l-Proline Accumulation and Freeze Tolerance in Saccharomyces cerevisiae Are Caused by a Mutation in the PRO1 Gene Encoding γ-Glutamyl Kinase

Yeast and bacterial strains.All yeasts used in this study were strains of S. cerevisiae. MB329-17C (MATα ura3-52 trp1 put1-54) and MB379-4C (MATahis4-42 put1-54), derived from crosses between S288C and Σ1278b, were mutant strains deficient in proline oxidase (put1 mutation) and were supplied by M. C. Brandriss (42). FH515, an AZC-resistant mutant strain with higher levels of intracellular proline, was isolated previously from strain MB329-17C after ethyl methanesulfonate treatment (38). In this study, put1 gene disruptant strain INVDput1 (MATahis3-Δ1 leu2 trp1-289 ura3-52 put1::URA3) and put1 and pro1 gene disruptant strain INVDput1pro1 (MATahis3-Δ1 leu2 trp1-289 ura3-52 put1::URA3 pro1::CgHIS3) were constructed from strain INVSc1 (MATahis3-Δ1 leu2 trp1-289 ura3-52) (Invitrogen, Carlsbad, Calif.), which is the wild-type strain with a S288C background. E. coli strain DH5α [F⁻ λ⁻ φ80lacZΔM15 Δ(lacZYA argF)U169 deoR recA1 endA1 hsdR17(r_k⁻_k⁺) supE44 thi-1 gyrA96] was used to construct the yeast genomic library and to subclone the yeast gene.

Plasmids.Two S. cerevisiae-E. coli shuttle vectors, p366 (ATCC 77163) and pRS415 (Stratagene, La Jolla, Calif.), both of which contain the bacterial ampicillin resistance gene and the S. cerevisiae LEU2 gene, were used for constructing a yeast genomic library and for cloning the PRO1 gene, respectively. Centromere plasmid p366 was derived from YCp50 (30) containing the centromere of chromosome IV (CEN4) by removing most of the URA3 gene as a 1.89-kb SalI/SmaI fragment and replacing it with a 2.23-kb SalI/XhoI fragment containing the LEU2 gene. Plasmid pCgHIS3 (supplied by S. Harashima) was used for disruption of the PRO1 gene. The plasmid vectors pBluescript II SK+ (Toyobo Biochemicals, Osaka, Japan) and YEp24 (3) harboring the URA3 gene were used for subcloning and for disrupting the PUT1 gene, respectively.

Culture media.The media used for growth of S. cerevisiae were SD medium (2% glucose, 0.67% Bacto Yeast Nitrogen Base without amino acids [Difco Laboratories, Detroit, Mich.]) and YPD (2% glucose, 1% Bacto Yeast Extract [Difco], 2% Bacto Peptone [Difco]). SD medium contains ammonium sulfate (0.1%) as the nitrogen source. When appropriate, required supplements were added to the media for auxotrophic strains. Yeast strains were also cultured on SD agar plates containing AZC (Sigma Chemical Co., St. Louis, Mo.). The E. coli recombinant strains were grown in Luria-Bertani medium (32) containing ampicillin (50 μg/ml). If necessary, 2% agar was added to solidify the medium.

Genetic analysis.Genetic crosses, sporulation, tetrad analysis, and complementation analysis were carried out by standard procedures as previously described (30).

Disruption of the PUT1 and PRO1 genes.The DNA fragment of the PUT1 gene encoding proline oxidase was prepared by PCR performed with genomic DNA from S. cerevisiae INVSc1 and oligonucleotide primers 5′-ACG CGT CGA CAA AGC CAG TTG TCC AGA C-3′ and 5′-GAC GAG CTC AGT CCT TCC CAC CTG ATA-3′ (the underlined sequences are the positions of SalI and SacI sites, respectively) based on the available nucleotide sequences by using a GeneAmp PCR system 2400 (PE Biosystems, Foster City, Calif.). A unique 2.6-kb amplified band for the PUT1 gene was digested with SalI and SacI and then ligated to the SalI and SacI sites of pBluescript II SK+. The plasmid harboring the PUT1 gene was designated pPUT1. Plasmid pPUT1U was then constructed by deleting the HindIII-HindlII fragment in the PUT1 gene from plasmid pPUT1 and inserting the 1.2-kb HindIII fragment containing the URA3 gene of plasmid YEp24 by ligation. For PUT1 gene disruption, the 2.0-kb SalI-SacI fragment containing put1::URA3 of pPUT1U was integrated into the PUT1 locus in strain INVSc1 by transformation. The Ura⁺ phenotype was selected, and the correct disruption event was confirmed by using the chromosomal PCR.

For PRO1 gene disruption, a DNA fragment containing the Candida glabrata HIS3 gene was amplified by PCR performed with plasmid pCgHIS3 and oligonucleotide primers 5′-TTA CTA CGT TGT CTT ATT GGG AGC CGA CGG GAT GAG GAA GGT TGT AAA ACG ACG GCC AGT-3′ and 5′-CCC CCA TCA CAT CCG GAT ACA AAT GGT CAG ACG ACC AAA TCA CAG GAA ACA GCT ATG ACC-3′ (the underlining indicates the sequences 400 bp upstream of the ATG initiation codon and 450 bp downstream of the TGA termination codon of the PRO1 gene, respectively). A unique 1.78-kb amplified band containing the C. glabrata HIS3 gene was purified and then integrated into the PRO1 locus in strain INVDput1 by transformation. The His⁺ Pro⁻ phenotype was selected, and the correct disruption was verified by chromosomal PCR analysis.

Isolation of the gene involved in intracellular l-proline accumulation.The enzymes used for DNA manipulation were obtained from Takara Shuzo (Kyoto, Japan). Conventional techniques (29) were used for S. cerevisiae genomic DNA preparation and transformation. Genomic DNA was prepared from mutant FH515 and partially digested with Sau3AI. The Sau3AI fragments larger than 5 kb were ligated into the unique BamHI site of the p366 vector. A CEN4- and LEU2-based FH515 genomic library containing more than 30,000 independent E. coli clones was introduced into S. cerevisiae INVDput1, and transformants were selected on SD medium containing 100 μg of AZC per ml. Plasmids prepared from colonies capable of growing on AZC-containing medium were then shuttled into E. coli DH5α and back into yeast strain INVDput1 to retest AZC resistance. The plasmids were sequenced to define the ends of the insert DNA by using primers YCP50+ (5′-TCC TGC TCG CTT CGC TAC TT-3′) and YCP50- (5′-AAA CAA GCG CTC ATG AGC CC-3′), which flank the BamHI site of p366, with a model 377 DNA sequencer from PE Biosystems by using the dideoxy chain termination method. The sequence obtained was compared to the yeast genome by using the BLAST program.

Construction of plasmids for expression of the PRO1 gene in S. cerevisiae.A fragment of the PRO1 gene including the probable TATA box sequences (20) and the tripartite terminator (20) was prepared by PCR performed with genomic DNA from strains MB329-17C and FH515 and oligonucleotide primers 5′-ACC CAA GCT TTG GTC AGT GGC ACA G-3′ and 5′-ACC CGA GCT CGA AGG ATT TTA ACG GAT CAC-3′ (the underlined sequences are the positions of a HindIII site and a SacI site, respectively). The unique 1.8-kb band amplified from genomic DNA of MB329-17C and FH515 was digested with HindIII and SacI and then ligated to the large fragment of pRS415 digested with HindIII and SacI to construct pRS-WTPRO1 and pRS-D154NPRO1, respectively. The nucleotide sequences of the PRO1 gene were confirmed by DNA sequencing.

Enzyme assays.To determine the activities of γ-GK (EC 2.7.2.11), γ-GPR (EC 1.2.1.41), and P5C reductase (EC 1.5.1.2), yeast cells were grown in 50 ml of SD medium containing ammonium sulfate as the nitrogen source at 30°C for 24 h with shaking. Whole-cell extracts were prepared by vortexing the cells with glass beads. Ammonium sulfate precipitates (70% saturation) of the extracts were then desalted with a PD-10 column (Amersham Pharmacia Biotech, Buckinghamshire, England) and used as enzyme sources. Protein concentrations were determined with a protein assay kit (Bio-Rad, Hercules, Calif.) by using bovine serum albumin as the standard protein.

γ-GK activity was assayed by the procedure of Smith et al. (35); the reaction mixture (final volume, 0.25 ml; pH 7.0) contained 50 mM l-glutamate, 10 mM ATP, 20 mM MgCl₂, 100 mM hydroxylamine-HCl, 50 mM Tris base, and the enzyme plus water. The reaction mixture was incubated at 37°C for 30 min to 2 h, and then the reaction was terminated by addition of 1 ml of stop solution (55 g of FeCl₃ · 6H₂O per liter, 20 g of trichloroacetic acid per liter, 21 ml of 12 N HCl per liter). Precipitated proteins were removed by centrifugation, and the absorbance at 535 nm was recorded by using a blank identical to the preparation described above except that it lacked ATP. The amount of γ-glutamyl hydroxamate was determined from the absorbance at 535 nm by comparison with a standard curve for γ-glutamyl hydroxamate (Sigma). One unit of activity was defined as the amount of enzyme required to produce 1 μmol of γ-glutamyl hydroxamate per h.

γ-GPR activity was assayed as described by Hayzer and Leisinger (12). The activity was not detectable in the forward (biosynthetic) direction due to the lability of γ-GP. To separate the P5C dehydrogenase activity which converts P5C to l-proline with the aid of NAD⁺, we measured the reverse reaction of γ-GPR by phosphate-dependent reduction of NADP⁺ with GSA (derived from equilibrium with P5C) as the substrate. The reaction mixture (final volume, 1 ml; pH 7.0) contained 2 mM P5C prepared as described previously (37), 1 mM NADP⁺, 100 mM KH₂PO₄, 50 mM imidazole base, and the enzyme plus water. The increase in the absorbance at 340 nm was determined at 30°C by using a blank identical to the preparation described above except that it lacked KH₂PO₄. One unit of activity was defined as the amount of enzyme required to produce 1 nmol of NADPH per min.

P5C reductase activity was assayed as described by Brandriss and Falvey (1). The reaction mixture (final volume, 1 ml; pH 7.0) contained 1 mM P5C, 0.4 mM NADH, 50 mM TAPS buffer, and the enzyme plus water. The decrease in the absorbance at 340 nm was determined at 30°C by using a blank identical to the preparation described above except that it lacked P5C. One unit of activity was defined as the amount of enzyme required to produce 1 nmol of NAD⁺ per min.

In addition, disruption of the PUT1 gene was confirmed by measurement of proline oxidase (EC 1.4.3.2) activity, which monitors the amount of the P5C-o-aminobenzaldehyde complex, as previously described (2). One unit of activity was defined as the amount of enzyme required to produce 1 nmol of P5C per min.

Intracellular contents of l-proline and freeze tolerance test.In a 500-ml flask, yeast cells were grown in 50 ml of SD medium at 30°C for 48 h with shaking to the stationary phase. For determination of the intracellular l-proline content, 5 ml of cell suspension (approximately 5 × 10⁸ cells) was removed, and the cells were washed twice with 0.9% NaCl and suspended in 0.5 ml of distilled water. A 1.5-ml microcentrifuge tube containing cells was transferred to a boiling water bath, and intracellular amino acids were extracted by boiling the preparation for 10 min. After centrifugation (5 min at 15,000 × g), the l-proline content in the supernatant was quantitated with an amino acid analyzer (l-8500A; Hitachi Co., Tokyo, Japan). The l-proline content was expressed as a percentage of the dry weight.

For the freeze tolerance test, 0.1 ml of cell suspension (approximately 1 × 10⁷ cells) was stored at −20°C. Under these conditions, it took about 1 h until the cells were frozen, assuming that the cooling rate was low (approximately 0.5 to 1.0°C/min). Samples of the frozen cells were thawed at room temperature for 15 min, serial dilutions in distilled water were prepared, and aliquots were plated immediately on YPD plates. After incubation at 30°C for 2 days, the survival rates were expressed as percentages, calculated as follows: (number of colonies after freezing at −20°C/(number of colonies before freezing) × 100.

Nucleotide sequence accession numbers.The GenBank accession numbers for PRO1, PUT1, and the E. coli γ-GK gene are M85293 , M18107 , and X00786 , respectively.

Analysis of the mutant with l-proline accumulation.First, AZC-resistant mutants were divided into three classes on the basis of an allelism test and growth characteristics. In particular, class I strains (exemplified by strain FH515) showed a prominent increase in cell viability compared to the viability of the parent strain after they were frozen in the medium and crossed with a wild-type strain of the opposite mating type (MB379-4C). Each heterozygous diploid grew on AZC-containing medium, indicating that the AZC-resistant growth phenotype of the mutant was dominant (Table 1).

Tetrad analyses of the diploid showed 2:2 segregation for AZC-resistant growth, indicating that the mutation was a single nuclear mutation (Table 1). The findings also suggested that l-proline accumulation was linked to AZC resistance (Table 1). Accordingly, mutant strain FH515 was used to clone the gene involved in l-proline accumulation. The members of the other classes of AZC-resistant mutants differed from the class I mutants in that they failed to produce l-proline and hence were not studied further.

Mutant that accumulated l-proline carried an allele of the PRO1 gene.To identify the mutated gene involved in l-proline accumulation in mutant FH515, we prepared an FH515-derived genome DNA library in single-copy vector p366 containing the LEU2 gene in E. coli strain DH5α. In addition, to accumulate l-proline in the host cells, a put1 disruptant that did not utilize l-proline was constructed from strain INVSc1. Proline oxidase activity was hardly detectable in put1 disruptant INVDput1 (data not shown). We then transformed INVDput1 with the genome DNA library and incubated the transformants on Leu⁻ solid medium containing 100 μg of AZC per ml at 30°C for several days. One plasmid (pYH1) was recovered from the only Leu⁺ transformant which showed the AZC-resistant phenotype. As shown in Fig. 2, when pYH1 was used to retransform strain INVDput1, all the Leu⁺ transformants were found to grow on AZC-containing medium and to accumulate an increased amount of l-proline (0.38% of dry weight), suggesting that the mutated gene resided in the plasmid.

• Open in new tab
• Download powerpoint

FIG. 2.

Growth phenotypes with AZC of S. cerevisiae strains. Strains INVSc1, INVDput1, INVDput1(pYH1), INVDput1(p366), INVDput1pro1(pRS-D154NPRO1), and INVDput1pro1(pRS-WTPRO1) were cultivated on SD medium containing AZC (50 μg/ml) at 30°C for 3 days.

The nucleotide sequence of the ends of the inserted DNA showed that pYH1 had an approximately 6.5-kb fragment of chromosome IV, containing three open reading frames (ATP5, BFR2, and PRO1 genes). Subcloning of this region indicated that the PRO1 gene encoding γ-GK in pYH1 conferred AZC resistance and l-proline accumulation to strain INVDput1, which had the wild-type PRO1 gene on the chromosome; these findings agree with the results of a genetic analysis which showed that the AZC-resistant growth phenotype of the FH515 mutant was dominant. We therefore sequenced the PRO1 gene in pYH1 and found that a single-base change from G to A at position 460 occurred, leading to replacement of aspartate with asparagine at position 154 in the γ-GK enzyme; this conclusion was based on the wild-type sequence (Fig. 3A). Similar results were obtained by direct sequencing of the PCR products from the chromosomal DNA of the wild-type (MB329-17C) and mutant (FH515) strains. The same single-amino-acid substitution at position 154 (Asp154Asn) was observed in the mutant, and no mutations were found in the 5′ upstream region and the 3′ untranslated sequence of γ-GK. Moreover, the mutation at position 154 in the S. cerevisiae γ-GK is novel because different mutations in other amino acid residues required for feedback inhibition of γ-GK by l-proline were identified in bacterial and plant genes that have been studied (Fig. 3B) (6, 7, 18, 27, 46).

• Open in new tab
• Download powerpoint

FIG. 3.

Amino acid replacement of Asp (GAC) at position 154 by Asn (AAC) in mutant FH515. (A) The numbers at the top and the bottom are the positions in the nucleotide sequence and the amino acid sequence, respectively (20). (B) Comparison of the S. cerevisiae γ-GK amino acid sequence (SC) (amino acids 107 to 157) with the E. coli γ-GK amino acid sequence (EC) (amino acids 101 to 151). Identical amino acids are indicated by vertical lines. The arrow at the top shows the mutation found in S. cerevisiae γ-GK in this study, and the arrows at the bottom indicate the mutations responsible for desensitization of l-proline feedback inhibition of γ-GK proteins in E. coli (6), S. marcescens (36), and V. aconitifolia (16).

Mutant which carried an allele of PRO1 showed an increase in both γ-GK and γ-GPR activities.Several studies of l-proline biosynthesis in bacterial cells have shown that γ-GK is the rate-limiting enzyme and is subject to feedback inhibition by l-proline (26, 35). In plants, the first steps of l-proline biosynthesis from l-glutamate are catalyzed by P5C synthetase (an EC number has not been not assigned), a bifunctional enzyme with γ-GK and γ-GPR activities (15, 44); the γ-GK activity of the P5C synthetase was also found to be sensitive to l-proline. In contrast, the PRO1 gene of S. cerevisiae was regulated by the general amino acid control system, which increased the expression of many amino acid biosynthetic genes in cells subjected to amino acid starvation conditions (20). However, feedback inhibition of γ-GK activity remains unknown. The crude extracts prepared from the wild-type (MB329-17C) and mutant (FH515) strains were assayed for γ-GK activity (Table 2). Unlike the bacterial and plant enzymes, the S. cerevisiae wild-type γ-GK showed no inhibition in the presence of 10 mM l-proline and 30% inhibition in the presence of 50 mM l-proline. These results indicate that the S. cerevisiae γ-GK, which has high sequence homology with E. coli γ-GK (level of identity, 38%) (20), was virtually insensitive to the physiological concentration of l-proline. In the mutant, the γ-GK activity was somewhat higher (approximately 25% higher) than the activity of the wild-type enzyme in the presence or absence of l-proline and was also inhibited only slightly by l-proline.

In E. coli, γ-GK and γ-GPR are thought to form a complex to provide protection to the labile γ-GP and to directly transfer the intermediate from one enzyme to the other (35). Vigna aconitifolia P5C synthetase is a fused protein with two separate catalytic domains, suggesting that the γ-GPR domain interacts with γ-GK, effecting the release of γ-GP, which exists in an enzyme-bound state. To study the influence of the γ-GK mutation on γ-GPR, we measured the γ-GPR activities in both strains. Table 2 shows that the mutant FH515 enzyme had greatly increased (approximately five- to sixfold increased) activity compared to the activity of the wild-type enzyme, even in the presence of l-proline. It was found that the S. cerevisiae γ-GPR was relatively insensitive to l-proline in a manner similar to that observed with γ-GK. On the other hand, the activity of P5C reductase, which converts P5C into l-proline, in the mutant was essentially equivalent to that in the wild-type strain (Table 2).

Expression of the mutated γ-GK in the put1-disrupted strain resulted in higher tolerance to freezing stress with l-proline accumulation.To further examine the effects of one amino acid substitution (Asp154Asn) in the γ-GK protein on the intracellular l-proline level and freeze tolerance in yeast, we constructed two centromere plasmids, one carrying the wild-type PRO1 gene and one carrying the mutated PRO1 gene (pRS-WTPRO1 and pRS-D154NPRO1, respectively), as described in Materials and Methods. These plasmids were introduced into the put1 pro1 disruptant (INVDput1pro1), and the growth phenotypes with AZC and the enzyme activities in the recombinant strains were examined. In agreement with the results obtained with mutant FH515, in strain INVDput1pro1 expressing the Asp154Asn mutant γ-GK AZC resistance was clearly evident (Fig. 2). Moreover, the activities of γ-GK and γ-GPR were apparently elevated, while the P5C reductase activity was virtually the same as that of the cells harboring the wild-type PRO1 gene (Table 2).

Table 3 shows the cellular proline levels and tolerance to freezing stress after 48 h of cultivation of the transformants in liquid SD medium. The wild-type (MB329-17C) and the mutant (FH515) strains were also cultivated as controls for l-proline accumulation and freeze tolerance. It was found that there was significant accumulation of l-proline in strain INVDput1pro1 expressing the Asp154Asn mutant γ-GK, as well as in FH515, whereas no intracellular l-proline was detected in strain INVDput1pro1 harboring the wild-type PRO1 gene and in strain MB329-17C. These results indicate that the elevated activities of γ-GK and γ-GPR, both of which are known to catalyze the first two steps of l-proline synthesis from l-glutamate, are important for increasing the level of l-proline in S. cerevisiae.

Strain FH515 exhibited increased cell viability compared with that of strain MB329-17C when the cell suspensions were frozen at −20°C for 1 to 3 days (Table 3). Similarly, cells expressing the mutant γ-GK showed approximately 14- to 33-fold increases in survival rates after freezing compared with the survival rates of cells harboring the wild-type PRO1 gene (Table 3). Prolonged storage of the cells at −20°C for up to 1 week caused a gradual loss of freeze tolerance, although a significant cryoprotective effect was observed in the mutant γ-GK-expressing cells (data not shown). In addition, the amount of accumulated l-proline in the cells remained relatively stable after cultivation in all cases due to the lower level of l-proline metabolism caused by mutation and disruption of the proline oxidase gene. These results indicate that a point mutation in the PRO1 gene, which elevated the activities of both γ-GK and γ-GPR, resulted in a higher level of l-proline accumulation and freeze tolerance in S. cerevisiae.