Zebrafish Slc5a12 Encodes an Electroneutral Sodium Monocarboxylate Transporter (SMCTn)

We have identified and characterized two different sodium-coupled monocarboxylate cotransporters (SMCT) from zebrafish (Danio rerio), electrogenic (zSMCTe) and electroneutral (zSMCTn). zSMCTn is the 12th member of the zebrafish Slc5 gene family (zSlc5a12). Both zSMCT sequences have ∼50% homology to human SLC5A8 (hSMCT). Transport function and kinetics were measured in Xenopus oocytes injected with zSMCT cRNAs by measurement of intracellular Na+ concentration ([Na+]i) and membrane potential. Both zSMCTs oocytes increased [Na+]i with addition of monocarboxylates (MC) such as lactate, pyruvate, nicotinate, and butyrate. By using two electrode voltage clamp experiments, we measured currents elicited from zSMCTe after MC addition. MC-elicited currents from zSMCTe were similar to hSMCT currents. In contrast, we found no significant MC-elicited current in either zSMCTn or control oocytes. Kinetic data show that zSMCTe has a higher affinity for lactate, nicotinate, and pyruvate (KmL-lactate = 0.17 ± 0.02 mm, Kmnicotinate = 0.54 ± 0.12 mm at -150 mV) than zSMCTn (KmL-lactate = 1.81 ± 0.19 mm, Kmnicotinate = 23.68 ± 4.88 mm). In situ hybridization showed that 1-, 3-, and 5-day-old zebrafish embryos abundantly express both zSMCTs in the brain, eyes, intestine, and kidney. Within the kidney, zSMCTn mRNA is expressed in pronephric tubules, whereas zSMCTe mRNA is more distal in pronephric ducts. zSMCTn is expressed in exocrine pancreas, but zSMCTe is not. Roles for Na+-coupled monocarboxylate cotransporters have not been described for the brain or eye. In summary, zSMCTe is the zebrafish SLC5A8 ortholog, and zSMCTn is a novel, electroneutral SMCT (zSlc5a12). Slc5a12 in higher vertebrates is likely responsible for the electroneutral Na+/lactate cotransport reported in mammalian and amphibian kidneys.

SLC5A8 was originally cloned by Rodriguez et al. (17) in an attempt to identify iodide transporters that might play a functional role in the physiology of the thyroid gland. This gene codes for a protein with 46% identity in amino acid sequence with NIS (Slc5a5) (18). The SLC5A8 protein is expressed in the apical membrane of thyroid follicular cells and was hypothesized to mediate passive transport of iodide independent of Na ϩ ; consequently, Slc5a8 was originally named the "apical iodide transporter" (17).
Recently, the SLC5A8 gene was shown to play an important role in controlling the development of colon cancers (19) and classical papillary thyroid carcinomas (PTC-cf) (20). Silencing of SLC5A8 is a common and early event in both tumors. Promoter methylation that inactivates SLC5A8 is found in ϳ60% of colon cancer cell lines and primary colon cancers and Ͼ80% adenomas (19). In the case of PTC-cf, SLC5A8 silencing by methylation was found in 90% of thyroid carcinomas (follicular, the most common type) and occurred in about 20% of other papillary thyroid carcinomas (20). SLC5A8 is thus also hypothesized to be a tumor suppressor gene. We reasoned that such important functions of SLC5A8 should be evolutionarily preserved and could be crucial in development and differentiation, particularly of epithelial tissues. In fact, Costa and co-workers (21) have found that Slc5a8 (Vito) is first expressed at the blastopore lip (the beginning of neurulation and gastrulation) in Xenopus tadpoles.
Slc5a8 was also cloned from mouse kidney (23). Not surprisingly, mouse Slc5a8 expressed in Xenopus oocytes is electrogenic, transporting the same substrates as human SMCT (13,14). Thus, Slc5a8 is the first transporter known to be expressed in mammalian colon and kidney that has the ability to mediate the electrogenic absorption of lactate and other monocarboxylates with Na ϩ (13,23,24). The K m and V max values for MC substrates of SLC5A8 indicate that it is a high affinity and low capacity transporter that could easily be saturated by MCs in the renal ultrafiltrate or in the colonic lumen. Thus, an electroneutral Na ϩ /MC cotransporter, particularly with high transport capacity, would allow MC absorption without excess loss in the urine or stool. Yet the molecular identity of such electroneutral Na ϩ /lactate transport proteins in kidney remains unknown.
In this study we identified, localized, and functionally characterized two zebrafish SMCTs. One is the electrogenic SMCT (Slc5a8), and the other is a novel electroneutral form of the Na ϩ /monocarboxylate cotransporter (SMCTn) with high transport capacity. SMCTn is the zebrafish ortholog of Slc5a12 recently reported as a second electrogenic Na ϩ /monocarboxylate cotransporter (32). Our data demonstrate that both SMCTn and the electrogenic SMCT (SMCTe/Slc5A8) are present in kidney and brain of the teleost Danio rerio (zebrafish) and have different kinetic properties. Both proteins increase intracellular Na ϩ in response to a variety of MCs as follows: short chain fatty acids, lactate, pyruvate, nicotinate, acetoacetate, and 3-␤-hydroxybutyrate. Slc5a8 mRNA is also found in the gut consistent with its original description in human.

EXPERIMENTAL PROCEDURES
cDNA Cloning of Zebrafish SMCT, zSlc5a8 and zSlc5a12-A BLAST search of the EST data base revealed two D. reiro cDNAs that were homologous to human SLC5A8, yet clearly distinct from one another. The cDNAs corresponding these ESTs (IMAGE clones zSMCTe (zSlc5a8, 7212813) and zSMCTn (Slc5a12, 6793401)) were obtained from the IMAGE consortium (Research Genetics, Genome Systems) and sequenced (GenBank TM accession numbers AY727859 for zSMCTe and AY727860 for zSMCTn). The complete EST inserts were sequenced (W. M. Keck Facility, New Haven, CT) and then subcloned in pGEMHE Xenopus oocyte expression vector between EcoRI (5Ј) and XbaI (3Ј).
Oocyte Isolation and Injection-Oocytes were removed from female X. laevis (Xenopus Express, Beverly Hills, FL) as described previously (25). Excised lobes of oocytes were placed into a Ca 2ϩ -free buffered saline solution (200 mosM) and defolliculated by collagenase digestion as described previously (26). Capped zSlc5a8 and zSlc5a12 cRNAs were synthesized using a linearized cDNA template and the T7 mMessage mMachine (Ambion, Austin, TX). Oocytes were injected with 50 nl of 0.5 ng/nl (25 ng/oocyte) cRNA of zSMCTe, zSMCTn, hSMCT, or water. The oocytes were maintained at 18°C in filtered ND96 medium (96 mM NaCl, 2 mM KCl, 1 mM MgCl 2 , 1.8 mM CaCl 2 , 5 mM HEPES-Tris (pH 7.5)) (13). Oocytes were studied 3-10 days after injection. All experiments were performed in accordance with the regulations of the Institutional Animal Care and Use Committee of Case Western Reserve University.
Ion-selective Microelectrode-Ion-selective microelectrodes were used to monitor intracellular Na ϩ concentration ([Na ϩ ] i ) of zSMCTs and water-injected oocytes as described previously (25)(26)(27). Intracellular [Na ϩ ] was measured as the difference between the Na ϩ electrode and a KCl voltage electrode impaled into the oocyte; and membrane potential (V m ) was the potential difference between the KCl microelectrode and an extracellular calomel (26,28). Na ϩ microelectrodes were calibrated using 10 mM NaCl and 100 mM NaCl solutions followed by point calibration in ND96 (96 mM Na ϩ ) as performed previously (26,28). Intracellular Na ϩ microelectrodes had slopes of at least Ϫ50 mV/p(Na ϩ ).
Two Electrode Voltage Clamp-Oocyte membrane currents were recorded using an OC-720C voltage clamp (Warner Instruments, Hamden, CT), filtered at 2-5 kHz, digitized at 10 kHz, and recorded with the Pulse software, and data were analyzed using the PulseFit program (HEKA, Germany) as described previously (25,29). For periods when I-V protocols were not being run, oocytes were clamped at a holding potential (V h ) of Ϫ50 mV; and the current was constantly monitored and recorded at 1 Hz. I-V protocols consisted of 400-ms steps from V h to Ϫ150 mV and ϩ50 mV in 20-mV steps as described previously (13).The I-V protocols were run in the absence and in the presence of a particular substrate. The substrate-specific current is determined by subtraction of the pre-substrate current from the substrate current. All of the experiments were performed at room temperature. Test substrates bathed oocytes for 1-3 min. For kinetic analysis, oocytes were exposed to only one monocarboxylate at one concentration and then discarded.
Calculations-Oocytes were perfused with ND96 for 5 min at which time "initial [Na ϩ ] i " is measured. The solution was switched to different monocarboxylates for 8 -10 min (i.e. [Na ϩ ] i and V m or I plateau), and the "final [Na ϩ ] i " is measured. Oocytes came from at least four separate donor animals. 22 Na ϩ Uptakes-Functions of zSMCTe and zSMCTn were assessed by measuring 22 Na ϩ uptake in groups of 15-20 oocytes 4 days after water or cRNA injection. 22 Na ϩ uptake was performed as described previously (30). Briefly, 22 Na ϩ uptake was measured with the following protocol: a 30-min incubation period ND96 with 1 mM ouabain, 100 M amiloride, 100 M bumetanide, followed by a 60-min uptake period in ND96. To analyze the pyruvate transport kinetics, we varied pyruvate concentration from 0 to 5 mM in ND96 for zSMCTe and from 0 to 20 mM for zSMCTn. These uptake solutions also contained 1 Ci/ml of 22 Na ϩ (PerkinElmer Life Sciences) and the same inhibitors used during the incubation period. Uptakes were performed at 30°C. At the end of the uptake period, oocytes were washed five times in ice-cold ND96 solution without isotope to remove extracellular tracer. Next, individual oocytes were dissolved in 10% SDS, and tracer activity was determined by ␤ scintillation counting. RNA-injected oocytes were compared with water controls subjected to identical conditions, using oocytes from the same donor. Kinetic analysis was performed by estimating the K m values for pyruvate of each SMCT. The K m values were calculated from log [ion] versus V/V max plots using GraphPad Prism (GraphPad Software, Inc., San Diego).
Fish Embryo Sections-After in situ hybridization, 5 dpf embryos were postfixed in 4% paraformaldehyde for 1 h at room temperature. Fixed embryos were incubated overnight in 30% sucrose in phosphate-buffered saline and then embedded in OCT. 15-m cryosections were cut with a Leica cryostat and mounted on gelatin-coated slides. Embryo sections were observed using a Zeiss Axiovert 25 microscope and acquired with an AxioCam digital camera and AxioVision software (Carl Zeiss, Germany) as described previously (27).
Statistical Analysis-The results are presented as means Ϯ S.E. The significance of the differences between groups was tested by one-way analysis of variance with multiple comparison using Bonferroni correction or by the Kruskal-Wallis oneway analysis of variance on ranks with the Dunn method for multiple comparison procedure, as needed.
Functional Expression of zSMCT Proteins in Xenopus Oocytes-We analyzed the function of both zSMCT clones by simultaneous measurement of membrane potential (V m ) and intracellular Na ϩ concentration ([Na ϩ ] i ) with ion-selective microelectrodes in Xenopus oocytes injected with zSMCTn or zSMCTe cRNA. Using two electrode voltage clamp experiments, we also examined currents elicited by monocarboxylates for both SMCT cotransporters.
When the MC tested was removed, the [Na ϩ ] i increase stopped. Fig. 2 also shows that the [Na ϩ ] i increment elicited by butyrate is dependent on extracellular Na ϩ , i.e. when extracellular Na ϩ was removed, it elicited a rapid fall of [Na ϩ ] i in all SMCT oocytes but not in controls despite the continued butyrate presence (zSMCTe ϭ Ϫ0.58 mM, zSMCTn ϭ Ϫ1.25 mM, hSMCTe ϭ Ϫ0.45 mM; water ϭ Ϫ0.13 mM). These data indicate that Na ϩ and MCs move together and are capable of moving into as well as out of the oocytes. Although it is tempting to compare Na ϩ transport rates of zSMCTe versus zSMCTn from Fig. 2, it is difficult, however, to compare these rates because the transporters have different affinities and capacities (see below). Additionally, zSMCTe and zSMCTn are subject to different electrochemical driving forces, i.e. zSMCTe is voltage-dependent and voltage-affected, whereas zSMCTn is not. These Na ϩ transport data in Fig. 2C versus data in Fig. 2D appear to indicate that zSMCTe does not readily reverse as zSMCTn with the removal of a given MC. A similar result is observed with human SLC5A8 (Fig. 2B). One consideration is that Slc5a8 clones are voltage-dependent, meaning that inward transport is higher at more negative voltages whereas outward transport is reduced at more negative voltages. On the other hand, an electroneutral transporter like zSMCTn/zSlc5a12 does not have its transport driven by voltage but rather is "concentration gradient"driven (ln{[substrates] out /[substrates] in }). Thus, the transport driving force for SMCTe clones would predict lower Na ϩ and MC outward transport because of the electrical driving force (2 ϫ V m ). This electrical driving force does not influence zSMCTn-mediated transport activity. If these SMCTe and SMCTn experiments were both performed under voltage clamp conditions, the data would be more comparable. Nevertheless, MC metabolism or altered intracellular MC affinity for transport site might also have a role in these differences.
The average V m change (⌬V m ) (Fig. 2, bottom columns) in the presence of various MCs for nine zSMCTn oocytes (Fig. 2H) was not different between the average ⌬V m for six control oocytes (Fig. 2E). Yet both data sets are significantly different from the average ⌬V m for eight zSMCTe oocytes (Fig. 2G) and for nine hSMCT oocytes (Fig. 2F).  mM, L-lactate ϭ 0.82 Ϯ 0.15 mM, pyruvate ϭ 1.03 Ϯ 0.24 mM). Interestingly, hSMCTe oocytes do not display any significant differences of [Na ϩ ] i movement for the MCs tested (⌬[Na ϩ ] i : butyrate ϭ 0.87 Ϯ 0.23 mM, propionate ϭ 0.83 Ϯ 0.12 mM, L-lactate ϭ 0.69 Ϯ 0.17 mM, pyruvate ϭ 0.85 Ϯ 0.22 mM). This analysis suggests that the affinity of zSMCTe for pyruvate and zSMCTn for pyruvate/L-lactate is different from other MCs.
These results indicate that like human SMCTe, zSMCTe electrogenically transports Na ϩ , and Na ϩ transport is elicited by monocarboxylates. Moreover, these results illustrate that zSMCTn is a new electroneutral Na ϩ /monocarboxylate cotransporter that is the molecular ortholog of mouse Slc5a12 (32).
A recent study by Srinivas et al. (32) reported the cloning of mouse Slc5a12 as a low affinity, electrogenic Na ϩ /monocarboxylate cotransporter (SMCT2). Our zSMCTn clone is most closely related to this mouse Slc5a12. Therefore, we tested if zSMCTn might be electrogenic if presented with 40 mM nicotinate (Fig. 3E) or 30 mM L-lactate (Fig. 3G) and compared the zSMCTn responses to those of water-injected control oocytes (Fig. 3, D and F, respectively). We verified that there was no osmolality differences in our test solutions by maintaining constant [Cl Ϫ ] and replacing gluconate with lactate or nicotinate. These voltage clamp data indicate that between Ϫ150 and ϩ50 mV there is no difference between water-injected and zSMCTn-injected oocytes (not shown), yet Fig. 2, D and H (32) reported that they had insufficient protein expression to allow kinetic analysis of mouse Slc5a12. Because we have robust Na ϩ transport mediated by zSMCTn (Slc5a12 ortholog), we compared the affinities of zSMCTn and zSMCTe for L-lactate and nicotinate by measuring either [Na ϩ ] i responses (Fig. 4) or current responses (Fig. 5) to varying concentrations. For zSMCTn, Na ϩ intracellular changes during a 5-min exposure (⌬Na ϩ i /5min) to varying concentrations of L-lactate (0 -30 mM) or nicotinate (0 -50 mM) were measured using Na ϩ -selective microelectrodes in a non-voltage-clamped configuration (Fig. 4, A and B). To account for potential biologic variation, we used at least seven zSMCTn-injected oocytes for each MC concentration from at least three different frogs. Fig. 4 shows steady-state kinetics calculated from the Michaelis-Menten equation for lactate (Fig. 4A), nicotinate (Fig. 4B), and pyruvate (Fig. 4, C and D). The data show that zSMCTn has ϳ15 times greater affinity for lactate (K m L-lactate ϭ 1.81 Ϯ 0.19 mM) and pyruvate (K m pyruvate ϭ 2.03 Ϯ 0.36 mM) than nicotinate (K m nicotinate ϭ 23.68 Ϯ 4.88 mM). The Na ϩ transport capacity of zSMCTn oocytes was lower for lactate (V max L-lactate ϭ 3.10 Ϯ 0.09 mM Na ϩ i /5 min) than nicotinate (V max nicotinate ϭ 4.80 Ϯ 0.05 mM Na ϩ i /5 min). 6 Fig. 4C shows that the maximal 22 Na ϩ transport of zSMCTe and zSMCTn for pyruvate is ϳ18 nmol/ 6 The [Na ϩ ] i response of zSMCTn-injected oocytes to MCs was linear over at least the first 5 min. These data further indicate that the transport of Na ϩ /MCs via zSMCTn is electroneutral.
To determine whether transport by zSMCTe differed in affinity or capacity from either zSMCTn or hSMCTe, we measured zSMCTe currents elicited by varying doses of L-lactate (Fig. 5, A and B) from 0 to 5 mM or nicotinate (Fig. 5, E and F) from 0 to 10 mM using a two electrode voltage clamp protocol. Fig. 5, A and E, shows the average IV curves from seven oocytes injected with zSMCTe and exposed to L-lactate or nicotinate solutions, respectively. Fig. 5, A and E, shows that the resulting zSMCTe currents are voltage-and concentration-dependent as observed for other MCs with hSMCTe (13). The analysis of steady-state kinetics calculated from the Michaelis-Menten equation for lactate data (Fig. 5B) and nicotinate data (Fig. 5F) show that the apparent K m value of each MC is constant for negative voltages (Ϫ150 to 0 mV) but tends to increase for voltages greater than ϩ10 mV for lactate (not statistically significant). This analysis also shows that zSMCTe has ϳ3-fold greater affinity for L-lactate than nicotinate (L-lactate ϭ 169 Ϯ 20 M; nicotinate ϭ 535 Ϯ 120 M at Ϫ150 mV; see Fig. 5, D and H). The I max value for L-lactate (Fig. 5C) and nicotinate (Fig. 5F) is smaller and constant for positive voltages, but I max values gradually increased at negative voltages. The zSMCTe-I max is slightly higher for L-lactate (L-lactate ϭ Ϫ697 Ϯ 21 nA at Ϫ150 mV) than nicotinate (Ϫ513 Ϯ 14 nA at Ϫ150 mV). Thus, the zSMCTe affinity for lactate and nicotinate are close to those reported for mammalian SMCT (Slc5a8), i.e. L-lactate ϭ 235 Ϯ 24 M and nicotinate ϭ 296 Ϯ 88 M (13,14,22,23). Taken together these data demonstrate that affinity of zSMCTn for L-lactate and nicotinate are ϳ10 times 7 and ϳ30 times lower than the affinity of zSMCTe for the same compounds, respectively (compare Fig. 4 with Fig. 5).
Nicotinate Alters L-Lactate Transport by zSMCTn-Under normal physiological circumstances multiple monocarboxylates 7 The average unclamped voltages for zSMCTn oocytes (Fig. 4) for these experiments was Ϫ45 Ϯ 6 mV. Thus, a comparison of the zSNCTn-K m values to the zSMCTe-K m values at Ϫ50 mV (Fig. 5) was used to calculate the 10-or 30-fold difference. are present. We were particularly interested in the interrelationship of nicotinate-(K m ϳ 24 mM) and L-lactate (K m ϳ 2 mM)-mediated Na ϩ transport via zSMCTn. To do this, we measured [Na ϩ ] i in unclamped zSMCTn oocytes and varied the order of L-lactate and nicotinate addition. Fig. 6A is a representative zSMCTn experiment in which we first exposed a zSMCTn oocyte to 10 mM nicotinate (below K m ) for 5 min and then to 10 mM L-lactate (above K m ). Fig. 6C is a similar zSMCTn experiment in which the oocyte was exposed first to 10 mM L-lactate and then 10 mM nicotinate. In both experiments, L-lactate elicits a large increase of [Na ϩ ] i (Fig. 6A ] i increase with L-lactate was decreased by ϳ30% after the exposure to nicotinate. The average rates for the "nicotinate-lactate" protocol were 0.15 Ϯ 0.05 mM Na ϩ i /min for nicotinate and 0.33 Ϯ 0.04 mM Na ϩ i /min for lactate. For the "lactate-nicotinate" protocol, the rate of nicotinate-elicited Na ϩ i transport was constant (0.16 Ϯ 0.02 mM Na ϩ i /min) whereas that rate of lactate-elicited [Na ϩ ] i transport was dramatically increased (0.57 Ϯ 0.04 mM Na ϩ i /min). Therefore, pre-exposure to L-lactate did not diminish the rate or extent of Na ϩ transport elicited by nicotinate. Fig. 6, B and D, shows average [Na ϩ ] i responses of five experiments similar to Fig. 6A (nicotinate 0.72 Ϯ 0.18 mM; L-lactate ϭ 1.58 Ϯ 0.19 mM) and six experiments similar to Fig. 6C (L-lactate ϭ 2.84 Ϯ 0.21 mM; nicotinate 0.79 Ϯ 0.08 mM). These data show that the Na ϩ transport amount by L-lactate was reduced in zSMCTn oocytes by ϳ40% after the exposure of nicotinate. These data also suggest that reduction of lactate movement via zSMCTn may be due to both MCs competing for the putative MC-binding site likely due to the obvious differences in K m values for the two substrates.
Ketone Body Transport-Ketone bodies are produced by the liver and used peripherally as an energy source during fasting and other lipolytic stress (1,2). The two main ketone bodies are acetoacetate (AcAc) and ␤-hydroxybutyrate (␤OH-buty), whereas acetone is the third, less abundant, ketone body. We therefore evaluated the ability of the zSMCT clones to transport Na ϩ in response to the presence of AcAc and ␤OH-buty (Fig. 7). These keto acids play important roles in basic metabolism (see "Discussion" and Fig. 10A).
In Situ Hybridization-The distribution of zSMCTe and zSMCTn expression was determined by in situ hybridization in zebrafish embryo whole mounts with a digoxigenin-labeled, antisense cRNA probe for zSMCTe and zSMCTn ( Fig. 8 and Fig. 9). Control zebrafish embryos show no signal in the organs with reactivity for either zSMCTe or zSMCTn transcripts (Fig. 8, A, D, and G). Signals for zSMCTn and zSMCTe, respectively, were detected in brain and eyes 24 h post-fertilization (hpf) (Fig. 8, B and C), 3 dpf (Fig. 8, E and F), and 5 dpf (Fig. 8, H and I). Both zSMCTe (Fig. 8I) and zSMCTn (Fig.  8H) transcripts are also present in the swim bladder of 5-day-old embryos. zSMCTe (Fig. 8, C, F, and I) and zSMCTn (Fig. 8, E and H) transcripts were also present in the pronephros (embryonic kidney). Although zSMCTn mRNA was in the pronephric tubule (early region of pronephros) (Fig. 8H), zSMCTe mRNA was in pronephric ducts (late region of pronephros) (Fig. 8I). Interestingly, zSMCTe transcripts appeared in the pronephros at 24 hpf (Fig. 8C), whereas zSMCTn transcripts were first evident at 3 dpf (Fig. 8E). We speculate that this message, and presumably protein distribution, is reflective of the absorptive capacity and function as well as the changing role of the pronephros between 24 and 96 hpf.
SLC5A8 was originally cloned from human colon, and thus its expression was expected in the teleost gut. Fig. 8I illustrates that zSMCTe mRNA is present in the gut at 5 dpf.
We further analyzed the zSMCTn and zSMCTe mRNA localization in different tissues by sectioning 5 dpf embryos FIGURE 7. Keto-acid transport by zSMCT, intracellular Na ؉ . Simultaneous measurement of intracellular Na ϩ concentration ([Na ϩ ] i ) (top panels) and membrane potential (V m ) (bottom panels) for a water-injected control oocyte (A and G), zSMCTn-injected oocyte (B and H), and zSMCTe-injected oocyte (C). Oocytes were continually superfused as indicated under "Experimental Procedures." 1 mM (A-C) or 5 mM (G and H) of AcAc, ␤OH-buty, or Buty were added as indicated by horizontal lines. Average Na ϩ transport response (⌬[Na ϩ ] i ϭ end[Na ϩ ] i Ϫ initial [Na ϩ ] i ) in presence of AcAc (black bar), ␤OH-buty (gray bars), and Buty (open bars) are indicated in the column graphs: water-injected oocytes n ϭ 5 (D), zSMCTn-injected oocytes, n ϭ 6 (E), and zSMCTe-injected oocytes, n ϭ 6 (F). Data were collected from at least two frogs. *, statistical significance (p Յ 0.05) for Na ϩ i increment compared with Buty response for that clone. Individual oocyte experiments are noted with numeric labels (below time bars) indicating the oocyte number used.

DISCUSSION
We have identified and characterized two Na ϩ /monocarboxylate cotransporters from D. rerio (zSMCT), one electrogenic (zSMCTe) and another that is electroneutral (zSMCTn). Molecular and functional characteristics indicate that zSMCTe is the ortholog of hSMCT/SLC5A8. The second transporter is a novel electroneutral SMCT, evidently the 12th member of the Slc5 gene/protein family (zSMCTn/Slc5a12). Like the hSMCT/SLC5A8 cotransporter (13,33), both zSMCTe and zSMCTn will transport a wide variety of monocarboxylates (short chain fatty acids, pyruvate, lactate, and nicotinate) in a Na ϩ -dependent manner.
Reports from the 1980s indicated electroneutral Na ϩ /MC transport in the luminal membrane of salamander proximal tubules (5) and membrane vesicles prepared from the cortex of rabbit kidney (4, 12, 34), with other data indicating the exist-  E, and H) and zSMCTe (C, F, and I) mRNAs are present in the pronephros (pronephric tubule (PT) and pronephric duct (PD)), eye (E), brain (Br), swim bladder (SB), and gut (G) (blue staining and arrows). Note absence of staining in control embryos treated identically, but not exposed to probe (A, D, and G). ence of electrogenic Na ϩ /MC transport (6 -8). Therefore, there has been experimental support for the existence of two kinds of Na ϩ -coupled monocarboxylate transport in kidney. Our studies in zebrafish are the first to experimentally determine the molecular entities for both of these processes, electroneutral transport (zSMCTn/zSlc5a12) and electrogenic transport (zSMCTe/zSlc5a8), and demonstrate their coexistence in the kidney (Figs. 8 -10).
Although in situ staining at both 3 and 5 dpf show pronephros staining, mRNAs for both zSMCTs are expressed by 24 hpf in the eye and brain. It is well known that the MCT transporters (Slc16) are expressed in the retinal pigmented epithelia and Müller cells (35,36), but the role of Na ϩ -coupled monocarboxylate cotransporters has not been incorporated into models of the retinal pigmented epithelia or visual system. Likewise, monocarboxylate cotransport has been shown in a variety of central nervous system tissues and cells, yet the possibility of Na ϩ coupling to this monocarboxylate transport has not been evaluated. Interestingly, mRNA reactivity seems labile in the brain and eyes. Messages in the brain and eye seem readily degraded in embryos that are not freshly isolated and fixed (within 9 months; data not shown).
Li et al. (19) first identified human SLC5A8 as a tumor suppressor in colon cancer. In fact, transfection of colon carcinoma cell lines with SLC5A8 dramatically reduced the growth rate of these cancer cells. This tumor suppressor role appears to be recapitulated in several other epithelial and nervous tissues (20,37,38). In support of a differentiation-promoting role, a search for genes involved in early development in Xenopus by Costa et al. (21) identified a transporter sequence Vito, also known as Xenopus Slc5a8, whose mRNA appears early in development (first evident at stage 11 in the superficial ectoderm of the blastopore lip). Vito (Slc5a8) is also abundant at the gastrula and neurulation stages. By stage 35 (organ development), Vito is abundant in the liver diverticulum, the pronephros, and the tail tip (21). Based on their results, these investigators postulated that Vito (two isoforms) is associated with early ectodermal cell determination (21). These data seem consistent with the view that Slc5a8 promotes differentiation. We also predict that demethylation of the Slc5a8 promoter region activates Slc5a8 transcription.
Our experiments here have only examined SMCT-mediated transport and SMCT-mRNA localization and thus do not discriminate between these possibilities.
It is well established that apoptosis provides a mechanism to remove damaged or oncogenic cells. It is plausible that restriction of metabolic fuels (lactate, pyruvate, etc.) is another form of cellular defense against growth dysregulation (initial oncogenic step). Slc5a8 (zSMCTe) and Slc5a12 (zSMCTn) are Na ϩcoupled monocarboxylate cotransporters carrying lactate, pyruvate, and other monocarboxylate metabolic fuels across epithelial membranes. Butyrate is a well known inhibitor of histone deacetylases, which directly regulate the extent of winding or unwinding of specific genes on chromosomes (39). Thus, butyrate transport per se may also play a key role in the differentiation or tumor suppressor effects associated with SLC5A8. These cause and effect relationships of monocarboxylate transport, differentiation, and tumor suppression await detailed studies to specifically address these critical issues.
Several studies have also reported the kinetics of renal Na ϩ /MC cotransport (7,8,12,34,40). These studies suggested the presence of two kinetically different Na ϩ /MC cotransport systems with one having higher affinity for MCs than the other. In this study we show that zSMCTn has a lower affinity for MCs than zSMCTe or hSMCTe (Figs. 4 and 5). Because zSMCTn is proximal to zSMCTe in the pronephros (see diagram in Fig.  10B), its lower MC affinity is anticipated. The pronephros localizations of zSMCTn and zSMCTe are likely reflective of their physiologic role in the uptake of MCs. The concentration of MCs from the glomerular filtrate that arrives in the early proximal tubule can be high (L-lactate, 1-10 mM in mammals). Thus, efficient absorption of MCs occurs by having a low affinity but high capacity transport of MCs (SMCTn). The MC concentration in the tubule fluid decreases as absorption occurs. Consequently, efficient MC absorption in the later proximal tubule is accomplished by high affinity MC transport (SMCTe) (Fig. 10B).
Studies with salamander proximal tubule showed that luminal, electroneutral lactate transport is Na ϩ -coupled and induces an intracellular alkalinization (41), whereas basolateral lactate transport is H ϩ -coupled (42). In our experiments we observed a clear Na ϩ -dependent, alkalinization mediated by short chain fatty acids (butyrate and propionate) in oocytes injected with zSMCTn. This alkalinization was not observed with the addition of 1 mM pyruvate, lactate, or nicotinate, perhaps because of some endogenous oocyte mechanism for H ϩ /pyruvate or H ϩ /lactate cotransport that dissipated the alkalinization produced by zSMCTn. In SMCTe-and SMCTninjected oocytes, alkalinization was consistently elicited by short chain fatty acids in our experiments.
There are 11 SLC5 genes identified in humans (15,16). Our   (43,44). It is attractive to speculate that defects in SMCT (SLC5A8) function could cause a saturation of the renal Na ϩ -coupled monocarboxylate transport system and might be manifest by monocarboxylates, especially lactate, in the urine. In fact, Thirumurugan et al. (45) have found that L-lactate excretion is increased in the Fanconi syndrome, which encompasses generalized defects in proximal tubule function. Transport by SMCTs may be the major mechanism by which lipid-modifying drugs gain entry into cells, i.e. clinically useful. Since the 1950s, nicotinic acid (niacin, Niasapan) has been used as a lipid-modifying drug (gram doses) in the treatment of dyslipidemias (46,47) or dyslipidemia with diabetes (11). Several hypotheses have been proposed regarding the basic mechanism for this niacin effect on blood lipids. These hypotheses include, among other things, a vasodilation effect, a vitamin effect, decreased hepatic cholesterol synthesis, increased cholesterol oxidation, etc. Regardless of the underlying mechanisms, it is clear that the hypolipidemic effect of nicotinate requires the recently identified nicotinate receptor (GPR109A/PUMA-G) (47)(48)(49). It is also well established that ligand-receptor interaction time is often determined by ligand degradation (acetylcholine receptors and acetylcholinesterase) or ligand uptake mechanisms. One particularly germane example is vesicular glutamate release at the central nervous system presynaptic terminals, stimulating post-synaptic glutamate receptors and rapid uptake of glutamate by adjacent glutamate transporters (SLC1 family) (50 -54). In humans genomic DNA, GPR109A is at 12q24.31 (49) and SLC5A8 is at 12q22-24 (19). It is attractive to speculate that these genes are located in the same region of the genome so that their activities may be coordinated.
Another explanation for the role of nicotinate in dyslipidemia therapy involves inhibition of lipolysis in adipose tissue, resulting in a decreased mobilization of free fatty acids (9 -11). Nicotinate is efficiently transported with Na ϩ by both Slc5a8 (zSMCTe) (13,22) and Slc5a12 (zSMCTn) but with varied affinities and capacities (Figs. [3][4][5]. Moreover, as illustrated in Fig. 6, the amount and rate of nicotinate transport by zSMCTn are significantly reduced compared with lactate transport. This transport reduction likely points to competition for the monocarboxylate-binding site of zSMCTn such that sterically "bulky" MCs compete with more simple MCs, including short chain fatty acids. These data support the idea that one possible therapeutic action of nicotinic acid in hyperlipidemia is to reduce the absorption of free fatty acid in tissues expressing SMCTs (Slc5a8 and Slc5a12) or MCTs (Slc16a1-a4).
Lactate transport is also reported in brain, but this transport is attributed to H ϩ /MC cotransport (MCT, Slc16a1-4) (55,56). Monocarboxylates such as lactate, pyruvate, and ketone bodies appear to play an important role in brain energy metabolism (Fig. 10A) as well as neuronal and glial intracellular pH buffering (57-59). Lactate has long been considered a waste product disposed of via the circulation; however, when lactate is accumulated it can be toxic for the brain. Monocarboxylates, together with other non-glucose substrates, have long been known to be substantial energy substrates for the developing brain (60). Na ϩ /MC cotransport in brain has not been reported previously. Nevertheless, we show clear evidence of both SMCT transcripts in the developing zebrafish brain (Figs. 8 and  9). The roles of the SMCT transporters in the brain and eye remain to be explored, yet central nervous system function of these transporters may provide some clues to early central nervous system development or cellular "programming." Ketone bodies or keto acids, such as AcAc and ␤OH-buty (Fig. 10A), are always present in the blood, and their levels increase during fasting and prolonged exercise. The metabolic pathway of hepatic ketone body formation (ketogenesis) and extrahepatic ketone body utilization (ketolysis) are especially important for the brain, which uses ketone bodies as a primary energy source when glucose is not available (3). Additionally, ketone bodies are used as substrates for lipid synthesis such as cholesterol for myelin and for neonates and during suckling (1).
Physiologic levels of ketone bodies circulating in plasma range from Ͻ0.1 mM (postprandial) to 6 mM (prolonged fasting) and can reach 25 mM in diabetic ketoacidosis. Diabetes is the most common pathological cause of elevated blood keto acids. In diabetic ketoacidosis, high levels of keto acids are produced in response to low insulin levels and high levels of counterregulatory hormones. In acute diabetic ketoacidosis, the keto acid ratio (␤OH-buty:AcAc) rises from normal (1:1) to as high as 10:1. In response to insulin therapy, ␤OH-buty levels commonly decrease long before AcAc levels (1)(2)(3). The H ϩ /MC cotransporters (MCT) are involved in the transport of ketone bodies (4). However, the chromosomal localization of the human homolog of zSMCT (SLC5a8 and SLC5a12) are positional candidates for proteins involved in diseases like noninsulin-dependent type 2 diabetes (NIDDM2 on 12q22-24, MIM 601407) (43,44) and familial combined hyperlipidemia (11p14) (61, 62), respectively.
Our current data suggest that zSMCTe (Slc5a8) and zSMCTn (Slc5a12) have significant differences in affinities for the keto acids AcAc and ␤OH-buty (Fig. 7). Furthermore, these mRNAs and presumably proteins are present in several organs, including the kidney. The chromosome localization of the human SLC5A8 and SLC5A12 orthologs are positional candidates for noninsulin-dependent type II diabetes and hyperlipidemia. We speculate that both SMCTe and SMCTn have important roles in the exacerbating diabetic ketoacidosis because reabsorption of these keto acids by SMCT proteins after glomerular filtration would prevent normal elimination or metabolic use leading to increasing plasma levels of these keto acids and acidosis. Finally, it should be noted that even though zSMCTe is only ϳ50% identical to hSMCT (SLC5A8), the affinity of zSMCTe for lactate and nicotinate were quite similar to those reported for mammalian SMCT (13,(22)(23)(24).
In summary, we have identified, cloned, functionally characterized, and localized two distinct Na ϩ /monocarboxylate cotransporters in the zebrafish, D. rerio. Our data demonstrate that zSMCTe is the molecular and functional ortholog of human high affinity SMCT (SLC5A8). In the 1st day of zebrafish development, zSMCTe is present, implicating that the presence of SMCT is important for normal development. The second transporter, zSMCTn, is a novel, electroneutral, and low affinity Na ϩ /monocarboxylate cotransporter, with a definite mammalian homolog (Slc5a12). Additionally, more than being merely renal or gut transporters, our zebrafish data demonstrate significant expression in the developing brain, eye, and ear. The roles of the SMCT proteins in these tissues will likely lead to novel biochemical, physiologic, and pathophysiologic insights.