An active mechanism for completion of the final stage of protein degradation in the liver, lysosomal transport of dipeptides.

Accumulation of products of proteolysis (e.g. dipeptides) in lysosomes may have pathological consequences. In the present experiment we have investigated the existence of a dipeptide transporter in a membrane preparation of liver lysosomes using Gly-3H-Gln as the probe. The results showed that (a) there was transport of Gly-Gln into an osmotically reactive space inside the lysosomal membrane vesicles; (b) transport was stimulated by acidification (pH 5.0) of the external medium; (c) there was a coupling between transport of protons and Gly-Gln with a stoichiometry of 1:1; (d) the presence of both acidic pH and membrane potential was necessary for uphill transport of Gly-Gln; (e) a single transporter with a Km of 4.67 mM mediated the uptake of Gly-Gln; and (f) Gly-Gln uptake was inhibited by dipeptides and tripeptides but not by amino acids. The results suggest the presence of a low affinity proton-coupled oligopeptide transporter in the liver lysosomal membrane which mediates transfer of dipeptides from a region of low dipeptidase activity (intralysosome) to a region of high dipeptidase activity (cytosol). In this manner, the transporter provides an active mechanism for completion of the final stage of protein degradation.

Lysosomes are major sites for protein degradation in the liver. Amino acids and dipeptides are the final products of proteolysis within this organelle (1,2). These products must be exported into the cytosol if lysosomes are to retain their biological integrity. Failure to export may result in pathological consequences. This is evidenced by the genetic defect in the export of cystine, which results in accumulation of a high concentration of this amino acid in the lysosomes causing cystinosis. Cystinosis is a progressive systemic disease with many pathological expressions (3).
Although as yet no genetic disease of dipeptide accumulation in lysosomes has been described, there is evidence that such an accumulation would also have pathological effects. This is evidenced by the osmotic protection experiments which determine the rates of release of lysosomal enzymes. Lloyd (4) showed that incubation of lysosomes isolated from rat liver with 0.25 M solutions of dipeptides or their constituent amino acids increased the release of their enzyme contents. Furthermore, the release was greater with dipeptides than with their constituent amino acids. He suggested that amino acids and dipeptides enter the lysosomes, and the entry is more efficient for dipeptides than for amino acids. Recently, Bird and Lloyd (5) found that the rate of enzyme release was greater with dipeptides containing L-isomers than D-analogs, suggesting a stereospecificity in lysosomal permeation of dipeptides. However, as reviewed by Forster and Lloyd (6), the osmotic protection technique does not allow any conclusion regarding the mechanism which a solute uses for entry into the lysosomes. For example, the greater rate of enzyme release by dipeptides containing Lrather than D-amino acids could be explained by the hydrolysis of dipeptides within the lysosomes which is stereospecific (7). This is consistent with a more recent study of Bird and Lloyd (8) who found a positive correlation between ability of dipeptides to disrupt lysosomes and their susceptibility to hydrolysis by lysosomal enzymes.
To determine whether the liver lysosomal membrane indeed possesses a dipeptide transporter, in the present experiments we have used a membrane vesicle uptake technique which has been validated for the studies of solute transport by the liver lysosomes. Jonas and Jobe (9) showed that properties of rat liver lysosomal membrane, like the proton-pumping ATPase activity, are preserved in vesicle preparations. For the present experiment we used labeled glycylglutamine (Gly-Gln) as the probe because we have previously validated the use of this dipeptide for characterization of oligopeptide transport in membrane vesicles prepared from kidney and intestine (10,11).
Preparation of Lysosomal Membrane Vesicles-Rats were killed and their livers were removed for preparation of lysosomal membrane vesicles (LMV). Vesicles were prepared as described by Symons and Jonas (12). Briefly, livers were homogenized in buffer 1 (250 mM sucrose, 20 mM Hepes, 1 mM EDTA, pH 7.0, with Tris base) containing 1 g/ml pepstatin and 2 g/ml leupeptin, using a Dounce homogenizer. The liver homogenate was suspended at an average protein concentration of 35 mg/ml in buffer 1 and centrifuged for 20 min at 750 ϫ g. The nuclear pellet was discarded and the centrifugation was repeated. The resulting supernatant was centrifuged for 10 min at 20,000 ϫ g, and the pellet was resuspended in buffer 2 (250 mM sucrose, 20 mM Hepes, pH 7.0, with Tris base) and recentrifuged for 10 min at 20,000 ϫ g. The final pellet was suspended in 9 ml of buffer 2 and mixed with 11 ml of isotonic Percoll (1 ml of 2.5 M sucrose, 200 mM Hepes/Tris, pH 7.0, mixed with 9 ml of Percoll). The resulting Percoll mixture was centrifuged for 90 min * 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  at 40,000 ϫ g. The dense brownish lysosomal band near the bottom of the gradient was removed and diluted with buffer 2 and centrifuged for 10 min at 20,000 ϫ g. The resulting pellet containing lysosomes was incubated in freshly prepared solution containing 5 mM methionine methyl ester, 2 mg/ml bovine serum albumin, 2 mM magnesium chloride, 2 g/ml leupeptin, and 1 g/ml pepstatin for 15-20 min at 37°C. Following this treatment, an equal volume of ice-cold isotonic Percoll was added. The mixture was centrifuged for 30 min at 35,000 ϫ g. A dark brown band located at the top of this dense gradient containing purified LMV was removed and diluted in buffer 2 and centrifuged at 20,000 ϫ g. The purified LMV pellet was resuspended in transport buffer sufficient to provide 3-5 mg/ml protein content. All of the above procedures were done at 4°C. Protein content of the preparation was assessed with the Bio-Rad protein assay. The purity of the LMV preparation was investigated by determining enrichment of organelle and membrane marker enzymes. The activities of ␤-hexosaminidase, 5Јnucleotidase, cytochrome c oxidase, and ␤-glucosidase in liver homogenate and vesicle suspension were measured by the methods described previously (13)(14)(15).
Gly-Gln Uptake by LMV-Uptake studies by LMV were performed at 23°C using a rapid filtration technique as described previously (10). Gly-Gln uptake was initiated by mixing 20 l of membrane suspension (preloaded with 100 mM KCl, 100 mM sucrose, 20 mM Hepes/Tris at pH 7.0) preincubated with 50 M valinomycin and incubated with 180 l of transport buffer containing 0.1 mM Gly-3 H-Gln. Composition of the transport buffer varied with different experiments and is described in the figure legends. Incubation of vesicles was terminated by injecting 20 l of vesicle mixture into 2 ml of ice-cold stop solution (same composition of the transport buffer but without Gly-3 H-Gln), followed by filtration. The filters were then washed with 5 ml of ice-cold stop solution. The radioactivity associated with filters was counted in a Beckman scintillation spectrometer. Nonspecific binding of the Gly-3 H-Gln was determined by adding the transport solution and vesicles directly to the respective ice-cold stop solution, followed by filtering, washing, and counting.
Calculation and Statistics-Kinetic constants of Gly-Gln transport were determined by applying a nonlinear regression method to the Michaelis-Menten kinetic equation using GRAFIT (Sigma): where V is Gly-Gln uptake in picomoles per milligram of protein per 45 s, [S] is the external Gly-Gln concentration in millimolar, V max is the maximal Gly-Gln uptake, and K m is the concentration of [S] that yielded one-half V max . Each rate of Gly-Gln uptake, which was corrected for nonspecific binding, is given as mean Ϯ S.E. of three replicates. Significant differences between values were determined by Student's t test. Table I, there was no enrichment of either cytochrome c oxidase (mitochondria) or 5Ј-nucleotidase (plasma membrane). In contrast, there was a more than 100-fold enrichment of a lysosomal membrane enzyme (␤-glucosidase). Furthermore, there was only a modest enrichment of our preparation with a non-membrane-associated lysosomal enzyme (␤-hexosaminidase).

Purity of LMV-As shown in
Establishing Uptake as Transport-Preliminary experiments revealed that there was uptake of Gly-Gln by LMV. To determine whether the observed Gly-Gln uptake by LMV was indeed transport into vesicles, the following experiment was performed. The rates of Gly-Gly uptake were measured as a function of medium osmolarity. Details of this experiment are presented in the legend to Fig. 1. The results showed that the rate of uptake decreased as a linear function of the reciprocal of medium osmolarity. Extrapolation of the data to infinite osmolarity showed no uptake, indicating the Gly-Gln was entirely or mostly transport into an osmotically reactive space. The data also indicate that the vesicles were properly resealed during preparation for uptake studies.
Driving Force for Dipeptide Transport-To investigate the driving force for dipeptide transport we determined lysosomal uptake of Gly-Gln in the presence and absence of a sodium or a proton gradient. Details of these experiments are presented in the legend to Fig. 2. As shown in this figure, the imposition of an inwardly directed sodium gradient (Na ϩ concentration out ϭ 100, Na ϩ concentration in ϭ 0) had no effect, but the imposition of an inwardly directed proton gradient (pH out 5.0/pH in 7.0) transiently stimulated Gly-Gln uptake by LMV. During the initial 2 min of incubation all of the uptake values were significantly (p Ͻ 0.01) greater when the outside pH was 5.0 and the inside pH was 7.0 than when both pH values were 7.0.
To investigate whether the above proton gradient indeed served as the driving force for Gly-Gln transport, the following experiment was performed. Prior to imposition of the proton gradient, the vesicles were preincubated with a proton ionophore (CCCP). As shown in Fig. 2, CCCP did not reduce Gly-Gln uptake by LMV, suggesting that the proton gradient was not the driving force in this case.
To investigate whether membrane potential influences transport of Gly-Gln by the liver LMV, the following experiment was performed. The effect of K ϩ -generated diffusion potential (inside negative) in the presence and absence of an acidic pH (5.0) in the transport medium was determined. The membrane potential was generated by replacing the KCl in the transport medium with 100 mM choline chloride while maintaining the inside concentration of KCl at 100 mM. Results showed that imposition of membrane potential, when the pH of the transport medium was 7.0, transiently stimulated Gly-Gln  uptake by the liver LMV (Fig. 3). All of the uptakes during the initial 2 min of incubation were significantly (p Ͻ 0.01) greater in the presence than in the absence of membrane potential. The stimulatory effect of membrane potential was greatly enhanced by acidification of the transport medium, resulting in a near 2-fold overshoot in transport of Gly-Gln at 2 min. Therefore, the presence of a membrane potential and an acidic pH was necessary for observing uphill transport of Gly-Gln.
Characterization of Transport-To investigate the affinity and the number of transporters involved in the uptake of Gly-Gln by membrane vesicles we investigated the rate of uptake as a function of dipeptide concentration. The rates of uptake by the liver LMV were determined in the presence of both a membrane potential and an acidic pH. Before this experiment, it was necessary to determine for how long the rate of Gly-Gln transport was linear. The results showed that for at least 50 s the rate was linear (data not shown); therefore, 45 s was chosen as the incubation time for the experiment detailed in the legend to Fig. 4. Eadie-Hofstee plot of data (Fig. 4) showed the presence of a single system for transport of Gly-Gln. The calculation of kinetic constants by the Michaelis-Menten equation showed a low affinity transporter with a K m value of 4.67 Ϯ 0.80 mM and a V max value of 880.38 Ϯ 3.81 pmol/mg protein/45 s.
Relationship between Transport of Gly-Gln and H ϩ -We used the fluorescent dye acridine orange technique to determine whether there is a coupling between proton and Gly-Gln transport. The results are shown in Fig. 5. There was a less than 5% change in the fluorescence when the liver LMV were added to the medium containing no dipeptide. The fluorescence change became appreciably greater (more than 11%) when the medium contained Gly-Gln.
The results of the above experiment suggested that Gly-Gln and H ϩ are cotransported into the vesicles. To investigate the kinetics of this interaction, we investigated the effect of a range of pH on Gly-Gln transport (Fig. 6). There was a steady decline in the rate of Gly-Gln uptake by the liver LMV when the pH of the transport medium was increased above 5.0 (Fig. 6, right   inset). There was a hyperbolic relationship between the rates of Gly-Gln uptake and H ϩ concentration (Fig. 6). The linearity of the Eadie-Hofstee plot of the data indicated a 1:1 stoichiometry between Gly-Gln and H ϩ uptake (Fig. 6, left inset). Calculation of the Hill coefficient also showed a coupling ratio of 1 (1.036 Ϯ 0.152). Calculation of kinetic constants by the Michaelis-Menten equation showed a K m of 0.36 Ϯ 0.02 M (pK a ϭ 6.4) and V max of 57.01 Ϯ 1.43 pmol/mg protein/45 s.
Substrate Specificity of the Transporter-To investigate the substrate specificity of the transporter in the liver LMV, we determined the inhibitory effect of representative groups of amino acids, dipeptides, and tripeptides on Gly-Gln uptake. The results showed (Table II) inhibition by dipeptides and tripeptides but not by amino acids. Due to poor solubility of tetrapeptides at 30 mM concentration, the effect of these oligopeptides on Gly-Gln uptake was not studied.

DISCUSSION
The present study is the first step in the functional characterization of a peptide transporter in the membrane of liver lysosomes. The information obtained reveals that there are functional similarities between this transporter and those located in the brush border membrane of kidney and intestine (16). For example, it is capable of uphill transport, it is stimulated by an acidic pH and membrane potential, its substrates include dipeptides and tripeptides but not amino acids, and it shows a 1:1 stoichiometry between cotransport of Gly-Gln and H ϩ . The uphill transport may have been more pronounced if the membrane orientation of all the vesicles was inside-out. It has been shown (17) that with the current technique of preparation, half of the LMV are oriented inside-out and the other half are oriented outside-out.
The present results suggest that the driving force for uphill transport is membrane potential which normally exists in lysosomes (inside positive/outside negative). At the pH used in the present experiment (4.5-7.0) Gly-Gln is in zwitterionic form. Therefore, its uptake does not add any electrical charge to the transporter. On the other hand, binding of a proton adds a positive charge to the transporter, making it more available for the influence of membrane potential.
The present results may provide an explanation for the disparity among the results of previous studies looking for the presence of an oligopeptide transporter in the liver. Lombardo et al. (18), using functional analysis, did not find any evidence for the presence of a dipeptide transporter in plasma membrane of rat liver. On the other hand, Fei et al. (19), using Northern analysis, found expression of Pept-1 mRNA in the liver. In view of the present data, it is possible that the basis for a positive Northern analysis in the liver (19) is the presence of an oligopeptide transporter in lysosomes. This suggestion is based on reports of a strong homology between genes encoding different oligopeptide transporters (20).
We think that the physiological function of the transporter we have characterized in the lysosomal membrane is mainly to export rather than to import dipeptides because (a) previous studies have shown that the lysosomal carriers can be demonstrated by either uptake or efflux studies (21); (b) as already mentioned, the vesicles have both inside-out as well as outsideout orientations; and (c) it is generally believed that the conditions across lysosomal membranes are the determinant of solute flux. For example, the inside pH is highly acidic, while the outside pH is above neutral. As we have shown in the present study, these conditions greatly favor transport in the direction of cytosol.
The K m of lysosomal oligopeptide transporter (4.67 mM) is considerably greater than the K m of lysosomal amino acid transporters, which ranges between 0.01 and 0.5 mM (6). Usually a greater K m is associated with a greater transport capacity. This may explain the efficiency of dipeptides over amino acids in causing the release of enzymes when lysosomes were incubated with these substrates in vitro (4).
Finally, the present finding leads to a new concept in protein catabolism. Hydrolysis of proteins, whether in the gut lumen, renal tubules, or in liver lysosomes may result in production of considerable amounts of dipeptides (1,2,22,23). However, the hydrolase activity against dipeptides as studied in the intestine (24,25) is either totally or mostly located in the cell cytoplasm. Furthermore, rates of hydrolysis are much greater when dipeptides are incubated with liver homogenates than with liver lysosomes (26). Therefore, oligopeptide transporters, whether located in the brush border or lysosomal membranes, appear to serve an important physiological function. They provide an FIG. 5. Effect of Gly-Gln on H ؉ transport. Using an ␣ scan fluorescence spectrophotometer, the fluorescence of the incubation medium containing either no (A) or 5 mM (B) Gly-Gln was monitored for 1.5 min before and for 3.5 min after the addition of 40 l of LMV to the cuvettes. Composition of the incubation medium was as follows: 6 M acridine orange, 100 mM sucrose, 100 mM KCl, and 20 mM Hepes/Tris (pH 6.5). Before the addition, LMV were preloaded with 100 mM sucrose, 100 mM KCl, and 20 mM Hepes/Tris (pH 7.0). The contents of the cuvettes were continuously stirred during the experiment.