Identification and Characterization of a Novel Nontranslated Sequence Variant of the Human Intestinal Di-/Tripeptide Transporter, hPEPT1

The human H+-coupled di-/tripeptide transporter (hPEPT1) mediates intestinal absorption of dietary di- and tripeptides, as well as several peptidomimetic drug compounds. The aim of the present study was to investigate the possible role of the hPEPT1 variant hPEPT1-RF in hPEPT1 regulation. However, the proposed hPEPT1-RF mRNA sequence could not be detected in Caco-2 cells or in human intestinal samples. Instead, a new sequence variant, hPEPT1-RFI, was found, which is almost identical to the proposed hPEPT1-RF, except for two nucleotide insertions and one deletion that resulted in a changed open reading frame as compared to hPEPT1-RF. In vitro translation analysis showed that hPEPT1-RFI was not translated. In conclusion, the existence of hPEPT1-RF could not be confirmed; furthermore, the identified sequence variant, hPEPT1-RFI, does not appear to be translated and is therefore unlikely to have a regulatory effect on hPEPT1 transport activity.


Introduction
The human SLC15A1 gene codes for the proton-coupled di-/tripeptide transporter hPEPT1. hPEPT1 mediates small intestinal absorption of dietary di-and tripeptides and a number of peptidomimetic drug compounds (see [1]). A putative splice variant of the SLC15A1 gene, the hPEPT1regulatory factor (hPEPT1-RF), has been described [2]. hPEPT1-RF was found by screening a human duodenum cDNA library with a hPEPT1 cDNA probe. The cDNA clone was 1724 bp and encoded a putative 208-amino-acid protein. The hPEPT1-RF protein shifted the pH optimum of hPEPT1-mediated peptide uptake slightly towards higher pH values, when hPEPT1 and hPEPT1-RF were coexpressed in Oocytes [2]. Using bioinformatics, the hPEPT1-RF sequence was shown to consist of 6 exons, three of which were completely shared with hPEPT1 mRNA and two partially shared [3]. No studies on the mechanism of the proposed hPEPT1-RF mediated hPEPT1 regulation are available. A few studies have addressed the mRNA expression of hPEPT1-RF in Caco-2 cells and human intestinal biopsy samples [4,5]. In the present study we tried to generate a construct of hPEPT1-RF from Caco-2 cells by PCR cloning. However, no hPEPT1-RF mRNA was detected. Instead, a novel mRNA sequence variant was identified, which we showed is transcribed in human intestinal tissues and Caco-2 cells, although at low levels. The novel mRNA sequence was highly similar to the putative hPEPT1-RF in the initial parts of the sequence but had a new open reading frame (ORF). We named this novel sequence variant human peptide transporter-regulatory factor inactive (hPEPT1-RFI) since western blotting, confocal laser scanning microscopy (CLSM), and in vitro translation analysis showed that hPEPT1-RFI mRNA was not translated. The hPEPT1-RFI protein is therefore not likely to be a regulator of hPEPT1 transport activity at the protein level.

DNA Transfection.
HEK 293 cells were plated two days prior to transfection on 12-well plates (Sigma-Aldrich) coated with 33 μg·mL −1 of polylysine, at a density of 7.95 × 10 4 cells/well. Transfections were performed at 90% confluency, using 4 μL Lipofectamine (Invitrogen) per well, according to the manufacturer's instructions. Transfection media were exchanged with supplemented MEM media four hours after transfection. Cells were lysed 48 hours following transfection and used for western blotting. An estimate of transfection efficiency of HEK 293 cells was done by estimating HEK 293 cells with eGFP-containing plasmids. The eGFP transfection efficiencies were between 40% and 50%.

Quantitative
Real-Time PCR Analysis. PEPT1-RFI and PEPT1 mRNA from Caco-2 cells and human cDNA samples from human jejunum (BioChain, AMS Biotechnology) were quantified using real-time PCR analysis. Caco-2 cell RNA was isolated using the RNeasy kit (Qiagen) according to the manufacturer's instructions. Samples were treated twice with RNase free DNase I to remove genomic DNA. Only RNA with an absorbance ratio of 1.8-2.0 as measured by the 260/ 280 nm ratio was used. Reverse transcription was performed on 0.5 μg RNA with an anchored oligo(dT)primer for first strand cDNA synthesis using the Reverse-iT kit (ABgene), using β-actin as an endogenous control. β-actin expression levels in Caco-2 cells were not different between 7 and 24 days in culture (mean C T -values ± SD: 16.14 ± 0.18 and 16.30 ± 0.09, resp.). PCR amplification of hPEPT1-RFI was performed by using the primers [5 -TGCTTCCTTCCCTGT-GAGTT] (sense) and  (antisense) resulting in an amplicon length of 104 bp. The primers used for hPEPT1 were [5 -TTGATGTAAACAAA-CTGACAAGGA] (sense) and [5 -ACTAGAAGCGTGTGGC-GTTG] (antisense) resulting in an amplicon length of 103 bp, and the primers for PCR amplification of β-actin were [5 -ATTGCCGACAGGATGCAG] (sense) and [5 -CGATCC-ACACGGAGTACTTG] (antisense) resulting in an amplicon length of 100 bp. All primers used were from DNA technology (Aarhus, DK). Real-time PCR amplifications were performed using an ABI GeneAmp 5700 (Applied biosystems, DK), in a total volume of 50 μL, containing 1 μL cDNA sample, 0.2-0.3 μM of each primer and 1 × SYBR green PCR Master Mix (Applied Biosystems). Reactions were performed in triplicate and nontemplate controls were included. Reaction parameters were 10 min at 95 • C, followed by 39 cycles of amplifications (20 s at 95 • C and 1 min at 60 • C), and ending with a dissociation curve. Plasmid standards were constructed by PCR cloning of the targets, as well as the pcDNA3.1[hPEPT1] plasmid (provided by Dr. Wolfgang Sadee, Ohio State University, Columbus OH, USA). Standards were prepared with 10-fold serial dilutions and included in triplicates in each run. Results were normalized to the endogenous control and the fold difference between two targets was obtained by dividing with a calibrator sample. Real-time PCR quantification was performed by using the relative standard curve method, and results are presented as means of the normalized target values relative to the calibrator sample ± standard deviation, as calculated using the relative standard curve method [http://www.appliedbiosystems.com/].

Restriction Endonuclease Analysis.
To verify that the cDNAs obtained from human were identical to the hPEPT1-RFI sequence from Caco-2 cells, we designed a restriction endonuclease assay to distinguish between the presence or absence of a T (848Tdel), where absence of the T indicates that the RNA transcript is the hPEPT1-RFI transcript. A PCR amplicon of 104 bp (generated by the same primers as used in the quantitative real-time PCR) was generated from reverse-transcribed DNase I treated RNA from Caco-2 cells grown for 24 days, human jejunum (BioChain, AMS Biotechnology), and ileum and colon. The ileum and colon sample was a gift provided by Dr. Jesper T. Troelsen (Panum Institute, University of Copenhagen, DK). The restriction endonuclease MnlI (NEB) recognizes the target sequence and an internal restriction site, generating two fragments of 34 bp and 70 bp for the hPEPT1-RFI amplicon and three fragments of 23 bp, 34 bp, and 47 bp for the hPEPT1-RF amplicon. The digested PCR amplicons were run on a 3% MetaPhor agarose gel (Cambrex).

Western Blot Analysis.
A peptide, identical to the last 13 C-terminal amino acids of the suggested hPEPT1-RF protein sequence, was synthesised (GKRYLKGTSVGLF) (Shafer-N, DK). The epitope is also present in the predicted hPEPT1-RFI amino acid sequence. A polyclonal anti-hPEPT1-RF/RFI antibody was raised in rabbits (Department of Experimental Medicine, University of Copenhagen, DK) and affinity purified. High affinity binding to the epitope was verified by ELISA (data not shown).
hPEPT1-RFI protein expression in transiently transfected HEK 293 cells was investigated using a mouse anti-HIS antibody. HEK 293 cells or Caco-2 cells were lysed in a buffer containing 2% NP40 buffer, as described previously [8]. The boiled samples, each containing 10-14 μg protein, and 250 ng of a positive control protein of the histidine tag (Positope, Invitrogen), were run on a 4%-20% duramide PAGEr trisglycine gel (Cambrex), together with the markers MagicMark and SeeBlue (Invitrogen). The gel was transferred to a nitrocellulose membrane (Bio-Rad), blocked in 2% BSA in TBS-T buffer, washed and incubated with a mouse anti-His (C-term) IgG antibody (1 : 5,000) (Invitrogen) or a rabbit anti-hPEPT1 antibody (1 : 1,000) or a rabbit anti-hPEPT1-RF/RFI antibody (1 : 400) diluted in TBS-T with 0.1% BSA. Bound antibody was detected with horseradish peroxidaseconjugated goat-anti-mouse (Medinova) or goat-anti-rabbit (Molecular Probes) IgG antibody and visualized by the ECL Advanced kit (Amersham Biosciences). Chemiluminescence was visualized by a Kodak Image Station 1000.

hPEPT1-RFI Is a Novel Variant of hPEPT1.
Initially we attempted to isolate mRNA for the proposed regulatory factor of hPEPT1, hPEPT1-RF, from Caco-2 cells by PCR cloning. However, it was not possible to identify a sequence corresponding to the proposed hPEPT1-RF sequence. We discovered, however, a novel sequence with great similarity to the published hPEPT1-RF sequence, but with three nucleotide changes. This novel sequence variant of hPEPT1 was named hPEPT1-RFI. The sequence carried two insertions (113 Tins and 693 Tins) and one deletion (848Tdel) as compared to hPEPT1-RF (accession number [Genbank: AB001328.1]) (Figure 1(a)). The identity of this sequence variation was confirmed by sequencing, in both directions, of two independently generated hPEPT1-RFI inserts cloned from Caco-2 cells. The hPEPT1-RFI sequence had a different open reading frame (ORF) as compared to hPEPT1-RF. The sequence yielded a predicted protein of 282 amino acids with a calculated molecular weight of ∼31 kDa (30,909 Da). The first 153 amino acids (Figure 1(b)) are identical with hPEPT1, while the last 129 amino acids are unique. Topology analysis by the HMMTOP transmembrane topology prediction server [6,7] suggested a protein with 7 transmembrane segments (TMS) of which the first four TMS are identical with TMS 2-5 of hPEPT1 (Figure 1(c)). ORF finder searches on the genomic clone that contains the SLC15A1 sequence revealed a 282-amino-acid ORF corresponding to hPEPT1-RFI [GenBank: AL353574.8], supporting the existence of the hPEPT1-RFI sequence. The presence of an ORF encoding hPEPT1-RFI in the SLC15A1 sequence thus indicates that hPEPT1-RFI can be generated by the existing sequence, but we cannot at present explain the lack of expression of hPEPT1-RF.

hPEPT1-RFI mRNA Is Found in Both Caco-2 Cells and Tissues from Healthy Humans.
To investigate if the hPEPT1-RFI and/or the hPEPT1-RF sequences are transcribed in human tissues, we constructed a restriction endonuclease assay. This assay cuts the PCR amplicon sequence at a control site and at a second site that is only present in the hPEPT1-RF sequence (848T). Thus, appearance of two bands in the gel identifies the hPEPT1-RFI sequence. PCR amplicons were generated from cDNA samples obtained from human jejunum, ileum and colon, and from Caco-2 cDNA, as a control, and digested by the restriction endonuclease MnlI. The result of the restriction enzyme analysis showed that all of the investigated amplicons had the 848Tdel mutation, which is present in the hPEPT1-RFI sequence (Figure 2(a)). Furthermore, the human adenocarcinoma cell lines OVCAR-5 and SK-OV-3 contained the 848Tdel mutation, the human glioblastoma cell line, SNB-19, and the colon carcinoma cell line HCT-116 (results not shown). These results indicate that the hPEPT1-RFI sequence and not hPEPT1-RF seems to be present in all tissues tested.   (c) Topographic orientation of the hPEPT1-RFI protein as predicted by the HMMTOP transmembrane topology prediction server [6,7]. The hPEPT1-RFI corresponds to hPEPT1 in the sequence shown in black circles. The image was generated using TOPO2, transmembrane protein display software (Johns S. J., http://www.sacs.ucsf.edu/TOPO2/). PCR was performed on reverse transcribed mRNA from Caco-2 cells at passage 23 sampled on days 7 and 24 of growth (Figure 2(b)). The relative expressions of hPEPT1-RFI and hPEPT1 mRNA in Caco-2 cells were 1.0 ± 0.3 and 31.9 ± 9.0 on day 7 and 1.6 ± 0.2 and 128.2 ± 6.3 on day 24, respectively (n = 3). Thus, hPEPT1-RFI showed a low mRNA expression profile in Caco-2 cells relative to hPEPT1. Furthermore, only a minor increase in hPEPT1-RFI expression during 24 days of cell growth was seen compared to the hPEPT1 expression pattern for the same time period. In human tissues, the hPEPT1-RFI and hPEPT1 expressions were investigated on cDNA from the jejunal part of the small intestine, from two healthy individuals (Figure 2(c)). The relative expressions of hPEPT1-RFI and hPEPT1 mRNA in the first individual were 3.95 ± 0.5 versus 15.5 ± 1.6 and in the second individual 1.00 ± 0.09 versus 10.4 ± 0.8, respectively. From the total hPEPT1related transcripts 20% of these were of hPEPT1-RFI origin in the first individual and 9% were of hPEPT1-RFI in the second individual, indicating that the hPEPT1-RFI transcript is expressed at low levels compared to hPEPT1 and that the ratio between hPEPT1 and hPEPT1-RFI of the total hPEPT1related transcripts shows high variation in the examined human tissues. Anderle et al. (2006) showed that hPEPT1-RF mRNA from human intestinal biopsies contributed with 2%-44% of the total hPEPT1-related transcripts [5]. In this study PCR was performed on hPEPT1-related transcripts, and the three nucleotide changes between hPEPT1-RF and hPEPT1-RFI may not have been noticed, or in other words, it seems possible that the detected mRNA was hPEPT1-RFI rather than hPEPT1-RF.

The hPEPT1 Sequence Variant hPEPT1-RFI Is Transcribed but Not Translated.
For investigating if the hPEPT1-RFI protein was present in Caco-2 cells, we used an antibody that was generated against an epitope present in the deduced 282 amino acids of the major ORF. However, this antibody did not show specific epitope staining of any protein band from Caco-2 cells (results not shown). For further investigations, we tagged the hPEPT1-RFI with the myc/HIS epitope and placed it in front of a strong promoter to be expressed in HEK 293 cells. The protein expression level of hPEPT1-RFI in HEK 293 cells expressing a myc/HIS tagged hPEPT1-RFI was investigated using Western blotting. We were not able to detect hPEPT1-RFI protein in any of the tested cell lysates. In protein lysates from HEK 293 cells, no difference could be observed between lanes with hPEPT1-RFI and empty (mock) plasmid using the anti-HIS antibody (Figure 3(a)). As controls proteins containing the myc/HIStag and HEK 293 cells transiently transfected with hPEPT1 were used (Figures 3(a) and 3(b)). Confocal laser scanning microscopy studies on hPEPT1-RFI transfected, and mock transfected HEK 293 cells could confirm the results from the Western blot experiments (Figures 3(c) and 3(d)). hPEPT1 expression in HEK 293 cells was used as a positive control for the transfection procedure (Figure 3(e)). We investigated whether the pcDNA3.  plasmid supplied with the kit, gave protein bands at the expected sizes (Lanes 1 and 3, Figure 4) as measured by the incorporation of radioactive labeled methionine. The two negative controls, pcDNA3.1[mock] and-DNA, showed no sign of protein bands. Also, the pcDNA3.1[hPEPT1-RFI]myc/HIS plasmid showed no protein band (Lane 2, Figure 4). These results, in support of the previous results, showed that no detectable protein is translated from the hPEPT1-RFI transcript. Interestingly, Saito and coworkers observed a band of low molecular weight (∼23 kDa) after in vitro translation of the putative hPEPT1-RF cRNA sequence. Possible this band could be an artifact of either radioactive labeled tRNAs, or small polypeptides from the translation of fragments remaining in the reticulocyte lysate.
In conclusion, we have revealed a novel variant of the SLC15A1 gene, named hPEPT1-RFI. Characterization of this novel sequence showed that it was present in several cell lines of human origin, including Caco-2 cells, and in human intestinal samples. We did not find evidence for the existence of the putative hPEPT1-RF sequence, questioning the proposed role of the hPEPT1-RF protein as a regulator of hPEPT1 pH sensitivity. Only a change of three nucleotides separates the hPEPT1-RF and hPEPT1-RFI sequence, but the impact of two insertions and one deletion was significant, since it changed the reading frame of the hPEPT1-RF sequence. Investigations of the putative protein of the hPEPT1-RFI sequence by Western blotting, CLSM, and in vitro translation gave no detectable results. As a consequence, no protein-protein interaction between hPEPT1-RFI and hPEPT1 can be anticipated.