The facC Gene of Aspergillus nidulans Encodes an Acetate-Inducible Carnitine Acetyltransferase

Abstract

Mutations in the facC gene of Aspergillus nidulans result in an inability to use acetate as a sole carbon source. This gene has been cloned by complementation. The proposed translation product of the facC gene has significant similarity to carnitine acetyltransferases (CAT) from other organisms. Total CAT activity was found to be inducible by acetate and fatty acids and repressed by glucose. Acetate-inducible activity was found to be absent in facC mutants, while fatty acid-inducible activity was absent in an acuJ mutant. Acetate induction of facC expression was dependent on the facB regulatory gene, and an expressed FacB fusion protein was demonstrated to bind to 5′ facC sequences. Carbon catabolite repression of facC expression was affected by mutations in the creA gene and a CreA fusion protein bound to 5′ facC sequences. Mutations in the acuJ gene led to increased acetate induction of facC expression and also of an amdS-lacZ reporter gene, and it is proposed that this results from accumulation of acetate, as well as increased expression of facB. A model is presented in which facC encodes a cytosolic CAT enzyme, while a different CAT enzyme, which is acuJ dependent, is present in peroxisomes and mitochondria, and these activities are required for the movement of acetyl groups between intracellular compartments.

Acetate and other two-carbon compounds, such as ethanol, are capable of acting as sole carbon sources for many microorganisms. Gluconeogenesis is required, and this leads to depletion of citric acid cycle intermediates. These are replenished by the anaplerotic glyoxalate bypass in which the action of the enzymes isocitrate lyase (ICL) and malate synthase (MS) results in the net production of intermediates from acetyl coenzyme A (acetyl-CoA) (5, 46). The glyoxalate bypass is present in bacteria, plants, and fungi, as well as some animals (reviewed in reference 19). ICL and MS are located in microbodies termed peroxisomes or glyoxysomes, depending on whether they contain the enzyme catalase (42, 69).

Therefore, for acetate utilization, three different subcellular locations are required—the cytosol, the mitochondrion, and the peroxisome. The activation of acetate to acetyl-CoA by acetyl-CoA synthase (ACS) occurs in the cytosol, as shown in Saccharomyces cerevisiae (43). Acetyl-CoA is imported into the mitochondrion, where it is metabolized via the citric acid cycle. A shuttle mechanism involving carnitine acetyltransferase (CAT) forming acetylcarnitine from carnitine and acetyl-CoA in a reversible reaction is required, since the outer mitochondrial membrane is permeable to acetylcarnitine but not to acetyl-CoA (45, 64). In addition, it has been recently shown that the peroxisomal membrane of S. cerevisiae is also impermeable to acetyl-CoA and that the acetylcarnitine shuttle is involved in the entry of acetyl-CoA into the peroxisome (70).

Acetate utilization mutants of a number of fungi have been isolated, e.g., Neurospora crassa (30, 31), S. cerevisiae (reviewed in reference 51), and Coprinus cinereus (9, 50). In Aspergillus nidulans, direct screening for acetate mutants led to the identification of twelve acu genes (3) and a further gene was identified by propionate resistance (62). In addition, acetate nonutilization mutants carrying mutations in the facA, facB, and facC genes have been isolated by resistance to fluoroacetate, which is toxic by virtue of conversion to fluorocitrate, an inhibitor of aconitase (2). The facA gene has been shown to encode ACS (13, 60). Cloning and characterization of the facB gene showed that it encodes a regulatory protein with a Zn(II)Cys binuclear cluster DNA binding domain (38, 65) which is involved in acetate induction of ACS (facA), ICL (acuD), and MS (acuE), as well as in induction of acetamidase (amdS) and NADP-specific isocitrate dehydrogenase (33, 66).

Some fungal mutants unable to grow on acetate are also unable to utilize fatty acids as sole carbon sources (3, 27). Selection of A. nidulans mutants unable to grow on fatty acids resulted in the isolation of new alleles of previously identified acu genes (20, 41). Fatty acids are first converted to their CoA esters and then, via the cyclic β-oxidation pathway, to acetyl-CoA (32, 42, 69). This pathway takes place exclusively in the peroxisomes, and fatty acids induce peroxisome biogenesis.

Since acetyl-CoA is produced by β-oxidation, enzymes of the glyoxalate bypass and of gluconeogenesis are required for growth on fatty acids as sole carbon sources. In A. nidulans, the acuJ gene has been shown to be required for growth on fatty acids (such as Tween 80 and oleate), as well as on acetate, and to affect CAT activity in response to fatty acid induction (53). This suggests that the formation of acetylcarnitine in the peroxisome is likely to be required for shuttling of acetyl groups into the cytosol and into the mitochondrion. The only acetate-specific mutations identified in A. nidulans are in the facA, facB, and facC genes (3). The growth responses of relevant A. nidulans mutants are summarized in Table 1, and the proposed pathways and enzyme locations are shown in Fig. 1.

TABLE 1.

Growth properties of relevant mutants^a

Genotype	Function	Growth with following sole carbon source:				Reaction to fluoroacetate
		Acetate	Butyrate	Tween 80	Oleate
Wild type		+	+	+	+	S
facA303	ACS	−	+	+	+	R
facB101	Regulatory gene	−	+	+	+	R
facC102	CAT	−	+	+	+	R
acuD306	ICL	−	−	−	−	S
acuE215	MS	−	−	−	−	S
acuJ211	CAT (?)	−	−	−	−	S

^aSole carbon sources were added to minimal medium with ammonium present as the nitrogen source. +, utilization of carbon source; − no utilization. Fluoroacetate was added to medium containing glycerol as the carbon source (see reference 2). R, resistant; S, sensitive. Results are from references 2, 3, and 53 and this study. The functions of the products of the relevant wild-type genes are shown. In the case of acuJ, it has not been demonstrated that this gene directly encodes a CAT. 

FIG. 1.

Proposed locations and functions of enzymes involved in fatty acid and acetate utilization in A. nidulans. The model for the role of CAT enzymes is presented in this paper. During growth on acetate, facC-encoded CAT activity in the cytosol produces acetylcarnitine, which enters peroxisomes and mitochondria, where acuJ-dependent CAT activity results in the formation of acetyl-CoA which is metabolized via the glyoxalate bypass and the citric acid cycle. During growth on fatty acids, acetyl-CoA formed by β-oxidation in the peroxisome is converted to acetylcarnitine by acuJ-dependent CAT activity and then can be shuttled to the mitochondria for metabolism via the citric acid cycle. Therefore, acuJ-dependent activity is required for growth on both fatty acids and acetate, while facC-dependent activity is required only for growth on acetate. A question mark adjacent to AcuJ indicates that it has not been demonstrated that the acuJ directly encodes a CAT activity.

The function of the facC gene has not been determined. The phenotypes of fluoroacetate resistance and acetate nonutilization in facC mutants clearly indicate an essential role in the conversion of acetate to citrate. However, facC mutants have not been found to have a specific enzyme defect (3, 33). We have now cloned and characterized the facC gene. The predicted polypeptide shows extended similarity to CAT proteins of eukaryotes and is most similar to one of two CAT enzymes (YAT1p) present in S. cerevisiae (44, 61). Mutations in the acuJ gene of A. nidulans lead to reduced CAT activity (53). We present evidence that there are two CAT activities present in A. nidulans—one that is encoded by facC, is inducible by acetate via FacB, and is predicted to be cytosolic and one that is dependent on acuJ, is inducible by fatty acids, and is predicted to be located in the mitochondria and the peroxisomes. Furthermore, we have demonstrated that both of these enzymes are subject to glucose repression mediated by the creA gene (21, 22) and demonstrated binding of expressed FacB and CreA fusion proteins to fragments of the 5′ untranslated region of facC containing predicted binding sites.

MATERIALS AND METHODS

Strains, media, and growth conditions.

The A. nidulans strains used in this work are shown in Table 2. The A. nidulans media and growth conditions used were described by Cove (16). Nitrogen sources were added to a final concentration of 10 mM. Carbon sources were generally added to 1% (wt/vol), except for acetate, which was added to 50 mM, and Tween 80, which was added to 0.5% (vol/vol). Growth of mycelium for DNA, RNA, and enzyme extracts was in 100 ml of medium in 250-ml Erlenmeyer flasks shaken at 37°C.

TABLE 2.

A. nidulans strains used in this study

Strain	Genotypea
MH0001	biA1
MH0058	pyroA4 facA303 nicB8 riboB2
MH0664	creA204 biA1 niiA4
MH0764	wA3 riboB2 facB101
MH2681	amdS368 prnA457 fwA1 niiA4
MH3018	pabA1 yA2 argB1
MH5840	biA1 amdS::lacZ pyroA4 niiA4 facB::Ble riboB2
MH7065	pyroA4 facC102 riboB2
MH8091	pyroA4 facC301 riboB2
MH8352	acuJ211 pabA1 riboB2
MH8361	pabA1 yA2 riboB2
MH8377	argB3 pyroA4
MH8462	acuJ211 amdS::lacZ facB::Ble riboB2
MH8463	acuJ211 ΔfacC::argB

^aThese gene designations are described by Clutterbuck (12), except for amdS::lacZ, which represents a translational fusion of lacZ to the amdS gene integrated by gene replacement at the amdS locus (18); facB::Ble, which represents an insertion of a bleomycin resistance gene inactivating the facB gene (65), and ΔfacC::argB, which represents an insertional replacement of facC sequences with the argB⁺ gene (Fig. 2B). 

General methods.

A. nidulans strains were transformed by the method of Andrianopoulos and Hynes (1). Plasmids were prepared for transformation by centrifugation on a cesium chloride gradient or by Magic Preps (Promega), Wizard Preps (Promega), or HPP columns (Boehringer Mannheim). Genetic manipulations were carried out as described by Clutterbuck (11). Genomic DNA was isolated from mycelium by the method of Lee and Taylor (48). Total RNA was isolated by using the FastRNA RED method (Bio 101, Inc.) in accordance with the manufacturer’s instructions.

DNA for Southern blot analysis was separated by electrophoresis through 1.0% agarose gels and transferred to Hybond N⁺ membrane (Amersham) by alkaline transfer (0.4 M NaOH) for 3 to 4 h as recommended by the manufacturer. Prior to transfer, the DNA was depurinated by treatment of the gels with 0.25 M HCl for 10 min.

RNA was separated through 1.2% agarose containing 0.6 M formaldehyde in 1× morpholinepropanesulfonic acid (MOPS) buffer (20 mM MOPS, 5 mM sodium acetate, 1 mM Na₂EDTA, pH 7.0). RNA samples were mixed in a solution of 0.5% formamide, 37% formaldehyde, 1× MOPS buffer, 1× loading dye, and 10-mg/ml ethidium bromide and heated to 68°C prior to loading. RNA was transferred to Hybond N⁺ (Amersham) by alkaline transfer in 0.04 M NaOH for 2 to 3 h.

DNA fragments were radioactively labelled for hybridization by using standard random hexanucleotide priming procedures using the Klenow fragment of DNA polymerase I (Promega Corp.) and [α-³²P]dATP (3,000 Ci/mmol; Bresatec).

For both Southern and Northern blot analyses, filters were prehybridized (at least 30 min) and hybridized with a labelled DNA probe overnight at 42°C in a solution of 50% formamide, 4× SSPE (1× SSPE is 0.18 M NaCl, 10 mM NaH₂PO₄, and 1 mM EDTA [pH 7.7]), 1% sodium dodecyl sulfate (SDS), 5% BLOTTO (10% skim milk powder), and 100-μg/ml sonicated herring sperm DNA. Filters were washed with 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)–0.1% SDS for 30 min at 42°C and then with 0.1× SSC–0.1% SDS for 30 min at 68°C, except where noted otherwise. Filters were then subjected to autoradiography.

DNA sequencing.

DNA fragments were subcloned into the pBluescriptSK+ or pBluescriptSK− phagemid (Stratagene) for sequencing. Single-stranded DNA was prepared for sequencing as described by Sambrook et al. (59). Double-stranded plasmid DNA was prepared for sequencing by polyethylene glycol precipitation or by use of an HPP column (Boehringer Mannheim). Automated sequencing was performed by using Dye Primer 373A DNA Cycle Sequencing Kits and Dye Terminator 373A DNA Cycle Sequencing Kits (Applied Biosystems, Inc.). The specific primers used in sequencing were 5′ CTC ATC CAC AAT CTC ACC 3′, 5′ CCA AGC GAG AGA GCTAGG 3′, and 5′ ATG GATCGTGTGCATGGC 3′.

A. nidulans libraries.

A chromosome VIII-specific A. nidulans genomic library (8) was kindly supplied by M. Katz (University of New England, Armidale, New South Wales, Australia), and an A. nidulans cDNA library in a λgt10 vector was provided by G. S. May (Baylor College of Medicine, Houston, Tex.).

EMSAs.

Probes for use in electrophoretic mobility shift assays (EMSAs) were labelled by end filling 5′ overhangs with the Klenow fragment of DNA polymerase I (Promega Corp.) and [α-³²P]dATP (3,000 Ci/mmol; Bresatec). They were then purified by electrophoresis on a 4% nondenaturing polyacrylamide gel in 1× Tris-borate-EDTA (TBE). Labelled fragments were localized by autoradiography, excised, eluted in dialysis tubing by electrophoresis in 0.2× TBE, and precipitated by standard techniques (59). Binding reaction mixtures comprised protein extract, labelled DNA probes, 1 μg of poly(dI-dC), and 1× binding buffer (25 mM HEPES · KOH [pH 7.6], 40 mM KCl, 1 mM EDTA, 50% glycerol) in 20-μl volumes. Binding reaction mixtures were incubated for 20 min at 25°C. DNA binding reaction mixtures were electrophoresed on 4% nondenaturing polyacrylamide gels containing (per 100 ml) 13.3 ml of a 30% acrylamide mixture (29:1 acrylamide-bisacrylamide ratio), 20 ml of 5× TBE, 66.2 ml of deionized H₂O, 0.52 μl of 10% ammonium persulfate, and 80 μl of N,N,N′,N′-tetramethylethylendiamine in 1× TBE at 100 to 150 V and 4°C (58). Gels were dried under a vacuum prior to autoradiography.

CAT assays.

Protein extracts were prepared by grinding 0.2 g of mycelium with glass beads under 1 ml of 50 mM Tris · HCl (pH 7.5) and centrifuging the resulting slurry at maximum speed in a microcentrifuge. CAT assays were done as described by Kawamoto et al. (40). The reaction was monitored spectrophotometrically at 30°C by monitoring the release of CoA-SH from acetyl-CoA using 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB). The reaction mixture contained 40 mM KH₂PO₄ (pH 8.0), 0.05 mM acetyl-CoA, 0.12 mM DTNB, and 0.1 ml of crude extract. The reaction was initiated by the addition of 2.2 mM dl-carnitine (β-hydroxy-γ-trimethylammonium butyrate) chloride, and the final reaction volume was 1.5 ml. A blank was used to which no carnitine was added, as a slight background rate of carnitine-independent release of CoA-SH from acetyl-CoA was observed. Protein concentrations were determined by using the Bio-Rad Protein Assay reagent, and CAT activities are expressed as nanomoles of CoA-SH produced per minute per microgram of protein, assuming an extinction coefficient of 13,600/M/cm for the chromophore formed from DTNB (26).

Nucleotide sequence accession number.

The GenBank accession number of the sequence presented here (see Fig. 3) is AF023156.

FIG. 3.

Nucleotide and deduced amino acid sequences of the facC gene. Nucleotide coordinates are given with respect to the start of translation, and introns are in lowercase. The ends of the cDNA are indicated by double underlining of the terminal bases.

RESULTS

Cloning and sequencing of the facC gene.

The veA and facC genes are 6 map units apart on chromosome VIII (12). A cosmid containing the veA gene on a 4-kb XhoI fragment was provided by L. Yager (Temple University). This fragment was used to probe a chromosome VIII-specific cosmid library (8), and cosmid SW17D08 was identified among a number hybridizing to the probe. Subsequently, this cosmid was shown to complement the facC102 mutation for growth on acetate in cotransformation experiments with a riboB2 facC102 strain (MH7065) by using riboB⁺ plasmid pPL3 (55) and selecting initially for RiboB⁺ transformants. The veA⁺-containing cosmid was then also found to complement the facC102 mutation.

Subcloning of cosmid SW17D08 yielded a 10-kb BamHI fragment in pUC13^cmr (Fig. 2A) capable of complementing the facC102 and facC301 alleles for growth on acetate in cotransformation experiments with pPL3. Restriction mapping (Fig. 2A) and Southern blotting of genomic DNA showed that the cloned DNA was unique and was not rearranged. Low-stringency Southern analysis (hybridization in 30% formamide and washing in 0.5× SSC at 37°C) failed to yield hybridizing bands other than those predicted. The 4-kb KpnI fragment (Fig. 2A) subcloned into pBluescriptSK⁺ (pCS3875) was found to be capable of complementing facC102 and was sequenced. This KpnI fragment was used as a probe to isolate two cDNA clones from a cDNA library in λgt10 (obtained from G. S. May). These were found to be identical, and sequencing revealed the 5′ and 3′ endpoints shown in Fig. 2A.

FIG. 2.

(A) Restriction map of the genomic facC region. The ends of the cDNA clone are indicated by the arrowheads. (B) facC deletion plasmid. A 4-kb PstI-BamHI fragment containing the argB⁺ gene (68) was inserted into the indicated PstI-BglII sites, resulting in the replacement of 1,165 bp of facC sequences with the argB⁺ gene.

The facC sequence is shown in Fig. 3. Comparison of the genomic and cDNA sequences revealed a long 5′ untranslated region (681 bp) in the predicted mRNA with an intron present prior to the predicted start codon.

The FacC product has similarity to CATs.

Sequence alignments showed that FacC has a high level of similarity throughout the sequence to CATs from fungi, Homo sapiens (Fig. 4), and other species (results not shown). A lower level of similarity to medium- and long-chain CATs of mammals was also observed (results not shown). The comparisons indicated that fungal CAT proteins fall into two classes based on sequence similarity. Cat2P (S. cerevisiae) and Cat1P (Candida tropicalis) fall into one class (62.6% similarity, 44.7% identity), while FacC and Yat1P (S. cerevisiae) fall into the other (65.3% similarity, 49% identity). There is approximately 50% similarity and 30% identity between these two classes. Therefore, there may be two functional classes of CAT enzymes in fungi.

FIG. 4.

Comparison of the amino acid sequence of A. nidulans (An) FacC with those of CAT-encoding genes. The proteins in the alignment are Yat1p (61) and Cat2p (44) from S. cerevisiae (Sc), Cat1p (39) from C. tropicalis (Ct), and Cat1 (15) from H. sapiens (Hs). The comparison used the Box shade program (http://ulrec3.unil.ch/software/Box_form.html). The black bars indicate similar sequences in FacC and Yat1p but not in the other proteins.

Since acuJ mutants are deficient in CAT activity (53), the ability of pCS3875 containing the facC gene to complement acuJ211 was tested by cotransformation with riboB⁺-containing plasmid pPL3. No complementation was observed in transformants shown to contain pCS3875 sequences by Southern blot analysis. This indicated that extra copies of the facC gene could not compensate for the lack of acuJ-dependent CAT activity.

Construction of a facC deletion mutant.

The central region of the facC gene was replaced with the A. nidulans argB⁺ gene (Fig. 2B). A plasmid containing this sequence was transformed into an argB1 strain selecting for ArgB⁺. Twenty-six of 176 transformants were unable to grow on acetate. Southern blot analysis of eight of these confirmed that replacement of the native facC⁺ gene with the deletion construct by a homologous double crossover had occurred in seven of these. One of these was shown in crosses to cosegregate ArgB⁺ and acetate-negative phenotypes and had phenotypes similar to those of facC102 mutant strains with respect to growth on acetate and fatty acids and resistance to fluoroacetate (Table 1). Northern blot analysis showed that no facC-specific RNA was present under acetate-induced conditions in a facC deletion mutant (data not shown).

Regulation of CAT activity.

CAT activity in A. nidulans has been shown to be induced by both acetate and fatty acids (53). We have confirmed this by using Tween 80 as a fatty acid source (Table 3), as well as oleate (results not shown). Both acuJ211 and facC mutations affected CAT levels (Table 3). The acuJ211 ΔfacC::argB double mutant had no significant CAT activity, indicating that these genes affect all CAT activity. The acuJ211 mutation produced very low levels of activity with Tween 80 induction but retention of acetate-inducible activity, while the facC102 and ΔfacC::argB mutations produced reduced levels with acetate induction but retention of Tween 80-inducible activity. Acetate-induced activity was restored to the facC102 mutant by transformation with pCS3875. These data strongly suggested that acuJ-dependent CAT activity was fatty acid induced while facC-dependent activity was acetate induced.

TABLE 3.

Effects of various mutations on regulation of CAT activity

Relevant genotype	CAT sp acta
	Glucose	Carbon starvation	Acetate	Tween 80
Wild type	6.5 ± 0.9	20.4 ± 1.3	61.0 ± 8.1	62.5 ± 7.5
facC102	7.6 ± 0.7	15.0 ± 0.5	23.2 ± 3.3	68.6 ± 3.8
ΔfacC::argB	6.4 ± 0.4	25.3 ± 3.3	26.8 ± 2.7	94.2 ± 1.1
acuJ211	NS	4.6 ± 1.3	22.8 ± 2.7	2.3 ± 0.2
acuJ211 ΔfacC::argB	NS	NS	NS	NS
facB::BleR	6.9 ± 0.5	25.6 ± 5.3	28.5 ± 3.9	89.2 ± 8.4
acuJ211 facB::BleR	NS	NS	NS	NS
facA303	7.2 ± 1.3	25.4 ± 3.8	70.1 ± 1.4	70.7 ± 4.6
facC102(pCS3875)b	7.6 ± 1.8	24.5 ± 3.9	48.3 ± 4.3	55.6 ± 6.9

^aAverage CAT specific activity is expressed as nanomoles of CoA-SH produced per minute per milligram of protein ± the standard error for three independent determinations. NS, no significant activity. Mycelium was grown in minimal medium containing 1% glucose for 16 h and then harvested (glucose) or transferred to minimal medium containing no carbon source (carbon starvation), 50 mM sodium acetate (acetate), or 0.5% Tween 80 (Tween 80) for 4 h before harvesting. All media contained 10 mM ammonium chloride as the sole nitrogen source. 

^bfacC102 strain cotransformed with facC⁺-containing plasmid pCS3875. 

Acetate-induced activity was observed in a facA303 mutant (lacking ACS activity), indicating that acetate does not need to be metabolized to acetyl-CoA for induction. The facB::Ble^R loss-of-function mutation resulted in loss of acetate induction, indicating dependence on this regulatory gene for induction of facC-dependent activity. Consistent with this, the acuJ211 facB::Ble^R double mutant completely lacked CAT activity.

Significant CAT activity occurred in the absence of an added inducer, and in the presence of glucose, this was acuJ dependent. Carbon starvation in the absence of an inducer resulted in increased CAT activity. The major component of this was acuJ dependent, but a minor component was observed in an acuJ211 background (Table 3). It is therefore suggested that expression of both activities is subject to carbon catabolite repression. The creA gene has been shown to encode a DNA binding protein involved in this process (4, 22, 35). Consistent with this, the creA204 mutation, which results from an amino acid substitution in the DNA-binding domain of CreA (63), affected CAT activity in the presence of glucose (Table 4).

TABLE 4.

Effect of the creA204 mutation on glucose repression of CAT activity

Relevant genotype	CAT sp acta
	Glucose (1%)	Glucose (0.1%)–acetate	Glucose (1%)–acetate	Glucose (1%)–Tween 80
Wild type	6.5 ± 0.9	10.3 ± 1.0	11.7 ± 2.0	8.2 ± 0.8
creA204	11.5 ± 3.9	31.6 ± 2.7	34.7 ± 1.1	18.7 ± 1.8

^aCAT specific activity is expressed as in Table 3. Mycelium was grown in the indicated concentration of glucose for 16 h, and then 50 mM sodium acetate or 0.5% Tween 80 was added and mycelium was harvested after a further 4 h of incubation. 

In the presence of glucose and in the absence of an added inducer, CAT activity was elevated about twofold in the creA204 strain. Acetate-induced activity in the presence of glucose was elevated about threefold in the creA204 strain, while Tween 80 induction was elevated about twofold. These data suggest that both facC- and acuJ-dependent activities are subject to CreA-mediated repression. The data also indicate that acetate induction of facC-dependent activity is subject to CreA-mediated repression.

Regulation of facC expression.

Northern blot analysis was used to analyze facC RNA (Fig. 5). Expression was induced by acetate, and this induction was reduced by the presence of glucose. The creA204 mutant was inducible by acetate and, in the presence of glucose, showed slightly higher levels of RNA than the wild type, but the sensitivity of the blot was not sufficient to clearly show the two- to threefold effects observed (Table 4). The expression of facC was clearly FacB dependent, since the facB101 loss-of-function mutation (66) produced low levels of expression under both noninduced and induced conditions. The facC102 mutant had low, weakly inducible levels of expression, while the acuJ211 mutant had significantly increased levels of induction by acetate.

FIG. 5.

Analysis of facC expression. RNAs were extracted from the wild type (MH0001) and the creA204 (MH0664), acuJ211 (MH8091), facB101 (MH0764), and facC102 (MH7065) mutant strains. Mycelium was grown for 16 h in 1% glucose–10 mM ammonium tartrate medium before transfer to medium containing glucose (1%) or no added carbon source (C-free) with or without acetate (added as 50 mM sodium acetate). facC RNA was detected by using the 4-kb KpnI fragment of pCS3875 (Fig. 1) as a probe. H3 represents RNA detected by probing with an EcoRI fragment of the histone H3 clone (23). rRNA refers to the large rRNA species observed by ethidium bromide staining of the gel.

EMSAs of the 5′ region of facC.

A fusion protein (FacB-MBP) containing amino acids 4 to 417 of FacB [which includes the ZnII(2)Cys6 DNA binding domain] fused to the maltose-binding protein (MBP) expressed in Escherichia coli has been found to bind in vitro to sequences in the 5′ regions of FacB-regulated genes (66). EMSA analysis showed that partially purified extracts containing FacB-MBP bound to two fragments (−552 to −19 and −220 to −19) from the 5′ region of facC (Fig. 6). Inspection of these sequences revealed two potential binding sites that conform to the two dissimilar consensus sequences for FacB-MBP binding (67).

FIG. 6.

EMSA of the facC 5′ region using the FacB-MBP fusion protein. (A) Two different fragments were labelled and used for EMSA—the EcoRI-HindIII and AccI-HindIII fragments, as indicated in panel B. For each experiment, lane 1 contained no added extract, lane 2 contained a control of expressed MBP, and lanes 3 to 5 contained increasing amounts of FacB-MBP fusion protein extract. The arrowheads labelled F and B indicate the positions of free and bound probe, respectively. (B) facC 5′ regions used as probes. The black rectangles indicate the positions of sequences (shown below) which conform to previously determined FacB binding site consensus sequences—TCC/GN8-10C/GGA and GCC/AN8-10G/TGC (67). Coordinates are relative to the 5′ end of the cloned cDNA.

A fusion between the DNA binding domain of CreA and glutathione S-transferase (CreA-GST) has been widely used for studies of CreA binding sites (17, 28, 47). EMSA analysis of the 5′ facC sequence detected specific binding to the −220 to −19 fragment (Fig. 7). This fragment contains two sequences (on the noncoding strand) consistent with the proposed consensus (5′-SYGGRG-3′) for CreA binding (17).

FIG. 7.

EMSA of the facC 5′ region using the CreA-GST fusion protein. (A) The AccI-HindIII fragment shown in panel B was used as a probe. Lane 1 contained no added extract, lane 2 contained a control extract of expressed GST, and lanes 3 and 4 contained increasing amounts of the CreA-GST extract. The arrowheads labelled F and B indicate free and bound probe, respectively. (B) facC 5′ region indicating the position of the probe used and the locations of potential consensus CreA binding sites (black rectangles) with the sequences indicated. Coordinates are relative to the 5′ end of the cloned cDNA.

These data are therefore consistent with the idea that facC is directly regulated by FacB and CreA.

Acetate induction of amdS expression in acuJ211 backgrounds.

The observation that the acuJ211 mutation resulted in increased acetate induction of facC expression (Fig. 5) led us to investigate the effects of this mutation on amdS expression. This was studied by using an amdS-lacZ fusion present in single copy at the amdS locus (18) and determining acetate induction of β-galactosidase levels (Table 5). The acuJ211 mutation resulted in slightly elevated uninduced activity and an increased response to acetate induction. This effect was abolished by the facB::Ble^R and facC102 loss-of-function mutations. Under uninduced conditions, it is likely that there is some endogenous induction, and the results are consistent with the idea that the acuJ211 mutation results in the accumulation of an acetate-derived inducer which acts via FacB.

TABLE 5.

Effects of the acuJ211 mutation on acetate induction of the amdS-lacZ reporter gene

Relevant genotypea	β-Galactosidase sp actb
	Uninduced	Induced
Wild type	30.9 ± 2.2	143.1 ± 10.0
acuJ211	62.6 ± 11.6	445.3 ± 41.6
acuJ211 facB::BleR	11.1 ± 2.3	14.4 ± 1.0
acuJ211 facC102	29.4 ± 4.8	46.8 ± 4.8

^aAll strains contained the amdS-lacZ reporter gene integrated in single copy by replacement of the amdS gene (18). The reporter gene contained the cis-acting amdI9 mutation in the 5′ region, which results in increased sensitivity to acetate induction (36). 

^bAverage β-galactosidase specific activity is expressed as Miller units per minute per milligram of protein ± the standard error for a minimum of three independent determinations (18). Mycelium was grown for 19 h in minimal medium containing 0.1% glucose as the carbon source with 10 mM ammonium tartrate as the nitrogen source. Induction was done by adding 50 mM sodium acetate at 16 h. 

DISCUSSION

We have established that the facC gene encodes a CAT enzyme required for acetate utilization but not for fatty acid utilization. Although no substrates other than acetyl-CoA have been tested, the similarity of FacC to other CAT protein sequences, together with the phenotype of facC mutants, clearly indicates that the major substrate is acetyl-CoA. This gene is directly regulated by acetate induction mediated by the facB gene, and binding of a FacB fusion protein to 5′ sequences of facC has been demonstrated. Binding of a CreA fusion protein to 5′ facC sequences has also been demonstrated, and the creA204 mutation results in elevated expression in the presence of glucose and acetate. The data are consistent with the proposal that CreA has a weak direct effect on facC expression and a strong indirect effect via effects on facB expression, which is subject to CreA-mediated glucose repression (38). Therefore, facC is regulated similarly to the facA (ACS-encoding) gene.

In addition to the facC-encoded activity, there is an additional CAT activity dependent on the acuJ gene, as suggested by Midgley (53). Our data suggest that this activity is present constitutively and is inducible by fatty acids and subject to catabolite repression. This pattern of regulation is consistent with the observation that this activity is required for growth on both acetate and fatty acids as sole carbon sources.

Since intracellular membranes are impermeable to acyl-bound CoA molecules, such as acetyl-CoA, acetyltransferase enzymes are required for shuttling of acyl groups across these membranes by formation of acylcarnitine derivatives which pass the membranes and then reform acyl-CoA derivatives in the organelle (7, 45, 64, 70). In vertebrates, multiple enzymes with overlapping activities have been isolated. CAT, specific for short-chain (C2 to C4) acyl derivatives, is localized in both mitochondria and peroxisomes (29, 54); carnitine palmitoyltransferase, specific for medium- and long-chain (C10 to C18) molecules, is mitochondrial (10); while carnitine octanoyltransferase (C6 to C8) is peroxisomal (29). This is consistent with the idea that β-oxidation of fatty acids occurs in both peroxisomes and mitochondria. In peroxisomes, oxidation terminates at the medium chain length (C6 to C8), and these shuttle to the mitochondria, where they are fully oxidized. In addition, acetyl-CoA formed in peroxisomes must enter mitochondria for entry into the citric acid cycle (reviewed in reference 7). In humans, a single gene encoding CAT proteins targeted to both mitochondria and peroxisomes has been found (15).

In fungi, β-oxidation can progress to completion in peroxisomes and does not occur in mitochondria (69; reviewed in reference 57). Consistent with this, only CAT activity has been detected in fungi and not carnitine transferases specific for longer-chain acyl derivatives. Cloning of CAT-encoding genes—YCAT1 from S. cerevisiae (25, 44) and CT-CAT from C. tropicalis (39)—has demonstrated that there is a single gene encoding the mitochondrial and peroxisomal CAT activities. Two different transcriptional start points result in two different proteins, one with an N-terminal mitochondrial localization sequence and one without. A C-terminal peroxisomal localization sequence results in peroxisomal localization of proteins lacking the mitochondrial localization sequence. A second gene, YAT1, encoding CAT activity was identified in S. cerevisiae, and this activity was found on the outer surface of intact mitochondria and lacked localization sequences (61).

It would be expected that CAT activity is required for shuttling of acetyl-CoA across mitochondrial and peroxisomal membranes and is therefore necessary for growth on acetate and fatty acids. However, in S. cerevisiae, deletion of either YCAT1 or YAT1 did not significantly alter growth rates on acetate or fatty acids (44, 61, 70). Either the presence of the second CAT gene could compensate for the loss of the first, or there is an additional shuttle system that could compensate for the carnitine-dependent shuttle. The properties of a double mutant have not been reported, but the finding that deletion of the CIT2 gene encoding peroxisomal citrate synthase prevented a YCAT null mutant from growing on acetate or fatty acids (70) indicates the second possibility. A second shuttle involving citrate can compensate for the loss of peroxisomal CAT activity.

It is clear that the situation is different in A. nidulans. Loss of facC-encoded CAT activity prevents growth on acetate but not on fatty acids, and this is consistent with acetate induction of facC expression. Therefore, this activity is proposed to be present in the cytosol, and this is supported by the lack of signal sequences. In addition, FacC shows greater similarity to YAT1p, which is proposed to be located on the outer surface of mitochondria (61), than it does to YCATp, which is located in peroxisomes and mitochondria.

Mutations in the acuJ gene result in loss of growth on both acetate and fatty acids as carbon sources. The acuJ-dependent activity is present at significant levels in noninduced cells and is inducible by fatty acids. We propose that the activity is localized in peroxisomes and mitochondria and is absolutely required for the formation of acetyl-CoA in peroxisomes and mitochondria from acetylcarnitine during growth on acetate and for the shuttling of acetyl groups from peroxisomes to mitochondria during growth on fatty acids. This implies that the citrate shuttle present in S. cerevisiae is absent in A. nidulans, and this is supported by the lack of peroxisomal citrate synthase activity (6, 56). It is predicted that the acuJ-dependent CAT enzyme protein will contain mitochondrial and peroxisomal localization signals and be more similar to YCATp and CT-CATp than to FacC. A model for the proposed pathways is presented in Fig. 1.

We have shown that acetate induction of facC and of an amdS-lacZ reporter is increased in an acuJ211 background (Fig. 5 and Table 5). Previous results have indicated that facB-dependent acetate induction is observed in a facA mutant background lacking ACS (33, 34). This indicates that acetate does not have to be metabolized to acetyl-CoA for induction to occur. Loss of peroxisomal and mitochondrial CAT activity in an acuJ211 background is proposed to result in elevated acetate levels and, hence, increased induction. In an acuJ mutant, acetate would be converted to acetyl-CoA (by ACS), and this would be converted to acetylcarnitine by facC-encoded CAT activity. The acetylcarnitine, however, would accumulate in mitochondria and peroxisomes due to the absence of acuJ-dependent CAT activity. Since the CAT shuttle is reversible, this would lead to reversal of the pathway leading to facC-dependent accumulation of acetyl-CoA in the cytosol. Acetyl-CoA hydrolase, which has been detected in N. crassa (14), might then convert this to acetate, which would accumulate. A gene encoding a protein with extensive similarity to the N. crassa acetyl-CoA hydrolase has recently been isolated from A. nidulans (34). This hypothesis is consistent with increased acetate induction of the amdS-lacZ gene in an acuJ211 background that is dependent on facB (Table 5).

In addition, facB expression itself is inducible by acetate and subject to carbon catabolite repression (38). However, acetate induction is not observed in either a facA or a facC mutant background, indicating that acetate must be metabolized to induce facB expression (34). Furthermore, preliminary data indicate that the acuJ211 mutation leads to increased levels of facB expression (34). Therefore, it is proposed that increased levels of FacB occur in an acuJ mutant, and this results from accumulation of acetylcarnitine (or a derivative), which is the true inducer of facB expression. The increased amount of FacB results in increased expression of facC and amdS. Consistent with this, the facC102 mutation, which prevents acetylcarnitine formation from acetate, is epistatic to the acuJ211 mutation with respect to acetate induction (Table 5). This model predicts that there is an additional regulatory gene regulating facB expression in response to the level of acetylcarnitine.

Overall, this interpretation is consistent with the previous suggestion that facB-mediated induction has two components—induction of FacB levels and increased activity of FacB caused by acetate itself (65). It should also be noted that catabolite repression of structural genes (e.g., facC and amdS) subject to facB control may be directly regulated by CreA as well as indirectly via CreA regulation of facB expression.

The acuJ-dependent CAT activity is inducible by fatty acids (53, this study). Fatty acid induction of gene expression in A. nidulans is not well characterized—in contrast to the situation in S. cerevisiae, where two transcription factors, OAF1p and OAF2p, are proposed to bind to oleate response elements in the promoters of oleate-inducible genes (24, 37, 49). It is clear that there are at least three classes of genes involved in acetate and fatty acid utilization. Acetate-specific genes such as facA and facC are regulated by facB-mediated acetate induction. Genes involved in the glyoxalate bypass (acuD and acuE) are acetate inducible via facB but are also required for growth on fatty acids. Since facB mutations do not affect growth on fatty acids (Table 1), it is likely that these genes are subject to facB-independent control. It is of interest that genes additional to facB have been found to influence acuD-dependent isocitrate lyase activity (see reference 52). The third class of genes is those encoding enzymes of β-oxidation and the acuJ-dependent CAT activity, and they are fatty acid inducible. Characterization of the acuJ gene is clearly of particular interest, and we have recently cloned the gene and found that this gene encodes a CAT.