Production of Pure Lipases

Production of Pure Lipases Most Rhizopus lipases described in the literature are produced by their host cells in very small amounts, which is not suitable for large-scale industrial applications. Another drawback of working with native microorganisms is that Rhizopus sp. usually produces different isolipases with different biocatalytic properties, which complicates biocatalytic application with unwanted side effects caused by unrelated enzymes (67). Thus, in order to implement more industrial lipase applications, one of the most promising strategies is to produce heterologous lipases in a host microorganism which have the advantages to minimize the complexity of the bioprocess, to facilitate the more economic production of large quantities of pure lipases and to reduce costs (68). Rhizopus species was mainly divided into three groups, including R. oryzae, R. microsporus, R. stolonifer (69). Until now, the lipase genes from all these three groups have been cloned. Consistent with the proposal that many Rhizopus isolates are actually the same organism (69), nearly complete identities were observed in the lipase genes from the genus Rhizopus. In the R. oryzae group, lipases from R. niveus IFO 9759 (70) (71) (72), R. delemar ATCC34612 (73) (74) and R. javanicus (75) have identical amino acid sequences, and the lipases from R. oryzae ATCC 853 (76, 77) (78) and R. arrhizus L-03-R-1(79) differ by only several substitutions. The lipase from R. stolonifer (GenBank No. AAZ66864.1) and R. chinensis (GenBank No. ABN59381.2)(80) were of about 80% amino acid sequence identity to R. oryzae lipase (GenBank No. BAG16821.1) (Fig). Rhizopus lipases are synthesized as pre-pro-proteins composed of signal sequence, prosequence and mature lipase sequence (76) (80). The production of Rhizopus lipases has been performed in Escherichia coli [8], in Saccharomyces cerevisiae [9] and in Pichia pastoris [10,11]. Due to either different protein post-translational modifications affecting the glycosylation pattern of these lipases, or to proteolytic cleavage products arising from the mature, the pro- or the pre-pro-enzyme, the lipases derived from recombinant strains showed different molecular mass and different enzymatic properties compared with those from wild strains (75) (81). E. coli as host When E. coli was selected as a host, overexpression of R. oryzae prolipase (ProROL) and mature R. oryzae lipase (mROL) was achieved by using E. coli BL321, the heat-inducible vector pCYTEXP1 and the preOmpA sequence to achieve secretion into the periplasmic space (77) (76). Both fusion proteins induced by a temperature shift to 42  °C were correctly processed and translocated to the periplasmic space, but both mROL and ProROL accumulated as insoluble aggregates. Furthermore, the enzyme activity of mROL damages cells even at low concentrations, whereas the expression of active ProROL has no toxic effects on E. coli and the amount of purified active ProROL (103 U/mL) is larger than that of mROL by a factor of more than 100(77). By in vitro refolding of the insoluble form of mROL, its prosequence was supposed to support the refolding of mROL in vitro (77). R. delemar lipase was also expressed as inactive and insoluble forms in the cytoplasm of E. coli BL21(DE3) using plasmid vector pE T11d (74). In order to achieve soluble production in E. coli, Di Lorenzo et al. (82) used a reductase-deficient host strain, E. coli Origami(DE3), which greatly enhance disulfide bond formation in the E. coli cytoplasm, and pET-11d as expression systems, yielded correctly folded and active mROL and ProROL only in the cytoplasm fraction. Although the specific activities of mROL and ProROL are comparable, the yield of ProROL (166U/mL) was higher than that of mROL (82). S. cerevisiae as host In contrast to E. coli expression system, eukaryotic expression system (eg. P. pastoris, S. cerevisiae) has the advantage of secretion into the cell culture supernatant, avoiding cell disruption, and post-translational modifications, including glycosylation and disulfide bond formation. Takahashi et al. (78) reported the production of the active R. oryzae lipase (ROL) by S. cerevisiae. When the ProROL gene fused with S. cerevisiae ÃŽ ±-factor signal sequence was expressed in S. cerevisiae using UPR-ICL as a promoter, two forms of ROL, ProROL (46 kDa) and r28ROL (35 kDa) having 28 amino acids of the prosequence, were secreted into the culture medium, but the activity of mROL was not detected in both the medium and cells (78). The N-terminal amino acid sequence analysis revealed that the processed form r28ROL was cleaved by Kex2-like protease in S. cerevisiae at the recognition site KR in the prosequence (78). The independent production of two different molecular forms of ROL was succe ssfully obtained using Kex2-engineered strains of S. cerevisiae (83). The amounts of enzyme secreted were estimated as 78.8 mg/l broth for ProROL and 38.3 mg/l broth for r28ROL. These two forms of ROL exhibited distinct properties. ProROL were more thermostable than the processed lipase r28ROL (83), which was also observed in the case of R. niveus lipase expressed in S. cerevisiae (84). The expressed R. niveus lipase activity in the culture supernatant reached 1600 U/mL, a value equivalent to 0.2~0.3 g lipase protein per 1 liter of culture supernatant (72). The role of the prosequence of ROL was further analyzed in S. cerevisiae. When mROL was co-expressed in trans with the prosequence fused to the pre-alpha-factor leader sequence, the activity of mROL was recovered, suggested that the covalent linkage of the prosequence to the mROL was not necessary for the function (85) (86) . From the expression of the ROL mutants with deletions at the N-terminal end of the prosequence together w ith mROL in trans, the residues from 20 to 37 in the prosequence were essential for the secretion, and those from 38 to 57 were essential for the formation of the active ROL and might play a role as an intramolecular chaperone (86). Apart from extracellular expression, R. oryzae lipase was also successfully displayed on the cell surface of S. cerevisiae with a lipase activity of 350.6 U/L (87-89). P. pastoris as host On the contrary to S. cerevisiae expression system when the prolipase gene from R. arrhizus was expressed in P. pastoris, only the mature lipase attached with 28 amino acids of the carboxy-terminal part of the prosequence was secreted in the supernatant (Niu, Li et al. 2006). Prolipase with complete prosequence was nearly found in the supernatant which may be the result that the activity of Kex2-like protease was higher in P. pastoris than in S. cerevisiae (Niu, Li et al. 2006). Interestingly, expression of the prolipase gene from R. chinensis in P. pastors produced two forms of lipase, mature lipase attached with 27 amino acids of the prosequence (r27RCL) and mature lipase (mRCL) (Yu, Wang et al. 2009). At the early stage of fermentation, the recombinant P. pastors produced only r27RCL, which was gradually degraded into mRCL along with the fermentation, probably processed by serine proteases presented in the culture medium (Yu, Wang et al. 2009). For R. oryzae lipase, r28ROL was als o observed to convert to mature lipase gradually incubated at 0-6  °C for a few days (Sayari, Frikha et al. 2005, Hama, Tamalampudi et al. 2006). It has been reported that the prosequence from R. chinensis lipases was more efficient than that from ROL in the role of facilitating the folding and secretion of an active lipase (Yu, Sha et al. 2013). The activity of the chimeric lipase from R. oryzae by replacing the prosequence with that from R. chinensis reached 4050 U/mL, which was 11 fold higher than that of the parent expressed in P. pastoris (Yu, Sha et al. 2013). mROL activity expressed in S. cerevisiae was not detected (Ueda, Takahashi et al. 2002), whereas mROL was functionally expressed and secreted in P. pastoris (Beer, McCarthy et al. 1998) (Minning, Schmidt-Dannert et al. 1998). The expression of mROL was compared under different regulated promoters (Serrano 2002) (Cos, Resina et al. 2005) (Resina, Serrano et al. 2004), i.e. alcohol oxidase 1 promoter (PAOX), glyceraldehyde-3-phosphate dehydrogenase promoter (PGAP) and the formaldehyde dehydrogenase 1 promoter (PFLD1). Expression of the mature form of ROL under the constitutive PGAP promoter resulted in extremely low extracellular lipase levels (Serrano 2002). PAOX is highly induced during growth on methanol, but tightly repressed during growth of the yeast on most common carbon sources such as glucose, glycerol or ethanol, while PFLD1 allows the design of methanol-free culture strategies, being methylamine a less volatile and flammable inducer. Valero and co-authors (Resina, Cos et al. 2005) (Cos, Resina et al. 2005) successfully expressed mROL in P. pastoris under the PFLD1 promoter using methylamine as nitrogen and sorbitol as carbon source. The productivity under PFLD1 promoter are comparable to the classic PAOX system (Resina, Cos et al. 2005) (Cos, Resina et al. 2005). However, limitationsincurrent knowledgein the regulation of methylamine metabolism in P. pastoris limit thedesignof efficientfermentationstrategies using PFLD1 promoter. The presence of multiple integrated copies of a desired expression cassette has been reported to be an important factor in increasing foreign protein production in P. pastoris (Cos, Serrano et al. 2005). The high-level expression of R. chinensis lipase was achieved by optimization of the lipase gene copy number in the host strain P. pastoris (Sha, Yu et al. 2013). Among 1, 3, 5 and 6-copy strains, the maximum lipase activity reached 12,500 U/mL in 5-cope strain, which was 6.2-fold higher than that in 1-copy strain (Sha, Yu et al. 2013). Maximizing gene copy number sometimes resulted in a decreased final productivity yield (Sha, Yu et al. 2013) or a negative effect on cell growth (Cos, Serrano et al. 2005), which indicated stress effects of Rhizopus lipases overexpression upon P. pastoris cells. Recent transcriptomic-based studies (Sha, Yu et al. 2013) (Resina, Bollok et al. 2007) strongly suggest that heterologous overexpression of Rhizopus lipases in P. pastoris indeed triggers the unfolded protein response (UPR). However, co-expression of certain chaperons could relief the cell stress caused by foreign protein overexpression. The productivity of R. chinensis lipase in P. pastoris was improved by co-expression with ERO1p and PDI (Kumar, Jahan et al. 2013) (Sha, Yu et al. 2013). These two chaperons are involved in the protein disulfide bond formation pathway in the endoplasmic reticulum, thus chaperon co-expression contribute to the relief of the protein overexpression stress in recombinant P. pastoris (Kumar, Jahan et al. 2013) (Sha, Yu et al. 2013). Co-expression of unfolded protein response transcriptional factor (Hac1) resulted in about a 3-fold increase in the overall specific productivity of mROL in P. pastoris while the deletion of GAS1, required for the cell wall assembly, achieved 4-fold increase. And the double mutant HAC1/Δgas1 strain yielded about a 7-fold increase (Resina, Maurer et al. 2009). Novel genetic factors enhancing heterologous protei n secretion in P. pastoris have been identified on the basis of transcriptomic analyses (Gasser, Sauer et al. 2007), which give a new insight into engineering P. pastoris for heterologous expression of foreign proteins. Another factor that should be considered for expression optimization is the Mut phenotype. When single copy mROL was expressed in P. pastoris under control of PAOX, the Maximal lipase activity (205U/mL), the specific production rate (qp, 63 U/gbiomass/h ) and the yield of lipase activity per biomass unit (YP/X, 5775U/gbiomass) by Muts strain were 1.37-, 1.29- and 2.34-fold higher than those by Mut+ strain, while the productivity (2246 U/L/h) by Muts strain was 1.34-fold lower than that by Mut+ strain (Cos, Serrano et al. 2005). For R. chinensis prolipase expressed in P. pastoris, Muts strain exibited 1.4-fold higher activity than that of Mut+ phenotype (Yu, Wang et al. 2009). Muts is not as sensitive as Mut+ to high transient methanol concentrations which make the bioprocess easier to control and enables scale up. However, Muts phenotype fermentation results in long induction times with low growth rates (Arnau, Casas et al. 2011). Different strategies have been implemented in the production of R. oryzae lipase under the PAOX promoter in P. pastoris Mut+ phenotype, eg. dissolved oxygen control (DO-stat), methanol limited fed-batch (MLFB, control of the substrate concentration close to zero), methanol non-limited fed-batch (MNLFB, maintenance of the substrate concentration at a constant value), as well as temperature limited fed-batch (TLFB). The MNLFB strategy for ROL expression led to 20–30 h reduction in the production time, a 11-fold higher final lipolytic activity, a 13.6-fold higher productivity and a 10.3-fold higher specific productivity compared to the DO-stat strategy (Minning, Serrano et al. 2001). After optimization of methanol concentration, the activity of Rhizopus chinensis lipase reached 2130 U/mL by maintaining methanol concentration at 1 g/L controlled by an on-line methanol analyzer (Wu, Yu et al. 2011). One of the main drawbacks of P. pastoris Mut+ phenotype is the high oxygen demand i n high cell density cultures using MNLFB strategy (Cos, Serrano et al. 2005). A TLFB strategy was thereafter applied to solve oxygen transfer limitations (Surribas, Stahn et al. 2007). Barrigà ³n et al. (Manuel Barrigon, Luis Montesinos et al. 2013) compared MLFB and MNLFB strategies for the production of mROL expressed in P. pastoris. Yields, productivities and specific production rate in all MLFB conditions were very low. Best results were obtained by MNLFB strategy at methanol set-point of 3 g/L, under which condition maximum lipase activity, productivity and specific productivity reached 280 U/mL, 5406 U/L/h, 102 U/g/h, respectively (Manuel Barrigon, Luis Montesinos et al. 2013) . Due to low specific growth rate of Muts phenotype, strategies to use mixed carbon sources in the induction phase are applied to improve the expression levels of mROL and to increase the productivity of the bioprocess. Glycerol and sorbitol are both frequently used as co-substrates. When glycerol was used as co-substrate, one of the key advantages is its higher ÃŽ ¼ (0.18 h−1 versus 0.02 h−1) compared with sorbitol. However, when the relation ÃŽ ¼Gly per ÃŽ ¼MeOH was larger than 4, a significant decrease of YP/X, volumetric and specific productivity was observed due to the repression of glycerol (Arnau, Casas et al. 2011). The sorbitol as an excellent non-repressive carbon source avoids the severe decrease of the specific production rate and it permits to achieve higher mROL production (Arnau, Ramon et al. 2010) (Ramon, Ferrer et al. 2007). During the induction stage, the ammonium ion released into the fermentation broth has a deep impact on cell growth and protein expression. In our previous study, the impact of NH4+ concentration on the expression of the R. oryzae chimeric lipase in P. pastoris co-expressed with ERO1p and PDI was investigated (Yu, Lu et al. 2013). Analysis of carbon metabolism and energy regeneration pattern revealed that under the optimum NH4+ concentration of 440 mmol/L more carbon source (methanol) was consumed with surged AOX activity and then the higher energy and amino acid precursors demand for recombinant protein synthesis is compensated for by the TCA cycle. Under the optimum NH4+ concentration the lipase activity reached 12,019 U/mL, which was 1.7-fold higher than that without supplement of NH4+ (Yu, Lu et al. 2013). It was also useful for other recombinant protein expression in P. pastoris by addition of inorganic nitrogen source (Kobayashi, Kuwae et al. 2000) (Zhang, Wang et al. 2008) (Yan g and Zhou 2004).