OUP user menu

Bdf1p deletion affects mitochondrial function and causes apoptotic cell death under salt stress

Xiangyong Liu, Hui Yang, Xiaohua Zhang, Liangyu Liu, Ming Lei, Zhaojie Zhang, Xiaoming Bao
DOI: http://dx.doi.org/10.1111/j.1567-1364.2008.00469.x 240-246 First published online: 1 March 2009


The Saccharomyces cerevisiae BDF1 gene encodes a bromodomain-containing transcription factor. We previously reported that deletion of Bdf1p in yeast cells resulted in increased sensitivity to NaCl stress. In this paper, we show that the function of Bdf1p in salt tolerance is not directly linked with the Ena1p-mediated Na+ extrusion system, and a number of other well-characterized stress-response pathways. Interestingly, however, our data demonstrate that, under the NaCl stress, the absence of Bdf1p leads to mitochondrial dysfunction, including decreasing of mitochondrial membrane potential (ΔΨ) and accumulation of reactive oxygen species, and chromatin fragmentation and condensation. These results indicate that the bromodomain-containing protein, Bdf1p, is involved in the regulation of apoptosis in yeast cells.

  • apoptosis
  • bromodomain
  • mitochondria
  • Saccharomyces cerevisiae
  • stress


Bromodomains are evolutionarily conserved and act as acetyl-lysine binding domains that process the molecular information conveyed by lysine acetylation modification (Zeng & Zhou, 2002; Yang, 2004). The bromodomain-containing proteins are a novel family of transcriptional regulators shown to be involved in a number of cellular processes. We previously reported that deletion of BDF1 gene confers increased sensitivity to NaCl-stress in Saccharomyces cerevisiae (Liu, 2007).

Apoptosis, or programmed cell death (PCD), is an intrinsic cell death process that plays critical roles in the normal development and health of multicellular organisms. It has been reported that stress conditions such as hyperosmotic (Silva, 2005), hydrogen peroxide (H2O2) (Madeo, 1999) and NaCl stress (Huh, 2002; Wadskog, 2004) can induce apoptosis in yeast cells. Apoptotic markers such as accumulation of reactive oxygen species (ROS), chromatin condensation, nuclear fragmentation and DNA breaks have been observed in NaCl stress-induced apoptotic yeast cells (Huh, 2002; Wadskog, 2004). The mitochondrial function has an intimate relationship with the NaCl-induced PCD (Butcher & Schreiber, 2003). We observed that deletion of BDF1 gene not only confers increased sensitivity to NaCl stress but also downregulates the expression of several genes associated with mitochondrial function (Liu, 2007).

Several important signal pathways and genes involved in salt tolerance have been identified, including the high osmolarity glycerol (HOG) pathway (de Nadal, 2002), which mediates the production and accumulation of the osmolyte glycerol (Rep, 2000), and the ENA1 gene, which encodes a plasma membrane P-type Na+-ATPase ion pump (Haro, 1991; Garciadeblas, 1993). Further investigation into the relationships between BDF1 and these pathways will be needed to better understand the function of Bdf1p on the yeast salt–stress response.

In this study, we explore the molecular mechanism underlying the Bdf1-dependent salt tolerance by examining the role of Bdf1 in a number of stress–response pathways, mitochondria functions, and apoptosis.

Materials and methods

Yeast strains and media

The yeast strains and genotypes are listed in Table 1. Yeast cells were routinely grown in YPD media (1% yeast extract, 2% peptone, 2% glucose) or in synthetic complete (SC) medium (0.17% yeast nitrogen base, 0.5% (NH4)2SO4, 2% glucose), and supplemented with all amino acids except those not to be added in order to maintain plasmids (uracil or leucine).

View this table:

Yeast strains used in this study

StrainGenotypeReference or source
W303-1AMATa, leu2-3/112 ura3-1 trp1-1 his3-11/15 ade2-1 can1-100 GAL SUC2 mal0de Jesus Ferreira (2001)
y05308MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 bdf1kanMX4de Jesus Ferreira (2001)
BLZ-ADerivative of W303-1A: bdf1kanMX4This study
YSH818Derivative of W303-1A: hog1LEU2Eberhardt & Hohmann (1995)
LZ-105Derivative of W303-1A: hog1LEU2 bdf1kanMX4This study
YN10Derivative of W303-1A: ena1HIS3Wieland (1995)
LZ-107Derivative of W303-1A: ena1HIS3 bdf1kanMX4This study
DHT23-1BDerivative of W303-1A: cmp1URA3 cmp2HIS3Nakamura (1993)
LZ-139Derivative of W303-1A: cmp1URA3 cmp2HIS3 bdf1kanMX4This study
TL01Derivative of W303-1A: trk1LEU2This study
TB02Derivative of W303-1A: trk1LEU2 bdf1kanMX4This study

Gene disruption

BDF1 was disrupted by transformation of yeast strains with a 2859 bp bdf1::kanMX4 gene deletion allele, which was amplified from the genomic DNA of strain y05308 using primers B1-1(5′-TTGTATCAGGCGCTTG-3′) and B1-2(5′-GTCTTTCAATACTTCGTC-3′). trk1Δ deletion mutants were constructed by long flanking homology -PCR as previously described (Wach, 1996). For disruption of TRK1, the primers B2-5(5′-ATTACCTATCAAGATAAAGAGC-3′) and B2-6(5′-GTCATCACCGAAACGCGCTGTTAGTTATCAAAAAAAGGT-3′) were used to amplify the 5′ homologous sequence of the TRK1 locus, and the primers B2-7(5′-AGAAAATACCGCATCAGGAAACATATATACTCATTCTTTAGAAAT-3′) and B2-8 (5′-GAGATTGTATGTCGTACATTAT-3′) were used to amplify the 3′ homologous sequence. Plasmid pRS315 was used as template to amplify the LEU2 disruption cassette. All the gene disruptions were confirmed by PCR.

Spot dilution growth assays

Cells were precultured in YPD liquid medium overnight. Cells were harvested by centrifugation, washed twice with water and resuspended in water. The cell density was normalized to OD600 nm=1.0. A 10-fold serial dilution of this culture was made and 4 μL of each dilution was spotted onto an appropriate solid medium.

Detection of ROS

ROS production was detected with dihydrorhodamine 123 (Sigma Chemical Co.). Dihydrorhodamine 123 (50 μM) was added into cell culture and incubated for 2 h and monitored with fluorescence microscope using a rhodamine optical filter. Fluorescence intensity was quantified using a fluorescence microplate reader. Cells were cultured at 30 °C for 2–3 days.

Assessment of mitochondrial membrane potential (ΔΨ)

Cells were collected at log-phase and washed once with phosphate-buffered saline (PBS). Cells were resuspended in 100 μL PBS containing 1 μM Mitotracker Red CMRos (Molecular Probes). Cells were mounted on a coverslip and checked immediately. A Nikon TE300 inverted microscope equipped with a Cascade 650 cooled monochrome digital camera (Roper Scientific) was used for image acquisition.

Superoxide dismutase (SOD) activities

Total SOD activity was measured using a SOD detection kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) (Zhang, 2006). The enzyme activity was monitored spectrophotometrically at 550 nm by hydroxylamine assay, which was generated from xanthine-oxidase assay. One unit (U) of SOD activity was defined as 50% reduction of the A550 nm. The protein concentration was measured using the Coomassie Blue G-250-binding assay with bovine serum albumin as the standard.

4′,6-Diamidino-2-phenylindole (DAPI) staining

Yeast cells were collected and washed twice with 1 × PBS buffer, resuspended in 80% ethanol for 10 min, washed with 1 × PBS and incubated with 1 μg mL−1 of DAPI for 15 min. The stained cells were mounted on a coverslip and checked immediately. A Nikon TE300 inverted microscope equipped with a Cascade 650 cooled monochrome digital camera was used for image acquisition.

Transmission electron microscopy (TEM)

TEM sample was prepared according to Yang (2006). Cells cultured at 30 °C in YPD medium to log-phase were harvested by gentle centrifugation, washed in PBS (pH 7.2), resuspended in 2.5% (v/v) glutaraldehyde in PBS and fixed for 40 min at room temperature. Cells were further fixed by 2% potassium permanganate in water for 1 h at room temperature. Fixed cells were dehydrated with 30%, 50%, 75%, 85%, 95%, and 100% ethanol. Cells were transitioned with propylene oxide, infiltrated in Spurr resin (Electron Microscopy Sciences, Hatfield, PA) in a microwave with vacuum. The Microwave (Pelco BioWave 34700, Ted Pella Inc., Redding, CA) setting was 650 W for 2 min and the restriction temperature was 43 °C. Resin was polymerized at 65 °C overnight in the oven. Ultrathin 60-nm sections were cut with a diamond knife, stained with 2% uranyl acetate and lead citrate, examined using a Hitachi H-7000 electron microscope, equipped with a high resolution (4 × 4K) cooled CCD digital camera (Gatan Inc.).

Results and discussion

The function of Bdf1p in yeast cells under salt stress is not directly linked with the currently known stress response pathways

In a previous paper, we reported that deletion of BDF1 results in increased sensitivity to NaCl stress (Liu, 2007). To further study this phenomenon, we investigated the relationship between BDF1 and genes that are known to play a role in salt tolerance.

Under high salt conditions, the yeast cells are mainly challenged by ion and hyperosmotic stress. Adaptation to hyperosmotic stress is dependent primarily on the HOG pathway (de Nadal, 2002). The mitogen-activated protein kinase Hog1p is the key factor of the HOG signaling pathway (Posas, 1996). To investigate the relationship between BDF1 and the HOG signaling pathway, the bdf1::kanMX4 disruption was introduced into a strain lacking HOG1. When the bdf1Δhog1Δ double-deletion strain was assayed for salt sensitivity, an increased Na+ sensitivity relative to either single mutant was observed (Fig. 1a). This result suggests that the Bdf1p-dependent NaCl resistance is independent of the HOG signaling pathway.


Genetic interactions between bdf1Δ and mutations in salt tolerance genes. (a,b,d,e) The function of Bdf1p on salt tolerance is independent of the HOG pathway, Cmp1pCmp2p, Ena1p and Trk1p.Ten-fold serial dilution of yeast strain were spotted onto YPD plates supplemented with the indicted concentrations of NaCl. The strains that were used are W303-1A (wild type), BLZ-A (bdf1Δ), YSH818 (hog1Δ), LZ-105 (bdf1Δhog1Δ), DHT23-1B (cmp1Δcmp2Δ), LZ-139(cmp1Δcmp2Δbdf1Δ), YN10(ena1Δ), LZ-107(ena1Δbdf1Δ), TL01(trk1Δ), TB02(trk1Δbdf1Δ). (c) Bdf1p and calcineurin exert different functions in Na+ tolerance. Ten-fold serial dilution of the W303-1A (wild type), BLZ-A (bdf1Δ) and DHT23-1B (cmp1Δcmp2Δ) were spotted onto YPD plates with or without FK506 (0.1 μg mL−1), and YPD containing 0.4 M NaCl with or without FK506 (0.1 μg mL−1), as indicated. Cells were culture at 30°C for 2 days. WT, wild type.

Calcineurin function is required for the normal salt stress resistance in S. cerevisiae. Under Na+ stress, an increase in intracellular Ca2+ concentration stimulates calmodulin, which in turn activates the calcineurin. Calcineurin functions through the Crz1p/Tcn1p transcription factor to induce the expression of several salt stress-responsive genes, such as ENA1 (Yoshimoto, 2002). We therefore examined whether there is a functional link between BDF1 and calcineurin. The calcineurin-defective mutant DHT23-1B, in which the calcineurin catalytic subunit-encoding genes CMP1 and CMP2 had been deleted, was used to construct the bdf1Δcmp1Δcmp2Δ triple mutant. Interestingly, the triple mutant showed enhanced sensitivity to NaCl, as compared with the bdf1Δ or Δcmp1Δcmp2Δ mutant (Fig. 1b).

We also monitored the growth behaviors of wild type, bdf1Δ and cmp1Δcmp2Δ subjected to NaCl stress in the presence or absence of the immunosuppressant drug FK506, a specific inhibitor of calcineurin (Nakamura, 1993). As shown in Fig. 1c, the growth of these strains was not affected on the YPD plate to which only 0.1 μg mL−1 FK506 was added. The salt sensitivity of bdf1Δ was similar to that of the cmp1Δcmp2Δ on medium containing 0.4 M NaCl in the absence of FK506. However, in the presence of FK506, the bdf1Δ mutant displays a more severe growth defect than the cmp1Δcmp2Δ mutant. Taken together, we conclude that Bdf1p and calcineurin exert their effects in Na+ tolerance through different mechanisms.

The controlled ion transport mechanism (influx and extrusion) is critical for salt tolerance. In S. cerevisiae, the ENA1 gene is the major determinant of the Na+ extrusion system (Haro, 1991; Garciadeblas, 1993). Deletion of ENA1 causes hypersensitivity to salt stress. Several different signal pathways have been implicated in the control of ENA1 expression (Márquez & Serrano, 1996). To test the possible genetic interaction between BDF1 and ENA1-encoded Na+ extrusion system, the bdf1Δena1Δ double mutant was constructed and the sensitivity to Na+ ions was analysed. Interestingly, the bdf1Δena1Δ double mutant showed enhanced sensitivity to NaCl, as compared with the single mutants (Fig. 1d). This result suggests that the function of BDF1 involved in Na+ tolerance is not directly linked with ENA1.

The primary K+ transport system in yeast is encoded by the TRK1 and TRK2 genes (Ramos, 1990; Ko & Gaber, 1991). This system also permits the influx of Na+ under salt stress conditions. It increases affinity for K+ to reduce Na+ permeability and increases Na+ tolerance (Gómez, 1996). Yeast mutants with impaired K+ transport systems showed increased sensitivity to salt stress due to accumulation of more Na+. To test whether there is some link between the Bdf1p and K+ transport system, the trk1Δ deletion mutant was constructed. As expected, the trk1Δ deletion mutant has a salt-sensitive phenotype (Fig. 1e). We performed the spot tests on YPD medium containing different salts and compared the growth of bdf1Δ, trk1Δ and bdf1Δtrk1Δ. As shown in Fig. 1e, the bdf1Δtrk1Δ double mutant showed enhanced sensitivity to NaCl compared with the single mutants. The result suggests that Bdf1p and K+ transport systems use different mechanisms in yeast salt tolerance.

Taken together, these results suggest that Bdf1p does not exert its role in salt tolerance through the HOG pathway, the calcineurin pathway, the K+ transportation system, or the Ena1p-mediated Na+ extrusion system.

Bdf1p affects mitochondrial function in cells under salt stress

bdf1Δ cells grow very slowly in a nonfermentable carbon source (Lygerou, 1994), suggesting a mitochondrial respiratory deficiency in these cells. Consistent with this notion, our recently published microarray data revealed that several genes associated with mitochondrial function were downregulated in the bdf1Δ strain after the salt treatment (Liu, 2007). Thus, it was interesting to investigate the status of the mitochondrial function in bdf1Δ cells under salt stress conditions. We measured mitochondrial membrane potential (ΔΨ) and ROS production as indicators of mitochondrial function (Butcher & Schreiber, 2003; Pozniakovsky, 2005).

To measure ΔΨ, Mitotracker red CMRos was used. Mitotracker red CMRos stains mitochondria in a ΔΨ-dependent fashion (Pozniakovsky, 2005). It specifically stains mitochondria when the ΔΨ is high, but stains in a diffused pattern when ΔΨ is lost. As shown in Fig. 2, after treatment with 0.6 M NaCl for 30 min, the wild-type cells displayed a linear MitoTracker staining, but about 20% of bdf1Δ cells showed a diffused staining (Fig. 2, arrow), indicating that absence of Bdf1p caused loss of ΔΨ under salt stress conditions.


Assessment of mitochondrial membrane potential in wild-type and BLZ-A (bdf1Δ) cells after treated with 0.6 M NaCl for 30 min. Cells were stained with MitoTracker Red for mitochondria membrane potential. About 20% of the bdf1Δ cells lost the Mitochondrial membrane potential and showed a diffused MitoTracker staining (marked by arrow) when treated with 0.6 M NaCl for 30 min. WT, wild type.

The slow growth of the bdf1Δ cells on nonfermentable carbon seems in contradiction to the results of the Mitotracker staining, in which the majority of the bdf1Δ cells appear to have functional mitochondria, as indicated by the linear staining. One possibility is that the mitochondrial function of bdf1Δ is impaired but not totally lost. Only the severely impaired mitochondria (about 20%) were revealed by the Mitotracker staining.

To monitor the ROS production, we used dihydrorhodamine 123(DHR), which becomes fluorescent chromophore rhodamine when oxidized by intracellular ROS. As shown in Fig. 3a, wild-type cells showed a dim, red fluorescence after incubation with DHR, whereas bdf1Δ cells showed an intense, red fluorescence after treatment with NaCl. Quantifying the relative fluorescence intensities with a fluorescent microplate reader showed a 1.5-fold increase in fluorescence intensities in the bdf1Δ mutant (Fig. 3b).


Reactive oxygen species production in wild-type and BLZ-A (bdf1Δ) cells treated with 0.6 M NaCl for 30 min. ROS production detected by dihydrorhodamine 123(DHR). (a) Bdf1Δ cells displayed more bright intense red fluorescence than wild-type cells staining with DHR after treatment with NaCl. (b) Quantification of ROS production. Relative fluorescence intensities were measure on a fluorescent microplate reader after staining with DHR.Values are means of three independent experiments. WT, wild type

SOD is the key enzyme providing cell protection from ROS toxicity. SOD catalyses the superoxide anion into H2O2 and molecular oxygen (Imlay & Linn, 1988; Esterbauer, 1991). We therefore investigated whether the accumulation of ROS in bdf1Δ cells was due to the inhibition of SOD activity in the mutant. SOD activities in the wild-type and bdf1Δ cells before and after salt treatment were determined. As shown in Fig. 4, after the salt treatment, the SOD activity increased significantly in bdf1Δ, and was 38 U higher than in wild-type cells. This indicated that the SOD activity is apparently unaffected by BDF1 deletion and responded normally in the bdf1Δ mutant.


SOD activities in wild-type and BLZ-A (bdf1Δ) cells treated with 0.6 M NaCl for 30 min. Values are means of three independent experiments. WT, wild type.

bdf1Δ cells display characteristics of apoptosis under salt stress conditions

Accumulation of ROS has been described as a key factor triggering apoptosis (Eisenberg, 2007; Fröhlich, 2007). Therefore, the high level of ROS in bdf1Δ led us to test whether apoptosis occurs in this mutant.

Cell survival of the wild-type and bdf1Δ cells after salt treatment was measured using phloxine-B, which stains only dead or dying yeast cells (Ahn, 2005). After treatment with 0.6 M NaCl for 30 min, about 30% of the bdf1Δ cells showed positive phloxine-B, compared with only about 2% of wild-type cells.

Chromatin fragmentation and condensation is a typical marker of apoptosis (Yamaki, 2001; Li, 2006). DAPI staining followed by fluorescence microscopy was used to observe nuclear structure. As shown in Fig. 5, after treatment with NaCl, wild-type cells had a normal nuclear morphology and the nucleus appeared as a single round spot. In contrast, several bdf1Δ mutant cells displayed randomly distributed nuclear or chromatin fragments aligned in a half-circle (Fig. 5).


DAPI staining of wild-type and BLZ-A (bdf1Δ) cells treated with 0.6 M NaCl for 30 min. Nuclei are fragmented and randomly distributed (arrow) in bdf1Δ cells. BF, bright field; WT, wild type.

Electron microscopy also revealed that when treated with NaCl, the bdf1Δ cells displayed extensive chromatin condensation and margination along the nuclear envelope, which is typical for apoptotic cells (Fig. 6d).


Electron microscopy images of exponentially growing cells of wt and BLZ-A (bdf1Ä) mutant with or without 0.6 M NaCl for 45 min. (a, b) wild-type control W303. (c–f) BLZ-A (bdf1Ä) mutant. Chromatin condensation is marked by arrows (d). Nuclear membrane widened is marked by arrowheads (f). M, mitochondria, N, nucleus, V, vacuole. WT, wild type. Scale bar 1 μm.

Bromodomain-containing proteins have been shown to be involved in cell differentiation, cell cycle progression, and signal transduction (Maruyama, 2002; Yang, 2004). To the best of our knowledge, these data provide, for the first time, evidence for a bromodomain-containing protein participating in the yeast apoptosis. It has been shown that the mitochondrial functions are important for yeast apoptosis. Our results indicate that the absence of Bdf1p leads to mitochondrial dysfunction, including decrease of mitochondrial membrane potential, accumulation of ROS and downregulation of mitochondrial gene expression (Liu, 2007). It is likely that Bdf1p participates in apoptosis indirectly by improving mitochondrial function. If this were the case, other respiration-deficient mutants (such as petite) would have the same apoptotic phenotype as the bdf1Δ mutant.


This work was supported by the Natural Science Foundation of China (grant Nos. 30170021, 30671143, 30570031 and 30740420552).


  • Editor: Ian Dawes


View Abstract