Promoter Rearrangements Cause Species-specific Hepatic Regulation of the Glyoxylate Reductase/Hydroxypyruvate Reductase Gene by the Peroxisome Proliferator-activated Receptor α*

In liver, the glyoxylate cycle contributes to two metabolic functions, urea and glucose synthesis. One of the key enzymes in this pathway is glyoxylate reductase/hydroxypyruvate reductase (GRHPR) whose dysfunction in human causes primary hyperoxaluria type 2, a disease resulting in oxalate accumulation and formation of kidney stones. In this study, we provide evidence for a transcriptional regulation by the peroxisome proliferator-activated receptor α (PPARα) of the mouse GRHPR gene in liver. Mice fed with a PPARα ligand or in which PPARα activity is enhanced by fasting increase their GRHPR gene expression via a peroxisome proliferator response element located in the promoter region of the gene. Consistent with these observations, mice deficient in PPARα present higher plasma levels of oxalate in comparison with their wild type counterparts. As expected, the administration of a PPARα ligand (Wy-14,643) reduces the plasma oxalate levels. Surprisingly, this effect is also observed in null mice, suggesting a PPARα-independent action of the compound. Despite a high degree of similarity between the transcribed region of the human and mouse GRHPR gene, the human promoter has been dramatically reorganized, which has resulted in a loss of PPARα regulation. Overall, these data indicate a species-specific regulation by PPARα of GRHPR, a key gene of the glyoxylate cycle.

As a major survival strategy, animals have developed metabolic pathways to store energy when food is abundant, which allows them to overcome periods of deprivation. Thus, they are able to switch from efficiently storing excess energy to its rapid mobilization to keep the organism alive when food is not available. Hormonal changes during fasting result in enhanced triglyceride hydrolysis in the adipose tissue, as well as glucose and ketone body production in the liver, to respond to the energy needs of the peripheral organs. Peroxisome proliferator-activated receptors (PPARs), 1 as members of the nuclear hormone receptor superfamily, are involved in the transcriptional control of these metabolic pathways, including ␤-oxidation, gluconeogenesis, and amino acid catabolism, which underscores their importance in energy homeostasis (1)(2)(3)(4). Upon activation by fatty acids and derivatives, they heterodimerize with the receptor for 9-cis-retinoid acid (retinoid X receptor, NR2B) and bind to peroxisome proliferator response elements (PPREs) in the promoter region of the genes whose expression they regulate. Three PPAR isotypes have been identified: PPAR␣ (NR1C1); PPAR␤/␦ (NR1C2); and PPAR␥ (NR1C3). They exhibit different expression patterns and functions in animals (1). PPAR␥ plays an important role in inflammation, lipid storage, and glucose homeostasis, whereas PPAR␤/␦ is important for skin functions, brain, and placenta development and also for energy homeostasis (3). PPAR␣, which is a part of this study, regulates peroxisomal and mitochondrial fatty acid oxidation, microsomal fatty acid hydroxylation, lipoprotein metabolism, bile and amino acid metabolism, glucose homeostasis, biotransformation, inflammation, hepatocarcinogenesis in rodents, and other pathways and processes (5). In particular, this PPAR has been implicated in the regulation of the expression of two enzymes involved in the glyoxylate pathway, namely alanine:glyoxylate aminotransferase (AGT) and glyoxylate reductase/hydroxypyruvate reductase (GRHPR) (2). AGT has a dual function, as alanine:glyoxylate aminotransferase and as serine: pyruvate aminotransferase, and is responsible for the conversion of glyoxylate into glycine and for the conversion of serine to hydroxypyruvate, respectively (see Fig. 1). Similarly, GRHPR functions both as glyoxylate reductase and as hydroxypyruvate reductase. GRHPR plays a key role in directing the carbon flux to gluconeogenesis by its ability to convert hydroxypyruvate into D-glycerate (see Fig. 1) (6). Therefore, regulation of this enzyme by PPAR␣ may contribute to the function of the receptor in energy homeostasis.
Linked to their role in metabolism, AGT and GRHPR are associated to primary hyperoxaluria type 1 and type 2, respectively, which are caused by an overproduction of oxalate. Oxalate is an inorganic acid that, when combined with calcium, produces insoluble calcium oxalate, which is the most common constituent of kidney stones (7). Deficiency in these two enzymes results in glyoxylate accumulation (see Fig. 1) and consequently to oxalate overproduction (8,9). Moreover, a decrease in glycolate oxidase activity and an increase in lactate dehydrogenase activity were observed after clofibrate treatment in rats, which results in an increase in oxalate concentration in urines (10). Therefore, we investigated the involvement of PPAR␣ in the regulation of glyoxylate metabolism and consequently in oxalate production.

MATERIALS AND METHODS
Animals-Mice were housed in a temperature-controlled room (23°C) on a 10-h dark, 14-h light cycle. Pure-bred wild type or PPAR␣ knock-out (11) mice of age 8 -10 weeks were used for all of the experiments. Animal experiments were approved by the Service Veterinaire of the Canton de Vaud. Fasted animals were deprived of food for 24 h starting at the beginning of the light cycle. The PPAR␣ ligand Wy-14,643 (50 mg/kg/day) or vehicle (0.5% carboxymethyl-cellulose) were administrated by gavage for 5 days (RNase protection assay (RPA) chromatin immunoprecipitation (ChIP)) or mixed to the food (0.1%) for 5 days (oxalate assay).
Similarly, the human promoter was amplified from genomic DNA extracted from HepG2 cells. A common downstream primer (5Ј-CCG-CTCGAGCGGCATGAGTCGCACCGGTCTCATC-3Ј) and the upstream primers (5Ј-GGGGTACCCCTCAAGGAAACCAACCCTGGTGC-3Ј, 5Ј-GGGGTCCCCCGGGACTCAGCCACCAC-AACCA-3Ј, and 5Ј-GGGGTA-CCCCGAGGCGGGAGGATCAT-TGGAGCAC-3Ј) were used for the amplification of the fragments subcloned in plasmids p4500-Luc, p3000-Luc, and p2000-Luc, respectively. PCR was performed using the following conditions: 95°C for 4 min; 35 cycles at 94°C for 45 s; 55°C for 45 s; and 72°C for 2 min with a final elongation step at 72°C for 7 min. PCR products were digested with KpnI and SmaI for the mouse fragments and with KpnI and XhoI for the human fragments and subcloned in front of the luciferase reporter gene into the ⌬pGL2 basic vector (Promega).
Primer Extension-␥-32 P-End-labeled primers 5Ј-CATGAGTCGC-GCCGGTTTCATAAGAC-3Ј (mouse) and 5Ј-ACCTGGCAGTACA-GAAGCTGGC-3Ј (human) were used for reverse transcription (RT) and sequencing. The sequencing was carried out using the fmol DNA cycle sequencing system (Promega). 50 g of total RNA was mixed with the labeled primer (3 M) in hybridization buffer (150 mM KCl, 10 mM Tris, pH 8.3, 1 mM EDTA). The mixture was heated at 95°C for 2 min and then placed at 65°C for 1 h and 30 min at room temperature for 1 h and 30 min and finally at 4°C overnight. Reverse transcription was then performed with the Superscript II RNase H-reverse transcriptase (Invitrogen).
RNA Extraction and Northern Blot-Total RNA was isolated using the TRIzol reagent (Invitrogen). Northern blot analysis was performed using 30 g of total RNA according to standard protocols (12). Mouse GRHPR cDNA probe was random-primed-labeled using the High Prime kit (Roche Applied Science).
RT-PCR/RPA-Gene-specific probes for the mouse GRHPR, AGT, and lactate dehydrogenase A were obtained by RT-PCR from mouse liver total RNA and cloned into the pGEM T-Easy vector (Promega). RT-PCR was performed with the Titan One Tube RT-PCR system (Roche Applied Science). 100 ng of total RNA was used for each PCR. The number of cycles was first determined to find the exponential phase of each set of primers.
Gene-specific antisense riboprobes were synthesized by in vitro transcription with either T7 or Sp6 RNA polymerase (Ambion). For all of the riboprobes with the exception of L27, a ratio of 1:1 of [␣-32 P]UTP to cold UTP was used, whereas a ratio of 1:20 was used for L27 probe.
RPA was carried out using the Direct Protect lysate RPA kit (Ambion) with the following modifications. 15 g of total RNA were resuspended in 45 l of lysis buffer and incubated with 5 ϫ 10 4 cpm of gene-specific and L27 riboprobes. RPA products were resolved in a 6% electrolyte gradient-denaturing polyacrylamide gel. Gels were dried and exposed to x-ray film.
Measure of Plasmatic Oxalate in Mice-Mouse plasma was acidified by dilution 1:1 with 0.15 N HCl to dissociate oxalate from plasmatic proteins. Plasmatic proteins were then removed by ultrafiltration and by centrifugation at 14,000 ϫ g (1 h, room temperature) in a Microcon YM30 column (Amicon). The measure of oxalate in ultrafiltrate was done using the oxalate oxidase/peroxidase method (Kit Oxalate, Dialine) on a Cobas FARA automate (Roche Applied Science).

RESULTS
Characterization of the Mouse GRHPR Gene-Using SABRE (selective amplification via biotin and restriction-mediated enrichment) (14), we previously isolated the cDNA of a novel target gene for PPAR␣ whose identity and regulation are studied herein. The full-length cDNA was amplified using 5Ј-and 3Ј-RACE PCR, cloned, and sequenced. A comparison of these mouse cDNA and derived protein sequences with those present in the NCBI data base indicated a very high similarity with the human GRHPR sequence. This analysis also revealed the presence of a region presenting a high homology with the D-isomer 2-hydroxyacid dehydrogenase consensus motif (PROSITE accession number PS00671) and a putative binding site for nicotinamide adenine dinucleotide (NAD). These characteristics suggested that the gene encodes a protein with an enzymatic activity, most likely mouse GRHPR. The amino acid sequence analysis revealed a Leu-Lys-Leu motif at the protein C terminus, which is close to the consensus motif of the peroxisomal targeting signal-1 that typically consists of these three amino acids or a conservative variant thereof (15). This observation suggested that the cloned cDNA might encode a peroxisomal protein.
To determine the subcellular localization of this gene product, fusion proteins with the green fluorescent protein (GFP) localized either at the C-terminal (GRHPR N2) or at the N-terminal (GRHPR C2) end of GRHPR were generated. When transfected into HEK 293 cells, the constructs produced cytoplasmic fusion proteins (Fig. 2). As a control, a GFP fusion acyl-CoA oxidase (ACO) was targeted to peroxisomes as expected, whereas the mutation of the peptide signal Ser-Lys-Leu into Leu-Lys-Leu, as found in GRHPR, resulted in a cytoplasmic localization of the mutated ACO. These results suggested that the cloned cDNA encodes a protein localized in the cytoplasm. We then examined the expression profile of the gene in different organs by Northern blot analysis. As shown in Fig. 3, high expression levels were detected in the liver and a much lower expression was also found in the kidney. Thus, based on the high percentage of identity, cytoplasmic localization, identical gene organization (see below), and expression profile in liver and kidney as human GRHPR (16), we concluded that the identified gene encodes the mouse GRHPR.
We then investigated the genomic organization of this GRHPR gene. Genomic DNA was obtained by screening a mouse -phage genomic library using two probes: one from the 5Ј end (nucleotides 40 -370) and one from the 3Ј end (nucleotides 938 -1237) of the cDNA. Two overlapping clones were identified containing the entire gene and ϳ10 kb of DNA upstream of the translation initiation codon ATG (data not shown). After sequencing, a comparison of the cDNA and the genomic sequences allowed determination of the organization of the gene (Fig. 4A), which was recently mapped on mouse chromosome 4 B2 (17). The gene spans over ϳ10 kb and is composed of 9 exons. The different exon lengths range from 73 (Exon 3) to 321 bp (Exon 9) in size corresponding to a total coding region of 987 bp for a transcribed region of 9369 bp (see FIG. 2

. Cellular localization of the GRHPR protein expression.
Expression vectors for GRHPR-GFP fusion proteins with GFP at the C terminus (GRHPR N2) or N terminus (GRHPR C2) were transfected in HEK 293 cells. 5 h after transfection, the culture medium was changed and 19 h later the localization of GRHPR-GFP was observed by fluorescent light microscopy. GFP-ACO located in peroxisomes and GFPmutated ACO (mACO) located in the cytoplasm were used as controls. below). The ATG start codon was located in Exon 1, 80 bp downstream of the transcription initiation site (see below). The last exon, Exon 9, contains the stop codon TAA and a 3Ј-UTR (3Ј-untranslated region) of 199 bp in which a polyadenylation signal, AUUAAA, was found 19 bp upstream of the poly(A) tail. The eight introns are ranging in size from 489 bp (Intron III) to 1568 bp (Intron VII). All of the exon-intron boundaries have the canonical GT/AG motif with the exception of the splice donor of Exon 2, which has a non-canonical site TA/AG (Fig. 4B).
A comparison with the human GRHPR gene localized on chromosome 9q12 (18) revealed an identical genomic organization with eight introns flanked by nine exons. The location of the eight introns is strictly conserved, and the length of the coding sequence of the human gene is the same (987 bp) as that of the mouse gene (Fig. 4B). However, the length of the introns varies between the two species. Furthermore, the 5Ј-UTR (106 bp) and the 3Ј-UTR (207 bp) of the human transcript are longer than in the mouse (80 and 199 bp, respectively).
Characterization of the Promoter of the Mouse and Human GRHPR Genes-Firstly, the transcription initiation site of the two genes was determined. Primer extension analysis showed that the transcription start sites were 24 bp upstream in mouse (Fig. 5A) and 26 bp upstream in human (Fig. 5B) of the 5Јends of cDNAs contained in GenBank TM data base (accession numbers BY353228 and CB998056). In human, two primer extension products separated by only 1 bp were detected, possibly reflecting 5Ј-cap methylation of the mRNA, which may cause incomplete reverse transcription of some of the mRNA molecules.
Secondly, the alignment of the two promoter sequences showed a surprisingly poor homology at first sight but further analysis revealed that the proximal region of the mouse promoter (ϳϪ300 to ϳϪ2000) is found in a reverse orientation in the human gene 4 kb upstream the transcription initiation site (ϳϪ3900 to ϳϪ5600) (Fig. 5C). Interestingly, the functional PPRE in the mouse promoter (mPPRE1) (see below) was also found but less conserved in human in this shifted region (Fig.  5C). To know whether this displacement and inversion of promoter region was specific to human, the GRHPR promoter of other species was also analyzed. Comparison with the rat (Ensemble data base accession number RNOR03291619), dog (Ensemble data base contig 55728.1.64020), and chimpanzee (Ensemble data base accession number AADA01236509) promoter showed that the rearrangement is also present in the chimpanzee but not in the rat and dog promoter (Fig. 5D). The alignment in the same orientation of the rearranged promoter fragment from the mouse, rat, chimpanzee, and human promoter revealed conserved regions among the four species, which may suggest that regulatory functions have been maintained after reconfiguration of the promoter in primates (Fig.   6). Interestingly, however, the functional PPRE identified in the mouse sequence (see below) was not conserved in any of the other species, not even in the rat where the half-sites of the response element are separated by 12 nucleotides (Fig. 6). Further analysis of the promoter demonstrated a high level of sequence repeats in the human promoter between the transcription initiation site and the boundary of the inverted region. Indeed, this region contains 64% repeated elements (41.5% short interspeed elements, 16.5% long interspeed elements, and 6% long terminal repeat elements), whereas the average of the human genome is 46%. In comparison, only 21% of the mouse promoter region is constituted of such elements (21% short interspeed elements), which is under the average of 37.5% for the complete genome (17). Interestingly, the percentage of repeats in the shifted region is 25 and 20% for the human and the mouse, respectively. These results suggest that even though the shifted region may have some regulatory function, the promoter region in primates underwent dramatic rearrangements also attested by its high content in repeated ele- Regulation of the Human and Mouse GRHPR Gene by PPAR␣-However, an analysis of the human and mouse 5Ј region over 4500 bp revealed several putative PPREs (NUBIScan algorithm) (19). The putative PPREs located between Ϫ633 and Ϫ621 (mPPRE1: GGGTTA A AGGTCA) and between Ϫ2363 and Ϫ2351 (mPPRE2: AGGTTA C AGGTGG) for the mouse and between Ϫ2172 and Ϫ2184 (hPPRE1: TG-GCCT G TTCCCA) and between Ϫ4159 and Ϫ4171 (hPPRE2 AGGGCA T TGGGCA) for the human promoter (Fig. 5C) suggested that GRHPR may be directly regulated by PPAR␣. These PPREs exhibit one (mPPRE1), four (mPPRE2), or three (hPPRE1; hPPRE2) mismatches (boldface and italics) when compared with the consensus PPRE sequence 5Ј-AGGNCA A AGGTCA-3Ј. Thus, only the mPPRE1 corresponds well to the consensus PPRE.
Regulation of the human GRHPR gene by PPAR␣ was initially tested using semi-quantitative RT-PCR in HepG2 cells. These cells, treated either with Wy-14,643 (10 M) or vehicle (Me 2 SO), showed no significant increase in GRHPR mRNA, whereas the expression of the known target gene of PPAR␣, FIAF (20), was increased by more than 2-fold by the PPAR␣ ligand (Fig. 7, A and B). To further examine possible PPAR␣dependent regulation of the human GRHPR promoter despite this negative result, transient transfections were performed. Three fragments of the promoter were subcloned upstream of the luciferase reporter gene. These constructs were transiently transfected in HepG2 cells. The empty vector (pGL2) and a reporter construct containing three PPREs served as negative and positive controls, respectively. The GRHPR promoter appeared functional in terms of basal activity, because transcription from all three constructs was higher than that seen with the parental promoter less reporter construct (Fig. 7C). However, ligand activation of transfected PPAR␣ failed to stimulate promoter activity in all of the constructs independently of the presence of the putative PPREs, indicating that they are not functional. Unexpectedly, the addition of PPAR␣ and Wy-14,643 down-regulated promoter activity for so far unknown reasons. As functional PPREs have been found located downstream of the transcription start site in some genes (13), the transcribed region was also scanned for putative PPREs. Three of them were identified in Intron 1 and 3 (hPPRE3, hPPRE4, and hPPRE5, see Fig. 7D). They were located between 1620 and 1632 bp (hPPRE3: TATCCT G TGTCCT), 2022 and 2034 bp (hPPRE4: TCACCT G TCACCT), and 3492 and 3504 bp (hPPRE5: TGTGCA C AGGTCA), and all presented three mismatches with respect to the consensus PPRE. To assess their functionality in vivo, ChIP was performed on HepG2 cells (Fig. 7D). The human FIAF PPRE was used as a control (20). A positive PCR amplification of ChIP indicated the binding of activated PPAR to the functional FIAF PPRE, which was significantly increased after stimulation of the cells with Wy-14,643. The DNA sequences containing both hPPRE4 and hPPRE5 were not amplified at all, whereas hPPRE3 showed some background level amplification in the cells transfected with PPAR␣ and treated with Me 2 SO. Transfection of a con- struct containing this hPPRE3, subcloned either upstream (p1620-Luc) or downstream (pLuc-1620) of the thymidine kinase-Luc reporter construct, failed to respond to PPAR␣ stimulation, confirming that this putative PPRE was not functional (Fig. 7D). The empty vector (pGL2) and a reporter construct containing three PPREs served as negative and positive controls, respectively. Altogether, these results suggest that the human gene is unlikely to be regulated by PPAR␣, at least in HepG2 cells in which PPAR target genes can be up-regulated as seen with the FIAF gene that served as positive control.
On the contrary, the mouse GRHPR gene was clearly regulated by PPAR␣. Indeed, the GRHPR mRNA level was upregulated in mouse liver either by fasting, which enhances PPAR␣ activity (21), or by treatment of the animals with Wy-14,643, whereas no difference was observed in the knock-out animals (Fig. 8, A and B). To determine which of the two putative PPREs identified in the promoter region of the mouse gene is functional, transient transfections in HepG2 cells were performed with different fragments of the promoter region (p2500-Luc, p1100-Luc, and p400-Luc) subcloned in front of the luciferase reporter gene (Fig. 8C). The results showed that the p400-Luc construct with no PPRE was sufficient to achieve basal promoter activity as expected (3.5-fold higher expression than the empty vector pGL2) but was neither responsive to the PPAR␣ ligand Wy-14,643 nor to cotransfection of exogenous PPAR␣, whereas the two other constructs responded to PPAR␣ stimulation. Taken together, this finding suggested that at least the mPPRE1 is functional.
In vivo binding of PPAR␣ to the chromosomal mPPREs was further verified by ChIP on the liver chromatin of wild type (PPAR␣ ϩ/ϩ ) and knock-out (PPAR␣ Ϫ/Ϫ ) mice, treated or not with Wy-14,643. The results showed a slight amplification of the DNA sequence spanning the PPRE1 but not PPRE2 in the PPAR␣ ϩ/ϩ mice, whereas no band was found in the PPAR␣ Ϫ/Ϫ mice (Fig. 8D). This amplification of PPRE1 was highly increased when the mice were treated with Wy-14,643. This result indicated that PPAR␣ binds only on PPRE1 that is closest to the consensus PPRE. The amplified band is specific, because no signal was obtained with preimmune serum. Taken together, both transfection assay and ChIP demonstrated that the mouse GRHPR gene is a direct PPAR␣ target gene and contains a functional PPRE in its promoter.
Effect of the PPAR␣ Agonist on Oxalate Production-As mentioned above, mutations in the GRHPR gene may result in the accumulation of oxalate, leading to the formation of kidney stones (9). The oxalate levels are regulated by many processes such as oxalate absorption, production (glyoxylate cycle and acid ascorbic breakdown), and secretion. Furthermore, oxalate synthesis is directly related to the pool of glyoxylate. Intravenous injection of glyoxylate showed that 65-80% of it is converted into glycine by AGT, 15-30% is metabolized to oxalate by lactate dehydrogenase, and 2-5% is reduced to glycolate by GRHPR (Fig. 9A) (23). Although the contribution of the glyoxylate reductase activity of GRHPR seems to play a minor role in the reduction of the glyoxylate pool by the conversion of glyoxylate into glycolate, its hydroxypyruvate reductase activity may have a greater effect on the glyoxylate level by targeting the flux of the cycle toward gluconeogenesis. Changes in expression FIG. 7. Regulation of the human GRHPR gene by PPAR␣. A, relative expression level of GRHPR mRNA in HepG2 cells. Total RNA (100 ng) from cells treated (10 M Wy-14,643) (right) or untreated (left) was used for RT-PCR reactions using GRHPR and FIAF primers. The products were analyzed on gels and quantified relative to the levels of L27 mRNA. B, representation of the quantification of the expression of GRHPR and FIAF mRNA expression level, n ϭ 3 Ϯ S.E. *, p Ͻ 0.05. Transient transfection studies with promoter (C) and intronic (E) putative PPREs were performed. HepG2 cells were transfected with different reporter constructs (p4500-Luc, p3000-Luc, p2000-Luc, p1620-Luc, and pLuc-1620) derived from the GRHPR promoter (see "Materials and Methods"). The results were normalized with the ␤-galactosidase activity. Values are the mean of three independent experiments. D, ChIP of PPREs. Chromatin from HepG2 cells, transfected with PPAR␣ and treated or not with Wy-16,463 (Wy), was immunoprecipitated with an antibody against PPAR␣ or with the preimmune serum (PI), and finally, the recovered DNA was amplified using PCR and analyzed on gels. Aliquots of chromatin were analyzed before immunoprecipitation (input). E, putative PPRE3 that showed a low background activity in the ChIP experiment was analyzed further in transfection experiments (see C) of constructs in which it was placed upstream or downstream of the thymidine kinase-Luc reporter gene. The reporter construct containing three PPREs was used as a positive control as in C. DMSO, Me 2 SO. of one of the enzyme may affect the flux of carbon through the cycle and modify the ratio of the different metabolites. For example, an increase of the GRHPR activity may decrease the glyoxylate pool and thus reduce the production of oxalate. Because the GRHPR gene is a target of PPAR␣, we examined whether this receptor has an effect on the production of oxalate and, therefore, measured its concentration in the plasma of PPAR␣ ϩ/ϩ and PPAR␣ Ϫ/Ϫ female and male mice. First, we observed that PPAR␣ Ϫ/Ϫ mice had a higher plasma oxalate levels than the PPAR␣ ϩ/ϩ mice (Fig. 9B). In addition, significantly greater variability in these levels was observed in the PPAR␣ Ϫ/Ϫ compared with the PPAR␣ ϩ/ϩ mice, also suggesting an effect of PPAR␣ on oxalate homeostasis (Fisher test: male, F ϭ 4.9; p Ͻ 0.05; female, F ϭ 6.7; p Ͻ 0.01). Consistent with the above results, mice treated with the PPAR␣ agonist Wy-14,643 showed a 3-fold increase in the mRNA level of GRHPR compared with treated-PPAR␣ Ϫ/Ϫ mice (Fig. 8A). The expression of AGT was decreased by half by Wy-14,643 as observed in many enzymes involved in amino acid metabolism (2), whereas that of lactate dehydrogenase A remained unchanged (Fig. 9C). This finding suggested that the effect of Wy-14,643 on oxalate homeostasis is quite complex, because on one hand reduced AGT expression would increase the glyoxylate pool and on the other hand increased GRHPR level would reduce this same pool. Following Wy-14,643 treatment, plasma oxalate levels were similarly reduced in the mutant and wild type mice (Fig.  9B). Fasting also showed a tendency of oxalate reduction that is observed in wild type and knock-out animals (Fig. 9B). These data highlight a dual effect on oxalate production. First, the increase of oxalate levels in PPAR␣ Ϫ/Ϫ showed that PPAR␣ is involved in lowering oxalate production. Second, the decrease observed after Wy-14,643 in both PPAR␣ ϩ/ϩ and PPAR␣ Ϫ/Ϫ suggested a PPAR␣-independent effect of this ligand. DISCUSSION Herein, we have characterized the structure, the expression profile, and the PPAR␣-mediated regulation of the mouse GRHPR gene. In contrast to the regulation of the mouse gene, the expression of the human GRHPR gene is PPAR␣-independent. This finding underscores a species-specific regulation of this important gene in mammals whose alteration in human causes hyperoxaluria, an autosomal human disorder resulting from an overproduction of oxalate.
A comparison of the mouse and human GRHPR orthologous genes revealed that they share exactly the same genomic organization, nine exons and eight introns with splicing sites located at precisely the same place in both genes. Both the cDNA and the protein sequences have an identity of ϳ86%, whereas the intronic regions are quite different. This level of identity suggests that GRHPR has been well conserved during evolution to an extent similar to the globins. In both mouse and human, GRPHR is cytoplasmic and its expression is highest in the liver (16). Previous reports on human GRHPR have demonstrated its presence also in kidney, fibroblasts, and peripheral blood lymphocytes (22). However, its role in some of these tissues remains to be elucidated (24). Both mouse and human promoter sequences lack a typical TATA-box but are GC-rich, which is characteristic of TATA-less promoters. Interestingly, FIG. 8. Regulation of the mouse GRHPR gene by PPAR␣. A, expression level. RNase protection assay was performed on total RNA (15 g) extracted from the liver of wild type (WT) and knock-out (KO) mice after treatment as indicated. L27 was used as control. B, representation of the quantification of GRHPR mRNA expression levels. C, characterization of the two mouse GRHPR gene putative PPREs. HepG2 cells were cotransfected with different reporter constructs containing the truncated promoter of the gene (p2500-Luc, p1100-Luc, and p400-Luc) as described under "Materials and Methods." The values obtained were normalized with the ␤-galactosidase activity. The means of three independent experiments are shown. D, ChIP. Chromatin from liver of WT and KO mice, treated or not with Wy-14,643, was used for chromatin immunoprecipitation. DNA was incubated with an antibody against PPAR␣ or with the preimmune serum, and the recovered DNA was used for PCR amplification and the reaction products were analyzed on gels. Aliquots of chromatin were analyzed before immunoprecipitation (Input). the proximal region of the mouse promoter was reversed in primates and shifted 4 kb upstream of the transcription initiation site in the human promoter. This shift was not observed in rat and dog, whereas it was present in chimpanzee, suggesting that the rearrangement occurred during primate evolution.
Clearly, the mouse and human GRHPR are differently regulated by PPAR␣. Interestingly, the functional PPRE of the mouse promoter is in the region that has been moved 5Ј and inverted in the human promoter. Sequence comparison of this region showed that the PPRE is not conserved in human. In addition to this translocation, the high level of insertion of repeat elements indicates that the human promoter underwent a profound reorganization, which corroborates the absence of regulation by PPAR␣ in human HepG2 cells. Other differences in the effect of PPAR␣ and its ligands in the liver of rodents and human have already been reported. For example, hepatocarcinogenic effects of PPAR␣ ligands are only seen in rodents (25). Similarly, genes such as ACO and apolipoprotein A-I are differently regulated in rodents and human (26,27), which can be explained, at least in part, by the fact that PPAR␣ is weakly expressed in human compared with mouse liver. Based on the loss of a function in the human GRHPR, it is tempting to speculate that, in the human liver, the levels of GRHPR are controlled by means other than PPAR␣.
Cross-genome comparisons are increasingly used as a screening method for important regulatory regions. In this context, it is commonly believed that sequences conferring a conserved regulatory mechanism to a gene stand out as strongly conserved islands in dot matrix comparisons of related genomes. This approach of identifying regulatory regions has been termed "phylogenetic footprinting" (28). The case of the GRHPR gene reported here is an interesting one and perhaps the first clear example of negative result when applying the phylogenetic footprinting approach. Initially, we expected that the mechanism and cis-acting regulatory elements regulating these GRHPR genes would be conserved between human and mouse. However, in agreement with the fact that the human gene is not induced in a PPAR␣-dependent manner, the PPRE in the mouse upstream region, which confers PPAR␣-inducibility to this gene in mouse, is not conserved in the homologous human upstream region. This shows that the phylogenetic footprinting approach is not only useful for the detection of conserved elements but also for detection of differences in the organization of regulatory regions of orthologous genes.
During fasting, the regulation of GRHPR by PPAR␣ in the mouse may be important for energy homeostasis. During this period, there is a reduction of urea cycle activity and, thus, a diminished urea production ( Fig. 1) (21). The demand for glycine supplied by the glyoxylate cycle is therefore less important. Conversely, the synthesis of glucose and ketone bodies is increased in the liver. These two pathways are controlled by PPAR␣ that represses amino acid catabolism and stimulates gluconeogenesis ( Fig. 1) (29,30). Therefore, by regulating the GRHPR gene, PPAR␣ may link the two pathways and, thus, direct the carbon flux toward glucose production (fasting state and high PPAR␣ activity) or toward the production of urea (fed state and low PPAR␣ activity) according to the need of the body (Fig. 1).
Primary hyperoxaluria type 2 in human is due to a genetic defect in GRHPR (9). In this condition, the increase of oxalate excretion can cause nephrolithiasis and nephrocalcinosis and may, in some cases, result in renal failure and systemic oxalate deposition. The coordinated effects of three functions regulate physiological oxalate level, namely oxalate synthesis, absorption, and secretion. It is not known yet whether primary hyperoxaluria type 2 occurs in mouse as well. However, using PPAR␣ ϩ/ϩ and PPAR␣ Ϫ/Ϫ mice, we showed that PPAR␣ decreases plasmatic oxalate levels. Indeed, PPAR␣ Ϫ/Ϫ mice showed variability and high oxalate levels, suggesting a reduced capacity to maintain physiological oxalate homeostasis. Interestingly, we observed that fasting or treatment with Wy-14,643 decreased oxalate levels in a PPAR␣-independent manner, showing that an additional Wy-14,643-inducible repression pathway may exist. PPAR␣-independent effects of Wy-14,643 have already been observed in the regulation of the glutamate dehydrogenase gene (2). Similarly, the stimulation of superoxide production in Kupffer cells isolated from PPAR␣ wild type and null mice is identical when treated with Wy-14,643, but here the absence of effect was attributed to the lack of PPAR␣ expression in these cells (22). During fasting, PPAR␣ represses amino acid catabolism and the urea cycle (2). Thus, in absence of PPAR␣, increased amino acid catabolism during fasting probably decreases glycine levels and depletes glyoxylate pools, thus leading to a decrease in oxalate production. The PPAR␣-independent effect of Wy-14,643 on oxalate level might be explained in a similar manner. PPAR␣ Ϫ/Ϫ mice show an absence of repressive control of PPAR␣ on amino acid catabolism and urea cycle. Wy-14,643 has already been shown to FIG. 9. Effect of PPAR␣ on oxalate production. A, percentage of glyoxylate transformation in the different products (23). B, oxalate level was measured in the plasma of WT and KO males and females fasted or not, treated or not with the Wy-14,643 (n ϭ 10 Ϯ S.E.; *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001). C, RNase protection assay performed on total RNA extracted from the liver of WT and KO mice, fasted or not, treated or not with the Wy-14,643. L27 was used as control.
repress glutamate dehydrogenase in a PPAR␣-independent manner (2), thus allowing an overstimulation of the urea cycle by entry of glutamate into the urea cycle through aspartate and argininosuccinate. Thus, the absence of a repressive control of amino acid catabolism and urea cycle and PPAR␣-independent overstimulation of the urea cycle by Wy-14,643 might decrease glycine levels, deplete glyoxylate pools, and thus decrease oxalate production. In rats, the PPRE of the GRHPR gene is not conserved but clofibrate was shown to increase the level of oxalate in the urine of treated rats (10), probably because of enhanced lactate dehydrogenase activity in the liver. Keeping in mind that the animal model and the ligand used are different, a comparison of the rat and mouse models may suggest that PPAR␣ ligand treatment may have an effect on oxalate secretion by several mechanisms. Some of them have yet to be elucidated as mentioned above. Therefore, further investigation evaluating the implication of PPAR␣ in the different pathways is needed to solve the question of the multifaceted effects PPAR␣ on oxalate levels.
In summary, we have demonstrated that the mouse GRHPR gene presents the same expression pattern, localization, and genomic organization as the human gene. However, the regulation of the gene by PPAR␣ occurs only in mouse due to dramatic differences in the promoter organization between these two orthologous genes. In mouse, the regulation of GRHPR expression by PPAR␣ may contribute to energy homeostasis by modulating the carbon supply for gluconeogenesis. Furthermore, we underscore that the regulation GRHPR in mouse, although important, does not entirely explain the effects of PPAR␣ and Wy-14,643 on oxalate production. However, a PPAR␣-dependent action is important in maintaining normal plasma oxalate level and a beneficial effect of PPAR␣ agonist in preventing oxalate accumulation in mouse is revealed by this study.