Biochemical Journal

Research article

Transaldolase deficiency influences the pentose phosphate pathway, mitochondrial homoeostasis and apoptosis signal processing

Yueming Qian, Sanjay Banerjee, Craig E. Grossman, Wendy Amidon, Gyorgy Nagy, Maureen Barcza, Brian Niland, David R. Karp, Frank A. Middleton, Katalin Banki, Andras Perl


TAL (transaldolase) was originally described in the yeast as an enzyme of the PPP (pentose phosphate pathway). However, certain organisms and mammalian tissues lack TAL, and the overall reason for its existence is unclear. Recently, deletion of Ser171 (TALΔS171) was found in five patients causing inactivation, proteasome-mediated degradation and complete deficiency of TAL. In the present study, microarray and follow-up Western-blot, enzyme-activity and metabolic studies of TALΔS171 TD (TAL-deficient) lymphoblasts revealed co-ordinated changes in the expression of genes involved in the PPP, mitochondrial biogenesis, oxidative stress, and Ca2+ fluxing. Sedoheptulose 7-phosphate was accumulated, whereas G6P (glucose 6-phosphate) was depleted, indicating a failure to recycle G6P for the oxidative branch of the PPP. Nucleotide analysis showed depletion of NADPH and NAD+ and accumulation of ADP-ribose. TD cells have diminished Δψm (mitochondrial transmembrane potential) and increased mitochondrial mass associated with increased production of nitric oxide and ATP. TAL deficiency resulted in enhanced spontaneous and H2O2-induced apoptosis. TD lymphoblasts showed increased expression of CD38, which hydrolyses NAD+ into ADP-ribose, a trigger of Ca2+ release from the endoplasmic reticulum that, in turn, facilitated CD20-induced apoptosis. By contrast, TD cells were resistant to CD95/Fas-induced apoptosis, owing to a dependence of caspase activity on redox-sensitive cysteine residues. Normalization of TAL activity by adeno-associated-virus-mediated gene transfer reversed the elevated CD38 expression, ATP and Ca2+ levels, suppressed H2O2- and CD20-induced apoptosis and enhanced Fas-induced cell death. The present study identified the TAL deficiency as a modulator of mitochondrial homoeostasis, Ca2+ fluxing and apoptosis.

  • apoptosis
  • Ca2+
  • mitochondrion
  • pentose phosphate pathway
  • transaldolase


Metabolism of glucose through the PPP (pentose phosphate pathway) fulfils two unique functions: formation of R5P (ribose 5-phosphate) for the synthesis of nucleotides, RNA and DNA, and generation of NADPH as a reducing equivalent for biosynthetic reactions and maintenance of a reducing environment [1]. ROI (reactive oxygen intermediates) have long been considered as toxic by-products of aerobic existence, but evidence is now accumulating that controlled levels of ROI modulate cellular function and are necessary for signal-transduction pathways, including those mediating PCD (programmed cell death) [2]. A normal reducing atmosphere, required for cellular integrity is provided by GSH, which protects cells from damage by ROI. Regeneration of GSH from its oxidized form, GSSG, is dependent on NADPH produced by the PPP [1].

Understanding regulation of the PPP has been complicated by the fact that the pathway is comprised of two separate, oxidative and non-oxidative, branches [1]. Reactions in the oxidative branch are irreversible, whereas all reactions of the non-oxidative phase are fully reversible. TAL (transaldolase) was, on the basis of metabolites and enzymatic activities detected in yeast [3], originally described as an enzyme of the non-oxidative branch of the PPP. However, certain organisms [4,5] and mammalian tissues do not express TAL [6] and the non-oxidative branch can function without this enzyme [7,8]. Hence the overall reason for the existence of TAL has not been clearly established. Mutations in TAL have been associated with liver cirrhosis in children from three families [911]. Deletion of Ser171 was noted in five out of six cases [9,11], and this was found to lead to inactivation, proteasome-mediated degradation and complete deficiency of TAL [12]. The pathogenesis of liver cirrhosis has been associated with increased cell death of hepatocytes and biliary epithelial cells [13]. To understand the consequences of TAL deficiency on cell-death-signal processing, B-cell LBs (lymphoblasts) available from the first TD (TAL-deficient) patient were investigated. Relative to control LBs, TD B-cells exhibit increased spontaneous, H2O2-, and CD20-induced apoptosis and reduced CD95/Fas-dependent apoptosis. Altered cell-death-signal processing is mediated by a unique pattern of metabolic changes affecting (1) the PPP (characterized by elevated S7P (sedoheptulose 7-phosphate) and diminished G6P (glucose 6-phosphate) levels, which reflect blocked recycling of G6P for the oxidative branch, (2) nucleotide metabolism (depletion of NADPH and elevation of ADP-ribose), (3) Ca2+ fluxing (increased cytoplasmic and mitochondrial Ca2+levels and elevated CD38 expresssion) and (4) decreased Δψm (mitochondrial transmembrane potential) and increased NO (nitric oxide) production, increased mitochondrial mass, increased ROI and ATP production. Normalization of TAL activity reversed the elevated CD38 expression and Ca2+ levels, the increased H2O2- and CD20-induced apoptosis and the diminished Fas-induced apoptosis. The results indicate that TAL regulates the PPP via G6P recycling and thus controls NADPH production, nucleotide metabolism, Ca2+ fluxing, the Δψm and pathway-specifically influences apoptosis signal processing of human B-cells.


Cell culture

Human EBV (Epstein–Barr virus)-transformed human B-cell LBs were cultured in RPMI 1640 medium supplemented with 10% (v/v) FBS (fetal bovine serum), 2 mM L-glutamine, 100 units/ml penicillin, 100 μg/ml streptomycin and 0.25 μg/ml amphotericin B. EBV-transformed LBs from TD patients and human control donors [GM13416 (Coriel Cell Repositories, Camden, NJ, U.S.A.) and JAB (a B-cell line we established in our laboratory from a human donor)] were established as described previously [12]. Human FBs (fibroblasts) were maintained in Hams F-10 medium supplemented with 20% FBS, 2 mM L-glutamine, 100 units/ml penicillin, 100 μg/ml streptomycin and 0.25 μg/ml amphotericin B. FBs were isolated from a skin biopsy obtained from the lateral aspect of the left thigh from a TAL-deficient patient [9] and age-matched female controls (#1 and #2). All cell lines were maintained in a humidified atmosphere with 5% CO2 at 37 °C. Cell-culture products were purchased from Cellgro (Mediatech, Herndon, VA, U.S.A.).

Induction of apoptosis and cell viability assays

Apoptosis was induced with H2O2 (100 μM; Sigma, St Louis, MO, U.S.A.) and Fas Ab (antibody) CH-11 (1 μg/ml; Upstate Biotechnology, Saranac Lake, NY, U.S.A.) as previously described [14,15]. CD20-mediated apoptosis was induced by cross-linking with goat anti-human IgG. Tissue-culture wells were pretreated with 100 μg/ml goat anti-human IgG (Jackson ImmunoResearch Laboratories, West Grove, PA, U.S.A.) at 37 °C for 2 h, washed twice with PBS, then treated with CD20 Ab (10, 100 or 1000 μg/ml rituximab; Genentech, South San Francisco, CA, U.S.A.), and washed again twice with PBS before the addition of cells at 106/ml. Apoptosis was quantified by staining with annexin V-Alexa 647 (Molecular Probes; excitation at 650 nm; emission at 665 nm recorded in FL-5) and exclusion of propidium iodide (excitation at 535 nm; emission at 617 nm recorded in FL-2).

Flow-cytometric analysis of Δψm, mitochondrial mass, ROI and NO production and cytoplasmic and mitochondrial Ca2+ level

Δψm was estimated by staining with 10 nM DiOC6 (3,3′-dihexyloxacarbocyanine iodide; Molecular Probes) [15] for 15 min at 37 °C in the dark before flow cytometry (excitation at 488 nm; emission at 525 nm recorded in FL-1). Δψm was also quantified using a potential-dependent J-aggregate-forming lipophilic cation, JC-1 (5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolocarbocyanine iodide). JC-1 selectively incorporates into mitochondria, where it forms monomers (fluoresces green at 527 nm) or aggregates, at high transmembrane potentials (fluoresces red at 590 nm). Co-treatment with a protonophore, 5 μM mClCCP (carbonyl cyanide m-chlorophenylhydrazone; Sigma) for 15 min at 37 °C resulted in decreased DHR (dihydrorhodamine 123), DiOC6, and JC-1 fluorescence and served as a positive control for disruption of Δψm [15]. Δψm was also assessed by staining with 1 μM TMRM (tetramethylrhodamine methyl ester; excitation at 543 nm; emission at 567 nm recorded in FL-2; all from Molecular Probes). Mitochondrial mass was assessed by staining with potential-insensitive mitochondrial dyes 50 nM NAO (nonyl Acridine Orange; excitation at 490 nm; emission at 540 nm recorded in FL-1; Molecular Probes) or with 100 nM MTG (MitoTracker Green-FM; excitation at 490 nm; emission at 516 nm recorded in FL-1; Molecular Probes). Production of ROI was assessed using oxidation-sensitive fluorescent probes DCFH-DA (5,6-carboxy-2′,7′-dichlorofluoresceindiacetate), DHR and HE (hydroethidine; Molecular Probes) as previously described [14]. Although rhodamine 123, the fluorescent product of DHR oxidation, binds selectively to the mitochondrial inner membrane, ethidium and DCF (5,6-carboxy-2′,7′dichlorofluorescein) remain in the cytosol of living cells. DCF and HE (hydroethidine) preferentially detect H2O2 and superoxide (O2) respectively [16]. Intracellular glutathione levels were assessed with 100 μM MCB (monochlorobimane; excitation at 380 nm; emission at 461 nm; FL-UV). Production of nitric oxide (NO) was assessed by using DAFFM (4-amino-5-methylamino-2′,7′-difluoroflourescein diacetate; Molecular Probes). Measurement of NO was calibrated by incubating cells with the NO donors NOC-18 {(Z)-1-[2-(2-aminoethyl)-N-(2-ammonioethyl)amino]diazen-1-ium 1,2-diolate; 200 μM–1.8 mM} and sodium nitroprusside (400 μM–10 mM). C-PTIO (carboxy-2-phenyl-4,4,5,5-tetramethyl-imidazoline-1-oxyl-3-oxide; 500 μM, 24 h), a specific NO chelator [17], was used to reduce NO levels and inhibit NO signalling. The [Ca2+]c (cytoplasmic Ca2+ concentration) was measured by loading the cells with 1 μM Fluo-3 acetoxymethyl ester (excitation at 506 nm, emission at 526 nm recorded in FL-1; Molecular Probes). The [Ca2+]m (mitochondrial Ca2+ concentration) was estimated by loading the cells with 4 μM Rhod-2 acetoxymethyl ester, which is compartmentalized into the mitochondria [18]. 2-APB (2-aminoethoxydiphenyl borane), a membrane-permeant antagonist of the IP3R [Ins(1,4,5)P3 receptor] and BAPTA-AM [1,2-bis-(o-aminophenoxy)ethane-N,N,N′,N′-tetra-acetic acid tetrakis(acetoxymethyl ester); 5 μM] were used to inhibit Ca2+ signalling [17].

Western-blot analysis

Cytosolic lysates of LBs and FBs were prepared by harvesting cells in a mixture comprising 40 mM triethanolamine, pH 7.6, 10 mM EDTA, 1 mM sodium orthovanadate, 0.1 mM sodium molybdate, 10 mM sodium pyrophosphate and 50 mM NaF. Protein concentrations were determined by the Bradford method using the Bio-Rad Protein Assay (Bio-Rad Laboratories, Hercules, CA, U.S.A.). A 10 μg portion of protein lysates, unless otherwise indicated, were separated on an SDS/12%-(w/v)-polyacrylamide gel and electroblotted on to nitrocellulose. For whole-cell lysates, 2×105 cells were resuspended in 10 μl of sample buffer and boiled at 95 °C for 5 min prior to loading. Nitrocellulose strips were immunoblotted with anti-TAL Ab 170 and β-actin-specific mouse Ab 1501R (Chemicon, Temecula, CA, U.S.A.) that had been treated overnight with blocking reagents as previously described [19]. G6PD (G6P dehydrogenase) was detected with a rabbit Ab specific to G6PD (catalogue no. A300-404A; Bethyl Laboratories, Montgomery, TX, U.S.A.). Horseradish-peroxidase-conjugated goat anti-rabbit IgG (Jackson ImmunoResearch Laboratories) was utilized as a secondary Ab for TAL, whereas biotin-conjugated goat anti-mouse IgG and streptavidin (Jackson ImmunoResearch Laboratories) were used as secondary and tertiary reagents respectively for the detection of β-actin. Protein bands were visualized by enhanced chemiluminescence using Western Lightning Chemiluminescence Reagent Plus (PerkinElmer, Boston, MA, U.S.A.) using a Kodak Image Station 440CF and quantified with Kodak 1D Image Analysis Software (Eastman Kodak Company, Rochester, NY, U.S.A.).

Enzyme-activity assays

TAL, TK (transketolase) and G6PD activities were measured as described previously [14]. GPX (glutathione peroxidase) activity was measured in the presence of a mixture containing 150 mM KH2PO4, pH 7.0, 1 mM GSH, 0.25 mM NADPH, 3 units of glutathione reductase and 1 mM H2O2 [20]. Catalase was measured in the presence of 50 mM KH2PO4/Na2HPO4, pH 7.0, and 10 mM H2O2 at room temperature (20 °C) by continuous reading of A240 for 3 min [21]. SOD (superoxide dismutase) activity was measured in the presence of 0.05 M sodium pyrophosphate buffer, pH 8.3, 186 μM phenazine methosulphate, 300 μM Nitroblue Tetrazolium and 780 μM NADH [22].

Measurement of phosphorylated sugars by HPLC

A total of 107 cells were resuspended in 1 ml of 10% perchloric acid, and after three freeze–thaw cycles the precipitate was removed by centrifugation (15000 g, 10 min, 4 °C). The supernatant was neutralized with 120 μl of 10 M KOH. After centrifugation, the supernatant was filtered through a 0.2-μm-pore-size PVDF membrane, dried in a Speedvac apparatus and dissolved in 100 μl of water. The total protein content of each sample was determined using the Lowry assay [23]. A 25 μl portion was injected into the HPLC chromatograph equipped with a Dionex ED40 amperometric detector and the material analysed on a Carbopack PAI column (Dionex, Sunnyvale, CA, U.S.A.) using a gradient of 100 mM NaOH/1 M sodium acetate in 100 mM NaOH [24].

Measurement of nucleotides by HPLC

Pellets from 107 cells washed in PBS were resuspended in 50 μl of 0.16 M KCl/5 mM glucose at 4 °C. Then 100 μl of ice-cold 0.5 M KOH was added. The sample was mixed, diluted with 100 μl of ice-cold water and neutralized with 30 μl of cold 1 M KH2PO4. The extracts were centrifuged for 10 min at 15000 g at 4 °C, the supernatants were filtered through a 0.45-μm-pore-size polypropylene membrane, and 10–20 μl aliquots were injected into a 125 mm (length)×4.6 mm (diameter) Hypersil column packed with 3-μm-particle-size ODS-2 (Millipore Waters Chromatography, Milford, MA, U.S.A.) attached to a Waters Alliance HPLC chromatograph equipped with a Model 996 photodiode array detector set at 254 nm. The mobile phase was made with buffers A and B. Buffer A consisted of 0.1 M KH2PO4, pH 5.5, and 8 mM tetrabutylammonium hydrogen sulfate. Buffer B consisted of 70% (v/v) buffer A and 30% (v/v) methanol. Gradient elution was performed as follows: from 10% B to 75% B in 11 min; to 100% B at 12 min; back to 10% B at 21 min [25].

ATP measurement by chemiluminescence

Intracellular ATP levels were determined using the luciferin–luciferase method [26].

Electron microscopy

Cell pellets were fixed overnight in PBS with 2.5% (w/v) glutaraldehyde, post-fixed in 1% OsO4, dehydrated in a graded series of ethanols, infiltrated with propylene oxide and embedded in Araldite 502 epoxy resin. Ultrathin sections were stained with uranyl acetate and Reynold's lead citrate prior to examination using a Tecnai BioTWIN 12 transmission electron microscope (FEI, Hillsboro, OR, U.S.A.).

Microarray analysis of gene expression

For comparative gene-expression studies, RNA was extracted from TD and control FBs and LBs using the RNeasy kit from Qiagen. The mRNA fraction from approx. 5 μg of total RNA was amplified and labelled using the MessageAmp kit (Ambion). A 20 μg portion of biotinylated cRNA was hydrolysed randomly to 35–200 nucleotides in a fragmentation buffer solution (at 94 °C for 35 min). A 15 μg portion of the fragmented cRNA was then added to a hybridization buffer (100 mM Mes, 1 M Na+ (as NaCl), 20mM EDTA, 0.01% Tween-20, 0.1 mg/ml herring sperm DNA and 0.5 mg/ml acetylated BSA), containing known concentrations of positive control genes (50 pM Oligo B2, 1.5, 5, 25 and 100 pM Escherichia coli bioB, bioC, bioD and cre provided by Affymetrix, Santa Clara, CA, U.S.A.). The entire hybridization solution was heated at 99 °C for 5 min, equilibrated at 45 °C for 5 min, then centrifuged at 15000 g for 5 min before being injected into a U133A GeneChip (Affymetrix). After washing and staining, fluorescent images were scanned at 2 μm resolution using the Agilent G2500A Gene Array Scanner. The Microarray Analysis Suite, version 5.0 (Affymetrix), was employed to generate the comparative analysis presented in this study. Distinct algorithms of the software were used to determine the absolute call that distinguishes the presence (P) or absence of a transcript (A), the differential change in gene expression [increase (I), decrease (D), marginal increase (MI), marginal decrease (MD) and no change (NC)], and the magnitude (fold) change. Fold-change calculations were based on the average difference of each probe set, since this output is directly related to its expression level. After scanning the GeneChips, the Affymetrix software (MicroArray Suite 5.0) was used to calculate the intensity of the signal from each perfect match probe relative to the signal for the mismatch probe and to determine whether or not the gene was present in the sample [and a probability value (P<0.05) associated with this determination], as well as to measure the expression level of the gene. The overall chip intensities for each sample were scaled by linear adjustment to the same target value (1000). Pivot tables containing the scaled data from each experiment were generated and imported into GeneSpring (Silicon Genetics, Redwood City, CA, U.S.A.). In GeneSpring, the individual gene chip data were normalized by adjusting the median intensity of each array to a value of 1.0.

Flow cytometry of cell-surface antigen expression

TD and control FBs were freshly plated or 107 LBs/ml were resuspended in fresh medium 24 h before analysis. A total of 107 cells were stained in 200 μl of PBS with 1% (v/v) fetal-calf serum with primary antibodies specific for CD95 (ZB4 and IgG1; Upstate Biotechnology) or GGT (γ-glutamyl transferase; 3A8 and IgG2a [27]) for 30 min at 4 °C. IgG1 CD38 Ab was obtained from Caltag (Burlingame, CA, U.S.A.). ME0.5 IgG1 (Serotec, Oxford, U.K.) and OKT3 IgG2a antibodies (A.T.C.C., Manassas, VA, U.S.A.) were used as isotype controls. After washing them three times, cells were stained with phycoerythrin-conjugated goat anti-mouse or anti-human IgG for 30 min at 4 °C and analysed by flow cytometry.

Transduction of wild-type TAL cDNA by AAV (adeno-associated virus) vector (pAAV-IRES-hrGFP)

The full-length wild-type TAL cDNA [19] was inserted upstream of the IRES (internal ribosomal entry site) of the pAAV-IRES-hrGFP vector (Stratagene, La Jolla, CA, U.S.A.; where hrGFP refers to humanized Renilla green fluorescent protein). TAL-expressing AAV was produced by transfection of HEK-293 (human embryonic kidney 293) cells with TAL-containing pAAV-IRES-hrGFP, pAAV-RC (containing AAV replication and capsid genes) and pHelper plasmids, which supply the necessary exogenous gene products for virus production (Stratagene). At a 1:1 multiplicity of infection, >99% of cells infected with pAAV-TAL-IRES-hrGFP or pAAV-IRES-hrGFP control virus were GFP-positive. Maximal GFP expression was observed 24 h after virus infection. GFP and TAL expression were diminished by 50% after 48 h and became undetectable at 5 days post infection. For flow-cytometric analysis with FL-1 fluorochrome excited at 488 nm, changes elicited by the wild-type TAL-producing AVV were compared with those infected by AAV lacking TAL. In these viruses, GFP was replaced with the 792-bp neomycin phosphotransferase gene transferred from pUHD172.1neo [28]. Using Western-blot analysis, both TAL and NPT (neomycin phosphotransferase) were detected in pAAV-TAL-IRES-neo-transduced cells using rabbit Ab to NPT II (United States Biological, Swampscott, MA, U.S.A.; results not shown).


Results were analysed using the Student t test or Mann–Whitney rank sum test for non-parametric data. ANOVA was assessed with Graphpad (San Diego, CA, U.S.A.) software. Changes were considered significant at P<0.05.

Human experimentation

Our research was carried out in accordance with the Declaration of Helsinki (2000) of the World Medical Association and had the approval of the ethical committees of our institutions.


Effect of TAL deficiency on metabolites and enzyme activities of the PPP and interconnected pathways

TAL catalyses the transfer of a three-carbon fragment, corresponding to dihydroxyacetone, to GA3P (D-glyceraldehyde 3-phosphate) and E4P (D-erythrose 4-phosphate) [1]. In the forward reaction of the PPP, TAL transfers dihydroxyacetone from S7P to GA3P, thus producing E4P and F6P (fructose 6-phosphate). The latter is converted by glucose-phosphate isomerase into G6P, which, in turn, becomes a substrate of G6PD. In the reverse reaction, E4P and G6P are converted into GA3P and S7P. To assess the impact of TAL deficiency on the PPP, unique substrates of the pathway were analysed by HPLC. As shown in Figure 1(A) and Supplementary Figure S1 at, S7P was markedly accumulated both in TD FBs (P=0.012) and LBs (P=0.0049). By contrast, G6P levels were reduced in TD FBs (0.062±0.017 pmol/μg of protein) with respect to control FBs #1 (0.246±0.044 pmol/μg of protein; P=0.0132) or control FBs #2 (0.279±0.066 pmol/μg of protein; P=0.047). G6P levels were also diminished in TD LBs relative to the control LBs #1 (P=0.034) or control LBs #2 (P=0.039) (Figure 1A).

Figure 1 HPLC analysis of (A) G6P and S7P levels in FBs and LBs of a TD patient (TDP) and two healthy controls and (B) nucleotides in LBs of a TD patient (TDP LB) and control LB JAB and GM13416

Results are means±S.E.M. for five or more independent experiments. *P<0.05; **P<0.01. (B) Bar charts show nucleotide levels in TDP and control LBs JAB and GM13416. Results are means±S.E.M. of four measurements. *P<0.05; **P<0.01.

Accumulation of S7P and depletion of G6P indicate that TAL deficiency blocked the forward reaction, i.e. recycling of G6P for the oxidative branch of the PPP. The oxidative PPP generates NADPH [1]. NADPH, as well as NADP+, AMP, cAMP and NAD+ were diminished, whereas ADP-ribose levels were elevated in TD cells (Figure 1B and Supplementary Figure S2 at Following the alkaline extraction required to preserve the integrity of NADPH and NADH [14], ATP was degraded. Therefore, ATP was measured by the luciferase assay [26]. The ATP content of TD LBs was increased (2.8±0.3 pmol/μg of protein) in comparison with JAB (0.9±0.2 pmol/μg of protein; P=0.0002) and GM13416 control LBs (1.5±0.4 pmol/μg of protein; P=0.004).

Profound depletion of G6P and NADPH suggested diminished activity of the oxidative PPP. As previously reported [12], TAL protein and enzyme activity were undetectable in LB and FB of the TALΔS171 patient (Figure 2A). To assess the impact of TAL deficiency on the oxidative PPP, expression and activity of G6PD, the rate-limiting enzyme of NADPH production, was measured. Both G6PD protein (Figure 2A) and enzyme activity levels were markedly diminished in LBs and FBs of the TD patient relative to control cells (Figures 2B and 2C). Although activities of TK and GPX were not significantly affected, catalase and SOD were reduced in TD LBs and FBs (Figures 2B and 2C).

Figure 2 Effect of TAL deficiency on the enzymatic activities of PPP and connected metabolic pathways

(A) Western-blot analysis of TAL, G6PD and β-actin levels in cytosolic protein lysates (20 μg/lane) of FBs and LBs from TD patient and controls. G6PD/β-actin levels relative to TD cells, normalized at 1, were determined by densitometry as described in [12]. (B) Activities of PPP enzymes TAL, TK and G6PD and antioxidant enzymes GPX, CAT and SOD in TD and control LBs 1 (GM13416) and 2 (JAB). (C) Activities of PPP enzymes TAL and G6PD and antioxidant enzymes CAT and SOD in TD FBs and control FBs 1 and 2. Results are means±S.E.M. for four or more independent experiments.

Loss of Δψm and increased mitochondrial biogenesis characterize mitochondrial dysfunction in TD cells

The oxidation–reduction equilibrium of pyridine nucleotides (NADH/NAD++NADPH/NADP+) regulates Δψm [29]. Since TAL deficiency may affect cell viability through controlling Δψm [14,15,30], annexin V-negative live cells were assessed for Δψm using potentiometric fluorescent dyes TMRM and JC-1 (Supplementary Figure S3 at and Table 1). A significant decrease in Δψm was detected by both TMRM and JC-1 fluorescence. As altered incorporation of potentiometric dyes may represent changes in mitochondrial mass, the latter was assessed by staining with the potential-insensitive mitochondrial probes NAO and MTG. Surprisingly, the mitochondrial mass in TD cells was increased, as determined by both NAO (+104%, P=0.01) and MTG (+39%, P=0.0003) fluorescence. Along the same lines, electron microscopy showed increased numbers of mitochondria in TD LBs (29.2±5.2/cell) with respect to control LBs (+10.2±2.2/cell, P=0.0107). Mitochondria also appeared larger in TD cells (Figure 3), exhibiting features of megamitochondria [31]. Whereas intracytosolic and mitochondrial ROI production, monitored by HE (−15.0%, P=0.02) and DHR (−23.0%, P<0.01) fluorescence were reduced, DCF fluorescence, preferentially detecting H2O2 [16], was enhanced in TD LBs (+23.8%, P=0.0013). TD LBs also produced increased amounts of NO, as monitored by DAF-FM fluorescence (+61.3%, P=0.0081; Figure 3), which is a key trigger of mitochondrial biogenesis [32]. Since mitochondria store Ca2+ [18], we investigated the [Ca2+]c and [Ca2+]m levels. [Ca2+]c and [Ca2+]m of TD LBs were elevated with respect to control LB, as determined by Fluo-3 (+53.0%, P=0.023) and Rhod-2 fluorescence (+29.0%, P=0.001) respectively (Table 1 and Supplementary Figure S3).

View this table:
Table 1 Cumulative analysis of the Δψm by TMRM and JC-1 fluorescence, mitochondrial mass by NAO and MTG fluorescence, cytoplasmic and mitochondrial ROI production by HE and DHR fluorescence, H2O2 levels by DCF fluorescence, cytoplasmic and mitochondrial Ca2+ levels by Fluo-3 and Rhod-2 fluorescence, and NO production by DAF-FM fluorescence in annexin V-negative cells using flow cytometry

A total of 5×106 LBs of a TD patient (TD LB) and normal donor cells (Normal LBs) were cultured with or without 50 μM H2O2 for 24 h. Results are expressed as relative fluorescence values with respect to those of unstimulated normal cells adjusted to 1.0 for each experiment. Results are means±S.E.M. for eight independent experiments. *P<0.05 reflects comparison of untreated TD LBs with untreated normal LBs, whereas #P<0.05 reflects a comparison of H2O2-treated TD LBs with H2O2-treated normal LBs for each parameter.

Figure 3 Electron microscopy of control (A and B) and TD LBs (C and D)

The arrow indicates megamitochondrion in (D). Note: 1 micron=1 μm.

Mitochondrial function was further tested by exposure of TAL-deficient and control LBs to a low dose (50 μM) of H2O2. Although H2O2 is freely diffusible, it is not, in itself, an ROI. Induction of apoptosis by H2O2, requires mitochondrial transformation into an ROI, e.g. OH, through the Fenton reaction [33]. After exposure to H2O2, the potentiometric dyes showed mitochondrial hyperpolarization in accordance with previous findings [26,34]. Mitochondrial hyperpolarization is dependent on NO production [17]. TD cells exhibited a blunted response in H2O2-induced NO production and mitochondrial hyperpolarization (Supplementary Figure S3 and Table 1). Maintenance of GSH in a reduced state is dependent on the availability of NADPH [1]. Although NADPH was profoundly depleted, GSH levels were normal in TD cells, suggesting compensatory changes in GSH metabolism.

TAL deficiency elicits co-ordinated changes in expression of genes involved in the PPP, mitochondrial biogenesis, GSH metabolism and Ca2+ fluxing

The impact of TAL deficiency on the PPP and connected metabolic pathways was investigated in the context of global gene expression using the Human Genome U133A chip with probe sets of 22283 genes. The specificity of TAL-deficiency-induced changes on gene expression was validated by analysis of TD LBs transduced with an AAV vector expressing wild-type TAL. Expression of 18 genes was significantly and specifically altered by TAL deficiency, i.e., normalization of TAL expression by AAV-mediated gene transfer reversed the changes in expression of each of these genes (Figure 4). Reduced TAL transcription was detected in both TD FBs and LBs in comparison with control FBs and LBs, in agreement with earlier Northern-blot studies [12]. G6PD mRNA levels were reduced in a manner that correlated with diminished G6PD protein levels (Figure 2A). G6PD, TK, 6PGD (6-phosphogluconate dehydrogenase), PGI (phosphoglucose isomerase), PFKM (phosphofructokinase), TPI1 (triosephosphate isomerase 1), ALDC (aldolase C) and PGK1 (phosphoglycerate kinase 1) mRNA levels were all reduced in TD LBs and FBs, following the pattern of TAL expression. On the basis of ANOVA, expression of these enzymes was co-ordinately regulated in TD LBs and FBs (P=0.005). Interestingly, genes regulating mitochondrial biogenesis [PGC-1 (peroxisome proliferator-activated receptor γ-co-activator 1)', NRF-1 (nuclear respiratory factor 1), HMOX1 (haem oxygenase), NUDT1 (nucleotide diphosphatase type motif 1/8-oxo-7,8-dihydroguanosine triphosphatase), PTK2B (protein tyrosine kinase 2, Ca2+-regulated) and BNIP3L (bcl-2-like interacting protein 3-like mitochondrial protein)] were also co-ordinately altered in TD FBs and LBs. Expression of the Fas/Apo-1/CD95 cell death receptor was only increased in TD FBs but not in LBs (results not shown). Flow-cytometric staining with monoclonal Ab ZB4 confirmed increased surface expression of the Fas/Apo-1/CD95 on FBs but not on LBs of the TD patient (Figure 5A). Increased surface expression levels of GGT (Figure 5B) and CD38/NAD+ hydrolase were also confirmed on TD FBs and LBs by flow cytometry (Figure 5C). Overexpression of GGT may help maintain GSH levels in cells with diminished NADPH production. Indeed, incubation with GGT Ab for 48 h reduced GSH content of TAL-deficient LB measured by MCB fluorescence (results not shown).

Figure 4 Relative mRNA levels of genes significantly altered in TD LBs (P<0.05 using GeneSpring) and reversed by correction of TAL activity based on microarray analysis

Gene expression was investigated in control and TD FBs and LBs as well as TD LBs infected with control AAV (TD LB-pAAV) and TAL-producing AAV (TD-LB-pAAV/TAL). Abbreviation: CLCA2, Ca2+-regulated chloride channel 2.

Figure 5 Flow-cytometric analyses

(A) Flow cytometry of cell-surface expression of the Fas/Apo-1/CD95 antigen on TD (grey areas) and control fibroblasts (control FBs 1 and 2; white areas) and LBs (control LBs 1 and 2; white areas). (B) Flow cytometry of GGT expression on TD (grey areas) and control cells (control FBs 1 and 2 and control LBs 1 and 2; white areas). (C) Flow cytometry of CD38 expression on TD (grey areas) and control cells (control FBs 1 and 2 and control LBs 1 and 2; white areas). The value mean channel fluorescence of TD cells is shown over each set of curves and the mean channel fluorescence of control cells is shown in parenthesis (white area overlays).

Expression of wild-type TAL reverses changes in ATP, Ca2+ fluxing and apoptosis susceptibility of TD cells

The phenotype of homozygous TALΔS171/TD patient's cells may be influenced by mutations in genes other than TAL. Therefore, we normalized TAL activity by transduction of wild-type TAL cDNA using an AAV vector. At 1:1 multiplicity of infection, >99% of AAV-infected cells were GFP-positive (Figure 6A) and expressed functionally active TAL (Figures 6B and 6C). After AAV infection, TAL expression (Figure 6B) and activity levels reached ∼100% of those in control LB (Figure 6C). Normalization of TAL protein expression and activity resulted in reduced ATP (Figure 6D), increased GSH (Figure 6E) and reduced [Ca2+]c in TD LBs (Figure 6F).

Figure 6 Effect of normalization of TAL protein and activity levels on ATP, Ca2+, GSH and apoptosis-susceptibility of TD LBs

(A) Flow-cytometric analyses of uninfected control TD LBs (TDLB, black), TDLB infected with TAL-expressing AAV (TDLB-pAAV/TAL, first grey curve; produces TAL and GFP) and TD LBs infected with GFP-expressing AAV (TD LB-pAAV, second grey curve; produces GFP only). (B) Western-blot analysis of TAL and β-actin levels in TD LBs, TD LBs infected with control AAV (TDLB-pAAV), and TD LBs infected with TAL-producing AAV (TDLB-pAAV/TAL) and control GM13416 and JAB LBs. (C) TAL enzyme activity levels in cells analysed in (B). (D) Measurement of ATP levels by the luciferase assay. (E) Assessment of intracellular Ca2+ by Fluo-3 relative fluorescence (RF). (F) Assessment of intracellular GSH by MCB relative fluorescence. (G) Apoptosis induced by 100 μM H2O2 12 h after AAV infection and measured by annexin-V/propidium iodide staining 24 h later. (H) Apoptosis induced by CD20 Ab 6 h after AAV infection and measured by annexin-V/PI staining 48 h later. Results are means±S.E.M. for four experiments. (I) Apoptosis induced by CH11 antibody to CD95 12 h after AAV infection and measured by annexin-V/PI staining 24 h later.

B-cell homoeostasis is maintained through H2O2- [35], Fas/CD95- [36] or CD20-mediated apoptosis [37]. To investigate the impact of TAL deficiency on the processing of these cell-death signals, cells were exposed to 100 μM H2O2 or Ab to Fas/CD95 or CD20 12 h after AAV-mediated transduction of wild-type TAL and survival was measured by annexin-V staining 24 h later. H2O2-induced apoptosis was enhanced in TD LBs as compared with control LBs, a situation that was reversed after infection with wild-type TAL-producing AAV (Figure 6G). Although microarray analysis and cell-surface staining showed similar CD20 expression levels (results not shown), cross-linking of CD20 resulted in accelerated apoptosis of TD LBs (Figure 6H). Enhanced CD20-induced cell death was reversed by expression of wild-type TAL (Figure 6H). In accordance with its dependence on Ca2+ fluxing [37], increased CD20-induced apoptosis of TD cells was also reversed by pretreatment with 5 μM BAPTA-AM for 30 min (results not shown). In contrast, Fas-induced apoptosis was reduced in TD LB; which was also reversed by infection with TAL-producing AAV (Figure 6I). These results reveal that the apoptosis-susceptibility of B-cells is regulated by TAL in a signal-dependent manner.


TAL is an enzyme of the non-oxidative branch of the PPP which is the principal supplier (1) of NADPH for reductive biosynthesis and regeneration of glutathione from its oxidized form, and (2) of ribose 5-phosphate for nucleotide production. G6PD has been uniformly regarded as the rate-controlling enzyme for both of these functions of the PPP [1]. TAL deficiency has been found to be associated with liver cirrhosis in children from three families [911]. Utilizing LB and FB cells available from the first patient [12], the present study provides evidence that deficiency of TAL influences apoptosis signal processing in a death-pathway-specific manner. TD LBs exhibit increased spontaneous and H2O2-induced apoptosis. Fas-induced apoptosis was reduced in LBs of the TD patient. Fas-induced caspase activity is dependent on maintaining of active-site cysteine residues in a reduced state [38]. Co-ordinate up-regulation of GSH and Fas-dependent apoptosis by normalization of TAL expression suggests that oxidative stress limits activity of caspases in TD LBs. NADPH depletion may underlie oxidative stress. Increased H2O2 levels in TD cells is also consistent with deficiency of NADPH, which is required for conversion of H2O2 into water by catalase [21].

The mechanism of oxidative stress in TD cells was investigated at the level of PPP enzymes and metabolites. S7P, a substrate of the forward TAL reaction, was accumulated, whereas G6P was depleted. Accumulation of S7P is compatible with a failure to recycle R5P into G6P through the non-oxidative branch of the PPP, thus reducing NADPH production by the oxidative branch. ADP-ribose and ATP levels were clearly elevated, indicating that nucleotide synthesis was not limited, but rather stimulated, by the availability of R5P. The five-carbon sugars R5P, xylulose 5-phosphate and ribulose-5-phosphate are likely to accumulate and inhibit 6PGD [39]. Co-ordinate accumulation of ADP-ribose and depletion of NAD+ may be attributed to overexpression of CD38 [40,41]. CD38 or NAD+ hydrolase metabolizes NAD+ into ADP-ribose and cADP-ribose, which, in turn, stimulate the release of Ca2+ from the endoplasmic reticulum [40]. Normalization of TAL activity reduced CD38 expression and [Ca2+]c, observations that are consistent with the notion that oxidative stress is responsible for overexpression of CD38 and the resultant NAD+ depletion, ADP-ribose accumulation and increased [Ca2+]c. Moreover, mitochondria constitute major Ca2+ stores [42]; thus enhanced mitochondrial mass also favours higher [Ca2+]m levels. Increased mitochondrial mass allows for increased production of ATP. In turn, excess ATP could also inhibit G6PD and 6PGD [43].

Expression of genes associated with the PPP and mitochondrial biogenesis were co-ordinately regulated with TAL deficiency and normalized by AAV-mediated transduction of the wild-type TAL gene. Overexpression of Fas/CD95, both at the mRNA level and on the surface of TD FBs, has been associated with increased Fas-dependent apoptosis of these cells (results not shown). Fas expression was not enhanced, and Fas-dependent apoptosis was diminished, in TD LBs. On the basis of microarray and confirmatory flow-cytometric studies, CD38 was seen to be overexpressed on both TD LBs and FBs. The promoter region of CD38 harbours several redox-sensitive AP-1 (activator protein 1) motifs [44]. Thus oxidative stress may account for increased expression of CD38 in TD B-cells. Normalization of TAL activity reversed the elevated CD38 expression and Ca2+ levels, the increased H2O2- and CD20-induced apoptosis and the diminished Fas-induced apoptosis of B cells, clearly indicating that each of these changes resulted from TAL deficiency. Increased CD20-induced apoptosis was also reversed by the Ca2+ chelator BAPTA-AM, suggesting that the enhancement of this cell-death pathway is mediated by high Ca2+ levels. Although the role of B-cell apoptosis in human TAL deficiency is presently unclear, enhanced spontaneous and H2O2-induced apoptosis and inhibited CD95/Fas apoptosis could contribute to the pathogenesis of liver disease. Clinical manifestations of TAL deficiency are dominated by the consequences of liver cirrhosis [911], which results from increased cell death of hepatocytes and biliary epithelial cells [8,13]. Increased expression of CD38 [45], elevation of [Ca2+]c and [Ca2+]m, as well as depletion of NAD+ and NADPH, have been implicated in mitochondrial dysfunction, oxidative stress [46] and cirrhosis of the liver [47]. There is growing evidence for the involvement of the CD95/Fas cell-death-receptor-initiated apoptosis pathway in physiological regulation of hepatocyte turnover [48], and defective Fas signalling predisposes to hepatocarcinogenesis [49]. As oxidative stress leads to cirrhosis [47], whereas CD95/Fas resistance facilitates hepatocarcinogenesis [49], the present study identifying TAL as a signal-specific regulator of apoptosis is likely to be relevant for the pathogenesis of TAL deficiency.


We thank Dr David Fernandez (Department of Medicine, SUNY Upstate Medical University, Syracuse, NY 13210, U.S.A.) for helpful discussions and Dr Paul Phillips (Department of Medicine, SUNY Upstate Medical University, Syracuse, NY 13210, U.S.A.) for continued encouragement and support. This work was supported in part by grant DK 49221 from the National Institutes of Health, the Central New York Community Foundation and the Children's Miracle Network.

Abbreviations: AAV, adeno-associated virus; Ab, antibody; ALDC, aldolase C; annexin V-FITC, fluorescein-conjugated annexin V; annexin V-PE, phycoerythrin-conjugated annexin V; 2-APB, 2-aminoethoxydiphenyl borane; BAPTA-AM, 1,2-bis-(o-aminophenoxy)ethane-N,N,N′,N′-tetra-acetic acid tetrakis(acetoxymethyl ester); BNIP3L, bcl-2-like interacting protein 3-like mitochondrial protein; [Ca2+]c, cytoplasmic Ca2+ concentration; [Ca2+]m, mitochondrial Ca2+ concentration; C-PTIO, carboxy-2-phenyl-4,4,5,5-tetramethyl-imidazoline-1-oxyl-3-oxide; DAF-FM, 4-amino-5-methylamino-2′,7′-difluorofluorescein diacetate; DCF, 5,6-carboxy-2′,7′-dichlorofluorescein; DCFH-DA, 5,6-carboxy-2′,7′-dichlorofluorescein diacetate; DHR, dihydrorhodamine 123; DiOC6, 3,3′-dihexyloxacarbocyanine iodide; EBV, Epstein–Barr virus; E4P, D-erythrose 4-phosphate; FBS, fetal bovine serum; F6P, fructose 6-phosphate; FB, fibroblast; GGT, γ-glutamyltransferase; G6P, glucose 6-phosphate; G6PD, G6P dehydrogenase; GA3P, D-glyceraldehyde 3-phosphate; GLH, gluconolactone hydrolase; GPX, glutathione peroxidase; HE, hydroethidine; HMOX1, haem oxygenase; (hr)GFP, (humanized Renilla) green fluorescent protein; IP3R, Ins(1,4,5)P3 receptor; IRES, internal ribosomal entry site; JC-1, 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolocarbocyanine iodide; LB, lymphoblast; MCB, monochlorobimane; mClCCP, carbonyl cyanide m-chlorophenylhydrazone; MTG, MitoTracker Green-FM; NAO, nonyl Acridine Orange; NPT, neomycin phosphotransferase; NRF-1, nuclear respiratory factor 1; NUDT1, nucleotide diphosphatase type motif 1/8-oxo-7,8-dihydroguanosine triphosphatase; PCD, programmed cell death; PFKM, phosphofructokinase; 6PG, 6-phosphogluconate; PGC-1, peroxisome proliferator-activated receptor γ-co-activator 1; 6PGD, 6-phosphogluconate dehydrogenase; PGK1, phosphoglycerate kinase 1; PGI, phosphoglucose isomerase; PPP, pentose phosphate pathway; PTK2B, protein tyrosine kinase 2; Ca2+-regulated, R5P, ribose 5-phosphate; ROI, reactive oxygen intermediates; SOD, superoxide dismutase; S7P, sedoheptulose 7-phosphate; TAL, transaldolase; TD, TAL-deficient; TK, transketolase; TMRM, tetramethylrhodamine methyl ester; Δψm, mitochondrial transmembrane potential


View Abstract