromatography and affinity chromagraphy. 2. After isoelectrofocusing ofthe preparation one major protein band could be detected besides a minor band. THe isoelectric point of the major protein band was found to be 4.7. 3. After purification the enzyme could be stabilized in a medium containing inorganic phosphate, glucose, glycerol and mercaptoethanol. 4. The molecular weight was determined by gel-filtration and was found to be 132 000+/-8000. 5. The enzyme shows a broad pH optimum ranging from 7.0 to 8.4. 6. The kinetic behavior of the purified enzyme at 37 degrees C was somewhat different from the normal Michaelis-Menten kinetics due to its instability. The affinity constants were 0.048--0.080 mM for glucose and 0.57--1.0 mM for Mg-ATP. 7. The enzyme was specific for Mg- ATP as the nucleotide substrate. Mg-UTP, Mg-ITP,Mg-GTP and Mg-CTP were not converted to corresponding diphosphates. Several hexoses could be phosphorylated by the enzyme. Mannose could be phosphorylated at the same rate as glucose, although the affinity for the enzyme was lower (5m=0.60mM). Much lower rates and lower affinities were found with 2-deoxy-D-glucose (5m=1.0mM), D(+)-glucosamine (5m=4.5 mM) and fructose (5m=10 mM). N-acetyl-D-glucosamine , galactose andsorbose were not phosphorylated at all.

Hexokinase isozyme distribution and regulatory properties in lymphoid cells.

The glycolytic enzyme hexokinase is studied in cultured leukemic lymphoblasts, in normal lymphocytes and in lymphoblasts obtained by stimulation of normal lymphocytes with phytohaemagglutinin. Hexokinase activity levels in cultured lymphoblasts and in normal lymphocytes are identical, but somewhat higher levels are found in stimulated lymphocytes. Cultured leukemic lymphoblasts differ in isozyme content in comparison to the other lymphoid cells. Besides hexokinase I, which is detected in all the lymphoid cells, they are characterized by the presence of hexokinase II. The concentration of type II increases during cell growth. Another difference between leukemic lymphoblasts and mature and stimulated lymphocytes is found in the regulatory properties of hexokinase I. Hexokinase I from both normal and stimulated lymphocytes is inhibited by glucose-1,6-diphosphate. This inhibition is decreased in part by addition of inorganic phosphate. Hexokinase I from leukemic lymphocytes, however, is inhibited to a lesser extent by glucose-1,6-diphosphate. Inorganic phosphate has no effect at all on this inhibition. In accordance with these findings a different pattern in the hexokinase I region was detected in electrophoresis with several cell types. The subisozyme hexokinase Ib, which appears to be the phosphate-regulated form, is predominant in lymphocytes, whereas it is present in a minor fraction in the cultured leukemic lymphoblasts. In these cells hexokinase Ic predominates.

Phosphorylation of pyruvate kinase type K is restricted to the dimeric form.

In the absence of glycolytic intermediate, fructose-1,6-bisphosphate, pyruvate kinase type K exists in the dimeric form and is readily phosphorylated, whereas in the same sample and the same conditions pyruvate kinase type M is present as a tetramer and is not phosphorylated. Addition of fructose-1,6-bisphosphate results in the association of dimeric K2 molecules to a tetrameric K4 enzyme as determined by gel filtration and cellulose acetate electrophoresis, with concomitant loss of the capacity of the K isozyme to become phosphorylated. Phosphorylated K2 dimers can also tetramerize, but with a low recovery of the radiolabel, suggesting a fructose-1,6-bisphosphate induced dephosphorylation or selective degradation. The dimeric K isozyme is enzymatically active; inactive K-type monomers can be detected by immunoblot analysis in the absence of fructose-1,6-bisphosphate, but no phosphorylated pyruvate kinase is present in this fraction. The formation of K4 tetramers can not be accomplished by the substrate phosphoenolpyruvate. Fructose-1,6-bisphosphate is an allosteric activator of pyruvate kinase type K and induces hyperbolic saturation curves for phosphoenolpyruvate. In contrast, in the absence of effectors, pyruvate kinase type M exhibits Michaelis-Menten kinetics, but sigmoidal curves can be induced by the amino acid phenylalanine. However, even in the presence of phenylalanine, the M-type maintained its tetrameric configuration and did not serve as a substrate in the phosphorylation reaction. These findings argue for the importance of subunit interaction in the regulation of phosphorylation of pyruvate kinase.

Mitochondrial and cytosolic hexokinase from rat brain: one and the same enzyme?

Rat brain mitochondrial hexokinase (ATP: D-hexose 6-phosphotransferase, EC 2.7.1.1) was solubilized by treatment of the mitochondria with glucose 6-phosphate and partly purified. The solubilized enzyme was compared with the cytosolic enzyme fraction. The solubilized and cytosolic enzymes were also compared with the enzyme bound to the mitochondrial membrane. The following observations were made. 1. There is no difference in electrophoretic mobility on cellulose-acetate between the cytosolic and the solubilized enzyme. Both fractions are hexokinase isoenzyme I. 2. There is no difference in kinetic parameters between the cytosolic or solubilized enzymes (P less than 0.001). For the cytosolic enzyme Km for glucose was 0.067 mM (S.E. = 0.024, n = 7); Km for MgATP2- was 0.42 mM (S.E. = 0.13, n = 7) and Ki,app for glucose 1,6-diphosphate was 0.084 mM (S.E. = 0.011, n = 5). For the solubilized enzyme Km for glucose was 0.071 mM (S.E. = 0.021, n = 6); Km for MgATP2- was 0.38 mM (S.E. = 0.11, n = 6) and Ki,app for glucose 1,6-diphosphate was 0.074 mM (S.E. = 0.010, n = 5). However when bound to the mitochondrial membrane, the enzyme has higher affinities for its substrates and a lower affinity for the inhibitor glucose 1,6-diphosphate. For the mitochondrial fraction Km for glucose was 0.045 mM (S.E. = 0.013, n = 7); Km for MgATP2- was 0.13 mM (S.E. = 0.02, n = 7) and Ki,app for glucose 1,6-diphosphate was 0.33 mM (S.E. = 0.03, n = 5). 3. The cytosolic and solubilized enzyme could be (re)-bound to depleted mitochondria to the same extent and with the same affinity. Limited proteolysis fully destroyed the enzyme's ability to bind to depleted mitochondria. 4. Our data support the hypothesis that soluble- and solubilizable enzyme from rat brain are one and the same enzyme, and that there is a simple equilibrium between the enzyme in these two pools.

Intracellular free magnesium and phosphorylated metabolites in hexokinase- and pyruvate kinase-deficient red cells measured using 31P-NMR spectroscopy.

The erythrocyte metabolism of two patients with nonspherocytic hemolytic anemia caused by a hexokinase deficiency, and a pyruvate kinase deficiency, respectively, were studied with NMR. The complexing of ATP and 2,3-diphosphoglycerate (2,3-DPG) with Mg2+ and hemoglobin (Hb) was determined using 31P-NMR on oxygenated and deoxygenated cells to investigate the influences of these enzyme defects on intracellular magnesium distribution and on Hb oxygen dissociation. In the pyruvate kinase-deficient red blood cells, the 2,3-DPG concentration was almost twice the normal value and the ATP concentration was near the lower limit of the normal range. In the hexokinase-deficient red cell population, the predominance of young cells masked the deficiency. Therefore, reticulocyte control cells were included in this study. In the oxygenated pyruvate kinase-deficient cells, the fraction of ATP that is complexed to magnesium as well as the free Mg2+ concentration were normal, despite the abnormal concentration of 2,3-DPG. In the deoxygenated cells the free Mg2+ concentration was lower than in normal cells. The fraction of Hb complexed with 2,3-DPG was higher than normal in both oxygenated and deoxygenated pyruvate kinase-deficient cells, in accordance with the high p50 of the oxygen-hemoglobin dissociation curve. In hexokinase-deficient cells, two major abnormalities are found: when the cells were deoxygenated, the concentration of ATP and 2,3-DPG fell. This was not observed for any other sample and could, therefore, be a consequence of the hexokinase deficiency. Despite almost normal levels of magnesium-binding metabolites, the free Mg2+ concentration in oxygenated and deoxygenated cels is much lower than in normal cells. This could be a cell-age-related phenomenon, since lower free Mg2+ concentrations were also found in reticulocyte control cells.