Yeast strains

The Saccharomyces cerevisiae strain used was W303 in this study. All acquired strains from original W303 strain were listed in S1 Table. For the DNA manipulation, Escherichia coli, DH5α strain, was used.

Cell culture

Yeast strains were pre-cultured in YPD medium (2% Glucose, 2% bactopeptone, and 1% yeast extract) and then shifted to main culture in YPD medium. Strains harboring plasmids were cultured in selective synthetic complete (SC) media according to the corresponding auxotrophic marker: SC-Ura medium for YCplac33 and YEplac195 plasmids, and SC-Leu medium for the YEp13 plasmid. The strains carrying a YCplac33-YGP1 or YCplac33-YGP1–3HA plasmid were pre-cultured in SC-Ura medium and then transferred to the main culture in YPD medium. Since the diauxic shift response was not clearly observed in SC medium, YPD medium was used for the main culture. For investigating non-fermentative growth, cells were incubated on a YPGlylac plate (3% Glycerol, 2% Sodium Lactate, 2% bactopeptone, and 1% yeast extract). All cell cultures were performed at 28 degrees. Samples were collected at the time set for each experiment.

Measurement of Glucose concentration

Glucose concentrations in the culture medium were measured using a Glucose (GO) Assay Kit (Sigma-Aldrich). Absorbance at 540 nm was measured with a Varioskan LUX microplate reader.

Gene deletion

All gene deletions were performed as previously described [24–26]. All primers used were listed in S2 Table. The amplified fragments by PCR were transformed into diploid cells, and transformants were selected on SC medium lacking each selective amino acid.

Plasmids construction

The plasmids used in this study were listed in S3 Table. All plasmids were constructed by inserting the indicated fragments between the EcoRI and SalI sites of the YCplac33 plasmid. YCplac33-YGP1 plasmid was constructed by inserting the YGP1promoter-ORF-3’UTR fragment. The YCplac33-MCM2promoter-GFP-ADH1 3’UTR and YCplac33-MCM2promoter-GFP-YGP1 3’UTR plasmids were constructed by inserting fragments containing the MCM2 promoter, GFP, and the corresponding 3′ UTR. For the construction of a YCplac33-YGP1–3HA plasmid, the YGP1promoter-ORF, 3HA, and YGP1 3’UTR fragments were inserted. A series of YGP1 promoter deletion plasmid was constructed by inserting two fragments comprising an upstream promoter region and a downstream promoter-YGP1 ORF-3’UTR fragment. YCplac33-YGP1-STRE1del and YCplac33-YGP1-STRE2del plasmids were constructed by inserting two fragments (upstream or downstream fragments from each STRE). The primers used for the amplification of fragments were listed in S4 Table.

RNA extraction and quantification

Total RNA samples were extracted utilizing ISOGEN reagent (Nippon Gene), followed by removal of genome DNA and reverse transcription using PrimeScript™ RT reagent Kit with gDNA Eraser (Perfect Real Time) (Takara). The expression of cDNA was quantified RT-qPCR using QuantStudio 5 for Fig 1 and QuantStudio 7 (Thermo Fisher Scientific) for other figures with SYBR Premix Ex Taq (Takara). Relative expression levels were calculated using the ΔΔCt method, with SCR1 used as the endogenous control. Rq values were expressed as fold changes relative to the indicated reference sample. The primers used for the amplification were listed in S5 Table. The YGP1–3HA mRNA was differentiated from endogenous YGP1 mRNA by using a set of primers from 3HA-tag sequences. All graphs were created by Microsoft Excel.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. YGP1 expression is highly induced at diauxic shift.

(A, B) The fold change of YGP1 (A) and HSP26 mRNA levels (B) in the wild-type and puf5Δ mutant along a time course relative to the wild-type 4 hours value. The expression pattern was confirmed using two biological replicates, each having three technical replicates. The data show the average Rq values ± SEM (n = 3 technical replicates from one representative biological replicate). (C) Yeast cells were pre-cultured and then transferred to YPD medium for the main culture. Glucose concentrations were measured at the indicated time points. For each strain, glucose levels were measured six times at each time point, and the mean value was taken as the glucose concentration. The graph shows the mean glucose concentrations ± SEM (n = 3 biological replicates each having n = 6 technical replicates).

https://doi.org/10.1371/journal.pone.0355681.g001

Protein extraction and quantification

Protein extraction was performed as previously described [27]. Protein samples were loaded onto a 12% SDS-PAGE gel for Ygp1–3HA and a 10% SDS-PAGE gel for Pgk1. The gel was transferred to a PDVF membrane (Millipore), and then membrane was reacted with the primary antibody, the anti-HA monoclonal antibody HA11 or anti-Pgk1 antibody, at 4°C overnight. Reacted with the secondary antibody, the anti-mouse IRDye® 800CW secondary antibodies and IRDye® 680RD secondary antibodies (LI-COR), the visualization and quantification were performed using ODYSSEY CLx (LI-COR). The acquired relative signal intensities of Ygp1–3HA protein were normalized by those of Pgk1, an endogenous control. All graphs were created by Microsoft Excel.

Cell cycle synchronization

Cell cycle synchronization was performed by α-factor induced G1-phase arrest and releasing as previously described [28]. The synthesized α-factor was purchased from GenScript and dissolved into sterilized water to 20 mM. Each bar1Δ background strain was exposed to 30 nM α-factor and reacted for 2 hours in YPD. After washing with YPD, samples were collected every 10 minutes.

Measurement of cell viability

Cells were cultured in SC medium until an appropriate time and diluted to final OD₆₀₀ = 2. Cell viability was counted using NucleoCounter® YC-100 (M&S TechnoSystems).

Statistical analysis

Biological replicates were defined as independent cultures initiated separately. Technical replicates were defined as repeated measurements performed on the same biological sample, including independent qPCR reactions and Western blot quantifications. For each experiment, analyses were performed using two independent biological replicates, each consisting of three technical replicates. Data in the graphs are presented as the mean ± SEM of the three technical replicates from a single representative biological replicate, unless stated otherwise. Exceptions to this replication scheme are as follows: for Figs 1C and 2G, three biological replicates were analyzed; for Figs 5B-C, 9B, and S5 Figs, a single biological replicate was used as this experiment was conducted for screening purposes. Data shown represent he mean ± SEM of the three technical replicates. Statistical significance was determined using a Welch t-test for comparisons between two samples, and a Kruskal–Wallis test followed by a Dwass–Steel–Critchlow–Fligner test for comparisons between more than three samples, considering heteroscedasticity. p < 0.05 was considered as statistically significant. For the comparison of three or more groups with equal variances, homogeneity of variances was first assessed using Bartlett’s test, followed by one-way ANOVA. Statistical analysis of Fig 9E was performed using aligned rank transform (ART)–based multifactor ANOVA, with multiple comparisons corrected using Holm’s method.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. YGP1 expression is regulated by Puf5 through its promoter.

(A) The fold change of YGP1 mRNA level in the puf5Δ mutant relative to that in wild-type strain. The expression pattern was confirmed using three biological replicates, each having three technical replicates. The data show the average Rq values ± SEM (n = 3 technical replicates from one representative biological replicate). **p < 0.01 is indicative of statistical significance. (B) The fold change of YGP1 mRNA level in the puf5Δ [YEplac195-PUF5] strain relative to that in puf5Δ [YEplac195 vector]. The expression pattern was confirmed by three biological replicates, each having three technical replicates. The data show the average Rq values ± SEM (n = 3 technical replicates from one representative biological replicate). **p < 0.01 is indicative of statistical significance. (C) The scheme of GFP reporter fragment on a YCplac33 plasmid. The mesh bar represents YGP1 promoter, the white bar does GFP ORF, and the black-dot bar does ADH1 3’UTR. (D) The fold change of GFP mRNA level in the puf5Δ mutant relative to that in wild-type strain harboring YCplac33-GFP-reporter plasmid. The expression pattern was confirmed by two biological replicates, each having three technical replicates. The data show the average Rq values ± SEM (n = 3 technical replicates from one representative biological replicate). **p < 0.01 is indicative of statistical significance. (E) The fold change of GFP mRNA level in the puf5Δ [YCplac33-GFP-reporter] [YEplac13-PUF5] strain relative to that in puf5Δ [YCplac33-GFP-reporter] [YEplac13 vector]. The expression pattern was confirmed by two biological replicates, each having three technical replicates. The data show the average Rq values ± SEM (n = 3 technical replicates from one representative biological replicate). **p < 0.01 is indicative of statistical significance. (F) The scheme of GFP reporter fragment on a YCplac33 plasmid. The mesh bar represents MCM2 promoter, the white bar does GFP ORF, and the black-dot bar does ADH1 3’UTR or YGP1 3’UTR. (G) The fold change of GFP mRNA level in the WT and puf5Δ mutant harboring a [YCplac33 MCM2promoter-GFP-ADH1 3’UTR] or [YCplac33 MCM2promoter-GFP-YGP1 3’UTR] plasmid relative to that in WT [YCplac33 MCM2promoter-GFP-ADH1 3’UTR]. The data show the average Rq values ± SEM of three biological replicates each having three technical replicates. ns is indicative of no statistical significance (p < 0.05).

https://doi.org/10.1371/journal.pone.0355681.g002

YGP1 expression is controlled by dual regulatory pathways in response to glucose availability

Glucose availability is a major determinant of yeast growth, and yeast cells dynamically regulate gene expression in response to environmental glucose levels to modulate cell-cycle progression, metabolism, and stress responses [1–6]. YGP1 is one of the genes whose expression is dynamically regulated under such conditions. Ygp1 is a secreted glycoprotein that is implicated in cell wall remodeling and potentially in stress responses triggered by environmental conditions such as glucose availability and acidic stress. Previous studies have reported that YGP1 expression is strongly induced under nutrient starvation conditions, including glucose deprivation [14,15]. Based on these observations, we hypothesized that the induction of YGP1 expression may contribute to the cellular responses to glucose starvation. In this study, we investigated the regulatory mechanisms governing YGP1 expression. We first examined the YGP1 expression pattern along a time course in YPD medium. YGP1 mRNA level peaked sharply at 12 hours in the wild-type and declined thereafter (Fig 1A). The expression of HSP26, a stationary phase marker [29], was induced at 12 hours and remained high for 3–4 days in the wild-type (Fig 1B). Investigating the glucose concentrations in the YPD medium, the wild-type strain exhausted glucose at 12h (Fig 1C), which is consistent with our previous report [30]. These results revealed that the expression of YGP1, which remained low during the glucose-replete logarithmic phase (4–9 h), was strongly induced at diauxic shift, and subsequently declined in the stationary phase. Given the pronounced glucose-dependent changes in expression, we analyzed the regulatory machinery of YGP1 expression. Our previous microarray analysis under logarithmic growth conditions revealed that YGP1 expression was 66% decreased in the puf5Δ mutant relative to the wild type [31] (Table 1). Puf5 encodes an RNA-binding protein regulating mRNA stability or translational efficiency through binding to 3’UTR region of its target mRNAs [32,33]. Accumulating evidence indicates that Puf5 targets LRG1 mRNA, which encodes a Rho1 GAP involved in the yeast cell wall integrity pathway [34–36], thereby repressing its expression [30,37–39]. This effect is more evident in the stationary phase than in log phase [30], and puf5Δ mutants exhibit a shortened chronological life span associated with aberrant CWI signaling [37]. In addition, previous studies have demonstrated that Puf5 contributes to glucose starvation-induced recruitment to P-bodies and their degradation of cell wall-related mRNAs [40]. These observations imply that Puf5 functions in a glucose-responsive manner. Confirming the microarray result by qPCR, we first examined the time-course YGP1 mRNA level in the puf5Δ mutant. YGP1 expression was decreased in the puf5Δ mutant compared to the wild-type during the log phase (4–9 h) (Fig 1A), while it reached similar levels at 12 hours in the two strains (Fig 1A). Expression of HSP26, a marker of the stationary phase, was comparable between the puf5Δ mutant and wild-type cells at 12h (Fig 1B). Consistently, measurement of glucose levels revealed that the puf5Δ mutant consumed glucose at a similar rate to the wild-type (Fig 1C). These observations indicate that Puf5 does not contribute to the induction of YGP1 expression at the diauxic shift but rather plays a specific role in the regulation of YGP1 during logarithmic growth. This suggests the existence of dual regulatory pathways controlling YGP1 expression under glucose-rich conditions and upon glucose depletion.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Microarray results of YGP1 in the wild-type and puf5Δ mutant.

https://doi.org/10.1371/journal.pone.0355681.t001

YGP1 expression is regulated by an RNA-binding protein Puf5 during the log phase

Aiming to clarify the mechanisms of the two distinct regulatory pathways of YGP1 expression, we first focused on the glucose-rich mid-log phase (4h) and investigated how Puf5 positively regulates YGP1 expression. Consistent with Fig 1A, YGP1 expression was significantly decreased in the puf5Δ mutant (Fig 2A). We also found that the multicopy-PUF5 increased the YGP1 mRNA level (Fig 2B). To analyze the responsible region for the Puf5-mediated regulation, we constructed YGP1 promoter-GFP reporter on a YCplac33 plasmid (Fig 2C) and investigated the expression of the GFP reporter. GFP expression showed similar expression pattern to YGP1 expression: its expression was decreased in the puf5Δ mutant and increased by multicopy-PUF5 (Figs 2D and E). While Puf5 influenced less efficiently on the YGP1 promoter-GFP-reporter than on endogenous YGP1 (Figs 2A, B, D, and E), the differences were significant (Figs 2D and E). To assess the contribution of the YGP1 3′ UTR to regulation mediated by Puf5, we introduced the YCplac33-MCM2promoter-GFP-ADH1 3’UTR and YCplac33-MCM2promoter-GFP-YGP1 3’UTR plasmids into wild-type and puf5Δ mutant and examined the GFP reporter expression levels. The expression levels of both the GFP-ADH1 3’UTR reporter and the GFP-YGP1 3’UTR reporter showed no significant difference between wild-type and puf5Δ mutant (Fig 2G). These results indicate that Puf5 positively regulates the expression of YGP1 via its promoter, rather than its 3’UTR.

YGP1 expression is cyclically regulated during M-phase in a Puf5-mediated manner

Previously, YGP1 expression was reported to be regulated in a cell cycle-specific manner [16]: YGP1 was allocated to the Mcm1-cluster gene which was induced from M to G1-phase during cell cycle. Given that Puf5 tunes cell cycle-specific expression of cyclin B genes CLB1 and CLB2 [31,41], we next examined whether the Puf5-mediated positive regulation is involved in the induction of cell cycle-specific YGP1 expression. Cells were synchronized in the G1-phase using α-factor in a bar1Δ mutant background where the protease inactivating α-factor is absent. In the bar1Δ puf5Δ double mutant, the expression of the S-phase marker RNR1 was induced upon similar timing to that in the bar1Δ mutant (Fig 3A and S1 Fig A). However, the peak expression of the late M-phase marker SIC1 was slightly delayed in the double mutant, suggesting a G2/M-phase delay in the bar1Δ puf5Δ double mutant (Fig 3B and S1 Fig B). The expression of YGP1 peaked after RNR1 but before SIC1 in the bar1Δ mutant, suggesting that YGP1 is specifically expressed during the M-phase (Fig 3C and S1 Fig C). Notably, this YGP1 induction was not observed in the bar1Δ puf5Δ double mutant (Fig 3C and S1 Fig C), indicating that Puf5 is required for the M-phase-specific induction of YGP1 expression.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. The rhythmic expression of YGP1 in M-phase is induced by Puf5.

The fold change of RNR1 (A), SIC1 (B), and YGP1 (C) mRNA levels in the bar1Δ and bar1Δ puf5Δ mutants relative to the bar1Δ 0-min value at the indicated time points after release from G1-phase arrest. Analysis was performed using two biological replicates, each having three technical replicates. Data shown here are the representative Rq values from three technical replicates for one of the biological replicates. The results for the other biological replicate are shown in S1 Figs.

https://doi.org/10.1371/journal.pone.0355681.g003

A discrete promoter segment mediates Puf5‑dependent regulation of YGP1

Given that YGP1 expression is regulated by Puf5 via its promoter (Figs 2D and E), we next aimed to clarify the detailed mechanism of this regulation. Since Puf5 is an RNA-binding protein which regulates its target mRNAs through binding to 3’UTR [32,33], we hypothesized that Puf5 indirectly controls YGP1 promoter via regulating its transcriptional factor-coding mRNAs. We have previously clarified that Puf5 positively regulates the expression of CLB1 and CLB2 through the negative regulation of its target IXR1 mRNA, which encodes a transcriptional repressor [31,41]. To assess whether the regulation of YGP1 expression is also mediated by Ixr1, we measured the YGP1 mRNA level in the wild-type, puf5Δ, ixr1Δ, and puf5Δ ixr1Δ mutants in an asynchronous culture. Interestingly, while the ixr1Δ mutant showed significantly higher YGP1 mRNA levels than the wild-type (Fig 4A, WT vs ixr1∆), the ixr1Δ mutation only trivially restored the decreased YGP1 expression caused by the puf5Δ mutation (Fig 4A, puf5∆ vs puf5∆ ixr1∆). In addition, the ixr1Δ mutation also failed to restore the cell cycle-specific induction of YGP1 in the bar1Δ puf5Δ background (Fig 4B and S2 Fig). These results suggest that Puf5 regulates YGP1 expression independently of Ixr1, while Ixr1 negatively regulates YGP1 expression.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. YGP1 expression is regulated by Puf5 independently of Ixr1.

(A) The fold change of YGP1 mRNA levels in the puf5Δ, ixr1Δ, and puf5Δ ixr1Δ mutants relative to that in wild-type strain. The expression pattern was confirmed by two biological replicates, each having three technical replicates. The data show the average Rq values ± SEM (n = 3 technical replicates from one representative biological replicate). *p < 0.05 and **p < 0.01 are indicative of statistical significance. (B) The fold change of YGP1 mRNA levels in the bar1Δ, bar1Δ puf5Δ, and bar1Δ puf5Δ ixr1Δ mutants relative to the bar1Δ 0-min value at the indicated time points after release from G1-phase arrest. Analysis was performed using two biological replicates, each having three technical replicates. Data shown here are the representative Rq values from three technical replicates for one of the biological replicates. The results for the other biological replicate are shown in S2 Fig.

https://doi.org/10.1371/journal.pone.0355681.g004

To clarify the regulatory mechanism by Puf5, we next aimed to map YGP1 promoter elements that mediate the action of Puf5. We constructed a YCplac33-YGP1 plasmid harboring a 960-nucleotide region upstream of the start codon as promoter, and generated a series of deletion constructs, each lacking a 60-nucleotide segment in the YGP1 promoter. In this process, the regions including each transcriptional start site (TSS) of two YGP1 transcripts were excluded (Fig 5A). We then introduced these plasmids into the ygp1Δ mutant and the puf5Δ ygp1Δ double mutant to assess the impact of each promoter region on exogenous YGP1 mRNA levels (Fig 5B). The scores in Figure 5B show the fold changes of YGP1 expression in the puf5Δ ygp1Δ double mutant relative to the ygp1Δ mutant, along with the corresponding p-values. The relative expression levels of YGP1 in each strain were also manifested by a bar graph (Fig 5C). The expression of full-length YGP1 was significantly decreased by the puf5Δ mutation (Figs 5B and C, Full). Interestingly, this decrease was canceled only in the del7 (−600 to −540) construct among 14 deletion constructs tested. These results indicate that the del7 region (−600 to −540) contains a critical segment for the Puf5-mediated control.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. A discrete promoter segment mediates Puf5‑dependent regulation of YGP1.

(A) A scheme of two YGP1 transcripts. TSS1 corresponds to the transcription start site (TSS) of transcript 1, and TSS2 to the TSS of transcript 2. (B) The scheme of a series of YGP1 promoter deletions. Every 60-nucleotide region was deleted from a YCplac33-YGP1 plasmid except for the region containing transcriptional start sites (TSSs). After introducing a full-length YCplac33-YGP1 plasmid or one of the deletion plasmids in the ygp1Δ and ygp1Δ puf5Δ strains, the strains were pre-cultured in SC-Ura medium followed by the 2-hour exposure to small amount of YPD and thereafter main culture in YPD medium until mid-log phase. The YGP1 expression was quantified by RT-qPCR. The right score shows the fold change of the YGP1 mRNA level in the ygp1Δ puf5Δ double mutant compared to that in the ygp1Δmutant (n = 3 technical replicates from one representative biological replicate), along with the corresponding p-values. (♰) p < 0.05 and (♰♰) p < 0.01 are indicative of statistical significance. ns, no statistical significance (p < 0.05). (C) The fold changes of YGP1 in the ygp1Δ and ygp1Δ puf5Δ strains harboring a full-length YCplac33-YGP1 plasmid or one of the deletion plasmids. The data show average Rq values ± SEM (n = 3 technical replicates from one representative biological replicate) relative to that in the ygp1Δ [YCplac33-YGP1 Full] sample.

https://doi.org/10.1371/journal.pone.0355681.g005

To identify the responsible factor which directly acts on the del7 region (−600 to −540), we searched YEASTRACT+ database, which catalogs transcriptional regulatory interactions in yeast (https://yeastract.com/). Analysis of the 60-nucleotide region (−600 to −540) revealed several transcription factors with binding motifs within this segment (Table 2). Comparing these factors with those having binding evidence to the YGP1 promoter (Table 3), we focused on the transcription activator Haa1. Haa1 is known to mediate the weak acid stress responses [20,21] through regulating the expression of 80% of acid stress-responsive genes [42]. Haa1 is also reported to localize the nucleus and positively regulate YGP1 expression [20,22]. To confirm the role of Haa1 in YGP1 expression, we investigated the YGP1 mRNA level in the wild-type, puf5Δ, haa1Δ, and puf5Δ haa1Δmutants. YGP1 expression was approximately 40% lower in the haa1Δ mutant than in the wild-type and notably showed further decrease in the puf5Δ haa1Δ double mutant (Fig 6A), indicating that Puf5 regulates YGP1 expression independently of Haa1. Then we next examined the Ygp1 protein levels using the exogenously expressed YGP1–3HA from a YCplac33 plasmid (Fig 6B). In this experiment, to minimize differences in culture conditions, the strains were pre-cultured in SC-Ura medium, followed by main culture in YPD medium. Even when main culture was performed in YPD medium, the plasmid carriage rate for each strain exceeded 95%, and no significant differences were observed between strains (Table 4). The Ygp1–3HA protein level was also significantly decreased in the puf5Δ and haa1Δ mutants compared to the wild-type (Fig 6C). In the puf5Δ haa1Δ double mutant, although the difference was not significant, the protein level was lower than the puf5Δ or haa1Δ single mutant (Fig 6C). While these data suggest that Puf5 functions independently of Haa1, the binding evidence of Puf5 to HAA1 mRNA in the previous genome-wide mapping of Puf5-mRNA interaction [43] implied the possibility that Haa1 acts downstream of Puf5. Indeed, HAA1 mRNA levels was ~ 20% decreased in the puf5Δ mutant (Fig 6D) and ~50% increased by the multicopy-PUF5 (Fig 6E). Taken together, it is suggested that Puf5 and Haa1 positively regulate the YGP1 expression through the 60-nucleotide promoter region (−600 to −540) and that the Puf5-mediated regulation is partly mediated by Haa1.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Transcriptional factors harboring binding motif within del7 region.

https://doi.org/10.1371/journal.pone.0355681.t002

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Transcriptional factors with binding evidence to YGP1 promoter.

https://doi.org/10.1371/journal.pone.0355681.t003

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Plasmid carriage rates in yeast strains harboring the YCplac33-YGP1-3HA plasmid after 24 hours of incubation in non-selective medium.

https://doi.org/10.1371/journal.pone.0355681.t004

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Transcriptional activator Haa1 positively regulates YGP1 expression.

(A) The fold changes of YGP1 mRNA levels in the puf5Δ, haa1Δ, and puf5Δ haa1Δ mutants relative to that in wild-type strain. The expression pattern was confirmed by two biological replicates, each having three technical replicates. The data show the average Rq values ± SEM (n = 5 technical replicates from one representative biological replicate). *p < 0.05 is indicative of statistical significance. (B) A scheme of YGP1–3HA fragment on a YCplac33. (C) The Ygp1–3HA protein levels expressed from a YCplac33 plasmid in the wild-type, puf5Δ, haa1Δ, and puf5Δ haa1Δ mutants. The image shows Ygp1–3HA and Pgk1 (loading control). The graph shows the fold changes of Ygp1–3HA protein levels in the puf5Δ, haa1Δ, and puf5Δ haa1Δ mutants relative to that in wild-type strain. The expression pattern was confirmed by two biological replicates, each having three technical replicates. The data show the average fold change ± SEM (n = 6 technical replicates from one representative biological replicate). *p < 0.05 are indicative of statistical significance. ns, no statistical significance (p < 0.05). (D) The fold change of HAA1 mRNA level in the puf5Δ mutant relative to that in wild-type strain. The expression pattern was confirmed by three biological replicates, each having three technical replicates. The data show the average Rq values ± SEM (n = 4 technical replicates). **p < 0.01 is indicative of statistical significance. (E) The fold change of HAA1 mRNA level in the puf5Δ [YEplac195-PUF5] strain relative to that in puf5Δ [YEplac195 vector]. The expression pattern was confirmed by two biological replicates, each having three technical replicates. The data show the average Rq values ± SEM (n = 3 technical replicates from one representative biological replicate). **p < 0.01 is indicative of statistical significance.

https://doi.org/10.1371/journal.pone.0355681.g006

Haa1 partially contributes to the periodic expression of YGP1 during the cell cycle

To test the role of Haa1 in the M-phase-specific induction of YGP1 expression, we synchronized the bar1Δ haa1Δ double mutant and examined YGP1 mRNA level. In the bar1Δ haa1Δ double mutant, the expression of the S-phase marker RNR1 and the late M-phase marker SIC1 peaked at similar time points to those observed in the bar1Δ mutant (Figs 7A and B, and S3 Figs A and B), suggesting that Haa1 has no major effect on cell cycle progression. While YGP1 induction was observed in the bar1Δ haa1Δ double mutant, the peak appeared to be modestly reduced and slightly delayed compared with that in the bar1Δ mutant (Fig 7C and S3 Fig C). Together with the results from asynchronous culture (Figs 6A and C), it is suggested that Haa1 partially contributes the induction of the M‑phase-specific induction of YGP1 expression.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. Haa1 partly regulates the expression of YGP1 during the cell cycle.

The fold change of RNR1 (A), SIC1 (B), and YGP1 (C) mRNA levels in the bar1Δ and bar1Δ haa1Δ mutants relative to the bar1Δ 0-min value at the indicated time points after release from G1-phase arrest. Analysis was performed using two biological replicates each having three technical replicates. Data shown here are the representative Rq values from three technical replicates for one of the biological replicates. The results for the other biological replicate are shown in S3 Figs.

https://doi.org/10.1371/journal.pone.0355681.g007

YGP1 expression is highly induced at the diauxic shift

Given that YGP1 expression is regulated by two distinct mechanisms depending on glucose availability, we next examined the regulatory mechanism governing YGP1 expression at the diauxic shift. YGP1 expression is strongly induced at the time of the diauxic shift (Figs 1A–C). To assess whether these expression patterns are reflected in those of Ygp1 protein levels, we exogenously expressed YGP1–3HA from a YCplac33 plasmid (Fig 6B). Since the plasmid retention rates remained above 95% after 24 h of cultivation in non-selective YPD medium (Table 4), main cultures were performed in YPD medium. YGP1–3HA mRNA levels, like endogenous YGP1 mRNA levels, rose at 12 hours and fell by 24 hours in wild-type (Fig 8A). Ygp1–3HA protein levels also showed a similar pattern (Fig 8B), indicating the diauxic shift-responsive induction of YGP1 is reflected to the protein level. Surprisingly, in this experiment, both YGP1 mRNA and protein levels at 12 hours were attenuated in the puf5Δ mutant compared to those in the wild-type, while the relative induction at 12 hours (12h/4h) was maintained (Fig 8A and B). The reason for the difference in expression patterns between endogenous YGP1 and YGP1–3HA was unclear. However, considering that the plasmid-carrying strain was pre-cultured in SC-Ura medium before the main culture in YPD medium, the impaired response to the diauxic shift in the puf5Δ mutant may be caused by the prior exposure to SC medium. Indeed, the stationary phase marker HSP26 induction level at 12 hours was also reduced in the puf5Δ mutant harboring a YGP1–3HA plasmid (Fig 8C).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 8. YGP1 expression is induced at the diauxic shift.

(A) The fold changes of YGP1–3HA mRNA expressed from a YCplac33 plasmid in the wild-type and puf5Δ mutant along a time course relative to the wild-type 4 hours value. The expression pattern was confirmed by two biological replicates, each having three technical replicates. The data show the average Rq values ± SEM (n = 3 technical replicates from one representative biological replicate). (B) The Ygp1–3HA protein levels expressed from a YCplac33 plasmid in the wild-type and puf5Δ mutant. The image shows Ygp1–3HA and Pgk1 (loading control). The expression pattern was confirmed by two biological replicates, each having three technical replicates. The graph shows the mean ± SEM of the fold changes of Ygp1–3HA protein levels relative to that in wild-type 4 hours (n = 4 technical replicates from one representative biological replicate). (C) The fold change of HSP26 mRNA levels in the wild-type and puf5Δ mutant harboring a Ycplac33 YGP1–3HA plasmid along a time course relative to the wild-type 4 hours value. The expression pattern was confirmed by two biological replicates, each having three technical replicates. The data show the average Rq values ± SEM (n = 3 technical replicates from one representative biological replicate).

https://doi.org/10.1371/journal.pone.0355681.g008

We next examined the potential involvement of Haa1 in the YGP1 induction at the diauxic shift. Time-course analysis revealed that endogenous YGP1 expression in the haa1Δ mutant was comparable to that in the wild-type (S4 Fig A), indicating Haa1 is not involved in this induction of YGP1 at the diauxic shift. Interestingly, when examined YGP1–3HA mRNA and Ygp1–3HA protein levels expressed from a plasmid, both mRNA and protein levels were slightly decreased in the haa1Δ mutant (S4 Figs B and C). In this experiment as well, since plasmid retention exceeded 95%, with no significant differences among strains (Table 4), the main cultures were performed in YPD medium. In addition, the expression of HSP26, a stationary phase marker, was mildly decreased in the haa1Δ mutant harboring a YGP1–3HA plasmid at 12h, while it was decreased more markedly in the puf5Δ mutant (S4 Fig D). These results suggest that Puf5 can function independently of Haa1, but the machinery involving Puf5 and Haa1 appears to positively regulate the response to the diauxic shift including YGP1 induction at least in the plasmid-carrying strains used in these experiments.

Transcriptional factors Msn2/Msn4 are responsible for the induction of YGP1 expression at the diauxic shift

To specify the responsible factor which induces YGP1 expression at the diauxic shift, we employed YGP1 promoter-deletion constructs again and measured mRNA levels at mid-log phase and the diauxic shift. Since no apparent plasmid loss was observed in both ygp1Δ and puf5Δ ygp1Δ strains after 24 h in non-selective YPD medium (Table 5), strains were pre-cultured in SC-Ura medium followed by the main culture in YPD medium. In S5 Fig A, relative fold change of YGP1 mRNA level (diauxic shift/ log phase) and the respective p-values were listed. In the wild-type strain, YGP1 expression was induced approximately 10 times at the diauxic shift (S5 Figs A and B, Full). This induction level was reduced in all deletion constructs but was particularly decreased to less than 1.5 times in the deletion constructs 7–11 (S5 Figs A and B). Table 6 lists transcription factors with predicted binding motifs within the regions (−600 to −300) corresponding to the deletion constructs 7–11. Among them, we focused on the stress-responsive transcription factors, Msn2 and Msn4, which harbor their binding element in the del 9 region (−480 to −420). The transcription factors Msn2 and Msn4 play central roles in regulating the general stress responses. Upon exposure to nutrient deprivation, organic acids, or heat shock, they translocate to the nucleus and activate transcription through binding to STREs in target gene promoters [7–10]. They are recognized as major regulators of the environmental stress response (ESR) [11] and are particularly responsive to glucose starvation, which induces their rapid nuclear accumulation [10,12,13]. The YGP1 promoter contains two STREs, STRE1 (upstream, −729 to −722) and STRE2 (downstream, −434 to −430) (Fig 9A), suggesting that Msn2 and Msn4 may regulate YGP1 under glucose starvation. To test this hypothesis, we measured endogenous YGP1 mRNA levels in the wild-type, msn2Δ, msn4Δ, and msn2Δ msn4Δ mutants over time. The YGP1 mRNA level at 12 hours was reduced in the msn2Δ and msn4Δ mutants and nearly abolished in the msn2Δ msn4Δ double mutant (Fig 9B). In this experiment, YGP1 expression started to increase from 9h (Fig 9B), exhibiting a slightly distinct pattern from that shown in Fig 1A, where expression levels remained relatively low at this time point. Nevertheless, the peak expression of YGP1 expression in wild-type cells was consistently detected at 12 h. As shown in Fig 1C, glucose levels at 9 h were close to depletion, suggesting that subtle differences in glucose concentrations during medium preparation may have caused small variations in the timing of glucose exhaustion. The YGP1–3HA mRNA and Ygp1–3HA protein levels expressed from a YCplac33 plasmids also showed similar expression patterns in the wild-type and msn2Δ msn4Δ double mutant (Figs 9C and D). Consistent with Fig 8A and 8B, the puf5Δ mutant showed an overall decrease in the YGP1–3HA mRNA and Ygp1–3HA protein levels (Figs 9C and D). In this experiment, we also confirmed that the msn2Δ msn4Δ double mutant exhibited the comparable plasmid carriage rate to that observed in wild-type and puf5Δ strains (Table 7). Deletion analysis of the STREs on a YCplac33-YGP1 plasmid showed partial loss of induction with STRE1 deletion and complete loss with STRE2 deletion, while the basal expression was retained in these deletions (Fig 9E). We conclude that Msn2 and Msn4 are the principal drivers for YGP1 expression by glucose-starvation, acting primarily through STRE2 (−434 to −430).

Download:

• PNG
  larger image
• TIFF
  original image

Table 5. Plasmid carriage rates in yeast strains harboring the YCplac33-YGP1 plasmid after 24 hours of incubation in non-selective medium.

https://doi.org/10.1371/journal.pone.0355681.t005

Download:

• PNG
  larger image
• TIFF
  original image

Table 6. Transcriptional factors harboring binding motif within each deletion region.

https://doi.org/10.1371/journal.pone.0355681.t006

Download:

• PNG
  larger image
• TIFF
  original image

Table 7. Plasmid carriage rates in yeast strains harboring the YCplac33-YGP1-3HA plasmid after 24 hours of incubation in non-selective medium.

https://doi.org/10.1371/journal.pone.0355681.t007

Download:

• PNG
  larger image
• TIFF
  original image

Fig 9. The induction of YGP1 expression is triggered by the transcriptional activators Msn2 and Msn4.

(A) A scheme of YGP1 promoter. Two stress response elements (STREs) exist within YGP1 promoter. The upstream one is labeled as STRE1, and the downstream one is as STRE2. (B) The fold change of YGP1 mRNA levels in the wild-type, msn2Δ, msn4Δ, and msn2Δ msn4Δ mutants along a time course relative to the wild-type 4-hours value. The data show the average Rq values ± SEM (n = 3 technical replicates from one representative biological replicate). (C) The fold change of YGP1–3HA mRNA levels expressed from a YCplac33 plasmid in the wild-type, puf5Δ, and msn2Δ msn4Δ mutants along a time course relative to the wild-type 4-hours value. The expression pattern was confirmed by two biological replicates, each having three technical replicates. The data show the average Rq values ± SEM (n = 3 technical replicates from one representative biological replicate). (D) The Ygp1–3HA protein levels expressed from a YCplac33 plasmid in the wild-type, puf5Δ, and msn2Δ msn4Δ mutants. The image shows Ygp1–3HA and Pgk1 (loading control). The expression pattern was confirmed by two biological replicates, each having three technical replicates. The graph shows the average fold change ± SEM of the fold changes of Ygp1–3HA protein levels relative to that in wild-type 4 hours (n = 4 technical replicates from one representative biological replicate). (E) The fold change of YGP1 mRNA levels in the ygp1Δ mutant harboring a YCplac33-YGP1, YCplac33-YGP1 STRE1-del, or YCplac33-YGP1 STRE2-del plasmid at the logarithmic growth phase or the diauxic shift. The expression pattern was confirmed by three biological replicates, each having three technical replicates. The data show the average Rq values ± SEM of the fold change relative to that in the logarithmic growth phase sample of ygp1Δ [YCplac33-YGP1] (n = 3 technical replicates from one representative biological replicate). Post-hoc pairwise comparisons were conducted using the aligned rank transform (ART) procedure followed by Holm-adjusted p-values to control the family-wise error rate.

https://doi.org/10.1371/journal.pone.0355681.g009

Ygp1 maintains long-term cell survival under the short-lived puf5Δ background

We next examined the physiological importance of the regulation of YGP1 expression by Puf5. The ygp1Δ mutant did not exhibit growth retardation or genetic interaction with the puf5Δ mutant under optimal conditions (YPD medium at 25–30 °C) (Fig 10A, 25 °C and 30 °C), at elevated temperature (Fig 10A, 35 °C and 37 °C) or non-fermentative condition (YPGlylac medium) (Fig 10B).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 10. YGP1 regulation mediated by Puf5 maintains long-term cell survival.

(A) Spot assay data of the wild-type, puf5Δ, ygp1Δ, and puf5Δ ygp1Δ mutants. The strains were cultured until mid-log phase, serially diluted, and spotted on YPD plates. The plates were incubated at an appropriate temperature. The representative images are shown. (B, C) Spot assay data of the wild-type, puf5Δ, ygp1Δ, and puf5Δ ygp1Δ mutants. The strains were cultured until mid-log phase, serially diluted, and spotted on YPD plates, YPGlylac plates (B), or SC plates (C). The plates were incubated at 25 degrees. The representative images are shown. (D, E) Growth curve of the wild-type, puf5Δ, ygp1Δ, and puf5Δ ygp1Δ mutants in YPD (D) or SC medium (E). The strains were cultured overnight and transferred in fresh medium to final OD₆₀₀ < 0.001. Optical density (OD₆₀₀) was measured every 1 hour. The data show the mean ± SE (n = 2 biological replicates) of the measured OD₆₀₀. (F) Cell viability during long-term culture in the wild-type, puf5Δ, ygp1Δ, and puf5Δ ygp1Δ mutants. The strains were cultured until mid-log phase (day 0) and continuously harvested until day 7 in SC medium. Cell viability was measured at each time point. A two-dimensional regression analysis was performed on the measurements (n = 2 biological replicates), and both the observed data and the fitted regression curve are presented. The graph was generated using the ggplot2 package in R [45,46]. (G) Spot assay data of the wild-type, puf5Δ, ygp1Δ, and puf5Δ ygp1Δ mutants. The strains were cultured until mid-log phase, serially diluted, and spotted on YPD plates, YPD-Acetic Acid 30 mM plates, and YPD-Acetic Acid 60 mM plates. The plates were incubated at 25 degrees. The representative images are shown. (H) Spot assay data of the wild-type, puf5Δ, haa1Δ, and puf5Δ haa1Δ mutants. The strains were cultured until mid-log phase, serially diluted, and spotted on YPD plates, YPD-Acetic Acid 30 mM plates, and YPD-Acetic Acid 60 mM plates. The plates were incubated at 25 degrees. The representative images are shown.

https://doi.org/10.1371/journal.pone.0355681.g010

Since we demonstrated that the Puf5-mediated regulation involving Haa1 contributes to the diauxic shift in the strain harboring a plasmid pre-cultured in SC medium (Figs 8A–C and S4 Figs A–C), we next examined cell growth or survival rate in SC medium. The puf5Δ, ygp1Δ, and puf5Δ ygp1Δ mutants grew similarly to the wild type in SC medium (Figs 10C–E). However, the long-term survival rate was decreased in the puf5Δ mutant and further decreased in the puf5Δ ygp1Δ double mutant compared to that in the puf5Δ single mutant (Fig 10F). Therefore, Ygp1 is suggested to support long-term cell survival in the puf5Δ background. Previously, it was reported that cell survival rate decreased during long-term culture in SC medium due to glucose depletion and rapid pH decline by the accumulation of the metabolic byproduct acetic acid [44]. Because of its lower buffering capacity, SC medium is known to exhibit a more dynamic pH decrease than YPD medium. Upon revision of the stationary-phase marker HSP26 during long-term culture in YPD medium, we found that HSP26 expression during the stationary phase (days 1–6) were lower in the puf5Δ mutant than in wild-type cells, while its induction at the diauxic shift was similar between the two strains (Fig 1B). Therefore, we hypothesized that Puf5 contributes to the acid stress responses and maintains long-term cell survival, with Ygp1 playing a positive role in this process in the puf5Δ mutant strain. Indeed, it was reported that Haa1-mediated YGP1 regulation is involved in the acid stress responses [21], and that YGP1 expression is induced by lactic acid, acetic acid, and hydrochloric acid stresses [23]. To confirm our hypothesis, we next investigated the sensitivity to acetic acid. Contrary to a previous report [21], the ygp1Δ mutant did not show sensitivity to acetic acid (Fig 10G). As expected, the puf5Δ mutant exhibited the sensitivity to acetic acid; however, the ygp1Δ deletion did not further enhance the acetic acid sensitivity of the puf5Δ mutant (Fig 10G). These observations suggest that Ygp1 is not a major determinant of acetic acid sensitivity on YPD plates containing acetic acid. In contrast, the haa1Δ deletion resulted in clear sensitivity to acetic acid. Moreover, the puf5Δ haa1Δ double mutant displayed more severe sensitivity than either the puf5Δ or haa1Δ single mutant (Fig 10H). Together, these results indicate that the Puf5-mediated regulation involving Haa1 contributes to the cellular response to acetic acid stress, while the specific contribution of Ygp1 as a downstream target remains to be determined.

YGP1 expression is regulated by Puf5 through its promoter mediated by Haa1

In this study, we investigated the regulation machinery of nutrition deprivation-responsive gene YGP1 (Fig 1A) and clarified that YGP1 expression is positively regulated by an RNA-binding protein Puf5 in the logarithmic growth phase (Figs 2A and B). This positive regulation also contributed to the cell cycle-dependent expression of YGP1 at the M-phase (Fig 3C and S1 Fig C). While the YGP1 promoter-GFP reporter assay revealed that Puf5 affects the YGP1 promoter (Figs 2D and E), the effect of the puf5Δ mutation on the reporter expression was smaller than that on the endogenous expression (Figs 2A–E). However, GFP-YGP1 3’UTR reporter expression was not significantly changed in the puf5Δ mutant (Fig 2G), suggesting that Puf5 specifically influences the YGP1 promoter. In the reporter construction, we fused ADH1 3’UTR to the YGP1promoter-GFP gene (Fig 2C), which is reported to be 1.3-times more stable than YGP1 3’UTR [47]. Therefore, it is possible that the promoter function was masked by the stabilization of reporter mRNA. Indeed, although the difference was not significant, replacing the YGP1 3’UTR with ADH1 3’UTR increased GFP reporter expression (Fig 2G). Nevertheless, the promoter deletion assay revealed the specific 60-nucleotide promoter region where Puf5 exerts its effect (Figs 5B and C). Therefore, we assume that the YGP1 promoter is the main responsible region for the regulation.

The positive regulation of YGP1 expression by Puf5 was independent of Ixr1 (Figs 4A and B, and S2 Fig). This result contrasts with previously reported Puf5-mediated CLB1/CLB2 regulation, which is dependent on Ixr1 [31,41]. From the promoter assay results, we specified a potential downstream factor Haa1, a transcriptional activator involved in weak acid responses [20–22] which regulates the expression of 80% of acid stress-responsive genes [42]. Haa1 positively regulates YGP1 expression during the logarithmic growth phase (Figs 6A and C) and also weakly affects M-phase specific induction (Fig 7C and S3 Fig C). Puf5 binds to HAA1 mRNA [43] and positively regulates its expression (Figs 6D and E), indicating that Haa1 functions downstream of Puf5. However, it remains unclear how Puf5 positively regulates HAA1 expression. Since it has been reported that Puf5 mainly functions as a negative regulator of its target mRNAs [32,33], our findings raise the possibility that the regulation of HAA1 expression by Puf5 may involve not only direct binding to HAA1 mRNA but also indirect mechanisms, such as regulation of transcription factors that regulate HAA1 expression.

In addition, Puf5 and Haa1 were suggested to regulate YGP1 expression both dependently and independently. Then, we questioned what another factor is functioning downstream of Puf5. Among transcriptional factors which have their own binding motif within 60-nucleotide promoter region (−600 to −540 from start codon) (Table 2), Rgt1, Stb5, and Yrr1 are factors whose mRNAs are bound by Puf5 [43,48]. However, none of them were reported to bind to YGP1 promoter or regulate YGP1 expression. Therefore, we speculate that another factor than ones listed in Table 2 or 3 possibly mediates the regulation. Mutant screening that identifies genes upregulating YGP1 expression in the puf5Δ background would be an effective approach to specify the downstream factor of Puf5.

YGP1 expression is regulated by dual mechanisms depending on glucose availability

We clarified that YGP1 expression was highly induced at the diauxic shift by Msn2 and Msn4. Contrary to the logarithmic growth phase, Puf5 was not involved in this burst of the YGP1 expression (Fig 1A). Therefore, the expression of the nutrition deprivation-responsive gene YGP1 is regulated by two regulatory mechanisms: a Puf5-dependent mechanism and a Msn2/Msn4-dependent mechanism. The former maintains YGP1 expression during logarithmic growth phase and also induces M-phase specific induction of YGP1 expression during cell cycle (Fig 11A), which is partly mediated by a transcriptional activator Haa1. The latter ensures the glucose-deprivation responsive induction of YGP1 expression via STRE2 within the YGP1 promoter (Fig 11B). We propose that, since glucose is still present in the medium at low levels up to 9 h (Fig 1C), the regulation mediated by Puf5 predominantly governs YGP1 expression during this glucose-available log phase condition. Upon entry into the diauxic shift at 12 h, Msn2 and Msn4 undergo rapid nuclear translocation, leading to robust transcriptional activation of YGP1, which likely overcomes regulation by Puf5. Moreover, when cells are exposed to the environment where acid metabolites accumulate, the Puf5-dependent mechanism is proposed to function in response to acid stress and ensure long-term survival.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 11. Working Model: YGP1 is regulated by two mechanisms.

(A) Puf5-dependent mechanism during glucose-rich condition, which is partly mediated by Haa1. This system regulates the YGP1 expression during the logarithmic growth phase, thereby inducting M-phase specific expression of YGP1. Moreover, this regulation contributes to the acid stress responses and ensures long-term cell survival in the acid environment. (B) Msn2/Msn4-dependent mechanism at the diauxic shift. This pathway induces YGP1 expression at the diauxic shift (glucose deprivation) by STRE2 within YGP1 promoter. The Puf5-mediated regulation possibly contributes to this glucose-deprivation-responsive induction of YGP1, especially during the recovery from prior acid stress response.

https://doi.org/10.1371/journal.pone.0355681.g011

YGP1 expression is required for long-term cell survival in the puf5Δ background

The Puf5-dependent regulatory pathway appears not to be involved in the diauxic shift-responsive induction of endogenous YGP1 expression. However, when using the strain harboring a YGP1–3HA plasmid, the expression of YGP1–3HA mRNA and Ygp1–3HA protein were decreased in the puf5Δ mutant, accompanied by the decreased and delayed expression of stationary phase-responsive HSP26 gene (Figs 8A–C). Similar expression patterns of YGP1–3HA and HSP26 were also observed in the haa1Δ mutant (S4B Fig–S4D Fig). When culturing the strain harboring a YCplac33 plasmid, we pre-cultured the cells in SC-Ura medium and then shifted them to main culture in YPD medium. Previously, cells were reported to undergo the diauxic shift at 12 hours both in SC and YPD medium [44]. However, it has been revealed that the duration stationary phase yeast cells can survive (chronological life span) is much shorter in SC medium than in YPD medium [49]. This shortened lifespan is attributed to the rapid pH decrease caused by the accumulation of the metabolic byproduct acetic acid in SC medium [44]. Indeed, both the puf5Δ and haa1Δ single mutants were sensitive to acetic acid (Figs 10G and H). Moreover, the puf5Δ mutant exhibited low viability during long-term culture in SC medium, and the puf5Δ ygp1Δ double mutant showed an even more pronounced decrease (Fig 10F). Therefore, Puf5 and Haa1 are suggested to be involved in the acid stress responses and maintained YGP1 expression contributes to maintaining long-term cell survival in the short-lived puf5Δ mutant background (Fig 11A). However, the ygp1Δ single mutant was not sensitive to acetic acid (Fig 10G). Previously, Haa1-mediated regulation of YGP1 expression was reported to contribute to various acid stress responses [21]. In that study, the ygp1Δ mutant showed slightly slower growth in 60 mM acetic acid, but the effect of ygp1Δ mutation was more significant in the response to benzoic acid and octanoic acid. Notably, YGP1 expression has been reported to be induced by propionic acid, lactic acid, acetic acid, and hydrochloric acid [21,23]. Therefore, we propose that Ygp1 acts downstream of the acetic acid stress response, although its physiological impact becomes evident only under relatively strong acid stress compared to acetic acid.

While the Puf5 and Haa1 are involved in the response to acetic acid, the sensitivity of the puf5Δ haa1Δ double mutant to acetic acid was stronger than the puf5Δ or haa1Δ single mutant (Fig 10H). Considering Puf5 functions both dependently and independently of Haa1, there appears to be the Haa1-independent regulatory pathway by Puf5 in acid stress responses. Previous genome wide screening of the mutation library for acid stress responses has revealed that various biological processes are involved in this process [23]. Interestingly, among the genes whose deletion enhances resistance or sensitivity to acid stress [23], approximately 50% genes showed the evidence of Puf5 binding to their transcribed mRNA [43,48]. In particular, LRG1, a previously identified target of Puf5 [30,37–39], was found to be a gene whose deletion increases resistance to acid stress [23]. LRG1 encodes a GTPase-activating protein for Rho1 and regulates the MAPK pathway controlling cell wall integrity (CWI) [34–36]. Indeed, cell wall-related genes were reported to be important for organic acid responses [23]. Therefore, these findings suggest that Puf5 contributes to the acid stress responses by regulating multiple targets, with Haa1 being one important downstream factor of Puf5 in this process.

In addition, the ygp1Δ mutation accelerated the mortality of the puf5Δ mutant (Fig 10F). However, the survival rate of the ygp1Δ single mutant was comparable to that of the wild-type (Fig 10F). In contrary, the ygp1Δ mutant has previously been reported as a short-lifespan mutant [50]. In that study, cells were grown to complete entry into quiescence and then transferred to water to measure chronological lifespan. This indicates that Ygp1 is critical for survival in the quiescent state. Therefore, it is possible that extending the measurement period would reveal a decrease in long-term viability in the ygp1Δ mutant.

The perspective of understanding Puf5-mediated regulatory network

The analysis of the YGP1 regulatory machinery revealed a novel function of Puf5 in the acid stress responses. This finding indicates the importance of Puf5-mediated balancing of mRNA expression during stress adaptation. Although YGP1 alone has minimal physiological significance, detailed analysis of such genes may lead to the discovery of new perspectives. Moreover, this study uncovered crosstalk between the acid stress response pathway mediated by an RNA-binding protein Puf5 and the general stress responsive transcriptional factors Msn2 and Msn4. Signals from accumulation of acidic metabolites or glucose deprivation appear to balance these two regulatory networks. Given that Ygp1 is a secretory glycoprotein, we propose that these pathways fine-tune YGP1 induction to coordinate stress responses and cell survival, thereby contributing to environmental sensing. Because Puf5 is an RNA-binding protein targeting more than 1,000 mRNAs [43,48], it is possible that Puf5 acts as an upstream regulator of this tuning system. Recently, Pum1 and Pum2, functional orthologs of Puf5 in humans, were reported to target metabolism-related mRNAs and possibly contribute to the progression of cardiovascular diseases and cancers [51]. The tumor microenvironment is known to be acidic, which favors cancer cell survival [52]. In addition, glucose availability is closely linked to both pathologies. We believe that understanding the Puf5-mediated stress-responsive machinery will help uncover the broader Puf family regulatory network and its pathological significance.