Genome Evolutionary Dynamics Meets Functional Genomics: A Case Story on the Identification of SLC25A44

Abstract

Gene clusters are becoming promising tools for gene identification. The study reveals the purposive genomic distribution of genes toward higher inheritance rates of intact metabolic pathways/phenotypes and, thereby, higher fitness. The co-localization of co-expressed, co-interacting, and functionally related genes was found as genome-wide trends in humans, mouse, golden eagle, rice fish, Drosophila, peanut, and Arabidopsis. As anticipated, the analyses verified the co-segregation of co-localized events. A negative correlation was notable between the likelihood of co-localization events and the inter-loci distances. The evolution of genomic blocks was also found convergent and uniform along the chromosomal arms. Calling a genomic block responsible for adjacent metabolic reactions is therefore recommended for identification of candidate genes and interpretation of cellular functions. As a case story, a function in the metabolism of energy and secondary metabolites was proposed for Slc25A44, based on its genomic local information. Slc25A44 was further characterized as an essential housekeeping gene which has been under evolutionary purifying pressure and belongs to the phylogenetic ETC-clade of SLC25s. Pathway enrichment mapped the Slc25A44s to the energy metabolism. The expression of peanut and human Slc25A44s in oocytes and Saccharomyces cerevisiae strains confirmed the transport of common precursors for secondary metabolites and ubiquinone. These results suggest that SLC25A44 is a mitochondrion-ER-nucleus zone transporter with biotechnological applications. Finally, a conserved three-amino acid signature on the cytosolic face of transport cavity was found important for rational engineering of SLC25s.

Keywords: SLC25 subfamily 44; candidate gene; electron transfer chains; flavonoids; gene clusters; gene organization; genomic co-localization; para-coumaric acid; resveratrol; ubiquinone.

Figure 1

Genome-wide trends in co-localization of co-expressed and co-interacting genes. (A) Number of co-expressed and co-interacting genes counted for every gene. The blue color is for the total number and the green color is for the cumulative number of events. (B) Percentage of genes with different fold-changes in likelihoods of co-localization with co-expressed and co-interacting partner genes. For every gene, the co-localization likelihood, i.e., the observed likelihood for co-expressed and co-interacting gene(s) located among the ±10 neighbor genes, was compared to its observed genome-wide likelihood of co-expressed and co-interacting genes, and fold-changes of ≥2 were considered as non-random distributions (see the methods). Homologs and duplicated events were excluded from the analyses. (C) Genes with co-localization likelihood fold-change of ≥2 had 1–20 co-expressed and co-interacting partner genes within the genomic window size of ±10 genes. (D) The co-localization likelihood of co-expressed and co-interacting genes across the inter-loci distances. GOI is the gene of interest and represents every gene in the genome. (E) Homolog genes between the human and mouse with co-localization likelihood fold-change of ≥2. The analyses were based on 16,974 homologs, and 254 homologs with missing data were excluded. (F) The frequency of overlapped co-expressed and co-interacting neighbor genes with identical physical gene order positions when comparing the human and mouse. The analyses were based on the 16,974 homolog genes between the human and mouse and within the genomic window size of ±10 genes from the homolog genes. (G) Total number of neighbor events and the number of overlapped neighbor events with identical physical gene order positions, both within the genomic window size of ±10 genes from the homolog genes, when comparing the human and mouse. (E–G) To exclude the effect of possible segmental inversions and differences in chromosomal orientations between the human and mouse, inverted but exactly same events were extracted and included in the analyses as overlapped events. FC: Fold change in the co-localization likelihood against the genome-wide likelihood, GOI: gene of interest.

Figure 2

Co-segregation of the co-localized events. The heatmap matrices of pairwise linkage disequilibrium statistics are shown. The analysis includes the SNPs in the flanking regions of the genes. The TLCD3B and ITPRIPL1 genes had interactions with all the 20 neighbor genes, and the co-segregation is demonstrated for the most distant neighbors. Complete linkage: D’ = 1, Random segregation: D’ = 0.

Figure 2

Co-segregation of the co-localized events. The heatmap matrices of pairwise linkage disequilibrium statistics are shown. The analysis includes the SNPs in the flanking regions of the genes. The TLCD3B and ITPRIPL1 genes had interactions with all the 20 neighbor genes, and the co-segregation is demonstrated for the most distant neighbors. Complete linkage: D’ = 1, Random segregation: D’ = 0.

Figure 3

Identification of SLC25A44 as a candidate transporter for the common precursors of ubiquinone and resveratrol. (A) The co-localization of membrane transporter genes with their co-expressed and co-interacting partner genes within the genomic window size of ±10 genes. (B) The identified genomic block in peanut. The metabolic relation among the genes is shown. (C) Metabolic pathway enrichment for the genes co-expressed with Slc25A44. (D) The phylogenetic tree of mitochondrial carriers. The tree includes SLC25 members from four primitive eukaryotes, four fungal species, two animals, one plant and one algal species, ancient SLC25 members from five bacterial species, and finally members of SLC25A44 from additional 18 animals and seven plant and algal species. See Table 1 and Table S5 for the species names and accession number of sequences. The clades are labeled by letters “a” for animal, “p” for plant, “f” for fungi, “pe” for primitive eukaryotes, and “b” for bacteria when there is at least one representative transporter member from the corresponding domains of life. CHS: chalcone synthase, ETC: electron transfer chains, FC: fold change in co-localization likelihood, NDUF: NADH ubiquinone oxidoreductase, and STS: resveratrol synthase.

Figure 3

Identification of SLC25A44 as a candidate transporter for the common precursors of ubiquinone and resveratrol. (A) The co-localization of membrane transporter genes with their co-expressed and co-interacting partner genes within the genomic window size of ±10 genes. (B) The identified genomic block in peanut. The metabolic relation among the genes is shown. (C) Metabolic pathway enrichment for the genes co-expressed with Slc25A44. (D) The phylogenetic tree of mitochondrial carriers. The tree includes SLC25 members from four primitive eukaryotes, four fungal species, two animals, one plant and one algal species, ancient SLC25 members from five bacterial species, and finally members of SLC25A44 from additional 18 animals and seven plant and algal species. See Table 1 and Table S5 for the species names and accession number of sequences. The clades are labeled by letters “a” for animal, “p” for plant, “f” for fungi, “pe” for primitive eukaryotes, and “b” for bacteria when there is at least one representative transporter member from the corresponding domains of life. CHS: chalcone synthase, ETC: electron transfer chains, FC: fold change in co-localization likelihood, NDUF: NADH ubiquinone oxidoreductase, and STS: resveratrol synthase.

Figure 4

The highly conserved Slc25A44 gene has transport activity when expressed in X. laevis oocytes. (A) The conserved three-amino acid signature in SLC25A44s and four other subclades within the phylogenetic ETC-clade. (B) The predicted whole and cytosolic face structure of the AdSLC25A44. Transmembrane helices (TH) and conserved tryptophan residues on TH4 are illustrated. (C) The structure of locus and polymorphism distribution for Arabidopsis and human Slc25A44s. (D–F) Functional transport studies by expressing AdSLC25A44 in oocytes. (D) Export assay: resveratrol and para-coumaric acid were injected into the oocytes and were quantified in the medium after 210 min. (E,F) Import assay: resveratrol and para-coumaric acid (E) or cinnamic and 4-aminobanzoic acids (F) were added into the medium and intracellular levels of these compounds in oocytes were quantified after 210 min. (D–F) Bars represent mean ± Std. n = 3–4 biological independent samples each with 20 (D) or 30 (E,F) oocytes. The two-tailed Student’s t-test was used to find the significant transport activities at 1% level (marked by two asterisks) when compared to the control with no heterologous expression (RNA-free solution was injected). ns: not significant.

Figure 4

The highly conserved Slc25A44 gene has transport activity when expressed in X. laevis oocytes. (A) The conserved three-amino acid signature in SLC25A44s and four other subclades within the phylogenetic ETC-clade. (B) The predicted whole and cytosolic face structure of the AdSLC25A44. Transmembrane helices (TH) and conserved tryptophan residues on TH4 are illustrated. (C) The structure of locus and polymorphism distribution for Arabidopsis and human Slc25A44s. (D–F) Functional transport studies by expressing AdSLC25A44 in oocytes. (D) Export assay: resveratrol and para-coumaric acid were injected into the oocytes and were quantified in the medium after 210 min. (E,F) Import assay: resveratrol and para-coumaric acid (E) or cinnamic and 4-aminobanzoic acids (F) were added into the medium and intracellular levels of these compounds in oocytes were quantified after 210 min. (D–F) Bars represent mean ± Std. n = 3–4 biological independent samples each with 20 (D) or 30 (E,F) oocytes. The two-tailed Student’s t-test was used to find the significant transport activities at 1% level (marked by two asterisks) when compared to the control with no heterologous expression (RNA-free solution was injected). ns: not significant.

Figure 5

The transport activities of AdSLC25A44 and HsSLC25A44 in S. cerevisiae. (A) Subcellular localization of AdSLC25A44_GFP in yeast. (B,C) Para-coumarate, 4-aminobenzoate, and 4-hydroxybenzoate concentrations in the fermentation broth after 72 h growth of the yeast strains producing para-coumaric acid from phenylalanine by the enzymes PAL and C4H (B) or from tyrosine by the enzyme TAL (C). Bars represent mean ± Std; n = 3 biological independent samples. The two-tailed Student’s t-test was used to compare the strains with and without the heterologous transporters. The significant events at 5% and 1% levels are labeled by one and two asterisks, respectively. C4H: Cinnamic acid hydroxylase, PAL: Phenylalanine ammonia-lyase, TAL: Tyrosine ammonia-lyase, and ns: not significant.

Figure 6

The SLC25A44 is a transporter of ubiquinone precursors with potential to boost the bio-based production of para-coumaric acid. (A,B) The effect of AdSLC25A44^LWW206IQF expression on the production rate of para-coumaric, 4-hydroxybenzoic, and 4-aminobenzoic acids in the PAL and TAL yeast strains. Bars represent mean ±Std. n = 6–9 biological independent samples. The two-tailed Student’s t-test was used to compare the strains with and without AdSLC25A44^LWW206IQF. The significant events at 1% level are marked by two asterisks. (C) Logarithmic cell growth was measured spectrophotometrically at A₆₀₀ nm. Significant differences, i.e., α vs. α′ and β vs. β′ (p = 10⁻⁴), were determined through one-way ANOVA followed by the least significant difference test. Bars represent mean ±Std. n = 3 biological independent samples. (D) A simplistic roadmap for the involvement of SLC25A44 and the other transporter members of ETC-clade in the mitochondrial electron transfer chains. 4HB: 4-hydroxybenzoate, 4AB: 4-aminobenzoate, I-IV: electron transfer chains complexes, CytC: cytochrome C, Q: ubiquinone, SAM: S-adenosylmethionine, and ns: not significant.