The intestinal peptide transporter PEPT1 mediates the uptake of di- and tripeptides derived from dietary protein breakdown into epithelial cells. Whereas the transporter appears to be essential to compensate for the reduced amino acid delivery in patients with mutations in amino acid transporter genes, such as in cystinuria or Hartnup disease, its physiological role in overall amino acid absorption is still not known. To assess the quantitative importance of PEPT1 in overall amino acid absorption and metabolism, PEPT1-deficient mice were studied by using brush border membrane vesicles, everted gut sacs, and Ussing chambers, as well as by transcriptome and proteome analysis of intestinal tissue samples. Neither gene expression nor proteome profiling nor functional analysis revealed evidence for any compensatory changes in the levels and/or function of transporters for free amino acids in the intestine. However, most plasma amino acid levels were increased in Pept1−/− compared with Pept1+/+ animals, suggesting that amino acid handling is altered. Plasma appearance rates of 15N-labeled amino acids determined after intragastric administration of a low dose of protein remained unchanged, whereas administration of a large protein load via gavage revealed marked differences in plasma appearance of selected amino acids. PEPT1 seems, therefore, important for overall amino acid absorption only after high dietary protein intake when amino acid transport processes are saturated and PEPT1 can provide additional absorption capacity. Since renal amino acid excretion remained unchanged, elevated basal concentrations of plasma amino acids in PEPT1-deficient animals seem to arise mainly from alterations in hepatic amino acid metabolism.
- peptide transporter 1
- high protein load
products of the hydrolysis of proteins in the small intestine are short-chain peptides and free amino acids. Free amino acids can be absorbed via various apical amino acid transporters with certain group specificities (9, 26), while di- and tripeptides are taken up into intestinal epithelial cells by the peptide transporter PEPT1 (SLC15A1). In the kidney, the paralog PEPT2 (SLC15A2) is responsible for the reabsorption of filtered di- and tripeptides, preventing amino acid loss in urine (25). PEPT1 and PEPT2 belong to the peptide transporter (PTR) family that comprises an evolutionary conserved transporter group present in all living organisms. A common feature of the PTR family proteins is that they couple substrate movement across the membranes to movement of protons down an inwardly directed electrochemical proton gradient, allowing transport of peptides against a concentration gradient (34). Due to the electrogenic nature of the peptide transporters, the transmembrane proton gradient contributes to the driving force. In the intestine, this gradient is predominantly generated by the sodium-proton exchanger, Na+/H+ exchanger 3 (SLC9A3) (10), which maintains the H+-electrochemical gradient through Na+-coupled H+ efflux and by regulation of intracellular pH. It was suggested that any maneuvers that alter the pH and change the membrane potential may result in changes in the uptake of substrates by PEPT1 (37). However, at high substrate concentrations, solely the membrane potential, but not the pH gradient, determines the transport rate of PEPT1 (8). Substrates of PEPT1 and PEPT2 are, with a few exceptions, all possible di- and tripeptides, but also peptidomimetic drugs, like aminocephalosporins, selected angiotensin I converting enzyme inhibitors, and various prodrugs (27, 28, 39). By affecting intestinal uptake and renal clearance of these drugs, both peptide transporters contribute to the pharmacokinetics of these compounds.
PEPT1 is a high-capacity and low-affinity transporter, whereas PEPT2 for the same substrates possesses, on average, a 10-fold higher affinity, but displays a lower maximal transport rate. PEPT2 is expressed highest in kidney tubules, but is also expressed in other organs, like brain and bronchial epithelium (4, 17, 31). In mice lacking PEPT2, we demonstrated that the tubular reabsorption of model dipeptides and their renal accumulation was significantly reduced. In particular, renal reabsorption of Cys-Gly, an important thiol and breakdown product of glutathione, was markedly impaired in PEPT2-deficient animals (13, 29). In addition, Ocheltree et al. demonstrated the prime role of PEPT2 in the clearance of dipeptides and related drugs from the cerebrospinal fluid (25). Intestinal uptake and perfusion experiments with the model dipeptide glycylsarcosine (Gly-Sar) have recently shown that >80% of Gly-Sar uptake in the small intestine is mediated by PEPT1 (20) and that PEPT1 may contribute significantly to intestinal fluid absorption (8).
There is a long-lasting dispute on the quantitative importance of intestinal transport of di- and tripeptides compared with that of free amino acids, dating back to the original discovery of peptide transport. The demonstration of transport of intact peptides was against the dogma that only monomers of all ingested nutrients are absorbed. However, it had been shown in intestinal perfusion studies in patients suffering from cystinuria or Hartnup disease (2, 19) that intestinal absorption of selected amino acids was abolished by the malfunctions of the amino acid transporters, while, when the amino acids were provided in the form of dipeptides, normal absorption occurred (3, 19, 33). In particular for lysine, an essential amino acid, this compensation of amino acid transporter function via peptide uptake established that PEPT1 under these condition is life-sustaining (19). Despite a comprehensive analysis of PEPT1 structure and functions with hundreds of publications over the last 40 yr, its overall importance in intestinal amino acid absorption is still unknown.
We, therefore, characterized the role of the transporter in amino acid absorption and metabolism in mice lacking PEPT1. For assessing to what extent the deletion of PEPT1 is compensated by changes in expression and function of amino acid transporters in intestinal epithelial cells, we also profiled intestinal tissues for changes in transcripts, proteins, and metabolites in transporter-deficient animals. Furthermore, intragastric administration of proteins, including 15N-labeled protein, with concomitant analysis of plasma and tissue amino acid levels was used to determine the role of PEPT1 in the absorptive phase in vivo.
Pept1 knockout mice were obtained from Deltagen (San Mateo, CA) (20) and backcrossed for 10 generations to C57BL/6J background. Animals were maintained at 22 ± 2°C and on a 12:12-h light-dark cycle. All procedures were conducted according to the German guidelines for animal care and approved by the state ethics committee (Reference no. 55.2-1-54-2531-140-08).
Plasma and urine amino acid analysis.
For 5 consecutive days, Pept1+/+ and Pept1−/− mice (n = 10) were fed a semisynthetic purified diet with medium (21% of energy) protein content (Ssniff EF Control, Ssniff Spezialdiäten, Soest, Germany). Spontaneous urine samples were collected daily between 8 and 9 AM and pooled (per animal) for analysis. After 4 days, plasma samples were collected. In 40 μl urine or plasma, amino acids and derivatives were labeled by the iTRAQ methodology using the AA45/32 starter kit, according to the manufacturer's instructions (Applied Biosystems) and analyzed via liquid chromatography/tandem mass spectrometry (LC-MS/MS) (3200QTRAP LC/MS/MS, Applied Biosystems). The data were analyzed using the Analyst 1.5 software.
Creatinine concentrations were determined with the Creatinine Liquicolor Jaffè Test no. G205117 (Rolf Greiner BioChemica, Flacht, Germany), and urea levels were analyzed with the Urea Liquicolor Test no. 10505 (Human Gesellschaft für Biochemica und Diagnostica, Wiesbaden, Germany), according to the manufacturer's instructions. Osmolarity in urine was determined with a semi-microosmometer (Knauer, Berlin, Germany).
Two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) of scraped mucosa samples was performed as described (16). Protein (300 μg) was loaded by passive rehydration, and isoelectric focusing was performed (pH 3–10). SDS gels (12.5%) were ran, fixed, and stained. Gels were scanned and analyzed using the Decodon Delta 2D software 4.0 (Decodon, Greifswald, Germany). Tryptic digestion of selected protein spots and peptide mass fingerprinting by matrix-assisted laser desorption time-of-flight mass spectrometer were performed as described (38).
Renal brush border membrane vesicles.
Brush border membrane vesicles (BBMV) were prepared from kidney cortex by Mg2+ aggregation, as described (5). Final pellets were suspended in Tris-HEPES-mannitol buffer comprising 1 mM Tris, 2 mM HEPES, and 100 mM mannitol (pH 7.4). l-Proline uptake was determined in buffers with or without 100 mM NaCl. The purity of BBMV preparation was determined by the enrichment of alkaline phosphatase activity. d-Glucose uptake into renal BBMV was similarly measured at 37°C in the presence of either 100 mM Na thiocyanate (SCN) or 100 mM KSCN by the rapid filtration technique (22).
Western blot analysis.
Brush border membranes from intestinal tissues were prepared as for renal BBMV by the Mg2+ precipitation technique (32). Twenty-five micrograms of the intestinal or renal membrane protein preparation were separated by 10% SDS-PAGE, followed by transfer to a nitrocellulose membrane with a semidry blotter (Bio-Rad Laboratories, München, Germany). After blocking, the primary antibody [rabbit anti-PEPT1, VGKENPYSSLEPVSQTNM, dilution 1:1,000 or rabbit anti-PEPT2, CKQIPHIQGNMINLETKNTRL, dilution 1:1,000; both Pineda, Berlin, Germany (35)] was added. After washing, the secondary antibody was added (SCD1, goat polyclonal IgG, dilution 1:10,000; Santa Cruz Biotechnology, or donkey anti-rabbit IRDye680, dilution 1:10,000; LI-COR, Bad Homburg, Germany). Chemiluminescence was detected with a radiographic film, and fluorescence was detected by the Odyssey Infrared Imaging System (LI-COR).
RNA preparation and microarray analysis.
Total RNA of scraped mucosa samples was isolated using TRIZOL (Invitrogen, Karlsruhe, Germany) and the RNeasy Mini Kit (Quiagen, Hilden, Germany). After reverse transcription, the corresponding cDNA was biotinylated and fragmented following the original protocol of Affymetrix (Affymetrix, Santa Clara, CA). The cRNA samples were hybridized overnight on Affymetrix mouse whole-genome arrays, and the arrays were washed and scanned following the instructions of the provider.
Quality control and statistical analysis were performed by a bioconductor and R based resource (https://madmax.bioinformatics.nl).
Uptake studies by the everted gut sac technique.
Jejunum (JJ) segments of Pept1+/+, Pept1+/−, and Pept1−/− were everted with a metal rod (n = 4). Sacs were filled with carbogen gassed Krebs buffer pH 7.4 and transferred into Krebs buffer pH 6.0 at 37°C with 10 μM [14C]Gly-Sar or 10 μM [3H]l-proline (GE Healthcare, München, Germany) and incubated for 10 min at 37°C. After incubation, serosal fluid was drained into small tubes. Each sac was weighed before and after fluid collection to calculate serosal volume. Tissues were dried, weighed, and lysed in NaOH for normalization of transport data.
Ussing chamber experiments.
Four stripped proximal JJ segments per animal were mounted into Ussing chambers (Warner Instruments) with an exposed surface area of 0.22 cm2. Tissues were short-circuited and equilibrated for 20–40 min before substrate-induced short-circuit currents (Isc) were measured. The apical side was constantly perfused (60 ml/min) with a MES-buffered Krebs solution (pH 6.4) containing 20 mM substrate or mannitol. The basolateral side was recirculated with a carbogen-gassed Krebs solution (pH 7.4). Substrates were applied for 2 min and washed out for 7 min. Substrate-induced currents were calculated as the difference of currents measured immediately before and at the end of the substrate application period and were corrected to a surface area of 1 cm2.
High-protein load and administration of 15N-labeled protein by gastric gavage.
Five to six mice per group and genotype received by gavage either 178 mg protein using a commercially available protein supplement with casein as the dominant fraction (Power System 90 Plus, Well Plus Trade, Hamburg, Germany), or 8.83 mg 15N-labeled protein. Yeast protein (Saccharomyces cerevisiae) was provided by Dr. W. Eisenreich (TU München, Garching, Germany).
For high-protein load, blood and tissue samples were collected at 0, 10, 30, and 60 min after gavage. In studies with 15N-labeled protein, 15 and 30 min after administration, blood was collected from portal vein and heart. For amino acid analysis of tissue, parts of duodenum (DD), JJ, and ileum (IL) were homogenized and diluted in MeOH/H2O.
Measurement of liver enzymes and urea.
In liver tissue samples, aspartate aminotransferase (AST), alanine aminotransferase, and glutamate dehydrogenase were determined. Tissue samples were homogenized, supernatant was diluted (1:100), and activities were determined with the alanine aminotransferase liquiUV test no. 12012 and AST liquiUV test no. 12011 (Human, Wiesbaden, Germany) and glutamate dehydrogenase FS no. G82100 (Rolf Greiner BioChemica, Flacht, Germany), according to the manufacturer's instructions. Urea levels were analyzed from undiluted supernatant with the Urea Liquicolor test no. 10505 (Human, Wiesbaden, Germany), according to the manufacturer's instructions. Arginase activity was measured by activation of the tissue supernatant in presence of glycine, incubation with l-arginine at 37°C for 10 min, and termination of reaction by adding 20% TCA. Protein was removed by centrifugation, and the supernatant was mixed with 9% H2SO4, 23% H3PO4, 100 mM FeCl3, and 3% isonitrosopropionphenone for urea determination (incubation: 100°C for 60 min).
Statistical analysis was performed using R 2.8 [R Foundation of Statistical Computing (36)] and GraphPad Prism 4.01 (GraphPad Software). One-way or two-way ANOVA and Tukey test or unpaired Student's t-test were used to test for statistical significance. Data are presented as means ± SD.
Haploinsufficiency in Pept1+/− animals.
The mouse Pept1 gene was disrupted by replacing part of the coding region of exon 3 with the LacZ-Neo cassette via homologous recombination in embryonic stem cells, which resulted in a significantly reduced Gly-Sar uptake in the intestine (20). We here demonstrate haploinsufficiency in jejunal sections of Pept1+/− animals using the everted gut sac technique and Gly-Sar (Fig. 1A). Pept1+/− animals displayed only 45% of uptake activity compared with Pept1+/+ animals, whereas, in Pept1−/− animals, uptake was reduced by 90%. Ussing chamber experiments confirmed that, in Pept1+/− animals, Isc was reduced by ∼40%, whereas, in Pept1−/− animals, the Gly-Sar-induced current was almost abolished (3% residual activity) compared with that in wild-type animals (Fig. 1B). Reduced transport function in Pept1+/− animals was associated with a reduced PEPT1 protein density in the brush border membrane (Fig. 1C).
Amino acid levels in plasma, urine, and liver in Pept1−/− mice.
In plasma and urine samples, 40 amino acids and derivatives could be quantified via LC-MS/MS analysis and revealed altered levels in Pept1−/− animals. In plasma, the sum of all amino acids and derivatives was increased in Pept1−/− mice compared with Pept1+/+ animals (4,971.13 ± 635.16 vs. 4,028.97 ± 469.93 μmol/l, P = 0.015). Among all 40 metabolites none displayed decreased levels, whereas 24 showed elevated concentrations in plasma (Table 1). Although branched-chain amino acids, such as valine, leucine, and isoleucine, and also aminobutyric acid displayed slightly elevated concentrations in Pept1−/− mice, differences did not reach significance. Plasma levels of the endogenous dipeptides anserine and carnosine remained unaltered as well.
Analysis of urea, creatinine, and osmolarity of urine samples revealed no differences between Pept1+/+ and Pept1−/− animals (data not shown). Amino acid levels in urine were corrected for creatinine concentration. The sum of all urinary amino acids showed no differences based on genotype (Fig. 2A). Only glycine and argininosuccinate, a urea cycle intermediate, displayed significantly lower levels in PEPT1-deficient compared with wild-type animals (glycine: 135.53 ± 9.56 vs. 159.28 ± 5.50 μmol/l, P < 0.001; arginosuccinate: 77.95 ± 21.97 vs. 104.15 ± 10.40 μmol/l, P = 0.024). The most striking finding was an almost 20-fold increase in urinary proline levels in Pept1−/− animals (429.20 ± 387.23 vs. 21.80 ± 9.52 μmol/l, P = 0.028). All other 37 amino acids and derivatives did not show significant changes in urinary levels. In liver tissue, the total concentration of all amino acids remained unchanged. Only citrulline, phosphoethanolamine, and ethanolamine (citrulline: 2.89 ± 1.41 vs. 1.28 ± 0.55 μmol/l, P = 0.026; phosphoethanolamine: 6.87 ± 2.32 vs. 4.00 ± 0.97 μmol/l, P = 0.019; ethanolamine: 0.56 ± 0.18 vs. 0.31 ± 0.09 μmol/l, P = 0.013) displayed significantly increased concentrations in Pept1−/− compared with Pept1+/+ mice (Fig. 2B). Analysis of selected enzyme activities in liver tissue of Pept1+/+ and Pept1−/− did not yield any significant differences, while urea concentration was significantly decreased in Pept1−/− animals (Table 2).
Transcriptome and proteome analysis in intestinal tissues.
For analysis of changes in transcriptome and proteome levels, jejunal sections of small intestine were removed from Pept1+/+ and Pept1−/− mice. 2D-PAGE-based proteome analysis of mucosal scrapings allowed, on average, 500 proteins to be detected. Thereof seven proteins showed major changes, and six could be identified by matrix-assisted laser desorption time-of-flight mass spectrometer (Table 3). The enzyme peptidylprolyl isomerase A (Ppia) and the translation elongation factor-1β increased in levels by ∼50% in Pept1−/− animals. Proteins such as ribonuclease inhibitor 1, suggested to control mRNA turnover, the coatomer subunit-ε, essential for the retrograde Golgi-to-endoplasmic reticulum transport of dilysine-tagged proteins, the ATP synthase subunit-β, and the adapter protein 14-3-3, implicated in the regulation of a large spectrum of signaling pathways, displayed lowered levels in Pept1−/− animals.
Differential gene expression analysis of scraped mucosa samples was performed by Affymetrix mouse whole genome arrays. Only four genes showed significantly altered mRNA levels in the transporter-deficient mice (q < 0.05). Except for Pept1, with a log2-fold change of −56, only subtle changes of transcript levels could be observed in Pept1−/− mice. Δ1-Pyrroline 5-carboxylate synthase (log2-fold change = −1.5) catalyzes an essential step in pathways by which ornithine, arginine, and proline are synthesized from glutamate. The other two transcripts derive from expressed sequence tag clones not yet associated to known proteins (transcripts identifiers: BE648536.1, BM119376.2). No changes in the transcript levels of any of the known amino acid transporters or of PEPT2, expressed in the gut enteric nervous system (30), were identified in tissues from PEPT1-deficient animals.
Electrogenic intestinal transport of selected amino acids and Gly-Sar in Pept1−/− animals.
Figure 3A shows representative current recordings obtained from jejunal segments of Pept1+/+ and Pept1−/− animals in Ussing chambers. The average glucose-induced Isc was not significantly different between Pept1−/− and Pept1+/+ animals (110.8 ± 23.0 vs. 92.8 ± 19.3 μA/cm2). Gly-Sar-mediated currents expressed as percentage of glucose-induced currents were 37 ± 3% (n = 8) in wild-type and virtually absent in Pept1−/− mice (1 ± 1%, n = 6). Currents induced by glycine or l-proline compared with those of glucose were not significantly different between Pept1+/+ and Pept1−/− animals (glycine: 34 ± 2 vs. 33 ± 2%; l-proline: 46 ± 2 vs. 50 ± 3%). Likewise, uptake and transepithelial transport of 3H-labeled l-proline using everted gut sacs of jejunal segments showed no differences between genotypes (Fig. 3B).
Amino acid uptake and peptide transporter expression in kidney tissue.
BBMV were prepared from kidney tissue of Pept1+/+ and Pept1−/− animals. The enrichment in the specific activity of alkaline phosphatase in the final membrane fraction was >18-fold. Although PEPT1 expression in mouse kidney was reported (20), we were not able to detect the PEPT1 protein in kidney by Western blot analysis (Fig. 4A), nor by immunofluorescence staining or by LacZ staining (data not shown). To assess whether PEPT2 expression in kidney was altered due to lack of PEPT1, we performed Western blot analysis, but could not observe any differences between wild-type and Pept1−/− animals (Fig. 4B). Since plasma proline levels showed marked differences between Pept1+/+ and Pept1−/− animals, uptake of l-proline into kidney BBMV was studied with d-glucose uptake serving as a control. In the presence of a Na+ gradient, similar overshoot phenomenona were observed for both l-proline (Fig. 4C) and d-glucose (Fig. 4D) uptake, independently of genotype.
Systemic amino acid appearance after administration of 15N-labeled protein by gavage.
After 12-h fasting, 15N-labeled protein (8.83 mg) was administered to Pept1+/+ and Pept1−/− animals by gavage. Blood samples were collected at 15 and 30 min from portal vein and heart. No differences were observed in blood collected from heart in any of the amino acids measured in plasma samples. A total of 10 amino acids (isoleucine, leucine, valine, glutamic acid, tyrosine, phenylalanine, alanine, citrulline, glutamine, tryptophan) displayed higher concentrations in portal compared with peripheral blood, but independently of genotype. However, in portal blood, isoleucine levels were significantly higher in Pept1+/+ compared with Pept1−/− animals after 15 min, but this difference disappeared after 30 min (Fig. 5A). No other labeled amino acid showed any difference in portal appearance rate, depending on genotype.
The administration of the 15N-labeled protein allowed us also to assess the prime site of absorption in the intestine. After 30 min, 21 labeled amino acids could be quantified in tissue samples of DD, JJ, and IL by LC-MS/MS analysis. Levels of almost all amino acids, except for cysteine, were highest in JJ compared with DD and IL segments (P < 0.001), but no differences between genotypes could be detected. Concentrations of proline, arginine, isoleucine, and cysteine in the different tissue segments are shown in Fig. 5B.
Plasma and liver amino acid levels after intragastric administration of a high-protein load.
Amino acids and derivatives were quantified in plasma after intragastric administration of a high-protein load (178 mg) with peripheral blood samples collected at 10, 30, and 60 min. Already after 10 min, samples obtained from Pept1+/+ animals showed significantly increased levels of glutamine, taurine, and glycine compared with Pept1−/− animals (Fig. 6). Thirty minutes after protein administration, five further amino acids displayed significantly elevated levels in wild-type mice, including the branched chain amino acids valine and isoleucine, as well as threonine, proline, and phosphoetanolamine. At 60 min, besides the sum of all amino acids (10,880.97 ± 1,404.39 vs. 7,344.08 ± 940.72 μmol/l, P < 0.001), a total of 15 amino acids and derivatives displayed significantly higher plasma concentrations in Pept1+/+ compared with Pept1−/− animals. Among them were, in addition to valine and isoleucine, aminobutyric acid, glutamate and aspartate, their amides glutamine and asparagine, serine, glycine, taurine, threonine, alanine, aminoadipic acid, and cystathione. From all amino acids, proline showed the largest concentration difference (430.17 ± 86.78 vs. 105.67 ± 15.43 μmol/l, P < 0.001).
In liver tissue, 60 min after the protein load, amino acid concentrations were significantly higher in wild-type animals for 12 amino acids and the sum of all amino acids (718.77 ± 84.55 vs. 541.59 ± 72.23 nmol/mg protein, P = 0.003), with most prominent changes in case of proline (Table 4). Measurements of enzyme activities and urea concentrations in liver tissue 60 min after gavage did not reveal major differences between Pept1+/+ and Pept1−/− animals (Table 2). However, independent of genotype, urea levels increased (19.72 ± 3.81 vs. 15.34 ± 4.51 μmol/mg, P = 0.02) after the protein load compared with basal levels, whereas AST activities decreased (1,737.27 ± 354.52 vs. 2,262.33 ± 421.81 mU/mg, P = 0.004).
The quantitative importance of peptide transport over that of transport of free amino acids has been a subject of controversial discussions after the initial discovery of peptide transport some 40 yr ago. Controversy arose from the dogma that only monomers as final degradation products of the luminal digestion of all nutrients could be absorbed (for a comprehensive review, see Ref. 21). However, human studies on protein digestion revealed that the majority of amino acids in the lumen are found in form of di- and tripeptides (1). Their uptake into tissues was finally demonstrated, and BBMV studies revealed that the transport process was proton dependent (15). Cloning of PEPT1 demonstrated that it is a high-capacity electrogenic system with the unique feature to transport literally thousands of possible dietary di- and tripeptides. Malfunction of PEPT1 leading to reduced amino acid absorption could, when uptake of essential amino acids becomes limited, cause a retardation of growth and development. Such a phenotype was indeed observed when PEPT1 in the nematode C. elegans was disrupted. PEPT1 deficiency here caused a severly retarded development, a reduced body size, and markedly reduced reproduction capacity (23).
Already in the 1970's, it was shown that peptide transport is essential for life in patients suffering from cystinuria or Hartnup disease. In both diseases, mutations in amino acid transporter genes cause not only a known renal phenotype, but also impaired intestinal amino acid absorption. Peptide transport, however, was shown to compensate for the loss of intestinal amino acid transporter function (2, 19, 33). Moreover, peptide transporters were shown to determine the oral availability of a large variety of peptidomimetic drugs and prodrugs (27). Thus PEPT1 was finally perceived as a system of physiological, but also of pharmacological, importance. However, even after 40 yr of intensive research on peptide transport, its nutritional importance and contribution to amino acid availability in mammals is still not known. With the availability of a PEPT1-deficient mouse, the role of the transporter in overall amino acid handling can be studied. But, in contrast to the marked phenotype in C. elegans, the transporter-deficient mouse line appears healthy, and no alterations in body weight, development, and fertility could be observed (20).
This finding asked whether there is a compensation with changes in amino acid transporter expression and/or increased transport function. Surprisingly, microarray-based profiling of intestinal tissues of PEPT1-deficient compared with wild-type animals did not reveal evidence for significant changes in mRNA levels of any of the amino acid transporters. In addition, electrogenic transport of glycine and proline in Ussing chambers, assessed at saturating substrate concentrations and compared with glucose transport, showed no differences between PEPT1-deficient and wild-type animals. In vivo studies with the administration of a 15N-labeled protein by gavage and sampling of portal and cardiac blood, followed by isotopologue analysis of amino acids, also failed to demonstrate any significant differences in the appearance rates of amino acids after digestion and absorption of the protein. Amino acid levels and tracer concentrations in intestinal tissue samples also did not reveal genotype-specific differences. Only when a large intragastric protein load was administered did Pept1−/− animals display reduced plasma appearance rates of a subset of amino acids compared with those in wild-type animals. Among these amino acids, proline, followed by valine, leucine, aminobutyric acid, glutamate, aspartate, and threonine, showed the most pronounced differences. As we could demonstrate that proline absorption rate via the proline transport pathways represented by SIT1/IMINOB (SLC6A20) and PAT1 (SLC36A1) transporters in the intestine (7) were not different between genotypes, the difference in proline plasma appearance may indicate that a larger quantity of proline is absorbed in peptide-bound form via PEPT1. This may not be surprising, as proline-containing peptides are generally more resistant to enzymatic hydrolysis by brush border and intracellular peptidases than all other peptides. Moreover, proline-containing peptides, such as Ile-Pro-Pro and Leu-Pro-Pro, were recently shown in humans to be absorbed in intact form after oral administration (18), yet with very low plasma appearance rates (12). A remarkable feature of all proline-containing peptides is that the peptide bond can be presented in cis or trans conformation, and PEPT1 was shown to transport only the trans peptide bond conformers (6). Interestingly, our proteome analysis identified the enzyme Ppia with twofold increased protein level in Pept1−/− animals. Ppia catalyze the cis-trans isomerization of proline imidic peptide bonds (14). This finding suggests that there could be a coregulation of Ppia and PEPT1 and that absorbed proline peptides in intestinal cells may participate in regulation of the isomerase levels. Taken together, the acute challenge tests providing different protein quantities to the mice established that PEPT1 is most important for amino acid absorption at high dietary protein intakes and that peptides with a higher enzymatic resistance against hydrolysis gain a kinetic advantage for uptake via PEPT1.
Among the most striking findings in the present study are the higher plasma amino acid levels present in Pept1−/− animals when kept on a standard diet. Although 24 out of 40 quantified amino acids and derivatives showed increased blood levels, most pronounced elevations were found for proline, arginine, and citrulline. Since this could result from alterations in renal excretion of these amino acids, we also determined urinary amino acid levels and observed a huge increase in proline and a decrease in glycine and argininosuccinate excretion. PEPT1 expression could not be detected here in mouse kidney, and for PEPT2 we showed that it does not undergo adaptive changes in protein level by the loss of PEPT1. Moreover, our studies in renal BBMV demonstrated that proline transport, like glucose transport, remains unchanged in PEPT1-deficient animals. The differences in glycine and proline urinary levels, therefore, seem to result predominantly from a different tubular load and a competition of these amino acids for renal reabsorption for common transporters with an overflow of proline, as shown in other models (24).
Amino acids and derivatives related to urea cycle are overrepresented in the subset of altered plasma amino acids. Amino acid oxidation requires the elimination of nitrogen by condensation of NH3 with CO2 for production of urea in liver. Likewise, urea production varies as a function of dietary protein intake (11). Urea levels in liver of Pept1−/− animals displayed lower basal concentrations compared with that in wild-type animals, but increased after the high-protein load with no genotype-specific effects. Arginine and citrulline showed increased plasma levels, and citrulline concentration in liver was increased as well. Argininosuccinate was found with decreased levels in urine. In addition, glutamate, glutamine, and alanine, which all contribute to interorgan transport of nitrogen and its delivery to liver for detoxification, showed increased plasma concentrations. However, neither liver transaminase levels nor the urea cycle enzyme arginase, serving as a marker enzyme of urea cycle capacity, revealed significant changes between genotypes, also not after a high-protein load. Taken together, these findings indicate that shuttling of amino acids into the urea cycle may be altered in Pept1−/− animals with increased plasma concentrations of almost all amino acids involved in nitrogen delivery and detoxification in liver. Thus these unexpected findings call for further experiments to determine the causality underlying these alterations in amino acid handling in Pept1−/− animals.
Based on our findings, we may conclude that the contribution of PEPT1 to overall intestinal amino acid absorption is negligible when low amounts of protein are ingested, but becomes visible when high-protein loads reach the intestine. Under these conditions, the maximum rate of hydrolysis in the lumen or at the brush border membrane becomes limited, leading to a higher concentration of intact di- and tripeptides that are then available for PEPT1. In addition, maximal amino acid absorption capacity may have been reached, allowing some amino acids in peptide form to gain a kinetic advantage, with higher absorption rates via PEPT1. These findings also suggest that a high protein diet administered to Pept1−/− mice may induce a stronger variant phenotype as under standard protein feeding conditions, in which animals may balance their amino acid homeostasis at its “limits”. Yet PEPT1 deficiency resembles a state of amino acid imbalance with distinct changes in plasma amino acid levels. Among those, amino acids that are directly or indirectly related to urea cycle are overrepresented, indicating an altered, yet partially compensated, hepatic detoxification capacity in animals deficient of PEPT1.
NuGO arrays were partly funded by the Nutrigenomics Organisation (EC funded Network of Excellence, Grant FOOD-2004-506360). This work was supported by a grant (DA 190/8-1) from the Deutsche Forschungsgemeinschaft.
No conflicts of interest, financial or otherwise, are declared by the author(s).
We thank Ronny Scheundel and Johanna Welzhofer for excellent technical assistance and Britta Spanier for the fruitful discussions.
- Copyright © 2011 the American Physiological Society