The Heavy Chain 4F2hc Modulates the Substrate Affinity and Specificity of the Light Chains LAT1 and LAT2

The human L-type amino acid transporters LAT1 and LAT2 mediate the transport of amino acids and amino acid derivatives across plasma membranes in a sodium-independent, obligatory antiport mode. In mammalian cells, LAT1 and LAT2 associate with the type-II membrane N-glycoprotein 4F2hc to form heteromeric amino acid transporters (HATs). The glycosylated ancillary protein 4F2hc is known to be important for successful trafficking of the unglycosylated transporters to the plasma membrane. The heavy (i.e., 4F2hc) and light (i.e., LAT1 and LAT2) chains belong to the solute carrier (SLC) families SLC3 and SLC7, and are covalently linked by a conserved disulfide bridge. Overexpression, absence, or malfunction of certain HATs is associated with human diseases and HATs are therefore considered therapeutic targets. Here, we present a comparative, functional characterization of the HATs 4F2hc-LAT1 and 4F2hc-LAT2, and their light chains LAT1 and LAT2. For this purpose, the HATs and the light chains were expressed in the methylotrophic yeast Pichia pastoris and a radiolabel transport assay was established. Importantly and in contrast to mammalian cells, P. pastoris has proven useful as eukaryotic expression system to successfully express human LAT1 and LAT2 in the plasma membrane without the requirement of co-expressed trafficking chaperone 4F2hc. Our results show a novel function of the heavy chain 4F2hc that impacts transport by modulating the substrate affinity and specificity of corresponding LATs. In addition, the presented data confirm that the light chains LAT1 and LAT2 constitute the substrate-transporting subunits of the HATs, and that light chains are also functional in the absence of the ancillary protein 4F2hc.


Introduction
Amino acids are essential biomolecules, which are involved in cellular processes ranging from energy production to protein synthesis and signaling. Membrane proteins belonging to different solute carrier (SLC) families mediate the transport of amino acids and their derivatives across biological membranes [1]. Among these transporter families, the SLC7 family consists of 15 genes [2], which encode amino acid transporters that belong to the amino acid, polyamine and organocation (APC) superfamily of transporters (transport classification (TC) system No. 2.A.3; http://www.tcdb.org) [3]. The SLC7 family consists of two subgroups: the cationic amino acid transporters (CATs; SLC7A1-A4 and SLC7A14) and the glycoprotein-associated L-type amino acid transporters (LATs; SLC7A5-A11, Slc7a12, SLC7A13, and Slc7a15) [2]. CATs are N-glycosylated, while LATs are not. In contrast to and characterize the transport function of the human light chains LAT1 and LAT2 in the absence of co-expressed heavy chain/ancillary protein 4F2hc.

Results and Discussion
The human HATs: 4F2hc-LAT1 and 4F2hc-LAT2, and LATs: LAT1 and LAT2 were expressed in the methylotrophic yeast Pichia pastoris. Western blot analysis indicated expression of the corresponding HATs and LATs ( Figure S1). Transport activities were determined by measuring the uptake of [ 3 H]L-leucine into P. pastoris cells expressing the corresponding HAT or LAT. Time-course experiments showed clear HAT-and LAT-specific transport activities, which were much higher than the [ 3 H]L-leucine uptake into untransformed host cells (Figure 1). In all cases, a saturation of the transport process was observed. Time-course experiments show clear LAT-specific transport, which is much higher than the uptake into untransformed host cells. Uptake assay times of 10 min for 4F2hc-LAT1, LAT1, and 4F2hc-LAT2, and of 2 min for LAT2 were chosen for all subsequently presented experiments (time points indicated by *). Data points are represented as mean with SD from a representative triplicate experiment. If not visible, error bars are smaller than symbols.
We determined the half maximal inhibitory concentrations (IC 50 s) of L-leucine by homologues competition for all constructs using the obtained time points, i.e., 10 min (4F2hc-LAT1, LAT1, 4F2hc-LAT2) and 2 min (LAT2) ( Figure S2). These IC 50 s ( Figure S2) gave first impressions of the affinities of the HATs and LATs for L-leucine. HAT-and LAT-mediated transport of [ 3 H]L-leucine was saturable and followed Michaelis-Menten kinetics with K m values of 25 µM (4F2hc-LAT1), 11 µM (LAT1), 249 µM (4F2hc-LAT2), and 42 µM (LAT2) ( Figure 2). As reflected from this data, the impact of the heavy chain 4F2hc on the affinity of the light chain for L-leucine was most pronounced for LAT2, where the K m value increases almost six-fold upon association with 4F2hc. The measured K m s of L-leucine for the two HATs were comparable with values from previous publications, i.e., 18 µM [34] and 20 µM [35] for 4F2hc-LAT1, and 220 µM for 4F2hc-LAT2 [13]. Recently, a study using proteoliposomes showed that human LAT1 has no transport activity in the absence of 4F2hc and concluded that the ancillary protein is essential for the transport activity of the complex [23]. In contrast, we demonstrate that LAT1 and LAT2 are able to transport [ 3 H]L-leucine in the absence of the heavy chain 4F2hc. This finding is further supported by studies, in which LAT1 was reconstituted into liposomes, and transport activity of this light chain was shown in the absence of the heavy chain 4F2hc [36,37].
Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 4 of 11 that LAT1 and LAT2 are able to transport [ 3 H]L-leucine in the absence of the heavy chain 4F2hc. This finding is further supported by studies, in which LAT1 was reconstituted into liposomes, and transport activity of this light chain was shown in the absence of the heavy chain 4F2hc [36,37]. The substrate specificities of 4F2hc-LAT1 and 4F2hc-LAT2, and LAT1 and LAT2 were determined by measuring the ability of proteinogenic amino acids and D-leucine at concentrations of about ten times Km to compete with [ 3 H]L-leucine uptake ( Figure 3). 4F2hc-LAT1 showed a relatively broad substrate specificity with highest for L-leucine and L-histidine, in line with previous reports [34][35][36]. The specificity for D-leucine was significantly lower compared to its L-isomer indicating stereospecificity of 4F2hc-LAT1 ( Figure 3). Having established an [ 3 H]L-leucine-based uptake assay and for comparison with kinetic values in the literature, we determined the half maximal inhibition concentration (IC50) for L-histidine to 23 µM using P. pastoris cells expressing human 4F2hc-LAT1 ( Figure 4). This IC50 is comparable to previously reported Km values for L-histidine, i.e., 12.7 µM [35] and 24.6 µM [36]. Competition data clearly showed that this HAT, 4F2hc-LAT1, has no considerable affinity (i.e., more than 50% residual [ 3 H]L-leucine uptake; Figure 3A) for glycine, L-proline, L-serine and the negatively charged amino acids L-aspartate and L-glutamate. The specificity for most tested amino acids decreased when [ 3 H]L-leucine transport competition was studied for the light chain LAT1 alone ( Figure 3B). The stereospecificity with respect to L-leucine increased significantly compared to 4F2hc-LAT1, i.e., almost no observed inhibition of [ 3 H]L-leucine transport by D-leucine ( Figure 3B). L-alanine, which was competing with HAT-mediated L-leucine transport, was not recognized by LAT1 at all. In summary, LAT1 is highly specific for L-leucine and has a modest affinity The substrate specificities of 4F2hc-LAT1 and 4F2hc-LAT2, and LAT1 and LAT2 were determined by measuring the ability of proteinogenic amino acids and D-leucine at concentrations of about ten times K m to compete with [ 3 H]L-leucine uptake ( Figure 3). 4F2hc-LAT1 showed a relatively broad substrate specificity with highest for L-leucine and L-histidine, in line with previous reports [34][35][36]. The specificity for D-leucine was significantly lower compared to its L-isomer indicating stereospecificity of 4F2hc-LAT1 ( Figure 3). Having established an [ 3 H]L-leucine-based uptake assay and for comparison with kinetic values in the literature, we determined the half maximal inhibition concentration (IC 50 ) for L-histidine to 23 µM using P. pastoris cells expressing human 4F2hc-LAT1 ( Figure 4). This IC 50 is comparable to previously reported K m values for L-histidine, i.e., 12.7 µM [35] and 24.6 µM [36]. Competition data clearly showed that this HAT, 4F2hc-LAT1, has no considerable affinity (i.e., more than 50% residual [ 3 H]L-leucine uptake; Figure 3A) for glycine, L-proline, L-serine and the negatively charged amino acids L-aspartate and L-glutamate. The specificity for most tested amino acids decreased when [ 3 H]L-leucine transport competition was studied for the light chain LAT1 alone ( Figure 3B). The stereospecificity with respect to L-leucine increased significantly compared to 4F2hc-LAT1, i.e., almost no observed inhibition of [ 3 H]L-leucine transport by D-leucine ( Figure 3B). L-alanine, which was competing with HAT-mediated L-leucine transport, was not recognized by LAT1 at all. In summary, LAT1 is highly specific for L-leucine and has a modest affinity for other amino acids ( Figure 3B). Interestingly, the effect of competition for L-histidine in LAT1 is significantly decreased in the absence of the heavy chain 4F2hc ( Figure 3A,B), which suggests a markedly lower affinity for this amino acid compared to the heterodimer. As observed for 4F2hc-LAT1, the HAT 4F2hc-LAT2 has also a relatively broad substrate specificity ( Figure 3C). Most amino acids with the exception of L-proline, L-serine, L-glutamate, and the positively charged amino acids L-lysine, L-arginine, and L-histidine reduced the residual [ 3 H]L-leucine uptake below 45%.
significantly decreased in the absence of the heavy chain 4F2hc ( Figure 3A,B), which suggests a markedly lower affinity for this amino acid compared to the heterodimer. As observed for 4F2hc-LAT1, the HAT 4F2hc-LAT2 has also a relatively broad substrate specificity ( Figure 3C). Most amino acids with the exception of L-proline, L-serine, L-glutamate, and the positively charged amino acids L-lysine, L-arginine, and L-histidine reduced the residual [ 3 H]L-leucine uptake below 45%. These concentrations correspond to about ten times the determined L-leucine Km values of the corresponding transporters ( Figure 2). Residual uptake in the presence of competitor was normalized with respect to control samples without competitor (Ctrl). The amino acids are abbreviated using their three-letter-code. Means with SD from normalized data of three independent experiments, each at least in triplicate are shown. If not visible, error bars are smaller than symbols.
In contrast to 4F2hc-LAT1, specificities of 4F2hc-LAT2 for the branched chain amino acids (BCAA) L-leucine, L-isoleucine, and L-valine were comparable ( Figure 3C). Again, in the absence of the heavy chain 4F2hc the competition pattern changed significantly ( Figure 3D). As observed for LAT1, L-leucine showed the strongest reduction of LAT2-mediated [ 3 H]L-leucine uptake in contrast to its D-isomer, reflecting stereospecificity of the transporter ( Figure 3D). Only L-isoleucine and Lvaline also reduced the residual [ 3 H]L-leucine uptake below 35%, which shows that LAT2 has a preferred affinity for BCAA. The strongest increase in competition associated with co-expression of the heavy chain 4F2hc is observed for L-tyrosine, L-cysteine, L-threonine, L-asparagine, L-glutamine, and L-aspartate. In general, association of the light chain LAT2 with the heavy chain 4F2hc expands the substrate specificity of the HAT compared to the BCAA-specific light chain LAT2 alone.  Figure 2). Residual uptake in the presence of competitor was normalized with respect to control samples without competitor (Ctrl). The amino acids are abbreviated using their three-letter-code. Means with SD from normalized data of three independent experiments, each at least in triplicate are shown. If not visible, error bars are smaller than symbols.
In contrast to 4F2hc-LAT1, specificities of 4F2hc-LAT2 for the branched chain amino acids (BCAA) L-leucine, L-isoleucine, and L-valine were comparable ( Figure 3C). Again, in the absence of the heavy chain 4F2hc the competition pattern changed significantly ( Figure 3D). As observed for LAT1, L-leucine showed the strongest reduction of LAT2-mediated [ 3 H]L-leucine uptake in contrast to its D-isomer, reflecting stereospecificity of the transporter ( Figure 3D). Only L-isoleucine and L-valine also reduced the residual [ 3 H]L-leucine uptake below 35%, which shows that LAT2 has a preferred affinity for BCAA. The strongest increase in competition associated with co-expression of the heavy chain 4F2hc is observed for L-tyrosine, L-cysteine, L-threonine, L-asparagine, L-glutamine, and L-aspartate. In general, association of the light chain LAT2 with the heavy chain 4F2hc expands the substrate specificity of the HAT compared to the BCAA-specific light chain LAT2 alone. Interestingly, the stereospecificity for leucine of the light chains LAT1 and LAT2 is more pronounced in the absence of the heavy chain 4F2hc (Figure 4). Interestingly, the stereospecificity for leucine of the light chains LAT1 and LAT2 is more pronounced in the absence of the heavy chain 4F2hc (Figure 4).

Conclusions
We have shown and confirmed that the light chains LAT1 and LAT2 are the substratetransporting subunits of the corresponding HATs 4F2hc-LAT1 and 4F2hc-LAT2, and that 4F2hc is not essential for the transport activity of the corresponding LATs. LAT1 and LAT2 have relatively high specificities for L-leucine, and modest specificities for other amino acids. That the ancillary protein 4F2hc is responsible for chaperoning the trafficking of light chains (i.e., LATs) to the plasma membrane of mammalian cells is a well-accepted and described concept. Our comparative transporter study revealed a novel function of 4F2hc, i.e., upon association of this ancillary protein with LAT1 and LAT2, the substrate affinity and specificity of these light subunits is modulated by significantly broadening their substrate specificities. The methylotrophic yeast P. pastoris has proven useful for the here-presented comparative transporter study. In contrast to mammalian cells, the substrate-transporting subunits LAT1 and LAT2 could be successfully expressed functional in Pichia in the absence of their ancillary N-glycoprotein 4F2hc. Therefore, this eukaryotic expression system, which also allows post-translational modifications such as glycosylation and disulfide bridge formation, opens the possibility to investigate the influence of pharmacologically relevant compounds on the function of heteromeric complexes and their transporters alone.

Cloning of Human 4F2hc-LAT1, 4F2hc-LAT2, LAT1, and LAT2
The making of the 4F2hc-LAT2 and LAT2 expression constructs, i.e., pPICZB-4F2hc-LAT2 and pPICZB-LAT2 in the pPICZB vector (Thermo Fisher Scientific, Waltham, MA, USA), and of the Pichia pastoris clones expressing human 4F2hc-LAT2 or LAT2 was described in detail previously [33]. These same two Pichia clones expressing 4F2hc-LAT2 and LAT2 were used in the here-presented study. We . For data analysis, the signal of the untransformed P. pastoris KM71H was subtracted from 4F2hc-LAT1 transporter. Cpm values of each experiment were normalized with respect to the determined upper plateau value, i.e., the fitted upper plateau value corresponds to 100%. A sigmoidal model curve (orange) was fitted to the net transport signals to obtain the IC 50 . Data points represent means with SD from normalized data of three independent experiments, each at least in triplicate are shown. If not visible, error bars are smaller than symbols.

Conclusions
We have shown and confirmed that the light chains LAT1 and LAT2 are the substrate-transporting subunits of the corresponding HATs 4F2hc-LAT1 and 4F2hc-LAT2, and that 4F2hc is not essential for the transport activity of the corresponding LATs. LAT1 and LAT2 have relatively high specificities for L-leucine, and modest specificities for other amino acids. That the ancillary protein 4F2hc is responsible for chaperoning the trafficking of light chains (i.e., LATs) to the plasma membrane of mammalian cells is a well-accepted and described concept. Our comparative transporter study revealed a novel function of 4F2hc, i.e., upon association of this ancillary protein with LAT1 and LAT2, the substrate affinity and specificity of these light subunits is modulated by significantly broadening their substrate specificities. The methylotrophic yeast P. pastoris has proven useful for the here-presented comparative transporter study. In contrast to mammalian cells, the substrate-transporting subunits LAT1 and LAT2 could be successfully expressed functional in Pichia in the absence of their ancillary N-glycoprotein 4F2hc. Therefore, this eukaryotic expression system, which also allows post-translational modifications such as glycosylation and disulfide bridge formation, opens the possibility to investigate the influence of pharmacologically relevant compounds on the function of heteromeric complexes and their transporters alone.

Cloning of Human 4F2hc-LAT1, 4F2hc-LAT2, LAT1, and LAT2
The making of the 4F2hc-LAT2 and LAT2 expression constructs, i.e., pPICZB-4F2hc-LAT2 and pPICZB-LAT2 in the pPICZB vector (Thermo Fisher Scientific, Waltham, MA, USA), and of the Pichia pastoris clones expressing human 4F2hc-LAT2 or LAT2 was described in detail previously [33]. These same two Pichia clones expressing 4F2hc-LAT2 and LAT2 were used in the here-presented study. We generated a pPICZB-based expression construct for human 4F2hc-LAT1 (pPICZB-4F2hc-LAT1) as described for 4F2hc-LAT2 [33], but using the cDNA of the light chain LAT1 (UniProt ID code Q01650) instead of LAT2. For the LAT1 construct, the human gene (UniProt ID code Q01650) was synthesized codon-optimized for expression in the methylotrophic yeast Pichia pastoris with 5 -HindIII and 3 -XhoI restriction sites (GenScript). In contrast to this codon-optimized LAT1 gene, the previously mentioned constructs were generated from cDNA. The LAT1 gene was ligated into the vector pZUDFPICZ-10His3C using 5 -HindIII and 3 -XhoI restriction sites yielding the construct pZUDFPICZ-10His3C-LAT1. pZUDFPICZ-10His3C is a modified version of the pPICZB plasmid (Thermo Fisher Scientific), which has been modified as follows. First, the single HindIII restriction site of pPICZB was removed by site-directed mutagenesis using the primer

[ 3 H]L-Leucine Radioligand Transport Assay
For transport experiments, 3 mL P. pastoris cells at OD 600 40 expressing the corresponding transporter were thawed, diluted 1:50 in transport buffer, and pelleted by centrifugation (3000× g, 15 min, room temperature). Subsequently, the pellet was washed by resuspending in 50 mL transport buffer and pelleted again. The washing step was repeated twice. Finally, the cell pellet was resuspended in 4 mL of transport buffer and incubated for 20 min at 30 • C under agitation (300 rpm, Multitron, Infors HT). The density of the yeast suspension was adjusted with transport buffer to OD 600 25 (LAT1, 4F2hc-LAT2 or 4F2hc-LAT1) and 7.5 (LAT2). All transport experiments were performed in a reaction volume of 100 µL. Final OD 600 values in uptake experiments were 10 for 4F2hc-LAT1, 4F2hc-LAT2, and LAT1, and 3 for LAT2. All transport reactions were done in 2 mL reaction tubes (Eppendorf) at 25 • C under agitation (1000 rpm, Thermomixer compact, Eppendorf, Hamburg, Germany). Transport was terminated after 10 min for 4F2hc-LAT1, 4F2hc-LAT2, or LAT, and 2 min for LAT2 by addition of 600 µL of pre-chilled transport buffer. Cells were rapidly separated from the buffer by transferring the stopped reactions on a 96-well 0.66 mm glass fiber filter plate (Corning FiltrEX, Corning, NY, USA) and vacuum filtration. Each well was washed with 2 mL of ice-cold transport buffer to remove free radioligand. The plate was then dried overnight at 37 • C and the backside was sealed with back seal (PerkinElmer, Waltham, MA, USA). The trapped radioligand was released by addition of 200 µL scintillation cocktail (MicroScint 40, PerkinElmer) to each well and the plate topside was sealed with Topseal TM -A Plus (PerkinElmer), followed by incubation for 30 min at 25 • C and 1000 rpm (Thermomixer compact, Eppendorf, Hamburg, Germany). Counts were measured in each well for 2 min with a scintillation counter (TopCount NXT, PerkinElmer).

Statistics
Experimental data points were performed at least in triplicate. For data analysis, the signal of the untransformed P. pastoris cells was subtracted from the transporter signal. Michaelis-Menten saturation curves were fitted into data points of independent experiments. Data points were then individually normalized using the corresponding V max values (i.e., the fitted upper plateau value corresponds to 100%). Data points from corresponding concentrations were averaged and SD obtained. Finally, Michaelis-Menten saturation curves were fitted to the averaged data yielding K m values. To determine the half maximal inhibitory concentration (IC 50 ) values of heterologous (i.e., L-histidine) L-leucine transport competition, a sigmoidal model curve was fitted to the net transport signals of independent experiments. Every experimental data point was individually normalized using the corresponding upper plateau values (i.e., the fitted upper plateau value corresponds to 100%). Data points from corresponding concentrations were averaged and a sigmoidal model curve was fitted to the data in order to obtain the IC 50 value. Prism6 (GraphPad Software) was used for data analysis.

Conflicts of Interest:
The authors declare no conflict of interest.

LAT
L-type amino acid transporter HAT Heteromeric amino acid transporters IC 50 Half maximal inhibitory concentration K m Michaelis-Menten constant SLC Solute carrier