Structural and Functional Characteristics of Two Sodium-coupled Dicarboxylate Transporters (ceNaDC1 and ceNaDC2) from Caenorhabditis elegans and Their Relevance to Life Span*

We have cloned and functionally characterized two Na (cid:1) -coupled dicarboxylate transporters, namely ceNaDC1 and ceNaDC2, from Caenorhabditis elegans . These two transporters show significant sequence homology with the product of the Indy gene identified in Drosophila melanogaster and with the Na (cid:1) -coupled dicarboxylate transporters NaDC1 and NaDC3 identified in mammals. In a mammalian cell heterologous expression system, the cloned ceNaDC1 and ceNaDC2 mediate Na (cid:1) coupled transport of various dicarboxylates. With succinate as the substrate, ceNaDC1 exhibits much lower affinity compared with ceNaDC2. Thus, ceNaDC1 and ceNaDC2 correspond at the functional level to the mammalian NaDC1 and NaDC3, respectively. The nadc1 and nadc2 genes are not expressed at the embryonic stage, but the expression is detectable all through the early larva stage to the adult stage. Tissue-specific expression pattern studies using a reporter gene fusion approach in transgenic C. elegans show that both genes are coexpressed in the intestinal tract, an organ responsible for not only the digestion and absorption of nutrients but also for the storage of energy in this organism. Inde-pendent knockdown of the function of these two transporters in C. elegans using the strategy of RNA interference

These are NaDC1 and NaDC3. NaDC1 is Na ϩ -coupled, electrogenic, and exhibits low affinity for its dicarboxylate substrates. The K t value (Michaelis-Menten constant) is in the range of 0.1-4.0 mM (1)(2)(3)(4). This isoform is expressed primarily in the brush border membrane of intestinal and renal epithelial cells. The physiological function of NaDC1 is to absorb the intermediates of the citric acid cycle, such as citrate, succinate, ␣-ketoglutarate, fumarate, and malate, in the intestine and kidney. NaDC3 is also a Na ϩ -coupled and electrogenic dicarboxylate transporter, but it exhibits relatively higher affinity for its substrates compared with NaDC1 (5-7). The K t value is in micromolar range. The NaDC3 is expressed primarily in the basolateral membrane of intestinal and renal epithelial cells. However, it is also found in tissues such as liver, placenta, and brain. NaDC3 in the kidney is involved in generating the driving force for the organic anion transporter OAT1 to facilitate the active entry of organic anions into the tubular cells across the basolateral membrane (8). In the brain, NaDC3 mediates the cellular uptake of N-acetylaspartate, a process closely linked to myelination (9). Therefore, the physiological functions of the NaDCs may extend beyond the mediation of cellular entry of citric acid cycle intermediates. Recently, we reported on the molecular identification of the third member of this family in mammals (10,11). This transporter, known as Na ϩcoupled citrate transporter (NaCT), mediates the cellular uptake of citrate in a Na ϩ -coupled manner.
In a recent study by Rogina et al. (12), a NaDC-like transporter, coded by the Indy (for I am not dead yet) gene, has been implicated in the regulation of life span in Drosophila. The investigators of this study suggested that defects in one copy of the Indy gene (heterozygosity) can lead to less efficient production of cellular energy and that, as a consequence, the metabolic profile of the fruit fly changes resulting in life span extension. The eating behavior of the organism is not altered, however. The decreased generation of cellular energy due to the heterozygous mutation in the Indy gene creates a biological situation resembling that of caloric restriction, which in other animal models leads to an extension of life span (13). Recently, we have identified (14) the transport function of Drosophila INDY. This transporter mediates the cellular uptake of a broad spectrum of citric acid cycle intermediates in a Na ϩ -independent manner. These characteristics of drINDY have now been confirmed independently by Knauf et al. (15).
Studies of life span extension are difficult, if not impossible, to conduct in mammals, particularly in humans. But it is relatively a simple task to monitor the mean and maximum life span in other animal models such as Caenorhabditis elegans. A number of features make C. elegans especially suitable for studies of life span extension. This organism has a short life span with a mean life span of ϳ15 days. In addition, there are techniques available to silence genes in this organism as a means of assessing the role of specific genes in the maintenance of life span. Therefore, with an aim to investigate the potential role of NaDC family in life span, we set out to clone the C. elegans counterparts of mammalian NaDCs and subsequently to monitor the influence of these transporters on life span by using the RNAi technique to down-regulate their function. These studies have successfully led to the molecular and functional identification of two different Na ϩ -coupled dicarboxylate transporters (ceNaDC1 and ceNaDC2) analogous to the mammalian NaDC1 and NaDC3. In addition, studies of the influence of these two transporters on life span have shown that disruption of the function of the high affinity transporter ceNaDC2, but not that of the low affinity transporter ceNaDC1, leads to a significant extension of the average life span in C. elegans.

EXPERIMENTAL PROCEDURES
Nematode Culture-A wild type nematode strain, C. elegans N2 (Bristol-Myers Squibb Co.), was obtained from the Caenorhabditis Genetics Center (St. Paul, MN). Nematode culture was carried out using a standard procedure with a large scale liquid cultivation protocol (16 -19). The nematodes were cleaned by sedimentation through 15% (w/v) Ficoll 400 in 0.1 M NaCl. The pellet was then used for total RNA preparation.
Extraction and Purification of Poly(A) ϩ RNA-Total RNA was isolated using the TRIzol reagent (Invitrogen). Poly(A) ϩ mRNA was purified by affinity chromatography using oligo(dT)-cellulose.
Reverse Transcription (RT)-PCR and Hybridization Probe Preparation-A pair of PCR primers specific for the putative C. elegans nadc1 gene was designed based on the sequence of the cosmid F31F6.6 (Gen-Bank TM accession number Z69884), 5Ј-GCC TCC AAG CAA AAT GTC TC-3Ј (forward primer) and 5Ј-CTA ACG CAA ATC CAC CTC C-3Ј (reverse primer). A second pair of PCR primers specific for putative C. elegans nadc2 gene was designed based on the sequence of the cosmid K08E5.2 (GenBank TM accession number Z30974), 5Ј-TCA TCC TTC CAA CAC CAT CC-3Ј (forward primer) and 5Ј-ACC ATT CCA CTT CCA AAC AC-3Ј (reverse primer). Poly(A) ϩ RNA (ϳ0.5 g) isolated from mixed stage C. elegans was taken as template to perform RT-PCR using an RT-PCR kit from PerkinElmer Life Sciences. A single RT-PCR product was obtained with an estimated size of ϳ1.0 and ϳ0.9 kb for the cenadc1 and the cenadc2 genes, respectively, as predicted by the distance between the two primers in each pair. The RT-PCR products were gel-purified and subcloned into pGEM-T Easy Vector (Promega, Madison, WI). The molecular identity of the inserts was established by sequencing. These cDNA fragments were used as probes to screen a C. elegans cDNA library.
Construction of a Directional C. elegans cDNA Library-The Super-Script Plasmid System from Invitrogen was used to establish the cDNA library using the poly(A) ϩ RNA from C. elegans. The transformation of the ligated cDNA into Escherichia coli was performed by electroporation using ElectroMAX DH10B competent cells. The bacteria plating, the filter lifting, the DNA fragment labeling, and the hybridization methods followed the routine procedure (20). The DNA sequencing of the full-length ceNaDC1 cDNA and ceNaDC2 cDNA clones was performed using an automated PerkinElmer Life Sciences Applied Biosystems 377 Prism DNA sequencer and the Taq DyeDeoxy terminator cycle sequencing protocol.
Vaccinia/T7 Expression System-Functional expression of the ceN-aDC cDNAs in mammalian cells was done in human retinal pigment epithelial (HRPE) cells using the vaccinia virus expression system as described previously (5,7,9). HRPE cells grown in 24-well plates were infected with a recombinant vaccinia virus (VTF 7-3 ) at a multiplicity of 10 plaque-forming units/cell. The virus was allowed to adsorb for 30 min at 37°C with gentle shaking of the plate. Cells were then transfected with the plasmid DNA (empty vector pSPORT or ceNaDC1 cDNA or ceNaDC2 cDNA constructs) using the lipofection procedure (Invitrogen). The cells were incubated at 37°C for 12 h and then used for determination of transport activity. Cells transfected with pSPORT alone without the cDNA insert were used as the control to determine endogenous transport activity in these cells. [ 3 H]Succinate uptake was determined at 37°C as described previously (5,7,9). In most experiments, the uptake medium was 25 mM Hepes/Tris (pH 7.5), containing 140 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl 2 , 0.8 mM MgSO 4 , and 5 mM glucose. In experiments in which the cation and anion dependence of the transport process was investigated, NaCl was replaced iso-osmotically by LiCl, KCl, sodium gluconate, or N-methyl-D-glucamine (NMDG) chloride. The transport activity in cDNA-transfected cells was adjusted for the endogenous activity to calculate the ceNaDC cDNA-specific activity. The endogenous succinate transport activity in vector-transfected cells was always less than 10% compared with the succinate transport activity measured in cells transfected with either ceNaDC1 cDNA or ceNaDC2 cDNA. Experiments were done in triplicate, and each experiment was repeated at least three times. Results are presented as means Ϯ S.E.
Semi-quantitative RT-PCR-An RT-PCR assay with the cenadc1-or cenadc2-specific primers described above was used to study the developmental stage-specific expression pattern of ceNaDC1 mRNA and ceNaDC2 mRNA. A Quantum RNA 18 S internal standard (Ambion, Austin, TX) was used for the semi-quantitative RT-PCR. Total RNA (ϳ1.0 g) isolated from different developmental stages of C. elegans (embryo, early larva, late larva, and adult) was taken as template to perform reverse transcription using an RT-PCR kit from PerkinElmer Life Sciences. The reverse transcription was initiated with random hexamers and carried out in a DNA thermal cycler (GeneAmp PCR System 9600) and thin-walled reaction tubes (PerkinElmer Life Sciences) at 42°C for 60 min, followed by incubation at 99°C for 5 min to inactivate the reverse transcriptase (Maloney murine leukemia virusreverse transcriptase). Reverse transcription was followed by PCR in a multiplex format, in which the gene-specific primers and the primers for the internal control (18 S rRNA) with their competimers were combined at a predefined ratio. The PCR cycle number was titrated according to the manufacturer's protocol to ensure that the reaction was within the linear range. A competimer was included to prevent the highly abundant rRNA from being overwhelmingly amplified during the reaction, and an optimal 18 S primer:competimer ratio was also pre-established by trial and error. The resultant multiplex PCR products were resolved in an 1.0% agarose gel, and the intensity of the gene-specific and the 18 S rRNA-specific bands was determined using SpectraImager 5000 Imaging system and AlphaEase 32-bit software (Alpha Innotech, San Leandro, CA). The steady state levels of ceNaDC1 mRNA and ceNaDC2 mRNA at different developmental stages were assessed from the relative ratios of the intensity of the ceNaDC1specific RT-PCR product or ceNaDC2-specific RT-PCR product to the intensity of the 18 S rRNA-specific RT-PCR product at each of these stages.
Analysis of Tissue-specific Expression Pattern of cenadc1 and cen-adc2-To study the tissue-specific expression pattern of the nadc1 and nadc2 genes in C. elegans, transcriptional cenadc1::gfp and cenadc2::gfp fusion genes were constructed, and transgenic animals expressing these transgenes were developed. The expression pattern of the cenadc1 and cenadc2 genes was investigated in live transgenic animals based on the expression pattern of the GFP reporter. A pair of primers for construction of a transcriptional cenadc1::gfp fusion gene was designed to amplify the cenadc1 promoter. The forward primer, 5Ј-CGC GTC GAC GCT TAC ATC ATT CTT GTA TTT TTC-3Ј, corresponds to the nucleotide positions 28,630 -28,659 of the cosmid F31F6 (GenBank TM accession number Z69884). The primer contains an incorporated SalI restriction site at its 5Ј end. The reverse primer, 5Ј-ATA GGA TCC ATG ATT GGA GGC TCT GCA ATA CTA-3Ј, corresponds to the nucleotide positions 29,797-29,772 in the same cosmid. A BamHI site was incorporated in this primer at the 5Ј end. The SalI and BamHI sites were introduced into these primers for subsequent cloning into a GFP expression vector. An ϳ1.2-kb DNA fragment of the cenadc1 promoter was amplified using the cosmid F31F6 DNA as template. Similarly, a pair of primers for construction of a transcriptional cenadc2::gfp fusion gene was also designed. The forward primer, 5Ј-GTC GAC AAA ATA TGT ATT AGC CAC ATA AAA CCC-3Ј, corresponds to the nucleotide positions 13,966 -13,998 of the cosmid K08E5 (GenBank TM accession number Z30974). The reverse primer, 5Ј-GGA TCC ATT TTC CGC ACA TGC CGA ATT TGC AT-3Ј, corresponds to the nucleotide positions 15,461-15,432 in the same cosmid. A SalI site and a BamHI site were incorporated in these primers for subsequent subcloning purposes. A ϳ1.5-kb DNA fragment of the cenadc2 promoter was amplified using the cosmid K08E5 as template. The PCR products were digested with SalI and BamHI and inserted into a GFP expression vector pPD117.01 (a generous gift from Dr. A. Fire, Carnegie Institution, Baltimore, MD) at a SalI/BamHI site. In these minigene constructs, a built-in mec7 promoter (ϳ0.9 kb) in the expression vector was replaced by the cenadc1 and cenadc2 promoter fragments in such a way that the GFP transcription is under control of the putative promoters of the cenadc1 and cenadc2 genes, respectively. The minigene fusion constructs were verified by sequence analysis. Transgenic lines were es-tablished using a standard germ line transformation protocol (17,18). Syncytial gonad injection was carried out according to a standard procedure (18). For microinjection, a computerized injection system, Transjector 5246 and Micromanipulator 5171 from Eppendorf (Hamburg, Germany), and a Nikon Eclipse TE 300 inverted microscope with Nomarski differential interference contrast optics were used. A cloned mutant collagen gene containing the rol-6 (plasmid pRF4, kindly provided by Dr. M. Koelle, Yale University School of Medicine, New Haven, CT) was used as a dominant genetic marker for DNA transformation. Coinjection of this dominant marker with the GFP fusion constructs allowed progeny selection of the transformed animals by their "roller" phenotype. The F1 rollers were picked up according to their characteristic rolling behavior and cultured individually to establish transformed lines. F2 rollers with extrachromosome arrays were selected for fluorescence microscopy to determine the GFP expression pattern. Stable transgenic lines were established by the ␥-irradiation method from the F2 rollers, and the background was cleaned up by several times of outcross (17,21).
Double Labeling Fluorescent Protein Expression System-Two modified versions of the Aequora victoria green fluorescent protein (GFP), designated as CFP and YFP with cyan-shifted and yellow-shifted spectra (22), respectively, were used to simultaneously follow the expression patterns of ceNaDC1 and ceNaDC2 in C. elegans. For the construction of the cenadc1 promoter-driven CFP and the cenadc2 promoter-driven YFP expression vectors, the GFP coding region in the GFP-expression vector cenadc1::gfp (pPD117) and cenadc2::gfp (pPD117) was substituted by the CFP and YFP coding regions, respectively. The CFP and YFP coding regions (ϳ950 bp) were obtained by an EcoRI/KpnI digestion of the vectors L4666 (pPD133.58) and L4664 (pPD133.51), respectively (kindly provided by Dr. A. Fire, Carnegie Institution, Baltimore, MD). The cenadc1::cfp and cenadc2::yfp expression vectors were linearized by SalI digestion and coinjected into the distal arms of the C. elegans syncytial gonads as described earlier. The extrachromosome arrays were used for fluorescence microscopy to compare the expression pattern of CFP and YFP in the same transgenic animal. Epi-fluorescence microscopic analysis of the expression of CFP and YFP was performed using an Axiophot microscope (Carl Zeiss, Thornwood, NY). Excitation and emission filter settings are as follows: for CFP examination, excitation at 436 Ϯ 20 nm, dichroic 455 nm LP, and emission at 480 Ϯ 40 nm; for YFP examination, excitation at 500 Ϯ 20 nm, dichroic 515 nm LP, and emission at 535 Ϯ 30 nm LP (22). The filter sets were purchased from Chroma Technology (Brattleboro, VT). The SPOTcooled CCD color digital camera (Diagnostic Instruments Inc., St. Sterling Heights, MI) and its associated data acquisition software were used to record the fluorescence images.
Bacteria-mediated RNA Interference (RNAi)-A fragment of the coding region of ceNaDC1 cDNA was generated by PCR and subcloned into a pGEM-T easy vector (Promega, Madison, WI). The DNA fragment was released by EcoRI digestion and inserted into a "double T7" plasmid (pPD129.36, a generous gift from Dr. A. Fire, Carnegie Institution, Baltimore, MD) at an EcoRI site within the multiple cloning site. A host strain DH5␣ was used for the first transformation. Competent host bacteria HT115 (DE3) (kindly provided by the Caenorhabditis Genetics Center, St. Paul, MN)) expressing T7-RNA polymerase from an inducible promoter was prepared using a standard CaCl 2 method (20). The double T7 promoter-containing plasmid with the cenadc1 gene-specific DNA fragment inserted between the two T7 promoter regions was transformed into the competent HT115 (DE3) cells and plated onto standard LB ϩ tetracycline (12.5 g/ml) ϩ ampicillin (100 g/ml) plates. HT115 cells harboring the double-T7 plasmid were cultured and induced to express dsRNA using 0.4 mM isopropyl-␤-D-1-thiogalactopyranoside at 37°C for 4 h. The experimental worms were transferred onto isopropyl-␤-D-1-thiogalactopyranoside-containing nematode growth medium plates with the induced bacteria HT115 lawn for bacteria feeding experiments. The empty vector pPD129 was processed similarly for use as a negative control, and the bacteria harboring this plasmid were used to feed the control worms. Sufficient quantities of bacteria HT115 were seeded on the testing plates for the worms to consume to prevent the worms from starving and to ensure that dsRNA was always present in the testing plates during the entire experimental period for the experimental worms. A similar experimental strategy was used for ceNaDC2. To serve as a positive control for the bacteriamediated RNAi in the assessment of the influence of ceNaDC1 and ceNaDC2 on life span, we monitored the influence of DAF-2 on life span using an identical experimental approach. Homozygous daf-2 Ϫ/Ϫ knockout in C. elegans is known to enhance the life span dramatically (ϳ2-fold) (23,24). Therefore, if the knockdown of DAF-2 function by bacteria-mediated RNAi in C. elegans doubles the life span, this can be taken as a positive control for the validity of the experimental approach to assess the role of ceNaDC1 and ceNaDC2 in life span. For this purpose, we obtained an ϳ0.8-kb DNA fragment specific for C. elegans DAF-2 by RT-PCR using the following primer pairs: 5Ј-CGAACAAAA-CACATCACAGAC-3Ј (forward primer) and 5Ј-TCCATCATTTCCATCA-CAACC-3Ј (reverse primer) using the nematode total RNA as the template. This fragment was then subcloned into pGEM-T easy vector. The fragment was then released from the vector by EcoRI digestion, and the released insert was cloned into the double T7 plasmid pPD129.36 at the EcoRI site at the multiple cloning region. Shuttling of this plasmid into HT115 (DE3) bacteria, induction of double-stranded RNA, and feeding of the worms with the bacteria were carried out as described earlier.
Life Span Measurement-Life span of age-synchronous nematodes was determined at 20°C. Eggs obtained from gravid hermaphrodites using an alkaline hypochlorite treatment procedure (18) were dispensed on nematode growth medium Petri dishes with bacteria lawn and allowed to hatch. Worms were inspected every day until death and were scored as dead when they were no longer able to move even in response to prodding with a platinum-wire pick. Each day, dead worms were removed from plates and the deaths were recorded. Experiments were started with 60 worms for each RNAi treatment (10 per plate). The worms were transferred to a new plate every day during the reproductive period and every 3 days afterward to avoid contamination by their offspring. Worms that died from matricidal hatching (the bag-of-worms phenotype) and the worms that crawled off the plates or burrowed into the agar were replaced by spare worms. A backup reservoir plate of ϳ30 spare worms was started at the same time as the experimental worms and was identically treated for this purpose. To avoid the influence of any potential subjective judgment of the experimenter in identifying the dead worms on the experimental outcome, the life span measurement studies were repeated with the experimenter blinded with regard to the identity of the individual experimental groups. Statistical analysis was performed using the Microsoft EXCEL 2000 analysis ToolPak. Mean life spans from different groups were compared using the Student's t test assuming unequal population variances.

Molecular Cloning and Structural
Characterization of ceN-aDC1 and ceNaDC2-The cloned ceNaDC1 cDNA is 1,989 bp long and contains a poly(A) tail (GenBank TM accession number AY090484). The 5Ј-and 3Ј-untranslated regions of this cDNA are 12 and 173 bp long, respectively. The ceNaDC1 protein, deduced from the coding region of the cDNA, contains 582 amino acids (Fig. 1) with a predicted molecular mass of 64 kDa and an isoelectric point of 6.64. The ceNaDC2 cDNA is 2,250 bp long and contains a poly(A) tail (GenBank TM accession number AY090485). The 5Ј-and 3Ј-untranslated regions of this cDNA are 50 and 500 bp long, respectively. The ceNaDC2 protein, deduced from the coding region of the cDNA, contains 566 amino acids (Fig. 1) with a predicted molecular mass of 62 kDa and an isoelectric point of 7.69. According to the Kyte-Doolittle plot with a 21-amino acid window size, ceNaDC1 as well as ceNaDC2 possess 12 putative transmembrane domains.
Following a multiple protein sequence alignment of the two ceNaDCs, the three members of the human NaDC family, and the drINDY using the PILEUP and in combination with the MOTIFS program in the GCG package, a sodium symporter family signature motif was identified within these transporter proteins (Fig. 1). A consensus pattern established for the signature sequence is as follows: (S)SX(2)FX(2)P(V)(G)X(3)NX-(I)V, where the X denotes the flexible amino acid residues preceding the number in parentheses, and the numerical value indicates the permitted number of the flexible amino acid residues in the consensus. This sodium symporter family is a group of integral membrane proteins that mediate the cellular uptake of a wide variety of molecules including di-or tricarboxylates and sulfate by a transport mechanism involving sodium cotransport (sodium symporters). They are grouped into a single gene family on the basis of sequence and functional similarities. This group consists of the following proteins: the sodium/sulfate cotransporters and sodium/dicarboxylate cotransporters identified in yeast, C. elegans, Drosophila, and mammals; the putative sulfur deprivation response regulator (SAC1) from Chlamydomonas reinhardtii; and the hypothetical protein YfbS from E. coli (25). These transporter proteins usually consist of 430 -620 amino acids. They are highly hydrophobic and contain 11 or 12 putative transmembrane regions. The highly conserved sodium symporter signature motif is located in or near the penultimate transmembrane domain.
The molecular identity of mammalian or C. elegans functional counterpart of drINDY is not known at present. Therefore, we compared the primary structure of ceNaDC1 and ce-NaDC2 with that of drINDY and mammalian NaDC1, NaDC3, and NaCT. With a pairwise comparison analysis, ceNaDC1 is more closely related to drINDY (51% similarity and 37% identity) than ceNaDC2 (46% similarity and 35% identity). Similarly, hNaDC1 and hNaCT are more closely related to drINDY (52% similarity and 40% identity) than hNaDC3 (50% similarity and 37% identity). However, the differences are small, and it is difficult to conclude whether ceNaDC1 or ceNaDC2 is the C. elegans functional counterpart of drINDY based on the structural comparison. Similarly, this structural analysis does not allow definitive conclusion with regard to the question of whether NaDC1or NaDC3 is the mammalian functional counterpart of drINDY. Structural comparison reveals that both ceNaDCs have similar sequence homology with hNaDC1, hNaDC3, and hNaCT. Thus, the sequence data have failed to provide any useful hint with respect to the functional identity of the two ceNaDCs.
The cenadc1 and cenadc2 genes are located on chromosomes X and III, respectively. Both genes, excluding the unidentified promoter region in respective genes, are ϳ3.5 kbp in size (C. elegans data base, ACeDB, data version WS57). The presence of 10 exons in the cenadc1 gene and 8 exons in cenadc2 gene was deduced by a comparison between the sequences of the cloned cDNAs and the respective genes in the GenBank TM data base (F31F6.6 and K08E5.2) from the nematode genome sequence project. The structural organization of the cenadc1 and cenadc2 genes is shown in Fig. 2.
Functional Characterization of ceNaDC1 and ceNaDC2 Using a Heterologous Expression System-The functional analysis of the cloned ceNaDCs was carried out by heterologous expression of the cDNAs in HRPE cells using the vaccinia virus expression system (5,7,9). Cells transfected with vector alone served as the control. The transport function was monitored by the uptake of [ 3 H]succinate. Initial studies on the time course of uptake indicated that the uptake was linear at least up to 5 min. All subsequent studies were therefore carried out with a 2-min incubation. With a succinate concentration of 10 M and in the presence of Na ϩ , the uptake of succinate increased 12-fold in cells expressing ceNaDC1 compared with control cells (Fig. 3A). Under similar conditions, the increase in succinate uptake was 24-fold in the case of ceNaDC2. Thus, both ceNaDC1 and ceNaDC2 mediate the uptake of succinate in the presence of Na ϩ . The uptake via these two transporters was, however, obligatorily dependent on the presence of Na ϩ because substitution of Na ϩ with Li ϩ , K ϩ , or NMDG abolished completely the cDNA-induced increase in succinate uptake. There was no involvement of anions in the uptake process as indicated by comparable uptake activities for both transporters in the presence of NaCl or sodium gluconate (data not shown).
The substrate selectivity of the uptake process mediated by ceNaDCs was then studied by competition analysis by monitoring the ability of various monocarboxylates and dicarboxylates (5 mM) to compete with succinate for the uptake process (Fig. 3B). Uptake measurements were made in parallel in vector-transfected cells and in cDNA-transfected cells, and then the cDNA-specific uptake was calculated by subtracting the uptake in vector-transfected cells from the uptake in cDNA-transfected cells. Only the cDNA-specific uptake was used in the analysis. Among the various dicarboxylates tested, the ceNaDC1-mediated succinate uptake was inhibited markedly by fumarate, malate, ␣-ketoglutarate, dimethylsuccinate, and N-acetylaspartate. In contrast to fumarate, its stereoisomer maleate failed to compete with succinate for transport via ceNaDC1. Malonate, a structural homolog of succinate, not only failed to inhibit the uptake of succinate but actually caused a significant stimulation of succinate uptake. The monocarboxylates pyruvate, lactate, and ␤-hydroxybutyrate caused only a minimal inhibition of succinate uptake.
The substrate selectivity of ceNaDC2 was more or less similar to that of ceNaDC1. The uptake of succinate mediated by ceNaDC2 was inhibited significantly by fumarate, malate, ␣-ketoglutarate, dimethylsuccinate, and N-acetylaspartate, whereas the monocarboxylates had only a minimal effect. However, there were some notable differences between ceNaDC1 and ceNaDC2. Maleate was able to inhibit ceNaDC2-mediated succinate uptake, whereas ceNaDC1-mediated succinate uptake was not affected. Malonate, which caused a significant stimulation of succinate uptake via ceNaDC1, had minimal effect on succinate uptake via ceNaDC2. Another notable feature was that fumarate and malate were much more potent in inhibiting ceNaDC2-mediated succinate uptake than in inhibiting ceNaDC1-mediated succinate uptake, suggesting that there may be significant differences in substrate affinities between the two transporters. But, interestingly the trend in the inhibitory potency was opposite for dimethylsuccinate and Nmethylaspartate. These two dicarboxylate derivatives were more potent in inhibiting ceNaDC1-mediated succinate uptake than in inhibiting ceNaDC2-mediated succinate uptake.
The cDNA-specific succinate uptake was saturable for ceN-aDC1 as well as for ceNaDC2, and the data conformed to the Michaelis-Menten equation describing a single saturable system (data not shown). The Michaelis-Menten constant (K t ) was 0.73 Ϯ 0.05 mM for ceNaDC1 and 60 Ϯ 9 M for ceNaDC2. Thus, with succinate as the substrate, ceNaDC2 exhibits a 10-fold greater affinity than ceINDY1. The competition studies suggest that the same may be true for other dicarboxylate substrates such as fumarate and malate. These data show that  3. A, ion dependence of ceNaDC-mediated succinate uptake in HRPE cells. Uptake of 10 M succinate was measured in buffers containing 140 mM concentrations of sodium, lithium, potassium, or NMDG (as chloride salts). Values represent means Ϯ S.E. for four determinations. Uptake of 10 M succinate measured in the vector (pSPORT)-transfected cells served as a control for endogenous uptake activity. The uptake in cDNA-transfected cells is given as percent of uptake in vector-transfected cells. B, substrate specificity of the ceN-aDC-mediated uptake. Uptake of 10 M [ 3 H]succinate was measured in the absence or presence of potential inhibitors (5 mM) in cells transfected with vector alone, ceNaDC1 cDNA, or ceNaDC2 cDNA. The cDNA-specific uptake was calculated by adjusting for the uptake in vector-transfected cells. The cDNA-specific uptake in the absence of inhibitors was taken as the control (100%), and the uptake in the presence of inhibitors is given as percent of this control value. ceNaDC1 is a low affinity Na ϩ /succinate cotransporter, and ceNaDC2 is a high affinity Na ϩ /succinate cotransporter. Therefore, these two C. elegans dicarboxylate transporters correspond at the functional level to mammalian NaDC1 and NaDC3, respectively. NaCT shows very little ability to transport succinate and thus is not related to either ceNaDC1 or ceNaDC2 in terms of transport function. Drosophila INDY does have the ability to transport various dicarboxylate intermediates of citric acid cycle (14,15). But the transport function is not Na ϩ -dependent. Furthermore, Drosophila INDY has a much higher affinity for citrate, a tricarboxylate, than for dicarboxylates (14).
The effect of Na ϩ on the uptake of succinate was then investigated by measuring the uptake in the presence of varying concentrations of extracellular Na ϩ in cells transfected with either ceNaDC1 cDNA or ceNaDC2 cDNA. Again, the uptake values were adjusted for the endogenous uptake activity measured under identical conditions in cells transfected with vector alone. The concentration of Na ϩ in the uptake medium was varied from 0 to 140 mM. The osmolality of the medium was maintained by adding appropriate concentrations of NMDG chloride as a substitute for NaCl. The relationship between the cDNA-specific uptake and Na ϩ concentration was sigmoidal for both ceNaDC1 and ceNaDC2, suggesting the involvement of multiple Na ϩ ions per succinate molecule transported. The uptake rates failed to reach saturation within the concentration range of Na ϩ employed in these studies.
Developmental Stage-specific Expression Pattern of ceNaDC1 mRNA and ceNaDC2 mRNA-To monitor the relative expression levels of ceNaDC1 mRNA and ceNaDC2 mRNA during different stages of C. elegans development, synchronized cultures were obtained, and total RNA was isolated at each of the following four stages of development: embryo, early larva (larva stages 1 and 2), late larva (larva stages 3 and 4), and adult. The steady state levels of mRNAs for ceNaDC1 and ceNaDC2 were then determined by semi-quantitative RT-PCR with 18 S rRNA as an internal control for variations in RNA input into RT-PCRs. The levels of ceNaDC1 mRNA and ceNaDC2 mRNA were compared at different developmental stages based on relative intensities of ceNaDC-specific RT-PCR products compared with that of 18 S rRNA-specific RT-PCR product (Fig. 4). ceNaDC1 mRNA expression was not detectable at the embryo stage. Abundant expression of this mRNA was evident, however, at the early larva stage. There was a transient decrease in ceNaDC1 mRNA levels at the late larva stage, but the levels increased again during subsequent development into the adult stage. The levels of ceNaDC1 mRNA as assessed by the relative band intensities of RT-PCR products for ceNaDC1 and 18 S rRNA at these four stages, namely embryo, early larva, late larva, and adult, were 0, 0.87, 0.24, and 0.77. In the case of ceNaDC2, the mRNA was below detectable levels in the embryo, but the expression was easily detectable at the early larva stage. The levels of mRNA reached the maximum at the late larva stage. The relative mRNA levels for ceNaDC2 at the four stages (embryo, early larva, late larva, and adult) were 0, 2.5, 4.2, and 3.7.
Tissue-specific Expression Pattern of cenadc1 and cenadc2 Genes-We first studied the tissue expression pattern of cen-adc1 and cenadc2 genes in C. elegans using the transgenic GFP fusion technique in which the transgene consisted of the cen -FIG. 4. Developmental stage-specific expression of ceNaDC1 and ceN-aDC2. Following RT-PCR, 10 l of the products were separated in a 1.2% (w/v) agarose gel to show the size and intensity of the cenadc gene-specific and the 18 S rRNA-specific fragments. A 1.0-kb DNA marker (Invitrogen) was used as a molecular mass standard. The upper bands (ϳ850 bp long) were derived from the ce-nadc1 (upper panel) and cenadc2 (lower panel) transcripts; the lower bands (ϳ480 bp) were amplified from 18 S rRNA and served as an internal control. The RNA samples that served as templates were prepared from different developmental stages of C. elegans: embryos (Emb), early larva stage (L1&2), late larva stage (L3&4), and adults. In addition, RNA prepared from a mixture of C. elegans at different developmental stages was also used (Mix). adc1 promoter fused with GFP cDNA or the cenadc2 promoter fused with GFP cDNA. In both cases, the expression of GFP is controlled by the respective promoter. Thus, the expression pattern of GFP would match the expression pattern of the cenadc1 and cenadc2 genes because of the control of the expression of the GFP reporter by the respective gene-specific promoters. With this technique, we found that GFP expression is restricted to the intestinal tract whether the expression of GFP is driven by the ceNaDC1 promoter or by the ceNaDC2 promoter (Fig. 5, A and B), indicating that both cenadc1 and cenadc2 genes are expressed in the intestinal tract. The expression pattern is evident from the early larva stage through the adult stage (data not shown). The GFP fluorescence is detectable throughout the intestinal tract, starting from the pharynx all the way through the anus. In the case of both promoters, the expression level of GFP is significantly greater in the anterior half of the intestine than in the posterior half.
Because both cenadc1 and cenadc2 are expressed in the same tissue, we employed a double-labeling approach to verify the coexpression pattern of the two genes. In this approach, we used two different fluorescent protein reporters, each driven independently by either cenadc1 promoter (CFP) or cenadc2 promoter (YFP). Transgenic animals were developed that expressed both of the reporter constructs. The expression of CFP as well as YFP was then examined in the same transgenic animal under a fluorescence microscope with different excitation and emission filter settings. These experiments showed that the cenadc2 promoter-controlled YFP and the cenadc1 promoter-controlled CFP were coexpressed in the intestinal tract (Fig. 5, C and D). This expression pattern was confirmed with at least 10 transgenic animals.
Influence of RNAi-mediated Knockdown of the Function of ceNaDC1 and ceNaDC2 on Average Life Span-The knockdown of the function of ceNaDC1 by feeding the wild type N2 worms on bacteria expressing the ceNaDC1-specific dsRNA did not show any significant influence on average life span nor on the maximal life span (Fig. 6). The average life span of these worms was same as that of the worms fed on bacteria harboring the empty vector pPD129 (pPD129 control, 15.3 days; ceNaDC1 knockdown, 14.8 days, p Ͼ 0.05). In contrast, the knockdown of the function of ceNaDC2 by feeding the wild type N2 worms on bacteria expressing the ceNaDC2-specific dsRNA enhanced significantly (p Ͻ 0.0001) the average life span of the worms (pPD129 control, 15.3 days; ceNaDC2 knockdown, 17.6 days). The increase in average life span induced by ceNaDC2 knockdown was 15%. We used DAF-2 knockdown as a positive control in these experiments. Worms feeding on bacteria expressing DAF-2-specific dsRNA exhibited an average life span of 30 days, showing that the knockdown of the function of DAF-2 doubles the average life span. This influence of DAF-2 knockdown on life span is similar to the influence of homozygous knockout of daf-2 gene function on life span (23,24). This attests to the validity of the experimental approach indicating that the bacteria-mediated RNAi strategy is as effective as homozygous knockout strategy. DISCUSSION We have described in this paper the cloning and functional characterization of two transporters in C. elegans that mediate the transport of several intermediates of the citric acid cycle. Both transporters are Na ϩ -coupled and exhibit broad substrate specificity for dicarboxylates. They do not interact with monocarboxylates. At the functional level, these two transporters, named ceNaDC1 and ceNaDC2, resemble the mammalian Na ϩ -coupled dicarboxylate transporters NaDC1 and NaDC3, respectively.
Even though ceNaDC1 and ceNaDC2 generally resemble NaDC1 and NaDC3, respectively, in terms of functional characteristics, there is one important difference. This difference relates to the interaction of these transporters with certain derivatives of succinate such as dimethylsuccinate and N-acetylaspartate. Dimethylsuccinate is considered to be a specific substrate for mammalian high affinity transporter NaDC3 (26,27). The low affinity transporter NaDC1 does not tolerate substitutions in the carbon backbone of succinate. Thus, dimethylsuccinate and dimercaptosuccinate are recognized preferentially by NaDC3. In contrast to the mammalian counterparts, it is ceNaDC1, the low affinity transporter in C. elegans, that interacts with dimethylsuccinate with much higher affinity compared with ceNaDC2. Interaction with N-acetylaspar- tate also follows a similar pattern. In mammals, NaDC3 shows high affinity for this succinate analog (9). In contrast, it is ceNaDC1, not ceNaDC2, that shows high affinity for this compound.
Na ϩ -activation kinetics of succinate uptake mediated by ce-NaDC1 and ceNaDC2 shows that multiple Na ϩ ions are involved in the transport mechanism. Succinate exists as a divalent anion under the experimental conditions (i.e. pH 7.5), and therefore the number of Na ϩ ions involved per transport cycle will determine whether or not the transport process is influenced by membrane potential. However, the exact number of Na ϩ ions transported with succinate per transport cycle could not be determined in the present studies because the activation of succinate uptake by Na ϩ did not saturate within the range of Na ϩ concentrations employed in the study. We tried to express ceNaDC1 and ceNaDC2 in Xenopus laevis oocytes to evaluate the electrogenic nature of these two transporters by using the two-microelectrode voltage clamp method, but the transporters were not functionally expressed in this heterologous system. We do not know the reasons for the lack of expression. We are currently trying different expression vectors for this purpose. Successful expression of these transporters in X. laevis oocytes may become essential to demonstrate unequivocally whether or not ceNaDCs are electrogenic.
In mammals, the expression of NaDC1 is restricted primarily to the intestine and kidney, whereas the expression of NaDC3 is evident not only in the intestine and kidney but also in the liver, brain, and placenta (26). Furthermore, NaDC1 and NaDC3 exhibit differential distribution in the apical versus basolateral membrane of the polarized cells in the intestine, kidney, liver, and placenta. NaDC1 is localized to the apical membrane of the intestinal and renal tubular cells. In contrast, NaDC3 is localized to the basolateral membrane of the renal tubular cells, sinusoidal membrane of the hepatocytes, and the brush border membrane of the placental syncytiotrophoblast (26). The physiological function of NaDC1 in the intestine and kidney is to facilitate the absorption of exogenous dicarboxylates in the intestine and the reabsorption of endogenous dicarboxylates in the kidney. In the liver and placenta, NaDC3 may play a role in the cellular entry of circulating dicarboxylates for subsequent metabolic utilization. Because these dicarboxylates are present in the circulation only in micromolar concentrations, the high affinity transporter NaDC3 has obvious advantages over the low affinity transporter NaDC1 to perform this function. In C. elegans, the low affinity trans-porter NaDC1 as well as the high affinity transporter NaDC2 are expressed predominantly in the intestinal tract. The C. elegans intestinal tract is a tubular structure made up of a single layer of 20 donut-shaped cells (28). Unlike in mammals, the intestinal tract in C. elegans performs a variety of functions in addition to the digestion and absorption of dietary nutrients. It is a primary site of synthesis and storage of fat as the energy source, a function similar to that of liver and adipose tissue in mammals. The cells of the intestinal tract in C. elegans are polarized, with numerous microvilli on the luminal surface analogous to the apical membrane of the enterocytes in mammals. The basolateral membrane of the intestinal cells is in contact with the pseudocoelomic space that is filled with fluid that supplies nutrients to the rest of the cells in the body. In this respect, there is a lot of similarity between the intestinal tract in C. elegans and the liver and adipose tissue in mammals. We have provided evidence in this paper in support of coexpression of NaDC1 and NaDC2 in the cells of the intestinal tract in C. elegans. It is not known, however, whether these two transporters are distributed differentially in the apical versus basolateral membrane of the intestinal cells.
The physiological functions of NaDC1 and NaDC2 in C. elegans intestinal tract are not known. We used the RNAi technique to silence the function of these two transporters to evaluate their influence on life span in this organism. This technique is very effective in silencing the function of any specific protein as evidenced by the doubling of the average life span by silencing the function of DAF-2. Homozygous mutations in daf-2 gene lead to doubling of life span in C. elegans (23,24). Since RNAi-mediated targeting of daf-2 also doubles the life span, we conclude that this technique is very effective in silencing the function of any targeted gene. RNAi-mediated interference of NaDC1 function does not have any noticeable effect on the average life span as well as on the maximal life span, whereas targeting NaDC2 by this approach results in a significant increase in the average and maximal life span. We speculate that NaDC2 is localized to the basolateral membrane of the intestinal cells where it functions in the cellular entry of endogenous dicarboxylates for subsequent metabolic utilization and energy production. Interference with this function leads to a metabolic state analogous to that of caloric restriction, thus resulting in life span extension. It has been well established in C. elegans that caloric restriction (29) or suppression of metabolic energy production within the mitochondria (30) is associated with a significant increase in life span. FIG. 6. Influence of the knockdown of ceNaDC1 and ceNaDC2 on life span in C. elegans. The knockdown of ceNaDC1 and ceNaDC2 was carried out by feeding the worms with bacteria producing ceNaDC1-or ceNaDC2-specific dsRNA. The knockdown of DAF-2 was used as a positive control. The survival curves were plotted according to the Kaplan-Meier algorithm using Sigma Plot (version 6.0, SPSS Inc., Chicago). These curves show the survival probability of the wild type animals at a given day after hatching under the influence of the genespecific dsRNAs. Each group was from a total of four experiments. The total number of worms in each group at the beginning of the life span experiment was 240.
It is interesting to note that Indy gene is expressed in Drosophila in tissues such as the fat body, midgut, and oenocytes (12,15). The fat body in this organism is involved in the metabolism and storage of major energy sources (fat, glycogen, and protein). The metabolic functions of this organ are similar to those of liver in mammals. The same is true with the intestinal tract in C. elegans where NaDC2 expression is seen. However, the transport characteristics of ceNaDC2 are very different from those of Drosophila INDY even though disruption of NaDC2 function enhances life span in C. elegans as disruption of INDY does in Drosophila. In addition to NaDC1 and NaDC2 reported in this paper, a recent search of the C. elegans data base has revealed that there are three other genes coding for putative transporters with structural similarity to Drosophila INDY. Cloning and functional characterization of these putative transporters will be required to establish the molecular identity of the gene that is the C. elegans functional counterpart of Drosophila INDY. The present studies have clearly shown that NaDC2 is involved in the regulation of life span in C. elegans, but it is likely that additional transporters with NaDC2-like transport function may exist in this organism and function in the regulation of life span.