Triosephosphate isomerase deficiency: a neurodegenerative misfolding disease

Biochemical Society Transactions (2002) Volume 30, part 2

42

43 44 45

Brand, M. D. and Ainscow, E. K. (2000) in Technological and Medical Implications of Metabolic Control Analysis (Cornish-Bowden, A. and Ca! rdenas, M. L., eds), pp. 131–138, Kluwer Academic Publishers, Dordrecht Krauss, S., Buttgereit, F. and Brand, M. D. (1999) Biochim. Biophys. Acta 1412, 129–138 Krauss, S. and Brand, M. D. (2000) FASEB J. 14, 2581–2588 Brand, M. D. and Krauss, S. (2000) in Animating the Cellular Map (Hofmeyr, J.-H. S., Rohwer, J. M. and Snoep,

46 47

J. L., eds), pp. 153–158, Stellenbosch University Press, Stellenbosch Krauss, S., Brand, M. D. and Buttgereit, F. (2001) Immunity 15, 497–502 Ideker, T., Thorsson, V., Ranish, J. A., Christmas, R., Buhler, J., Eng, J. K., Bumgarner, R., Goodlett, D. R., Aebersold, R., and Hood, L. (2001) Science 292, 929–934

Received 20 November 2001

Triosephosphate isomerase deficiency: a neurodegenerative misfolding disease Judit Ola! h*, Ferenc Orosz*, Gyo$ rgy M. Keseru% †, Zolta! n Kova! ri†, Ja! nos Kova! cs‡, Susan Holla! n§ and Judit Ova! di*1 *Institute of Enzymology, Hungarian Academy of Sciences, Budapest, H-1518, P.O. Box 7, Hungary, †Chemical Works of Gedeon Richter Ltd., Budapest, Hungary, ‡Department of General Zoology, University of Eo$ tvo$ s Lora! nd, Budapest, Hungary, and §National Institute of Blood Transfusion, Budapest, Hungary ultrastructural (immunoelectron microscopy) data for characterization of mutant isomerase structures and for the TPI-related metabolic processes in normal and deficient cells. The relationships between mutation-induced TPI misfolding and formation of aberrant protein aggregates are discussed.

Abstract A number of neurodegenerative diseases are mediated by mutation-induced protein misfolding. The resulting genetic defects, however, are expressed in varying phenotypes. Of the several wellestablished glycolytic enzyme deficiencies, triosephosphate isomerase (TPI) deficiency is the only one in which haemolytic anaemia is coupled with progressive, severe neurological disorder. In a Hungarian family with severe decrease in TPI activity, two germ line-identical but phenotypically differing compound heterozygote brothers inherited two independent (Phe#%! Leu and Glu"%& stop codon) mutations. We have demonstrated recently [Orosz, Ola! h, Alvarez, Keseru% , Szabo! , Wa! gner, Kova! ri, Hora! nyi, Baro! ti, Martial, Holla! n and Ova! di (2001) Blood 98, 3106–3112] that the mutations of TPI explain in themselves neither the severe decrease in the enzyme activity characteristic of TPI deficiency nor the enhanced ability of the mutant enzyme from haemolysate of the propositus to associate with subcellular particles. Here we present kinetic (flux analysis), thermodynamic (microcalorimetry and fluorescence spectroscopy), structural (in silico) and

Relationship between protein misfolding and neurodegeneration Proteins exhibit a variety of motions due to their conformational flexibility. These macromolecules undergo fast dynamic interconversion between different conformational substates. These interconversions coupled with subunit exchange between oligomers may be rather slow on the biological time scale, resulting in a long-lived heterogeneous population, which is an important issue for their biological function [1]. The conformational changes of non-toxic forms of specific proteins or fragments can produce pathological conditions such as Alzheimer’s disease, Parkinson’s disease, Huntington’s disease and other neurodegenerative disorders. In the development of these diseases mutations of proteins with misfolded structures play an important role. With the aging of society, neurodegenerative diseases are becoming more widespread. These serious chronic diseases are the consequences of the destruction of groups of neurons in various areas of the brain. Research in the last 10 years has produced significant results in the understanding of the mechanism of the evolution of these diseases. The studies have revealed that the development of neurodegeneration is a multistep

Key words : kinetics, metabolism, microtubules, mutant protein. Abbreviations used : DHAP, dihydroxyacetone phosphate ; DSC, differential scanning calorimetry ; Glu-6-PDH, glucose-6phosphate dehydrogenase ; MD, molecular dynamics ; MT, microtubule ; TPI, triosephosphate isomerase ; mTPI ; human recombinant mutant (Phe240 Leu) TPI ; PFK, phosphofructokinase ; PPP, pentose phosphate pathway ; wTPI, human recombinant wild-type TPI ; βAPP, β-amyloid precursor protein. 1 To whom correspondence should be addressed (e-mail ovadi!enzim.hu).

# 2002 Biochemical Society

30

Metabolite Channelling and Metabolic Complexity

In this paper we present data on the structural and associative consequences and metabolic processes related to TPI mutation obtained with human recombinant enzymes and with haemolysates. We focus on the mutation-related structural and functional alterations of TPI occurring at the atomic, molecular, ultrastructural and metabolic levels.

process during which one or a few specific mutant protein species of altered conformation initiate non-physiological protein–protein interactions, resulting in enhanced aggregation. The aggregates form plaques, which are typical morphological hallmarks of the neurodegenerative process. The hetero-associations of the mutant proteins with subcellular structures, such as cytoskeleton, cell membranes or glycolytic enzymes, may be crucial in the initiation of neurodegeneration. Aberrant associations of mutant proteins with glycolytic enzymes are well documented; for example, glyceraldehyde-3-phosphate dehydrogenase associates with mutant huntingtin protein [2], and phosphofructokinase (PFK) binds to βamyloid fragments [3] or to ataxin proteins [4]. The potential involvement of a truncated fragment of the mutant huntingtin protein in the formation of Huntington’s disease was suggested recently [5]. Nevertheless, details of these processes, the mechanisms causing differences in the phenotypes of identical mutations and the different sensitivity of the neuronal structures towards neurotoxic effects are still unclear.

Catalytic and associative properties of the mutant enzymes We have recently reported that the specific activity of human recombinant mutant (Phe#%! Leu) TPI (mTPI) relative to human recombinant wildtype TPI (wTPI) is much higher (30 %) than expected from the activity (3 %) measured in haemolysates of normal control and deficient red cells (Table 1). Concerning the associative properties of the normal and mutant enzymes in purified form compared with in haemolysates, similar discrepancies were observed with both enzymes. Associations of TPIs with red cell membranes and microtubules (MTs) were investigated. The latter, which are the major constituent of the axon in neuronal cells, are likely to be a potent target of TPIs, as suggested by data obtained with a reconstituted bovine brain system [9]. We found that wTPI and mTPI showed similar capabilities of associating with MTs. The binding of the mutant enzyme from the haemolysate of deficient cells was more extensive compared with the normal control (Table 1). Qualitatively similar binding results were obtained with inside-outvesicles prepared from red cell membranes instead of MTs (results not shown). At high physiological protein concentrations one would expect a much higher binding ratio. This assumption was supported by similar experiments carried out in the presence of poly(ethylene glycol), a crowding agent that mimics the high physiological protein concentration favouring protein–protein interactions [10]. The hetero-association of TPIs with subcellular membranes decreases the catalytic activity of both wild-type and mutant enzymes [9]. This decrease of the catalytic activity may not be important in the case of normal cells, where the wild-type enzyme exhibits high excess catalytic activity compared with other glycolytic enzymes. However, in the case of deficient cells the additional decrease may result in metabolic blockage at the interconversion of triosephosphates. The hetero-association of TPI with MTs forms a superstructure, which can be visualized by electron microscopy. MTs assembled in the pres-

Triosephosphate isomerase (TPI) deficiency TPI is a glycolytic enzyme, which catalyses the interconversion of -glyceraldehyde 3-phosphate to dihydroxyacetone phosphate (DHAP). The rate of the catalysis is diffusion-limited [6], and the equilibrium favours the formation of DHAP by 1 : 20. Large numbers of glycolytic enzymopathies were found in red blood cells, resulting in haemolysis and other aberrations [7]. However, TPI deficiency is unique among the glycolytic enzyme defects since the chronic haemolytic anaemia in this case is associated with progressive neurological dysfunction and death in early childhood [8]. Patients with various inherited mutations have been identified and all but one of the homozygotes and several compound heterozygotes carry the Glu"!% Asp mutation. Compound heterozygote Hungarian brothers with their missense (Phe#%! Leu) and nonsense (Glu"%& stop) mutations are unique because (i) the sites and their combination of the mutations differ from all others identified so far, (ii) the elder brother of the two germline-identical compound heterozygotes is free of neurological manifestations and (iii) both, even the propositus with neurological symptoms, are over 20 years old. Mother, father and a simple heterozygote brother are symptom-free heterozygotes.

31

# 2002 Biochemical Society

Kinetic and binding properties of recombinant TPIs and TPIs from blood cells DHAP levels in whole blood [27], enzyme activities [28], TPI binding to MTs [9] and heat stability [18] were measured. The pentose phosphate pathway (PPP) was measured with 5 mM ribose 5-phosphate as a substrate, in the presence of 0.1 mM thiamine pyrophosphate, 1 mM NADP+ and 5 mM MgCl2 [29]. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; PEG, poly(ethylene glycol).

Binding to MTs (%)

Heat stability (1/min*)

PPP

PEGk

PEG+

MTk

MT+

– –

1.6 1.7

– 11.6

0.014 0.099

0.007 0.074

Catalytic properties (units/mg)

32

wTPI mTPI

DHAP metabolic level (µM)

TPI

Aldolase

GAPDH

– –

8000 2400

– –

– –

Glu-6-PDH – –

Catalytic properties (units/g of Hb)

Control Propositus

Binding to MTs (%)

Heat stability (1/min*)

DHAP metabolic level (µM)

TPI

Aldolase

GAPDH

Glu-6-PDH

PPP

PEGk

PEG+

MTk

MT+

11.4 701

1400 33

3.2 5.2

156 208

10 15.7

9.5 14.8

1.5 11

3.1 11

0.011 0.111

– –

*First-order rate constant of inactivation.

Biochemical Society Transactions (2002) Volume 30, part 2

# 2002 Biochemical Society

Table 1

Metabolite Channelling and Metabolic Complexity

Figure 1 Transmission (A, B) and immuno- (C, D) electron microscopic images of MTs assembled in the absence (A) or presence of wild-type (C) or mutant TPI (B, D) The structure of MTs is similar on sections prepared from glutaraldehyde/osmium tetroxide-fixed samples (A, B). Immunogold decoration suggests that both types of the enzyme can bind to the surface of MTs (C, D). Scale bar, 150 nm. Transmission electron microscopy and immunoelectron microscopy were carried out as described in [11] and [12], respectively.

ence of either mutant or wild-type TPI, as described previously [11,12], were apparently intact and similar to those polymerized without TPI (Figures 1A and 1B). As TPI is a relatively small protein with a molecular mass of 53 kDa, its direct detection by conventional transmission electron microscopy is rather difficult. To see whether wTPI or mTPI was present on the surface of MTs, immunoelectron microscopic images were taken by the use of an anti-TPI antibody, which recognizes both TPIs. As shown in Figure 1, the immunogold conjugates are clearly seen on the surface of MTs prepared with addition of wTPI (Figure 1C) and mTPI (Figure 1D). Nonspecific labelling by gold-linked secondary antibody was minimal on control sections incubated in the absence of primary antibody and no immunostaining was seen on MTs polymerized without addition of TPI (results not shown).

from persistent conformational heterogeneity in the ensemble of native oligomers [17]. A prerequisite for such heterogeneity is that the dynamics of subunit exchange\unfolding should be slow relative to the experimental time scale [1]. Our recent heat-inactivation experiments showed that the stability of mTPI is significantly reduced compared with wTPI (the t . !& value of heat inactivation was approximately one order of magnitude lower) [18]. In the present study the unfolding process of the recombinant enzymes were investigated by differential scanning calorimetry (DSC). The thermograms recorded in the temperature range of 10–75 mC (see Figure 2, upper panel) indicate distinct thermal unfolding processes. Both enzymes undergo a temperature-induced transition from the native state to the unfolded one. Whereas in the case of wTPI the heat-absorption curve shows a single sharp transition with its melting point at 66.2 mC, the unfolding of the mutant enzyme is induced at a lower temperature (65.4 mC) ; furthermore, it shows an additional transition at 54.8 mC. These findings suggest that the mutation of Phe#%! results in a protein with lowered stability, probably due to an altered conformational state. Different mechanisms could be responsible for the complex behaviour of the mutant enzyme. A plausible explanation is that the heat-assisted conformational changes of the mutant enzyme are coupled with dissociation of the dimeric species, which is more extensive in the case of the mutant enzyme. Such

Folding and stability of mutant TPI TPI, which conserved its structural and functional properties across species [13], is a stable homodimer [14]. This is supported by its marked resistance against oxidation and proteolysis, during which it maintains the native dimeric structure [15]. Unfolding\refolding experiments revealed that the refolding of TPI proceeds from unfolded monomer to folded inactive monomer, and then to native, active dimers [16]. The dissociation\ unfolding processes are independent of the protein concentration [15]. This behaviour can be derived

33

# 2002 Biochemical Society

Biochemical Society Transactions (2002) Volume 30, part 2

a complex thermal-stability profile has not been documented for TPI previously, although similar types of study were carried out with other mutant TPIs [19,20]. The conformational stability of mTPI compared with the wild-type enzyme was also investigated by fluorescence spectroscopy. The effect of the mutation on TPI stability was assessed by monitoring changes in the fluorescence emission peak wavelength as a function of temperature elevation. The heat-induced denaturation resulted

in a red-shifted emission peak wavelength from 325 to 348 nm for both recombinant enzymes (Figure 2, lower panel); however, the inflection points of the unfolding curves were different, but similar to those obtained with DSC. Our recent CD measurements have indicated that the α-helix content of mTPI, which is similar to that of wTPI at 30 mC, is almost abolished at 60 mC, whereas only a small decrease in α-helix content of wTPI could be detected at this temperature [18]. All these data reveal the crucial role of Phe at

Figure 2 Thermal stability of recombinant mTPI and wTPI monitored by DSC (upper panel) and intrinsic fluorescence (lower panel) Upper panel : DSC thermograms of wild-type (solid line) and mutant (dotted line) TPIs in 20 mM Tris buffer, pH 8.0. Enzyme concentration, 10 µg/ml ; heating rate, 1 mC/min. Lower panel : dependence of fluorescence emission peak wavelength (λmax) on temperature. The concentration of TPIs was 50 µg/ml. Excitation wavelength was 280 nm. Curves are drawn to visualize the difference between mTPI (#, dashed line) and wTPI ($, solid line).

(°C)

# 2002 Biochemical Society

34

Metabolite Channelling and Metabolic Complexity

duced by the nonsense mutation (Glu"%& stop) can bind to wild-type monomer, forming a heterodimer, has been supported by our molecular dynamics (MD) simulation [18]. The calculations, based upon the X-ray crystal structure of wTPI, suggested that the wild-type monomer–truncatedfragment heterodimer is a stable species (Figure 3). However, in the deficient cells two kinds of polypeptide chain could be synthesized : a whole chain with a missense mutation and a truncated fragment. Figure 3 also shows that a heterodimer can be formed from these polypeptide chains. Our MD model reveals a distinct conformational state

position of 240 in maintaining the conformational stability of TPI.

Searching for misfolding promoting factors Two experimental observations need explanation concerning the mutation-induced effects : (i) the activity of TPI in deficient cells is significantly lower, as compared with that caused by mutation and (ii) the association of the mutant TPI with subcellular particles is enhanced in haemolysate compared with the relevant recombinant mutant enzyme. The idea that the truncated peptide pro-

Figure 3 MD structures of the truncated heterodimeric TPIs MD simulations were based on the crystal structure of human recombinant TPI dimer [30]. The crystal structure was mutated to have the mutant starting models (one dimer from a wild-type and a truncated monomer, another dimer from a Phe240 Leu mutant and a truncated monomer). The backbone atoms of the dimer structures were fitted with a root mean square distance of 6.54 AH , using the Sybyl 6.7 software package (Tripos, St Louis, MO, U.S.A.). The dimers consist of a wild-type (red) plus a truncated monomer (orange) and a mutant (Phe240 Leu ; blue) plus a truncated (cyan) monomer. MD calculations were carried out as described previously [18].

35

# 2002 Biochemical Society

Biochemical Society Transactions (2002) Volume 30, part 2

the disease. Generation of βAPP starts a vicious circle because it binds in nanomolar concentrations to the glycolytic key enzyme PFK ([3] and references therein). The inhibition of PFK decreases the glucose turnover and subsequently the oxidative phosphorylation. Once oxidative phosphorylation has dropped below a critical level, βAPP can only partly be inserted into the synaptic membranes and β-amyloid is generated. In this way the decrease of PFK activity in senile demented brain cortex can be caused by the PFK– β-amyloid interaction. In brain the major energy source is the glucose that is metabolised by glycolysis. Similarly, mature red cells depend almost solely on anaerobic glycolysis to produce the energy required for their functions. Therefore, it is obvious that deficiency of glycolysis may induce pathological processes. Consequently, the glycolytic pathway has been the objective of many theoretical studies, and computer models that simulate this network in normal and deficient red blood cells have been evaluated

in the active-site region due to the mutation (results not shown) ; however, it may not extend to the subunit contact surface since the two heterodimeric structures display similar global conformational states. Therefore, our MD simulation provided in silico evidence for the formation of the mutation-containing heterodimeric TPI. This artificial species, if present in vivo, may display very different catalytic and aggregation properties.

Metabolism in the TPI-deficient diseased cell In many diseases the structural changes of proteins result in functional alteration, from which metabolic disorders originate. However, the causeconsequence relations in these processes remain obscure in many cases. A typical example is Alzheimer’s disease, in which alteration of the critical levels of glucose and ATP turnovers ensuring the proper insertion of β-amyloid precursor protein (βAPP) into the cellular or synaptic membranes seems to be crucial in the evolution of

Scheme 1 The PPP cross-talking with glycolysis Glu-6-PDH, glucose-6-phosphate dehydrogenase ; GAP, D-glyceraldehyde 3-phosphate.

# 2002 Biochemical Society

36

Metabolite Channelling and Metabolic Complexity

[21–23]. Nevertheless, no clear data are available on the energy metabolism of patients with severe TPI deficiency. It is an extensively discussed issue that the low TPI activity in deficient cells may lead to a metabolic block in the glycolytic pathway that results in an increased (20–60-fold) concentration of DHAP (for reviews see [24,25]), especially if the hetero-association of the mutant TPI further reduces the catalytic activity, as we have suggested previously [9]. This phenomenon can explain why the DHAP level is so high in red cells. However, in platelets from TPI-deficient individuals only modest elevation of DHAP level was detected [18], probably because in platelets, just as in many other cells including brain cells, DHAP is not a dead-end product but can be converted in the direction of lipid synthesis. This fact, plus an early finding [26] that another enzymopathy coupled with high DHAP levels does not cause neurological disorder, suggest that the high level of DHAP may not be responsible for the clinical symptoms. On the other hand, it has been reported that the concentrations of 2,3-diphosphoglycerate, and more importantly the ATP level, in TPIdeficient red blood cells do not differ significantly from those of normal controls [27]. To assess this apparent inconsistency, the activities of other glycolytic enzymes as well as the flux of the pentose phosphate pathway (PPP) cross-talking with glycolysis via common intermediates (Scheme 1) were measured in normal and in deficient cell-free haemolysates. As shown in Table 1, we have found that while the activity of TPI was much lower in the deficient cell haemolysate as compared with the normal control, the activity of aldolase and glyceraldehyde-3-phosphate dehydrogenase was higher. These two enzymes catalyse the conversion of hexose (fructose 6-phosphate) and triose (-glyceraldehyde 3-phosphate) phosphates, which are involved in both pathways. The key regulatory enzyme of the PPP is glucose6-phosphate dehydrogenase (Glu-6-PDH), catalysing the oxidative conversion of glucose 6-phosphate in the direction of ribose 5-phosphate formation. The further conversion of this metabolite results in the formation of the common intermediates of the two pathways. We have found that the activity of Glu-6-PDH in the haemolysate of the deficient cells is significantly higher than that of the control, which may result in a more active PPP in the deficient cells. This idea is supported by direct flux measurement when the reduction of NADPH to NADP+ catalysed by

Glu-6-PDH was monitored, initiating the pathway by adding ribose 5-phosphate as substrate. As shown in Table 1 the PPP flux is significantly higher in deficient cells as compared with the normal controls. Therefore, the interconnection of the two pathways can compensate for the reduced TPI activity of the deficient cells, and it could be responsible for the normal ATP level found in the TPI-deficient erythrocytes.

Outlook In recent years many observations have revealed that the protein aggregation induced by protein misfolding into fibrils and deposition in senile plaques is primarily related to neurotoxicity in the cases of several neurological diseases. TPI enzymopathy is a unique glycolytic enzyme deficiency coupled with neurodegeneration. We present data on the mutation-induced misfolding process, which probably plays an important role in the enhanced associations of mTPI. The mutant enzyme may interact with the truncated fragment via subunit exchange, or with MTs in the brain. No dependence on ionic strength was found in these protein–protein associations. Aberrant hetero-associations primarily involved in the process may lead to fibrillar aggregation, a potent constituent of the senile plaque formation similar to that detected in the case of Alzheimer’s disease and other neurodegenerative disorders. However, on the basis of recent clinical and experimental results obtained with the compound heterozygote Hungarian brother (results not shown), it became obvious that the mutations alone are not sufficient to explain the extreme variability of neurodegenerative diseases. That phenotype is not determined necessarily by genotype is demonstrated by our recent data. Additional investigations of protein conformational dynamics and their relation to folding and protein– protein interaction by means of basic physicochemical methods are necessary to understand the human syndrome. Our studies presented here may be relevant to the understanding of the molecular basis of human ‘ conformational ’ or ‘ misfolding ’ diseases. This work was supported by grants from the Hungarian National Science Foundation OTKA (T-025291 and T-031892 to J. Ova! di, T-029910 to J. K., T-029924 and T-035019 to F. O., T-30044 to G. M. K. and T-033138 to S. H.) and from the Hungarian Ministry of Education (NKFP 1/47 to J. Ova! di). We thank Professor Joseph A. Martial for kindly providing recombinant TPIs.

37

# 2002 Biochemical Society

Biochemical Society Transactions (2002) Volume 30, part 2

17

References 1 2

3 4

5 6 7 8 9

10 11 12 13 14

15 16

Ferreira, S. T. and De Felice, F. G. (2001) FEBS Lett. 498, 129–134 Burke, J. R., Enghild, J. J., Martin, M. E., Jou, Y.-S., Myers, R. M., Roses, A. D., Vance, J. M. and Strittmatter, W. J. (1996) Nat. Med. 2, 347–350 Meier-Ruge, W. A. and Bertoni-Freddari, C. (1997) Ann. N.Y. Acad. Sci. 826, 229–241 Koshy, B., Matilla, T., Burright, E. N., Merry, D. E., Fischbeck, K. H., Orr, H. T. and Zoghbi, H. Y. (1996) Hum. Mol. Genet. 5, 1311–1318 Cha, J.-H. J. (2000) Trends Neurosci. 23, 387–392 Rose, I. A., Fung, W. J. and Warms, J. V. B. (1990) Biochemistry 29, 4312–4317 Tanaka, K. R. and Zerez, C. R. (1990) Semin. Hematol. 27, 165–185 Arya, R., Layton, D. M. and Bellingham, A. J. (1995) Blood Rev. 9, 165–175 Orosz, F., Wa! gner, G., Liliom, K., Kova! cs, J., Baro! ti, K., Hora! nyi, M., Farkas, T., Holla! n, S. and Ova! di, J. (2000) Proc. Natl. Acad. Sci. U.S.A. 97, 1026–1031 Minton, A. P. and Wilf, J. (1981) Biochemistry 20, 4821–4826 Lehotzky, A., Pa! lfia, Z., Kova! cs, J., Molna! r, A. and Ova! di, J. (1994) Biochem. Biophys. Res. Commun. 204, 585–591 Ve! rtessy, B. G., Kova! cs, J., Lo% w, P., Lehotzky, A., Molna! r, A., Orosz, F. and Ova! di, J. (1997) Biochemistry 36, 2051–2062 Lu, H. S., Yuan, P. M. and Gracy, R. W. (1984) J. Biol. Chem. 259, 11958–11968 Mainfroid, V., Terpstra, P., Beauregard, M., Fre! re, J.-M., Mande, S., Hol, W. G. J., Martial, J. and Goraj, K. (1996) J. Mol. Biol. 257, 441–456 Rietveld, A. W. M. and Ferreira, S. (1998) Biochemistry 37, 933–937 Borchert, T. V., Zeelen, J. P., Schliebs, W., Callens, M., Minke, W., Jaenicke, R. and Wierenga, R. K. (1995) FEBS Lett. 367, 315–318

18

19

20

21 22 23

24 25 26 27

28

29 30

Erijman, L. and Weber, G. (1991) Biochemistry 30, 1595–1599 Orosz, F., Ola! h, J., Alvarez, M., Keseru% , G. M., Szabo! , B., Wa! gner, G., Kova! ri, Z., Hora! nyi, M., Baro! ti, K., Martial, J. A. et al. (2001) Blood 98, 3106–3112 Alvarez, M., Zeelen, J. P., Mainfroid, V., Rentier-Delrue, F., Martial, J. A., Wyns, L., Wierenga, R. K. and Maes, D. (1998) J. Biol. Chem. 273, 2199–2206 Norledge, B. V., Lambeir, A. M., Abagyan, R. A., Rottman, A., Fernandez, A. M., Filimonov, V. V., Peter, M. G. and Wierenga, R. K. (2001) Proteins Struct. Funct. Genet. 42, 383–389 Schuster, R. and Holzhu$ tter, H.-G. (1995) Eur. J. Biochem. 229, 403–418 Heinrich, R., Montero, F., Klipp, E., Waddell, T. G. and Melendez-Hevia, E. (1997) Eur. J. Biochem. 243, 191–201 Martinov, M. V., Plotnikov, A. G., Vitvitsky, V. M. and Ataullakhanov, F. I. (2000) Biochim. Biophys. Acta 1474, 75–87 Valentine, W. N. and Paglia, D. E. (1984) Blood 64, 583–591 Orosz, F., Ve! rtessy, B., Holla! n, S., Hora! nyi, M. and Ova! di, J. (1996) J. Theor. Biol. 182, 437–447 Rosa, R., Prehu, M.-O. and Beuzard, Y. (1978) J. Clin. Invest. 62, 907–915 Holla! n, S., Fujii, H., Hirono, A., Hirono, K., Karro, H., Miwa, S., Harsanyi, V., Gyodi, E. and Inselt-Kovacs, M. (1993) Hum. Genet. 92, 486–490 Beutler, E., Blume, K. G., Kaplan, J. C., Lo$ hr, G. W., Ramot, B. and Valentine, W. N. (1977) Br. J. Haematol. 35, 331–340 Cabezas, H., Raposo, R. R. and Melendez-Hevia, E. (1999) Mol. Cell. Biochem. 201, 57–63 Mande, S. H., Mainfroid, V., Kalk, K. H., Goraj, K., Martial, J. A. and Hol, W. G. J. (1994) Protein Sci. 3, 810–821

Received 19 November 2001

Multiple glucose 6-phosphate pools or channelling of flux in diverse pathways ? Loranne Agius*1, Josep Centelles† and Marta Cascante† *Department of Diabetes and Metabolism, University of Newcastle upon Tyne, Newcastle upon Tyne NE2 4HH, U.K., and †Departament de Bioquı! mica i Biologia Molecular, Facultat de Quı! mica, Universitat de Barcelona, Martı! i Franque' s, 1, 08028 Barcelona, Spain have suggested that there are multiple subcellular pools of glucose 6-phosphate. It is proposed that this data can be interpreted in terms of channelling of metabolic intermediates through multiple pathways of glucose metabolism with leakage of glucose 6-phosphate from the channels into a single free pool. It is also proposed that measurement of total tissue content of glucose 6-phosphate approximates the free pool.

Abstract Glucose 6-phosphate is an intermediate of pathways of glucose utilization and production as well as a regulator of enzyme activity and gene expression. Studies on the latter functions are in part based on measurement of the glucose 6-phosphate content in a whole-cell extract. Several studies Key words : compartmentation, glycolysis, intermediates. Abbreviations used : glucose 6-P, glucose 6-phosphate ; glucose 1-P, glucose 1-phosphate ; fructose 1-P, fructose 1-phosphate ; fructose 6-P, fructose 6-phosphate. 1 To whom correspondence should be addressed (e-mail Loranne.Agius!ncl.ac.uk).

# 2002 Biochemical Society

Introduction Glucose 6-phosphate (glucose 6-P) is the first intermediate in the metabolism of glucose by

38