Profiling of Cytosolic and Peroxisomal Acetyl-CoA Metabolism in Saccharomyces cerevisiae

As a key intracellular metabolite, acetyl-coenzyme A (acetyl-CoA) plays a major role in various metabolic pathways that link anabolism and catabolism. In the yeast Saccharomyces cerevisiae, acetyl-CoA involving metabolism is compartmentalized, and may vary with the nutrient supply of a cell. Membranes separating intracellular compartments are impermeable to acetyl-CoA and no direct transport between the compartments occurs. Thus, without carnitine supply the glyoxylate shunt is the sole possible route for transferring acetyl-CoA from the cytosol or the peroxisomes into the mitochondria. Here, we investigate the physiological profiling of different deletion mutants of ACS1, ACS2, CIT2 and MLS1 individually or in combination under alternative carbon sources, and study how various mutations alter carbon distribution. Based on our results a detailed model of carbon distribution about cytosolic and peroxisomal acetyl-CoA metabolism in yeast is suggested. This will be useful to further develop yeast as a cell factory for the biosynthesis of acetyl-CoA-derived products.


Introduction
Acetyl-coenzyme A (acetyl-CoA) is involved in many different metabolic pathways and plays a key role in linking anabolism and catabolism. It is a precursor metabolite for de novo biosynthesis of fatty acids and sterols. It is also the end product of lipid degradation by b-oxidation, and it is used for synthesis of C 4 metabolites via the glyoxylate shunt during utilization of carbon sources such as fatty acids and C 2 compounds, e.g. acetate and ethanol. Furthermore, it is a crucial substrate for energy production in the tricarboxylic acid (TCA) cycle. In addition to serving as an important metabolite intermediate in central carbon metabolism, acetyl-CoA is also an important building block for the biosynthesis of many other, biotechnologically relevant metabolites, such as waxes, flavonoids, certain amino acids, carotenoids, polyhydroxyalkanoates, polyketides, butanol and isoprenoids.
In Saccharomyces cerevisiae, acetyl-CoA metabolism takes place in at least four subcellular compartments: nucleus, mitochondria, cytosol and peroxisomes. Nuclear acetyl-CoA serves as acetyl donor for histone acetylation [1], but besides this role in chromatin regulation, acetyl-CoA is mainly important for the central carbon metabolism. Depending on the supply of substrate to the cell, various mechanisms may lead to the formation and utilization of acetyl-CoA. With sugars as carbon source, direct formation of acetyl-CoA from pyruvate catalyzed by the pyruvate dehydrogenase complex (PDH) in the mitochondria serves as fuel for the TCA cycle. On the other hand, acetyl-CoA generated via direct activation of acetate in an ATP-dependent reaction by an acetyl-CoA synthetase (ACS) in the cytosol is the only source for fatty acid and sterol biosynthesis [2]. During growth on oleate, acetyl-CoA is the end product of b-oxidation of straight chain fatty acids, which in yeast exclusively takes place in the peroxisomes [3]. When cells grow on a non-fermentable substrate such as ethanol or acetate, an additional extra-mitochondrial pool of acetyl-CoA is required for the glyoxylate shunt where it is consumed by citrate synthase and malate synthase.
Since the membranes of organelles are impermeable for acetyl-CoA [4], this molecule must either be synthesized within each subcellular compartment where it is required or imported using specific transport mechanisms. Two transport systems have been identified [4]. In the first, the carnitine/acetyl-carnitine shuttle, acetyl-CoA produced in the peroxisomes or the cytosol is converted into acetylcarnitine, which is subsequently transported into the mitochondria. However, it has been shown that S. cerevisiae is not capable of de novo synthesis of carnitine and unless carnitine is supplied with the medium this transport system is non-functional [5]. The second pathway is synthesis of C 4 dicarboxylic acids from acetyl-CoA via the glyoxylate shunt, followed by transport of the C 4 dicarboxylic acids to the mitochondria or the cytosol where they can serve as precursor for formation of pyruvate or phosphoenolpyruvate required for either the TCA cycle or for gluconeogenesis.
A hallmark of eukaryotic cells is the organization of various specific anabolic and catabolic pathways into subcellular compartments. Cast as a central molecule in yeast metabolism, it is not surprising that the biochemical mechanism for the generation and utilization of acetyl-CoA has been studied in detail. However, the compartmental demands for acetyl-CoA vary with the nutrient supply of a cell, and it has been difficult to ascertain the cellular location and function of certain enzymes, and further complicated by multiple isoforms of some of these enzymes related to acetyl-CoA metabolism in S. cerevisiae. This holds in particular for acetyl-CoA synthetase (EC 6.2.1.1), which catalyzes the ligation of acetate and CoA. Two structural genes, ACS1 and ACS2, encoding this enzyme have been characterized in S. cerevisiae [6]. At high concentrations of glucose, ACS1 is transcriptionally repressed, whereas a severe de-repression occurs under glucose limitation or supplementation with non-fermentable carbon sources such as ethanol or acetate. Acs1p was originally thought to be located in the mitochondria [7,8], and possibly also in the peroxisomes [9]. Green fluorescent protein (GFP) tagging however suggested a distribution between the nucleus and the cytoplasm [10]. With respect to substrate specificity Acs1p was shown to have a 30-fold higher affinity to acetate than Acs2p supporting the idea that Acs1p is primarily responsible for acetate activation during growth on C 2 compounds [6]. ACS2 is constitutively expressed irrespectively of the kind of carbon source and type of metabolism (respiratory or fermentative) [2,11]. Thus, it is essential for growth on glucose containing medium. It has also been demonstrated that ACS2 is required for histone acetylation and the biology of the endoplasmic reticulum (ER), the Golgi apparatus and the vacuole [1]. Hence, although Acs2p was predicted to be cytosolic, it may also be present in the nucleus and possibly in the ER with respect to its function in histone acetylation and global transcriptional regulation. To allow for growth on fatty acids, ethanol or acetate, the glyoxylate shunt enzymes play an essential role because of the net synthesis of C 4 metabolites from acetyl-CoA units, which are either the end product of fatty acid degradation or result from activation of C 2 compounds (ethanol or acetate). Of the glyoxylate shunt enzymes, citrate synthase 2 (Cit2p) and malate synthase 1 (Mls1p) catalyze the condensation reaction of acetyl-CoA with oxaloacetate or glyoxylate, respectively. It has been difficult to establish whether all glyoxylate shunt enzymes are localized in the peroxisomes or in the cytosol. Cit2p was firstly determined as extra-mitochondrial by using immunological criteria [12][13][14]. From cellular fractionation and metabolic changes of a deletion mutant, different authors suggested that Cit2p is a cytosolic enzyme [12,15]. However, Lewin and coworkers [16] concluded that it was peroxisomal and involved in the glyoxylate shunt by sedimentation on density gradients after cultivating the yeast cells on an oleate medium. An altering distribution has also been reported for Mls1p. Immunoelectron microscopy in combination with cell fractionation showed that it was peroxisomal when cells were grown on fatty acids, while it remained in the cytosol when yeast utilized C 2 compounds such as ethanol or acetate [17].
Here we map the carbon distribution of acetyl-CoA metabolism in yeast under different nutrient conditions. Among a variety of S. cerevisiae strains, the CEN.PK strain family was selected based on the physiological and genetic properties [18]. We investigate the growth phenotypes and metabolite processing in different mutants affected in acetyl-CoA metabolism. Various location and transport possibilities are discussed under different nutrient situations, and a detailed map for acetyl-CoA metabolism in yeast is proposed.

Materials and Methods
Yeast strains, plasmid construction and gene deletions S. cerevisiae strains and plasmids used in this study are summarized in Table 1. All experiments were performed in the background of CEN.PK 113-5D (MATa SUC2 MAL2-8 c ura3-52; P. Kötter, University of Frankfurt, Germany).
Fragments ACS1-UP/KanMX-UP and KanMX-DOWN/ ACS1-DOWN, respectively, were fused to each other via PCR. The two fusion fragments were integrated into the genome by homologous recombination yielding strain SCIYC02.
For the loop out of the KanMX expression module, procedures were followed as described previously [21]. Subsequently, correct transformants were re-streaked on synthetic complete medium containing 50 mg L 21 uracil and 750 mg L 21 5-fluoroorotic acid (5-FOA) for the loss of the plasmid pSH47 containing the URA3 marker and the reestablishment of uracil auxotrophy. Thus, strain SCIYC05 (acs1D) was obtained.
The same approach was used for all gene deletions in this study. All oligonucleotides used to perform gene deletions can be found in Table S1.
The deletion of each gene was verified by diagnostic PCR, which was performed by using the genomic DNA from each strain as template, and for each gene two pairs of primers were used, one outside of the integration locus and one inside the integrated fragment (see Table S2 for the primers used for strain confirmation).
Growth characterization in liquid cultures was performed in 250 ml baffled shake flasks containing 50 ml of minimal medium with 20 g L 21 glucose as carbon source, which were inoculated at OD 600 0.02 from pre-cultures. Pre-culture cells were grown in YPD medium and washed twice with sterile water. Cultures were incubated at 30uC in a rotary shaker at 180 rpm. Samples were taken during the process and the growth was measured by determining the optical density at 600 nm (OD 600 ). The filtrate was analyzed using a Summit HPLC (Dionex, Sunnyvale, CA, USA) with an Aminex HPX-87H column (Bio-Rad, Hercules, CA, USA) to measure the residual glucose and external metabolites including glycerol, ethanol, acetate, pyruvate, glyoxylate and succinate [22].

Spot assay
Strains were grown to late log-phase in liquid minimal medium at 30uC. The cells were harvested and washed twice with sterile water. After measuring the OD 600 serial dilutions with final cell concentrations of 1610 5 , 1610 4 , 1610 3 , 1610 2 cells per ml were prepared. From each dilution, 5 ml were spotted on minimal medium plates with glucose, acetate, ethanol or glycerol as carbon source. The plates were recorded photographically following 10 days incubation at 30uC.

Results
ACS1 deletion mutant fails to grow on non-fermentable carbon sources while acs2D shows the opposite phenotype Although two acetyl-CoA synthetases in S. cerevisiae, Acs1p and Acs2p, have been studied well for their function, biochemical properties and transcription regulation, there are conflicting reports on the growth phenotype of ACS1 deletion mutants [9,11] and some of media used before were not devoid of carnitine (using complex medium) [11]. Therefore, ACS1 and ACS2 were deleted in the strain CEN.PK 113-5D to generate two mutant strains, SCIYC05 (acs1D) and SCIYC23 (acs2D), respectively. The growth phenotypes of the mutant strains were re-characterized by a spotting assay on synthetic minimal medium with different carbon sources together with the reference strain CEN.PK 113-5D (Fig. 1). SCIYC05 (acs1D) showed reduced growth with glucose as carbon source compared with the reference strain. However, it failed to grow on acetate, ethanol, or glycerol as sole carbon source. This might be due to the fact that: (i) acetyl-CoA produced by Acs2p cannot enter the peroxisomes, and hence cannot enter the glyoxylate shunt which will lead to formation of C 4 units and from here further to glycolytic intermediates required for cell growth; or (ii) the Acs2p activity is at an insufficient rate to provide the C 4 units needed for gluconeogenesis and other biosynthetic pathways to allow cells to utilize C 2 compounds as sole carbon source. In contrast, the ACS2 deletion strain SCIYC23 (acs2D) was unable to grow on glucose media. But it grew normally on nonfermentable carbon sources (acetate, ethanol and glycerol) compared with the reference strain.
To further examine the physiological features of the acs1D strain, liquid growth assays were conducted. SCIYC05 (acs1D) and the reference strain were cultured in shake flasks with minimal media containing 20 g L 21 glucose as carbon source. As shown in Fig. 2A and B, SCIYC05 (acs1D) showed an extended lag-phase, a lower growth rate, and a lower final biomass concentration during the glucose phase compared to the reference strain. The extended lag phase for an acs1 mutant during growth on glucose has been observed in previous studies as well [11]. Although ACS1 is reportedly subjected to glucose repression [6], the strongly delayed glucose consumption indicates further that Acs1p may have other function as well, e.g. Acs1p is probably involved in chromatin regulation since a nuclear localization has been reported as well [10]. After glucose depletion, SCIYC05 (acs1D) stopped growing, while the reference strain continued growing. HPLC analysis of metabolites revealed that SCIYC05 (acs1D) did not utilize ethanol and glycerol, which were produced during growth on glucose ( Fig. 2B); whereas the reference strain could metabolize all nonfermentable carbon sources derived from overflow metabolism.
Deletion of CIT2 results in differences in growth on solid and liquid culture with non-fermentable carbon sources In S. cerevisiae, the citrate synthase isoenzymes Cit1p and Cit3p have been shown to be exclusively located in the mitochondria, whereas it has not been conclusively shown whether Cit2p is present in the cytoplasm or in the peroxisomes. Furthermore, there are conflicting reports in the literature on the growth phenotype of cit2D mutant strains [4,13,15,23]. To investigate the effect of CIT2 disruption on cell growth and metabolism strain SCIYC06 (cit2D) was re-characterized on minimal medium with glucose, glycerol, ethanol or acetate as carbon source. As shown in Fig. 1, SCIYC06 (cit2D) did not display any severe difference in growth on glucose, but a slight growth reduction on nonfermentable carbon sources compared to the reference strain. As carnitine is not present in the media, meaning that the carnitine shuttle is not active, these results imply that CIT2 is dispensable for growth on C 2 carbon sources, which is in contrast to the study by Swiegers et al. [23].  To reinforce the observation that CIT2 deletion still allows cell growth, SCIYC06 (cit2D) as well as the reference strain were tested in liquid culture with minimal medium including 20 g L 21 glucose as carbon source. SCIYC06 (cit2D) did not exhibit any growth difference when utilizing glucose as carbon source, whereas it failed to continue growing when glucose was depleted (Fig. 2C). As shown in Fig. 2C, metabolite analysis of the shake flask cultivation with SCIYC06 (cit2D) revealed that acetate was formed while ethanol was being consumed. This conversion results in net NAD(P)H production in the cells and hence allows the cells to generate ATP. The phenomenon of accumulating acetate, however, indicates that despite the presence of intact ACS1 and ACS2 the cells were not able to further catabolise acetyl-CoA resulting in acetate accumulation and no biomass formation.

MLS1 deletion affects the utilization of non-fermentable carbon sources
In yeast, two enzymes, Mls1p and Dal7p, have a 81% protein sequence identity to each other. Dal7p is thought to be involved in the degradation of allantoic acid to urea, whereas only MLS1 is highly transcribed upon growth on non-fermentable carbon sources [24]. Hence, malate synthase Mls1p is believed to be specifically involved in the glyoxylate shunt necessary for the utilization of C 2 carbon sources. Although Mls1p ends with a Cterminal SKL tripeptide representing a peroxisomal targeting signal PTS1, selective distribution to the peroxisomes or the cytosol depending on the carbon source has been reported [17]. As expected for the MLS1 deletion strain, no significant changes in phenotype were observed on plate assays with glucose as carbon source (Fig. 1). In contrast, SCIYC07 (mls1D) was unable to grow on minimal media with glycerol, ethanol or acetate as sole carbon sources. Our results are consistent with previous work [17]. It was therefore confirmed that MLS1 is indispensable for the utilization of non-fermentable carbon sources. This also applies to the other key enzyme (Icl1p) unique to the glyoxylate shunt (data not shown).
In liquid culture, while SCIYC07 (mls1D) behaved similar during glucose as carbon source; the growth of SCIYC07 (mls1D) was remarkably slow after depletion of glucose, compared with the reference strain ( Fig. 2A and 2D). However, extracellular metabolite analysis revealed that ethanol was almost completely metabolized, and after glucose consumption acetate first increased slightly until 24 hours after inoculation, and then decreased slowly. The concentration of glycerol was also shown to decrease very slowly, from 1.14 g L 21 to 0.53 g L 21 after glucose depletion. Comparing the consumption of (non-fermentable) carbon sources between SCIYC06 (cit2D) (Fig. 2C) and SCIYC07 (mls1D) (Fig. 2D), it was found that SCIYC06 (cit2D) metabolized ethanol to form acetate and the biomass concentration did not increase. On the other hand SCIYC07 (mls1D) consumed both C 2 carbon sources, but still the OD 600 only increased by 1.46, compared to the reference strain, for which the OD 600 increased by 6.15 ( Fig. 2A). This raised the question why both the accumulated ethanol and acetate were consumed, while only a small increase in biomass was observed.
Although MLS1 was deleted, isocitrate lyase was still active in these cells producing glyoxylate, which has been reported to be a toxic intermediate for mammalian cells [25] and bacteria [26]. To test the possibility of glyoxylate accumulation, extracellular glyoxylate and pyruvate were analyzed for SCIYC07 (mls1D) as well as reference strain. No accumulation of glyoxylate or pyruvate was detected for the reference strain. For SCIYC07 (mls1D), as shown in Fig. 3, when cells started to consume both C 2 carbon sources, glyoxylate and pyruvate accumulated. Glyoxylate reached a concentration of 0.4 g L 21 at 40 hours after inoculation, then decreasing a little, whilst pyruvate continuously increased from the time point of glucose depletion to the end of the cultivation.

CIT2 deletion restores growth in an acs1D strain
It was observed that deleting ACS1, CIT2 or MLS1, which are related to the acetyl-CoA metabolism, led to different changes in carbon source utilization. To investigate if the double deletion or triple deletions of these three genes cause different phenotypic changes in growth characteristics compared with single deletion mutants, we constructed the mutants SCIYC14 (acs1D cit2D), SCIYC15 (cit2D mls1D) and SCIYC16 (mls1D acs1D). Repeating the ACS1 deletion in SCIYC15 (cit2D mls1D) led to the triple deletion strain SCIYC21 (acs1D cit2D mls1D).
As shown in Fig. 4, no apparent growth was exhibited in the double deletion strains SCIYC15 (cit2D mls1D), SCIYC16 (mls1D acs1D), and the triple deletion mutant SCIYC21 (acs1D cit2D mls1D), supplemented with non-fermentable carbon sources. However, SCIYC14 (acs1D cit2D) was surprisingly able to grow on glycerol, ethanol or acetate as sole carbon source. Although the growth was slower compared to the reference strain, it clearly showed growth was restored in the acs1D and cit2D double deletion strain, whereas acs1D single deletion led to an inability to grow on such carbon sources.

Truncated Mls1p suppresses the growth defects of the acs1D mutant
The inability of the acs1D single deletion strain to grow on C 2 carbon source was overcome by further deleting another gene (CIT2) related to the acetyl-CoA metabolism. This phenotype implied that our previous hypothesis on the inefficiency of Acs2p in SCIYC05 (acs1D) was incorrect. This phenotypic change also indicates the possibility of flexible distribution of glyoxylate shunt enzymes. One possible assumption to explain these phenotypes would be that on non-fermentable carbon sources both Cit2p and Mls1p are localized in the peroxisomes when ACS1 is deleted, whereas Mls1p mainly stays in the cytosol in the case of the acs1D and cit2D double deletion strain. To validate this prediction, the plasmid pIYC15 encoding a truncated Mls1p (without the -SKL tripeptide), which is supposed to be located in the cytoplasm, was transformed into SCIYC07 (mls1D) and SCIYC05 (acs1D), yielding strains SCIYC28 and SCIYC29, respectively. Harboring the plasmid pIYC15 restored the growth of SCIYC07 (mls1D) on nonfermentable carbon sources, which confirmed that expressing a cytosolic malate synthase was efficient for maintaining a functional glyoxylate shunt (Fig. 5). As shown in Fig. 5, also SCIYC29 expressing the truncated Mls1p was able to grow on glycerol, ethanol or acetate as sole carbon source. As a control, the wild type Mls1p was also expressed in SCIYC07 (mls1D) and SCIYC05 (acs1D), resulting in strains SCIYC35 and SCIYC36, respectively. While SCIYC35 grew well on ethanol or acetate, SCIYC36 was unable to grow on C 2 carbon sources (Fig. 5). Taking into account of the growth difference of SCIYC29 and SCIYC36, the results indeed support the hypothesis that both Cit2p and Mls1p are localized in the peroxisomes during growth on non-fermentable carbon sources when ACS1 is deleted.

Discussion
Acetyl-CoA is a key intermediate in cellular metabolism. It is involved as a precursor metabolite; it is used for generation of Gibbs free energy, and for production of many industrially relevant compounds. In the yeast S. cerevisiae metabolism involving acetyl-CoA is compartmentalized and may vary with the nutrient supply of a cell. Where acetyl-CoA is metabolized on nonfermentable carbon sources is not fully understood, but as S. cerevisiae cannot synthesize carnitine and we did not provide this via the medium, the glyoxylate shunt is the sole possible route for transferring acetyl-CoA (in form of tricarboxylic acid cycle intermediates) into the mitochondria. Cit2p, Mls1p, Acs1p and Acs2p, are the main enzymes relevant to acetyl-CoA metabolism-either consuming or producing this molecule outside of mitochondria. Therefore, the aim of this study was to investigate the physiological effects of different mutant strains with combinations of these genes deletion (ACS1, ACS2, CIT2 and MLS1) on different carbon sources in S. cerevisiae, and try to understand intracellular transport of intermediates in various mutations with alternative carbon sources.
Irrespective of the nutrient supply to the cell the cytosolic acetyl-CoA pool is indispensable for synthesis of fatty acids and sterols, and it relies completely on acetyl-CoA synthetase activity on all carbon sources. In view of its essential role in the central carbon metabolism, at least one functional isoenzyme of acetyl-CoA synthetase must be located in the cytosol. SCIYC23 (acs2D) lacking ACS2 can grow on minimal medium containing a non-fermentable carbon source like ethanol or acetate, which clearly indicates that Acs1p is subjected to cytosolic distribution under these conditions. If Acs1p was exclusively localized in the peroxisomes, it could not activate acetate to form acetyl-CoA in the cytosol; and acetyl-CoA produced in the peroxisomes cannot be shuttled without the carnitine transport system. Our findings for the acs2D mutant are consistent with earlier reports by van den Berg and Steensma [11], i.e. that inactivation of ACS2 did not cause inability to grow on Figure 4. Growth phenotypes of double deletion and the triple deletion strains. Plate assay for SCIYC14 (acs1D cit2D), SCIYC15 (cit2D mls1D) and SCIYC16 (mls1D acs1D), and the triple deletion strain SCIYC21 (acs1D cit2D mls1D) on minimal medium with glucose, glycerol, ethanol or acetate as carbon source. The plates were recorded photographically after 10 days incubation at 30uC. doi:10.1371/journal.pone.0042475.g004 Figure 5. Growth phenotypes of mls1D and acs1D strains expressing an SKL-less or intact Mls1p variant. Cells were grown on minimal medium supplemented with glucose, glycerol, ethanol or acetate as carbon source. Uracil (20 mg ml 21 ) was added as required (reference, SCIYC05 (acs1D) and SCIYC07 (mls1D)). The plates were recorded photographically after 10 days of incubation at 30uC. doi:10.1371/journal.pone.0042475.g005 media with acetate or ethanol as sole carbon source (in complex medium with yeast extract). Since the media used in our study are devoid of carnitine it further confirms that Acs1p is distributed to the cytosol on a non-fermentable carbon source. The ability of acs2D mutants to grow in aerobic, glucose-limited chemostat cultures has also been reported [6], which further supports the conclusion that Acs1p is subjected to cytosolic distribution under de-repressing conditions. Even though it is reported that Cit2p has both N-and C-terminal signal for mitochondrion and peroxisomes, the potential of N-terminal signal can only be recognized by removing the C-terminal sequence [27]. Taking this into account together with that recently reported CIT2 is in peroxisomes regardless of synthetic or complex medium [28], strongly indicates that Cit2p is the peroxisomal isoenzyme of citrate synthase that is involved in the glyoxylate shunt, and it therefore seems likely that Acs1p functions in the peroxisomes in synthesis of acetyl-CoA required for use by Cit2p. Furthermore, according to the results of Kunze et al. [17], Mls1p remains in the cytosol in ethanol-grown cells and it therefore seems reasonable that there is a dual distribution of Acs1p for cells growing under non-repressing conditions. Similarly, it was also recently shown that fumarase and aconitase were dually localized in the mitochondria and the cytoplasm [29,30]. Overall, Acs1p is suggested to show a dual distribution pattern between the cytoplasm and the peroxisomes, as shown in Fig. 6A.
The scenario in Fig. 6A does probably not apply to the acs1D strain. For mutant SCIYC05 (acs1D), the inability to grow on a non-fermentable carbon source was observed. Besides the requirement of 1.8 mmol cytosolic acetyl-CoA per gram cells for lipid synthesis in the cytosol [31], metabolism on C 2 carbon sources passes through acetyl-CoA for both biosynthesis and energy metabolism. Therefore, inefficient conversion of C 2 compounds (ethanol or acetate) to acetyl-CoA could be explained by low Acs2p activity, which may explain the acs1D strain's inability to grow on non-fermentable carbon sources. However, the double mutant strain lacking both ACS1 and CIT2 was shown to be able to grow on non-fermentable carbon sources, which contradicts this hypothesis. Furthermore, expressing the truncated  but not the intact Mls1p in the cytoplasm of SCIYC05 (acs1D) restored the growth capacity. This observation strongly suggests that both Cit2p and Mls1p are located in the peroxisomes of cells lacking ACS1 (Fig. 6B), and low Acs2p activity therefore cannot explain the inability of growth of the acs1D strain.
Outside the mitochondria, acetyl-CoA synthetases serve as the main producers of acetyl-CoA, while citrate synthase 2 catalyzes one of the acetyl-CoA consuming reactions. Deletion of CIT2 breaks the integrity of the glyoxylate shunt. However, in our study the cit2D strain SCIYC06 (cit2D) was shown to be able to grow on minimal media plates containing ethanol, acetate or glycerol as sole carbon source. As shown in Fig. 6A and 5C, the disruption in the glyoxylate shunt resulting from the CIT2 deletion can be bypassed by (i) succinate being imported into the mitochondria followed by decarboxylation of malate to pyruvate via malic enzyme (as shown pathway I in Fig. 6A); or (ii) malate being converted by cytosolic malate dehydrogenase Mdh2p to form oxaloacetate, followed by synthesis of phosphoenolpyruvate (PEP) involving PEP carboxykinase (as shown pathway II in Fig. 6A). Thus, citrate or isocitrate from the mitochondria might bypass the Cit2p reaction in the glyoxylate shunt. This is supported by the fact that the mitochondrial citrate synthase Cit1p could act as a shunt for the missing Cit2p via the citrate transporter Ctp1p [32]. Although the discrepancy of growth phenotype on solid and liquid medium is unclear so far, the growth conditions are different, which might lead to the different physiology.
Being an essential enzyme, elimination of Mls1p also breaks the integrity of the glyoxylate shunt. According to the plate assay, the MLS1 deletion strain SCIYC07 (mls1D) was unable to grow on non-fermentable nutrient sources, indicating that Mls1p is specific and absolutely required for growth on C 2 compounds as carbon source. Surprisingly, this mutant utilized ethanol and acetate derived from the overflow metabolism during growth on glucose in shake flask cultivation, while the biomass was increasing only slightly. After conversion of C 2 carbon sources to acetyl-CoA, this compound can be metabolized by different pathways such as the glyoxylate shunt, or the biosynthesis of fatty acids and sterols. Analysis of the extracellular metabolite profile revealed that glyoxylate accumulated. This observation indeed showed that in absence of MLS1, the glyoxylate shunt was working partially, which is also consistent with the phenomenon that all metabolites from overflow metabolism are slowly consumed. These results suggested an alternative pathway to the activities of Mls1p: the counter-exchange of succinate and fumarate and subsequent intramitochondrial conversion of succinate to fumarate, together with the canalization of fumarate to malate. This is supported by the fact of that the mitochondrial succinate-fumarate carrier Sfc1p is indispensible for acetate utilization on minimal medium [28].
In contrast to the inability of SCIYC05 (acs1D) to grow on nonfermentable carbon sources, SCIYC14 (acs1D cit2D) was observed to be able to grow under such carbon conditions. Although the membranes of subcellular organelles are permeable to C 2 compounds such as ethanol and acetate, metabolism on C 2 carbon source has to pass through acetyl-CoA for biosynthesis and energy requirement. In SCIYC14 (acs1D cit2D), the conversion of a C 2 carbon source to acetyl-CoA can be realized only by the sole extra-mitochondrial producer, cytosolic Acs2p. The acetyl-CoA units so produced cannot be transported into other compartments without the help of carnitine. Furthermore, the resulting acetyl-CoA units in the cytosol must be channeled into the synthesis of C 4 dicarboxylic acids. Taken together with the deficiency of Cit2p activity, Mls1p which is the only candidate enzyme left can condense acetyl-CoA to form C 4 metabolites, may well be localized in the cytoplasm of SCIYC14 (acs1D cit2D) (Fig. 6D).
In this study, we have investigated the physiological profiling of S. cerevisiae mutants affected in acetyl-CoA metabolism. Through analyzing the growth phenotypes and metabolite processing changes in different mutants on non-fermentable carbon sources, we have proposed a model of acetyl-CoA metabolism to speculate how various mutations alter carbon distribution. This study should be of interest to those studying acetyl-CoA production in eukaryotic cells and central carbon metabolism in fungi.