Main

The hyperornithinemia-hyperammonemia-homocitrullinuria (HHH) syndrome (OMIM 238970), an autosomal recessive disorder of the urea cycle and ornithine degradation pathway, is caused by mutations in the mitochondrial ornithine transporter (ORNT1) (1,2). In general, patients with HHH syndrome have a later onset and milder clinical presentation compared with patients with other urea cycle disorders such as deficiency of ornithine transcarbamylase (OTC). Moreover, a salient feature of HHH syndrome is the marked clinical variability seen among patients carrying the same mutant ORNT1 allele (3,4).

Biochemically, patients with HHH syndrome are characterized by chronic elevation in plasma levels of ornithine, episodic or postprandial hyperammonemia, and urinary excretion of homocitrulline and orotic acid (2). The chronic hyperornithinemia of HHH patients may be indistinguishable from the plasma ornithine levels of patients with retinopathy secondary to a deficiency in ornithine aminotransferase (OAT) (2). Presenting symptoms for HHH patients may include liver dysfunction, coagulopathies, unexplained seizure activity, developmental and growth delay, academic difficulties, gait disturbances, and stroke-like episodes (1–4). The cognitive development of HHH patients may vary from normal intelligence to mild and severe mental retardation. In general, HHH patients are maintained on a protein restricted diet and supplemented with special formulas, citrulline and sodium phenylbutyrate.

We believe that the mechanisms of disease underlying the variable phenotype of patients with HHH syndrome may include differences in mitochondrial physiology, the existence of additional transporters with redundant function to ORNT1, and environmental or epigenetic factors (3). Recent evidence that additional factors may influence the clinical presentation and course of HHH syndrome or other mitochondrial carrier disorders is the observation of posttranslational modifications, such as lysine acetylation, of murine mitochondrial carrier proteins, and Otc (5). Furthermore, we identified human ORNT2, a functional retroposon that has overlapping functions with ORNT1, as a gene that may potentially modify the phenotype of patients with HHH syndrome (6). Both ORNT2 and ORNT1 belong to the mitochondrial carrier subfamily that includes the carnitine/acyl-carnitine transporter (CACT), SLC25A29 and SLC25A45, a gene whose function is not known (7,8). By using an in vitro assay of ornithine metabolism in cultured fibroblasts of HHH patients, we previously showed that CACT, the mitochondrial transporter defective in a neonatal onset fatty acid oxidation defect, does not have overlapping function with ORNT1 (6). For this work, we reasoned that SLC25A29 may have overlapping functions with ORNT1 given that it contains the mitochondrial signature sequence P29FDT, a region that is conserved among several mitochondrial ornithine transporters from yeast to human (6). Interestingly, Sekoguchi et al. (9) previously reported that the murine Slc25a29 functions as carnitine/acyl-carnitine transporter-like (CACL) and is up-regulated in response to partial hepatectomy and fasting.

In this report, we show that the human and mouse SLC25A29 transporters, but not the mouse Slc25a45, can rescue the defective ornithine metabolism in cultured fibroblasts of patients with HHH syndrome. In addition to being CACL, SLC25A29 is the third human mitochondrial ornithine transporter, designated as ORNT3; a gene that, although expressed at relatively low levels in the liver and highest levels in the CNS, may potentially modify the clinical phenotype of patients with HHH syndrome and perhaps be pharmacologically manipulated to compensate for a defective ORNT1.

METHODS

Total RNA isolation and real time PCR.

We obtained total RNA (DNase treated, RNase-free) of different mouse tissues from CLONTECH (BD Biosciences, San Francisco, CA) and mouse cerebellar type II astrocytes from ATCC (Manassas, VA). To extract total RNA from cultured astrocytes, we used a combination of TRIzol reagent (Invitrogen, Carlsbad, CA) for isolation and Qiagen columns (Valencia, CA) for final purification and elution after on-column DNase I treatment. We performed reverse transcription with BioRad i-Script cDNA Synthesis kit (Hercules, CA) using 1 μg of total RNA in accordance with the manufacturer's recommendation. By utilizing gene specific primers (Table 1) and cDNA as template, we amplified Ornt1, Ornt3, Slc25a45, Cact, Otc, and glyceraldehyde 3-phosphate dehydrogenase (G3pdh) from different mouse tissues. The complete open reading frame (ORF) and genomic structure of all amplified mouse genes used in designing PCR primers were obtained from the Mouse Ensemble website (www.ensembl.org/Mus_musculus/index.html).

Table 1 Primer combinations for RT-PCR, qPCR, and expression constructs
Full size table

For semi-quantitative RT-PCR, serial dilutions of cDNA from mouse tissues were made to determine amplification conditions within the exponential range for each target gene. A final amount of 200 ng cDNA was amplified at 96°C (1 min), [94°C (15 s), 62°C (30 s), 72°C (20 s)] × 32 cycles and 72°C (20 s) for Ornt1, Ornt3, and Slc25a45, whereas 31 and 28 cycles were used for Cact and G3pdh, respectively. PCR products were quantified with an AlphaImager 2200 Documentation and Analysis System (Alpha Innotech Corporation, San Leandro, CA) at nonsaturating intensity and expressed as ratios of target gene to normalizer (G3pdh).

Next, we performed quantitative real-time PCR (qPCR) using qPCR primers for Ornt1, Ornt3, TATA box-binding protein (Tbp), and ribosomal protein L27 (Rpl27). We obtained validated primers from Qiagen for Ornt1 and Tbp. To confirm the identity of the amplified region of Ornt1 and Tbp, we TA cloned (Invitrogen) and sequenced the amplified PCR products. The amplified regions of both sets of primers cross an intron-exon junction; namely, intron 6 that interrupts the amplified ORF of Ornt1 (764–873 bp) and introns 7 and 8, which interrupt the amplified region of Tbp (608–722 bp). The primer sequences for Ornt3 (JCUCI-135 and JCUCI-136) and Rpl27 (JCUCI-141 and JCUCI-142) listed in Table 1 were obtained from the Massachusetts General Hospital and the Harvard School of Medicine Center for Computational and Integrative Biology website (pga.mgh.harvard.edu/primerbank/index.html). For Ornt3, the primers corresponded to the ORF region covering 36 to 109 bp (introns 2 and 3); whereas, for Rpl27, the primers corresponded to the ORF region covering 64 to 164 bp (intron 3). Optimization of each primer pair was carried out using serial dilutions and a range of annealing temperatures. The amplification efficiency of all primers was close to 1.0. We carried out first strand synthesis in duplicate reaction volume as mentioned earlier for RT-PCR. All samples were transcribed at the same time to minimize variation. We obtained nucleic acid concentrations on each cDNA sample and used 100 ng of cDNA template in each 25 μL qPCR reaction using the iQ SYBR Green Supermix (BioRad) in accordance with the manufacturer's recommendation. A final primer concentration of 500 nM was used in all reactions. We normalized CT (threshold cycle) results with two different housekeeping genes, Tbp and Rpl27. As documented by other investigators, we observed that, although Tbp and Rpl27 produced a relatively constant expression across mouse tissues, slight differences between reference genes yielded differences in relative quantification among tissues after normalization and comparison to the heart, which was used as a calibrator (10).

Generation of mouse and human expression constructs.

We first cloned the wild-type human and mouse ORNT3 cDNAs into the EcoR1/NotI site of pcDNA3.1 (Invitrogen). We next generated the N-myc (EQKLISEEDLL)-tagged human and mouse ORNT3 and mouse UCP-like (UniGene-Mm.202653) constructs, the latter which served as a negative control, using the primers outlined in Table 1 and a pcDNA3.1 backbone as previously published (6). To generate both tagged and untagged Ornt3 and ORNT3 constructs, we used the mouse AA445192 (1,400 bp with ORF = 918 bp) and human BE791151 (1,440 bp with ORF = 909 bp) cDNAs as templates, respectively. For the mouse Slc25a45 construct, we used a 1,453 bp cDNA (AW988159, ORF = 864 bp) obtained from Invitrogen and subcloned it into the EcoRI/NotI site of pcDNA3.1. The mouse UCP-like (ORF = 310aa) N-myc construct was generated in a one-step RT-PCR (Qiagen) reaction with mouse liver total RNA as a template following the manufacturer's recommendations. The mouse Cact expression construct has been previously reported (6). All constructs were sequenced as previously published (6).

Transfections, ornithine transport assay, and indirect immunofluorescence.

For these experiments, we used established transformed primary fibroblasts that carry the ORNT1-F188Δ allele, which produces an unstable protein, or the wild-type ORNT1 allele, which served as a positive control (1). The UCI Institutional Review Board approved the use of human fibroblast cell lines. We cultured National Institutes of Health-3T3 fibroblasts and mouse cerebellar type II astrocytes in media containing DMEM, sodium pyruvate (1 mM), and 10% FCS. We performed electroporation, the ornithine transport assay, and immunofluorescence as previously described (6,11). Briefly, in the ornithine incorporation assays, transfected cells plated on glass coverslips were incubated for 6 h at 37°C in earle's balanced salt solution containing 0.1 μCi/mL each of 3H-leucine [L-(3,4,5-3H(N)), 173 mCi/mmole] and 14C-ornithine [L-(1-14C), 47.70 mCi/mmole] (Perkin Elmer, Boston, MA). The final ornithine concentration of 2.1 μM is below the Km obtained for the purified human ORNT1 and cultured wild-type fibroblasts (220 μM) and human liver (80 μM) (8,12). For transient transfection studies, we assayed three to six wells per transfected plasmid and expressed results as mean ± SD. We expressed the ability to incorporate 14C-ornithine into protein as glutamate and proline relative to total protein synthesis (leucine-3H) as 14C/3H ratio. We analyzed data using the unpaired t test and presented our results in graphs as percent increase compared with control (untransfected) cells. Transfection efficiency varied between 25 and 30%.

RESULTS

Functional characterization of SLC25A29 (ORNT3).

Previously, we hypothesized that both the SLC25A29 and SLC25A45 were potential genes with overlapping function to ORNT1 based on phylogeny and the conservation of amino acid sequences P29-F/L-D-T-Xaa-K-V-R/K (6). To test our hypothesis, we studied whether the human and mouse SLC25A29 constructs and murine Slc25a45 could rescue the deficient ornithine metabolism by using an established ornithine transport assay in transformed skin fibroblasts from control and HHH patients (6). For these experiments, we used the mouse UCP-like gene and Cact as negative controls and the human N-myc-ORNT2 and mouse N-myc-Ornt1 (wild-type) constructs as positive controls (6). Results from overexpression studies shown in Figure 1A clearly demonstrate that both the N-myc-tagged human SLC25A29 and mouse Slc25a29 constructs rescue the deficient ornithine transport by approximately 5-fold relative to untransfected HHH patient fibroblasts. Similar results were obtained using the untagged SLC25A29 (data not shown) and Slc25a29 (Fig. 1B) constructs. Hence, SLC25A29, hitherto also known as CACL, was identified as the third human mitochondrial ornithine transporter, ORNT3. Conversely, ornithine incorporation assay results in Figure 1B illustrate that neither the mouse Cact, Slc25a45 nor UCP-like protein (data not shown) is capable of restoring the ornithine transport that is deficient in cultured HHH fibroblasts.

Figure 1
figure 1

Transient expression of human and mouse ORNT3 constructs restore ornithine metabolism in cultured HHH fibroblasts. Transformed HHH fibroblasts carrying the nonfunctional ORNT1-F188Δ allele were used to study the ability to incorporate 14C-ornithine and 3H-leucine 48 h after electroporation. The data are expressed as the ratio of 14C/3H in precipitated protein and represent the mean ± SD. For all studies, we electroporated fibroblasts with 25 μg of N-myc tagged or untagged construct or empty vector (pcDNA3.1). (A) Graph compares the ornithine incorporation of untransfected control and HHH-F188Δ fibroblasts vs. HHH-F188Δ fibroblasts transfected with either vector, N-myc tagged murine Ornt1, Ornt3, or human ORNT3 constructs. (B) Data show the ornithine metabolism of untransfected control and HHH-F188Δ fibroblasts vs HHH-F188Δ fibroblasts transiently transfected with either vector, N-myc tagged human ORNT2 or untagged mouse Ornt3, Slc25a45, or Cact constructs. Data represent six separate experiments with four measurements per data point per experiment. *p < 0.05 vs. HHH + vector.

Full size image

Concomitant with ornithine transport studies, we performed immunofluorescence to confirm that both the human and mouse N-myc-tagged ORNT3 proteins target to the mitochondria. Wild-type N-myc-tagged Ornt1 served as a positive control. Results shown in Figure 2 demonstrate that both the human ORNT3 (data not shown) and mouse Ornt3 transporters have a normal mitochondrial targeting pattern as evidenced by its localization in multiple, elongated, cytoplasmic structures that coincide with the location of MitoTracker-labeled mitochondria.

Figure 2
figure 2

Ornt3 is localized to the mitochondria. HHH-F188Δ cells overexpressing mouse N-myc-Ornt3 (A–C) or N-myc-Ornt1 (D–F) transporter were prelabeled with MitoTracker (B and E), which stains mitochondria red, and processed for indirect immunofluorescence using a monoclonal FITC-labeled anti-myc antibody (A and D). Overlay of the images (C and F) show the mitochondrial localization of both transporters. Bars denote 20 μm.

Full size image

Comparison of Ornt3, Slc25a45, Ornt1, and CACT protein structure.

In humans, the subfamily of mitochondrial carrier proteins that includes ORNT1, ORNT2, CACT, SLC25A45, and ORNT3 is characterized by the transport of charged amino compounds such as ornithine, arginine, carnitine, and citrulline (8). Interestingly, in the mouse, the Ornt2 gene is nonfunctional because it is not translated unlike the functional ORNT2 present in the human and other mammalian species (6). Moreover, there is very high amino acid sequence similarity between human and mouse orthologues of this subfamily. Given that Ornt3 has overlapping functions with Cact and Ornt1, we thought it important to compare the protein structure of four members of this subfamily of mitochondrial transporters. As shown in Figure 3, comparison of the amino acid sequences of the mouse Ornt1, Cact, Ornt3, and Slc25a45 demonstrate that the regions of the first hydrophilic loop and the fourth, fifth, and sixth transmembrane domains (TMDs) contain the highest number of conserved amino acids. Interestingly, recently published reports have demonstrated that the region previously thought to be the beginning of the first, third, and fifth hydrophilic loop (after the proline in the mitochondrial motif) actually forms part of a TMD (8). At the amino acid level, Ornt3 is 34% identical to Ornt1 and 37% identical to Cact (9). Most importantly, some of the conserved amino acids (Arg and Asp/Glu) in the fourth and sixth TMD that are thought to play a potential role in solute recognition in Cact and Ornt1 are also conserved in Ornt3 (13). In the case of human ORNT1, several of these conserved amino acids (Arg179, Glu180, and Arg275) were mutated in patients with HHH syndrome (1,8). Conversely, though Glu77 has been implicated as part of the ornithine binding site of Ornt1, there is no such conserved amino acid in the second TMD of Ornt3 (13). Lastly, similar to Ornt1 and most mitochondrial carrier proteins, Ornt3 has consensus sites for serine/threonine phosphorylation by protein kinases, such as PKC (S/T-Xaa-R/K) (3).

Figure 3
figure 3

Sequence alignment of the mouse mitochondrial carrier proteins Ornt1, Ornt3, Slc25a45, and Cact. Amino acids that are conserved relative to Ornt1 are highlighted in black. Filled lines (Roman numerals I–VI) indicate the predicted TMDs. The mitochondrial carrier signature motif [P-h-D/E-Xaa-h-K/R-Xaa-K/R-(20–30aa)-D/E-G-(4Xaa)-a-K/R-G (h, hydrophobic; Xaa, any amino acid)] begins in the region of the first, third, and fifth TMD of each protein. Amino acids in asterisks (*) for Ornt1 and Cact are predicted to be involved in solute recognition.

Full size image

Finally, examination of the genomic structure of the Ornt3 revealed that it has three introns interrupting the ORF and, surprisingly, the region corresponding to TMD II-VI is contained within a single exon, exon 4. This is in sharp contrast to Ornt1, Cact, and Slc25a45 that have several introns (five for Ornt1 and Slc25a45; eight for Cact) that interrupt the ORF of the structural genes of these three transporters.

mRNA tissue distribution of mouse Ornt3.

We initially used semi-quantitative RT-PCR to investigate the mRNA tissue distribution of Ornt3 vs. that of Ornt1, Cact, and Slc25a45. We compared CNS derived tissues (whole brain, cerebellum, cortex, hypothalamus, and hippocampus) with peripheral tissues such as liver, kidney, heart, and testis. As shown in Figure 4, the expression of Ornt1 is much more predominant in the liver, kidney, and testis when compared with the CNS and the heart; whereas, for Ornt3, after normalization with G3pdh, the level of mRNA expression in brain derived tissues and is higher relative to heart, liver, kidney, and testis. The distribution of Cact mRNA follows the pattern expected for tissues heavily dependent on fatty acid oxidation such as the liver, kidney, testis, and heart. Interestingly, the mRNA distribution pattern of the Slc25a45 gene was similar to that of Cact, except for the heart that exhibited low Slc25a45 expression. Slc25a45 was also previously shown to be expressed in brain tissues (cortex and pons) of the rat (7). Finally, using similar RT-PCR conditions, we obtained a restricted expression in the liver for Otc (data not shown).

Figure 4
figure 4

Expression pattern of Ornt1, Ornt3, Slc25a45, and Cact across mouse tissues. RT-PCR amplification was performed using gene specific primers and commercially obtained total RNA (DNase treated, RNase-free) from different murine tissues. Panel A represents 20 μL of final PCR product run on a 2% agarose gel. Panel B represents the ratio of target gene cDNA/G3pdh cDNA for each mouse tissue. This figure is representative of experiments performed in duplicate from two different first strand amplifications using 200 ng cDNA per 25 μL PCR reaction.

Full size image

Given the relatively higher level of Ornt3 expression in CNS derived tissue, we next performed a more accurate determination using quantitative real time PCR. For these experiments, we obtained the relative quantification of Ornt3 and Ornt1 using Tbp and Rpl27 as reference genes and the heart as a calibrator. Results shown in Figure 5 demonstrate marked Ornt1 expression in the liver (27-fold–37-fold) and kidney (∼6-fold) relative to the heart, CNS tissues, and cultured murine astrocytes. A similar expression pattern was previously obtained using Northern blot analysis (1). The Ornt3 mRNA, on the other hand, is significantly expressed in CNS tissue, heart, and cultured astrocytes relative to the liver and kidney albeit at a much lower fold difference. In general, the increase in Ornt3 expression in CNS tissues varied from 1.35-fold in the whole brain to 4.8-fold in the hippocampus relative to the heart (1.0-fold) using Tbp as a normalizer. These results confirm and expand previously published Northern blot results of Slc25a29 in mouse brain and heart (9). A similar pattern of expression was also observed for both Ornt1 and Ornt3 after normalizing with Rpl27 (data not shown) except that Ornt3 fold differences seen in CNS derived tissues were more pronounced (i.e., 3.8-fold in brain and 5.15-fold in hippocampus) with the cerebellum (8.2-fold), not the hippocampus, showing maximum expression. Again, these results were relative to the heart, which was used as a calibrator in all real time PCR experiments (Fig. 5).

Figure 5
figure 5

Real time PCR amplification of mouse Ornt1 (white) and Ornt3 (black) from peripheral and CNS tissues and cultured mouse astrocytes. We performed qPCR for Ornt1, Ornt3, and Tbp using validated primers and a BIORAD MiniOpticon Real Time Thermocycler. For SybrGreen based amplification, we used 100 ng of first strand DNA synthesized from 2 μg of total RNA. The CT values were normalized to those of Tbp following the ΔΔCT method, and the heart was used as a calibrator (*). Figure represents three different experiments performed in duplicate.

Full size image

Ornithine metabolism in mouse cerebellar astrocytes.

Because Ornt3 is a mitochondrial ornithine transporter that is significantly expressed in CNS tissues, we thought it important to investigate the level of ornithine transport in cultured mouse cerebellar type II astrocytes, which have a lower expression level of Ornt1 message relative to the liver and kidney and lack a functional Ornt2. Results show that ornithine transport for both astrocytes and mouse National Institutes of Health-3T3 fibroblasts are similar but vary from 30 to 40% when compared with established wild-type human fibroblasts (data not shown). The reason for the difference between human and mouse cells may relate to the differences in metabolic rate between species, variable OAT activity, or the fact that only two functional ornithine transporters are present in mouse vs. three in the humans.

DISCUSSION

In this work, we demonstrate that both the human and mouse SLC25A29 genes have overlapping function with ORNT1 and is therefore the third known human mitochondrial ornithine transporter, ORNT3. The ability of SLC25A29 to also function as carnitine/acyl-carnitine translocator-like (CACL) should not come as a surprise given that it is a member of the ORNT1, ORNT2, CACT, and SLC25A45 subfamily of mitochondrial carrier proteins, which transports positively charged amino compounds. Although the function of SLC25A45 is unknown, we speculate that it will function as a mitochondrial choline transporter given the very close structural similarity of choline to carnitine (13). Contrary to humans, there are only four functional subfamily members in mice and yeast. In mice, Ornt2 is nonfunctional; whereas, in yeast, no ORNT2 exists. The yeast subfamily is composed of ARG11 (ORNT1 orthologue), CACT, YMC1, and YMC2 (14). Surprisingly, although yeast lacks ORNT2, the existence of a second (redundant) mitochondrial cationic amino acid transport system was reported in the arg11 KO (14). This suggests that, throughout evolution, gene redundancy for cationic amino acid transport has existed to provide an advantage for species to adapt to metabolic compartmentalization given the very fundamental role of these amino acids transporters in bioenergetics, ammonia detoxification, mitochondrial protein synthesis, and polyamine biosynthesis. Lastly, additional support is the recent characterization of two cationic amino acid transporters (AtmBAC1 and AtmBAC2) in Arabidopsis thaliana (15).

SLC25A29's redundant function to CACT, on the other hand, seems to be a more recent evolutionary event. This assertion is based on the observation of lethal phenotypes of Caenorhabditis elegans (dif-1) and Drosophila melanogaster (Colt) Cact KOs and on the absence of residual transport activity in a yeast Cact KO (Δyor100c) (9,16). Although it is clear that under experimental conditions SLC25A29 works as CACL, it seems that its expression in heart, liver, and brain is not sufficient to have a physiologic effect in the great majority of patients with CACT deficiency who usually succumb to cardiomyopathy and liver dysfunction early in life. In addition to its role in amino acid metabolism, it is possible that SLC25A29 plays a role during embryonic and fetal development because CACT deficient patients do not show any developmental defects unlike the C. elegans dif-1 and D. melanogaster colt mutants. Thus, we speculate that SLC25A29 may also have an important role in structural lipid turnover especially in the brain and heart where it is expressed (9,17).

From a structure-function perspective, it is important to understand how ORNT3 can have overlapping functions with both ORNT1 and CACT. The recent elucidation of the bovine ADP/ATP-1 crystal structure and follow-up work by others has provided some insight into how these transporters may work at the molecular level (8,13). Published conclusions about the solute recognition sites (second, fourth, and sixth TMDs) for both the ornithine and carnitine transporters were deduced using the amino acid structure of the yeast orthologues of these two mitochondrial carriers and other mitochondrial carrier proteins. Given that the findings on the ADP/ATP-1 are based on a monomeric structure, there is always room for subtle but significant differences in these transporters' molecular structure if we consider that they are thought to work as homodimers or possibly as heterodimers (i.e., ORNT1-ORNT3). Perhaps differences in a combinatorial process (protein-protein interactions) mediated in part by posttranslational modifications may explain differences in the phenotypic variability of patients with HHH syndrome (3,5).

At the mRNA level, Ornt3 seems to have higher expression in the CNS when compared with the liver, thus suggesting a potential role for ORNT3 in the CNS cellular pathophysiology of HHH syndrome (9). Although speculative, it may be that ORNT3 serves a compensatory role in CNS tissues by metabolizing excess ornithine accumulated from an ORNT1 deficiency; however, this process could potentially disrupt its additional role in the CNS metabolism of fatty acids with a secondary effect on brain growth and development. Although the CNS has traditionally been thought of as an organ dependent on ketone bodies, glucose, and lactate as energy sources necessary for growth and development, this notion is starting to change given recent reports, which detail the expression and functionality of components of the fatty acid oxidation pathway in astrocytes (17,18). Our work demonstrated that cultured cerebellar type II mouse astrocytes possess an ornithine transport system that should be physiologically relevant given the fundamental role that cationic amino acids play in brain function. For example, recent work, using a glutaryl-CoA dehydrogenase KO mouse model, highlighted the importance of ORNT1 in brain physiology as it relates to the mitochondrial entry of lysine before its degradation into glutaric acid (19).

To understand the mechanisms of disease underlying the CNS dysfunction variability in patients with HHH syndrome who are under adequate metabolic control, a precise understanding of the biologic functions of ORNT1, ORNT2, and ORNT3 are required. Because ORNT3 transports ornithine in vitro, localization of ORNT3 in the periportal hepatocytes will illuminate its potential role in the urea cycle in addition to its likely role in the ornithine oxidation pathway. Liposomal and KO animal studies, however, are needed to elucidate the physiologic relevance of the dual identity of SLC25A29.