Genetic Analysis of the Role of Peroxisomes in the Utilization of Acetate and Fatty Acids in Aspergillus nidulans

Abstract

Peroxisomes are organelles containing a diverse array of enzymes. In fungi they are important for carbon source utilization, pathogenesis, development, and secondary metabolism. We have studied Aspergillus nidulans peroxin (pex) mutants isolated by virtue of their inability to grow on butyrate or by the inactivation of specific pex genes. While all pex mutants are able to form colonies, those unable to import PTS1 proteins are partially defective in asexual and sexual development. The pex mutants are able to grow on acetate but are affected in growth on fatty acids, indicating a requirement for the peroxisomal localization of β-oxidation enzymes. However, mislocalization of malate synthase does not prevent growth on either fatty acids or acetate, showing that the glyoxylate cycle does not require peroxisomal localization. Proliferation of peroxisomes is dependent on fatty acids, but not on acetate, and on PexK (Pex11), expression of which is activated by the FarA transcription factor. Proliferation was greatly reduced in a farAΔ strain. A mutation affecting a mitochodrial ketoacyl-CoA thiolase and disruption of a mitochondrial hydroxy-acyl-CoA dehydrogenase gene prevented growth on short-chain but not long-chain fatty acids. Together with previous results, this is consistent with growth on even-numbered short-chain fatty acids requiring a mitochondrial as well as a peroxisomal β-oxidation pathway. The mitochondrial pathway is not required for growth on valerate or for long-chain fatty acid utilization.

EUKARYOTES contain single membrane organelles called microbodies containing specialized enzymes involved in a wide range of metabolic activities. Commonly these enzymes include oxidases generating hydrogen peroxide. When this occurs, the microbodies contain catalase and other antioxidative activities to remove reactive oxygen species and they are termed peroxisomes. In some organisms and plant tissues, microbodies lack catalase but contain enzymes of the glyoxylate bypass and these microbodies are often called glyoxysomes (for reviews see Titorenko and Rachubinski 2001; Platta and Erdmann 2007). In the filamentous fungus Neurospora crassa, microbodies contain glyoxylate cycle enzymes but lack catalase as well as peroxisomal oxidases but also contain enzymes for fatty acid β-oxidation with acyl-CoA dehydrogenase substituting for acyl-CoA oxidase (Kionka and Kunau 1985; Thieringer and Kunau 1991; Gainey et al. 1992; Schliebs et al. 2006). It is unlikely that the two classes of microbodies differ fundamentally and this is supported by the finding in silico of conserved orthologs of the full range of proteins that constitute peroxisomes (peroxins) in N. crassa (Kiel et al. 2006). This is also likely to be the case for plants (e.g., Nishimura et al. 1986).

Peroxins are proteins required for peroxisome division, for biogenesis from the endoplasmic reticulum, and for the import of proteins into the peroxisomal matrix (reviewed by Platta and Erdmann 2007). Mutations in pex genes can result in the absence of peroxisomes, abnormal peroxisomal structures, mistargeting of matrix proteins, or an inability to respond to stimuli that cause increased numbers of peroxisomes. Two major classes of peroxisomal targeting signals occur in matrix proteins. PTS1 sequences comprise three C-terminal amino acids (aa) usually of the form S/A R/K L/M although the context of the C-terminal sequence can greatly affect targeting (Brocard and Hartig 2006) and some peroxisomal proteins have cryptic PTS1 sequences (e.g., Klein et al. 2002). Other matrix proteins have PTS2 sequences close to the N terminus with the consensus R/K L/V/I X5 H/Q L/A/F/I (Petriv et al. 2004). A large number of peroxins are involved in the import of all matrix proteins while others are specific to each PTS class; e.g., Pex5 and Pex7 are the specific receptors for PTS1 and PTS2 proteins, respectively (Lazarow 2006; Stanley and Wilmanns 2006).

Many microorganisms are able to use two carbon compounds and fatty acids as sole carbon sources. In fungi, there is increasing interest in the role of lipid catabolism in both animal and plant pathogenesis (e.g., Barelle et al. 2006; Ramírez and Lorenz 2007; Schöbel et al. 2007; Wang et al. 2007) as well as in the involvement of peroxisomes in secondary metabolism and development (e.g., Berteaux-Lecellier et al. 1995; Kiel et al. 2005; Maggio-Hall et al. 2005).

In Saccharomyces cerevisiae, all steps of β-oxidation are carried out in peroxisomes, leading to the breakdown of long-chain fatty acids such as oleate to acetyl-CoA, which is then utilized via the glyoxylate cycle (Hiltunen et al. 2003). PEX mutants are unable to use oleate but are still able to use acetate or ethanol as carbon sources, indicating that the glyoxylate cycle does not depend on peroxisomal localization (Erdmann et al. 1989). The enzymes isocitrate lyase (ICL) and malate synthase (MAS) are unique to the glyoxylate cycle and are required for growth on oleate. ICL is not peroxisomal in S. cerevisiae while MAS contains a PTS1 and, during growth on oleate, is in peroxisomes where the substrate acetyl-CoA is produced by β-oxidation. However, mislocalization of MAS to the cytoplasm does not prevent oleate utilization and it is cytoplasmic during growth on acetate or ethanol (Kunze et al. 2002, 2006). Genes specific to oleate utilization, including some PEX genes, are induced by fatty acids via the Oaf1-Pip2 heterodimer (Smith et al. 2002, 2007; Gurvitz and Rottensteiner 2006) while enzymes of the glyoxylate cycle are regulated separately by the gluconeogenic control mechanism involving the Cat8 and Sip4 activators (reviewed in Schuller 2003).

The filamentous ascomycete, Aspergillus nidulans, is able to grow on both short-chain (C4-6) and long-chain fatty acids as well as on two carbon compounds as sole sources of carbon and energy (Armitt et al. 1976; Hynes et al. 2006). Peroxisomes, which contain catalase, proliferate in response to oleate, and some enzymes of β-oxidation are inducible by oleate (Valenciano et al. 1996, 1998). Peroxisomal β-oxidation is indicated by the deletion of foxA encoding a peroxisomal multifunctional enzyme affecting growth on long-chain fatty acids (Maggio-Hall and Keller 2004). The glyoxylate cycle enzymes ICL (encoded by acuD) and MAS (encoded by acuE) have been found to be present in peroxisomes (Gainey et al. 1992; Szewczyk et al. 2001). While AcuE has a PTS1, AcuD lacks an obvious PTS (Sandeman et al. 1991; Gainey et al. 1992). Here we show that, unexpectedly, AcuD peroxisomal localization is dependent on the PTS2 receptor Pex7(G) and that deletion of an internal sequence results in mislocalization.

By the isolation and complementation of mutants unable to grow on butyrate we have discovered that the farA, farB, and scfA genes specify transcription factors required for fatty acid induction of genes of fatty acid utilization (Hynes et al. 2006). We report here the cloning by complementation of further genes required for growth on butyrate and identify some of these as pex genes encoding peroxins. We have also inactivated the genes predicted to encode the orthologs of Pex5, the PTS1 receptor: Pex3 required for peroxisome biogenesis and Pex11 involved in peroxisome proliferation. With GFP-tagged proteins containing PTS1 or PTS2 sequences we have investigated peroxisomal protein localization in these mutants. The pex mutants have been found to affect growth on both short- and long-chain fatty acids. However, growth on acetate is not abolished in the mutants and we have also found that mislocalization of AcuE (MAS) to the cytoplasm does not prevent growth on either fatty acids or acetate, showing that the glyoxylate cycle does not depend on peroxisomal localization. Proliferation of peroxisomes in response to fatty acids but not acetate has been observed and this is dependent on the Pex11 ortholog, PexK, and on the fatty acid regulator FarA. Overall, our results show that peroxisomes are dispensable for acetate metabolism but crucial for fatty acid utilization.

None of the mutants isolated had mutations in genes encoding peroxisomal enzymes of β-oxidation. However, one mutant affecting growth on short- but not long-chain fatty acids was predicted to result from a mutation affecting a mitochodrial ketoacyl-CoA thiolase. Maggio-Hall and Keller (2004) have demonstrated the existence of a mitochondrial pathway for the β-oxidation of short-chain fatty acids. We have disrupted a gene for a predicted mitochondrial hydroxy-acyl-CoA dehydrogenase and found that this prevents growth on short-chain fatty acids. The properties of these mutants are consistent with growth on even-numbered short-chain fatty acids requiring a mitochondrial as well as a peroxisomal β-oxidation pathway. This mitochondrial pathway is not required for growth on the odd-numbered valerate (C5) or for long-chain fatty acid utilization.

While the pex mutants form colonies on standard laboratory media, all mutants affected in PTS1 protein import have reduced conidiation and are subfertile in homozygous sexual crosses. These effects were shown to be independent of the VelA pathway, which affects the balance between asexual and sexual development (Kim et al. 2002; Tsitsigiannis et al. 2004a). The possibility that these effects were due to an inability to form Woronin bodies containing the PTS1 HexA protein was eliminated by the isolation of a hexA deletion mutant that had only a minor conidiation defect and was capable of normal sexual development. The pex mutants defective in PTS1 protein import were found to be auxotrophic for biotin due to an inability to synthesize the intermediate pimelic acid. This collection of pex mutants now provides valuable tools for the study of the requirements for the correct compartmentalization of enzymes and metabolite shuttling between the cytoplasm and organelles during the growth and development of a filamentous fungus.

MATERIALS AND METHODS

A. nidulans strains, media, and transformation:

Media and conditions for growth of A. nidulans were as described by Cove (1966). Carbon and nitrogen sources were added as appropriate to minimal salts. The pH of these was adjusted to 6.5 where necessary. Long-chain fatty acids were dispersed in media by the addition of 0.5% tergitol (NP-40, Sigma, St. Louis) before melting or autoclaving. All strains were derived from the original Glasgow strain and contained the velA1 mutation except where specifically mentioned and standard genetic manipulations were as previously described (Clutterbuck 1974, 1994; Todd et al. 2007). Strains containing the scdAΔ and echAΔ mutations were obtained from Lori Maggio-Hall. Preparation of protoplasts and transformation were as described (Andrianopoulos and Hynes 1988).

Molecular techniques:

Standard methods for DNA manipulations, nucleic acid blotting, and hybridization have been described (Sambrook et al. 1989; Hynes et al. 2006). For all strains generated by DNA manipulations followed by transformation of A. nidulans, the genome integration events were characterized by Southern blotting of restriction-enzyme-digested genomic DNA isolated from transformants.

Bioinformatic analysis:

All A. nidulans DNA sequences were derived from the genome sequence at the Broad Institute (http://www.broad.mit.edu/annotation/genome/aspergillus_group/MultiHome.html). Predicted protein-targeting sequences were analyzed by the use of the programs TargetP (http://www.cbs.dtu.dk/services/TargetP/) and Wolf Psort (http://wolfpsort.org/).

Isolation of mutants unable to grow on butyrate and molecular cloning of genes:

Mutants unable to utilize 20 mm butyrate were isolated as previously described (Hynes et al. 2006). For each mutant, it was shown by outcrossing that the phenotype resulted from a mutation in a single gene. PyrG⁻ double mutants were generated by crossing to a strain containing the pyrG89 mutation and these strains were used for cloning the genes by functional complementation using the genomic library in the autonomously replicating vector pRG3AMA1 (Osherov and May 2000) as described (Hynes et al. 2006). Sequences corresponding to the complementing genes were recovered and analyzed either through rescue by transforming into Escherichia coli or by direct PCR on genomic DNA from complementing transformants using primers flanking the insert DNA in the vector (Hynes et al. 2006). Sequences were compared with the A. nidulans DNA sequence (http://www.broad.mit.edu/annotation/genome/aspergillus_group/MultiHome.html) to identify the mutant genes. The cloned genes are listed in Table 1 together with the corresponding orthologs from S. cerevisiae (http://www.yeastgenome.org/). The A. nidulans system of gene nomenclature has been followed with letters corresponding to the numbers assigned to the S. cerevisiae genes.

TABLE 1.

Summary of mutants affecting peroxisomal functions

Gene	Allele	Genome locusa	S. cerevisiae orthologb	% identity (similarity)c	Length (aa)a	Proposed function
pexA	pexA9	AN5991.3	PEX1	39 (54)	1178	PTS1 and PTS2 protein import
pexF	pexF23	AN2925.3	PEX6	34 (50)	1477	PTS1 and PTS2 protein import
pexM	pexM15	AN1511.3	PEX13	35 (53)	437	PTS1 and PTS2 protein import
pexC	pexC∷bar	AN2281.3	PEX3	23 (39)	531	Peroxisome biogenesis
pexE	pexEΔ	AN10215.3	PEX5	33 (48)	656	PTS1 protein import receptor
pexG	pexG14	AN0880.3	PEX7	41 (56)	356	PTS2 protein import receptor
pexK	pexKΔ	AN1921.3	PEX11	27 (50)	236	Peroxisome proliferation
antA	antA15	AN0257.3	ANT1	37 (55)	337	ATP carrier

^aPredicted gene and length of protein based on gene annotations (http://www.broad.mit.edu/annotation/genome/aspergillus_group/MultiHome.html).

^bSequences from the Saccharomyces Genome Database (http://www.yeastgenome.org/).

^cBlastP (Altschul et al. 1990) of S. cerevisiae protein against predicted A. nidulans sequence.

The insert DNA in the complementing clones corresponded to the following coordinates in the genome sequence: pexA, contig 102—84,400–90,000; pexM, contig 23—52,763–58,765; antA, contig 5—338,689–341,912. For pexF, insert DNA from a complementing clone was subcloned and the complementing sequence identified as corresponding to AN2925.3. In addition, a strain containing a deletion of pexF was constructed by replacing a BamHI–EcoRI fragment (corresponding to encoded amino acid coordinates 343–1081) with the pyrG gene as a BamHI–EcoRI fragment (Borneman et al. 2001) and transformation of a pyrG89 strain selecting for PyrG⁺. The resulting strain had an identical phenotype to the original pexF23 mutant and was found to be allelic in crosses between the strains. The pexG14 mutation was complemented by a subclone, corresponding to contig 14 (47,157–49,109) derived from an initial complementing plasmid, and contained AN0880.3 as the only intact gene. DNA corresponding to the mutant gene was cloned from genomic DNA using the PCR primers 5′-ACCGTACGTATAAGTCATTCG-3′ and 5′-AAGAGAGGACCCAGTTGTAG-3′ and sequenced, revealing the presence of a T insertion corresponding to codon 261 and resulting in an in-frame stop codon after a further 18 codons, thereby eliminating two conserved WD domains. Another gene was found to be complemented by a plasmid containing a 6-kb insert, and sequencing from one end showed that this began at contig 68 (18,464) and contained the gene AN10512.3 predicted to specify a keto-acyl thiolase with a mitochondrial targeting sequence. This gene was designated mthA.

Gene deletions and disruptions:

AN10215.3 is predicted to encode an ortholog of Pex5 (Table 1). The gene was cloned using the PCR primers 5′-GTTATATGGGATTGGCGTGG-3′ and 5′-TCCGAATTCCGAGCTCTGC-3′, generating a 3.86-kb fragment, which was cloned into pBluescript SK+ (Stratagene, La Jolla, CA). A deletion strain was made by replacing a SphI–SalI fragment (corresponding to amino acids 191–633 of the predicted protein) with a SphI–XhoI fragment of the riboB gene (Oakley et al. 1987) and transforming the linear insert into a riboB2 strain selecting for RiboB⁺ transformants. The resulting phenotypes described in the results were complemented by cotransformation experiments using the original clone. AN1921.3 is predicted to encode an ortholog of Pex11 (Table 1). The gene was cloned using the PCR primers 5′-CATAACTAAGATGTGTCTCTGG-3′ and 5′-GTGCTGTGTCTGAAATATAGG-3′, generating a 2-kb fragment, which was cloned into pGEMTEasy (Promega, Madison, WI). A deletion strain was made by replacing a HindIII–XhoI fragment (corresponding to amino acids 27–181 of the predicted protein) with a HindIII–XhoI fragment of the riboB gene and transforming the linear insert into a riboB2 strain selecting for RiboB⁺ transformants. AN2281.3 is predicted to encode an ortholog of Pex3 (Table 1). A 3.2-kb fragment containing this gene was amplified with the primers 5′-GAAGGGAAAGAATGAGAGA-3′ and 5′-ACTCAATTGCATCTAGGTCC-3′ and inserted into pGEMTEasy. An internal 1-kb XbaI–EcoRI fragment lacking 5′ and 3′ sequences of the gene was cloned into SpeI–EcoRI-cut pMT1612 containing the bar selectable marker (Nayak et al. 2006). The resulting plasmid was transformed into an nkuAΔ strain (Nayak et al. 2006) selecting transformants for glufosinate resistance. Transformants were screened for growth on butyrate and a transformant showing inhibited growth was isolated. Southern blot analysis showed that the predicted disruption event resulting from homologous recombination had occurred, yielding two tandem truncated copies of the gene—one lacking 390 bp of 3′ coding sequence corresponding to the last 129 amino acids and the other lacking 445 bp of 5′-UTR and the coding sequences corresponding to the first 123 amino acids of the predicted protein. The resulting phenotypes described in the results were complemented by cotransformation experiments using the original cloned DNA. AN7008.3 was chosen for disruption because it was predicted to encode a mitochondrial protein containing 3-hydroxyacyl-CoA dehydrogenase domains. An internal fragment of this gene was obtained using the PCR primers 5′-TATCCGCCCATTGCTAACTC-3′ and 5′-AGATGCGATCTTGCCCATAC-3′ and cloned into a plasmid containing the A. fumigatus riboB gene as a heterologous marker (Nayak et al. 2006). This plasmid was transformed into an nkuAΔ strain selecting for Ribo⁺ transformants. A transformant showing inhibited growth on short-chain fatty acids was selected for further analysis. The predicted homologous disruption event generates truncated genes—one encoding a protein lacking the last 103 amino acids and one lacking the 5′ region and the sequence corresponding to the first 216 amino acids. AN4695.3 is predicted to encode the Woronin body protein HexA (see also accession no. AAF67173). A 2.9-kb fragment encompassing this gene was cloned using the PCR primers 5′-CCATCCATGCCTACGATACC-3′ and 5′-CAGTTCGGGAGAGATTCGAG-3′ and inserted into EcoRV-digested pBluescriptSK+. A deletion strain was made by replacing an NsiI-partial SalI fragment (corresponding to amino acids 22–221 of the predicted protein) with a partial XhoI-partial PstI fragment of the bar gene from pMT1612 and transforming the linear insert into a nkuAΔ strain selecting for glufosinate-resistant transformants.

Construction of strains expressing GFP-tagged proteins:

A strain expressing a GFP-AcuE fusion protein from the gpd promoter has been described (Szewczyk et al. 2001). This strain was crossed with relevant strains to obtain strains expressing the fusion protein in various mutant backgrounds. A strain expressing a GFP-AcuD fusion protein from the alcA promoter (Maggio-Hall and Keller 2004; K. Weeradechapon and G. Turner, personal communication) was also used for crosses with various mutant strains. Sources of gfp coding sequences for the following constructs were the plasmids pALX125, -213, or -215 with gfp flanked by the gpd promoter and the trpC terminator. AN1050.3 is predicted to encode a keto-acyl-CoA thiolase with an N-terminal PTS2 (RLNSILSHL) sequence. A 1.5-kb PCR fragment containing 100 bp of the 5′ sequence of this gene together with the entire coding region was generated with the primers 5′-CAACTCCTCATGTCTTCTCG-3′ and 5′-ACAACAAAGCGCAATGCATAAACC-3′. The second primer removed the stop codon inserting an NsiI site, allowing a C-terminal in-frame fusion with GFP by cloning into EcoRV-cut pALX215. A ClaI fragment containing this fusion sequence was cloned into ClaI-digested pMT1612 containing the bar selectable marker. This plasmid was transformed into an nkuAΔ strain selecting for glufosinate resistance. A transformant was identified as arising by a single homologous integration event at AN1050.3, resulting in tandem copies of a truncated gene and the fusion gene driven by the AN1050.3 promoter and encoding 417 amino acids fused to GFP. This strain was used for crosses into pex mutant backgrounds. An N-terminal GFP fusion with AcuE (AN6653.3; Sandeman et al. 1991) was made by generating a PCR fragment with a primer containing an EcoRI site (5′-GGAATTCCGACCGCCCAGCTTAAGGATGTG-3′) and the standard M13 reverse primer using as template a PGEMTEasy clone of acuE. This was cloned as an EcoRI–SpeI fragment into pALX213, resulting in a fusion gene driven by the gpd promoter producing GFP fused to amino acids 4–540 including the C-terminal SKL and containing 671 bp 3′-UTR. A mutation of the C-terminal S codon (TCA) was mutated to a stop codon (TAA) using the primers 5′-GTACCCGGAGACGAAATCTC-3′ and 5′-TGCTTAAAAGCTCTAATTTC-3′ for inverse PCR and recircularizing using as template an acuE clone in pGEMTEasy containing the entire gene together with 1533 bp of 5′-UTR. An NcoI–SpeI fragment containing the mutation was used to replace the corresponding fragment in the gfp-acuE clone described above, resulting in a predicted GFP-AcuE fusion protein lacking the C-terminal SKL sequence. For complementation and GFP localization studies, the resulting constructs were cotransformed into a riboB2 stain containing an acuEΔ (Nayak et al. 2006) selecting for RiboB⁺. The entire ORF of AcuD (AN5634.3) was amplified with the primers 5′-GCTGAAAGCTTCCATCATGTCTT-3′ and 5′-CTGCAGTTGAACTGATCCTCTGTC-3′ with the second primer mutating the stop codon. This fragment was cloned into pALX215 at the EcoRV site, creating a coding region for the entire ORF of AcuD with GFP fused at the C terminus driven by the gpd promoter with the trpC terminator. This plasmid was used as a template for inverse PCR to delete sequences corresponding to AcuD amino acids 148–158 and 315–338 using the primer pairs 5′-GGGGTTGACTACCTTCGTCC-3′/5′-GGGAGTCATGCGCTCCTCGC-3′ and 5′-GGGTCCAACCTCGAGGCCCG-3′/5′-GGGATCAACGACAGCGTCGT-3′, respectively. NotI–XhoI fragments from each of the three plasmids were inserted into a plasmid containing the A. fumigatus pyroA4 selectable marker (Nayak et al. 2006) as well as an internal fragment of the wA gene. The resulting plasmids were transformed into a pyroA4 nkuAΔ strain selecting for Pyro⁺ and screening of transformants for white conidial colonies arising from homologous integration of the plasmid at the wA locus. The pexK gene (AN1921.3) was amplified with the primers 5′-CATAACTAAGATGTGTCTCTGG-3′ and 5′-GTAAGCTGTCTTTCGCCA-3′ to generate a product with a 728-bp 5′-UTR and with the stop codon changed to TAC. An in-frame fusion was made by cloning this fragment into a plasmid containing gfp and the trpC terminator. The insert was amplified using the first primer and the standard vector primer M13 forward and cloned into a plasmid containing the A. fumigatus pyroA4 gene cut with EcoICRI. This was transformed into an nkuAΔ pyroA4 strain selecting for PyroA⁺ and a transformant identified in which integration had occurred at the trpC locus. This strain expressed PexK-GFP from the pexK promoter and was crossed into different backgrounds.

Microscopy:

Mycelium for microscopy was grown on coverslips in liquid media at 25°. Alternatively, conidia were grown on solid media (1% agar)-coated microscope slides. Microscopy and capture of images were as described (Szewczyk et al. 2001).

RESULTS

Isolation of mutants affected in peroxisomal functions:

Mutants specifically showing reduced ability to grow on media containing butyrate as the sole carbon source but able to utilize acetate and glutamate were isolated as described previously (Hynes et al. 2006). The phenotype of mutants on butyrate medium ranged from complete inhibition to poorer utilization. The genes corresponding to some of these mutants were cloned by complementation using a genomic library in the pRG3AMA1 autonomously replicating vector as described in materials and methods. Sequencing of complementing clones and comparison with the A. nidulans genome sequence showed that these genes corresponded to those encoding characterized peroxisomal functions, including the peroxins Pex1, -6, -7, and -13, designated PexA, -F, -G, and -M to follow the A. nidulans gene nomenclature (Table 1). In addition, a gene encoding an ATP carrier (AntA) was identified. Notably, with one exception (see below), no mutants affected in fatty acid metabolic enzymes were identified.

Mutations affecting additional peroxisome functions were generated by cloning genes by PCR using primers based on the A. nidulans genome sequence and generating deletions or disruptions (materials and methods). A deletion of the pexF gene resulted in a phenotype identical to that of the pexF23 mutation and was shown to be allelic to pexF23. The genes predicted to encode Pex5 (designated PexE) and Pex11 (PexK) were deleted and the gene predicted to encode Pex3 (PexC) was disrupted by insertion of the bar gene encoding glufosinate resistance (Table 1). The predicted defects in peroxisomal functions based on characterized S. cerevisiae genes are summarized in Table 1.

Effects of mutations on peroxisomal protein targeting:

The effects of pex mutations on the localization of the glyoxylate bypass enzymes, ICL (AcuD) and MAS (AcuE), were investigated by the use of GFP-tagged proteins (Figure 1A). As shown previously (Szewczyk et al. 2001; Maggio-Hall and Keller 2004), GFP-AcuD and GFP-AcuE localized to punctate dots. This pattern was lost in pexF, pexM, and pexC mutant backgrounds, consistent with the predicted roles of these genes in PTS1 and PTS2 protein import and peroxisomal biogenesis. AcuE contains a predicted PTS1 (SKL), and mislocalization was observed in the pexE mutant background (PTS1 receptor) but not in the pexG mutant (PTS2 receptor). Surprisingly, the reverse was observed for AcuD, which has no characterized peroxisomal targeting sequence, indicating that this protein might have a cryptic PTS2 sequence recognized by PexG (see below). The function of PexG in the targeting of PTS2 proteins was confirmed by the finding that a GFP fusion with the product of the gene AN1050.3, predicted to encode a ketoacyl-CoA thiolase with a standard N-terminal PTS2 sequence, was mislocalized in the pexG but not the pexE mutant (Figure 1B).

Figure 1.—

Effects of pex mutations on the peroxisomal localization of GFP-tagged proteins. Strains expressing the indicated GFP fusion proteins in different pex mutant backgrounds were grown for 15–20 hr on microscope slides coated with solid media. (A) AcuD-GFP stains were grown with 0.25% ethanol except for the pexEΔ strain, which was grown on 10 mm threonine GFP-AcuE with 1% glucose and (B) AN1050-GFP with 0.25% Tween80 as carbon sources with ammonium as the nitrogen source. (C) Deletion of amino acids 315–338 of AcuD results in loss of peroxisomal localization. Mycelium of strains containing plasmids expressing the gfp fusion genes integrated at the wA locus were grown in liquid glucose minimal media on coverslips for microscopy. DIC, differential interference contrast; FITC, standard fluorescein isothiocyanate filter set. Bar, 20 μm.

The ICL of fungi show an unusually diverse range of peroxisomal targeting signals (reviewed in Kunze et al. 2006). AcuD has no obvious peroxisomal targeting signals as also found for ICL in other filamentous ascomycetes. ICL sequences from bacteria to plants are highly conserved and Gainey et al. (1992) identified a region relatively conserved in filamentous ascomycetes but absent from bacteria, plants, and S. cerevisiae. By a similar analysis, we identified two regions in AcuD for analysis. In-frame deletions corresponding to these sequences were made in a gfp-acuD fusion construct and transformed into A. nidulans with integration at the wA locus (materials and methods). The deletion corresponding to aa 315–338 resulted in mislocalization (Figure 1C), indicating that the 315–338 region may contain a cryptic sequence acting as a PTS2 sequence recognized by the pexG receptor. The gfp-acuD strains were crossed to an acuD loss-of-function mutant, and it was found that the wild-type fusion gene complemented for growth on acetate while both deletion genes showed lack of complementation, suggesting that the deletions resulted in loss of enzyme activity. We cannot exclude the possibility that AcuD is localized via piggybacking with an unknown protein with a standard PTS2 sequence.

Proliferation of peroxisomes is dependent on PexK:

In S. cerevisiae and other fungi, including A. nidulans, peroxisome numbers increase in response to fatty acids (Valenciano et al. 1996). Pex11 has been shown to be a peroxisomal membrane protein that recruits dynamin-like proteins necessary for peroxisome division and proliferation in many species (Erdmann and Blobel 1995; Li et al. 2002; Li and Gould 2003; Kiel et al. 2005; Thoms and Erdmann 2005; Yan et al. 2005; Lingard and Trelease 2006; Orth et al. 2007). When a strain expressing GFP-AcuE from the constitutive gpd promoter was grown on glucose and transferred to media containing different carbon sources, increased numbers of peroxisomes were observed in the presence of oleate or butyrate in comparison with glucose or acetate media (Figure 2A). This was not observed in the pexKΔ background where there was no response to fatty acid media with only a few larger intensely fluorescent peroxisomes observed.

Figure 2.—

Peroxisomal biogenesis and proliferation. (A) Proliferation of GFP-AcuE-labeled peroxisomes in reponse to fatty acids is dependent on PexK and FarA. (B) FarA-dependent proliferation of peroxisomes in response to fatty acids visualized by peroxisomal membrane labeling with PexK-GFP. Enlargements (arrowheads) show increased numbers of peroxisomes resulting from fatty acid induction. In glucose media only 1–2 peroxisomes are observed while butyrate and oleate result in 5–10 and >10 peroxisomes in each cluster, respectively. In farAΔ, single peroxisomes are observed except for oleate where clusters of up to 5 peroxisomes are infrequently observed. (C) Peroxisome biogenesis visualized by PexK-GFP is dependent on PexC and is aberrant in the pexF23 mutant. Mycelium was grown on coverslips in 1% glucose minimal liquid media for 18 hr and then transferred to the same media or to minimal media containing 10 mm acetate, 10 mm butyrate, or 0.5% Tween80 (as a source of oleate). All media contained ammonium chloride as the nitrogen source.

A sequence encoding PexK labeled at the C terminus with GFP was inserted into the A. nidulans genome at the trpC locus (materials and methods) and it was found that this fusion protein was able to almost fully complement the pexKΔ for growth on fatty acid media. A pexKgfp pexK⁺ strain was grown for 16 hr in glucose medium and transferred to glucose-, butyrate-, or oleate-containing media for 4 hr for microscopy. During growth on glucose, labeling of the peroxisomal membrane by GFP would be dependent on basal expression of the fusion gene. Fatty acid induction would be expected to increase expression of both tagged and untagged PexK, potentially resulting in peroxisomal proliferation. This was observed with both butyrate and oleate resulting in increased numbers of peroxisomes present in distinct clusters spaced along the hyphae (Figure 2B). Consistent with PexC being required for de novo production of peroxisomes from the endoplasmic reticulum (Hoepfner et al. 2005; Kragt et al. 2005), the pexC∷bar disruption resulted in mislocalization of GFP-AcuE (Figure 1A) and also in loss of the formation of peroxisomes labeled with PexK-GFP (Figure 2C). The pexF23 mutation resulted in a few small punctate fluorescent structures, presumably peroxisomal membrane ghosts, as well as increased levels of mislocalized PexK-GFP (Figure 2C).

Deletion of farA, encoding a transcription factor required for fatty acid induction of enzymes involved in fatty acid utilization and of expression of some pex genes (Hynes et al. 2006), greatly reduced the expansion of the number of peroxisomes in response to butyrate and oleate as visualized by both GFP-AcuE and PexK-GFP labeling (Figure 2, A and B). This indicated that expansion of the numbers of peroxisomes in response to fatty acids results from FarA-dependent increased levels of the proteins required for peroxisomal biogenesis and proliferation.

Growth of mutants on carbon sources:

All mutants were able to utilize glucose, a strong carbon source, as well as the poorer carbon sources, proline and lactose, almost as well as wild type (Figure 3A). The only observable effect was a delay in growth on all media observable at 1 day of incubation, resulting in somewhat smaller colony diameters. Therefore, peroxisomal functions are not essential for growth. The mutants were also able to utilize acetate with the exception of pexEΔ (discussed in more detail below). However, growth of all pex mutants was clearly affected to some extent. Growth on ethanol, which is metabolized via acetyl-CoA, was similar to that on acetate except for the somewhat greater effects of the pex mutants. Growth on the short-chain fatty acids butyrate (C4), valerate (C5), and hexanoate (C6) was greatly inhibited in all mutants defective in the peroxisomal targeting of PTS1 proteins and this was observed in the presence of lactose (Figure 3A). Utilization of these fatty acids was clearly reduced but not inhibited in the pexG mutant defective in PTS2 protein targeting as well as in the pexKΔ and the antA15 mutants.

Figure 3.—

Growth of mutants on carbon sources. The following carbon sources were added to minimal media (1% agar) with 10 mm ammonium chloride as the nitrogen source at the following concentrations: glucose (1%), proline and acetate (50 mm), ethanol and lactose (0.5%), butyrate and valerate (10 mm), hexanoate (5 mm), and C12-C22 chain-length fatty acids (2.5 mm). Acids were adjusted to pH 6.5 with sodium hydroxide where necessary and tergitol (NP-40; Sigma) was added before melting media to disperse longer-chain fatty acids. Growth was for 2–3 days at 37°. (A) Growth of mutants on short-chain fatty acids. (B) An enlargement of colonies of a wild-type strain on oleate and lactose to show inhibition of conidiation. (C) Growth of mutants on long-chain fatty acids.

As observed previously (Maggio-Hall and Keller 2004), A. nidulans is unable to grow on fatty acids of chain length 7–10 due to strong inhibition but is able to utilize longer-chain fatty acids. However, on long-chain fatty acids conidiation was greatly reduced as illustrated for oleate compared with the poor carbon source lactose (Figure 3B). Growth inhibition by long-chain fatty acids was observed for all mutants affected in PTS1 protein localization while growth was reduced but not inhibited in the pexF, pexK, and antA mutants (Figure 3C). These results showed that peroxisomes are involved in both short- and long-chain fatty acid utilization.

Developmental defects in peroxisomal mutants:

Despite strong growth on standard glucose containing complete and minimal media, pex mutants defective in PTS1 protein localization conidiated poorly as shown by a brown colony appearance. As for wild type, this was partly remediated by the presence of 1 m sorbitol (Figure 4A) and 0.6 m KCl (not shown). The effects of the pexG mutation were much less extreme. Conidial counts supported these observations (Figure 4B). Mutants containing the pexKΔ were not detectably affected in conidiation (not shown).

Figure 4.—

Developmental phenotypes of pex mutants. (A) Colony morphology of mutants. Reduced conidiation is shown by the brown appearance of colonies growing on complete and glucose minimal media. This is partially restored by the presence of sorbitol (1 m) in complete medium. (B) Conidial counts of mutants grown on complete media or complete media with 1 m sorbitol. Conidia were harvested from 1.7-cm-diameter circular regions and counted with a hemocytometer. Error bars represent the standard errors of three replicates. (C) Colony morphology of pex mutants showing a further reduction of conidiation in velA⁺ backgrounds. (D) Magnified view of surface of colonies growing on glucose minimal medium showing the production of small cleistothecia in pexF and pexC mutants in a velA⁺ background. The cleistothecia are observed as black spheres coated by Hulle cells. (E) Colony morphology of mutants containing a deletion of the hexA gene. Conidiation is only slightly affected in comparison with the pexF23 mutant. Both wild-type (green) and yA1 (yellow) conidial colonies are shown. Increased sensitivity to sorbose (0.1% in the presence of 50 mm proline as carbon source) is shown by the hexAΔ and pexF23 mutants. (F) Cleistothecia (white arrowheads) produced by wild-type and hexAΔ but not pexF23 strains in homozygous crosses viewed with an inverted microscope. (G) Induction of the production of cleistothecia by oleate in wild-type and hexAΔ but not in pexF23 strains. Discs (1 cm) saturated with 1 m oleate were placed in the middle of glucose minimal plates spread with conidia. Plates were taped to exclude air after 2 days and incubated for another 5 days and viewed with an inverted microscope. Cleistothecia coated with Hulle cells are seen at the periphery of the filters.

These effects on conidiation were observed in the velA1 mutant background that is standard for most laboratory work. In a velA⁺ background, conidiation is greatly reduced (Mooney and Yager 1990; Kim et al. 2002; Tsitsigiannis et al. 2004a). We therefore tested the effects of the pexF and pexC mutations on conidiation in velA⁺-containing strains and found virtually no conidiation (Figure 4C). Again, this was relieved by the presence of sorbitol. No effects of velA⁺ on growth inhibition by fatty acids were observed (not shown). The pex mutations affecting PTS1-localized proteins, but not the pexG or pexK mutations, were also observed to affect sexual reproduction with homozygous but not heterozygous crosses being subfertile with very few small fruiting bodies (cleistothecia) being produced, indicating a recessive defect. A velA⁺ background results in greatly increased cleistothecial production (Kim et al. 2002) and in this background pexF- and pexC-containing strains produced increased numbers of small cleistothecia (Figure 4D).

Woronin bodies that are used to plug septal pores as well as broken hyphal tips, preventing cytoplasmic leakage (Jedd and Chua 2000; Tenney et al. 2000; Tey et al. 2005), contain the Hex1 PTS1 protein, and Woronin bodies have been shown to be derived from peroxisomes in N. crassa (Yuan et al. 2003; Managadze et al. 2007). Therefore, the gene AN4695.3 (designated hexA), predicted to encode the Woronin body protein, was deleted (materials and methods). Growth and conidiation was only slightly affected by the hexAΔ (Figure 4E) and homozygous crosses were fully fertile (Figure 4F). The pexF23 hexAΔ double mutant was not more affected than the pexF23 single mutant in conidiation (Figure 4E). Sensitivity to sorbose has been observed in other fungi lacking the Woronin body protein and, although A. nidulans is highly sensitive to sorbose, some sensitivity of both pexF23 and hexAΔ mutants was observable (Figure 4E).

Signaling by the oxylipin Psi factors derived from fatty acids has been shown to affect the balance between asexual and sexual development (for review see Tsitsigiannis and Keller 2007). Oleate was found to enhance cleistothecial production in both wild-type and hexAΔ strains but not at all in a pexF23 mutant. This might indicate that peroxisomal fatty acid metabolism is required for oxylipin production but it cannot be excluded that this is the result of the inhibitory effects of oleate in pexF23. The pex mutants tested were also found to have reduced rates of conidial germination (Table 2). It is likely that this explains the initial slower growth and somewhat smaller colony size even on glucose media.

TABLE 2.

Conidial germination in pex mutants

		Strain
Time (hr)	Wild type	pexF2323	pexC∷bar	pexEΔ	pexG14	pexKΔ
4	53 (15)	23 (4)	10 (2)	30 (11)	8 (5)	13 (5)
6	100 (0)	90 (7)	70 (12)	86 (9)	80 (7)	93 (4)

Conidia were diluted and germinated in glucose minimal medium and viewed with an inverted microscope. Conidia visibly germinating were counted and are expressed as a percentage of the total. Values are the average of three replicates with standard errors shown in parentheses.

It was found that all pex mutants affected in PTS1 protein targeting, but not the pexG mutant, were auxotrophic for biotin. The pex mutants all responded to the addition of the intermediate pimelic acid while a biA1 mutant, predicted to lack biotin synthase, the final step in biotin synthesis, did not respond to pimelic acid. It has been shown that pimelic acid is synthesized from oleate and other long-chain fatty acids in some fungal species (Ohsugi et al. 1988).

The effects of mislocalization of glyoxylate cycle enzymes:

The unique enzymes of the glyoxylate pathway, ICL and MAS, are peroxisomal in plants and in some fungi (for review see Kunze et al. 2006), as shown here for A. nidulans. However, in S. cerevisiae, ICL is always cytoplasmic and MAS, which contains a PTS1, is only peroxisomal during growth on oleate, but this localization is not essential for oleate utilization (Kunze et al. 2002).

We investigated the effects of mislocalization of MAS on growth on acetate and fatty acids by mutating the acuE gene to eliminate the encoded C-terminal PTS1 (materials and methods). Cotransformation of an acuEΔ riboB2 strain with plasmids encoding wild-type or the PTS1 mutant AcuE, together with a riboB⁺-containing plasmid selecting for RiboB⁺, yielded ∼50% transformants capable of growth on acetate or oleate as carbon sources, indicating complementation of the acuEΔ. Complementation was similarly found for constructs expressing GFP-AcuE fusions with or without the PTS1. The PTS1 mutation did not affect growth on acetate and only slightly reduced growth on butyrate and long-chain fatty acids (Figure 5A). Growth on valerate was not affected by acuE mutations, consistent with one cycle of β-oxidation resulting in the formation of propionyl-CoA, which is metabolized via the methyl-citrate pathway, thereby bypassing the glyoxylate cycle (Brock et al. 2000; Brock 2005). The PTS1 mutation was found to result in mislocalization of GFP-AcuE to the cytoplasm (Figure 5B). These results indicated that, like S. cerevisiae, peroxisomal localization of MAS is not essential for a functional glyoxylate cycle during growth on either acetate or oleate (Kunze et al. 2006). However, in A. nidulans ICL is peroxisomal with the consequence that the two enzymes would be predicted to be in separate compartments in acuE PTS1 mutants. It is likely that localization of ICL is inefficient with some enzyme being present in the cytoplasm; this was supported by the observation of some cytoplasmic fluorescence in GFP-AcuD strains.

Figure 5.—

Effects of mislocalization of peroxisomal enzymes on growth on acetate and fatty acids. (A) Mislocalization of MAS encoded by acuE does not prevent growth on acetate or fatty acids. The phenotypes of representative cotransformants expressing GFP-AcuE with (SKL) and without (stop) PTS1 are compared with wild type and an acuE mutant. (B) Mislocalization of GFP-AcuE resulting from loss of PTS1. Mycelium for microscopy was grown in liquid glucose minimal media on coverslips. Similar results were observed for mycelia grown in acetate media and for five other transformants for each construct. (C) Growth inhibition on acetate and ethanol media in a pexEΔ mutant is partially restored in a double mutant with the pexG14 mutation.

Surprisingly, GFP-AcuD was mislocalized in a pexG but not a pexE background (Figure 1), indicating a requirement for the PTS2 receptor peroxin. However, growth on acetate or ethanol utilization was only slightly affected in the pexG14 mutant (Figure 5C), showing that peroxisomal localization of ICL is not required. However, pexEΔ strains, where AcuE and other PTS1 proteins are cytoplasmic, were inhibited on acetate and ethanol media (Figure 5C). Loss of targeting of all peroxisomal proteins in pexEΔ pexG14 double mutants resulted in improved growth on these media (Figure 5C) and this was also observed in pexEΔ double mutants with pexF23, pexA9, pexM15, and pexC∷bar (not shown). This indicated that, in the presence of acetate, ICL activity in the peroxisome in the absence of MAS and other PTS1 proteins might result in the toxic accumulation of glyoxylate.

Mitochondrial pathways for fatty acid utilization:

A mitochochondrial pathway for fatty acid β-oxidation has been discovered in A. nidulans by Maggio-Hall and Keller (2004). Deletion of the echA gene, encoding a mitochondrial enoyl-CoA hydratase, results in inhibited growth on butyrate, hexanoate as well as oleate and erucic acid. In contrast, deletion of scdA, predicted to encode a mitochondrial acyl-CoA dehydrogenase, does not result in inhibited growth on fatty acids but in an inability to utilize hexanoate and butyrate (Maggio-Hall et al. 2007). We have confirmed these results (Figure 6A). One of the mutants that we isolated was complemented by a clone containing a gene, AN10512.3, predicted to encode a mitochondrial ketoacyl-CoA thiolase (materials and methods). The phenotype of this mutant (designated mthA25) was identical to that of the scdAΔ: inability to grow on hexanoate and butyrate and no inhibition by long-chain fatty acids (Figure 6A). These results supported the existence of a mitochondrial pathway for β-oxidation of even-numbered short-chain fatty acids (Figure 6C). Accumulation of enoyl-CoA in an echA mutant results in growth inhibition, while the scdA and mthA mutations result in loss of a functional pathway without the accumulation of toxic metabolites. Consistent with this, an scdAΔ echAΔ double mutant is not inhibited by fatty acids (Maggio-Hall et al. 2007) and we have isolated a putative scdA mutation as a suppressor of the echAΔ (our unpublished data).

Figure 6.—

Effects of mutations in the mitochondrial short-chain fatty acid β-oxidation pathway. (A) Growth on even- but not odd-numbered chain fatty acids is inhibited in the echAΔ mutant. Growth on butyrate and hexanoate but not valerate or long-chain fatty acids is greatly reduced in the scdAΔ and mthA25 mutants. (B) The hadA disruption mutant has properties similar to echAΔ. Growth was for 2–3 days at 37°. (C) Proposed mitochondrial β-oxidation pathway. Accumulation of intermediates in echA and hadA mutants leads to growth inhibition.

The gene AN7008.3 is predicted to encode a mitochondrial hydroxy-acylCoA dehydrogenase. We cloned this gene (designated hadA) and generated a disruption mutant by insertion of the riboB gene (materials and methods). This mutation resulted in a phenotype similar to the echAΔ: inhibition by butyrate and the long-chain fatty acids elaidic, oleate, and erucic (Figure 6B). It was found that scdAΔ hadA∷riboB double mutants were not sensitive to butyrate and oleate. These results showed that the predicted hydroxyacyl-CoA dehydrogenase functions in this pathway and that accumulation of the hydroxyacyl-CoA substrate results in inhibition (Figure 6C). The mutants affected in the mitochodrial pathway do not prevent growth on the odd-numbered valerate (C5) nor is inhibition observed with tridecanoic (C13) or heptadecanoic (C17) (Figure 6, A and B).

DISCUSSION

By the isolation of mutants unable to use the short-chain fatty acid butyrate as a carbon source, we have identified mutants affected in peroxisomal functioning. These include pexA and pexF mutants predicted to be defective in the recycling of PTS receptors to the cytoplasm due to a loss of the AAA ATPase complex as well as a pexM mutant defective in an essential protein of the membrane complex necessary for docking of receptor–cargo protein complexes (for review see Platta and Erdmann 2007). An additional mutant due to a loss-of-function mutation in pexG, predicted to encode the receptor for PTS2 proteins, results in an inability to target a PTS2-containing protein but is unaffected in PTS1 protein localization. Deletion of the pexE gene results in loss of PTS1 but not PTS2 protein import, showing that the gene product is the specific PTS1 receptor. These results confirm the proposal that the PTS receptors function independently in filamentous fungi unlike in mammals and plants (see Kiel et al. 2006). The pexC gene (AN2281.3) is predicted to encode a protein orthologous to Pex3, which together with Pex16 and Pex19, have been found to be essential for peroxisomal membrane formation in S. cerevisiae and mammals (Hettema et al. 2000; Hoepfner et al. 2005; Kragt et al. 2005; Fujiki et al. 2006). Disruption of pexC resulted in mislocalization of both PTS1 and PTS2 proteins as well as the inability to form peroxisomal structures as visualized by PexK-GFP.

Developmental phenotypes of mutants:

Complete loss of import of matrix proteins, or indeed an inability to form peroxisomes as in the pexC mutant, does not result in extreme phenotypes under laboratory conditions. Mutants are able to form colonies both on rich complete media and on minimal media where the only essential function missing is the ability to synthesize pimelic acid in the biotin biosynthetic pathway.

Incomplete developmental defects were observed in the mutants unable to localize PTS1 proteins. Conidiation was significantly reduced and homozygous crosses produced very few small cleistothecia. In velA⁺ backgrounds, where the balance between sexual and asexual development is altered (Kim et al. 2002), conidiation was further reduced and production of cleistothecia increased, indicating that these effects were independent of the VelA-signaling pathway. In view of the major role for lipid metabolism in the generation of oxylipin-signaling molecules regulating development as well as secondary metabolism (Tsitsigiannis et al. 2004b, 2005; Tsitsigiannis and Keller 2007), it is surprising that a complete lack of peroxisomes does not have a more drastic effect on A. nidulans. The observation of an almost complete absence of conidiation in wild-type strains growing on long-chain fatty acids as sole carbon sources might reflect gross alterations in oxylipin factor signaling. Only a minor conidiation defect was observed in a mutant lacking the HexA Woronin body protein. In the heterothallic fungus Podospora anserina, loss of Pex2 results in an inability to switch from mitotic to meiotic development (Berteaux-Lecellier et al. 1995) and pex5Δ mutants have nuclear, mitochondrial, and sexual abnormalities (Bonnet et al. 2006). Interestingly, a pex7Δ is epistatic to the effects of the pex5Δ for all phenotypes, raising the possibility that mislocalization of only PTS1 proteins is more detrimental than mislocalization of all peroxisomal proteins (Bonnet et al. 2006). The developmental phenotype of the pexEΔ described here is not more extreme than that of other pex mutants or the pexEΔ pexG14 double mutant. Reactive oxygen species have been found to be important in sexual development in A. nidulans (Scherer et al. 2002; Lara-Ortiz et al. 2003). It is possible that some developmental abnormalities in the pex mutants result from impairment of normal reactive oxygen metabolism.

Fungal sexual and asexual spores have high lipid contents (e.g., Goodrich-Tanrikulu et al. 1998; Tsitsigiannis et al. 2004b; Schöbel et al. 2007). Messenger RNAs for glyoxylate cycle and gluconeogenic genes are found in A. nidulans conidia (Osherov and May 2000, 2001). Although rates of conidial germination in pex mutants are reduced, both conidia and ascospores of the pex mutants can germinate, showing that peroxisomal fatty acid metabolism is not essential for signaling the initiation of germination and that lipids do not provide the only initial carbon source for growth. It is likely that the other major source of carbon in the spore is trehalose (Fillinger et al. 2001).

Regulation of peroxisome proliferation:

The number of peroxisomes greatly increases in response to both short- and long-chain fatty acids and this is dependent on PexK. GFP-PexK labels peroxisomes in the absence of proliferation, indicating that it is a normal component of the peroxisomal membrane. Strains containing the pexKΔ grow more poorly than wild type on both long- and short-chain fatty acids but no inhibition by fatty acids is observed and no detectable growth or developmental defects are observed in pexKΔ strains. Overexpression of Pex11 in Penicillium chrysogenum has been found to result in peroxisome proliferation (Kiel et al. 2005). While in S. cerevisiae there is one Pex11 ortholog required for peroxisome proliferation (Erdmann and Blobel 1995), multiple Pex11 proteins are found in humans and plants (Li et al. 2002; Lingard and Trelease 2006). It has been suggested that there are two additional paralogs of Pex11 in A. nidulans as well as in A. fumigatus and P. chrysogenum (Kiel et al. 2006). It is not known if these additional proteins play a role in the response to fatty acids or in the response to other stimuli.

The expression of pexK is induced in response to fatty acids but not acetate, and this is dependent on the trancription factors ScfA and FarB for short-chain fatty acid induction and on FarA for both long- and short-chain induction (Hynes et al. 2006). Both FarA and FarB have been found to bind in vitro to DNA containing the core CCGAGG sequence found in the 5′ upstream region of pexK and other genes inducible by fatty acids. The greatly reduced peroxisome proliferation observed in farAΔ is consistent with proliferation dependent on fatty acid induction of PexK expression. The 5′ sequences of many A. nidulans pex genes contain one or more copies of the CCGAGG-binding sequence. These genes include pexE and pexG encoding the specific PTS receptors well as pexC and AN5113.3 (the predicted Pex16 ortholog) but not AN4899.3 (Pex19), all of which are required for peroxisome biogenesis from the endoplasmic reticulum (Hettema et al. 2000; Fujiki et al. 2006). Genes for enzymes of β-oxidation (both mitochondrial and peroxisomal), the methyl-citrate pathway of propionyl-CoA metabolism, lipases, and cutinases also have 5′-CCGAGG sequences (Hynes et al. 2006). Therefore, during growth in the presence of fatty acids it is predicted that there is a major reorganization of metabolism with expansion of the number of peroxisomes containing enzymes relevant to fatty acid catabolism resulting from gene induction.

The role of peroxisomes in acetate utilization:

In S. cerevisiae, the unique enzymes of the glyoxylate cycle, MAS and ICL, are cytoplasmic during growth on acetate or ethanol and pex mutants are able to grow on these carbon sources (Kunze et al. 2006). In A. nidulans, ICL and MAS are present in peroxisomes independently of the carbon source. However, we have shown that mutants with loss of peroxisomal targeting of matrix proteins, loss of peroxisome biogenesis, or peroxisome proliferation are able to grow on acetate. Mislocalization of MAS resulting from loss of the PTS1 of AcuE does not abolish growth on acetate. AcuD targeting requires the PTS2 receptor PexG and yet the pexG14 mutant can grow on acetate, indicating that ICL does not have to be peroxisomal. However, pexG-dependent ICL localization results in the pexE mutant having ICL in peroxisomes in the absence of MAS as well as other PTS1-containing proteins. This results in inhibited growth on acetate and ethanol, which is restored by loss of targeting of ICL in double mutants with other pex mutations. This might be explained if peroxisomal ICL activity results in the production of glyoxylate, which, in the absence of the PTS1-dependent MAS, cannot be used for malate production and is unable to escape the peroxisome, resulting in toxicity. This scenario depends on the import into the peroxisome of isocitrate generated from citrate in the cytoplasm by aconitase.

Peroxisomes and fatty acid utilization:

Complete loss of protein targeting to peroxisomes or a lack of peroxisomes results in growth inhibition by fatty acids as does the mislocalization of PTS1-containing proteins in the pexEΔ mutant. The pexG14 mutant, where only PTS2 proteins are mislocalized, also resulted in reduced growth, but not inhibition, on fatty acid medium with the greatest effects observed with short-chain compared with long-chain fatty acids. The lack of a strong phenotype of pexG14 on long-chain fatty acids might reflect the involvement of multiple ketoacyl-CoA thiolases with either PTS1 or PTS2 signals. Similar effects of pex mutants in P. anserina have been reported with inhibited growth on oleate in pex2 and pex5 mutants and reduced oleate utilization in a pex7 mutant (Bonnet et al. 2006). Magnaporthe grisea and Colletotrichum lagenarium pex6 deletions show complete loss of growth on long-chain fatty acid media (Kimura et al. 2001; Ramos-Pamplona and Naqvi 2006). One possible mechanism for fatty acid inhibition is that cytoplasmic β-oxidation enzymes result in the generation of hydrogen peroxide, which cannot satisfactorily be detoxified by mislocalized peroxisomal catalase (Kawasaki and Aguirre 2001) or glutathione peroxidase (AN5440.3) as suggested for P. anserina (Ruprich-Robert et al. 2002). We have previously reported that downregulation of the expression of genes involved in fatty acid breakdown leads to some resistance to oleate inhibition of a pexFΔ strain by the isolation of mutations in the farA regulatory gene as partial suppressors of this phenotype (Hynes et al. 2006). An additional possibility is that pex mutants are sensitive to fatty acids because of pertubations in membrane composition as observed for S. cerevisiae mutants (Lockshon et al. 2007). However, this is not an obvious consequence of exposure to short-chain fatty acids and, in the S. cerevisiae studies, sensitivity was observed in a pex7Δ, which was not observed here.

Further evidence for peroxisomal metabolism of fatty acids is provided by the requirement for PexK, which is necessary for peroxisomal proliferation. However, the effects of pexKΔ are not complete with growth being less affected on long- than on short-chain fatty acids, indicating that proliferation is not absolutely essential. In addition, mutation of the antA gene predicted to encode a peroxisomal ATP carrier necessary for β-oxidation (Palmieri et al. 2001; van Roermund et al. 2001) results in greatly reduced growth on all fatty acids.

In our screen for butyrate-nonutilizing mutants we have failed to isolate any affected in peroxisomal enzymes and previous attempts to isolate mutants in A. nidulans have not yielded any lacking long-chain β-oxidation enzymes although an uncharacterised pex mutant was isolated (Kawasaki et al. 1995; De Lucas et al. 1997). A deletion mutant of foxA encoding a peroxisomal multifunctional enzyme is unable to grow on erucic acid (C22) but is only partially defective for growth on oleate and is unaffected for growth on short-chain fatty acids (Maggio-Hall and Keller 2004). Furthermore, analysis of the genome sequences of A. nidulans and other filamentous fungi reveals the presence of many possible genes coding for β-oxidation enzymes with potential PTS sequences, and deletion or disruption of some of these has failed to result in complete loss of growth on any one fatty acid (K. Reiser, M. A. Davis and M. J. Hynes, unpublished results). It is likely that filamentous fungi have evolved multiple pathways allowing growth on a diversity of fatty acid substrates.

The normal situation for fatty acid utilization in A. nidulans is the production of acetyl-CoA by β-oxidation and its metabolism via the glyoxylate cycle within peroxisomes, resulting in the production of malate and succinate. It is not clear how these metabolites exit the peroxisome and in all organisms there are many unanswered questions relating to metabolite trafficking between the cytoplasm, peroxisomes, and mitochondria (for discussion see Kunze et al. 2006; Visser et al. 2007).

Mitochondrial β-oxidation:

Our data suggest that a complete peroxisomal β-oxidation pathway exists as for S. cerevisiae. However, as shown by Maggio-Hall and Keller (2004), an additional mitochondrial pathway exists. Mutations of either echA or hadA genes lead to inhibited growth on short-chain fatty acids as well as on even-numbered long-chain fatty acids due to accumulation of toxic intermediates. With even-numbered long-chain fatty acids, butyryl-CoA produced in the penultimate cycle of peroxisomal β-oxidation can enter the mitochondria for β-oxidation, accounting for the toxicity observed. The lack of butyrate and hexanoate utilization (but not inhibition) in the scdA and mthA mutants shows that the mitochondrial pathway is essential for growth on these fatty acids. However, the scdA and mthA mutants do not affect growth on long-chain fatty acids, indicating that this mitochondrial pathway is not required for their utilization. A possible reason for an essential mitochondrial short-chain β-oxidation pathway is that only one or two cycles of the peroxisomal pathway are insufficient for efficient carbon flux and so the direct generation of acetyl-CoA within the mitochondrion is required for energy production. The mitochondrial pathway is also required for the utilization of the amino acids valine and isoleucine as carbon sources (Maggio-Hall et al. 2007).

Mutants affected in the mitochondrial pathway are able to grow on the odd-numbered valerate (C5) and on tridecanoic (C13) and heptadecanoic (C17). This is consistent with valerate not being a substrate for the defined mitochondrial pathway but undergoing one round of β-oxidation in the peroxisome to produce acetyl-CoA and propionyl-CoA. The final cycle of β-oxidation of odd-numbered long-chain fatty acids would also result in acetyl-CoA and propionyl-CoA in the peroxisome with propionyl-CoA shuttling to the mitochondria where it can be metabolized via the methyl-citrate pathway (Brock et al. 2000; Brock 2005). The possibility that there is a distinct mitochondrial β-oxidation pathway for valerate remains.