Peroxisome biogenesis occurs in late dorsal-anterior structures in the development of Xenopus laevis



Metabolism and development are two important processes not often examined in the same context. The focus of the present study is the expression of specific peroxisomal genes, the subsequent biogenesis of peroxisomes, and their potential role in the metabolism associated with the development of Xenopus laevis embryos. The temporal and expression patterns of six peroxisomal genes (PEX5, ACO, PEX19, PMP70, PEX16, and catalase) were elucidated using RT-PCR. Functionally related peroxisomal genes exhibited similar expression patterns with their RNA levels elevated relatively late during embryogenesis. Using immunohistochemistry PMP70 and catalase protein was localized largely to dorsal-anterior structures. Peroxisomal function was assayed with peroxisomal targeted-GFP, which when microinjected, revealed peroxisomes in dorsal-anterior structures at stage 45. A requirement for peroxisomal function appears to be present only late in development as organogenesis is finishing, yolk stores are depleted, and ingestion commences. Developmental Dynamics 236:3554–3561, 2007. © 2007 Wiley-Liss, Inc.


While our knowledge of metabolic processes and their regulation is extensive, their role in development receives little attention. Embryonic development is one of the most fragile periods an organism endures, and metabolic regulation must play an important part in this delicate process. Whereas many early mammalian metabolic cues are maternal, no such regulation is possible in oviparous animals such as frogs, where embryonic development takes place in the dichotomous environment of the egg. In Xenopus laevis, early embryo provisions are stored in the form of yolk proteins in yolk platelets. Yolk is composed principally of phospho-lipo-glycoproteins derived from a larger maternal precursor, vitellogenin (Wallace and Selman,1985). Interestingly, while yolk stores are found in the eggs of many animals, yolk is not degraded or utilized until late embryogenesis (Armant et al.,1986; Medina,1989; Fagotto,1990). This late trigger of yolk utilization implies that metabolic regulatory mechanisms exist and are triggered at appropriate times during development. Past metabolic studies have focused primarily on biochemical and physiological aspects of fat and yolk utilization. However, research has begun to examine lipid metabolism at the sub-cellular and molecular level. Recently, lipid bodies have been shown to directly interact with peroxisomes at the sub-cellular level, forming structures termed gnarls (Binns et al.,2006). Here we investigate the relationship between early metabolic events and the expression of genes associated with peroxisome biogenesis.

Peroxisomes are organelles responsible for very long chain fatty acid (VLCFA) metabolism through the β-oxidation pathway (Lazarow and Fujiki,1985). As fatty acids are metabolized, the corrosive by-product H2O2 is produced. However, it is converted to water and oxygen by catalase (Deisseroth and Dounce,1970). Though numerous genes and proteins are associated with peroxisomal function, here we focus on six (PEX5, PEX16, PEX19, ACO, PMP70, and catalase) that together are representative of peroxisome biogenesis, import, and enzymatic function. Biogenesis and the associated import and compilation of peroxisomal proteins into the organelle are largely accomplished by proteins termed peroxins (PEX proteins) (Distel et al.,1996). PEX5 is a cytosolic receptor for proteins containing a peroxisomal targeting signal (PTS1), such as catalase and acetyl-CoA oxidase (ACO) (Fournier et al.,1994). PEX19 is another cytosolic receptor responsible for the targeting of peroxisomal membrane proteins (PMPs), such as the 70-kDa protein PMP70 and the x-adrenoleukodystrophy-associated protein, ALD, to the peroxisomal membrane (Sacksteder et al.,2000). An additional PEX protein that has received much attention recently is PEX16, which is believed to be the key player in peroxisomal biogenesis at the endoplasmic reticulum (Kim et al.,2006; Schluter et al.,2006). Aside from metabolism, peroxisomes are also known for their impact on cell senescence (Terlecky et al.,2006), disease (Gould and Valle,2000), and to a much lesser extent, development (Baumgart et al.,2003).

While relatively little is known about the distribution of peroxisomes in early development, data are available regarding peroxisome proliferator-activated receptors (PPARs), which belong to the nuclear hormone receptor family (Issemann and Green,1990; Dreyer et al.,1992; Michalik et al.,2002). The three members of the PPAR family are; α, β, and γ, which act as key transcription factors to regulate aspects of lipid metabolism by acting on genomic sequences termed peroxisome proliferator response elements (PPRE) (Tugwood et al.,1992). The precise function of PPARβ remains elusive. However, PPARα has been shown to induce fatty acid metabolism, and PPARγ has been associated with adipogenesis in mouse (Lemberger et al.,1996). PPAR α, β, and γ have been cloned in X. laevis and their mRNAs are expressed differentially, such that PPARα and β are expressed in oocytes and embryos at several stages, whereas PPARγ is not (Dreyer et al.,1992). PPARα has been shown to directly upregulate enzymes necessary for peroxisomal β-oxidation, such as ACO (Tugwood et al.,1992), which catalyzes the first step of the peroxisomal β-oxidation of fatty acids (Osumi and Hashimoto,1978). As the regulation of PPARs is complex, involving many potential ligands (Wahli et al.,1995; Ito et al.,2006), and as this is compounded by the fact that PPARs act upon numerous metabolism genes, this illustrates that fatty acid metabolism in development is a complicated multifaceted process.

Recent theories suggest peroxisomes could act as intracellular signaling compartments and organizing platforms that orchestrate developmental decisions from inside the cell (Masters,1996; Titorenko and Rachubinski,2004). In development, mammalian studies have been restricted to immunohistochemistry and mRNA localization studies in tissues of mouse and rat (Pollard et al.,1995; Espeel et al.,1997; Grabenbauer et al.,2001). As mammalian studies use relatively late developmental stages, peroxisomes were detected in all stages and tissues examined. However, Xenopus embryological approaches afford us the possibility to observe peroxisome biogenesis in much earlier stages.

In this study, the temporal expression of X. laevis PEX5, 16, 19, PMP70, catalase, and ACO was examined using RT-PCR. Similar patterns of expression were seen for ACO and its cytosolic receptor PEX5, as well as PMP70 and its cytosolic receptor PEX19, while the expression of catalase and PEX16 increased gradually over development. The cellular localization of catalase and PMP70 protein was examined with immunohistochemistry and showed punctate staining in late stages, indicative of peroxisome formation primarily in dorsal-anterior structures. This localization was further supported with microinjected GFP constructs targeted to peroxisomes. The GFP tag consisted of the amino acids KANL (GFP-KANL), which is the C-terminal sequence within catalase for import into the peroxisome (Purdue and Lazarow,1996). The selected developmental stages were chosen to determine peroxisomal gene expression, protein localization, and functional import during early differentiation, before the organism shifts from yolk metabolism to the metabolism of ingested food following stage 45. Together, the expression, localization, and import studies shown here give an indication for a requirement of peroxisomal function in non-endodermal tissues of the developing embryo prior to ingestion.


RT-PCR Analysis of Peroxisomal Gene RNA Levels During Embryogenesis

The temporal analysis of RNA levels of all six genes was examined during early development (Fig. 1). Developmental stages examined included embryos completing gastrulation (stage 12), and continuing to develop through neurulation and neural tube closure (stages 20, 25), as well as axis elongation and early organogenesis (stage 30). The final stage examined was as free swimming embryos prepare to feed (stage 45). All six genes displayed changes in RNA levels, with three general patterns identified. ACO and PEX5 levels began relatively low and displayed a peak early in development. There was a peak of ACO RNA at stage 20, while there was a peak of PEX5 RNA at stage 25. The second pattern resulted from PEX19 and PMP70, where early expression at stage 12 was relatively high, and then fell by stage 20, only to gradually increase to higher levels. The third pattern was exhibited by PEX16 and catalase, where RNA levels remained more or less constant until stage 30 after which time the RNA levels of these genes rise substantially. This RT-PCR data was complemented by whole mount in situ hybridization data for each gene examined, which showed the presence of transcripts primarily in dorsal and anterior structures, displaying constitutive expression following stage 12 (data not shown).

Figure 1.

Expression analysis of peroxisomal gene RNA levels during developmental stages of X. laevis. The results were obtained by quantifying band intensities of RT-PCR products for peroxisomal RNA as compared to EF1α on a 0.8% agarose gel. PEX5 expression is at its highest expression is a peak at stage 25, while other stages are low in relative expression (A). The expression of ACO exhibits a similar pattern to PEX5, however its early peak occurs at stage 20, and there is a steady increase at after the peak from stage 25 to 45 (B). PEX19 expression is highest as stage 45, however, stage 12 shows the second highest expression (C). PMP70 follows the same expression as PEX19 (D). The expression of PEX16 remains relatively low compared to its peak at Stage 45 (E). The expression of catalase is essentially the same as PEX16, having relatively low expression until stage 45 (F). Error bars are representative of three replications.

Immunohistochemistry Localizes PMP70 and Catalase Proteins to Punctate Structures in Stage-45 Embryos

Immunohistochemistry was used to examine the distribution of PMP70 and catalase protein and more precisely allow for the identification of peroxisomal structures. Specific cross-reactivity of antibodies to X. laevis proteins was first determined using Western blotting. Bands of expected size were identified for both PMP70 (70 kDa) and catalase (55 kDa) (Fig. 2A). Furthermore, these antibodies were used to localize these proteins in X. laevis A6 cells. As expected these antibodies were localized to distinct punctate organelles, characteristic of peroxisomes, within the cells (Fig. 2B,C). Having determined that these antibodies cross-reacted appropriately in both Western analysis and cell culture, the distribution of PMP70 and catalase proteins in histological sections of embryos was subsequently examined. No specific peroxisomal staining or punctate structures were seen prior to stage 45. Feeding begins shortly following this stage and as metabolism shifts from digestion of embryonic food stores to include ingested food, peroxisome regulation was not examined beyond stage 45 as it would not be solely regulated based on endogenous metabolic signals. At stage 45, PMP70 protein was localized to punctate structures, whereas no punctate structures were seen at stage 30. At stage 45, punctate structures were seen in ectodermal and sub-ectodermal cells of the posterior (Fig. 3A,B) and anterior head (Fig. 3C,D), and ventral ectodermal cells (Fig. 3E,F). This staining pattern was also seen when using a catalase antibody in the ectodermal and sub-ectodermal cells of the posterior (Fig. 4A,B) and anterior head (Fig. 4C,D), and ventral ectodermal cells (Fig. 4E,F). Neither PMP70 nor catalase antibody was able to detect specific signals within endodermal yolk cells.

Figure 2.

Cross reactivity of mammalian peroxisomal antibodies with X. laevis protein. Western blot analysis for cross reactivity between X. laevis protein and antibodies for PMP70, Catalase. Each lane represents a independent protein sample. GAPDH was used as a control (A). Immunohistochemistry on X. laevis A6 cells using primary antibodies for PMP70 (B), catalase (C), and no primary antibody control (D). DAPI (blue), peroxisomal antibodies (red). Scale bar = 10μm.

Figure 3.

PMP70 immunohistochemistry on sectioned X. laevis embryos. The stages present include 30 (A,C,E) and 45 (B,D,F). Tissue samples are of ectodermal and sub-ectodermal cells in the posterior head (A–B), ectodermal and sub-ectodermal cells in the anterior head (C–D), ventral ectodermal cells (E–F). DAPI (blue), actin filaments (green), peroxisomal signal (red). Scale bar = 10 μm.

Figure 4.

Catalase immunohistochemistry on sectioned X. laevis embryos. The stages present include 30 (A,C,E) and 45 (B,D,F). Tissue samples are of ectodermal and sub-ectodermal cells in the posterior head (A–B), ectodermal and sub-ectodermal cells in the anterior head (C–D), ventral ectodermal cells (E–F). DAPI (blue), actin filaments (green), peroxisomal signal (red). Scale bar = 10 μm.

Peroxisomal Targeted GFP Is Localized to Punctate Structures in Stage-45 Embryos

The data obtained from our immunohistological studies were confirmed through the microinjection of KANL targeted GFP into embryos. If peroxisomes were indeed present in the tissues identified by PMP70 and catalase antibodies, then microinjected targeted-GFP should also be localized to those tissues, showing not only peroxisome localization, but functional import. In support of our immunohistological data, KANL targeted GFP was localized to distinct punctate organelles within the anterior head, posterior head, and ventral ectoderm in stage-45 embryos (Fig. 5A,C,E). As with the immunohistochemical data, no specific punctate structures were seen prior to stage 45, nor were they seen in the endoderm. Control injections of untargeted GFP resulted in non-specific, cytosolic distribution of GFP protein within cells of the posterior head, anterior head, and ventral ectoderm (Fig. 5B,D,F). Regions of X. laevis where high magnification images were taken for immunohistochemistry and targeted GFP expression and highlighted on low-magnification embryo images (Fig. 6).

Figure 5.

Microinjections of tagged and control GFP. Sections of X. laevis embryos microinjected with either GFP-KANL (A,C,E) or control GFP (B,D,FO. Tissue samples are of sub-ectodermal cells in the posterior head (A–B),k sub-ectodermal cells in the anterior head (C–D), and ventral ectodermal cells (E–F). DAPI (blue), actin filaments (red), GFP (green). Scale bar = 10 μm.

Figure 6.

Stage 30 and 45 X. laevis embryos depicting location of histological sections used in Figures 3, 4 and 5. A lateral view of stage 30 (i) and stage 45 (ii) embryos demonstrating the approximate location and plane of histological section. Areas are cells of the ectoderm and sub-ectoderm in the posterior head (A), cells of the ectoderm and sub-ectoderm in the anterior head (B), and cells of the ventral ectoderm (C). The approximate size of the embryos are approximately 5mm and 10mm for stage 30 and 45, respectively.


Peroxisomal Gene RNA Levels Are Dynamic Throughout Development

The RNA levels of all six genes examined changed temporally during early development in three distinct patterns. The first pattern is exhibited by ACO and PEX5, where ACO RNA levels are the first to rise, at stage 20. This is expected as PPARα, which has been shown to directly upregulate ACO, accumulates just prior to stage 20 (Dreyer et al.,1992). After ACO peak RNA levels are seen at stage 20, a rise in the levels of PEX5 RNA occurs at stage 25. This is surprising as it has been well documented that PEX5 is the cytosolic receptor for ACO and should be present prior (Purdue and Lazarow,1996; Stanley and Wilmanns,2006). However, it is important to note that the balance of transcription factors responsible for peroxisome biogenesis is complex, and not necessarily chronological. The temporal expression patterns shown here suggest that PEX5 is not directly regulated by PPARα, but perhaps by a factor further down the PPARα pathway. It is also important to note that the time difference between stage 20 and 25 is approximately 5.5 hr, and this difference would likely not impact the timeframe for creating active peroxisomes. Indeed, the relative levels of transcript at these early stages give an indication for gene regulation, and the relative pathways governing these genes and may not yet reflect peroxisomal function.

Interestingly, both PEX19 and PMP70 share similar RNA level patterns suggesting that their regulation is correlated or perhaps controlled by the same mechanism. Previously, regulation of these genes has only been associated with oleate and fenofibrate, but a molecular mechanism for their regulation is unknown. PEX19 was shown to increase fivefold after the addition of oleate in yeast cell culture (Gotte et al.,1998) while PMP70 was upregulated by fenofibrate in mouse liver (Berger et al.,1999). The early presence of relatively high levels of RNA of PEX19 and PMP70 shows that their regulation may occur separately from the other peroxisomal genes in this study.

Lastly, PEX16 and catalase RNA levels rise gradually over development until their peak at stage 45 when we show the appearance of peroxisomes. There has been little prior work on PEX16 gene regulation, while catalase gene regulation has been extensively characterized with a multitude of known proteins and substrates that either increase or decrease its expression levels (Nenoi et al.,2001; Park et al.,2004; Zhu et al.,2005). Here we describe the expression of these genes during X. laevis development. The requirement for peroxisomal function is expected to rise at stage 30 as metabolic demand increases to culminate near stage 45 with the formation of functional peroxisomes. Stage 45 is the approximate time when tadpoles shift from regulating the metabolism of endogenous stored yolk, to regulating the metabolism associated with ingested food. We demonstrate in this study an increase in transcription for all peroxisomal genes examined prior to feeding. In addition, correlating with our RT-PCR data, our whole mount in situ data, although not quantitative, show the presence of these transcripts in all stages examined (data not shown).

Immunohistochemistry Localizes Peroxisomal Protein in Dorsal-Anterior Structures

Apart from RNA expression patterns, immunohistochemistry gives both temporal and spatial data for the presence of proteins and, therefore, a better indication of functional peroxisomes. At late X. laevis stages, staining is localized to punctate structures suggestive of peroxisomes, predominately in developing brain and other dorsal-anterior structures. The presence of peroxisomes in brain tissue can be correlated to the high levels of plasmalogens that are synthesized in peroxisomes and are required for central white matter formation (Gould et al.,2001). Unexpectedly, an immunohistochemical signal was not detected in late endoderm, where yolk utilization in this region has been biochemically demonstrated for late development (Fagotto and Maxfield,1994). The lack of signal here is perhaps due to antibody permeability restrictions in the lipo-protein dense endoderm; however, the lack of signal in endoderm correlates with targeted GFP microinjection studies.

Targeted GFP Microinjections Correlate With Immunohistochemistry

Confirming that the punctate structures seen in late embryonic stages were active peroxisomes required a functional approach. Although punctate protein structures suggest the presence of peroxisomes, the import of proteins indicates that they are functional. Detection of functional peroxisomes is important as studies have found structures resembling peroxisomes, which lack any functionality (Yamasaki et al.,1999). An active peroxisome would import proteins containing a KANL signal, indicating that several proteins, including PEX5, 7, 13, and 14, are present and functional (Girzalsky et al.,1999; Dansen et al.,2001; Mukai and Fujiki,2006). The localization of targeted GFP would be expected in the same tissues seen with immunohistochemistry, with perhaps an increase in cytosolic staining due to the import properties shown previously with KANL targeting (Wood et al.,2006). Targeted GFP was seen in punctate structures that coincided with our immunohistochemical data. Signals were again most prevalent in anterior and ventral ectodermal tissues and not readily detectable in deeper endodermal tissues. Indeed, our data indicate that peroxisomes are present predominantly in ectodermal sub-ectodermal tissues starting at stage 45, perhaps where β-oxidation is occurring, as opposed to in the endoderm, where the preliminary biochemical breakdown of yolk is taking place.


Our data indicate that peroxisomes are assembled in specific tissues and stages during X. laevis development. While functional peroxisomes were not found in the earliest stages examined, the expression of the peroxisomal genes examined was dynamic, indicating developmental cues act to coordinate peroxisomal biogenesis in development. Furthermore, the presence of functional peroxisomes in given ectodermal tissues suggests an important function for these tissues in metabolism that is independent of the yolk storage capabilities of the endoderm. As well, while peroxisomes are implicated in cell signaling, they are not likely aiding in the differentiation of tissues at stage 30 when organogenesis is at its peak, as functional peroxisomes were not detected at these stages using immunohistochemistry or targeted GFP injections. Indeed, it appears as though no functional peroxisomes are needed to regulate metabolic or signaling pathways prior to feeding, and their role in metabolism may be critical as the embryo shifts from utilizing stored food reserves to ingested food. While these organelles may not play early developmental roles, an understanding of their regulation and biogenesis is still key as their formation late in development is likely crucial as the tadpole shifts from one food source to another.


Animal Care

Adult X. laevis were reared in accordance with Canadian Council on Animal Care. Fertilizations were performed according to Wu and Gerhart (1991), and embryos were staged according to Nieuwkoop and Faber (1956). Embryos to be sectioned were fixed in 4% formaldehyde at desired stages, paraffin-embedded, and sectioned to 7μm along a longitudinal plane.

Cell Culture

A6 cells derived from X. laevis epithelial cells (gift from Dr. John Heikkila) were grown in Leibowitz-15 media at 22°C until fixation in 4% formaldehyde for staining.

RNA Isolations and Reverse Transcriptase PCR (RT-PCR)

Total RNA was isolated with an RNeasy kit (Qiagen) from stages 12, 20, 25, 30, 45, and evaluated on a 1.0% agarose formaldehyde gel. Synthesis of cDNA was performed with SSII reverse transcriptase (Invitrogen). To analyze RNA expression levels during development, RT-PCR primers were designed against known Xenopus peroxisomal genes with the following NCBI accession numbers: PEX5 (NP_001011381), ACO (AAH63727), PEX19 (AAH87526), PMP70 (EF070607), PEX16 (AAH72248), and catalase (BC054964). The gene-specific primers used for RT-PCR were as listed (Table 1). The EF1α control primers flank an intron in genomic DNA; as a result, a larger band would be present in the event of genomic contamination. Aliquots were taken at different cycles throughout the PCR to ensure that bands were increasing in intensity and thus in the log phase at cycle 35. RT-PCR products at cycle 35 were then visualized on a 0.8% TAE agarose gel and unsaturated band intensities were quantified against EF1α (gene of interest value/EF1α) with Quantity One software (Version 4.4.0 Bio-Rad). All quantified PCRs were repeated three times to generate error bars. Note that the scale of the Y-axis is not standardized between genes. This is to highlight the pattern of expression, and, therefore, genes are not quantified against each other.

Table 1. Primers Used for PCR

Cloning, mRNA Synthesis, and Microinjection

The amplicons of peroxisomal genes listed above were obtained through RT-PCR and cloned with the TOPO-TA cloning system as described by the manufacturer's protocol (Invitrogen) and sequenced to ensure gene identity (Robart's Research Institute, London, ON). Full-length eGFP (Clontech, CA) and GFP-KANL were amplified by RT-PCR, then ligated into the T7TS plasmid such that the sequence was flanked by 5′ and 3′ UTRs of the X. laevis β-globin gene (gift from Dr. PA Krieg). Capped polyadenylated RNA was synthesized using mMachine (Ambion) and visualized on a 1.0% agarose formaldehyde gel to ensure quality and transcription validity. Embryos in 4% ficoll in MMR at the one-cell stage were microinjected with approximately 2 ng of GFP or GFP-KANL mRNA per embryo.

Western Blotting

PMP70 (Cedarlane) and catalase (Calbiochem) polyclonal antibodies were tested for cross-reactivity with X. laevis protein by Western blotting. Protein was isolated from X. laevis at stage 45 and quantified using the Bradford assay, to ensure that 10 mg of protein was loaded into each lane (Bradford,1976). Samples were run in duplicate. Antibodies were used at a dilution of 1 in 5,000, and blots were developed using enhanced chemiluminescence(Amersham).

Immunohistochemistry and Fluorescence Microscopy

Paraffin-embedded embryo sections were re-hydrated by washing in Xylene, 100, 90, and 70% ethanol each for 2 min twice, followed by 5 min in 0.1% Tween-20 in PBS twice. Embryos and A6 cells were stained with DAPI (Invitrogen), FITC-phalloidin (Sigma), and rhodamine-phalloidin (Sigma) according to the manufacturer's protocol. A phalloidin stain was used on histological sections in order to visualize cellular and tissue organization more accurately. Embryo histological sections and A6 cells to be immunostained were incubated as previously described, with the exception of a 24-hr primary and secondary (Texas red conjugated) antibody incubation and a 1 in 100 antibody dilution (Legakis et al.,2002). All embryos were visualized with a Zeiss AxioStop 2 Mot. Images were captured with a Retiga 1600 camera (Qimaging) using Northern Eclipse image capture and analysis software (Empix).


We thank Dr. Paul A. Walton for constructive discussions and input and Rick Harris for his aid with microscopy. This research was supported by a Discovery Grant from Canada's Natural Sciences and Engineering Research Council (NSERC) to S.D. C.A.C. and L.A.W. acknowledge Ontario Graduate Scholarships for funding their research.