Transcriptional Regulation of the Genes Encoding Branched-Chain Aminotransferases in Kluyveromyces lactis and Lachancea kluyveri Is Independent of Chromatin Remodeling

Abstract

In yeasts, the Leu3 transcriptional factor regulates the expression of genes encoding enzymes of the leucine biosynthetic pathway, in which the first committed step is catalyzed by α-isopropylmalate synthase (α-IPMS). This enzyme is feedback inhibited by leucine, and its product, α-isopropylmalate (α-IPM), constitutes a Leu3 co-activator. In S. cerevisiae, the ScBAT1 and ScBAT2 genes encode branched-chain aminotransferase isozymes. ScBAT1 transcriptional activation is dependent on the α-IPM concentration and independent of chromatin organization, while that of ScBAT2 is α-IPM-independent but dependent on chromatin organization. This study aimed at understanding whether chromatin remodeling determines the transcriptional regulation of orthologous KlBAT1 and LkBAT1 genes in Kluyveromyces lactis and Lachancea kluyveri under conditions in which the branched-chain amino acids are synthesized or degraded. The results indicate that, in K. lactis, KlBAT1 expression is reduced under catabolic conditions, while in L. kluyveri, LkBAT1 displays a constitutive expression profile. The chromatin organization of KlBAT1 and LkBAT1 promoters did not change, maintaining the Leu3-binding sites free of nucleosomes. Comparison of the α-IPMS sensitivities to feedback inhibition suggested that the main determinant of transcriptional activation of the KlBAT1 and LkBAT1 genes might be the availability of the α-IPM co-activator, as reported previously for the ScBAT1 gene of S. cerevisiae.

1. Introduction

Metabolism and transcription constitute two functional levels, which are interconnected and are essential for the response to modifications of external nutrient availability of all living organisms. Metabolites can control gene expression, which modulates enzyme concentrations and metabolic activity [1]. Fluctuations in metabolism can activate particular transcriptional factors, which evoke specific gene expression profiles [1]. In yeast, the metabolites leucine and α-isopropylmalate (α-IPM) and the transcriptional factor Leu3 jointly participate in the regulation of the leucine biosynthetic pathway [2,3,4]. The first committed step in this pathway is catalyzed by α-isopropylmalate synthase (α-IPMS, EC 2.3.3.13), an enzyme that is feedback inhibited by leucine and whose activity influences transcriptional regulation because its product α-IPM is also a co-activator of the Leu3 protein. The complex Leu3-α-IPM activates the transcription of genes of the leucine, valine, isoleucine, and glutamate synthetic pathways [5,6]. Leucine biosynthesis and its transcriptional regulation in Saccharomyces cerevisiae are summarized in Figure 1a. Leucine is synthesized from pyruvate by the sequential reactions catalyzed by Ilv2/Ilv6, Ilv5, Ilv3, Leu4/Leu9, Leu1, Leu2, and Bat1 enzymes [7] and can be degraded to α-IPM by the reverse pathway Bat2-Leu2-Leu1 [8]. Leucine concentration can control the pathway by feedback inhibition of the α-IPMS isozymes Leu4/Leu9 [9], while the complex Leu3-α-IPM activates the expression of the nine genes encoding for the enzymes of the leucine biosynthetic pathway [4,5,10]. The canonical mechanism proposed is that Leu3 can act as an activator or repressor, depending on the α-IPM concentration. Transcription of the Leu3-regulated genes is active at high levels of α-IPM and inactive at low α-IPM levels [5,10].

It has been documented that Leu3 is permanently bound to its binding site located in the nucleosome-free region (NFR) of the ScBAT1 promoter, and its transcriptional activation is dependent on α-IPM availability [8]. However, feedback inhibition of the Leu4/Leu4 and Leu4/Leu9 isoenzymes reduces α-IPM synthesis, which in turn impairs ScBAT1 expression due to depletion of the complex Leu3-α-IPM [8,11,12]. Conversely, ScBAT2 paralogue expression is not dependent on α-IPM availability; instead, its transcriptional regulation requires chromatin remodeling [8]. S. cerevisiae paralogous genes ScBAT1 and ScBAT2 encode ScBat1 and ScBat2 branched-chain aminotransferases (BCATs, EC 2.6.1.42), which catalyze the first step of the catabolism and the last step of the biosynthesis of branched-chain amino acids (BCAAs), namely valine, isoleucine, and leucine [13,14]. It has been proposed that the ScBAT1 and ScBAT2 genes arose from a hybridization and whole-genome duplication event (WGD) [15] (Figure 1b). It has been hypothesized that the yeasts Kluyveromyces lactis and Lachancea kluyveri descend from the pre-WGD ancestor and have a single BAT-encoding gene (Figure 1b).

It has been shown that the BCATs from K. lactis (KlBat1) and L. kluyveri (LkBat1) represent bifunctional enzymes that are involved in both BCAAs biosynthesis and catabolism [12,16]. When glutamine (Gln) is provided as a nitrogen source, the synthesis of branched-chain amino acids is active, and the transcription of the KlBAT1 gene is induced. Contrastingly, when the culture media is supplemented with high concentrations of BCAAs, transcription of the KlBAT1 gene is repressed [12]. Thus, the KlBAT1 gene displays a similar expression response to that found for the S. cerevisiae ScBAT1 paralogue, while LkBAT1 displays constitutive expression on Gln or high concentrations of BCAAs as nitrogen sources, indicating biosynthetic and catabolic roles of the encoded enzyme [12,16].

The objectives of this work were as follows: (i) to analyze the evolutionary history of the BCATs in the Saccharomycotina, (ii) to determine the role of chromatin organization in gene expression profiles of orthologous KlBAT1 and LkBAT1, and (iii) to correlate the α-IPMS sensitivities to feedback inhibition by leucine, with the occurrence of transcriptional repression of the genes encoding BCATs in the presence of high concentrations of BCAAs. Our results indicate that although ScBAT1, KlBAT1, and LkBAT1 orthologous genes are phylogenetically closely related, the transcriptional regulatory mechanisms of the LkBAT1 gene have significantly diverged to maintain high expression levels even when high concentrations of BCAAs are present in the culture media. Here, we show that chromatin remodeling is not involved in the transcriptional regulation of the KlBAT1 and LkBAT1 genes during biosynthetic and catabolic conditions and that Leu3-binding sites are present in NFR of their promoter sequences, as has been reported for the ScBAT1 gene of S. cerevisiae. A comparison of the sensitivities to leucine inhibition of the α-IPMS activities in L. kluyveri and K. lactis revealed that in the former, the α-IPMS enzymes are significantly less sensitive than in the latter. These results suggest that the lack of transcriptional repression of the LkBAT1 gene under catabolic conditions in L. kluyveri is most likely due to an increased synthesis of the transcriptional co-activator α-IPM. Our results support the notion that the availability of the α-IPM co-activator is the main determinant factor of transcriptional activation of the Leu3-regulated genes in L. kluyveri and K. lactis, highlighting the importance of metabolite concentrations as determinants of transcription levels.

2. Materials and Methods

2.1. Growth Conditions

The K. lactis, L. kluyveri, and S. cerevisiae wild-type and mutant strains used in this study are described in

Table 1. Yeast strains used in this study.

Species/Strain	Genotype	Source
K. lactis/Y155	MATα ade2 his3 ura3	[12]
L. kluyveri/Y156	MATαura3	[17]
K. lactis/Y155-1 Klleu4Δ	MATαade2 his3 ura3 KlLEU4::KanMX4	[18]
K. lactis/Y155-2 Klleu4bisΔ	MATαade2 his3 ura3 KlLEU4bis::KanMX4	[18]
L. kluyveri/Y156-3 Lkleu4Δ	MATα ura3 LkLEU4::NatMX	[19]
L. kluyveri/Y156-4 Lkleu4bisΔ	MATαura3 LkLEU4bis::KanMX4	[19]
S. cerevisiae/CLA11-700	MATα ura3	[8]

. All of them were grown on minimal media (MM) containing salts, trace elements, and vitamins without amino acids according to the formula for yeast nitrogen base (Difco, BD, USA). Glucose (2% w v⁻¹) was used as a carbon source, and glutamine (Gln) (7 mM) or VIL (valine 150 mg L⁻¹ + leucine 100 mg L⁻¹ + isoleucine 30 mg L⁻¹) was used as a nitrogen source as reported previously [12,16]. Uracil (20 mg L⁻¹), adenine (20 mg L⁻¹), and histidine (20 mg L⁻¹) were added as auxotrophic requirements when needed. All the cultures were incubated at 30 °C with shaking (250 rpm). Unless otherwise stated, reagents were purchased from Sigma-Aldrich-Merck (Darmstadt, Germany).

2.2. Strains

The single mutants Klleu4Δ (Y155-1), Klleu4bisΔ (Y155-2), Lkleu4Δ (Y156-3), and Lkleu4bisΔ (Y156-4) were respectively obtained from the strains Y155 or Y156 by gene replacement. The corresponding coding sequences were replaced by the KanMX4 or NatMX modules, as described previously [18,19]. To construct the K. lactis Klleu4Δ mutant strain (Y155-1, Table 1), the coding sequence of KlLEU4 (KLLA0F23529g) was replaced by homologous recombination in the K. lactis wild-type strain Y155 using a module containing the KanMX4 cassette flanked by 676 bp of 5′ untranslated region (UTR) (−676 to −1) and 474 bp of 3′ UTR (+1831 to +2305) sequences of KlLEU4 [18]. The same strategy was used to construct the K. lactis Klleu4bisΔ mutant strain (Y155-2, Table 1); the coding sequence KlLEU4bis (KLLA0D14201g) was replaced by the KanMX4 cassette flanked by 735 bp of 5′ UTR (−735 to −1) and 682 bp of 3′ UTR (+1848 to +2530) of the KlLEU4bis gene [18]. The L. kluyveri Lkleu4Δ single mutant (Y156-3, Table 1) was constructed by replacing the LkLEU4 (SAKL0E10472) coding sequence by the NatMX cassette flanked by 927 bp of 5′ UTR and 954 bp of 3′ UTR sequences of the LkLEU4 gene by homologous recombination in strain L. kluyveri Y156 [19]. Similarly, the L. kluyveri Lkleu4bisΔ single mutant (Y156-4, Table 1) was constructed replacing the coding sequence of LkLEU4bis (SAKL0F05170g) by the KanMX4 cassette flanked by ≈1000 bp of 5′ and 3′ UTR sequences of the LkLEU4bis gene in the wild-type L. kluyveri strain Y156 [19].

2.3. Phylogeny of Yeast Proteins

The PhylomeDB (http://phylomedb.org/ accessed on 9 February 2024) database was used to analyze BCAT phylogeny [20]. PhylomeDB is a public database for complete catalogs of gene phylogenies (phylomes) that allows users to interactively explore the evolutionary history of genes through the visualization of phylogenetic trees and multiple sequence alignments. The proteomes used for the reconstruction of Saccharomycotina phylomes can be consulted at http://phylomedb.org/phylome_206 accessed on 9 February 2024.

2.4. Quantitative Polymerase Chain Reaction for Differential Transcript Expression Analysis

Quantitative polymerase chain reaction (qPCR) and total yeast RNA extraction were performed as described earlier [21]. Cultures were grown to an OD₆₀₀~0.5 in MM with glutamine (Gln) or VIL as the sole nitrogen source and 2% glucose as the carbon source. The cells were harvested by centrifugation, and the pellet was washed with water and then disrupted with a mortar and pestle with liquid nitrogen. The TRIzol reagent (Invitrogen, Waltham, MA, USA) was used to extract total RNA, which was treated with RQ1 RNAse-Free DNAse (Promega, Madison, WI, USA) to degrade DNA from the samples. The RevertAid H Minus First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA) was used to obtain cDNA, of which 30 ng µL−1 was used to perform PCR using the KAPA SYBR Fast kit (Roche, Basel, Switzerland) and the Corbett Research Rotor-Gene 6000 (Qiagen, Hilden, Germany). Transcript levels were determined for ScBAT1, ScBAT2, KlBAT1, and LkBAT1 using deoxyoligonucleotides FW. ScB1qPCR to RV. Lk18SqPCR (Table S1). To compare the expression levels, the 2^−ΔΔCT method was used, employing the 18S gene as a housekeeping control for S. cerevisiae, K. lactis, and L. kluyveri. Statistical analysis was performed using GraphPad Prism 9.0 (GraphPad Software Inc., Boston, MA, USA). All experiments were performed in triplicate, and data are presented as mean ± standard error of the mean.

2.5. Nucleosome Scanning Assay

Nucleosome scanning experiments were performed as previously described [8]. Wild-type K. lactis and L. kluyveri strains were grown to a OD₆₀₀~0.5 in 50 mL MM 2% glucose with 7 mM glutamine (Gln) or VIL as a nitrogen source. DNA and its associated proteins were cross-linked with formaldehyde, internucleosomal regions were digested with MNase, and then, proteins with proteinase K. Extracted DNA was separated by 1.5% agarose gel electrophoresis and monosomal bands were cut and purified. qPCR was performed to independently determine relative MNase protection of KlBAT1 (KLLA0A10307g) and LkBAT1 (SAKL0B12496g) templates which was calculated as a ratio considering amplification of a region of KlVCX1 and LkVCX1 respectively, with the following deoxyoligonucletide pairs: KlFw 5′-CGA GCG AGC ATT TTG GAC CCG TTT-3′ and KlRv 5′-TGG AGA CAG TAG TAC CTG AGA TGA TC-3′ (for KlVCX1), and LkFw 5′-GGT GGG TAC AAC AGA ATC CAA CAG-3′ and LkRv 5′-GGC ACA GGA AAT GGC CAA CAA GGA-3′ (for LkVCX1). These regions within the KlVCX1 and LkVCX1 genes were used as controls because a well-positioned nucleosome was observed in the presence of either Gln or VIL as sole nitrogen sources (Figure S1). PCR deoxyoligonucleotides are described in Tables S2 and S3, which amplify from around −600 pb to +250 bp of KlBAT1 or LkBAT1 loci whose coordinates are given relative to the +1 ATG. All presented nucleosome scanning assays represent the mean values and standard errors of at least three independent biological replicates.

2.6. In Silico Promoter Analysis

We examined a 600 bp intergenic region upstream of the start codon of the branched-chain amino acid transaminase genes of K. lactis and L. kluyveri genes. The 600 bp sequences upstream of the predicted start codon were subject to in silico promoter analysis. All genomic sequences analyzed in this study were obtained from the YGOB database (http://ygob.ucd.ie/ygob/ accessed on 16 February 2024) [22]. Sequences were subject to motif scanning using the “Matrix Scan” program, a member of the “RSA tools” package [23]. The yeast transcription factor matrix motifs used for this analysis were downloaded from the “YeTFaSCo” database (http://yetfasco.ccbr.utoronto.ca/ accessed on 16 February 2024) [24]. Sequence alignments were made with the SIM tool (https://web.expasy.org/sim/ accessed on 16 February 2024).

2.7. α-IPMS Enzyme Assay and IC50 Determination

Analysis of α-IPMS activity was carried out as described [11]. For the determination of the half-maximal (50%) inhibitory concentration (IC50), a dose–response curve was obtained using increasing concentrations of leucine and plotted on a logarithmic scale. The substrate concentrations were 0.25 mM acetyl coenzyme A (AcCoA) and 2.5 mM α-ketoisovalerate (α-KIV). Data were analyzed using the GraphPad Prism software version 8.2.1 (USA).

3. Results

3.1. Phylogeny of Branched-Chain Aminotransferases in the Saccharomycotina

BCATs are conserved throughout the tree of life. The S. cerevisiae ScBAT1 syntenic copy shows analogous genomic positions with KlBAT1 from K. lactis and LKBAT1 from L. kluyveri. These three genes encode for mitochondrial BCATs [12,16]. BCATs phylogenetic tree obtained from PhylomeDB showed at least one ancestral duplication in the Saccharomycotina clade, so the two paralogous genes present in S. cerevisiae seemed to have originated before the WGD event (Figure 1b). ScBat1, KlBat1, and LkBat1 proteins cluster together with BCATs from Kluyveromyces and Lachancea clades, which conserved one of the ancestral duplicated proteins. Contrastingly, the Bat2 orthologous proteins are only present in post-WGD species such as Saccharomyces bayanus, Saccharomyces paradoxus, and Candida glabrata. Yeast species harboring only Bat1 may have undergone selective pressures favoring a streamlined metabolic pathway specific to certain branched-chain amino acids. In contrast, the presence of both Bat1 and Bat2 in certain yeast lineages could imply a more diversified metabolic strategy.

3.2. KlBAT1 and LkBAT1 Expression Profiles Are Not Dependent on Chromatin Remodeling

Gene expression and chromatin organization of promoters were analyzed using Gln or VIL as the sole nitrogen source. qPCR analysis and Nucleosome Scanning Assay (NuSA) were carried out on the pertinent samples simultaneously obtained from exponentially growing yeast cultures (OD₆₀₀ = 0.5), as described in Section 2. The results confirmed a biosynthetic expression profile for KlBAT1 (Gln-induced and VIL-repressed) (Figure 2a) [12] and a constitutive expression for LkBAT1 (Figure 2b) [16]. In S. cerevisiae, it has been reported that transcript levels of the ScBAT1 gene are lower on VIL than on Gln and that those of the ScBAT2 gene show the opposite expression pattern [8]. These genes were used as controls (Figure 2c,d); as expected, the ScBAT1 and ScBAT2 genes from S. cerevisiae showed divergent expression profiles.

In order to analyze the chromatin structure, NuSA assays were carried out to determine nucleosome positioning across the KlBAT1 and LkBAT1 promoters in cells grown on Gln or VIL as sole nitrogen sources. DNA isolated from the MNase digestion was first examined in the KlVCX1 locus (KLLA0F05632g) and LkVCX1 locus (SAKL0F09878g) to find a protected region to normalize our data. A region with significant protection in KlVCX1 or LkVCX1 locus was observed (Figure S1). qPCR was performed to amplify overlapping regions of KlBAT1 or LkBAT1 promoters using 25 and 26 primer pairs, respectively (Tables S2 and S3). Peaks of relative protection indicated that in either Gln or VIL, four nucleosomes were positioned around the KlBAT1 promoter (−2, −1, +1, and +2) (Figure 3a). Nucleosomes −1 and +1 constitute the border of the 100 bp MNase-sensitive NFR, as was found for S. cerevisiae ScBAT1 paralogous [8], which spans from around −200 to −100 with respect to the KlBAT1 +1 ATG (Figure 3a). Therefore, the KlBAT1 promoter has a similar chromatin organization in samples obtained from Gln or VIL, indicating that in K. lactis, no chromatin remodeling is required to attain VIL-dependent repression.

For the LkBAT1 promoter, NuSA analysis revealed that on Gln or VIL, at least three nucleosomes designated −2, −1, and +2 are clearly positioned (Figure 3b). The region that constitutes the +1 nucleosome contains the TATA_BOX. Protection of this region was low under both conditions. The LkBAT1 NFR spans from −280 to −50 bp. As expected, LkBAT1 showed almost identical chromatin organization on either Gln or VIL, in agreement with the observed constitutive expression (Figure 2b and Figure 3b). Therefore, the expression profiles of KlBAT1 and LkBAT1 are not dependent on chromatin remodeling in either condition.

3.3. The Putative LEU3-Binding Site Is Located within the NFR of KlBAT1 and LkBAT1 Promoters

To test for the functionality of the predicted consensus transcription factor (TF) binding sites that could be involved in KlBAT1 and LkBAT1 expression profiles, we performed a comparative in silico analysis of the location of cis-regulatory elements in KlBAT1 and LkBAT1 promoters (Figure 4) as described in Section 2. It was found that, on both promoters, in the NFR, a putative Leu3-binding site is located (Figure 4a,b), which is known to play a role in the expression of the ScBAT1 gene from S. cerevisiae [8]. The putative Leu3-binding sites within the NFR in ScBAT1, KlBAT1, and LkBAT1 are 5′-GCCGGTACCGGC-3′ at −147 bp, 5′-GGCAATCTGCC-3′ at −230 bp, and 5′-CCGGTAGCGG-3′ at −171 bp, respectively (Figure 4c,d, Table S4). In these three binding sites, the everted CGG half-sites are shown in bold letters [8,25].

To explore the possibility that the function and regulation of the Leu3 proteins from K. lactis, L. kluyveri, and S. cerevisiae were conserved, the corresponding protein sequences were compared. For the S. cerevisiae protein, it has been reported that the residues A36, R41, Q43, K44, R75, K78, and R79 recognize the everted CGG half-sites and eight adjacent phosphodiester bonds [25]. Notably, these residues are conserved in the Leu3 proteins of K. lactis and L. kluyveri, as well as the six cysteine residues that ligate the two zinc ions in the DNA-binding domain (Figure 5) [25]. Zhou and coworkers demonstrated that, in S. cerevisiae, the DNA-binding domain of Leu3 encompasses the 173 N-terminal residues, and the activation domain comprises the 32 C-terminal residues, whereas the central part of the protein (residues 174 to 773) determines its response to the α-isopropylmalate co-activator; this central part is connected to the activation domain by a stretch of 81 residues [2,3]. Pairwise comparisons of the Leu3 protein sequences of the three species indicated that the identity in the DNA-binding domain ranges from 75 to 76.5%, in the central part of the protein from 60.5 to 65.6%, and in the last 26 residues from 84.6 to 91.7% (Figure 5). Putative Leu3-binding sites were also identified in the promoter regions of the genes encoding the enzymes of the leucine biosynthetic pathway in K. lactis and L. kluyveri (Supplementary Table S4). This analysis suggests that the function and regulation of Leu3 are conserved and that this transcription factor could be positioned on the NFR of the ScBAT1, KlBAT1, and LKBAT1 genes, ready to activate under conditions in which the α-IPM subcellular concentration is high enough to allow transcriptional activation.

3.4. Leucine Feedback Sensitivity of the α-IPMS Enzymes Inversely Correlates with KlBAT1 and LkBAT1 Expression Levels

The capacity of the Leu3 protein to activate transcription is dependent on the availability of the co-activator α-IPM, which is produced by the enzyme α-IPMS. The amino acid leucine exerts feedback inhibition on α-IPMS activity. In K. lactis, this activity is the result of the function of the KlLeu4 and KlLeu4bis isoenzymes [18], while in L. kluyveri, it is the result of the LkLeu4 and LkLeu4bis isoenzymes [19]. For S. cerevisiae, it has been previously determined that the transcriptional activation of the ScBAT1 gene by Leu3 is proportional to the capacity of α-IPM synthesis by the Leu4 and Leu9 α-IPMS isoforms [8].

To explore the possibility that the capacity of α-IPM synthesis in K. lactis and L. kluyveri influences transcriptional activation of the KlBAT1 and LKBAT1 genes, the IC50 for the KlLeu4 (LEU4 syntenic isoform), KlLeu4bis (LEU4 non-syntenic isoform), LkLeu4 (LEU4 syntenic isoform) and LkLeu4bis (LEU4 non-syntenic isoform) isoforms, was determined in cell extracts obtained from exponentially grown cultures of the Klleu4bisΔ, Klleu4Δ, Lkleu4bisΔ, and Lkleu4Δ single mutants as described in Section 2. As Figure 3c,d shows, the IC50 for both of the K. lactis isoforms displayed higher leucine sensitivity (IC50 = 0.038 mM and 0.023 mM) than that observed for the LkLeu4 and LkLeu4bis isoforms (IC50 = 0.27 mM and 0.13 mM). These results are in accordance with previous reports [18,19] and support the notion that the repression of the KlBAT1 gene on VIL (Figure 2a) could be the result of feedback inhibition of the α-IPMS isoenzymes, thereby reducing the α-IPM intracellular pool and thus the amount of the Leu3-α-IPM activator complex; while the low sensitivity of the α-IPMS isoenzymes in L. kluyveri could sustain enough α-IPM synthesis in the presence of VIL with the corresponding transcriptional activation of the Leu3-α-IPM-dependent genes like LkBAT1 (Figure 2b).

4. Discussion

The leucine biosynthetic pathway in yeast illustrates the interplay of metabolism and transcriptional regulation. The role of α-IPM as a Leu3 co-activator represents a direct link between feedback inhibition of the first committed step of the leucine biosynthetic pathway and transcriptional regulation of genes of a regulon that influences the synthesis of valine, isoleucine, leucine, and glutamate in yeast and other fungi [10,26]. Our results show evidence that highlights the role of the α-IPM co-activator as a key modulator of gene expression and suggests that chromatin remodeling could not play a role as the main determinant of transcriptional regulation of the Leu3-regulated genes KlBAT1 and LkBAT1 on either glutamine or VIL as nitrogen sources. These observations emphasize the role of small molecules acting as transcriptional coregulators under metabolic conditions in which chromatin organization remodeling could not be achieved. Our results support the notion that such a determinant role is conserved among pre- and post-WGD yeast species.

In S. cerevisiae, the Leu3-α-IPM complex, besides activating BCAA biosynthesis pathway genes, also activates expression of the GDH1 gene, which encodes the NADP⁺-dependent glutamate dehydrogenase, an enzyme crucial for ammonia assimilation and glutamate biosynthesis [6], this amino acid participates in many transamination reactions distributing amino groups throughout metabolism. The Leu3-α-IPM complex also activates the expression of the ILV5 gene, which encodes a moonlighting protein that participates in the synthesis of branched-chain amino acids and in the maintenance of mitochondrial DNA stability [27]. The intracellular concentrations of leucine and valine, in turn, influence the regulation of other biosynthetic pathways through the activation of the TORC1 kinase, a key regulator of cell division in response to nutrient availability [28]. The BCAA intracellular pools also promote chronological longevity [29]. These cellular processes most likely are also influenced by BCAAs in K. lactis and L. kluyveri, although the regulatory mechanisms needed to achieve homeostasis of BCAAs in each species may be different. Interestingly, leucine has been selected also in bacteria and archaea to link the environmental nutritional status with a variety of cellular processes [30].

The concentration of BCAAs is determined by the rate of synthesis, catabolism, and transport. In S. cerevisiae, the availability of two BAT genes allowed the specialization of ScBAT1 for branched-chain amino acid biosynthesis and ScBAT2 for their catabolism. However, in K. lactis and L. kluyveri, the regulation of the single gene encoding for BCATs must warrant the proper concentrations of BCAAs in response to changes in nutrient availability. As expected, the results presented in this study show that the presence of high concentrations of BCAAs in the culture media reduced the rate of valine, isoleucine, and leucine synthesis, triggering repression of the KlBAT1 gene in K. lactis, confirming previous observations [12]. However, in L. kluyveri, this repression was not observed in accordance with Montalvo-Arredondo et al., 2015 [16]. Here, we demonstrate that changes in chromatin organization are not involved in differential transcriptional regulation and that the Leu3-binding sites in the KlBAT1 and LkBAT1 promoters are located in NFR.

The high conservation of the amino acid sequence of the Leu3 proteins and the presence of putative Leu3-binding sites in the promoter regions of the genes encoding for enzymes of the leucine biosynthetic pathway suggest that the function and regulation of the Leu3 proteins are similar in S. cerevisiae, K. lactis, and L. kluyveri. Thus, in the case of the KlBAT1 and LkBAT1 genes, it is likely that the Leu3 protein remains bound to the NFR in the promoters and that the key factor determining the rate of KlBAT1 and LkBAT1 transcription is actually the α-IPM concentration, as in the case of the ScBAT1 gene from S. cerevisiae [8]. In line with this proposition, the lower sensitivity to leucine inhibition of the α-IPMS activity in L. kluyveri, as compared to that found in K. lactis, could maintain enough amount of the Leu3-α-IPM-activating complex, thus achieving constitutive expression of the LkBAT1 gene even in the presence of VIL. Contrastingly, the strong inhibition of the α-IPMS activity observed in K. lactis may have resulted in low levels of transcription of the KlBAT1 gene due to lack of the co-activator α-IPM. Apparently, in the lineage of L. kluyveri, a selection pressure favored the mild feedback inhibition of the α-IPMS activity, resulting in the constitutive expression of the Leu3-regulated genes. Considering the multiple connections of leucine metabolism with other cellular processes, it is intriguing to ponder the consequences of this regulatory peculiarity.

The assumption that the function and regulation of the Leu3 transcriptional factors are similar in K. lactis, L. kluyveri, and S. cerevisiae is based on the highly conserved sequences of the corresponding proteins; however, presented results suggest further functional studies could be interesting to follow in order to establish which could be the mechanisms involved in transcriptional regulation of the KlBAT1 and LkBAT1 genes in leu3Δ mutant strains.

5. Conclusions

Although the genes ScBAT1, KlBAT1, and LkBAT1 cluster together in the same branch of a phylogenetic tree and the Leu3 protein is highly conserved, the transcriptional regulation of the genes encoding branched-chain aminotransferases in S. cerevisiae, K. lactis, and L. kluyveri shows significant differences. The ScBAT1 gene is specialized for the synthesis of BCAAs, while the ScBAT2 gene is specialized for their catabolism. The KlBAT1 and LKBAT1 genes are involved in both synthesis and catabolism; however, the KlBAT1 gene is repressed in the presence of VIL, whereas the LkBAT1 gene is constitutively transcribed. Chromatin remodeling in the promoter regions of these genes does not play a significant role in the transcriptional regulation of the ScBAT1, KlBAT1, and LkBAT1 genes, but Leu3-binding sites are in NFR, suggesting that this transcriptional factor remains bound to DNA. The observed inverse correlation between the sensitivity of the α-IPMS activities to inhibition by leucine and transcriptional levels of the KlBAT1 and LkBAT1 genes on VIL indicates that the availability of the α-IPM co-activator could modulate the transcription levels.