MATERIALS AND METHODS

LLC-PK₁-F⁺cells (12) and LLC-PK₁ cells obtained from the American Type Culture Collection were grown to confluence on 60-mm plastic dishes in DMEM containing (in mM) 1.8l-Gln, 0.034l-Glu, 28${HCO}_3^{-}$, and either 5 or 20d-glucose (5 for F⁺ cells) as well as 10% fetal bovine serum (Hyclone, Ogden, UT) in 5% CO₂-95% atmosphere. Experiments were performed 4–8 days after seeding with fresh medium exchanged daily. Seed plates are routinely screened forMycoplasma with contamination monitored by use of a Mycoplasma PCR primer kit (Stratagene Cloning Systems, La Jolla, CA).

Experimental design.

These experiments were designed to characterize each component of the γ-GT-Glu transporter unit depicted in Fig.1. The two components, extrinsic plasma membrane γ-GT activity and the intrinsic membrane Glu transporter, were separately assessed under physiological conditions, i.e., medium Glu and Gln concentration at 0.034 and 0.9 mM, respectively. The balance between extracellular Glu formation and uptake was then determined from the medium Glu content after 45 min of incubation in DMEM. To confirm that Gln conversion to Glu indeed occurs extracellularly and that this conversion is an expression of γ-GT activity, formation of [¹⁴C]Glu from [¹⁴C]Gln was measured in the absence and presence of the γ-GT inhibitor, acivicin (AT-125) (16). Monolayers were presented with DMEM containing 0.9 mMl-Gln and tracer amounts of radiolabeled [U-¹⁴C]Gln (2 μCi/ml) for 1 and 2.5 min with breakdown rates determined in the presence of Glu transport blockers,d-Asp (5 mM) anddl-threo-β-hydroxyaspartate (THA, 0.5 mM), which are potent inhibitors of the high-affinity Glu transporters (1, 2). The medium was then removed and exchanged for the same medium minus the tracer plus 0.75 mM AT-125 for 15 min, after which the tracer-containing medium plus AT-125 was exchanged for a second determination of the 1- and 2.5- min rates. The rate of extracellular Gln conversion to Glu was then estimated from the counts per minute in Glu divided by the Gln specific activity (cpm/nmol; rates expressed in nmol Gln converted to Glu ⋅ min⁻¹ ⋅ mg protein⁻¹). Simultaneously measured Glu accumulation, determined from the increment in medium Glu concentration times volume (2.5 ml), provided a comparison between γ-GT-generated Glu and that accumulating via efflux from the cells.

To determine the Glu transporter activity under physiological conditions, monolayers were incubated with DMEM containing 0.034 mM Glu plus tracer amounts of [U-¹⁴C]Glu (2 μCi/ml, 158 mCi/mmol, Sigma, St. Louis, MO). After 1 min, the radioactive medium was suctioned off, and dishes were placed on ice and repeatedly (6 times) washed with ice-cold PBS (pH 7.4). Ice-cold 5% TCA (1 ml) was added to the cells scraped into tubes and homogenized with a Polytron (half-speed for 20 s). After transfer of to 1.5-ml Microfuge tubes, the homogenates were centrifuged (10,000g for 10 min), the supernatants were retained for analysis of radioactive Glu, and the pellets were processed for protein determination by the dye binding assay with BSA as the standard (4). One-minute Glu uptake rate was determined from the monolayer radioactive Glu (cpm) divided by the medium Glu specific activity (rates expressed in nmol ⋅ min⁻¹ ⋅ mg protein⁻¹). Radiolabeled Glu accounted for 92 ± 4% of the total cellular radioactivity after 1 min.

To assess the role of extracellular Glu formation on intracellular Gln conversion to Glu and flux through the GDH pathway, 1 mM PAH was added to DMEM containing 0.9 mM l-Gln over 45 min. PAH increases the maximal velocity of γ-GT-catalyzed hydrolysis of Gln (30). To confirm that indeed more Glu was being formed extracellularly, parallel experiments were carried out in which the Glu transport step was blocked with the rise in medium Glu accumulating in the presence of PAH taken as the enhancement of γ-GT-Glu production. To confirm that more Glu entered the cells in the presence of PAH, the medium and cellular Glu contents were measured after 45 min. A demonstrable increase in extracellular Glu accumulation with Glu uptake blocked and the rise in cellular Glu without extracellular accumulation were then taken to indicate an increased flux of extracellular Glu through the γ-GT-Glu transporter unit.

To assess the effect of modulating Glu delivery on intracellular Gln and Glu metabolism, estimates of PDG and GDH fluxes were made. Intracellular Gln-to-Glu conversion rate was determined as an index of PDG flux as previously described (32). Briefly, the monolayers were first incubated for 45 min at 37°C in DMEM containing either 1.8 or 0.9 mM l-Gln, after which samples were taken, and the medium was replaced with fresh DMEM containing radiolabeled [U-¹⁴C]Gln (2 μCi/ml, 250 mCi/mmol, NEN, Boston, MA) plus Glu transport blockers (see above) to prevent uptake of extracellularly formed Glu. After 1 min of incubation (37°C, 5% CO₂), the plates were processed as above, and the supernatant was fractioned by HPLC to isolate radioactive Glu and Gln (32). The rate of intracellular Glu formation from Gln was then determined from monolayer radioactive Glu (cpm) divided by medium Gln specific activity (rates expressed in nmol ⋅ min⁻¹ ⋅ mg protein⁻¹) and taken as an estimate of flux through the PDG (Fig. 1, reaction 1). The rate of Gln uptake (expressed in nmol ⋅ min⁻¹ ⋅ mg protein⁻¹) was also determined in these experiments (32). Note that unidirectional Gln transport into the cell measured at 1 min exceeded the intracellular conversion to Glu by at least a factor of 3 in both the cell lines; consequently the transport step is not rate limiting for the intracellular conversion. In addition, >82% of the monolayer radioactivity could be recovered in combined Gln plus Glu peaks at 1.0 min, consistent with limited conversion of radiolabeled Glu over this time course; a longer time course would, of course, reflect the contribution of downstream reactions rather than the initial glutaminase flux. Although the intracellular conversion rates obtained are only an estimate of the intracellular glutaminase flux, they do provide a comparison of the same estimated flux between the two cell lines (under similar conditions) as well as within the same cell line (before and after lowering or raising cellular Glu).

We previously utilized d-Asp (10 mM for 18 h) to block l-Glu uptake and lower intracellular Glu (32). In the present acute 45-min study, d-Asp (5 mM) was used to both block l-Glu uptake and to deliver an acid load. To confirm thatd-Asp effectively blocks the uptake, radiolabeledl-[U-¹⁴C]Glu (2 μCi/ml) was added to the DMEM and disappearance of the labeledl-Glu was determined after 45 min. In the presence of d-Asp, 97 ± 2 and 95 ± 4% of the radiolabeled l-Glu remained in the medium in the LLC-PK₁-F⁺and LLC-PK₁ cell lines, respectively; control monolayers, on the other hand, removed 51 ± 2 and 60 ± 4 of the radiolabeledl-Glu, respectively (n = 3 pairs for each line). These results are in agreement with our previous study showing >95% inhibition ofl-[U-¹⁴C]Glu uptake by 10 mM d-Asp (32). To confirm that 5 mM d-Asp would lower cell pH as well as block Glu uptake under these conditions, monolayers were grown to confluence in especially designed chambers and loaded with 2,7-biscarboxyethyl-5(6)-carboxyfluorescein (BCECF, Molecular Probes, Eugene OR) by exposing them to 10 μM of the ester BCECF-AM for 15 min in HEPES-buffered DMEM minus phenol red and FCS. The chamber was then viewed with an epifluoroscope (Olympus IMT-2) equipped with a fluorescence detector (Photon Technology Instrument 710 PMT), and the fluorescence was measured at 440 and 495 nm with a 530-nm emission wavelength. The high-K⁺-nigericin technique, essentially as described by Thomas et al. (29), was used to calibrate cell pH. Estimated cell pH was 7.55 ± 0.09 in HEPES-buffered DMEM at 35°C and promptly decreased 0.23 ± 0.08 units with exchange of the DMEM to one with 5 mMd-Asp medium; this decline in estimated cell pH was maintained for at least 15 min. Thus the Glu transporter’s hypothesized dual role in modulating the glutaminase and GDH fluxes was tested, since both the block ofl-Glu uptake and decrease in cellular pH were demonstrated.

To assess flux through GDH vs. transamination (reaction 4 vs.5), ${NH}_4^{+}$and alanine (Ala) formation were measured in the absence and presence of d-Asp or PAH. The steady-state uptake of Gln and formation of${NH}_4^{+}$, Ala, and Glu were determined from changes in the metabolite concentration over the 45-min incubation times medium volume (3 ml); medium incubated in the absence of monolayers served as a blank (rates, corrected for spontaneous breakdown, expressed in nmol ⋅ min⁻¹ ⋅ mg protein⁻¹). A shift in the${NH}_4^{+}$-to-Ala formation ratio was taken as evidence consistent with a fall in cell pH.

Analyses.

Medium and monolayer Glu, Gln, Ala, and Asp concentrations were determined on TCA extracts as previously described (32). Briefly, amino acids in the supernatants were derivatized witho-phthaldialdehyde, separated by HPLC, and quantitated with a fluorometric detector; retention times for Asp, Glu, Gln, and Ala were 5.0, 7.8, 11.3, and 16.0 min, respectively. Samples were spiked with homoserine as an internal standard. Recoveries of stock radiolabeledl-[¹⁴C]Glu andl-[¹⁴C]Gln added to the column were 95 ± 6 and 87 ± 5%, respectively (n = 3).${NH}_4^{+}$ concentration was measured by the microdiffusion method (32) and protein by dye binding (4) with BSA as a standard. γ-GT activity was assayed in situ as described (19). Medium${HCO}_3^{-}$ concentration was routinely measured at the termination of these 45-min experiments as a check on spontaneous acidosis, which could influence the metabolic pathways (19,21). However, over this 45-min interval, there was little change in${HCO}_3^{-}$ concentration (27.3 ± 0.3 and 27.4 ± 0.3 vs. initial 28 mM for LLC-PK₁-F⁺and LLC-PK₁ monolayers, respectively), which would not be expected to affect Gln utilization under these conditions.

Statistical comparisons were made between the two cell lines with the unpaired Student’s t-test or between untreated monolayers and those exposed tod-Asp or PAH for 45 min with the paired Student’s t-test; for multiple group comparisons, control, AT-125, andd-Asp, ANOVA (repeated measurements) and a corrected t-test (Bonferroni) were used. When directional changes were predicted based on the a priori hypothesis (Fig. 1), a one-tailedt table was employed; otherwise a two-tailed t table was consulted.

RESULTS

The Glu content measured after 45 min of incubation in medium containing 1.8 mM l-Gln and 34 μM l-Glu is shown in Fig.2 (control). In the LLC-PK₁-F⁺cells, Glu accumulated in excess of that preformed at the rate of 0.52 ± 0.18 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹, which contrasted with a deficit in medium Glu content observed in the LLC-PK₁ cell line (−0.91 ± 0.07 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹,P < 0.001). The Glu transporter activity measured over 1 min at the preformed medium Glu concentration (34 μM) was not different in the two cell lines (1.08 ± 0.13 and 1.38 ± 0.08 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹ for LLC-PK₁-F⁺and LLC-PK₁, respectively; seematerials and methods for details). Therefore the accumulation as opposed to deficit in the medium Glu content cannot be explained by a difference in transporter activity operating on the preformed medium Glu. On the other hand, γ-GT activity was very different in the two cell lines (51 ± 6 and 28 ± 5 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹) when assayed in situ with the artificial substrate γ-glutamyl-p-nitroanalide (seematerials and methods) and in confirmation of the originally observed difference between the two cell lines when assayed in homogenates (11, 12). In support of the medium Glu differences being attributable to the different γ-GT activity, AT-125 was added to the medium and Glu content was determined after 45 min; γ-GT activity measured after 45 min was reduced >90% in both cell lines (3 ± 3 and 2 ± 3 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹ for the LLC-PK₁-F⁺and LLC-PK₁, respectively). With almost complete elimination of γ-GT activity in the LLC-PK₁-F⁺cells (Fig. 2, AT-125), medium Glu accumulation reversed to uptake (0.52 ± 0.18 to −0.41 ± 0.08 nmol ⋅ min⁻¹ ⋅ mg⁻¹, control vs. AT-125, P < 0.001) and further exacerbated the Glu deficit already existing in the medium in the LLC-PK₁ cell line (−0.91 ± 0.07 to −1.26 ± 0.06 nmol ⋅ min⁻¹ ⋅ mg⁻¹,P < 0.01).

To confirm that extracellular Glu was indeed formed from Gln via γ-GT activity, the initial rate of extracellular radiolabeledl-[¹⁴C]Gln conversion tol-[¹⁴C]Glu was monitored in the absence and presence of AT-125; in addition, 5 mMd-Asp and 0.5 mM THA were present to prevent uptake of the extracellularly generatedl-[¹⁴C]Glu (see materials and methods,Experimental design). As shown in Fig. 3, Gln is really broken down to Glu extracellularly, which accumulates under these conditions at a surprisingly high rate (2.1 ± 0.4 and 1.6 ± 0.2 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹ at 1 and 2.5 min). In the presence of AT-125, this extracellular conversion was virtually eliminated, falling to only 15 and 8% of the control rates (0.31 ± 0.32 and 0.12 ± 20 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹ at 1 and 2.5 min). Furthermore Glu accumulated in the medium at a rate equal to the extracellular hydrolysis rate at 1 min (2.9 ± 0.6 vs. 2.1 ± 0.4 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹, respectively) and significantly higher at 2.5 min (3.5 ± 0.6 vs. 1.6 ± 0.2 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹,P < 0.05). In the AT-125-treated monolayers, Glu accumulation far exceeded the negligible extracellular formation rate at 2.5 min (1.5 ± 0.3 vs. 0.12 ± 0.20 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹), demonstrating cellular Glu efflux in the presence ofd-Asp as previously shown in Glu-loaded membrane vesicles (8, 28). Collectively, these results show that extracellular Glu generated from Gln via γ-GT (Fig. 1) provides substrate for the Glu transporter and that extracellular Glu may reflect de novo formation or transporter efflux or both.

The role that γ-GT and Glu transporter operating as a unit play in maintaining the cellular Glu content in these two cell lines is shown in Fig. 4. In the LLC-PK₁-F⁺cells, Glu content was significantly higher than in the LLC-PK₁ cells (175 ± 6 vs. 131 ± 33 nmol/mg, P < 0.05) in line with their greater γ-GT activity and Glu transport. With the extracellular source of Glu eliminated, either by inhibition of γ-GT with AT-125 or by blocking ofl-Glu uptake with 5 mMd-Asp, cellular Glu decreased 13 and 32%, respectively, in the LLC-PK₁-F⁺cell line; the same maneuvers reduced cellular Glu 16 and 42% in the LLC-PK₁ cell line, demonstrating that extracellular formation and transport into both cell lines maintains cellular Glu concentration, as depicted in Fig. 1. Becausel-Glu competes with Gln in inhibiting PDG (10, 27), the cellularl-Gln content was measured in these cells and shown in Fig. 5. Surprisingly, the Gln-adapted LLC-PK₁-F⁺cell line showed a nearly threefold higher Gln content compared with the LLC-PK₁ cell line (140 ± 4 vs. 50 ± 19 nmol/mg protein, P < 0.001). Furthermore, the rate at which Gln was taken up into the cells was measured over 1 min by usingl-[¹⁴C]Gln (see materials and methods) and was determined to be 17 ± 1 and 6 ± 2 nmol/mg protein for the LLC-PK₁-F⁺and LLC-PK₁ cells, respectively (P < 0.01). Thus the threefold greater Gln transport rate under these conditions appears to maintain the higher cellular Gln concentration in the LLC-PK₁-F⁺cell line. Neither AT-125 nord-Asp significantly affected the Gln content of either cell line. Consequently, the ratio of the glutaminase inhibitor l-Glu to the natural substrate l-Gln is actually highest in the LLC-PK₁cells (2.6 vs. 1.25 in the LLC-PK₁-F⁺cells) and reduced with d-Asp to 2.0 and 1.03 in the two cell lines, respectively, entirely as the result of the fall in l-Glu.

In line with the inhibitor-to-substrate ratio, the estimated intracellular glutaminase flux (Fig. 6) was significantly lower in the LLC-PK₁compared with LLC-PK₁-F⁺cells (1.7 ± 0.4 vs. 4.2 ± 0.7 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹,P < 0.05).d-Asp accelerated this estimated glutaminase flux 47% in the LLC-PK₁-F⁺cells (4.2 ± 0.7 to 6.1 ± 1.1 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹,P < 0.05) and 118% in the LLC-PK₁ cells (1.7 ± 0.4 to 3.7 ± 0.5 nmol ⋅ min⁻¹ ⋅ mg⁻¹,P < 0.05). These results show that the glutaminase flux is under Glu transporter control in both cell lines even though strongly modulated by the upregulated Gln transporter activity in the LLC-PK₁-F⁺cells and, moreover, that because of this the LLC-PK₁ cell line is under tighter Glu transporter control, at least with the 1.8 mMl-Gln DMEM.

The rates for Gln uptake and ${NH}_4^{+}$ and Ala formation measured over the 45-min time course are presented in Fig.7. Steady-state Gln uptake was 2.2-fold higher in the LLC-PK₁-F⁺cell line (17.6 ± 0.8 vs. 8.1 ± 0.5 nmol ⋅ min⁻¹ ⋅ mg⁻¹,P < 0.001), consistent with the 2.8-fold higher unidirectional uptake flux measured over 1 min (above). Ala formation was 3.2-fold higher in the LLC-PK₁-F⁺cell line (11.8 ± 1.2 vs. 3.7 ± 0.6 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹,P < 0.01) in contrast to${NH}_4^{+}$ formation, which was only 1.6-fold greater in the LLC-PK₁-F⁺cell line (6.5 ± 0.5 vs. 4.1 ± 0.2 nmol ⋅ min⁻¹ ⋅ mg⁻¹,P < 0.05). The large disparity between Gln uptake and ${NH}_4^{+}$ formation in the LLC-PK₁-F⁺cell line as opposed to Ala formation suggests that the Gln is largely metabolized via the Gln transamination pathway under these conditions (25). The effect of d-Asp on the rates of ${NH}_4^{+}$ and Ala formation is also shown in Fig. 7. In the LLC-PK₁-F⁺cells, d-Asp increased${NH}_4^{+}$ formation 58% (10.3 ± 1.0 vs. 6.5 ± 0.5 nmol ⋅ min⁻¹ ⋅ mg⁻¹,P < 0.001) and in the LLC-PK₁ cells 83% (7.5 ± 0.4 vs. 4.1 ± 0.2 nmol ⋅ min⁻¹ ⋅ mg⁻¹,P < 0.001). The Ala formation rate was unchanged in the LLC-PK₁-F⁺cell line (11.8 ± 1.2 and 12.1 ± 1.3 nmol ⋅ min⁻¹ ⋅ mg⁻¹), so that the ratio of ${NH}_4^{+}$ to Ala formation rose consistent with the intracellular Glu formed being metabolized predominantly through the GDH pathway and${NH}_4^{+}$ production rather than Ala formation. In the LLC-PK₁ cells, this effect is more clear-cut due to the lower Ala formation apparently largely coupled to the PDG flux and less to the Gln transamination pathway activity (13). In these cellsd-Asp resulted in a small but significant (P < 0.01) rise in Ala formation (3.7 ± 0.6 to 4.5 ± 0.6 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹) consistent with the increased Glu formation via PDG followed by transamination (above Fig. 6); nevertheless, as in the LLC-PK₁-F⁺monolayers, ${NH}_4^{+}$ formation increased even further, so that the ${NH}_4^{+}$-to-Ala formation ratio now rose from 1.1 to 1.7. Thus both cell lines respond tod-Asp with an accelerated PDG flux coupled predominantly to GDH and ${NH}_4^{+}$formation.

To test the Glu transporter regulation under conditions approximating those found in vivo, LLC-PK₁-F⁺monolayers were exposed to 0.9 mMl-Gln, and the above measurements were carried out. Interestingly, cellular Glu concentration remained unchanged (Fig. 8) compared with incubation in 1.8 mMl-Gln (166 ± 13 vs. 175 nmol/mg protein); in contrast, cellular Gln content decreased nearly 50% compared with 1.8 mM l-Gln (69 ± 12 vs. 140 ± 4 nmol/mg, respectively,P < 0.001). Consequently, the ratio of Glu to Gln inside these cells doubled (2.4 vs. 1.2), consistent with a greater degree of Glu transporter control over the glutaminase flux. Indeed the estimated Gln-to-Glu conversion rate measured over 1 min was markedly reduced at 0.9 mM (2.8 ± 0.3 vs. 6.5 ± 0.5 nmol ⋅ min⁻¹ ⋅ mg⁻¹,P < 0.001). Note that now Gln uptake (8.4 ± 1.0 vs. 17.6 ± 0.85 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹,P < 0.01) and Ala formation (7.7 ± 0.2 vs. 11.8 ± 1.2 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹,P < 0.01) were both reduced compared with the respective rates at 1.8 mMl-Gln. After 45 min in 5 mMd-Asp, cellular Glu content had fallen 28% (166 ± 13 to 120 ± 10 nmol/mg,P < 0.01) without a change in the Gln content, effectively lowering the Glu-to-Gln ratio (2.4 to 1.7). The reduction in this ratio was associated with a 45% increase in the glutaminase flux (2.9 ± 0.2 to 4.2 ± 0.3 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹,P < 0.01), a 71% increase in${NH}_4^{+}$ formation (3.6 ± 0.5 to 6.1 ± 0.2 nmol ⋅ min⁻¹ ⋅ mg⁻¹,P < 0.01), and a tendency for Ala production to increase (7.7 ± 0.2 to 8.9 ± 0.7 nmol ⋅ min⁻¹ ⋅ mg protein⁻¹,P = 0.06), but nevertheless resulting in a marked rise in the ${NH}_4^{+}$-to-Ala formation ratio (0.47 to 0.69). Thus, under near-physiological medium Gln and Glu concentrations, Glu transporter control of glutaminase flux appears tighter as the Gln uptake rate falls off.

To assess the effect of upregulating extracellular Glu formation on PDG and GDH fluxes in LLC-PK₁-F⁺cells incubated in 0.9 mM l-Gln, the γ-GT-mediated conversion of extracellular Gln to Glu was enhanced by PAH (1 mM). The effect of adding PAH to the medium was an elevation in the monolayer Glu content measured after 45 min of incubation (189 ± 12 vs. 174 ± 18 nmol/mg for PAH and control, respectively,P < 0.05) consistent with increased extracellular formation and uptake. Although the medium Glu content showed a net uptake (−0.37 ± 0.21 vs. −0.52 ± 0.18 nmol ⋅ min⁻¹ ⋅ mg⁻¹for PAH and control, respectively), an increased rate of Glu formation could in fact be demonstrated behind ad-Asp block. In these experiments monolayers were incubated withd-Asp (5 mM) ord-Asp plus PAH and the rate of medium Glu accumulation was measured. In the presence of PAH, Glu accumulation rate was 39% higher than in thed-Asp-treated monolayers (3.40 ± 0.40 vs. 2.44 ± 0.26 nmol ⋅ min⁻¹ ⋅ mg⁻¹,P < 0.02). The effect that increasing the Glu uptake has on the steady-state formation of${NH}_4^{+}$ and Ala is presented in Fig.9. In the presence of PAH overall${NH}_4^{+}$ formation increased 41% (4.1 ± 0.2 to 5.8 ± 0.4 nmol ⋅ min⁻¹ ⋅ mg⁻¹,P < 0.01) without increasing flux through the transamination pathway (5.2 ± 0.5 vs. 5.6 ± 0.6 nmol ⋅ min⁻¹ ⋅ mg⁻¹for control and PAH, respectively). Note that the increased${NH}_4^{+}$ formation could not be accounted for by the increased extracellular Gln breakdown (1.7 vs. 1.0 nmol ⋅ min⁻¹ ⋅ mg⁻¹for increased overall ${NH}_4^{+}$ formation vs. extracellular Gln hydrolysis). Consequently, the difference must be derived from either the intracellular glutaminase flux or flux through the GDH pathway. However the estimated intracellular Gln breakdown rate actually decreased 31% (2.2 ± 0.2 vs. 3.2 ± 0.2 nmol ⋅ min⁻¹ ⋅ mg⁻¹,P < 0.05) consistent with the elevated cellular Glu content and the shift of intracellular Glu into the ammoniagenic pathway.

DISCUSSION

Our goal was to elucidate the role of the Glu transporter in modulating cellular Gln metabolism, specifically its role in regulating flux through the PDG and GDH pathways (Fig. 1, reactions 1 and 4). Our previous studies using LLC-PK₁-F⁺cells grown on porous supports had shown that either blocking or slowing Glu entrance into these monolayers resulted in a 40–50% reduction in cellular Glu concentration and a two- to threefold increase in the intracellular conversion of Gln to Glu (32). In those studies, Glu uptake was limited by using AT-125 to inhibit γ-GT-generated extracellular Glu production or, alternatively, by using d-Asp to inhibit Glu uptake, both treatments for an 18-h period before assessing the PDG flux. In the present design the parental LLC-PK₁ cell line was utilized as a model for Glu availability-limited Glu transport, since γ-GT activity was only one-half that expressed in the Gln-adapted LLC-PK₁-F⁺cell line (11, 12). Furthermore, these responses were studied after only 45 min to assess whether acute regulation could, in fact, be demonstrated. Consequently, on the basis of the available extracellular Glu and the above hypothesis, we expected that the Gln flux through PDG should be more suppressed in the Gln-dependent LLC-PK₁-F⁺cells, a paradoxical scenario given that these cells were selected to grow on Gln as their sole energy source (12). Indeed, we too observed the twofold higher γ-GT activity as originally reported (11, 12) and, as expected from the model shown in Fig. 1, an increase in cellular Glu concentration in the Gln-dependent cell line (Fig. 4 and Ref. 32). However, unexpected on the basis of Glu transport alone was the greater flux through the PDG reaction in the Gln-adapted LLC-PK₁-F⁺cells. How is this discrepancy to be explained? First, we had overlooked the obviously implied possibility that competitive inhibition requires factoring Gln uptake into the model, and second, if the cellular l-Gln concentration were higher in the LLC-PK₁-F⁺cells as the result of Gln transporter activity, then the actual PDG flux would depend on the existing ratio of Glu to Gln. In fact, the ratio of inhibitor Glu to substrate Gln was approximately half that found in the LLC-PK₁ cell line, and as a consequence flux through the PDG reaction was actually higher in the LLC-PK₁-F⁺cells in line with a greater dependence on Gln as a fuel source.

The benefit of framing the hypothesis shown in Fig. 1 and having two closely related cell lines expressing very different γ-GT activity is that potentially more of the underlying control mechanism is revealed, in this case the importance of the previously unrecognized adaptive increase in l-Gln uptake. In the original hypothesis (32), we had overlooked the adaptive nature of the role of Gln transporter(s) in maintaining cellular Gln simply because our studies were limited to the LLC-PK₁-F⁺cell line. In the present comparative study, the Gln-adapted cells expressed a two- to threefold higher Gln uptake (Fig. 1,reactions 6 and7), whether measured as unidirectional Gln transport or as steady-state uptake (Fig. 7). Of course, the two- to threefold increase in Gln transport would be expected to elevate the cellular Gln concentration, overriding the 40–50% increase in cellular Glu, as in fact we observed (Figs. 4and 5). That this was dependent primarily on Gln transport activity was supported by the fall in cellular Gln with reduced transport activity at the near-physiological medium Gln concentration. Previous studies (24, 26) had shown that the LLC-PK₁ cell line expressed a number of transport systems capable of transporting Gln, namely, systems A, ASC, and L. The fact that the adaptive Gln uptake was paralleled by similarly elevated Ala release (Fig. 1,reaction 6) is consistent with transstimulation of Gln uptake by intracellular Ala (5). Note that although both γ-GT and PDG pathways were enhanced in the LLC-PK₁-F⁺cells and may yield Ala, their combined activity would not account for the large Ala production, raising the question as to the actual pathway metabolizing the Gln. A likely possibility is Gln transamination by the so-called glutaminase II pathway, actually an initial transamination followed by deamidation (6, 25). According to these reactions (Fig. 1,reaction 3), Gln would first undergo transamination, forming Ala and α-ketoglutaramate, a product that secondarily undergoes deamidation via an ω-amidase. That this pathway, at least the transamination step, is operative in the parental LLC-PK₁ cell line was clearly shown by use of¹⁵N-amino-labeled Gln; Sahai et al. (25) found ¹⁵N-labeled Ala and accumulation of α-ketoglutaramate, suggesting that the secondary ω-amidase reaction is only loosely coupled to the transamination step (7). If so, the elevated Gln uptake exhibited by the LLC-PK₁-F⁺cells might be metabolized to Ala via this pathway with Ala efflux coupled to and driving the high Gln influx (Fig. 1,reaction 6) (5). Note that, in the LLC-PK₁ cell line, this pathway was not ammoniagenic, so that a considerable discrepancy exists between Ala and ${NH}_4^{+}$ formation consistent with what we observed (Fig. 7). Thus an adaptive increase in Gln transport coupled to Gln transamination yielding Ala would readily account for the enhanced Gln utilization observed to occur in the LLC-PK₁-F⁺cell line.

A truer assessment of the Glu transport regulation of PDG flux required experiments performed at the near-physiological Gln concentration (0.9 vs. ∼0.8 mM for in vivo) (3). Under this condition, Glu transport regulation of PDG flux is much tighter as both Gln uptake and the cellular concentration declined ∼50% (Fig. 8). In line with suppressed glutaminase flux, both intracellular Glu formation and${NH}_4^{+}$ production were reduced ∼50% compared with the rates at 1.8 mM l-Gln. Note that the fall in cellular Gln concentration was associated with a decrease in Ala release, suggesting that Gln utilization by the Gln transamination pathway is concentration dependent as previously noted (6). Under these conditions, blocking Glu uptake withd-Asp or enhancing uptake by PAH activation of extracellular Glu production resulted in a corresponding drop or elevation in cellular Glu. In response, Gln flux through PDG either rose or fell, demonstrating regulatory control of this rate-limiting reaction via the Glu transporter.

We also considered an additional role of the Glu transporter activity, apart from the transfer ofl-Glu, that potentially accelerates Glu flux through the GDH pathway (Fig. 1,reaction 4). The basis for this is the Glu transporter’s acidifying effect (15) and the enhancement of flux through the GDH pathway with cellular acidosis (19, 21, 25). In cells expressing competing GDH and alanine aminotransferase pathways (Fig. 1, reactions 4 and5), acidosis would favor flux through GDH, so that the ratio of ${NH}_4^{+}$ to Ala formation rises as a consequence (19). We thus deployed both the${NH}_4^{+}$ formation rate and this ratio as indexes of GDH flux under these conditions in the two cell lines; in addition, these indexes were monitored at 0.9 mMl-Gln to reduce the high “background” Ala formation in the LLC-PK₁-F⁺cell line. To ensure a high transporter turnover, we used 5 mMd-Asp, which is readily transported (1), or enhanced extracellular Glu formation to drive the transporter. Under this condition (fall in cellular pH),${NH}_4^{+}$ production increased and to an extent greater than that attributed to either the increased flux through the PDG pathway, as is the case ford-Asp, or to the increased γ-GT-associated ${NH}_4^{+}$ formation, as is the case for PAH. These findings are therefore consistent with increased Glu deamination and ${NH}_4^{+}$ formation as a consequence of Glu transporter activity and cellular acidosis. In contrast, Ala formation remained unchanged or barely increased, indicative of a near-constant flux through the transamination pathway. Consequently, in the LLC-PK₁ cell line, the ${NH}_4^{+}$-to-Ala formation ratio rose from 1.1 to 1.7, whereas for LLC-PK₁-F⁺cells, at 0.9 mM l-Gln, the ratio rose from 0.5 to 0.7. Thus these findings are consistent with and predicted from the Glu transporter’s role in acidifying the cells and accelerating the GDH ammoniagenic flux.

Finally, d-Asp clearly promoted the efflux of l-Glu from the cells as shown by the accumulation of medium Glu with both γ-GT and Glu uptake blocked. In this regard, it should be noted that the Glu transporter expressed in plasma membrane vesicles exhibits an exchange reaction of extracellular d-Asp for l-Glu (8, 28); thus an exchange reaction might account for the high Glu efflux and accumulation in the medium. On the other hand, its also possible that Glu efflux is proton driven (20), so thatd-Asp transport, by delivering an acid load, might indirectly effect Glu efflux as well as flux through GDH. Noteworthy, acidogenic hormones, i.e., growth and parathyroid hormones (22, 31) as well as prostaglandin E₂ (unpublished observation), have been shown to mimic these effects on cellular metabolism, suggesting that this mechanism may have physiological and pathophysiological relevance. Thus the Glu transporter may play yet another role in cellular Glu metabolism by shunting metabolically generated Glu out into the medium. Although we do not know the source of the large Glu efflux from these cells in the presence ofd-Asp, it is not unreasonable to suspect that a major fraction is derived from the intracellular Gln conversion to Glu; if so, that would place PDG-generated Glu in a cytosolic pool accessible to the transporter as suggested by Kvamme et al. (18) and possibly explain why plasma membrane transporters exert this extraordinary control over a major metabolic pathway. Further studies utilizing labeled Gln and cellular pH measurements will be required to elucidate the source and mechanism underlying this obviously important transporter-mediated Glu efflux.