Molecular Genetics and Metabolism 86 (2005) 401–411

www.elsevier.com/locate/ymgme

Regulation of human erythrocyte metabolism by insulin: Cellular distribution of 6-phosphofructo-1-kinase and its implication for red

blood cell function

Patricia Zancan a,b, Mauro Sola-Penna a,¤

a Laboratório de Enzimologia e Controle do Metabolismo (LabECoM), Departamento de Fármacos, Faculdade de Farmácia, Universidade Federal do Rio de Janeiro, Ilha do Fundão, Rio de Janeiro, RJ 21941-590, Brazil

b Instituto de Bioquímica Médica, Universidade Federal do Rio de Janeiro, Ilha do Fundão, Rio de Janeiro, RJ 21941-590, Brazil

Received 9 May 2005; received in revised form 17 June 2005; accepted 21 June 2005Available online 15 August 2005

Abstract

Human erythrocytes are highly specialized cells whose function is oxygen transport. These cells’ sole metabolic source of energy isthe fermentation of glucose via glycolysis. They contain an active insulin receptor and respond to insulin by increasing phosphoryla-tion of tyrosine residues in several proteins. However, no metabolic eVects have yet been associated with activation of this receptor inhuman erythrocytes. Here, we show that insulin increases the rate of glycolysis in human erythrocytes. Lactate production increased56 and 173% in the presence of 10 and 100 nM insulin, respectively. A higher insulin concentration (1000 nM) partially reversed thestimulation of glycolysis. These eVects occur through activation of the key glycolytic enzyme 6-phosphofructo-1-kinase, which exhib-its the same pattern of modulation by insulin as seen for glycolytic Xux. This modulation also occurs physiologically since ex vivoexperiments revealed 50% stimulation of 6-phosphofructo-1-kinase (PFK) activity following a high carbohydrate meal. Insulinincreases phosphorylation of PFK and redistributes the enzyme in red blood cells, causing it to detach from the erythrocyte mem-brane: upon insulin stimulation, the amount of enzyme associated with the plasma decreases by 86%. Detachment is a commonmechanism of enzyme activation. As a consequence, insulin prevents up to 68% of red cells hemolysis. These results show that insulinregulates erythrocyte glycolysis and viability and suggest that this regulation is associated to other erythrocyte functions such as oxy-gen transport. Finally, we suggest that this regulatory mechanism might be compromised in patients with diabetes mellitus. 2005 Elsevier Inc. All rights reserved.

Keywords: Glycolysis; Diabetes; Erythrocytes; Plasma membrane; Phosphorylation; Metabolism

Introduction

The primary physiological function of the red bloodcell is to transport oxygen. In addition, there are severalmetabolic functions that a red blood cell must performfor its own survival. These include generation of meta-bolic energy (e.g., ATP), generation of reducing agents(e.g., NADH and NADPH), generation of 2,3-diphos-

* Corresponding author. Fax: +55 21 2562 6445.E-mail address: maurosp@ufrj.br (M. Sola-Penna).

1096-7192/$ - see front matter 2005 Elsevier Inc. All rights reserved.doi:10.1016/j.ymgme.2005.06.011

phoglycerate, and maintenance of ionic and concentra-tion gradients across the cell membrane. These functionsinvolve a minimum set of metabolic pathways, includingthe glycolytic sequence, the pentose phosphate pathway,the 2,3-diphosphoglycerate shunt, and the pathways ofnucleotide metabolism; in the mature cell, the ability tosynthesize RNA, protein, lipid, and purines has beenlost. Erythrocytes have a life span in blood of about 120days. They lack organelles such as the nucleus and mito-chondria; energy is generated exclusively by glycolysis[1]. Glucose is taken up in an insulin-independent,

402 P. Zancan, M. Sola-Penna / Molecular Genetics and Metabolism 86 (2005) 401–411

saturable, passive way by erythrocytes via the glucosetransporter GLUT1. Regardless of the high glucosetransport capacity of erythrocytes, these cells consumeglucose at a slow rate compared to other tissues; forexample, glucose consumption by white blood cells isabout 500 times higher than that of erythrocytes [1].

Human erythrocytes contain speciWc insulinreceptors, which have binding characteristics similar tothose found in other classical insulin target cells [2].These receptors have been extensively used to evaluatein vitro sensitivity of human subjects [2]. This sensitivityis altered in diabetes mellitus [3], obesity [4], very highcarbohydrate intake [4], and other metabolic alterations[5–7]. However, the physiological signiWcance of thisreceptor remains unclear.

Classical insulin-responsive cells exhibit an increasedrate of the glycolysis as a response to insulin signaling[8]. This event occurs mainly through the activation of6-phosphofructo-1-kinase (PFK), the key regulatoryenzyme of the glycolytic pathway. In insulin-responsivetissues, the hormone activates PFK through two mainmechanisms: raising the levels of the activator fructose-2,6-bisphosphate [9,10]; and altering the intracellular dis-tribution of the enzyme [11]. Mammalian erythrocyteslack the enzyme 6-phosphofructo-2-kinase (PFK-2)responsible for fructose-2,6-bisphosphate synthesis.However, it has been shown that erythrocyte PFKreversibly binds to the membrane cytoskeleton, which isproposed to modulate its activity [12–14]. This bindingof PFK to the erythrocyte membrane cytoskeletonappears to be regulated by several events, including Ca2+

concentration [15], serotonin signaling [16], competitionwith deoxy-hemoglobin [17], and tyrosine kinase activi-ties [18,19].

The existence of an active insulin receptor in humanerythrocytes is well documented. Moreover, there isevidence suggesting that this receptor modulates someof the erythrocyte functions. However, no eVect ofinsulin on the major physiologic function of thehormone, the regulation of glucose metabolism, hasever been demonstrated in human erythrocytes. Theresults presented in this paper show for the Wrst timethat insulin is responsible for the regulation ofglycolysis in human erythrocytes and that this phe-nomenon is correlated with the functioning of the redblood cells. Glycolysis rate enhancement is due to acti-vation of 6-phosphofructo-1-kinase by insulin. Wedemonstrate that the activation occurs through adecrease in the membrane-bound PFK levels, whichaugments enzyme activity. These phenomena aredependent on the protein tyrosine kinase activity of theinsulin receptor. Finally, it is established that modula-tion of human erythrocyte glycolysis is directly relatedto the ability of the red blood cells to maintain theirelementary functions, while avoiding cellular damageand hemolysis.

Material and methods

Materials

Human insulin (10 mg/ml in 25 mM Hepes (2-(4-(2-hydroxyethyl)-1-piperazinyl)ethanesulfonic acid)buVer, pH 8.2), ATP and fructose-6-phosphate werepurchased from Sigma Chemical (St. Louis, MO,USA). 32Pi was purchased from Instituto de PesquisasEnergéticas e Nucleares (São Paulo, Brazil). [�-32P]ATPwas prepared according to Maia et al. [20]. Genistein(4�,5,7-trihydroxyisoXavone) from Calbiochem Bio-chemical (CA, USA) was the generous gift of Prof.Mario A. Cardoso Silva-Neto. The reagents were of thehighest purity available. PuriWed PFK was obtainedfrom rabbit skeletal muscle according to the methoddeveloped by Kemp [21], with modiWcations intro-duced by Kuo et al. [22].

Samples collection

Blood samples were collected in the morning afteran overnight fast, and were drawn into blood collectiontubes containing 15% ethylenediaminetetraacetic acid(EDTA). After centrifugation, plasma and buVy coatwere removed by aspiration and the remainingerythrocytes were washed twice with saline. Thisprotocol to obtain packed erythrocytes has beenextensively used by many groups [2,12,15,16,23–25].Protein quantiWcations were performed as described byLowry et al. [26] using bovine serum albumin as a stan-dard.

Radioassay for 6-phosphofructo-1-kinase activity

Enzyme activity was measured according to theradiometric method described previously [27]. SomemodiWcations were introduced which improve theaccuracy and sensitivity of the method as follows. Theassays were performed at 37 °C in a reaction mediumcontaining 50 mM Tris–HCl (pH 7.4), 5 mM (NH4)2SO4,5 mM MgCl2, 1 mM [�-32P]ATP (4 �Ci/�mol), 1 mMfructose-6-phosphate and puriWed PFK (10 �g/ml) orerythrocyte PFK. After diVerent time intervals, 0.4 mlaliquots were withdrawn and added to 1 ml of activatedcharcoal suspension (250 g/L) containing 0.1 N HCl and0.5 M mannitol. The material was centrifuged at 1500gfor 10 min, and an aliquot of the supernatant wascounted in a liquid scintillation counter to evaluate theamount of [1-32P]fructose-1,6-bisphosphate formed.Activity rate was calculated by the linear regression ofthe amount of [1-32P]fructose-1,6-bisphosphate formedduring the linear phase of the reaction in function of thereaction time. Duplicates were performed for all samplesand blanks were obtained in the absence of fructose-6-phosphate.

P. Zancan, M. Sola-Penna / Molecular Genetics and Metabolism 86 (2005) 401–411 403

Phosphorylation of 6-phosphofructo-1-kinase—evaluation of phosphorylated enzyme

The phosphorylation assay was assessed by twoprotocols. Protocol 1: fresh human red blood cells (25 �Lat 70–80% hematocrit) were mixed with 25 �l of amedium containing insulin (0 or 100 nM) in phosphatebuVer saline (PBS) (pH 7.2), 5 mM MgCl2, 5 mM glucoseand [32P]PO4 (1.5 mCi/�mol) during 1 h at 37 °C. Afterincubation, samples were analysed by sodium dodecylsulfate-containing polyacrylamide gel electrophoresis(SDS–PAGE) according to Laemmli [28]. Gels werestained with Coomassie blue, dried and examined byautoradiography on a PhosphoImage apparatus (StormTechnologies). Phosphorylation was evaluated on the85 kDa band, which corresponds to the PFK monomersmolecular mass. Band intensity was quantiWed with thesoftware Image Quant 5.2 (Molecular Dynamics).Protocol 2: puriWed rabbit skeletal muscle PFK (1 mg/ml) was incubated at 37 °C with 5 mM MgCl2 and 5�Lhuman red cells lysed with 10 mM potassium phosphate,pH 8.0 in the absence or presence of 6 nM insulin. Thephosphorylation reaction was initiated with 3 mM [�-32P]ATP (100�Ci/�mol). The reaction was terminatedafter 10 min, and phosphoprotein was analyzed asdescribed above.

Activity assay of erythrocyte 6-phosphofructo-1-kinase—an eVect of insulin on the enzyme activity

Human red cells (0.5 ml at 70–80% hematocrit) wereincubated with 0.5 ml of a medium containing diVerentconcentrations of insulin (0–1000 nM) in phosphatebuVer (pH 7.2) during 1 h at 37 °C. Erythrocytes werelysed according to [23] with 3 volumes of ice-cold hypo-tonic lysis medium (10 mM potassium phosphate, pH8.0) and centrifuged at 90,000g for 20 min. The mem-brane pellets were resuspended in 0.3 ml of dilutionbuVer containing 100 mM Tris–HCl (pH 8.2), 120 mMKCl, 2 mM NaF, and 1 mM dithiothreitol (DTT). Eryth-rocyte PFK activity from the whole hemolysate and thefraction bound to the membrane were measured byradioassay method by addition of approximately 30 and70 �g/ml protein, respectively. Reaction was stoppedafter 20, 30, and 40 min by addition of the activatedcharcoal suspension, as described above. PFK activitywas calculated by the linear regression of the productformed versus the reaction time.

Ex vivo analysis of erythrocyte 6-phosphofructo-1-kinase activity

Human blood samples collected after an 8-h fastingperiod (starved group) or 45 min after a hypercaloricmeal (fed group) were centrifuged at 1000g for 10 min toremove plasma and white cells. The packed red blood

cells were lysed with 3 volumes of ice-cold hypotoniclysis medium (10 mM potassium phosphate, pH 8.0). Thehemolysate was ultracentrifuged at 90,000g for 20 minand the pellet was resuspended in 0.3 ml of dilutionbuVer containing 100 mM Tris–HCl (pH 8.2), 120 mMKCl, 2 mM NaF, and 1 mM DTT. Erythrocyte PFKactivity was measured by addition of approximately70�g/ml protein. Reaction was stopped after 10, 20, and30 min. PFK activity was calculated by the linear regres-sion of the product formed versus the reaction time.

Preparation of membranes

Red blood cell ghosts were prepared by a methodsimilar to the one described by Alves et al. [29]. BrieXy,fasting blood was collected from young healthy subjectsin polypropylene tubes containing 15% EDTA. Theblood was centrifuged at 750g for 15 min at 0–4 °C.Plasma and buVy coat were aspirated to remove polynu-clear and mononuclear cells and erythrocytes were resus-pended in ice-cold buVer containing 121 mM NaCl,25.3 mM NaHCO3, and 1.3 mM CaCl2 (pH 7.8) to 50%hematocrit. The supernatant was removed after centrifu-gation at 750g for 15 min. The pellet was washed Wvetimes with the same buVer to remove lymphocytes orgranulocytes. Then, the cells were lysed in ice-cold dis-tilled water for 30 min with gentle shaking. Erythrocytemembranes were obtained by homogenization of thelysate in a Potter–Elvehjem glass homogenizer and werecentrifuged at 30,000g for 30 min. The pellet was washedWve times with buVer and was stored at 4 °C.

Modulation of puriWed 6-phosphofructo-1-kinase activity by insulin in the presence of puriWed erythrocyte membranes—partially puriWed system

Partially puriWed rabbit skeletal muscle PFK(0.52 mg/ml) and erythrocyte membrane preparation(0.1 mg/ml) were incubated in 25 �l of a medium contain-ing diVerent insulin concentrations (0–1000 nM) in50 mM Tris–HCl (pH 8.2), 5 mM (NH4)2SO4, 5 mMMgCl2, and 1 mM ATP in the presence or absence ofgenistein (26�M) during 1 h at 37 °C. After incubation,4 �l mixture was used to determine PFK activity throughthe radioassay method described above.

Enzyme–membrane interactions

The interaction between 6-phosphofructo-1-kinaseand erythrocyte membranes was assessed by measuringthe enzyme activity associated with the membrane afterseparation of this material from soluble matter. PartiallypuriWed rabbit skeletal muscle PFK (2.7 mg/ml) anderythrocyte membrane preparation (0.1 mg/ml) wereincubated in 100 �l of a medium containing diVerentinsulin concentrations (0–1000 nM) in 50 mM Tris–HCl

404 P. Zancan, M. Sola-Penna / Molecular Genetics and Metabolism 86 (2005) 401–411

(pH 8.2), 5 mM (NH4)2SO4, 5 mM MgCl2, and 1 mMATP in the presence or absence of genistein (26 �M) dur-ing 1 h at 37 °C. After incubation, mixtures were centri-fuged at 20 p.s.i for 20 min on a Beckman Airfuge(»90,000g, Beckman Instruments, CA, USA) for separa-tion of erythrocyte membranes. The membrane pelletswere suspended in 10 �l of 10 mM potassium phosphate(pH 8.0). The amount of PFK bound to the erythrocytemembrane was evaluated by measuring the PFK activityin the pellet using the radioassay described above.

Lactate production and hemolysis measurements

Fresh collected erythrocytes were centrifuged andplasma and white blood cells were removed byaspiration. The packed red blood cells were washed threetimes with ice-cold phosphate buVer saline. The entireprocedure was conducted at 0–4 °C. Washederythrocytes (8–10% hematocrit) were incubated in thepresence of 50 ml of an isotonic medium containing100 mM Tris–HCl (pH 7.4), 10 mM glucose, 150 mMNaCl, and 0–1000 nM insulin. Incubation was per-formed in a culture shaker, at 37 °C, at 150 rpm. Aliquotswere withdrawn after diVerent time periods and centri-fuged at 6500g. Supernatant was used to measure thelactate content and hemolysis. Lactate was measured byadding 20�l of the supernatant to 1 ml of a reactionmedium containing 100 mM Tris–HCl (pH 8.2), 2 mMNAD+ and 2 U/ml lactate dehydrogenase. NADH for-mation was followed on a spectrophotometer at 340 nm.Hemolysis was measured by the absorbance of thesupernatant at 415 nm, where hemoglobin absorbs.

Statistical analysis

All results in the Wgures are expressed asmeans § standard errors of the mean (SEM). Analysis ofthe data was done using the Sigma Plot software pack-ages (Jandel ScientiWc). Values for each group were com-pared by either a paired and a non-paired student’s t testor two-way analysis of variance (ANOVA). P values lessthan 0.05 was considered to indicate signiWcance in allcases.

Results

Modulation of human erythrocyte 6-phosphofructo-1-kinase activity by insulin

To determine whether human erythrocyte PFK activitycould be modulated by insulin, intact erythrocytesobtained from volunteers after an overnight fast wereincubated in the presence of increasing concentrations ofhuman insulin (0–1000 nM) for 1h at 37 °C. These concen-trations of insulin were employed to cover a large spec-

trum of physiologic insulinemia, varying from basal levels(up to 10 nM), including the hyperglycemic-induced con-centrations (10–500 nM), and pathologic insulin levels(around 1000 nM). Incubation period was followed by dis-ruption of cells by the addition of 3 volumes of the hypo-tonic solution described under Materials and methods andPFK activity was measured in the lysate. Insulin inducedactivation, albeit small (»10%), of PFK catalytic rate in adose-dependent manner up to 100 nM insulin (Fig. 1,black bars, P < 0.01, two-tailed ANOVA). PFK activityraised from 11.4§0.9 nmol fruc-1,6-P2 mg¡1 min¡1 in theabsence of insulin to 12.0§1.1 nmol fruc-1,6-P2 mg¡1 min¡1 and 12.4§1.2 nmol fruc-1,6-P2 mg¡1 min¡1

in the presence of 10 and 100 nM insulin, respectively(nD4, P < 0.05, two-tailed ANOVA, Student–Newman–Keuls post-test). However, this stimulation was notobserved when incubation was done in the presence of1000 nM insulin. When the same experiments were per-formed in the presence of the tyrosine kinase inhibitor,genistein, insulin promoted no signiWcant eVects on PFKactivity (Fig. 1, blank bars, P >0.05, two-tailed ANOVA).These results show that insulin modulates human erythro-cyte PFK activity and suggest that this modulation occursthrough insulin receptor tyrosine kinase activity.

To conWrm whether this modulation of human eryth-rocyte PFK occurs within the physiological variations ofthe insulinemia, we compared the PFK activity fromfreshly collected blood from healthy volunteers after anovernight fast and 45 min after a high-carbohydrate

Fig. 1. EVects of insulin on 6-phosphofructo-1-kinase from humanerythrocytes. Freshly collected erythrocytes were incubated for 1 h at37 °C in the presence of the insulin concentrations indicated on theabscissa. Cells were lysed and PFK activity was measured in a mediumcontaining 50 mM Tris–HCl (pH 8.2), 5 mM MgCl2, 5 mM (NH4)2SO4,1 mM [�-32P]ATP (4 �Ci/�mol), and 1 mM fructose-6-phosphate.Blanks for all conditions were performed in the absence of fructose-6-phosphate and were subtracted from total activity. These experimentswere performed in the absence (black bars) and in the presence of26 �M genistein (white bars), a tyrosine kinase inhibitor. Values aremeans § standard errors of four independent experiments performedin duplicate. *P < 0.05 compared to control without insulin (two-tailedANOVA, Student–Newman–Keuls post-test, n D 4).

P. Zancan, M. Sola-Penna / Molecular Genetics and Metabolism 86 (2005) 401–411 405

food intake. The time interval after meal was around thehyperglycemic peak, where the glycemia increased atleast two fold (data not shown). Fig. 2 summarizes theseresults where it can be seen that blood collected after thehigh-carbohydrate meal presented a PFK activity 56%higher than the PFK activity of the blood collected afterthe starvation period (P < 0.05, Student’s t test). Theseresults corroborate that, physiologically, when insuline-mia rises, the hormone augments erythrocyte PFK activ-ity, which indicates that insulin may modulate theglycolytic Xux in these cells.

Phosphorylation of human erythrocyte 6-phosphofructo-1-kinase

Since the insulin eVects of human erythrocyte PFKactivity depend upon tyrosine kinase activity, we testedthe ability of the hormone to modulate the direct phos-phorylation of PFK. As described in Material and meth-ods, two types of experiments were performed to test thishypothesis. First (protocol 1), fresh collected erythro-cytes from overnight-starved volunteers were incubatedfor 1 h at 37 °C in the presence of the desired concentra-tions of insulin and [32P]PO4. After this period, an ali-quot was withdrawn and mixed with denaturing samplebuVer for SDS–PAGE. The electrophoresis was run andincorporation of radioactive phosphate into the 85 kDaband was measured by autoradiography. Appropriatecontrols were performed to conWrm that the 85 kDaband corresponds to PFK (data not shown). Results are

Fig. 2. 6-Phosphofructo-1-kinase activity from human erythrocytesfrom starved and fed subjects. Erythrocytes were collected from volun-teers after an overnight fast (starved) and after a high carbohydratemeal (fed). Cells were lysed and PFK activity was measured in amedium containing 50 mM Tris–HCl (pH 8.2), 5 mM MgCl2, 5 mM(NH4)2SO4, 1 mM [�-32P]ATP (4 �Ci/�mol), and 1 mM fructose-6-phosphate. Blanks for all conditions were performed in the absence offructose-6-phosphate and were subtracted from total activity. Valuesare means § standard errors of four independent experiments per-formed in duplicate. *P < 0.05 compared with the starved condition(Student’s t test, n D 4).

presented in Table 1 (protocol 1 column). In the presenceof 100 nM insulin we observed an increase of 15.6 § 2.3%(P < 0.05, Student’s t test, n D 3) of phosphate incorpora-tion into PFK band. However, whole erythrocyte lysatespresent a small amount of PFK, compared to total pro-tein content, which could explain this low phosphateincorporation. For this reason, we performed the secondprotocol where erythrocyte lysates were added to puri-Wed PFK, [� 32P]ATP and 6 nM insulin, and the mixturewas incubated for 10 min at 37 °C. Then the sampleswere mixed with denaturing sample buVer for SDS–PAGE and electrophoresis was performed as above.Incorporation of radioactive phosphate was analyzedand results are presented in Table 1 (protocol 2 column).In this case, we observed an increase of 32.8 § 4.0%(P < 0.05, Student’s t test, n D 2) in phosphate incorpo-ration in the presence of insulin, when compared to con-trol in the absence of the hormone. These results conWrmthat isulin increases phosphorylation of human erythro-cytes PFK.

Dependence on intracellular components controlling insulin modulation of 6-phosphofructo-1-kinase

Since we demonstrated that insulin induces thephosphorylation of PFK, we next set out to determinewhether other intracellular components are necessaryto achieve the stimulation of the enzyme activityinduced by insulin in intact human erythrocytes. Inthis context, we tested the eVects of insulin on PFKactivity using a partially puriWed system containingpuriWed 6-phosphofructo-1-kinase and membranefragments isolated from human erythrocytes. This sys-tem contains the whole membrane components andlacks all intracellular components other than PFK,which was exogenously added. Insulin promoted anactivation of the enzyme when this partially puriWed

Table 1Phosphate incorporation into 6-phosphofructo-1-kinase

Incorporation of [32P]PO4 was analyzed as described in Material andmethods according to two distinct protocols. In protocol 1, fresh col-lected erythrocytes were incubated in the presence of a medium con-taining [32P]PO4 in the absence and in the presence of insulin (n D 3). Inprotocol 2, lysed red blood cells with puriWed PFK added were incu-bated in the presence of [�-32P]ATP in the absence and in the presenceof insulin (n D 2). Phosphate incorporation was analyzed by autoradi-ography of the 85-kDa band revealed after SDS–PAGE of the sam-ples. Incorporation was quantiWed using a PhosphoImage apparatus(Storm Technologies) and the above values represent themeans § standard errors of relative incorporation analyzed for eachexperiment.

¤ P < 0.05 compared to control (Student’s t test).

Condition Phosphate incorporation (protocol 1) (%)

Phosphate incorporation (protocol 2) (%)

Control 99.8 § 2.6 99.6 § 2.9Insulin 115.4 § 4.9¤ 132.4 § 6.9¤

406 P. Zancan, M. Sola-Penna / Molecular Genetics and Metabolism 86 (2005) 401–411

system was incubated in the presence ofconcentrations up to 100 nM of the hormone (Fig. 3,black bars, *P < 0.01 compared to control withoutinsulin, two-tailed ANOVA, Student–Newman–Keuls

Fig. 3. Insulin stimulates puriWed 6-phosphofructo-1-kinase in thepresence of erythrocyte membranes. PuriWed PFK and erythrocytemembranes were incubated for 1 h at 37 °C in the presence of the insu-lin concentrations indicated on the abscissa. After incubation, PFKactivity was measured in a medium containing 50 mM Tris–HCl (pH8.2), 5 mM MgCl2, 5 mM (NH4)2SO4, 1 mM [�-32P]ATP (4 �Ci/�mol),and 1 mM fructose-6-phosphate. Blanks for all conditions were per-formed in the absence of fructose-6-phosphate and were subtractedfrom total activity. These experiments were performed in the absence(black bars) and in the presence of 26 �M genistein (white bars), atyrosine kinase inhibitor. Values are means § standard errors of fourindependent experiments performed in duplicate. *P < 0.01 comparedto control without insulin, **P < 0.05 compared to 100 nM insulin(two-tailed ANOVA, Student–Newman–Keuls post-test, n D 4).

post-test, n D 4). On the other hand, the activation ofPFK by insulin was partially reversed in the presenceof 1000 nM hormone (P < 0.05 compared to 100 nMinsulin, two-tailed ANOVA, Student– Newman–Keulspost-test, n D 4). In addition, these insulin eVects werecompletely abolished when the assays were performedin the presence of the tyrosine kinase inhibitor,genistein (Fig. 3, white bars P < 0.05 compared to100 nM insulin, two-tailed ANOVA, Student–New-man–Keuls post-test, n D 4). Actually, the eVects ofinsulin on this partially puriWed system presented anidentical pattern to those obtained with intacterythrocytes, discharging the requirement for anyintracellular component other than tyrosine kinase totransduce the observed eVects of insulin on activationof erythrocyte PFK.

Mechanism of activation of human erythrocyte 6-phosphofructo-1-kinase by insulin

Erythrocyte PFK is described to be modulated by areversible binding to cell membrane components. Totest whether insulin increases the enzyme activitythrough modulation of its binding to erythrocyte mem-branes, we measured the activity of PFK that remainsassociated to the erythrocyte membrane upon incuba-tion of the partially puriWed system in the presence ofinsulin. After incubation for 1 h at 37 °C, the mixtureswere centrifuged at 90,000g for 20 min and the pelletcontaining erythrocyte membranes were resuspendedin dilution buVer as described in Materials and

Fig. 4. Detachment of 6-phosphofructo-1-kinase from erythrocyte membranes. PuriWed PFK and erythrocyte membranes were incubated for 1 h at37 °C in the presence of the insulin concentrations indicated on the abscissa. After incubation, mixture was centrifuged for 20 min at 96,000g andPFK activity was measured in the pellet in a medium containing 50 mM Tris–HCl (pH 8.2), 5 mM MgCl2, 5 mM (NH4)2SO4, 1 mM [�-32P]ATP(4 �Ci/mol), and 1 mM fructose-6-phosphate. Blanks for all conditions were performed in the absence of fructose-6-phosphate and were subtractedfrom total activity. These experiments were performed in the absence (black bars) and in the presence of 26 �M genistein (white bars), a tyrosinekinase inhibitor. *P < 0.01 compared to control without insulin (two-tailed ANOVA, Student–Newman–Keuls post-test, n D 4). Inset, incubation wasperformed in combination of the absence and in the presence of 100 nM insulin and 26 �M genistein, a tyrosine kinase inhibitor. All values aremeans § standard errors of four independent experiments performed in duplicate.

P. Zancan, M. Sola-Penna / Molecular Genetics and Metabolism 86 (2005) 401–411 407

methods. The amount of PFK associated to the mem-branes was evaluated by the enzyme activity remainingin this pellet. Insulin induced the detachment of PFKfrom erythrocyte membranes in function of the hor-mone concentration (Fig. 4). The enzyme activityshifted from 21.2 § 2.6 nmol fruc-1,6-P2 mg¡1 min¡1 inthe absence of the hormone to 8.8 § 1.4 (42% of con-trol), 3.8 § 2.0 (18% of control) and 7.6 § 0.6 nmol fruc-1,6-P2 mg¡1 min¡1 (36% of control) in the presence of10, 100, and 1000 nM insulin, respectively (P < 0.01,two-tailed ANOVA, n D 4). This detachment is notobserved when the assay was performed in the presenceof genistein (Fig. 4, inset).

To conWrm whether the detachment of PFK fromerythrocyte membranes occurs in a whole-cell system, weincubated intact erythrocytes freshly collected fromovernight-starved volunteers, in the absence and in the

Table 2Detachment of 6-phosphofructo-1-kinase from erythrocyte mem-branes

Freshly collected erythrocytes were incubated in the absence and inthe presence of insulin. After incubation, cells were lysed and mem-branes were separated by centrifugation. PFK activity associated withthe membranes was measured as described in Material and methods.Values represent means § standard errors of four independent experi-ments performed in duplicate.

¤ P < 0.05 compared to control (Student’s t test).

Condition PFK activity nmol Fruc 1,6-P2 mg¡1 min¡1

Control 3.21 § 0.3Insulin (10 nM) 1.86 § 0.25¤

presence of 10 nM insulin for 1 h at 37 °C. After thisincubation, erythrocytes were lysed by the addition of ahypotonic solution as described for the experiments inFig. 1. Then, lysate was centrifuged at 90,000g for 20 minand the PFK that still remains associated to the mem-branes was evaluated as described above. The resultspresented in Table 2 conWrm that insulin induced adetachment of PFK from the erythrocyte membrane. Itis interesting to note that insulin was able to promote a42% decrease in the PFK activity in the pellet (P < 0.05,Student’s t test, n D 4). Such a large modulation reXectsthe capacity of the hormone to change the localizationof the endogenous enzyme in erythrocytes, since noexogenous enzyme was added to the reaction medium inthis experiment.

Activation of glycolytic Xux of human erythrocytes by insulin

To obtain more experimental support for insulinmodulation of PFK activity in red blood cells and itsphysiological relevance, the Wnal product of erythrocyteglucose consumption was quantiWed. Glycolytic Xux wasevaluated measuring lactate production as a function ofthe time of incubation in various insulin concentrations(0–1000 nM). Lactate content increased linearly duringthe Wrst 4 h of incubation in all conditions (Fig. 5). Therate of lactate formation was also calculated from thelinear regression of lactate content against time mea-sured in this experiment. Insulin increased the rate of

Fig. 5. Lactate production by human erythrocytes. Fresh collected erythrocytes were incubated in the presence of 100 mM Tris–HCl (pH 7.4), 10 mMglucose, 150 mM NaCl and the concentrations of insulin indicated. Incubation was performed in a culture shaker, at 37 °C and 150 rpm. Aliquotswere withdrawn after the time periods indicated on the abscissa and centrifuged at 6500g and supernatant was used to measure the lactate content.Lactate was measured by adding 20 �l of the supernatant to 1 ml of a reaction medium containing 100 mM Tris–HCl (pH 8.2), 2 mM NAD+ and 2 U/ml lactate dehydrogenase. NADH formation was followed at 340 nm on a spectrophotometer. Results are expressed as the lactate amount formed atthe diVerent time periods (L) divided by the lactate amount measured in aliquots withdrawn prior to incubation (L0). Inset, rate of lactate formationwas calculated by the linear regression of the values presented in main panel as a function of incubation time. All values are means § standard errorsof three independent experiments performed in duplicate. *P < 0.01 compared to the absence of insulin and **P < 0.01 compared to the absence ofinsulin and the presence of 10 or 1000 nM insulin (two-tailed ANOVA, Student–Newman–Keuls post-test, n D 3).

408 P. Zancan, M. Sola-Penna / Molecular Genetics and Metabolism 86 (2005) 401–411

erythrocyte glycolysis (Fig. 5, inset). Incubation in thepresence of 10 nM insulin promoted 54% activation ofthe rate of lactate production by erythrocytes (P < 0.01,ANOVA, Student–Neman–Keuls post-test, n D 3). Thisstimulation was potentiated by incubation with 100 nMinsulin, which promoted 173% activation compared tocontrol and 76% activation compared to the results inthe presence of 10 nM insulin (P < 0.01, ANOVA, Stu-dent–Newman–Keuls post-test, n D 3). In contrast, theeVect was partially blocked by the presence of 1000 nMinsulin, which promoted a 45% stimulation of the rate ofglycolysis. The result obtained with 1000 nM insulin wasnot diVerent from those observed in the presence of10 nM insulin but is higher than control and lower thanthe stimulation obtained in the presence of 100 nMinsulin (P < 0.01, ANOVA, Student–Newman–Keulspost-test, n D 3).

EVects of insulin on hemolysis

To evaluate whether insulin has an eVect on hemoly-sis during incubation for long periods, washed erythro-cytes were incubated during 4 h in the presence ofdiVerent insulin concentrations (0–1000 nM). Aliquotswere withdrawn periodically to evaluate the amount ofhemoglobin released to the medium. Hemolysis occurredlinearly during this time interval for all conditions tested(Fig. 6). The rate of hemolysis was calculated by a linearregression of the measured hemolysis as a function ofincubation time (Fig. 6, inset). The eVects of insulin onhemolysis were dependent on insulin concentration. The

presence of 10 nM insulin was able to reduce hemolysisto 57% of the value observed in control (P < 0.01,ANOVA, Student–Newman–Keuls post test, n D 3). Pre-vention of hemolysis was greater on incubation with100 nM insulin (68% prevention, P < 0.01, ANOVA, Stu-dent–Newman–Keuls post test, n D 3). This protectionwas completely blocked in the presence of 1000 nM insu-lin. In this condition the rate of hemolysis was not statis-tically diVerent from control (P > 0.05, ANOVA,Student–Newman–Keuls post test, n D 3).

Discussion

Insulin is the most important hormone in the regula-tion of blood glucose concentrations and is essential inthe postprandial state. As blood sugar concentrationsrise, insulin is secreted into the blood stream by the betacells of the endocrine pancreas [30]. Insulin initiates itsaction by binding to a glycoprotein receptor on the cellsurface. The insulin receptor is a transmembrane glyco-protein possessing insulin-stimulable, tyrosine-speciWcprotein kinase activity [31]. The main physiologic func-tion of insulin is to reduce the glycemia, increasing glu-cose uptake in insulin-dependent tissues such as skeletalmuscle and adipose tissue [30], and increasing the rate ofglucose breakdown in several tissues, including thosecited above and others like liver, brain and heart muscle.

Erythrocytes maintain the glycolytic Xux at a ratewell below the maximal possible rate [1]. The regulationof glycolysis in these cells as well as the physiological sig-

Fig. 6. EVects of insulin on human erythrocyte hemolysis. Fresh collected erythrocytes were incubated in the presence of 100 mM Tris–HCl (pH 7.4),10 mM glucose, 150 mM NaCl and the concentrations of insulin indicated. Incubation was performed in a culture shaker, at 37 °C and 150 rpm. Ali-quots were withdrawn after the time periods indicated on the abscissa and centrifuged at 6500g and supernatant was used to measure hemolysis bythe absorbance of the supernatant at 415 nm, where hemoglobin absorbs. Values are presented as the absorbance at the diVerent time periods (H)divided by the absorbance of aliquots withdrawn prior to incubation (H0). Inset, rate of hemolysis was calculated by the linear regression of the val-ues presented in main panel as a function of the incubation time. All values are means § standard errors of three independent experiments performedin duplicate. *P < 0.01 compared to the absence of insulin and **P < 0.01 compared to the absence of insulin and the presence of 10 or 1000 nM insu-lin (two-tailed ANOVA, Student–Newman–Keuls post-test, n D 3).

P. Zancan, M. Sola-Penna / Molecular Genetics and Metabolism 86 (2005) 401–411 409

niWcance of this regulation is still not understood. How-ever, it is assumed that the rate of glycolysis aVects theprimary erythrocyte physiological function: the abilityof hemoglobin to bind and transport oxygen [24].

Several studies have pointed out that insulin may reg-ulate the glycolytic Xux in erythrocytes [3,32–34]. Theevidence includes the fact that insulin signaling increasesthe tyrosine phosphorylation of the cytoplasmic domainof band 3, which prevents the binding of glycolyticenzymes, PFK and others [18]. Erythrocyte band 3 is thepredominant protein of the human red blood cell mem-brane and it binds several peripheral proteins, includingPFK and hemoglobin [18]. Binding of PFK to band 3 isaccomplished to enzyme inhibition [35] and this fact sup-ports the hypothesis of regulation of erythrocyte glycol-ysis by insulin. However, it has never been directlydemonstrated that insulin modulates the glycolytic Xuxor any of glycolysis regulating enzymes in mammalianerythrocytes.

In this study, we have reported for the Wrst time thatinsulin increases the rate of glucose consumption inhuman erythrocytes through the activation of the keyglycolytic enzyme, 6-phosphofructo-1-kinase (Fig. 1).The eVect on PFK activity promoted by insulin in theseexperiments is very small (»10%). However, this activa-tion is statistically signiWcant (P < 0.01) as revealed by theanalysis of multiple variance (ANOVA, with Student–Newman–Keuls post-test). This result must be taken intoaccount since it is strongly corroborated by the data ofFig. 3. These latter data obtained with a partially puriWedsystem containing only the erythrocyte membranes andpuriWed PFK revealed that insulin promoted the samepattern of stimulation observed in Fig. 1. However, thehormone was able to augment the PFK activity up to93%, revealing that the results obtained in Fig. 1 wereunderestimated. Using the partially puriWed system weoptimized the measurements of insulin eVects on theerythrocyte PFK activity by exogenous addition of theenzyme. In addition, this magnitude of eVects is sup-ported by the data presented in Fig. 4, where insulin pro-moted a marked reduction (82%) in the amount of PFKbound to red-cell membranes. Since the detachment ofPFK from red-cell membranes is described as an impor-tant mechanism of the enzyme activation, 82% of reduc-tion in the membrane-bound enzyme is not consistentwith only 10% of enzyme activation observed in Fig. 1,but is very consistent with the 93% activation shown inFig. 4. We attribute this diVerence to the experimentallimitations used in enzyme activity determination inFig. 1. In this Wgure, red blood cells must be disruptedprior to measure the enzyme activity by hypotonic shock,which dilutes several times the cytosolic content of redblood cells. This dilution might detach the normallybound enzyme, which would increase the activity of ourcontrols, burying the eVects of insulin. Actually, this is themain diYculty to observe any eVect of insulin on erythro-

cytes glucose metabolism, as has been pointed out by oth-ers [3,19,24]. Here, our goal to determine these eVects isdue to the higher sensitivity of the radiometric methodfor measurements of PFK activity that we have previ-ously described [27]. This method allowed us to deter-mine PFK activity in subcellular fractionation,containing undersized amounts of PFK, derived fromdiVerent biological materials [27,36–38]. On the otherhand, an improvement of the method is described, whichallowed us to maximize the sensibility of the radiometricmethod and observe the insulin-induced stimulation ofPFK in intact human erythrocytes.

Other evidence is presented to conWrm the augmentedactivity of erythrocyte PFK upon insulin stimulation ofthese cells. The ex vivo experiments shown on Fig. 2revealed that the PFK activity from erythrocytes col-lected during hyperglycemia (which coincides with theinsulin peak in healthy subjects) is 56% higher than thosecollected after an overnight fast (P < 0.05, Student’s ttest). Moreover, in this study we clearly demonstratedthat insulin increased the rate of lactate formation byhuman erythrocytes (Fig. 5, inset). Lactate is the Wnalproduct of glucose fermentation by erythrocytes and itsproduction rate is indicative of the glycolytic Xux. Takentogether, our results indicate a physiological activationpromoted by insulin on erythrocyte glucose metabolismthrough modulation of PFK activity and localization.

The mechanism of PFK activation by insulin involvesthe displacement of the enzyme from the membrane, asshown in Fig. 4 and in Table 2. The binding of erythrocytePFK to the membrane, more speciWcally to the predomi-nant membrane protein band 3, has been associated withenzyme inhibition [18]. Here we proposed that the dis-placement of PFK from erythrocyte membrane may beregulated by tyrosine-phosphorylation of the enzyme.This phosphorylation is increased by insulin-stimulationof human red blood cells, as shown in Table 1. Besides, theexperiments presented in Figs. 3 and 4 show that insulinmodulation of PFK activity and localization does notdepend on other cytosolic components, but does dependexclusively on membrane-enzyme interaction. Indeed,Clari et al. [34], studying the tyrosine-phosphorylation ofcytosolic proteins in human erythrocytes, have shown aphosphorylation of an 85 kDa protein. These authors didnot focus on this protein, since the major site of phosphor-ylation was a 19kDa protein, their study object. However,the 85kDa protein appears to be a good phosphorylationsite and coincides with the molecular mass of humanPFK, which conWrms our results. Additionally, we havepreviously demonstrated that phosphorylation of PFK attyrosine or serine residues aVects enzyme interaction withother structural cellular components in skeletal muscle[11,36]. These results corroborate the hypothesis that theincreased phosphorylation measured here may be respon-sible for induction of the detachment from the membraneand activity of red blood cells PFK.

410 P. Zancan, M. Sola-Penna / Molecular Genetics and Metabolism 86 (2005) 401–411

Altogether, our results indicate that erythrocyte gly-colysis is under regulation by the most important hor-mone of glucose homeostasis control. Insulin, having anactive participation, regulates the erythrocyte glycolysis,and the cell functions. This hypothesis is strongly cor-roborated by the data presented in Fig. 6, where insulinwas able to prevent hemolysis. The pattern of hemolysisprevention is directly correlated to the activation oferythrocyte glycolysis (Fig. 5), activation of PFK (Figs. 1and 3) and the detachment of the enzyme from erythro-cyte membranes (Fig. 5 and Table 2).

A decreased erythrocyte insulin binding has been dem-onstrated in patients with hyperinsulinemia, a conditionwhich occurs in diabetes mellitus type II and is main-tained even after a 12-h fast [7]. Actually, erythrocyteinsulin receptor is down-regulated upon incubation withinsulin, and this down-regulation is dependent on thetime of incubation and on the concentration of insulin[39–41]. This fact explains the data presented here in thepresence of 1000 nM insulin, where the eVects of the hor-mone were partially or completely blocked. In fact, thelack of eVects at the higher insulin concentration mightbe due to the down-regulation of insulin receptor pro-moted by the hormone and allow us to correlate the dataobtained in the presence of 1000 nM insulin with theinsulin resistance observed in diabetes mellitus type II.

Acknowledgments

Grateful thanks are due to Dr. Martha M. Sorensonfor for critical reading of the manuscript. This work wassupported by grants from Fundação Carlos ChagasFilho de Amparo a Pesquisa do Estado do Rio deJaneiro (FAPERJ), Fundação Ary Frausino/FundaçãoEducacional Chales Darwin (FAF/FECD Programa deOncobiologia), Conselho Nacional de DesenvolvimentoCientíWco e Tecnológico (CNPq) and Programa deNúcleos de Excelência (PRONEX).

References

[1] J.C. Grundy, Aspects of blood cell biochemistry: erythrocytes,platelets, and stem cells, in: K.B. Storey (Ed.), Functional Metabo-lism: Regulation and Adaptation, John Wiley, New Jersey, 2004,pp. 505–528.

[2] S. Suzuki, T. Toyota, Y. Goto, Characterization of the insulinreceptor kinase from human erythrocytes, Endocrinology 121(1987) 972–979.

[3] F. Marques, M.E. Crespo, Z.I. Silva, M. Bicho, Insulin and highglucose modulation of phosphatase and reductase enzymes in thehuman erythrocytes: a comparative analysis in normal and dia-betic states, Diabetes Res. Clin. Pract. 47 (2000) 191–198.

[4] S.W. Rizkalla, Y. Le Bouc, P. Serog, M. Apfelbaum, Carbohydrateintake aVects insulin binding to human erythrocytes in normalweight subjects but not in subjects with family obesity, Metabo-lism 30 (1981) 900–907.

[5] R.F. Santos, R. Nomizo, E. Oliveira, M. Ursich, B. Wajchenberg,G.M. Reaven, S. Azhar, Erythrocyte insulin receptor tyrosinekinase activity is increased in glyburide-treated patients with type2 diabetes in good glycaemic control, Diabetes Obes. Metab. 2(2000) 237–241.

[6] M. Sauvage, P. Maziere, H. Fathallah, F. Giraud, Insulin stimu-lates NHE1 activity by sequential activation of phosphatidylinosi-tol 3-kinase and protein kinase C zeta in human erythrocytes, Eur.J. Biochem. 267 (2000) 955–962.

[7] D.B. Corry, F.S. Joolhar, M.T. Hori, M.L. Tuck, Decreased eryth-rocyte insulin binding in hypertensive subjects with hyperinsuline-mia, Am. J. Hypertens. 15 (2002) 296–301.

[8] L. Riera, A. Manzano, A. Navarro-Sabaté, J.C. Perales, R. Bar-trons, Insulin induces PFKFB3 gene expression in HT29 humancolon adenocarcinoma cells, Biochim. Biophys. Acta 1589 (2002)89–92.

[9] L. Bosca, G.G. Rousseau, L. Hue, Phorbol 12-myristate 13-acetateand insulin increase the concentration of fructose 2,6-bisphos-phate and stimulate glycolysis in chicken embryo Wbroblasts, Proc.Natl. Acad. Sci. USA 82 (1985) 6440–6444.

[10] C. Denis-Pouxviel, T. Gauthier, D. Daviaud, J.C. Murat, Phospho-fructokinase 2 and glycolysis in HT29 human colon adenocarci-noma cell line. Regulation by insulin and phorbol esters, Biochem.J. 268 (1990) 465–470.

[11] A.P.P. Silva, G.G. Alves, A.H.B. Araújo, M. Sola-Penna, EVects ofinsulin and actin on phosphofructokinase activity and cellular dis-tribution in skeletal muscle, An. Acad. Bras. Ciências 76 (2004)541–548.

[12] N.S. Karadsheh, K. Uyeda, Changes in allosteric properties ofphosphofructokinase bound to erythrocyte membranes, J. Biol.Chem. 252 (1977) 7418–7420.

[13] T. Higashi, C.S. Richards, K. Uyeda, The interaction of phospho-fructokinase with erythrocyte membranes, J. Biol. Chem. 254(1979) 9542–9550.

[14] J.D. Jenkins, F.J. Kezdy, T.L. Steck, Mode of interaction of phos-phofructokinase with the erythrocyte membrane, J. Biol. Chem.260 (1985) 10426–10433.

[15] M. Assouline-Cohen, R. Beitner, EVects of Ca2+ on erythrocytemembrane skeleton-bound phosphofructokinase, ATP levels, andhemolysis, Mol. Genet. Metab. 66 (1999) 56–61.

[16] M. Assouline-Cohen, H. Ben-Porat, R. Beitner, Activation ofmembrane skeleton-bound phosphofructokinase in erythrocytesinduced by serotonin, Mol. Genet. Metab. 63 (1998) 235–238.

[17] I. Messana, M. Orlando, L. Cassiano, L. Pennacchietti, C. Zuppi,M. Castagnola, B. Giardina, Human erythrocyte metabolism ismodulated by the O2-linked transition of hemoglobin, FEBS Lett.390 (1996) 25–28.

[18] P.S. Low, D.P. Allen, T.F. Zioncheck, P. Chari, B.M. Willardson,R.L. Geahlen, M.L. Harrison, Tyrosine phosphorylation of band 3inhibits peripheral protein binding, J. Biol. Chem. 262 (1987)4592–4596.

[19] M.L. Harrison, P. Rathinavelu, P. Arese, R.L. Geahlen, P.S. Low,Role of band 3 tyrosine phosphorylation in the regulation oferythrocyte glycolysis, J. Biol. Chem. 266 (1991) 4106–4111.

[20] J.C.C. Maia, S.L. Gomes, M.H. Juliani, Preparation of [�-32P] and[�-32P]-ucleoside triphosphate, with high speciWc activity, in: C.M.Morel (Ed.), Genes and Antigenes of Parasites, a LaboratoryManual, Fundação Oswaldo Cruz. Rio de Janeiro, RJ, Brazil,1983, pp. 146–157.

[21] R.G. Kemp, Phosphofructokinase, Methods Enzymol. 42 (1975)72–77.

[22] H.-J. Kuo, D.A. Malencik, R.-S. Liou, S.R. Anderson, FactorsaVecting the activation of rabbit muscle phosphofructokinase byactin, Biochemistry 25 (1986) 1278–1286.

[23] K.S. Heard, N. Fidyk, A. Carruthers, ATP-dependent substrateocclusion by the human erythrocyte sugar transporter, Biochemis-try 39 (2000) 3005–3014.

P. Zancan, M. Sola-Penna / Molecular Genetics and Metabolism 86 (2005) 401–411 411

[24] P.S. Low, P. Rathinavelu, M.L. Harrison, Regulation of glycolysisvia reversible enzyme binding to the membrane protein, band 3, J.Biol. Chem. 268 (1993) 14627–14631.

[25] S. Renner, V. Prohaska, C. Gerber, D. Niethammer, G. Bruchelt,Analysis of metabolites of glucose pathways in human erythrocytesby analytical isotachophoresis, J. Chromatogr. A 916 (2001) 247–253.

[26] O.H. Lowry, N.L. Rosenbrough, A.L. Farr, R.J. Randall, Proteinmeasurement with the Folin phenol reagent, J. Biol. Chem. 193(1951) 265–275.

[27] M. Sola-Penna, A.C. Santos, F.C. Serejo, G.G. Alves, T. El-Bacha,J. Faber-Barata, M.F. Pereira, A.T. Da Poian, M.M. Sorenson, Aradioassay for phosphofructokinase-1 activity in cell extracts andpuriWed enzyme, J. Biochem. Biophys. Methods 50 (2002) 129–140.

[28] U.K. Laemmli, Cleavage of structural proteins during the assem-bly of the head of bacteriophage T4, Nature 227 (1970) 680–685.

[29] G.G. Alves, L.M.T.R. Lima, M.P. Favero-Retto, A.P. Lemos, C.E.Peres-Sampaio, J.R. Meyer-Fernandes, A. Vieyra, M. Sola-Penna,p-Nitrophenylphosphatase activity catalyzed by plasma mem-brane (Ca2+ + Mg2+)ATPase: correlation with structural changesmodulated by glycerol and Ca2+, Biosci. Rep. 21 (2001) 25–32.

[30] A.H. Khan, J.E. Pessin, Insulin regulation of glucose uptake: acomplex interplay of intracellular signaling pathways, Diabetolo-gia 45 (2002) 1475–1483.

[31] A. Ullrich, J.R. Bell, E.Y. Chen, R. Herrera, L.M. Petruzzeli, T.J.Dull, A. Gray, L. Coussens, Y.C. Liao, M. Tsubokawa, A. Maron,P.H. Seeburg, C. Grunfeld, O. Rosen, J. Ramachandran, Humaninsulin receptor and its relationship to the tyrosine kinase familyof oncogenes, Nature 313 (1985) 756–761.

[32] P. Boivin, C. Galand, The human red cell acid phosphatase is aphosphotyrosine protein phosphatase which dephosphorylates themembrane protein band 3, Biochem. Biophys. Res. Commun. 134(1986) 557–564.

[33] P. Boivin, Role of the phosphorylation of red blood cell mem-brane proteins, Biochem. J. 256 (1988) 689–695.

[34] G. Clari, L.D. Libera, V. Moret, Tyrosine phosphorylation ofcytosolic proteins in human erythrocytes, Biochem. Biophys. Res.Commun. 166 (1990) 1378–1383.

[35] G. Chétrite, R. Cassoly, AYnity of hemoglobin for the cyto-plasmic fragment of human erythrocyte membrane band 3.Equilibrium measurements at physiological pH using matrix-bound proteins: the eVects of ionic strength, deoxygenationand of 2,3-diphosphoglycerate, J. Mol. Biol. 185 (1985) 639–644.

[36] G.G. Alves, M. Sola-Penna, Epinephrine modulates cellular distri-bution of muscle phosphofructokinase, Mol. Genet. Metab. 78(2003) 302–306.

[37] T. El-Bacha, M.S. Freitas, M. Sola-Penna, Cellular distribution ofphosphofructokinase activity and implications to metabolic regu-lation in human breast cancer, Mol. Genet. Metab. 79 (2003) 294–299.

[38] T. El-Bacha, M.M. Menezes, M.C. Azevedo e Silva, M. Sola-Penna, A.T. Da Poian, Mayaro virus infection alters glucosemetabolism in cultured cells through activation of the enzyme6-phosphofructo 1-kinase, Mol. Cell. Biochem. 266 (2004) 191–198.

[39] M.L. Dustin, G.R. Jacobson, S.W. Peterson, EVects of insulinreceptor down-regulation on hexose transport in human erythro-cytes, J. Biol. Chem. 259 (1984) 13660–13663.

[40] S.W. Peterson, A.L. Miller, R.S. Kelleher, E.F. Murray, Insulinreceptor down-regulation in human erythrocytes, J. Biol. Chem.258 (1983) 9605–9607.

[41] C. Wilson, S.W. Peterson, Insulin receptor processing as afunction of erythrocyte age, J. Biol. Chem. 261 (1986) 2123–2128.