and pharmacogenetic studies (see below).
3.3.3 Inhibitors
MATE transport can be inhibited by a number of chemicals. Structural examples are shown in Fig. 3.2. The antiparasitic drug, pyrimethamine, is considered one of the most potent and specific inhibitors. Inhibition of mMate1/hMATE1 is achieved at nanomolar concentrations of pyrimethamine, whereas micromolar concentrations are required to block mOct1/hOCT2 and mOct2/hOCT2 [52]. Cis‐inhibition of TEA transport by hMATE1 is observed with cimetidine, quinidine, procainamide, and verapamil and to a lesser degree nicotine, serotonin, and choline [5, 6]. Likewise, a number of hormones can inhibit TEA transport in hMATE1‐ and mMate1‐expressing cells including corticosterone, testosterone, progesterone, and androstenedione [15]. Notably, no effects on hMATE1 or hMATE2‐K activity by organic anions such as p‐aminohippurate, uric acid, and beta‐lactam antibiotics were observed [5, 6].
3.4 STRUCTURE
3.4.1 Modeling of Ligand Interactions
A variety of experimental approaches, including Bayesian machine learning, binary classification modeling, molecular docking and pharmacophore modeling, have been employed to define physicochemical and structural properties of MATE ligands. Inhibitors of MATE1 have been defined by their cationic charge, high molecular weight, and lipophilicity [53]. Using a combination of in vitro and computational approaches, a pharmacophore for hMATE1 and hMATE2‐K inhibitors was defined that favors shared features of high‐affinity inhibitors (such as pyrimethamine and quinidine) and avoids structures observed in low‐affinity inhibitors (such as histamine, caffeine, and chloramphenicol) [54]. Refinement of the model resulted in a pharmacophore for MATE inhibition that included two hydrophobes, a hydrogen‐bond acceptor, and an ionizable feature [54]. A subsequent combinatorial pharmacophore model for hMATE1 inhibition predicted multiple sites for ligand interaction. These included two regions for competitive inhibition by smaller molecules, as well as accommodation of large inhibitors within the central cavity of the transporter where they act noncompetitively to prevent the conformational changes needed for organic cation transport [55].
Transporter‐based drug–drug interactions can often be dependent upon the identity of the victim substrate being evaluated, as has been demonstrated for OCT [56, 57]. Interestingly, while some early studies with a small number of substrates suggested this may be the case for MATE transporters [58], a more robust subsequent analysis has suggested otherwise. Using four structurally distinct organic cation substrates (metformin, cimetidine, MPP, and N,N,N‐trimethyl‐2‐[methyl(7‐nitrobenzo[c][1,2,5]oxadiazol‐4‐yl)amino]ethanaminium iodide) and over 400 drugs, it was determined that the identity of the substrate has little influence on ligand interaction and inhibition of hMATE1‐mediated uptake of substrates in overexpressing cells [59]. Similar results were observed in a separate study using metformin and atenolol as substrates [60]. Notably, there were no significant differences in the apparent Michaelis constant of the transported substrate and the 50% inhibitor concentration calculated from the inhibition of transport across the four substrates. These data would suggest that there is a shared binding site for interaction of substrates and inhibitors on the external surface of hMATE1.
3.4.2 Secondary Structure
With the original cloning of hMATE1, it was predicted that the secondary structure contained 12 transmembrane helices with internal‐facing NH2‐ and COOH‐termini [5]. Subsequent predictions pointed to a 13th transmembrane domain for hMATE and rbMate1 that would result in extracellular localization of the carboxy‐terminus. This additional domain was not anticipated for mMate1 protein and instead the additional amino acids were thought to form a long cytoplasmic COOH terminus [20]. These predictions were supported by experimental evidence that showed extracellular accessibility of the carboxy‐terminus of rbMate1, but not mMate1 [20]. Epitope tagging and cysteine accessibility scanning affirmed that rbMate1 protein includes 13 transmembrane domains with an intracellular NH2‐terminus and extracellular COOH‐terminus (Fig. 3.3) [61]. Subsequently, a functionally active variant of mMate1, mMate1b, was identified and revealed to contain a long hydrophobic tail, similar to other MATE transporters. This region of mMate1b encodes a 13th transmembrane domain that results in an extracellular carboxy terminus [62]. The exact purpose of the 13th transmembrane domain has remained unclear as the first 12 domains form the “functional core” of the rbMate1, hMATE1, and mMate1b orthologs [63]. Truncated mutant forms of MATE1/Mate1 proteins retain functional activity, ligand binding, and multi‐selectivity of substrates, leading researchers to posit other potential roles for this domain in substrate translocation, stabilization in the membrane bilayer, oligomerization, or protein–protein interactions [61, 62].
Using the X‐ray structure of the NorM transporter (3.65 Å) [64], a homology model for hMATE1 was developed [63]. This model positions hMATE1 in two bundles of six transmembrane helices (N lobe: transmembrane domains 1–6, C lobe: transmembrane domains 7–12) with an internal cavity of ~4,000 Å that is open to the extracellular space [63]. A relatively short cytoplasmic loop between domains 6 and 7 is positioned to connect the two halves and is consistent with hydropathy plots for N or M. Evaluation of the crystal structure of a MATE transporter from Arabidopsis thaliana (2.6 Å), known as AtDTX14, has also provided insights into hMATE1 structure and function [65]. The amino acid sequence identity between hMATE1 and AtDTX14 is 32%. A key hydrogen bonding network in the C‐lobe demonstrated in AtDTX14 is conserved in hMATE1 and is considered to be the substrate‐binding site [65]. While key insights into the likely structure of hMATE1 (and orthologs) have been made, there is little insight into similarities and differences for MATE2/2‐K, as well as a need for a MATE structure from crystallography or cryo‐electron microscopy.
FIGURE 3.3 Predicted structure of the human MATE1 transporter. The membrane topology of the human MATE1 transporter has been adapted from [128].
3.4.3 Structural Features
Treatment of cells expressing rMate1 with p‐chloromercuribenzene sulfonate, an organic mercurial chemical, significantly reduced uptake of TEA. This inhibition of transport could be rescued by dithiothreitol suggesting that reduced sulfhydryl groups are important for the activity of rMate1 [17]. Subsequent site‐directed mutagenesis identified key residues [66]. One histidine (rMate1: His‐385, hMATE1: His‐386, hMATE2‐K: His‐382) and two cysteine residues (rMate1: Cys‐62 and Cys‐126, hMATE1: Cys‐63 and Cys‐127, hMATE2‐K: Cys‐59 and Cys‐123) were essential for transport activity. The impaired function of mutants in these residues was not due to improper trafficking or reduced expression as all localized to the plasma membrane to similar degrees as wild‐type proteins [66]. Interestingly, unlabeled TEA was able to protect against the transport inhibition achieved by the sulfhydryl reagent PCMBS suggesting that rMate1 substrates interact directly with sulfhydryl‐containing Cys‐62 and Cys‐126 [66]. By comparison, unlabeled TEA had no effect on the ability of the histidine residue modifier DEPC to block rMate1 activity in vitro.
There has also been interest in elucidating the role of negatively charged glutamates in the recognition of substrates that are largely cationic in nature. Mutation of the glutamate residue at position 273 to a glycine reduced hMATE1 function without altering its insertion into the plasma membrane [5]. Notably, this glutamate residue is conserved across rMate1, mMate1, and hMATE2‐K. Substitution of glutamate at position 273 with alanine resulted in the absence of hMATE1 protein, whereas amino acid changes at the remaining glutamates differentially altered activity
