SLC22A11 Inserts the Uremic Toxins Indoxyl Sulfate and P-Cresol Sulfate into the Plasma Membrane

Chronic kidney disease (CKD) is a global health concern affecting millions worldwide. One of the critical challenges in CKD is the accumulation of uremic toxins such as p-cresol sulfate (pCS) and indoxyl sulfate (IS), which contribute to systemic damage and CKD progression. Understanding the transport mechanisms of these prominent toxins is essential for developing effective treatments. Here, we investigated whether pCS and IS are routed to the plasma membrane or to the cytosol by two key transporters, SLC22A11 and OAT1. To distinguish between cytosolic transport and plasma membrane insertion, we used a hyperosmolarity assay in which the accumulation of substrates into HEK-293 cells in isotonic and hypertonic buffers was measured in parallel using LC-MS/MS. Judging from the efficiency of transport (TE), pCS is a relevant substrate of SLC22A11 at 7.8 ± 1.4 µL min−1 mg protein−1 but not as good as estrone-3-sulfate; OAT1 translocates pCS less efficiently. The TE of SLC22A11 for IS was similar to pCS. For OAT1, however, IS is an excellent substrate. With OAT1 and p-aminohippuric acid, our study revealed an influence of transporter abundance on the outcomes of the hyperosmolarity assay; very high transport activity confounded results. SLC22A11 was found to insert both pCS and IS into the plasma membrane, whereas OAT1 conveys these toxins to the cytosol. These disparate transport mechanisms bear profound ramifications for toxicity. Membrane insertion might promote membrane damage and microvesicle release. Our results underscore the imperative for detailed structural inquiries into the translocation of small molecules.


Introduction
With diminished kidney function, metabolic waste products that are excreted by the kidney under normal conditions accumulate in the body and cause systemic damage.The increased retention of so-called uremic toxins further deteriorates renal function, contributing to the development of chronic kidney disease (CKD) [1,2].CKD is estimated to affect 8-16% of the population worldwide [3,4].Largely protein-bound uremic toxins such as p-cresol sulfate (pCS) and indoxyl sulfate (IS) are a major challenge in the treatment of CKD as they are difficult to remove via dialysis [5] and are associated with various co-morbidities including cardiovascular disease and osteoporosis [6].pCS and IS (see Figure 1 for structures) are formed in the human body by the metabolism of amino acids by intestinal bacteria.Tyrosine and phenylalanine, for example, are degraded to p-cresol [7], which, after absorption, is sulfated in the liver by sulfotransferase 1A1 (SULT1A1) to pCS [8].In contrast, tryptophan is metabolized by bacterial tryptophanase to indole [9], which is hydroxylated in the liver by cytochrome P450 2E1 (CYP2E1) to 3-hydroxy indole [10] and subsequently sulfated by SULT1A1 to IS [11].
pCS and IS are normally excreted from the human body through the renal proximal tubule.The two toxins are mainly cleared from the blood compartment via uptake catalyzed by organic anion transporter 1 (OAT1) and OAT3 [12][13][14].OAT1 (human gene symbol SLC22A6) and OAT3 (SLC22A8) are found in the basolateral membrane of proximal tubular cells; their substrate spectrum, besides pCS and IS, encompasses many other anionic compounds, including metabolites, toxins, and drugs [15].In addition, IS is a substrate of SLC22A11, also known as OAT4 [13].SLC22A11 is located in the apical membrane of renal proximal tubule cells and transports uric acid (UA), glutamate, steroid sulfate conjugates such as estrone-3 sulfate (E3S), and other compounds, including drugs and toxins [15][16][17].Therefore, SLC22A11 was suggested to mediate efflux and reabsorption of IS on the apical side of the proximal tubule [13].It is unknown whether SLC22A11 transports pCS.pCS and IS are normally excreted from the human body through the renal proximal tubule.The two toxins are mainly cleared from the blood compartment via uptake catalyzed by organic anion transporter 1 (OAT1) and OAT3 [12][13][14].OAT1 (human gene symbol SLC22A6) and OAT3 (SLC22A8) are found in the basolateral membrane of proximal tubular cells; their substrate spectrum, besides pCS and IS, encompasses many other anionic compounds, including metabolites, toxins, and drugs [15].In addition, IS is a substrate of SLC22A11, also known as OAT4 [13].SLC22A11 is located in the apical membrane of renal proximal tubule cells and transports uric acid (UA), glutamate, steroid sulfate conjugates such as estrone-3 sulfate (E3S), and other compounds, including drugs and toxins [15][16][17].Therefore, SLC22A11 was suggested to mediate efflux and reabsorption of IS on the apical side of the proximal tubule [13].It is unknown whether SLC22A11 transports pCS.
We have previously uncovered that SLC22A11 transports the substrates E3S and UA via fundamentally different mechanisms [18].Our data suggest that SLC22A11 inserts E3S into the plasma membrane while UA, a regular substrate, is transported into the cytosol [18].Indeed, some transporters, like the long-chain fatty acid transporter FadL of Escherichia coli, can insert certain amphipathic or predominantly hydrophobic molecules into the membrane through a lateral opening [19].In addition, the ability of ABC transporters We have previously uncovered that SLC22A11 transports the substrates E3S and UA via fundamentally different mechanisms [18].Our data suggest that SLC22A11 inserts E3S into the plasma membrane while UA, a regular substrate, is transported into the cytosol [18].Indeed, some transporters, like the long-chain fatty acid transporter FadL of Escherichia coli, can insert certain amphipathic or predominantly hydrophobic molecules into the membrane through a lateral opening [19].In addition, the ability of ABC transporters to retrieve and transport membrane-soluble substrates from the phospholipid bilayer is well known [20].Surprisingly, the insertion into the plasma membrane of small molecules by SLC transport proteins has hardly been explored, yet this alternative route has important implications for transcellular trafficking and the intracellular processing of drugs and toxins.
To better understand the pathophysiological mechanisms of toxicity in humans caused by the uremic toxins pCS and IS and their roles in CKD progression, it is important to identify the underlying transport mechanisms in detail.The current belief invariably holds that uremic toxins such as pCS and IS are transported by organic anion transporters into and out of the cytosol.However, our recent results might challenge this assumption: due to the structural similarity to E3S (see Figure 1), transport proteins could insert pCS (in fact, an exact substructure of E3S) and IS into the plasma membrane.
We have recently established a hyperosmolarity assay to distinguish between regular transport of small molecules into the cytosol and insertion into the plasma membrane [21].In this assay, the uptake of substrates into HEK-293 cells (293 cells) by specific transport proteins is measured in isotonic and hypertonic buffers in parallel.Inhibition of substrate accumulation due to hyperosmolarity indicates transport into the cytosol, whereas stimulation of accumulation indicates insertion into the plasma membrane [21].The decrease in the accumulation of regular substrates in hyperosmolar buffer might be caused by the reduction in cell volume and endocytosis of plasma membrane areas, including transporters, during membrane remodeling [22][23][24][25][26].In contrast, the increase in substrate accumulation could be explained by the formation of invaginations and a decrease in membrane tension due to increased extracellular osmotic pressure.These changes in membrane properties could provide more space for small molecules to be inserted into the plasma membrane [21,27].
The primary aim of the present study was to investigate the mechanisms of transport of pCS and IS via human OAT1 and SLC22A11 in 293 cells using the hyperosmolarity assay.Our results indicate that OAT1 transports pCS and IS into the cytosol, whereas SLC22A11 transports pCS and IS into the plasma membrane.

Results
The mechanisms of transport of the uremic toxins IS and pCS by SLC22A11 and OAT1 were investigated through hyperosmolarity experiments developed previously [21].Transport proteins were inducibly expressed in 293 cells, which were then incubated in the hyperosmolarity assay; the resulting cell lysates were analyzed via LC-MS/MS.A stimulation of substrate accumulation in the hyperosmolarity assay indicates integration into the plasma membrane, whereas inhibition of substrate accumulation by hyperosmolarity indicates transport into the cytosol [21].

OAT1 Abundance Affects the Hyperosmolarity Assay
During a hyperosmolarity assay verification, transport of p-aminohippuric acid (pAH), a model substrate for OAT1, was examined.Surprisingly, OAT1-expressing cells showed a slight increase in accumulation under hyperosmolar conditions after incubation with 10 µM pAH for 20 min, from 3.2 ± 0.1 nmol mg protein −1 to a maximum of 3.8 ± 0.1 nmol mg protein −1 at 400 mM mannitol (Figure 2A).This would indicate insertion into the membrane.However, because of the polar structure of pAH (Figure 1), insertion into the membrane is very unlikely.One possible reason for this discrepancy could be the particularly efficient transport of pAH by OAT1; we have previously measured, for example, 59 µL min −1 mg protein −1 (unpublished).We therefore repeated the assay with reduced expression of OAT1.The reduction was achieved by titrating doxycycline in the culture medium (Figure 3).At 3 ng mL −1 , the accumulation of pAH by OAT1 in 293 cells was reduced to approximately 10% of the maximal accumulation at 1000 ng mL −1 doxycycline.

Figure 2.
Hyperosmolarity assay with reduced concentrations of doxycycline.The analyte content of cell lysates was determined via LC-MS.Each symbol represents a dish.(A) Stably transfected 293 cells with (on; n = 3) or without (off; n = 2) expression of OAT1 were incubated with 10 µM pAH for 20 min at 37 °C in uptake buffer with mannitol, as indicated.The expression of OAT1 was either normal (1000 ng mL −1 doxycycline) or reduced (3 ng mL −1 doxycycline).(B) Stably transfected 293 cells with (on; n = 3) or without (off; n = 2) expression of SLC22A11 were incubated with 10 µM E3S for 60 min at 37 °C in uptake buffer with mannitol, as indicated.The expression of SLC22A11 was either normal (1000 ng mL −1 doxycycline; adjusted presentation of previous data [21]) or reduced (1.4 ng mL −1 doxycycline).(C) As in (B), but with 10 µM UA instead of 10 µM E3S.With reduced expression, pAH levels in OAT1-expressing cells decreased from 44 ± 1 pmol mg protein −1 to a minimum of 19 ± 1 pmol mg protein −1 at 800 mM mannitol (Figure 2A).This result indicates regular transport of pAH via OAT1 into the cytosol.To verify the hyperosmolarity assay in the presence of reduced expression of another transport protein, the transport of the model compounds E3S and UA via SLC22A11 was examined.After a titration experiment with E3S and several doxycycline concentrations, a doxycycline concentration of 1.4 ng mL −1 was determined to yield 10% uptake of E3S into cells via SLC22A11 (Figure 3).At reduced expression of SLC22A11, the UA content of cells decreased linearly with increasing mannitol concentration, whereas E3S levels increased with increasing mannitol concentration (Figure 2B,C).Thus, for SLC22A11, the results at reduced expression, in principle, conform to the established results at normal expression levels; the curves clearly differ in shape, however.

Comparison of IS Transport by SLC22A11 and OAT1
In time course experiments, cells expressing the transporter SLC22A11 or OAT1 (high expression) were incubated with 10 µM IS in uptake buffer with or without 400 mM mannitol (Figure 4A).The accumulation of IS by SLC22A11 over time was stimulated by hyperosmolarity; after 60 min, the accumulation of IS was 0.35 ± 0.01 nmol mg protein −1 (control) and 0.58 ± 0.01 nmol mg protein −1 (400 mM mannitol).In contrast, the accumulation of IS by OAT1 over time was inhibited by hyperosmolarity; after 60 min, the accumulation of IS was 4.6 ± 0.1 nmol mg protein −1 (control) and 3.3 ± 0.1 nmol mg protein −1 (400 mM mannitol) (Figure 4A).This implies that SLC22A11 and OAT1 differ fundamentally in their mechanism of transport of IS.With reduced expression, pAH levels in OAT1-expressing cells decreased from 44 ± 1 pmol mg protein −1 to a minimum of 19 ± 1 pmol mg protein −1 at 800 mM mannitol (Figure 2A).This result indicates regular transport of pAH via OAT1 into the cytosol.To verify the hyperosmolarity assay in the presence of reduced expression of another transport protein, the transport of the model compounds E3S and UA via SLC22A11 was examined.After a titration experiment with E3S and several doxycycline concentrations, a doxycycline concentration of 1.4 ng mL −1 was determined to yield 10% uptake of E3S into cells via SLC22A11 (Figure 3).At reduced expression of SLC22A11, the UA content of cells decreased linearly with increasing mannitol concentration, whereas E3S levels increased with increasing mannitol concentration (Figure 2B,C).Thus, for SLC22A11, the results at reduced expression, in principle, conform to the established results at normal expression levels; the curves clearly differ in shape, however.

Comparison of IS Transport by SLC22A11 and OAT1
In time course experiments, cells expressing the transporter SLC22A11 or OAT1 (high expression) were incubated with 10 µM IS in uptake buffer with or without 400 mM mannitol (Figure 4A).The accumulation of IS by SLC22A11 over time was stimulated by hyperosmolarity; after 60 min, the accumulation of IS was 0.35 ± 0.01 nmol mg protein −1 (control) and 0.58 ± 0.01 nmol mg protein −1 (400 mM mannitol).In contrast, the accumulation of IS by OAT1 over time was inhibited by hyperosmolarity; after 60 min, the accumulation of IS was 4.6 ± 0.1 nmol mg protein −1 (control) and 3.3 ± 0.1 nmol mg protein −1 (400 mM mannitol) (Figure 4A).This implies that SLC22A11 and OAT1 differ fundamentally in their mechanism of transport of IS.In long-term incubations (60 min) in which the uptake of IS into SLC22A11-expressing cells (high expression) was examined as a function of mannitol concentration, the IS levels increased from 0.20 ± 0.01 nmol mg protein −1 (0 mM mannitol) to a maximum of 0.45 ± 0.03 nmol mg protein −1 at about 400 mM mannitol (Figure 4B).In contrast, IS levels in OAT1-expressing cells decreased from 3.0 ± 0.1 nmol mg protein −1 (0 mM mannitol) to a minimum of 2.0 ± 0.1 nmol mg protein −1 at about 600 mM mannitol.In experiments with expression reduced to approximately 10% uptake (3 ng mL −1 doxycycline, see above), IS levels in OAT1-expressing cells actually decreased by more than half, from 0.68 ± 0.01 nmol mg protein −1 (0 mM mannitol) to 0.28 ± 0.01 nmol mg protein −1 (800 mM mannitol) (Figure 4C).In SLC22A11-expressing cells, again the opposite effect was observed, with an increase in IS levels from 0.38 ± 0.02 nmol mg protein −1 (0 mM mannitol) to a maximum of 0.67 ± 0.01 nmol mg protein −1 at about 600 mM mannitol (Figure 4C).These results suggest that SLC22A11 inserts IS into the plasma membrane, whereas OAT1 transports IS into the cytosol.

Comparison of pCS Transport by SLC22A11 and OAT1
In time course experiments (high expression), the accumulation of pCS by SLC22A11 over time was stimulated by hyperosmolarity; after 20 min, the accumulation of pCS was 0.26 ± 0.01 nmol mg protein −1 (control) and 0.37 ± 0.01 nmol mg protein −1 (400 mM mannitol) (Figure 5A).In contrast, the accumulation of pCS by OAT1 over time was inhibited by hyperosmolarity; after 20 min, the accumulation of pCS was 0.77 ± 0.01 nmol mg protein −1 (control) and 0.61 ± 0.01 nmol mg protein −1 (400 mM mannitol) (Figure 5).These results are similar to those obtained for IS (see above) and suggest that there are also different types of transport mechanism for pCS via SLC22A11 and OAT1.

Transport of IP, IC, and IAA by SLC22A11
To investigate whether the sulfate residue in indoxyl sulfate (IS) can be substituted by other moieties with a negative charge in the membrane insertion by SLC22A11, the transport of the uremic toxin indole-3-acetic acid (IAA) as well as indole-3-carboxylic acid (IC) and indoxyl phosphate (IP) (Figure 1) was examined by comparing long-term incubations (60 min) of stably transfected 293 cells in uptake buffer with or without 400 mM mannitol (Figure 6).These experiments showed that the accumulation of IP in SLC22A11expressing cells was stimulated by hyperosmolarity after 60 min; the accumulation of IP was 0.013 ± 0.001 nmol mg protein −1 (control) and 0.077 ± 0.003 nmol mg protein −1 (mannitol).This suggests that IP, like IS and pCS, is inserted into the membrane by SLC22A11.No uptake of IC or IAA into 293 cells via SLC22A11 was observed (Figure 6).(IC) and indoxyl phosphate (IP) (Figure 1) was examined by comparing long-term incubations (60 min) of stably transfected 293 cells in uptake buffer with or without 400 mM mannitol (Figure 6).These experiments showed that the accumulation of IP in SLC22A11expressing cells was stimulated by hyperosmolarity after 60 min; the accumulation of IP was 0.013 ± 0.001 nmol mg protein −1 (control) and 0.077 ± 0.003 nmol mg protein −1 (mannitol).This suggests that IP, like IS and pCS, is inserted into the membrane by SLC22A11.
No uptake of IC or IAA into 293 cells via SLC22A11 was observed (Figure 6).

Discussion
In this study, we examined whether the prominent uremic toxins pCS and IS are actually inserted into the plasma membrane by SLC22A11 or OAT1.This hypothesis was based on the structural similarity of pCS and IS to the steroid hormone precursor E3S (Figure 1); these compounds, when projected on a disc, consist of a large, purely hydrophobic sector and a small, very hydrophilic sector with a negative charge.Our previous data suggest that E3S is inserted into the plasma membrane by several SLC transporters, namely SLC22A11, MATE1, OAT3, SLC10A6, and SLC22A9 [18,21].In contrast, SLC10A1

Discussion
In this study, we examined whether the prominent uremic toxins pCS and IS are actually inserted into the plasma membrane by SLC22A11 or OAT1.This hypothesis was based on the structural similarity of pCS and IS to the steroid hormone precursor E3S (Figure 1); these compounds, when projected on a disc, consist of a large, purely hydrophobic sector and a small, very hydrophilic sector with a negative charge.Our previous data suggest that E3S is inserted into the plasma membrane by several SLC transporters, namely SLC22A11, MATE1, OAT3, SLC10A6, and SLC22A9 [18,21].In contrast, SLC10A1 transports E3S-as a regular substrate-into the cytosol.These assumptions are based in particular on results from the hyperosmolarity assay [21], a technically simple test used to distinguish between small molecule transport into the cytosol and insertion into the plasma membrane.Here, we used this assay to compare transport of pCS and IS via SLC22A11 and OAT1.
Results for pCS and IS are consistent.Most importantly, the transport of both uremic toxins by OAT1 and SLC22A11 was fundamentally different.With SLC22A11, the uptake curves of pCS and IS (Figures 4B and 5B) resemble the previous uptake curves of E3S [21].By contrast, with OAT1, the uptake curves of pCS and IS (Figures 4B and 5B) resemble the uptake curves of regular substrates, for example, the transport of UA by SLC22A11 [21].Therefore, we suggest that SLC22A11 inserts pCS and IS into the plasma membrane, while OAT1 transports pCS and IS as regular substrates into the cytosol.
Transport to these separate compartments likely results in different mechanisms of toxicity.When pCS and IS enter the cytosol of tubule cells, for example, they can bind to and activate the human aryl hydrocarbon receptor (AHR) [6,28], inducing apoptotic and necrotic cell death [29].In addition, oxidative stress is increased, and antioxidant capacity is decreased, leading to tubule cell damage and interstitial inflammation [29][30][31].The damaged renal tubule activates transforming growth factor-β1 (TGF-β1) signaling, further driving interstitial inflammation and renal fibrosis [29].Membrane insertion should have different effects.Indeed, treatment of cells with IS caused membrane damage [32] and the release of microvesicles [33].
Surprisingly, when we tested OAT1 with the model substrate pAH [34] in the hyperosmolarity assay, the uptake was not decreased by the addition of mannitol, but even slightly increased (Figure 2A).Clearly, because of the polar structure of pAH (Figure 1), insertion into the membrane is highly unlikely.So, what went wrong with the assay?The expression of OAT1 in our system, measured via peptide quantification, is good, but not outstanding (external Figure 7 [35]).However, the efficiency of transport of pAH by OAT1 is very high, as confirmed by our present results (Figure 2A); this is based largely on high affinity, since the K m values reported at 5 and 9 µM [36,37] are rather low for a transporter.In our previous paper, we proposed that with regular substrates, the rapid decrease in accumulation could reflect a loss of accessible, active transporters from the cell surface due to the rapid reduction in the cell surface caused by hyperosmolarity [21].However, if the transport activity (= number of active transporters × transport catalyzed by a single transporter molecule) is very high, a transport activity reserve (not detected in the isotonic control) might exist which could counteract the loss of cell surface.We tested this hypothesis by reducing the number of transporter molecules per cell in our expression system.This can be achieved easily by decreasing the concentration of the expression inducer doxycycline (Figure 3).When transporter expression was reduced to approximately 10% of maximal expression, pAH uptake indeed decreased as a function of mannitol concentration (Figure 2A), confirming regular transport of pAH by OAT1 into the cytosol.The hyperosmolarity assay at reduced transporter expression also worked to demonstrate the disparate handling of E3S and UA by SLC22A11 (Figure 2B,C); however, the curves at different transporter numbers clearly differ in shape.For pCS and IS, results are consistent at high and low expression (Figures 4C and 5C).Our present results reveal that the number of transporters per cell may strongly affect the outcomes of the hyperosmolarity assay.A lower level of transporter expression would be preferred, but the activity must stand out from the background.
It is noteworthy that background accumulation of pCS (average k in for transporter expression off, 0.46 µL min −1 mg protein −1 ) was higher compared to that of IS (0.23 µL min −1 mg protein −1 ) or E3S (0.24 µL min −1 mg protein −1 ; data from [18]).This contributes to a poorer signal-to-background ratio; in the uptake equilibrium after 10 min, accumulation of pCS mediated by SLC22A11 above the background (expression off) was much smaller (as low as f = 1.2; Figure 5B, isotonic buffer) than that observed with E3S as the substrate (f = 19; Figure 2B).
In this study, we demonstrated for the first time that pCS is a substrate of SLC22A11 (Figure 5).The efficiency of transport of pCS, extracted from the time courses of accumulation (Figure 5A, isotonic buffer; delta of k in values for expression on and off), was 7.8 ± 1.4 µL min −1 mg protein −1 .Thus, pCS is a relevant substrate of SLC22A11 but not as good as E3S (40 ± 6 µL min −1 mg protein −1 [18]); OAT1-which does not transport E3S [18] -translocates pCS less efficiently (3.6 ± 0.5 µL min −1 mg protein −1 ).The efficiency of transport of IS by SLC22A11 was similar to that of pCS at 5.6 ± 0.3 µL min −1 mg protein −1 (Figure 4A).For OAT1, however, IS is an excellent substrate at 55 ± 3 µL min −1 mg protein −1 .The marked preference of OAT1 for IS may contribute to the fact that in CKD patients, the renal clearance of IS is approximately three times higher than the renal clearance of pCS [38].
Our data agree with a previous report where uptake and efflux of IS by human SLC22A11 alias OAT4 was demonstrated in transfected cells [13]; however, our interpretation is entirely different.It is unclear whether SLC22A11, which is located on the apical side of proximal tubule cells, has a role at all in moving cytosolic IS or pCS out of the cells after basolateral entry via OAT1 or whether its prime effect is to transfer the toxins from the luminal fluid into the cell membrane.Instead, MRP transporters could mediate the apical exit from the cytosol [1,2].
To investigate whether the sulfate residue in indoxyl sulfate (IS) can be substituted by other moieties with a negative charge in the membrane insertion by SLC22A11, the transport of the indole derivatives IAA, IC, and IP was examined using the hyperosmolarity assay.For IP, the uptake into 293 cells by SLC22A11 was stimulated by hyperosmolarity after 60 min of incubation (Figure 6), indicating the insertion of IP into the membrane by SLC22A11.This means that the sulfate group, which represents the polar part in the structures of E3S, pCS, and IS, is not essential for an insertion into the membrane by SLC22A11; a phosphate group is also accepted, albeit at a lower level of equilibrium accumulation.In contrast, a carboxylate moiety is not tolerated since IAA and IC were not transported into 293 cells by SLC22A11 (Figure 6).
Transporter-mediated membrane insertion of suitable compounds is apparently not limited to anionic solutes.Based on the cryo-EM structure of the organic cation transporter 3 (OCT3 alias EMT; gene symbol SLC22A3), a lateral access site (the V-site) was identified through which OCT3 could transport its hydrophobic substrates out of or into the plasma membrane [39].Further cryo-EM structures of SLC transporters, particularly SLC22A11, would help to elucidate how small molecules enter the plasma membrane, providing important insights into the distribution of drugs and toxins in the human body.
In conclusion, our present results, obtained using the hyperosmolarity assay, reveal that SLC22A11 inserts the important uremic toxins pCS and IS into the plasma membrane.By contrast, OAT1 catalyzes movement into the cytosol.Transport to these separate compartments is expected to result in different mechanisms of toxicity.

Plasmids and cDNAs
The transporter cDNAs used in this study were of human origin.The SLC22A11 cDNA was expressed from pEBTetLNC [18], and OAT1 cDNA was expressed from the pEBTetD vector [40].Both vectors are Epstein-Barr virus-derived plasmid vectors that allow doxycycline-inducible protein expression in human cell lines.Construction of pEBTetLNC/SLC22A11h [18] and pEBTetD/OAT1h [41] was described previously.

Cell Culture
A total of 293 cells (ATCC CRL-1573; also known as HEK-293 cells), a transformed cell line derived from human embryonic kidney, were grown as adherent culture in plastic culture flasks (Falcon 353110 and 353112, Becton Dickinson, Heidelberg, Germany) at 37 • C in a humidified 5% CO 2 atmosphere.The growth medium was Dulbecco's Modified Eagle Medium (Life Technologies 31885-023, Thermo Fisher Scientific, Dreieich, Germany) supplemented with 10% fetal bovine serum (FBS Analog, 2224SAMPLE, neoFroxx, Einhausen, Germany), 100 U mL −1 penicillin, and 0.1 mg mL −1 streptomycin (P4333, Sigma-Aldrich); in some experiments with SLC22A11, 10 µg mL −1 ciprofloxacin was used instead of penicillin/streptomycin.The medium was changed every 2-3 days, and the culture was split every 5 days.Stably transfected cell lines were generated, as reported previously [40], using Turbofect (R0531, Thermo Fisher Scientific, Dreieich, Germany).Since pEBTet-derived vectors [18,40] are propagated episomally, we used cell pools rather than single-cell clones.Cell culture medium always contained 3 µg mL −1 puromycin (13884, Cayman chemical, Ann Arbor, MI, USA) to maintain plasmids for up to 8 weeks in culture.Nevertheless, only cells whose transfection was not older than 5 weeks were used for the experiments.

Transport Assays
For measurement of solute uptake, cells were seeded in 6 cm diameter polystyrol dishes (83.3901,Sarstedt, Nümbrecht, Germany) precoated with 0.1 g l −1 poly-L-ornithine (P3655, Merck, Darmstadt, Germany) in 0.15 M boric acid-NaOH pH 8.4 and grown to a confluence of at least 70%.To turn protein expression on, cells were cultivated for at least 20 h with 1 µg mL −1 doxycycline (195044, MP Biomedicals, Eschwege, Germany) in growth medium unless otherwise indicated.Uptake buffer contained 125 mM NaCl, In osmolarity experiments, we either used the 3rd-order polynomial function of GraphPad Prism or simple linear regression.In doxycycline titration experiments, we used the sigmoidal, 4PL (four-parameter logistic) curve with Equation (2).

15 Figure 1 .
Figure 1.Structures of compounds used for investigation.Charges are drawn to indicate the dominating species at pH = 7.4.

Figure 1 .
Figure 1.Structures of compounds used for investigation.Charges are drawn to indicate the dominating species at pH = 7.4.

− 1 )Figure 2 .
Figure 2. Hyperosmolarity assay with reduced concentrations of doxycycline.The analyte content of cell lysates was determined via LC-MS.Each symbol represents a dish.(A) Stably transfected 293 cells with (on; n = 3) or without (off; n = 2) expression of OAT1 were incubated with 10 µM pAH for 20 min at 37• C in uptake buffer with mannitol, as indicated.The expression of OAT1 was either normal (1000 ng mL −1 doxycycline) or reduced (3 ng mL −1 doxycycline).(B) Stably transfected 293 cells with (on; n = 3) or without (off; n = 2) expression of SLC22A11 were incubated with 10 µM E3S for 60 min at 37 • C in uptake buffer with mannitol, as indicated.The expression of SLC22A11 was either normal (1000 ng mL −1 doxycycline; adjusted presentation of previous data[21]) or reduced (1.4 ng mL −1 doxycycline).(C) As in (B), but with 10 µM UA instead of 10 µM E3S.

Figure 4 .
Figure 4. Hyperosmolarity assay reveals different mechanisms for IS transport by OAT1 and SLC22A11.The analyte content of cell lysates was determined via LC-MS.Each symbol represents a dish.(A) Stably transfected 293 cells with (on) or without (off) high expression (1000 ng mL −1 doxycycline) of OAT1 or SLC22A11 were incubated at 37 °C with 10 µM IS for the indicated time in uptake buffer (control) or in uptake buffer plus 400 mM mannitol (n = 2).(B) Stably transfected 293 cells with (on, n = 3) or without (off, n = 2) high expression (1000 ng mL −1 doxycycline) of OAT1 or SLC22A11 were incubated with 10 µM IS for 60 min at 37 °C in uptake buffer with mannitol, as indicated.(C) As in (B), but with reduced transporter expression (3 ng mL −1 doxycycline).

− 1 )Figure 4 .
Figure 4. Hyperosmolarity assay reveals different mechanisms for IS transport by OAT1 and SLC22A11.The analyte content of cell lysates was determined via LC-MS.Each symbol represents a dish.(A) Stably transfected 293 cells with (on) or without (off) high expression (1000 ng mL −1 doxycycline) of OAT1 or SLC22A11 were incubated at 37 • C with 10 µM IS for the indicated time in uptake buffer (control) or in uptake buffer plus 400 mM mannitol (n = 2).(B) Stably transfected 293 cells with (on, n = 3) or without (off, n = 2) high expression (1000 ng mL −1 doxycycline) of OAT1 or SLC22A11 were incubated with 10 µM IS for 60 min at 37 • C in uptake buffer with mannitol, as indicated.(C) As in (B), but with reduced transporter expression (3 ng mL −1 doxycycline).

Figure 5 .
Figure5.Hyperosmolarity assay reveals different mechanisms for pCS transport by OAT1 and SLC22A11.The analyte content of cell lysates was determined via LC-MS.Each symbol represents a dish.(A) Stably transfected 293 cells with (on) or without (off) high expression (1000 ng mL −1 doxycycline) of OAT1 or SLC22A11 were incubated at 37 °C with 10 µM pCS for the indicated time in uptake buffer (control) or in uptake buffer plus 400 mM mannitol (n = 2).(B) Stably transfected 293 cells with (on, n = 3) or without (off, n = 2) high expression (1000 ng mL −1 doxycycline) of OAT1 or SLC22A11 were incubated with 10 µM pCS for 10 min (SLC22A11) or 60 min (OAT1) at 37 °C in uptake buffer with mannitol, as indicated.(C) As in (B), but with reduced expression (3 ng mL −1 doxycycline).

− 1 )Figure 5 .
Figure 5. Hyperosmolarity assay reveals different mechanisms for pCS transport by OAT1 and SLC22A11.The analyte content of cell lysates was determined via LC-MS.Each symbol represents a dish.(A) Stably transfected 293 cells with (on) or without (off) high expression (1000 ng mL −1 doxycycline) of OAT1 or SLC22A11 were incubated at 37 • C with 10 µM pCS for the indicated time in uptake buffer (control) or in uptake buffer plus 400 mM mannitol (n = 2).(B) Stably transfected 293 cells with (on, n = 3) or without (off, n = 2) high expression (1000 ng mL −1 doxycycline) of OAT1 or SLC22A11 were incubated with 10 µM pCS for 10 min (SLC22A11) or 60 min (OAT1) at 37 • C in uptake buffer with mannitol, as indicated.(C) As in (B), but with reduced expression (3 ng mL −1 doxycycline).

with Equation ( 1 )
, where c out represents the substrate concentration and k in and k out are rate constants.y = offset + k in /k out * c out * [1 − exp(−k out * x)]