l-Carnitine plays a well-documented role in eukaryotic energy homeostasis by acting as a shuttling molecule for activated acyl residues across intracellular membranes. This activity, supported by carnitine acyl-transferases and transporters, is referred to as the carnitine shuttle. However, several pleiotropic and often beneficial effects of carnitine in humans have been reported that appear to be unrelated to shuttling activity, but little conclusive evidence regarding molecular mechanisms exists. We have recently demonstrated a role of carnitine, independent of the carnitine shuttle, in yeast stress protection. Here, we show that carnitine specifically protects against oxidative stress caused by H2O2 and the superoxide-generating agent menadione. Surprisingly, carnitine has a detrimental effect on survival when combined with thiol-modifying agents. Central elements of the oxidative stress response, specifically the transcription factors yap1p and Skn7p, are shown to be required for carnitine's protective effect, but several downstream effectors are dispensable. A DNA microarray-based analysis identifies Cyc3p, a cytochrome c heme lyase, as being important for carnitine's impact during oxidative stress. These findings establish a direct genetic link to a carnitine-related phenotype that is independent of the shuttle system and suggests that Saccharomyces cerevisiae should provide a useful model for further elucidation of carnitine's physiological roles.
The role of l-carnitine in cellular metabolism has been well characterized for its function in the transfer of the activated products of energy metabolism between intracellular compartments (Bieber, 1988; Reddy & Mannaerts, 1994). Activated acyl residues are transferred to carnitine by the activities of various carnitine acyl-transferases and transported by carnitine/acyl-carnitine carriers across the membranes of the mitochondria and peroxisomes. Collectively, this molecular network is referred to as the carnitine shuttle. The carnitine shuttle of Saccharomyces cerevisiae closely resembles that of higher eukaryotes in function, with minor differences in the composition. Most notably, yeast appears to harbor only short-chain carnitine acyl-transferase activity, in contrast to the long-chain acyl-transferase activities present in higher eukaryotes (Kispal, 1993). In yeast, this activity is ascribed to three carnitine acetyl-transferases (CATs), Yat1p, Yat2p and Cat2p, compared with a single mitochondrial CAT in mammalian systems (Schmalix & Bandlow, 1993; Swiegers, 2001). Furthermore, in yeast, the glyoxylate cycle provides an alternative route to support energy generation from peroxisomally generated acetyl-CoA (van Roermund, 1995; Palmieri, 1997). A final difference between yeast and mammals is that in yeast, acetyl-coenzyme A is generated both in the cytosol and in the peroxisomes, compared with the peroxisomal and mitochondrial acetyl-CoA generation in higher eukaryotes (Kunau, 1995).
Carnitine supplementation results in various beneficial effects in human subjects or cell lines. Most notably, supplementation of carnitine and of various acyl-carnitines has been associated with protecting against neurodegeneration and mitochondrial decay resulting from aging and against the onset of apoptosis in various cell lines (Rani & Panneerselvam, 2001; Calabrese, 2005, 2006), also suggesting a link to oxidative stress protection. Several hypotheses have been generated to explain these effects. These include the upregulation of the mammalian stress-responsive gene HO-1, encoding a heme oxygenase (Calabrese, 2006; Calo, 2006). Furthermore, carnitine has also been postulated to have a possible intrinsic antioxidant capacity, which has been proposed to be involved in some of the observed effects in neuronal apoptosis protection (Gulcin, 2006; Silva-Adaya, 2008). Carnitine, through its involvement in energy metabolism, has been reported to have a stimulatory effect on mitochondrial metabolism, which has been linked to the benefits of carnitine and acetyl-carnitine supplementation. In particular, feeding carnitine to rats was shown to reverse the age-dependent decline in mitochondrial membrane potential and the decline in cardiolipin levels (Hagen, 1998). In thio-redoxin-deficient DT 40 cells, acetyl-carnitine was shown to specifically protect against oxidative stress in and around the mitochondria, including the release of cytochrome c and SOD1 (Zhu, 2008). However, it is unclear whether these changes are influenced directly or indirectly by the presence of carnitine and acyl-carnitines. Moreover, no direct mechanisms have thus far been proposed to explain the different, generally beneficial effects of dietary carnitine supplementation.
Recently, we have demonstrated a role for carnitine in the protection against weak organic acid stress induced by for example acetate and oxidative stress induced by hydrogen peroxide in the eukaryotic model organism, S. cerevisiae (Franken, 2008). This effect was found to be independent of all yeast CATs and therefore also of the carnitine shuttle. Here, we present an analysis of the protective effect of carnitine against oxidative stress in yeast. The data indicate that carnitine specifically protects against the effects of the reactive oxygen species (ROS)-generating agents H2O2 and menadione. In liquid media, this effect was obvious at carnitine concentrations below 1 mg L−1, whereas a visual impact on plates required higher levels. Intriguingly, carnitine has an opposite and detrimental effect when combined with the thiol-modifying agents diamide and dithiothrietol. This observation is the first report of a damaging effect associated with carnitine supplementation. The data also confirm that carnitine has no free radical scavenging activity. A genetic analysis reveals that all the observed impacts of carnitine require genetic mediation, in particular the general regulators of the oxidative stress response, encoded by yap1 and SKN7. A global gene expression analysis by DNA microarray comparing stressed and unstressed yeast grown in the presence and absence of carnitine led to the identification of the cytochrome c heme lysase Cyc3p as being important for carnitine-mediated stress protection. This finding links carnitine to mitochondrial functions, and suggests several possible pathways through which carnitine may exert its molecular effects.
Materials and methods
Yeast strains and media
The wild-type strain BY4742 and all single gene knockouts used in this study were obtained from Euroscarf (Frankfurt, Germany). The strains in which yap1p and Msn2p have been replaced with green fluorescent protein (GFP)-tagged versions of the same proteins were purchased from Invitrogen. All strains are derived from the S288c genetic background and were grown either on yeast extract peptone dextrose or on minimal yeast nitrogen base with dextrose media containing a 0.67% w/v yeast nitrogen base without amino acids (Difco) and 2% w/v glucose supplemented with amino acids according to the specific requirements of the respective strains to standardized concentrations described previously (Ausubel, 1994).
Multicopy expression of CYC3 and creation of the Δcyc1Δcyc7 double mutant
The plasmids and constructs used in this study are listed in Table 1, and the primers for the amplification of CYC3 and also the CYC1 integration cassette are listed in Table 2. Standard DNA techniques were used for the isolation and manipulation of DNA throughout the study (Sambrook, 1989; Ausubel, 1994). The restriction enzymes, T4 DNA-ligase and Expand Hi-Fidelity polymerase used in the enzymatic manipulation of DNA were obtained from Roche Diagnostics (Randburg, South Africa) and used according to the specifications of the supplier. Escherichia coli DH5α (Gibco-BRL/Life Technologies) was used as the host for the construction and propagation of all plasmids. Sequencing of all plasmids was carried out on an ABI PRISMTM automated sequencer.
Sequences with homology to the URA3 region of the plasmid YDp-U are underlined.
For the cloning of CYC3, the primer pair CYC3-F and CYC3-R was used, which amplifies a 1551-bp fragment containing the CYC3 promoter, ORF and terminator. The amplified fragment was ligated into the cloning vector pGEM-T-easy (Promega). The CYC3 gene cassette was excised using the EcoRI site from the pGEM-Teasy vector and a SalI site upstream of the CYC3 stop codon. The excised fragment was ligated into the plasmid YEpLac195 (Gietz & Sugino, 1988) using EcoRI and SalI restriction sites.
For the amplification of the CYC1 disruption cassette, the primer pair CYC1-URA3int-Fp and CYC1-URA3int-Rp was used, which amplifies the URA3 cassette from the plasmid YDp-U and incorporates 5′ and 3′ flanking regions homologous to regions outside the CYC1 ORF. The amplified disruption cassette was transformed into the BY4742 Δcyc7 strain to generate the Δcyc1Δcyc7 double mutant. Integration was verified by PCR using the primers CYC1-F and CYC7-R.
ABTS antioxidant assay
To determine the antioxidant capacity of carnitine, the ABTS [2,2′-azinobis-(-3-ethyl-benzothiazolin-6-sulfonic acid)] radical cation decoloration assay was used as described previously (Re, 1999; De Beer, 2003). The assay can be effectively used for lipophilic and hydrophilic antioxidants and is based on the preformation of the blue/green ABTS·+chromophore by oxidation with potassium persulfate, which is then reduced in the presence of hydrogen-donating antioxidants. The resulting decoloration can subsequently be measured spectrophotometrically. The ABTS·+radical monocation was formed by addition of 88 μL of a 140 mM potassium persulfate solution to 5 mL of 7 mM ABTS, which was protected from light and incubated overnight at room temperature. The ABTS solution was diluted to OD734 nm 0.7 for use in the assay. Dilution series of 50, 100, 150, 200, 300 and 400 mM were set up for Trolox, ascorbic acid and l-carnitine. The reaction mixture consisted of 200 μL of the diluted ABTS solution to which 10 μL of each of the samples from the mentioned dilution series was added. ABTS decoloration was measured after 4 min at OD734 nm. Assays were performed in triplicate.
Preparation of plates containing redox stress-inducing agents and growth assays
The plates used for the assessment of the effect of various oxidants [H2O2, menadione, linoleic acid hydroperoxide (LoaOOH), cumene hydroperoxide (CHP) and diamide] and also the thiol-reducing agent dithiothrietol in combination with carnitine were prepared on the day before cultures were spotted. Soybean casein digest (SCD) agar medium was left to cool to 50 °C before the addition of the oxidants (0.5 and 1.5 mM H2O2; 0.1 and 0.2 mM menadione; 1.0 and 1.5 mM LoaOOH; 0.1 and 0.25 CHP; 0.8 and 1.2 mM diamide; 8.0 and 16 mM dithiothrietol) and also l-carnitine (added in a concentration range between 50 and 1500 mg L−1). Plates were stored overnight at 4 °C in the dark. LoaOOH was prepared as described by Evans (1998).
In order to monitor the effects of the various oxidants on growth and survival, cultures of the strains to be tested were grown to an OD600 nm of 1.0, washed and spotted in four 10 × serial dilutions on the plates. The plates were incubated at 30 °C for 3 days and monitored for growth. To quantify the effect of carnitine supplementation in combination with H2O2 on the deletion mutants, growth was calculated according to the dilutions for strains grown on plates containing H2O2 and compared with growth on H2O2 with carnitine supplementation, calculated as fold difference. Experiments were repeated in triplicate and data were analyzed for significance using Student's t-test.
Determination of intracellular ROS
The oxidant-sensitive probe 2′,7′-dichlorofluorescein diacetate (DCFH-DA, Sigma) was used for the detection of intracellular ROS. The dye is cell permeable and is widely used for the rapid quantification of ROS in eukaryotic cells (Jakubowski & Bartosz, 1997). The DCFH-DA probe is nonfluorescent until acetate groups are removed by intracellular esterases and can then be oxidized by intracellular ROS to the fluorescent compound 2′,7′-dichlorofluorescein, which can be detected as an indirect measure of intracellular ROS. Cultures of BY4742 were grown overnight to inoculate fresh SCD and media and also SCD containing carnitine (1000 mg L−1) to an OD600 nm of 0.1. A duplicate set was also inoculated, which was to be treated with 0.6 mM H2O2 for 30 and 90 min, respectively, when the cultures reached the mid-log phase. After treatment, the cultures were harvested, washed once and resuspended in phosphate-buffered saline. Cells were diluted to ∼106 mL−1; DCFH-DA was added to a concentration of 100 μM and incubated for 30 min at 28 °C. Data were acquired using the BD FACSAria cell sorter, equipped with 407, 488 and 633 nm lasers and the bd facsdiva 6.1 software. The samples were acquired with an event rate of 600 s−1 using a 70-μm nozzle. Intensity histogram overlays were generated using flowjo 2.1.1. The experiment was performed in triplicate.
Cultures of the wild-type strain BY4742 was grown overnight in SCD-containing tubes to serve as precultures. Two flasks containing 50 mL of freshly prepared SCD media and also two containing SCD with carnitine (1000 mg L−1) were inoculated to an OD600 nm∼0.1 and grown to the mid-log phase (OD600 nm 0.4–0.5). A duplicate set was also inoculated and exposed to 0.4 mM H2O2 for 30 min before harvesting and isolation of total RNA. RNA was isolated as described previously (Schmitt, 1990). Probe preparation and hybridization to Affymetrix Genechip® microarrays were performed according to Affymetrix instructions, starting with 6 μg of total RNA. The results for each condition were derived from two independent culture replicates. The quality of total RNA, cDNA, cRNA and fragmented cRNA was confirmed using the Agilent Bioanalyzer 2100.
Transcriptomics data acquisition and statistical analysis
Acquisition and quantification of array images and data filtering were performed using affymetrix genechip®operating software (gcos) version 1.4. All arrays were scaled to a target value of 500 using the average signal from all gene features using gcos. Genes with expression values below 12 were set to 12+ the expression value as described previously (Boer, 2003) in order to eliminate insignificant variations. Determination of differential gene expression between experimental parameters was conducted using significance analysis of microarray (sam) version 2. The two-class, unpaired setting was used and genes with a Q-value of 0.5 (P<0.005) were considered differentially expressed. Only genes with a fold change >1.8 (positive or negative) were taken into consideration.
Impact of carnitine concentration on stress resistance and the relationship of carnitine with known redox stressors
It has been established previously that carnitine supplementation enhances the growth of yeast strains under conditions of organic acid stress and also oxidative stress induced by hydrogen peroxide (Franken, 2008). Here, the data show that this effect of carnitine is concentration dependent. On plates, the effect is visible from below 100 mg L−1 upwards, as observed in the case of protection against H2O2 (Fig. 1). Furthermore, when liquid (SCD) cultures were grown to the mid-log phase in the presence of carnitine at concentrations of 1 mg L−1, harvested, washed and spotted on peroxide plates (without carnitine supplementation), a clear protective impact was obvious when compared with cultures grown under the same conditions in the absence of carnitine (Fig. 2). No effect could be observed for cultures grown in the presence of 100 mg L−1γ-butyrobetaïne, the direct precursor of carnitine in the eukaryotic carnitine biosynthesis pathway, which differs only by a hydroxyl group on the third carbon from carnitine. Interestingly, for all concentrations above 1 up to 1000 mg L−1, growth was comparable. For concentrations below 1 mg L−1, no effect could be observed (data not shown).
Carnitine supplementation (100 and 1000 mg L−1) in combination with redox stress-inducing agents. (a) ROS-inducing stressors: H2O2 (1.5 mM), menadione (0.1 mM), LoaOOH (1.0 mM) and CHP (0.8 mM). (b) Thiol-modifying agents: the thiol-oxidizing agent diamide and the thiol-reducing agent dithiothrietol (DTT) were used at final concentrations of 1.2 and 8 mM, respectively.
Liquid cultures supplemented with carnitine (1, 10 and 100 mg L−1) and γ-butyrobetaïne (γ-BB; 100 mg L−1) were grown to the mid-log phase and spotted on SCD and SCD containing 0.75 mM H2O2. No effect of carnitine supplementation could be observed below a concentration of 1 mg L−1 (not shown).
Since a range of oxidative stress-inducing agents have been identified and their effects on the physiology of yeast cultures have been reasonably well characterized (Gasch, 2000; Thorpe, 2004), it was of interest to investigate the effect of carnitine supplementation of cultures exposed to these compounds. The compounds that were chosen for this study include the ROS-generating oxidants hydrogen peroxide, the superoxide-generating agent menadione, LoaOOH (a byproduct of lipid peroxidation), CHP (an aromatic hydroperoxide) and also the thiol-oxidizing agent diamide.
With regard to ROS-generating oxidants, carnitine visibly enhances the growth of exposed cultures in all cases, except for exposure to CHP (Fig. 1a). Although there is a growth-enhancing effect in the case of LoaOOH, it is not as clear as in the case of H2O2 and menadione.
Interestingly, carnitine addition in combination with the thiol oxidant diamide leads to a significant decrease in growth compared with cultures that are only exposed to the oxidant (Fig. 1b). Similar results are observed in combination with the thiol-reducing agent dithiothrietol. For all of the plate phenotypes, carnitine was active at media concentrations similar to those described for hydrogen peroxide above (data not shown).
Carnitine does not scavenge free radicals, but behaves like an antioxidant in a biological context
A role for carnitine and some of its esters has been indicated in protecting against the onset of apoptosis, aging and other states associated with the buildup of oxidative stress (Rani & Panneerselvam, 2001). As indicated, carnitine addition to yeast cultures exposed to 0.6 mM H2O2 was, similar to observations in higher organisms, able to drastically decrease the amount of intracellular ROS in yeast, as detected by the fluorescent probe DCFH-DA (Fig. 3a). One line of argumentation suggests that this could be due to the antioxidant effect of carnitine itself (Gulcin, 2006). To investigate whether l-carnitine is indeed able to scavenge free radicals, the molecule was used in an ABTS free radical scavenging assay. As is clear in Fig. 3b, carnitine is unable to scavenge free radicals and would, therefore, be unable to function as a molecular antioxidant.
Assessment of carnitine's possible role as an antioxidant. (a) Flow-cytometer profiles of ROS accumulation in strains treated with 0.6 mM H2O2 for 30 and 90 min with and without carnitine supplementation detected by the fluorescent probe DCFH-DA. (b) Carnitine does not exhibit the ability to scavenge free radicals using the ABTS decoloration assay; ascorbate and Trolox were used as controls. (c) Carnitine supplementation rescues the synthetic methionine auxotrophy of a Δsod2 strain under aerobic growth conditions in a concentration-dependent manner.
Using carnitine in a yeast-based biological antioxidant screen proposed by Z̆yracka (2005) (Fig. 3c) illustrates that carnitine addition to cultures of a Δsod1 strain abolishes the synthetic methionine auxotrophy of the strain under aerobic conditions. The effect of carnitine is concentration dependent and comparable to the effects reported for antioxidants such as ascorbate, glutathione, cystein and N-acetylcystein (Z̆yracka, 2005). γ-Butyrobetaïne did not show any free radical scavenging activity and was unable to complement the auxotrophy of Δsod1.
The protective effect of carnitine requires the major pathways involved in oxidative stress protection
In yeast, the main responses to oxidative stress are regulated by the transcription factors yap1p, Msn2p and also Skn7p (Jamieson, 1998). These responses include the upregulation of antioxidant enzymes, such as SOD1, TRX2 and TSA1, and enzymes involved in glutathione metabolism, such as GLR1. To assess whether the action of carnitine requires these pathways, strains bearing deletions of genes involved in these pathways were assessed for their oxidative stress response in the presence or absence of carnitine. The genes assessed included yap1 as well as the transcription factor Skn7p and also the downstream antioxidant enzymes Sod2p (superoxide dismutase, Saffi, 2006), Tsa1p (a thio-redoxin peroxidase, Chae, 1994), Trx2p (a thioredoxin, Pedrajas, 1999), Glr1p (a glutathione oxidoreductase, Grant, 2001) and Gsh1p (which catalyzes the first step in glutathione biosynthesis, Ohtake & Yabuuchi, 1991). The strains were spotted on plates containing H2O2 (in concentrations ranging between 0.1 and 1.0 mM) with and without the addition of carnitine to a final concentration of 1000 mg L−1. The results (Fig. 4) indicate that carnitine supplementation leads to significantly (P<0.05) enhanced survival in the case of Δsod2 (266-fold), Δtsa1 (35-fold) and Δtrx2 (10-fold) strains. In the case of gene deletion of yap1, SKN7 and the genes required for glutathione metabolism, carnitine was unable to compensate for the deletion of the respective genes. In fact, for strains lacking the central transcription factors required for the yeast's oxidative stress defense, namely yap1 and SKN7, a significant (34-fold for Δyap1 and 26-fold for Δskn7; P<0.5) decrease in survival was observed on plates containing 0.25 mM H2O2 and carnitine (1000 mg L−1). This effect was not apparent on lower (0.1 mM) concentrations of H2O2, indicating that it is due to the effect of carnitine in combination with concentrations of H2O2 that are sufficient to cause cellular stress and not resulting from a toxic consequence of carnitine supplementation alone.
The effect of carnitine supplementation on strains with deletions of genes required for the cells' defense against oxidative stress. The transcription factors yap1p and Skn7p and genes involved in glutathione metabolism (GSH1 and GLR1) are required for carnitine's protective effect, whereas carnitine supplementation results in enhanced growth of strains with deletions of the antioxidant enzymes SOD2, TRX2 and TSA1. Wt, wild type.
It has been indicated that certain molecules, such as green tea polyphenols, are able to induce the activation of yap1p and also Msn2p, localized in the cytosol under nonstress conditions, which leads to the nuclear localization of the two transcription factors in order to activate target gene sets (Maeta, 2007). Considering that carnitine has been proposed to act through the mammalian regulator of stress-induced genes, namely Nrf2 (Calabrese, 2005), the effect of carnitine on the localization of yap1p and Msn2p using GFP fusions of the two transcription factors was investigated. Cultures exposed to carnitine, however, did not cause a change in the localization of either yap1p or Msn2p (data not shown).
Screening for carnitine-specific genetic links to the protection against oxidative stress by carnitine
To screen for possible genetic mediators of the protective effect of carnitine, microarray analysis was performed, comparing the wild-type strain BY4742 grown in SCD with and without carnitine supplementation. In addition, global expression analysis of cultures grown to the mid-log phase (with and without carnitine addition) and then exposed to H2O2 was also performed. The data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus (Edgar, 2002) and are accessible through GEO Series accession number GSE16346 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE16346).
Using a fold cut-off value of 1.8, 40 genes were found to be upregulated and also 40 downregulated when comparing the peroxide-treated cultures supplemented with carnitine with the unsupplemented, peroxide-treated set. Comparison of the carnitine-treated and -untreated sets revealed only five genes up- and 12 genes downregulated. Deletion mutants representing all of the identified, nonessential, genes upregulated by the presence of carnitine during peroxide stress were spotted in serial dilutions on plates containing 0.75 and 1.6 mM H2O2 and also on a duplicate set that contained carnitine (added to 1000 mg L−1). Also included in the set of mutants were strains bearing deletions of AFT1, AFT2, CAD1, RPN4 and SOK2, which were identified as over-represented transcription factors within the set of upregulated genes using the yeast transcription factor database, Yeastract (http://www.yeastract.com/). A list of the identified genes and the deletion strains selected for the screen is given in the supplementary material. In addition, strains bearing deletions of HES1 and TSA2 were included, because both genes were upregulated by the presence of carnitine under nonstressed conditions. The plates were incubated at 30 °C for 3 days and inspected for growth. Fourteen out of the 38 upregulated genes that were screened were sensitive to H2O2 present in the media. Out of the five over-represented transcription factors, Δaft1, Δsok2 and Δrpn4 also tested sensitive to peroxide. For all the strains tested, except for the Δcyc3 mutant, growth enhancement by carnitine supplementation could be observed on both 0.75 and 1.6 mM H2O2. In fact, growth enhancement on 1.6 mM H2O2 was observed even when no growth was detectable in the presence of only H2O2 (as shown for the wild type in Fig. 5b), for all strains tested, except Δcyc3.
A role for CYC3 in protective and detrimental effects of carnitine. (a) Deletion of the cytochrome c heme lyase, CYC3, renders the yeast strain more sensitive to H2O2 treatment compared with the wild type (Wt) and is less responsive to growth enhancement by carnitine addition as compared with deletion of SOD2 and also all other gene deletions screened. (b) The effect of CYC3 deletion is comparable to a strain with no cytochrome c (Δcyc1Δcyc7) for carnitine's protection in the presence of H2O2. (c) CYC3- and both cytochrome c-encoding genes are enhancing carnitine's toxicity in combination with dithiothrietol.
Carnitine protection requires the cytochrome c heme lyase, Cyc3p
Deletion of the cytochrome c heme lyase CYC3 leads to increased sensitivity compared with the wild-type control on plates containing H2O2 (Fig. 5a). In addition, the deletion of CYC3 results in a significant (P<0.05) decrease in the strains' responsiveness to carnitine treatment when compared with mutants with similar H2O2 sensitivity (2.3-fold for Δcyc3, compared with 10-fold for Δtrx2, 35-fold for Δtsa1 and 266-fold for Δsod2 on 0.75 mM H2O2). Cyc3p, a conserved eukaryotic protein, is required for the maturation of the two yeast cytochrome c's, Cyc1p and Cyc7p, through the attachment of prosthetic heme groups (Dumont, 1991; Tong & Margoliash, 1998; Schwarz & Cox, 2002). Deletion of Cyc3p results in a strain lacking mature cytochrome c. Interestingly, CYC3 expression is induced by carnitine and not H2O2, whereas the expression of CYC1 and to a lesser degree CYC7 is induced by the presence by H2O2 (Table 3). To assess whether the observed phenotype is due to the lack of functional cytochrome c, or a separate function of Cyc3p, a Δcyc3 strain was compared with a strain with both cytochrome c-encoding genes, CYC1 and CYC7, deleted. The strains were spotted on plates containing H2O2 and also dithiothrietol, with and without carnitine supplementation. The double mutant's sensitivity to the stress-inducing agents and responsiveness to carnitine supplementation was comparable to that of the Δcyc3 strain in all cases (Fig. 5b). Both strains exhibit increased sensitivity toward H2O2 in comparison with the wild-type strain and a decreased responsiveness to carnitine. Furthermore, both the Δcyc3 and the Δcyc1Δcyc7 double mutant grow slightly better than the wild type when exposed to 16 mM dithiothrietol (Fig. 5c). This difference is enhanced on plates with dithiothrietol and carnitine. The absence of functional cytochrome c therefore suppresses both carnitine-related stress phenotypes, the improvement in the presence of H2O2 and the detrimental impact in the presence of thiol-modifying agents. Expressing CYC3 from the multicopy plasmid, YEpLac195, resulted in complementation of the Δcyc3 strain with regard to sensitivity to oxidants and carnitine responsiveness. However, no effect was observed when the same construct was expressed in the wild-type strain or the Δsod2, Δtrx2 or Δyap1 strains (data not shown).
Shaded blocks are used to highlight upregulation. The respective P-values are indicated in brackets.
Wt, wild type.
A beneficial role for carnitine has been indicated in various human disease states associated with oxidative damage and mitochondrial decay (Calabrese, 2006). The molecular mechanisms behind these observations, however, remain poorly defined. In addition, a role for carnitine in protection against oxidative stress in yeast has been indicated recently (Franken, 2008). This study presents an analysis of the effect of l-carnitine supplementation under oxidative stress conditions and aims to establish the yeast, S. cerevisiae, as a model system for further studies in this field.
Carnitine specifically protected yeast cells from the damaging effects of the ROS-generating oxidants H2O2 and menadione. Interestingly, the effects of both oxidants on global gene expression were shown to be largely identical (Gasch, 2000). No or little effect was observed in the case of CHP and only a marginal difference in the case of LoaOOH-induced stress, which suggests a specific and not general function for carnitine in oxidative stress protection.
Intriguingly, carnitine supplementation in combination with the thiol oxidant diamide or the thiol-reducing agent dithiothrietol resulted in an enhanced toxicity. It is noteworthy that in a yeast deletion library screen of the compounds used in this study, it was reported that deletion of genes involved in mitochondrial function resulted in sensitivity toward H2O2 and the same mutants caused resistance to diamide (Thorpe, 2004). As far as we are aware, this is the first report suggesting a detrimental or a toxic effect of carnitine. This finding may be of significant importance when considering that dithiothrietol as well as carnitine have been proposed as therapeutic agents for the treatment of Alzheimer's and other neurodegenerative diseases (Offen, 1996; Thal, 1996; Marcum, 2005).
For both the protective and the detrimental effects observed under carnitine-supplemented conditions, a correlation to carnitine concentration was observed. In all cases, a carnitine-related effect was observed from concentrations of 100 mg L−1 (0.5 mM) upwards in solid media. In liquid media, cells that were pregrown at concentrations of 1 mg L−1 of carnitine displayed significantly better survival than cells grown in the absence of carnitine, indicating that carnitine exerts protective effects at concentrations found in natural systems. The requirement for higher carnitine levels on solid media can possibly be accounted for by limited carnitine uptake by the carnitine carrier Agp2, which is subject to glucose repression. Carnitine uptake under these conditions was reported to be only ∼5% of carnitine uptake under nonrepressed conditions (Van Roermund, 1999; Stella, 2005). In the same studies, it was also shown that carnitine added at higher concentrations (0.5 mM) could compensate for the deletion of Agp2p, which may account for the concentration dependence of the carnitine phenotypes on solid glucose media. It therefore seems likely that carnitine is required at higher concentrations in solid media to provide a sufficient localized carnitine concentration around the growing colonies.
A protective effect for carnitine against the generation of intracellular ROS has been extensively described in mammalian systems (Sachan, 2005; Savitha, 2005; Binienda, 2006). Using the ROS-sensitive probe DCFH-DA, we show that carnitine indeed protects against ROS formation in S. cerevisiae. Such findings have in some cases been explained by a possible antioxidant activity of carnitine (Calo, 2006), while other studies have indicated no such activity (Rhemrev, 2000). Our data confirm that carnitine has no free radical scavenging activity when used in the ABTS decoloration assay. The 1,1-diphenyl-2-picryl-hydrazyl free radical scavenging assay that was used in a previous study to indicate carnitine's free radical scavenging capacity has been indicated to be a less chemically sound and valid measurement than the ABTS scavenging assay used in this study (Huang, 2005; Gulcin, 2006). When using carnitine in a yeast-based system to screen for antioxidant potential, the molecule exhibited phenotypes strikingly similar to the phenotypes reported previously for known antioxidants. In summary, carnitine performs similar to an antioxidant in a biological context, but is not an antioxidant itself. This suggests the effect of carnitine to be mediated by genetic factors, similar to the activity of the antioxidant selenium (Rayman, 2000).
The data show that carnitine requires the central regulators yap1 or SKN7 as well as genes central to the metabolism of the cellular antioxidant glutathione, namely GSH1 and GLR1, to protect against oxidative stress. Interestingly, carnitine supplementation resulted in an increased sensitivity in the absence of the cell's oxidative stress response in strains with deletions of the regulators of this response, yap1p and Skn7p. This finding further strengthens the assumption that the effect of carnitine is not due to a chemical activity of the compound itself. It is also interesting that this phenotype is similar to that observed in the presence of dithiothrietol and diamide. This suggests that the oxidative stress response controlled by these factors is responsible for suppressing the negative impact of carnitine, and that the target proteins responsible for this negative impact are likely directly or indirectly controlled by those two factors. Furthermore, from these data, it seems likely that although carnitine requires the presence of the cellular oxidative stress protection network, the impact of carnitine may also be directed toward the activity of genes, proteins or networks distinct from the targets of yap1 and Skn7p. The presence of carnitine was, however, able to enhance the growth of strains that did not produce the enzymes Sod2p, Tsa1p and Trx2p, which function in an antioxidant capacity downstream of the regulators, yap1p and Skn7p.
The effect of carnitine supplementation on the global gene expression of peroxide-stressed cultures was evaluated and several genes were found to be differentially expressed, confirming carnitine's involvement in the regulation of gene expression. To establish whether any of the genes upregulated above a 1.8-fold cut-off is required for carnitine-dependent protection, we evaluated representative deletion strains of the identified genes on H2O2- and dithiothrietol-containing plates for responsiveness to carnitine. Of all the strains tested, only deletion of the cytochrome c heme lyase, CYC3, led to a notable reduction in response to carnitine. This was found to be the case for both H2O2 and dithiothrietol. In addition, the effect of CYC3 deletion is similar to the deletion of both cytochrome c's, CYC1 and CYC7, indicating the effect to be caused by the absence of cytochrome c in the Δcyc3 strain. Because deletion of CYC3 leads to a reduced responsiveness to carnitine and multicopy expression has no discernable effect, it can be concluded that CYC3 is required for growth enhancement by carnitine, but is not exclusively involved in the effect of carnitine. Interestingly, both CYC1 and CYC7 are upregulated by H2O2 stress, but CYC3 expression is only induced once carnitine is added to the media. If the function of CYC3 is rate limiting under these conditions, its increased expression would explain its contribution to the protection of carnitine. In general, the integrity of the electron transport chain is required for resistance to H2O2. Increased cytochrome c content by enhancing the expression of CYC3 or CYC1 and CYC7 can specifically be beneficial under oxidative stress conditions by (1) playing a role in the regulation of apoptosis caused by oxidative damage (Eisenberg, 2007), (2) a suggested antioxidant function of cytochrome c itself (Forman & Azzi, 1997; Skulachev, 1998) and (3) its role in the respiratory chain helping to maintain H2O2 and superoxide at a lower physiological level (Zhao, 2003).
Previous data have established a positive effect of carnitine on mitochondrial metabolism. This has been suggested to be caused by the stimulatory effect of the carnitine shuttle on mitochondrial functions (Hagen, 1998; Zhu, 2008). The results of this study present the first report that carnitine may act more specifically through other genetic systems, specifically by inducing the expression of CYC3, which directly impacts on mitochondrial function and also on resistance to oxidative stresses. Future work will focus on further elucidation of the role played by CYC3 in this context and also the identification of upstream factors involved in the regulation of the effects that carnitine has on cellular physiology.
Table S1. Genes upregulated by carnitine and genes of over-represented transcription factors screened on H2O2 and H2O2 with carnitine (1000 mg L−1). Only deletion of CYC3, showed decreased responsiveness to carnitine. The two strains, Δtsa2 and Δhes1, were included because the genes showed significant up-regulation in the microarrays comparing yeast cells in the presence or absence of carnitine without H2O2. Strains highlighted in grey displayed H2O2 sensitivity comparable to the wild type strain, whereas the remaining strains where more sensitive to H2O2 than the wild type strain.
(2003) The genome-wide transcriptional responses of Saccharomyces cerevisiae grown on glucose in aerobic chemostat cultures limited for carbon, nitrogen, phosphorus, or sulfur. J Biol Chem 278: 3265–3274.
(1999) Molecular characterization of carnitine-dependent transport of acetyl-CoA from peroxisomes to mitochondria in Saccharomyces cerevisiae and identification of a plasma membrane carnitine transporter, Agp2p. EMBO J 18: 5843–5452.