Synthesis and Characterization of High Affinity Inhibitors of the H /Peptide Transporter PEPT2*

In this study, we describe the rational synthesis and functional analysis of novel high affinity inhibitors for the mammalian peptide transporter PEPT2. Moreover, we demonstrate which structural properties convert a transported compound into a non-translocated inhibitor. Starting from Lys[Z(NO2)]-Pro (where Z is benzyloxycarbonyl), which we recently identified as the first competitive high affinity inhibitor of the intestinal peptide transporter PEPT1, a series of different lysine-containing dipeptide derivatives was synthesized and studied for interaction with PEPT2 based on transport competition assays in Pichia pastoris yeast cells expressing PEPT2 heterologously and in renal SKPT cells expressing PEPT2. In addition, the two-electrode voltage clamp technique in Xenopus laevis oocytes expressing PEPT2 was used to determine whether the compounds are transported electrogenically or block the uptake of dipeptides. Synthesis and functional analysis of Lys-Lys derivatives containing benzyloxycarbonyl or 4-nitrobenzyloxycarbonyl side chain protections provided a set of inhibitors that reversibly inhibited the uptake of dipeptides by PEPT2 with Ki values as low as 10 1 nM. This is the highest affinity of a ligand of PEPT2 ever reported. Moreover, based on the structurefunction relationship, we conclude that the spatial location of the side chain amino protecting group in a dipeptide containing a diaminocarbonic acid and its intramolecular distance from the C atom are key factors for the transformation of a substrate into an inhibitor of PEPT2.

In this study, we describe the rational synthesis and functional analysis of novel high affinity inhibitors for the mammalian peptide transporter PEPT2. Moreover, we demonstrate which structural properties convert a transported compound into a non-translocated inhibitor. Starting from Lys[Z(NO 2 )]-Pro (where Z is benzyloxycarbonyl), which we recently identified as the first competitive high affinity inhibitor of the intestinal peptide transporter PEPT1, a series of different lysine-containing dipeptide derivatives was synthesized and studied for interaction with PEPT2 based on transport competition assays in Pichia pastoris yeast cells expressing PEPT2 heterologously and in renal SKPT cells expressing PEPT2. In addition, the two-electrode voltage clamp technique in Xenopus laevis oocytes expressing PEPT2 was used to determine whether the compounds are transported electrogenically or block the uptake of dipeptides. Synthesis and functional analysis of Lys-Lys derivatives containing benzyloxycarbonyl or 4-nitrobenzyloxycarbonyl side chain protections provided a set of inhibitors that reversibly inhibited the uptake of dipeptides by PEPT2 with K i values as low as 10 ؎ 1 nM. This is the highest affinity of a ligand of PEPT2 ever reported. Moreover, based on the structurefunction relationship, we conclude that the spatial location of the side chain amino protecting group in a dipeptide containing a diaminocarbonic acid and its intramolecular distance from the C␣ atom are key factors for the transformation of a substrate into an inhibitor of PEPT2.
The mammalian H ϩ /peptide cotransporter PEPT2 was initially identified in the brush border membrane of renal proximal tubular cells as the high affinity subtype of the mammalian proton-coupled di-and tripeptide transporters (1). In kidney, PEPT2 is mainly responsible for the rapid and efficient uptake of a large number of different di-and tripeptides as well as various peptidomimetics from the tubular fluids into the cell. As recent studies show, PEPT2 is also expressed in a variety of other tissues such as lung and the central nervous system (2,3), but its primary physiological function in these organs is still not known. Specific inhibitors have been extremely useful in the identification of the function and structure of numerous receptors, transporters, and enzymes, but so far no inhibitors of PEPT2 with the required specificity and affinity have been found. Recently, we identified Lys[Z(NO 2 )]-Pro 1 as the first reversible and competitive inhibitor of the intestinal peptide transporter PEPT1 with a high apparent affinity constant (5-10 M) (4). As PEPT2 is known to have similar but not identical structural requirements for substrate recognition and transport, this study was initiated to analyze whether Lys[Z(NO 2 )]-Pro also inhibits PEPT2 or whether even more effective inhibitors can be obtained by alterations in the structure and hydrophobicity of this compound. Varying lysinecontaining dipeptide derivatives were synthesized and submitted to structure-function analysis in three different expression systems employing competitive uptake studies with the radiolabeled dipeptides D-Phe-Ala and Gly-Sar and two-electrode voltage clamp measurements of transport-generated currents. This procedure led to a new class of high affinity inhibitors for PEPT2. It also allowed the identification of some structural features that turn a Lys-containing dipeptide substrate of PEPT2 into an inhibitor.
Synthesis of Dipeptide Derivatives-All non-commercially available dipeptide derivatives were synthesized according to laboratory standard procedures. Briefly, from Boc-Ala-OH, Boc-Val-OH and N ␣ -Boc-N ⑀ -X-Lys-OH, and N ␣ -Boc-N ␦ -X-Orn-OH and N ␣ -Boc-N ␥ -X-Dab-OH (with X ϭ Z(NO 2 ), Z, Ac, where Dab is diaminobutyric acid), respectively, we prepared the respective N-hydroxysuccinimid esters (ONSu). The activated amino acid derivatives were coupled with 1.5-fold excess of the C-terminal amino acid (Lys(X), Orn(X), Ala, Pro, Val, Sar, or Lys(Boc)). * This work was supported by Grant 2880A/0028G and a fellowship from the Land Sachsen-Anhalt (to I. K.), by Grant Da 190/6-1 from the Deutsche Forschungsgemeinschaft, and by the Fonds der Chemischen Industrie. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The crude products were purified by flash chromatography on silica gel. After removal of the Boc groups by treatment with concentrated formic acid at room temperature, the resulting dipeptide derivatives were characterized as formiates. In the case of Lys(Boc)-Ala, the Z-Lys(Boc)-ONSu was coupled with alanine, and the Z-protecting group was removed by catalytic hydrogenation. The purity of the final products was evaluated by analytical reversed phase-high pressure liquid chromatography, mass spectrometry, and capillary electrophoresis, and it was found to exceed 98% in all cases. The dipeptide derivatives shown in Tables I and III were dissolved at high concentrations in dimethyl sulfoxide before preparing the respective uptake buffers.
Pichia pastoris Strains and Transport Assays in Yeast-Cultures of P. pastoris strains expressing PEPT2 were prepared as described previously (6,7). Cells were pelleted at 3000 ϫ g for 10 min, washed twice with 100 mM potassium phosphate buffer (PPB, pH 6.5), and resuspended to 5 ϫ 10 7 cells ϫ 20 l Ϫ1 PPB. Uptake measurements were performed at 22-24°C by using a rapid filtration technique on 96-well filter plates (HATF type, 0.45-m pore size; Millipore, Eschborn, Germany). In brief, uptake was initiated by mixing 20 l of cell suspension with 30 l of PPB containing [ 3 H]D-Phe-Ala (final concentration, 25 nM) either with or without competitors (final concentration, 0.01-200 M). After 15 min of incubation, uptake was terminated by the addition of 200 l of ice-cold PPB followed by filtration. The filters were washed four more times with 200 l of PPB, removed from the plate with a punch, and transferred into vials. Radioactivity associated with the filter was measured by liquid scintillation counting.
Culture of SKPT Cells and Uptake Measurements-SKPT cells were cultured in Dulbecco's Modified Eagle Medium:Nutrient Mixture F12 (Ham) 1:1 and L-glutamine, fetal bovine serum, recombinant insulin, epidermal growth factor, apotransferrin, dexamethasone and gentamicin as described (5,8). SKPT cells at passage numbers 56 -77 were maintained in 75 cm 2 culture flasks at 37°C in a humidified atmosphere with 5% CO 2 . With a starting cell density of 0.8 ϫ 10 6 cells per dish, the cultures reached confluence within 20 h. Uptake was measured in these cells 4 days after seeding as described previously (5,8). The uptake medium was 25 mM Mes/Tris (pH 6.0) containing 140 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl 2 , 0.8 mM MgSO 4 , and 5 mM glucose. Uptake was initiated after washing the cells for 30 s in uptake buffer by adding 1 ml of uptake medium containing radiolabeled Gly-Sar (10 M) with increasing concentrations of test compounds (0 -0.316 mM). If necessary, the pH of the solutions was corrected before preparing the required dilutions. After incubation for 10 min, the cells were quickly washed four times with ice-cold buffer, solubilized in 1 ml of Igepal ® Ca-630 (0.5% v/v, Sigma) in buffer (50 mM Tris, pH 8.0, 140 mM NaCl, 1.5 mM MgSO 4 ), and prepared for liquid scintillation spectrometry. For each experiment, the samples for protein measurement were prepared and measured as described earlier (5).
Xenopus laevis Oocytes Expressing PEPT2 and Electrophysiology-Female X. laevis were oocytes purchased from Nasco (Fort Atkinson, WI). Surgically removed oocytes were separated by collagenase treatment and handled as described previously (9). Individual oocytes were injected with 30 nl of RNA solution containing 30 ng of PEPT2 cRNA. All electrophysiological measurements were performed after 3-6 days by incubation of oocytes in a buffer composed of 88 mM NaCl, 1 mM KCl, 0.82 mM CaCl 2 , 0.41 mM MgCl 2 , 0.33 mM Ca(NO 3 ) 2 , 2.4 mM NaHCO 3 , and 10 mM MES/Tris at pH 6.5 (modified Barth solution).
The two-electrode voltage clamp technique was applied to characterize responses in current (I) and transmembrane potential (V m ) to substrate addition in oocytes expressing PEPT2. In short, oocytes were placed in an open chamber with a volume of 0.5 ml and continuously superfused with modified Barth solution or with solutions of Gly-Gln and/or the corresponding dipeptide derivatives. Electrodes with resistances between 1 and 10 megaohm were connected to a TEC-05 amplifier (NPI Electronic Instruments, Tamm, Germany). PEPT2-expressing oocytes were voltage-clamped at Ϫ100 mV, and current-voltage (I-V m ) relationships were measured using short (100 ms) pulses separated by 200-ms pauses in the potential range from Ϫ160 to ϩ20 mV. I-V m measurements were made immediately before and 30 s after substrate application when current flow reached steady state. Currents evoked by PEPT2 at a given membrane potential were calculated as the difference of the currents measured in the presence and absence of substrate.
Calculations and Statistics-All data are given as the mean ϮS.E. of three or four independent experiments. Significance of differences between the uptake rates and calculated constants were determined by a nonpaired U test. IC 50 values (i.e. concentration of the unlabeled compound necessary to inhibit 50% of radiolabeled dipeptide carrier-mediated uptake) were determined by non-linear regression using the logistical equation for an asymmetric sigmoid (allosteric Hill kinetics): y ϭ minimum ϩ (maximum Ϫ minimum)/(1 ϩ (X/IC 50 ) Ϫp ), where maximum is the initial Y value, minimum is the final Y value, and the exponent P represents Hill coefficient. Inhibition constants (K i ) were calculated from IC 50 values according to the method published by Cheng and Prusoff (10).

Interaction of Lys[Z(NO 2 )] Dipeptide
Derivatives with PEPT2-We first determined the affinity of Lys[Z(NO 2 )]-Pro (LZNP) and three related dipeptide derivatives containing the N-terminal Lys[Z(NO 2 )] moiety but varying C-terminal amino acids (Ala, Val, Sar). Using competition assays based on PEPT2-expressing P. pastoris yeast cells and the renal cell line SKPT constitutively expressing PEPT2, it became obvious that the four dipeptide derivatives compete with radiolabeled D-Phe-Ala or Gly-Sar uptake via PEPT2 in dose-dependent manners ( Fig. 1A and Table I Lys[Z(NO 2 )]-Val (LZNV) (Fig. 1B), no substrate-evoked inward current could be recorded. Moreover, LZNV was able to inhibit the inward current evoked by 0.2 mM of the dipeptide Gly-Gln. This was also the case for the other Lys[Z(NO 2 )] dipeptide derivatives (data not shown). The inhibition of Gly-Gln-evoked inward current by Lys[Z(NO 2 )]-Val was found to be dose-dependent and reversible, suggesting a competitive mode of action at the substrate binding site of PEPT2 (Fig. 1B). To demonstrate this, we investigated the kinetics of the inhibition of Gly-Sar uptake in SKPT cells caused by LZNV. Gly-Sar uptake was measured over the concentration range of 0.02-5 mM in the absence or presence of LZNV at a concentration of 200 nM. In the absence of the inhibitor, the Michaelis-Menten constant, K t , for Gly-Sar was 108 Ϯ 17 M, and the maximal velocity, V max , was 5.4 Ϯ 0.8 nmol ϫ mg of protein Ϫ1 per 10 min. The corresponding kinetic constants obtained in the presence of LZNV were K t ϭ 168 Ϯ 14 M and V max ϭ 4.5 Ϯ 0.4 nmol ϫ mg of protein Ϫ1 per 10 min. Thus, the presence of the inhibitor at a concentration close to its K i value increased the K t value for Gly-Sar with the V max not altered significantly. This confirmed that Lys[Z(NO 2 )]-Val inhibits PEPT2-mediated Gly-Sar uptake into SKPT cells in a competitive manner. The same experiment was performed in Xenopus oocytes expressing PEPT2. In this system, the K t value for Gly-Gln transport was increased from 92 to 353 M in the presence of LZNV (10 M). The V max values were affected only insignificantly (control, 118 nA; LZNV, 106 nA). Moreover, we determined the inhibition constant (K i ) for LZNV in a Dixon type experiment by measuring current in oocytes at two different Gly-Gln concentrations in the presence of increasing concentrations of LZNV (0 -50 M). The results revealed linearity at all Gly-Gln concentrations with lines intersecting above the abscissa in the fourth quadrant, as expected for a competitive inhibitor. A K i value of 2.1 Ϯ 0.1 M was calculated from the point of intersection.
By moving the Lys[Z(NO 2 )] moiety into the C-terminal position with either an N-terminal Ala or Val residue attached, two compounds (ALZN and VLZN) that also inhibited dipeptide influx in the Pichia and SKPT competition assays with apparent K i values of 10 Ϯ 4 and 3.7 Ϯ 0.1 M (ALZN) and 13 Ϯ 1 M and 0.5 Ϯ 0.02 M (VLZN) were obtained. However, in contrast to LZNA, which did not show any electrogenic transport in oocytes, the reversed peptide sequence ALZN produced an inward current of 11 nA, 46% as high as that generated by Gly-Gln (24 nA) at a membrane potential of Ϫ100 mV (Fig. 2). Moreover, whereas LZNA could completely block Gly-Gln-induced inward currents in oocytes, ALZN failed to cause the full inhibition of Gly-Gln-generated transport currents (Fig. 2). Similar results were obtained for VLZN (data not shown). These data demonstrate that the Lys[Z(NO 2 )] residue provides a potent inhibitor only when present in the N-terminal position of a dipeptide.
Influence of the Physicochemical Properties and the Position of Protecting Groups on Inhibition of PEPT2-To understand why the Lys[Z(NO 2 )] moiety when present in the N-terminal position of a dipeptide is able to block dipeptide transport activity, we synthesized a set of dipeptide derivatives differing in the structure and spatial position of the -amino protecting group. As shown in Table II, the dipeptide derivatives Lys(Z)-Ala, Lys(Boc)-Ala, and Lys(Ac)-Ala competed for uptake of radiolabeled D-Phe-Ala in yeast cells with affinities of 23 Ϯ 6, 26 Ϯ 8, and 64 Ϯ 7 M and for uptake of radiolabeled Gly-Sar in SKPT cells with affinities of 0.9 Ϯ 0.1, 3.2 Ϯ 0.2, and 16 Ϯ 2 M, respectively. Whereas 5 mM Lys(Z)-Ala did not produce any inward current in oocytes, it inhibited the current evoked by 5 mM Gly-Gln in oocytes almost completely by 98% (data not shown). Lys(Boc)-Ala displayed the same characteristics as Lys(Z)-Ala with no transport currents when provided alone, but the dose-dependent inhibition of Gly-Gln generated currents. In contrast, 5 mM Lys(Ac)-Ala showed electrogenic transport with inward currents as high as that of the dipeptide (102% I Gly-Gln ) but failed to inhibit Gly-Gln currents when both compounds were perfused together. The rather hydrophobic Z and Boc functions can therefore, similarly to the Z(NO 2 ) group, Percent I Gly-Gln was taken from the recordings of the I-V relationships representing the current evoked by 5 mM of the test compounds and compared to that generated by 5 mM Gly-Gln at a membrane potential of Ϫ100 mV. ϩ/Ϫ indicates whether or not the compound was able to inhibit the current evoked by 5 mM Gly-Gln.   render the natural dipeptide Lys-Ala into an inhibitor. That the blocked ⑀-amino group of the lysine residue has to be in the N-terminal position for the inhibition of PEPT2 could be confirmed by comparing Lys(Z)-Ala with Ala-Lys(Z). The latter showed an affinity in the yeast transport competition assay of 9 Ϯ 2 M and 1.7 Ϯ 0.3 M using SKPT cells and was transported in oocytes with maximal currents of 82% of that induced by 5 mM Gly-Gln. Employing this series of derivatives, we therefore established that a hydrophobic blocking group at the N ⑀ -amino function of lysine is required for the generation of a transport inhibitor and simultaneously confirmed that this side chain construct only works as an inhibitor when present at the N terminus of a dipeptide.

)]-Xaa dipeptide derivatives for inhibition of [ 3 H]D-Phe-Ala uptake by PEPT2 expressed in yeast and [ 14 C]Gly-Sar uptake into SKPT cells
Next, we synthesized ornithine derivatives with the side chain blocked by a Z(NO 2 ) or a Z group and an alanine residue in the C-terminal position. Orn[Z(NO 2 )]-Ala proved to be a slightly weaker inhibitor for PEPT2 than the corresponding Lys derivative with affinities of 2.0 Ϯ 0.8 (yeast) and 1.7 Ϯ 0.04 M (SKPT), respectively. The affinity of Orn(Z)-Ala in the competition assays was almost identical to that of Lys(Z)-Ala (yeast, 26 Ϯ 10 versus 23 Ϯ 6 M; SKPT, 1.9 Ϯ 0.2 versus 0.9 Ϯ 0.1 M), too. However, in contrast to the Lys derivative, which also clearly was an inhibitor, Orn(Z)-Ala failed to inhibit Gly-Gln-evoked transport currents in oocytes but generated transport currents when perfused alone at 5 mM that reached 65% of currents of 5 mM of Gly-Gln in the same oocytes. In this special case, shortening the side chain with the protecting group just by one CH 2 unit (Orn versus Lys) prevents the inhibition of PEPT2 and maintains the compound's capability for transport. That the distance between C␣ and the protecting group of the side chain of the N-terminal amino acid is indeed a key factor in rendering a compound into an inhibitor was further confirmed by the dipeptide derivatives Dab(Z)-Ala and Dab[Z(NO 2 )]-Ala. They contain a Dab residue instead of an ornithine or lysine residue with the protecting group in N ␥amino position. When provided at a concentration of 5 mM, Dab(Z)-Ala is electrogenically transported by PEPT2 with inward currents of 97% I Gly-Gln and failed to inhibit the current evoked by 5 mM Gly-Gln. Again, adding an NO 2 group renders the substrate into an inhibitor. Dab[Z(NO 2 )]-Ala does not generate inward current. This observation suggests that both the structure as well as the spatial position of the N-protecting group within the substrate binding domain of PEPT2 is crucial for the ability of a compound to block the transport cycle.
Optimizing the Lysine-Dipeptide Derivatives for Inhibition of PEPT2-The demonstration that the rather hydrophobic  dipeptides Lys(Z)-Ala and Ala-Lys(Z) displayed a higher affinity for PEPT2 than most of our earlier tested naturally occurring dipeptides led us to assume that an N ⑀ -protected lysine derivative in both the N-and C-terminal position could be advantageous for obtaining more effective inhibitors. We therefore synthesized lysyl-lysine derivatives containing either a Zor a Z(NO 2 ) group attached at the N ⑀ -amino groups, resulting in Lys(Z)-Lys[Z(NO 2 )] (LZLZN) and Lys[Z(NO 2 )]-Lys[Z(NO 2 )] (LZNLZN). Fig. 3 shows an electrophysiological analysis of PEPT2 function in X. laevis oocytes with Lys[Z(NO 2 )]-Lys[Z(NO 2 )] as the transport inhibitor. Table III summarizes the data obtained on the affinities of these compounds for PEPT2 in yeast as well as SKPT cells. As expected, adding a Z-or Z(NO 2 ) group to a Lys residue, resulting in a higher hydrophobicity of the corresponding dipeptide, increased its affinity for interaction with PEPT2 dramatically. The apparent affinity of  Fig. 4. When these dipeptide derivatives were analyzed for transport in oocytes expressing PEPT2 (Fig. 3A, shown for LZLZN), they did not produce any inward currents but were able to inhibit the current evoked by Gly-Gln completely but reversibly. In the present study, we demonstrate that lysine-containing dipeptides can be converted from transported substrates of PEPT2 into effective transport inhibitors by blocking the N ⑀amino group of an N-terminal lysine residue with either a Boc, Z, or Z(NO 2 ) function. Moving such a modified Lys residue into the C-terminal position of a dipeptide structure surprisingly retains the compound's capacity for electrogenic transport. This strongly suggests that two distinctly different pockets (N-terminal P1 and C-terminal P2 pocket) within the substrate binding domain of PEPT2 accommodate the side chains of dipeptides and derivatives. Although it has been shown that the physicochemical characteristics of the N-and C-terminal residues of dipeptides differently affect their affinity for interaction with PEPT2 (11), the present data demonstrate for the first time that the specificity within the side chains can also discriminate between substrates and inhibitors. To obtain a high affinity inhibitor, a sufficiently long side chain spacer in the N-terminal position carrying a hydrophobic function like the Z or Z(NO 2 ) groups must be provided for the proposed P1 pocket. Shortening the spacer by just one CH 2 unit, as in the ornithine homologue with the Z-function, prevents the inhibition of PEPT2 and allows transport. This suggests that a definite amino acid side chain of the PEPT2 protein specifically interacts with the altered side chain structure in P1, preventing the conformational change of the protein in the initiation of the substrate translocation step. In contrast, the P2 site in PEPT2 appears to be less restrictive to all substrate side chain modifications for transport, although they clearly affect substrate affinity. Fig. 5 provides a model for the proposed asymmetry for binding and transport or inhibition by the modified lysyl-dipeptides. With LZLZN and LZNLZN, we obtained substrates with the highest affinity (40 and 10 nM, respectively, in the yeast assay) reported so far for a ligand of PEPT2. These compounds should be very useful for probing the protein structure of PEPT2 as well as for identifying its biological role in the various cell types and tissues where the protein is expressed.