Promoter Trapping of a Novel Medium-chain Acyl-CoA Oxidase, Which Is Induced Transcriptionally during Arabidopsis Seed Germination*

The first step of peroxisomal fatty acid β-oxidation is catalyzed by a family of acyl-CoA oxidase isozymes with distinct fatty acyl-CoA chain-length specificities. Here we identify a new acyl-CoA oxidase gene from Arabidopsis(AtACX3) following the isolation of a promoter-trapped mutant in which β-glucuronidase expression was initially detected in the root meristem. In acx3 mutant seedlings medium-chain acyl-CoA oxidase activity was reduced by 95%, whereas long- and short-chain activities were unchanged. Despite this reduction in activity lipid catabolism and seedling development were not perturbed.AtACX3 was cloned and expressed in Escherichia coli. The recombinant enzyme displayed medium-chain acyl-CoA substrate specificity. Analysis of β-glucuronidase activity inacx3 revealed that, in addition to constitutive expression in the root axis, AtACX3 is also up-regulated strongly in the hypocotyl and cotyledons of germinating seedlings. This suggests that β-oxidation is regulated predominantly at the level of transcription in germinating oilseeds. After the discovery ofAtACX3, the Arabidopsis acyl-CoA oxidase gene family now comprises four isozymes with substrate specificities that encompass the full range of acyl-CoA chain lengths that exist in vivo.

The first step of peroxisomal fatty acid ␤-oxidation is catalyzed by a family of acyl-CoA oxidase isozymes with distinct fatty acyl-CoA chain-length specificities. Here we identify a new acyl-CoA oxidase gene from Arabidopsis (AtACX3) following the isolation of a promotertrapped mutant in which ␤-glucuronidase expression was initially detected in the root meristem. In acx3 mutant seedlings medium-chain acyl-CoA oxidase activity was reduced by 95%, whereas long-and short-chain activities were unchanged. Despite this reduction in activity lipid catabolism and seedling development were not perturbed. AtACX3 was cloned and expressed in Escherichia coli. The recombinant enzyme displayed mediumchain acyl-CoA substrate specificity. Analysis of ␤-glucuronidase activity in acx3 revealed that, in addition to constitutive expression in the root axis, AtACX3 is also up-regulated strongly in the hypocotyl and cotyledons of germinating seedlings. This suggests that ␤-oxidation is regulated predominantly at the level of transcription in germinating oilseeds. After the discovery of AtACX3, the Arabidopsis acyl-CoA oxidase gene family now comprises four isozymes with substrate specificities that encompass the full range of acyl-CoA chain lengths that exist in vivo.
Peroxisomal ␤-oxidation is the primary pathway of fatty acid catabolism in plants. The pathway plays a fundamental role in breaking down stored lipid reserves to provide metabolic energy and carbon skeletons during processes such as oilseed germination, leaf senescence, and starvation (1,2). ␤-Oxidation may also play a constitutive role in membrane lipid turnover and be involved in the synthesis of important fatty acid-derived signals such as jasmonic acid (3) and traumatin (4). Recently an Arabidopsis mutant (aim1) has been isolated that is deficient in a multifunctional protein isoform (5). The fact that aim1 displays an altered inflorescence meristem phenotype has lead to the suggestion that ␤-oxidation may be involved in flower development (5).
Acyl-CoA oxidase (ACX, 1 EC 1.3.3.6) catalyzes the conversion of fatty acyl-CoAs to trans-2-enoyl-CoAs. The reaction requires FAD as a cofactor, which is subsequently re-oxidized by O 2 to form H 2 O 2 . This first step is believed to be predominant in exerting control over the rate of carbon flux through the pathway (7,8). Biochemical evidence suggests that plants contain a family of acyl-CoA oxidase isozymes with distinct but partially overlapping substrate specificities (9 -11). cDNA clones of several acyl-CoA oxidase homologues have been identified in plants (12)(13)(14).
However, direct evidence of their identity has only recently been obtained by overexpression and characterization of the recombinant proteins. Hooks et al. (15) have identified and characterized two long-chain acyl-CoA oxidases from the oilseed Arabidopsis thaliana (AtACX1 and AtACX2). The preferred substrate of AtACX1 is myristoyl-CoA (C14:0), whereas that of AtACX2 is oleoyl-CoA (C18:1). In addition Hayashi et al. (16) have recently described a gene (referred to here as At-ACX4) that encodes an enzyme with specificity for the shortchain substrate hexanoyl-CoA (C6:0). A comparison of the substrate specificities of known acyl-CoA oxidases from Arabidopsis suggests that a gene encoding a medium-chain acyl-CoA oxidase with a preference for decanoyl-CoA (C10:0) and lauroyl-CoA (C12:0) remains to be discovered (15).
Three acyl-CoA oxidase genes, a multifunctional protein, and a 3-keto acyl-CoA thiolase are up-regulated co-ordinately dur-ing Arabidopsis seed germination and post-germinative growth, correlating with the period of most rapid fatty acid degradation (Refs. 15-18, respectively). In addition genes encoding glyoxylate cycle and gluconeogenesis enzymes are also induced during oilseed germination (19,20). Although transcription is known to play a major role in the regulation of the glyoxylate cycle (19,21,22), the level at which peroxisomal ␤-oxidation is regulated has not yet been determined.
The regulation of ␤-oxidation in plants may be important for a variety of biotechnological applications. In the majority of cases where crops have been genetically engineered to produce novel fatty acids, the accumulation of these products is much lower than that required for commercial exploitation (23,24). It is believed that in these plants the novel fatty acids are synthesized at significant rates but are subsequently degraded by peroxisomal ␤-oxidation (25,26). Moreover, Eccleston and Ohlrogge (25) provide evidence that in the case of plants producing medium-chain fatty acids, medium-chain acyl-CoA oxidase activity is up-regulated to facilitate this degradation. The downregulation of acyl-CoA oxidases may therefore promote the accumulation of unusual fatty acids in genetically modified crops.
In this study we identify and characterize a promotertrapped Arabidopsis mutant disrupted in a gene encoding a new member of the acyl-CoA oxidase family with mediumchain substrate specificity. This gene is up-regulated at the level of transcription during seed germination. Combined with the three genes previously identified and characterized, our data reveal that Arabidopsis comprises a family of isoforms that together are capable of utilizing the full range of fatty acyl-CoA chain lengths present in vivo. Isolation of T-DNA Flanking Sequences-The sequence flanking the right border of the T-DNA inserts in lines displaying GUS expression in the roots was obtained by inverse PCR. Genomic DNA was digested with EcoRI and ligated using T4 DNA ligase. PCR was then carried out using the primers GUS1 (5Ј-CCAGACTGAATGCCCACACGCCGTC) and inverse PCR 2200 (5Ј-GTATCACCGCGTCTTTGATCGCGTC). The conditions for amplification were 15 s at 94°C, 30 s at 65°C, and 30 s at 68°C, repeated 38 times followed by a 2-min extension. The products were sequenced directly using the nested primer GUS2 (5Ј-TCACGG-GTTGGGGTTTCTACAGGAC).

Plant Material and Growth Conditions-A T-DNA-mutagenized
Isolation of cDNA Clones-A partial clone of AtACX3 was isolated from a ZAPII cDNA library constructed from 3-day-old etiolated hypocotyls (30) by 3Ј RACE using the gene specific primer ACX3S (5Ј-AGGAGTATTTATCACAAACTCTTCAG) in combination with the T7 primer. The library was then screened by colony hybridization according to Sambrook et al. (31) using the partial clone as a probe (15). Four full-length clones were identified and sequenced. RNA was prepared from 2-day-old seedlings (see below), and the 5Ј-UTR was mapped by 5Ј RACE using the 5Ј RACE System Version 2 (Life Technologies, Inc.) following the manufacturer's protocols. GSP1 and GSP2 were (5Ј-ACT-TTACAGCATTACCCCACAGGA) and (5Ј-GTGATCATAAATCCCGCA-AACCT), respectively. Three 5Ј RACE products from separate PCR reactions were cloned and sequenced.
Expression of AtACX3-One AtACX3 cDNA clone was determined to be in-frame with the N terminus of the pBluescript ␤-galactosidase gene and could be expressed as a fusion protein following induction of Escherichia coli XL1-blue MRF TM cells with isopropyl-␤-D-thiogalactoside. Cells were grown at 37°C to an optical density of 0.5 at 600 nm in Luria broth media. Isopropyl-␤-D-thiogalactoside (0.4 mM) was then added to the culture, and the cells grown overnight at 28°C. The culture was centrifuged at 700 ϫ g for 10 min, and the pellet was resuspended in extraction buffer (150 mM Tris/HCl, pH 7.5, 10 mM KCl, 1 mm EDTA, 1 mm dithiothreitol, 10 mm FAD, 10% (v/v) glycerol). The cells were lysed by sonication using a Soniprep 150 ultrasonic disintegrator (Sanyo Gallenkamp PLC, Leicester, UK), and cell debris was removed by centrifugation at 21,000 ϫ g for 10 min. Assays were performed on the supernatant. Tissue Extraction and Subcellular Fractionation-Crude tissue extracts were prepared from approximately 2000 2-day-old seedlings. The tissue was ground in 1 ml of extraction buffer using a glass homogenizer. The extract was then centrifuged at 13,000 ϫ g for 10 min, and the supernatant was desalted using a Sephadex G-50 spin column as described by Hooks et al. (15).
For subcellular fractionation experiments, approximately 10,000 2-day-old seedlings were homogenized in 3 ml of a medium containing: 150 Tricine/KOH, pH 7.5, 1 mM EDTA, 0.5 M sucrose using an Ultra-Turrax (Janke and Kunkel KG). The homogenate was filtered through four layers of mira cloth, and 2 ml of homogenate was layered on top of a gradient consisting of a 1-ml cushion of 60% (w/v) sucrose plus 8 ml of a 60 to 30% linear sucrose gradient. The gradient was centrifuged at 30,000 ϫ g for 3 h in a Sorvall OTD55B centrifuge using a TST 41.14 swing out rotor. 0.5-ml fractions were removed and assayed for various enzyme activities. The sucrose concentration of the fractions was determined using a refractometer.
Enzyme Assays, Histochemical Staining, and Fatty Acid Measurement-ACX assays were performed on plant tissue and bacterial cell extracts according to the method of Hryb and Hogg (32) using 50 M acyl-CoAs as substrate. GUS assays and histochemical staining were carried out as described by Jefferson (33). Catalase and cytochrome c oxidase were assayed according to Takahashi et al. (34) and Denyer and Smith (35), respectively. Protein content was determined as described by Bradford (36) using bovine serum albumin as a standard. Fatty acids were measured using the method of Browse et al. (37).
Northern Analysis, Southern Analysis, and Reverse Transcriptase-PCR-Total RNA from various tissues was isolated using the Purescript RNA isolation kit (Flowgen) or the hot phenol method (38). Ten g of total RNA was separated by electrophoresis using a 1.1% formaldehyde gel and alkaline-blotted onto Zeta-Probe membrane (Bio-Rad) using 50 mM NaOH. Genomic DNA was purified using the Puregene DNA isolation kit (Flowgen). Three g of DNA were digested using EcoRI, separated on a 0.8% agarose gel, and blotted using 0.4 M NaOH. Probes were prepared from AtACX3 and gusA, and membranes were hybridized using the digoxygenin system (Roche Molecular Biochemicals) following the manufacturer protocols. Bands were detected using a ChemiImager (Alpha Innotech Corp.). Reverse transcriptase-PCR was performed using the Reverse-iT kit from Advanced Biotechnologies Ltd. The primers used were GUS1, ACX3S, and ACX3A (5Ј-GAAACAT-CAGCAACTGCATTCAAC).

RESULTS
Mutant Isolation-As part of a research program focused on root-rhizosphere interactions, Arabidopsis promoter trapped lines that displayed GUS expression in the roots of 7-day-old seedlings were isolated from the INRA-Versailles T-DNA-mutagenized population (27). Genomic DNA sequences that flank the right borders of the T-DNAs were obtained by inverse PCR. One of two inverse PCR products from a line with GUS expression in the root meristem (Rm 328) was found to be homologous to acyl-CoA oxidases. These enzymes catalyze the first step in the pathway of peroxisomal ␤-oxidation and are known to be highly active in the root tips of maize (2). This putative acyl-CoA oxidase was different from those previously identified in Arabidopsis, and the new gene and mutant were designated AtACX3 and acx3, respectively.
Cloning of AtACX3-A partial AtACX3 cDNA clone was isolated by performing 3Ј RACE on a cDNA library constructed from 3-day-old etiolated seedling hypocotyls (30). This clone was used as a homologous probe to screen the same library by colony hybridization. Four full-length cDNAs were isolated and sequenced. Finally the 5Ј-UTR was mapped using 5Ј RACE. The assembled sequence was submitted to the GenBank data base. The cDNA was 2246 bp long and contained a 2028-bp putative open reading frame (Fig. 1). The 5Ј-UTR extended 53 bp 5Ј of the start of translation. The 3Ј-UTR contained a putative polyadenylation signal (AATAAA) 96 bp 3Ј of the stop codon.
The deduced protein was 675 amino acids long (Fig. 1) with a calculated molecular mass of 75676.33 Da and an isoelectric point (pI) of 8.17. Both the molecular weight and pI values are similar to those of long-chain acyl-CoA oxidases from Arabidopsis (15). A similarity search of available data bases revealed that expressed sequence tags homologous to AtACX3 (Ͼ70% amino acid identity) are present in a variety of plant species including Glycine max and Gossypium hirsutum. Comparison of the Arabidopsis acyl-CoA oxidase proteins shows that At-ACX3 shares 28, 23, and 14% identity at the amino acid level with AtACX2, AtACX1 (15), and AtACX4 (16) (Fig. 1). Consensus motifs characteristic of type 1 or 2 peroxisomal targeting signals (14,40) are not obvious in the AtACX3 amino acid sequence.
Characterization of the acx3 Locus and Expression of gusA-Southern analysis using a probe to gusA revealed that acx3 contained three copies of the T-DNA (data not shown). As shown in Fig. 2A, one copy was segregated out by back-crossing acx3. The remaining two copies formed a tandem inverted repeat inserted in AtACX3. In Fig. 2B this was demonstrated by a PCR experiment on acx3 genomic DNA. Primers 5Ј or 3Ј of the site of insertion in AtACX3 were used in combination with a gusA primer to demonstrate that gusA is present at both borders. The PCR products were sequenced, and comparison with the AtACX3 cDNA sequence revealed that the insertion is situated in an exon, 806 bp 3Ј of the putative start of transcription.
In Fig. 2B a reverse transcriptase-PCR experiment using the same primer combinations on RNA from 2-day-old acx3 seedlings demonstrated also that the 5Ј end of AtACX3 is expressed in vivo as a gusA transcriptional fusion. In contrast, the gusA copy bordering the 3Ј end of AtACX3 was not expressed. No product was detected when primers specific to a region of AtACX3 that is 3Ј of the insertion site were used in combination (data not shown). These data showed that wild type transcripts are absent from acx3 mutant seedlings.
It has also been demonstrated that the post-germinative growth of Arabidopsis mutants disrupted in storage lipid breakdown can be prevented if exogenous sugars are not supplied to the seedling (18,42). To determine if the germination or post-germinative growth of acx3 is impaired, the rate of hypocotyl elongation and fatty acid breakdown were measured in etiolated seedlings in the absence of exogenous sucrose. The data in Fig. 4 show that the rate of hypocotyl growth and fatty acid breakdown were not significantly different from wild type over the course of 5 days following germination. Furthermore, no visible vegetative or reproductive phenotype was observed throughout the life cycle of acx3 plants.
Activity of AtACX3-To investigate whether the acx3 mutant displayed altered acyl-CoA oxidase activity, the enzyme was measured in 2-day-old germinating seedlings (Table I). Saturated acyl-CoAs ranging from 4 to 20 carbons in length were used at a saturating concentration (50 M). The acx3 mutant was almost deficient in medium-chain acyl-CoA oxidase activity (Ͻ5% wild type), whereas short-and long-chain activities were unchanged (Table I). The maximal difference between mutant and wild type activity was observed using lauroyl-CoA (C12:0) as substrate (Table I). In wild type seedlings, acyl-CoA oxidase activity declined with increasing substrate chain length. Long-chain acyl-CoA oxidase activity was approximately 8-and 10-fold lower than that of medium-and shortchain activities, respectively (Table I).
To confirm the function of AtACX3, the cDNA was expressed in E. coli as a ␤-galactosidase fusion protein transcribed from the pBluescript cloning vector upon induction with isopropyl-␤-D-thiogalactoside. As previously reported (15), no inducible acyl-CoA oxidase activity was observed in extracts of E. coli harboring pBluescript without insert and endogenous levels of acyl-CoA oxidase activity were below the limits of detection. There was also no induction when the AtACX3 cDNA is in the antisense orientation (data not shown). As shown in Fig. 5A, AtACX3 expressed in the correct orientation encoded a protein with medium-chain acyl-CoA oxidase activity. The optimum substrate was lauroyl-CoA (C12:0). No activity was detected with substrates of chain length greater than C14:0. In Fig. 5B, kinetic analysis of recombinant AtACX3 showed that the apparent K m value of the enzyme for lauroyl-CoA was 3.7 M. The optimum pH was between 8.5 and 9.0 and activity was dependent on the provision of the cofactor FAD (data not shown).
Subcellular Localization-To investigate whether AtACX3 is a peroxisomal protein, the subcellular location of mediumchain acyl-CoA oxidase activity was determined. The acx3 mutant specifically lacks this activity (Table I), making it a reliable marker for the subcellular localization of the protein in wild type. A homogenate of two-day-old wild type seedlings was fractionated on a sucrose density gradient by centrifugation to separate the subcellular compartments. Catalase and cytochrome c oxidase activities were used as peroxisomal and mitochondrial markers, respectively. As shown in Fig. 6, mediumchain acyl-CoA oxidase activity co-localized with that of catalase (fraction 16), suggesting that the majority of AtACX3 is located in the peroxisome. Both activities were present in the supernatant as well as in the peroxisomal fraction. This is likely to be due to a proportion of the organelles rupturing during the tissue homogenization step.
Expression of AtACX3-In Fig. 7, Northern blot analysis of total RNA from imbibed seeds, germinating seedlings, and various tissues from wild type plants showed that AtACX3 is expressed at low levels in all tissues but is up-regulated strongly during germination and leaf senescence. AtACX3 transcripts were detectable in imbibed seeds before radicle emergence. Steady-state AtACX3 mRNA levels increased during germination to a maximum between 2 and 3 days after imbibition (DAI) and subsequently decreased (Fig. 7A).
The analysis of GUS expression in acx3, displayed in Fig. 8A, revealed that the AtACX::gusA transcriptional fusion is ex-  pressed in the cotyledons, hypocotyl, and root tip of young seedlings. GUS expression in the root axis was constitutive, whereas the cotyledons and hypocotyl showed transient expression during early post-germinative growth. This is reflected by the change in the level of total GUS activity, as shown in Fig.  8B. The activity was detectable before radicle emergence, increased rapidly during germination, peaked at 2 to 3 days after imbibition (DAI), and then declined (Fig. 8B). This pattern of expression correlated positively with medium-chain acyl-CoA oxidase activity in wild type seedlings, although GUS levels in acx3 declined more slowly after the activity peaked at day 3 (Fig. 8B). The retention of GUS is likely to be a result of the relative stability of the protein in vivo. An increase in GUS expression was also observed in senescing leaves (data not shown).

DISCUSSION
In this study we identify a new member of the ACX gene family from Arabidopsis using a promoter-trapping strategy. Hooks et al. (2) previously showed that in maize, acyl-CoA oxidases are highly active in the root tips. A promoter-less GUS T-DNA line was identified in which GUS activity was localized to the root meristem of 7-day-old seedlings. This line contained an in vivo transcriptional fusion between the gusA gene and the 5Ј end of a new acyl-CoA oxidase homologue designated AtACX3.
The AtACX3 gene was cloned and characterized. A comparison of the predicted amino acid sequence with those of At-ACX1, AtACX2 (15), and AtACX4 (16) revealed that AtACX3 shares a significant level of homology (28 -14% identity). Furthermore, all four Arabidopsis acyl-CoA oxidase isoforms contain common motifs that are homologous to the conserved signatures from mammalian acyl-CoA dehydrogenases (39). It has been suggested that these regions may be important for interaction with the substrates (16).
Two independent lines of evidence demonstrated that At-ACX3 encodes an acyl-CoA oxidase with medium-chain acyl-CoA substrate specificity. First, the acx3 mutant was specifically deficient in medium-chain activity, whereas long-and short-chain activities were normal. Second, when expressed in E. coli, the recombinant protein displayed maximum activity using medium-chain acyl-CoAs as a substrate. The K m value, pH optima, and cofactor requirements of AtACX3 are very similar to those previously reported for both long-and shortchain acyl-CoA oxidases from Arabidopsis (15,16).
Subcellular fractionation experiments indicate that AtACX3 is localized to the peroxisome. However, AtACX3 lacks motifs clearly diagnostic of either PTS1 or PTS2 (14,40). Both At-ACX1 and AtACX4 contain a C-terminal type 1 peroxisomal targeting signal (PTS1), whereas AtACX2 is likely to be targeted via an N-terminal type 2 motif that is cleaved on transit (PTS2) (15,16). Further work will be required to establish the mechanism of AtACX3 import into the peroxisome.
Interestingly, despite containing less than 5% of wild type medium-chain acyl-CoA oxidase activity, acx3 mutant seeds were able to germinate, and the seedlings developed normally. No inhibition of storage lipid breakdown was observed following germination. This is in contrast to ped1, a mutant disrupted in a putative 3-keto acyl-CoA thiolase, which is incapable of post-germinative growth without the provision of exogenous sugar (16). It is probable that the residual mediumchain acyl-CoA oxidase activity detected in acx3 is derived from other acyl-CoA oxidase isoforms whose substrate specificities partially overlap AtACX3 (15,16). This activity is very small but apparently sufficient to allow fatty acid degradation. It can be concluded that medium-chain acyl-CoA oxidase activity exerts little control over the breakdown of storage lipid in germinating seeds. Long-chain acyl-CoA oxidase activity is about 8-fold less than that of medium-chain acyl-CoA oxidase in wild type seedlings (Table I) and leaves (26). If acyl-CoA oxidase plays a significant role in the control of peroxisomal ␤-oxidation (7,8), then long-chain acyl-CoA oxidase is likely to predominate.
Disruption of AtACX3 causes resistance to the pro-herbicide 2,4-DB. This chemical is metabolized to the herbicide (and auxin analogue) 2,4-D by ␤-oxidation (5,18,41). Long-and short-chain acyl-CoA oxidase activities are unchanged in acx3, suggesting that 2,4-DB is perceived as a medium-chain substrate and selectively metabolized by medium-chain acyl-CoA oxidase in wild type seedlings. acx3 is sensitive to 2,4-DB concentrations greater than 2 M, whereas a second ped1 mutant allele (ped1-2) is resistant to 10 M 2,4-DB. 2 Furthermore Hayashi et al. (18) report that ped1 is a dominant mutation with respect to 2,4-DB resistance. In contrast, acx3 is recessive. These differences are consistent with the observation that acx3 retains a low level of medium-chain acyl-CoA oxidase activity and, therefore, a diminished rather than blocked capacity to metabolize 2,4-DB.
In addition to constitutive expression in the root tip, AtACX3 was also strongly induced in the cotyledons and hypocotyl of seedlings during germination and in senescing leaves. These tissues are highly active in fatty acid breakdown (1). The temporal and spatial pattern of AtACX3 expression is similar to other Arabidopsis genes encoding enzymes of peroxisomal ␤-oxidation. These include three additional acyl-CoA oxidase isoforms, a multifunctional enzyme (AtMFP2), and a 3-keto acyl-CoA thiolase (Refs. 15-18, respectively). However, not all the genes of ␤-oxidation are co-ordinately expressed during germination. For example AtMFP1 transcripts are more abundant in mature plant tissues than in young seedlings (5).
The glyoxylate-cycle enzymes malate synthase and isocitrate lyase are co-ordinately induced during Arabidopsis seed germination (42). Reporter gene fusion experiments with both MS and ICL have revealed that the induction of the glyoxylate cycle occurs at the level of transcription in germinating oilseeds (19,21,22). Importantly, in this study the activity of the GUS reporter in acx3 correlated positively with medium-chain acyl-CoA oxidase activity levels during germination and early post-germinative growth in the wild type, showing that AtACX3 is also transcriptionally regulated. This is the first demonstration that an enzyme of peroxisomal ␤-oxidation is primarily controlled at the level of transcription during germination.
In mammals and yeast the enzymes of peroxisomal ␤-oxidation are co-ordinately regulated by lipid-based signals (either free fatty acids or acyl-CoAs) (43)(44)(45). This regulation is mediated via transcription but in each case involves a unique signal transduction pathway (43,44,46). Recently Eccleston and Ohlrogge (25) reported that, in transgenic Brassica napus embryos engineered to synthesize unusual medium-chain fatty acids, the activity of medium-chain acyl-CoA oxidase is up-regulated. It will be interesting to establish if metabolic signaling from fatty acids (or acyl-CoAs) regulates peroxisomal ␤-oxidation in higher plants. Alternatively, different mechanisms may operate that involve unique plant signaling molecules.
In conclusion we identified and characterized a promoter trapped Arabidopsis mutant disrupted in a gene encoding a new member of the acyl-CoA oxidase family with mediumchain substrate specificity. This enzyme is induced transcriptionally during germination and is also expressed constitutively in the root axis. The discovery of AtACX3 fills a gap in our knowledge of the acyl-CoA oxidase gene family that was apparent from a comparison of the substrate specificities of previously characterized genes (15). In combination, the four acyl-CoA oxidase isoforms identified to date can metabolize a broad range of acyl-CoA chain lengths (C4:0 to C22:0). It will be interesting to determine if this gene family is now complete or whether new acyl-CoA oxidase isoforms have yet to be identified in Arabidopsis.