Reconstitution of human peroxisomal β-oxidation in yeast

Abstract

We report the permanent introduction of the human peroxisomal β-oxidation enzymatic machinery required for straight chain degradation of fatty acids into the yeast, Saccharomyces cerevisiae. Peroxisomal β-oxidation encompasses four sequential reactions that are confined to three enzymes. The genes encoding human acyl-CoA oxidase 1, peroxisomal multifunctional enzyme type 2 and 3-ketoacyl-CoA thiolase were introduced into the genomic loci of their yeast gene equivalents. The human β-oxidation genes were individually tagged with sequence coding for GFP and expression of the protein chimeras as well as their targeting to peroxisomes was confirmed. Functional complementation of the β-oxidation pathway was assessed by growth on media containing fatty acids of different chain lengths. Yeast cells exhibited distinctive substrate specificities depending on whether they expressed the human or their endogenous β-oxidation machinery. The genetic engineering of yeast to contain a ‘humanized’ organelle is a first step for the in vivo study of human peroxisome disorders in a model organism.

We have constructed yeast strains containing peroxisomes that exhibit a substrate specificity for fatty acids resembling that of human, and not yeast, peroxisomes, which constitutes a first step to making fully ‘humanized’ peroxisomes in yeast as a tool for investigating human peroxisome disease.

INTRODUCTION

Peroxisomes are ubiquitous organelles involved in a variety of cellular processes. Distinctive protein machineries direct the import of peroxisomal matrix and membrane proteins, whereas peroxisomal enzymes catalyze the various metabolic reactions carried out by peroxisomes (Platta and Erdmann 2007). Peroxisomes are essential for human development and survival as exemplified by the existence of lethal congenital peroxisome biogenesis disorders (Steinberg et al.2006). A hallmark of peroxisome disorders is the accumulation of very-long chain fatty acids (VLCFAs), as these types of fatty acids are typically broken down in peroxisomes. VLCFAs increase the micro-viscosity of membranes and may lead to demyelination, inflammatory responses and migrational defects of neurons in the central nervous system (Wanders et al.2017).

Peroxisomal β-oxidation is an enzymatic pathway wherein straight-chain fatty acids are incrementally shortened by two carbons until they have been converted into acetyl-CoA (Poirier et al.2006). The pathway consists of four sequential reactions that are performed by three enzymes. Acyl-CoA oxidase catalyzes the desaturation of acyl-CoAs to 2-trans-enoyl-CoAs, multifunctional enzyme catalyzes the second (hydratase) and third (dehydrogenase) enzymatic steps, and 3-ketoacyl-CoA thiolase is responsible for the final thiolytic cleavage of the fatty acid.

The β-oxidation pathway has been conserved in peroxisomes of virtually all organisms (Waterham, Ferdinandusse and Wanders 2016), although a number of differences exist between yeast and human (Fig. 1). The mammalian machinery for the peroxisomal breakdown of fatty acids is more elaborate, with a large number of transporters for the uptake of fatty acids into the peroxisome lumen and the existence of a pathway for the degradation of branched chain fatty acids. This secondary pathway, which is referred to as α-oxidation, is absent in yeast. Moreover, β-oxidation occurs both in peroxisomes and mitochondria in mammals, whereas it is confined to peroxisomes in yeast. As a result, yeast peroxisomes are nondiscriminatory with regards to the chain length of acyl-CoAs broken down by β-oxidation. In contrast, mammalian peroxisomes preferentially carry out the breakdown of VLCFAs. Polyunsaturated VLCFAs are metabolized efficiently in peroxisomes but not in mitochondria, and some of these fatty acids can even inhibit mitochondrial β-oxidation in mammals (Osmundsen and Bjornstad 1985; Hiltunen et al.1986). Peroxisomal β-oxidation products can be fully oxidized to CO₂ and H₂O only after they have been shuttled to mitochondria in mammals (Fransen, Lismont and Walton 2017).

Figure 1.

Human and yeast peroxisomal β-oxidation pathways. The human peroxisomal β-oxidation pathway for the degradation of straight-chain fatty acids is encoded by the ACOX1 (encoding acyl-CoA oxidase), MFE2 (encoding multifunctional enzyme) and ACAA1 (encoding 3-ketoacyl-CoA thiolase) genes (pink boxes, black font). These genes were transplanted into the yeast system at their corresponding yeast gene loci POX1, FOX2 and POT1, respectively. Human peroxisomes contain additional pathways for α-oxidation of branched chain fatty acids, plasmalogen synthesis, bile acid synthesis and glyoxylate detoxification. Additional pathways that were not investigated in this study are presented in semitransparent color. Modified from Waterham, Ferdinandusse and Wanders (2016).

Yeast peroxisome biogenesis mutants are conditionally viable as long as cells are not grown on medium containing fatty acid as the main carbon source. This feature, along with its ease of genetic manipulation, has made yeast an organism of choice for the identification and characterization of genes involved in peroxisome function and biogenesis (Smith and Aitchison 2013). To expand further the utility of yeast as a model organism for the study of peroxisomal biology, we permanently expressed genes for human peroxisomal proteins in yeast to create ‘humanized’ peroxisomes. Here we describe the reconstitution of human peroxisomal β-oxidation in yeast, which was accomplished by replacing the endogenous yeast peroxisomal β-oxidation machinery with the human peroxisomal β-oxidation machinery.

MATERIALS AND METHODS

Yeast strains and cell culture conditions

Saccharomyces cerevisiae strains used in this study are listed in Table 1. Unless otherwise noted, strains were grown in YPD medium at 30°C. For the growth assay presented in Fig. 3, cells were grown in YPD medium to an OD₆₀₀ of ∼0.5, harvested by centrifugation, washed in H₂O and adjusted to an OD₆₀₀ of 1.0. 1 μL of a 1:10 dilution series was spotted onto plates containing glucose (YPD), behenic acid (YPBB), oleic acid (YPBO), or myristic acid (YPBM).

Table 1.

S. cerevisiae strains used in this study.

Strain	Genotype	Reference
BY4742	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0	Giaever et al.2002
BY4742, mCherry-PTS1	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, can1::mCherry-PTS1	This work
BY4742, pex3Δ, mCherry-PTS1	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, pex3::LEU2, can1::mCherry-PTS1	This work
BY4742, GFP-POX1, mCherry-PTS1	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, pox1::GFP-POX1, can1::mCherry-PTS1	This work
BY4742, GFP-FOX2, mCherry-PTS1	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, fox2::GFP-FOX2, can1::mCherry-PTS1	This work
BY4742, POT1-GFP, mCherry-PTS1	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, pot1::POT1-GFP, can1::mCherry-PTS1	This work
BY4742, ACOX1, mCherry-PTS1	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, pox1::ACOX1, can1::mCherry-PTS1	This work
BY4742, MFE2, mCherry-PTS1	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, fox2::MFE2, can1::mCherry-PTS1	This work
BY4742, ACAA1, mCherry-PTS1	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, pot1::ACAA1, can1::mCherry-PTS1	This work
BY4742, GFP-ACOX1, mCherry-PTS1	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, pox1::GFP-ACOX1, can1::mCherry-PTS1	This work
BY4742, GFP-MFE2, mCherry-PTS1	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, fox2::GFP-MFE2, can1::mCherry-PTS1	This work
BY4742, GFP-ACAA1, mCherry-PTS1	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, pot1::ACAA1-GFP, can1::mCherry-PTS1	This work
BY4742, pox1Δ, mCherry-PTS1	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, pox1::URA3, can1::mCherry-PTS1	This work
BY4742, fox2Δ, mCherry-PTS1	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, fox2::URA3, can1::mCherry-PTS1	This work
BY4742, pot1Δ, mCherry-PTS1	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, pot1::URA3, can1::mCherry-PTS1	This work
BY4742, ACOX1, MFE2, ACAA1, mCherry-PTS1	MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, pox1::ACOX1, fox2::MFE2, pot1::ACAA1, can1::mCherry-PTS1	This work

Figure 3.

Metabolic activity of ‘humanized’ yeast peroxisomes. (A) Yeast cells expressing either their endogenous peroxisomal β-oxidation machinery (Scβ-oxidation, top rows), containing a deletion of the PEX3 gene (middle rows), or expressing the human peroxisomal β-oxidation machinery (Hsβ-oxidation, bottom rows) were grown on solid medium that contained glucose, behenic acid, oleic acid or myristic acid as the main carbon source. Note that residual growth of the pex3Δ strain on fatty acid-containing media was due to the presence of small amounts of other carbon sources in the media. Presented are series of 1:10 dilutions of cells of cells that had been adjusted to an OD₆₀₀ of one. (B) Yeast cells expressing combinations of human and yeast peroxisomal β-oxidation enzymes were grown on solid medium containing myristic acid as the main carbon source. Reconstitution of a functional β-oxidation pathway and growth of cells on myristic acid solid medium was observed only in cells expressing ScACOX, ScMFE and HsTHI. Presented are series of 1:10 dilutions as in (A).

Media components were: YPD, 1% yeast extract, 2% peptone, 2% glucose; YPBB, 0.67% yeast nitrogen base (YNB) complete, 0.1% yeast extract, 6.61 mM KH₂PO₄, 1.32 mM K₂HPO₄, 0.5% Tween 40, 0.15% (w/v) behenic acid; YPBO, 0.67% yeast nitrogen base (YNB) complete, 0.1% yeast extract, 6.61 mM KH₂PO₄, 1.32 mM K₂HPO₄, 0.5% Tween 40, 0.15% (v/v) oleic acid; YPBM, 0.67% yeast nitrogen base (YNB) complete, 0.1% yeast extract, 6.61 mM KH₂PO₄, 1.32 mM K₂HPO₄, 0.5% Tween 40, 0.15% (w/v) myristic acid; SCIM, 0.67% YNB without amino acids, 0.5% yeast extract, 0.5% peptone, 3.3% Brij 35, 0.3% glucose, 0.3% oleic acid, 1 × complete supplement mixture (CSM); 5-FOA, 0.67% YNB without amino acids, 1 × CSM-uracil, 0.1% 5-fluoroorotic acid, 0.0012% uracil, 2% glucose; SM-arg + can, 0.67% YNB without amino acids, 1 × CSM-arginine, 0.006% canavanine, 2% glucose, 2% agar. For solid media, 2% agar was added.

Genetic manipulation of yeast

The delitto perfetto methodology (Storici and Resnick 2006) was used to introduce permanently genes for human peroxisomal β-oxidation enzymes into their equivalent gene loci in yeast. In a first transformation, the endogenous open reading frame (ORF) of the yeast gene was replaced by a counter selectable cassette. In a second transformation, this cassette was replaced by the gene for the corresponding human peroxisomal β-oxidation enzyme. All human gene sequences were codon-optimized for yeast expression (Figs S1 to S3, Supporting Information). Transformants were counter selected by growth on 5-FOA medium. All integrations were confirmed by sequencing the genomic locus. Integrations were done sequentially to construct strains that carry genes for multiple human peroxisomal β-oxidation enzymes.

To tag β-oxidation genes with sequence coding for GFP, a counter selectable cassette was introduced at either the 3΄-end (genes for ScTHI and HsTHI) or the 5΄-end (all other genes) of the ORF. In a second transformation, sequence for GFP was inserted in-frame with the ORF of the β-oxidation gene.

Strains expressing a genomically integrated peroxisomal reporter (mCherry-PTS1) were constructed by replacing the ORF of the CAN1 locus with sequence coding for mCherry (Shaner et al., 2004) fused at its C-terminus to the PTS1, Ser-Lys-Leu (SKL) and selection by growth on SM-arg + can plates (Hampsey 1997).

Confocal fluorescence microscopy

Live-cell imaging was performed with a LSM710 confocal fluorescence microscope (Carl Zeiss) equipped with a 63 × 1.4 NA Plan-Apo chromate objective and a piezoelectric stage to allow for rapid acquisition of z-stacks. Images were collected with a z-resolution of 0.2 μm to a total stack height of 5 μm. The GFP emission was collected within a range of 493–552 nm, whereas mCherry was excited with a 561-nm laser and its emission collected between 574 and 735 nm. Image processing was done as previously described (Knoblach et al.2013).

Cells used for live-cell imaging were diluted 1:100 from an overnight culture into fresh YPD medium and grown for 4–5 h. Cells were washed with water, and 2 μL of the cell suspension were spotted onto a thin agarose pad prepared from hot nonfluorescent medium, covered with a cover slip and sealed with VALAP (Fagarasanu et al.2009).

Antibodies

GFP-tagged β-oxidation enzymes were detected with affinity-purified anti-GFP rabbit antibodies (kind gift of Dr. Luc Berthiaume, University of Alberta). Rabbit antibodies to S. cerevisiae glucose-6-phosphate dehydrogenase (G6PDH) were from Sigma–Aldrich. Horseradish peroxidase-conjugated donkey antirabbit secondary antibodies were used to detect primary antibodies in immunoblot analysis. Antigen-antibody complexes in immunoblots were detected by enhanced chemiluminescence (GE Healthcare).

RESULTS AND DISCUSSION

To construct yeast strains that permanently express human β-oxidation genes, we inserted yeast codon-optimized sequence coding for the human β-oxidation enzymes acyl-CoA oxidase (here designated as HsACOX; encoded by the ACOX1 gene), multifunctional enzyme 2 (here designated as HsMFE; encoded by the MFE2 gene) and 3-ketoacyl-Co thiolase (here designated as HsTHI; encoded by the ACAA1 gene) into their respective yeast equivalent gene loci, i.e. POX1 encoding ScACOX, FOX2 encoding ScMFE, and POT1 encoding ScTHI. The human β-oxidation genes were thus under transcriptional control of their yeast gene equivalent promoters, which are responsive to nutritional cues and are upregulated in the presence of fatty acids (Schuller 2003).

Strains that had been genetically engineered to express human peroxisomal β-oxidation enzymes tagged with GFP were tested for the expression of the recombinant genes and the proper targeting of their gene products to peroxisomes. β-oxidation enzymes are peroxisomal matrix proteins that are imported into the lumen of the peroxisome via cytosolic receptors that recognize their cargo based on conserved targeting signals (Francisco et al.2017). Peroxisomal targeting signal 1 (PTS1) is a C-terminal tripeptide, whereas peroxisomal targeting signal 2 (PTS2) is an N-terminal nonapeptide. We generated a C-terminal fusion of GFP to ScTHI and to HsTHI, which are both PTS2-containing proteins. Yeast acyl-CoA oxidase (ScACOX) does not contain a PTS1 but is imported into the lumen of the peroxisome via the PTS1-receptor, Pex5p (Klein et al., 2002). ScACOX, ScMFE, HsACOX and HsMFE were tagged at their N-termini with GFP.

We first tested for expression of the heterologous genes by immunoblotting with an antibody to GFP for detection of the protein chimeras. All yeast β-oxidation enzymes, as well as their human counterparts, robustly expressed after growth of cells in medium containing fatty acid as the carbon source. The protein products migrated at their anticipated molecular mass in SDS-PAGE (Fig. 2A).

Figure 2.

Expression and localization of GFP-tagged yeast and human peroxisomal β-oxidation enzymes in yeast. (A) Total cell lysates from yeast strains expressing GFP-ScACOX, GFP-HsACOX, GFP-ScMFE, GFP-HsMFE, ScTHI-GFP or HsTHI-GFP were separated by SDS-PAGE. Immunoblotting was performed with antibodies directed against GFP (top) or against glucose-6-phosphate dehydrogenase (G6PDH) as a control for protein loading (bottom). The migrations of molecular weight markers are indicated at left. (B) Subcellular localization of yeast and human peroxisomal β-oxidation enzymes. Cells expressing GFP-tagged yeast and human peroxisomal β-oxidation enzymes and the peroxisomal reporter mCherry-PTS1 were imaged by confocal fluorescence microscopy. Maximum intensity projections of individual z-stacks are presented. Bar, 2 μm.

We next performed confocal fluorescence microscopy to visualize the subcellular localization of the β-oxidation enzymes in living cells. The GFP signal of the β-oxidation enzymes was not uniformly distributed, but localized to puncta (Fig. 2B). The puncta were also decorated by the peroxisomal reporter mCherry-PTS1, as is evident in the merged images of the green and red channels (Fig. 2B). While all other β-oxidation enzymes were strictly confined to peroxisomes, HsTHI was localized primarily to peroxisomes but also exhibited some localization to the cytosol, which suggests the presence of peroxisomal import signals other than the N-terminal PTS2 in HsTHI that are not recognized in yeast. Overall, we conclude that human β-oxidation enzymes properly targeted to yeast peroxisomes.

To study how the introduction of the human β-oxidation machinery into yeast affects the metabolic activity of the so-modified yeast peroxisomes, we conducted growth assays on glucose medium that is nonselective for peroxisome metabolic activity, as well as on fatty acid containing media that require peroxisome metabolic activity for yeast growth (Fig. 3A). We compared the growth of cells that express the yeast β-oxidation machinery with those of cells that express the human β-oxidation machinery. A pex3Δ strain, which fails to assemble peroxisomes (Höhfeld, Veenhuis and Kunau 1991), was included as a control. When cells were grown on nonselective glucose-containing medium, all three strains exhibited similar growth. This was expected, as yeast cells do not require peroxisomes for growth on fermentative carbon sources like glucose. Notable differences in yeast growth were observed when cells were grown on media containing different fatty acids as the main carbon source. Cells expressing either the yeast or the human β-oxidation machinery exhibited poor, but uniform growth on a C22 saturated fatty acid (behenic acid). Oleic acid is a mono-unsaturated C18 fatty acid. While strains expressing either the yeast or the human β-oxidation machinery were able to grow on oleic acid medium, the strain with the yeast peroxisomal β-oxidation machinery exhibited stronger growth on this carbon source than the strain with the human peroxisomal β-oxidation machinery. The most striking differences in growth between strains with the yeast peroxisomal β-oxidation machinery and the human peroxisomal β-oxidation machinery were observed when cells were grown on the C14 fatty acid, myristic acid. Yeast cells expressing their endogenous β-oxidation enzymes grew strongly on myristic acid medium, as evident by the size of the colonies and the depletion of the myristic acid around the colonies to form haloes. On the other hand, cells expressing the human β-oxidation machinery grew poorly on myristic acid medium. As expected, growth of the pex3Δ strain was compromised on all three fatty acid-containing media. We conclude that the yeast and human peroxisomal β-oxidation machineries have different specificities for different fatty acid substrates. While yeast cells with their endogenous β-oxidation enzymes can process fatty acids of different chain lengths, yeast cells with human β-oxidation enzymes grow preferentially on VLCFAs.

We further investigated the VLCFA substrate specificity of the human peroxisomal β-oxidation enzymes in yeast by conducting a growth assay on myristic acid medium with cells expressing one of the three human peroxisomal β-oxidation enzymes and the yeast forms of the two other peroxisomal β-oxidation enzymes (Fig. 3B). Cells deleted for POX1, FOX2 or POT1, i.e. were missing one of the three yeast β-oxidation enzymes, were unable to metabolize myristic acid (Fig. 3B). Cells expressing either HsACOX or HsMFE and the yeast forms of the remaining two peroxisomal β-oxidation enzymes were also unable to metabolize myristic acid. The first enzymatic step catalyzed by acyl-CoA oxidase has been shown to be the rate-limiting step in peroxisomal β-oxidation. Mammals contain three acyl-CoA oxidases that differ in their substrate specificities, whereas the yeast acyl-CoA oxidase Pox1p (ScACOX) metabolizes a broad range of substrates (Poirier et al.2006). Notably, myristic acid was consumed in cells where the endogenous thiolase gene POT1 was replaced by the gene for HsTHI, which liberates acyl-CoA in the final reaction of peroxisomal β-oxidation (Fig. 3B). Thus, HsTHI metabolizes fatty acids of different chain lengths in yeast comparably to ScTHI.

In summary, reconstitution of the human peroxisomal β-oxidation pathway in yeast was accomplished by swapping the endogenous yeast peroxisomal β-oxidation machinery with the human peroxisomal β-oxidation machinery. The so-modified yeast peroxisomes are metabolically active but display a specificity for fatty acid substrates that is more characteristic of human peroxisomes. This work is a first step towards making fully ‘humanized’ peroxisomes in yeast, which will provide a valuable tool for the analysis of human peroxisome biogenesis disorders in a simple unicellular model organism.

SUPPLEMENTARY DATA

Supplementary data are available at FEMSYR online.

FUNDING

This work was supported by Foundation Grant FDN-143289 from the Canadian Institutes of Health Research to RAR.

Conflict of interest. None declared.