Yeast Cell-Based Transport Assay for the Functional Characterization of Human 4F2hc-LAT1 and ‐LAT2, and LAT1 and LAT2 Substrates and Inhibitors

In mammalian cells, the L-type amino acid transporters (LATs) LAT1 (SLC7A5) and LAT2 (SLC7A8) form heterodimeric amino acid transporters (HATs) with the ancillary protein 4F2hc and are involved in the cellular uptake of specific amino acids. The HAT 4F2hc-LAT1 is found upregulated in various cancer cell types, while 4F2hc-LAT2 is a transporter for non-cancer cells. Preclinical studies have highlighted that 4F2hc-LAT1 plays an important role in tumor progression representing a valid anticancer target. Consequently, current research is focusing on the development of potent and specific human 4F2hc-LAT1 inhibitors. On the other hand, 4F2hc-LAT2 is emerging as target of other diseases, thus also gaining clinical interest. To determine affinity and specificity of substrates and inhibitors for 4F2hc-LAT1 or 4F2hc-LAT2, robust transport cell assays are indispensable. We have optimized and validated a transport assay using cells of the methylotrophic yeast Pichia pastoris stably overexpressing the human HATs 4F2hc-LAT1 or -LAT2, and the LATs LAT1 or LAT2 alone. The radioligand [3H]L-leucine was used as reporter and the substrates L-leucine, triiodothyronine (T3) and thyroxine (T4) as well as the inhibitors BCH and JPH203 (KYT-0353) for assay validation. Obtained half-maximal inhibitory concentrations also provided new insights, e.g., into the LAT specificity of the potent inhibitor JPH203 and on the potency of the thyroid hormones T3 and T4 to inhibit transport through human 4F2hc-LAT2. The LAT1 and LAT2 assays are of particular interest to determine possible implications and influences of 4F2hc in ligand binding and transport. In summary, the presented assays are valuable for characterization of ligands, e.g., towards 4F2hc-LAT1 specificity, and can also be applied for compound screening. Finally, our established approach and assay would also be applicable to other HATs and LATs of interest.


INTRODUCTION
Amino acids have diverse and essential roles in cell function, e.g., for protein synthesis, metabolism, signal transduction, neural transmission, and cellular growth and proliferation. Transport of amino acids across biological membranes is mediated by amino acid transporters, which are embedded in lipid bilayers of cells (Christensen, 1990;McGivan and Pastor-Anglada, 1994). Malfunction, absence or overexpression of amino acid transporters can affect homeostasis in the body leading to human diseases. The solute carrier (SLC) superfamily includes currently eleven families containing amino acid transporters (Kandasamy et al., 2018). The SLC7 family of amino acid transporters consists of fifteen genes and is split into two subgroups: the cationic amino acid transporters (CATs) and the L-type amino acid transporters (LATs) (Verrey et al., 2004;Fotiadis et al., 2013). CATs comprise the SLC7A1-A4 and SLC7A14 genes, and LATs the SLC7A5-A11, Slc7a12, SLC7A13, and Slc7a15 genes . In contrast to CATs, LATs are not glycosylated. For correct trafficking to the plasma membrane in mammalian cells, LATs associated with type II membrane N-glycoproteins from the SLC3 family, i.e., 4F2hc (SLC3A2; CD98) and rBAT (SLC3A1) (Palacin and . These ancillary proteins (the heavy chains) are covalently connected to the corresponding LATs (the light subunits) through a conserved disulfide bridge to form heterodimeric amino acid transporters (HATs) (Chillaron et al., 2001;Wagner et al., 2001;Palacin and Kanai, 2004;Verrey et al., 2004;Fotiadis et al., 2013). The light subunits are the catalytic subunits of HATs (Reig et al., 2002;Rosell et al., 2014;Napolitano et al., 2015). LAT1 (SLC7A5) and LAT2 (SLC7A8) are isoforms of the system L of amino acid transporters requiring the heavy chain 4F2 (4F2hc) for functional expression at the plasma membrane (Kanai et al., 1998;Pineda et al., 1999;Segawa et al., 1999). Furthermore, we recently showed that 4F2hc can modulate the substrate affinity and specificity of the light chains LAT1 and LAT2 (Kantipudi et al., 2020). In addition to these two LAT specific functions, the ancillary protein 4F2hc has multifunctional roles such as in cell adhesion, cell fusion, integrin signaling and regulation of macrophage activation via galectin-3 (Fenczik et al., 1997;Tsurudome and Ito, 2000;Feral et al., 2005;MacKinnon et al., 2008). 4F2hc-LAT1 is expressed in different tissues and organs (e.g., brain, ovary, placenta and testis), and in relatively high levels at the blood-brain barrier and in several types of tumors Scalise et al., 2018;Häfliger and Charles, 2019). The location and high expression levels make 4F2hc-LAT1 an interesting vehicle for drug delivery into the brain and for cancer cell targeting (Häfliger and Charles, 2019;Puris et al., 2020). In cancer cells, 4F2hc-LAT1 provides neutral and essential amino acids for nutrition and regulation of the mTOR signaling pathway (Nicklin et al., 2009). Thus, inhibition of this HAT represents a valid approach to block migration and invasion of cancer cells, and to induce apoptosis. In contrast, 4F2hc-LAT2 is ubiquitously expressed in the human body and highly expressed in polarized epithelia suggesting a major role of this HAT in transepithelial transport of amino acids (Bröer, 2008;Fotiadis et al., 2013). Thus, both transporters have evolved towards specific functions, e.g., LAT1 for uptake of specific amino acids into growing cells, and LAT2 towards normal cell-type and transcellular amino acid transport. LAT1 and LAT2 are sodium-independent transporters that exchange substrates across membranes with a one-to-one stoichiometry (Verrey et al., 2004;Fotiadis et al., 2013). The substrate specificities of both HATs are comparable, but 4F2hc-LAT2 accepts in addition to large neutral also small neutral amino acids (Pineda et al., 1999;Rossier et al., 1999;Meier et al., 2002). Other substrates of 4F2hc-LAT1 and -LAT2 represent amino acid derivatives such as the thyroid hormones T3 and T4 (Friesema et al., 2001;Zevenbergen et al., 2015). The compound 2aminobicyclo-(2,2,1)-heptane-2-carboxylic acid (BCH)  was described as specific inhibitor of system L inhibiting both, 4F2hc-LAT1 and -LAT2 (Kanai et al., 1998;Segawa et al., 1999). On the other hand, the tyrosine-based JPH203 (KYT-0353) molecule was reported as a competitive, potent and highly specific 4F2hc-LAT1 inhibitor with strong inhibitory effects on the growth of different cancer cells (Oda et al., 2010;Yun et al., 2014;Häfliger et al., 2018). Therefore, transport inhibitors with high specificity towards 4F2hc-LAT1 but not -LAT2 represent promising drug candidates for cancer therapy and diagnosis. In crescentic glomerulonephritis pathogenesis, LAT2 was shown to be upregulated activating the mTORC1 pathway (Kurayama et al., 2011). Thus, LAT2-specific inhibitors might also be interesting and considered therapeutically for crescentic glomerulonephritis and other emerging LAT2-related diseases. Towards discovery of potent and selective inhibitors against 4F2hc-LAT1 or -LAT2, robust assays for ligand screening and functional characterization using cells overexpressing corresponding LATs separately are crucial. Establishment of mammalian cell lines for stable expression of 4F2hc-LAT1 and -LAT2 is not straightforward since most host cell lines express HATs endogenously. As a consequence, the activity of the exogenous LAT is difficult to distinguish from the endogenous one and the limited endogenous pool of 4F2hc is used for both, endogenous and exogenous LATs, thus introducing ambiguities in the assays. Khunweeraphong et al. reported the establishment of stable HEK293 cell lines expressing exogenous LAT1 or LAT2, and using endogenous 4F2hc of the cells to form HATs (Khunweeraphong et al., 2012). HEK293 cells indicated reduced backgrounds of amino acid transport, e.g., reduced contamination of endogenous LAT1 activity (Khunweeraphong et al., 2012), and other advantages compared to the murine S2 cells previously used in a similar endeavor (Morimoto et al., 2008).
The methylotrophic yeast Pichia pastoris represents a wellestablished system for the expression of recombinant human membrane proteins (Byrne, 2015). Cholesterol and derivatives thereof were shown to play an important role in function and stability of LATs Dickens et al., 2017;Cosco et al., 2020). Yeasts such as P. pastoris produce ergosterol (Nes et al., 1978) providing a valuable cholesterol derivate for interaction with heterologously expressed membrane proteins. Towards establishment of a robust assay for ligand screening and functional characterization, we have optimized, applied and validated a previously reported radioligand assay using P. pastoris overexpressing human 4F2hc-LAT1 or -LAT2, and the substrate [ 3 H]L-leucine as radioligand. In contrast to the previously reported HEK293 cell assay (Khunweeraphong et al., 2012), 4F2hc is co-expressed with LAT1 or LAT2 in the Pichia-based assay, thus not limiting 4F2hc availability and boosting expression of HATs (Costa et al., 2013;Rosell et al., 2014;Kantipudi et al., 2020). Interestingly, and in absence of 4F2hc co-expression, Pichia is also able to express functionally the light subunits LAT1 and LAT2 alone (Costa et al., 2013;Rosell et al., 2014;Kantipudi et al., 2020). This allows evaluating possible contributions of the heavy chain 4F2hc on ligand binding and transport inhibition through selected substrates and inhibitors.

[ 3 H]L-Leucine Radioligand Transport Assay
For transport studies, 1 ml of thawed P. pastoris cells (OD 600 40) expressing the corresponding transporter were diluted in 50 ml transport buffer and pelleted by centrifugation (3,000 × g, 15 min, room temperature). The pellet was then washed by resuspending it in 50 mL transport buffer and by repeating the washing procedure (centrifugation and resuspension) two times. Finally, the cell pellet was resuspended in 2 ml of transport buffer and incubated for 20 min at 30°C under agitation (300 rpm, Multitron, Infors HT, Bottmingen, Switzerland). The yeast suspension density was adjusted with transport buffer to an OD 600 of 1.875 (4F2hc-LAT1, LAT1, 4F2hc-LAT2 or LAT2, and untransformed P. pastoris KM71H cells). All transport experiments were performed in a reaction volume of 100 µL.

Data Analysis, Curve Fitting and Statistics
Experiments were performed at least in triplicate. For data analysis, the signal of the untransformed P. pastoris cells was subtracted from the transporter signal to obtain the net transport signal. Data from three independent experiments were taken. In each of these experiments, the net transport signals were averaged and the half maximal inhibitory concentration (IC 50 ) values of homologous (L-leucine) and heterologous (BCH, JPH203, T3, and T4) L-leucine transport competition experiments were determined by fitting a sigmoidal model curve to these data. Every experimental data point was then individually normalized using the corresponding upper plateau values (i.e., the fitted upper plateau value that corresponds to 100%). Single, normalized data points from the three independent experiments were averaged and a sigmoidal model curve was fitted to the data in order to obtain the IC 50 values. Prism6 (GraphPad Software) was used for data analysis.

RESULTS
Human 4F2hc-LAT1, LAT1, 4F2hc-LAT2 or LAT2 were expressed in the methylotrophic yeast P. pastoris. We showed in previous reports that not only the HATs 4F2hc-LAT1 and -LAT2, but also the light subunits LAT1 and LAT2 in absence of ancillary protein are properly folded, correctly trafficked to the plasma membrane and functional in P. pastoris (Rosell et al., 2014;Kantipudi et al., 2020). Transport activities were determined by measuring the uptake of the substrate [ 3 H]L-leucine into Pichia cells at OD 600 0.75 expressing the corresponding HAT or LAT. Time-course experiments showed clear HAT-and LAT-dependent transport activities, which were much higher than the [ 3 H]L-leucine uptake into untransformed host cells (Figure 1). In all cases, saturation of the transport process was observed. Differences in radioligand transport ( Figure 1) are due to different expression levels of corresponding recombinant LATs as estimated from V max /OD values using previously determined kinetic parameters (Kantipudi et al., 2020). Uptake assay times of 10 min (i.e., in the linear regimes-time points indicated by asterisks in Figure 1) and corresponding Pichia cells at OD 600 0.75 were taken for the subsequently presented experiments with 4F2hc-LAT1, LAT1, 4F2hc-LAT2, and LAT2 (Figures 2-6).
As previously reported (Kantipudi et al., 2020), comparison of IC 50 s indicated a modulatory effect of the heavy subunit 4F2 on light subunits, which is most striking between 4F2hc-LAT2 and LAT2 (Figure 2). After validation of the yeast-cell based transport assay with the substrate L-leucine (Figures 1, 2), we pursued validation using the described system L transport inhibitor BCH  and the 4F2hc-LAT1 specific inhibitor JPH203 (Oda et al., 2010). We obtained IC 50 values of 72 µM (4F2hc-LAT1), 78 µM (LAT1), 195 µM (4F2hc-LAT2), and 184 µM (LAT2) for BCH ( Figure 3). These results indicated that the compound BCH inhibits both human HATs being about 2.7-fold more specific for 4F2hc-LAT1. Interestingly, and in contrast to L-leucine (Figure 2), LAT1 and LAT2 in absence of 4F2hc had comparable IC 50 s as their heterodimeric forms (Figure 3) indicating no significant influence of the ancillary protein on BCH binding to LATs of HATs.

DISCUSSION
A cell-based transport assay using the methylotrophic yeast P. pastoris overexpressing the human HATs 4F2hc-LAT1 or -LAT2, and LATs LAT1 or LAT2, and the radiolabeled substrate [ 3 H]L-leucine was optimized and validated using selected substrates and inhibitors. In contrast to our previous study (Kantipudi et al., 2020), we adjusted uptake time and OD 600 for the four stably expressing Pichia clones to same conditions, i.e., to 10 min and OD 600 0.75 (Figure 1). For comparison and validation of the adjusted uptake conditions, IC 50 s were determined for the substrate L-leucine (Figure 2)-the here studied transporters have high affinities for L-leucine relative to the other proteinogenic amino acids (Kantipudi et al., 2020). The obtained IC 50 s for L-leucine ( Figure 2) were comparable to previously published values using oocytes (Kanai et al., 1998;Friesema et al., 2001;Yanagida et al., 2001), and P. pastoris cells at OD 600 10 (4F2hc-LAT1, 4F2hc-LAT2, and LAT1) and OD 600 3 (LAT2), and uptake times of 10 min (4F2hc-LAT1, 4F2hc-LAT2, and LAT1) and 2 min (LAT2) (Kantipudi et al., 2020). The modulatory effect of the heavy chain 4F2 on the L-leucine affinity of light subunits, this being most pronounced for 4F2hc-LAT2 and LAT2 ( Figures 2C,D), was also observed in line with our previous report (Kantipudi et al., 2020). Newly, we tested with the P. pastoris cell-based assay, described transport inhibitors, i.e., the system L and 4F2hc-LAT1 uptake inhibitors BCH  and JPH203 (Oda et al., 2010), respectively. The IC 50 s of BCH for the two HATs were comparable with values from the literature ( Table 1) and almost identical to the IC 50 s from corresponding LATs in absence of 4F2hc ( Figure 3). The latter indicated no contribution of the ancillary protein on BCH binding to light subunits. JPH203 was described as a competitive, potent and highly specific 4F2hc-LAT1 inhibitor Oda et al., 2010). The authors of these two studies determined IC 50 s of 0.14 µM (Oda et al., 2010), and 0.19 and 0.2 µM  for human LAT1 HATs expressed in S2 cells derived from mouse renal proximal tubules (Morimoto et al., 2008) ( Table 1). The IC 50 of 0.197 µM for 4F2hc-LAT1 obtained with the here presented assay ( Figure 4A) is similar to the values (0.14, 0.19, and 0.2 µM) determined for the same HAT using S2 cells (Table 1). Furthermore, most kinetic parameters for JPH203 and 4F2hc-LAT1 determined using other cell types than S2 cells were also in good agreement with the IC 50 value from P. pastoris cells (Table 1). However, a marked difference is found between the IC 50 of JPH203 determined for human 4F2hc-LAT2 using the P. pastoris (1.5 µM; Figure 4C) and the S2 cell-based assays (>10 µM (Oda et al., 2010) and almost no inhibition at 10 µM ). Furthermore, Oda et al. indicated an IC 50 (S2-LAT2)/IC 50 (S2-LAT1) ratio of >500, which would be indicative of a high specificity of JPH203 towards human 4F2hc-LAT1 and not 4F2hc-LAT2 (Oda et al., 2010). In strong contrast, the IC 50 (Pichia-4F2hc-LAT2)/IC 50 (Pichia-4F2hc-LAT1) ratio was only about 7.6, indicating a significantly more moderate specificity of JPH203 towards 4F2hc-LAT1. This finding was further corroborated independently using Pichia cells expressing the light subunits alone, which yielded a comparable IC 50 (Pichia-LAT2)/IC 50 (Pichia-LAT1) ratio of about 4.9. Future studies on the specificity of JPH203 towards 4F2hc-LAT1 and -LAT2 will clarify this ambiguity. For further validation of our yeast cell-based transport assay, we tested amino acid derivatives, specifically the thyroid hormones T3 and T4, which are substrates of 4F2hc-LAT1 and -LAT2 (Friesema et al., 2001;Zevenbergen et al., 2015). For 4F2hc-LAT1, T3 reflected a higher potency in transport inhibition than T4 ( Figures 5A, 6A) in line with previous studies ( not have an important influence on the binding of the two hormones to the light subunits ( Figures 5, 6).

CONCLUSION
We have provided a robust yeast cell-based transport assay for the functional characterization of human 4F2hc-LAT1 and -LAT2, and LAT1 and LAT2 substrates and inhibitors. P. pastoris cells expressing HATs are valuable for determining kinetic parameters of ligands, e.g., potency and selectivity of transport inhibitors towards 4F2hc-LAT1 and -LAT2. Such kinetic information will help evaluating the structure-activity relationship of new ligands. Moreover, assays performed with yeast cells expressing light subunits provide kinetic parameters that, when compared with those of corresponding HATs, are of interest towards identification of possible effects of the ancillary protein 4F2hc on ligand affinity and specificity. First kinetic parameters on T3 and T4 were successfully provided for human 4F2hc-LAT2 and LAT2 using the here presented yeast cell-based transport assay. Finally, the established approach and assay could also be used for compound screening and applied to other HATs and LATs of interest.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
DF conceived and designed the study. SK performed the experiments, collected and analyzed the data. SK and DF wrote the manuscript. DF obtained the funding. SK and DF read, and approved the submitted version.  Oda et al. (2010)