Transport of the phosphonodipeptide alafosfalin by the H+/peptide cotransporters PEPT1 and PEPT2 in intestinal and renal epithelial cells


M. Brandsch, Biozentrum, Martin-Luther- University Halle-Wittenberg, Membrane Transport Group, Weinbergweg 22, D-06120 Halle, Germany. Fax: + 49 345 552 7258, Tel.: + 49 345 552 1630, E-mail:


The interaction of the antibacterial phosphonodipeptide alafosfalin with mammalian H+/peptide cotransporters was studied in Caco-2 cells, expressing the low-affinity intestinal type peptide transporter 1 (PEPT1), and SKPT cells, expressing the high-affinity renal type peptide transporter 2 (PEPT2). Alafosfalin strongly inhibited the uptake of [14C]glycylsarcosine with Ki values of 0.19 ± 0.01 mm and 0.07 ± 0.01 mm for PEPT1 and PEPT2, respectively. Saturation kinetic studies revealed that in both cell types alafosfalin affected only the affinity constant (Kt) but not the maximal velocity (Vmax) of glycylsarcosine (Gly-Sar) uptake. The inhibition constants and the competitive nature of inhibition were confirmed in Dixon-type experiments. Caco-2 cells and SKPT cells were also cultured on permeable filters: apical uptake and transepithelial apical to basolateral flux of [14C]Gly-Sar across Caco-2 cell monolayers were reduced by alafosfalin (3 mm) by 73%. In SKPT cells, uptake of [14C]Gly-Sar but not flux was inhibited by 61%. We found no evidence for an inhibition of the basolateral to apical uptake or flux of [14C]Gly-Sar by alafosfalin. Alafosfalin (3 mm) did not affect the apical to basolateral [14C]mannitol flux. Determined in an Ussing-type experiment with Caco-2 cells cultured in Snapwells™, alafosfalin increased the short-circuit current through Caco-2 cell monolayers. We conclude that alafosfalin interacts with both H+/peptide symporters and that alafosfalin is actively transported across the intestinal epithelium in a H+-symport, explaining its oral availability. The results also demonstrate that dipeptides where the C-terminal carboxyl group is substituted by a phosphonic function represent high-affinity substrates for mammalian H+/peptide cotransporters.




intestinal type peptide transporter 1


renal type peptide transporter 2

Alafosfalin (alaphosphin, l-alanyl-l-1-aminoethylphosphonic acid) is an antibacterial dipeptide analogue where the carboxyl group at the C-terminal alanine is replaced with a phosphonic [P(O)(OH)2] function. The compound was one of the most promising aminophosphonic acids obtained in an extensive study synthesizing more than 300 di- to penta-peptide alanine mimetics with varying stereometry and different substituents for the C-terminal carboxyl function [1]. It displays good oral availability, substantial antibacterial activity mainly against Gram-negative bacteria and synergism with β-lactam antibiotics [1–5]. In clinical studies alafosfalin was tested for the treatment of gastrointestinal [3] and urinary tract infections [2,4]. Studies demonstrated the competitive effect of food on its enzymatic breakdown in the intestinal lumen [6]. In a recent publication Kafarski & Lejczak [7] reviewed the potential medical importance of aminophosphonic acids and conclude that, due to their negligible mammalian toxicity and the fact that they very efficiently mimic aminocarboxylic acids making them extremely important antimetabolites, aminophosphonic acids play an important role in drug development. Very recently Tsopelas and coworkers assessed [99mTc]alafosfalin as an infection imaging agent and concluded that its distribution characteristics are advantageous in imaging abdominal and soft tissue infections [8]. Moreover, alafosfalin has been used as a selective agent in bacterial growth media for isolation of strong pathogens like Salmonella[9,10].

Several authors have studied bacterial uptake and intracellular metabolism of phosphonopeptides. They have demonstrated the uptake of alafosfalin into bacteria by peptide permeases located in the cytoplasmic membrane [1,11–14]. However, there are no detailed studies regarding the specific uptake mechanism at mammalian cells for phosphonodipeptides such as alafosfalin.

We hypothesized that these compounds, even though they do not possess a carboxyl group at the C-terminus, are substrates for the mammalian peptide transporters PEPT1 and PEPT2. These carriers are not only responsible for the uptake of nutritional di- and tripeptides across the intestinal and renal epithelium and in other cell types [15–18], they also accept several pharmacologically relevant peptidomimetics as substrates such as many β-lactam antibiotics, enzyme inhibitors and δ-aminolevulinic acid [17–23]. In the present study we used the intestinal cell line Caco-2 which exclusively expresses PEPT1 [22] and the renal SKPT cell line which expresses PEPT2 but not PEPT1 [22,24,25] to investigate interaction and transport of alafosfalin with mammalian peptide transporters.

Materials and methods


The human colon carcinoma cell line Caco-2 was obtained from the German Collection of Microorganisms and Cell Cultures (Braunschweig, Germany). The renal cell line SKPT-0193 Cl.2, established from isolated cells of rat proximal tubules, was provided by U. Hopfer (Case Western Reserve University, Cleveland, OH, USA) [24,25]. Cell culture media and most supplements were purchased from Life Technologies, Inc. (Karlsruhe, Germany). Fetal bovine serum was obtained from Biochrom (Berlin, Germany). Insulin, dexamethasone and glycylsarcosine (Gly-Sar) were from Sigma Chemie (Deisenhofen, Germany). d-[1-14C]Mannitol (specific radioactivity, 56 mCi·mmol−1) and [glycine-1-14C]glycylsarcosine ([14C]Gly-Sar; specific radioactivity, 53 mCi·mmol−1) were supplied from Amersham Biosciences UK, Ltd (Little Chalfont, UK). Bovine apo-transferrin was purchased from ICN Biomedicals (Eschwege, Germany). l-Alanyl-l-1-aminoethylphosphonic acid (alafosfalin) was from Fluka (Buchs, Switzerland).

Cell culture

Cells were cultured in 75 cm2 culture flasks as described previously [20,24,26–28]. The intestinal cell line Caco-2 was cultured in minimum essential medium supplemented with nonessential amino acid solution (1%, v/v), fetal bovine serum (10%, v/v) and gentamicin (45 µg·mL−1). For uptake studies, cells were subcultured in 35 mm disposable Petri dishes (BD Biosciences, Heidelberg, Germany). With a starting cell density of 83 000 cells·cm−2, cultures reached confluence the next day. Uptake experiments were carried out seven days after seeding. For flux studies, cells were cultured on permeable polycarbonate Transwell® cell culture inserts (24 mm diameter, 3 µm pore size; Corning Costar Bodenheim GmbH, Germany) with a starting cell density of 44 000 cells·cm−2 for 21 days [20].

Culture medium for SKPT cells was Dulbecco's modified Eagle's medium/F12 nutrient mixture (1 : 1, v/v) supplemented with fetal bovine serum (10%, w/v), gentamicin (45 µg·mL−1), epidermal growth factor (10 ng·mL−1), insulin (4 µg·mL−1), dexamethasone (5 µg·mL−1) and apotransferrin (5 µg·mL−1). Cells were subcultured in Petri dishes with a starting cell density of 83 000 cells·cm−2 or in Transwell® chambers with a seeding density of 133 000 cells·cm−2. Uptake and flux studies in SKPT cells were performed four days after seeding.

Transport studies

Uptake of [14C]Gly-Sar was determined at room temperature [22,24,26,27]. The uptake buffer was 25 mm2-(N-morpholino)ethanesulfonic acid/tris(hydroxymethyl) aminomethane (Mes/Tris, pH 6.0) with 140 mm NaCl, 5.4 mm KCl, 1.8 mm CaCl2, 0.8 mm MgSO4 and 5 mm glucose. Uptake experiments were initiated by removing the culture medium from the dishes, washing the cell monolayers twice with 1 mL buffer and adding 1 mL of uptake buffer containing [14C]Gly-Sar and unlabeled compounds. After incubation for the desired time (typically 10 min) buffer was removed and the monolayers were quickly washed four times with ice-cold uptake buffer, dissolved and prepared for liquid scintillation spectrometry.

Transepithelial flux of [14C]mannitol and [14C]Gly-Sar across cell monolayers cultured on permeable filters was measured at 37 °C at pH 6.0 at the apical (1.5 mL) and 7.5 at the basolateral side (2.6 mL) [20]. After washing the inserts with buffer, the Transwell® chambers were preincubated for 10 min and the transepithelial electrical resistance was measured. Flux experiments were started by adding incubation buffer containing radiolabeled compounds and/or alafosfalin to the donor side (apical or basolateral compartment). At time intervals of 10, 30, 60 and 120 min, samples were taken from the receiver compartment and replaced with fresh buffer. After 2 h, the filters were quickly washed twice with ice-cold uptake buffer and cut out of the plastic inserts. Uptake and transepithelial flux of [14C]mannitol and [14C]Gly-Sar were quantified by liquid scintillation counting. Protein was measured according to the procedure of Bradford.

Short-circuit current measurement

Caco-2 cell monolayers cultured for 14 days in Snapwells™ (1.13 cm2, 3 µm pore size, starting cell seeding density 200 000 cells per well; Corning Costar GmbH) were mounted in Ussing chambers (K. Mussler, Aachen, Germany), kept at 37 °C and supplied with 5% CO2/95% O2 (v/v). Experiments were started by adding 5 mL incubation buffer, pH 6.0, to the mucosal side and 5 mL buffer, pH 7.5, to the serosal side. An automated voltage/current clamp apparatus with Ag/AgCl electrodes was used for continuous monitoring of tissue potential difference, current and tissue resistance. Chambers were short-circuited 15 min after mounting the cells. After a further 5–6 min, alafosfalin was added by replacing 1 mL buffer at the mucosal side by 1 mL of a 35 mm solution of alafosfalin (pH 6.0) to obtain a final alafosfalin concentration of 7 mm. Currents with a resolution of 1 µA·cm−2 were recorded every 6 s for a time period of 5 min.

Data analysis

Each experimental point represents the mean ± SE of three to six measurements. Inhibition constants (Ki) were calculated from IC50 values as described previously [24,26].Flux data were calculated after correction for the amount removed by linear regression of appearance rates in the receiver well vs. time [20]. Statistical analysis was performed by the two-tailed nonparametric U-test. A P < 0.05 was considered significant.

Results and Discussion

Affinity of alafosfalin for PEPT1 and PEPT2

Caco-2 and SKPT cell cultures are well established systems for intestinal or renal peptide transport studies. It has been unequivocally shown in several investigations, by functional studies, RT-PCR, Northern blot analyses and using antibodies, that Caco-2 cells express the low-affinity, high-capacity (‘intestinal’) type peptide transport system PEPT1 whereas SKPT cells express the high-affinity, low-capacity (‘renal’) type system PEPT2 but not PEPT1 [22,24,25]. In the present investigation, we first determined the inhibition of [14C]Gly-Sar uptake by alafosfalin. Gly-Sar is used as the reference substrate for peptide transport studies because of its relatively high enzymatic stability [20,24–27]. Alafosfalin strongly inhibited [14C]Gly-Sar (10 µm) uptake in both cell lines. From the inhibition curves shown in Fig. 1, IC50 values were calculated, i.e. the alafosfalin concentration necessary to inhibit the carrier-mediated [14C]Gly-Sar uptake by 50%. From the resulting IC50 values the apparent Ki values of alafosfalin were calculated using Kt values of Gly-Sar transport of 580 µm for PEPT1 at Caco-2 cells and 71 µm for PEPT2 at SKPT cells as described previously [24,26](see below). Alafosfalin showed a remarkably high affinity to both carriers: Ki values of 194 ± 13 µm for PEPT1 and 74 ± 10 µm for PEPT2 were obtained. For comparison, the inhibition constants (Ki) of Ala-Ala are 80 ± 10 µm for PEPT1 and 6.3 ± 0.3 µm for PEPT2 and those of Ala-Lys are 210 ± 20 µm for PEPT1 and 11.7 ± 0.1 µm for PEPT2, respectively [24,27,28].

Figure 1.

Inhibition of [14C]Gly-Sar uptake by alafosfalin in Caco-2 cells (A) and SKPT cells (B). Uptake of 10 µm[14C]Gly-Sar was measured at pH 6.0 for 10 min in confluent cell monolayers in the presence of increasing concentrations of unlabeled alafosfalin (Caco-2: 0–10 mm, SKPT: 0–3.16 mm). Uptake of [14C]Gly-Sar measured in the absence of the inhibitor [Caco-2: 215.3 ± 18.2 pmol·(10 min)−1·(mg protein)−1, SKPT: 241.7 ± 32.1 pmol·(10 min)−1·(mg protein)−1] was designated as 100% (n = 4). Insets: Uptake rate of [14C]Gly-Sar (10 µm, 10 min, pH 6.0) was measured at two different concentrations of unlabeled Gly-Sar: 50 µm (•) and 500 µm (○) (Caco-2) and 20 µm (•) and 50 µm (○) (SKPT), respectively, in the presence of increasing concentrations of alafosfalin (Dixon-plots).

Next, uptake of [14C]Gly-Sar (10 µm) was measured for 10 min at pH 6.0 at two different Gly-Sar concentrations (50 and 500 µm for Caco-2 cells, 20 and 50 µm for SKPT cells) in the presence of increasing amounts of alafosfalin. The linear, nonmediated transport components, i.e. simple diffusion plus tracer binding, of 19.2% and 5.1% were determined with an excess amount of unlabeled Gly-Sar (31.6 mm Gly-Sar for Caco-2 cells and 20 mm Gly-Sar for SKPT cells) and subtracted from total uptake rates. The results are presented as Dixon plots (Fig. 1 insets). They reveal linearity at both Gly-Sar concentrations (r2 values > 0.99) with lines intersecting above the abscissa in the forth quadrant as expected for a competitive inhibitor. For alafosfalin, Ki values of 140 µm for PEPT1 and 73 µm for PEPT2 were calculated from the point of intersection. In a third series of experiments, the effects of alafosfalin on the kinetic parameters, Kt and Vmax, of Gly-Sar uptake were determined in both cell lines. The uptake of [14C]Gly-Sar (10 µm) was measured at increasing concentrations of unlabeled Gly-Sar in the absence and presence of alafosfalin in the uptake medium at a fixed concentration of 260 µm (Caco-2) or 80 µm (SKPT). The Eadie–Hofstee plots (uptake rate vs. uptake rate/substrate concentration, data not shown) gave two straight lines for each cell type intersecting near the ordinate. The maximal velocities of Gly-Sar uptake (Vmax) in both cell lines were not affected by alafosfalin. In contrast, the apparent affinity constants of Gly-Sar uptake (Kt) were strongly affected by alafosfalin; in the presence of alafosfalin the Kt value of Gly-Sar uptake in Caco-2 cells was 1160 µm compared to 580 µm in the absence of the inhibiting phosphonodipeptide. The same alafosfalin effect was observed in SKPT cells where the Vmax values were 2.2 ± 0.1 nmol·(10 min−1)·(mg protein)−1 in the absence and 2.0 ± 0.2 nmol·(10 min−1)·(mg protein)−1 in the presence of alafosfalin, but the Kt value of Gly-Sar uptake was doubled at the alafosfalin concentration close to its Ki value from (Kt =) 71 µm to 138 µm. Both types of experiments show that the inhibition of Gly-Sar uptake by alafosfalin is of the competitive type.

The disappearance of alafosfalin from the luminal compartment during these experiments was determined by HPLC after incubating the cells for 10, 30, 60 and 120 min with alafosfalin at a concentration of 1 mm. The recovery rates were 93 ± 0.7% (Caco-2) and 96 ± 0.3% (SKPT) after 10 min and 81 ± 3% (Caco-2) and 90 ± 0.8% (SKPT) after 2 h, respectively.

Effect of alafosfalin on transepithelial [14C]Gly-Sar flux

PEPT1 and PEPT2 are expressed in the apical membranes of intestinal (PEPT1) and renal (PEPT1 or PEPT2) epithelial cells. Absorption of intact di- and tri-peptides and related mimetics at the intestinal epithelium or their reabsorption at the renal epithelium requires both their uptake from the luminal compartment into the cells and their transport (efflux) across the basolateral membranes to the blood side. Basolateral peptide transport systems have been described both at the intestine and in the kidney [25, 29–31]. Therefore, alafosfalin should not only inhibit the uptake but also the transepithelial net flux of PEPT1 and PEPT2 substrates. For such flux studies, Caco-2 and SKPT cells were cultured on permeable Transwell® filters for 21 days (Caco-2) and four days (SKPT), respectively. After these times, the transepithelial electrical resistances reached maximal values of 803 ± 29 Ω·cm2 (Caco-2) and 452 ± 13 Ω·cm2 (SKPT). To rule out that alafosfalin has any negative effect on the integrity of the monolayers, the transepithelial [14C]mannitol flux (10 µm) was measured for up to 2 h in the absence (control) and presence of 3 mm alafosfalin. At Caco-2 cells, the apical to basolateral [14C]mannitol flux was 0.07–0.08 ± 0.005%·cm−2·h−1 without and with 3 mm alafosfalin in the uptake medium. At SKPT cells mannitol flux values of 0.28 ± 0.01%·cm−2·h−1 in the absence and 0.27 ± 0.01%·cm−2·h−1 in the presence of 3 mm alafosfalin were obtained by linear regression of flux data. Also, the mean [14C]mannitol uptake of 0.05 ± 0.004%·cm−2·(2 h)−1 into both cell types was not affected.

In contrast, alafosfalin (3 mm) strongly reduced the transepithelial flux of [14C]Gly-Sar (10 µm) across Caco-2 cell monolayers from 165.8 ± 5.5 pmol·cm−2·h−1 (= 1.11 ± 0.04%·cm−2·h−1) by 73% to 45.1 ± 0.7 pmol·cm−2·h−1(Table 1). [14C]Gly-Sar uptake into the cells was also reduced by 73% (Table 1). In SKPT cells flux rates were 45.1 ± 2.1 pmol·cm−2·h−1 (= 0.3 ± 0.01%·cm−2·h−1) in the absence and 45.7 ± 3.2 pmol·cm−2·h−1 in the presence of 3 mm alafosfalin. Hence, in SKPT cells we did not find an inhibitory effect of alafosfalin on the [14C]Gly-Sar flux. The [14C]Gly-Sar uptake into the cells, however, was strongly reduced by alafosfalin by 61% (Table 1).

Table 1. Effect of alafosfalin on the transepithelial flux and the intracellular accumulation of [14C]Gly-Sar at Caco-2 and SKPT cell monolayers. Transepithelial flux and uptake of 10 µm[14C]Gly-Sar from apical to basolateral (Ja-b) and from basolateral to apical (Jb-a) side, respectively, was measured in the absence (control) and presence of 3 mm alafosfalin (n = 4–6).
pmol·cm−2·(2 h)−1
pmol·cm−2·(2 h)−1
Control165.8 ± 5.5100135.6 ± 2.710045.1 ± 2.11009.6 ± 1.3100
Alafosfalin45.1 ± 0.72736.1 ± 1.02745.7 ± 3.21013.7 ± 0.439
Control22.9 ± 0.2147.4 ± 0.25.435.2 ± 2.5782.8 ± 0.129
Alafosfalin19.3 ± 0.3126.2 ± 0.34.637.1 ± 2.7822.6 ± 0.127

We also studied basolateral uptake and transepithelial basolateral to apical flux of Gly-Sar (10 µm). In Caco-2 cells, [14C]Gly-Sar flux values were 22.9 ± 0.2 pmol·cm−2·h−1(= 0.09 ± 0.001%·cm−2·h−1) without and 19.3 ± 0.3 pmol·cm−2·h−1 with 3 mm alafosfalin (Table 1). In SKPT cells, the basolateral to apical [14C]Gly-Sar flux rates were 35.2 ± 2.5 pmol·cm−2·h−1 (= 0.14 ± 0.009%·cm−2·h−1) without and 37.1 ± 2.7 pmol·cm−2·h−1 with 3 mm alafosfalin, respectively. Moreover, uptake rates into the cells were inhibited only insignificantly (Table 1). The results show that in both cell types basolateral uptake and basolateral to apical flux rates of Gly-Sar were very much lower compared to the apical to basolateral transport and that the basolateral [14C]Gly-Sar uptake in Caco-2 and SKPT cells could not be inhibited by an excess amount of alafosfalin. Hence, we found no evidence for the interaction of alafosfalin with the putative basolateral peptide transporters.

Transport of alafosfalin – short-circuit current measurements

Inhibition of PEPT1- and PEPT2-mediated Gly-Sar uptake and flux by alafosfalin is important new information on structural requirements of H+/peptide symporters. It shows that phosphonodipeptides interact with mammalian PEPT1 and PEPT2 with high affinity. It does not necessarily mean that the phosphonodipeptide is actually transported into the cells. Interaction with PEPT1 and PEPT2, the fact that bacterial permeases take up alafosfalin [11–14] and the known oral availability of alafosfalin, can only be regarded as circumstantial evidence for actual transport. Therefore, we exploited the electrogenic nature of H+/peptide symporters in an Ussing-type of experiment to obtain more direct evidence. Figure 2 shows the inwardly directed short-circuit current (Isc) across Caco-2 cell monolayers cultured in Snapwells® and mounted in Ussing chambers. Current increased rapidly after addition of alafosfalin (7 mm) to the apical (luminal) compartment by 4.3 ± 0.3 µA·cm−2 whereas transepithelial resistance of 175 ± 2 Ω·cm2 remained unchanged. As expected, Gly-Sar but not phenylalanine (both 10 mm) also induced an increase in current (ΔIsc = 3.4 ± 0.2 µA·cm−2, Fig. 2 inset). Thwaites and coworkers reported that net Gly-Sar transport (20 mm) was associated with an increase of the inward short-circuit current ΔIsc of 6 µA·cm−2[29].

Figure 2.

Stimulatory effect of alafosfalin (7 mm) on the short-circuit current (Isc) across Caco-2 cells. Alafosfalin was added 20 min after mounting the Caco-2 Snapwell™ inserts in the Ussing chambers. pH values were 6.0 at the apical (mucosal) side and 7.5 at the basolateral (serosal) side. Inset: Total increase of short-circuit current by alafosfalin (7 mm, n = 3), Gly-Sar (10 mm, n = 6) and phenylalanine (10 mm, n = 3).

We conclude that the antibacterial phosphonodipeptide alafosfalin interacts with both mammalian H+/peptide symporters with high affinity. PEPT1 and PEPT2 do not seem to differentiate very much between a dipeptide and its derivative where the C-terminal carboxyl group is substituted by a phosphonic function. Moreover, alafosfalin is transported electrogenically across Caco-2 cell monolayers, most likely in a H+-symport. This would explain the high oral availability of small antibacterial peptides of this type. Phosphonodipeptides are interesting compounds for studies on structure-affinity relationships of substrates for PEPT1 and PEPT2.


This study was supported by Land Sachsen-Anhalt grant 3505 A/0403 L and by the Federal Ministry of Education and Research grant # BMBF 0312750 A. We thank Ingelore Hamann for excellent technical assistance and Prof. Martin Luckner (BioService Halle GmbH) for his support. This work will form part of the doctoral thesis of J. N.