The alkane-assimilating yeast Yarrowia lipolytica degrades very efficiently hydrophobic substrates such as n-alkanes, fatty acids, fats and oils for which it has specific metabolic pathways. An overview of the oxidative degradation pathways for alkanes and triglycerides in Y. lipolytica is given, with new insights arising from the recent genome sequencing of this yeast. This includes the interaction of hydrophobic substrates with yeast cells, their uptake and transport, the primary alkane oxidation to the corresponding fatty alcohols and then by different enzymes to fatty acids, and the subsequent degradation in peroxisomal β-oxidation or storage into lipid bodies. Several enzymes involved in hydrophobic substrate utilisation belong to multigene families, such as lipases/esterases (LIP genes), cytochromes P450 (ALK genes) and peroxisomal acyl-CoA oxidases (POX genes). Examples are presented demonstrating that wild-type and genetically engineered strains of Y. lipolytica can be used for alkane and fatty-acid bioconversion, such as aroma production, for production of SCP and SCO, for citric acid production, in bioremediation, in fine chemistry, for steroid biotransformation, and in food industry. These examples demonstrate distinct advantages of Y. lipolytica for their use in bioconversion reactions of biotechnologically interesting hydrophobic substrates.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Yarrowia lipolytica is one of the most extensively studied “non-conventional” yeasts which is currently used as a model for the study of protein secretion, peroxisome biogenesis, dimorphism, degradation of hydrophobic substrates (HS), and several new fields. Recently, the entire sequence of the six Y. lipolytica chromosomes has been determined , allowing its admission into the “omic” disciplines such as genomics, transcriptomics and proteomics. Several reviews have already been published on its physiology and genetics [2–4], secretion [5–7], dimorphism , peroxisome biogenesis [9,10] and mitochondrial complex I .
This review will focus on the physiology and genetics in relation to HS utilisation, although without intention to be comprehensive, and will give some new insights arising from the recent Y. lipolytica genome sequencing. It will also show some potential applications of this yeast linked to its specific capacity to efficiently utilise HS.
In the mid-1960s, interest in Y. lipolytica was due to its ability to utilise HS, especially alkanes, for the production of single-cell protein (SCP) as well as intermediate metabolites such as citric acid and 2-ketoglutaric acid. This resulted in a broad knowledge of its cultivation in large-scale fermenters. Y. lipolytica has been classified as Generally Regarded As Safe (GRAS) by the American Food and Drug Administration (FDA) for citric acid production.
In the early 1990s, the development of genetic tools and the observation that this yeast was able to highly express and secrete proteins such as alkaline extracellular protease (AEP), triggered analysis of protein secretion in the laboratories of Gaillardin (France), Ogrydziak (USA) and Pfizer Inc. (USA), joined later by the Dominguez group (Spain). These groups proposed that Y. lipolytica could be an alternative host for heterologous protein production. In the mid-1990s, due to its ability to grow on oleic acid, Rachubinsky's group (Canada) has used Y. lipolytica for the analysis of genes involved in peroxisome biogenesis, joined by the Nicaud group (France), and later by the groups of Barth and Sibirny (Germany and Ukraine, respectively) for the analysis of pexophagy. Nicaud (France) and Mauersberger (Germany) have performed genetic analysis of the HS degradation pathway while Belin and Nicaud (France) have genetically engineered Y. lipolytica for improved lactone production.
In the late-1990s, several groups also have initiated studies on dimorphism, since Y. lipolytica is able to grow as yeast or hyphal forms depending on growth conditions. The number of groups using Y. lipolytica as alternative yeast model is increasing rapidly with the development of molecular and genetic tools. Therefore, we will not give a comprehensive list of recent Y. lipolytica“aficionados”.
In the early-2000s, several international groups have joined efforts for the sequencing of the Y. lipolytica genome. This included chromosome analysis , gene mapping and the construction of a BAC library (Casaregola et al., unpublished). These efforts launched the partial genome sequencing of strain W29 . Recently, the complete genome sequence of the Y. lipolytica strain E150 (CLIB99) has been determined through the Génolevures Consortium, an active community of yeast researchers based in several French laboratories, assisted by an informal network of international partners [1,14]. More information can be found at the Génolevures web site (http://cbi.labri.fr/Genolevures/).
1.2Yarrowia lipolytica phylogeny and ecology
Y. lipolytica is an ascomycetous yeast which has been assigned to the family Dipodascaceae. Other current members of this family are Dipodascus, Galactomyces, Sporopachyderma, Stephanoascus, Wickerhamiella and Zygoascus. Except for Dipodascus and Galactomyces, the assignment of other genera to the Dipodascaceae, is uncertain . By analysis of the D1/D2 domain of the 26S-rDNA sequences of ascomycetous yeasts, Y. lipolytica was placed, together with Arxula and some Candida species, in the Metschnikowia/Stephanoascus clade . Although the phylogenetic position of Y. lipolytica is uncertain, it is clear that it is phylogenetically distant from other well-known and thoroughly studied alkane-degrading yeasts such as C. maltosa, C. albicans, C. tropicalis, Debaryomyces (Schwanniomyces) occidentalis, Debaryomyces hansenii and Pichia guilliermondii (Fig. 1). The latter species are all relatively closely related. The teleomorphic genera Debaryomyces and Pichia have been assigned to the Saccharomycetaceae. Not much is known about alkane degradation by species belonging to the Dipodascaceae, but some species of Metschnikowia, Stephanoascus and Arxula (i.e., M. pulcherrima, M. agaves, S. ciferrii, A. adeninivorans and A. terrestis) are known to degrade hexadecane [15,18].
Y. lipolytica, unlike C. albicans and C. tropicalis, is not considered as pathogenic . Most strains are unable to grow above 32 °C and the species is strictly aerobic. This yeast can be isolated from substrates rich in lipids and proteins  such as dairy products [2,21,22], cheese [23,24] or sausages . Y. lipolytica is one of the predominant species in Camembert and blue-veined cheeses and is frequently found at concentrations exceeding 106–107 cfug−1 (Section 3.7). This occurrence of Y. lipolytica correlates with its extracellular lipolytic and proteolytic activities, and also with its strong growth at 5–10 °C . Strains have also been isolated from soil, sewage and oil-polluted environments [2,26–29].
The knowledge on HS utilisation pathways in alkane-assimilating yeasts has been previously reviewed (see for example [17,30,31]) and partially for Y. lipolytica. Therefore, and due to limits of space, we will give here only a short overview and do not intend to be comprehensive.
The catabolism of HS, such as alkanes, fatty acids and triglycerides, is a quite complex metabolism which involves several metabolic pathways taking place in different subcellular compartments (Fig. 2). An important characteristic of alkane assimilation by yeasts is the metabolic flow of carbon from alkane substrates to synthesis of all cellular components via fatty acids, which is quite different from the case of conventional substrates like carbohydrates. Alkane assimilation by yeasts has been found to occur mainly via the monoterminal and diterminal oxidation pathways (Figs. 4 and 5). The main steps of HS degradation pathways in yeasts are: (i) The uptake and transport of HS to the site of primary oxidation, for which mechanisms are mostly unknown. Since these substrates are not miscible with water, their uptake requires morphological and physiological modifications, notably in cell adhesion properties (surface hydrophobicity) or in the production of emulsifiers (surfactants) (Section 2.1). Triglycerides are first hydrolysed by lipolytic enzymes into free fatty acids (FFA) (Section 2.2.1), which are then taken up by the cell (Section 2.2.2), whereas alkanes enter directly into the cell. (ii) The primary or monoterminal oxidation of alkanes in ER and peroxisomes to corresponding fatty acids of the same chain lengths, initiated in the ER by a cytochrome P450 (P450) catalysed terminal hydroxylation. Additionally diterminal or ω-oxidation leading to dicarboxylic acids (DCA) can occur (Section 2.3). (iii) The activation of FFA to their corresponding CoA esters which are subsequently degraded to acetyl-CoA and propionyl-CoA (in case of odd-chain alkanes) via peroxisomalβ-oxidation, or the direct incorporation of fatty-acyl moieties into cellular lipids after chain elongation and desaturation (Sections 2.3 and 2.4). Depending on environmental conditions, cells may accumulate FFA into lipid bodies (Section 2.5). (iv) The synthesis of tricarboxylic-acid (TCA) cycle intermediates from acetyl-CoA via the glyoxylate cycle followed by gluconeogenesis, and activation of the methyl citrate cycle for propionyl-CoA utilisation when using odd-chain alkanes (Section 2.4).
In order to get insight into these pathways, classical and reverse genetics have been used. The first attempts to identify genes involved in HS utilisation were initiated in the 1970–1980s in the groups of Mortimer, Numa, Barth and Mauersberger, by the isolation of mutants affected in alkane utilisation (for review see [2,4]). Later, mutants affected in oleic acid utilisation have been isolated by Rachubinsky's group focusing on peroxisome biogenesis mutants (for review see ), and alkane chain length-specific mutants have been studied by the Takagi/Ohta group [32–34]. Recently, new tagged mutants affected in HS utilisation have been isolated after either transposon insertion  or by a simple URA3-based mutagenesis cassette insertion [36,37], and subsequently new genes involved in these pathways were characterised (Nicaud and Mauersberger, unpublished results). Reverse genetics were also used for the characterisation of genes belonging to several gene families including P450 [32,55], lipases  and acyl-CoA oxidases .
2.1Contact between cells and substrate
To enter cells, HS must interact with the cell surface (Fig. 2). Two hypotheses have been formulated to explain this step of the transport of poorly water-miscible substrates into microorganisms. The compounds can be solubilised (or pseudo-solubilised) in the presence of surface-active compounds (surfactant-mediated transport) or they can adhere directly to the cell wall (direct interfacial transport, for review see [17,31,40], cf. Section 2.2.2).
In Y. lipolytica, evidence for both mechanisms has been described. This yeast species can produce surfactants during growth on HS [29,41] and a correlation has been observed between the induction of adhesion between cells and HS (Fig. 3), and the increase in apolar properties of the cell surface [27,42]. This modification of surface hydrophobicity appears to be linked to the presence of protrusions on the surface of cells growing on HS. These structures, which also have been observed in alkane-grown C. tropicalis, resemble channels that connect the cell wall to the interior of the cell and probably participate in HS uptake (cf. Section 2.2.2). Their presence depends on the growth phase  and seems to be regulated in an oleate-responsive manner, i.e., inducible by alkane and oleic acid and repressed by glucose (Albertyn and Smit, unpublished results). From a biotechnological point of view, these phenomena of adhesion, at least of lipids, appear to be less important than the interfacial surface between the substrate and the aqueous medium . However, the surface properties of the yeast may also have an impact on the emulsification of substrates .
2.2Uptake and transport of hydrophobic substrates
Y. lipolytica can utilise triglycerides as carbon source. The first step of their catabolism involves hydrolysis into free fatty acid and glycerol by lipolytic enzymes (lipases), first reported by Peters and Nelson. Later, Ota and collaborators have described an extracellular lipase activity and two cell-bound lipases of 39 and 44 kDa (for review see [2,47]). Y. lipolytica is able to produce several lipases (extracellular, membrane-bound, and intracellular activities), and lipase production depends on media composition and environmental conditions [48,49]. It is only recently that their corresponding genes have been identified. The LIP2 gene encodes the major extracellular lipase, Lip2p. It is synthesised as a 334-amino-acid (aa) preproprotein containing a signal sequence, a stretch of four dipeptides X-Ala, X-Pro (substrate of diaminopeptidase), followed by a 12-aa pro-region ending with a Lys-Arg dipeptide (substrate of the endoprotease encoded by XPR6) . This extracellular lipase Lip2p was previously reported to hydrolyse long-chain triglycerides, with preference for oleyl residues . LIP1 (Z50020), LIP3 (AJ249751) and LIP6, which code for carboxylesterases, were cloned and disrupted by Dominguez and collaborators (Dominguez et al., personal communication). More recently, two genes, LIP7 (AJ549519) and LIP8 (AJ549520) were shown to encode membrane-bound lipases, with Lip7p presenting a maximum activity for caproate (C6) while that of Lip8p was for caprate (C10) (Fickers et al., unpublished). A recent survey of the Y. lipolytica genome has revealed that there are four genes coding for carboxylesterase/lipase of type B, like LIP1, and sixteen genes presenting similarities with triacylglycerol lipases, like LIP2 (Nicaud et al., unpublished). The large number of genes will complicate their complete functional analysis.
2.2.2Alkane and fatty-acid transport
With the exception of triglycerides, utilisation of HS starts with the uptake of the substrates by yeast cells. Despite many efforts, the basic principles of this process have not yet been elucidated. Most findings support the suggestion that uptake of alkanes by yeast cells is a passive, diffusion-like process, which is facilitated by special hydrophobic properties and structures of the yeast cell (for reviews see [40,50]). In yeasts utilising HS as carbon source, several modifications in cell structure occurred, probably related to HS transport, e.g., protrusions at the cell surface, decreased thickness of cell wall and periplasmic space, membrane invaginations and electron-dense channels with associated ER structures beneath the plasma membrane, connecting the outside cell wall protrusions to the cell interior [27,43,44]. These observations have led to the hypothesis that alkanes attached to the protrusions or hydrophobic outgrowths may migrate through the channels via the plasma membrane to the ER, the site of alkane hydroxylation by P450 monooxygenase systems (, cf. Section 2.3). The exact mechanism by which hydrophobic compounds pass through a membrane is in all organisms still highly controversial.
Concerning fatty-acid uptake, investigations carried out by Kohlwein and Paltauf in S. uvarum and Saccharomycopsis (Y.) lipolytica concluded that below a threshold of 10 μM, an energy-free transporter was required, whereas above this concentration fatty acids like lauric or oleic acid diffused freely . Moreover, these authors proposed that at least two different chain-length-selective transporters are present. The above-mentioned structures resembling channels crossing the membrane (Section 2.1.) might play this role.
Once inside the cytoplasm, fatty acids might interact with fatty-acid-binding proteins (FABP) and Y. lipolytica possesses at least one palmitate-inducible FABP . Recent attempts to clone the gene coding for this FABP or lipid-binding protein in Y. lipolytica led to a cDNA coding for a putative sterol carrier protein of the SCP gene family (SCP2 gene, GenBank Accession No. AJ431362, Milikowski et al., unpublished). Nevertheless, this protein with fatty-acid-binding activity seems to be important for the efficient use of fatty acids by Y. lipolytica (Ermacora, personal communication).
Apart from the FABP which is absent from S. cerevisiae, several other genes (FAA1, FAA4, FAT1, ACB1, ABC transporter PXA1 and PXA2, PEX11) suggested to be implicated in fatty-acid uptake and intracellular transport in S. cerevisiae are probably present in Y. lipolytica, although, their function in HS transport processes remains to be established for Y. lipolytica.
2.3Primary alkane oxidation to fatty acids and dicarboxylic acids
The first step of both monoterminal or primary alkane oxidation and fatty-acid diterminal or ω-oxidation in yeast involves a terminal hydroxylation by a P450-dependent alkane monooxygenase system (AMOS, or alkane 1-hydroxylase) or fatty-acid ω-hydroxylase (FAH), respectively. This results in fatty-alcohol production from alkane, or ω-hydroxy fatty-acid production from fatty acid (Figs. 2, 4, and 5). The AMOS or FAH systems involve P450 (ALK genes belonging to CYP52 family) which are located in the ER together with an NADPH-dependent P450 reductase (CPR, NCP gene) for the electron transport. The second step is performed either by membrane-bound or soluble NAD+- or NADP+-dependent fatty-alcohol dehydrogenases (FADH, ADH genes) or by hydrogen peroxide-producing fatty-alcohol oxidases (FAOD; FAO genes), which convert the terminal hydroxy groups of 1-alkanols, 1,ω-diols, or ω-hydroxy fatty acids into corresponding fatty aldehydes (Figs. 4 and 5). The third step involves the oxidation of the fatty aldehyde to an FFA catalysed by ER or peroxisomal NAD(P)+-dependent fatty-aldehyde dehydrogenase (FALDH, ALD(H) genes, Figs. 4 and 5). These oxidation steps finally result in fatty-acid production from alkane, or dicarboxylic acid (DCA) production from fatty acid (Figs. 2, 4, and 5).
AMOS and FAH. P450 has been shown to be induced during growth of Y. lipolytica on alkanes and fatty acids  and to be involved in AMOS and FAH activities (Mauersberger, unpublished). Eight alkane- or fatty-acid-inducible P450 isoforms of the CYP52 gene family (ALK1 to ALK8, [32,55]) and one CPR gene (Iida and Takagi, personal communication) were identified in Y. lipolytica strain CX161-1B. Recent exploration of the genome of strain E150 revealed twelve ALK genes. The four new genes, ALK9 to ALK12, presented higher homology to ALK1-ALK3 (Nicaud et al., unpublished). A gene for cytochrome b5 (CYB5) and two NADH-b5 reductase genes (CBR1, MCR1), probably involved in electron transport to P450, as well as other P450 genes involved in ergosterol biosynthesis (CYP51, CYP61) are also present (Mauersberger, unpublished).
ALK1 is involved in short-chain alkane (C10) degradation. A strain disrupted in ALK1 was unable to grow on decane while it grows on hexadecane . Single deletions of ALK2, ALK3, ALK4 or ALK6 did not affect growth on alkanes. A double-disruptant alk1KO alk2-KO is unable to grow on both alkanes , indicating their role as AMOS in terminal alkane hydroxylation (cf. Fig. 4). Recent studies on the induction of ALK1 by decane identified an alkane-responsive region (ARR1), consisting of two alkane-responsive elements (ARE1 and 2), and revealed that peroxisome deficiency repressed ALK1 induction, involving PEX10, PEX5, and PEX6 gene products [34,56]. They also identified the yeast alkane-signaling gene YAS1 encoding a transcription factor for alkane signaling, essential for induction of ALK1. Recently, eight ALK genes from Y. lipolytica strain ATTC-8661 have been expressed in plant leaves of Nicotiana benthamiana, indicating that ALK3, ALK5, and ALK7 code for lauric acid ω-hydroxylases (FAH), whereas no ω-hydroxylase activity was detected with ALK1, ALK2, ALK4, and ALK6. Whether these P450s themselves are able to catalyse efficiently a cascade of sequential mono- and diterminal monooxygenation reactions, thus oxidising alkanes directly to fatty acids and DCA (shown in vitro for C. maltosa P450 52A3 ), and whether this alternative P450 function plays a role in vivo in diterminal alkane oxidation (Fig. 5), it remains to be clarified.
FAOD, FADH, and ADH. It is well established now that in alkane-utilising yeasts the long-chain 1-alkanol oxidation to fatty aldehyde is mainly catalysed by alkane-inducible, peroxisomal membrane-bound FAOD, and not by particulate NAD(P)-FADH, as was previously assumed. H2O2-producing FAOD activity was demonstrated for Y. lipolytica. Later, a 70-kDa FAOD protein was purified from alkane-grown Y. lipolytica H222 and studies on FAOD substrate specificity indicated the presence of probably several FAOD enzymes . However, surprisingly no gene coding for an FAOD could be detected in Y. lipolytica using the FAOD protein sequences from different Candida yeasts  as query (Smit, Nicaud and Mauersberger, unpublished). Furthermore, at least four NAD(P)+-dependent soluble alcohol dehydrogenase activities (ADH or FADH) with substrate specificities towards ethanol to 1-octanol were described . More recently probable short-chain specific ADH1-3 genes have been cloned (Kim et al., unpublished, GenBank Accession Nos. AF175271 to AF175273). Additionally, at least one putative long-chain-specific FADH gene (ADH4) was detected in the Y. lipolytica genome sequence (Mauersberger, unpublished). Whether some of these soluble (F)ADH are also involved in long-chain fatty-alcohol oxidation remains to be clarified (Figs. 4 and 5).
FALDH. Alkane-inducible and membrane-bound NAD(P)+-FALDH as well as soluble (F)ALDH activities were detected in several alkane-assimilating yeasts, including Y. lipolytica[17,60]. When the aa sequence of human FALDH  was used as query, four putative FALDH genes named FALDH1 to FALDH4 were found. The deduced proteins showed 34–39% aa identity with the confirmed human FALDH, and lower identity (<25%) with other classes of ALDHs (Smit and Nicaud, unpublished). Northern-blot analysis showed that FALDH4 is induced in cells grown in alkane medium (Smit et al., unpublished). Functional analysis by sequential gene disruption using the new gene disruption method developed for Y. lipolytica is currently in progress. Thus, the enzymatic systems for oxidation of higher alcohols and aldehydes (membrane-bound FAOD and FALDH) occur predominantly in peroxisomes, although activity of isoenzymes in ER membranes or in cytosol could not be excluded (Fig. 5). Both alkanes and triglycerides result in the production of FFAs, which are further degraded in the peroxisome by the β-oxidation pathway (Section 2.4). Considering the different intracellular localisation of mono- and diterminal alkane oxidation reactions, alkane or fatty-acid hydroxylation in ER and 1-alkanol and aldehyde oxidation in peroxisomes, the fatty alcohol, and not fatty acid or DCA, must be the essential transport form for long alkyl chains in the main substrate flux in alkane-growing yeast cells. This suggests a direct functional importance of the proliferation of ER membrane structures and their tight association with peroxisomes during yeast growth on n-alkane ([17,31,60,66], Figs. 2 and 5).
Although a diterminal orω-oxidation pathway of alkane (Fig. 5) is not achieved to any noticeable extent in growing cells , in β-oxidation mutants it can become the main pathway allowing the industrial production of long-chain DCA ([67,68], cf. Section 3).
In addition to these catabolic pathways, a part of the fatty acids derived from long-chain alkanes (C14-C18) is directly used for lipid synthesis (Fig. 5). After activation by acyl-CoA synthetase I (ACS I in ER, Section 2.4), and partial fatty-acid elongation and desaturation, the resulting saturated and unsaturated fatty acids of membrane chain-lengths (C16–C18), are incorporated into cellular phospholipids by glycerol-3-phosphate acyltransferase (GAT), or into triglycerides for transport and storage into lipid bodies as triacylglycerol (TAG) or steryl esters (STE) (Figs. 2 and 4, cf. Section 2.5).
2.4Fatty-acid degradation by β-oxidation
Inside the cell, fatty acids are energised as acyl-CoA esters. In Y. lipolytica long-chain-specific acyl-CoA synthetases I and II (ACS I and II) have been detected and studied about 25 years ago [69,70]. ACS I is distributed among different subcellular fractions and appears to be involved in lipid synthesis, whereas ACS II is present in peroxisomes where β-oxidation, the main pathway for the breakdown of these fatty acid-CoA esters, takes place (Figs. 2 and 5). This degradation system consists in a four-reaction sequence resulting in a two-carbon shortening of the CoA ester of fatty acid (Fig. 6). This cycle is repeated several times, theoretically until the complete breakdown of the acyl-CoA. Practically, the pathway is often stopped earlier depending on the physico-chemical properties of the acyl-CoA, on the chain length and concentration of the substrate, on the presence of CoA, acetyl-CoA and on the NAD+/NADH ratio. From the four enzymatic activities of the pathway, the two oxidations appear to exert the highest control. For the first reaction, some yeast species possess more than one acyl-CoA oxidase-encoding gene: three for C. tropicalis, two for C. maltosa, but only one for S. cerevisiae. In Y. lipolytica, the large number of genes indicate the adaptation of this species to hydrophobic substrates. From the genome study, a sixth gene (POX6) coding for an acyl-CoA oxidase (Aox6) has recently been added to the five genes (POX1 to POX5 GenBank Accession Nos. AJ001299 to AJ001303) previously detected . The global acyl-CoA oxidase enzyme is imported as a heteropentamer into the peroxisome . Studies with mutants  and with purified enzymes [73,74] have shown that two acyl-CoA oxidases (Aox2 and Aox3) were chain-length-specific, whereas the other exhibited a weak activity towards a large spectrum of chain-lengths (Aox4 and Aox5) or had no detectable activity (Aox1). Protein shuffling and mutagenesis experiments broadened the substrate selectivity of Aox3 (short-chain-specific) to long-chain acyl-CoA by only a single aa modification [44,74]. In contrast to acyl-CoA oxidases, which are encoded by numerous genes, the two other enzymes, the multifunctional enzyme bearing the hydratase and dehydrogenase activities (MFE2, AF198225) and the 3-ketoacyl-thiolase (POT1, X69988), are each encoded by only one gene (Figs. 2 and 5). More recently, a decane-inducible peroxisomal acetoacetyl-CoA thiolase (PAT1, AB042276) was cloned by complementing a C10−mutant. This enzyme showed a high homology to the corresponding enzymes in C. tropicalis and S. cerevisiae ERG10-encoded proteins . Both thiolase types are thought to be involved in the last step of β-oxidation in Y. lipolytica. As in other yeasts, a second acetoacetyl-CoA thiolase gene was found in the Y. lipolytica genome sequence (Mauersberger, unpublished).
The acetyl-CoA formed in β-oxidation enters into the glyoxylate cycle (key enzyme isocitrate lyase, ICL1), and is linked to the tricarboxylic acid (TCA) and methylcitrate (when using odd-chain HS) cycles, located in mitochondria ([4,31], Fig. 2). In accordance, the key enzyme encoding the methylisocitrate lyase gene ICL2 was found in the Y. lipolytica genome sequence (Mauersberger, unpublished).
2.5Lipid accumulation into lipid bodies
Depending on environmental conditions, yeast cells are able to mobilise FFA or to store them as triacylglycerols (TAG) and steryl esters (STE) into LB (for review see ). Lipid bodies (LB) consist of a hydrophobic core formed from neutral lipids, mainly TAG and to a lesser extent STE, which is surrounded by a phospholipid monolayer with a few embedded proteins (for a review see ). Few LB could be observed in Y. lipolytica when grown in glucose medium whereas LB accumulation was observed during culture in fatty-acid or triglyceride medium (Fig. 7). After 6 h of growth in YNBO medium, small LBs could be observed (Fig. 7(a)). Then, LB size increased and they began to coalesce, giving larger but fewer LB in the cells (Fig. 7(b)) and only few, large LBs could be observed during the stationary phase (Fig. 7(c)). LB content and composition depend on growth conditions and substrate composition [77,78]. It was also demonstrated that yeast genotype could improve LB accumulation (; see also cf. 3.2).
Due to specific properties of Y. lipolytica, such as the very efficient utilisation of HS and the high secretion capacity of metabolites and proteins, several projects to use this yeast in different applications arose rapidly, mainly focusing on bioconversion processes in the chemical and food industries. In the past the main application was the production of single-cell protein (SCP) starting in the 60s (recent reviews [2,3,17]). Thereafter several processes were developed, some of them patented, notably for the production of citric acid, aromas, starters for the food industry, for biotransformation of steroids, or waste treatment. Some of these applications are described below.
3.1Alkane and fatty-acid bioconversion
Although initial interest in bioconversion with Y. lipolytica focused mainly on the production of SCP or citric acid from n-alkanes (cf. Sections 3.2 and 3.3), later attention turned towards the production of primary oxidation products and other metabolites directly from alkanes, fatty acids or triglycerides. The aim was to exploit mutants or even wild-type yeast strains by stimulating accumulation of commercially valuable intermediates [2,3,17,40,67,79,80,115].
Several mono- and diterminal oxidation products of long- and middle-chain alkanes and fatty acids (fatty alcohols, aldehydes and acids, ω-hydroxy fatty acids (HFA), wax esters, methyl citrate, and especially dicarboxylic acids (DCA)) are of industrial importance in the production of detergents, surfactants, lubricants, and cosmetics . In the 80s DCA and HFA production directly from HS (alkanes, fatty-acid derivatives, or renewable plant oils) with Y. lipolytica was studied by the Henkel company in Germany [40,79] and by institutions in China . Further patents of different companies for the application of Y. lipolytica, C. tropicalis and other yeasts in DCA and HFA production were filed. We will focus here on fatty-acid conversion into aroma compounds.
3.1.1Production of the aroma γ-lactone
γ-Decalactone exhibits oily fruity aroma notes. It is used in the formulation of fruit aromas such as strawberry, apricot and peach, but is also present in many fermented products such as bread and whiskies (recent reviews [80,81]). It is the most important lactone for flavour application and its market volume attains several hundred tons per year. It is relatively easy to produce by microbial conversion from castor oil and Y. lipolytica is the organism, yielding the highest product concentration .
The pathway involves the β-oxidation of the ricinoleyl-CoA (C18) until the production of the 4-hydroxydecanoic acid (C10) which lactonises readily to yield γ-decalactone (4-decanolide) . As β-oxidation controls an important part of the production , experiments carried out with acyl-CoA oxidase-deficient mutants of Y. lipolytica improved our understanding of the process and increased the yields significantly. Two main strategies have been used, the first one aimed to decrease the β-oxidation of the C10 lactone-precursor and the degradation of the aroma compound by deleting enzymes involved in β-oxidation of acyl-CoA shorter than C10 . For the second strategy, the exit of the metabolite between two β-oxidation cycles was favoured by decreasing the global acyl-CoA oxidase activity. This significantly decreased the yield-lowering intra-cycle exit which results in 3-hydroxy-γ-decalactone .
Apart from β-oxidation activity, another factor influencing the yields is the toxicity of the lactone. The mechanisms involved in this toxicity have been characterised showing that the carbon-lateral chain of the lactone interacts with membranes, increasing their fluidity and decreasing their integrity .
3.1.2Green-notes aroma production
A thorough knowledge of the behaviour of Y. lipolytica towards HS and the improvement of genetic tools have resulted in the selection of this species to express a heterologous bioconversion system for production of lipid-derived green notes. These compounds are the products of the lipoxygenase/hydroperoxide lyase (HPL) pathway which is induced in plants in response to stress. The first enzyme converts polyunsaturated C18-fatty acids into hydroperoxides that are cleaved by a second enzyme, giving rise to volatile aldehydes that play an important role in the flavour of vegetables due to the green and fresh sensorial notes they exhibit. Some processes have been developed to produce these compounds from plant sources  but the production stops rapidly due to the great instability of the lyase. In order to increase the productivity, attempts were made to produce and stabilise the enzyme in microorganisms. The expression of HPL in Y. lipolytica resulted in the production of much higher yields of hexanal .
3.2Production of SCP and SCO
From the mid 60s to the 80s, Y. lipolytica was used in large-scale fermentation for production of single-cell protein (SCP) from n-alkanes, allowing process development for this yeast in very large fermenters with high cell densities . Recently, these technologies were applied for single-cell oil (SCO) production including a new cocoa-butter substitute . Lipid accumulation was investigated by Aggelis and co-workers  using industrial fats or glycerol (see for example ). They improved lipid accumulation or modified the lipid composition, depending on the substrates used, growth conditions and media composition. It has been demonstrated that lipid accumulation could also be obtained by genetic manipulation of Y. lipolytica. Potential application of lipids and lipid fractions might include the production of cerebrosides from the sphingolipid fraction of hexadecane-grown yeast cells .
3.3Overproduction of intermediate metabolites
The yeast Y. lipolytica is unique in its ability to produce and excrete into the medium a broad range of organic acids, including the TCA-cycle intermediates citric (CA), isocitric (ICA) and 2-ketoglutaric (KG), as well as pyruvic (PY) acids [2–4,88,115]. These acids can be produced from low-cost carbon sources (alkanes, vegetables oils, fats, ethanol, molasses, starch hydrolysates) by growth limitation caused by different nutrition factors, like nitrogen source, thiamine, phosphate, or mineral salt compounds (P, S, Mg). Nitrogen exhaustion triggers secretion of CA and ICA, while limitation of thiamine at low pH values causes secretion of mainly KG and PY. Y. lipolytica offers an alternative to the traditionally used Aspergillus niger citrate production processes with molasses, which are associated with the accumulation of significant amounts of solid and liquid wastes. Indeed, Y. lipolytica could utilise different substrates, exhibit a higher maximal product formation rate (up to 3 g CA l−1 h−1) and a higher substrate-related yield (up to 1.5 g g−1, especially on HS). Thus, up to 250 g l−1 CA could be obtained using fed-batch fermentation conditions with 20% alkanes or sunflower oil as carbon source [88,91]. Furthermore it shows advantages in sugar tolerance, product isolation, waste and sewage minimisation, and the possibility to develop a process with closed circuits (Stottmeister, personal communication; see http://www.ufz.de). Therefore, processes are under development for Y. lipolytica using renewable low-cost substrates such as plant oils, fat and glycerol. Indeed, a CA process using rape seed oil has been introduced recently by Archer Daniels Midland (Decatur, Ill., USA; see ).
One disadvantage of using Y. lipolytica for CA production is that it simultaneously produces CA and ICA, the ratio depending on the carbon source and strain. Wild-type strains produce CA/ICA in a ratio of about 90:10 on glucose, glycerol or ethanol, while this ratio is 60:40 on alkanes or triglycerides. However, this ratio was changed in selected mutant strains used for process developments ([2,3,89,90]). This ratio was also modified by gene amplification of the ICL1 gene (10–25 copies) resulting in a ratio of nearly 95:5 on all substrates tested, including HS (; Mauersberger et al., unpublished). Addition of acetate to the fermentation medium with plant oil also increased the ratio from 60:40 to about 80:20 .
To extend the substrate spectrum of Y. lipolytica to sucrose-containing mixtures like molasses, the S. cerevisiae SUC2 gene, expressed with the XPR2 promoter and signal sequence, was introduced in different strains. Such Suc+-transformants were able to grow on sucrose and CA could be produced from molasses ([36,93]; Mauersberger et al., unpublished).
Furthermore, KG could be accumulated at up to 195 gl−1 with selected Y. lipolytica strains and alkanes in fed-batch conditions with a yield of 90% and a productivity of 1.4 gl−1h−1[2,3,94]. These production levels and processes developments demonstrate that Y. lipolytica may be used for CA and KG production.
3.4Application of Y. lipolytica in bioremediation processes
Given its ability to utilise alkanes, fatty acids and oil, it is not surprising that Y. lipolytica strains emerged from several independent studies as a very promising agent for the treatment of both mineral oil pollution and plant oil waste. In Italy, approximately 2 million m3 of wastewater are produced each year by the lipolytic industry. Oil mill wastewater (OMW) is an important source of water pollution when released into local rivers without treatment. OMW forms a stable emulsion, composed of water from the processing, olive pulp and oil. This emulsion presents a serious pollution problem as the waste is a dark acidic liquid with a strong oily smell containing fats, sugars, phosphate, phenol and metals. Different studies have reported the successful use of OMW as growth substrates for Y. lipolytica, resulting in a significant reduction of chemical oxygen demand (COD), together with lipolytic enzyme and biomass production. Scioli and Vollaro  reported a reduction of the COD level by approximately 80% in olive oil mill wastewater after 24 h of fermentation, whereas Oswal  reported a reduction of more than 90% in the COD level in palm oil mill effluent. These wastes (fats and oils) were also used as carbon source for biomass production (growth) or for bioconversion into valuable products like citric acid (see Sections 3.2 and 3.3).
Mineral oil contamination of soil and water occur frequently. Bioremediation has become the major method employed in the restoration of oil-polluted environments. A strain isolated from tropical marine samples gave optimal degradation of crude oil at 30 °C , while a strain isolated from diesel oil-contaminated alpine soils was able to degrade diesel oil at temperatures ranging from 4 to 30 °C with maximum activity between 10 and 20 °C. The latter strain was also used for the development of a biosensor for the determination of middle-chain alkanes at temperatures between 5 and 25 °C . A Y. lipolytica strain isolated from a polluted site in Korea  was immobilised in polyurethane foam and worked very well for the bioremediation of oil films on surface waters , degrading 50% of absorbed oil (7–9 g oil/g foam) within five days.
Y. lipolytica was also successfully used in the treatment of sewage sludge from the food industries. Such effluent, containing up to 90% of grease, was used to grow different strains selected for their ability to grow on hydrophobic substrates and to secrete large amounts of hydrolases. The best extracellular-lipase-producing strains were then selected and tested in a 6 m3grease tank with a feeding rate of 6 m3 every 24 h. The COD level was maintained at a value of 3000 mg l−1 during 33 weeks of treatment, indicating that most of the fatty compound was metabolised in situ [99,100]. Based on these observations, a starter composed of freeze-dried Y. lipolytica and extracellular lipase enzyme was developed and commercialised by Artechno S.A. (http://www.artechno.be).
Whole cells or lipases of Y. lipolytica have been applied in enantioselective resolution (hydrolysis, oxidation or reduction) and re-esterification reactions. The need for enantiomerically pure molecules, especially in the pharmaceutical industry, has grown since the legislation required investigations on the pharmacological effects of both enantiomers. The market of drugs sold as single-enantiomer has a 10% growth per year ($160 billion worldwide in 2002). The enzymatic method of resolving a racemic mixture is very attractive. The extracellular lipase Lip2p from Y. lipolytica was found to be very effective for the resolution of 2-halogeno-carboxylic acids, important intermediates in the synthetic pathways for a number of drugs (analgesics, prostaglandin, prostacyclin, semi-synthetic penicillin). Resolution of 2-bromo-p-tolylacetic acid ethyl ester catalysed by Y. lipolytica Lip2 lipase showed an S-enantiopreference of 28, which is similar to the best result obtained with Burkholderia cepacia lipase (E= 30). For the 2-bromo-o-tolylacetic acid, a precursor of analgesics and non-peptide angiotensin II-receptor antagonists, none of the commercial lipases was able to resolve the racemic mixture . Y. lipolytica lipase is the only enzyme able to perform this resolution (E= 27) . From a kinetic point of view, Y. lipolytica lipase is ten to one hundred times more active than the lipase from B. cepacia. This impressive catalytic activity is confirmed by the kcat value (60.000 s−1) observed with triolein.
To obtain large amounts of this lipase, an expression system was developed [6,103]. Strains with the protease-encoding genes deleted, and containing multicopy integration of the expression cassettes (3–16 copies) carrying the LIP2 gene expressed under control of the oleic acid-inducible POX2 promoter, were obtained using ura3d4 as selection marker gene . In parallel, a mineral medium was developed for lipase production which fulfilled nutritional requirements of Y. lipolytica and enabled, in fed-batch mode, 100 g l−1 of biomass to be obtained with the tremendous production of 60.000 lipase U ml−1(1 U = 1 micromole triolein hydrolysed per minute). After concentration by ultrafiltration and lyophilisation, an active powder was obtained (10 g l−1). The 3D-structure of Lip2p was determined in silico by homology modelling with the lipase from Rhizomucor miehei (29% identity). Mutants of Lip2p were obtained with improved enantioselectivity (E > 200) and higher activities (Marty et al., unpublished).
Furthermore, wild-type strains, mutants or enzymes of Y. lipolytica were used by Fantin and co-workers for the enantioselective oxidation or reduction of 2-alkanols or cyclic alcohols as well as for the efficient enantioselective hydrolysis of esters, enol-esters, epoxides or lactones . The enantioselective oxidations of alcohols are probably catalysed by fatty-alcohol oxidases (FAOD) or alcohol dehydrogenases (FADH/ADH), as described for the FAOD from C. maltosa. Kinetic resolution of racemic secondary alcohols via oxidation with whole cells of Y. lipolytica offers an alternative approach to the synthesis of these sought-after intermediates for organic synthesis . Otherwise, stereoselective reduction of ketones presents the possibility of asymmetric synthesis of single enantiomers of secondary alcohols, thus offering the possibility of producing the desired enantiomer in 100% yield (while the maximum yield from a kinetic resolution is only 50%). In a study on the reduction of aryloxy-halo-2-propanones, Y. lipolytica displayed superior enantioselectivity compared to S. cerevisiae, which is widely used for this type of biotransformation . Y. lipolytica also sometimes displays opposite enantioselectivity to S. cerevisiae, giving the R- enantiomer, while S. cerevisiae usually gives the S-enantiomer .
Additionally, whole cells of Y. lipolytica gave enantioselective hydrolysis of a range of racemic acetyl esters of secondary alcohols as well as enol esters, lactones and epoxides . These activities can probably be ascribed to the membrane-bound lipases or carboxylesterases of Y. lipolytica (Section 2.2.1). The reported hydrolysis of styrene oxide is of particular interest, since annotation of the Y. lipolytica genome revealed three putative epoxide hydrolases (GenBank Accession Nos. XP504164.1, XP502171.1, XP499652.1) that showed 39–45% aa identity with each other and 21–25% aa identity with mammalian cytosolic soluble epoxide hydrolases (Nicaud and Smit, unpublished results).
3.6Cytochrome P450-catalysed biotransformation by recombinant Y. lipolytica cells
Alkane-utilising yeasts, such as Y. lipolytica or Candida spp., exhibit a high catalytic activity of their P450s (AMOS, FAH), catalysing terminal hydroxylations of alkanes or fatty acids. This is supported by an efficient subcellular organisation, facilitating substrate and product transport processes, a proliferation of the ER as well as an efficient electron transfer system ([17,50,66], cf. Section 2.3). Therefore, heterologous P450 expression and bioconversion experiments were tested in Y. lipolytica.
The first attempts involved expression of bovine P45017α (CYP17A cDNA) and results were compared with those obtained in S. cerevisiae. Functional expression of P45017α in Y. lipolytica under control of the strong and regulated ICL1 promoter  was first demonstrated using an ARS/CEN low-copy plasmid. The expressed P45017α was highly active, especially in alkane-grown cells [106,107]. Subsequently, multicopy (8–35) transformants were obtained by multicopy integration into rDNA or LTR zeta of Ylt1 according to . By the same approach, strains of the opposite mating type expressing high levels of the homologous NADPH-P450 reductase (YlCPR) under pICL1 control were constructed. These two types of haploid multicopy transformants were used to construct diploid strains containing multiple cassettes for both P45017α and P450 reductase. Simultaneous overexpression of P45017α and P450 reductase in diploid strains resulted in increased steroid bioconversion rates. The P45017 expressing strains catalysed the biotransformations of progesterone, pregnenolone and related derivatives into the corresponding 17α-hydroxy-steroids ([107,116], Mauersberger et al, unpublished results). By using a combination of yeast cell biotransformation and chemical oxidation, androstendione can be obtained starting from progesterone, thus extending the abilities of enzymatic-chemical steroid synthesis with recombinant yeast cells .
More recently, functional expression of human P4501A1 (CYP1A1, involved in drug oxidation) in Y. lipolytica, with or without overproduction of Y. lipolytica NADPH-P450 reductase (YlCPR), was demonstrated. The highest P4501A1 activity was observed in whole-cell biotransformation of hydroxyresorufin into resorufin when multiple copies of CYP1A1 were inserted, and YlCPR was co-expressed ().
The stable high-level and functional expression of heterologous P450s together with its NADPH-P450 reductase opens new perspectives for further improvement of the efficiency of biotransformation reactions with recombinant Y. lipolytica cells, a system which seems to be useful especially for hydrophobic substrates.
3.7Y. lipolytica in the agro-food industry
An increasing number of studies clearly demonstrate the natural occurrence of Y. lipolytica in different kinds of food, underlining the importance of this yeast in the agro-food industry. The appearance of different yeasts, including very often Y. lipolytica, and their role in the production, ripening or spoilage of traditional dairy products (cheese, yoghurt) and of sausages and other meat products, has been studied during the last decade. The participation and influence of Y. lipolytica on the pigment formation during Portuguese cheese production was recently studied [110,111]. It is currently recognised that yeasts contribute to the maturation of cheese. Addition of Y. lipolytica was shown to accelerate ripening and to improve the quality of cheeses [21,48,112,113]. Recently, Ferriera and Viljoen  proposed to add Y. lipolytica in addition to Debaryomyces hansenii as part of the starter culture for the production of matured Cheddar cheese . The selection of the Y. lipolytica strain to be used is important, since a large variation in lipase and protease activities is observed depending on the strain [21,24].
Due to the ability of Y. lipolytica to degrade and oxidise very efficiently HS, such as fats, oils, alkanes, and fatty acids, and the development of powerful tools for genetics and molecular biology, this yeast recently has been used as a model organism to investigate the metabolic pathways involved in HS metabolism, including HS transport processes, biogenesis and degradation of organelles (peroxisome and lipid body), and the identification of genes involved in these processes and their regulation. Classical and tagged mutants were isolated and used for the identification of these genes. Although the main HS degradation pathways and the genes involved are known, genes encoding lipases/esterases (LIP1-8), alkane- and fatty-acid-hydroxylating P450 systems (ALK1-12, CPR), peroxisomal β-oxidation enzymes, acyl-CoA-oxidases (POX1-6), multifunctional enzymes (MFE2), thiolases (POT1 and PAT1), and the glyoxylate-cycle isocitrate lyase (ICL1), there is still a lack of understanding of their regulation, structure-function relationships and in particular HS transport processes.
One of the most striking features in Y. lipolytica is the presence of several multigene families involved in these pathways (LIP, P450 ALK, FALDH, POX). This multiplicity of genes is obviously the basis for a wide range of substrate and chain-lengths specificities of enzymatic steps involved in HS degradation in this yeast. The recent availability of the complete genome sequence of Y. lipolytica, and the development of efficient gene expression and gene disruption methods, will certainly speed up characterisation of the genes involved in HS utilisation. These and subsequent studies will significantly contribute to a deeper understanding of HS-metabolising pathways in this yeasts.
Since initial industrial utilisation of Y. lipolytica in the 60s for the production of SCP, several new lines of application have been tested. We have presented here the potential and actual utilisations of wild-type, mutant or recombinant strains in bioconversion, aroma production, SCO production, intermediate-metabolite production, in bioremediation processes and in food technology. Recombinant Y. lipolytica strains expressing heterologous P450s have been tested as new powerful tools for the development of efficient P450-catalysed biotransformation systems. The expected increase in knowledge of HS metabolism in Y. lipolytica consequent upon the availability of its genome sequence will demonstrate its great utility and will offer new perspectives for the construction of new yeast “cell factories” with high potential for biotechnological applications.
We were not able to cite all of the contributions made to the field of HS utilisation by Y. lipolytica, and apologise to those authors whose publications might have been omitted. This work was supported by the Institut National de la Recherche Agronomique, and by the Centre National de la Recherche Scientifique (France). P.F. is a recipient of a fellowship from the Fond pour la Formation à la Recherche dans l'Industrie et l'Agriculture (Belgium). S.M. was partially supported by Dresden University of Technology, by DAAD (Deutscher Akademischer Austauschdienst – PROCOPE, a German-French exchange program), and by grants from the Sächsisches Staatsministerium für Umwelt und Landwirtschaft (SMUL, Saxony), and the Bundesministerium für Bildung und Forschung (BMBF) of Germany. M.S. recently spent a six-month sabbatical, funded by the CNRS and the South African National Research Foundation, in the laboratory of J.M.N. We thank Colin Tinsley for editing the English language of the text.