| dc.description.abstract | Mutations and single-nucleotide polymorphisms (SNPs) (mutations that occur in more than one percent of a population) can occur throughout the human genome. The study of polymorphisms, single-nucleotide mutations, and predictive pharmacogenomics has become a focus for both researchers and physicians in recent years. However, knowing about the existence of such variants is only part of the challenge. Mutations in DNA can impact the function of the gene, protein, enzyme, etc. that is encoded by that region of DNA to varying degrees. For decades, computer algorithms have attempted to predict outcomes of amino acid substitutions based on assumptions about important amino acids that were derived from decades of experimentation: 1) most amino acid substitutions result in damaged protein function or structure 2) few substitutions allowed for normal protein function and 3) the same amino acid substitutions will behave similarly in homologous proteins. Thus, there were three types of functional outcomes that positions were generally expected: neutral, deleterious, or catastrophic. A mutation that causes little to no effect is considered a neutral substitution. In contrast, if a mutation causes decreased protein function that results in a clinical phenotype it is considered to be a deleterious substitution. Further, if a mutation results in a complete lack of function and even degradation of the protein, it is classified as a catastrophic substitution.However, one major flaw of the current algorithms is that they are largely based on data and rules obtained from mutating conserved amino acid positions. Very few studies systematically studied whether these rules apply to nonconserved positions. This leaves mutations occurring in nonconserved locations enigmatic. Recent studies in the LacI/GalR transcriptional regulator family as well as in globular proteins have shown that multiple substitutions at some nonconserved positions result in intermediate outcomes in addition to neutral and catastrophic. Further, when taken together, multiple substitutions at these amino acid positions result in a continuum of outcomes akin to a dimmer switch, thus they are referred to as rheostat positions. In addition, many of these intermediate substitutions do not always damage the protein, nor do the same amino acid substitution in different homologs always produce the same functional outcomes. Because these substitutions do not follow these canonical assumptions, they are currently unpredictable. Furthermore, the existence and behavior of rheostat positions has yet to be elucidated in other types of proteins, including integral transmembrane proteins. These proteins are particularly important for the disposition and metabolism of therapeutic drugs. Thus, alterations in the function of these proteins are particularly important for the prediction of drug response. Therefore, our aim is to predict and characterize rheostat positions in an integral transmembrane protein to improve computational predictions of rheostat-like functional outcomes. Our original hypothesis was that these positions would be more likely to occur at nonconserved amino acid locations. We decided to use the Na+/taurocholate cotransporting polypeptide (NTCP) as a model integral transmembrane protein. Human NTCP is a transporter highly expressed at the basolateral membrane of hepatocytes and is involved in the enterohepatic circulation of bile acids. In addition, NTCP can transport sulfated hormones and certain statins. Given the diversity of these substrates, we decided to utilize three model substrates to study the function of NTCP: taurocholate (a bile acid), estrone-3-sulfate (a sulfated hormone), and rosuvastatin (a statin). We then selected potential rheostat positions within human NTCP using three methods. The first was to select a known polymorphic location shown to have varying levels of functional impact depending on the substrate, position 267. Next, we selected a position, 271, that was hypothesized to be a rheostat based on protein structure modeling and energy calculations. Finally, we selected two additional positions based on evolutionary differences calculated using a multiple sequence alignment consisting of over 1500 sequences from both pro- and eukaryotes. Using this final method along with other criteria, we selected a highly conserved and a highly nonconserved amino acid positions, 102 and 146, respectively. Our studies concluded that position 102 was the strongest rheostat for both expression and function, while positions 267 and 271 were neutral for expression but rheostatic for all substrates, except for estrone-3-sulfate transport by N271 variants. Further, position 146 was neutral for most variants for both expression and function. In addition, characterization of select S267 and N271 variants revealed impacts on Michaelis-Menten kinetics to varying degrees. For example, S267 variants demonstrated more alterations in NTCP’s capacity while at position 271 most alterations affected the transporter's velocity or turnover. We next related the function and expression of these S267 and N271 variants to simulated structure model energy predictions to assess if they could aid in the prediction of the observed functional outcomes. While there were some correlations between the surface expression and calculated energy scores, more studies are needed to determine if improved computational models could be used to predict complex rheostatic outcomes. | |
