Rhizopus oryzae Lipase, a Promising Industrial Enzyme: Biochemical Characteristics, Production and Biocatalytic Applications

Abstract

Lipases are biocatalysts with a significant potential to enable a shift from current pollutant manufacturing processes to environmentally sustainable approaches. The main reason of this prospect is their catalytic versatility as they carry out several industrially relevant reactions as hydrolysis of fats in water/lipid interface and synthesis reactions in solvent-free or non-aqueous media such as transesterification, interesterification and esterification. Because of the outstanding traits of Rhizopus oryzae lipase (ROL), 1,3-specificity, high enantioselectivity and stability in organic media, its application in energy, food and pharmaceutical industrial sector has been widely studied. Significant advances have been made in the biochemical characterisation of ROL particularly in how its activity and stability are affected by the presence of its prosequence. In addition, native and heterologous production of ROL, the latter in cell factories like Escherichia coli, Saccharomyces cerevisiae and Komagataella phaffii (Pichia pastoris), have been thoroughly described. Therefore, in this review, we summarise the current knowledge about R. oryzae lipase (i) biochemical characteristics, (ii) production strategies and (iii) potential industrial applications.

1. Introduction

Enzymes are highly efficient biocatalysts with several current and potential future applications in industrial-scale catalysis because of their widely researched advantages, such as their operation in milder conditions, outstanding selectivity and specificity and their lower environmental and physiological toxicity [1,2,3,4]. Likewise, social concerns and public policies are forcing industries to shift from manufacturing pollutant processes to more eco-friendly approaches in which biocatalysis can develop a significant role [5,6,7,8,9]. Consequently, the market for industrially relevant enzymes is expected to reach $7.0 billion by 2023 from $5.5 billion in 2018 at a remarkable compound annual growth rate (CAGR) of 4.9% for the period 2018–2023 [10].

Currently, hydrolytic enzymes dominate the market of industrial enzymes, accounting for more than 70% of all the sales [11]. Amongst them, lipases (triacylglycerol hydrolases E.C. 3.1.1.3) are the third largest group after proteases and amylases [12,13,14]. The main reason of their importance is based on the wide variety of industrially relevant reactions that they can carry out (Figure 1). In aqueous media, lipases catalyse the hydrolysis of fats at the water/lipid interface by cleaving the ester bonds with the concomitant consumption of water molecules and producing the corresponding fatty acids and glycerol or alcohol. Additionally, because of the principle of microreversibility, in non-aqueous media they might perform the reverse reaction, the so-called synthesis reactions. These reactions, in turn, can be classified as esterification, where a double displacement reaction between alcohols and carboxylic acids occurs resulting in esters and water; transesterification, where the acyl group of an ester switches with an alkyl group of an alcohol, acid or another ester resulting in alcoholysis, acidolysis and interesterification reactions respectively; and rarely, aminolysis, where carboxylic esters are converted to the corresponding carboxylic amides [15,16,17,18,19,20]. This vast versatility makes lipases suitable for several applications such as flavours, detergents, pharmaceuticals, bioenergy- and fine-chemicals-producing industries [14,21].

Lipases can be fundamentally classified based on two traits (Figure 2), specificity and source [15,22]. The former includes three different subclasses; substrate specific lipases, which can selectively act on a specific substrate between a mixture of crude raw materials [23,24,25,26], enantioselective lipases which preferentially hydrolyse one of the isomers of a racemate over the other [27,28,29,30] and regioselective lipases which favour their activity depending on positional specificity and can be divided, in turn, in three classes [31,32,33]. Non-specific lipases catalyse the hydrolysis of triacylglycerols (TAGs) into free fatty acids and glycerol with mono- and di-glycerides as intermediates. These lipases can act over all the fatty acids from any position of the substrate. For instance, Candida antartica lipase is one of the most employed and studied non-specific lipase [34,35]. Fatty acid-specific lipases show fatty acid selectivity and they tend to hydrolyse esters containing long-chain fatty acids with double bonds in between C-9 and C-10. Some examples of these enzymes are lipases from Geotrichum candidum [36] and Penicillium citrinum [37]. 1,3-specific lipases, such as Rhizopus oryzae [38,39] and Yarrowia lipolytica lipases [40], hydrolyse the acylglycerols present in positions sn-1 and sn-3 of the TAG producing free fatty acids, 2-monoacylglycerols and 1,2- or 2,3-diacylglycerols, avoiding the final production of glycerol. During this reaction, formation of diacylglycerols has been reported to occur faster than monoacylglycerols formation [33]. Moreover, with 1,3-specific lipases, acyl migration phenomena might take place, which is a non-enzymatic process that involves spontaneous movement of an acyl group from one hydroxyl group to an adjacent one. Thus, the formed product after the hydrolysis with 1,3-specific lipases might suffer acyl migration yielding 1,3-diacylglycerols and 1-monoacylglycerols from 1,2- or 2,3-diacylglycerols and 2-monoacylglycerols, respectively. In addition, glycerol might be obtained as the final product after extended reaction times [32].

In terms of lipases source, they can be classified as animals, plants, insects and microbial lipases which includes yeasts, bacteria and fungi lipases [15,41]. Microbial lipases have been described to have some desirable traits in comparison to animals and plants lipases, for instance: higher stability, chemo-selectivity, enantio-selectivity and the absence of cofactor requirements [16,22,31,42,43]. Concretely, fungal lipases have been broadly studied because of the known potential of native microorganisms to produce these enzymes and their notable unique catalytic properties. At industrial scale, most employed lipase-producing fungi belong to the genera of Rhizopus sp., Aspergillus sp., Penicillium sp., Geotrichum sp. and Mucor sp. [22,44,45].

Focusing on Rhizopus genus, owing to its industrial relevance, some attempts have been carried out in order to classify the species forming it and to reflect the key differences between them. However, the classification has changed over time and might have had a negative effect on lipase notation accuracy. Traditionally, the species have been differentiated based on morphological and physiological features [46]. For instance, Schipper [47] described the existence of three groups—R. microsporus, R. stolonifer and R. arrhizus (=oryzae). Lately, Abe et al. [48] carried out the first molecular phylogenetic study and concurred with the classification performed by Schipper. Subsequent studies differed from that classification and organised the genus into 10 [49] and 8 [50] species until Abe et al. [51] finally confirmed the 8 species division by using rDNA ITS, actin-1 and translation elongation factor 1α (EF-1α) sequences as differential traits. Notwithstanding these classifications, currently, most of the Rhizopus samples in culture collections belong to four species or species complexes: R. microsporus, R. stolonifer, R. arrhizus (or R. oryzae) and R. delemar (or R. arrhizus var. delemar) [46]. Amongst them, according to GenBank records, R. oryzae or arrhizus (including R. arrhizus var. delemar) is the most significant one as more than 7000 identified isolates have been reported [46].

R. oryzae is broadly employed in industry because it can carry out the synthesis of a great variety of products like organic acids (lactic and fumaric acids), volatile compounds and enzymes (cellulases, proteases, tannases, xylanasas, pyruvate decarboxylases, lipases etc.,) [52,53,54,55]. Concretely, according to Web of Knowledge data, R. oryzae lipase (ROL) is one of the most studied enzymes of this fungi. There are three major commercial formulations of this lipase (

Table 1. Major commercial suppliers of Rhizopus oryzae lipase and some lipase properties [16].

Supplier	Name	Application	Lipase Properties
Amano	Lipase DF “Amano” 15	Oil and fats	Optimum pH range 6–7; stable pH range 4–7, optimum temperature range 35–40 °C, relatively specific to fatty acids
Sigma	Lipase from R. oryzae (no. 62305)	Oil and fats	Optimum pH 8, optimum temperature 40 °C
Sigma	Lipase, immobilised on Immobead 150 from R. oryzae (no. 89445)	Pharmaceutical and bioenergy	Optimum pH 7.5, optimum temperature 40 °C

) and more than 200 scientific works have been published in the last 5 years highlighting the relevance of this enzyme. Therefore, the aim of this review is to provide a complete overview of ROL in terms of biochemical properties, enzyme native and heterologous production and its industrial applications.

2. Biochemical Properties

R. oryzae lipase (ROL) is a protein synthesised as a precursor form containing a presequence of 26 amino acids, followed by a prosequence of 97 attached to the N-terminal of a 269 amino acids mature sequence (Figure 3) [56]. All known lipases from Rhizopus genus follow the same identical structure even though some amino acidic substitutions can be detected when their primary sequences are compared, not only between different species but also between different isolated strains of the same species (Figure 4). For instance, Ben Salah et al. [57] addressed the presence of several substitutions in the sequences of Rhizopus lipases published by his group and Sayari et al. [58], Beer et al. [56], Derewenda et al. [59] and Khono et al. [60].

ROL contains four potential N-glycosylations sites (Figure 3) that follow the consensus sequence Asn-X-Ser/Thr, where X is any amino acid instead of proline. One of these putative sites is found in the prosequence where modifications in glycosylation patterns have been described to have an effect on protein secretion [61]. For instance, Yu et al. [62] added two extra N-glycosylation sites to ROL prosequence and expressed this mutant in Komagataella phaffii (Pichia pastoris). The extracellular activity and total protein were 218- and 6.25-fold higher respectively in the strain harbouring the two extra N-glycosylation sites than in the non-modified one highlighting the relevance of glycosylation.

The presequence of ROL has been described to act as signal peptide promoting enzyme secretion, while the prosequence has been reported to exhibit diverse functions that are still under research. Beer et al. [63] depicted the significance of the latter in lowering lipase toxicity during its synthesis and in acting as intramolecular chaperone enabling the proper folding of the enzyme. In fact, genetically modified E. coli strains producing heterologous ROL without the prosequence resulted in cell lysis. To date, a large number of prosequences of different enzymes have also been identified to function as intramolecular chaperone and to assist the folding of their respective proteins [64]. In addition, several scientific works have related ROL prosequence with the translocation of the protein across the endoplasmic reticulum membrane, enhancement of free lipase stability and changes in enzyme substrate specificity. Nevertheless, the mechanisms that allow these traits are yet unknown despite the broad research carried out [65,66,67,68,69,70]. In any case, both the presequence and the prosequence are expected to be proteolytically removed to form the mature lipase. In spite of this, the native microorganism secretes a lipase that is attached to the N-terminal of mature sequence the 28 C-terminal amino acids of the prosequence (proROL), which then are cleaved via limited proteolysis catalysed by extracellular proteases [56,60,71]. However, some studies have indicated that the presence of these 28 amino acids of the prosequence alongside the mature sequence is enough for some of the presumed features of the prosequence to occur. For instance, higher free lipase stability and changes in enzyme specificity have been described when the 28 amino acids of the prosequence were expressed together with the mature sequence in K. phaffii [39]. In addition, these amino acids have also enabled lowering the toxicity of ROL production in E. coli [63] and they have been related to direct proteins to secretory pathway in Aspergillus oryzae [72].

The mature sequence of R. oryzae lipase (rROL) is constituted by 269 amino acids and the protein formed by them has a molecular weight (MW) of 29.542 kDa and a isoelectric point (pI) of 8—calculated by Expasy Proteomics Server [57]. These results agree with the published experimental data (

Table 2. Biochemical properties and substrate specificity of different published works dealing with ROL.

Lipase Name 1	MW (kDa)	Isoelectric Point	pH Optimum	T Optimum (°C)	Substrate Specificity	Ref.
rROL	29		8/7.25 2	30/40 2	C12 > C10 > C8 > C4 4	[39]
proROL	32		7.25	40	C8 > C12 > C10 > C4 4	[39]
rROL	30		8.5			[56]
entire-proROL	40		8			[56]
pre-entire-proROL 3	42		8			[56]
rROL	29		8	37		[57]
rROL	29					[58]
proROL	32					[58]
proROL	34		6–6.5	35		[60]
rROL	30		6	40		[60]
proROL	35		9	40	C16 > C18 > C12 > C8 > C4 5 C16 > C12 > C8 > C18 > C4 6	[68]
proROL	32	6.9				[70]
rROL	30	9.3	8.25	30	C8 > C10 > C6 > C4 > C12 > C16,C14 > C2 6	[73]
proROL	35		5.2	30	C12 > C10 > C8 > C6 > C16 > C5 > C4 > C3 > C2 4	[74]
proROL	32	7.6	7.5	35	C8 > C6 > C4 > C2 6	[75]
rROL	29				C12 > C10 > C8 > C6 > C4 > C3 > C2 4 C8 > C10 > C18 > C4 > C6 6	[90]
proROL	34				C2 > C3 > C8 > C6 > C12 > C10 > C4 4 C8 > C10 > C4 > C6 > C18 6	[90]
proROL			8	40		[92]
rROL	30.3	8.6	8–8.5	30		[93]
proROL			8.5	30		[94]
proROL	37		8.5	40		[95]
rROL	29		8			[96]
ROL	17	4.2	7	40		[97]
ROL			7	40		[98]
ROL			6	45	C8 > C4 > C6 > C2 6 C8 > C12 > C14 > C16 > C18 5	[99]
proROL	32		7	35		[100]
ROL			6	30	C7,C8,C12,C16 > C2,C3,C4,C18 5	[101]
ROL			7.5	50		[102]
proROL	32		7.5	30–40		[103]
ROL	14.45	6.5	9	30–40	C16 > C18 > C12 > C8 > C4 > C2 4	[104]
			8.3	35–37		[105]
proROL	35				C10 > C14 > C12 > C8 > C6 > C4 > C16 6	[65]
entire-proROL	46					[65]

¹ Names are based on the established nomenclature in this review. ROL indicates that the lipase cannot be classified under the determined parameters in this work; ² different values caused by the employment of 200 or 400 mM tris-HCl buffer; ³ pre-entire-proROL includes the presequence as well as entire-proROL; ⁴ p-nitrophenol esters were employed for substrate specificity analysis; ⁵ methyl esters of different carbon chain length were employed for substrate specificity analysis. Saturated methyl esters are just considered; ⁶ homotriacylglycerols were employed for substrate specificity analysis.

) in which MW and pI values around 29 kDa and 8 have been respectively reported [58,60,73,74]. However, variations in these values can be found because of the presence of the 28 amino acids of the prosequence described above [39,60,70,75,76]. In this case, MW increases to 32 kDa and pI decreases roughly to 7, highlighting the average acid nature of these 28 amino acids. Besides, the production of a lipase including the whole prosequence and close to 40 kDa has also been described (entire-proROL) [56,65].

The 3D structure of the lipase from R. oryzae [59,60] (Figure 5) and several microorganisms more, such as Geotrichum candidum [77], Candida rugosa [78], Pseudomonas glumae [79] and Penicillium camemberti [80] have been crystallographically resolved and showed that all lipases have a common α/β hydrolase fold structure that can also be found in other hydrolases [45]. Regarding ROL, it contains nine α-helixes and eight β-strands forming a molecule that it is stabilised by three disulfide bonds between residues 29–269, 40–43 and 235–244 [60]. In addition, this structure contains three key components that can be also found in most lipases besides ROL, the lid, the active site and the oxyanion hole [15]. The lid is an amphiphilic loop—also called flap—that covers the active site preventing the access of the substrate while the enzyme is in aqueous medium [81]. The active site, in turn, is mainly responsible for carrying out enzyme catalysis and consists, in all α/β hydrolases, of a highly conserved catalytic triad formed by a nucleophilic, a catalytic acidic and a histidine residues. In lipases, this triad is composed of nucleophilic serine residue and an aspartic or glutamic acid residue that it is bonded to a histidine; hence, lipases are classified as serine hydrolases. In the specific case of ROL, the lid domain is a short α-helix structure formed by six amino acids (FRSAIT) and the active site is formed by three amino acids Ser¹⁴⁵, Asp²⁰⁴ and His²⁵⁷ [59,60,82,83]. The function of these two elements is crucial during catalysis in which the lipase binds to the water/lipid interface and the lid opening occurs by a concomitant structural change in the substrate-binding site that enables the coupling of the substrate to the active site—lid-closed and partially opened 3D structures of Rhizopus delemar (=oryzae) lipase have been described by Derewenda et al. [59]. The structural change undergone is known as ‘interfacial activation’ and it is a unique property of lipases that enables them to hydrolyse insoluble esters and to distinguish them from esterases that can hydrolyse water-soluble esters [15,84,85,86]. It must be highlighted that the 28 amino acids of the prosequence introduced above have been deemed to interfere in this process as they are located next to the lid region and contain 50% hydrophobic residues. Therefore, this sequence extends the hydrophobic patch created in the open lipase by the open lid and the catalytic crevice influencing the interaction with the lipidic substrate [58]. This role might explain some of the assumed properties of these 28 amino acids, however, the mechanism remains unknown. Additionally, together with the catalytic triad and the lid, the oxyanion hole plays an important role and it is also a highly conserved sequence that largely influences the catalytic efficiency of the enzyme. During the hydrolysis reaction, a negatively charged tetrahedral intermediate is generated and it gets stabilised by hydrogen binding with the oxyanion hole [15,87,88]. This function has been described to be presumably performed by the hydroxyl and main-chain amide groups of Thr⁸³ in ROL [59,60,89].

Because of the relevance of the lipolytic activity of this enzyme, it has been widely researched in order to know how it is affected by the conditions of reaction medium. Guillen et al. [90] described that ionic strength has a remarkable impact. Actually, the relative activity of ROL in 200 mM Tris-HCl was reported to be twice the activity observed in 400 mM. Moreover, as all enzymes, ROL activity is highly influenced by the pH and temperature. Optimum activity pH values of 8 have been principally reported [56,57,73,75,90,91,92,93,94,95,96]. However, other studies have also stated more acid [74,97,98,99,100,101,102,103] and basic [68,104,105] optimums. Regarding temperature, most of the published optimum values can be found between 30 and 45 °C. In fact, 40 °C has been the most commonly reported optimum [39,60,68,90,92,95,97,98,103,104] although lower [57,60,74,75,90,93,94,100,101] and higher [99,102] values have also been described (Table 2). Nevertheless, for both pH and temperature, as can be observed in Table 2, some of the differences are based on the presence of the 28 amino acids of the prosequence. In this line, Kohno et al. [60] reported these differences and afterwards, other works [39,63,90] described similar results highlighting the relevance of these amino acids in lipase catalytic performance.

The presence of metal ions in reaction medium has been extensively studied as they play different and important roles influencing the structure and activity of enzymes. These ions may bind to some of the amino acid side chains of the lipase and participate in catalysis, interfere with the bonds between amino acid side chains and cause denaturation of the active site or alter enzyme activity by stabilising or destabilising enzyme conformation [100,106,107,108]. Amongst the different published works some contradictory information can be found. Nevertheless, there are some metal ions that have been clearly described to enhance or worsen ROL and other lipases performance. Wang et al. [68] and other authors [97,100,104] found that Ca²⁺ increases ROL activity as it might create electrostatic interactions that mask the repulsions either between the enzyme and its emulsified substrate or between the enzyme and product-free fatty acids [93]. On the other hand, Hg²⁺ has been reported to act like a ROL activity inhibitor suggesting that thiol groups are required for the adequate function of the enzyme [74]. Similar results have been reported with other lipases from Pseudomonas aeruginosa AAU2 [109], Galactomyces geotrichum Y05 [110], Yarrowia lipolytica [111] and Candida rugosa [112]. In addition, no significant effects have been observed with the chelating agent EDTA, indicating that ROL activity is independent of any metal, hence, it is not a metalloprotein [74,104].

ROL activity has also been analysed in presence of amino-acid-modifying agents in order to elucidate the relevance of the different amino acids in protein catalytic performance. N-Bromosuccinimide (NBS), which acts over tryptophan residues, have been reported to strongly inhibit enzyme activity indicating that the protein might have a tryptophan residue involved in its activity [74,97]. In the case of phenylmethylsulfonyl fluoride (PMSF), a serine protease inhibitor whose activity is related to serine residues modification, no clear results have been reported. Kantak et al. [97] indicated that this agent has a relevant effect while Hiol et al. [75] stated exactly the opposite. However, these differences might be caused by the different disposition of the lipase lid during the assay, that is, if it was open or not, it could allow or not the interaction of PMSF with serine residue of the active site [113].

ROL activity—as most lipases from Rhizopus genus—has a strong 1,3-regiospecificity that makes its activity interesting for several industrial processes such as fat and oil modification for structured lipids production [16,55,75,99]. Nevertheless, Li et al. [114] reported, while studying ROL methanolysis performance, that this lipase was not regiospecific although showed a preference to 1,3-positions. These results were lately confirmed with Rhizopus arrhizus (=oryzae) lipase [115]. However, Okumura et al. and Song et al. [101,116] stated that Rhizopus delemar (=oryzae) and R. oryzae lipases, respectively, do not hydrolyse the ester bond in position 2. Afterwards, Canet et al. and Cao et al. [117,118] proved that mature ROL exhibits a negligible activity on 2-monoolein highlighting that the lipase has a strong 1,3-regioespecificity. The observed dissimilarities amongst different works might be due to the different employed reaction conditions that could enhance spontaneous acyl-migration, or the presence of the 28 amino acids of the prosequence that has already been described to have an effect on lipase specificity [39,63]. Besides 1,3-regiospecificity, substrate specificity of ROL has been also widely studied. Many of the published works are based on the employment of p-nitrophenol esters of different carbon-chain length. For instance, ROL isolated and characterised by Adak et al. [104] was reported to be more specific to long carbon-chain p-nitrophenol esters, concretely to p-nitrophenol palmitate (C16). Guillen et al. [90] reported a similar trend for rROL produced in K. phaffii and, although a higher specificity to short carbon-chain p-nitrophenol esters was detected for proROL, the presence of esterases in the commercial product was concluded to be the reason. In fact, Tako et al. [74] also observed that the longer the carbon-chain, the higher the specificity of ROL. However, in this last case, the maximum was obtained with p-nitrophenol dodecanoate (C12) and not palmitate. ROL substrate specificity has been also analysed with homotriacylglycerols, that is, triacylglycerols in which the three fatty acids are identical. C8 and C10 homotriacylglycerols—triacylglycerols containing three C8 and C10 fatty acids respectively—are preferably hydrolysed by ROL while it barely acts over C2 and C4 homotriacylglycerols. In contrast to some of the published works, some authors have also described that no significant differences were observed with those substrates between rROL and proROL [65,73,75,90].

Lipases are widely known for their capacity to carry out synthesis reactions in non-aqueous mediums. In fact, as previously mentioned, this capacity makes them relevant for many industrial processes in which these reactions are needed, or the solubility of substrates/products requires the use of organic solvents. Therefore, the higher the lipase stability in these solvents, the more suitable the lipase for industrial application [43]. ROL has been extensively described as a tolerant enzyme to non-aqueous solvents [68,75,94], particularly in alkanes and long-chain alcohols such as hexane and dodecanol respectively. However, polar solvents like acetone or short-chain alcohols have an important negative effect on the enzyme because they strip off the crucial bound water from the enzyme’s surface [119]. In some cases, it is remarkable the different results that can be obtained between the stability of the enzyme in a solvent, such as methanol and ethanol, and the operational stability employing that solvents as substrate. For instance, methanol has proven to be more detrimental than ethanol during biodiesel synthesis while during stability assays exactly the opposite result was obtained [39].

3. Rhizopus oryzae Lipase Production and Bioprocess Engineering

First attempts of ROL production were made with the original fungi isolated from palm fruit [75,94,120]. R. oryzae secretes, as previously mentioned, one form of lipase with a molecular weight close to 32 kDa—the mature sequence including 28 amino acids of its prosequence. However, a second form of ROL with a molecular weight around 29 kDa was detected after keeping the supernatant at 6 °C for few days; i.e., the lipase form corresponding to the loss of the 28 amino acids [58]. Consequently, the distinct lipases derived from R. oryzae described in the literature are originated because of the different proteolytic processing and not because of the presence of different genes [56].

To increase ROL industrial production, its expression in a cell factory is mandatory. This way, production cost, bioprocess engineering and downstream complexity are minimised [16].

In Escherichia coli, the presence of disulphide bonds in ROL structure and the lack of the necessary enzymes to process fungal maturation signals were the main causes that led to the production of enzymatically inactive protein as insoluble aggregates [56]. Thereafter, active lipase was obtained at lab scale by subjecting these aggregates to a refolding process. However, the large-scale production was not implemented because of the high cost of the procedure [63]. Despite that, Di Lorenzo et al. [121], achieved the production of an active and soluble ROL and proROL using the E. coli Origami (DE3) strain and pET-11d expression system. The final specific activities of both enzymes were quite similar but the yield of proROL production was higher than ROL, likely because of the toxic effect of the latter towards the host cells.

To avoid the inherent problems of prokaryotic cell factories producing eukaryotic proteins, particularly those related to post-translational processing, eukaryotic cell factories were tested for ROL production.

The extracellular production of ROL has been studied in S. cerevisiae and K. phaffii (P. pastoris) by expressing essentially three different genes. A gene encoding the prosequence of 97 amino acids fused to the N-terminal of the mature lipase region of 269 amino acids (proROL-gene), a gene encoding a truncated prosequence of its 28 C-terminal amino acids fused to the N-terminal of the mature lipase region (28proROL-gene) and a gene encoding the mature lipase (rROL-gene). Regardless of proROL-gene or 28proROL-gene expression, a protein with only 28 amino acids of the prosequence plus the mature lipase (proROL) was detected. Exceptionally, the complete prosequence plus the mature lipase region (entire-proROL) was also reported with proROL-gene construction. With respect to the rROL-gene construction, jus the mature lipase (rROL) was obtained.

First attempts of producing ROL in eukaryotic platforms were made with the widely used cell factory S. cerevisiae. Takahasi et al. [65] reported that S. cerevisiae secreted two active lipases when it was transformed with the proROL-gene fused to the pre-α-factor, the entire-proROL and proROL—the lipase formed after Kex2-like protease cleavage of the prosequence. In parallel, when S. cerevisiae strains were transformed with rROL-gene fused to the pre-α or prepro-α factor encoding gene, almost no activity was detected, highlighting the mentioned relevance of ROL prosequence during lipase production [66,69,71].

A summary of the results obtained with these cell factories is shown in

Table 3. Summary of E. coli and S. cerevisiae cell factories expressing Rhizopus oryzae lipase.

Cell Factory	Promotor/Vector	Lipase	Production	Lipolytic Activity	Reference
E. coli Origami DE3	pET11	proROL	Intracellular	166 U mL−1	[121]
	pET22	proROL	Intracellular	82 U mL−1
S. cerevisiae	UPR-ICL	rROL	Extracellular	0.29 U flask−1	[65]
	UPR-ICL	proROL	Extracellular	191 U flask−1

.

3.1. Komogatella Phaffii Cell Factory

Unlike the reported results with S. cerevisiae, when proROL-gene was expressed in K. phaffii cell factory, only proROL was detected in the medium, which might indicate that the activity of the Kex2-like protease is higher in this cell factory than in S. cerevisiae [67]. Moreover, rROL-gene was satisfactorily expressed and the corresponding lipase was detected in the supernatant [73].

This appropriate performance on ROL secretion, jointly with the well-known excellent characteristics of K. phaffii, make this yeast the most suitable cell factory for heterologous ROL production [122,123,124,125]. In addition, K. phaffi does not produce endogenous extracellular lipases or esterases [126]. Thus, downstream processes might be easier and cheaper. However, two of the bottlenecks of K. phaffii cell factory are transformed clones screening and selecting the best operational strategy to maximise production. To minimise this problem, the use of microbioreactor devices has been successfully implemented [127].

In the following sections, the heterologous ROL production in K. phaffi and its improvement are treated. However, it must be highlighted that owing to the different lipolytic assay conditions employed amongst the published works, a comparison between the obtained values of ROL activity in U ml⁻¹ is challenging. This value is highly dependent on substrate type, enzymatic reaction conditions as well as the equipment used [128].

3.1.1. rROL P_AOX

Regarding K. phaffii’s promoters, the inducible alcohol oxidase promoter (P_AOX) remains as the most used promoter for the heterologous expression of ROL in this cell factory [16]. Since the first article, in which the functional expression of rROL in Mut⁺ phenotype K. phaffii strain under P_AOX producing 180 U ml⁻¹ was described [73], an important effort to assess the effect of different promoters, phenotypes, gene dosage, co-expression of different chaperones to reduce stress burden and different operational bioprocess strategies has been made.

All the data of production parameters and bioprocess conditions under K. phaffii are summarised in

Table 4. Summary of K. phaffii cell factory expressing Rhizopus oryzae lipase.

Gene	Prom.	Co-expr. Δ Delet.	Phenot.	Operational Mode	Pmax AU mL−1	Μ h−1	qp AU gX−1 h−1	YP/X AU gX−1	Vol. Prod. AU L−1 h−1	Spec. Vol. Prod. AU gX−1 h−1	Ref.
rROL	AOX		Mut+ SC	Fed-batch MeOH added when CO2 decrease	500				5435		[73]
rROL	AOX		Mut+ SC	Fed-batch MeOH added manual control	1334				12,888	268	[128]
rROL	AOX		Mut+ SC	Fed-batch MeOH added manual control 2 g L−1	150	0.036	130	2470	3000	49	[129]
rROL	AOX		Mut+ MC	Fed-batch MeOH added manual control 2 g L−1	175	0.012	71	3578	1894	40	[129]
rROL	AOX		MutS His+ SC	Fed-batch MeOH added manual control 2 g L−1	205	0.005	83	5775	2246	63	[129]
rROL	AOX		MutS His- MC	Fed-batch MeOH added manual control 2 g L−1	270	0.004	59	7500	1500	46	[129]
rROL	AOX		Mut+ SC	Fed-batch MeOH controlled 3 g L−1 MNLFB + MLFB to maintain 25% DO	644	0.03	277	7800	8110	98	[130]
rROL	AOX		Mut+ SC	Fed-batch Temperature limited	534	0.13	130	5830	5200	57	[130]
rROL	AOX		Mut+ SC	Fed-batch MeOH controlled 3 g L−1 MNLFB + temperature limited	713	0.02	161	6960	5980	58	[130]
rROL	AOX		Mut+ SC	Fed-batch MeOH controlled 3 g L−1 MNLFB	280	0.046	322	5282	5406	102	[131]
rROL	AOX		Mut+ SC	Fed-batch MNLFB µ = 0.015 h−1	135	0.014	46	2644	1857	36	[131]
rROL	AOX		Mut+ SC	Fed-batch MeOH controlled 3 g L−1 MNLFB DO 25%	368	0.034	256	7700	5490		[132]
rROL	AOX		Mut+ SC	Continuous MeOH controlled 3 g L−1		0.02	0.91	45.7	1.6		[133]
rROL	AOX		Mut+ SC	Continuous MeOH controlled 3 g L−1		0.05	0.98	19.6	2.7		[133]
rROL	AOX		MutS SC	Fed-batch MeOH controlled 1 g L−1	490	0.004	148	11,236	4901	112	[134]
rROL	AOX		MutS SC	Fed-batch MeOH controlled 2 g L−1 Sorbitol exponential feeding rate 0.02 h−1	488	0.02	326	10,369	6346	135	[135]
rROL	AOX		MutS SC	Fed-batch MeOH controlled 2 g L−1 Glycerol exponential feeding rate 0.02 h−1	471	0.025	149	6373	5416	73	[136]
rROL	FLD1		Mut+ SC	Fed-batch Manual SNLFB 8 g L−1	385	0.02	244	7634	4379	87	[137]
rROL	FLD1	GAP HAC1 S. cerevisiae	Mut+ SC	Fed-batch MLFB µ = 0.005 h−1	73	0.005	52	3630	1147	57	[138]
rROL	FLD1	Δ GAS1	Mut+ SC	Fed-batch MLFB µ = 0.005 h−1	240	0.006	145	7172	2790	82	[138]
rROL	FLD1	Δ GAS1 GAP HAC1 S. cerevisiae	Mut+ SC	Fed-batch MLFB µ = 0.005 h−1	206	0.005	206	13,186	3274	144	[138]
proROL	AOX		Mut+ SC	Fed-batch Methanol controlled at 0.1 v/v	12,019						[139]
proROL	AOX		Mut+ MC	Fed-batch Methanol controlled at 3 g L−1	140			4375	1521		[67]
rROL	AOX		Mut+ MC	Fed-batch Methanol controlled at 3 g L−1	195			5417	2130		[67]
proROL	AOX		Mut+ MC	Fed-batch Methanol-sorbitol to maintain DO 20–50%	20,500			175,213	178,261	1524	[140]
proROL	AOX	Ubc1	Mut+ MC	Fed-batch Methanol-sorbitol to maintain DO 20–50%	28,600			223,883	223,833	1865	[140]
proROL	AOX	Hrd1	Mut+ MC	Fed-batch Methanol-sorbitol to maintain DO 20–50%	29,600			229,922	233,543	1810	[140]
proROL	AOX	Ubc1 + Hrd1	Mut+ MC	Fed-batch Methanol-sorbitol to maintain DO 20–50%	33,900			280,165	266,929	2206	[140]
proROL	AOX	Ssa4 + Sso2 + Bmh2	Mut+ MC	Fed-batch Methanol-sorbitol to maintain DO 20–50%	36,578			302,000	290,000	2206	[141]
proROL	AOX	Ssa4 + Sso2 + Bmh2 + Vhb9	Mut+ MC	Fed-batch Methanol-sorbitol to maintain DO 20–50%	41,700			345,000	331,000	2206	[141]
proROL	AOX		Mut+	Fed-batch Methanol to maintain DO over 20%	21,000			107,000	125,000		[68]
proROL	GAP		Mut+ SC	Fed-batch Constant feeding rate to maintain DO 20–40%	11				131	1	[62]
proROL	GAP	Glycosylation Mutant A + B	Mut+ SC	Fed-batch Constant feeding rate to maintain DO 20–40%	2600				30,952	221	[62]
28proROL	AOX		Mut+ SC	Fed-batch MNLFB µ = 0.015 h−1	219	0.011	57	5264	2763	49	[39]
28proROL	AOX		Mut+ SC	Fed-batch MNLFB µ = 0.045 h−1	147	0.038	69	1908	2782	44	[39]
28proROL	AOX		Mut+ SC	Fed-batch MNLFB 3 g L−1	358	0.065	308	4972	7160	99	[39]

.

A comparative study of the effect of ROL gene dosage and K. phaffii strain phenotype demonstrated that, for a single rROL-gene copy strain, the highest production and specific volumetric productivity were reached with the methanol slow consumption phenotype (Mut^S). Despite that, the maximum volumetric productivity was reached with methanol plus consumption phenotype (Mut⁺), because of the greater specific growth rate associated to this phenotype in contrast to Mut^S. Additionally, regarding gene dosage, it was described that the higher the amount of gene copies transformed into the K. phaffii strain genome, the higher the lipase production. However, this increase showed a negative effect on the specific production rate and productivities, particularly for Mut^S phenotype strain. This result was probably caused by the great amount of rROL-gene copies contained in Mut^S strain, 23–24, in comparison to Mut⁺ (5–6 copies) [129]. Moreover, this study was carried out without a robust methanol concentration control, and it is well-known that the maintenance of a constant optimal methanol concentration during the fed-batch (methanol non-limited fed-batch, MNLFB) is necessary to get not only the maximal production but also a reproducible bioprocess.

In this sense, the implementation of a robust methanol on-line control strategy to ensure a constant methanol concentration (MNLFB) in the bioreactor [134] allowed a substantial increase of lipase production, specific production rates and productivities in Mut⁺ phenotype strain. In fact, a methanol set-point concentration of 3 g L⁻¹ was found to be the optimum for rROL production, specific production rate, product-biomass yield and volumetric productivity [131]. These values were almost twice the results obtained in previous experiments with manual methanol control [129]. Additionally, when a methanol limiting fed-batch (MLFB) strategy was applied, that is, a pre-programed exponential feeding rate of methanol to maintain a constant specific growth rate, all production parameters were dramatically lower than with MNLFB strategy. These results suggested a possible negative effect of rROL production on K. phaffii depending on the operational strategy selected [131].

Dissolved oxygen (DO) set-point has been described as a significant parameter during rROL production. A DO set-point around 25% was found to be the best operational condition to achieve the highest specific growth, methanol consumption, oxygen uptake, CO₂ production and rROL-specific production rates. However, the highest product-biomass yield (Y_P/X) was achieved with a lower DO value (10%) [132].

Alternative operational strategies have been assessed to avoid oxygen limitation issues in high cell densities fed-batch cultures, for instance, the combination of MNLFB with methanol or temperature-limited fed-batch strategies. In this line, Surribas et al. [130] reported higher rROL production and Y_P/X, although lower specific production rate.

Moreover, despite oxygen limitation problems, the feeding strategy, as mentioned before, is a key factor in rROL production in K. phaffii. Therefore, in order to improve the substrate feeding strategy, a mathematical model was applied to select the optimal methanol feeding rate achieving significant improvements in final titre and productivity—2.2- and 3.4-fold higher respectively—in comparison to the best standard pre-programmed exponential feeding strategy (MLFB). Nevertheless, the obtained values were quite similar to the most promising control algorithm of the optimal methanol concentration strategy (MNLFB, 3 g L⁻¹) [142].

In the case of Mut^S phenotype strain, when the MNLFB strategy was applied, the greatest ROL production, Y_P/X and productivities were obtained with a constant methanol concentration of 1 g L⁻¹, while the optimal value for Mut⁺ phenotype strain was 3 g L⁻¹. In these conditions, although the lipase production and Y_P/X were around two-fold higher than with Mut⁺ phenotype strain, the specific production rate was 2-fold lower, because of the lower specific growth rate (0.013 h⁻¹) of Mut^S compared to Mut⁺ (0.046 h⁻¹). Therefore, specific and volumetric productivities of both phenotypes were quite similar [134]. Considering these results, in order to select the best methanol consumption phenotype, not only the production parameters, but also the operational implementation feasibility should be taken into account. In consequence, as methanol consumption, oxygen demand and heat generation are lower in Mut^S than in Mut⁺, the former has a greater implementation potential from the bioprocess point of view.

In order to improve rROL production and minimise the proteolysis in Mut^S phenotype strain cultures, different operational strategies have been studied. Li et al. [143] described that maintaining a constant methanol concentration during the fed-batch cultivation showed a better lipase production than adjusting methanol feeding rate to keep DO over 35% (DO-stat). However, the former strategy needs to be complemented to avoid proteases expression and lipase degradation. Besides, modification of medium composition, pre-induction, osmotic stress and temperature have been also described as an important factor in rROL production.

The productivity of Mut^S phenotype strain cultures can be improved by using co-substrates jointly with the methanol during the bioprocess. In this sense, because of the repression of P_AOX by glycerol and glucose, sorbitol was chosen as the most suitable co-substrate—P_AOX repression by sorbitol happened only at concentrations over 40 g L⁻¹. Interestingly, the sorbitol assimilation profile was different depending on the selected Mut phenotype. In Mut^S, both substrates were consumed simultaneously, getting not only higher specific growth rates, but also higher secretion levels. However, this effect was not observed in Mut⁺ phenotype strains in which substrates consumption was sequential [144].

When a systematic study comparing fed-batch cultures at different constant methanol concentration set-points and different specific growth rates of sorbitol was made with Mut^S, the optimal conditions were at methanol set-point of 2 g L⁻¹ and sorbitol specific growth rate of 0.01 h⁻¹. Under these conditions, all production parameters were increased in comparison to the best results obtained using methanol as the sole carbon source at a methanol set-point of 1 g L⁻¹. Interestingly, the use of co-substrate strategy overcame partially the physiological bottlenecks detected at different levels in the rROL production bioprocess. According to Arnau et al. [135] the addition of sorbitol during the methanol induction phase avoided the severe decrease of the specific production rate obtained when methanol was used as a sole carbon source. However, the drawback of using sorbitol as the co-substrate is its maximum specific growth rate in K. phaffii, which is too low (0.02 h⁻¹), similar to the value growing in methanol for Mut^S strain. Thus, the productivity could be affected because of the long-time bioprocess. Consequently, to overcome this problem, glycerol was also tested as co-substrate. This substrate has a µ_MAX around 0.2 h⁻¹, 10-fold higher than K. phaffii growing in sorbitol. However, although glycerol was fed under limiting conditions, the productivity of the bioprocess decreased dramatically when the specific growth rate of glycerol was higher than 0.02 h⁻¹, highlighting the repressive effect of this compound on P_AOX [136]. Besides, the effect of dilution rate on chemostat cultures with a Mut⁺ phenotype strain, using glycerol as the co-substrate, revealed that the optimal dilution rate was between 0.04 and 0.06 h⁻¹ in terms of maximum productivity. When these results were compared to the optimal dilution rate of rROL production (0.02 h⁻¹) using methanol as the sole carbon source, the productivity was close to two-fold higher, but Y_P/X followed the opposite trend. Thus, in terms of productivity, the mixed feeding strategy (glycerol-methanol) might be potentially applied in fed-batch cultures with promising results opening new lines of research for the future [133].

3.1.2. rROL P_FLD1

rROL-gene has been also expressed under the inducible formaldehyde dehydrogenase 1 promoter (P_FLD1), which has been reported as a strong and independent promoter induced by both methanol as sole carbon source and methylamine as the sole nitrogen source. Besides, P_FLD1 has been described to be strongly repressed by an excess of either glycerol or glucose but not sorbitol. In fact, even though the combination of methanol-methylamine showed a synergetic effect on ROL production levels under this promotor, the combination of sorbitol as sole carbon source and methylamine as sole nitrogen source enabled, for the very first time, ROL production in a methanol-free high cell density fed-batch culture. This way, the use of methanol is avoided and substituted by methylamine, which is a less flammable and volatile P_FLD1 inducer than the former [145,146].

Different operational fed-batch strategies were tested to assess the performance of P_FLD1. The best outcome was obtained by maintaining a constant concentration of 8 g L⁻¹ of sorbitol by feeding the culture media with a stock solution of sorbitol and methylamine—stoichiometric ratio of 0.118 g methylamine per g of sorbitol [137]. Under these conditions, the production was 1.37-fold higher than the best result reported with Mut⁺ phenotype strain by Barrigon et al. [131], although 0.58-fold lower than the best outcome described for Mut^S phenotype by Arnau et al. [135], both under P_AOX. In any case, P_FLD1 was successfully employed for rROL production under methanol-free high cell density fed-batch strategy [137].

Transcriptomic studies using P_FLD1 revealed that rROL overexpression and secretion seem to trigger the unfolding protein response (UPR) in K. phaffii, resulting in a physiological bottleneck for the bioprocess production. The study pointed out a complex response to both dynamic environmental conditions (bioprocess) and endogenous stress factors (omics) [147]. In this sense, the co-expression of rROL under inducible P_FLD1 and the S. cerevisiae UPR transcriptional factor (HAC1) under constitutive P_GAP (ROL_HAC1) showed—working at the same specific growth rate (0.005 h⁻¹)—two-fold higher values of lipase production and volumetric productivity; and close to three-fold higher values on Y_P/X and specific productivity. In parallel, this co-expression system reduced intracellular binding proteins (BiP) levels which has been related to UPR triggering [138]. On the other hand, in order to improve the possible secretion bottlenecks of rROL, a new rROL-producing strain was constructed with deletion of the GAS1 gene (ROL_ΔGAS1), a β 1–3-glucanosytransglycosylase which plays a key role in yeast cell assembly. The rROL production parameters of this strain were clearly increased although a substantial reduction of µ_MAX was observed [148]. Finally, the double mutant ROL_HAC1-ROL_ΔGAS1 was tested and the synergetic effect of these engineering strategies was confirmed [138]. However, the specific production rate and intracellular rROL levels reflected a complex interaction of these modifications with physiological aspects, including energy metabolism and growth conditions that require further research. Nevertheless, considering all these results, transcriptomics studies have provided new insights into microorganisms’ modification in general and K. phaffii in particular [149]. Consequently, further research in this field will allow a better understanding of all the bottlenecks present in the heterologous production of proteins.

3.1.3. proROL P_AOX

Wang et al. [68] reported proROL production under P_AOX in Mut⁺ phenotype strain. Fifty-litre bioreactors were employed for high cell density fed-batch cultures with methanol, whose concentration was kept stable by monitoring the dissolved oxygen (DO) content and maintaining it over 20%. A maximum activity of 21,000 U mL⁻¹ was obtained with a final biomass concentration of 197 g L⁻¹ DCW.

Nitrogen concentration (NH₄⁺) has been described as a relevant factor during proROL production in high cell density fed-batch cultures with Mut⁺ phenotype strain [139]. The optimum NH₄⁺ concentration to maximise lipase production (12,019 U mL⁻¹) was found to be 440 mmol L⁻¹. Furthermore, this concentration of NH₄⁺ prevented protease production and has a significant impact on the carbon metabolism and energy regeneration.

Interestingly, with proROL-producing strain, as it was observed with rROL, the recombinant lipase produced with Mut^S phenotype was more active and more stable to changes of induction time and methanol concentration than the lipase produced with Mut⁺ strain [100].

The comparison of strains harbouring multicopy of proROL and rROL-genes has also been studied in fed-batch cultures under P_AOX and Mut⁺ phenotype with methanol concentration control at 3 g L⁻¹. As it was previously commented, the truncated lipase proROL, with only the 28 C-terminal amino acids of the prosequence fused to the mature sequence was detected in the medium with proROL-gene harbouring strain. The amounts of enzyme secreted were quite similar, 91 mg L⁻¹ for proROL and 80 mg L⁻¹ for rROL. The characterisation of both lipases showed the importance of the prosequence in terms of thermostability and relative activity towards triacylglycerols [67].

Jiao et al. [140], proved that proROL activities were enhanced along with the increase of proROL-gene copy number up to five copies. At higher gene copy number, lipolytic activity decreased. Authors also tried to alleviate cellular stress via ER-associated protein degradation (ERAD) expression. In this sense, two ERAD-related genes Ubc1 and Hrd1 were co-expressed together with the proROL-gene in high cell densities fed-batch cultures. Sorbitol was employed as a co-substrate jointly with methanol and DO was maintained between 25 and 50%. The production reached with the double mutant Ubc1-Hrd1 was higher than the strains co-expressing only one of the ERAD-relates genes. Therefore, the double mutant showed a synergetic effect on proROL production [140].

Furthermore, the co-expression of molecular chaperones ERO1p and PDI has been described to increase the production of proROL [150]. However, no effect was observed by increasing Pdi gene copy number. In parallel, the effect on proROL production of 10 helper factors was studied. Between them Ssa4, Bmh2 and Sso2—all participating in protein secretion—and Vhb—to alleviate oxygen intake problems—were tested for proROL expression enhancement. Results confirmed that proROL production was improved in fed-batch bioprocess with a strain containing 5 copies of proROL and the co-expression of these protein helpers [141].

3.1.4. rROL and proROL P_GAP

rROL and proROL production under the constitutive promoter GAP has not been extensively studied. In fact, in our lab, the expression of rROL under P_GAP has not been successful. However, Yu et al. [62] reported significant proROL production levels after engineering the N-glycosylation pattern of the prosequence. In fact, the production of the modified proROL was 218-fold higher than the parent proROL, highlighting the rational design of N-glycosylation sites as an efficient way to enhance proROL secretion under P_GAP.

3.1.5. 28proROL P_AOX

Recently a new gen construction (28proROL-gene) consisting in the 28 C-terminal amino acids of the prosequence fused to the N-terminal of mature sequence of ROL has been expressed under P_AOX. In batch cultures, growth kinetics, production and yields were higher for proROL-producing strain than rROL. In addition, under similar fed-batch operational strategies, lipase production and volumetric productivity were always higher in 28proROL-gene harbouring strain than rROL-gene, suggesting a stress reduction during lipase expression caused by the presence of the 28 amino acids of the prosequence. Consequently, UPR phenomena might not be triggered to the same extent as in rROL production. After fed-batch culture downstream, the obtained free proROL showed a higher stability than rROL, making it more appropriate as an industrial biocatalyst [39].

The 28proROL-gene has been expressed under P_GAP at similar production levels than under P_AOX in Valero’s group (data not published). These results are relevant in order to carry out friendlier bioprocess engineering avoiding the use of methanol.

3.1.6. Whole Cells

Hama et al. reported that rROL and proROL are located in different regions in R. oryzae cells, proROL in the cell wall and rROL bound to the cell membrane. Besides, these cells have been successfully employed as whole cells biocatalysts (WCB) in many relevant biotransformations, for instance, for enzymatic biodiesel production [151,152]. It must be highlighted that the fatty acid composition of the membrane has been reported to influence lipase activity and stability during biodiesel reactions [153].

Modified S. cerevisiae strains producing ROL have also been used as WCB [154,155]. Matsumoto et al. [154] reported the intracellular production of proROL in S. cerevisiae under the 5′ upstream region of the isocitrate lyase gene of Candida tropicalis (UPR-ICL). Additionally, the expression of the lipase under the constitutive promoter glyceraldehyde-3-phosphate dehydrogenase was also studied. However, this system did not improve the results obtained with UPR-ICL.

proROL was successfully expressed under P_AOX control and displayed on Mut⁺ phenotype K. phaffii cell surface using the Flo1P anchor system previously developed in S. cerevisiae. The obtained WCB showed higher thermal stability than free enzyme [156]. Additionally, a similar approach using Sed1p anchor protein was studied in a Mut^S phenotype. In the same sense, the obtained biocatalysts was stable in a wide range of temperatures and pH [157].

4. Industrial Applications of Rhizopus oryzae Lipase

The 1,3-regiospecificity and catalytic versatility make ROL appropriate for improving the sustainability of food, pharmaceutical and energy industry [16,21,158].

4.1. Biodiesel Production

Because of petroleum depletion and environmental concerns, in the past decade, biodiesel (mono-alkyl esters of long chain fatty acids) is gathering significant interest as a renewable, biodegradable and more environmentally friendly alternative to fossil fuels. Biodiesel can be classified into three different generations based on the source from which it is derived. First-generation biodiesel is synthesised with edible-oils such as soybean or sunflower oils. Therefore, it might cause the so-called “food vs. fuel” ethical issue because of the use of food and agricultural lands for biofuel production [159]. In order to prevent this problem, alternative substrates have emerged for biodiesel synthesis, leading to second- and third-generation biodiesel production. The former uses non-edible oils that are not considered for human consumption and are produced from crops that, even if they require lands, are generally from poor lands not useful for agriculture. Meanwhile, third-generation biodiesel completely avoids ethical issues by using microbial lipids and oleaginous wastes such as oils from microalgae or oleaginous yeasts and waste cooking oils (WCO) respectively [160,161,162,163]. Additionally, there is a fourth-generation biodiesel that is at its preliminary research stages and is based on man-made biological tools, that is, biodiesel producing genetically modified microorganisms [164,165].

Typically, these alternative substrates, those yielding second- and third-generation biodiesel, have a higher free fatty acid (FFA) content, which can make biodiesel production through chemical synthesis—the most common process for current industrial biodiesel production—more complex because previous operation of FFA neutralisation is required to avoid soap formation, an usual side reaction when substrates with high FFA content and basic catalysts are employed [161,166,167]. In this context, enzymatic biodiesel synthesis with lipases arises as an alternative owing to its several advantages such as the milder reaction conditions, less water consumption, easier downstream and particularly, the absence of side reactions and consequently the capacity of employing substrates with high FFA content [168,169]. In fact, substrates with initial high amounts of FFA have been reported to enhance enzymatic biodiesel synthesis reaction rate and biocatalysts operational stability [168,170,171]. Given all the advantages, numerous lipases have been studied in this biotransformation with significant results, such as the lipases of Candida rugosa [172,173], Candida antartica [174,175] and Burkholderia cepacia [176,177].

In the seeking for the best lipase to make enzymatic biodiesel feasible at industrial scale, lipases’ regiospecificity has become a crucial trait. Non-specific enzymes produce mono-alkyl esters and glycerol, which is an undesired by-product of the transesterification reaction that has been described to hinder reaction progress or even affect negatively on enzymes stability and biodiesel downstream [178]. Conversely, 1,3-regioespecific lipases, avoid glycerol formation by producing 2-monoacylglycerol which acts as a lubricant and in certain amount, upgrades biodiesel characteristics [179,180,181]. Furthermore, monoacylglycerols can improve the cost-effectiveness of a biodiesel biorefinery as they are more valuable products than glycerol because of their utility in pharmaceutical and food industry as emulsifiers [182,183,184]. Consequently, ROL has been widely studied in biodiesel production because of its regiospecificity.

Considering biodiesel ethical issues, even if several studies have employed ROL with edible oils such as olive [185], rapeseed [186,187], soybean [188,189,190,191] and sunflower [192,193] oils—commonly as model substrates for research—most of the published works have focused on the use of alternative substrates (

Table 5. Summary of biodiesel production with Rhizopus oryzae lipase as main biocatalyst.

Substrates	Lipase	Immobilisation Technique	Reactor Type	Stepwise Addition	Biodiesel Generation	Yield-Conversion/Operational Stability	Ref.
OO + MeOH	rROL	IA onto ReliZymeTM OD 403M	PBR	Yes	1st	Y: PBR 49.1% OS: second batch 44.8%	[185]
OO + MeOH	rROL	IA onto ReliZymeTM OD 403M	STR	Yes	1st	Y: STR 33.56% OS: second batch 7.7%	[185]
RO + MeOH	proROL	WCB over agar plate	SLLB	No	1st	No biodiesel production	[186]
RO + EtOH	proROL	WCB over agar plate	SLLB	No	1st	No biodiesel production	[186]
RO + MeOH	proROL	WCB over agar plate	SGLB	No	1st	Y: 58%	[186]
RO + EtOH	proROL	WCB over agar plate	SGLB	No	1st	Y: 72%	[186]
Crude CO + MeOH	proROL	Free enzymes	BR	Yes	1st	Y: 68.56%	[187]
Crude CO + MeOH	proROL-CRL	Free enzymes	BR	Yes	1st	Y: 84.25%	[187]
Crude CO + MeOH	proROL-CRL	CI onto functionalised silica gel	BR	Yes	1st	Y: 88.9%	[187]
SYO + MeOH	proROL	WCB immobilised into BSPs	BR	Yes	1st	Y: 82.2% OS: after 6 cycles almost all activity loss	[188]
SYO + MeOH	proROL	CI WCB immobilised onto BSPs	BR	Yes	1st	Y: 92.2% OS: after 6 cycles no loss of activity	[188]
SYO + EtOH	proROL	IA onto microporous resin NKA (polystyrene)	BR	Yes	1st	Y: 58.5%	[189]
SYO + EtOH	proROL-CRL	IA onto microporous resin NKA (polystyrene)	BR	Yes	1st	Y: 80.8%	[189]
SYO + EtOH	proROL-Novozyme 435	proROL: IA onto microporous resin NKA (polystyrene). Novozyme 435: IA onto Lewatit VP OC 1600	BR	Yes	1st	Y: 98.5% OS: after 20 cycles Y decreased to 78.3%	[189]
SYO + EtOH	proROL-PFL	IA onto microporous resin NKA (polystyrene)	BR	Yes	1st	Y: 55.8%	[189]
SYO + MeOH	proROL	CI onto magnetic chitosan microspheres	MSFBR	Yes	1st	Y: 91.3% OS: after 6 reaction cycles Y decreased to around 80%	[190]
SYO + MeoH	proROL	WCB immobilised into BSPs	BR	Yes	1st	Y: over 90% OS: after 10 reaction cycles Y decreased to 10%	[191]
SYO + MeoH	proROL	WCB immobilised into BSPs	PBR	Yes	1st	Y: over 90% OS: after 10 reaction cycles Y decreased to 80%	[191]
SO + EtOH	proROL	CI onto modified sepiolite with p-hydroxybenzaldehyde linker	BR	No	1st	C: 84.3% OS: after 9 cycles C decreased to 21.4%	[192]
SO + EtOH	proROL	CI onto modified sepiolite with benzylamine-terephthalic aldehyde linker	BR	No	1st		[192]
SO + EtOH	proROL	IE onto demineralised sepiolite	BR	No	1st	Y: 90.2% OS: proROL IE after 9 cycles C decreased to 18.1%	[192]
Pistacia chinensis bge seed oil + MeOH	rROL	CI onto Amberlite IRA-93	BR	Yes	2nd	Y: 92% OS: after 8 cycles Y decreased to 60%	[196]
Pistacia chinensis bge seed oil + MeOH	rROL	IA microporous resin HPD-400	BR	Yes	2nd	Y: 94% OS: after 8 cycles Y decreased to 50%	[196]
Calophyllum inophyllum linn oil + MeOH	proROL	WCB immobilised into BSPs	PBR	Yes	2nd	Y: 92% OS: after 6 cycles Y decreased a 4.9%	[197]
Oil extracted from Nannochloropsis gaditana + MeOH	proROL	WCB	BR	Yes	3rd	Y: 83% OS: after 3 cycles Y decreased to 71%	[198]
Oil extracted from Nannochloropsis gaditana + MeOH	proROL	WCB immobilised into BSPs	BR	Yes	3rd	Y: 70% OS: second cycle Y decreased to 43%	[198]
Oil extracted from Nannochloropsis gaditana + MeOH	proROL	WCB immobilised into BSPs	BR	Yes	3rd	Y: 83% OS: after 3 cycles Y decreased to 71%	[199]
Oil extracted from Nannochloropsis gaditana + MeOH	proROL	WCB	TPB	No	3rd	Y: 58%	[200]
Oil extracted from Nannochloropsis gaditana + EtOH	proROL	WCB	TPB	No	3rd	Y: 92%	[200]
Oil extracted from Botryococcus braunii + MeOH	proROL	WCB	TPB	No	3rd	Y: 58%	[200]
Oil extracted from Botryococcus braunii + EtOH	proROL	WCB	TPB	No	3rd	Y: 68%	[200]
Oil extracted from Chlorella vulgaris + MeOH	proROL	Free enzyme	BR	Yes	3rd	C: 75%	[201]
Oil extracted from Chlorella vulgaris + MeOH	proROL	IA onto MNP	BR	Yes	3rd	C: 46% OS: after 5 cycles decreased to 10%	[201]
Oil extracted from Chlorella vulgaris + MeOH	proROL	CI onto AP modified MNP	BR	Yes	3rd	C: 53% OS: after 5 cycles C decreased to 25%	[201]
Oil extracted from Chlorella vulgaris + MeOH	proROL	CI onto AP-GA modified MNP	BR	Yes	3rd	C: 69.8% OS: after 5 cycles C decreased to 45%	[201]
Sludge palm oil + MeOH	proROL	IE into alginate-polyvinyl alcohol beads	BR	No	3rd	Y: 91.30% OS: no activity loss after 15 cycles	[202]
Oil extracted from SCG + MeOH	R. delemar (= oryzae) lipase	Free enzyme	BR	No	3rd	Y: 18%	[203]
WCO + MeOH	proROL	Free enzyme	BR		3rd	Y: 93%	[204]
WCO + iso-propanol	proROL	Free enzyme	BR		3rd	Y: 86.8%	[204]
WCO + iso-butanol	proROL	Free enzyme	BR		3rd	Y: 80.2%	[204]
WCO + iso-amyl alcohol	proROL	Free enzyme	BR		3rd	Y: 64%	[204]
WCO + MeOH	proROL	WCB IE into calcium alginate beads	BR		3rd	Y: 84%	[204]
WCO + iso-propanol	proROL	WCB IE into calcium alginate beads	BR		3rd	Y: 71%	[204]
WCO + iso-butanol	proROL	WCB IE into calcium alginate beads	BR		3rd	Y: 62%	[204]
WCO+ iso-amyl alcohol	proROL	WCB IE into calcium alginate beads	BR		3rd	Y: 43%	[204]
JO + MeOH	proROL	WCB IE into sodium alginate beads	BR	No	2nd	Y: 80.5% OS: after 6 cycles Y decreased to 61.5%	[205]
KO + MeOH	proROL	WCB IE into sodium alginate beads	BR	No	2nd	Y: 78.3% OS: after 6 cycles Y decreased to 63.4%	[205]
SYO + MeOH	proROL	WCB	BR	Yes	1st	Y: 80% OS: after 3 cycles Y decreased to 18%	[206]
SYO + MeOH	proROL	WCB immobilised into BSPs	BR	Yes	1st	Y: 82% OS: after 10 cycles Y decreased to 10%	[206]
SYO + MeOH	proROL	CI WCB immobilised into BSPs	BR	Yes	1st	Y: 74% OS: after 35 cycles Y decreased to 65%	[206]
SYO + MeOH	proROL	WCB immobilised into BSPs	BR	Yes	1st	Y: 82% OS: after 6 cycles Y decreased to 48%	[207]
SYO + MeOH	proROL	CI WCB immobilised into BSPs	BR	Yes	1st	Y: 80% OS: after 6 cycles Y decreased to 70%	[207]
ALO + MeOH	rROL	IA onto rice husk	BR	Yes	2nd		[208]
ALO + MeOH	rROL	IA onto ReliZymeTM OD403	BR	Yes	2nd	Y: 64.5% OS: after 7 cycles Y decreased to 41.3%	[208]
Crude microbial oil from Candida sp. LEB-M3 + MeOH	rROL	IA onto ReliZymeTM OD403	BR	Yes	3rd	Y: 38% OS: after 7 cycles Y decreased to 26.6%	[209]
Neutralised microbial oil from Candida sp. LEB-M3 + MeOH	rROL	IA onto ReliZymeTM OD403	BR	Yes	3rd	Y: 38%	[209]
OO + MeOH	rROL	IA onto ReliZymeTM OD403	BR	Yes	1st	Y: 54.3% OS: after 7 cycles Y decreased to 40%	[209]
OA + MeOH	rROL	IA onto ReliZymeTM OD403	BR	Yes	1st	Y: 68%	[209]
RO + EtOH	proROL	IA onto microporous resin NKA	BR	No	1st	Y: above 98% OS: After 10 cycles Y decreased to 60%	[210]
JO + MeOH	proROL-CRL	WCB (proROL) and free enzyme (CRL) IE into sodium alginate beads	PBR	No	2nd	Y: 84.2%	[211]
KO + MeOH	proROL-CRL	WCB (proROL) and free enzyme (CRL) IE into sodium alginate beads	PBR	No	2nd	Y: 81%	[211]
WCO + MeOH	proROL	WCB IE into sodium alginate beads	BR	No	3rd	Y: 94.01%	[212]
WCO + Methyl acetate	proROL	WCB IE into sodium alginate beads	BR	No	3rd	Y: 91.11%	[212]
WCO + Ethyl acetate	proROL	WCB IE into sodium alginate beads	BR	No	3rd	Y: 90.06	[212]
WCO + MeOH	proROL	IE into sodium alginate beads	BR	No	3rd	Y: 83%	[212]
WCO + Methyl acetate	proROL	IE into sodium alginate beads	BR	No	3rd	Y: 80%	[212]
WCO + Ethyl acetate	proROL	IE into sodium alginate beads	BR	No	3rd	Y: 78%	[212]
Oil extracted from Chlorella vulgaris + MeOH	proROL	IA into MNP	BR	Yes	3rd	Y: 45% OS: after 5 cycles Y decreased to 10%	[213]
Oil extracted from Chlorella vulgaris + MeOH	proROL	IA into MGO	BR	Yes	3rd	Y: 51% OS: after 5 cycles Y decreased to 16%	[213]
Oil extracted from Chlorella vulgaris + MeOH	proROL	IA into MGO-AP	BR	Yes	3rd	Y: 54% OS: after 5 cycles Y decreased to 25%	[213]
Oil extracted from Chlorella vulgaris + MeOH	proROL	CI into MGO-AP-GA	BR	Yes	3rd	Y: 68% OS: after 5 cycles Y decreased to 58.77%	[213]
Cottonseed oil + MeOH	proROL	WCB immobilised into BSPs	BR	Yes	1st	Y: 27.9%	[214]
Rubber seed oil + MeOHe	proROL	Free enzyme	BR	Yes	2nd	Y: 31%	[215]
Rubber seed oil + Ethyl acetate	proROL	Free enzyme	BR	No	2nd	Y: 33.3%
SYO + MeOH	proROL-CRL	CI onto silica gel pretreated with AP and GA	BR	Yes	1st	Y: 99.99% OS: after 20 cycles Y decreased to 85%	[216]
RO deodoriser distillate + MeOH	proROL	Free enzyme	BR	Yes	1st	Y: 93.07%	[217]
RO deodoriser distillate + MeOH	proROL-CRL	Free enzyme	BR	Yes	1st	Y: 98.16%	[217]
ALO + MeOH	rROL	CI onto ET, AP and GA pretreated ReliZymeTM HFA403	BR	Yes	2nd	Y: 57.16% OS: after 5 cycles Y decreased a 12.31%	[181]
ALO + EtOH	rROL	CI onto ET, AP and GA pretreated ReliZymeTM HFA403	BR	Yes	2nd	Y: 60.25% OS: after 7 cycles Y decreased a 11.89%	[181]
Triolein + MeOH	rROL	Free enzyme	BR	No	1st	Y: 71.2%	[117]
Triolein + EtOH	rROL	Free enzyme	BR	No	1st	Y: 64.2%	[117]
Triolein + MeOH	rROL	IA onto RelyZymeTM OD403S	BR	No	1st	Y: 82.6%	[117]
Triolein + EtOH	rROL	IA onto RelyZymeTM OD403S	BR	No	1st	Y:100.7%	[117]
JO + MeOH	rROL	IA onto Lewatit VP OC 1600	BR	Yes	2nd	Y: 61% OS: after 10 cycles Y decreased a 40%	[218]
JO + MeOH	rROL	IA onto LifetechTM ECR1030M	BR	Yes	2nd	Y: 63% OS: after 10 cycles Y decreased a 40%	[218]
JO + MeOH	rROL	IA onto LifetechTM AP1090M	BR	Yes	2nd	Y: 55% OS: after 10 cycles Y decreased a 25%	[218]
JO + MeOH	rROL	CI onto LifetechTM ECR8285M	BR	Yes	2nd	Y: 63% OS: after 10 cycles Y decreased a 60%	[218]
JO + MeOH	rROL	CI onto Amberlita IRA 96	BR	Yes	2nd	Y: 68% OS: after 10 cycles Y decreased a 20%	[218]
OO + MeOH	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 77%	[219]
OO + EtOH	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 62%	[219]
OO + Propanol	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 46%	[219]
OO + Butanol	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 18%	[219]
SYO + MeOH	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 50%	[219]
SYO + EtOH	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 46%	[219]
SYO + Propanol	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 35%	[219]
SYO + Butanol	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 10%	[219]
CO + MeOH	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 70%	[219]
CO + EtOH	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 56%	[219]
CO + Propanol	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 43%	[219]
CO + Butanol	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 16%	[219]
SO + MeOH	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 32%	[219]
SO + EtOH	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 28%	[219]
SO + Propanol	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 17%	[219]
SO + Butanol	prorROL	IA onto Amberlite XAD 761	BR	No	1st	Y: 7%	[219]
Algal oil + MeOH	prorROL	IA onto Amberlite XAD 761	BR	No	3rd	Y: 63%	[219]
Algal oil + EtOH	prorROL	IA onto Amberlite XAD 761	BR	No	3rd	Y: 55%	[219]
Algal oil + Propanol	prorROL	IA onto Amberlite XAD 761	BR	No	3rd	Y: 40%	[219]
Algal oil + Butanol	prorROL	IA onto Amberlite XAD 761	BR	No	3rd	Y: 13%	[219]
ALO + MeOH	rROL	CI onto AP and GA treated ReliZymeTM HFA403	BR	Yes	2nd	Y: 28.62% OS: after 9 cycles, Y decreased a 43%	[170]
JO + MeOH	proROL	WCB immobilised into BSPs	BR	Yes	2nd	Y: 88.6% OS: after 6 cycles Y decreased a 21%	[220]
OA + MeOH	proROL	WCB immobilised into BSPs	BR	No	1st	Y: 80% OS: after 8 cycles, almost no activity loss.	[221]
Rice bran oil + MeOH	proROL	IA onto rod-like mesoporous silica	BR	No	1st	Y: 81.7% OS: after 3 cycles Y decreased to 67.7%	[222]
JO + MeOH	proROL	IE into polyvinyl alcohol—alginate matrix	BR	No	2nd	Yield: 87.1%	[223]
ALO + MeOH	rROL	IA Octadecyl-Sepabeads	BR	Yes	2nd	Y: 58.31% OS: after 2 cycles Y decreased to 54.67%	[224]
Tung oil + MeOH	proROL	CI onto Amberlite IRA 93	BR	Yes	2nd	Y: 91.9% OS: after 6 cycles Y decreased to 85.1%	[95]
Babassu oil + EtOH	proROL	WCB immobilised into BSPs	BR	No	1st	Y: 74.15%	[225]

). Jatropha curcas oil is one of the non-edible oils with higher potential for second-generation biodiesel production, probably because of the easy cultivation process and worldwide spread of the plant [194]. Rodrigues et al. [195] reported yields close to the theoretical 100%—real 66% considering ROL 1,3-regioespecificity—and high operational stability of the biocatalysts. In Table 5 are detailed other studies with promising results using this substrate as well as other non-edible oils like Pistacia chinensis bge oil [196], Tung oil [95], Calophyllum inophyllum oil [197] and alperujo oil (olive pomace) [181].

Regarding the third-generation biodiesel, microalgae and waste oils have been the most studied substrates. The former has several advantages that make the overall process of biodiesel production more environmentally friendly as microalgae oil production involves atmospheric CO₂ fixation and can use domestic wastewater like growth substrate facilitating its posterior treatment. However, the main drawbacks for microalgae oil employment are the scale-up of photobioreactors and lipids extraction [226,227]. Nevertheless, ROL has been satisfactorily employed with this substrate, for instance, with oils extracted from Nannochloropsis gaditana [198,199,200], Botryococcus braunii [200] and Chlorella vulgaris [201]. Actually, with the last one, fatty acid methyl esters (FAME) conversions over 70% were obtained indicating ROL suitability for biodiesel production with microalgae oil. Additionally, oils extracted from oleaginous yeasts, such as Candida sp. LEB-M3, have been also employed. The use of yeasts becomes important in biodiesel refineries as they might grow in the glycerol coming from this biofuel production [209]. Regarding waste oils, they have a significant potential in biodiesel industry because of their relevance in circular economy strategies which aim to avoid residue generation by seeking new applications to waste [228,229]. Moreover, considering the tight economic competition between biodiesel and fossil fuels, cheap raw materials are required. In fact, the cost of the feedstocks is more than the 70% of the total cost of biodiesel. Thus, oleaginous wastes might help lowering these percentage and making enzymatic biodiesel production feasible [230]. Sludge from palm oil [202] and spent coffee grounds [203] can be found amongst some of the oleaginous residues studied in biodiesel production with ROL. However, waste cooking oil is the foremost substrate of this category because it is inexpensive and, through its employment in biodiesel synthesis, public institutions avoid the great cost of its management [231,232]. Relevant results have been published with WCO, for instance, Bharathiraja et al. [204] reported maximum triglyceride conversion of 94%. Nevertheless, not many studies dealing with ROL and WCO have been published and, considering the great relevance of this substrate for biodiesel industry, it could be a possible research target for future projects.

Biocatalysts operational stability, reusability and price are related and essential traits that must be considered in enzymatic biodiesel production because of the high cost of enzymes and the tight economic competition with conventional diesel. Some approaches have focused on cutting prices of the enzymes through heterologous production, as it has been explained in the previous section. Besides, other strategies have centred on lipase immobilisation. This technique allows enzyme reutilisation and generally enhances enzyme stability [6,233,234]. In the following paragraphs, the different immobilisation strategies assessed with ROL in biodiesel production will be introduced.

Earlier attempts of employing this enzyme in biodiesel synthesis were principally based on whole-cell biocatalysts (WCBs). Thus, the enzyme acts confined in its natural cellular environment, which protects the lipase from inactivation and degradation. Moreover, as no downstream processes of the biocatalyst are needed, its final cost is considerably lowered [235]. Syed et al. [205] immobilised lipase-producing R. oryzae cells into alginate beads and employed them in biodiesel production with jatropha and karanja oil. A response surface optimisation was applied and under the best conditions, biodiesel yields of 73.5% and 72.5% with each respective oil were obtained. In addition, operational stability of the biocatalyst was evaluated and after six cycles, just an activity loss of 20% was reported. Even if free cells, without immobilisation into alginate beads, could have been used in biodiesel production, Sun et al. [206] stated the suitability of cell immobilisation to avoid enzyme leakage and denaturation. This author immobilised R. oryzae fungus cells onto biomass support particles (BSPs) and obtained higher operational stability than using free cells. Moreover, to further minimise the enzyme leakage and deactivation, the crosslinking agent glutaraldehyde was used for immobilised cells treatment. The crosslinked biocatalyst obtained better FAMEs yields and operational stability. In the same sense, glutaraldehyde treatment of WCBs—also called WCBs stabilisation—was reported by Ban et al. [207] as well. Lately, He et al. [188] employed this strategy too and obtained a ROL biocatalyst with increased operational stability. After six reactions cycles, more than 90% of initial activity was maintained. However, WCBs show higher complexity in being reused and worse conversion rates than free lipases immobilised onto acrylic resins [235]. For instance, Bharathiraja et al. [204] published that WCBs exhibit worse reaction rate than immobilised purified proROL because of diffusional problems. Therefore, considering these inconveniences and how heterologous production of ROL has been improved, the use of free ROL and its subsequent immobilisation have gained importance amongst the published works.

Traditionally, lipases have been immobilised through adsorption, particularly onto hydrophobic supports—generally acrylic resins with hydrophobic superficial groups such as octadecyl or divinylbenzene—because of the presence of a large hydrophobic patch around the catalytic triad of the lipases that enables an easy immobilisation and might lead to their hyperactivation [236,237]. However, during biodiesel enzymatic synthesis, highly non-polar reaction mediums are employed that might cause enzyme desorption and in consequence, poor biocatalyst operational stability [84]. Nevertheless, some authors have used ROL with this immobilisation technique and obtained outstanding stability results. For instance, Bonet-Ragel et al. [208] reported that after six consecutive reaction cycles, the biocatalysts retained more than the 60% of the initial activity, in accordance with the results published by Duarte et al. [209] and Su et al. [210]. Moreover, in order to overcome the potential enzyme leakage when adsorption techniques are used, some published works have treated the obtained biocatalysts with crosslinking agents like glutaraldehyde, as it was previously explained for WCBs [196,238]. Notwithstanding these mentioned works and other listed in Table 5, ROL entrapment and covalent immobilisation are the most common immobilisation techniques. The former has been used not only with free ROL but with WCBs because it is an easy, fast and cheap immobilisation technique [239]. The most common entrapment strategies are based on polyvinyl alcohol and alginate employment [211,212,223,240]. Muanruksa et al. [202] obtained outstanding results with free proROL immobilised into alginate-polyvinyl alcohol beads. Esterification degrees over 90% were reported and the biocatalyst exhibited a high operational stability, 15 reaction cycles were done with almost no loss of activity.

Regarding covalent immobilisation, since the binding forces between the lipase and the supports are strong, obtained biocatalysts tend to show high stability, high resistance to extreme pH and temperature conditions and almost no enzyme leakage. However, these strong links between the enzyme and the support, as well as the harsh conditions employed during immobilisation process, might have a negative impact on the enzyme activity [241,242,243]. In any case, there are several studies that employ this immobilisation technique in biodiesel synthesis. Nematian et al. [213] immobilised proROL onto a superparamagnetic nanostructure and described that amongst the three different biocatalysts studied—two based on lipase-support electrostatic interactions and the third one on covalent-linkage—the covalently immobilised proROL showed higher conversion and operational stability. Bonet-Ragel et al. [181] covalently immobilised rROL onto glutaraldehyde pre-treated epoxide acrylic resins and studied its reaction performance and operational stability in biodiesel synthesis with methanol and ethanol as acyl-acceptors. Under the best conditions, yields close to the theoretical 100% were obtained after 360 min for methanol and 260 min for ethanol. In addition, regarding operational stability, no significant activity loss was observed after five consecutive reaction cycles with both alcohols. Besides, Luna et al. [192] described similar operational stability results with ethanol and sunflower oil as substrates, indicating that covalent immobilisation is an adequate technique for biodiesel synthesis with ROL.

In terms of operational strategy in biodiesel synthesis, although ROL has been described as a suitable industrial and solvent-tolerant enzyme, some improvements have been reported to obtain better reaction yields, higher stability or enhance the scale-up of the bioprocess. One of the most commonly employed approach is based on the stepwise addition of the alcohol as the interaction between the lipase and the alcohol is the main enzyme-deactivating factor [244,245]. Several authors have published works in which ROL and stepwise addition strategy have been employed [181,214]. Additionally, other authors have focused on seeking the most adequate acyl-acceptor—the one that has fewer negative effect on the enzyme—by testing different alcohols [204,240] and even the short-esters of the corresponding alcohols performing interesterification reactions [212,215]. Besides, regarding solvents employment, their absence in solvent-free systems has arisen as an interesting operational alternative because of the minimisation of biodiesel downstream processes and the avoidance of hazardous solvents, making the overall biotransformations more cost-effective and environmentally friendly [170,208,215,218].

Lately, the joint employment of both 1,3-regiospecific and non-specific lipases have been researched in order to accelerate biodiesel reaction rates and obtain higher yields [187]. Lee et al. [216] reported yields close to 100% in 2-h reaction and outstanding operational stabilities when using proROL and Candida rugosa lipase (CRL). Actually, the conversion yield was still 85% after 20 reaction cycles. In line with these results, Zeng et al. [217] described higher biodiesel production rates when employing together proROL and CRL.

Regarding the scale-up of biodiesel production using ROL, Canet et al. [185] compared packed bed reactor (PBR) with stirred tank reactor (STR) in biodiesel synthesis with rROL immobilised through hydrophobic adsorption. Results showed a higher reaction rate with STR than PBR but, just the opposite outcome when operational stability was the analysed trait. Other authors have also employed PBRs [191,197,211] or even more genuine reactors such as the magnetically stabilised fluidised bed reactor [190] or three-phase bioreactors [186]. However, there are not many works related to the scale-up of biodiesel production with ROL considering the vast amount of research papers published dealing with this biocatalyst. Therefore, more research in this field could be relevant for future projects.

4.2. Structured Lipids Production

Fats and oils are consumed in daily diets as an important source of energy, essential fatty acids and fat-soluble nutrients. Their functional, nutritional and organoleptic properties depend on their composition in saturated and polyunsaturated fatty acids, fatty acid chain length and on the distribution of the different fatty acids in the triacylglycerols (TAGs) (position sn-1, sn-3 or sn-2). Therefore, by modifying the fatty acids composition or its profile, lipids with improved properties might be obtained, the so-called structured lipids (SL). Currently, there are various SLs of commercial interest whose properties have been widely described (

Table 6. Definition and properties of the main commercially relevant structured lipids.

SL Type	Definition	Properties	Ref.
Low caloric and dietetic TAGs	Present lower caloric value than conventional oils and fats. Include SLS-, MLM- and MMM- type TAGs.	M and S fatty acids present lower caloric value than their counterparts L. M fatty acids have lower tendency to get accumulated. Released M fatty acids can be directly absorbed and provide readily energy in the liver.	[246,247,248,249]
Human milk fat substitutes (HMFS)	Mimic the fatty acid profile of human milk. Contain oleic (30–35%), palmitic (20–30%), linoleic (7–14%) and stearic acids (5.7–8%). Palmitic acid mainly in sn-2 position.	Promote palmitic acid absorption as 2-monoacylpalmitate Promote calcium absorption	[246,247,250,251]
Cocoa butter equivalents (CBE)	Mimic the scarce natural cocoa butter Mainly formed by saturated fatty acids (stearic and palmitic acids) in sn-1,3 and monounsaturated fatty acids (oleic acid) in sn-2 position.	Desirable polymorph is β form Similar organoleptic properties to cocoa butter	[246,247,252,253]
Trans-free plastic fats	Mimic trans fatty acids containing hydrogenated vegetable oils.	Avoid potential cardiovascular diseases caused by trans fatty acids.	[247,254,255]
TAGs rich in specific long-chain and polyunsaturated fatty acids (PUFAs)	Modified TAGs containing a combination of n-3 and n-6 PUFAs to enhance nutritional values. Mainly eicosapentaenoic (EPA) and docosahexaenoic acid (DHA) are employed.	EPA decreases blood viscosity, platelets aggregation and promotes vasodilation. DHA promotes sensorial and neuronal maturation in babies.	[246,256]
MAGs and DAGs	Modified lipids containing one or two fatty acids linked to a glycerol	Non-ionic surfactants capable of using as emulsifiers in the food industry. 1,3-DAGs reduce serum TAGs level and supress body fat accumulation	[182,184,237,257]

), (i) low caloric and dietetic triacylglycerols that include TAGs with medium-chains (MMM) and TAGs with short- and medium-chain fatty acids in sn-1 and sn-3 and a long-chain fatty acids in sn-2 position, SLS and MLM respectively; (ii) human milk fat substitutes (HMFS), (iii) cocoa butter equivalents (CBE), (iv) trans- free plastic fats, (v) triacylglycerols rich in specific long-chain and polyunsaturated fatty acids (PUFAs) and recently, even (vi) diacylglycerols (DAG) and monoacylglycerols (MAG) have been considered as SLs [246,247].

SLs production can be carried out through chemical or enzymatical processes, the latter having several advantages when compared to chemical catalysis [258]. Hence, in the same way as stated for biodiesel synthesis in the previous section, enzymatically catalysed reactions allow milder reaction conditions which in this case, as well as lowering energy consumption, might lead to a reduction in the loss of original attributes of temperature-sensitive substrates and products. Moreover, through enzymatic catalysis, the use of deleterious solvents can be avoided enabling a safer and more environmentally friendly food production. However, the most remarkable advantage of lipase employment in this biotransformation is their specificity and selectivity [259,260]. Concretely, 1,3-regiospecific lipases like ROL arise a keen interest because of their capacity to only modify the sn-1 and sn-3 positions of TAGs—even though acyl-migration phenomena might occur depending on reaction conditions.

Table 7. Summary of structured lipids production with Rhizopus oryzae lipase as main biotcatalyst.

Product	Substrates	Reaction Type	Lipase	Immobilisation Technique	ID/OS	Ref.
MLM	OO + CRA	Acidolysis	proROL/rROL	CI onto Eupergit®C/sepiolite (AlPO4-sepiolite)	ID: 21.6%. OS: half-life 159 h	[261]
MLM	OO + CA	Acidolysis	proROL/rROL	CI onto Eupergit®C/sepiolite (AlPO4-sepiolite)	ID: 34.82%. OS: half-life 136 h	[261]
MLM	SCG + CA	Acidolysis	proROL	CI onto GA treated MNP	ID: 50%	[262]
MLM	SCG + ethyl caprate	Interesterification	proROL	CI onto GA treated MNP	ID: 26%	[262]
MLM	OP + CA	Acidolysis	proROL	CI onto GA treated MNP	ID: 51% OS: 6.8 batches	[262]
MLM	OP + ethyl caprate	Interesterification	proROL	CI onto GA treated MNP	ID: 46%. OS: 9.1 batches	[262]
MLM	Grapeseed oil + CRA	Acidolysis	rROL	CI onto Amberlite IRA 96	ID: 54%. OS: half-life 166 h	[263]
MLM	Grapeseed oil + CA	Acidolysis	rROL	CI onto Amberlite IRA 96	ID: 69% OS: half-life 118 h	[263]
MLM	TGA58F + CA	Acidolysis	proROL	IA onto Dowex WBA	ID: 64.6%	[264]
MLM	TGA40 + CA	Acidolysis	proROL	IA onto Dowex WBA	ID: 62.8%	[264]
MLM	TGA55E + CA	Acidolysis	proROL	IA onto Dowex WBA	ID: 64.8% OS: 90 days in PBR1 dropped 10%	[264]
MLM	OO + CRA	Acidolysis	rROL	CI onto Eupergit® C/IA onto Lewatit VP OC 1600	OS: half time 2.4 batches (54.3 h) with Eupergit®C	[265]
MLM	OO + CA	Acidolysis	rROL	CI onto Eupergit® C/IA onto Lewatit VP OC 1600	OS: half time 10.2 batches (234 h) with Lewatit VP OC 1600	[265]
MLM	OO + CRA	Acidolysis	rROL	CI onto Eupergit® C	ID: 15.5%	[266]
MLM	OO + CA	Acidolysis	rROL	CI onto Eupergit® C	ID: 33.3%	[266]
MLM	OO + CRA	Acidolysis	rROL	CI onto Amberlite IRA 96	ID: 76.9	[267]
MLM	OO + CA	Acidolysis	rROL	CI onto Amberlite IRA 96	ID: 85.6%	[267]
HMFS	PA enriched TAGs + OA enriched mixtures	Acidolysis	proROL	IA onto Accurel® MP-1000	ID: OA in sn-1,3 67.2% - PA in sn-2 67.8%. OS: no activity loss in 10 uses (190 h)	[268]
HMFS	Lard + FFA from EPAX 1050TG	Acidolysis	rROL	CI onto Accurel® MP-1000	ID: 24 mol%. OS: after 4 batches, 55% of original activity	[269]
HMFS	Tripalmitin + FFA from camelina oil	Acidolysis	rROL	AI onto RelizymeTM OD403/S/CI onto Lewatit VP OC 1600	ID: 52%	[270]
TAGs rich in PUFAs	cod liver + tuna oil + ethanol.	Alcoholysis	proROL	IA onto Accurel® MP-1000	Alcoholysis ID: 72% OS: after 6 cycles, complete deactivation.	[271]
	2-MAG from alcoholysis + CRA	Esterification	proROL	IA onto Accurel® MP-1000	ID: 95%. OS: after 5 cycles, no activity loss.	[271]
TAGs rich in PUFAs	Tuna oil + CRA	Acidolysis	proROL	IA onto Accurel® MP-1000	OS: over one week	[272]
TAGs rich in PUFAs	cod liver oil + ethanol 96%	Alcoholysis	proROL	IA onto Accurel® MP-1000	Alcoholysis Y: 78%. OS: after 3 cycles, a 57% decrease	[273]
	cod liver oil + 1-butanol	Alcoholysis	proROL	IA onto Accurel® MP-1000	Alcoholysis Y: 78%. OS: after 3 cycles, no activity decrease	[273]
	Esterification: 2-MAG from alcoholysis + CRA	Esterification	proROL	IA onto Accurel® MP-1000	Esterification Y: 71%.	[273]
TAGs rich in PUFAs	Fish oil + CRA	Acidolysis	proROL	Non-immobilised	ID: 2.5%	[274]
HMFS	Milkfat + SYO	Interesterification	proROL	EI into polysiloxane-PVA	ID: 8.14%. OS: after 10 batches, no activity loss	[275]
CBE	SO + SA-PA mixtures	Acidolysis	proROL	IA onto Accurel® MP-1000		[276]

shows a summary of the latest published works about SLs synthesis employing ROL. Nunes et al. [261] produced MLM-type SLs by acidolysis of olive oil with capric and caprylic acids. The employed biocatalysts were rROL produced in K. phaffii and commercial native ROL (proROL), both of them covalently immobilised onto Eupergit© C and modified Sepiolite. Noticeably, rROL showed a better performance than the native lipase, the percentages of incorporated capric and caprylic acids were higher as well as the operational stability. In spite of the use of pure or commercial substrates, oleaginous wastes or even non-commercially profitable oils might also be employed for MLM-type SLs synthesis with ROL. For instance, Mota et al. [262] described how low-calorie SLs of MLM-type can be produced using oil extracted from spent coffee grounds and oil from olive pomace with proROL immobilised onto magnetic nanoparticles. In the same line, Costa et al. [263] synthesised MLM-type SLs with the oil extracted from grapeseeds of Vitis vinifera L., which are a by-product of the wine industry. Moreover, instead of residual oils, Nagao et al. [264] employed the oil from the oleaginous microorganism Mortierella alpina to produce MLMs rich in arachidonic acid, a precursor of several hormones.

Regarding HMFS, Esteban et al. [268] used several commercial lipases, amongst them proROL immobilised onto Accurel^® MP-1000, to produce a TAG rich in palmitic acid in sn-2 and oleic acid in sn-1,3; the so called OPO, which is the main component of human milk TAGs. proROL showed the best performance in oleic acid incorporation and exhibited a high operational stability, after ten reuse cycles almost no activity loss was found. Simões et al. [269] also tested different lipases for HMFS production and reported that rROL immobilised onto Accurel^® MP-1000 showed a similar performance to Novozymes 435 and Lipozyme RM IM in acidolysis reaction between lard and FFA mixture from fish oil rich in docosahexaenoic acid. Besides, Faustino et al. [270] immobilised rROL produced in K. phaffii onto two different supports, Lewatit VP OC 1600 and Relizyme OD403/S, and applied the formed biocatalysts in the production of HMFS rich in polyunsaturated fatty acids (PUFAs). The acidolysis reaction was carried out in solvent-free system between tripalmitin and FFAs (mainly linoleic and linolenic acids) from camelina oil, which proved to be a good source of PUFAs. According to the authors, the results obtained with rROL immobilised onto Lewatit VP OC 1600 were comparable to the commonly used commercial lipase Lipozyme RM IM.

Triacylglycerols rich in long-chain and polyunsaturated fatty acids have also been produced with ROL. In most of the cases, these SLs’ production is based on a two-step process in order to minimise the acyl migration phenomena [277]. In the first step, through alcoholysis reaction, 2-monoacylglycerols (2-MAGs) are obtained from oils containing TAGs rich in PUFAs or long-chain fatty acids in the mentioned sn-2 position, usually fish oils. Then, these 2-MAGs are esterified with other relevant FFA to obtain the nutritionally interesting TAGs rich in PUFAs. For instance, Muñio et al. [271] studied the performance of different commercial lipases, including proROL immobilised onto Accurel^® MP-1000, in the process of alcoholysis of tuna and cod oil to obtain 2-MAGs and then, carry out their subsequent esterification with capric acid. In alcoholysis reaction the commercial lipase Novozyme 435 showed a better operational stability than Lipase D (commercial proROL), although the latter exhibited higher reaction yield. During esterification reaction, Lipase D obtained the highest SLs percentage (over 90%) in the mixture. Moreover, no loss in proROL activity was observed after at least five reaction cycles. Hita et al. [272] and Rodriguez et al. [273] reported similar results with immobilised proROL.

With respect to CBE, although Ray et al. [276] described the kinetics of the acidolysis of high oleic sunflower oil with stearic–palmitic acid mixtures that, after further fractionation of the product, could be potentially used in CBE formulations, ROL has not been extensively used for CBE production. Therefore, this subject might be a great research target for future projects, as well as DAG and MAG synthesis, which have not been specifically treated but just as a minor topic during other products synthesis, like biodiesel.

4.3. Flavour Esters Production

Flavour and aromatic esters are widely found in nature and have pleasant organoleptic attributes, including fruity, floral, spicy, creamy or nutty aromas. These traits made them suitable as ingredients for food, beverages, cosmetics, pharmaceuticals, chemicals and personal care products, like perfumes, body lotions, shampoos and other toiletries [278,279]. In general, most of the flavour and fragrance compounds are produced through extraction from their natural source, usually fruits, plants and flowers. However, they are found in the environment in low concentrations making the extraction a costly process and not viable to fulfil their growing demand. Therefore, chemical and enzymatic synthesis procedures have arisen to solve flavour esters scarcity [280,281]. Noticeably, the latter exhibits a significant advantage—notwithstanding the already explained benefits of enzymatic synthesis over chemical one in the previous sections—which is the capacity to label the obtained products as natural according to European Legislation (EC 1334/2008) if and when the employed reactants are also natural. Thereby, the use of enzymes satisfies consumers trend towards natural products and boosts economic value of the obtained flavour esters [278]. In fact, as well as ROL, other lipases have been used for flavour esters production, for instance, the commercial Novozym^® 435 (Candida antartica lipase B) [280,281], Candida rugosa lipase [282,283] and Burkholderia cepacia lipase [284].

Ethyl butyrate is an important component of many fruit flavours such as pineapple, passion fruit and strawberry [285]. The enzymatic synthesis of this compound can be carried out through esterification of butyric acid and ethanol. Guillen et al. [286] immobilised rROL onto three different supports, EP100, Eupergit^®CM and Octadecyl-Sepabeads to test them in this esterification reaction. In terms of reaction rate and yield, rROL immobilised onto EP100 supports showed the best performance. However, rROL immobilised onto Octadecyl-Sepabeads exhibited the highest operational stability. Consequently, this biocatalyst was used for further research in which the effects of butyric acid and ethanol concentration were studied through DoE strategy to maximise the reaction rate and final yield [287]. The obtained results indicated that the suitable acid:alcohol ratio for maximum yield was 1.45 and that the higher the butyric acid concentration the higher the reaction rate. However, as previously described by Grosso et al. [288], elevated concentrations of butyric acid led to enzyme deactivation.

Butyl acetate is another flavour ester with resembling organoleptic properties to pineapple flavour whose production with ROL was reported by Ben Salah et al. [289]. The synthesis of this compound was carried out through esterification reaction of butanol and acetic acid with immobilised proROL onto Celite 545—as preliminary results of the reaction with free enzyme showed poor yield and they were clearly exceeded by the immobilised biocatalyst. According to the authors, solvent-free reaction was chosen as the most suitable strategy because of the easier product purification and lower toxicity and inflammability. In these conditions, a maximum yield of 60% was obtained and the biocatalyst was stable for three consecutive cycles without a decrease in synthesis activity.

Besides esterification, transesterification reaction catalysed by ROL has also been employed for flavour esters synthesis, for example, Kumari et al. [290] reported isoamyl acetate ester synthesis—pleasant banana flavour—through isoamyl alcohol and vinyl acetate transesterification with immobilised proROL. Furthermore, as stated by these authors, the inhibitory effect of the acid [291] was avoided through the use of transesterification reaction with vinyl acetate ester instead of esterification reaction with the corresponding acid. Under optimal conditions, a conversion of 95% in 8 h of reaction was obtained including a great operational stability, after three reaction cycles no activity loss was detected. Garlapati et al. [292] described the use of covalently immobilised proROL onto activated silica to produce through transesterification reactions methyl butyrate and octyl acetate, flavour esters with pineapple and orange odours respectively. As a result of an optimisation process, authors reached high reaction yields in solvent-free system, 70.42% in 14 h and 92.35% in 12 h for methyl butyrate and octyl acetate respectively. Moreover, in both cases, the biocatalyst was reusable for five times retaining a relative activity of more than 95%. Transesterification reaction was as well employed for citronellol esters synthesis with immobilised proROL into HPMC–PVA polymer (hydroxypropyl methyl cellulose—polyvinyl alcohol) and in supercritical carbon dioxide reaction medium [293]. For the three studied flavour esters (citronellol acetate, citronellol butyrate and citronellol laurate) final yields over 90% were achieved indicating the suitability of this biocatalysts and the proposed system for these biotransformations.

4.4. Resolution of Racemic Mixtures

Enantiomerically pure compounds are very attractive for the preparation of a wide range of products, particularly in food and pharmaceutical industries where the desired organoleptic properties or effects might be only related to one of the isomers. Therefore, racemic resolution processes become relevant and arise the interest in lipases considering the enantioselectivity and specificity of these enzymes [294,295].

Palomo et al. [296] employed proROL to carry out the enzymatic resolution of (R)-glycidyl butyrate because of its importance in linezolid synthesis. This product is already sold as a treatment for multidrug resistant Gram-positive infections. According to these authors, they followed the ‘conformational engineering’ strategy, that is, different techniques for proROL immobilisation were employed. This way, the enzyme structure would have different rigidity or the microenvironment surrounding the enzyme would alter the exact shape of the open form of the lipase influencing its catalytic performance. Amongst the three different biocatalysts formed, the best enantiomeric excess (ee) was obtained with proROL immobilised through adsorption on dextran sulphate-coated sepabeads, 99% ee with a 55% conversion.

Benzoin is a relevant α-hydroxy ketone that might act as building block in organic synthesis. Songür et al. [297] described its enantioselective production from benzoin acetate through the employment of R. oryzae cell homogenates (Figure 6). The objective of using cell homogenates was to combine the enantioselective hydrolysis of proROL with the racemisation process of the racemase of R. oryzae in order to increase the ee and conversion values. This way, a final conversion of (S)-benzoin close to the 100% and 96% ee was achieved.

Covalently immobilised proROL onto Lewatit-aldehyde support has been reported as an adequate biocatalyst for asymmetric hydrolysis of dimethyl 3-phenylglutarate [298]. Under the best conditions, it was possible to obtain the (R)-methyl-3-phenylglutarate with a 92% ee and an yield in monoester of 97%.

(S)-enantiomer of ibuprofen is 160 more active than its (R)-enantiomer, which can even cause side effects in the gastrointestinal tract. Therefore, obtaining the adequate enantiomer becomes crucial in this case. Yousefi et al. [299] reported the use of immobilised proROL onto octadecyl sepharose to carry out the enantioselective resolution of racemic ibuprofens esters (Figure 7).

The racemic resolution of (R,S)-1-phenylethanol to produce (S)-1-phenylethanol, a chiral building block, was carried out with proROL-displaying yeast whole cell biocatalyst, that is, a S. cerevisiae strain genetically modified to display proROL on the cell surface. After 36 h of reaction, significant results were obtained, 97.3% yield and 93.3% ee [300]. The same biocatalyst was employed to catalyse the optical resolution of the pharmaceutical precursor (R,S)-1-benzyloxy-3-chloro-2-propyl monosuccinate. In this case, the operational stability of the biocatalysts was assessed and it was stable after at least eight reaction cycles [301].