Glutamate-Sensing Genes Are Conserved among Populations Compared to Glutamate Metabolism Genes

Introduction: Glutamate is a representative taste molecule with an umami flavor and is a major nutrient found abundantly in nature. Furthermore, it plays a significant role in the human body as a key metabolic intermediate and neurotransmitter. Therefore, the divergence of glutamate functions among populations during their evolution is of particular interest with a hypothesis that the genetic variation can lead to understanding divergence in taste perception. To elucidate variation in glutamate applications and to deepen our understanding of taste perception, we examined the nucleotide diversity of genes associated with glutamate sensing and metabolism among human populations. Methods: We first established 67 genes related to glutamate sensing and metabolism based on the database and literature survey. Then, for those genes, we used a population genomics approach based on ten populations over 76,156 human genomes in the gnomAD database. Results: Statistical tests of means and medians of the minor allele frequencies did not show any significant difference among populations. However, we observed substantial differences between two functional groups, glutamate sensing and glutamate metabolism, in populations of Latino/admixed American, Ashkenazi Jewish, and Others. Interestingly, we could find significant differences between the African population and the East Asian population at the single nucleotide polymorphism level of glutamate metabolism genes, but no clear differences were noted in glutamate-sensing genes. These suggest that glutamate-sensing genes are under the functional constraint compared to glutamate metabolism genes. Conclusion: Thus, glutamate-sensing genes and metabolism genes have a contrasting mode of the evolution, and glutamate-sensing genes are conservatively evolved, indicating its functional importance.


Introduction
Taste perception varies among populations, which may be associated with cultural and genetic backgrounds [1][2][3].Food diversity worldwide is well described and explained by cultural differences among human populations living in a region and local environmental conditions.However, there are limited studies on the genetic background of taste perception among populations [4].Different human populations have distinct genome structures, and the degree of diversity varies among genomic regions and genes based on their functional relevance.Ancestral or regional differences in taste perception are fundamental to understanding food preference and nutritional intake; therefore, genetic analyses should be conducted to reveal the genetic diversity underlying these differences.
Glutamate is an anion form of a glutamic acid which is one of the primary amino acids in the body.It is naturally abundant in various foods, such as meat, cheese, and tomato; free glutamate is responsible for their savory or umami taste [5].The flavor-enhancing characteristics of free glutamate can be traced back to ancient times.A fermented fish sauce known as "Garum" was rich in free glutamate and was commonly used as a seasoning by the Romans [6].Furthermore, several studies [7][8][9][10][11][12][13] have reported the use of umami substances in traditional flavorings throughout China, Korea, India, and Southeast Asia, revealing that most of the fermented products contain high amounts of free glutamate.This indicates that the umami taste of glutamate is essential for understanding the basis of food preference and nutritional intake.It is reported that heterodimeric G protein-coupled receptors, including the taste receptor type 1/member 1 (TAS1R1) and taste receptor type 1/member 3 (TAS1R3), and variants of metabotropic glutamate receptors, such as mGluR1 and mGluR4, are the umami taste receptors for glutamate [14].Variants of TAS1R1 among populations were reported, and they were associated with umami taste perception [14].Interestingly, the variants in TAS1R1 are enriched in the Japanese population [14] but not in the French population [15].This potentially corresponds to the differences in the taste perception of umami between Japanese and French populations, suggesting that genetic differences among populations can reveal the differences in taste perception.In addition, it is reported that taste perception links to the nutritional intervention [16], showing the importance of the comprehensive study on the glutamate metabolisms.
Humans ingest glutamate from two main sources, either through dietary protein or foods that contain significant amounts of free glutamate.Dietary protein is digested into free amino acids and small peptides, which are absorbed into intestinal mucosal cells, where peptides are hydrolyzed into free amino acids, including glutamate [17][18][19].When glutamate is ingested as a free amino acid, it is absorbed from the gut by an active transport system specific for dicarboxylic amino acids [20,21].The metabolism of glutamate involves oxidative deamination, transamination, decarboxylation, and amidation, all of which are well established in mammals [22].As previous experiments have shown the primary importance of the gut in metabolizing dietary glutamate before it reaches the liver [23,24], only excess glutamate leaving the gastrointestinal tract is metabolized in the liver [25] into urea or glutamine [26].In a series of processes as above, glutamate transporters as well as enzymes in the TCA cycle, urea cycle, and nitrogen metabolism play important roles in utilizing glutamate obtained from outside the body.In contrast, endogenous glutamate facilitates biochemical and metabolic processes that play major roles in nitrogen metabolism in mammals.Glutamate in the brain is an important excitatory neurotransmitter and precursor of the inhibitory neurotransmitter γ-aminobutyric acid.Low extracellular fluid concentrations, which are essential for optimal brain function, are maintained by neurons, astrocytes, and the blood-brain barrier (BBB).The BBB effectively excludes the passive brain influx of plasma glutamate [27].Therefore, glutamate acting as a neurotransmitter has entirely different origins compared with that ingested from the diet.These differences in endogenous functions through brain receptors and transporters indicate the variability in glutamate turnover in the body.
Under these situations, it is of particular interest that how the glutamate functions are established and diverged in human populations in terms of the taste perception.We hypothesize that the genetic characteristics can lead understand the differences in food preference and nutritional intake among human populations.Though humans have evolved distinct genetic traits worldwide, research on the nucleotide diversity of genes involved in glutamate metabolism at individual and population levels is limited.Therefore, in this study, we focused on genes associated with glutamate functions to clarify their characteristics in human populations and examined our hypothesis to understand the trends in food preference and nutritional intake among human populations.

Selection of Genes Associated with Glutamate Functions
A literature review was conducted to select genes associated with glutamate functions.We also referred to pathways in the Kyoto Encyclopedia of Genes and Genomes [28] to obtain Population Genomics of Glutamate-Sensing and Metabolism Genes  [29].We selected missense variants with global alternate allele frequency of >1% and <99% (i.e., minor allele frequency >1%) and marked as a missense variant using annotation by variant effective predictor to focus on nonrare variants with potential functional differences.For this analysis, we used the global minor allele to calculate the allele frequency of each population.We did not consider conserved genes in this analysis as they do not contain missense variants.We compared these allele frequencies among variants, populations, and functional categories.We used eleven population groups from gnomAD (i.e., ten pop-ulation groups and one global population group) and seven functional categories (i.e., the abovementioned six categories and one category containing all other genes in gnomAD as controls).Analysis of variance and median test were used to statistically compare the population groups and functional categories.Fisher's exact test was used to compare each variant among populations.Fisher's exact test was performed by R ver.3.4.2,and the median test and ANOVA were conducted using Python ver.3.7.0 with scipy ver.1.3.1 library [30].We drew violin plots with matplotlib ver.3.2.1 and seaborn ver.0.11 libraries.We performed multiple testing correction based on the Benjamini-Hochberg procedure.

Results
We selected genes associated with glutamate functions, as shown in Table 1.This list contains 67 genes belonging to 6 categories.We examined the variants and their allele frequencies as described in the materials and methods section.Based on our criteria, 145 missense variants were identified among these 67 genes (Table 1; online suppl.Table 1; for all online suppl.material, see https://doi.org/10.1159/000535181) and used at the following analyses.GnomAD is one of the comprehensive variant databases to date, containing 759,302,267 variants in 76,156 genomes among ten population groups (gnomAD v.3.1.1),so that we assume that our dataset covers variants enough among the global populations.Also, as described, our analysis was focusing only on the missense variants, thus we could not count the variations in well-conserved Goto/Masuzawa/Kohmura/Takumi/ Takeyama/Miyazaki/Gojobori/Mineta genes.For example, as shown in Table 1, AMPA-type glutamate receptors do not show any variants and thus we did not consider those genes in our analysis.
Figure 1 shows the distribution of allele frequencies among populations and functional categories.We calculated the mean and median allele frequencies of each population (Fig. 1a) and functional category (Fig. 1b) and compared them statistically.This analysis did not indicate any significant difference among populations or functional categories (q < 0.05) for either the mean or median allele frequency.We assume that the average or the median value lose the information at the variant level and need more focusing analysis.Then, we conducted statistical analysis of each variant.We first performed an overall assessment of statistical significance of each variant at the entire population by Fisher's exact test.Among 145 variants, 117 variants were significant.Then, we tested them under the pairwise Fisher's exact test between populations.The pairwise comparisons revealed that 63% (74/117) of the variants showed a significant difference in allele frequency in any pairwise combination of populations (Fisher's exact test; q < 0.05) (online suppl.Tables 1, 2).In particular, the average ratio of significant variants was 32% in the African/ African American population, 20% in the East Asian population, and 6-14% in other populations (online suppl.Table 2), indicating variants were distributed differently in these populations.
Considering the functions of glutamate, we hypothesized that there are several differences between genes associated with glutamate sensing and those associated with glutamate metabolism.The distribution of allele frequencies based on these two functional groups is shown in Figure 2. We observed significant differences in the allele frequencies of genes between these two functional groups (median test; q < 0.05) in Latino/admixed American, Ashkenazi Jewish, and other populations (Fig. 2a), indicating the migration history of these populations.We also examined the allele frequencies of genes in the glutamate-sensing and metabolism groups among populations (Fig. 2b) and found that only genes belonging to the metabolism group showed a significant difference (median test; q < 0.05) among populations.Furthermore, we performed pairwise median test among populations belonging to the metabolism group and found a significant difference between African/African American and East Asian populations.However, we could not observe any significant differences in the allele frequencies of glutamate-sensing genes among populations.

Discussion
We identified 145 missense variants among these 67 genes (Table 1).Our analysis was focusing only on the missense variants whose minor allele frequencies are >1%.Therefore, the variations in well-conserved genes were excluded from our analyses.As mentioned, AMPA-type glutamate receptors do not show any variants.However, this clearly showed AMPA-type glutamate receptors is under the strong selective pressure comparing to other genes listed, suggesting evolutionary importance of AMPA-type glutamate receptors for human.This is a potential limitation of our analysis, but we took this approach to identify variants with the phenotypes and established in each population.Also, it should be noted that we may have overestimated the variance in particular categories which contain conserved genes and do not consider the rare variants in populations.
The pairwise Fisher's exact test between variants revealed that 63% of the variants showed a significant difference in allele frequency in any pairwise combination of populations (online suppl.Tables 1, 2).It is noted that the average ratio of significant variants was relatively high in the African/African American population and in the East Asian population (online suppl.Table 2).As introduced in the introduction, for example, variants of TAS1R1 are associated with umami taste perception [14].These variants enriched in the Japanese population [14] were not in the French population [15], potentially causing differences in the taste perception of umami between these populations.This example illustrates the functional and regional differences in variants associated with the genetic background.In this study, TAS1R1 was included in the oral glutamate receptors, and the French population and the Japanese population were categorized into European (non-Finnish) and East Asian, respectively, following the gnomAD database.The pairwise Fisher's exact tests of variants in the oral glutamate receptors between the European (non-Finnish) and all the other populations showed that there are statistically significant variants only between the European  (non-Finnish) and the East Asian and between the European (non-Finnish) and African/African American populations (online suppl.Table 2).The average ratios of numbers of significant variants in the African/ African American and in the East Asian population were higher than the ratios in the other populations.These results implied the variants in the oral glutamate receptors were distributed in a different way in these two populations.Therefore, our results supported the findings earlier in [14,15].Furthermore, Raliou et al. [15] examined possible relations between phenotype (sensitivity to glutamate) and genotype (polymorphisms in glutamate taste receptors tas1r1, tas1r3, mGluR4, and mGluR1) at the individual level.Interestingly, 10 non-synonymous variants were identified in the French populations with various umami sensitivities including nontasters.They suggested molecules for umami perception, but synonymous variants linked to umami receptors in this study could provide novel insights into the umami perception among populations in future.
In this study, we classified genes according to their functions: the genes associated with oral glutamate receptors and glutamate transporters as glutamatesensing genes and the remaining genes associated with the TCA cycle, urea cycle, nitrogen metabolism, and brain glutamate receptors as glutamate metabolism genes.We examined the allele frequencies of genes in the glutamate-sensing and metabolism groups among populations and found that only genes belonging to the metabolism group showed a significant difference among populations (Fig. 2b).We could not observe any significant differences in the allele frequencies of glutamate-sensing genes among populations (Fig. 2b).These results clearly indicated that genes involved in glutamate sensing are conserved among populations.Jiang et al. [31] identified gene ontology categories conserved and differentiated in human populations.They reported that "cellular process," "catalytic activity," and "binding" were genetically highly diverged among populations."Multicellular organismal process" and "molecular transducer activity" showed low levels of genetic diversity.Comparing to our study, genes in the metabolic group such as genes in TCA cycle are associated with "cellular process," and glutamate-sensing genes such as TAS1R1/TAS1R3 are associated with "molecular transducer activity."Our findings follow this global tendency.Glutamate sensing is directly associated with the preference or desire to ingest glutamate-containing foods, and thus these genes are under the strong functional constraint at each population.Recently, it was reported that TAS1R1/TAS1R3 receptor gene was adaptively evolved to sense glutamate so that humans can consume leafy diets [32].This finding also suggests that glutamate-sensing genes are important to characterize and to conserve eating behavior as human.Contrary, glutamate metabolism is involved in the utilization of glutamate in the body.This utilization should be flexible to fit the needs of each population with different environmental conditions, cultural backgrounds, and diets during human evolution.
This study describes the genetic diversity of glutamate-related genes among distinct populations.Our analysis suggests that the allele frequencies of glutamate-related genes are conserved across groups of genes but divergent at the single nucleotide polymorphism level among populations, suggesting the importance of single nucleotide polymorphism-level surveys.Additionally, genes associated with glutamate sensing and metabolism showed distinct features among populations, indicating their significance in different populations.This study provides a comprehensive atlas of genes involved in glutamate functions in human populations worldwide and a basis for future studies on glutamate and taste perception at the population level.

Fig. 1 .
Fig. 1.Distribution of allele frequencies across 11 populations and 6 functional categories.The violin plots show the distribution of allele frequencies for populations belonging to 6 functional categories (a) and functional categories among 11 populations (b).The red triangle indicates the median, the blue circle shows the mean, and colored area represents the distribution of data based on kernel probability density.

Fig. 2 .
Fig. 2. Distribution of allele frequencies across 11 populations and 2 functional groups.The violin plots show the distribution of allele frequencies for populations belonging to 2 functional groups (a) and functional categories among 11 populations (b).The red triangle indicates the median, the blue circle shows the mean, and colored area represents the distribution of data based on kernel probability density.Statistical significance (indicated by the asterisk) was determined using median test.

Table 1 .
List of representative genes associated with glutamate functions Ann Nutr Metab 2023;79:502-510 DOI: 10.1159/000535181Goto/Masuzawa/Kohmura/Takumi/ Takeyama/Miyazaki/Gojobori/Mineta information on gene annotations, sequences, and glutamate metabolic pathways.Based on the relationship of genes with glutamate functions, we selected genes belonging to the following functional categories: (1) oral glutamate receptors, (2) TCA cycle, (3) urea cycle, (4) nitrogen metabolism, (5) brain glutamate receptors, and (6) glutamate transporters.Further, we categorized the genes associated with oral glutamate receptors and glutamate transporters as glutamate-sensing genes and the remaining genes associated with the TCA cycle, urea cycle, nitrogen metabolism, and brain glutamate receptors as glutamate metabolism genes.We used all genes in the Genome Aggregation Database (gnomAD) as control genes except for the abovementioned selected genes.

Table 1
(continued) include the truncated forms of metabotropic glutamate receptors in this study because it is unable to distinguish the intact and truncated forms in our analysis.Number of variants for these genes are shown as hyphens.