3-Hydroxyglutaric Acid Is Transported Via the Sodium-Dependent Dicarboxylate Transporter Nadc3

J Mol Med (2007) 85:763–770 DOI 10.1007/s00109-007-0174-5

RAPID COMMUNICATION

3-Hydroxyglutaric acid is transported via the sodium-dependent dicarboxylate transporter NaDC3

Franziska Stellmer & Britta Keyser & Birgitta C. Burckhardt & Hermann Koepsell & Thomas Streichert & Markus Glatzel & Sabrina Jabs & Joachim Thiem & Wilhelm Herdering & David M. Koeller & Stephen I. Goodman & Zoltan Lukacs & Kurt Ullrich & Gerhard Burckhardt & Thomas Braulke & Chris Mühlhausen

Received: 20 October 2006 /Revised: 16 January 2007 /Accepted: 8 February 2007 / Published online: 14 March 2007 # Springer-Verlag 2007

Abstract Patients with glutaryl-CoA dehydrogenase (GCDH) deficiency accumulate glutaric acid (GA) and 3- hydroxyglutaric acid (3OH-GA) in their blood and urine. To identify the transporter mediating the translocation of 3OH-

Electronic supplementary material The online version of this article (doi:10.1007/s00109-007-0174-5) contains supplementary material, which is available to authorized users. F. Stellmer : B. Keyser : S. Jabs : Z. Lukacs : K. Ullrich : T. Braulke : C. Mühlhausen (*) Department of Pediatrics, University Medical Center Hamburg-Eppendorf, Martinistrasse 52, 20246 Hamburg, Germany FRANZISKA STELLMER CHRIS MÜHLHAUSEN e-mail: [email protected] just completed her MD thesis in received his MD from the Georg- the Molecular Biology Lab of August-University of Göttingen B. C. Burckhardt : G. Burckhardt the Department of Pediatrics, after having performed a thesis Center of Physiology and Pathophysiology, University Medical Center about transport of lysosomal Georg-August-University, Hamburg-Eppendorf. She will proteins in the Department of Humboldtallee 23, qualify as a physician shortly. Biochemistry with Prof. K. von 37073 Göttingen, Germany Her research projects include Figura. He is presently a spe- renal abnormalities in a mouse cialist pediatrician and research H. Koepsell model of glutaryl-CoA dehy- project leader at the Department Institute of Anatomy and Cell Biology, University of Würzburg, drogenase deficiency, especial- of Pediatrics, University Medical Koellikerstrasse 6, ly renal transport mechanisms. Center Hamburg-Eppendorf. His 97070 Würzburg, Germany research interests include inborn metabolic diseases, especially T. Streichert : Z. Lukacs glutaryl-CoA dehydrogenase de- Department of Clinical Chemistry, University Medical Center ficiency as well as lysosomal Hamburg-Eppendorf, storage diseases. Martinistrasse 52, 20246 Hamburg, Germany GA through membranes, kidney tissue of Gcdh−/− mice have M. Glatzel been investigated because of its central role in urinary Department of Neuropathology, excretion of this metabolite. Using microarray analyses of University Medical Center Hamburg-Eppendorf, − − Martinistrasse 52, kidney-expressed genes in Gcdh / mice, several differen- 20246 Hamburg, Germany tially expressed genes encoding transporter proteins were 764 J Mol Med (2007) 85:763–770 identified. Real-time polymerase chain reaction analysis tissues and body fluids is accompanied by irreversible confirmed the upregulation of the sodium-dependent dicar- destruction of striatal neurons with subsequent development boxylate cotransporter 3 (NaDC3) and the organic cation of an irreversible disabling movement disorder [1]. transporter 2 (OCT2). Expression analysis of NaDC3 in Whereas the pathomechanisms in GA1 causing striatal Xenopus laevis oocytes by the two-electrode-voltage-clamp dysfunction and degeneration are not fully understood, both technique demonstrated the sodium-dependent translocation N-methyl-D-aspartate receptor-mediated neurotoxicity and of 3OH-GA with a KM value of 0.95 mM. Furthermore, disintegration of endothelial barriers by 3OH-GA are tracer flux measurements in Chinese hamster ovary cells discussed [2–5]. However, cellular mechanisms facilitating overexpressing OCT2 showed that 3OH-GA inhibited the transport of 3OH-GA through membranes have not significantly the uptake of methyl-4-phenylpyridinium, been investigated yet. whereas 3OH-GA is not transported by OCT2. The data A mouse model of GA1 was generated by a targeted demonstrate for the first time the membrane translocation of disruption of the Gcdh gene with a biochemical phenotype 3OH-GA mediated by NaDC3 and the cis-inhibitory effect similar to human GA1 patients, including high excretion on OCT2-mediated transport of cations. rates of GA and 3OH-GA into the urine [6]. Additionally, diffuse spongiform myelinopathy and an unexplained Keywords Glutaric aciduria type 1 . Slc22a2 . OCT2 . hypertrophy of kidneys were seen in Gcdh−/− mice. Glutaryl-CoA dehydrogenase deficiency. Slc13a3 . NaDC3 To examine transporters involved in the translocation of 3OH-GA via membranes, we performed oligonucleotide Abbreviations microarray analyses in kidneys of Gcdh-deficient mice and GA1 glutaric aciduria type 1 found several solute carrier protein (Slc) genes differential- GA glutaric acid ly regulated. Confirmed by quantitative real-time polymer- Gcdh glutaryl-CoA dehydrogenase ase chain reaction (qRT-PCR), immunohistochemistry, NaDC3 sodium-dependent dicarboxylate expression analyses in Xenopus laevis oocytes and two- cotransporter 3 electrode-voltage-clamp measurements, the sodium-depen- 3OH- 3-hydroxyglutaric acid dent dicarboxylate cotransporter 3 (NaDC3, Slc13a3) was GA identified as a potential transporter mediating the low OCT2 organic cation transporter 2 affinity translocation of 3OH-GA at the basolateral mem- Slc solute carrier brane of kidney proximal tubule cells.

Materials and methods Introduction Materials 1-[3H]-Methyl-4-phenylpyridinium (MPP) and Glutaric aciduria type 1 (GA1) is caused by the deficiency [3H]-sodium borohydride (NaBH ) were purchased from of the mitochondrial matrix protein glutaryl-CoA dehydro- 4 Biotrend (Köln, Germany) and Amersham Pharmacia genase (GCDH). The enzyme catalyzes the oxidative (Freiburg, Germany) respectively. Dimethyl-3-ketoglutaric decarboxylation of glutaryl-CoA to crotonyl-CoA in the and GA acids and diaminobenzidine were obtained from catabolic pathway of lysine, hydroxylysine, and tryptophan. Sigma Chemical (St. Louis, MO). Silica gel plates (HPTLC During catabolic crises, the accumulation of large amounts F254) and silica gel (0.015–0.063 mesh) were from Merck of glutaric (GA) and 3-hydroxyglutaric (3OH-GA) acids in (Darmstadt, Germany). 3OH-GA was synthesized as 3 : reported previously [4]. The synthesis of [ H]-labeled J. Thiem W. Herdering 3OH-GA is described in the supplementary methods online. Institute of Organic Chemistry, University of Hamburg, The human (h) NaDC3 complementary deoxyribonucleic Martin-Luther-King-Platz 6, 20146 Hamburg, Germany acid (cDNA) was kindly provided by Dr. V. Ganapathy (Medical College of Georgia, Augusta, GA). All other D. M. Koeller chemicals were of analytical grade or higher. Department of Pediatrics, Oregon Health and Science University, 707 SW Gaines Road, Portland, OR 97239, USA Antibodies Polyclonal anti-OCT2 and anti-NaDC3 antibodies were from Alpha Diagnostic International (San S. I. Goodman Antonio, TX), anti-actin from Santa Cruz (Heidelberg, Department of Pediatrics, Germany), and horseradish peroxidase (HRP)- and alkaline- University of Colorado Health Sciences Center, P.O. Box C233, 4200 East Ninth Avenue, phosphatase-conjugated anti-rabbit and anti-guinea-pig Denver, CO 80262, USA immunoglobulin G (IgG), respectively, from Jackson J Mol Med (2007) 85:763–770 765

Immunoresearch (West Grove, PA). The polyclonal guinea- cRNA synthesis and oocyte injection Stage V–VI oocytes pig antibody against the basolateral K+/Cl− (K-Cl) cotrans- from X. laevis (Nasco, Fort Atkinson, WI) were defollicu- porter Kcc4 (Slc12a7) [7] was kindly provided by Dr. C. lated by an overnight incubation at 18°C with collagenase Hübner (University Medical Center Hamburg-Eppendorf). (Type CLSII, Biochrom KG, Germany; 0.5 mg/ml) in oocyte Ringer’s solution (ORI) containing 110 mM NaCl,

Mice-Gcdh −/− mice and wild-type littermate controls were 3 mM KCl, 2 mM CaCl2, 5 mM HEPES/Tris, pH 7.5. One generated from heterozygotes [6]. The genetic background day after removal from the ovaries, oocytes were injected of all mice groups used in this study was C57BL6/SJ129 with 23 nl of water or 23 ng hNaDC3-cRNA in an hybrid. The genotypes were confirmed by polymerase chain equivalent volume, synthesized from NotI-linearized plas- reaction (PCR) and measurements of glutarylcarnitine mid (mMessage mMachine-T7 in vitro transcription kit; concentration in dried blood spots. The mice were housed Ambion, Austin, TX). After injection, the oocytes were in an animal facility of the University Hospital with a incubated for 3 days at 18°C in ORI, supplemented with 12-h light–dark cycle and allowed water and food ad 50 μg/ml gentamycin and 2.5 mM sodium pyruvate. libitum. Animal care was provided in accordance with institutional guidelines. Anesthesized mice were either used Electrophysiological studies Current recordings were per- for preparation of kidneys or perfused with phosphate- formed in ORI using the two-electrode-voltage-clamp buffered saline, pH 7.4, containing 50 U/100 ml heparin technique with a commercial amplifier (OC725, Warner, followed by cryosectioning. Hambden, CT). Borosilicate glass microelectrodes were filled with 3 M KCl and had resistances of ∼1MΩ. The Target labeling and microarray hybridization Procedures resting membrane potential of the oocytes ranged between for cDNA synthesis, labeling, and hybridization were −22 and −36 mV, and the holding currents to achieve a carried out according to the manufacturer’sprotocol potential of −60 mV were in the range of −10 to −40 nA. (Affymetrix) and are described in the supplementary methods online. Immunohistochemistry and Western blotting Cryosections from mouse kidneys were cut to a thickness of 7 μm and RNA extraction, cDNA preparation, and real-time PCR To- double-immunostained with anti-NaDC3 antiserum (1:50) tal ribonucleic acid (RNA) was prepared from mice kidney or anti-Kcc4 antiserum (1:200). Visualization was achieved as described previously [8]. RNA was reverse transcribed using HRP-coupled anti-rabbit IgG (NaDC3) and alkaline into cDNA using Superscript III Reverse Transcriptase phosphatase-coupled anti-guinea-pig IgG (Kcc4), respec- (Invitrogen, Carlsbad, CA). qRT-PCR was performed using tively, and diaminobenzidine and fast red as chromogens. 1.5 μl cDNA and 13.5 μl of a LightCycler FastStart DNA Alternatively, the fast red staining was visualized by Master SYBR Green I mix (Roche Diagnostics, India- fluorescence microscopy using an Olympus BH2 micro- napolis, IN) containing 0.3 μM gene-specific PCR primers. scope (Hamburg, Germany). Extracts (50 μg of protein) of

Sequences of primers and Tm for each gene are shown in 100,000×g membrane fractions of mice kidney were used the supplementary methods online Table I. Analyses were for Western blotting with anti-OCT2 antibody (1:1,000) or performed in a LightCycler instrument (Roche) according anti-actin (1:250) and visualized with HRP-conjugated anti- to the manufacturer’s instructions. For organic anion-trans- rabbit IgG (1:10,000) and enhanced chemiluminescence porting polypeptide 1a1 (Slc21a1), 1a6 (Slc21a13), 7 reagents (Pierce Biotechnology, Rockford, IL). (Slc22a7), urea transporter 2 (Slc14a2), and phosphoenol- pyruvate carboxykinase 1 (Pck1) murine TaqMan MGB Data analysis Microarray data analyses were carried out as probes, 6-carboxy-fluorescein dye-labeled (Applied Bio- described in the supplementary methods online. All other systems) was used in 96 well optical reaction plates, and data were analyzed using one-way analysis of variance triplicates were quantified in an ABI Prism 7000 sequence followed by Scheffe’s test. Significance was accepted at detector. Primer sequences and TaqMan assay ID numbers p<0.05. SPSS 12.0 software (SPSS, Chicago, IL) was used are listed in the supplementary methods online Table I. The for calculations. relative level of each mRNA was determined using the comparative CT method [9].

Transport experiments The transport of [3H]-MPP and Results [3H]-3OH-GA in the presence or absence of 2 mM 3OH- GA or 2 mM MPP was measured in Chinese hamster ovary To identify potential transporters involved in the transmem- (CHO) cells stably expressing hOCT2 as described previ- brane transport of 3OH-GA, kidney tissue was used for ously [10]. oligonucleotide microarray analysis because of its central 766 J Mol Med (2007) 85:763–770 role in secretion of 3OH-GA into urine. The microarray results). Western blot analyses using extracts from mem- analyses covering a total of 14,000 mouse genes revealed brane fractions of kidneys prepared at 42 days of age that in the kidney of 42-day-old male Gcdh−/− mice, 61 revealed an elevated level of OCT2 protein in Gcdh−/− genes were upregulated, whereas 55 were downregulated. mice (Fig. 1b). To examine whether OCT2 may be affected These 116 genes were categorized in groups of encoding by 3OH-GA, uptake experiments were performed in a proteins playing roles in signal transduction, energy and heterologous expression system. First, the uptake of [3H]- metabolism, transcription and cell cycle control, or function MPP was measured in the absence and presence of 3OH- in transport and cell structure as molecular chaperones or GA in stably hOCT2-expressing CHO cells. The uptake of encoding Slc proteins (supplementary Table II online). The MPP in these cells is OCT2- and potential-dependent, with expression profile of upregulated genes was dominated by an apparent KM value of 16 μM[13]. The presence of genes involved in metabolism and those belonging to the 2 mM 3OH-GA inhibited significantly the OCT2-mediated Slc transporter group. influx of 0.2 μM[3H]-MPP by 25–30% (Fig. 2a). The Because of the variability inherited in microarray observed partial inhibition is consistent with the finding technology, qRT-PCR analysis was applied to validate the that polyspecific cation transporters have a complex expression of a subset of dysregulated carriers suspected as binding pocket containing several overlapping interaction potential candidates directly or indirectly involved in the domains for different substrates [14]. As a control, the transport of GA and 3OH-GA. qRT-PCR data for organic addition of 2 mM MPP inhibited the uptake of [3H]-MPP anion transporter 2 (Slc22a7), NaDC3 (Slc13a3), the completely (Fig. 2a). Using 10 μM[3H]-MPP, no inhibition organic cation transporter 2 (OCT2, Slc22a2), and the urea of MPP uptake by 2 mM 3OH-GA was detected (Fig. 2b). transporter 2 (Slc14a2) estimated each in six Gcdh−/− and control mice were in good agreement with the microarray results and show a relative expression of 1.6±0.4, 3.8±2.2, 1.7±0.3, and 0.1±0.06, respectively (Figs. 1a and 3a). In addition, the relative mRNA expression of cytosolic Pck1 was significantly increased 3.3±0.04 fold in Gcdh−/− mice, confirming the microarray data. Two genes, Slc21a1 (organic anion-transporting polypeptide 1a1) and Slc21a13 (organic anion-transporting polypeptide 1a6) were found to be increased only 1.3- and 1.2-fold in their expression in Gcdh−/− mice. The discrepancy, in particular for the expression of Slc21a1, may be the result of an unexplained high variance between different animals. NaDC3 and the polyspecific OCT2 were examined in more detail because of their high expression in the brain and kidney [11]. In addition, NaDC3 transports dicarboxylated substrates such as α-ketoglutarate and GA [12]. Furthermore, OCT2 was also found to be upregulated in brain microarray analysis (C. Mühlhausen, T. Streichert, and T. Braulke, unpublished

Fig. 2 Cis-inhibition of OCT2-mediated [3H]-MPP-uptake by 3OH- GA. OCT2-overexpressing CHO cells were incubated with 0.2 μM [3H]-MPP (a)or10μM[3H]-MPP (b) in the absence (−) and presence (+) of 2 mM 3OH-GA or 2 mM MPP for 1 s. c OCT2-overexpressing Fig. 1 Renal expression analysis of OCT2. a The relative mRNA cells were incubated with 0.2 μM[3H]-3OH-GA for 1 s in the absence expression of OCT2 was estimated by qRT-PCR analysis. The (−) and presence (+) of 2 mM 3OH-GA. Internalized radioactivity was expression levels were standardized by expression levels of glyceral- measured and related to cell protein. The results were expressed in dehyde-3-phosphate dehydrogenase in Gcdh−/− mice at 42 days of relation to unaffected uptake, which was estimated to be 400,000 cpm age compared with age-matched control mice. Data are mean±SEM per mg protein (a, b) and 350 cpm per mg protein (c). The values are (n=6). Double asterisk indicates p<0.001. b OCT2 Western blot the mean±SE of four (a, c) or one experiment (b) carried out in analysis of kidney membrane extracts of control and Gcdh−/− mice. quadruplicates. Double asterisks indicate p<0.001 for difference to β-Actin immunoreactivity was used as loading control control J Mol Med (2007) 85:763–770 767

These experiments suggest that, at low concentrations of α-ketoglutarate>GA>succinate. As ΔImax values depend [3H]-MPP, 3OH-GA may compete for binding to OCT2. on the level of NaDC3 expression, these values were No uptake of [3H]-3OH-GA was observed in OCT2- normalized to the currents induced at a saturating succinate overexpressing CHO cells (Fig. 2c). The data indicate that concentration (ΔIsucc, 1 mM) measured in the same oocyte OCT2 does not directly transport 3OH-GA but appears to (Table 1). The normalized ΔImax values decrease in the be regulated in functional properties by the organic anion. order GA>α-ketoglutarate>oxaloacetate>3OH-GA. Under Expression of NaDC3 in the kidney was demonstrated by immunohistochemical staining showing prominent and specific expression in proximal tubules, colocalizing with the basolateral membrane K+/Cl− (K-Cl) cotransporter Kcc4 (Slc12a7; [7]; supplementary online Fig. 1). The antibody appeared to react with a protein distributed all over proximal tubule cells. Higher magnification revealed staining in the infoldings of the basolateral membrane, the site of NaDC3 (not shown). The sensitivity of the immunohistochemical staining, however, did not allow for the assessment of quantitative differences in NaDC3 expression between Gcdh−/− and control mice. Unfortunately, the antibody available against NaDC3 did not function in Western blotting. The electrogenic NaDC3-mediated transport of 3OH-GA was examined in hNaDC3-expressing oocytes. At −60 mV, addition of either 1 mM GA or 3OH-GA induced inward currents compared with ORI alone (Fig. 3b). These currents differed in their magnitude. In paired experiments, inward currents of −80±44 nA with 1 mM GA and of −13±4 nA with 1 mM 3OH-GA were observed. These substrate- associated currents were abolished when sodium was replaced by N-methyl-D-glucamine, indicating sodium- dependent translocation of GA and 3OH-GA (Fig. 3c). The kinetics of hNaDC3-induced 3OH-GA translocation was determined by measuring the 3OH-GA-associated inward currents at varying concentrations of 3OH-GA

(Fig. 3d). The process was saturable with a KM of 0.954± 0.182 mM, and the maximal substrate inducible current

ΔImax was −45±4 nA (Table 1). The KM for other dicarboxylates decreased in the order oxaloacetate>

Fig. 3 Expression analysis and transport of GA and 3OH-GA mediated by hNaDC3. The relative mRNA expression of NaDC3 in kidneys of Gcdh−/− mice at 42 days of age (a) was estimated by standardized qRT-PCR analysis. Data are mean±SEM (n=6). Single asterisk indicates p<0.05. b Substrate-associated inward currents induced by GA and 3OH-GA (each 1 mM). c 3OH-GA-dependent steady-state currents expressed as a percentage of the steady-state current evoked by GA in the same hNaDC3-expressing oocyte in the presence (grey columns) and absence (black columns) of sodium. Substrate concentrations of 1 mM were used, and sodium was replaced by equimolar concentrations of N-methyl-D-glucamine. The results shown were obtained in eight oocytes from four different frogs. Double asterisk indicates p<0.001. d Concentration-dependence of 3OH-GA-evoked currents. 3OH-GA concentrations as indicated were applied either in decreasing or in increasing order in five oocytes from three donors. Experiments (b, c, d) were performed at −60 mV, and data represent the mean±SE (c, d) 768 J Mol Med (2007) 85:763–770

Table 1 Substrate specificity of hNaDC3 intermediates in a competitive manner. Under basal con- c ditions cerebral concentrations of GA and 3OH-GA in post Substrate KM (μM) ΔImax (nA) ΔImax/ΔIsucc n/m mortem brain biopsy material of GA1 patients have been Succinatea 25±16 −42±21 1.0 4/3 calculated to be 500–5,000 μM and 200 μM, respectively, b Oxaloacetate 152±30 −22±7 0.43±0.12 5/4 similar to GA and 3OH-GA concentrations in brains of a − GA 40±12 46±9 1.33±0.09 5/3 Gcdh−/− mice [21]. During encephalopathic crises, however, α-ketoglutaratea 45±13 −26±3 0.99±0.23 4/4 induced in Gcdh−/− mice [5] brain levels of 3OH-GA 3OH-GAb 950±180 −47±4 0.35±0.06 5/3 increased further by 90% (S. I. Goodman, unpublished

Values are means±SE. KM is the Michaelis–Menten constant, results). At these high concentrations, 3OH-GA may expressing the substrate concentration where half-maximal currents considerably interact with NaDC3 in the brain despite of are achieved at −60 mV, and ΔImax is the substrate-inducible current its low affinity for this dicarboxylate (KM 950 μM). at saturating substrate concentrations. ΔImax/ΔIsucc represents the substrate-inducible current at saturating substrate concentrations In kidney proximal tubule cells, NaDC3 is believed to divided by the current induced at 1 mM succinate in the same oocyte mediate the uptake of α-ketoglutarate at the basolateral to normalize for individual NaDC3 expression. membrane [22] required indirectly for the secretion of a a μ Data were obtained using 10, 50, 100, 500, and 1,000 M variety of endogenous and exogenous organic anions, such dicarboxylate; b Data from 50, 100, 200, 500, 1,000, and 2,000 μM dicarboxylate; as ß-lactam antibiotics, diuretics, and nonsteroidal anti- c Number of oocytes/number of frogs used in the study. inflammatory drugs, translocated at the basolateral membrane by organic anion transporters OAT1 and OAT3 [12]. In agreement with the basolateral localization of NaDC3 in all conditions tested so far, inward currents were absent in human proximal tubule cells [22], in the mouse kidney, oocytes injected with an equivalent amount of water, NaDC3 staining was found in the basolateral membrane indicating that the substrate-associated currents depend on (supplementary online Fig. 1). The increased expression of the expression of hNaDC-3 (not shown). NaDC3 in kidneys of Gcdh−/− mice reported here may be one component of an adaptive system participating in urinary secretion of GA/3OH-GA. This system, however, Discussion requires additional luminal GA/3OH-GA transporters, which remain to be identified. Because NaDC3 translocates GA1 patients are characterized by high levels of the oxaloacetate (Table 1), its upregulation may be linked to an dicarboxylates GA and 3-hydroxyglutaric acid in all tissues, increased requirement of TCA cycle intermediates for plasma, and urine because of the deficiency of the Pck1-mediated gluconeogenesis. Pck1 expression was mitochondrial enzyme Gcdh [1]. It is believed that these found to be significantly elevated in kidneys of Gcdh−/− dicarboxylates, in particular 3OH-GA, contribute to striatal mice (supplementary online Table II). neurodegeneration through excitotoxic mechanisms and The second important result of this report is the cis- disrupt structural vascular integrity [3–5]. This study inhibitory effect of 3OH-GA on OCT2-mediated uptake of describes the identification of the first transporter, the MPP (Fig. 2a). OCT2 (Slc22a2) is localized primarily at the Na+-dependent dicarboxylate cotransporter NaDC3, able basolateral membranes of epithelial cells of S2 and S3 to transport 3OH-GA, and the capability of 3OH-GA to act segments of the proximal tubules, and in neurons of the as a cis-inhibitor of the organic cation transporter OCT2. cerebral cortex [13], mediating the electrogenic transport of NaDC3 is highly expressed in the brain and kidneys and a broad range of structurally diverse cationic compounds. exhibits variable binding properties for a broad range of These include the prototypic organic cation tetraethylam- dicarboxylate substrates, such as α-ketoglutarate, GA, and monium, MPP, antidiabetics, neurotoxins, and endogenous succinate as well as its derivatives 2,3-dimethyl- and 2,3- substrates such as choline and monoamine neurotransmit- dimercaptosuccinate [12, 15, 16]. Recent studies investigat- ters [13, 23, 24]. Although MPP and 3OH-GA are ing cultured rodent brain cells provided evidence that structurally different, the data presented in Fig. 2a and b NaDC3 is expressed in astrocytes but not in neurons [17, suggest that 3OH-GA acts as a competitive inhibitor in a 18]. The primary function of NaDC3 in the brain is likely to complex binding pocket of the polyspecific OCT2 contain- supply tricarboxylic acid (TCA) cycle intermediates gener- ing several overlapping interaction domains for different ated in astrocytes to neurons required for amino acid substrates [14, 24]. It is tempting to speculate that cis- neurotransmitter synthesis of glutamate and gamma amino- inhibitory effects of 3OH-GA may impair the OCT2- butyric acid and energy metabolism [19, 20]. At present, it is mediated removal of cationic toxins adding to the unknown whether increased concentrations of 3OH-GA and complexity of neurodegenerative mechanism in GA1. GA in the brain during catabolic crises of GA1 patients The molecular mechanism regulating the mRNA expres- impair the supply of neuronal cells with TCA cycle sion of various genes in the kidneys of Gcdh−/− mice are J Mol Med (2007) 85:763–770 769 yet unknown. However, under experimental conditions of 8. Chirgwin JM, Przybyla AE, MacDonald RJ, Rutter WJ (1979) induced chronic metabolic acidosis, several renal mRNAs Isolation of biologically active ribonucleic acid from sources – were found to be increased, including Pck1 and the apical enriched in ribonuclease. Biochemistry 18:5294 5299 9. Livak KJ, Schmittgen TD (2001) Analysis of relative gene NaDC1 transporter (Slc13a2) [25, 26], which were also expression data using real-time quantitative PCR and the 2-DDCT upregulated in kidneys of Gcdh−/− mice (supplementary method. Methods 25:402–408 online Table II). Furthermore, the reduced urinary pH in 10. Lips KS, Volk C, Schmitt BM, Pfeil U, Arndt P, Miska D, Ermert Gcdh−/− mice (F. Stellmer, unpublished results) suggests L, Kummer W, Koepsell H (2005) Polyspecific cation transporters mediate luminal release of acetylcholine from bronchial epitheli- an acid–base imbalance that may contribute to renal um. Am J Respir Cell Mol Biol 33:79–88 hypertrophy in Gcdh−/− mice [6, 27]. It remains to be 11. Koepsell H (1998) Organic cation transporters in intestine, kidney, investigated whether the expression of NaDC3 and OCT2 liver, and brain. Annu Rev Physiol 60:243–266 in Gcdh−/− mice are part of a coordinate response to the 12. Burckhardt BC, Burckhardt G (2003) Transport of organic anions across the basolateral membrane of proximal tubule cells. Rev acidification of the fluid of the tubular lumen and decrease Physiol Biochem Pharmacol 146:95–158 in intracellular pH. 13. Busch AE, Karbach U, Miska D, Gorboulev V, Akhoundova A, Taken together, the accumulation of major dicarboxyl- Volk C, Arndt P, Ulzheimer JC, Sonders MS, Baumann C, ates, 3OH-GA and GA, in the circulation and tissues in Waldegger S, Lang F, Koepsell H (1998) Human neurons express the polyspecific cation transporter hOCT2, which translocates mice deficient for Gcdh leads to complex renal adaption monoamine neurotransmitters, amantadine, and memantine. Mol processes affecting the abundance and activity of several Pharmacol 54:342–352 transporter proteins including NaDC3 and OCT2. Such 14. Popp C, Gorboulev V, Müller TD, Gorbunov D, Shatskaya N, adaptions may contribute both to restore the acid–base Koepsell H (2005) Amino acids critical for substrate affinity of rat organic cation transporter 1 line the substrate binding region in a balance and to enhance renal extraction of nonmetaboliz- model derived from the tertiary structure of lactose permease. Mol able organic acids. Pharmacol 67:1600–1611 15. Pajor AM, Gangula R, Yao X (2001) Cloning and functional characterization of a high-affinity Na+/dicarboxylate cotrans- Acknowledgements This work was supported by the Deutsche porter from mouse brain. Am J Physiol Cell Physiol 280: Forschungsgemeinschaft (GRK 336 to Keyser and Stellmer; grant C1215–C1223 BU998/2-3 to Burckhardt) and the Arbeitsgemeinschaft für Pädiatrische 16. Burckhardt BC, Lorenz J, Kobbe C, Burckhardt G (2005) Stoffwechselstörungen (APS). Substrate specificity of the human renal sodium dicarboxylate cotransporter, hNaDC-3, under voltage-clamp conditions. Am J Physiol Renal Physiol 288:F792–F799 17. Wada M, Shimada A, Fujita T (2006) Functional characterization References of Na+-coupled citrate transporter NaC2/NaCT expressed in primary cultures of neurons from mouse cerebral cortex. Brain Res 1081:92–100 1. Goodman SI, Frerman FE (2001) Organic acidemias due to 18. Yodoya E, Wada M, Shimada A, Katsukawa H, Okada N, defects in lysine oxidation: 2-ketoadipic acidemia and glutaric Yamamoto A, Ganapathy V, Fujita T (2006) Functional and acidemia. In: Scriver CR, Beaudet AL, Sly WS, Valle D, Childs molecular identification of sodium-coupled dicarboxylate trans- B, Kinzler KW, Vogelstein B (eds) The metabolic and molecular porters in rat primary cultured cerebrocortical astrocytes and bases of inherited disease. McGraw-Hill, New York, pp 2195– neurons. J Neurochem 97:162–173 2204 19. Schousboe A, Westergaard N, Waagepentersen HS, Larsson OM, 2. Ullrich K, Flott-Rahmel B, Schluff P, Musshoff U, Das A, Lücke Bakken IJ, Sonnewald U (1997) Trafficking between glia and T, Steinfeld R, Christensen E, Jakobs C, Ludolph A, Neu A, neurons of TCA cycle intermediates and related metabolites. Glia Röper R (1999) Glutaric aciduria type I: pathomechanism of 21:99–105 neurodegeneration. J Inherit Metab Dis 22:392–403 20. Peng LA, Schousboe A, Hertz L (1991) Utilization of α- 3. Kölker S, Koeller DM, Okun JG, Hoffmann GF (2004) ketoglutarate as a precursor for transmitter glutamate in cultured Pathomechanisms of neurodegeneration in glutaryl-CoA dehydro- cerebellar granule cells. Neurochem Res 16:29–34 genase deficiency. Ann Neurol 55:7–12 21. Sauer SW, Okun JG, Fricker G, Mahringer A, Müller I, Crnic LS, 4. Mühlhausen C, Ott N, Chalajour F, Tilki D, Freudenberg F, Mühlhausen C, Hoffmann GF, Hörster F, Goodman SI, Harding Shahhossini M, Thiem J, Ullrich K, Braulke T, Ergün S (2006) CO, Koeller DM, Kölker S (2006) Intracerebral accumulation of Endothelial effects of 3-hydroxyglutaric acid: implications for glutaric and 3-hydroxyglutaric acids secondary to limited flux glutaric aciduria type I. Pediatr Res 59:196–202 across the blood-brain barrier constitute a biochemical risk factor 5. Zinnanti WJ, Lazovic J, Wolpert EB, Antonetti DA, Smith MB, for neurodegeneration in glutaryl-CoA dehydrogenase deficiency. Connor JR, Woontner M, Goodman SI, Cheng KC (2006) A diet- J Neurochem 97:899–910 induced mouse model for glutaric aciduria type I. Brain 129:899– 22. Bai X, Chen X, Feng Z, Hou K, Zhang P, Fu B, Shi S (2006) 910 Identification of basolateral membrane targeting signal of human 6. Koeller DM, Woontner M, Crnic LS, Kleinschmidt-DeMasters B, sodium-dependent dicarboxylate transporter 3. J Cell Physiol Stephens J, Hunt EL, Goodman SI (2002) Biochemical, patho- 206:821–830 logic and behavioral analysis of a mouse model of glutaric 23. Gorboulev V, Ulzheimer JC, Akhoundova A, Ulzheimer-Teuber acidemia type I. Hum Mol Genet 11:347–357 I, Karbach U, Quester S, Baumann C, Lang F, Busch AE, 7. Boettger T, Hübner CA, Maier H, Rust MB, Beck FX, Jentsch TJ Koepsell H (1997) Cloning and characterization of two human (2002) Deafness and renal tubular acidosis in mice lacking the polyspecific organic cation transporters. DNA Cell Biol 16:871– K-Cl co-transporter Kcc4. Nature 416:874–878 881 770 J Mol Med (2007) 85:763–770

24. Arndt P, Volk C, Gorboulev V, Budiman T, Popp C, Ulzheimer- mRNA and protein abundance in rat kidney. Kidney Int 58:206– Teuber I, Akhoundova A, Koppatz S, Bamberg E, Nagel G, 215 Koepsell H (2001) Interaction of cations, anions, and weak base 26. Curthoys NP, Gstraunthaler G (2001) Mechanism of increased quinine with rat renal cation transporter rOCT2 compared with renal gene expression during metabolic acidosis. Am J Physiol rOCT1. Am J Physiol Renal Physiol 281:F454–F468 Renal Physiol 281:F381–F390 25. Aruga S, Wehrli S, Kaissling B, Moe OW, Preisig PA, Pajor AM, 27. Preisig PA (1999) A cell cycle-dependent mechanism of renal Alpern RJ (2000) Chronic metabolic acidosis increases NaDC-1 tubule epithelial cell hypertrophy. Kidney Int 56:1193–1198